From 025adbe8e58290798001b472aec3eb618d8fc930 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 5 May 2008 01:20:04 -0700 Subject: tipc: Simplify log buffer resizing This patch simplifies & standardizes the way TIPC's print buffer log is resized. Code to terminate use of the log buffer is eliminated by simply setting the log buffer size to 0 bytes. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/config.c | 2 +- net/tipc/core.c | 4 ++-- net/tipc/dbg.c | 26 +++++++------------------- net/tipc/dbg.h | 6 +++--- 4 files changed, 13 insertions(+), 25 deletions(-) (limited to 'net') diff --git a/net/tipc/config.c b/net/tipc/config.c index c71337a22d33..91d56f8fee9f 100644 --- a/net/tipc/config.c +++ b/net/tipc/config.c @@ -529,7 +529,7 @@ struct sk_buff *tipc_cfg_do_cmd(u32 orig_node, u16 cmd, const void *request_area break; #endif case TIPC_CMD_SET_LOG_SIZE: - rep_tlv_buf = tipc_log_resize(req_tlv_area, req_tlv_space); + rep_tlv_buf = tipc_log_resize_cmd(req_tlv_area, req_tlv_space); break; case TIPC_CMD_DUMP_LOG: rep_tlv_buf = tipc_log_dump(); diff --git a/net/tipc/core.c b/net/tipc/core.c index 740aac5cdfb6..862d41541615 100644 --- a/net/tipc/core.c +++ b/net/tipc/core.c @@ -182,7 +182,7 @@ static int __init tipc_init(void) { int res; - tipc_log_reinit(CONFIG_TIPC_LOG); + tipc_log_resize(CONFIG_TIPC_LOG); info("Activated (version " TIPC_MOD_VER " compiled " __DATE__ " " __TIME__ ")\n"); @@ -209,7 +209,7 @@ static void __exit tipc_exit(void) tipc_core_stop_net(); tipc_core_stop(); info("Deactivated\n"); - tipc_log_stop(); + tipc_log_resize(0); } module_init(tipc_init); diff --git a/net/tipc/dbg.c b/net/tipc/dbg.c index e809d2a2ce06..1a8d0a24d9e3 100644 --- a/net/tipc/dbg.c +++ b/net/tipc/dbg.c @@ -350,43 +350,31 @@ void tipc_dump(struct print_buf *pb, const char *fmt, ...) } /** - * tipc_log_stop - free up TIPC log print buffer + * tipc_log_resize - change the size of the TIPC log buffer + * @log_size: print buffer size to use */ -void tipc_log_stop(void) +void tipc_log_resize(int log_size) { spin_lock_bh(&print_lock); if (TIPC_LOG->buf) { kfree(TIPC_LOG->buf); TIPC_LOG->buf = NULL; } - spin_unlock_bh(&print_lock); -} - -/** - * tipc_log_reinit - (re)initialize TIPC log print buffer - * @log_size: print buffer size to use - */ - -void tipc_log_reinit(int log_size) -{ - tipc_log_stop(); - if (log_size) { if (log_size < TIPC_PB_MIN_SIZE) log_size = TIPC_PB_MIN_SIZE; - spin_lock_bh(&print_lock); tipc_printbuf_init(TIPC_LOG, kmalloc(log_size, GFP_ATOMIC), log_size); - spin_unlock_bh(&print_lock); } + spin_unlock_bh(&print_lock); } /** - * tipc_log_resize - reconfigure size of TIPC log buffer + * tipc_log_resize_cmd - reconfigure size of TIPC log buffer */ -struct sk_buff *tipc_log_resize(const void *req_tlv_area, int req_tlv_space) +struct sk_buff *tipc_log_resize_cmd(const void *req_tlv_area, int req_tlv_space) { u32 value; @@ -397,7 +385,7 @@ struct sk_buff *tipc_log_resize(const void *req_tlv_area, int req_tlv_space) if (value != delimit(value, 0, 32768)) return tipc_cfg_reply_error_string(TIPC_CFG_INVALID_VALUE " (log size must be 0-32768)"); - tipc_log_reinit(value); + tipc_log_resize(value); return tipc_cfg_reply_none(); } diff --git a/net/tipc/dbg.h b/net/tipc/dbg.h index c01b085000e0..718a5db245b5 100644 --- a/net/tipc/dbg.h +++ b/net/tipc/dbg.h @@ -61,10 +61,10 @@ int tipc_printbuf_empty(struct print_buf *pb); int tipc_printbuf_validate(struct print_buf *pb); void tipc_printbuf_move(struct print_buf *pb_to, struct print_buf *pb_from); -void tipc_log_reinit(int log_size); -void tipc_log_stop(void); +void tipc_log_resize(int log_size); -struct sk_buff *tipc_log_resize(const void *req_tlv_area, int req_tlv_space); +struct sk_buff *tipc_log_resize_cmd(const void *req_tlv_area, + int req_tlv_space); struct sk_buff *tipc_log_dump(void); #endif -- cgit v1.2.2 From fb98ec71c7f81b6db9b793aeb9d53823b6960d8b Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 5 May 2008 01:20:42 -0700 Subject: tipc: Provide feedback when log buffer resizing fails This patch provides feedback to the user when TIPC is unable to set its log buffer to the requested size. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/dbg.c | 13 ++++++++++--- net/tipc/dbg.h | 4 ++-- 2 files changed, 12 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/tipc/dbg.c b/net/tipc/dbg.c index 1a8d0a24d9e3..c6806910598c 100644 --- a/net/tipc/dbg.c +++ b/net/tipc/dbg.c @@ -2,7 +2,7 @@ * net/tipc/dbg.c: TIPC print buffer routines for debugging * * Copyright (c) 1996-2006, Ericsson AB - * Copyright (c) 2005-2006, Wind River Systems + * Copyright (c) 2005-2007, Wind River Systems * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -354,8 +354,10 @@ void tipc_dump(struct print_buf *pb, const char *fmt, ...) * @log_size: print buffer size to use */ -void tipc_log_resize(int log_size) +int tipc_log_resize(int log_size) { + int res = 0; + spin_lock_bh(&print_lock); if (TIPC_LOG->buf) { kfree(TIPC_LOG->buf); @@ -366,8 +368,11 @@ void tipc_log_resize(int log_size) log_size = TIPC_PB_MIN_SIZE; tipc_printbuf_init(TIPC_LOG, kmalloc(log_size, GFP_ATOMIC), log_size); + res = !TIPC_LOG->buf; } spin_unlock_bh(&print_lock); + + return res; } /** @@ -385,7 +390,9 @@ struct sk_buff *tipc_log_resize_cmd(const void *req_tlv_area, int req_tlv_space) if (value != delimit(value, 0, 32768)) return tipc_cfg_reply_error_string(TIPC_CFG_INVALID_VALUE " (log size must be 0-32768)"); - tipc_log_resize(value); + if (tipc_log_resize(value)) + return tipc_cfg_reply_error_string( + "unable to create specified log (log size is now 0)"); return tipc_cfg_reply_none(); } diff --git a/net/tipc/dbg.h b/net/tipc/dbg.h index 718a5db245b5..6b00062bf040 100644 --- a/net/tipc/dbg.h +++ b/net/tipc/dbg.h @@ -2,7 +2,7 @@ * net/tipc/dbg.h: Include file for TIPC print buffer routines * * Copyright (c) 1997-2006, Ericsson AB - * Copyright (c) 2005-2006, Wind River Systems + * Copyright (c) 2005-2007, Wind River Systems * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -61,7 +61,7 @@ int tipc_printbuf_empty(struct print_buf *pb); int tipc_printbuf_validate(struct print_buf *pb); void tipc_printbuf_move(struct print_buf *pb_to, struct print_buf *pb_from); -void tipc_log_resize(int log_size); +int tipc_log_resize(int log_size); struct sk_buff *tipc_log_resize_cmd(const void *req_tlv_area, int req_tlv_space); -- cgit v1.2.2 From 93758c3da19e99f5377cc1413c27320882b18f4b Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 5 May 2008 01:21:12 -0700 Subject: tipc: Fix recursive spinlock invocation in print buffer code This patch fixes two routines that allow the global print buffer spinlock to be taken recursively. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/dbg.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/tipc/dbg.c b/net/tipc/dbg.c index c6806910598c..834319e3b7e7 100644 --- a/net/tipc/dbg.c +++ b/net/tipc/dbg.c @@ -179,8 +179,10 @@ void tipc_printbuf_move(struct print_buf *pb_to, struct print_buf *pb_from) } if (pb_to->size < pb_from->size) { - tipc_printbuf_reset(pb_to); - tipc_printf(pb_to, "*** PRINT BUFFER MOVE ERROR ***"); + strcpy(pb_to->buf, "*** PRINT BUFFER MOVE ERROR ***"); + pb_to->buf[pb_to->size - 1] = ~0; + pb_to->crs = strchr(pb_to->buf, 0); + pb_to->next = NULL; return; } @@ -405,27 +407,32 @@ struct sk_buff *tipc_log_dump(void) struct sk_buff *reply; spin_lock_bh(&print_lock); - if (!TIPC_LOG->buf) + if (!TIPC_LOG->buf) { + spin_unlock_bh(&print_lock); reply = tipc_cfg_reply_ultra_string("log not activated\n"); - else if (tipc_printbuf_empty(TIPC_LOG)) + } else if (tipc_printbuf_empty(TIPC_LOG)) { + spin_unlock_bh(&print_lock); reply = tipc_cfg_reply_ultra_string("log is empty\n"); + } else { struct tlv_desc *rep_tlv; struct print_buf pb; int str_len; str_len = min(TIPC_LOG->size, 32768u); + spin_unlock_bh(&print_lock); reply = tipc_cfg_reply_alloc(TLV_SPACE(str_len)); if (reply) { rep_tlv = (struct tlv_desc *)reply->data; tipc_printbuf_init(&pb, TLV_DATA(rep_tlv), str_len); + spin_lock_bh(&print_lock); tipc_printbuf_move(&pb, TIPC_LOG); + spin_unlock_bh(&print_lock); str_len = strlen(TLV_DATA(rep_tlv)) + 1; skb_put(reply, TLV_SPACE(str_len)); TLV_SET(rep_tlv, TIPC_TLV_ULTRA_STRING, NULL, str_len); } } - spin_unlock_bh(&print_lock); return reply; } -- cgit v1.2.2 From 40dbfae440abe6860167f12e0296bd7a1a599839 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 5 May 2008 01:21:54 -0700 Subject: tipc: Fix null pointer dereference in debug code This patch eliminates an obsolete use of the DBG_OUTPUT print buffer which could lead to a null pointer crash in tipc_printf() if TIPC's debugging capabilities are configured. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/name_distr.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'net') diff --git a/net/tipc/name_distr.c b/net/tipc/name_distr.c index 39fd1619febf..aecba5cd87d6 100644 --- a/net/tipc/name_distr.c +++ b/net/tipc/name_distr.c @@ -41,9 +41,6 @@ #include "msg.h" #include "name_distr.h" -#undef DBG_OUTPUT -#define DBG_OUTPUT NULL - #define ITEM_SIZE sizeof(struct distr_item) /** -- cgit v1.2.2 From c89039850bdf8047472b4ee6132048dacef2cf5a Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 5 May 2008 01:22:30 -0700 Subject: tipc: Elimination of print buffer chaining This patch revamps TIPC's print buffer subsystem to eliminate support for arbitrary chains of print buffers, which were rarely needed and difficult to use safely. In its place, print buffers can now be configured to echo their output to the system console. This provides an equivalent for the only chaining currently utilized by TIPC, in a faster and more compact manner. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/core.h | 7 +-- net/tipc/dbg.c | 163 +++++++++++++++++++++++++------------------------------- net/tipc/dbg.h | 4 +- 3 files changed, 80 insertions(+), 94 deletions(-) (limited to 'net') diff --git a/net/tipc/core.h b/net/tipc/core.h index 325404fd4eb5..7042ef377262 100644 --- a/net/tipc/core.h +++ b/net/tipc/core.h @@ -66,7 +66,6 @@ struct tipc_msg; extern struct print_buf *TIPC_NULL, *TIPC_CONS, *TIPC_LOG; -extern struct print_buf *TIPC_TEE(struct print_buf *, struct print_buf *); void tipc_msg_print(struct print_buf*,struct tipc_msg *,const char*); void tipc_printf(struct print_buf *, const char *fmt, ...); void tipc_dump(struct print_buf*,const char *fmt, ...); @@ -98,11 +97,13 @@ void tipc_dump(struct print_buf*,const char *fmt, ...); * TIPC_CONS : system console * TIPC_LOG : TIPC log buffer * &buf : user-defined buffer (struct print_buf *) - * TIPC_TEE(&buf_a,&buf_b) : list of buffers (eg. TIPC_TEE(TIPC_CONS,TIPC_LOG)) + * + * Note: TIPC_LOG is configured to echo its output to the system console; + * user-defined buffers can be configured to do the same thing. */ #ifndef TIPC_OUTPUT -#define TIPC_OUTPUT TIPC_TEE(TIPC_CONS,TIPC_LOG) +#define TIPC_OUTPUT TIPC_LOG #endif #ifndef DBG_OUTPUT diff --git a/net/tipc/dbg.c b/net/tipc/dbg.c index 834319e3b7e7..8ca9457250be 100644 --- a/net/tipc/dbg.c +++ b/net/tipc/dbg.c @@ -38,18 +38,44 @@ #include "config.h" #include "dbg.h" -static char print_string[TIPC_PB_MAX_STR]; -static DEFINE_SPINLOCK(print_lock); +/* + * TIPC pre-defines the following print buffers: + * + * TIPC_NULL : null buffer (i.e. print nowhere) + * TIPC_CONS : system console + * TIPC_LOG : TIPC log buffer + * + * Additional user-defined print buffers are also permitted. + */ -static struct print_buf null_buf = { NULL, 0, NULL, NULL }; +static struct print_buf null_buf = { NULL, 0, NULL, 0 }; struct print_buf *TIPC_NULL = &null_buf; -static struct print_buf cons_buf = { NULL, 0, NULL, NULL }; +static struct print_buf cons_buf = { NULL, 0, NULL, 1 }; struct print_buf *TIPC_CONS = &cons_buf; -static struct print_buf log_buf = { NULL, 0, NULL, NULL }; +static struct print_buf log_buf = { NULL, 0, NULL, 1 }; struct print_buf *TIPC_LOG = &log_buf; +/* + * Locking policy when using print buffers. + * + * 1) tipc_printf() uses 'print_lock' to protect against concurrent access to + * 'print_string' when writing to a print buffer. This also protects against + * concurrent writes to the print buffer being written to. + * + * 2) tipc_dump() and tipc_log_XXX() leverage the aforementioned + * use of 'print_lock' to protect against all types of concurrent operations + * on their associated print buffer (not just write operations). + * + * Note: All routines of the form tipc_printbuf_XXX() are lock-free, and rely + * on the caller to prevent simultaneous use of the print buffer(s) being + * manipulated. + */ + +static char print_string[TIPC_PB_MAX_STR]; +static DEFINE_SPINLOCK(print_lock); + #define FORMAT(PTR,LEN,FMT) \ {\ @@ -60,27 +86,14 @@ struct print_buf *TIPC_LOG = &log_buf; *(PTR + LEN) = '\0';\ } -/* - * Locking policy when using print buffers. - * - * The following routines use 'print_lock' for protection: - * 1) tipc_printf() - to protect its print buffer(s) and 'print_string' - * 2) TIPC_TEE() - to protect its print buffer(s) - * 3) tipc_dump() - to protect its print buffer(s) and 'print_string' - * 4) tipc_log_XXX() - to protect TIPC_LOG - * - * All routines of the form tipc_printbuf_XXX() rely on the caller to prevent - * simultaneous use of the print buffer(s) being manipulated. - */ - /** * tipc_printbuf_init - initialize print buffer to empty * @pb: pointer to print buffer structure * @raw: pointer to character array used by print buffer * @size: size of character array * - * Makes the print buffer a null device that discards anything written to it - * if the character array is too small (or absent). + * Note: If the character array is too small (or absent), the print buffer + * becomes a null device that discards anything written to it. */ void tipc_printbuf_init(struct print_buf *pb, char *raw, u32 size) @@ -88,7 +101,7 @@ void tipc_printbuf_init(struct print_buf *pb, char *raw, u32 size) pb->buf = raw; pb->crs = raw; pb->size = size; - pb->next = NULL; + pb->echo = 0; if (size < TIPC_PB_MIN_SIZE) { pb->buf = NULL; @@ -105,7 +118,11 @@ void tipc_printbuf_init(struct print_buf *pb, char *raw, u32 size) void tipc_printbuf_reset(struct print_buf *pb) { - tipc_printbuf_init(pb, pb->buf, pb->size); + if (pb->buf != NULL) { + pb->crs = pb->buf; + pb->buf[0] = 0; + pb->buf[pb->size - 1] = ~0; + } } /** @@ -182,7 +199,6 @@ void tipc_printbuf_move(struct print_buf *pb_to, struct print_buf *pb_from) strcpy(pb_to->buf, "*** PRINT BUFFER MOVE ERROR ***"); pb_to->buf[pb_to->size - 1] = ~0; pb_to->crs = strchr(pb_to->buf, 0); - pb_to->next = NULL; return; } @@ -205,8 +221,8 @@ void tipc_printbuf_move(struct print_buf *pb_to, struct print_buf *pb_from) } /** - * tipc_printf - append formatted output to print buffer chain - * @pb: pointer to chain of print buffers (may be NULL) + * tipc_printf - append formatted output to print buffer + * @pb: pointer to print buffer * @fmt: formatted info to be printed */ @@ -215,66 +231,36 @@ void tipc_printf(struct print_buf *pb, const char *fmt, ...) int chars_to_add; int chars_left; char save_char; - struct print_buf *pb_next; spin_lock_bh(&print_lock); + FORMAT(print_string, chars_to_add, fmt); if (chars_to_add >= TIPC_PB_MAX_STR) strcpy(print_string, "*** PRINT BUFFER STRING TOO LONG ***"); - while (pb) { - if (pb == TIPC_CONS) - printk(print_string); - else if (pb->buf) { - chars_left = pb->buf + pb->size - pb->crs - 1; - if (chars_to_add <= chars_left) { - strcpy(pb->crs, print_string); - pb->crs += chars_to_add; - } else if (chars_to_add >= (pb->size - 1)) { - strcpy(pb->buf, print_string + chars_to_add + 1 - - pb->size); - pb->crs = pb->buf + pb->size - 1; - } else { - strcpy(pb->buf, print_string + chars_left); - save_char = print_string[chars_left]; - print_string[chars_left] = 0; - strcpy(pb->crs, print_string); - print_string[chars_left] = save_char; - pb->crs = pb->buf + chars_to_add - chars_left; - } + if (pb->buf) { + chars_left = pb->buf + pb->size - pb->crs - 1; + if (chars_to_add <= chars_left) { + strcpy(pb->crs, print_string); + pb->crs += chars_to_add; + } else if (chars_to_add >= (pb->size - 1)) { + strcpy(pb->buf, print_string + chars_to_add + 1 + - pb->size); + pb->crs = pb->buf + pb->size - 1; + } else { + strcpy(pb->buf, print_string + chars_left); + save_char = print_string[chars_left]; + print_string[chars_left] = 0; + strcpy(pb->crs, print_string); + print_string[chars_left] = save_char; + pb->crs = pb->buf + chars_to_add - chars_left; } - pb_next = pb->next; - pb->next = NULL; - pb = pb_next; } - spin_unlock_bh(&print_lock); -} -/** - * TIPC_TEE - perform next output operation on both print buffers - * @b0: pointer to chain of print buffers (may be NULL) - * @b1: pointer to print buffer to add to chain - * - * Returns pointer to print buffer chain. - */ - -struct print_buf *TIPC_TEE(struct print_buf *b0, struct print_buf *b1) -{ - struct print_buf *pb = b0; + if (pb->echo) + printk(print_string); - if (!b0 || (b0 == b1)) - return b1; - - spin_lock_bh(&print_lock); - while (pb->next) { - if ((pb->next == b1) || (pb->next == b0)) - pb->next = pb->next->next; - else - pb = pb->next; - } - pb->next = b1; spin_unlock_bh(&print_lock); - return b0; } /** @@ -323,31 +309,28 @@ static void printbuf_dump(struct print_buf *pb) } /** - * tipc_dump - dump non-console print buffer(s) to console - * @pb: pointer to chain of print buffers + * tipc_dump - dump (non-console) print buffer to console + * @pb: pointer to print buffer */ void tipc_dump(struct print_buf *pb, const char *fmt, ...) { - struct print_buf *pb_next; int len; + if (pb == TIPC_CONS) + return; + spin_lock_bh(&print_lock); + FORMAT(print_string, len, fmt); printk(print_string); - for (; pb; pb = pb->next) { - if (pb != TIPC_CONS) { - printk("\n---- Start of %s log dump ----\n\n", - (pb == TIPC_LOG) ? "global" : "local"); - printbuf_dump(pb); - tipc_printbuf_reset(pb); - printk("\n---- End of dump ----\n"); - } - pb_next = pb->next; - pb->next = NULL; - pb = pb_next; - } + printk("\n---- Start of %s log dump ----\n\n", + (pb == TIPC_LOG) ? "global" : "local"); + printbuf_dump(pb); + tipc_printbuf_reset(pb); + printk("\n---- End of dump ----\n"); + spin_unlock_bh(&print_lock); } @@ -368,8 +351,10 @@ int tipc_log_resize(int log_size) if (log_size) { if (log_size < TIPC_PB_MIN_SIZE) log_size = TIPC_PB_MIN_SIZE; + res = TIPC_LOG->echo; tipc_printbuf_init(TIPC_LOG, kmalloc(log_size, GFP_ATOMIC), log_size); + TIPC_LOG->echo = res; res = !TIPC_LOG->buf; } spin_unlock_bh(&print_lock); diff --git a/net/tipc/dbg.h b/net/tipc/dbg.h index 6b00062bf040..5ef1bc8f64ef 100644 --- a/net/tipc/dbg.h +++ b/net/tipc/dbg.h @@ -42,14 +42,14 @@ * @buf: pointer to character array containing print buffer contents * @size: size of character array * @crs: pointer to first unused space in character array (i.e. final NUL) - * @next: used to link print buffers when printing to more than one at a time + * @echo: echo output to system console if non-zero */ struct print_buf { char *buf; u32 size; char *crs; - struct print_buf *next; + int echo; }; #define TIPC_PB_MIN_SIZE 64 /* minimum size for a print buffer's array */ -- cgit v1.2.2 From 7d3aa71239f588215b5a7c359f05155b192d8081 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 5 May 2008 01:22:59 -0700 Subject: [TIPC]: Cosmetic cleanup of print buffer code This patch contains changes to make TIPC's print buffer code conform more closely to Linux kernel coding guidelines. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/core.h | 6 ++++-- net/tipc/dbg.c | 14 +++++++------- 2 files changed, 11 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/tipc/core.h b/net/tipc/core.h index 7042ef377262..bc633552e9ff 100644 --- a/net/tipc/core.h +++ b/net/tipc/core.h @@ -2,7 +2,7 @@ * net/tipc/core.h: Include file for TIPC global declarations * * Copyright (c) 2005-2006, Ericsson AB - * Copyright (c) 2005-2006, Wind River Systems + * Copyright (c) 2005-2007, Wind River Systems * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -65,7 +65,9 @@ #define assert(i) BUG_ON(!(i)) struct tipc_msg; -extern struct print_buf *TIPC_NULL, *TIPC_CONS, *TIPC_LOG; +extern struct print_buf *const TIPC_NULL; +extern struct print_buf *const TIPC_CONS; +extern struct print_buf *const TIPC_LOG; void tipc_msg_print(struct print_buf*,struct tipc_msg *,const char*); void tipc_printf(struct print_buf *, const char *fmt, ...); void tipc_dump(struct print_buf*,const char *fmt, ...); diff --git a/net/tipc/dbg.c b/net/tipc/dbg.c index 8ca9457250be..cda496815a05 100644 --- a/net/tipc/dbg.c +++ b/net/tipc/dbg.c @@ -49,13 +49,13 @@ */ static struct print_buf null_buf = { NULL, 0, NULL, 0 }; -struct print_buf *TIPC_NULL = &null_buf; +struct print_buf *const TIPC_NULL = &null_buf; static struct print_buf cons_buf = { NULL, 0, NULL, 1 }; -struct print_buf *TIPC_CONS = &cons_buf; +struct print_buf *const TIPC_CONS = &cons_buf; static struct print_buf log_buf = { NULL, 0, NULL, 1 }; -struct print_buf *TIPC_LOG = &log_buf; +struct print_buf *const TIPC_LOG = &log_buf; /* * Locking policy when using print buffers. @@ -107,7 +107,7 @@ void tipc_printbuf_init(struct print_buf *pb, char *raw, u32 size) pb->buf = NULL; } else if (raw) { pb->buf[0] = 0; - pb->buf[size-1] = ~0; + pb->buf[size - 1] = ~0; } } @@ -118,7 +118,7 @@ void tipc_printbuf_init(struct print_buf *pb, char *raw, u32 size) void tipc_printbuf_reset(struct print_buf *pb) { - if (pb->buf != NULL) { + if (pb->buf) { pb->crs = pb->buf; pb->buf[0] = 0; pb->buf[pb->size - 1] = ~0; @@ -158,7 +158,7 @@ int tipc_printbuf_validate(struct print_buf *pb) if (pb->buf[pb->size - 1] == 0) { cp_buf = kmalloc(pb->size, GFP_ATOMIC); - if (cp_buf != NULL){ + if (cp_buf) { tipc_printbuf_init(&cb, cp_buf, pb->size); tipc_printbuf_move(&cb, pb); tipc_printbuf_move(pb, &cb); @@ -205,7 +205,7 @@ void tipc_printbuf_move(struct print_buf *pb_to, struct print_buf *pb_from) /* Copy data from char after cursor to end (if used) */ len = pb_from->buf + pb_from->size - pb_from->crs - 2; - if ((pb_from->buf[pb_from->size-1] == 0) && (len > 0)) { + if ((pb_from->buf[pb_from->size - 1] == 0) && (len > 0)) { strcpy(pb_to->buf, pb_from->crs + 1); pb_to->crs = pb_to->buf + len; } else -- cgit v1.2.2 From 6063da9d74d4da812ae0d8f233b7e320e15765e3 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 5 May 2008 01:23:34 -0700 Subject: tipc: Cosmetic cleanup of system & debug output declarations This patch contains changes to make TIPC's system & debug message declarations more readable. Declarations have been regrouped and recommented to make it easier to understand what output is generated in both standard and debugging modes. In addition, oversize lines have been fixed to respect the 80 character upper bound used in the kernel. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/core.h | 117 ++++++++++++++++++++++++++++++++------------------------ 1 file changed, 67 insertions(+), 50 deletions(-) (limited to 'net') diff --git a/net/tipc/core.h b/net/tipc/core.h index bc633552e9ff..d7f3b3c96b97 100644 --- a/net/tipc/core.h +++ b/net/tipc/core.h @@ -59,87 +59,104 @@ #include /* - * TIPC debugging code + * TIPC sanity test macros */ #define assert(i) BUG_ON(!(i)) -struct tipc_msg; -extern struct print_buf *const TIPC_NULL; -extern struct print_buf *const TIPC_CONS; -extern struct print_buf *const TIPC_LOG; -void tipc_msg_print(struct print_buf*,struct tipc_msg *,const char*); -void tipc_printf(struct print_buf *, const char *fmt, ...); -void tipc_dump(struct print_buf*,const char *fmt, ...); - -#ifdef CONFIG_TIPC_DEBUG - /* - * TIPC debug support included: - * - system messages are printed to TIPC_OUTPUT print buffer - * - debug messages are printed to DBG_OUTPUT print buffer + * TIPC system monitoring code */ -#define err(fmt, arg...) tipc_printf(TIPC_OUTPUT, KERN_ERR "TIPC: " fmt, ## arg) -#define warn(fmt, arg...) tipc_printf(TIPC_OUTPUT, KERN_WARNING "TIPC: " fmt, ## arg) -#define info(fmt, arg...) tipc_printf(TIPC_OUTPUT, KERN_NOTICE "TIPC: " fmt, ## arg) - -#define dbg(fmt, arg...) do {if (DBG_OUTPUT != TIPC_NULL) tipc_printf(DBG_OUTPUT, fmt, ## arg);} while(0) -#define msg_dbg(msg, txt) do {if (DBG_OUTPUT != TIPC_NULL) tipc_msg_print(DBG_OUTPUT, msg, txt);} while(0) -#define dump(fmt, arg...) do {if (DBG_OUTPUT != TIPC_NULL) tipc_dump(DBG_OUTPUT, fmt, ##arg);} while(0) - - /* - * By default, TIPC_OUTPUT is defined to be system console and TIPC log buffer, - * while DBG_OUTPUT is the null print buffer. These defaults can be changed - * here, or on a per .c file basis, by redefining these symbols. The following - * print buffer options are available: + * TIPC's print buffer subsystem supports the following print buffers: * - * TIPC_NULL : null buffer (i.e. print nowhere) - * TIPC_CONS : system console - * TIPC_LOG : TIPC log buffer - * &buf : user-defined buffer (struct print_buf *) + * TIPC_NULL : null buffer (i.e. print nowhere) + * TIPC_CONS : system console + * TIPC_LOG : TIPC log buffer + * &buf : user-defined buffer (struct print_buf *) * * Note: TIPC_LOG is configured to echo its output to the system console; * user-defined buffers can be configured to do the same thing. */ +extern struct print_buf *const TIPC_NULL; +extern struct print_buf *const TIPC_CONS; +extern struct print_buf *const TIPC_LOG; + +void tipc_printf(struct print_buf *, const char *fmt, ...); +void tipc_msg_print(struct print_buf *, struct tipc_msg *, const char *); +void tipc_dump(struct print_buf *, const char *fmt, ...); + +/* + * TIPC_OUTPUT is the destination print buffer for system messages. + */ + #ifndef TIPC_OUTPUT #define TIPC_OUTPUT TIPC_LOG #endif -#ifndef DBG_OUTPUT -#define DBG_OUTPUT TIPC_NULL -#endif - -#else - /* - * TIPC debug support not included: - * - system messages are printed to system console - * - debug messages are not printed + * TIPC can be configured to send system messages to TIPC_OUTPUT + * or to the system console only. */ +#ifdef CONFIG_TIPC_DEBUG + +#define err(fmt, arg...) tipc_printf(TIPC_OUTPUT, \ + KERN_ERR "TIPC: " fmt, ## arg) +#define warn(fmt, arg...) tipc_printf(TIPC_OUTPUT, \ + KERN_WARNING "TIPC: " fmt, ## arg) +#define info(fmt, arg...) tipc_printf(TIPC_OUTPUT, \ + KERN_NOTICE "TIPC: " fmt, ## arg) + +#else + #define err(fmt, arg...) printk(KERN_ERR "TIPC: " fmt , ## arg) #define info(fmt, arg...) printk(KERN_INFO "TIPC: " fmt , ## arg) #define warn(fmt, arg...) printk(KERN_WARNING "TIPC: " fmt , ## arg) -#define dbg(fmt, arg...) do {} while (0) -#define msg_dbg(msg,txt) do {} while (0) -#define dump(fmt,arg...) do {} while (0) +#endif +/* + * DBG_OUTPUT is the destination print buffer for debug messages. + * It defaults to the the null print buffer, but can be redefined + * (typically in the individual .c files being debugged) to allow + * selected debug messages to be generated where needed. + */ + +#ifndef DBG_OUTPUT +#define DBG_OUTPUT TIPC_NULL +#endif /* - * TIPC_OUTPUT is defined to be the system console, while DBG_OUTPUT is - * the null print buffer. Thes ensures that any system or debug messages - * that are generated without using the above macros are handled correctly. + * TIPC can be configured to send debug messages to the specified print buffer + * (typically DBG_OUTPUT) or to suppress them entirely. */ -#undef TIPC_OUTPUT -#define TIPC_OUTPUT TIPC_CONS +#ifdef CONFIG_TIPC_DEBUG -#undef DBG_OUTPUT -#define DBG_OUTPUT TIPC_NULL +#define dbg(fmt, arg...) \ + do { \ + if (DBG_OUTPUT != TIPC_NULL) \ + tipc_printf(DBG_OUTPUT, fmt, ## arg); \ + } while (0) +#define msg_dbg(msg, txt) \ + do { \ + if (DBG_OUTPUT != TIPC_NULL) \ + tipc_msg_print(DBG_OUTPUT, msg, txt); \ + } while (0) +#define dump(fmt, arg...) \ + do { \ + if (DBG_OUTPUT != TIPC_NULL) \ + tipc_dump(DBG_OUTPUT, fmt, ##arg); \ + } while (0) + +#else + +#define dbg(fmt, arg...) do {} while (0) +#define msg_dbg(msg, txt) do {} while (0) +#define dump(fmt, arg...) do {} while (0) #endif -- cgit v1.2.2 From 48c971394626173eaf1c33441ea1d900c88b21a3 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 5 May 2008 01:24:06 -0700 Subject: tipc: Exclude debug-only print buffer code when not debugging This patch modifies TIPC to only exclude debug-related print buffer routines when debugging capabilities are not required. It also fixes up some related #defines that exceed 80 characters. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/core.h | 12 ++++++++---- net/tipc/dbg.c | 8 ++++++-- net/tipc/link.c | 20 ++++++++++++++++---- net/tipc/msg.c | 10 +++++++--- 4 files changed, 37 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/tipc/core.h b/net/tipc/core.h index d7f3b3c96b97..0d783bcc6f9a 100644 --- a/net/tipc/core.h +++ b/net/tipc/core.h @@ -85,8 +85,6 @@ extern struct print_buf *const TIPC_CONS; extern struct print_buf *const TIPC_LOG; void tipc_printf(struct print_buf *, const char *fmt, ...); -void tipc_msg_print(struct print_buf *, struct tipc_msg *, const char *); -void tipc_dump(struct print_buf *, const char *fmt, ...); /* * TIPC_OUTPUT is the destination print buffer for system messages. @@ -144,20 +142,26 @@ void tipc_dump(struct print_buf *, const char *fmt, ...); #define msg_dbg(msg, txt) \ do { \ if (DBG_OUTPUT != TIPC_NULL) \ - tipc_msg_print(DBG_OUTPUT, msg, txt); \ + tipc_msg_dbg(DBG_OUTPUT, msg, txt); \ } while (0) #define dump(fmt, arg...) \ do { \ if (DBG_OUTPUT != TIPC_NULL) \ - tipc_dump(DBG_OUTPUT, fmt, ##arg); \ + tipc_dump_dbg(DBG_OUTPUT, fmt, ##arg); \ } while (0) +void tipc_msg_dbg(struct print_buf *, struct tipc_msg *, const char *); +void tipc_dump_dbg(struct print_buf *, const char *fmt, ...); + #else #define dbg(fmt, arg...) do {} while (0) #define msg_dbg(msg, txt) do {} while (0) #define dump(fmt, arg...) do {} while (0) +#define tipc_msg_dbg(...) do {} while (0) +#define tipc_dump_dbg(...) do {} while (0) + #endif diff --git a/net/tipc/dbg.c b/net/tipc/dbg.c index cda496815a05..29ecae851668 100644 --- a/net/tipc/dbg.c +++ b/net/tipc/dbg.c @@ -263,6 +263,8 @@ void tipc_printf(struct print_buf *pb, const char *fmt, ...) spin_unlock_bh(&print_lock); } +#ifdef CONFIG_TIPC_DEBUG + /** * print_to_console - write string of bytes to console in multiple chunks */ @@ -309,11 +311,11 @@ static void printbuf_dump(struct print_buf *pb) } /** - * tipc_dump - dump (non-console) print buffer to console + * tipc_dump_dbg - dump (non-console) print buffer to console * @pb: pointer to print buffer */ -void tipc_dump(struct print_buf *pb, const char *fmt, ...) +void tipc_dump_dbg(struct print_buf *pb, const char *fmt, ...) { int len; @@ -334,6 +336,8 @@ void tipc_dump(struct print_buf *pb, const char *fmt, ...) spin_unlock_bh(&print_lock); } +#endif + /** * tipc_log_resize - change the size of the TIPC log buffer * @log_size: print buffer size to use diff --git a/net/tipc/link.c b/net/tipc/link.c index 2a26a16e269f..bd206ebe4eea 100644 --- a/net/tipc/link.c +++ b/net/tipc/link.c @@ -147,9 +147,21 @@ static void link_print(struct link *l_ptr, struct print_buf *buf, #define LINK_LOG_BUF_SIZE 0 -#define dbg_link(fmt, arg...) do {if (LINK_LOG_BUF_SIZE) tipc_printf(&l_ptr->print_buf, fmt, ## arg); } while(0) -#define dbg_link_msg(msg, txt) do {if (LINK_LOG_BUF_SIZE) tipc_msg_print(&l_ptr->print_buf, msg, txt); } while(0) -#define dbg_link_state(txt) do {if (LINK_LOG_BUF_SIZE) link_print(l_ptr, &l_ptr->print_buf, txt); } while(0) +#define dbg_link(fmt, arg...) \ + do { \ + if (LINK_LOG_BUF_SIZE) \ + tipc_printf(&l_ptr->print_buf, fmt, ## arg); \ + } while (0) +#define dbg_link_msg(msg, txt) \ + do { \ + if (LINK_LOG_BUF_SIZE) \ + tipc_msg_dbg(&l_ptr->print_buf, msg, txt); \ + } while (0) +#define dbg_link_state(txt) \ + do { \ + if (LINK_LOG_BUF_SIZE) \ + link_print(l_ptr, &l_ptr->print_buf, txt); \ + } while (0) #define dbg_link_dump() do { \ if (LINK_LOG_BUF_SIZE) { \ tipc_printf(LOG, "\n\nDumping link <%s>:\n", l_ptr->name); \ @@ -1651,7 +1663,7 @@ static void link_retransmit_failure(struct link *l_ptr, struct sk_buff *buf) struct tipc_msg *msg = buf_msg(buf); warn("Retransmission failure on link <%s>\n", l_ptr->name); - tipc_msg_print(TIPC_OUTPUT, msg, ">RETR-FAIL>"); + tipc_msg_dbg(TIPC_OUTPUT, msg, ">RETR-FAIL>"); if (l_ptr->addr) { diff --git a/net/tipc/msg.c b/net/tipc/msg.c index 696a8633df75..38abebaae889 100644 --- a/net/tipc/msg.c +++ b/net/tipc/msg.c @@ -41,7 +41,9 @@ #include "bearer.h" -void tipc_msg_print(struct print_buf *buf, struct tipc_msg *msg, const char *str) +#ifdef CONFIG_TIPC_DEBUG + +void tipc_msg_dbg(struct print_buf *buf, struct tipc_msg *msg, const char *str) { u32 usr = msg_user(msg); tipc_printf(buf, str); @@ -315,9 +317,11 @@ void tipc_msg_print(struct print_buf *buf, struct tipc_msg *msg, const char *str } tipc_printf(buf, "\n"); if ((usr == CHANGEOVER_PROTOCOL) && (msg_msgcnt(msg))) { - tipc_msg_print(buf,msg_get_wrapped(msg)," /"); + tipc_msg_dbg(buf, msg_get_wrapped(msg), " /"); } if ((usr == MSG_FRAGMENTER) && (msg_type(msg) == FIRST_FRAGMENT)) { - tipc_msg_print(buf,msg_get_wrapped(msg)," /"); + tipc_msg_dbg(buf, msg_get_wrapped(msg), " /"); } } + +#endif -- cgit v1.2.2 From a7ca0268b5dfffcaa8a1fe40c6eccdeac50fa3ea Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 5 May 2008 01:24:36 -0700 Subject: tipc: Update version number to TIPC 1.6.4 This patch updates TIPC's version number to 1.6.4. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/tipc/core.c b/net/tipc/core.c index 862d41541615..6d6aa5a3c240 100644 --- a/net/tipc/core.c +++ b/net/tipc/core.c @@ -49,7 +49,7 @@ #include "config.h" -#define TIPC_MOD_VER "1.6.3" +#define TIPC_MOD_VER "1.6.4" #ifndef CONFIG_TIPC_ZONES #define CONFIG_TIPC_ZONES 3 -- cgit v1.2.2 From 1c014420583564ac09e3b67006f2e7050861e66b Mon Sep 17 00:00:00 2001 From: Ivo van Doorn Date: Thu, 17 Apr 2008 19:41:02 +0200 Subject: mac80211: Replace ieee80211_tx_control->key_idx with ieee80211_key_conf The hw_key_idx inside the ieee80211_key_conf structure does not provide all the information drivers might need to perform hardware encryption. This is in particular true for rt2x00 who needs to know the key algorithm and whether it is a shared or pairwise key. By passing the ieee80211_key_conf pointer it assures us that drivers can make full use of all information that it should know about a particular key. Additionally this patch updates all drivers to grab the hw_key_idx from the ieee80211_key_conf structure. v2: Removed bogus u16 cast v3: Add warning about ieee80211_tx_control pointers v4: Update warning about ieee80211_tx_control pointers Signed-off-by: Ivo van Doorn Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/wep.c | 2 +- net/mac80211/wpa.c | 8 ++++---- 2 files changed, 5 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/mac80211/wep.c b/net/mac80211/wep.c index affcecd78c10..3cbae42ec504 100644 --- a/net/mac80211/wep.c +++ b/net/mac80211/wep.c @@ -337,7 +337,7 @@ static int wep_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) if (ieee80211_wep_encrypt(tx->local, skb, tx->key)) return -1; } else { - tx->control->key_idx = tx->key->conf.hw_key_idx; + tx->control->hw_key = &tx->key->conf; if (tx->key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_IV) { if (!ieee80211_wep_add_iv(tx->local, skb, tx->key)) return -1; diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index 45709ada8fee..42f3654e1c5e 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -228,7 +228,7 @@ static int tkip_encrypt_skb(struct ieee80211_tx_data *tx, 0x7f), (u8) key->u.tkip.iv16); - tx->control->key_idx = tx->key->conf.hw_key_idx; + tx->control->hw_key = &tx->key->conf; return 0; } @@ -256,7 +256,7 @@ ieee80211_crypto_tkip_encrypt(struct ieee80211_tx_data *tx) !(tx->key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_IV) && !wpa_test) { /* hwaccel - with no need for preallocated room for IV/ICV */ - tx->control->key_idx = tx->key->conf.hw_key_idx; + tx->control->hw_key = &tx->key->conf; return TX_CONTINUE; } @@ -478,7 +478,7 @@ static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, if (key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) { /* hwaccel - with preallocated room for CCMP header */ - tx->control->key_idx = key->conf.hw_key_idx; + tx->control->hw_key = &tx->key->conf; return 0; } @@ -505,7 +505,7 @@ ieee80211_crypto_ccmp_encrypt(struct ieee80211_tx_data *tx) !(tx->key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_IV)) { /* hwaccel - with no need for preallocated room for CCMP " * header or MIC fields */ - tx->control->key_idx = tx->key->conf.hw_key_idx; + tx->control->hw_key = &tx->key->conf; return TX_CONTINUE; } -- cgit v1.2.2 From 988c0f723d0b1abb399e6e71d8bf3f8bf1949a70 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 17 Apr 2008 19:21:22 +0200 Subject: mac80211: a few code cleanups This has some code cleanups (some inspired by checkpatch), I got bored at probably a third of the output though so if somebody else wants to... Signed-off-by: Johannes Berg --- net/mac80211/aes_ccm.c | 2 +- net/mac80211/aes_ccm.h | 2 +- net/mac80211/debugfs_netdev.c | 2 +- net/mac80211/ieee80211_i.h | 6 +++--- net/mac80211/iface.c | 3 +-- net/mac80211/main.c | 6 +++--- net/mac80211/mlme.c | 31 ++++++++++++++----------------- net/mac80211/tkip.h | 4 ++-- net/mac80211/wep.h | 2 +- 9 files changed, 27 insertions(+), 31 deletions(-) (limited to 'net') diff --git a/net/mac80211/aes_ccm.c b/net/mac80211/aes_ccm.c index 59f1691f62c8..4d4c2dfcf9a0 100644 --- a/net/mac80211/aes_ccm.c +++ b/net/mac80211/aes_ccm.c @@ -134,7 +134,7 @@ int ieee80211_aes_ccm_decrypt(struct crypto_cipher *tfm, u8 *scratch, } -struct crypto_cipher * ieee80211_aes_key_setup_encrypt(const u8 key[]) +struct crypto_cipher *ieee80211_aes_key_setup_encrypt(const u8 key[]) { struct crypto_cipher *tfm; diff --git a/net/mac80211/aes_ccm.h b/net/mac80211/aes_ccm.h index 885f19030b29..8cd0f14aab4d 100644 --- a/net/mac80211/aes_ccm.h +++ b/net/mac80211/aes_ccm.h @@ -14,7 +14,7 @@ #define AES_BLOCK_LEN 16 -struct crypto_cipher * ieee80211_aes_key_setup_encrypt(const u8 key[]); +struct crypto_cipher *ieee80211_aes_key_setup_encrypt(const u8 key[]); void ieee80211_aes_ccm_encrypt(struct crypto_cipher *tfm, u8 *scratch, u8 *b_0, u8 *aad, u8 *data, size_t data_len, u8 *cdata, u8 *mic); diff --git a/net/mac80211/debugfs_netdev.c b/net/mac80211/debugfs_netdev.c index e3326d046944..3ae5493d728b 100644 --- a/net/mac80211/debugfs_netdev.c +++ b/net/mac80211/debugfs_netdev.c @@ -528,7 +528,7 @@ void ieee80211_debugfs_change_if_type(struct ieee80211_sub_if_data *sdata, add_files(sdata); } -static int netdev_notify(struct notifier_block * nb, +static int netdev_notify(struct notifier_block *nb, unsigned long state, void *ndev) { diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index c7314bf4bec2..a05e1db36993 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -919,9 +919,9 @@ ieee80211_rx_result ieee80211_sta_rx_scan( void ieee80211_rx_bss_list_init(struct net_device *dev); void ieee80211_rx_bss_list_deinit(struct net_device *dev); int ieee80211_sta_set_extra_ie(struct net_device *dev, char *ie, size_t len); -struct sta_info * ieee80211_ibss_add_sta(struct net_device *dev, - struct sk_buff *skb, u8 *bssid, - u8 *addr); +struct sta_info *ieee80211_ibss_add_sta(struct net_device *dev, + struct sk_buff *skb, u8 *bssid, + u8 *addr); int ieee80211_sta_deauthenticate(struct net_device *dev, u16 reason); int ieee80211_sta_disassociate(struct net_device *dev, u16 reason); void ieee80211_bss_info_change_notify(struct ieee80211_sub_if_data *sdata, diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c index 80954a512185..f41c7e0de622 100644 --- a/net/mac80211/iface.c +++ b/net/mac80211/iface.c @@ -33,9 +33,8 @@ static void ieee80211_if_sdata_deinit(struct ieee80211_sub_if_data *sdata) { int i; - for (i = 0; i < IEEE80211_FRAGMENT_MAX; i++) { + for (i = 0; i < IEEE80211_FRAGMENT_MAX; i++) __skb_queue_purge(&sdata->fragments[i].skb_list); - } } /* Must be called with rtnl lock held. */ diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 9ad4e3631b6b..e19be27a3def 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -385,8 +385,8 @@ static int ieee80211_open(struct net_device *dev) * yet be effective. Trigger execution of ieee80211_sta_work * to fix this. */ - if(sdata->vif.type == IEEE80211_IF_TYPE_STA || - sdata->vif.type == IEEE80211_IF_TYPE_IBSS) { + if (sdata->vif.type == IEEE80211_IF_TYPE_STA || + sdata->vif.type == IEEE80211_IF_TYPE_IBSS) { struct ieee80211_if_sta *ifsta = &sdata->u.sta; queue_work(local->hw.workqueue, &ifsta->work); } @@ -1482,7 +1482,7 @@ void ieee80211_tx_status(struct ieee80211_hw *hw, struct sk_buff *skb, return; } - rthdr = (struct ieee80211_tx_status_rtap_hdr*) + rthdr = (struct ieee80211_tx_status_rtap_hdr *) skb_push(skb, sizeof(*rthdr)); memset(rthdr, 0, sizeof(*rthdr)); diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index a5e5c31c23ab..de09f58d9683 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -306,28 +306,24 @@ static void ieee80211_sta_wmm_params(struct net_device *dev, switch (aci) { case 1: queue = IEEE80211_TX_QUEUE_DATA3; - if (acm) { + if (acm) local->wmm_acm |= BIT(0) | BIT(3); - } break; case 2: queue = IEEE80211_TX_QUEUE_DATA1; - if (acm) { + if (acm) local->wmm_acm |= BIT(4) | BIT(5); - } break; case 3: queue = IEEE80211_TX_QUEUE_DATA0; - if (acm) { + if (acm) local->wmm_acm |= BIT(6) | BIT(7); - } break; case 0: default: queue = IEEE80211_TX_QUEUE_DATA2; - if (acm) { + if (acm) local->wmm_acm |= BIT(1) | BIT(2); - } break; } @@ -706,9 +702,8 @@ static void ieee80211_send_assoc(struct net_device *dev, if (bss) { if (bss->capability & WLAN_CAPABILITY_PRIVACY) capab |= WLAN_CAPABILITY_PRIVACY; - if (bss->wmm_ie) { + if (bss->wmm_ie) wmm = 1; - } ieee80211_rx_bss_put(dev, bss); } @@ -1805,9 +1800,8 @@ static void ieee80211_rx_mgmt_deauth(struct net_device *dev, " (reason=%d)\n", dev->name, print_mac(mac, mgmt->sa), reason_code); - if (ifsta->flags & IEEE80211_STA_AUTHENTICATED) { + if (ifsta->flags & IEEE80211_STA_AUTHENTICATED) printk(KERN_DEBUG "%s: deauthenticated\n", dev->name); - } if (ifsta->state == IEEE80211_AUTHENTICATE || ifsta->state == IEEE80211_ASSOCIATE || @@ -3517,10 +3511,12 @@ static int ieee80211_sta_create_ibss(struct net_device *dev, bss->beacon_int = local->hw.conf.beacon_int; bss->last_update = jiffies; bss->capability = WLAN_CAPABILITY_IBSS; - if (sdata->default_key) { + + if (sdata->default_key) bss->capability |= WLAN_CAPABILITY_PRIVACY; - } else + else sdata->drop_unencrypted = 0; + bss->supp_rates_len = sband->n_bitrates; pos = bss->supp_rates; for (i = 0; i < sband->n_bitrates; i++) { @@ -4203,6 +4199,7 @@ int ieee80211_sta_set_extra_ie(struct net_device *dev, char *ie, size_t len) { struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); struct ieee80211_if_sta *ifsta = &sdata->u.sta; + kfree(ifsta->extra_ie); if (len == 0) { ifsta->extra_ie = NULL; @@ -4220,9 +4217,9 @@ int ieee80211_sta_set_extra_ie(struct net_device *dev, char *ie, size_t len) } -struct sta_info * ieee80211_ibss_add_sta(struct net_device *dev, - struct sk_buff *skb, u8 *bssid, - u8 *addr) +struct sta_info *ieee80211_ibss_add_sta(struct net_device *dev, + struct sk_buff *skb, u8 *bssid, + u8 *addr) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct sta_info *sta; diff --git a/net/mac80211/tkip.h b/net/mac80211/tkip.h index b7c2ee763d9d..1fa0bb4dba3c 100644 --- a/net/mac80211/tkip.h +++ b/net/mac80211/tkip.h @@ -13,8 +13,8 @@ #include #include "key.h" -u8 * ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, - u8 iv0, u8 iv1, u8 iv2); +u8 *ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, + u8 iv0, u8 iv1, u8 iv2); void ieee80211_tkip_gen_phase1key(struct ieee80211_key *key, u8 *ta, u16 *phase1key); void ieee80211_tkip_gen_rc4key(struct ieee80211_key *key, u8 *ta, diff --git a/net/mac80211/wep.h b/net/mac80211/wep.h index 363779c50658..e587172115b8 100644 --- a/net/mac80211/wep.h +++ b/net/mac80211/wep.h @@ -26,7 +26,7 @@ int ieee80211_wep_encrypt(struct ieee80211_local *local, struct sk_buff *skb, struct ieee80211_key *key); int ieee80211_wep_decrypt(struct ieee80211_local *local, struct sk_buff *skb, struct ieee80211_key *key); -u8 * ieee80211_wep_is_weak_iv(struct sk_buff *skb, struct ieee80211_key *key); +u8 *ieee80211_wep_is_weak_iv(struct sk_buff *skb, struct ieee80211_key *key); ieee80211_rx_result ieee80211_crypto_wep_decrypt(struct ieee80211_rx_data *rx); -- cgit v1.2.2 From c6adbd2158fee972adcc6232de5e2ef375f1f782 Mon Sep 17 00:00:00 2001 From: Ivo van Doorn Date: Thu, 17 Apr 2008 21:11:18 +0200 Subject: mac80211: Add IEEE80211_KEY_FLAG_PAIRWISE This adds a new flag to the ieee80211_key_conf structure. This flag will inform the driver the key is pairwise rather then a shared key. This is important for drivers who support both types of keys, and need to be informed which type of key this is. Alternative would be drivers checking the address argument of set_key(), but it will be safer when mac80211 is more explicit. Signed-off-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/mac80211/key.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'net') diff --git a/net/mac80211/key.c b/net/mac80211/key.c index 150d66dbda9d..88b211af7c1f 100644 --- a/net/mac80211/key.c +++ b/net/mac80211/key.c @@ -323,6 +323,13 @@ void ieee80211_key_link(struct ieee80211_key *key, */ if (sta->flags & WLAN_STA_WME) key->conf.flags |= IEEE80211_KEY_FLAG_WMM_STA; + + /* + * This key is for a specific sta interface, + * inform the driver that it should try to store + * this key as pairwise key. + */ + key->conf.flags |= IEEE80211_KEY_FLAG_PAIRWISE; } else { if (sdata->vif.type == IEEE80211_IF_TYPE_STA) { struct sta_info *ap; -- cgit v1.2.2 From 3acea5b616c6d85008700a9d51cb02a81b2d0c67 Mon Sep 17 00:00:00 2001 From: Ester Kummer Date: Thu, 17 Apr 2008 16:05:14 -0700 Subject: mac80211: correct skb allocation This patch corrects the allocation of skb in ADDBA req/resp and DELBA it removes redundant space u.addba_* are already counted in sizeof(*mgmt) Signed-off-by: Ester Kummer Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index de09f58d9683..63b391d01251 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -1093,8 +1093,8 @@ static void ieee80211_send_addba_resp(struct net_device *dev, u8 *da, u16 tid, struct ieee80211_mgmt *mgmt; u16 capab; - skb = dev_alloc_skb(sizeof(*mgmt) + local->hw.extra_tx_headroom + 1 + - sizeof(mgmt->u.action.u.addba_resp)); + skb = dev_alloc_skb(sizeof(*mgmt) + local->hw.extra_tx_headroom); + if (!skb) { printk(KERN_DEBUG "%s: failed to allocate buffer " "for addba resp frame\n", dev->name); @@ -1142,9 +1142,7 @@ void ieee80211_send_addba_request(struct net_device *dev, const u8 *da, struct ieee80211_mgmt *mgmt; u16 capab; - skb = dev_alloc_skb(sizeof(*mgmt) + local->hw.extra_tx_headroom + 1 + - sizeof(mgmt->u.action.u.addba_req)); - + skb = dev_alloc_skb(sizeof(*mgmt) + local->hw.extra_tx_headroom); if (!skb) { printk(KERN_ERR "%s: failed to allocate buffer " @@ -1406,8 +1404,7 @@ void ieee80211_send_delba(struct net_device *dev, const u8 *da, u16 tid, struct ieee80211_mgmt *mgmt; u16 params; - skb = dev_alloc_skb(sizeof(*mgmt) + local->hw.extra_tx_headroom + 1 + - sizeof(mgmt->u.action.u.delba)); + skb = dev_alloc_skb(sizeof(*mgmt) + local->hw.extra_tx_headroom); if (!skb) { printk(KERN_ERR "%s: failed to allocate buffer " -- cgit v1.2.2 From 712590de5e5c977ff96efa3a16fcc03b0e976c3c Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Mon, 21 Apr 2008 11:47:51 +0300 Subject: make sta_rx_agg_session_timer_expired() static sta_rx_agg_session_timer_expired() can now become static. Signed-off-by: Adrian Bunk Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 1 - net/mac80211/mlme.c | 3 ++- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index a05e1db36993..cabdc1439d57 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -940,7 +940,6 @@ void ieee80211_send_delba(struct net_device *dev, const u8 *da, u16 tid, void ieee80211_sta_stop_rx_ba_session(struct net_device *dev, u8 *da, u16 tid, u16 initiator, u16 reason); -void sta_rx_agg_session_timer_expired(unsigned long data); void sta_addba_resp_timer_expired(unsigned long data); void ieee80211_sta_tear_down_BA_sessions(struct net_device *dev, u8 *addr); u64 ieee80211_sta_get_rates(struct ieee80211_local *local, diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 63b391d01251..e860d0bacea9 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -87,6 +87,7 @@ static int ieee80211_sta_start_scan(struct net_device *dev, u8 *ssid, size_t ssid_len); static int ieee80211_sta_config_auth(struct net_device *dev, struct ieee80211_if_sta *ifsta); +static void sta_rx_agg_session_timer_expired(unsigned long data); void ieee802_11_parse_elems(u8 *start, size_t len, @@ -1608,7 +1609,7 @@ timer_expired_exit: * resetting it after each frame that arrives from the originator. * if this timer expires ieee80211_sta_stop_rx_ba_session will be executed. */ -void sta_rx_agg_session_timer_expired(unsigned long data) +static void sta_rx_agg_session_timer_expired(unsigned long data) { /* not an elegant detour, but there is no choice as the timer passes * only one argument, and verious sta_info are needed here, so init -- cgit v1.2.2 From c12cf2109702c052688391f4171f239effb241bf Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Mon, 21 Apr 2008 11:48:28 +0300 Subject: remove ieee80211_tx_frame() After the softmac removal ieee80211_tx_frame() was no longer used. Signed-off-by: Adrian Bunk Signed-off-by: John W. Linville --- net/ieee80211/ieee80211_tx.c | 86 -------------------------------------------- 1 file changed, 86 deletions(-) (limited to 'net') diff --git a/net/ieee80211/ieee80211_tx.c b/net/ieee80211/ieee80211_tx.c index d8b02603cbe5..d996547f7a62 100644 --- a/net/ieee80211/ieee80211_tx.c +++ b/net/ieee80211/ieee80211_tx.c @@ -542,90 +542,4 @@ int ieee80211_xmit(struct sk_buff *skb, struct net_device *dev) return 1; } -/* Incoming 802.11 strucure is converted to a TXB - * a block of 802.11 fragment packets (stored as skbs) */ -int ieee80211_tx_frame(struct ieee80211_device *ieee, - struct ieee80211_hdr *frame, int hdr_len, int total_len, - int encrypt_mpdu) -{ - struct ieee80211_txb *txb = NULL; - unsigned long flags; - struct net_device_stats *stats = &ieee->stats; - struct sk_buff *skb_frag; - int priority = -1; - int fraglen = total_len; - int headroom = ieee->tx_headroom; - struct ieee80211_crypt_data *crypt = ieee->crypt[ieee->tx_keyidx]; - - spin_lock_irqsave(&ieee->lock, flags); - - if (encrypt_mpdu && (!ieee->sec.encrypt || !crypt)) - encrypt_mpdu = 0; - - /* If there is no driver handler to take the TXB, dont' bother - * creating it... */ - if (!ieee->hard_start_xmit) { - printk(KERN_WARNING "%s: No xmit handler.\n", ieee->dev->name); - goto success; - } - - if (unlikely(total_len < 24)) { - printk(KERN_WARNING "%s: skb too small (%d).\n", - ieee->dev->name, total_len); - goto success; - } - - if (encrypt_mpdu) { - frame->frame_ctl |= cpu_to_le16(IEEE80211_FCTL_PROTECTED); - fraglen += crypt->ops->extra_mpdu_prefix_len + - crypt->ops->extra_mpdu_postfix_len; - headroom += crypt->ops->extra_mpdu_prefix_len; - } - - /* When we allocate the TXB we allocate enough space for the reserve - * and full fragment bytes (bytes_per_frag doesn't include prefix, - * postfix, header, FCS, etc.) */ - txb = ieee80211_alloc_txb(1, fraglen, headroom, GFP_ATOMIC); - if (unlikely(!txb)) { - printk(KERN_WARNING "%s: Could not allocate TXB\n", - ieee->dev->name); - goto failed; - } - txb->encrypted = 0; - txb->payload_size = fraglen; - - skb_frag = txb->fragments[0]; - - memcpy(skb_put(skb_frag, total_len), frame, total_len); - - if (ieee->config & - (CFG_IEEE80211_COMPUTE_FCS | CFG_IEEE80211_RESERVE_FCS)) - skb_put(skb_frag, 4); - - /* To avoid overcomplicating things, we do the corner-case frame - * encryption in software. The only real situation where encryption is - * needed here is during software-based shared key authentication. */ - if (encrypt_mpdu) - ieee80211_encrypt_fragment(ieee, skb_frag, hdr_len); - - success: - spin_unlock_irqrestore(&ieee->lock, flags); - - if (txb) { - if ((*ieee->hard_start_xmit) (txb, ieee->dev, priority) == 0) { - stats->tx_packets++; - stats->tx_bytes += txb->payload_size; - return 0; - } - ieee80211_txb_free(txb); - } - return 0; - - failed: - spin_unlock_irqrestore(&ieee->lock, flags); - stats->tx_errors++; - return 1; -} - -EXPORT_SYMBOL(ieee80211_tx_frame); EXPORT_SYMBOL(ieee80211_txb_free); -- cgit v1.2.2 From 7eafd25d9559bd0f652449c222d38d63412e3d4a Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Mon, 21 Apr 2008 11:48:34 +0300 Subject: remove ieee80211_wx_{get,set}_auth() After the bcm43xx removal ieee80211_wx_{get,set}_auth() were no longer used. Signed-off-by: Adrian Bunk Signed-off-by: John W. Linville --- net/ieee80211/ieee80211_wx.c | 89 -------------------------------------------- 1 file changed, 89 deletions(-) (limited to 'net') diff --git a/net/ieee80211/ieee80211_wx.c b/net/ieee80211/ieee80211_wx.c index 623489afa62c..822606b615ca 100644 --- a/net/ieee80211/ieee80211_wx.c +++ b/net/ieee80211/ieee80211_wx.c @@ -744,98 +744,9 @@ int ieee80211_wx_get_encodeext(struct ieee80211_device *ieee, return 0; } -int ieee80211_wx_set_auth(struct net_device *dev, - struct iw_request_info *info, - union iwreq_data *wrqu, - char *extra) -{ - struct ieee80211_device *ieee = netdev_priv(dev); - unsigned long flags; - int err = 0; - - spin_lock_irqsave(&ieee->lock, flags); - - switch (wrqu->param.flags & IW_AUTH_INDEX) { - case IW_AUTH_WPA_VERSION: - case IW_AUTH_CIPHER_PAIRWISE: - case IW_AUTH_CIPHER_GROUP: - case IW_AUTH_KEY_MGMT: - /* - * Host AP driver does not use these parameters and allows - * wpa_supplicant to control them internally. - */ - break; - case IW_AUTH_TKIP_COUNTERMEASURES: - break; /* FIXME */ - case IW_AUTH_DROP_UNENCRYPTED: - ieee->drop_unencrypted = !!wrqu->param.value; - break; - case IW_AUTH_80211_AUTH_ALG: - break; /* FIXME */ - case IW_AUTH_WPA_ENABLED: - ieee->privacy_invoked = ieee->wpa_enabled = !!wrqu->param.value; - break; - case IW_AUTH_RX_UNENCRYPTED_EAPOL: - ieee->ieee802_1x = !!wrqu->param.value; - break; - case IW_AUTH_PRIVACY_INVOKED: - ieee->privacy_invoked = !!wrqu->param.value; - break; - default: - err = -EOPNOTSUPP; - break; - } - spin_unlock_irqrestore(&ieee->lock, flags); - return err; -} - -int ieee80211_wx_get_auth(struct net_device *dev, - struct iw_request_info *info, - union iwreq_data *wrqu, - char *extra) -{ - struct ieee80211_device *ieee = netdev_priv(dev); - unsigned long flags; - int err = 0; - - spin_lock_irqsave(&ieee->lock, flags); - - switch (wrqu->param.flags & IW_AUTH_INDEX) { - case IW_AUTH_WPA_VERSION: - case IW_AUTH_CIPHER_PAIRWISE: - case IW_AUTH_CIPHER_GROUP: - case IW_AUTH_KEY_MGMT: - case IW_AUTH_TKIP_COUNTERMEASURES: /* FIXME */ - case IW_AUTH_80211_AUTH_ALG: /* FIXME */ - /* - * Host AP driver does not use these parameters and allows - * wpa_supplicant to control them internally. - */ - err = -EOPNOTSUPP; - break; - case IW_AUTH_DROP_UNENCRYPTED: - wrqu->param.value = ieee->drop_unencrypted; - break; - case IW_AUTH_WPA_ENABLED: - wrqu->param.value = ieee->wpa_enabled; - break; - case IW_AUTH_RX_UNENCRYPTED_EAPOL: - wrqu->param.value = ieee->ieee802_1x; - break; - default: - err = -EOPNOTSUPP; - break; - } - spin_unlock_irqrestore(&ieee->lock, flags); - return err; -} - EXPORT_SYMBOL(ieee80211_wx_set_encodeext); EXPORT_SYMBOL(ieee80211_wx_get_encodeext); EXPORT_SYMBOL(ieee80211_wx_get_scan); EXPORT_SYMBOL(ieee80211_wx_set_encode); EXPORT_SYMBOL(ieee80211_wx_get_encode); - -EXPORT_SYMBOL_GPL(ieee80211_wx_set_auth); -EXPORT_SYMBOL_GPL(ieee80211_wx_get_auth); -- cgit v1.2.2 From e100bb64bf7cdeae7f742a65ee1985649a7fd1b4 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Wed, 30 Apr 2008 18:51:21 +0200 Subject: mac80211: QoS related cleanups This * makes the queue number passed to drivers a u16 (as it will be with skb_get_queue_mapping) * removes the useless queue number defines * splits hw->queues into hw->queues/ampdu_queues * removes the debugfs files for per-queue counters * removes some dead QoS code * removes the beacon queue configuration for IBSS so that the drivers now never get a queue number bigger than (hw->queues + hw->ampdu_queues - 1) for tx and only in the range 0..hw->queues-1 for conf_tx. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/debugfs.c | 43 ---------------- net/mac80211/debugfs_sta.c | 38 -------------- net/mac80211/ieee80211_i.h | 8 +-- net/mac80211/main.c | 5 ++ net/mac80211/mlme.c | 23 +++------ net/mac80211/rx.c | 5 -- net/mac80211/sta_info.c | 2 +- net/mac80211/wme.c | 120 +++++++++++++++++++-------------------------- 8 files changed, 64 insertions(+), 180 deletions(-) (limited to 'net') diff --git a/net/mac80211/debugfs.c b/net/mac80211/debugfs.c index 1cccbfd781f6..d20d90eead1f 100644 --- a/net/mac80211/debugfs.c +++ b/net/mac80211/debugfs.c @@ -197,45 +197,6 @@ DEBUGFS_STATS_FILE(rx_handlers_fragments, 20, "%u", DEBUGFS_STATS_FILE(tx_status_drop, 20, "%u", local->tx_status_drop); -static ssize_t stats_wme_rx_queue_read(struct file *file, - char __user *userbuf, - size_t count, loff_t *ppos) -{ - struct ieee80211_local *local = file->private_data; - char buf[NUM_RX_DATA_QUEUES*15], *p = buf; - int i; - - for (i = 0; i < NUM_RX_DATA_QUEUES; i++) - p += scnprintf(p, sizeof(buf)+buf-p, - "%u\n", local->wme_rx_queue[i]); - - return simple_read_from_buffer(userbuf, count, ppos, buf, p-buf); -} - -static const struct file_operations stats_wme_rx_queue_ops = { - .read = stats_wme_rx_queue_read, - .open = mac80211_open_file_generic, -}; - -static ssize_t stats_wme_tx_queue_read(struct file *file, - char __user *userbuf, - size_t count, loff_t *ppos) -{ - struct ieee80211_local *local = file->private_data; - char buf[NUM_TX_DATA_QUEUES*15], *p = buf; - int i; - - for (i = 0; i < NUM_TX_DATA_QUEUES; i++) - p += scnprintf(p, sizeof(buf)+buf-p, - "%u\n", local->wme_tx_queue[i]); - - return simple_read_from_buffer(userbuf, count, ppos, buf, p-buf); -} - -static const struct file_operations stats_wme_tx_queue_ops = { - .read = stats_wme_tx_queue_read, - .open = mac80211_open_file_generic, -}; #endif DEBUGFS_DEVSTATS_FILE(dot11ACKFailureCount); @@ -303,8 +264,6 @@ void debugfs_hw_add(struct ieee80211_local *local) DEBUGFS_STATS_ADD(rx_expand_skb_head2); DEBUGFS_STATS_ADD(rx_handlers_fragments); DEBUGFS_STATS_ADD(tx_status_drop); - DEBUGFS_STATS_ADD(wme_tx_queue); - DEBUGFS_STATS_ADD(wme_rx_queue); #endif DEBUGFS_STATS_ADD(dot11ACKFailureCount); DEBUGFS_STATS_ADD(dot11RTSFailureCount); @@ -356,8 +315,6 @@ void debugfs_hw_del(struct ieee80211_local *local) DEBUGFS_STATS_DEL(rx_expand_skb_head2); DEBUGFS_STATS_DEL(rx_handlers_fragments); DEBUGFS_STATS_DEL(tx_status_drop); - DEBUGFS_STATS_DEL(wme_tx_queue); - DEBUGFS_STATS_DEL(wme_rx_queue); #endif DEBUGFS_STATS_DEL(dot11ACKFailureCount); DEBUGFS_STATS_DEL(dot11RTSFailureCount); diff --git a/net/mac80211/debugfs_sta.c b/net/mac80211/debugfs_sta.c index 6d47a1d31b37..676a93202ff9 100644 --- a/net/mac80211/debugfs_sta.c +++ b/net/mac80211/debugfs_sta.c @@ -123,36 +123,6 @@ static ssize_t sta_last_seq_ctrl_read(struct file *file, char __user *userbuf, } STA_OPS(last_seq_ctrl); -#ifdef CONFIG_MAC80211_DEBUG_COUNTERS -static ssize_t sta_wme_rx_queue_read(struct file *file, char __user *userbuf, - size_t count, loff_t *ppos) -{ - char buf[15*NUM_RX_DATA_QUEUES], *p = buf; - int i; - struct sta_info *sta = file->private_data; - for (i = 0; i < NUM_RX_DATA_QUEUES; i++) - p += scnprintf(p, sizeof(buf)+buf-p, "%u ", - sta->wme_rx_queue[i]); - p += scnprintf(p, sizeof(buf)+buf-p, "\n"); - return simple_read_from_buffer(userbuf, count, ppos, buf, p - buf); -} -STA_OPS(wme_rx_queue); - -static ssize_t sta_wme_tx_queue_read(struct file *file, char __user *userbuf, - size_t count, loff_t *ppos) -{ - char buf[15*NUM_TX_DATA_QUEUES], *p = buf; - int i; - struct sta_info *sta = file->private_data; - for (i = 0; i < NUM_TX_DATA_QUEUES; i++) - p += scnprintf(p, sizeof(buf)+buf-p, "%u ", - sta->wme_tx_queue[i]); - p += scnprintf(p, sizeof(buf)+buf-p, "\n"); - return simple_read_from_buffer(userbuf, count, ppos, buf, p - buf); -} -STA_OPS(wme_tx_queue); -#endif - static ssize_t sta_agg_status_read(struct file *file, char __user *userbuf, size_t count, loff_t *ppos) { @@ -293,10 +263,6 @@ void ieee80211_sta_debugfs_add(struct sta_info *sta) DEBUGFS_ADD(num_ps_buf_frames); DEBUGFS_ADD(inactive_ms); DEBUGFS_ADD(last_seq_ctrl); -#ifdef CONFIG_MAC80211_DEBUG_COUNTERS - DEBUGFS_ADD(wme_rx_queue); - DEBUGFS_ADD(wme_tx_queue); -#endif DEBUGFS_ADD(agg_status); } @@ -306,10 +272,6 @@ void ieee80211_sta_debugfs_remove(struct sta_info *sta) DEBUGFS_DEL(num_ps_buf_frames); DEBUGFS_DEL(inactive_ms); DEBUGFS_DEL(last_seq_ctrl); -#ifdef CONFIG_MAC80211_DEBUG_COUNTERS - DEBUGFS_DEL(wme_rx_queue); - DEBUGFS_DEL(wme_tx_queue); -#endif DEBUGFS_DEL(agg_status); debugfs_remove(sta->debugfs.dir); diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index cabdc1439d57..d82ed20a344c 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -610,8 +610,8 @@ struct ieee80211_local { struct sta_info *sta_hash[STA_HASH_SIZE]; struct timer_list sta_cleanup; - unsigned long state[NUM_TX_DATA_QUEUES_AMPDU]; - struct ieee80211_tx_stored_packet pending_packet[NUM_TX_DATA_QUEUES_AMPDU]; + unsigned long state[IEEE80211_MAX_AMPDU_QUEUES + IEEE80211_MAX_AMPDU_QUEUES]; + struct ieee80211_tx_stored_packet pending_packet[IEEE80211_MAX_AMPDU_QUEUES + IEEE80211_MAX_AMPDU_QUEUES]; struct tasklet_struct tx_pending_tasklet; /* number of interfaces with corresponding IFF_ flags */ @@ -705,8 +705,6 @@ struct ieee80211_local { unsigned int rx_expand_skb_head2; unsigned int rx_handlers_fragments; unsigned int tx_status_drop; - unsigned int wme_rx_queue[NUM_RX_DATA_QUEUES]; - unsigned int wme_tx_queue[NUM_RX_DATA_QUEUES]; #define I802_DEBUG_INC(c) (c)++ #else /* CONFIG_MAC80211_DEBUG_COUNTERS */ #define I802_DEBUG_INC(c) do { } while (0) @@ -764,8 +762,6 @@ struct ieee80211_local { struct dentry *rx_expand_skb_head2; struct dentry *rx_handlers_fragments; struct dentry *tx_status_drop; - struct dentry *wme_tx_queue; - struct dentry *wme_rx_queue; #endif struct dentry *dot11ACKFailureCount; struct dentry *dot11RTSFailureCount; diff --git a/net/mac80211/main.c b/net/mac80211/main.c index e19be27a3def..55e76117da9e 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -1745,6 +1745,11 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) goto fail_wep; } + if (hw->queues > IEEE80211_MAX_QUEUES) + hw->queues = IEEE80211_MAX_QUEUES; + if (hw->ampdu_queues > IEEE80211_MAX_AMPDU_QUEUES) + hw->ampdu_queues = IEEE80211_MAX_AMPDU_QUEUES; + ieee80211_install_qdisc(local->mdev); /* add one default STA interface */ diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index e860d0bacea9..55b85ae5bc11 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -257,19 +257,8 @@ static void ieee80211_sta_def_wmm_params(struct net_device *dev, qparam.cw_max = 1023; qparam.txop = 0; - for (i = IEEE80211_TX_QUEUE_DATA0; i < NUM_TX_DATA_QUEUES; i++) - local->ops->conf_tx(local_to_hw(local), - i + IEEE80211_TX_QUEUE_DATA0, - &qparam); - - if (ibss) { - /* IBSS uses different parameters for Beacon sending */ - qparam.cw_min++; - qparam.cw_min *= 2; - qparam.cw_min--; - local->ops->conf_tx(local_to_hw(local), - IEEE80211_TX_QUEUE_BEACON, &qparam); - } + for (i = 0; i < local_to_hw(local)->queues; i++) + local->ops->conf_tx(local_to_hw(local), i, &qparam); } } @@ -306,23 +295,23 @@ static void ieee80211_sta_wmm_params(struct net_device *dev, switch (aci) { case 1: - queue = IEEE80211_TX_QUEUE_DATA3; + queue = 3; if (acm) local->wmm_acm |= BIT(0) | BIT(3); break; case 2: - queue = IEEE80211_TX_QUEUE_DATA1; + queue = 1; if (acm) local->wmm_acm |= BIT(4) | BIT(5); break; case 3: - queue = IEEE80211_TX_QUEUE_DATA0; + queue = 0; if (acm) local->wmm_acm |= BIT(6) | BIT(7); break; case 0: default: - queue = IEEE80211_TX_QUEUE_DATA2; + queue = 2; if (acm) local->wmm_acm |= BIT(1) | BIT(2); break; diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index 02f436a86061..e8b89c89e875 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -275,11 +275,6 @@ static void ieee80211_parse_qos(struct ieee80211_rx_data *rx) } } - I802_DEBUG_INC(rx->local->wme_rx_queue[tid]); - /* only a debug counter, sta might not be assigned properly yet */ - if (rx->sta) - I802_DEBUG_INC(rx->sta->wme_rx_queue[tid]); - rx->queue = tid; /* Set skb->priority to 1d tag if highest order bit of TID is not set. * For now, set skb->priority to 0 for other cases. */ diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index 7d4fe4a52929..631943e8af8b 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -257,7 +257,7 @@ struct sta_info *sta_info_alloc(struct ieee80211_sub_if_data *sdata, * sta_rx_agg_session_timer_expired for useage */ sta->timer_to_tid[i] = i; /* tid to tx queue: initialize according to HW (0 is valid) */ - sta->tid_to_tx_q[i] = local->hw.queues; + sta->tid_to_tx_q[i] = local->hw.queues + local->hw.ampdu_queues; /* rx */ sta->ampdu_mlme.tid_state_rx[i] = HT_AGG_STATE_IDLE; sta->ampdu_mlme.tid_rx[i] = NULL; diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 64faa3dc488f..5eddf1f32ed9 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -19,16 +19,22 @@ #include "wme.h" /* maximum number of hardware queues we support. */ -#define TC_80211_MAX_QUEUES 16 +#define QD_MAX_QUEUES (IEEE80211_MAX_AMPDU_QUEUES + IEEE80211_MAX_QUEUES) +/* current number of hardware queues we support. */ +#define QD_NUM(hw) ((hw)->queues + (hw)->ampdu_queues) +/* + * Default mapping in classifier to work with default + * queue setup. + */ const int ieee802_1d_to_ac[8] = { 2, 3, 3, 2, 1, 1, 0, 0 }; struct ieee80211_sched_data { - unsigned long qdisc_pool[BITS_TO_LONGS(TC_80211_MAX_QUEUES)]; + unsigned long qdisc_pool[BITS_TO_LONGS(QD_MAX_QUEUES)]; struct tcf_proto *filter_list; - struct Qdisc *queues[TC_80211_MAX_QUEUES]; - struct sk_buff_head requeued[TC_80211_MAX_QUEUES]; + struct Qdisc *queues[QD_MAX_QUEUES]; + struct sk_buff_head requeued[QD_MAX_QUEUES]; }; static const char llc_ip_hdr[8] = {0xAA, 0xAA, 0x3, 0, 0, 0, 0x08, 0}; @@ -95,7 +101,7 @@ static inline int wme_downgrade_ac(struct sk_buff *skb) /* positive return value indicates which queue to use * negative return value indicates to drop the frame */ -static inline int classify80211(struct sk_buff *skb, struct Qdisc *qd) +static int classify80211(struct sk_buff *skb, struct Qdisc *qd) { struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; @@ -106,7 +112,7 @@ static inline int classify80211(struct sk_buff *skb, struct Qdisc *qd) if (unlikely((fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_DATA)) { /* management frames go on AC_VO queue, but are sent * without QoS control fields */ - return IEEE80211_TX_QUEUE_DATA0; + return 0; } if (0 /* injected */) { @@ -141,14 +147,16 @@ static inline int classify80211(struct sk_buff *skb, struct Qdisc *qd) static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) { struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_hw *hw = &local->hw; struct ieee80211_sched_data *q = qdisc_priv(qd); struct ieee80211_tx_packet_data *pkt_data = (struct ieee80211_tx_packet_data *) skb->cb; struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; unsigned short fc = le16_to_cpu(hdr->frame_control); struct Qdisc *qdisc; - int err, queue; struct sta_info *sta; + int err; + u16 queue; u8 tid; if (pkt_data->flags & IEEE80211_TXPD_REQUEUE) { @@ -158,7 +166,7 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) tid = skb->priority & QOS_CONTROL_TAG1D_MASK; if (sta) { int ampdu_queue = sta->tid_to_tx_q[tid]; - if ((ampdu_queue < local->hw.queues) && + if ((ampdu_queue < QD_NUM(hw)) && test_bit(ampdu_queue, q->qdisc_pool)) { queue = ampdu_queue; pkt_data->flags |= IEEE80211_TXPD_AMPDU; @@ -174,6 +182,9 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) queue = classify80211(skb, qd); + if (unlikely(queue >= local->hw.queues)) + queue = local->hw.queues - 1; + /* now we know the 1d priority, fill in the QoS header if there is one */ if (WLAN_FC_IS_QOS_DATA(fc)) { @@ -193,8 +204,8 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) sta = sta_info_get(local, hdr->addr1); if (sta) { int ampdu_queue = sta->tid_to_tx_q[tid]; - if ((ampdu_queue < local->hw.queues) && - test_bit(ampdu_queue, q->qdisc_pool)) { + if ((ampdu_queue < QD_NUM(hw)) && + test_bit(ampdu_queue, q->qdisc_pool)) { queue = ampdu_queue; pkt_data->flags |= IEEE80211_TXPD_AMPDU; } else { @@ -205,17 +216,6 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) rcu_read_unlock(); } - if (unlikely(queue >= local->hw.queues)) { -#if 0 - if (net_ratelimit()) { - printk(KERN_DEBUG "%s - queue=%d (hw does not " - "support) -> %d\n", - __func__, queue, local->hw.queues - 1); - } -#endif - queue = local->hw.queues - 1; - } - if (unlikely(queue < 0)) { kfree_skb(skb); err = NET_XMIT_DROP; @@ -270,7 +270,7 @@ static struct sk_buff *wme_qdiscop_dequeue(struct Qdisc* qd) int queue; /* check all the h/w queues in numeric/priority order */ - for (queue = 0; queue < hw->queues; queue++) { + for (queue = 0; queue < QD_NUM(hw); queue++) { /* see if there is room in this hardware queue */ if ((test_bit(IEEE80211_LINK_STATE_XOFF, &local->state[queue])) || @@ -308,7 +308,7 @@ static void wme_qdiscop_reset(struct Qdisc* qd) /* QUESTION: should we have some hardware flush functionality here? */ - for (queue = 0; queue < hw->queues; queue++) { + for (queue = 0; queue < QD_NUM(hw); queue++) { skb_queue_purge(&q->requeued[queue]); qdisc_reset(q->queues[queue]); } @@ -326,7 +326,7 @@ static void wme_qdiscop_destroy(struct Qdisc* qd) tcf_destroy_chain(q->filter_list); q->filter_list = NULL; - for (queue=0; queue < hw->queues; queue++) { + for (queue = 0; queue < QD_NUM(hw); queue++) { skb_queue_purge(&q->requeued[queue]); qdisc_destroy(q->queues[queue]); q->queues[queue] = &noop_qdisc; @@ -337,17 +337,6 @@ static void wme_qdiscop_destroy(struct Qdisc* qd) /* called whenever parameters are updated on existing qdisc */ static int wme_qdiscop_tune(struct Qdisc *qd, struct nlattr *opt) { -/* struct ieee80211_sched_data *q = qdisc_priv(qd); -*/ - /* check our options block is the right size */ - /* copy any options to our local structure */ -/* Ignore options block for now - always use static mapping - struct tc_ieee80211_qopt *qopt = nla_data(opt); - - if (opt->nla_len < nla_attr_size(sizeof(*qopt))) - return -EINVAL; - memcpy(q->tag2queue, qopt->tag2queue, sizeof(qopt->tag2queue)); -*/ return 0; } @@ -358,7 +347,7 @@ static int wme_qdiscop_init(struct Qdisc *qd, struct nlattr *opt) struct ieee80211_sched_data *q = qdisc_priv(qd); struct net_device *dev = qd->dev; struct ieee80211_local *local; - int queues; + struct ieee80211_hw *hw; int err = 0, i; /* check that device is a mac80211 device */ @@ -366,29 +355,26 @@ static int wme_qdiscop_init(struct Qdisc *qd, struct nlattr *opt) dev->ieee80211_ptr->wiphy->privid != mac80211_wiphy_privid) return -EINVAL; - /* check this device is an ieee80211 master type device */ - if (dev->type != ARPHRD_IEEE80211) + local = wdev_priv(dev->ieee80211_ptr); + hw = &local->hw; + + /* only allow on master dev */ + if (dev != local->mdev) return -EINVAL; - /* check that there is no qdisc currently attached to device - * this ensures that we will be the root qdisc. (I can't find a better - * way to test this explicitly) */ - if (dev->qdisc_sleeping != &noop_qdisc) + /* ensure that we are root qdisc */ + if (qd->parent != TC_H_ROOT) return -EINVAL; if (qd->flags & TCQ_F_INGRESS) return -EINVAL; - local = wdev_priv(dev->ieee80211_ptr); - queues = local->hw.queues; - /* if options were passed in, set them */ - if (opt) { + if (opt) err = wme_qdiscop_tune(qd, opt); - } /* create child queues */ - for (i = 0; i < queues; i++) { + for (i = 0; i < QD_NUM(hw); i++) { skb_queue_head_init(&q->requeued[i]); q->queues[i] = qdisc_create_dflt(qd->dev, &pfifo_qdisc_ops, qd->handle); @@ -398,8 +384,8 @@ static int wme_qdiscop_init(struct Qdisc *qd, struct nlattr *opt) } } - /* reserve all legacy QoS queues */ - for (i = 0; i < min(IEEE80211_TX_QUEUE_DATA4, queues); i++) + /* non-aggregation queues: reserve/mark as used */ + for (i = 0; i < local->hw.queues; i++) set_bit(i, q->qdisc_pool); return err; @@ -407,16 +393,6 @@ static int wme_qdiscop_init(struct Qdisc *qd, struct nlattr *opt) static int wme_qdiscop_dump(struct Qdisc *qd, struct sk_buff *skb) { -/* struct ieee80211_sched_data *q = qdisc_priv(qd); - unsigned char *p = skb->tail; - struct tc_ieee80211_qopt opt; - - memcpy(&opt.tag2queue, q->tag2queue, TC_80211_MAX_TAG + 1); - NLA_PUT(skb, TCA_OPTIONS, sizeof(opt), &opt); -*/ return skb->len; -/* -nla_put_failure: - skb_trim(skb, p - skb->data);*/ return -1; } @@ -429,7 +405,7 @@ static int wme_classop_graft(struct Qdisc *qd, unsigned long arg, struct ieee80211_hw *hw = &local->hw; unsigned long queue = arg - 1; - if (queue >= hw->queues) + if (queue >= QD_NUM(hw)) return -EINVAL; if (!new) @@ -453,7 +429,7 @@ wme_classop_leaf(struct Qdisc *qd, unsigned long arg) struct ieee80211_hw *hw = &local->hw; unsigned long queue = arg - 1; - if (queue >= hw->queues) + if (queue >= QD_NUM(hw)) return NULL; return q->queues[queue]; @@ -466,7 +442,7 @@ static unsigned long wme_classop_get(struct Qdisc *qd, u32 classid) struct ieee80211_hw *hw = &local->hw; unsigned long queue = TC_H_MIN(classid); - if (queue - 1 >= hw->queues) + if (queue - 1 >= QD_NUM(hw)) return 0; return queue; @@ -492,7 +468,7 @@ static int wme_classop_change(struct Qdisc *qd, u32 handle, u32 parent, struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; - if (cl - 1 > hw->queues) + if (cl - 1 > QD_NUM(hw)) return -ENOENT; /* TODO: put code to program hardware queue parameters here, @@ -509,7 +485,7 @@ static int wme_classop_delete(struct Qdisc *qd, unsigned long cl) struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; - if (cl - 1 > hw->queues) + if (cl - 1 > QD_NUM(hw)) return -ENOENT; return 0; } @@ -522,7 +498,7 @@ static int wme_classop_dump_class(struct Qdisc *qd, unsigned long cl, struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; - if (cl - 1 > hw->queues) + if (cl - 1 > QD_NUM(hw)) return -ENOENT; tcm->tcm_handle = TC_H_MIN(cl); tcm->tcm_parent = qd->handle; @@ -540,7 +516,7 @@ static void wme_classop_walk(struct Qdisc *qd, struct qdisc_walker *arg) if (arg->stop) return; - for (queue = 0; queue < hw->queues; queue++) { + for (queue = 0; queue < QD_NUM(hw); queue++) { if (arg->count < arg->skip) { arg->count++; continue; @@ -657,10 +633,13 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, DECLARE_MAC_BUF(mac); /* prepare the filter and save it for the SW queue - * matching the recieved HW queue */ + * matching the received HW queue */ + + if (!local->hw.ampdu_queues) + return -EPERM; /* try to get a Qdisc from the pool */ - for (i = IEEE80211_TX_QUEUE_BEACON; i < local->hw.queues; i++) + for (i = local->hw.queues; i < QD_NUM(&local->hw); i++) if (!test_and_set_bit(i, q->qdisc_pool)) { ieee80211_stop_queue(local_to_hw(local), i); sta->tid_to_tx_q[tid] = i; @@ -689,13 +668,14 @@ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, struct sta_info *sta, u16 tid, u8 requeue) { + struct ieee80211_hw *hw = &local->hw; struct ieee80211_sched_data *q = qdisc_priv(local->mdev->qdisc_sleeping); int agg_queue = sta->tid_to_tx_q[tid]; /* return the qdisc to the pool */ clear_bit(agg_queue, q->qdisc_pool); - sta->tid_to_tx_q[tid] = local->hw.queues; + sta->tid_to_tx_q[tid] = QD_NUM(hw); if (requeue) ieee80211_requeue(local, agg_queue); -- cgit v1.2.2 From 3df5ee60f1ee559b1417397461891f8b483e8089 Mon Sep 17 00:00:00 2001 From: "John W. Linville" Date: Thu, 1 May 2008 17:07:32 -0400 Subject: wireless: fix warning introduced by "mac80211: QoS related cleanups" MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit net/mac80211/wme.c: In function ‘wme_qdiscop_enqueue’: net/mac80211/wme.c:219: warning: comparison is always false due to limited range of data type drivers/net/wireless/p54/p54common.c: In function ‘p54_conf_tx’: drivers/net/wireless/p54/p54common.c:947: warning: comparison is always false due to limited range of data type Signed-off-by: John W. Linville --- net/mac80211/wme.c | 23 +++++++++-------------- 1 file changed, 9 insertions(+), 14 deletions(-) (limited to 'net') diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 5eddf1f32ed9..b1e20ca03ffe 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -216,20 +216,15 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) rcu_read_unlock(); } - if (unlikely(queue < 0)) { - kfree_skb(skb); - err = NET_XMIT_DROP; - } else { - tid = skb->priority & QOS_CONTROL_TAG1D_MASK; - pkt_data->queue = (unsigned int) queue; - qdisc = q->queues[queue]; - err = qdisc->enqueue(skb, qdisc); - if (err == NET_XMIT_SUCCESS) { - qd->q.qlen++; - qd->bstats.bytes += skb->len; - qd->bstats.packets++; - return NET_XMIT_SUCCESS; - } + tid = skb->priority & QOS_CONTROL_TAG1D_MASK; + pkt_data->queue = (unsigned int) queue; + qdisc = q->queues[queue]; + err = qdisc->enqueue(skb, qdisc); + if (err == NET_XMIT_SUCCESS) { + qd->q.qlen++; + qd->bstats.bytes += skb->len; + qd->bstats.packets++; + return NET_XMIT_SUCCESS; } qd->qstats.drops++; return err; -- cgit v1.2.2 From 5c5e12898af0978a780991950be12d0d73c02f04 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Sat, 3 May 2008 00:44:09 +0200 Subject: mac80211: fix wme code In commit e100bb64bf7cdeae7f742a65ee1985649a7fd1b4 (mac80211: QoS related cleanups) I accidentally changed a variable from int to u16 causing a warning that a comparison for < 0 was always false. John thought this was a missing deletion of code and removed the warning by deleting the never executed branch of code in commit 3df5ee60f1ee559b1417397461891f8b483e8089 (wireless: fix warning introduced by "mac80211: QoS related cleanups") but the problem really was my mistake of using a u16 variable for the queue variable when that variable can also contain an error code. This patch restores the original code and variable type. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/wme.c | 26 +++++++++++++++----------- 1 file changed, 15 insertions(+), 11 deletions(-) (limited to 'net') diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index b1e20ca03ffe..8ffff27fe000 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -155,8 +155,7 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) unsigned short fc = le16_to_cpu(hdr->frame_control); struct Qdisc *qdisc; struct sta_info *sta; - int err; - u16 queue; + int err, queue; u8 tid; if (pkt_data->flags & IEEE80211_TXPD_REQUEUE) { @@ -216,15 +215,20 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) rcu_read_unlock(); } - tid = skb->priority & QOS_CONTROL_TAG1D_MASK; - pkt_data->queue = (unsigned int) queue; - qdisc = q->queues[queue]; - err = qdisc->enqueue(skb, qdisc); - if (err == NET_XMIT_SUCCESS) { - qd->q.qlen++; - qd->bstats.bytes += skb->len; - qd->bstats.packets++; - return NET_XMIT_SUCCESS; + if (unlikely(queue < 0)) { + kfree_skb(skb); + err = NET_XMIT_DROP; + } else { + tid = skb->priority & QOS_CONTROL_TAG1D_MASK; + pkt_data->queue = (unsigned int) queue; + qdisc = q->queues[queue]; + err = qdisc->enqueue(skb, qdisc); + if (err == NET_XMIT_SUCCESS) { + qd->q.qlen++; + qd->bstats.bytes += skb->len; + qd->bstats.packets++; + return NET_XMIT_SUCCESS; + } } qd->qstats.drops++; return err; -- cgit v1.2.2 From d364d9276b54af16fcb4db83f1315b620daec102 Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Fri, 9 May 2008 15:13:26 -0700 Subject: sctp: Bring SCTP_DELAYED_ACK socket option into API compliance Brings delayed_ack socket option set/get into line with the latest ietf socket extensions API draft, while maintaining backwards compatibility. Signed-off-by: Wei Yongjun Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/associola.c | 3 + net/sctp/sm_sideeffect.c | 17 +++- net/sctp/socket.c | 242 ++++++++++++++++++++++++++++++----------------- 3 files changed, 169 insertions(+), 93 deletions(-) (limited to 'net') diff --git a/net/sctp/associola.c b/net/sctp/associola.c index b4cd2b71953f..7b79d1e781a8 100644 --- a/net/sctp/associola.c +++ b/net/sctp/associola.c @@ -136,6 +136,7 @@ static struct sctp_association *sctp_association_init(struct sctp_association *a /* Set association default SACK delay */ asoc->sackdelay = msecs_to_jiffies(sp->sackdelay); + asoc->sackfreq = sp->sackfreq; /* Set the association default flags controlling * Heartbeat, SACK delay, and Path MTU Discovery. @@ -261,6 +262,7 @@ static struct sctp_association *sctp_association_init(struct sctp_association *a * already received one packet.] */ asoc->peer.sack_needed = 1; + asoc->peer.sack_cnt = 0; /* Assume that the peer will tell us if he recognizes ASCONF * as part of INIT exchange. @@ -615,6 +617,7 @@ struct sctp_transport *sctp_assoc_add_peer(struct sctp_association *asoc, * association configured value. */ peer->sackdelay = asoc->sackdelay; + peer->sackfreq = asoc->sackfreq; /* Enable/disable heartbeat, SACK delay, and path MTU discovery * based on association setting. diff --git a/net/sctp/sm_sideeffect.c b/net/sctp/sm_sideeffect.c index 23a9f1a95b7d..b083312c725a 100644 --- a/net/sctp/sm_sideeffect.c +++ b/net/sctp/sm_sideeffect.c @@ -190,20 +190,28 @@ static int sctp_gen_sack(struct sctp_association *asoc, int force, * unacknowledged DATA chunk. ... */ if (!asoc->peer.sack_needed) { - /* We will need a SACK for the next packet. */ - asoc->peer.sack_needed = 1; + asoc->peer.sack_cnt++; /* Set the SACK delay timeout based on the * SACK delay for the last transport * data was received from, or the default * for the association. */ - if (trans) + if (trans) { + /* We will need a SACK for the next packet. */ + if (asoc->peer.sack_cnt >= trans->sackfreq - 1) + asoc->peer.sack_needed = 1; + asoc->timeouts[SCTP_EVENT_TIMEOUT_SACK] = trans->sackdelay; - else + } else { + /* We will need a SACK for the next packet. */ + if (asoc->peer.sack_cnt >= asoc->sackfreq - 1) + asoc->peer.sack_needed = 1; + asoc->timeouts[SCTP_EVENT_TIMEOUT_SACK] = asoc->sackdelay; + } /* Restart the SACK timer. */ sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_RESTART, @@ -216,6 +224,7 @@ static int sctp_gen_sack(struct sctp_association *asoc, int force, goto nomem; asoc->peer.sack_needed = 0; + asoc->peer.sack_cnt = 0; sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(sack)); diff --git a/net/sctp/socket.c b/net/sctp/socket.c index e7e3baf7009e..66985871401b 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -2305,74 +2305,98 @@ static int sctp_setsockopt_peer_addr_params(struct sock *sk, return 0; } -/* 7.1.23. Delayed Ack Timer (SCTP_DELAYED_ACK_TIME) - * - * This options will get or set the delayed ack timer. The time is set - * in milliseconds. If the assoc_id is 0, then this sets or gets the - * endpoints default delayed ack timer value. If the assoc_id field is - * non-zero, then the set or get effects the specified association. - * - * struct sctp_assoc_value { - * sctp_assoc_t assoc_id; - * uint32_t assoc_value; - * }; - * - * assoc_id - This parameter, indicates which association the - * user is preforming an action upon. Note that if - * this field's value is zero then the endpoints - * default value is changed (effecting future - * associations only). - * - * assoc_value - This parameter contains the number of milliseconds - * that the user is requesting the delayed ACK timer - * be set to. Note that this value is defined in - * the standard to be between 200 and 500 milliseconds. - * - * Note: a value of zero will leave the value alone, - * but disable SACK delay. A non-zero value will also - * enable SACK delay. +/* + * 7.1.23. Get or set delayed ack timer (SCTP_DELAYED_SACK) + * + * This option will effect the way delayed acks are performed. This + * option allows you to get or set the delayed ack time, in + * milliseconds. It also allows changing the delayed ack frequency. + * Changing the frequency to 1 disables the delayed sack algorithm. If + * the assoc_id is 0, then this sets or gets the endpoints default + * values. If the assoc_id field is non-zero, then the set or get + * effects the specified association for the one to many model (the + * assoc_id field is ignored by the one to one model). Note that if + * sack_delay or sack_freq are 0 when setting this option, then the + * current values will remain unchanged. + * + * struct sctp_sack_info { + * sctp_assoc_t sack_assoc_id; + * uint32_t sack_delay; + * uint32_t sack_freq; + * }; + * + * sack_assoc_id - This parameter, indicates which association the user + * is performing an action upon. Note that if this field's value is + * zero then the endpoints default value is changed (effecting future + * associations only). + * + * sack_delay - This parameter contains the number of milliseconds that + * the user is requesting the delayed ACK timer be set to. Note that + * this value is defined in the standard to be between 200 and 500 + * milliseconds. + * + * sack_freq - This parameter contains the number of packets that must + * be received before a sack is sent without waiting for the delay + * timer to expire. The default value for this is 2, setting this + * value to 1 will disable the delayed sack algorithm. */ -static int sctp_setsockopt_delayed_ack_time(struct sock *sk, +static int sctp_setsockopt_delayed_ack(struct sock *sk, char __user *optval, int optlen) { - struct sctp_assoc_value params; + struct sctp_sack_info params; struct sctp_transport *trans = NULL; struct sctp_association *asoc = NULL; struct sctp_sock *sp = sctp_sk(sk); - if (optlen != sizeof(struct sctp_assoc_value)) - return - EINVAL; + if (optlen == sizeof(struct sctp_sack_info)) { + if (copy_from_user(¶ms, optval, optlen)) + return -EFAULT; - if (copy_from_user(¶ms, optval, optlen)) - return -EFAULT; + if (params.sack_delay == 0 && params.sack_freq == 0) + return 0; + } else if (optlen == sizeof(struct sctp_assoc_value)) { + printk(KERN_WARNING "SCTP: Use of struct sctp_sack_info " + "in delayed_ack socket option deprecated\n"); + printk(KERN_WARNING "SCTP: struct sctp_sack_info instead\n"); + if (copy_from_user(¶ms, optval, optlen)) + return -EFAULT; + + if (params.sack_delay == 0) + params.sack_freq = 1; + else + params.sack_freq = 0; + } else + return - EINVAL; /* Validate value parameter. */ - if (params.assoc_value > 500) + if (params.sack_delay > 500) return -EINVAL; - /* Get association, if assoc_id != 0 and the socket is a one + /* Get association, if sack_assoc_id != 0 and the socket is a one * to many style socket, and an association was not found, then * the id was invalid. */ - asoc = sctp_id2assoc(sk, params.assoc_id); - if (!asoc && params.assoc_id && sctp_style(sk, UDP)) + asoc = sctp_id2assoc(sk, params.sack_assoc_id); + if (!asoc && params.sack_assoc_id && sctp_style(sk, UDP)) return -EINVAL; - if (params.assoc_value) { + if (params.sack_delay) { if (asoc) { asoc->sackdelay = - msecs_to_jiffies(params.assoc_value); + msecs_to_jiffies(params.sack_delay); asoc->param_flags = (asoc->param_flags & ~SPP_SACKDELAY) | SPP_SACKDELAY_ENABLE; } else { - sp->sackdelay = params.assoc_value; + sp->sackdelay = params.sack_delay; sp->param_flags = (sp->param_flags & ~SPP_SACKDELAY) | SPP_SACKDELAY_ENABLE; } - } else { + } + + if (params.sack_freq == 1) { if (asoc) { asoc->param_flags = (asoc->param_flags & ~SPP_SACKDELAY) | @@ -2382,22 +2406,40 @@ static int sctp_setsockopt_delayed_ack_time(struct sock *sk, (sp->param_flags & ~SPP_SACKDELAY) | SPP_SACKDELAY_DISABLE; } + } else if (params.sack_freq > 1) { + if (asoc) { + asoc->sackfreq = params.sack_freq; + asoc->param_flags = + (asoc->param_flags & ~SPP_SACKDELAY) | + SPP_SACKDELAY_ENABLE; + } else { + sp->sackfreq = params.sack_freq; + sp->param_flags = + (sp->param_flags & ~SPP_SACKDELAY) | + SPP_SACKDELAY_ENABLE; + } } /* If change is for association, also apply to each transport. */ if (asoc) { list_for_each_entry(trans, &asoc->peer.transport_addr_list, transports) { - if (params.assoc_value) { + if (params.sack_delay) { trans->sackdelay = - msecs_to_jiffies(params.assoc_value); + msecs_to_jiffies(params.sack_delay); trans->param_flags = (trans->param_flags & ~SPP_SACKDELAY) | SPP_SACKDELAY_ENABLE; - } else { + } + if (params.sack_delay == 1) { trans->param_flags = (trans->param_flags & ~SPP_SACKDELAY) | SPP_SACKDELAY_DISABLE; + } else if (params.sack_freq > 1) { + trans->sackfreq = params.sack_freq; + trans->param_flags = + (trans->param_flags & ~SPP_SACKDELAY) | + SPP_SACKDELAY_ENABLE; } } } @@ -3186,8 +3228,8 @@ SCTP_STATIC int sctp_setsockopt(struct sock *sk, int level, int optname, retval = sctp_setsockopt_peer_addr_params(sk, optval, optlen); break; - case SCTP_DELAYED_ACK_TIME: - retval = sctp_setsockopt_delayed_ack_time(sk, optval, optlen); + case SCTP_DELAYED_ACK: + retval = sctp_setsockopt_delayed_ack(sk, optval, optlen); break; case SCTP_PARTIAL_DELIVERY_POINT: retval = sctp_setsockopt_partial_delivery_point(sk, optval, optlen); @@ -3446,6 +3488,7 @@ SCTP_STATIC int sctp_init_sock(struct sock *sk) sp->pathmaxrxt = sctp_max_retrans_path; sp->pathmtu = 0; // allow default discovery sp->sackdelay = sctp_sack_timeout; + sp->sackfreq = 3; sp->param_flags = SPP_HB_ENABLE | SPP_PMTUD_ENABLE | SPP_SACKDELAY_ENABLE; @@ -3999,70 +4042,91 @@ static int sctp_getsockopt_peer_addr_params(struct sock *sk, int len, return 0; } -/* 7.1.23. Delayed Ack Timer (SCTP_DELAYED_ACK_TIME) - * - * This options will get or set the delayed ack timer. The time is set - * in milliseconds. If the assoc_id is 0, then this sets or gets the - * endpoints default delayed ack timer value. If the assoc_id field is - * non-zero, then the set or get effects the specified association. - * - * struct sctp_assoc_value { - * sctp_assoc_t assoc_id; - * uint32_t assoc_value; - * }; +/* + * 7.1.23. Get or set delayed ack timer (SCTP_DELAYED_SACK) + * + * This option will effect the way delayed acks are performed. This + * option allows you to get or set the delayed ack time, in + * milliseconds. It also allows changing the delayed ack frequency. + * Changing the frequency to 1 disables the delayed sack algorithm. If + * the assoc_id is 0, then this sets or gets the endpoints default + * values. If the assoc_id field is non-zero, then the set or get + * effects the specified association for the one to many model (the + * assoc_id field is ignored by the one to one model). Note that if + * sack_delay or sack_freq are 0 when setting this option, then the + * current values will remain unchanged. + * + * struct sctp_sack_info { + * sctp_assoc_t sack_assoc_id; + * uint32_t sack_delay; + * uint32_t sack_freq; + * }; * - * assoc_id - This parameter, indicates which association the - * user is preforming an action upon. Note that if - * this field's value is zero then the endpoints - * default value is changed (effecting future - * associations only). + * sack_assoc_id - This parameter, indicates which association the user + * is performing an action upon. Note that if this field's value is + * zero then the endpoints default value is changed (effecting future + * associations only). * - * assoc_value - This parameter contains the number of milliseconds - * that the user is requesting the delayed ACK timer - * be set to. Note that this value is defined in - * the standard to be between 200 and 500 milliseconds. + * sack_delay - This parameter contains the number of milliseconds that + * the user is requesting the delayed ACK timer be set to. Note that + * this value is defined in the standard to be between 200 and 500 + * milliseconds. * - * Note: a value of zero will leave the value alone, - * but disable SACK delay. A non-zero value will also - * enable SACK delay. + * sack_freq - This parameter contains the number of packets that must + * be received before a sack is sent without waiting for the delay + * timer to expire. The default value for this is 2, setting this + * value to 1 will disable the delayed sack algorithm. */ -static int sctp_getsockopt_delayed_ack_time(struct sock *sk, int len, +static int sctp_getsockopt_delayed_ack(struct sock *sk, int len, char __user *optval, int __user *optlen) { - struct sctp_assoc_value params; + struct sctp_sack_info params; struct sctp_association *asoc = NULL; struct sctp_sock *sp = sctp_sk(sk); - if (len < sizeof(struct sctp_assoc_value)) - return - EINVAL; + if (len >= sizeof(struct sctp_sack_info)) { + len = sizeof(struct sctp_sack_info); - len = sizeof(struct sctp_assoc_value); - - if (copy_from_user(¶ms, optval, len)) - return -EFAULT; + if (copy_from_user(¶ms, optval, len)) + return -EFAULT; + } else if (len == sizeof(struct sctp_assoc_value)) { + printk(KERN_WARNING "SCTP: Use of struct sctp_sack_info " + "in delayed_ack socket option deprecated\n"); + printk(KERN_WARNING "SCTP: struct sctp_sack_info instead\n"); + if (copy_from_user(¶ms, optval, len)) + return -EFAULT; + } else + return - EINVAL; - /* Get association, if assoc_id != 0 and the socket is a one + /* Get association, if sack_assoc_id != 0 and the socket is a one * to many style socket, and an association was not found, then * the id was invalid. */ - asoc = sctp_id2assoc(sk, params.assoc_id); - if (!asoc && params.assoc_id && sctp_style(sk, UDP)) + asoc = sctp_id2assoc(sk, params.sack_assoc_id); + if (!asoc && params.sack_assoc_id && sctp_style(sk, UDP)) return -EINVAL; if (asoc) { /* Fetch association values. */ - if (asoc->param_flags & SPP_SACKDELAY_ENABLE) - params.assoc_value = jiffies_to_msecs( + if (asoc->param_flags & SPP_SACKDELAY_ENABLE) { + params.sack_delay = jiffies_to_msecs( asoc->sackdelay); - else - params.assoc_value = 0; + params.sack_freq = asoc->sackfreq; + + } else { + params.sack_delay = 0; + params.sack_freq = 1; + } } else { /* Fetch socket values. */ - if (sp->param_flags & SPP_SACKDELAY_ENABLE) - params.assoc_value = sp->sackdelay; - else - params.assoc_value = 0; + if (sp->param_flags & SPP_SACKDELAY_ENABLE) { + params.sack_delay = sp->sackdelay; + params.sack_freq = sp->sackfreq; + } else { + params.sack_delay = 0; + params.sack_freq = 1; + } } if (copy_to_user(optval, ¶ms, len)) @@ -5218,8 +5282,8 @@ SCTP_STATIC int sctp_getsockopt(struct sock *sk, int level, int optname, retval = sctp_getsockopt_peer_addr_params(sk, len, optval, optlen); break; - case SCTP_DELAYED_ACK_TIME: - retval = sctp_getsockopt_delayed_ack_time(sk, len, optval, + case SCTP_DELAYED_ACK: + retval = sctp_getsockopt_delayed_ack(sk, len, optval, optlen); break; case SCTP_INITMSG: -- cgit v1.2.2 From 88a0a948e752bb9b617a8c55417a9fd9b0257199 Mon Sep 17 00:00:00 2001 From: Vlad Yasevich Date: Fri, 9 May 2008 15:14:11 -0700 Subject: sctp: Support the new specification of sctp_connectx() The specification of sctp_connectx() has been changed to return an association id. We've added a new socket option that will return the association id as the return value from the setsockopt() call. The library that implements sctp_connectx() interface will implement both socket options. Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/socket.c | 70 ++++++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 59 insertions(+), 11 deletions(-) (limited to 'net') diff --git a/net/sctp/socket.c b/net/sctp/socket.c index 66985871401b..81600eea05d2 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -956,7 +956,8 @@ out: */ static int __sctp_connect(struct sock* sk, struct sockaddr *kaddrs, - int addrs_size) + int addrs_size, + sctp_assoc_t *assoc_id) { struct sctp_sock *sp; struct sctp_endpoint *ep; @@ -1111,6 +1112,8 @@ static int __sctp_connect(struct sock* sk, timeo = sock_sndtimeo(sk, f_flags & O_NONBLOCK); err = sctp_wait_for_connect(asoc, &timeo); + if (!err && assoc_id) + *assoc_id = asoc->assoc_id; /* Don't free association on exit. */ asoc = NULL; @@ -1128,7 +1131,8 @@ out_free: /* Helper for tunneling sctp_connectx() requests through sctp_setsockopt() * * API 8.9 - * int sctp_connectx(int sd, struct sockaddr *addrs, int addrcnt); + * int sctp_connectx(int sd, struct sockaddr *addrs, int addrcnt, + * sctp_assoc_t *asoc); * * If sd is an IPv4 socket, the addresses passed must be IPv4 addresses. * If the sd is an IPv6 socket, the addresses passed can either be IPv4 @@ -1144,8 +1148,10 @@ out_free: * representation is termed a "packed array" of addresses). The caller * specifies the number of addresses in the array with addrcnt. * - * On success, sctp_connectx() returns 0. On failure, sctp_connectx() returns - * -1, and sets errno to the appropriate error code. + * On success, sctp_connectx() returns 0. It also sets the assoc_id to + * the association id of the new association. On failure, sctp_connectx() + * returns -1, and sets errno to the appropriate error code. The assoc_id + * is not touched by the kernel. * * For SCTP, the port given in each socket address must be the same, or * sctp_connectx() will fail, setting errno to EINVAL. @@ -1182,11 +1188,12 @@ out_free: * addrs The pointer to the addresses in user land * addrssize Size of the addrs buffer * - * Returns 0 if ok, <0 errno code on error. + * Returns >=0 if ok, <0 errno code on error. */ -SCTP_STATIC int sctp_setsockopt_connectx(struct sock* sk, +SCTP_STATIC int __sctp_setsockopt_connectx(struct sock* sk, struct sockaddr __user *addrs, - int addrs_size) + int addrs_size, + sctp_assoc_t *assoc_id) { int err = 0; struct sockaddr *kaddrs; @@ -1209,13 +1216,46 @@ SCTP_STATIC int sctp_setsockopt_connectx(struct sock* sk, if (__copy_from_user(kaddrs, addrs, addrs_size)) { err = -EFAULT; } else { - err = __sctp_connect(sk, kaddrs, addrs_size); + err = __sctp_connect(sk, kaddrs, addrs_size, assoc_id); } kfree(kaddrs); + return err; } +/* + * This is an older interface. It's kept for backward compatibility + * to the option that doesn't provide association id. + */ +SCTP_STATIC int sctp_setsockopt_connectx_old(struct sock* sk, + struct sockaddr __user *addrs, + int addrs_size) +{ + return __sctp_setsockopt_connectx(sk, addrs, addrs_size, NULL); +} + +/* + * New interface for the API. The since the API is done with a socket + * option, to make it simple we feed back the association id is as a return + * indication to the call. Error is always negative and association id is + * always positive. + */ +SCTP_STATIC int sctp_setsockopt_connectx(struct sock* sk, + struct sockaddr __user *addrs, + int addrs_size) +{ + sctp_assoc_t assoc_id = 0; + int err = 0; + + err = __sctp_setsockopt_connectx(sk, addrs, addrs_size, &assoc_id); + + if (err) + return err; + else + return assoc_id; +} + /* API 3.1.4 close() - UDP Style Syntax * Applications use close() to perform graceful shutdown (as described in * Section 10.1 of [SCTP]) on ALL the associations currently represented @@ -3206,10 +3246,18 @@ SCTP_STATIC int sctp_setsockopt(struct sock *sk, int level, int optname, optlen, SCTP_BINDX_REM_ADDR); break; + case SCTP_SOCKOPT_CONNECTX_OLD: + /* 'optlen' is the size of the addresses buffer. */ + retval = sctp_setsockopt_connectx_old(sk, + (struct sockaddr __user *)optval, + optlen); + break; + case SCTP_SOCKOPT_CONNECTX: /* 'optlen' is the size of the addresses buffer. */ - retval = sctp_setsockopt_connectx(sk, (struct sockaddr __user *)optval, - optlen); + retval = sctp_setsockopt_connectx(sk, + (struct sockaddr __user *)optval, + optlen); break; case SCTP_DISABLE_FRAGMENTS: @@ -3336,7 +3384,7 @@ SCTP_STATIC int sctp_connect(struct sock *sk, struct sockaddr *addr, /* Pass correct addr len to common routine (so it knows there * is only one address being passed. */ - err = __sctp_connect(sk, addr, af->sockaddr_len); + err = __sctp_connect(sk, addr, af->sockaddr_len, NULL); } sctp_release_sock(sk); -- cgit v1.2.2 From 20c2c1fd6c842caf70dcb1d94b9d58861949fd3d Mon Sep 17 00:00:00 2001 From: Neil Horman Date: Fri, 9 May 2008 15:14:50 -0700 Subject: sctp: add sctp/remaddr table to complete RFC remote address table OID Add support for RFC3873 remote address table OID. +--(5) sctpAssocRemAddrTable | | | |--(-) sctpAssocId (shared index) | | | +--(1) sctpAssocRemAddrType (index) . | . +--(2) sctpAssocRemAddr (index) . | +--(3) sctpAssocRemAddrActive | +--(4) sctpAssocRemAddrHBActive | +--(5) sctpAssocRemAddrRTO | +--(6) sctpAssocRemAddrMaxPathRtx | +--(7) sctpAssocRemAddrRtx | +--(8) sctpAssocRemAddrStartTime This patch places all the requsite data in /proc/net/sctp/remaddr. Signed-off-by: Neil Horman Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/proc.c | 141 ++++++++++++++++++++++++++++++++++++++++++++++++++++ net/sctp/protocol.c | 3 ++ 2 files changed, 144 insertions(+) (limited to 'net') diff --git a/net/sctp/proc.c b/net/sctp/proc.c index 0aba759cb9b7..5dd89831eceb 100644 --- a/net/sctp/proc.c +++ b/net/sctp/proc.c @@ -383,3 +383,144 @@ void sctp_assocs_proc_exit(void) { remove_proc_entry("assocs", proc_net_sctp); } + +static void *sctp_remaddr_seq_start(struct seq_file *seq, loff_t *pos) +{ + if (*pos >= sctp_assoc_hashsize) + return NULL; + + if (*pos < 0) + *pos = 0; + + if (*pos == 0) + seq_printf(seq, "ADDR ASSOC_ID HB_ACT RTO MAX_PATH_RTX " + "REM_ADDR_RTX START\n"); + + return (void *)pos; +} + +static void *sctp_remaddr_seq_next(struct seq_file *seq, void *v, loff_t *pos) +{ + if (++*pos >= sctp_assoc_hashsize) + return NULL; + + return pos; +} + +static void sctp_remaddr_seq_stop(struct seq_file *seq, void *v) +{ + return; +} + +static int sctp_remaddr_seq_show(struct seq_file *seq, void *v) +{ + struct sctp_hashbucket *head; + struct sctp_ep_common *epb; + struct sctp_association *assoc; + struct hlist_node *node; + struct sctp_transport *tsp; + int hash = *(loff_t *)v; + + if (hash >= sctp_assoc_hashsize) + return -ENOMEM; + + head = &sctp_assoc_hashtable[hash]; + sctp_local_bh_disable(); + read_lock(&head->lock); + sctp_for_each_hentry(epb, node, &head->chain) { + assoc = sctp_assoc(epb); + list_for_each_entry(tsp, &assoc->peer.transport_addr_list, + transports) { + /* + * The remote address (ADDR) + */ + tsp->af_specific->seq_dump_addr(seq, &tsp->ipaddr); + seq_printf(seq, " "); + + /* + * The association ID (ASSOC_ID) + */ + seq_printf(seq, "%d ", tsp->asoc->assoc_id); + + /* + * If the Heartbeat is active (HB_ACT) + * Note: 1 = Active, 0 = Inactive + */ + seq_printf(seq, "%d ", timer_pending(&tsp->hb_timer)); + + /* + * Retransmit time out (RTO) + */ + seq_printf(seq, "%lu ", tsp->rto); + + /* + * Maximum path retransmit count (PATH_MAX_RTX) + */ + seq_printf(seq, "%d ", tsp->pathmaxrxt); + + /* + * remote address retransmit count (REM_ADDR_RTX) + * Note: We don't have a way to tally this at the moment + * so lets just leave it as zero for the moment + */ + seq_printf(seq, "0 "); + + /* + * remote address start time (START). This is also not + * currently implemented, but we can record it with a + * jiffies marker in a subsequent patch + */ + seq_printf(seq, "0"); + + seq_printf(seq, "\n"); + } + } + + read_unlock(&head->lock); + sctp_local_bh_enable(); + + return 0; + +} + +static const struct seq_operations sctp_remaddr_ops = { + .start = sctp_remaddr_seq_start, + .next = sctp_remaddr_seq_next, + .stop = sctp_remaddr_seq_stop, + .show = sctp_remaddr_seq_show, +}; + +/* Cleanup the proc fs entry for 'remaddr' object. */ +void sctp_remaddr_proc_exit(void) +{ + remove_proc_entry("remaddr", proc_net_sctp); +} + +static int sctp_remaddr_seq_open(struct inode *inode, struct file *file) +{ + return seq_open(file, &sctp_remaddr_ops); +} + +static const struct file_operations sctp_remaddr_seq_fops = { + .open = sctp_remaddr_seq_open, + .read = seq_read, + .llseek = seq_lseek, + .release = seq_release, +}; + +int __init sctp_remaddr_proc_init(void) +{ + struct proc_dir_entry *p; + + p = create_proc_entry("remaddr", S_IRUGO, proc_net_sctp); + if (!p) + return -ENOMEM; + p->proc_fops = &sctp_remaddr_seq_fops; + + return 0; +} + +void sctp_assoc_proc_exit(void) +{ + remove_proc_entry("remaddr", proc_net_sctp); +} diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c index 0ec234b762c2..b8bd9e014498 100644 --- a/net/sctp/protocol.c +++ b/net/sctp/protocol.c @@ -113,6 +113,8 @@ static __init int sctp_proc_init(void) goto out_nomem; if (sctp_assocs_proc_init()) goto out_nomem; + if (sctp_remaddr_proc_init()) + goto out_nomem; return 0; @@ -129,6 +131,7 @@ static void sctp_proc_exit(void) sctp_snmp_proc_exit(); sctp_eps_proc_exit(); sctp_assocs_proc_exit(); + sctp_remaddr_proc_exit(); if (proc_net_sctp) { proc_net_sctp = NULL; -- cgit v1.2.2 From 4e3e6dcb43c3669a8817cb3d0f920f91661afd98 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 12 May 2008 15:41:53 -0700 Subject: tipc: Enhancements to name table initialization This patch enhances the initialization of TIPC's name table by removing a pointless spinlock operation, and by using kcalloc() to detect requests for an oversized name table. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/name_table.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/tipc/name_table.c b/net/tipc/name_table.c index ac7dfdda7973..892373e498e4 100644 --- a/net/tipc/name_table.c +++ b/net/tipc/name_table.c @@ -1050,15 +1050,12 @@ void tipc_nametbl_dump(void) int tipc_nametbl_init(void) { - int array_size = sizeof(struct hlist_head) * tipc_nametbl_size; - - table.types = kzalloc(array_size, GFP_ATOMIC); + table.types = kcalloc(tipc_nametbl_size, sizeof(struct hlist_head), + GFP_ATOMIC); if (!table.types) return -ENOMEM; - write_lock_bh(&tipc_nametbl_lock); table.local_publ_count = 0; - write_unlock_bh(&tipc_nametbl_lock); return 0; } -- cgit v1.2.2 From 7ef43ebaa538e0cc9063cbf84593a05091bcace2 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 12 May 2008 15:42:28 -0700 Subject: tipc: Fix race condition when creating socket or native port This patch eliminates the (very remote) chance of a crash resulting from a partially initialized socket or native port unexpectedly receiving a message. Now, during the creation of a socket or native port, the underlying generic port's lock is not released until all initialization required to handle incoming messages has been done. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/port.c | 20 ++++++++++++-------- net/tipc/ref.c | 12 ++++++++---- net/tipc/socket.c | 5 ++++- net/tipc/subscr.c | 1 + 4 files changed, 25 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/tipc/port.c b/net/tipc/port.c index 2f5806410c64..757de38fe6af 100644 --- a/net/tipc/port.c +++ b/net/tipc/port.c @@ -211,15 +211,18 @@ exit: } /** - * tipc_createport_raw - create a native TIPC port + * tipc_createport_raw - create a generic TIPC port * - * Returns local port reference + * Returns port reference, or 0 if unable to create it + * + * Note: The newly created port is returned in the locked state. */ u32 tipc_createport_raw(void *usr_handle, u32 (*dispatcher)(struct tipc_port *, struct sk_buff *), void (*wakeup)(struct tipc_port *), - const u32 importance) + const u32 importance, + struct tipc_port **tp_ptr) { struct port *p_ptr; struct tipc_msg *msg; @@ -237,7 +240,6 @@ u32 tipc_createport_raw(void *usr_handle, return 0; } - tipc_port_lock(ref); p_ptr->publ.usr_handle = usr_handle; p_ptr->publ.max_pkt = MAX_PKT_DEFAULT; p_ptr->publ.ref = ref; @@ -262,7 +264,7 @@ u32 tipc_createport_raw(void *usr_handle, INIT_LIST_HEAD(&p_ptr->port_list); list_add_tail(&p_ptr->port_list, &ports); spin_unlock_bh(&tipc_port_list_lock); - tipc_port_unlock(p_ptr); + *tp_ptr = &p_ptr->publ; return ref; } @@ -1053,6 +1055,7 @@ int tipc_createport(u32 user_ref, { struct user_port *up_ptr; struct port *p_ptr; + struct tipc_port *tp_ptr; u32 ref; up_ptr = kmalloc(sizeof(*up_ptr), GFP_ATOMIC); @@ -1060,12 +1063,13 @@ int tipc_createport(u32 user_ref, warn("Port creation failed, no memory\n"); return -ENOMEM; } - ref = tipc_createport_raw(NULL, port_dispatcher, port_wakeup, importance); - p_ptr = tipc_port_lock(ref); - if (!p_ptr) { + ref = tipc_createport_raw(NULL, port_dispatcher, port_wakeup, + importance, &tp_ptr); + if (ref == 0) { kfree(up_ptr); return -ENOMEM; } + p_ptr = (struct port *)tp_ptr; p_ptr->user_port = up_ptr; up_ptr->user_ref = user_ref; diff --git a/net/tipc/ref.c b/net/tipc/ref.c index 89cbab24d08f..a101de86824d 100644 --- a/net/tipc/ref.c +++ b/net/tipc/ref.c @@ -142,9 +142,13 @@ void tipc_ref_table_stop(void) /** * tipc_ref_acquire - create reference to an object * - * Return a unique reference value which can be translated back to the pointer - * 'object' at a later time. Also, pass back a pointer to the lock protecting - * the object, but without locking it. + * Register an object pointer in reference table and lock the object. + * Returns a unique reference value that is used from then on to retrieve the + * object pointer, or to determine that the object has been deregistered. + * + * Note: The object is returned in the locked state so that the caller can + * register a partially initialized object, without running the risk that + * the object will be accessed before initialization is complete. */ u32 tipc_ref_acquire(void *object, spinlock_t **lock) @@ -178,13 +182,13 @@ u32 tipc_ref_acquire(void *object, spinlock_t **lock) ref = (next_plus_upper & ~index_mask) + index; entry->ref = ref; entry->object = object; - spin_unlock_bh(&entry->lock); *lock = &entry->lock; } else if (tipc_ref_table.init_point < tipc_ref_table.capacity) { index = tipc_ref_table.init_point++; entry = &(tipc_ref_table.entries[index]); spin_lock_init(&entry->lock); + spin_lock_bh(&entry->lock); ref = tipc_ref_table.start_mask + index; entry->ref = ref; entry->object = object; diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 230f9ca2ad6b..38f48795b40e 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -188,6 +188,7 @@ static int tipc_create(struct net *net, struct socket *sock, int protocol) const struct proto_ops *ops; socket_state state; struct sock *sk; + struct tipc_port *tp_ptr; u32 portref; /* Validate arguments */ @@ -225,7 +226,7 @@ static int tipc_create(struct net *net, struct socket *sock, int protocol) /* Allocate TIPC port for socket to use */ portref = tipc_createport_raw(sk, &dispatch, &wakeupdispatch, - TIPC_LOW_IMPORTANCE); + TIPC_LOW_IMPORTANCE, &tp_ptr); if (unlikely(portref == 0)) { sk_free(sk); return -ENOMEM; @@ -241,6 +242,8 @@ static int tipc_create(struct net *net, struct socket *sock, int protocol) sk->sk_backlog_rcv = backlog_rcv; tipc_sk(sk)->p = tipc_get_port(portref); + spin_unlock_bh(tp_ptr->lock); + if (sock->state == SS_READY) { tipc_set_portunreturnable(portref, 1); if (sock->type == SOCK_DGRAM) diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c index 8c01ccd3626c..8f8d0a6c1c16 100644 --- a/net/tipc/subscr.c +++ b/net/tipc/subscr.c @@ -474,6 +474,7 @@ static void subscr_named_msg_event(void *usr_handle, kfree(subscriber); return; } + spin_unlock_bh(subscriber->lock); /* Establish a connection to subscriber */ -- cgit v1.2.2 From ae7245cbf27ee6b6423bc363cbe01c93e57befda Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Thu, 1 May 2008 22:19:33 -0700 Subject: wireless: use get/put_unaligned_* helpers Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/mesh_hwmp.c | 2 +- net/wireless/radiotap.c | 16 +++++++--------- 2 files changed, 8 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/mac80211/mesh_hwmp.c b/net/mac80211/mesh_hwmp.c index 3df809222d1c..23689733e29a 100644 --- a/net/mac80211/mesh_hwmp.c +++ b/net/mac80211/mesh_hwmp.c @@ -26,7 +26,7 @@ static inline u32 u32_field_get(u8 *preq_elem, int offset, bool ae) { if (ae) offset += 6; - return le32_to_cpu(get_unaligned((__le32 *) (preq_elem + offset))); + return get_unaligned_le32(preq_elem + offset); } /* HWMP IE processing macros */ diff --git a/net/wireless/radiotap.c b/net/wireless/radiotap.c index 28fbd0b0b568..f591871a7b4f 100644 --- a/net/wireless/radiotap.c +++ b/net/wireless/radiotap.c @@ -59,23 +59,21 @@ int ieee80211_radiotap_iterator_init( return -EINVAL; /* sanity check for allowed length and radiotap length field */ - if (max_length < le16_to_cpu(get_unaligned(&radiotap_header->it_len))) + if (max_length < get_unaligned_le16(&radiotap_header->it_len)) return -EINVAL; iterator->rtheader = radiotap_header; - iterator->max_length = le16_to_cpu(get_unaligned( - &radiotap_header->it_len)); + iterator->max_length = get_unaligned_le16(&radiotap_header->it_len); iterator->arg_index = 0; - iterator->bitmap_shifter = le32_to_cpu(get_unaligned( - &radiotap_header->it_present)); + iterator->bitmap_shifter = get_unaligned_le32(&radiotap_header->it_present); iterator->arg = (u8 *)radiotap_header + sizeof(*radiotap_header); iterator->this_arg = NULL; /* find payload start allowing for extended bitmap(s) */ if (unlikely(iterator->bitmap_shifter & (1<arg)) & - (1<arg) & + (1 << IEEE80211_RADIOTAP_EXT)) { iterator->arg += sizeof(u32); /* @@ -241,8 +239,8 @@ int ieee80211_radiotap_iterator_next( if (iterator->bitmap_shifter & 1) { /* b31 was set, there is more */ /* move to next u32 bitmap */ - iterator->bitmap_shifter = le32_to_cpu( - get_unaligned(iterator->next_bitmap)); + iterator->bitmap_shifter = + get_unaligned_le32(iterator->next_bitmap); iterator->next_bitmap++; } else /* no more bitmaps: end */ -- cgit v1.2.2 From 8c046c8c64ba81dd87468ddaf2db4a5d926b988b Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Fri, 2 May 2008 13:47:45 -0700 Subject: mac80211: tkip.c use kernel-provided infrastructure Use kernel-provided bit rotation and unaligned access infrastructure rather than opencoding it. Some minor spacing adjustments as well. Signed-off-by: Harvey Harrison Cc: "John W. Linville" Signed-off-by: Andrew Morton Signed-off-by: John W. Linville --- net/mac80211/tkip.c | 145 +++++++++++++++++----------------------------------- 1 file changed, 46 insertions(+), 99 deletions(-) (limited to 'net') diff --git a/net/mac80211/tkip.c b/net/mac80211/tkip.c index 09093da24af6..a7c3febc5a45 100644 --- a/net/mac80211/tkip.c +++ b/net/mac80211/tkip.c @@ -8,23 +8,22 @@ */ #include +#include #include #include +#include #include #include "key.h" #include "tkip.h" #include "wep.h" - -/* TKIP key mixing functions */ - - #define PHASE1_LOOP_COUNT 8 - -/* 2-byte by 2-byte subset of the full AES S-box table; second part of this - * table is identical to first part but byte-swapped */ +/* + * 2-byte by 2-byte subset of the full AES S-box table; second part of this + * table is identical to first part but byte-swapped + */ static const u16 tkip_sbox[256] = { 0xC6A5, 0xF884, 0xEE99, 0xF68D, 0xFF0D, 0xD6BD, 0xDEB1, 0x9154, @@ -61,53 +60,13 @@ static const u16 tkip_sbox[256] = 0x82C3, 0x29B0, 0x5A77, 0x1E11, 0x7BCB, 0xA8FC, 0x6DD6, 0x2C3A, }; - -static inline u16 Mk16(u8 x, u8 y) -{ - return ((u16) x << 8) | (u16) y; -} - - -static inline u8 Hi8(u16 v) -{ - return v >> 8; -} - - -static inline u8 Lo8(u16 v) -{ - return v & 0xff; -} - - -static inline u16 Hi16(u32 v) -{ - return v >> 16; -} - - -static inline u16 Lo16(u32 v) -{ - return v & 0xffff; -} - - -static inline u16 RotR1(u16 v) +static u16 tkipS(u16 val) { - return (v >> 1) | ((v & 0x0001) << 15); -} - - -static inline u16 tkip_S(u16 val) -{ - u16 a = tkip_sbox[Hi8(val)]; - - return tkip_sbox[Lo8(val)] ^ Hi8(a) ^ (Lo8(a) << 8); + return tkip_sbox[val & 0xff] ^ swab16(tkip_sbox[val >> 8]); } - - -/* P1K := Phase1(TA, TK, TSC) +/* + * P1K := Phase1(TA, TK, TSC) * TA = transmitter address (48 bits) * TK = dot11DefaultKeyValue or dot11KeyMappingValue (128 bits) * TSC = TKIP sequence counter (48 bits, only 32 msb bits used) @@ -118,23 +77,22 @@ static void tkip_mixing_phase1(const u8 *ta, const u8 *tk, u32 tsc_IV32, { int i, j; - p1k[0] = Lo16(tsc_IV32); - p1k[1] = Hi16(tsc_IV32); - p1k[2] = Mk16(ta[1], ta[0]); - p1k[3] = Mk16(ta[3], ta[2]); - p1k[4] = Mk16(ta[5], ta[4]); + p1k[0] = tsc_IV32 & 0xFFFF; + p1k[1] = tsc_IV32 >> 16; + p1k[2] = get_unaligned_le16(ta + 0); + p1k[3] = get_unaligned_le16(ta + 2); + p1k[4] = get_unaligned_le16(ta + 4); for (i = 0; i < PHASE1_LOOP_COUNT; i++) { j = 2 * (i & 1); - p1k[0] += tkip_S(p1k[4] ^ Mk16(tk[ 1 + j], tk[ 0 + j])); - p1k[1] += tkip_S(p1k[0] ^ Mk16(tk[ 5 + j], tk[ 4 + j])); - p1k[2] += tkip_S(p1k[1] ^ Mk16(tk[ 9 + j], tk[ 8 + j])); - p1k[3] += tkip_S(p1k[2] ^ Mk16(tk[13 + j], tk[12 + j])); - p1k[4] += tkip_S(p1k[3] ^ Mk16(tk[ 1 + j], tk[ 0 + j])) + i; + p1k[0] += tkipS(p1k[4] ^ get_unaligned_le16(tk + 0 + j)); + p1k[1] += tkipS(p1k[0] ^ get_unaligned_le16(tk + 4 + j)); + p1k[2] += tkipS(p1k[1] ^ get_unaligned_le16(tk + 8 + j)); + p1k[3] += tkipS(p1k[2] ^ get_unaligned_le16(tk + 12 + j)); + p1k[4] += tkipS(p1k[3] ^ get_unaligned_le16(tk + 0 + j)) + i; } } - static void tkip_mixing_phase2(const u16 *p1k, const u8 *tk, u16 tsc_IV16, u8 *rc4key) { @@ -148,31 +106,29 @@ static void tkip_mixing_phase2(const u16 *p1k, const u8 *tk, u16 tsc_IV16, ppk[4] = p1k[4]; ppk[5] = p1k[4] + tsc_IV16; - ppk[0] += tkip_S(ppk[5] ^ Mk16(tk[ 1], tk[ 0])); - ppk[1] += tkip_S(ppk[0] ^ Mk16(tk[ 3], tk[ 2])); - ppk[2] += tkip_S(ppk[1] ^ Mk16(tk[ 5], tk[ 4])); - ppk[3] += tkip_S(ppk[2] ^ Mk16(tk[ 7], tk[ 6])); - ppk[4] += tkip_S(ppk[3] ^ Mk16(tk[ 9], tk[ 8])); - ppk[5] += tkip_S(ppk[4] ^ Mk16(tk[11], tk[10])); - ppk[0] += RotR1(ppk[5] ^ Mk16(tk[13], tk[12])); - ppk[1] += RotR1(ppk[0] ^ Mk16(tk[15], tk[14])); - ppk[2] += RotR1(ppk[1]); - ppk[3] += RotR1(ppk[2]); - ppk[4] += RotR1(ppk[3]); - ppk[5] += RotR1(ppk[4]); - - rc4key[0] = Hi8(tsc_IV16); - rc4key[1] = (Hi8(tsc_IV16) | 0x20) & 0x7f; - rc4key[2] = Lo8(tsc_IV16); - rc4key[3] = Lo8((ppk[5] ^ Mk16(tk[1], tk[0])) >> 1); - - for (i = 0; i < 6; i++) { - rc4key[4 + 2 * i] = Lo8(ppk[i]); - rc4key[5 + 2 * i] = Hi8(ppk[i]); - } + ppk[0] += tkipS(ppk[5] ^ get_unaligned_le16(tk + 0)); + ppk[1] += tkipS(ppk[0] ^ get_unaligned_le16(tk + 2)); + ppk[2] += tkipS(ppk[1] ^ get_unaligned_le16(tk + 4)); + ppk[3] += tkipS(ppk[2] ^ get_unaligned_le16(tk + 6)); + ppk[4] += tkipS(ppk[3] ^ get_unaligned_le16(tk + 8)); + ppk[5] += tkipS(ppk[4] ^ get_unaligned_le16(tk + 10)); + ppk[0] += ror16(ppk[5] ^ get_unaligned_le16(tk + 12), 1); + ppk[1] += ror16(ppk[0] ^ get_unaligned_le16(tk + 14), 1); + ppk[2] += ror16(ppk[1], 1); + ppk[3] += ror16(ppk[2], 1); + ppk[4] += ror16(ppk[3], 1); + ppk[5] += ror16(ppk[4], 1); + + rc4key[0] = tsc_IV16 >> 8; + rc4key[1] = ((tsc_IV16 >> 8) | 0x20) & 0x7f; + rc4key[2] = tsc_IV16 & 0xFF; + rc4key[3] = ((ppk[5] ^ get_unaligned_le16(tk)) >> 1) & 0xFF; + + rc4key += 4; + for (i = 0; i < 6; i++) + put_unaligned_le16(ppk[i], rc4key + 2 * i); } - /* Add TKIP IV and Ext. IV at @pos. @iv0, @iv1, and @iv2 are the first octets * of the IV. Returns pointer to the octet following IVs (i.e., beginning of * the packet payload). */ @@ -183,14 +139,10 @@ u8 * ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, *pos++ = iv1; *pos++ = iv2; *pos++ = (key->conf.keyidx << 6) | (1 << 5) /* Ext IV */; - *pos++ = key->u.tkip.iv32 & 0xff; - *pos++ = (key->u.tkip.iv32 >> 8) & 0xff; - *pos++ = (key->u.tkip.iv32 >> 16) & 0xff; - *pos++ = (key->u.tkip.iv32 >> 24) & 0xff; - return pos; + put_unaligned_le32(key->u.tkip.iv32, pos); + return pos + 4; } - void ieee80211_tkip_gen_phase1key(struct ieee80211_key *key, u8 *ta, u16 *phase1key) { @@ -228,10 +180,8 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, u16 iv16; u32 iv32; - iv16 = data[hdr_len] << 8; - iv16 += data[hdr_len + 2]; - iv32 = data[hdr_len + 4] | (data[hdr_len + 5] << 8) | - (data[hdr_len + 6] << 16) | (data[hdr_len + 7] << 24); + iv16 = data[hdr_len + 2] | (data[hdr_len] << 8); + iv32 = get_unaligned_le32(data + hdr_len + 4); #ifdef CONFIG_TKIP_DEBUG printk(KERN_DEBUG "TKIP encrypt: iv16 = 0x%04x, iv32 = 0x%08x\n", @@ -281,7 +231,6 @@ void ieee80211_tkip_encrypt_data(struct crypto_blkcipher *tfm, ieee80211_wep_encrypt_data(tfm, rc4key, 16, pos, payload_len); } - /* Decrypt packet payload with TKIP using @key. @pos is a pointer to the * beginning of the buffer containing IEEE 802.11 header payload, i.e., * including IV, Ext. IV, real data, Michael MIC, ICV. @payload_len is the @@ -302,7 +251,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, iv16 = (pos[0] << 8) | pos[2]; keyid = pos[3]; - iv32 = pos[4] | (pos[5] << 8) | (pos[6] << 16) | (pos[7] << 24); + iv32 = get_unaligned_le32(pos + 4); pos += 8; #ifdef CONFIG_TKIP_DEBUG { @@ -409,5 +358,3 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, return res; } - - -- cgit v1.2.2 From 3434fbd39862d471c92b66c28cd449deea8e9f90 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Sat, 3 May 2008 00:59:37 +0200 Subject: mac80211: require four hardware queues for QoS/HT This patch makes mac80211 only announce QoS/HT support when the underlying hardware has four (or more) queues. Signed-off-by: Johannes Berg Cc: Ron Rindjunksi Signed-off-by: John W. Linville --- net/mac80211/iface.c | 3 ++- net/mac80211/mlme.c | 21 ++++++++++++++------- net/mac80211/tx.c | 4 ++-- 3 files changed, 18 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c index f41c7e0de622..5a9a3c6ef482 100644 --- a/net/mac80211/iface.c +++ b/net/mac80211/iface.c @@ -157,9 +157,10 @@ void ieee80211_if_set_type(struct net_device *dev, int type) ifsta->auth_algs = IEEE80211_AUTH_ALG_OPEN | IEEE80211_AUTH_ALG_SHARED_KEY; ifsta->flags |= IEEE80211_STA_CREATE_IBSS | - IEEE80211_STA_WMM_ENABLED | IEEE80211_STA_AUTO_BSSID_SEL | IEEE80211_STA_AUTO_CHANNEL_SEL; + if (sdata->local->hw.queues >= 4) + ifsta->flags |= IEEE80211_STA_WMM_ENABLED; msdata = IEEE80211_DEV_TO_SUB_IF(sdata->local->mdev); sdata->bss = &msdata->u.ap; diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 55b85ae5bc11..108c6fc42fe2 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -272,6 +272,12 @@ static void ieee80211_sta_wmm_params(struct net_device *dev, int count; u8 *pos; + if (!(ifsta->flags & IEEE80211_STA_WMM_ENABLED)) + return; + + if (!wmm_param) + return; + if (wmm_param_len < 8 || wmm_param[5] /* version */ != 1) return; count = wmm_param[6] & 0x0f; @@ -763,8 +769,10 @@ static void ieee80211_send_assoc(struct net_device *dev, *pos++ = 1; /* WME ver */ *pos++ = 0; } + /* wmm support is a must to HT */ - if (wmm && sband->ht_info.ht_supported) { + if (wmm && (ifsta->flags & IEEE80211_STA_WMM_ENABLED) && + sband->ht_info.ht_supported) { __le16 tmp = cpu_to_le16(sband->ht_info.cap); pos = skb_put(skb, sizeof(struct ieee80211_ht_cap)+2); *pos++ = WLAN_EID_HT_CAPABILITY; @@ -2021,7 +2029,8 @@ static void ieee80211_rx_mgmt_assoc_resp(struct ieee80211_sub_if_data *sdata, else sdata->flags &= ~IEEE80211_SDATA_OPERATING_GMODE; - if (elems.ht_cap_elem && elems.ht_info_elem && elems.wmm_param) { + if (elems.ht_cap_elem && elems.ht_info_elem && elems.wmm_param && + (ifsta->flags & IEEE80211_STA_WMM_ENABLED)) { struct ieee80211_ht_bss_info bss_info; ieee80211_ht_cap_ie_to_ht_info( (struct ieee80211_ht_cap *) @@ -2034,7 +2043,7 @@ static void ieee80211_rx_mgmt_assoc_resp(struct ieee80211_sub_if_data *sdata, rate_control_rate_init(sta, local); - if (elems.wmm_param && (ifsta->flags & IEEE80211_STA_WMM_ENABLED)) { + if (elems.wmm_param) { sta->flags |= WLAN_STA_WME; rcu_read_unlock(); ieee80211_sta_wmm_params(dev, ifsta, elems.wmm_param, @@ -2817,10 +2826,8 @@ static void ieee80211_rx_mgmt_beacon(struct net_device *dev, ieee802_11_parse_elems(mgmt->u.beacon.variable, len - baselen, &elems); - if (elems.wmm_param && (ifsta->flags & IEEE80211_STA_WMM_ENABLED)) { - ieee80211_sta_wmm_params(dev, ifsta, elems.wmm_param, - elems.wmm_param_len); - } + ieee80211_sta_wmm_params(dev, ifsta, elems.wmm_param, + elems.wmm_param_len); /* Do not send changes to driver if we are scanning. This removes * requirement that driver's bss_info_changed function needs to be diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index f35eaea98e73..3a03d7807c39 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -1490,8 +1490,8 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, rcu_read_unlock(); } - /* receiver is QoS enabled, use a QoS type frame */ - if (sta_flags & WLAN_STA_WME) { + /* receiver and we are QoS enabled, use a QoS type frame */ + if (sta_flags & WLAN_STA_WME && local->hw.queues >= 4) { fc |= IEEE80211_STYPE_QOS_DATA; hdrlen += 2; } -- cgit v1.2.2 From 07346f81e87d6e4cca7ae9adfa711d0c61c87b56 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Sat, 3 May 2008 01:02:02 +0200 Subject: mac80211: proper STA info locking As discussed earlier, we can unify locking in struct sta_info and use just a single spinlock protecting all members of the structure that need protection. Many don't, but one of the especially bad ones is the 'flags' member that can currently be clobbered when RX and TX is being processed on different CPUs at the same time. Because having four spinlocks for different, mostly exclusive parts of a single structure is overkill, this patch also kills the ampdu and mesh plink spinlocks and uses just a single one for everything. Because none of the spinlocks are nested, this is safe. It remains to be seen whether or not we should make the sta flags use atomic bit operations instead, for now though this is a safe thing and using atomic operations instead will be very simple using the new static inline functions this patch introduces for accessing sta->flags. Since spin_lock_bh() is used with this lock, there shouldn't be any contention even if aggregation is enabled at around the same time as both requires frame transmission/reception which is in a bh context. Signed-off-by: Johannes Berg Cc: Tomas Winkler Cc: Ron Rindjunsky Cc: Luis Carlos Cobo Signed-off-by: John W. Linville --- net/mac80211/cfg.c | 2 ++ net/mac80211/debugfs_sta.c | 15 ++++---- net/mac80211/key.c | 4 +-- net/mac80211/main.c | 31 ++++++++-------- net/mac80211/mesh_plink.c | 88 +++++++++++++++++++++++----------------------- net/mac80211/mlme.c | 38 ++++++++++---------- net/mac80211/rx.c | 17 ++++----- net/mac80211/sta_info.c | 14 +++----- net/mac80211/sta_info.h | 72 ++++++++++++++++++++++++++++++++----- net/mac80211/tx.c | 21 +++++------ 10 files changed, 179 insertions(+), 123 deletions(-) (limited to 'net') diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c index 699d97b8de5e..3cef80dcd0e5 100644 --- a/net/mac80211/cfg.c +++ b/net/mac80211/cfg.c @@ -602,6 +602,7 @@ static void sta_apply_parameters(struct ieee80211_local *local, */ if (params->station_flags & STATION_FLAG_CHANGED) { + spin_lock_bh(&sta->lock); sta->flags &= ~WLAN_STA_AUTHORIZED; if (params->station_flags & STATION_FLAG_AUTHORIZED) sta->flags |= WLAN_STA_AUTHORIZED; @@ -613,6 +614,7 @@ static void sta_apply_parameters(struct ieee80211_local *local, sta->flags &= ~WLAN_STA_WME; if (params->station_flags & STATION_FLAG_WME) sta->flags |= WLAN_STA_WME; + spin_unlock_bh(&sta->lock); } /* diff --git a/net/mac80211/debugfs_sta.c b/net/mac80211/debugfs_sta.c index 676a93202ff9..1f00273d99c2 100644 --- a/net/mac80211/debugfs_sta.c +++ b/net/mac80211/debugfs_sta.c @@ -74,14 +74,15 @@ static ssize_t sta_flags_read(struct file *file, char __user *userbuf, { char buf[100]; struct sta_info *sta = file->private_data; + u32 staflags = get_sta_flags(sta); int res = scnprintf(buf, sizeof(buf), "%s%s%s%s%s%s%s", - sta->flags & WLAN_STA_AUTH ? "AUTH\n" : "", - sta->flags & WLAN_STA_ASSOC ? "ASSOC\n" : "", - sta->flags & WLAN_STA_PS ? "PS\n" : "", - sta->flags & WLAN_STA_AUTHORIZED ? "AUTHORIZED\n" : "", - sta->flags & WLAN_STA_SHORT_PREAMBLE ? "SHORT PREAMBLE\n" : "", - sta->flags & WLAN_STA_WME ? "WME\n" : "", - sta->flags & WLAN_STA_WDS ? "WDS\n" : ""); + staflags & WLAN_STA_AUTH ? "AUTH\n" : "", + staflags & WLAN_STA_ASSOC ? "ASSOC\n" : "", + staflags & WLAN_STA_PS ? "PS\n" : "", + staflags & WLAN_STA_AUTHORIZED ? "AUTHORIZED\n" : "", + staflags & WLAN_STA_SHORT_PREAMBLE ? "SHORT PREAMBLE\n" : "", + staflags & WLAN_STA_WME ? "WME\n" : "", + staflags & WLAN_STA_WDS ? "WDS\n" : ""); return simple_read_from_buffer(userbuf, count, ppos, buf, res); } STA_OPS(flags); diff --git a/net/mac80211/key.c b/net/mac80211/key.c index 88b211af7c1f..d4893bd17754 100644 --- a/net/mac80211/key.c +++ b/net/mac80211/key.c @@ -321,7 +321,7 @@ void ieee80211_key_link(struct ieee80211_key *key, * some hardware cannot handle TKIP with QoS, so * we indicate whether QoS could be in use. */ - if (sta->flags & WLAN_STA_WME) + if (test_sta_flags(sta, WLAN_STA_WME)) key->conf.flags |= IEEE80211_KEY_FLAG_WMM_STA; /* @@ -342,7 +342,7 @@ void ieee80211_key_link(struct ieee80211_key *key, /* same here, the AP could be using QoS */ ap = sta_info_get(key->local, key->sdata->u.sta.bssid); if (ap) { - if (ap->flags & WLAN_STA_WME) + if (test_sta_flags(ap, WLAN_STA_WME)) key->conf.flags |= IEEE80211_KEY_FLAG_WMM_STA; } diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 55e76117da9e..f277407f0f5a 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -346,6 +346,7 @@ static int ieee80211_open(struct net_device *dev) goto err_del_interface; } + /* no locking required since STA is not live yet */ sta->flags |= WLAN_STA_AUTHORIZED; res = sta_info_insert(sta); @@ -588,7 +589,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) return -ENOENT; } - spin_lock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_lock_bh(&sta->lock); /* we have tried too many times, receiver does not want A-MPDU */ if (sta->ampdu_mlme.addba_req_num[tid] > HT_AGG_MAX_RETRIES) { @@ -691,7 +692,7 @@ start_ba_err: spin_unlock_bh(&local->mdev->queue_lock); ret = -EBUSY; start_ba_exit: - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); rcu_read_unlock(); return ret; } @@ -719,7 +720,7 @@ int ieee80211_stop_tx_ba_session(struct ieee80211_hw *hw, /* check if the TID is in aggregation */ state = &sta->ampdu_mlme.tid_state_tx[tid]; - spin_lock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_lock_bh(&sta->lock); if (*state != HT_AGG_STATE_OPERATIONAL) { ret = -ENOENT; @@ -749,7 +750,7 @@ int ieee80211_stop_tx_ba_session(struct ieee80211_hw *hw, } stop_BA_exit: - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); rcu_read_unlock(); return ret; } @@ -778,12 +779,12 @@ void ieee80211_start_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u16 tid) } state = &sta->ampdu_mlme.tid_state_tx[tid]; - spin_lock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_lock_bh(&sta->lock); if (!(*state & HT_ADDBA_REQUESTED_MSK)) { printk(KERN_DEBUG "addBA was not requested yet, state is %d\n", *state); - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); rcu_read_unlock(); return; } @@ -796,7 +797,7 @@ void ieee80211_start_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u16 tid) printk(KERN_DEBUG "Aggregation is on for tid %d \n", tid); ieee80211_wake_queue(hw, sta->tid_to_tx_q[tid]); } - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); rcu_read_unlock(); } EXPORT_SYMBOL(ieee80211_start_tx_ba_cb); @@ -830,10 +831,10 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) } state = &sta->ampdu_mlme.tid_state_tx[tid]; - spin_lock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_lock_bh(&sta->lock); if ((*state & HT_AGG_STATE_REQ_STOP_BA_MSK) == 0) { printk(KERN_DEBUG "unexpected callback to A-MPDU stop\n"); - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); rcu_read_unlock(); return; } @@ -860,7 +861,7 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) sta->ampdu_mlme.addba_req_num[tid] = 0; kfree(sta->ampdu_mlme.tid_tx[tid]); sta->ampdu_mlme.tid_tx[tid] = NULL; - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); rcu_read_unlock(); } @@ -1315,7 +1316,7 @@ static void ieee80211_handle_filtered_frame(struct ieee80211_local *local, * packet. If the STA went to power save mode, this will happen * happen when it wakes up for the next time. */ - sta->flags |= WLAN_STA_CLEAR_PS_FILT; + set_sta_flags(sta, WLAN_STA_CLEAR_PS_FILT); /* * This code races in the following way: @@ -1347,7 +1348,7 @@ static void ieee80211_handle_filtered_frame(struct ieee80211_local *local, * can be unknown, for example with different interrupt status * bits. */ - if (sta->flags & WLAN_STA_PS && + if (test_sta_flags(sta, WLAN_STA_PS) && skb_queue_len(&sta->tx_filtered) < STA_MAX_TX_BUFFER) { ieee80211_remove_tx_extra(local, sta->key, skb, &status->control); @@ -1355,7 +1356,7 @@ static void ieee80211_handle_filtered_frame(struct ieee80211_local *local, return; } - if (!(sta->flags & WLAN_STA_PS) && + if (!test_sta_flags(sta, WLAN_STA_PS) && !(status->control.flags & IEEE80211_TXCTL_REQUEUE)) { /* Software retry the packet once */ status->control.flags |= IEEE80211_TXCTL_REQUEUE; @@ -1370,7 +1371,7 @@ static void ieee80211_handle_filtered_frame(struct ieee80211_local *local, "queue_len=%d PS=%d @%lu\n", wiphy_name(local->hw.wiphy), skb_queue_len(&sta->tx_filtered), - !!(sta->flags & WLAN_STA_PS), jiffies); + !!test_sta_flags(sta, WLAN_STA_PS), jiffies); dev_kfree_skb(skb); } @@ -1399,7 +1400,7 @@ void ieee80211_tx_status(struct ieee80211_hw *hw, struct sk_buff *skb, struct sta_info *sta; sta = sta_info_get(local, hdr->addr1); if (sta) { - if (sta->flags & WLAN_STA_PS) { + if (test_sta_flags(sta, WLAN_STA_PS)) { /* * The STA is in power save mode, so assume * that this TX packet failed because of that. diff --git a/net/mac80211/mesh_plink.c b/net/mac80211/mesh_plink.c index 37f0c2b94ae7..9efeb1f07025 100644 --- a/net/mac80211/mesh_plink.c +++ b/net/mac80211/mesh_plink.c @@ -79,7 +79,7 @@ void mesh_plink_dec_estab_count(struct ieee80211_sub_if_data *sdata) * * @sta: mes peer link to restart * - * Locking: this function must be called holding sta->plink_lock + * Locking: this function must be called holding sta->lock */ static inline void mesh_plink_fsm_restart(struct sta_info *sta) { @@ -105,7 +105,7 @@ static struct sta_info *mesh_plink_alloc(struct ieee80211_sub_if_data *sdata, if (!sta) return NULL; - sta->flags |= WLAN_STA_AUTHORIZED; + sta->flags = WLAN_STA_AUTHORIZED; sta->supp_rates[local->hw.conf.channel->band] = rates; return sta; @@ -118,7 +118,7 @@ static struct sta_info *mesh_plink_alloc(struct ieee80211_sub_if_data *sdata, * * All mesh paths with this peer as next hop will be flushed * - * Locking: the caller must hold sta->plink_lock + * Locking: the caller must hold sta->lock */ static void __mesh_plink_deactivate(struct sta_info *sta) { @@ -139,9 +139,9 @@ static void __mesh_plink_deactivate(struct sta_info *sta) */ void mesh_plink_deactivate(struct sta_info *sta) { - spin_lock_bh(&sta->plink_lock); + spin_lock_bh(&sta->lock); __mesh_plink_deactivate(sta); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); } static int mesh_plink_frame_tx(struct net_device *dev, @@ -270,10 +270,10 @@ static void mesh_plink_timer(unsigned long data) */ sta = (struct sta_info *) data; - spin_lock_bh(&sta->plink_lock); + spin_lock_bh(&sta->lock); if (sta->ignore_plink_timer) { sta->ignore_plink_timer = false; - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); return; } mpl_dbg("Mesh plink timer for %s fired on state %d\n", @@ -298,7 +298,7 @@ static void mesh_plink_timer(unsigned long data) rand % sta->plink_timeout; ++sta->plink_retries; mod_plink_timer(sta, sta->plink_timeout); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(dev, PLINK_OPEN, sta->addr, llid, 0, 0); break; @@ -311,7 +311,7 @@ static void mesh_plink_timer(unsigned long data) reason = cpu_to_le16(MESH_CONFIRM_TIMEOUT); sta->plink_state = PLINK_HOLDING; mod_plink_timer(sta, dot11MeshHoldingTimeout(sdata)); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(dev, PLINK_CLOSE, sta->addr, llid, plid, reason); break; @@ -319,10 +319,10 @@ static void mesh_plink_timer(unsigned long data) /* holding timer */ del_timer(&sta->plink_timer); mesh_plink_fsm_restart(sta); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); break; default: - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); break; } } @@ -344,16 +344,16 @@ int mesh_plink_open(struct sta_info *sta) DECLARE_MAC_BUF(mac); #endif - spin_lock_bh(&sta->plink_lock); + spin_lock_bh(&sta->lock); get_random_bytes(&llid, 2); sta->llid = llid; if (sta->plink_state != PLINK_LISTEN) { - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); return -EBUSY; } sta->plink_state = PLINK_OPN_SNT; mesh_plink_timer_set(sta, dot11MeshRetryTimeout(sdata)); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mpl_dbg("Mesh plink: starting establishment with %s\n", print_mac(mac, sta->addr)); @@ -367,10 +367,10 @@ void mesh_plink_block(struct sta_info *sta) DECLARE_MAC_BUF(mac); #endif - spin_lock_bh(&sta->plink_lock); + spin_lock_bh(&sta->lock); __mesh_plink_deactivate(sta); sta->plink_state = PLINK_BLOCKED; - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); } int mesh_plink_close(struct sta_info *sta) @@ -383,14 +383,14 @@ int mesh_plink_close(struct sta_info *sta) mpl_dbg("Mesh plink: closing link with %s\n", print_mac(mac, sta->addr)); - spin_lock_bh(&sta->plink_lock); + spin_lock_bh(&sta->lock); sta->reason = cpu_to_le16(MESH_LINK_CANCELLED); reason = sta->reason; if (sta->plink_state == PLINK_LISTEN || sta->plink_state == PLINK_BLOCKED) { mesh_plink_fsm_restart(sta); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); return 0; } else if (sta->plink_state == PLINK_ESTAB) { __mesh_plink_deactivate(sta); @@ -402,7 +402,7 @@ int mesh_plink_close(struct sta_info *sta) sta->plink_state = PLINK_HOLDING; llid = sta->llid; plid = sta->plid; - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(sta->sdata->dev, PLINK_CLOSE, sta->addr, llid, plid, reason); return 0; @@ -490,7 +490,7 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, /* avoid warning */ break; } - spin_lock_bh(&sta->plink_lock); + spin_lock_bh(&sta->lock); } else if (!sta) { /* ftype == PLINK_OPEN */ u64 rates; @@ -512,9 +512,9 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, return; } event = OPN_ACPT; - spin_lock_bh(&sta->plink_lock); + spin_lock_bh(&sta->lock); } else { - spin_lock_bh(&sta->plink_lock); + spin_lock_bh(&sta->lock); switch (ftype) { case PLINK_OPEN: if (!mesh_plink_free_count(sdata) || @@ -551,7 +551,7 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, break; default: mpl_dbg("Mesh plink: unknown frame subtype\n"); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); rcu_read_unlock(); return; } @@ -568,7 +568,7 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, switch (event) { case CLS_ACPT: mesh_plink_fsm_restart(sta); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); break; case OPN_ACPT: sta->plink_state = PLINK_OPN_RCVD; @@ -576,14 +576,14 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, get_random_bytes(&llid, 2); sta->llid = llid; mesh_plink_timer_set(sta, dot11MeshRetryTimeout(sdata)); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(dev, PLINK_OPEN, sta->addr, llid, 0, 0); mesh_plink_frame_tx(dev, PLINK_CONFIRM, sta->addr, llid, plid, 0); break; default: - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); break; } break; @@ -603,7 +603,7 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, sta->ignore_plink_timer = true; llid = sta->llid; - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(dev, PLINK_CLOSE, sta->addr, llid, plid, reason); break; @@ -612,7 +612,7 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, sta->plink_state = PLINK_OPN_RCVD; sta->plid = plid; llid = sta->llid; - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(dev, PLINK_CONFIRM, sta->addr, llid, plid, 0); break; @@ -622,10 +622,10 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, dot11MeshConfirmTimeout(sdata))) sta->ignore_plink_timer = true; - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); break; default: - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); break; } break; @@ -645,13 +645,13 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, sta->ignore_plink_timer = true; llid = sta->llid; - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(dev, PLINK_CLOSE, sta->addr, llid, plid, reason); break; case OPN_ACPT: llid = sta->llid; - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(dev, PLINK_CONFIRM, sta->addr, llid, plid, 0); break; @@ -659,12 +659,12 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, del_timer(&sta->plink_timer); sta->plink_state = PLINK_ESTAB; mesh_plink_inc_estab_count(sdata); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mpl_dbg("Mesh plink with %s ESTABLISHED\n", print_mac(mac, sta->addr)); break; default: - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); break; } break; @@ -684,7 +684,7 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, sta->ignore_plink_timer = true; llid = sta->llid; - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(dev, PLINK_CLOSE, sta->addr, llid, plid, reason); break; @@ -692,14 +692,14 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, del_timer(&sta->plink_timer); sta->plink_state = PLINK_ESTAB; mesh_plink_inc_estab_count(sdata); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mpl_dbg("Mesh plink with %s ESTABLISHED\n", print_mac(mac, sta->addr)); mesh_plink_frame_tx(dev, PLINK_CONFIRM, sta->addr, llid, plid, 0); break; default: - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); break; } break; @@ -713,18 +713,18 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, sta->plink_state = PLINK_HOLDING; llid = sta->llid; mod_plink_timer(sta, dot11MeshHoldingTimeout(sdata)); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(dev, PLINK_CLOSE, sta->addr, llid, plid, reason); break; case OPN_ACPT: llid = sta->llid; - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(dev, PLINK_CONFIRM, sta->addr, llid, plid, 0); break; default: - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); break; } break; @@ -734,7 +734,7 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, if (del_timer(&sta->plink_timer)) sta->ignore_plink_timer = 1; mesh_plink_fsm_restart(sta); - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); break; case OPN_ACPT: case CNF_ACPT: @@ -742,19 +742,19 @@ void mesh_rx_plink_frame(struct net_device *dev, struct ieee80211_mgmt *mgmt, case CNF_RJCT: llid = sta->llid; reason = sta->reason; - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); mesh_plink_frame_tx(dev, PLINK_CLOSE, sta->addr, llid, plid, reason); break; default: - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); } break; default: /* should not get here, PLINK_BLOCKED is dealt with at the * beggining of the function */ - spin_unlock_bh(&sta->plink_lock); + spin_unlock_bh(&sta->lock); break; } diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 108c6fc42fe2..6c4b27be35da 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -1241,7 +1241,7 @@ static void ieee80211_sta_process_addba_request(struct net_device *dev, /* examine state machine */ - spin_lock_bh(&sta->ampdu_mlme.ampdu_rx); + spin_lock_bh(&sta->lock); if (sta->ampdu_mlme.tid_state_rx[tid] != HT_AGG_STATE_IDLE) { #ifdef CONFIG_MAC80211_HT_DEBUG @@ -1308,7 +1308,7 @@ static void ieee80211_sta_process_addba_request(struct net_device *dev, tid_agg_rx->stored_mpdu_num = 0; status = WLAN_STATUS_SUCCESS; end: - spin_unlock_bh(&sta->ampdu_mlme.ampdu_rx); + spin_unlock_bh(&sta->lock); end_no_lock: ieee80211_send_addba_resp(sta->sdata->dev, sta->addr, tid, @@ -1340,10 +1340,10 @@ static void ieee80211_sta_process_addba_resp(struct net_device *dev, state = &sta->ampdu_mlme.tid_state_tx[tid]; - spin_lock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_lock_bh(&sta->lock); if (!(*state & HT_ADDBA_REQUESTED_MSK)) { - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); printk(KERN_DEBUG "state not HT_ADDBA_REQUESTED_MSK:" "%d\n", *state); goto addba_resp_exit; @@ -1351,7 +1351,7 @@ static void ieee80211_sta_process_addba_resp(struct net_device *dev, if (mgmt->u.action.u.addba_resp.dialog_token != sta->ampdu_mlme.tid_tx[tid]->dialog_token) { - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); #ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "wrong addBA response token, tid %d\n", tid); #endif /* CONFIG_MAC80211_HT_DEBUG */ @@ -1375,7 +1375,7 @@ static void ieee80211_sta_process_addba_resp(struct net_device *dev, ieee80211_wake_queue(hw, sta->tid_to_tx_q[tid]); } - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); printk(KERN_DEBUG "recipient accepted agg: tid %d \n", tid); } else { printk(KERN_DEBUG "recipient rejected agg: tid %d \n", tid); @@ -1383,7 +1383,7 @@ static void ieee80211_sta_process_addba_resp(struct net_device *dev, sta->ampdu_mlme.addba_req_num[tid]++; /* this will allow the state check in stop_BA_session */ *state = HT_AGG_STATE_OPERATIONAL; - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); ieee80211_stop_tx_ba_session(hw, sta->addr, tid, WLAN_BACK_INITIATOR); } @@ -1453,17 +1453,17 @@ void ieee80211_sta_stop_rx_ba_session(struct net_device *dev, u8 *ra, u16 tid, } /* check if TID is in operational state */ - spin_lock_bh(&sta->ampdu_mlme.ampdu_rx); + spin_lock_bh(&sta->lock); if (sta->ampdu_mlme.tid_state_rx[tid] != HT_AGG_STATE_OPERATIONAL) { - spin_unlock_bh(&sta->ampdu_mlme.ampdu_rx); + spin_unlock_bh(&sta->lock); rcu_read_unlock(); return; } sta->ampdu_mlme.tid_state_rx[tid] = HT_AGG_STATE_REQ_STOP_BA_MSK | (initiator << HT_AGG_STATE_INITIATOR_SHIFT); - spin_unlock_bh(&sta->ampdu_mlme.ampdu_rx); + spin_unlock_bh(&sta->lock); /* stop HW Rx aggregation. ampdu_action existence * already verified in session init so we add the BUG_ON */ @@ -1540,10 +1540,10 @@ static void ieee80211_sta_process_delba(struct net_device *dev, ieee80211_sta_stop_rx_ba_session(dev, sta->addr, tid, WLAN_BACK_INITIATOR, 0); else { /* WLAN_BACK_RECIPIENT */ - spin_lock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_lock_bh(&sta->lock); sta->ampdu_mlme.tid_state_tx[tid] = HT_AGG_STATE_OPERATIONAL; - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); ieee80211_stop_tx_ba_session(&local->hw, sta->addr, tid, WLAN_BACK_RECIPIENT); } @@ -1580,9 +1580,9 @@ void sta_addba_resp_timer_expired(unsigned long data) state = &sta->ampdu_mlme.tid_state_tx[tid]; /* check if the TID waits for addBA response */ - spin_lock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_lock_bh(&sta->lock); if (!(*state & HT_ADDBA_REQUESTED_MSK)) { - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); *state = HT_AGG_STATE_IDLE; printk(KERN_DEBUG "timer expired on tid %d but we are not " "expecting addBA response there", tid); @@ -1593,7 +1593,7 @@ void sta_addba_resp_timer_expired(unsigned long data) /* go through the state check in stop_BA_session */ *state = HT_AGG_STATE_OPERATIONAL; - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); ieee80211_stop_tx_ba_session(hw, temp_sta->addr, tid, WLAN_BACK_INITIATOR); @@ -1984,8 +1984,8 @@ static void ieee80211_rx_mgmt_assoc_resp(struct ieee80211_sub_if_data *sdata, * to between the sta_info_alloc() and sta_info_insert() above. */ - sta->flags |= WLAN_STA_AUTH | WLAN_STA_ASSOC | WLAN_STA_ASSOC_AP | - WLAN_STA_AUTHORIZED; + set_sta_flags(sta, WLAN_STA_AUTH | WLAN_STA_ASSOC | WLAN_STA_ASSOC_AP | + WLAN_STA_AUTHORIZED); rates = 0; basic_rates = 0; @@ -2044,7 +2044,7 @@ static void ieee80211_rx_mgmt_assoc_resp(struct ieee80211_sub_if_data *sdata, rate_control_rate_init(sta, local); if (elems.wmm_param) { - sta->flags |= WLAN_STA_WME; + set_sta_flags(sta, WLAN_STA_WME); rcu_read_unlock(); ieee80211_sta_wmm_params(dev, ifsta, elems.wmm_param, elems.wmm_param_len); @@ -4237,7 +4237,7 @@ struct sta_info *ieee80211_ibss_add_sta(struct net_device *dev, if (!sta) return NULL; - sta->flags |= WLAN_STA_AUTHORIZED; + set_sta_flags(sta, WLAN_STA_AUTHORIZED); sta->supp_rates[local->hw.conf.channel->band] = sdata->u.sta.supp_rates_bits[local->hw.conf.channel->band]; diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index e8b89c89e875..b399e09ec7aa 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -479,7 +479,7 @@ ieee80211_rx_h_check(struct ieee80211_rx_data *rx) ((rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_CTL && (rx->fc & IEEE80211_FCTL_STYPE) == IEEE80211_STYPE_PSPOLL)) && rx->sdata->vif.type != IEEE80211_IF_TYPE_IBSS && - (!rx->sta || !(rx->sta->flags & WLAN_STA_ASSOC)))) { + (!rx->sta || !test_sta_flags(rx->sta, WLAN_STA_ASSOC)))) { if ((!(rx->fc & IEEE80211_FCTL_FROMDS) && !(rx->fc & IEEE80211_FCTL_TODS) && (rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA) @@ -630,8 +630,7 @@ static void ap_sta_ps_start(struct net_device *dev, struct sta_info *sta) if (sdata->bss) atomic_inc(&sdata->bss->num_sta_ps); - sta->flags |= WLAN_STA_PS; - sta->flags &= ~WLAN_STA_PSPOLL; + set_and_clear_sta_flags(sta, WLAN_STA_PS, WLAN_STA_PSPOLL); #ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG printk(KERN_DEBUG "%s: STA %s aid %d enters power save mode\n", dev->name, print_mac(mac, sta->addr), sta->aid); @@ -652,7 +651,7 @@ static int ap_sta_ps_end(struct net_device *dev, struct sta_info *sta) if (sdata->bss) atomic_dec(&sdata->bss->num_sta_ps); - sta->flags &= ~(WLAN_STA_PS | WLAN_STA_PSPOLL); + clear_sta_flags(sta, WLAN_STA_PS | WLAN_STA_PSPOLL); if (!skb_queue_empty(&sta->ps_tx_buf)) sta_info_clear_tim_bit(sta); @@ -727,9 +726,10 @@ ieee80211_rx_h_sta_process(struct ieee80211_rx_data *rx) if (!(rx->fc & IEEE80211_FCTL_MOREFRAGS)) { /* Change STA power saving mode only in the end of a frame * exchange sequence */ - if ((sta->flags & WLAN_STA_PS) && !(rx->fc & IEEE80211_FCTL_PM)) + if (test_sta_flags(sta, WLAN_STA_PS) && + !(rx->fc & IEEE80211_FCTL_PM)) rx->sent_ps_buffered += ap_sta_ps_end(dev, sta); - else if (!(sta->flags & WLAN_STA_PS) && + else if (!test_sta_flags(sta, WLAN_STA_PS) && (rx->fc & IEEE80211_FCTL_PM)) ap_sta_ps_start(dev, sta); } @@ -983,7 +983,7 @@ ieee80211_rx_h_ps_poll(struct ieee80211_rx_data *rx) * Tell TX path to send one frame even though the STA may * still remain is PS mode after this frame exchange. */ - rx->sta->flags |= WLAN_STA_PSPOLL; + set_sta_flags(rx->sta, WLAN_STA_PSPOLL); #ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG printk(KERN_DEBUG "STA %s aid %d: PS Poll (entries after %d)\n", @@ -1046,7 +1046,8 @@ ieee80211_rx_h_remove_qos_control(struct ieee80211_rx_data *rx) static int ieee80211_802_1x_port_control(struct ieee80211_rx_data *rx) { - if (unlikely(!rx->sta || !(rx->sta->flags & WLAN_STA_AUTHORIZED))) { + if (unlikely(!rx->sta || + !test_sta_flags(rx->sta, WLAN_STA_AUTHORIZED))) { #ifdef CONFIG_MAC80211_DEBUG if (net_ratelimit()) printk(KERN_DEBUG "%s: dropped frame " diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index 631943e8af8b..baf5e4746884 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -202,14 +202,12 @@ void sta_info_destroy(struct sta_info *sta) dev_kfree_skb_any(skb); for (i = 0; i < STA_TID_NUM; i++) { - spin_lock_bh(&sta->ampdu_mlme.ampdu_rx); + spin_lock_bh(&sta->lock); if (sta->ampdu_mlme.tid_rx[i]) del_timer_sync(&sta->ampdu_mlme.tid_rx[i]->session_timer); - spin_unlock_bh(&sta->ampdu_mlme.ampdu_rx); - spin_lock_bh(&sta->ampdu_mlme.ampdu_tx); if (sta->ampdu_mlme.tid_tx[i]) del_timer_sync(&sta->ampdu_mlme.tid_tx[i]->addba_resp_timer); - spin_unlock_bh(&sta->ampdu_mlme.ampdu_tx); + spin_unlock_bh(&sta->lock); } __sta_info_free(local, sta); @@ -236,6 +234,8 @@ struct sta_info *sta_info_alloc(struct ieee80211_sub_if_data *sdata, if (!sta) return NULL; + spin_lock_init(&sta->lock); + memcpy(sta->addr, addr, ETH_ALEN); sta->local = local; sta->sdata = sdata; @@ -249,8 +249,6 @@ struct sta_info *sta_info_alloc(struct ieee80211_sub_if_data *sdata, return NULL; } - spin_lock_init(&sta->ampdu_mlme.ampdu_rx); - spin_lock_init(&sta->ampdu_mlme.ampdu_tx); for (i = 0; i < STA_TID_NUM; i++) { /* timer_to_tid must be initialized with identity mapping to * enable session_timer's data differentiation. refer to @@ -276,7 +274,6 @@ struct sta_info *sta_info_alloc(struct ieee80211_sub_if_data *sdata, #ifdef CONFIG_MAC80211_MESH sta->plink_state = PLINK_LISTEN; - spin_lock_init(&sta->plink_lock); init_timer(&sta->plink_timer); #endif @@ -437,8 +434,7 @@ void __sta_info_unlink(struct sta_info **sta) list_del(&(*sta)->list); - if ((*sta)->flags & WLAN_STA_PS) { - (*sta)->flags &= ~WLAN_STA_PS; + if (test_and_clear_sta_flags(*sta, WLAN_STA_PS)) { if (sdata->bss) atomic_dec(&sdata->bss->num_sta_ps); __sta_info_clear_tim_bit(sdata->bss, *sta); diff --git a/net/mac80211/sta_info.h b/net/mac80211/sta_info.h index f8c95bc9659c..97a61c39ad90 100644 --- a/net/mac80211/sta_info.h +++ b/net/mac80211/sta_info.h @@ -129,23 +129,19 @@ enum plink_state { * * @tid_state_rx: TID's state in Rx session state machine. * @tid_rx: aggregation info for Rx per TID - * @ampdu_rx: for locking sections in aggregation Rx flow * @tid_state_tx: TID's state in Tx session state machine. * @tid_tx: aggregation info for Tx per TID * @addba_req_num: number of times addBA request has been sent. - * @ampdu_tx: for locking sectionsi in aggregation Tx flow * @dialog_token_allocator: dialog token enumerator for each new session; */ struct sta_ampdu_mlme { /* rx */ u8 tid_state_rx[STA_TID_NUM]; struct tid_ampdu_rx *tid_rx[STA_TID_NUM]; - spinlock_t ampdu_rx; /* tx */ u8 tid_state_tx[STA_TID_NUM]; struct tid_ampdu_tx *tid_tx[STA_TID_NUM]; u8 addba_req_num[STA_TID_NUM]; - spinlock_t ampdu_tx; u8 dialog_token_allocator; }; @@ -177,6 +173,8 @@ struct sta_ampdu_mlme { * @rx_bytes: Number of bytes received from this STA * @supp_rates: Bitmap of supported rates (per band) * @ht_info: HT capabilities of this STA + * @lock: used for locking all fields that require locking, see comments + * in the header file. */ struct sta_info { /* General information, mostly static */ @@ -187,6 +185,7 @@ struct sta_info { struct ieee80211_key *key; struct rate_control_ref *rate_ctrl; void *rate_ctrl_priv; + spinlock_t lock; struct ieee80211_ht_info ht_info; u64 supp_rates[IEEE80211_NUM_BANDS]; u8 addr[ETH_ALEN]; @@ -199,7 +198,7 @@ struct sta_info { */ u8 pin_status; - /* frequently updated information, needs locking? */ + /* frequently updated information, locked with lock spinlock */ u32 flags; /* @@ -251,7 +250,7 @@ struct sta_info { int channel_use_raw; /* - * Aggregation information, comes with own locking. + * Aggregation information, locked with lock. */ struct sta_ampdu_mlme ampdu_mlme; u8 timer_to_tid[STA_TID_NUM]; /* identity mapping to ID timers */ @@ -270,9 +269,6 @@ struct sta_info { enum plink_state plink_state; u32 plink_timeout; struct timer_list plink_timer; - spinlock_t plink_lock; /* For peer_state reads / updates and other - updates in the structure. Ensures robust - transitions for the peerlink FSM */ #endif #ifdef CONFIG_MAC80211_DEBUGFS @@ -299,6 +295,64 @@ static inline enum plink_state sta_plink_state(struct sta_info *sta) return PLINK_LISTEN; } +static inline void set_sta_flags(struct sta_info *sta, const u32 flags) +{ + spin_lock_bh(&sta->lock); + sta->flags |= flags; + spin_unlock_bh(&sta->lock); +} + +static inline void clear_sta_flags(struct sta_info *sta, const u32 flags) +{ + spin_lock_bh(&sta->lock); + sta->flags &= ~flags; + spin_unlock_bh(&sta->lock); +} + +static inline void set_and_clear_sta_flags(struct sta_info *sta, + const u32 set, const u32 clear) +{ + spin_lock_bh(&sta->lock); + sta->flags |= set; + sta->flags &= ~clear; + spin_unlock_bh(&sta->lock); +} + +static inline u32 test_sta_flags(struct sta_info *sta, const u32 flags) +{ + u32 ret; + + spin_lock_bh(&sta->lock); + ret = sta->flags & flags; + spin_unlock_bh(&sta->lock); + + return ret; +} + +static inline u32 test_and_clear_sta_flags(struct sta_info *sta, + const u32 flags) +{ + u32 ret; + + spin_lock_bh(&sta->lock); + ret = sta->flags & flags; + sta->flags &= ~flags; + spin_unlock_bh(&sta->lock); + + return ret; +} + +static inline u32 get_sta_flags(struct sta_info *sta) +{ + u32 ret; + + spin_lock_bh(&sta->lock); + ret = sta->flags; + spin_unlock_bh(&sta->lock); + + return ret; +} + /* Maximum number of concurrently registered stations */ #define MAX_STA_COUNT 2007 diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 3a03d7807c39..5f31a6233e13 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -256,7 +256,7 @@ ieee80211_tx_h_check_assoc(struct ieee80211_tx_data *tx) if (tx->flags & IEEE80211_TX_PS_BUFFERED) return TX_CONTINUE; - sta_flags = tx->sta ? tx->sta->flags : 0; + sta_flags = tx->sta ? get_sta_flags(tx->sta) : 0; if (likely(tx->flags & IEEE80211_TX_UNICAST)) { if (unlikely(!(sta_flags & WLAN_STA_ASSOC) && @@ -391,6 +391,7 @@ static ieee80211_tx_result ieee80211_tx_h_unicast_ps_buf(struct ieee80211_tx_data *tx) { struct sta_info *sta = tx->sta; + u32 staflags; DECLARE_MAC_BUF(mac); if (unlikely(!sta || @@ -398,8 +399,10 @@ ieee80211_tx_h_unicast_ps_buf(struct ieee80211_tx_data *tx) (tx->fc & IEEE80211_FCTL_STYPE) == IEEE80211_STYPE_PROBE_RESP))) return TX_CONTINUE; - if (unlikely((sta->flags & WLAN_STA_PS) && - !(sta->flags & WLAN_STA_PSPOLL))) { + staflags = get_sta_flags(sta); + + if (unlikely((staflags & WLAN_STA_PS) && + !(staflags & WLAN_STA_PSPOLL))) { struct ieee80211_tx_packet_data *pkt_data; #ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG printk(KERN_DEBUG "STA %s aid %d: PS buffer (entries " @@ -430,13 +433,13 @@ ieee80211_tx_h_unicast_ps_buf(struct ieee80211_tx_data *tx) return TX_QUEUED; } #ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG - else if (unlikely(sta->flags & WLAN_STA_PS)) { + else if (unlikely(test_sta_flags(sta, WLAN_STA_PS))) { printk(KERN_DEBUG "%s: STA %s in PS mode, but pspoll " "set -> send frame\n", tx->dev->name, print_mac(mac, sta->addr)); } #endif /* CONFIG_MAC80211_VERBOSE_PS_DEBUG */ - sta->flags &= ~WLAN_STA_PSPOLL; + clear_sta_flags(sta, WLAN_STA_PSPOLL); return TX_CONTINUE; } @@ -697,7 +700,7 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) if (((fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA) && (tx->rate->flags & IEEE80211_RATE_SHORT_PREAMBLE) && tx->sdata->bss_conf.use_short_preamble && - (!tx->sta || (tx->sta->flags & WLAN_STA_SHORT_PREAMBLE))) { + (!tx->sta || test_sta_flags(tx->sta, WLAN_STA_SHORT_PREAMBLE))) { tx->control->flags |= IEEE80211_TXCTL_SHORT_PREAMBLE; } @@ -1025,10 +1028,8 @@ __ieee80211_tx_prepare(struct ieee80211_tx_data *tx, if (!tx->sta) control->flags |= IEEE80211_TXCTL_CLEAR_PS_FILT; - else if (tx->sta->flags & WLAN_STA_CLEAR_PS_FILT) { + else if (test_and_clear_sta_flags(tx->sta, WLAN_STA_CLEAR_PS_FILT)) control->flags |= IEEE80211_TXCTL_CLEAR_PS_FILT; - tx->sta->flags &= ~WLAN_STA_CLEAR_PS_FILT; - } hdrlen = ieee80211_get_hdrlen(tx->fc); if (skb->len > hdrlen + sizeof(rfc1042_header) + 2) { @@ -1486,7 +1487,7 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, rcu_read_lock(); sta = sta_info_get(local, hdr.addr1); if (sta) - sta_flags = sta->flags; + sta_flags = get_sta_flags(sta); rcu_read_unlock(); } -- cgit v1.2.2 From fac371d9f09f461dfe9fbbceb2a38e2e12164dda Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Wed, 7 May 2008 23:18:09 +0200 Subject: mac80211: fix queue constant confusion In commit 31ccc476b77234f6afb3 (mac80211: QoS related cleanups) I accidentally changed these to use IEEE80211_MAX_AMPDU_QUEUES twice which obviously is wrong, it should be IEEE80211_MAX_QUEUES once. Currently harmless as they're both the same value anyway. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index d82ed20a344c..173b7fcb3029 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -610,8 +610,8 @@ struct ieee80211_local { struct sta_info *sta_hash[STA_HASH_SIZE]; struct timer_list sta_cleanup; - unsigned long state[IEEE80211_MAX_AMPDU_QUEUES + IEEE80211_MAX_AMPDU_QUEUES]; - struct ieee80211_tx_stored_packet pending_packet[IEEE80211_MAX_AMPDU_QUEUES + IEEE80211_MAX_AMPDU_QUEUES]; + unsigned long state[IEEE80211_MAX_QUEUES + IEEE80211_MAX_AMPDU_QUEUES]; + struct ieee80211_tx_stored_packet pending_packet[IEEE80211_MAX_QUEUES + IEEE80211_MAX_AMPDU_QUEUES]; struct tasklet_struct tx_pending_tasklet; /* number of interfaces with corresponding IFF_ flags */ -- cgit v1.2.2 From 566bfe5a8bcde13188a356f77666f8115813cf31 Mon Sep 17 00:00:00 2001 From: Bruno Randolf Date: Thu, 8 May 2008 19:15:40 +0200 Subject: mac80211: use hardware flags for signal/noise units trying to clean up the signal/noise code. the previous code in mac80211 had confusing names for the related variables, did not have much definition of what units of signal and noise were provided and used implicit mechanisms from the wireless extensions. this patch introduces hardware capability flags to let the hardware specify clearly if it can provide signal and noise level values and which units it can provide. this also anticipates possible new units like RCPI in the future. for signal: IEEE80211_HW_SIGNAL_UNSPEC - unspecified, unknown, hw specific IEEE80211_HW_SIGNAL_DB - dB difference to unspecified reference point IEEE80211_HW_SIGNAL_DBM - dBm, difference to 1mW for noise we currently only have dBm: IEEE80211_HW_NOISE_DBM - dBm, difference to 1mW if IEEE80211_HW_SIGNAL_UNSPEC or IEEE80211_HW_SIGNAL_DB is used the driver has to provide the maximum value (max_signal) it reports in order for applications to make sense of the signal values. i tried my best to find out for each driver what it can provide and update it but i'm not sure (?) for some of them and used the more conservative guess in doubt. this can be fixed easily after this patch has been merged by changing the hardware flags of the driver. DRIVER SIGNAL MAX NOISE QUAL ----------------------------------------------------------------- adm8211 unspec(?) 100 n/a missing at76_usb unspec(?) (?) unused missing ath5k dBm dBm percent rssi b43legacy dBm dBm percent jssi(?) b43 dBm dBm percent jssi(?) iwl-3945 dBm dBm percent snr+more iwl-4965 dBm dBm percent snr+more p54 unspec 127 n/a missing rt2x00 dBm n/a percent rssi+tx/rx frame success rt2400 dBm n/a rt2500pci dBm n/a rt2500usb dBm n/a rt61pci dBm n/a rt73usb dBm n/a rtl8180 unspec(?) 65 n/a (?) rtl8187 unspec(?) 65 (?) noise(?) zd1211 dB(?) 100 n/a percent drivers/net/wireless/ath5k/base.c: Changes-licensed-under: 3-Clause-BSD Signed-off-by: Bruno Randolf Signed-off-by: John W. Linville --- net/ieee80211/ieee80211_rx.c | 2 +- net/mac80211/debugfs_sta.c | 2 +- net/mac80211/ieee80211_i.h | 2 +- net/mac80211/main.c | 10 +++++----- net/mac80211/mlme.c | 12 ++++++------ net/mac80211/rx.c | 4 ++-- net/mac80211/sta_info.h | 2 +- net/mac80211/wext.c | 28 ++++++++++++++++++++-------- 8 files changed, 37 insertions(+), 25 deletions(-) (limited to 'net') diff --git a/net/ieee80211/ieee80211_rx.c b/net/ieee80211/ieee80211_rx.c index 200ee1e63728..69dbc342a464 100644 --- a/net/ieee80211/ieee80211_rx.c +++ b/net/ieee80211/ieee80211_rx.c @@ -391,7 +391,7 @@ int ieee80211_rx(struct ieee80211_device *ieee, struct sk_buff *skb, wstats.updated = 0; if (rx_stats->mask & IEEE80211_STATMASK_RSSI) { - wstats.level = rx_stats->rssi; + wstats.level = rx_stats->signal; wstats.updated |= IW_QUAL_LEVEL_UPDATED; } else wstats.updated |= IW_QUAL_LEVEL_INVALID; diff --git a/net/mac80211/debugfs_sta.c b/net/mac80211/debugfs_sta.c index 1f00273d99c2..a2cc0284c9d0 100644 --- a/net/mac80211/debugfs_sta.c +++ b/net/mac80211/debugfs_sta.c @@ -63,8 +63,8 @@ STA_FILE(tx_fragments, tx_fragments, LU); STA_FILE(tx_filtered, tx_filtered_count, LU); STA_FILE(tx_retry_failed, tx_retry_failed, LU); STA_FILE(tx_retry_count, tx_retry_count, LU); -STA_FILE(last_rssi, last_rssi, D); STA_FILE(last_signal, last_signal, D); +STA_FILE(last_qual, last_qual, D); STA_FILE(last_noise, last_noise, D); STA_FILE(channel_use, channel_use, D); STA_FILE(wep_weak_iv_count, wep_weak_iv_count, LU); diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 173b7fcb3029..ed0d9b35ae6f 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -82,7 +82,7 @@ struct ieee80211_sta_bss { u16 capability; /* host byte order */ enum ieee80211_band band; int freq; - int rssi, signal, noise; + int signal, noise, qual; u8 *wpa_ie; size_t wpa_ie_len; u8 *rsn_ie; diff --git a/net/mac80211/main.c b/net/mac80211/main.c index f277407f0f5a..390df48dfd40 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -1702,13 +1702,13 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) local->hw.conf.beacon_int = 1000; - local->wstats_flags |= local->hw.max_rssi ? - IW_QUAL_LEVEL_UPDATED : IW_QUAL_LEVEL_INVALID; - local->wstats_flags |= local->hw.max_signal ? + local->wstats_flags |= local->hw.flags & (IEEE80211_HW_SIGNAL_UNSPEC | + IEEE80211_HW_SIGNAL_DB | + IEEE80211_HW_SIGNAL_DBM) ? IW_QUAL_QUAL_UPDATED : IW_QUAL_QUAL_INVALID; - local->wstats_flags |= local->hw.max_noise ? + local->wstats_flags |= local->hw.flags & IEEE80211_HW_NOISE_DBM ? IW_QUAL_NOISE_UPDATED : IW_QUAL_NOISE_INVALID; - if (local->hw.max_rssi < 0 || local->hw.max_noise < 0) + if (local->hw.flags & IEEE80211_HW_SIGNAL_DBM) local->wstats_flags |= IW_QUAL_DBM; result = sta_info_start(local); diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 6c4b27be35da..dda30e98aa1e 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -1959,8 +1959,8 @@ static void ieee80211_rx_mgmt_assoc_resp(struct ieee80211_sub_if_data *sdata, local->hw.conf.channel->center_freq, ifsta->ssid, ifsta->ssid_len); if (bss) { - sta->last_rssi = bss->rssi; sta->last_signal = bss->signal; + sta->last_qual = bss->qual; sta->last_noise = bss->noise; ieee80211_rx_bss_put(dev, bss); } @@ -2629,9 +2629,9 @@ static void ieee80211_rx_bss_info(struct net_device *dev, bss->timestamp = beacon_timestamp; bss->last_update = jiffies; - bss->rssi = rx_status->ssi; bss->signal = rx_status->signal; bss->noise = rx_status->noise; + bss->qual = rx_status->qual; if (!beacon && !bss->probe_resp) bss->probe_resp = true; @@ -3427,9 +3427,9 @@ static int ieee80211_sta_config_auth(struct net_device *dev, !ieee80211_sta_match_ssid(ifsta, bss->ssid, bss->ssid_len)) continue; - if (!selected || top_rssi < bss->rssi) { + if (!selected || top_rssi < bss->signal) { selected = bss; - top_rssi = bss->rssi; + top_rssi = bss->signal; } } if (selected) @@ -4060,8 +4060,8 @@ ieee80211_sta_scan_result(struct net_device *dev, memset(&iwe, 0, sizeof(iwe)); iwe.cmd = IWEVQUAL; - iwe.u.qual.qual = bss->signal; - iwe.u.qual.level = bss->rssi; + iwe.u.qual.qual = bss->qual; + iwe.u.qual.level = bss->signal; iwe.u.qual.noise = bss->noise; iwe.u.qual.updated = local->wstats_flags; current_ev = iwe_stream_add_event(current_ev, end_buf, &iwe, diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index b399e09ec7aa..474f13662c84 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -209,7 +209,7 @@ ieee80211_rx_monitor(struct ieee80211_local *local, struct sk_buff *origskb, cpu_to_le16(IEEE80211_CHAN_DYN | IEEE80211_CHAN_2GHZ); - rtfixed->antsignal = status->ssi; + rtfixed->antsignal = status->signal; rthdr->it_len = cpu_to_le16(rtap_len); } @@ -719,8 +719,8 @@ ieee80211_rx_h_sta_process(struct ieee80211_rx_data *rx) sta->rx_fragments++; sta->rx_bytes += rx->skb->len; - sta->last_rssi = rx->status->ssi; sta->last_signal = rx->status->signal; + sta->last_qual = rx->status->qual; sta->last_noise = rx->status->noise; if (!(rx->fc & IEEE80211_FCTL_MOREFRAGS)) { diff --git a/net/mac80211/sta_info.h b/net/mac80211/sta_info.h index 97a61c39ad90..e89cc1655547 100644 --- a/net/mac80211/sta_info.h +++ b/net/mac80211/sta_info.h @@ -216,8 +216,8 @@ struct sta_info { * from this STA */ unsigned long rx_fragments; /* number of received MPDUs */ unsigned long rx_dropped; /* number of dropped MPDUs from this STA */ - int last_rssi; /* RSSI of last received frame from this STA */ int last_signal; /* signal of last received frame from this STA */ + int last_qual; /* qual of last received frame from this STA */ int last_noise; /* noise of last received frame from this STA */ /* last received seq/frag number from this STA (per RX queue) */ __le16 last_seq_ctrl[NUM_RX_DATA_QUEUES]; diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c index 76e1de1dc735..6a342a9a40cd 100644 --- a/net/mac80211/wext.c +++ b/net/mac80211/wext.c @@ -169,14 +169,26 @@ static int ieee80211_ioctl_giwrange(struct net_device *dev, range->num_encoding_sizes = 2; range->max_encoding_tokens = NUM_DEFAULT_KEYS; - range->max_qual.qual = local->hw.max_signal; - range->max_qual.level = local->hw.max_rssi; - range->max_qual.noise = local->hw.max_noise; + if (local->hw.flags & IEEE80211_HW_SIGNAL_UNSPEC || + local->hw.flags & IEEE80211_HW_SIGNAL_DB) + range->max_qual.level = local->hw.max_signal; + else if (local->hw.flags & IEEE80211_HW_SIGNAL_DBM) + range->max_qual.level = -110; + else + range->max_qual.level = 0; + + if (local->hw.flags & IEEE80211_HW_NOISE_DBM) + range->max_qual.noise = -110; + else + range->max_qual.noise = 0; + + range->max_qual.qual = 100; range->max_qual.updated = local->wstats_flags; - range->avg_qual.qual = local->hw.max_signal/2; - range->avg_qual.level = 0; - range->avg_qual.noise = 0; + range->avg_qual.qual = 50; + /* not always true but better than nothing */ + range->avg_qual.level = range->max_qual.level / 2; + range->avg_qual.noise = range->max_qual.noise / 2; range->avg_qual.updated = local->wstats_flags; range->enc_capa = IW_ENC_CAPA_WPA | IW_ENC_CAPA_WPA2 | @@ -996,8 +1008,8 @@ static struct iw_statistics *ieee80211_get_wireless_stats(struct net_device *dev wstats->qual.noise = 0; wstats->qual.updated = IW_QUAL_ALL_INVALID; } else { - wstats->qual.level = sta->last_rssi; - wstats->qual.qual = sta->last_signal; + wstats->qual.level = sta->last_signal; + wstats->qual.qual = sta->last_qual; wstats->qual.noise = sta->last_noise; wstats->qual.updated = local->wstats_flags; } -- cgit v1.2.2 From 601ae7f25aea58f208a7f640f6174aac0652403a Mon Sep 17 00:00:00 2001 From: Bruno Randolf Date: Thu, 8 May 2008 19:22:43 +0200 Subject: mac80211: make rx radiotap header more flexible use hw flags and rx flags to determine which fields are present in the header and use all available information from the driver. make sure radiotap header starts at a naturally aligned address (mod 8) for all radiotap fields. Signed-off-by: Bruno Randolf Signed-off-by: John W. Linville --- net/mac80211/rx.c | 195 +++++++++++++++++++++++++++++++++++++----------------- 1 file changed, 133 insertions(+), 62 deletions(-) (limited to 'net') diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index 474f13662c84..1159a43a3df6 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -77,6 +77,134 @@ static inline int should_drop_frame(struct ieee80211_rx_status *status, return 0; } +static int +ieee80211_rx_radiotap_len(struct ieee80211_local *local, + struct ieee80211_rx_status *status) +{ + int len; + + /* always present fields */ + len = sizeof(struct ieee80211_radiotap_header) + 9; + + if (status->flag & RX_FLAG_TSFT) + len += 8; + if (local->hw.flags & IEEE80211_HW_SIGNAL_DB || + local->hw.flags & IEEE80211_HW_SIGNAL_DBM) + len += 1; + if (local->hw.flags & IEEE80211_HW_NOISE_DBM) + len += 1; + + if (len & 1) /* padding for RX_FLAGS if necessary */ + len++; + + /* make sure radiotap starts at a naturally aligned address */ + if (len % 8) + len = roundup(len, 8); + + return len; +} + +/** + * ieee80211_add_rx_radiotap_header - add radiotap header + * + * add a radiotap header containing all the fields which the hardware provided. + */ +static void +ieee80211_add_rx_radiotap_header(struct ieee80211_local *local, + struct sk_buff *skb, + struct ieee80211_rx_status *status, + struct ieee80211_rate *rate, + int rtap_len) +{ + struct ieee80211_radiotap_header *rthdr; + unsigned char *pos; + + rthdr = (struct ieee80211_radiotap_header *)skb_push(skb, rtap_len); + memset(rthdr, 0, rtap_len); + + /* radiotap header, set always present flags */ + rthdr->it_present = + cpu_to_le32((1 << IEEE80211_RADIOTAP_FLAGS) | + (1 << IEEE80211_RADIOTAP_RATE) | + (1 << IEEE80211_RADIOTAP_CHANNEL) | + (1 << IEEE80211_RADIOTAP_ANTENNA) | + (1 << IEEE80211_RADIOTAP_RX_FLAGS)); + rthdr->it_len = cpu_to_le16(rtap_len); + + pos = (unsigned char *)(rthdr+1); + + /* the order of the following fields is important */ + + /* IEEE80211_RADIOTAP_TSFT */ + if (status->flag & RX_FLAG_TSFT) { + *(__le64 *)pos = cpu_to_le64(status->mactime); + rthdr->it_present |= + cpu_to_le32(1 << IEEE80211_RADIOTAP_TSFT); + pos += 8; + } + + /* IEEE80211_RADIOTAP_FLAGS */ + if (local->hw.flags & IEEE80211_HW_RX_INCLUDES_FCS) + *pos |= IEEE80211_RADIOTAP_F_FCS; + pos++; + + /* IEEE80211_RADIOTAP_RATE */ + *pos = rate->bitrate / 5; + pos++; + + /* IEEE80211_RADIOTAP_CHANNEL */ + *(__le16 *)pos = cpu_to_le16(status->freq); + pos += 2; + if (status->band == IEEE80211_BAND_5GHZ) + *(__le16 *)pos = cpu_to_le16(IEEE80211_CHAN_OFDM | + IEEE80211_CHAN_5GHZ); + else + *(__le16 *)pos = cpu_to_le16(IEEE80211_CHAN_DYN | + IEEE80211_CHAN_2GHZ); + pos += 2; + + /* IEEE80211_RADIOTAP_DBM_ANTSIGNAL */ + if (local->hw.flags & IEEE80211_HW_SIGNAL_DBM) { + *pos = status->signal; + rthdr->it_present |= + cpu_to_le32(1 << IEEE80211_RADIOTAP_DBM_ANTSIGNAL); + pos++; + } + + /* IEEE80211_RADIOTAP_DBM_ANTNOISE */ + if (local->hw.flags & IEEE80211_HW_NOISE_DBM) { + *pos = status->noise; + rthdr->it_present |= + cpu_to_le32(1 << IEEE80211_RADIOTAP_DBM_ANTNOISE); + pos++; + } + + /* IEEE80211_RADIOTAP_LOCK_QUALITY is missing */ + + /* IEEE80211_RADIOTAP_ANTENNA */ + *pos = status->antenna; + pos++; + + /* IEEE80211_RADIOTAP_DB_ANTSIGNAL */ + if (local->hw.flags & IEEE80211_HW_SIGNAL_DB) { + *pos = status->signal; + rthdr->it_present |= + cpu_to_le32(1 << IEEE80211_RADIOTAP_DB_ANTSIGNAL); + pos++; + } + + /* IEEE80211_RADIOTAP_DB_ANTNOISE is not used */ + + /* IEEE80211_RADIOTAP_RX_FLAGS */ + /* ensure 2 byte alignment for the 2 byte field as required */ + if ((pos - (unsigned char *)rthdr) & 1) + pos++; + /* FIXME: when radiotap gets a 'bad PLCP' flag use it here */ + if (status->flag & (RX_FLAG_FAILED_FCS_CRC | RX_FLAG_FAILED_PLCP_CRC)) + *(__le16 *)pos |= cpu_to_le16(IEEE80211_RADIOTAP_F_RX_BADFCS); + pos += 2; +} + /* * This function copies a received frame to all monitor interfaces and * returns a cleaned-up SKB that no longer includes the FCS nor the @@ -89,17 +217,6 @@ ieee80211_rx_monitor(struct ieee80211_local *local, struct sk_buff *origskb, { struct ieee80211_sub_if_data *sdata; int needed_headroom = 0; - struct ieee80211_radiotap_header *rthdr; - __le64 *rttsft = NULL; - struct ieee80211_rtap_fixed_data { - u8 flags; - u8 rate; - __le16 chan_freq; - __le16 chan_flags; - u8 antsignal; - u8 padding_for_rxflags; - __le16 rx_flags; - } __attribute__ ((packed)) *rtfixed; struct sk_buff *skb, *skb2; struct net_device *prev_dev = NULL; int present_fcs_len = 0; @@ -116,8 +233,8 @@ ieee80211_rx_monitor(struct ieee80211_local *local, struct sk_buff *origskb, if (status->flag & RX_FLAG_RADIOTAP) rtap_len = ieee80211_get_radiotap_len(origskb->data); else - /* room for radiotap header, always present fields and TSFT */ - needed_headroom = sizeof(*rthdr) + sizeof(*rtfixed) + 8; + /* room for the radiotap header based on driver features */ + needed_headroom = ieee80211_rx_radiotap_len(local, status); if (local->hw.flags & IEEE80211_HW_RX_INCLUDES_FCS) present_fcs_len = FCS_LEN; @@ -163,55 +280,9 @@ ieee80211_rx_monitor(struct ieee80211_local *local, struct sk_buff *origskb, } /* if necessary, prepend radiotap information */ - if (!(status->flag & RX_FLAG_RADIOTAP)) { - rtfixed = (void *) skb_push(skb, sizeof(*rtfixed)); - rtap_len = sizeof(*rthdr) + sizeof(*rtfixed); - if (status->flag & RX_FLAG_TSFT) { - rttsft = (void *) skb_push(skb, sizeof(*rttsft)); - rtap_len += 8; - } - rthdr = (void *) skb_push(skb, sizeof(*rthdr)); - memset(rthdr, 0, sizeof(*rthdr)); - memset(rtfixed, 0, sizeof(*rtfixed)); - rthdr->it_present = - cpu_to_le32((1 << IEEE80211_RADIOTAP_FLAGS) | - (1 << IEEE80211_RADIOTAP_RATE) | - (1 << IEEE80211_RADIOTAP_CHANNEL) | - (1 << IEEE80211_RADIOTAP_DB_ANTSIGNAL) | - (1 << IEEE80211_RADIOTAP_RX_FLAGS)); - rtfixed->flags = 0; - if (local->hw.flags & IEEE80211_HW_RX_INCLUDES_FCS) - rtfixed->flags |= IEEE80211_RADIOTAP_F_FCS; - - if (rttsft) { - *rttsft = cpu_to_le64(status->mactime); - rthdr->it_present |= - cpu_to_le32(1 << IEEE80211_RADIOTAP_TSFT); - } - - /* FIXME: when radiotap gets a 'bad PLCP' flag use it here */ - rtfixed->rx_flags = 0; - if (status->flag & - (RX_FLAG_FAILED_FCS_CRC | RX_FLAG_FAILED_PLCP_CRC)) - rtfixed->rx_flags |= - cpu_to_le16(IEEE80211_RADIOTAP_F_RX_BADFCS); - - rtfixed->rate = rate->bitrate / 5; - - rtfixed->chan_freq = cpu_to_le16(status->freq); - - if (status->band == IEEE80211_BAND_5GHZ) - rtfixed->chan_flags = - cpu_to_le16(IEEE80211_CHAN_OFDM | - IEEE80211_CHAN_5GHZ); - else - rtfixed->chan_flags = - cpu_to_le16(IEEE80211_CHAN_DYN | - IEEE80211_CHAN_2GHZ); - - rtfixed->antsignal = status->signal; - rthdr->it_len = cpu_to_le16(rtap_len); - } + if (!(status->flag & RX_FLAG_RADIOTAP)) + ieee80211_add_rx_radiotap_header(local, skb, status, rate, + needed_headroom); skb_reset_mac_header(skb); skb->ip_summed = CHECKSUM_UNNECESSARY; -- cgit v1.2.2 From 2940bb69fd84047e78fdb1868a8b894df5584255 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Thu, 8 May 2008 14:30:18 -0700 Subject: wireless: Add missing locking to cfg80211_dev_rename device_rename only performs useful and race free validity checking at the optional sysfs level so depending on it for all of the validity checking in cfg80211_dev_rename is racy. Instead implement all of the needed validity checking and locking in cfg80211_dev_rename. Signed-off-by: Eric W. Biederman Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/wireless/core.c | 33 ++++++++++++++++++++++++++++----- 1 file changed, 28 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/wireless/core.c b/net/wireless/core.c index 80afacdae46c..f1da0b93bc56 100644 --- a/net/wireless/core.c +++ b/net/wireless/core.c @@ -143,8 +143,11 @@ void cfg80211_put_dev(struct cfg80211_registered_device *drv) int cfg80211_dev_rename(struct cfg80211_registered_device *rdev, char *newname) { + struct cfg80211_registered_device *drv; int idx, taken = -1, result, digits; + mutex_lock(&cfg80211_drv_mutex); + /* prohibit calling the thing phy%d when %d is not its number */ sscanf(newname, PHY_NAME "%d%n", &idx, &taken); if (taken == strlen(newname) && idx != rdev->idx) { @@ -156,14 +159,30 @@ int cfg80211_dev_rename(struct cfg80211_registered_device *rdev, * deny the name if it is phy where is printed * without leading zeroes. taken == strlen(newname) here */ + result = -EINVAL; if (taken == strlen(PHY_NAME) + digits) - return -EINVAL; + goto out_unlock; + } + + + /* Ignore nop renames */ + result = 0; + if (strcmp(newname, dev_name(&rdev->wiphy.dev)) == 0) + goto out_unlock; + + /* Ensure another device does not already have this name. */ + list_for_each_entry(drv, &cfg80211_drv_list, list) { + result = -EINVAL; + if (strcmp(newname, dev_name(&drv->wiphy.dev)) == 0) + goto out_unlock; } - /* this will check for collisions */ + /* this will only check for collisions in sysfs + * which is not even always compiled in. + */ result = device_rename(&rdev->wiphy.dev, newname); if (result) - return result; + goto out_unlock; if (!debugfs_rename(rdev->wiphy.debugfsdir->d_parent, rdev->wiphy.debugfsdir, @@ -172,9 +191,13 @@ int cfg80211_dev_rename(struct cfg80211_registered_device *rdev, printk(KERN_ERR "cfg80211: failed to rename debugfs dir to %s!\n", newname); - nl80211_notify_dev_rename(rdev); + result = 0; +out_unlock: + mutex_unlock(&cfg80211_drv_mutex); + if (result == 0) + nl80211_notify_dev_rename(rdev); - return 0; + return result; } /* exported functions */ -- cgit v1.2.2 From b0a6717994a4e00ee19372e1bdaab53572ae025c Mon Sep 17 00:00:00 2001 From: Ivo van Doorn Date: Tue, 13 May 2008 15:03:02 +0200 Subject: mac80211: Set IEEE80211_TXPD_REQ_TX_STATUS for all TX frames All interfaces should set the IEEE80211_TXPD_REQ_TX_STATUS flag for all TX frames which will force the master interface to set the IEEE80211_TX_CTL_REQ_TX_STATUS flag. This in turn will allow drivers to check for that flag before reporting the TX status to mac80211. This is very usefull when frames (like beacons, RTS and CTS-to-self) should not be reported back to mac80211. Later we could add more extensive checks to exclude more frames from being reported, or let mac80211 decide if it wants the frame for status reporting or not. v2: Monitor interfaces should also set IEEE80211_TXPD_REQ_TX_STATUS Signed-off-by: Ivo van Doorn Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/tx.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 5f31a6233e13..4c25fd5d76a7 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -1337,6 +1337,8 @@ int ieee80211_monitor_start_xmit(struct sk_buff *skb, pkt_data->ifindex = dev->ifindex; pkt_data->flags |= IEEE80211_TXPD_DO_NOT_ENCRYPT; + /* Interfaces should always request a status report */ + pkt_data->flags |= IEEE80211_TXPD_REQ_TX_STATUS; /* * fix up the pointers accounting for the radiotap @@ -1618,6 +1620,9 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, if (ethertype == ETH_P_PAE) pkt_data->flags |= IEEE80211_TXPD_EAPOL_FRAME; + /* Interfaces should always request a status report */ + pkt_data->flags |= IEEE80211_TXPD_REQ_TX_STATUS; + skb->dev = local->mdev; dev->stats.tx_packets++; dev->stats.tx_bytes += skb->len; -- cgit v1.2.2 From e15f880409c807bb589e9492263564e80f0de6e9 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 19 May 2008 13:27:31 -0700 Subject: tipc: Add support for customized subscription overlap handling This patch enables TIPC's topology server code to do customized overlap detection handling on a per-subscription basis. (This capability is needed to support the upcoming introduction of multi-cluster TIPC networks.) Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/subscr.c | 4 +++- net/tipc/subscr.h | 8 ++++++++ 2 files changed, 11 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c index 8f8d0a6c1c16..81e2bd5f2413 100644 --- a/net/tipc/subscr.c +++ b/net/tipc/subscr.c @@ -157,7 +157,8 @@ void tipc_subscr_report_overlap(struct subscription *sub, return; if (!must && !(sub->filter & TIPC_SUB_PORTS)) return; - subscr_send_event(sub, found_lower, found_upper, event, port_ref, node); + + sub->event_cb(sub, found_lower, found_upper, event, port_ref, node); } /** @@ -372,6 +373,7 @@ static void subscr_subscribe(struct tipc_subscr *s, subscr_terminate(subscriber); return; } + sub->event_cb = subscr_send_event; memcpy(&sub->evt.s, s, sizeof(struct tipc_subscr)); INIT_LIST_HEAD(&sub->subscription_list); INIT_LIST_HEAD(&sub->nameseq_list); diff --git a/net/tipc/subscr.h b/net/tipc/subscr.h index 93a8e674fac1..d95536832907 100644 --- a/net/tipc/subscr.h +++ b/net/tipc/subscr.h @@ -37,11 +37,18 @@ #ifndef _TIPC_SUBSCR_H #define _TIPC_SUBSCR_H +struct subscription; + +typedef void (*tipc_subscr_event) (struct subscription *sub, + u32 found_lower, u32 found_upper, + u32 event, u32 port_ref, u32 node); + /** * struct subscription - TIPC network topology subscription object * @seq: name sequence associated with subscription * @timeout: duration of subscription (in ms) * @filter: event filtering to be done for subscription + * @event_cb: routine invoked when a subscription event is detected * @evt: template for events generated by subscription * @subscription_list: adjacent subscriptions in subscriber's subscription list * @nameseq_list: adjacent subscriptions in name sequence's subscription list @@ -53,6 +60,7 @@ struct subscription { struct tipc_name_seq seq; u32 timeout; u32 filter; + tipc_subscr_event event_cb; struct tipc_event evt; struct list_head subscription_list; struct list_head nameseq_list; -- cgit v1.2.2 From 8e9501f5188d90eed737240453c32cad01849c96 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 19 May 2008 13:28:32 -0700 Subject: tipc: Add support for customized subscription endianness This patch enables TIPC's topology server code to do customized endianness conversions on a per-subscription basis. (This capability is needed to support the upcoming consolidation of subscriber and subscription object references.) Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/subscr.c | 35 ++++++++++++++++------------------- net/tipc/subscr.h | 2 ++ 2 files changed, 18 insertions(+), 19 deletions(-) (limited to 'net') diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c index 81e2bd5f2413..a62e5d30638c 100644 --- a/net/tipc/subscr.c +++ b/net/tipc/subscr.c @@ -47,7 +47,6 @@ * @subscriber_list: adjacent subscribers in top. server's list of subscribers * @subscription_list: list of subscription objects for this subscriber * @port_ref: object reference to port used to communicate with subscriber - * @swap: indicates if subscriber uses opposite endianness in its messages */ struct subscriber { @@ -56,7 +55,6 @@ struct subscriber { struct list_head subscriber_list; struct list_head subscription_list; u32 port_ref; - int swap; }; /** @@ -109,11 +107,11 @@ static void subscr_send_event(struct subscription *sub, msg_sect.iov_base = (void *)&sub->evt; msg_sect.iov_len = sizeof(struct tipc_event); - sub->evt.event = htohl(event, sub->owner->swap); - sub->evt.found_lower = htohl(found_lower, sub->owner->swap); - sub->evt.found_upper = htohl(found_upper, sub->owner->swap); - sub->evt.port.ref = htohl(port_ref, sub->owner->swap); - sub->evt.port.node = htohl(node, sub->owner->swap); + sub->evt.event = htohl(event, sub->swap); + sub->evt.found_lower = htohl(found_lower, sub->swap); + sub->evt.found_upper = htohl(found_upper, sub->swap); + sub->evt.port.ref = htohl(port_ref, sub->swap); + sub->evt.port.node = htohl(node, sub->swap); tipc_send(sub->owner->port_ref, 1, &msg_sect); } @@ -324,18 +322,16 @@ static void subscr_subscribe(struct tipc_subscr *s, struct subscriber *subscriber) { struct subscription *sub; + int swap; - /* Determine/update subscriber's endianness */ + /* Determine subscriber's endianness */ - if (s->filter & (TIPC_SUB_PORTS | TIPC_SUB_SERVICE)) - subscriber->swap = 0; - else - subscriber->swap = 1; + swap = !(s->filter & (TIPC_SUB_PORTS | TIPC_SUB_SERVICE)); /* Detect & process a subscription cancellation request */ - if (s->filter & htohl(TIPC_SUB_CANCEL, subscriber->swap)) { - s->filter &= ~htohl(TIPC_SUB_CANCEL, subscriber->swap); + if (s->filter & htohl(TIPC_SUB_CANCEL, swap)) { + s->filter &= ~htohl(TIPC_SUB_CANCEL, swap); subscr_cancel(s, subscriber); return; } @@ -360,11 +356,11 @@ static void subscr_subscribe(struct tipc_subscr *s, /* Initialize subscription object */ - sub->seq.type = htohl(s->seq.type, subscriber->swap); - sub->seq.lower = htohl(s->seq.lower, subscriber->swap); - sub->seq.upper = htohl(s->seq.upper, subscriber->swap); - sub->timeout = htohl(s->timeout, subscriber->swap); - sub->filter = htohl(s->filter, subscriber->swap); + sub->seq.type = htohl(s->seq.type, swap); + sub->seq.lower = htohl(s->seq.lower, swap); + sub->seq.upper = htohl(s->seq.upper, swap); + sub->timeout = htohl(s->timeout, swap); + sub->filter = htohl(s->filter, swap); if ((!(sub->filter & TIPC_SUB_PORTS) == !(sub->filter & TIPC_SUB_SERVICE)) || (sub->seq.lower > sub->seq.upper)) { @@ -378,6 +374,7 @@ static void subscr_subscribe(struct tipc_subscr *s, INIT_LIST_HEAD(&sub->subscription_list); INIT_LIST_HEAD(&sub->nameseq_list); list_add(&sub->subscription_list, &subscriber->subscription_list); + sub->swap = swap; atomic_inc(&topsrv.subscription_count); if (sub->timeout != TIPC_WAIT_FOREVER) { k_init_timer(&sub->timer, diff --git a/net/tipc/subscr.h b/net/tipc/subscr.h index d95536832907..3e3e0265146e 100644 --- a/net/tipc/subscr.h +++ b/net/tipc/subscr.h @@ -49,6 +49,7 @@ typedef void (*tipc_subscr_event) (struct subscription *sub, * @timeout: duration of subscription (in ms) * @filter: event filtering to be done for subscription * @event_cb: routine invoked when a subscription event is detected + * @swap: indicates if subscriber uses opposite endianness in its messages * @evt: template for events generated by subscription * @subscription_list: adjacent subscriptions in subscriber's subscription list * @nameseq_list: adjacent subscriptions in name sequence's subscription list @@ -61,6 +62,7 @@ struct subscription { u32 timeout; u32 filter; tipc_subscr_event event_cb; + int swap; struct tipc_event evt; struct list_head subscription_list; struct list_head nameseq_list; -- cgit v1.2.2 From fc5ad582709ce9c7b9ab7b70c1e5b5e2cfc384db Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 19 May 2008 13:29:06 -0700 Subject: tipc: Fix bug in topology server byte swapping routine This patch fixes TIPC's topology server so that it does byte swapping correctly when endianness conversion is required. (Note: This bug only impacted an application if it issues a subscription request to a topology server on another node, rather than the server on it's own node; since the topology server is normally not accessible by off-node applications, most TIPC applications were not impacted by the bug.) Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/subscr.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'net') diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c index a62e5d30638c..dde23f1e7542 100644 --- a/net/tipc/subscr.c +++ b/net/tipc/subscr.c @@ -86,9 +86,7 @@ static struct top_srv topsrv = { 0 }; static u32 htohl(u32 in, int swap) { - char *c = (char *)∈ - - return swap ? ((c[3] << 3) + (c[2] << 2) + (c[1] << 1) + c[0]) : in; + return swap ? (u32)___constant_swab32(in) : in; } /** -- cgit v1.2.2 From 28353e7fad1d224687220a448950dc552645a50a Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 19 May 2008 13:29:47 -0700 Subject: tipc: Consolidate subscriber & subscriber port references This patch modifies TIPC's network topology service so that it only requires a single reference table entry per subscriber connection, rather than two. This is achieved by letting the reference to the server port communicating with the subscriber act as the reference to the subscriber object itself. (Since the subscriber cannot exist without its port, and vice versa, this dual role for the reference is perfectly natural.) This consolidation reduces the size of the reference table by 50% in the default configuration. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/core.c | 3 +- net/tipc/subscr.c | 199 +++++++++++++++++++++++++++++++----------------------- net/tipc/subscr.h | 4 +- 3 files changed, 116 insertions(+), 90 deletions(-) (limited to 'net') diff --git a/net/tipc/core.c b/net/tipc/core.c index 6d6aa5a3c240..3d97386af095 100644 --- a/net/tipc/core.c +++ b/net/tipc/core.c @@ -164,8 +164,7 @@ int tipc_core_start(void) tipc_mode = TIPC_NODE_MODE; if ((res = tipc_handler_start()) || - (res = tipc_ref_table_init(tipc_max_ports + tipc_max_subscriptions, - tipc_random)) || + (res = tipc_ref_table_init(tipc_max_ports, tipc_random)) || (res = tipc_reg_start()) || (res = tipc_nametbl_init()) || (res = tipc_k_signal((Handler)tipc_subscr_start, 0)) || diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c index dde23f1e7542..7c62791eb0cc 100644 --- a/net/tipc/subscr.c +++ b/net/tipc/subscr.c @@ -38,23 +38,22 @@ #include "dbg.h" #include "subscr.h" #include "name_table.h" +#include "port.h" #include "ref.h" /** * struct subscriber - TIPC network topology subscriber - * @ref: object reference to subscriber object itself - * @lock: pointer to spinlock controlling access to subscriber object + * @port_ref: object reference to server port connecting to subscriber + * @lock: pointer to spinlock controlling access to subscriber's server port * @subscriber_list: adjacent subscribers in top. server's list of subscribers * @subscription_list: list of subscription objects for this subscriber - * @port_ref: object reference to port used to communicate with subscriber */ struct subscriber { - u32 ref; + u32 port_ref; spinlock_t *lock; struct list_head subscriber_list; struct list_head subscription_list; - u32 port_ref; }; /** @@ -91,6 +90,9 @@ static u32 htohl(u32 in, int swap) /** * subscr_send_event - send a message containing a tipc_event to the subscriber + * + * Note: Must not hold subscriber's server port lock, since tipc_send() will + * try to take the lock if the message is rejected and returned! */ static void subscr_send_event(struct subscription *sub, @@ -110,7 +112,7 @@ static void subscr_send_event(struct subscription *sub, sub->evt.found_upper = htohl(found_upper, sub->swap); sub->evt.port.ref = htohl(port_ref, sub->swap); sub->evt.port.node = htohl(node, sub->swap); - tipc_send(sub->owner->port_ref, 1, &msg_sect); + tipc_send(sub->server_ref, 1, &msg_sect); } /** @@ -163,20 +165,18 @@ void tipc_subscr_report_overlap(struct subscription *sub, static void subscr_timeout(struct subscription *sub) { - struct subscriber *subscriber; - u32 subscriber_ref; + struct port *server_port; - /* Validate subscriber reference (in case subscriber is terminating) */ + /* Validate server port reference (in case subscriber is terminating) */ - subscriber_ref = sub->owner->ref; - subscriber = (struct subscriber *)tipc_ref_lock(subscriber_ref); - if (subscriber == NULL) + server_port = tipc_port_lock(sub->server_ref); + if (server_port == NULL) return; /* Validate timeout (in case subscription is being cancelled) */ if (sub->timeout == TIPC_WAIT_FOREVER) { - tipc_ref_unlock(subscriber_ref); + tipc_port_unlock(server_port); return; } @@ -184,19 +184,21 @@ static void subscr_timeout(struct subscription *sub) tipc_nametbl_unsubscribe(sub); - /* Notify subscriber of timeout, then unlink subscription */ + /* Unlink subscription from subscriber */ - subscr_send_event(sub, - sub->evt.s.seq.lower, - sub->evt.s.seq.upper, - TIPC_SUBSCR_TIMEOUT, - 0, - 0); list_del(&sub->subscription_list); + /* Release subscriber's server port */ + + tipc_port_unlock(server_port); + + /* Notify subscriber of timeout */ + + subscr_send_event(sub, sub->evt.s.seq.lower, sub->evt.s.seq.upper, + TIPC_SUBSCR_TIMEOUT, 0, 0); + /* Now destroy subscription */ - tipc_ref_unlock(subscriber_ref); k_term_timer(&sub->timer); kfree(sub); atomic_dec(&topsrv.subscription_count); @@ -205,7 +207,7 @@ static void subscr_timeout(struct subscription *sub) /** * subscr_del - delete a subscription within a subscription list * - * Called with subscriber locked. + * Called with subscriber port locked. */ static void subscr_del(struct subscription *sub) @@ -219,7 +221,7 @@ static void subscr_del(struct subscription *sub) /** * subscr_terminate - terminate communication with a subscriber * - * Called with subscriber locked. Routine must temporarily release this lock + * Called with subscriber port locked. Routine must temporarily release lock * to enable subscription timeout routine(s) to finish without deadlocking; * the lock is then reclaimed to allow caller to release it upon return. * (This should work even in the unlikely event some other thread creates @@ -229,14 +231,21 @@ static void subscr_del(struct subscription *sub) static void subscr_terminate(struct subscriber *subscriber) { + u32 port_ref; struct subscription *sub; struct subscription *sub_temp; /* Invalidate subscriber reference */ - tipc_ref_discard(subscriber->ref); + port_ref = subscriber->port_ref; + subscriber->port_ref = 0; spin_unlock_bh(subscriber->lock); + /* Sever connection to subscriber */ + + tipc_shutdown(port_ref); + tipc_deleteport(port_ref); + /* Destroy any existing subscriptions for subscriber */ list_for_each_entry_safe(sub, sub_temp, &subscriber->subscription_list, @@ -250,27 +259,25 @@ static void subscr_terminate(struct subscriber *subscriber) subscr_del(sub); } - /* Sever connection to subscriber */ - - tipc_shutdown(subscriber->port_ref); - tipc_deleteport(subscriber->port_ref); - /* Remove subscriber from topology server's subscriber list */ spin_lock_bh(&topsrv.lock); list_del(&subscriber->subscriber_list); spin_unlock_bh(&topsrv.lock); - /* Now destroy subscriber */ + /* Reclaim subscriber lock */ spin_lock_bh(subscriber->lock); + + /* Now destroy subscriber */ + kfree(subscriber); } /** * subscr_cancel - handle subscription cancellation request * - * Called with subscriber locked. Routine must temporarily release this lock + * Called with subscriber port locked. Routine must temporarily release lock * to enable the subscription timeout routine to finish without deadlocking; * the lock is then reclaimed to allow caller to release it upon return. * @@ -313,11 +320,11 @@ static void subscr_cancel(struct tipc_subscr *s, /** * subscr_subscribe - create subscription for subscriber * - * Called with subscriber locked + * Called with subscriber port locked. */ -static void subscr_subscribe(struct tipc_subscr *s, - struct subscriber *subscriber) +static struct subscription *subscr_subscribe(struct tipc_subscr *s, + struct subscriber *subscriber) { struct subscription *sub; int swap; @@ -331,7 +338,7 @@ static void subscr_subscribe(struct tipc_subscr *s, if (s->filter & htohl(TIPC_SUB_CANCEL, swap)) { s->filter &= ~htohl(TIPC_SUB_CANCEL, swap); subscr_cancel(s, subscriber); - return; + return NULL; } /* Refuse subscription if global limit exceeded */ @@ -340,16 +347,16 @@ static void subscr_subscribe(struct tipc_subscr *s, warn("Subscription rejected, subscription limit reached (%u)\n", tipc_max_subscriptions); subscr_terminate(subscriber); - return; + return NULL; } /* Allocate subscription object */ - sub = kzalloc(sizeof(*sub), GFP_ATOMIC); + sub = kmalloc(sizeof(*sub), GFP_ATOMIC); if (!sub) { warn("Subscription rejected, no memory\n"); subscr_terminate(subscriber); - return; + return NULL; } /* Initialize subscription object */ @@ -365,40 +372,41 @@ static void subscr_subscribe(struct tipc_subscr *s, warn("Subscription rejected, illegal request\n"); kfree(sub); subscr_terminate(subscriber); - return; + return NULL; } sub->event_cb = subscr_send_event; - memcpy(&sub->evt.s, s, sizeof(struct tipc_subscr)); - INIT_LIST_HEAD(&sub->subscription_list); INIT_LIST_HEAD(&sub->nameseq_list); list_add(&sub->subscription_list, &subscriber->subscription_list); + sub->server_ref = subscriber->port_ref; sub->swap = swap; + memcpy(&sub->evt.s, s, sizeof(struct tipc_subscr)); atomic_inc(&topsrv.subscription_count); if (sub->timeout != TIPC_WAIT_FOREVER) { k_init_timer(&sub->timer, (Handler)subscr_timeout, (unsigned long)sub); k_start_timer(&sub->timer, sub->timeout); } - sub->owner = subscriber; - tipc_nametbl_subscribe(sub); + + return sub; } /** * subscr_conn_shutdown_event - handle termination request from subscriber + * + * Called with subscriber's server port unlocked. */ static void subscr_conn_shutdown_event(void *usr_handle, - u32 portref, + u32 port_ref, struct sk_buff **buf, unsigned char const *data, unsigned int size, int reason) { - struct subscriber *subscriber; + struct subscriber *subscriber = usr_handle; spinlock_t *subscriber_lock; - subscriber = tipc_ref_lock((u32)(unsigned long)usr_handle); - if (subscriber == NULL) + if (tipc_port_lock(port_ref) == NULL) return; subscriber_lock = subscriber->lock; @@ -408,6 +416,8 @@ static void subscr_conn_shutdown_event(void *usr_handle, /** * subscr_conn_msg_event - handle new subscription request from subscriber + * + * Called with subscriber's server port unlocked. */ static void subscr_conn_msg_event(void *usr_handle, @@ -416,20 +426,46 @@ static void subscr_conn_msg_event(void *usr_handle, const unchar *data, u32 size) { - struct subscriber *subscriber; + struct subscriber *subscriber = usr_handle; spinlock_t *subscriber_lock; + struct subscription *sub; + + /* + * Lock subscriber's server port (& make a local copy of lock pointer, + * in case subscriber is deleted while processing subscription request) + */ - subscriber = tipc_ref_lock((u32)(unsigned long)usr_handle); - if (subscriber == NULL) + if (tipc_port_lock(port_ref) == NULL) return; subscriber_lock = subscriber->lock; - if (size != sizeof(struct tipc_subscr)) - subscr_terminate(subscriber); - else - subscr_subscribe((struct tipc_subscr *)data, subscriber); - spin_unlock_bh(subscriber_lock); + if (size != sizeof(struct tipc_subscr)) { + subscr_terminate(subscriber); + spin_unlock_bh(subscriber_lock); + } else { + sub = subscr_subscribe((struct tipc_subscr *)data, subscriber); + spin_unlock_bh(subscriber_lock); + if (sub != NULL) { + + /* + * We must release the server port lock before adding a + * subscription to the name table since TIPC needs to be + * able to (re)acquire the port lock if an event message + * issued by the subscription process is rejected and + * returned. The subscription cannot be deleted while + * it is being added to the name table because: + * a) the single-threading of the native API port code + * ensures the subscription cannot be cancelled and + * the subscriber connection cannot be broken, and + * b) the name table lock ensures the subscription + * timeout code cannot delete the subscription, + * so the subscription object is still protected. + */ + + tipc_nametbl_subscribe(sub); + } + } } /** @@ -445,16 +481,10 @@ static void subscr_named_msg_event(void *usr_handle, struct tipc_portid const *orig, struct tipc_name_seq const *dest) { - struct subscriber *subscriber; - struct iovec msg_sect = {NULL, 0}; - spinlock_t *subscriber_lock; + static struct iovec msg_sect = {NULL, 0}; - dbg("subscr_named_msg_event: orig = %x own = %x,\n", - orig->node, tipc_own_addr); - if (size && (size != sizeof(struct tipc_subscr))) { - warn("Subscriber rejected, invalid subscription size\n"); - return; - } + struct subscriber *subscriber; + u32 server_port_ref; /* Create subscriber object */ @@ -465,18 +495,11 @@ static void subscr_named_msg_event(void *usr_handle, } INIT_LIST_HEAD(&subscriber->subscription_list); INIT_LIST_HEAD(&subscriber->subscriber_list); - subscriber->ref = tipc_ref_acquire(subscriber, &subscriber->lock); - if (subscriber->ref == 0) { - warn("Subscriber rejected, reference table exhausted\n"); - kfree(subscriber); - return; - } - spin_unlock_bh(subscriber->lock); - /* Establish a connection to subscriber */ + /* Create server port & establish connection to subscriber */ tipc_createport(topsrv.user_ref, - (void *)(unsigned long)subscriber->ref, + subscriber, importance, NULL, NULL, @@ -488,32 +511,36 @@ static void subscr_named_msg_event(void *usr_handle, &subscriber->port_ref); if (subscriber->port_ref == 0) { warn("Subscriber rejected, unable to create port\n"); - tipc_ref_discard(subscriber->ref); kfree(subscriber); return; } tipc_connect2port(subscriber->port_ref, orig); + /* Lock server port (& save lock address for future use) */ + + subscriber->lock = tipc_port_lock(subscriber->port_ref)->publ.lock; /* Add subscriber to topology server's subscriber list */ - tipc_ref_lock(subscriber->ref); spin_lock_bh(&topsrv.lock); list_add(&subscriber->subscriber_list, &topsrv.subscriber_list); spin_unlock_bh(&topsrv.lock); - /* - * Subscribe now if message contains a subscription, - * otherwise send an empty response to complete connection handshaking - */ + /* Unlock server port */ - subscriber_lock = subscriber->lock; - if (size) - subscr_subscribe((struct tipc_subscr *)data, subscriber); - else - tipc_send(subscriber->port_ref, 1, &msg_sect); + server_port_ref = subscriber->port_ref; + spin_unlock_bh(subscriber->lock); - spin_unlock_bh(subscriber_lock); + /* Send an ACK- to complete connection handshaking */ + + tipc_send(server_port_ref, 1, &msg_sect); + + /* Handle optional subscription request */ + + if (size != 0) { + subscr_conn_msg_event(subscriber, server_port_ref, + buf, data, size); + } } int tipc_subscr_start(void) @@ -572,8 +599,8 @@ void tipc_subscr_stop(void) list_for_each_entry_safe(subscriber, subscriber_temp, &topsrv.subscriber_list, subscriber_list) { - tipc_ref_lock(subscriber->ref); subscriber_lock = subscriber->lock; + spin_lock_bh(subscriber_lock); subscr_terminate(subscriber); spin_unlock_bh(subscriber_lock); } diff --git a/net/tipc/subscr.h b/net/tipc/subscr.h index 3e3e0265146e..b9af687b6368 100644 --- a/net/tipc/subscr.h +++ b/net/tipc/subscr.h @@ -49,12 +49,12 @@ typedef void (*tipc_subscr_event) (struct subscription *sub, * @timeout: duration of subscription (in ms) * @filter: event filtering to be done for subscription * @event_cb: routine invoked when a subscription event is detected + * @server_ref: object reference of server port associated with subscription * @swap: indicates if subscriber uses opposite endianness in its messages * @evt: template for events generated by subscription * @subscription_list: adjacent subscriptions in subscriber's subscription list * @nameseq_list: adjacent subscriptions in name sequence's subscription list * @timer_ref: reference to timer governing subscription duration (may be NULL) - * @owner: pointer to subscriber object associated with this subscription */ struct subscription { @@ -62,12 +62,12 @@ struct subscription { u32 timeout; u32 filter; tipc_subscr_event event_cb; + u32 server_ref; int swap; struct tipc_event evt; struct list_head subscription_list; struct list_head nameseq_list; struct timer_list timer; - struct subscriber *owner; }; int tipc_subscr_overlap(struct subscription * sub, -- cgit v1.2.2 From 5b06c85c3b96fa8db632f1ee94f99a2bd0215f3a Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 19 May 2008 13:30:13 -0700 Subject: tipc: Cosmetic cleanup of topology service code This patch contains a set of cosmetic changes to TIPC's network topology service subsystem, including: - updates to comments (including copyright dates) - re-ordering structure fields to group them more logically - removal of optional debugging code that is no longer required - minor changes to whitespace to conform to Linux coding conventions Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/subscr.c | 8 +++----- net/tipc/subscr.h | 20 ++++++++++---------- 2 files changed, 13 insertions(+), 15 deletions(-) (limited to 'net') diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c index 7c62791eb0cc..0326d3060bc7 100644 --- a/net/tipc/subscr.c +++ b/net/tipc/subscr.c @@ -1,8 +1,8 @@ /* - * net/tipc/subscr.c: TIPC subscription service + * net/tipc/subscr.c: TIPC network topology service * * Copyright (c) 2000-2006, Ericsson AB - * Copyright (c) 2005, Wind River Systems + * Copyright (c) 2005-2007, Wind River Systems * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -36,10 +36,10 @@ #include "core.h" #include "dbg.h" -#include "subscr.h" #include "name_table.h" #include "port.h" #include "ref.h" +#include "subscr.h" /** * struct subscriber - TIPC network topology subscriber @@ -149,8 +149,6 @@ void tipc_subscr_report_overlap(struct subscription *sub, u32 node, int must) { - dbg("Rep overlap %u:%u,%u<->%u,%u\n", sub->seq.type, sub->seq.lower, - sub->seq.upper, found_lower, found_upper); if (!tipc_subscr_overlap(sub, found_lower, found_upper)) return; if (!must && !(sub->filter & TIPC_SUB_PORTS)) diff --git a/net/tipc/subscr.h b/net/tipc/subscr.h index b9af687b6368..45d89bf4d202 100644 --- a/net/tipc/subscr.h +++ b/net/tipc/subscr.h @@ -1,8 +1,8 @@ /* - * net/tipc/subscr.h: Include file for TIPC subscription service + * net/tipc/subscr.h: Include file for TIPC network topology service * * Copyright (c) 2003-2006, Ericsson AB - * Copyright (c) 2005, Wind River Systems + * Copyright (c) 2005-2007, Wind River Systems * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -49,12 +49,12 @@ typedef void (*tipc_subscr_event) (struct subscription *sub, * @timeout: duration of subscription (in ms) * @filter: event filtering to be done for subscription * @event_cb: routine invoked when a subscription event is detected + * @timer: timer governing subscription duration (optional) + * @nameseq_list: adjacent subscriptions in name sequence's subscription list + * @subscription_list: adjacent subscriptions in subscriber's subscription list * @server_ref: object reference of server port associated with subscription * @swap: indicates if subscriber uses opposite endianness in its messages * @evt: template for events generated by subscription - * @subscription_list: adjacent subscriptions in subscriber's subscription list - * @nameseq_list: adjacent subscriptions in name sequence's subscription list - * @timer_ref: reference to timer governing subscription duration (may be NULL) */ struct subscription { @@ -62,19 +62,19 @@ struct subscription { u32 timeout; u32 filter; tipc_subscr_event event_cb; + struct timer_list timer; + struct list_head nameseq_list; + struct list_head subscription_list; u32 server_ref; int swap; struct tipc_event evt; - struct list_head subscription_list; - struct list_head nameseq_list; - struct timer_list timer; }; -int tipc_subscr_overlap(struct subscription * sub, +int tipc_subscr_overlap(struct subscription *sub, u32 found_lower, u32 found_upper); -void tipc_subscr_report_overlap(struct subscription * sub, +void tipc_subscr_report_overlap(struct subscription *sub, u32 found_lower, u32 found_upper, u32 event, -- cgit v1.2.2 From d62c612ef8a66be534a3ada598cfa28d40cd0b3c Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 19 May 2008 13:45:33 -0700 Subject: netns: Introduce sysctl root for read-only net sysctls. This one stores all ctl-heads in one list and restricts the permissions not give write access to non-init net namespaces. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/sysctl_net.c | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+) (limited to 'net') diff --git a/net/sysctl_net.c b/net/sysctl_net.c index b4f0525f91af..d8e79162724c 100644 --- a/net/sysctl_net.c +++ b/net/sysctl_net.c @@ -40,6 +40,27 @@ static struct ctl_table_root net_sysctl_root = { .lookup = net_ctl_header_lookup, }; +static LIST_HEAD(net_sysctl_ro_tables); +static struct list_head *net_ctl_ro_header_lookup(struct ctl_table_root *root, + struct nsproxy *namespaces) +{ + return &net_sysctl_ro_tables; +} + +static int net_ctl_ro_header_perms(struct ctl_table_root *root, + struct nsproxy *namespaces, struct ctl_table *table) +{ + if (namespaces->net_ns == &init_net) + return table->mode; + else + return table->mode & ~0222; +} + +static struct ctl_table_root net_sysctl_ro_root = { + .lookup = net_ctl_ro_header_lookup, + .permissions = net_ctl_ro_header_perms, +}; + static int sysctl_net_init(struct net *net) { INIT_LIST_HEAD(&net->sysctl_table_headers); @@ -64,6 +85,7 @@ static __init int sysctl_init(void) if (ret) goto out; register_sysctl_root(&net_sysctl_root); + register_sysctl_root(&net_sysctl_ro_root); out: return ret; } @@ -80,6 +102,14 @@ struct ctl_table_header *register_net_sysctl_table(struct net *net, } EXPORT_SYMBOL_GPL(register_net_sysctl_table); +struct ctl_table_header *register_net_sysctl_rotable(const + struct ctl_path *path, struct ctl_table *table) +{ + return __register_sysctl_paths(&net_sysctl_ro_root, + &init_nsproxy, path, table); +} +EXPORT_SYMBOL_GPL(register_net_sysctl_rotable); + void unregister_net_sysctl_table(struct ctl_table_header *header) { unregister_sysctl_table(header); -- cgit v1.2.2 From d5a4502e9efa534212484fd691339f6469cf95ff Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 19 May 2008 13:49:52 -0700 Subject: netns: Register net/core/ sysctls at read-only root. Most of the net/core/xxx sysctls are read-only now, but this goal is achieved with excessive memory consumption in each namespace - the whole table is cloned and most of the entries in it are ~= 0222. Split it into two parts and register (the largest) one at the read-only root. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/core/sysctl_net_core.c | 39 +++++++++++++++++++-------------------- 1 file changed, 19 insertions(+), 20 deletions(-) (limited to 'net') diff --git a/net/core/sysctl_net_core.c b/net/core/sysctl_net_core.c index 5fc801057244..a570e2af22cb 100644 --- a/net/core/sysctl_net_core.c +++ b/net/core/sysctl_net_core.c @@ -124,14 +124,6 @@ static struct ctl_table net_core_table[] = { }, #endif /* CONFIG_XFRM */ #endif /* CONFIG_NET */ - { - .ctl_name = NET_CORE_SOMAXCONN, - .procname = "somaxconn", - .data = &init_net.core.sysctl_somaxconn, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = &proc_dointvec - }, { .ctl_name = NET_CORE_BUDGET, .procname = "netdev_budget", @@ -151,6 +143,18 @@ static struct ctl_table net_core_table[] = { { .ctl_name = 0 } }; +static struct ctl_table netns_core_table[] = { + { + .ctl_name = NET_CORE_SOMAXCONN, + .procname = "somaxconn", + .data = &init_net.core.sysctl_somaxconn, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec + }, + { .ctl_name = 0 } +}; + static __net_initdata struct ctl_path net_core_path[] = { { .procname = "net", .ctl_name = CTL_NET, }, { .procname = "core", .ctl_name = NET_CORE, }, @@ -159,23 +163,17 @@ static __net_initdata struct ctl_path net_core_path[] = { static __net_init int sysctl_core_net_init(struct net *net) { - struct ctl_table *tbl, *tmp; + struct ctl_table *tbl; net->core.sysctl_somaxconn = SOMAXCONN; - tbl = net_core_table; + tbl = netns_core_table; if (net != &init_net) { - tbl = kmemdup(tbl, sizeof(net_core_table), GFP_KERNEL); + tbl = kmemdup(tbl, sizeof(netns_core_table), GFP_KERNEL); if (tbl == NULL) goto err_dup; - for (tmp = tbl; tmp->procname; tmp++) { - if (tmp->data >= (void *)&init_net && - tmp->data < (void *)(&init_net + 1)) - tmp->data += (char *)net - (char *)&init_net; - else - tmp->mode &= ~0222; - } + tbl[0].data = &net->core.sysctl_somaxconn; } net->core.sysctl_hdr = register_net_sysctl_table(net, @@ -186,7 +184,7 @@ static __net_init int sysctl_core_net_init(struct net *net) return 0; err_reg: - if (tbl != net_core_table) + if (tbl != netns_core_table) kfree(tbl); err_dup: return -ENOMEM; @@ -198,7 +196,7 @@ static __net_exit void sysctl_core_net_exit(struct net *net) tbl = net->core.sysctl_hdr->ctl_table_arg; unregister_net_sysctl_table(net->core.sysctl_hdr); - BUG_ON(tbl == net_core_table); + BUG_ON(tbl == netns_core_table); kfree(tbl); } @@ -209,6 +207,7 @@ static __net_initdata struct pernet_operations sysctl_core_ops = { static __init int sysctl_core_init(void) { + register_net_sysctl_rotable(net_core_path, net_core_table); return register_pernet_subsys(&sysctl_core_ops); } -- cgit v1.2.2 From 0a64b4b811025ce0386ad84d81504e4ff7985856 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 19 May 2008 13:51:29 -0700 Subject: inet: Rename fragmentation sysctl-related functions/variables. The fragments sysctls also contains some, that are to be visible, but read-only in net namespaces. The naming in net/core/sysctl_net_core.c is - tables, that are to be registered in namespaces have a "ns" word in their names. So rename ones in ipv4/ip_fragment.c and ipv6/reassembly.c to fit this. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/ip_fragment.c | 18 +++++++++--------- net/ipv6/reassembly.c | 18 +++++++++--------- 2 files changed, 18 insertions(+), 18 deletions(-) (limited to 'net') diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c index cd6ce6ac6358..7f102eeb618e 100644 --- a/net/ipv4/ip_fragment.c +++ b/net/ipv4/ip_fragment.c @@ -598,7 +598,7 @@ int ip_defrag(struct sk_buff *skb, u32 user) #ifdef CONFIG_SYSCTL static int zero; -static struct ctl_table ip4_frags_ctl_table[] = { +static struct ctl_table ip4_frags_ns_ctl_table[] = { { .ctl_name = NET_IPV4_IPFRAG_HIGH_THRESH, .procname = "ipfrag_high_thresh", @@ -644,14 +644,14 @@ static struct ctl_table ip4_frags_ctl_table[] = { { } }; -static int ip4_frags_ctl_register(struct net *net) +static int ip4_frags_ns_ctl_register(struct net *net) { struct ctl_table *table; struct ctl_table_header *hdr; - table = ip4_frags_ctl_table; + table = ip4_frags_ns_ctl_table; if (net != &init_net) { - table = kmemdup(table, sizeof(ip4_frags_ctl_table), GFP_KERNEL); + table = kmemdup(table, sizeof(ip4_frags_ns_ctl_table), GFP_KERNEL); if (table == NULL) goto err_alloc; @@ -676,7 +676,7 @@ err_alloc: return -ENOMEM; } -static void ip4_frags_ctl_unregister(struct net *net) +static void ip4_frags_ns_ctl_unregister(struct net *net) { struct ctl_table *table; @@ -685,12 +685,12 @@ static void ip4_frags_ctl_unregister(struct net *net) kfree(table); } #else -static inline int ip4_frags_ctl_register(struct net *net) +static inline int ip4_frags_ns_ctl_register(struct net *net) { return 0; } -static inline void ip4_frags_ctl_unregister(struct net *net) +static inline void ip4_frags_ns_ctl_unregister(struct net *net) { } #endif @@ -714,12 +714,12 @@ static int ipv4_frags_init_net(struct net *net) inet_frags_init_net(&net->ipv4.frags); - return ip4_frags_ctl_register(net); + return ip4_frags_ns_ctl_register(net); } static void ipv4_frags_exit_net(struct net *net) { - ip4_frags_ctl_unregister(net); + ip4_frags_ns_ctl_unregister(net); inet_frags_exit_net(&net->ipv4.frags, &ip4_frags); } diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c index 798cabc7535b..7e008de87117 100644 --- a/net/ipv6/reassembly.c +++ b/net/ipv6/reassembly.c @@ -632,7 +632,7 @@ static struct inet6_protocol frag_protocol = }; #ifdef CONFIG_SYSCTL -static struct ctl_table ip6_frags_ctl_table[] = { +static struct ctl_table ip6_frags_ns_ctl_table[] = { { .ctl_name = NET_IPV6_IP6FRAG_HIGH_THRESH, .procname = "ip6frag_high_thresh", @@ -670,14 +670,14 @@ static struct ctl_table ip6_frags_ctl_table[] = { { } }; -static int ip6_frags_sysctl_register(struct net *net) +static int ip6_frags_ns_sysctl_register(struct net *net) { struct ctl_table *table; struct ctl_table_header *hdr; - table = ip6_frags_ctl_table; + table = ip6_frags_ns_ctl_table; if (net != &init_net) { - table = kmemdup(table, sizeof(ip6_frags_ctl_table), GFP_KERNEL); + table = kmemdup(table, sizeof(ip6_frags_ns_ctl_table), GFP_KERNEL); if (table == NULL) goto err_alloc; @@ -701,7 +701,7 @@ err_alloc: return -ENOMEM; } -static void ip6_frags_sysctl_unregister(struct net *net) +static void ip6_frags_ns_sysctl_unregister(struct net *net) { struct ctl_table *table; @@ -710,12 +710,12 @@ static void ip6_frags_sysctl_unregister(struct net *net) kfree(table); } #else -static inline int ip6_frags_sysctl_register(struct net *net) +static inline int ip6_frags_ns_sysctl_register(struct net *net) { return 0; } -static inline void ip6_frags_sysctl_unregister(struct net *net) +static inline void ip6_frags_ns_sysctl_unregister(struct net *net) { } #endif @@ -728,12 +728,12 @@ static int ipv6_frags_init_net(struct net *net) inet_frags_init_net(&net->ipv6.frags); - return ip6_frags_sysctl_register(net); + return ip6_frags_ns_sysctl_register(net); } static void ipv6_frags_exit_net(struct net *net) { - ip6_frags_sysctl_unregister(net); + ip6_frags_ns_sysctl_unregister(net); inet_frags_exit_net(&net->ipv6.frags, &ip6_frags); } -- cgit v1.2.2 From 0002c630c4ee7a3c6b1d87e34bfd6ce9694b49be Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 19 May 2008 13:52:28 -0700 Subject: ipv6: In fragmentation code, handle error returned from register_pernet_subsys. The error code is ignored now, but ipv6 is a module and one can be loaded under memory pressure, so the error may occur (in theory). Besides, I'm going to handle error returned from registering a read-only part of the table, so ignoring this one, while handing the other one would look strange. (However, this possibility of error is rather small, so I'm not sure whether this is a candidate for current net tree). Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv6/reassembly.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c index 7e008de87117..130d6f6b6a6e 100644 --- a/net/ipv6/reassembly.c +++ b/net/ipv6/reassembly.c @@ -750,7 +750,9 @@ int __init ipv6_frag_init(void) if (ret) goto out; - register_pernet_subsys(&ip6_frags_ops); + ret = register_pernet_subsys(&ip6_frags_ops); + if (ret) + goto err_pernet; ip6_frags.hashfn = ip6_hashfn; ip6_frags.constructor = ip6_frag_init; @@ -763,6 +765,10 @@ int __init ipv6_frag_init(void) inet_frags_init(&ip6_frags); out: return ret; + +err_pernet: + inet6_del_protocol(&frag_protocol, IPPROTO_FRAGMENT); + goto out; } void ipv6_frag_exit(void) -- cgit v1.2.2 From 7d291ebb834278e30c211b26fb7076adcb636ad9 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 19 May 2008 13:53:02 -0700 Subject: inet: Register fragmentation some ctls at read-only root. Parts of fragments-related sysctls are read-only, but this is done by cloning all the tables and dropping write-bits from mode. Do the same but with read-only root. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/ip_fragment.c | 16 ++++++++++++++-- net/ipv6/reassembly.c | 35 ++++++++++++++++++++++++++++++++++- 2 files changed, 48 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c index 7f102eeb618e..be1cb89a8d5a 100644 --- a/net/ipv4/ip_fragment.c +++ b/net/ipv4/ip_fragment.c @@ -624,6 +624,10 @@ static struct ctl_table ip4_frags_ns_ctl_table[] = { .proc_handler = &proc_dointvec_jiffies, .strategy = &sysctl_jiffies }, + { } +}; + +static struct ctl_table ip4_frags_ctl_table[] = { { .ctl_name = NET_IPV4_IPFRAG_SECRET_INTERVAL, .procname = "ipfrag_secret_interval", @@ -658,8 +662,6 @@ static int ip4_frags_ns_ctl_register(struct net *net) table[0].data = &net->ipv4.frags.high_thresh; table[1].data = &net->ipv4.frags.low_thresh; table[2].data = &net->ipv4.frags.timeout; - table[3].mode &= ~0222; - table[4].mode &= ~0222; } hdr = register_net_sysctl_table(net, net_ipv4_ctl_path, table); @@ -684,6 +686,11 @@ static void ip4_frags_ns_ctl_unregister(struct net *net) unregister_net_sysctl_table(net->ipv4.frags_hdr); kfree(table); } + +static void ip4_frags_ctl_register(void) +{ + register_net_sysctl_rotable(net_ipv4_ctl_path, ip4_frags_ctl_table); +} #else static inline int ip4_frags_ns_ctl_register(struct net *net) { @@ -693,6 +700,10 @@ static inline int ip4_frags_ns_ctl_register(struct net *net) static inline void ip4_frags_ns_ctl_unregister(struct net *net) { } + +static inline void ip4_frags_ctl_register(void) +{ +} #endif static int ipv4_frags_init_net(struct net *net) @@ -730,6 +741,7 @@ static struct pernet_operations ip4_frags_ops = { void __init ipfrag_init(void) { + ip4_frags_ctl_register(); register_pernet_subsys(&ip4_frags_ops); ip4_frags.hashfn = ip4_hashfn; ip4_frags.constructor = ip4_frag_init; diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c index 130d6f6b6a6e..9391a6949b96 100644 --- a/net/ipv6/reassembly.c +++ b/net/ipv6/reassembly.c @@ -658,6 +658,10 @@ static struct ctl_table ip6_frags_ns_ctl_table[] = { .proc_handler = &proc_dointvec_jiffies, .strategy = &sysctl_jiffies, }, + { } +}; + +static struct ctl_table ip6_frags_ctl_table[] = { { .ctl_name = NET_IPV6_IP6FRAG_SECRET_INTERVAL, .procname = "ip6frag_secret_interval", @@ -684,7 +688,6 @@ static int ip6_frags_ns_sysctl_register(struct net *net) table[0].data = &net->ipv6.frags.high_thresh; table[1].data = &net->ipv6.frags.low_thresh; table[2].data = &net->ipv6.frags.timeout; - table[3].mode &= ~0222; } hdr = register_net_sysctl_table(net, net_ipv6_ctl_path, table); @@ -709,6 +712,20 @@ static void ip6_frags_ns_sysctl_unregister(struct net *net) unregister_net_sysctl_table(net->ipv6.sysctl.frags_hdr); kfree(table); } + +static struct ctl_table_header *ip6_ctl_header; + +static int ip6_frags_sysctl_register(void) +{ + ip6_ctl_header = register_net_sysctl_rotable(net_ipv6_ctl_path, + ip6_frags_ctl_table); + return ip6_ctl_header == NULL ? -ENOMEM : 0; +} + +static void ip6_frags_sysctl_unregister(void) +{ + unregister_net_sysctl_table(ip6_ctl_header); +} #else static inline int ip6_frags_ns_sysctl_register(struct net *net) { @@ -718,6 +735,15 @@ static inline int ip6_frags_ns_sysctl_register(struct net *net) static inline void ip6_frags_ns_sysctl_unregister(struct net *net) { } + +static inline int ip6_frags_sysctl_register(void) +{ + return 0; +} + +static inline void ip6_frags_sysctl_unregister(void) +{ +} #endif static int ipv6_frags_init_net(struct net *net) @@ -750,6 +776,10 @@ int __init ipv6_frag_init(void) if (ret) goto out; + ret = ip6_frags_sysctl_register(); + if (ret) + goto err_sysctl; + ret = register_pernet_subsys(&ip6_frags_ops); if (ret) goto err_pernet; @@ -767,6 +797,8 @@ out: return ret; err_pernet: + ip6_frags_sysctl_unregister(); +err_sysctl: inet6_del_protocol(&frag_protocol, IPPROTO_FRAGMENT); goto out; } @@ -774,6 +806,7 @@ err_pernet: void ipv6_frag_exit(void) { inet_frags_fini(&ip6_frags); + ip6_frags_sysctl_unregister(); unregister_pernet_subsys(&ip6_frags_ops); inet6_del_protocol(&frag_protocol, IPPROTO_FRAGMENT); } -- cgit v1.2.2 From 34ac2573e88c4f80fc5e219d8012ea383a788803 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 19 May 2008 13:53:30 -0700 Subject: ipv6: Register some net/ipv6/ core sysctls at read-only root. There are some sysctls left to be switched to read-only, but they are all in ipv6, so complete with them. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv6/sysctl_net_ipv6.c | 29 ++++++++++++++++++++++------- 1 file changed, 22 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/ipv6/sysctl_net_ipv6.c b/net/ipv6/sysctl_net_ipv6.c index 3804dcbbfab0..5c99274558bf 100644 --- a/net/ipv6/sysctl_net_ipv6.c +++ b/net/ipv6/sysctl_net_ipv6.c @@ -37,6 +37,10 @@ static ctl_table ipv6_table_template[] = { .mode = 0644, .proc_handler = &proc_dointvec }, + { .ctl_name = 0 } +}; + +static ctl_table ipv6_table[] = { { .ctl_name = NET_IPV6_MLD_MAX_MSF, .procname = "mld_max_msf", @@ -80,12 +84,6 @@ static int ipv6_sysctl_net_init(struct net *net) ipv6_table[2].data = &net->ipv6.sysctl.bindv6only; - /* We don't want this value to be per namespace, it should be global - to all namespaces, so make it read-only when we are not in the - init network namespace */ - if (net != &init_net) - ipv6_table[3].mode = 0444; - net->ipv6.sysctl.table = register_net_sysctl_table(net, net_ipv6_ctl_path, ipv6_table); if (!net->ipv6.sysctl.table) @@ -126,12 +124,29 @@ static struct pernet_operations ipv6_sysctl_net_ops = { .exit = ipv6_sysctl_net_exit, }; +static struct ctl_table_header *ip6_header; + int ipv6_sysctl_register(void) { - return register_pernet_subsys(&ipv6_sysctl_net_ops); + int err = -ENOMEM;; + + ip6_header = register_net_sysctl_rotable(net_ipv6_ctl_path, ipv6_table); + if (ip6_header == NULL) + goto out; + + err = register_pernet_subsys(&ipv6_sysctl_net_ops); + if (err) + goto err_pernet; +out: + return err; + +err_pernet: + unregister_net_sysctl_table(ip6_header); + goto out; } void ipv6_sysctl_unregister(void) { + unregister_net_sysctl_table(ip6_header); unregister_pernet_subsys(&ipv6_sysctl_net_ops); } -- cgit v1.2.2 From 96e74088f1da4d9a53735a4a57a4f984f86b75c6 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 21 May 2008 14:12:46 -0700 Subject: net: The dev->get_stats pointer is not NULL nowadays. And so does the pointer is returns, but sysfs and netlinks still check for both cases. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/core/net-sysfs.c | 9 ++++----- net/core/rtnetlink.c | 20 ++++++++------------ 2 files changed, 12 insertions(+), 17 deletions(-) (limited to 'net') diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index 90e2177af081..dccd737ea2e3 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -242,11 +242,11 @@ static ssize_t netstat_show(const struct device *d, offset % sizeof(unsigned long) != 0); read_lock(&dev_base_lock); - if (dev_isalive(dev) && dev->get_stats && - (stats = (*dev->get_stats)(dev))) + if (dev_isalive(dev)) { + stats = dev->get_stats(dev); ret = sprintf(buf, fmt_ulong, *(unsigned long *)(((u8 *) stats) + offset)); - + } read_unlock(&dev_base_lock); return ret; } @@ -457,8 +457,7 @@ int netdev_register_kobject(struct net_device *net) strlcpy(dev->bus_id, net->name, BUS_ID_SIZE); #ifdef CONFIG_SYSFS - if (net->get_stats) - *groups++ = &netstat_group; + *groups++ = &netstat_group; #ifdef CONFIG_WIRELESS_EXT if (net->wireless_handlers && net->wireless_handlers->get_wireless_stats) diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index cf857c4dc7b1..ca32ddb8ad17 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -606,6 +606,8 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev, { struct ifinfomsg *ifm; struct nlmsghdr *nlh; + struct net_device_stats *stats; + struct nlattr *attr; nlh = nlmsg_put(skb, pid, seq, type, sizeof(*ifm), flags); if (nlh == NULL) @@ -652,19 +654,13 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev, NLA_PUT(skb, IFLA_BROADCAST, dev->addr_len, dev->broadcast); } - if (dev->get_stats) { - struct net_device_stats *stats = dev->get_stats(dev); - if (stats) { - struct nlattr *attr; + attr = nla_reserve(skb, IFLA_STATS, + sizeof(struct rtnl_link_stats)); + if (attr == NULL) + goto nla_put_failure; - attr = nla_reserve(skb, IFLA_STATS, - sizeof(struct rtnl_link_stats)); - if (attr == NULL) - goto nla_put_failure; - - copy_rtnl_link_stats(nla_data(attr), stats); - } - } + stats = dev->get_stats(dev); + copy_rtnl_link_stats(nla_data(attr), stats); if (dev->rtnl_link_ops) { if (rtnl_link_fill(skb, dev) < 0) -- cgit v1.2.2 From a339f1c881fdb8092ef9b118610307e10e885fc8 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 21 May 2008 14:13:47 -0700 Subject: bridge: Use on-device stats instead of private ones. Even though bridges require 6 fields from struct net_device_stats, the on-device stats are always there, so we may just use them. The br_dev_get_stats is no longer required after this. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/bridge/br_device.c | 11 ++--------- net/bridge/br_forward.c | 2 +- net/bridge/br_input.c | 10 +++++----- net/bridge/br_private.h | 1 - 4 files changed, 8 insertions(+), 16 deletions(-) (limited to 'net') diff --git a/net/bridge/br_device.c b/net/bridge/br_device.c index bf7787395fe0..626c7795ae30 100644 --- a/net/bridge/br_device.c +++ b/net/bridge/br_device.c @@ -21,12 +21,6 @@ #include #include "br_private.h" -static struct net_device_stats *br_dev_get_stats(struct net_device *dev) -{ - struct net_bridge *br = netdev_priv(dev); - return &br->statistics; -} - /* net device transmit always called with no BH (preempt_disabled) */ int br_dev_xmit(struct sk_buff *skb, struct net_device *dev) { @@ -34,8 +28,8 @@ int br_dev_xmit(struct sk_buff *skb, struct net_device *dev) const unsigned char *dest = skb->data; struct net_bridge_fdb_entry *dst; - br->statistics.tx_packets++; - br->statistics.tx_bytes += skb->len; + dev->stats.tx_packets++; + dev->stats.tx_bytes += skb->len; skb_reset_mac_header(skb); skb_pull(skb, ETH_HLEN); @@ -161,7 +155,6 @@ void br_dev_setup(struct net_device *dev) ether_setup(dev); dev->do_ioctl = br_dev_ioctl; - dev->get_stats = br_dev_get_stats; dev->hard_start_xmit = br_dev_xmit; dev->open = br_dev_open; dev->set_multicast_list = br_dev_set_multicast_list; diff --git a/net/bridge/br_forward.c b/net/bridge/br_forward.c index bdd7c35c3c7b..a4711674b3df 100644 --- a/net/bridge/br_forward.c +++ b/net/bridge/br_forward.c @@ -115,7 +115,7 @@ static void br_flood(struct net_bridge *br, struct sk_buff *skb, struct sk_buff *skb2; if ((skb2 = skb_clone(skb, GFP_ATOMIC)) == NULL) { - br->statistics.tx_dropped++; + br->dev->stats.tx_dropped++; kfree_skb(skb); return; } diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c index 255c00f60ce7..fa0f5711a996 100644 --- a/net/bridge/br_input.c +++ b/net/bridge/br_input.c @@ -24,13 +24,13 @@ const u8 br_group_address[ETH_ALEN] = { 0x01, 0x80, 0xc2, 0x00, 0x00, 0x00 }; static void br_pass_frame_up(struct net_bridge *br, struct sk_buff *skb) { - struct net_device *indev; + struct net_device *indev, *brdev = br->dev; - br->statistics.rx_packets++; - br->statistics.rx_bytes += skb->len; + brdev->stats.rx_packets++; + brdev->stats.rx_bytes += skb->len; indev = skb->dev; - skb->dev = br->dev; + skb->dev = brdev; NF_HOOK(PF_BRIDGE, NF_BR_LOCAL_IN, skb, indev, NULL, netif_receive_skb); @@ -64,7 +64,7 @@ int br_handle_frame_finish(struct sk_buff *skb) dst = NULL; if (is_multicast_ether_addr(dest)) { - br->statistics.multicast++; + br->dev->stats.multicast++; skb2 = skb; } else if ((dst = __br_fdb_get(br, dest)) && dst->is_local) { skb2 = skb; diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h index c11b554fd109..0243cb489edd 100644 --- a/net/bridge/br_private.h +++ b/net/bridge/br_private.h @@ -90,7 +90,6 @@ struct net_bridge spinlock_t lock; struct list_head port_list; struct net_device *dev; - struct net_device_stats statistics; spinlock_t hash_lock; struct hlist_head hash[BR_HASH_SIZE]; struct list_head age_list; -- cgit v1.2.2 From addd68eb6fe6f10017974947c90cc85ab2848cf2 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 21 May 2008 14:14:22 -0700 Subject: ipgre: Use on-device stats instead of private ones. Just switch from tunnel->stat to tunnel->dev->stats. The ip_tunnel->stat member itself will be removed after I fix its other users (very soon). Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/ip_gre.c | 32 ++++++++++++++------------------ 1 file changed, 14 insertions(+), 18 deletions(-) (limited to 'net') diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c index 2ada033406de..eede36e55702 100644 --- a/net/ipv4/ip_gre.c +++ b/net/ipv4/ip_gre.c @@ -617,6 +617,8 @@ static int ipgre_rcv(struct sk_buff *skb) read_lock(&ipgre_lock); if ((tunnel = ipgre_tunnel_lookup(dev_net(skb->dev), iph->saddr, iph->daddr, key)) != NULL) { + struct net_device_stats *stats = &tunnel->dev->stats; + secpath_reset(skb); skb->protocol = *(__be16*)(h + 2); @@ -641,28 +643,28 @@ static int ipgre_rcv(struct sk_buff *skb) /* Looped back packet, drop it! */ if (skb->rtable->fl.iif == 0) goto drop; - tunnel->stat.multicast++; + stats->multicast++; skb->pkt_type = PACKET_BROADCAST; } #endif if (((flags&GRE_CSUM) && csum) || (!(flags&GRE_CSUM) && tunnel->parms.i_flags&GRE_CSUM)) { - tunnel->stat.rx_crc_errors++; - tunnel->stat.rx_errors++; + stats->rx_crc_errors++; + stats->rx_errors++; goto drop; } if (tunnel->parms.i_flags&GRE_SEQ) { if (!(flags&GRE_SEQ) || (tunnel->i_seqno && (s32)(seqno - tunnel->i_seqno) < 0)) { - tunnel->stat.rx_fifo_errors++; - tunnel->stat.rx_errors++; + stats->rx_fifo_errors++; + stats->rx_errors++; goto drop; } tunnel->i_seqno = seqno + 1; } - tunnel->stat.rx_packets++; - tunnel->stat.rx_bytes += skb->len; + stats->rx_packets++; + stats->rx_bytes += skb->len; skb->dev = tunnel->dev; dst_release(skb->dst); skb->dst = NULL; @@ -684,7 +686,7 @@ drop_nolock: static int ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) { struct ip_tunnel *tunnel = netdev_priv(dev); - struct net_device_stats *stats = &tunnel->stat; + struct net_device_stats *stats = &tunnel->dev->stats; struct iphdr *old_iph = ip_hdr(skb); struct iphdr *tiph; u8 tos; @@ -698,7 +700,7 @@ static int ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) int mtu; if (tunnel->recursion++) { - tunnel->stat.collisions++; + stats->collisions++; goto tx_error; } @@ -714,7 +716,7 @@ static int ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) /* NBMA tunnel */ if (skb->dst == NULL) { - tunnel->stat.tx_fifo_errors++; + stats->tx_fifo_errors++; goto tx_error; } @@ -765,7 +767,7 @@ static int ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) .tos = RT_TOS(tos) } }, .proto = IPPROTO_GRE }; if (ip_route_output_key(dev_net(dev), &rt, &fl)) { - tunnel->stat.tx_carrier_errors++; + stats->tx_carrier_errors++; goto tx_error; } } @@ -773,7 +775,7 @@ static int ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) if (tdev == dev) { ip_rt_put(rt); - tunnel->stat.collisions++; + stats->collisions++; goto tx_error; } @@ -1098,11 +1100,6 @@ done: return err; } -static struct net_device_stats *ipgre_tunnel_get_stats(struct net_device *dev) -{ - return &(((struct ip_tunnel*)netdev_priv(dev))->stat); -} - static int ipgre_tunnel_change_mtu(struct net_device *dev, int new_mtu) { struct ip_tunnel *tunnel = netdev_priv(dev); @@ -1228,7 +1225,6 @@ static void ipgre_tunnel_setup(struct net_device *dev) dev->uninit = ipgre_tunnel_uninit; dev->destructor = free_netdev; dev->hard_start_xmit = ipgre_tunnel_xmit; - dev->get_stats = ipgre_tunnel_get_stats; dev->do_ioctl = ipgre_tunnel_ioctl; dev->change_mtu = ipgre_tunnel_change_mtu; -- cgit v1.2.2 From 50f59cea075875d84018a5fc62cf2f5e6173a919 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 21 May 2008 14:15:16 -0700 Subject: ipip: Use on-device stats instead of private ones. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/ipip.c | 22 ++++++++-------------- 1 file changed, 8 insertions(+), 14 deletions(-) (limited to 'net') diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c index 149111f08e8d..26c85c23ca4f 100644 --- a/net/ipv4/ipip.c +++ b/net/ipv4/ipip.c @@ -496,8 +496,8 @@ static int ipip_rcv(struct sk_buff *skb) skb->protocol = htons(ETH_P_IP); skb->pkt_type = PACKET_HOST; - tunnel->stat.rx_packets++; - tunnel->stat.rx_bytes += skb->len; + tunnel->dev->stats.rx_packets++; + tunnel->dev->stats.rx_bytes += skb->len; skb->dev = tunnel->dev; dst_release(skb->dst); skb->dst = NULL; @@ -520,7 +520,7 @@ static int ipip_rcv(struct sk_buff *skb) static int ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) { struct ip_tunnel *tunnel = netdev_priv(dev); - struct net_device_stats *stats = &tunnel->stat; + struct net_device_stats *stats = &tunnel->dev->stats; struct iphdr *tiph = &tunnel->parms.iph; u8 tos = tunnel->parms.iph.tos; __be16 df = tiph->frag_off; @@ -533,7 +533,7 @@ static int ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) int mtu; if (tunnel->recursion++) { - tunnel->stat.collisions++; + stats->collisions++; goto tx_error; } @@ -546,7 +546,7 @@ static int ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) if (!dst) { /* NBMA tunnel */ if ((rt = skb->rtable) == NULL) { - tunnel->stat.tx_fifo_errors++; + stats->tx_fifo_errors++; goto tx_error; } if ((dst = rt->rt_gateway) == 0) @@ -561,7 +561,7 @@ static int ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) .tos = RT_TOS(tos) } }, .proto = IPPROTO_IPIP }; if (ip_route_output_key(dev_net(dev), &rt, &fl)) { - tunnel->stat.tx_carrier_errors++; + stats->tx_carrier_errors++; goto tx_error_icmp; } } @@ -569,7 +569,7 @@ static int ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) if (tdev == dev) { ip_rt_put(rt); - tunnel->stat.collisions++; + stats->collisions++; goto tx_error; } @@ -579,7 +579,7 @@ static int ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) mtu = skb->dst ? dst_mtu(skb->dst) : dev->mtu; if (mtu < 68) { - tunnel->stat.collisions++; + stats->collisions++; ip_rt_put(rt); goto tx_error; } @@ -813,11 +813,6 @@ done: return err; } -static struct net_device_stats *ipip_tunnel_get_stats(struct net_device *dev) -{ - return &(((struct ip_tunnel*)netdev_priv(dev))->stat); -} - static int ipip_tunnel_change_mtu(struct net_device *dev, int new_mtu) { if (new_mtu < 68 || new_mtu > 0xFFF8 - sizeof(struct iphdr)) @@ -830,7 +825,6 @@ static void ipip_tunnel_setup(struct net_device *dev) { dev->uninit = ipip_tunnel_uninit; dev->hard_start_xmit = ipip_tunnel_xmit; - dev->get_stats = ipip_tunnel_get_stats; dev->do_ioctl = ipip_tunnel_ioctl; dev->change_mtu = ipip_tunnel_change_mtu; dev->destructor = free_netdev; -- cgit v1.2.2 From 4eecc107a85a5882e253bd97310d4e96300a2068 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 21 May 2008 14:15:46 -0700 Subject: sit: Use on-device stats instead of private ones. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv6/sit.c | 24 +++++++++--------------- 1 file changed, 9 insertions(+), 15 deletions(-) (limited to 'net') diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c index 5a6fab95569f..b0ee96187633 100644 --- a/net/ipv6/sit.c +++ b/net/ipv6/sit.c @@ -578,13 +578,13 @@ static int ipip6_rcv(struct sk_buff *skb) if ((tunnel->dev->priv_flags & IFF_ISATAP) && !isatap_chksrc(skb, iph, tunnel)) { - tunnel->stat.rx_errors++; + tunnel->dev->stats.rx_errors++; read_unlock(&ipip6_lock); kfree_skb(skb); return 0; } - tunnel->stat.rx_packets++; - tunnel->stat.rx_bytes += skb->len; + tunnel->dev->stats.rx_packets++; + tunnel->dev->stats.rx_bytes += skb->len; skb->dev = tunnel->dev; dst_release(skb->dst); skb->dst = NULL; @@ -624,7 +624,7 @@ static inline __be32 try_6to4(struct in6_addr *v6dst) static int ipip6_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) { struct ip_tunnel *tunnel = netdev_priv(dev); - struct net_device_stats *stats = &tunnel->stat; + struct net_device_stats *stats = &tunnel->dev->stats; struct iphdr *tiph = &tunnel->parms.iph; struct ipv6hdr *iph6 = ipv6_hdr(skb); u8 tos = tunnel->parms.iph.tos; @@ -638,7 +638,7 @@ static int ipip6_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) int addr_type; if (tunnel->recursion++) { - tunnel->stat.collisions++; + stats->collisions++; goto tx_error; } @@ -705,20 +705,20 @@ static int ipip6_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) .oif = tunnel->parms.link, .proto = IPPROTO_IPV6 }; if (ip_route_output_key(dev_net(dev), &rt, &fl)) { - tunnel->stat.tx_carrier_errors++; + stats->tx_carrier_errors++; goto tx_error_icmp; } } if (rt->rt_type != RTN_UNICAST) { ip_rt_put(rt); - tunnel->stat.tx_carrier_errors++; + stats->tx_carrier_errors++; goto tx_error_icmp; } tdev = rt->u.dst.dev; if (tdev == dev) { ip_rt_put(rt); - tunnel->stat.collisions++; + stats->collisions++; goto tx_error; } @@ -728,7 +728,7 @@ static int ipip6_tunnel_xmit(struct sk_buff *skb, struct net_device *dev) mtu = skb->dst ? dst_mtu(skb->dst) : dev->mtu; if (mtu < 68) { - tunnel->stat.collisions++; + stats->collisions++; ip_rt_put(rt); goto tx_error; } @@ -1003,11 +1003,6 @@ done: return err; } -static struct net_device_stats *ipip6_tunnel_get_stats(struct net_device *dev) -{ - return &(((struct ip_tunnel*)netdev_priv(dev))->stat); -} - static int ipip6_tunnel_change_mtu(struct net_device *dev, int new_mtu) { if (new_mtu < IPV6_MIN_MTU || new_mtu > 0xFFF8 - sizeof(struct iphdr)) @@ -1021,7 +1016,6 @@ static void ipip6_tunnel_setup(struct net_device *dev) dev->uninit = ipip6_tunnel_uninit; dev->destructor = free_netdev; dev->hard_start_xmit = ipip6_tunnel_xmit; - dev->get_stats = ipip6_tunnel_get_stats; dev->do_ioctl = ipip6_tunnel_ioctl; dev->change_mtu = ipip6_tunnel_change_mtu; -- cgit v1.2.2 From 2f4c02d40c0d59f4efc59801ad1498cb0f8550ea Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 21 May 2008 14:16:14 -0700 Subject: ipmr: Ipip tunnel uses on-device stats. The ipmr uses ipip tunnels for its purposes and updates the tunnels' stats, but the ipip driver is already switched to use on-device ones. Actually, this is a part of the patch #4 from this set. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/ipmr.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index 11700a4dcd95..65f12005785e 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c @@ -1230,8 +1230,8 @@ static void ipmr_queue_xmit(struct sk_buff *skb, struct mfc_cache *c, int vifi) if (vif->flags & VIFF_TUNNEL) { ip_encap(skb, vif->local, vif->remote); /* FIXME: extra output firewall step used to be here. --RR */ - ((struct ip_tunnel *)netdev_priv(vif->dev))->stat.tx_packets++; - ((struct ip_tunnel *)netdev_priv(vif->dev))->stat.tx_bytes+=skb->len; + vif->dev->stats.tx_packets++; + vif->dev->stats.tx_bytes += skb->len; } IPCB(skb)->flags |= IPSKB_FORWARDED; -- cgit v1.2.2 From 3dca02af38c11a970160387ab36ae6043feb03cd Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 21 May 2008 14:17:05 -0700 Subject: ip6tnl: Use on-device stats instead of private ones. This tunnel uses its own private structure and requires separate patch to switch from private stats to on-device ones. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv6/ip6_tunnel.c | 26 ++++++-------------------- 1 file changed, 6 insertions(+), 20 deletions(-) (limited to 'net') diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c index 2bda3ba100b1..37814810ac49 100644 --- a/net/ipv6/ip6_tunnel.c +++ b/net/ipv6/ip6_tunnel.c @@ -711,7 +711,7 @@ static int ip6_tnl_rcv(struct sk_buff *skb, __u16 protocol, } if (!ip6_tnl_rcv_ctl(t)) { - t->stat.rx_dropped++; + t->dev->stats.rx_dropped++; read_unlock(&ip6_tnl_lock); goto discard; } @@ -728,8 +728,8 @@ static int ip6_tnl_rcv(struct sk_buff *skb, __u16 protocol, dscp_ecn_decapsulate(t, ipv6h, skb); - t->stat.rx_packets++; - t->stat.rx_bytes += skb->len; + t->dev->stats.rx_packets++; + t->dev->stats.rx_bytes += skb->len; netif_rx(skb); read_unlock(&ip6_tnl_lock); return 0; @@ -849,7 +849,7 @@ static int ip6_tnl_xmit2(struct sk_buff *skb, __u32 *pmtu) { struct ip6_tnl *t = netdev_priv(dev); - struct net_device_stats *stats = &t->stat; + struct net_device_stats *stats = &t->dev->stats; struct ipv6hdr *ipv6h = ipv6_hdr(skb); struct ipv6_tel_txoption opt; struct dst_entry *dst; @@ -1043,11 +1043,11 @@ static int ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev) { struct ip6_tnl *t = netdev_priv(dev); - struct net_device_stats *stats = &t->stat; + struct net_device_stats *stats = &t->dev->stats; int ret; if (t->recursion++) { - t->stat.collisions++; + stats->collisions++; goto tx_err; } @@ -1288,19 +1288,6 @@ ip6_tnl_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) return err; } -/** - * ip6_tnl_get_stats - return the stats for tunnel device - * @dev: virtual device associated with tunnel - * - * Return: stats for device - **/ - -static struct net_device_stats * -ip6_tnl_get_stats(struct net_device *dev) -{ - return &(((struct ip6_tnl *)netdev_priv(dev))->stat); -} - /** * ip6_tnl_change_mtu - change mtu manually for tunnel device * @dev: virtual device associated with tunnel @@ -1334,7 +1321,6 @@ static void ip6_tnl_dev_setup(struct net_device *dev) dev->uninit = ip6_tnl_dev_uninit; dev->destructor = free_netdev; dev->hard_start_xmit = ip6_tnl_xmit; - dev->get_stats = ip6_tnl_get_stats; dev->do_ioctl = ip6_tnl_ioctl; dev->change_mtu = ip6_tnl_change_mtu; -- cgit v1.2.2 From cf3677ae19c2f62979b39143f5d2f6b3dfb3b3e4 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 21 May 2008 14:17:33 -0700 Subject: ipmr: Use on-device stats instead of private ones. These devices use the private area of appropriate size for statistics. Turning them to use on-device ones make them "privless" and thus - really small wrt kmalloc cache, they are allocated from. Besides, code looks nicer, because of absence of multi-braced type casts and dereferences. [ Fix build failures -DaveM ] Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/ipmr.c | 25 +++++++++---------------- 1 file changed, 9 insertions(+), 16 deletions(-) (limited to 'net') diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index 65f12005785e..a34da4977c73 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c @@ -181,26 +181,20 @@ static int reg_vif_num = -1; static int reg_vif_xmit(struct sk_buff *skb, struct net_device *dev) { read_lock(&mrt_lock); - ((struct net_device_stats*)netdev_priv(dev))->tx_bytes += skb->len; - ((struct net_device_stats*)netdev_priv(dev))->tx_packets++; + dev->stats.tx_bytes += skb->len; + dev->stats.tx_packets++; ipmr_cache_report(skb, reg_vif_num, IGMPMSG_WHOLEPKT); read_unlock(&mrt_lock); kfree_skb(skb); return 0; } -static struct net_device_stats *reg_vif_get_stats(struct net_device *dev) -{ - return (struct net_device_stats*)netdev_priv(dev); -} - static void reg_vif_setup(struct net_device *dev) { dev->type = ARPHRD_PIMREG; dev->mtu = ETH_DATA_LEN - sizeof(struct iphdr) - 8; dev->flags = IFF_NOARP; dev->hard_start_xmit = reg_vif_xmit; - dev->get_stats = reg_vif_get_stats; dev->destructor = free_netdev; } @@ -209,8 +203,7 @@ static struct net_device *ipmr_reg_vif(void) struct net_device *dev; struct in_device *in_dev; - dev = alloc_netdev(sizeof(struct net_device_stats), "pimreg", - reg_vif_setup); + dev = alloc_netdev(0, "pimreg", reg_vif_setup); if (dev == NULL) return NULL; @@ -1170,8 +1163,8 @@ static void ipmr_queue_xmit(struct sk_buff *skb, struct mfc_cache *c, int vifi) if (vif->flags & VIFF_REGISTER) { vif->pkt_out++; vif->bytes_out+=skb->len; - ((struct net_device_stats*)netdev_priv(vif->dev))->tx_bytes += skb->len; - ((struct net_device_stats*)netdev_priv(vif->dev))->tx_packets++; + vif->dev->stats.tx_bytes += skb->len; + vif->dev->stats.tx_packets++; ipmr_cache_report(skb, vifi, IGMPMSG_WHOLEPKT); kfree_skb(skb); return; @@ -1487,8 +1480,8 @@ int pim_rcv_v1(struct sk_buff * skb) skb->pkt_type = PACKET_HOST; dst_release(skb->dst); skb->dst = NULL; - ((struct net_device_stats*)netdev_priv(reg_dev))->rx_bytes += skb->len; - ((struct net_device_stats*)netdev_priv(reg_dev))->rx_packets++; + reg_dev->stats.rx_bytes += skb->len; + reg_dev->stats.rx_packets++; nf_reset(skb); netif_rx(skb); dev_put(reg_dev); @@ -1542,8 +1535,8 @@ static int pim_rcv(struct sk_buff * skb) skb->ip_summed = 0; skb->pkt_type = PACKET_HOST; dst_release(skb->dst); - ((struct net_device_stats*)netdev_priv(reg_dev))->rx_bytes += skb->len; - ((struct net_device_stats*)netdev_priv(reg_dev))->rx_packets++; + reg_dev->stats.rx_bytes += skb->len; + reg_dev->stats.rx_packets++; skb->dst = NULL; nf_reset(skb); netif_rx(skb); -- cgit v1.2.2 From dc58c78c047fb01f4c13e7de91abc5eb931920b3 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 21 May 2008 14:17:54 -0700 Subject: ip6mr: Use on-device stats instead of private ones. Similar to ipmr. [ Fix build failures -DaveM ] Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv6/ip6mr.c | 22 +++++++--------------- 1 file changed, 7 insertions(+), 15 deletions(-) (limited to 'net') diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c index 2de3c464fe75..bf268b386963 100644 --- a/net/ipv6/ip6mr.c +++ b/net/ipv6/ip6mr.c @@ -388,8 +388,8 @@ static int pim6_rcv(struct sk_buff *skb) skb->ip_summed = 0; skb->pkt_type = PACKET_HOST; dst_release(skb->dst); - ((struct net_device_stats *)netdev_priv(reg_dev))->rx_bytes += skb->len; - ((struct net_device_stats *)netdev_priv(reg_dev))->rx_packets++; + reg_dev->stats.rx_bytes += skb->len; + reg_dev->stats.rx_packets++; skb->dst = NULL; nf_reset(skb); netif_rx(skb); @@ -409,26 +409,20 @@ static struct inet6_protocol pim6_protocol = { static int reg_vif_xmit(struct sk_buff *skb, struct net_device *dev) { read_lock(&mrt_lock); - ((struct net_device_stats *)netdev_priv(dev))->tx_bytes += skb->len; - ((struct net_device_stats *)netdev_priv(dev))->tx_packets++; + dev->stats.tx_bytes += skb->len; + dev->stats.tx_packets++; ip6mr_cache_report(skb, reg_vif_num, MRT6MSG_WHOLEPKT); read_unlock(&mrt_lock); kfree_skb(skb); return 0; } -static struct net_device_stats *reg_vif_get_stats(struct net_device *dev) -{ - return (struct net_device_stats *)netdev_priv(dev); -} - static void reg_vif_setup(struct net_device *dev) { dev->type = ARPHRD_PIMREG; dev->mtu = 1500 - sizeof(struct ipv6hdr) - 8; dev->flags = IFF_NOARP; dev->hard_start_xmit = reg_vif_xmit; - dev->get_stats = reg_vif_get_stats; dev->destructor = free_netdev; } @@ -436,9 +430,7 @@ static struct net_device *ip6mr_reg_vif(void) { struct net_device *dev; - dev = alloc_netdev(sizeof(struct net_device_stats), "pim6reg", - reg_vif_setup); - + dev = alloc_netdev(0, "pim6reg", reg_vif_setup); if (dev == NULL) return NULL; @@ -1377,8 +1369,8 @@ static int ip6mr_forward2(struct sk_buff *skb, struct mfc6_cache *c, int vifi) if (vif->flags & MIFF_REGISTER) { vif->pkt_out++; vif->bytes_out += skb->len; - ((struct net_device_stats *)netdev_priv(vif->dev))->tx_bytes += skb->len; - ((struct net_device_stats *)netdev_priv(vif->dev))->tx_packets++; + vif->dev->stats.tx_bytes += skb->len; + vif->dev->stats.tx_packets++; ip6mr_cache_report(skb, vifi, MRT6MSG_WHOLEPKT); kfree_skb(skb); return 0; -- cgit v1.2.2 From 59f0c4523fdea865fab7d69d878269992a9d08dd Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 21 May 2008 14:52:30 -0700 Subject: tipc: Fix skb_under_panic when configuring TIPC without privileges This patch prevents a TIPC configuration command requiring network administrator privileges from triggering an skbuff underrun if it is issued by a process lacking those privileges. The revised error handling code avoids the use of a potentially uninitialized global variable by transforming the unauthorized command into a new command, then following the standard command processing path to generate the required error message. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/config.c | 6 +++++- net/tipc/netlink.c | 16 +++++++++------- 2 files changed, 14 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/tipc/config.c b/net/tipc/config.c index 91d56f8fee9f..16e7cb74969b 100644 --- a/net/tipc/config.c +++ b/net/tipc/config.c @@ -2,7 +2,7 @@ * net/tipc/config.c: TIPC configuration management code * * Copyright (c) 2002-2006, Ericsson AB - * Copyright (c) 2004-2006, Wind River Systems + * Copyright (c) 2004-2007, Wind River Systems * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -602,6 +602,10 @@ struct sk_buff *tipc_cfg_do_cmd(u32 orig_node, u16 cmd, const void *request_area case TIPC_CMD_GET_NETID: rep_tlv_buf = tipc_cfg_reply_unsigned(tipc_net_id); break; + case TIPC_CMD_NOT_NET_ADMIN: + rep_tlv_buf = + tipc_cfg_reply_error_string(TIPC_CFG_NOT_NET_ADMIN); + break; default: rep_tlv_buf = tipc_cfg_reply_error_string(TIPC_CFG_NOT_SUPPORTED " (unknown command)"); diff --git a/net/tipc/netlink.c b/net/tipc/netlink.c index 6a7f7b4c2595..c387217bb230 100644 --- a/net/tipc/netlink.c +++ b/net/tipc/netlink.c @@ -2,7 +2,7 @@ * net/tipc/netlink.c: TIPC configuration handling * * Copyright (c) 2005-2006, Ericsson AB - * Copyright (c) 2005, Wind River Systems + * Copyright (c) 2005-2007, Wind River Systems * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -45,15 +45,17 @@ static int handle_cmd(struct sk_buff *skb, struct genl_info *info) struct nlmsghdr *req_nlh = info->nlhdr; struct tipc_genlmsghdr *req_userhdr = info->userhdr; int hdr_space = NLMSG_SPACE(GENL_HDRLEN + TIPC_GENL_HDRLEN); + u16 cmd; if ((req_userhdr->cmd & 0xC000) && (!capable(CAP_NET_ADMIN))) - rep_buf = tipc_cfg_reply_error_string(TIPC_CFG_NOT_NET_ADMIN); + cmd = TIPC_CMD_NOT_NET_ADMIN; else - rep_buf = tipc_cfg_do_cmd(req_userhdr->dest, - req_userhdr->cmd, - NLMSG_DATA(req_nlh) + GENL_HDRLEN + TIPC_GENL_HDRLEN, - NLMSG_PAYLOAD(req_nlh, GENL_HDRLEN + TIPC_GENL_HDRLEN), - hdr_space); + cmd = req_userhdr->cmd; + + rep_buf = tipc_cfg_do_cmd(req_userhdr->dest, cmd, + NLMSG_DATA(req_nlh) + GENL_HDRLEN + TIPC_GENL_HDRLEN, + NLMSG_PAYLOAD(req_nlh, GENL_HDRLEN + TIPC_GENL_HDRLEN), + hdr_space); if (rep_buf) { skb_push(rep_buf, hdr_space); -- cgit v1.2.2 From 2ecb0924d7791372a70ef8f1174e37b329b955c3 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 21 May 2008 14:53:00 -0700 Subject: tipc: Prevent node object duplication due to simultaneous discovery This patch ensures that the simultaneous discovery of the same neighboring node by multiple interfaces does not cause TIPC to add the node into its internal data structures more than once. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/node.c | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+) (limited to 'net') diff --git a/net/tipc/node.c b/net/tipc/node.c index 598f4d3a0098..34e9a2bb7c19 100644 --- a/net/tipc/node.c +++ b/net/tipc/node.c @@ -52,16 +52,40 @@ static void node_established_contact(struct node *n_ptr); struct node *tipc_nodes = NULL; /* sorted list of nodes within cluster */ +static DEFINE_SPINLOCK(node_create_lock); + u32 tipc_own_tag = 0; +/** + * tipc_node_create - create neighboring node + * + * Currently, this routine is called by neighbor discovery code, which holds + * net_lock for reading only. We must take node_create_lock to ensure a node + * isn't created twice if two different bearers discover the node at the same + * time. (It would be preferable to switch to holding net_lock in write mode, + * but this is a non-trivial change.) + */ + struct node *tipc_node_create(u32 addr) { struct cluster *c_ptr; struct node *n_ptr; struct node **curr_node; + spin_lock_bh(&node_create_lock); + + for (n_ptr = tipc_nodes; n_ptr; n_ptr = n_ptr->next) { + if (addr < n_ptr->addr) + break; + if (addr == n_ptr->addr) { + spin_unlock_bh(&node_create_lock); + return n_ptr; + } + } + n_ptr = kzalloc(sizeof(*n_ptr),GFP_ATOMIC); if (!n_ptr) { + spin_unlock_bh(&node_create_lock); warn("Node creation failed, no memory\n"); return NULL; } @@ -71,6 +95,7 @@ struct node *tipc_node_create(u32 addr) c_ptr = tipc_cltr_create(addr); } if (!c_ptr) { + spin_unlock_bh(&node_create_lock); kfree(n_ptr); return NULL; } @@ -91,6 +116,7 @@ struct node *tipc_node_create(u32 addr) } } (*curr_node) = n_ptr; + spin_unlock_bh(&node_create_lock); return n_ptr; } -- cgit v1.2.2 From 811102ca23dfdde5ee8b782b3a4bbff44c499cb2 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 21 May 2008 14:53:34 -0700 Subject: tipc: Optimize null pointer check during neighbor discovery This patch optimizes TIPC neighbor discovery code to avoid testing for a null node pointer when the pointer is already known to be non-null. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/discover.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/tipc/discover.c b/net/tipc/discover.c index 5d643e5721eb..faeaf06d377e 100644 --- a/net/tipc/discover.c +++ b/net/tipc/discover.c @@ -200,9 +200,8 @@ void tipc_disc_recv_msg(struct sk_buff *buf) dbg(" in own cluster\n"); if (n_ptr == NULL) { n_ptr = tipc_node_create(orig); - } - if (n_ptr == NULL) { - return; + if (!n_ptr) + return; } spin_lock_bh(&n_ptr->lock); link = n_ptr->links[b_ptr->identity]; -- cgit v1.2.2 From 6d4a6672c8263f98544d2b91690dc7074b144090 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 21 May 2008 14:54:12 -0700 Subject: tipc: Update "previous node" indicators when node address changes This patch ensures that the "previous node" field in any existing TIPC port message header templates is updated properly when a TIPC network address is assigned to the node. (Previously, only the "originating node" field was updated.) Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/port.c | 1 + 1 file changed, 1 insertion(+) (limited to 'net') diff --git a/net/tipc/port.c b/net/tipc/port.c index 757de38fe6af..2c64ad88e3cf 100644 --- a/net/tipc/port.c +++ b/net/tipc/port.c @@ -780,6 +780,7 @@ void tipc_port_reinit(void) msg = &p_ptr->publ.phdr; if (msg_orignode(msg) == tipc_own_addr) break; + msg_set_prevnode(msg, tipc_own_addr); msg_set_orignode(msg, tipc_own_addr); } spin_unlock_bh(&tipc_port_list_lock); -- cgit v1.2.2 From 03194379a77b02df3404ec4848a50c6784e9a8a5 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 21 May 2008 14:55:04 -0700 Subject: tipc: Fix initialization sequence problems when entering network mode This patch ensures that TIPC's topology service and configuration service are shut down before switching into "network mode". This ensures that TIPC does not mistakenly try to send unnecessary "publication withdraw" messages to other nodes before it is fully initialized for sending off-node messages. Note that the node's current network address is now updated only after the two services are shut down; this ensures that any existing connections to the topology server are terminated correctly using the old address. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/config.c | 3 +-- net/tipc/core.c | 4 ++-- net/tipc/core.h | 2 +- net/tipc/net.c | 10 +++++++--- net/tipc/net.h | 2 +- 5 files changed, 12 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/tipc/config.c b/net/tipc/config.c index 16e7cb74969b..ca3544d030c7 100644 --- a/net/tipc/config.c +++ b/net/tipc/config.c @@ -293,7 +293,6 @@ static struct sk_buff *cfg_set_own_addr(void) if (tipc_mode == TIPC_NET_MODE) return tipc_cfg_reply_error_string(TIPC_CFG_NOT_SUPPORTED " (cannot change node address once assigned)"); - tipc_own_addr = addr; /* * Must release all spinlocks before calling start_net() because @@ -306,7 +305,7 @@ static struct sk_buff *cfg_set_own_addr(void) */ spin_unlock_bh(&config_lock); - tipc_core_start_net(); + tipc_core_start_net(addr); spin_lock_bh(&config_lock); return tipc_cfg_reply_none(); } diff --git a/net/tipc/core.c b/net/tipc/core.c index 3d97386af095..3256bd7d398f 100644 --- a/net/tipc/core.c +++ b/net/tipc/core.c @@ -117,11 +117,11 @@ void tipc_core_stop_net(void) * start_net - start TIPC networking sub-systems */ -int tipc_core_start_net(void) +int tipc_core_start_net(unsigned long addr) { int res; - if ((res = tipc_net_start()) || + if ((res = tipc_net_start(addr)) || (res = tipc_eth_media_start())) { tipc_core_stop_net(); } diff --git a/net/tipc/core.h b/net/tipc/core.h index bd78d1705c0f..a881f92a8537 100644 --- a/net/tipc/core.h +++ b/net/tipc/core.h @@ -202,7 +202,7 @@ extern atomic_t tipc_user_count; extern int tipc_core_start(void); extern void tipc_core_stop(void); -extern int tipc_core_start_net(void); +extern int tipc_core_start_net(unsigned long addr); extern void tipc_core_stop_net(void); extern int tipc_handler_start(void); extern void tipc_handler_stop(void); diff --git a/net/tipc/net.c b/net/tipc/net.c index c39c76201e8e..cc51fa483672 100644 --- a/net/tipc/net.c +++ b/net/tipc/net.c @@ -266,7 +266,7 @@ void tipc_net_route_msg(struct sk_buff *buf) tipc_link_send(buf, dnode, msg_link_selector(msg)); } -int tipc_net_start(void) +int tipc_net_start(u32 addr) { char addr_string[16]; int res; @@ -274,6 +274,10 @@ int tipc_net_start(void) if (tipc_mode != TIPC_NODE_MODE) return -ENOPROTOOPT; + tipc_subscr_stop(); + tipc_cfg_stop(); + + tipc_own_addr = addr; tipc_mode = TIPC_NET_MODE; tipc_named_reinit(); tipc_port_reinit(); @@ -284,10 +288,10 @@ int tipc_net_start(void) (res = tipc_bclink_init())) { return res; } - tipc_subscr_stop(); - tipc_cfg_stop(); + tipc_k_signal((Handler)tipc_subscr_start, 0); tipc_k_signal((Handler)tipc_cfg_init, 0); + info("Started in network mode\n"); info("Own node address %s, network identity %u\n", addr_string_fill(addr_string, tipc_own_addr), tipc_net_id); diff --git a/net/tipc/net.h b/net/tipc/net.h index a6a0e9976ac9..d154ac2bda9a 100644 --- a/net/tipc/net.h +++ b/net/tipc/net.h @@ -58,7 +58,7 @@ void tipc_net_route_msg(struct sk_buff *buf); struct node *tipc_net_select_remote_node(u32 addr, u32 ref); u32 tipc_net_select_router(u32 addr, u32 ref); -int tipc_net_start(void); +int tipc_net_start(u32 addr); void tipc_net_stop(void); #endif -- cgit v1.2.2 From 402d7752ed253369b7ab037e2d778e52d59c19ed Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Tue, 6 May 2008 18:53:43 +0400 Subject: mac80211: Brush up error paths in mesh_path_add. There are already tree paths, that do incremental rollbacks, so merge them together, rename labels and format the code to look a bit nicer. (I do not mind dropping/delaying this patch however). Signed-off-by: Pavel Emelyanov Signed-off-by: John W. Linville --- net/mac80211/mesh_pathtbl.c | 42 ++++++++++++++++++++---------------------- 1 file changed, 20 insertions(+), 22 deletions(-) (limited to 'net') diff --git a/net/mac80211/mesh_pathtbl.c b/net/mac80211/mesh_pathtbl.c index 99c2d360888e..7097ef981991 100644 --- a/net/mac80211/mesh_pathtbl.c +++ b/net/mac80211/mesh_pathtbl.c @@ -158,19 +158,14 @@ int mesh_path_add(u8 *dst, struct net_device *dev) if (atomic_add_unless(&sdata->u.sta.mpaths, 1, MESH_MAX_MPATHS) == 0) return -ENOSPC; + err = -ENOMEM; new_mpath = kzalloc(sizeof(struct mesh_path), GFP_KERNEL); - if (!new_mpath) { - atomic_dec(&sdata->u.sta.mpaths); - err = -ENOMEM; - goto endadd2; - } + if (!new_mpath) + goto err_path_alloc; + new_node = kmalloc(sizeof(struct mpath_node), GFP_KERNEL); - if (!new_node) { - kfree(new_mpath); - atomic_dec(&sdata->u.sta.mpaths); - err = -ENOMEM; - goto endadd2; - } + if (!new_node) + goto err_node_alloc; read_lock(&pathtbl_resize_lock); memcpy(new_mpath->dst, dst, ETH_ALEN); @@ -189,16 +184,11 @@ int mesh_path_add(u8 *dst, struct net_device *dev) spin_lock(&mesh_paths->hashwlock[hash_idx]); + err = -EEXIST; hlist_for_each_entry(node, n, bucket, list) { mpath = node->mpath; - if (mpath->dev == dev && memcmp(dst, mpath->dst, ETH_ALEN) - == 0) { - err = -EEXIST; - atomic_dec(&sdata->u.sta.mpaths); - kfree(new_node); - kfree(new_mpath); - goto endadd; - } + if (mpath->dev == dev && memcmp(dst, mpath->dst, ETH_ALEN) == 0) + goto err_exists; } hlist_add_head_rcu(&new_node->list, bucket); @@ -206,10 +196,9 @@ int mesh_path_add(u8 *dst, struct net_device *dev) mesh_paths->mean_chain_len * (mesh_paths->hash_mask + 1)) grow = 1; -endadd: spin_unlock(&mesh_paths->hashwlock[hash_idx]); read_unlock(&pathtbl_resize_lock); - if (!err && grow) { + if (grow) { struct mesh_table *oldtbl, *newtbl; write_lock(&pathtbl_resize_lock); @@ -225,7 +214,16 @@ endadd: synchronize_rcu(); mesh_table_free(oldtbl, false); } -endadd2: + return 0; + +err_exists: + spin_unlock(&mesh_paths->hashwlock[hash_idx]); + read_unlock(&pathtbl_resize_lock); + kfree(new_node); +err_node_alloc: + kfree(new_mpath); +err_path_alloc: + atomic_dec(&sdata->u.sta.mpaths); return err; } -- cgit v1.2.2 From 5194ee82b4aafc35b32c96db11bdacea9bf548be Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 7 May 2008 19:44:20 +0400 Subject: mac80211: Fix one more call to synchronize_rcu in atomic context. (This set applies OK without the previous one of 4 patches, but with some fuzz in the 7th one) The mesh_path_node_free() does so under hashwlock. But, this one is called 1. from mesh_path_add() after an old hash is hidden and synchronize_rcu() is calld 2. mesh_pathtbl_unregister(), when the module is being unloaded and no devices exist to mess with this hash. So, it seems to me, that simply removing the call is OK. Signed-off-by: Pavel Emelyanov Signed-off-by: John W. Linville --- net/mac80211/mesh_pathtbl.c | 1 - 1 file changed, 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/mesh_pathtbl.c b/net/mac80211/mesh_pathtbl.c index 7097ef981991..0b6c4bfe3e78 100644 --- a/net/mac80211/mesh_pathtbl.c +++ b/net/mac80211/mesh_pathtbl.c @@ -458,7 +458,6 @@ static void mesh_path_node_free(struct hlist_node *p, bool free_leafs) struct mpath_node *node = hlist_entry(p, struct mpath_node, list); mpath = node->mpath; hlist_del_rcu(p); - synchronize_rcu(); if (free_leafs) kfree(mpath); kfree(node); -- cgit v1.2.2 From 4caf86c6928cfaca270327bc944f901c2e2a8f50 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 7 May 2008 19:47:01 +0400 Subject: mac80211: Prepare mesh_table_grow to failing copy_node callback. The mesh_path_node_copy() performs kmalloc() and thus - may fail (well, it does not now, but I'm fixing this right now). Its caller - the mesh_table_grow() - isn't prepared for such a trick yet. This preparation is just flush the new hash and make copy_node() return an int value. Signed-off-by: Pavel Emelyanov Signed-off-by: John W. Linville --- net/mac80211/mesh.c | 15 +++++++++++++-- net/mac80211/mesh.h | 2 +- net/mac80211/mesh_pathtbl.c | 3 ++- 3 files changed, 16 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/mac80211/mesh.c b/net/mac80211/mesh.c index 697ef67f96b6..ca81d0065eb8 100644 --- a/net/mac80211/mesh.c +++ b/net/mac80211/mesh.c @@ -349,7 +349,7 @@ struct mesh_table *mesh_table_grow(struct mesh_table *tbl) { struct mesh_table *newtbl; struct hlist_head *oldhash; - struct hlist_node *p; + struct hlist_node *p, *q; int err = 0; int i; @@ -373,13 +373,24 @@ struct mesh_table *mesh_table_grow(struct mesh_table *tbl) oldhash = tbl->hash_buckets; for (i = 0; i <= tbl->hash_mask; i++) hlist_for_each(p, &oldhash[i]) - tbl->copy_node(p, newtbl); + if (tbl->copy_node(p, newtbl) < 0) + goto errcopy; endgrow: if (err) return NULL; else return newtbl; + +errcopy: + for (i = 0; i <= newtbl->hash_mask; i++) { + hlist_for_each_safe(p, q, &newtbl->hash_buckets[i]) + tbl->free_node(p, 0); + } + kfree(newtbl->hash_buckets); + kfree(newtbl->hashwlock); + kfree(newtbl); + return NULL; } /** diff --git a/net/mac80211/mesh.h b/net/mac80211/mesh.h index 2e161f6d8288..669eafafe497 100644 --- a/net/mac80211/mesh.h +++ b/net/mac80211/mesh.h @@ -109,7 +109,7 @@ struct mesh_table { __u32 hash_rnd; /* Used for hash generation */ atomic_t entries; /* Up to MAX_MESH_NEIGHBOURS */ void (*free_node) (struct hlist_node *p, bool free_leafs); - void (*copy_node) (struct hlist_node *p, struct mesh_table *newtbl); + int (*copy_node) (struct hlist_node *p, struct mesh_table *newtbl); int size_order; int mean_chain_len; }; diff --git a/net/mac80211/mesh_pathtbl.c b/net/mac80211/mesh_pathtbl.c index 0b6c4bfe3e78..512bfa112c6a 100644 --- a/net/mac80211/mesh_pathtbl.c +++ b/net/mac80211/mesh_pathtbl.c @@ -463,7 +463,7 @@ static void mesh_path_node_free(struct hlist_node *p, bool free_leafs) kfree(node); } -static void mesh_path_node_copy(struct hlist_node *p, struct mesh_table *newtbl) +static int mesh_path_node_copy(struct hlist_node *p, struct mesh_table *newtbl) { struct mesh_path *mpath; struct mpath_node *node, *new_node; @@ -476,6 +476,7 @@ static void mesh_path_node_copy(struct hlist_node *p, struct mesh_table *newtbl) hash_idx = mesh_table_hash(mpath->dst, mpath->dev, newtbl); hlist_add_head(&new_node->list, &newtbl->hash_buckets[hash_idx]); + return 0; } int mesh_pathtbl_init(void) -- cgit v1.2.2 From 00242c40a167113688dd7ed46ec94e8a4c47f603 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 7 May 2008 19:48:14 +0400 Subject: mac80211: Report allocation failure from mesh_path_node_copy. Now - return the -ENOMEM in case kmalloc fails. Signed-off-by: Pavel Emelyanov Signed-off-by: John W. Linville --- net/mac80211/mesh_pathtbl.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/mesh_pathtbl.c b/net/mac80211/mesh_pathtbl.c index 512bfa112c6a..fe1ceeea07f1 100644 --- a/net/mac80211/mesh_pathtbl.c +++ b/net/mac80211/mesh_pathtbl.c @@ -469,9 +469,12 @@ static int mesh_path_node_copy(struct hlist_node *p, struct mesh_table *newtbl) struct mpath_node *node, *new_node; u32 hash_idx; + new_node = kmalloc(sizeof(struct mpath_node), GFP_KERNEL); + if (new_node == NULL) + return -ENOMEM; + node = hlist_entry(p, struct mpath_node, list); mpath = node->mpath; - new_node = kmalloc(sizeof(struct mpath_node), GFP_KERNEL); new_node->mpath = mpath; hash_idx = mesh_table_hash(mpath->dst, mpath->dev, newtbl); hlist_add_head(&new_node->list, -- cgit v1.2.2 From 8566dc3fca470454461b161677a5ae3bb3a8c1b8 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 7 May 2008 19:51:51 +0400 Subject: mac80211: Fix sleeping allocation under lock in mesh_path_node_copy. The mesh_path_node_copy() can be called like this: mesh_path_add `- write_lock(&pathtbl_resize_lock); /* ! */ `- mesh_table_grow `- ->copy_node `- mesh_path_node_copy thus, the GFP_KERNEL is not suitable here. The acceptable fix, I suppose, is make this allocation GPF_ATOMIC - the mpath_node being allocated is 4 pointers, i.e. this allocation is small enough to survive even under a moderate memory pressure. Signed-off-by: Pavel Emelyanov Signed-off-by: John W. Linville --- net/mac80211/mesh_pathtbl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/mesh_pathtbl.c b/net/mac80211/mesh_pathtbl.c index fe1ceeea07f1..1154398eceaa 100644 --- a/net/mac80211/mesh_pathtbl.c +++ b/net/mac80211/mesh_pathtbl.c @@ -469,7 +469,7 @@ static int mesh_path_node_copy(struct hlist_node *p, struct mesh_table *newtbl) struct mpath_node *node, *new_node; u32 hash_idx; - new_node = kmalloc(sizeof(struct mpath_node), GFP_KERNEL); + new_node = kmalloc(sizeof(struct mpath_node), GFP_ATOMIC); if (new_node == NULL) return -ENOMEM; -- cgit v1.2.2 From 3282aea9ea5644a5b0161ad0fbd70fbf1099a470 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 7 May 2008 19:54:04 +0400 Subject: mac80211: Do not report false error from mesh_path_add. In case the hash grow failed, it is not fair to return error - the new node _was_ _actually_ added in this case. Besides, after my previous patch, this grow is more likely to fail on large hashes. Signed-off-by: Pavel Emelyanov Signed-off-by: John W. Linville --- net/mac80211/mesh_pathtbl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/mesh_pathtbl.c b/net/mac80211/mesh_pathtbl.c index 1154398eceaa..947b13b40726 100644 --- a/net/mac80211/mesh_pathtbl.c +++ b/net/mac80211/mesh_pathtbl.c @@ -206,7 +206,7 @@ int mesh_path_add(u8 *dst, struct net_device *dev) newtbl = mesh_table_grow(mesh_paths); if (!newtbl) { write_unlock(&pathtbl_resize_lock); - return -ENOMEM; + return 0; } rcu_assign_pointer(mesh_paths, newtbl); write_unlock(&pathtbl_resize_lock); -- cgit v1.2.2 From a3538b19a6d226f9d3d9b18865468370009dec55 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 7 May 2008 19:55:59 +0400 Subject: mac80211: Merge error paths in mesh_table_grow(). This is the first (of two) clean ups after the fixes above. The err variable is not even required after this cleaning. Signed-off-by: Pavel Emelyanov Signed-off-by: John W. Linville --- net/mac80211/mesh.c | 16 ++++------------ 1 file changed, 4 insertions(+), 12 deletions(-) (limited to 'net') diff --git a/net/mac80211/mesh.c b/net/mac80211/mesh.c index ca81d0065eb8..cbce001f8f23 100644 --- a/net/mac80211/mesh.c +++ b/net/mac80211/mesh.c @@ -350,20 +350,15 @@ struct mesh_table *mesh_table_grow(struct mesh_table *tbl) struct mesh_table *newtbl; struct hlist_head *oldhash; struct hlist_node *p, *q; - int err = 0; int i; if (atomic_read(&tbl->entries) - < tbl->mean_chain_len * (tbl->hash_mask + 1)) { - err = -EPERM; + < tbl->mean_chain_len * (tbl->hash_mask + 1)) goto endgrow; - } newtbl = mesh_table_alloc(tbl->size_order + 1); - if (!newtbl) { - err = -ENOMEM; + if (!newtbl) goto endgrow; - } newtbl->free_node = tbl->free_node; newtbl->mean_chain_len = tbl->mean_chain_len; @@ -376,11 +371,7 @@ struct mesh_table *mesh_table_grow(struct mesh_table *tbl) if (tbl->copy_node(p, newtbl) < 0) goto errcopy; -endgrow: - if (err) - return NULL; - else - return newtbl; + return newtbl; errcopy: for (i = 0; i <= newtbl->hash_mask; i++) { @@ -390,6 +381,7 @@ errcopy: kfree(newtbl->hash_buckets); kfree(newtbl->hashwlock); kfree(newtbl); +endgrow: return NULL; } -- cgit v1.2.2 From bd9b448f4c0a514559bdae4ca18ca3e8cd999c6d Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 7 May 2008 19:57:11 +0400 Subject: mac80211: Consolidate hash kfree-ing in mesh.c. There are already two places, that kfree the mesh_table and its buckets. Signed-off-by: Pavel Emelyanov Signed-off-by: John W. Linville --- net/mac80211/mesh.c | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/mac80211/mesh.c b/net/mac80211/mesh.c index cbce001f8f23..b5933b271491 100644 --- a/net/mac80211/mesh.c +++ b/net/mac80211/mesh.c @@ -315,6 +315,13 @@ struct mesh_table *mesh_table_alloc(int size_order) return newtbl; } +static void __mesh_table_free(struct mesh_table *tbl) +{ + kfree(tbl->hash_buckets); + kfree(tbl->hashwlock); + kfree(tbl); +} + void mesh_table_free(struct mesh_table *tbl, bool free_leafs) { struct hlist_head *mesh_hash; @@ -330,9 +337,7 @@ void mesh_table_free(struct mesh_table *tbl, bool free_leafs) } spin_unlock(&tbl->hashwlock[i]); } - kfree(tbl->hash_buckets); - kfree(tbl->hashwlock); - kfree(tbl); + __mesh_table_free(tbl); } static void ieee80211_mesh_path_timer(unsigned long data) @@ -378,9 +383,7 @@ errcopy: hlist_for_each_safe(p, q, &newtbl->hash_buckets[i]) tbl->free_node(p, 0); } - kfree(newtbl->hash_buckets); - kfree(newtbl->hashwlock); - kfree(newtbl); + __mesh_table_free(tbl); endgrow: return NULL; } -- cgit v1.2.2 From 1bd3dff549537d3d9b92d0b284f4cc2be264a56d Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 14 May 2008 16:26:16 -0700 Subject: mac80211: michael.c use kernel-provided infrastructure Replace private implementation of bit rotation and unaligned access helpers with kernel-provided implementation. Fold xswap helper in its one usage in the michael_block macro. Signed-off-by: Harvey Harrison Cc: "John W. Linville" Signed-off-by: Andrew Morton Signed-off-by: John W. Linville --- net/mac80211/michael.c | 60 ++++++++++++-------------------------------------- 1 file changed, 14 insertions(+), 46 deletions(-) (limited to 'net') diff --git a/net/mac80211/michael.c b/net/mac80211/michael.c index 0f844f7895f1..c1e5897fcf7c 100644 --- a/net/mac80211/michael.c +++ b/net/mac80211/michael.c @@ -8,71 +8,39 @@ */ #include +#include +#include #include "michael.h" -static inline u32 rotr(u32 val, int bits) -{ - return (val >> bits) | (val << (32 - bits)); -} - - -static inline u32 rotl(u32 val, int bits) -{ - return (val << bits) | (val >> (32 - bits)); -} - - -static inline u32 xswap(u32 val) -{ - return ((val & 0xff00ff00) >> 8) | ((val & 0x00ff00ff) << 8); -} - - #define michael_block(l, r) \ do { \ - r ^= rotl(l, 17); \ + r ^= rol32(l, 17); \ l += r; \ - r ^= xswap(l); \ + r ^= ((l & 0xff00ff00) >> 8) | ((l & 0x00ff00ff) << 8); \ l += r; \ - r ^= rotl(l, 3); \ + r ^= rol32(l, 3); \ l += r; \ - r ^= rotr(l, 2); \ + r ^= ror32(l, 2); \ l += r; \ } while (0) - -static inline u32 michael_get32(u8 *data) -{ - return data[0] | (data[1] << 8) | (data[2] << 16) | (data[3] << 24); -} - - -static inline void michael_put32(u32 val, u8 *data) -{ - data[0] = val & 0xff; - data[1] = (val >> 8) & 0xff; - data[2] = (val >> 16) & 0xff; - data[3] = (val >> 24) & 0xff; -} - - void michael_mic(u8 *key, u8 *da, u8 *sa, u8 priority, u8 *data, size_t data_len, u8 *mic) { u32 l, r, val; size_t block, blocks, left; - l = michael_get32(key); - r = michael_get32(key + 4); + l = get_unaligned_le32(key); + r = get_unaligned_le32(key + 4); /* A pseudo header (DA, SA, Priority, 0, 0, 0) is used in Michael MIC * calculation, but it is _not_ transmitted */ - l ^= michael_get32(da); + l ^= get_unaligned_le32(da); michael_block(l, r); - l ^= da[4] | (da[5] << 8) | (sa[0] << 16) | (sa[1] << 24); + l ^= get_unaligned_le16(&da[4]) | (get_unaligned_le16(sa) << 16); michael_block(l, r); - l ^= michael_get32(&sa[2]); + l ^= get_unaligned_le32(&sa[2]); michael_block(l, r); l ^= priority; michael_block(l, r); @@ -82,7 +50,7 @@ void michael_mic(u8 *key, u8 *da, u8 *sa, u8 priority, left = data_len % 4; for (block = 0; block < blocks; block++) { - l ^= michael_get32(&data[block * 4]); + l ^= get_unaligned_le32(&data[block * 4]); michael_block(l, r); } @@ -99,6 +67,6 @@ void michael_mic(u8 *key, u8 *da, u8 *sa, u8 priority, /* last block is zero, so l ^ 0 = l */ michael_block(l, r); - michael_put32(l, mic); - michael_put32(r, mic + 4); + put_unaligned_le32(l, mic); + put_unaligned_le32(r, mic + 4); } -- cgit v1.2.2 From 1b19ca396621dcba573b20d4625526f5a460c886 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 14 May 2008 16:26:17 -0700 Subject: mac80211: introduce struct michael_mic_ctx and static helpers Replace the existing macro with a static function, significantly shrinks the size of the produced object file. Signed-off-by: Harvey Harrison Cc: "John W. Linville" Signed-off-by: Andrew Morton Signed-off-by: John W. Linville --- net/mac80211/michael.c | 75 +++++++++++++++++++++++++++----------------------- net/mac80211/michael.h | 4 +++ 2 files changed, 44 insertions(+), 35 deletions(-) (limited to 'net') diff --git a/net/mac80211/michael.c b/net/mac80211/michael.c index c1e5897fcf7c..4e151e0caeb7 100644 --- a/net/mac80211/michael.c +++ b/net/mac80211/michael.c @@ -13,46 +13,52 @@ #include "michael.h" -#define michael_block(l, r) \ -do { \ - r ^= rol32(l, 17); \ - l += r; \ - r ^= ((l & 0xff00ff00) >> 8) | ((l & 0x00ff00ff) << 8); \ - l += r; \ - r ^= rol32(l, 3); \ - l += r; \ - r ^= ror32(l, 2); \ - l += r; \ -} while (0) +static void michael_block(struct michael_mic_ctx *mctx, u32 val) +{ + mctx->l ^= val; + mctx->r ^= rol32(mctx->l, 17); + mctx->l += mctx->r; + mctx->r ^= ((mctx->l & 0xff00ff00) >> 8) | + ((mctx->l & 0x00ff00ff) << 8); + mctx->l += mctx->r; + mctx->r ^= rol32(mctx->l, 3); + mctx->l += mctx->r; + mctx->r ^= ror32(mctx->l, 2); + mctx->l += mctx->r; +} + +static void michael_mic_hdr(struct michael_mic_ctx *mctx, + u8 *key, u8 *da, u8 *sa, u8 priority) +{ + mctx->l = get_unaligned_le32(key); + mctx->r = get_unaligned_le32(key + 4); + + /* + * A pseudo header (DA, SA, Priority, 0, 0, 0) is used in Michael MIC + * calculation, but it is _not_ transmitted + */ + michael_block(mctx, get_unaligned_le32(da)); + michael_block(mctx, get_unaligned_le16(&da[4]) | + (get_unaligned_le16(sa) << 16)); + michael_block(mctx, get_unaligned_le32(&sa[2])); + michael_block(mctx, priority); +} void michael_mic(u8 *key, u8 *da, u8 *sa, u8 priority, u8 *data, size_t data_len, u8 *mic) { - u32 l, r, val; + u32 val; size_t block, blocks, left; + struct michael_mic_ctx mctx; - l = get_unaligned_le32(key); - r = get_unaligned_le32(key + 4); - - /* A pseudo header (DA, SA, Priority, 0, 0, 0) is used in Michael MIC - * calculation, but it is _not_ transmitted */ - l ^= get_unaligned_le32(da); - michael_block(l, r); - l ^= get_unaligned_le16(&da[4]) | (get_unaligned_le16(sa) << 16); - michael_block(l, r); - l ^= get_unaligned_le32(&sa[2]); - michael_block(l, r); - l ^= priority; - michael_block(l, r); + michael_mic_hdr(&mctx, key, da, sa, priority); /* Real data */ blocks = data_len / 4; left = data_len % 4; - for (block = 0; block < blocks; block++) { - l ^= get_unaligned_le32(&data[block * 4]); - michael_block(l, r); - } + for (block = 0; block < blocks; block++) + michael_block(&mctx, get_unaligned_le32(&data[block * 4])); /* Partial block of 0..3 bytes and padding: 0x5a + 4..7 zeros to make * total length a multiple of 4. */ @@ -62,11 +68,10 @@ void michael_mic(u8 *key, u8 *da, u8 *sa, u8 priority, left--; val |= data[blocks * 4 + left]; } - l ^= val; - michael_block(l, r); - /* last block is zero, so l ^ 0 = l */ - michael_block(l, r); - put_unaligned_le32(l, mic); - put_unaligned_le32(r, mic + 4); + michael_block(&mctx, val); + michael_block(&mctx, 0); + + put_unaligned_le32(mctx.l, mic); + put_unaligned_le32(mctx.r, mic + 4); } diff --git a/net/mac80211/michael.h b/net/mac80211/michael.h index 2e6aebabeea1..7d5707bf8843 100644 --- a/net/mac80211/michael.h +++ b/net/mac80211/michael.h @@ -14,6 +14,10 @@ #define MICHAEL_MIC_LEN 8 +struct michael_mic_ctx { + u32 l, r; +}; + void michael_mic(u8 *key, u8 *da, u8 *sa, u8 priority, u8 *data, size_t data_len, u8 *mic); -- cgit v1.2.2 From a7b6f0c5558ad03281b8064d6a4ab2e124dea991 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 14 May 2008 16:26:18 -0700 Subject: mac80211: add const, remove unused function, make one function static Signed-off-by: Harvey Harrison Cc: "John W. Linville" Signed-off-by: Andrew Morton Signed-off-by: John W. Linville --- net/mac80211/michael.c | 7 +++---- net/mac80211/michael.h | 4 ++-- net/mac80211/tkip.c | 12 ++---------- net/mac80211/tkip.h | 4 ---- 4 files changed, 7 insertions(+), 20 deletions(-) (limited to 'net') diff --git a/net/mac80211/michael.c b/net/mac80211/michael.c index 4e151e0caeb7..1fcdf38cf60c 100644 --- a/net/mac80211/michael.c +++ b/net/mac80211/michael.c @@ -6,7 +6,6 @@ * it under the terms of the GNU General Public License version 2 as * published by the Free Software Foundation. */ - #include #include #include @@ -28,7 +27,7 @@ static void michael_block(struct michael_mic_ctx *mctx, u32 val) } static void michael_mic_hdr(struct michael_mic_ctx *mctx, - u8 *key, u8 *da, u8 *sa, u8 priority) + const u8 *key, const u8 *da, const u8 *sa, u8 priority) { mctx->l = get_unaligned_le32(key); mctx->r = get_unaligned_le32(key + 4); @@ -44,8 +43,8 @@ static void michael_mic_hdr(struct michael_mic_ctx *mctx, michael_block(mctx, priority); } -void michael_mic(u8 *key, u8 *da, u8 *sa, u8 priority, - u8 *data, size_t data_len, u8 *mic) +void michael_mic(const u8 *key, const u8 *da, const u8 *sa, u8 priority, + const u8 *data, size_t data_len, u8 *mic) { u32 val; size_t block, blocks, left; diff --git a/net/mac80211/michael.h b/net/mac80211/michael.h index 7d5707bf8843..69b4501f13ba 100644 --- a/net/mac80211/michael.h +++ b/net/mac80211/michael.h @@ -18,7 +18,7 @@ struct michael_mic_ctx { u32 l, r; }; -void michael_mic(u8 *key, u8 *da, u8 *sa, u8 priority, - u8 *data, size_t data_len, u8 *mic); +void michael_mic(const u8 *key, const u8 *da, const u8 *sa, u8 priority, + const u8 *data, size_t data_len, u8 *mic); #endif /* MICHAEL_H */ diff --git a/net/mac80211/tkip.c b/net/mac80211/tkip.c index a7c3febc5a45..8cdf053cb835 100644 --- a/net/mac80211/tkip.c +++ b/net/mac80211/tkip.c @@ -6,7 +6,6 @@ * it under the terms of the GNU General Public License version 2 as * published by the Free Software Foundation. */ - #include #include #include @@ -132,7 +131,7 @@ static void tkip_mixing_phase2(const u16 *p1k, const u8 *tk, u16 tsc_IV16, /* Add TKIP IV and Ext. IV at @pos. @iv0, @iv1, and @iv2 are the first octets * of the IV. Returns pointer to the octet following IVs (i.e., beginning of * the packet payload). */ -u8 * ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, +u8 *ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, u8 iv0, u8 iv1, u8 iv2) { *pos++ = iv0; @@ -143,14 +142,7 @@ u8 * ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, return pos + 4; } -void ieee80211_tkip_gen_phase1key(struct ieee80211_key *key, u8 *ta, - u16 *phase1key) -{ - tkip_mixing_phase1(ta, &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], - key->u.tkip.iv32, phase1key); -} - -void ieee80211_tkip_gen_rc4key(struct ieee80211_key *key, u8 *ta, +static void ieee80211_tkip_gen_rc4key(struct ieee80211_key *key, u8 *ta, u8 *rc4key) { /* Calculate per-packet key */ diff --git a/net/mac80211/tkip.h b/net/mac80211/tkip.h index 1fa0bb4dba3c..b890427fc959 100644 --- a/net/mac80211/tkip.h +++ b/net/mac80211/tkip.h @@ -15,10 +15,6 @@ u8 *ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, u8 iv0, u8 iv1, u8 iv2); -void ieee80211_tkip_gen_phase1key(struct ieee80211_key *key, u8 *ta, - u16 *phase1key); -void ieee80211_tkip_gen_rc4key(struct ieee80211_key *key, u8 *ta, - u8 *rc4key); void ieee80211_tkip_encrypt_data(struct crypto_blkcipher *tfm, struct ieee80211_key *key, u8 *pos, size_t payload_len, u8 *ta); -- cgit v1.2.2 From b0f76b335f8b1c324b4b2be06369d391b26a7cc9 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 14 May 2008 16:26:19 -0700 Subject: mac80211: add a struct to hold tkip context Signed-off-by: Harvey Harrison Cc: "John W. Linville" Signed-off-by: Andrew Morton Signed-off-by: John W. Linville --- net/mac80211/cfg.c | 4 ++-- net/mac80211/debugfs_key.c | 8 +++---- net/mac80211/key.h | 17 ++++++++------- net/mac80211/tkip.c | 54 +++++++++++++++++++++++----------------------- net/mac80211/wpa.c | 16 +++++++------- 5 files changed, 50 insertions(+), 49 deletions(-) (limited to 'net') diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c index 3cef80dcd0e5..dbf0563c397d 100644 --- a/net/mac80211/cfg.c +++ b/net/mac80211/cfg.c @@ -256,8 +256,8 @@ static int ieee80211_get_key(struct wiphy *wiphy, struct net_device *dev, case ALG_TKIP: params.cipher = WLAN_CIPHER_SUITE_TKIP; - iv32 = key->u.tkip.iv32; - iv16 = key->u.tkip.iv16; + iv32 = key->u.tkip.tx.iv32; + iv16 = key->u.tkip.tx.iv16; if (key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE && sdata->local->ops->get_tkip_seq) diff --git a/net/mac80211/debugfs_key.c b/net/mac80211/debugfs_key.c index 19efc3a6a932..7439b63df5d0 100644 --- a/net/mac80211/debugfs_key.c +++ b/net/mac80211/debugfs_key.c @@ -97,8 +97,8 @@ static ssize_t key_tx_spec_read(struct file *file, char __user *userbuf, break; case ALG_TKIP: len = scnprintf(buf, sizeof(buf), "%08x %04x\n", - key->u.tkip.iv32, - key->u.tkip.iv16); + key->u.tkip.tx.iv32, + key->u.tkip.tx.iv16); break; case ALG_CCMP: tpn = key->u.ccmp.tx_pn; @@ -128,8 +128,8 @@ static ssize_t key_rx_spec_read(struct file *file, char __user *userbuf, for (i = 0; i < NUM_RX_DATA_QUEUES; i++) p += scnprintf(p, sizeof(buf)+buf-p, "%08x %04x\n", - key->u.tkip.iv32_rx[i], - key->u.tkip.iv16_rx[i]); + key->u.tkip.rx[i].iv32, + key->u.tkip.rx[i].iv16); len = p - buf; break; case ALG_CCMP: diff --git a/net/mac80211/key.h b/net/mac80211/key.h index f52c3df1fe9a..a0f774aafa45 100644 --- a/net/mac80211/key.h +++ b/net/mac80211/key.h @@ -69,6 +69,13 @@ enum ieee80211_internal_key_flags { KEY_FLAG_TODO_ADD_DEBUGFS = BIT(5), }; +struct tkip_ctx { + u32 iv32; + u16 iv16; + u16 p1k[5]; + int initialized; +}; + struct ieee80211_key { struct ieee80211_local *local; struct ieee80211_sub_if_data *sdata; @@ -85,16 +92,10 @@ struct ieee80211_key { union { struct { /* last used TSC */ - u32 iv32; - u16 iv16; - u16 p1k[5]; - int tx_initialized; + struct tkip_ctx tx; /* last received RSC */ - u32 iv32_rx[NUM_RX_DATA_QUEUES]; - u16 iv16_rx[NUM_RX_DATA_QUEUES]; - u16 p1k_rx[NUM_RX_DATA_QUEUES][5]; - int rx_initialized[NUM_RX_DATA_QUEUES]; + struct tkip_ctx rx[NUM_RX_DATA_QUEUES]; } tkip; struct { u8 tx_pn[6]; diff --git a/net/mac80211/tkip.c b/net/mac80211/tkip.c index 8cdf053cb835..d74c91e23a7f 100644 --- a/net/mac80211/tkip.c +++ b/net/mac80211/tkip.c @@ -138,7 +138,7 @@ u8 *ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, *pos++ = iv1; *pos++ = iv2; *pos++ = (key->conf.keyidx << 6) | (1 << 5) /* Ext IV */; - put_unaligned_le32(key->u.tkip.iv32, pos); + put_unaligned_le32(key->u.tkip.tx.iv32, pos); return pos + 4; } @@ -146,16 +146,16 @@ static void ieee80211_tkip_gen_rc4key(struct ieee80211_key *key, u8 *ta, u8 *rc4key) { /* Calculate per-packet key */ - if (key->u.tkip.iv16 == 0 || !key->u.tkip.tx_initialized) { + if (key->u.tkip.tx.iv16 == 0 || !key->u.tkip.tx.initialized) { /* IV16 wrapped around - perform TKIP phase 1 */ tkip_mixing_phase1(ta, &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], - key->u.tkip.iv32, key->u.tkip.p1k); - key->u.tkip.tx_initialized = 1; + key->u.tkip.tx.iv32, key->u.tkip.tx.p1k); + key->u.tkip.tx.initialized = 1; } - tkip_mixing_phase2(key->u.tkip.p1k, + tkip_mixing_phase2(key->u.tkip.tx.p1k, &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], - key->u.tkip.iv16, rc4key); + key->u.tkip.tx.iv16, rc4key); } void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, @@ -179,9 +179,9 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, printk(KERN_DEBUG "TKIP encrypt: iv16 = 0x%04x, iv32 = 0x%08x\n", iv16, iv32); - if (iv32 != key->u.tkip.iv32) { + if (iv32 != key->u.tkip.tx.iv32) { printk(KERN_DEBUG "skb: iv32 = 0x%08x key: iv32 = 0x%08x\n", - iv32, key->u.tkip.iv32); + iv32, key->u.tkip.tx.iv32); printk(KERN_DEBUG "Wrap around of iv16 in the middle of a " "fragmented packet\n"); } @@ -190,19 +190,19 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, /* Update the p1k only when the iv16 in the packet wraps around, this * might occur after the wrap around of iv16 in the key in case of * fragmented packets. */ - if (iv16 == 0 || !key->u.tkip.tx_initialized) { + if (iv16 == 0 || !key->u.tkip.tx.initialized) { /* IV16 wrapped around - perform TKIP phase 1 */ tkip_mixing_phase1(ta, &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], - iv32, key->u.tkip.p1k); - key->u.tkip.tx_initialized = 1; + iv32, key->u.tkip.tx.p1k); + key->u.tkip.tx.initialized = 1; } if (type == IEEE80211_TKIP_P1_KEY) { - memcpy(outkey, key->u.tkip.p1k, sizeof(u16) * 5); + memcpy(outkey, key->u.tkip.tx.p1k, sizeof(u16) * 5); return; } - tkip_mixing_phase2(key->u.tkip.p1k, + tkip_mixing_phase2(key->u.tkip.tx.p1k, &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], iv16, outkey); } EXPORT_SYMBOL(ieee80211_get_tkip_key); @@ -263,33 +263,33 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, if ((keyid >> 6) != key->conf.keyidx) return TKIP_DECRYPT_INVALID_KEYIDX; - if (key->u.tkip.rx_initialized[queue] && - (iv32 < key->u.tkip.iv32_rx[queue] || - (iv32 == key->u.tkip.iv32_rx[queue] && - iv16 <= key->u.tkip.iv16_rx[queue]))) { + if (key->u.tkip.rx[queue].initialized && + (iv32 < key->u.tkip.rx[queue].iv32 || + (iv32 == key->u.tkip.rx[queue].iv32 && + iv16 <= key->u.tkip.rx[queue].iv16))) { #ifdef CONFIG_TKIP_DEBUG DECLARE_MAC_BUF(mac); printk(KERN_DEBUG "TKIP replay detected for RX frame from " "%s (RX IV (%04x,%02x) <= prev. IV (%04x,%02x)\n", print_mac(mac, ta), - iv32, iv16, key->u.tkip.iv32_rx[queue], - key->u.tkip.iv16_rx[queue]); + iv32, iv16, key->u.tkip.rx[queue].iv32, + key->u.tkip.rx[queue].iv16); #endif /* CONFIG_TKIP_DEBUG */ return TKIP_DECRYPT_REPLAY; } if (only_iv) { res = TKIP_DECRYPT_OK; - key->u.tkip.rx_initialized[queue] = 1; + key->u.tkip.rx[queue].initialized = 1; goto done; } - if (!key->u.tkip.rx_initialized[queue] || - key->u.tkip.iv32_rx[queue] != iv32) { - key->u.tkip.rx_initialized[queue] = 1; + if (!key->u.tkip.rx[queue].initialized || + key->u.tkip.rx[queue].iv32 != iv32) { + key->u.tkip.rx[queue].initialized = 1; /* IV16 wrapped around - perform TKIP phase 1 */ tkip_mixing_phase1(ta, &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], - iv32, key->u.tkip.p1k_rx[queue]); + iv32, key->u.tkip.rx[queue].p1k); #ifdef CONFIG_TKIP_DEBUG { int i; @@ -303,7 +303,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, printk("\n"); printk(KERN_DEBUG "TKIP decrypt: P1K="); for (i = 0; i < 5; i++) - printk("%04x ", key->u.tkip.p1k_rx[queue][i]); + printk("%04x ", key->u.tkip.rx[queue].p1k[i]); printk("\n"); } #endif /* CONFIG_TKIP_DEBUG */ @@ -318,11 +318,11 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, key->local->ops->update_tkip_key( local_to_hw(key->local), &key->conf, - sta_addr, iv32, key->u.tkip.p1k_rx[queue]); + sta_addr, iv32, key->u.tkip.rx[queue].p1k); } } - tkip_mixing_phase2(key->u.tkip.p1k_rx[queue], + tkip_mixing_phase2(key->u.tkip.rx[queue].p1k, &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], iv16, rc4key); #ifdef CONFIG_TKIP_DEBUG diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index 42f3654e1c5e..d7304490d2ec 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -176,8 +176,8 @@ ieee80211_rx_h_michael_mic_verify(struct ieee80211_rx_data *rx) skb_trim(skb, skb->len - MICHAEL_MIC_LEN); /* update IV in key information to be able to detect replays */ - rx->key->u.tkip.iv32_rx[rx->queue] = rx->tkip_iv32; - rx->key->u.tkip.iv16_rx[rx->queue] = rx->tkip_iv16; + rx->key->u.tkip.rx[rx->queue].iv32 = rx->tkip_iv32; + rx->key->u.tkip.rx[rx->queue].iv16 = rx->tkip_iv16; return RX_CONTINUE; } @@ -214,19 +214,19 @@ static int tkip_encrypt_skb(struct ieee80211_tx_data *tx, pos += hdrlen; /* Increase IV for the frame */ - key->u.tkip.iv16++; - if (key->u.tkip.iv16 == 0) - key->u.tkip.iv32++; + key->u.tkip.tx.iv16++; + if (key->u.tkip.tx.iv16 == 0) + key->u.tkip.tx.iv32++; if (tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) { hdr = (struct ieee80211_hdr *)skb->data; /* hwaccel - with preallocated room for IV */ ieee80211_tkip_add_iv(pos, key, - (u8) (key->u.tkip.iv16 >> 8), - (u8) (((key->u.tkip.iv16 >> 8) | 0x20) & + (u8) (key->u.tkip.tx.iv16 >> 8), + (u8) (((key->u.tkip.tx.iv16 >> 8) | 0x20) & 0x7f), - (u8) key->u.tkip.iv16); + (u8) key->u.tkip.tx.iv16); tx->control->hw_key = &tx->key->conf; return 0; -- cgit v1.2.2 From 82a57447fa66bf138cd55206f33eea21ee257335 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 14 May 2008 16:26:19 -0700 Subject: mac80211: tkip.c use struct tkip_ctx in phase 1 key mixing Signed-off-by: Harvey Harrison Cc: "John W. Linville" Signed-off-by: Andrew Morton Signed-off-by: John W. Linville --- net/mac80211/tkip.c | 27 ++++++++++----------------- 1 file changed, 10 insertions(+), 17 deletions(-) (limited to 'net') diff --git a/net/mac80211/tkip.c b/net/mac80211/tkip.c index d74c91e23a7f..e9ee247e13f7 100644 --- a/net/mac80211/tkip.c +++ b/net/mac80211/tkip.c @@ -71,10 +71,12 @@ static u16 tkipS(u16 val) * TSC = TKIP sequence counter (48 bits, only 32 msb bits used) * P1K: 80 bits */ -static void tkip_mixing_phase1(const u8 *ta, const u8 *tk, u32 tsc_IV32, - u16 *p1k) +static void tkip_mixing_phase1(struct ieee80211_key *key, const u8 *ta, + struct tkip_ctx *ctx, u32 tsc_IV32) { int i, j; + const u8 *tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; + u16 *p1k = ctx->p1k; p1k[0] = tsc_IV32 & 0xFFFF; p1k[1] = tsc_IV32 >> 16; @@ -90,6 +92,7 @@ static void tkip_mixing_phase1(const u8 *ta, const u8 *tk, u32 tsc_IV32, p1k[3] += tkipS(p1k[2] ^ get_unaligned_le16(tk + 12 + j)); p1k[4] += tkipS(p1k[3] ^ get_unaligned_le16(tk + 0 + j)) + i; } + ctx->initialized = 1; } static void tkip_mixing_phase2(const u16 *p1k, const u8 *tk, u16 tsc_IV16, @@ -146,12 +149,8 @@ static void ieee80211_tkip_gen_rc4key(struct ieee80211_key *key, u8 *ta, u8 *rc4key) { /* Calculate per-packet key */ - if (key->u.tkip.tx.iv16 == 0 || !key->u.tkip.tx.initialized) { - /* IV16 wrapped around - perform TKIP phase 1 */ - tkip_mixing_phase1(ta, &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], - key->u.tkip.tx.iv32, key->u.tkip.tx.p1k); - key->u.tkip.tx.initialized = 1; - } + if (key->u.tkip.tx.iv16 == 0 || !key->u.tkip.tx.initialized) + tkip_mixing_phase1(key, ta, &key->u.tkip.tx, key->u.tkip.tx.iv32); tkip_mixing_phase2(key->u.tkip.tx.p1k, &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], @@ -190,12 +189,8 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, /* Update the p1k only when the iv16 in the packet wraps around, this * might occur after the wrap around of iv16 in the key in case of * fragmented packets. */ - if (iv16 == 0 || !key->u.tkip.tx.initialized) { - /* IV16 wrapped around - perform TKIP phase 1 */ - tkip_mixing_phase1(ta, &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], - iv32, key->u.tkip.tx.p1k); - key->u.tkip.tx.initialized = 1; - } + if (iv16 == 0 || !key->u.tkip.tx.initialized) + tkip_mixing_phase1(key, ta, &key->u.tkip.tx, iv32); if (type == IEEE80211_TKIP_P1_KEY) { memcpy(outkey, key->u.tkip.tx.p1k, sizeof(u16) * 5); @@ -286,10 +281,8 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, if (!key->u.tkip.rx[queue].initialized || key->u.tkip.rx[queue].iv32 != iv32) { - key->u.tkip.rx[queue].initialized = 1; /* IV16 wrapped around - perform TKIP phase 1 */ - tkip_mixing_phase1(ta, &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], - iv32, key->u.tkip.rx[queue].p1k); + tkip_mixing_phase1(key, ta, &key->u.tkip.rx[queue], iv32); #ifdef CONFIG_TKIP_DEBUG { int i; -- cgit v1.2.2 From 3c83809917dce9bbd880f6b08edc2d0ccac14a25 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 14 May 2008 16:26:20 -0700 Subject: mac80211: tkip.c use struct tkip_ctx in phase 2 key mixing Signed-off-by: Harvey Harrison Cc: "John W. Linville" Signed-off-by: Andrew Morton Signed-off-by: John W. Linville --- net/mac80211/tkip.c | 17 +++++++---------- 1 file changed, 7 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/mac80211/tkip.c b/net/mac80211/tkip.c index e9ee247e13f7..a00cf1ea7719 100644 --- a/net/mac80211/tkip.c +++ b/net/mac80211/tkip.c @@ -95,10 +95,12 @@ static void tkip_mixing_phase1(struct ieee80211_key *key, const u8 *ta, ctx->initialized = 1; } -static void tkip_mixing_phase2(const u16 *p1k, const u8 *tk, u16 tsc_IV16, - u8 *rc4key) +static void tkip_mixing_phase2(struct ieee80211_key *key, struct tkip_ctx *ctx, + u16 tsc_IV16, u8 *rc4key) { u16 ppk[6]; + const u16 *p1k = ctx->p1k; + const u8 *tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; int i; ppk[0] = p1k[0]; @@ -152,9 +154,7 @@ static void ieee80211_tkip_gen_rc4key(struct ieee80211_key *key, u8 *ta, if (key->u.tkip.tx.iv16 == 0 || !key->u.tkip.tx.initialized) tkip_mixing_phase1(key, ta, &key->u.tkip.tx, key->u.tkip.tx.iv32); - tkip_mixing_phase2(key->u.tkip.tx.p1k, - &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], - key->u.tkip.tx.iv16, rc4key); + tkip_mixing_phase2(key, &key->u.tkip.tx, key->u.tkip.tx.iv16, rc4key); } void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, @@ -197,8 +197,7 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, return; } - tkip_mixing_phase2(key->u.tkip.tx.p1k, - &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], iv16, outkey); + tkip_mixing_phase2(key, &key->u.tkip.tx, iv16, outkey); } EXPORT_SYMBOL(ieee80211_get_tkip_key); @@ -315,9 +314,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, } } - tkip_mixing_phase2(key->u.tkip.rx[queue].p1k, - &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY], - iv16, rc4key); + tkip_mixing_phase2(key, &key->u.tkip.rx[queue], iv16, rc4key); #ifdef CONFIG_TKIP_DEBUG { int i; -- cgit v1.2.2 From edcdf8b21ac920e06b4180246123fe43b022e020 Mon Sep 17 00:00:00 2001 From: Ron Rindjunsky Date: Thu, 15 May 2008 13:53:55 +0800 Subject: mac80211: separate Tx and Rx MCS when configuring HT This patch follows the 11n spec in separation between Tx and Rx MCS capabilities. Up until now, when configuring the HT possible set of Tx MCS only Rx MCS were considered, assuming they are the same as the Tx MCS. This patch fixed this by looking at low level driver Tx capabilities. Signed-off-by: Ron Rindjunsky Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/main.c | 96 ++++++++++++++++++++++++++++++++++------------------- 1 file changed, 61 insertions(+), 35 deletions(-) (limited to 'net') diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 36016363d225..b0fddb7de549 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -35,8 +35,6 @@ #include "debugfs.h" #include "debugfs_netdev.h" -#define SUPP_MCS_SET_LEN 16 - /* * For seeing transmitted packets on monitor interfaces * we have a radiotap header too. @@ -1068,56 +1066,84 @@ u32 ieee80211_handle_ht(struct ieee80211_local *local, int enable_ht, struct ieee80211_supported_band *sband; struct ieee80211_ht_info ht_conf; struct ieee80211_ht_bss_info ht_bss_conf; - int i; u32 changed = 0; + int i; + u8 max_tx_streams = IEEE80211_HT_CAP_MAX_STREAMS; + u8 tx_mcs_set_cap; sband = local->hw.wiphy->bands[conf->channel->band]; + memset(&ht_conf, 0, sizeof(struct ieee80211_ht_info)); + memset(&ht_bss_conf, 0, sizeof(struct ieee80211_ht_bss_info)); + /* HT is not supported */ if (!sband->ht_info.ht_supported) { conf->flags &= ~IEEE80211_CONF_SUPPORT_HT_MODE; - return 0; + goto out; } - memset(&ht_conf, 0, sizeof(struct ieee80211_ht_info)); - memset(&ht_bss_conf, 0, sizeof(struct ieee80211_ht_bss_info)); - - if (enable_ht) { - if (!(conf->flags & IEEE80211_CONF_SUPPORT_HT_MODE)) + /* disable HT */ + if (!enable_ht) { + if (conf->flags & IEEE80211_CONF_SUPPORT_HT_MODE) changed |= BSS_CHANGED_HT; + conf->flags &= ~IEEE80211_CONF_SUPPORT_HT_MODE; + conf->ht_conf.ht_supported = 0; + goto out; + } - conf->flags |= IEEE80211_CONF_SUPPORT_HT_MODE; - ht_conf.ht_supported = 1; - ht_conf.cap = req_ht_cap->cap & sband->ht_info.cap; - ht_conf.cap &= ~(IEEE80211_HT_CAP_MIMO_PS); - ht_conf.cap |= sband->ht_info.cap & IEEE80211_HT_CAP_MIMO_PS; + if (!(conf->flags & IEEE80211_CONF_SUPPORT_HT_MODE)) + changed |= BSS_CHANGED_HT; - for (i = 0; i < SUPP_MCS_SET_LEN; i++) - ht_conf.supp_mcs_set[i] = - sband->ht_info.supp_mcs_set[i] & - req_ht_cap->supp_mcs_set[i]; + conf->flags |= IEEE80211_CONF_SUPPORT_HT_MODE; + ht_conf.ht_supported = 1; - ht_bss_conf.primary_channel = req_bss_cap->primary_channel; - ht_bss_conf.bss_cap = req_bss_cap->bss_cap; - ht_bss_conf.bss_op_mode = req_bss_cap->bss_op_mode; + ht_conf.cap = req_ht_cap->cap & sband->ht_info.cap; + ht_conf.cap &= ~(IEEE80211_HT_CAP_MIMO_PS); + ht_conf.cap |= sband->ht_info.cap & IEEE80211_HT_CAP_MIMO_PS; + ht_bss_conf.primary_channel = req_bss_cap->primary_channel; + ht_bss_conf.bss_cap = req_bss_cap->bss_cap; + ht_bss_conf.bss_op_mode = req_bss_cap->bss_op_mode; - ht_conf.ampdu_factor = req_ht_cap->ampdu_factor; - ht_conf.ampdu_density = req_ht_cap->ampdu_density; + ht_conf.ampdu_factor = req_ht_cap->ampdu_factor; + ht_conf.ampdu_density = req_ht_cap->ampdu_density; - /* if bss configuration changed store the new one */ - if (memcmp(&conf->ht_conf, &ht_conf, sizeof(ht_conf)) || - memcmp(&conf->ht_bss_conf, &ht_bss_conf, sizeof(ht_bss_conf))) { - changed |= BSS_CHANGED_HT; - memcpy(&conf->ht_conf, &ht_conf, sizeof(ht_conf)); - memcpy(&conf->ht_bss_conf, &ht_bss_conf, sizeof(ht_bss_conf)); - } - } else { - if (conf->flags & IEEE80211_CONF_SUPPORT_HT_MODE) - changed |= BSS_CHANGED_HT; - conf->flags &= ~IEEE80211_CONF_SUPPORT_HT_MODE; - } + /* Bits 96-100 */ + tx_mcs_set_cap = sband->ht_info.supp_mcs_set[12]; + /* configure suppoerted Tx MCS according to requested MCS + * (based in most cases on Rx capabilities of peer) and self + * Tx MCS capabilities (as defined by low level driver HW + * Tx capabilities) */ + if (!(tx_mcs_set_cap & IEEE80211_HT_CAP_MCS_TX_DEFINED)) + goto check_changed; + + /* Counting from 0 therfore + 1 */ + if (tx_mcs_set_cap & IEEE80211_HT_CAP_MCS_TX_RX_DIFF) + max_tx_streams = ((tx_mcs_set_cap & + IEEE80211_HT_CAP_MCS_TX_STREAMS) >> 2) + 1; + + for (i = 0; i < max_tx_streams; i++) + ht_conf.supp_mcs_set[i] = + sband->ht_info.supp_mcs_set[i] & + req_ht_cap->supp_mcs_set[i]; + + if (tx_mcs_set_cap & IEEE80211_HT_CAP_MCS_TX_UEQM) + for (i = IEEE80211_SUPP_MCS_SET_UEQM; + i < IEEE80211_SUPP_MCS_SET_LEN; i++) + ht_conf.supp_mcs_set[i] = + sband->ht_info.supp_mcs_set[i] & + req_ht_cap->supp_mcs_set[i]; + +check_changed: + /* if bss configuration changed store the new one */ + if (memcmp(&conf->ht_conf, &ht_conf, sizeof(ht_conf)) || + memcmp(&conf->ht_bss_conf, &ht_bss_conf, sizeof(ht_bss_conf))) { + changed |= BSS_CHANGED_HT; + memcpy(&conf->ht_conf, &ht_conf, sizeof(ht_conf)); + memcpy(&conf->ht_bss_conf, &ht_bss_conf, sizeof(ht_bss_conf)); + } +out: return changed; } -- cgit v1.2.2 From c4680470a34a4f39af3d0a5c40f70befd8701908 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 15 May 2008 12:55:25 +0200 Subject: mac80211: fix bugs in queue handling functions Commit 55c308c1315bc7267dbb88011c208fd743cdce31 ("mac80211: QoS related cleanups") introduced another bug, the queue handling functions that operate on all queues now only operated on the first queues, not the A-MPDU queues as expected. This patch fixes this. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/util.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/mac80211/util.c b/net/mac80211/util.c index 24a465c4df09..9cd07e1031af 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -355,7 +355,7 @@ void ieee80211_start_queues(struct ieee80211_hw *hw) struct ieee80211_local *local = hw_to_local(hw); int i; - for (i = 0; i < local->hw.queues; i++) + for (i = 0; i < hw->queues + hw->ampdu_queues; i++) clear_bit(IEEE80211_LINK_STATE_XOFF, &local->state[i]); if (!ieee80211_qdisc_installed(local->mdev)) netif_start_queue(local->mdev); @@ -366,7 +366,7 @@ void ieee80211_stop_queues(struct ieee80211_hw *hw) { int i; - for (i = 0; i < hw->queues; i++) + for (i = 0; i < hw->queues + hw->ampdu_queues; i++) ieee80211_stop_queue(hw, i); } EXPORT_SYMBOL(ieee80211_stop_queues); @@ -375,7 +375,7 @@ void ieee80211_wake_queues(struct ieee80211_hw *hw) { int i; - for (i = 0; i < hw->queues; i++) + for (i = 0; i < hw->queues + hw->ampdu_queues; i++) ieee80211_wake_queue(hw, i); } EXPORT_SYMBOL(ieee80211_wake_queues); -- cgit v1.2.2 From 36d6825b91bc492b65b6333c369cd96a2fc8c903 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 15 May 2008 12:55:26 +0200 Subject: mac80211: let drivers wake but not start queues Having drivers start queues is just confusing, their ->start() callback can block and do whatever is necessary, so let mac80211 start queues and have drivers wake queues when necessary (to get packets flowing again right away.) Signed-off-by: Johannes Berg Acked-by: David S. Miller Signed-off-by: John W. Linville --- net/mac80211/main.c | 8 +++++++- net/mac80211/util.c | 12 ------------ 2 files changed, 7 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/mac80211/main.c b/net/mac80211/main.c index b0fddb7de549..9761d9bd5a79 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -110,7 +110,13 @@ static int ieee80211_master_open(struct net_device *dev) break; } } - return res; + + if (res) + return res; + + netif_start_queue(local->mdev); + + return 0; } static int ieee80211_master_stop(struct net_device *dev) diff --git a/net/mac80211/util.c b/net/mac80211/util.c index 9cd07e1031af..800c15aff6e7 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -350,18 +350,6 @@ void ieee80211_stop_queue(struct ieee80211_hw *hw, int queue) } EXPORT_SYMBOL(ieee80211_stop_queue); -void ieee80211_start_queues(struct ieee80211_hw *hw) -{ - struct ieee80211_local *local = hw_to_local(hw); - int i; - - for (i = 0; i < hw->queues + hw->ampdu_queues; i++) - clear_bit(IEEE80211_LINK_STATE_XOFF, &local->state[i]); - if (!ieee80211_qdisc_installed(local->mdev)) - netif_start_queue(local->mdev); -} -EXPORT_SYMBOL(ieee80211_start_queues); - void ieee80211_stop_queues(struct ieee80211_hw *hw) { int i; -- cgit v1.2.2 From 2e92e6f2c50b4baf85cca968f0e6f1b5c0df7d39 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 15 May 2008 12:55:27 +0200 Subject: mac80211: use rate index in TX control This patch modifies struct ieee80211_tx_control to give band info and the rate index (instead of rate pointers) to drivers. This mostly serves to reduce the TX control structure size to make it fit into skb->cb so that the fragmentation code can put it there and we can think about passing it to drivers that way in the future. The rt2x00 driver update was done by Ivo, thanks. Signed-off-by: Ivo van Doorn Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 8 ++-- net/mac80211/mlme.c | 6 +-- net/mac80211/rate.c | 12 +++-- net/mac80211/rate.h | 25 ++++------ net/mac80211/rc80211_pid_algo.c | 6 +-- net/mac80211/tx.c | 104 ++++++++++++++++++++++------------------ net/mac80211/util.c | 10 +++- 7 files changed, 93 insertions(+), 78 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index ed0d9b35ae6f..a4cccd1b7d53 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -159,11 +159,11 @@ struct ieee80211_tx_data { struct ieee80211_tx_control *control; struct ieee80211_channel *channel; - struct ieee80211_rate *rate; + s8 rate_idx; /* use this rate (if set) for last fragment; rate can * be set to lower rate for the first fragments, e.g., * when using CTS protection with IEEE 802.11g. */ - struct ieee80211_rate *last_frag_rate; + s8 last_frag_rate_idx; /* Extra fragments (in addition to the first fragment * in skb) */ @@ -225,9 +225,9 @@ struct ieee80211_tx_stored_packet { struct ieee80211_tx_control control; struct sk_buff *skb; struct sk_buff **extra_frag; - struct ieee80211_rate *last_frag_rate; + s8 last_frag_rate_idx; int num_extra_frag; - unsigned int last_frag_rate_ctrl_probe; + bool last_frag_rate_ctrl_probe; }; struct beacon_data { diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 7877d3b3f4cb..604149369dc9 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -2406,15 +2406,15 @@ static int ieee80211_sta_join_ibss(struct net_device *dev, memset(&control, 0, sizeof(control)); rate_control_get_rate(dev, sband, skb, &ratesel); - if (!ratesel.rate) { + if (ratesel.rate_idx < 0) { printk(KERN_DEBUG "%s: Failed to determine TX rate " "for IBSS beacon\n", dev->name); break; } control.vif = &sdata->vif; - control.tx_rate = ratesel.rate; + control.tx_rate_idx = ratesel.rate_idx; if (sdata->bss_conf.use_short_preamble && - ratesel.rate->flags & IEEE80211_RATE_SHORT_PREAMBLE) + sband->bitrates[ratesel.rate_idx].flags & IEEE80211_RATE_SHORT_PREAMBLE) control.flags |= IEEE80211_TXCTL_SHORT_PREAMBLE; control.antenna_sel_tx = local->hw.conf.antenna_sel_tx; control.flags |= IEEE80211_TXCTL_NO_ACK; diff --git a/net/mac80211/rate.c b/net/mac80211/rate.c index 841df93807fc..0388c090dfe9 100644 --- a/net/mac80211/rate.c +++ b/net/mac80211/rate.c @@ -176,20 +176,24 @@ void rate_control_get_rate(struct net_device *dev, rcu_read_lock(); sta = sta_info_get(local, hdr->addr1); - memset(sel, 0, sizeof(struct rate_selection)); + sel->rate_idx = -1; + sel->nonerp_idx = -1; + sel->probe_idx = -1; ref->ops->get_rate(ref->priv, dev, sband, skb, sel); + BUG_ON(sel->rate_idx < 0); + /* Select a non-ERP backup rate. */ - if (!sel->nonerp) { + if (sel->nonerp_idx < 0) { for (i = 0; i < sband->n_bitrates; i++) { struct ieee80211_rate *rate = &sband->bitrates[i]; - if (sel->rate->bitrate < rate->bitrate) + if (sband->bitrates[sel->rate_idx].bitrate < rate->bitrate) break; if (rate_supported(sta, sband->band, i) && !(rate->flags & IEEE80211_RATE_ERP_G)) - sel->nonerp = rate; + sel->nonerp_idx = i; } } diff --git a/net/mac80211/rate.h b/net/mac80211/rate.h index 5b45f33cb766..a29148dcca99 100644 --- a/net/mac80211/rate.h +++ b/net/mac80211/rate.h @@ -19,14 +19,15 @@ #include "ieee80211_i.h" #include "sta_info.h" -/* TODO: kdoc */ +/** + * struct rate_selection - rate selection for rate control algos + * @rate: selected transmission rate index + * @nonerp: Non-ERP rate to use instead if ERP cannot be used + * @probe: rate for probing (or -1) + * + */ struct rate_selection { - /* Selected transmission rate */ - struct ieee80211_rate *rate; - /* Non-ERP rate to use if mac80211 decides it cannot use an ERP rate */ - struct ieee80211_rate *nonerp; - /* probe with this rate, or NULL for no probing */ - struct ieee80211_rate *probe; + s8 rate_idx, nonerp_idx, probe_idx; }; struct rate_control_ops { @@ -138,7 +139,7 @@ static inline int rate_supported(struct sta_info *sta, return (sta == NULL || sta->supp_rates[band] & BIT(index)); } -static inline int +static inline s8 rate_lowest_index(struct ieee80211_local *local, struct ieee80211_supported_band *sband, struct sta_info *sta) @@ -155,14 +156,6 @@ rate_lowest_index(struct ieee80211_local *local, return 0; } -static inline struct ieee80211_rate * -rate_lowest(struct ieee80211_local *local, - struct ieee80211_supported_band *sband, - struct sta_info *sta) -{ - return &sband->bitrates[rate_lowest_index(local, sband, sta)]; -} - /* functions for rate control related to a device */ int ieee80211_init_rate_ctrl_alg(struct ieee80211_local *local, diff --git a/net/mac80211/rc80211_pid_algo.c b/net/mac80211/rc80211_pid_algo.c index a849b745bdb5..14cde36f5070 100644 --- a/net/mac80211/rc80211_pid_algo.c +++ b/net/mac80211/rc80211_pid_algo.c @@ -266,7 +266,7 @@ static void rate_control_pid_tx_status(void *priv, struct net_device *dev, /* Ignore all frames that were sent with a different rate than the rate * we currently advise mac80211 to use. */ - if (status->control.tx_rate != &sband->bitrates[sta->txrate_idx]) + if (status->control.tx_rate_idx != sta->txrate_idx) goto unlock; spinfo = sta->rate_ctrl_priv; @@ -330,7 +330,7 @@ static void rate_control_pid_get_rate(void *priv, struct net_device *dev, fc = le16_to_cpu(hdr->frame_control); if ((fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_DATA || is_multicast_ether_addr(hdr->addr1) || !sta) { - sel->rate = rate_lowest(local, sband, sta); + sel->rate_idx = rate_lowest_index(local, sband, sta); rcu_read_unlock(); return; } @@ -349,7 +349,7 @@ static void rate_control_pid_get_rate(void *priv, struct net_device *dev, rcu_read_unlock(); - sel->rate = &sband->bitrates[rateidx]; + sel->rate_idx = rateidx; #ifdef CONFIG_MAC80211_DEBUGFS rate_control_pid_event_tx_rate( diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index aecec2a72b08..99c3860bc0e2 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -91,11 +91,12 @@ static u16 ieee80211_duration(struct ieee80211_tx_data *tx, int group_addr, int next_frag_len) { int rate, mrate, erp, dur, i; - struct ieee80211_rate *txrate = tx->rate; + struct ieee80211_rate *txrate; struct ieee80211_local *local = tx->local; struct ieee80211_supported_band *sband; - sband = local->hw.wiphy->bands[local->hw.conf.channel->band]; + sband = local->hw.wiphy->bands[tx->channel->band]; + txrate = &sband->bitrates[tx->rate_idx]; erp = 0; if (tx->sdata->flags & IEEE80211_SDATA_OPERATING_GMODE) @@ -610,40 +611,40 @@ ieee80211_tx_h_rate_ctrl(struct ieee80211_tx_data *tx) struct rate_selection rsel; struct ieee80211_supported_band *sband; - sband = tx->local->hw.wiphy->bands[tx->local->hw.conf.channel->band]; + sband = tx->local->hw.wiphy->bands[tx->channel->band]; - if (likely(!tx->rate)) { + if (likely(tx->rate_idx < 0)) { rate_control_get_rate(tx->dev, sband, tx->skb, &rsel); - tx->rate = rsel.rate; - if (unlikely(rsel.probe)) { + tx->rate_idx = rsel.rate_idx; + if (unlikely(rsel.probe_idx >= 0)) { tx->control->flags |= IEEE80211_TXCTL_RATE_CTRL_PROBE; tx->flags |= IEEE80211_TX_PROBE_LAST_FRAG; - tx->control->alt_retry_rate = tx->rate; - tx->rate = rsel.probe; + tx->control->alt_retry_rate_idx = tx->rate_idx; + tx->rate_idx = rsel.probe_idx; } else - tx->control->alt_retry_rate = NULL; + tx->control->alt_retry_rate_idx = -1; - if (!tx->rate) + if (unlikely(tx->rate_idx < 0)) return TX_DROP; } else - tx->control->alt_retry_rate = NULL; + tx->control->alt_retry_rate_idx = -1; if (tx->sdata->bss_conf.use_cts_prot && - (tx->flags & IEEE80211_TX_FRAGMENTED) && rsel.nonerp) { - tx->last_frag_rate = tx->rate; - if (rsel.probe) + (tx->flags & IEEE80211_TX_FRAGMENTED) && (rsel.nonerp_idx >= 0)) { + tx->last_frag_rate_idx = tx->rate_idx; + if (rsel.probe_idx >= 0) tx->flags &= ~IEEE80211_TX_PROBE_LAST_FRAG; else tx->flags |= IEEE80211_TX_PROBE_LAST_FRAG; - tx->rate = rsel.nonerp; - tx->control->tx_rate = rsel.nonerp; + tx->rate_idx = rsel.nonerp_idx; + tx->control->tx_rate_idx = rsel.nonerp_idx; tx->control->flags &= ~IEEE80211_TXCTL_RATE_CTRL_PROBE; } else { - tx->last_frag_rate = tx->rate; - tx->control->tx_rate = tx->rate; + tx->last_frag_rate_idx = tx->rate_idx; + tx->control->tx_rate_idx = tx->rate_idx; } - tx->control->tx_rate = tx->rate; + tx->control->tx_rate_idx = tx->rate_idx; return TX_CONTINUE; } @@ -655,6 +656,9 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) u16 fc = le16_to_cpu(hdr->frame_control); u16 dur; struct ieee80211_tx_control *control = tx->control; + struct ieee80211_supported_band *sband; + + sband = tx->local->hw.wiphy->bands[tx->channel->band]; if (!control->retry_limit) { if (!is_multicast_ether_addr(hdr->addr1)) { @@ -681,14 +685,14 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) * frames. * TODO: The last fragment could still use multiple retry * rates. */ - control->alt_retry_rate = NULL; + control->alt_retry_rate_idx = -1; } /* Use CTS protection for unicast frames sent using extended rates if * there are associated non-ERP stations and RTS/CTS is not configured * for the frame. */ if ((tx->sdata->flags & IEEE80211_SDATA_OPERATING_GMODE) && - (tx->rate->flags & IEEE80211_RATE_ERP_G) && + (sband->bitrates[tx->rate_idx].flags & IEEE80211_RATE_ERP_G) && (tx->flags & IEEE80211_TX_UNICAST) && tx->sdata->bss_conf.use_cts_prot && !(control->flags & IEEE80211_TXCTL_USE_RTS_CTS)) @@ -698,7 +702,7 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) * short preambles at the selected rate and short preambles are * available on the network at the current point in time. */ if (((fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA) && - (tx->rate->flags & IEEE80211_RATE_SHORT_PREAMBLE) && + (sband->bitrates[tx->rate_idx].flags & IEEE80211_RATE_SHORT_PREAMBLE) && tx->sdata->bss_conf.use_short_preamble && (!tx->sta || test_sta_flags(tx->sta, WLAN_STA_SHORT_PREAMBLE))) { tx->control->flags |= IEEE80211_TXCTL_SHORT_PREAMBLE; @@ -715,32 +719,32 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) if ((control->flags & IEEE80211_TXCTL_USE_RTS_CTS) || (control->flags & IEEE80211_TXCTL_USE_CTS_PROTECT)) { struct ieee80211_supported_band *sband; - struct ieee80211_rate *rate, *baserate; + struct ieee80211_rate *rate; + s8 baserate = -1; int idx; - sband = tx->local->hw.wiphy->bands[ - tx->local->hw.conf.channel->band]; + sband = tx->local->hw.wiphy->bands[tx->channel->band]; /* Do not use multiple retry rates when using RTS/CTS */ - control->alt_retry_rate = NULL; + control->alt_retry_rate_idx = -1; /* Use min(data rate, max base rate) as CTS/RTS rate */ - rate = tx->rate; - baserate = NULL; + rate = &sband->bitrates[tx->rate_idx]; for (idx = 0; idx < sband->n_bitrates; idx++) { if (sband->bitrates[idx].bitrate > rate->bitrate) continue; if (tx->sdata->basic_rates & BIT(idx) && - (!baserate || - (baserate->bitrate < sband->bitrates[idx].bitrate))) - baserate = &sband->bitrates[idx]; + (baserate < 0 || + (sband->bitrates[baserate].bitrate + < sband->bitrates[idx].bitrate))) + baserate = idx; } - if (baserate) - control->rts_cts_rate = baserate; + if (baserate >= 0) + control->rts_cts_rate_idx = baserate; else - control->rts_cts_rate = &sband->bitrates[0]; + control->rts_cts_rate_idx = 0; } if (tx->sta) { @@ -768,7 +772,11 @@ ieee80211_tx_h_load_stats(struct ieee80211_tx_data *tx) struct sk_buff *skb = tx->skb; struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; u32 load = 0, hdrtime; - struct ieee80211_rate *rate = tx->rate; + struct ieee80211_rate *rate; + struct ieee80211_supported_band *sband; + + sband = tx->local->hw.wiphy->bands[tx->channel->band]; + rate = &sband->bitrates[tx->rate_idx]; /* TODO: this could be part of tx_status handling, so that the number * of retries would be known; TX rate should in that case be stored @@ -803,7 +811,7 @@ ieee80211_tx_h_load_stats(struct ieee80211_tx_data *tx) for (i = 0; i < tx->num_extra_frag; i++) { load += 2 * hdrtime; load += tx->extra_frag[i]->len * - tx->rate->bitrate; + rate->bitrate; } } @@ -859,7 +867,7 @@ __ieee80211_parse_tx_radiotap(struct ieee80211_tx_data *tx, int ret = ieee80211_radiotap_iterator_init(&iterator, rthdr, skb->len); struct ieee80211_tx_control *control = tx->control; - sband = tx->local->hw.wiphy->bands[tx->local->hw.conf.channel->band]; + sband = tx->local->hw.wiphy->bands[tx->channel->band]; control->flags |= IEEE80211_TXCTL_DO_NOT_ENCRYPT; tx->flags |= IEEE80211_TX_INJECTED; @@ -899,7 +907,7 @@ __ieee80211_parse_tx_radiotap(struct ieee80211_tx_data *tx, r = &sband->bitrates[i]; if (r->bitrate == target_rate) { - tx->rate = r; + tx->rate_idx = i; break; } } @@ -1097,7 +1105,7 @@ static int __ieee80211_tx(struct ieee80211_local *local, struct sk_buff *skb, if (__ieee80211_queue_stopped(local, control->queue)) return IEEE80211_TX_FRAG_AGAIN; if (i == tx->num_extra_frag) { - control->tx_rate = tx->last_frag_rate; + control->tx_rate_idx = tx->last_frag_rate_idx; if (tx->flags & IEEE80211_TX_PROBE_LAST_FRAG) control->flags |= @@ -1155,6 +1163,7 @@ static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb, sta = tx.sta; tx.channel = local->hw.conf.channel; + control->band = tx.channel->band; for (handler = ieee80211_tx_handlers; *handler != NULL; handler++) { @@ -1187,7 +1196,7 @@ static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb, next_len = tx.extra_frag[i + 1]->len; } else { next_len = 0; - tx.rate = tx.last_frag_rate; + tx.rate_idx = tx.last_frag_rate_idx; } dur = ieee80211_duration(&tx, 0, next_len); hdr->duration_id = cpu_to_le16(dur); @@ -1224,7 +1233,7 @@ retry: store->skb = skb; store->extra_frag = tx.extra_frag; store->num_extra_frag = tx.num_extra_frag; - store->last_frag_rate = tx.last_frag_rate; + store->last_frag_rate_idx = tx.last_frag_rate_idx; store->last_frag_rate_ctrl_probe = !!(tx.flags & IEEE80211_TX_PROBE_LAST_FRAG); } @@ -1685,7 +1694,7 @@ void ieee80211_tx_pending(unsigned long data) tx.control = &store->control; tx.extra_frag = store->extra_frag; tx.num_extra_frag = store->num_extra_frag; - tx.last_frag_rate = store->last_frag_rate; + tx.last_frag_rate_idx = store->last_frag_rate_idx; tx.flags = 0; if (store->last_frag_rate_ctrl_probe) tx.flags |= IEEE80211_TX_PROBE_LAST_FRAG; @@ -1789,9 +1798,10 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, struct ieee80211_mgmt *mgmt; int *num_beacons; bool err = true; + enum ieee80211_band band = local->hw.conf.channel->band; u8 *pos; - sband = local->hw.wiphy->bands[local->hw.conf.channel->band]; + sband = local->hw.wiphy->bands[band]; rcu_read_lock(); @@ -1885,8 +1895,9 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, } if (control) { + control->band = band; rate_control_get_rate(local->mdev, sband, skb, &rsel); - if (!rsel.rate) { + if (unlikely(rsel.rate_idx < 0)) { if (net_ratelimit()) { printk(KERN_DEBUG "%s: ieee80211_beacon_get: " "no rate found\n", @@ -1898,9 +1909,9 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, } control->vif = vif; - control->tx_rate = rsel.rate; + control->tx_rate_idx = rsel.rate_idx; if (sdata->bss_conf.use_short_preamble && - rsel.rate->flags & IEEE80211_RATE_SHORT_PREAMBLE) + sband->bitrates[rsel.rate_idx].flags & IEEE80211_RATE_SHORT_PREAMBLE) control->flags |= IEEE80211_TXCTL_SHORT_PREAMBLE; control->antenna_sel_tx = local->hw.conf.antenna_sel_tx; control->flags |= IEEE80211_TXCTL_NO_ACK; @@ -2006,6 +2017,7 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw, sta = tx.sta; tx.flags |= IEEE80211_TX_PS_BUFFERED; tx.channel = local->hw.conf.channel; + control->band = tx.channel->band; for (handler = ieee80211_tx_handlers; *handler != NULL; handler++) { res = (*handler)(&tx); diff --git a/net/mac80211/util.c b/net/mac80211/util.c index 800c15aff6e7..65a34fddeb00 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -266,10 +266,13 @@ __le16 ieee80211_rts_duration(struct ieee80211_hw *hw, bool short_preamble; int erp; u16 dur; + struct ieee80211_supported_band *sband; + + sband = local->hw.wiphy->bands[local->hw.conf.channel->band]; short_preamble = sdata->bss_conf.use_short_preamble; - rate = frame_txctl->rts_cts_rate; + rate = &sband->bitrates[frame_txctl->rts_cts_rate_idx]; erp = 0; if (sdata->flags & IEEE80211_SDATA_OPERATING_GMODE) @@ -300,10 +303,13 @@ __le16 ieee80211_ctstoself_duration(struct ieee80211_hw *hw, bool short_preamble; int erp; u16 dur; + struct ieee80211_supported_band *sband; + + sband = local->hw.wiphy->bands[local->hw.conf.channel->band]; short_preamble = sdata->bss_conf.use_short_preamble; - rate = frame_txctl->rts_cts_rate; + rate = &sband->bitrates[frame_txctl->rts_cts_rate_idx]; erp = 0; if (sdata->flags & IEEE80211_SDATA_OPERATING_GMODE) erp = rate->flags & IEEE80211_RATE_ERP_G; -- cgit v1.2.2 From e24549485f859be6518929bb1c9c0257d79f033d Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 15 May 2008 12:55:28 +0200 Subject: mac80211: reorder some transmit handlers The next patch will require that transmit handlers that are after fragmentation are aware of the fact that the control info is also fragmented. To make that easier, this patch moves a number of transmit handlers before fragmentation. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/tx.c | 234 +++++++++++++++++++++++++++--------------------------- 1 file changed, 118 insertions(+), 116 deletions(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 99c3860bc0e2..666158f02a89 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -505,106 +505,6 @@ ieee80211_tx_h_select_key(struct ieee80211_tx_data *tx) return TX_CONTINUE; } -static ieee80211_tx_result -ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) -{ - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) tx->skb->data; - size_t hdrlen, per_fragm, num_fragm, payload_len, left; - struct sk_buff **frags, *first, *frag; - int i; - u16 seq; - u8 *pos; - int frag_threshold = tx->local->fragmentation_threshold; - - if (!(tx->flags & IEEE80211_TX_FRAGMENTED)) - return TX_CONTINUE; - - first = tx->skb; - - hdrlen = ieee80211_get_hdrlen(tx->fc); - payload_len = first->len - hdrlen; - per_fragm = frag_threshold - hdrlen - FCS_LEN; - num_fragm = DIV_ROUND_UP(payload_len, per_fragm); - - frags = kzalloc(num_fragm * sizeof(struct sk_buff *), GFP_ATOMIC); - if (!frags) - goto fail; - - hdr->frame_control |= cpu_to_le16(IEEE80211_FCTL_MOREFRAGS); - seq = le16_to_cpu(hdr->seq_ctrl) & IEEE80211_SCTL_SEQ; - pos = first->data + hdrlen + per_fragm; - left = payload_len - per_fragm; - for (i = 0; i < num_fragm - 1; i++) { - struct ieee80211_hdr *fhdr; - size_t copylen; - - if (left <= 0) - goto fail; - - /* reserve enough extra head and tail room for possible - * encryption */ - frag = frags[i] = - dev_alloc_skb(tx->local->tx_headroom + - frag_threshold + - IEEE80211_ENCRYPT_HEADROOM + - IEEE80211_ENCRYPT_TAILROOM); - if (!frag) - goto fail; - /* Make sure that all fragments use the same priority so - * that they end up using the same TX queue */ - frag->priority = first->priority; - skb_reserve(frag, tx->local->tx_headroom + - IEEE80211_ENCRYPT_HEADROOM); - fhdr = (struct ieee80211_hdr *) skb_put(frag, hdrlen); - memcpy(fhdr, first->data, hdrlen); - if (i == num_fragm - 2) - fhdr->frame_control &= cpu_to_le16(~IEEE80211_FCTL_MOREFRAGS); - fhdr->seq_ctrl = cpu_to_le16(seq | ((i + 1) & IEEE80211_SCTL_FRAG)); - copylen = left > per_fragm ? per_fragm : left; - memcpy(skb_put(frag, copylen), pos, copylen); - - pos += copylen; - left -= copylen; - } - skb_trim(first, hdrlen + per_fragm); - - tx->num_extra_frag = num_fragm - 1; - tx->extra_frag = frags; - - return TX_CONTINUE; - - fail: - printk(KERN_DEBUG "%s: failed to fragment frame\n", tx->dev->name); - if (frags) { - for (i = 0; i < num_fragm - 1; i++) - if (frags[i]) - dev_kfree_skb(frags[i]); - kfree(frags); - } - I802_DEBUG_INC(tx->local->tx_handlers_drop_fragment); - return TX_DROP; -} - -static ieee80211_tx_result -ieee80211_tx_h_encrypt(struct ieee80211_tx_data *tx) -{ - if (!tx->key) - return TX_CONTINUE; - - switch (tx->key->conf.alg) { - case ALG_WEP: - return ieee80211_crypto_wep_encrypt(tx); - case ALG_TKIP: - return ieee80211_crypto_tkip_encrypt(tx); - case ALG_CCMP: - return ieee80211_crypto_ccmp_encrypt(tx); - } - - /* not reached */ - WARN_ON(1); - return TX_DROP; -} - static ieee80211_tx_result ieee80211_tx_h_rate_ctrl(struct ieee80211_tx_data *tx) { @@ -747,26 +647,114 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) control->rts_cts_rate_idx = 0; } - if (tx->sta) { + if (tx->sta) control->aid = tx->sta->aid; - tx->sta->tx_packets++; - tx->sta->tx_fragments++; - tx->sta->tx_bytes += tx->skb->len; - if (tx->extra_frag) { - int i; - tx->sta->tx_fragments += tx->num_extra_frag; - for (i = 0; i < tx->num_extra_frag; i++) { - tx->sta->tx_bytes += - tx->extra_frag[i]->len; - } - } + + return TX_CONTINUE; +} + +static ieee80211_tx_result +ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) +{ + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) tx->skb->data; + size_t hdrlen, per_fragm, num_fragm, payload_len, left; + struct sk_buff **frags, *first, *frag; + int i; + u16 seq; + u8 *pos; + int frag_threshold = tx->local->fragmentation_threshold; + + if (!(tx->flags & IEEE80211_TX_FRAGMENTED)) + return TX_CONTINUE; + + first = tx->skb; + + hdrlen = ieee80211_get_hdrlen(tx->fc); + payload_len = first->len - hdrlen; + per_fragm = frag_threshold - hdrlen - FCS_LEN; + num_fragm = DIV_ROUND_UP(payload_len, per_fragm); + + frags = kzalloc(num_fragm * sizeof(struct sk_buff *), GFP_ATOMIC); + if (!frags) + goto fail; + + hdr->frame_control |= cpu_to_le16(IEEE80211_FCTL_MOREFRAGS); + seq = le16_to_cpu(hdr->seq_ctrl) & IEEE80211_SCTL_SEQ; + pos = first->data + hdrlen + per_fragm; + left = payload_len - per_fragm; + for (i = 0; i < num_fragm - 1; i++) { + struct ieee80211_hdr *fhdr; + size_t copylen; + + if (left <= 0) + goto fail; + + /* reserve enough extra head and tail room for possible + * encryption */ + frag = frags[i] = + dev_alloc_skb(tx->local->tx_headroom + + frag_threshold + + IEEE80211_ENCRYPT_HEADROOM + + IEEE80211_ENCRYPT_TAILROOM); + if (!frag) + goto fail; + /* Make sure that all fragments use the same priority so + * that they end up using the same TX queue */ + frag->priority = first->priority; + skb_reserve(frag, tx->local->tx_headroom + + IEEE80211_ENCRYPT_HEADROOM); + fhdr = (struct ieee80211_hdr *) skb_put(frag, hdrlen); + memcpy(fhdr, first->data, hdrlen); + if (i == num_fragm - 2) + fhdr->frame_control &= cpu_to_le16(~IEEE80211_FCTL_MOREFRAGS); + fhdr->seq_ctrl = cpu_to_le16(seq | ((i + 1) & IEEE80211_SCTL_FRAG)); + copylen = left > per_fragm ? per_fragm : left; + memcpy(skb_put(frag, copylen), pos, copylen); + + pos += copylen; + left -= copylen; } + skb_trim(first, hdrlen + per_fragm); + + tx->num_extra_frag = num_fragm - 1; + tx->extra_frag = frags; return TX_CONTINUE; + + fail: + printk(KERN_DEBUG "%s: failed to fragment frame\n", tx->dev->name); + if (frags) { + for (i = 0; i < num_fragm - 1; i++) + if (frags[i]) + dev_kfree_skb(frags[i]); + kfree(frags); + } + I802_DEBUG_INC(tx->local->tx_handlers_drop_fragment); + return TX_DROP; } static ieee80211_tx_result -ieee80211_tx_h_load_stats(struct ieee80211_tx_data *tx) +ieee80211_tx_h_encrypt(struct ieee80211_tx_data *tx) +{ + if (!tx->key) + return TX_CONTINUE; + + switch (tx->key->conf.alg) { + case ALG_WEP: + return ieee80211_crypto_wep_encrypt(tx); + case ALG_TKIP: + return ieee80211_crypto_tkip_encrypt(tx); + case ALG_CCMP: + return ieee80211_crypto_ccmp_encrypt(tx); + } + + /* not reached */ + WARN_ON(1); + return TX_DROP; +} + +static ieee80211_tx_result +ieee80211_tx_h_stats(struct ieee80211_tx_data *tx) { struct ieee80211_local *local = tx->local; struct sk_buff *skb = tx->skb; @@ -822,6 +810,20 @@ ieee80211_tx_h_load_stats(struct ieee80211_tx_data *tx) tx->sta->channel_use_raw += load; tx->sdata->channel_use_raw += load; + if (tx->sta) { + tx->sta->tx_packets++; + tx->sta->tx_fragments++; + tx->sta->tx_bytes += tx->skb->len; + if (tx->extra_frag) { + int i; + tx->sta->tx_fragments += tx->num_extra_frag; + for (i = 0; i < tx->num_extra_frag; i++) { + tx->sta->tx_bytes += + tx->extra_frag[i]->len; + } + } + } + return TX_CONTINUE; } @@ -834,11 +836,11 @@ static ieee80211_tx_handler ieee80211_tx_handlers[] = ieee80211_tx_h_ps_buf, ieee80211_tx_h_select_key, ieee80211_tx_h_michael_mic_add, - ieee80211_tx_h_fragment, - ieee80211_tx_h_encrypt, ieee80211_tx_h_rate_ctrl, ieee80211_tx_h_misc, - ieee80211_tx_h_load_stats, + ieee80211_tx_h_fragment, + ieee80211_tx_h_encrypt, + ieee80211_tx_h_stats, NULL }; -- cgit v1.2.2 From e039fa4a4195ac4ee895e6f3d1334beed63256fe Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 15 May 2008 12:55:29 +0200 Subject: mac80211: move TX info into skb->cb This patch converts mac80211 and all drivers to have transmit information and status in skb->cb rather than allocating extra memory for it and copying all the data around. To make it fit, a union is used where only data that is necessary for all steps is kept outside of the union. A number of fixes were done by Ivo, as well as the rt2x00 part of this patch. Signed-off-by: Ivo van Doorn Signed-off-by: Johannes Berg Acked-by: David S. Miller Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 18 +-- net/mac80211/main.c | 128 +++++---------- net/mac80211/mlme.c | 30 ++-- net/mac80211/rate.h | 8 +- net/mac80211/rc80211_pid.h | 4 +- net/mac80211/rc80211_pid_algo.c | 18 +-- net/mac80211/rc80211_pid_debugfs.c | 8 +- net/mac80211/rx.c | 10 +- net/mac80211/sta_info.c | 6 +- net/mac80211/sta_info.h | 2 +- net/mac80211/tx.c | 310 ++++++++++++++++++------------------- net/mac80211/util.c | 10 +- net/mac80211/wep.c | 9 +- net/mac80211/wme.c | 22 ++- net/mac80211/wpa.c | 65 ++++---- 15 files changed, 285 insertions(+), 363 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index a4cccd1b7d53..79a65b3ee02b 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -2,6 +2,7 @@ * Copyright 2002-2005, Instant802 Networks, Inc. * Copyright 2005, Devicescape Software, Inc. * Copyright 2006-2007 Jiri Benc + * Copyright 2007-2008 Johannes Berg * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License version 2 as @@ -147,7 +148,6 @@ typedef unsigned __bitwise__ ieee80211_tx_result; #define IEEE80211_TX_UNICAST BIT(1) #define IEEE80211_TX_PS_BUFFERED BIT(2) #define IEEE80211_TX_PROBE_LAST_FRAG BIT(3) -#define IEEE80211_TX_INJECTED BIT(4) struct ieee80211_tx_data { struct sk_buff *skb; @@ -157,7 +157,6 @@ struct ieee80211_tx_data { struct sta_info *sta; struct ieee80211_key *key; - struct ieee80211_tx_control *control; struct ieee80211_channel *channel; s8 rate_idx; /* use this rate (if set) for last fragment; rate can @@ -207,22 +206,7 @@ struct ieee80211_rx_data { u16 tkip_iv16; }; -/* flags used in struct ieee80211_tx_packet_data.flags */ -#define IEEE80211_TXPD_REQ_TX_STATUS BIT(0) -#define IEEE80211_TXPD_DO_NOT_ENCRYPT BIT(1) -#define IEEE80211_TXPD_REQUEUE BIT(2) -#define IEEE80211_TXPD_EAPOL_FRAME BIT(3) -#define IEEE80211_TXPD_AMPDU BIT(4) -/* Stored in sk_buff->cb */ -struct ieee80211_tx_packet_data { - int ifindex; - unsigned long jiffies; - unsigned int flags; - u8 queue; -}; - struct ieee80211_tx_stored_packet { - struct ieee80211_tx_control control; struct sk_buff *skb; struct sk_buff **extra_frag; s8 last_frag_rate_idx; diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 9761d9bd5a79..8f1ff7ef1667 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -971,8 +971,7 @@ void ieee80211_if_setup(struct net_device *dev) /* everything else */ static int __ieee80211_if_config(struct net_device *dev, - struct sk_buff *beacon, - struct ieee80211_tx_control *control) + struct sk_buff *beacon) { struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); @@ -990,13 +989,11 @@ static int __ieee80211_if_config(struct net_device *dev, conf.ssid_len = sdata->u.sta.ssid_len; } else if (ieee80211_vif_is_mesh(&sdata->vif)) { conf.beacon = beacon; - conf.beacon_control = control; ieee80211_start_mesh(dev); } else if (sdata->vif.type == IEEE80211_IF_TYPE_AP) { conf.ssid = sdata->u.ap.ssid; conf.ssid_len = sdata->u.ap.ssid_len; conf.beacon = beacon; - conf.beacon_control = control; } return local->ops->config_interface(local_to_hw(local), &sdata->vif, &conf); @@ -1009,23 +1006,21 @@ int ieee80211_if_config(struct net_device *dev) if (sdata->vif.type == IEEE80211_IF_TYPE_MESH_POINT && (local->hw.flags & IEEE80211_HW_HOST_GEN_BEACON_TEMPLATE)) return ieee80211_if_config_beacon(dev); - return __ieee80211_if_config(dev, NULL, NULL); + return __ieee80211_if_config(dev, NULL); } int ieee80211_if_config_beacon(struct net_device *dev) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee80211_tx_control control; struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); struct sk_buff *skb; if (!(local->hw.flags & IEEE80211_HW_HOST_GEN_BEACON_TEMPLATE)) return 0; - skb = ieee80211_beacon_get(local_to_hw(local), &sdata->vif, - &control); + skb = ieee80211_beacon_get(local_to_hw(local), &sdata->vif); if (!skb) return -ENOMEM; - return __ieee80211_if_config(dev, skb, &control); + return __ieee80211_if_config(dev, skb); } int ieee80211_hw_config(struct ieee80211_local *local) @@ -1180,38 +1175,20 @@ void ieee80211_reset_erp_info(struct net_device *dev) } void ieee80211_tx_status_irqsafe(struct ieee80211_hw *hw, - struct sk_buff *skb, - struct ieee80211_tx_status *status) + struct sk_buff *skb) { struct ieee80211_local *local = hw_to_local(hw); - struct ieee80211_tx_status *saved; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); int tmp; skb->dev = local->mdev; - saved = kmalloc(sizeof(struct ieee80211_tx_status), GFP_ATOMIC); - if (unlikely(!saved)) { - if (net_ratelimit()) - printk(KERN_WARNING "%s: Not enough memory, " - "dropping tx status", skb->dev->name); - /* should be dev_kfree_skb_irq, but due to this function being - * named _irqsafe instead of just _irq we can't be sure that - * people won't call it from non-irq contexts */ - dev_kfree_skb_any(skb); - return; - } - memcpy(saved, status, sizeof(struct ieee80211_tx_status)); - /* copy pointer to saved status into skb->cb for use by tasklet */ - memcpy(skb->cb, &saved, sizeof(saved)); - skb->pkt_type = IEEE80211_TX_STATUS_MSG; - skb_queue_tail(status->control.flags & IEEE80211_TXCTL_REQ_TX_STATUS ? + skb_queue_tail(info->flags & IEEE80211_TX_CTL_REQ_TX_STATUS ? &local->skb_queue : &local->skb_queue_unreliable, skb); tmp = skb_queue_len(&local->skb_queue) + skb_queue_len(&local->skb_queue_unreliable); while (tmp > IEEE80211_IRQSAFE_QUEUE_LIMIT && (skb = skb_dequeue(&local->skb_queue_unreliable))) { - memcpy(&saved, skb->cb, sizeof(saved)); - kfree(saved); dev_kfree_skb_irq(skb); tmp--; I802_DEBUG_INC(local->tx_status_drop); @@ -1225,7 +1202,6 @@ static void ieee80211_tasklet_handler(unsigned long data) struct ieee80211_local *local = (struct ieee80211_local *) data; struct sk_buff *skb; struct ieee80211_rx_status rx_status; - struct ieee80211_tx_status *tx_status; struct ieee80211_ra_tid *ra_tid; while ((skb = skb_dequeue(&local->skb_queue)) || @@ -1240,12 +1216,8 @@ static void ieee80211_tasklet_handler(unsigned long data) __ieee80211_rx(local_to_hw(local), skb, &rx_status); break; case IEEE80211_TX_STATUS_MSG: - /* get pointer to saved status out of skb->cb */ - memcpy(&tx_status, skb->cb, sizeof(tx_status)); skb->pkt_type = 0; - ieee80211_tx_status(local_to_hw(local), - skb, tx_status); - kfree(tx_status); + ieee80211_tx_status(local_to_hw(local), skb); break; case IEEE80211_DELBA_MSG: ra_tid = (struct ieee80211_ra_tid *) &skb->cb; @@ -1274,24 +1246,15 @@ static void ieee80211_tasklet_handler(unsigned long data) * Also, tx_packet_data in cb is restored from tx_control. */ static void ieee80211_remove_tx_extra(struct ieee80211_local *local, struct ieee80211_key *key, - struct sk_buff *skb, - struct ieee80211_tx_control *control) + struct sk_buff *skb) { int hdrlen, iv_len, mic_len; - struct ieee80211_tx_packet_data *pkt_data; - - pkt_data = (struct ieee80211_tx_packet_data *)skb->cb; - pkt_data->ifindex = vif_to_sdata(control->vif)->dev->ifindex; - pkt_data->flags = 0; - if (control->flags & IEEE80211_TXCTL_REQ_TX_STATUS) - pkt_data->flags |= IEEE80211_TXPD_REQ_TX_STATUS; - if (control->flags & IEEE80211_TXCTL_DO_NOT_ENCRYPT) - pkt_data->flags |= IEEE80211_TXPD_DO_NOT_ENCRYPT; - if (control->flags & IEEE80211_TXCTL_REQUEUE) - pkt_data->flags |= IEEE80211_TXPD_REQUEUE; - if (control->flags & IEEE80211_TXCTL_EAPOL_FRAME) - pkt_data->flags |= IEEE80211_TXPD_EAPOL_FRAME; - pkt_data->queue = control->queue; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); + + info->flags &= IEEE80211_TX_CTL_REQ_TX_STATUS | + IEEE80211_TX_CTL_DO_NOT_ENCRYPT | + IEEE80211_TX_CTL_REQUEUE | + IEEE80211_TX_CTL_EAPOL_FRAME; hdrlen = ieee80211_get_hdrlen_from_skb(skb); @@ -1338,9 +1301,10 @@ no_key: static void ieee80211_handle_filtered_frame(struct ieee80211_local *local, struct sta_info *sta, - struct sk_buff *skb, - struct ieee80211_tx_status *status) + struct sk_buff *skb) { + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); + sta->tx_filtered_count++; /* @@ -1382,18 +1346,16 @@ static void ieee80211_handle_filtered_frame(struct ieee80211_local *local, */ if (test_sta_flags(sta, WLAN_STA_PS) && skb_queue_len(&sta->tx_filtered) < STA_MAX_TX_BUFFER) { - ieee80211_remove_tx_extra(local, sta->key, skb, - &status->control); + ieee80211_remove_tx_extra(local, sta->key, skb); skb_queue_tail(&sta->tx_filtered, skb); return; } if (!test_sta_flags(sta, WLAN_STA_PS) && - !(status->control.flags & IEEE80211_TXCTL_REQUEUE)) { + !(info->flags & IEEE80211_TX_CTL_REQUEUE)) { /* Software retry the packet once */ - status->control.flags |= IEEE80211_TXCTL_REQUEUE; - ieee80211_remove_tx_extra(local, sta->key, skb, - &status->control); + info->flags |= IEEE80211_TX_CTL_REQUEUE; + ieee80211_remove_tx_extra(local, sta->key, skb); dev_queue_xmit(skb); return; } @@ -1407,28 +1369,20 @@ static void ieee80211_handle_filtered_frame(struct ieee80211_local *local, dev_kfree_skb(skb); } -void ieee80211_tx_status(struct ieee80211_hw *hw, struct sk_buff *skb, - struct ieee80211_tx_status *status) +void ieee80211_tx_status(struct ieee80211_hw *hw, struct sk_buff *skb) { struct sk_buff *skb2; struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; struct ieee80211_local *local = hw_to_local(hw); + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); u16 frag, type; struct ieee80211_tx_status_rtap_hdr *rthdr; struct ieee80211_sub_if_data *sdata; struct net_device *prev_dev = NULL; - if (!status) { - printk(KERN_ERR - "%s: ieee80211_tx_status called with NULL status\n", - wiphy_name(local->hw.wiphy)); - dev_kfree_skb(skb); - return; - } - rcu_read_lock(); - if (status->excessive_retries) { + if (info->status.excessive_retries) { struct sta_info *sta; sta = sta_info_get(local, hdr->addr1); if (sta) { @@ -1437,27 +1391,23 @@ void ieee80211_tx_status(struct ieee80211_hw *hw, struct sk_buff *skb, * The STA is in power save mode, so assume * that this TX packet failed because of that. */ - status->excessive_retries = 0; - status->flags |= IEEE80211_TX_STATUS_TX_FILTERED; - ieee80211_handle_filtered_frame(local, sta, - skb, status); + ieee80211_handle_filtered_frame(local, sta, skb); rcu_read_unlock(); return; } } } - if (status->flags & IEEE80211_TX_STATUS_TX_FILTERED) { + if (info->flags & IEEE80211_TX_STAT_TX_FILTERED) { struct sta_info *sta; sta = sta_info_get(local, hdr->addr1); if (sta) { - ieee80211_handle_filtered_frame(local, sta, skb, - status); + ieee80211_handle_filtered_frame(local, sta, skb); rcu_read_unlock(); return; } } else - rate_control_tx_status(local->mdev, skb, status); + rate_control_tx_status(local->mdev, skb); rcu_read_unlock(); @@ -1471,14 +1421,14 @@ void ieee80211_tx_status(struct ieee80211_hw *hw, struct sk_buff *skb, frag = le16_to_cpu(hdr->seq_ctrl) & IEEE80211_SCTL_FRAG; type = le16_to_cpu(hdr->frame_control) & IEEE80211_FCTL_FTYPE; - if (status->flags & IEEE80211_TX_STATUS_ACK) { + if (info->flags & IEEE80211_TX_STAT_ACK) { if (frag == 0) { local->dot11TransmittedFrameCount++; if (is_multicast_ether_addr(hdr->addr1)) local->dot11MulticastTransmittedFrameCount++; - if (status->retry_count > 0) + if (info->status.retry_count > 0) local->dot11RetryCount++; - if (status->retry_count > 1) + if (info->status.retry_count > 1) local->dot11MultipleRetryCount++; } @@ -1524,17 +1474,17 @@ void ieee80211_tx_status(struct ieee80211_hw *hw, struct sk_buff *skb, cpu_to_le32((1 << IEEE80211_RADIOTAP_TX_FLAGS) | (1 << IEEE80211_RADIOTAP_DATA_RETRIES)); - if (!(status->flags & IEEE80211_TX_STATUS_ACK) && + if (!(info->flags & IEEE80211_TX_STAT_ACK) && !is_multicast_ether_addr(hdr->addr1)) rthdr->tx_flags |= cpu_to_le16(IEEE80211_RADIOTAP_F_TX_FAIL); - if ((status->control.flags & IEEE80211_TXCTL_USE_RTS_CTS) && - (status->control.flags & IEEE80211_TXCTL_USE_CTS_PROTECT)) + if ((info->flags & IEEE80211_TX_CTL_USE_RTS_CTS) && + (info->flags & IEEE80211_TX_CTL_USE_CTS_PROTECT)) rthdr->tx_flags |= cpu_to_le16(IEEE80211_RADIOTAP_F_TX_CTS); - else if (status->control.flags & IEEE80211_TXCTL_USE_RTS_CTS) + else if (info->flags & IEEE80211_TX_CTL_USE_RTS_CTS) rthdr->tx_flags |= cpu_to_le16(IEEE80211_RADIOTAP_F_TX_RTS); - rthdr->data_retries = status->retry_count; + rthdr->data_retries = info->status.retry_count; /* XXX: is this sufficient for BPF? */ skb_set_mac_header(skb, 0); @@ -1895,7 +1845,9 @@ static int __init ieee80211_init(void) struct sk_buff *skb; int ret; - BUILD_BUG_ON(sizeof(struct ieee80211_tx_packet_data) > sizeof(skb->cb)); + BUILD_BUG_ON(sizeof(struct ieee80211_tx_info) > sizeof(skb->cb)); + BUILD_BUG_ON(offsetof(struct ieee80211_tx_info, driver_data) + + IEEE80211_TX_INFO_DRIVER_DATA_SIZE > sizeof(skb->cb)); ret = rc80211_pid_init(); if (ret) diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 604149369dc9..9a264379d7b1 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -578,7 +578,7 @@ void ieee80211_sta_tx(struct net_device *dev, struct sk_buff *skb, int encrypt) { struct ieee80211_sub_if_data *sdata; - struct ieee80211_tx_packet_data *pkt_data; + struct ieee80211_tx_info *info; sdata = IEEE80211_DEV_TO_SUB_IF(dev); skb->dev = sdata->local->mdev; @@ -586,11 +586,11 @@ void ieee80211_sta_tx(struct net_device *dev, struct sk_buff *skb, skb_set_network_header(skb, 0); skb_set_transport_header(skb, 0); - pkt_data = (struct ieee80211_tx_packet_data *) skb->cb; - memset(pkt_data, 0, sizeof(struct ieee80211_tx_packet_data)); - pkt_data->ifindex = sdata->dev->ifindex; + info = IEEE80211_SKB_CB(skb); + memset(info, 0, sizeof(struct ieee80211_tx_info)); + info->control.ifindex = sdata->dev->ifindex; if (!encrypt) - pkt_data->flags |= IEEE80211_TXPD_DO_NOT_ENCRYPT; + info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; dev_queue_xmit(skb); } @@ -2314,7 +2314,7 @@ static int ieee80211_sta_join_ibss(struct net_device *dev, int res, rates, i, j; struct sk_buff *skb; struct ieee80211_mgmt *mgmt; - struct ieee80211_tx_control control; + struct ieee80211_tx_info *control; struct rate_selection ratesel; u8 *pos; struct ieee80211_sub_if_data *sdata; @@ -2404,21 +2404,22 @@ static int ieee80211_sta_join_ibss(struct net_device *dev, memcpy(pos, &bss->supp_rates[8], rates); } - memset(&control, 0, sizeof(control)); + control = IEEE80211_SKB_CB(skb); + rate_control_get_rate(dev, sband, skb, &ratesel); if (ratesel.rate_idx < 0) { printk(KERN_DEBUG "%s: Failed to determine TX rate " "for IBSS beacon\n", dev->name); break; } - control.vif = &sdata->vif; - control.tx_rate_idx = ratesel.rate_idx; + control->control.vif = &sdata->vif; + control->tx_rate_idx = ratesel.rate_idx; if (sdata->bss_conf.use_short_preamble && sband->bitrates[ratesel.rate_idx].flags & IEEE80211_RATE_SHORT_PREAMBLE) - control.flags |= IEEE80211_TXCTL_SHORT_PREAMBLE; - control.antenna_sel_tx = local->hw.conf.antenna_sel_tx; - control.flags |= IEEE80211_TXCTL_NO_ACK; - control.retry_limit = 1; + control->flags |= IEEE80211_TX_CTL_SHORT_PREAMBLE; + control->antenna_sel_tx = local->hw.conf.antenna_sel_tx; + control->flags |= IEEE80211_TX_CTL_NO_ACK; + control->control.retry_limit = 1; ifsta->probe_resp = skb_copy(skb, GFP_ATOMIC); if (ifsta->probe_resp) { @@ -2433,8 +2434,7 @@ static int ieee80211_sta_join_ibss(struct net_device *dev, } if (local->ops->beacon_update && - local->ops->beacon_update(local_to_hw(local), - skb, &control) == 0) { + local->ops->beacon_update(local_to_hw(local), skb) == 0) { printk(KERN_DEBUG "%s: Configured IBSS beacon " "template\n", dev->name); skb = NULL; diff --git a/net/mac80211/rate.h b/net/mac80211/rate.h index a29148dcca99..0ed9c8a2f56f 100644 --- a/net/mac80211/rate.h +++ b/net/mac80211/rate.h @@ -34,8 +34,7 @@ struct rate_control_ops { struct module *module; const char *name; void (*tx_status)(void *priv, struct net_device *dev, - struct sk_buff *skb, - struct ieee80211_tx_status *status); + struct sk_buff *skb); void (*get_rate)(void *priv, struct net_device *dev, struct ieee80211_supported_band *band, struct sk_buff *skb, @@ -77,13 +76,12 @@ struct rate_control_ref *rate_control_get(struct rate_control_ref *ref); void rate_control_put(struct rate_control_ref *ref); static inline void rate_control_tx_status(struct net_device *dev, - struct sk_buff *skb, - struct ieee80211_tx_status *status) + struct sk_buff *skb) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct rate_control_ref *ref = local->rate_ctrl; - ref->ops->tx_status(ref->priv, dev, skb, status); + ref->ops->tx_status(ref->priv, dev, skb); } diff --git a/net/mac80211/rc80211_pid.h b/net/mac80211/rc80211_pid.h index 04afc13ed825..2078803d3581 100644 --- a/net/mac80211/rc80211_pid.h +++ b/net/mac80211/rc80211_pid.h @@ -61,7 +61,7 @@ enum rc_pid_event_type { union rc_pid_event_data { /* RC_PID_EVENT_TX_STATUS */ struct { - struct ieee80211_tx_status tx_status; + struct ieee80211_tx_info tx_status; }; /* RC_PID_EVENT_TYPE_RATE_CHANGE */ /* RC_PID_EVENT_TYPE_TX_RATE */ @@ -158,7 +158,7 @@ struct rc_pid_debugfs_entries { }; void rate_control_pid_event_tx_status(struct rc_pid_event_buffer *buf, - struct ieee80211_tx_status *stat); + struct ieee80211_tx_info *stat); void rate_control_pid_event_rate_change(struct rc_pid_event_buffer *buf, int index, int rate); diff --git a/net/mac80211/rc80211_pid_algo.c b/net/mac80211/rc80211_pid_algo.c index 14cde36f5070..e8945413e4a2 100644 --- a/net/mac80211/rc80211_pid_algo.c +++ b/net/mac80211/rc80211_pid_algo.c @@ -237,8 +237,7 @@ static void rate_control_pid_sample(struct rc_pid_info *pinfo, } static void rate_control_pid_tx_status(void *priv, struct net_device *dev, - struct sk_buff *skb, - struct ieee80211_tx_status *status) + struct sk_buff *skb) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; @@ -248,6 +247,7 @@ static void rate_control_pid_tx_status(void *priv, struct net_device *dev, struct rc_pid_sta_info *spinfo; unsigned long period; struct ieee80211_supported_band *sband; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); rcu_read_lock(); @@ -266,28 +266,28 @@ static void rate_control_pid_tx_status(void *priv, struct net_device *dev, /* Ignore all frames that were sent with a different rate than the rate * we currently advise mac80211 to use. */ - if (status->control.tx_rate_idx != sta->txrate_idx) + if (info->tx_rate_idx != sta->txrate_idx) goto unlock; spinfo = sta->rate_ctrl_priv; spinfo->tx_num_xmit++; #ifdef CONFIG_MAC80211_DEBUGFS - rate_control_pid_event_tx_status(&spinfo->events, status); + rate_control_pid_event_tx_status(&spinfo->events, info); #endif /* We count frames that totally failed to be transmitted as two bad * frames, those that made it out but had some retries as one good and * one bad frame. */ - if (status->excessive_retries) { + if (info->status.excessive_retries) { spinfo->tx_num_failed += 2; spinfo->tx_num_xmit++; - } else if (status->retry_count) { + } else if (info->status.retry_count) { spinfo->tx_num_failed++; spinfo->tx_num_xmit++; } - if (status->excessive_retries) { + if (info->status.excessive_retries) { sta->tx_retry_failed++; sta->tx_num_consecutive_failures++; sta->tx_num_mpdu_fail++; @@ -295,8 +295,8 @@ static void rate_control_pid_tx_status(void *priv, struct net_device *dev, sta->tx_num_consecutive_failures = 0; sta->tx_num_mpdu_ok++; } - sta->tx_retry_count += status->retry_count; - sta->tx_num_mpdu_fail += status->retry_count; + sta->tx_retry_count += info->status.retry_count; + sta->tx_num_mpdu_fail += info->status.retry_count; /* Update PID controller state. */ period = (HZ * pinfo->sampling_period + 500) / 1000; diff --git a/net/mac80211/rc80211_pid_debugfs.c b/net/mac80211/rc80211_pid_debugfs.c index ff5c380f3c13..8121d3bc6835 100644 --- a/net/mac80211/rc80211_pid_debugfs.c +++ b/net/mac80211/rc80211_pid_debugfs.c @@ -39,11 +39,11 @@ static void rate_control_pid_event(struct rc_pid_event_buffer *buf, } void rate_control_pid_event_tx_status(struct rc_pid_event_buffer *buf, - struct ieee80211_tx_status *stat) + struct ieee80211_tx_info *stat) { union rc_pid_event_data evd; - memcpy(&evd.tx_status, stat, sizeof(struct ieee80211_tx_status)); + memcpy(&evd.tx_status, stat, sizeof(struct ieee80211_tx_info)); rate_control_pid_event(buf, RC_PID_EVENT_TYPE_TX_STATUS, &evd); } @@ -167,8 +167,8 @@ static ssize_t rate_control_pid_events_read(struct file *file, char __user *buf, switch (ev->type) { case RC_PID_EVENT_TYPE_TX_STATUS: p += snprintf(pb + p, length - p, "tx_status %u %u", - ev->data.tx_status.excessive_retries, - ev->data.tx_status.retry_count); + ev->data.tx_status.status.excessive_retries, + ev->data.tx_status.status.retry_count); break; case RC_PID_EVENT_TYPE_RATE_CHANGE: p += snprintf(pb + p, length - p, "rate_change %d %d", diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index fa68305fd59e..cf0de3b0fe24 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -714,7 +714,7 @@ static int ap_sta_ps_end(struct net_device *dev, struct sta_info *sta) struct sk_buff *skb; int sent = 0; struct ieee80211_sub_if_data *sdata; - struct ieee80211_tx_packet_data *pkt_data; + struct ieee80211_tx_info *info; DECLARE_MAC_BUF(mac); sdata = sta->sdata; @@ -734,13 +734,13 @@ static int ap_sta_ps_end(struct net_device *dev, struct sta_info *sta) /* Send all buffered frames to the station */ while ((skb = skb_dequeue(&sta->tx_filtered)) != NULL) { - pkt_data = (struct ieee80211_tx_packet_data *) skb->cb; + info = IEEE80211_SKB_CB(skb); sent++; - pkt_data->flags |= IEEE80211_TXPD_REQUEUE; + info->flags |= IEEE80211_TX_CTL_REQUEUE; dev_queue_xmit(skb); } while ((skb = skb_dequeue(&sta->ps_tx_buf)) != NULL) { - pkt_data = (struct ieee80211_tx_packet_data *) skb->cb; + info = IEEE80211_SKB_CB(skb); local->total_ps_buffered--; sent++; #ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG @@ -748,7 +748,7 @@ static int ap_sta_ps_end(struct net_device *dev, struct sta_info *sta) "since STA not sleeping anymore\n", dev->name, print_mac(mac, sta->addr), sta->aid); #endif /* CONFIG_MAC80211_VERBOSE_PS_DEBUG */ - pkt_data->flags |= IEEE80211_TXPD_REQUEUE; + info->flags |= IEEE80211_TX_CTL_REQUEUE; dev_queue_xmit(skb); } diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index baf5e4746884..ef3149324d54 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -511,20 +511,20 @@ static inline int sta_info_buffer_expired(struct ieee80211_local *local, struct sta_info *sta, struct sk_buff *skb) { - struct ieee80211_tx_packet_data *pkt_data; + struct ieee80211_tx_info *info; int timeout; if (!skb) return 0; - pkt_data = (struct ieee80211_tx_packet_data *) skb->cb; + info = IEEE80211_SKB_CB(skb); /* Timeout: (2 * listen_interval * beacon_int * 1024 / 1000000) sec */ timeout = (sta->listen_interval * local->hw.conf.beacon_int * 32 / 15625) * HZ; if (timeout < STA_TX_BUFFER_EXPIRE) timeout = STA_TX_BUFFER_EXPIRE; - return time_after(jiffies, pkt_data->jiffies + timeout); + return time_after(jiffies, info->control.jiffies + timeout); } diff --git a/net/mac80211/sta_info.h b/net/mac80211/sta_info.h index e89cc1655547..f592290e42b9 100644 --- a/net/mac80211/sta_info.h +++ b/net/mac80211/sta_info.h @@ -32,7 +32,7 @@ * @WLAN_STA_WDS: Station is one of our WDS peers. * @WLAN_STA_PSPOLL: Station has just PS-polled us. * @WLAN_STA_CLEAR_PS_FILT: Clear PS filter in hardware (using the - * IEEE80211_TXCTL_CLEAR_PS_FILT control flag) when the next + * IEEE80211_TX_CTL_CLEAR_PS_FILT control flag) when the next * frame to this station is transmitted. */ enum ieee80211_sta_info_flags { diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 666158f02a89..ac9a4af7ad42 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -238,12 +238,12 @@ static ieee80211_tx_result ieee80211_tx_h_check_assoc(struct ieee80211_tx_data *tx) { #ifdef CONFIG_MAC80211_VERBOSE_DEBUG - struct sk_buff *skb = tx->skb; - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; #endif /* CONFIG_MAC80211_VERBOSE_DEBUG */ + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb); u32 sta_flags; - if (unlikely(tx->flags & IEEE80211_TX_INJECTED)) + if (unlikely(info->flags & IEEE80211_TX_CTL_INJECTED)) return TX_CONTINUE; if (unlikely(tx->local->sta_sw_scanning) && @@ -348,6 +348,8 @@ static void purge_old_ps_buffers(struct ieee80211_local *local) static ieee80211_tx_result ieee80211_tx_h_multicast_ps_buf(struct ieee80211_tx_data *tx) { + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb); + /* * broadcast/multicast frame * @@ -383,7 +385,7 @@ ieee80211_tx_h_multicast_ps_buf(struct ieee80211_tx_data *tx) } /* buffered in hardware */ - tx->control->flags |= IEEE80211_TXCTL_SEND_AFTER_DTIM; + info->flags |= IEEE80211_TX_CTL_SEND_AFTER_DTIM; return TX_CONTINUE; } @@ -392,6 +394,7 @@ static ieee80211_tx_result ieee80211_tx_h_unicast_ps_buf(struct ieee80211_tx_data *tx) { struct sta_info *sta = tx->sta; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb); u32 staflags; DECLARE_MAC_BUF(mac); @@ -404,7 +407,6 @@ ieee80211_tx_h_unicast_ps_buf(struct ieee80211_tx_data *tx) if (unlikely((staflags & WLAN_STA_PS) && !(staflags & WLAN_STA_PSPOLL))) { - struct ieee80211_tx_packet_data *pkt_data; #ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG printk(KERN_DEBUG "STA %s aid %d: PS buffer (entries " "before %d)\n", @@ -428,8 +430,7 @@ ieee80211_tx_h_unicast_ps_buf(struct ieee80211_tx_data *tx) if (skb_queue_empty(&sta->ps_tx_buf)) sta_info_set_tim_bit(sta); - pkt_data = (struct ieee80211_tx_packet_data *)tx->skb->cb; - pkt_data->jiffies = jiffies; + info->control.jiffies = jiffies; skb_queue_tail(&sta->ps_tx_buf, tx->skb); return TX_QUEUED; } @@ -461,17 +462,18 @@ static ieee80211_tx_result ieee80211_tx_h_select_key(struct ieee80211_tx_data *tx) { struct ieee80211_key *key; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb); u16 fc = tx->fc; - if (unlikely(tx->control->flags & IEEE80211_TXCTL_DO_NOT_ENCRYPT)) + if (unlikely(info->flags & IEEE80211_TX_CTL_DO_NOT_ENCRYPT)) tx->key = NULL; else if (tx->sta && (key = rcu_dereference(tx->sta->key))) tx->key = key; else if ((key = rcu_dereference(tx->sdata->default_key))) tx->key = key; else if (tx->sdata->drop_unencrypted && - !(tx->control->flags & IEEE80211_TXCTL_EAPOL_FRAME) && - !(tx->flags & IEEE80211_TX_INJECTED)) { + !(info->flags & IEEE80211_TX_CTL_EAPOL_FRAME) && + !(info->flags & IEEE80211_TX_CTL_INJECTED)) { I802_DEBUG_INC(tx->local->tx_handlers_drop_unencrypted); return TX_DROP; } else @@ -500,7 +502,7 @@ ieee80211_tx_h_select_key(struct ieee80211_tx_data *tx) } if (!tx->key || !(tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE)) - tx->control->flags |= IEEE80211_TXCTL_DO_NOT_ENCRYPT; + info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; return TX_CONTINUE; } @@ -510,6 +512,7 @@ ieee80211_tx_h_rate_ctrl(struct ieee80211_tx_data *tx) { struct rate_selection rsel; struct ieee80211_supported_band *sband; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb); sband = tx->local->hw.wiphy->bands[tx->channel->band]; @@ -517,18 +520,17 @@ ieee80211_tx_h_rate_ctrl(struct ieee80211_tx_data *tx) rate_control_get_rate(tx->dev, sband, tx->skb, &rsel); tx->rate_idx = rsel.rate_idx; if (unlikely(rsel.probe_idx >= 0)) { - tx->control->flags |= - IEEE80211_TXCTL_RATE_CTRL_PROBE; + info->flags |= IEEE80211_TX_CTL_RATE_CTRL_PROBE; tx->flags |= IEEE80211_TX_PROBE_LAST_FRAG; - tx->control->alt_retry_rate_idx = tx->rate_idx; + info->control.alt_retry_rate_idx = tx->rate_idx; tx->rate_idx = rsel.probe_idx; } else - tx->control->alt_retry_rate_idx = -1; + info->control.alt_retry_rate_idx = -1; if (unlikely(tx->rate_idx < 0)) return TX_DROP; } else - tx->control->alt_retry_rate_idx = -1; + info->control.alt_retry_rate_idx = -1; if (tx->sdata->bss_conf.use_cts_prot && (tx->flags & IEEE80211_TX_FRAGMENTED) && (rsel.nonerp_idx >= 0)) { @@ -538,13 +540,13 @@ ieee80211_tx_h_rate_ctrl(struct ieee80211_tx_data *tx) else tx->flags |= IEEE80211_TX_PROBE_LAST_FRAG; tx->rate_idx = rsel.nonerp_idx; - tx->control->tx_rate_idx = rsel.nonerp_idx; - tx->control->flags &= ~IEEE80211_TXCTL_RATE_CTRL_PROBE; + info->tx_rate_idx = rsel.nonerp_idx; + info->flags &= ~IEEE80211_TX_CTL_RATE_CTRL_PROBE; } else { tx->last_frag_rate_idx = tx->rate_idx; - tx->control->tx_rate_idx = tx->rate_idx; + info->tx_rate_idx = tx->rate_idx; } - tx->control->tx_rate_idx = tx->rate_idx; + info->tx_rate_idx = tx->rate_idx; return TX_CONTINUE; } @@ -555,28 +557,32 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) tx->skb->data; u16 fc = le16_to_cpu(hdr->frame_control); u16 dur; - struct ieee80211_tx_control *control = tx->control; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb); struct ieee80211_supported_band *sband; sband = tx->local->hw.wiphy->bands[tx->channel->band]; - if (!control->retry_limit) { + if (tx->sta) + info->control.aid = tx->sta->aid; + + if (!info->control.retry_limit) { if (!is_multicast_ether_addr(hdr->addr1)) { - if (tx->skb->len + FCS_LEN > tx->local->rts_threshold + int len = min_t(int, tx->skb->len + FCS_LEN, + tx->local->fragmentation_threshold); + if (len > tx->local->rts_threshold && tx->local->rts_threshold < - IEEE80211_MAX_RTS_THRESHOLD) { - control->flags |= - IEEE80211_TXCTL_USE_RTS_CTS; - control->flags |= - IEEE80211_TXCTL_LONG_RETRY_LIMIT; - control->retry_limit = + IEEE80211_MAX_RTS_THRESHOLD) { + info->flags |= IEEE80211_TX_CTL_USE_RTS_CTS; + info->flags |= + IEEE80211_TX_CTL_LONG_RETRY_LIMIT; + info->control.retry_limit = tx->local->long_retry_limit; } else { - control->retry_limit = + info->control.retry_limit = tx->local->short_retry_limit; } } else { - control->retry_limit = 1; + info->control.retry_limit = 1; } } @@ -585,7 +591,7 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) * frames. * TODO: The last fragment could still use multiple retry * rates. */ - control->alt_retry_rate_idx = -1; + info->control.alt_retry_rate_idx = -1; } /* Use CTS protection for unicast frames sent using extended rates if @@ -595,8 +601,8 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) (sband->bitrates[tx->rate_idx].flags & IEEE80211_RATE_ERP_G) && (tx->flags & IEEE80211_TX_UNICAST) && tx->sdata->bss_conf.use_cts_prot && - !(control->flags & IEEE80211_TXCTL_USE_RTS_CTS)) - control->flags |= IEEE80211_TXCTL_USE_CTS_PROTECT; + !(info->flags & IEEE80211_TX_CTL_USE_RTS_CTS)) + info->flags |= IEEE80211_TX_CTL_USE_CTS_PROTECT; /* Transmit data frames using short preambles if the driver supports * short preambles at the selected rate and short preambles are @@ -605,7 +611,7 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) (sband->bitrates[tx->rate_idx].flags & IEEE80211_RATE_SHORT_PREAMBLE) && tx->sdata->bss_conf.use_short_preamble && (!tx->sta || test_sta_flags(tx->sta, WLAN_STA_SHORT_PREAMBLE))) { - tx->control->flags |= IEEE80211_TXCTL_SHORT_PREAMBLE; + info->flags |= IEEE80211_TX_CTL_SHORT_PREAMBLE; } /* Setup duration field for the first fragment of the frame. Duration @@ -616,8 +622,8 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) tx->extra_frag[0]->len : 0); hdr->duration_id = cpu_to_le16(dur); - if ((control->flags & IEEE80211_TXCTL_USE_RTS_CTS) || - (control->flags & IEEE80211_TXCTL_USE_CTS_PROTECT)) { + if ((info->flags & IEEE80211_TX_CTL_USE_RTS_CTS) || + (info->flags & IEEE80211_TX_CTL_USE_CTS_PROTECT)) { struct ieee80211_supported_band *sband; struct ieee80211_rate *rate; s8 baserate = -1; @@ -626,7 +632,7 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) sband = tx->local->hw.wiphy->bands[tx->channel->band]; /* Do not use multiple retry rates when using RTS/CTS */ - control->alt_retry_rate_idx = -1; + info->control.alt_retry_rate_idx = -1; /* Use min(data rate, max base rate) as CTS/RTS rate */ rate = &sband->bitrates[tx->rate_idx]; @@ -642,13 +648,13 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) } if (baserate >= 0) - control->rts_cts_rate_idx = baserate; + info->control.rts_cts_rate_idx = baserate; else - control->rts_cts_rate_idx = 0; + info->control.rts_cts_rate_idx = 0; } if (tx->sta) - control->aid = tx->sta->aid; + info->control.aid = tx->sta->aid; return TX_CONTINUE; } @@ -762,6 +768,7 @@ ieee80211_tx_h_stats(struct ieee80211_tx_data *tx) u32 load = 0, hdrtime; struct ieee80211_rate *rate; struct ieee80211_supported_band *sband; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); sband = tx->local->hw.wiphy->bands[tx->channel->band]; rate = &sband->bitrates[tx->rate_idx]; @@ -786,9 +793,9 @@ ieee80211_tx_h_stats(struct ieee80211_tx_data *tx) if (!is_multicast_ether_addr(hdr->addr1)) load += hdrtime; - if (tx->control->flags & IEEE80211_TXCTL_USE_RTS_CTS) + if (info->flags & IEEE80211_TX_CTL_USE_RTS_CTS) load += 2 * hdrtime; - else if (tx->control->flags & IEEE80211_TXCTL_USE_CTS_PROTECT) + else if (info->flags & IEEE80211_TX_CTL_USE_CTS_PROTECT) load += hdrtime; /* TODO: optimise again */ @@ -839,6 +846,7 @@ static ieee80211_tx_handler ieee80211_tx_handlers[] = ieee80211_tx_h_rate_ctrl, ieee80211_tx_h_misc, ieee80211_tx_h_fragment, + /* handlers after fragment must be aware of tx info fragmentation! */ ieee80211_tx_h_encrypt, ieee80211_tx_h_stats, NULL @@ -867,12 +875,12 @@ __ieee80211_parse_tx_radiotap(struct ieee80211_tx_data *tx, (struct ieee80211_radiotap_header *) skb->data; struct ieee80211_supported_band *sband; int ret = ieee80211_radiotap_iterator_init(&iterator, rthdr, skb->len); - struct ieee80211_tx_control *control = tx->control; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); sband = tx->local->hw.wiphy->bands[tx->channel->band]; - control->flags |= IEEE80211_TXCTL_DO_NOT_ENCRYPT; - tx->flags |= IEEE80211_TX_INJECTED; + info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; + info->flags |= IEEE80211_TX_CTL_INJECTED; tx->flags &= ~IEEE80211_TX_FRAGMENTED; /* @@ -920,7 +928,7 @@ __ieee80211_parse_tx_radiotap(struct ieee80211_tx_data *tx, * radiotap uses 0 for 1st ant, mac80211 is 1 for * 1st ant */ - control->antenna_sel_tx = (*iterator.this_arg) + 1; + info->antenna_sel_tx = (*iterator.this_arg) + 1; break; #if 0 @@ -944,8 +952,8 @@ __ieee80211_parse_tx_radiotap(struct ieee80211_tx_data *tx, skb_trim(skb, skb->len - FCS_LEN); } if (*iterator.this_arg & IEEE80211_RADIOTAP_F_WEP) - control->flags &= - ~IEEE80211_TXCTL_DO_NOT_ENCRYPT; + info->flags &= + ~IEEE80211_TX_CTL_DO_NOT_ENCRYPT; if (*iterator.this_arg & IEEE80211_RADIOTAP_F_FRAG) tx->flags |= IEEE80211_TX_FRAGMENTED; break; @@ -980,12 +988,12 @@ __ieee80211_parse_tx_radiotap(struct ieee80211_tx_data *tx, static ieee80211_tx_result __ieee80211_tx_prepare(struct ieee80211_tx_data *tx, struct sk_buff *skb, - struct net_device *dev, - struct ieee80211_tx_control *control) + struct net_device *dev) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct ieee80211_hdr *hdr; struct ieee80211_sub_if_data *sdata; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); int hdrlen; @@ -994,7 +1002,7 @@ __ieee80211_tx_prepare(struct ieee80211_tx_data *tx, tx->dev = dev; /* use original interface */ tx->local = local; tx->sdata = IEEE80211_DEV_TO_SUB_IF(dev); - tx->control = control; + tx->channel = local->hw.conf.channel; /* * Set this flag (used below to indicate "automatic fragmentation"), * it will be cleared/left by radiotap as desired. @@ -1021,10 +1029,10 @@ __ieee80211_tx_prepare(struct ieee80211_tx_data *tx, if (is_multicast_ether_addr(hdr->addr1)) { tx->flags &= ~IEEE80211_TX_UNICAST; - control->flags |= IEEE80211_TXCTL_NO_ACK; + info->flags |= IEEE80211_TX_CTL_NO_ACK; } else { tx->flags |= IEEE80211_TX_UNICAST; - control->flags &= ~IEEE80211_TXCTL_NO_ACK; + info->flags &= ~IEEE80211_TX_CTL_NO_ACK; } if (tx->flags & IEEE80211_TX_FRAGMENTED) { @@ -1037,16 +1045,16 @@ __ieee80211_tx_prepare(struct ieee80211_tx_data *tx, } if (!tx->sta) - control->flags |= IEEE80211_TXCTL_CLEAR_PS_FILT; + info->flags |= IEEE80211_TX_CTL_CLEAR_PS_FILT; else if (test_and_clear_sta_flags(tx->sta, WLAN_STA_CLEAR_PS_FILT)) - control->flags |= IEEE80211_TXCTL_CLEAR_PS_FILT; + info->flags |= IEEE80211_TX_CTL_CLEAR_PS_FILT; hdrlen = ieee80211_get_hdrlen(tx->fc); if (skb->len > hdrlen + sizeof(rfc1042_header) + 2) { u8 *pos = &skb->data[hdrlen + sizeof(rfc1042_header)]; tx->ethertype = (pos[0] << 8) | pos[1]; } - control->flags |= IEEE80211_TXCTL_FIRST_FRAGMENT; + info->flags |= IEEE80211_TX_CTL_FIRST_FRAGMENT; return TX_CONTINUE; } @@ -1056,14 +1064,12 @@ __ieee80211_tx_prepare(struct ieee80211_tx_data *tx, */ static int ieee80211_tx_prepare(struct ieee80211_tx_data *tx, struct sk_buff *skb, - struct net_device *mdev, - struct ieee80211_tx_control *control) + struct net_device *mdev) { - struct ieee80211_tx_packet_data *pkt_data; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); struct net_device *dev; - pkt_data = (struct ieee80211_tx_packet_data *)skb->cb; - dev = dev_get_by_index(&init_net, pkt_data->ifindex); + dev = dev_get_by_index(&init_net, info->control.ifindex); if (unlikely(dev && !is_ieee80211_device(dev, mdev))) { dev_put(dev); dev = NULL; @@ -1071,7 +1077,7 @@ static int ieee80211_tx_prepare(struct ieee80211_tx_data *tx, if (unlikely(!dev)) return -ENODEV; /* initialises tx with control */ - __ieee80211_tx_prepare(tx, skb, dev, control); + __ieee80211_tx_prepare(tx, skb, dev); dev_put(dev); return 0; } @@ -1079,7 +1085,7 @@ static int ieee80211_tx_prepare(struct ieee80211_tx_data *tx, static int __ieee80211_tx(struct ieee80211_local *local, struct sk_buff *skb, struct ieee80211_tx_data *tx) { - struct ieee80211_tx_control *control = tx->control; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); int ret, i; if (!ieee80211_qdisc_installed(local->mdev) && @@ -1090,39 +1096,39 @@ static int __ieee80211_tx(struct ieee80211_local *local, struct sk_buff *skb, if (skb) { ieee80211_dump_frame(wiphy_name(local->hw.wiphy), "TX to low-level driver", skb); - ret = local->ops->tx(local_to_hw(local), skb, control); + ret = local->ops->tx(local_to_hw(local), skb); if (ret) return IEEE80211_TX_AGAIN; local->mdev->trans_start = jiffies; ieee80211_led_tx(local, 1); } if (tx->extra_frag) { - control->flags &= ~(IEEE80211_TXCTL_USE_RTS_CTS | - IEEE80211_TXCTL_USE_CTS_PROTECT | - IEEE80211_TXCTL_CLEAR_PS_FILT | - IEEE80211_TXCTL_FIRST_FRAGMENT); for (i = 0; i < tx->num_extra_frag; i++) { if (!tx->extra_frag[i]) continue; - if (__ieee80211_queue_stopped(local, control->queue)) + info = IEEE80211_SKB_CB(tx->extra_frag[i]); + info->flags &= ~(IEEE80211_TX_CTL_USE_RTS_CTS | + IEEE80211_TX_CTL_USE_CTS_PROTECT | + IEEE80211_TX_CTL_CLEAR_PS_FILT | + IEEE80211_TX_CTL_FIRST_FRAGMENT); + if (__ieee80211_queue_stopped(local, info->queue)) return IEEE80211_TX_FRAG_AGAIN; if (i == tx->num_extra_frag) { - control->tx_rate_idx = tx->last_frag_rate_idx; + info->tx_rate_idx = tx->last_frag_rate_idx; if (tx->flags & IEEE80211_TX_PROBE_LAST_FRAG) - control->flags |= - IEEE80211_TXCTL_RATE_CTRL_PROBE; + info->flags |= + IEEE80211_TX_CTL_RATE_CTRL_PROBE; else - control->flags &= - ~IEEE80211_TXCTL_RATE_CTRL_PROBE; + info->flags &= + ~IEEE80211_TX_CTL_RATE_CTRL_PROBE; } ieee80211_dump_frame(wiphy_name(local->hw.wiphy), "TX to low-level driver", tx->extra_frag[i]); ret = local->ops->tx(local_to_hw(local), - tx->extra_frag[i], - control); + tx->extra_frag[i]); if (ret) return IEEE80211_TX_FRAG_AGAIN; local->mdev->trans_start = jiffies; @@ -1135,17 +1141,18 @@ static int __ieee80211_tx(struct ieee80211_local *local, struct sk_buff *skb, return IEEE80211_TX_OK; } -static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb, - struct ieee80211_tx_control *control) +static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct sta_info *sta; ieee80211_tx_handler *handler; struct ieee80211_tx_data tx; ieee80211_tx_result res = TX_DROP, res_prepare; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); int ret, i; + int queue = info->queue; - WARN_ON(__ieee80211_queue_pending(local, control->queue)); + WARN_ON(__ieee80211_queue_pending(local, queue)); if (unlikely(skb->len < 10)) { dev_kfree_skb(skb); @@ -1155,7 +1162,7 @@ static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb, rcu_read_lock(); /* initialises tx */ - res_prepare = __ieee80211_tx_prepare(&tx, skb, dev, control); + res_prepare = __ieee80211_tx_prepare(&tx, skb, dev); if (res_prepare == TX_DROP) { dev_kfree_skb(skb); @@ -1165,7 +1172,7 @@ static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb, sta = tx.sta; tx.channel = local->hw.conf.channel; - control->band = tx.channel->band; + info->band = tx.channel->band; for (handler = ieee80211_tx_handlers; *handler != NULL; handler++) { @@ -1174,7 +1181,8 @@ static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb, break; } - skb = tx.skb; /* handlers are allowed to change skb */ + if (WARN_ON(tx.skb != skb)) + goto drop; if (unlikely(res == TX_DROP)) { I802_DEBUG_INC(local->tx_handlers_drop); @@ -1209,12 +1217,12 @@ retry: ret = __ieee80211_tx(local, skb, &tx); if (ret) { struct ieee80211_tx_stored_packet *store = - &local->pending_packet[control->queue]; + &local->pending_packet[info->queue]; if (ret == IEEE80211_TX_FRAG_AGAIN) skb = NULL; set_bit(IEEE80211_LINK_STATE_PENDING, - &local->state[control->queue]); + &local->state[queue]); smp_mb(); /* When the driver gets out of buffers during sending of * fragments and calls ieee80211_stop_queue, there is @@ -1225,13 +1233,11 @@ retry: * called with IEEE80211_LINK_STATE_PENDING. Prevent this by * continuing transmitting here when that situation is * possible to have happened. */ - if (!__ieee80211_queue_stopped(local, control->queue)) { + if (!__ieee80211_queue_stopped(local, queue)) { clear_bit(IEEE80211_LINK_STATE_PENDING, - &local->state[control->queue]); + &local->state[queue]); goto retry; } - memcpy(&store->control, control, - sizeof(struct ieee80211_tx_control)); store->skb = skb; store->extra_frag = tx.extra_frag; store->num_extra_frag = tx.num_extra_frag; @@ -1258,21 +1264,14 @@ retry: int ieee80211_master_start_xmit(struct sk_buff *skb, struct net_device *dev) { - struct ieee80211_tx_control control; - struct ieee80211_tx_packet_data *pkt_data; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); struct net_device *odev = NULL; struct ieee80211_sub_if_data *osdata; int headroom; int ret; - /* - * copy control out of the skb so other people can use skb->cb - */ - pkt_data = (struct ieee80211_tx_packet_data *)skb->cb; - memset(&control, 0, sizeof(struct ieee80211_tx_control)); - - if (pkt_data->ifindex) - odev = dev_get_by_index(&init_net, pkt_data->ifindex); + if (info->control.ifindex) + odev = dev_get_by_index(&init_net, info->control.ifindex); if (unlikely(odev && !is_ieee80211_device(odev, dev))) { dev_put(odev); odev = NULL; @@ -1285,6 +1284,7 @@ int ieee80211_master_start_xmit(struct sk_buff *skb, dev_kfree_skb(skb); return 0; } + osdata = IEEE80211_DEV_TO_SUB_IF(odev); headroom = osdata->local->tx_headroom + IEEE80211_ENCRYPT_HEADROOM; @@ -1296,21 +1296,8 @@ int ieee80211_master_start_xmit(struct sk_buff *skb, } } - control.vif = &osdata->vif; - control.type = osdata->vif.type; - if (pkt_data->flags & IEEE80211_TXPD_REQ_TX_STATUS) - control.flags |= IEEE80211_TXCTL_REQ_TX_STATUS; - if (pkt_data->flags & IEEE80211_TXPD_DO_NOT_ENCRYPT) - control.flags |= IEEE80211_TXCTL_DO_NOT_ENCRYPT; - if (pkt_data->flags & IEEE80211_TXPD_REQUEUE) - control.flags |= IEEE80211_TXCTL_REQUEUE; - if (pkt_data->flags & IEEE80211_TXPD_EAPOL_FRAME) - control.flags |= IEEE80211_TXCTL_EAPOL_FRAME; - if (pkt_data->flags & IEEE80211_TXPD_AMPDU) - control.flags |= IEEE80211_TXCTL_AMPDU; - control.queue = pkt_data->queue; - - ret = ieee80211_tx(odev, skb, &control); + info->control.vif = &osdata->vif; + ret = ieee80211_tx(odev, skb); dev_put(odev); return ret; @@ -1320,7 +1307,7 @@ int ieee80211_monitor_start_xmit(struct sk_buff *skb, struct net_device *dev) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee80211_tx_packet_data *pkt_data; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); struct ieee80211_radiotap_header *prthdr = (struct ieee80211_radiotap_header *)skb->data; u16 len_rthdr; @@ -1342,14 +1329,12 @@ int ieee80211_monitor_start_xmit(struct sk_buff *skb, skb->dev = local->mdev; - pkt_data = (struct ieee80211_tx_packet_data *)skb->cb; - memset(pkt_data, 0, sizeof(*pkt_data)); /* needed because we set skb device to master */ - pkt_data->ifindex = dev->ifindex; + info->control.ifindex = dev->ifindex; - pkt_data->flags |= IEEE80211_TXPD_DO_NOT_ENCRYPT; + info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; /* Interfaces should always request a status report */ - pkt_data->flags |= IEEE80211_TXPD_REQ_TX_STATUS; + info->flags |= IEEE80211_TX_CTL_REQ_TX_STATUS; /* * fix up the pointers accounting for the radiotap @@ -1393,7 +1378,7 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, struct net_device *dev) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee80211_tx_packet_data *pkt_data; + struct ieee80211_tx_info *info; struct ieee80211_sub_if_data *sdata; int ret = 1, head_need; u16 ethertype, hdrlen, meshhdrlen = 0, fc; @@ -1625,14 +1610,14 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, nh_pos += hdrlen; h_pos += hdrlen; - pkt_data = (struct ieee80211_tx_packet_data *)skb->cb; - memset(pkt_data, 0, sizeof(struct ieee80211_tx_packet_data)); - pkt_data->ifindex = dev->ifindex; + info = IEEE80211_SKB_CB(skb); + memset(info, 0, sizeof(*info)); + info->control.ifindex = dev->ifindex; if (ethertype == ETH_P_PAE) - pkt_data->flags |= IEEE80211_TXPD_EAPOL_FRAME; + info->flags |= IEEE80211_TX_CTL_EAPOL_FRAME; /* Interfaces should always request a status report */ - pkt_data->flags |= IEEE80211_TXPD_REQ_TX_STATUS; + info->flags |= IEEE80211_TX_CTL_REQ_TX_STATUS; skb->dev = local->mdev; dev->stats.tx_packets++; @@ -1693,7 +1678,6 @@ void ieee80211_tx_pending(unsigned long data) continue; } store = &local->pending_packet[i]; - tx.control = &store->control; tx.extra_frag = store->extra_frag; tx.num_extra_frag = store->num_extra_frag; tx.last_frag_rate_idx = store->last_frag_rate_idx; @@ -1786,11 +1770,11 @@ static void ieee80211_beacon_add_tim(struct ieee80211_local *local, } struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, - struct ieee80211_vif *vif, - struct ieee80211_tx_control *control) + struct ieee80211_vif *vif) { struct ieee80211_local *local = hw_to_local(hw); struct sk_buff *skb; + struct ieee80211_tx_info *info; struct net_device *bdev; struct ieee80211_sub_if_data *sdata = NULL; struct ieee80211_if_ap *ap = NULL; @@ -1896,31 +1880,32 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, goto out; } - if (control) { - control->band = band; - rate_control_get_rate(local->mdev, sband, skb, &rsel); - if (unlikely(rsel.rate_idx < 0)) { - if (net_ratelimit()) { - printk(KERN_DEBUG "%s: ieee80211_beacon_get: " - "no rate found\n", - wiphy_name(local->hw.wiphy)); - } - dev_kfree_skb(skb); - skb = NULL; - goto out; - } + info = IEEE80211_SKB_CB(skb); + + info->band = band; + rate_control_get_rate(local->mdev, sband, skb, &rsel); - control->vif = vif; - control->tx_rate_idx = rsel.rate_idx; - if (sdata->bss_conf.use_short_preamble && - sband->bitrates[rsel.rate_idx].flags & IEEE80211_RATE_SHORT_PREAMBLE) - control->flags |= IEEE80211_TXCTL_SHORT_PREAMBLE; - control->antenna_sel_tx = local->hw.conf.antenna_sel_tx; - control->flags |= IEEE80211_TXCTL_NO_ACK; - control->flags |= IEEE80211_TXCTL_DO_NOT_ENCRYPT; - control->retry_limit = 1; - control->flags |= IEEE80211_TXCTL_CLEAR_PS_FILT; + if (unlikely(rsel.rate_idx < 0)) { + if (net_ratelimit()) { + printk(KERN_DEBUG "%s: ieee80211_beacon_get: " + "no rate found\n", + wiphy_name(local->hw.wiphy)); + } + dev_kfree_skb(skb); + skb = NULL; + goto out; } + + info->control.vif = vif; + info->tx_rate_idx = rsel.rate_idx; + if (sdata->bss_conf.use_short_preamble && + sband->bitrates[rsel.rate_idx].flags & IEEE80211_RATE_SHORT_PREAMBLE) + info->flags |= IEEE80211_TX_CTL_SHORT_PREAMBLE; + info->antenna_sel_tx = local->hw.conf.antenna_sel_tx; + info->flags |= IEEE80211_TX_CTL_NO_ACK; + info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; + info->control.retry_limit = 1; + info->flags |= IEEE80211_TX_CTL_CLEAR_PS_FILT; (*num_beacons)++; out: rcu_read_unlock(); @@ -1930,7 +1915,7 @@ EXPORT_SYMBOL(ieee80211_beacon_get); void ieee80211_rts_get(struct ieee80211_hw *hw, struct ieee80211_vif *vif, const void *frame, size_t frame_len, - const struct ieee80211_tx_control *frame_txctl, + const struct ieee80211_tx_info *frame_txctl, struct ieee80211_rts *rts) { const struct ieee80211_hdr *hdr = frame; @@ -1947,7 +1932,7 @@ EXPORT_SYMBOL(ieee80211_rts_get); void ieee80211_ctstoself_get(struct ieee80211_hw *hw, struct ieee80211_vif *vif, const void *frame, size_t frame_len, - const struct ieee80211_tx_control *frame_txctl, + const struct ieee80211_tx_info *frame_txctl, struct ieee80211_cts *cts) { const struct ieee80211_hdr *hdr = frame; @@ -1963,8 +1948,7 @@ EXPORT_SYMBOL(ieee80211_ctstoself_get); struct sk_buff * ieee80211_get_buffered_bc(struct ieee80211_hw *hw, - struct ieee80211_vif *vif, - struct ieee80211_tx_control *control) + struct ieee80211_vif *vif) { struct ieee80211_local *local = hw_to_local(hw); struct sk_buff *skb; @@ -1976,6 +1960,7 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw, struct ieee80211_sub_if_data *sdata; struct ieee80211_if_ap *bss = NULL; struct beacon_data *beacon; + struct ieee80211_tx_info *info; sdata = vif_to_sdata(vif); bdev = sdata->dev; @@ -1995,7 +1980,7 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw, if (bss->dtim_count != 0) return NULL; /* send buffered bc/mc only after DTIM beacon */ - memset(control, 0, sizeof(*control)); + while (1) { skb = skb_dequeue(&bss->ps_bc_buf); if (!skb) @@ -2012,21 +1997,26 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw, cpu_to_le16(IEEE80211_FCTL_MOREDATA); } - if (!ieee80211_tx_prepare(&tx, skb, local->mdev, control)) + if (!ieee80211_tx_prepare(&tx, skb, local->mdev)) break; dev_kfree_skb_any(skb); } + + info = IEEE80211_SKB_CB(skb); + sta = tx.sta; tx.flags |= IEEE80211_TX_PS_BUFFERED; tx.channel = local->hw.conf.channel; - control->band = tx.channel->band; + info->band = tx.channel->band; for (handler = ieee80211_tx_handlers; *handler != NULL; handler++) { res = (*handler)(&tx); if (res == TX_DROP || res == TX_QUEUED) break; } - skb = tx.skb; /* handlers are allowed to change skb */ + + if (WARN_ON(tx.skb != skb)) + return NULL; if (res == TX_DROP) { I802_DEBUG_INC(local->tx_handlers_drop); diff --git a/net/mac80211/util.c b/net/mac80211/util.c index 65a34fddeb00..d9109dee461f 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -258,7 +258,7 @@ EXPORT_SYMBOL(ieee80211_generic_frame_duration); __le16 ieee80211_rts_duration(struct ieee80211_hw *hw, struct ieee80211_vif *vif, size_t frame_len, - const struct ieee80211_tx_control *frame_txctl) + const struct ieee80211_tx_info *frame_txctl) { struct ieee80211_local *local = hw_to_local(hw); struct ieee80211_rate *rate; @@ -272,7 +272,7 @@ __le16 ieee80211_rts_duration(struct ieee80211_hw *hw, short_preamble = sdata->bss_conf.use_short_preamble; - rate = &sband->bitrates[frame_txctl->rts_cts_rate_idx]; + rate = &sband->bitrates[frame_txctl->control.rts_cts_rate_idx]; erp = 0; if (sdata->flags & IEEE80211_SDATA_OPERATING_GMODE) @@ -295,7 +295,7 @@ EXPORT_SYMBOL(ieee80211_rts_duration); __le16 ieee80211_ctstoself_duration(struct ieee80211_hw *hw, struct ieee80211_vif *vif, size_t frame_len, - const struct ieee80211_tx_control *frame_txctl) + const struct ieee80211_tx_info *frame_txctl) { struct ieee80211_local *local = hw_to_local(hw); struct ieee80211_rate *rate; @@ -309,7 +309,7 @@ __le16 ieee80211_ctstoself_duration(struct ieee80211_hw *hw, short_preamble = sdata->bss_conf.use_short_preamble; - rate = &sband->bitrates[frame_txctl->rts_cts_rate_idx]; + rate = &sband->bitrates[frame_txctl->control.rts_cts_rate_idx]; erp = 0; if (sdata->flags & IEEE80211_SDATA_OPERATING_GMODE) erp = rate->flags & IEEE80211_RATE_ERP_G; @@ -317,7 +317,7 @@ __le16 ieee80211_ctstoself_duration(struct ieee80211_hw *hw, /* Data frame duration */ dur = ieee80211_frame_duration(local, frame_len, rate->bitrate, erp, short_preamble); - if (!(frame_txctl->flags & IEEE80211_TXCTL_NO_ACK)) { + if (!(frame_txctl->flags & IEEE80211_TX_CTL_NO_ACK)) { /* ACK duration */ dur += ieee80211_frame_duration(local, 10, rate->bitrate, erp, short_preamble); diff --git a/net/mac80211/wep.c b/net/mac80211/wep.c index 3cbae42ec504..1e7f03dd8f6b 100644 --- a/net/mac80211/wep.c +++ b/net/mac80211/wep.c @@ -333,11 +333,16 @@ ieee80211_crypto_wep_decrypt(struct ieee80211_rx_data *rx) static int wep_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) { + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); + + info->control.iv_len = WEP_IV_LEN; + info->control.icv_len = WEP_ICV_LEN; + if (!(tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE)) { + info->control.hw_key = &tx->key->conf; if (ieee80211_wep_encrypt(tx->local, skb, tx->key)) return -1; } else { - tx->control->hw_key = &tx->key->conf; if (tx->key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_IV) { if (!ieee80211_wep_add_iv(tx->local, skb, tx->key)) return -1; @@ -349,8 +354,6 @@ static int wep_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) ieee80211_tx_result ieee80211_crypto_wep_encrypt(struct ieee80211_tx_data *tx) { - tx->control->iv_len = WEP_IV_LEN; - tx->control->icv_len = WEP_ICV_LEN; ieee80211_tx_set_protected(tx); if (wep_encrypt_skb(tx, tx->skb) < 0) { diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index c87baf4ce979..477690f4dca7 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -149,8 +149,7 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_tx_packet_data *pkt_data = - (struct ieee80211_tx_packet_data *) skb->cb; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; unsigned short fc = le16_to_cpu(hdr->frame_control); struct Qdisc *qdisc; @@ -158,8 +157,8 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) int err, queue; u8 tid; - if (pkt_data->flags & IEEE80211_TXPD_REQUEUE) { - queue = pkt_data->queue; + if (info->flags & IEEE80211_TX_CTL_REQUEUE) { + queue = info->queue; rcu_read_lock(); sta = sta_info_get(local, hdr->addr1); tid = skb->priority & QOS_CONTROL_TAG1D_MASK; @@ -168,9 +167,9 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) if ((ampdu_queue < QD_NUM(hw)) && test_bit(ampdu_queue, q->qdisc_pool)) { queue = ampdu_queue; - pkt_data->flags |= IEEE80211_TXPD_AMPDU; + info->flags |= IEEE80211_TX_CTL_AMPDU; } else { - pkt_data->flags &= ~IEEE80211_TXPD_AMPDU; + info->flags &= ~IEEE80211_TX_CTL_AMPDU; } } rcu_read_unlock(); @@ -206,9 +205,9 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) if ((ampdu_queue < QD_NUM(hw)) && test_bit(ampdu_queue, q->qdisc_pool)) { queue = ampdu_queue; - pkt_data->flags |= IEEE80211_TXPD_AMPDU; + info->flags |= IEEE80211_TX_CTL_AMPDU; } else { - pkt_data->flags &= ~IEEE80211_TXPD_AMPDU; + info->flags &= ~IEEE80211_TX_CTL_AMPDU; } } @@ -220,7 +219,7 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) err = NET_XMIT_DROP; } else { tid = skb->priority & QOS_CONTROL_TAG1D_MASK; - pkt_data->queue = (unsigned int) queue; + info->queue = (unsigned int) queue; qdisc = q->queues[queue]; err = qdisc->enqueue(skb, qdisc); if (err == NET_XMIT_SUCCESS) { @@ -241,13 +240,12 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) static int wme_qdiscop_requeue(struct sk_buff *skb, struct Qdisc* qd) { struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_tx_packet_data *pkt_data = - (struct ieee80211_tx_packet_data *) skb->cb; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); struct Qdisc *qdisc; int err; /* we recorded which queue to use earlier! */ - qdisc = q->queues[pkt_data->queue]; + qdisc = q->queues[info->queue]; if ((err = qdisc->ops->requeue(skb, qdisc)) == 0) { qd->q.qlen++; diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index d7304490d2ec..d6635f6e5618 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -183,15 +183,25 @@ ieee80211_rx_h_michael_mic_verify(struct ieee80211_rx_data *rx) } -static int tkip_encrypt_skb(struct ieee80211_tx_data *tx, - struct sk_buff *skb, int test) +static int tkip_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) { struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; struct ieee80211_key *key = tx->key; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); int hdrlen, len, tailneed; u16 fc; u8 *pos; + info->control.icv_len = TKIP_ICV_LEN; + info->control.iv_len = TKIP_IV_LEN; + + if ((tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) && + !(tx->key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_IV)) { + /* hwaccel - with no need for preallocated room for IV/ICV */ + info->control.hw_key = &tx->key->conf; + return TX_CONTINUE; + } + fc = le16_to_cpu(hdr->frame_control); hdrlen = ieee80211_get_hdrlen(fc); len = skb->len - hdrlen; @@ -228,7 +238,7 @@ static int tkip_encrypt_skb(struct ieee80211_tx_data *tx, 0x7f), (u8) key->u.tkip.tx.iv16); - tx->control->hw_key = &tx->key->conf; + info->control.hw_key = &tx->key->conf; return 0; } @@ -246,28 +256,16 @@ ieee80211_tx_result ieee80211_crypto_tkip_encrypt(struct ieee80211_tx_data *tx) { struct sk_buff *skb = tx->skb; - int wpa_test = 0, test = 0; - tx->control->icv_len = TKIP_ICV_LEN; - tx->control->iv_len = TKIP_IV_LEN; ieee80211_tx_set_protected(tx); - if ((tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) && - !(tx->key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_IV) && - !wpa_test) { - /* hwaccel - with no need for preallocated room for IV/ICV */ - tx->control->hw_key = &tx->key->conf; - return TX_CONTINUE; - } - - if (tkip_encrypt_skb(tx, skb, test) < 0) + if (tkip_encrypt_skb(tx, skb) < 0) return TX_DROP; if (tx->extra_frag) { int i; for (i = 0; i < tx->num_extra_frag; i++) { - if (tkip_encrypt_skb(tx, tx->extra_frag[i], test) - < 0) + if (tkip_encrypt_skb(tx, tx->extra_frag[i]) < 0) return TX_DROP; } } @@ -429,16 +427,27 @@ static inline int ccmp_hdr2pn(u8 *pn, u8 *hdr) } -static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, - struct sk_buff *skb, int test) +static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) { struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; struct ieee80211_key *key = tx->key; + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); int hdrlen, len, tailneed; u16 fc; u8 *pos, *pn, *b_0, *aad, *scratch; int i; + info->control.icv_len = CCMP_MIC_LEN; + info->control.iv_len = CCMP_HDR_LEN; + + if ((tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) && + !(tx->key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_IV)) { + /* hwaccel - with no need for preallocated room for CCMP " + * header or MIC fields */ + info->control.hw_key = &tx->key->conf; + return TX_CONTINUE; + } + scratch = key->u.ccmp.tx_crypto_buf; b_0 = scratch + 3 * AES_BLOCK_LEN; aad = scratch + 4 * AES_BLOCK_LEN; @@ -478,7 +487,7 @@ static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, if (key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) { /* hwaccel - with preallocated room for CCMP header */ - tx->control->hw_key = &tx->key->conf; + info->control.hw_key = &tx->key->conf; return 0; } @@ -495,28 +504,16 @@ ieee80211_tx_result ieee80211_crypto_ccmp_encrypt(struct ieee80211_tx_data *tx) { struct sk_buff *skb = tx->skb; - int test = 0; - tx->control->icv_len = CCMP_MIC_LEN; - tx->control->iv_len = CCMP_HDR_LEN; ieee80211_tx_set_protected(tx); - if ((tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) && - !(tx->key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_IV)) { - /* hwaccel - with no need for preallocated room for CCMP " - * header or MIC fields */ - tx->control->hw_key = &tx->key->conf; - return TX_CONTINUE; - } - - if (ccmp_encrypt_skb(tx, skb, test) < 0) + if (ccmp_encrypt_skb(tx, skb) < 0) return TX_DROP; if (tx->extra_frag) { int i; for (i = 0; i < tx->num_extra_frag; i++) { - if (ccmp_encrypt_skb(tx, tx->extra_frag[i], test) - < 0) + if (ccmp_encrypt_skb(tx, tx->extra_frag[i]) < 0) return TX_DROP; } } -- cgit v1.2.2 From eefce91a384a64c7bbf913eb08c4adfb911c3639 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Sat, 17 May 2008 00:57:13 +0200 Subject: mac80211: dont allow fragmentation and requeuing on A-MPDU queues There really is no reason for a driver to reject a frame on an A-MPDU queue when it can stop that queue for any period of time and is given frames one by one. Hence, disallow it with a big warning and reduce mac80211-internal state. Also add a warning when we try to fragment a frame destined for an A-MPDU queue and drop it, the actual bug needs to be fixed elsewhere but I'm not exactly sure how to yet. Signed-off-by: Johannes Berg Cc: Ron Rindjunsky Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 2 +- net/mac80211/tx.c | 23 +++++++++++++++++++++-- 2 files changed, 22 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 79a65b3ee02b..86a861251e8c 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -595,7 +595,7 @@ struct ieee80211_local { struct timer_list sta_cleanup; unsigned long state[IEEE80211_MAX_QUEUES + IEEE80211_MAX_AMPDU_QUEUES]; - struct ieee80211_tx_stored_packet pending_packet[IEEE80211_MAX_QUEUES + IEEE80211_MAX_AMPDU_QUEUES]; + struct ieee80211_tx_stored_packet pending_packet[IEEE80211_MAX_QUEUES]; struct tasklet_struct tx_pending_tasklet; /* number of interfaces with corresponding IFF_ flags */ diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index ac9a4af7ad42..6268bbca148e 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -673,6 +673,16 @@ ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) if (!(tx->flags & IEEE80211_TX_FRAGMENTED)) return TX_CONTINUE; + /* + * Warn when submitting a fragmented A-MPDU frame and drop it. + * This is an error and needs to be fixed elsewhere, but when + * done needs to take care of monitor interfaces (injection) + * etc. + */ + if (WARN_ON(tx->flags & IEEE80211_TX_CTL_AMPDU || + IEEE80211_SKB_CB(tx->skb)->queue >= tx->local->hw.queues)) + return TX_DROP; + first = tx->skb; hdrlen = ieee80211_get_hdrlen(tx->fc); @@ -1216,8 +1226,17 @@ static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb) retry: ret = __ieee80211_tx(local, skb, &tx); if (ret) { - struct ieee80211_tx_stored_packet *store = - &local->pending_packet[info->queue]; + struct ieee80211_tx_stored_packet *store; + + /* + * Since there are no fragmented frames on A-MPDU + * queues, there's no reason for a driver to reject + * a frame there, warn and drop it. + */ + if (WARN_ON(queue >= local->hw.queues)) + goto drop; + + store = &local->pending_packet[queue]; if (ret == IEEE80211_TX_FRAG_AGAIN) skb = NULL; -- cgit v1.2.2 From e2530083609148a7835b54c431f6b8956407c1f6 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Sat, 17 May 2008 00:57:14 +0200 Subject: mac80211: use multi-queue master netdevice This patch updates mac80211 and drivers to be multi-queue aware and use that instead of the internal queue mapping. Also does a number of cleanups in various pieces of the code that fall out and reduces internal mac80211 state size. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/Kconfig | 14 ++++++- net/mac80211/Makefile | 2 +- net/mac80211/ieee80211_i.h | 16 +++++--- net/mac80211/iface.c | 2 +- net/mac80211/main.c | 29 ++++++++++---- net/mac80211/sta_info.c | 2 +- net/mac80211/tx.c | 97 +++++++++++++++++++++------------------------- net/mac80211/util.c | 29 +++++++------- net/mac80211/wme.c | 14 +++---- net/mac80211/wme.h | 2 +- 10 files changed, 114 insertions(+), 93 deletions(-) (limited to 'net') diff --git a/net/mac80211/Kconfig b/net/mac80211/Kconfig index a24b459dd45a..590e00b2766c 100644 --- a/net/mac80211/Kconfig +++ b/net/mac80211/Kconfig @@ -7,11 +7,23 @@ config MAC80211 select CRC32 select WIRELESS_EXT select CFG80211 - select NET_SCH_FIFO ---help--- This option enables the hardware independent IEEE 802.11 networking stack. +config MAC80211_QOS + def_bool y + depends on MAC80211 + depends on NET_SCHED + depends on NETDEVICES_MULTIQUEUE + +comment "QoS/HT support disabled" + depends on MAC80211 && !MAC80211_QOS +comment "QoS/HT support needs CONFIG_NET_SCHED" + depends on MAC80211 && !NET_SCHED +comment "QoS/HT support needs CONFIG_NETDEVICES_MULTIQUEUE" + depends on MAC80211 && !NETDEVICES_MULTIQUEUE + menu "Rate control algorithm selection" depends on MAC80211 != n diff --git a/net/mac80211/Makefile b/net/mac80211/Makefile index 4e5847fd316c..1d2a4e010e5c 100644 --- a/net/mac80211/Makefile +++ b/net/mac80211/Makefile @@ -29,7 +29,7 @@ mac80211-y := \ event.o mac80211-$(CONFIG_MAC80211_LEDS) += led.o -mac80211-$(CONFIG_NET_SCHED) += wme.o +mac80211-$(CONFIG_MAC80211_QOS) += wme.o mac80211-$(CONFIG_MAC80211_DEBUGFS) += \ debugfs.o \ debugfs_sta.o \ diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 86a861251e8c..7d614cdcefcb 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -594,7 +594,7 @@ struct ieee80211_local { struct sta_info *sta_hash[STA_HASH_SIZE]; struct timer_list sta_cleanup; - unsigned long state[IEEE80211_MAX_QUEUES + IEEE80211_MAX_AMPDU_QUEUES]; + unsigned long queues_pending[BITS_TO_LONGS(IEEE80211_MAX_QUEUES)]; struct ieee80211_tx_stored_packet pending_packet[IEEE80211_MAX_QUEUES]; struct tasklet_struct tx_pending_tasklet; @@ -758,6 +758,15 @@ struct ieee80211_local { #endif }; +static inline int ieee80211_is_multiqueue(struct ieee80211_local *local) +{ +#ifdef CONFIG_MAC80211_QOS + return netif_is_multiqueue(local->mdev); +#else + return 0; +#endif +} + /* this struct represents 802.11n's RA/TID combination */ struct ieee80211_ra_tid { u8 ra[ETH_ALEN]; @@ -827,11 +836,6 @@ static inline struct ieee80211_hw *local_to_hw( return &local->hw; } -enum ieee80211_link_state_t { - IEEE80211_LINK_STATE_XOFF = 0, - IEEE80211_LINK_STATE_PENDING, -}; - struct sta_attribute { struct attribute attr; ssize_t (*show)(const struct sta_info *, char *buf); diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c index 3c64e42eb12e..984472702381 100644 --- a/net/mac80211/iface.c +++ b/net/mac80211/iface.c @@ -168,7 +168,7 @@ void ieee80211_if_set_type(struct net_device *dev, int type) ifsta->flags |= IEEE80211_STA_CREATE_IBSS | IEEE80211_STA_AUTO_BSSID_SEL | IEEE80211_STA_AUTO_CHANNEL_SEL; - if (sdata->local->hw.queues >= 4) + if (ieee80211_num_regular_queues(&sdata->local->hw) >= 4) ifsta->flags |= IEEE80211_STA_WMM_ENABLED; msdata = IEEE80211_DEV_TO_SUB_IF(sdata->local->mdev); diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 8f1ff7ef1667..97d4a537ca2f 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -1634,12 +1634,32 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) if (result < 0) return result; + /* + * We use the number of queues for feature tests (QoS, HT) internally + * so restrict them appropriately. + */ +#ifdef CONFIG_MAC80211_QOS + if (hw->queues > IEEE80211_MAX_QUEUES) + hw->queues = IEEE80211_MAX_QUEUES; + if (hw->ampdu_queues > IEEE80211_MAX_AMPDU_QUEUES) + hw->ampdu_queues = IEEE80211_MAX_AMPDU_QUEUES; + if (hw->queues < 4) + hw->ampdu_queues = 0; +#else + hw->queues = 1; + hw->ampdu_queues = 0; +#endif + /* for now, mdev needs sub_if_data :/ */ - mdev = alloc_netdev(sizeof(struct ieee80211_sub_if_data), - "wmaster%d", ether_setup); + mdev = alloc_netdev_mq(sizeof(struct ieee80211_sub_if_data), + "wmaster%d", ether_setup, + ieee80211_num_queues(hw)); if (!mdev) goto fail_mdev_alloc; + if (ieee80211_num_queues(hw) > 1) + mdev->features |= NETIF_F_MULTI_QUEUE; + sdata = IEEE80211_DEV_TO_SUB_IF(mdev); mdev->ieee80211_ptr = &sdata->wdev; sdata->wdev.wiphy = local->hw.wiphy; @@ -1728,11 +1748,6 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) goto fail_wep; } - if (hw->queues > IEEE80211_MAX_QUEUES) - hw->queues = IEEE80211_MAX_QUEUES; - if (hw->ampdu_queues > IEEE80211_MAX_AMPDU_QUEUES) - hw->ampdu_queues = IEEE80211_MAX_AMPDU_QUEUES; - ieee80211_install_qdisc(local->mdev); /* add one default STA interface */ diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index ef3149324d54..c24770cb02c5 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -255,7 +255,7 @@ struct sta_info *sta_info_alloc(struct ieee80211_sub_if_data *sdata, * sta_rx_agg_session_timer_expired for useage */ sta->timer_to_tid[i] = i; /* tid to tx queue: initialize according to HW (0 is valid) */ - sta->tid_to_tx_q[i] = local->hw.queues + local->hw.ampdu_queues; + sta->tid_to_tx_q[i] = ieee80211_num_queues(&local->hw); /* rx */ sta->ampdu_mlme.tid_state_rx[i] = HT_AGG_STATE_IDLE; sta->ampdu_mlme.tid_rx[i] = NULL; diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 6268bbca148e..9273651d3d7c 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -213,18 +213,6 @@ static u16 ieee80211_duration(struct ieee80211_tx_data *tx, int group_addr, return dur; } -static inline int __ieee80211_queue_stopped(const struct ieee80211_local *local, - int queue) -{ - return test_bit(IEEE80211_LINK_STATE_XOFF, &local->state[queue]); -} - -static inline int __ieee80211_queue_pending(const struct ieee80211_local *local, - int queue) -{ - return test_bit(IEEE80211_LINK_STATE_PENDING, &local->state[queue]); -} - static int inline is_ieee80211_device(struct net_device *dev, struct net_device *master) { @@ -680,7 +668,8 @@ ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) * etc. */ if (WARN_ON(tx->flags & IEEE80211_TX_CTL_AMPDU || - IEEE80211_SKB_CB(tx->skb)->queue >= tx->local->hw.queues)) + skb_get_queue_mapping(tx->skb) >= + ieee80211_num_regular_queues(&tx->local->hw))) return TX_DROP; first = tx->skb; @@ -1098,11 +1087,9 @@ static int __ieee80211_tx(struct ieee80211_local *local, struct sk_buff *skb, struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); int ret, i; - if (!ieee80211_qdisc_installed(local->mdev) && - __ieee80211_queue_stopped(local, 0)) { - netif_stop_queue(local->mdev); + if (netif_subqueue_stopped(local->mdev, skb)) return IEEE80211_TX_AGAIN; - } + if (skb) { ieee80211_dump_frame(wiphy_name(local->hw.wiphy), "TX to low-level driver", skb); @@ -1121,7 +1108,8 @@ static int __ieee80211_tx(struct ieee80211_local *local, struct sk_buff *skb, IEEE80211_TX_CTL_USE_CTS_PROTECT | IEEE80211_TX_CTL_CLEAR_PS_FILT | IEEE80211_TX_CTL_FIRST_FRAGMENT); - if (__ieee80211_queue_stopped(local, info->queue)) + if (netif_subqueue_stopped(local->mdev, + tx->extra_frag[i])) return IEEE80211_TX_FRAG_AGAIN; if (i == tx->num_extra_frag) { info->tx_rate_idx = tx->last_frag_rate_idx; @@ -1160,9 +1148,11 @@ static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb) ieee80211_tx_result res = TX_DROP, res_prepare; struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); int ret, i; - int queue = info->queue; + u16 queue; - WARN_ON(__ieee80211_queue_pending(local, queue)); + queue = skb_get_queue_mapping(skb); + + WARN_ON(test_bit(queue, local->queues_pending)); if (unlikely(skb->len < 10)) { dev_kfree_skb(skb); @@ -1233,28 +1223,28 @@ retry: * queues, there's no reason for a driver to reject * a frame there, warn and drop it. */ - if (WARN_ON(queue >= local->hw.queues)) + if (WARN_ON(queue >= ieee80211_num_regular_queues(&local->hw))) goto drop; store = &local->pending_packet[queue]; if (ret == IEEE80211_TX_FRAG_AGAIN) skb = NULL; - set_bit(IEEE80211_LINK_STATE_PENDING, - &local->state[queue]); + set_bit(queue, local->queues_pending); smp_mb(); - /* When the driver gets out of buffers during sending of - * fragments and calls ieee80211_stop_queue, there is - * a small window between IEEE80211_LINK_STATE_XOFF and - * IEEE80211_LINK_STATE_PENDING flags are set. If a buffer + /* + * When the driver gets out of buffers during sending of + * fragments and calls ieee80211_stop_queue, the netif + * subqueue is stopped. There is, however, a small window + * in which the PENDING bit is not yet set. If a buffer * gets available in that window (i.e. driver calls * ieee80211_wake_queue), we would end up with ieee80211_tx - * called with IEEE80211_LINK_STATE_PENDING. Prevent this by + * called with the PENDING bit still set. Prevent this by * continuing transmitting here when that situation is - * possible to have happened. */ - if (!__ieee80211_queue_stopped(local, queue)) { - clear_bit(IEEE80211_LINK_STATE_PENDING, - &local->state[queue]); + * possible to have happened. + */ + if (!__netif_subqueue_stopped(local->mdev, queue)) { + clear_bit(queue, local->queues_pending); goto retry; } store->skb = skb; @@ -1509,7 +1499,8 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, } /* receiver and we are QoS enabled, use a QoS type frame */ - if (sta_flags & WLAN_STA_WME && local->hw.queues >= 4) { + if (sta_flags & WLAN_STA_WME && + ieee80211_num_regular_queues(&local->hw) >= 4) { fc |= IEEE80211_STYPE_QOS_DATA; hdrlen += 2; } @@ -1661,41 +1652,51 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, return ret; } -/* helper functions for pending packets for when queues are stopped */ +/* + * ieee80211_clear_tx_pending may not be called in a context where + * it is possible that it packets could come in again. + */ void ieee80211_clear_tx_pending(struct ieee80211_local *local) { int i, j; struct ieee80211_tx_stored_packet *store; - for (i = 0; i < local->hw.queues; i++) { - if (!__ieee80211_queue_pending(local, i)) + for (i = 0; i < ieee80211_num_regular_queues(&local->hw); i++) { + if (!test_bit(i, local->queues_pending)) continue; store = &local->pending_packet[i]; kfree_skb(store->skb); for (j = 0; j < store->num_extra_frag; j++) kfree_skb(store->extra_frag[j]); kfree(store->extra_frag); - clear_bit(IEEE80211_LINK_STATE_PENDING, &local->state[i]); + clear_bit(i, local->queues_pending); } } +/* + * Transmit all pending packets. Called from tasklet, locks master device + * TX lock so that no new packets can come in. + */ void ieee80211_tx_pending(unsigned long data) { struct ieee80211_local *local = (struct ieee80211_local *)data; struct net_device *dev = local->mdev; struct ieee80211_tx_stored_packet *store; struct ieee80211_tx_data tx; - int i, ret, reschedule = 0; + int i, ret; netif_tx_lock_bh(dev); - for (i = 0; i < local->hw.queues; i++) { - if (__ieee80211_queue_stopped(local, i)) + for (i = 0; i < ieee80211_num_regular_queues(&local->hw); i++) { + /* Check that this queue is ok */ + if (__netif_subqueue_stopped(local->mdev, i)) continue; - if (!__ieee80211_queue_pending(local, i)) { - reschedule = 1; + + if (!test_bit(i, local->queues_pending)) { + ieee80211_wake_queue(&local->hw, i); continue; } + store = &local->pending_packet[i]; tx.extra_frag = store->extra_frag; tx.num_extra_frag = store->num_extra_frag; @@ -1708,19 +1709,11 @@ void ieee80211_tx_pending(unsigned long data) if (ret == IEEE80211_TX_FRAG_AGAIN) store->skb = NULL; } else { - clear_bit(IEEE80211_LINK_STATE_PENDING, - &local->state[i]); - reschedule = 1; + clear_bit(i, local->queues_pending); + ieee80211_wake_queue(&local->hw, i); } } netif_tx_unlock_bh(dev); - if (reschedule) { - if (!ieee80211_qdisc_installed(dev)) { - if (!__ieee80211_queue_stopped(local, 0)) - netif_wake_queue(dev); - } else - netif_schedule(dev); - } } /* functions for drivers to get certain frames */ diff --git a/net/mac80211/util.c b/net/mac80211/util.c index d9109dee461f..4f7180b287da 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -331,17 +331,15 @@ void ieee80211_wake_queue(struct ieee80211_hw *hw, int queue) { struct ieee80211_local *local = hw_to_local(hw); - if (test_and_clear_bit(IEEE80211_LINK_STATE_XOFF, - &local->state[queue])) { - if (test_bit(IEEE80211_LINK_STATE_PENDING, - &local->state[queue])) - tasklet_schedule(&local->tx_pending_tasklet); - else - if (!ieee80211_qdisc_installed(local->mdev)) { - if (queue == 0) - netif_wake_queue(local->mdev); - } else - __netif_schedule(local->mdev); + if (test_bit(queue, local->queues_pending)) { + tasklet_schedule(&local->tx_pending_tasklet); + } else { + if (ieee80211_is_multiqueue(local)) { + netif_wake_subqueue(local->mdev, queue); + } else { + WARN_ON(queue != 0); + netif_wake_queue(local->mdev); + } } } EXPORT_SYMBOL(ieee80211_wake_queue); @@ -350,9 +348,12 @@ void ieee80211_stop_queue(struct ieee80211_hw *hw, int queue) { struct ieee80211_local *local = hw_to_local(hw); - if (!ieee80211_qdisc_installed(local->mdev) && queue == 0) + if (ieee80211_is_multiqueue(local)) { + netif_stop_subqueue(local->mdev, queue); + } else { + WARN_ON(queue != 0); netif_stop_queue(local->mdev); - set_bit(IEEE80211_LINK_STATE_XOFF, &local->state[queue]); + } } EXPORT_SYMBOL(ieee80211_stop_queue); @@ -360,7 +361,7 @@ void ieee80211_stop_queues(struct ieee80211_hw *hw) { int i; - for (i = 0; i < hw->queues + hw->ampdu_queues; i++) + for (i = 0; i < ieee80211_num_queues(hw); i++) ieee80211_stop_queue(hw, i); } EXPORT_SYMBOL(ieee80211_stop_queues); diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 477690f4dca7..14a9ff10a1e9 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -158,7 +158,7 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) u8 tid; if (info->flags & IEEE80211_TX_CTL_REQUEUE) { - queue = info->queue; + queue = skb_get_queue_mapping(skb); rcu_read_lock(); sta = sta_info_get(local, hdr->addr1); tid = skb->priority & QOS_CONTROL_TAG1D_MASK; @@ -219,7 +219,7 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) err = NET_XMIT_DROP; } else { tid = skb->priority & QOS_CONTROL_TAG1D_MASK; - info->queue = (unsigned int) queue; + skb_set_queue_mapping(skb, queue); qdisc = q->queues[queue]; err = qdisc->enqueue(skb, qdisc); if (err == NET_XMIT_SUCCESS) { @@ -240,12 +240,11 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) static int wme_qdiscop_requeue(struct sk_buff *skb, struct Qdisc* qd) { struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); struct Qdisc *qdisc; int err; /* we recorded which queue to use earlier! */ - qdisc = q->queues[info->queue]; + qdisc = q->queues[skb_get_queue_mapping(skb)]; if ((err = qdisc->ops->requeue(skb, qdisc)) == 0) { qd->q.qlen++; @@ -269,11 +268,8 @@ static struct sk_buff *wme_qdiscop_dequeue(struct Qdisc* qd) /* check all the h/w queues in numeric/priority order */ for (queue = 0; queue < QD_NUM(hw); queue++) { /* see if there is room in this hardware queue */ - if ((test_bit(IEEE80211_LINK_STATE_XOFF, - &local->state[queue])) || - (test_bit(IEEE80211_LINK_STATE_PENDING, - &local->state[queue])) || - (!test_bit(queue, q->qdisc_pool))) + if (__netif_subqueue_stopped(local->mdev, queue) || + !test_bit(queue, q->qdisc_pool)) continue; /* there is space - try and get a frame */ diff --git a/net/mac80211/wme.h b/net/mac80211/wme.h index fcc6b05508cc..bbdb53344817 100644 --- a/net/mac80211/wme.h +++ b/net/mac80211/wme.h @@ -31,7 +31,7 @@ static inline int WLAN_FC_IS_QOS_DATA(u16 fc) return (fc & 0x8C) == 0x88; } -#ifdef CONFIG_NET_SCHED +#ifdef CONFIG_MAC80211_QOS void ieee80211_install_qdisc(struct net_device *dev); int ieee80211_qdisc_installed(struct net_device *dev); int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, -- cgit v1.2.2 From 9e72ebd686a7f39facdfff639386055f1ad7dc88 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Wed, 21 May 2008 17:33:42 +0200 Subject: mac80211: remove channel use statistics The useless channel use statistics are quite a lot of code, currently use integer divisions in the packet fast path, are rather inaccurate since they do not account for retries and finally nobody even cares. Hence, remove them completely. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/debugfs_netdev.c | 9 ------ net/mac80211/debugfs_sta.c | 1 - net/mac80211/ieee80211_i.h | 32 ------------------- net/mac80211/rx.c | 62 ++---------------------------------- net/mac80211/sta_info.h | 4 --- net/mac80211/tx.c | 73 ++++++------------------------------------- 6 files changed, 12 insertions(+), 169 deletions(-) (limited to 'net') diff --git a/net/mac80211/debugfs_netdev.c b/net/mac80211/debugfs_netdev.c index 3ae5493d728b..b2089b2da48a 100644 --- a/net/mac80211/debugfs_netdev.c +++ b/net/mac80211/debugfs_netdev.c @@ -155,7 +155,6 @@ static const struct file_operations name##_ops = { \ __IEEE80211_IF_WFILE(name) /* common attributes */ -IEEE80211_IF_FILE(channel_use, channel_use, DEC); IEEE80211_IF_FILE(drop_unencrypted, drop_unencrypted, DEC); /* STA/IBSS attributes */ @@ -248,7 +247,6 @@ IEEE80211_IF_WFILE(min_discovery_timeout, static void add_sta_files(struct ieee80211_sub_if_data *sdata) { - DEBUGFS_ADD(channel_use, sta); DEBUGFS_ADD(drop_unencrypted, sta); DEBUGFS_ADD(state, sta); DEBUGFS_ADD(bssid, sta); @@ -269,7 +267,6 @@ static void add_sta_files(struct ieee80211_sub_if_data *sdata) static void add_ap_files(struct ieee80211_sub_if_data *sdata) { - DEBUGFS_ADD(channel_use, ap); DEBUGFS_ADD(drop_unencrypted, ap); DEBUGFS_ADD(num_sta_ps, ap); DEBUGFS_ADD(dtim_count, ap); @@ -281,14 +278,12 @@ static void add_ap_files(struct ieee80211_sub_if_data *sdata) static void add_wds_files(struct ieee80211_sub_if_data *sdata) { - DEBUGFS_ADD(channel_use, wds); DEBUGFS_ADD(drop_unencrypted, wds); DEBUGFS_ADD(peer, wds); } static void add_vlan_files(struct ieee80211_sub_if_data *sdata) { - DEBUGFS_ADD(channel_use, vlan); DEBUGFS_ADD(drop_unencrypted, vlan); } @@ -376,7 +371,6 @@ static void add_files(struct ieee80211_sub_if_data *sdata) static void del_sta_files(struct ieee80211_sub_if_data *sdata) { - DEBUGFS_DEL(channel_use, sta); DEBUGFS_DEL(drop_unencrypted, sta); DEBUGFS_DEL(state, sta); DEBUGFS_DEL(bssid, sta); @@ -397,7 +391,6 @@ static void del_sta_files(struct ieee80211_sub_if_data *sdata) static void del_ap_files(struct ieee80211_sub_if_data *sdata) { - DEBUGFS_DEL(channel_use, ap); DEBUGFS_DEL(drop_unencrypted, ap); DEBUGFS_DEL(num_sta_ps, ap); DEBUGFS_DEL(dtim_count, ap); @@ -409,14 +402,12 @@ static void del_ap_files(struct ieee80211_sub_if_data *sdata) static void del_wds_files(struct ieee80211_sub_if_data *sdata) { - DEBUGFS_DEL(channel_use, wds); DEBUGFS_DEL(drop_unencrypted, wds); DEBUGFS_DEL(peer, wds); } static void del_vlan_files(struct ieee80211_sub_if_data *sdata) { - DEBUGFS_DEL(channel_use, vlan); DEBUGFS_DEL(drop_unencrypted, vlan); } diff --git a/net/mac80211/debugfs_sta.c b/net/mac80211/debugfs_sta.c index a2cc0284c9d0..79a062782d52 100644 --- a/net/mac80211/debugfs_sta.c +++ b/net/mac80211/debugfs_sta.c @@ -66,7 +66,6 @@ STA_FILE(tx_retry_count, tx_retry_count, LU); STA_FILE(last_signal, last_signal, D); STA_FILE(last_qual, last_qual, D); STA_FILE(last_noise, last_noise, D); -STA_FILE(channel_use, channel_use, D); STA_FILE(wep_weak_iv_count, wep_weak_iv_count, LU); static ssize_t sta_flags_read(struct file *file, char __user *userbuf, diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 7d614cdcefcb..3f8601cafffb 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -201,7 +201,6 @@ struct ieee80211_rx_data { unsigned int flags; int sent_ps_buffered; int queue; - int load; u32 tkip_iv32; u16 tkip_iv16; }; @@ -448,14 +447,11 @@ struct ieee80211_sub_if_data { struct ieee80211_if_sta sta; u32 mntr_flags; } u; - int channel_use; - int channel_use_raw; #ifdef CONFIG_MAC80211_DEBUGFS struct dentry *debugfsdir; union { struct { - struct dentry *channel_use; struct dentry *drop_unencrypted; struct dentry *state; struct dentry *bssid; @@ -474,7 +470,6 @@ struct ieee80211_sub_if_data { struct dentry *num_beacons_sta; } sta; struct { - struct dentry *channel_use; struct dentry *drop_unencrypted; struct dentry *num_sta_ps; struct dentry *dtim_count; @@ -484,12 +479,10 @@ struct ieee80211_sub_if_data { struct dentry *num_buffered_multicast; } ap; struct { - struct dentry *channel_use; struct dentry *drop_unencrypted; struct dentry *peer; } wds; struct { - struct dentry *channel_use; struct dentry *drop_unencrypted; } vlan; struct { @@ -661,9 +654,6 @@ struct ieee80211_local { assoc_led_name[32], radio_led_name[32]; #endif - u32 channel_use; - u32 channel_use_raw; - #ifdef CONFIG_MAC80211_DEBUGFS struct work_struct sta_debugfs_add; #endif @@ -861,28 +851,6 @@ u32 ieee80211_handle_ht(struct ieee80211_local *local, int enable_ht, /* ieee80211_ioctl.c */ extern const struct iw_handler_def ieee80211_iw_handler_def; - - -/* Least common multiple of the used rates (in 100 kbps). This is used to - * calculate rate_inv values for each rate so that only integers are needed. */ -#define CHAN_UTIL_RATE_LCM 95040 -/* 1 usec is 1/8 * (95040/10) = 1188 */ -#define CHAN_UTIL_PER_USEC 1188 -/* Amount of bits to shift the result right to scale the total utilization - * to values that will not wrap around 32-bit integers. */ -#define CHAN_UTIL_SHIFT 9 -/* Theoretical maximum of channel utilization counter in 10 ms (stat_time=1): - * (CHAN_UTIL_PER_USEC * 10000) >> CHAN_UTIL_SHIFT = 23203. So dividing the - * raw value with about 23 should give utilization in 10th of a percentage - * (1/1000). However, utilization is only estimated and not all intervals - * between frames etc. are calculated. 18 seems to give numbers that are closer - * to the real maximum. */ -#define CHAN_UTIL_PER_10MS 18 -#define CHAN_UTIL_HDR_LONG (202 * CHAN_UTIL_PER_USEC) -#define CHAN_UTIL_HDR_SHORT (40 * CHAN_UTIL_PER_USEC) - - -/* ieee80211_ioctl.c */ int ieee80211_set_freq(struct ieee80211_local *local, int freq); /* ieee80211_sta.c */ void ieee80211_sta_timer(unsigned long data); diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index cf0de3b0fe24..9400a9766a77 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -387,50 +387,8 @@ static void ieee80211_verify_ip_alignment(struct ieee80211_rx_data *rx) } -static u32 ieee80211_rx_load_stats(struct ieee80211_local *local, - struct sk_buff *skb, - struct ieee80211_rx_status *status, - struct ieee80211_rate *rate) -{ - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; - u32 load = 0, hdrtime; - - /* Estimate total channel use caused by this frame */ - - /* 1 bit at 1 Mbit/s takes 1 usec; in channel_use values, - * 1 usec = 1/8 * (1080 / 10) = 13.5 */ - - if (status->band == IEEE80211_BAND_5GHZ || - (status->band == IEEE80211_BAND_5GHZ && - rate->flags & IEEE80211_RATE_ERP_G)) - hdrtime = CHAN_UTIL_HDR_SHORT; - else - hdrtime = CHAN_UTIL_HDR_LONG; - - load = hdrtime; - if (!is_multicast_ether_addr(hdr->addr1)) - load += hdrtime; - - /* TODO: optimise again */ - load += skb->len * CHAN_UTIL_RATE_LCM / rate->bitrate; - - /* Divide channel_use by 8 to avoid wrapping around the counter */ - load >>= CHAN_UTIL_SHIFT; - - return load; -} - /* rx handlers */ -static ieee80211_rx_result -ieee80211_rx_h_if_stats(struct ieee80211_rx_data *rx) -{ - if (rx->sta) - rx->sta->channel_use_raw += rx->load; - rx->sdata->channel_use_raw += rx->load; - return RX_CONTINUE; -} - static ieee80211_rx_result ieee80211_rx_h_passive_scan(struct ieee80211_rx_data *rx) { @@ -1780,7 +1738,6 @@ static void ieee80211_rx_cooked_monitor(struct ieee80211_rx_data *rx) typedef ieee80211_rx_result (*ieee80211_rx_handler)(struct ieee80211_rx_data *); static ieee80211_rx_handler ieee80211_rx_handlers[] = { - ieee80211_rx_h_if_stats, ieee80211_rx_h_passive_scan, ieee80211_rx_h_check, ieee80211_rx_h_decrypt, @@ -1939,7 +1896,6 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata, static void __ieee80211_rx_handle_packet(struct ieee80211_hw *hw, struct sk_buff *skb, struct ieee80211_rx_status *status, - u32 load, struct ieee80211_rate *rate) { struct ieee80211_local *local = hw_to_local(hw); @@ -1958,7 +1914,6 @@ static void __ieee80211_rx_handle_packet(struct ieee80211_hw *hw, rx.local = local; rx.status = status; - rx.load = load; rx.rate = rate; rx.fc = le16_to_cpu(hdr->frame_control); type = rx.fc & IEEE80211_FCTL_FTYPE; @@ -2067,7 +2022,6 @@ u8 ieee80211_sta_manage_reorder_buf(struct ieee80211_hw *hw, struct ieee80211_rx_status status; u16 head_seq_num, buf_size; int index; - u32 pkt_load; struct ieee80211_supported_band *sband; struct ieee80211_rate *rate; @@ -2102,12 +2056,9 @@ u8 ieee80211_sta_manage_reorder_buf(struct ieee80211_hw *hw, sizeof(status)); sband = local->hw.wiphy->bands[status.band]; rate = &sband->bitrates[status.rate_idx]; - pkt_load = ieee80211_rx_load_stats(local, - tid_agg_rx->reorder_buf[index], - &status, rate); __ieee80211_rx_handle_packet(hw, tid_agg_rx->reorder_buf[index], - &status, pkt_load, rate); + &status, rate); tid_agg_rx->stored_mpdu_num--; tid_agg_rx->reorder_buf[index] = NULL; } @@ -2149,11 +2100,8 @@ u8 ieee80211_sta_manage_reorder_buf(struct ieee80211_hw *hw, sizeof(status)); sband = local->hw.wiphy->bands[status.band]; rate = &sband->bitrates[status.rate_idx]; - pkt_load = ieee80211_rx_load_stats(local, - tid_agg_rx->reorder_buf[index], - &status, rate); __ieee80211_rx_handle_packet(hw, tid_agg_rx->reorder_buf[index], - &status, pkt_load, rate); + &status, rate); tid_agg_rx->stored_mpdu_num--; tid_agg_rx->reorder_buf[index] = NULL; tid_agg_rx->head_seq_num = seq_inc(tid_agg_rx->head_seq_num); @@ -2232,7 +2180,6 @@ void __ieee80211_rx(struct ieee80211_hw *hw, struct sk_buff *skb, struct ieee80211_rx_status *status) { struct ieee80211_local *local = hw_to_local(hw); - u32 pkt_load; struct ieee80211_rate *rate = NULL; struct ieee80211_supported_band *sband; @@ -2272,11 +2219,8 @@ void __ieee80211_rx(struct ieee80211_hw *hw, struct sk_buff *skb, return; } - pkt_load = ieee80211_rx_load_stats(local, skb, status, rate); - local->channel_use_raw += pkt_load; - if (!ieee80211_rx_reorder_ampdu(local, skb)) - __ieee80211_rx_handle_packet(hw, skb, status, pkt_load, rate); + __ieee80211_rx_handle_packet(hw, skb, status, rate); rcu_read_unlock(); } diff --git a/net/mac80211/sta_info.h b/net/mac80211/sta_info.h index f592290e42b9..95753f860acf 100644 --- a/net/mac80211/sta_info.h +++ b/net/mac80211/sta_info.h @@ -245,10 +245,6 @@ struct sta_info { unsigned int wme_tx_queue[NUM_RX_DATA_QUEUES]; #endif - /* Debug counters, no locking doesn't matter */ - int channel_use; - int channel_use_raw; - /* * Aggregation information, locked with lock. */ diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 9273651d3d7c..baa1be0671e5 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -761,73 +761,18 @@ ieee80211_tx_h_encrypt(struct ieee80211_tx_data *tx) static ieee80211_tx_result ieee80211_tx_h_stats(struct ieee80211_tx_data *tx) { - struct ieee80211_local *local = tx->local; - struct sk_buff *skb = tx->skb; - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; - u32 load = 0, hdrtime; - struct ieee80211_rate *rate; - struct ieee80211_supported_band *sband; - struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); - - sband = tx->local->hw.wiphy->bands[tx->channel->band]; - rate = &sband->bitrates[tx->rate_idx]; - - /* TODO: this could be part of tx_status handling, so that the number - * of retries would be known; TX rate should in that case be stored - * somewhere with the packet */ - - /* Estimate total channel use caused by this frame */ - - /* 1 bit at 1 Mbit/s takes 1 usec; in channel_use values, - * 1 usec = 1/8 * (1080 / 10) = 13.5 */ - - if (tx->channel->band == IEEE80211_BAND_5GHZ || - (tx->channel->band == IEEE80211_BAND_2GHZ && - rate->flags & IEEE80211_RATE_ERP_G)) - hdrtime = CHAN_UTIL_HDR_SHORT; - else - hdrtime = CHAN_UTIL_HDR_LONG; - - load = hdrtime; - if (!is_multicast_ether_addr(hdr->addr1)) - load += hdrtime; - - if (info->flags & IEEE80211_TX_CTL_USE_RTS_CTS) - load += 2 * hdrtime; - else if (info->flags & IEEE80211_TX_CTL_USE_CTS_PROTECT) - load += hdrtime; + int i; - /* TODO: optimise again */ - load += skb->len * CHAN_UTIL_RATE_LCM / rate->bitrate; + if (!tx->sta) + return TX_CONTINUE; + tx->sta->tx_packets++; + tx->sta->tx_fragments++; + tx->sta->tx_bytes += tx->skb->len; if (tx->extra_frag) { - int i; - for (i = 0; i < tx->num_extra_frag; i++) { - load += 2 * hdrtime; - load += tx->extra_frag[i]->len * - rate->bitrate; - } - } - - /* Divide channel_use by 8 to avoid wrapping around the counter */ - load >>= CHAN_UTIL_SHIFT; - local->channel_use_raw += load; - if (tx->sta) - tx->sta->channel_use_raw += load; - tx->sdata->channel_use_raw += load; - - if (tx->sta) { - tx->sta->tx_packets++; - tx->sta->tx_fragments++; - tx->sta->tx_bytes += tx->skb->len; - if (tx->extra_frag) { - int i; - tx->sta->tx_fragments += tx->num_extra_frag; - for (i = 0; i < tx->num_extra_frag; i++) { - tx->sta->tx_bytes += - tx->extra_frag[i]->len; - } - } + tx->sta->tx_fragments += tx->num_extra_frag; + for (i = 0; i < tx->num_extra_frag; i++) + tx->sta->tx_bytes += tx->extra_frag[i]->len; } return TX_CONTINUE; -- cgit v1.2.2 From 866988edacebb6f1982266b51fc344c969f481ec Mon Sep 17 00:00:00 2001 From: Alan Cox Date: Sun, 25 May 2008 23:41:40 -0700 Subject: wanrouter: Push down BKL Signed-off-by: Alan Cox Signed-off-by: David S. Miller --- net/wanrouter/wanmain.c | 6 ++++-- net/wanrouter/wanproc.c | 2 +- 2 files changed, 5 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/wanrouter/wanmain.c b/net/wanrouter/wanmain.c index 9ab31a3ce3ad..b210a88d0960 100644 --- a/net/wanrouter/wanmain.c +++ b/net/wanrouter/wanmain.c @@ -350,9 +350,9 @@ __be16 wanrouter_type_trans(struct sk_buff *skb, struct net_device *dev) * o execute requested action or pass command to the device driver */ -int wanrouter_ioctl(struct inode *inode, struct file *file, - unsigned int cmd, unsigned long arg) +long wanrouter_ioctl(struct file *file, unsigned int cmd, unsigned long arg) { + struct inode *inode = file->f_path.dentry->d_inode; int err = 0; struct proc_dir_entry *dent; struct wan_device *wandev; @@ -372,6 +372,7 @@ int wanrouter_ioctl(struct inode *inode, struct file *file, if (wandev->magic != ROUTER_MAGIC) return -EINVAL; + lock_kernel(); switch (cmd) { case ROUTER_SETUP: err = wanrouter_device_setup(wandev, data); @@ -403,6 +404,7 @@ int wanrouter_ioctl(struct inode *inode, struct file *file, err = wandev->ioctl(wandev, cmd, arg); else err = -EINVAL; } + unlock_kernel(); return err; } diff --git a/net/wanrouter/wanproc.c b/net/wanrouter/wanproc.c index 5bebe40bf4e6..267f7ff49827 100644 --- a/net/wanrouter/wanproc.c +++ b/net/wanrouter/wanproc.c @@ -278,7 +278,7 @@ static const struct file_operations wandev_fops = { .read = seq_read, .llseek = seq_lseek, .release = single_release, - .ioctl = wanrouter_ioctl, + .unlocked_ioctl = wanrouter_ioctl, }; /* -- cgit v1.2.2 From 54064600981ab0fe977b0e760732c30f4472d693 Mon Sep 17 00:00:00 2001 From: Alan Cox Date: Sun, 25 May 2008 23:43:11 -0700 Subject: irda: Push BKL down into irda ioctl handlers Signed-off-by: Alan Cox Signed-off-by: David S. Miller --- net/irda/irnet/irnet_ppp.c | 54 +++++++++++++++++++++++++--------------------- net/irda/irnet/irnet_ppp.h | 7 +++--- 2 files changed, 33 insertions(+), 28 deletions(-) (limited to 'net') diff --git a/net/irda/irnet/irnet_ppp.c b/net/irda/irnet/irnet_ppp.c index e0eab5927c4f..f6e54fa97f47 100644 --- a/net/irda/irnet/irnet_ppp.c +++ b/net/irda/irnet/irnet_ppp.c @@ -628,8 +628,8 @@ dev_irnet_poll(struct file * file, * This is the way pppd configure us and control us while the PPP * instance is active. */ -static int -dev_irnet_ioctl(struct inode * inode, +static long +dev_irnet_ioctl( struct file * file, unsigned int cmd, unsigned long arg) @@ -660,6 +660,7 @@ dev_irnet_ioctl(struct inode * inode, { DEBUG(FS_INFO, "Entering PPP discipline.\n"); /* PPP channel setup (ap->chan in configued in dev_irnet_open())*/ + lock_kernel(); err = ppp_register_channel(&ap->chan); if(err == 0) { @@ -672,12 +673,14 @@ dev_irnet_ioctl(struct inode * inode, } else DERROR(FS_ERROR, "Can't setup PPP channel...\n"); + unlock_kernel(); } else { /* In theory, should be N_TTY */ DEBUG(FS_INFO, "Exiting PPP discipline.\n"); /* Disconnect from the generic PPP layer */ + lock_kernel(); if(ap->ppp_open) { ap->ppp_open = 0; @@ -686,24 +689,20 @@ dev_irnet_ioctl(struct inode * inode, else DERROR(FS_ERROR, "Channel not registered !\n"); err = 0; + unlock_kernel(); } break; /* Query PPP channel and unit number */ case PPPIOCGCHAN: - if(!ap->ppp_open) - break; - if(put_user(ppp_channel_index(&ap->chan), (int __user *)argp)) - break; - DEBUG(FS_INFO, "Query channel.\n"); - err = 0; + if(ap->ppp_open && !put_user(ppp_channel_index(&ap->chan), + (int __user *)argp)) + err = 0; break; case PPPIOCGUNIT: - if(!ap->ppp_open) - break; - if(put_user(ppp_unit_number(&ap->chan), (int __user *)argp)) - break; - DEBUG(FS_INFO, "Query unit number.\n"); + lock_kernel(); + if(ap->ppp_open && !put_user(ppp_unit_number(&ap->chan), + (int __user *)argp)) err = 0; break; @@ -723,34 +722,39 @@ dev_irnet_ioctl(struct inode * inode, DEBUG(FS_INFO, "Standard PPP ioctl.\n"); if(!capable(CAP_NET_ADMIN)) err = -EPERM; - else + else { + lock_kernel(); err = ppp_irnet_ioctl(&ap->chan, cmd, arg); + unlock_kernel(); + } break; /* TTY IOCTLs : Pretend that we are a tty, to keep pppd happy */ /* Get termios */ case TCGETS: DEBUG(FS_INFO, "Get termios.\n"); + lock_kernel(); #ifndef TCGETS2 - if(kernel_termios_to_user_termios((struct termios __user *)argp, &ap->termios)) - break; + if(!kernel_termios_to_user_termios((struct termios __user *)argp, &ap->termios)) + err = 0; #else if(kernel_termios_to_user_termios_1((struct termios __user *)argp, &ap->termios)) - break; + err = 0; #endif - err = 0; + unlock_kernel(); break; /* Set termios */ case TCSETSF: DEBUG(FS_INFO, "Set termios.\n"); + lock_kernel(); #ifndef TCGETS2 - if(user_termios_to_kernel_termios(&ap->termios, (struct termios __user *)argp)) - break; + if(!user_termios_to_kernel_termios(&ap->termios, (struct termios __user *)argp)) + err = 0; #else - if(user_termios_to_kernel_termios_1(&ap->termios, (struct termios __user *)argp)) - break; + if(!user_termios_to_kernel_termios_1(&ap->termios, (struct termios __user *)argp)) + err = 0; #endif - err = 0; + unlock_kernel(); break; /* Set DTR/RTS */ @@ -773,7 +777,9 @@ dev_irnet_ioctl(struct inode * inode, * We should also worry that we don't accept junk here and that * we get rid of our own buffers */ #ifdef FLUSH_TO_PPP + lock_kernel(); ppp_output_wakeup(&ap->chan); + unlock_kernel(); #endif /* FLUSH_TO_PPP */ err = 0; break; @@ -788,7 +794,7 @@ dev_irnet_ioctl(struct inode * inode, default: DERROR(FS_ERROR, "Unsupported ioctl (0x%X)\n", cmd); - err = -ENOIOCTLCMD; + err = -ENOTTY; } DEXIT(FS_TRACE, " - err = 0x%X\n", err); diff --git a/net/irda/irnet/irnet_ppp.h b/net/irda/irnet/irnet_ppp.h index d2beb7df8f7f..d9f8bd4ebd05 100644 --- a/net/irda/irnet/irnet_ppp.h +++ b/net/irda/irnet/irnet_ppp.h @@ -76,9 +76,8 @@ static ssize_t static unsigned int dev_irnet_poll(struct file *, poll_table *); -static int - dev_irnet_ioctl(struct inode *, - struct file *, +static long + dev_irnet_ioctl(struct file *, unsigned int, unsigned long); /* ------------------------ PPP INTERFACE ------------------------ */ @@ -102,7 +101,7 @@ static struct file_operations irnet_device_fops = .read = dev_irnet_read, .write = dev_irnet_write, .poll = dev_irnet_poll, - .ioctl = dev_irnet_ioctl, + .unlocked_ioctl = dev_irnet_ioctl, .open = dev_irnet_open, .release = dev_irnet_close /* Also : llseek, readdir, mmap, flush, fsync, fasync, lock, readv, writev */ -- cgit v1.2.2 From 23c0752a25d73ccc4547700e8a57d5ae2f2edf56 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 29 May 2008 10:38:53 +0200 Subject: mac80211: clean up skb reallocation code This cleans up the skb reallocation code to avoid problems with skb->truesize, not resize an skb twice for a single output path because we didn't expand it enough during the first copy and also removes the code to further expand it during crypto operations which will no longer be necessary. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/tx.c | 97 ++++++++++++++++++++++++++++++++++++++---------------- net/mac80211/wep.c | 10 ++---- net/mac80211/wpa.c | 58 ++++++++++++++------------------ 3 files changed, 95 insertions(+), 70 deletions(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index baa1be0671e5..dac44cbd036f 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -1215,6 +1215,45 @@ retry: /* device xmit handlers */ +static int ieee80211_skb_resize(struct ieee80211_local *local, + struct sk_buff *skb, + int head_need, bool may_encrypt) +{ + int tail_need = 0; + + /* + * This could be optimised, devices that do full hardware + * crypto (including TKIP MMIC) need no tailroom... But we + * have no drivers for such devices currently. + */ + if (may_encrypt) { + tail_need = IEEE80211_ENCRYPT_TAILROOM; + tail_need -= skb_tailroom(skb); + tail_need = max_t(int, tail_need, 0); + } + + if (head_need || tail_need) { + /* Sorry. Can't account for this any more */ + skb_orphan(skb); + } + + if (skb_header_cloned(skb)) + I802_DEBUG_INC(local->tx_expand_skb_head_cloned); + else + I802_DEBUG_INC(local->tx_expand_skb_head); + + if (pskb_expand_head(skb, head_need, tail_need, GFP_ATOMIC)) { + printk(KERN_DEBUG "%s: failed to reallocate TX buffer\n", + wiphy_name(local->hw.wiphy)); + return -ENOMEM; + } + + /* update truesize too */ + skb->truesize += head_need + tail_need; + + return 0; +} + int ieee80211_master_start_xmit(struct sk_buff *skb, struct net_device *dev) { @@ -1222,6 +1261,7 @@ int ieee80211_master_start_xmit(struct sk_buff *skb, struct net_device *odev = NULL; struct ieee80211_sub_if_data *osdata; int headroom; + bool may_encrypt; int ret; if (info->control.ifindex) @@ -1241,13 +1281,18 @@ int ieee80211_master_start_xmit(struct sk_buff *skb, osdata = IEEE80211_DEV_TO_SUB_IF(odev); - headroom = osdata->local->tx_headroom + IEEE80211_ENCRYPT_HEADROOM; - if (skb_headroom(skb) < headroom) { - if (pskb_expand_head(skb, headroom, 0, GFP_ATOMIC)) { - dev_kfree_skb(skb); - dev_put(odev); - return 0; - } + may_encrypt = !(info->flags & IEEE80211_TX_CTL_DO_NOT_ENCRYPT); + + headroom = osdata->local->tx_headroom; + if (may_encrypt) + headroom += IEEE80211_ENCRYPT_HEADROOM; + headroom -= skb_headroom(skb); + headroom = max_t(int, 0, headroom); + + if (ieee80211_skb_resize(osdata->local, skb, headroom, may_encrypt)) { + dev_kfree_skb(skb); + dev_put(odev); + return 0; } info->control.vif = &osdata->vif; @@ -1509,32 +1554,26 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, * build in headroom in __dev_alloc_skb() (linux/skbuff.h) and * alloc_skb() (net/core/skbuff.c) */ - head_need = hdrlen + encaps_len + meshhdrlen + local->tx_headroom; - head_need -= skb_headroom(skb); + head_need = hdrlen + encaps_len + meshhdrlen - skb_headroom(skb); - /* We are going to modify skb data, so make a copy of it if happens to - * be cloned. This could happen, e.g., with Linux bridge code passing - * us broadcast frames. */ + /* + * So we need to modify the skb header and hence need a copy of + * that. The head_need variable above doesn't, so far, include + * the needed header space that we don't need right away. If we + * can, then we don't reallocate right now but only after the + * frame arrives at the master device (if it does...) + * + * If we cannot, however, then we will reallocate to include all + * the ever needed space. Also, if we need to reallocate it anyway, + * make it big enough for everything we may ever need. + */ if (head_need > 0 || skb_header_cloned(skb)) { -#if 0 - printk(KERN_DEBUG "%s: need to reallocate buffer for %d bytes " - "of headroom\n", dev->name, head_need); -#endif - - if (skb_header_cloned(skb)) - I802_DEBUG_INC(local->tx_expand_skb_head_cloned); - else - I802_DEBUG_INC(local->tx_expand_skb_head); - /* Since we have to reallocate the buffer, make sure that there - * is enough room for possible WEP IV/ICV and TKIP (8 bytes - * before payload and 12 after). */ - if (pskb_expand_head(skb, (head_need > 0 ? head_need + 8 : 8), - 12, GFP_ATOMIC)) { - printk(KERN_DEBUG "%s: failed to reallocate TX buffer" - "\n", dev->name); + head_need += IEEE80211_ENCRYPT_HEADROOM; + head_need += local->tx_headroom; + head_need = max_t(int, 0, head_need); + if (ieee80211_skb_resize(local, skb, head_need, true)) goto fail; - } } if (encaps_data) { diff --git a/net/mac80211/wep.c b/net/mac80211/wep.c index 1e7f03dd8f6b..c9fd1291b19d 100644 --- a/net/mac80211/wep.c +++ b/net/mac80211/wep.c @@ -93,13 +93,9 @@ static u8 *ieee80211_wep_add_iv(struct ieee80211_local *local, fc |= IEEE80211_FCTL_PROTECTED; hdr->frame_control = cpu_to_le16(fc); - if ((skb_headroom(skb) < WEP_IV_LEN || - skb_tailroom(skb) < WEP_ICV_LEN)) { - I802_DEBUG_INC(local->tx_expand_skb_head); - if (unlikely(pskb_expand_head(skb, WEP_IV_LEN, WEP_ICV_LEN, - GFP_ATOMIC))) - return NULL; - } + if (WARN_ON(skb_tailroom(skb) < WEP_ICV_LEN || + skb_headroom(skb) < WEP_IV_LEN)) + return NULL; hdrlen = ieee80211_get_hdrlen(fc); newhdr = skb_push(skb, WEP_IV_LEN); diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index d6635f6e5618..9f6fd20374e1 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -79,6 +79,7 @@ ieee80211_tx_h_michael_mic_add(struct ieee80211_tx_data *tx) struct sk_buff *skb = tx->skb; int authenticator; int wpa_test = 0; + int tail; fc = tx->fc; @@ -98,16 +99,13 @@ ieee80211_tx_h_michael_mic_add(struct ieee80211_tx_data *tx) return TX_CONTINUE; } - if (skb_tailroom(skb) < MICHAEL_MIC_LEN) { - I802_DEBUG_INC(tx->local->tx_expand_skb_head); - if (unlikely(pskb_expand_head(skb, TKIP_IV_LEN, - MICHAEL_MIC_LEN + TKIP_ICV_LEN, - GFP_ATOMIC))) { - printk(KERN_DEBUG "%s: failed to allocate more memory " - "for Michael MIC\n", tx->dev->name); - return TX_DROP; - } - } + tail = MICHAEL_MIC_LEN; + if (!(tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE)) + tail += TKIP_ICV_LEN; + + if (WARN_ON(skb_tailroom(skb) < tail || + skb_headroom(skb) < TKIP_IV_LEN)) + return TX_DROP; #if 0 authenticator = fc & IEEE80211_FCTL_FROMDS; /* FIX */ @@ -188,7 +186,7 @@ static int tkip_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; struct ieee80211_key *key = tx->key; struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); - int hdrlen, len, tailneed; + int hdrlen, len, tail; u16 fc; u8 *pos; @@ -199,7 +197,7 @@ static int tkip_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) !(tx->key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_IV)) { /* hwaccel - with no need for preallocated room for IV/ICV */ info->control.hw_key = &tx->key->conf; - return TX_CONTINUE; + return 0; } fc = le16_to_cpu(hdr->frame_control); @@ -207,17 +205,13 @@ static int tkip_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) len = skb->len - hdrlen; if (tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) - tailneed = 0; + tail = 0; else - tailneed = TKIP_ICV_LEN; - - if ((skb_headroom(skb) < TKIP_IV_LEN || - skb_tailroom(skb) < tailneed)) { - I802_DEBUG_INC(tx->local->tx_expand_skb_head); - if (unlikely(pskb_expand_head(skb, TKIP_IV_LEN, tailneed, - GFP_ATOMIC))) - return -1; - } + tail = TKIP_ICV_LEN; + + if (WARN_ON(skb_tailroom(skb) < tail || + skb_headroom(skb) < TKIP_IV_LEN)) + return -1; pos = skb_push(skb, TKIP_IV_LEN); memmove(pos, pos + TKIP_IV_LEN, hdrlen); @@ -432,7 +426,7 @@ static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; struct ieee80211_key *key = tx->key; struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); - int hdrlen, len, tailneed; + int hdrlen, len, tail; u16 fc; u8 *pos, *pn, *b_0, *aad, *scratch; int i; @@ -445,7 +439,7 @@ static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) /* hwaccel - with no need for preallocated room for CCMP " * header or MIC fields */ info->control.hw_key = &tx->key->conf; - return TX_CONTINUE; + return 0; } scratch = key->u.ccmp.tx_crypto_buf; @@ -457,17 +451,13 @@ static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) len = skb->len - hdrlen; if (key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) - tailneed = 0; + tail = 0; else - tailneed = CCMP_MIC_LEN; - - if ((skb_headroom(skb) < CCMP_HDR_LEN || - skb_tailroom(skb) < tailneed)) { - I802_DEBUG_INC(tx->local->tx_expand_skb_head); - if (unlikely(pskb_expand_head(skb, CCMP_HDR_LEN, tailneed, - GFP_ATOMIC))) - return -1; - } + tail = CCMP_MIC_LEN; + + if (WARN_ON(skb_tailroom(skb) < tail || + skb_headroom(skb) < CCMP_HDR_LEN)) + return -1; pos = skb_push(skb, CCMP_HDR_LEN); memmove(pos, pos + CCMP_HDR_LEN, hdrlen); -- cgit v1.2.2 From 747cf5e924a469a15a454b88a813236460b30975 Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Tue, 27 May 2008 17:50:51 +0300 Subject: mac80211: fix ieee80211_get_buffered_bc fix bss not initialized in ieee80211_get_buffered_bc and unbalanced locking Signed-off-by: Tomas Winkler Signed-off-by: Zhu Yi Signed-off-by: John W. Linville --- net/mac80211/tx.c | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index dac44cbd036f..16af30811f98 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -1947,7 +1947,7 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw, struct ieee80211_vif *vif) { struct ieee80211_local *local = hw_to_local(hw); - struct sk_buff *skb; + struct sk_buff *skb = NULL; struct sta_info *sta; ieee80211_tx_handler *handler; struct ieee80211_tx_data tx; @@ -1960,7 +1960,7 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw, sdata = vif_to_sdata(vif); bdev = sdata->dev; - + bss = &sdata->u.ap; if (!bss) return NULL; @@ -1968,19 +1968,16 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw, rcu_read_lock(); beacon = rcu_dereference(bss->beacon); - if (sdata->vif.type != IEEE80211_IF_TYPE_AP || !beacon || - !beacon->head) { - rcu_read_unlock(); - return NULL; - } + if (sdata->vif.type != IEEE80211_IF_TYPE_AP || !beacon || !beacon->head) + goto out; if (bss->dtim_count != 0) - return NULL; /* send buffered bc/mc only after DTIM beacon */ + goto out; /* send buffered bc/mc only after DTIM beacon */ while (1) { skb = skb_dequeue(&bss->ps_bc_buf); if (!skb) - return NULL; + goto out; local->total_ps_buffered--; if (!skb_queue_empty(&bss->ps_bc_buf) && skb->len >= 2) { @@ -2023,6 +2020,7 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw, skb = NULL; } +out: rcu_read_unlock(); return skb; -- cgit v1.2.2 From b83f4e15e65d94f6f56924b0b06a77a7ca2b4d8a Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Tue, 27 May 2008 17:50:52 +0300 Subject: mac80211: fix deadlock in sta->lock This patch fixes a deadlock of sta->lock use, occurring while changing tx aggregation states, as dev_queue_xmit end up in new function test_and_clear_sta_flags that uses that lock thus leading to deadlock Signed-off-by: Tomas Winkler Signed-off-by: Ron Rindjunsky Signed-off-by: John W. Linville --- net/mac80211/main.c | 30 +++++++++++++++++------------- 1 file changed, 17 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 97d4a537ca2f..f79f6b9938a6 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -589,8 +589,8 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) sta = sta_info_get(local, ra); if (!sta) { printk(KERN_DEBUG "Could not find the station\n"); - rcu_read_unlock(); - return -ENOENT; + ret = -ENOENT; + goto exit; } spin_lock_bh(&sta->lock); @@ -598,7 +598,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) /* we have tried too many times, receiver does not want A-MPDU */ if (sta->ampdu_mlme.addba_req_num[tid] > HT_AGG_MAX_RETRIES) { ret = -EBUSY; - goto start_ba_exit; + goto err_unlock_sta; } state = &sta->ampdu_mlme.tid_state_tx[tid]; @@ -609,7 +609,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) "idle on tid %u\n", tid); #endif /* CONFIG_MAC80211_HT_DEBUG */ ret = -EAGAIN; - goto start_ba_exit; + goto err_unlock_sta; } /* prepare A-MPDU MLME for Tx aggregation */ @@ -620,7 +620,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) printk(KERN_ERR "allocate tx mlme to tid %d failed\n", tid); ret = -ENOMEM; - goto start_ba_exit; + goto err_unlock_sta; } /* Tx timer */ sta->ampdu_mlme.tid_tx[tid]->addba_resp_timer.function = @@ -643,7 +643,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) printk(KERN_DEBUG "BA request denied - queue unavailable for" " tid %d\n", tid); #endif /* CONFIG_MAC80211_HT_DEBUG */ - goto start_ba_err; + goto err_unlock_queue; } sdata = sta->sdata; @@ -665,12 +665,13 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) " tid %d\n", tid); #endif /* CONFIG_MAC80211_HT_DEBUG */ *state = HT_AGG_STATE_IDLE; - goto start_ba_err; + goto err_unlock_queue; } /* Will put all the packets in the new SW queue */ ieee80211_requeue(local, ieee802_1d_to_ac[tid]); spin_unlock_bh(&local->mdev->queue_lock); + spin_unlock_bh(&sta->lock); /* send an addBA request */ sta->ampdu_mlme.dialog_token_allocator++; @@ -678,25 +679,26 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) sta->ampdu_mlme.dialog_token_allocator; sta->ampdu_mlme.tid_tx[tid]->ssn = start_seq_num; + ieee80211_send_addba_request(sta->sdata->dev, ra, tid, sta->ampdu_mlme.tid_tx[tid]->dialog_token, sta->ampdu_mlme.tid_tx[tid]->ssn, 0x40, 5000); - /* activate the timer for the recipient's addBA response */ sta->ampdu_mlme.tid_tx[tid]->addba_resp_timer.expires = jiffies + ADDBA_RESP_INTERVAL; add_timer(&sta->ampdu_mlme.tid_tx[tid]->addba_resp_timer); printk(KERN_DEBUG "activated addBA response timer on tid %d\n", tid); - goto start_ba_exit; + goto exit; -start_ba_err: +err_unlock_queue: kfree(sta->ampdu_mlme.tid_tx[tid]); sta->ampdu_mlme.tid_tx[tid] = NULL; spin_unlock_bh(&local->mdev->queue_lock); ret = -EBUSY; -start_ba_exit: +err_unlock_sta: spin_unlock_bh(&sta->lock); +exit: rcu_read_unlock(); return ret; } @@ -835,10 +837,11 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) } state = &sta->ampdu_mlme.tid_state_tx[tid]; - spin_lock_bh(&sta->lock); + /* NOTE: no need to use sta->lock in this state check, as + * ieee80211_stop_tx_ba_session will let only + * one stop call to pass through per sta/tid */ if ((*state & HT_AGG_STATE_REQ_STOP_BA_MSK) == 0) { printk(KERN_DEBUG "unexpected callback to A-MPDU stop\n"); - spin_unlock_bh(&sta->lock); rcu_read_unlock(); return; } @@ -861,6 +864,7 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) * ieee80211_wake_queue is not used here as this queue is not * necessarily stopped */ netif_schedule(local->mdev); + spin_lock_bh(&sta->lock); *state = HT_AGG_STATE_IDLE; sta->ampdu_mlme.addba_req_num[tid] = 0; kfree(sta->ampdu_mlme.tid_tx[tid]); -- cgit v1.2.2 From e623157b8d778a63736b0f41c04acc57c4f61ae0 Mon Sep 17 00:00:00 2001 From: Emmanuel Grumbach Date: Tue, 27 May 2008 20:00:11 +0300 Subject: mac80211: sends HT IE to user level through wext This patch adds HT IE in the scan list that is returned to user level through wext. This is useful to let wpa_supplicant if a bss supports 11n or not: WEP and TKIP are not supported in 11n. Signed-off-by: Emmanuel Grumbach Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 9a264379d7b1..6faa7006681a 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -4128,6 +4128,14 @@ ieee80211_sta_scan_result(struct net_device *dev, bss->rsn_ie); } + if (bss && bss->ht_ie) { + memset(&iwe, 0, sizeof(iwe)); + iwe.cmd = IWEVGENIE; + iwe.u.data.length = bss->ht_ie_len; + current_ev = iwe_stream_add_point(current_ev, end_buf, &iwe, + bss->ht_ie); + } + if (bss && bss->supp_rates_len > 0) { /* display all supported rates in readable format */ char *p = current_ev + IW_EV_LCP_LEN; -- cgit v1.2.2 From 9306102ea5696a3815f8d24ac0c0fbd1e19be7d3 Mon Sep 17 00:00:00 2001 From: Emmanuel Grumbach Date: Thu, 29 May 2008 16:35:23 +0800 Subject: mac80211: allow disable FAT in specific configurations This patch allows to disable FAT channel in specific configurations. For example the configuration (8, +1), (primary channel 8, extension channel 12) isn't permitted in U.S., but (8, -1), (primary channel 8, extension channel 4) is. When FAT channel configuration is not permitted, FAT channel should be reported as not supported in the capabilities of the HT IE in association request. And sssociation is performed on 20Mhz channel. Signed-off-by: Emmanuel Grumbach Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 2 ++ net/mac80211/mlme.c | 46 ++++++++++++++++++++++++++++++++++++++++++++-- 2 files changed, 46 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 3f8601cafffb..432011cd3647 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -92,6 +92,8 @@ struct ieee80211_sta_bss { size_t wmm_ie_len; u8 *ht_ie; size_t ht_ie_len; + u8 *ht_add_ie; + size_t ht_add_ie_len; #ifdef CONFIG_MAC80211_MESH u8 *mesh_id; size_t mesh_id_len; diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 6faa7006681a..d30c11337b04 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -808,8 +808,29 @@ static void ieee80211_send_assoc(struct net_device *dev, /* wmm support is a must to HT */ if (wmm && (ifsta->flags & IEEE80211_STA_WMM_ENABLED) && - sband->ht_info.ht_supported) { - __le16 tmp = cpu_to_le16(sband->ht_info.cap); + sband->ht_info.ht_supported && bss->ht_add_ie) { + struct ieee80211_ht_addt_info *ht_add_info = + (struct ieee80211_ht_addt_info *)bss->ht_add_ie; + u16 cap = sband->ht_info.cap; + __le16 tmp; + u32 flags = local->hw.conf.channel->flags; + + switch (ht_add_info->ht_param & IEEE80211_HT_IE_CHA_SEC_OFFSET) { + case IEEE80211_HT_IE_CHA_SEC_ABOVE: + if (flags & IEEE80211_CHAN_NO_FAT_ABOVE) { + cap &= ~IEEE80211_HT_CAP_SUP_WIDTH; + cap &= ~IEEE80211_HT_CAP_SGI_40; + } + break; + case IEEE80211_HT_IE_CHA_SEC_BELOW: + if (flags & IEEE80211_CHAN_NO_FAT_BELOW) { + cap &= ~IEEE80211_HT_CAP_SUP_WIDTH; + cap &= ~IEEE80211_HT_CAP_SGI_40; + } + break; + } + + tmp = cpu_to_le16(cap); pos = skb_put(skb, sizeof(struct ieee80211_ht_cap)+2); *pos++ = WLAN_EID_HT_CAPABILITY; *pos++ = sizeof(struct ieee80211_ht_cap); @@ -2264,6 +2285,7 @@ static void ieee80211_rx_bss_free(struct ieee80211_sta_bss *bss) kfree(bss->rsn_ie); kfree(bss->wmm_ie); kfree(bss->ht_ie); + kfree(bss->ht_add_ie); kfree(bss_mesh_id(bss)); kfree(bss_mesh_cfg(bss)); kfree(bss); @@ -2640,6 +2662,26 @@ static void ieee80211_rx_bss_info(struct net_device *dev, bss->ht_ie_len = 0; } + if (elems.ht_info_elem && + (!bss->ht_add_ie || + bss->ht_add_ie_len != elems.ht_info_elem_len || + memcmp(bss->ht_add_ie, elems.ht_info_elem, + elems.ht_info_elem_len))) { + kfree(bss->ht_add_ie); + bss->ht_add_ie = + kmalloc(elems.ht_info_elem_len + 2, GFP_ATOMIC); + if (bss->ht_add_ie) { + memcpy(bss->ht_add_ie, elems.ht_info_elem - 2, + elems.ht_info_elem_len + 2); + bss->ht_add_ie_len = elems.ht_info_elem_len + 2; + } else + bss->ht_add_ie_len = 0; + } else if (!elems.ht_info_elem && bss->ht_add_ie) { + kfree(bss->ht_add_ie); + bss->ht_add_ie = NULL; + bss->ht_add_ie_len = 0; + } + bss->beacon_int = le16_to_cpu(mgmt->u.beacon.beacon_int); bss->capability = le16_to_cpu(mgmt->u.beacon.capab_info); -- cgit v1.2.2 From 5854a32e6cb672d182ce378c69f0f7470137a062 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Mon, 2 Jun 2008 09:38:04 +0200 Subject: mac80211: fix rate control initialisation In commit 2e92e6f2c50b4baf85cca968f0e6f1b5c0df7d39 ("mac80211: use rate index in TX control") I forgot to initialise a few new variables to -1 which means that the rate control algorithm is never triggered and 0 is used as the only rate index, effectively fixing the transmit bitrate at the lowest supported. This patch adds the missing initialisation. Signed-off-by: Johannes Berg Bisected-by: Sitsofe Wheeler Signed-off-by: John W. Linville --- net/mac80211/tx.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 16af30811f98..332ddcb1068a 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -947,6 +947,8 @@ __ieee80211_tx_prepare(struct ieee80211_tx_data *tx, tx->local = local; tx->sdata = IEEE80211_DEV_TO_SUB_IF(dev); tx->channel = local->hw.conf.channel; + tx->rate_idx = -1; + tx->last_frag_rate_idx = -1; /* * Set this flag (used below to indicate "automatic fragmentation"), * it will be cleared/left by radiotap as desired. -- cgit v1.2.2 From 2b2121417eff64125bdb7f322d3b533e06d73dae Mon Sep 17 00:00:00 2001 From: Pavel Roskin Date: Mon, 2 Jun 2008 07:54:50 -0400 Subject: mac80211: fix panic when using hardware WEP e039fa4a4195ac4ee895e6f3d1334beed63256fe ("mac80211: move TX info into skb->cb") misplaced code for setting hardware WEP keys. Move it back. This fixes kernel panic in b43 if WEP is used and hardware encryption is enabled. Signed-off-by: Pavel Roskin Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/wep.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/wep.c b/net/mac80211/wep.c index c9fd1291b19d..e7b6344c900a 100644 --- a/net/mac80211/wep.c +++ b/net/mac80211/wep.c @@ -335,10 +335,10 @@ static int wep_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) info->control.icv_len = WEP_ICV_LEN; if (!(tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE)) { - info->control.hw_key = &tx->key->conf; if (ieee80211_wep_encrypt(tx->local, skb, tx->key)) return -1; } else { + info->control.hw_key = &tx->key->conf; if (tx->key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_IV) { if (!ieee80211_wep_add_iv(tx->local, skb, tx->key)) return -1; -- cgit v1.2.2 From b97e77e0446f0702de7fa0f5d2c52acf42d0289f Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Mon, 2 Jun 2008 20:31:56 +0300 Subject: mac80211: fix unbalanced locking in ieee80211_get_buffered_bc This patch fixes unbalanced locking in ieee80211_get_buffered_bc Signed-off-by: Tomas Winkler Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/tx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 332ddcb1068a..4214d039fbc6 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -2011,7 +2011,7 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw, } if (WARN_ON(tx.skb != skb)) - return NULL; + res = TX_DROP; if (res == TX_DROP) { I802_DEBUG_INC(local->tx_handlers_drop); -- cgit v1.2.2 From 2d892986e82306b8ad96285fb54b9999523331e0 Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Mon, 2 Jun 2008 20:31:57 +0300 Subject: mac80211: removing shadowed sband This patch removes doubly defined sband variable Signed-off-by: Tomas Winkler Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/tx.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 4214d039fbc6..1ad9e664f287 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -612,13 +612,10 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) if ((info->flags & IEEE80211_TX_CTL_USE_RTS_CTS) || (info->flags & IEEE80211_TX_CTL_USE_CTS_PROTECT)) { - struct ieee80211_supported_band *sband; struct ieee80211_rate *rate; s8 baserate = -1; int idx; - sband = tx->local->hw.wiphy->bands[tx->channel->band]; - /* Do not use multiple retry rates when using RTS/CTS */ info->control.alt_retry_rate_idx = -1; -- cgit v1.2.2 From 84b07c1638c36ae937d4930b467001a0d22904e5 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:28:21 -0700 Subject: tipc: Fix bug in connection setup via native API This patch fixes a bug that prevented TIPC from receiving a connection setup request message on a native TIPC port. The revised connection setup logic ensures that validation of the source of a connection-based message is skipped if the port is not yet connected to a peer. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/port.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/tipc/port.c b/net/tipc/port.c index 2c64ad88e3cf..0bd3e6192c43 100644 --- a/net/tipc/port.c +++ b/net/tipc/port.c @@ -842,13 +842,10 @@ static void port_dispatcher_sigh(void *dummy) tipc_port_unlock(p_ptr); if (unlikely(!connected)) { - if (unlikely(published)) + if (tipc_connect2port(dref, &orig)) goto reject; - tipc_connect2port(dref,&orig); - } - if (unlikely(msg_origport(msg) != peer_port)) - goto reject; - if (unlikely(msg_orignode(msg) != peer_node)) + } else if ((msg_origport(msg) != peer_port) || + (msg_orignode(msg) != peer_node)) goto reject; if (unlikely(!cb)) goto reject; -- cgit v1.2.2 From 5307e46957e76d71f02d2d736030ad92cdb3dd8c Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:28:45 -0700 Subject: tipc: Standardize error checking on incoming messages via native API This patch re-orders & re-groups the error checks performed on messages being delivered to native API ports, in order to clarify the similarities and differences required for the various message types. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/port.c | 25 +++++++++---------------- 1 file changed, 9 insertions(+), 16 deletions(-) (limited to 'net') diff --git a/net/tipc/port.c b/net/tipc/port.c index 0bd3e6192c43..4dfef9e798e3 100644 --- a/net/tipc/port.c +++ b/net/tipc/port.c @@ -841,14 +841,14 @@ static void port_dispatcher_sigh(void *dummy) u32 peer_node = port_peernode(p_ptr); tipc_port_unlock(p_ptr); + if (unlikely(!cb)) + goto reject; if (unlikely(!connected)) { if (tipc_connect2port(dref, &orig)) goto reject; } else if ((msg_origport(msg) != peer_port) || (msg_orignode(msg) != peer_node)) goto reject; - if (unlikely(!cb)) - goto reject; if (unlikely(++p_ptr->publ.conn_unacked >= TIPC_FLOW_CONTROL_WIN)) tipc_acknowledge(dref, @@ -862,9 +862,7 @@ static void port_dispatcher_sigh(void *dummy) tipc_msg_event cb = up_ptr->msg_cb; tipc_port_unlock(p_ptr); - if (unlikely(connected)) - goto reject; - if (unlikely(!cb)) + if (unlikely(!cb || connected)) goto reject; skb_pull(buf, msg_hdr_sz(msg)); cb(usr_handle, dref, &buf, msg_data(msg), @@ -877,11 +875,7 @@ static void port_dispatcher_sigh(void *dummy) tipc_named_msg_event cb = up_ptr->named_msg_cb; tipc_port_unlock(p_ptr); - if (unlikely(connected)) - goto reject; - if (unlikely(!cb)) - goto reject; - if (unlikely(!published)) + if (unlikely(!cb || connected || !published)) goto reject; dseq.type = msg_nametype(msg); dseq.lower = msg_nameinst(msg); @@ -908,11 +902,10 @@ err: u32 peer_node = port_peernode(p_ptr); tipc_port_unlock(p_ptr); - if (!connected || !cb) - break; - if (msg_origport(msg) != peer_port) + if (!cb || !connected) break; - if (msg_orignode(msg) != peer_node) + if ((msg_origport(msg) != peer_port) || + (msg_orignode(msg) != peer_node)) break; tipc_disconnect(dref); skb_pull(buf, msg_hdr_sz(msg)); @@ -924,7 +917,7 @@ err: tipc_msg_err_event cb = up_ptr->err_cb; tipc_port_unlock(p_ptr); - if (connected || !cb) + if (!cb || connected) break; skb_pull(buf, msg_hdr_sz(msg)); cb(usr_handle, dref, &buf, msg_data(msg), @@ -937,7 +930,7 @@ err: up_ptr->named_err_cb; tipc_port_unlock(p_ptr); - if (connected || !cb) + if (!cb || connected) break; dseq.type = msg_nametype(msg); dseq.lower = msg_nameinst(msg); -- cgit v1.2.2 From e0d4e3d0d72cfae7b7eac14e39e12dfc6b406313 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:29:09 -0700 Subject: tipc: Fix bugs in message error code display when debugging This patch corrects two problems in the display of error code information in TIPC messages when debugging: - no longer tries to display error code in NAME_DISTRIBUTOR messages, which don't have the error field - now displays error code in 24 byte data messages, which do have the error field Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/msg.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'net') diff --git a/net/tipc/msg.c b/net/tipc/msg.c index 38abebaae889..73dcd00d674e 100644 --- a/net/tipc/msg.c +++ b/net/tipc/msg.c @@ -230,13 +230,10 @@ void tipc_msg_dbg(struct print_buf *buf, struct tipc_msg *msg, const char *str) switch (usr) { case CONN_MANAGER: - case NAME_DISTRIBUTOR: case TIPC_LOW_IMPORTANCE: case TIPC_MEDIUM_IMPORTANCE: case TIPC_HIGH_IMPORTANCE: case TIPC_CRITICAL_IMPORTANCE: - if (msg_short(msg)) - break; /* No error */ switch (msg_errcode(msg)) { case TIPC_OK: break; -- cgit v1.2.2 From a686e6859e976712e28f6af927cd52a6a3bb372a Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:29:39 -0700 Subject: tipc: Fix minor bugs in link session number handling This patch introduces a new, out-of-range value to indicate that a link endpoint does not have an existing session established with its peer, eliminating the risk that the previously used "invalid session number" value (i.e. zero) might eventually be assigned as a valid session number and cause incorrect link behavior. The patch also introduces explicit bit masking when assigning a new link session number to ensure it does not exceed 16 bits. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/link.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/tipc/link.c b/net/tipc/link.c index bd206ebe4eea..b8c1231e3147 100644 --- a/net/tipc/link.c +++ b/net/tipc/link.c @@ -50,6 +50,12 @@ #include "bcast.h" +/* + * Out-of-range value for link session numbers + */ + +#define INVALID_SESSION 0x10000 + /* * Limit for deferred reception queue: */ @@ -464,7 +470,7 @@ struct link *tipc_link_create(struct bearer *b_ptr, const u32 peer, msg = l_ptr->pmsg; msg_init(msg, LINK_PROTOCOL, RESET_MSG, TIPC_OK, INT_H_SIZE, l_ptr->addr); msg_set_size(msg, sizeof(l_ptr->proto_msg)); - msg_set_session(msg, tipc_random); + msg_set_session(msg, (tipc_random & 0xffff)); msg_set_bearer_id(msg, b_ptr->identity); strcpy((char *)msg_data(msg), if_name); @@ -705,10 +711,10 @@ void tipc_link_reset(struct link *l_ptr) u32 checkpoint = l_ptr->next_in_no; int was_active_link = tipc_link_is_active(l_ptr); - msg_set_session(l_ptr->pmsg, msg_session(l_ptr->pmsg) + 1); + msg_set_session(l_ptr->pmsg, ((msg_session(l_ptr->pmsg) + 1) & 0xffff)); - /* Link is down, accept any session: */ - l_ptr->peer_session = 0; + /* Link is down, accept any session */ + l_ptr->peer_session = INVALID_SESSION; /* Prepare for max packet size negotiation */ link_init_max_pkt(l_ptr); @@ -2275,7 +2281,8 @@ static void link_recv_proto_msg(struct link *l_ptr, struct sk_buff *buf) switch (msg_type(msg)) { case RESET_MSG: - if (!link_working_unknown(l_ptr) && l_ptr->peer_session) { + if (!link_working_unknown(l_ptr) && + (l_ptr->peer_session != INVALID_SESSION)) { if (msg_session(msg) == l_ptr->peer_session) { dbg("Duplicate RESET: %u<->%u\n", msg_session(msg), l_ptr->peer_session); -- cgit v1.2.2 From 1265a02108c508b508112cdeac922aad03e0146a Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:32:35 -0700 Subject: tipc: Minor optimizations to received message processing This patch enhances TIPC's handler for incoming messages in two ways: - the trivial, single-use routine for processing non-sequenced messages has been merged into the main handler - the interface that received a message is now identified without having to access and/or modify the associated sk_buff Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/discover.c | 4 ++-- net/tipc/discover.h | 2 +- net/tipc/link.c | 27 +++++---------------------- 3 files changed, 8 insertions(+), 25 deletions(-) (limited to 'net') diff --git a/net/tipc/discover.c b/net/tipc/discover.c index faeaf06d377e..ada213aac4d4 100644 --- a/net/tipc/discover.c +++ b/net/tipc/discover.c @@ -156,11 +156,11 @@ static void disc_dupl_alert(struct bearer *b_ptr, u32 node_addr, /** * tipc_disc_recv_msg - handle incoming link setup message (request or response) * @buf: buffer containing message + * @b_ptr: bearer that message arrived on */ -void tipc_disc_recv_msg(struct sk_buff *buf) +void tipc_disc_recv_msg(struct sk_buff *buf, struct bearer *b_ptr) { - struct bearer *b_ptr = (struct bearer *)TIPC_SKB_CB(buf)->handle; struct link *link; struct tipc_media_addr media_addr; struct tipc_msg *msg = buf_msg(buf); diff --git a/net/tipc/discover.h b/net/tipc/discover.h index 9fd7587b143a..c36eaeb7d5d0 100644 --- a/net/tipc/discover.h +++ b/net/tipc/discover.h @@ -48,7 +48,7 @@ struct link_req *tipc_disc_init_link_req(struct bearer *b_ptr, void tipc_disc_update_link_req(struct link_req *req); void tipc_disc_stop_link_req(struct link_req *req); -void tipc_disc_recv_msg(struct sk_buff *buf); +void tipc_disc_recv_msg(struct sk_buff *buf, struct bearer *b_ptr); void tipc_disc_link_event(u32 addr, char *name, int up); #if 0 diff --git a/net/tipc/link.c b/net/tipc/link.c index b8c1231e3147..c62ebfea9304 100644 --- a/net/tipc/link.c +++ b/net/tipc/link.c @@ -1766,21 +1766,6 @@ void tipc_link_retransmit(struct link *l_ptr, struct sk_buff *buf, l_ptr->retransm_queue_head = l_ptr->retransm_queue_size = 0; } -/* - * link_recv_non_seq: Receive packets which are outside - * the link sequence flow - */ - -static void link_recv_non_seq(struct sk_buff *buf) -{ - struct tipc_msg *msg = buf_msg(buf); - - if (msg_user(msg) == LINK_CONFIG) - tipc_disc_recv_msg(buf); - else - tipc_bclink_recv_pkt(buf); -} - /** * link_insert_deferred_queue - insert deferred messages back into receive chain */ @@ -1857,7 +1842,7 @@ void tipc_recv_msg(struct sk_buff *head, struct tipc_bearer *tb_ptr) { read_lock_bh(&tipc_net_lock); while (head) { - struct bearer *b_ptr; + struct bearer *b_ptr = (struct bearer *)tb_ptr; struct node *n_ptr; struct link *l_ptr; struct sk_buff *crs; @@ -1868,9 +1853,6 @@ void tipc_recv_msg(struct sk_buff *head, struct tipc_bearer *tb_ptr) u32 released = 0; int type; - b_ptr = (struct bearer *)tb_ptr; - TIPC_SKB_CB(buf)->handle = b_ptr; - head = head->next; /* Ensure message is well-formed */ @@ -1889,7 +1871,10 @@ void tipc_recv_msg(struct sk_buff *head, struct tipc_bearer *tb_ptr) msg = buf_msg(buf); if (unlikely(msg_non_seq(msg))) { - link_recv_non_seq(buf); + if (msg_user(msg) == LINK_CONFIG) + tipc_disc_recv_msg(buf, b_ptr); + else + tipc_bclink_recv_pkt(buf); continue; } @@ -1996,8 +1981,6 @@ deliver: if (link_recv_changeover_msg(&l_ptr, &buf)) { msg = buf_msg(buf); seq_no = msg_seqno(msg); - TIPC_SKB_CB(buf)->handle - = b_ptr; if (type == ORIGINAL_MSG) goto deliver; goto protocol_check; -- cgit v1.2.2 From 9c396a7bfb4fe74e444be09069651280da520944 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:36:58 -0700 Subject: tipc: Prevent access of non-existent field in short message header This patch eliminates a case where TIPC's link code could try reading a field that is not present in a short message header. (The random value obtained was not being used, but the read operation could result in an invalid memory access exception in extremely rare circumstances.) Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/link.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/tipc/link.c b/net/tipc/link.c index c62ebfea9304..022cb2f107ac 100644 --- a/net/tipc/link.c +++ b/net/tipc/link.c @@ -2674,10 +2674,12 @@ int tipc_link_send_long_buf(struct link *l_ptr, struct sk_buff *buf) u32 pack_sz = link_max_pkt(l_ptr); u32 fragm_sz = pack_sz - INT_H_SIZE; u32 fragm_no = 1; - u32 destaddr = msg_destnode(inmsg); + u32 destaddr; if (msg_short(inmsg)) destaddr = l_ptr->addr; + else + destaddr = msg_destnode(inmsg); if (msg_routed(inmsg)) msg_set_prevnode(inmsg, tipc_own_addr); -- cgit v1.2.2 From 757152175666681d54d370500e41a756cfedd4fc Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:37:34 -0700 Subject: tipc: Optimize message initialization routine This patch eliminates the rarely-used "error code" argument when initializing a TIPC message header, since the default value of zero is the desired result in most cases; the few exceptional cases now set the error code explicitly. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/bcast.c | 2 +- net/tipc/cluster.c | 2 +- net/tipc/discover.c | 3 +-- net/tipc/link.c | 12 ++++++------ net/tipc/msg.h | 3 +-- net/tipc/name_distr.c | 3 +-- net/tipc/port.c | 9 +++++---- 7 files changed, 16 insertions(+), 18 deletions(-) (limited to 'net') diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c index e7880172ef19..0052c0747d05 100644 --- a/net/tipc/bcast.c +++ b/net/tipc/bcast.c @@ -276,7 +276,7 @@ static void bclink_send_nack(struct node *n_ptr) if (buf) { msg = buf_msg(buf); msg_init(msg, BCAST_PROTOCOL, STATE_MSG, - TIPC_OK, INT_H_SIZE, n_ptr->addr); + INT_H_SIZE, n_ptr->addr); msg_set_mc_netid(msg, tipc_net_id); msg_set_bcast_ack(msg, mod(n_ptr->bclink.last_in)); msg_set_bcgap_after(msg, n_ptr->bclink.gap_after); diff --git a/net/tipc/cluster.c b/net/tipc/cluster.c index 4bb3404f610b..bc1db474fe01 100644 --- a/net/tipc/cluster.c +++ b/net/tipc/cluster.c @@ -238,7 +238,7 @@ static struct sk_buff *tipc_cltr_prepare_routing_msg(u32 data_size, u32 dest) if (buf) { msg = buf_msg(buf); memset((char *)msg, 0, size); - msg_init(msg, ROUTE_DISTRIBUTOR, 0, TIPC_OK, INT_H_SIZE, dest); + msg_init(msg, ROUTE_DISTRIBUTOR, 0, INT_H_SIZE, dest); } return buf; } diff --git a/net/tipc/discover.c b/net/tipc/discover.c index ada213aac4d4..64c20284e0f7 100644 --- a/net/tipc/discover.c +++ b/net/tipc/discover.c @@ -120,8 +120,7 @@ static struct sk_buff *tipc_disc_init_msg(u32 type, if (buf) { msg = buf_msg(buf); - msg_init(msg, LINK_CONFIG, type, TIPC_OK, DSC_H_SIZE, - dest_domain); + msg_init(msg, LINK_CONFIG, type, DSC_H_SIZE, dest_domain); msg_set_non_seq(msg); msg_set_req_links(msg, req_links); msg_set_dest_domain(msg, dest_domain); diff --git a/net/tipc/link.c b/net/tipc/link.c index 022cb2f107ac..9784a8e963b4 100644 --- a/net/tipc/link.c +++ b/net/tipc/link.c @@ -468,7 +468,7 @@ struct link *tipc_link_create(struct bearer *b_ptr, const u32 peer, l_ptr->pmsg = (struct tipc_msg *)&l_ptr->proto_msg; msg = l_ptr->pmsg; - msg_init(msg, LINK_PROTOCOL, RESET_MSG, TIPC_OK, INT_H_SIZE, l_ptr->addr); + msg_init(msg, LINK_PROTOCOL, RESET_MSG, INT_H_SIZE, l_ptr->addr); msg_set_size(msg, sizeof(l_ptr->proto_msg)); msg_set_session(msg, (tipc_random & 0xffff)); msg_set_bearer_id(msg, b_ptr->identity); @@ -1128,7 +1128,7 @@ int tipc_link_send_buf(struct link *l_ptr, struct sk_buff *buf) if (bundler) { msg_init(&bundler_hdr, MSG_BUNDLER, OPEN_MSG, - TIPC_OK, INT_H_SIZE, l_ptr->addr); + INT_H_SIZE, l_ptr->addr); skb_copy_to_linear_data(bundler, &bundler_hdr, INT_H_SIZE); skb_trim(bundler, INT_H_SIZE); @@ -1392,7 +1392,7 @@ again: msg_dbg(hdr, ">FRAGMENTING>"); msg_init(&fragm_hdr, MSG_FRAGMENTER, FIRST_FRAGMENT, - TIPC_OK, INT_H_SIZE, msg_destnode(hdr)); + INT_H_SIZE, msg_destnode(hdr)); msg_set_link_selector(&fragm_hdr, sender->publ.ref); msg_set_size(&fragm_hdr, max_pkt); msg_set_fragm_no(&fragm_hdr, 1); @@ -2426,7 +2426,7 @@ void tipc_link_changeover(struct link *l_ptr) } msg_init(&tunnel_hdr, CHANGEOVER_PROTOCOL, - ORIGINAL_MSG, TIPC_OK, INT_H_SIZE, l_ptr->addr); + ORIGINAL_MSG, INT_H_SIZE, l_ptr->addr); msg_set_bearer_id(&tunnel_hdr, l_ptr->peer_bearer_id); msg_set_msgcnt(&tunnel_hdr, msgcount); dbg("Link changeover requires %u tunnel messages\n", msgcount); @@ -2481,7 +2481,7 @@ void tipc_link_send_duplicate(struct link *l_ptr, struct link *tunnel) struct tipc_msg tunnel_hdr; msg_init(&tunnel_hdr, CHANGEOVER_PROTOCOL, - DUPLICATE_MSG, TIPC_OK, INT_H_SIZE, l_ptr->addr); + DUPLICATE_MSG, INT_H_SIZE, l_ptr->addr); msg_set_msgcnt(&tunnel_hdr, l_ptr->out_queue_size); msg_set_bearer_id(&tunnel_hdr, l_ptr->peer_bearer_id); iter = l_ptr->first_out; @@ -2687,7 +2687,7 @@ int tipc_link_send_long_buf(struct link *l_ptr, struct sk_buff *buf) /* Prepare reusable fragment header: */ msg_init(&fragm_hdr, MSG_FRAGMENTER, FIRST_FRAGMENT, - TIPC_OK, INT_H_SIZE, destaddr); + INT_H_SIZE, destaddr); msg_set_link_selector(&fragm_hdr, msg_link_selector(inmsg)); msg_set_long_msgno(&fragm_hdr, mod(l_ptr->long_msg_seq_no++)); msg_set_fragm_no(&fragm_hdr, fragm_no); diff --git a/net/tipc/msg.h b/net/tipc/msg.h index ad487e8abcc2..3418ffa72c94 100644 --- a/net/tipc/msg.h +++ b/net/tipc/msg.h @@ -696,7 +696,7 @@ static inline u32 msg_tot_importance(struct tipc_msg *m) static inline void msg_init(struct tipc_msg *m, u32 user, u32 type, - u32 err, u32 hsize, u32 destnode) + u32 hsize, u32 destnode) { memset(m, 0, hsize); msg_set_version(m); @@ -705,7 +705,6 @@ static inline void msg_init(struct tipc_msg *m, u32 user, u32 type, msg_set_size(m, hsize); msg_set_prevnode(m, tipc_own_addr); msg_set_type(m, type); - msg_set_errcode(m, err); if (!msg_short(m)) { msg_set_orignode(m, tipc_own_addr); msg_set_destnode(m, destnode); diff --git a/net/tipc/name_distr.c b/net/tipc/name_distr.c index aecba5cd87d6..10a69894e2fd 100644 --- a/net/tipc/name_distr.c +++ b/net/tipc/name_distr.c @@ -103,8 +103,7 @@ static struct sk_buff *named_prepare_buf(u32 type, u32 size, u32 dest) if (buf != NULL) { msg = buf_msg(buf); - msg_init(msg, NAME_DISTRIBUTOR, type, TIPC_OK, - LONG_H_SIZE, dest); + msg_init(msg, NAME_DISTRIBUTOR, type, LONG_H_SIZE, dest); msg_set_size(msg, LONG_H_SIZE + size); } return buf; diff --git a/net/tipc/port.c b/net/tipc/port.c index 4dfef9e798e3..e3e9c121afbe 100644 --- a/net/tipc/port.c +++ b/net/tipc/port.c @@ -244,8 +244,7 @@ u32 tipc_createport_raw(void *usr_handle, p_ptr->publ.max_pkt = MAX_PKT_DEFAULT; p_ptr->publ.ref = ref; msg = &p_ptr->publ.phdr; - msg_init(msg, TIPC_LOW_IMPORTANCE, TIPC_NAMED_MSG, TIPC_OK, LONG_H_SIZE, - 0); + msg_init(msg, TIPC_LOW_IMPORTANCE, TIPC_NAMED_MSG, LONG_H_SIZE, 0); msg_set_orignode(msg, tipc_own_addr); msg_set_prevnode(msg, tipc_own_addr); msg_set_origport(msg, ref); @@ -404,7 +403,8 @@ static struct sk_buff *port_build_proto_msg(u32 destport, u32 destnode, buf = buf_acquire(LONG_H_SIZE); if (buf) { msg = buf_msg(buf); - msg_init(msg, usr, type, err, LONG_H_SIZE, destnode); + msg_init(msg, usr, type, LONG_H_SIZE, destnode); + msg_set_errcode(msg, err); msg_set_destport(msg, destport); msg_set_origport(msg, origport); msg_set_destnode(msg, destnode); @@ -448,7 +448,8 @@ int tipc_reject_msg(struct sk_buff *buf, u32 err) return data_sz; } rmsg = buf_msg(rbuf); - msg_init(rmsg, imp, msg_type(msg), err, hdr_sz, msg_orignode(msg)); + msg_init(rmsg, imp, msg_type(msg), hdr_sz, msg_orignode(msg)); + msg_set_errcode(rmsg, err); msg_set_destport(rmsg, msg_origport(msg)); msg_set_prevnode(rmsg, tipc_own_addr); msg_set_origport(rmsg, msg_destport(msg)); -- cgit v1.2.2 From 0f15d36453bbd02d404445dace49f4ae072f44e5 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:37:59 -0700 Subject: tipc: Prevent display of name table types with no publications This patch adds a check to prevent TIPC's name table display code from listing a name type entry if it exists only to hold subscription info, rather than published names. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/name_table.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'net') diff --git a/net/tipc/name_table.c b/net/tipc/name_table.c index 892373e498e4..4455f13cfaf6 100644 --- a/net/tipc/name_table.c +++ b/net/tipc/name_table.c @@ -905,6 +905,9 @@ static void nameseq_list(struct name_seq *seq, struct print_buf *buf, u32 depth, struct sub_seq *sseq; char typearea[11]; + if (seq->first_free == 0) + return; + sprintf(typearea, "%-10u", seq->type); if (depth == 1) { -- cgit v1.2.2 From 307fdf5e7defcacf84db21cda18770b2bf5f5196 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:38:22 -0700 Subject: tipc: Add missing spinlock in name table display code This patch ensures that the display code that traverses the publication lists belonging to a name table entry take its associated spinlock, to protect against a possible change to one of its "head of list" pointers caused by a simultaneous name table lookup operation by another thread of control. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/name_table.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/tipc/name_table.c b/net/tipc/name_table.c index 4455f13cfaf6..096f7bd240a0 100644 --- a/net/tipc/name_table.c +++ b/net/tipc/name_table.c @@ -74,7 +74,7 @@ struct sub_seq { * @first_free: array index of first unused sub-sequence entry * @ns_list: links to adjacent name sequences in hash chain * @subscriptions: list of subscriptions for this 'type' - * @lock: spinlock controlling access to name sequence structure + * @lock: spinlock controlling access to publication lists of all sub-sequences */ struct name_seq { @@ -918,7 +918,9 @@ static void nameseq_list(struct name_seq *seq, struct print_buf *buf, u32 depth, for (sseq = seq->sseqs; sseq != &seq->sseqs[seq->first_free]; sseq++) { if ((lowbound <= sseq->upper) && (upbound >= sseq->lower)) { tipc_printf(buf, "%s ", typearea); + spin_lock_bh(&seq->lock); subseq_list(sseq, buf, depth, index); + spin_unlock_bh(&seq->lock); sprintf(typearea, "%10s", " "); } } -- cgit v1.2.2 From bd7845337b105e090dd18912d511139945fa7586 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:47:30 -0700 Subject: tipc: Expand link sequence gap field to 13 bits This patch increases the "sequence gap" field of the LINK_PROTOCOL message header from 8 bits to 13 bits (utilizing 5 previously unused 0 bits). This ensures that the field is big enough to indicate the loss of up to 8191 consecutive messages on the link, thereby accommodating the current worst-case scenario of 4000 lost messages. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/msg.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/tipc/msg.h b/net/tipc/msg.h index 3418ffa72c94..b23619fab968 100644 --- a/net/tipc/msg.h +++ b/net/tipc/msg.h @@ -2,7 +2,7 @@ * net/tipc/msg.h: Include file for TIPC message header routines * * Copyright (c) 2000-2007, Ericsson AB - * Copyright (c) 2005-2007, Wind River Systems + * Copyright (c) 2005-2008, Wind River Systems * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -325,7 +325,7 @@ static inline struct tipc_msg *msg_get_wrapped(struct tipc_msg *m) +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ w0:|vers |msg usr|hdr sz |n|resrv| packet size | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - w1:|m typ|rsv=0| sequence gap | broadcast ack no | + w1:|m typ| sequence gap | broadcast ack no | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ w2:| link level ack no/bc_gap_from | seq no / bcast_gap_to | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ @@ -388,12 +388,12 @@ static inline struct tipc_msg *msg_get_wrapped(struct tipc_msg *m) static inline u32 msg_seq_gap(struct tipc_msg *m) { - return msg_bits(m, 1, 16, 0xff); + return msg_bits(m, 1, 16, 0x1fff); } static inline void msg_set_seq_gap(struct tipc_msg *m, u32 n) { - msg_set_bits(m, 1, 16, 0xff, n); + msg_set_bits(m, 1, 16, 0x1fff, n); } static inline u32 msg_req_links(struct tipc_msg *m) -- cgit v1.2.2 From 9bef54383d16568da19cfe46bdc52cdedb9bb8da Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:47:55 -0700 Subject: tipc: Message header creation optimizations This patch eliminates several cases where message header fields were being set to the same value twice. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/port.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) (limited to 'net') diff --git a/net/tipc/port.c b/net/tipc/port.c index e3e9c121afbe..93014f9bc95f 100644 --- a/net/tipc/port.c +++ b/net/tipc/port.c @@ -244,11 +244,8 @@ u32 tipc_createport_raw(void *usr_handle, p_ptr->publ.max_pkt = MAX_PKT_DEFAULT; p_ptr->publ.ref = ref; msg = &p_ptr->publ.phdr; - msg_init(msg, TIPC_LOW_IMPORTANCE, TIPC_NAMED_MSG, LONG_H_SIZE, 0); - msg_set_orignode(msg, tipc_own_addr); - msg_set_prevnode(msg, tipc_own_addr); + msg_init(msg, importance, TIPC_NAMED_MSG, LONG_H_SIZE, 0); msg_set_origport(msg, ref); - msg_set_importance(msg,importance); p_ptr->last_in_seqno = 41; p_ptr->sent = 1; INIT_LIST_HEAD(&p_ptr->wait_list); @@ -407,7 +404,6 @@ static struct sk_buff *port_build_proto_msg(u32 destport, u32 destnode, msg_set_errcode(msg, err); msg_set_destport(msg, destport); msg_set_origport(msg, origport); - msg_set_destnode(msg, destnode); msg_set_orignode(msg, orignode); msg_set_transp_seqno(msg, seqno); msg_set_msgcnt(msg, ack); @@ -451,7 +447,6 @@ int tipc_reject_msg(struct sk_buff *buf, u32 err) msg_init(rmsg, imp, msg_type(msg), hdr_sz, msg_orignode(msg)); msg_set_errcode(rmsg, err); msg_set_destport(rmsg, msg_origport(msg)); - msg_set_prevnode(rmsg, tipc_own_addr); msg_set_origport(rmsg, msg_destport(msg)); if (msg_short(msg)) msg_set_orignode(rmsg, tipc_own_addr); -- cgit v1.2.2 From 99c145939bc1f65f9b946f2b9dd7bfc1f44783d6 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:48:25 -0700 Subject: tipc: Fix bugs in rejection of message with short header This patch ensures that TIPC doesn't try to access non-existent message header fields when rejecting a message with a short header. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/port.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/tipc/port.c b/net/tipc/port.c index 93014f9bc95f..2e0cff408ff9 100644 --- a/net/tipc/port.c +++ b/net/tipc/port.c @@ -448,13 +448,15 @@ int tipc_reject_msg(struct sk_buff *buf, u32 err) msg_set_errcode(rmsg, err); msg_set_destport(rmsg, msg_origport(msg)); msg_set_origport(rmsg, msg_destport(msg)); - if (msg_short(msg)) + if (msg_short(msg)) { msg_set_orignode(rmsg, tipc_own_addr); - else + /* leave name type & instance as zeroes */ + } else { msg_set_orignode(rmsg, msg_destnode(msg)); + msg_set_nametype(rmsg, msg_nametype(msg)); + msg_set_nameinst(rmsg, msg_nameinst(msg)); + } msg_set_size(rmsg, data_sz + hdr_sz); - msg_set_nametype(rmsg, msg_nametype(msg)); - msg_set_nameinst(rmsg, msg_nameinst(msg)); skb_copy_to_linear_data_offset(rbuf, hdr_sz, msg_data(msg), data_sz); /* send self-abort message when rejecting on a connected port */ -- cgit v1.2.2 From 40aecb1b13f50d96616abb612c17e59457f54263 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Wed, 4 Jun 2008 17:54:48 -0700 Subject: tipc: Message rejection rework preparatory changes This patch defines a few new message header manipulation routines, and generalizes the usefulness of another, in preparation for upcoming rework of TIPC's message rejection code. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/bcast.c | 2 +- net/tipc/discover.c | 2 +- net/tipc/msg.h | 31 +++++++++++++++++++++++++++++-- 3 files changed, 31 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c index 0052c0747d05..a5883b1452ff 100644 --- a/net/tipc/bcast.c +++ b/net/tipc/bcast.c @@ -571,7 +571,7 @@ static int tipc_bcbearer_send(struct sk_buff *buf, assert(tipc_cltr_bcast_nodes.count != 0); bcbuf_set_acks(buf, tipc_cltr_bcast_nodes.count); msg = buf_msg(buf); - msg_set_non_seq(msg); + msg_set_non_seq(msg, 1); msg_set_mc_netid(msg, tipc_net_id); } diff --git a/net/tipc/discover.c b/net/tipc/discover.c index 64c20284e0f7..1657f0e795ff 100644 --- a/net/tipc/discover.c +++ b/net/tipc/discover.c @@ -121,7 +121,7 @@ static struct sk_buff *tipc_disc_init_msg(u32 type, if (buf) { msg = buf_msg(buf); msg_init(msg, LINK_CONFIG, type, DSC_H_SIZE, dest_domain); - msg_set_non_seq(msg); + msg_set_non_seq(msg, 1); msg_set_req_links(msg, req_links); msg_set_dest_domain(msg, dest_domain); msg_set_bc_netid(msg, tipc_net_id); diff --git a/net/tipc/msg.h b/net/tipc/msg.h index b23619fab968..7ee6ae238147 100644 --- a/net/tipc/msg.h +++ b/net/tipc/msg.h @@ -75,6 +75,14 @@ static inline void msg_set_bits(struct tipc_msg *m, u32 w, m->hdr[w] |= htonl(val); } +static inline void msg_swap_words(struct tipc_msg *msg, u32 a, u32 b) +{ + u32 temp = msg->hdr[a]; + + msg->hdr[a] = msg->hdr[b]; + msg->hdr[b] = temp; +} + /* * Word 0 */ @@ -119,9 +127,9 @@ static inline int msg_non_seq(struct tipc_msg *m) return msg_bits(m, 0, 20, 1); } -static inline void msg_set_non_seq(struct tipc_msg *m) +static inline void msg_set_non_seq(struct tipc_msg *m, u32 n) { - msg_set_bits(m, 0, 20, 1, 1); + msg_set_bits(m, 0, 20, 1, n); } static inline int msg_dest_droppable(struct tipc_msg *m) @@ -224,6 +232,25 @@ static inline void msg_set_seqno(struct tipc_msg *m, u32 n) msg_set_bits(m, 2, 0, 0xffff, n); } +/* + * TIPC may utilize the "link ack #" and "link seq #" fields of a short + * message header to hold the destination node for the message, since the + * normal "dest node" field isn't present. This cache is only referenced + * when required, so populating the cache of a longer message header is + * harmless (as long as the header has the two link sequence fields present). + * + * Note: Host byte order is OK here, since the info never goes off-card. + */ + +static inline u32 msg_destnode_cache(struct tipc_msg *m) +{ + return m->hdr[2]; +} + +static inline void msg_set_destnode_cache(struct tipc_msg *m, u32 dnode) +{ + m->hdr[2] = dnode; +} /* * Words 3-10 -- cgit v1.2.2 From 9457afee85e0dfc2b5075a391d6f34463b4c2b90 Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Thu, 5 Jun 2008 11:23:39 -0700 Subject: netlink: Remove nonblock parameter from netlink_attachskb Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/netlink/af_netlink.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/netlink/af_netlink.c b/net/netlink/af_netlink.c index 9b97f8006c9c..6507c02dbe0d 100644 --- a/net/netlink/af_netlink.c +++ b/net/netlink/af_netlink.c @@ -759,7 +759,7 @@ struct sock *netlink_getsockbyfilp(struct file *filp) * 0: continue * 1: repeat lookup - reference dropped while waiting for socket memory. */ -int netlink_attachskb(struct sock *sk, struct sk_buff *skb, int nonblock, +int netlink_attachskb(struct sock *sk, struct sk_buff *skb, long *timeo, struct sock *ssk) { struct netlink_sock *nlk; @@ -892,7 +892,7 @@ retry: return err; } - err = netlink_attachskb(sk, skb, nonblock, &timeo, ssk); + err = netlink_attachskb(sk, skb, &timeo, ssk); if (err == 1) goto retry; if (err) -- cgit v1.2.2 From 7bfe8bdb80d5988483b01177b09b9c8d1605dffb Mon Sep 17 00:00:00 2001 From: Vlad Yasevich Date: Mon, 9 Jun 2008 15:45:05 -0700 Subject: sctp: Fix problems with the new SCTP_DELAYED_ACK code The default sack frequency should be 2. Also fix copy/paste error when updating all transports. Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/socket.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/sctp/socket.c b/net/sctp/socket.c index 81600eea05d2..253e5ea7e1e8 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -2471,7 +2471,7 @@ static int sctp_setsockopt_delayed_ack(struct sock *sk, (trans->param_flags & ~SPP_SACKDELAY) | SPP_SACKDELAY_ENABLE; } - if (params.sack_delay == 1) { + if (params.sack_freq == 1) { trans->param_flags = (trans->param_flags & ~SPP_SACKDELAY) | SPP_SACKDELAY_DISABLE; @@ -3536,7 +3536,7 @@ SCTP_STATIC int sctp_init_sock(struct sock *sk) sp->pathmaxrxt = sctp_max_retrans_path; sp->pathmtu = 0; // allow default discovery sp->sackdelay = sctp_sack_timeout; - sp->sackfreq = 3; + sp->sackfreq = 2; sp->param_flags = SPP_HB_ENABLE | SPP_PMTUD_ENABLE | SPP_SACKDELAY_ENABLE; -- cgit v1.2.2 From f1494ed1d318542baa9480cfd44d040a92635129 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Mon, 9 Jun 2008 15:49:57 -0700 Subject: iucv: fix section mismatch warning. WARNING: net/iucv/built-in.o(.exit.text+0x9c): Section mismatch in reference from the function iucv_exit() to the variable .cpuinit.data:iucv_cpu_notifier This warning is caused by a reference from unregister_hotcpu_notifier() from an exit function to a cpuinitdata annotated data structurre. This is a false positive warning since for the non CPU_HOTPLUG case unregister_hotcpu_notifier() is a nop. Use __refdata instead of __cpuinitdata to get rid of the warning. Signed-off-by: Heiko Carstens Signed-off-by: Ursula Braun Signed-off-by: David S. Miller --- net/iucv/iucv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/iucv/iucv.c b/net/iucv/iucv.c index 918970762131..2d43175b0cdb 100644 --- a/net/iucv/iucv.c +++ b/net/iucv/iucv.c @@ -598,7 +598,7 @@ static int __cpuinit iucv_cpu_notify(struct notifier_block *self, return NOTIFY_OK; } -static struct notifier_block __cpuinitdata iucv_cpu_notifier = { +static struct notifier_block __refdata iucv_cpu_notifier = { .notifier_call = iucv_cpu_notify, }; -- cgit v1.2.2 From 7b9d1b22a382aa221018c19880ee22c44467feec Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Mon, 9 Jun 2008 15:50:30 -0700 Subject: iucv: prevent cpu hotplug when walking cpu_online_map. The code used preempt_disable() to prevent cpu hotplug, however that doesn't protect for cpus being added. So use get_online_cpus() instead. Signed-off-by: Heiko Carstens Signed-off-by: Martin Schwidefsky Signed-off-by: Ursula Braun Signed-off-by: David S. Miller --- net/iucv/iucv.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/iucv/iucv.c b/net/iucv/iucv.c index 2d43175b0cdb..531a206ce7a6 100644 --- a/net/iucv/iucv.c +++ b/net/iucv/iucv.c @@ -474,14 +474,14 @@ static void iucv_setmask_mp(void) { int cpu; - preempt_disable(); + get_online_cpus(); for_each_online_cpu(cpu) /* Enable all cpus with a declared buffer. */ if (cpu_isset(cpu, iucv_buffer_cpumask) && !cpu_isset(cpu, iucv_irq_cpumask)) smp_call_function_single(cpu, iucv_allow_cpu, NULL, 0, 1); - preempt_enable(); + put_online_cpus(); } /** @@ -521,16 +521,17 @@ static int iucv_enable(void) goto out; /* Declare per cpu buffers. */ rc = -EIO; - preempt_disable(); + get_online_cpus(); for_each_online_cpu(cpu) smp_call_function_single(cpu, iucv_declare_cpu, NULL, 0, 1); - preempt_enable(); if (cpus_empty(iucv_buffer_cpumask)) /* No cpu could declare an iucv buffer. */ goto out_path; + put_online_cpus(); return 0; out_path: + put_online_cpus(); kfree(iucv_path_table); out: return rc; @@ -545,7 +546,9 @@ out: */ static void iucv_disable(void) { + get_online_cpus(); on_each_cpu(iucv_retrieve_cpu, NULL, 0, 1); + put_online_cpus(); kfree(iucv_path_table); } -- cgit v1.2.2 From 469689a4dd476c1be6750deea5f59528a17b8b4a Mon Sep 17 00:00:00 2001 From: Ursula Braun Date: Mon, 9 Jun 2008 15:51:03 -0700 Subject: af_iucv: exploit target message class support of IUCV The first 4 bytes of data to be sent are stored additionally into the message class field of the send request. A receiving target program (not an af_iucv socket program) can make use of this information to pre-screen incoming messages. Signed-off-by: Ursula Braun Signed-off-by: David S. Miller --- net/iucv/af_iucv.c | 1 + 1 file changed, 1 insertion(+) (limited to 'net') diff --git a/net/iucv/af_iucv.c b/net/iucv/af_iucv.c index 7b0038f45b16..58e4aee3e696 100644 --- a/net/iucv/af_iucv.c +++ b/net/iucv/af_iucv.c @@ -644,6 +644,7 @@ static int iucv_sock_sendmsg(struct kiocb *iocb, struct socket *sock, } txmsg.class = 0; + memcpy(&txmsg.class, skb->data, skb->len >= 4 ? 4 : skb->len); txmsg.tag = iucv->send_tag++; memcpy(skb->cb, &txmsg.tag, 4); skb_queue_tail(&iucv->send_skb_q, skb); -- cgit v1.2.2 From 93f65158723ceb7078ee9a0fd4830c0de00f4b9e Mon Sep 17 00:00:00 2001 From: Kuo-lang Tseng Date: Mon, 9 Jun 2008 15:55:45 -0700 Subject: netfilter: ebtables: add IPv6 support It implements matching functions for IPv6 address & traffic class (merged from the patch sent by Jan Engelhardt [jengelh@computergmbh.de] http://marc.info/?l=netfilter-devel&m=120182168424052&w=2), protocol, and layer-4 port id. Corresponding watcher logging function is also added for IPv6. Signed-off-by: Kuo-lang Tseng Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/bridge/netfilter/Kconfig | 9 +++ net/bridge/netfilter/Makefile | 1 + net/bridge/netfilter/ebt_ip6.c | 144 +++++++++++++++++++++++++++++++++++++++++ net/bridge/netfilter/ebt_log.c | 64 +++++++++++++----- 4 files changed, 202 insertions(+), 16 deletions(-) create mode 100644 net/bridge/netfilter/ebt_ip6.c (limited to 'net') diff --git a/net/bridge/netfilter/Kconfig b/net/bridge/netfilter/Kconfig index 7beeefa0f9c0..fb684c2ff8b6 100644 --- a/net/bridge/netfilter/Kconfig +++ b/net/bridge/netfilter/Kconfig @@ -83,6 +83,15 @@ config BRIDGE_EBT_IP To compile it as a module, choose M here. If unsure, say N. +config BRIDGE_EBT_IP6 + tristate "ebt: IP6 filter support" + depends on BRIDGE_NF_EBTABLES + help + This option adds the IP6 match, which allows basic IPV6 header field + filtering. + + To compile it as a module, choose M here. If unsure, say N. + config BRIDGE_EBT_LIMIT tristate "ebt: limit match support" depends on BRIDGE_NF_EBTABLES diff --git a/net/bridge/netfilter/Makefile b/net/bridge/netfilter/Makefile index 83715d73a503..dd960645b413 100644 --- a/net/bridge/netfilter/Makefile +++ b/net/bridge/netfilter/Makefile @@ -14,6 +14,7 @@ obj-$(CONFIG_BRIDGE_EBT_802_3) += ebt_802_3.o obj-$(CONFIG_BRIDGE_EBT_AMONG) += ebt_among.o obj-$(CONFIG_BRIDGE_EBT_ARP) += ebt_arp.o obj-$(CONFIG_BRIDGE_EBT_IP) += ebt_ip.o +obj-$(CONFIG_BRIDGE_EBT_IP) += ebt_ip6.o obj-$(CONFIG_BRIDGE_EBT_LIMIT) += ebt_limit.o obj-$(CONFIG_BRIDGE_EBT_MARK) += ebt_mark_m.o obj-$(CONFIG_BRIDGE_EBT_PKTTYPE) += ebt_pkttype.o diff --git a/net/bridge/netfilter/ebt_ip6.c b/net/bridge/netfilter/ebt_ip6.c new file mode 100644 index 000000000000..36efb3a75249 --- /dev/null +++ b/net/bridge/netfilter/ebt_ip6.c @@ -0,0 +1,144 @@ +/* + * ebt_ip6 + * + * Authors: + * Manohar Castelino + * Kuo-Lang Tseng + * Jan Engelhardt + * + * Summary: + * This is just a modification of the IPv4 code written by + * Bart De Schuymer + * with the changes required to support IPv6 + * + * Jan, 2008 + */ + +#include +#include +#include +#include +#include +#include +#include + +struct tcpudphdr { + __be16 src; + __be16 dst; +}; + +static int ebt_filter_ip6(const struct sk_buff *skb, + const struct net_device *in, + const struct net_device *out, const void *data, + unsigned int datalen) +{ + const struct ebt_ip6_info *info = (struct ebt_ip6_info *)data; + const struct ipv6hdr *ih6; + struct ipv6hdr _ip6h; + const struct tcpudphdr *pptr; + struct tcpudphdr _ports; + struct in6_addr tmp_addr; + int i; + + ih6 = skb_header_pointer(skb, 0, sizeof(_ip6h), &_ip6h); + if (ih6 == NULL) + return EBT_NOMATCH; + if (info->bitmask & EBT_IP6_TCLASS && + FWINV(info->tclass != ipv6_get_dsfield(ih6), EBT_IP6_TCLASS)) + return EBT_NOMATCH; + for (i = 0; i < 4; i++) + tmp_addr.in6_u.u6_addr32[i] = ih6->saddr.in6_u.u6_addr32[i] & + info->smsk.in6_u.u6_addr32[i]; + if (info->bitmask & EBT_IP6_SOURCE && + FWINV((ipv6_addr_cmp(&tmp_addr, &info->saddr) != 0), + EBT_IP6_SOURCE)) + return EBT_NOMATCH; + for (i = 0; i < 4; i++) + tmp_addr.in6_u.u6_addr32[i] = ih6->daddr.in6_u.u6_addr32[i] & + info->dmsk.in6_u.u6_addr32[i]; + if (info->bitmask & EBT_IP6_DEST && + FWINV((ipv6_addr_cmp(&tmp_addr, &info->daddr) != 0), EBT_IP6_DEST)) + return EBT_NOMATCH; + if (info->bitmask & EBT_IP6_PROTO) { + uint8_t nexthdr = ih6->nexthdr; + int offset_ph; + + offset_ph = ipv6_skip_exthdr(skb, sizeof(_ip6h), &nexthdr); + if (offset_ph == -1) + return EBT_NOMATCH; + if (FWINV(info->protocol != nexthdr, EBT_IP6_PROTO)) + return EBT_NOMATCH; + if (!(info->bitmask & EBT_IP6_DPORT) && + !(info->bitmask & EBT_IP6_SPORT)) + return EBT_MATCH; + pptr = skb_header_pointer(skb, offset_ph, sizeof(_ports), + &_ports); + if (pptr == NULL) + return EBT_NOMATCH; + if (info->bitmask & EBT_IP6_DPORT) { + u32 dst = ntohs(pptr->dst); + if (FWINV(dst < info->dport[0] || + dst > info->dport[1], EBT_IP6_DPORT)) + return EBT_NOMATCH; + } + if (info->bitmask & EBT_IP6_SPORT) { + u32 src = ntohs(pptr->src); + if (FWINV(src < info->sport[0] || + src > info->sport[1], EBT_IP6_SPORT)) + return EBT_NOMATCH; + } + return EBT_MATCH; + } + return EBT_MATCH; +} + +static int ebt_ip6_check(const char *tablename, unsigned int hookmask, + const struct ebt_entry *e, void *data, unsigned int datalen) +{ + struct ebt_ip6_info *info = (struct ebt_ip6_info *)data; + + if (datalen != EBT_ALIGN(sizeof(struct ebt_ip6_info))) + return -EINVAL; + if (e->ethproto != htons(ETH_P_IPV6) || e->invflags & EBT_IPROTO) + return -EINVAL; + if (info->bitmask & ~EBT_IP6_MASK || info->invflags & ~EBT_IP6_MASK) + return -EINVAL; + if (info->bitmask & (EBT_IP6_DPORT | EBT_IP6_SPORT)) { + if (info->invflags & EBT_IP6_PROTO) + return -EINVAL; + if (info->protocol != IPPROTO_TCP && + info->protocol != IPPROTO_UDP && + info->protocol != IPPROTO_UDPLITE && + info->protocol != IPPROTO_SCTP && + info->protocol != IPPROTO_DCCP) + return -EINVAL; + } + if (info->bitmask & EBT_IP6_DPORT && info->dport[0] > info->dport[1]) + return -EINVAL; + if (info->bitmask & EBT_IP6_SPORT && info->sport[0] > info->sport[1]) + return -EINVAL; + return 0; +} + +static struct ebt_match filter_ip6 = +{ + .name = EBT_IP6_MATCH, + .match = ebt_filter_ip6, + .check = ebt_ip6_check, + .me = THIS_MODULE, +}; + +static int __init ebt_ip6_init(void) +{ + return ebt_register_match(&filter_ip6); +} + +static void __exit ebt_ip6_fini(void) +{ + ebt_unregister_match(&filter_ip6); +} + +module_init(ebt_ip6_init); +module_exit(ebt_ip6_fini); +MODULE_DESCRIPTION("Ebtables: IPv6 protocol packet match"); +MODULE_LICENSE("GPL"); diff --git a/net/bridge/netfilter/ebt_log.c b/net/bridge/netfilter/ebt_log.c index 0b209e4aad0a..c883ec8a28b4 100644 --- a/net/bridge/netfilter/ebt_log.c +++ b/net/bridge/netfilter/ebt_log.c @@ -18,6 +18,9 @@ #include #include #include +#include +#include +#include static DEFINE_SPINLOCK(ebt_log_lock); @@ -58,6 +61,27 @@ static void print_MAC(const unsigned char *p) printk("%02x%c", *p, i == ETH_ALEN - 1 ? ' ':':'); } +static void +print_ports(const struct sk_buff *skb, uint8_t protocol, int offset) +{ + if (protocol == IPPROTO_TCP || + protocol == IPPROTO_UDP || + protocol == IPPROTO_UDPLITE || + protocol == IPPROTO_SCTP || + protocol == IPPROTO_DCCP) { + const struct tcpudphdr *pptr; + struct tcpudphdr _ports; + + pptr = skb_header_pointer(skb, offset, + sizeof(_ports), &_ports); + if (pptr == NULL) { + printk(" INCOMPLETE TCP/UDP header"); + return; + } + printk(" SPT=%u DPT=%u", ntohs(pptr->src), ntohs(pptr->dst)); + } +} + #define myNIPQUAD(a) a[0], a[1], a[2], a[3] static void ebt_log_packet(unsigned int pf, unsigned int hooknum, @@ -95,23 +119,31 @@ ebt_log_packet(unsigned int pf, unsigned int hooknum, printk(" IP SRC=%u.%u.%u.%u IP DST=%u.%u.%u.%u, IP " "tos=0x%02X, IP proto=%d", NIPQUAD(ih->saddr), NIPQUAD(ih->daddr), ih->tos, ih->protocol); - if (ih->protocol == IPPROTO_TCP || - ih->protocol == IPPROTO_UDP || - ih->protocol == IPPROTO_UDPLITE || - ih->protocol == IPPROTO_SCTP || - ih->protocol == IPPROTO_DCCP) { - const struct tcpudphdr *pptr; - struct tcpudphdr _ports; - - pptr = skb_header_pointer(skb, ih->ihl*4, - sizeof(_ports), &_ports); - if (pptr == NULL) { - printk(" INCOMPLETE TCP/UDP header"); - goto out; - } - printk(" SPT=%u DPT=%u", ntohs(pptr->src), - ntohs(pptr->dst)); + print_ports(skb, ih->protocol, ih->ihl*4); + goto out; + } + + if ((bitmask & EBT_LOG_IP6) && eth_hdr(skb)->h_proto == + htons(ETH_P_IPV6)) { + const struct ipv6hdr *ih; + struct ipv6hdr _iph; + uint8_t nexthdr; + int offset_ph; + + ih = skb_header_pointer(skb, 0, sizeof(_iph), &_iph); + if (ih == NULL) { + printk(" INCOMPLETE IPv6 header"); + goto out; } + printk(" IPv6 SRC=%x:%x:%x:%x:%x:%x:%x:%x " + "IPv6 DST=%x:%x:%x:%x:%x:%x:%x:%x, IPv6 " + "priority=0x%01X, Next Header=%d", NIP6(ih->saddr), + NIP6(ih->daddr), ih->priority, ih->nexthdr); + nexthdr = ih->nexthdr; + offset_ph = ipv6_skip_exthdr(skb, sizeof(_iph), &nexthdr); + if (offset_ph == -1) + goto out; + print_ports(skb, nexthdr, offset_ph); goto out; } -- cgit v1.2.2 From 0adf9d67489cd30bab8eb93f7de81a674e44e1c3 Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Mon, 9 Jun 2008 15:56:20 -0700 Subject: netfilter: ctnetlink: group errors into logical errno sets This patch groups ctnetlink errors into three logical sets: * Malformed messages: if ctnetlink receives a message without some mandatory attribute, then it returns EINVAL. * Unsupported operations: if userspace tries to perform an unsupported operation, then it returns EOPNOTSUPP. * Unchangeable: if userspace tries to change some attribute of the conntrack object that can only be set once, then it returns EBUSY. This patch reduces the number of -EINVAL from 23 to 14 and it results in 5 -EBUSY and 6 -EOPNOTSUPP. Signed-off-by: Pablo Neira Ayuso Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/netfilter/nf_conntrack_netlink.c | 21 ++++++++++----------- 1 file changed, 10 insertions(+), 11 deletions(-) (limited to 'net') diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index 0edefcfc5949..13918c1fbf66 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -4,7 +4,7 @@ * (C) 2001 by Jay Schulist * (C) 2002-2006 by Harald Welte * (C) 2003 by Patrick Mchardy - * (C) 2005-2007 by Pablo Neira Ayuso + * (C) 2005-2008 by Pablo Neira Ayuso * * Initial connection tracking via netlink development funded and * generally made possible by Network Robots, Inc. (www.networkrobots.com) @@ -891,20 +891,19 @@ ctnetlink_change_status(struct nf_conn *ct, struct nlattr *cda[]) if (d & (IPS_EXPECTED|IPS_CONFIRMED|IPS_DYING)) /* unchangeable */ - return -EINVAL; + return -EBUSY; if (d & IPS_SEEN_REPLY && !(status & IPS_SEEN_REPLY)) /* SEEN_REPLY bit can only be set */ - return -EINVAL; - + return -EBUSY; if (d & IPS_ASSURED && !(status & IPS_ASSURED)) /* ASSURED bit can only be set */ - return -EINVAL; + return -EBUSY; if (cda[CTA_NAT_SRC] || cda[CTA_NAT_DST]) { #ifndef CONFIG_NF_NAT_NEEDED - return -EINVAL; + return -EOPNOTSUPP; #else struct nf_nat_range range; @@ -945,7 +944,7 @@ ctnetlink_change_helper(struct nf_conn *ct, struct nlattr *cda[]) /* don't change helper of sibling connections */ if (ct->master) - return -EINVAL; + return -EBUSY; err = ctnetlink_parse_help(cda[CTA_HELP], &helpname); if (err < 0) @@ -963,7 +962,7 @@ ctnetlink_change_helper(struct nf_conn *ct, struct nlattr *cda[]) helper = __nf_conntrack_helper_find_byname(helpname); if (helper == NULL) - return -EINVAL; + return -EOPNOTSUPP; if (help) { if (help->helper == helper) @@ -1258,12 +1257,12 @@ ctnetlink_new_conntrack(struct sock *ctnl, struct sk_buff *skb, if (!(nlh->nlmsg_flags & NLM_F_EXCL)) { /* we only allow nat config for new conntracks */ if (cda[CTA_NAT_SRC] || cda[CTA_NAT_DST]) { - err = -EINVAL; + err = -EOPNOTSUPP; goto out_unlock; } /* can't link an existing conntrack to a master */ if (cda[CTA_TUPLE_MASTER]) { - err = -EINVAL; + err = -EOPNOTSUPP; goto out_unlock; } err = ctnetlink_change_conntrack(nf_ct_tuplehash_to_ctrack(h), @@ -1608,7 +1607,7 @@ ctnetlink_del_expect(struct sock *ctnl, struct sk_buff *skb, h = __nf_conntrack_helper_find_byname(name); if (!h) { spin_unlock_bh(&nf_conntrack_lock); - return -EINVAL; + return -EOPNOTSUPP; } for (i = 0; i < nf_ct_expect_hsize; i++) { hlist_for_each_entry_safe(exp, n, next, -- cgit v1.2.2 From a258860e01b80e8f554a4ab1a6c95e6042eb8b73 Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Mon, 9 Jun 2008 15:56:39 -0700 Subject: netfilter: ctnetlink: add full support for SCTP to ctnetlink This patch adds full support for SCTP to ctnetlink. This includes three new attributes: state, original vtag and reply vtag. Signed-off-by: Pablo Neira Ayuso Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/netfilter/nf_conntrack_proto_sctp.c | 80 +++++++++++++++++++++++++++++++++ 1 file changed, 80 insertions(+) (limited to 'net') diff --git a/net/netfilter/nf_conntrack_proto_sctp.c b/net/netfilter/nf_conntrack_proto_sctp.c index cbf2e27a22b2..41183a4d2d62 100644 --- a/net/netfilter/nf_conntrack_proto_sctp.c +++ b/net/netfilter/nf_conntrack_proto_sctp.c @@ -463,6 +463,82 @@ static bool sctp_new(struct nf_conn *ct, const struct sk_buff *skb, return true; } +#if defined(CONFIG_NF_CT_NETLINK) || defined(CONFIG_NF_CT_NETLINK_MODULE) + +#include +#include + +static int sctp_to_nlattr(struct sk_buff *skb, struct nlattr *nla, + const struct nf_conn *ct) +{ + struct nlattr *nest_parms; + + read_lock_bh(&sctp_lock); + nest_parms = nla_nest_start(skb, CTA_PROTOINFO_SCTP | NLA_F_NESTED); + if (!nest_parms) + goto nla_put_failure; + + NLA_PUT_U8(skb, CTA_PROTOINFO_SCTP_STATE, ct->proto.sctp.state); + + NLA_PUT_BE32(skb, + CTA_PROTOINFO_SCTP_VTAG_ORIGINAL, + htonl(ct->proto.sctp.vtag[IP_CT_DIR_ORIGINAL])); + + NLA_PUT_BE32(skb, + CTA_PROTOINFO_SCTP_VTAG_REPLY, + htonl(ct->proto.sctp.vtag[IP_CT_DIR_REPLY])); + + read_unlock_bh(&sctp_lock); + + nla_nest_end(skb, nest_parms); + + return 0; + +nla_put_failure: + read_unlock_bh(&sctp_lock); + return -1; +} + +static const struct nla_policy sctp_nla_policy[CTA_PROTOINFO_SCTP_MAX+1] = { + [CTA_PROTOINFO_SCTP_STATE] = { .type = NLA_U8 }, + [CTA_PROTOINFO_SCTP_VTAG_ORIGINAL] = { .type = NLA_U32 }, + [CTA_PROTOINFO_SCTP_VTAG_REPLY] = { .type = NLA_U32 }, +}; + +static int nlattr_to_sctp(struct nlattr *cda[], struct nf_conn *ct) +{ + struct nlattr *attr = cda[CTA_PROTOINFO_SCTP]; + struct nlattr *tb[CTA_PROTOINFO_SCTP_MAX+1]; + int err; + + /* updates may not contain the internal protocol info, skip parsing */ + if (!attr) + return 0; + + err = nla_parse_nested(tb, + CTA_PROTOINFO_SCTP_MAX, + attr, + sctp_nla_policy); + if (err < 0) + return err; + + if (!tb[CTA_PROTOINFO_SCTP_STATE] || + !tb[CTA_PROTOINFO_SCTP_VTAG_ORIGINAL] || + !tb[CTA_PROTOINFO_SCTP_VTAG_REPLY]) + return -EINVAL; + + write_lock_bh(&sctp_lock); + ct->proto.sctp.state = nla_get_u8(tb[CTA_PROTOINFO_SCTP_STATE]); + ct->proto.sctp.vtag[IP_CT_DIR_ORIGINAL] = + ntohl(nla_get_be32(tb[CTA_PROTOINFO_SCTP_VTAG_ORIGINAL])); + ct->proto.sctp.vtag[IP_CT_DIR_REPLY] = + ntohl(nla_get_be32(tb[CTA_PROTOINFO_SCTP_VTAG_REPLY])); + write_unlock_bh(&sctp_lock); + + return 0; +} +#endif + #ifdef CONFIG_SYSCTL static unsigned int sctp_sysctl_table_users; static struct ctl_table_header *sctp_sysctl_header; @@ -591,6 +667,8 @@ static struct nf_conntrack_l4proto nf_conntrack_l4proto_sctp4 __read_mostly = { .new = sctp_new, .me = THIS_MODULE, #if defined(CONFIG_NF_CT_NETLINK) || defined(CONFIG_NF_CT_NETLINK_MODULE) + .to_nlattr = sctp_to_nlattr, + .from_nlattr = nlattr_to_sctp, .tuple_to_nlattr = nf_ct_port_tuple_to_nlattr, .nlattr_to_tuple = nf_ct_port_nlattr_to_tuple, .nla_policy = nf_ct_port_nla_policy, @@ -617,6 +695,8 @@ static struct nf_conntrack_l4proto nf_conntrack_l4proto_sctp6 __read_mostly = { .new = sctp_new, .me = THIS_MODULE, #if defined(CONFIG_NF_CT_NETLINK) || defined(CONFIG_NF_CT_NETLINK_MODULE) + .to_nlattr = sctp_to_nlattr, + .from_nlattr = nlattr_to_sctp, .tuple_to_nlattr = nf_ct_port_tuple_to_nlattr, .nlattr_to_tuple = nf_ct_port_nlattr_to_tuple, .nla_policy = nf_ct_port_nla_policy, -- cgit v1.2.2 From 560ee653b67074b805f1b661988a72a0e58811a5 Mon Sep 17 00:00:00 2001 From: James Morris Date: Mon, 9 Jun 2008 15:57:24 -0700 Subject: netfilter: ip_tables: add iptables security table for mandatory access control rules The following patch implements a new "security" table for iptables, so that MAC (SELinux etc.) networking rules can be managed separately to standard DAC rules. This is to help with distro integration of the new secmark-based network controls, per various previous discussions. The need for a separate table arises from the fact that existing tools and usage of iptables will likely clash with centralized MAC policy management. The SECMARK and CONNSECMARK targets will still be valid in the mangle table to prevent breakage of existing users. Signed-off-by: James Morris Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/ipv4/netfilter/Kconfig | 12 +++ net/ipv4/netfilter/Makefile | 1 + net/ipv4/netfilter/iptable_security.c | 180 ++++++++++++++++++++++++++++++++++ net/netfilter/xt_CONNSECMARK.c | 10 +- net/netfilter/xt_SECMARK.c | 10 +- 5 files changed, 207 insertions(+), 6 deletions(-) create mode 100644 net/ipv4/netfilter/iptable_security.c (limited to 'net') diff --git a/net/ipv4/netfilter/Kconfig b/net/ipv4/netfilter/Kconfig index 2767841a8cef..6e251402506e 100644 --- a/net/ipv4/netfilter/Kconfig +++ b/net/ipv4/netfilter/Kconfig @@ -365,6 +365,18 @@ config IP_NF_RAW If you want to compile it as a module, say M here and read . If unsure, say `N'. +# security table for MAC policy +config IP_NF_SECURITY + tristate "Security table" + depends on IP_NF_IPTABLES + depends on SECURITY + default m if NETFILTER_ADVANCED=n + help + This option adds a `security' table to iptables, for use + with Mandatory Access Control (MAC) policy. + + If unsure, say N. + # ARP tables config IP_NF_ARPTABLES tristate "ARP tables support" diff --git a/net/ipv4/netfilter/Makefile b/net/ipv4/netfilter/Makefile index d9b92fbf5579..3f31291f37ce 100644 --- a/net/ipv4/netfilter/Makefile +++ b/net/ipv4/netfilter/Makefile @@ -42,6 +42,7 @@ obj-$(CONFIG_IP_NF_FILTER) += iptable_filter.o obj-$(CONFIG_IP_NF_MANGLE) += iptable_mangle.o obj-$(CONFIG_NF_NAT) += iptable_nat.o obj-$(CONFIG_IP_NF_RAW) += iptable_raw.o +obj-$(CONFIG_IP_NF_SECURITY) += iptable_security.o # matches obj-$(CONFIG_IP_NF_MATCH_ADDRTYPE) += ipt_addrtype.o diff --git a/net/ipv4/netfilter/iptable_security.c b/net/ipv4/netfilter/iptable_security.c new file mode 100644 index 000000000000..2b472ac2263a --- /dev/null +++ b/net/ipv4/netfilter/iptable_security.c @@ -0,0 +1,180 @@ +/* + * "security" table + * + * This is for use by Mandatory Access Control (MAC) security models, + * which need to be able to manage security policy in separate context + * to DAC. + * + * Based on iptable_mangle.c + * + * Copyright (C) 1999 Paul `Rusty' Russell & Michael J. Neuling + * Copyright (C) 2000-2004 Netfilter Core Team netfilter.org> + * Copyright (C) 2008 Red Hat, Inc., James Morris redhat.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ +#include +#include +#include + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("James Morris redhat.com>"); +MODULE_DESCRIPTION("iptables security table, for MAC rules"); + +#define SECURITY_VALID_HOOKS (1 << NF_INET_LOCAL_IN) | \ + (1 << NF_INET_FORWARD) | \ + (1 << NF_INET_LOCAL_OUT) + +static struct +{ + struct ipt_replace repl; + struct ipt_standard entries[3]; + struct ipt_error term; +} initial_table __initdata = { + .repl = { + .name = "security", + .valid_hooks = SECURITY_VALID_HOOKS, + .num_entries = 4, + .size = sizeof(struct ipt_standard) * 3 + sizeof(struct ipt_error), + .hook_entry = { + [NF_INET_LOCAL_IN] = 0, + [NF_INET_FORWARD] = sizeof(struct ipt_standard), + [NF_INET_LOCAL_OUT] = sizeof(struct ipt_standard) * 2, + }, + .underflow = { + [NF_INET_LOCAL_IN] = 0, + [NF_INET_FORWARD] = sizeof(struct ipt_standard), + [NF_INET_LOCAL_OUT] = sizeof(struct ipt_standard) * 2, + }, + }, + .entries = { + IPT_STANDARD_INIT(NF_ACCEPT), /* LOCAL_IN */ + IPT_STANDARD_INIT(NF_ACCEPT), /* FORWARD */ + IPT_STANDARD_INIT(NF_ACCEPT), /* LOCAL_OUT */ + }, + .term = IPT_ERROR_INIT, /* ERROR */ +}; + +static struct xt_table security_table = { + .name = "security", + .valid_hooks = SECURITY_VALID_HOOKS, + .lock = __RW_LOCK_UNLOCKED(security_table.lock), + .me = THIS_MODULE, + .af = AF_INET, +}; + +static unsigned int +ipt_local_in_hook(unsigned int hook, + struct sk_buff *skb, + const struct net_device *in, + const struct net_device *out, + int (*okfn)(struct sk_buff *)) +{ + return ipt_do_table(skb, hook, in, out, + nf_local_in_net(in, out)->ipv4.iptable_security); +} + +static unsigned int +ipt_forward_hook(unsigned int hook, + struct sk_buff *skb, + const struct net_device *in, + const struct net_device *out, + int (*okfn)(struct sk_buff *)) +{ + return ipt_do_table(skb, hook, in, out, + nf_forward_net(in, out)->ipv4.iptable_security); +} + +static unsigned int +ipt_local_out_hook(unsigned int hook, + struct sk_buff *skb, + const struct net_device *in, + const struct net_device *out, + int (*okfn)(struct sk_buff *)) +{ + /* Somebody is playing with raw sockets. */ + if (skb->len < sizeof(struct iphdr) + || ip_hdrlen(skb) < sizeof(struct iphdr)) { + if (net_ratelimit()) + printk(KERN_INFO "iptable_security: ignoring short " + "SOCK_RAW packet.\n"); + return NF_ACCEPT; + } + return ipt_do_table(skb, hook, in, out, + nf_local_out_net(in, out)->ipv4.iptable_security); +} + +static struct nf_hook_ops ipt_ops[] __read_mostly = { + { + .hook = ipt_local_in_hook, + .owner = THIS_MODULE, + .pf = PF_INET, + .hooknum = NF_INET_LOCAL_IN, + .priority = NF_IP_PRI_SECURITY, + }, + { + .hook = ipt_forward_hook, + .owner = THIS_MODULE, + .pf = PF_INET, + .hooknum = NF_INET_FORWARD, + .priority = NF_IP_PRI_SECURITY, + }, + { + .hook = ipt_local_out_hook, + .owner = THIS_MODULE, + .pf = PF_INET, + .hooknum = NF_INET_LOCAL_OUT, + .priority = NF_IP_PRI_SECURITY, + }, +}; + +static int __net_init iptable_security_net_init(struct net *net) +{ + net->ipv4.iptable_security = + ipt_register_table(net, &security_table, &initial_table.repl); + + if (IS_ERR(net->ipv4.iptable_security)) + return PTR_ERR(net->ipv4.iptable_security); + + return 0; +} + +static void __net_exit iptable_security_net_exit(struct net *net) +{ + ipt_unregister_table(net->ipv4.iptable_security); +} + +static struct pernet_operations iptable_security_net_ops = { + .init = iptable_security_net_init, + .exit = iptable_security_net_exit, +}; + +static int __init iptable_security_init(void) +{ + int ret; + + ret = register_pernet_subsys(&iptable_security_net_ops); + if (ret < 0) + return ret; + + ret = nf_register_hooks(ipt_ops, ARRAY_SIZE(ipt_ops)); + if (ret < 0) + goto cleanup_table; + + return ret; + +cleanup_table: + unregister_pernet_subsys(&iptable_security_net_ops); + return ret; +} + +static void __exit iptable_security_fini(void) +{ + nf_unregister_hooks(ipt_ops, ARRAY_SIZE(ipt_ops)); + unregister_pernet_subsys(&iptable_security_net_ops); +} + +module_init(iptable_security_init); +module_exit(iptable_security_fini); diff --git a/net/netfilter/xt_CONNSECMARK.c b/net/netfilter/xt_CONNSECMARK.c index 211189eb2b67..76ca1f2421eb 100644 --- a/net/netfilter/xt_CONNSECMARK.c +++ b/net/netfilter/xt_CONNSECMARK.c @@ -8,7 +8,7 @@ * Copyright (C) 2002,2004 MARA Systems AB * by Henrik Nordstrom * - * (C) 2006 Red Hat, Inc., James Morris + * (C) 2006,2008 Red Hat, Inc., James Morris * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License version 2 as @@ -94,6 +94,12 @@ connsecmark_tg_check(const char *tablename, const void *entry, { const struct xt_connsecmark_target_info *info = targinfo; + if (strcmp(tablename, "mangle") && strcmp(tablename, "security")) { + printk(KERN_INFO PFX "target only valid in the \'mangle\' " + "or \'security\' tables, not \'%s\'.\n", tablename); + return false; + } + switch (info->mode) { case CONNSECMARK_SAVE: case CONNSECMARK_RESTORE: @@ -126,7 +132,6 @@ static struct xt_target connsecmark_tg_reg[] __read_mostly = { .destroy = connsecmark_tg_destroy, .target = connsecmark_tg, .targetsize = sizeof(struct xt_connsecmark_target_info), - .table = "mangle", .me = THIS_MODULE, }, { @@ -136,7 +141,6 @@ static struct xt_target connsecmark_tg_reg[] __read_mostly = { .destroy = connsecmark_tg_destroy, .target = connsecmark_tg, .targetsize = sizeof(struct xt_connsecmark_target_info), - .table = "mangle", .me = THIS_MODULE, }, }; diff --git a/net/netfilter/xt_SECMARK.c b/net/netfilter/xt_SECMARK.c index c0284856ccd4..94f87ee7552b 100644 --- a/net/netfilter/xt_SECMARK.c +++ b/net/netfilter/xt_SECMARK.c @@ -5,7 +5,7 @@ * Based on the nfmark match by: * (C) 1999-2001 Marc Boucher * - * (C) 2006 Red Hat, Inc., James Morris + * (C) 2006,2008 Red Hat, Inc., James Morris * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License version 2 as @@ -89,6 +89,12 @@ secmark_tg_check(const char *tablename, const void *entry, { struct xt_secmark_target_info *info = targinfo; + if (strcmp(tablename, "mangle") && strcmp(tablename, "security")) { + printk(KERN_INFO PFX "target only valid in the \'mangle\' " + "or \'security\' tables, not \'%s\'.\n", tablename); + return false; + } + if (mode && mode != info->mode) { printk(KERN_INFO PFX "mode already set to %hu cannot mix with " "rules for mode %hu\n", mode, info->mode); @@ -127,7 +133,6 @@ static struct xt_target secmark_tg_reg[] __read_mostly = { .destroy = secmark_tg_destroy, .target = secmark_tg, .targetsize = sizeof(struct xt_secmark_target_info), - .table = "mangle", .me = THIS_MODULE, }, { @@ -137,7 +142,6 @@ static struct xt_target secmark_tg_reg[] __read_mostly = { .destroy = secmark_tg_destroy, .target = secmark_tg, .targetsize = sizeof(struct xt_secmark_target_info), - .table = "mangle", .me = THIS_MODULE, }, }; -- cgit v1.2.2 From 17e6e59f0a1d7188d783c15dc3ccebd95a0840cd Mon Sep 17 00:00:00 2001 From: James Morris Date: Mon, 9 Jun 2008 15:58:05 -0700 Subject: netfilter: ip6_tables: add ip6tables security table This is a port of the IPv4 security table for IPv6. Signed-off-by: James Morris Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/ipv6/netfilter/Kconfig | 12 +++ net/ipv6/netfilter/Makefile | 1 + net/ipv6/netfilter/ip6table_security.c | 172 +++++++++++++++++++++++++++++++++ 3 files changed, 185 insertions(+) create mode 100644 net/ipv6/netfilter/ip6table_security.c (limited to 'net') diff --git a/net/ipv6/netfilter/Kconfig b/net/ipv6/netfilter/Kconfig index 6cae5475737e..689dec899c57 100644 --- a/net/ipv6/netfilter/Kconfig +++ b/net/ipv6/netfilter/Kconfig @@ -208,5 +208,17 @@ config IP6_NF_RAW If you want to compile it as a module, say M here and read . If unsure, say `N'. +# security table for MAC policy +config IP6_NF_SECURITY + tristate "Security table" + depends on IP6_NF_IPTABLES + depends on SECURITY + default m if NETFILTER_ADVANCED=n + help + This option adds a `security' table to iptables, for use + with Mandatory Access Control (MAC) policy. + + If unsure, say N. + endmenu diff --git a/net/ipv6/netfilter/Makefile b/net/ipv6/netfilter/Makefile index fbf2c14ed887..3f17c948eefb 100644 --- a/net/ipv6/netfilter/Makefile +++ b/net/ipv6/netfilter/Makefile @@ -8,6 +8,7 @@ obj-$(CONFIG_IP6_NF_FILTER) += ip6table_filter.o obj-$(CONFIG_IP6_NF_MANGLE) += ip6table_mangle.o obj-$(CONFIG_IP6_NF_QUEUE) += ip6_queue.o obj-$(CONFIG_IP6_NF_RAW) += ip6table_raw.o +obj-$(CONFIG_IP6_NF_SECURITY) += ip6table_security.o # objects for l3 independent conntrack nf_conntrack_ipv6-objs := nf_conntrack_l3proto_ipv6.o nf_conntrack_proto_icmpv6.o nf_conntrack_reasm.o diff --git a/net/ipv6/netfilter/ip6table_security.c b/net/ipv6/netfilter/ip6table_security.c new file mode 100644 index 000000000000..063a3d9c3c67 --- /dev/null +++ b/net/ipv6/netfilter/ip6table_security.c @@ -0,0 +1,172 @@ +/* + * "security" table for IPv6 + * + * This is for use by Mandatory Access Control (MAC) security models, + * which need to be able to manage security policy in separate context + * to DAC. + * + * Based on iptable_mangle.c + * + * Copyright (C) 1999 Paul `Rusty' Russell & Michael J. Neuling + * Copyright (C) 2000-2004 Netfilter Core Team netfilter.org> + * Copyright (C) 2008 Red Hat, Inc., James Morris redhat.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ +#include +#include + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("James Morris redhat.com>"); +MODULE_DESCRIPTION("ip6tables security table, for MAC rules"); + +#define SECURITY_VALID_HOOKS (1 << NF_INET_LOCAL_IN) | \ + (1 << NF_INET_FORWARD) | \ + (1 << NF_INET_LOCAL_OUT) + +static struct +{ + struct ip6t_replace repl; + struct ip6t_standard entries[3]; + struct ip6t_error term; +} initial_table __initdata = { + .repl = { + .name = "security", + .valid_hooks = SECURITY_VALID_HOOKS, + .num_entries = 4, + .size = sizeof(struct ip6t_standard) * 3 + sizeof(struct ip6t_error), + .hook_entry = { + [NF_INET_LOCAL_IN] = 0, + [NF_INET_FORWARD] = sizeof(struct ip6t_standard), + [NF_INET_LOCAL_OUT] = sizeof(struct ip6t_standard) * 2, + }, + .underflow = { + [NF_INET_LOCAL_IN] = 0, + [NF_INET_FORWARD] = sizeof(struct ip6t_standard), + [NF_INET_LOCAL_OUT] = sizeof(struct ip6t_standard) * 2, + }, + }, + .entries = { + IP6T_STANDARD_INIT(NF_ACCEPT), /* LOCAL_IN */ + IP6T_STANDARD_INIT(NF_ACCEPT), /* FORWARD */ + IP6T_STANDARD_INIT(NF_ACCEPT), /* LOCAL_OUT */ + }, + .term = IP6T_ERROR_INIT, /* ERROR */ +}; + +static struct xt_table security_table = { + .name = "security", + .valid_hooks = SECURITY_VALID_HOOKS, + .lock = __RW_LOCK_UNLOCKED(security_table.lock), + .me = THIS_MODULE, + .af = AF_INET6, +}; + +static unsigned int +ip6t_local_in_hook(unsigned int hook, + struct sk_buff *skb, + const struct net_device *in, + const struct net_device *out, + int (*okfn)(struct sk_buff *)) +{ + return ip6t_do_table(skb, hook, in, out, + init_net.ipv6.ip6table_security); +} + +static unsigned int +ip6t_forward_hook(unsigned int hook, + struct sk_buff *skb, + const struct net_device *in, + const struct net_device *out, + int (*okfn)(struct sk_buff *)) +{ + return ip6t_do_table(skb, hook, in, out, + init_net.ipv6.ip6table_security); +} + +static unsigned int +ip6t_local_out_hook(unsigned int hook, + struct sk_buff *skb, + const struct net_device *in, + const struct net_device *out, + int (*okfn)(struct sk_buff *)) +{ + /* TBD: handle short packets via raw socket */ + return ip6t_do_table(skb, hook, in, out, + init_net.ipv6.ip6table_security); +} + +static struct nf_hook_ops ip6t_ops[] __read_mostly = { + { + .hook = ip6t_local_in_hook, + .owner = THIS_MODULE, + .pf = PF_INET6, + .hooknum = NF_INET_LOCAL_IN, + .priority = NF_IP6_PRI_SECURITY, + }, + { + .hook = ip6t_forward_hook, + .owner = THIS_MODULE, + .pf = PF_INET6, + .hooknum = NF_INET_FORWARD, + .priority = NF_IP6_PRI_SECURITY, + }, + { + .hook = ip6t_local_out_hook, + .owner = THIS_MODULE, + .pf = PF_INET6, + .hooknum = NF_INET_LOCAL_OUT, + .priority = NF_IP6_PRI_SECURITY, + }, +}; + +static int __net_init ip6table_security_net_init(struct net *net) +{ + net->ipv6.ip6table_security = + ip6t_register_table(net, &security_table, &initial_table.repl); + + if (IS_ERR(net->ipv6.ip6table_security)) + return PTR_ERR(net->ipv6.ip6table_security); + + return 0; +} + +static void __net_exit ip6table_security_net_exit(struct net *net) +{ + ip6t_unregister_table(net->ipv6.ip6table_security); +} + +static struct pernet_operations ip6table_security_net_ops = { + .init = ip6table_security_net_init, + .exit = ip6table_security_net_exit, +}; + +static int __init ip6table_security_init(void) +{ + int ret; + + ret = register_pernet_subsys(&ip6table_security_net_ops); + if (ret < 0) + return ret; + + ret = nf_register_hooks(ip6t_ops, ARRAY_SIZE(ip6t_ops)); + if (ret < 0) + goto cleanup_table; + + return ret; + +cleanup_table: + unregister_pernet_subsys(&ip6table_security_net_ops); + return ret; +} + +static void __exit ip6table_security_fini(void) +{ + nf_unregister_hooks(ip6t_ops, ARRAY_SIZE(ip6t_ops)); + unregister_pernet_subsys(&ip6table_security_net_ops); +} + +module_init(ip6table_security_init); +module_exit(ip6table_security_fini); -- cgit v1.2.2 From 31d8519c9cf87e9d8a0cc5b9734fda02af66d7e2 Mon Sep 17 00:00:00 2001 From: Pekka Enberg Date: Mon, 9 Jun 2008 15:58:39 -0700 Subject: netfilter: nf_conntrack_extend: use krealloc() in nf_conntrack_extend.c V2 The ksize() API is going away because it is being abused and it doesn't even work consistenly across different allocators. Therefore, convert net/netfilter/nf_conntrack_extend.c to use krealloc(). Signed-off-by: Pekka Enberg Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/netfilter/nf_conntrack_extend.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/netfilter/nf_conntrack_extend.c b/net/netfilter/nf_conntrack_extend.c index bcc19fa4ed1e..ba1c4915e9eb 100644 --- a/net/netfilter/nf_conntrack_extend.c +++ b/net/netfilter/nf_conntrack_extend.c @@ -88,13 +88,11 @@ void *__nf_ct_ext_add(struct nf_conn *ct, enum nf_ct_ext_id id, gfp_t gfp) newlen = newoff + t->len; rcu_read_unlock(); - if (newlen >= ksize(ct->ext)) { - new = kmalloc(newlen, gfp); - if (!new) - return NULL; - - memcpy(new, ct->ext, ct->ext->len); + new = krealloc(ct->ext, newlen, gfp); + if (!new) + return NULL; + if (new != ct->ext) { for (i = 0; i < NF_CT_EXT_NUM; i++) { if (!nf_ct_ext_exist(ct, i)) continue; -- cgit v1.2.2 From 51091764f26ec36c02e35166f083193a30f426fc Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 9 Jun 2008 15:59:06 -0700 Subject: netfilter: nf_conntrack: add nf_ct_kill() Encapsulate the common if (del_timer(&ct->timeout)) ct->timeout.function((unsigned long)ct) sequence in a new function. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/ipv4/netfilter/nf_conntrack_proto_icmp.c | 5 ++--- net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c | 5 ++--- net/netfilter/nf_conntrack_core.c | 7 +++++++ net/netfilter/nf_conntrack_netlink.c | 3 +-- net/netfilter/nf_conntrack_proto_dccp.c | 3 +-- net/netfilter/nf_conntrack_proto_tcp.c | 9 +++------ 6 files changed, 16 insertions(+), 16 deletions(-) (limited to 'net') diff --git a/net/ipv4/netfilter/nf_conntrack_proto_icmp.c b/net/ipv4/netfilter/nf_conntrack_proto_icmp.c index 78ab19accace..0e21a46184fb 100644 --- a/net/ipv4/netfilter/nf_conntrack_proto_icmp.c +++ b/net/ipv4/netfilter/nf_conntrack_proto_icmp.c @@ -87,9 +87,8 @@ static int icmp_packet(struct nf_conn *ct, means this will only run once even if count hits zero twice (theoretically possible with SMP) */ if (CTINFO2DIR(ctinfo) == IP_CT_DIR_REPLY) { - if (atomic_dec_and_test(&ct->proto.icmp.count) - && del_timer(&ct->timeout)) - ct->timeout.function((unsigned long)ct); + if (atomic_dec_and_test(&ct->proto.icmp.count)) + nf_ct_kill(ct); } else { atomic_inc(&ct->proto.icmp.count); nf_conntrack_event_cache(IPCT_PROTOINFO_VOLATILE, skb); diff --git a/net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c b/net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c index ee713b03e9ec..fe081b90e05e 100644 --- a/net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c +++ b/net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c @@ -89,9 +89,8 @@ static int icmpv6_packet(struct nf_conn *ct, means this will only run once even if count hits zero twice (theoretically possible with SMP) */ if (CTINFO2DIR(ctinfo) == IP_CT_DIR_REPLY) { - if (atomic_dec_and_test(&ct->proto.icmp.count) - && del_timer(&ct->timeout)) - ct->timeout.function((unsigned long)ct); + if (atomic_dec_and_test(&ct->proto.icmp.count)) + nf_ct_kill(ct); } else { atomic_inc(&ct->proto.icmp.count); nf_conntrack_event_cache(IPCT_PROTOINFO_VOLATILE, skb); diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c index c4b1799da5d7..79b07c35eb87 100644 --- a/net/netfilter/nf_conntrack_core.c +++ b/net/netfilter/nf_conntrack_core.c @@ -848,6 +848,13 @@ acct: } EXPORT_SYMBOL_GPL(__nf_ct_refresh_acct); +void nf_ct_kill(struct nf_conn *ct) +{ + if (del_timer(&ct->timeout)) + ct->timeout.function((unsigned long)ct); +} +EXPORT_SYMBOL_GPL(nf_ct_kill); + #if defined(CONFIG_NF_CT_NETLINK) || defined(CONFIG_NF_CT_NETLINK_MODULE) #include diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index 13918c1fbf66..ab655f660dfa 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -812,9 +812,8 @@ ctnetlink_del_conntrack(struct sock *ctnl, struct sk_buff *skb, return -ENOENT; } } - if (del_timer(&ct->timeout)) - ct->timeout.function((unsigned long)ct); + nf_ct_kill(ct); nf_ct_put(ct); return 0; diff --git a/net/netfilter/nf_conntrack_proto_dccp.c b/net/netfilter/nf_conntrack_proto_dccp.c index afb4a1861d2c..223742f371f9 100644 --- a/net/netfilter/nf_conntrack_proto_dccp.c +++ b/net/netfilter/nf_conntrack_proto_dccp.c @@ -475,8 +475,7 @@ static int dccp_packet(struct nf_conn *ct, const struct sk_buff *skb, if (type == DCCP_PKT_RESET && !test_bit(IPS_SEEN_REPLY_BIT, &ct->status)) { /* Tear down connection immediately if only reply is a RESET */ - if (del_timer(&ct->timeout)) - ct->timeout.function((unsigned long)ct); + nf_ct_kill(ct); return NF_ACCEPT; } diff --git a/net/netfilter/nf_conntrack_proto_tcp.c b/net/netfilter/nf_conntrack_proto_tcp.c index ba94004fe323..c4aa11e01405 100644 --- a/net/netfilter/nf_conntrack_proto_tcp.c +++ b/net/netfilter/nf_conntrack_proto_tcp.c @@ -843,8 +843,7 @@ static int tcp_packet(struct nf_conn *ct, /* Attempt to reopen a closed/aborted connection. * Delete this connection and look up again. */ write_unlock_bh(&tcp_lock); - if (del_timer(&ct->timeout)) - ct->timeout.function((unsigned long)ct); + nf_ct_kill(ct); return -NF_REPEAT; } /* Fall through */ @@ -877,8 +876,7 @@ static int tcp_packet(struct nf_conn *ct, if (LOG_INVALID(IPPROTO_TCP)) nf_log_packet(pf, 0, skb, NULL, NULL, NULL, "nf_ct_tcp: killing out of sync session "); - if (del_timer(&ct->timeout)) - ct->timeout.function((unsigned long)ct); + nf_ct_kill(ct); return -NF_DROP; } ct->proto.tcp.last_index = index; @@ -961,8 +959,7 @@ static int tcp_packet(struct nf_conn *ct, problem case, so we can delete the conntrack immediately. --RR */ if (th->rst) { - if (del_timer(&ct->timeout)) - ct->timeout.function((unsigned long)ct); + nf_ct_kill(ct); return NF_ACCEPT; } } else if (!test_bit(IPS_ASSURED_BIT, &ct->status) -- cgit v1.2.2 From 718d4ad98e272daebc258e49dc02f52a6a8de9d3 Mon Sep 17 00:00:00 2001 From: Fabian Hugelshofer Date: Mon, 9 Jun 2008 15:59:40 -0700 Subject: netfilter: nf_conntrack: properly account terminating packets Currently the last packet of a connection isn't accounted when its causing abnormal termination. Introduces nf_ct_kill_acct() which increments the accounting counters on conntrack kill. The new function was necessary, because there are calls to nf_ct_kill() which don't need accounting: nf_conntrack_proto_tcp.c line ~847: Kills ct and returns NF_REPEAT. We don't want to count twice. nf_conntrack_proto_tcp.c line ~880: Kills ct and returns NF_DROP. I think we don't want to count dropped packets. nf_conntrack_netlink.c line ~824: As far as I can see ctnetlink_del_conntrack() is used to destroy a conntrack on behalf of the user. There is an sk_buff, but I don't think this is an actual packet. Incrementing counters here is therefore not desired. Signed-off-by: Fabian Hugelshofer Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/ipv4/netfilter/nf_conntrack_proto_icmp.c | 2 +- net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c | 2 +- net/netfilter/nf_conntrack_core.c | 16 ++++++++++++++-- net/netfilter/nf_conntrack_proto_dccp.c | 2 +- net/netfilter/nf_conntrack_proto_tcp.c | 2 +- 5 files changed, 18 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/ipv4/netfilter/nf_conntrack_proto_icmp.c b/net/ipv4/netfilter/nf_conntrack_proto_icmp.c index 0e21a46184fb..97791048fa9b 100644 --- a/net/ipv4/netfilter/nf_conntrack_proto_icmp.c +++ b/net/ipv4/netfilter/nf_conntrack_proto_icmp.c @@ -88,7 +88,7 @@ static int icmp_packet(struct nf_conn *ct, (theoretically possible with SMP) */ if (CTINFO2DIR(ctinfo) == IP_CT_DIR_REPLY) { if (atomic_dec_and_test(&ct->proto.icmp.count)) - nf_ct_kill(ct); + nf_ct_kill_acct(ct, ctinfo, skb); } else { atomic_inc(&ct->proto.icmp.count); nf_conntrack_event_cache(IPCT_PROTOINFO_VOLATILE, skb); diff --git a/net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c b/net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c index fe081b90e05e..14d47d833545 100644 --- a/net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c +++ b/net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c @@ -90,7 +90,7 @@ static int icmpv6_packet(struct nf_conn *ct, (theoretically possible with SMP) */ if (CTINFO2DIR(ctinfo) == IP_CT_DIR_REPLY) { if (atomic_dec_and_test(&ct->proto.icmp.count)) - nf_ct_kill(ct); + nf_ct_kill_acct(ct, ctinfo, skb); } else { atomic_inc(&ct->proto.icmp.count); nf_conntrack_event_cache(IPCT_PROTOINFO_VOLATILE, skb); diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c index 79b07c35eb87..e6d645221d5c 100644 --- a/net/netfilter/nf_conntrack_core.c +++ b/net/netfilter/nf_conntrack_core.c @@ -848,12 +848,24 @@ acct: } EXPORT_SYMBOL_GPL(__nf_ct_refresh_acct); -void nf_ct_kill(struct nf_conn *ct) +void __nf_ct_kill_acct(struct nf_conn *ct, + enum ip_conntrack_info ctinfo, + const struct sk_buff *skb, + int do_acct) { +#ifdef CONFIG_NF_CT_ACCT + if (do_acct) { + spin_lock_bh(&nf_conntrack_lock); + ct->counters[CTINFO2DIR(ctinfo)].packets++; + ct->counters[CTINFO2DIR(ctinfo)].bytes += + skb->len - skb_network_offset(skb); + spin_unlock_bh(&nf_conntrack_lock); + } +#endif if (del_timer(&ct->timeout)) ct->timeout.function((unsigned long)ct); } -EXPORT_SYMBOL_GPL(nf_ct_kill); +EXPORT_SYMBOL_GPL(__nf_ct_kill_acct); #if defined(CONFIG_NF_CT_NETLINK) || defined(CONFIG_NF_CT_NETLINK_MODULE) diff --git a/net/netfilter/nf_conntrack_proto_dccp.c b/net/netfilter/nf_conntrack_proto_dccp.c index 223742f371f9..e7866dd3cde6 100644 --- a/net/netfilter/nf_conntrack_proto_dccp.c +++ b/net/netfilter/nf_conntrack_proto_dccp.c @@ -475,7 +475,7 @@ static int dccp_packet(struct nf_conn *ct, const struct sk_buff *skb, if (type == DCCP_PKT_RESET && !test_bit(IPS_SEEN_REPLY_BIT, &ct->status)) { /* Tear down connection immediately if only reply is a RESET */ - nf_ct_kill(ct); + nf_ct_kill_acct(ct, ctinfo, skb); return NF_ACCEPT; } diff --git a/net/netfilter/nf_conntrack_proto_tcp.c b/net/netfilter/nf_conntrack_proto_tcp.c index c4aa11e01405..8db13fba10bc 100644 --- a/net/netfilter/nf_conntrack_proto_tcp.c +++ b/net/netfilter/nf_conntrack_proto_tcp.c @@ -959,7 +959,7 @@ static int tcp_packet(struct nf_conn *ct, problem case, so we can delete the conntrack immediately. --RR */ if (th->rst) { - nf_ct_kill(ct); + nf_ct_kill_acct(ct, ctinfo, skb); return NF_ACCEPT; } } else if (!test_bit(IPS_ASSURED_BIT, &ct->status) -- cgit v1.2.2 From e57dce60c7478fdeeb9a1ebd311261ec901afe4d Mon Sep 17 00:00:00 2001 From: Fabian Hugelshofer Date: Mon, 9 Jun 2008 15:59:58 -0700 Subject: netfilter: ctnetlink: include conntrack status in destroy event message When a conntrack is destroyed, the connection status does not get exported to netlink. I don't see a reason for not doing so. This patch exports the status on all conntrack events. Signed-off-by: Fabian Hugelshofer Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/netfilter/nf_conntrack_netlink.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index ab655f660dfa..63c4e1f299b8 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -475,14 +475,14 @@ static int ctnetlink_conntrack_event(struct notifier_block *this, if (ctnetlink_dump_id(skb, ct) < 0) goto nla_put_failure; + if (ctnetlink_dump_status(skb, ct) < 0) + goto nla_put_failure; + if (events & IPCT_DESTROY) { if (ctnetlink_dump_counters(skb, ct, IP_CT_DIR_ORIGINAL) < 0 || ctnetlink_dump_counters(skb, ct, IP_CT_DIR_REPLY) < 0) goto nla_put_failure; } else { - if (ctnetlink_dump_status(skb, ct) < 0) - goto nla_put_failure; - if (ctnetlink_dump_timeout(skb, ct) < 0) goto nla_put_failure; -- cgit v1.2.2 From e64bda89b8fe81cce9b4a20885d2c204c2d52532 Mon Sep 17 00:00:00 2001 From: Rami Rosen Date: Mon, 9 Jun 2008 16:00:45 -0700 Subject: netfilter: {ip,ip6,nfnetlink}_queue: misc cleanups - No need to perform data_len = 0 in the switch command, since data_len is initialized to 0 in the beginning of the ipq_build_packet_message() method. - {ip,ip6}_queue: We can reach nlmsg_failure only from one place; skb is sure to be NULL when getting there; since skb is NULL, there is no need to check this fact and call kfree_skb(). Signed-off-by: Rami Rosen Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/ipv4/netfilter/ip_queue.c | 3 --- net/ipv6/netfilter/ip6_queue.c | 3 --- net/netfilter/nfnetlink_queue.c | 1 - 3 files changed, 7 deletions(-) (limited to 'net') diff --git a/net/ipv4/netfilter/ip_queue.c b/net/ipv4/netfilter/ip_queue.c index 26a37cedcf2e..aa33a4a7a715 100644 --- a/net/ipv4/netfilter/ip_queue.c +++ b/net/ipv4/netfilter/ip_queue.c @@ -156,7 +156,6 @@ ipq_build_packet_message(struct nf_queue_entry *entry, int *errp) case IPQ_COPY_META: case IPQ_COPY_NONE: size = NLMSG_SPACE(sizeof(*pmsg)); - data_len = 0; break; case IPQ_COPY_PACKET: @@ -224,8 +223,6 @@ ipq_build_packet_message(struct nf_queue_entry *entry, int *errp) return skb; nlmsg_failure: - if (skb) - kfree_skb(skb); *errp = -EINVAL; printk(KERN_ERR "ip_queue: error creating packet message\n"); return NULL; diff --git a/net/ipv6/netfilter/ip6_queue.c b/net/ipv6/netfilter/ip6_queue.c index 2eff3ae8977d..1b8815f6153d 100644 --- a/net/ipv6/netfilter/ip6_queue.c +++ b/net/ipv6/netfilter/ip6_queue.c @@ -159,7 +159,6 @@ ipq_build_packet_message(struct nf_queue_entry *entry, int *errp) case IPQ_COPY_META: case IPQ_COPY_NONE: size = NLMSG_SPACE(sizeof(*pmsg)); - data_len = 0; break; case IPQ_COPY_PACKET: @@ -226,8 +225,6 @@ ipq_build_packet_message(struct nf_queue_entry *entry, int *errp) return skb; nlmsg_failure: - if (skb) - kfree_skb(skb); *errp = -EINVAL; printk(KERN_ERR "ip6_queue: error creating packet message\n"); return NULL; diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c index 3447025ce068..04e9c965f8ca 100644 --- a/net/netfilter/nfnetlink_queue.c +++ b/net/netfilter/nfnetlink_queue.c @@ -243,7 +243,6 @@ nfqnl_build_packet_message(struct nfqnl_instance *queue, switch ((enum nfqnl_config_mode)queue->copy_mode) { case NFQNL_COPY_META: case NFQNL_COPY_NONE: - data_len = 0; break; case NFQNL_COPY_PACKET: -- cgit v1.2.2 From 9a727a250c676334efdcb71a5b2ad4603addda06 Mon Sep 17 00:00:00 2001 From: "John W. Linville" Date: Tue, 10 Jun 2008 13:31:23 -0400 Subject: net/mac80211/ieee80211_i.h: fix-up merge damage These definitions were originally removed in "mac80211: remove channel use statistics". Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 21 --------------------- 1 file changed, 21 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 884be4d100f1..b19bd16703b2 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -853,27 +853,6 @@ u32 ieee80211_handle_ht(struct ieee80211_local *local, int enable_ht, /* ieee80211_ioctl.c */ extern const struct iw_handler_def ieee80211_iw_handler_def; - -/* Least common multiple of the used rates (in 100 kbps). This is used to - * calculate rate_inv values for each rate so that only integers are needed. */ -#define CHAN_UTIL_RATE_LCM 95040 -/* 1 usec is 1/8 * (95040/10) = 1188 */ -#define CHAN_UTIL_PER_USEC 1188 -/* Amount of bits to shift the result right to scale the total utilization - * to values that will not wrap around 32-bit integers. */ -#define CHAN_UTIL_SHIFT 9 -/* Theoretical maximum of channel utilization counter in 10 ms (stat_time=1): - * (CHAN_UTIL_PER_USEC * 10000) >> CHAN_UTIL_SHIFT = 23203. So dividing the - * raw value with about 23 should give utilization in 10th of a percentage - * (1/1000). However, utilization is only estimated and not all intervals - * between frames etc. are calculated. 18 seems to give numbers that are closer - * to the real maximum. */ -#define CHAN_UTIL_PER_10MS 18 -#define CHAN_UTIL_HDR_LONG (202 * CHAN_UTIL_PER_USEC) -#define CHAN_UTIL_HDR_SHORT (40 * CHAN_UTIL_PER_USEC) - - -/* ieee80211_ioctl.c */ int ieee80211_set_freq(struct net_device *dev, int freq); /* ieee80211_sta.c */ -- cgit v1.2.2 From 573bf470e693f73a6ac437b17a64a10902ba54bf Mon Sep 17 00:00:00 2001 From: Thomas Graf Date: Tue, 10 Jun 2008 15:40:04 -0700 Subject: ipv4 addr: Send netlink notification for address label changes Makes people happy who try to keep a list of addresses up to date by listening to notifications. Signed-off-by: Thomas Graf Signed-off-by: David S. Miller --- net/ipv4/devinet.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c index 79a7ef6209ff..61011e1d580e 100644 --- a/net/ipv4/devinet.c +++ b/net/ipv4/devinet.c @@ -1013,7 +1013,7 @@ static void inetdev_changename(struct net_device *dev, struct in_device *in_dev) memcpy(old, ifa->ifa_label, IFNAMSIZ); memcpy(ifa->ifa_label, dev->name, IFNAMSIZ); if (named++ == 0) - continue; + goto skip; dot = strchr(old, ':'); if (dot == NULL) { sprintf(old, ":%d", named); @@ -1024,6 +1024,8 @@ static void inetdev_changename(struct net_device *dev, struct in_device *in_dev) } else { strcpy(ifa->ifa_label + (IFNAMSIZ - strlen(dot) - 1), dot); } +skip: + rtmsg_ifa(RTM_NEWADDR, ifa, NULL, 0); } } -- cgit v1.2.2 From 9cba632e24ef18e2905c18997a8f24e8d7a29e71 Mon Sep 17 00:00:00 2001 From: Rami Rosen Date: Wed, 23 Apr 2008 14:34:00 +0300 Subject: ipv6 mcast: Remove unused macro (MLDV2_QQIC) from mcast.c. This patch removes MLDV2_QQIC macro from mcast.c as it is unused. Signed-off-by: Rami Rosen Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/mcast.c | 1 - 1 file changed, 1 deletion(-) (limited to 'net') diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c index fd632dd7f98d..ee30ec4c3a6e 100644 --- a/net/ipv6/mcast.c +++ b/net/ipv6/mcast.c @@ -164,7 +164,6 @@ static int ip6_mc_leave_src(struct sock *sk, struct ipv6_mc_socklist *iml, ((MLDV2_MASK(value, nbmant) | (1<<(nbmant))) << \ (MLDV2_MASK((value) >> (nbmant), nbexp) + (nbexp)))) -#define MLDV2_QQIC(value) MLDV2_EXP(0x80, 4, 3, value) #define MLDV2_MRC(value) MLDV2_EXP(0x8000, 12, 3, value) #define IPV6_MLD_MAX_MSF 64 -- cgit v1.2.2 From 7d120c55df02a2b87f4aa317f1f04e398398dcdc Mon Sep 17 00:00:00 2001 From: Rami Rosen Date: Wed, 23 Apr 2008 14:35:13 +0300 Subject: ipv6 mroute: Use MRT6_VERSION instead of MRT_VERSION in ip6mr.c. MRT6_VERSION should be used instead of MRT_VERSION in ip6mr.c. Signed-off-by: Rami Rosen Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/ip6mr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c index bf268b386963..0b11b378d89a 100644 --- a/net/ipv6/ip6mr.c +++ b/net/ipv6/ip6mr.c @@ -1240,7 +1240,7 @@ int ip6_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, int #endif /* - * Spurious command, or MRT_VERSION which you cannot + * Spurious command, or MRT6_VERSION which you cannot * set. */ default: -- cgit v1.2.2 From 2b5ead46442d80928cce987ae6acf3fe99968ad8 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Tue, 13 May 2008 01:16:24 +0900 Subject: ipv6 addrconf: Introduce addrconf_is_prefix_route() helper. This inline function, for readability, returns if the route is a "prefix" route regardless if it was installed by RA or by hand. Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/addrconf.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 147588f4c7c0..9ea4e62741ee 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -231,6 +231,12 @@ static inline int addrconf_qdisc_ok(struct net_device *dev) return (dev->qdisc != &noop_qdisc); } +/* Check if a route is valid prefix route */ +static inline int addrconf_is_prefix_route(const struct rt6_info *rt) +{ + return ((rt->rt6i_flags & (RTF_GATEWAY | RTF_DEFAULT)) == 0); +} + static void addrconf_del_timer(struct inet6_ifaddr *ifp) { if (del_timer(&ifp->timer)) @@ -777,7 +783,7 @@ static void ipv6_del_addr(struct inet6_ifaddr *ifp) ipv6_addr_prefix(&prefix, &ifp->addr, ifp->prefix_len); rt = rt6_lookup(net, &prefix, NULL, ifp->idev->dev->ifindex, 1); - if (rt && ((rt->rt6i_flags & (RTF_GATEWAY | RTF_DEFAULT)) == 0)) { + if (rt && addrconf_is_prefix_route(rt)) { if (onlink == 0) { ip6_del_rt(rt); rt = NULL; @@ -1788,7 +1794,7 @@ void addrconf_prefix_rcv(struct net_device *dev, u8 *opt, int len) rt = rt6_lookup(dev_net(dev), &pinfo->prefix, NULL, dev->ifindex, 1); - if (rt && ((rt->rt6i_flags & (RTF_GATEWAY | RTF_DEFAULT)) == 0)) { + if (rt && addrconf_is_prefix_route(rt)) { /* Autoconf prefix route */ if (valid_lft == 0) { ip6_del_rt(rt); -- cgit v1.2.2 From 3de232554a91adc74e80dc15c304be806bd7e1f9 Mon Sep 17 00:00:00 2001 From: Benjamin Thery Date: Wed, 28 May 2008 14:51:24 +0200 Subject: ipv6 netns: Address labels per namespace This pacth makes IPv6 address labels per network namespace. It keeps the global label tables, ip6addrlbl_table, but adds a 'net' member to each ip6addrlbl_entry. This new member is taken into account when matching labels. Changelog ========= * v1: Initial version * v2: * Minize the penalty when network namespaces are not configured: * the 'net' member is added only if CONFIG_NET_NS is defined. This saves space when network namespaces are not configured. * 'net' value is retrieved with the inlined function ip6addrlbl_net() that always return &init_net when CONFIG_NET_NS is not defined. * 'net' member in ip6addrlbl_entry renamed to the less generic 'lbl_net' name (helps code search). Signed-off-by: Benjamin Thery Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/addrconf.c | 12 +++--- net/ipv6/addrlabel.c | 106 ++++++++++++++++++++++++++++++++++++--------------- 2 files changed, 82 insertions(+), 36 deletions(-) (limited to 'net') diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 9ea4e62741ee..fa43374e85c5 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -964,7 +964,8 @@ static inline int ipv6_saddr_preferred(int type) return 0; } -static int ipv6_get_saddr_eval(struct ipv6_saddr_score *score, +static int ipv6_get_saddr_eval(struct net *net, + struct ipv6_saddr_score *score, struct ipv6_saddr_dst *dst, int i) { @@ -1043,7 +1044,8 @@ static int ipv6_get_saddr_eval(struct ipv6_saddr_score *score, break; case IPV6_SADDR_RULE_LABEL: /* Rule 6: Prefer matching label */ - ret = ipv6_addr_label(&score->ifa->addr, score->addr_type, + ret = ipv6_addr_label(net, + &score->ifa->addr, score->addr_type, score->ifa->idev->dev->ifindex) == dst->label; break; #ifdef CONFIG_IPV6_PRIVACY @@ -1097,7 +1099,7 @@ int ipv6_dev_get_saddr(struct net_device *dst_dev, dst.addr = daddr; dst.ifindex = dst_dev ? dst_dev->ifindex : 0; dst.scope = __ipv6_addr_src_scope(dst_type); - dst.label = ipv6_addr_label(daddr, dst_type, dst.ifindex); + dst.label = ipv6_addr_label(net, daddr, dst_type, dst.ifindex); dst.prefs = prefs; hiscore->rule = -1; @@ -1165,8 +1167,8 @@ int ipv6_dev_get_saddr(struct net_device *dst_dev, for (i = 0; i < IPV6_SADDR_RULE_MAX; i++) { int minihiscore, miniscore; - minihiscore = ipv6_get_saddr_eval(hiscore, &dst, i); - miniscore = ipv6_get_saddr_eval(score, &dst, i); + minihiscore = ipv6_get_saddr_eval(net, hiscore, &dst, i); + miniscore = ipv6_get_saddr_eval(net, score, &dst, i); if (minihiscore > miniscore) { if (i == IPV6_SADDR_RULE_SCOPE && diff --git a/net/ipv6/addrlabel.c b/net/ipv6/addrlabel.c index 9bfa8846f262..08909039d87b 100644 --- a/net/ipv6/addrlabel.c +++ b/net/ipv6/addrlabel.c @@ -29,6 +29,9 @@ */ struct ip6addrlbl_entry { +#ifdef CONFIG_NET_NS + struct net *lbl_net; +#endif struct in6_addr prefix; int prefixlen; int ifindex; @@ -46,6 +49,16 @@ static struct ip6addrlbl_table u32 seq; } ip6addrlbl_table; +static inline +struct net *ip6addrlbl_net(const struct ip6addrlbl_entry *lbl) +{ +#ifdef CONFIG_NET_NS + return lbl->lbl_net; +#else + return &init_net; +#endif +} + /* * Default policy table (RFC3484 + extensions) * @@ -65,7 +78,7 @@ static struct ip6addrlbl_table #define IPV6_ADDR_LABEL_DEFAULT 0xffffffffUL -static const __initdata struct ip6addrlbl_init_table +static const __net_initdata struct ip6addrlbl_init_table { const struct in6_addr *prefix; int prefixlen; @@ -108,6 +121,9 @@ static const __initdata struct ip6addrlbl_init_table /* Object management */ static inline void ip6addrlbl_free(struct ip6addrlbl_entry *p) { +#ifdef CONFIG_NET_NS + release_net(p->lbl_net); +#endif kfree(p); } @@ -128,10 +144,13 @@ static inline void ip6addrlbl_put(struct ip6addrlbl_entry *p) } /* Find label */ -static int __ip6addrlbl_match(struct ip6addrlbl_entry *p, +static int __ip6addrlbl_match(struct net *net, + struct ip6addrlbl_entry *p, const struct in6_addr *addr, int addrtype, int ifindex) { + if (!net_eq(ip6addrlbl_net(p), net)) + return 0; if (p->ifindex && p->ifindex != ifindex) return 0; if (p->addrtype && p->addrtype != addrtype) @@ -141,19 +160,21 @@ static int __ip6addrlbl_match(struct ip6addrlbl_entry *p, return 1; } -static struct ip6addrlbl_entry *__ipv6_addr_label(const struct in6_addr *addr, +static struct ip6addrlbl_entry *__ipv6_addr_label(struct net *net, + const struct in6_addr *addr, int type, int ifindex) { struct hlist_node *pos; struct ip6addrlbl_entry *p; hlist_for_each_entry_rcu(p, pos, &ip6addrlbl_table.head, list) { - if (__ip6addrlbl_match(p, addr, type, ifindex)) + if (__ip6addrlbl_match(net, p, addr, type, ifindex)) return p; } return NULL; } -u32 ipv6_addr_label(const struct in6_addr *addr, int type, int ifindex) +u32 ipv6_addr_label(struct net *net, + const struct in6_addr *addr, int type, int ifindex) { u32 label; struct ip6addrlbl_entry *p; @@ -161,7 +182,7 @@ u32 ipv6_addr_label(const struct in6_addr *addr, int type, int ifindex) type &= IPV6_ADDR_MAPPED | IPV6_ADDR_COMPATv4 | IPV6_ADDR_LOOPBACK; rcu_read_lock(); - p = __ipv6_addr_label(addr, type, ifindex); + p = __ipv6_addr_label(net, addr, type, ifindex); label = p ? p->label : IPV6_ADDR_LABEL_DEFAULT; rcu_read_unlock(); @@ -174,7 +195,8 @@ u32 ipv6_addr_label(const struct in6_addr *addr, int type, int ifindex) } /* allocate one entry */ -static struct ip6addrlbl_entry *ip6addrlbl_alloc(const struct in6_addr *prefix, +static struct ip6addrlbl_entry *ip6addrlbl_alloc(struct net *net, + const struct in6_addr *prefix, int prefixlen, int ifindex, u32 label) { @@ -216,6 +238,9 @@ static struct ip6addrlbl_entry *ip6addrlbl_alloc(const struct in6_addr *prefix, newp->addrtype = addrtype; newp->label = label; INIT_HLIST_NODE(&newp->list); +#ifdef CONFIG_NET_NS + newp->lbl_net = hold_net(net); +#endif atomic_set(&newp->refcnt, 1); return newp; } @@ -237,6 +262,7 @@ static int __ip6addrlbl_add(struct ip6addrlbl_entry *newp, int replace) hlist_for_each_entry_safe(p, pos, n, &ip6addrlbl_table.head, list) { if (p->prefixlen == newp->prefixlen && + net_eq(ip6addrlbl_net(p), ip6addrlbl_net(newp)) && p->ifindex == newp->ifindex && ipv6_addr_equal(&p->prefix, &newp->prefix)) { if (!replace) { @@ -261,7 +287,8 @@ out: } /* add a label */ -static int ip6addrlbl_add(const struct in6_addr *prefix, int prefixlen, +static int ip6addrlbl_add(struct net *net, + const struct in6_addr *prefix, int prefixlen, int ifindex, u32 label, int replace) { struct ip6addrlbl_entry *newp; @@ -274,7 +301,7 @@ static int ip6addrlbl_add(const struct in6_addr *prefix, int prefixlen, (unsigned int)label, replace); - newp = ip6addrlbl_alloc(prefix, prefixlen, ifindex, label); + newp = ip6addrlbl_alloc(net, prefix, prefixlen, ifindex, label); if (IS_ERR(newp)) return PTR_ERR(newp); spin_lock(&ip6addrlbl_table.lock); @@ -286,7 +313,8 @@ static int ip6addrlbl_add(const struct in6_addr *prefix, int prefixlen, } /* remove a label */ -static int __ip6addrlbl_del(const struct in6_addr *prefix, int prefixlen, +static int __ip6addrlbl_del(struct net *net, + const struct in6_addr *prefix, int prefixlen, int ifindex) { struct ip6addrlbl_entry *p = NULL; @@ -300,6 +328,7 @@ static int __ip6addrlbl_del(const struct in6_addr *prefix, int prefixlen, hlist_for_each_entry_safe(p, pos, n, &ip6addrlbl_table.head, list) { if (p->prefixlen == prefixlen && + net_eq(ip6addrlbl_net(p), net) && p->ifindex == ifindex && ipv6_addr_equal(&p->prefix, prefix)) { hlist_del_rcu(&p->list); @@ -311,7 +340,8 @@ static int __ip6addrlbl_del(const struct in6_addr *prefix, int prefixlen, return ret; } -static int ip6addrlbl_del(const struct in6_addr *prefix, int prefixlen, +static int ip6addrlbl_del(struct net *net, + const struct in6_addr *prefix, int prefixlen, int ifindex) { struct in6_addr prefix_buf; @@ -324,13 +354,13 @@ static int ip6addrlbl_del(const struct in6_addr *prefix, int prefixlen, ipv6_addr_prefix(&prefix_buf, prefix, prefixlen); spin_lock(&ip6addrlbl_table.lock); - ret = __ip6addrlbl_del(&prefix_buf, prefixlen, ifindex); + ret = __ip6addrlbl_del(net, &prefix_buf, prefixlen, ifindex); spin_unlock(&ip6addrlbl_table.lock); return ret; } /* add default label */ -static __init int ip6addrlbl_init(void) +static int __net_init ip6addrlbl_net_init(struct net *net) { int err = 0; int i; @@ -338,7 +368,8 @@ static __init int ip6addrlbl_init(void) ADDRLABEL(KERN_DEBUG "%s()\n", __func__); for (i = 0; i < ARRAY_SIZE(ip6addrlbl_init_table); i++) { - int ret = ip6addrlbl_add(ip6addrlbl_init_table[i].prefix, + int ret = ip6addrlbl_add(net, + ip6addrlbl_init_table[i].prefix, ip6addrlbl_init_table[i].prefixlen, 0, ip6addrlbl_init_table[i].label, 0); @@ -349,11 +380,32 @@ static __init int ip6addrlbl_init(void) return err; } +static void __net_exit ip6addrlbl_net_exit(struct net *net) +{ + struct ip6addrlbl_entry *p = NULL; + struct hlist_node *pos, *n; + + /* Remove all labels belonging to the exiting net */ + spin_lock(&ip6addrlbl_table.lock); + hlist_for_each_entry_safe(p, pos, n, &ip6addrlbl_table.head, list) { + if (net_eq(ip6addrlbl_net(p), net)) { + hlist_del_rcu(&p->list); + ip6addrlbl_put(p); + } + } + spin_unlock(&ip6addrlbl_table.lock); +} + +static struct pernet_operations ipv6_addr_label_ops = { + .init = ip6addrlbl_net_init, + .exit = ip6addrlbl_net_exit, +}; + int __init ipv6_addr_label_init(void) { spin_lock_init(&ip6addrlbl_table.lock); - return ip6addrlbl_init(); + return register_pernet_subsys(&ipv6_addr_label_ops); } static const struct nla_policy ifal_policy[IFAL_MAX+1] = { @@ -371,9 +423,6 @@ static int ip6addrlbl_newdel(struct sk_buff *skb, struct nlmsghdr *nlh, u32 label; int err = 0; - if (net != &init_net) - return 0; - err = nlmsg_parse(nlh, sizeof(*ifal), tb, IFAL_MAX, ifal_policy); if (err < 0) return err; @@ -385,7 +434,7 @@ static int ip6addrlbl_newdel(struct sk_buff *skb, struct nlmsghdr *nlh, return -EINVAL; if (ifal->ifal_index && - !__dev_get_by_index(&init_net, ifal->ifal_index)) + !__dev_get_by_index(net, ifal->ifal_index)) return -EINVAL; if (!tb[IFAL_ADDRESS]) @@ -403,12 +452,12 @@ static int ip6addrlbl_newdel(struct sk_buff *skb, struct nlmsghdr *nlh, switch(nlh->nlmsg_type) { case RTM_NEWADDRLABEL: - err = ip6addrlbl_add(pfx, ifal->ifal_prefixlen, + err = ip6addrlbl_add(net, pfx, ifal->ifal_prefixlen, ifal->ifal_index, label, nlh->nlmsg_flags & NLM_F_REPLACE); break; case RTM_DELADDRLABEL: - err = ip6addrlbl_del(pfx, ifal->ifal_prefixlen, + err = ip6addrlbl_del(net, pfx, ifal->ifal_prefixlen, ifal->ifal_index); break; default: @@ -458,12 +507,10 @@ static int ip6addrlbl_dump(struct sk_buff *skb, struct netlink_callback *cb) int idx = 0, s_idx = cb->args[0]; int err; - if (net != &init_net) - return 0; - rcu_read_lock(); hlist_for_each_entry_rcu(p, pos, &ip6addrlbl_table.head, list) { - if (idx >= s_idx) { + if (idx >= s_idx && + net_eq(ip6addrlbl_net(p), net)) { if ((err = ip6addrlbl_fill(skb, p, ip6addrlbl_table.seq, NETLINK_CB(cb->skb).pid, @@ -499,9 +546,6 @@ static int ip6addrlbl_get(struct sk_buff *in_skb, struct nlmsghdr* nlh, struct ip6addrlbl_entry *p; struct sk_buff *skb; - if (net != &init_net) - return 0; - err = nlmsg_parse(nlh, sizeof(*ifal), tb, IFAL_MAX, ifal_policy); if (err < 0) return err; @@ -513,7 +557,7 @@ static int ip6addrlbl_get(struct sk_buff *in_skb, struct nlmsghdr* nlh, return -EINVAL; if (ifal->ifal_index && - !__dev_get_by_index(&init_net, ifal->ifal_index)) + !__dev_get_by_index(net, ifal->ifal_index)) return -EINVAL; if (!tb[IFAL_ADDRESS]) @@ -524,7 +568,7 @@ static int ip6addrlbl_get(struct sk_buff *in_skb, struct nlmsghdr* nlh, return -EINVAL; rcu_read_lock(); - p = __ipv6_addr_label(addr, ipv6_addr_type(addr), ifal->ifal_index); + p = __ipv6_addr_label(net, addr, ipv6_addr_type(addr), ifal->ifal_index); if (p && ip6addrlbl_hold(p)) p = NULL; lseq = ip6addrlbl_table.seq; @@ -552,7 +596,7 @@ static int ip6addrlbl_get(struct sk_buff *in_skb, struct nlmsghdr* nlh, goto out; } - err = rtnl_unicast(skb, &init_net, NETLINK_CB(in_skb).pid); + err = rtnl_unicast(skb, net, NETLINK_CB(in_skb).pid); out: return err; } -- cgit v1.2.2 From 9e8b4ed8bb7d2f0f91dccf0abf648771d76e7a01 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Mon, 28 Apr 2008 00:50:45 +0900 Subject: key: Introduce pfkey_sockaddr_len() for raw sockaddr{} length. Signed-off-by: YOSHIFUJI Hideaki --- net/key/af_key.c | 50 +++++++++++++++++++------------------------------- 1 file changed, 19 insertions(+), 31 deletions(-) (limited to 'net') diff --git a/net/key/af_key.c b/net/key/af_key.c index 9bba7ac5fee0..02c8aba4239e 100644 --- a/net/key/af_key.c +++ b/net/key/af_key.c @@ -579,6 +579,19 @@ static uint8_t pfkey_proto_from_xfrm(uint8_t proto) return (proto ? proto : IPSEC_PROTO_ANY); } +static inline int pfkey_sockaddr_len(sa_family_t family) +{ + switch (family) { + case AF_INET: + return sizeof(struct sockaddr_in); +#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) + case AF_INET6: + return sizeof(struct sockaddr_in6); +#endif + } + return 0; +} + static int pfkey_sadb_addr2xfrm_addr(struct sadb_address *addr, xfrm_address_t *xaddr) { @@ -642,20 +655,11 @@ static struct xfrm_state *pfkey_xfrm_state_lookup(struct sadb_msg *hdr, void ** } #define PFKEY_ALIGN8(a) (1 + (((a) - 1) | (8 - 1))) + static int pfkey_sockaddr_size(sa_family_t family) { - switch (family) { - case AF_INET: - return PFKEY_ALIGN8(sizeof(struct sockaddr_in)); -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - case AF_INET6: - return PFKEY_ALIGN8(sizeof(struct sockaddr_in6)); -#endif - default: - return 0; - } - /* NOTREACHED */ + return PFKEY_ALIGN8(pfkey_sockaddr_len(family)); } static inline int pfkey_mode_from_xfrm(int mode) @@ -1952,9 +1956,7 @@ static int pfkey_xfrm_policy2msg_size(struct xfrm_policy *xp) for (i=0; ixfrm_nr; i++) { t = xp->xfrm_vec + i; - socklen += (t->encap_family == AF_INET ? - sizeof(struct sockaddr_in) : - sizeof(struct sockaddr_in6)); + socklen += pfkey_sockaddr_len(t->encap_family); } return sizeof(struct sadb_msg) + @@ -1996,9 +1998,7 @@ static int pfkey_xfrm_policy2msg(struct sk_buff *skb, struct xfrm_policy *xp, in int i; int size; int sockaddr_size = pfkey_sockaddr_size(xp->family); - int socklen = (xp->family == AF_INET ? - sizeof(struct sockaddr_in) : - sizeof(struct sockaddr_in6)); + int socklen = pfkey_sockaddr_len(xp->family); size = pfkey_xfrm_policy2msg_size(xp); @@ -2122,9 +2122,7 @@ static int pfkey_xfrm_policy2msg(struct sk_buff *skb, struct xfrm_policy *xp, in req_size = sizeof(struct sadb_x_ipsecrequest); if (t->mode == XFRM_MODE_TUNNEL) - req_size += ((t->encap_family == AF_INET ? - sizeof(struct sockaddr_in) : - sizeof(struct sockaddr_in6)) * 2); + req_size += pfkey_sockaddr_len(t->encap_family) * 2; else size -= 2*socklen; rq = (void*)skb_put(skb, req_size); @@ -2459,17 +2457,7 @@ out: #ifdef CONFIG_NET_KEY_MIGRATE static int pfkey_sockaddr_pair_size(sa_family_t family) { - switch (family) { - case AF_INET: - return PFKEY_ALIGN8(sizeof(struct sockaddr_in) * 2); -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - case AF_INET6: - return PFKEY_ALIGN8(sizeof(struct sockaddr_in6) * 2); -#endif - default: - return 0; - } - /* NOTREACHED */ + return PFKEY_ALIGN8(pfkey_sockaddr_len(family) * 2); } static int parse_sockaddr_pair(struct sadb_x_ipsecrequest *rq, -- cgit v1.2.2 From e5b56652c11baf144eb713c33ff70a1b5d6d09bc Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Mon, 28 Apr 2008 01:46:41 +0900 Subject: key: Share common code path to fill sockaddr{}. Signed-off-by: YOSHIFUJI Hideaki --- net/key/af_key.c | 458 ++++++++++++++----------------------------------------- 1 file changed, 111 insertions(+), 347 deletions(-) (limited to 'net') diff --git a/net/key/af_key.c b/net/key/af_key.c index 02c8aba4239e..2b05cb6c15c6 100644 --- a/net/key/af_key.c +++ b/net/key/af_key.c @@ -691,6 +691,36 @@ static inline int pfkey_mode_to_xfrm(int mode) } } +static unsigned int pfkey_sockaddr_fill(xfrm_address_t *xaddr, __be16 port, + struct sockaddr *sa, + unsigned short family) +{ + switch (family) { + case AF_INET: + { + struct sockaddr_in *sin = (struct sockaddr_in *)sa; + sin->sin_family = AF_INET; + sin->sin_port = port; + sin->sin_addr.s_addr = xaddr->a4; + memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); + return 32; + } +#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) + case AF_INET6: + { + struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)sa; + sin6->sin6_family = AF_INET6; + sin6->sin6_port = port; + sin6->sin6_flowinfo = 0; + ipv6_addr_copy(&sin6->sin6_addr, (struct in6_addr *)xaddr->a6); + sin6->sin6_scope_id = 0; + return 128; + } +#endif + } + return 0; +} + static struct sk_buff *__pfkey_xfrm_state2msg(struct xfrm_state *x, int add_keys, int hsc) { @@ -701,13 +731,9 @@ static struct sk_buff *__pfkey_xfrm_state2msg(struct xfrm_state *x, struct sadb_address *addr; struct sadb_key *key; struct sadb_x_sa2 *sa2; - struct sockaddr_in *sin; struct sadb_x_sec_ctx *sec_ctx; struct xfrm_sec_ctx *xfrm_ctx; int ctx_size = 0; -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - struct sockaddr_in6 *sin6; -#endif int size; int auth_key_size = 0; int encrypt_key_size = 0; @@ -865,29 +891,12 @@ static struct sk_buff *__pfkey_xfrm_state2msg(struct xfrm_state *x, protocol's number." - RFC2367 */ addr->sadb_address_proto = 0; addr->sadb_address_reserved = 0; - if (x->props.family == AF_INET) { - addr->sadb_address_prefixlen = 32; - sin = (struct sockaddr_in *) (addr + 1); - sin->sin_family = AF_INET; - sin->sin_addr.s_addr = x->props.saddr.a4; - sin->sin_port = 0; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - } -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - else if (x->props.family == AF_INET6) { - addr->sadb_address_prefixlen = 128; - - sin6 = (struct sockaddr_in6 *) (addr + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - memcpy(&sin6->sin6_addr, x->props.saddr.a6, - sizeof(struct in6_addr)); - sin6->sin6_scope_id = 0; - } -#endif - else + addr->sadb_address_prefixlen = + pfkey_sockaddr_fill(&x->props.saddr, 0, + (struct sockaddr *) (addr + 1), + x->props.family); + if (!addr->sadb_address_prefixlen) BUG(); /* dst address */ @@ -898,70 +907,32 @@ static struct sk_buff *__pfkey_xfrm_state2msg(struct xfrm_state *x, sizeof(uint64_t); addr->sadb_address_exttype = SADB_EXT_ADDRESS_DST; addr->sadb_address_proto = 0; - addr->sadb_address_prefixlen = 32; /* XXX */ addr->sadb_address_reserved = 0; - if (x->props.family == AF_INET) { - sin = (struct sockaddr_in *) (addr + 1); - sin->sin_family = AF_INET; - sin->sin_addr.s_addr = x->id.daddr.a4; - sin->sin_port = 0; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - if (x->sel.saddr.a4 != x->props.saddr.a4) { - addr = (struct sadb_address*) skb_put(skb, - sizeof(struct sadb_address)+sockaddr_size); - addr->sadb_address_len = - (sizeof(struct sadb_address)+sockaddr_size)/ - sizeof(uint64_t); - addr->sadb_address_exttype = SADB_EXT_ADDRESS_PROXY; - addr->sadb_address_proto = - pfkey_proto_from_xfrm(x->sel.proto); - addr->sadb_address_prefixlen = x->sel.prefixlen_s; - addr->sadb_address_reserved = 0; - - sin = (struct sockaddr_in *) (addr + 1); - sin->sin_family = AF_INET; - sin->sin_addr.s_addr = x->sel.saddr.a4; - sin->sin_port = x->sel.sport; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - } - } -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - else if (x->props.family == AF_INET6) { - addr->sadb_address_prefixlen = 128; + addr->sadb_address_prefixlen = + pfkey_sockaddr_fill(&x->id.daddr, 0, + (struct sockaddr *) (addr + 1), + x->props.family); + if (!addr->sadb_address_prefixlen) + BUG(); - sin6 = (struct sockaddr_in6 *) (addr + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - memcpy(&sin6->sin6_addr, x->id.daddr.a6, sizeof(struct in6_addr)); - sin6->sin6_scope_id = 0; + if (xfrm_addr_cmp(&x->sel.saddr, &x->props.saddr, + x->props.family)) { + addr = (struct sadb_address*) skb_put(skb, + sizeof(struct sadb_address)+sockaddr_size); + addr->sadb_address_len = + (sizeof(struct sadb_address)+sockaddr_size)/ + sizeof(uint64_t); + addr->sadb_address_exttype = SADB_EXT_ADDRESS_PROXY; + addr->sadb_address_proto = + pfkey_proto_from_xfrm(x->sel.proto); + addr->sadb_address_prefixlen = x->sel.prefixlen_s; + addr->sadb_address_reserved = 0; - if (memcmp (x->sel.saddr.a6, x->props.saddr.a6, - sizeof(struct in6_addr))) { - addr = (struct sadb_address *) skb_put(skb, - sizeof(struct sadb_address)+sockaddr_size); - addr->sadb_address_len = - (sizeof(struct sadb_address)+sockaddr_size)/ - sizeof(uint64_t); - addr->sadb_address_exttype = SADB_EXT_ADDRESS_PROXY; - addr->sadb_address_proto = - pfkey_proto_from_xfrm(x->sel.proto); - addr->sadb_address_prefixlen = x->sel.prefixlen_s; - addr->sadb_address_reserved = 0; - - sin6 = (struct sockaddr_in6 *) (addr + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = x->sel.sport; - sin6->sin6_flowinfo = 0; - memcpy(&sin6->sin6_addr, x->sel.saddr.a6, - sizeof(struct in6_addr)); - sin6->sin6_scope_id = 0; - } + pfkey_sockaddr_fill(&x->sel.saddr, x->sel.sport, + (struct sockaddr *) (addr + 1), + x->props.family); } -#endif - else - BUG(); /* auth key */ if (add_keys && auth_key_size) { @@ -1989,12 +1960,8 @@ static int pfkey_xfrm_policy2msg(struct sk_buff *skb, struct xfrm_policy *xp, in struct sadb_address *addr; struct sadb_lifetime *lifetime; struct sadb_x_policy *pol; - struct sockaddr_in *sin; struct sadb_x_sec_ctx *sec_ctx; struct xfrm_sec_ctx *xfrm_ctx; -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - struct sockaddr_in6 *sin6; -#endif int i; int size; int sockaddr_size = pfkey_sockaddr_size(xp->family); @@ -2016,26 +1983,10 @@ static int pfkey_xfrm_policy2msg(struct sk_buff *skb, struct xfrm_policy *xp, in addr->sadb_address_proto = pfkey_proto_from_xfrm(xp->selector.proto); addr->sadb_address_prefixlen = xp->selector.prefixlen_s; addr->sadb_address_reserved = 0; - /* src address */ - if (xp->family == AF_INET) { - sin = (struct sockaddr_in *) (addr + 1); - sin->sin_family = AF_INET; - sin->sin_addr.s_addr = xp->selector.saddr.a4; - sin->sin_port = xp->selector.sport; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - } -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - else if (xp->family == AF_INET6) { - sin6 = (struct sockaddr_in6 *) (addr + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = xp->selector.sport; - sin6->sin6_flowinfo = 0; - memcpy(&sin6->sin6_addr, xp->selector.saddr.a6, - sizeof(struct in6_addr)); - sin6->sin6_scope_id = 0; - } -#endif - else + if (!pfkey_sockaddr_fill(&xp->selector.saddr, + xp->selector.sport, + (struct sockaddr *) (addr + 1), + xp->family)) BUG(); /* dst address */ @@ -2048,26 +1999,10 @@ static int pfkey_xfrm_policy2msg(struct sk_buff *skb, struct xfrm_policy *xp, in addr->sadb_address_proto = pfkey_proto_from_xfrm(xp->selector.proto); addr->sadb_address_prefixlen = xp->selector.prefixlen_d; addr->sadb_address_reserved = 0; - if (xp->family == AF_INET) { - sin = (struct sockaddr_in *) (addr + 1); - sin->sin_family = AF_INET; - sin->sin_addr.s_addr = xp->selector.daddr.a4; - sin->sin_port = xp->selector.dport; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - } -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - else if (xp->family == AF_INET6) { - sin6 = (struct sockaddr_in6 *) (addr + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = xp->selector.dport; - sin6->sin6_flowinfo = 0; - memcpy(&sin6->sin6_addr, xp->selector.daddr.a6, - sizeof(struct in6_addr)); - sin6->sin6_scope_id = 0; - } -#endif - else - BUG(); + + pfkey_sockaddr_fill(&xp->selector.daddr, xp->selector.dport, + (struct sockaddr *) (addr + 1), + xp->family); /* hard time */ lifetime = (struct sadb_lifetime *) skb_put(skb, @@ -2121,10 +2056,13 @@ static int pfkey_xfrm_policy2msg(struct sk_buff *skb, struct xfrm_policy *xp, in int mode; req_size = sizeof(struct sadb_x_ipsecrequest); - if (t->mode == XFRM_MODE_TUNNEL) - req_size += pfkey_sockaddr_len(t->encap_family) * 2; - else + if (t->mode == XFRM_MODE_TUNNEL) { + socklen = pfkey_sockaddr_len(t->encap_family); + req_size += socklen * 2; + } else { size -= 2*socklen; + socklen = 0; + } rq = (void*)skb_put(skb, req_size); pol->sadb_x_policy_len += req_size/8; memset(rq, 0, sizeof(*rq)); @@ -2139,42 +2077,15 @@ static int pfkey_xfrm_policy2msg(struct sk_buff *skb, struct xfrm_policy *xp, in if (t->optional) rq->sadb_x_ipsecrequest_level = IPSEC_LEVEL_USE; rq->sadb_x_ipsecrequest_reqid = t->reqid; + if (t->mode == XFRM_MODE_TUNNEL) { - switch (t->encap_family) { - case AF_INET: - sin = (void*)(rq+1); - sin->sin_family = AF_INET; - sin->sin_addr.s_addr = t->saddr.a4; - sin->sin_port = 0; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - sin++; - sin->sin_family = AF_INET; - sin->sin_addr.s_addr = t->id.daddr.a4; - sin->sin_port = 0; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - break; -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - case AF_INET6: - sin6 = (void*)(rq+1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - memcpy(&sin6->sin6_addr, t->saddr.a6, - sizeof(struct in6_addr)); - sin6->sin6_scope_id = 0; - - sin6++; - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - memcpy(&sin6->sin6_addr, t->id.daddr.a6, - sizeof(struct in6_addr)); - sin6->sin6_scope_id = 0; - break; -#endif - default: - break; - } + u8 *sa = (void *)(rq + 1); + pfkey_sockaddr_fill(&t->saddr, 0, + (struct sockaddr *)sa, + t->encap_family); + pfkey_sockaddr_fill(&t->id.daddr, 0, + (struct sockaddr *) (sa + socklen), + t->encap_family); } } @@ -3079,10 +2990,6 @@ static int pfkey_send_acquire(struct xfrm_state *x, struct xfrm_tmpl *t, struct struct sadb_msg *hdr; struct sadb_address *addr; struct sadb_x_policy *pol; - struct sockaddr_in *sin; -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - struct sockaddr_in6 *sin6; -#endif int sockaddr_size; int size; struct sadb_x_sec_ctx *sec_ctx; @@ -3131,29 +3038,11 @@ static int pfkey_send_acquire(struct xfrm_state *x, struct xfrm_tmpl *t, struct addr->sadb_address_exttype = SADB_EXT_ADDRESS_SRC; addr->sadb_address_proto = 0; addr->sadb_address_reserved = 0; - if (x->props.family == AF_INET) { - addr->sadb_address_prefixlen = 32; - - sin = (struct sockaddr_in *) (addr + 1); - sin->sin_family = AF_INET; - sin->sin_addr.s_addr = x->props.saddr.a4; - sin->sin_port = 0; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - } -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - else if (x->props.family == AF_INET6) { - addr->sadb_address_prefixlen = 128; - - sin6 = (struct sockaddr_in6 *) (addr + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - memcpy(&sin6->sin6_addr, - x->props.saddr.a6, sizeof(struct in6_addr)); - sin6->sin6_scope_id = 0; - } -#endif - else + addr->sadb_address_prefixlen = + pfkey_sockaddr_fill(&x->props.saddr, 0, + (struct sockaddr *) (addr + 1), + x->props.family); + if (!addr->sadb_address_prefixlen) BUG(); /* dst address */ @@ -3165,29 +3054,11 @@ static int pfkey_send_acquire(struct xfrm_state *x, struct xfrm_tmpl *t, struct addr->sadb_address_exttype = SADB_EXT_ADDRESS_DST; addr->sadb_address_proto = 0; addr->sadb_address_reserved = 0; - if (x->props.family == AF_INET) { - addr->sadb_address_prefixlen = 32; - - sin = (struct sockaddr_in *) (addr + 1); - sin->sin_family = AF_INET; - sin->sin_addr.s_addr = x->id.daddr.a4; - sin->sin_port = 0; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - } -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - else if (x->props.family == AF_INET6) { - addr->sadb_address_prefixlen = 128; - - sin6 = (struct sockaddr_in6 *) (addr + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - memcpy(&sin6->sin6_addr, - x->id.daddr.a6, sizeof(struct in6_addr)); - sin6->sin6_scope_id = 0; - } -#endif - else + addr->sadb_address_prefixlen = + pfkey_sockaddr_fill(&x->id.daddr, 0, + (struct sockaddr *) (addr + 1), + x->props.family); + if (!addr->sadb_address_prefixlen) BUG(); pol = (struct sadb_x_policy *) skb_put(skb, sizeof(struct sadb_x_policy)); @@ -3313,10 +3184,6 @@ static int pfkey_send_new_mapping(struct xfrm_state *x, xfrm_address_t *ipaddr, struct sadb_sa *sa; struct sadb_address *addr; struct sadb_x_nat_t_port *n_port; - struct sockaddr_in *sin; -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - struct sockaddr_in6 *sin6; -#endif int sockaddr_size; int size; __u8 satype = (x->id.proto == IPPROTO_ESP ? SADB_SATYPE_ESP : 0); @@ -3380,29 +3247,11 @@ static int pfkey_send_new_mapping(struct xfrm_state *x, xfrm_address_t *ipaddr, addr->sadb_address_exttype = SADB_EXT_ADDRESS_SRC; addr->sadb_address_proto = 0; addr->sadb_address_reserved = 0; - if (x->props.family == AF_INET) { - addr->sadb_address_prefixlen = 32; - - sin = (struct sockaddr_in *) (addr + 1); - sin->sin_family = AF_INET; - sin->sin_addr.s_addr = x->props.saddr.a4; - sin->sin_port = 0; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - } -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - else if (x->props.family == AF_INET6) { - addr->sadb_address_prefixlen = 128; - - sin6 = (struct sockaddr_in6 *) (addr + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - memcpy(&sin6->sin6_addr, - x->props.saddr.a6, sizeof(struct in6_addr)); - sin6->sin6_scope_id = 0; - } -#endif - else + addr->sadb_address_prefixlen = + pfkey_sockaddr_fill(&x->props.saddr, 0, + (struct sockaddr *) (addr + 1), + x->props.family); + if (!addr->sadb_address_prefixlen) BUG(); /* NAT_T_SPORT (old port) */ @@ -3421,28 +3270,11 @@ static int pfkey_send_new_mapping(struct xfrm_state *x, xfrm_address_t *ipaddr, addr->sadb_address_exttype = SADB_EXT_ADDRESS_DST; addr->sadb_address_proto = 0; addr->sadb_address_reserved = 0; - if (x->props.family == AF_INET) { - addr->sadb_address_prefixlen = 32; - - sin = (struct sockaddr_in *) (addr + 1); - sin->sin_family = AF_INET; - sin->sin_addr.s_addr = ipaddr->a4; - sin->sin_port = 0; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - } -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - else if (x->props.family == AF_INET6) { - addr->sadb_address_prefixlen = 128; - - sin6 = (struct sockaddr_in6 *) (addr + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - memcpy(&sin6->sin6_addr, &ipaddr->a6, sizeof(struct in6_addr)); - sin6->sin6_scope_id = 0; - } -#endif - else + addr->sadb_address_prefixlen = + pfkey_sockaddr_fill(ipaddr, 0, + (struct sockaddr *) (addr + 1), + x->props.family); + if (!addr->sadb_address_prefixlen) BUG(); /* NAT_T_DPORT (new port) */ @@ -3460,10 +3292,6 @@ static int set_sadb_address(struct sk_buff *skb, int sasize, int type, struct xfrm_selector *sel) { struct sadb_address *addr; - struct sockaddr_in *sin; -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - struct sockaddr_in6 *sin6; -#endif addr = (struct sadb_address *)skb_put(skb, sizeof(struct sadb_address) + sasize); addr->sadb_address_len = (sizeof(struct sadb_address) + sasize)/8; addr->sadb_address_exttype = type; @@ -3472,50 +3300,16 @@ static int set_sadb_address(struct sk_buff *skb, int sasize, int type, switch (type) { case SADB_EXT_ADDRESS_SRC: - if (sel->family == AF_INET) { - addr->sadb_address_prefixlen = sel->prefixlen_s; - sin = (struct sockaddr_in *)(addr + 1); - sin->sin_family = AF_INET; - memcpy(&sin->sin_addr.s_addr, &sel->saddr, - sizeof(sin->sin_addr.s_addr)); - sin->sin_port = 0; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - } -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - else if (sel->family == AF_INET6) { - addr->sadb_address_prefixlen = sel->prefixlen_s; - sin6 = (struct sockaddr_in6 *)(addr + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - sin6->sin6_scope_id = 0; - memcpy(&sin6->sin6_addr.s6_addr, &sel->saddr, - sizeof(sin6->sin6_addr.s6_addr)); - } -#endif + addr->sadb_address_prefixlen = sel->prefixlen_s; + pfkey_sockaddr_fill(&sel->saddr, 0, + (struct sockaddr *)(addr + 1), + sel->family); break; case SADB_EXT_ADDRESS_DST: - if (sel->family == AF_INET) { - addr->sadb_address_prefixlen = sel->prefixlen_d; - sin = (struct sockaddr_in *)(addr + 1); - sin->sin_family = AF_INET; - memcpy(&sin->sin_addr.s_addr, &sel->daddr, - sizeof(sin->sin_addr.s_addr)); - sin->sin_port = 0; - memset(sin->sin_zero, 0, sizeof(sin->sin_zero)); - } -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - else if (sel->family == AF_INET6) { - addr->sadb_address_prefixlen = sel->prefixlen_d; - sin6 = (struct sockaddr_in6 *)(addr + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - sin6->sin6_scope_id = 0; - memcpy(&sin6->sin6_addr.s6_addr, &sel->daddr, - sizeof(sin6->sin6_addr.s6_addr)); - } -#endif + addr->sadb_address_prefixlen = sel->prefixlen_d; + pfkey_sockaddr_fill(&sel->daddr, 0, + (struct sockaddr *)(addr + 1), + sel->family); break; default: return -EINVAL; @@ -3530,10 +3324,8 @@ static int set_ipsecrequest(struct sk_buff *skb, xfrm_address_t *src, xfrm_address_t *dst) { struct sadb_x_ipsecrequest *rq; - struct sockaddr_in *sin; -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - struct sockaddr_in6 *sin6; -#endif + u8 *sa; + int socklen = pfkey_sockaddr_len(family); int size_req; size_req = sizeof(struct sadb_x_ipsecrequest) + @@ -3547,38 +3339,10 @@ static int set_ipsecrequest(struct sk_buff *skb, rq->sadb_x_ipsecrequest_level = level; rq->sadb_x_ipsecrequest_reqid = reqid; - switch (family) { - case AF_INET: - sin = (struct sockaddr_in *)(rq + 1); - sin->sin_family = AF_INET; - memcpy(&sin->sin_addr.s_addr, src, - sizeof(sin->sin_addr.s_addr)); - sin++; - sin->sin_family = AF_INET; - memcpy(&sin->sin_addr.s_addr, dst, - sizeof(sin->sin_addr.s_addr)); - break; -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - case AF_INET6: - sin6 = (struct sockaddr_in6 *)(rq + 1); - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - sin6->sin6_scope_id = 0; - memcpy(&sin6->sin6_addr.s6_addr, src, - sizeof(sin6->sin6_addr.s6_addr)); - sin6++; - sin6->sin6_family = AF_INET6; - sin6->sin6_port = 0; - sin6->sin6_flowinfo = 0; - sin6->sin6_scope_id = 0; - memcpy(&sin6->sin6_addr.s6_addr, dst, - sizeof(sin6->sin6_addr.s6_addr)); - break; -#endif - default: + sa = (u8 *) (rq + 1); + if (!pfkey_sockaddr_fill(src, 0, (struct sockaddr *)sa, family) || + !pfkey_sockaddr_fill(dst, 0, (struct sockaddr *)(sa + socklen), family)) return -EINVAL; - } return 0; } -- cgit v1.2.2 From 5f95ac9111f75aa240dc3bcabffc0f047f13cb64 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Mon, 28 Apr 2008 02:46:24 +0900 Subject: key: Share common code path to extract address from sockaddr{}. Signed-off-by: YOSHIFUJI Hideaki --- net/key/af_key.c | 107 ++++++++++++++++++++----------------------------------- 1 file changed, 38 insertions(+), 69 deletions(-) (limited to 'net') diff --git a/net/key/af_key.c b/net/key/af_key.c index 2b05cb6c15c6..771bd61d6303 100644 --- a/net/key/af_key.c +++ b/net/key/af_key.c @@ -592,25 +592,30 @@ static inline int pfkey_sockaddr_len(sa_family_t family) return 0; } -static int pfkey_sadb_addr2xfrm_addr(struct sadb_address *addr, - xfrm_address_t *xaddr) +static +int pfkey_sockaddr_extract(const struct sockaddr *sa, xfrm_address_t *xaddr) { - switch (((struct sockaddr*)(addr + 1))->sa_family) { + switch (sa->sa_family) { case AF_INET: xaddr->a4 = - ((struct sockaddr_in *)(addr + 1))->sin_addr.s_addr; + ((struct sockaddr_in *)sa)->sin_addr.s_addr; return AF_INET; #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) case AF_INET6: memcpy(xaddr->a6, - &((struct sockaddr_in6 *)(addr + 1))->sin6_addr, + &((struct sockaddr_in6 *)sa)->sin6_addr, sizeof(struct in6_addr)); return AF_INET6; #endif - default: - return 0; } - /* NOTREACHED */ + return 0; +} + +static +int pfkey_sadb_addr2xfrm_addr(struct sadb_address *addr, xfrm_address_t *xaddr) +{ + return pfkey_sockaddr_extract((struct sockaddr *)(addr + 1), + xaddr); } static struct xfrm_state *pfkey_xfrm_state_lookup(struct sadb_msg *hdr, void **ext_hdrs) @@ -1828,10 +1833,6 @@ static int parse_ipsecrequest(struct xfrm_policy *xp, struct sadb_x_ipsecrequest *rq) { struct xfrm_tmpl *t = xp->xfrm_vec + xp->xfrm_nr; - struct sockaddr_in *sin; -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - struct sockaddr_in6 *sin6; -#endif int mode; if (xp->xfrm_nr >= XFRM_MAX_DEPTH) @@ -1856,31 +1857,19 @@ parse_ipsecrequest(struct xfrm_policy *xp, struct sadb_x_ipsecrequest *rq) /* addresses present only in tunnel mode */ if (t->mode == XFRM_MODE_TUNNEL) { - struct sockaddr *sa; - sa = (struct sockaddr *)(rq+1); - switch(sa->sa_family) { - case AF_INET: - sin = (struct sockaddr_in*)sa; - t->saddr.a4 = sin->sin_addr.s_addr; - sin++; - if (sin->sin_family != AF_INET) - return -EINVAL; - t->id.daddr.a4 = sin->sin_addr.s_addr; - break; -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - case AF_INET6: - sin6 = (struct sockaddr_in6*)sa; - memcpy(t->saddr.a6, &sin6->sin6_addr, sizeof(struct in6_addr)); - sin6++; - if (sin6->sin6_family != AF_INET6) - return -EINVAL; - memcpy(t->id.daddr.a6, &sin6->sin6_addr, sizeof(struct in6_addr)); - break; -#endif - default: + u8 *sa = (u8 *) (rq + 1); + int family, socklen; + + family = pfkey_sockaddr_extract((struct sockaddr *)sa, + &t->saddr); + if (!family) return -EINVAL; - } - t->encap_family = sa->sa_family; + + socklen = pfkey_sockaddr_len(family); + if (pfkey_sockaddr_extract((struct sockaddr *)(sa + socklen), + &t->id.daddr) != family) + return -EINVAL; + t->encap_family = family; } else t->encap_family = xp->family; @@ -2375,44 +2364,24 @@ static int parse_sockaddr_pair(struct sadb_x_ipsecrequest *rq, xfrm_address_t *saddr, xfrm_address_t *daddr, u16 *family) { - struct sockaddr *sa = (struct sockaddr *)(rq + 1); + u8 *sa = (u8 *) (rq + 1); + int af, socklen; + if (rq->sadb_x_ipsecrequest_len < - pfkey_sockaddr_pair_size(sa->sa_family)) + pfkey_sockaddr_pair_size(((struct sockaddr *)sa)->sa_family)) return -EINVAL; - switch (sa->sa_family) { - case AF_INET: - { - struct sockaddr_in *sin; - sin = (struct sockaddr_in *)sa; - if ((sin+1)->sin_family != AF_INET) - return -EINVAL; - memcpy(&saddr->a4, &sin->sin_addr, sizeof(saddr->a4)); - sin++; - memcpy(&daddr->a4, &sin->sin_addr, sizeof(daddr->a4)); - *family = AF_INET; - break; - } -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - case AF_INET6: - { - struct sockaddr_in6 *sin6; - sin6 = (struct sockaddr_in6 *)sa; - if ((sin6+1)->sin6_family != AF_INET6) - return -EINVAL; - memcpy(&saddr->a6, &sin6->sin6_addr, - sizeof(saddr->a6)); - sin6++; - memcpy(&daddr->a6, &sin6->sin6_addr, - sizeof(daddr->a6)); - *family = AF_INET6; - break; - } -#endif - default: + af = pfkey_sockaddr_extract((struct sockaddr *) sa, + saddr); + if (!af) + return -EINVAL; + + socklen = pfkey_sockaddr_len(af); + if (pfkey_sockaddr_extract((struct sockaddr *) (sa + socklen), + daddr) != af) return -EINVAL; - } + *family = af; return 0; } -- cgit v1.2.2 From 81b302a321a0d99ff172b8cb2a8de17bff2f9499 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Mon, 28 Apr 2008 03:17:38 +0900 Subject: key: Use xfrm_addr_cmp() where appropriate. Signed-off-by: YOSHIFUJI Hideaki --- net/key/af_key.c | 9 +-------- 1 file changed, 1 insertion(+), 8 deletions(-) (limited to 'net') diff --git a/net/key/af_key.c b/net/key/af_key.c index 771bd61d6303..841af9f2d5e0 100644 --- a/net/key/af_key.c +++ b/net/key/af_key.c @@ -767,14 +767,7 @@ static struct sk_buff *__pfkey_xfrm_state2msg(struct xfrm_state *x, } /* identity & sensitivity */ - - if ((x->props.family == AF_INET && - x->sel.saddr.a4 != x->props.saddr.a4) -#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) - || (x->props.family == AF_INET6 && - memcmp (x->sel.saddr.a6, x->props.saddr.a6, sizeof (struct in6_addr))) -#endif - ) + if (xfrm_addr_cmp(&x->sel.saddr, &x->props.saddr, x->props.family)) size += sizeof(struct sadb_address) + sockaddr_size; if (add_keys) { -- cgit v1.2.2 From 7d5d5525bd88313e6fd90c0659665aee5114bc2d Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Thu, 17 Apr 2008 12:29:53 +0900 Subject: tcp md5sig: Share MD5 Signature option parser between IPv4 and IPv6. Signed-off-by: YOSHIFUJI Hideaki --- net/ipv4/tcp_input.c | 40 ++++++++++++++++++++++++++++++++++++++++ net/ipv4/tcp_ipv4.c | 42 +----------------------------------------- net/ipv6/tcp_ipv6.c | 33 +-------------------------------- 3 files changed, 42 insertions(+), 73 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index eba873e9b560..f331e67f2328 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -3450,6 +3450,43 @@ static int tcp_fast_parse_options(struct sk_buff *skb, struct tcphdr *th, return 1; } +#ifdef CONFIG_TCP_MD5SIG +/* + * Parse MD5 Signature option + */ +u8 *tcp_parse_md5sig_option(struct tcphdr *th) +{ + int length = (th->doff << 2) - sizeof (*th); + u8 *ptr = (u8*)(th + 1); + + /* If the TCP option is too short, we can short cut */ + if (length < TCPOLEN_MD5SIG) + return NULL; + + while (length > 0) { + int opcode = *ptr++; + int opsize; + + switch(opcode) { + case TCPOPT_EOL: + return NULL; + case TCPOPT_NOP: + length--; + continue; + default: + opsize = *ptr++; + if (opsize < 2 || opsize > length) + return NULL; + if (opcode == TCPOPT_MD5SIG) + return ptr; + } + ptr += opsize - 2; + length -= opsize; + } + return NULL; +} +#endif + static inline void tcp_store_ts_recent(struct tcp_sock *tp) { tp->rx_opt.ts_recent = tp->rx_opt.rcv_tsval; @@ -5467,6 +5504,9 @@ EXPORT_SYMBOL(sysctl_tcp_ecn); EXPORT_SYMBOL(sysctl_tcp_reordering); EXPORT_SYMBOL(sysctl_tcp_adv_win_scale); EXPORT_SYMBOL(tcp_parse_options); +#ifdef CONFIG_TCP_MD5SIG +EXPORT_SYMBOL(tcp_parse_md5sig_option); +#endif EXPORT_SYMBOL(tcp_rcv_established); EXPORT_SYMBOL(tcp_rcv_state_process); EXPORT_SYMBOL(tcp_initialize_rcv_mss); diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index cd601a866c2f..56f550933644 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -1134,52 +1134,12 @@ static int tcp_v4_inbound_md5_hash(struct sock *sk, struct sk_buff *skb) struct tcp_md5sig_key *hash_expected; const struct iphdr *iph = ip_hdr(skb); struct tcphdr *th = tcp_hdr(skb); - int length = (th->doff << 2) - sizeof(struct tcphdr); int genhash; - unsigned char *ptr; unsigned char newhash[16]; hash_expected = tcp_v4_md5_do_lookup(sk, iph->saddr); + hash_location = tcp_parse_md5sig_option(th); - /* - * If the TCP option length is less than the TCP_MD5SIG - * option length, then we can shortcut - */ - if (length < TCPOLEN_MD5SIG) { - if (hash_expected) - return 1; - else - return 0; - } - - /* Okay, we can't shortcut - we have to grub through the options */ - ptr = (unsigned char *)(th + 1); - while (length > 0) { - int opcode = *ptr++; - int opsize; - - switch (opcode) { - case TCPOPT_EOL: - goto done_opts; - case TCPOPT_NOP: - length--; - continue; - default: - opsize = *ptr++; - if (opsize < 2) - goto done_opts; - if (opsize > length) - goto done_opts; - - if (opcode == TCPOPT_MD5SIG) { - hash_location = ptr; - goto done_opts; - } - } - ptr += opsize-2; - length -= opsize; - } -done_opts: /* We've parsed the options - do we have a hash? */ if (!hash_expected && !hash_location) return 0; diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 715965f0fac0..dd4ddb30a3a9 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -844,43 +844,12 @@ static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb) struct tcp_md5sig_key *hash_expected; struct ipv6hdr *ip6h = ipv6_hdr(skb); struct tcphdr *th = tcp_hdr(skb); - int length = (th->doff << 2) - sizeof (*th); int genhash; - u8 *ptr; u8 newhash[16]; hash_expected = tcp_v6_md5_do_lookup(sk, &ip6h->saddr); + hash_location = tcp_parse_md5sig_option(th); - /* If the TCP option is too short, we can short cut */ - if (length < TCPOLEN_MD5SIG) - return hash_expected ? 1 : 0; - - /* parse options */ - ptr = (u8*)(th + 1); - while (length > 0) { - int opcode = *ptr++; - int opsize; - - switch(opcode) { - case TCPOPT_EOL: - goto done_opts; - case TCPOPT_NOP: - length--; - continue; - default: - opsize = *ptr++; - if (opsize < 2 || opsize > length) - goto done_opts; - if (opcode == TCPOPT_MD5SIG) { - hash_location = ptr; - goto done_opts; - } - } - ptr += opsize - 2; - length -= opsize; - } - -done_opts: /* do we have a hash as expected? */ if (!hash_expected) { if (!hash_location) -- cgit v1.2.2 From 076fb7223357769c39f3ddf900bba6752369c76a Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Thu, 17 Apr 2008 12:48:12 +0900 Subject: tcp md5sig: Remove redundant protocol argument. Protocol is always TCP, so remove useless protocol argument. Signed-off-by: YOSHIFUJI Hideaki --- net/ipv4/tcp_ipv4.c | 20 ++++++++------------ net/ipv4/tcp_output.c | 3 +-- net/ipv6/tcp_ipv6.c | 17 +++++++---------- 3 files changed, 16 insertions(+), 24 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 56f550933644..f25445d21bdf 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -95,8 +95,7 @@ static struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr); static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, __be32 saddr, __be32 daddr, - struct tcphdr *th, int protocol, - unsigned int tcplen); + struct tcphdr *th, unsigned int tcplen); #endif struct inet_hashinfo __cacheline_aligned tcp_hashinfo = { @@ -586,8 +585,7 @@ static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb) key, ip_hdr(skb)->daddr, ip_hdr(skb)->saddr, - &rep.th, IPPROTO_TCP, - arg.iov[0].iov_len); + &rep.th, arg.iov[0].iov_len); } #endif arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr, @@ -680,8 +678,7 @@ static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk, key, ip_hdr(skb)->daddr, ip_hdr(skb)->saddr, - &rep.th, IPPROTO_TCP, - arg.iov[0].iov_len); + &rep.th, arg.iov[0].iov_len); } #endif arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr, @@ -1006,7 +1003,7 @@ static int tcp_v4_parse_md5_keys(struct sock *sk, char __user *optval, static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, __be32 saddr, __be32 daddr, - struct tcphdr *th, int protocol, + struct tcphdr *th, unsigned int tcplen) { struct scatterlist sg[4]; @@ -1039,7 +1036,7 @@ static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, bp->saddr = saddr; bp->daddr = daddr; bp->pad = 0; - bp->protocol = protocol; + bp->protocol = IPPROTO_TCP; bp->len = htons(tcplen); sg_init_table(sg, 4); @@ -1099,7 +1096,7 @@ int tcp_v4_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, struct sock *sk, struct dst_entry *dst, struct request_sock *req, - struct tcphdr *th, int protocol, + struct tcphdr *th, unsigned int tcplen) { __be32 saddr, daddr; @@ -1115,7 +1112,7 @@ int tcp_v4_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, } return tcp_v4_do_calc_md5_hash(md5_hash, key, saddr, daddr, - th, protocol, tcplen); + th, tcplen); } EXPORT_SYMBOL(tcp_v4_calc_md5_hash); @@ -1166,8 +1163,7 @@ static int tcp_v4_inbound_md5_hash(struct sock *sk, struct sk_buff *skb) genhash = tcp_v4_do_calc_md5_hash(newhash, hash_expected, iph->saddr, iph->daddr, - th, sk->sk_protocol, - skb->len); + th, skb->len); if (genhash || memcmp(hash_location, newhash, 16) != 0) { if (net_ratelimit()) { diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index ad993ecb4810..f3ffd674edde 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -607,7 +607,6 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, md5, sk, NULL, NULL, tcp_hdr(skb), - sk->sk_protocol, skb->len); } #endif @@ -2266,7 +2265,7 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst, tp->af_specific->calc_md5_hash(md5_hash_location, md5, NULL, dst, req, - tcp_hdr(skb), sk->sk_protocol, + tcp_hdr(skb), skb->len); } #endif diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index dd4ddb30a3a9..334d21c23da9 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -736,8 +736,7 @@ static int tcp_v6_parse_md5_keys (struct sock *sk, char __user *optval, static int tcp_v6_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, struct in6_addr *saddr, struct in6_addr *daddr, - struct tcphdr *th, int protocol, - unsigned int tcplen) + struct tcphdr *th, unsigned int tcplen) { struct scatterlist sg[4]; __u16 data_len; @@ -761,7 +760,7 @@ static int tcp_v6_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, ipv6_addr_copy(&bp->saddr, saddr); ipv6_addr_copy(&bp->daddr, daddr); bp->len = htonl(tcplen); - bp->protocol = htonl(protocol); + bp->protocol = htonl(IPPROTO_TCP); sg_init_table(sg, 4); @@ -821,8 +820,7 @@ static int tcp_v6_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, struct sock *sk, struct dst_entry *dst, struct request_sock *req, - struct tcphdr *th, int protocol, - unsigned int tcplen) + struct tcphdr *th, unsigned int tcplen) { struct in6_addr *saddr, *daddr; @@ -835,7 +833,7 @@ static int tcp_v6_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, } return tcp_v6_do_calc_md5_hash(md5_hash, key, saddr, daddr, - th, protocol, tcplen); + th, tcplen); } static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb) @@ -879,8 +877,7 @@ static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb) genhash = tcp_v6_do_calc_md5_hash(newhash, hash_expected, &ip6h->saddr, &ip6h->daddr, - th, sk->sk_protocol, - skb->len); + th, skb->len); if (genhash || memcmp(hash_location, newhash, 16) != 0) { if (net_ratelimit()) { printk(KERN_INFO "MD5 Hash %s for " @@ -1020,7 +1017,7 @@ static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb) tcp_v6_do_calc_md5_hash((__u8 *)&opt[1], key, &ipv6_hdr(skb)->daddr, &ipv6_hdr(skb)->saddr, - t1, IPPROTO_TCP, tot_len); + t1, tot_len); } #endif @@ -1126,7 +1123,7 @@ static void tcp_v6_send_ack(struct tcp_timewait_sock *tw, tcp_v6_do_calc_md5_hash((__u8 *)topt, key, &ipv6_hdr(skb)->daddr, &ipv6_hdr(skb)->saddr, - t1, IPPROTO_TCP, tot_len); + t1, tot_len); } #endif -- cgit v1.2.2 From 8d26d76dd4a4c87ef037a44a42a0608ffc730199 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Thu, 17 Apr 2008 13:19:16 +0900 Subject: tcp md5sig: Share most of hash calcucaltion bits between IPv4 and IPv6. We can share most part of the hash calculation code because the only difference between IPv4 and IPv6 is their pseudo headers. Signed-off-by: YOSHIFUJI Hideaki --- net/ipv4/tcp.c | 70 +++++++++++++++++++++++++++++++++++++++++++++++++++++ net/ipv4/tcp_ipv4.c | 52 +++------------------------------------ net/ipv6/tcp_ipv6.c | 53 ++++------------------------------------ 3 files changed, 79 insertions(+), 96 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index ab66683b8043..6efbae0e5512 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -2459,6 +2459,76 @@ static unsigned long tcp_md5sig_users; static struct tcp_md5sig_pool **tcp_md5sig_pool; static DEFINE_SPINLOCK(tcp_md5sig_pool_lock); +int tcp_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, + int bplen, + struct tcphdr *th, unsigned int tcplen, + struct tcp_md5sig_pool *hp) +{ + struct scatterlist sg[4]; + __u16 data_len; + int block = 0; + __sum16 cksum; + struct hash_desc *desc = &hp->md5_desc; + int err; + unsigned int nbytes = 0; + + sg_init_table(sg, 4); + + /* 1. The TCP pseudo-header */ + sg_set_buf(&sg[block++], &hp->md5_blk, bplen); + nbytes += bplen; + + /* 2. The TCP header, excluding options, and assuming a + * checksum of zero + */ + cksum = th->check; + th->check = 0; + sg_set_buf(&sg[block++], th, sizeof(*th)); + nbytes += sizeof(*th); + + /* 3. The TCP segment data (if any) */ + data_len = tcplen - (th->doff << 2); + if (data_len > 0) { + u8 *data = (u8 *)th + (th->doff << 2); + sg_set_buf(&sg[block++], data, data_len); + nbytes += data_len; + } + + /* 4. an independently-specified key or password, known to both + * TCPs and presumably connection-specific + */ + sg_set_buf(&sg[block++], key->key, key->keylen); + nbytes += key->keylen; + + sg_mark_end(&sg[block - 1]); + + /* Now store the hash into the packet */ + err = crypto_hash_init(desc); + if (err) { + if (net_ratelimit()) + printk(KERN_WARNING "%s(): hash_init failed\n", __func__); + return -1; + } + err = crypto_hash_update(desc, sg, nbytes); + if (err) { + if (net_ratelimit()) + printk(KERN_WARNING "%s(): hash_update failed\n", __func__); + return -1; + } + err = crypto_hash_final(desc, md5_hash); + if (err) { + if (net_ratelimit()) + printk(KERN_WARNING "%s(): hash_final failed\n", __func__); + return -1; + } + + /* Reset header */ + th->check = cksum; + + return 0; +} +EXPORT_SYMBOL(tcp_calc_md5_hash); + static void __tcp_free_md5sig_pool(struct tcp_md5sig_pool **pool) { int cpu; diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index f25445d21bdf..e331cdbd0953 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -1006,15 +1006,9 @@ static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, struct tcphdr *th, unsigned int tcplen) { - struct scatterlist sg[4]; - __u16 data_len; - int block = 0; - __sum16 old_checksum; struct tcp_md5sig_pool *hp; struct tcp4_pseudohdr *bp; - struct hash_desc *desc; int err; - unsigned int nbytes = 0; /* * Okay, so RFC2385 is turned on for this connection, @@ -1026,10 +1020,9 @@ static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, goto clear_hash_noput; bp = &hp->md5_blk.ip4; - desc = &hp->md5_desc; /* - * 1. the TCP pseudo-header (in the order: source IP address, + * The TCP pseudo-header (in the order: source IP address, * destination IP address, zero-padded protocol number, and * segment length) */ @@ -1039,50 +1032,13 @@ static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, bp->protocol = IPPROTO_TCP; bp->len = htons(tcplen); - sg_init_table(sg, 4); - - sg_set_buf(&sg[block++], bp, sizeof(*bp)); - nbytes += sizeof(*bp); - - /* 2. the TCP header, excluding options, and assuming a - * checksum of zero/ - */ - old_checksum = th->check; - th->check = 0; - sg_set_buf(&sg[block++], th, sizeof(struct tcphdr)); - nbytes += sizeof(struct tcphdr); - - /* 3. the TCP segment data (if any) */ - data_len = tcplen - (th->doff << 2); - if (data_len > 0) { - unsigned char *data = (unsigned char *)th + (th->doff << 2); - sg_set_buf(&sg[block++], data, data_len); - nbytes += data_len; - } - - /* 4. an independently-specified key or password, known to both - * TCPs and presumably connection-specific - */ - sg_set_buf(&sg[block++], key->key, key->keylen); - nbytes += key->keylen; - - sg_mark_end(&sg[block - 1]); - - /* Now store the Hash into the packet */ - err = crypto_hash_init(desc); - if (err) - goto clear_hash; - err = crypto_hash_update(desc, sg, nbytes); - if (err) - goto clear_hash; - err = crypto_hash_final(desc, md5_hash); + err = tcp_calc_md5_hash(md5_hash, key, sizeof(*bp), + th, tcplen, hp); if (err) goto clear_hash; - /* Reset header, and free up the crypto */ + /* Free up the crypto pool */ tcp_put_md5sig_pool(); - th->check = old_checksum; - out: return 0; clear_hash: diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 334d21c23da9..0ae0311082fb 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -738,23 +738,17 @@ static int tcp_v6_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, struct in6_addr *daddr, struct tcphdr *th, unsigned int tcplen) { - struct scatterlist sg[4]; - __u16 data_len; - int block = 0; - __sum16 cksum; struct tcp_md5sig_pool *hp; struct tcp6_pseudohdr *bp; - struct hash_desc *desc; int err; - unsigned int nbytes = 0; hp = tcp_get_md5sig_pool(); if (!hp) { printk(KERN_WARNING "%s(): hash pool not found...\n", __func__); goto clear_hash_noput; } + bp = &hp->md5_blk.ip6; - desc = &hp->md5_desc; /* 1. TCP pseudo-header (RFC2460) */ ipv6_addr_copy(&bp->saddr, saddr); @@ -762,51 +756,14 @@ static int tcp_v6_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, bp->len = htonl(tcplen); bp->protocol = htonl(IPPROTO_TCP); - sg_init_table(sg, 4); - - sg_set_buf(&sg[block++], bp, sizeof(*bp)); - nbytes += sizeof(*bp); - - /* 2. TCP header, excluding options */ - cksum = th->check; - th->check = 0; - sg_set_buf(&sg[block++], th, sizeof(*th)); - nbytes += sizeof(*th); - - /* 3. TCP segment data (if any) */ - data_len = tcplen - (th->doff << 2); - if (data_len > 0) { - u8 *data = (u8 *)th + (th->doff << 2); - sg_set_buf(&sg[block++], data, data_len); - nbytes += data_len; - } - - /* 4. shared key */ - sg_set_buf(&sg[block++], key->key, key->keylen); - nbytes += key->keylen; - - sg_mark_end(&sg[block - 1]); + err = tcp_calc_md5_hash(md5_hash, key, sizeof(*bp), + th, tcplen, hp); - /* Now store the hash into the packet */ - err = crypto_hash_init(desc); - if (err) { - printk(KERN_WARNING "%s(): hash_init failed\n", __func__); - goto clear_hash; - } - err = crypto_hash_update(desc, sg, nbytes); - if (err) { - printk(KERN_WARNING "%s(): hash_update failed\n", __func__); - goto clear_hash; - } - err = crypto_hash_final(desc, md5_hash); - if (err) { - printk(KERN_WARNING "%s(): hash_final failed\n", __func__); + if (err) goto clear_hash; - } - /* Reset header, and free up the crypto */ + /* Free up the crypto pool */ tcp_put_md5sig_pool(); - th->check = cksum; out: return 0; clear_hash: -- cgit v1.2.2 From 9501f9722922f2e80e1f9dc6682311d65c2b5690 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Fri, 18 Apr 2008 12:45:16 +0900 Subject: tcp md5sig: Let the caller pass appropriate key for tcp_v{4,6}_do_calc_md5_hash(). As we do for other socket/timewait-socket specific parameters, let the callers pass appropriate arguments to tcp_v{4,6}_do_calc_md5_hash(). Signed-off-by: YOSHIFUJI Hideaki --- net/ipv4/tcp_ipv4.c | 50 ++++++++++++++++++++------------------------------ net/ipv6/tcp_ipv6.c | 33 ++++++++++++--------------------- 2 files changed, 32 insertions(+), 51 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index e331cdbd0953..f7ff2a64a7f0 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -96,6 +96,12 @@ static struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk, static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, __be32 saddr, __be32 daddr, struct tcphdr *th, unsigned int tcplen); +#else +static inline +struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr) +{ + return NULL; +} #endif struct inet_hashinfo __cacheline_aligned tcp_hashinfo = { @@ -604,9 +610,9 @@ static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb) outside socket context is ugly, certainly. What can I do? */ -static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk, - struct sk_buff *skb, u32 seq, u32 ack, - u32 win, u32 ts) +static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack, + u32 win, u32 ts, int oif, + struct tcp_md5sig_key *key) { struct tcphdr *th = tcp_hdr(skb); struct { @@ -618,10 +624,6 @@ static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk, ]; } rep; struct ip_reply_arg arg; -#ifdef CONFIG_TCP_MD5SIG - struct tcp_md5sig_key *key; - struct tcp_md5sig_key tw_key; -#endif memset(&rep.th, 0, sizeof(struct tcphdr)); memset(&arg, 0, sizeof(arg)); @@ -647,23 +649,6 @@ static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk, rep.th.window = htons(win); #ifdef CONFIG_TCP_MD5SIG - /* - * The SKB holds an imcoming packet, but may not have a valid ->sk - * pointer. This is especially the case when we're dealing with a - * TIME_WAIT ack, because the sk structure is long gone, and only - * the tcp_timewait_sock remains. So the md5 key is stashed in that - * structure, and we use it in preference. I believe that (twsk || - * skb->sk) holds true, but we program defensively. - */ - if (!twsk && skb->sk) { - key = tcp_v4_md5_do_lookup(skb->sk, ip_hdr(skb)->daddr); - } else if (twsk && twsk->tw_md5_keylen) { - tw_key.key = twsk->tw_md5_key; - tw_key.keylen = twsk->tw_md5_keylen; - key = &tw_key; - } else - key = NULL; - if (key) { int offset = (ts) ? 3 : 0; @@ -685,8 +670,8 @@ static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk, ip_hdr(skb)->saddr, /* XXX */ arg.iov[0].iov_len, IPPROTO_TCP, 0); arg.csumoffset = offsetof(struct tcphdr, check) / 2; - if (twsk) - arg.bound_dev_if = twsk->tw_sk.tw_bound_dev_if; + if (oif) + arg.bound_dev_if = oif; ip_send_reply(dev_net(skb->dev)->ipv4.tcp_sock, skb, &arg, arg.iov[0].iov_len); @@ -699,9 +684,12 @@ static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb) struct inet_timewait_sock *tw = inet_twsk(sk); struct tcp_timewait_sock *tcptw = tcp_twsk(sk); - tcp_v4_send_ack(tcptw, skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt, + tcp_v4_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt, tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale, - tcptw->tw_ts_recent); + tcptw->tw_ts_recent, + tw->tw_bound_dev_if, + tcp_twsk_md5_key(tcptw) + ); inet_twsk_put(tw); } @@ -709,9 +697,11 @@ static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb) static void tcp_v4_reqsk_send_ack(struct sk_buff *skb, struct request_sock *req) { - tcp_v4_send_ack(NULL, skb, tcp_rsk(req)->snt_isn + 1, + tcp_v4_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, - req->ts_recent); + req->ts_recent, + 0, + tcp_v4_md5_do_lookup(skb->sk, ip_hdr(skb)->daddr)); } /* diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 0ae0311082fb..ecdbb9f46541 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -82,6 +82,12 @@ static struct inet_connection_sock_af_ops ipv6_specific; #ifdef CONFIG_TCP_MD5SIG static struct tcp_sock_af_ops tcp_sock_ipv6_specific; static struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific; +#else +static struct tcp_md5sig_key *tcp_v6_md5_do_lookup(struct sock *sk, + struct in6_addr *addr) +{ + return NULL; +} #endif static void tcp_v6_hash(struct sock *sk) @@ -1011,8 +1017,8 @@ static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb) kfree_skb(buff); } -static void tcp_v6_send_ack(struct tcp_timewait_sock *tw, - struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 ts) +static void tcp_v6_send_ack(struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 ts, + struct tcp_md5sig_key *key) { struct tcphdr *th = tcp_hdr(skb), *t1; struct sk_buff *buff; @@ -1021,22 +1027,6 @@ static void tcp_v6_send_ack(struct tcp_timewait_sock *tw, struct sock *ctl_sk = net->ipv6.tcp_sk; unsigned int tot_len = sizeof(struct tcphdr); __be32 *topt; -#ifdef CONFIG_TCP_MD5SIG - struct tcp_md5sig_key *key; - struct tcp_md5sig_key tw_key; -#endif - -#ifdef CONFIG_TCP_MD5SIG - if (!tw && skb->sk) { - key = tcp_v6_md5_do_lookup(skb->sk, &ipv6_hdr(skb)->daddr); - } else if (tw && tw->tw_md5_keylen) { - tw_key.key = tw->tw_md5_key; - tw_key.keylen = tw->tw_md5_keylen; - key = &tw_key; - } else { - key = NULL; - } -#endif if (ts) tot_len += TCPOLEN_TSTAMP_ALIGNED; @@ -1116,16 +1106,17 @@ static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb) struct inet_timewait_sock *tw = inet_twsk(sk); struct tcp_timewait_sock *tcptw = tcp_twsk(sk); - tcp_v6_send_ack(tcptw, skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt, + tcp_v6_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt, tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale, - tcptw->tw_ts_recent); + tcptw->tw_ts_recent, tcp_twsk_md5_key(tcptw)); inet_twsk_put(tw); } static void tcp_v6_reqsk_send_ack(struct sk_buff *skb, struct request_sock *req) { - tcp_v6_send_ack(NULL, skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, req->ts_recent); + tcp_v6_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, req->ts_recent, + tcp_v6_md5_do_lookup(skb->sk, &ipv6_hdr(skb)->daddr)); } -- cgit v1.2.2 From 0b040829952d84bf2a62526f0e24b624e0699447 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Tue, 10 Jun 2008 22:46:50 -0700 Subject: net: remove CVS keywords This patch removes CVS keywords that weren't updated for a long time from comments. Signed-off-by: Adrian Bunk Signed-off-by: David S. Miller --- net/bluetooth/bnep/bnep.h | 4 ---- net/bluetooth/bnep/core.c | 4 ---- net/bluetooth/bnep/netdev.c | 4 ---- net/bluetooth/bnep/sock.c | 4 ---- net/bluetooth/rfcomm/core.c | 2 -- net/bluetooth/rfcomm/sock.c | 2 -- net/bluetooth/rfcomm/tty.c | 2 -- net/bridge/br.c | 2 -- net/bridge/br_device.c | 2 -- net/bridge/br_fdb.c | 2 -- net/bridge/br_forward.c | 2 -- net/bridge/br_if.c | 2 -- net/bridge/br_input.c | 2 -- net/bridge/br_ioctl.c | 2 -- net/bridge/br_notify.c | 2 -- net/bridge/br_private.h | 2 -- net/bridge/br_private_stp.h | 2 -- net/bridge/br_stp.c | 2 -- net/bridge/br_stp_bpdu.c | 2 -- net/bridge/br_stp_if.c | 2 -- net/bridge/br_stp_timer.c | 2 -- net/core/skbuff.c | 2 -- net/core/sock.c | 2 -- net/ipv4/af_inet.c | 2 -- net/ipv4/arp.c | 2 -- net/ipv4/devinet.c | 2 -- net/ipv4/fib_frontend.c | 2 -- net/ipv4/fib_hash.c | 2 -- net/ipv4/fib_semantics.c | 2 -- net/ipv4/fib_trie.c | 2 -- net/ipv4/icmp.c | 2 -- net/ipv4/igmp.c | 2 -- net/ipv4/inet_diag.c | 2 -- net/ipv4/inetpeer.c | 2 -- net/ipv4/ip_forward.c | 2 -- net/ipv4/ip_fragment.c | 2 -- net/ipv4/ip_input.c | 2 -- net/ipv4/ip_options.c | 2 -- net/ipv4/ip_output.c | 2 -- net/ipv4/ip_sockglue.c | 2 -- net/ipv4/ipconfig.c | 2 -- net/ipv4/ipip.c | 2 -- net/ipv4/ipmr.c | 2 -- net/ipv4/ipvs/ip_vs_app.c | 2 -- net/ipv4/ipvs/ip_vs_conn.c | 2 -- net/ipv4/ipvs/ip_vs_core.c | 2 -- net/ipv4/ipvs/ip_vs_ctl.c | 2 -- net/ipv4/ipvs/ip_vs_dh.c | 2 -- net/ipv4/ipvs/ip_vs_est.c | 2 -- net/ipv4/ipvs/ip_vs_ftp.c | 2 -- net/ipv4/ipvs/ip_vs_lblc.c | 2 -- net/ipv4/ipvs/ip_vs_lblcr.c | 2 -- net/ipv4/ipvs/ip_vs_lc.c | 2 -- net/ipv4/ipvs/ip_vs_nq.c | 2 -- net/ipv4/ipvs/ip_vs_proto.c | 2 -- net/ipv4/ipvs/ip_vs_proto_ah.c | 2 -- net/ipv4/ipvs/ip_vs_proto_esp.c | 2 -- net/ipv4/ipvs/ip_vs_proto_tcp.c | 2 -- net/ipv4/ipvs/ip_vs_proto_udp.c | 2 -- net/ipv4/ipvs/ip_vs_rr.c | 2 -- net/ipv4/ipvs/ip_vs_sched.c | 2 -- net/ipv4/ipvs/ip_vs_sed.c | 2 -- net/ipv4/ipvs/ip_vs_sh.c | 2 -- net/ipv4/ipvs/ip_vs_sync.c | 2 -- net/ipv4/ipvs/ip_vs_wlc.c | 2 -- net/ipv4/ipvs/ip_vs_wrr.c | 2 -- net/ipv4/ipvs/ip_vs_xmit.c | 2 -- net/ipv4/proc.c | 2 -- net/ipv4/protocol.c | 2 -- net/ipv4/raw.c | 2 -- net/ipv4/route.c | 2 -- net/ipv4/syncookies.c | 2 -- net/ipv4/sysctl_net_ipv4.c | 2 -- net/ipv4/tcp.c | 2 -- net/ipv4/tcp_diag.c | 2 -- net/ipv4/tcp_input.c | 2 -- net/ipv4/tcp_ipv4.c | 2 -- net/ipv4/tcp_minisocks.c | 2 -- net/ipv4/tcp_output.c | 2 -- net/ipv4/tcp_timer.c | 2 -- net/ipv4/udp.c | 2 -- net/ipv4/udplite.c | 2 -- net/ipv6/addrconf.c | 2 -- net/ipv6/af_inet6.c | 2 -- net/ipv6/datagram.c | 2 -- net/ipv6/exthdrs.c | 2 -- net/ipv6/icmp.c | 2 -- net/ipv6/ip6_fib.c | 2 -- net/ipv6/ip6_input.c | 2 -- net/ipv6/ip6_output.c | 2 -- net/ipv6/ip6_tunnel.c | 2 -- net/ipv6/ipv6_sockglue.c | 2 -- net/ipv6/mcast.c | 2 -- net/ipv6/proc.c | 2 -- net/ipv6/protocol.c | 2 -- net/ipv6/raw.c | 2 -- net/ipv6/reassembly.c | 2 -- net/ipv6/route.c | 2 -- net/ipv6/sit.c | 2 -- net/ipv6/tcp_ipv6.c | 2 -- net/ipv6/udp.c | 2 -- net/ipv6/udplite.c | 2 -- net/packet/af_packet.c | 2 -- net/sched/sch_htb.c | 2 -- net/sunrpc/auth_gss/auth_gss.c | 2 -- net/sysctl_net.c | 1 - net/unix/af_unix.c | 2 -- 107 files changed, 221 deletions(-) (limited to 'net') diff --git a/net/bluetooth/bnep/bnep.h b/net/bluetooth/bnep/bnep.h index e69244dd8de8..b69bf4e7c48b 100644 --- a/net/bluetooth/bnep/bnep.h +++ b/net/bluetooth/bnep/bnep.h @@ -16,10 +16,6 @@ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ -/* - * $Id: bnep.h,v 1.5 2002/08/04 21:23:58 maxk Exp $ - */ - #ifndef _BNEP_H #define _BNEP_H diff --git a/net/bluetooth/bnep/core.c b/net/bluetooth/bnep/core.c index f85d94643aaf..1d98a1b80da7 100644 --- a/net/bluetooth/bnep/core.c +++ b/net/bluetooth/bnep/core.c @@ -25,10 +25,6 @@ SOFTWARE IS DISCLAIMED. */ -/* - * $Id: core.c,v 1.20 2002/08/04 21:23:58 maxk Exp $ - */ - #include #include diff --git a/net/bluetooth/bnep/netdev.c b/net/bluetooth/bnep/netdev.c index 95e3837e4312..d9fa0ab2c87f 100644 --- a/net/bluetooth/bnep/netdev.c +++ b/net/bluetooth/bnep/netdev.c @@ -25,10 +25,6 @@ SOFTWARE IS DISCLAIMED. */ -/* - * $Id: netdev.c,v 1.8 2002/08/04 21:23:58 maxk Exp $ - */ - #include #include diff --git a/net/bluetooth/bnep/sock.c b/net/bluetooth/bnep/sock.c index 201e5b1ce473..8ffb57f2303a 100644 --- a/net/bluetooth/bnep/sock.c +++ b/net/bluetooth/bnep/sock.c @@ -24,10 +24,6 @@ SOFTWARE IS DISCLAIMED. */ -/* - * $Id: sock.c,v 1.4 2002/08/04 21:23:58 maxk Exp $ - */ - #include #include diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c index 0c2c93735e93..b4fb84e398e5 100644 --- a/net/bluetooth/rfcomm/core.c +++ b/net/bluetooth/rfcomm/core.c @@ -23,8 +23,6 @@ /* * Bluetooth RFCOMM core. - * - * $Id: core.c,v 1.42 2002/10/01 23:26:25 maxk Exp $ */ #include diff --git a/net/bluetooth/rfcomm/sock.c b/net/bluetooth/rfcomm/sock.c index 5083adcbfae5..c9054487670a 100644 --- a/net/bluetooth/rfcomm/sock.c +++ b/net/bluetooth/rfcomm/sock.c @@ -23,8 +23,6 @@ /* * RFCOMM sockets. - * - * $Id: sock.c,v 1.24 2002/10/03 01:00:34 maxk Exp $ */ #include diff --git a/net/bluetooth/rfcomm/tty.c b/net/bluetooth/rfcomm/tty.c index c9191871c1e0..be84f4fc1477 100644 --- a/net/bluetooth/rfcomm/tty.c +++ b/net/bluetooth/rfcomm/tty.c @@ -23,8 +23,6 @@ /* * RFCOMM TTY. - * - * $Id: tty.c,v 1.24 2002/10/03 01:54:38 holtmann Exp $ */ #include diff --git a/net/bridge/br.c b/net/bridge/br.c index 8f3c58e5f7a5..cede010f4ddd 100644 --- a/net/bridge/br.c +++ b/net/bridge/br.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br.c,v 1.47 2001/12/24 00:56:41 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_device.c b/net/bridge/br_device.c index 626c7795ae30..a6ffc6c2a69f 100644 --- a/net/bridge/br_device.c +++ b/net/bridge/br_device.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_device.c,v 1.6 2001/12/24 00:59:55 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_fdb.c b/net/bridge/br_fdb.c index 72c5976a5ce3..4de74cdd091d 100644 --- a/net/bridge/br_fdb.c +++ b/net/bridge/br_fdb.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_fdb.c,v 1.6 2002/01/17 00:57:07 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_forward.c b/net/bridge/br_forward.c index a4711674b3df..512645727f51 100644 --- a/net/bridge/br_forward.c +++ b/net/bridge/br_forward.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_forward.c,v 1.4 2001/08/14 22:05:57 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c index c2397f503b0f..143c954681b8 100644 --- a/net/bridge/br_if.c +++ b/net/bridge/br_if.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_if.c,v 1.7 2001/12/24 00:59:55 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c index fa0f5711a996..0145e9416714 100644 --- a/net/bridge/br_input.c +++ b/net/bridge/br_input.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_input.c,v 1.10 2001/12/24 04:50:20 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_ioctl.c b/net/bridge/br_ioctl.c index 0655a5f07f58..eeee218eed80 100644 --- a/net/bridge/br_ioctl.c +++ b/net/bridge/br_ioctl.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_ioctl.c,v 1.4 2000/11/08 05:16:40 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_notify.c b/net/bridge/br_notify.c index 00644a544e3c..88d8ec7b3142 100644 --- a/net/bridge/br_notify.c +++ b/net/bridge/br_notify.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_notify.c,v 1.2 2000/02/21 15:51:34 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h index 0243cb489edd..83ff5861c2d2 100644 --- a/net/bridge/br_private.h +++ b/net/bridge/br_private.h @@ -4,8 +4,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_private.h,v 1.7 2001/12/24 00:59:55 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_private_stp.h b/net/bridge/br_private_stp.h index e29f01ac1adf..8b650f7fbfa0 100644 --- a/net/bridge/br_private_stp.h +++ b/net/bridge/br_private_stp.h @@ -4,8 +4,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_private_stp.h,v 1.3 2001/02/05 06:03:47 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_stp.c b/net/bridge/br_stp.c index e38034aa56f5..284d1b2fa1ff 100644 --- a/net/bridge/br_stp.c +++ b/net/bridge/br_stp.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_stp.c,v 1.4 2000/06/19 10:13:35 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_stp_bpdu.c b/net/bridge/br_stp_bpdu.c index ddeb6e5d45d6..9dc2de656965 100644 --- a/net/bridge/br_stp_bpdu.c +++ b/net/bridge/br_stp_bpdu.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_stp_bpdu.c,v 1.3 2001/11/10 02:35:25 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_stp_if.c b/net/bridge/br_stp_if.c index 1a430eccec9b..1a4e5c37a0cf 100644 --- a/net/bridge/br_stp_if.c +++ b/net/bridge/br_stp_if.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_stp_if.c,v 1.4 2001/04/14 21:14:39 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/bridge/br_stp_timer.c b/net/bridge/br_stp_timer.c index 77f5255e6915..772a140bfdf0 100644 --- a/net/bridge/br_stp_timer.c +++ b/net/bridge/br_stp_timer.c @@ -5,8 +5,6 @@ * Authors: * Lennert Buytenhek * - * $Id: br_stp_timer.c,v 1.3 2000/05/05 02:17:17 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 1e556d312117..3e18f8525e82 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -4,8 +4,6 @@ * Authors: Alan Cox * Florian La Roche * - * Version: $Id: skbuff.c,v 1.90 2001/11/07 05:56:19 davem Exp $ - * * Fixes: * Alan Cox : Fixed the worst of the load * balancer bugs. diff --git a/net/core/sock.c b/net/core/sock.c index 88094cb09c06..3879bf65897e 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -7,8 +7,6 @@ * handler for protocols to use and generic option handler. * * - * Version: $Id: sock.c,v 1.117 2002/02/01 22:01:03 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Florian La Roche, diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 24eca23c2db3..42bd24b64b57 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -5,8 +5,6 @@ * * PF_INET protocol family socket handler. * - * Version: $Id: af_inet.c,v 1.137 2002/02/01 22:01:03 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Florian La Roche, diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c index 9b539fa9fe18..20c515a1be28 100644 --- a/net/ipv4/arp.c +++ b/net/ipv4/arp.c @@ -1,6 +1,4 @@ /* linux/net/ipv4/arp.c - * - * Version: $Id: arp.c,v 1.99 2001/08/30 22:55:42 davem Exp $ * * Copyright (C) 1994 by Florian La Roche * diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c index 61011e1d580e..f8c0b0aea93a 100644 --- a/net/ipv4/devinet.c +++ b/net/ipv4/devinet.c @@ -1,8 +1,6 @@ /* * NET3 IP device support routines. * - * Version: $Id: devinet.c,v 1.44 2001/10/31 21:55:54 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c index 0b2ac6a3d903..5ad01d63f83b 100644 --- a/net/ipv4/fib_frontend.c +++ b/net/ipv4/fib_frontend.c @@ -5,8 +5,6 @@ * * IPv4 Forwarding Information Base: FIB frontend. * - * Version: $Id: fib_frontend.c,v 1.26 2001/10/31 21:55:54 davem Exp $ - * * Authors: Alexey Kuznetsov, * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/fib_hash.c b/net/ipv4/fib_hash.c index 2e2fc3376ac9..eeec4bf982b8 100644 --- a/net/ipv4/fib_hash.c +++ b/net/ipv4/fib_hash.c @@ -5,8 +5,6 @@ * * IPv4 FIB: lookup engine and maintenance routines. * - * Version: $Id: fib_hash.c,v 1.13 2001/10/31 21:55:54 davem Exp $ - * * Authors: Alexey Kuznetsov, * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c index 3b83c34019fc..9335eba683c3 100644 --- a/net/ipv4/fib_semantics.c +++ b/net/ipv4/fib_semantics.c @@ -5,8 +5,6 @@ * * IPv4 Forwarding Information Base: semantics. * - * Version: $Id: fib_semantics.c,v 1.19 2002/01/12 07:54:56 davem Exp $ - * * Authors: Alexey Kuznetsov, * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c index 4b02d14e7ab9..394db9c941a1 100644 --- a/net/ipv4/fib_trie.c +++ b/net/ipv4/fib_trie.c @@ -22,8 +22,6 @@ * IP-address lookup using LC-tries. Stefan Nilsson and Gunnar Karlsson * IEEE Journal on Selected Areas in Communications, 17(6):1083-1092, June 1999 * - * Version: $Id: fib_trie.c,v 1.3 2005/06/08 14:20:01 robert Exp $ - * * * Code from fib_hash has been reused which includes the following header: * diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c index 87397351ddac..aa7cf46853b7 100644 --- a/net/ipv4/icmp.c +++ b/net/ipv4/icmp.c @@ -3,8 +3,6 @@ * * Alan Cox, * - * Version: $Id: icmp.c,v 1.85 2002/02/01 22:01:03 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c index 2769dc4a4c84..68e84a933e90 100644 --- a/net/ipv4/igmp.c +++ b/net/ipv4/igmp.c @@ -8,8 +8,6 @@ * the older version didn't come out right using gcc 2.5.8, the newer one * seems to fall out with gcc 2.6.2. * - * Version: $Id: igmp.c,v 1.47 2002/02/01 22:01:03 davem Exp $ - * * Authors: * Alan Cox * diff --git a/net/ipv4/inet_diag.c b/net/ipv4/inet_diag.c index da97695e7096..c10036e7a463 100644 --- a/net/ipv4/inet_diag.c +++ b/net/ipv4/inet_diag.c @@ -1,8 +1,6 @@ /* * inet_diag.c Module for monitoring INET transport protocols sockets. * - * Version: $Id: inet_diag.c,v 1.3 2002/02/01 22:01:04 davem Exp $ - * * Authors: Alexey Kuznetsov, * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/inetpeer.c b/net/ipv4/inetpeer.c index af995198f643..a456ceeac3f2 100644 --- a/net/ipv4/inetpeer.c +++ b/net/ipv4/inetpeer.c @@ -3,8 +3,6 @@ * * This source is covered by the GNU GPL, the same as all kernel sources. * - * Version: $Id: inetpeer.c,v 1.7 2001/09/20 21:22:50 davem Exp $ - * * Authors: Andrey V. Savochkin */ diff --git a/net/ipv4/ip_forward.c b/net/ipv4/ip_forward.c index 4813c39b438b..37d36a3f33cd 100644 --- a/net/ipv4/ip_forward.c +++ b/net/ipv4/ip_forward.c @@ -5,8 +5,6 @@ * * The IP forwarding functionality. * - * Version: $Id: ip_forward.c,v 1.48 2000/12/13 18:31:48 davem Exp $ - * * Authors: see ip.c * * Fixes: diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c index be1cb89a8d5a..91e321407313 100644 --- a/net/ipv4/ip_fragment.c +++ b/net/ipv4/ip_fragment.c @@ -5,8 +5,6 @@ * * The IP fragmentation functionality. * - * Version: $Id: ip_fragment.c,v 1.59 2002/01/12 07:54:56 davem Exp $ - * * Authors: Fred N. van Kempen * Alan Cox * diff --git a/net/ipv4/ip_input.c b/net/ipv4/ip_input.c index ff77a4a7f9ec..7c26428ea67b 100644 --- a/net/ipv4/ip_input.c +++ b/net/ipv4/ip_input.c @@ -5,8 +5,6 @@ * * The Internet Protocol (IP) module. * - * Version: $Id: ip_input.c,v 1.55 2002/01/12 07:39:45 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Donald Becker, diff --git a/net/ipv4/ip_options.c b/net/ipv4/ip_options.c index 33126ad2cfdc..be3f18a7a40e 100644 --- a/net/ipv4/ip_options.c +++ b/net/ipv4/ip_options.c @@ -5,8 +5,6 @@ * * The options processing module for ip.c * - * Version: $Id: ip_options.c,v 1.21 2001/09/01 00:31:50 davem Exp $ - * * Authors: A.N.Kuznetsov * */ diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c index e527628f56cf..f1278eecf56d 100644 --- a/net/ipv4/ip_output.c +++ b/net/ipv4/ip_output.c @@ -5,8 +5,6 @@ * * The Internet Protocol (IP) output module. * - * Version: $Id: ip_output.c,v 1.100 2002/02/01 22:01:03 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Donald Becker, diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c index e0514e82308e..105d92a039b9 100644 --- a/net/ipv4/ip_sockglue.c +++ b/net/ipv4/ip_sockglue.c @@ -5,8 +5,6 @@ * * The IP to API glue. * - * Version: $Id: ip_sockglue.c,v 1.62 2002/02/01 22:01:04 davem Exp $ - * * Authors: see ip.c * * Fixes: diff --git a/net/ipv4/ipconfig.c b/net/ipv4/ipconfig.c index ed45037ce9be..b88aa9afa42e 100644 --- a/net/ipv4/ipconfig.c +++ b/net/ipv4/ipconfig.c @@ -1,6 +1,4 @@ /* - * $Id: ipconfig.c,v 1.46 2002/02/01 22:01:04 davem Exp $ - * * Automatic Configuration of IP -- use DHCP, BOOTP, RARP, or * user-supplied information to configure own IP address and routes. * diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c index 86d8836551b9..4c6d2caf9203 100644 --- a/net/ipv4/ipip.c +++ b/net/ipv4/ipip.c @@ -1,8 +1,6 @@ /* * Linux NET3: IP/IP protocol decoder. * - * Version: $Id: ipip.c,v 1.50 2001/10/02 02:22:36 davem Exp $ - * * Authors: * Sam Lantinga (slouken@cs.ucdavis.edu) 02/01/95 * diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index a34da4977c73..300ab0c2919e 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c @@ -9,8 +9,6 @@ * as published by the Free Software Foundation; either version * 2 of the License, or (at your option) any later version. * - * Version: $Id: ipmr.c,v 1.65 2001/10/31 21:55:54 davem Exp $ - * * Fixes: * Michael Chastain : Incorrect size of copying. * Alan Cox : Added the cache manager code diff --git a/net/ipv4/ipvs/ip_vs_app.c b/net/ipv4/ipvs/ip_vs_app.c index 535abe0c45e7..1f1897a1a702 100644 --- a/net/ipv4/ipvs/ip_vs_app.c +++ b/net/ipv4/ipvs/ip_vs_app.c @@ -1,8 +1,6 @@ /* * ip_vs_app.c: Application module support for IPVS * - * Version: $Id: ip_vs_app.c,v 1.17 2003/03/22 06:31:21 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_conn.c b/net/ipv4/ipvs/ip_vs_conn.c index 65f1ba112752..f8bdae47a77f 100644 --- a/net/ipv4/ipvs/ip_vs_conn.c +++ b/net/ipv4/ipvs/ip_vs_conn.c @@ -5,8 +5,6 @@ * high-performance and highly available server based on a * cluster of servers. * - * Version: $Id: ip_vs_conn.c,v 1.31 2003/04/18 09:03:16 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * Julian Anastasov diff --git a/net/ipv4/ipvs/ip_vs_core.c b/net/ipv4/ipvs/ip_vs_core.c index 963981a9d501..bcf6276ba4b2 100644 --- a/net/ipv4/ipvs/ip_vs_core.c +++ b/net/ipv4/ipvs/ip_vs_core.c @@ -5,8 +5,6 @@ * high-performance and highly available server based on a * cluster of servers. * - * Version: $Id: ip_vs_core.c,v 1.34 2003/05/10 03:05:23 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * Julian Anastasov diff --git a/net/ipv4/ipvs/ip_vs_ctl.c b/net/ipv4/ipvs/ip_vs_ctl.c index 94c5767c8e01..9a5ace0b4dd6 100644 --- a/net/ipv4/ipvs/ip_vs_ctl.c +++ b/net/ipv4/ipvs/ip_vs_ctl.c @@ -5,8 +5,6 @@ * high-performance and highly available server based on a * cluster of servers. * - * Version: $Id: ip_vs_ctl.c,v 1.36 2003/06/08 09:31:19 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * Julian Anastasov diff --git a/net/ipv4/ipvs/ip_vs_dh.c b/net/ipv4/ipvs/ip_vs_dh.c index dcf5d46aaa5e..8afc1503ed20 100644 --- a/net/ipv4/ipvs/ip_vs_dh.c +++ b/net/ipv4/ipvs/ip_vs_dh.c @@ -1,8 +1,6 @@ /* * IPVS: Destination Hashing scheduling module * - * Version: $Id: ip_vs_dh.c,v 1.5 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * Inspired by the consistent hashing scheduler patch from diff --git a/net/ipv4/ipvs/ip_vs_est.c b/net/ipv4/ipvs/ip_vs_est.c index dfa0d713c801..bc04eedd6dbb 100644 --- a/net/ipv4/ipvs/ip_vs_est.c +++ b/net/ipv4/ipvs/ip_vs_est.c @@ -1,8 +1,6 @@ /* * ip_vs_est.c: simple rate estimator for IPVS * - * Version: $Id: ip_vs_est.c,v 1.4 2002/11/30 01:50:35 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_ftp.c b/net/ipv4/ipvs/ip_vs_ftp.c index 59aa166b7678..c1c758e4f733 100644 --- a/net/ipv4/ipvs/ip_vs_ftp.c +++ b/net/ipv4/ipvs/ip_vs_ftp.c @@ -1,8 +1,6 @@ /* * ip_vs_ftp.c: IPVS ftp application module * - * Version: $Id: ip_vs_ftp.c,v 1.13 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * Changes: diff --git a/net/ipv4/ipvs/ip_vs_lblc.c b/net/ipv4/ipvs/ip_vs_lblc.c index 3888642706ad..0efa3db4b180 100644 --- a/net/ipv4/ipvs/ip_vs_lblc.c +++ b/net/ipv4/ipvs/ip_vs_lblc.c @@ -1,8 +1,6 @@ /* * IPVS: Locality-Based Least-Connection scheduling module * - * Version: $Id: ip_vs_lblc.c,v 1.10 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_lblcr.c b/net/ipv4/ipvs/ip_vs_lblcr.c index daa260eb21cf..8e3bbeb45138 100644 --- a/net/ipv4/ipvs/ip_vs_lblcr.c +++ b/net/ipv4/ipvs/ip_vs_lblcr.c @@ -1,8 +1,6 @@ /* * IPVS: Locality-Based Least-Connection with Replication scheduler * - * Version: $Id: ip_vs_lblcr.c,v 1.11 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_lc.c b/net/ipv4/ipvs/ip_vs_lc.c index d88fef90a641..ac9f08e065d5 100644 --- a/net/ipv4/ipvs/ip_vs_lc.c +++ b/net/ipv4/ipvs/ip_vs_lc.c @@ -1,8 +1,6 @@ /* * IPVS: Least-Connection Scheduling module * - * Version: $Id: ip_vs_lc.c,v 1.10 2003/04/18 09:03:16 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_nq.c b/net/ipv4/ipvs/ip_vs_nq.c index bc2a9e5f2a7b..a46bf258d420 100644 --- a/net/ipv4/ipvs/ip_vs_nq.c +++ b/net/ipv4/ipvs/ip_vs_nq.c @@ -1,8 +1,6 @@ /* * IPVS: Never Queue scheduling module * - * Version: $Id: ip_vs_nq.c,v 1.2 2003/06/08 09:31:19 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_proto.c b/net/ipv4/ipvs/ip_vs_proto.c index 4b1c16cbb16b..876714f23d65 100644 --- a/net/ipv4/ipvs/ip_vs_proto.c +++ b/net/ipv4/ipvs/ip_vs_proto.c @@ -1,8 +1,6 @@ /* * ip_vs_proto.c: transport protocol load balancing support for IPVS * - * Version: $Id: ip_vs_proto.c,v 1.2 2003/04/18 09:03:16 wensong Exp $ - * * Authors: Wensong Zhang * Julian Anastasov * diff --git a/net/ipv4/ipvs/ip_vs_proto_ah.c b/net/ipv4/ipvs/ip_vs_proto_ah.c index 4bf835e1d86d..73e0ea87c1f5 100644 --- a/net/ipv4/ipvs/ip_vs_proto_ah.c +++ b/net/ipv4/ipvs/ip_vs_proto_ah.c @@ -1,8 +1,6 @@ /* * ip_vs_proto_ah.c: AH IPSec load balancing support for IPVS * - * Version: $Id: ip_vs_proto_ah.c,v 1.1 2003/07/04 15:04:37 wensong Exp $ - * * Authors: Julian Anastasov , February 2002 * Wensong Zhang * diff --git a/net/ipv4/ipvs/ip_vs_proto_esp.c b/net/ipv4/ipvs/ip_vs_proto_esp.c index db6a6b7b1a0b..21d70c8ffa54 100644 --- a/net/ipv4/ipvs/ip_vs_proto_esp.c +++ b/net/ipv4/ipvs/ip_vs_proto_esp.c @@ -1,8 +1,6 @@ /* * ip_vs_proto_esp.c: ESP IPSec load balancing support for IPVS * - * Version: $Id: ip_vs_proto_esp.c,v 1.1 2003/07/04 15:04:37 wensong Exp $ - * * Authors: Julian Anastasov , February 2002 * Wensong Zhang * diff --git a/net/ipv4/ipvs/ip_vs_proto_tcp.c b/net/ipv4/ipvs/ip_vs_proto_tcp.c index b83dc14b0a4d..d0ea467986a0 100644 --- a/net/ipv4/ipvs/ip_vs_proto_tcp.c +++ b/net/ipv4/ipvs/ip_vs_proto_tcp.c @@ -1,8 +1,6 @@ /* * ip_vs_proto_tcp.c: TCP load balancing support for IPVS * - * Version: $Id: ip_vs_proto_tcp.c,v 1.3 2002/11/30 01:50:35 wensong Exp $ - * * Authors: Wensong Zhang * Julian Anastasov * diff --git a/net/ipv4/ipvs/ip_vs_proto_udp.c b/net/ipv4/ipvs/ip_vs_proto_udp.c index 75771cb3cd6f..c6be5d56823f 100644 --- a/net/ipv4/ipvs/ip_vs_proto_udp.c +++ b/net/ipv4/ipvs/ip_vs_proto_udp.c @@ -1,8 +1,6 @@ /* * ip_vs_proto_udp.c: UDP load balancing support for IPVS * - * Version: $Id: ip_vs_proto_udp.c,v 1.3 2002/11/30 01:50:35 wensong Exp $ - * * Authors: Wensong Zhang * Julian Anastasov * diff --git a/net/ipv4/ipvs/ip_vs_rr.c b/net/ipv4/ipvs/ip_vs_rr.c index 433f8a947924..c8db12d39e61 100644 --- a/net/ipv4/ipvs/ip_vs_rr.c +++ b/net/ipv4/ipvs/ip_vs_rr.c @@ -1,8 +1,6 @@ /* * IPVS: Round-Robin Scheduling module * - * Version: $Id: ip_vs_rr.c,v 1.9 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * diff --git a/net/ipv4/ipvs/ip_vs_sched.c b/net/ipv4/ipvs/ip_vs_sched.c index 121a32b1b756..b64767309855 100644 --- a/net/ipv4/ipvs/ip_vs_sched.c +++ b/net/ipv4/ipvs/ip_vs_sched.c @@ -5,8 +5,6 @@ * high-performance and highly available server based on a * cluster of servers. * - * Version: $Id: ip_vs_sched.c,v 1.13 2003/05/10 03:05:23 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * diff --git a/net/ipv4/ipvs/ip_vs_sed.c b/net/ipv4/ipvs/ip_vs_sed.c index dd7c128f9db3..2a7d31358181 100644 --- a/net/ipv4/ipvs/ip_vs_sed.c +++ b/net/ipv4/ipvs/ip_vs_sed.c @@ -1,8 +1,6 @@ /* * IPVS: Shortest Expected Delay scheduling module * - * Version: $Id: ip_vs_sed.c,v 1.1 2003/05/10 03:06:08 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_sh.c b/net/ipv4/ipvs/ip_vs_sh.c index 1b25b00ef1e1..b8fdfac65001 100644 --- a/net/ipv4/ipvs/ip_vs_sh.c +++ b/net/ipv4/ipvs/ip_vs_sh.c @@ -1,8 +1,6 @@ /* * IPVS: Source Hashing scheduling module * - * Version: $Id: ip_vs_sh.c,v 1.5 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_sync.c b/net/ipv4/ipvs/ip_vs_sync.c index eff54efe0351..2d4a86f73325 100644 --- a/net/ipv4/ipvs/ip_vs_sync.c +++ b/net/ipv4/ipvs/ip_vs_sync.c @@ -5,8 +5,6 @@ * high-performance and highly available server based on a * cluster of servers. * - * Version: $Id: ip_vs_sync.c,v 1.13 2003/06/08 09:31:19 wensong Exp $ - * * Authors: Wensong Zhang * * ip_vs_sync: sync connection info from master load balancer to backups diff --git a/net/ipv4/ipvs/ip_vs_wlc.c b/net/ipv4/ipvs/ip_vs_wlc.c index 8a9d913261d8..772c3cb4eca1 100644 --- a/net/ipv4/ipvs/ip_vs_wlc.c +++ b/net/ipv4/ipvs/ip_vs_wlc.c @@ -1,8 +1,6 @@ /* * IPVS: Weighted Least-Connection Scheduling module * - * Version: $Id: ip_vs_wlc.c,v 1.13 2003/04/18 09:03:16 wensong Exp $ - * * Authors: Wensong Zhang * Peter Kese * diff --git a/net/ipv4/ipvs/ip_vs_wrr.c b/net/ipv4/ipvs/ip_vs_wrr.c index 85c680add6df..1d6932d7dc97 100644 --- a/net/ipv4/ipvs/ip_vs_wrr.c +++ b/net/ipv4/ipvs/ip_vs_wrr.c @@ -1,8 +1,6 @@ /* * IPVS: Weighted Round-Robin Scheduling module * - * Version: $Id: ip_vs_wrr.c,v 1.12 2002/09/15 08:14:08 wensong Exp $ - * * Authors: Wensong Zhang * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/ipvs/ip_vs_xmit.c b/net/ipv4/ipvs/ip_vs_xmit.c index f63006caea03..9892d4aca42e 100644 --- a/net/ipv4/ipvs/ip_vs_xmit.c +++ b/net/ipv4/ipvs/ip_vs_xmit.c @@ -1,8 +1,6 @@ /* * ip_vs_xmit.c: various packet transmitters for IPVS * - * Version: $Id: ip_vs_xmit.c,v 1.2 2002/11/30 01:50:35 wensong Exp $ - * * Authors: Wensong Zhang * Julian Anastasov * diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 552169b41b16..eb5cee279c5f 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -7,8 +7,6 @@ * PROC file system. It is mainly used for debugging and * statistics. * - * Version: $Id: proc.c,v 1.45 2001/05/16 16:45:35 davem Exp $ - * * Authors: Fred N. van Kempen, * Gerald J. Heim, * Fred Baumgarten, diff --git a/net/ipv4/protocol.c b/net/ipv4/protocol.c index 971ab9356e51..ea50da0649fd 100644 --- a/net/ipv4/protocol.c +++ b/net/ipv4/protocol.c @@ -5,8 +5,6 @@ * * INET protocol dispatch tables. * - * Version: $Id: protocol.c,v 1.14 2001/05/18 02:25:49 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c index e7e091d365ff..1d0c97c8712d 100644 --- a/net/ipv4/raw.c +++ b/net/ipv4/raw.c @@ -5,8 +5,6 @@ * * RAW - implementation of IP "raw" sockets. * - * Version: $Id: raw.c,v 1.64 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 96be336064fb..fe3a02237286 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -5,8 +5,6 @@ * * ROUTE - implementation of the IP router. * - * Version: $Id: route.c,v 1.103 2002/01/12 07:44:09 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Alan Cox, diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c index 73ba98921d64..6317d3c8dc0d 100644 --- a/net/ipv4/syncookies.c +++ b/net/ipv4/syncookies.c @@ -8,8 +8,6 @@ * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version * 2 of the License, or (at your option) any later version. - * - * $Id: syncookies.c,v 1.18 2002/02/01 22:01:04 davem Exp $ */ #include diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c index c437f804ee38..901607003205 100644 --- a/net/ipv4/sysctl_net_ipv4.c +++ b/net/ipv4/sysctl_net_ipv4.c @@ -1,8 +1,6 @@ /* * sysctl_net_ipv4.c: sysctl interface to net IPV4 subsystem. * - * $Id: sysctl_net_ipv4.c,v 1.50 2001/10/20 00:00:11 davem Exp $ - * * Begun April 1, 1996, Mike Shaver. * Added /proc/sys/net/ipv4 directory entry (empty =) ). [MS] */ diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index ab66683b8043..ad66b09e0bcd 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp.c,v 1.216 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Mark Evans, diff --git a/net/ipv4/tcp_diag.c b/net/ipv4/tcp_diag.c index 2fbcc7d1b1a0..838d491dfda7 100644 --- a/net/ipv4/tcp_diag.c +++ b/net/ipv4/tcp_diag.c @@ -1,8 +1,6 @@ /* * tcp_diag.c Module for monitoring TCP transport protocols sockets. * - * Version: $Id: tcp_diag.c,v 1.3 2002/02/01 22:01:04 davem Exp $ - * * Authors: Alexey Kuznetsov, * * This program is free software; you can redistribute it and/or diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index eba873e9b560..b68c3c7d906b 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp_input.c,v 1.243 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Mark Evans, diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index cd601a866c2f..f2926ae1de57 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp_ipv4.c,v 1.240 2002/02/01 22:01:04 davem Exp $ - * * IPv4 specific functions * * diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index 019c8c16e5cc..1276cab85e3e 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp_minisocks.c,v 1.15 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Mark Evans, diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index ad993ecb4810..b171ac65ccab 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp_output.c,v 1.146 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Mark Evans, diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c index 4de68cf5f2aa..e77e7ae0bf2c 100644 --- a/net/ipv4/tcp_timer.c +++ b/net/ipv4/tcp_timer.c @@ -5,8 +5,6 @@ * * Implementation of the Transmission Control Protocol(TCP). * - * Version: $Id: tcp_timer.c,v 1.88 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Mark Evans, diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 56fcda3694ba..355e6d62d483 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -5,8 +5,6 @@ * * The User Datagram Protocol (UDP). * - * Version: $Id: udp.c,v 1.102 2002/02/01 22:01:04 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Arnt Gulbrandsen, diff --git a/net/ipv4/udplite.c b/net/ipv4/udplite.c index 72ce26b6c4d3..4ad16b6d5138 100644 --- a/net/ipv4/udplite.c +++ b/net/ipv4/udplite.c @@ -1,8 +1,6 @@ /* * UDPLITE An implementation of the UDP-Lite protocol (RFC 3828). * - * Version: $Id: udplite.c,v 1.25 2006/10/19 07:22:36 gerrit Exp $ - * * Authors: Gerrit Renker * * Changes: diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 147588f4c7c0..deb38bf03376 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -6,8 +6,6 @@ * Pedro Roque * Alexey Kuznetsov * - * $Id: addrconf.c,v 1.69 2001/10/31 21:55:54 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c index e84b3fd17fb4..350457c761e6 100644 --- a/net/ipv6/af_inet6.c +++ b/net/ipv6/af_inet6.c @@ -7,8 +7,6 @@ * * Adapted from linux/net/ipv4/af_inet.c * - * $Id: af_inet6.c,v 1.66 2002/02/01 22:01:04 davem Exp $ - * * Fixes: * piggy, Karl Knutson : Socket protocol table * Hideaki YOSHIFUJI : sin6_scope_id support diff --git a/net/ipv6/datagram.c b/net/ipv6/datagram.c index b9c2de84a8a2..8cdb6b65ee96 100644 --- a/net/ipv6/datagram.c +++ b/net/ipv6/datagram.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: datagram.c,v 1.24 2002/02/01 22:01:04 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/ipv6/exthdrs.c b/net/ipv6/exthdrs.c index 3cd1c993d52b..602ea826f0a5 100644 --- a/net/ipv6/exthdrs.c +++ b/net/ipv6/exthdrs.c @@ -7,8 +7,6 @@ * Andi Kleen * Alexey Kuznetsov * - * $Id: exthdrs.c,v 1.13 2001/06/19 15:58:56 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/ipv6/icmp.c b/net/ipv6/icmp.c index d42dd16d3487..399d41f65437 100644 --- a/net/ipv6/icmp.c +++ b/net/ipv6/icmp.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: icmp.c,v 1.38 2002/02/08 03:57:19 davem Exp $ - * * Based on net/ipv4/icmp.c * * RFC 1885 diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c index 1ee4fa17c129..4de2b9efcacb 100644 --- a/net/ipv6/ip6_fib.c +++ b/net/ipv6/ip6_fib.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: ip6_fib.c,v 1.25 2001/10/31 21:55:55 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/ipv6/ip6_input.c b/net/ipv6/ip6_input.c index 4e5c8615832c..f77a6011c302 100644 --- a/net/ipv6/ip6_input.c +++ b/net/ipv6/ip6_input.c @@ -6,8 +6,6 @@ * Pedro Roque * Ian P. Morris * - * $Id: ip6_input.c,v 1.19 2000/12/13 18:31:50 davem Exp $ - * * Based in linux/net/ipv4/ip_input.c * * This program is free software; you can redistribute it and/or diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 48cdce9c696c..40a2813a63d1 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: ip6_output.c,v 1.34 2002/02/01 22:01:04 davem Exp $ - * * Based on linux/net/ipv4/ip_output.c * * This program is free software; you can redistribute it and/or diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c index 37814810ac49..17c7b098cdb0 100644 --- a/net/ipv6/ip6_tunnel.c +++ b/net/ipv6/ip6_tunnel.c @@ -6,8 +6,6 @@ * Ville Nuorvala * Yasuyuki Kozakai * - * $Id$ - * * Based on: * linux/net/ipv6/sit.c and linux/net/ipv4/ipip.c * diff --git a/net/ipv6/ipv6_sockglue.c b/net/ipv6/ipv6_sockglue.c index 26b83e512a09..237ebbb9383b 100644 --- a/net/ipv6/ipv6_sockglue.c +++ b/net/ipv6/ipv6_sockglue.c @@ -7,8 +7,6 @@ * * Based on linux/net/ipv4/ip_sockglue.c * - * $Id: ipv6_sockglue.c,v 1.41 2002/02/01 22:01:04 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c index fd632dd7f98d..fbb2d12c41bc 100644 --- a/net/ipv6/mcast.c +++ b/net/ipv6/mcast.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: mcast.c,v 1.40 2002/02/08 03:57:19 davem Exp $ - * * Based on linux/ipv4/igmp.c and linux/ipv4/ip_sockglue.c * * This program is free software; you can redistribute it and/or diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c index df0736a4cafa..cbc7e514d3ec 100644 --- a/net/ipv6/proc.c +++ b/net/ipv6/proc.c @@ -7,8 +7,6 @@ * PROC file system. This is very similar to the IPv4 version, * except it reports the sockets in the INET6 address family. * - * Version: $Id: proc.c,v 1.17 2002/02/01 22:01:04 davem Exp $ - * * Authors: David S. Miller (davem@caip.rutgers.edu) * YOSHIFUJI Hideaki * diff --git a/net/ipv6/protocol.c b/net/ipv6/protocol.c index f929f47b925e..9ab789159913 100644 --- a/net/ipv6/protocol.c +++ b/net/ipv6/protocol.c @@ -5,8 +5,6 @@ * * PF_INET6 protocol dispatch tables. * - * Version: $Id: protocol.c,v 1.10 2001/05/18 02:25:49 davem Exp $ - * * Authors: Pedro Roque * * This program is free software; you can redistribute it and/or diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c index 8fee9a15b2d3..e03c1898ab2e 100644 --- a/net/ipv6/raw.c +++ b/net/ipv6/raw.c @@ -7,8 +7,6 @@ * * Adapted from linux/net/ipv4/raw.c * - * $Id: raw.c,v 1.51 2002/02/01 22:01:04 davem Exp $ - * * Fixes: * Hideaki YOSHIFUJI : sin6_scope_id support * YOSHIFUJI,H.@USAGI : raw checksum (RFC2292(bis) compliance) diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c index 9391a6949b96..13509f906d89 100644 --- a/net/ipv6/reassembly.c +++ b/net/ipv6/reassembly.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: reassembly.c,v 1.26 2001/03/07 22:00:57 davem Exp $ - * * Based on: net/ipv4/ip_fragment.c * * This program is free software; you can redistribute it and/or diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 220cffe9e63b..edae81319b51 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: route.c,v 1.56 2001/10/31 21:55:55 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c index 6b8f0583b637..b0c5080420a8 100644 --- a/net/ipv6/sit.c +++ b/net/ipv6/sit.c @@ -6,8 +6,6 @@ * Pedro Roque * Alexey Kuznetsov * - * $Id: sit.c,v 1.53 2001/09/25 05:09:53 davem Exp $ - * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 715965f0fac0..155499197fc5 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -5,8 +5,6 @@ * Authors: * Pedro Roque * - * $Id: tcp_ipv6.c,v 1.144 2002/02/01 22:01:04 davem Exp $ - * * Based on: * linux/net/ipv4/tcp.c * linux/net/ipv4/tcp_input.c diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index dd309626ae9a..e0693fffc9bd 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -7,8 +7,6 @@ * * Based on linux/ipv4/udp.c * - * $Id: udp.c,v 1.65 2002/02/01 22:01:04 davem Exp $ - * * Fixes: * Hideaki YOSHIFUJI : sin6_scope_id support * YOSHIFUJI Hideaki @USAGI and: Support IPV6_V6ONLY socket option, which diff --git a/net/ipv6/udplite.c b/net/ipv6/udplite.c index 491efd00a866..f6cdcb348e05 100644 --- a/net/ipv6/udplite.c +++ b/net/ipv6/udplite.c @@ -2,8 +2,6 @@ * UDPLITEv6 An implementation of the UDP-Lite protocol over IPv6. * See also net/ipv4/udplite.c * - * Version: $Id: udplite.c,v 1.9 2006/10/19 08:28:10 gerrit Exp $ - * * Authors: Gerrit Renker * * Changes: diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index 2cee87da4441..beca6402f1cf 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -5,8 +5,6 @@ * * PACKET - implements raw packet sockets. * - * Version: $Id: af_packet.c,v 1.61 2002/02/08 03:57:19 davem Exp $ - * * Authors: Ross Biro * Fred N. van Kempen, * Alan Cox, diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 5bc1ed490180..213071859030 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -24,8 +24,6 @@ * Jiri Fojtasek * fixed requeue routine * and many others. thanks. - * - * $Id: sch_htb.c,v 1.25 2003/12/07 11:08:25 devik Exp devik $ */ #include #include diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c index cc12d5f5d5da..019d4b4478c9 100644 --- a/net/sunrpc/auth_gss/auth_gss.c +++ b/net/sunrpc/auth_gss/auth_gss.c @@ -33,8 +33,6 @@ * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. - * - * $Id$ */ diff --git a/net/sysctl_net.c b/net/sysctl_net.c index d8e79162724c..007c1a6708ee 100644 --- a/net/sysctl_net.c +++ b/net/sysctl_net.c @@ -4,7 +4,6 @@ * Begun April 1, 1996, Mike Shaver. * Added /proc/sys/net directories for each protocol family. [MS] * - * $Log: sysctl_net.c,v $ * Revision 1.2 1996/05/08 20:24:40 shaver * Added bits for NET_BRIDGE and the NET_IPV4_ARP stuff and * NET_IPV4_IP_FORWARD. diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index e18cd3628db4..392e80e3268d 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -8,8 +8,6 @@ * as published by the Free Software Foundation; either version * 2 of the License, or (at your option) any later version. * - * Version: $Id: af_unix.c,v 1.133 2002/02/08 03:57:19 davem Exp $ - * * Fixes: * Linus Torvalds : Assorted bug cures. * Niibe Yutaka : async I/O support. -- cgit v1.2.2 From d2c3cc0070d32bf6cabe6b82942c3e80eae0bfc3 Mon Sep 17 00:00:00 2001 From: Tony Vroon Date: Wed, 11 Jun 2008 16:23:56 -0400 Subject: mac80211: implement EU regulatory domain Implement missing EU regulatory domain for mac80211. Based on the information in IEEE 802.11-2007 (specifically pages 1142, 1143 & 1148) and ETSI 301 893 (V1.4.1). With thanks to Johannes Berg. Signed-off-by: Tony Vroon Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/wireless/reg.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) (limited to 'net') diff --git a/net/wireless/reg.c b/net/wireless/reg.c index 185488da2466..855bff4b3250 100644 --- a/net/wireless/reg.c +++ b/net/wireless/reg.c @@ -80,6 +80,23 @@ static const struct ieee80211_channel_range ieee80211_JP_channels[] = { IEEE80211_CHAN_RADAR), }; +static const struct ieee80211_channel_range ieee80211_EU_channels[] = { + /* IEEE 802.11b/g, channels 1..13 */ + RANGE_PWR(2412, 2472, 20, 6, 0), + /* IEEE 802.11a, channel 36*/ + RANGE_PWR(5180, 5180, 23, 6, IEEE80211_CHAN_PASSIVE_SCAN), + /* IEEE 802.11a, channel 40*/ + RANGE_PWR(5200, 5200, 23, 6, IEEE80211_CHAN_PASSIVE_SCAN), + /* IEEE 802.11a, channel 44*/ + RANGE_PWR(5220, 5220, 23, 6, IEEE80211_CHAN_PASSIVE_SCAN), + /* IEEE 802.11a, channels 48..64 */ + RANGE_PWR(5240, 5320, 23, 6, IEEE80211_CHAN_NO_IBSS | + IEEE80211_CHAN_RADAR), + /* IEEE 802.11a, channels 100..140 */ + RANGE_PWR(5500, 5700, 30, 6, IEEE80211_CHAN_NO_IBSS | + IEEE80211_CHAN_RADAR), +}; + #define REGDOM(_code) \ { \ .code = __stringify(_code), \ @@ -90,6 +107,7 @@ static const struct ieee80211_channel_range ieee80211_JP_channels[] = { static const struct ieee80211_regdomain ieee80211_regdoms[] = { REGDOM(US), REGDOM(JP), + REGDOM(EU), }; -- cgit v1.2.2 From 8d5e0d58b39eed9b0f1064f4a7f2b215869b7e71 Mon Sep 17 00:00:00 2001 From: Ron Rindjunsky Date: Thu, 12 Jun 2008 15:42:29 +0300 Subject: mac80211: do not fragment while aggregation is in use This patch denies the use of framentation while ampdu is used. Signed-off-by: Ron Rindjunsky Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/tx.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 1ad9e664f287..195cb6dd02a0 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -660,9 +660,8 @@ ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) /* * Warn when submitting a fragmented A-MPDU frame and drop it. - * This is an error and needs to be fixed elsewhere, but when - * done needs to take care of monitor interfaces (injection) - * etc. + * This scenario is handled in __ieee80211_tx_prepare but extra + * caution taken here as fragmented ampdu may cause Tx stop. */ if (WARN_ON(tx->flags & IEEE80211_TX_CTL_AMPDU || skb_get_queue_mapping(tx->skb) >= @@ -981,7 +980,8 @@ __ieee80211_tx_prepare(struct ieee80211_tx_data *tx, if (tx->flags & IEEE80211_TX_FRAGMENTED) { if ((tx->flags & IEEE80211_TX_UNICAST) && skb->len + FCS_LEN > local->fragmentation_threshold && - !local->ops->set_frag_threshold) + !local->ops->set_frag_threshold && + !(info->flags & IEEE80211_TX_CTL_AMPDU)) tx->flags |= IEEE80211_TX_FRAGMENTED; else tx->flags &= ~IEEE80211_TX_FRAGMENTED; -- cgit v1.2.2 From dc0ae30c31ee6ef83992bb692f37dfbba08a2ef6 Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Thu, 12 Jun 2008 22:38:37 +0300 Subject: mac80211: fix beacon interval value This patch fixes setting beacon interval 1. in register_hw it honors value requested by the driver 2. It uses default 100 instead of 1000 or 10000. Scanning for beacon interval ~1sec and above is not sane Signed-off-by: Tomas Winkler Signed-off-by: Emmanuel Grumbach Signed-off-by: John W. Linville --- net/mac80211/main.c | 3 ++- net/mac80211/mlme.c | 2 +- 2 files changed, 3 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/mac80211/main.c b/net/mac80211/main.c index b182f018a187..5c5396edad32 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -1707,7 +1707,8 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) debugfs_hw_add(local); - local->hw.conf.beacon_int = 1000; + if (local->hw.conf.beacon_int < 10) + local->hw.conf.beacon_int = 100; local->wstats_flags |= local->hw.flags & (IEEE80211_HW_SIGNAL_UNSPEC | IEEE80211_HW_SIGNAL_DB | diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 7f05820dc629..46f203740834 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -3583,7 +3583,7 @@ static int ieee80211_sta_create_ibss(struct net_device *dev, sband = local->hw.wiphy->bands[bss->band]; if (local->hw.conf.beacon_int == 0) - local->hw.conf.beacon_int = 10000; + local->hw.conf.beacon_int = 100; bss->beacon_int = local->hw.conf.beacon_int; bss->last_update = jiffies; bss->capability = WLAN_CAPABILITY_IBSS; -- cgit v1.2.2 From c9c6950c14ffc0e30e592fec1ebcb203ad3dff10 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 11 Jun 2008 14:21:57 -0700 Subject: mac80211: make ieee80211_get_hdrlen_from_skb return unsigned Many callers already expect it to. Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/util.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/mac80211/util.c b/net/mac80211/util.c index 6513bc2d2707..f3c30d00b21e 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -133,14 +133,14 @@ int ieee80211_get_hdrlen(u16 fc) } EXPORT_SYMBOL(ieee80211_get_hdrlen); -int ieee80211_get_hdrlen_from_skb(const struct sk_buff *skb) +unsigned int ieee80211_get_hdrlen_from_skb(const struct sk_buff *skb) { - const struct ieee80211_hdr *hdr = (const struct ieee80211_hdr *) skb->data; - int hdrlen; + const struct ieee80211_hdr *hdr = (const struct ieee80211_hdr *)skb->data; + unsigned int hdrlen; if (unlikely(skb->len < 10)) return 0; - hdrlen = ieee80211_get_hdrlen(le16_to_cpu(hdr->frame_control)); + hdrlen = ieee80211_hdrlen(hdr->frame_control); if (unlikely(hdrlen > skb->len)) return 0; return hdrlen; -- cgit v1.2.2 From 6693be7124cb8e4f15f0d80ed6e3e50678771737 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 11 Jun 2008 14:21:57 -0700 Subject: mac80211: add utility function to get header length Take a __le16 directly rather than a host-endian value. Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/util.c | 32 ++++++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+) (limited to 'net') diff --git a/net/mac80211/util.c b/net/mac80211/util.c index f3c30d00b21e..9f365a3af96f 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -133,6 +133,38 @@ int ieee80211_get_hdrlen(u16 fc) } EXPORT_SYMBOL(ieee80211_get_hdrlen); +unsigned int ieee80211_hdrlen(__le16 fc) +{ + unsigned int hdrlen = 24; + + if (ieee80211_is_data(fc)) { + if (ieee80211_has_a4(fc)) + hdrlen = 30; + if (ieee80211_is_data_qos(fc)) + hdrlen += IEEE80211_QOS_CTL_LEN; + goto out; + } + + if (ieee80211_is_ctl(fc)) { + /* + * ACK and CTS are 10 bytes, all others 16. To see how + * to get this condition consider + * subtype mask: 0b0000000011110000 (0x00F0) + * ACK subtype: 0b0000000011010000 (0x00D0) + * CTS subtype: 0b0000000011000000 (0x00C0) + * bits that matter: ^^^ (0x00E0) + * value of those: 0b0000000011000000 (0x00C0) + */ + if ((fc & cpu_to_le16(0x00E0)) == cpu_to_le16(0x00C0)) + hdrlen = 10; + else + hdrlen = 16; + } +out: + return hdrlen; +} +EXPORT_SYMBOL(ieee80211_hdrlen); + unsigned int ieee80211_get_hdrlen_from_skb(const struct sk_buff *skb) { const struct ieee80211_hdr *hdr = (const struct ieee80211_hdr *)skb->data; -- cgit v1.2.2 From d5184cacf3eeaeb6ae0c7a02aa44fe589beeda23 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 11 Jun 2008 14:21:58 -0700 Subject: mac80211: wpa.c use new access helpers Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/wpa.c | 63 ++++++++++++++---------------------------------------- 1 file changed, 16 insertions(+), 47 deletions(-) (limited to 'net') diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index 9f6fd20374e1..9834cecaca35 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -24,46 +24,22 @@ static int ieee80211_get_hdr_info(const struct sk_buff *skb, u8 **sa, u8 **da, { struct ieee80211_hdr *hdr; size_t hdrlen; - u16 fc; - int a4_included; - u8 *pos; + __le16 fc; - hdr = (struct ieee80211_hdr *) skb->data; - fc = le16_to_cpu(hdr->frame_control); - - hdrlen = 24; - if ((fc & (IEEE80211_FCTL_FROMDS | IEEE80211_FCTL_TODS)) == - (IEEE80211_FCTL_FROMDS | IEEE80211_FCTL_TODS)) { - hdrlen += ETH_ALEN; - *sa = hdr->addr4; - *da = hdr->addr3; - } else if (fc & IEEE80211_FCTL_FROMDS) { - *sa = hdr->addr3; - *da = hdr->addr1; - } else if (fc & IEEE80211_FCTL_TODS) { - *sa = hdr->addr2; - *da = hdr->addr3; - } else { - *sa = hdr->addr2; - *da = hdr->addr1; - } + hdr = (struct ieee80211_hdr *)skb->data; + fc = hdr->frame_control; - if (fc & 0x80) - hdrlen += 2; + hdrlen = ieee80211_hdrlen(fc); + + *sa = ieee80211_get_SA(hdr); + *da = ieee80211_get_DA(hdr); *data = skb->data + hdrlen; *data_len = skb->len - hdrlen; - a4_included = (fc & (IEEE80211_FCTL_TODS | IEEE80211_FCTL_FROMDS)) == - (IEEE80211_FCTL_TODS | IEEE80211_FCTL_FROMDS); - if ((fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA && - fc & IEEE80211_STYPE_QOS_DATA) { - pos = (u8 *) &hdr->addr4; - if (a4_included) - pos += 6; - *qos_tid = pos[0] & 0x0f; - *qos_tid |= 0x80; /* qos_included flag */ - } else + if (ieee80211_is_data_qos(fc)) + *qos_tid = (*ieee80211_get_qos_ctl(hdr) & 0x0f) | 0x80; + else *qos_tid = 0; return skb->len < hdrlen ? -1 : 0; @@ -186,8 +162,8 @@ static int tkip_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; struct ieee80211_key *key = tx->key; struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); - int hdrlen, len, tail; - u16 fc; + unsigned int hdrlen; + int len, tail; u8 *pos; info->control.icv_len = TKIP_ICV_LEN; @@ -200,8 +176,7 @@ static int tkip_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) return 0; } - fc = le16_to_cpu(hdr->frame_control); - hdrlen = ieee80211_get_hdrlen(fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); len = skb->len - hdrlen; if (tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) @@ -272,14 +247,12 @@ ieee80211_rx_result ieee80211_crypto_tkip_decrypt(struct ieee80211_rx_data *rx) { struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) rx->skb->data; - u16 fc; int hdrlen, res, hwaccel = 0, wpa_test = 0; struct ieee80211_key *key = rx->key; struct sk_buff *skb = rx->skb; DECLARE_MAC_BUF(mac); - fc = le16_to_cpu(hdr->frame_control); - hdrlen = ieee80211_get_hdrlen(fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); if ((rx->fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_DATA) return RX_CONTINUE; @@ -427,7 +400,6 @@ static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) struct ieee80211_key *key = tx->key; struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); int hdrlen, len, tail; - u16 fc; u8 *pos, *pn, *b_0, *aad, *scratch; int i; @@ -446,8 +418,7 @@ static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) b_0 = scratch + 3 * AES_BLOCK_LEN; aad = scratch + 4 * AES_BLOCK_LEN; - fc = le16_to_cpu(hdr->frame_control); - hdrlen = ieee80211_get_hdrlen(fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); len = skb->len - hdrlen; if (key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) @@ -516,7 +487,6 @@ ieee80211_rx_result ieee80211_crypto_ccmp_decrypt(struct ieee80211_rx_data *rx) { struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) rx->skb->data; - u16 fc; int hdrlen; struct ieee80211_key *key = rx->key; struct sk_buff *skb = rx->skb; @@ -524,8 +494,7 @@ ieee80211_crypto_ccmp_decrypt(struct ieee80211_rx_data *rx) int data_len; DECLARE_MAC_BUF(mac); - fc = le16_to_cpu(hdr->frame_control); - hdrlen = ieee80211_get_hdrlen(fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); if ((rx->fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_DATA) return RX_CONTINUE; -- cgit v1.2.2 From a494bb1cae40dd0a98682826d91ddf533cbc864e Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 11 Jun 2008 14:21:58 -0700 Subject: mac80211: use new helpers in util.c - ieee80211_get_bssid() Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/util.c | 41 +++++++++++++++++++---------------------- 1 file changed, 19 insertions(+), 22 deletions(-) (limited to 'net') diff --git a/net/mac80211/util.c b/net/mac80211/util.c index 9f365a3af96f..ce62b163b82c 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -45,38 +45,37 @@ const unsigned char bridge_tunnel_header[] __aligned(2) = u8 *ieee80211_get_bssid(struct ieee80211_hdr *hdr, size_t len, enum ieee80211_if_types type) { - u16 fc; + __le16 fc = hdr->frame_control; /* drop ACK/CTS frames and incorrect hdr len (ctrl) */ if (len < 16) return NULL; - fc = le16_to_cpu(hdr->frame_control); - - switch (fc & IEEE80211_FCTL_FTYPE) { - case IEEE80211_FTYPE_DATA: + if (ieee80211_is_data(fc)) { if (len < 24) /* drop incorrect hdr len (data) */ return NULL; - switch (fc & (IEEE80211_FCTL_TODS | IEEE80211_FCTL_FROMDS)) { - case IEEE80211_FCTL_TODS: - return hdr->addr1; - case (IEEE80211_FCTL_TODS | IEEE80211_FCTL_FROMDS): + + if (ieee80211_has_a4(fc)) return NULL; - case IEEE80211_FCTL_FROMDS: + if (ieee80211_has_tods(fc)) + return hdr->addr1; + if (ieee80211_has_fromds(fc)) return hdr->addr2; - case 0: - return hdr->addr3; - } - break; - case IEEE80211_FTYPE_MGMT: + + return hdr->addr3; + } + + if (ieee80211_is_mgmt(fc)) { if (len < 24) /* drop incorrect hdr len (mgmt) */ return NULL; return hdr->addr3; - case IEEE80211_FTYPE_CTL: - if ((fc & IEEE80211_FCTL_STYPE) == IEEE80211_STYPE_PSPOLL) + } + + if (ieee80211_is_ctl(fc)) { + if(ieee80211_is_pspoll(fc)) return hdr->addr1; - else if ((fc & IEEE80211_FCTL_STYPE) == - IEEE80211_STYPE_BACK_REQ) { + + if (ieee80211_is_back_req(fc)) { switch (type) { case IEEE80211_IF_TYPE_STA: return hdr->addr2; @@ -84,11 +83,9 @@ u8 *ieee80211_get_bssid(struct ieee80211_hdr *hdr, size_t len, case IEEE80211_IF_TYPE_VLAN: return hdr->addr1; default: - return NULL; + break; /* fall through to the return */ } } - else - return NULL; } return NULL; -- cgit v1.2.2 From 002aaf4ea6be3247c246d274979359c3bc93c82a Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 11 Jun 2008 14:21:59 -0700 Subject: mac80211: wme.c use new helpers Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/wme.c | 18 +++++------------- 1 file changed, 5 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 14a9ff10a1e9..d8c2f9688b25 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -105,11 +105,8 @@ static int classify80211(struct sk_buff *skb, struct Qdisc *qd) { struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; - unsigned short fc = le16_to_cpu(hdr->frame_control); - int qos; - /* see if frame is data or non data frame */ - if (unlikely((fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_DATA)) { + if (!ieee80211_is_data(hdr->frame_control)) { /* management frames go on AC_VO queue, but are sent * without QoS control fields */ return 0; @@ -119,10 +116,7 @@ static int classify80211(struct sk_buff *skb, struct Qdisc *qd) /* use AC from radiotap */ } - /* is this a QoS frame? */ - qos = fc & IEEE80211_STYPE_QOS_DATA; - - if (!qos) { + if (!ieee80211_is_data_qos(hdr->frame_control)) { skb->priority = 0; /* required for correct WPA/11i MIC */ return ieee802_1d_to_ac[skb->priority]; } @@ -151,7 +145,6 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) struct ieee80211_sched_data *q = qdisc_priv(qd); struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; - unsigned short fc = le16_to_cpu(hdr->frame_control); struct Qdisc *qdisc; struct sta_info *sta; int err, queue; @@ -185,16 +178,15 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) /* now we know the 1d priority, fill in the QoS header if there is one */ - if (WLAN_FC_IS_QOS_DATA(fc)) { - u8 *p = skb->data + ieee80211_get_hdrlen(fc) - 2; + if (ieee80211_is_data_qos(hdr->frame_control)) { + u8 *p = ieee80211_get_qos_ctl(hdr); u8 ack_policy = 0; tid = skb->priority & QOS_CONTROL_TAG1D_MASK; if (local->wifi_wme_noack_test) ack_policy |= QOS_CONTROL_ACK_POLICY_NOACK << QOS_CONTROL_ACK_POLICY_SHIFT; /* qos header is 2 bytes, second reserved */ - *p = ack_policy | tid; - p++; + *p++ = ack_policy | tid; *p = 0; rcu_read_lock(); -- cgit v1.2.2 From 87228f57434108d8463ff10fd408d8d1273a23d2 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 11 Jun 2008 14:21:59 -0700 Subject: mac80211: rx.c use new helpers Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/rx.c | 19 +++++++------------ 1 file changed, 7 insertions(+), 12 deletions(-) (limited to 'net') diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index a3643fd86af9..e80336f8f1ea 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -67,12 +67,9 @@ static inline int should_drop_frame(struct ieee80211_rx_status *status, return 1; if (unlikely(skb->len < 16 + present_fcs_len + radiotap_len)) return 1; - if (((hdr->frame_control & cpu_to_le16(IEEE80211_FCTL_FTYPE)) == - cpu_to_le16(IEEE80211_FTYPE_CTL)) && - ((hdr->frame_control & cpu_to_le16(IEEE80211_FCTL_STYPE)) != - cpu_to_le16(IEEE80211_STYPE_PSPOLL)) && - ((hdr->frame_control & cpu_to_le16(IEEE80211_FCTL_STYPE)) != - cpu_to_le16(IEEE80211_STYPE_BACK_REQ))) + if (ieee80211_is_ctl(hdr->frame_control) && + !ieee80211_is_pspoll(hdr->frame_control) && + !ieee80211_is_back_req(hdr->frame_control)) return 1; return 0; } @@ -2118,7 +2115,7 @@ static u8 ieee80211_rx_reorder_ampdu(struct ieee80211_local *local, struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; struct sta_info *sta; struct tid_ampdu_rx *tid_agg_rx; - u16 fc, sc; + u16 sc; u16 mpdu_seq_num; u8 ret = 0, *qc; int tid; @@ -2127,14 +2124,12 @@ static u8 ieee80211_rx_reorder_ampdu(struct ieee80211_local *local, if (!sta) return ret; - fc = le16_to_cpu(hdr->frame_control); - /* filter the QoS data rx stream according to * STA/TID and check if this STA/TID is on aggregation */ - if (!WLAN_FC_IS_QOS_DATA(fc)) + if (!ieee80211_is_data_qos(hdr->frame_control)) goto end_reorder; - qc = skb->data + ieee80211_get_hdrlen(fc) - QOS_CONTROL_LEN; + qc = ieee80211_get_qos_ctl(hdr); tid = qc[0] & QOS_CONTROL_TID_MASK; if (sta->ampdu_mlme.tid_state_rx[tid] != HT_AGG_STATE_OPERATIONAL) @@ -2143,7 +2138,7 @@ static u8 ieee80211_rx_reorder_ampdu(struct ieee80211_local *local, tid_agg_rx = sta->ampdu_mlme.tid_rx[tid]; /* null data frames are excluded */ - if (unlikely(fc & IEEE80211_STYPE_NULLFUNC)) + if (unlikely(ieee80211_is_nullfunc(hdr->frame_control))) goto end_reorder; /* new un-ordered ampdu frame - process it */ -- cgit v1.2.2 From c801242c38de247d82f12f6bf28bd19a280a12ae Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 11 Jun 2008 14:22:00 -0700 Subject: mac80211: tkip.c consolidate tkip IV writing in helper Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/tkip.c | 25 ++++++++++++++----------- net/mac80211/tkip.h | 4 ++-- net/mac80211/wpa.c | 8 +------- 3 files changed, 17 insertions(+), 20 deletions(-) (limited to 'net') diff --git a/net/mac80211/tkip.c b/net/mac80211/tkip.c index a00cf1ea7719..a2c8ca1100b2 100644 --- a/net/mac80211/tkip.c +++ b/net/mac80211/tkip.c @@ -64,6 +64,15 @@ static u16 tkipS(u16 val) return tkip_sbox[val & 0xff] ^ swab16(tkip_sbox[val >> 8]); } +static u8 *write_tkip_iv(u8 *pos, u16 iv16) +{ + *pos++ = iv16 >> 8; + *pos++ = ((iv16 >> 8) | 0x20) & 0x7f; + *pos++ = iv16 & 0xFF; + return pos; +} + + /* * P1K := Phase1(TA, TK, TSC) * TA = transmitter address (48 bits) @@ -123,12 +132,9 @@ static void tkip_mixing_phase2(struct ieee80211_key *key, struct tkip_ctx *ctx, ppk[4] += ror16(ppk[3], 1); ppk[5] += ror16(ppk[4], 1); - rc4key[0] = tsc_IV16 >> 8; - rc4key[1] = ((tsc_IV16 >> 8) | 0x20) & 0x7f; - rc4key[2] = tsc_IV16 & 0xFF; - rc4key[3] = ((ppk[5] ^ get_unaligned_le16(tk)) >> 1) & 0xFF; + rc4key = write_tkip_iv(rc4key, tsc_IV16); + *rc4key++ = ((ppk[5] ^ get_unaligned_le16(tk)) >> 1) & 0xFF; - rc4key += 4; for (i = 0; i < 6; i++) put_unaligned_le16(ppk[i], rc4key + 2 * i); } @@ -136,12 +142,9 @@ static void tkip_mixing_phase2(struct ieee80211_key *key, struct tkip_ctx *ctx, /* Add TKIP IV and Ext. IV at @pos. @iv0, @iv1, and @iv2 are the first octets * of the IV. Returns pointer to the octet following IVs (i.e., beginning of * the packet payload). */ -u8 *ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, - u8 iv0, u8 iv1, u8 iv2) +u8 *ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, u16 iv16) { - *pos++ = iv0; - *pos++ = iv1; - *pos++ = iv2; + pos = write_tkip_iv(pos, iv16); *pos++ = (key->conf.keyidx << 6) | (1 << 5) /* Ext IV */; put_unaligned_le32(key->u.tkip.tx.iv32, pos); return pos + 4; @@ -213,7 +216,7 @@ void ieee80211_tkip_encrypt_data(struct crypto_blkcipher *tfm, u8 rc4key[16]; ieee80211_tkip_gen_rc4key(key, ta, rc4key); - pos = ieee80211_tkip_add_iv(pos, key, rc4key[0], rc4key[1], rc4key[2]); + pos = ieee80211_tkip_add_iv(pos, key, key->u.tkip.tx.iv16); ieee80211_wep_encrypt_data(tfm, rc4key, 16, pos, payload_len); } diff --git a/net/mac80211/tkip.h b/net/mac80211/tkip.h index b890427fc959..d4714383f5fc 100644 --- a/net/mac80211/tkip.h +++ b/net/mac80211/tkip.h @@ -13,8 +13,8 @@ #include #include "key.h" -u8 *ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, - u8 iv0, u8 iv1, u8 iv2); +u8 *ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, u16 iv16); + void ieee80211_tkip_encrypt_data(struct crypto_blkcipher *tfm, struct ieee80211_key *key, u8 *pos, size_t payload_len, u8 *ta); diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index 9834cecaca35..345e10e9b313 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -198,14 +198,8 @@ static int tkip_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) key->u.tkip.tx.iv32++; if (tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) { - hdr = (struct ieee80211_hdr *)skb->data; - /* hwaccel - with preallocated room for IV */ - ieee80211_tkip_add_iv(pos, key, - (u8) (key->u.tkip.tx.iv16 >> 8), - (u8) (((key->u.tkip.tx.iv16 >> 8) | 0x20) & - 0x7f), - (u8) key->u.tkip.tx.iv16); + ieee80211_tkip_add_iv(pos, key, key->u.tkip.tx.iv16); info->control.hw_key = &tx->key->conf; return 0; -- cgit v1.2.2 From 7c70537f97fe35f46762247a4bda72c16d585736 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 11 Jun 2008 14:22:00 -0700 Subject: mac80211: tkip.c fold ieee80211_gen_rc4key into its one caller Also change the arguments of the phase1, 2 key mixing to take a pointer to the encrytion key and the tkip_ctx in the same order. Do the dereference of the encryption key in the callers. Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/tkip.c | 38 +++++++++++++++++--------------------- 1 file changed, 17 insertions(+), 21 deletions(-) (limited to 'net') diff --git a/net/mac80211/tkip.c b/net/mac80211/tkip.c index a2c8ca1100b2..8a7dac742b71 100644 --- a/net/mac80211/tkip.c +++ b/net/mac80211/tkip.c @@ -72,7 +72,6 @@ static u8 *write_tkip_iv(u8 *pos, u16 iv16) return pos; } - /* * P1K := Phase1(TA, TK, TSC) * TA = transmitter address (48 bits) @@ -80,11 +79,10 @@ static u8 *write_tkip_iv(u8 *pos, u16 iv16) * TSC = TKIP sequence counter (48 bits, only 32 msb bits used) * P1K: 80 bits */ -static void tkip_mixing_phase1(struct ieee80211_key *key, const u8 *ta, - struct tkip_ctx *ctx, u32 tsc_IV32) +static void tkip_mixing_phase1(const u8 *tk, struct tkip_ctx *ctx, + const u8 *ta, u32 tsc_IV32) { int i, j; - const u8 *tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; u16 *p1k = ctx->p1k; p1k[0] = tsc_IV32 & 0xFFFF; @@ -104,12 +102,11 @@ static void tkip_mixing_phase1(struct ieee80211_key *key, const u8 *ta, ctx->initialized = 1; } -static void tkip_mixing_phase2(struct ieee80211_key *key, struct tkip_ctx *ctx, +static void tkip_mixing_phase2(const u8 *tk, struct tkip_ctx *ctx, u16 tsc_IV16, u8 *rc4key) { u16 ppk[6]; const u16 *p1k = ctx->p1k; - const u8 *tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; int i; ppk[0] = p1k[0]; @@ -150,16 +147,6 @@ u8 *ieee80211_tkip_add_iv(u8 *pos, struct ieee80211_key *key, u16 iv16) return pos + 4; } -static void ieee80211_tkip_gen_rc4key(struct ieee80211_key *key, u8 *ta, - u8 *rc4key) -{ - /* Calculate per-packet key */ - if (key->u.tkip.tx.iv16 == 0 || !key->u.tkip.tx.initialized) - tkip_mixing_phase1(key, ta, &key->u.tkip.tx, key->u.tkip.tx.iv32); - - tkip_mixing_phase2(key, &key->u.tkip.tx, key->u.tkip.tx.iv16, rc4key); -} - void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, struct sk_buff *skb, enum ieee80211_tkip_key_type type, u8 *outkey) @@ -170,6 +157,7 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, u8 *data = (u8 *) hdr; u16 fc = le16_to_cpu(hdr->frame_control); int hdr_len = ieee80211_get_hdrlen(fc); + u8 *tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; u8 *ta = hdr->addr2; u16 iv16; u32 iv32; @@ -193,14 +181,14 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, * might occur after the wrap around of iv16 in the key in case of * fragmented packets. */ if (iv16 == 0 || !key->u.tkip.tx.initialized) - tkip_mixing_phase1(key, ta, &key->u.tkip.tx, iv32); + tkip_mixing_phase1(tk, &key->u.tkip.tx, ta, iv32); if (type == IEEE80211_TKIP_P1_KEY) { memcpy(outkey, key->u.tkip.tx.p1k, sizeof(u16) * 5); return; } - tkip_mixing_phase2(key, &key->u.tkip.tx, iv16, outkey); + tkip_mixing_phase2(tk, &key->u.tkip.tx, iv16, outkey); } EXPORT_SYMBOL(ieee80211_get_tkip_key); @@ -214,8 +202,15 @@ void ieee80211_tkip_encrypt_data(struct crypto_blkcipher *tfm, u8 *pos, size_t payload_len, u8 *ta) { u8 rc4key[16]; + struct tkip_ctx *ctx = &key->u.tkip.tx; + const u8 *tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; + + /* Calculate per-packet key */ + if (ctx->iv16 == 0 || !ctx->initialized) + tkip_mixing_phase1(tk, ctx, ta, ctx->iv32); + + tkip_mixing_phase2(tk, ctx, ctx->iv16, rc4key); - ieee80211_tkip_gen_rc4key(key, ta, rc4key); pos = ieee80211_tkip_add_iv(pos, key, key->u.tkip.tx.iv16); ieee80211_wep_encrypt_data(tfm, rc4key, 16, pos, payload_len); } @@ -234,6 +229,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, u32 iv16; u8 rc4key[16], keyid, *pos = payload; int res; + const u8 *tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; if (payload_len < 12) return -1; @@ -284,7 +280,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, if (!key->u.tkip.rx[queue].initialized || key->u.tkip.rx[queue].iv32 != iv32) { /* IV16 wrapped around - perform TKIP phase 1 */ - tkip_mixing_phase1(key, ta, &key->u.tkip.rx[queue], iv32); + tkip_mixing_phase1(tk, &key->u.tkip.rx[queue], ta, iv32); #ifdef CONFIG_TKIP_DEBUG { int i; @@ -317,7 +313,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, } } - tkip_mixing_phase2(key, &key->u.tkip.rx[queue], iv16, rc4key); + tkip_mixing_phase2(tk, &key->u.tkip.rx[queue], iv16, rc4key); #ifdef CONFIG_TKIP_DEBUG { int i; -- cgit v1.2.2 From c644bce95f287e763a0b49e5d03f0fe6256f6d2e Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 11 Jun 2008 14:22:02 -0700 Subject: mac80211: tkip.c use a local struct tkip_ctx in ieee80211_get_tkip_key Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/tkip.c | 30 ++++++++++++++++-------------- 1 file changed, 16 insertions(+), 14 deletions(-) (limited to 'net') diff --git a/net/mac80211/tkip.c b/net/mac80211/tkip.c index 8a7dac742b71..e710243d82e2 100644 --- a/net/mac80211/tkip.c +++ b/net/mac80211/tkip.c @@ -153,25 +153,27 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, { struct ieee80211_key *key = (struct ieee80211_key *) container_of(keyconf, struct ieee80211_key, conf); - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; - u8 *data = (u8 *) hdr; - u16 fc = le16_to_cpu(hdr->frame_control); - int hdr_len = ieee80211_get_hdrlen(fc); - u8 *tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; - u8 *ta = hdr->addr2; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data; + u8 *data; + const u8 *tk; + struct tkip_ctx *ctx; u16 iv16; u32 iv32; - iv16 = data[hdr_len + 2] | (data[hdr_len] << 8); - iv32 = get_unaligned_le32(data + hdr_len + 4); + data = (u8 *)hdr + ieee80211_hdrlen(hdr->frame_control); + iv16 = data[2] | (data[0] << 8); + iv32 = get_unaligned_le32(&data[4]); + + tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; + ctx = &key->u.tkip.tx; #ifdef CONFIG_TKIP_DEBUG printk(KERN_DEBUG "TKIP encrypt: iv16 = 0x%04x, iv32 = 0x%08x\n", iv16, iv32); - if (iv32 != key->u.tkip.tx.iv32) { + if (iv32 != ctx->iv32) { printk(KERN_DEBUG "skb: iv32 = 0x%08x key: iv32 = 0x%08x\n", - iv32, key->u.tkip.tx.iv32); + iv32, ctx->iv32); printk(KERN_DEBUG "Wrap around of iv16 in the middle of a " "fragmented packet\n"); } @@ -180,15 +182,15 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, /* Update the p1k only when the iv16 in the packet wraps around, this * might occur after the wrap around of iv16 in the key in case of * fragmented packets. */ - if (iv16 == 0 || !key->u.tkip.tx.initialized) - tkip_mixing_phase1(tk, &key->u.tkip.tx, ta, iv32); + if (iv16 == 0 || !ctx->initialized) + tkip_mixing_phase1(tk, ctx, hdr->addr2, iv32); if (type == IEEE80211_TKIP_P1_KEY) { - memcpy(outkey, key->u.tkip.tx.p1k, sizeof(u16) * 5); + memcpy(outkey, ctx->p1k, sizeof(u16) * 5); return; } - tkip_mixing_phase2(tk, &key->u.tkip.tx, iv16, outkey); + tkip_mixing_phase2(tk, ctx, iv16, outkey); } EXPORT_SYMBOL(ieee80211_get_tkip_key); -- cgit v1.2.2 From 87291c0269e77b029282676448fed3706a54211a Mon Sep 17 00:00:00 2001 From: Vladimir Koutny Date: Fri, 13 Jun 2008 16:50:44 +0200 Subject: mac80211: eliminate IBSS warning in rate_lowest_index() In IBSS mode prior to join/creation of new IBSS it is possible that a frame from unknown station is received and an ibss_add_sta() is called. This will cause a warning in rate_lowest_index() since the list of supported rates of our station is not initialized yet. The fix is to add ibss stations with a rate we received that frame at; this single-element set will be extended later based on beacon data. Also there is no need to store stations from a foreign IBSS. Signed-off-by: Vladimir Koutny Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 2 +- net/mac80211/mlme.c | 15 +++++++++++---- net/mac80211/rx.c | 10 ++++++++-- 3 files changed, 20 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index b19bd16703b2..14fccf16b80f 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -876,7 +876,7 @@ void ieee80211_rx_bss_list_deinit(struct net_device *dev); int ieee80211_sta_set_extra_ie(struct net_device *dev, char *ie, size_t len); struct sta_info *ieee80211_ibss_add_sta(struct net_device *dev, struct sk_buff *skb, u8 *bssid, - u8 *addr); + u8 *addr, u64 supp_rates); int ieee80211_sta_deauthenticate(struct net_device *dev, u16 reason); int ieee80211_sta_disassociate(struct net_device *dev, u16 reason); void ieee80211_bss_info_change_notify(struct ieee80211_sub_if_data *sdata, diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 46f203740834..55659a730dc1 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -2863,7 +2863,8 @@ static void ieee80211_rx_bss_info(struct net_device *dev, dev->name, print_mac(mac, mgmt->bssid)); ieee80211_sta_join_ibss(dev, &sdata->u.sta, bss); ieee80211_ibss_add_sta(dev, NULL, - mgmt->bssid, mgmt->sa); + mgmt->bssid, mgmt->sa, + BIT(rx_status->rate_idx)); } } @@ -4307,12 +4308,13 @@ int ieee80211_sta_set_extra_ie(struct net_device *dev, char *ie, size_t len) struct sta_info *ieee80211_ibss_add_sta(struct net_device *dev, struct sk_buff *skb, u8 *bssid, - u8 *addr) + u8 *addr, u64 supp_rates) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct sta_info *sta; struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); DECLARE_MAC_BUF(mac); + int band = local->hw.conf.channel->band; /* TODO: Could consider removing the least recently used entry and * allow new one to be added. */ @@ -4324,6 +4326,9 @@ struct sta_info *ieee80211_ibss_add_sta(struct net_device *dev, return NULL; } + if (!ieee80211_bssid_match(bssid, sdata->u.sta.bssid)) + return NULL; + printk(KERN_DEBUG "%s: Adding new IBSS station %s (dev=%s)\n", wiphy_name(local->hw.wiphy), print_mac(mac, addr), dev->name); @@ -4333,8 +4338,10 @@ struct sta_info *ieee80211_ibss_add_sta(struct net_device *dev, set_sta_flags(sta, WLAN_STA_AUTHORIZED); - sta->supp_rates[local->hw.conf.channel->band] = - sdata->u.sta.supp_rates_bits[local->hw.conf.channel->band]; + if (supp_rates) + sta->supp_rates[band] = supp_rates; + else + sta->supp_rates[band] = sdata->u.sta.supp_rates_bits[band]; rate_control_rate_init(sta, local); diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index e80336f8f1ea..c32a0bcd53b7 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -1823,8 +1823,13 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata, if (!bssid) return 0; if ((rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_MGMT && - (rx->fc & IEEE80211_FCTL_STYPE) == IEEE80211_STYPE_BEACON) + (rx->fc & IEEE80211_FCTL_STYPE) == IEEE80211_STYPE_BEACON) { + if (!rx->sta) + rx->sta = ieee80211_ibss_add_sta(sdata->dev, + rx->skb, bssid, hdr->addr2, + BIT(rx->status->rate_idx)); return 1; + } else if (!ieee80211_bssid_match(bssid, sdata->u.sta.bssid)) { if (!(rx->flags & IEEE80211_RX_IN_SCAN)) return 0; @@ -1837,7 +1842,8 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata, rx->flags &= ~IEEE80211_RX_RA_MATCH; } else if (!rx->sta) rx->sta = ieee80211_ibss_add_sta(sdata->dev, rx->skb, - bssid, hdr->addr2); + bssid, hdr->addr2, + BIT(rx->status->rate_idx)); break; case IEEE80211_IF_TYPE_MESH_POINT: if (!multicast && -- cgit v1.2.2 From 7d06b2e053d2d536348e3a0f6bb02982a41bea37 Mon Sep 17 00:00:00 2001 From: Brian Haley Date: Sat, 14 Jun 2008 17:04:49 -0700 Subject: net: change proto destroy method to return void Change struct proto destroy function pointer to return void. Noticed by Al Viro. Signed-off-by: Brian Haley Signed-off-by: David S. Miller --- net/dccp/dccp.h | 2 +- net/dccp/ipv6.c | 4 ++-- net/dccp/proto.c | 4 +--- net/ipv4/raw.c | 3 +-- net/ipv4/tcp_ipv4.c | 4 +--- net/ipv4/udp.c | 3 +-- net/ipv4/udp_impl.h | 2 +- net/ipv6/af_inet6.c | 4 +--- net/ipv6/raw.c | 4 ++-- net/ipv6/tcp_ipv6.c | 4 ++-- net/ipv6/udp.c | 4 +--- net/ipv6/udp_impl.h | 2 +- net/sctp/socket.c | 3 +-- 13 files changed, 16 insertions(+), 27 deletions(-) (limited to 'net') diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h index f44d492d3b74..1b2cea244e12 100644 --- a/net/dccp/dccp.h +++ b/net/dccp/dccp.h @@ -262,7 +262,7 @@ extern int dccp_rcv_established(struct sock *sk, struct sk_buff *skb, const struct dccp_hdr *dh, const unsigned len); extern int dccp_init_sock(struct sock *sk, const __u8 ctl_sock_initialized); -extern int dccp_destroy_sock(struct sock *sk); +extern void dccp_destroy_sock(struct sock *sk); extern void dccp_close(struct sock *sk, long timeout); extern struct sk_buff *dccp_make_response(struct sock *sk, diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c index f7fe2a572d7b..eec3c4717890 100644 --- a/net/dccp/ipv6.c +++ b/net/dccp/ipv6.c @@ -1091,10 +1091,10 @@ static int dccp_v6_init_sock(struct sock *sk) return err; } -static int dccp_v6_destroy_sock(struct sock *sk) +static void dccp_v6_destroy_sock(struct sock *sk) { dccp_destroy_sock(sk); - return inet6_destroy_sock(sk); + inet6_destroy_sock(sk); } static struct timewait_sock_ops dccp6_timewait_sock_ops = { diff --git a/net/dccp/proto.c b/net/dccp/proto.c index 9dfe2470962c..a0b56009611f 100644 --- a/net/dccp/proto.c +++ b/net/dccp/proto.c @@ -237,7 +237,7 @@ int dccp_init_sock(struct sock *sk, const __u8 ctl_sock_initialized) EXPORT_SYMBOL_GPL(dccp_init_sock); -int dccp_destroy_sock(struct sock *sk) +void dccp_destroy_sock(struct sock *sk) { struct dccp_sock *dp = dccp_sk(sk); struct dccp_minisock *dmsk = dccp_msk(sk); @@ -268,8 +268,6 @@ int dccp_destroy_sock(struct sock *sk) /* clean up feature negotiation state */ dccp_feat_clean(dmsk); - - return 0; } EXPORT_SYMBOL_GPL(dccp_destroy_sock); diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c index 1d0c97c8712d..36035a0c6dc2 100644 --- a/net/ipv4/raw.c +++ b/net/ipv4/raw.c @@ -606,12 +606,11 @@ static void raw_close(struct sock *sk, long timeout) sk_common_release(sk); } -static int raw_destroy(struct sock *sk) +static void raw_destroy(struct sock *sk) { lock_sock(sk); ip_flush_pending_frames(sk); release_sock(sk); - return 0; } /* This gets rid of all the nasties in af_inet. -DaveM */ diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index b219a7a7cd08..64b385f65930 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -1775,7 +1775,7 @@ static int tcp_v4_init_sock(struct sock *sk) return 0; } -int tcp_v4_destroy_sock(struct sock *sk) +void tcp_v4_destroy_sock(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); @@ -1819,8 +1819,6 @@ int tcp_v4_destroy_sock(struct sock *sk) } atomic_dec(&tcp_sockets_allocated); - - return 0; } EXPORT_SYMBOL(tcp_v4_destroy_sock); diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 355e6d62d483..eba790dcd16b 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -1253,12 +1253,11 @@ int udp_rcv(struct sk_buff *skb) return __udp4_lib_rcv(skb, udp_hash, IPPROTO_UDP); } -int udp_destroy_sock(struct sock *sk) +void udp_destroy_sock(struct sock *sk) { lock_sock(sk); udp_flush_pending_frames(sk); release_sock(sk); - return 0; } /* diff --git a/net/ipv4/udp_impl.h b/net/ipv4/udp_impl.h index 7288bf7977fb..2e9bad2fa1bc 100644 --- a/net/ipv4/udp_impl.h +++ b/net/ipv4/udp_impl.h @@ -26,7 +26,7 @@ extern int udp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg, extern int udp_sendpage(struct sock *sk, struct page *page, int offset, size_t size, int flags); extern int udp_queue_rcv_skb(struct sock * sk, struct sk_buff *skb); -extern int udp_destroy_sock(struct sock *sk); +extern void udp_destroy_sock(struct sock *sk); #ifdef CONFIG_PROC_FS extern int udp4_seq_show(struct seq_file *seq, void *v); diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c index 350457c761e6..3ce8d2f318c6 100644 --- a/net/ipv6/af_inet6.c +++ b/net/ipv6/af_inet6.c @@ -371,7 +371,7 @@ int inet6_release(struct socket *sock) EXPORT_SYMBOL(inet6_release); -int inet6_destroy_sock(struct sock *sk) +void inet6_destroy_sock(struct sock *sk) { struct ipv6_pinfo *np = inet6_sk(sk); struct sk_buff *skb; @@ -389,8 +389,6 @@ int inet6_destroy_sock(struct sock *sk) if ((opt = xchg(&np->opt, NULL)) != NULL) sock_kfree_s(sk, opt, opt->tot_len); - - return 0; } EXPORT_SYMBOL_GPL(inet6_destroy_sock); diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c index 70a57e45bf0e..456777d7a407 100644 --- a/net/ipv6/raw.c +++ b/net/ipv6/raw.c @@ -1162,13 +1162,13 @@ static void rawv6_close(struct sock *sk, long timeout) sk_common_release(sk); } -static int raw6_destroy(struct sock *sk) +static void raw6_destroy(struct sock *sk) { lock_sock(sk); ip6_flush_pending_frames(sk); release_sock(sk); - return inet6_destroy_sock(sk); + inet6_destroy_sock(sk); } static int rawv6_init_sk(struct sock *sk) diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index ebed5d3adb82..daefc18d50a0 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1872,7 +1872,7 @@ static int tcp_v6_init_sock(struct sock *sk) return 0; } -static int tcp_v6_destroy_sock(struct sock *sk) +static void tcp_v6_destroy_sock(struct sock *sk) { #ifdef CONFIG_TCP_MD5SIG /* Clean up the MD5 key list */ @@ -1880,7 +1880,7 @@ static int tcp_v6_destroy_sock(struct sock *sk) tcp_v6_clear_md5_list(sk); #endif tcp_v4_destroy_sock(sk); - return inet6_destroy_sock(sk); + inet6_destroy_sock(sk); } #ifdef CONFIG_PROC_FS diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index e0693fffc9bd..09687f7a8564 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -879,15 +879,13 @@ do_confirm: goto out; } -int udpv6_destroy_sock(struct sock *sk) +void udpv6_destroy_sock(struct sock *sk) { lock_sock(sk); udp_v6_flush_pending_frames(sk); release_sock(sk); inet6_destroy_sock(sk); - - return 0; } /* diff --git a/net/ipv6/udp_impl.h b/net/ipv6/udp_impl.h index 321b81a4d418..92dd7da766d8 100644 --- a/net/ipv6/udp_impl.h +++ b/net/ipv6/udp_impl.h @@ -29,7 +29,7 @@ extern int udpv6_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg, size_t len, int noblock, int flags, int *addr_len); extern int udpv6_queue_rcv_skb(struct sock * sk, struct sk_buff *skb); -extern int udpv6_destroy_sock(struct sock *sk); +extern void udpv6_destroy_sock(struct sock *sk); #ifdef CONFIG_PROC_FS extern int udp6_seq_show(struct seq_file *seq, void *v); diff --git a/net/sctp/socket.c b/net/sctp/socket.c index 253e5ea7e1e8..f98650cc48d8 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -3588,7 +3588,7 @@ SCTP_STATIC int sctp_init_sock(struct sock *sk) } /* Cleanup any SCTP per socket resources. */ -SCTP_STATIC int sctp_destroy_sock(struct sock *sk) +SCTP_STATIC void sctp_destroy_sock(struct sock *sk) { struct sctp_endpoint *ep; @@ -3598,7 +3598,6 @@ SCTP_STATIC int sctp_destroy_sock(struct sock *sk) ep = sctp_sk(sk)->ep; sctp_endpoint_free(ep); atomic_dec(&sctp_sockets_allocated); - return 0; } /* API 4.1.7 shutdown() - TCP Style Syntax -- cgit v1.2.2 From d6266281f8175e3ad68c28b20a609b278b47ade5 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 16 Jun 2008 17:11:50 -0700 Subject: udp: introduce a udp_hashfn function Currently the chain to store a UDP socket is calculated with simple (x & (UDP_HTABLE_SIZE - 1)). But taking net into account would make this calculation a bit more complex, so moving it into a function would help. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/udp.c | 12 ++++++------ net/ipv6/udp.c | 4 ++-- 2 files changed, 8 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index eba790dcd16b..d8f527d15701 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -134,7 +134,7 @@ static inline int __udp_lib_lport_inuse(struct net *net, __u16 num, struct sock *sk; struct hlist_node *node; - sk_for_each(sk, node, &udptable[num & (UDP_HTABLE_SIZE - 1)]) + sk_for_each(sk, node, &udptable[udp_hashfn(num)]) if (net_eq(sock_net(sk), net) && sk->sk_hash == num) return 1; return 0; @@ -174,7 +174,7 @@ int udp_lib_get_port(struct sock *sk, unsigned short snum, for (i = 0; i < UDP_HTABLE_SIZE; i++) { int size = 0; - head = &udptable[rover & (UDP_HTABLE_SIZE - 1)]; + head = &udptable[udp_hashfn(rover)]; if (hlist_empty(head)) goto gotit; @@ -211,7 +211,7 @@ int udp_lib_get_port(struct sock *sk, unsigned short snum, gotit: snum = rover; } else { - head = &udptable[snum & (UDP_HTABLE_SIZE - 1)]; + head = &udptable[udp_hashfn(snum)]; sk_for_each(sk2, node, head) if (sk2->sk_hash == snum && @@ -227,7 +227,7 @@ gotit: inet_sk(sk)->num = snum; sk->sk_hash = snum; if (sk_unhashed(sk)) { - head = &udptable[snum & (UDP_HTABLE_SIZE - 1)]; + head = &udptable[udp_hashfn(snum)]; sk_add_node(sk, head); sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1); } @@ -264,7 +264,7 @@ static struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr, int badness = -1; read_lock(&udp_hash_lock); - sk_for_each(sk, node, &udptable[hnum & (UDP_HTABLE_SIZE - 1)]) { + sk_for_each(sk, node, &udptable[udp_hashfn(hnum)]) { struct inet_sock *inet = inet_sk(sk); if (net_eq(sock_net(sk), net) && sk->sk_hash == hnum && @@ -1068,7 +1068,7 @@ static int __udp4_lib_mcast_deliver(struct sk_buff *skb, int dif; read_lock(&udp_hash_lock); - sk = sk_head(&udptable[ntohs(uh->dest) & (UDP_HTABLE_SIZE - 1)]); + sk = sk_head(&udptable[udp_hashfn(ntohs(uh->dest))]); dif = skb->dev->ifindex; sk = udp_v4_mcast_next(sk, uh->dest, daddr, uh->source, saddr, dif); if (sk) { diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index 09687f7a8564..6e4a822ba651 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -65,7 +65,7 @@ static struct sock *__udp6_lib_lookup(struct net *net, int badness = -1; read_lock(&udp_hash_lock); - sk_for_each(sk, node, &udptable[hnum & (UDP_HTABLE_SIZE - 1)]) { + sk_for_each(sk, node, &udptable[udp_hashfn(hnum)]) { struct inet_sock *inet = inet_sk(sk); if (net_eq(sock_net(sk), net) && sk->sk_hash == hnum && @@ -361,7 +361,7 @@ static int __udp6_lib_mcast_deliver(struct sk_buff *skb, struct in6_addr *saddr, int dif; read_lock(&udp_hash_lock); - sk = sk_head(&udptable[ntohs(uh->dest) & (UDP_HTABLE_SIZE - 1)]); + sk = sk_head(&udptable[udp_hashfn(ntohs(uh->dest))]); dif = inet6_iif(skb); sk = udp_v6_mcast_next(sk, uh->dest, daddr, uh->source, saddr, dif); if (!sk) { -- cgit v1.2.2 From e31634931d00081c75e3fb3f3ec51a50dbf108bb Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 16 Jun 2008 17:12:11 -0700 Subject: udp: provide a struct net pointer for __udp[46]_lib_mcast_deliver They both calculate the hash chain, but currently do not have a struct net pointer, so pass one there via additional argument, all the more so their callers already have such. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/udp.c | 9 ++++++--- net/ipv6/udp.c | 12 ++++++++---- 2 files changed, 14 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index d8f527d15701..6b0acb438f34 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -1059,7 +1059,7 @@ drop: * Note: called only from the BH handler context, * so we don't need to lock the hashes. */ -static int __udp4_lib_mcast_deliver(struct sk_buff *skb, +static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb, struct udphdr *uh, __be32 saddr, __be32 daddr, struct hlist_head udptable[]) @@ -1156,6 +1156,7 @@ int __udp4_lib_rcv(struct sk_buff *skb, struct hlist_head udptable[], struct rtable *rt = (struct rtable*)skb->dst; __be32 saddr = ip_hdr(skb)->saddr; __be32 daddr = ip_hdr(skb)->daddr; + struct net *net; /* * Validate the packet. @@ -1177,10 +1178,12 @@ int __udp4_lib_rcv(struct sk_buff *skb, struct hlist_head udptable[], if (udp4_csum_init(skb, uh, proto)) goto csum_error; + net = dev_net(skb->dev); if (rt->rt_flags & (RTCF_BROADCAST|RTCF_MULTICAST)) - return __udp4_lib_mcast_deliver(skb, uh, saddr, daddr, udptable); + return __udp4_lib_mcast_deliver(net, skb, uh, + saddr, daddr, udptable); - sk = __udp4_lib_lookup(dev_net(skb->dev), saddr, uh->source, daddr, + sk = __udp4_lib_lookup(net, saddr, uh->source, daddr, uh->dest, inet_iif(skb), udptable); if (sk != NULL) { diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index 6e4a822ba651..80fb72c48976 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -353,8 +353,9 @@ static struct sock *udp_v6_mcast_next(struct sock *sk, * Note: called only from the BH handler context, * so we don't need to lock the hashes. */ -static int __udp6_lib_mcast_deliver(struct sk_buff *skb, struct in6_addr *saddr, - struct in6_addr *daddr, struct hlist_head udptable[]) +static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb, + struct in6_addr *saddr, struct in6_addr *daddr, + struct hlist_head udptable[]) { struct sock *sk, *sk2; const struct udphdr *uh = udp_hdr(skb); @@ -435,6 +436,7 @@ int __udp6_lib_rcv(struct sk_buff *skb, struct hlist_head udptable[], struct net_device *dev = skb->dev; struct in6_addr *saddr, *daddr; u32 ulen = 0; + struct net *net; if (!pskb_may_pull(skb, sizeof(struct udphdr))) goto short_packet; @@ -469,11 +471,13 @@ int __udp6_lib_rcv(struct sk_buff *skb, struct hlist_head udptable[], if (udp6_csum_init(skb, uh, proto)) goto discard; + net = dev_net(skb->dev); /* * Multicast receive code */ if (ipv6_addr_is_multicast(daddr)) - return __udp6_lib_mcast_deliver(skb, saddr, daddr, udptable); + return __udp6_lib_mcast_deliver(net, skb, + saddr, daddr, udptable); /* Unicast */ @@ -481,7 +485,7 @@ int __udp6_lib_rcv(struct sk_buff *skb, struct hlist_head udptable[], * check socket cache ... must talk to Alan about his plans * for sock caches... i'll skip this for now. */ - sk = __udp6_lib_lookup(dev_net(skb->dev), saddr, uh->source, + sk = __udp6_lib_lookup(net, saddr, uh->source, daddr, uh->dest, inet6_iif(skb), udptable); if (sk == NULL) { -- cgit v1.2.2 From 19c7578fb22b0aef103222cae9b522f03ae489d6 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 16 Jun 2008 17:12:29 -0700 Subject: udp: add struct net argument to udp_hashfn Every caller already has this one. The new argument is currently unused, but this will be fixed shortly. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/udp.c | 12 ++++++------ net/ipv6/udp.c | 4 ++-- 2 files changed, 8 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 6b0acb438f34..11eabf136144 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -134,7 +134,7 @@ static inline int __udp_lib_lport_inuse(struct net *net, __u16 num, struct sock *sk; struct hlist_node *node; - sk_for_each(sk, node, &udptable[udp_hashfn(num)]) + sk_for_each(sk, node, &udptable[udp_hashfn(net, num)]) if (net_eq(sock_net(sk), net) && sk->sk_hash == num) return 1; return 0; @@ -174,7 +174,7 @@ int udp_lib_get_port(struct sock *sk, unsigned short snum, for (i = 0; i < UDP_HTABLE_SIZE; i++) { int size = 0; - head = &udptable[udp_hashfn(rover)]; + head = &udptable[udp_hashfn(net, rover)]; if (hlist_empty(head)) goto gotit; @@ -211,7 +211,7 @@ int udp_lib_get_port(struct sock *sk, unsigned short snum, gotit: snum = rover; } else { - head = &udptable[udp_hashfn(snum)]; + head = &udptable[udp_hashfn(net, snum)]; sk_for_each(sk2, node, head) if (sk2->sk_hash == snum && @@ -227,7 +227,7 @@ gotit: inet_sk(sk)->num = snum; sk->sk_hash = snum; if (sk_unhashed(sk)) { - head = &udptable[udp_hashfn(snum)]; + head = &udptable[udp_hashfn(net, snum)]; sk_add_node(sk, head); sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1); } @@ -264,7 +264,7 @@ static struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr, int badness = -1; read_lock(&udp_hash_lock); - sk_for_each(sk, node, &udptable[udp_hashfn(hnum)]) { + sk_for_each(sk, node, &udptable[udp_hashfn(net, hnum)]) { struct inet_sock *inet = inet_sk(sk); if (net_eq(sock_net(sk), net) && sk->sk_hash == hnum && @@ -1068,7 +1068,7 @@ static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb, int dif; read_lock(&udp_hash_lock); - sk = sk_head(&udptable[udp_hashfn(ntohs(uh->dest))]); + sk = sk_head(&udptable[udp_hashfn(net, ntohs(uh->dest))]); dif = skb->dev->ifindex; sk = udp_v4_mcast_next(sk, uh->dest, daddr, uh->source, saddr, dif); if (sk) { diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index 80fb72c48976..432edaa882f6 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -65,7 +65,7 @@ static struct sock *__udp6_lib_lookup(struct net *net, int badness = -1; read_lock(&udp_hash_lock); - sk_for_each(sk, node, &udptable[udp_hashfn(hnum)]) { + sk_for_each(sk, node, &udptable[udp_hashfn(net, hnum)]) { struct inet_sock *inet = inet_sk(sk); if (net_eq(sock_net(sk), net) && sk->sk_hash == hnum && @@ -362,7 +362,7 @@ static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb, int dif; read_lock(&udp_hash_lock); - sk = sk_head(&udptable[udp_hashfn(ntohs(uh->dest))]); + sk = sk_head(&udptable[udp_hashfn(net, ntohs(uh->dest))]); dif = inet6_iif(skb); sk = udp_v6_mcast_next(sk, uh->dest, daddr, uh->source, saddr, dif); if (!sk) { -- cgit v1.2.2 From 7f635ab71eef8da012320c0092b662d6af8c1e69 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 16 Jun 2008 17:12:49 -0700 Subject: inet: add struct net argument to inet_bhashfn Binding to some port in many namespaces may create too long chains in bhash-es, so prepare the hashfn to take struct net into account. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/inet_connection_sock.c | 6 ++++-- net/ipv4/inet_hashtables.c | 11 +++++++---- net/ipv4/inet_timewait_sock.c | 6 ++++-- 3 files changed, 15 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c index 045e799d3e1d..4c804b3c287b 100644 --- a/net/ipv4/inet_connection_sock.c +++ b/net/ipv4/inet_connection_sock.c @@ -103,7 +103,8 @@ int inet_csk_get_port(struct sock *sk, unsigned short snum) rover = net_random() % remaining + low; do { - head = &hashinfo->bhash[inet_bhashfn(rover, hashinfo->bhash_size)]; + head = &hashinfo->bhash[inet_bhashfn(net, rover, + hashinfo->bhash_size)]; spin_lock(&head->lock); inet_bind_bucket_for_each(tb, node, &head->chain) if (tb->ib_net == net && tb->port == rover) @@ -130,7 +131,8 @@ int inet_csk_get_port(struct sock *sk, unsigned short snum) */ snum = rover; } else { - head = &hashinfo->bhash[inet_bhashfn(snum, hashinfo->bhash_size)]; + head = &hashinfo->bhash[inet_bhashfn(net, snum, + hashinfo->bhash_size)]; spin_lock(&head->lock); inet_bind_bucket_for_each(tb, node, &head->chain) if (tb->ib_net == net && tb->port == snum) diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index 2023d37b2708..dc1b78de8990 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -70,7 +70,8 @@ void inet_bind_hash(struct sock *sk, struct inet_bind_bucket *tb, static void __inet_put_port(struct sock *sk) { struct inet_hashinfo *hashinfo = sk->sk_prot->h.hashinfo; - const int bhash = inet_bhashfn(inet_sk(sk)->num, hashinfo->bhash_size); + const int bhash = inet_bhashfn(sock_net(sk), inet_sk(sk)->num, + hashinfo->bhash_size); struct inet_bind_hashbucket *head = &hashinfo->bhash[bhash]; struct inet_bind_bucket *tb; @@ -95,7 +96,8 @@ EXPORT_SYMBOL(inet_put_port); void __inet_inherit_port(struct sock *sk, struct sock *child) { struct inet_hashinfo *table = sk->sk_prot->h.hashinfo; - const int bhash = inet_bhashfn(inet_sk(child)->num, table->bhash_size); + const int bhash = inet_bhashfn(sock_net(sk), inet_sk(child)->num, + table->bhash_size); struct inet_bind_hashbucket *head = &table->bhash[bhash]; struct inet_bind_bucket *tb; @@ -438,7 +440,8 @@ int __inet_hash_connect(struct inet_timewait_death_row *death_row, local_bh_disable(); for (i = 1; i <= remaining; i++) { port = low + (i + offset) % remaining; - head = &hinfo->bhash[inet_bhashfn(port, hinfo->bhash_size)]; + head = &hinfo->bhash[inet_bhashfn(net, port, + hinfo->bhash_size)]; spin_lock(&head->lock); /* Does not bother with rcv_saddr checks, @@ -493,7 +496,7 @@ ok: goto out; } - head = &hinfo->bhash[inet_bhashfn(snum, hinfo->bhash_size)]; + head = &hinfo->bhash[inet_bhashfn(net, snum, hinfo->bhash_size)]; tb = inet_csk(sk)->icsk_bind_hash; spin_lock_bh(&head->lock); if (sk_head(&tb->owners) == sk && !sk->sk_bind_node.next) { diff --git a/net/ipv4/inet_timewait_sock.c b/net/ipv4/inet_timewait_sock.c index ce16e9ac24c1..06006a5ac8b9 100644 --- a/net/ipv4/inet_timewait_sock.c +++ b/net/ipv4/inet_timewait_sock.c @@ -32,7 +32,8 @@ static void __inet_twsk_kill(struct inet_timewait_sock *tw, write_unlock(lock); /* Disassociate with bind bucket. */ - bhead = &hashinfo->bhash[inet_bhashfn(tw->tw_num, hashinfo->bhash_size)]; + bhead = &hashinfo->bhash[inet_bhashfn(twsk_net(tw), tw->tw_num, + hashinfo->bhash_size)]; spin_lock(&bhead->lock); tb = tw->tw_tb; __hlist_del(&tw->tw_bind_node); @@ -81,7 +82,8 @@ void __inet_twsk_hashdance(struct inet_timewait_sock *tw, struct sock *sk, Note, that any socket with inet->num != 0 MUST be bound in binding cache, even if it is closed. */ - bhead = &hashinfo->bhash[inet_bhashfn(inet->num, hashinfo->bhash_size)]; + bhead = &hashinfo->bhash[inet_bhashfn(twsk_net(tw), inet->num, + hashinfo->bhash_size)]; spin_lock(&bhead->lock); tw->tw_tb = icsk->icsk_bind_hash; BUG_TRAP(icsk->icsk_bind_hash); -- cgit v1.2.2 From 2086a65078bd24682bdcf413d9c91d81988b8359 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 16 Jun 2008 17:13:08 -0700 Subject: inet: add struct net argument to inet_lhashfn Listening-on-one-port sockets in many namespaces produce long chains in the listening_hash-es, so prepare the inet_lhashfn to take struct net into account. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/inet_hashtables.c | 2 +- net/ipv6/inet6_hashtables.c | 3 ++- 2 files changed, 3 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index dc1b78de8990..4f597b3175e7 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -194,7 +194,7 @@ struct sock *__inet_lookup_listener(struct net *net, const struct hlist_head *head; read_lock(&hashinfo->lhash_lock); - head = &hashinfo->listening_hash[inet_lhashfn(hnum)]; + head = &hashinfo->listening_hash[inet_lhashfn(net, hnum)]; if (!hlist_empty(head)) { const struct inet_sock *inet = inet_sk((sk = __sk_head(head))); diff --git a/net/ipv6/inet6_hashtables.c b/net/ipv6/inet6_hashtables.c index 580014aea4d6..b940156ca4f5 100644 --- a/net/ipv6/inet6_hashtables.c +++ b/net/ipv6/inet6_hashtables.c @@ -104,7 +104,8 @@ struct sock *inet6_lookup_listener(struct net *net, int score, hiscore = 0; read_lock(&hashinfo->lhash_lock); - sk_for_each(sk, node, &hashinfo->listening_hash[inet_lhashfn(hnum)]) { + sk_for_each(sk, node, + &hashinfo->listening_hash[inet_lhashfn(net, hnum)]) { if (net_eq(sock_net(sk), net) && inet_sk(sk)->num == hnum && sk->sk_family == PF_INET6) { const struct ipv6_pinfo *np = inet6_sk(sk); -- cgit v1.2.2 From 9f26b3add3783c0af869ea2207871da5dafefffa Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 16 Jun 2008 17:13:27 -0700 Subject: inet: add struct net argument to inet_ehashfn Although this hash takes addresses into account, the ehash chains can also be too long when, for instance, communications via lo occur. So, prepare the inet_hashfn to take struct net into account. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/inet_hashtables.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index 4f597b3175e7..eca5899729e3 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -227,7 +227,7 @@ struct sock * __inet_lookup_established(struct net *net, /* Optimize here for direct hit, only listening connections can * have wildcards anyways. */ - unsigned int hash = inet_ehashfn(daddr, hnum, saddr, sport); + unsigned int hash = inet_ehashfn(net, daddr, hnum, saddr, sport); struct inet_ehash_bucket *head = inet_ehash_bucket(hashinfo, hash); rwlock_t *lock = inet_ehash_lockp(hashinfo, hash); @@ -267,13 +267,13 @@ static int __inet_check_established(struct inet_timewait_death_row *death_row, int dif = sk->sk_bound_dev_if; INET_ADDR_COOKIE(acookie, saddr, daddr) const __portpair ports = INET_COMBINED_PORTS(inet->dport, lport); - unsigned int hash = inet_ehashfn(daddr, lport, saddr, inet->dport); + struct net *net = sock_net(sk); + unsigned int hash = inet_ehashfn(net, daddr, lport, saddr, inet->dport); struct inet_ehash_bucket *head = inet_ehash_bucket(hinfo, hash); rwlock_t *lock = inet_ehash_lockp(hinfo, hash); struct sock *sk2; const struct hlist_node *node; struct inet_timewait_sock *tw; - struct net *net = sock_net(sk); prefetch(head->chain.first); write_lock(lock); -- cgit v1.2.2 From 33de014c63646f69f36f3673e3b4676f931dc878 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 16 Jun 2008 17:13:48 -0700 Subject: inet6: add struct net argument to inet6_ehashfn Same as for inet_hashfn, prepare its ipv6 incarnation. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv6/inet6_hashtables.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv6/inet6_hashtables.c b/net/ipv6/inet6_hashtables.c index b940156ca4f5..a9cc8ab33a49 100644 --- a/net/ipv6/inet6_hashtables.c +++ b/net/ipv6/inet6_hashtables.c @@ -68,7 +68,7 @@ struct sock *__inet6_lookup_established(struct net *net, /* Optimize here for direct hit, only listening connections can * have wildcards anyways. */ - unsigned int hash = inet6_ehashfn(daddr, hnum, saddr, sport); + unsigned int hash = inet6_ehashfn(net, daddr, hnum, saddr, sport); struct inet_ehash_bucket *head = inet_ehash_bucket(hashinfo, hash); rwlock_t *lock = inet_ehash_lockp(hashinfo, hash); @@ -166,14 +166,14 @@ static int __inet6_check_established(struct inet_timewait_death_row *death_row, const struct in6_addr *saddr = &np->daddr; const int dif = sk->sk_bound_dev_if; const __portpair ports = INET_COMBINED_PORTS(inet->dport, lport); - const unsigned int hash = inet6_ehashfn(daddr, lport, saddr, + struct net *net = sock_net(sk); + const unsigned int hash = inet6_ehashfn(net, daddr, lport, saddr, inet->dport); struct inet_ehash_bucket *head = inet_ehash_bucket(hinfo, hash); rwlock_t *lock = inet_ehash_lockp(hinfo, hash); struct sock *sk2; const struct hlist_node *node; struct inet_timewait_sock *tw; - struct net *net = sock_net(sk); prefetch(head->chain.first); write_lock(lock); -- cgit v1.2.2 From 25519a2a769d42fc2733a8f119682272d99b1304 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 21 Dec 2007 03:22:38 -0800 Subject: wext: Remove inline from get_priv_size() and adjust_priv_size(). The compiler inlines when appropriate. Signed-off-by: David S. Miller --- net/wireless/wext.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/wireless/wext.c b/net/wireless/wext.c index 947188a5b937..e83d74affd72 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -500,7 +500,7 @@ static int call_commit_handler(struct net_device *dev) /* * Calculate size of private arguments */ -static inline int get_priv_size(__u16 args) +static int get_priv_size(__u16 args) { int num = args & IW_PRIV_SIZE_MASK; int type = (args & IW_PRIV_TYPE_MASK) >> 12; @@ -512,8 +512,7 @@ static inline int get_priv_size(__u16 args) /* * Re-calculate the size of private arguments */ -static inline int adjust_priv_size(__u16 args, - union iwreq_data * wrqu) +static int adjust_priv_size(__u16 args, union iwreq_data *wrqu) { int num = wrqu->data.length; int max = args & IW_PRIV_SIZE_MASK; -- cgit v1.2.2 From 208887d4cc5a5c1eeb68bd170e21e32b1129cd94 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 21 Dec 2007 03:24:24 -0800 Subject: wext: Make adjust_priv_size() take a "struct iw_point *". Signed-off-by: David S. Miller --- net/wireless/wext.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/wireless/wext.c b/net/wireless/wext.c index e83d74affd72..cd2cf9fec10f 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -512,9 +512,9 @@ static int get_priv_size(__u16 args) /* * Re-calculate the size of private arguments */ -static int adjust_priv_size(__u16 args, union iwreq_data *wrqu) +static int adjust_priv_size(__u16 args, struct iw_point *iwp) { - int num = wrqu->data.length; + int num = iwp->length; int max = args & IW_PRIV_SIZE_MASK; int type = (args & IW_PRIV_TYPE_MASK) >> 12; @@ -976,7 +976,7 @@ static int ioctl_private_call(struct net_device *dev, struct ifreq *ifr, * avoid leaking kernel bits outside. */ if (!(descr->get_args & IW_PRIV_SIZE_FIXED)) { extra_size = adjust_priv_size(descr->get_args, - &(iwr->u)); + &(iwr->u.data)); } err = copy_to_user(iwr->u.data.pointer, extra, -- cgit v1.2.2 From 84149b0fca08f9ec554dfc28dabc39839fdf8a06 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 21 Dec 2007 03:27:17 -0800 Subject: wext: Extract standard call iw_point handling into seperate function. Signed-off-by: David S. Miller --- net/wireless/wext.c | 258 +++++++++++++++++++++++++++------------------------- 1 file changed, 134 insertions(+), 124 deletions(-) (limited to 'net') diff --git a/net/wireless/wext.c b/net/wireless/wext.c index cd2cf9fec10f..d17c0f45accb 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -694,6 +694,138 @@ void wext_proc_exit(struct net *net) */ /* ---------------------------------------------------------------- */ +static int ioctl_standard_iw_point(struct iw_point *iwp, unsigned int cmd, + const struct iw_ioctl_description *descr, + iw_handler handler, struct net_device *dev, + struct iw_request_info *info) +{ + int err, extra_size, user_length = 0, essid_compat = 0; + char *extra; + + /* Calculate space needed by arguments. Always allocate + * for max space. + */ + extra_size = descr->max_tokens * descr->token_size; + + /* Check need for ESSID compatibility for WE < 21 */ + switch (cmd) { + case SIOCSIWESSID: + case SIOCGIWESSID: + case SIOCSIWNICKN: + case SIOCGIWNICKN: + if (iwp->length == descr->max_tokens + 1) + essid_compat = 1; + else if (IW_IS_SET(cmd) && (iwp->length != 0)) { + char essid[IW_ESSID_MAX_SIZE + 1]; + + err = copy_from_user(essid, iwp->pointer, + iwp->length * + descr->token_size); + if (err) + return -EFAULT; + + if (essid[iwp->length - 1] == '\0') + essid_compat = 1; + } + break; + default: + break; + } + + iwp->length -= essid_compat; + + /* Check what user space is giving us */ + if (IW_IS_SET(cmd)) { + /* Check NULL pointer */ + if (!iwp->pointer && iwp->length != 0) + return -EFAULT; + /* Check if number of token fits within bounds */ + if (iwp->length > descr->max_tokens) + return -E2BIG; + if (iwp->length < descr->min_tokens) + return -EINVAL; + } else { + /* Check NULL pointer */ + if (!iwp->pointer) + return -EFAULT; + /* Save user space buffer size for checking */ + user_length = iwp->length; + + /* Don't check if user_length > max to allow forward + * compatibility. The test user_length < min is + * implied by the test at the end. + */ + + /* Support for very large requests */ + if ((descr->flags & IW_DESCR_FLAG_NOMAX) && + (user_length > descr->max_tokens)) { + /* Allow userspace to GET more than max so + * we can support any size GET requests. + * There is still a limit : -ENOMEM. + */ + extra_size = user_length * descr->token_size; + + /* Note : user_length is originally a __u16, + * and token_size is controlled by us, + * so extra_size won't get negative and + * won't overflow... + */ + } + } + + /* kzalloc() ensures NULL-termination for essid_compat. */ + extra = kzalloc(extra_size, GFP_KERNEL); + if (!extra) + return -ENOMEM; + + /* If it is a SET, get all the extra data in here */ + if (IW_IS_SET(cmd) && (iwp->length != 0)) { + if (copy_from_user(extra, iwp->pointer, + iwp->length * + descr->token_size)) { + err = -EFAULT; + goto out; + } + } + + err = handler(dev, info, (union iwreq_data *) iwp, extra); + + iwp->length += essid_compat; + + /* If we have something to return to the user */ + if (!err && IW_IS_GET(cmd)) { + /* Check if there is enough buffer up there */ + if (user_length < iwp->length) { + err = -E2BIG; + goto out; + } + + if (copy_to_user(iwp->pointer, extra, + iwp->length * + descr->token_size)) { + err = -EFAULT; + goto out; + } + } + + /* Generate an event to notify listeners of the change */ + if ((descr->flags & IW_DESCR_FLAG_EVENT) && err == -EIWCOMMIT) { + union iwreq_data *data = (union iwreq_data *) iwp; + + if (descr->flags & IW_DESCR_FLAG_RESTRICT) + /* If the event is restricted, don't + * export the payload. + */ + wireless_send_event(dev, cmd, data, NULL); + else + wireless_send_event(dev, cmd, data, extra); + } + +out: + kfree(extra); + return err; +} + /* * Wrapper to call a standard Wireless Extension handler. * We do various checks and also take care of moving data between @@ -729,130 +861,8 @@ static int ioctl_standard_call(struct net_device * dev, ((ret == 0) || (ret == -EIWCOMMIT))) wireless_send_event(dev, cmd, &(iwr->u), NULL); } else { - char * extra; - int extra_size; - int user_length = 0; - int err; - int essid_compat = 0; - - /* Calculate space needed by arguments. Always allocate - * for max space. Easier, and won't last long... */ - extra_size = descr->max_tokens * descr->token_size; - - /* Check need for ESSID compatibility for WE < 21 */ - switch (cmd) { - case SIOCSIWESSID: - case SIOCGIWESSID: - case SIOCSIWNICKN: - case SIOCGIWNICKN: - if (iwr->u.data.length == descr->max_tokens + 1) - essid_compat = 1; - else if (IW_IS_SET(cmd) && (iwr->u.data.length != 0)) { - char essid[IW_ESSID_MAX_SIZE + 1]; - - err = copy_from_user(essid, iwr->u.data.pointer, - iwr->u.data.length * - descr->token_size); - if (err) - return -EFAULT; - - if (essid[iwr->u.data.length - 1] == '\0') - essid_compat = 1; - } - break; - default: - break; - } - - iwr->u.data.length -= essid_compat; - - /* Check what user space is giving us */ - if (IW_IS_SET(cmd)) { - /* Check NULL pointer */ - if ((iwr->u.data.pointer == NULL) && - (iwr->u.data.length != 0)) - return -EFAULT; - /* Check if number of token fits within bounds */ - if (iwr->u.data.length > descr->max_tokens) - return -E2BIG; - if (iwr->u.data.length < descr->min_tokens) - return -EINVAL; - } else { - /* Check NULL pointer */ - if (iwr->u.data.pointer == NULL) - return -EFAULT; - /* Save user space buffer size for checking */ - user_length = iwr->u.data.length; - - /* Don't check if user_length > max to allow forward - * compatibility. The test user_length < min is - * implied by the test at the end. */ - - /* Support for very large requests */ - if ((descr->flags & IW_DESCR_FLAG_NOMAX) && - (user_length > descr->max_tokens)) { - /* Allow userspace to GET more than max so - * we can support any size GET requests. - * There is still a limit : -ENOMEM. */ - extra_size = user_length * descr->token_size; - /* Note : user_length is originally a __u16, - * and token_size is controlled by us, - * so extra_size won't get negative and - * won't overflow... */ - } - } - - /* Create the kernel buffer */ - /* kzalloc ensures NULL-termination for essid_compat */ - extra = kzalloc(extra_size, GFP_KERNEL); - if (extra == NULL) - return -ENOMEM; - - /* If it is a SET, get all the extra data in here */ - if (IW_IS_SET(cmd) && (iwr->u.data.length != 0)) { - err = copy_from_user(extra, iwr->u.data.pointer, - iwr->u.data.length * - descr->token_size); - if (err) { - kfree(extra); - return -EFAULT; - } - } - - /* Call the handler */ - ret = handler(dev, &info, &(iwr->u), extra); - - iwr->u.data.length += essid_compat; - - /* If we have something to return to the user */ - if (!ret && IW_IS_GET(cmd)) { - /* Check if there is enough buffer up there */ - if (user_length < iwr->u.data.length) { - kfree(extra); - return -E2BIG; - } - - err = copy_to_user(iwr->u.data.pointer, extra, - iwr->u.data.length * - descr->token_size); - if (err) - ret = -EFAULT; - } - - /* Generate an event to notify listeners of the change */ - if ((descr->flags & IW_DESCR_FLAG_EVENT) && - ((ret == 0) || (ret == -EIWCOMMIT))) { - if (descr->flags & IW_DESCR_FLAG_RESTRICT) - /* If the event is restricted, don't - * export the payload */ - wireless_send_event(dev, cmd, &(iwr->u), NULL); - else - wireless_send_event(dev, cmd, &(iwr->u), - extra); - } - - /* Cleanup - I told you it wasn't that long ;-) */ - kfree(extra); + ret = ioctl_standard_iw_point(&iwr->u.data, cmd, descr, + handler, dev, &info); } /* Call commit handler if needed and defined */ -- cgit v1.2.2 From d88174e4d295f0880e5f9cb6d42f26b0367c8fd9 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 21 Dec 2007 03:33:46 -0800 Subject: wext: Extract private call iw_point handling into seperate functions. Signed-off-by: David S. Miller --- net/wireless/wext.c | 141 +++++++++++++++++++++++++++------------------------- 1 file changed, 74 insertions(+), 67 deletions(-) (limited to 'net') diff --git a/net/wireless/wext.c b/net/wireless/wext.c index d17c0f45accb..a1cd19add6d8 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -890,25 +890,22 @@ static int ioctl_standard_call(struct net_device * dev, * a iw_handler but process it in your ioctl handler (i.e. use the * old driver API). */ -static int ioctl_private_call(struct net_device *dev, struct ifreq *ifr, - unsigned int cmd, iw_handler handler) +static int get_priv_descr_and_size(struct net_device *dev, unsigned int cmd, + const struct iw_priv_args **descrp) { - struct iwreq * iwr = (struct iwreq *) ifr; - const struct iw_priv_args * descr = NULL; - struct iw_request_info info; - int extra_size = 0; - int i; - int ret = -EINVAL; + const struct iw_priv_args *descr; + int i, extra_size; - /* Get the description of the IOCTL */ - for (i = 0; i < dev->wireless_handlers->num_private_args; i++) + descr = NULL; + for (i = 0; i < dev->wireless_handlers->num_private_args; i++) { if (cmd == dev->wireless_handlers->private_args[i].cmd) { - descr = &(dev->wireless_handlers->private_args[i]); + descr = &dev->wireless_handlers->private_args[i]; break; } + } - /* Compute the size of the set/get arguments */ - if (descr != NULL) { + extra_size = 0; + if (descr) { if (IW_IS_SET(cmd)) { int offset = 0; /* For sub-ioctls */ /* Check for sub-ioctl handler */ @@ -933,72 +930,82 @@ static int ioctl_private_call(struct net_device *dev, struct ifreq *ifr, extra_size = 0; } } + *descrp = descr; + return extra_size; +} - /* Prepare the call */ - info.cmd = cmd; - info.flags = 0; +static int ioctl_private_iw_point(struct iw_point *iwp, unsigned int cmd, + const struct iw_priv_args *descr, + iw_handler handler, struct net_device *dev, + struct iw_request_info *info, int extra_size) +{ + char *extra; + int err; - /* Check if we have a pointer to user space data or not. */ - if (extra_size == 0) { - /* No extra arguments. Trivial to handle */ - ret = handler(dev, &info, &(iwr->u), (char *) &(iwr->u)); - } else { - char * extra; - int err; + /* Check what user space is giving us */ + if (IW_IS_SET(cmd)) { + if (!iwp->pointer && iwp->length != 0) + return -EFAULT; - /* Check what user space is giving us */ - if (IW_IS_SET(cmd)) { - /* Check NULL pointer */ - if ((iwr->u.data.pointer == NULL) && - (iwr->u.data.length != 0)) - return -EFAULT; + if (iwp->length > (descr->set_args & IW_PRIV_SIZE_MASK)) + return -E2BIG; + } else if (!iwp->pointer) + return -EFAULT; - /* Does it fits within bounds ? */ - if (iwr->u.data.length > (descr->set_args & - IW_PRIV_SIZE_MASK)) - return -E2BIG; - } else if (iwr->u.data.pointer == NULL) - return -EFAULT; + extra = kmalloc(extra_size, GFP_KERNEL); + if (!extra) + return -ENOMEM; - /* Always allocate for max space. Easier, and won't last - * long... */ - extra = kmalloc(extra_size, GFP_KERNEL); - if (extra == NULL) - return -ENOMEM; - - /* If it is a SET, get all the extra data in here */ - if (IW_IS_SET(cmd) && (iwr->u.data.length != 0)) { - err = copy_from_user(extra, iwr->u.data.pointer, - extra_size); - if (err) { - kfree(extra); - return -EFAULT; - } + /* If it is a SET, get all the extra data in here */ + if (IW_IS_SET(cmd) && (iwp->length != 0)) { + if (copy_from_user(extra, iwp->pointer, extra_size)) { + err = -EFAULT; + goto out; } + } - /* Call the handler */ - ret = handler(dev, &info, &(iwr->u), extra); + /* Call the handler */ + err = handler(dev, info, (union iwreq_data *) iwp, extra); - /* If we have something to return to the user */ - if (!ret && IW_IS_GET(cmd)) { + /* If we have something to return to the user */ + if (!err && IW_IS_GET(cmd)) { + /* Adjust for the actual length if it's variable, + * avoid leaking kernel bits outside. + */ + if (!(descr->get_args & IW_PRIV_SIZE_FIXED)) + extra_size = adjust_priv_size(descr->get_args, iwp); - /* Adjust for the actual length if it's variable, - * avoid leaking kernel bits outside. */ - if (!(descr->get_args & IW_PRIV_SIZE_FIXED)) { - extra_size = adjust_priv_size(descr->get_args, - &(iwr->u.data)); - } + if (copy_to_user(iwp->pointer, extra, extra_size)) + err = -EFAULT; + } - err = copy_to_user(iwr->u.data.pointer, extra, - extra_size); - if (err) - ret = -EFAULT; - } +out: + kfree(extra); + return err; +} - /* Cleanup - I told you it wasn't that long ;-) */ - kfree(extra); - } +static int ioctl_private_call(struct net_device *dev, struct ifreq *ifr, + unsigned int cmd, iw_handler handler) +{ + struct iwreq *iwr = (struct iwreq *) ifr; + int extra_size = 0, ret = -EINVAL; + const struct iw_priv_args *descr; + struct iw_request_info info; + extra_size = get_priv_descr_and_size(dev, cmd, &descr); + + /* Prepare the call */ + info.cmd = cmd; + info.flags = 0; + + /* Check if we have a pointer to user space data or not. */ + if (extra_size == 0) { + /* No extra arguments. Trivial to handle */ + ret = handler(dev, &info, &(iwr->u), (char *) &(iwr->u)); + } else { + ret = ioctl_private_iw_point(&iwr->u.data, cmd, descr, + handler, dev, &info, extra_size); + } /* Call commit handler if needed and defined */ if (ret == -EIWCOMMIT) -- cgit v1.2.2 From 67dd7608078b17f63f29ff2108fc5bf2407ddcec Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 21 Dec 2007 03:36:31 -0800 Subject: wext: Pull ioctl permission checking out into helper function. Signed-off-by: David S. Miller --- net/wireless/wext.c | 22 +++++++++++++++------- 1 file changed, 15 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/wireless/wext.c b/net/wireless/wext.c index a1cd19add6d8..e96559ea0fb3 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -1061,18 +1061,26 @@ static int wireless_process_ioctl(struct net *net, struct ifreq *ifr, unsigned i return -EOPNOTSUPP; } +/* If command is `set a parameter', or `get the encoding parameters', + * check if the user has the right to do it. + */ +static int wext_permission_check(unsigned int cmd) +{ + if ((IW_IS_SET(cmd) || cmd == SIOCGIWENCODE || cmd == SIOCGIWENCODEEXT) + && !capable(CAP_NET_ADMIN)) + return -EPERM; + + return 0; +} + /* entry point from dev ioctl */ int wext_handle_ioctl(struct net *net, struct ifreq *ifr, unsigned int cmd, void __user *arg) { - int ret; + int ret = wext_permission_check(cmd); - /* If command is `set a parameter', or - * `get the encoding parameters', check if - * the user has the right to do it */ - if ((IW_IS_SET(cmd) || cmd == SIOCGIWENCODE || cmd == SIOCGIWENCODEEXT) - && !capable(CAP_NET_ADMIN)) - return -EPERM; + if (ret) + return ret; dev_load(net, ifr->ifr_name); rtnl_lock(); -- cgit v1.2.2 From ca1e8bb8e4e89e2769e2b39eb29fdcfc5c19cf89 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 21 Dec 2007 03:41:45 -0800 Subject: wext: Parameterize the standard/private handlers. The WEXT standard and private handlers to use are now arguments to wireless_process_ioctl(). Signed-off-by: David S. Miller --- net/wireless/wext.c | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/wireless/wext.c b/net/wireless/wext.c index e96559ea0fb3..929d24120f37 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -1015,11 +1015,17 @@ static int ioctl_private_call(struct net_device *dev, struct ifreq *ifr, } /* ---------------------------------------------------------------- */ +typedef int (*wext_ioctl_func)(struct net_device *, struct ifreq *, + unsigned int, iw_handler); + /* * Main IOCTl dispatcher. * Check the type of IOCTL and call the appropriate wrapper... */ -static int wireless_process_ioctl(struct net *net, struct ifreq *ifr, unsigned int cmd) +static int wireless_process_ioctl(struct net *net, struct ifreq *ifr, + unsigned int cmd, + wext_ioctl_func standard, + wext_ioctl_func private) { struct net_device *dev; iw_handler handler; @@ -1035,12 +1041,12 @@ static int wireless_process_ioctl(struct net *net, struct ifreq *ifr, unsigned i * Note that 'cmd' is already filtered in dev_ioctl() with * (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) */ if (cmd == SIOCGIWSTATS) - return ioctl_standard_call(dev, ifr, cmd, - &iw_handler_get_iwstats); + return standard(dev, ifr, cmd, + &iw_handler_get_iwstats); if (cmd == SIOCGIWPRIV && dev->wireless_handlers) - return ioctl_standard_call(dev, ifr, cmd, - &iw_handler_get_private); + return standard(dev, ifr, cmd, + &iw_handler_get_private); /* Basic check */ if (!netif_device_present(dev)) @@ -1051,9 +1057,9 @@ static int wireless_process_ioctl(struct net *net, struct ifreq *ifr, unsigned i if (handler) { /* Standard and private are not the same */ if (cmd < SIOCIWFIRSTPRIV) - return ioctl_standard_call(dev, ifr, cmd, handler); + return standard(dev, ifr, cmd, handler); else - return ioctl_private_call(dev, ifr, cmd, handler); + return private(dev, ifr, cmd, handler); } /* Old driver API : call driver ioctl handler */ if (dev->do_ioctl) @@ -1084,7 +1090,9 @@ int wext_handle_ioctl(struct net *net, struct ifreq *ifr, unsigned int cmd, dev_load(net, ifr->ifr_name); rtnl_lock(); - ret = wireless_process_ioctl(net, ifr, cmd); + ret = wireless_process_ioctl(net, ifr, cmd, + ioctl_standard_call, + ioctl_private_call); rtnl_unlock(); if (IW_IS_GET(cmd) && copy_to_user(arg, ifr, sizeof(struct iwreq))) return -EFAULT; -- cgit v1.2.2 From d2911255590d9ca561a481b9dbebcfcbbf38fa4e Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 21 Dec 2007 03:46:01 -0800 Subject: wext: Pass iwreq pointer down into standard/private handlers. They have no need to see the object as an ifreq. Signed-off-by: David S. Miller --- net/wireless/wext.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/wireless/wext.c b/net/wireless/wext.c index 929d24120f37..e9c88172ec55 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -832,11 +832,10 @@ out: * user space and kernel space. */ static int ioctl_standard_call(struct net_device * dev, - struct ifreq * ifr, + struct iwreq *iwr, unsigned int cmd, iw_handler handler) { - struct iwreq * iwr = (struct iwreq *) ifr; const struct iw_ioctl_description * descr; struct iw_request_info info; int ret = -EINVAL; @@ -984,10 +983,9 @@ out: return err; } -static int ioctl_private_call(struct net_device *dev, struct ifreq *ifr, +static int ioctl_private_call(struct net_device *dev, struct iwreq *iwr, unsigned int cmd, iw_handler handler) { - struct iwreq *iwr = (struct iwreq *) ifr; int extra_size = 0, ret = -EINVAL; const struct iw_priv_args *descr; struct iw_request_info info; @@ -1015,7 +1013,7 @@ static int ioctl_private_call(struct net_device *dev, struct ifreq *ifr, } /* ---------------------------------------------------------------- */ -typedef int (*wext_ioctl_func)(struct net_device *, struct ifreq *, +typedef int (*wext_ioctl_func)(struct net_device *, struct iwreq *, unsigned int, iw_handler); /* @@ -1027,6 +1025,7 @@ static int wireless_process_ioctl(struct net *net, struct ifreq *ifr, wext_ioctl_func standard, wext_ioctl_func private) { + struct iwreq *iwr = (struct iwreq *) ifr; struct net_device *dev; iw_handler handler; @@ -1041,11 +1040,11 @@ static int wireless_process_ioctl(struct net *net, struct ifreq *ifr, * Note that 'cmd' is already filtered in dev_ioctl() with * (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) */ if (cmd == SIOCGIWSTATS) - return standard(dev, ifr, cmd, + return standard(dev, iwr, cmd, &iw_handler_get_iwstats); if (cmd == SIOCGIWPRIV && dev->wireless_handlers) - return standard(dev, ifr, cmd, + return standard(dev, iwr, cmd, &iw_handler_get_private); /* Basic check */ @@ -1057,9 +1056,9 @@ static int wireless_process_ioctl(struct net *net, struct ifreq *ifr, if (handler) { /* Standard and private are not the same */ if (cmd < SIOCIWFIRSTPRIV) - return standard(dev, ifr, cmd, handler); + return standard(dev, iwr, cmd, handler); else - return private(dev, ifr, cmd, handler); + return private(dev, iwr, cmd, handler); } /* Old driver API : call driver ioctl handler */ if (dev->do_ioctl) -- cgit v1.2.2 From a67fa76d8be4e24e2d61cd76438a893d4c2886f7 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 3 Jun 2008 07:36:30 -0700 Subject: wext: Pull top-level ioctl dispatch logic into helper function. Signed-off-by: David S. Miller --- net/wireless/wext.c | 26 ++++++++++++++++++++------ 1 file changed, 20 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/wireless/wext.c b/net/wireless/wext.c index e9c88172ec55..09022cbb58ba 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -1079,8 +1079,10 @@ static int wext_permission_check(unsigned int cmd) } /* entry point from dev ioctl */ -int wext_handle_ioctl(struct net *net, struct ifreq *ifr, unsigned int cmd, - void __user *arg) +static int wext_ioctl_dispatch(struct net *net, struct ifreq *ifr, + unsigned int cmd, + wext_ioctl_func standard, + wext_ioctl_func private) { int ret = wext_permission_check(cmd); @@ -1089,12 +1091,24 @@ int wext_handle_ioctl(struct net *net, struct ifreq *ifr, unsigned int cmd, dev_load(net, ifr->ifr_name); rtnl_lock(); - ret = wireless_process_ioctl(net, ifr, cmd, - ioctl_standard_call, - ioctl_private_call); + ret = wireless_process_ioctl(net, ifr, cmd, standard, private); rtnl_unlock(); - if (IW_IS_GET(cmd) && copy_to_user(arg, ifr, sizeof(struct iwreq))) + + return ret; +} + +int wext_handle_ioctl(struct net *net, struct ifreq *ifr, unsigned int cmd, + void __user *arg) +{ + int ret = wext_ioctl_dispatch(net, ifr, cmd, + ioctl_standard_call, + ioctl_private_call); + + if (ret >= 0 && + IW_IS_GET(cmd) && + copy_to_user(arg, ifr, sizeof(struct iwreq))) return -EFAULT; + return ret; } -- cgit v1.2.2 From 87de87d5e47f94b4ea647a5bd1bc8dc1f7930db4 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 3 Jun 2008 09:14:03 -0700 Subject: wext: Dispatch and handle compat ioctls entirely in net/wireless/wext.c Next we can kill the hacks in fs/compat_ioctl.c and also dispatch compat ioctls down into the driver and 80211 protocol helper layers in order to handle iw_point objects embedded in stream replies which need to be translated. Signed-off-by: David S. Miller --- net/socket.c | 10 +++++ net/wireless/wext.c | 104 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 114 insertions(+) (limited to 'net') diff --git a/net/socket.c b/net/socket.c index 66c4a8cf6db9..81fe82513046 100644 --- a/net/socket.c +++ b/net/socket.c @@ -90,6 +90,7 @@ #include #include +#include #include #include @@ -2210,10 +2211,19 @@ static long compat_sock_ioctl(struct file *file, unsigned cmd, { struct socket *sock = file->private_data; int ret = -ENOIOCTLCMD; + struct sock *sk; + struct net *net; + + sk = sock->sk; + net = sock_net(sk); if (sock->ops->compat_ioctl) ret = sock->ops->compat_ioctl(sock, cmd, arg); + if (ret == -ENOIOCTLCMD && + (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST)) + ret = compat_wext_handle_ioctl(net, cmd, arg); + return ret; } #endif diff --git a/net/wireless/wext.c b/net/wireless/wext.c index 09022cbb58ba..1a4636a9fcde 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -1112,6 +1112,110 @@ int wext_handle_ioctl(struct net *net, struct ifreq *ifr, unsigned int cmd, return ret; } +#ifdef CONFIG_COMPAT +static int compat_standard_call(struct net_device *dev, + struct iwreq *iwr, + unsigned int cmd, + iw_handler handler) +{ + const struct iw_ioctl_description *descr; + struct compat_iw_point *iwp_compat; + struct iw_request_info info; + struct iw_point iwp; + int err; + + descr = standard_ioctl + (cmd - SIOCIWFIRST); + + if (descr->header_type != IW_HEADER_TYPE_POINT) + return ioctl_standard_call(dev, iwr, cmd, handler); + + iwp_compat = (struct compat_iw_point *) &iwr->u.data; + iwp.pointer = compat_ptr(iwp_compat->pointer); + iwp.length = iwp_compat->length; + iwp.flags = iwp_compat->flags; + + info.cmd = cmd; + info.flags = 0; + + err = ioctl_standard_iw_point(&iwp, cmd, descr, handler, dev, &info); + + iwp_compat->pointer = ptr_to_compat(iwp.pointer); + iwp_compat->length = iwp.length; + iwp_compat->flags = iwp.flags; + + return err; +} + +static int compat_private_call(struct net_device *dev, struct iwreq *iwr, + unsigned int cmd, iw_handler handler) +{ + const struct iw_priv_args *descr; + struct iw_request_info info; + int ret, extra_size; + + extra_size = get_priv_descr_and_size(dev, cmd, &descr); + + /* Prepare the call */ + info.cmd = cmd; + info.flags = 0; + + /* Check if we have a pointer to user space data or not. */ + if (extra_size == 0) { + /* No extra arguments. Trivial to handle */ + ret = handler(dev, &info, &(iwr->u), (char *) &(iwr->u)); + } else { + struct compat_iw_point *iwp_compat; + struct iw_point iwp; + + iwp_compat = (struct compat_iw_point *) &iwr->u.data; + iwp.pointer = compat_ptr(iwp_compat->pointer); + iwp.length = iwp_compat->length; + iwp.flags = iwp_compat->flags; + + ret = ioctl_private_iw_point(&iwp, cmd, descr, + handler, dev, &info, extra_size); + + iwp_compat->pointer = ptr_to_compat(iwp.pointer); + iwp_compat->length = iwp.length; + iwp_compat->flags = iwp.flags; + } + + /* Call commit handler if needed and defined */ + if (ret == -EIWCOMMIT) + ret = call_commit_handler(dev); + + return ret; +} + +int compat_wext_handle_ioctl(struct net *net, unsigned int cmd, + unsigned long arg) +{ + void __user *argp = (void __user *)arg; + struct iwreq iwr; + char *colon; + int ret; + + if (copy_from_user(&iwr, argp, sizeof(struct iwreq))) + return -EFAULT; + + iwr.ifr_name[IFNAMSIZ-1] = 0; + colon = strchr(iwr.ifr_name, ':'); + if (colon) + *colon = 0; + + ret = wext_ioctl_dispatch(net, (struct ifreq *) &iwr, cmd, + compat_standard_call, + compat_private_call); + + if (ret >= 0 && + IW_IS_GET(cmd) && + copy_to_user(argp, &iwr, sizeof(struct iwreq))) + return -EFAULT; + + return ret; +} +#endif + /************************* EVENT PROCESSING *************************/ /* * Process events generated by the wireless layer or the driver. -- cgit v1.2.2 From 0f5cabba49021d36e9f76bd97d7fa0f4a408063f Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 3 Jun 2008 07:39:16 -0700 Subject: wext: Create IW_REQUEST_FLAG_COMPAT and set it as needed. Now low-level WEXT ioctl handlers can do compat handling when necessary. Signed-off-by: David S. Miller --- net/wireless/wext.c | 73 ++++++++++++++++++++++++----------------------------- 1 file changed, 33 insertions(+), 40 deletions(-) (limited to 'net') diff --git a/net/wireless/wext.c b/net/wireless/wext.c index 1a4636a9fcde..273a84359998 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -834,10 +834,10 @@ out: static int ioctl_standard_call(struct net_device * dev, struct iwreq *iwr, unsigned int cmd, + struct iw_request_info *info, iw_handler handler) { const struct iw_ioctl_description * descr; - struct iw_request_info info; int ret = -EINVAL; /* Get the description of the IOCTL */ @@ -845,15 +845,11 @@ static int ioctl_standard_call(struct net_device * dev, return -EOPNOTSUPP; descr = &(standard_ioctl[cmd - SIOCIWFIRST]); - /* Prepare the call */ - info.cmd = cmd; - info.flags = 0; - /* Check if we have a pointer to user space data or not */ if (descr->header_type != IW_HEADER_TYPE_POINT) { /* No extra arguments. Trivial to handle */ - ret = handler(dev, &info, &(iwr->u), NULL); + ret = handler(dev, info, &(iwr->u), NULL); /* Generate an event to notify listeners of the change */ if ((descr->flags & IW_DESCR_FLAG_EVENT) && @@ -861,7 +857,7 @@ static int ioctl_standard_call(struct net_device * dev, wireless_send_event(dev, cmd, &(iwr->u), NULL); } else { ret = ioctl_standard_iw_point(&iwr->u.data, cmd, descr, - handler, dev, &info); + handler, dev, info); } /* Call commit handler if needed and defined */ @@ -984,25 +980,21 @@ out: } static int ioctl_private_call(struct net_device *dev, struct iwreq *iwr, - unsigned int cmd, iw_handler handler) + unsigned int cmd, struct iw_request_info *info, + iw_handler handler) { int extra_size = 0, ret = -EINVAL; const struct iw_priv_args *descr; - struct iw_request_info info; extra_size = get_priv_descr_and_size(dev, cmd, &descr); - /* Prepare the call */ - info.cmd = cmd; - info.flags = 0; - /* Check if we have a pointer to user space data or not. */ if (extra_size == 0) { /* No extra arguments. Trivial to handle */ - ret = handler(dev, &info, &(iwr->u), (char *) &(iwr->u)); + ret = handler(dev, info, &(iwr->u), (char *) &(iwr->u)); } else { ret = ioctl_private_iw_point(&iwr->u.data, cmd, descr, - handler, dev, &info, extra_size); + handler, dev, info, extra_size); } /* Call commit handler if needed and defined */ @@ -1014,7 +1006,8 @@ static int ioctl_private_call(struct net_device *dev, struct iwreq *iwr, /* ---------------------------------------------------------------- */ typedef int (*wext_ioctl_func)(struct net_device *, struct iwreq *, - unsigned int, iw_handler); + unsigned int, struct iw_request_info *, + iw_handler); /* * Main IOCTl dispatcher. @@ -1022,6 +1015,7 @@ typedef int (*wext_ioctl_func)(struct net_device *, struct iwreq *, */ static int wireless_process_ioctl(struct net *net, struct ifreq *ifr, unsigned int cmd, + struct iw_request_info *info, wext_ioctl_func standard, wext_ioctl_func private) { @@ -1040,11 +1034,11 @@ static int wireless_process_ioctl(struct net *net, struct ifreq *ifr, * Note that 'cmd' is already filtered in dev_ioctl() with * (cmd >= SIOCIWFIRST && cmd <= SIOCIWLAST) */ if (cmd == SIOCGIWSTATS) - return standard(dev, iwr, cmd, + return standard(dev, iwr, cmd, info, &iw_handler_get_iwstats); if (cmd == SIOCGIWPRIV && dev->wireless_handlers) - return standard(dev, iwr, cmd, + return standard(dev, iwr, cmd, info, &iw_handler_get_private); /* Basic check */ @@ -1056,9 +1050,9 @@ static int wireless_process_ioctl(struct net *net, struct ifreq *ifr, if (handler) { /* Standard and private are not the same */ if (cmd < SIOCIWFIRSTPRIV) - return standard(dev, iwr, cmd, handler); + return standard(dev, iwr, cmd, info, handler); else - return private(dev, iwr, cmd, handler); + return private(dev, iwr, cmd, info, handler); } /* Old driver API : call driver ioctl handler */ if (dev->do_ioctl) @@ -1080,7 +1074,7 @@ static int wext_permission_check(unsigned int cmd) /* entry point from dev ioctl */ static int wext_ioctl_dispatch(struct net *net, struct ifreq *ifr, - unsigned int cmd, + unsigned int cmd, struct iw_request_info *info, wext_ioctl_func standard, wext_ioctl_func private) { @@ -1091,7 +1085,7 @@ static int wext_ioctl_dispatch(struct net *net, struct ifreq *ifr, dev_load(net, ifr->ifr_name); rtnl_lock(); - ret = wireless_process_ioctl(net, ifr, cmd, standard, private); + ret = wireless_process_ioctl(net, ifr, cmd, info, standard, private); rtnl_unlock(); return ret; @@ -1100,10 +1094,12 @@ static int wext_ioctl_dispatch(struct net *net, struct ifreq *ifr, int wext_handle_ioctl(struct net *net, struct ifreq *ifr, unsigned int cmd, void __user *arg) { - int ret = wext_ioctl_dispatch(net, ifr, cmd, - ioctl_standard_call, - ioctl_private_call); + struct iw_request_info info = { .cmd = cmd, .flags = 0 }; + int ret; + ret = wext_ioctl_dispatch(net, ifr, cmd, &info, + ioctl_standard_call, + ioctl_private_call); if (ret >= 0 && IW_IS_GET(cmd) && copy_to_user(arg, ifr, sizeof(struct iwreq))) @@ -1116,28 +1112,25 @@ int wext_handle_ioctl(struct net *net, struct ifreq *ifr, unsigned int cmd, static int compat_standard_call(struct net_device *dev, struct iwreq *iwr, unsigned int cmd, + struct iw_request_info *info, iw_handler handler) { const struct iw_ioctl_description *descr; struct compat_iw_point *iwp_compat; - struct iw_request_info info; struct iw_point iwp; int err; descr = standard_ioctl + (cmd - SIOCIWFIRST); if (descr->header_type != IW_HEADER_TYPE_POINT) - return ioctl_standard_call(dev, iwr, cmd, handler); + return ioctl_standard_call(dev, iwr, cmd, info, handler); iwp_compat = (struct compat_iw_point *) &iwr->u.data; iwp.pointer = compat_ptr(iwp_compat->pointer); iwp.length = iwp_compat->length; iwp.flags = iwp_compat->flags; - info.cmd = cmd; - info.flags = 0; - - err = ioctl_standard_iw_point(&iwp, cmd, descr, handler, dev, &info); + err = ioctl_standard_iw_point(&iwp, cmd, descr, handler, dev, info); iwp_compat->pointer = ptr_to_compat(iwp.pointer); iwp_compat->length = iwp.length; @@ -1147,22 +1140,18 @@ static int compat_standard_call(struct net_device *dev, } static int compat_private_call(struct net_device *dev, struct iwreq *iwr, - unsigned int cmd, iw_handler handler) + unsigned int cmd, struct iw_request_info *info, + iw_handler handler) { const struct iw_priv_args *descr; - struct iw_request_info info; int ret, extra_size; extra_size = get_priv_descr_and_size(dev, cmd, &descr); - /* Prepare the call */ - info.cmd = cmd; - info.flags = 0; - /* Check if we have a pointer to user space data or not. */ if (extra_size == 0) { /* No extra arguments. Trivial to handle */ - ret = handler(dev, &info, &(iwr->u), (char *) &(iwr->u)); + ret = handler(dev, info, &(iwr->u), (char *) &(iwr->u)); } else { struct compat_iw_point *iwp_compat; struct iw_point iwp; @@ -1173,7 +1162,7 @@ static int compat_private_call(struct net_device *dev, struct iwreq *iwr, iwp.flags = iwp_compat->flags; ret = ioctl_private_iw_point(&iwp, cmd, descr, - handler, dev, &info, extra_size); + handler, dev, info, extra_size); iwp_compat->pointer = ptr_to_compat(iwp.pointer); iwp_compat->length = iwp.length; @@ -1191,6 +1180,7 @@ int compat_wext_handle_ioctl(struct net *net, unsigned int cmd, unsigned long arg) { void __user *argp = (void __user *)arg; + struct iw_request_info info; struct iwreq iwr; char *colon; int ret; @@ -1203,7 +1193,10 @@ int compat_wext_handle_ioctl(struct net *net, unsigned int cmd, if (colon) *colon = 0; - ret = wext_ioctl_dispatch(net, (struct ifreq *) &iwr, cmd, + info.cmd = cmd; + info.flags = IW_REQUEST_FLAG_COMPAT; + + ret = wext_ioctl_dispatch(net, (struct ifreq *) &iwr, cmd, &info, compat_standard_call, compat_private_call); -- cgit v1.2.2 From ccc580571cf0799d0460a085a7632b77753f083e Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Mon, 16 Jun 2008 18:50:49 -0700 Subject: wext: Emit event stream entries correctly when compat. Three major portions to this change: 1) Add IW_EV_COMPAT_LCP_LEN, IW_EV_COMPAT_POINT_OFF, and IW_EV_COMPAT_POINT_LEN helper defines. 2) Delete iw_stream_check_add_*(), they are unused. 3) Add iw_request_info argument to iwe_stream_add_*(), and use it to size the event and pointer lengths correctly depending upon whether IW_REQUEST_FLAG_COMPAT is set or not. 4) The mechanical transformations to the drivers and wireless stack bits to get the iw_request_info passed down into the routines modified in #3. Also, explicit references to IW_EV_LCP_LEN are replaced with iwe_stream_lcp_len(info). With a lot of help and bug fixes from Masakazu Mokuno. Signed-off-by: David S. Miller --- net/ieee80211/ieee80211_wx.c | 48 ++++++++++++++++++-------------- net/mac80211/ieee80211_i.h | 5 +++- net/mac80211/mlme.c | 66 +++++++++++++++++++++++++------------------- net/mac80211/wext.c | 2 +- 4 files changed, 71 insertions(+), 50 deletions(-) (limited to 'net') diff --git a/net/ieee80211/ieee80211_wx.c b/net/ieee80211/ieee80211_wx.c index 822606b615ca..973832dd7faf 100644 --- a/net/ieee80211/ieee80211_wx.c +++ b/net/ieee80211/ieee80211_wx.c @@ -43,8 +43,9 @@ static const char *ieee80211_modes[] = { #define MAX_CUSTOM_LEN 64 static char *ieee80211_translate_scan(struct ieee80211_device *ieee, - char *start, char *stop, - struct ieee80211_network *network) + char *start, char *stop, + struct ieee80211_network *network, + struct iw_request_info *info) { char custom[MAX_CUSTOM_LEN]; char *p; @@ -57,7 +58,7 @@ static char *ieee80211_translate_scan(struct ieee80211_device *ieee, iwe.cmd = SIOCGIWAP; iwe.u.ap_addr.sa_family = ARPHRD_ETHER; memcpy(iwe.u.ap_addr.sa_data, network->bssid, ETH_ALEN); - start = iwe_stream_add_event(start, stop, &iwe, IW_EV_ADDR_LEN); + start = iwe_stream_add_event(info, start, stop, &iwe, IW_EV_ADDR_LEN); /* Remaining entries will be displayed in the order we provide them */ @@ -66,17 +67,19 @@ static char *ieee80211_translate_scan(struct ieee80211_device *ieee, iwe.u.data.flags = 1; if (network->flags & NETWORK_EMPTY_ESSID) { iwe.u.data.length = sizeof(""); - start = iwe_stream_add_point(start, stop, &iwe, ""); + start = iwe_stream_add_point(info, start, stop, + &iwe, ""); } else { iwe.u.data.length = min(network->ssid_len, (u8) 32); - start = iwe_stream_add_point(start, stop, &iwe, network->ssid); + start = iwe_stream_add_point(info, start, stop, + &iwe, network->ssid); } /* Add the protocol name */ iwe.cmd = SIOCGIWNAME; snprintf(iwe.u.name, IFNAMSIZ, "IEEE 802.11%s", ieee80211_modes[network->mode]); - start = iwe_stream_add_event(start, stop, &iwe, IW_EV_CHAR_LEN); + start = iwe_stream_add_event(info, start, stop, &iwe, IW_EV_CHAR_LEN); /* Add mode */ iwe.cmd = SIOCGIWMODE; @@ -86,7 +89,8 @@ static char *ieee80211_translate_scan(struct ieee80211_device *ieee, else iwe.u.mode = IW_MODE_ADHOC; - start = iwe_stream_add_event(start, stop, &iwe, IW_EV_UINT_LEN); + start = iwe_stream_add_event(info, start, stop, + &iwe, IW_EV_UINT_LEN); } /* Add channel and frequency */ @@ -95,7 +99,7 @@ static char *ieee80211_translate_scan(struct ieee80211_device *ieee, iwe.u.freq.m = ieee80211_channel_to_freq(ieee, network->channel); iwe.u.freq.e = 6; iwe.u.freq.i = 0; - start = iwe_stream_add_event(start, stop, &iwe, IW_EV_FREQ_LEN); + start = iwe_stream_add_event(info, start, stop, &iwe, IW_EV_FREQ_LEN); /* Add encryption capability */ iwe.cmd = SIOCGIWENCODE; @@ -104,12 +108,13 @@ static char *ieee80211_translate_scan(struct ieee80211_device *ieee, else iwe.u.data.flags = IW_ENCODE_DISABLED; iwe.u.data.length = 0; - start = iwe_stream_add_point(start, stop, &iwe, network->ssid); + start = iwe_stream_add_point(info, start, stop, + &iwe, network->ssid); /* Add basic and extended rates */ /* Rate : stuffing multiple values in a single event require a bit * more of magic - Jean II */ - current_val = start + IW_EV_LCP_LEN; + current_val = start + iwe_stream_lcp_len(info); iwe.cmd = SIOCGIWRATE; /* Those two flags are ignored... */ iwe.u.bitrate.fixed = iwe.u.bitrate.disabled = 0; @@ -124,17 +129,19 @@ static char *ieee80211_translate_scan(struct ieee80211_device *ieee, /* Bit rate given in 500 kb/s units (+ 0x80) */ iwe.u.bitrate.value = ((rate & 0x7f) * 500000); /* Add new value to event */ - current_val = iwe_stream_add_value(start, current_val, stop, &iwe, IW_EV_PARAM_LEN); + current_val = iwe_stream_add_value(info, start, current_val, + stop, &iwe, IW_EV_PARAM_LEN); } for (; j < network->rates_ex_len; j++) { rate = network->rates_ex[j] & 0x7F; /* Bit rate given in 500 kb/s units (+ 0x80) */ iwe.u.bitrate.value = ((rate & 0x7f) * 500000); /* Add new value to event */ - current_val = iwe_stream_add_value(start, current_val, stop, &iwe, IW_EV_PARAM_LEN); + current_val = iwe_stream_add_value(info, start, current_val, + stop, &iwe, IW_EV_PARAM_LEN); } /* Check if we added any rate */ - if((current_val - start) > IW_EV_LCP_LEN) + if ((current_val - start) > iwe_stream_lcp_len(info)) start = current_val; /* Add quality statistics */ @@ -181,14 +188,14 @@ static char *ieee80211_translate_scan(struct ieee80211_device *ieee, iwe.u.qual.level = network->stats.signal; } - start = iwe_stream_add_event(start, stop, &iwe, IW_EV_QUAL_LEN); + start = iwe_stream_add_event(info, start, stop, &iwe, IW_EV_QUAL_LEN); iwe.cmd = IWEVCUSTOM; p = custom; iwe.u.data.length = p - custom; if (iwe.u.data.length) - start = iwe_stream_add_point(start, stop, &iwe, custom); + start = iwe_stream_add_point(info, start, stop, &iwe, custom); memset(&iwe, 0, sizeof(iwe)); if (network->wpa_ie_len) { @@ -196,7 +203,7 @@ static char *ieee80211_translate_scan(struct ieee80211_device *ieee, memcpy(buf, network->wpa_ie, network->wpa_ie_len); iwe.cmd = IWEVGENIE; iwe.u.data.length = network->wpa_ie_len; - start = iwe_stream_add_point(start, stop, &iwe, buf); + start = iwe_stream_add_point(info, start, stop, &iwe, buf); } memset(&iwe, 0, sizeof(iwe)); @@ -205,7 +212,7 @@ static char *ieee80211_translate_scan(struct ieee80211_device *ieee, memcpy(buf, network->rsn_ie, network->rsn_ie_len); iwe.cmd = IWEVGENIE; iwe.u.data.length = network->rsn_ie_len; - start = iwe_stream_add_point(start, stop, &iwe, buf); + start = iwe_stream_add_point(info, start, stop, &iwe, buf); } /* Add EXTRA: Age to display seconds since last beacon/probe response @@ -217,7 +224,7 @@ static char *ieee80211_translate_scan(struct ieee80211_device *ieee, jiffies_to_msecs(jiffies - network->last_scanned)); iwe.u.data.length = p - custom; if (iwe.u.data.length) - start = iwe_stream_add_point(start, stop, &iwe, custom); + start = iwe_stream_add_point(info, start, stop, &iwe, custom); /* Add spectrum management information */ iwe.cmd = -1; @@ -238,7 +245,7 @@ static char *ieee80211_translate_scan(struct ieee80211_device *ieee, if (iwe.cmd == IWEVCUSTOM) { iwe.u.data.length = p - custom; - start = iwe_stream_add_point(start, stop, &iwe, custom); + start = iwe_stream_add_point(info, start, stop, &iwe, custom); } return start; @@ -272,7 +279,8 @@ int ieee80211_wx_get_scan(struct ieee80211_device *ieee, if (ieee->scan_age == 0 || time_after(network->last_scanned + ieee->scan_age, jiffies)) - ev = ieee80211_translate_scan(ieee, ev, stop, network); + ev = ieee80211_translate_scan(ieee, ev, stop, network, + info); else IEEE80211_DEBUG_SCAN("Not showing network '%s (" "%s)' due to age (%dms).\n", diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 14fccf16b80f..80a9e7c07b47 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -24,6 +24,7 @@ #include #include #include +#include #include "key.h" #include "sta_info.h" @@ -867,7 +868,9 @@ int ieee80211_sta_set_bssid(struct net_device *dev, u8 *bssid); int ieee80211_sta_req_scan(struct net_device *dev, u8 *ssid, size_t ssid_len); void ieee80211_sta_req_auth(struct net_device *dev, struct ieee80211_if_sta *ifsta); -int ieee80211_sta_scan_results(struct net_device *dev, char *buf, size_t len); +int ieee80211_sta_scan_results(struct net_device *dev, + struct iw_request_info *info, + char *buf, size_t len); ieee80211_rx_result ieee80211_sta_rx_scan( struct net_device *dev, struct sk_buff *skb, struct ieee80211_rx_status *rx_status); diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 55659a730dc1..e06d6450f215 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -4087,6 +4087,7 @@ int ieee80211_sta_req_scan(struct net_device *dev, u8 *ssid, size_t ssid_len) static char * ieee80211_sta_scan_result(struct net_device *dev, + struct iw_request_info *info, struct ieee80211_sta_bss *bss, char *current_ev, char *end_buf) { @@ -4101,7 +4102,7 @@ ieee80211_sta_scan_result(struct net_device *dev, iwe.cmd = SIOCGIWAP; iwe.u.ap_addr.sa_family = ARPHRD_ETHER; memcpy(iwe.u.ap_addr.sa_data, bss->bssid, ETH_ALEN); - current_ev = iwe_stream_add_event(current_ev, end_buf, &iwe, + current_ev = iwe_stream_add_event(info, current_ev, end_buf, &iwe, IW_EV_ADDR_LEN); memset(&iwe, 0, sizeof(iwe)); @@ -4109,13 +4110,13 @@ ieee80211_sta_scan_result(struct net_device *dev, if (bss_mesh_cfg(bss)) { iwe.u.data.length = bss_mesh_id_len(bss); iwe.u.data.flags = 1; - current_ev = iwe_stream_add_point(current_ev, end_buf, &iwe, - bss_mesh_id(bss)); + current_ev = iwe_stream_add_point(info, current_ev, end_buf, + &iwe, bss_mesh_id(bss)); } else { iwe.u.data.length = bss->ssid_len; iwe.u.data.flags = 1; - current_ev = iwe_stream_add_point(current_ev, end_buf, &iwe, - bss->ssid); + current_ev = iwe_stream_add_point(info, current_ev, end_buf, + &iwe, bss->ssid); } if (bss->capability & (WLAN_CAPABILITY_ESS | WLAN_CAPABILITY_IBSS) @@ -4128,22 +4129,22 @@ ieee80211_sta_scan_result(struct net_device *dev, iwe.u.mode = IW_MODE_MASTER; else iwe.u.mode = IW_MODE_ADHOC; - current_ev = iwe_stream_add_event(current_ev, end_buf, &iwe, - IW_EV_UINT_LEN); + current_ev = iwe_stream_add_event(info, current_ev, end_buf, + &iwe, IW_EV_UINT_LEN); } memset(&iwe, 0, sizeof(iwe)); iwe.cmd = SIOCGIWFREQ; iwe.u.freq.m = ieee80211_frequency_to_channel(bss->freq); iwe.u.freq.e = 0; - current_ev = iwe_stream_add_event(current_ev, end_buf, &iwe, + current_ev = iwe_stream_add_event(info, current_ev, end_buf, &iwe, IW_EV_FREQ_LEN); memset(&iwe, 0, sizeof(iwe)); iwe.cmd = SIOCGIWFREQ; iwe.u.freq.m = bss->freq; iwe.u.freq.e = 6; - current_ev = iwe_stream_add_event(current_ev, end_buf, &iwe, + current_ev = iwe_stream_add_event(info, current_ev, end_buf, &iwe, IW_EV_FREQ_LEN); memset(&iwe, 0, sizeof(iwe)); iwe.cmd = IWEVQUAL; @@ -4151,7 +4152,7 @@ ieee80211_sta_scan_result(struct net_device *dev, iwe.u.qual.level = bss->signal; iwe.u.qual.noise = bss->noise; iwe.u.qual.updated = local->wstats_flags; - current_ev = iwe_stream_add_event(current_ev, end_buf, &iwe, + current_ev = iwe_stream_add_event(info, current_ev, end_buf, &iwe, IW_EV_QUAL_LEN); memset(&iwe, 0, sizeof(iwe)); @@ -4161,35 +4162,36 @@ ieee80211_sta_scan_result(struct net_device *dev, else iwe.u.data.flags = IW_ENCODE_DISABLED; iwe.u.data.length = 0; - current_ev = iwe_stream_add_point(current_ev, end_buf, &iwe, ""); + current_ev = iwe_stream_add_point(info, current_ev, end_buf, + &iwe, ""); if (bss && bss->wpa_ie) { memset(&iwe, 0, sizeof(iwe)); iwe.cmd = IWEVGENIE; iwe.u.data.length = bss->wpa_ie_len; - current_ev = iwe_stream_add_point(current_ev, end_buf, &iwe, - bss->wpa_ie); + current_ev = iwe_stream_add_point(info, current_ev, end_buf, + &iwe, bss->wpa_ie); } if (bss && bss->rsn_ie) { memset(&iwe, 0, sizeof(iwe)); iwe.cmd = IWEVGENIE; iwe.u.data.length = bss->rsn_ie_len; - current_ev = iwe_stream_add_point(current_ev, end_buf, &iwe, - bss->rsn_ie); + current_ev = iwe_stream_add_point(info, current_ev, end_buf, + &iwe, bss->rsn_ie); } if (bss && bss->ht_ie) { memset(&iwe, 0, sizeof(iwe)); iwe.cmd = IWEVGENIE; iwe.u.data.length = bss->ht_ie_len; - current_ev = iwe_stream_add_point(current_ev, end_buf, &iwe, - bss->ht_ie); + current_ev = iwe_stream_add_point(info, current_ev, end_buf, + &iwe, bss->ht_ie); } if (bss && bss->supp_rates_len > 0) { /* display all supported rates in readable format */ - char *p = current_ev + IW_EV_LCP_LEN; + char *p = current_ev + iwe_stream_lcp_len(info); int i; memset(&iwe, 0, sizeof(iwe)); @@ -4200,7 +4202,7 @@ ieee80211_sta_scan_result(struct net_device *dev, for (i = 0; i < bss->supp_rates_len; i++) { iwe.u.bitrate.value = ((bss->supp_rates[i] & 0x7f) * 500000); - p = iwe_stream_add_value(current_ev, p, + p = iwe_stream_add_value(info, current_ev, p, end_buf, &iwe, IW_EV_PARAM_LEN); } current_ev = p; @@ -4214,7 +4216,8 @@ ieee80211_sta_scan_result(struct net_device *dev, iwe.cmd = IWEVCUSTOM; sprintf(buf, "tsf=%016llx", (unsigned long long)(bss->timestamp)); iwe.u.data.length = strlen(buf); - current_ev = iwe_stream_add_point(current_ev, end_buf, + current_ev = iwe_stream_add_point(info, current_ev, + end_buf, &iwe, buf); kfree(buf); } @@ -4229,31 +4232,36 @@ ieee80211_sta_scan_result(struct net_device *dev, iwe.cmd = IWEVCUSTOM; sprintf(buf, "Mesh network (version %d)", cfg[0]); iwe.u.data.length = strlen(buf); - current_ev = iwe_stream_add_point(current_ev, end_buf, + current_ev = iwe_stream_add_point(info, current_ev, + end_buf, &iwe, buf); sprintf(buf, "Path Selection Protocol ID: " "0x%02X%02X%02X%02X", cfg[1], cfg[2], cfg[3], cfg[4]); iwe.u.data.length = strlen(buf); - current_ev = iwe_stream_add_point(current_ev, end_buf, + current_ev = iwe_stream_add_point(info, current_ev, + end_buf, &iwe, buf); sprintf(buf, "Path Selection Metric ID: " "0x%02X%02X%02X%02X", cfg[5], cfg[6], cfg[7], cfg[8]); iwe.u.data.length = strlen(buf); - current_ev = iwe_stream_add_point(current_ev, end_buf, + current_ev = iwe_stream_add_point(info, current_ev, + end_buf, &iwe, buf); sprintf(buf, "Congestion Control Mode ID: " "0x%02X%02X%02X%02X", cfg[9], cfg[10], cfg[11], cfg[12]); iwe.u.data.length = strlen(buf); - current_ev = iwe_stream_add_point(current_ev, end_buf, + current_ev = iwe_stream_add_point(info, current_ev, + end_buf, &iwe, buf); sprintf(buf, "Channel Precedence: " "0x%02X%02X%02X%02X", cfg[13], cfg[14], cfg[15], cfg[16]); iwe.u.data.length = strlen(buf); - current_ev = iwe_stream_add_point(current_ev, end_buf, + current_ev = iwe_stream_add_point(info, current_ev, + end_buf, &iwe, buf); kfree(buf); } @@ -4263,7 +4271,9 @@ ieee80211_sta_scan_result(struct net_device *dev, } -int ieee80211_sta_scan_results(struct net_device *dev, char *buf, size_t len) +int ieee80211_sta_scan_results(struct net_device *dev, + struct iw_request_info *info, + char *buf, size_t len) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); char *current_ev = buf; @@ -4276,8 +4286,8 @@ int ieee80211_sta_scan_results(struct net_device *dev, char *buf, size_t len) spin_unlock_bh(&local->sta_bss_lock); return -E2BIG; } - current_ev = ieee80211_sta_scan_result(dev, bss, current_ev, - end_buf); + current_ev = ieee80211_sta_scan_result(dev, info, bss, + current_ev, end_buf); } spin_unlock_bh(&local->sta_bss_lock); return current_ev - buf; diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c index 5af3862e7191..f47d13bdf7f7 100644 --- a/net/mac80211/wext.c +++ b/net/mac80211/wext.c @@ -567,7 +567,7 @@ static int ieee80211_ioctl_giwscan(struct net_device *dev, if (local->sta_sw_scanning || local->sta_hw_scanning) return -EAGAIN; - res = ieee80211_sta_scan_results(dev, extra, data->length); + res = ieee80211_sta_scan_results(dev, info, extra, data->length); if (res >= 0) { data->length = res; return 0; -- cgit v1.2.2 From 22196d3648581b253f927186b30075fb005287b0 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 01:06:01 -0700 Subject: decnet: Remove SOCK_SLEEP_{PRE,POST} usage. Just expand the wait sequence. And as a nice side-effect the timeout is respected now. Signed-off-by: David S. Miller --- net/decnet/af_decnet.c | 30 +++++++++++++++--------------- 1 file changed, 15 insertions(+), 15 deletions(-) (limited to 'net') diff --git a/net/decnet/af_decnet.c b/net/decnet/af_decnet.c index fc2efe899e91..931bdf9cb756 100644 --- a/net/decnet/af_decnet.c +++ b/net/decnet/af_decnet.c @@ -1719,6 +1719,8 @@ static int dn_recvmsg(struct kiocb *iocb, struct socket *sock, * See if there is data ready to read, sleep if there isn't */ for(;;) { + DEFINE_WAIT(wait); + if (sk->sk_err) goto out; @@ -1748,14 +1750,11 @@ static int dn_recvmsg(struct kiocb *iocb, struct socket *sock, goto out; } - set_bit(SOCK_ASYNC_WAITDATA, &sock->flags); - SOCK_SLEEP_PRE(sk) - - if (!dn_data_ready(sk, queue, flags, target)) - schedule(); - - SOCK_SLEEP_POST(sk) - clear_bit(SOCK_ASYNC_WAITDATA, &sock->flags); + prepare_to_wait(sk->sk_sleep, &wait, TASK_INTERRUPTIBLE); + set_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags); + sk_wait_event(sk, &timeo, dn_data_ready(sk, queue, flags, target)); + clear_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags); + finish_wait(sk->sk_sleep, &wait); } for(skb = queue->next; skb != (struct sk_buff *)queue; skb = nskb) { @@ -2002,18 +2001,19 @@ static int dn_sendmsg(struct kiocb *iocb, struct socket *sock, * size. */ if (dn_queue_too_long(scp, queue, flags)) { + DEFINE_WAIT(wait); + if (flags & MSG_DONTWAIT) { err = -EWOULDBLOCK; goto out; } - SOCK_SLEEP_PRE(sk) - - if (dn_queue_too_long(scp, queue, flags)) - schedule(); - - SOCK_SLEEP_POST(sk) - + prepare_to_wait(sk->sk_sleep, &wait, TASK_INTERRUPTIBLE); + set_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags); + sk_wait_event(sk, &timeo, + !dn_queue_too_long(scp, queue, flags)); + clear_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags); + finish_wait(sk->sk_sleep, &wait); continue; } -- cgit v1.2.2 From ee5850defcbd98d3a9cb3e0ae93511e7c89bdecd Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 01:21:03 -0700 Subject: llc: Use sock_graft() instead of by-hand version. Signed-off-by: David S. Miller --- net/llc/af_llc.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'net') diff --git a/net/llc/af_llc.c b/net/llc/af_llc.c index 97101dcde4c0..5bcc452a247f 100644 --- a/net/llc/af_llc.c +++ b/net/llc/af_llc.c @@ -128,10 +128,8 @@ static int llc_ui_send_data(struct sock* sk, struct sk_buff *skb, int noblock) static void llc_ui_sk_init(struct socket *sock, struct sock *sk) { + sock_graft(sk, sock); sk->sk_type = sock->type; - sk->sk_sleep = &sock->wait; - sk->sk_socket = sock; - sock->sk = sk; sock->ops = &llc_ui_ops; } -- cgit v1.2.2 From 9375cb8a1232d2a15fe34bec4d3474872e02faec Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 02:20:54 -0700 Subject: ax25: Use sock_graft() and remove bogus sk_socket and sk_sleep init. The way that listening sockets work in ax25 is that the packet input code path creates new socks via ax25_make_new() and attaches them to the incoming SKB. This SKB gets queued up into the listening socket's receive queue. When accept()'d the sock gets hooked up to the real parent socket. Alternatively, if the listening socket is closed and released, any unborn socks stuff up in the receive queue get released. So during this time period these sockets are unreachable in any other way, so no wakeup events nor references to their ->sk_socket and ->sk_sleep members can occur. And even if they do, all such paths have to make NULL checks. So do not deceptively initialize them in ax25_make_new() to the values in the listening socket. Leave them at NULL. Finally, use sock_graft() in ax25_accept(). Signed-off-by: David S. Miller --- net/ax25/af_ax25.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) (limited to 'net') diff --git a/net/ax25/af_ax25.c b/net/ax25/af_ax25.c index 2712544cf0ca..97eaa23ad9ea 100644 --- a/net/ax25/af_ax25.c +++ b/net/ax25/af_ax25.c @@ -893,13 +893,11 @@ struct sock *ax25_make_new(struct sock *osk, struct ax25_dev *ax25_dev) sk->sk_destruct = ax25_free_sock; sk->sk_type = osk->sk_type; - sk->sk_socket = osk->sk_socket; sk->sk_priority = osk->sk_priority; sk->sk_protocol = osk->sk_protocol; sk->sk_rcvbuf = osk->sk_rcvbuf; sk->sk_sndbuf = osk->sk_sndbuf; sk->sk_state = TCP_ESTABLISHED; - sk->sk_sleep = osk->sk_sleep; sock_copy_flags(sk, osk); oax25 = ax25_sk(osk); @@ -1361,13 +1359,11 @@ static int ax25_accept(struct socket *sock, struct socket *newsock, int flags) goto out; newsk = skb->sk; - newsk->sk_socket = newsock; - newsk->sk_sleep = &newsock->wait; + sock_graft(newsk, newsock); /* Now attach up the new socket */ kfree_skb(skb); sk->sk_ack_backlog--; - newsock->sk = newsk; newsock->state = SS_CONNECTED; out: -- cgit v1.2.2 From 7b66767f969edcbdd573aca8063beee7534d242b Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 02:36:44 -0700 Subject: netrom: Use sock_graft() and remove bogus sk_socket and sk_sleep init. This is the netrom variant of changeset 9375cb8a1232d2a15fe34bec4d3474872e02faec ("ax25: Use sock_graft() and remove bogus sk_socket and sk_sleep init.") Signed-off-by: David S. Miller --- net/netrom/af_netrom.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) (limited to 'net') diff --git a/net/netrom/af_netrom.c b/net/netrom/af_netrom.c index 4bae8b998cab..58779624cdbe 100644 --- a/net/netrom/af_netrom.c +++ b/net/netrom/af_netrom.c @@ -475,13 +475,11 @@ static struct sock *nr_make_new(struct sock *osk) sock_init_data(NULL, sk); sk->sk_type = osk->sk_type; - sk->sk_socket = osk->sk_socket; sk->sk_priority = osk->sk_priority; sk->sk_protocol = osk->sk_protocol; sk->sk_rcvbuf = osk->sk_rcvbuf; sk->sk_sndbuf = osk->sk_sndbuf; sk->sk_state = TCP_ESTABLISHED; - sk->sk_sleep = osk->sk_sleep; sock_copy_flags(sk, osk); skb_queue_head_init(&nr->ack_queue); @@ -810,13 +808,11 @@ static int nr_accept(struct socket *sock, struct socket *newsock, int flags) goto out_release; newsk = skb->sk; - newsk->sk_socket = newsock; - newsk->sk_sleep = &newsock->wait; + sock_graft(newsk, newsock); /* Now attach up the new socket */ kfree_skb(skb); sk_acceptq_removed(sk); - newsock->sk = newsk; out_release: release_sock(sk); -- cgit v1.2.2 From 44ccff1f539c8c5bbfc1eacd41cb9ef65022a4ca Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 02:39:21 -0700 Subject: rose: Use sock_graft() and remove bogus sk_socket and sk_sleep init. This is the rose variant of changeset 9375cb8a1232d2a15fe34bec4d3474872e02faec ("ax25: Use sock_graft() and remove bogus sk_socket and sk_sleep init.") Signed-off-by: David S. Miller --- net/rose/af_rose.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) (limited to 'net') diff --git a/net/rose/af_rose.c b/net/rose/af_rose.c index 1ebf65294405..af86bcb604e5 100644 --- a/net/rose/af_rose.c +++ b/net/rose/af_rose.c @@ -566,13 +566,11 @@ static struct sock *rose_make_new(struct sock *osk) #endif sk->sk_type = osk->sk_type; - sk->sk_socket = osk->sk_socket; sk->sk_priority = osk->sk_priority; sk->sk_protocol = osk->sk_protocol; sk->sk_rcvbuf = osk->sk_rcvbuf; sk->sk_sndbuf = osk->sk_sndbuf; sk->sk_state = TCP_ESTABLISHED; - sk->sk_sleep = osk->sk_sleep; sock_copy_flags(sk, osk); init_timer(&rose->timer); @@ -924,14 +922,12 @@ static int rose_accept(struct socket *sock, struct socket *newsock, int flags) goto out_release; newsk = skb->sk; - newsk->sk_socket = newsock; - newsk->sk_sleep = &newsock->wait; + sock_graft(newsk, newsock); /* Now attach up the new socket */ skb->sk = NULL; kfree_skb(skb); sk->sk_ack_backlog--; - newsock->sk = newsk; out_release: release_sock(sk); -- cgit v1.2.2 From b61d38e05542200f6d001b5ea9975dc0dc5bab27 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 02:44:35 -0700 Subject: x25: Use sock_graft() and remove bogus sk_socket and sk_sleep init. This is the x25 variant of changeset 9375cb8a1232d2a15fe34bec4d3474872e02faec ("ax25: Use sock_graft() and remove bogus sk_socket and sk_sleep init.") Signed-off-by: David S. Miller --- net/x25/af_x25.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) (limited to 'net') diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c index 6ba67c523c16..bcb091f713ec 100644 --- a/net/x25/af_x25.c +++ b/net/x25/af_x25.c @@ -555,13 +555,11 @@ static struct sock *x25_make_new(struct sock *osk) x25 = x25_sk(sk); sk->sk_type = osk->sk_type; - sk->sk_socket = osk->sk_socket; sk->sk_priority = osk->sk_priority; sk->sk_protocol = osk->sk_protocol; sk->sk_rcvbuf = osk->sk_rcvbuf; sk->sk_sndbuf = osk->sk_sndbuf; sk->sk_state = TCP_ESTABLISHED; - sk->sk_sleep = osk->sk_sleep; sk->sk_backlog_rcv = osk->sk_backlog_rcv; sock_copy_flags(sk, osk); @@ -808,14 +806,12 @@ static int x25_accept(struct socket *sock, struct socket *newsock, int flags) if (!skb->sk) goto out2; newsk = skb->sk; - newsk->sk_socket = newsock; - newsk->sk_sleep = &newsock->wait; + sock_graft(newsk, newsock); /* Now attach up the new socket */ skb->sk = NULL; kfree_skb(skb); sk->sk_ack_backlog--; - newsock->sk = newsk; newsock->state = SS_CONNECTED; rc = 0; out2: -- cgit v1.2.2 From 0efffaf9d57f54a864b739176d1f659a69486e8b Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 03:01:47 -0700 Subject: econet: Use sock_orphan() instead of open-coded (and buggy) variant. It doesn't grab the sk_callback_lock, it doesn't NULL out the sk->sk_sleep waitqueue pointer, etc. Signed-off-by: David S. Miller --- net/econet/af_econet.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'net') diff --git a/net/econet/af_econet.c b/net/econet/af_econet.c index 7c9bb13b1539..d35127bb84e1 100644 --- a/net/econet/af_econet.c +++ b/net/econet/af_econet.c @@ -573,9 +573,7 @@ static int econet_release(struct socket *sock) sk->sk_state_change(sk); /* It is useless. Just for sanity. */ - sock->sk = NULL; - sk->sk_socket = NULL; - sock_set_flag(sk, SOCK_DEAD); + sock_orphan(sk); /* Purge queues */ -- cgit v1.2.2 From c751e4f8b32a3869bb4fec12100952abd9baa0e1 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 03:05:13 -0700 Subject: x25: Use sock_orphan() instead of open-coded (and buggy) variant. It doesn't grab the sk_callback_lock, it doesn't NULL out the sk->sk_sleep waitqueue pointer, etc. Signed-off-by: David S. Miller --- net/x25/af_x25.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'net') diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c index bcb091f713ec..7b1c6ef04553 100644 --- a/net/x25/af_x25.c +++ b/net/x25/af_x25.c @@ -612,8 +612,7 @@ static int x25_release(struct socket *sock) break; } - sock->sk = NULL; - sk->sk_socket = NULL; /* Not used, but we should do this */ + sock_orphan(sk); out: return 0; } -- cgit v1.2.2 From 48c5732f4ac0621a2fdde006d55a6621a47a728f Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 03:19:58 -0700 Subject: netrom: Kill spurious NULL'ing of sk->sk_socket. In nr_release(), one code path calls sock_orphan() which will NULL out sk->sk_socket already. In the other case, handling states other than NR_STATE_{0,1,2,3}, seems to not be possible other than due to bugs. Even for an uninitialized nr->state value, that would be zero or NR_STATE_0. It might be wise to stick a WARN_ON() here. Signed-off-by: David S. Miller --- net/netrom/af_netrom.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'net') diff --git a/net/netrom/af_netrom.c b/net/netrom/af_netrom.c index 58779624cdbe..74884f4a6255 100644 --- a/net/netrom/af_netrom.c +++ b/net/netrom/af_netrom.c @@ -536,11 +536,9 @@ static int nr_release(struct socket *sock) sk->sk_state_change(sk); sock_orphan(sk); sock_set_flag(sk, SOCK_DESTROY); - sk->sk_socket = NULL; break; default: - sk->sk_socket = NULL; break; } -- cgit v1.2.2 From 3d00fb9eb11ac49d4035f756d116deeeaf99a26b Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Tue, 17 Jun 2008 15:54:14 -0700 Subject: sctp: fix error path in sctp_proc_init After the sctp_remaddr_proc_init failed, the proper rollback is not the sctp_remaddr_proc_exit, but the sctp_assocs_proc_exit. Signed-off-by: Pavel Emelyanov Acked-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/protocol.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c index 23aaffb97ca3..98c6a882016a 100644 --- a/net/sctp/protocol.c +++ b/net/sctp/protocol.c @@ -119,7 +119,7 @@ static __init int sctp_proc_init(void) return 0; out_remaddr_proc_init: - sctp_remaddr_proc_exit(); + sctp_assocs_proc_exit(); out_assocs_proc_init: sctp_eps_proc_exit(); out_eps_proc_init: -- cgit v1.2.2 From 43aa1920117801fe9ae3d1fad886b62511e09bee Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Tue, 17 Jun 2008 16:09:45 -0700 Subject: bridge: handle process all link-local frames Any frame addressed to link-local addresses should be processed by local receive path. The earlier code would process them only if STP was enabled. Since there are other frames like LACP for bonding, we should always process them. Signed-off-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/bridge/br_input.c | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c index 0145e9416714..30b88777c3df 100644 --- a/net/bridge/br_input.c +++ b/net/bridge/br_input.c @@ -134,14 +134,11 @@ struct sk_buff *br_handle_frame(struct net_bridge_port *p, struct sk_buff *skb) if (skb->protocol == htons(ETH_P_PAUSE)) goto drop; - /* Process STP BPDU's through normal netif_receive_skb() path */ - if (p->br->stp_enabled != BR_NO_STP) { - if (NF_HOOK(PF_BRIDGE, NF_BR_LOCAL_IN, skb, skb->dev, - NULL, br_handle_local_finish)) - return NULL; - else - return skb; - } + if (NF_HOOK(PF_BRIDGE, NF_BR_LOCAL_IN, skb, skb->dev, + NULL, br_handle_local_finish)) + return NULL; /* frame consumed by filter */ + else + return skb; /* continue processing */ } switch (p->state) { -- cgit v1.2.2 From 92c0574f11598c8036f81e27d2e8bdd6eed7d76d Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Tue, 17 Jun 2008 16:10:06 -0700 Subject: bridge: make bridge address settings sticky Normally, the bridge just chooses the smallest mac address as the bridge id and mac address of bridge device. But if the administrator has explictly set the interface address then don't change it. Signed-off-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/bridge/br_device.c | 1 + net/bridge/br_private.h | 2 ++ net/bridge/br_stp_if.c | 4 ++++ 3 files changed, 7 insertions(+) (limited to 'net') diff --git a/net/bridge/br_device.c b/net/bridge/br_device.c index a6ffc6c2a69f..d9449df7cad5 100644 --- a/net/bridge/br_device.c +++ b/net/bridge/br_device.c @@ -87,6 +87,7 @@ static int br_set_mac_address(struct net_device *dev, void *p) spin_lock_bh(&br->lock); memcpy(dev->dev_addr, addr->sa_data, ETH_ALEN); br_stp_change_bridge_id(br, addr->sa_data); + br->flags |= BR_SET_MAC_ADDR; spin_unlock_bh(&br->lock); return 0; diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h index 83ff5861c2d2..8593c9f6a302 100644 --- a/net/bridge/br_private.h +++ b/net/bridge/br_private.h @@ -92,6 +92,8 @@ struct net_bridge struct hlist_head hash[BR_HASH_SIZE]; struct list_head age_list; unsigned long feature_mask; + unsigned long flags; +#define BR_SET_MAC_ADDR 0x00000001 /* STP */ bridge_id designated_root; diff --git a/net/bridge/br_stp_if.c b/net/bridge/br_stp_if.c index 1a4e5c37a0cf..9a52ac5b4525 100644 --- a/net/bridge/br_stp_if.c +++ b/net/bridge/br_stp_if.c @@ -214,6 +214,10 @@ void br_stp_recalculate_bridge_id(struct net_bridge *br) const unsigned char *addr = br_mac_zero; struct net_bridge_port *p; + /* user has chosen a value so keep it */ + if (br->flags & BR_SET_MAC_ADDR) + return; + list_for_each_entry(p, &br->port_list, list) { if (addr == br_mac_zero || memcmp(p->dev->dev_addr, addr, ETH_ALEN) < 0) -- cgit v1.2.2 From f586287e0fed366d80822666f70487472ab8793a Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Tue, 17 Jun 2008 16:16:13 -0700 Subject: bridge: fix IPV6=n build Fix bridge netfilter code so that it uses CONFIG_IPV6 as needed: net/built-in.o: In function `ebt_filter_ip6': ebt_ip6.c:(.text+0x87c37): undefined reference to `ipv6_skip_exthdr' net/built-in.o: In function `ebt_log_packet': ebt_log.c:(.text+0x88dee): undefined reference to `ipv6_skip_exthdr' make[1]: *** [.tmp_vmlinux1] Error 1 Signed-off-by: Randy Dunlap Signed-off-by: David S. Miller --- net/bridge/netfilter/Kconfig | 2 +- net/bridge/netfilter/Makefile | 2 +- net/bridge/netfilter/ebt_log.c | 2 ++ 3 files changed, 4 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/bridge/netfilter/Kconfig b/net/bridge/netfilter/Kconfig index fb684c2ff8b6..540df4106bec 100644 --- a/net/bridge/netfilter/Kconfig +++ b/net/bridge/netfilter/Kconfig @@ -85,7 +85,7 @@ config BRIDGE_EBT_IP config BRIDGE_EBT_IP6 tristate "ebt: IP6 filter support" - depends on BRIDGE_NF_EBTABLES + depends on BRIDGE_NF_EBTABLES && IPV6 help This option adds the IP6 match, which allows basic IPV6 header field filtering. diff --git a/net/bridge/netfilter/Makefile b/net/bridge/netfilter/Makefile index dd960645b413..0718699540b0 100644 --- a/net/bridge/netfilter/Makefile +++ b/net/bridge/netfilter/Makefile @@ -14,7 +14,7 @@ obj-$(CONFIG_BRIDGE_EBT_802_3) += ebt_802_3.o obj-$(CONFIG_BRIDGE_EBT_AMONG) += ebt_among.o obj-$(CONFIG_BRIDGE_EBT_ARP) += ebt_arp.o obj-$(CONFIG_BRIDGE_EBT_IP) += ebt_ip.o -obj-$(CONFIG_BRIDGE_EBT_IP) += ebt_ip6.o +obj-$(CONFIG_BRIDGE_EBT_IP6) += ebt_ip6.o obj-$(CONFIG_BRIDGE_EBT_LIMIT) += ebt_limit.o obj-$(CONFIG_BRIDGE_EBT_MARK) += ebt_mark_m.o obj-$(CONFIG_BRIDGE_EBT_PKTTYPE) += ebt_pkttype.o diff --git a/net/bridge/netfilter/ebt_log.c b/net/bridge/netfilter/ebt_log.c index c883ec8a28b4..2f430d4ae911 100644 --- a/net/bridge/netfilter/ebt_log.c +++ b/net/bridge/netfilter/ebt_log.c @@ -123,6 +123,7 @@ ebt_log_packet(unsigned int pf, unsigned int hooknum, goto out; } +#if defined(CONFIG_BRIDGE_EBT_IP6) || defined(CONFIG_BRIDGE_EBT_IP6_MODULE) if ((bitmask & EBT_LOG_IP6) && eth_hdr(skb)->h_proto == htons(ETH_P_IPV6)) { const struct ipv6hdr *ih; @@ -146,6 +147,7 @@ ebt_log_packet(unsigned int pf, unsigned int hooknum, print_ports(skb, nexthdr, offset_ph); goto out; } +#endif if ((bitmask & EBT_LOG_ARP) && ((eth_hdr(skb)->h_proto == htons(ETH_P_ARP)) || -- cgit v1.2.2 From 61c33e012964ce358b42d2a1e9cd309af5dab02b Mon Sep 17 00:00:00 2001 From: Mitchell Blank Jr Date: Tue, 17 Jun 2008 16:20:06 -0700 Subject: atm: use const where reasonable From: Mitchell Blank Jr Signed-off-by: Chas Williams Signed-off-by: David S. Miller --- net/atm/addr.c | 10 +++++----- net/atm/addr.h | 4 ++-- net/atm/br2684.c | 14 +++++++------- net/atm/common.c | 8 ++++---- net/atm/lec.c | 55 +++++++++++++++++++++++++++---------------------------- net/atm/lec.h | 10 +++++----- 6 files changed, 50 insertions(+), 51 deletions(-) (limited to 'net') diff --git a/net/atm/addr.c b/net/atm/addr.c index 6afa77d63bb5..82e85abc303d 100644 --- a/net/atm/addr.c +++ b/net/atm/addr.c @@ -9,7 +9,7 @@ #include "signaling.h" #include "addr.h" -static int check_addr(struct sockaddr_atmsvc *addr) +static int check_addr(const struct sockaddr_atmsvc *addr) { int i; @@ -23,7 +23,7 @@ static int check_addr(struct sockaddr_atmsvc *addr) return -EINVAL; } -static int identical(struct sockaddr_atmsvc *a, struct sockaddr_atmsvc *b) +static int identical(const struct sockaddr_atmsvc *a, const struct sockaddr_atmsvc *b) { if (*a->sas_addr.prv) if (memcmp(a->sas_addr.prv, b->sas_addr.prv, ATM_ESA_LEN)) @@ -35,7 +35,7 @@ static int identical(struct sockaddr_atmsvc *a, struct sockaddr_atmsvc *b) return !strcmp(a->sas_addr.pub, b->sas_addr.pub); } -static void notify_sigd(struct atm_dev *dev) +static void notify_sigd(const struct atm_dev *dev) { struct sockaddr_atmpvc pvc; @@ -63,7 +63,7 @@ void atm_reset_addr(struct atm_dev *dev, enum atm_addr_type_t atype) notify_sigd(dev); } -int atm_add_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr, +int atm_add_addr(struct atm_dev *dev, const struct sockaddr_atmsvc *addr, enum atm_addr_type_t atype) { unsigned long flags; @@ -98,7 +98,7 @@ int atm_add_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr, return 0; } -int atm_del_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr, +int atm_del_addr(struct atm_dev *dev, const struct sockaddr_atmsvc *addr, enum atm_addr_type_t atype) { unsigned long flags; diff --git a/net/atm/addr.h b/net/atm/addr.h index f39433ad45da..6837e9e7eb13 100644 --- a/net/atm/addr.h +++ b/net/atm/addr.h @@ -10,9 +10,9 @@ #include void atm_reset_addr(struct atm_dev *dev, enum atm_addr_type_t type); -int atm_add_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr, +int atm_add_addr(struct atm_dev *dev, const struct sockaddr_atmsvc *addr, enum atm_addr_type_t type); -int atm_del_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr, +int atm_del_addr(struct atm_dev *dev, const struct sockaddr_atmsvc *addr, enum atm_addr_type_t type); int atm_get_addr(struct atm_dev *dev, struct sockaddr_atmsvc __user *buf, size_t size, enum atm_addr_type_t type); diff --git a/net/atm/br2684.c b/net/atm/br2684.c index 05fafdc2eea3..8d9a6f158880 100644 --- a/net/atm/br2684.c +++ b/net/atm/br2684.c @@ -52,12 +52,12 @@ static void skb_debug(const struct sk_buff *skb) #define ETHERTYPE_IPV6 0x86, 0xdd #define PAD_BRIDGED 0x00, 0x00 -static unsigned char ethertype_ipv4[] = { ETHERTYPE_IPV4 }; -static unsigned char ethertype_ipv6[] = { ETHERTYPE_IPV6 }; -static unsigned char llc_oui_pid_pad[] = +static const unsigned char ethertype_ipv4[] = { ETHERTYPE_IPV4 }; +static const unsigned char ethertype_ipv6[] = { ETHERTYPE_IPV6 }; +static const unsigned char llc_oui_pid_pad[] = { LLC, SNAP_BRIDGED, PID_ETHERNET, PAD_BRIDGED }; -static unsigned char llc_oui_ipv4[] = { LLC, SNAP_ROUTED, ETHERTYPE_IPV4 }; -static unsigned char llc_oui_ipv6[] = { LLC, SNAP_ROUTED, ETHERTYPE_IPV6 }; +static const unsigned char llc_oui_ipv4[] = { LLC, SNAP_ROUTED, ETHERTYPE_IPV4 }; +static const unsigned char llc_oui_ipv6[] = { LLC, SNAP_ROUTED, ETHERTYPE_IPV6 }; enum br2684_encaps { e_vc = BR2684_ENCAPS_VC, @@ -217,8 +217,8 @@ static int br2684_xmit_vcc(struct sk_buff *skb, struct br2684_dev *brdev, return 1; } -static inline struct br2684_vcc *pick_outgoing_vcc(struct sk_buff *skb, - struct br2684_dev *brdev) +static inline struct br2684_vcc *pick_outgoing_vcc(const struct sk_buff *skb, + const struct br2684_dev *brdev) { return list_empty(&brdev->brvccs) ? NULL : list_entry_brvcc(brdev->brvccs.next); /* 1 vcc/dev right now */ } diff --git a/net/atm/common.c b/net/atm/common.c index c865517ba449..d34edbe754c8 100644 --- a/net/atm/common.c +++ b/net/atm/common.c @@ -262,7 +262,7 @@ static int adjust_tp(struct atm_trafprm *tp,unsigned char aal) } -static int check_ci(struct atm_vcc *vcc, short vpi, int vci) +static int check_ci(const struct atm_vcc *vcc, short vpi, int vci) { struct hlist_head *head = &vcc_hash[vci & (VCC_HTABLE_SIZE - 1)]; @@ -290,7 +290,7 @@ static int check_ci(struct atm_vcc *vcc, short vpi, int vci) } -static int find_ci(struct atm_vcc *vcc, short *vpi, int *vci) +static int find_ci(const struct atm_vcc *vcc, short *vpi, int *vci) { static short p; /* poor man's per-device cache */ static int c; @@ -646,7 +646,7 @@ static int atm_change_qos(struct atm_vcc *vcc,struct atm_qos *qos) } -static int check_tp(struct atm_trafprm *tp) +static int check_tp(const struct atm_trafprm *tp) { /* @@@ Should be merged with adjust_tp */ if (!tp->traffic_class || tp->traffic_class == ATM_ANYCLASS) return 0; @@ -663,7 +663,7 @@ static int check_tp(struct atm_trafprm *tp) } -static int check_qos(struct atm_qos *qos) +static int check_qos(const struct atm_qos *qos) { int error; diff --git a/net/atm/lec.c b/net/atm/lec.c index 653aca3573ac..5799fb52365a 100644 --- a/net/atm/lec.c +++ b/net/atm/lec.c @@ -65,36 +65,36 @@ static int lec_close(struct net_device *dev); static struct net_device_stats *lec_get_stats(struct net_device *dev); static void lec_init(struct net_device *dev); static struct lec_arp_table *lec_arp_find(struct lec_priv *priv, - unsigned char *mac_addr); + const unsigned char *mac_addr); static int lec_arp_remove(struct lec_priv *priv, struct lec_arp_table *to_remove); /* LANE2 functions */ -static void lane2_associate_ind(struct net_device *dev, u8 *mac_address, - u8 *tlvs, u32 sizeoftlvs); -static int lane2_resolve(struct net_device *dev, u8 *dst_mac, int force, +static void lane2_associate_ind(struct net_device *dev, const u8 *mac_address, + const u8 *tlvs, u32 sizeoftlvs); +static int lane2_resolve(struct net_device *dev, const u8 *dst_mac, int force, u8 **tlvs, u32 *sizeoftlvs); -static int lane2_associate_req(struct net_device *dev, u8 *lan_dst, - u8 *tlvs, u32 sizeoftlvs); +static int lane2_associate_req(struct net_device *dev, const u8 *lan_dst, + const u8 *tlvs, u32 sizeoftlvs); -static int lec_addr_delete(struct lec_priv *priv, unsigned char *atm_addr, +static int lec_addr_delete(struct lec_priv *priv, const unsigned char *atm_addr, unsigned long permanent); static void lec_arp_check_empties(struct lec_priv *priv, struct atm_vcc *vcc, struct sk_buff *skb); static void lec_arp_destroy(struct lec_priv *priv); static void lec_arp_init(struct lec_priv *priv); static struct atm_vcc *lec_arp_resolve(struct lec_priv *priv, - unsigned char *mac_to_find, + const unsigned char *mac_to_find, int is_rdesc, struct lec_arp_table **ret_entry); -static void lec_arp_update(struct lec_priv *priv, unsigned char *mac_addr, - unsigned char *atm_addr, unsigned long remoteflag, +static void lec_arp_update(struct lec_priv *priv, const unsigned char *mac_addr, + const unsigned char *atm_addr, unsigned long remoteflag, unsigned int targetless_le_arp); static void lec_flush_complete(struct lec_priv *priv, unsigned long tran_id); static int lec_mcast_make(struct lec_priv *priv, struct atm_vcc *vcc); static void lec_set_flush_tran_id(struct lec_priv *priv, - unsigned char *atm_addr, + const unsigned char *atm_addr, unsigned long tran_id); -static void lec_vcc_added(struct lec_priv *priv, struct atmlec_ioc *ioc_data, +static void lec_vcc_added(struct lec_priv *priv, const struct atmlec_ioc *ioc_data, struct atm_vcc *vcc, void (*old_push) (struct atm_vcc *vcc, struct sk_buff *skb)); @@ -634,7 +634,7 @@ static struct atm_dev lecatm_dev = { */ static int send_to_lecd(struct lec_priv *priv, atmlec_msg_type type, - unsigned char *mac_addr, unsigned char *atm_addr, + const unsigned char *mac_addr, const unsigned char *atm_addr, struct sk_buff *data) { struct sock *sk; @@ -705,10 +705,9 @@ static void lec_init(struct net_device *dev) dev->set_multicast_list = lec_set_multicast_list; dev->do_ioctl = NULL; printk("%s: Initialized!\n", dev->name); - return; } -static unsigned char lec_ctrl_magic[] = { +static const unsigned char lec_ctrl_magic[] = { 0xff, 0x00, 0x01, @@ -1276,7 +1275,7 @@ module_exit(lane_module_cleanup); * lec will be used. * If dst_mac == NULL, targetless LE_ARP will be sent */ -static int lane2_resolve(struct net_device *dev, u8 *dst_mac, int force, +static int lane2_resolve(struct net_device *dev, const u8 *dst_mac, int force, u8 **tlvs, u32 *sizeoftlvs) { unsigned long flags; @@ -1322,8 +1321,8 @@ static int lane2_resolve(struct net_device *dev, u8 *dst_mac, int force, * Returns 1 for success, 0 for failure (out of memory) * */ -static int lane2_associate_req(struct net_device *dev, u8 *lan_dst, - u8 *tlvs, u32 sizeoftlvs) +static int lane2_associate_req(struct net_device *dev, const u8 *lan_dst, + const u8 *tlvs, u32 sizeoftlvs) { int retval; struct sk_buff *skb; @@ -1358,8 +1357,8 @@ static int lane2_associate_req(struct net_device *dev, u8 *lan_dst, * LANE2: 3.1.5, LE_ASSOCIATE.indication * */ -static void lane2_associate_ind(struct net_device *dev, u8 *mac_addr, - u8 *tlvs, u32 sizeoftlvs) +static void lane2_associate_ind(struct net_device *dev, const u8 *mac_addr, + const u8 *tlvs, u32 sizeoftlvs) { #if 0 int i = 0; @@ -1744,7 +1743,7 @@ static void lec_arp_destroy(struct lec_priv *priv) * Find entry by mac_address */ static struct lec_arp_table *lec_arp_find(struct lec_priv *priv, - unsigned char *mac_addr) + const unsigned char *mac_addr) { struct hlist_node *node; struct hlist_head *head; @@ -1764,7 +1763,7 @@ static struct lec_arp_table *lec_arp_find(struct lec_priv *priv, } static struct lec_arp_table *make_entry(struct lec_priv *priv, - unsigned char *mac_addr) + const unsigned char *mac_addr) { struct lec_arp_table *to_return; @@ -1921,7 +1920,7 @@ restart: * */ static struct atm_vcc *lec_arp_resolve(struct lec_priv *priv, - unsigned char *mac_to_find, int is_rdesc, + const unsigned char *mac_to_find, int is_rdesc, struct lec_arp_table **ret_entry) { unsigned long flags; @@ -2017,7 +2016,7 @@ out: } static int -lec_addr_delete(struct lec_priv *priv, unsigned char *atm_addr, +lec_addr_delete(struct lec_priv *priv, const unsigned char *atm_addr, unsigned long permanent) { unsigned long flags; @@ -2047,8 +2046,8 @@ lec_addr_delete(struct lec_priv *priv, unsigned char *atm_addr, * Notifies: Response to arp_request (atm_addr != NULL) */ static void -lec_arp_update(struct lec_priv *priv, unsigned char *mac_addr, - unsigned char *atm_addr, unsigned long remoteflag, +lec_arp_update(struct lec_priv *priv, const unsigned char *mac_addr, + const unsigned char *atm_addr, unsigned long remoteflag, unsigned int targetless_le_arp) { unsigned long flags; @@ -2148,7 +2147,7 @@ out: * Notifies: Vcc setup ready */ static void -lec_vcc_added(struct lec_priv *priv, struct atmlec_ioc *ioc_data, +lec_vcc_added(struct lec_priv *priv, const struct atmlec_ioc *ioc_data, struct atm_vcc *vcc, void (*old_push) (struct atm_vcc *vcc, struct sk_buff *skb)) { @@ -2336,7 +2335,7 @@ restart: static void lec_set_flush_tran_id(struct lec_priv *priv, - unsigned char *atm_addr, unsigned long tran_id) + const unsigned char *atm_addr, unsigned long tran_id) { unsigned long flags; struct hlist_node *node; diff --git a/net/atm/lec.h b/net/atm/lec.h index b41cda7ea1e1..0d376682c1a3 100644 --- a/net/atm/lec.h +++ b/net/atm/lec.h @@ -42,12 +42,12 @@ struct lecdatahdr_8025 { * */ struct lane2_ops { - int (*resolve) (struct net_device *dev, u8 *dst_mac, int force, + int (*resolve) (struct net_device *dev, const u8 *dst_mac, int force, u8 **tlvs, u32 *sizeoftlvs); - int (*associate_req) (struct net_device *dev, u8 *lan_dst, - u8 *tlvs, u32 sizeoftlvs); - void (*associate_indicator) (struct net_device *dev, u8 *mac_addr, - u8 *tlvs, u32 sizeoftlvs); + int (*associate_req) (struct net_device *dev, const u8 *lan_dst, + const u8 *tlvs, u32 sizeoftlvs); + void (*associate_indicator) (struct net_device *dev, const u8 *mac_addr, + const u8 *tlvs, u32 sizeoftlvs); }; /* -- cgit v1.2.2 From fe2c802ab62aa63d276deafa905875f3455f2621 Mon Sep 17 00:00:00 2001 From: Bernard Pidoux Date: Tue, 17 Jun 2008 17:08:32 -0700 Subject: rose: improving AX25 routing frames via ROSE network ROSE network is organized through nodes connected via hamradio or Internet. AX25 packet radio frames sent to a remote ROSE address destination are routed through these nodes. Without the present patch, automatic routing mechanism did not work optimally due to an improper parameter checking. rose_get_neigh() function is called either by rose_connect() or by rose_route_frame(). In the case of a call from rose_connect(), f0 timer is checked to find if a connection is already pending. In that case it returns the address of the neighbour, or returns a NULL otherwise. When called by rose_route_frame() the purpose was to route a packet AX25 frame through an adjacent node given a destination rose address. However, in that case, t0 timer checked does not indicate if the adjacent node is actually connected even if the timer is not null. Thus, for each frame sent, the function often tried to start a new connexion even if the adjacent node was already connected. The patch adds a "new" parameter that is true when the function is called by rose route_frame(). This instructs rose_get_neigh() to check node parameter "restarted". If restarted is true it means that the route to the destination address is opened via a neighbour node already connected. If "restarted" is false the function returns a NULL. In that case the calling function will initiate a new connection as before. This results in a fast routing of frames, from nodes to nodes, until destination is reached, as originaly specified by ROSE protocole. Signed-off-by: Bernard Pidoux Signed-off-by: David S. Miller --- net/rose/af_rose.c | 4 ++-- net/rose/rose_route.c | 29 ++++++++++++++++++----------- 2 files changed, 20 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/rose/af_rose.c b/net/rose/af_rose.c index af86bcb604e5..46461a69cd0f 100644 --- a/net/rose/af_rose.c +++ b/net/rose/af_rose.c @@ -757,7 +757,7 @@ static int rose_connect(struct socket *sock, struct sockaddr *uaddr, int addr_le sock->state = SS_UNCONNECTED; rose->neighbour = rose_get_neigh(&addr->srose_addr, &cause, - &diagnostic); + &diagnostic, 0); if (!rose->neighbour) { err = -ENETUNREACH; goto out_release; @@ -853,7 +853,7 @@ rose_try_next_neigh: if (sk->sk_state != TCP_ESTABLISHED) { /* Try next neighbour */ - rose->neighbour = rose_get_neigh(&addr->srose_addr, &cause, &diagnostic); + rose->neighbour = rose_get_neigh(&addr->srose_addr, &cause, &diagnostic, 0); if (rose->neighbour) goto rose_try_next_neigh; diff --git a/net/rose/rose_route.c b/net/rose/rose_route.c index bd593871c81e..a81066a1010a 100644 --- a/net/rose/rose_route.c +++ b/net/rose/rose_route.c @@ -662,27 +662,34 @@ struct rose_route *rose_route_free_lci(unsigned int lci, struct rose_neigh *neig } /* - * Find a neighbour given a ROSE address. + * Find a neighbour or a route given a ROSE address. */ struct rose_neigh *rose_get_neigh(rose_address *addr, unsigned char *cause, - unsigned char *diagnostic) + unsigned char *diagnostic, int new) { struct rose_neigh *res = NULL; struct rose_node *node; int failed = 0; int i; - spin_lock_bh(&rose_node_list_lock); + if (!new) spin_lock_bh(&rose_node_list_lock); for (node = rose_node_list; node != NULL; node = node->next) { if (rosecmpm(addr, &node->address, node->mask) == 0) { for (i = 0; i < node->count; i++) { - if (!rose_ftimer_running(node->neighbour[i])) { - res = node->neighbour[i]; - goto out; - } else - failed = 1; + if (new) { + if (node->neighbour[i]->restarted) { + res = node->neighbour[i]; + goto out; + } + } + else { + if (!rose_ftimer_running(node->neighbour[i])) { + res = node->neighbour[i]; + goto out; + } else + failed = 1; + } } - break; } } @@ -695,7 +702,7 @@ struct rose_neigh *rose_get_neigh(rose_address *addr, unsigned char *cause, } out: - spin_unlock_bh(&rose_node_list_lock); + if (!new) spin_unlock_bh(&rose_node_list_lock); return res; } @@ -1018,7 +1025,7 @@ int rose_route_frame(struct sk_buff *skb, ax25_cb *ax25) rose_route = rose_route->next; } - if ((new_neigh = rose_get_neigh(dest_addr, &cause, &diagnostic)) == NULL) { + if ((new_neigh = rose_get_neigh(dest_addr, &cause, &diagnostic, 1)) == NULL) { rose_transmit_clear_request(rose_neigh, lci, cause, diagnostic); goto out; } -- cgit v1.2.2 From c1da4ac752b8b0411791d26c678fcf23d2eed242 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Fri, 13 Jun 2008 18:12:00 -0700 Subject: net/core: add NETDEV_BONDING_FAILOVER event Add NETDEV_BONDING_FAILOVER event to be used in a successive patch by bonding to announce fail-over for the active-backup mode through the netdev events notifier chain mechanism. Such an event can be of use for the RDMA CM (communication manager) to let native RDMA ULPs (eg NFS-RDMA, iSER) always be aligned with the IP stack, in the sense that they use the same ports/links as the stack does. More usages can be done to allow monitoring tools based on netlink events being aware to bonding fail-over. Signed-off-by: Or Gerlitz Signed-off-by: Jay Vosburgh Signed-off-by: Jeff Garzik --- net/core/dev.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 68d8df0992ab..0e45742e7158 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -961,6 +961,12 @@ void netdev_state_change(struct net_device *dev) } } +void netdev_bonding_change(struct net_device *dev) +{ + call_netdevice_notifiers(NETDEV_BONDING_FAILOVER, dev); +} +EXPORT_SYMBOL(netdev_bonding_change); + /** * dev_load - load a network module * @net: the applicable net namespace -- cgit v1.2.2 From b8a9787eddb0e4665f31dd1d64584732b2b5d051 Mon Sep 17 00:00:00 2001 From: Jay Vosburgh Date: Fri, 13 Jun 2008 18:12:04 -0700 Subject: bonding: Allow setting max_bonds to zero Permit bonding to function rationally if max_bonds is set to zero. This will load the module, but create no master devices (which can be created via sysfs). Requires some change to bond_create_sysfs; currently, the netdev sysfs directory is determined from the first bonding device created, but this is no longer possible. Instead, an interface from net/core is created to create and destroy files in net_class. Based on a patch submitted by Phil Oester . Modified by Jay Vosburgh to fix the sysfs issue mentioned above and to update the documentation. Signed-off-by: Phil Oester Signed-off-by: Jay Vosburgh Signed-off-by: Jeff Garzik --- net/core/net-sysfs.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'net') diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index dccd737ea2e3..3f7941319217 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -468,6 +468,19 @@ int netdev_register_kobject(struct net_device *net) return device_add(dev); } +int netdev_class_create_file(struct class_attribute *class_attr) +{ + return class_create_file(&net_class, class_attr); +} + +void netdev_class_remove_file(struct class_attribute *class_attr) +{ + class_remove_file(&net_class, class_attr); +} + +EXPORT_SYMBOL(netdev_class_create_file); +EXPORT_SYMBOL(netdev_class_remove_file); + void netdev_initialize_kobject(struct net_device *net) { struct device *device = &(net->dev); -- cgit v1.2.2 From cb61cb9b8b5ef6c2697d84e5015e314626eb2fba Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 17 Jun 2008 21:04:56 -0700 Subject: udp: sk_drops handling In commits 33c732c36169d7022ad7d6eb474b0c9be43a2dc1 ([IPV4]: Add raw drops counter) and a92aa318b4b369091fd80433c80e62838db8bc1c ([IPV6]: Add raw drops counter), Wang Chen added raw drops counter for /proc/net/raw & /proc/net/raw6 This patch adds this capability to UDP sockets too (/proc/net/udp & /proc/net/udp6). This means that 'RcvbufErrors' errors found in /proc/net/snmp can be also be examined for each udp socket. # grep Udp: /proc/net/snmp Udp: InDatagrams NoPorts InErrors OutDatagrams RcvbufErrors SndbufErrors Udp: 23971006 75 899420 16390693 146348 0 # cat /proc/net/udp sl local_address rem_address st tx_queue rx_queue tr tm->when retrnsmt --- uid timeout inode ref pointer drops 75: 00000000:02CB 00000000:0000 07 00000000:00000000 00:00000000 00000000 --- 0 0 2358 2 ffff81082a538c80 0 111: 00000000:006F 00000000:0000 07 00000000:00000000 00:00000000 00000000 --- 0 0 2286 2 ffff81042dd35c80 146348 In this example, only port 111 (0x006F) was flooded by messages that user program could not read fast enough. 146348 messages were lost. Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- net/ipv4/raw.c | 2 +- net/ipv4/udp.c | 11 +++++++---- net/ipv6/raw.c | 2 +- net/ipv6/udp.c | 11 +++++++---- 4 files changed, 16 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c index 7d449468409e..925fdf18cf92 100644 --- a/net/ipv4/raw.c +++ b/net/ipv4/raw.c @@ -944,7 +944,7 @@ static int raw_seq_show(struct seq_file *seq, void *v) if (v == SEQ_START_TOKEN) seq_printf(seq, " sl local_address rem_address st tx_queue " "rx_queue tr tm->when retrnsmt uid timeout " - "inode drops\n"); + "inode ref pointer drops\n"); else raw_sock_seq_show(seq, v, raw_seq_private(seq)->bucket); return 0; diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 11eabf136144..3bbf6fb6e4f5 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -1040,8 +1040,10 @@ int udp_queue_rcv_skb(struct sock * sk, struct sk_buff *skb) if ((rc = sock_queue_rcv_skb(sk,skb)) < 0) { /* Note that an ENOMEM error is charged twice */ - if (rc == -ENOMEM) + if (rc == -ENOMEM) { UDP_INC_STATS_BH(UDP_MIB_RCVBUFERRORS, is_udplite); + atomic_inc(&sk->sk_drops); + } goto drop; } @@ -1629,12 +1631,13 @@ static void udp4_format_sock(struct sock *sp, struct seq_file *f, __u16 srcp = ntohs(inet->sport); seq_printf(f, "%4d: %08X:%04X %08X:%04X" - " %02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p%n", + " %02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p %d%n", bucket, src, srcp, dest, destp, sp->sk_state, atomic_read(&sp->sk_wmem_alloc), atomic_read(&sp->sk_rmem_alloc), 0, 0L, 0, sock_i_uid(sp), 0, sock_i_ino(sp), - atomic_read(&sp->sk_refcnt), sp, len); + atomic_read(&sp->sk_refcnt), sp, + atomic_read(&sp->sk_drops), len); } int udp4_seq_show(struct seq_file *seq, void *v) @@ -1643,7 +1646,7 @@ int udp4_seq_show(struct seq_file *seq, void *v) seq_printf(seq, "%-127s\n", " sl local_address rem_address st tx_queue " "rx_queue tr tm->when retrnsmt uid timeout " - "inode"); + "inode ref pointer drops"); else { struct udp_iter_state *state = seq->private; int len; diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c index 456777d7a407..34cfb3f41c2c 100644 --- a/net/ipv6/raw.c +++ b/net/ipv6/raw.c @@ -1251,7 +1251,7 @@ static int raw6_seq_show(struct seq_file *seq, void *v) "local_address " "remote_address " "st tx_queue rx_queue tr tm->when retrnsmt" - " uid timeout inode drops\n"); + " uid timeout inode ref pointer drops\n"); else raw6_sock_seq_show(seq, v, raw_seq_private(seq)->bucket); return 0; diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index 432edaa882f6..f91e1df0d25e 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -297,8 +297,10 @@ int udpv6_queue_rcv_skb(struct sock * sk, struct sk_buff *skb) if ((rc = sock_queue_rcv_skb(sk,skb)) < 0) { /* Note that an ENOMEM error is charged twice */ - if (rc == -ENOMEM) + if (rc == -ENOMEM) { UDP6_INC_STATS_BH(UDP_MIB_RCVBUFERRORS, is_udplite); + atomic_inc(&sk->sk_drops); + } goto drop; } @@ -955,7 +957,7 @@ static void udp6_sock_seq_show(struct seq_file *seq, struct sock *sp, int bucket srcp = ntohs(inet->sport); seq_printf(seq, "%4d: %08X%08X%08X%08X:%04X %08X%08X%08X%08X:%04X " - "%02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p\n", + "%02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p %d\n", bucket, src->s6_addr32[0], src->s6_addr32[1], src->s6_addr32[2], src->s6_addr32[3], srcp, @@ -967,7 +969,8 @@ static void udp6_sock_seq_show(struct seq_file *seq, struct sock *sp, int bucket 0, 0L, 0, sock_i_uid(sp), 0, sock_i_ino(sp), - atomic_read(&sp->sk_refcnt), sp); + atomic_read(&sp->sk_refcnt), sp, + atomic_read(&sp->sk_drops)); } int udp6_seq_show(struct seq_file *seq, void *v) @@ -978,7 +981,7 @@ int udp6_seq_show(struct seq_file *seq, void *v) "local_address " "remote_address " "st tx_queue rx_queue tr tm->when retrnsmt" - " uid timeout inode\n"); + " uid timeout inode ref pointer drops\n"); else udp6_sock_seq_show(seq, v, ((struct udp_iter_state *)seq->private)->bucket); return 0; -- cgit v1.2.2 From 30902dc3cb0ea1cfc7ac2b17bcf478ff98420d74 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 21:26:37 -0700 Subject: ax25: Fix std timer socket destroy handling. Tihomir Heidelberg - 9a4gl, reports: -------------------- I would like to direct you attention to one problem existing in ax.25 kernel since 2.4. If listening socket is closed and its SKB queue is released but those sockets get weird. Those "unAccepted()" sockets should be destroyed in ax25_std_heartbeat_expiry, but it will not happen. And there is also a note about that in ax25_std_timer.c: /* Magic here: If we listen() and a new link dies before it is accepted() it isn't 'dead' so doesn't get removed. */ This issue cause ax25d to stop accepting new connections and I had to restarted ax25d approximately each day and my services were unavailable. Also netstat -n -l shows invalid source and device for those listening sockets. It is strange why ax25d's listening socket get weird because of this issue, but definitely when I solved this bug I do not have problems with ax25d anymore and my ax25d can run for months without problems. -------------------- Actually as far as I can see, this problem is even in releases as far back as 2.2.x as well. It seems senseless to special case this test on TCP_LISTEN state. Anything still stuck in state 0 has no external references and we can just simply kill it off directly. Signed-off-by: David S. Miller --- net/ax25/ax25_std_timer.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/ax25/ax25_std_timer.c b/net/ax25/ax25_std_timer.c index 96e4b9273250..cdc7e751ef36 100644 --- a/net/ax25/ax25_std_timer.c +++ b/net/ax25/ax25_std_timer.c @@ -39,11 +39,9 @@ void ax25_std_heartbeat_expiry(ax25_cb *ax25) switch (ax25->state) { case AX25_STATE_0: - /* Magic here: If we listen() and a new link dies before it - is accepted() it isn't 'dead' so doesn't get removed. */ - if (!sk || sock_flag(sk, SOCK_DESTROY) || - (sk->sk_state == TCP_LISTEN && - sock_flag(sk, SOCK_DEAD))) { + if (!sk || + sock_flag(sk, SOCK_DESTROY) || + sock_flag(sk, SOCK_DEAD)) { if (sk) { sock_hold(sk); ax25_destroy_socket(ax25); -- cgit v1.2.2 From 972692e0db9b0a62329ca394062b58917ddbd03c Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 17 Jun 2008 22:41:38 -0700 Subject: net: Add sk_set_socket() helper. In order to more easily grep for all things that set sk->sk_socket, add sk_set_socket() helper inline function. Suggested (although only half-seriously) by Evgeniy Polyakov. Signed-off-by: David S. Miller --- net/core/sock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/core/sock.c b/net/core/sock.c index 3879bf65897e..2c0ba52e5303 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -1066,7 +1066,7 @@ struct sock *sk_clone(const struct sock *sk, const gfp_t priority) * to be taken into account in all callers. -acme */ sk_refcnt_debug_inc(newsk); - newsk->sk_socket = NULL; + sk_set_socket(newsk, NULL); newsk->sk_sleep = NULL; if (newsk->sk_prot->sockets_allocated) @@ -1702,7 +1702,7 @@ void sock_init_data(struct socket *sock, struct sock *sk) sk->sk_rcvbuf = sysctl_rmem_default; sk->sk_sndbuf = sysctl_wmem_default; sk->sk_state = TCP_CLOSE; - sk->sk_socket = sock; + sk_set_socket(sk, sock); sock_set_flag(sk, SOCK_ZAPPED); -- cgit v1.2.2 From dd574dbfcc9e74e7dd8fd59ae0075d23e71a3da1 Mon Sep 17 00:00:00 2001 From: Rami Rosen Date: Wed, 18 Jun 2008 00:51:09 -0700 Subject: ipv6: minor cleanup in net/ipv6/tcp_ipv6.c [RESEND ]. In net/ipv6/tcp_ipv6.c: - Remove unneeded tcp_v6_send_check() declaration. Signed-off-by: Rami Rosen Signed-off-by: David S. Miller --- net/ipv6/tcp_ipv6.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'net') diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index daefc18d50a0..09be09cc1aa6 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -70,8 +70,6 @@ static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb); static void tcp_v6_reqsk_send_ack(struct sk_buff *skb, struct request_sock *req); -static void tcp_v6_send_check(struct sock *sk, int len, - struct sk_buff *skb); static int tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb); -- cgit v1.2.2 From dad9b335c6940de2746a9788eb456d09cf102f81 Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Wed, 18 Jun 2008 01:48:28 -0700 Subject: netdevice: Fix promiscuity and allmulti overflow Max of promiscuity and allmulti plus positive @inc can cause overflow. Fox example: when allmulti=0xFFFFFFFF, any caller give dev_set_allmulti() a positive @inc will cause allmulti be off. This is not what we want, though it's rare case. The fix is that only negative @inc will cause allmulti or promiscuity be off and when any caller makes the counters touch the roof, we return error. Change of v2: Change void function dev_set_promiscuity/allmulti to return int. So callers can get the overflow error. Caller's fix will be done later. Change of v3: 1. Since we return error to caller, we don't need to print KERN_ERROR, KERN_WARNING is enough. 2. In dev_set_promiscuity(), if __dev_set_promiscuity() failed, we return at once. Signed-off-by: Wang Chen Acked-by: Patrick McHardy Signed-off-by: David S. Miller --- net/core/dev.c | 55 +++++++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 45 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 0e45742e7158..a495f712d38c 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2771,16 +2771,29 @@ int netdev_set_master(struct net_device *slave, struct net_device *master) return 0; } -static void __dev_set_promiscuity(struct net_device *dev, int inc) +static int __dev_set_promiscuity(struct net_device *dev, int inc) { unsigned short old_flags = dev->flags; ASSERT_RTNL(); - if ((dev->promiscuity += inc) == 0) - dev->flags &= ~IFF_PROMISC; - else - dev->flags |= IFF_PROMISC; + dev->flags |= IFF_PROMISC; + dev->promiscuity += inc; + if (dev->promiscuity == 0) { + /* + * Avoid overflow. + * If inc causes overflow, untouch promisc and return error. + */ + if (inc < 0) + dev->flags &= ~IFF_PROMISC; + else { + dev->promiscuity -= inc; + printk(KERN_WARNING "%s: promiscuity touches roof, " + "set promiscuity failed, promiscuity feature " + "of device might be broken.\n", dev->name); + return -EOVERFLOW; + } + } if (dev->flags != old_flags) { printk(KERN_INFO "device %s %s promiscuous mode\n", dev->name, (dev->flags & IFF_PROMISC) ? "entered" : @@ -2798,6 +2811,7 @@ static void __dev_set_promiscuity(struct net_device *dev, int inc) if (dev->change_rx_flags) dev->change_rx_flags(dev, IFF_PROMISC); } + return 0; } /** @@ -2809,14 +2823,19 @@ static void __dev_set_promiscuity(struct net_device *dev, int inc) * remains above zero the interface remains promiscuous. Once it hits zero * the device reverts back to normal filtering operation. A negative inc * value is used to drop promiscuity on the device. + * Return 0 if successful or a negative errno code on error. */ -void dev_set_promiscuity(struct net_device *dev, int inc) +int dev_set_promiscuity(struct net_device *dev, int inc) { unsigned short old_flags = dev->flags; + int err; - __dev_set_promiscuity(dev, inc); + err = __dev_set_promiscuity(dev, inc); + if (!err) + return err; if (dev->flags != old_flags) dev_set_rx_mode(dev); + return err; } /** @@ -2829,22 +2848,38 @@ void dev_set_promiscuity(struct net_device *dev, int inc) * to all interfaces. Once it hits zero the device reverts back to normal * filtering operation. A negative @inc value is used to drop the counter * when releasing a resource needing all multicasts. + * Return 0 if successful or a negative errno code on error. */ -void dev_set_allmulti(struct net_device *dev, int inc) +int dev_set_allmulti(struct net_device *dev, int inc) { unsigned short old_flags = dev->flags; ASSERT_RTNL(); dev->flags |= IFF_ALLMULTI; - if ((dev->allmulti += inc) == 0) - dev->flags &= ~IFF_ALLMULTI; + dev->allmulti += inc; + if (dev->allmulti == 0) { + /* + * Avoid overflow. + * If inc causes overflow, untouch allmulti and return error. + */ + if (inc < 0) + dev->flags &= ~IFF_ALLMULTI; + else { + dev->allmulti -= inc; + printk(KERN_WARNING "%s: allmulti touches roof, " + "set allmulti failed, allmulti feature of " + "device might be broken.\n", dev->name); + return -EOVERFLOW; + } + } if (dev->flags ^ old_flags) { if (dev->change_rx_flags) dev->change_rx_flags(dev, IFF_ALLMULTI); dev_set_rx_mode(dev); } + return 0; } /* -- cgit v1.2.2 From 7115e632f90952454ab6426e0d2151327162a30f Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Thu, 19 Jun 2008 16:07:48 -0700 Subject: sctp: Validate Initiate Tag when handling ICMP message This patch add to validate initiate tag and chunk type if verification tag is 0 when handling ICMP message. RFC 4960, Appendix C. ICMP Handling ICMP6) An implementation MUST validate that the Verification Tag contained in the ICMP message matches the Verification Tag of the peer. If the Verification Tag is not 0 and does NOT match, discard the ICMP message. If it is 0 and the ICMP message contains enough bytes to verify that the chunk type is an INIT chunk and that the Initiate Tag matches the tag of the peer, continue with ICMP7. If the ICMP message is too short or the chunk type or the Initiate Tag does not match, silently discard the packet. Signed-off-by: Wei Yongjun Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/input.c | 27 +++++++++++++++++++++++++-- 1 file changed, 25 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/sctp/input.c b/net/sctp/input.c index ca6b022b1df2..d354a23972d4 100644 --- a/net/sctp/input.c +++ b/net/sctp/input.c @@ -430,6 +430,9 @@ struct sock *sctp_err_lookup(int family, struct sk_buff *skb, struct sock *sk = NULL; struct sctp_association *asoc; struct sctp_transport *transport = NULL; + struct sctp_init_chunk *chunkhdr; + __u32 vtag = ntohl(sctphdr->vtag); + int len = skb->len - ((void *)sctphdr - (void *)skb->data); *app = NULL; *tpp = NULL; @@ -451,8 +454,28 @@ struct sock *sctp_err_lookup(int family, struct sk_buff *skb, sk = asoc->base.sk; - if (ntohl(sctphdr->vtag) != asoc->c.peer_vtag) { - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + /* RFC 4960, Appendix C. ICMP Handling + * + * ICMP6) An implementation MUST validate that the Verification Tag + * contained in the ICMP message matches the Verification Tag of + * the peer. If the Verification Tag is not 0 and does NOT + * match, discard the ICMP message. If it is 0 and the ICMP + * message contains enough bytes to verify that the chunk type is + * an INIT chunk and that the Initiate Tag matches the tag of the + * peer, continue with ICMP7. If the ICMP message is too short + * or the chunk type or the Initiate Tag does not match, silently + * discard the packet. + */ + if (vtag == 0) { + chunkhdr = (struct sctp_init_chunk *)((void *)sctphdr + + sizeof(struct sctphdr)); + if (len < sizeof(struct sctphdr) + sizeof(sctp_chunkhdr_t) + + sizeof(__be32) || + chunkhdr->chunk_hdr.type != SCTP_CID_INIT || + ntohl(chunkhdr->init_hdr.init_tag) != asoc->c.my_vtag) { + goto out; + } + } else if (vtag != asoc->c.peer_vtag) { goto out; } -- cgit v1.2.2 From 2e3216cd54b142ba605e87522e15f42e0c4e3996 Mon Sep 17 00:00:00 2001 From: Vlad Yasevich Date: Thu, 19 Jun 2008 16:08:18 -0700 Subject: sctp: Follow security requirement of responding with 1 packet RFC 4960, Section 11.4. Protection of Non-SCTP-Capable Hosts When an SCTP stack receives a packet containing multiple control or DATA chunks and the processing of the packet requires the sending of multiple chunks in response, the sender of the response chunk(s) MUST NOT send more than one packet. If bundling is supported, multiple response chunks that fit into a single packet MAY be bundled together into one single response packet. If bundling is not supported, then the sender MUST NOT send more than one response chunk and MUST discard all other responses. Note that this rule does NOT apply to a SACK chunk, since a SACK chunk is, in itself, a response to DATA and a SACK does not require a response of more DATA. We implement this by not servicing our outqueue until we reach the end of the packet. This enables maximum bundling. We also identify 'response' chunks and make sure that we only send 1 packet when sending such chunks. Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/associola.c | 1 + net/sctp/output.c | 7 +++++-- net/sctp/outqueue.c | 34 ++++++++++++++++++++++++---------- net/sctp/sm_sideeffect.c | 27 ++++++++++++++++++--------- net/sctp/sm_statefuns.c | 16 +++++----------- 5 files changed, 53 insertions(+), 32 deletions(-) (limited to 'net') diff --git a/net/sctp/associola.c b/net/sctp/associola.c index 35b6a023a6d0..ff1dc5bd936c 100644 --- a/net/sctp/associola.c +++ b/net/sctp/associola.c @@ -1025,6 +1025,7 @@ static void sctp_assoc_bh_rcv(struct work_struct *work) struct sctp_chunk *chunk; struct sock *sk; struct sctp_inq *inqueue; + struct sctp_outq *outq; int state; sctp_subtype_t subtype; int error = 0; diff --git a/net/sctp/output.c b/net/sctp/output.c index 6d45bae93b46..abcd00dc05eb 100644 --- a/net/sctp/output.c +++ b/net/sctp/output.c @@ -157,7 +157,8 @@ void sctp_packet_free(struct sctp_packet *packet) * packet can be sent only after receiving the COOKIE_ACK. */ sctp_xmit_t sctp_packet_transmit_chunk(struct sctp_packet *packet, - struct sctp_chunk *chunk) + struct sctp_chunk *chunk, + int one_packet) { sctp_xmit_t retval; int error = 0; @@ -175,7 +176,9 @@ sctp_xmit_t sctp_packet_transmit_chunk(struct sctp_packet *packet, /* If we have an empty packet, then we can NOT ever * return PMTU_FULL. */ - retval = sctp_packet_append_chunk(packet, chunk); + if (!one_packet) + retval = sctp_packet_append_chunk(packet, + chunk); } break; diff --git a/net/sctp/outqueue.c b/net/sctp/outqueue.c index ace6770e9048..70ead8dc3485 100644 --- a/net/sctp/outqueue.c +++ b/net/sctp/outqueue.c @@ -702,6 +702,7 @@ int sctp_outq_uncork(struct sctp_outq *q) return error; } + /* * Try to flush an outqueue. * @@ -725,6 +726,7 @@ int sctp_outq_flush(struct sctp_outq *q, int rtx_timeout) sctp_xmit_t status; int error = 0; int start_timer = 0; + int one_packet = 0; /* These transports have chunks to send. */ struct list_head transport_list; @@ -830,20 +832,33 @@ int sctp_outq_flush(struct sctp_outq *q, int rtx_timeout) if (sctp_test_T_bit(chunk)) { packet->vtag = asoc->c.my_vtag; } - case SCTP_CID_SACK: - case SCTP_CID_HEARTBEAT: + /* The following chunks are "response" chunks, i.e. + * they are generated in response to something we + * received. If we are sending these, then we can + * send only 1 packet containing these chunks. + */ case SCTP_CID_HEARTBEAT_ACK: - case SCTP_CID_SHUTDOWN: case SCTP_CID_SHUTDOWN_ACK: - case SCTP_CID_ERROR: - case SCTP_CID_COOKIE_ECHO: case SCTP_CID_COOKIE_ACK: - case SCTP_CID_ECN_ECNE: + case SCTP_CID_COOKIE_ECHO: + case SCTP_CID_ERROR: case SCTP_CID_ECN_CWR: - case SCTP_CID_ASCONF: case SCTP_CID_ASCONF_ACK: + one_packet = 1; + /* Fall throught */ + + case SCTP_CID_SACK: + case SCTP_CID_HEARTBEAT: + case SCTP_CID_SHUTDOWN: + case SCTP_CID_ECN_ECNE: + case SCTP_CID_ASCONF: case SCTP_CID_FWD_TSN: - sctp_packet_transmit_chunk(packet, chunk); + status = sctp_packet_transmit_chunk(packet, chunk, + one_packet); + if (status != SCTP_XMIT_OK) { + /* put the chunk back */ + list_add(&chunk->list, &q->control_chunk_list); + } break; default: @@ -974,7 +989,7 @@ int sctp_outq_flush(struct sctp_outq *q, int rtx_timeout) atomic_read(&chunk->skb->users) : -1); /* Add the chunk to the packet. */ - status = sctp_packet_transmit_chunk(packet, chunk); + status = sctp_packet_transmit_chunk(packet, chunk, 0); switch (status) { case SCTP_XMIT_PMTU_FULL: @@ -1239,7 +1254,6 @@ int sctp_outq_sack(struct sctp_outq *q, struct sctp_sackhdr *sack) * Make sure the empty queue handler will get run later. */ q->empty = (list_empty(&q->out_chunk_list) && - list_empty(&q->control_chunk_list) && list_empty(&q->retransmit)); if (!q->empty) goto finish; diff --git a/net/sctp/sm_sideeffect.c b/net/sctp/sm_sideeffect.c index b083312c725a..9732c797e8ed 100644 --- a/net/sctp/sm_sideeffect.c +++ b/net/sctp/sm_sideeffect.c @@ -664,7 +664,7 @@ static int sctp_cmd_process_sack(sctp_cmd_seq_t *cmds, struct sctp_association *asoc, struct sctp_sackhdr *sackh) { - int err; + int err = 0; if (sctp_outq_sack(&asoc->outqueue, sackh)) { /* There are no more TSNs awaiting SACK. */ @@ -672,11 +672,6 @@ static int sctp_cmd_process_sack(sctp_cmd_seq_t *cmds, SCTP_ST_OTHER(SCTP_EVENT_NO_PENDING_TSN), asoc->state, asoc->ep, asoc, NULL, GFP_ATOMIC); - } else { - /* Windows may have opened, so we need - * to check if we have DATA to transmit - */ - err = sctp_outq_flush(&asoc->outqueue, 0); } return err; @@ -1481,8 +1476,15 @@ static int sctp_cmd_interpreter(sctp_event_t event_type, break; case SCTP_CMD_DISCARD_PACKET: - /* We need to discard the whole packet. */ + /* We need to discard the whole packet. + * Uncork the queue since there might be + * responses pending + */ chunk->pdiscard = 1; + if (asoc) { + sctp_outq_uncork(&asoc->outqueue); + local_cork = 0; + } break; case SCTP_CMD_RTO_PENDING: @@ -1553,8 +1555,15 @@ static int sctp_cmd_interpreter(sctp_event_t event_type, } out: - if (local_cork) - sctp_outq_uncork(&asoc->outqueue); + /* If this is in response to a received chunk, wait until + * we are done with the packet to open the queue so that we don't + * send multiple packets in response to a single request. + */ + if (asoc && SCTP_EVENT_T_CHUNK == event_type && chunk) { + if (chunk->end_of_packet || chunk->singleton) + sctp_outq_uncork(&asoc->outqueue); + } else if (local_cork) + sctp_outq_uncork(&asoc->outqueue); return error; nomem: error = -ENOMEM; diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c index 0c9d5a6950fe..b66a41d03c0d 100644 --- a/net/sctp/sm_statefuns.c +++ b/net/sctp/sm_statefuns.c @@ -795,8 +795,6 @@ sctp_disposition_t sctp_sf_do_5_1D_ce(const struct sctp_endpoint *ep, sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_START, SCTP_TO(SCTP_EVENT_TIMEOUT_AUTOCLOSE)); - sctp_add_cmd_sf(commands, SCTP_CMD_TRANSMIT, SCTP_NULL()); - /* This will send the COOKIE ACK */ sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(repl)); @@ -883,7 +881,6 @@ sctp_disposition_t sctp_sf_do_5_1E_ca(const struct sctp_endpoint *ep, if (asoc->autoclose) sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_START, SCTP_TO(SCTP_EVENT_TIMEOUT_AUTOCLOSE)); - sctp_add_cmd_sf(commands, SCTP_CMD_TRANSMIT, SCTP_NULL()); /* It may also notify its ULP about the successful * establishment of the association with a Communication Up @@ -1781,7 +1778,6 @@ static sctp_disposition_t sctp_sf_do_dupcook_b(const struct sctp_endpoint *ep, goto nomem; sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(repl)); - sctp_add_cmd_sf(commands, SCTP_CMD_TRANSMIT, SCTP_NULL()); /* RFC 2960 5.1 Normal Establishment of an Association * @@ -1898,12 +1894,13 @@ static sctp_disposition_t sctp_sf_do_dupcook_d(const struct sctp_endpoint *ep, } } - sctp_add_cmd_sf(commands, SCTP_CMD_TRANSMIT, SCTP_NULL()); repl = sctp_make_cookie_ack(new_asoc, chunk); if (!repl) goto nomem; + sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(repl)); + if (ev) sctp_add_cmd_sf(commands, SCTP_CMD_EVENT_ULP, SCTP_ULPEVENT(ev)); @@ -1911,9 +1908,6 @@ static sctp_disposition_t sctp_sf_do_dupcook_d(const struct sctp_endpoint *ep, sctp_add_cmd_sf(commands, SCTP_CMD_EVENT_ULP, SCTP_ULPEVENT(ai_ev)); - sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(repl)); - sctp_add_cmd_sf(commands, SCTP_CMD_TRANSMIT, SCTP_NULL()); - return SCTP_DISPOSITION_CONSUME; nomem: @@ -3970,9 +3964,6 @@ sctp_disposition_t sctp_sf_unk_chunk(const struct sctp_endpoint *ep, return sctp_sf_pdiscard(ep, asoc, type, arg, commands); break; case SCTP_CID_ACTION_DISCARD_ERR: - /* Discard the packet. */ - sctp_sf_pdiscard(ep, asoc, type, arg, commands); - /* Generate an ERROR chunk as response. */ hdr = unk_chunk->chunk_hdr; err_chunk = sctp_make_op_error(asoc, unk_chunk, @@ -3982,6 +3973,9 @@ sctp_disposition_t sctp_sf_unk_chunk(const struct sctp_endpoint *ep, sctp_add_cmd_sf(commands, SCTP_CMD_REPLY, SCTP_CHUNK(err_chunk)); } + + /* Discard the packet. */ + sctp_sf_pdiscard(ep, asoc, type, arg, commands); return SCTP_DISPOSITION_CONSUME; break; case SCTP_CID_ACTION_SKIP: -- cgit v1.2.2 From 0187bdfb05674147774ca79a79942537f3ad54bd Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Thu, 19 Jun 2008 16:15:47 -0700 Subject: net: Disable LRO on devices that are forwarding Large Receive Offload (LRO) is only appropriate for packets that are destined for the host, and should be disabled if received packets may be forwarded. It can also confuse the GSO on output. Add dev_disable_lro() function which uses the appropriate ethtool ops to disable LRO if enabled. Add calls to dev_disable_lro() in br_add_if() and functions that enable IPv4 and IPv6 forwarding. Signed-off-by: Ben Hutchings Signed-off-by: David S. Miller --- net/bridge/br_if.c | 1 + net/core/dev.c | 24 ++++++++++++++++++++++++ net/ipv4/devinet.c | 21 ++++++++++++++++----- net/ipv6/addrconf.c | 6 ++++++ 4 files changed, 47 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c index 143c954681b8..832a561500d9 100644 --- a/net/bridge/br_if.c +++ b/net/bridge/br_if.c @@ -387,6 +387,7 @@ int br_add_if(struct net_bridge *br, struct net_device *dev) goto err2; rcu_assign_pointer(dev->br_port, p); + dev_disable_lro(dev); dev_set_promiscuity(dev, 1); list_add_rcu(&p->list, &br->port_list); diff --git a/net/core/dev.c b/net/core/dev.c index a495f712d38c..f6944ecd5b2e 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -90,6 +90,7 @@ #include #include #include +#include #include #include #include @@ -1123,6 +1124,29 @@ int dev_close(struct net_device *dev) } +/** + * dev_disable_lro - disable Large Receive Offload on a device + * @dev: device + * + * Disable Large Receive Offload (LRO) on a net device. Must be + * called under RTNL. This is needed if received packets may be + * forwarded to another interface. + */ +void dev_disable_lro(struct net_device *dev) +{ + if (dev->ethtool_ops && dev->ethtool_ops->get_flags && + dev->ethtool_ops->set_flags) { + u32 flags = dev->ethtool_ops->get_flags(dev); + if (flags & ETH_FLAG_LRO) { + flags &= ~ETH_FLAG_LRO; + dev->ethtool_ops->set_flags(dev, flags); + } + } + WARN_ON(dev->features & NETIF_F_LRO); +} +EXPORT_SYMBOL(dev_disable_lro); + + static int dev_boot_phase = 1; /* diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c index f8c0b0aea93a..9de2514946ca 100644 --- a/net/ipv4/devinet.c +++ b/net/ipv4/devinet.c @@ -168,6 +168,8 @@ static struct in_device *inetdev_init(struct net_device *dev) in_dev->dev = dev; if ((in_dev->arp_parms = neigh_parms_alloc(dev, &arp_tbl)) == NULL) goto out_kfree; + if (IPV4_DEVCONF(in_dev->cnf, FORWARDING)) + dev_disable_lro(dev); /* Reference in_dev->dev */ dev_hold(dev); /* Account for reference dev->ip_ptr (below) */ @@ -1241,6 +1243,8 @@ static void inet_forward_change(struct net *net) read_lock(&dev_base_lock); for_each_netdev(net, dev) { struct in_device *in_dev; + if (on) + dev_disable_lro(dev); rcu_read_lock(); in_dev = __in_dev_get_rcu(dev); if (in_dev) @@ -1248,8 +1252,6 @@ static void inet_forward_change(struct net *net) rcu_read_unlock(); } read_unlock(&dev_base_lock); - - rt_cache_flush(0); } static int devinet_conf_proc(ctl_table *ctl, int write, @@ -1335,10 +1337,19 @@ static int devinet_sysctl_forward(ctl_table *ctl, int write, if (write && *valp != val) { struct net *net = ctl->extra2; - if (valp == &IPV4_DEVCONF_ALL(net, FORWARDING)) - inet_forward_change(net); - else if (valp != &IPV4_DEVCONF_DFLT(net, FORWARDING)) + if (valp != &IPV4_DEVCONF_DFLT(net, FORWARDING)) { + rtnl_lock(); + if (valp == &IPV4_DEVCONF_ALL(net, FORWARDING)) { + inet_forward_change(net); + } else if (*valp) { + struct ipv4_devconf *cnf = ctl->extra1; + struct in_device *idev = + container_of(cnf, struct in_device, cnf); + dev_disable_lro(idev->dev); + } + rtnl_unlock(); rt_cache_flush(0); + } } return ret; diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 9be6be3a7ff3..84127d854cfc 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -348,6 +348,8 @@ static struct inet6_dev * ipv6_add_dev(struct net_device *dev) kfree(ndev); return NULL; } + if (ndev->cnf.forwarding) + dev_disable_lro(dev); /* We refer to the device */ dev_hold(dev); @@ -442,6 +444,8 @@ static void dev_forward_change(struct inet6_dev *idev) if (!idev) return; dev = idev->dev; + if (idev->cnf.forwarding) + dev_disable_lro(dev); if (dev && (dev->flags & IFF_MULTICAST)) { if (idev->cnf.forwarding) ipv6_dev_mc_inc(dev, &in6addr_linklocal_allrouters); @@ -487,12 +491,14 @@ static void addrconf_fixup_forwarding(struct ctl_table *table, int *p, int old) if (p == &net->ipv6.devconf_dflt->forwarding) return; + rtnl_lock(); if (p == &net->ipv6.devconf_all->forwarding) { __s32 newf = net->ipv6.devconf_all->forwarding; net->ipv6.devconf_dflt->forwarding = newf; addrconf_forward_change(net, newf); } else if ((!*p) ^ (!old)) dev_forward_change((struct inet6_dev *)table->extra1); + rtnl_unlock(); if (*p) rt6_purge_dflt_routers(net); -- cgit v1.2.2 From 4497b0763cb1afae463f5e144c28b5d806e28b60 Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Thu, 19 Jun 2008 16:22:28 -0700 Subject: net: Discard and warn about LRO'd skbs received for forwarding Add skb_warn_if_lro() to test whether an skb was received with LRO and warn if so. Change br_forward(), ip_forward() and ip6_forward() to call it) and discard the skb if it returns true. Signed-off-by: Ben Hutchings Signed-off-by: David S. Miller --- net/bridge/br_forward.c | 2 +- net/core/skbuff.c | 8 ++++++++ net/ipv4/ip_forward.c | 3 +++ net/ipv6/ip6_output.c | 3 +++ 4 files changed, 15 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/bridge/br_forward.c b/net/bridge/br_forward.c index 512645727f51..bdd9ccea17ce 100644 --- a/net/bridge/br_forward.c +++ b/net/bridge/br_forward.c @@ -89,7 +89,7 @@ void br_deliver(const struct net_bridge_port *to, struct sk_buff *skb) /* called with rcu_read_lock */ void br_forward(const struct net_bridge_port *to, struct sk_buff *skb) { - if (should_deliver(to, skb)) { + if (!skb_warn_if_lro(skb) && should_deliver(to, skb)) { __br_forward(to, skb); return; } diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 3e18f8525e82..2df012be973d 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -2583,6 +2583,13 @@ bool skb_partial_csum_set(struct sk_buff *skb, u16 start, u16 off) return true; } +void __skb_warn_lro_forwarding(const struct sk_buff *skb) +{ + if (net_ratelimit()) + pr_warning("%s: received packets cannot be forwarded" + " while LRO is enabled\n", skb->dev->name); +} + EXPORT_SYMBOL(___pskb_trim); EXPORT_SYMBOL(__kfree_skb); EXPORT_SYMBOL(kfree_skb); @@ -2616,6 +2623,7 @@ EXPORT_SYMBOL(skb_seq_read); EXPORT_SYMBOL(skb_abort_seq_read); EXPORT_SYMBOL(skb_find_text); EXPORT_SYMBOL(skb_append_datato_frags); +EXPORT_SYMBOL(__skb_warn_lro_forwarding); EXPORT_SYMBOL_GPL(skb_to_sgvec); EXPORT_SYMBOL_GPL(skb_cow_data); diff --git a/net/ipv4/ip_forward.c b/net/ipv4/ip_forward.c index 37d36a3f33cd..da14725916d3 100644 --- a/net/ipv4/ip_forward.c +++ b/net/ipv4/ip_forward.c @@ -56,6 +56,9 @@ int ip_forward(struct sk_buff *skb) struct rtable *rt; /* Route we use */ struct ip_options * opt = &(IPCB(skb)->opt); + if (skb_warn_if_lro(skb)) + goto drop; + if (!xfrm4_policy_check(NULL, XFRM_POLICY_FWD, skb)) goto drop; diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 40a2813a63d1..fd7cd1bfe151 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -407,6 +407,9 @@ int ip6_forward(struct sk_buff *skb) if (ipv6_devconf.forwarding == 0) goto error; + if (skb_warn_if_lro(skb)) + goto drop; + if (!xfrm6_policy_check(NULL, XFRM_POLICY_FWD, skb)) { IP6_INC_STATS(ip6_dst_idev(dst), IPSTATS_MIB_INDISCARDS); goto drop; -- cgit v1.2.2 From 0f474d9bc59e23f1952e75d2f6edcaeb358eb32d Mon Sep 17 00:00:00 2001 From: Vlad Yasevich Date: Fri, 20 Jun 2008 10:34:47 -0700 Subject: sctp: Kill unused variable in sctp_assoc_bh_rcv() Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/associola.c | 1 - 1 file changed, 1 deletion(-) (limited to 'net') diff --git a/net/sctp/associola.c b/net/sctp/associola.c index ff1dc5bd936c..35b6a023a6d0 100644 --- a/net/sctp/associola.c +++ b/net/sctp/associola.c @@ -1025,7 +1025,6 @@ static void sctp_assoc_bh_rcv(struct work_struct *work) struct sctp_chunk *chunk; struct sock *sk; struct sctp_inq *inqueue; - struct sctp_outq *outq; int state; sctp_subtype_t subtype; int error = 0; -- cgit v1.2.2 From c8fcd905a59a535bff93a120ac44b09ce24e13e6 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 23 Jun 2008 17:22:57 -0300 Subject: rfkill: fix minor typo in kernel doc Fix a minor typo in an exported function documentation Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Cc: Dmitry Torokhov Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index 4e10a95de832..f95081a4a024 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -412,7 +412,7 @@ int rfkill_register(struct rfkill *rfkill) EXPORT_SYMBOL(rfkill_register); /** - * rfkill_unregister - Uegister a rfkill structure. + * rfkill_unregister - Unregister a rfkill structure. * @rfkill: rfkill structure to be unregistered * * This function should be called by the network driver during device -- cgit v1.2.2 From 28f089c18464810ec9e91ee10a89adbb02ad7765 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 23 Jun 2008 17:22:58 -0300 Subject: rfkill: handle SW_RFKILL_ALL events Teach rfkill-input how to handle SW_RFKILL_ALL events (new name for the SW_RADIO event). SW_RFKILL_ALL is an absolute enable-or-disable command that is tied to all radios in a system. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Cc: Dmitry Torokhov Signed-off-by: John W. Linville --- net/rfkill/rfkill-input.c | 45 +++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 43 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill-input.c b/net/rfkill/rfkill-input.c index e4b051dbed61..9d6c9255bf2c 100644 --- a/net/rfkill/rfkill-input.c +++ b/net/rfkill/rfkill-input.c @@ -55,6 +55,22 @@ static void rfkill_task_handler(struct work_struct *work) mutex_unlock(&task->mutex); } +static void rfkill_schedule_set(struct rfkill_task *task, + enum rfkill_state desired_state) +{ + unsigned long flags; + + spin_lock_irqsave(&task->lock, flags); + + if (time_after(jiffies, task->last + msecs_to_jiffies(200))) { + task->desired_state = desired_state; + task->last = jiffies; + schedule_work(&task->work); + } + + spin_unlock_irqrestore(&task->lock, flags); +} + static void rfkill_schedule_toggle(struct rfkill_task *task) { unsigned long flags; @@ -87,9 +103,9 @@ static DEFINE_RFKILL_TASK(rfkill_uwb, RFKILL_TYPE_UWB); static DEFINE_RFKILL_TASK(rfkill_wimax, RFKILL_TYPE_WIMAX); static void rfkill_event(struct input_handle *handle, unsigned int type, - unsigned int code, int down) + unsigned int code, int data) { - if (type == EV_KEY && down == 1) { + if (type == EV_KEY && data == 1) { switch (code) { case KEY_WLAN: rfkill_schedule_toggle(&rfkill_wlan); @@ -106,6 +122,26 @@ static void rfkill_event(struct input_handle *handle, unsigned int type, default: break; } + } else if (type == EV_SW) { + switch (code) { + case SW_RFKILL_ALL: + /* EVERY radio type. data != 0 means radios ON */ + rfkill_schedule_set(&rfkill_wimax, + (data)? RFKILL_STATE_ON: + RFKILL_STATE_OFF); + rfkill_schedule_set(&rfkill_uwb, + (data)? RFKILL_STATE_ON: + RFKILL_STATE_OFF); + rfkill_schedule_set(&rfkill_bt, + (data)? RFKILL_STATE_ON: + RFKILL_STATE_OFF); + rfkill_schedule_set(&rfkill_wlan, + (data)? RFKILL_STATE_ON: + RFKILL_STATE_OFF); + break; + default: + break; + } } } @@ -168,6 +204,11 @@ static const struct input_device_id rfkill_ids[] = { .evbit = { BIT_MASK(EV_KEY) }, .keybit = { [BIT_WORD(KEY_WIMAX)] = BIT_MASK(KEY_WIMAX) }, }, + { + .flags = INPUT_DEVICE_ID_MATCH_EVBIT | INPUT_DEVICE_ID_MATCH_SWBIT, + .evbit = { BIT(EV_SW) }, + .swbit = { [BIT_WORD(SW_RFKILL_ALL)] = BIT_MASK(SW_RFKILL_ALL) }, + }, { } }; -- cgit v1.2.2 From e954b0b85b9e737564b8ad9738de5816747b5901 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 23 Jun 2008 17:22:59 -0300 Subject: rfkill: add parameter to disable radios by default Currently, radios are always enabled when their rfkill interface is registered. This is not optimal, the safest state for a radio is to be offline unless the user turns it on. Add a module parameter that causes all radios to be disabled when their rfkill interface is registered. The module default is not changed so unless the parameter is used, radios will still be forced to their enabled state when they are registered. The new rfkill module parameter is called "default_state". Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index f95081a4a024..3edc585dcfa6 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -39,6 +39,11 @@ MODULE_LICENSE("GPL"); static LIST_HEAD(rfkill_list); /* list of registered rf switches */ static DEFINE_MUTEX(rfkill_mutex); +static unsigned int rfkill_default_state = RFKILL_STATE_ON; +module_param_named(default_state, rfkill_default_state, uint, 0444); +MODULE_PARM_DESC(default_state, + "Default initial state for all radio types, 0 = radio off"); + static enum rfkill_state rfkill_states[RFKILL_TYPE_MAX]; @@ -436,8 +441,12 @@ static int __init rfkill_init(void) int error; int i; + if (rfkill_default_state != RFKILL_STATE_OFF && + rfkill_default_state != RFKILL_STATE_ON) + return -EINVAL; + for (i = 0; i < ARRAY_SIZE(rfkill_states); i++) - rfkill_states[i] = RFKILL_STATE_ON; + rfkill_states[i] = rfkill_default_state; error = class_register(&rfkill_class); if (error) { -- cgit v1.2.2 From 801e49af4c1a9b988ba0d25de2b368c99c3bf2b3 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 23 Jun 2008 17:23:00 -0300 Subject: rfkill: add read-write rfkill switch support Currently, rfkill support for read/write rfkill switches is hacked through a round-trip over the input layer and rfkill-input to let a driver sync rfkill->state to hardware changes. This is buggy and sub-optimal. It causes real problems. It is best to think of the rfkill class as supporting only write-only switches at the moment. In order to implement the read/write functionality properly: Add a get_state() hook that is called by the class every time it needs to fetch the current state of the switch. Add a call to this hook every time the *current* state of the radio plays a role in a decision. Also add a force_state() method that can be used to forcefully syncronize the class' idea of the current state of the switch. This allows for a faster implementation of the read/write functionality, as a driver which get events on switch changes can avoid the need for a get_state() hook. If the get_state() hook is left as NULL, current behaviour is maintained, so this change is fully backwards compatible with the current rfkill drivers. For hardware that issues events when the rfkill state changes, leave get_state() NULL in the rfkill struct, set the initial state properly before registering with the rfkill class, and use the force_state() method in the driver to keep the rfkill interface up-to-date. get_state() can be called by the class from atomic context. It must not sleep. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Cc: Dmitry Torokhov Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 53 ++++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 50 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index 3edc585dcfa6..4ae4486c77ea 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -62,19 +62,39 @@ static void rfkill_led_trigger(struct rfkill *rfkill, #endif /* CONFIG_RFKILL_LEDS */ } +static void update_rfkill_state(struct rfkill *rfkill) +{ + enum rfkill_state newstate; + + if (rfkill->get_state) { + mutex_lock(&rfkill->mutex); + if (!rfkill->get_state(rfkill->data, &newstate)) + rfkill->state = newstate; + mutex_unlock(&rfkill->mutex); + } +} + static int rfkill_toggle_radio(struct rfkill *rfkill, enum rfkill_state state) { int retval = 0; + enum rfkill_state oldstate, newstate; + + oldstate = rfkill->state; + + if (rfkill->get_state && + !rfkill->get_state(rfkill->data, &newstate)) + rfkill->state = newstate; if (state != rfkill->state) { retval = rfkill->toggle_radio(rfkill->data, state); - if (!retval) { + if (!retval) rfkill->state = state; - rfkill_led_trigger(rfkill, state); - } } + if (rfkill->state != oldstate) + rfkill_led_trigger(rfkill, rfkill->state); + return retval; } @@ -105,6 +125,32 @@ void rfkill_switch_all(enum rfkill_type type, enum rfkill_state state) } EXPORT_SYMBOL(rfkill_switch_all); +/** + * rfkill_force_state - Force the internal rfkill radio state + * @rfkill: pointer to the rfkill class to modify. + * @state: the current radio state the class should be forced to. + * + * This function updates the internal state of the radio cached + * by the rfkill class. It should be used when the driver gets + * a notification by the firmware/hardware of the current *real* + * state of the radio rfkill switch. + * + * It may not be called from an atomic context. + */ +int rfkill_force_state(struct rfkill *rfkill, enum rfkill_state state) +{ + if (state != RFKILL_STATE_OFF && + state != RFKILL_STATE_ON) + return -EINVAL; + + mutex_lock(&rfkill->mutex); + rfkill->state = state; + mutex_unlock(&rfkill->mutex); + + return 0; +} +EXPORT_SYMBOL(rfkill_force_state); + static ssize_t rfkill_name_show(struct device *dev, struct device_attribute *attr, char *buf) @@ -147,6 +193,7 @@ static ssize_t rfkill_state_show(struct device *dev, { struct rfkill *rfkill = to_rfkill(dev); + update_rfkill_state(rfkill); return sprintf(buf, "%d\n", rfkill->state); } -- cgit v1.2.2 From 477576a073699783abb53ae14993d5d41c66301d Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 23 Jun 2008 17:23:01 -0300 Subject: rfkill: add the WWAN radio type MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Unfortunately, instead of adding a generic Wireless WAN type, a technology- specific type (WiMAX) was added. That's useless for other WWAN devices, such as EDGE, UMTS, X-RTT and other such radios. Add a WWAN rfkill type for generic wireless WAN devices. No keys are added as most devices really want to use KEY_WLAN for WWAN control (in a cycle of none, WLAN, WWAN, WLAN+WWAN) and need no specific keycode added. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Cc: Iñaky Pérez-González Cc: David S. Miller Signed-off-by: John W. Linville --- net/rfkill/rfkill-input.c | 4 ++++ net/rfkill/rfkill.c | 3 +++ 2 files changed, 7 insertions(+) (limited to 'net') diff --git a/net/rfkill/rfkill-input.c b/net/rfkill/rfkill-input.c index 9d6c9255bf2c..29c13d308b31 100644 --- a/net/rfkill/rfkill-input.c +++ b/net/rfkill/rfkill-input.c @@ -101,6 +101,7 @@ static DEFINE_RFKILL_TASK(rfkill_wlan, RFKILL_TYPE_WLAN); static DEFINE_RFKILL_TASK(rfkill_bt, RFKILL_TYPE_BLUETOOTH); static DEFINE_RFKILL_TASK(rfkill_uwb, RFKILL_TYPE_UWB); static DEFINE_RFKILL_TASK(rfkill_wimax, RFKILL_TYPE_WIMAX); +static DEFINE_RFKILL_TASK(rfkill_wwan, RFKILL_TYPE_WWAN); static void rfkill_event(struct input_handle *handle, unsigned int type, unsigned int code, int data) @@ -126,6 +127,9 @@ static void rfkill_event(struct input_handle *handle, unsigned int type, switch (code) { case SW_RFKILL_ALL: /* EVERY radio type. data != 0 means radios ON */ + rfkill_schedule_set(&rfkill_wwan, + (data)? RFKILL_STATE_ON: + RFKILL_STATE_OFF); rfkill_schedule_set(&rfkill_wimax, (data)? RFKILL_STATE_ON: RFKILL_STATE_OFF); diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index 4ae4486c77ea..79f3bbb027ff 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -180,6 +180,9 @@ static ssize_t rfkill_type_show(struct device *dev, case RFKILL_TYPE_WIMAX: type = "wimax"; break; + case RFKILL_TYPE_WWAN: + type = "wwan"; + break; default: BUG(); } -- cgit v1.2.2 From 526324b61a9667ed9a71f0a8a8899cf675346c76 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 23 Jun 2008 17:23:02 -0300 Subject: rfkill: rework suspend and resume handlers The resume handler should reset the wireless transmitter rfkill state to exactly what it was when the system was suspended. Do it, and do it using the normal routines for state change while at it. The suspend handler should force-switch the transmitter to blocked state, ignoring caches. Do it. Also take an opportunity shot to rfkill_remove_switch() and also force the transmitter to blocked state there, bypassing caches. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 35 ++++++++++++++++++----------------- 1 file changed, 18 insertions(+), 17 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index 79f3bbb027ff..fb566902030a 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -75,24 +75,25 @@ static void update_rfkill_state(struct rfkill *rfkill) } static int rfkill_toggle_radio(struct rfkill *rfkill, - enum rfkill_state state) + enum rfkill_state state, + int force) { int retval = 0; enum rfkill_state oldstate, newstate; oldstate = rfkill->state; - if (rfkill->get_state && + if (rfkill->get_state && !force && !rfkill->get_state(rfkill->data, &newstate)) rfkill->state = newstate; - if (state != rfkill->state) { + if (force || state != rfkill->state) { retval = rfkill->toggle_radio(rfkill->data, state); if (!retval) rfkill->state = state; } - if (rfkill->state != oldstate) + if (force || rfkill->state != oldstate) rfkill_led_trigger(rfkill, rfkill->state); return retval; @@ -107,7 +108,6 @@ static int rfkill_toggle_radio(struct rfkill *rfkill, * a specific switch is claimed by userspace in which case it is * left alone. */ - void rfkill_switch_all(enum rfkill_type type, enum rfkill_state state) { struct rfkill *rfkill; @@ -118,7 +118,7 @@ void rfkill_switch_all(enum rfkill_type type, enum rfkill_state state) list_for_each_entry(rfkill, &rfkill_list, node) { if ((!rfkill->user_claim) && (rfkill->type == type)) - rfkill_toggle_radio(rfkill, state); + rfkill_toggle_radio(rfkill, state, 0); } mutex_unlock(&rfkill_mutex); @@ -214,7 +214,8 @@ static ssize_t rfkill_state_store(struct device *dev, if (mutex_lock_interruptible(&rfkill->mutex)) return -ERESTARTSYS; error = rfkill_toggle_radio(rfkill, - state ? RFKILL_STATE_ON : RFKILL_STATE_OFF); + state ? RFKILL_STATE_ON : RFKILL_STATE_OFF, + 0); mutex_unlock(&rfkill->mutex); return error ? error : count; @@ -255,7 +256,8 @@ static ssize_t rfkill_claim_store(struct device *dev, if (rfkill->user_claim != claim) { if (!claim) rfkill_toggle_radio(rfkill, - rfkill_states[rfkill->type]); + rfkill_states[rfkill->type], + 0); rfkill->user_claim = claim; } @@ -288,12 +290,11 @@ static int rfkill_suspend(struct device *dev, pm_message_t state) if (dev->power.power_state.event != state.event) { if (state.event & PM_EVENT_SLEEP) { - mutex_lock(&rfkill->mutex); - - if (rfkill->state == RFKILL_STATE_ON) - rfkill->toggle_radio(rfkill->data, - RFKILL_STATE_OFF); + /* Stop transmitter, keep state, no notifies */ + update_rfkill_state(rfkill); + mutex_lock(&rfkill->mutex); + rfkill->toggle_radio(rfkill->data, RFKILL_STATE_OFF); mutex_unlock(&rfkill->mutex); } @@ -310,8 +311,8 @@ static int rfkill_resume(struct device *dev) if (dev->power.power_state.event != PM_EVENT_ON) { mutex_lock(&rfkill->mutex); - if (rfkill->state == RFKILL_STATE_ON) - rfkill->toggle_radio(rfkill->data, RFKILL_STATE_ON); + /* restore radio state AND notify everybody */ + rfkill_toggle_radio(rfkill, rfkill->state, 1); mutex_unlock(&rfkill->mutex); } @@ -338,7 +339,7 @@ static int rfkill_add_switch(struct rfkill *rfkill) mutex_lock(&rfkill_mutex); - error = rfkill_toggle_radio(rfkill, rfkill_states[rfkill->type]); + error = rfkill_toggle_radio(rfkill, rfkill_states[rfkill->type], 0); if (!error) list_add_tail(&rfkill->node, &rfkill_list); @@ -351,7 +352,7 @@ static void rfkill_remove_switch(struct rfkill *rfkill) { mutex_lock(&rfkill_mutex); list_del_init(&rfkill->node); - rfkill_toggle_radio(rfkill, RFKILL_STATE_OFF); + rfkill_toggle_radio(rfkill, RFKILL_STATE_OFF, 1); mutex_unlock(&rfkill_mutex); } -- cgit v1.2.2 From 79399a8d1908f6a406e82d23c5a9937e1722ed3a Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 23 Jun 2008 17:23:03 -0300 Subject: rfkill: add notifier chains support Add a notifier chain for use by the rfkill class. This notifier chain signals the following events (more to be added when needed): 1. rfkill: rfkill device state has changed A pointer to the rfkill struct will be passed as a parameter. The notifier message types have been added to include/linux/rfkill.h instead of to include/linux/notifier.h in order to avoid the madness of modifying a header used globally (and that triggers an almost full tree rebuild every time it is touched) with information that is of interest only to code that includes the rfkill.h header. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 70 ++++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 67 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index fb566902030a..a561e350a70a 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -46,6 +46,49 @@ MODULE_PARM_DESC(default_state, static enum rfkill_state rfkill_states[RFKILL_TYPE_MAX]; +static BLOCKING_NOTIFIER_HEAD(rfkill_notifier_list); + + +/** + * register_rfkill_notifier - Add notifier to rfkill notifier chain + * @nb: pointer to the new entry to add to the chain + * + * See blocking_notifier_chain_register() for return value and further + * observations. + * + * Adds a notifier to the rfkill notifier chain. The chain will be + * called with a pointer to the relevant rfkill structure as a parameter, + * refer to include/linux/rfkill.h for the possible events. + * + * Notifiers added to this chain are to always return NOTIFY_DONE. This + * chain is a blocking notifier chain: notifiers can sleep. + * + * Calls to this chain may have been done through a workqueue. One must + * assume unordered asynchronous behaviour, there is no way to know if + * actions related to the event that generated the notification have been + * carried out already. + */ +int register_rfkill_notifier(struct notifier_block *nb) +{ + return blocking_notifier_chain_register(&rfkill_notifier_list, nb); +} +EXPORT_SYMBOL_GPL(register_rfkill_notifier); + +/** + * unregister_rfkill_notifier - remove notifier from rfkill notifier chain + * @nb: pointer to the entry to remove from the chain + * + * See blocking_notifier_chain_unregister() for return value and further + * observations. + * + * Removes a notifier from the rfkill notifier chain. + */ +int unregister_rfkill_notifier(struct notifier_block *nb) +{ + return blocking_notifier_chain_unregister(&rfkill_notifier_list, nb); +} +EXPORT_SYMBOL_GPL(unregister_rfkill_notifier); + static void rfkill_led_trigger(struct rfkill *rfkill, enum rfkill_state state) @@ -62,14 +105,25 @@ static void rfkill_led_trigger(struct rfkill *rfkill, #endif /* CONFIG_RFKILL_LEDS */ } +static void notify_rfkill_state_change(struct rfkill *rfkill) +{ + blocking_notifier_call_chain(&rfkill_notifier_list, + RFKILL_STATE_CHANGED, + rfkill); +} + static void update_rfkill_state(struct rfkill *rfkill) { - enum rfkill_state newstate; + enum rfkill_state newstate, oldstate; if (rfkill->get_state) { mutex_lock(&rfkill->mutex); - if (!rfkill->get_state(rfkill->data, &newstate)) + if (!rfkill->get_state(rfkill->data, &newstate)) { + oldstate = rfkill->state; rfkill->state = newstate; + if (oldstate != newstate) + notify_rfkill_state_change(rfkill); + } mutex_unlock(&rfkill->mutex); } } @@ -93,8 +147,10 @@ static int rfkill_toggle_radio(struct rfkill *rfkill, rfkill->state = state; } - if (force || rfkill->state != oldstate) + if (force || rfkill->state != oldstate) { rfkill_led_trigger(rfkill, rfkill->state); + notify_rfkill_state_change(rfkill); + } return retval; } @@ -139,12 +195,20 @@ EXPORT_SYMBOL(rfkill_switch_all); */ int rfkill_force_state(struct rfkill *rfkill, enum rfkill_state state) { + enum rfkill_state oldstate; + if (state != RFKILL_STATE_OFF && state != RFKILL_STATE_ON) return -EINVAL; mutex_lock(&rfkill->mutex); + + oldstate = rfkill->state; rfkill->state = state; + + if (state != oldstate) + notify_rfkill_state_change(rfkill); + mutex_unlock(&rfkill->mutex); return 0; -- cgit v1.2.2 From 99c632e5a304e1f76350eb9e8b2493514de8b60c Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 23 Jun 2008 17:23:04 -0300 Subject: rfkill: add type string helper We will need access to the rfkill switch type in string format for more than just sysfs. Therefore, move it to a generic helper. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 33 +++++++++++++++------------------ 1 file changed, 15 insertions(+), 18 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index a561e350a70a..3c7773475ea6 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -224,34 +224,31 @@ static ssize_t rfkill_name_show(struct device *dev, return sprintf(buf, "%s\n", rfkill->name); } -static ssize_t rfkill_type_show(struct device *dev, - struct device_attribute *attr, - char *buf) +static const char *rfkill_get_type_str(enum rfkill_type type) { - struct rfkill *rfkill = to_rfkill(dev); - const char *type; - - switch (rfkill->type) { + switch (type) { case RFKILL_TYPE_WLAN: - type = "wlan"; - break; + return "wlan"; case RFKILL_TYPE_BLUETOOTH: - type = "bluetooth"; - break; + return "bluetooth"; case RFKILL_TYPE_UWB: - type = "ultrawideband"; - break; + return "ultrawideband"; case RFKILL_TYPE_WIMAX: - type = "wimax"; - break; + return "wimax"; case RFKILL_TYPE_WWAN: - type = "wwan"; - break; + return "wwan"; default: BUG(); } +} + +static ssize_t rfkill_type_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct rfkill *rfkill = to_rfkill(dev); - return sprintf(buf, "%s\n", type); + return sprintf(buf, "%s\n", rfkill_get_type_str(rfkill->type)); } static ssize_t rfkill_state_show(struct device *dev, -- cgit v1.2.2 From ffb67c34e436fb163c4067936ccec797354fa6c6 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 23 Jun 2008 17:23:05 -0300 Subject: rfkill: add uevent notifications Use the notification chains to also send uevents, so that userspace can be notified of state changes of every rfkill switch. Userspace should use these events for OSD/status report applications and rfkill GUI frontends. HAL might want to broadcast them over DBUS, for example. It might be also useful for userspace implementations of rfkill-input, or to use HAL as the platform driver which promotes rfkill switch change events into input events (to synchronize all other switches) when necessary for platforms that lack a convenient platform-specific kernel module to do it. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Cc: Dmitry Torokhov Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 42 ++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 42 insertions(+) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index 3c7773475ea6..dd1c3f18f31d 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -386,12 +386,51 @@ static int rfkill_resume(struct device *dev) #define rfkill_resume NULL #endif +static int rfkill_blocking_uevent_notifier(struct notifier_block *nb, + unsigned long eventid, + void *data) +{ + struct rfkill *rfkill = (struct rfkill *)data; + + switch (eventid) { + case RFKILL_STATE_CHANGED: + kobject_uevent(&rfkill->dev.kobj, KOBJ_CHANGE); + break; + default: + break; + } + + return NOTIFY_DONE; +} + +static struct notifier_block rfkill_blocking_uevent_nb = { + .notifier_call = rfkill_blocking_uevent_notifier, + .priority = 0, +}; + +static int rfkill_dev_uevent(struct device *dev, struct kobj_uevent_env *env) +{ + struct rfkill *rfkill = to_rfkill(dev); + int error; + + error = add_uevent_var(env, "RFKILL_NAME=%s", rfkill->name); + if (error) + return error; + error = add_uevent_var(env, "RFKILL_TYPE=%s", + rfkill_get_type_str(rfkill->type)); + if (error) + return error; + error = add_uevent_var(env, "RFKILL_STATE=%d", rfkill->state); + return error; +} + static struct class rfkill_class = { .name = "rfkill", .dev_release = rfkill_release, .dev_attrs = rfkill_dev_attrs, .suspend = rfkill_suspend, .resume = rfkill_resume, + .dev_uevent = rfkill_dev_uevent, }; static int rfkill_add_switch(struct rfkill *rfkill) @@ -566,11 +605,14 @@ static int __init rfkill_init(void) return error; } + register_rfkill_notifier(&rfkill_blocking_uevent_nb); + return 0; } static void __exit rfkill_exit(void) { + unregister_rfkill_notifier(&rfkill_blocking_uevent_nb); class_unregister(&rfkill_class); } -- cgit v1.2.2 From fbc6af2f3c46df4722f5161d0ad20dd87cd7dfa9 Mon Sep 17 00:00:00 2001 From: Fabien Crespel Date: Mon, 23 Jun 2008 17:23:06 -0300 Subject: rfkill: drop current_state from tasks in rfkill-input The whole current_state thing seems completely useless and a source of problems in rfkill-input, since state comparison is already done in rfkill, and rfkill-input is more than likely to become out of sync with the real state. Signed-off-by: Fabien Crespel Acked-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Cc: Dmitry Torokhov Signed-off-by: John W. Linville --- net/rfkill/rfkill-input.c | 15 +-------------- 1 file changed, 1 insertion(+), 14 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill-input.c b/net/rfkill/rfkill-input.c index 29c13d308b31..d285f9a9d829 100644 --- a/net/rfkill/rfkill-input.c +++ b/net/rfkill/rfkill-input.c @@ -30,27 +30,15 @@ struct rfkill_task { spinlock_t lock; /* for accessing last and desired state */ unsigned long last; /* last schedule */ enum rfkill_state desired_state; /* on/off */ - enum rfkill_state current_state; /* on/off */ }; static void rfkill_task_handler(struct work_struct *work) { struct rfkill_task *task = container_of(work, struct rfkill_task, work); - enum rfkill_state state; mutex_lock(&task->mutex); - /* - * Use temp variable to fetch desired state to keep it - * consistent even if rfkill_schedule_toggle() runs in - * another thread or interrupts us. - */ - state = task->desired_state; - - if (state != task->current_state) { - rfkill_switch_all(task->type, state); - task->current_state = state; - } + rfkill_switch_all(task->type, task->desired_state); mutex_unlock(&task->mutex); } @@ -94,7 +82,6 @@ static void rfkill_schedule_toggle(struct rfkill_task *task) .mutex = __MUTEX_INITIALIZER(n.mutex), \ .lock = __SPIN_LOCK_UNLOCKED(n.lock), \ .desired_state = RFKILL_STATE_ON, \ - .current_state = RFKILL_STATE_ON, \ } static DEFINE_RFKILL_TASK(rfkill_wlan, RFKILL_TYPE_WLAN); -- cgit v1.2.2 From 4081f00dc45abce6bdac352a6354c07ce15db45b Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 23 Jun 2008 17:23:07 -0300 Subject: rfkill: do not allow userspace to override ALL RADIOS OFF SW_RFKILL_ALL is the "emergency power-off all radios" input event. It must be handled, and must always do the same thing as far as the rfkill system is concerned: all transmitters are to go *immediately* offline. For safety, do NOT allow userspace to override EV_SW SW_RFKILL_ALL OFF. As long as rfkill-input is loaded, that event will *always* be processed, and it will *always* force all rfkill switches to disable all wireless transmitters, regardless of user_claim attribute or anything else. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Cc: Dmitry Torokhov Signed-off-by: John W. Linville --- net/rfkill/rfkill-input.c | 47 ++++++++++++++++++++++++++++++++--------------- net/rfkill/rfkill-input.h | 1 + net/rfkill/rfkill.c | 18 ++++++++++++++++++ 3 files changed, 51 insertions(+), 15 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill-input.c b/net/rfkill/rfkill-input.c index d285f9a9d829..5d4c8b2446f7 100644 --- a/net/rfkill/rfkill-input.c +++ b/net/rfkill/rfkill-input.c @@ -43,11 +43,26 @@ static void rfkill_task_handler(struct work_struct *work) mutex_unlock(&task->mutex); } +static void rfkill_task_epo_handler(struct work_struct *work) +{ + rfkill_epo(); +} + +static DECLARE_WORK(epo_work, rfkill_task_epo_handler); + +static void rfkill_schedule_epo(void) +{ + schedule_work(&epo_work); +} + static void rfkill_schedule_set(struct rfkill_task *task, enum rfkill_state desired_state) { unsigned long flags; + if (unlikely(work_pending(&epo_work))) + return; + spin_lock_irqsave(&task->lock, flags); if (time_after(jiffies, task->last + msecs_to_jiffies(200))) { @@ -63,6 +78,9 @@ static void rfkill_schedule_toggle(struct rfkill_task *task) { unsigned long flags; + if (unlikely(work_pending(&epo_work))) + return; + spin_lock_irqsave(&task->lock, flags); if (time_after(jiffies, task->last + msecs_to_jiffies(200))) { @@ -114,21 +132,20 @@ static void rfkill_event(struct input_handle *handle, unsigned int type, switch (code) { case SW_RFKILL_ALL: /* EVERY radio type. data != 0 means radios ON */ - rfkill_schedule_set(&rfkill_wwan, - (data)? RFKILL_STATE_ON: - RFKILL_STATE_OFF); - rfkill_schedule_set(&rfkill_wimax, - (data)? RFKILL_STATE_ON: - RFKILL_STATE_OFF); - rfkill_schedule_set(&rfkill_uwb, - (data)? RFKILL_STATE_ON: - RFKILL_STATE_OFF); - rfkill_schedule_set(&rfkill_bt, - (data)? RFKILL_STATE_ON: - RFKILL_STATE_OFF); - rfkill_schedule_set(&rfkill_wlan, - (data)? RFKILL_STATE_ON: - RFKILL_STATE_OFF); + /* handle EPO (emergency power off) through shortcut */ + if (data) { + rfkill_schedule_set(&rfkill_wwan, + RFKILL_STATE_ON); + rfkill_schedule_set(&rfkill_wimax, + RFKILL_STATE_ON); + rfkill_schedule_set(&rfkill_uwb, + RFKILL_STATE_ON); + rfkill_schedule_set(&rfkill_bt, + RFKILL_STATE_ON); + rfkill_schedule_set(&rfkill_wlan, + RFKILL_STATE_ON); + } else + rfkill_schedule_epo(); break; default: break; diff --git a/net/rfkill/rfkill-input.h b/net/rfkill/rfkill-input.h index 4dae5006fc77..f63d05045685 100644 --- a/net/rfkill/rfkill-input.h +++ b/net/rfkill/rfkill-input.h @@ -12,5 +12,6 @@ #define __RFKILL_INPUT_H void rfkill_switch_all(enum rfkill_type type, enum rfkill_state state); +void rfkill_epo(void); #endif /* __RFKILL_INPUT_H */ diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index dd1c3f18f31d..7d07175c407f 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -181,6 +181,24 @@ void rfkill_switch_all(enum rfkill_type type, enum rfkill_state state) } EXPORT_SYMBOL(rfkill_switch_all); +/** + * rfkill_epo - emergency power off all transmitters + * + * This kicks all rfkill devices to RFKILL_STATE_OFF, ignoring + * everything in its path but rfkill_mutex. + */ +void rfkill_epo(void) +{ + struct rfkill *rfkill; + + mutex_lock(&rfkill_mutex); + list_for_each_entry(rfkill, &rfkill_list, node) { + rfkill_toggle_radio(rfkill, RFKILL_STATE_OFF, 1); + } + mutex_unlock(&rfkill_mutex); +} +EXPORT_SYMBOL_GPL(rfkill_epo); + /** * rfkill_force_state - Force the internal rfkill radio state * @rfkill: pointer to the rfkill class to modify. -- cgit v1.2.2 From 5005657cbd0fd6f277f807c0612a6b6d4396a02c Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Mon, 23 Jun 2008 17:46:42 -0300 Subject: rfkill: rename the rfkill_state states and add block-locked state The current naming of rfkill_state causes a lot of confusion: not only the "kill" in rfkill suggests negative logic, but also the fact that rfkill cannot turn anything on (it can just force something off or stop forcing something off) is often forgotten. Rename RFKILL_STATE_OFF to RFKILL_STATE_SOFT_BLOCKED (transmitter is blocked and will not operate; state can be changed by a toggle_radio request), and RFKILL_STATE_ON to RFKILL_STATE_UNBLOCKED (transmitter is not blocked, and may operate). Also, add a new third state, RFKILL_STATE_HARD_BLOCKED (transmitter is blocked and will not operate; state cannot be changed through a toggle_radio request), which is used by drivers to indicate a wireless transmiter was blocked by a hardware rfkill line that accepts no overrides. Keep the old names as #defines, but document them as deprecated. This way, drivers can be converted to the new names *and* verified to actually use rfkill correctly one by one. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/rfkill/rfkill-input.c | 29 +++++++++--------- net/rfkill/rfkill.c | 75 ++++++++++++++++++++++++++++++++++++++--------- 2 files changed, 76 insertions(+), 28 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill-input.c b/net/rfkill/rfkill-input.c index 5d4c8b2446f7..8aa822730145 100644 --- a/net/rfkill/rfkill-input.c +++ b/net/rfkill/rfkill-input.c @@ -84,7 +84,8 @@ static void rfkill_schedule_toggle(struct rfkill_task *task) spin_lock_irqsave(&task->lock, flags); if (time_after(jiffies, task->last + msecs_to_jiffies(200))) { - task->desired_state = !task->desired_state; + task->desired_state = + rfkill_state_complement(task->desired_state); task->last = jiffies; schedule_work(&task->work); } @@ -92,14 +93,14 @@ static void rfkill_schedule_toggle(struct rfkill_task *task) spin_unlock_irqrestore(&task->lock, flags); } -#define DEFINE_RFKILL_TASK(n, t) \ - struct rfkill_task n = { \ - .work = __WORK_INITIALIZER(n.work, \ - rfkill_task_handler), \ - .type = t, \ - .mutex = __MUTEX_INITIALIZER(n.mutex), \ - .lock = __SPIN_LOCK_UNLOCKED(n.lock), \ - .desired_state = RFKILL_STATE_ON, \ +#define DEFINE_RFKILL_TASK(n, t) \ + struct rfkill_task n = { \ + .work = __WORK_INITIALIZER(n.work, \ + rfkill_task_handler), \ + .type = t, \ + .mutex = __MUTEX_INITIALIZER(n.mutex), \ + .lock = __SPIN_LOCK_UNLOCKED(n.lock), \ + .desired_state = RFKILL_STATE_UNBLOCKED, \ } static DEFINE_RFKILL_TASK(rfkill_wlan, RFKILL_TYPE_WLAN); @@ -135,15 +136,15 @@ static void rfkill_event(struct input_handle *handle, unsigned int type, /* handle EPO (emergency power off) through shortcut */ if (data) { rfkill_schedule_set(&rfkill_wwan, - RFKILL_STATE_ON); + RFKILL_STATE_UNBLOCKED); rfkill_schedule_set(&rfkill_wimax, - RFKILL_STATE_ON); + RFKILL_STATE_UNBLOCKED); rfkill_schedule_set(&rfkill_uwb, - RFKILL_STATE_ON); + RFKILL_STATE_UNBLOCKED); rfkill_schedule_set(&rfkill_bt, - RFKILL_STATE_ON); + RFKILL_STATE_UNBLOCKED); rfkill_schedule_set(&rfkill_wlan, - RFKILL_STATE_ON); + RFKILL_STATE_UNBLOCKED); } else rfkill_schedule_epo(); break; diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index 7d07175c407f..ce0e23148cdd 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -39,7 +39,7 @@ MODULE_LICENSE("GPL"); static LIST_HEAD(rfkill_list); /* list of registered rf switches */ static DEFINE_MUTEX(rfkill_mutex); -static unsigned int rfkill_default_state = RFKILL_STATE_ON; +static unsigned int rfkill_default_state = RFKILL_STATE_UNBLOCKED; module_param_named(default_state, rfkill_default_state, uint, 0444); MODULE_PARM_DESC(default_state, "Default initial state for all radio types, 0 = radio off"); @@ -98,7 +98,7 @@ static void rfkill_led_trigger(struct rfkill *rfkill, if (!led->name) return; - if (state == RFKILL_STATE_OFF) + if (state != RFKILL_STATE_UNBLOCKED) led_trigger_event(led, LED_OFF); else led_trigger_event(led, LED_FULL); @@ -128,6 +128,28 @@ static void update_rfkill_state(struct rfkill *rfkill) } } +/** + * rfkill_toggle_radio - wrapper for toggle_radio hook + * calls toggle_radio taking into account a lot of "small" + * details. + * @rfkill: the rfkill struct to use + * @force: calls toggle_radio even if cache says it is not needed, + * and also makes sure notifications of the state will be + * sent even if it didn't change + * @state: the new state to call toggle_radio() with + * + * This wrappen protects and enforces the API for toggle_radio + * calls. Note that @force cannot override a (possibly cached) + * state of RFKILL_STATE_HARD_BLOCKED. Any device making use of + * RFKILL_STATE_HARD_BLOCKED implements either get_state() or + * rfkill_force_state(), so the cache either is bypassed or valid. + * + * Note that we do call toggle_radio for RFKILL_STATE_SOFT_BLOCKED + * even if the radio is in RFKILL_STATE_HARD_BLOCKED state, so as to + * give the driver a hint that it should double-BLOCK the transmitter. + * + * Caller must have aquired rfkill_mutex. + */ static int rfkill_toggle_radio(struct rfkill *rfkill, enum rfkill_state state, int force) @@ -141,9 +163,28 @@ static int rfkill_toggle_radio(struct rfkill *rfkill, !rfkill->get_state(rfkill->data, &newstate)) rfkill->state = newstate; + switch (state) { + case RFKILL_STATE_HARD_BLOCKED: + /* typically happens when refreshing hardware state, + * such as on resume */ + state = RFKILL_STATE_SOFT_BLOCKED; + break; + case RFKILL_STATE_UNBLOCKED: + /* force can't override this, only rfkill_force_state() can */ + if (rfkill->state == RFKILL_STATE_HARD_BLOCKED) + return -EPERM; + break; + case RFKILL_STATE_SOFT_BLOCKED: + /* nothing to do, we want to give drivers the hint to double + * BLOCK even a transmitter that is already in state + * RFKILL_STATE_HARD_BLOCKED */ + break; + } + if (force || state != rfkill->state) { retval = rfkill->toggle_radio(rfkill->data, state); - if (!retval) + /* never allow a HARD->SOFT downgrade! */ + if (!retval && rfkill->state != RFKILL_STATE_HARD_BLOCKED) rfkill->state = state; } @@ -184,7 +225,7 @@ EXPORT_SYMBOL(rfkill_switch_all); /** * rfkill_epo - emergency power off all transmitters * - * This kicks all rfkill devices to RFKILL_STATE_OFF, ignoring + * This kicks all rfkill devices to RFKILL_STATE_SOFT_BLOCKED, ignoring * everything in its path but rfkill_mutex. */ void rfkill_epo(void) @@ -193,7 +234,7 @@ void rfkill_epo(void) mutex_lock(&rfkill_mutex); list_for_each_entry(rfkill, &rfkill_list, node) { - rfkill_toggle_radio(rfkill, RFKILL_STATE_OFF, 1); + rfkill_toggle_radio(rfkill, RFKILL_STATE_SOFT_BLOCKED, 1); } mutex_unlock(&rfkill_mutex); } @@ -215,8 +256,9 @@ int rfkill_force_state(struct rfkill *rfkill, enum rfkill_state state) { enum rfkill_state oldstate; - if (state != RFKILL_STATE_OFF && - state != RFKILL_STATE_ON) + if (state != RFKILL_STATE_SOFT_BLOCKED && + state != RFKILL_STATE_UNBLOCKED && + state != RFKILL_STATE_HARD_BLOCKED) return -EINVAL; mutex_lock(&rfkill->mutex); @@ -290,11 +332,14 @@ static ssize_t rfkill_state_store(struct device *dev, if (!capable(CAP_NET_ADMIN)) return -EPERM; + /* RFKILL_STATE_HARD_BLOCKED is illegal here... */ + if (state != RFKILL_STATE_UNBLOCKED && + state != RFKILL_STATE_SOFT_BLOCKED) + return -EINVAL; + if (mutex_lock_interruptible(&rfkill->mutex)) return -ERESTARTSYS; - error = rfkill_toggle_radio(rfkill, - state ? RFKILL_STATE_ON : RFKILL_STATE_OFF, - 0); + error = rfkill_toggle_radio(rfkill, state, 0); mutex_unlock(&rfkill->mutex); return error ? error : count; @@ -373,7 +418,8 @@ static int rfkill_suspend(struct device *dev, pm_message_t state) update_rfkill_state(rfkill); mutex_lock(&rfkill->mutex); - rfkill->toggle_radio(rfkill->data, RFKILL_STATE_OFF); + rfkill->toggle_radio(rfkill->data, + RFKILL_STATE_SOFT_BLOCKED); mutex_unlock(&rfkill->mutex); } @@ -470,7 +516,7 @@ static void rfkill_remove_switch(struct rfkill *rfkill) { mutex_lock(&rfkill_mutex); list_del_init(&rfkill->node); - rfkill_toggle_radio(rfkill, RFKILL_STATE_OFF, 1); + rfkill_toggle_radio(rfkill, RFKILL_STATE_SOFT_BLOCKED, 1); mutex_unlock(&rfkill_mutex); } @@ -610,8 +656,9 @@ static int __init rfkill_init(void) int error; int i; - if (rfkill_default_state != RFKILL_STATE_OFF && - rfkill_default_state != RFKILL_STATE_ON) + /* RFKILL_STATE_HARD_BLOCKED is illegal here... */ + if (rfkill_default_state != RFKILL_STATE_SOFT_BLOCKED && + rfkill_default_state != RFKILL_STATE_UNBLOCKED) return -EINVAL; for (i = 0; i < ARRAY_SIZE(rfkill_states); i++) -- cgit v1.2.2 From f2df38596a81b6c24f4586b0b4befeaebf3e02db Mon Sep 17 00:00:00 2001 From: Assaf Krauss Date: Sun, 15 Jun 2008 18:23:29 +0300 Subject: mac80211: 11h Infrastructure - Parsing This patch introduces parsing of 11h and 11d related elements from incoming management frames. Signed-off-by: Assaf Krauss Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 9 +++++++++ net/mac80211/mlme.c | 19 +++++++++++++++++++ 2 files changed, 28 insertions(+) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 80a9e7c07b47..af352c05c983 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -791,6 +791,10 @@ struct ieee802_11_elems { u8 *preq; u8 *prep; u8 *perr; + u8 *ch_switch_elem; + u8 *country_elem; + u8 *pwr_constr_elem; + u8 *quiet_elem; /* first quite element */ /* length of them, respectively */ u8 ssid_len; @@ -815,6 +819,11 @@ struct ieee802_11_elems { u8 preq_len; u8 prep_len; u8 perr_len; + u8 ch_switch_elem_len; + u8 country_elem_len; + u8 pwr_constr_elem_len; + u8 quiet_elem_len; + u8 num_of_quiet_elem; /* can be more the one */ }; static inline struct ieee80211_local *hw_to_local( diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index e06d6450f215..32453561fe32 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -204,6 +204,25 @@ void ieee802_11_parse_elems(u8 *start, size_t len, elems->perr = pos; elems->perr_len = elen; break; + case WLAN_EID_CHANNEL_SWITCH: + elems->ch_switch_elem = pos; + elems->ch_switch_elem_len = elen; + break; + case WLAN_EID_QUIET: + if (!elems->quiet_elem) { + elems->quiet_elem = pos; + elems->quiet_elem_len = elen; + } + elems->num_of_quiet_elem++; + break; + case WLAN_EID_COUNTRY: + elems->country_elem = pos; + elems->country_elem_len = elen; + break; + case WLAN_EID_PWR_CONSTRAINT: + elems->pwr_constr_elem = pos; + elems->pwr_constr_elem_len = elen; + break; default: break; } -- cgit v1.2.2 From b662348662f9661f9259c7186c1bdb65620045f1 Mon Sep 17 00:00:00 2001 From: Assaf Krauss Date: Mon, 16 Jun 2008 16:09:49 +0300 Subject: mac80211: 11h - Handling measurement request This patch handles the 11h measurement request information element. This is minimal requested implementation - refuse measurement. Signed-off-by: Assaf Krauss Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 78 +++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 78 insertions(+) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 32453561fe32..6d9a4facf9dd 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -1720,6 +1720,71 @@ void ieee80211_sta_tear_down_BA_sessions(struct net_device *dev, u8 *addr) } } +static void ieee80211_send_refuse_measurement_request(struct net_device *dev, + struct ieee80211_msrment_ie *request_ie, + const u8 *da, const u8 *bssid, + u8 dialog_token) +{ + struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); + struct sk_buff *skb; + struct ieee80211_mgmt *msr_report; + + skb = dev_alloc_skb(sizeof(*msr_report) + local->hw.extra_tx_headroom + + sizeof(struct ieee80211_msrment_ie)); + + if (!skb) { + printk(KERN_ERR "%s: failed to allocate buffer for " + "measurement report frame\n", dev->name); + return; + } + + skb_reserve(skb, local->hw.extra_tx_headroom); + msr_report = (struct ieee80211_mgmt *)skb_put(skb, 24); + memset(msr_report, 0, 24); + memcpy(msr_report->da, da, ETH_ALEN); + memcpy(msr_report->sa, dev->dev_addr, ETH_ALEN); + memcpy(msr_report->bssid, bssid, ETH_ALEN); + msr_report->frame_control = IEEE80211_FC(IEEE80211_FTYPE_MGMT, + IEEE80211_STYPE_ACTION); + + skb_put(skb, 1 + sizeof(msr_report->u.action.u.measurement)); + msr_report->u.action.category = WLAN_CATEGORY_SPECTRUM_MGMT; + msr_report->u.action.u.measurement.action_code = + WLAN_ACTION_SPCT_MSR_RPRT; + msr_report->u.action.u.measurement.dialog_token = dialog_token; + + msr_report->u.action.u.measurement.element_id = WLAN_EID_MEASURE_REPORT; + msr_report->u.action.u.measurement.length = + sizeof(struct ieee80211_msrment_ie); + + memset(&msr_report->u.action.u.measurement.msr_elem, 0, + sizeof(struct ieee80211_msrment_ie)); + msr_report->u.action.u.measurement.msr_elem.token = request_ie->token; + msr_report->u.action.u.measurement.msr_elem.mode |= + IEEE80211_SPCT_MSR_RPRT_MODE_REFUSED; + msr_report->u.action.u.measurement.msr_elem.type = request_ie->type; + + ieee80211_sta_tx(dev, skb, 0); +} + +static void ieee80211_sta_process_measurement_req(struct net_device *dev, + struct ieee80211_mgmt *mgmt, + size_t len) +{ + /* + * Ignoring measurement request is spec violation. + * Mandatory measurements must be reported optional + * measurements might be refused or reported incapable + * For now just refuse + * TODO: Answer basic measurement as unmeasured + */ + ieee80211_send_refuse_measurement_request(dev, + &mgmt->u.action.u.measurement.msr_elem, + mgmt->sa, mgmt->bssid, + mgmt->u.action.u.measurement.dialog_token); +} + + static void ieee80211_rx_mgmt_auth(struct net_device *dev, struct ieee80211_if_sta *ifsta, struct ieee80211_mgmt *mgmt, @@ -3044,11 +3109,24 @@ static void ieee80211_rx_mgmt_action(struct net_device *dev, struct ieee80211_rx_status *rx_status) { struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); + struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); if (len < IEEE80211_MIN_ACTION_SIZE) return; switch (mgmt->u.action.category) { + case WLAN_CATEGORY_SPECTRUM_MGMT: + if (local->hw.conf.channel->band != IEEE80211_BAND_5GHZ) + break; + switch (mgmt->u.action.u.chan_switch.action_code) { + case WLAN_ACTION_SPCT_MSR_REQ: + if (len < (IEEE80211_MIN_ACTION_SIZE + + sizeof(mgmt->u.action.u.measurement))) + break; + ieee80211_sta_process_measurement_req(dev, mgmt, len); + break; + } + break; case WLAN_CATEGORY_BACK: switch (mgmt->u.action.u.addba_req.action_code) { case WLAN_ACTION_ADDBA_REQ: -- cgit v1.2.2 From 135a2110c55c71d7ccaf5ac66968b993347fe8e2 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Mon, 16 Jun 2008 20:55:29 +0200 Subject: mac80211: remove shared key todo Adding shared key authentication is not going to happen anyway. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 6d9a4facf9dd..75c4e9743fb4 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -1837,11 +1837,12 @@ static void ieee80211_rx_mgmt_auth(struct net_device *dev, auth_transaction, status_code); if (sdata->vif.type == IEEE80211_IF_TYPE_IBSS) { - /* IEEE 802.11 standard does not require authentication in IBSS + /* + * IEEE 802.11 standard does not require authentication in IBSS * networks and most implementations do not seem to use it. * However, try to reply to authentication attempts if someone * has actually implemented this. - * TODO: Could implement shared key authentication. */ + */ if (auth_alg != WLAN_AUTH_OPEN || auth_transaction != 1) { printk(KERN_DEBUG "%s: unexpected IBSS authentication " "frame (alg=%d transaction=%d)\n", -- cgit v1.2.2 From 5a9f7b047e81a73a1ce3e42ef87c28a61fd4df24 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Wed, 18 Jun 2008 14:58:09 +0200 Subject: mac80211: use separate spinlock for sta flags MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit David Ellingsworth posted a bug that was only noticable on UP/NO-PREEMPT and Michael correctly analysed it to be a spin_lock_bh() section within a spin_lock_irqsave() section. This adds a separate spinlock for the sta_info flags to fix that issue and avoid having to take much care about where the sta flag manipulation functions are called. Signed-off-by: Johannes Berg Reported-By: David Ellingsworth Signed-off-by: John W. Linville --- net/mac80211/sta_info.c | 1 + net/mac80211/sta_info.h | 40 +++++++++++++++++++++++++++------------- 2 files changed, 28 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index c24770cb02c5..b3c733162fc1 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -235,6 +235,7 @@ struct sta_info *sta_info_alloc(struct ieee80211_sub_if_data *sdata, return NULL; spin_lock_init(&sta->lock); + spin_lock_init(&sta->flaglock); memcpy(sta->addr, addr, ETH_ALEN); sta->local = local; diff --git a/net/mac80211/sta_info.h b/net/mac80211/sta_info.h index 95753f860acf..fd228c198e31 100644 --- a/net/mac80211/sta_info.h +++ b/net/mac80211/sta_info.h @@ -164,6 +164,7 @@ struct sta_ampdu_mlme { * @aid: STA's unique AID (1..2007, 0 = not assigned yet), * only used in AP (and IBSS?) mode * @flags: STA flags, see &enum ieee80211_sta_info_flags + * @flaglock: spinlock for flags accesses * @ps_tx_buf: buffer of frames to transmit to this station * when it leaves power saving state * @tx_filtered: buffer of frames we already tried to transmit @@ -186,6 +187,7 @@ struct sta_info { struct rate_control_ref *rate_ctrl; void *rate_ctrl_priv; spinlock_t lock; + spinlock_t flaglock; struct ieee80211_ht_info ht_info; u64 supp_rates[IEEE80211_NUM_BANDS]; u8 addr[ETH_ALEN]; @@ -198,7 +200,10 @@ struct sta_info { */ u8 pin_status; - /* frequently updated information, locked with lock spinlock */ + /* + * frequently updated, locked with own spinlock (flaglock), + * use the accessors defined below + */ u32 flags; /* @@ -293,34 +298,41 @@ static inline enum plink_state sta_plink_state(struct sta_info *sta) static inline void set_sta_flags(struct sta_info *sta, const u32 flags) { - spin_lock_bh(&sta->lock); + unsigned long irqfl; + + spin_lock_irqsave(&sta->flaglock, irqfl); sta->flags |= flags; - spin_unlock_bh(&sta->lock); + spin_unlock_irqrestore(&sta->flaglock, irqfl); } static inline void clear_sta_flags(struct sta_info *sta, const u32 flags) { - spin_lock_bh(&sta->lock); + unsigned long irqfl; + + spin_lock_irqsave(&sta->flaglock, irqfl); sta->flags &= ~flags; - spin_unlock_bh(&sta->lock); + spin_unlock_irqrestore(&sta->flaglock, irqfl); } static inline void set_and_clear_sta_flags(struct sta_info *sta, const u32 set, const u32 clear) { - spin_lock_bh(&sta->lock); + unsigned long irqfl; + + spin_lock_irqsave(&sta->flaglock, irqfl); sta->flags |= set; sta->flags &= ~clear; - spin_unlock_bh(&sta->lock); + spin_unlock_irqrestore(&sta->flaglock, irqfl); } static inline u32 test_sta_flags(struct sta_info *sta, const u32 flags) { u32 ret; + unsigned long irqfl; - spin_lock_bh(&sta->lock); + spin_lock_irqsave(&sta->flaglock, irqfl); ret = sta->flags & flags; - spin_unlock_bh(&sta->lock); + spin_unlock_irqrestore(&sta->flaglock, irqfl); return ret; } @@ -329,11 +341,12 @@ static inline u32 test_and_clear_sta_flags(struct sta_info *sta, const u32 flags) { u32 ret; + unsigned long irqfl; - spin_lock_bh(&sta->lock); + spin_lock_irqsave(&sta->flaglock, irqfl); ret = sta->flags & flags; sta->flags &= ~flags; - spin_unlock_bh(&sta->lock); + spin_unlock_irqrestore(&sta->flaglock, irqfl); return ret; } @@ -341,10 +354,11 @@ static inline u32 test_and_clear_sta_flags(struct sta_info *sta, static inline u32 get_sta_flags(struct sta_info *sta) { u32 ret; + unsigned long irqfl; - spin_lock_bh(&sta->lock); + spin_lock_irqsave(&sta->flaglock, irqfl); ret = sta->flags; - spin_unlock_bh(&sta->lock); + spin_unlock_irqrestore(&sta->flaglock, irqfl); return ret; } -- cgit v1.2.2 From 97b045d62bffae5a91a286b56ac51db0c4385687 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Fri, 20 Jun 2008 01:22:30 +0200 Subject: mac80211: add single function calling tx handlers This modifies mac80211 to only have a single function calling the TX handlers rather than them being invoked in multiple places. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/tx.c | 82 +++++++++++++++++++++++++++---------------------------- 1 file changed, 40 insertions(+), 42 deletions(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index ce06e791bf43..7a14a39ebd78 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -1083,13 +1083,46 @@ static int __ieee80211_tx(struct ieee80211_local *local, struct sk_buff *skb, return IEEE80211_TX_OK; } +/* + * Invoke TX handlers, return 0 on success and non-zero if the + * frame was dropped or queued. + */ +static int invoke_tx_handlers(struct ieee80211_tx_data *tx) +{ + struct ieee80211_local *local = tx->local; + struct sk_buff *skb = tx->skb; + ieee80211_tx_handler *handler; + ieee80211_tx_result res = TX_DROP; + int i; + + for (handler = ieee80211_tx_handlers; *handler != NULL; handler++) { + res = (*handler)(tx); + if (res != TX_CONTINUE) + break; + } + + if (unlikely(res == TX_DROP)) { + I802_DEBUG_INC(local->tx_handlers_drop); + dev_kfree_skb(skb); + for (i = 0; i < tx->num_extra_frag; i++) + if (tx->extra_frag[i]) + dev_kfree_skb(tx->extra_frag[i]); + kfree(tx->extra_frag); + return -1; + } else if (unlikely(res == TX_QUEUED)) { + I802_DEBUG_INC(local->tx_handlers_queued); + return -1; + } + + return 0; +} + static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct sta_info *sta; - ieee80211_tx_handler *handler; struct ieee80211_tx_data tx; - ieee80211_tx_result res = TX_DROP, res_prepare; + ieee80211_tx_result res_prepare; struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); int ret, i; u16 queue; @@ -1118,26 +1151,8 @@ static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb) tx.channel = local->hw.conf.channel; info->band = tx.channel->band; - for (handler = ieee80211_tx_handlers; *handler != NULL; - handler++) { - res = (*handler)(&tx); - if (res != TX_CONTINUE) - break; - } - - if (WARN_ON(tx.skb != skb)) - goto drop; - - if (unlikely(res == TX_DROP)) { - I802_DEBUG_INC(local->tx_handlers_drop); - goto drop; - } - - if (unlikely(res == TX_QUEUED)) { - I802_DEBUG_INC(local->tx_handlers_queued); - rcu_read_unlock(); - return 0; - } + if (invoke_tx_handlers(&tx)) + goto out; if (tx.extra_frag) { for (i = 0; i < tx.num_extra_frag; i++) { @@ -1198,6 +1213,7 @@ retry: store->last_frag_rate_ctrl_probe = !!(tx.flags & IEEE80211_TX_PROBE_LAST_FRAG); } + out: rcu_read_unlock(); return 0; @@ -1948,9 +1964,7 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw, struct ieee80211_local *local = hw_to_local(hw); struct sk_buff *skb = NULL; struct sta_info *sta; - ieee80211_tx_handler *handler; struct ieee80211_tx_data tx; - ieee80211_tx_result res = TX_DROP; struct net_device *bdev; struct ieee80211_sub_if_data *sdata; struct ieee80211_if_ap *bss = NULL; @@ -2001,25 +2015,9 @@ ieee80211_get_buffered_bc(struct ieee80211_hw *hw, tx.channel = local->hw.conf.channel; info->band = tx.channel->band; - for (handler = ieee80211_tx_handlers; *handler != NULL; handler++) { - res = (*handler)(&tx); - if (res == TX_DROP || res == TX_QUEUED) - break; - } - - if (WARN_ON(tx.skb != skb)) - res = TX_DROP; - - if (res == TX_DROP) { - I802_DEBUG_INC(local->tx_handlers_drop); - dev_kfree_skb(skb); - skb = NULL; - } else if (res == TX_QUEUED) { - I802_DEBUG_INC(local->tx_handlers_queued); + if (invoke_tx_handlers(&tx)) skb = NULL; - } - -out: + out: rcu_read_unlock(); return skb; -- cgit v1.2.2 From 9ae705cfd390f9077eec856ea4dff374d166de33 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Fri, 20 Jun 2008 11:54:24 +0200 Subject: mac80211: rename TKIP debugging Kconfig symbol ... to MAC80211_TKIP_DEBUG rather than TKIP_DEBUG. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/Kconfig | 2 +- net/mac80211/tkip.c | 20 ++++++++++---------- 2 files changed, 11 insertions(+), 11 deletions(-) (limited to 'net') diff --git a/net/mac80211/Kconfig b/net/mac80211/Kconfig index 590e00b2766c..0d3661d9b6a0 100644 --- a/net/mac80211/Kconfig +++ b/net/mac80211/Kconfig @@ -150,7 +150,7 @@ config MAC80211_LOWTX_FRAME_DUMP If unsure, say N and insert the debugging code you require into the driver you are debugging. -config TKIP_DEBUG +config MAC80211_TKIP_DEBUG bool "TKIP debugging" depends on MAC80211_DEBUG diff --git a/net/mac80211/tkip.c b/net/mac80211/tkip.c index e710243d82e2..699807889988 100644 --- a/net/mac80211/tkip.c +++ b/net/mac80211/tkip.c @@ -167,7 +167,7 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; ctx = &key->u.tkip.tx; -#ifdef CONFIG_TKIP_DEBUG +#ifdef CONFIG_MAC80211_TKIP_DEBUG printk(KERN_DEBUG "TKIP encrypt: iv16 = 0x%04x, iv32 = 0x%08x\n", iv16, iv32); @@ -177,7 +177,7 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, printk(KERN_DEBUG "Wrap around of iv16 in the middle of a " "fragmented packet\n"); } -#endif /* CONFIG_TKIP_DEBUG */ +#endif /* Update the p1k only when the iv16 in the packet wraps around, this * might occur after the wrap around of iv16 in the key in case of @@ -240,7 +240,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, keyid = pos[3]; iv32 = get_unaligned_le32(pos + 4); pos += 8; -#ifdef CONFIG_TKIP_DEBUG +#ifdef CONFIG_MAC80211_TKIP_DEBUG { int i; printk(KERN_DEBUG "TKIP decrypt: data(len=%zd)", payload_len); @@ -250,7 +250,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, printk(KERN_DEBUG "TKIP decrypt: iv16=%04x iv32=%08x\n", iv16, iv32); } -#endif /* CONFIG_TKIP_DEBUG */ +#endif if (!(keyid & (1 << 5))) return TKIP_DECRYPT_NO_EXT_IV; @@ -262,14 +262,14 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, (iv32 < key->u.tkip.rx[queue].iv32 || (iv32 == key->u.tkip.rx[queue].iv32 && iv16 <= key->u.tkip.rx[queue].iv16))) { -#ifdef CONFIG_TKIP_DEBUG +#ifdef CONFIG_MAC80211_TKIP_DEBUG DECLARE_MAC_BUF(mac); printk(KERN_DEBUG "TKIP replay detected for RX frame from " "%s (RX IV (%04x,%02x) <= prev. IV (%04x,%02x)\n", print_mac(mac, ta), iv32, iv16, key->u.tkip.rx[queue].iv32, key->u.tkip.rx[queue].iv16); -#endif /* CONFIG_TKIP_DEBUG */ +#endif return TKIP_DECRYPT_REPLAY; } @@ -283,7 +283,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, key->u.tkip.rx[queue].iv32 != iv32) { /* IV16 wrapped around - perform TKIP phase 1 */ tkip_mixing_phase1(tk, &key->u.tkip.rx[queue], ta, iv32); -#ifdef CONFIG_TKIP_DEBUG +#ifdef CONFIG_MAC80211_TKIP_DEBUG { int i; DECLARE_MAC_BUF(mac); @@ -299,7 +299,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, printk("%04x ", key->u.tkip.rx[queue].p1k[i]); printk("\n"); } -#endif /* CONFIG_TKIP_DEBUG */ +#endif if (key->local->ops->update_tkip_key && key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) { u8 bcast[ETH_ALEN] = @@ -316,7 +316,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, } tkip_mixing_phase2(tk, &key->u.tkip.rx[queue], iv16, rc4key); -#ifdef CONFIG_TKIP_DEBUG +#ifdef CONFIG_MAC80211_TKIP_DEBUG { int i; printk(KERN_DEBUG "TKIP decrypt: Phase2 rc4key="); @@ -324,7 +324,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, printk("%02x ", rc4key[i]); printk("\n"); } -#endif /* CONFIG_TKIP_DEBUG */ +#endif res = ieee80211_wep_decrypt_data(tfm, rc4key, 16, pos, payload_len - 12); done: -- cgit v1.2.2 From ffd7891dc909b3648e87f7cf8f84a6dc12fc1cc6 Mon Sep 17 00:00:00 2001 From: "Luis R. Rodriguez" Date: Sat, 21 Jun 2008 10:02:46 -0400 Subject: mac80211: Let drivers have access to TKIP key offets for TX and RX MIC Some drivers may want to to use the TKIP key offsets for TX and RX MIC so lets move this out. Lets also clear up a bit how this is used internally in mac80211. Signed-off-by: Luis R. Rodriguez Signed-off-by: John W. Linville --- net/mac80211/key.h | 37 ++++++++++++------------------------- net/mac80211/tkip.c | 10 +++++----- net/mac80211/wpa.c | 20 ++++++++++++++------ 3 files changed, 31 insertions(+), 36 deletions(-) (limited to 'net') diff --git a/net/mac80211/key.h b/net/mac80211/key.h index a0f774aafa45..425816e0996c 100644 --- a/net/mac80211/key.h +++ b/net/mac80211/key.h @@ -16,31 +16,18 @@ #include #include -/* ALG_TKIP - * struct ieee80211_key::key is encoded as a 256-bit (32 byte) data block: - * Temporal Encryption Key (128 bits) - * Temporal Authenticator Tx MIC Key (64 bits) - * Temporal Authenticator Rx MIC Key (64 bits) - */ - -#define WEP_IV_LEN 4 -#define WEP_ICV_LEN 4 - -#define ALG_TKIP_KEY_LEN 32 -/* Starting offsets for each key */ -#define ALG_TKIP_TEMP_ENCR_KEY 0 -#define ALG_TKIP_TEMP_AUTH_TX_MIC_KEY 16 -#define ALG_TKIP_TEMP_AUTH_RX_MIC_KEY 24 -#define TKIP_IV_LEN 8 -#define TKIP_ICV_LEN 4 - -#define ALG_CCMP_KEY_LEN 16 -#define CCMP_HDR_LEN 8 -#define CCMP_MIC_LEN 8 -#define CCMP_TK_LEN 16 -#define CCMP_PN_LEN 6 - -#define NUM_RX_DATA_QUEUES 17 +#define WEP_IV_LEN 4 +#define WEP_ICV_LEN 4 +#define ALG_TKIP_KEY_LEN 32 +#define ALG_CCMP_KEY_LEN 16 +#define CCMP_HDR_LEN 8 +#define CCMP_MIC_LEN 8 +#define CCMP_TK_LEN 16 +#define CCMP_PN_LEN 6 +#define TKIP_IV_LEN 8 +#define TKIP_ICV_LEN 4 + +#define NUM_RX_DATA_QUEUES 17 struct ieee80211_local; struct ieee80211_sub_if_data; diff --git a/net/mac80211/tkip.c b/net/mac80211/tkip.c index 699807889988..995f7af3d25e 100644 --- a/net/mac80211/tkip.c +++ b/net/mac80211/tkip.c @@ -164,7 +164,7 @@ void ieee80211_get_tkip_key(struct ieee80211_key_conf *keyconf, iv16 = data[2] | (data[0] << 8); iv32 = get_unaligned_le32(&data[4]); - tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; + tk = &key->conf.key[NL80211_TKIP_DATA_OFFSET_ENCR_KEY]; ctx = &key->u.tkip.tx; #ifdef CONFIG_MAC80211_TKIP_DEBUG @@ -205,7 +205,7 @@ void ieee80211_tkip_encrypt_data(struct crypto_blkcipher *tfm, { u8 rc4key[16]; struct tkip_ctx *ctx = &key->u.tkip.tx; - const u8 *tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; + const u8 *tk = &key->conf.key[NL80211_TKIP_DATA_OFFSET_ENCR_KEY]; /* Calculate per-packet key */ if (ctx->iv16 == 0 || !ctx->initialized) @@ -231,7 +231,7 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, u32 iv16; u8 rc4key[16], keyid, *pos = payload; int res; - const u8 *tk = &key->conf.key[ALG_TKIP_TEMP_ENCR_KEY]; + const u8 *tk = &key->conf.key[NL80211_TKIP_DATA_OFFSET_ENCR_KEY]; if (payload_len < 12) return -1; @@ -286,13 +286,13 @@ int ieee80211_tkip_decrypt_data(struct crypto_blkcipher *tfm, #ifdef CONFIG_MAC80211_TKIP_DEBUG { int i; + u8 key_offset = NL80211_TKIP_DATA_OFFSET_ENCR_KEY; DECLARE_MAC_BUF(mac); printk(KERN_DEBUG "TKIP decrypt: Phase1 TA=%s" " TK=", print_mac(mac, ta)); for (i = 0; i < 16; i++) printk("%02x ", - key->conf.key[ - ALG_TKIP_TEMP_ENCR_KEY + i]); + key->conf.key[key_offset + i]); printk("\n"); printk(KERN_DEBUG "TKIP decrypt: P1K="); for (i = 0; i < 5; i++) diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index 345e10e9b313..f809761fbfb5 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -49,7 +49,7 @@ static int ieee80211_get_hdr_info(const struct sk_buff *skb, u8 **sa, u8 **da, ieee80211_tx_result ieee80211_tx_h_michael_mic_add(struct ieee80211_tx_data *tx) { - u8 *data, *sa, *da, *key, *mic, qos_tid; + u8 *data, *sa, *da, *key, *mic, qos_tid, key_offset; size_t data_len; u16 fc; struct sk_buff *skb = tx->skb; @@ -88,8 +88,12 @@ ieee80211_tx_h_michael_mic_add(struct ieee80211_tx_data *tx) #else authenticator = 1; #endif - key = &tx->key->conf.key[authenticator ? ALG_TKIP_TEMP_AUTH_TX_MIC_KEY : - ALG_TKIP_TEMP_AUTH_RX_MIC_KEY]; + /* At this point we know we're using ALG_TKIP. To get the MIC key + * we now will rely on the offset from the ieee80211_key_conf::key */ + key_offset = authenticator ? + NL80211_TKIP_DATA_OFFSET_TX_MIC_KEY : + NL80211_TKIP_DATA_OFFSET_RX_MIC_KEY; + key = &tx->key->conf.key[key_offset]; mic = skb_put(skb, MICHAEL_MIC_LEN); michael_mic(key, da, sa, qos_tid & 0x0f, data, data_len, mic); @@ -100,7 +104,7 @@ ieee80211_tx_h_michael_mic_add(struct ieee80211_tx_data *tx) ieee80211_rx_result ieee80211_rx_h_michael_mic_verify(struct ieee80211_rx_data *rx) { - u8 *data, *sa, *da, *key = NULL, qos_tid; + u8 *data, *sa, *da, *key = NULL, qos_tid, key_offset; size_t data_len; u16 fc; u8 mic[MICHAEL_MIC_LEN]; @@ -131,8 +135,12 @@ ieee80211_rx_h_michael_mic_verify(struct ieee80211_rx_data *rx) #else authenticator = 1; #endif - key = &rx->key->conf.key[authenticator ? ALG_TKIP_TEMP_AUTH_RX_MIC_KEY : - ALG_TKIP_TEMP_AUTH_TX_MIC_KEY]; + /* At this point we know we're using ALG_TKIP. To get the MIC key + * we now will rely on the offset from the ieee80211_key_conf::key */ + key_offset = authenticator ? + NL80211_TKIP_DATA_OFFSET_RX_MIC_KEY : + NL80211_TKIP_DATA_OFFSET_TX_MIC_KEY; + key = &rx->key->conf.key[key_offset]; michael_mic(key, da, sa, qos_tid & 0x0f, data, data_len, mic); if (memcmp(mic, data + data_len, MICHAEL_MIC_LEN) != 0 || wpa_test) { if (!(rx->flags & IEEE80211_RX_RA_MATCH)) -- cgit v1.2.2 From 70217d7f83a12e0a16c726951ddf8f7f0ba7d1a4 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Sun, 22 Jun 2008 16:45:23 -0700 Subject: mac80211: wep.c use new frame control helpers Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/wep.c | 39 +++++++++++++++------------------------ 1 file changed, 15 insertions(+), 24 deletions(-) (limited to 'net') diff --git a/net/mac80211/wep.c b/net/mac80211/wep.c index e7b6344c900a..35b664d00e23 100644 --- a/net/mac80211/wep.c +++ b/net/mac80211/wep.c @@ -84,20 +84,17 @@ static u8 *ieee80211_wep_add_iv(struct ieee80211_local *local, struct sk_buff *skb, struct ieee80211_key *key) { - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; - u16 fc; - int hdrlen; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data; + unsigned int hdrlen; u8 *newhdr; - fc = le16_to_cpu(hdr->frame_control); - fc |= IEEE80211_FCTL_PROTECTED; - hdr->frame_control = cpu_to_le16(fc); + hdr->frame_control |= cpu_to_le16(IEEE80211_FCTL_PROTECTED); if (WARN_ON(skb_tailroom(skb) < WEP_ICV_LEN || skb_headroom(skb) < WEP_IV_LEN)) return NULL; - hdrlen = ieee80211_get_hdrlen(fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); newhdr = skb_push(skb, WEP_IV_LEN); memmove(newhdr, newhdr + WEP_IV_LEN, hdrlen); ieee80211_wep_get_iv(local, key, newhdr + hdrlen); @@ -109,12 +106,10 @@ static void ieee80211_wep_remove_iv(struct ieee80211_local *local, struct sk_buff *skb, struct ieee80211_key *key) { - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; - u16 fc; - int hdrlen; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data; + unsigned int hdrlen; - fc = le16_to_cpu(hdr->frame_control); - hdrlen = ieee80211_get_hdrlen(fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); memmove(skb->data + WEP_IV_LEN, skb->data, hdrlen); skb_pull(skb, WEP_IV_LEN); } @@ -224,17 +219,15 @@ int ieee80211_wep_decrypt(struct ieee80211_local *local, struct sk_buff *skb, u32 klen; u8 *rc4key; u8 keyidx; - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; - u16 fc; - int hdrlen; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data; + unsigned int hdrlen; size_t len; int ret = 0; - fc = le16_to_cpu(hdr->frame_control); - if (!(fc & IEEE80211_FCTL_PROTECTED)) + if (!ieee80211_has_protected(hdr->frame_control)) return -1; - hdrlen = ieee80211_get_hdrlen(fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); if (skb->len < 8 + hdrlen) return -1; @@ -281,17 +274,15 @@ int ieee80211_wep_decrypt(struct ieee80211_local *local, struct sk_buff *skb, u8 * ieee80211_wep_is_weak_iv(struct sk_buff *skb, struct ieee80211_key *key) { - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; - u16 fc; - int hdrlen; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data; + unsigned int hdrlen; u8 *ivpos; u32 iv; - fc = le16_to_cpu(hdr->frame_control); - if (!(fc & IEEE80211_FCTL_PROTECTED)) + if (!ieee80211_has_protected(hdr->frame_control)) return NULL; - hdrlen = ieee80211_get_hdrlen(fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); ivpos = skb->data + hdrlen; iv = (ivpos[0] << 16) | (ivpos[1] << 8) | ivpos[2]; -- cgit v1.2.2 From 065e9605f941b8bc4dbfa1f14ba98eb0da7e3fbe Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Sun, 22 Jun 2008 16:45:27 -0700 Subject: mac80211: tx.c use new frame control helpers Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/tx.c | 54 +++++++++++++++++++++++++----------------------------- 1 file changed, 25 insertions(+), 29 deletions(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 7a14a39ebd78..bf3600a04776 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -52,9 +52,8 @@ static inline void ieee80211_include_sequence(struct ieee80211_sub_if_data *sdat static void ieee80211_dump_frame(const char *ifname, const char *title, const struct sk_buff *skb) { - const struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; - u16 fc; - int hdrlen; + const struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data; + unsigned int hdrlen; DECLARE_MAC_BUF(mac); printk(KERN_DEBUG "%s: %s (len=%d)", ifname, title, skb->len); @@ -63,13 +62,12 @@ static void ieee80211_dump_frame(const char *ifname, const char *title, return; } - fc = le16_to_cpu(hdr->frame_control); - hdrlen = ieee80211_get_hdrlen(fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); if (hdrlen > skb->len) hdrlen = skb->len; if (hdrlen >= 4) printk(" FC=0x%04x DUR=0x%04x", - fc, le16_to_cpu(hdr->duration_id)); + le16_to_cpu(hdr->frame_control), le16_to_cpu(hdr->duration_id)); if (hdrlen >= 10) printk(" A1=%s", print_mac(mac, hdr->addr1)); if (hdrlen >= 16) @@ -281,7 +279,7 @@ ieee80211_tx_h_sequence(struct ieee80211_tx_data *tx) { struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; - if (ieee80211_get_hdrlen(le16_to_cpu(hdr->frame_control)) >= 24) + if (ieee80211_hdrlen(hdr->frame_control) >= 24) ieee80211_include_sequence(tx->sdata, hdr); return TX_CONTINUE; @@ -542,8 +540,7 @@ ieee80211_tx_h_rate_ctrl(struct ieee80211_tx_data *tx) static ieee80211_tx_result ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) { - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) tx->skb->data; - u16 fc = le16_to_cpu(hdr->frame_control); + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; u16 dur; struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb); struct ieee80211_supported_band *sband; @@ -595,7 +592,7 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) /* Transmit data frames using short preambles if the driver supports * short preambles at the selected rate and short preambles are * available on the network at the current point in time. */ - if (((fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA) && + if (ieee80211_is_data(hdr->frame_control) && (sband->bitrates[tx->rate_idx].flags & IEEE80211_RATE_SHORT_PREAMBLE) && tx->sdata->bss_conf.use_short_preamble && (!tx->sta || test_sta_flags(tx->sta, WLAN_STA_SHORT_PREAMBLE))) { @@ -647,7 +644,7 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) static ieee80211_tx_result ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) { - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) tx->skb->data; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; size_t hdrlen, per_fragm, num_fragm, payload_len, left; struct sk_buff **frags, *first, *frag; int i; @@ -670,7 +667,7 @@ ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) first = tx->skb; - hdrlen = ieee80211_get_hdrlen(tx->fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); payload_len = first->len - hdrlen; per_fragm = frag_threshold - hdrlen - FCS_LEN; num_fragm = DIV_ROUND_UP(payload_len, per_fragm); @@ -1395,7 +1392,8 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, struct ieee80211_tx_info *info; struct ieee80211_sub_if_data *sdata; int ret = 1, head_need; - u16 ethertype, hdrlen, meshhdrlen = 0, fc; + u16 ethertype, hdrlen, meshhdrlen = 0; + __le16 fc; struct ieee80211_hdr hdr; struct ieee80211s_hdr mesh_hdr; const u8 *encaps_data; @@ -1418,12 +1416,12 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, /* convert Ethernet header to proper 802.11 header (based on * operation mode) */ ethertype = (skb->data[12] << 8) | skb->data[13]; - fc = IEEE80211_FTYPE_DATA | IEEE80211_STYPE_DATA; + fc = cpu_to_le16(IEEE80211_FTYPE_DATA | IEEE80211_STYPE_DATA); switch (sdata->vif.type) { case IEEE80211_IF_TYPE_AP: case IEEE80211_IF_TYPE_VLAN: - fc |= IEEE80211_FCTL_FROMDS; + fc |= cpu_to_le16(IEEE80211_FCTL_FROMDS); /* DA BSSID SA */ memcpy(hdr.addr1, skb->data, ETH_ALEN); memcpy(hdr.addr2, dev->dev_addr, ETH_ALEN); @@ -1431,7 +1429,7 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, hdrlen = 24; break; case IEEE80211_IF_TYPE_WDS: - fc |= IEEE80211_FCTL_FROMDS | IEEE80211_FCTL_TODS; + fc |= cpu_to_le16(IEEE80211_FCTL_FROMDS | IEEE80211_FCTL_TODS); /* RA TA DA SA */ memcpy(hdr.addr1, sdata->u.wds.remote_addr, ETH_ALEN); memcpy(hdr.addr2, dev->dev_addr, ETH_ALEN); @@ -1441,7 +1439,7 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, break; #ifdef CONFIG_MAC80211_MESH case IEEE80211_IF_TYPE_MESH_POINT: - fc |= IEEE80211_FCTL_FROMDS | IEEE80211_FCTL_TODS; + fc |= cpu_to_le16(IEEE80211_FCTL_FROMDS | IEEE80211_FCTL_TODS); /* RA TA DA SA */ if (is_multicast_ether_addr(skb->data)) memcpy(hdr.addr1, skb->data, ETH_ALEN); @@ -1471,7 +1469,7 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, break; #endif case IEEE80211_IF_TYPE_STA: - fc |= IEEE80211_FCTL_TODS; + fc |= cpu_to_le16(IEEE80211_FCTL_TODS); /* BSSID SA DA */ memcpy(hdr.addr1, sdata->u.sta.bssid, ETH_ALEN); memcpy(hdr.addr2, skb->data + ETH_ALEN, ETH_ALEN); @@ -1506,7 +1504,7 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, /* receiver and we are QoS enabled, use a QoS type frame */ if (sta_flags & WLAN_STA_WME && ieee80211_num_regular_queues(&local->hw) >= 4) { - fc |= IEEE80211_STYPE_QOS_DATA; + fc |= cpu_to_le16(IEEE80211_STYPE_QOS_DATA); hdrlen += 2; } @@ -1534,7 +1532,7 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, goto fail; } - hdr.frame_control = cpu_to_le16(fc); + hdr.frame_control = fc; hdr.duration_id = 0; hdr.seq_ctrl = 0; @@ -1603,7 +1601,7 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, h_pos += meshhdrlen; } - if (fc & IEEE80211_STYPE_QOS_DATA) { + if (ieee80211_is_data_qos(fc)) { __le16 *qos_control; qos_control = (__le16*) skb_push(skb, 2); @@ -1861,8 +1859,8 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, mgmt = (struct ieee80211_mgmt *) skb_put(skb, 24 + sizeof(mgmt->u.beacon)); memset(mgmt, 0, 24 + sizeof(mgmt->u.beacon)); - mgmt->frame_control = IEEE80211_FC(IEEE80211_FTYPE_MGMT, - IEEE80211_STYPE_BEACON); + mgmt->frame_control = + cpu_to_le16(IEEE80211_FTYPE_MGMT | IEEE80211_STYPE_BEACON); memset(mgmt->da, 0xff, ETH_ALEN); memcpy(mgmt->sa, sdata->dev->dev_addr, ETH_ALEN); /* BSSID is left zeroed, wildcard value */ @@ -1930,10 +1928,9 @@ void ieee80211_rts_get(struct ieee80211_hw *hw, struct ieee80211_vif *vif, struct ieee80211_rts *rts) { const struct ieee80211_hdr *hdr = frame; - u16 fctl; - fctl = IEEE80211_FTYPE_CTL | IEEE80211_STYPE_RTS; - rts->frame_control = cpu_to_le16(fctl); + rts->frame_control = + cpu_to_le16(IEEE80211_FTYPE_CTL | IEEE80211_STYPE_RTS); rts->duration = ieee80211_rts_duration(hw, vif, frame_len, frame_txctl); memcpy(rts->ra, hdr->addr1, sizeof(rts->ra)); @@ -1947,10 +1944,9 @@ void ieee80211_ctstoself_get(struct ieee80211_hw *hw, struct ieee80211_vif *vif, struct ieee80211_cts *cts) { const struct ieee80211_hdr *hdr = frame; - u16 fctl; - fctl = IEEE80211_FTYPE_CTL | IEEE80211_STYPE_CTS; - cts->frame_control = cpu_to_le16(fctl); + cts->frame_control = + cpu_to_le16(IEEE80211_FTYPE_CTL | IEEE80211_STYPE_CTS); cts->duration = ieee80211_ctstoself_duration(hw, vif, frame_len, frame_txctl); memcpy(cts->ra, hdr->addr1, sizeof(cts->ra)); -- cgit v1.2.2 From 182503abf450d39417c2cc6a2c49b4731117d21b Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Sun, 22 Jun 2008 16:45:29 -0700 Subject: mac80211: rx.c use new frame control helpers Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/rx.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index c32a0bcd53b7..8962d1355f04 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -61,7 +61,7 @@ static inline int should_drop_frame(struct ieee80211_rx_status *status, int present_fcs_len, int radiotap_len) { - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data; if (status->flag & (RX_FLAG_FAILED_FCS_CRC | RX_FLAG_FAILED_PLCP_CRC)) return 1; @@ -2123,7 +2123,7 @@ static u8 ieee80211_rx_reorder_ampdu(struct ieee80211_local *local, struct tid_ampdu_rx *tid_agg_rx; u16 sc; u16 mpdu_seq_num; - u8 ret = 0, *qc; + u8 ret = 0; int tid; sta = sta_info_get(local, hdr->addr2); @@ -2135,8 +2135,7 @@ static u8 ieee80211_rx_reorder_ampdu(struct ieee80211_local *local, if (!ieee80211_is_data_qos(hdr->frame_control)) goto end_reorder; - qc = ieee80211_get_qos_ctl(hdr); - tid = qc[0] & QOS_CONTROL_TID_MASK; + tid = *ieee80211_get_qos_ctl(hdr) & QOS_CONTROL_TID_MASK; if (sta->ampdu_mlme.tid_state_rx[tid] != HT_AGG_STATE_OPERATIONAL) goto end_reorder; -- cgit v1.2.2 From 4e3996fe899651e00d3085110cc6e92f6a78ee3e Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Sun, 22 Jun 2008 16:45:32 -0700 Subject: mac80211: mlme.c use new frame control helpers Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 39 +++++++++++++++++++-------------------- 1 file changed, 19 insertions(+), 20 deletions(-) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 75c4e9743fb4..7b4d4d46843b 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -3271,33 +3271,32 @@ ieee80211_sta_rx_scan(struct net_device *dev, struct sk_buff *skb, struct ieee80211_rx_status *rx_status) { struct ieee80211_mgmt *mgmt; - u16 fc; + __le16 fc; if (skb->len < 2) return RX_DROP_UNUSABLE; mgmt = (struct ieee80211_mgmt *) skb->data; - fc = le16_to_cpu(mgmt->frame_control); + fc = mgmt->frame_control; - if ((fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_CTL) + if (ieee80211_is_ctl(fc)) return RX_CONTINUE; if (skb->len < 24) return RX_DROP_MONITOR; - if ((fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_MGMT) { - if ((fc & IEEE80211_FCTL_STYPE) == IEEE80211_STYPE_PROBE_RESP) { - ieee80211_rx_mgmt_probe_resp(dev, mgmt, - skb->len, rx_status); - dev_kfree_skb(skb); - return RX_QUEUED; - } else if ((fc & IEEE80211_FCTL_STYPE) == IEEE80211_STYPE_BEACON) { - ieee80211_rx_mgmt_beacon(dev, mgmt, skb->len, - rx_status); - dev_kfree_skb(skb); - return RX_QUEUED; - } + if (ieee80211_is_probe_resp(fc)) { + ieee80211_rx_mgmt_probe_resp(dev, mgmt, skb->len, rx_status); + dev_kfree_skb(skb); + return RX_QUEUED; } + + if (ieee80211_is_beacon(fc)) { + ieee80211_rx_mgmt_beacon(dev, mgmt, skb->len, rx_status); + dev_kfree_skb(skb); + return RX_QUEUED; + } + return RX_CONTINUE; } @@ -3875,7 +3874,7 @@ static void ieee80211_send_nullfunc(struct ieee80211_local *local, { struct sk_buff *skb; struct ieee80211_hdr *nullfunc; - u16 fc; + __le16 fc; skb = dev_alloc_skb(local->hw.extra_tx_headroom + 24); if (!skb) { @@ -3887,11 +3886,11 @@ static void ieee80211_send_nullfunc(struct ieee80211_local *local, nullfunc = (struct ieee80211_hdr *) skb_put(skb, 24); memset(nullfunc, 0, 24); - fc = IEEE80211_FTYPE_DATA | IEEE80211_STYPE_NULLFUNC | - IEEE80211_FCTL_TODS; + fc = cpu_to_le16(IEEE80211_FTYPE_DATA | IEEE80211_STYPE_NULLFUNC | + IEEE80211_FCTL_TODS); if (powersave) - fc |= IEEE80211_FCTL_PM; - nullfunc->frame_control = cpu_to_le16(fc); + fc |= cpu_to_le16(IEEE80211_FCTL_PM); + nullfunc->frame_control = fc; memcpy(nullfunc->addr1, sdata->u.sta.bssid, ETH_ALEN); memcpy(nullfunc->addr2, sdata->dev->dev_addr, ETH_ALEN); memcpy(nullfunc->addr3, sdata->u.sta.bssid, ETH_ALEN); -- cgit v1.2.2 From fa6adfe9e625a6a843a1abed5f4e7a000c11952c Mon Sep 17 00:00:00 2001 From: Emmanuel Grumbach Date: Tue, 24 Jun 2008 13:37:58 +0300 Subject: mac80211: don't return -EINVAL upon iwconfig wlan0 rts auto This patch avoids returning -EINVAL upon iwconfig wlan0 rts auto. If rts->fixed is 0, then we should choose a default value instead of failing. Signed-off-by: Emmanuel Grumbach Signed-off-by: Tomas Winkler Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/wext.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'net') diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c index f47d13bdf7f7..3cbaf5301d00 100644 --- a/net/mac80211/wext.c +++ b/net/mac80211/wext.c @@ -721,6 +721,9 @@ static int ieee80211_ioctl_siwrts(struct net_device *dev, if (rts->disabled) local->rts_threshold = IEEE80211_MAX_RTS_THRESHOLD; + else if (!rts->fixed) + /* if the rts value is not fixed, then take default */ + local->rts_threshold = IEEE80211_MAX_RTS_THRESHOLD; else if (rts->value < 0 || rts->value > IEEE80211_MAX_RTS_THRESHOLD) return -EINVAL; else -- cgit v1.2.2 From b9fcc4f2987a757acb3af43aa31dc860bb957970 Mon Sep 17 00:00:00 2001 From: Emmanuel Grumbach Date: Tue, 24 Jun 2008 13:37:59 +0300 Subject: mac80211: update the authentication method This patch updates the authentication method upon giwencode ioctl. Signed-off-by: Emmanuel Grumbach Signed-off-by: Tomas Winkler Acked-by: Johannes Berg Acked-by: Dan Williams Signed-off-by: John W. Linville --- net/mac80211/wext.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'net') diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c index 3cbaf5301d00..5e76ab120cd6 100644 --- a/net/mac80211/wext.c +++ b/net/mac80211/wext.c @@ -952,6 +952,19 @@ static int ieee80211_ioctl_giwencode(struct net_device *dev, erq->length = sdata->keys[idx]->conf.keylen; erq->flags |= IW_ENCODE_ENABLED; + if (sdata->vif.type == IEEE80211_IF_TYPE_STA) { + struct ieee80211_if_sta *ifsta = &sdata->u.sta; + switch (ifsta->auth_alg) { + case WLAN_AUTH_OPEN: + case WLAN_AUTH_LEAP: + erq->flags |= IW_ENCODE_OPEN; + break; + case WLAN_AUTH_SHARED_KEY: + erq->flags |= IW_ENCODE_RESTRICTED; + break; + } + } + return 0; } -- cgit v1.2.2 From f37d08bddc5cb8de18e55f2b0a401b3eb6269af4 Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Tue, 24 Jun 2008 15:50:17 +0300 Subject: mac80211: add phy information to giwname This patch add phy information to giwname. Quoting: It's not useless, it's supposed to tell you about the protocol capability of the device, like "IEEE 802.11b" or "IEEE 802.11abg" Jean Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/wext.c | 32 ++++++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+) (limited to 'net') diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c index 5e76ab120cd6..df0531c28141 100644 --- a/net/mac80211/wext.c +++ b/net/mac80211/wext.c @@ -135,7 +135,39 @@ static int ieee80211_ioctl_giwname(struct net_device *dev, struct iw_request_info *info, char *name, char *extra) { + struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); + struct ieee80211_supported_band *sband; + u8 is_ht = 0, is_a = 0, is_b = 0, is_g = 0; + + + sband = local->hw.wiphy->bands[IEEE80211_BAND_5GHZ]; + if (sband) { + is_a = 1; + is_ht |= sband->ht_info.ht_supported; + } + + sband = local->hw.wiphy->bands[IEEE80211_BAND_2GHZ]; + if (sband) { + int i; + /* Check for mandatory rates */ + for (i = 0; i < sband->n_bitrates; i++) { + if (sband->bitrates[i].bitrate == 10) + is_b = 1; + if (sband->bitrates[i].bitrate == 60) + is_g = 1; + } + is_ht |= sband->ht_info.ht_supported; + } + strcpy(name, "IEEE 802.11"); + if (is_a) + strcat(name, "a"); + if (is_b) + strcat(name, "b"); + if (is_g) + strcat(name, "g"); + if (is_ht) + strcat(name, "n"); return 0; } -- cgit v1.2.2 From 59959a6150c8af737898e83f727e824dbed7b0fa Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 26 Jun 2008 19:59:56 +0200 Subject: mac80211: make workqueue freezable This patch makes the mac80211 workqueue freezable making it interact a bit better with system suspend and not try to ping the AP while the hardware is down. This doesn't really help with implementing proper suspend in any way but makes some bad things trigger less. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 5c5396edad32..b661ee5bb824 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -1691,7 +1691,7 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) list_add_tail(&sdata->list, &local->interfaces); name = wiphy_dev(local->hw.wiphy)->driver->name; - local->hw.workqueue = create_singlethread_workqueue(name); + local->hw.workqueue = create_freezeable_workqueue(name); if (!local->hw.workqueue) { result = -ENOMEM; goto fail_workqueue; -- cgit v1.2.2 From 03f93c3d4c8aa9ed2e2b0a949ece658053527d71 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Wed, 25 Jun 2008 14:36:27 +0300 Subject: mac80211: fix tx fragmentation This patch fixes TX fragmentation caused by tx handlers reordering and 'tx info to cb' patches Signed-off-by: Johannes Berg Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/tx.c | 68 ++++++++++++++++++++++++++++++------------------------- 1 file changed, 37 insertions(+), 31 deletions(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index bf3600a04776..52ab85c4341b 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -85,8 +85,8 @@ static inline void ieee80211_dump_frame(const char *ifname, const char *title, } #endif /* CONFIG_MAC80211_LOWTX_FRAME_DUMP */ -static u16 ieee80211_duration(struct ieee80211_tx_data *tx, int group_addr, - int next_frag_len) +static __le16 ieee80211_duration(struct ieee80211_tx_data *tx, int group_addr, + int next_frag_len) { int rate, mrate, erp, dur, i; struct ieee80211_rate *txrate; @@ -138,7 +138,7 @@ static u16 ieee80211_duration(struct ieee80211_tx_data *tx, int group_addr, /* data/mgmt */ if (0 /* FIX: data/mgmt during CFP */) - return 32768; + return cpu_to_le16(32768); if (group_addr) /* Group address as the destination - no ACK */ return 0; @@ -208,7 +208,7 @@ static u16 ieee80211_duration(struct ieee80211_tx_data *tx, int group_addr, tx->sdata->bss_conf.use_short_preamble); } - return dur; + return cpu_to_le16(dur); } static int inline is_ieee80211_device(struct net_device *dev, @@ -541,7 +541,6 @@ static ieee80211_tx_result ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) { struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; - u16 dur; struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb); struct ieee80211_supported_band *sband; @@ -599,14 +598,6 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) info->flags |= IEEE80211_TX_CTL_SHORT_PREAMBLE; } - /* Setup duration field for the first fragment of the frame. Duration - * for remaining fragments will be updated when they are being sent - * to low-level driver in ieee80211_tx(). */ - dur = ieee80211_duration(tx, is_multicast_ether_addr(hdr->addr1), - (tx->flags & IEEE80211_TX_FRAGMENTED) ? - tx->extra_frag[0]->len : 0); - hdr->duration_id = cpu_to_le16(dur); - if ((info->flags & IEEE80211_TX_CTL_USE_RTS_CTS) || (info->flags & IEEE80211_TX_CTL_USE_CTS_PROTECT)) { struct ieee80211_rate *rate; @@ -708,6 +699,8 @@ ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) fhdr->seq_ctrl = cpu_to_le16(seq | ((i + 1) & IEEE80211_SCTL_FRAG)); copylen = left > per_fragm ? per_fragm : left; memcpy(skb_put(frag, copylen), pos, copylen); + memcpy(frag->cb, first->cb, sizeof(frag->cb)); + skb_copy_queue_mapping(frag, first); pos += copylen; left -= copylen; @@ -751,6 +744,36 @@ ieee80211_tx_h_encrypt(struct ieee80211_tx_data *tx) return TX_DROP; } +static ieee80211_tx_result +ieee80211_tx_h_calculate_duration(struct ieee80211_tx_data *tx) +{ + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; + int next_len, i; + int group_addr = is_multicast_ether_addr(hdr->addr1); + + if (!(tx->flags & IEEE80211_TX_FRAGMENTED)) { + hdr->duration_id = ieee80211_duration(tx, group_addr, 0); + return TX_CONTINUE; + } + + hdr->duration_id = ieee80211_duration(tx, group_addr, + tx->extra_frag[0]->len); + + for (i = 0; i < tx->num_extra_frag; i++) { + if (i + 1 < tx->num_extra_frag) { + next_len = tx->extra_frag[i + 1]->len; + } else { + next_len = 0; + tx->rate_idx = tx->last_frag_rate_idx; + } + + hdr = (struct ieee80211_hdr *)tx->extra_frag[i]->data; + hdr->duration_id = ieee80211_duration(tx, 0, next_len); + } + + return TX_CONTINUE; +} + static ieee80211_tx_result ieee80211_tx_h_stats(struct ieee80211_tx_data *tx) { @@ -785,6 +808,7 @@ static ieee80211_tx_handler ieee80211_tx_handlers[] = ieee80211_tx_h_fragment, /* handlers after fragment must be aware of tx info fragmentation! */ ieee80211_tx_h_encrypt, + ieee80211_tx_h_calculate_duration, ieee80211_tx_h_stats, NULL }; @@ -1151,24 +1175,6 @@ static int ieee80211_tx(struct net_device *dev, struct sk_buff *skb) if (invoke_tx_handlers(&tx)) goto out; - if (tx.extra_frag) { - for (i = 0; i < tx.num_extra_frag; i++) { - int next_len, dur; - struct ieee80211_hdr *hdr = - (struct ieee80211_hdr *) - tx.extra_frag[i]->data; - - if (i + 1 < tx.num_extra_frag) { - next_len = tx.extra_frag[i + 1]->len; - } else { - next_len = 0; - tx.rate_idx = tx.last_frag_rate_idx; - } - dur = ieee80211_duration(&tx, 0, next_len); - hdr->duration_id = cpu_to_le16(dur); - } - } - retry: ret = __ieee80211_tx(local, skb, &tx); if (ret) { -- cgit v1.2.2 From 7b1e78d5052a92e95d851fbb0236a712264fe7e8 Mon Sep 17 00:00:00 2001 From: Yi Zhu Date: Wed, 18 Jun 2008 17:53:43 +0300 Subject: mac80211: add MAC80211_VERBOSE_SPECT_MGMT_DEBUG Kconfig option The patch introduces MAC80211_VERBOSE_SPECT_MGMT_DEBUG Kconfig option to suppress Spectrum Management 802.11h related debug logs. Signed-off-by: Zhu Yi Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/Kconfig | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'net') diff --git a/net/mac80211/Kconfig b/net/mac80211/Kconfig index 0d3661d9b6a0..594bc7319ca6 100644 --- a/net/mac80211/Kconfig +++ b/net/mac80211/Kconfig @@ -177,3 +177,10 @@ config MAC80211_VERBOSE_MPL_DEBUG ---help--- Say Y here to print out verbose mesh peer link debug messages. + +config MAC80211_VERBOSE_SPECT_MGMT_DEBUG + bool "Verbose Spectrum Management (IEEE 802.11h)debugging" + depends on MAC80211_DEBUG + ---help--- + Say Y here to print out verbose Spectrum Management (IEEE 802.11h) + debug messages. -- cgit v1.2.2 From 06ff47bc9595848b818ac79e7d8069337c6e58b1 Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Wed, 18 Jun 2008 17:53:44 +0300 Subject: mac80211: add spectrum capabilities This patch add spectrum capability and required information elements to association request providing AP has requested it and it is supported by the driver Signed-off-by: Tomas Winkler Signed-off-by: Assaf Krauss Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 7b4d4d46843b..c9f7c13d2e7e 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -747,6 +747,10 @@ static void ieee80211_send_assoc(struct net_device *dev, * b-only mode) */ rates_len = ieee80211_compatible_rates(bss, sband, &rates); + if ((bss->capability & WLAN_CAPABILITY_SPECTRUM_MGMT) && + (local->hw.flags & IEEE80211_HW_SPECTRUM_MGMT)) + capab |= WLAN_CAPABILITY_SPECTRUM_MGMT; + ieee80211_rx_bss_put(dev, bss); } else { rates = ~0; @@ -814,6 +818,26 @@ static void ieee80211_send_assoc(struct net_device *dev, } } + if (capab & WLAN_CAPABILITY_SPECTRUM_MGMT) { + /* 1. power capabilities */ + pos = skb_put(skb, 4); + *pos++ = WLAN_EID_PWR_CAPABILITY; + *pos++ = 2; + *pos++ = 0; /* min tx power */ + *pos++ = local->hw.conf.channel->max_power; /* max tx power */ + + /* 2. supported channels */ + /* TODO: get this in reg domain format */ + pos = skb_put(skb, 2 * sband->n_channels + 2); + *pos++ = WLAN_EID_SUPPORTED_CHANNELS; + *pos++ = 2 * sband->n_channels; + for (i = 0; i < sband->n_channels; i++) { + *pos++ = ieee80211_frequency_to_channel( + sband->channels[i].center_freq); + *pos++ = 1; /* one channel in the subband*/ + } + } + if (ifsta->extra_ie) { pos = skb_put(skb, ifsta->extra_ie_len); memcpy(pos, ifsta->extra_ie, ifsta->extra_ie_len); -- cgit v1.2.2 From bf998f686430107fb8790ef6713f8e352a3deede Mon Sep 17 00:00:00 2001 From: Emmanuel Grumbach Date: Tue, 24 Jun 2008 13:38:00 +0300 Subject: mac80211: add last beacon time in scan list This patch adds the interval between the scan results and the last time a beacon was received in the result of the scan. Signed-off-by: Emmanuel Grumbach Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index c9f7c13d2e7e..cffef44cec05 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -4340,6 +4340,13 @@ ieee80211_sta_scan_result(struct net_device *dev, current_ev = iwe_stream_add_point(info, current_ev, end_buf, &iwe, buf); + memset(&iwe, 0, sizeof(iwe)); + iwe.cmd = IWEVCUSTOM; + sprintf(buf, " Last beacon: %dms ago", + jiffies_to_msecs(jiffies - bss->last_update)); + iwe.u.data.length = strlen(buf); + current_ev = iwe_stream_add_point(info, current_ev, + end_buf, &iwe, buf); kfree(buf); } } -- cgit v1.2.2 From e5f5e7339cd95d07937e6f8081b46fba86c742a7 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Thu, 26 Jun 2008 13:38:13 +0300 Subject: build algorithms into the mac80211 module The old infrastructure was: - the default algorithm is built into mac80211 - other algorithms get into their own modules The implementation of this complicated scheme was horrible (just look at net/mac80211/Makefile), and anyone adding a new algorithm would most likely not get it right at his first attempt. This patch therefore builds all enabled algorithms into the mac80211 module. The user interface for the rate control algorithms changes as follows: - first the user can choose which algorithms to enable (currently only MAC80211_RC_PID is available) - if more than one algorithm is enabled (currently not possible since only one algorithm is present) the user then chooses the default one Note: - MAC80211_RC_PID is always enables for CONFIG_EMBEDDED=n Technical changes: - all selected algorithms get into the mac80211 module - net/mac80211/Makefile can now become much less complicated - support for rc80211_pid_algo.c being modular is no longer required - this includes unexporting mesh_plink_broken Signed-off-by: Adrian Bunk Signed-off-by: John W. Linville --- net/mac80211/Kconfig | 31 +++++++++---------------------- net/mac80211/Makefile | 18 ++++-------------- net/mac80211/mesh_pathtbl.c | 1 - net/mac80211/rate.h | 4 +--- net/mac80211/rc80211_pid_algo.c | 10 ---------- 5 files changed, 14 insertions(+), 50 deletions(-) (limited to 'net') diff --git a/net/mac80211/Kconfig b/net/mac80211/Kconfig index 594bc7319ca6..661d3c291482 100644 --- a/net/mac80211/Kconfig +++ b/net/mac80211/Kconfig @@ -27,6 +27,14 @@ comment "QoS/HT support needs CONFIG_NETDEVICES_MULTIQUEUE" menu "Rate control algorithm selection" depends on MAC80211 != n +config MAC80211_RC_PID + bool "PID controller based rate control algorithm" if EMBEDDED + default y + ---help--- + This option enables a TX rate control algorithm for + mac80211 that uses a PID controller to select the TX + rate. + choice prompt "Default rate control algorithm" default MAC80211_RC_DEFAULT_PID @@ -38,40 +46,19 @@ choice config MAC80211_RC_DEFAULT_PID bool "PID controller based rate control algorithm" - select MAC80211_RC_PID + depends on MAC80211_RC_PID ---help--- Select the PID controller based rate control as the default rate control algorithm. You should choose this unless you know what you are doing. -config MAC80211_RC_DEFAULT_NONE - bool "No default algorithm" - depends on EMBEDDED - help - Selecting this option will select no default algorithm - and allow you to not build any. Do not choose this - option unless you know your driver comes with another - suitable algorithm. endchoice -comment "Selecting 'y' for an algorithm will" -comment "build the algorithm into mac80211." - config MAC80211_RC_DEFAULT string default "pid" if MAC80211_RC_DEFAULT_PID default "" -config MAC80211_RC_PID - tristate "PID controller based rate control algorithm" - ---help--- - This option enables a TX rate control algorithm for - mac80211 that uses a PID controller to select the TX - rate. - - Say Y or M unless you're sure you want to use a - different rate control algorithm. - endmenu config MAC80211_MESH diff --git a/net/mac80211/Makefile b/net/mac80211/Makefile index 1d2a4e010e5c..fa47438e338f 100644 --- a/net/mac80211/Makefile +++ b/net/mac80211/Makefile @@ -1,13 +1,5 @@ obj-$(CONFIG_MAC80211) += mac80211.o -# objects for PID algorithm -rc80211_pid-y := rc80211_pid_algo.o -rc80211_pid-$(CONFIG_MAC80211_DEBUGFS) += rc80211_pid_debugfs.o - -# build helper for PID algorithm -rc-pid-y := $(rc80211_pid-y) -rc-pid-m := rc80211_pid.o - # mac80211 objects mac80211-y := \ main.o \ @@ -42,10 +34,8 @@ mac80211-$(CONFIG_MAC80211_MESH) += \ mesh_plink.o \ mesh_hwmp.o +# objects for PID algorithm +rc80211_pid-y := rc80211_pid_algo.o +rc80211_pid-$(CONFIG_MAC80211_DEBUGFS) += rc80211_pid_debugfs.o -# Build rate control algorithm(s) -CFLAGS_rc80211_pid_algo.o += -DRC80211_PID_COMPILE -mac80211-$(CONFIG_MAC80211_RC_PID) += $(rc-pid-$(CONFIG_MAC80211_RC_PID)) - -# Modular rate algorithms are assigned to mac80211-m - make separate modules -obj-m += $(mac80211-m) +mac80211-$(CONFIG_MAC80211_RC_PID) += $(rc80211_pid-y) diff --git a/net/mac80211/mesh_pathtbl.c b/net/mac80211/mesh_pathtbl.c index 947b13b40726..5f88a2e6ee50 100644 --- a/net/mac80211/mesh_pathtbl.c +++ b/net/mac80211/mesh_pathtbl.c @@ -262,7 +262,6 @@ void mesh_plink_broken(struct sta_info *sta) } rcu_read_unlock(); } -EXPORT_SYMBOL(mesh_plink_broken); /** * mesh_path_flush_by_nexthop - Deletes mesh paths if their next hop matches diff --git a/net/mac80211/rate.h b/net/mac80211/rate.h index 0ed9c8a2f56f..ede7ab56f65b 100644 --- a/net/mac80211/rate.h +++ b/net/mac80211/rate.h @@ -162,9 +162,7 @@ void rate_control_deinitialize(struct ieee80211_local *local); /* Rate control algorithms */ -#if defined(RC80211_PID_COMPILE) || \ - (defined(CONFIG_MAC80211_RC_PID) && \ - !defined(CONFIG_MAC80211_RC_PID_MODULE)) +#ifdef CONFIG_MAC80211_RC_PID extern int rc80211_pid_init(void); extern void rc80211_pid_exit(void); #else diff --git a/net/mac80211/rc80211_pid_algo.c b/net/mac80211/rc80211_pid_algo.c index e8945413e4a2..62388f8e9024 100644 --- a/net/mac80211/rc80211_pid_algo.c +++ b/net/mac80211/rc80211_pid_algo.c @@ -540,11 +540,6 @@ static struct rate_control_ops mac80211_rcpid = { #endif }; -MODULE_DESCRIPTION("PID controller based rate control algorithm"); -MODULE_AUTHOR("Stefano Brivio"); -MODULE_AUTHOR("Mattias Nissler"); -MODULE_LICENSE("GPL"); - int __init rc80211_pid_init(void) { return ieee80211_rate_control_register(&mac80211_rcpid); @@ -554,8 +549,3 @@ void rc80211_pid_exit(void) { ieee80211_rate_control_unregister(&mac80211_rcpid); } - -#ifdef CONFIG_MAC80211_RC_PID_MODULE -module_init(rc80211_pid_init); -module_exit(rc80211_pid_exit); -#endif -- cgit v1.2.2 From ae6a44e3afae0e813e9ab707a2317b08e21f4ab2 Mon Sep 17 00:00:00 2001 From: Ester Kummer Date: Fri, 27 Jun 2008 18:54:48 +0300 Subject: mac80211: removing duplicated parsing of information elements This patch removes the duplicated parsing of information elements in ieee80211_rx_bss_info and in ieee_rx_mgmt_beacon Signed-off-by: Tomas Winkler Signed-off-by: Ester Kummer Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 188 +++++++++++++++++++++++++++------------------------- 1 file changed, 97 insertions(+), 91 deletions(-) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index cffef44cec05..1e4054b273dd 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -2646,11 +2646,10 @@ static void ieee80211_rx_bss_info(struct net_device *dev, struct ieee80211_mgmt *mgmt, size_t len, struct ieee80211_rx_status *rx_status, + struct ieee802_11_elems *elems, int beacon) { struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee802_11_elems elems; - size_t baselen; int freq, clen; struct ieee80211_sta_bss *bss; struct sta_info *sta; @@ -2669,29 +2668,24 @@ static void ieee80211_rx_bss_info(struct net_device *dev, print_mac(mac, mgmt->sa), print_mac(mac2, mgmt->da)); #endif - baselen = (u8 *) mgmt->u.beacon.variable - (u8 *) mgmt; - if (baselen > len) - return; - beacon_timestamp = le64_to_cpu(mgmt->u.beacon.timestamp); - ieee802_11_parse_elems(mgmt->u.beacon.variable, len - baselen, &elems); - if (ieee80211_vif_is_mesh(&sdata->vif) && elems.mesh_id && - elems.mesh_config && mesh_matches_local(&elems, dev)) { - u64 rates = ieee80211_sta_get_rates(local, &elems, + if (ieee80211_vif_is_mesh(&sdata->vif) && elems->mesh_id && + elems->mesh_config && mesh_matches_local(elems, dev)) { + u64 rates = ieee80211_sta_get_rates(local, elems, rx_status->band); mesh_neighbour_update(mgmt->sa, rates, dev, - mesh_peer_accepts_plinks(&elems, dev)); + mesh_peer_accepts_plinks(elems, dev)); } rcu_read_lock(); - if (sdata->vif.type == IEEE80211_IF_TYPE_IBSS && elems.supp_rates && + if (sdata->vif.type == IEEE80211_IF_TYPE_IBSS && elems->supp_rates && memcmp(mgmt->bssid, sdata->u.sta.bssid, ETH_ALEN) == 0 && (sta = sta_info_get(local, mgmt->sa))) { u64 prev_rates; - u64 supp_rates = ieee80211_sta_get_rates(local, &elems, + u64 supp_rates = ieee80211_sta_get_rates(local, elems, rx_status->band); prev_rates = sta->supp_rates[rx_status->band]; @@ -2716,8 +2710,8 @@ static void ieee80211_rx_bss_info(struct net_device *dev, rcu_read_unlock(); - if (elems.ds_params && elems.ds_params_len == 1) - freq = ieee80211_channel_to_frequency(elems.ds_params[0]); + if (elems->ds_params && elems->ds_params_len == 1) + freq = ieee80211_channel_to_frequency(elems->ds_params[0]); else freq = rx_status->freq; @@ -2727,23 +2721,23 @@ static void ieee80211_rx_bss_info(struct net_device *dev, return; #ifdef CONFIG_MAC80211_MESH - if (elems.mesh_config) - bss = ieee80211_rx_mesh_bss_get(dev, elems.mesh_id, - elems.mesh_id_len, elems.mesh_config, freq); + if (elems->mesh_config) + bss = ieee80211_rx_mesh_bss_get(dev, elems->mesh_id, + elems->mesh_id_len, elems->mesh_config, freq); else #endif bss = ieee80211_rx_bss_get(dev, mgmt->bssid, freq, - elems.ssid, elems.ssid_len); + elems->ssid, elems->ssid_len); if (!bss) { #ifdef CONFIG_MAC80211_MESH - if (elems.mesh_config) - bss = ieee80211_rx_mesh_bss_add(dev, elems.mesh_id, - elems.mesh_id_len, elems.mesh_config, - elems.mesh_config_len, freq); + if (elems->mesh_config) + bss = ieee80211_rx_mesh_bss_add(dev, elems->mesh_id, + elems->mesh_id_len, elems->mesh_config, + elems->mesh_config_len, freq); else #endif bss = ieee80211_rx_bss_add(dev, mgmt->bssid, freq, - elems.ssid, elems.ssid_len); + elems->ssid, elems->ssid_len); if (!bss) return; } else { @@ -2756,43 +2750,43 @@ static void ieee80211_rx_bss_info(struct net_device *dev, } /* save the ERP value so that it is available at association time */ - if (elems.erp_info && elems.erp_info_len >= 1) { - bss->erp_value = elems.erp_info[0]; + if (elems->erp_info && elems->erp_info_len >= 1) { + bss->erp_value = elems->erp_info[0]; bss->has_erp_value = 1; } - if (elems.ht_cap_elem && - (!bss->ht_ie || bss->ht_ie_len != elems.ht_cap_elem_len || - memcmp(bss->ht_ie, elems.ht_cap_elem, elems.ht_cap_elem_len))) { + if (elems->ht_cap_elem && + (!bss->ht_ie || bss->ht_ie_len != elems->ht_cap_elem_len || + memcmp(bss->ht_ie, elems->ht_cap_elem, elems->ht_cap_elem_len))) { kfree(bss->ht_ie); - bss->ht_ie = kmalloc(elems.ht_cap_elem_len + 2, GFP_ATOMIC); + bss->ht_ie = kmalloc(elems->ht_cap_elem_len + 2, GFP_ATOMIC); if (bss->ht_ie) { - memcpy(bss->ht_ie, elems.ht_cap_elem - 2, - elems.ht_cap_elem_len + 2); - bss->ht_ie_len = elems.ht_cap_elem_len + 2; + memcpy(bss->ht_ie, elems->ht_cap_elem - 2, + elems->ht_cap_elem_len + 2); + bss->ht_ie_len = elems->ht_cap_elem_len + 2; } else bss->ht_ie_len = 0; - } else if (!elems.ht_cap_elem && bss->ht_ie) { + } else if (!elems->ht_cap_elem && bss->ht_ie) { kfree(bss->ht_ie); bss->ht_ie = NULL; bss->ht_ie_len = 0; } - if (elems.ht_info_elem && + if (elems->ht_info_elem && (!bss->ht_add_ie || - bss->ht_add_ie_len != elems.ht_info_elem_len || - memcmp(bss->ht_add_ie, elems.ht_info_elem, - elems.ht_info_elem_len))) { + bss->ht_add_ie_len != elems->ht_info_elem_len || + memcmp(bss->ht_add_ie, elems->ht_info_elem, + elems->ht_info_elem_len))) { kfree(bss->ht_add_ie); bss->ht_add_ie = - kmalloc(elems.ht_info_elem_len + 2, GFP_ATOMIC); + kmalloc(elems->ht_info_elem_len + 2, GFP_ATOMIC); if (bss->ht_add_ie) { - memcpy(bss->ht_add_ie, elems.ht_info_elem - 2, - elems.ht_info_elem_len + 2); - bss->ht_add_ie_len = elems.ht_info_elem_len + 2; + memcpy(bss->ht_add_ie, elems->ht_info_elem - 2, + elems->ht_info_elem_len + 2); + bss->ht_add_ie_len = elems->ht_info_elem_len + 2; } else bss->ht_add_ie_len = 0; - } else if (!elems.ht_info_elem && bss->ht_add_ie) { + } else if (!elems->ht_info_elem && bss->ht_add_ie) { kfree(bss->ht_add_ie); bss->ht_add_ie = NULL; bss->ht_add_ie_len = 0; @@ -2802,20 +2796,20 @@ static void ieee80211_rx_bss_info(struct net_device *dev, bss->capability = le16_to_cpu(mgmt->u.beacon.capab_info); bss->supp_rates_len = 0; - if (elems.supp_rates) { + if (elems->supp_rates) { clen = IEEE80211_MAX_SUPP_RATES - bss->supp_rates_len; - if (clen > elems.supp_rates_len) - clen = elems.supp_rates_len; - memcpy(&bss->supp_rates[bss->supp_rates_len], elems.supp_rates, + if (clen > elems->supp_rates_len) + clen = elems->supp_rates_len; + memcpy(&bss->supp_rates[bss->supp_rates_len], elems->supp_rates, clen); bss->supp_rates_len += clen; } - if (elems.ext_supp_rates) { + if (elems->ext_supp_rates) { clen = IEEE80211_MAX_SUPP_RATES - bss->supp_rates_len; - if (clen > elems.ext_supp_rates_len) - clen = elems.ext_supp_rates_len; + if (clen > elems->ext_supp_rates_len) + clen = elems->ext_supp_rates_len; memcpy(&bss->supp_rates[bss->supp_rates_len], - elems.ext_supp_rates, clen); + elems->ext_supp_rates, clen); bss->supp_rates_len += clen; } @@ -2839,33 +2833,33 @@ static void ieee80211_rx_bss_info(struct net_device *dev, return; } - if (elems.wpa && - (!bss->wpa_ie || bss->wpa_ie_len != elems.wpa_len || - memcmp(bss->wpa_ie, elems.wpa, elems.wpa_len))) { + if (elems->wpa && + (!bss->wpa_ie || bss->wpa_ie_len != elems->wpa_len || + memcmp(bss->wpa_ie, elems->wpa, elems->wpa_len))) { kfree(bss->wpa_ie); - bss->wpa_ie = kmalloc(elems.wpa_len + 2, GFP_ATOMIC); + bss->wpa_ie = kmalloc(elems->wpa_len + 2, GFP_ATOMIC); if (bss->wpa_ie) { - memcpy(bss->wpa_ie, elems.wpa - 2, elems.wpa_len + 2); - bss->wpa_ie_len = elems.wpa_len + 2; + memcpy(bss->wpa_ie, elems->wpa - 2, elems->wpa_len + 2); + bss->wpa_ie_len = elems->wpa_len + 2; } else bss->wpa_ie_len = 0; - } else if (!elems.wpa && bss->wpa_ie) { + } else if (!elems->wpa && bss->wpa_ie) { kfree(bss->wpa_ie); bss->wpa_ie = NULL; bss->wpa_ie_len = 0; } - if (elems.rsn && - (!bss->rsn_ie || bss->rsn_ie_len != elems.rsn_len || - memcmp(bss->rsn_ie, elems.rsn, elems.rsn_len))) { + if (elems->rsn && + (!bss->rsn_ie || bss->rsn_ie_len != elems->rsn_len || + memcmp(bss->rsn_ie, elems->rsn, elems->rsn_len))) { kfree(bss->rsn_ie); - bss->rsn_ie = kmalloc(elems.rsn_len + 2, GFP_ATOMIC); + bss->rsn_ie = kmalloc(elems->rsn_len + 2, GFP_ATOMIC); if (bss->rsn_ie) { - memcpy(bss->rsn_ie, elems.rsn - 2, elems.rsn_len + 2); - bss->rsn_ie_len = elems.rsn_len + 2; + memcpy(bss->rsn_ie, elems->rsn - 2, elems->rsn_len + 2); + bss->rsn_ie_len = elems->rsn_len + 2; } else bss->rsn_ie_len = 0; - } else if (!elems.rsn && bss->rsn_ie) { + } else if (!elems->rsn && bss->rsn_ie) { kfree(bss->rsn_ie); bss->rsn_ie = NULL; bss->rsn_ie_len = 0; @@ -2885,20 +2879,21 @@ static void ieee80211_rx_bss_info(struct net_device *dev, * inclusion of the WMM Parameters in beacons, however, is optional. */ - if (elems.wmm_param && - (!bss->wmm_ie || bss->wmm_ie_len != elems.wmm_param_len || - memcmp(bss->wmm_ie, elems.wmm_param, elems.wmm_param_len))) { + if (elems->wmm_param && + (!bss->wmm_ie || bss->wmm_ie_len != elems->wmm_param_len || + memcmp(bss->wmm_ie, elems->wmm_param, elems->wmm_param_len))) { kfree(bss->wmm_ie); - bss->wmm_ie = kmalloc(elems.wmm_param_len + 2, GFP_ATOMIC); + bss->wmm_ie = kmalloc(elems->wmm_param_len + 2, GFP_ATOMIC); if (bss->wmm_ie) { - memcpy(bss->wmm_ie, elems.wmm_param - 2, - elems.wmm_param_len + 2); - bss->wmm_ie_len = elems.wmm_param_len + 2; + memcpy(bss->wmm_ie, elems->wmm_param - 2, + elems->wmm_param_len + 2); + bss->wmm_ie_len = elems->wmm_param_len + 2; } else bss->wmm_ie_len = 0; - } else if (elems.wmm_info && - (!bss->wmm_ie || bss->wmm_ie_len != elems.wmm_info_len || - memcmp(bss->wmm_ie, elems.wmm_info, elems.wmm_info_len))) { + } else if (elems->wmm_info && + (!bss->wmm_ie || bss->wmm_ie_len != elems->wmm_info_len || + memcmp(bss->wmm_ie, elems->wmm_info, + elems->wmm_info_len))) { /* As for certain AP's Fifth bit is not set in WMM IE in * beacon frames.So while parsing the beacon frame the * wmm_info structure is used instead of wmm_param. @@ -2908,14 +2903,14 @@ static void ieee80211_rx_bss_info(struct net_device *dev, * n-band association. */ kfree(bss->wmm_ie); - bss->wmm_ie = kmalloc(elems.wmm_info_len + 2, GFP_ATOMIC); + bss->wmm_ie = kmalloc(elems->wmm_info_len + 2, GFP_ATOMIC); if (bss->wmm_ie) { - memcpy(bss->wmm_ie, elems.wmm_info - 2, - elems.wmm_info_len + 2); - bss->wmm_ie_len = elems.wmm_info_len + 2; + memcpy(bss->wmm_ie, elems->wmm_info - 2, + elems->wmm_info_len + 2); + bss->wmm_ie_len = elems->wmm_info_len + 2; } else bss->wmm_ie_len = 0; - } else if (!elems.wmm_param && !elems.wmm_info && bss->wmm_ie) { + } else if (!elems->wmm_param && !elems->wmm_info && bss->wmm_ie) { kfree(bss->wmm_ie); bss->wmm_ie = NULL; bss->wmm_ie_len = 0; @@ -2926,8 +2921,9 @@ static void ieee80211_rx_bss_info(struct net_device *dev, !local->sta_sw_scanning && !local->sta_hw_scanning && bss->capability & WLAN_CAPABILITY_IBSS && bss->freq == local->oper_channel->center_freq && - elems.ssid_len == sdata->u.sta.ssid_len && - memcmp(elems.ssid, sdata->u.sta.ssid, sdata->u.sta.ssid_len) == 0) { + elems->ssid_len == sdata->u.sta.ssid_len && + memcmp(elems->ssid, sdata->u.sta.ssid, + sdata->u.sta.ssid_len) == 0) { if (rx_status->flag & RX_FLAG_TSFT) { /* in order for correct IBSS merging we need mactime * @@ -2986,7 +2982,17 @@ static void ieee80211_rx_mgmt_probe_resp(struct net_device *dev, size_t len, struct ieee80211_rx_status *rx_status) { - ieee80211_rx_bss_info(dev, mgmt, len, rx_status, 0); + size_t baselen; + struct ieee802_11_elems elems; + + baselen = (u8 *) mgmt->u.probe_resp.variable - (u8 *) mgmt; + if (baselen > len) + return; + + ieee802_11_parse_elems(mgmt->u.probe_resp.variable, len - baselen, + &elems); + + ieee80211_rx_bss_info(dev, mgmt, len, rx_status, &elems, 0); } @@ -3003,7 +3009,14 @@ static void ieee80211_rx_mgmt_beacon(struct net_device *dev, struct ieee80211_conf *conf = &local->hw.conf; u32 changed = 0; - ieee80211_rx_bss_info(dev, mgmt, len, rx_status, 1); + /* Process beacon from the current BSS */ + baselen = (u8 *) mgmt->u.beacon.variable - (u8 *) mgmt; + if (baselen > len) + return; + + ieee802_11_parse_elems(mgmt->u.beacon.variable, len - baselen, &elems); + + ieee80211_rx_bss_info(dev, mgmt, len, rx_status, &elems, 1); sdata = IEEE80211_DEV_TO_SUB_IF(dev); if (sdata->vif.type != IEEE80211_IF_TYPE_STA) @@ -3014,13 +3027,6 @@ static void ieee80211_rx_mgmt_beacon(struct net_device *dev, memcmp(ifsta->bssid, mgmt->bssid, ETH_ALEN) != 0) return; - /* Process beacon from the current BSS */ - baselen = (u8 *) mgmt->u.beacon.variable - (u8 *) mgmt; - if (baselen > len) - return; - - ieee802_11_parse_elems(mgmt->u.beacon.variable, len - baselen, &elems); - ieee80211_sta_wmm_params(dev, ifsta, elems.wmm_param, elems.wmm_param_len); -- cgit v1.2.2 From 5479d0e73958710b9a255337eaa2ba47eb492def Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Sat, 28 Jun 2008 03:15:03 +0300 Subject: mac80211: fix warning: unused variable invoke_tx_handlers This patch fixes warning: unused variable in invoke_tx_handlers when compiling without MAC80211_DEBUG option Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/tx.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 52ab85c4341b..8f1c574bc8fb 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -1110,7 +1110,6 @@ static int __ieee80211_tx(struct ieee80211_local *local, struct sk_buff *skb, */ static int invoke_tx_handlers(struct ieee80211_tx_data *tx) { - struct ieee80211_local *local = tx->local; struct sk_buff *skb = tx->skb; ieee80211_tx_handler *handler; ieee80211_tx_result res = TX_DROP; @@ -1123,7 +1122,7 @@ static int invoke_tx_handlers(struct ieee80211_tx_data *tx) } if (unlikely(res == TX_DROP)) { - I802_DEBUG_INC(local->tx_handlers_drop); + I802_DEBUG_INC(tx->local->tx_handlers_drop); dev_kfree_skb(skb); for (i = 0; i < tx->num_extra_frag; i++) if (tx->extra_frag[i]) @@ -1131,7 +1130,7 @@ static int invoke_tx_handlers(struct ieee80211_tx_data *tx) kfree(tx->extra_frag); return -1; } else if (unlikely(res == TX_QUEUED)) { - I802_DEBUG_INC(local->tx_handlers_queued); + I802_DEBUG_INC(tx->local->tx_handlers_queued); return -1; } -- cgit v1.2.2 From 4faeb86070c38c0df9f5a23c3f6acf5538492a33 Mon Sep 17 00:00:00 2001 From: Assaf Krauss Date: Mon, 30 Jun 2008 17:23:16 +0800 Subject: mac80211: add beacon timestamp to beacon template in IBSS This patch adds a beacon timestamp to the beacon template used in IBSS mode. This way the underlying driver can update its TSF accordingly. According the spec station should adopt the highest TSF from an incoming beacons in the cell. Signed-off-by: Assaf Krauss Signed-off-by: Zhu Yi Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 1 + 1 file changed, 1 insertion(+) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 1e4054b273dd..0a310d09ab07 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -2503,6 +2503,7 @@ static int ieee80211_sta_join_ibss(struct net_device *dev, memcpy(mgmt->bssid, ifsta->bssid, ETH_ALEN); mgmt->u.beacon.beacon_int = cpu_to_le16(local->hw.conf.beacon_int); + mgmt->u.beacon.timestamp = cpu_to_le64(bss->timestamp); mgmt->u.beacon.capab_info = cpu_to_le16(bss->capability); pos = skb_put(skb, 2 + ifsta->ssid_len); -- cgit v1.2.2 From 6dbf4bcac98bbc76ef425b3a2b4169f31199f6c7 Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Tue, 1 Jul 2008 19:29:07 -0700 Subject: icmp: fix units for ratelimit Convert the sysctl values for icmp ratelimit to use milliseconds instead of jiffies which is based on kernel configured HZ. Internal kernel jiffies are not a proper unit for any userspace API. Signed-off-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/ipv4/sysctl_net_ipv4.c | 3 ++- net/ipv6/icmp.c | 3 ++- 2 files changed, 4 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c index 901607003205..14ef202a2254 100644 --- a/net/ipv4/sysctl_net_ipv4.c +++ b/net/ipv4/sysctl_net_ipv4.c @@ -793,7 +793,8 @@ static struct ctl_table ipv4_net_table[] = { .data = &init_net.ipv4.sysctl_icmp_ratelimit, .maxlen = sizeof(int), .mode = 0644, - .proc_handler = &proc_dointvec + .proc_handler = &proc_dointvec_ms_jiffies, + .strategy = &sysctl_ms_jiffies }, { .ctl_name = NET_IPV4_ICMP_RATEMASK, diff --git a/net/ipv6/icmp.c b/net/ipv6/icmp.c index 399d41f65437..abedf95fdf2d 100644 --- a/net/ipv6/icmp.c +++ b/net/ipv6/icmp.c @@ -954,7 +954,8 @@ ctl_table ipv6_icmp_table_template[] = { .data = &init_net.ipv6.sysctl.icmpv6_time, .maxlen = sizeof(int), .mode = 0644, - .proc_handler = &proc_dointvec + .proc_handler = &proc_dointvec_ms_jiffies, + .strategy = &sysctl_ms_jiffies }, { .ctl_name = 0 }, }; -- cgit v1.2.2 From ecbed6a41900126e7b9509e12a8d0cc22176e3eb Mon Sep 17 00:00:00 2001 From: Vlad Yasevich Date: Tue, 1 Jul 2008 20:06:22 -0700 Subject: sctp: Mark GET_PEER|LOCAL_ADDR_OLD deprecated. Socket options SCTP_GET_PEER_ADDR_OLD, SCTP_GET_PEER_ADDR_NUM_OLD, SCTP_GET_LOCAL_ADDR_OLD, and SCTP_GET_PEER_LOCAL_ADDR_NUM_OLD have been replaced by newer versions a since 2005. It's time to officially deprecate them and schedule them for removal. Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/socket.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'net') diff --git a/net/sctp/socket.c b/net/sctp/socket.c index 43460a1cb6d0..df5572c39f0c 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -4223,6 +4223,8 @@ static int sctp_getsockopt_peer_addrs_num_old(struct sock *sk, int len, if (copy_from_user(&id, optval, sizeof(sctp_assoc_t))) return -EFAULT; + printk(KERN_WARNING "SCTP: Use of SCTP_GET_PEER_ADDRS_NUM_OLD " + "socket option deprecated\n"); /* For UDP-style sockets, id specifies the association to query. */ asoc = sctp_id2assoc(sk, id); if (!asoc) @@ -4262,6 +4264,9 @@ static int sctp_getsockopt_peer_addrs_old(struct sock *sk, int len, if (getaddrs.addr_num <= 0) return -EINVAL; + printk(KERN_WARNING "SCTP: Use of SCTP_GET_PEER_ADDRS_OLD " + "socket option deprecated\n"); + /* For UDP-style sockets, id specifies the association to query. */ asoc = sctp_id2assoc(sk, getaddrs.assoc_id); if (!asoc) @@ -4355,6 +4360,9 @@ static int sctp_getsockopt_local_addrs_num_old(struct sock *sk, int len, if (copy_from_user(&id, optval, sizeof(sctp_assoc_t))) return -EFAULT; + printk(KERN_WARNING "SCTP: Use of SCTP_GET_LOCAL_ADDRS_NUM_OLD " + "socket option deprecated\n"); + /* * For UDP-style sockets, id specifies the association to query. * If the id field is set to the value '0' then the locally bound @@ -4515,6 +4523,10 @@ static int sctp_getsockopt_local_addrs_old(struct sock *sk, int len, if (getaddrs.addr_num <= 0 || getaddrs.addr_num >= (INT_MAX / sizeof(union sctp_addr))) return -EINVAL; + + printk(KERN_WARNING "SCTP: Use of SCTP_GET_LOCAL_ADDRS_OLD " + "socket option deprecated\n"); + /* * For UDP-style sockets, id specifies the association to query. * If the id field is set to the value '0' then the locally bound -- cgit v1.2.2 From 0853ad66b14feb12acde7ac13b7c3b75770a0adc Mon Sep 17 00:00:00 2001 From: Santwona Behera Date: Wed, 2 Jul 2008 03:47:41 -0700 Subject: netdev: Add support for rx flow hash configuration, using ethtool. Added new interfaces to ethtool to configure receive network flow distribution across multiple rx rings using hashing. Signed-off-by: Santwona Behera Signed-off-by: David S. Miller --- net/core/ethtool.c | 37 +++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+) (limited to 'net') diff --git a/net/core/ethtool.c b/net/core/ethtool.c index 0133b5ebd545..14ada537f895 100644 --- a/net/core/ethtool.c +++ b/net/core/ethtool.c @@ -209,6 +209,36 @@ static int ethtool_get_drvinfo(struct net_device *dev, void __user *useraddr) return 0; } +static int ethtool_set_rxhash(struct net_device *dev, void __user *useraddr) +{ + struct ethtool_rxnfc cmd; + + if (!dev->ethtool_ops->set_rxhash) + return -EOPNOTSUPP; + + if (copy_from_user(&cmd, useraddr, sizeof(cmd))) + return -EFAULT; + + return dev->ethtool_ops->set_rxhash(dev, &cmd); +} + +static int ethtool_get_rxhash(struct net_device *dev, void __user *useraddr) +{ + struct ethtool_rxnfc info; + + if (!dev->ethtool_ops->get_rxhash) + return -EOPNOTSUPP; + + if (copy_from_user(&info, useraddr, sizeof(info))) + return -EFAULT; + + dev->ethtool_ops->get_rxhash(dev, &info); + + if (copy_to_user(useraddr, &info, sizeof(info))) + return -EFAULT; + return 0; +} + static int ethtool_get_regs(struct net_device *dev, char __user *useraddr) { struct ethtool_regs regs; @@ -826,6 +856,7 @@ int dev_ethtool(struct net *net, struct ifreq *ifr) case ETHTOOL_GGSO: case ETHTOOL_GFLAGS: case ETHTOOL_GPFLAGS: + case ETHTOOL_GRXFH: break; default: if (!capable(CAP_NET_ADMIN)) @@ -977,6 +1008,12 @@ int dev_ethtool(struct net *net, struct ifreq *ifr) rc = ethtool_set_value(dev, useraddr, dev->ethtool_ops->set_priv_flags); break; + case ETHTOOL_GRXFH: + rc = ethtool_get_rxhash(dev, useraddr); + break; + case ETHTOOL_SRXFH: + rc = ethtool_set_rxhash(dev, useraddr); + break; default: rc = -EOPNOTSUPP; } -- cgit v1.2.2 From d9e8a70fa20dc3eaa00859a6eac0adfaef910c77 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Mon, 30 Jun 2008 15:10:44 +0200 Subject: mac80211: get rid of function pointers in TX path This changes the TX path to no longer use function pointers for TX handlers but rather invoke them directly. If debugging is enabled, mark the TX handlers noinline because otherwise they all get inlined into invoke_tx_handlers() which makes it harder to see where a bug is. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 6 +++++ net/mac80211/tx.c | 63 +++++++++++++++++++++------------------------- 2 files changed, 35 insertions(+), 34 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index af352c05c983..46705ae7a847 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -954,4 +954,10 @@ int ieee80211_frame_duration(struct ieee80211_local *local, size_t len, void mac80211_ev_michael_mic_failure(struct net_device *dev, int keyidx, struct ieee80211_hdr *hdr); +#ifdef CONFIG_MAC80211_DEBUG +#define debug_noinline noinline +#else +#define debug_noinline +#endif + #endif /* IEEE80211_I_H */ diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 8f1c574bc8fb..7b930d3c2fba 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -220,7 +220,7 @@ static int inline is_ieee80211_device(struct net_device *dev, /* tx handlers */ -static ieee80211_tx_result +static ieee80211_tx_result debug_noinline ieee80211_tx_h_check_assoc(struct ieee80211_tx_data *tx) { #ifdef CONFIG_MAC80211_VERBOSE_DEBUG @@ -274,7 +274,7 @@ ieee80211_tx_h_check_assoc(struct ieee80211_tx_data *tx) return TX_CONTINUE; } -static ieee80211_tx_result +static ieee80211_tx_result debug_noinline ieee80211_tx_h_sequence(struct ieee80211_tx_data *tx) { struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; @@ -432,7 +432,7 @@ ieee80211_tx_h_unicast_ps_buf(struct ieee80211_tx_data *tx) return TX_CONTINUE; } -static ieee80211_tx_result +static ieee80211_tx_result debug_noinline ieee80211_tx_h_ps_buf(struct ieee80211_tx_data *tx) { if (unlikely(tx->flags & IEEE80211_TX_PS_BUFFERED)) @@ -444,7 +444,7 @@ ieee80211_tx_h_ps_buf(struct ieee80211_tx_data *tx) return ieee80211_tx_h_multicast_ps_buf(tx); } -static ieee80211_tx_result +static ieee80211_tx_result debug_noinline ieee80211_tx_h_select_key(struct ieee80211_tx_data *tx) { struct ieee80211_key *key; @@ -493,7 +493,7 @@ ieee80211_tx_h_select_key(struct ieee80211_tx_data *tx) return TX_CONTINUE; } -static ieee80211_tx_result +static ieee80211_tx_result debug_noinline ieee80211_tx_h_rate_ctrl(struct ieee80211_tx_data *tx) { struct rate_selection rsel; @@ -537,7 +537,7 @@ ieee80211_tx_h_rate_ctrl(struct ieee80211_tx_data *tx) return TX_CONTINUE; } -static ieee80211_tx_result +static ieee80211_tx_result debug_noinline ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) { struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; @@ -632,7 +632,7 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) return TX_CONTINUE; } -static ieee80211_tx_result +static ieee80211_tx_result debug_noinline ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) { struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; @@ -724,7 +724,7 @@ ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) return TX_DROP; } -static ieee80211_tx_result +static ieee80211_tx_result debug_noinline ieee80211_tx_h_encrypt(struct ieee80211_tx_data *tx) { if (!tx->key) @@ -744,7 +744,7 @@ ieee80211_tx_h_encrypt(struct ieee80211_tx_data *tx) return TX_DROP; } -static ieee80211_tx_result +static ieee80211_tx_result debug_noinline ieee80211_tx_h_calculate_duration(struct ieee80211_tx_data *tx) { struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; @@ -774,7 +774,7 @@ ieee80211_tx_h_calculate_duration(struct ieee80211_tx_data *tx) return TX_CONTINUE; } -static ieee80211_tx_result +static ieee80211_tx_result debug_noinline ieee80211_tx_h_stats(struct ieee80211_tx_data *tx) { int i; @@ -795,24 +795,6 @@ ieee80211_tx_h_stats(struct ieee80211_tx_data *tx) } -typedef ieee80211_tx_result (*ieee80211_tx_handler)(struct ieee80211_tx_data *); -static ieee80211_tx_handler ieee80211_tx_handlers[] = -{ - ieee80211_tx_h_check_assoc, - ieee80211_tx_h_sequence, - ieee80211_tx_h_ps_buf, - ieee80211_tx_h_select_key, - ieee80211_tx_h_michael_mic_add, - ieee80211_tx_h_rate_ctrl, - ieee80211_tx_h_misc, - ieee80211_tx_h_fragment, - /* handlers after fragment must be aware of tx info fragmentation! */ - ieee80211_tx_h_encrypt, - ieee80211_tx_h_calculate_duration, - ieee80211_tx_h_stats, - NULL -}; - /* actual transmit path */ /* @@ -1111,16 +1093,29 @@ static int __ieee80211_tx(struct ieee80211_local *local, struct sk_buff *skb, static int invoke_tx_handlers(struct ieee80211_tx_data *tx) { struct sk_buff *skb = tx->skb; - ieee80211_tx_handler *handler; ieee80211_tx_result res = TX_DROP; int i; - for (handler = ieee80211_tx_handlers; *handler != NULL; handler++) { - res = (*handler)(tx); - if (res != TX_CONTINUE) - break; - } +#define CALL_TXH(txh) \ + res = txh(tx); \ + if (res != TX_CONTINUE) \ + goto txh_done; + + CALL_TXH(ieee80211_tx_h_check_assoc) + CALL_TXH(ieee80211_tx_h_sequence) + CALL_TXH(ieee80211_tx_h_ps_buf) + CALL_TXH(ieee80211_tx_h_select_key) + CALL_TXH(ieee80211_tx_h_michael_mic_add) + CALL_TXH(ieee80211_tx_h_rate_ctrl) + CALL_TXH(ieee80211_tx_h_misc) + CALL_TXH(ieee80211_tx_h_fragment) + /* handlers after fragment must be aware of tx info fragmentation! */ + CALL_TXH(ieee80211_tx_h_encrypt) + CALL_TXH(ieee80211_tx_h_calculate_duration) + CALL_TXH(ieee80211_tx_h_stats) +#undef CALL_TXH + txh_done: if (unlikely(res == TX_DROP)) { I802_DEBUG_INC(tx->local->tx_handlers_drop); dev_kfree_skb(skb); -- cgit v1.2.2 From 49461622edf74cd1e1a1056cee3ca8dd90cd9556 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Mon, 30 Jun 2008 15:10:45 +0200 Subject: mac80211: get rid of function pointers in RX path This changes the RX path to no longer use function pointers for RX handlers but rather invoke them directly. If debugging is enabled, mark the RX handlers noinline because otherwise they all get inlined into ieee80211_invoke_rx_handlers() which makes it harder to see where a bug is. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/rx.c | 97 +++++++++++++++++++++++++------------------------------ 1 file changed, 44 insertions(+), 53 deletions(-) (limited to 'net') diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index 8962d1355f04..289112777e90 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -386,7 +386,7 @@ static void ieee80211_verify_ip_alignment(struct ieee80211_rx_data *rx) /* rx handlers */ -static ieee80211_rx_result +static ieee80211_rx_result debug_noinline ieee80211_rx_h_passive_scan(struct ieee80211_rx_data *rx) { struct ieee80211_local *local = rx->local; @@ -463,7 +463,7 @@ ieee80211_rx_mesh_check(struct ieee80211_rx_data *rx) } -static ieee80211_rx_result +static ieee80211_rx_result debug_noinline ieee80211_rx_h_check(struct ieee80211_rx_data *rx) { struct ieee80211_hdr *hdr; @@ -522,7 +522,7 @@ ieee80211_rx_h_check(struct ieee80211_rx_data *rx) } -static ieee80211_rx_result +static ieee80211_rx_result debug_noinline ieee80211_rx_h_decrypt(struct ieee80211_rx_data *rx) { struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) rx->skb->data; @@ -710,7 +710,7 @@ static int ap_sta_ps_end(struct net_device *dev, struct sta_info *sta) return sent; } -static ieee80211_rx_result +static ieee80211_rx_result debug_noinline ieee80211_rx_h_sta_process(struct ieee80211_rx_data *rx) { struct sta_info *sta = rx->sta; @@ -858,7 +858,7 @@ ieee80211_reassemble_find(struct ieee80211_sub_if_data *sdata, return NULL; } -static ieee80211_rx_result +static ieee80211_rx_result debug_noinline ieee80211_rx_h_defragment(struct ieee80211_rx_data *rx) { struct ieee80211_hdr *hdr; @@ -974,7 +974,7 @@ ieee80211_rx_h_defragment(struct ieee80211_rx_data *rx) return RX_CONTINUE; } -static ieee80211_rx_result +static ieee80211_rx_result debug_noinline ieee80211_rx_h_ps_poll(struct ieee80211_rx_data *rx) { struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(rx->dev); @@ -1049,7 +1049,7 @@ ieee80211_rx_h_ps_poll(struct ieee80211_rx_data *rx) return RX_QUEUED; } -static ieee80211_rx_result +static ieee80211_rx_result debug_noinline ieee80211_rx_h_remove_qos_control(struct ieee80211_rx_data *rx) { u16 fc = rx->fc; @@ -1367,7 +1367,7 @@ ieee80211_deliver_skb(struct ieee80211_rx_data *rx) } } -static ieee80211_rx_result +static ieee80211_rx_result debug_noinline ieee80211_rx_h_amsdu(struct ieee80211_rx_data *rx) { struct net_device *dev = rx->dev; @@ -1484,7 +1484,7 @@ ieee80211_rx_h_amsdu(struct ieee80211_rx_data *rx) return RX_QUEUED; } -static ieee80211_rx_result +static ieee80211_rx_result debug_noinline ieee80211_rx_h_data(struct ieee80211_rx_data *rx) { struct net_device *dev = rx->dev; @@ -1515,7 +1515,7 @@ ieee80211_rx_h_data(struct ieee80211_rx_data *rx) return RX_QUEUED; } -static ieee80211_rx_result +static ieee80211_rx_result debug_noinline ieee80211_rx_h_ctrl(struct ieee80211_rx_data *rx) { struct ieee80211_local *local = rx->local; @@ -1559,7 +1559,7 @@ ieee80211_rx_h_ctrl(struct ieee80211_rx_data *rx) return RX_CONTINUE; } -static ieee80211_rx_result +static ieee80211_rx_result debug_noinline ieee80211_rx_h_mgmt(struct ieee80211_rx_data *rx) { struct ieee80211_sub_if_data *sdata; @@ -1732,66 +1732,57 @@ static void ieee80211_rx_cooked_monitor(struct ieee80211_rx_data *rx) dev_kfree_skb(skb); } -typedef ieee80211_rx_result (*ieee80211_rx_handler)(struct ieee80211_rx_data *); -static ieee80211_rx_handler ieee80211_rx_handlers[] = -{ - ieee80211_rx_h_passive_scan, - ieee80211_rx_h_check, - ieee80211_rx_h_decrypt, - ieee80211_rx_h_sta_process, - ieee80211_rx_h_defragment, - ieee80211_rx_h_ps_poll, - ieee80211_rx_h_michael_mic_verify, - /* this must be after decryption - so header is counted in MPDU mic - * must be before pae and data, so QOS_DATA format frames - * are not passed to user space by these functions - */ - ieee80211_rx_h_remove_qos_control, - ieee80211_rx_h_amsdu, - ieee80211_rx_h_data, - ieee80211_rx_h_ctrl, - ieee80211_rx_h_mgmt, - NULL -}; static void ieee80211_invoke_rx_handlers(struct ieee80211_sub_if_data *sdata, struct ieee80211_rx_data *rx, struct sk_buff *skb) { - ieee80211_rx_handler *handler; ieee80211_rx_result res = RX_DROP_MONITOR; rx->skb = skb; rx->sdata = sdata; rx->dev = sdata->dev; - for (handler = ieee80211_rx_handlers; *handler != NULL; handler++) { - res = (*handler)(rx); - - switch (res) { - case RX_CONTINUE: - continue; - case RX_DROP_UNUSABLE: - case RX_DROP_MONITOR: - I802_DEBUG_INC(sdata->local->rx_handlers_drop); - if (rx->sta) - rx->sta->rx_dropped++; - break; - case RX_QUEUED: - I802_DEBUG_INC(sdata->local->rx_handlers_queued); - break; - } - break; - } - +#define CALL_RXH(rxh) \ + res = rxh(rx); \ + if (res != RX_CONTINUE) \ + goto rxh_done; + + CALL_RXH(ieee80211_rx_h_passive_scan) + CALL_RXH(ieee80211_rx_h_check) + CALL_RXH(ieee80211_rx_h_decrypt) + CALL_RXH(ieee80211_rx_h_sta_process) + CALL_RXH(ieee80211_rx_h_defragment) + CALL_RXH(ieee80211_rx_h_ps_poll) + CALL_RXH(ieee80211_rx_h_michael_mic_verify) + /* must be after MMIC verify so header is counted in MPDU mic */ + CALL_RXH(ieee80211_rx_h_remove_qos_control) + CALL_RXH(ieee80211_rx_h_amsdu) + CALL_RXH(ieee80211_rx_h_data) + CALL_RXH(ieee80211_rx_h_ctrl) + CALL_RXH(ieee80211_rx_h_mgmt) + +#undef CALL_RXH + + rxh_done: switch (res) { - case RX_CONTINUE: case RX_DROP_MONITOR: + I802_DEBUG_INC(sdata->local->rx_handlers_drop); + if (rx->sta) + rx->sta->rx_dropped++; + /* fall through */ + case RX_CONTINUE: ieee80211_rx_cooked_monitor(rx); break; case RX_DROP_UNUSABLE: + I802_DEBUG_INC(sdata->local->rx_handlers_drop); + if (rx->sta) + rx->sta->rx_dropped++; dev_kfree_skb(rx->skb); break; + case RX_QUEUED: + I802_DEBUG_INC(sdata->local->rx_handlers_queued); + break; } } -- cgit v1.2.2 From f4ea83dd743d3e1bec8fdf954ac911c6b12ae87a Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Mon, 30 Jun 2008 15:10:46 +0200 Subject: mac80211: rework debug settings and make debugging safer This patch reworks the mac80211 debug settings making them more focused and adding help text for those that didn't have one. It also removes a number of printks that can be triggered remotely and add no value, e.g. "too short deauthentication frame received - ignoring". If somebody really needs to debug that they should just add a monitor interface and look at the frames in wireshark. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/Kconfig | 130 +++++++++++++++++++--------- net/mac80211/ieee80211_i.h | 2 +- net/mac80211/iface.c | 6 +- net/mac80211/main.c | 35 ++++++-- net/mac80211/mlme.c | 206 +++++++++++---------------------------------- net/mac80211/rx.c | 106 +++-------------------- net/mac80211/sta_info.c | 2 + net/mac80211/tx.c | 9 +- net/mac80211/wep.c | 13 +-- net/mac80211/wme.c | 1 - net/mac80211/wpa.c | 28 +----- 11 files changed, 196 insertions(+), 342 deletions(-) (limited to 'net') diff --git a/net/mac80211/Kconfig b/net/mac80211/Kconfig index 661d3c291482..11a1e7fa195d 100644 --- a/net/mac80211/Kconfig +++ b/net/mac80211/Kconfig @@ -88,10 +88,16 @@ config MAC80211_DEBUGFS Say N unless you know you need this. +menuconfig MAC80211_DEBUG_MENU + bool "Select mac80211 debugging features" + depends on MAC80211 + ---help--- + This option collects various mac80211 debug settings. + config MAC80211_DEBUG_PACKET_ALIGNMENT bool "Enable packet alignment debugging" - depends on MAC80211 - help + depends on MAC80211_DEBUG_MENU + ---help--- This option is recommended for driver authors and strongly discouraged for everybody else, it will trigger a warning when a driver hands mac80211 a buffer that is aligned in @@ -100,33 +106,95 @@ config MAC80211_DEBUG_PACKET_ALIGNMENT Say N unless you're writing a mac80211 based driver. -config MAC80211_DEBUG - bool "Enable debugging output" - depends on MAC80211 +config MAC80211_NOINLINE + bool "Do not inline TX/RX handlers" + depends on MAC80211_DEBUG_MENU ---help--- - This option will enable debug tracing output for the - ieee80211 network stack. + This option affects code generation in mac80211, when + selected some functions are marked "noinline" to allow + easier debugging of problems in the transmit and receive + paths. + + This option increases code size a bit and inserts a lot + of function calls in the code, but is otherwise safe to + enable. - If you are not trying to debug or develop the ieee80211 - subsystem, you most likely want to say N here. + If unsure, say N unless you expect to be finding problems + in mac80211. + +config MAC80211_VERBOSE_DEBUG + bool "Verbose debugging output" + depends on MAC80211_DEBUG_MENU + ---help--- + Selecting this option causes mac80211 to print out + many debugging messages. It should not be selected + on production systems as some of the messages are + remotely triggerable. + + Do not select this option. config MAC80211_HT_DEBUG - bool "Enable HT debugging output" - depends on MAC80211_DEBUG + bool "Verbose HT debugging" + depends on MAC80211_DEBUG_MENU ---help--- This option enables 802.11n High Throughput features debug tracing output. - If you are not trying to debug of develop the ieee80211 - subsystem, you most likely want to say N here. + It should not be selected on production systems as some + of the messages are remotely triggerable. -config MAC80211_VERBOSE_DEBUG - bool "Verbose debugging output" - depends on MAC80211_DEBUG + Do not select this option. + +config MAC80211_TKIP_DEBUG + bool "Verbose TKIP debugging" + depends on MAC80211_DEBUG_MENU + ---help--- + Selecting this option causes mac80211 to print out + very verbose TKIP debugging messages. It should not + be selected on production systems as those messages + are remotely triggerable. + + Do not select this option. + +config MAC80211_IBSS_DEBUG + bool "Verbose IBSS debugging" + depends on MAC80211_DEBUG_MENU + ---help--- + Selecting this option causes mac80211 to print out + very verbose IBSS debugging messages. It should not + be selected on production systems as those messages + are remotely triggerable. + + Do not select this option. + +config MAC80211_VERBOSE_PS_DEBUG + bool "Verbose powersave mode debugging" + depends on MAC80211_DEBUG_MENU + ---help--- + Selecting this option causes mac80211 to print out very + verbose power save mode debugging messages (when mac80211 + is an AP and has power saving stations.) + It should not be selected on production systems as those + messages are remotely triggerable. + + Do not select this option. + +config MAC80211_VERBOSE_MPL_DEBUG + bool "Verbose mesh peer link debugging" + depends on MAC80211_DEBUG_MENU + depends on MAC80211_MESH + ---help--- + Selecting this option causes mac80211 to print out very + verbose mesh peer link debugging messages (when mac80211 + is taking part in a mesh network). + It should not be selected on production systems as those + messages are remotely triggerable. + + Do not select this option. config MAC80211_LOWTX_FRAME_DUMP bool "Debug frame dumping" - depends on MAC80211_DEBUG + depends on MAC80211_DEBUG_MENU ---help--- Selecting this option will cause the stack to print a message for each frame that is handed @@ -137,33 +205,17 @@ config MAC80211_LOWTX_FRAME_DUMP If unsure, say N and insert the debugging code you require into the driver you are debugging. -config MAC80211_TKIP_DEBUG - bool "TKIP debugging" - depends on MAC80211_DEBUG - config MAC80211_DEBUG_COUNTERS bool "Extra statistics for TX/RX debugging" depends on MAC80211_DEBUG - -config MAC80211_IBSS_DEBUG - bool "Support for IBSS testing" - depends on MAC80211_DEBUG - ---help--- - Say Y here if you intend to debug the IBSS code. - -config MAC80211_VERBOSE_PS_DEBUG - bool "Verbose powersave mode debugging" - depends on MAC80211_DEBUG + depends on MAC80211_DEBUG_MENU + depends on MAC80211_DEBUGFS ---help--- - Say Y here to print out verbose powersave - mode debug messages. + Selecting this option causes mac80211 to keep additional + and very verbose statistics about TX and RX handler use + and show them in debugfs. -config MAC80211_VERBOSE_MPL_DEBUG - bool "Verbose mesh peer link debugging" - depends on MAC80211_DEBUG && MAC80211_MESH - ---help--- - Say Y here to print out verbose mesh peer link - debug messages. + If unsure, say N. config MAC80211_VERBOSE_SPECT_MGMT_DEBUG bool "Verbose Spectrum Management (IEEE 802.11h)debugging" diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 46705ae7a847..f90da1bbec49 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -954,7 +954,7 @@ int ieee80211_frame_duration(struct ieee80211_local *local, size_t len, void mac80211_ev_michael_mic_failure(struct net_device *dev, int keyidx, struct ieee80211_hdr *hdr); -#ifdef CONFIG_MAC80211_DEBUG +#ifdef CONFIG_MAC80211_NOINLINE #define debug_noinline noinline #else #define debug_noinline diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c index 984472702381..eeb16926aa7d 100644 --- a/net/mac80211/iface.c +++ b/net/mac80211/iface.c @@ -184,9 +184,9 @@ void ieee80211_if_set_type(struct net_device *dev, int type) sdata->u.mntr_flags = MONITOR_FLAG_CONTROL | MONITOR_FLAG_OTHER_BSS; break; - default: - printk(KERN_WARNING "%s: %s: Unknown interface type 0x%x", - dev->name, __func__, type); + case IEEE80211_IF_TYPE_INVALID: + BUG(); + break; } ieee80211_debugfs_change_if_type(sdata, oldtype); } diff --git a/net/mac80211/main.c b/net/mac80211/main.c index b661ee5bb824..f18cfd727872 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -151,9 +151,7 @@ static int ieee80211_change_mtu(struct net_device *dev, int new_mtu) /* FIX: what would be proper limits for MTU? * This interface uses 802.3 frames. */ if (new_mtu < 256 || - new_mtu > IEEE80211_MAX_DATA_LEN - 24 - 6 - meshhdrlen) { - printk(KERN_WARNING "%s: invalid MTU %d\n", - dev->name, new_mtu); + new_mtu > IEEE80211_MAX_DATA_LEN - 24 - 6 - meshhdrlen) { return -EINVAL; } @@ -589,7 +587,9 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) sta = sta_info_get(local, ra); if (!sta) { +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "Could not find the station\n"); +#endif ret = -ENOENT; goto exit; } @@ -617,9 +617,11 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) sta->ampdu_mlme.tid_tx[tid] = kmalloc(sizeof(struct tid_ampdu_tx), GFP_ATOMIC); if (!sta->ampdu_mlme.tid_tx[tid]) { +#ifdef CONFIG_MAC80211_HT_DEBUG if (net_ratelimit()) printk(KERN_ERR "allocate tx mlme to tid %d failed\n", tid); +#endif ret = -ENOMEM; goto err_unlock_sta; } @@ -689,7 +691,9 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) sta->ampdu_mlme.tid_tx[tid]->addba_resp_timer.expires = jiffies + ADDBA_RESP_INTERVAL; add_timer(&sta->ampdu_mlme.tid_tx[tid]->addba_resp_timer); +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "activated addBA response timer on tid %d\n", tid); +#endif goto exit; err_unlock_queue: @@ -771,8 +775,10 @@ void ieee80211_start_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u16 tid) DECLARE_MAC_BUF(mac); if (tid >= STA_TID_NUM) { +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "Bad TID value: tid = %d (>= %d)\n", tid, STA_TID_NUM); +#endif return; } @@ -780,8 +786,10 @@ void ieee80211_start_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u16 tid) sta = sta_info_get(local, ra); if (!sta) { rcu_read_unlock(); +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "Could not find station: %s\n", print_mac(mac, ra)); +#endif return; } @@ -789,8 +797,10 @@ void ieee80211_start_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u16 tid) spin_lock_bh(&sta->lock); if (!(*state & HT_ADDBA_REQUESTED_MSK)) { +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "addBA was not requested yet, state is %d\n", *state); +#endif spin_unlock_bh(&sta->lock); rcu_read_unlock(); return; @@ -801,7 +811,9 @@ void ieee80211_start_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u16 tid) *state |= HT_ADDBA_DRV_READY_MSK; if (*state == HT_AGG_STATE_OPERATIONAL) { +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "Aggregation is on for tid %d \n", tid); +#endif ieee80211_wake_queue(hw, sta->tid_to_tx_q[tid]); } spin_unlock_bh(&sta->lock); @@ -818,8 +830,10 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) DECLARE_MAC_BUF(mac); if (tid >= STA_TID_NUM) { +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "Bad TID value: tid = %d (>= %d)\n", tid, STA_TID_NUM); +#endif return; } @@ -831,8 +845,10 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) rcu_read_lock(); sta = sta_info_get(local, ra); if (!sta) { +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "Could not find station: %s\n", print_mac(mac, ra)); +#endif rcu_read_unlock(); return; } @@ -842,7 +858,9 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) * ieee80211_stop_tx_ba_session will let only * one stop call to pass through per sta/tid */ if ((*state & HT_AGG_STATE_REQ_STOP_BA_MSK) == 0) { +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "unexpected callback to A-MPDU stop\n"); +#endif rcu_read_unlock(); return; } @@ -884,9 +902,11 @@ void ieee80211_start_tx_ba_cb_irqsafe(struct ieee80211_hw *hw, struct sk_buff *skb = dev_alloc_skb(0); if (unlikely(!skb)) { +#ifdef CONFIG_MAC80211_HT_DEBUG if (net_ratelimit()) printk(KERN_WARNING "%s: Not enough memory, " "dropping start BA session", skb->dev->name); +#endif return; } ra_tid = (struct ieee80211_ra_tid *) &skb->cb; @@ -907,9 +927,11 @@ void ieee80211_stop_tx_ba_cb_irqsafe(struct ieee80211_hw *hw, struct sk_buff *skb = dev_alloc_skb(0); if (unlikely(!skb)) { +#ifdef CONFIG_MAC80211_HT_DEBUG if (net_ratelimit()) printk(KERN_WARNING "%s: Not enough memory, " "dropping stop BA session", skb->dev->name); +#endif return; } ra_tid = (struct ieee80211_ra_tid *) &skb->cb; @@ -1236,9 +1258,8 @@ static void ieee80211_tasklet_handler(unsigned long data) ra_tid->ra, ra_tid->tid); dev_kfree_skb(skb); break ; - default: /* should never get here! */ - printk(KERN_ERR "%s: Unknown message type (%d)\n", - wiphy_name(local->hw.wiphy), skb->pkt_type); + default: + WARN_ON(1); dev_kfree_skb(skb); break; } @@ -1365,12 +1386,14 @@ static void ieee80211_handle_filtered_frame(struct ieee80211_local *local, return; } +#ifdef CONFIG_MAC80211_VERBOSE_DEBUG if (net_ratelimit()) printk(KERN_DEBUG "%s: dropped TX filtered frame, " "queue_len=%d PS=%d @%lu\n", wiphy_name(local->hw.wiphy), skb_queue_len(&sta->tx_filtered), !!test_sta_flags(sta, WLAN_STA_PS), jiffies); +#endif dev_kfree_skb(skb); } diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 0a310d09ab07..4a3bddd206d8 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -346,7 +346,7 @@ static void ieee80211_sta_wmm_params(struct net_device *dev, params.cw_max = ecw2cw((pos[1] & 0xf0) >> 4); params.cw_min = ecw2cw(pos[1] & 0x0f); params.txop = pos[2] | (pos[3] << 8); -#ifdef CONFIG_MAC80211_DEBUG +#ifdef CONFIG_MAC80211_VERBOSE_DEBUG printk(KERN_DEBUG "%s: WMM queue=%d aci=%d acm=%d aifs=%d " "cWmin=%d cWmax=%d txop=%d\n", dev->name, queue, aci, acm, params.aifs, params.cw_min, @@ -371,6 +371,7 @@ static u32 ieee80211_handle_protect_preamb(struct ieee80211_sub_if_data *sdata, u32 changed = 0; if (use_protection != bss_conf->use_cts_prot) { +#ifdef CONFIG_MAC80211_VERBOSE_DEBUG if (net_ratelimit()) { printk(KERN_DEBUG "%s: CTS protection %s (BSSID=" "%s)\n", @@ -378,11 +379,13 @@ static u32 ieee80211_handle_protect_preamb(struct ieee80211_sub_if_data *sdata, use_protection ? "enabled" : "disabled", print_mac(mac, ifsta->bssid)); } +#endif bss_conf->use_cts_prot = use_protection; changed |= BSS_CHANGED_ERP_CTS_PROT; } if (use_short_preamble != bss_conf->use_short_preamble) { +#ifdef CONFIG_MAC80211_VERBOSE_DEBUG if (net_ratelimit()) { printk(KERN_DEBUG "%s: switched to %s barker preamble" " (BSSID=%s)\n", @@ -390,6 +393,7 @@ static u32 ieee80211_handle_protect_preamb(struct ieee80211_sub_if_data *sdata, use_short_preamble ? "short" : "long", print_mac(mac, ifsta->bssid)); } +#endif bss_conf->use_short_preamble = use_short_preamble; changed |= BSS_CHANGED_ERP_PREAMBLE; } @@ -1175,14 +1179,10 @@ static void ieee80211_auth_challenge(struct net_device *dev, u8 *pos; struct ieee802_11_elems elems; - printk(KERN_DEBUG "%s: replying to auth challenge\n", dev->name); pos = mgmt->u.auth.variable; ieee802_11_parse_elems(pos, len - (pos - (u8 *) mgmt), &elems); - if (!elems.challenge) { - printk(KERN_DEBUG "%s: no challenge IE in shared key auth " - "frame\n", dev->name); + if (!elems.challenge) return; - } ieee80211_send_auth(dev, ifsta, 3, elems.challenge - 2, elems.challenge_len + 2, 1); } @@ -1364,9 +1364,11 @@ static void ieee80211_sta_process_addba_request(struct net_device *dev, sta->ampdu_mlme.tid_rx[tid] = kmalloc(sizeof(struct tid_ampdu_rx), GFP_ATOMIC); if (!sta->ampdu_mlme.tid_rx[tid]) { +#ifdef CONFIG_MAC80211_HT_DEBUG if (net_ratelimit()) printk(KERN_ERR "allocate rx mlme to tid %d failed\n", tid); +#endif goto end; } /* rx timer */ @@ -1382,9 +1384,11 @@ static void ieee80211_sta_process_addba_request(struct net_device *dev, tid_agg_rx->reorder_buf = kmalloc(buf_size * sizeof(struct sk_buff *), GFP_ATOMIC); if (!tid_agg_rx->reorder_buf) { +#ifdef CONFIG_MAC80211_HT_DEBUG if (net_ratelimit()) printk(KERN_ERR "can not allocate reordering buffer " "to tid %d\n", tid); +#endif kfree(sta->ampdu_mlme.tid_rx[tid]); goto end; } @@ -1451,8 +1455,6 @@ static void ieee80211_sta_process_addba_resp(struct net_device *dev, if (!(*state & HT_ADDBA_REQUESTED_MSK)) { spin_unlock_bh(&sta->lock); - printk(KERN_DEBUG "state not HT_ADDBA_REQUESTED_MSK:" - "%d\n", *state); goto addba_resp_exit; } @@ -1471,22 +1473,14 @@ static void ieee80211_sta_process_addba_resp(struct net_device *dev, #endif /* CONFIG_MAC80211_HT_DEBUG */ if (le16_to_cpu(mgmt->u.action.u.addba_resp.status) == WLAN_STATUS_SUCCESS) { - if (*state & HT_ADDBA_RECEIVED_MSK) - printk(KERN_DEBUG "double addBA response\n"); - *state |= HT_ADDBA_RECEIVED_MSK; sta->ampdu_mlme.addba_req_num[tid] = 0; - if (*state == HT_AGG_STATE_OPERATIONAL) { - printk(KERN_DEBUG "Aggregation on for tid %d \n", tid); + if (*state == HT_AGG_STATE_OPERATIONAL) ieee80211_wake_queue(hw, sta->tid_to_tx_q[tid]); - } spin_unlock_bh(&sta->lock); - printk(KERN_DEBUG "recipient accepted agg: tid %d \n", tid); } else { - printk(KERN_DEBUG "recipient rejected agg: tid %d \n", tid); - sta->ampdu_mlme.addba_req_num[tid]++; /* this will allow the state check in stop_BA_session */ *state = HT_AGG_STATE_OPERATIONAL; @@ -1585,7 +1579,7 @@ void ieee80211_sta_stop_rx_ba_session(struct net_device *dev, u8 *ra, u16 tid, ra, tid, NULL); if (ret) printk(KERN_DEBUG "HW problem - can not stop rx " - "aggergation for tid %d\n", tid); + "aggregation for tid %d\n", tid); /* shutdown timer has not expired */ if (initiator != WLAN_BACK_TIMER) @@ -1691,12 +1685,16 @@ void sta_addba_resp_timer_expired(unsigned long data) if (!(*state & HT_ADDBA_REQUESTED_MSK)) { spin_unlock_bh(&sta->lock); *state = HT_AGG_STATE_IDLE; +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "timer expired on tid %d but we are not " "expecting addBA response there", tid); +#endif goto timer_expired_exit; } +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "addBA response timer expired on tid %d\n", tid); +#endif /* go through the state check in stop_BA_session */ *state = HT_AGG_STATE_OPERATIONAL; @@ -1724,7 +1722,9 @@ static void sta_rx_agg_session_timer_expired(unsigned long data) struct sta_info *sta = container_of(timer_to_id, struct sta_info, timer_to_tid[0]); +#ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "rx session timer expired on tid %d\n", (u16)*ptid); +#endif ieee80211_sta_stop_rx_ba_session(sta->sdata->dev, sta->addr, (u16)*ptid, WLAN_BACK_TIMER, WLAN_REASON_QSTA_TIMEOUT); @@ -1819,47 +1819,24 @@ static void ieee80211_rx_mgmt_auth(struct net_device *dev, DECLARE_MAC_BUF(mac); if (ifsta->state != IEEE80211_AUTHENTICATE && - sdata->vif.type != IEEE80211_IF_TYPE_IBSS) { - printk(KERN_DEBUG "%s: authentication frame received from " - "%s, but not in authenticate state - ignored\n", - dev->name, print_mac(mac, mgmt->sa)); + sdata->vif.type != IEEE80211_IF_TYPE_IBSS) return; - } - if (len < 24 + 6) { - printk(KERN_DEBUG "%s: too short (%zd) authentication frame " - "received from %s - ignored\n", - dev->name, len, print_mac(mac, mgmt->sa)); + if (len < 24 + 6) return; - } if (sdata->vif.type != IEEE80211_IF_TYPE_IBSS && - memcmp(ifsta->bssid, mgmt->sa, ETH_ALEN) != 0) { - printk(KERN_DEBUG "%s: authentication frame received from " - "unknown AP (SA=%s BSSID=%s) - " - "ignored\n", dev->name, print_mac(mac, mgmt->sa), - print_mac(mac, mgmt->bssid)); + memcmp(ifsta->bssid, mgmt->sa, ETH_ALEN) != 0) return; - } if (sdata->vif.type != IEEE80211_IF_TYPE_IBSS && - memcmp(ifsta->bssid, mgmt->bssid, ETH_ALEN) != 0) { - printk(KERN_DEBUG "%s: authentication frame received from " - "unknown BSSID (SA=%s BSSID=%s) - " - "ignored\n", dev->name, print_mac(mac, mgmt->sa), - print_mac(mac, mgmt->bssid)); + memcmp(ifsta->bssid, mgmt->bssid, ETH_ALEN) != 0) return; - } auth_alg = le16_to_cpu(mgmt->u.auth.auth_alg); auth_transaction = le16_to_cpu(mgmt->u.auth.auth_transaction); status_code = le16_to_cpu(mgmt->u.auth.status_code); - printk(KERN_DEBUG "%s: RX authentication from %s (alg=%d " - "transaction=%d status=%d)\n", - dev->name, print_mac(mac, mgmt->sa), auth_alg, - auth_transaction, status_code); - if (sdata->vif.type == IEEE80211_IF_TYPE_IBSS) { /* * IEEE 802.11 standard does not require authentication in IBSS @@ -1867,26 +1844,16 @@ static void ieee80211_rx_mgmt_auth(struct net_device *dev, * However, try to reply to authentication attempts if someone * has actually implemented this. */ - if (auth_alg != WLAN_AUTH_OPEN || auth_transaction != 1) { - printk(KERN_DEBUG "%s: unexpected IBSS authentication " - "frame (alg=%d transaction=%d)\n", - dev->name, auth_alg, auth_transaction); + if (auth_alg != WLAN_AUTH_OPEN || auth_transaction != 1) return; - } ieee80211_send_auth(dev, ifsta, 2, NULL, 0, 0); } if (auth_alg != ifsta->auth_alg || - auth_transaction != ifsta->auth_transaction) { - printk(KERN_DEBUG "%s: unexpected authentication frame " - "(alg=%d transaction=%d)\n", - dev->name, auth_alg, auth_transaction); + auth_transaction != ifsta->auth_transaction) return; - } if (status_code != WLAN_STATUS_SUCCESS) { - printk(KERN_DEBUG "%s: AP denied authentication (auth_alg=%d " - "code=%d)\n", dev->name, ifsta->auth_alg, status_code); if (status_code == WLAN_STATUS_NOT_SUPPORTED_AUTH_ALG) { u8 algs[3]; const int num_algs = ARRAY_SIZE(algs); @@ -1915,9 +1882,6 @@ static void ieee80211_rx_mgmt_auth(struct net_device *dev, !ieee80211_sta_wep_configured(dev)) continue; ifsta->auth_alg = algs[pos]; - printk(KERN_DEBUG "%s: set auth_alg=%d for " - "next try\n", - dev->name, ifsta->auth_alg); break; } } @@ -1947,27 +1911,14 @@ static void ieee80211_rx_mgmt_deauth(struct net_device *dev, u16 reason_code; DECLARE_MAC_BUF(mac); - if (len < 24 + 2) { - printk(KERN_DEBUG "%s: too short (%zd) deauthentication frame " - "received from %s - ignored\n", - dev->name, len, print_mac(mac, mgmt->sa)); + if (len < 24 + 2) return; - } - if (memcmp(ifsta->bssid, mgmt->sa, ETH_ALEN) != 0) { - printk(KERN_DEBUG "%s: deauthentication frame received from " - "unknown AP (SA=%s BSSID=%s) - " - "ignored\n", dev->name, print_mac(mac, mgmt->sa), - print_mac(mac, mgmt->bssid)); + if (memcmp(ifsta->bssid, mgmt->sa, ETH_ALEN)) return; - } reason_code = le16_to_cpu(mgmt->u.deauth.reason_code); - printk(KERN_DEBUG "%s: RX deauthentication from %s" - " (reason=%d)\n", - dev->name, print_mac(mac, mgmt->sa), reason_code); - if (ifsta->flags & IEEE80211_STA_AUTHENTICATED) printk(KERN_DEBUG "%s: deauthenticated\n", dev->name); @@ -1992,27 +1943,14 @@ static void ieee80211_rx_mgmt_disassoc(struct net_device *dev, u16 reason_code; DECLARE_MAC_BUF(mac); - if (len < 24 + 2) { - printk(KERN_DEBUG "%s: too short (%zd) disassociation frame " - "received from %s - ignored\n", - dev->name, len, print_mac(mac, mgmt->sa)); + if (len < 24 + 2) return; - } - if (memcmp(ifsta->bssid, mgmt->sa, ETH_ALEN) != 0) { - printk(KERN_DEBUG "%s: disassociation frame received from " - "unknown AP (SA=%s BSSID=%s) - " - "ignored\n", dev->name, print_mac(mac, mgmt->sa), - print_mac(mac, mgmt->bssid)); + if (memcmp(ifsta->bssid, mgmt->sa, ETH_ALEN)) return; - } reason_code = le16_to_cpu(mgmt->u.disassoc.reason_code); - printk(KERN_DEBUG "%s: RX disassociation from %s" - " (reason=%d)\n", - dev->name, print_mac(mac, mgmt->sa), reason_code); - if (ifsta->flags & IEEE80211_STA_ASSOCIATED) printk(KERN_DEBUG "%s: disassociated\n", dev->name); @@ -2048,27 +1986,14 @@ static void ieee80211_rx_mgmt_assoc_resp(struct ieee80211_sub_if_data *sdata, /* AssocResp and ReassocResp have identical structure, so process both * of them in this function. */ - if (ifsta->state != IEEE80211_ASSOCIATE) { - printk(KERN_DEBUG "%s: association frame received from " - "%s, but not in associate state - ignored\n", - dev->name, print_mac(mac, mgmt->sa)); + if (ifsta->state != IEEE80211_ASSOCIATE) return; - } - if (len < 24 + 6) { - printk(KERN_DEBUG "%s: too short (%zd) association frame " - "received from %s - ignored\n", - dev->name, len, print_mac(mac, mgmt->sa)); + if (len < 24 + 6) return; - } - if (memcmp(ifsta->bssid, mgmt->sa, ETH_ALEN) != 0) { - printk(KERN_DEBUG "%s: association frame received from " - "unknown AP (SA=%s BSSID=%s) - " - "ignored\n", dev->name, print_mac(mac, mgmt->sa), - print_mac(mac, mgmt->bssid)); + if (memcmp(ifsta->bssid, mgmt->sa, ETH_ALEN) != 0) return; - } capab_info = le16_to_cpu(mgmt->u.assoc_resp.capab_info); status_code = le16_to_cpu(mgmt->u.assoc_resp.status_code); @@ -2663,12 +2588,6 @@ static void ieee80211_rx_bss_info(struct net_device *dev, if (!beacon && memcmp(mgmt->da, dev->dev_addr, ETH_ALEN)) return; /* ignore ProbeResp to foreign address */ -#if 0 - printk(KERN_DEBUG "%s: RX %s from %s to %s\n", - dev->name, beacon ? "Beacon" : "Probe Response", - print_mac(mac, mgmt->sa), print_mac(mac2, mgmt->da)); -#endif - beacon_timestamp = le64_to_cpu(mgmt->u.beacon.timestamp); if (ieee80211_vif_is_mesh(&sdata->vif) && elems->mesh_id && @@ -2698,15 +2617,6 @@ static void ieee80211_rx_bss_info(struct net_device *dev, sta->supp_rates[rx_status->band] = sdata->u.sta.supp_rates_bits[rx_status->band]; } - if (sta->supp_rates[rx_status->band] != prev_rates) { - printk(KERN_DEBUG "%s: updated supp_rates set for " - "%s based on beacon info (0x%llx & 0x%llx -> " - "0x%llx)\n", - dev->name, print_mac(mac, sta->addr), - (unsigned long long) prev_rates, - (unsigned long long) supp_rates, - (unsigned long long) sta->supp_rates[rx_status->band]); - } } rcu_read_unlock(); @@ -2962,11 +2872,10 @@ static void ieee80211_rx_bss_info(struct net_device *dev, #endif /* CONFIG_MAC80211_IBSS_DEBUG */ if (beacon_timestamp > rx_timestamp) { #ifndef CONFIG_MAC80211_IBSS_DEBUG - if (net_ratelimit()) + printk(KERN_DEBUG "%s: beacon TSF higher than " + "local TSF - IBSS merge with BSSID %s\n", + dev->name, print_mac(mac, mgmt->bssid)); #endif - printk(KERN_DEBUG "%s: beacon TSF higher than " - "local TSF - IBSS merge with BSSID %s\n", - dev->name, print_mac(mac, mgmt->bssid)); ieee80211_sta_join_ibss(dev, &sdata->u.sta, bss); ieee80211_ibss_add_sta(dev, NULL, mgmt->bssid, mgmt->sa, @@ -3106,11 +3015,11 @@ static void ieee80211_rx_mgmt_probe_req(struct net_device *dev, pos = mgmt->u.probe_req.variable; if (pos[0] != WLAN_EID_SSID || pos + 2 + pos[1] > end) { - if (net_ratelimit()) { - printk(KERN_DEBUG "%s: Invalid SSID IE in ProbeReq " - "from %s\n", - dev->name, print_mac(mac, mgmt->sa)); - } +#ifdef CONFIG_MAC80211_IBSS_DEBUG + printk(KERN_DEBUG "%s: Invalid SSID IE in ProbeReq " + "from %s\n", + dev->name, print_mac(mac, mgmt->sa)); +#endif return; } if (pos[1] != 0 && @@ -3179,11 +3088,6 @@ static void ieee80211_rx_mgmt_action(struct net_device *dev, break; ieee80211_sta_process_delba(dev, mgmt, len); break; - default: - if (net_ratelimit()) - printk(KERN_DEBUG "%s: Rx unknown A-MPDU action\n", - dev->name); - break; } break; case PLINK_CATEGORY: @@ -3194,11 +3098,6 @@ static void ieee80211_rx_mgmt_action(struct net_device *dev, if (ieee80211_vif_is_mesh(&sdata->vif)) mesh_rx_path_sel_frame(dev, mgmt, len); break; - default: - if (net_ratelimit()) - printk(KERN_DEBUG "%s: Rx unknown action frame - " - "category=%d\n", dev->name, mgmt->u.action.category); - break; } } @@ -3234,11 +3133,6 @@ void ieee80211_sta_rx_mgmt(struct net_device *dev, struct sk_buff *skb, skb_queue_tail(&ifsta->skb_queue, skb); queue_work(local->hw.workqueue, &ifsta->work); return; - default: - printk(KERN_DEBUG "%s: received unknown management frame - " - "stype=%d\n", dev->name, - (fc & IEEE80211_FCTL_STYPE) >> 4); - break; } fail: @@ -3367,8 +3261,10 @@ static void ieee80211_sta_expire(struct net_device *dev, unsigned long exp_time) spin_lock_irqsave(&local->sta_lock, flags); list_for_each_entry_safe(sta, tmp, &local->sta_list, list) if (time_after(jiffies, sta->last_rx + exp_time)) { +#ifdef CONFIG_MAC80211_IBSS_DEBUG printk(KERN_DEBUG "%s: expiring inactive STA %s\n", dev->name, print_mac(mac, sta->addr)); +#endif __sta_info_unlink(&sta); if (sta) list_add(&sta->list, &tmp_list); @@ -3451,13 +3347,10 @@ void ieee80211_sta_work(struct work_struct *work) if (local->sta_sw_scanning || local->sta_hw_scanning) return; - if (sdata->vif.type != IEEE80211_IF_TYPE_STA && - sdata->vif.type != IEEE80211_IF_TYPE_IBSS && - sdata->vif.type != IEEE80211_IF_TYPE_MESH_POINT) { - printk(KERN_DEBUG "%s: ieee80211_sta_work: non-STA interface " - "(type=%d)\n", dev->name, sdata->vif.type); + if (WARN_ON(sdata->vif.type != IEEE80211_IF_TYPE_STA && + sdata->vif.type != IEEE80211_IF_TYPE_IBSS && + sdata->vif.type != IEEE80211_IF_TYPE_MESH_POINT)) return; - } ifsta = &sdata->u.sta; while ((skb = skb_dequeue(&ifsta->skb_queue))) @@ -3511,8 +3404,7 @@ void ieee80211_sta_work(struct work_struct *work) break; #endif default: - printk(KERN_DEBUG "ieee80211_sta_work: Unknown state %d\n", - ifsta->state); + WARN_ON(1); break; } @@ -3547,8 +3439,6 @@ static void ieee80211_sta_reset_auth(struct net_device *dev, ifsta->auth_alg = WLAN_AUTH_LEAP; else ifsta->auth_alg = WLAN_AUTH_OPEN; - printk(KERN_DEBUG "%s: Initial auth_alg=%d\n", dev->name, - ifsta->auth_alg); ifsta->auth_transaction = -1; ifsta->flags &= ~IEEE80211_STA_ASSOCIATED; ifsta->auth_tries = ifsta->assoc_tries = 0; @@ -4474,8 +4364,10 @@ struct sta_info *ieee80211_ibss_add_sta(struct net_device *dev, if (!ieee80211_bssid_match(bssid, sdata->u.sta.bssid)) return NULL; +#ifdef CONFIG_MAC80211_VERBOSE_DEBUG printk(KERN_DEBUG "%s: Adding new IBSS station %s (dev=%s)\n", wiphy_name(local->hw.wiphy), print_mac(mac, addr), dev->name); +#endif sta = sta_info_alloc(sdata, addr, GFP_ATOMIC); if (!sta) @@ -4502,7 +4394,7 @@ int ieee80211_sta_deauthenticate(struct net_device *dev, u16 reason) struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); struct ieee80211_if_sta *ifsta = &sdata->u.sta; - printk(KERN_DEBUG "%s: deauthenticate(reason=%d)\n", + printk(KERN_DEBUG "%s: deauthenticating by local choice (reason=%d)\n", dev->name, reason); if (sdata->vif.type != IEEE80211_IF_TYPE_STA && @@ -4520,7 +4412,7 @@ int ieee80211_sta_disassociate(struct net_device *dev, u16 reason) struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); struct ieee80211_if_sta *ifsta = &sdata->u.sta; - printk(KERN_DEBUG "%s: disassociate(reason=%d)\n", + printk(KERN_DEBUG "%s: disassociating by local choice (reason=%d)\n", dev->name, reason); if (sdata->vif.type != IEEE80211_IF_TYPE_STA) diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index 289112777e90..6a88e8f9bff0 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -613,11 +613,6 @@ ieee80211_rx_h_decrypt(struct ieee80211_rx_data *rx) rx->key->tx_rx_count++; /* TODO: add threshold stuff again */ } else { -#ifdef CONFIG_MAC80211_DEBUG - if (net_ratelimit()) - printk(KERN_DEBUG "%s: RX protected frame," - " but have no key\n", rx->dev->name); -#endif /* CONFIG_MAC80211_DEBUG */ return RX_DROP_MONITOR; } @@ -789,7 +784,7 @@ ieee80211_reassemble_add(struct ieee80211_sub_if_data *sdata, sdata->fragment_next = 0; if (!skb_queue_empty(&entry->skb_list)) { -#ifdef CONFIG_MAC80211_DEBUG +#ifdef CONFIG_MAC80211_VERBOSE_DEBUG struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) entry->skb_list.next->data; DECLARE_MAC_BUF(mac); @@ -801,7 +796,7 @@ ieee80211_reassemble_add(struct ieee80211_sub_if_data *sdata, jiffies - entry->first_frag_time, entry->seq, entry->last_frag, print_mac(mac, hdr->addr1), print_mac(mac2, hdr->addr2)); -#endif /* CONFIG_MAC80211_DEBUG */ +#endif __skb_queue_purge(&entry->skb_list); } @@ -922,18 +917,8 @@ ieee80211_rx_h_defragment(struct ieee80211_rx_data *rx) break; } rpn = rx->key->u.ccmp.rx_pn[rx->queue]; - if (memcmp(pn, rpn, CCMP_PN_LEN) != 0) { - if (net_ratelimit()) - printk(KERN_DEBUG "%s: defrag: CCMP PN not " - "sequential A2=%s" - " PN=%02x%02x%02x%02x%02x%02x " - "(expected %02x%02x%02x%02x%02x%02x)\n", - rx->dev->name, print_mac(mac, hdr->addr2), - rpn[0], rpn[1], rpn[2], rpn[3], rpn[4], - rpn[5], pn[0], pn[1], pn[2], pn[3], - pn[4], pn[5]); + if (memcmp(pn, rpn, CCMP_PN_LEN)) return RX_DROP_UNUSABLE; - } memcpy(entry->last_pn, pn, CCMP_PN_LEN); } @@ -1037,7 +1022,7 @@ ieee80211_rx_h_ps_poll(struct ieee80211_rx_data *rx) * have nothing buffered for it? */ printk(KERN_DEBUG "%s: STA %s sent PS Poll even " - "though there is no buffered frames for it\n", + "though there are no buffered frames for it\n", rx->dev->name, print_mac(mac, rx->sta->addr)); #endif /* CONFIG_MAC80211_VERBOSE_PS_DEBUG */ } @@ -1073,14 +1058,8 @@ static int ieee80211_802_1x_port_control(struct ieee80211_rx_data *rx) { if (unlikely(!rx->sta || - !test_sta_flags(rx->sta, WLAN_STA_AUTHORIZED))) { -#ifdef CONFIG_MAC80211_DEBUG - if (net_ratelimit()) - printk(KERN_DEBUG "%s: dropped frame " - "(unauthorized port)\n", rx->dev->name); -#endif /* CONFIG_MAC80211_DEBUG */ + !test_sta_flags(rx->sta, WLAN_STA_AUTHORIZED))) return -EACCES; - } return 0; } @@ -1160,16 +1139,8 @@ ieee80211_data_to_8023(struct ieee80211_rx_data *rx) memcpy(src, hdr->addr2, ETH_ALEN); if (unlikely(sdata->vif.type != IEEE80211_IF_TYPE_AP && - sdata->vif.type != IEEE80211_IF_TYPE_VLAN)) { - if (net_ratelimit()) - printk(KERN_DEBUG "%s: dropped ToDS frame " - "(BSSID=%s SA=%s DA=%s)\n", - dev->name, - print_mac(mac, hdr->addr1), - print_mac(mac2, hdr->addr2), - print_mac(mac3, hdr->addr3)); + sdata->vif.type != IEEE80211_IF_TYPE_VLAN)) return -1; - } break; case (IEEE80211_FCTL_TODS | IEEE80211_FCTL_FROMDS): /* RA TA DA SA */ @@ -1177,17 +1148,8 @@ ieee80211_data_to_8023(struct ieee80211_rx_data *rx) memcpy(src, hdr->addr4, ETH_ALEN); if (unlikely(sdata->vif.type != IEEE80211_IF_TYPE_WDS && - sdata->vif.type != IEEE80211_IF_TYPE_MESH_POINT)) { - if (net_ratelimit()) - printk(KERN_DEBUG "%s: dropped FromDS&ToDS " - "frame (RA=%s TA=%s DA=%s SA=%s)\n", - rx->dev->name, - print_mac(mac, hdr->addr1), - print_mac(mac2, hdr->addr2), - print_mac(mac3, hdr->addr3), - print_mac(mac4, hdr->addr4)); + sdata->vif.type != IEEE80211_IF_TYPE_MESH_POINT)) return -1; - } break; case IEEE80211_FCTL_FROMDS: /* DA BSSID SA */ @@ -1204,27 +1166,13 @@ ieee80211_data_to_8023(struct ieee80211_rx_data *rx) memcpy(dst, hdr->addr1, ETH_ALEN); memcpy(src, hdr->addr2, ETH_ALEN); - if (sdata->vif.type != IEEE80211_IF_TYPE_IBSS) { - if (net_ratelimit()) { - printk(KERN_DEBUG "%s: dropped IBSS frame " - "(DA=%s SA=%s BSSID=%s)\n", - dev->name, - print_mac(mac, hdr->addr1), - print_mac(mac2, hdr->addr2), - print_mac(mac3, hdr->addr3)); - } + if (sdata->vif.type != IEEE80211_IF_TYPE_IBSS) return -1; - } break; } - if (unlikely(skb->len - hdrlen < 8)) { - if (net_ratelimit()) { - printk(KERN_DEBUG "%s: RX too short data frame " - "payload\n", dev->name); - } + if (unlikely(skb->len - hdrlen < 8)) return -1; - } payload = skb->data + hdrlen; ethertype = (payload[6] << 8) | payload[7]; @@ -1416,10 +1364,8 @@ ieee80211_rx_h_amsdu(struct ieee80211_rx_data *rx) padding = ((4 - subframe_len) & 0x3); /* the last MSDU has no padding */ - if (subframe_len > remaining) { - printk(KERN_DEBUG "%s: wrong buffer size\n", dev->name); + if (subframe_len > remaining) return RX_DROP_UNUSABLE; - } skb_pull(skb, sizeof(struct ethhdr)); /* if last subframe reuse skb */ @@ -1440,8 +1386,6 @@ ieee80211_rx_h_amsdu(struct ieee80211_rx_data *rx) eth = (struct ethhdr *) skb_pull(skb, ntohs(len) + padding); if (!eth) { - printk(KERN_DEBUG "%s: wrong buffer size\n", - dev->name); dev_kfree_skb(frame); return RX_DROP_UNUSABLE; } @@ -1593,31 +1537,16 @@ static void ieee80211_rx_michael_mic_report(struct net_device *dev, else keyidx = -1; - if (net_ratelimit()) - printk(KERN_DEBUG "%s: TKIP hwaccel reported Michael MIC " - "failure from %s to %s keyidx=%d\n", - dev->name, print_mac(mac, hdr->addr2), - print_mac(mac2, hdr->addr1), keyidx); - if (!rx->sta) { /* * Some hardware seem to generate incorrect Michael MIC * reports; ignore them to avoid triggering countermeasures. */ - if (net_ratelimit()) - printk(KERN_DEBUG "%s: ignored spurious Michael MIC " - "error for unknown address %s\n", - dev->name, print_mac(mac, hdr->addr2)); goto ignore; } - if (!(rx->fc & IEEE80211_FCTL_PROTECTED)) { - if (net_ratelimit()) - printk(KERN_DEBUG "%s: ignored spurious Michael MIC " - "error for a frame with no PROTECTED flag (src " - "%s)\n", dev->name, print_mac(mac, hdr->addr2)); + if (!(rx->fc & IEEE80211_FCTL_PROTECTED)) goto ignore; - } if (rx->sdata->vif.type == IEEE80211_IF_TYPE_AP && keyidx) { /* @@ -1626,24 +1555,13 @@ static void ieee80211_rx_michael_mic_report(struct net_device *dev, * group keys and only the AP is sending real multicast * frames in the BSS. */ - if (net_ratelimit()) - printk(KERN_DEBUG "%s: ignored Michael MIC error for " - "a frame with non-zero keyidx (%d)" - " (src %s)\n", dev->name, keyidx, - print_mac(mac, hdr->addr2)); goto ignore; } if ((rx->fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_DATA && ((rx->fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_MGMT || - (rx->fc & IEEE80211_FCTL_STYPE) != IEEE80211_STYPE_AUTH)) { - if (net_ratelimit()) - printk(KERN_DEBUG "%s: ignored spurious Michael MIC " - "error for a frame that cannot be encrypted " - "(fc=0x%04x) (src %s)\n", - dev->name, rx->fc, print_mac(mac, hdr->addr2)); + (rx->fc & IEEE80211_FCTL_STYPE) != IEEE80211_STYPE_AUTH)) goto ignore; - } mac80211_ev_michael_mic_failure(rx->dev, keyidx, hdr); ignore: diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index b3c733162fc1..d8a16b7f6a6b 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -554,8 +554,10 @@ static void sta_info_cleanup_expire_buffered(struct ieee80211_local *local, sdata = sta->sdata; local->total_ps_buffered--; +#ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG printk(KERN_DEBUG "Buffered frame expired (STA " "%s)\n", print_mac(mac, sta->addr)); +#endif dev_kfree_skb(skb); if (skb_queue_empty(&sta->ps_tx_buf)) diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 7b930d3c2fba..9bd9faac3c3c 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -327,8 +327,10 @@ static void purge_old_ps_buffers(struct ieee80211_local *local) rcu_read_unlock(); local->total_ps_buffered = total; +#ifdef MAC80211_VERBOSE_PS_DEBUG printk(KERN_DEBUG "%s: PS buffers full - purged %d frames\n", wiphy_name(local->hw.wiphy), purged); +#endif } static ieee80211_tx_result @@ -358,11 +360,13 @@ ieee80211_tx_h_multicast_ps_buf(struct ieee80211_tx_data *tx) purge_old_ps_buffers(tx->local); if (skb_queue_len(&tx->sdata->bss->ps_bc_buf) >= AP_MAX_BC_BUFFER) { +#ifdef MAC80211_VERBOSE_PS_DEBUG if (net_ratelimit()) { printk(KERN_DEBUG "%s: BC TX buffer full - " "dropping the oldest frame\n", tx->dev->name); } +#endif dev_kfree_skb(skb_dequeue(&tx->sdata->bss->ps_bc_buf)); } else tx->local->total_ps_buffered++; @@ -403,11 +407,13 @@ ieee80211_tx_h_unicast_ps_buf(struct ieee80211_tx_data *tx) purge_old_ps_buffers(tx->local); if (skb_queue_len(&sta->ps_tx_buf) >= STA_MAX_TX_BUFFER) { struct sk_buff *old = skb_dequeue(&sta->ps_tx_buf); +#ifdef MAC80211_VERBOSE_PS_DEBUG if (net_ratelimit()) { printk(KERN_DEBUG "%s: STA %s TX " "buffer full - dropping oldest frame\n", tx->dev->name, print_mac(mac, sta->addr)); } +#endif dev_kfree_skb(old); } else tx->local->total_ps_buffered++; @@ -713,7 +719,6 @@ ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) return TX_CONTINUE; fail: - printk(KERN_DEBUG "%s: failed to fragment frame\n", tx->dev->name); if (frags) { for (i = 0; i < num_fragm - 1; i++) if (frags[i]) @@ -1404,8 +1409,6 @@ int ieee80211_subif_start_xmit(struct sk_buff *skb, sdata = IEEE80211_DEV_TO_SUB_IF(dev); if (unlikely(skb->len < ETH_HLEN)) { - printk(KERN_DEBUG "%s: short skb (len=%d)\n", - dev->name, skb->len); ret = 0; goto fail; } diff --git a/net/mac80211/wep.c b/net/mac80211/wep.c index 35b664d00e23..872d2fcd1a5b 100644 --- a/net/mac80211/wep.c +++ b/net/mac80211/wep.c @@ -253,11 +253,8 @@ int ieee80211_wep_decrypt(struct ieee80211_local *local, struct sk_buff *skb, if (ieee80211_wep_decrypt_data(local->wep_rx_tfm, rc4key, klen, skb->data + hdrlen + WEP_IV_LEN, - len)) { - if (net_ratelimit()) - printk(KERN_DEBUG "WEP decrypt failed (ICV)\n"); + len)) ret = -1; - } kfree(rc4key); @@ -301,14 +298,8 @@ ieee80211_crypto_wep_decrypt(struct ieee80211_rx_data *rx) return RX_CONTINUE; if (!(rx->status->flag & RX_FLAG_DECRYPTED)) { - if (ieee80211_wep_decrypt(rx->local, rx->skb, rx->key)) { -#ifdef CONFIG_MAC80211_DEBUG - if (net_ratelimit()) - printk(KERN_DEBUG "%s: RX WEP frame, decrypt " - "failed\n", rx->dev->name); -#endif /* CONFIG_MAC80211_DEBUG */ + if (ieee80211_wep_decrypt(rx->local, rx->skb, rx->key)) return RX_DROP_UNUSABLE; - } } else if (!(rx->status->flag & RX_FLAG_IV_STRIPPED)) { ieee80211_wep_remove_iv(rx->local, rx->skb, rx->key); /* remove ICV */ diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index cfa8fbb0736a..f23b5a4d4ac9 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -680,7 +680,6 @@ void ieee80211_requeue(struct ieee80211_local *local, int queue) if (!qdisc || !qdisc->dequeue) return; - printk(KERN_DEBUG "requeue: qlen = %d\n", qdisc->q.qlen); for (len = qdisc->q.qlen; len > 0; len--) { skb = qdisc->dequeue(qdisc); root_qd->q.qlen--; diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index f809761fbfb5..b414d5d92f38 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -146,9 +146,6 @@ ieee80211_rx_h_michael_mic_verify(struct ieee80211_rx_data *rx) if (!(rx->flags & IEEE80211_RX_RA_MATCH)) return RX_DROP_UNUSABLE; - printk(KERN_DEBUG "%s: invalid Michael MIC in data frame from " - "%s\n", rx->dev->name, print_mac(mac, sa)); - mac80211_ev_michael_mic_failure(rx->dev, rx->key->conf.keyidx, (void *) skb->data); return RX_DROP_UNUSABLE; @@ -282,15 +279,8 @@ ieee80211_crypto_tkip_decrypt(struct ieee80211_rx_data *rx) hdr->addr1, hwaccel, rx->queue, &rx->tkip_iv32, &rx->tkip_iv16); - if (res != TKIP_DECRYPT_OK || wpa_test) { -#ifdef CONFIG_MAC80211_DEBUG - if (net_ratelimit()) - printk(KERN_DEBUG "%s: TKIP decrypt failed for RX " - "frame from %s (res=%d)\n", rx->dev->name, - print_mac(mac, rx->sta->addr), res); -#endif /* CONFIG_MAC80211_DEBUG */ + if (res != TKIP_DECRYPT_OK || wpa_test) return RX_DROP_UNUSABLE; - } /* Trim ICV */ skb_trim(skb, skb->len - TKIP_ICV_LEN); @@ -512,16 +502,6 @@ ieee80211_crypto_ccmp_decrypt(struct ieee80211_rx_data *rx) (void) ccmp_hdr2pn(pn, skb->data + hdrlen); if (memcmp(pn, key->u.ccmp.rx_pn[rx->queue], CCMP_PN_LEN) <= 0) { -#ifdef CONFIG_MAC80211_DEBUG - u8 *ppn = key->u.ccmp.rx_pn[rx->queue]; - - printk(KERN_DEBUG "%s: CCMP replay detected for RX frame from " - "%s (RX PN %02x%02x%02x%02x%02x%02x <= prev. PN " - "%02x%02x%02x%02x%02x%02x)\n", rx->dev->name, - print_mac(mac, rx->sta->addr), - pn[0], pn[1], pn[2], pn[3], pn[4], pn[5], - ppn[0], ppn[1], ppn[2], ppn[3], ppn[4], ppn[5]); -#endif /* CONFIG_MAC80211_DEBUG */ key->u.ccmp.replays++; return RX_DROP_UNUSABLE; } @@ -541,12 +521,6 @@ ieee80211_crypto_ccmp_decrypt(struct ieee80211_rx_data *rx) skb->data + hdrlen + CCMP_HDR_LEN, data_len, skb->data + skb->len - CCMP_MIC_LEN, skb->data + hdrlen + CCMP_HDR_LEN)) { -#ifdef CONFIG_MAC80211_DEBUG - if (net_ratelimit()) - printk(KERN_DEBUG "%s: CCMP decrypt failed " - "for RX frame from %s\n", rx->dev->name, - print_mac(mac, rx->sta->addr)); -#endif /* CONFIG_MAC80211_DEBUG */ return RX_DROP_UNUSABLE; } } -- cgit v1.2.2 From 40b215e594b65a3488576c9d24b367548e18902a Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Thu, 3 Jul 2008 01:05:41 -0700 Subject: tcp: de-bloat a bit with factoring NET_INC_STATS_BH out There are some places in TCP that select one MIB index to bump snmp statistics like this: if () NET_INC_STATS_BH(); else if () NET_INC_STATS_BH(); ... else NET_INC_STATS_BH(); or in a more tricky but still similar way. On the other hand, this NET_INC_STATS_BH is a camouflaged increment of percpu variable, which is not that small. Factoring those cases out de-bloats 235 bytes on non-preemptible i386 config and drives parts of the code into 80 columns. add/remove: 0/0 grow/shrink: 0/7 up/down: 0/-235 (-235) function old new delta tcp_fastretrans_alert 1437 1424 -13 tcp_dsack_set 137 124 -13 tcp_xmit_retransmit_queue 690 676 -14 tcp_try_undo_recovery 283 265 -18 tcp_sacktag_write_queue 1550 1515 -35 tcp_update_reordering 162 106 -56 tcp_retransmit_timer 990 904 -86 Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/tcp_input.c | 46 ++++++++++++++++++++++++++++++++-------------- net/ipv4/tcp_output.c | 7 +++++-- net/ipv4/tcp_timer.c | 15 +++++++++------ 3 files changed, 46 insertions(+), 22 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index de30e70ff256..d6ea970a1513 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -947,17 +947,21 @@ static void tcp_update_reordering(struct sock *sk, const int metric, { struct tcp_sock *tp = tcp_sk(sk); if (metric > tp->reordering) { + int mib_idx; + tp->reordering = min(TCP_MAX_REORDERING, metric); /* This exciting event is worth to be remembered. 8) */ if (ts) - NET_INC_STATS_BH(LINUX_MIB_TCPTSREORDER); + mib_idx = LINUX_MIB_TCPTSREORDER; else if (tcp_is_reno(tp)) - NET_INC_STATS_BH(LINUX_MIB_TCPRENOREORDER); + mib_idx = LINUX_MIB_TCPRENOREORDER; else if (tcp_is_fack(tp)) - NET_INC_STATS_BH(LINUX_MIB_TCPFACKREORDER); + mib_idx = LINUX_MIB_TCPFACKREORDER; else - NET_INC_STATS_BH(LINUX_MIB_TCPSACKREORDER); + mib_idx = LINUX_MIB_TCPSACKREORDER; + + NET_INC_STATS_BH(mib_idx); #if FASTRETRANS_DEBUG > 1 printk(KERN_DEBUG "Disorder%d %d %u f%u s%u rr%d\n", tp->rx_opt.sack_ok, inet_csk(sk)->icsk_ca_state, @@ -1456,18 +1460,22 @@ tcp_sacktag_write_queue(struct sock *sk, struct sk_buff *ack_skb, if (!tcp_is_sackblock_valid(tp, dup_sack, sp[used_sacks].start_seq, sp[used_sacks].end_seq)) { + int mib_idx; + if (dup_sack) { if (!tp->undo_marker) - NET_INC_STATS_BH(LINUX_MIB_TCPDSACKIGNOREDNOUNDO); + mib_idx = LINUX_MIB_TCPDSACKIGNOREDNOUNDO; else - NET_INC_STATS_BH(LINUX_MIB_TCPDSACKIGNOREDOLD); + mib_idx = LINUX_MIB_TCPDSACKIGNOREDOLD; } else { /* Don't count olds caused by ACK reordering */ if ((TCP_SKB_CB(ack_skb)->ack_seq != tp->snd_una) && !after(sp[used_sacks].end_seq, tp->snd_una)) continue; - NET_INC_STATS_BH(LINUX_MIB_TCPSACKDISCARD); + mib_idx = LINUX_MIB_TCPSACKDISCARD; } + + NET_INC_STATS_BH(mib_idx); if (i == 0) first_sack_index = -1; continue; @@ -2380,15 +2388,19 @@ static int tcp_try_undo_recovery(struct sock *sk) struct tcp_sock *tp = tcp_sk(sk); if (tcp_may_undo(tp)) { + int mib_idx; + /* Happy end! We did not retransmit anything * or our original transmission succeeded. */ DBGUNDO(sk, inet_csk(sk)->icsk_ca_state == TCP_CA_Loss ? "loss" : "retrans"); tcp_undo_cwr(sk, 1); if (inet_csk(sk)->icsk_ca_state == TCP_CA_Loss) - NET_INC_STATS_BH(LINUX_MIB_TCPLOSSUNDO); + mib_idx = LINUX_MIB_TCPLOSSUNDO; else - NET_INC_STATS_BH(LINUX_MIB_TCPFULLUNDO); + mib_idx = LINUX_MIB_TCPFULLUNDO; + + NET_INC_STATS_BH(mib_idx); tp->undo_marker = 0; } if (tp->snd_una == tp->high_seq && tcp_is_reno(tp)) { @@ -2560,7 +2572,7 @@ static void tcp_fastretrans_alert(struct sock *sk, int pkts_acked, int flag) int is_dupack = !(flag & (FLAG_SND_UNA_ADVANCED | FLAG_NOT_DUP)); int do_lost = is_dupack || ((flag & FLAG_DATA_SACKED) && (tcp_fackets_out(tp) > tp->reordering)); - int fast_rexmit = 0; + int fast_rexmit = 0, mib_idx; if (WARN_ON(!tp->packets_out && tp->sacked_out)) tp->sacked_out = 0; @@ -2683,9 +2695,11 @@ static void tcp_fastretrans_alert(struct sock *sk, int pkts_acked, int flag) /* Otherwise enter Recovery state */ if (tcp_is_reno(tp)) - NET_INC_STATS_BH(LINUX_MIB_TCPRENORECOVERY); + mib_idx = LINUX_MIB_TCPRENORECOVERY; else - NET_INC_STATS_BH(LINUX_MIB_TCPSACKRECOVERY); + mib_idx = LINUX_MIB_TCPSACKRECOVERY; + + NET_INC_STATS_BH(mib_idx); tp->high_seq = tp->snd_nxt; tp->prior_ssthresh = 0; @@ -3700,10 +3714,14 @@ static inline int tcp_sack_extend(struct tcp_sack_block *sp, u32 seq, static void tcp_dsack_set(struct tcp_sock *tp, u32 seq, u32 end_seq) { if (tcp_is_sack(tp) && sysctl_tcp_dsack) { + int mib_idx; + if (before(seq, tp->rcv_nxt)) - NET_INC_STATS_BH(LINUX_MIB_TCPDSACKOLDSENT); + mib_idx = LINUX_MIB_TCPDSACKOLDSENT; else - NET_INC_STATS_BH(LINUX_MIB_TCPDSACKOFOSENT); + mib_idx = LINUX_MIB_TCPDSACKOFOSENT; + + NET_INC_STATS_BH(mib_idx); tp->rx_opt.dsack = 1; tp->duplicate_sack[0].start_seq = seq; diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 8f83ab432705..edef2afe905e 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1985,14 +1985,17 @@ void tcp_xmit_retransmit_queue(struct sock *sk) if (sacked & TCPCB_LOST) { if (!(sacked & (TCPCB_SACKED_ACKED|TCPCB_SACKED_RETRANS))) { + int mib_idx; + if (tcp_retransmit_skb(sk, skb)) { tp->retransmit_skb_hint = NULL; return; } if (icsk->icsk_ca_state != TCP_CA_Loss) - NET_INC_STATS_BH(LINUX_MIB_TCPFASTRETRANS); + mib_idx = LINUX_MIB_TCPFASTRETRANS; else - NET_INC_STATS_BH(LINUX_MIB_TCPSLOWSTARTRETRANS); + mib_idx = LINUX_MIB_TCPSLOWSTARTRETRANS; + NET_INC_STATS_BH(mib_idx); if (skb == tcp_write_queue_head(sk)) inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c index 3e358cbb1247..6a480d1fd8f6 100644 --- a/net/ipv4/tcp_timer.c +++ b/net/ipv4/tcp_timer.c @@ -326,24 +326,27 @@ static void tcp_retransmit_timer(struct sock *sk) goto out; if (icsk->icsk_retransmits == 0) { + int mib_idx; + if (icsk->icsk_ca_state == TCP_CA_Disorder || icsk->icsk_ca_state == TCP_CA_Recovery) { if (tcp_is_sack(tp)) { if (icsk->icsk_ca_state == TCP_CA_Recovery) - NET_INC_STATS_BH(LINUX_MIB_TCPSACKRECOVERYFAIL); + mib_idx = LINUX_MIB_TCPSACKRECOVERYFAIL; else - NET_INC_STATS_BH(LINUX_MIB_TCPSACKFAILURES); + mib_idx = LINUX_MIB_TCPSACKFAILURES; } else { if (icsk->icsk_ca_state == TCP_CA_Recovery) - NET_INC_STATS_BH(LINUX_MIB_TCPRENORECOVERYFAIL); + mib_idx = LINUX_MIB_TCPRENORECOVERYFAIL; else - NET_INC_STATS_BH(LINUX_MIB_TCPRENOFAILURES); + mib_idx = LINUX_MIB_TCPRENOFAILURES; } } else if (icsk->icsk_ca_state == TCP_CA_Loss) { - NET_INC_STATS_BH(LINUX_MIB_TCPLOSSFAILURES); + mib_idx = LINUX_MIB_TCPLOSSFAILURES; } else { - NET_INC_STATS_BH(LINUX_MIB_TCPTIMEOUTS); + mib_idx = LINUX_MIB_TCPTIMEOUTS; } + NET_INC_STATS_BH(mib_idx); } if (tcp_use_frto(sk)) { -- cgit v1.2.2 From d68b82705a4a754e5773f412c6b8f1e65259bc8b Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Wed, 25 Jun 2008 16:26:47 +0900 Subject: ipv6: Do not assign non-valid address on interface. Check the type of the address when adding a new one on interface. - the unspecified address (::) is always disallowed (RFC4291 2.5.2) - the loopback address is disallowed unless the interface is (one of) loopback (RFC4291 2.5.3). - multicast addresses are disallowed. Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/addrconf.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'net') diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 84127d854cfc..8b6875f02039 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -578,6 +578,13 @@ ipv6_add_addr(struct inet6_dev *idev, const struct in6_addr *addr, int pfxlen, struct rt6_info *rt; int hash; int err = 0; + int addr_type = ipv6_addr_type(addr); + + if (addr_type == IPV6_ADDR_ANY || + addr_type & IPV6_ADDR_MULTICAST || + (!(idev->dev->flags & IFF_LOOPBACK) && + addr_type & IPV6_ADDR_LOOPBACK)) + return ERR_PTR(-EADDRNOTAVAIL); rcu_read_lock_bh(); if (idev->dead) { -- cgit v1.2.2 From f81b2e7d8cf8c6a52b7a5224c3b89cee5aeb6811 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Wed, 25 Jun 2008 16:55:26 +0900 Subject: ipv6: Do not forward packets with the unspecified source address. RFC4291 2.5.2. Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/ip6_output.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index fd7cd1bfe151..871bdec09edb 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -498,7 +498,8 @@ int ip6_forward(struct sk_buff *skb) int addrtype = ipv6_addr_type(&hdr->saddr); /* This check is security critical. */ - if (addrtype & (IPV6_ADDR_MULTICAST|IPV6_ADDR_LOOPBACK)) + if (addrtype == IPV6_ADDR_ANY || + addrtype & (IPV6_ADDR_MULTICAST | IPV6_ADDR_LOOPBACK)) goto error; if (addrtype & IPV6_ADDR_LINKLOCAL) { icmpv6_send(skb, ICMPV6_DEST_UNREACH, -- cgit v1.2.2 From 5ce83afaac956238c3c25f60a899c511e9d8cbf4 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Wed, 25 Jun 2008 16:58:17 +0900 Subject: ipv6: Assume the loopback address in link-local scope. Handle interface property strictly when looking up a route for the loopback address (RFC4291 2.5.3). Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/route.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 751e98f9b8b4..dbad96c58baa 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -228,7 +228,7 @@ static __inline__ int rt6_check_expired(const struct rt6_info *rt) static inline int rt6_need_strict(struct in6_addr *daddr) { return (ipv6_addr_type(daddr) & - (IPV6_ADDR_MULTICAST | IPV6_ADDR_LINKLOCAL)); + (IPV6_ADDR_MULTICAST | IPV6_ADDR_LINKLOCAL | IPV6_ADDR_LOOPBACK)); } /* -- cgit v1.2.2 From 778d80be52699596bf70e0eb0761cf5e1e46088d Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Sat, 28 Jun 2008 14:17:11 +0900 Subject: ipv6: Add disable_ipv6 sysctl to disable IPv6 operaion on specific interface. Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/addrconf.c | 11 +++++++++++ net/ipv6/ip6_input.c | 3 ++- net/ipv6/ip6_output.c | 7 +++++++ 3 files changed, 20 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 8b6875f02039..8c5cff50bbed 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -183,6 +183,7 @@ struct ipv6_devconf ipv6_devconf __read_mostly = { #endif .proxy_ndp = 0, .accept_source_route = 0, /* we do not accept RH0 by default. */ + .disable_ipv6 = 0, }; static struct ipv6_devconf ipv6_devconf_dflt __read_mostly = { @@ -215,6 +216,7 @@ static struct ipv6_devconf ipv6_devconf_dflt __read_mostly = { #endif .proxy_ndp = 0, .accept_source_route = 0, /* we do not accept RH0 by default. */ + .disable_ipv6 = 0, }; /* IPv6 Wildcard Address and Loopback Address defined by RFC2553 */ @@ -3657,6 +3659,7 @@ static inline void ipv6_store_devconf(struct ipv6_devconf *cnf, #ifdef CONFIG_IPV6_MROUTE array[DEVCONF_MC_FORWARDING] = cnf->mc_forwarding; #endif + array[DEVCONF_DISABLE_IPV6] = cnf->disable_ipv6; } static inline size_t inet6_if_nlmsg_size(void) @@ -4215,6 +4218,14 @@ static struct addrconf_sysctl_table .proc_handler = &proc_dointvec, }, #endif + { + .ctl_name = CTL_UNNUMBERED, + .procname = "disable_ipv6", + .data = &ipv6_devconf.disable_ipv6, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec, + }, { .ctl_name = 0, /* sentinel */ } diff --git a/net/ipv6/ip6_input.c b/net/ipv6/ip6_input.c index 34e5a96623ae..ea81c614dde2 100644 --- a/net/ipv6/ip6_input.c +++ b/net/ipv6/ip6_input.c @@ -71,7 +71,8 @@ int ipv6_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt IP6_INC_STATS_BH(idev, IPSTATS_MIB_INRECEIVES); - if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL) { + if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL || + !idev || unlikely(idev->cnf.disable_ipv6)) { IP6_INC_STATS_BH(idev, IPSTATS_MIB_INDISCARDS); rcu_read_unlock(); goto out; diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 871bdec09edb..0981c1ef3057 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -173,6 +173,13 @@ static inline int ip6_skb_dst_mtu(struct sk_buff *skb) int ip6_output(struct sk_buff *skb) { + struct inet6_dev *idev = ip6_dst_idev(skb->dst); + if (unlikely(idev->cnf.disable_ipv6)) { + IP6_INC_STATS(idev, IPSTATS_MIB_OUTDISCARDS); + kfree_skb(skb); + return 0; + } + if ((skb->len > ip6_skb_dst_mtu(skb) && !skb_is_gso(skb)) || dst_allfrag(skb->dst)) return ip6_fragment(skb, ip6_output2); -- cgit v1.2.2 From 1b34be74cbf18f5d58cc85c7c4afcd9f7d74accd Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Sat, 28 Jun 2008 14:18:38 +0900 Subject: ipv6 addrconf: add accept_dad sysctl to control DAD operation. - If 0, disable DAD. - If 1, perform DAD (default). - If >1, perform DAD and disable IPv6 operation if DAD for MAC-based link-local address has been failed (RFC4862 5.4.5). We do not follow RFC4862 by default. Refer to the netdev thread entitled "Linux IPv6 DAD not full conform to RFC 4862 ?" http://www.spinics.net/lists/netdev/msg52027.html Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/addrconf.c | 35 +++++++++++++++++++++++++++++++++++ 1 file changed, 35 insertions(+) (limited to 'net') diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 8c5cff50bbed..2ec73e62202c 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -119,6 +119,7 @@ static void ipv6_regen_rndid(unsigned long data); static int desync_factor = MAX_DESYNC_FACTOR * HZ; #endif +static int ipv6_generate_eui64(u8 *eui, struct net_device *dev); static int ipv6_count_addresses(struct inet6_dev *idev); /* @@ -184,6 +185,7 @@ struct ipv6_devconf ipv6_devconf __read_mostly = { .proxy_ndp = 0, .accept_source_route = 0, /* we do not accept RH0 by default. */ .disable_ipv6 = 0, + .accept_dad = 1, }; static struct ipv6_devconf ipv6_devconf_dflt __read_mostly = { @@ -217,6 +219,7 @@ static struct ipv6_devconf ipv6_devconf_dflt __read_mostly = { .proxy_ndp = 0, .accept_source_route = 0, /* we do not accept RH0 by default. */ .disable_ipv6 = 0, + .accept_dad = 1, }; /* IPv6 Wildcard Address and Loopback Address defined by RFC2553 */ @@ -380,6 +383,9 @@ static struct inet6_dev * ipv6_add_dev(struct net_device *dev) */ in6_dev_hold(ndev); + if (dev->flags & (IFF_NOARP | IFF_LOOPBACK)) + ndev->cnf.accept_dad = -1; + #if defined(CONFIG_IPV6_SIT) || defined(CONFIG_IPV6_SIT_MODULE) if (dev->type == ARPHRD_SIT && (dev->priv_flags & IFF_ISATAP)) { printk(KERN_INFO @@ -1421,6 +1427,20 @@ static void addrconf_dad_stop(struct inet6_ifaddr *ifp) void addrconf_dad_failure(struct inet6_ifaddr *ifp) { + struct inet6_dev *idev = ifp->idev; + if (idev->cnf.accept_dad > 1 && !idev->cnf.disable_ipv6) { + struct in6_addr addr; + + addr.s6_addr32[0] = htonl(0xfe800000); + addr.s6_addr32[1] = 0; + + if (!ipv6_generate_eui64(addr.s6_addr + 8, idev->dev) && + ipv6_addr_equal(&ifp->addr, &addr)) { + /* DAD failed for link-local based on MAC address */ + idev->cnf.disable_ipv6 = 1; + } + } + if (net_ratelimit()) printk(KERN_INFO "%s: duplicate address detected!\n", ifp->idev->dev->name); addrconf_dad_stop(ifp); @@ -2753,6 +2773,7 @@ static void addrconf_dad_start(struct inet6_ifaddr *ifp, u32 flags) spin_lock_bh(&ifp->lock); if (dev->flags&(IFF_NOARP|IFF_LOOPBACK) || + idev->cnf.accept_dad < 1 || !(ifp->flags&IFA_F_TENTATIVE) || ifp->flags & IFA_F_NODAD) { ifp->flags &= ~(IFA_F_TENTATIVE|IFA_F_OPTIMISTIC); @@ -2800,6 +2821,11 @@ static void addrconf_dad_timer(unsigned long data) read_unlock_bh(&idev->lock); goto out; } + if (idev->cnf.accept_dad > 1 && idev->cnf.disable_ipv6) { + read_unlock_bh(&idev->lock); + addrconf_dad_failure(ifp); + return; + } spin_lock_bh(&ifp->lock); if (ifp->probes == 0) { /* @@ -3660,6 +3686,7 @@ static inline void ipv6_store_devconf(struct ipv6_devconf *cnf, array[DEVCONF_MC_FORWARDING] = cnf->mc_forwarding; #endif array[DEVCONF_DISABLE_IPV6] = cnf->disable_ipv6; + array[DEVCONF_ACCEPT_DAD] = cnf->accept_dad; } static inline size_t inet6_if_nlmsg_size(void) @@ -4226,6 +4253,14 @@ static struct addrconf_sysctl_table .mode = 0644, .proc_handler = &proc_dointvec, }, + { + .ctl_name = CTL_UNNUMBERED, + .procname = "accept_dad", + .data = &ipv6_devconf.accept_dad, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec, + }, { .ctl_name = 0, /* sentinel */ } -- cgit v1.2.2 From dd3abc4ef52597ec8268274222574b2700ba3ded Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Wed, 2 Jul 2008 18:30:18 +0900 Subject: ipv6 route: Prefer outgoing interface with source address assigned. Outgoing interface is selected by the route decision if unspecified. Let's prefer routes via interface(s) with the address assigned if we have multiple routes with same cost. With help from Naohiro Ooiwa . Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/route.c | 20 ++++++++++++++++---- 1 file changed, 16 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/ipv6/route.c b/net/ipv6/route.c index dbad96c58baa..5d6c166dfbb6 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -237,15 +237,20 @@ static inline int rt6_need_strict(struct in6_addr *daddr) static inline struct rt6_info *rt6_device_match(struct net *net, struct rt6_info *rt, + struct in6_addr *saddr, int oif, int flags) { struct rt6_info *local = NULL; struct rt6_info *sprt; - if (oif) { - for (sprt = rt; sprt; sprt = sprt->u.dst.rt6_next) { - struct net_device *dev = sprt->rt6i_dev; + if (!oif && ipv6_addr_any(saddr)) + goto out; + + for (sprt = rt; sprt; sprt = sprt->u.dst.rt6_next) { + struct net_device *dev = sprt->rt6i_dev; + + if (oif) { if (dev->ifindex == oif) return sprt; if (dev->flags & IFF_LOOPBACK) { @@ -259,14 +264,21 @@ static inline struct rt6_info *rt6_device_match(struct net *net, } local = sprt; } + } else { + if (ipv6_chk_addr(net, saddr, dev, + flags & RT6_LOOKUP_F_IFACE)) + return sprt; } + } + if (oif) { if (local) return local; if (flags & RT6_LOOKUP_F_IFACE) return net->ipv6.ip6_null_entry; } +out: return rt; } @@ -539,7 +551,7 @@ static struct rt6_info *ip6_pol_route_lookup(struct net *net, fn = fib6_lookup(&table->tb6_root, &fl->fl6_dst, &fl->fl6_src); restart: rt = fn->leaf; - rt = rt6_device_match(net, rt, fl->oif, flags); + rt = rt6_device_match(net, rt, &fl->fl6_src, fl->oif, flags); BACKTRACK(net, &fl->fl6_src); out: dst_use(&rt->u.dst, jiffies); -- cgit v1.2.2 From 623d1a1af77bd52a389c6eda5920e28eb2ee468b Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Thu, 3 Jul 2008 12:13:30 +0800 Subject: ipv6: Do cleanup for ip6_mr_init. If do not do it, we will get following issues: 1. Leaving junks after inet6_init failing halfway. 2. Leaving proc and notifier junks after ipv6 modules unloading. Signed-off-by: Wang Chen Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/af_inet6.c | 11 ++++++++++- net/ipv6/ip6mr.c | 38 +++++++++++++++++++++++++++++++++----- 2 files changed, 43 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c index 3ce8d2f318c6..6b39af1acb5e 100644 --- a/net/ipv6/af_inet6.c +++ b/net/ipv6/af_inet6.c @@ -953,7 +953,9 @@ static int __init inet6_init(void) if (err) goto icmp_fail; #ifdef CONFIG_IPV6_MROUTE - ip6_mr_init(); + err = ip6_mr_init(); + if (err) + goto ipmr_fail; #endif err = ndisc_init(); if (err) @@ -1057,6 +1059,10 @@ netfilter_fail: igmp_fail: ndisc_cleanup(); ndisc_fail: +#ifdef CONFIG_IPV6_MROUTE + ip6_mr_cleanup(); +ipmr_fail: +#endif icmpv6_cleanup(); icmp_fail: unregister_pernet_subsys(&inet6_net_ops); @@ -1111,6 +1117,9 @@ static void __exit inet6_exit(void) ipv6_netfilter_fini(); igmp6_cleanup(); ndisc_cleanup(); +#ifdef CONFIG_IPV6_MROUTE + ip6_mr_cleanup(); +#endif icmpv6_cleanup(); rawv6_exit(); diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c index 90e763073dc5..cfac26d674ed 100644 --- a/net/ipv6/ip6mr.c +++ b/net/ipv6/ip6mr.c @@ -948,23 +948,51 @@ static struct notifier_block ip6_mr_notifier = { * Setup for IP multicast routing */ -void __init ip6_mr_init(void) +int __init ip6_mr_init(void) { + int err; + mrt_cachep = kmem_cache_create("ip6_mrt_cache", sizeof(struct mfc6_cache), 0, SLAB_HWCACHE_ALIGN, NULL); if (!mrt_cachep) - panic("cannot allocate ip6_mrt_cache"); + return -ENOMEM; setup_timer(&ipmr_expire_timer, ipmr_expire_process, 0); - register_netdevice_notifier(&ip6_mr_notifier); + err = register_netdevice_notifier(&ip6_mr_notifier); + if (err) + goto reg_notif_fail; +#ifdef CONFIG_PROC_FS + err = -ENOMEM; + if (!proc_net_fops_create(&init_net, "ip6_mr_vif", 0, &ip6mr_vif_fops)) + goto proc_vif_fail; + if (!proc_net_fops_create(&init_net, "ip6_mr_cache", + 0, &ip6mr_mfc_fops)) + goto proc_cache_fail; +#endif + return 0; +reg_notif_fail: + kmem_cache_destroy(mrt_cachep); #ifdef CONFIG_PROC_FS - proc_net_fops_create(&init_net, "ip6_mr_vif", 0, &ip6mr_vif_fops); - proc_net_fops_create(&init_net, "ip6_mr_cache", 0, &ip6mr_mfc_fops); +proc_vif_fail: + unregister_netdevice_notifier(&ip6_mr_notifier); +proc_cache_fail: + proc_net_remove(&init_net, "ip6_mr_vif"); #endif + return err; } +void ip6_mr_cleanup(void) +{ +#ifdef CONFIG_PROC_FS + proc_net_remove(&init_net, "ip6_mr_cache"); + proc_net_remove(&init_net, "ip6_mr_vif"); +#endif + unregister_netdevice_notifier(&ip6_mr_notifier); + del_timer(&ipmr_expire_timer); + kmem_cache_destroy(mrt_cachep); +} static int ip6mr_mfc_add(struct mf6cctl *mfc, int mrtsock) { -- cgit v1.2.2 From 03d2f897e9fb3218989baa2139a951ce7f5414bf Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Thu, 3 Jul 2008 12:13:36 +0800 Subject: ipv4: Do cleanup for ip_mr_init Same as ip6_mr_init(), make ip_mr_init() return errno if fails. But do not do error handling in inet_init(), just print a msg. Signed-off-by: Wang Chen Signed-off-by: YOSHIFUJI Hideaki --- net/ipv4/af_inet.c | 5 +++-- net/ipv4/ipmr.c | 28 ++++++++++++++++++++++++---- 2 files changed, 27 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 42bd24b64b57..dc411335c14f 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -1479,14 +1479,15 @@ static int __init inet_init(void) * Initialise the multicast router */ #if defined(CONFIG_IP_MROUTE) - ip_mr_init(); + if (ip_mr_init()) + printk(KERN_CRIT "inet_init: Cannot init ipv4 mroute\n"); #endif /* * Initialise per-cpu ipv4 mibs */ if (init_ipv4_mibs()) - printk(KERN_CRIT "inet_init: Cannot init ipv4 mibs\n"); ; + printk(KERN_CRIT "inet_init: Cannot init ipv4 mibs\n"); ipv4_proc_init(); diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index 300ab0c2919e..438fab9c62a0 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c @@ -1878,16 +1878,36 @@ static struct net_protocol pim_protocol = { * Setup for IP multicast routing */ -void __init ip_mr_init(void) +int __init ip_mr_init(void) { + int err; + mrt_cachep = kmem_cache_create("ip_mrt_cache", sizeof(struct mfc_cache), 0, SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL); + if (!mrt_cachep) + return -ENOMEM; + setup_timer(&ipmr_expire_timer, ipmr_expire_process, 0); - register_netdevice_notifier(&ip_mr_notifier); + err = register_netdevice_notifier(&ip_mr_notifier); + if (err) + goto reg_notif_fail; #ifdef CONFIG_PROC_FS - proc_net_fops_create(&init_net, "ip_mr_vif", 0, &ipmr_vif_fops); - proc_net_fops_create(&init_net, "ip_mr_cache", 0, &ipmr_mfc_fops); + err = -ENOMEM; + if (!proc_net_fops_create(&init_net, "ip_mr_vif", 0, &ipmr_vif_fops)) + goto proc_vif_fail; + if (!proc_net_fops_create(&init_net, "ip_mr_cache", 0, &ipmr_mfc_fops)) + goto proc_cache_fail; #endif + return 0; +reg_notif_fail: + kmem_cache_destroy(mrt_cachep); +#ifdef CONFIG_PROC_FS +proc_vif_fail: + unregister_netdevice_notifier(&ip_mr_notifier); +proc_cache_fail: + proc_net_remove(&init_net, "ip_mr_vif"); +#endif + return err; } -- cgit v1.2.2 From e0835f8fa56d2d308486f8a34cf1c4480cd27f4e Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Thu, 3 Jul 2008 16:51:22 +0900 Subject: ipv4,ipv6 mroute: Add some helper inline functions to remove ugly ifdefs. ip{,v6}_mroute_{set,get}sockopt() should not matter by optimization but it would be better not to depend on optimization semantically. Signed-off-by: YOSHIFUJI Hideaki --- net/ipv6/af_inet6.c | 8 -------- 1 file changed, 8 deletions(-) (limited to 'net') diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c index 6b39af1acb5e..3d828bc4b1cf 100644 --- a/net/ipv6/af_inet6.c +++ b/net/ipv6/af_inet6.c @@ -59,9 +59,7 @@ #include #include -#ifdef CONFIG_IPV6_MROUTE #include -#endif MODULE_AUTHOR("Cast of dozens"); MODULE_DESCRIPTION("IPv6 protocol stack for Linux"); @@ -952,11 +950,9 @@ static int __init inet6_init(void) err = icmpv6_init(); if (err) goto icmp_fail; -#ifdef CONFIG_IPV6_MROUTE err = ip6_mr_init(); if (err) goto ipmr_fail; -#endif err = ndisc_init(); if (err) goto ndisc_fail; @@ -1059,10 +1055,8 @@ netfilter_fail: igmp_fail: ndisc_cleanup(); ndisc_fail: -#ifdef CONFIG_IPV6_MROUTE ip6_mr_cleanup(); ipmr_fail: -#endif icmpv6_cleanup(); icmp_fail: unregister_pernet_subsys(&inet6_net_ops); @@ -1117,9 +1111,7 @@ static void __exit inet6_exit(void) ipv6_netfilter_fini(); igmp6_cleanup(); ndisc_cleanup(); -#ifdef CONFIG_IPV6_MROUTE ip6_mr_cleanup(); -#endif icmpv6_cleanup(); rawv6_exit(); -- cgit v1.2.2 From 76e6ebfb40a2455c18234dcb0f9df37533215461 Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Sat, 5 Jul 2008 19:00:44 -0700 Subject: netns: add namespace parameter to rt_cache_flush Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/arp.c | 2 +- net/ipv4/devinet.c | 8 +++++--- net/ipv4/fib_frontend.c | 17 +++++++++-------- net/ipv4/fib_hash.c | 6 +++--- net/ipv4/fib_rules.c | 2 +- net/ipv4/fib_trie.c | 6 +++--- net/ipv4/route.c | 8 ++++---- 7 files changed, 26 insertions(+), 23 deletions(-) (limited to 'net') diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c index 20c515a1be28..29df75a6bcc7 100644 --- a/net/ipv4/arp.c +++ b/net/ipv4/arp.c @@ -1197,7 +1197,7 @@ static int arp_netdev_event(struct notifier_block *this, unsigned long event, vo switch (event) { case NETDEV_CHANGEADDR: neigh_changeaddr(&arp_tbl, dev); - rt_cache_flush(0); + rt_cache_flush(dev_net(dev), 0); break; default: break; diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c index 9de2514946ca..2e667e2f90df 100644 --- a/net/ipv4/devinet.c +++ b/net/ipv4/devinet.c @@ -1348,7 +1348,7 @@ static int devinet_sysctl_forward(ctl_table *ctl, int write, dev_disable_lro(idev->dev); } rtnl_unlock(); - rt_cache_flush(0); + rt_cache_flush(net, 0); } } @@ -1362,9 +1362,10 @@ int ipv4_doint_and_flush(ctl_table *ctl, int write, int *valp = ctl->data; int val = *valp; int ret = proc_dointvec(ctl, write, filp, buffer, lenp, ppos); + struct net *net = ctl->extra2; if (write && *valp != val) - rt_cache_flush(0); + rt_cache_flush(net, 0); return ret; } @@ -1375,9 +1376,10 @@ int ipv4_doint_and_flush_strategy(ctl_table *table, int __user *name, int nlen, { int ret = devinet_conf_sysctl(table, name, nlen, oldval, oldlenp, newval, newlen); + struct net *net = table->extra2; if (ret == 1) - rt_cache_flush(0); + rt_cache_flush(net, 0); return ret; } diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c index 5ad01d63f83b..65c1503f8cc8 100644 --- a/net/ipv4/fib_frontend.c +++ b/net/ipv4/fib_frontend.c @@ -144,7 +144,7 @@ static void fib_flush(struct net *net) } if (flushed) - rt_cache_flush(-1); + rt_cache_flush(net, -1); } /* @@ -897,21 +897,22 @@ static void fib_disable_ip(struct net_device *dev, int force) { if (fib_sync_down_dev(dev, force)) fib_flush(dev_net(dev)); - rt_cache_flush(0); + rt_cache_flush(dev_net(dev), 0); arp_ifdown(dev); } static int fib_inetaddr_event(struct notifier_block *this, unsigned long event, void *ptr) { struct in_ifaddr *ifa = (struct in_ifaddr*)ptr; + struct net_device *dev = ifa->ifa_dev->dev; switch (event) { case NETDEV_UP: fib_add_ifaddr(ifa); #ifdef CONFIG_IP_ROUTE_MULTIPATH - fib_sync_up(ifa->ifa_dev->dev); + fib_sync_up(dev); #endif - rt_cache_flush(-1); + rt_cache_flush(dev_net(dev), -1); break; case NETDEV_DOWN: fib_del_ifaddr(ifa); @@ -919,9 +920,9 @@ static int fib_inetaddr_event(struct notifier_block *this, unsigned long event, /* Last address was deleted from this interface. Disable IP. */ - fib_disable_ip(ifa->ifa_dev->dev, 1); + fib_disable_ip(dev, 1); } else { - rt_cache_flush(-1); + rt_cache_flush(dev_net(dev), -1); } break; } @@ -949,14 +950,14 @@ static int fib_netdev_event(struct notifier_block *this, unsigned long event, vo #ifdef CONFIG_IP_ROUTE_MULTIPATH fib_sync_up(dev); #endif - rt_cache_flush(-1); + rt_cache_flush(dev_net(dev), -1); break; case NETDEV_DOWN: fib_disable_ip(dev, 0); break; case NETDEV_CHANGEMTU: case NETDEV_CHANGE: - rt_cache_flush(0); + rt_cache_flush(dev_net(dev), 0); break; } return NOTIFY_DONE; diff --git a/net/ipv4/fib_hash.c b/net/ipv4/fib_hash.c index eeec4bf982b8..c8cac6c7f881 100644 --- a/net/ipv4/fib_hash.c +++ b/net/ipv4/fib_hash.c @@ -472,7 +472,7 @@ static int fn_hash_insert(struct fib_table *tb, struct fib_config *cfg) fib_release_info(fi_drop); if (state & FA_S_ACCESSED) - rt_cache_flush(-1); + rt_cache_flush(cfg->fc_nlinfo.nl_net, -1); rtmsg_fib(RTM_NEWROUTE, key, fa, cfg->fc_dst_len, tb->tb_id, &cfg->fc_nlinfo, NLM_F_REPLACE); return 0; @@ -532,7 +532,7 @@ static int fn_hash_insert(struct fib_table *tb, struct fib_config *cfg) if (new_f) fz->fz_nent++; - rt_cache_flush(-1); + rt_cache_flush(cfg->fc_nlinfo.nl_net, -1); rtmsg_fib(RTM_NEWROUTE, key, new_fa, cfg->fc_dst_len, tb->tb_id, &cfg->fc_nlinfo, 0); @@ -614,7 +614,7 @@ static int fn_hash_delete(struct fib_table *tb, struct fib_config *cfg) write_unlock_bh(&fib_hash_lock); if (fa->fa_state & FA_S_ACCESSED) - rt_cache_flush(-1); + rt_cache_flush(cfg->fc_nlinfo.nl_net, -1); fn_free_alias(fa, f); if (kill_fn) { fn_free_node(f); diff --git a/net/ipv4/fib_rules.c b/net/ipv4/fib_rules.c index 1fb56876be54..bc05de413087 100644 --- a/net/ipv4/fib_rules.c +++ b/net/ipv4/fib_rules.c @@ -260,7 +260,7 @@ static size_t fib4_rule_nlmsg_payload(struct fib_rule *rule) static void fib4_rule_flush_cache(void) { - rt_cache_flush(-1); + rt_cache_flush(&init_net, -1); } static struct fib_rules_ops fib4_rules_ops_template = { diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c index 394db9c941a1..d16ae4623be6 100644 --- a/net/ipv4/fib_trie.c +++ b/net/ipv4/fib_trie.c @@ -1271,7 +1271,7 @@ static int fn_trie_insert(struct fib_table *tb, struct fib_config *cfg) fib_release_info(fi_drop); if (state & FA_S_ACCESSED) - rt_cache_flush(-1); + rt_cache_flush(cfg->fc_nlinfo.nl_net, -1); rtmsg_fib(RTM_NEWROUTE, htonl(key), new_fa, plen, tb->tb_id, &cfg->fc_nlinfo, NLM_F_REPLACE); @@ -1316,7 +1316,7 @@ static int fn_trie_insert(struct fib_table *tb, struct fib_config *cfg) list_add_tail_rcu(&new_fa->fa_list, (fa ? &fa->fa_list : fa_head)); - rt_cache_flush(-1); + rt_cache_flush(cfg->fc_nlinfo.nl_net, -1); rtmsg_fib(RTM_NEWROUTE, htonl(key), new_fa, plen, tb->tb_id, &cfg->fc_nlinfo, 0); succeeded: @@ -1664,7 +1664,7 @@ static int fn_trie_delete(struct fib_table *tb, struct fib_config *cfg) trie_leaf_remove(t, l); if (fa->fa_state & FA_S_ACCESSED) - rt_cache_flush(-1); + rt_cache_flush(cfg->fc_nlinfo.nl_net, -1); fib_release_info(fa->fa_info); alias_free_mem_rcu(fa); diff --git a/net/ipv4/route.c b/net/ipv4/route.c index fe3a02237286..cedc366505bd 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -791,7 +791,7 @@ static void rt_cache_invalidate(void) * delay < 0 : invalidate cache (fast : entries will be deleted later) * delay >= 0 : invalidate & flush cache (can be long) */ -void rt_cache_flush(int delay) +void rt_cache_flush(struct net *net, int delay) { rt_cache_invalidate(); if (delay >= 0) @@ -2825,7 +2825,7 @@ done: void ip_rt_multicast_event(struct in_device *in_dev) { - rt_cache_flush(0); + rt_cache_flush(dev_net(in_dev->dev), 0); } #ifdef CONFIG_SYSCTL @@ -2837,7 +2837,7 @@ static int ipv4_sysctl_rtcache_flush(ctl_table *ctl, int write, { if (write) { proc_dointvec(ctl, write, filp, buffer, lenp, ppos); - rt_cache_flush(flush_delay); + rt_cache_flush(&init_net, flush_delay); return 0; } @@ -2857,7 +2857,7 @@ static int ipv4_sysctl_rtcache_flush_strategy(ctl_table *table, return -EINVAL; if (get_user(delay, (int __user *)newval)) return -EFAULT; - rt_cache_flush(delay); + rt_cache_flush(&init_net, delay); return 0; } -- cgit v1.2.2 From ae299fc051aa68ca6ef1807c37bb92d9b6ff817c Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Sat, 5 Jul 2008 19:01:28 -0700 Subject: net: add fib_rules_ops to flush_cache method This is required to pass namespace context into rt_cache_flush called from ->flush_cache. Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/core/fib_rules.c | 2 +- net/decnet/dn_rules.c | 2 +- net/ipv4/fib_rules.c | 4 ++-- 3 files changed, 4 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c index e3e9ab0f74e3..1c2943a119f3 100644 --- a/net/core/fib_rules.c +++ b/net/core/fib_rules.c @@ -69,7 +69,7 @@ static void rules_ops_put(struct fib_rules_ops *ops) static void flush_route_cache(struct fib_rules_ops *ops) { if (ops->flush_cache) - ops->flush_cache(); + ops->flush_cache(ops); } int fib_rules_register(struct fib_rules_ops *ops) diff --git a/net/decnet/dn_rules.c b/net/decnet/dn_rules.c index 5b7539b7fe0c..14fbca55e908 100644 --- a/net/decnet/dn_rules.c +++ b/net/decnet/dn_rules.c @@ -229,7 +229,7 @@ static u32 dn_fib_rule_default_pref(struct fib_rules_ops *ops) return 0; } -static void dn_fib_rule_flush_cache(void) +static void dn_fib_rule_flush_cache(struct fib_rules_ops *ops) { dn_rt_cache_flush(-1); } diff --git a/net/ipv4/fib_rules.c b/net/ipv4/fib_rules.c index bc05de413087..6080d7120821 100644 --- a/net/ipv4/fib_rules.c +++ b/net/ipv4/fib_rules.c @@ -258,9 +258,9 @@ static size_t fib4_rule_nlmsg_payload(struct fib_rule *rule) + nla_total_size(4); /* flow */ } -static void fib4_rule_flush_cache(void) +static void fib4_rule_flush_cache(struct fib_rules_ops *ops) { - rt_cache_flush(&init_net, -1); + rt_cache_flush(ops->fro_net, -1); } static struct fib_rules_ops fib4_rules_ops_template = { -- cgit v1.2.2 From 639e104facec20f64f2eb940851ae45e5f255e6b Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Sat, 5 Jul 2008 19:02:06 -0700 Subject: ipv4: remove static flush_delay variable flush delay is used as an external storage for net.ipv4.route.flush sysctl entry. It is write-only. The ctl_table->data for this entry is used once. Fix this case to point to the stack to remove global variable. Do this to avoid additional variable on struct net in the next patch. Possible race (as it was before) accessing this local variable is removed using flush_mutex. Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/route.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index cedc366505bd..790de32cd7d4 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -2829,14 +2829,20 @@ void ip_rt_multicast_event(struct in_device *in_dev) } #ifdef CONFIG_SYSCTL -static int flush_delay; - static int ipv4_sysctl_rtcache_flush(ctl_table *ctl, int write, struct file *filp, void __user *buffer, size_t *lenp, loff_t *ppos) { if (write) { + int flush_delay; + static DEFINE_MUTEX(flush_mutex); + + mutex_lock(&flush_mutex); + ctl->data = &flush_delay; proc_dointvec(ctl, write, filp, buffer, lenp, ppos); + ctl->data = NULL; + mutex_unlock(&flush_mutex); + rt_cache_flush(&init_net, flush_delay); return 0; } @@ -2865,7 +2871,6 @@ ctl_table ipv4_route_table[] = { { .ctl_name = NET_IPV4_ROUTE_FLUSH, .procname = "flush", - .data = &flush_delay, .maxlen = sizeof(int), .mode = 0200, .proc_handler = &ipv4_sysctl_rtcache_flush, -- cgit v1.2.2 From 39a23e75087ce815abbddbd565b9a2e567ac47da Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Sat, 5 Jul 2008 19:02:33 -0700 Subject: netns: register net.ipv4.route.flush in each namespace Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/route.c | 79 +++++++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 69 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 790de32cd7d4..6fe799de9b96 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -2835,6 +2835,7 @@ static int ipv4_sysctl_rtcache_flush(ctl_table *ctl, int write, { if (write) { int flush_delay; + struct net *net; static DEFINE_MUTEX(flush_mutex); mutex_lock(&flush_mutex); @@ -2843,7 +2844,8 @@ static int ipv4_sysctl_rtcache_flush(ctl_table *ctl, int write, ctl->data = NULL; mutex_unlock(&flush_mutex); - rt_cache_flush(&init_net, flush_delay); + net = (struct net *)ctl->extra1; + rt_cache_flush(net, flush_delay); return 0; } @@ -2859,23 +2861,17 @@ static int ipv4_sysctl_rtcache_flush_strategy(ctl_table *table, size_t newlen) { int delay; + struct net *net; if (newlen != sizeof(int)) return -EINVAL; if (get_user(delay, (int __user *)newval)) return -EFAULT; - rt_cache_flush(&init_net, delay); + net = (struct net *)table->extra1; + rt_cache_flush(net, delay); return 0; } ctl_table ipv4_route_table[] = { - { - .ctl_name = NET_IPV4_ROUTE_FLUSH, - .procname = "flush", - .maxlen = sizeof(int), - .mode = 0200, - .proc_handler = &ipv4_sysctl_rtcache_flush, - .strategy = &ipv4_sysctl_rtcache_flush_strategy, - }, { .ctl_name = NET_IPV4_ROUTE_GC_THRESH, .procname = "gc_thresh", @@ -3014,6 +3010,66 @@ ctl_table ipv4_route_table[] = { }, { .ctl_name = 0 } }; + +static __net_initdata struct ctl_path ipv4_route_path[] = { + { .procname = "net", .ctl_name = CTL_NET, }, + { .procname = "ipv4", .ctl_name = NET_IPV4, }, + { .procname = "route", .ctl_name = NET_IPV4_ROUTE, }, + { }, +}; + + +static struct ctl_table ipv4_route_flush_table[] = { + { + .ctl_name = NET_IPV4_ROUTE_FLUSH, + .procname = "flush", + .maxlen = sizeof(int), + .mode = 0200, + .proc_handler = &ipv4_sysctl_rtcache_flush, + .strategy = &ipv4_sysctl_rtcache_flush_strategy, + }, + { .ctl_name = 0 }, +}; + +static __net_init int sysctl_route_net_init(struct net *net) +{ + struct ctl_table *tbl; + + tbl = ipv4_route_flush_table; + if (net != &init_net) { + tbl = kmemdup(tbl, sizeof(ipv4_route_flush_table), GFP_KERNEL); + if (tbl == NULL) + goto err_dup; + } + tbl[0].extra1 = net; + + net->ipv4.route_hdr = + register_net_sysctl_table(net, ipv4_route_path, tbl); + if (net->ipv4.route_hdr == NULL) + goto err_reg; + return 0; + +err_reg: + if (tbl != ipv4_route_flush_table) + kfree(tbl); +err_dup: + return -ENOMEM; +} + +static __net_exit void sysctl_route_net_exit(struct net *net) +{ + struct ctl_table *tbl; + + tbl = net->ipv4.route_hdr->ctl_table_arg; + unregister_net_sysctl_table(net->ipv4.route_hdr); + BUG_ON(tbl == ipv4_route_flush_table); + kfree(tbl); +} + +static __net_initdata struct pernet_operations sysctl_route_ops = { + .init = sysctl_route_net_init, + .exit = sysctl_route_net_exit, +}; #endif #ifdef CONFIG_NET_CLS_ROUTE @@ -3090,6 +3146,9 @@ int __init ip_rt_init(void) #endif rtnl_register(PF_INET, RTM_GETROUTE, inet_rtm_getroute, NULL); +#ifdef CONFIG_SYSCTL + register_pernet_subsys(&sysctl_route_ops); +#endif return rc; } -- cgit v1.2.2 From 9f5e97e53675caeda48e9988122a30470f4d309d Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Sat, 5 Jul 2008 19:02:59 -0700 Subject: netns: make rt_secret_rebuild timer per namespace Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/route.c | 40 ++++++++++++++++++++++++++++++---------- 1 file changed, 30 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 6fe799de9b96..f99d9dbb7720 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -132,7 +132,6 @@ static int ip_rt_secret_interval __read_mostly = 10 * 60 * HZ; static void rt_worker_func(struct work_struct *work); static DECLARE_DELAYED_WORK(expires_work, rt_worker_func); -static struct timer_list rt_secret_timer; /* * Interface to generic destination cache. @@ -801,10 +800,11 @@ void rt_cache_flush(struct net *net, int delay) /* * We change rt_genid and let gc do the cleanup */ -static void rt_secret_rebuild(unsigned long dummy) +static void rt_secret_rebuild(unsigned long __net) { + struct net *net = (struct net *)__net; rt_cache_invalidate(); - mod_timer(&rt_secret_timer, jiffies + ip_rt_secret_interval); + mod_timer(&net->ipv4.rt_secret_timer, jiffies + ip_rt_secret_interval); } /* @@ -3072,6 +3072,31 @@ static __net_initdata struct pernet_operations sysctl_route_ops = { }; #endif + +static __net_init int rt_secret_timer_init(struct net *net) +{ + net->ipv4.rt_secret_timer.function = rt_secret_rebuild; + net->ipv4.rt_secret_timer.data = (unsigned long)net; + init_timer_deferrable(&net->ipv4.rt_secret_timer); + + net->ipv4.rt_secret_timer.expires = + jiffies + net_random() % ip_rt_secret_interval + + ip_rt_secret_interval; + add_timer(&net->ipv4.rt_secret_timer); + return 0; +} + +static __net_exit void rt_secret_timer_exit(struct net *net) +{ + del_timer_sync(&net->ipv4.rt_secret_timer); +} + +static __net_initdata struct pernet_operations rt_secret_timer_ops = { + .init = rt_secret_timer_init, + .exit = rt_secret_timer_exit, +}; + + #ifdef CONFIG_NET_CLS_ROUTE struct ip_rt_acct *ip_rt_acct __read_mostly; #endif /* CONFIG_NET_CLS_ROUTE */ @@ -3124,19 +3149,14 @@ int __init ip_rt_init(void) devinet_init(); ip_fib_init(); - rt_secret_timer.function = rt_secret_rebuild; - rt_secret_timer.data = 0; - init_timer_deferrable(&rt_secret_timer); - /* All the timers, started at system startup tend to synchronize. Perturb it a bit. */ schedule_delayed_work(&expires_work, net_random() % ip_rt_gc_interval + ip_rt_gc_interval); - rt_secret_timer.expires = jiffies + net_random() % ip_rt_secret_interval + - ip_rt_secret_interval; - add_timer(&rt_secret_timer); + if (register_pernet_subsys(&rt_secret_timer_ops)) + printk(KERN_ERR "Unable to setup rt_secret_timer\n"); if (ip_rt_proc_init()) printk(KERN_ERR "Unable to create route proc files\n"); -- cgit v1.2.2 From 86c657f6b5bbf2f3ec2213eca528998134a9b344 Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Sat, 5 Jul 2008 19:03:31 -0700 Subject: netns: add struct net parameter to rt_cache_invalidate Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/route.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index f99d9dbb7720..9725223ffe9d 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -778,7 +778,7 @@ static void rt_worker_func(struct work_struct *work) * many times (2^24) without giving recent rt_genid. * Jenkins hash is strong enough that litle changes of rt_genid are OK. */ -static void rt_cache_invalidate(void) +static void rt_cache_invalidate(struct net *net) { unsigned char shuffle; @@ -792,7 +792,7 @@ static void rt_cache_invalidate(void) */ void rt_cache_flush(struct net *net, int delay) { - rt_cache_invalidate(); + rt_cache_invalidate(net); if (delay >= 0) rt_do_flush(!in_softirq()); } @@ -803,7 +803,7 @@ void rt_cache_flush(struct net *net, int delay) static void rt_secret_rebuild(unsigned long __net) { struct net *net = (struct net *)__net; - rt_cache_invalidate(); + rt_cache_invalidate(net); mod_timer(&net->ipv4.rt_secret_timer, jiffies + ip_rt_secret_interval); } -- cgit v1.2.2 From b00180defdeeac8e07e3dc02e53e7395d42bbd19 Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Sat, 5 Jul 2008 19:04:09 -0700 Subject: ipv4: pass current value of rt_genid into rt_hash Basically, there is no difference to atomic_read internally or pass it as a parameter as rt_hash is inline. Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/route.c | 28 +++++++++++++++++----------- 1 file changed, 17 insertions(+), 11 deletions(-) (limited to 'net') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 9725223ffe9d..e4e37edbad60 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -256,11 +256,12 @@ static DEFINE_PER_CPU(struct rt_cache_stat, rt_cache_stat); #define RT_CACHE_STAT_INC(field) \ (__raw_get_cpu_var(rt_cache_stat).field++) -static inline unsigned int rt_hash(__be32 daddr, __be32 saddr, int idx) +static inline unsigned int rt_hash(__be32 daddr, __be32 saddr, int idx, + int genid) { return jhash_3words((__force u32)(__be32)(daddr), (__force u32)(__be32)(saddr), - idx, atomic_read(&rt_genid)) + idx, genid) & rt_hash_mask; } @@ -1180,7 +1181,8 @@ void ip_rt_redirect(__be32 old_gw, __be32 daddr, __be32 new_gw, for (i = 0; i < 2; i++) { for (k = 0; k < 2; k++) { - unsigned hash = rt_hash(daddr, skeys[i], ikeys[k]); + unsigned hash = rt_hash(daddr, skeys[i], ikeys[k], + atomic_read(&rt_genid)); rthp=&rt_hash_table[hash].chain; @@ -1295,7 +1297,8 @@ static struct dst_entry *ipv4_negative_advice(struct dst_entry *dst) } else if ((rt->rt_flags & RTCF_REDIRECTED) || rt->u.dst.expires) { unsigned hash = rt_hash(rt->fl.fl4_dst, rt->fl.fl4_src, - rt->fl.oif); + rt->fl.oif, + atomic_read(&rt_genid)); #if RT_CACHE_DEBUG >= 1 printk(KERN_DEBUG "ipv4_negative_advice: redirect to " NIPQUAD_FMT "/%02x dropped\n", @@ -1444,7 +1447,8 @@ unsigned short ip_rt_frag_needed(struct net *net, struct iphdr *iph, for (k = 0; k < 2; k++) { for (i = 0; i < 2; i++) { - unsigned hash = rt_hash(daddr, skeys[i], ikeys[k]); + unsigned hash = rt_hash(daddr, skeys[i], ikeys[k], + atomic_read(&rt_genid)); rcu_read_lock(); for (rth = rcu_dereference(rt_hash_table[hash].chain); rth; @@ -1709,7 +1713,7 @@ static int ip_route_input_mc(struct sk_buff *skb, __be32 daddr, __be32 saddr, RT_CACHE_STAT_INC(in_slow_mc); in_dev_put(in_dev); - hash = rt_hash(daddr, saddr, dev->ifindex); + hash = rt_hash(daddr, saddr, dev->ifindex, atomic_read(&rt_genid)); return rt_intern_hash(hash, rth, &skb->rtable); e_nobufs: @@ -1870,7 +1874,7 @@ static int ip_mkroute_input(struct sk_buff *skb, return err; /* put it into the cache */ - hash = rt_hash(daddr, saddr, fl->iif); + hash = rt_hash(daddr, saddr, fl->iif, atomic_read(&rt_genid)); return rt_intern_hash(hash, rth, &skb->rtable); } @@ -2026,7 +2030,7 @@ local_input: rth->rt_flags &= ~RTCF_LOCAL; } rth->rt_type = res.type; - hash = rt_hash(daddr, saddr, fl.iif); + hash = rt_hash(daddr, saddr, fl.iif, atomic_read(&rt_genid)); err = rt_intern_hash(hash, rth, &skb->rtable); goto done; @@ -2077,7 +2081,7 @@ int ip_route_input(struct sk_buff *skb, __be32 daddr, __be32 saddr, net = dev_net(dev); tos &= IPTOS_RT_MASK; - hash = rt_hash(daddr, saddr, iif); + hash = rt_hash(daddr, saddr, iif, atomic_read(&rt_genid)); rcu_read_lock(); for (rth = rcu_dereference(rt_hash_table[hash].chain); rth; @@ -2266,7 +2270,8 @@ static int ip_mkroute_output(struct rtable **rp, int err = __mkroute_output(&rth, res, fl, oldflp, dev_out, flags); unsigned hash; if (err == 0) { - hash = rt_hash(oldflp->fl4_dst, oldflp->fl4_src, oldflp->oif); + hash = rt_hash(oldflp->fl4_dst, oldflp->fl4_src, oldflp->oif, + atomic_read(&rt_genid)); err = rt_intern_hash(hash, rth, rp); } @@ -2478,7 +2483,8 @@ int __ip_route_output_key(struct net *net, struct rtable **rp, unsigned hash; struct rtable *rth; - hash = rt_hash(flp->fl4_dst, flp->fl4_src, flp->oif); + hash = rt_hash(flp->fl4_dst, flp->fl4_src, flp->oif, + atomic_read(&rt_genid)); rcu_read_lock_bh(); for (rth = rcu_dereference(rt_hash_table[hash].chain); rth; -- cgit v1.2.2 From e84f84f276473dcc673f360e8ff3203148bdf0e2 Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Sat, 5 Jul 2008 19:04:32 -0700 Subject: netns: place rt_genid into struct net Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/route.c | 76 ++++++++++++++++++++++++++++++++------------------------ 1 file changed, 43 insertions(+), 33 deletions(-) (limited to 'net') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index e4e37edbad60..67c3ed772c27 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -250,7 +250,6 @@ static inline void rt_hash_lock_init(void) static struct rt_hash_bucket *rt_hash_table __read_mostly; static unsigned rt_hash_mask __read_mostly; static unsigned int rt_hash_log __read_mostly; -static atomic_t rt_genid __read_mostly; static DEFINE_PER_CPU(struct rt_cache_stat, rt_cache_stat); #define RT_CACHE_STAT_INC(field) \ @@ -265,6 +264,11 @@ static inline unsigned int rt_hash(__be32 daddr, __be32 saddr, int idx, & rt_hash_mask; } +static inline int rt_genid(struct net *net) +{ + return atomic_read(&net->ipv4.rt_genid); +} + #ifdef CONFIG_PROC_FS struct rt_cache_iter_state { struct seq_net_private p; @@ -334,7 +338,7 @@ static void *rt_cache_seq_start(struct seq_file *seq, loff_t *pos) struct rt_cache_iter_state *st = seq->private; if (*pos) return rt_cache_get_idx(seq, *pos - 1); - st->genid = atomic_read(&rt_genid); + st->genid = rt_genid(seq_file_net(seq)); return SEQ_START_TOKEN; } @@ -681,6 +685,11 @@ static inline int compare_netns(struct rtable *rt1, struct rtable *rt2) return dev_net(rt1->u.dst.dev) == dev_net(rt2->u.dst.dev); } +static inline int rt_is_expired(struct rtable *rth) +{ + return rth->rt_genid != rt_genid(dev_net(rth->u.dst.dev)); +} + /* * Perform a full scan of hash table and free all entries. * Can be called by a softirq or a process. @@ -736,7 +745,7 @@ static void rt_check_expire(void) continue; spin_lock_bh(rt_hash_lock_addr(i)); while ((rth = *rthp) != NULL) { - if (rth->rt_genid != atomic_read(&rt_genid)) { + if (rt_is_expired(rth)) { *rthp = rth->u.dst.rt_next; rt_free(rth); continue; @@ -784,7 +793,7 @@ static void rt_cache_invalidate(struct net *net) unsigned char shuffle; get_random_bytes(&shuffle, sizeof(shuffle)); - atomic_add(shuffle + 1U, &rt_genid); + atomic_add(shuffle + 1U, &net->ipv4.rt_genid); } /* @@ -881,7 +890,7 @@ static int rt_garbage_collect(struct dst_ops *ops) rthp = &rt_hash_table[k].chain; spin_lock_bh(rt_hash_lock_addr(k)); while ((rth = *rthp) != NULL) { - if (rth->rt_genid == atomic_read(&rt_genid) && + if (!rt_is_expired(rth) && !rt_may_expire(rth, tmo, expire)) { tmo >>= 1; rthp = &rth->u.dst.rt_next; @@ -963,7 +972,7 @@ restart: spin_lock_bh(rt_hash_lock_addr(hash)); while ((rth = *rthp) != NULL) { - if (rth->rt_genid != atomic_read(&rt_genid)) { + if (rt_is_expired(rth)) { *rthp = rth->u.dst.rt_next; rt_free(rth); continue; @@ -1139,7 +1148,7 @@ static void rt_del(unsigned hash, struct rtable *rt) spin_lock_bh(rt_hash_lock_addr(hash)); ip_rt_put(rt); while ((aux = *rthp) != NULL) { - if (aux == rt || (aux->rt_genid != atomic_read(&rt_genid))) { + if (aux == rt || rt_is_expired(aux)) { *rthp = aux->u.dst.rt_next; rt_free(aux); continue; @@ -1182,7 +1191,7 @@ void ip_rt_redirect(__be32 old_gw, __be32 daddr, __be32 new_gw, for (i = 0; i < 2; i++) { for (k = 0; k < 2; k++) { unsigned hash = rt_hash(daddr, skeys[i], ikeys[k], - atomic_read(&rt_genid)); + rt_genid(net)); rthp=&rt_hash_table[hash].chain; @@ -1194,7 +1203,7 @@ void ip_rt_redirect(__be32 old_gw, __be32 daddr, __be32 new_gw, rth->fl.fl4_src != skeys[i] || rth->fl.oif != ikeys[k] || rth->fl.iif != 0 || - rth->rt_genid != atomic_read(&rt_genid) || + rt_is_expired(rth) || !net_eq(dev_net(rth->u.dst.dev), net)) { rthp = &rth->u.dst.rt_next; continue; @@ -1233,7 +1242,7 @@ void ip_rt_redirect(__be32 old_gw, __be32 daddr, __be32 new_gw, rt->u.dst.neighbour = NULL; rt->u.dst.hh = NULL; rt->u.dst.xfrm = NULL; - rt->rt_genid = atomic_read(&rt_genid); + rt->rt_genid = rt_genid(net); rt->rt_flags |= RTCF_REDIRECTED; /* Gateway is different ... */ @@ -1298,7 +1307,7 @@ static struct dst_entry *ipv4_negative_advice(struct dst_entry *dst) rt->u.dst.expires) { unsigned hash = rt_hash(rt->fl.fl4_dst, rt->fl.fl4_src, rt->fl.oif, - atomic_read(&rt_genid)); + rt_genid(dev_net(dst->dev))); #if RT_CACHE_DEBUG >= 1 printk(KERN_DEBUG "ipv4_negative_advice: redirect to " NIPQUAD_FMT "/%02x dropped\n", @@ -1448,7 +1457,7 @@ unsigned short ip_rt_frag_needed(struct net *net, struct iphdr *iph, for (k = 0; k < 2; k++) { for (i = 0; i < 2; i++) { unsigned hash = rt_hash(daddr, skeys[i], ikeys[k], - atomic_read(&rt_genid)); + rt_genid(net)); rcu_read_lock(); for (rth = rcu_dereference(rt_hash_table[hash].chain); rth; @@ -1463,7 +1472,7 @@ unsigned short ip_rt_frag_needed(struct net *net, struct iphdr *iph, rth->fl.iif != 0 || dst_metric_locked(&rth->u.dst, RTAX_MTU) || !net_eq(dev_net(rth->u.dst.dev), net) || - rth->rt_genid != atomic_read(&rt_genid)) + !rt_is_expired(rth)) continue; if (new_mtu < 68 || new_mtu >= old_mtu) { @@ -1698,7 +1707,7 @@ static int ip_route_input_mc(struct sk_buff *skb, __be32 daddr, __be32 saddr, rth->fl.oif = 0; rth->rt_gateway = daddr; rth->rt_spec_dst= spec_dst; - rth->rt_genid = atomic_read(&rt_genid); + rth->rt_genid = rt_genid(dev_net(dev)); rth->rt_flags = RTCF_MULTICAST; rth->rt_type = RTN_MULTICAST; if (our) { @@ -1713,7 +1722,7 @@ static int ip_route_input_mc(struct sk_buff *skb, __be32 daddr, __be32 saddr, RT_CACHE_STAT_INC(in_slow_mc); in_dev_put(in_dev); - hash = rt_hash(daddr, saddr, dev->ifindex, atomic_read(&rt_genid)); + hash = rt_hash(daddr, saddr, dev->ifindex, rt_genid(dev_net(dev))); return rt_intern_hash(hash, rth, &skb->rtable); e_nobufs: @@ -1839,7 +1848,7 @@ static int __mkroute_input(struct sk_buff *skb, rth->u.dst.input = ip_forward; rth->u.dst.output = ip_output; - rth->rt_genid = atomic_read(&rt_genid); + rth->rt_genid = rt_genid(dev_net(rth->u.dst.dev)); rt_set_nexthop(rth, res, itag); @@ -1874,7 +1883,8 @@ static int ip_mkroute_input(struct sk_buff *skb, return err; /* put it into the cache */ - hash = rt_hash(daddr, saddr, fl->iif, atomic_read(&rt_genid)); + hash = rt_hash(daddr, saddr, fl->iif, + rt_genid(dev_net(rth->u.dst.dev))); return rt_intern_hash(hash, rth, &skb->rtable); } @@ -2000,7 +2010,7 @@ local_input: goto e_nobufs; rth->u.dst.output= ip_rt_bug; - rth->rt_genid = atomic_read(&rt_genid); + rth->rt_genid = rt_genid(net); atomic_set(&rth->u.dst.__refcnt, 1); rth->u.dst.flags= DST_HOST; @@ -2030,7 +2040,7 @@ local_input: rth->rt_flags &= ~RTCF_LOCAL; } rth->rt_type = res.type; - hash = rt_hash(daddr, saddr, fl.iif, atomic_read(&rt_genid)); + hash = rt_hash(daddr, saddr, fl.iif, rt_genid(net)); err = rt_intern_hash(hash, rth, &skb->rtable); goto done; @@ -2081,7 +2091,7 @@ int ip_route_input(struct sk_buff *skb, __be32 daddr, __be32 saddr, net = dev_net(dev); tos &= IPTOS_RT_MASK; - hash = rt_hash(daddr, saddr, iif, atomic_read(&rt_genid)); + hash = rt_hash(daddr, saddr, iif, rt_genid(net)); rcu_read_lock(); for (rth = rcu_dereference(rt_hash_table[hash].chain); rth; @@ -2093,7 +2103,7 @@ int ip_route_input(struct sk_buff *skb, __be32 daddr, __be32 saddr, (rth->fl.fl4_tos ^ tos)) == 0 && rth->fl.mark == skb->mark && net_eq(dev_net(rth->u.dst.dev), net) && - rth->rt_genid == atomic_read(&rt_genid)) { + !rt_is_expired(rth)) { dst_use(&rth->u.dst, jiffies); RT_CACHE_STAT_INC(in_hit); rcu_read_unlock(); @@ -2221,7 +2231,7 @@ static int __mkroute_output(struct rtable **result, rth->rt_spec_dst= fl->fl4_src; rth->u.dst.output=ip_output; - rth->rt_genid = atomic_read(&rt_genid); + rth->rt_genid = rt_genid(dev_net(dev_out)); RT_CACHE_STAT_INC(out_slow_tot); @@ -2271,7 +2281,7 @@ static int ip_mkroute_output(struct rtable **rp, unsigned hash; if (err == 0) { hash = rt_hash(oldflp->fl4_dst, oldflp->fl4_src, oldflp->oif, - atomic_read(&rt_genid)); + rt_genid(dev_net(dev_out))); err = rt_intern_hash(hash, rth, rp); } @@ -2483,8 +2493,7 @@ int __ip_route_output_key(struct net *net, struct rtable **rp, unsigned hash; struct rtable *rth; - hash = rt_hash(flp->fl4_dst, flp->fl4_src, flp->oif, - atomic_read(&rt_genid)); + hash = rt_hash(flp->fl4_dst, flp->fl4_src, flp->oif, rt_genid(net)); rcu_read_lock_bh(); for (rth = rcu_dereference(rt_hash_table[hash].chain); rth; @@ -2497,7 +2506,7 @@ int __ip_route_output_key(struct net *net, struct rtable **rp, !((rth->fl.fl4_tos ^ flp->fl4_tos) & (IPTOS_RT_MASK | RTO_ONLINK)) && net_eq(dev_net(rth->u.dst.dev), net) && - rth->rt_genid == atomic_read(&rt_genid)) { + !rt_is_expired(rth)) { dst_use(&rth->u.dst, jiffies); RT_CACHE_STAT_INC(out_hit); rcu_read_unlock_bh(); @@ -2528,7 +2537,7 @@ static struct dst_ops ipv4_dst_blackhole_ops = { }; -static int ipv4_dst_blackhole(struct rtable **rp, struct flowi *flp) +static int ipv4_dst_blackhole(struct net *net, struct rtable **rp, struct flowi *flp) { struct rtable *ort = *rp; struct rtable *rt = (struct rtable *) @@ -2552,7 +2561,7 @@ static int ipv4_dst_blackhole(struct rtable **rp, struct flowi *flp) rt->idev = ort->idev; if (rt->idev) in_dev_hold(rt->idev); - rt->rt_genid = atomic_read(&rt_genid); + rt->rt_genid = rt_genid(net); rt->rt_flags = ort->rt_flags; rt->rt_type = ort->rt_type; rt->rt_dst = ort->rt_dst; @@ -2588,7 +2597,7 @@ int ip_route_output_flow(struct net *net, struct rtable **rp, struct flowi *flp, err = __xfrm_lookup((struct dst_entry **)rp, flp, sk, flags ? XFRM_LOOKUP_WAIT : 0); if (err == -EREMOTE) - err = ipv4_dst_blackhole(rp, flp); + err = ipv4_dst_blackhole(net, rp, flp); return err; } @@ -2807,7 +2816,7 @@ int ip_rt_dump(struct sk_buff *skb, struct netlink_callback *cb) rt = rcu_dereference(rt->u.dst.rt_next), idx++) { if (!net_eq(dev_net(rt->u.dst.dev), net) || idx < s_idx) continue; - if (rt->rt_genid != atomic_read(&rt_genid)) + if (rt_is_expired(rt)) continue; skb->dst = dst_clone(&rt->u.dst); if (rt_fill_info(skb, NETLINK_CB(cb->skb).pid, @@ -3081,6 +3090,10 @@ static __net_initdata struct pernet_operations sysctl_route_ops = { static __net_init int rt_secret_timer_init(struct net *net) { + atomic_set(&net->ipv4.rt_genid, + (int) ((num_physpages ^ (num_physpages>>8)) ^ + (jiffies ^ (jiffies >> 7)))); + net->ipv4.rt_secret_timer.function = rt_secret_rebuild; net->ipv4.rt_secret_timer.data = (unsigned long)net; init_timer_deferrable(&net->ipv4.rt_secret_timer); @@ -3121,9 +3134,6 @@ int __init ip_rt_init(void) { int rc = 0; - atomic_set(&rt_genid, (int) ((num_physpages ^ (num_physpages>>8)) ^ - (jiffies ^ (jiffies >> 7)))); - #ifdef CONFIG_NET_CLS_ROUTE ip_rt_acct = __alloc_percpu(256 * sizeof(struct ip_rt_acct)); if (!ip_rt_acct) -- cgit v1.2.2 From 32cb5b4e035e3d7b52f1e9de87920645a00e5234 Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Sat, 5 Jul 2008 19:06:12 -0700 Subject: netns: selective flush of rt_cache dst cache is marked as expired on the per/namespace basis by previous path. Right now we have to implement selective cache shrinking. This procedure has been ported from older OpenVz codebase. Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/route.c | 31 ++++++++++++++++++++++++++++++- 1 file changed, 30 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 67c3ed772c27..113cd2512ba7 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -699,6 +699,7 @@ static void rt_do_flush(int process_context) { unsigned int i; struct rtable *rth, *next; + struct rtable * tail; for (i = 0; i <= rt_hash_mask; i++) { if (process_context && need_resched()) @@ -708,11 +709,39 @@ static void rt_do_flush(int process_context) continue; spin_lock_bh(rt_hash_lock_addr(i)); +#ifdef CONFIG_NET_NS + { + struct rtable ** prev, * p; + + rth = rt_hash_table[i].chain; + + /* defer releasing the head of the list after spin_unlock */ + for (tail = rth; tail; tail = tail->u.dst.rt_next) + if (!rt_is_expired(tail)) + break; + if (rth != tail) + rt_hash_table[i].chain = tail; + + /* call rt_free on entries after the tail requiring flush */ + prev = &rt_hash_table[i].chain; + for (p = *prev; p; p = next) { + next = p->u.dst.rt_next; + if (!rt_is_expired(p)) { + prev = &p->u.dst.rt_next; + } else { + *prev = next; + rt_free(p); + } + } + } +#else rth = rt_hash_table[i].chain; rt_hash_table[i].chain = NULL; + tail = NULL; +#endif spin_unlock_bh(rt_hash_lock_addr(i)); - for (; rth; rth = next) { + for (; rth != tail; rth = next) { next = rth->u.dst.rt_next; rt_free(rth); } -- cgit v1.2.2 From 629ca23c331ec75ac87b016debbb3c4d2fe62650 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Sat, 5 Jul 2008 21:18:07 -0700 Subject: MIB: add struct net to UDP_INC_STATS_USER Nothing special - all the places already have a struct sock at hands, so use the sock_net() net. Signed-off-by: Pavel Emelyanov Acked-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/udp.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 3bbf6fb6e4f5..c97da5394d70 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -526,7 +526,8 @@ out: up->len = 0; up->pending = 0; if (!err) - UDP_INC_STATS_USER(UDP_MIB_OUTDATAGRAMS, is_udplite); + UDP_INC_STATS_USER(sock_net(sk), + UDP_MIB_OUTDATAGRAMS, is_udplite); return err; } @@ -725,7 +726,8 @@ out: * seems like overkill. */ if (err == -ENOBUFS || test_bit(SOCK_NOSPACE, &sk->sk_socket->flags)) { - UDP_INC_STATS_USER(UDP_MIB_SNDBUFERRORS, is_udplite); + UDP_INC_STATS_USER(sock_net(sk), + UDP_MIB_SNDBUFERRORS, is_udplite); } return err; @@ -888,7 +890,8 @@ try_again: goto out_free; if (!peeked) - UDP_INC_STATS_USER(UDP_MIB_INDATAGRAMS, is_udplite); + UDP_INC_STATS_USER(sock_net(sk), + UDP_MIB_INDATAGRAMS, is_udplite); sock_recv_timestamp(msg, sk, skb); @@ -917,7 +920,7 @@ out: csum_copy_err: lock_sock(sk); if (!skb_kill_datagram(sk, skb, flags)) - UDP_INC_STATS_USER(UDP_MIB_INERRORS, is_udplite); + UDP_INC_STATS_USER(sock_net(sk), UDP_MIB_INERRORS, is_udplite); release_sock(sk); if (noblock) -- cgit v1.2.2 From 0283328e2360bbf3081e97f1b720dd4c4dbae111 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Sat, 5 Jul 2008 21:18:48 -0700 Subject: MIB: add struct net to UDP_INC_STATS_BH Two special cases here - one is rxrpc - I put init_net there explicitly, since we haven't touched this part yet. The second place is in __udp4_lib_rcv - we already have a struct net there, but I have to move its initialization above to make it ready at the "drop" label. Signed-off-by: Pavel Emelyanov Acked-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/udp.c | 18 ++++++++++-------- net/rxrpc/ar-input.c | 5 +++-- 2 files changed, 13 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index c97da5394d70..7187121e922d 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -991,7 +991,8 @@ int udp_queue_rcv_skb(struct sock * sk, struct sk_buff *skb) ret = (*up->encap_rcv)(sk, skb); if (ret <= 0) { - UDP_INC_STATS_BH(UDP_MIB_INDATAGRAMS, + UDP_INC_STATS_BH(sock_net(sk), + UDP_MIB_INDATAGRAMS, is_udplite); return -ret; } @@ -1044,7 +1045,8 @@ int udp_queue_rcv_skb(struct sock * sk, struct sk_buff *skb) if ((rc = sock_queue_rcv_skb(sk,skb)) < 0) { /* Note that an ENOMEM error is charged twice */ if (rc == -ENOMEM) { - UDP_INC_STATS_BH(UDP_MIB_RCVBUFERRORS, is_udplite); + UDP_INC_STATS_BH(sock_net(sk), + UDP_MIB_RCVBUFERRORS, is_udplite); atomic_inc(&sk->sk_drops); } goto drop; @@ -1053,7 +1055,7 @@ int udp_queue_rcv_skb(struct sock * sk, struct sk_buff *skb) return 0; drop: - UDP_INC_STATS_BH(UDP_MIB_INERRORS, is_udplite); + UDP_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite); kfree_skb(skb); return -1; } @@ -1161,7 +1163,7 @@ int __udp4_lib_rcv(struct sk_buff *skb, struct hlist_head udptable[], struct rtable *rt = (struct rtable*)skb->dst; __be32 saddr = ip_hdr(skb)->saddr; __be32 daddr = ip_hdr(skb)->daddr; - struct net *net; + struct net *net = dev_net(skb->dev); /* * Validate the packet. @@ -1183,7 +1185,6 @@ int __udp4_lib_rcv(struct sk_buff *skb, struct hlist_head udptable[], if (udp4_csum_init(skb, uh, proto)) goto csum_error; - net = dev_net(skb->dev); if (rt->rt_flags & (RTCF_BROADCAST|RTCF_MULTICAST)) return __udp4_lib_mcast_deliver(net, skb, uh, saddr, daddr, udptable); @@ -1217,7 +1218,7 @@ int __udp4_lib_rcv(struct sk_buff *skb, struct hlist_head udptable[], if (udp_lib_checksum_complete(skb)) goto csum_error; - UDP_INC_STATS_BH(UDP_MIB_NOPORTS, proto == IPPROTO_UDPLITE); + UDP_INC_STATS_BH(net, UDP_MIB_NOPORTS, proto == IPPROTO_UDPLITE); icmp_send(skb, ICMP_DEST_UNREACH, ICMP_PORT_UNREACH, 0); /* @@ -1251,7 +1252,7 @@ csum_error: ntohs(uh->dest), ulen); drop: - UDP_INC_STATS_BH(UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE); + UDP_INC_STATS_BH(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE); kfree_skb(skb); return 0; } @@ -1458,7 +1459,8 @@ unsigned int udp_poll(struct file *file, struct socket *sock, poll_table *wait) spin_lock_bh(&rcvq->lock); while ((skb = skb_peek(rcvq)) != NULL && udp_lib_checksum_complete(skb)) { - UDP_INC_STATS_BH(UDP_MIB_INERRORS, is_lite); + UDP_INC_STATS_BH(sock_net(sk), + UDP_MIB_INERRORS, is_lite); __skb_unlink(skb, rcvq); kfree_skb(skb); } diff --git a/net/rxrpc/ar-input.c b/net/rxrpc/ar-input.c index f8a699e92962..f98c8027e5c1 100644 --- a/net/rxrpc/ar-input.c +++ b/net/rxrpc/ar-input.c @@ -21,6 +21,7 @@ #include #include #include +#include #include "ar-internal.h" unsigned long rxrpc_ack_timeout = 1; @@ -708,12 +709,12 @@ void rxrpc_data_ready(struct sock *sk, int count) if (skb_checksum_complete(skb)) { rxrpc_free_skb(skb); rxrpc_put_local(local); - UDP_INC_STATS_BH(UDP_MIB_INERRORS, 0); + UDP_INC_STATS_BH(&init_net, UDP_MIB_INERRORS, 0); _leave(" [CSUM failed]"); return; } - UDP_INC_STATS_BH(UDP_MIB_INDATAGRAMS, 0); + UDP_INC_STATS_BH(&init_net, UDP_MIB_INDATAGRAMS, 0); /* the socket buffer we have is owned by UDP, with UDP's data all over * it, but we really want our own */ -- cgit v1.2.2 From 235b9f7ac53489011d32efeb89e12e308fdd2c64 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Sat, 5 Jul 2008 21:19:20 -0700 Subject: MIB: add struct net to UDP6_INC_STATS_USER As simple as the patch #1 in this set. Signed-off-by: Pavel Emelyanov Acked-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv6/udp.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index f91e1df0d25e..833f715e4bff 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -166,7 +166,8 @@ try_again: goto out_free; if (!peeked) - UDP6_INC_STATS_USER(UDP_MIB_INDATAGRAMS, is_udplite); + UDP6_INC_STATS_USER(sock_net(sk), + UDP_MIB_INDATAGRAMS, is_udplite); sock_recv_timestamp(msg, sk, skb); @@ -213,7 +214,7 @@ out: csum_copy_err: lock_sock(sk); if (!skb_kill_datagram(sk, skb, flags)) - UDP6_INC_STATS_USER(UDP_MIB_INERRORS, is_udplite); + UDP6_INC_STATS_USER(sock_net(sk), UDP_MIB_INERRORS, is_udplite); release_sock(sk); if (flags & MSG_DONTWAIT) @@ -591,7 +592,8 @@ out: up->len = 0; up->pending = 0; if (!err) - UDP6_INC_STATS_USER(UDP_MIB_OUTDATAGRAMS, is_udplite); + UDP6_INC_STATS_USER(sock_net(sk), + UDP_MIB_OUTDATAGRAMS, is_udplite); return err; } @@ -873,7 +875,8 @@ out: * seems like overkill. */ if (err == -ENOBUFS || test_bit(SOCK_NOSPACE, &sk->sk_socket->flags)) { - UDP6_INC_STATS_USER(UDP_MIB_SNDBUFERRORS, is_udplite); + UDP6_INC_STATS_USER(sock_net(sk), + UDP_MIB_SNDBUFERRORS, is_udplite); } return err; -- cgit v1.2.2 From ef28d1a20f9f18ebf1be15ef6f097a76f9a63499 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Sat, 5 Jul 2008 21:19:40 -0700 Subject: MIB: add struct net to UDP6_INC_STATS_BH Signed-off-by: Pavel Emelyanov Acked-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv6/udp.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index 833f715e4bff..d1477b350f76 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -299,7 +299,8 @@ int udpv6_queue_rcv_skb(struct sock * sk, struct sk_buff *skb) if ((rc = sock_queue_rcv_skb(sk,skb)) < 0) { /* Note that an ENOMEM error is charged twice */ if (rc == -ENOMEM) { - UDP6_INC_STATS_BH(UDP_MIB_RCVBUFERRORS, is_udplite); + UDP6_INC_STATS_BH(sock_net(sk), + UDP_MIB_RCVBUFERRORS, is_udplite); atomic_inc(&sk->sk_drops); } goto drop; @@ -307,7 +308,7 @@ int udpv6_queue_rcv_skb(struct sock * sk, struct sk_buff *skb) return 0; drop: - UDP6_INC_STATS_BH(UDP_MIB_INERRORS, is_udplite); + UDP6_INC_STATS_BH(sock_net(sk), UDP_MIB_INERRORS, is_udplite); kfree_skb(skb); return -1; } @@ -439,7 +440,7 @@ int __udp6_lib_rcv(struct sk_buff *skb, struct hlist_head udptable[], struct net_device *dev = skb->dev; struct in6_addr *saddr, *daddr; u32 ulen = 0; - struct net *net; + struct net *net = dev_net(skb->dev); if (!pskb_may_pull(skb, sizeof(struct udphdr))) goto short_packet; @@ -474,7 +475,6 @@ int __udp6_lib_rcv(struct sk_buff *skb, struct hlist_head udptable[], if (udp6_csum_init(skb, uh, proto)) goto discard; - net = dev_net(skb->dev); /* * Multicast receive code */ @@ -497,7 +497,8 @@ int __udp6_lib_rcv(struct sk_buff *skb, struct hlist_head udptable[], if (udp_lib_checksum_complete(skb)) goto discard; - UDP6_INC_STATS_BH(UDP_MIB_NOPORTS, proto == IPPROTO_UDPLITE); + UDP6_INC_STATS_BH(net, UDP_MIB_NOPORTS, + proto == IPPROTO_UDPLITE); icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_PORT_UNREACH, 0, dev); @@ -522,7 +523,7 @@ short_packet: ulen, skb->len); discard: - UDP6_INC_STATS_BH(UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE); + UDP6_INC_STATS_BH(net, UDP_MIB_INERRORS, proto == IPPROTO_UDPLITE); kfree_skb(skb); return 0; } -- cgit v1.2.2 From a19800d704177caaa5874baf5819307c5b7d5e4f Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 21:25:39 -0700 Subject: net: Add STP demux layer Add small STP demux layer for demuxing STP PDUs based on MAC address. This is needed to run both GARP and STP in parallel (or even load the modules) since both use LLC_SAP_BSPAN. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/802/Kconfig | 3 ++ net/802/Makefile | 1 + net/802/stp.c | 102 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ net/Kconfig | 1 + 4 files changed, 107 insertions(+) create mode 100644 net/802/Kconfig create mode 100644 net/802/stp.c (limited to 'net') diff --git a/net/802/Kconfig b/net/802/Kconfig new file mode 100644 index 000000000000..01cb0943626b --- /dev/null +++ b/net/802/Kconfig @@ -0,0 +1,3 @@ +config STP + tristate + select LLC diff --git a/net/802/Makefile b/net/802/Makefile index 68569ffddea1..c441d895ac25 100644 --- a/net/802/Makefile +++ b/net/802/Makefile @@ -10,3 +10,4 @@ obj-$(CONFIG_FDDI) += fddi.o obj-$(CONFIG_HIPPI) += hippi.o obj-$(CONFIG_IPX) += p8022.o psnap.o p8023.o obj-$(CONFIG_ATALK) += p8022.o psnap.o +obj-$(CONFIG_STP) += stp.o diff --git a/net/802/stp.c b/net/802/stp.c new file mode 100644 index 000000000000..0b7a24452d11 --- /dev/null +++ b/net/802/stp.c @@ -0,0 +1,102 @@ +/* + * STP SAP demux + * + * Copyright (c) 2008 Patrick McHardy + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * version 2 as published by the Free Software Foundation. + */ +#include +#include +#include +#include +#include +#include +#include + +/* 01:80:c2:00:00:20 - 01:80:c2:00:00:2F */ +#define GARP_ADDR_MIN 0x20 +#define GARP_ADDR_MAX 0x2F +#define GARP_ADDR_RANGE (GARP_ADDR_MAX - GARP_ADDR_MIN) + +static const struct stp_proto *garp_protos[GARP_ADDR_RANGE + 1] __read_mostly; +static const struct stp_proto *stp_proto __read_mostly; + +static struct llc_sap *sap __read_mostly; +static unsigned int sap_registered; +static DEFINE_MUTEX(stp_proto_mutex); + +/* Called under rcu_read_lock from LLC */ +static int stp_pdu_rcv(struct sk_buff *skb, struct net_device *dev, + struct packet_type *pt, struct net_device *orig_dev) +{ + const struct ethhdr *eh = eth_hdr(skb); + const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb); + const struct stp_proto *proto; + + if (pdu->ssap != LLC_SAP_BSPAN || + pdu->dsap != LLC_SAP_BSPAN || + pdu->ctrl_1 != LLC_PDU_TYPE_U) + goto err; + + if (eh->h_dest[5] >= GARP_ADDR_MIN && eh->h_dest[5] <= GARP_ADDR_MAX) { + proto = rcu_dereference(garp_protos[eh->h_dest[5] - + GARP_ADDR_MIN]); + if (proto && + compare_ether_addr(eh->h_dest, proto->group_address)) + goto err; + } else + proto = rcu_dereference(stp_proto); + + if (!proto) + goto err; + + proto->rcv(proto, skb, dev); + return 0; + +err: + kfree_skb(skb); + return 0; +} + +int stp_proto_register(const struct stp_proto *proto) +{ + int err = 0; + + mutex_lock(&stp_proto_mutex); + if (sap_registered++ == 0) { + sap = llc_sap_open(LLC_SAP_BSPAN, stp_pdu_rcv); + if (!sap) { + err = -ENOMEM; + goto out; + } + } + if (is_zero_ether_addr(proto->group_address)) + rcu_assign_pointer(stp_proto, proto); + else + rcu_assign_pointer(garp_protos[proto->group_address[5] - + GARP_ADDR_MIN], proto); +out: + mutex_unlock(&stp_proto_mutex); + return err; +} +EXPORT_SYMBOL_GPL(stp_proto_register); + +void stp_proto_unregister(const struct stp_proto *proto) +{ + mutex_lock(&stp_proto_mutex); + if (is_zero_ether_addr(proto->group_address)) + rcu_assign_pointer(stp_proto, NULL); + else + rcu_assign_pointer(garp_protos[proto->group_address[5] - + GARP_ADDR_MIN], NULL); + synchronize_rcu(); + + if (--sap_registered == 0) + llc_sap_put(sap); + mutex_unlock(&stp_proto_mutex); +} +EXPORT_SYMBOL_GPL(stp_proto_unregister); + +MODULE_LICENSE("GPL"); diff --git a/net/Kconfig b/net/Kconfig index acbf7c60e89b..b98668751749 100644 --- a/net/Kconfig +++ b/net/Kconfig @@ -181,6 +181,7 @@ source "net/dccp/Kconfig" source "net/sctp/Kconfig" source "net/tipc/Kconfig" source "net/atm/Kconfig" +source "net/802/Kconfig" source "net/bridge/Kconfig" source "net/8021q/Kconfig" source "net/decnet/Kconfig" -- cgit v1.2.2 From 7c85fbf0657f216557b0c9c4a2e4e07f37d8bb8c Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 21:25:56 -0700 Subject: bridge: Use STP demux Use the STP demux layer for receiving STP PDUs instead of directly registering with LLC. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/bridge/Kconfig | 1 + net/bridge/br.c | 16 +++++++++------- net/bridge/br_private.h | 5 +++-- net/bridge/br_stp_bpdu.c | 12 +++--------- 4 files changed, 16 insertions(+), 18 deletions(-) (limited to 'net') diff --git a/net/bridge/Kconfig b/net/bridge/Kconfig index 12265aff7099..e143ca678881 100644 --- a/net/bridge/Kconfig +++ b/net/bridge/Kconfig @@ -5,6 +5,7 @@ config BRIDGE tristate "802.1d Ethernet Bridging" select LLC + select STP ---help--- If you say Y here, then your Linux box will be able to act as an Ethernet bridge, which means that the different Ethernet segments it diff --git a/net/bridge/br.c b/net/bridge/br.c index cede010f4ddd..573acdf6f9ff 100644 --- a/net/bridge/br.c +++ b/net/bridge/br.c @@ -18,21 +18,24 @@ #include #include #include +#include #include "br_private.h" int (*br_should_route_hook)(struct sk_buff *skb); -static struct llc_sap *br_stp_sap; +static const struct stp_proto br_stp_proto = { + .rcv = br_stp_rcv, +}; static int __init br_init(void) { int err; - br_stp_sap = llc_sap_open(LLC_SAP_BSPAN, br_stp_rcv); - if (!br_stp_sap) { + err = stp_proto_register(&br_stp_proto); + if (err < 0) { printk(KERN_ERR "bridge: can't register sap for STP\n"); - return -EADDRINUSE; + return err; } err = br_fdb_init(); @@ -65,13 +68,13 @@ err_out2: err_out1: br_fdb_fini(); err_out: - llc_sap_put(br_stp_sap); + stp_proto_unregister(&br_stp_proto); return err; } static void __exit br_deinit(void) { - rcu_assign_pointer(br_stp_sap->rcv_func, NULL); + stp_proto_unregister(&br_stp_proto); br_netlink_fini(); unregister_netdevice_notifier(&br_device_notifier); @@ -82,7 +85,6 @@ static void __exit br_deinit(void) synchronize_net(); br_netfilter_fini(); - llc_sap_put(br_stp_sap); br_fdb_get_hook = NULL; br_fdb_put_hook = NULL; diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h index 8593c9f6a302..815ed38925b2 100644 --- a/net/bridge/br_private.h +++ b/net/bridge/br_private.h @@ -226,8 +226,9 @@ extern void br_stp_set_path_cost(struct net_bridge_port *p, extern ssize_t br_show_bridge_id(char *buf, const struct bridge_id *id); /* br_stp_bpdu.c */ -extern int br_stp_rcv(struct sk_buff *skb, struct net_device *dev, - struct packet_type *pt, struct net_device *orig_dev); +struct stp_proto; +extern void br_stp_rcv(const struct stp_proto *proto, struct sk_buff *skb, + struct net_device *dev); /* br_stp_timer.c */ extern void br_stp_timer_init(struct net_bridge *br); diff --git a/net/bridge/br_stp_bpdu.c b/net/bridge/br_stp_bpdu.c index 9dc2de656965..996476174517 100644 --- a/net/bridge/br_stp_bpdu.c +++ b/net/bridge/br_stp_bpdu.c @@ -18,6 +18,7 @@ #include #include #include +#include #include #include "br_private.h" @@ -131,10 +132,9 @@ void br_send_tcn_bpdu(struct net_bridge_port *p) * * NO locks, but rcu_read_lock (preempt_disabled) */ -int br_stp_rcv(struct sk_buff *skb, struct net_device *dev, - struct packet_type *pt, struct net_device *orig_dev) +void br_stp_rcv(const struct stp_proto *proto, struct sk_buff *skb, + struct net_device *dev) { - const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb); const unsigned char *dest = eth_hdr(skb)->h_dest; struct net_bridge_port *p = rcu_dereference(dev->br_port); struct net_bridge *br; @@ -146,11 +146,6 @@ int br_stp_rcv(struct sk_buff *skb, struct net_device *dev, if (!p) goto err; - if (pdu->ssap != LLC_SAP_BSPAN - || pdu->dsap != LLC_SAP_BSPAN - || pdu->ctrl_1 != LLC_PDU_TYPE_U) - goto err; - if (!pskb_may_pull(skb, 4)) goto err; @@ -224,5 +219,4 @@ int br_stp_rcv(struct sk_buff *skb, struct net_device *dev, spin_unlock(&br->lock); err: kfree_skb(skb); - return 0; } -- cgit v1.2.2 From eca9ebac651f774d8b10fce7c5d173c3c3d3394f Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 21:26:13 -0700 Subject: net: Add GARP applicant-only participant Add an implementation of the GARP (Generic Attribute Registration Protocol) applicant-only participant. This will be used by the following patch to add GVRP support to the VLAN code. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/802/Kconfig | 4 + net/802/Makefile | 1 + net/802/garp.c | 633 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 638 insertions(+) create mode 100644 net/802/garp.c (limited to 'net') diff --git a/net/802/Kconfig b/net/802/Kconfig index 01cb0943626b..be33d27c8e69 100644 --- a/net/802/Kconfig +++ b/net/802/Kconfig @@ -1,3 +1,7 @@ config STP tristate select LLC + +config GARP + tristate + select STP diff --git a/net/802/Makefile b/net/802/Makefile index c441d895ac25..7893d679910c 100644 --- a/net/802/Makefile +++ b/net/802/Makefile @@ -11,3 +11,4 @@ obj-$(CONFIG_HIPPI) += hippi.o obj-$(CONFIG_IPX) += p8022.o psnap.o p8023.o obj-$(CONFIG_ATALK) += p8022.o psnap.o obj-$(CONFIG_STP) += stp.o +obj-$(CONFIG_GARP) += garp.o diff --git a/net/802/garp.c b/net/802/garp.c new file mode 100644 index 000000000000..3b78f7b74fd4 --- /dev/null +++ b/net/802/garp.c @@ -0,0 +1,633 @@ +/* + * IEEE 802.1D Generic Attribute Registration Protocol (GARP) + * + * Copyright (c) 2008 Patrick McHardy + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * version 2 as published by the Free Software Foundation. + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +static unsigned int garp_join_time __read_mostly = 200; +module_param(garp_join_time, uint, 0644); +MODULE_PARM_DESC(garp_join_time, "Join time in ms (default 200ms)"); +MODULE_LICENSE("GPL"); + +static const struct garp_state_trans { + u8 state; + u8 action; +} garp_applicant_state_table[GARP_APPLICANT_MAX + 1][GARP_EVENT_MAX + 1] = { + [GARP_APPLICANT_VA] = { + [GARP_EVENT_TRANSMIT_PDU] = { .state = GARP_APPLICANT_AA, + .action = GARP_ACTION_S_JOIN_IN }, + [GARP_EVENT_R_JOIN_IN] = { .state = GARP_APPLICANT_AA }, + [GARP_EVENT_R_JOIN_EMPTY] = { .state = GARP_APPLICANT_VA }, + [GARP_EVENT_R_EMPTY] = { .state = GARP_APPLICANT_VA }, + [GARP_EVENT_R_LEAVE_IN] = { .state = GARP_APPLICANT_VA }, + [GARP_EVENT_R_LEAVE_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_REQ_JOIN] = { .state = GARP_APPLICANT_INVALID }, + [GARP_EVENT_REQ_LEAVE] = { .state = GARP_APPLICANT_LA }, + }, + [GARP_APPLICANT_AA] = { + [GARP_EVENT_TRANSMIT_PDU] = { .state = GARP_APPLICANT_QA, + .action = GARP_ACTION_S_JOIN_IN }, + [GARP_EVENT_R_JOIN_IN] = { .state = GARP_APPLICANT_QA }, + [GARP_EVENT_R_JOIN_EMPTY] = { .state = GARP_APPLICANT_VA }, + [GARP_EVENT_R_EMPTY] = { .state = GARP_APPLICANT_VA }, + [GARP_EVENT_R_LEAVE_IN] = { .state = GARP_APPLICANT_VA }, + [GARP_EVENT_R_LEAVE_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_REQ_JOIN] = { .state = GARP_APPLICANT_INVALID }, + [GARP_EVENT_REQ_LEAVE] = { .state = GARP_APPLICANT_LA }, + }, + [GARP_APPLICANT_QA] = { + [GARP_EVENT_TRANSMIT_PDU] = { .state = GARP_APPLICANT_INVALID }, + [GARP_EVENT_R_JOIN_IN] = { .state = GARP_APPLICANT_QA }, + [GARP_EVENT_R_JOIN_EMPTY] = { .state = GARP_APPLICANT_VA }, + [GARP_EVENT_R_EMPTY] = { .state = GARP_APPLICANT_VA }, + [GARP_EVENT_R_LEAVE_IN] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_R_LEAVE_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_REQ_JOIN] = { .state = GARP_APPLICANT_INVALID }, + [GARP_EVENT_REQ_LEAVE] = { .state = GARP_APPLICANT_LA }, + }, + [GARP_APPLICANT_LA] = { + [GARP_EVENT_TRANSMIT_PDU] = { .state = GARP_APPLICANT_VO, + .action = GARP_ACTION_S_LEAVE_EMPTY }, + [GARP_EVENT_R_JOIN_IN] = { .state = GARP_APPLICANT_LA }, + [GARP_EVENT_R_JOIN_EMPTY] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_R_EMPTY] = { .state = GARP_APPLICANT_LA }, + [GARP_EVENT_R_LEAVE_IN] = { .state = GARP_APPLICANT_LA }, + [GARP_EVENT_R_LEAVE_EMPTY] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_REQ_JOIN] = { .state = GARP_APPLICANT_VA }, + [GARP_EVENT_REQ_LEAVE] = { .state = GARP_APPLICANT_INVALID }, + }, + [GARP_APPLICANT_VP] = { + [GARP_EVENT_TRANSMIT_PDU] = { .state = GARP_APPLICANT_AA, + .action = GARP_ACTION_S_JOIN_IN }, + [GARP_EVENT_R_JOIN_IN] = { .state = GARP_APPLICANT_AP }, + [GARP_EVENT_R_JOIN_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_R_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_R_LEAVE_IN] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_R_LEAVE_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_REQ_JOIN] = { .state = GARP_APPLICANT_INVALID }, + [GARP_EVENT_REQ_LEAVE] = { .state = GARP_APPLICANT_VO }, + }, + [GARP_APPLICANT_AP] = { + [GARP_EVENT_TRANSMIT_PDU] = { .state = GARP_APPLICANT_QA, + .action = GARP_ACTION_S_JOIN_IN }, + [GARP_EVENT_R_JOIN_IN] = { .state = GARP_APPLICANT_QP }, + [GARP_EVENT_R_JOIN_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_R_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_R_LEAVE_IN] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_R_LEAVE_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_REQ_JOIN] = { .state = GARP_APPLICANT_INVALID }, + [GARP_EVENT_REQ_LEAVE] = { .state = GARP_APPLICANT_AO }, + }, + [GARP_APPLICANT_QP] = { + [GARP_EVENT_TRANSMIT_PDU] = { .state = GARP_APPLICANT_INVALID }, + [GARP_EVENT_R_JOIN_IN] = { .state = GARP_APPLICANT_QP }, + [GARP_EVENT_R_JOIN_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_R_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_R_LEAVE_IN] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_R_LEAVE_EMPTY] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_REQ_JOIN] = { .state = GARP_APPLICANT_INVALID }, + [GARP_EVENT_REQ_LEAVE] = { .state = GARP_APPLICANT_QO }, + }, + [GARP_APPLICANT_VO] = { + [GARP_EVENT_TRANSMIT_PDU] = { .state = GARP_APPLICANT_INVALID }, + [GARP_EVENT_R_JOIN_IN] = { .state = GARP_APPLICANT_AO }, + [GARP_EVENT_R_JOIN_EMPTY] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_R_EMPTY] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_R_LEAVE_IN] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_R_LEAVE_EMPTY] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_REQ_JOIN] = { .state = GARP_APPLICANT_VP }, + [GARP_EVENT_REQ_LEAVE] = { .state = GARP_APPLICANT_INVALID }, + }, + [GARP_APPLICANT_AO] = { + [GARP_EVENT_TRANSMIT_PDU] = { .state = GARP_APPLICANT_INVALID }, + [GARP_EVENT_R_JOIN_IN] = { .state = GARP_APPLICANT_QO }, + [GARP_EVENT_R_JOIN_EMPTY] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_R_EMPTY] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_R_LEAVE_IN] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_R_LEAVE_EMPTY] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_REQ_JOIN] = { .state = GARP_APPLICANT_AP }, + [GARP_EVENT_REQ_LEAVE] = { .state = GARP_APPLICANT_INVALID }, + }, + [GARP_APPLICANT_QO] = { + [GARP_EVENT_TRANSMIT_PDU] = { .state = GARP_APPLICANT_INVALID }, + [GARP_EVENT_R_JOIN_IN] = { .state = GARP_APPLICANT_QO }, + [GARP_EVENT_R_JOIN_EMPTY] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_R_EMPTY] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_R_LEAVE_IN] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_R_LEAVE_EMPTY] = { .state = GARP_APPLICANT_VO }, + [GARP_EVENT_REQ_JOIN] = { .state = GARP_APPLICANT_QP }, + [GARP_EVENT_REQ_LEAVE] = { .state = GARP_APPLICANT_INVALID }, + }, +}; + +static int garp_attr_cmp(const struct garp_attr *attr, + const void *data, u8 len, u8 type) +{ + if (attr->type != type) + return attr->type - type; + if (attr->dlen != len) + return attr->dlen - len; + return memcmp(attr->data, data, len); +} + +static struct garp_attr *garp_attr_lookup(const struct garp_applicant *app, + const void *data, u8 len, u8 type) +{ + struct rb_node *parent = app->gid.rb_node; + struct garp_attr *attr; + int d; + + while (parent) { + attr = rb_entry(parent, struct garp_attr, node); + d = garp_attr_cmp(attr, data, len, type); + if (d < 0) + parent = parent->rb_left; + else if (d > 0) + parent = parent->rb_right; + else + return attr; + } + return NULL; +} + +static void garp_attr_insert(struct garp_applicant *app, struct garp_attr *new) +{ + struct rb_node *parent = NULL, **p = &app->gid.rb_node; + struct garp_attr *attr; + int d; + + while (*p) { + parent = *p; + attr = rb_entry(parent, struct garp_attr, node); + d = garp_attr_cmp(attr, new->data, new->dlen, new->type); + if (d < 0) + p = &parent->rb_left; + else if (d > 0) + p = &parent->rb_right; + } + rb_link_node(&new->node, parent, p); + rb_insert_color(&new->node, &app->gid); +} + +static struct garp_attr *garp_attr_create(struct garp_applicant *app, + const void *data, u8 len, u8 type) +{ + struct garp_attr *attr; + + attr = kmalloc(sizeof(*attr) + len, GFP_ATOMIC); + if (!attr) + return attr; + attr->state = GARP_APPLICANT_VO; + attr->type = type; + attr->dlen = len; + memcpy(attr->data, data, len); + garp_attr_insert(app, attr); + return attr; +} + +static void garp_attr_destroy(struct garp_applicant *app, struct garp_attr *attr) +{ + rb_erase(&attr->node, &app->gid); + kfree(attr); +} + +static int garp_pdu_init(struct garp_applicant *app) +{ + struct sk_buff *skb; + struct garp_pdu_hdr *gp; + +#define LLC_RESERVE sizeof(struct llc_pdu_un) + skb = alloc_skb(app->dev->mtu + LL_RESERVED_SPACE(app->dev), + GFP_ATOMIC); + if (!skb) + return -ENOMEM; + + skb->dev = app->dev; + skb->protocol = htons(ETH_P_802_2); + skb_reserve(skb, LL_RESERVED_SPACE(app->dev) + LLC_RESERVE); + + gp = (struct garp_pdu_hdr *)__skb_put(skb, sizeof(*gp)); + put_unaligned(htons(GARP_PROTOCOL_ID), &gp->protocol); + + app->pdu = skb; + return 0; +} + +static int garp_pdu_append_end_mark(struct garp_applicant *app) +{ + if (skb_tailroom(app->pdu) < sizeof(u8)) + return -1; + *(u8 *)__skb_put(app->pdu, sizeof(u8)) = GARP_END_MARK; + return 0; +} + +static void garp_pdu_queue(struct garp_applicant *app) +{ + if (!app->pdu) + return; + + garp_pdu_append_end_mark(app); + garp_pdu_append_end_mark(app); + + llc_pdu_header_init(app->pdu, LLC_PDU_TYPE_U, LLC_SAP_BSPAN, + LLC_SAP_BSPAN, LLC_PDU_CMD); + llc_pdu_init_as_ui_cmd(app->pdu); + llc_mac_hdr_init(app->pdu, app->dev->dev_addr, + app->app->proto.group_address); + + skb_queue_tail(&app->queue, app->pdu); + app->pdu = NULL; +} + +static void garp_queue_xmit(struct garp_applicant *app) +{ + struct sk_buff *skb; + + while ((skb = skb_dequeue(&app->queue))) + dev_queue_xmit(skb); +} + +static int garp_pdu_append_msg(struct garp_applicant *app, u8 attrtype) +{ + struct garp_msg_hdr *gm; + + if (skb_tailroom(app->pdu) < sizeof(*gm)) + return -1; + gm = (struct garp_msg_hdr *)__skb_put(app->pdu, sizeof(*gm)); + gm->attrtype = attrtype; + garp_cb(app->pdu)->cur_type = attrtype; + return 0; +} + +static int garp_pdu_append_attr(struct garp_applicant *app, + const struct garp_attr *attr, + enum garp_attr_event event) +{ + struct garp_attr_hdr *ga; + unsigned int len; + int err; +again: + if (!app->pdu) { + err = garp_pdu_init(app); + if (err < 0) + return err; + } + + if (garp_cb(app->pdu)->cur_type != attr->type) { + if (garp_cb(app->pdu)->cur_type && + garp_pdu_append_end_mark(app) < 0) + goto queue; + if (garp_pdu_append_msg(app, attr->type) < 0) + goto queue; + } + + len = sizeof(*ga) + attr->dlen; + if (skb_tailroom(app->pdu) < len) + goto queue; + ga = (struct garp_attr_hdr *)__skb_put(app->pdu, len); + ga->len = len; + ga->event = event; + memcpy(ga->data, attr->data, attr->dlen); + return 0; + +queue: + garp_pdu_queue(app); + goto again; +} + +static void garp_attr_event(struct garp_applicant *app, + struct garp_attr *attr, enum garp_event event) +{ + enum garp_applicant_state state; + + state = garp_applicant_state_table[attr->state][event].state; + if (state == GARP_APPLICANT_INVALID) + return; + + switch (garp_applicant_state_table[attr->state][event].action) { + case GARP_ACTION_NONE: + break; + case GARP_ACTION_S_JOIN_IN: + garp_pdu_append_attr(app, attr, GARP_JOIN_IN); + break; + case GARP_ACTION_S_LEAVE_EMPTY: + garp_pdu_append_attr(app, attr, GARP_LEAVE_EMPTY); + /* As a pure applicant, sending a leave message implies that + * the attribute was unregistered and can be destroyed. */ + garp_attr_destroy(app, attr); + return; + default: + WARN_ON(1); + } + + attr->state = state; +} + +int garp_request_join(const struct net_device *dev, + const struct garp_application *appl, + const void *data, u8 len, u8 type) +{ + struct garp_port *port = dev->garp_port; + struct garp_applicant *app = port->applicants[appl->type]; + struct garp_attr *attr; + + spin_lock_bh(&app->lock); + attr = garp_attr_create(app, data, len, type); + if (!attr) { + spin_unlock_bh(&app->lock); + return -ENOMEM; + } + garp_attr_event(app, attr, GARP_EVENT_REQ_JOIN); + spin_unlock_bh(&app->lock); + return 0; +} +EXPORT_SYMBOL_GPL(garp_request_join); + +void garp_request_leave(const struct net_device *dev, + const struct garp_application *appl, + const void *data, u8 len, u8 type) +{ + struct garp_port *port = dev->garp_port; + struct garp_applicant *app = port->applicants[appl->type]; + struct garp_attr *attr; + + spin_lock_bh(&app->lock); + attr = garp_attr_lookup(app, data, len, type); + if (!attr) { + spin_unlock_bh(&app->lock); + return; + } + garp_attr_event(app, attr, GARP_EVENT_REQ_LEAVE); + spin_unlock_bh(&app->lock); +} +EXPORT_SYMBOL_GPL(garp_request_leave); + +static void garp_gid_event(struct garp_applicant *app, enum garp_event event) +{ + struct rb_node *node, *next; + struct garp_attr *attr; + + for (node = rb_first(&app->gid); + next = node ? rb_next(node) : NULL, node != NULL; + node = next) { + attr = rb_entry(node, struct garp_attr, node); + garp_attr_event(app, attr, event); + } +} + +static void garp_join_timer_arm(struct garp_applicant *app) +{ + unsigned long delay; + + delay = (u64)msecs_to_jiffies(garp_join_time) * net_random() >> 32; + mod_timer(&app->join_timer, jiffies + delay); +} + +static void garp_join_timer(unsigned long data) +{ + struct garp_applicant *app = (struct garp_applicant *)data; + + spin_lock(&app->lock); + garp_gid_event(app, GARP_EVENT_TRANSMIT_PDU); + garp_pdu_queue(app); + spin_unlock(&app->lock); + + garp_queue_xmit(app); + garp_join_timer_arm(app); +} + +static int garp_pdu_parse_end_mark(struct sk_buff *skb) +{ + if (!pskb_may_pull(skb, sizeof(u8))) + return -1; + if (*skb->data == GARP_END_MARK) { + skb_pull(skb, sizeof(u8)); + return -1; + } + return 0; +} + +static int garp_pdu_parse_attr(struct garp_applicant *app, struct sk_buff *skb, + u8 attrtype) +{ + const struct garp_attr_hdr *ga; + struct garp_attr *attr; + enum garp_event event; + unsigned int dlen; + + if (!pskb_may_pull(skb, sizeof(*ga))) + return -1; + ga = (struct garp_attr_hdr *)skb->data; + if (ga->len < sizeof(*ga)) + return -1; + + if (!pskb_may_pull(skb, ga->len)) + return -1; + skb_pull(skb, ga->len); + dlen = sizeof(*ga) - ga->len; + + if (attrtype > app->app->maxattr) + return 0; + + switch (ga->event) { + case GARP_LEAVE_ALL: + if (dlen != 0) + return -1; + garp_gid_event(app, GARP_EVENT_R_LEAVE_EMPTY); + return 0; + case GARP_JOIN_EMPTY: + event = GARP_EVENT_R_JOIN_EMPTY; + break; + case GARP_JOIN_IN: + event = GARP_EVENT_R_JOIN_IN; + break; + case GARP_LEAVE_EMPTY: + event = GARP_EVENT_R_LEAVE_EMPTY; + break; + case GARP_EMPTY: + event = GARP_EVENT_R_EMPTY; + break; + default: + return 0; + } + + if (dlen == 0) + return -1; + attr = garp_attr_lookup(app, ga->data, dlen, attrtype); + if (attr == NULL) + return 0; + garp_attr_event(app, attr, event); + return 0; +} + +static int garp_pdu_parse_msg(struct garp_applicant *app, struct sk_buff *skb) +{ + const struct garp_msg_hdr *gm; + + if (!pskb_may_pull(skb, sizeof(*gm))) + return -1; + gm = (struct garp_msg_hdr *)skb->data; + if (gm->attrtype == 0) + return -1; + skb_pull(skb, sizeof(*gm)); + + while (skb->len > 0) { + if (garp_pdu_parse_attr(app, skb, gm->attrtype) < 0) + return -1; + if (garp_pdu_parse_end_mark(skb) < 0) + break; + } + return 0; +} + +static void garp_pdu_rcv(const struct stp_proto *proto, struct sk_buff *skb, + struct net_device *dev) +{ + struct garp_application *appl = proto->data; + struct garp_port *port; + struct garp_applicant *app; + const struct garp_pdu_hdr *gp; + + port = rcu_dereference(dev->garp_port); + if (!port) + goto err; + app = rcu_dereference(port->applicants[appl->type]); + if (!app) + goto err; + + if (!pskb_may_pull(skb, sizeof(*gp))) + goto err; + gp = (struct garp_pdu_hdr *)skb->data; + if (get_unaligned(&gp->protocol) != htons(GARP_PROTOCOL_ID)) + goto err; + skb_pull(skb, sizeof(*gp)); + + spin_lock(&app->lock); + while (skb->len > 0) { + if (garp_pdu_parse_msg(app, skb) < 0) + break; + if (garp_pdu_parse_end_mark(skb) < 0) + break; + } + spin_unlock(&app->lock); +err: + kfree_skb(skb); +} + +static int garp_init_port(struct net_device *dev) +{ + struct garp_port *port; + + port = kzalloc(sizeof(*port), GFP_KERNEL); + if (!port) + return -ENOMEM; + rcu_assign_pointer(dev->garp_port, port); + return 0; +} + +static void garp_release_port(struct net_device *dev) +{ + struct garp_port *port = dev->garp_port; + unsigned int i; + + for (i = 0; i <= GARP_APPLICATION_MAX; i++) { + if (port->applicants[i]) + return; + } + rcu_assign_pointer(dev->garp_port, NULL); + synchronize_rcu(); + kfree(port); +} + +int garp_init_applicant(struct net_device *dev, struct garp_application *appl) +{ + struct garp_applicant *app; + int err; + + ASSERT_RTNL(); + + if (!dev->garp_port) { + err = garp_init_port(dev); + if (err < 0) + goto err1; + } + + err = -ENOMEM; + app = kzalloc(sizeof(*app), GFP_KERNEL); + if (!app) + goto err2; + + err = dev_mc_add(dev, appl->proto.group_address, ETH_ALEN, 0); + if (err < 0) + goto err3; + + app->dev = dev; + app->app = appl; + app->gid = RB_ROOT; + spin_lock_init(&app->lock); + skb_queue_head_init(&app->queue); + rcu_assign_pointer(dev->garp_port->applicants[appl->type], app); + setup_timer(&app->join_timer, garp_join_timer, (unsigned long)app); + garp_join_timer_arm(app); + return 0; + +err3: + kfree(app); +err2: + garp_release_port(dev); +err1: + return err; +} +EXPORT_SYMBOL_GPL(garp_init_applicant); + +void garp_uninit_applicant(struct net_device *dev, struct garp_application *appl) +{ + struct garp_port *port = dev->garp_port; + struct garp_applicant *app = port->applicants[appl->type]; + + ASSERT_RTNL(); + + rcu_assign_pointer(port->applicants[appl->type], NULL); + synchronize_rcu(); + + /* Delete timer and generate a final TRANSMIT_PDU event to flush out + * all pending messages before the applicant is gone. */ + del_timer_sync(&app->join_timer); + garp_gid_event(app, GARP_EVENT_TRANSMIT_PDU); + garp_pdu_queue(app); + garp_queue_xmit(app); + + dev_mc_delete(dev, appl->proto.group_address, ETH_ALEN, 0); + kfree(app); + garp_release_port(dev); +} +EXPORT_SYMBOL_GPL(garp_uninit_applicant); + +int garp_register_application(struct garp_application *appl) +{ + appl->proto.rcv = garp_pdu_rcv; + appl->proto.data = appl; + return stp_proto_register(&appl->proto); +} +EXPORT_SYMBOL_GPL(garp_register_application); + +void garp_unregister_application(struct garp_application *appl) +{ + stp_proto_unregister(&appl->proto); +} +EXPORT_SYMBOL_GPL(garp_unregister_application); -- cgit v1.2.2 From b3ce0325f2bf1bb737cb8fc7c349ce8fefdd9e40 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 21:26:27 -0700 Subject: vlan: Change vlan_dev_set_vlan_flag() to handle multiple flags at once Change vlan_dev_set_vlan_flag() to handle multiple flags at once and rename to vlan_dev_change_flags(). This allows to to use it from the netlink interface, which in turn allows to handle necessary adjustments when changing flags centrally. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan.c | 6 +++--- net/8021q/vlan.h | 3 +-- net/8021q/vlan_dev.c | 20 +++++++++----------- net/8021q/vlan_netlink.c | 4 +--- 4 files changed, 14 insertions(+), 19 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c index ab2225da0ee2..b591bfca1ab2 100644 --- a/net/8021q/vlan.c +++ b/net/8021q/vlan.c @@ -591,9 +591,9 @@ static int vlan_ioctl_handler(struct net *net, void __user *arg) err = -EPERM; if (!capable(CAP_NET_ADMIN)) break; - err = vlan_dev_set_vlan_flag(dev, - args.u.flag, - args.vlan_qos); + err = vlan_dev_change_flags(dev, + args.vlan_qos ? args.u.flag : 0, + args.u.flag); break; case SET_VLAN_NAME_TYPE_CMD: diff --git a/net/8021q/vlan.h b/net/8021q/vlan.h index 5229a72c7ea1..639e2544a804 100644 --- a/net/8021q/vlan.h +++ b/net/8021q/vlan.h @@ -28,8 +28,7 @@ void vlan_dev_set_ingress_priority(const struct net_device *dev, u32 skb_prio, short vlan_prio); int vlan_dev_set_egress_priority(const struct net_device *dev, u32 skb_prio, short vlan_prio); -int vlan_dev_set_vlan_flag(const struct net_device *dev, - u32 flag, short flag_val); +int vlan_dev_change_flags(const struct net_device *dev, u32 flag, u32 mask); void vlan_dev_get_realdev_name(const struct net_device *dev, char *result); void vlan_dev_get_vid(const struct net_device *dev, unsigned short *result); diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 5d055c242ed8..76c665cdab66 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -507,18 +507,16 @@ int vlan_dev_set_egress_priority(const struct net_device *dev, } /* Flags are defined in the vlan_flags enum in include/linux/if_vlan.h file. */ -int vlan_dev_set_vlan_flag(const struct net_device *dev, - u32 flag, short flag_val) +int vlan_dev_change_flags(const struct net_device *dev, u32 flags, u32 mask) { - /* verify flag is supported */ - if (flag == VLAN_FLAG_REORDER_HDR) { - if (flag_val) - vlan_dev_info(dev)->flags |= VLAN_FLAG_REORDER_HDR; - else - vlan_dev_info(dev)->flags &= ~VLAN_FLAG_REORDER_HDR; - return 0; - } - return -EINVAL; + struct vlan_dev_info *vlan = vlan_dev_info(dev); + u32 old_flags = vlan->flags; + + if (mask & ~VLAN_FLAG_REORDER_HDR) + return -EINVAL; + + vlan->flags = (old_flags & ~mask) | (flags & mask); + return 0; } void vlan_dev_get_realdev_name(const struct net_device *dev, char *result) diff --git a/net/8021q/vlan_netlink.c b/net/8021q/vlan_netlink.c index c93e69ec28ed..fd7cb195d53f 100644 --- a/net/8021q/vlan_netlink.c +++ b/net/8021q/vlan_netlink.c @@ -75,7 +75,6 @@ static int vlan_validate(struct nlattr *tb[], struct nlattr *data[]) static int vlan_changelink(struct net_device *dev, struct nlattr *tb[], struct nlattr *data[]) { - struct vlan_dev_info *vlan = vlan_dev_info(dev); struct ifla_vlan_flags *flags; struct ifla_vlan_qos_mapping *m; struct nlattr *attr; @@ -83,8 +82,7 @@ static int vlan_changelink(struct net_device *dev, if (data[IFLA_VLAN_FLAGS]) { flags = nla_data(data[IFLA_VLAN_FLAGS]); - vlan->flags = (vlan->flags & ~flags->mask) | - (flags->flags & flags->mask); + vlan_dev_change_flags(dev, flags->flags, flags->mask); } if (data[IFLA_VLAN_INGRESS_QOS]) { nla_for_each_nested(attr, data[IFLA_VLAN_INGRESS_QOS], rem) { -- cgit v1.2.2 From ce305002e1c9b90c2c151ce18bab0b895dd55ae6 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 21:26:41 -0700 Subject: vlan: Move device unregistration before lower dev cleanup Move the unregister_netdevice() call for the VLAN device before cleanup for the lower device. This is needed by GVRP so it can send a leave message before the applicant on the lower device is cleaned up. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c index b591bfca1ab2..8cae2daeb1cc 100644 --- a/net/8021q/vlan.c +++ b/net/8021q/vlan.c @@ -165,6 +165,8 @@ void unregister_vlan_dev(struct net_device *dev) synchronize_net(); + unregister_netdevice(dev); + /* If the group is now empty, kill off the group. */ if (grp->nr_vlans == 0) { if (real_dev->features & NETIF_F_HW_VLAN_RX) @@ -178,8 +180,6 @@ void unregister_vlan_dev(struct net_device *dev) /* Get rid of the vlan's reference to real_dev */ dev_put(real_dev); - - unregister_netdevice(dev); } static void vlan_transfer_operstate(const struct net_device *dev, -- cgit v1.2.2 From 70c03b49b80ba3634958acc31853771019c0ebd3 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 21:26:57 -0700 Subject: vlan: Add GVRP support Add GVRP support for dynamically registering VLANs with switches. By default GVRP is disabled because we only support the applicant-only participant model, which means it should not be enabled on vlans that are members of a bridge. Since there is currently no way to cleanly determine that, the user is responsible for enabling it. The code is pretty small and low impact, its wrapped in a config option though because it depends on the GARP implementation and the STP core. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/Kconfig | 10 ++++++++ net/8021q/Makefile | 9 +++---- net/8021q/vlan.c | 23 ++++++++++++++--- net/8021q/vlan.h | 16 ++++++++++++ net/8021q/vlan_dev.c | 18 +++++++++++-- net/8021q/vlan_gvrp.c | 66 ++++++++++++++++++++++++++++++++++++++++++++++++ net/8021q/vlan_netlink.c | 3 ++- 7 files changed, 132 insertions(+), 13 deletions(-) create mode 100644 net/8021q/vlan_gvrp.c (limited to 'net') diff --git a/net/8021q/Kconfig b/net/8021q/Kconfig index c4a382e450e2..fa073a54963e 100644 --- a/net/8021q/Kconfig +++ b/net/8021q/Kconfig @@ -17,3 +17,13 @@ config VLAN_8021Q will be called 8021q. If unsure, say N. + +config VLAN_8021Q_GVRP + bool "GVRP (GARP VLAN Registration Protocol) support" + depends on VLAN_8021Q + select GARP + help + Select this to enable GVRP end-system support. GVRP is used for + automatic propagation of registered VLANs to switches. + + If unsure, say N. diff --git a/net/8021q/Makefile b/net/8021q/Makefile index 10ca7f486c3a..3006e9ed7b08 100644 --- a/net/8021q/Makefile +++ b/net/8021q/Makefile @@ -4,9 +4,6 @@ obj-$(CONFIG_VLAN_8021Q) += 8021q.o -8021q-objs := vlan.o vlan_dev.o vlan_netlink.o - -ifeq ($(CONFIG_PROC_FS),y) -8021q-objs += vlanproc.o -endif - +8021q-y := vlan.o vlan_dev.o vlan_netlink.o +8021q-$(CONFIG_VLAN_8021Q_GVRP) += vlan_gvrp.o +8021q-$(CONFIG_PROC_FS) += vlanproc.o diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c index 8cae2daeb1cc..b529110c9355 100644 --- a/net/8021q/vlan.c +++ b/net/8021q/vlan.c @@ -169,6 +169,8 @@ void unregister_vlan_dev(struct net_device *dev) /* If the group is now empty, kill off the group. */ if (grp->nr_vlans == 0) { + vlan_gvrp_uninit_applicant(real_dev); + if (real_dev->features & NETIF_F_HW_VLAN_RX) real_dev->vlan_rx_register(real_dev, NULL); @@ -249,15 +251,18 @@ int register_vlan_dev(struct net_device *dev) ngrp = grp = vlan_group_alloc(real_dev); if (!grp) return -ENOBUFS; + err = vlan_gvrp_init_applicant(real_dev); + if (err < 0) + goto out_free_group; } err = vlan_group_prealloc_vid(grp, vlan_id); if (err < 0) - goto out_free_group; + goto out_uninit_applicant; err = register_netdevice(dev); if (err < 0) - goto out_free_group; + goto out_uninit_applicant; /* Account for reference in struct vlan_dev_info */ dev_hold(real_dev); @@ -278,6 +283,9 @@ int register_vlan_dev(struct net_device *dev) return 0; +out_uninit_applicant: + if (ngrp) + vlan_gvrp_uninit_applicant(real_dev); out_free_group: if (ngrp) vlan_group_free(ngrp); @@ -713,14 +721,20 @@ static int __init vlan_proto_init(void) if (err < 0) goto err2; - err = vlan_netlink_init(); + err = vlan_gvrp_init(); if (err < 0) goto err3; + err = vlan_netlink_init(); + if (err < 0) + goto err4; + dev_add_pack(&vlan_packet_type); vlan_ioctl_set(vlan_ioctl_handler); return 0; +err4: + vlan_gvrp_uninit(); err3: unregister_netdevice_notifier(&vlan_notifier_block); err2: @@ -745,8 +759,9 @@ static void __exit vlan_cleanup_module(void) BUG_ON(!hlist_empty(&vlan_group_hash[i])); unregister_pernet_gen_device(vlan_net_id, &vlan_net_ops); - synchronize_net(); + + vlan_gvrp_uninit(); } module_init(vlan_proto_init); diff --git a/net/8021q/vlan.h b/net/8021q/vlan.h index 639e2544a804..097b2e04c928 100644 --- a/net/8021q/vlan.h +++ b/net/8021q/vlan.h @@ -37,6 +37,22 @@ void vlan_setup(struct net_device *dev); int register_vlan_dev(struct net_device *dev); void unregister_vlan_dev(struct net_device *dev); +#ifdef CONFIG_VLAN_8021Q_GVRP +extern int vlan_gvrp_request_join(const struct net_device *dev); +extern void vlan_gvrp_request_leave(const struct net_device *dev); +extern int vlan_gvrp_init_applicant(struct net_device *dev); +extern void vlan_gvrp_uninit_applicant(struct net_device *dev); +extern int vlan_gvrp_init(void); +extern void vlan_gvrp_uninit(void); +#else +static inline int vlan_gvrp_request_join(const struct net_device *dev) { return 0; } +static inline void vlan_gvrp_request_leave(const struct net_device *dev) {} +static inline int vlan_gvrp_init_applicant(struct net_device *dev) { return 0; } +static inline void vlan_gvrp_uninit_applicant(struct net_device *dev) {} +static inline int vlan_gvrp_init(void) { return 0; } +static inline void vlan_gvrp_uninit(void) {} +#endif + int vlan_netlink_init(void); void vlan_netlink_fini(void); diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 76c665cdab66..a0617bf7cec6 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -512,10 +512,17 @@ int vlan_dev_change_flags(const struct net_device *dev, u32 flags, u32 mask) struct vlan_dev_info *vlan = vlan_dev_info(dev); u32 old_flags = vlan->flags; - if (mask & ~VLAN_FLAG_REORDER_HDR) + if (mask & ~(VLAN_FLAG_REORDER_HDR | VLAN_FLAG_GVRP)) return -EINVAL; vlan->flags = (old_flags & ~mask) | (flags & mask); + + if (netif_running(dev) && (vlan->flags ^ old_flags) & VLAN_FLAG_GVRP) { + if (vlan->flags & VLAN_FLAG_GVRP) + vlan_gvrp_request_join(dev); + else + vlan_gvrp_request_leave(dev); + } return 0; } @@ -550,12 +557,19 @@ static int vlan_dev_open(struct net_device *dev) if (dev->flags & IFF_PROMISC) dev_set_promiscuity(real_dev, 1); + if (vlan->flags & VLAN_FLAG_GVRP) + vlan_gvrp_request_join(dev); + return 0; } static int vlan_dev_stop(struct net_device *dev) { - struct net_device *real_dev = vlan_dev_info(dev)->real_dev; + struct vlan_dev_info *vlan = vlan_dev_info(dev); + struct net_device *real_dev = vlan->real_dev; + + if (vlan->flags & VLAN_FLAG_GVRP) + vlan_gvrp_request_leave(dev); dev_mc_unsync(real_dev, dev); dev_unicast_unsync(real_dev, dev); diff --git a/net/8021q/vlan_gvrp.c b/net/8021q/vlan_gvrp.c new file mode 100644 index 000000000000..db9781608362 --- /dev/null +++ b/net/8021q/vlan_gvrp.c @@ -0,0 +1,66 @@ +/* + * IEEE 802.1Q GARP VLAN Registration Protocol (GVRP) + * + * Copyright (c) 2008 Patrick McHardy + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * version 2 as published by the Free Software Foundation. + */ +#include +#include +#include +#include "vlan.h" + +#define GARP_GVRP_ADDRESS { 0x01, 0x80, 0xc2, 0x00, 0x00, 0x21 } + +enum gvrp_attributes { + GVRP_ATTR_INVALID, + GVRP_ATTR_VID, + __GVRP_ATTR_MAX +}; +#define GVRP_ATTR_MAX (__GVRP_ATTR_MAX - 1) + +static struct garp_application vlan_gvrp_app __read_mostly = { + .proto.group_address = GARP_GVRP_ADDRESS, + .maxattr = GVRP_ATTR_MAX, + .type = GARP_APPLICATION_GVRP, +}; + +int vlan_gvrp_request_join(const struct net_device *dev) +{ + const struct vlan_dev_info *vlan = vlan_dev_info(dev); + __be16 vid = htons(vlan->vlan_id); + + return garp_request_join(vlan->real_dev, &vlan_gvrp_app, + &vid, sizeof(vid), GVRP_ATTR_VID); +} + +void vlan_gvrp_request_leave(const struct net_device *dev) +{ + const struct vlan_dev_info *vlan = vlan_dev_info(dev); + __be16 vid = htons(vlan->vlan_id); + + garp_request_leave(vlan->real_dev, &vlan_gvrp_app, + &vid, sizeof(vid), GVRP_ATTR_VID); +} + +int vlan_gvrp_init_applicant(struct net_device *dev) +{ + return garp_init_applicant(dev, &vlan_gvrp_app); +} + +void vlan_gvrp_uninit_applicant(struct net_device *dev) +{ + garp_uninit_applicant(dev, &vlan_gvrp_app); +} + +int __init vlan_gvrp_init(void) +{ + return garp_register_application(&vlan_gvrp_app); +} + +void vlan_gvrp_uninit(void) +{ + garp_unregister_application(&vlan_gvrp_app); +} diff --git a/net/8021q/vlan_netlink.c b/net/8021q/vlan_netlink.c index fd7cb195d53f..e9c91dcecc9b 100644 --- a/net/8021q/vlan_netlink.c +++ b/net/8021q/vlan_netlink.c @@ -59,7 +59,8 @@ static int vlan_validate(struct nlattr *tb[], struct nlattr *data[]) } if (data[IFLA_VLAN_FLAGS]) { flags = nla_data(data[IFLA_VLAN_FLAGS]); - if ((flags->flags & flags->mask) & ~VLAN_FLAG_REORDER_HDR) + if ((flags->flags & flags->mask) & + ~(VLAN_FLAG_REORDER_HDR | VLAN_FLAG_GVRP)) return -EINVAL; } -- cgit v1.2.2 From 6fe1c7a5556807e9d7154a2d2fb938d8a9e47e5f Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 23:21:31 -0700 Subject: net-sched: add dynamically sized qdisc class hash helpers Currently all qdiscs which allow to create classes uses a fixed sized hash table with size 16 to hash the classes. This causes a large bottleneck when using thousands of classes and unbound filters. Add helpers for dynamically sized class hashes to fix this. The following patches will convert the qdiscs to use them. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/sched/sch_api.c | 104 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 104 insertions(+) (limited to 'net') diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 10f01ad04380..e9ebc7af049e 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -316,6 +316,110 @@ void qdisc_watchdog_cancel(struct qdisc_watchdog *wd) } EXPORT_SYMBOL(qdisc_watchdog_cancel); +struct hlist_head *qdisc_class_hash_alloc(unsigned int n) +{ + unsigned int size = n * sizeof(struct hlist_head), i; + struct hlist_head *h; + + if (size <= PAGE_SIZE) + h = kmalloc(size, GFP_KERNEL); + else + h = (struct hlist_head *) + __get_free_pages(GFP_KERNEL, get_order(size)); + + if (h != NULL) { + for (i = 0; i < n; i++) + INIT_HLIST_HEAD(&h[i]); + } + return h; +} + +static void qdisc_class_hash_free(struct hlist_head *h, unsigned int n) +{ + unsigned int size = n * sizeof(struct hlist_head); + + if (size <= PAGE_SIZE) + kfree(h); + else + free_pages((unsigned long)h, get_order(size)); +} + +void qdisc_class_hash_grow(struct Qdisc *sch, struct Qdisc_class_hash *clhash) +{ + struct Qdisc_class_common *cl; + struct hlist_node *n, *next; + struct hlist_head *nhash, *ohash; + unsigned int nsize, nmask, osize; + unsigned int i, h; + + /* Rehash when load factor exceeds 0.75 */ + if (clhash->hashelems * 4 <= clhash->hashsize * 3) + return; + nsize = clhash->hashsize * 2; + nmask = nsize - 1; + nhash = qdisc_class_hash_alloc(nsize); + if (nhash == NULL) + return; + + ohash = clhash->hash; + osize = clhash->hashsize; + + sch_tree_lock(sch); + for (i = 0; i < osize; i++) { + hlist_for_each_entry_safe(cl, n, next, &ohash[i], hnode) { + h = qdisc_class_hash(cl->classid, nmask); + hlist_add_head(&cl->hnode, &nhash[h]); + } + } + clhash->hash = nhash; + clhash->hashsize = nsize; + clhash->hashmask = nmask; + sch_tree_unlock(sch); + + qdisc_class_hash_free(ohash, osize); +} +EXPORT_SYMBOL(qdisc_class_hash_grow); + +int qdisc_class_hash_init(struct Qdisc_class_hash *clhash) +{ + unsigned int size = 4; + + clhash->hash = qdisc_class_hash_alloc(size); + if (clhash->hash == NULL) + return -ENOMEM; + clhash->hashsize = size; + clhash->hashmask = size - 1; + clhash->hashelems = 0; + return 0; +} +EXPORT_SYMBOL(qdisc_class_hash_init); + +void qdisc_class_hash_destroy(struct Qdisc_class_hash *clhash) +{ + qdisc_class_hash_free(clhash->hash, clhash->hashsize); +} +EXPORT_SYMBOL(qdisc_class_hash_destroy); + +void qdisc_class_hash_insert(struct Qdisc_class_hash *clhash, + struct Qdisc_class_common *cl) +{ + unsigned int h; + + INIT_HLIST_NODE(&cl->hnode); + h = qdisc_class_hash(cl->classid, clhash->hashmask); + hlist_add_head(&cl->hnode, &clhash->hash[h]); + clhash->hashelems++; +} +EXPORT_SYMBOL(qdisc_class_hash_insert); + +void qdisc_class_hash_remove(struct Qdisc_class_hash *clhash, + struct Qdisc_class_common *cl) +{ + hlist_del(&cl->hnode); + clhash->hashelems--; +} +EXPORT_SYMBOL(qdisc_class_hash_remove); + /* Allocate an unique handle from space managed by kernel */ static u32 qdisc_alloc_handle(struct net_device *dev) -- cgit v1.2.2 From be0d39d52ce35554e856de7e9ea37ac1fa4a7f91 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 23:21:47 -0700 Subject: net-sched: sch_hfsc: use dynamic class hash helpers Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/sched/sch_hfsc.c | 81 ++++++++++++++++++++++++++-------------------------- 1 file changed, 40 insertions(+), 41 deletions(-) (limited to 'net') diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index e817aa00441d..3a8267246a4f 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -113,7 +113,7 @@ enum hfsc_class_flags struct hfsc_class { - u32 classid; /* class id */ + struct Qdisc_class_common cl_common; unsigned int refcnt; /* usage count */ struct gnet_stats_basic bstats; @@ -134,7 +134,6 @@ struct hfsc_class struct rb_node vt_node; /* parent's vt_tree member */ struct rb_root cf_tree; /* active children sorted by cl_f */ struct rb_node cf_node; /* parent's cf_heap member */ - struct list_head hlist; /* hash list member */ struct list_head dlist; /* drop list member */ u64 cl_total; /* total work in bytes */ @@ -177,13 +176,11 @@ struct hfsc_class unsigned long cl_nactive; /* number of active children */ }; -#define HFSC_HSIZE 16 - struct hfsc_sched { u16 defcls; /* default class id */ struct hfsc_class root; /* root class */ - struct list_head clhash[HFSC_HSIZE]; /* class hash */ + struct Qdisc_class_hash clhash; /* class hash */ struct rb_root eligible; /* eligible tree */ struct list_head droplist; /* active leaf class list (for dropping) */ @@ -933,26 +930,16 @@ hfsc_adjust_levels(struct hfsc_class *cl) } while ((cl = cl->cl_parent) != NULL); } -static inline unsigned int -hfsc_hash(u32 h) -{ - h ^= h >> 8; - h ^= h >> 4; - - return h & (HFSC_HSIZE - 1); -} - static inline struct hfsc_class * hfsc_find_class(u32 classid, struct Qdisc *sch) { struct hfsc_sched *q = qdisc_priv(sch); - struct hfsc_class *cl; + struct Qdisc_class_common *clc; - list_for_each_entry(cl, &q->clhash[hfsc_hash(classid)], hlist) { - if (cl->classid == classid) - return cl; - } - return NULL; + clc = qdisc_class_find(&q->clhash, classid); + if (clc == NULL) + return NULL; + return container_of(clc, struct hfsc_class, cl_common); } static void @@ -1032,7 +1019,8 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, if (cl != NULL) { if (parentid) { - if (cl->cl_parent && cl->cl_parent->classid != parentid) + if (cl->cl_parent && + cl->cl_parent->cl_common.classid != parentid) return -EINVAL; if (cl->cl_parent == NULL && parentid != TC_H_ROOT) return -EINVAL; @@ -1091,8 +1079,8 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, if (usc != NULL) hfsc_change_usc(cl, usc, 0); + cl->cl_common.classid = classid; cl->refcnt = 1; - cl->classid = classid; cl->sched = q; cl->cl_parent = parent; cl->qdisc = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, classid); @@ -1103,7 +1091,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, cl->cf_tree = RB_ROOT; sch_tree_lock(sch); - list_add_tail(&cl->hlist, &q->clhash[hfsc_hash(classid)]); + qdisc_class_hash_insert(&q->clhash, &cl->cl_common); list_add_tail(&cl->siblings, &parent->children); if (parent->level == 0) hfsc_purge_queue(sch, parent); @@ -1111,6 +1099,8 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, cl->cl_pcvtoff = parent->cl_cvtoff; sch_tree_unlock(sch); + qdisc_class_hash_grow(sch, &q->clhash); + if (tca[TCA_RATE]) gen_new_estimator(&cl->bstats, &cl->rate_est, &sch->dev->queue_lock, tca[TCA_RATE]); @@ -1145,7 +1135,7 @@ hfsc_delete_class(struct Qdisc *sch, unsigned long arg) hfsc_adjust_levels(cl->cl_parent); hfsc_purge_queue(sch, cl); - list_del(&cl->hlist); + qdisc_class_hash_remove(&q->clhash, &cl->cl_common); if (--cl->refcnt == 0) hfsc_destroy_class(sch, cl); @@ -1212,7 +1202,7 @@ hfsc_graft_class(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, return -EINVAL; if (new == NULL) { new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, - cl->classid); + cl->cl_common.classid); if (new == NULL) new = &noop_qdisc; } @@ -1345,8 +1335,9 @@ hfsc_dump_class(struct Qdisc *sch, unsigned long arg, struct sk_buff *skb, struct hfsc_class *cl = (struct hfsc_class *)arg; struct nlattr *nest; - tcm->tcm_parent = cl->cl_parent ? cl->cl_parent->classid : TC_H_ROOT; - tcm->tcm_handle = cl->classid; + tcm->tcm_parent = cl->cl_parent ? cl->cl_parent->cl_common.classid : + TC_H_ROOT; + tcm->tcm_handle = cl->cl_common.classid; if (cl->level == 0) tcm->tcm_info = cl->qdisc->handle; @@ -1390,14 +1381,16 @@ static void hfsc_walk(struct Qdisc *sch, struct qdisc_walker *arg) { struct hfsc_sched *q = qdisc_priv(sch); + struct hlist_node *n; struct hfsc_class *cl; unsigned int i; if (arg->stop) return; - for (i = 0; i < HFSC_HSIZE; i++) { - list_for_each_entry(cl, &q->clhash[i], hlist) { + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry(cl, n, &q->clhash.hash[i], + cl_common.hnode) { if (arg->count < arg->skip) { arg->count++; continue; @@ -1433,21 +1426,22 @@ hfsc_init_qdisc(struct Qdisc *sch, struct nlattr *opt) { struct hfsc_sched *q = qdisc_priv(sch); struct tc_hfsc_qopt *qopt; - unsigned int i; + int err; if (opt == NULL || nla_len(opt) < sizeof(*qopt)) return -EINVAL; qopt = nla_data(opt); q->defcls = qopt->defcls; - for (i = 0; i < HFSC_HSIZE; i++) - INIT_LIST_HEAD(&q->clhash[i]); + err = qdisc_class_hash_init(&q->clhash); + if (err < 0) + return err; q->eligible = RB_ROOT; INIT_LIST_HEAD(&q->droplist); skb_queue_head_init(&q->requeue); + q->root.cl_common.classid = sch->handle; q->root.refcnt = 1; - q->root.classid = sch->handle; q->root.sched = q; q->root.qdisc = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, sch->handle); @@ -1457,7 +1451,8 @@ hfsc_init_qdisc(struct Qdisc *sch, struct nlattr *opt) q->root.vt_tree = RB_ROOT; q->root.cf_tree = RB_ROOT; - list_add(&q->root.hlist, &q->clhash[hfsc_hash(q->root.classid)]); + qdisc_class_hash_insert(&q->clhash, &q->root.cl_common); + qdisc_class_hash_grow(sch, &q->clhash); qdisc_watchdog_init(&q->watchdog, sch); @@ -1520,10 +1515,11 @@ hfsc_reset_qdisc(struct Qdisc *sch) { struct hfsc_sched *q = qdisc_priv(sch); struct hfsc_class *cl; + struct hlist_node *n; unsigned int i; - for (i = 0; i < HFSC_HSIZE; i++) { - list_for_each_entry(cl, &q->clhash[i], hlist) + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry(cl, n, &q->clhash.hash[i], cl_common.hnode) hfsc_reset_class(cl); } __skb_queue_purge(&q->requeue); @@ -1537,17 +1533,20 @@ static void hfsc_destroy_qdisc(struct Qdisc *sch) { struct hfsc_sched *q = qdisc_priv(sch); - struct hfsc_class *cl, *next; + struct hlist_node *n, *next; + struct hfsc_class *cl; unsigned int i; - for (i = 0; i < HFSC_HSIZE; i++) { - list_for_each_entry(cl, &q->clhash[i], hlist) + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry(cl, n, &q->clhash.hash[i], cl_common.hnode) tcf_destroy_chain(&cl->filter_list); } - for (i = 0; i < HFSC_HSIZE; i++) { - list_for_each_entry_safe(cl, next, &q->clhash[i], hlist) + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry_safe(cl, n, next, &q->clhash.hash[i], + cl_common.hnode) hfsc_destroy_class(sch, cl); } + qdisc_class_hash_destroy(&q->clhash); __skb_queue_purge(&q->requeue); qdisc_watchdog_cancel(&q->watchdog); } -- cgit v1.2.2 From d77fea2eb9206833c7aa1b013044ddeb5225b92c Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 23:22:05 -0700 Subject: net-sched: sch_cbq: use dynamic class hash helpers Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/sched/sch_cbq.c | 112 ++++++++++++++++++++++++++-------------------------- 1 file changed, 55 insertions(+), 57 deletions(-) (limited to 'net') diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 2a3c97f7dc63..968b4c73c9c1 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -73,11 +73,10 @@ struct cbq_sched_data; struct cbq_class { - struct cbq_class *next; /* hash table link */ + struct Qdisc_class_common common; struct cbq_class *next_alive; /* next class with backlog in this priority band */ /* Parameters */ - u32 classid; unsigned char priority; /* class priority */ unsigned char priority2; /* priority to be used after overlimit */ unsigned char ewma_log; /* time constant for idle time calculation */ @@ -144,7 +143,7 @@ struct cbq_class struct cbq_sched_data { - struct cbq_class *classes[16]; /* Hash table of all classes */ + struct Qdisc_class_hash clhash; /* Hash table of all classes */ int nclasses[TC_CBQ_MAXPRIO+1]; unsigned quanta[TC_CBQ_MAXPRIO+1]; @@ -177,23 +176,15 @@ struct cbq_sched_data #define L2T(cl,len) qdisc_l2t((cl)->R_tab,len) - -static __inline__ unsigned cbq_hash(u32 h) -{ - h ^= h>>8; - h ^= h>>4; - return h&0xF; -} - static __inline__ struct cbq_class * cbq_class_lookup(struct cbq_sched_data *q, u32 classid) { - struct cbq_class *cl; + struct Qdisc_class_common *clc; - for (cl = q->classes[cbq_hash(classid)]; cl; cl = cl->next) - if (cl->classid == classid) - return cl; - return NULL; + clc = qdisc_class_find(&q->clhash, classid); + if (clc == NULL) + return NULL; + return container_of(clc, struct cbq_class, common); } #ifdef CONFIG_NET_CLS_ACT @@ -1071,13 +1062,14 @@ static void cbq_adjust_levels(struct cbq_class *this) static void cbq_normalize_quanta(struct cbq_sched_data *q, int prio) { struct cbq_class *cl; - unsigned h; + struct hlist_node *n; + unsigned int h; if (q->quanta[prio] == 0) return; - for (h=0; h<16; h++) { - for (cl = q->classes[h]; cl; cl = cl->next) { + for (h = 0; h < q->clhash.hashsize; h++) { + hlist_for_each_entry(cl, n, &q->clhash.hash[h], common.hnode) { /* BUGGGG... Beware! This expression suffer of arithmetic overflows! */ @@ -1086,7 +1078,7 @@ static void cbq_normalize_quanta(struct cbq_sched_data *q, int prio) q->quanta[prio]; } if (cl->quantum <= 0 || cl->quantum>32*cl->qdisc->dev->mtu) { - printk(KERN_WARNING "CBQ: class %08x has bad quantum==%ld, repaired.\n", cl->classid, cl->quantum); + printk(KERN_WARNING "CBQ: class %08x has bad quantum==%ld, repaired.\n", cl->common.classid, cl->quantum); cl->quantum = cl->qdisc->dev->mtu/2 + 1; } } @@ -1114,10 +1106,12 @@ static void cbq_sync_defmap(struct cbq_class *cl) if (split->defaults[i]) continue; - for (h=0; h<16; h++) { + for (h = 0; h < q->clhash.hashsize; h++) { + struct hlist_node *n; struct cbq_class *c; - for (c = q->classes[h]; c; c = c->next) { + hlist_for_each_entry(c, n, &q->clhash.hash[h], + common.hnode) { if (c->split == split && c->level < level && c->defmap&(1<defaults[i] = c; @@ -1135,12 +1129,12 @@ static void cbq_change_defmap(struct cbq_class *cl, u32 splitid, u32 def, u32 ma if (splitid == 0) { if ((split = cl->split) == NULL) return; - splitid = split->classid; + splitid = split->common.classid; } - if (split == NULL || split->classid != splitid) { + if (split == NULL || split->common.classid != splitid) { for (split = cl->tparent; split; split = split->tparent) - if (split->classid == splitid) + if (split->common.classid == splitid) break; } @@ -1163,13 +1157,7 @@ static void cbq_unlink_class(struct cbq_class *this) struct cbq_class *cl, **clp; struct cbq_sched_data *q = qdisc_priv(this->qdisc); - for (clp = &q->classes[cbq_hash(this->classid)]; (cl = *clp) != NULL; clp = &cl->next) { - if (cl == this) { - *clp = cl->next; - cl->next = NULL; - break; - } - } + qdisc_class_hash_remove(&q->clhash, &this->common); if (this->tparent) { clp=&this->sibling; @@ -1195,12 +1183,10 @@ static void cbq_unlink_class(struct cbq_class *this) static void cbq_link_class(struct cbq_class *this) { struct cbq_sched_data *q = qdisc_priv(this->qdisc); - unsigned h = cbq_hash(this->classid); struct cbq_class *parent = this->tparent; this->sibling = this; - this->next = q->classes[h]; - q->classes[h] = this; + qdisc_class_hash_insert(&q->clhash, &this->common); if (parent == NULL) return; @@ -1242,6 +1228,7 @@ cbq_reset(struct Qdisc* sch) { struct cbq_sched_data *q = qdisc_priv(sch); struct cbq_class *cl; + struct hlist_node *n; int prio; unsigned h; @@ -1258,8 +1245,8 @@ cbq_reset(struct Qdisc* sch) for (prio = 0; prio <= TC_CBQ_MAXPRIO; prio++) q->active[prio] = NULL; - for (h = 0; h < 16; h++) { - for (cl = q->classes[h]; cl; cl = cl->next) { + for (h = 0; h < q->clhash.hashsize; h++) { + hlist_for_each_entry(cl, n, &q->clhash.hash[h], common.hnode) { qdisc_reset(cl->q); cl->next_alive = NULL; @@ -1406,9 +1393,13 @@ static int cbq_init(struct Qdisc *sch, struct nlattr *opt) if ((q->link.R_tab = qdisc_get_rtab(r, tb[TCA_CBQ_RTAB])) == NULL) return -EINVAL; + err = qdisc_class_hash_init(&q->clhash); + if (err < 0) + goto put_rtab; + q->link.refcnt = 1; q->link.sibling = &q->link; - q->link.classid = sch->handle; + q->link.common.classid = sch->handle; q->link.qdisc = sch; if (!(q->link.q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, sch->handle))) @@ -1441,6 +1432,10 @@ static int cbq_init(struct Qdisc *sch, struct nlattr *opt) cbq_addprio(q, &q->link); return 0; + +put_rtab: + qdisc_put_rtab(q->link.R_tab); + return err; } static __inline__ int cbq_dump_rate(struct sk_buff *skb, struct cbq_class *cl) @@ -1521,7 +1516,7 @@ static __inline__ int cbq_dump_fopt(struct sk_buff *skb, struct cbq_class *cl) struct tc_cbq_fopt opt; if (cl->split || cl->defmap) { - opt.split = cl->split ? cl->split->classid : 0; + opt.split = cl->split ? cl->split->common.classid : 0; opt.defmap = cl->defmap; opt.defchange = ~0; NLA_PUT(skb, TCA_CBQ_FOPT, sizeof(opt), &opt); @@ -1602,10 +1597,10 @@ cbq_dump_class(struct Qdisc *sch, unsigned long arg, struct nlattr *nest; if (cl->tparent) - tcm->tcm_parent = cl->tparent->classid; + tcm->tcm_parent = cl->tparent->common.classid; else tcm->tcm_parent = TC_H_ROOT; - tcm->tcm_handle = cl->classid; + tcm->tcm_handle = cl->common.classid; tcm->tcm_info = cl->q->handle; nest = nla_nest_start(skb, TCA_OPTIONS); @@ -1650,8 +1645,9 @@ static int cbq_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, if (cl) { if (new == NULL) { - if ((new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, - cl->classid)) == NULL) + new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + cl->common.classid); + if (new == NULL) return -ENOBUFS; } else { #ifdef CONFIG_NET_CLS_ACT @@ -1716,6 +1712,7 @@ static void cbq_destroy(struct Qdisc* sch) { struct cbq_sched_data *q = qdisc_priv(sch); + struct hlist_node *n, *next; struct cbq_class *cl; unsigned h; @@ -1727,18 +1724,16 @@ cbq_destroy(struct Qdisc* sch) * classes from root to leafs which means that filters can still * be bound to classes which have been destroyed already. --TGR '04 */ - for (h = 0; h < 16; h++) { - for (cl = q->classes[h]; cl; cl = cl->next) + for (h = 0; h < q->clhash.hashsize; h++) { + hlist_for_each_entry(cl, n, &q->clhash.hash[h], common.hnode) tcf_destroy_chain(&cl->filter_list); } - for (h = 0; h < 16; h++) { - struct cbq_class *next; - - for (cl = q->classes[h]; cl; cl = next) { - next = cl->next; + for (h = 0; h < q->clhash.hashsize; h++) { + hlist_for_each_entry_safe(cl, n, next, &q->clhash.hash[h], + common.hnode) cbq_destroy_class(sch, cl); - } } + qdisc_class_hash_destroy(&q->clhash); } static void cbq_put(struct Qdisc *sch, unsigned long arg) @@ -1781,7 +1776,8 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t if (cl) { /* Check parent */ if (parentid) { - if (cl->tparent && cl->tparent->classid != parentid) + if (cl->tparent && + cl->tparent->common.classid != parentid) return -EINVAL; if (!cl->tparent && parentid != TC_H_ROOT) return -EINVAL; @@ -1883,7 +1879,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t cl->refcnt = 1; if (!(cl->q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, classid))) cl->q = &noop_qdisc; - cl->classid = classid; + cl->common.classid = classid; cl->tparent = parent; cl->qdisc = sch; cl->allot = parent->allot; @@ -1916,6 +1912,8 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t cbq_set_fopt(cl, nla_data(tb[TCA_CBQ_FOPT])); sch_tree_unlock(sch); + qdisc_class_hash_grow(sch, &q->clhash); + if (tca[TCA_RATE]) gen_new_estimator(&cl->bstats, &cl->rate_est, &sch->dev->queue_lock, tca[TCA_RATE]); @@ -2008,15 +2006,15 @@ static void cbq_unbind_filter(struct Qdisc *sch, unsigned long arg) static void cbq_walk(struct Qdisc *sch, struct qdisc_walker *arg) { struct cbq_sched_data *q = qdisc_priv(sch); + struct cbq_class *cl; + struct hlist_node *n; unsigned h; if (arg->stop) return; - for (h = 0; h < 16; h++) { - struct cbq_class *cl; - - for (cl = q->classes[h]; cl; cl = cl->next) { + for (h = 0; h < q->clhash.hashsize; h++) { + hlist_for_each_entry(cl, n, &q->clhash.hash[h], common.hnode) { if (arg->count < arg->skip) { arg->count++; continue; -- cgit v1.2.2 From fbd8f1379aeeb3e44a59302a6b2850636130bb2a Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 23:22:19 -0700 Subject: net-sched: sch_htb: move hash and sibling list removal to htb_delete Hash list removal currently happens twice (once in htb_delete, once in htb_destroy_class), which makes it harder to use the dynamically sized class hash without adding special cases for HTB. The reason is that qdisc destruction destroys classes in hierarchical order, which is not necessary if filters are destroyed in a separate iteration during qdisc destruction. Adjust qdisc destruction to follow the same scheme as other hierarchical qdiscs by first performing a filter destruction pass, then destroying all classes in hash order. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/sched/sch_htb.c | 40 +++++++++++++++++----------------------- 1 file changed, 17 insertions(+), 23 deletions(-) (limited to 'net') diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 0284791169c9..d01fe3a1a623 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -1226,8 +1226,6 @@ static void htb_parent_to_leaf(struct htb_sched *q, struct htb_class *cl, static void htb_destroy_class(struct Qdisc *sch, struct htb_class *cl) { - struct htb_sched *q = qdisc_priv(sch); - if (!cl->level) { BUG_TRAP(cl->un.leaf.q); qdisc_destroy(cl->un.leaf.q); @@ -1237,21 +1235,6 @@ static void htb_destroy_class(struct Qdisc *sch, struct htb_class *cl) qdisc_put_rtab(cl->ceil); tcf_destroy_chain(&cl->filter_list); - - while (!list_empty(&cl->children)) - htb_destroy_class(sch, list_entry(cl->children.next, - struct htb_class, sibling)); - - /* note: this delete may happen twice (see htb_delete) */ - hlist_del_init(&cl->hlist); - list_del(&cl->sibling); - - if (cl->prio_activity) - htb_deactivate(q, cl); - - if (cl->cmode != HTB_CAN_SEND) - htb_safe_rb_erase(&cl->pq_node, q->wait_pq + cl->level); - kfree(cl); } @@ -1259,6 +1242,9 @@ static void htb_destroy_class(struct Qdisc *sch, struct htb_class *cl) static void htb_destroy(struct Qdisc *sch) { struct htb_sched *q = qdisc_priv(sch); + struct hlist_node *n, *next; + struct htb_class *cl; + unsigned int i; qdisc_watchdog_cancel(&q->watchdog); /* This line used to be after htb_destroy_class call below @@ -1267,10 +1253,14 @@ static void htb_destroy(struct Qdisc *sch) unbind_filter on it (without Oops). */ tcf_destroy_chain(&q->filter_list); - while (!list_empty(&q->root)) - htb_destroy_class(sch, list_entry(q->root.next, - struct htb_class, sibling)); - + for (i = 0; i < HTB_HSIZE; i++) { + hlist_for_each_entry(cl, n, q->hash + i, hlist) + tcf_destroy_chain(&cl->filter_list); + } + for (i = 0; i < HTB_HSIZE; i++) { + hlist_for_each_entry_safe(cl, n, next, q->hash + i, hlist) + htb_destroy_class(sch, cl); + } __skb_queue_purge(&q->direct_queue); } @@ -1302,12 +1292,16 @@ static int htb_delete(struct Qdisc *sch, unsigned long arg) qdisc_tree_decrease_qlen(cl->un.leaf.q, qlen); } - /* delete from hash and active; remainder in destroy_class */ - hlist_del_init(&cl->hlist); + /* delete from hash, sibling list and active */ + hlist_del(&cl->hlist); + list_del(&cl->sibling); if (cl->prio_activity) htb_deactivate(q, cl); + if (cl->cmode != HTB_CAN_SEND) + htb_safe_rb_erase(&cl->pq_node, q->wait_pq + cl->level); + if (last_child) htb_parent_to_leaf(q, cl, new_q); -- cgit v1.2.2 From f4c1f3e0c59be0e6566d9c00b1d8b204ffb861a2 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 23:22:35 -0700 Subject: net-sched: sch_htb: use dynamic class hash helpers Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/sched/sch_htb.c | 100 ++++++++++++++++++++++------------------------------ 1 file changed, 42 insertions(+), 58 deletions(-) (limited to 'net') diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index d01fe3a1a623..e7461a17d71e 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -51,7 +51,6 @@ one less than their parent. */ -#define HTB_HSIZE 16 /* classid hash size */ static int htb_hysteresis __read_mostly = 0; /* whether to use mode hysteresis for speedup */ #define HTB_VER 0x30011 /* major must be matched with number suplied by TC as version */ @@ -72,8 +71,8 @@ enum htb_cmode { /* interior & leaf nodes; props specific to leaves are marked L: */ struct htb_class { + struct Qdisc_class_common common; /* general class parameters */ - u32 classid; struct gnet_stats_basic bstats; struct gnet_stats_queue qstats; struct gnet_stats_rate_est rate_est; @@ -83,7 +82,6 @@ struct htb_class { /* topology */ int level; /* our level (see above) */ struct htb_class *parent; /* parent class */ - struct hlist_node hlist; /* classid hash list item */ struct list_head sibling; /* sibling list item */ struct list_head children; /* children list */ @@ -141,7 +139,7 @@ static inline long L2T(struct htb_class *cl, struct qdisc_rate_table *rate, struct htb_sched { struct list_head root; /* root classes list */ - struct hlist_head hash[HTB_HSIZE]; /* hashed by classid */ + struct Qdisc_class_hash clhash; struct list_head drops[TC_HTB_NUMPRIO];/* active leaves (for drops) */ /* self list - roots of self generating tree */ @@ -176,32 +174,16 @@ struct htb_sched { long direct_pkts; }; -/* compute hash of size HTB_HSIZE for given handle */ -static inline int htb_hash(u32 h) -{ -#if HTB_HSIZE != 16 -#error "Declare new hash for your HTB_HSIZE" -#endif - h ^= h >> 8; /* stolen from cbq_hash */ - h ^= h >> 4; - return h & 0xf; -} - /* find class in global hash table using given handle */ static inline struct htb_class *htb_find(u32 handle, struct Qdisc *sch) { struct htb_sched *q = qdisc_priv(sch); - struct hlist_node *p; - struct htb_class *cl; + struct Qdisc_class_common *clc; - if (TC_H_MAJ(handle) != sch->handle) + clc = qdisc_class_find(&q->clhash, handle); + if (clc == NULL) return NULL; - - hlist_for_each_entry(cl, p, q->hash + htb_hash(handle), hlist) { - if (cl->classid == handle) - return cl; - } - return NULL; + return container_of(clc, struct htb_class, common); } /** @@ -282,7 +264,7 @@ static void htb_add_to_id_tree(struct rb_root *root, parent = *p; c = rb_entry(parent, struct htb_class, node[prio]); - if (cl->classid > c->classid) + if (cl->common.classid > c->common.classid) p = &parent->rb_right; else p = &parent->rb_left; @@ -446,7 +428,7 @@ static void htb_deactivate_prios(struct htb_sched *q, struct htb_class *cl) /* we are removing child which is pointed to from parent feed - forget the pointer but remember classid */ - p->un.inner.last_ptr_id[prio] = cl->classid; + p->un.inner.last_ptr_id[prio] = cl->common.classid; p->un.inner.ptr[prio] = NULL; } @@ -751,10 +733,10 @@ static struct rb_node *htb_id_find_next_upper(int prio, struct rb_node *n, while (n) { struct htb_class *cl = rb_entry(n, struct htb_class, node[prio]); - if (id == cl->classid) + if (id == cl->common.classid) return n; - if (id > cl->classid) { + if (id > cl->common.classid) { n = n->rb_right; } else { r = n; @@ -864,7 +846,7 @@ next: if (!cl->warned) { printk(KERN_WARNING "htb: class %X isn't work conserving ?!\n", - cl->classid); + cl->common.classid); cl->warned = 1; } q->nwc_hit++; @@ -975,13 +957,12 @@ static unsigned int htb_drop(struct Qdisc *sch) static void htb_reset(struct Qdisc *sch) { struct htb_sched *q = qdisc_priv(sch); - int i; - - for (i = 0; i < HTB_HSIZE; i++) { - struct hlist_node *p; - struct htb_class *cl; + struct htb_class *cl; + struct hlist_node *n; + unsigned int i; - hlist_for_each_entry(cl, p, q->hash + i, hlist) { + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry(cl, n, &q->clhash.hash[i], common.hnode) { if (cl->level) memset(&cl->un.inner, 0, sizeof(cl->un.inner)); else { @@ -1040,8 +1021,9 @@ static int htb_init(struct Qdisc *sch, struct nlattr *opt) } INIT_LIST_HEAD(&q->root); - for (i = 0; i < HTB_HSIZE; i++) - INIT_HLIST_HEAD(q->hash + i); + err = qdisc_class_hash_init(&q->clhash); + if (err < 0) + return err; for (i = 0; i < TC_HTB_NUMPRIO; i++) INIT_LIST_HEAD(q->drops + i); @@ -1096,8 +1078,8 @@ static int htb_dump_class(struct Qdisc *sch, unsigned long arg, struct tc_htb_opt opt; spin_lock_bh(&sch->dev->queue_lock); - tcm->tcm_parent = cl->parent ? cl->parent->classid : TC_H_ROOT; - tcm->tcm_handle = cl->classid; + tcm->tcm_parent = cl->parent ? cl->parent->common.classid : TC_H_ROOT; + tcm->tcm_handle = cl->common.classid; if (!cl->level && cl->un.leaf.q) tcm->tcm_info = cl->un.leaf.q->handle; @@ -1152,7 +1134,7 @@ static int htb_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, if (cl && !cl->level) { if (new == NULL && (new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, - cl->classid)) + cl->common.classid)) == NULL) return -ENOBUFS; sch_tree_lock(sch); @@ -1253,14 +1235,16 @@ static void htb_destroy(struct Qdisc *sch) unbind_filter on it (without Oops). */ tcf_destroy_chain(&q->filter_list); - for (i = 0; i < HTB_HSIZE; i++) { - hlist_for_each_entry(cl, n, q->hash + i, hlist) + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry(cl, n, &q->clhash.hash[i], common.hnode) tcf_destroy_chain(&cl->filter_list); } - for (i = 0; i < HTB_HSIZE; i++) { - hlist_for_each_entry_safe(cl, n, next, q->hash + i, hlist) + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry_safe(cl, n, next, &q->clhash.hash[i], + common.hnode) htb_destroy_class(sch, cl); } + qdisc_class_hash_destroy(&q->clhash); __skb_queue_purge(&q->direct_queue); } @@ -1280,7 +1264,7 @@ static int htb_delete(struct Qdisc *sch, unsigned long arg) if (!cl->level && htb_parent_last_child(cl)) { new_q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, - cl->parent->classid); + cl->parent->common.classid); last_child = 1; } @@ -1292,8 +1276,8 @@ static int htb_delete(struct Qdisc *sch, unsigned long arg) qdisc_tree_decrease_qlen(cl->un.leaf.q, qlen); } - /* delete from hash, sibling list and active */ - hlist_del(&cl->hlist); + /* delete from hash and active; remainder in destroy_class */ + qdisc_class_hash_remove(&q->clhash, &cl->common); list_del(&cl->sibling); if (cl->prio_activity) @@ -1390,7 +1374,6 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, tca[TCA_RATE] ? : &est.nla); cl->refcnt = 1; INIT_LIST_HEAD(&cl->sibling); - INIT_HLIST_NODE(&cl->hlist); INIT_LIST_HEAD(&cl->children); INIT_LIST_HEAD(&cl->un.leaf.drop_list); RB_CLEAR_NODE(&cl->pq_node); @@ -1425,7 +1408,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, /* leaf (we) needs elementary qdisc */ cl->un.leaf.q = new_q ? new_q : &noop_qdisc; - cl->classid = classid; + cl->common.classid = classid; cl->parent = parent; /* set class to be in HTB_CAN_SEND state */ @@ -1436,7 +1419,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, cl->cmode = HTB_CAN_SEND; /* attach to the hash list and parent's family */ - hlist_add_head(&cl->hlist, q->hash + htb_hash(classid)); + qdisc_class_hash_insert(&q->clhash, &cl->common); list_add_tail(&cl->sibling, parent ? &parent->children : &q->root); } else { @@ -1454,13 +1437,13 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, if (!hopt->quantum && cl->un.leaf.quantum < 1000) { printk(KERN_WARNING "HTB: quantum of class %X is small. Consider r2q change.\n", - cl->classid); + cl->common.classid); cl->un.leaf.quantum = 1000; } if (!hopt->quantum && cl->un.leaf.quantum > 200000) { printk(KERN_WARNING "HTB: quantum of class %X is big. Consider r2q change.\n", - cl->classid); + cl->common.classid); cl->un.leaf.quantum = 200000; } if (hopt->quantum) @@ -1483,6 +1466,8 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, cl->ceil = ctab; sch_tree_unlock(sch); + qdisc_class_hash_grow(sch, &q->clhash); + *arg = (unsigned long)cl; return 0; @@ -1539,16 +1524,15 @@ static void htb_unbind_filter(struct Qdisc *sch, unsigned long arg) static void htb_walk(struct Qdisc *sch, struct qdisc_walker *arg) { struct htb_sched *q = qdisc_priv(sch); - int i; + struct htb_class *cl; + struct hlist_node *n; + unsigned int i; if (arg->stop) return; - for (i = 0; i < HTB_HSIZE; i++) { - struct hlist_node *p; - struct htb_class *cl; - - hlist_for_each_entry(cl, p, q->hash + i, hlist) { + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry(cl, n, &q->clhash.hash[i], common.hnode) { if (arg->count < arg->skip) { arg->count++; continue; -- cgit v1.2.2 From 4207759939c208f24d7e09596753e67b6bc188f9 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 23:22:53 -0700 Subject: net-sched: sch_htb: remove child and sibling lists Now that the qdisc isn't destroyed in hierarchical order anymore, the only user of the child lists left is htb_parent_last_child(). This can be easily changed to use a counter of children to save a few bytes. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/sched/sch_htb.c | 21 +++++++-------------- 1 file changed, 7 insertions(+), 14 deletions(-) (limited to 'net') diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index e7461a17d71e..128a5ab2e376 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -81,9 +81,8 @@ struct htb_class { /* topology */ int level; /* our level (see above) */ + unsigned int children; struct htb_class *parent; /* parent class */ - struct list_head sibling; /* sibling list item */ - struct list_head children; /* children list */ union { struct htb_class_leaf { @@ -138,7 +137,6 @@ static inline long L2T(struct htb_class *cl, struct qdisc_rate_table *rate, } struct htb_sched { - struct list_head root; /* root classes list */ struct Qdisc_class_hash clhash; struct list_head drops[TC_HTB_NUMPRIO];/* active leaves (for drops) */ @@ -1020,7 +1018,6 @@ static int htb_init(struct Qdisc *sch, struct nlattr *opt) return -EINVAL; } - INIT_LIST_HEAD(&q->root); err = qdisc_class_hash_init(&q->clhash); if (err < 0) return err; @@ -1175,12 +1172,9 @@ static inline int htb_parent_last_child(struct htb_class *cl) if (!cl->parent) /* the root class */ return 0; - - if (!(cl->parent->children.next == &cl->sibling && - cl->parent->children.prev == &cl->sibling)) + if (cl->parent->children > 1) /* not the last child */ return 0; - return 1; } @@ -1259,7 +1253,7 @@ static int htb_delete(struct Qdisc *sch, unsigned long arg) // TODO: why don't allow to delete subtree ? references ? does // tc subsys quarantee us that in htb_destroy it holds no class // refs so that we can remove children safely there ? - if (!list_empty(&cl->children) || cl->filter_cnt) + if (cl->children || cl->filter_cnt) return -EBUSY; if (!cl->level && htb_parent_last_child(cl)) { @@ -1278,7 +1272,7 @@ static int htb_delete(struct Qdisc *sch, unsigned long arg) /* delete from hash and active; remainder in destroy_class */ qdisc_class_hash_remove(&q->clhash, &cl->common); - list_del(&cl->sibling); + cl->parent->children--; if (cl->prio_activity) htb_deactivate(q, cl); @@ -1373,8 +1367,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, &sch->dev->queue_lock, tca[TCA_RATE] ? : &est.nla); cl->refcnt = 1; - INIT_LIST_HEAD(&cl->sibling); - INIT_LIST_HEAD(&cl->children); + cl->children = 0; INIT_LIST_HEAD(&cl->un.leaf.drop_list); RB_CLEAR_NODE(&cl->pq_node); @@ -1420,8 +1413,8 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, /* attach to the hash list and parent's family */ qdisc_class_hash_insert(&q->clhash, &cl->common); - list_add_tail(&cl->sibling, - parent ? &parent->children : &q->root); + if (parent) + parent->children++; } else { if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, -- cgit v1.2.2 From aee18a8cf28808b7302ef698d77fa73883e60f1b Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 23:23:27 -0700 Subject: net-sched: sch_htb: remove write-only qdisc filter_cnt The filter_cnt is supposed to count filter references to a class. Since the qdisc can't be the target of a filter, it doesn't need a filter_cnt. In fact the counter is never decreased since cls_api considers a return value of zero a failure and doesn't unbind again. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/sched/sch_htb.c | 7 ------- 1 file changed, 7 deletions(-) (limited to 'net') diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 128a5ab2e376..ee8b4ffe110c 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -159,7 +159,6 @@ struct htb_sched { /* filters for qdisc itself */ struct tcf_proto *filter_list; - int filter_cnt; int rate2quantum; /* quant = rate / rate2quantum */ psched_time_t now; /* cached dequeue time */ @@ -1484,7 +1483,6 @@ static struct tcf_proto **htb_find_tcf(struct Qdisc *sch, unsigned long arg) static unsigned long htb_bind_filter(struct Qdisc *sch, unsigned long parent, u32 classid) { - struct htb_sched *q = qdisc_priv(sch); struct htb_class *cl = htb_find(classid, sch); /*if (cl && !cl->level) return 0; @@ -1498,20 +1496,15 @@ static unsigned long htb_bind_filter(struct Qdisc *sch, unsigned long parent, */ if (cl) cl->filter_cnt++; - else - q->filter_cnt++; return (unsigned long)cl; } static void htb_unbind_filter(struct Qdisc *sch, unsigned long arg) { - struct htb_sched *q = qdisc_priv(sch); struct htb_class *cl = (struct htb_class *)arg; if (cl) cl->filter_cnt--; - else - q->filter_cnt--; } static void htb_walk(struct Qdisc *sch, struct qdisc_walker *arg) -- cgit v1.2.2 From fb0305ce1b03f6ff17f84f2c63daccecb45f2805 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sat, 5 Jul 2008 23:40:21 -0700 Subject: net-sched: consolidate default fifo qdisc setup Signed-off-by: Patrick McHardy Acked-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/sched/sch_fifo.c | 42 ++++++++++++++++++++++++++++++++++++++++++ net/sched/sch_netem.c | 24 +----------------------- net/sched/sch_red.c | 33 +++------------------------------ net/sched/sch_tbf.c | 33 ++++----------------------------- 4 files changed, 50 insertions(+), 82 deletions(-) (limited to 'net') diff --git a/net/sched/sch_fifo.c b/net/sched/sch_fifo.c index 95ed48221652..82d7d7bbbb16 100644 --- a/net/sched/sch_fifo.c +++ b/net/sched/sch_fifo.c @@ -107,3 +107,45 @@ struct Qdisc_ops bfifo_qdisc_ops __read_mostly = { .owner = THIS_MODULE, }; EXPORT_SYMBOL(bfifo_qdisc_ops); + +/* Pass size change message down to embedded FIFO */ +int fifo_set_limit(struct Qdisc *q, unsigned int limit) +{ + struct nlattr *nla; + int ret = -ENOMEM; + + /* Hack to avoid sending change message to non-FIFO */ + if (strncmp(q->ops->id + 1, "fifo", 4) != 0) + return 0; + + nla = kmalloc(nla_attr_size(sizeof(struct tc_fifo_qopt)), GFP_KERNEL); + if (nla) { + nla->nla_type = RTM_NEWQDISC; + nla->nla_len = nla_attr_size(sizeof(struct tc_fifo_qopt)); + ((struct tc_fifo_qopt *)nla_data(nla))->limit = limit; + + ret = q->ops->change(q, nla); + kfree(nla); + } + return ret; +} +EXPORT_SYMBOL(fifo_set_limit); + +struct Qdisc *fifo_create_dflt(struct Qdisc *sch, struct Qdisc_ops *ops, + unsigned int limit) +{ + struct Qdisc *q; + int err = -ENOMEM; + + q = qdisc_create_dflt(sch->dev, ops, TC_H_MAKE(sch->handle, 1)); + if (q) { + err = fifo_set_limit(q, limit); + if (err < 0) { + qdisc_destroy(q); + q = NULL; + } + } + + return q ? : ERR_PTR(err); +} +EXPORT_SYMBOL(fifo_create_dflt); diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index c9c649b26eaa..24697667247c 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -310,28 +310,6 @@ static void netem_reset(struct Qdisc *sch) qdisc_watchdog_cancel(&q->watchdog); } -/* Pass size change message down to embedded FIFO */ -static int set_fifo_limit(struct Qdisc *q, int limit) -{ - struct nlattr *nla; - int ret = -ENOMEM; - - /* Hack to avoid sending change message to non-FIFO */ - if (strncmp(q->ops->id + 1, "fifo", 4) != 0) - return 0; - - nla = kmalloc(nla_attr_size(sizeof(struct tc_fifo_qopt)), GFP_KERNEL); - if (nla) { - nla->nla_type = RTM_NEWQDISC; - nla->nla_len = nla_attr_size(sizeof(struct tc_fifo_qopt)); - ((struct tc_fifo_qopt *)nla_data(nla))->limit = limit; - - ret = q->ops->change(q, nla); - kfree(nla); - } - return ret; -} - /* * Distribution data is a variable size payload containing * signed 16 bit values. @@ -416,7 +394,7 @@ static int netem_change(struct Qdisc *sch, struct nlattr *opt) if (ret < 0) return ret; - ret = set_fifo_limit(q->qdisc, qopt->limit); + ret = fifo_set_limit(q->qdisc, qopt->limit); if (ret) { pr_debug("netem: can't set fifo limit\n"); return ret; diff --git a/net/sched/sch_red.c b/net/sched/sch_red.c index 5c569853b9c0..77098acf0adc 100644 --- a/net/sched/sch_red.c +++ b/net/sched/sch_red.c @@ -174,33 +174,6 @@ static void red_destroy(struct Qdisc *sch) qdisc_destroy(q->qdisc); } -static struct Qdisc *red_create_dflt(struct Qdisc *sch, u32 limit) -{ - struct Qdisc *q; - struct nlattr *nla; - int ret; - - q = qdisc_create_dflt(sch->dev, &bfifo_qdisc_ops, - TC_H_MAKE(sch->handle, 1)); - if (q) { - nla = kmalloc(nla_attr_size(sizeof(struct tc_fifo_qopt)), - GFP_KERNEL); - if (nla) { - nla->nla_type = RTM_NEWQDISC; - nla->nla_len = nla_attr_size(sizeof(struct tc_fifo_qopt)); - ((struct tc_fifo_qopt *)nla_data(nla))->limit = limit; - - ret = q->ops->change(q, nla); - kfree(nla); - - if (ret == 0) - return q; - } - qdisc_destroy(q); - } - return NULL; -} - static const struct nla_policy red_policy[TCA_RED_MAX + 1] = { [TCA_RED_PARMS] = { .len = sizeof(struct tc_red_qopt) }, [TCA_RED_STAB] = { .len = RED_STAB_SIZE }, @@ -228,9 +201,9 @@ static int red_change(struct Qdisc *sch, struct nlattr *opt) ctl = nla_data(tb[TCA_RED_PARMS]); if (ctl->limit > 0) { - child = red_create_dflt(sch, ctl->limit); - if (child == NULL) - return -ENOMEM; + child = fifo_create_dflt(sch, &bfifo_qdisc_ops, ctl->limit); + if (IS_ERR(child)) + return PTR_ERR(child); } sch_tree_lock(sch); diff --git a/net/sched/sch_tbf.c b/net/sched/sch_tbf.c index 0b7d78f59d8c..444c227fcb6b 100644 --- a/net/sched/sch_tbf.c +++ b/net/sched/sch_tbf.c @@ -242,34 +242,6 @@ static void tbf_reset(struct Qdisc* sch) qdisc_watchdog_cancel(&q->watchdog); } -static struct Qdisc *tbf_create_dflt_qdisc(struct Qdisc *sch, u32 limit) -{ - struct Qdisc *q; - struct nlattr *nla; - int ret; - - q = qdisc_create_dflt(sch->dev, &bfifo_qdisc_ops, - TC_H_MAKE(sch->handle, 1)); - if (q) { - nla = kmalloc(nla_attr_size(sizeof(struct tc_fifo_qopt)), - GFP_KERNEL); - if (nla) { - nla->nla_type = RTM_NEWQDISC; - nla->nla_len = nla_attr_size(sizeof(struct tc_fifo_qopt)); - ((struct tc_fifo_qopt *)nla_data(nla))->limit = limit; - - ret = q->ops->change(q, nla); - kfree(nla); - - if (ret == 0) - return q; - } - qdisc_destroy(q); - } - - return NULL; -} - static const struct nla_policy tbf_policy[TCA_TBF_MAX + 1] = { [TCA_TBF_PARMS] = { .len = sizeof(struct tc_tbf_qopt) }, [TCA_TBF_RTAB] = { .type = NLA_BINARY, .len = TC_RTAB_SIZE }, @@ -322,8 +294,11 @@ static int tbf_change(struct Qdisc* sch, struct nlattr *opt) goto done; if (qopt->limit > 0) { - if ((child = tbf_create_dflt_qdisc(sch, qopt->limit)) == NULL) + child = fifo_create_dflt(sch, &bfifo_qdisc_ops, qopt->limit); + if (IS_ERR(child)) { + err = PTR_ERR(child); goto done; + } } sch_tree_lock(sch); -- cgit v1.2.2 From 4b5a698ef423eebc37cfacc6d3376d6dffd5bf83 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Sun, 6 Jul 2008 15:49:08 -0700 Subject: net: fix dev_set_promiscuity() breakage Commit dad9b335 (netdevice: Fix promiscuity and allmulti overflow) broke dev_set_promiscuity() by returning on success without reprogramming the device. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/core/dev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index bfa9a6a951dd..75933932463d 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2859,7 +2859,7 @@ int dev_set_promiscuity(struct net_device *dev, int inc) int err; err = __dev_set_promiscuity(dev, inc); - if (!err) + if (err < 0) return err; if (dev->flags != old_flags) dev_set_rx_mode(dev); -- cgit v1.2.2 From d2789312cc6d875462d1d248e07a8a9caf8a6ae3 Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Tue, 8 Jul 2008 02:34:52 -0700 Subject: netfilter: use correct namespace in ip6table_security Signed-off-by: Alexey Dobriyan Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/ipv6/netfilter/ip6table_security.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv6/netfilter/ip6table_security.c b/net/ipv6/netfilter/ip6table_security.c index 063a3d9c3c67..a07abee30497 100644 --- a/net/ipv6/netfilter/ip6table_security.c +++ b/net/ipv6/netfilter/ip6table_security.c @@ -72,7 +72,7 @@ ip6t_local_in_hook(unsigned int hook, int (*okfn)(struct sk_buff *)) { return ip6t_do_table(skb, hook, in, out, - init_net.ipv6.ip6table_security); + nf_local_in_net(in, out)->ipv6.ip6table_security); } static unsigned int @@ -83,7 +83,7 @@ ip6t_forward_hook(unsigned int hook, int (*okfn)(struct sk_buff *)) { return ip6t_do_table(skb, hook, in, out, - init_net.ipv6.ip6table_security); + nf_forward_net(in, out)->ipv6.ip6table_security); } static unsigned int @@ -95,7 +95,7 @@ ip6t_local_out_hook(unsigned int hook, { /* TBD: handle short packets via raw socket */ return ip6t_do_table(skb, hook, in, out, - init_net.ipv6.ip6table_security); + nf_local_out_net(in, out)->ipv6.ip6table_security); } static struct nf_hook_ops ip6t_ops[] __read_mostly = { -- cgit v1.2.2 From b11c16beb92112885edccc79e17d39c5d218f441 Mon Sep 17 00:00:00 2001 From: Russ Dill Date: Tue, 8 Jul 2008 02:35:27 -0700 Subject: netfilter: Get rid of refrences to no longer existant Fast NAT. Get rid of refrences to no longer existant Fast NAT. IP_ROUTE_NAT support was removed in August of 2004, but references to Fast NAT were left in a couple of config options. Signed-off-by: Russ Dill Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/ipv4/netfilter/Kconfig | 3 +-- net/netfilter/Kconfig | 5 ++--- 2 files changed, 3 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/ipv4/netfilter/Kconfig b/net/ipv4/netfilter/Kconfig index 6e251402506e..f23e60c93ef9 100644 --- a/net/ipv4/netfilter/Kconfig +++ b/net/ipv4/netfilter/Kconfig @@ -213,8 +213,7 @@ config IP_NF_TARGET_NETMAP help NETMAP is an implementation of static 1:1 NAT mapping of network addresses. It maps the network address part, while keeping the host - address part intact. It is similar to Fast NAT, except that - Netfilter's connection tracking doesn't work well with Fast NAT. + address part intact. To compile it as a module, choose M here. If unsure, say N. diff --git a/net/netfilter/Kconfig b/net/netfilter/Kconfig index aa8d80c35e28..316c7af1d2b1 100644 --- a/net/netfilter/Kconfig +++ b/net/netfilter/Kconfig @@ -33,9 +33,8 @@ config NF_CONNTRACK into connections. This is required to do Masquerading or other kinds of Network - Address Translation (except for Fast NAT). It can also be used to - enhance packet filtering (see `Connection state match support' - below). + Address Translation. It can also be used to enhance packet + filtering (see `Connection state match support' below). To compile it as a module, choose M here. If unsure, say N. -- cgit v1.2.2 From b891c5a831b13f74989dcbd7b39d04537b2a05d9 Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Tue, 8 Jul 2008 02:35:55 -0700 Subject: netfilter: nf_conntrack: add allocation flag to nf_conntrack_alloc ctnetlink does not need to allocate the conntrack entries with GFP_ATOMIC as its code is executed in user context. Signed-off-by: Pablo Neira Ayuso Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/netfilter/nf_conntrack_core.c | 7 ++++--- net/netfilter/nf_conntrack_netlink.c | 2 +- 2 files changed, 5 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c index f27c99246a4c..212a0888408d 100644 --- a/net/netfilter/nf_conntrack_core.c +++ b/net/netfilter/nf_conntrack_core.c @@ -464,7 +464,8 @@ static noinline int early_drop(unsigned int hash) } struct nf_conn *nf_conntrack_alloc(const struct nf_conntrack_tuple *orig, - const struct nf_conntrack_tuple *repl) + const struct nf_conntrack_tuple *repl, + gfp_t gfp) { struct nf_conn *ct = NULL; @@ -489,7 +490,7 @@ struct nf_conn *nf_conntrack_alloc(const struct nf_conntrack_tuple *orig, } } - ct = kmem_cache_zalloc(nf_conntrack_cachep, GFP_ATOMIC); + ct = kmem_cache_zalloc(nf_conntrack_cachep, gfp); if (ct == NULL) { pr_debug("nf_conntrack_alloc: Can't alloc conntrack.\n"); atomic_dec(&nf_conntrack_count); @@ -542,7 +543,7 @@ init_conntrack(const struct nf_conntrack_tuple *tuple, return NULL; } - ct = nf_conntrack_alloc(tuple, &repl_tuple); + ct = nf_conntrack_alloc(tuple, &repl_tuple, GFP_ATOMIC); if (ct == NULL || IS_ERR(ct)) { pr_debug("Can't allocate conntrack.\n"); return (struct nf_conntrack_tuple_hash *)ct; diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index 63c4e1f299b8..dd233393f695 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -1128,7 +1128,7 @@ ctnetlink_create_conntrack(struct nlattr *cda[], struct nf_conn_help *help; struct nf_conntrack_helper *helper; - ct = nf_conntrack_alloc(otuple, rtuple); + ct = nf_conntrack_alloc(otuple, rtuple, GFP_KERNEL); if (ct == NULL || IS_ERR(ct)) return -ENOMEM; -- cgit v1.2.2 From 43de9dfeaa30f7ed801dc1c38bdb63b1738bddcc Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Tue, 8 Jul 2008 02:36:18 -0700 Subject: netfilter: ip6table_filter in netns for real One still needs to remove checks in nf_hook_slow() and nf_sockopt_find() to test this, though. Signed-off-by: Alexey Dobriyan Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/ipv6/netfilter/ip6table_filter.c | 31 ++++++++++++++++++++++--------- 1 file changed, 22 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/ipv6/netfilter/ip6table_filter.c b/net/ipv6/netfilter/ip6table_filter.c index f979e48b469b..55a2c290bad4 100644 --- a/net/ipv6/netfilter/ip6table_filter.c +++ b/net/ipv6/netfilter/ip6table_filter.c @@ -61,13 +61,25 @@ static struct xt_table packet_filter = { /* The work comes in here from netfilter.c. */ static unsigned int -ip6t_hook(unsigned int hook, - struct sk_buff *skb, - const struct net_device *in, - const struct net_device *out, - int (*okfn)(struct sk_buff *)) +ip6t_local_in_hook(unsigned int hook, + struct sk_buff *skb, + const struct net_device *in, + const struct net_device *out, + int (*okfn)(struct sk_buff *)) +{ + return ip6t_do_table(skb, hook, in, out, + nf_local_in_net(in, out)->ipv6.ip6table_filter); +} + +static unsigned int +ip6t_forward_hook(unsigned int hook, + struct sk_buff *skb, + const struct net_device *in, + const struct net_device *out, + int (*okfn)(struct sk_buff *)) { - return ip6t_do_table(skb, hook, in, out, init_net.ipv6.ip6table_filter); + return ip6t_do_table(skb, hook, in, out, + nf_forward_net(in, out)->ipv6.ip6table_filter); } static unsigned int @@ -87,19 +99,20 @@ ip6t_local_out_hook(unsigned int hook, } #endif - return ip6t_do_table(skb, hook, in, out, init_net.ipv6.ip6table_filter); + return ip6t_do_table(skb, hook, in, out, + nf_local_out_net(in, out)->ipv6.ip6table_filter); } static struct nf_hook_ops ip6t_ops[] __read_mostly = { { - .hook = ip6t_hook, + .hook = ip6t_local_in_hook, .owner = THIS_MODULE, .pf = PF_INET6, .hooknum = NF_INET_LOCAL_IN, .priority = NF_IP6_PRI_FILTER, }, { - .hook = ip6t_hook, + .hook = ip6t_forward_hook, .owner = THIS_MODULE, .pf = PF_INET6, .hooknum = NF_INET_FORWARD, -- cgit v1.2.2 From 58de7862e61cb71251a25314d1b3d7260af1448a Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Tue, 8 Jul 2008 02:37:07 -0700 Subject: netfilter: ebt_nflog: fix Kconfig typo The help text should refer to nflog instead of ulog. Noticed by Krzysztof Halasa . Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/bridge/netfilter/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/bridge/netfilter/Kconfig b/net/bridge/netfilter/Kconfig index 540df4106bec..909479794999 100644 --- a/net/bridge/netfilter/Kconfig +++ b/net/bridge/netfilter/Kconfig @@ -230,7 +230,7 @@ config BRIDGE_EBT_NFLOG either the old LOG target, the old ULOG target or nfnetlink_log as backend. - This option adds the ulog watcher, that you can use in any rule + This option adds the nflog watcher, that you can use in any rule in any ebtables table. To compile it as a module, choose M here. If unsure, say N. -- cgit v1.2.2 From 4ad3f26162ece5aca3045fd45e15dd99acea4a0e Mon Sep 17 00:00:00 2001 From: Joonwoo Park Date: Tue, 8 Jul 2008 02:38:56 -0700 Subject: netfilter: fix string extension for case insensitive pattern matching The flag XT_STRING_FLAG_IGNORECASE indicates case insensitive string matching. netfilter can find cmd.exe, Cmd.exe, cMd.exe and etc easily. A new revision 1 was added, in the meantime invert of xt_string_info was moved into flags as a flag. If revision is 1, The flag XT_STRING_FLAG_INVERT indicates invert matching. Signed-off-by: Joonwoo Park Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/netfilter/xt_string.c | 38 ++++++++++++++++++++++++++++++++++++-- 1 file changed, 36 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/netfilter/xt_string.c b/net/netfilter/xt_string.c index 72f694d947f4..4903182a062b 100644 --- a/net/netfilter/xt_string.c +++ b/net/netfilter/xt_string.c @@ -29,12 +29,16 @@ string_mt(const struct sk_buff *skb, const struct net_device *in, { const struct xt_string_info *conf = matchinfo; struct ts_state state; + int invert; memset(&state, 0, sizeof(struct ts_state)); + invert = (match->revision == 0 ? conf->u.v0.invert : + conf->u.v1.flags & XT_STRING_FLAG_INVERT); + return (skb_find_text((struct sk_buff *)skb, conf->from_offset, conf->to_offset, conf->config, &state) - != UINT_MAX) ^ conf->invert; + != UINT_MAX) ^ invert; } #define STRING_TEXT_PRIV(m) ((struct xt_string_info *)(m)) @@ -46,6 +50,7 @@ string_mt_check(const char *tablename, const void *ip, { struct xt_string_info *conf = matchinfo; struct ts_config *ts_conf; + int flags = TS_AUTOLOAD; /* Damn, can't handle this case properly with iptables... */ if (conf->from_offset > conf->to_offset) @@ -54,8 +59,15 @@ string_mt_check(const char *tablename, const void *ip, return false; if (conf->patlen > XT_STRING_MAX_PATTERN_SIZE) return false; + if (match->revision == 1) { + if (conf->u.v1.flags & + ~(XT_STRING_FLAG_IGNORECASE | XT_STRING_FLAG_INVERT)) + return false; + if (conf->u.v1.flags & XT_STRING_FLAG_IGNORECASE) + flags |= TS_IGNORECASE; + } ts_conf = textsearch_prepare(conf->algo, conf->pattern, conf->patlen, - GFP_KERNEL, TS_AUTOLOAD); + GFP_KERNEL, flags); if (IS_ERR(ts_conf)) return false; @@ -72,6 +84,17 @@ static void string_mt_destroy(const struct xt_match *match, void *matchinfo) static struct xt_match string_mt_reg[] __read_mostly = { { .name = "string", + .revision = 0, + .family = AF_INET, + .checkentry = string_mt_check, + .match = string_mt, + .destroy = string_mt_destroy, + .matchsize = sizeof(struct xt_string_info), + .me = THIS_MODULE + }, + { + .name = "string", + .revision = 1, .family = AF_INET, .checkentry = string_mt_check, .match = string_mt, @@ -81,6 +104,17 @@ static struct xt_match string_mt_reg[] __read_mostly = { }, { .name = "string", + .revision = 0, + .family = AF_INET6, + .checkentry = string_mt_check, + .match = string_mt, + .destroy = string_mt_destroy, + .matchsize = sizeof(struct xt_string_info), + .me = THIS_MODULE + }, + { + .name = "string", + .revision = 1, .family = AF_INET6, .checkentry = string_mt_check, .match = string_mt, -- cgit v1.2.2 From 81c684d12ddc05bba4953e36e9cdd5939dde344b Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Tue, 8 Jul 2008 03:05:28 -0700 Subject: ipv4: remove flush_mutex from ipv4_sysctl_rtcache_flush It is possible to avoid locking at all in ipv4_sysctl_rtcache_flush by defining local ctl_table on the stack. The patch is based on the suggestion from Eric W. Biederman. Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv4/route.c | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 113cd2512ba7..79c1e74263a1 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -2873,22 +2873,20 @@ void ip_rt_multicast_event(struct in_device *in_dev) } #ifdef CONFIG_SYSCTL -static int ipv4_sysctl_rtcache_flush(ctl_table *ctl, int write, +static int ipv4_sysctl_rtcache_flush(ctl_table *__ctl, int write, struct file *filp, void __user *buffer, size_t *lenp, loff_t *ppos) { if (write) { int flush_delay; + ctl_table ctl; struct net *net; - static DEFINE_MUTEX(flush_mutex); - mutex_lock(&flush_mutex); - ctl->data = &flush_delay; - proc_dointvec(ctl, write, filp, buffer, lenp, ppos); - ctl->data = NULL; - mutex_unlock(&flush_mutex); + memcpy(&ctl, __ctl, sizeof(ctl)); + ctl.data = &flush_delay; + proc_dointvec(&ctl, write, filp, buffer, lenp, ppos); - net = (struct net *)ctl->extra1; + net = (struct net *)__ctl->extra1; rt_cache_flush(net, flush_delay); return 0; } -- cgit v1.2.2 From acc81e1465d29e0284008770cc4b8bc90bd93bd7 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Tue, 8 Jul 2008 03:21:27 -0700 Subject: vlan: fix network_header/mac_header adjustments Lennert Buytenhek points out that the VLAN code incorrectly adjusts skb->network_header to point in the middle of the VLAN header and additionally tries to adjust skb->mac_header without checking for validity. The network_header should not be touched at all since we're only adding headers in front of it, mac_header adjustments are not necessary at all. Based on patch by Lennert Buytenhek . Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan_dev.c | 1 - 1 file changed, 1 deletion(-) (limited to 'net') diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index a0617bf7cec6..4a8525927c27 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -308,7 +308,6 @@ static int vlan_dev_hard_header(struct sk_buff *skb, struct net_device *dev, vhdr->h_vlan_encapsulated_proto = htons(len); skb->protocol = htons(ETH_P_8021Q); - skb_reset_network_header(skb); } /* Before delegating work to the lower layer, enter our MAC-address */ -- cgit v1.2.2 From 26a25239d7a660cc7162e2463b48b40d544364d0 Mon Sep 17 00:00:00 2001 From: Joonwoo Park Date: Tue, 8 Jul 2008 03:22:16 -0700 Subject: vlan: Use is_vlan_dev() Use simplified is_vlan_dev function. Signed-off-by: Joonwoo Park Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan.c | 3 +-- net/8021q/vlan_dev.c | 2 +- net/8021q/vlanproc.c | 2 +- 3 files changed, 3 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c index b529110c9355..8141e2dc510b 100644 --- a/net/8021q/vlan.c +++ b/net/8021q/vlan.c @@ -570,8 +570,7 @@ static int vlan_ioctl_handler(struct net *net, void __user *arg) goto out; err = -EINVAL; - if (args.cmd != ADD_VLAN_CMD && - !(dev->priv_flags & IFF_802_1Q_VLAN)) + if (args.cmd != ADD_VLAN_CMD && !is_vlan_dev(dev)) goto out; } diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 4a8525927c27..88f318a9601d 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -694,7 +694,7 @@ static int vlan_dev_init(struct net_device *dev) dev->hard_start_xmit = vlan_dev_hard_start_xmit; } - if (real_dev->priv_flags & IFF_802_1Q_VLAN) + if (is_vlan_dev(real_dev)) subclass = 1; lockdep_set_class_and_subclass(&dev->_xmit_lock, diff --git a/net/8021q/vlanproc.c b/net/8021q/vlanproc.c index 08b54b593d56..6073a888b6f9 100644 --- a/net/8021q/vlanproc.c +++ b/net/8021q/vlanproc.c @@ -290,7 +290,7 @@ static int vlandev_seq_show(struct seq_file *seq, void *offset) static const char fmt[] = "%30s %12lu\n"; int i; - if (!(vlandev->priv_flags & IFF_802_1Q_VLAN)) + if (!is_vlan_dev(vlandev)) return 0; seq_printf(seq, -- cgit v1.2.2 From 75b8846acd11ad3fc736d4df3413fe946bbf367c Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Tue, 8 Jul 2008 03:22:42 -0700 Subject: vlan: Add ethtool support Add ethtool support for querying the device for offload settings. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan_dev.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) (limited to 'net') diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 88f318a9601d..722697d31e4f 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -28,6 +28,7 @@ #include #include #include +#include #include #include #include @@ -716,6 +717,22 @@ static void vlan_dev_uninit(struct net_device *dev) } } +static u32 vlan_ethtool_get_rx_csum(struct net_device *dev) +{ + const struct vlan_dev_info *vlan = vlan_dev_info(dev); + struct net_device *real_dev = vlan->real_dev; + + if (real_dev->ethtool_ops == NULL || + real_dev->ethtool_ops->get_rx_csum == NULL) + return 0; + return real_dev->ethtool_ops->get_rx_csum(real_dev); +} + +static const struct ethtool_ops vlan_ethtool_ops = { + .get_link = ethtool_op_get_link, + .get_rx_csum = vlan_ethtool_get_rx_csum, +}; + void vlan_setup(struct net_device *dev) { ether_setup(dev); @@ -734,6 +751,7 @@ void vlan_setup(struct net_device *dev) dev->change_rx_flags = vlan_dev_change_rx_flags; dev->do_ioctl = vlan_dev_ioctl; dev->destructor = free_netdev; + dev->ethtool_ops = &vlan_ethtool_ops; memset(dev->broadcast, 0, ETH_ALEN); } -- cgit v1.2.2 From 7750f403cbe56971336d575b354365190b4e3227 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Tue, 8 Jul 2008 03:23:36 -0700 Subject: vlan: uninline __vlan_hwaccel_rx The function is huge and included at least once in every VLAN acceleration capable driver. Uninline it; to avoid having drivers depend on the VLAN module, the function is always built in statically when VLAN is enabled. With all VLAN acceleration capable drivers that build on x86_64 enabled, this results in: text data bss dec hex filename 6515227 854044 343968 7713239 75b1d7 vmlinux.inlined 6505637 854044 343968 7703649 758c61 vmlinux.uninlined ---------------------------------------------------------- -9590 Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/Makefile | 9 +++++---- net/8021q/vlan.h | 8 ++++++++ net/8021q/vlan_core.c | 48 ++++++++++++++++++++++++++++++++++++++++++++++++ net/Makefile | 4 +++- 4 files changed, 64 insertions(+), 5 deletions(-) create mode 100644 net/8021q/vlan_core.c (limited to 'net') diff --git a/net/8021q/Makefile b/net/8021q/Makefile index 3006e9ed7b08..9f4f174ead1c 100644 --- a/net/8021q/Makefile +++ b/net/8021q/Makefile @@ -1,9 +1,10 @@ # # Makefile for the Linux VLAN layer. # +obj-$(subst m,y,$(CONFIG_VLAN_8021Q)) += vlan_core.o +obj-$(CONFIG_VLAN_8021Q) += 8021q.o -obj-$(CONFIG_VLAN_8021Q) += 8021q.o +8021q-y := vlan.o vlan_dev.o vlan_netlink.o +8021q-$(CONFIG_VLAN_8021Q_GVRP) += vlan_gvrp.o +8021q-$(CONFIG_PROC_FS) += vlanproc.o -8021q-y := vlan.o vlan_dev.o vlan_netlink.o -8021q-$(CONFIG_VLAN_8021Q_GVRP) += vlan_gvrp.o -8021q-$(CONFIG_PROC_FS) += vlanproc.o diff --git a/net/8021q/vlan.h b/net/8021q/vlan.h index 097b2e04c928..7cc1a97c42fc 100644 --- a/net/8021q/vlan.h +++ b/net/8021q/vlan.h @@ -37,6 +37,14 @@ void vlan_setup(struct net_device *dev); int register_vlan_dev(struct net_device *dev); void unregister_vlan_dev(struct net_device *dev); +static inline u32 vlan_get_ingress_priority(struct net_device *dev, + unsigned short vlan_tag) +{ + struct vlan_dev_info *vip = vlan_dev_info(dev); + + return vip->ingress_priority_map[(vlan_tag >> 13) & 0x7]; +} + #ifdef CONFIG_VLAN_8021Q_GVRP extern int vlan_gvrp_request_join(const struct net_device *dev); extern void vlan_gvrp_request_leave(const struct net_device *dev); diff --git a/net/8021q/vlan_core.c b/net/8021q/vlan_core.c new file mode 100644 index 000000000000..85c94edb000f --- /dev/null +++ b/net/8021q/vlan_core.c @@ -0,0 +1,48 @@ +#include +#include +#include +#include "vlan.h" + +/* VLAN rx hw acceleration helper. This acts like netif_{rx,receive_skb}(). */ +int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp, + unsigned short vlan_tag, int polling) +{ + struct net_device_stats *stats; + + if (skb_bond_should_drop(skb)) { + dev_kfree_skb_any(skb); + return NET_RX_DROP; + } + + skb->dev = vlan_group_get_device(grp, vlan_tag & VLAN_VID_MASK); + if (skb->dev == NULL) { + dev_kfree_skb_any(skb); + /* Not NET_RX_DROP, this is not being dropped + * due to congestion. */ + return NET_RX_SUCCESS; + } + skb->dev->last_rx = jiffies; + + stats = &skb->dev->stats; + stats->rx_packets++; + stats->rx_bytes += skb->len; + + skb->priority = vlan_get_ingress_priority(skb->dev, vlan_tag); + switch (skb->pkt_type) { + case PACKET_BROADCAST: + break; + case PACKET_MULTICAST: + stats->multicast++; + break; + case PACKET_OTHERHOST: + /* Our lower layer thinks this is not local, let's make sure. + * This allows the VLAN to have a different MAC than the + * underlying device, and still route correctly. */ + if (!compare_ether_addr(eth_hdr(skb)->h_dest, + skb->dev->dev_addr)) + skb->pkt_type = PACKET_HOST; + break; + }; + return (polling ? netif_receive_skb(skb) : netif_rx(skb)); +} +EXPORT_SYMBOL(__vlan_hwaccel_rx); diff --git a/net/Makefile b/net/Makefile index b7a13643b549..4f43e7f874f3 100644 --- a/net/Makefile +++ b/net/Makefile @@ -42,7 +42,9 @@ obj-$(CONFIG_AF_RXRPC) += rxrpc/ obj-$(CONFIG_ATM) += atm/ obj-$(CONFIG_DECNET) += decnet/ obj-$(CONFIG_ECONET) += econet/ -obj-$(CONFIG_VLAN_8021Q) += 8021q/ +ifneq ($(CONFIG_VLAN_8021Q),) +obj-y += 8021q/ +endif obj-$(CONFIG_IP_DCCP) += dccp/ obj-$(CONFIG_IP_SCTP) += sctp/ obj-y += wireless/ -- cgit v1.2.2 From 22d1ba74bbafa96d3f425cc12714d3fe8675183f Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Tue, 8 Jul 2008 03:23:57 -0700 Subject: vlan: move struct vlan_dev_info to private header Hide struct vlan_dev_info from drivers to prevent them from growing more creative ways to use it. Provide accessors for the two drivers that currently use it. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan.c | 4 +--- net/8021q/vlan.h | 50 +++++++++++++++++++++++++++++++++++++++++++++++++- net/8021q/vlan_core.c | 12 ++++++++++++ net/8021q/vlan_dev.c | 5 ----- 4 files changed, 62 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c index 8141e2dc510b..7a2625d2f9a0 100644 --- a/net/8021q/vlan.c +++ b/net/8021q/vlan.c @@ -543,7 +543,6 @@ static struct notifier_block vlan_notifier_block __read_mostly = { static int vlan_ioctl_handler(struct net *net, void __user *arg) { int err; - unsigned short vid = 0; struct vlan_ioctl_args args; struct net_device *dev = NULL; @@ -644,8 +643,7 @@ static int vlan_ioctl_handler(struct net *net, void __user *arg) case GET_VLAN_VID_CMD: err = 0; - vlan_dev_get_vid(dev, &vid); - args.u.VID = vid; + args.u.VID = vlan_dev_vlan_id(dev); if (copy_to_user(arg, &args, sizeof(struct vlan_ioctl_args))) err = -EFAULT; diff --git a/net/8021q/vlan.h b/net/8021q/vlan.h index 7cc1a97c42fc..14c421e033f2 100644 --- a/net/8021q/vlan.h +++ b/net/8021q/vlan.h @@ -3,6 +3,55 @@ #include + +/** + * struct vlan_priority_tci_mapping - vlan egress priority mappings + * @priority: skb priority + * @vlan_qos: vlan priority: (skb->priority << 13) & 0xE000 + * @next: pointer to next struct + */ +struct vlan_priority_tci_mapping { + u32 priority; + unsigned short vlan_qos; + struct vlan_priority_tci_mapping *next; +}; + +/** + * struct vlan_dev_info - VLAN private device data + * @nr_ingress_mappings: number of ingress priority mappings + * @ingress_priority_map: ingress priority mappings + * @nr_egress_mappings: number of egress priority mappings + * @egress_priority_map: hash of egress priority mappings + * @vlan_id: VLAN identifier + * @flags: device flags + * @real_dev: underlying netdevice + * @real_dev_addr: address of underlying netdevice + * @dent: proc dir entry + * @cnt_inc_headroom_on_tx: statistic - number of skb expansions on TX + * @cnt_encap_on_xmit: statistic - number of skb encapsulations on TX + */ +struct vlan_dev_info { + unsigned int nr_ingress_mappings; + u32 ingress_priority_map[8]; + unsigned int nr_egress_mappings; + struct vlan_priority_tci_mapping *egress_priority_map[16]; + + unsigned short vlan_id; + unsigned short flags; + + struct net_device *real_dev; + unsigned char real_dev_addr[ETH_ALEN]; + + struct proc_dir_entry *dent; + unsigned long cnt_inc_headroom_on_tx; + unsigned long cnt_encap_on_xmit; +}; + +static inline struct vlan_dev_info *vlan_dev_info(const struct net_device *dev) +{ + return netdev_priv(dev); +} + #define VLAN_GRP_HASH_SHIFT 5 #define VLAN_GRP_HASH_SIZE (1 << VLAN_GRP_HASH_SHIFT) #define VLAN_GRP_HASH_MASK (VLAN_GRP_HASH_SIZE - 1) @@ -30,7 +79,6 @@ int vlan_dev_set_egress_priority(const struct net_device *dev, u32 skb_prio, short vlan_prio); int vlan_dev_change_flags(const struct net_device *dev, u32 flag, u32 mask); void vlan_dev_get_realdev_name(const struct net_device *dev, char *result); -void vlan_dev_get_vid(const struct net_device *dev, unsigned short *result); int vlan_check_real_dev(struct net_device *real_dev, unsigned short vlan_id); void vlan_setup(struct net_device *dev); diff --git a/net/8021q/vlan_core.c b/net/8021q/vlan_core.c index 85c94edb000f..f980b9154cc3 100644 --- a/net/8021q/vlan_core.c +++ b/net/8021q/vlan_core.c @@ -46,3 +46,15 @@ int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp, return (polling ? netif_receive_skb(skb) : netif_rx(skb)); } EXPORT_SYMBOL(__vlan_hwaccel_rx); + +struct net_device *vlan_dev_real_dev(const struct net_device *dev) +{ + return vlan_dev_info(dev)->real_dev; +} +EXPORT_SYMBOL_GPL(vlan_dev_real_dev); + +u16 vlan_dev_vlan_id(const struct net_device *dev) +{ + return vlan_dev_info(dev)->vlan_id; +} +EXPORT_SYMBOL_GPL(vlan_dev_vlan_id); diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 722697d31e4f..2aab294c5744 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -531,11 +531,6 @@ void vlan_dev_get_realdev_name(const struct net_device *dev, char *result) strncpy(result, vlan_dev_info(dev)->real_dev->name, 23); } -void vlan_dev_get_vid(const struct net_device *dev, unsigned short *result) -{ - *result = vlan_dev_info(dev)->vlan_id; -} - static int vlan_dev_open(struct net_device *dev) { struct vlan_dev_info *vlan = vlan_dev_info(dev); -- cgit v1.2.2 From 9bb8582efb555521c7eec595ebd34e835ddc34b8 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Tue, 8 Jul 2008 03:24:44 -0700 Subject: vlan: TCI related type and naming cleanups The VLAN code contains multiple spots that use tag, id and tci as identifiers for arguments and variables incorrectly and they actually contain or are expected to contain something different. Additionally types are used inconsistently (unsigned short vs u16) and identifiers are sometimes capitalized. - consistently use u16 for storing TCI, ID or QoS values - consistently use vlan_id and vlan_tci for storing the respective values - remove capitalization - add kdoc comment to netif_hwaccel_{rx,receive_skb} Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan.c | 34 ++++++++++++++++------------------ net/8021q/vlan.h | 19 +++++++++---------- net/8021q/vlan_core.c | 6 +++--- net/8021q/vlan_dev.c | 47 +++++++++++++++++++++++------------------------ net/8021q/vlan_gvrp.c | 8 ++++---- 5 files changed, 55 insertions(+), 59 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c index 7a2625d2f9a0..68bdcf4a795c 100644 --- a/net/8021q/vlan.c +++ b/net/8021q/vlan.c @@ -83,13 +83,12 @@ static struct vlan_group *__vlan_find_group(struct net_device *real_dev) * * Must be invoked with RCU read lock (no preempt) */ -struct net_device *__find_vlan_dev(struct net_device *real_dev, - unsigned short VID) +struct net_device *__find_vlan_dev(struct net_device *real_dev, u16 vlan_id) { struct vlan_group *grp = __vlan_find_group(real_dev); if (grp) - return vlan_group_get_device(grp, VID); + return vlan_group_get_device(grp, vlan_id); return NULL; } @@ -117,14 +116,14 @@ static struct vlan_group *vlan_group_alloc(struct net_device *real_dev) return grp; } -static int vlan_group_prealloc_vid(struct vlan_group *vg, int vid) +static int vlan_group_prealloc_vid(struct vlan_group *vg, u16 vlan_id) { struct net_device **array; unsigned int size; ASSERT_RTNL(); - array = vg->vlan_devices_arrays[vid / VLAN_GROUP_ARRAY_PART_LEN]; + array = vg->vlan_devices_arrays[vlan_id / VLAN_GROUP_ARRAY_PART_LEN]; if (array != NULL) return 0; @@ -133,7 +132,7 @@ static int vlan_group_prealloc_vid(struct vlan_group *vg, int vid) if (array == NULL) return -ENOBUFS; - vg->vlan_devices_arrays[vid / VLAN_GROUP_ARRAY_PART_LEN] = array; + vg->vlan_devices_arrays[vlan_id / VLAN_GROUP_ARRAY_PART_LEN] = array; return 0; } @@ -147,7 +146,7 @@ void unregister_vlan_dev(struct net_device *dev) struct vlan_dev_info *vlan = vlan_dev_info(dev); struct net_device *real_dev = vlan->real_dev; struct vlan_group *grp; - unsigned short vlan_id = vlan->vlan_id; + u16 vlan_id = vlan->vlan_id; ASSERT_RTNL(); @@ -205,7 +204,7 @@ static void vlan_transfer_operstate(const struct net_device *dev, } } -int vlan_check_real_dev(struct net_device *real_dev, unsigned short vlan_id) +int vlan_check_real_dev(struct net_device *real_dev, u16 vlan_id) { char *name = real_dev->name; @@ -242,7 +241,7 @@ int register_vlan_dev(struct net_device *dev) { struct vlan_dev_info *vlan = vlan_dev_info(dev); struct net_device *real_dev = vlan->real_dev; - unsigned short vlan_id = vlan->vlan_id; + u16 vlan_id = vlan->vlan_id; struct vlan_group *grp, *ngrp = NULL; int err; @@ -295,8 +294,7 @@ out_free_group: /* Attach a VLAN device to a mac address (ie Ethernet Card). * Returns 0 if the device was created or a negative error code otherwise. */ -static int register_vlan_device(struct net_device *real_dev, - unsigned short VLAN_ID) +static int register_vlan_device(struct net_device *real_dev, u16 vlan_id) { struct net_device *new_dev; struct net *net = dev_net(real_dev); @@ -304,10 +302,10 @@ static int register_vlan_device(struct net_device *real_dev, char name[IFNAMSIZ]; int err; - if (VLAN_ID >= VLAN_VID_MASK) + if (vlan_id >= VLAN_VID_MASK) return -ERANGE; - err = vlan_check_real_dev(real_dev, VLAN_ID); + err = vlan_check_real_dev(real_dev, vlan_id); if (err < 0) return err; @@ -315,26 +313,26 @@ static int register_vlan_device(struct net_device *real_dev, switch (vn->name_type) { case VLAN_NAME_TYPE_RAW_PLUS_VID: /* name will look like: eth1.0005 */ - snprintf(name, IFNAMSIZ, "%s.%.4i", real_dev->name, VLAN_ID); + snprintf(name, IFNAMSIZ, "%s.%.4i", real_dev->name, vlan_id); break; case VLAN_NAME_TYPE_PLUS_VID_NO_PAD: /* Put our vlan.VID in the name. * Name will look like: vlan5 */ - snprintf(name, IFNAMSIZ, "vlan%i", VLAN_ID); + snprintf(name, IFNAMSIZ, "vlan%i", vlan_id); break; case VLAN_NAME_TYPE_RAW_PLUS_VID_NO_PAD: /* Put our vlan.VID in the name. * Name will look like: eth0.5 */ - snprintf(name, IFNAMSIZ, "%s.%i", real_dev->name, VLAN_ID); + snprintf(name, IFNAMSIZ, "%s.%i", real_dev->name, vlan_id); break; case VLAN_NAME_TYPE_PLUS_VID: /* Put our vlan.VID in the name. * Name will look like: vlan0005 */ default: - snprintf(name, IFNAMSIZ, "vlan%.4i", VLAN_ID); + snprintf(name, IFNAMSIZ, "vlan%.4i", vlan_id); } new_dev = alloc_netdev(sizeof(struct vlan_dev_info), name, @@ -349,7 +347,7 @@ static int register_vlan_device(struct net_device *real_dev, */ new_dev->mtu = real_dev->mtu; - vlan_dev_info(new_dev)->vlan_id = VLAN_ID; /* 1 through VLAN_VID_MASK */ + vlan_dev_info(new_dev)->vlan_id = vlan_id; vlan_dev_info(new_dev)->real_dev = real_dev; vlan_dev_info(new_dev)->dent = NULL; vlan_dev_info(new_dev)->flags = VLAN_FLAG_REORDER_HDR; diff --git a/net/8021q/vlan.h b/net/8021q/vlan.h index 14c421e033f2..a6603a4d917f 100644 --- a/net/8021q/vlan.h +++ b/net/8021q/vlan.h @@ -12,7 +12,7 @@ */ struct vlan_priority_tci_mapping { u32 priority; - unsigned short vlan_qos; + u16 vlan_qos; struct vlan_priority_tci_mapping *next; }; @@ -36,8 +36,8 @@ struct vlan_dev_info { unsigned int nr_egress_mappings; struct vlan_priority_tci_mapping *egress_priority_map[16]; - unsigned short vlan_id; - unsigned short flags; + u16 vlan_id; + u16 flags; struct net_device *real_dev; unsigned char real_dev_addr[ETH_ALEN]; @@ -67,30 +67,29 @@ static inline struct vlan_dev_info *vlan_dev_info(const struct net_device *dev) * Must be invoked with rcu_read_lock (ie preempt disabled) * or with RTNL. */ -struct net_device *__find_vlan_dev(struct net_device *real_dev, - unsigned short VID); /* vlan.c */ +struct net_device *__find_vlan_dev(struct net_device *real_dev, u16 vlan_id); /* found in vlan_dev.c */ int vlan_skb_recv(struct sk_buff *skb, struct net_device *dev, struct packet_type *ptype, struct net_device *orig_dev); void vlan_dev_set_ingress_priority(const struct net_device *dev, - u32 skb_prio, short vlan_prio); + u32 skb_prio, u16 vlan_prio); int vlan_dev_set_egress_priority(const struct net_device *dev, - u32 skb_prio, short vlan_prio); + u32 skb_prio, u16 vlan_prio); int vlan_dev_change_flags(const struct net_device *dev, u32 flag, u32 mask); void vlan_dev_get_realdev_name(const struct net_device *dev, char *result); -int vlan_check_real_dev(struct net_device *real_dev, unsigned short vlan_id); +int vlan_check_real_dev(struct net_device *real_dev, u16 vlan_id); void vlan_setup(struct net_device *dev); int register_vlan_dev(struct net_device *dev); void unregister_vlan_dev(struct net_device *dev); static inline u32 vlan_get_ingress_priority(struct net_device *dev, - unsigned short vlan_tag) + u16 vlan_tci) { struct vlan_dev_info *vip = vlan_dev_info(dev); - return vip->ingress_priority_map[(vlan_tag >> 13) & 0x7]; + return vip->ingress_priority_map[(vlan_tci >> 13) & 0x7]; } #ifdef CONFIG_VLAN_8021Q_GVRP diff --git a/net/8021q/vlan_core.c b/net/8021q/vlan_core.c index f980b9154cc3..68df12d3664b 100644 --- a/net/8021q/vlan_core.c +++ b/net/8021q/vlan_core.c @@ -5,7 +5,7 @@ /* VLAN rx hw acceleration helper. This acts like netif_{rx,receive_skb}(). */ int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp, - unsigned short vlan_tag, int polling) + u16 vlan_tci, int polling) { struct net_device_stats *stats; @@ -14,7 +14,7 @@ int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp, return NET_RX_DROP; } - skb->dev = vlan_group_get_device(grp, vlan_tag & VLAN_VID_MASK); + skb->dev = vlan_group_get_device(grp, vlan_tci & VLAN_VID_MASK); if (skb->dev == NULL) { dev_kfree_skb_any(skb); /* Not NET_RX_DROP, this is not being dropped @@ -27,7 +27,7 @@ int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp, stats->rx_packets++; stats->rx_bytes += skb->len; - skb->priority = vlan_get_ingress_priority(skb->dev, vlan_tag); + skb->priority = vlan_get_ingress_priority(skb->dev, vlan_tci); switch (skb->pkt_type) { case PACKET_BROADCAST: break; diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 2aab294c5744..2ccac6bea57e 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -150,9 +150,9 @@ int vlan_skb_recv(struct sk_buff *skb, struct net_device *dev, struct packet_type *ptype, struct net_device *orig_dev) { struct vlan_hdr *vhdr; - unsigned short vid; struct net_device_stats *stats; - unsigned short vlan_TCI; + u16 vlan_id; + u16 vlan_tci; skb = skb_share_check(skb, GFP_ATOMIC); if (skb == NULL) @@ -162,14 +162,14 @@ int vlan_skb_recv(struct sk_buff *skb, struct net_device *dev, goto err_free; vhdr = (struct vlan_hdr *)skb->data; - vlan_TCI = ntohs(vhdr->h_vlan_TCI); - vid = (vlan_TCI & VLAN_VID_MASK); + vlan_tci = ntohs(vhdr->h_vlan_TCI); + vlan_id = vlan_tci & VLAN_VID_MASK; rcu_read_lock(); - skb->dev = __find_vlan_dev(dev, vid); + skb->dev = __find_vlan_dev(dev, vlan_id); if (!skb->dev) { pr_debug("%s: ERROR: No net_device for VID: %u on dev: %s\n", - __func__, (unsigned int)vid, dev->name); + __func__, vlan_id, dev->name); goto err_unlock; } @@ -181,11 +181,10 @@ int vlan_skb_recv(struct sk_buff *skb, struct net_device *dev, skb_pull_rcsum(skb, VLAN_HLEN); - skb->priority = vlan_get_ingress_priority(skb->dev, - ntohs(vhdr->h_vlan_TCI)); + skb->priority = vlan_get_ingress_priority(skb->dev, vlan_tci); pr_debug("%s: priority: %u for TCI: %hu\n", - __func__, skb->priority, ntohs(vhdr->h_vlan_TCI)); + __func__, skb->priority, vlan_tci); switch (skb->pkt_type) { case PACKET_BROADCAST: /* Yeah, stats collect these together.. */ @@ -228,7 +227,7 @@ err_free: return NET_RX_DROP; } -static inline unsigned short +static inline u16 vlan_dev_get_egress_qos_mask(struct net_device *dev, struct sk_buff *skb) { struct vlan_priority_tci_mapping *mp; @@ -260,7 +259,7 @@ static int vlan_dev_hard_header(struct sk_buff *skb, struct net_device *dev, unsigned int len) { struct vlan_hdr *vhdr; - unsigned short veth_TCI = 0; + u16 vlan_tci = 0; int rc = 0; int build_vlan_header = 0; struct net_device *vdev = dev; @@ -292,10 +291,10 @@ static int vlan_dev_hard_header(struct sk_buff *skb, struct net_device *dev, * VLAN ID 12 bits (low bits) * */ - veth_TCI = vlan_dev_info(dev)->vlan_id; - veth_TCI |= vlan_dev_get_egress_qos_mask(dev, skb); + vlan_tci = vlan_dev_info(dev)->vlan_id; + vlan_tci |= vlan_dev_get_egress_qos_mask(dev, skb); - vhdr->h_vlan_TCI = htons(veth_TCI); + vhdr->h_vlan_TCI = htons(vlan_tci); /* * Set the protocol type. For a packet of type ETH_P_802_3 we @@ -373,7 +372,7 @@ static int vlan_dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev) if (veth->h_vlan_proto != htons(ETH_P_8021Q) || vlan_dev_info(dev)->flags & VLAN_FLAG_REORDER_HDR) { int orig_headroom = skb_headroom(skb); - unsigned short veth_TCI; + u16 vlan_tci; /* This is not a VLAN frame...but we can fix that! */ vlan_dev_info(dev)->cnt_encap_on_xmit++; @@ -386,10 +385,10 @@ static int vlan_dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev) * CFI 1 bit * VLAN ID 12 bits (low bits) */ - veth_TCI = vlan_dev_info(dev)->vlan_id; - veth_TCI |= vlan_dev_get_egress_qos_mask(dev, skb); + vlan_tci = vlan_dev_info(dev)->vlan_id; + vlan_tci |= vlan_dev_get_egress_qos_mask(dev, skb); - skb = __vlan_put_tag(skb, veth_TCI); + skb = __vlan_put_tag(skb, vlan_tci); if (!skb) { stats->tx_dropped++; return 0; @@ -422,7 +421,7 @@ static int vlan_dev_hwaccel_hard_start_xmit(struct sk_buff *skb, struct net_device *dev) { struct net_device_stats *stats = &dev->stats; - unsigned short veth_TCI; + u16 vlan_tci; /* Construct the second two bytes. This field looks something * like: @@ -430,9 +429,9 @@ static int vlan_dev_hwaccel_hard_start_xmit(struct sk_buff *skb, * CFI 1 bit * VLAN ID 12 bits (low bits) */ - veth_TCI = vlan_dev_info(dev)->vlan_id; - veth_TCI |= vlan_dev_get_egress_qos_mask(dev, skb); - skb = __vlan_hwaccel_put_tag(skb, veth_TCI); + vlan_tci = vlan_dev_info(dev)->vlan_id; + vlan_tci |= vlan_dev_get_egress_qos_mask(dev, skb); + skb = __vlan_hwaccel_put_tag(skb, vlan_tci); stats->tx_packets++; stats->tx_bytes += skb->len; @@ -457,7 +456,7 @@ static int vlan_dev_change_mtu(struct net_device *dev, int new_mtu) } void vlan_dev_set_ingress_priority(const struct net_device *dev, - u32 skb_prio, short vlan_prio) + u32 skb_prio, u16 vlan_prio) { struct vlan_dev_info *vlan = vlan_dev_info(dev); @@ -470,7 +469,7 @@ void vlan_dev_set_ingress_priority(const struct net_device *dev, } int vlan_dev_set_egress_priority(const struct net_device *dev, - u32 skb_prio, short vlan_prio) + u32 skb_prio, u16 vlan_prio) { struct vlan_dev_info *vlan = vlan_dev_info(dev); struct vlan_priority_tci_mapping *mp = NULL; diff --git a/net/8021q/vlan_gvrp.c b/net/8021q/vlan_gvrp.c index db9781608362..061ceceeef12 100644 --- a/net/8021q/vlan_gvrp.c +++ b/net/8021q/vlan_gvrp.c @@ -30,19 +30,19 @@ static struct garp_application vlan_gvrp_app __read_mostly = { int vlan_gvrp_request_join(const struct net_device *dev) { const struct vlan_dev_info *vlan = vlan_dev_info(dev); - __be16 vid = htons(vlan->vlan_id); + __be16 vlan_id = htons(vlan->vlan_id); return garp_request_join(vlan->real_dev, &vlan_gvrp_app, - &vid, sizeof(vid), GVRP_ATTR_VID); + &vlan_id, sizeof(vlan_id), GVRP_ATTR_VID); } void vlan_gvrp_request_leave(const struct net_device *dev) { const struct vlan_dev_info *vlan = vlan_dev_info(dev); - __be16 vid = htons(vlan->vlan_id); + __be16 vlan_id = htons(vlan->vlan_id); garp_request_leave(vlan->real_dev, &vlan_gvrp_app, - &vid, sizeof(vid), GVRP_ATTR_VID); + &vlan_id, sizeof(vlan_id), GVRP_ATTR_VID); } int vlan_gvrp_init_applicant(struct net_device *dev) -- cgit v1.2.2 From b2898a27809f54a33050a70d0eaa4a78194163a0 Mon Sep 17 00:00:00 2001 From: Ivo van Doorn Date: Tue, 1 Jul 2008 10:45:13 +0200 Subject: mac80211: Don't request encryption for probe response Probe responses shouldn't be encrypted, and mac80211 doesn't set the crypto key accordingly. However it didn't set the IEEE80211_TX_CTL_DO_NOT_ENCRYPT flag which means drivers could make an attempt to encrypt it, and causing a NULL pointer dereference when accessing the provided hw_key field. Signed-off-by: Ivo van Doorn Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 1 + 1 file changed, 1 insertion(+) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 4a3bddd206d8..86abdf96390c 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -2481,6 +2481,7 @@ static int ieee80211_sta_join_ibss(struct net_device *dev, control->flags |= IEEE80211_TX_CTL_SHORT_PREAMBLE; control->antenna_sel_tx = local->hw.conf.antenna_sel_tx; control->flags |= IEEE80211_TX_CTL_NO_ACK; + control->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; control->control.retry_limit = 1; ifsta->probe_resp = skb_copy(skb, GFP_ATOMIC); -- cgit v1.2.2 From 429a380571a6e6b8525b93161544eafc9b227e44 Mon Sep 17 00:00:00 2001 From: Ron Rindjunsky Date: Tue, 1 Jul 2008 14:16:03 +0300 Subject: mac80211: add block ack request capability This patch adds block ack request capability Signed-off-by: Ester Kummer Signed-off-by: Tomas Winkler Signed-off-by: Ron Rindjunsky Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 1 + net/mac80211/main.c | 21 +++++++++++++++++++-- net/mac80211/mlme.c | 29 +++++++++++++++++++++++++++++ 3 files changed, 49 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index f90da1bbec49..175cbdd36d7f 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -904,6 +904,7 @@ void ieee80211_send_addba_request(struct net_device *dev, const u8 *da, u16 agg_size, u16 timeout); void ieee80211_send_delba(struct net_device *dev, const u8 *da, u16 tid, u16 initiator, u16 reason_code); +void ieee80211_send_bar(struct net_device *dev, u8 *ra, u16 tid, u16 ssn); void ieee80211_sta_stop_rx_ba_session(struct net_device *dev, u8 *da, u16 tid, u16 initiator, u16 reason); diff --git a/net/mac80211/main.c b/net/mac80211/main.c index f18cfd727872..074f71a62a6d 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -1404,14 +1404,15 @@ void ieee80211_tx_status(struct ieee80211_hw *hw, struct sk_buff *skb) struct ieee80211_local *local = hw_to_local(hw); struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); u16 frag, type; + __le16 fc; struct ieee80211_tx_status_rtap_hdr *rthdr; struct ieee80211_sub_if_data *sdata; struct net_device *prev_dev = NULL; + struct sta_info *sta; rcu_read_lock(); if (info->status.excessive_retries) { - struct sta_info *sta; sta = sta_info_get(local, hdr->addr1); if (sta) { if (test_sta_flags(sta, WLAN_STA_PS)) { @@ -1426,8 +1427,24 @@ void ieee80211_tx_status(struct ieee80211_hw *hw, struct sk_buff *skb) } } + fc = hdr->frame_control; + + if ((info->flags & IEEE80211_TX_STAT_AMPDU_NO_BACK) && + (ieee80211_is_data_qos(fc))) { + u16 tid, ssn; + u8 *qc; + sta = sta_info_get(local, hdr->addr1); + if (sta) { + qc = ieee80211_get_qos_ctl(hdr); + tid = qc[0] & 0xf; + ssn = ((le16_to_cpu(hdr->seq_ctrl) + 0x10) + & IEEE80211_SCTL_SEQ); + ieee80211_send_bar(sta->sdata->dev, hdr->addr1, + tid, ssn); + } + } + if (info->flags & IEEE80211_TX_STAT_TX_FILTERED) { - struct sta_info *sta; sta = sta_info_get(local, hdr->addr1); if (sta) { ieee80211_handle_filtered_frame(local, sta, skb); diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 86abdf96390c..e080482b63cd 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -1536,6 +1536,35 @@ void ieee80211_send_delba(struct net_device *dev, const u8 *da, u16 tid, ieee80211_sta_tx(dev, skb, 0); } +void ieee80211_send_bar(struct net_device *dev, u8 *ra, u16 tid, u16 ssn) +{ + struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); + struct sk_buff *skb; + struct ieee80211_bar *bar; + u16 bar_control = 0; + + skb = dev_alloc_skb(sizeof(*bar) + local->hw.extra_tx_headroom); + if (!skb) { + printk(KERN_ERR "%s: failed to allocate buffer for " + "bar frame\n", dev->name); + return; + } + skb_reserve(skb, local->hw.extra_tx_headroom); + bar = (struct ieee80211_bar *)skb_put(skb, sizeof(*bar)); + memset(bar, 0, sizeof(*bar)); + bar->frame_control = IEEE80211_FC(IEEE80211_FTYPE_CTL, + IEEE80211_STYPE_BACK_REQ); + memcpy(bar->ra, ra, ETH_ALEN); + memcpy(bar->ta, dev->dev_addr, ETH_ALEN); + bar_control |= (u16)IEEE80211_BAR_CTRL_ACK_POLICY_NORMAL; + bar_control |= (u16)IEEE80211_BAR_CTRL_CBMTID_COMPRESSED_BA; + bar_control |= (u16)(tid << 12); + bar->control = cpu_to_le16(bar_control); + bar->start_seq_num = cpu_to_le16(ssn); + + ieee80211_sta_tx(dev, skb, 0); +} + void ieee80211_sta_stop_rx_ba_session(struct net_device *dev, u8 *ra, u16 tid, u16 initiator, u16 reason) { -- cgit v1.2.2 From d96a7bc0499d0332cecb0a1d7d7d0d44f9c8cc28 Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Tue, 1 Jul 2008 14:29:20 +0300 Subject: mac80211: remove useless tid assignment for management and control frames This patch removes useless tid assignment for management and control frames Signed-off-by: Tomas Winkler Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/wme.c | 1 - 1 file changed, 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index ffe1af82fa4d..a1a53a4f2b99 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -210,7 +210,6 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) kfree_skb(skb); err = NET_XMIT_DROP; } else { - tid = skb->priority & QOS_CONTROL_TAG1D_MASK; skb_set_queue_mapping(skb, queue); qdisc = q->queues[queue]; err = qdisc->enqueue(skb, qdisc); -- cgit v1.2.2 From ebd74487d4b7a48ab8513ecfe3d321346d7c602e Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Tue, 1 Jul 2008 10:44:50 +0300 Subject: mac80211: fix warning: unused variable ifsta This patch fixes warning unused variable ifsta when compiling without CONFIG_MAC80211_VERBOSE_DEBUG Signed-off-by: Tomas Winkler Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index e080482b63cd..2a927089f4db 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -366,8 +366,10 @@ static u32 ieee80211_handle_protect_preamb(struct ieee80211_sub_if_data *sdata, bool use_short_preamble) { struct ieee80211_bss_conf *bss_conf = &sdata->bss_conf; +#ifdef CONFIG_MAC80211_VERBOSE_DEBUG struct ieee80211_if_sta *ifsta = &sdata->u.sta; DECLARE_MAC_BUF(mac); +#endif u32 changed = 0; if (use_protection != bss_conf->use_cts_prot) { -- cgit v1.2.2 From 4e887d5b2fb909f18d153f47d2f3fdc2c76bb83a Mon Sep 17 00:00:00 2001 From: Rami Rosen Date: Wed, 2 Jul 2008 10:53:57 +0300 Subject: mac80211: remove MAC80211_DEBUG from net/mac80211/Kconfig. This patch removes MAC80211_DEBUG from /net/mac80211/Kconfig (in MAC80211_DEBUG_COUNTERS config entry), and replaces MAC80211_DEBUG_MENU instead of MAC80211_DEBUG (in MAC80211_VERBOSE_SPECT_MGMT_DEBUG config entry). Signed-off-by: Rami Rosen Signed-off-by: John W. Linville --- net/mac80211/Kconfig | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'net') diff --git a/net/mac80211/Kconfig b/net/mac80211/Kconfig index 11a1e7fa195d..40f1add17753 100644 --- a/net/mac80211/Kconfig +++ b/net/mac80211/Kconfig @@ -207,7 +207,6 @@ config MAC80211_LOWTX_FRAME_DUMP config MAC80211_DEBUG_COUNTERS bool "Extra statistics for TX/RX debugging" - depends on MAC80211_DEBUG depends on MAC80211_DEBUG_MENU depends on MAC80211_DEBUGFS ---help--- @@ -219,7 +218,7 @@ config MAC80211_DEBUG_COUNTERS config MAC80211_VERBOSE_SPECT_MGMT_DEBUG bool "Verbose Spectrum Management (IEEE 802.11h)debugging" - depends on MAC80211_DEBUG + depends on MAC80211_DEBUG_MENU ---help--- Say Y here to print out verbose Spectrum Management (IEEE 802.11h) debug messages. -- cgit v1.2.2 From 238f74a227fd7de8ea1bc66dcbbd36cf9920d1cb Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 2 Jul 2008 11:05:34 -0700 Subject: mac80211: move QOS control helpers into ieee80211.h Also remove the WLAN_IS_QOS_DATA inline after removing the last two users. This starts moving away from using rx->fc to using the header frame_control directly. Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/rx.c | 28 ++++++++++++++-------------- net/mac80211/wme.c | 4 ++-- net/mac80211/wme.h | 8 -------- 3 files changed, 16 insertions(+), 24 deletions(-) (limited to 'net') diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index 6a88e8f9bff0..a3a26e557274 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -321,20 +321,20 @@ ieee80211_rx_monitor(struct ieee80211_local *local, struct sk_buff *origskb, static void ieee80211_parse_qos(struct ieee80211_rx_data *rx) { - u8 *data = rx->skb->data; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)rx->skb->data; int tid; /* does the frame have a qos control field? */ - if (WLAN_FC_IS_QOS_DATA(rx->fc)) { - u8 *qc = data + ieee80211_get_hdrlen(rx->fc) - QOS_CONTROL_LEN; + if (ieee80211_is_data_qos(hdr->frame_control)) { + u8 *qc = ieee80211_get_qos_ctl(hdr); /* frame has qos control */ - tid = qc[0] & QOS_CONTROL_TID_MASK; - if (qc[0] & IEEE80211_QOS_CONTROL_A_MSDU_PRESENT) + tid = *qc & IEEE80211_QOS_CTL_TID_MASK; + if (*qc & IEEE80211_QOS_CONTROL_A_MSDU_PRESENT) rx->flags |= IEEE80211_RX_AMSDU; else rx->flags &= ~IEEE80211_RX_AMSDU; } else { - if (unlikely((rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_MGMT)) { + if (unlikely(ieee80211_is_mgmt(hdr->frame_control))) { /* Separate TID for management frames */ tid = NUM_RX_DATA_QUEUES - 1; } else { @@ -1037,19 +1037,19 @@ ieee80211_rx_h_ps_poll(struct ieee80211_rx_data *rx) static ieee80211_rx_result debug_noinline ieee80211_rx_h_remove_qos_control(struct ieee80211_rx_data *rx) { - u16 fc = rx->fc; u8 *data = rx->skb->data; - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) data; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)data; - if (!WLAN_FC_IS_QOS_DATA(fc)) + if (!ieee80211_is_data_qos(hdr->frame_control)) return RX_CONTINUE; /* remove the qos control field, update frame type and meta-data */ - memmove(data + 2, data, ieee80211_get_hdrlen(fc) - 2); - hdr = (struct ieee80211_hdr *) skb_pull(rx->skb, 2); + memmove(data + IEEE80211_QOS_CTL_LEN, data, + ieee80211_hdrlen(hdr->frame_control) - IEEE80211_QOS_CTL_LEN); + hdr = (struct ieee80211_hdr *)skb_pull(rx->skb, IEEE80211_QOS_CTL_LEN); /* change frame type to non QOS */ - rx->fc = fc &= ~IEEE80211_STYPE_QOS_DATA; - hdr->frame_control = cpu_to_le16(fc); + rx->fc &= ~IEEE80211_STYPE_QOS_DATA; + hdr->frame_control &= ~cpu_to_le16(IEEE80211_STYPE_QOS_DATA); return RX_CONTINUE; } @@ -2044,7 +2044,7 @@ static u8 ieee80211_rx_reorder_ampdu(struct ieee80211_local *local, if (!ieee80211_is_data_qos(hdr->frame_control)) goto end_reorder; - tid = *ieee80211_get_qos_ctl(hdr) & QOS_CONTROL_TID_MASK; + tid = *ieee80211_get_qos_ctl(hdr) & IEEE80211_QOS_CTL_TID_MASK; if (sta->ampdu_mlme.tid_state_rx[tid] != HT_AGG_STATE_OPERATIONAL) goto end_reorder; diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index a1a53a4f2b99..5c666f7eda8f 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -154,7 +154,7 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) queue = skb_get_queue_mapping(skb); rcu_read_lock(); sta = sta_info_get(local, hdr->addr1); - tid = skb->priority & QOS_CONTROL_TAG1D_MASK; + tid = skb->priority & IEEE80211_QOS_CTL_TAG1D_MASK; if (sta) { int ampdu_queue = sta->tid_to_tx_q[tid]; if ((ampdu_queue < QD_NUM(hw)) && @@ -181,7 +181,7 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) if (ieee80211_is_data_qos(hdr->frame_control)) { u8 *p = ieee80211_get_qos_ctl(hdr); u8 ack_policy = 0; - tid = skb->priority & QOS_CONTROL_TAG1D_MASK; + tid = skb->priority & IEEE80211_QOS_CTL_TAG1D_MASK; if (local->wifi_wme_noack_test) ack_policy |= QOS_CONTROL_ACK_POLICY_NOACK << QOS_CONTROL_ACK_POLICY_SHIFT; diff --git a/net/mac80211/wme.h b/net/mac80211/wme.h index bbdb53344817..1aca609eccfc 100644 --- a/net/mac80211/wme.h +++ b/net/mac80211/wme.h @@ -19,18 +19,10 @@ #define QOS_CONTROL_ACK_POLICY_NORMAL 0 #define QOS_CONTROL_ACK_POLICY_NOACK 1 -#define QOS_CONTROL_TID_MASK 0x0f #define QOS_CONTROL_ACK_POLICY_SHIFT 5 -#define QOS_CONTROL_TAG1D_MASK 0x07 - extern const int ieee802_1d_to_ac[8]; -static inline int WLAN_FC_IS_QOS_DATA(u16 fc) -{ - return (fc & 0x8C) == 0x88; -} - #ifdef CONFIG_MAC80211_QOS void ieee80211_install_qdisc(struct net_device *dev); int ieee80211_qdisc_installed(struct net_device *dev); -- cgit v1.2.2 From 73e1f7c823252d671361cddde0b498bf2c0fe4e1 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 2 Jul 2008 11:05:34 -0700 Subject: mac80211: use symbolic defines in wpa.c ETH_ALEN and IEEE80211_QOS_CTL_LEN Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/wpa.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index b414d5d92f38..222001c47d89 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -38,7 +38,7 @@ static int ieee80211_get_hdr_info(const struct sk_buff *skb, u8 **sa, u8 **da, *data_len = skb->len - hdrlen; if (ieee80211_is_data_qos(fc)) - *qos_tid = (*ieee80211_get_qos_ctl(hdr) & 0x0f) | 0x80; + *qos_tid = (*ieee80211_get_qos_ctl(hdr) & IEEE80211_QOS_CTL_TID_MASK) | 0x80; else *qos_tid = 0; @@ -312,7 +312,7 @@ static void ccmp_special_blocks(struct sk_buff *skb, u8 *pn, u8 *b_0, u8 *aad, data_len -= CCMP_HDR_LEN + (encrypted ? CCMP_MIC_LEN : 0); if (qos_tid & 0x80) { qos_included = 1; - qos_tid &= 0x0f; + qos_tid &= IEEE80211_QOS_CTL_TID_MASK; } else qos_included = 0; /* First block, b_0 */ @@ -320,7 +320,7 @@ static void ccmp_special_blocks(struct sk_buff *skb, u8 *pn, u8 *b_0, u8 *aad, b_0[0] = 0x59; /* flags: Adata: 1, M: 011, L: 001 */ /* Nonce: QoS Priority | A2 | PN */ b_0[1] = qos_tid; - memcpy(&b_0[2], hdr->addr2, 6); + memcpy(&b_0[2], hdr->addr2, ETH_ALEN); memcpy(&b_0[8], pn, CCMP_PN_LEN); /* l(m) */ b_0[14] = (data_len >> 8) & 0xff; @@ -332,7 +332,7 @@ static void ccmp_special_blocks(struct sk_buff *skb, u8 *pn, u8 *b_0, u8 *aad, len_a = a4_included ? 28 : 22; if (qos_included) - len_a += 2; + len_a += IEEE80211_QOS_CTL_LEN; aad[0] = 0; /* (len_a >> 8) & 0xff; */ aad[1] = len_a & 0xff; @@ -340,17 +340,17 @@ static void ccmp_special_blocks(struct sk_buff *skb, u8 *pn, u8 *b_0, u8 *aad, aad[2] = fc_pos[0] & ~(BIT(4) | BIT(5) | BIT(6)); /* Retry, PwrMgt, MoreData; set Protected */ aad[3] = (fc_pos[1] & ~(BIT(3) | BIT(4) | BIT(5))) | BIT(6); - memcpy(&aad[4], &hdr->addr1, 18); + memcpy(&aad[4], &hdr->addr1, 3 * ETH_ALEN); /* Mask Seq#, leave Frag# */ aad[22] = *((u8 *) &hdr->seq_ctrl) & 0x0f; aad[23] = 0; if (a4_included) { - memcpy(&aad[24], hdr->addr4, 6); + memcpy(&aad[24], hdr->addr4, ETH_ALEN); aad[30] = 0; aad[31] = 0; } else - memset(&aad[24], 0, 8); + memset(&aad[24], 0, ETH_ALEN + IEEE80211_QOS_CTL_LEN); if (qos_included) { u8 *dpos = &aad[a4_included ? 30 : 24]; -- cgit v1.2.2 From f14df8049f9c9f34164dd598772fecea83a394a2 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 2 Jul 2008 11:05:35 -0700 Subject: mac80211: remove one user of ieee80211_get_hdr_info ccmp_special_blocks was only using it to calculate data_len, calculate that directly. Use unaligned helpers rather than masking/shifting. Use symbolic constants for the masked frame_control, and do it directly on a le16 value. Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/wpa.c | 78 ++++++++++++++++++++++++++---------------------------- 1 file changed, 37 insertions(+), 41 deletions(-) (limited to 'net') diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index 222001c47d89..919e30ce2980 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -11,6 +11,8 @@ #include #include #include +#include +#include #include #include "ieee80211_i.h" @@ -296,67 +298,61 @@ ieee80211_crypto_tkip_decrypt(struct ieee80211_rx_data *rx) static void ccmp_special_blocks(struct sk_buff *skb, u8 *pn, u8 *b_0, u8 *aad, int encrypted) { - u16 fc; - int a4_included, qos_included; - u8 qos_tid, *fc_pos, *data, *sa, *da; - int len_a; - size_t data_len; - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; + __le16 mask_fc; + int a4_included; + u8 qos_tid; + u16 data_len, len_a; + unsigned int hdrlen; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data; - fc_pos = (u8 *) &hdr->frame_control; - fc = fc_pos[0] ^ (fc_pos[1] << 8); - a4_included = (fc & (IEEE80211_FCTL_TODS | IEEE80211_FCTL_FROMDS)) == - (IEEE80211_FCTL_TODS | IEEE80211_FCTL_FROMDS); - - ieee80211_get_hdr_info(skb, &sa, &da, &qos_tid, &data, &data_len); - data_len -= CCMP_HDR_LEN + (encrypted ? CCMP_MIC_LEN : 0); - if (qos_tid & 0x80) { - qos_included = 1; - qos_tid &= IEEE80211_QOS_CTL_TID_MASK; - } else - qos_included = 0; - /* First block, b_0 */ + /* + * Mask FC: zero subtype b4 b5 b6 + * Retry, PwrMgt, MoreData; set Protected + */ + mask_fc = hdr->frame_control; + mask_fc &= ~cpu_to_le16(0x0070 | IEEE80211_FCTL_RETRY | + IEEE80211_FCTL_PM | IEEE80211_FCTL_MOREDATA); + mask_fc |= cpu_to_le16(IEEE80211_FCTL_PROTECTED); + + hdrlen = ieee80211_hdrlen(hdr->frame_control); + len_a = hdrlen - 2; + a4_included = ieee80211_has_a4(hdr->frame_control); + if (ieee80211_is_data_qos(hdr->frame_control)) + qos_tid = *ieee80211_get_qos_ctl(hdr) & IEEE80211_QOS_CTL_TID_MASK; + else + qos_tid = 0; + + data_len = skb->len - hdrlen - CCMP_HDR_LEN; + if (encrypted) + data_len -= CCMP_MIC_LEN; + + /* First block, b_0 */ b_0[0] = 0x59; /* flags: Adata: 1, M: 011, L: 001 */ /* Nonce: QoS Priority | A2 | PN */ b_0[1] = qos_tid; memcpy(&b_0[2], hdr->addr2, ETH_ALEN); memcpy(&b_0[8], pn, CCMP_PN_LEN); /* l(m) */ - b_0[14] = (data_len >> 8) & 0xff; - b_0[15] = data_len & 0xff; - + put_unaligned_be16(data_len, &b_0[14]); /* AAD (extra authenticate-only data) / masked 802.11 header * FC | A1 | A2 | A3 | SC | [A4] | [QC] */ - - len_a = a4_included ? 28 : 22; - if (qos_included) - len_a += IEEE80211_QOS_CTL_LEN; - - aad[0] = 0; /* (len_a >> 8) & 0xff; */ - aad[1] = len_a & 0xff; - /* Mask FC: zero subtype b4 b5 b6 */ - aad[2] = fc_pos[0] & ~(BIT(4) | BIT(5) | BIT(6)); - /* Retry, PwrMgt, MoreData; set Protected */ - aad[3] = (fc_pos[1] & ~(BIT(3) | BIT(4) | BIT(5))) | BIT(6); + put_unaligned_be16(len_a, &aad[0]); + put_unaligned(mask_fc, (__le16 *)&aad[2]); memcpy(&aad[4], &hdr->addr1, 3 * ETH_ALEN); /* Mask Seq#, leave Frag# */ aad[22] = *((u8 *) &hdr->seq_ctrl) & 0x0f; aad[23] = 0; + if (a4_included) { memcpy(&aad[24], hdr->addr4, ETH_ALEN); - aad[30] = 0; + aad[30] = qos_tid; aad[31] = 0; - } else + } else { memset(&aad[24], 0, ETH_ALEN + IEEE80211_QOS_CTL_LEN); - if (qos_included) { - u8 *dpos = &aad[a4_included ? 30 : 24]; - - /* Mask QoS Control field */ - dpos[0] = qos_tid; - dpos[1] = 0; + aad[24] = qos_tid; } } -- cgit v1.2.2 From 8e8862b79d2ce9177bfddd85b8328a86a25c69b2 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 2 Jul 2008 11:05:35 -0700 Subject: mac80211: remove ieee80211_get_hdr_info Do the check for sufficient skb->len explicitly and pass a pointer to the struct ieee80211_hdr directly to the michael_mic calculation. Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/michael.c | 20 ++++++++++++++----- net/mac80211/michael.h | 2 +- net/mac80211/wpa.c | 54 ++++++++++++++++++-------------------------------- 3 files changed, 35 insertions(+), 41 deletions(-) (limited to 'net') diff --git a/net/mac80211/michael.c b/net/mac80211/michael.c index 1fcdf38cf60c..408649bd4702 100644 --- a/net/mac80211/michael.c +++ b/net/mac80211/michael.c @@ -8,6 +8,7 @@ */ #include #include +#include #include #include "michael.h" @@ -26,9 +27,18 @@ static void michael_block(struct michael_mic_ctx *mctx, u32 val) mctx->l += mctx->r; } -static void michael_mic_hdr(struct michael_mic_ctx *mctx, - const u8 *key, const u8 *da, const u8 *sa, u8 priority) +static void michael_mic_hdr(struct michael_mic_ctx *mctx, const u8 *key, + struct ieee80211_hdr *hdr) { + u8 *da, *sa, tid; + + da = ieee80211_get_DA(hdr); + sa = ieee80211_get_SA(hdr); + if (ieee80211_is_data_qos(hdr->frame_control)) + tid = *ieee80211_get_qos_ctl(hdr) & IEEE80211_QOS_CTL_TID_MASK; + else + tid = 0; + mctx->l = get_unaligned_le32(key); mctx->r = get_unaligned_le32(key + 4); @@ -40,17 +50,17 @@ static void michael_mic_hdr(struct michael_mic_ctx *mctx, michael_block(mctx, get_unaligned_le16(&da[4]) | (get_unaligned_le16(sa) << 16)); michael_block(mctx, get_unaligned_le32(&sa[2])); - michael_block(mctx, priority); + michael_block(mctx, tid); } -void michael_mic(const u8 *key, const u8 *da, const u8 *sa, u8 priority, +void michael_mic(const u8 *key, struct ieee80211_hdr *hdr, const u8 *data, size_t data_len, u8 *mic) { u32 val; size_t block, blocks, left; struct michael_mic_ctx mctx; - michael_mic_hdr(&mctx, key, da, sa, priority); + michael_mic_hdr(&mctx, key, hdr); /* Real data */ blocks = data_len / 4; diff --git a/net/mac80211/michael.h b/net/mac80211/michael.h index 69b4501f13ba..3b848dad9587 100644 --- a/net/mac80211/michael.h +++ b/net/mac80211/michael.h @@ -18,7 +18,7 @@ struct michael_mic_ctx { u32 l, r; }; -void michael_mic(const u8 *key, const u8 *da, const u8 *sa, u8 priority, +void michael_mic(const u8 *key, struct ieee80211_hdr *hdr, const u8 *data, size_t data_len, u8 *mic); #endif /* MICHAEL_H */ diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index 919e30ce2980..241c932d3b6c 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -21,38 +21,13 @@ #include "aes_ccm.h" #include "wpa.h" -static int ieee80211_get_hdr_info(const struct sk_buff *skb, u8 **sa, u8 **da, - u8 *qos_tid, u8 **data, size_t *data_len) -{ - struct ieee80211_hdr *hdr; - size_t hdrlen; - __le16 fc; - - hdr = (struct ieee80211_hdr *)skb->data; - fc = hdr->frame_control; - - hdrlen = ieee80211_hdrlen(fc); - - *sa = ieee80211_get_SA(hdr); - *da = ieee80211_get_DA(hdr); - - *data = skb->data + hdrlen; - *data_len = skb->len - hdrlen; - - if (ieee80211_is_data_qos(fc)) - *qos_tid = (*ieee80211_get_qos_ctl(hdr) & IEEE80211_QOS_CTL_TID_MASK) | 0x80; - else - *qos_tid = 0; - - return skb->len < hdrlen ? -1 : 0; -} - - ieee80211_tx_result ieee80211_tx_h_michael_mic_add(struct ieee80211_tx_data *tx) { - u8 *data, *sa, *da, *key, *mic, qos_tid, key_offset; + u8 *data, *key, *mic, key_offset; size_t data_len; + unsigned int hdrlen; + struct ieee80211_hdr *hdr; u16 fc; struct sk_buff *skb = tx->skb; int authenticator; @@ -65,9 +40,14 @@ ieee80211_tx_h_michael_mic_add(struct ieee80211_tx_data *tx) !WLAN_FC_DATA_PRESENT(fc)) return TX_CONTINUE; - if (ieee80211_get_hdr_info(skb, &sa, &da, &qos_tid, &data, &data_len)) + hdr = (struct ieee80211_hdr *)skb->data; + hdrlen = ieee80211_hdrlen(hdr->frame_control); + if (skb->len < hdrlen) return TX_DROP; + data = skb->data + hdrlen; + data_len = skb->len - hdrlen; + if ((tx->key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) && !(tx->flags & IEEE80211_TX_FRAGMENTED) && !(tx->key->conf.flags & IEEE80211_KEY_FLAG_GENERATE_MMIC) && @@ -97,7 +77,7 @@ ieee80211_tx_h_michael_mic_add(struct ieee80211_tx_data *tx) NL80211_TKIP_DATA_OFFSET_RX_MIC_KEY; key = &tx->key->conf.key[key_offset]; mic = skb_put(skb, MICHAEL_MIC_LEN); - michael_mic(key, da, sa, qos_tid & 0x0f, data, data_len, mic); + michael_mic(key, hdr, data, data_len, mic); return TX_CONTINUE; } @@ -106,8 +86,10 @@ ieee80211_tx_h_michael_mic_add(struct ieee80211_tx_data *tx) ieee80211_rx_result ieee80211_rx_h_michael_mic_verify(struct ieee80211_rx_data *rx) { - u8 *data, *sa, *da, *key = NULL, qos_tid, key_offset; + u8 *data, *key = NULL, key_offset; size_t data_len; + unsigned int hdrlen; + struct ieee80211_hdr *hdr; u16 fc; u8 mic[MICHAEL_MIC_LEN]; struct sk_buff *skb = rx->skb; @@ -126,11 +108,13 @@ ieee80211_rx_h_michael_mic_verify(struct ieee80211_rx_data *rx) !(rx->fc & IEEE80211_FCTL_PROTECTED) || !WLAN_FC_DATA_PRESENT(fc)) return RX_CONTINUE; - if (ieee80211_get_hdr_info(skb, &sa, &da, &qos_tid, &data, &data_len) - || data_len < MICHAEL_MIC_LEN) + hdr = (struct ieee80211_hdr *)skb->data; + hdrlen = ieee80211_hdrlen(hdr->frame_control); + if (skb->len < hdrlen + MICHAEL_MIC_LEN) return RX_DROP_UNUSABLE; - data_len -= MICHAEL_MIC_LEN; + data = skb->data + hdrlen; + data_len = skb->len - hdrlen - MICHAEL_MIC_LEN; #if 0 authenticator = fc & IEEE80211_FCTL_TODS; /* FIX */ @@ -143,7 +127,7 @@ ieee80211_rx_h_michael_mic_verify(struct ieee80211_rx_data *rx) NL80211_TKIP_DATA_OFFSET_RX_MIC_KEY : NL80211_TKIP_DATA_OFFSET_TX_MIC_KEY; key = &rx->key->conf.key[key_offset]; - michael_mic(key, da, sa, qos_tid & 0x0f, data, data_len, mic); + michael_mic(key, hdr, data, data_len, mic); if (memcmp(mic, data + data_len, MICHAEL_MIC_LEN) != 0 || wpa_test) { if (!(rx->flags & IEEE80211_RX_RA_MATCH)) return RX_DROP_UNUSABLE; -- cgit v1.2.2 From fc32f9243dc93e75f81457e95d9cb90ee6136d94 Mon Sep 17 00:00:00 2001 From: Tomas Winkler Date: Thu, 3 Jul 2008 01:27:13 +0300 Subject: mac80211: call bss_info_change only once upon disassociation This patch removes call of ieee80211_bss_info_change_notify from within ieee80211_reset_erp_info. This allows gathering all bss info changes into one call to the driver in the disassociation flow. Signed-off-by: Tomas Winkler Acked-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 2 +- net/mac80211/main.c | 12 ++++++------ net/mac80211/mlme.c | 2 +- 3 files changed, 8 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 175cbdd36d7f..02a8753a4eca 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -893,7 +893,7 @@ int ieee80211_sta_deauthenticate(struct net_device *dev, u16 reason); int ieee80211_sta_disassociate(struct net_device *dev, u16 reason); void ieee80211_bss_info_change_notify(struct ieee80211_sub_if_data *sdata, u32 changed); -void ieee80211_reset_erp_info(struct net_device *dev); +u32 ieee80211_reset_erp_info(struct net_device *dev); int ieee80211_ht_cap_ie_to_ht_info(struct ieee80211_ht_cap *ht_cap_ie, struct ieee80211_ht_info *ht_info); int ieee80211_ht_addt_info_ie_to_ht_bss_info( diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 074f71a62a6d..cc756e93e6c8 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -182,10 +182,11 @@ static int ieee80211_open(struct net_device *dev) { struct ieee80211_sub_if_data *sdata, *nsdata; struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); + struct sta_info *sta; struct ieee80211_if_init_conf conf; + u32 changed = 0; int res; bool need_hw_reconfig = 0; - struct sta_info *sta; sdata = IEEE80211_DEV_TO_SUB_IF(dev); @@ -329,7 +330,8 @@ static int ieee80211_open(struct net_device *dev) goto err_stop; ieee80211_if_config(dev); - ieee80211_reset_erp_info(dev); + changed |= ieee80211_reset_erp_info(dev); + ieee80211_bss_info_change_notify(sdata, changed); ieee80211_enable_keys(sdata); if (sdata->vif.type == IEEE80211_IF_TYPE_STA && @@ -1190,15 +1192,13 @@ void ieee80211_bss_info_change_notify(struct ieee80211_sub_if_data *sdata, changed); } -void ieee80211_reset_erp_info(struct net_device *dev) +u32 ieee80211_reset_erp_info(struct net_device *dev) { struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); sdata->bss_conf.use_cts_prot = 0; sdata->bss_conf.use_short_preamble = 0; - ieee80211_bss_info_change_notify(sdata, - BSS_CHANGED_ERP_CTS_PROT | - BSS_CHANGED_ERP_PREAMBLE); + return BSS_CHANGED_ERP_CTS_PROT | BSS_CHANGED_ERP_PREAMBLE; } void ieee80211_tx_status_irqsafe(struct ieee80211_hw *hw, diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 2a927089f4db..37ea04f5bab9 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -573,7 +573,7 @@ static void ieee80211_set_associated(struct net_device *dev, ieee80211_sta_tear_down_BA_sessions(dev, ifsta->bssid); ifsta->flags &= ~IEEE80211_STA_ASSOCIATED; netif_carrier_off(dev); - ieee80211_reset_erp_info(dev); + changed |= ieee80211_reset_erp_info(dev); sdata->bss_conf.assoc_ht = 0; sdata->bss_conf.ht_conf = NULL; -- cgit v1.2.2 From a7767f958a3b09a1bcd0ddcb21ef6f9a8ebd782c Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 2 Jul 2008 16:30:51 -0700 Subject: mac80211: remove trivial rx_data->fc users Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/rx.c | 107 ++++++++++++++++++++++++++---------------------------- 1 file changed, 51 insertions(+), 56 deletions(-) (limited to 'net') diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index a3a26e557274..fab443d717eb 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -352,9 +352,10 @@ static void ieee80211_parse_qos(struct ieee80211_rx_data *rx) static void ieee80211_verify_ip_alignment(struct ieee80211_rx_data *rx) { #ifdef CONFIG_MAC80211_DEBUG_PACKET_ALIGNMENT + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)rx->skb->data; int hdrlen; - if (!WLAN_FC_DATA_PRESENT(rx->fc)) + if (!ieee80211_is_data_present(hdr->frame_control)) return; /* @@ -376,7 +377,7 @@ static void ieee80211_verify_ip_alignment(struct ieee80211_rx_data *rx) * header and the payload is not supported, the driver is required * to move the 802.11 header further back in that case. */ - hdrlen = ieee80211_get_hdrlen(rx->fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); if (rx->flags & IEEE80211_RX_AMSDU) hdrlen += ETH_HLEN; WARN_ON_ONCE(((unsigned long)(rx->skb->data + hdrlen)) & 3); @@ -415,14 +416,11 @@ ieee80211_rx_h_passive_scan(struct ieee80211_rx_data *rx) static ieee80211_rx_result ieee80211_rx_mesh_check(struct ieee80211_rx_data *rx) { - int hdrlen = ieee80211_get_hdrlen(rx->fc); - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) rx->skb->data; - -#define msh_h_get(h, l) ((struct ieee80211s_hdr *) ((u8 *)h + l)) + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)rx->skb->data; + unsigned int hdrlen = ieee80211_hdrlen(hdr->frame_control); - if ((rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA) { - if (!((rx->fc & IEEE80211_FCTL_FROMDS) && - (rx->fc & IEEE80211_FCTL_TODS))) + if (ieee80211_is_data(hdr->frame_control)) { + if (!ieee80211_has_a4(hdr->frame_control)) return RX_DROP_MONITOR; if (memcmp(hdr->addr4, rx->dev->dev_addr, ETH_ALEN) == 0) return RX_DROP_MONITOR; @@ -435,27 +433,30 @@ ieee80211_rx_mesh_check(struct ieee80211_rx_data *rx) if (!rx->sta || sta_plink_state(rx->sta) != PLINK_ESTAB) { struct ieee80211_mgmt *mgmt; - if ((rx->fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_MGMT) + if (!ieee80211_is_mgmt(hdr->frame_control)) return RX_DROP_MONITOR; - switch (rx->fc & IEEE80211_FCTL_STYPE) { - case IEEE80211_STYPE_ACTION: + if (ieee80211_is_action(hdr->frame_control)) { mgmt = (struct ieee80211_mgmt *)hdr; if (mgmt->u.action.category != PLINK_CATEGORY) return RX_DROP_MONITOR; - /* fall through on else */ - case IEEE80211_STYPE_PROBE_REQ: - case IEEE80211_STYPE_PROBE_RESP: - case IEEE80211_STYPE_BEACON: return RX_CONTINUE; - break; - default: - return RX_DROP_MONITOR; } - } else if ((rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA && - is_multicast_ether_addr(hdr->addr1) && - mesh_rmc_check(hdr->addr4, msh_h_get(hdr, hdrlen), rx->dev)) + if (ieee80211_is_probe_req(hdr->frame_control) || + ieee80211_is_probe_resp(hdr->frame_control) || + ieee80211_is_beacon(hdr->frame_control)) + return RX_CONTINUE; + + return RX_DROP_MONITOR; + + } + +#define msh_h_get(h, l) ((struct ieee80211s_hdr *) ((u8 *)h + l)) + + if (ieee80211_is_data(hdr->frame_control) && + is_multicast_ether_addr(hdr->addr1) && + mesh_rmc_check(hdr->addr4, msh_h_get(hdr, hdrlen), rx->dev)) return RX_DROP_MONITOR; #undef msh_h_get @@ -466,13 +467,11 @@ ieee80211_rx_mesh_check(struct ieee80211_rx_data *rx) static ieee80211_rx_result debug_noinline ieee80211_rx_h_check(struct ieee80211_rx_data *rx) { - struct ieee80211_hdr *hdr; - - hdr = (struct ieee80211_hdr *) rx->skb->data; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)rx->skb->data; /* Drop duplicate 802.11 retransmissions (IEEE 802.11 Chap. 9.2.9) */ if (rx->sta && !is_multicast_ether_addr(hdr->addr1)) { - if (unlikely(rx->fc & IEEE80211_FCTL_RETRY && + if (unlikely(ieee80211_has_retry(hdr->frame_control) && rx->sta->last_seq_ctrl[rx->queue] == hdr->seq_ctrl)) { if (rx->flags & IEEE80211_RX_RA_MATCH) { @@ -501,15 +500,14 @@ ieee80211_rx_h_check(struct ieee80211_rx_data *rx) if (ieee80211_vif_is_mesh(&rx->sdata->vif)) return ieee80211_rx_mesh_check(rx); - if (unlikely(((rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA || - ((rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_CTL && - (rx->fc & IEEE80211_FCTL_STYPE) == IEEE80211_STYPE_PSPOLL)) && + if (unlikely((ieee80211_is_data(hdr->frame_control) || + ieee80211_is_pspoll(hdr->frame_control)) && rx->sdata->vif.type != IEEE80211_IF_TYPE_IBSS && (!rx->sta || !test_sta_flags(rx->sta, WLAN_STA_ASSOC)))) { - if ((!(rx->fc & IEEE80211_FCTL_FROMDS) && - !(rx->fc & IEEE80211_FCTL_TODS) && - (rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA) - || !(rx->flags & IEEE80211_RX_RA_MATCH)) { + if ((!ieee80211_has_fromds(hdr->frame_control) && + !ieee80211_has_tods(hdr->frame_control) && + ieee80211_is_data(hdr->frame_control)) || + !(rx->flags & IEEE80211_RX_RA_MATCH)) { /* Drop IBSS frames and frames for other hosts * silently. */ return RX_DROP_MONITOR; @@ -525,7 +523,7 @@ ieee80211_rx_h_check(struct ieee80211_rx_data *rx) static ieee80211_rx_result debug_noinline ieee80211_rx_h_decrypt(struct ieee80211_rx_data *rx) { - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) rx->skb->data; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)rx->skb->data; int keyidx; int hdrlen; ieee80211_rx_result result = RX_DROP_UNUSABLE; @@ -557,7 +555,7 @@ ieee80211_rx_h_decrypt(struct ieee80211_rx_data *rx) * possible. */ - if (!(rx->fc & IEEE80211_FCTL_PROTECTED)) + if (!ieee80211_has_protected(hdr->frame_control)) return RX_CONTINUE; /* @@ -586,7 +584,7 @@ ieee80211_rx_h_decrypt(struct ieee80211_rx_data *rx) (rx->status->flag & RX_FLAG_IV_STRIPPED)) return RX_CONTINUE; - hdrlen = ieee80211_get_hdrlen(rx->fc); + hdrlen = ieee80211_hdrlen(hdr->frame_control); if (rx->skb->len < 8 + hdrlen) return RX_DROP_UNUSABLE; /* TODO: count this? */ @@ -618,7 +616,7 @@ ieee80211_rx_h_decrypt(struct ieee80211_rx_data *rx) /* Check for weak IVs if possible */ if (rx->sta && rx->key->conf.alg == ALG_WEP && - ((rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA) && + ieee80211_is_data(hdr->frame_control) && (!(rx->status->flag & RX_FLAG_IV_STRIPPED) || !(rx->status->flag & RX_FLAG_DECRYPTED)) && ieee80211_wep_is_weak_iv(rx->skb, rx->key)) @@ -710,7 +708,7 @@ ieee80211_rx_h_sta_process(struct ieee80211_rx_data *rx) { struct sta_info *sta = rx->sta; struct net_device *dev = rx->dev; - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) rx->skb->data; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)rx->skb->data; if (!sta) return RX_CONTINUE; @@ -744,21 +742,20 @@ ieee80211_rx_h_sta_process(struct ieee80211_rx_data *rx) sta->last_qual = rx->status->qual; sta->last_noise = rx->status->noise; - if (!(rx->fc & IEEE80211_FCTL_MOREFRAGS)) { + if (!ieee80211_has_morefrags(hdr->frame_control)) { /* Change STA power saving mode only in the end of a frame * exchange sequence */ if (test_sta_flags(sta, WLAN_STA_PS) && - !(rx->fc & IEEE80211_FCTL_PM)) + !ieee80211_has_pm(hdr->frame_control)) rx->sent_ps_buffered += ap_sta_ps_end(dev, sta); else if (!test_sta_flags(sta, WLAN_STA_PS) && - (rx->fc & IEEE80211_FCTL_PM)) + ieee80211_has_pm(hdr->frame_control)) ap_sta_ps_start(dev, sta); } /* Drop data::nullfunc frames silently, since they are used only to * control station power saving mode. */ - if ((rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_DATA && - (rx->fc & IEEE80211_FCTL_STYPE) == IEEE80211_STYPE_NULLFUNC) { + if (ieee80211_is_nullfunc(hdr->frame_control)) { I802_DEBUG_INC(rx->local->rx_handlers_drop_nullfunc); /* Update counter and free packet here to avoid counting this * as a dropped packed. */ @@ -1465,15 +1462,15 @@ ieee80211_rx_h_ctrl(struct ieee80211_rx_data *rx) struct ieee80211_local *local = rx->local; struct ieee80211_hw *hw = &local->hw; struct sk_buff *skb = rx->skb; - struct ieee80211_bar *bar = (struct ieee80211_bar *) skb->data; + struct ieee80211_bar *bar = (struct ieee80211_bar *)skb->data; struct tid_ampdu_rx *tid_agg_rx; u16 start_seq_num; u16 tid; - if (likely((rx->fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_CTL)) + if (likely(!ieee80211_is_ctl(bar->frame_control))) return RX_CONTINUE; - if ((rx->fc & IEEE80211_FCTL_STYPE) == IEEE80211_STYPE_BACK_REQ) { + if (ieee80211_is_back_req(bar->frame_control)) { if (!rx->sta) return RX_CONTINUE; tid = le16_to_cpu(bar->control) >> 12; @@ -1527,11 +1524,12 @@ static void ieee80211_rx_michael_mic_report(struct net_device *dev, struct ieee80211_hdr *hdr, struct ieee80211_rx_data *rx) { - int keyidx, hdrlen; + int keyidx; + unsigned int hdrlen; DECLARE_MAC_BUF(mac); DECLARE_MAC_BUF(mac2); - hdrlen = ieee80211_get_hdrlen_from_skb(rx->skb); + hdrlen = ieee80211_hdrlen(hdr->frame_control); if (rx->skb->len >= hdrlen + 4) keyidx = rx->skb->data[hdrlen + 3] >> 6; else @@ -1545,7 +1543,7 @@ static void ieee80211_rx_michael_mic_report(struct net_device *dev, goto ignore; } - if (!(rx->fc & IEEE80211_FCTL_PROTECTED)) + if (!ieee80211_has_protected(hdr->frame_control)) goto ignore; if (rx->sdata->vif.type == IEEE80211_IF_TYPE_AP && keyidx) { @@ -1558,9 +1556,8 @@ static void ieee80211_rx_michael_mic_report(struct net_device *dev, goto ignore; } - if ((rx->fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_DATA && - ((rx->fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_MGMT || - (rx->fc & IEEE80211_FCTL_STYPE) != IEEE80211_STYPE_AUTH)) + if (!ieee80211_is_data(hdr->frame_control) && + !ieee80211_is_auth(hdr->frame_control)) goto ignore; mac80211_ev_michael_mic_failure(rx->dev, keyidx, hdr); @@ -1731,8 +1728,7 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata, case IEEE80211_IF_TYPE_IBSS: if (!bssid) return 0; - if ((rx->fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_MGMT && - (rx->fc & IEEE80211_FCTL_STYPE) == IEEE80211_STYPE_BEACON) { + if (ieee80211_is_beacon(hdr->frame_control)) { if (!rx->sta) rx->sta = ieee80211_ibss_add_sta(sdata->dev, rx->skb, bssid, hdr->addr2, @@ -1783,8 +1779,7 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata, return 0; break; case IEEE80211_IF_TYPE_WDS: - if (bssid || - (rx->fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_DATA) + if (bssid || !ieee80211_is_data(hdr->frame_control)) return 0; if (compare_ether_addr(sdata->u.wds.remote_addr, hdr->addr2)) return 0; -- cgit v1.2.2 From c34498b9e633baa3266af98106502633b6bc371b Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 2 Jul 2008 16:30:52 -0700 Subject: mac80211: wpa.c remove rx/tx_data ->fc users Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/wpa.c | 21 ++++++++------------- 1 file changed, 8 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index 241c932d3b6c..ec2ae86d64aa 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -28,19 +28,16 @@ ieee80211_tx_h_michael_mic_add(struct ieee80211_tx_data *tx) size_t data_len; unsigned int hdrlen; struct ieee80211_hdr *hdr; - u16 fc; struct sk_buff *skb = tx->skb; int authenticator; int wpa_test = 0; int tail; - fc = tx->fc; - + hdr = (struct ieee80211_hdr *)skb->data; if (!tx->key || tx->key->conf.alg != ALG_TKIP || skb->len < 24 || - !WLAN_FC_DATA_PRESENT(fc)) + !ieee80211_is_data_present(hdr->frame_control)) return TX_CONTINUE; - hdr = (struct ieee80211_hdr *)skb->data; hdrlen = ieee80211_hdrlen(hdr->frame_control); if (skb->len < hdrlen) return TX_DROP; @@ -90,25 +87,23 @@ ieee80211_rx_h_michael_mic_verify(struct ieee80211_rx_data *rx) size_t data_len; unsigned int hdrlen; struct ieee80211_hdr *hdr; - u16 fc; u8 mic[MICHAEL_MIC_LEN]; struct sk_buff *skb = rx->skb; int authenticator = 1, wpa_test = 0; DECLARE_MAC_BUF(mac); - fc = rx->fc; - /* * No way to verify the MIC if the hardware stripped it */ if (rx->status->flag & RX_FLAG_MMIC_STRIPPED) return RX_CONTINUE; + hdr = (struct ieee80211_hdr *)skb->data; if (!rx->key || rx->key->conf.alg != ALG_TKIP || - !(rx->fc & IEEE80211_FCTL_PROTECTED) || !WLAN_FC_DATA_PRESENT(fc)) + !ieee80211_has_protected(hdr->frame_control) || + !ieee80211_is_data_present(hdr->frame_control)) return RX_CONTINUE; - hdr = (struct ieee80211_hdr *)skb->data; hdrlen = ieee80211_hdrlen(hdr->frame_control); if (skb->len < hdrlen + MICHAEL_MIC_LEN) return RX_DROP_UNUSABLE; @@ -239,7 +234,7 @@ ieee80211_crypto_tkip_decrypt(struct ieee80211_rx_data *rx) hdrlen = ieee80211_hdrlen(hdr->frame_control); - if ((rx->fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_DATA) + if (!ieee80211_is_data(hdr->frame_control)) return RX_CONTINUE; if (!rx->sta || skb->len - hdrlen < 12) @@ -458,7 +453,7 @@ ieee80211_crypto_ccmp_encrypt(struct ieee80211_tx_data *tx) ieee80211_rx_result ieee80211_crypto_ccmp_decrypt(struct ieee80211_rx_data *rx) { - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) rx->skb->data; + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)rx->skb->data; int hdrlen; struct ieee80211_key *key = rx->key; struct sk_buff *skb = rx->skb; @@ -468,7 +463,7 @@ ieee80211_crypto_ccmp_decrypt(struct ieee80211_rx_data *rx) hdrlen = ieee80211_hdrlen(hdr->frame_control); - if ((rx->fc & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_DATA) + if (!ieee80211_is_data(hdr->frame_control)) return RX_CONTINUE; data_len = skb->len - hdrlen - CCMP_HDR_LEN - CCMP_MIC_LEN; -- cgit v1.2.2 From feccb466944cb1aeaabc701cfde6771f3be74919 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 2 Jul 2008 16:30:52 -0700 Subject: mac80211: pass scratch buffer directly, remove additional pointers Recalculate the offset pointers in the ccmp calculations rather than in the callers. Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/aes_ccm.c | 13 ++++++++----- net/mac80211/aes_ccm.h | 4 ++-- net/mac80211/wpa.c | 26 ++++++++++---------------- 3 files changed, 20 insertions(+), 23 deletions(-) (limited to 'net') diff --git a/net/mac80211/aes_ccm.c b/net/mac80211/aes_ccm.c index 4d4c2dfcf9a0..e756ed931164 100644 --- a/net/mac80211/aes_ccm.c +++ b/net/mac80211/aes_ccm.c @@ -53,15 +53,17 @@ static inline void aes_ccm_prepare(struct crypto_cipher *tfm, u8 *b_0, u8 *aad, void ieee80211_aes_ccm_encrypt(struct crypto_cipher *tfm, u8 *scratch, - u8 *b_0, u8 *aad, u8 *data, size_t data_len, + u8 *data, size_t data_len, u8 *cdata, u8 *mic) { int i, j, last_len, num_blocks; - u8 *pos, *cpos, *b, *s_0, *e; + u8 *pos, *cpos, *b, *s_0, *e, *b_0, *aad; b = scratch; s_0 = scratch + AES_BLOCK_LEN; e = scratch + 2 * AES_BLOCK_LEN; + b_0 = scratch + 3 * AES_BLOCK_LEN; + aad = scratch + 4 * AES_BLOCK_LEN; num_blocks = DIV_ROUND_UP(data_len, AES_BLOCK_LEN); last_len = data_len % AES_BLOCK_LEN; @@ -92,15 +94,16 @@ void ieee80211_aes_ccm_encrypt(struct crypto_cipher *tfm, u8 *scratch, int ieee80211_aes_ccm_decrypt(struct crypto_cipher *tfm, u8 *scratch, - u8 *b_0, u8 *aad, u8 *cdata, size_t data_len, - u8 *mic, u8 *data) + u8 *cdata, size_t data_len, u8 *mic, u8 *data) { int i, j, last_len, num_blocks; - u8 *pos, *cpos, *b, *s_0, *a; + u8 *pos, *cpos, *b, *s_0, *a, *b_0, *aad; b = scratch; s_0 = scratch + AES_BLOCK_LEN; a = scratch + 2 * AES_BLOCK_LEN; + b_0 = scratch + 3 * AES_BLOCK_LEN; + aad = scratch + 4 * AES_BLOCK_LEN; num_blocks = DIV_ROUND_UP(data_len, AES_BLOCK_LEN); last_len = data_len % AES_BLOCK_LEN; diff --git a/net/mac80211/aes_ccm.h b/net/mac80211/aes_ccm.h index 8cd0f14aab4d..6e7820ef3448 100644 --- a/net/mac80211/aes_ccm.h +++ b/net/mac80211/aes_ccm.h @@ -16,10 +16,10 @@ struct crypto_cipher *ieee80211_aes_key_setup_encrypt(const u8 key[]); void ieee80211_aes_ccm_encrypt(struct crypto_cipher *tfm, u8 *scratch, - u8 *b_0, u8 *aad, u8 *data, size_t data_len, + u8 *data, size_t data_len, u8 *cdata, u8 *mic); int ieee80211_aes_ccm_decrypt(struct crypto_cipher *tfm, u8 *scratch, - u8 *b_0, u8 *aad, u8 *cdata, size_t data_len, + u8 *cdata, size_t data_len, u8 *mic, u8 *data); void ieee80211_aes_key_free(struct crypto_cipher *tfm); diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c index ec2ae86d64aa..2f33df0dcccf 100644 --- a/net/mac80211/wpa.c +++ b/net/mac80211/wpa.c @@ -274,16 +274,20 @@ ieee80211_crypto_tkip_decrypt(struct ieee80211_rx_data *rx) } -static void ccmp_special_blocks(struct sk_buff *skb, u8 *pn, u8 *b_0, u8 *aad, +static void ccmp_special_blocks(struct sk_buff *skb, u8 *pn, u8 *scratch, int encrypted) { __le16 mask_fc; int a4_included; u8 qos_tid; + u8 *b_0, *aad; u16 data_len, len_a; unsigned int hdrlen; struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data; + b_0 = scratch + 3 * AES_BLOCK_LEN; + aad = scratch + 4 * AES_BLOCK_LEN; + /* * Mask FC: zero subtype b4 b5 b6 * Retry, PwrMgt, MoreData; set Protected @@ -367,7 +371,7 @@ static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) struct ieee80211_key *key = tx->key; struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); int hdrlen, len, tail; - u8 *pos, *pn, *b_0, *aad, *scratch; + u8 *pos, *pn; int i; info->control.icv_len = CCMP_MIC_LEN; @@ -381,10 +385,6 @@ static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) return 0; } - scratch = key->u.ccmp.tx_crypto_buf; - b_0 = scratch + 3 * AES_BLOCK_LEN; - aad = scratch + 4 * AES_BLOCK_LEN; - hdrlen = ieee80211_hdrlen(hdr->frame_control); len = skb->len - hdrlen; @@ -420,8 +420,8 @@ static int ccmp_encrypt_skb(struct ieee80211_tx_data *tx, struct sk_buff *skb) } pos += CCMP_HDR_LEN; - ccmp_special_blocks(skb, pn, b_0, aad, 0); - ieee80211_aes_ccm_encrypt(key->u.ccmp.tfm, scratch, b_0, aad, pos, len, + ccmp_special_blocks(skb, pn, key->u.ccmp.tx_crypto_buf, 0); + ieee80211_aes_ccm_encrypt(key->u.ccmp.tfm, key->u.ccmp.tx_crypto_buf, pos, len, pos, skb_put(skb, CCMP_MIC_LEN)); return 0; @@ -483,16 +483,10 @@ ieee80211_crypto_ccmp_decrypt(struct ieee80211_rx_data *rx) if (!(rx->status->flag & RX_FLAG_DECRYPTED)) { /* hardware didn't decrypt/verify MIC */ - u8 *scratch, *b_0, *aad; - - scratch = key->u.ccmp.rx_crypto_buf; - b_0 = scratch + 3 * AES_BLOCK_LEN; - aad = scratch + 4 * AES_BLOCK_LEN; - - ccmp_special_blocks(skb, pn, b_0, aad, 1); + ccmp_special_blocks(skb, pn, key->u.ccmp.rx_crypto_buf, 1); if (ieee80211_aes_ccm_decrypt( - key->u.ccmp.tfm, scratch, b_0, aad, + key->u.ccmp.tfm, key->u.ccmp.rx_crypto_buf, skb->data + hdrlen + CCMP_HDR_LEN, data_len, skb->data + skb->len - CCMP_MIC_LEN, skb->data + hdrlen + CCMP_HDR_LEN)) { -- cgit v1.2.2 From 5fdae6b37e9346dbaea1e51ed6235ed0269d445d Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 2 Jul 2008 16:30:53 -0700 Subject: mac80211: aes_ccm.c remove crypto wrapper and extra args Signed-off-by: Harvey Harrison Signed-off-by: John W. Linville --- net/mac80211/aes_ccm.c | 38 +++++++++++++++++--------------------- 1 file changed, 17 insertions(+), 21 deletions(-) (limited to 'net') diff --git a/net/mac80211/aes_ccm.c b/net/mac80211/aes_ccm.c index e756ed931164..a87cb3ba2df6 100644 --- a/net/mac80211/aes_ccm.c +++ b/net/mac80211/aes_ccm.c @@ -16,31 +16,28 @@ #include "key.h" #include "aes_ccm.h" - -static void ieee80211_aes_encrypt(struct crypto_cipher *tfm, - const u8 pt[16], u8 ct[16]) -{ - crypto_cipher_encrypt_one(tfm, ct, pt); -} - - -static inline void aes_ccm_prepare(struct crypto_cipher *tfm, u8 *b_0, u8 *aad, - u8 *b, u8 *s_0, u8 *a) +static void aes_ccm_prepare(struct crypto_cipher *tfm, u8 *scratch, u8 *a) { int i; + u8 *b_0, *aad, *b, *s_0; - ieee80211_aes_encrypt(tfm, b_0, b); + b_0 = scratch + 3 * AES_BLOCK_LEN; + aad = scratch + 4 * AES_BLOCK_LEN; + b = scratch; + s_0 = scratch + AES_BLOCK_LEN; + + crypto_cipher_encrypt_one(tfm, b, b_0); /* Extra Authenticate-only data (always two AES blocks) */ for (i = 0; i < AES_BLOCK_LEN; i++) aad[i] ^= b[i]; - ieee80211_aes_encrypt(tfm, aad, b); + crypto_cipher_encrypt_one(tfm, b, aad); aad += AES_BLOCK_LEN; for (i = 0; i < AES_BLOCK_LEN; i++) aad[i] ^= b[i]; - ieee80211_aes_encrypt(tfm, aad, a); + crypto_cipher_encrypt_one(tfm, a, aad); /* Mask out bits from auth-only-b_0 */ b_0[0] &= 0x07; @@ -48,7 +45,7 @@ static inline void aes_ccm_prepare(struct crypto_cipher *tfm, u8 *b_0, u8 *aad, /* S_0 is used to encrypt T (= MIC) */ b_0[14] = 0; b_0[15] = 0; - ieee80211_aes_encrypt(tfm, b_0, s_0); + crypto_cipher_encrypt_one(tfm, s_0, b_0); } @@ -67,7 +64,7 @@ void ieee80211_aes_ccm_encrypt(struct crypto_cipher *tfm, u8 *scratch, num_blocks = DIV_ROUND_UP(data_len, AES_BLOCK_LEN); last_len = data_len % AES_BLOCK_LEN; - aes_ccm_prepare(tfm, b_0, aad, b, s_0, b); + aes_ccm_prepare(tfm, scratch, b); /* Process payload blocks */ pos = data; @@ -79,11 +76,11 @@ void ieee80211_aes_ccm_encrypt(struct crypto_cipher *tfm, u8 *scratch, /* Authentication followed by encryption */ for (i = 0; i < blen; i++) b[i] ^= pos[i]; - ieee80211_aes_encrypt(tfm, b, b); + crypto_cipher_encrypt_one(tfm, b, b); b_0[14] = (j >> 8) & 0xff; b_0[15] = j & 0xff; - ieee80211_aes_encrypt(tfm, b_0, e); + crypto_cipher_encrypt_one(tfm, e, b_0); for (i = 0; i < blen; i++) *cpos++ = *pos++ ^ e[i]; } @@ -107,7 +104,7 @@ int ieee80211_aes_ccm_decrypt(struct crypto_cipher *tfm, u8 *scratch, num_blocks = DIV_ROUND_UP(data_len, AES_BLOCK_LEN); last_len = data_len % AES_BLOCK_LEN; - aes_ccm_prepare(tfm, b_0, aad, b, s_0, a); + aes_ccm_prepare(tfm, scratch, a); /* Process payload blocks */ cpos = cdata; @@ -119,13 +116,12 @@ int ieee80211_aes_ccm_decrypt(struct crypto_cipher *tfm, u8 *scratch, /* Decryption followed by authentication */ b_0[14] = (j >> 8) & 0xff; b_0[15] = j & 0xff; - ieee80211_aes_encrypt(tfm, b_0, b); + crypto_cipher_encrypt_one(tfm, b, b_0); for (i = 0; i < blen; i++) { *pos = *cpos++ ^ b[i]; a[i] ^= *pos++; } - - ieee80211_aes_encrypt(tfm, a, a); + crypto_cipher_encrypt_one(tfm, a, a); } for (i = 0; i < CCMP_MIC_LEN; i++) { -- cgit v1.2.2 From e4abd4d49d6df6d5e94564c5e831c61ac722f6ec Mon Sep 17 00:00:00 2001 From: Emmanuel Grumbach Date: Thu, 3 Jul 2008 18:02:27 +0300 Subject: mac80211: add support for iwconfig wlanX frag auto Signed-off-by: Emmanuel Grumbach Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/wext.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'net') diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c index 1babb979fe00..736c32e340f2 100644 --- a/net/mac80211/wext.c +++ b/net/mac80211/wext.c @@ -800,6 +800,8 @@ static int ieee80211_ioctl_siwfrag(struct net_device *dev, if (frag->disabled) local->fragmentation_threshold = IEEE80211_MAX_FRAG_THRESHOLD; + else if (!frag->fixed) + local->fragmentation_threshold = IEEE80211_MAX_FRAG_THRESHOLD; else if (frag->value < 256 || frag->value > IEEE80211_MAX_FRAG_THRESHOLD) return -EINVAL; -- cgit v1.2.2 From 0f687e9aeb590e9581709379f47dd13ee9357258 Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Thu, 3 Jul 2008 13:14:56 -0300 Subject: rfkill: some minor kernel-doc changes for rfkill_toggle_radio Improve rfkill_toggle_radio's kernel-doc header a bit. Signed-off-by: Henrique de Moraes Holschuh Acked-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index ce0e23148cdd..aa7039dfa19d 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -130,17 +130,19 @@ static void update_rfkill_state(struct rfkill *rfkill) /** * rfkill_toggle_radio - wrapper for toggle_radio hook - * calls toggle_radio taking into account a lot of "small" - * details. + * * @rfkill: the rfkill struct to use * @force: calls toggle_radio even if cache says it is not needed, * and also makes sure notifications of the state will be * sent even if it didn't change * @state: the new state to call toggle_radio() with * - * This wrappen protects and enforces the API for toggle_radio - * calls. Note that @force cannot override a (possibly cached) - * state of RFKILL_STATE_HARD_BLOCKED. Any device making use of + * Calls rfkill->toggle_radio, enforcing the API for toggle_radio + * calls and handling all the red tape such as issuing notifications + * if the call is successful. + * + * Note that @force cannot override a (possibly cached) state of + * RFKILL_STATE_HARD_BLOCKED. Any device making use of * RFKILL_STATE_HARD_BLOCKED implements either get_state() or * rfkill_force_state(), so the cache either is bypassed or valid. * -- cgit v1.2.2 From fd4484af7c02b31bcb6090eeb0d85cf947719f2d Mon Sep 17 00:00:00 2001 From: Henrique de Moraes Holschuh Date: Thu, 3 Jul 2008 13:14:57 -0300 Subject: rfkill: ignore errors from rfkill_toggle_radio in rfkill_add_switch rfkill_add_switch() calls rfkill_toggle_radio() to set the state of a recently registered rfkill class to the current global state [for that rfkill->type]. The rfkill_toggle_radio() call is going to error out if the hardware is RFKILL_STATE_HARD_BLOCKED, and the global state is RFKILL_STATE_UNBLOCKED. That is a quite normal situation which I missed to account for. As things stand, the error return from rfkill_toggle_radio ends up causing rfkill_register to bail out with an error (de-registering the new switch in the process), which is Not Nice. Change rfkill_add_switch() to not return errors because of a failed call to rfkill_toggle_radio(). We can go back to returning errors again (if that's indeed the right thing to do) if we define the exact error codes the rfkill->toggle_radio callbacks are to return in each situation, so that we can ignore the right ones only. Bug reported by "kionez ". Signed-off-by: Henrique de Moraes Holschuh Cc: kionez Acked-by: Ivo van Doorn Signed-off-by: John W. Linville --- net/rfkill/rfkill.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/rfkill/rfkill.c b/net/rfkill/rfkill.c index aa7039dfa19d..7a560b785097 100644 --- a/net/rfkill/rfkill.c +++ b/net/rfkill/rfkill.c @@ -501,17 +501,15 @@ static struct class rfkill_class = { static int rfkill_add_switch(struct rfkill *rfkill) { - int error; - mutex_lock(&rfkill_mutex); - error = rfkill_toggle_radio(rfkill, rfkill_states[rfkill->type], 0); - if (!error) - list_add_tail(&rfkill->node, &rfkill_list); + rfkill_toggle_radio(rfkill, rfkill_states[rfkill->type], 0); + + list_add_tail(&rfkill->node, &rfkill_list); mutex_unlock(&rfkill_mutex); - return error; + return 0; } static void rfkill_remove_switch(struct rfkill *rfkill) -- cgit v1.2.2 From 6ef307bc561911c8cdda98ef3896b5982b602a43 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Thu, 3 Jul 2008 13:52:18 -0700 Subject: mac80211: fix lots of kernel-doc Fix more than 50 kernel-doc warnings in ieee80211/mac80211 kernel-doc notation. Fix a few typos also. Note: Some fields are marked as TBD and need to have their description corrected. Signed-off-by: Randy Dunlap Signed-off-by: John W. Linville --- net/mac80211/sta_info.c | 1 + net/mac80211/sta_info.h | 81 ++++++++++++++++++++++++++++++++++++------------- 2 files changed, 61 insertions(+), 21 deletions(-) (limited to 'net') diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index d8a16b7f6a6b..47d2c1bbfcaa 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -135,6 +135,7 @@ struct sta_info *sta_info_get_by_idx(struct ieee80211_local *local, int idx, /** * __sta_info_free - internal STA free helper * + * @local: pointer to the global information * @sta: STA info to free * * This function must undo everything done by sta_info_alloc() diff --git a/net/mac80211/sta_info.h b/net/mac80211/sta_info.h index fd228c198e31..94311dcfe043 100644 --- a/net/mac80211/sta_info.h +++ b/net/mac80211/sta_info.h @@ -160,11 +160,21 @@ struct sta_ampdu_mlme { * @list: global linked list entry * @hnext: hash table linked list pointer * @local: pointer to the global information + * @sdata: TBD + * @key: TBD + * @rate_ctrl: TBD + * @rate_ctrl_priv: TBD + * @lock: used for locking all fields that require locking, see comments + * in the header file. + * @flaglock: spinlock for flags accesses + * @ht_info: HT capabilities of this STA + * @supp_rates: Bitmap of supported rates (per band) * @addr: MAC address of this STA * @aid: STA's unique AID (1..2007, 0 = not assigned yet), * only used in AP (and IBSS?) mode + * @listen_interval: TBD + * @pin_status: TBD * @flags: STA flags, see &enum ieee80211_sta_info_flags - * @flaglock: spinlock for flags accesses * @ps_tx_buf: buffer of frames to transmit to this station * when it leaves power saving state * @tx_filtered: buffer of frames we already tried to transmit @@ -172,10 +182,41 @@ struct sta_ampdu_mlme { * power saving state * @rx_packets: Number of MSDUs received from this STA * @rx_bytes: Number of bytes received from this STA - * @supp_rates: Bitmap of supported rates (per band) - * @ht_info: HT capabilities of this STA - * @lock: used for locking all fields that require locking, see comments - * in the header file. + * @wep_weak_iv_count: TBD + * @last_rx: TBD + * @num_duplicates: number of duplicate frames received from this STA + * @rx_fragments: number of received MPDUs + * @rx_dropped: number of dropped MPDUs from this STA + * @last_signal: signal of last received frame from this STA + * @last_qual: qual of last received frame from this STA + * @last_noise: noise of last received frame from this STA + * @last_seq_ctrl: last received seq/frag number from this STA (per RX queue) + * @wme_rx_queue: TBD + * @tx_filtered_count: TBD + * @tx_retry_failed: TBD + * @tx_retry_count: TBD + * @tx_num_consecutive_failures: TBD + * @tx_num_mpdu_ok: TBD + * @tx_num_mpdu_fail: TBD + * @fail_avg: moving percentage of failed MSDUs + * @tx_packets: number of RX/TX MSDUs + * @tx_bytes: TBD + * @tx_fragments: number of transmitted MPDUs + * @txrate_idx: TBD + * @last_txrate_idx: TBD + * @wme_tx_queue: TBD + * @ampdu_mlme: TBD + * @timer_to_tid: identity mapping to ID timers + * @tid_to_tx_q: map tid to tx queue + * @llid: Local link ID + * @plid: Peer link ID + * @reason: Cancel reason on PLINK_HOLDING state + * @plink_retries: Retries in establishment + * @ignore_plink_timer: TBD + * @plink_state plink_state: TBD + * @plink_timeout: TBD + * @plink_timer: TBD + * @debugfs: debug filesystem info */ struct sta_info { /* General information, mostly static */ @@ -217,14 +258,12 @@ struct sta_info { unsigned long rx_packets, rx_bytes; unsigned long wep_weak_iv_count; unsigned long last_rx; - unsigned long num_duplicates; /* number of duplicate frames received - * from this STA */ - unsigned long rx_fragments; /* number of received MPDUs */ - unsigned long rx_dropped; /* number of dropped MPDUs from this STA */ - int last_signal; /* signal of last received frame from this STA */ - int last_qual; /* qual of last received frame from this STA */ - int last_noise; /* noise of last received frame from this STA */ - /* last received seq/frag number from this STA (per RX queue) */ + unsigned long num_duplicates; + unsigned long rx_fragments; + unsigned long rx_dropped; + int last_signal; + int last_qual; + int last_noise; __le16 last_seq_ctrl[NUM_RX_DATA_QUEUES]; #ifdef CONFIG_MAC80211_DEBUG_COUNTERS unsigned int wme_rx_queue[NUM_RX_DATA_QUEUES]; @@ -241,9 +280,9 @@ struct sta_info { unsigned int fail_avg; /* Updated from TX path only, no locking requirements */ - unsigned long tx_packets; /* number of RX/TX MSDUs */ + unsigned long tx_packets; unsigned long tx_bytes; - unsigned long tx_fragments; /* number of transmitted MPDUs */ + unsigned long tx_fragments; int txrate_idx; int last_txrate_idx; #ifdef CONFIG_MAC80211_DEBUG_COUNTERS @@ -254,18 +293,18 @@ struct sta_info { * Aggregation information, locked with lock. */ struct sta_ampdu_mlme ampdu_mlme; - u8 timer_to_tid[STA_TID_NUM]; /* identity mapping to ID timers */ - u8 tid_to_tx_q[STA_TID_NUM]; /* map tid to tx queue */ + u8 timer_to_tid[STA_TID_NUM]; + u8 tid_to_tx_q[STA_TID_NUM]; #ifdef CONFIG_MAC80211_MESH /* * Mesh peer link attributes * TODO: move to a sub-structure that is referenced with pointer? */ - __le16 llid; /* Local link ID */ - __le16 plid; /* Peer link ID */ - __le16 reason; /* Cancel reason on PLINK_HOLDING state */ - u8 plink_retries; /* Retries in establishment */ + __le16 llid; + __le16 plid; + __le16 reason; + u8 plink_retries; bool ignore_plink_timer; enum plink_state plink_state; u32 plink_timeout; -- cgit v1.2.2 From b46372710ab536c0967f76be5dc41341583d4a54 Mon Sep 17 00:00:00 2001 From: Julius Volz Date: Tue, 8 Jul 2008 14:02:19 +0200 Subject: net/wireless/nl80211.c: fix endless Netlink callback loop. Although I only tested similar code (I don't use any of this wireless code), the state maintainance between Netlink dump callback invocations seems wrong here and should lead to an endless loop. There are also other examples in the same file which might have the same problem. Perhaps someone can actually test this (or refute my logic). Take the simple example with only one element in the list (which should fit into the message): 1. invocation: Start: idx = 0, start = 0 Loop: condition (++idx < start) => (1 < 0) => false => no continue, fill one entry, exit loop, return skb->len > 0 2. invocation: Start: idx = 0, start = 1 Loop: condition (++idx < start) => (1 < 1) => false => no continue, fill the same entry again, exit loop, return skb->len > 0 3. invocation: Same as 2. invocation, endless invocation of callback. Also, iterations where the filling of an element fails should not be counted as completed, so idx should not be incremented in this case. Signed-off-by: Julius Volz Signed-off-by: John W. Linville --- net/wireless/nl80211.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/wireless/nl80211.c b/net/wireless/nl80211.c index fb75f265b39c..b7fefffd2d0d 100644 --- a/net/wireless/nl80211.c +++ b/net/wireless/nl80211.c @@ -199,12 +199,14 @@ static int nl80211_dump_wiphy(struct sk_buff *skb, struct netlink_callback *cb) mutex_lock(&cfg80211_drv_mutex); list_for_each_entry(dev, &cfg80211_drv_list, list) { - if (++idx < start) + if (++idx <= start) continue; if (nl80211_send_wiphy(skb, NETLINK_CB(cb->skb).pid, cb->nlh->nlmsg_seq, NLM_F_MULTI, - dev) < 0) + dev) < 0) { + idx--; break; + } } mutex_unlock(&cfg80211_drv_mutex); -- cgit v1.2.2 From 11a100f844f6096787ab20e19f17d72abc957a8f Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Tue, 8 Jul 2008 15:36:57 -0700 Subject: vlan: avoid header copying and linearisation where possible - vlan_dev_reorder_header() is only called on the receive path after calling skb_share_check(). This means we can use skb_cow() since all we need is a writable header. - vlan_dev_hard_header() includes a work-around for some apparently broken out of tree MPLS code. The hard_header functions can expect to always have a headroom of at least there own hard_header_len available, so the reallocation check is unnecessary. - __vlan_put_tag() can use skb_cow_head() to avoid the skb_unshare() copy when the header is writable. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan_dev.c | 34 +++++----------------------------- 1 file changed, 5 insertions(+), 29 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 2ccac6bea57e..b6e52c025fd8 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -74,11 +74,8 @@ static int vlan_dev_rebuild_header(struct sk_buff *skb) static inline struct sk_buff *vlan_check_reorder_header(struct sk_buff *skb) { if (vlan_dev_info(skb->dev)->flags & VLAN_FLAG_REORDER_HDR) { - if (skb_shared(skb) || skb_cloned(skb)) { - struct sk_buff *nskb = skb_copy(skb, GFP_ATOMIC); - kfree_skb(skb); - skb = nskb; - } + if (skb_cow(skb, skb_headroom(skb)) < 0) + skb = NULL; if (skb) { /* Lifted from Gleb's VLAN code... */ memmove(skb->data - ETH_HLEN, @@ -262,12 +259,14 @@ static int vlan_dev_hard_header(struct sk_buff *skb, struct net_device *dev, u16 vlan_tci = 0; int rc = 0; int build_vlan_header = 0; - struct net_device *vdev = dev; pr_debug("%s: skb: %p type: %hx len: %u vlan_id: %hx, daddr: %p\n", __func__, skb, type, len, vlan_dev_info(dev)->vlan_id, daddr); + if (WARN_ON(skb_headroom(skb) < dev->hard_header_len)) + return -ENOSPC; + /* build vlan header only if re_order_header flag is NOT set. This * fixes some programs that get confused when they see a VLAN device * sending a frame that is VLAN encoded (the consensus is that the VLAN @@ -316,29 +315,6 @@ static int vlan_dev_hard_header(struct sk_buff *skb, struct net_device *dev, dev = vlan_dev_info(dev)->real_dev; - /* MPLS can send us skbuffs w/out enough space. This check will grow - * the skb if it doesn't have enough headroom. Not a beautiful solution, - * so I'll tick a counter so that users can know it's happening... - * If they care... - */ - - /* NOTE: This may still break if the underlying device is not the final - * device (and thus there are more headers to add...) It should work for - * good-ole-ethernet though. - */ - if (skb_headroom(skb) < dev->hard_header_len) { - struct sk_buff *sk_tmp = skb; - skb = skb_realloc_headroom(sk_tmp, dev->hard_header_len); - kfree_skb(sk_tmp); - if (skb == NULL) { - struct net_device_stats *stats = &vdev->stats; - stats->tx_dropped++; - return -ENOMEM; - } - vlan_dev_info(vdev)->cnt_inc_headroom_on_tx++; - pr_debug("%s: %s: had to grow skb\n", __func__, vdev->name); - } - if (build_vlan_header) { /* Now make the underlying real hard header */ rc = dev_hard_header(skb, dev, ETH_P_8021Q, daddr, saddr, -- cgit v1.2.2 From e65d22e18038eed7307276e46810d884c402d57d Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 16:46:01 -0700 Subject: pkt_sched: Remove comment reference to old style TX locking. We haven't had netdev->tbusy in many years :) Signed-off-by: David S. Miller --- net/sched/sch_api.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index e9ebc7af049e..69e918bb4278 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -99,7 +99,7 @@ static int tclass_notify(struct sk_buff *oskb, struct nlmsghdr *n, ---requeue requeues once dequeued packet. It is used for non-standard or - just buggy devices, which can defer output even if dev->tbusy=0. + just buggy devices, which can defer output even if netif_queue_stopped()=0. ---reset -- cgit v1.2.2 From bb949fbd1878973c3539d9aecff52f284482a937 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 16:55:56 -0700 Subject: netdev: Create netdev_queue abstraction. A netdev_queue is an entity managed by a qdisc. Currently there is one RX and one TX queue, and a netdev_queue merely contains a backpointer to the net_device. The Qdisc struct is augmented with a netdev_queue pointer as well. Eventually the 'dev' Qdisc member will go away and we will have the resulting hierarchy: net_device --> netdev_queue --> Qdisc Also, qdisc_alloc() and qdisc_create_dflt() now take a netdev_queue pointer argument. Signed-off-by: David S. Miller --- net/core/dev.c | 8 ++++++++ net/mac80211/wme.c | 6 ++++-- net/sched/sch_api.c | 12 +++++++----- net/sched/sch_atm.c | 6 ++++-- net/sched/sch_cbq.c | 9 ++++++--- net/sched/sch_dsmark.c | 6 ++++-- net/sched/sch_fifo.c | 3 ++- net/sched/sch_generic.c | 14 ++++++++++---- net/sched/sch_hfsc.c | 9 ++++++--- net/sched/sch_htb.c | 11 +++++++---- net/sched/sch_netem.c | 3 ++- net/sched/sch_prio.c | 3 ++- 12 files changed, 62 insertions(+), 28 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 75933932463d..9b281c906eb0 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4072,6 +4072,12 @@ static struct net_device_stats *internal_stats(struct net_device *dev) return &dev->stats; } +static void netdev_init_queues(struct net_device *dev) +{ + dev->rx_queue.dev = dev; + dev->tx_queue.dev = dev; +} + /** * alloc_netdev_mq - allocate network device * @sizeof_priv: size of private data to allocate space for @@ -4124,6 +4130,8 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, dev->egress_subqueue_count = queue_count; dev->gso_max_size = GSO_MAX_SIZE; + netdev_init_queues(dev); + dev->get_stats = internal_stats; netpoll_netdev_init(dev); setup(dev); diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 5c666f7eda8f..770f1c09b793 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -359,7 +359,8 @@ static int wme_qdiscop_init(struct Qdisc *qd, struct nlattr *opt) /* create child queues */ for (i = 0; i < QD_NUM(hw); i++) { skb_queue_head_init(&q->requeued[i]); - q->queues[i] = qdisc_create_dflt(qd->dev, &pfifo_qdisc_ops, + q->queues[i] = qdisc_create_dflt(qd->dev, qd->dev_queue, + &pfifo_qdisc_ops, qd->handle); if (!q->queues[i]) { q->queues[i] = &noop_qdisc; @@ -575,7 +576,8 @@ void ieee80211_install_qdisc(struct net_device *dev) { struct Qdisc *qdisc; - qdisc = qdisc_create_dflt(dev, &wme_qdisc_ops, TC_H_ROOT); + qdisc = qdisc_create_dflt(dev, &dev->tx_queue, + &wme_qdisc_ops, TC_H_ROOT); if (!qdisc) { printk(KERN_ERR "%s: qdisc installation failed\n", dev->name); return; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 69e918bb4278..b86c98bd06a3 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -552,8 +552,8 @@ static int qdisc_graft(struct net_device *dev, struct Qdisc *parent, */ static struct Qdisc * -qdisc_create(struct net_device *dev, u32 parent, u32 handle, - struct nlattr **tca, int *errp) +qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, + u32 parent, u32 handle, struct nlattr **tca, int *errp) { int err; struct nlattr *kind = tca[TCA_KIND]; @@ -593,7 +593,7 @@ qdisc_create(struct net_device *dev, u32 parent, u32 handle, if (ops == NULL) goto err_out; - sch = qdisc_alloc(dev, ops); + sch = qdisc_alloc(dev, dev_queue, ops); if (IS_ERR(sch)) { err = PTR_ERR(sch); goto err_out2; @@ -892,10 +892,12 @@ create_n_graft: if (!(n->nlmsg_flags&NLM_F_CREATE)) return -ENOENT; if (clid == TC_H_INGRESS) - q = qdisc_create(dev, tcm->tcm_parent, tcm->tcm_parent, + q = qdisc_create(dev, &dev->rx_queue, + tcm->tcm_parent, tcm->tcm_parent, tca, &err); else - q = qdisc_create(dev, tcm->tcm_parent, tcm->tcm_handle, + q = qdisc_create(dev, &dev->tx_queue, + tcm->tcm_parent, tcm->tcm_handle, tca, &err); if (q == NULL) { if (err == -EAGAIN) diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c index db0e23ae85f8..3dddab531d5a 100644 --- a/net/sched/sch_atm.c +++ b/net/sched/sch_atm.c @@ -296,7 +296,8 @@ static int atm_tc_change(struct Qdisc *sch, u32 classid, u32 parent, goto err_out; } flow->filter_list = NULL; - flow->q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, classid); + flow->q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, classid); if (!flow->q) flow->q = &noop_qdisc; pr_debug("atm_tc_change: qdisc %p\n", flow->q); @@ -555,7 +556,8 @@ static int atm_tc_init(struct Qdisc *sch, struct nlattr *opt) pr_debug("atm_tc_init(sch %p,[qdisc %p],opt %p)\n", sch, p, opt); p->flows = &p->link; - p->link.q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, sch->handle); + p->link.q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, sch->handle); if (!p->link.q) p->link.q = &noop_qdisc; pr_debug("atm_tc_init: link (%p) qdisc %p\n", &p->link, p->link.q); diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 968b4c73c9c1..d360dcd0818b 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -1401,7 +1401,8 @@ static int cbq_init(struct Qdisc *sch, struct nlattr *opt) q->link.sibling = &q->link; q->link.common.classid = sch->handle; q->link.qdisc = sch; - if (!(q->link.q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + if (!(q->link.q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, sch->handle))) q->link.q = &noop_qdisc; @@ -1645,7 +1646,8 @@ static int cbq_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, if (cl) { if (new == NULL) { - new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + new = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, cl->common.classid); if (new == NULL) return -ENOBUFS; @@ -1877,7 +1879,8 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t cl->R_tab = rtab; rtab = NULL; cl->refcnt = 1; - if (!(cl->q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, classid))) + if (!(cl->q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, classid))) cl->q = &noop_qdisc; cl->common.classid = classid; cl->tparent = parent; diff --git a/net/sched/sch_dsmark.c b/net/sched/sch_dsmark.c index c4c1317cd47d..c955ba24e5cf 100644 --- a/net/sched/sch_dsmark.c +++ b/net/sched/sch_dsmark.c @@ -60,7 +60,8 @@ static int dsmark_graft(struct Qdisc *sch, unsigned long arg, sch, p, new, old); if (new == NULL) { - new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + new = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, sch->handle); if (new == NULL) new = &noop_qdisc; @@ -390,7 +391,8 @@ static int dsmark_init(struct Qdisc *sch, struct nlattr *opt) p->default_index = default_index; p->set_tc_index = nla_get_flag(tb[TCA_DSMARK_SET_TC_INDEX]); - p->q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, sch->handle); + p->q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, sch->handle); if (p->q == NULL) p->q = &noop_qdisc; diff --git a/net/sched/sch_fifo.c b/net/sched/sch_fifo.c index 82d7d7bbbb16..779eae85faf0 100644 --- a/net/sched/sch_fifo.c +++ b/net/sched/sch_fifo.c @@ -137,7 +137,8 @@ struct Qdisc *fifo_create_dflt(struct Qdisc *sch, struct Qdisc_ops *ops, struct Qdisc *q; int err = -ENOMEM; - q = qdisc_create_dflt(sch->dev, ops, TC_H_MAKE(sch->handle, 1)); + q = qdisc_create_dflt(sch->dev, sch->dev_queue, + ops, TC_H_MAKE(sch->handle, 1)); if (q) { err = fifo_set_limit(q, limit); if (err < 0) { diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 13afa7214392..d97086480893 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -440,7 +440,9 @@ static struct Qdisc_ops pfifo_fast_ops __read_mostly = { .owner = THIS_MODULE, }; -struct Qdisc *qdisc_alloc(struct net_device *dev, struct Qdisc_ops *ops) +struct Qdisc *qdisc_alloc(struct net_device *dev, + struct netdev_queue *dev_queue, + struct Qdisc_ops *ops) { void *p; struct Qdisc *sch; @@ -462,6 +464,7 @@ struct Qdisc *qdisc_alloc(struct net_device *dev, struct Qdisc_ops *ops) sch->ops = ops; sch->enqueue = ops->enqueue; sch->dequeue = ops->dequeue; + sch->dev_queue = dev_queue; sch->dev = dev; dev_hold(dev); atomic_set(&sch->refcnt, 1); @@ -471,12 +474,14 @@ errout: return ERR_PTR(err); } -struct Qdisc * qdisc_create_dflt(struct net_device *dev, struct Qdisc_ops *ops, +struct Qdisc * qdisc_create_dflt(struct net_device *dev, + struct netdev_queue *dev_queue, + struct Qdisc_ops *ops, unsigned int parentid) { struct Qdisc *sch; - sch = qdisc_alloc(dev, ops); + sch = qdisc_alloc(dev, dev_queue, ops); if (IS_ERR(sch)) goto errout; sch->stats_lock = &dev->queue_lock; @@ -545,7 +550,8 @@ void dev_activate(struct net_device *dev) if (dev->qdisc_sleeping == &noop_qdisc) { struct Qdisc *qdisc; if (dev->tx_queue_len) { - qdisc = qdisc_create_dflt(dev, &pfifo_fast_ops, + qdisc = qdisc_create_dflt(dev, &dev->tx_queue, + &pfifo_fast_ops, TC_H_ROOT); if (qdisc == NULL) { printk(KERN_INFO "%s: activation failed\n", dev->name); diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 3a8267246a4f..5a22fec4eadd 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1083,7 +1083,8 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, cl->refcnt = 1; cl->sched = q; cl->cl_parent = parent; - cl->qdisc = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, classid); + cl->qdisc = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, classid); if (cl->qdisc == NULL) cl->qdisc = &noop_qdisc; INIT_LIST_HEAD(&cl->children); @@ -1201,7 +1202,8 @@ hfsc_graft_class(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, if (cl->level > 0) return -EINVAL; if (new == NULL) { - new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + new = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, cl->cl_common.classid); if (new == NULL) new = &noop_qdisc; @@ -1443,7 +1445,8 @@ hfsc_init_qdisc(struct Qdisc *sch, struct nlattr *opt) q->root.cl_common.classid = sch->handle; q->root.refcnt = 1; q->root.sched = q; - q->root.qdisc = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + q->root.qdisc = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, sch->handle); if (q->root.qdisc == NULL) q->root.qdisc = &noop_qdisc; diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index ee8b4ffe110c..956a67f66b9c 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -1129,7 +1129,8 @@ static int htb_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, if (cl && !cl->level) { if (new == NULL && - (new = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + (new = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, cl->common.classid)) == NULL) return -ENOBUFS; @@ -1256,8 +1257,9 @@ static int htb_delete(struct Qdisc *sch, unsigned long arg) return -EBUSY; if (!cl->level && htb_parent_last_child(cl)) { - new_q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, - cl->parent->common.classid); + new_q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, + cl->parent->common.classid); last_child = 1; } @@ -1376,7 +1378,8 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, /* create leaf qdisc early because it uses kmalloc(GFP_KERNEL) so that can't be used inside of sch_tree_lock -- thanks to Karlis Peisenieks */ - new_q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, classid); + new_q = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, classid); sch_tree_lock(sch); if (parent && !parent->level) { unsigned int qlen = parent->un.leaf.q->q.qlen; diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index 24697667247c..aa7a04e32ae9 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -536,7 +536,8 @@ static int netem_init(struct Qdisc *sch, struct nlattr *opt) qdisc_watchdog_init(&q->watchdog, sch); - q->qdisc = qdisc_create_dflt(sch->dev, &tfifo_qdisc_ops, + q->qdisc = qdisc_create_dflt(sch->dev, sch->dev_queue, + &tfifo_qdisc_ops, TC_H_MAKE(sch->handle, 1)); if (!q->qdisc) { pr_debug("netem: qdisc create failed\n"); diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c index 5532f1031ab5..ca58a039208e 100644 --- a/net/sched/sch_prio.c +++ b/net/sched/sch_prio.c @@ -281,7 +281,8 @@ static int prio_tune(struct Qdisc *sch, struct nlattr *opt) for (i=0; ibands; i++) { if (q->queues[i] == &noop_qdisc) { struct Qdisc *child; - child = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops, + child = qdisc_create_dflt(sch->dev, sch->dev_queue, + &pfifo_qdisc_ops, TC_H_MAKE(sch->handle, i + 1)); if (child) { sch_tree_lock(sch); -- cgit v1.2.2 From 5ce2d488fe039ddd86a638496cf704df86c74eeb Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 17:06:30 -0700 Subject: pkt_sched: Remove 'dev' member of struct Qdisc. It can be obtained via the netdev_queue. So create a helper routine, qdisc_dev(), to make the transformations nicer looking. Now, qdisc_alloc() now no longer needs a net_device pointer argument. Signed-off-by: David S. Miller --- net/mac80211/wme.c | 28 ++++++++++++++-------------- net/sched/cls_api.c | 2 +- net/sched/cls_route.c | 4 ++-- net/sched/sch_api.c | 10 +++++----- net/sched/sch_atm.c | 4 ++-- net/sched/sch_cbq.c | 22 +++++++++++----------- net/sched/sch_dsmark.c | 4 ++-- net/sched/sch_fifo.c | 6 +++--- net/sched/sch_generic.c | 12 +++++------- net/sched/sch_gred.c | 2 +- net/sched/sch_hfsc.c | 10 +++++----- net/sched/sch_htb.c | 24 ++++++++++++------------ net/sched/sch_netem.c | 10 +++++----- net/sched/sch_prio.c | 15 ++++++++------- net/sched/sch_sfq.c | 4 ++-- net/sched/sch_teql.c | 12 ++++++------ 16 files changed, 84 insertions(+), 85 deletions(-) (limited to 'net') diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 770f1c09b793..2fbc171130bf 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -103,7 +103,7 @@ static inline int wme_downgrade_ac(struct sk_buff *skb) * negative return value indicates to drop the frame */ static int classify80211(struct sk_buff *skb, struct Qdisc *qd) { - struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; if (!ieee80211_is_data(hdr->frame_control)) { @@ -140,7 +140,7 @@ static int classify80211(struct sk_buff *skb, struct Qdisc *qd) static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) { - struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; struct ieee80211_sched_data *q = qdisc_priv(qd); struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); @@ -249,7 +249,7 @@ static int wme_qdiscop_requeue(struct sk_buff *skb, struct Qdisc* qd) static struct sk_buff *wme_qdiscop_dequeue(struct Qdisc* qd) { struct ieee80211_sched_data *q = qdisc_priv(qd); - struct net_device *dev = qd->dev; + struct net_device *dev = qdisc_dev(qd); struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; struct sk_buff *skb; @@ -286,7 +286,7 @@ static struct sk_buff *wme_qdiscop_dequeue(struct Qdisc* qd) static void wme_qdiscop_reset(struct Qdisc* qd) { struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; int queue; @@ -303,7 +303,7 @@ static void wme_qdiscop_reset(struct Qdisc* qd) static void wme_qdiscop_destroy(struct Qdisc* qd) { struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; int queue; @@ -328,7 +328,7 @@ static int wme_qdiscop_tune(struct Qdisc *qd, struct nlattr *opt) static int wme_qdiscop_init(struct Qdisc *qd, struct nlattr *opt) { struct ieee80211_sched_data *q = qdisc_priv(qd); - struct net_device *dev = qd->dev; + struct net_device *dev = qdisc_dev(qd); struct ieee80211_local *local; struct ieee80211_hw *hw; int err = 0, i; @@ -359,7 +359,7 @@ static int wme_qdiscop_init(struct Qdisc *qd, struct nlattr *opt) /* create child queues */ for (i = 0; i < QD_NUM(hw); i++) { skb_queue_head_init(&q->requeued[i]); - q->queues[i] = qdisc_create_dflt(qd->dev, qd->dev_queue, + q->queues[i] = qdisc_create_dflt(qdisc_dev(qd), qd->dev_queue, &pfifo_qdisc_ops, qd->handle); if (!q->queues[i]) { @@ -386,7 +386,7 @@ static int wme_classop_graft(struct Qdisc *qd, unsigned long arg, struct Qdisc *new, struct Qdisc **old) { struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; unsigned long queue = arg - 1; @@ -410,7 +410,7 @@ static struct Qdisc * wme_classop_leaf(struct Qdisc *qd, unsigned long arg) { struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; unsigned long queue = arg - 1; @@ -423,7 +423,7 @@ wme_classop_leaf(struct Qdisc *qd, unsigned long arg) static unsigned long wme_classop_get(struct Qdisc *qd, u32 classid) { - struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; unsigned long queue = TC_H_MIN(classid); @@ -450,7 +450,7 @@ static int wme_classop_change(struct Qdisc *qd, u32 handle, u32 parent, struct nlattr **tca, unsigned long *arg) { unsigned long cl = *arg; - struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; if (cl - 1 > QD_NUM(hw)) @@ -467,7 +467,7 @@ static int wme_classop_change(struct Qdisc *qd, u32 handle, u32 parent, * when we add WMM-SA support - TSPECs may be deleted here */ static int wme_classop_delete(struct Qdisc *qd, unsigned long cl) { - struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; if (cl - 1 > QD_NUM(hw)) @@ -480,7 +480,7 @@ static int wme_classop_dump_class(struct Qdisc *qd, unsigned long cl, struct sk_buff *skb, struct tcmsg *tcm) { struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; if (cl - 1 > QD_NUM(hw)) @@ -494,7 +494,7 @@ static int wme_classop_dump_class(struct Qdisc *qd, unsigned long cl, static void wme_classop_walk(struct Qdisc *qd, struct qdisc_walker *arg) { - struct ieee80211_local *local = wdev_priv(qd->dev->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); struct ieee80211_hw *hw = &local->hw; int queue; diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c index 9360fc81e8c7..e2389f161e46 100644 --- a/net/sched/cls_api.c +++ b/net/sched/cls_api.c @@ -334,7 +334,7 @@ static int tcf_fill_node(struct sk_buff *skb, struct tcf_proto *tp, tcm->tcm_family = AF_UNSPEC; tcm->tcm__pad1 = 0; tcm->tcm__pad1 = 0; - tcm->tcm_ifindex = tp->q->dev->ifindex; + tcm->tcm_ifindex = qdisc_dev(tp->q)->ifindex; tcm->tcm_parent = tp->classid; tcm->tcm_info = TC_H_MAKE(tp->prio, tp->protocol); NLA_PUT_STRING(skb, TCA_KIND, tp->ops->kind); diff --git a/net/sched/cls_route.c b/net/sched/cls_route.c index 784dcb870b98..5a16ca28aa3d 100644 --- a/net/sched/cls_route.c +++ b/net/sched/cls_route.c @@ -302,7 +302,7 @@ static int route4_delete(struct tcf_proto *tp, unsigned long arg) *fp = f->next; tcf_tree_unlock(tp); - route4_reset_fastmap(tp->q->dev, head, f->id); + route4_reset_fastmap(qdisc_dev(tp->q), head, f->id); route4_delete_filter(tp, f); /* Strip tree */ @@ -500,7 +500,7 @@ reinsert: } tcf_tree_unlock(tp); - route4_reset_fastmap(tp->q->dev, head, f->id); + route4_reset_fastmap(qdisc_dev(tp->q), head, f->id); *arg = (unsigned long)f; return 0; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index b86c98bd06a3..1f893082a4f6 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -281,7 +281,7 @@ static enum hrtimer_restart qdisc_watchdog(struct hrtimer *timer) { struct qdisc_watchdog *wd = container_of(timer, struct qdisc_watchdog, timer); - struct net_device *dev = wd->qdisc->dev; + struct net_device *dev = qdisc_dev(wd->qdisc); wd->qdisc->flags &= ~TCQ_F_THROTTLED; smp_wmb(); @@ -493,7 +493,7 @@ void qdisc_tree_decrease_qlen(struct Qdisc *sch, unsigned int n) if (TC_H_MAJ(parentid) == TC_H_MAJ(TC_H_INGRESS)) return; - sch = qdisc_lookup(sch->dev, TC_H_MAJ(parentid)); + sch = qdisc_lookup(qdisc_dev(sch), TC_H_MAJ(parentid)); if (sch == NULL) { WARN_ON(parentid != TC_H_ROOT); return; @@ -593,7 +593,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, if (ops == NULL) goto err_out; - sch = qdisc_alloc(dev, dev_queue, ops); + sch = qdisc_alloc(dev_queue, ops); if (IS_ERR(sch)) { err = PTR_ERR(sch); goto err_out2; @@ -940,7 +940,7 @@ static int tc_fill_qdisc(struct sk_buff *skb, struct Qdisc *q, u32 clid, tcm->tcm_family = AF_UNSPEC; tcm->tcm__pad1 = 0; tcm->tcm__pad2 = 0; - tcm->tcm_ifindex = q->dev->ifindex; + tcm->tcm_ifindex = qdisc_dev(q)->ifindex; tcm->tcm_parent = clid; tcm->tcm_handle = q->handle; tcm->tcm_info = atomic_read(&q->refcnt); @@ -1186,7 +1186,7 @@ static int tc_fill_tclass(struct sk_buff *skb, struct Qdisc *q, nlh = NLMSG_NEW(skb, pid, seq, event, sizeof(*tcm), flags); tcm = NLMSG_DATA(nlh); tcm->tcm_family = AF_UNSPEC; - tcm->tcm_ifindex = q->dev->ifindex; + tcm->tcm_ifindex = qdisc_dev(q)->ifindex; tcm->tcm_parent = q->handle; tcm->tcm_handle = q->handle; tcm->tcm_info = 0; diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c index 3dddab531d5a..0de757e3be4a 100644 --- a/net/sched/sch_atm.c +++ b/net/sched/sch_atm.c @@ -296,7 +296,7 @@ static int atm_tc_change(struct Qdisc *sch, u32 classid, u32 parent, goto err_out; } flow->filter_list = NULL; - flow->q = qdisc_create_dflt(sch->dev, sch->dev_queue, + flow->q = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, classid); if (!flow->q) flow->q = &noop_qdisc; @@ -556,7 +556,7 @@ static int atm_tc_init(struct Qdisc *sch, struct nlattr *opt) pr_debug("atm_tc_init(sch %p,[qdisc %p],opt %p)\n", sch, p, opt); p->flows = &p->link; - p->link.q = qdisc_create_dflt(sch->dev, sch->dev_queue, + p->link.q = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, sch->handle); if (!p->link.q) p->link.q = &noop_qdisc; diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index d360dcd0818b..9f2ace585fd6 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -650,7 +650,7 @@ static enum hrtimer_restart cbq_undelay(struct hrtimer *timer) } sch->flags &= ~TCQ_F_THROTTLED; - netif_schedule(sch->dev); + netif_schedule(qdisc_dev(sch)); return HRTIMER_NORESTART; } @@ -1077,9 +1077,9 @@ static void cbq_normalize_quanta(struct cbq_sched_data *q, int prio) cl->quantum = (cl->weight*cl->allot*q->nclasses[prio])/ q->quanta[prio]; } - if (cl->quantum <= 0 || cl->quantum>32*cl->qdisc->dev->mtu) { + if (cl->quantum <= 0 || cl->quantum>32*qdisc_dev(cl->qdisc)->mtu) { printk(KERN_WARNING "CBQ: class %08x has bad quantum==%ld, repaired.\n", cl->common.classid, cl->quantum); - cl->quantum = cl->qdisc->dev->mtu/2 + 1; + cl->quantum = qdisc_dev(cl->qdisc)->mtu/2 + 1; } } } @@ -1401,7 +1401,7 @@ static int cbq_init(struct Qdisc *sch, struct nlattr *opt) q->link.sibling = &q->link; q->link.common.classid = sch->handle; q->link.qdisc = sch; - if (!(q->link.q = qdisc_create_dflt(sch->dev, sch->dev_queue, + if (!(q->link.q = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, sch->handle))) q->link.q = &noop_qdisc; @@ -1411,7 +1411,7 @@ static int cbq_init(struct Qdisc *sch, struct nlattr *opt) q->link.cpriority = TC_CBQ_MAXPRIO-1; q->link.ovl_strategy = TC_CBQ_OVL_CLASSIC; q->link.overlimit = cbq_ovl_classic; - q->link.allot = psched_mtu(sch->dev); + q->link.allot = psched_mtu(qdisc_dev(sch)); q->link.quantum = q->link.allot; q->link.weight = q->link.R_tab->rate.rate; @@ -1646,7 +1646,7 @@ static int cbq_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, if (cl) { if (new == NULL) { - new = qdisc_create_dflt(sch->dev, sch->dev_queue, + new = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, cl->common.classid); if (new == NULL) @@ -1746,10 +1746,10 @@ static void cbq_put(struct Qdisc *sch, unsigned long arg) #ifdef CONFIG_NET_CLS_ACT struct cbq_sched_data *q = qdisc_priv(sch); - spin_lock_bh(&sch->dev->queue_lock); + spin_lock_bh(&qdisc_dev(sch)->queue_lock); if (q->rx_class == cl) q->rx_class = NULL; - spin_unlock_bh(&sch->dev->queue_lock); + spin_unlock_bh(&qdisc_dev(sch)->queue_lock); #endif cbq_destroy_class(sch, cl); @@ -1828,7 +1828,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &sch->dev->queue_lock, + &qdisc_dev(sch)->queue_lock, tca[TCA_RATE]); return 0; } @@ -1879,7 +1879,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t cl->R_tab = rtab; rtab = NULL; cl->refcnt = 1; - if (!(cl->q = qdisc_create_dflt(sch->dev, sch->dev_queue, + if (!(cl->q = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, classid))) cl->q = &noop_qdisc; cl->common.classid = classid; @@ -1919,7 +1919,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t if (tca[TCA_RATE]) gen_new_estimator(&cl->bstats, &cl->rate_est, - &sch->dev->queue_lock, tca[TCA_RATE]); + &qdisc_dev(sch)->queue_lock, tca[TCA_RATE]); *arg = (unsigned long)cl; return 0; diff --git a/net/sched/sch_dsmark.c b/net/sched/sch_dsmark.c index c955ba24e5cf..3aafbd17393a 100644 --- a/net/sched/sch_dsmark.c +++ b/net/sched/sch_dsmark.c @@ -60,7 +60,7 @@ static int dsmark_graft(struct Qdisc *sch, unsigned long arg, sch, p, new, old); if (new == NULL) { - new = qdisc_create_dflt(sch->dev, sch->dev_queue, + new = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, sch->handle); if (new == NULL) @@ -391,7 +391,7 @@ static int dsmark_init(struct Qdisc *sch, struct nlattr *opt) p->default_index = default_index; p->set_tc_index = nla_get_flag(tb[TCA_DSMARK_SET_TC_INDEX]); - p->q = qdisc_create_dflt(sch->dev, sch->dev_queue, + p->q = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, sch->handle); if (p->q == NULL) p->q = &noop_qdisc; diff --git a/net/sched/sch_fifo.c b/net/sched/sch_fifo.c index 779eae85faf0..1d97fa42c902 100644 --- a/net/sched/sch_fifo.c +++ b/net/sched/sch_fifo.c @@ -48,10 +48,10 @@ static int fifo_init(struct Qdisc *sch, struct nlattr *opt) struct fifo_sched_data *q = qdisc_priv(sch); if (opt == NULL) { - u32 limit = sch->dev->tx_queue_len ? : 1; + u32 limit = qdisc_dev(sch)->tx_queue_len ? : 1; if (sch->ops == &bfifo_qdisc_ops) - limit *= sch->dev->mtu; + limit *= qdisc_dev(sch)->mtu; q->limit = limit; } else { @@ -137,7 +137,7 @@ struct Qdisc *fifo_create_dflt(struct Qdisc *sch, struct Qdisc_ops *ops, struct Qdisc *q; int err = -ENOMEM; - q = qdisc_create_dflt(sch->dev, sch->dev_queue, + q = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, ops, TC_H_MAKE(sch->handle, 1)); if (q) { err = fifo_set_limit(q, limit); diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index d97086480893..b626a4f32b6b 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -364,7 +364,7 @@ static int pfifo_fast_enqueue(struct sk_buff *skb, struct Qdisc* qdisc) { struct sk_buff_head *list = prio2list(skb, qdisc); - if (skb_queue_len(list) < qdisc->dev->tx_queue_len) { + if (skb_queue_len(list) < qdisc_dev(qdisc)->tx_queue_len) { qdisc->q.qlen++; return __qdisc_enqueue_tail(skb, qdisc, list); } @@ -440,8 +440,7 @@ static struct Qdisc_ops pfifo_fast_ops __read_mostly = { .owner = THIS_MODULE, }; -struct Qdisc *qdisc_alloc(struct net_device *dev, - struct netdev_queue *dev_queue, +struct Qdisc *qdisc_alloc(struct netdev_queue *dev_queue, struct Qdisc_ops *ops) { void *p; @@ -465,8 +464,7 @@ struct Qdisc *qdisc_alloc(struct net_device *dev, sch->enqueue = ops->enqueue; sch->dequeue = ops->dequeue; sch->dev_queue = dev_queue; - sch->dev = dev; - dev_hold(dev); + dev_hold(qdisc_dev(sch)); atomic_set(&sch->refcnt, 1); return sch; @@ -481,7 +479,7 @@ struct Qdisc * qdisc_create_dflt(struct net_device *dev, { struct Qdisc *sch; - sch = qdisc_alloc(dev, dev_queue, ops); + sch = qdisc_alloc(dev_queue, ops); if (IS_ERR(sch)) goto errout; sch->stats_lock = &dev->queue_lock; @@ -534,7 +532,7 @@ void qdisc_destroy(struct Qdisc *qdisc) ops->destroy(qdisc); module_put(ops->owner); - dev_put(qdisc->dev); + dev_put(qdisc_dev(qdisc)); call_rcu(&qdisc->q_rcu, __qdisc_destroy); } EXPORT_SYMBOL(qdisc_destroy); diff --git a/net/sched/sch_gred.c b/net/sched/sch_gred.c index c89fba56db56..39fa28511f07 100644 --- a/net/sched/sch_gred.c +++ b/net/sched/sch_gred.c @@ -164,7 +164,7 @@ static int gred_enqueue(struct sk_buff *skb, struct Qdisc* sch) * if no default DP has been configured. This * allows for DP flows to be left untouched. */ - if (skb_queue_len(&sch->q) < sch->dev->tx_queue_len) + if (skb_queue_len(&sch->q) < qdisc_dev(sch)->tx_queue_len) return qdisc_enqueue_tail(skb, sch); else goto drop; diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 5a22fec4eadd..333525422f45 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1045,7 +1045,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &sch->dev->queue_lock, + &qdisc_dev(sch)->queue_lock, tca[TCA_RATE]); return 0; } @@ -1083,7 +1083,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, cl->refcnt = 1; cl->sched = q; cl->cl_parent = parent; - cl->qdisc = qdisc_create_dflt(sch->dev, sch->dev_queue, + cl->qdisc = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, classid); if (cl->qdisc == NULL) cl->qdisc = &noop_qdisc; @@ -1104,7 +1104,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, if (tca[TCA_RATE]) gen_new_estimator(&cl->bstats, &cl->rate_est, - &sch->dev->queue_lock, tca[TCA_RATE]); + &qdisc_dev(sch)->queue_lock, tca[TCA_RATE]); *arg = (unsigned long)cl; return 0; } @@ -1202,7 +1202,7 @@ hfsc_graft_class(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, if (cl->level > 0) return -EINVAL; if (new == NULL) { - new = qdisc_create_dflt(sch->dev, sch->dev_queue, + new = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, cl->cl_common.classid); if (new == NULL) @@ -1445,7 +1445,7 @@ hfsc_init_qdisc(struct Qdisc *sch, struct nlattr *opt) q->root.cl_common.classid = sch->handle; q->root.refcnt = 1; q->root.sched = q; - q->root.qdisc = qdisc_create_dflt(sch->dev, sch->dev_queue, + q->root.qdisc = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, sch->handle); if (q->root.qdisc == NULL) diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 956a67f66b9c..31f7d1536e6d 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -1026,7 +1026,7 @@ static int htb_init(struct Qdisc *sch, struct nlattr *opt) qdisc_watchdog_init(&q->watchdog, sch); skb_queue_head_init(&q->direct_queue); - q->direct_qlen = sch->dev->tx_queue_len; + q->direct_qlen = qdisc_dev(sch)->tx_queue_len; if (q->direct_qlen < 2) /* some devices have zero tx_queue_len */ q->direct_qlen = 2; @@ -1043,7 +1043,7 @@ static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) struct nlattr *nest; struct tc_htb_glob gopt; - spin_lock_bh(&sch->dev->queue_lock); + spin_lock_bh(&qdisc_dev(sch)->queue_lock); gopt.direct_pkts = q->direct_pkts; gopt.version = HTB_VER; @@ -1057,11 +1057,11 @@ static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) NLA_PUT(skb, TCA_HTB_INIT, sizeof(gopt), &gopt); nla_nest_end(skb, nest); - spin_unlock_bh(&sch->dev->queue_lock); + spin_unlock_bh(&qdisc_dev(sch)->queue_lock); return skb->len; nla_put_failure: - spin_unlock_bh(&sch->dev->queue_lock); + spin_unlock_bh(&qdisc_dev(sch)->queue_lock); nla_nest_cancel(skb, nest); return -1; } @@ -1073,7 +1073,7 @@ static int htb_dump_class(struct Qdisc *sch, unsigned long arg, struct nlattr *nest; struct tc_htb_opt opt; - spin_lock_bh(&sch->dev->queue_lock); + spin_lock_bh(&qdisc_dev(sch)->queue_lock); tcm->tcm_parent = cl->parent ? cl->parent->common.classid : TC_H_ROOT; tcm->tcm_handle = cl->common.classid; if (!cl->level && cl->un.leaf.q) @@ -1095,11 +1095,11 @@ static int htb_dump_class(struct Qdisc *sch, unsigned long arg, NLA_PUT(skb, TCA_HTB_PARMS, sizeof(opt), &opt); nla_nest_end(skb, nest); - spin_unlock_bh(&sch->dev->queue_lock); + spin_unlock_bh(&qdisc_dev(sch)->queue_lock); return skb->len; nla_put_failure: - spin_unlock_bh(&sch->dev->queue_lock); + spin_unlock_bh(&qdisc_dev(sch)->queue_lock); nla_nest_cancel(skb, nest); return -1; } @@ -1129,7 +1129,7 @@ static int htb_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, if (cl && !cl->level) { if (new == NULL && - (new = qdisc_create_dflt(sch->dev, sch->dev_queue, + (new = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, cl->common.classid)) == NULL) @@ -1257,7 +1257,7 @@ static int htb_delete(struct Qdisc *sch, unsigned long arg) return -EBUSY; if (!cl->level && htb_parent_last_child(cl)) { - new_q = qdisc_create_dflt(sch->dev, sch->dev_queue, + new_q = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, cl->parent->common.classid); last_child = 1; @@ -1365,7 +1365,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, goto failure; gen_new_estimator(&cl->bstats, &cl->rate_est, - &sch->dev->queue_lock, + &qdisc_dev(sch)->queue_lock, tca[TCA_RATE] ? : &est.nla); cl->refcnt = 1; cl->children = 0; @@ -1378,7 +1378,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, /* create leaf qdisc early because it uses kmalloc(GFP_KERNEL) so that can't be used inside of sch_tree_lock -- thanks to Karlis Peisenieks */ - new_q = qdisc_create_dflt(sch->dev, sch->dev_queue, + new_q = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, classid); sch_tree_lock(sch); if (parent && !parent->level) { @@ -1420,7 +1420,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, } else { if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &sch->dev->queue_lock, + &qdisc_dev(sch)->queue_lock, tca[TCA_RATE]); sch_tree_lock(sch); } diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index aa7a04e32ae9..790582960444 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -180,7 +180,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) * skb will be queued. */ if (count > 1 && (skb2 = skb_clone(skb, GFP_ATOMIC)) != NULL) { - struct Qdisc *rootq = sch->dev->qdisc; + struct Qdisc *rootq = qdisc_dev(sch)->qdisc; u32 dupsave = q->duplicate; /* prevent duplicating a dup... */ q->duplicate = 0; @@ -333,9 +333,9 @@ static int get_dist_table(struct Qdisc *sch, const struct nlattr *attr) for (i = 0; i < n; i++) d->table[i] = data[i]; - spin_lock_bh(&sch->dev->queue_lock); + spin_lock_bh(&qdisc_dev(sch)->queue_lock); d = xchg(&q->delay_dist, d); - spin_unlock_bh(&sch->dev->queue_lock); + spin_unlock_bh(&qdisc_dev(sch)->queue_lock); kfree(d); return 0; @@ -495,7 +495,7 @@ static int tfifo_init(struct Qdisc *sch, struct nlattr *opt) q->limit = ctl->limit; } else - q->limit = max_t(u32, sch->dev->tx_queue_len, 1); + q->limit = max_t(u32, qdisc_dev(sch)->tx_queue_len, 1); q->oldest = PSCHED_PASTPERFECT; return 0; @@ -536,7 +536,7 @@ static int netem_init(struct Qdisc *sch, struct nlattr *opt) qdisc_watchdog_init(&q->watchdog, sch); - q->qdisc = qdisc_create_dflt(sch->dev, sch->dev_queue, + q->qdisc = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &tfifo_qdisc_ops, TC_H_MAKE(sch->handle, 1)); if (!q->qdisc) { diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c index ca58a039208e..39157f7bc046 100644 --- a/net/sched/sch_prio.c +++ b/net/sched/sch_prio.c @@ -136,7 +136,8 @@ prio_dequeue(struct Qdisc* sch) * pulling an skb. This way we avoid excessive requeues * for slower queues. */ - if (!__netif_subqueue_stopped(sch->dev, (q->mq ? prio : 0))) { + if (!__netif_subqueue_stopped(qdisc_dev(sch), + (q->mq ? prio : 0))) { qdisc = q->queues[prio]; skb = qdisc->dequeue(qdisc); if (skb) { @@ -165,8 +166,8 @@ static struct sk_buff *rr_dequeue(struct Qdisc* sch) * for slower queues. If the queue is stopped, try the * next queue. */ - if (!__netif_subqueue_stopped(sch->dev, - (q->mq ? q->curband : 0))) { + if (!__netif_subqueue_stopped(qdisc_dev(sch), + (q->mq ? q->curband : 0))) { qdisc = q->queues[q->curband]; skb = qdisc->dequeue(qdisc); if (skb) { @@ -249,10 +250,10 @@ static int prio_tune(struct Qdisc *sch, struct nlattr *opt) if (q->mq) { if (sch->parent != TC_H_ROOT) return -EINVAL; - if (netif_is_multiqueue(sch->dev)) { + if (netif_is_multiqueue(qdisc_dev(sch))) { if (q->bands == 0) - q->bands = sch->dev->egress_subqueue_count; - else if (q->bands != sch->dev->egress_subqueue_count) + q->bands = qdisc_dev(sch)->egress_subqueue_count; + else if (q->bands != qdisc_dev(sch)->egress_subqueue_count) return -EINVAL; } else return -EOPNOTSUPP; @@ -281,7 +282,7 @@ static int prio_tune(struct Qdisc *sch, struct nlattr *opt) for (i=0; ibands; i++) { if (q->queues[i] == &noop_qdisc) { struct Qdisc *child; - child = qdisc_create_dflt(sch->dev, sch->dev_queue, + child = qdisc_create_dflt(qdisc_dev(sch), sch->dev_queue, &pfifo_qdisc_ops, TC_H_MAKE(sch->handle, i + 1)); if (child) { diff --git a/net/sched/sch_sfq.c b/net/sched/sch_sfq.c index 6a97afbfb952..8458f630fac4 100644 --- a/net/sched/sch_sfq.c +++ b/net/sched/sch_sfq.c @@ -461,7 +461,7 @@ static int sfq_change(struct Qdisc *sch, struct nlattr *opt) return -EINVAL; sch_tree_lock(sch); - q->quantum = ctl->quantum ? : psched_mtu(sch->dev); + q->quantum = ctl->quantum ? : psched_mtu(qdisc_dev(sch)); q->perturb_period = ctl->perturb_period * HZ; if (ctl->limit) q->limit = min_t(u32, ctl->limit, SFQ_DEPTH - 1); @@ -502,7 +502,7 @@ static int sfq_init(struct Qdisc *sch, struct nlattr *opt) q->max_depth = 0; q->tail = SFQ_DEPTH; if (opt == NULL) { - q->quantum = psched_mtu(sch->dev); + q->quantum = psched_mtu(qdisc_dev(sch)); q->perturb_period = 0; q->perturbation = net_random(); } else { diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index 0444fd0f0d22..b3fc82623fc6 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -78,7 +78,7 @@ struct teql_sched_data static int teql_enqueue(struct sk_buff *skb, struct Qdisc* sch) { - struct net_device *dev = sch->dev; + struct net_device *dev = qdisc_dev(sch); struct teql_sched_data *q = qdisc_priv(sch); if (q->q.qlen < dev->tx_queue_len) { @@ -111,7 +111,7 @@ teql_dequeue(struct Qdisc* sch) skb = __skb_dequeue(&dat->q); if (skb == NULL) { - struct net_device *m = dat->m->dev->qdisc->dev; + struct net_device *m = qdisc_dev(dat->m->dev->qdisc); if (m) { dat->m->slaves = sch; netif_wake_queue(m); @@ -170,7 +170,7 @@ teql_destroy(struct Qdisc* sch) static int teql_qdisc_init(struct Qdisc *sch, struct nlattr *opt) { - struct net_device *dev = sch->dev; + struct net_device *dev = qdisc_dev(sch); struct teql_master *m = (struct teql_master*)sch->ops; struct teql_sched_data *q = qdisc_priv(sch); @@ -282,7 +282,7 @@ restart: goto drop; do { - struct net_device *slave = q->dev; + struct net_device *slave = qdisc_dev(q); if (slave->qdisc_sleeping != q) continue; @@ -352,7 +352,7 @@ static int teql_master_open(struct net_device *dev) q = m->slaves; do { - struct net_device *slave = q->dev; + struct net_device *slave = qdisc_dev(q); if (slave == NULL) return -EUNATCH; @@ -403,7 +403,7 @@ static int teql_master_mtu(struct net_device *dev, int new_mtu) q = m->slaves; if (q) { do { - if (new_mtu > q->dev->mtu) + if (new_mtu > qdisc_dev(q)->mtu) return -EINVAL; } while ((q=NEXT_SLAVE(q)) != m->slaves); } -- cgit v1.2.2 From dc2b48475a0a36f8b3bbb2da60d3a006dc5c2c84 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 17:18:23 -0700 Subject: netdev: Move queue_lock into struct netdev_queue. The lock is now an attribute of the device queue. One thing to notice is that "suspicious" places emerge which will need specific training about multiple queue handling. They are so marked with explicit "netdev->rx_queue" and "netdev->tx_queue" references. Signed-off-by: David S. Miller --- net/core/dev.c | 33 ++++++++++++++++++++++----------- net/mac80211/main.c | 10 +++++----- net/mac80211/wme.c | 2 +- net/sched/sch_api.c | 2 +- net/sched/sch_cbq.c | 8 ++++---- net/sched/sch_generic.c | 40 ++++++++++++++++++++-------------------- net/sched/sch_hfsc.c | 4 ++-- net/sched/sch_htb.c | 16 ++++++++-------- net/sched/sch_netem.c | 4 ++-- net/sched/sch_teql.c | 4 ++-- 10 files changed, 67 insertions(+), 56 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 9b281c906eb0..05011048b86c 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1667,6 +1667,7 @@ out_kfree_skb: int dev_queue_xmit(struct sk_buff *skb) { struct net_device *dev = skb->dev; + struct netdev_queue *txq; struct Qdisc *q; int rc = -ENOMEM; @@ -1699,14 +1700,15 @@ int dev_queue_xmit(struct sk_buff *skb) } gso: - spin_lock_prefetch(&dev->queue_lock); + txq = &dev->tx_queue; + spin_lock_prefetch(&txq->lock); /* Disable soft irqs for various locks below. Also * stops preemption for RCU. */ rcu_read_lock_bh(); - /* Updates of qdisc are serialized by queue_lock. + /* Updates of qdisc are serialized by queue->lock. * The struct Qdisc which is pointed to by qdisc is now a * rcu structure - it may be accessed without acquiring * a lock (but the structure may be stale.) The freeing of the @@ -1714,7 +1716,7 @@ gso: * more references to it. * * If the qdisc has an enqueue function, we still need to - * hold the queue_lock before calling it, since queue_lock + * hold the queue->lock before calling it, since queue->lock * also serializes access to the device queue. */ @@ -1724,19 +1726,19 @@ gso: #endif if (q->enqueue) { /* Grab device queue */ - spin_lock(&dev->queue_lock); + spin_lock(&txq->lock); q = dev->qdisc; if (q->enqueue) { /* reset queue_mapping to zero */ skb_set_queue_mapping(skb, 0); rc = q->enqueue(skb, q); qdisc_run(dev); - spin_unlock(&dev->queue_lock); + spin_unlock(&txq->lock); rc = rc == NET_XMIT_BYPASS ? NET_XMIT_SUCCESS : rc; goto out; } - spin_unlock(&dev->queue_lock); + spin_unlock(&txq->lock); } /* The device has no queue. Common case for software devices: @@ -1919,14 +1921,17 @@ static void net_tx_action(struct softirq_action *h) while (head) { struct net_device *dev = head; + struct netdev_queue *txq; head = head->next_sched; + txq = &dev->tx_queue; + smp_mb__before_clear_bit(); clear_bit(__LINK_STATE_SCHED, &dev->state); - if (spin_trylock(&dev->queue_lock)) { + if (spin_trylock(&txq->lock)) { qdisc_run(dev); - spin_unlock(&dev->queue_lock); + spin_unlock(&txq->lock); } else { netif_schedule(dev); } @@ -3787,7 +3792,6 @@ int register_netdevice(struct net_device *dev) BUG_ON(!dev_net(dev)); net = dev_net(dev); - spin_lock_init(&dev->queue_lock); spin_lock_init(&dev->_xmit_lock); netdev_set_lockdep_class(&dev->_xmit_lock, dev->type); dev->xmit_lock_owner = -1; @@ -4072,10 +4076,17 @@ static struct net_device_stats *internal_stats(struct net_device *dev) return &dev->stats; } +static void netdev_init_one_queue(struct net_device *dev, + struct netdev_queue *queue) +{ + spin_lock_init(&queue->lock); + queue->dev = dev; +} + static void netdev_init_queues(struct net_device *dev) { - dev->rx_queue.dev = dev; - dev->tx_queue.dev = dev; + netdev_init_one_queue(dev, &dev->rx_queue); + netdev_init_one_queue(dev, &dev->tx_queue); } /** diff --git a/net/mac80211/main.c b/net/mac80211/main.c index cf477ad39dac..12aeaf78ae75 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -636,7 +636,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) /* ensure that TX flow won't interrupt us * until the end of the call to requeue function */ - spin_lock_bh(&local->mdev->queue_lock); + spin_lock_bh(&local->mdev->tx_queue.lock); /* create a new queue for this aggregation */ ret = ieee80211_ht_agg_queue_add(local, sta, tid); @@ -675,7 +675,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) /* Will put all the packets in the new SW queue */ ieee80211_requeue(local, ieee802_1d_to_ac[tid]); - spin_unlock_bh(&local->mdev->queue_lock); + spin_unlock_bh(&local->mdev->tx_queue.lock); spin_unlock_bh(&sta->lock); /* send an addBA request */ @@ -701,7 +701,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) err_unlock_queue: kfree(sta->ampdu_mlme.tid_tx[tid]); sta->ampdu_mlme.tid_tx[tid] = NULL; - spin_unlock_bh(&local->mdev->queue_lock); + spin_unlock_bh(&local->mdev->tx_queue.lock); ret = -EBUSY; err_unlock_sta: spin_unlock_bh(&sta->lock); @@ -875,10 +875,10 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) /* avoid ordering issues: we are the only one that can modify * the content of the qdiscs */ - spin_lock_bh(&local->mdev->queue_lock); + spin_lock_bh(&local->mdev->tx_queue.lock); /* remove the queue for this aggregation */ ieee80211_ht_agg_queue_remove(local, sta, tid, 1); - spin_unlock_bh(&local->mdev->queue_lock); + spin_unlock_bh(&local->mdev->tx_queue.lock); /* we just requeued the all the frames that were in the removed * queue, and since we might miss a softirq we do netif_schedule. diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 2fbc171130bf..59ed9cae66b9 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -648,7 +648,7 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, } /** - * the caller needs to hold local->mdev->queue_lock + * the caller needs to hold local->mdev->tx_queue.lock */ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, struct sta_info *sta, u16 tid, diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 1f893082a4f6..2a1834f8c7d8 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -606,7 +606,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, sch->stats_lock = &dev->ingress_lock; handle = TC_H_MAKE(TC_H_INGRESS, 0); } else { - sch->stats_lock = &dev->queue_lock; + sch->stats_lock = &dev_queue->lock; if (handle == 0) { handle = qdisc_alloc_handle(dev); err = -ENOMEM; diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 9f2ace585fd6..99ce3da2b0a4 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -1746,10 +1746,10 @@ static void cbq_put(struct Qdisc *sch, unsigned long arg) #ifdef CONFIG_NET_CLS_ACT struct cbq_sched_data *q = qdisc_priv(sch); - spin_lock_bh(&qdisc_dev(sch)->queue_lock); + spin_lock_bh(&sch->dev_queue->lock); if (q->rx_class == cl) q->rx_class = NULL; - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); #endif cbq_destroy_class(sch, cl); @@ -1828,7 +1828,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, + &sch->dev_queue->lock, tca[TCA_RATE]); return 0; } @@ -1919,7 +1919,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t if (tca[TCA_RATE]) gen_new_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, tca[TCA_RATE]); + &sch->dev_queue->lock, tca[TCA_RATE]); *arg = (unsigned long)cl; return 0; diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index b626a4f32b6b..ee8f9f78a095 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -29,31 +29,31 @@ /* Main transmission queue. */ /* Modifications to data participating in scheduling must be protected with - * dev->queue_lock spinlock. + * queue->lock spinlock. * * The idea is the following: * - enqueue, dequeue are serialized via top level device - * spinlock dev->queue_lock. + * spinlock queue->lock. * - ingress filtering is serialized via top level device * spinlock dev->ingress_lock. * - updates to tree and tree walking are only done under the rtnl mutex. */ void qdisc_lock_tree(struct net_device *dev) - __acquires(dev->queue_lock) + __acquires(dev->tx_queue.lock) __acquires(dev->ingress_lock) { - spin_lock_bh(&dev->queue_lock); + spin_lock_bh(&dev->tx_queue.lock); spin_lock(&dev->ingress_lock); } EXPORT_SYMBOL(qdisc_lock_tree); void qdisc_unlock_tree(struct net_device *dev) __releases(dev->ingress_lock) - __releases(dev->queue_lock) + __releases(dev->tx_queue.lock) { spin_unlock(&dev->ingress_lock); - spin_unlock_bh(&dev->queue_lock); + spin_unlock_bh(&dev->tx_queue.lock); } EXPORT_SYMBOL(qdisc_unlock_tree); @@ -118,15 +118,15 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, } /* - * NOTE: Called under dev->queue_lock with locally disabled BH. + * NOTE: Called under queue->lock with locally disabled BH. * * __LINK_STATE_QDISC_RUNNING guarantees only one CPU can process this - * device at a time. dev->queue_lock serializes queue accesses for + * device at a time. queue->lock serializes queue accesses for * this device AND dev->qdisc pointer itself. * * netif_tx_lock serializes accesses to device driver. * - * dev->queue_lock and netif_tx_lock are mutually exclusive, + * queue->lock and netif_tx_lock are mutually exclusive, * if one is grabbed, another must be free. * * Note, that this procedure can be called by a watchdog timer @@ -148,14 +148,14 @@ static inline int qdisc_restart(struct net_device *dev) /* And release queue */ - spin_unlock(&dev->queue_lock); + spin_unlock(&q->dev_queue->lock); HARD_TX_LOCK(dev, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) ret = dev_hard_start_xmit(skb, dev); HARD_TX_UNLOCK(dev); - spin_lock(&dev->queue_lock); + spin_lock(&q->dev_queue->lock); q = dev->qdisc; switch (ret) { @@ -482,7 +482,7 @@ struct Qdisc * qdisc_create_dflt(struct net_device *dev, sch = qdisc_alloc(dev_queue, ops); if (IS_ERR(sch)) goto errout; - sch->stats_lock = &dev->queue_lock; + sch->stats_lock = &dev_queue->lock; sch->parent = parentid; if (!ops->init || ops->init(sch, NULL) == 0) @@ -494,7 +494,7 @@ errout: } EXPORT_SYMBOL(qdisc_create_dflt); -/* Under dev->queue_lock and BH! */ +/* Under queue->lock and BH! */ void qdisc_reset(struct Qdisc *qdisc) { @@ -514,7 +514,7 @@ static void __qdisc_destroy(struct rcu_head *head) kfree((char *) qdisc - qdisc->padded); } -/* Under dev->queue_lock and BH! */ +/* Under queue->lock and BH! */ void qdisc_destroy(struct Qdisc *qdisc) { @@ -566,13 +566,13 @@ void dev_activate(struct net_device *dev) /* Delay activation until next carrier-on event */ return; - spin_lock_bh(&dev->queue_lock); + spin_lock_bh(&dev->tx_queue.lock); rcu_assign_pointer(dev->qdisc, dev->qdisc_sleeping); if (dev->qdisc != &noqueue_qdisc) { dev->trans_start = jiffies; dev_watchdog_up(dev); } - spin_unlock_bh(&dev->queue_lock); + spin_unlock_bh(&dev->tx_queue.lock); } void dev_deactivate(struct net_device *dev) @@ -581,7 +581,7 @@ void dev_deactivate(struct net_device *dev) struct sk_buff *skb; int running; - spin_lock_bh(&dev->queue_lock); + spin_lock_bh(&dev->tx_queue.lock); qdisc = dev->qdisc; dev->qdisc = &noop_qdisc; @@ -589,7 +589,7 @@ void dev_deactivate(struct net_device *dev) skb = dev->gso_skb; dev->gso_skb = NULL; - spin_unlock_bh(&dev->queue_lock); + spin_unlock_bh(&dev->tx_queue.lock); kfree_skb(skb); @@ -607,9 +607,9 @@ void dev_deactivate(struct net_device *dev) * Double-check inside queue lock to ensure that all effects * of the queue run are visible when we return. */ - spin_lock_bh(&dev->queue_lock); + spin_lock_bh(&dev->tx_queue.lock); running = test_bit(__LINK_STATE_QDISC_RUNNING, &dev->state); - spin_unlock_bh(&dev->queue_lock); + spin_unlock_bh(&dev->tx_queue.lock); /* * The running flag should never be set at this point because diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 333525422f45..997d520ca580 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1045,7 +1045,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, + &sch->dev_queue->lock, tca[TCA_RATE]); return 0; } @@ -1104,7 +1104,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, if (tca[TCA_RATE]) gen_new_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, tca[TCA_RATE]); + &sch->dev_queue->lock, tca[TCA_RATE]); *arg = (unsigned long)cl; return 0; } diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 31f7d1536e6d..c8ca54cc26b0 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -1043,7 +1043,7 @@ static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) struct nlattr *nest; struct tc_htb_glob gopt; - spin_lock_bh(&qdisc_dev(sch)->queue_lock); + spin_lock_bh(&sch->dev_queue->lock); gopt.direct_pkts = q->direct_pkts; gopt.version = HTB_VER; @@ -1057,11 +1057,11 @@ static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) NLA_PUT(skb, TCA_HTB_INIT, sizeof(gopt), &gopt); nla_nest_end(skb, nest); - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); return skb->len; nla_put_failure: - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); nla_nest_cancel(skb, nest); return -1; } @@ -1073,7 +1073,7 @@ static int htb_dump_class(struct Qdisc *sch, unsigned long arg, struct nlattr *nest; struct tc_htb_opt opt; - spin_lock_bh(&qdisc_dev(sch)->queue_lock); + spin_lock_bh(&sch->dev_queue->lock); tcm->tcm_parent = cl->parent ? cl->parent->common.classid : TC_H_ROOT; tcm->tcm_handle = cl->common.classid; if (!cl->level && cl->un.leaf.q) @@ -1095,11 +1095,11 @@ static int htb_dump_class(struct Qdisc *sch, unsigned long arg, NLA_PUT(skb, TCA_HTB_PARMS, sizeof(opt), &opt); nla_nest_end(skb, nest); - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); return skb->len; nla_put_failure: - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); nla_nest_cancel(skb, nest); return -1; } @@ -1365,7 +1365,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, goto failure; gen_new_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, + &sch->dev_queue->lock, tca[TCA_RATE] ? : &est.nla); cl->refcnt = 1; cl->children = 0; @@ -1420,7 +1420,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, } else { if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &qdisc_dev(sch)->queue_lock, + &sch->dev_queue->lock, tca[TCA_RATE]); sch_tree_lock(sch); } diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index 790582960444..71b73c528f9b 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -333,9 +333,9 @@ static int get_dist_table(struct Qdisc *sch, const struct nlattr *attr) for (i = 0; i < n; i++) d->table[i] = data[i]; - spin_lock_bh(&qdisc_dev(sch)->queue_lock); + spin_lock_bh(&sch->dev_queue->lock); d = xchg(&q->delay_dist, d); - spin_unlock_bh(&qdisc_dev(sch)->queue_lock); + spin_unlock_bh(&sch->dev_queue->lock); kfree(d); return 0; diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index b3fc82623fc6..4f3054e8e1ab 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -154,9 +154,9 @@ teql_destroy(struct Qdisc* sch) master->slaves = NEXT_SLAVE(q); if (q == master->slaves) { master->slaves = NULL; - spin_lock_bh(&master->dev->queue_lock); + spin_lock_bh(&master->dev->tx_queue.lock); qdisc_reset(master->dev->qdisc); - spin_unlock_bh(&master->dev->queue_lock); + spin_unlock_bh(&master->dev->tx_queue.lock); } } skb_queue_purge(&dat->q); -- cgit v1.2.2 From 555353cfa1aee293de445bfa6de43276138ddd82 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 17:33:13 -0700 Subject: netdev: The ingress_lock member is no longer needed. Every qdisc is assosciated with a queue, and in the case of ingress qdiscs that will now be netdev->rx_queue so using that queue's lock is the thing to do. Signed-off-by: David S. Miller --- net/core/dev.c | 12 +++++++----- net/sched/sch_api.c | 3 +-- net/sched/sch_generic.c | 10 +++++----- 3 files changed, 13 insertions(+), 12 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 05011048b86c..2322fb69fd53 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2014,10 +2014,11 @@ static inline struct sk_buff *handle_macvlan(struct sk_buff *skb, */ static int ing_filter(struct sk_buff *skb) { - struct Qdisc *q; struct net_device *dev = skb->dev; - int result = TC_ACT_OK; u32 ttl = G_TC_RTTL(skb->tc_verd); + struct netdev_queue *rxq; + int result = TC_ACT_OK; + struct Qdisc *q; if (MAX_RED_LOOP < ttl++) { printk(KERN_WARNING @@ -2029,10 +2030,12 @@ static int ing_filter(struct sk_buff *skb) skb->tc_verd = SET_TC_RTTL(skb->tc_verd, ttl); skb->tc_verd = SET_TC_AT(skb->tc_verd, AT_INGRESS); - spin_lock(&dev->ingress_lock); + rxq = &dev->rx_queue; + + spin_lock(&rxq->lock); if ((q = dev->qdisc_ingress) != NULL) result = q->enqueue(skb, q); - spin_unlock(&dev->ingress_lock); + spin_unlock(&rxq->lock); return result; } @@ -3795,7 +3798,6 @@ int register_netdevice(struct net_device *dev) spin_lock_init(&dev->_xmit_lock); netdev_set_lockdep_class(&dev->_xmit_lock, dev->type); dev->xmit_lock_owner = -1; - spin_lock_init(&dev->ingress_lock); dev->iflink = -1; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 2a1834f8c7d8..570cef2a9c5f 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -601,12 +601,11 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, sch->parent = parent; + sch->stats_lock = &dev_queue->lock; if (handle == TC_H_INGRESS) { sch->flags |= TCQ_F_INGRESS; - sch->stats_lock = &dev->ingress_lock; handle = TC_H_MAKE(TC_H_INGRESS, 0); } else { - sch->stats_lock = &dev_queue->lock; if (handle == 0) { handle = qdisc_alloc_handle(dev); err = -ENOMEM; diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index ee8f9f78a095..804d44b00348 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -35,24 +35,24 @@ * - enqueue, dequeue are serialized via top level device * spinlock queue->lock. * - ingress filtering is serialized via top level device - * spinlock dev->ingress_lock. + * spinlock dev->rx_queue.lock. * - updates to tree and tree walking are only done under the rtnl mutex. */ void qdisc_lock_tree(struct net_device *dev) __acquires(dev->tx_queue.lock) - __acquires(dev->ingress_lock) + __acquires(dev->rx_queue.lock) { spin_lock_bh(&dev->tx_queue.lock); - spin_lock(&dev->ingress_lock); + spin_lock(&dev->rx_queue.lock); } EXPORT_SYMBOL(qdisc_lock_tree); void qdisc_unlock_tree(struct net_device *dev) - __releases(dev->ingress_lock) + __releases(dev->rx_queue.lock) __releases(dev->tx_queue.lock) { - spin_unlock(&dev->ingress_lock); + spin_unlock(&dev->rx_queue.lock); spin_unlock_bh(&dev->tx_queue.lock); } EXPORT_SYMBOL(qdisc_unlock_tree); -- cgit v1.2.2 From b0e1e6462df3c5944010b3328a546d8fe5d932cd Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 17:42:10 -0700 Subject: netdev: Move rest of qdisc state into struct netdev_queue Now qdisc, qdisc_sleeping, and qdisc_list also live there. Signed-off-by: David S. Miller --- net/core/dev.c | 4 +-- net/core/link_watch.c | 8 +++-- net/core/rtnetlink.c | 6 ++-- net/ipv6/addrconf.c | 3 +- net/mac80211/wme.c | 20 +++++++---- net/sched/cls_api.c | 7 ++-- net/sched/sch_api.c | 34 ++++++++++++------- net/sched/sch_generic.c | 90 +++++++++++++++++++++++++++++++------------------ net/sched/sch_netem.c | 2 +- net/sched/sch_teql.c | 14 ++++---- 10 files changed, 120 insertions(+), 68 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 2322fb69fd53..ce79c28d739d 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1720,14 +1720,14 @@ gso: * also serializes access to the device queue. */ - q = rcu_dereference(dev->qdisc); + q = rcu_dereference(txq->qdisc); #ifdef CONFIG_NET_CLS_ACT skb->tc_verd = SET_TC_AT(skb->tc_verd,AT_EGRESS); #endif if (q->enqueue) { /* Grab device queue */ spin_lock(&txq->lock); - q = dev->qdisc; + q = txq->qdisc; if (q->enqueue) { /* reset queue_mapping to zero */ skb_set_queue_mapping(skb, 0); diff --git a/net/core/link_watch.c b/net/core/link_watch.c index a5e372b9ec4d..50218218445b 100644 --- a/net/core/link_watch.c +++ b/net/core/link_watch.c @@ -79,8 +79,10 @@ static void rfc2863_policy(struct net_device *dev) static int linkwatch_urgent_event(struct net_device *dev) { + struct netdev_queue *txq = &dev->tx_queue; + return netif_running(dev) && netif_carrier_ok(dev) && - dev->qdisc != dev->qdisc_sleeping; + txq->qdisc != txq->qdisc_sleeping; } @@ -181,7 +183,9 @@ static void __linkwatch_run_queue(int urgent_only) rfc2863_policy(dev); if (dev->flags & IFF_UP) { if (netif_carrier_ok(dev)) { - WARN_ON(dev->qdisc_sleeping == &noop_qdisc); + struct netdev_queue *txq = &dev->tx_queue; + + WARN_ON(txq->qdisc_sleeping == &noop_qdisc); dev_activate(dev); } else dev_deactivate(dev); diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 6c8d7f0ea01a..8ef9f1db610e 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -605,6 +605,7 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev, int type, u32 pid, u32 seq, u32 change, unsigned int flags) { + struct netdev_queue *txq; struct ifinfomsg *ifm; struct nlmsghdr *nlh; struct net_device_stats *stats; @@ -635,8 +636,9 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev, if (dev->master) NLA_PUT_U32(skb, IFLA_MASTER, dev->master->ifindex); - if (dev->qdisc_sleeping) - NLA_PUT_STRING(skb, IFLA_QDISC, dev->qdisc_sleeping->ops->id); + txq = &dev->tx_queue; + if (txq->qdisc_sleeping) + NLA_PUT_STRING(skb, IFLA_QDISC, txq->qdisc_sleeping->ops->id); if (1) { struct rtnl_link_ifmap map = { diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 8572cb05fc21..5c84c798331d 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -231,7 +231,8 @@ const struct in6_addr in6addr_linklocal_allrouters = IN6ADDR_LINKLOCAL_ALLROUTER /* Check if a valid qdisc is available */ static inline int addrconf_qdisc_ok(struct net_device *dev) { - return (dev->qdisc != &noop_qdisc); + struct netdev_queue *txq = &dev->tx_queue; + return (txq->qdisc != &noop_qdisc); } /* Check if a route is valid prefix route */ diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 59ed9cae66b9..6ae43a3c7726 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -574,9 +574,10 @@ static struct Qdisc_ops wme_qdisc_ops __read_mostly = void ieee80211_install_qdisc(struct net_device *dev) { + struct netdev_queue *txq = &dev->tx_queue; struct Qdisc *qdisc; - qdisc = qdisc_create_dflt(dev, &dev->tx_queue, + qdisc = qdisc_create_dflt(dev, txq, &wme_qdisc_ops, TC_H_ROOT); if (!qdisc) { printk(KERN_ERR "%s: qdisc installation failed\n", dev->name); @@ -587,15 +588,17 @@ void ieee80211_install_qdisc(struct net_device *dev) qdisc->handle = 0x80010000; qdisc_lock_tree(dev); - list_add_tail(&qdisc->list, &dev->qdisc_list); - dev->qdisc_sleeping = qdisc; + list_add_tail(&qdisc->list, &txq->qdisc_list); + txq->qdisc_sleeping = qdisc; qdisc_unlock_tree(dev); } int ieee80211_qdisc_installed(struct net_device *dev) { - return dev->qdisc_sleeping->ops == &wme_qdisc_ops; + struct netdev_queue *txq = &dev->tx_queue; + + return txq->qdisc_sleeping->ops == &wme_qdisc_ops; } @@ -614,8 +617,9 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, struct sta_info *sta, u16 tid) { int i; + struct netdev_queue *txq = &local->mdev->tx_queue; struct ieee80211_sched_data *q = - qdisc_priv(local->mdev->qdisc_sleeping); + qdisc_priv(txq->qdisc_sleeping); DECLARE_MAC_BUF(mac); /* prepare the filter and save it for the SW queue @@ -655,8 +659,9 @@ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, u8 requeue) { struct ieee80211_hw *hw = &local->hw; + struct netdev_queue *txq = &local->mdev->tx_queue; struct ieee80211_sched_data *q = - qdisc_priv(local->mdev->qdisc_sleeping); + qdisc_priv(txq->qdisc_sleeping); int agg_queue = sta->tid_to_tx_q[tid]; /* return the qdisc to the pool */ @@ -671,7 +676,8 @@ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, void ieee80211_requeue(struct ieee80211_local *local, int queue) { - struct Qdisc *root_qd = local->mdev->qdisc_sleeping; + struct netdev_queue *txq = &local->mdev->tx_queue; + struct Qdisc *root_qd = txq->qdisc_sleeping; struct ieee80211_sched_data *q = qdisc_priv(root_qd); struct Qdisc *qdisc = q->queues[queue]; struct sk_buff *skb = NULL; diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c index e2389f161e46..b483bbea6118 100644 --- a/net/sched/cls_api.c +++ b/net/sched/cls_api.c @@ -166,7 +166,8 @@ replay: /* Find qdisc */ if (!parent) { - q = dev->qdisc_sleeping; + struct netdev_queue *dev_queue = &dev->tx_queue; + q = dev_queue->qdisc_sleeping; parent = q->handle; } else { q = qdisc_lookup(dev, TC_H_MAJ(t->tcm_parent)); @@ -390,6 +391,7 @@ static int tcf_node_dump(struct tcf_proto *tp, unsigned long n, static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb) { struct net *net = sock_net(skb->sk); + struct netdev_queue *dev_queue; int t; int s_t; struct net_device *dev; @@ -408,8 +410,9 @@ static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb) if ((dev = dev_get_by_index(&init_net, tcm->tcm_ifindex)) == NULL) return skb->len; + dev_queue = &dev->tx_queue; if (!tcm->tcm_parent) - q = dev->qdisc_sleeping; + q = dev_queue->qdisc_sleeping; else q = qdisc_lookup(dev, TC_H_MAJ(tcm->tcm_parent)); if (!q) diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 570cef2a9c5f..2313fa7c97be 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -185,9 +185,10 @@ EXPORT_SYMBOL(unregister_qdisc); struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) { + struct netdev_queue *dev_queue = &dev->tx_queue; struct Qdisc *q; - list_for_each_entry(q, &dev->qdisc_list, list) { + list_for_each_entry(q, &dev_queue->qdisc_list, list) { if (q->handle == handle) return q; } @@ -441,6 +442,7 @@ static u32 qdisc_alloc_handle(struct net_device *dev) static struct Qdisc * dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) { + struct netdev_queue *dev_queue; struct Qdisc *oqdisc; if (dev->flags & IFF_UP) @@ -459,8 +461,8 @@ dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) } } else { - - oqdisc = dev->qdisc_sleeping; + dev_queue = &dev->tx_queue; + oqdisc = dev_queue->qdisc_sleeping; /* Prune old scheduler */ if (oqdisc && atomic_read(&oqdisc->refcnt) <= 1) @@ -469,8 +471,8 @@ dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) /* ... and graft new one */ if (qdisc == NULL) qdisc = &noop_qdisc; - dev->qdisc_sleeping = qdisc; - dev->qdisc = &noop_qdisc; + dev_queue->qdisc_sleeping = qdisc; + dev_queue->qdisc = &noop_qdisc; } qdisc_unlock_tree(dev); @@ -633,7 +635,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, } } qdisc_lock_tree(dev); - list_add_tail(&sch->list, &dev->qdisc_list); + list_add_tail(&sch->list, &dev_queue->qdisc_list); qdisc_unlock_tree(dev); return sch; @@ -740,7 +742,8 @@ static int tc_get_qdisc(struct sk_buff *skb, struct nlmsghdr *n, void *arg) q = dev->qdisc_ingress; } } else { - q = dev->qdisc_sleeping; + struct netdev_queue *dev_queue = &dev->tx_queue; + q = dev_queue->qdisc_sleeping; } if (!q) return -ENOENT; @@ -814,7 +817,8 @@ replay: q = dev->qdisc_ingress; } } else { - q = dev->qdisc_sleeping; + struct netdev_queue *dev_queue = &dev->tx_queue; + q = dev_queue->qdisc_sleeping; } /* It may be default qdisc, ignore it */ @@ -1015,12 +1019,14 @@ static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb) read_lock(&dev_base_lock); idx = 0; for_each_netdev(&init_net, dev) { + struct netdev_queue *dev_queue; if (idx < s_idx) goto cont; if (idx > s_idx) s_q_idx = 0; q_idx = 0; - list_for_each_entry(q, &dev->qdisc_list, list) { + dev_queue = &dev->tx_queue; + list_for_each_entry(q, &dev_queue->qdisc_list, list) { if (q_idx < s_q_idx) { q_idx++; continue; @@ -1054,6 +1060,7 @@ done: static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n, void *arg) { struct net *net = sock_net(skb->sk); + struct netdev_queue *dev_queue; struct tcmsg *tcm = NLMSG_DATA(n); struct nlattr *tca[TCA_MAX + 1]; struct net_device *dev; @@ -1091,6 +1098,7 @@ static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n, void *arg) /* Step 1. Determine qdisc handle X:0 */ + dev_queue = &dev->tx_queue; if (pid != TC_H_ROOT) { u32 qid1 = TC_H_MAJ(pid); @@ -1101,7 +1109,7 @@ static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n, void *arg) } else if (qid1) { qid = qid1; } else if (qid == 0) - qid = dev->qdisc_sleeping->handle; + qid = dev_queue->qdisc_sleeping->handle; /* Now qid is genuine qdisc handle consistent both with parent and child. @@ -1112,7 +1120,7 @@ static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n, void *arg) pid = TC_H_MAKE(qid, pid); } else { if (qid == 0) - qid = dev->qdisc_sleeping->handle; + qid = dev_queue->qdisc_sleeping->handle; } /* OK. Locate qdisc */ @@ -1248,6 +1256,7 @@ static int qdisc_class_dump(struct Qdisc *q, unsigned long cl, struct qdisc_walk static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) { struct net *net = sock_net(skb->sk); + struct netdev_queue *dev_queue; int t; int s_t; struct net_device *dev; @@ -1266,7 +1275,8 @@ static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) s_t = cb->args[0]; t = 0; - list_for_each_entry(q, &dev->qdisc_list, list) { + dev_queue = &dev->tx_queue; + list_for_each_entry(q, &dev_queue->qdisc_list, list) { if (t < s_t || !q->ops->cl_ops || (tcm->tcm_parent && TC_H_MAJ(tcm->tcm_parent) != q->handle)) { diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 804d44b00348..3223e5ba76aa 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -122,7 +122,7 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * * __LINK_STATE_QDISC_RUNNING guarantees only one CPU can process this * device at a time. queue->lock serializes queue accesses for - * this device AND dev->qdisc pointer itself. + * this device AND txq->qdisc pointer itself. * * netif_tx_lock serializes accesses to device driver. * @@ -138,7 +138,8 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, */ static inline int qdisc_restart(struct net_device *dev) { - struct Qdisc *q = dev->qdisc; + struct netdev_queue *txq = &dev->tx_queue; + struct Qdisc *q = txq->qdisc; struct sk_buff *skb; int ret = NETDEV_TX_BUSY; @@ -148,15 +149,15 @@ static inline int qdisc_restart(struct net_device *dev) /* And release queue */ - spin_unlock(&q->dev_queue->lock); + spin_unlock(&txq->lock); HARD_TX_LOCK(dev, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) ret = dev_hard_start_xmit(skb, dev); HARD_TX_UNLOCK(dev); - spin_lock(&q->dev_queue->lock); - q = dev->qdisc; + spin_lock(&txq->lock); + q = txq->qdisc; switch (ret) { case NETDEV_TX_OK: @@ -207,9 +208,10 @@ void __qdisc_run(struct net_device *dev) static void dev_watchdog(unsigned long arg) { struct net_device *dev = (struct net_device *)arg; + struct netdev_queue *txq = &dev->tx_queue; netif_tx_lock(dev); - if (dev->qdisc != &noop_qdisc) { + if (txq->qdisc != &noop_qdisc) { if (netif_device_present(dev) && netif_running(dev) && netif_carrier_ok(dev)) { @@ -539,53 +541,63 @@ EXPORT_SYMBOL(qdisc_destroy); void dev_activate(struct net_device *dev) { + struct netdev_queue *txq = &dev->tx_queue; + /* No queueing discipline is attached to device; create default one i.e. pfifo_fast for devices, which need queueing and noqueue_qdisc for virtual interfaces */ - if (dev->qdisc_sleeping == &noop_qdisc) { + if (txq->qdisc_sleeping == &noop_qdisc) { struct Qdisc *qdisc; if (dev->tx_queue_len) { - qdisc = qdisc_create_dflt(dev, &dev->tx_queue, + qdisc = qdisc_create_dflt(dev, txq, &pfifo_fast_ops, TC_H_ROOT); if (qdisc == NULL) { printk(KERN_INFO "%s: activation failed\n", dev->name); return; } - list_add_tail(&qdisc->list, &dev->qdisc_list); + list_add_tail(&qdisc->list, &txq->qdisc_list); } else { qdisc = &noqueue_qdisc; } - dev->qdisc_sleeping = qdisc; + txq->qdisc_sleeping = qdisc; } if (!netif_carrier_ok(dev)) /* Delay activation until next carrier-on event */ return; - spin_lock_bh(&dev->tx_queue.lock); - rcu_assign_pointer(dev->qdisc, dev->qdisc_sleeping); - if (dev->qdisc != &noqueue_qdisc) { + spin_lock_bh(&txq->lock); + rcu_assign_pointer(txq->qdisc, txq->qdisc_sleeping); + if (txq->qdisc != &noqueue_qdisc) { dev->trans_start = jiffies; dev_watchdog_up(dev); } - spin_unlock_bh(&dev->tx_queue.lock); + spin_unlock_bh(&txq->lock); +} + +static void dev_deactivate_queue(struct net_device *dev, + struct netdev_queue *dev_queue, + struct Qdisc *qdisc_default) +{ + struct Qdisc *qdisc = dev_queue->qdisc; + + if (qdisc) { + dev_queue->qdisc = qdisc_default; + qdisc_reset(qdisc); + } } void dev_deactivate(struct net_device *dev) { - struct Qdisc *qdisc; struct sk_buff *skb; int running; spin_lock_bh(&dev->tx_queue.lock); - qdisc = dev->qdisc; - dev->qdisc = &noop_qdisc; - - qdisc_reset(qdisc); + dev_deactivate_queue(dev, &dev->tx_queue, &noop_qdisc); skb = dev->gso_skb; dev->gso_skb = NULL; @@ -622,32 +634,44 @@ void dev_deactivate(struct net_device *dev) } while (WARN_ON_ONCE(running)); } +static void dev_init_scheduler_queue(struct net_device *dev, + struct netdev_queue *dev_queue, + struct Qdisc *qdisc) +{ + dev_queue->qdisc = qdisc; + dev_queue->qdisc_sleeping = qdisc; + INIT_LIST_HEAD(&dev_queue->qdisc_list); +} + void dev_init_scheduler(struct net_device *dev) { qdisc_lock_tree(dev); - dev->qdisc = &noop_qdisc; - dev->qdisc_sleeping = &noop_qdisc; - INIT_LIST_HEAD(&dev->qdisc_list); + dev_init_scheduler_queue(dev, &dev->tx_queue, &noop_qdisc); + dev_init_scheduler_queue(dev, &dev->rx_queue, NULL); qdisc_unlock_tree(dev); setup_timer(&dev->watchdog_timer, dev_watchdog, (unsigned long)dev); } -void dev_shutdown(struct net_device *dev) +static void dev_shutdown_scheduler_queue(struct net_device *dev, + struct netdev_queue *dev_queue, + struct Qdisc *qdisc_default) { - struct Qdisc *qdisc; + struct Qdisc *qdisc = dev_queue->qdisc_sleeping; + + if (qdisc) { + dev_queue->qdisc = qdisc_default; + dev_queue->qdisc_sleeping = qdisc_default; - qdisc_lock_tree(dev); - qdisc = dev->qdisc_sleeping; - dev->qdisc = &noop_qdisc; - dev->qdisc_sleeping = &noop_qdisc; - qdisc_destroy(qdisc); -#if defined(CONFIG_NET_SCH_INGRESS) || defined(CONFIG_NET_SCH_INGRESS_MODULE) - if ((qdisc = dev->qdisc_ingress) != NULL) { - dev->qdisc_ingress = NULL; qdisc_destroy(qdisc); } -#endif +} + +void dev_shutdown(struct net_device *dev) +{ + qdisc_lock_tree(dev); + dev_shutdown_scheduler_queue(dev, &dev->tx_queue, &noop_qdisc); + dev_shutdown_scheduler_queue(dev, &dev->rx_queue, NULL); BUG_TRAP(!timer_pending(&dev->watchdog_timer)); qdisc_unlock_tree(dev); } diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index 71b73c528f9b..4093f1eaaf60 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -180,7 +180,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) * skb will be queued. */ if (count > 1 && (skb2 = skb_clone(skb, GFP_ATOMIC)) != NULL) { - struct Qdisc *rootq = qdisc_dev(sch)->qdisc; + struct Qdisc *rootq = qdisc_dev(sch)->tx_queue.qdisc; u32 dupsave = q->duplicate; /* prevent duplicating a dup... */ q->duplicate = 0; diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index 4f3054e8e1ab..8ac05981be20 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -107,17 +107,19 @@ static struct sk_buff * teql_dequeue(struct Qdisc* sch) { struct teql_sched_data *dat = qdisc_priv(sch); + struct netdev_queue *dat_queue; struct sk_buff *skb; skb = __skb_dequeue(&dat->q); + dat_queue = &dat->m->dev->tx_queue; if (skb == NULL) { - struct net_device *m = qdisc_dev(dat->m->dev->qdisc); + struct net_device *m = qdisc_dev(dat_queue->qdisc); if (m) { dat->m->slaves = sch; netif_wake_queue(m); } } - sch->q.qlen = dat->q.qlen + dat->m->dev->qdisc->q.qlen; + sch->q.qlen = dat->q.qlen + dat_queue->qdisc->q.qlen; return skb; } @@ -155,7 +157,7 @@ teql_destroy(struct Qdisc* sch) if (q == master->slaves) { master->slaves = NULL; spin_lock_bh(&master->dev->tx_queue.lock); - qdisc_reset(master->dev->qdisc); + qdisc_reset(master->dev->tx_queue.qdisc); spin_unlock_bh(&master->dev->tx_queue.lock); } } @@ -216,7 +218,7 @@ static int teql_qdisc_init(struct Qdisc *sch, struct nlattr *opt) static int __teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device *dev) { - struct teql_sched_data *q = qdisc_priv(dev->qdisc); + struct teql_sched_data *q = qdisc_priv(dev->tx_queue.qdisc); struct neighbour *mn = skb->dst->neighbour; struct neighbour *n = q->ncache; @@ -252,7 +254,7 @@ __teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device * static inline int teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device *dev) { - if (dev->qdisc == &noop_qdisc) + if (dev->tx_queue.qdisc == &noop_qdisc) return -ENODEV; if (dev->header_ops == NULL || @@ -284,7 +286,7 @@ restart: do { struct net_device *slave = qdisc_dev(q); - if (slave->qdisc_sleeping != q) + if (slave->tx_queue.qdisc_sleeping != q) continue; if (netif_queue_stopped(slave) || __netif_subqueue_stopped(slave, subq) || -- cgit v1.2.2 From 816f3258e70db38d6d92c8d871377179fd69160f Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 22:49:00 -0700 Subject: netdev: Kill qdisc_ingress, use netdev->rx_queue.qdisc instead. Now that our qdisc management is bi-directional, per-queue, and fully orthogonal, there is no reason to have a special ingress qdisc pointer in struct net_device. Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++-- net/sched/sch_api.c | 11 ++++++----- 2 files changed, 8 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index ce79c28d739d..ab760a954d99 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2033,7 +2033,7 @@ static int ing_filter(struct sk_buff *skb) rxq = &dev->rx_queue; spin_lock(&rxq->lock); - if ((q = dev->qdisc_ingress) != NULL) + if ((q = rxq->qdisc) != NULL) result = q->enqueue(skb, q); spin_unlock(&rxq->lock); @@ -2044,7 +2044,7 @@ static inline struct sk_buff *handle_ing(struct sk_buff *skb, struct packet_type **pt_prev, int *ret, struct net_device *orig_dev) { - if (!skb->dev->qdisc_ingress) + if (!skb->dev->rx_queue.qdisc) goto out; if (*pt_prev) { diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 2313fa7c97be..4003c280b69f 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -450,14 +450,15 @@ dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) qdisc_lock_tree(dev); if (qdisc && qdisc->flags&TCQ_F_INGRESS) { - oqdisc = dev->qdisc_ingress; + dev_queue = &dev->rx_queue; + oqdisc = dev_queue->qdisc; /* Prune old scheduler */ if (oqdisc && atomic_read(&oqdisc->refcnt) <= 1) { /* delete */ qdisc_reset(oqdisc); - dev->qdisc_ingress = NULL; + dev_queue->qdisc = NULL; } else { /* new */ - dev->qdisc_ingress = qdisc; + dev_queue->qdisc = qdisc; } } else { @@ -739,7 +740,7 @@ static int tc_get_qdisc(struct sk_buff *skb, struct nlmsghdr *n, void *arg) return -ENOENT; q = qdisc_leaf(p, clid); } else { /* ingress */ - q = dev->qdisc_ingress; + q = dev->rx_queue.qdisc; } } else { struct netdev_queue *dev_queue = &dev->tx_queue; @@ -814,7 +815,7 @@ replay: return -ENOENT; q = qdisc_leaf(p, clid); } else { /*ingress */ - q = dev->qdisc_ingress; + q = dev->rx_queue.qdisc; } } else { struct netdev_queue *dev_queue = &dev->tx_queue; -- cgit v1.2.2 From 68dfb42798e1eb2d42acbf872925cc75f1487d9b Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 22:57:31 -0700 Subject: pkt_sched: Kill stats_lock member of struct Qdisc. It is always equal to qdisc->dev_queue->lock Signed-off-by: David S. Miller --- net/sched/sch_api.c | 9 ++++----- net/sched/sch_generic.c | 1 - 2 files changed, 4 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 4003c280b69f..e73bd68aa7ae 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -604,7 +604,6 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, sch->parent = parent; - sch->stats_lock = &dev_queue->lock; if (handle == TC_H_INGRESS) { sch->flags |= TCQ_F_INGRESS; handle = TC_H_MAKE(TC_H_INGRESS, 0); @@ -622,7 +621,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, if (!ops->init || (err = ops->init(sch, tca[TCA_OPTIONS])) == 0) { if (tca[TCA_RATE]) { err = gen_new_estimator(&sch->bstats, &sch->rate_est, - sch->stats_lock, + &sch->dev_queue->lock, tca[TCA_RATE]); if (err) { /* @@ -664,7 +663,7 @@ static int qdisc_change(struct Qdisc *sch, struct nlattr **tca) } if (tca[TCA_RATE]) gen_replace_estimator(&sch->bstats, &sch->rate_est, - sch->stats_lock, tca[TCA_RATE]); + &sch->dev_queue->lock, tca[TCA_RATE]); return 0; } @@ -954,7 +953,7 @@ static int tc_fill_qdisc(struct sk_buff *skb, struct Qdisc *q, u32 clid, q->qstats.qlen = q->q.qlen; if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS, - TCA_XSTATS, q->stats_lock, &d) < 0) + TCA_XSTATS, &q->dev_queue->lock, &d) < 0) goto nla_put_failure; if (q->ops->dump_stats && q->ops->dump_stats(q, &d) < 0) @@ -1203,7 +1202,7 @@ static int tc_fill_tclass(struct sk_buff *skb, struct Qdisc *q, goto nla_put_failure; if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS, - TCA_XSTATS, q->stats_lock, &d) < 0) + TCA_XSTATS, &q->dev_queue->lock, &d) < 0) goto nla_put_failure; if (cl_ops->dump_stats && cl_ops->dump_stats(q, cl, &d) < 0) diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 3223e5ba76aa..dda78ee314ec 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -484,7 +484,6 @@ struct Qdisc * qdisc_create_dflt(struct net_device *dev, sch = qdisc_alloc(dev_queue, ops); if (IS_ERR(sch)) goto errout; - sch->stats_lock = &dev_queue->lock; sch->parent = parentid; if (!ops->init || ops->init(sch, NULL) == 0) -- cgit v1.2.2 From 74d58a0c1d5b348a8d4ea9643b573a6ab455a3f3 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 22:57:51 -0700 Subject: pkt_sched: Make netem queue agnostic. It just wants the root qdisc given an arbitrary qdisc, and that is simply qdisc->dev_queue->qdisc Signed-off-by: David S. Miller Acked-by: Stephen Hemminger --- net/sched/sch_netem.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index 4093f1eaaf60..bc585f2089ff 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -180,7 +180,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) * skb will be queued. */ if (count > 1 && (skb2 = skb_clone(skb, GFP_ATOMIC)) != NULL) { - struct Qdisc *rootq = qdisc_dev(sch)->tx_queue.qdisc; + struct Qdisc *rootq = sch->dev_queue->qdisc; u32 dupsave = q->duplicate; /* prevent duplicating a dup... */ q->duplicate = 0; -- cgit v1.2.2 From ee609cb36220d18c0cf476b066a5ab7e6f6d3a69 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 22:58:37 -0700 Subject: netdev: Move next_sched into struct netdev_queue. We schedule queues, not the device, for output queue processing in BH. Signed-off-by: David S. Miller --- net/core/dev.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index ab760a954d99..d6b8d3c3e6ec 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1323,13 +1323,14 @@ static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev) void __netif_schedule(struct net_device *dev) { if (!test_and_set_bit(__LINK_STATE_SCHED, &dev->state)) { + struct netdev_queue *txq = &dev->tx_queue; unsigned long flags; struct softnet_data *sd; local_irq_save(flags); sd = &__get_cpu_var(softnet_data); - dev->next_sched = sd->output_queue; - sd->output_queue = dev; + txq->next_sched = sd->output_queue; + sd->output_queue = txq; raise_softirq_irqoff(NET_TX_SOFTIRQ); local_irq_restore(flags); } @@ -1912,7 +1913,7 @@ static void net_tx_action(struct softirq_action *h) } if (sd->output_queue) { - struct net_device *head; + struct netdev_queue *head; local_irq_disable(); head = sd->output_queue; @@ -1920,12 +1921,10 @@ static void net_tx_action(struct softirq_action *h) local_irq_enable(); while (head) { - struct net_device *dev = head; - struct netdev_queue *txq; + struct netdev_queue *txq = head; + struct net_device *dev = txq->dev; head = head->next_sched; - txq = &dev->tx_queue; - smp_mb__before_clear_bit(); clear_bit(__LINK_STATE_SCHED, &dev->state); @@ -4346,7 +4345,7 @@ static int dev_cpu_callback(struct notifier_block *nfb, void *ocpu) { struct sk_buff **list_skb; - struct net_device **list_net; + struct netdev_queue **list_net; struct sk_buff *skb; unsigned int cpu, oldcpu = (unsigned long)ocpu; struct softnet_data *sd, *oldsd; -- cgit v1.2.2 From 6fa9864b53f0680e432a2c431c2cf2055daa3a88 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:01:06 -0700 Subject: net: Clean up explicit ->tx_queue references in link watch. First, we add a qdisc_tx_changing() helper which returns true if the qdisc attachment is in transition. Second, we remove an assertion warning which is of limited value and is hard to express precisely in a multiqueue environment. Signed-off-by: David S. Miller --- net/core/link_watch.c | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/core/link_watch.c b/net/core/link_watch.c index 50218218445b..bf8f7af699d7 100644 --- a/net/core/link_watch.c +++ b/net/core/link_watch.c @@ -77,12 +77,10 @@ static void rfc2863_policy(struct net_device *dev) } -static int linkwatch_urgent_event(struct net_device *dev) +static bool linkwatch_urgent_event(struct net_device *dev) { - struct netdev_queue *txq = &dev->tx_queue; - return netif_running(dev) && netif_carrier_ok(dev) && - txq->qdisc != txq->qdisc_sleeping; + qdisc_tx_changing(dev); } @@ -182,12 +180,9 @@ static void __linkwatch_run_queue(int urgent_only) rfc2863_policy(dev); if (dev->flags & IFF_UP) { - if (netif_carrier_ok(dev)) { - struct netdev_queue *txq = &dev->tx_queue; - - WARN_ON(txq->qdisc_sleeping == &noop_qdisc); + if (netif_carrier_ok(dev)) dev_activate(dev); - } else + else dev_deactivate(dev); netdev_state_change(dev); @@ -218,7 +213,7 @@ static void linkwatch_event(struct work_struct *dummy) void linkwatch_fire_event(struct net_device *dev) { - int urgent = linkwatch_urgent_event(dev); + bool urgent = linkwatch_urgent_event(dev); if (!test_and_set_bit(__LINK_STATE_LINKWATCH_PENDING, &dev->state)) { dev_hold(dev); -- cgit v1.2.2 From 052979499c767268b912d25031ae524c451679d0 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:01:27 -0700 Subject: pkt_sched: Add qdisc_tx_is_noop() helper and use in IPV6. This indicates if the NOOP scheduler is what is active for TX on a given device. Signed-off-by: David S. Miller --- net/ipv6/addrconf.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 5c84c798331d..30184e0dd74c 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -229,10 +229,9 @@ const struct in6_addr in6addr_linklocal_allnodes = IN6ADDR_LINKLOCAL_ALLNODES_IN const struct in6_addr in6addr_linklocal_allrouters = IN6ADDR_LINKLOCAL_ALLROUTERS_INIT; /* Check if a valid qdisc is available */ -static inline int addrconf_qdisc_ok(struct net_device *dev) +static inline bool addrconf_qdisc_ok(const struct net_device *dev) { - struct netdev_queue *txq = &dev->tx_queue; - return (txq->qdisc != &noop_qdisc); + return !qdisc_tx_is_noop(dev); } /* Check if a route is valid prefix route */ -- cgit v1.2.2 From c2aa288548a29d909ec875e81137fb0dbbb420b7 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:01:52 -0700 Subject: mac80211: Decrease number of explicit ->tx_queue references. Accomplish this by using local variables. Signed-off-by: David S. Miller --- net/mac80211/main.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 12aeaf78ae75..2968baa66b91 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -570,6 +570,7 @@ static int ieee80211_stop(struct net_device *dev) int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) { struct ieee80211_local *local = hw_to_local(hw); + struct netdev_queue *txq; struct sta_info *sta; struct ieee80211_sub_if_data *sdata; u16 start_seq_num = 0; @@ -636,7 +637,8 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) /* ensure that TX flow won't interrupt us * until the end of the call to requeue function */ - spin_lock_bh(&local->mdev->tx_queue.lock); + txq = &local->mdev->tx_queue; + spin_lock_bh(&txq->lock); /* create a new queue for this aggregation */ ret = ieee80211_ht_agg_queue_add(local, sta, tid); @@ -675,7 +677,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) /* Will put all the packets in the new SW queue */ ieee80211_requeue(local, ieee802_1d_to_ac[tid]); - spin_unlock_bh(&local->mdev->tx_queue.lock); + spin_unlock_bh(&txq->lock); spin_unlock_bh(&sta->lock); /* send an addBA request */ @@ -701,7 +703,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) err_unlock_queue: kfree(sta->ampdu_mlme.tid_tx[tid]); sta->ampdu_mlme.tid_tx[tid] = NULL; - spin_unlock_bh(&local->mdev->tx_queue.lock); + spin_unlock_bh(&txq->lock); ret = -EBUSY; err_unlock_sta: spin_unlock_bh(&sta->lock); @@ -826,6 +828,7 @@ EXPORT_SYMBOL(ieee80211_start_tx_ba_cb); void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) { struct ieee80211_local *local = hw_to_local(hw); + struct netdev_queue *txq; struct sta_info *sta; u8 *state; int agg_queue; @@ -875,10 +878,11 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) /* avoid ordering issues: we are the only one that can modify * the content of the qdiscs */ - spin_lock_bh(&local->mdev->tx_queue.lock); + txq = &local->mdev->tx_queue; + spin_lock_bh(&txq->lock); /* remove the queue for this aggregation */ ieee80211_ht_agg_queue_remove(local, sta, tid, 1); - spin_unlock_bh(&local->mdev->tx_queue.lock); + spin_unlock_bh(&txq->lock); /* we just requeued the all the frames that were in the removed * queue, and since we might miss a softirq we do netif_schedule. -- cgit v1.2.2 From 970565bbad0c7b98db0d14131a69e5a0f4445d49 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:10:33 -0700 Subject: netdev: Move gso_skb into netdev_queue. Signed-off-by: David S. Miller --- net/sched/sch_generic.c | 42 +++++++++++++++++++++++------------------- 1 file changed, 23 insertions(+), 19 deletions(-) (limited to 'net') diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index dda78ee314ec..8247a406a401 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -63,10 +63,11 @@ static inline int qdisc_qlen(struct Qdisc *q) } static inline int dev_requeue_skb(struct sk_buff *skb, struct net_device *dev, + struct netdev_queue *dev_queue, struct Qdisc *q) { if (unlikely(skb->next)) - dev->gso_skb = skb; + dev_queue->gso_skb = skb; else q->ops->requeue(skb, q); @@ -75,12 +76,13 @@ static inline int dev_requeue_skb(struct sk_buff *skb, struct net_device *dev, } static inline struct sk_buff *dev_dequeue_skb(struct net_device *dev, + struct netdev_queue *dev_queue, struct Qdisc *q) { struct sk_buff *skb; - if ((skb = dev->gso_skb)) - dev->gso_skb = NULL; + if ((skb = dev_queue->gso_skb)) + dev_queue->gso_skb = NULL; else skb = q->dequeue(q); @@ -89,6 +91,7 @@ static inline struct sk_buff *dev_dequeue_skb(struct net_device *dev, static inline int handle_dev_cpu_collision(struct sk_buff *skb, struct net_device *dev, + struct netdev_queue *dev_queue, struct Qdisc *q) { int ret; @@ -111,7 +114,7 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * some time. */ __get_cpu_var(netdev_rx_stat).cpu_collision++; - ret = dev_requeue_skb(skb, dev, q); + ret = dev_requeue_skb(skb, dev, dev_queue, q); } return ret; @@ -144,7 +147,7 @@ static inline int qdisc_restart(struct net_device *dev) int ret = NETDEV_TX_BUSY; /* Dequeue packet */ - if (unlikely((skb = dev_dequeue_skb(dev, q)) == NULL)) + if (unlikely((skb = dev_dequeue_skb(dev, txq, q)) == NULL)) return 0; @@ -167,7 +170,7 @@ static inline int qdisc_restart(struct net_device *dev) case NETDEV_TX_LOCKED: /* Driver try lock failed */ - ret = handle_dev_cpu_collision(skb, dev, q); + ret = handle_dev_cpu_collision(skb, dev, txq, q); break; default: @@ -176,7 +179,7 @@ static inline int qdisc_restart(struct net_device *dev) printk(KERN_WARNING "BUG %s code %d qlen %d\n", dev->name, ret, q->q.qlen); - ret = dev_requeue_skb(skb, dev, q); + ret = dev_requeue_skb(skb, dev, txq, q); break; } @@ -578,31 +581,32 @@ void dev_activate(struct net_device *dev) spin_unlock_bh(&txq->lock); } -static void dev_deactivate_queue(struct net_device *dev, - struct netdev_queue *dev_queue, +static void dev_deactivate_queue(struct netdev_queue *dev_queue, struct Qdisc *qdisc_default) { - struct Qdisc *qdisc = dev_queue->qdisc; + struct Qdisc *qdisc; + struct sk_buff *skb; + + spin_lock_bh(&dev_queue->lock); + qdisc = dev_queue->qdisc; if (qdisc) { dev_queue->qdisc = qdisc_default; qdisc_reset(qdisc); } + skb = dev_queue->gso_skb; + dev_queue->gso_skb = NULL; + + spin_unlock_bh(&dev_queue->lock); + + kfree_skb(skb); } void dev_deactivate(struct net_device *dev) { - struct sk_buff *skb; int running; - spin_lock_bh(&dev->tx_queue.lock); - dev_deactivate_queue(dev, &dev->tx_queue, &noop_qdisc); - - skb = dev->gso_skb; - dev->gso_skb = NULL; - spin_unlock_bh(&dev->tx_queue.lock); - - kfree_skb(skb); + dev_deactivate_queue(&dev->tx_queue, &noop_qdisc); dev_watchdog_down(dev); -- cgit v1.2.2 From 86d804e10a37cd86f16bf72386c37e843a98a74b Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:11:25 -0700 Subject: netdev: Make netif_schedule() routines work with netdev_queue objects. Only plain netif_schedule() remains taking a net_device, mostly as a compatability item while we transition the rest of these interfaces. Everything else calls netif_schedule_queue() or __netif_schedule(), both of which take a netdev_queue pointer. Signed-off-by: David S. Miller --- net/core/dev.c | 9 +++++---- net/mac80211/main.c | 4 ++-- net/sched/sch_api.c | 4 ++-- net/sched/sch_cbq.c | 2 +- net/sched/sch_generic.c | 10 +++++----- 5 files changed, 15 insertions(+), 14 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index d6b8d3c3e6ec..0dc888ad4217 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1320,12 +1320,13 @@ static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev) } -void __netif_schedule(struct net_device *dev) +void __netif_schedule(struct netdev_queue *txq) { + struct net_device *dev = txq->dev; + if (!test_and_set_bit(__LINK_STATE_SCHED, &dev->state)) { - struct netdev_queue *txq = &dev->tx_queue; - unsigned long flags; struct softnet_data *sd; + unsigned long flags; local_irq_save(flags); sd = &__get_cpu_var(softnet_data); @@ -1932,7 +1933,7 @@ static void net_tx_action(struct softirq_action *h) qdisc_run(dev); spin_unlock(&txq->lock); } else { - netif_schedule(dev); + netif_schedule_queue(txq); } } } diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 2968baa66b91..1c4d3ba6b878 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -885,10 +885,10 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) spin_unlock_bh(&txq->lock); /* we just requeued the all the frames that were in the removed - * queue, and since we might miss a softirq we do netif_schedule. + * queue, and since we might miss a softirq we do netif_schedule_queue. * ieee80211_wake_queue is not used here as this queue is not * necessarily stopped */ - netif_schedule(local->mdev); + netif_schedule_queue(txq); spin_lock_bh(&sta->lock); *state = HT_AGG_STATE_IDLE; sta->ampdu_mlme.addba_req_num[tid] = 0; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index e73bd68aa7ae..95873f8dd37c 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -282,11 +282,11 @@ static enum hrtimer_restart qdisc_watchdog(struct hrtimer *timer) { struct qdisc_watchdog *wd = container_of(timer, struct qdisc_watchdog, timer); - struct net_device *dev = qdisc_dev(wd->qdisc); + struct netdev_queue *txq = wd->qdisc->dev_queue; wd->qdisc->flags &= ~TCQ_F_THROTTLED; smp_wmb(); - netif_schedule(dev); + netif_schedule_queue(txq); return HRTIMER_NORESTART; } diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 99ce3da2b0a4..4efc836cbf38 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -650,7 +650,7 @@ static enum hrtimer_restart cbq_undelay(struct hrtimer *timer) } sch->flags &= ~TCQ_F_THROTTLED; - netif_schedule(qdisc_dev(sch)); + netif_schedule_queue(sch->dev_queue); return HRTIMER_NORESTART; } diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 8247a406a401..407dfdb142a4 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -62,7 +62,7 @@ static inline int qdisc_qlen(struct Qdisc *q) return q->q.qlen; } -static inline int dev_requeue_skb(struct sk_buff *skb, struct net_device *dev, +static inline int dev_requeue_skb(struct sk_buff *skb, struct netdev_queue *dev_queue, struct Qdisc *q) { @@ -71,7 +71,7 @@ static inline int dev_requeue_skb(struct sk_buff *skb, struct net_device *dev, else q->ops->requeue(skb, q); - netif_schedule(dev); + netif_schedule_queue(dev_queue); return 0; } @@ -114,7 +114,7 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * some time. */ __get_cpu_var(netdev_rx_stat).cpu_collision++; - ret = dev_requeue_skb(skb, dev, dev_queue, q); + ret = dev_requeue_skb(skb, dev_queue, q); } return ret; @@ -179,7 +179,7 @@ static inline int qdisc_restart(struct net_device *dev) printk(KERN_WARNING "BUG %s code %d qlen %d\n", dev->name, ret, q->q.qlen); - ret = dev_requeue_skb(skb, dev, txq, q); + ret = dev_requeue_skb(skb, txq, q); break; } @@ -200,7 +200,7 @@ void __qdisc_run(struct net_device *dev) * 2. we've been doing it for too long. */ if (need_resched() || jiffies != start_time) { - netif_schedule(dev); + netif_schedule_queue(&dev->tx_queue); break; } } -- cgit v1.2.2 From eb6aafe3f843cb0e939546c03540a3b4911b6964 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:12:38 -0700 Subject: pkt_sched: Make qdisc_run take a netdev_queue. This allows us to use this calling convention all the way down into qdisc_restart(). Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++-- net/sched/sch_generic.c | 26 ++++++++++++++------------ 2 files changed, 16 insertions(+), 14 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 0dc888ad4217..0218b0b9be80 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1734,7 +1734,7 @@ gso: /* reset queue_mapping to zero */ skb_set_queue_mapping(skb, 0); rc = q->enqueue(skb, q); - qdisc_run(dev); + qdisc_run(txq); spin_unlock(&txq->lock); rc = rc == NET_XMIT_BYPASS ? NET_XMIT_SUCCESS : rc; @@ -1930,7 +1930,7 @@ static void net_tx_action(struct softirq_action *h) clear_bit(__LINK_STATE_SCHED, &dev->state); if (spin_trylock(&txq->lock)) { - qdisc_run(dev); + qdisc_run(txq); spin_unlock(&txq->lock); } else { netif_schedule_queue(txq); diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 407dfdb142a4..fcc7533f0bcc 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -75,9 +75,8 @@ static inline int dev_requeue_skb(struct sk_buff *skb, return 0; } -static inline struct sk_buff *dev_dequeue_skb(struct net_device *dev, - struct netdev_queue *dev_queue, - struct Qdisc *q) +static inline struct sk_buff *dequeue_skb(struct netdev_queue *dev_queue, + struct Qdisc *q) { struct sk_buff *skb; @@ -90,10 +89,10 @@ static inline struct sk_buff *dev_dequeue_skb(struct net_device *dev, } static inline int handle_dev_cpu_collision(struct sk_buff *skb, - struct net_device *dev, struct netdev_queue *dev_queue, struct Qdisc *q) { + struct net_device *dev = dev_queue->dev; int ret; if (unlikely(dev->xmit_lock_owner == smp_processor_id())) { @@ -139,21 +138,23 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * >0 - queue is not empty. * */ -static inline int qdisc_restart(struct net_device *dev) +static inline int qdisc_restart(struct netdev_queue *txq) { - struct netdev_queue *txq = &dev->tx_queue; struct Qdisc *q = txq->qdisc; - struct sk_buff *skb; int ret = NETDEV_TX_BUSY; + struct net_device *dev; + struct sk_buff *skb; /* Dequeue packet */ - if (unlikely((skb = dev_dequeue_skb(dev, txq, q)) == NULL)) + if (unlikely((skb = dequeue_skb(txq, q)) == NULL)) return 0; /* And release queue */ spin_unlock(&txq->lock); + dev = txq->dev; + HARD_TX_LOCK(dev, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) ret = dev_hard_start_xmit(skb, dev); @@ -170,7 +171,7 @@ static inline int qdisc_restart(struct net_device *dev) case NETDEV_TX_LOCKED: /* Driver try lock failed */ - ret = handle_dev_cpu_collision(skb, dev, txq, q); + ret = handle_dev_cpu_collision(skb, txq, q); break; default: @@ -186,11 +187,12 @@ static inline int qdisc_restart(struct net_device *dev) return ret; } -void __qdisc_run(struct net_device *dev) +void __qdisc_run(struct netdev_queue *txq) { + struct net_device *dev = txq->dev; unsigned long start_time = jiffies; - while (qdisc_restart(dev)) { + while (qdisc_restart(txq)) { if (netif_queue_stopped(dev)) break; @@ -200,7 +202,7 @@ void __qdisc_run(struct net_device *dev) * 2. we've been doing it for too long. */ if (need_resched() || jiffies != start_time) { - netif_schedule_queue(&dev->tx_queue); + netif_schedule_queue(txq); break; } } -- cgit v1.2.2 From c773e847ea8f6812804e40f52399c6921a00eab1 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:13:53 -0700 Subject: netdev: Move _xmit_lock and xmit_lock_owner into netdev_queue. Accesses are mostly structured such that when there are multiple TX queues the code transformations will be a little bit simpler. Signed-off-by: David S. Miller --- net/8021q/vlan_dev.c | 15 +++++++++++++-- net/core/dev.c | 28 ++++++++++++++++++++-------- net/netrom/af_netrom.c | 12 +++++++++++- net/rose/af_rose.c | 12 +++++++++++- net/sched/sch_generic.c | 9 ++++----- 5 files changed, 59 insertions(+), 17 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index b6e52c025fd8..8efa399823e3 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -627,6 +627,18 @@ static void vlan_dev_set_rx_mode(struct net_device *vlan_dev) */ static struct lock_class_key vlan_netdev_xmit_lock_key; +static void vlan_dev_set_lockdep_one(struct netdev_queue *txq, + int subclass) +{ + lockdep_set_class_and_subclass(&txq->_xmit_lock, + &vlan_netdev_xmit_lock_key, subclass); +} + +static void vlan_dev_set_lockdep_class(struct net_device *dev, int subclass) +{ + vlan_dev_set_lockdep_one(&dev->tx_queue, subclass); +} + static const struct header_ops vlan_header_ops = { .create = vlan_dev_hard_header, .rebuild = vlan_dev_rebuild_header, @@ -668,8 +680,7 @@ static int vlan_dev_init(struct net_device *dev) if (is_vlan_dev(real_dev)) subclass = 1; - lockdep_set_class_and_subclass(&dev->_xmit_lock, - &vlan_netdev_xmit_lock_key, subclass); + vlan_dev_set_lockdep_class(dev, subclass); return 0; } diff --git a/net/core/dev.c b/net/core/dev.c index 0218b0b9be80..a29a359b15d1 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -258,7 +258,7 @@ DEFINE_PER_CPU(struct softnet_data, softnet_data); #ifdef CONFIG_DEBUG_LOCK_ALLOC /* - * register_netdevice() inits dev->_xmit_lock and sets lockdep class + * register_netdevice() inits txq->_xmit_lock and sets lockdep class * according to dev->type */ static const unsigned short netdev_lock_type[] = @@ -1758,19 +1758,19 @@ gso: if (dev->flags & IFF_UP) { int cpu = smp_processor_id(); /* ok because BHs are off */ - if (dev->xmit_lock_owner != cpu) { + if (txq->xmit_lock_owner != cpu) { - HARD_TX_LOCK(dev, cpu); + HARD_TX_LOCK(dev, txq, cpu); if (!netif_queue_stopped(dev) && !netif_subqueue_stopped(dev, skb)) { rc = 0; if (!dev_hard_start_xmit(skb, dev)) { - HARD_TX_UNLOCK(dev); + HARD_TX_UNLOCK(dev, txq); goto out; } } - HARD_TX_UNLOCK(dev); + HARD_TX_UNLOCK(dev, txq); if (net_ratelimit()) printk(KERN_CRIT "Virtual device %s asks to " "queue packet!\n", dev->name); @@ -3761,6 +3761,20 @@ static void rollback_registered(struct net_device *dev) dev_put(dev); } +static void __netdev_init_queue_locks_one(struct netdev_queue *dev_queue, + struct net_device *dev) +{ + spin_lock_init(&dev_queue->_xmit_lock); + netdev_set_lockdep_class(&dev_queue->_xmit_lock, dev->type); + dev_queue->xmit_lock_owner = -1; +} + +static void netdev_init_queue_locks(struct net_device *dev) +{ + __netdev_init_queue_locks_one(&dev->tx_queue, dev); + __netdev_init_queue_locks_one(&dev->rx_queue, dev); +} + /** * register_netdevice - register a network device * @dev: device to register @@ -3795,9 +3809,7 @@ int register_netdevice(struct net_device *dev) BUG_ON(!dev_net(dev)); net = dev_net(dev); - spin_lock_init(&dev->_xmit_lock); - netdev_set_lockdep_class(&dev->_xmit_lock, dev->type); - dev->xmit_lock_owner = -1; + netdev_init_queue_locks(dev); dev->iflink = -1; diff --git a/net/netrom/af_netrom.c b/net/netrom/af_netrom.c index 74884f4a6255..819afc449e1e 100644 --- a/net/netrom/af_netrom.c +++ b/net/netrom/af_netrom.c @@ -74,6 +74,16 @@ static const struct proto_ops nr_proto_ops; */ static struct lock_class_key nr_netdev_xmit_lock_key; +static void nr_set_lockdep_one(struct netdev_queue *txq) +{ + lockdep_set_class(&txq->_xmit_lock, &nr_netdev_xmit_lock_key); +} + +static void nr_set_lockdep_key(struct net_device *dev) +{ + nr_set_lockdep_one(&dev->tx_queue); +} + /* * Socket removal during an interrupt is now safe. */ @@ -1430,7 +1440,7 @@ static int __init nr_proto_init(void) free_netdev(dev); goto fail; } - lockdep_set_class(&dev->_xmit_lock, &nr_netdev_xmit_lock_key); + nr_set_lockdep_key(dev); dev_nr[i] = dev; } diff --git a/net/rose/af_rose.c b/net/rose/af_rose.c index 46461a69cd0f..7dbbc0891623 100644 --- a/net/rose/af_rose.c +++ b/net/rose/af_rose.c @@ -75,6 +75,16 @@ ax25_address rose_callsign; */ static struct lock_class_key rose_netdev_xmit_lock_key; +static void rose_set_lockdep_one(struct netdev_queue *txq) +{ + lockdep_set_class(&txq->_xmit_lock, &rose_netdev_xmit_lock_key); +} + +static void rose_set_lockdep_key(struct net_device *dev) +{ + rose_set_lockdep_one(&dev->tx_queue); +} + /* * Convert a ROSE address into text. */ @@ -1576,7 +1586,7 @@ static int __init rose_proto_init(void) free_netdev(dev); goto fail; } - lockdep_set_class(&dev->_xmit_lock, &rose_netdev_xmit_lock_key); + rose_set_lockdep_key(dev); dev_rose[i] = dev; } diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index fcc7533f0bcc..b6a36d394663 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -92,10 +92,9 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, struct netdev_queue *dev_queue, struct Qdisc *q) { - struct net_device *dev = dev_queue->dev; int ret; - if (unlikely(dev->xmit_lock_owner == smp_processor_id())) { + if (unlikely(dev_queue->xmit_lock_owner == smp_processor_id())) { /* * Same CPU holding the lock. It may be a transient * configuration error, when hard_start_xmit() recurses. We @@ -105,7 +104,7 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, kfree_skb(skb); if (net_ratelimit()) printk(KERN_WARNING "Dead loop on netdevice %s, " - "fix it urgently!\n", dev->name); + "fix it urgently!\n", dev_queue->dev->name); ret = qdisc_qlen(q); } else { /* @@ -155,10 +154,10 @@ static inline int qdisc_restart(struct netdev_queue *txq) dev = txq->dev; - HARD_TX_LOCK(dev, smp_processor_id()); + HARD_TX_LOCK(dev, txq, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) ret = dev_hard_start_xmit(skb, dev); - HARD_TX_UNLOCK(dev); + HARD_TX_UNLOCK(dev, txq); spin_lock(&txq->lock); q = txq->qdisc; -- cgit v1.2.2 From b19fa1fa91845234961c64dbd564671aa7c0fd27 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:14:24 -0700 Subject: net: Delete NETDEVICES_MULTIQUEUE kconfig option. Multiple TX queue support is a core networking feature. Signed-off-by: David S. Miller --- net/mac80211/Kconfig | 3 --- 1 file changed, 3 deletions(-) (limited to 'net') diff --git a/net/mac80211/Kconfig b/net/mac80211/Kconfig index 40f1add17753..d2038418e2bd 100644 --- a/net/mac80211/Kconfig +++ b/net/mac80211/Kconfig @@ -15,14 +15,11 @@ config MAC80211_QOS def_bool y depends on MAC80211 depends on NET_SCHED - depends on NETDEVICES_MULTIQUEUE comment "QoS/HT support disabled" depends on MAC80211 && !MAC80211_QOS comment "QoS/HT support needs CONFIG_NET_SCHED" depends on MAC80211 && !NET_SCHED -comment "QoS/HT support needs CONFIG_NETDEVICES_MULTIQUEUE" - depends on MAC80211 && !NETDEVICES_MULTIQUEUE menu "Rate control algorithm selection" depends on MAC80211 != n -- cgit v1.2.2 From 79d16385c7f287a33ea771c4dbe60ae43f791b49 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Jul 2008 23:14:46 -0700 Subject: netdev: Move atomic queue state bits into netdev_queue. Signed-off-by: David S. Miller --- net/sched/sch_generic.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index b6a36d394663..243de935b182 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -121,9 +121,9 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, /* * NOTE: Called under queue->lock with locally disabled BH. * - * __LINK_STATE_QDISC_RUNNING guarantees only one CPU can process this - * device at a time. queue->lock serializes queue accesses for - * this device AND txq->qdisc pointer itself. + * __QUEUE_STATE_QDISC_RUNNING guarantees only one CPU can process + * this queue at a time. queue->lock serializes queue accesses for + * this queue AND txq->qdisc pointer itself. * * netif_tx_lock serializes accesses to device driver. * @@ -206,7 +206,7 @@ void __qdisc_run(struct netdev_queue *txq) } } - clear_bit(__LINK_STATE_QDISC_RUNNING, &dev->state); + clear_bit(__QUEUE_STATE_QDISC_RUNNING, &txq->state); } static void dev_watchdog(unsigned long arg) @@ -605,9 +605,10 @@ static void dev_deactivate_queue(struct netdev_queue *dev_queue, void dev_deactivate(struct net_device *dev) { + struct netdev_queue *dev_queue = &dev->tx_queue; int running; - dev_deactivate_queue(&dev->tx_queue, &noop_qdisc); + dev_deactivate_queue(dev_queue, &noop_qdisc); dev_watchdog_down(dev); @@ -616,16 +617,17 @@ void dev_deactivate(struct net_device *dev) /* Wait for outstanding qdisc_run calls. */ do { - while (test_bit(__LINK_STATE_QDISC_RUNNING, &dev->state)) + while (test_bit(__QUEUE_STATE_QDISC_RUNNING, &dev_queue->state)) yield(); /* * Double-check inside queue lock to ensure that all effects * of the queue run are visible when we return. */ - spin_lock_bh(&dev->tx_queue.lock); - running = test_bit(__LINK_STATE_QDISC_RUNNING, &dev->state); - spin_unlock_bh(&dev->tx_queue.lock); + spin_lock_bh(&dev_queue->lock); + running = test_bit(__QUEUE_STATE_QDISC_RUNNING, + &dev_queue->state); + spin_unlock_bh(&dev_queue->lock); /* * The running flag should never be set at this point because -- cgit v1.2.2 From 2013c7e35aeba39777f9b3eef8a70207b3931152 Mon Sep 17 00:00:00 2001 From: Gerrit Renker Date: Sun, 13 Jul 2008 11:51:40 +0100 Subject: dccp ccid-3: Fix error in loss detection The TFRC loss detection code used the wrong loss condition (RFC 4340, 7.7.1): * the difference between sequence numbers s1 and s2 instead of * the number of packets missing between s1 and s2 (one less than the distance). Since this condition appears in many places of the code, it has been put into a separate function, dccp_loss_free(). Further changes: ---------------- * tidied up incorrect typing (it was using `int' for u64/s64 types); * optimised conditional statements for common case of non-reordered packets; * rewrote comments/documentation to match the changes. Signed-off-by: Gerrit Renker --- net/dccp/ccids/lib/packet_history.c | 73 ++++++++++++++----------------------- net/dccp/dccp.h | 15 ++++++++ 2 files changed, 43 insertions(+), 45 deletions(-) (limited to 'net') diff --git a/net/dccp/ccids/lib/packet_history.c b/net/dccp/ccids/lib/packet_history.c index 20af1a693427..8b5c41ec7ee1 100644 --- a/net/dccp/ccids/lib/packet_history.c +++ b/net/dccp/ccids/lib/packet_history.c @@ -215,22 +215,19 @@ static void __one_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n2 u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno, s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno, s2 = DCCP_SKB_CB(skb)->dccpd_seq; - int n1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_ndp, - d12 = dccp_delta_seqno(s1, s2), d2; - if (d12 > 0) { /* S1 < S2 */ + if (likely(dccp_delta_seqno(s1, s2) > 0)) { /* S1 < S2 */ h->loss_count = 2; tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 2), skb, n2); return; } /* S0 < S2 < S1 */ - d2 = dccp_delta_seqno(s0, s2); - if (d2 == 1 || n2 >= d2) { /* S2 is direct successor of S0 */ - int d21 = -d12; + if (dccp_loss_free(s0, s2, n2)) { + u64 n1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_ndp; - if (d21 == 1 || n1 >= d21) { + if (dccp_loss_free(s2, s1, n1)) { /* hole is filled: S0, S2, and S1 are consecutive */ h->loss_count = 0; h->loss_start = tfrc_rx_hist_index(h, 1); @@ -238,9 +235,9 @@ static void __one_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n2 /* gap between S2 and S1: just update loss_prev */ tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_loss_prev(h), skb, n2); - } else { /* hole between S0 and S2 */ + } else { /* gap between S0 and S2 */ /* - * Reorder history to insert S2 between S0 and s1 + * Reorder history to insert S2 between S0 and S1 */ tfrc_rx_hist_swap(h, 0, 3); h->loss_start = tfrc_rx_hist_index(h, 3); @@ -256,22 +253,18 @@ static int __two_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n3) s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno, s2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_seqno, s3 = DCCP_SKB_CB(skb)->dccpd_seq; - int n1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_ndp, - d23 = dccp_delta_seqno(s2, s3), d13, d3, d31; - if (d23 > 0) { /* S2 < S3 */ + if (likely(dccp_delta_seqno(s2, s3) > 0)) { /* S2 < S3 */ h->loss_count = 3; tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 3), skb, n3); return 1; } /* S3 < S2 */ - d13 = dccp_delta_seqno(s1, s3); - if (d13 > 0) { + if (dccp_delta_seqno(s1, s3) > 0) { /* S1 < S3 < S2 */ /* - * The sequence number order is S1, S3, S2 - * Reorder history to insert entry between S1 and S2 + * Reorder history to insert S3 between S1 and S2 */ tfrc_rx_hist_swap(h, 2, 3); tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 2), skb, n3); @@ -280,17 +273,15 @@ static int __two_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n3) } /* S0 < S3 < S1 */ - d31 = -d13; - d3 = dccp_delta_seqno(s0, s3); - if (d3 == 1 || n3 >= d3) { /* S3 is a successor of S0 */ + if (dccp_loss_free(s0, s3, n3)) { + u64 n1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_ndp; - if (d31 == 1 || n1 >= d31) { + if (dccp_loss_free(s3, s1, n1)) { /* hole between S0 and S1 filled by S3 */ - int d2 = dccp_delta_seqno(s1, s2), - n2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_ndp; + u64 n2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_ndp; - if (d2 == 1 || n2 >= d2) { + if (dccp_loss_free(s1, s2, n2)) { /* entire hole filled by S0, S3, S1, S2 */ h->loss_start = tfrc_rx_hist_index(h, 2); h->loss_count = 0; @@ -307,8 +298,8 @@ static int __two_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n3) } /* - * The remaining case: S3 is not a successor of S0. - * Sequence order is S0, S3, S1, S2; reorder to insert between S0 and S1 + * The remaining case: S0 < S3 < S1 < S2; gap between S0 and S3 + * Reorder history to insert S3 between S0 and S1. */ tfrc_rx_hist_swap(h, 0, 3); h->loss_start = tfrc_rx_hist_index(h, 3); @@ -318,33 +309,25 @@ static int __two_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n3) return 1; } -/* return the signed modulo-2^48 sequence number distance from entry e1 to e2 */ -static s64 tfrc_rx_hist_delta_seqno(struct tfrc_rx_hist *h, u8 e1, u8 e2) -{ - DCCP_BUG_ON(e1 > h->loss_count || e2 > h->loss_count); - - return dccp_delta_seqno(tfrc_rx_hist_entry(h, e1)->tfrchrx_seqno, - tfrc_rx_hist_entry(h, e2)->tfrchrx_seqno); -} - /* recycle RX history records to continue loss detection if necessary */ static void __three_after_loss(struct tfrc_rx_hist *h) { /* - * The distance between S0 and S1 is always greater than 1 and the NDP - * count of S1 is smaller than this distance. Otherwise there would - * have been no loss. Hence it is only necessary to see whether there - * are further missing data packets between S1/S2 and S2/S3. + * At this stage we know already that there is a gap between S0 and S1 + * (since S0 was the highest sequence number received before detecting + * the loss). To recycle the loss record, it is thus only necessary to + * check for other possible gaps between S1/S2 and between S2/S3. */ - int d2 = tfrc_rx_hist_delta_seqno(h, 1, 2), - d3 = tfrc_rx_hist_delta_seqno(h, 2, 3), - n2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_ndp, + u64 s1 = tfrc_rx_hist_entry(h, 1)->tfrchrx_seqno, + s2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_seqno, + s3 = tfrc_rx_hist_entry(h, 3)->tfrchrx_seqno; + u64 n2 = tfrc_rx_hist_entry(h, 2)->tfrchrx_ndp, n3 = tfrc_rx_hist_entry(h, 3)->tfrchrx_ndp; - if (d2 == 1 || n2 >= d2) { /* S2 is successor to S1 */ + if (dccp_loss_free(s1, s2, n2)) { - if (d3 == 1 || n3 >= d3) { - /* S3 is successor of S2: entire hole is filled */ + if (dccp_loss_free(s2, s3, n3)) { + /* no gap between S2 and S3: entire hole is filled */ h->loss_start = tfrc_rx_hist_index(h, 3); h->loss_count = 0; } else { @@ -353,7 +336,7 @@ static void __three_after_loss(struct tfrc_rx_hist *h) h->loss_count = 1; } - } else { /* gap between S1 and S2 */ + } else { /* gap between S1 and S2 */ h->loss_start = tfrc_rx_hist_index(h, 1); h->loss_count = 2; } diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h index 1b2cea244e12..32617e0576cb 100644 --- a/net/dccp/dccp.h +++ b/net/dccp/dccp.h @@ -153,6 +153,21 @@ static inline u64 max48(const u64 seq1, const u64 seq2) return after48(seq1, seq2) ? seq1 : seq2; } +/** + * dccp_loss_free - Evaluates condition for data loss from RFC 4340, 7.7.1 + * @s1: start sequence number + * @s2: end sequence number + * @ndp: NDP count on packet with sequence number @s2 + * Returns true if the sequence range s1...s2 has no data loss. + */ +static inline bool dccp_loss_free(const u64 s1, const u64 s2, const u64 ndp) +{ + s64 delta = dccp_delta_seqno(s1, s2); + + BUG_TRAP(delta >= 0); + return (u64)delta <= ndp + 1; +} + enum { DCCP_MIB_NUM = 0, DCCP_MIB_ACTIVEOPENS, /* ActiveOpens */ -- cgit v1.2.2 From 5b5d0e704880addfd979c262e6441f126708539c Mon Sep 17 00:00:00 2001 From: Gerrit Renker Date: Sun, 13 Jul 2008 11:51:40 +0100 Subject: dccp: Upgrade NDP count from 3 to 6 bytes RFC 4340, 7.7 specifies up to 6 bytes for the NDP Count option, whereas the code is currently limited to up to 3 bytes. This seems to be a relict of an earlier draft version and is brought up to date by the patch. Signed-off-by: Gerrit Renker --- net/dccp/ccids/ccid3.c | 2 +- net/dccp/ccids/lib/packet_history.c | 6 +++--- net/dccp/ccids/lib/packet_history.h | 6 +++--- net/dccp/options.c | 14 ++++++++------ 4 files changed, 15 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c index a1929f33d703..523db262c18f 100644 --- a/net/dccp/ccids/ccid3.c +++ b/net/dccp/ccids/ccid3.c @@ -794,7 +794,7 @@ static void ccid3_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb) { struct ccid3_hc_rx_sock *hcrx = ccid3_hc_rx_sk(sk); enum ccid3_fback_type do_feedback = CCID3_FBACK_NONE; - const u32 ndp = dccp_sk(sk)->dccps_options_received.dccpor_ndp; + const u64 ndp = dccp_sk(sk)->dccps_options_received.dccpor_ndp; const bool is_data_packet = dccp_data_packet(skb); if (unlikely(hcrx->ccid3hcrx_state == TFRC_RSTATE_NO_DATA)) { diff --git a/net/dccp/ccids/lib/packet_history.c b/net/dccp/ccids/lib/packet_history.c index 8b5c41ec7ee1..712930564c38 100644 --- a/net/dccp/ccids/lib/packet_history.c +++ b/net/dccp/ccids/lib/packet_history.c @@ -153,7 +153,7 @@ void tfrc_rx_packet_history_exit(void) static inline void tfrc_rx_hist_entry_from_skb(struct tfrc_rx_hist_entry *entry, const struct sk_buff *skb, - const u32 ndp) + const u64 ndp) { const struct dccp_hdr *dh = dccp_hdr(skb); @@ -166,7 +166,7 @@ static inline void tfrc_rx_hist_entry_from_skb(struct tfrc_rx_hist_entry *entry, void tfrc_rx_hist_add_packet(struct tfrc_rx_hist *h, const struct sk_buff *skb, - const u32 ndp) + const u64 ndp) { struct tfrc_rx_hist_entry *entry = tfrc_rx_hist_last_rcv(h); @@ -356,7 +356,7 @@ static void __three_after_loss(struct tfrc_rx_hist *h) */ int tfrc_rx_handle_loss(struct tfrc_rx_hist *h, struct tfrc_loss_hist *lh, - struct sk_buff *skb, u32 ndp, + struct sk_buff *skb, const u64 ndp, u32 (*calc_first_li)(struct sock *), struct sock *sk) { int is_new_loss = 0; diff --git a/net/dccp/ccids/lib/packet_history.h b/net/dccp/ccids/lib/packet_history.h index c7eeda49cb20..6976156cda3c 100644 --- a/net/dccp/ccids/lib/packet_history.h +++ b/net/dccp/ccids/lib/packet_history.h @@ -64,7 +64,7 @@ struct tfrc_rx_hist_entry { u64 tfrchrx_seqno:48, tfrchrx_ccval:4, tfrchrx_type:4; - u32 tfrchrx_ndp; /* In fact it is from 8 to 24 bits */ + u64 tfrchrx_ndp:48; ktime_t tfrchrx_tstamp; }; @@ -145,14 +145,14 @@ static inline int tfrc_rx_hist_new_loss_indicated(struct tfrc_rx_hist *h, } extern void tfrc_rx_hist_add_packet(struct tfrc_rx_hist *h, - const struct sk_buff *skb, const u32 ndp); + const struct sk_buff *skb, const u64 ndp); extern int tfrc_rx_hist_duplicate(struct tfrc_rx_hist *h, struct sk_buff *skb); struct tfrc_loss_hist; extern int tfrc_rx_handle_loss(struct tfrc_rx_hist *h, struct tfrc_loss_hist *lh, - struct sk_buff *skb, u32 ndp, + struct sk_buff *skb, const u64 ndp, u32 (*first_li)(struct sock *sk), struct sock *sk); extern u32 tfrc_rx_hist_sample_rtt(struct tfrc_rx_hist *h, diff --git a/net/dccp/options.c b/net/dccp/options.c index 43bc24e761d0..dc7c158a2f4b 100644 --- a/net/dccp/options.c +++ b/net/dccp/options.c @@ -124,12 +124,12 @@ int dccp_parse_options(struct sock *sk, struct dccp_request_sock *dreq, mandatory = 1; break; case DCCPO_NDP_COUNT: - if (len > 3) + if (len > 6) goto out_invalid_option; opt_recv->dccpor_ndp = dccp_decode_value_var(value, len); - dccp_pr_debug("%s rx opt: NDP count=%d\n", dccp_role(sk), - opt_recv->dccpor_ndp); + dccp_pr_debug("%s opt: NDP count=%llu\n", dccp_role(sk), + (unsigned long long)opt_recv->dccpor_ndp); break; case DCCPO_CHANGE_L: /* fall through */ @@ -307,9 +307,11 @@ static void dccp_encode_value_var(const u32 value, unsigned char *to, *to++ = (value & 0xFF); } -static inline int dccp_ndp_len(const int ndp) +static inline u8 dccp_ndp_len(const u64 ndp) { - return likely(ndp <= 0xFF) ? 1 : ndp <= 0xFFFF ? 2 : 3; + if (likely(ndp <= 0xFF)) + return 1; + return likely(ndp <= USHORT_MAX) ? 2 : (ndp <= UINT_MAX ? 4 : 6); } int dccp_insert_option(struct sock *sk, struct sk_buff *skb, @@ -336,7 +338,7 @@ EXPORT_SYMBOL_GPL(dccp_insert_option); static int dccp_insert_option_ndp(struct sock *sk, struct sk_buff *skb) { struct dccp_sock *dp = dccp_sk(sk); - int ndp = dp->dccps_ndp_count; + u64 ndp = dp->dccps_ndp_count; if (dccp_non_data_packet(skb)) ++dp->dccps_ndp_count; -- cgit v1.2.2 From b552c6231f19d50165bbf59e8b34d3f713ab5c01 Mon Sep 17 00:00:00 2001 From: Gerrit Renker Date: Sun, 13 Jul 2008 11:51:40 +0100 Subject: dccp ccid-3: Fix a loss detection bug This fixes a bug in the logic of the TFRC loss detection: * new_loss_indicated() should not be called while a loss is pending; * but the code allows this; * thus, for two subsequent gaps in the sequence space, when loss_count has not yet reached NDUPACK=3, the loss_count is falsely reduced to 1. To avoid further and similar problems, all loss handling and loss detection is now done inside tfrc_rx_hist_handle_loss(), using an appropriate routine to track new losses. Further changes: ---------------- * added a reminder that no RX history operations should be performed when rx_handle_loss() has identified a (new) loss, since the function takes care of packet reordering during loss detection; * made tfrc_rx_hist_loss_pending() bool (thanks to an earlier suggestion by Arnaldo); * removed unused functions. Signed-off-by: Gerrit Renker --- net/dccp/ccids/ccid3.c | 12 +++++------- net/dccp/ccids/lib/packet_history.c | 24 ++++++++++++++++++++---- net/dccp/ccids/lib/packet_history.h | 24 ++---------------------- 3 files changed, 27 insertions(+), 33 deletions(-) (limited to 'net') diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c index 523db262c18f..f6756e0c9e69 100644 --- a/net/dccp/ccids/ccid3.c +++ b/net/dccp/ccids/ccid3.c @@ -825,18 +825,16 @@ static void ccid3_hc_rx_packet_recv(struct sock *sk, struct sk_buff *skb) } /* - * Handle pending losses and otherwise check for new loss + * Perform loss detection and handle pending losses */ - if (tfrc_rx_hist_loss_pending(&hcrx->ccid3hcrx_hist) && - tfrc_rx_handle_loss(&hcrx->ccid3hcrx_hist, - &hcrx->ccid3hcrx_li_hist, - skb, ndp, ccid3_first_li, sk) ) { + if (tfrc_rx_handle_loss(&hcrx->ccid3hcrx_hist, &hcrx->ccid3hcrx_li_hist, + skb, ndp, ccid3_first_li, sk)) { do_feedback = CCID3_FBACK_PARAM_CHANGE; goto done_receiving; } - if (tfrc_rx_hist_new_loss_indicated(&hcrx->ccid3hcrx_hist, skb, ndp)) - goto update_records; + if (tfrc_rx_hist_loss_pending(&hcrx->ccid3hcrx_hist)) + return; /* done receiving */ /* * Handle data packets: RTT sampling and monitoring p diff --git a/net/dccp/ccids/lib/packet_history.c b/net/dccp/ccids/lib/packet_history.c index 712930564c38..6cc108afdc3b 100644 --- a/net/dccp/ccids/lib/packet_history.c +++ b/net/dccp/ccids/lib/packet_history.c @@ -206,10 +206,21 @@ static void tfrc_rx_hist_swap(struct tfrc_rx_hist *h, const u8 a, const u8 b) * * In the descriptions, `Si' refers to the sequence number of entry number i, * whose NDP count is `Ni' (lower case is used for variables). - * Note: All __after_loss functions expect that a test against duplicates has - * been performed already: the seqno of the skb must not be less than the - * seqno of loss_prev; and it must not equal that of any valid hist_entry. + * Note: All __xxx_loss functions expect that a test against duplicates has been + * performed already: the seqno of the skb must not be less than the seqno + * of loss_prev; and it must not equal that of any valid history entry. */ +static void __do_track_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u64 n1) +{ + u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno, + s1 = DCCP_SKB_CB(skb)->dccpd_seq; + + if (!dccp_loss_free(s0, s1, n1)) { /* gap between S0 and S1 */ + h->loss_count = 1; + tfrc_rx_hist_entry_from_skb(tfrc_rx_hist_entry(h, 1), skb, n1); + } +} + static void __one_after_loss(struct tfrc_rx_hist *h, struct sk_buff *skb, u32 n2) { u64 s0 = tfrc_rx_hist_loss_prev(h)->tfrchrx_seqno, @@ -353,6 +364,9 @@ static void __three_after_loss(struct tfrc_rx_hist *h) * Chooses action according to pending loss, updates LI database when a new * loss was detected, and does required post-processing. Returns 1 when caller * should send feedback, 0 otherwise. + * Since it also takes care of reordering during loss detection and updates the + * records accordingly, the caller should not perform any more RX history + * operations when loss_count is greater than 0 after calling this function. */ int tfrc_rx_handle_loss(struct tfrc_rx_hist *h, struct tfrc_loss_hist *lh, @@ -361,7 +375,9 @@ int tfrc_rx_handle_loss(struct tfrc_rx_hist *h, { int is_new_loss = 0; - if (h->loss_count == 1) { + if (h->loss_count == 0) { + __do_track_loss(h, skb, ndp); + } else if (h->loss_count == 1) { __one_after_loss(h, skb, ndp); } else if (h->loss_count != 2) { DCCP_BUG("invalid loss_count %d", h->loss_count); diff --git a/net/dccp/ccids/lib/packet_history.h b/net/dccp/ccids/lib/packet_history.h index 6976156cda3c..461cc91cce88 100644 --- a/net/dccp/ccids/lib/packet_history.h +++ b/net/dccp/ccids/lib/packet_history.h @@ -118,30 +118,10 @@ static inline struct tfrc_rx_hist_entry * return h->ring[h->loss_start]; } -/* initialise loss detection and disable RTT sampling */ -static inline void tfrc_rx_hist_loss_indicated(struct tfrc_rx_hist *h) -{ - h->loss_count = 1; -} - /* indicate whether previously a packet was detected missing */ -static inline int tfrc_rx_hist_loss_pending(const struct tfrc_rx_hist *h) -{ - return h->loss_count; -} - -/* any data packets missing between last reception and skb ? */ -static inline int tfrc_rx_hist_new_loss_indicated(struct tfrc_rx_hist *h, - const struct sk_buff *skb, - u32 ndp) +static inline bool tfrc_rx_hist_loss_pending(const struct tfrc_rx_hist *h) { - int delta = dccp_delta_seqno(tfrc_rx_hist_last_rcv(h)->tfrchrx_seqno, - DCCP_SKB_CB(skb)->dccpd_seq); - - if (delta > 1 && ndp < delta) - tfrc_rx_hist_loss_indicated(h); - - return tfrc_rx_hist_loss_pending(h); + return h->loss_count > 0; } extern void tfrc_rx_hist_add_packet(struct tfrc_rx_hist *h, -- cgit v1.2.2 From 2eeea7ba6b4b65ed27b7646a1bdea3b45973c861 Mon Sep 17 00:00:00 2001 From: Gerrit Renker Date: Sun, 13 Jul 2008 11:51:40 +0100 Subject: dccp ccid-3: Length of loss intervals This corrects an error in the computation of the open loss interval I_0: * the interval length is (highest_seqno - start_seqno) + 1 * and not (highest_seqno - start_seqno). This condition was not fully clear in RFC 3448, but reflects the current revision state of rfc3448bis and is also consistent with RFC 4340, 6.1.1. Further changes: ---------------- * variable renamed due to line length constraints; * explicit typecast to `s64' to avoid implicit signed/unsigned casting. Signed-off-by: Gerrit Renker --- net/dccp/ccids/lib/loss_interval.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/dccp/ccids/lib/loss_interval.c b/net/dccp/ccids/lib/loss_interval.c index 849e181e698f..bcd6ac415bb9 100644 --- a/net/dccp/ccids/lib/loss_interval.c +++ b/net/dccp/ccids/lib/loss_interval.c @@ -90,14 +90,14 @@ u8 tfrc_lh_update_i_mean(struct tfrc_loss_hist *lh, struct sk_buff *skb) { struct tfrc_loss_interval *cur = tfrc_lh_peek(lh); u32 old_i_mean = lh->i_mean; - s64 length; + s64 len; if (cur == NULL) /* not initialised */ return 0; - length = dccp_delta_seqno(cur->li_seqno, DCCP_SKB_CB(skb)->dccpd_seq); + len = dccp_delta_seqno(cur->li_seqno, DCCP_SKB_CB(skb)->dccpd_seq) + 1; - if (length - cur->li_length <= 0) /* duplicate or reordered */ + if (len - (s64)cur->li_length <= 0) /* duplicate or reordered */ return 0; if (SUB16(dccp_hdr(skb)->dccph_ccval, cur->li_ccval) > 4) @@ -114,7 +114,7 @@ u8 tfrc_lh_update_i_mean(struct tfrc_loss_hist *lh, struct sk_buff *skb) if (tfrc_lh_length(lh) == 1) /* due to RFC 3448, 6.3.1 */ return 0; - cur->li_length = length; + cur->li_length = len; tfrc_lh_calc_i_mean(lh); return (lh->i_mean < old_i_mean); @@ -159,7 +159,7 @@ int tfrc_lh_interval_add(struct tfrc_loss_hist *lh, struct tfrc_rx_hist *rh, else { cur->li_length = dccp_delta_seqno(cur->li_seqno, new->li_seqno); new->li_length = dccp_delta_seqno(new->li_seqno, - tfrc_rx_hist_last_rcv(rh)->tfrchrx_seqno); + tfrc_rx_hist_last_rcv(rh)->tfrchrx_seqno) + 1; if (lh->counter > (2*LIH_SIZE)) lh->counter -= LIH_SIZE; -- cgit v1.2.2 From 79d554a6976a295aa9212172b218f29ca71c3b3d Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:44 +0200 Subject: [Bluetooth] Change retrieval of L2CAP features mask Getting the remote L2CAP features mask is really important, but doing this as less intrusive as possible is tricky. To play nice with older systems and Bluetooth qualification testing, the features mask is now only retrieved in two specific cases and only once per lifetime of an ACL link. When trying to establish a L2CAP connection and the remote features mask is unknown, the L2CAP information request is sent when the ACL link goes into connected state. This applies only to outgoing connections and also only for the connection oriented channels. The second case is when a connection request has been received. In this case a connection response with the result pending and the information request will be send. After receiving an information response or if the timeout gets triggered, the normal connection setup process with security setup will be initiated. Signed-off-by: Marcel Holtmann --- net/bluetooth/l2cap.c | 166 +++++++++++++++++++++++++++++++++++--------------- 1 file changed, 118 insertions(+), 48 deletions(-) (limited to 'net') diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c index 6e180d255505..2e3abdfbd69d 100644 --- a/net/bluetooth/l2cap.c +++ b/net/bluetooth/l2cap.c @@ -55,7 +55,7 @@ #define BT_DBG(D...) #endif -#define VERSION "2.9" +#define VERSION "2.10" static u32 l2cap_feat_mask = 0x0000; @@ -253,6 +253,21 @@ static void l2cap_chan_del(struct sock *sk, int err) sk->sk_state_change(sk); } +/* Service level security */ +static inline int l2cap_check_link_mode(struct sock *sk) +{ + struct l2cap_conn *conn = l2cap_pi(sk)->conn; + + if ((l2cap_pi(sk)->link_mode & L2CAP_LM_ENCRYPT) || + (l2cap_pi(sk)->link_mode & L2CAP_LM_SECURE)) + return hci_conn_encrypt(conn->hcon); + + if (l2cap_pi(sk)->link_mode & L2CAP_LM_AUTH) + return hci_conn_auth(conn->hcon); + + return 1; +} + static inline u8 l2cap_get_ident(struct l2cap_conn *conn) { u8 id; @@ -287,6 +302,34 @@ static inline int l2cap_send_cmd(struct l2cap_conn *conn, u8 ident, u8 code, u16 return hci_send_acl(conn->hcon, skb, 0); } +static void l2cap_do_start(struct sock *sk) +{ + struct l2cap_conn *conn = l2cap_pi(sk)->conn; + + if (conn->info_state & L2CAP_INFO_FEAT_MASK_REQ_SENT) { + struct l2cap_conn_req req; + req.scid = cpu_to_le16(l2cap_pi(sk)->scid); + req.psm = l2cap_pi(sk)->psm; + + l2cap_pi(sk)->ident = l2cap_get_ident(conn); + + l2cap_send_cmd(conn, l2cap_pi(sk)->ident, + L2CAP_CONN_REQ, sizeof(req), &req); + } else { + struct l2cap_info_req req; + req.type = cpu_to_le16(L2CAP_IT_FEAT_MASK); + + conn->info_state |= L2CAP_INFO_FEAT_MASK_REQ_SENT; + conn->info_ident = l2cap_get_ident(conn); + + mod_timer(&conn->info_timer, jiffies + + msecs_to_jiffies(L2CAP_INFO_TIMEOUT)); + + l2cap_send_cmd(conn, conn->info_ident, + L2CAP_INFO_REQ, sizeof(req), &req); + } +} + /* ---- L2CAP connections ---- */ static void l2cap_conn_start(struct l2cap_conn *conn) { @@ -301,16 +344,35 @@ static void l2cap_conn_start(struct l2cap_conn *conn) bh_lock_sock(sk); if (sk->sk_type != SOCK_SEQPACKET) { - l2cap_sock_clear_timer(sk); - sk->sk_state = BT_CONNECTED; - sk->sk_state_change(sk); - } else if (sk->sk_state == BT_CONNECT) { + bh_unlock_sock(sk); + continue; + } + + if (sk->sk_state == BT_CONNECT) { struct l2cap_conn_req req; - l2cap_pi(sk)->ident = l2cap_get_ident(conn); req.scid = cpu_to_le16(l2cap_pi(sk)->scid); req.psm = l2cap_pi(sk)->psm; + + l2cap_pi(sk)->ident = l2cap_get_ident(conn); + l2cap_send_cmd(conn, l2cap_pi(sk)->ident, L2CAP_CONN_REQ, sizeof(req), &req); + } else if (sk->sk_state == BT_CONNECT2) { + struct l2cap_conn_rsp rsp; + rsp.scid = cpu_to_le16(l2cap_pi(sk)->dcid); + rsp.dcid = cpu_to_le16(l2cap_pi(sk)->scid); + + if (l2cap_check_link_mode(sk)) { + sk->sk_state = BT_CONFIG; + rsp.result = cpu_to_le16(L2CAP_CR_SUCCESS); + rsp.status = cpu_to_le16(L2CAP_CS_NO_INFO); + } else { + rsp.result = cpu_to_le16(L2CAP_CR_PEND); + rsp.status = cpu_to_le16(L2CAP_CS_AUTHEN_PEND); + } + + l2cap_send_cmd(conn, l2cap_pi(sk)->ident, + L2CAP_CONN_RSP, sizeof(rsp), &rsp); } bh_unlock_sock(sk); @@ -321,22 +383,27 @@ static void l2cap_conn_start(struct l2cap_conn *conn) static void l2cap_conn_ready(struct l2cap_conn *conn) { - BT_DBG("conn %p", conn); + struct l2cap_chan_list *l = &conn->chan_list; + struct sock *sk; - if (conn->chan_list.head || !hlist_empty(&l2cap_sk_list.head)) { - struct l2cap_info_req req; + BT_DBG("conn %p", conn); - req.type = cpu_to_le16(L2CAP_IT_FEAT_MASK); + read_lock(&l->lock); - conn->info_state |= L2CAP_INFO_FEAT_MASK_REQ_SENT; - conn->info_ident = l2cap_get_ident(conn); + for (sk = l->head; sk; sk = l2cap_pi(sk)->next_c) { + bh_lock_sock(sk); - mod_timer(&conn->info_timer, - jiffies + msecs_to_jiffies(L2CAP_INFO_TIMEOUT)); + if (sk->sk_type != SOCK_SEQPACKET) { + l2cap_sock_clear_timer(sk); + sk->sk_state = BT_CONNECTED; + sk->sk_state_change(sk); + } else if (sk->sk_state == BT_CONNECT) + l2cap_do_start(sk); - l2cap_send_cmd(conn, conn->info_ident, - L2CAP_INFO_REQ, sizeof(req), &req); + bh_unlock_sock(sk); } + + read_unlock(&l->lock); } /* Notify sockets that we cannot guaranty reliability anymore */ @@ -729,22 +796,11 @@ static int l2cap_do_connect(struct sock *sk) l2cap_sock_set_timer(sk, sk->sk_sndtimeo); if (hcon->state == BT_CONNECTED) { - if (!(conn->info_state & L2CAP_INFO_FEAT_MASK_REQ_SENT)) { - l2cap_conn_ready(conn); - goto done; - } - - if (sk->sk_type == SOCK_SEQPACKET) { - struct l2cap_conn_req req; - l2cap_pi(sk)->ident = l2cap_get_ident(conn); - req.scid = cpu_to_le16(l2cap_pi(sk)->scid); - req.psm = l2cap_pi(sk)->psm; - l2cap_send_cmd(conn, l2cap_pi(sk)->ident, - L2CAP_CONN_REQ, sizeof(req), &req); - } else { + if (sk->sk_type != SOCK_SEQPACKET) { l2cap_sock_clear_timer(sk); sk->sk_state = BT_CONNECTED; - } + } else + l2cap_do_start(sk); } done: @@ -1477,7 +1533,7 @@ static inline int l2cap_connect_req(struct l2cap_conn *conn, struct l2cap_cmd_hd struct l2cap_conn_req *req = (struct l2cap_conn_req *) data; struct l2cap_conn_rsp rsp; struct sock *sk, *parent; - int result = 0, status = 0; + int result, status = 0; u16 dcid = 0, scid = __le16_to_cpu(req->scid); __le16 psm = req->psm; @@ -1526,25 +1582,24 @@ static inline int l2cap_connect_req(struct l2cap_conn *conn, struct l2cap_cmd_hd l2cap_sock_set_timer(sk, sk->sk_sndtimeo); - /* Service level security */ - result = L2CAP_CR_PEND; - status = L2CAP_CS_AUTHEN_PEND; - sk->sk_state = BT_CONNECT2; l2cap_pi(sk)->ident = cmd->ident; - if ((l2cap_pi(sk)->link_mode & L2CAP_LM_ENCRYPT) || - (l2cap_pi(sk)->link_mode & L2CAP_LM_SECURE)) { - if (!hci_conn_encrypt(conn->hcon)) - goto done; - } else if (l2cap_pi(sk)->link_mode & L2CAP_LM_AUTH) { - if (!hci_conn_auth(conn->hcon)) - goto done; + if (conn->info_state & L2CAP_INFO_FEAT_MASK_REQ_SENT) { + if (l2cap_check_link_mode(sk)) { + sk->sk_state = BT_CONFIG; + result = L2CAP_CR_SUCCESS; + status = L2CAP_CS_NO_INFO; + } else { + sk->sk_state = BT_CONNECT2; + result = L2CAP_CR_PEND; + status = L2CAP_CS_AUTHEN_PEND; + } + } else { + sk->sk_state = BT_CONNECT2; + result = L2CAP_CR_PEND; + status = L2CAP_CS_NO_INFO; } - sk->sk_state = BT_CONFIG; - result = status = 0; - -done: write_unlock_bh(&list->lock); response: @@ -1556,6 +1611,21 @@ sendresp: rsp.result = cpu_to_le16(result); rsp.status = cpu_to_le16(status); l2cap_send_cmd(conn, cmd->ident, L2CAP_CONN_RSP, sizeof(rsp), &rsp); + + if (result == L2CAP_CR_PEND && status == L2CAP_CS_NO_INFO) { + struct l2cap_info_req info; + info.type = cpu_to_le16(L2CAP_IT_FEAT_MASK); + + conn->info_state |= L2CAP_INFO_FEAT_MASK_REQ_SENT; + conn->info_ident = l2cap_get_ident(conn); + + mod_timer(&conn->info_timer, jiffies + + msecs_to_jiffies(L2CAP_INFO_TIMEOUT)); + + l2cap_send_cmd(conn, conn->info_ident, + L2CAP_INFO_REQ, sizeof(info), &info); + } + return 0; } @@ -1664,9 +1734,9 @@ static inline int l2cap_config_req(struct l2cap_conn *conn, struct l2cap_cmd_hdr } if (!(l2cap_pi(sk)->conf_state & L2CAP_CONF_REQ_SENT)) { - u8 req[64]; + u8 buf[64]; l2cap_send_cmd(conn, l2cap_get_ident(conn), L2CAP_CONF_REQ, - l2cap_build_conf_req(sk, req), req); + l2cap_build_conf_req(sk, buf), buf); } unlock: -- cgit v1.2.2 From 77db1980565626471a980f0d2d17299e4bd5e7a5 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:45 +0200 Subject: [Bluetooth] Enforce security for outgoing RFCOMM connections Recent tests with various Bluetooth headsets have shown that some of them don't enforce authentication and encryption when connecting. All of them leave it up to the host stack to enforce it. Non of them should allow unencrypted connections, but that is how it is. So in case the link mode settings require authentication and/or encryption it will now also be enforced on outgoing RFCOMM connections. Previously this was only done for incoming connections. This support has a small drawback from a protocol level point of view since the host stack can't really tell with 100% certainty if a remote side is already authenticated or not. So if both sides are configured to enforce authentication it will be requested twice. Most Bluetooth chips are caching this information and thus no extra authentication procedure has to be triggered over-the-air, but it can happen. Signed-off-by: Marcel Holtmann --- net/bluetooth/rfcomm/core.c | 77 +++++++++++++++++++++++++++------------------ net/bluetooth/rfcomm/sock.c | 8 +++-- 2 files changed, 51 insertions(+), 34 deletions(-) (limited to 'net') diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c index 0c2c93735e93..1f92f9ab4959 100644 --- a/net/bluetooth/rfcomm/core.c +++ b/net/bluetooth/rfcomm/core.c @@ -53,7 +53,7 @@ #define BT_DBG(D...) #endif -#define VERSION "1.8" +#define VERSION "1.9" static int disable_cfc = 0; static int channel_mtu = -1; @@ -230,6 +230,21 @@ static int rfcomm_l2sock_create(struct socket **sock) return err; } +static inline int rfcomm_check_link_mode(struct rfcomm_dlc *d) +{ + struct sock *sk = d->session->sock->sk; + + if (d->link_mode & (RFCOMM_LM_ENCRYPT | RFCOMM_LM_SECURE)) { + if (!hci_conn_encrypt(l2cap_pi(sk)->conn->hcon)) + return 1; + } else if (d->link_mode & RFCOMM_LM_AUTH) { + if (!hci_conn_auth(l2cap_pi(sk)->conn->hcon)) + return 1; + } + + return 0; +} + /* ---- RFCOMM DLCs ---- */ static void rfcomm_dlc_timeout(unsigned long arg) { @@ -371,15 +386,23 @@ static int __rfcomm_dlc_open(struct rfcomm_dlc *d, bdaddr_t *src, bdaddr_t *dst, d->addr = __addr(s->initiator, dlci); d->priority = 7; - d->state = BT_CONFIG; + d->state = BT_CONFIG; rfcomm_dlc_link(s, d); + d->out = 1; + d->mtu = s->mtu; d->cfc = (s->cfc == RFCOMM_CFC_UNKNOWN) ? 0 : s->cfc; - if (s->state == BT_CONNECTED) - rfcomm_send_pn(s, 1, d); + if (s->state == BT_CONNECTED) { + if (rfcomm_check_link_mode(d)) + set_bit(RFCOMM_AUTH_PENDING, &d->flags); + else + rfcomm_send_pn(s, 1, d); + } + rfcomm_dlc_set_timer(d, RFCOMM_CONN_TIMEOUT); + return 0; } @@ -1146,21 +1169,6 @@ static int rfcomm_recv_disc(struct rfcomm_session *s, u8 dlci) return 0; } -static inline int rfcomm_check_link_mode(struct rfcomm_dlc *d) -{ - struct sock *sk = d->session->sock->sk; - - if (d->link_mode & (RFCOMM_LM_ENCRYPT | RFCOMM_LM_SECURE)) { - if (!hci_conn_encrypt(l2cap_pi(sk)->conn->hcon)) - return 1; - } else if (d->link_mode & RFCOMM_LM_AUTH) { - if (!hci_conn_auth(l2cap_pi(sk)->conn->hcon)) - return 1; - } - - return 0; -} - static void rfcomm_dlc_accept(struct rfcomm_dlc *d) { struct sock *sk = d->session->sock->sk; @@ -1205,10 +1213,8 @@ static int rfcomm_recv_sabm(struct rfcomm_session *s, u8 dlci) if (rfcomm_check_link_mode(d)) { set_bit(RFCOMM_AUTH_PENDING, &d->flags); rfcomm_dlc_set_timer(d, RFCOMM_AUTH_TIMEOUT); - return 0; - } - - rfcomm_dlc_accept(d); + } else + rfcomm_dlc_accept(d); } return 0; } @@ -1223,10 +1229,8 @@ static int rfcomm_recv_sabm(struct rfcomm_session *s, u8 dlci) if (rfcomm_check_link_mode(d)) { set_bit(RFCOMM_AUTH_PENDING, &d->flags); rfcomm_dlc_set_timer(d, RFCOMM_AUTH_TIMEOUT); - return 0; - } - - rfcomm_dlc_accept(d); + } else + rfcomm_dlc_accept(d); } else { rfcomm_send_dm(s, dlci); } @@ -1636,7 +1640,11 @@ static void rfcomm_process_connect(struct rfcomm_session *s) d = list_entry(p, struct rfcomm_dlc, list); if (d->state == BT_CONFIG) { d->mtu = s->mtu; - rfcomm_send_pn(s, 1, d); + if (rfcomm_check_link_mode(d)) { + set_bit(RFCOMM_AUTH_PENDING, &d->flags); + rfcomm_dlc_set_timer(d, RFCOMM_AUTH_TIMEOUT); + } else + rfcomm_send_pn(s, 1, d); } } } @@ -1709,7 +1717,11 @@ static inline void rfcomm_process_dlcs(struct rfcomm_session *s) if (test_and_clear_bit(RFCOMM_AUTH_ACCEPT, &d->flags)) { rfcomm_dlc_clear_timer(d); - rfcomm_dlc_accept(d); + if (d->out) { + rfcomm_send_pn(s, 1, d); + rfcomm_dlc_set_timer(d, RFCOMM_CONN_TIMEOUT); + } else + rfcomm_dlc_accept(d); if (d->link_mode & RFCOMM_LM_SECURE) { struct sock *sk = s->sock->sk; hci_conn_change_link_key(l2cap_pi(sk)->conn->hcon); @@ -1717,7 +1729,10 @@ static inline void rfcomm_process_dlcs(struct rfcomm_session *s) continue; } else if (test_and_clear_bit(RFCOMM_AUTH_REJECT, &d->flags)) { rfcomm_dlc_clear_timer(d); - rfcomm_send_dm(s, d->dlci); + if (!d->out) + rfcomm_send_dm(s, d->dlci); + else + d->state = BT_CLOSED; __rfcomm_dlc_close(d, ECONNREFUSED); continue; } @@ -1726,7 +1741,7 @@ static inline void rfcomm_process_dlcs(struct rfcomm_session *s) continue; if ((d->state == BT_CONNECTED || d->state == BT_DISCONN) && - d->mscex == RFCOMM_MSCEX_OK) + d->mscex == RFCOMM_MSCEX_OK) rfcomm_process_tx(d); } } diff --git a/net/bluetooth/rfcomm/sock.c b/net/bluetooth/rfcomm/sock.c index 5083adcbfae5..cacb1ab51f99 100644 --- a/net/bluetooth/rfcomm/sock.c +++ b/net/bluetooth/rfcomm/sock.c @@ -309,13 +309,13 @@ static struct sock *rfcomm_sock_alloc(struct net *net, struct socket *sock, int sk->sk_destruct = rfcomm_sock_destruct; sk->sk_sndtimeo = RFCOMM_CONN_TIMEOUT; - sk->sk_sndbuf = RFCOMM_MAX_CREDITS * RFCOMM_DEFAULT_MTU * 10; - sk->sk_rcvbuf = RFCOMM_MAX_CREDITS * RFCOMM_DEFAULT_MTU * 10; + sk->sk_sndbuf = RFCOMM_MAX_CREDITS * RFCOMM_DEFAULT_MTU * 10; + sk->sk_rcvbuf = RFCOMM_MAX_CREDITS * RFCOMM_DEFAULT_MTU * 10; sock_reset_flag(sk, SOCK_ZAPPED); sk->sk_protocol = proto; - sk->sk_state = BT_OPEN; + sk->sk_state = BT_OPEN; bt_sock_link(&rfcomm_sk_list, sk); @@ -413,6 +413,8 @@ static int rfcomm_sock_connect(struct socket *sock, struct sockaddr *addr, int a bacpy(&bt_sk(sk)->dst, &sa->rc_bdaddr); rfcomm_pi(sk)->channel = sa->rc_channel; + d->link_mode = rfcomm_pi(sk)->link_mode; + err = rfcomm_dlc_open(d, &bt_sk(sk)->src, &sa->rc_bdaddr, sa->rc_channel); if (!err) err = bt_sock_wait_state(sk, BT_CONNECTED, -- cgit v1.2.2 From 9719f8afce34d3d04e884873a8a5e3483e30974c Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:45 +0200 Subject: [Bluetooth] Disconnect when encryption gets disabled The Bluetooth specification allows to enable or disable the encryption of an ACL link at any time by either the peer or the remote device. If a L2CAP or RFCOMM connection requested an encrypted link, they will now disconnect that link if the encryption gets disabled. Higher protocols that don't care about encryption (like SDP) are not affected. Signed-off-by: Marcel Holtmann --- net/bluetooth/l2cap.c | 13 ++++++++++++- net/bluetooth/rfcomm/core.c | 8 ++++++++ 2 files changed, 20 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c index 2e3abdfbd69d..252264062f59 100644 --- a/net/bluetooth/l2cap.c +++ b/net/bluetooth/l2cap.c @@ -2197,7 +2197,7 @@ static int l2cap_auth_cfm(struct hci_conn *hcon, u8 status) return 0; } -static int l2cap_encrypt_cfm(struct hci_conn *hcon, u8 status) +static int l2cap_encrypt_cfm(struct hci_conn *hcon, u8 status, u8 encrypt) { struct l2cap_chan_list *l; struct l2cap_conn *conn = hcon->l2cap_data; @@ -2215,8 +2215,19 @@ static int l2cap_encrypt_cfm(struct hci_conn *hcon, u8 status) read_lock(&l->lock); for (sk = l->head; sk; sk = l2cap_pi(sk)->next_c) { + struct l2cap_pinfo *pi = l2cap_pi(sk); + bh_lock_sock(sk); + if ((pi->link_mode & (L2CAP_LM_ENCRYPT | L2CAP_LM_SECURE)) && + (sk->sk_state == BT_CONNECTED || + sk->sk_state == BT_CONFIG) && + !status && encrypt == 0x00) { + __l2cap_sock_close(sk, ECONNREFUSED); + bh_unlock_sock(sk); + continue; + } + if (sk->sk_state != BT_CONNECT2) { bh_unlock_sock(sk); continue; diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c index 1f92f9ab4959..e7a6a03cea37 100644 --- a/net/bluetooth/rfcomm/core.c +++ b/net/bluetooth/rfcomm/core.c @@ -2003,6 +2003,14 @@ static void rfcomm_encrypt_cfm(struct hci_conn *conn, u8 status, u8 encrypt) list_for_each_safe(p, n, &s->dlcs) { d = list_entry(p, struct rfcomm_dlc, list); + if ((d->link_mode & (RFCOMM_LM_ENCRYPT | RFCOMM_LM_SECURE)) && + (d->state == BT_CONNECTED || + d->state == BT_CONFIG) && + !status && encrypt == 0x00) { + __rfcomm_dlc_close(d, ECONNREFUSED); + continue; + } + if (!test_and_clear_bit(RFCOMM_AUTH_PENDING, &d->flags)) continue; -- cgit v1.2.2 From ae29319649b80ed9d28d7b4f164e3f5f75020fc8 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:45 +0200 Subject: [Bluetooth] Update authentication status after successful encryption The authentication status is not communicated to both parties. This is actually a flaw in the Bluetooth specification. Only the requesting side really knows if the authentication was successful or not. This piece of information is however needed on the other side to know if it has to trigger the authentication procedure or not. Worst case is that both sides will request authentication at different times, but this should be avoided since it costs extra time when setting up a new connection. For Bluetooth encryption it is required to authenticate the link first and the encryption status is communicated to both sides. So when a link is switched to encryption it is possible to update the authentication status since it implies an authenticated link. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_event.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index 6aef8f24e581..0aba21a03b3c 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -883,9 +883,11 @@ static inline void hci_encrypt_change_evt(struct hci_dev *hdev, struct sk_buff * conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(ev->handle)); if (conn) { if (!ev->status) { - if (ev->encrypt) + if (ev->encrypt) { + /* Encryption implies authentication */ + conn->link_mode |= HCI_LM_AUTH; conn->link_mode |= HCI_LM_ENCRYPT; - else + } else conn->link_mode &= ~HCI_LM_ENCRYPT; } -- cgit v1.2.2 From 9dc0a3afc08d6c20c284994dcd84531787d00ec2 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:46 +0200 Subject: [Bluetooth] Support the case when headset falls back to SCO link When trying to establish an eSCO link between two devices then it can happen that the remote device falls back to a SCO link. Currently this case is not handled correctly and the message dispatching will break since it is looking for eSCO packets. So in case the configured link falls back to SCO overwrite the link type with the correct value. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_event.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index 0aba21a03b3c..6bc5a0506c6c 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -1314,8 +1314,16 @@ static inline void hci_sync_conn_complete_evt(struct hci_dev *hdev, struct sk_bu hci_dev_lock(hdev); conn = hci_conn_hash_lookup_ba(hdev, ev->link_type, &ev->bdaddr); - if (!conn) - goto unlock; + if (!conn) { + if (ev->link_type == ESCO_LINK) + goto unlock; + + conn = hci_conn_hash_lookup_ba(hdev, ESCO_LINK, &ev->bdaddr); + if (!conn) + goto unlock; + + conn->type = SCO_LINK; + } if (!ev->status) { conn->handle = __le16_to_cpu(ev->handle); -- cgit v1.2.2 From a8746417e864da1ed36dd2432a399fbeb843c2a0 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:46 +0200 Subject: [Bluetooth] Track connection packet type changes The connection packet type can be changed after the connection has been established and thus needs to be properly tracked to ensure that the host stack has always correct and valid information about it. On incoming connections the Bluetooth core switches the supported packet types to the configured list for this controller. However the usefulness of this feature has been questioned a lot. The general consent is that every Bluetooth host stack should enable as many packet types as the hardware actually supports and leave the decision to the link manager software running on the Bluetooth chip. When running on Bluetooth 2.0 or later hardware, don't change the packet type for incoming connections anymore. This hardware likely supports Enhanced Data Rate and thus leave it completely up to the link manager to pick the best packet type. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_conn.c | 32 ++++++++++++++++++++++++-------- net/bluetooth/hci_event.c | 32 +++++++++++++++++++++++++------- 2 files changed, 49 insertions(+), 15 deletions(-) (limited to 'net') diff --git a/net/bluetooth/hci_conn.c b/net/bluetooth/hci_conn.c index f8880261da0e..69c64ce054fb 100644 --- a/net/bluetooth/hci_conn.c +++ b/net/bluetooth/hci_conn.c @@ -59,7 +59,8 @@ void hci_acl_connect(struct hci_conn *conn) BT_DBG("%p", conn); conn->state = BT_CONNECT; - conn->out = 1; + conn->out = 1; + conn->link_mode = HCI_LM_MASTER; conn->attempt++; @@ -76,7 +77,7 @@ void hci_acl_connect(struct hci_conn *conn) memcpy(conn->dev_class, ie->data.dev_class, 3); } - cp.pkt_type = cpu_to_le16(hdev->pkt_type & ACL_PTYPE_MASK); + cp.pkt_type = cpu_to_le16(conn->pkt_type); if (lmp_rswitch_capable(hdev) && !(hdev->link_mode & HCI_LM_MASTER)) cp.role_switch = 0x01; else @@ -122,7 +123,7 @@ void hci_add_sco(struct hci_conn *conn, __u16 handle) conn->out = 1; cp.handle = cpu_to_le16(handle); - cp.pkt_type = cpu_to_le16(hdev->pkt_type & SCO_PTYPE_MASK); + cp.pkt_type = cpu_to_le16(conn->pkt_type); hci_send_cmd(hdev, HCI_OP_ADD_SCO, sizeof(cp), &cp); } @@ -138,7 +139,7 @@ void hci_setup_sync(struct hci_conn *conn, __u16 handle) conn->out = 1; cp.handle = cpu_to_le16(handle); - cp.pkt_type = cpu_to_le16(hdev->esco_type); + cp.pkt_type = cpu_to_le16(conn->pkt_type); cp.tx_bandwidth = cpu_to_le32(0x00001f40); cp.rx_bandwidth = cpu_to_le32(0x00001f40); @@ -199,13 +200,28 @@ struct hci_conn *hci_conn_add(struct hci_dev *hdev, int type, bdaddr_t *dst) return NULL; bacpy(&conn->dst, dst); - conn->hdev = hdev; - conn->type = type; - conn->mode = HCI_CM_ACTIVE; - conn->state = BT_OPEN; + conn->hdev = hdev; + conn->type = type; + conn->mode = HCI_CM_ACTIVE; + conn->state = BT_OPEN; conn->power_save = 1; + switch (type) { + case ACL_LINK: + conn->pkt_type = hdev->pkt_type & ACL_PTYPE_MASK; + break; + case SCO_LINK: + if (lmp_esco_capable(hdev)) + conn->pkt_type = hdev->esco_type & SCO_ESCO_MASK; + else + conn->pkt_type = hdev->pkt_type & SCO_PTYPE_MASK; + break; + case ESCO_LINK: + conn->pkt_type = hdev->esco_type; + break; + } + skb_queue_head_init(&conn->data_q); setup_timer(&conn->disc_timer, hci_conn_timeout, (unsigned long)conn); diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index 6bc5a0506c6c..d4d2dcc40fc7 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -699,14 +699,12 @@ static inline void hci_conn_complete_evt(struct hci_dev *hdev, struct sk_buff *s } /* Set packet type for incoming connection */ - if (!conn->out) { + if (!conn->out && hdev->hci_ver < 3) { struct hci_cp_change_conn_ptype cp; cp.handle = ev->handle; - cp.pkt_type = (conn->type == ACL_LINK) ? - cpu_to_le16(hdev->pkt_type & ACL_PTYPE_MASK): - cpu_to_le16(hdev->pkt_type & SCO_PTYPE_MASK); - - hci_send_cmd(hdev, HCI_OP_CHANGE_CONN_PTYPE, sizeof(cp), &cp); + cp.pkt_type = cpu_to_le16(conn->pkt_type); + hci_send_cmd(hdev, HCI_OP_CHANGE_CONN_PTYPE, + sizeof(cp), &cp); } else { /* Update disconnect timer */ hci_conn_hold(conn); @@ -786,7 +784,7 @@ static inline void hci_conn_request_evt(struct hci_dev *hdev, struct sk_buff *sk struct hci_cp_accept_sync_conn_req cp; bacpy(&cp.bdaddr, &ev->bdaddr); - cp.pkt_type = cpu_to_le16(hdev->esco_type); + cp.pkt_type = cpu_to_le16(conn->pkt_type); cp.tx_bandwidth = cpu_to_le32(0x00001f40); cp.rx_bandwidth = cpu_to_le32(0x00001f40); @@ -1237,6 +1235,22 @@ static inline void hci_clock_offset_evt(struct hci_dev *hdev, struct sk_buff *sk hci_dev_unlock(hdev); } +static inline void hci_pkt_type_change_evt(struct hci_dev *hdev, struct sk_buff *skb) +{ + struct hci_ev_pkt_type_change *ev = (void *) skb->data; + struct hci_conn *conn; + + BT_DBG("%s status %d", hdev->name, ev->status); + + hci_dev_lock(hdev); + + conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(ev->handle)); + if (conn && !ev->status) + conn->pkt_type = __le16_to_cpu(ev->pkt_type); + + hci_dev_unlock(hdev); +} + static inline void hci_pscan_rep_mode_evt(struct hci_dev *hdev, struct sk_buff *skb) { struct hci_ev_pscan_rep_mode *ev = (void *) skb->data; @@ -1480,6 +1494,10 @@ void hci_event_packet(struct hci_dev *hdev, struct sk_buff *skb) hci_clock_offset_evt(hdev, skb); break; + case HCI_EV_PKT_TYPE_CHANGE: + hci_pkt_type_change_evt(hdev, skb); + break; + case HCI_EV_PSCAN_REP_MODE: hci_pscan_rep_mode_evt(hdev, skb); break; -- cgit v1.2.2 From e4e8e37c42bdaaefcb84eeaef0dc1bc3f696f8f6 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:47 +0200 Subject: [Bluetooth] Make use of the default link policy settings The Bluetooth specification supports the default link policy settings on a per host controller basis. For every new connection the link manager would then use these settings. It is better to use this instead of bothering the controller on every connection setup to overwrite the default settings. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_conn.c | 2 ++ net/bluetooth/hci_core.c | 35 +++++++++++++++------- net/bluetooth/hci_event.c | 75 +++++++++++++++++++++++++++++++++++++++-------- 3 files changed, 89 insertions(+), 23 deletions(-) (limited to 'net') diff --git a/net/bluetooth/hci_conn.c b/net/bluetooth/hci_conn.c index 69c64ce054fb..6175ce841e9e 100644 --- a/net/bluetooth/hci_conn.c +++ b/net/bluetooth/hci_conn.c @@ -65,6 +65,8 @@ void hci_acl_connect(struct hci_conn *conn) conn->attempt++; + conn->link_policy = hdev->link_policy; + memset(&cp, 0, sizeof(cp)); bacpy(&cp.bdaddr, &conn->dst); cp.pscan_rep_mode = 0x02; diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c index aec6929f5c16..69b2c1aac08a 100644 --- a/net/bluetooth/hci_core.c +++ b/net/bluetooth/hci_core.c @@ -279,10 +279,20 @@ static void hci_encrypt_req(struct hci_dev *hdev, unsigned long opt) BT_DBG("%s %x", hdev->name, encrypt); - /* Authentication */ + /* Encryption */ hci_send_cmd(hdev, HCI_OP_WRITE_ENCRYPT_MODE, 1, &encrypt); } +static void hci_linkpol_req(struct hci_dev *hdev, unsigned long opt) +{ + __le16 policy = cpu_to_le16(opt); + + BT_DBG("%s %x", hdev->name, opt); + + /* Default link policy */ + hci_send_cmd(hdev, HCI_OP_WRITE_DEF_LINK_POLICY, 2, &policy); +} + /* Get HCI device by index. * Device is held on return. */ struct hci_dev *hci_dev_get(int index) @@ -694,32 +704,35 @@ int hci_dev_cmd(unsigned int cmd, void __user *arg) msecs_to_jiffies(HCI_INIT_TIMEOUT)); break; - case HCISETPTYPE: - hdev->pkt_type = (__u16) dr.dev_opt; - break; - case HCISETLINKPOL: - hdev->link_policy = (__u16) dr.dev_opt; + err = hci_request(hdev, hci_linkpol_req, dr.dev_opt, + msecs_to_jiffies(HCI_INIT_TIMEOUT)); break; case HCISETLINKMODE: - hdev->link_mode = ((__u16) dr.dev_opt) & (HCI_LM_MASTER | HCI_LM_ACCEPT); + hdev->link_mode = ((__u16) dr.dev_opt) & + (HCI_LM_MASTER | HCI_LM_ACCEPT); + break; + + case HCISETPTYPE: + hdev->pkt_type = (__u16) dr.dev_opt; break; case HCISETACLMTU: - hdev->acl_mtu = *((__u16 *)&dr.dev_opt + 1); - hdev->acl_pkts = *((__u16 *)&dr.dev_opt + 0); + hdev->acl_mtu = *((__u16 *) &dr.dev_opt + 1); + hdev->acl_pkts = *((__u16 *) &dr.dev_opt + 0); break; case HCISETSCOMTU: - hdev->sco_mtu = *((__u16 *)&dr.dev_opt + 1); - hdev->sco_pkts = *((__u16 *)&dr.dev_opt + 0); + hdev->sco_mtu = *((__u16 *) &dr.dev_opt + 1); + hdev->sco_pkts = *((__u16 *) &dr.dev_opt + 0); break; default: err = -EINVAL; break; } + hci_dev_put(hdev); return err; } diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index d4d2dcc40fc7..9af181a61650 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -110,6 +110,25 @@ static void hci_cc_role_discovery(struct hci_dev *hdev, struct sk_buff *skb) hci_dev_unlock(hdev); } +static void hci_cc_read_link_policy(struct hci_dev *hdev, struct sk_buff *skb) +{ + struct hci_rp_read_link_policy *rp = (void *) skb->data; + struct hci_conn *conn; + + BT_DBG("%s status 0x%x", hdev->name, rp->status); + + if (rp->status) + return; + + hci_dev_lock(hdev); + + conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(rp->handle)); + if (conn) + conn->link_policy = __le16_to_cpu(rp->policy); + + hci_dev_unlock(hdev); +} + static void hci_cc_write_link_policy(struct hci_dev *hdev, struct sk_buff *skb) { struct hci_rp_write_link_policy *rp = (void *) skb->data; @@ -128,13 +147,41 @@ static void hci_cc_write_link_policy(struct hci_dev *hdev, struct sk_buff *skb) hci_dev_lock(hdev); conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(rp->handle)); - if (conn) { + if (conn) conn->link_policy = get_unaligned_le16(sent + 2); - } hci_dev_unlock(hdev); } +static void hci_cc_read_def_link_policy(struct hci_dev *hdev, struct sk_buff *skb) +{ + struct hci_rp_read_def_link_policy *rp = (void *) skb->data; + + BT_DBG("%s status 0x%x", hdev->name, rp->status); + + if (rp->status) + return; + + hdev->link_policy = __le16_to_cpu(rp->policy); +} + +static void hci_cc_write_def_link_policy(struct hci_dev *hdev, struct sk_buff *skb) +{ + __u8 status = *((__u8 *) skb->data); + void *sent; + + BT_DBG("%s status 0x%x", hdev->name, status); + + sent = hci_sent_cmd_data(hdev, HCI_OP_WRITE_DEF_LINK_POLICY); + if (!sent) + return; + + if (!status) + hdev->link_policy = get_unaligned_le16(sent); + + hci_req_complete(hdev, status); +} + static void hci_cc_reset(struct hci_dev *hdev, struct sk_buff *skb) { __u8 status = *((__u8 *) skb->data); @@ -347,8 +394,8 @@ static void hci_cc_read_local_version(struct hci_dev *hdev, struct sk_buff *skb) return; hdev->hci_ver = rp->hci_ver; - hdev->hci_rev = btohs(rp->hci_rev); - hdev->manufacturer = btohs(rp->manufacturer); + hdev->hci_rev = __le16_to_cpu(rp->hci_rev); + hdev->manufacturer = __le16_to_cpu(rp->manufacturer); BT_DBG("%s manufacturer %d hci ver %d:%d", hdev->name, hdev->manufacturer, @@ -690,14 +737,6 @@ static inline void hci_conn_complete_evt(struct hci_dev *hdev, struct sk_buff *s hci_send_cmd(hdev, HCI_OP_READ_REMOTE_FEATURES, sizeof(cp), &cp); } - /* Set link policy */ - if (conn->type == ACL_LINK && hdev->link_policy) { - struct hci_cp_write_link_policy cp; - cp.handle = ev->handle; - cp.policy = cpu_to_le16(hdev->link_policy); - hci_send_cmd(hdev, HCI_OP_WRITE_LINK_POLICY, sizeof(cp), &cp); - } - /* Set packet type for incoming connection */ if (!conn->out && hdev->hci_ver < 3) { struct hci_cp_change_conn_ptype cp; @@ -974,10 +1013,22 @@ static inline void hci_cmd_complete_evt(struct hci_dev *hdev, struct sk_buff *sk hci_cc_role_discovery(hdev, skb); break; + case HCI_OP_READ_LINK_POLICY: + hci_cc_read_link_policy(hdev, skb); + break; + case HCI_OP_WRITE_LINK_POLICY: hci_cc_write_link_policy(hdev, skb); break; + case HCI_OP_READ_DEF_LINK_POLICY: + hci_cc_read_def_link_policy(hdev, skb); + break; + + case HCI_OP_WRITE_DEF_LINK_POLICY: + hci_cc_write_def_link_policy(hdev, skb); + break; + case HCI_OP_RESET: hci_cc_reset(hdev, skb); break; -- cgit v1.2.2 From f383f2750af19fe6f820edf40d8729f9741c5b37 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:47 +0200 Subject: [Bluetooth] Some cleanups for HCI event handling Some minor cosmetic cleanups to the HCI event handling to make the code easier to read and understand. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_event.c | 39 +++++++++++++++++++++++---------------- 1 file changed, 23 insertions(+), 16 deletions(-) (limited to 'net') diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index 9af181a61650..bf3fbf9817b4 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -198,12 +198,14 @@ static void hci_cc_write_local_name(struct hci_dev *hdev, struct sk_buff *skb) BT_DBG("%s status 0x%x", hdev->name, status); + if (status) + return; + sent = hci_sent_cmd_data(hdev, HCI_OP_WRITE_LOCAL_NAME); if (!sent) return; - if (!status) - memcpy(hdev->dev_name, sent, 248); + memcpy(hdev->dev_name, sent, 248); } static void hci_cc_read_local_name(struct hci_dev *hdev, struct sk_buff *skb) @@ -313,12 +315,14 @@ static void hci_cc_write_class_of_dev(struct hci_dev *hdev, struct sk_buff *skb) BT_DBG("%s status 0x%x", hdev->name, status); + if (status) + return; + sent = hci_sent_cmd_data(hdev, HCI_OP_WRITE_CLASS_OF_DEV); if (!sent) return; - if (!status) - memcpy(hdev->dev_class, sent, 3); + memcpy(hdev->dev_class, sent, 3); } static void hci_cc_read_voice_setting(struct hci_dev *hdev, struct sk_buff *skb) @@ -333,7 +337,7 @@ static void hci_cc_read_voice_setting(struct hci_dev *hdev, struct sk_buff *skb) setting = __le16_to_cpu(rp->voice_setting); - if (hdev->voice_setting == setting ) + if (hdev->voice_setting == setting) return; hdev->voice_setting = setting; @@ -350,28 +354,31 @@ static void hci_cc_read_voice_setting(struct hci_dev *hdev, struct sk_buff *skb) static void hci_cc_write_voice_setting(struct hci_dev *hdev, struct sk_buff *skb) { __u8 status = *((__u8 *) skb->data); + __u16 setting; void *sent; BT_DBG("%s status 0x%x", hdev->name, status); + if (status) + return; + sent = hci_sent_cmd_data(hdev, HCI_OP_WRITE_VOICE_SETTING); if (!sent) return; - if (!status) { - __u16 setting = get_unaligned_le16(sent); + setting = get_unaligned_le16(sent); - if (hdev->voice_setting != setting) { - hdev->voice_setting = setting; + if (hdev->voice_setting == setting) + return; - BT_DBG("%s voice setting 0x%04x", hdev->name, setting); + hdev->voice_setting = setting; - if (hdev->notify) { - tasklet_disable(&hdev->tx_task); - hdev->notify(hdev, HCI_NOTIFY_VOICE_SETTING); - tasklet_enable(&hdev->tx_task); - } - } + BT_DBG("%s voice setting 0x%04x", hdev->name, setting); + + if (hdev->notify) { + tasklet_disable(&hdev->tx_task); + hdev->notify(hdev, HCI_NOTIFY_VOICE_SETTING); + tasklet_enable(&hdev->tx_task); } } -- cgit v1.2.2 From c7bdd5026d28d178238bd794c61612602a54d55e Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:47 +0200 Subject: [Bluetooth] Update class of device value whenever possible The class of device value can only be retrieved via inquiry or during an incoming connection request. Outgoing connections can't ask for the class of device. To compensate for this the value is stored and copied via the inquiry cache, but currently only updated via inquiry. This update should also happen during an incoming connection request. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_event.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'net') diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index bf3fbf9817b4..e47676128bb5 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -796,10 +796,14 @@ static inline void hci_conn_request_evt(struct hci_dev *hdev, struct sk_buff *sk if (mask & HCI_LM_ACCEPT) { /* Connection accepted */ + struct inquiry_entry *ie; struct hci_conn *conn; hci_dev_lock(hdev); + if ((ie = hci_inquiry_cache_lookup(hdev, &ev->bdaddr))) + memcpy(ie->data.dev_class, ev->dev_class, 3); + conn = hci_conn_hash_lookup_ba(hdev, ev->link_type, &ev->bdaddr); if (!conn) { if (!(conn = hci_conn_add(hdev, ev->link_type, &ev->bdaddr))) { -- cgit v1.2.2 From 0493684ed2397e111574f343534d8e4ec440dfa5 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:48 +0200 Subject: [Bluetooth] Disable disconnect timer during Simple Pairing During the Simple Pairing process the HCI disconnect timer must be disabled. The way to do this is by holding a reference count of the HCI connection. The Simple Pairing process on both sides starts with an IO Capabilities Request and ends with Simple Pairing Complete. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_event.c | 40 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 40 insertions(+) (limited to 'net') diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index e47676128bb5..7c9ac01cd8f9 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -1464,6 +1464,38 @@ static inline void hci_extended_inquiry_result_evt(struct hci_dev *hdev, struct hci_dev_unlock(hdev); } +static inline void hci_io_capa_request_evt(struct hci_dev *hdev, struct sk_buff *skb) +{ + struct hci_ev_io_capa_request *ev = (void *) skb->data; + struct hci_conn *conn; + + BT_DBG("%s", hdev->name); + + hci_dev_lock(hdev); + + conn = hci_conn_hash_lookup_ba(hdev, ACL_LINK, &ev->bdaddr); + if (conn) + hci_conn_hold(conn); + + hci_dev_unlock(hdev); +} + +static inline void hci_simple_pair_complete_evt(struct hci_dev *hdev, struct sk_buff *skb) +{ + struct hci_ev_simple_pair_complete *ev = (void *) skb->data; + struct hci_conn *conn; + + BT_DBG("%s", hdev->name); + + hci_dev_lock(hdev); + + conn = hci_conn_hash_lookup_ba(hdev, ACL_LINK, &ev->bdaddr); + if (conn) + hci_conn_put(conn); + + hci_dev_unlock(hdev); +} + void hci_event_packet(struct hci_dev *hdev, struct sk_buff *skb) { struct hci_event_hdr *hdr = (void *) skb->data; @@ -1588,6 +1620,14 @@ void hci_event_packet(struct hci_dev *hdev, struct sk_buff *skb) hci_extended_inquiry_result_evt(hdev, skb); break; + case HCI_EV_IO_CAPA_REQUEST: + hci_io_capa_request_evt(hdev, skb); + break; + + case HCI_EV_SIMPLE_PAIR_COMPLETE: + hci_simple_pair_complete_evt(hdev, skb); + break; + default: BT_DBG("%s event 0x%x", hdev->name, event); break; -- cgit v1.2.2 From 333140b57fa7867bc92e5ee879b6ac4ef5e1d867 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:48 +0200 Subject: [Bluetooth] Track status of Simple Pairing mode The Simple Pairing feature is optional and needs to be enabled by the host stack first. The Linux kernel relies on the Bluetooth daemon to either enable or disable it, but at any time it needs to know the current state of the Simple Pairing mode. So track any changes made by external entities and store the current mode in the HCI device structure. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_event.c | 37 +++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+) (limited to 'net') diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index 7c9ac01cd8f9..6077a651aac6 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -391,6 +391,35 @@ static void hci_cc_host_buffer_size(struct hci_dev *hdev, struct sk_buff *skb) hci_req_complete(hdev, status); } +static void hci_cc_read_ssp_mode(struct hci_dev *hdev, struct sk_buff *skb) +{ + struct hci_rp_read_ssp_mode *rp = (void *) skb->data; + + BT_DBG("%s status 0x%x", hdev->name, rp->status); + + if (rp->status) + return; + + hdev->ssp_mode = rp->mode; +} + +static void hci_cc_write_ssp_mode(struct hci_dev *hdev, struct sk_buff *skb) +{ + __u8 status = *((__u8 *) skb->data); + void *sent; + + BT_DBG("%s status 0x%x", hdev->name, status); + + if (status) + return; + + sent = hci_sent_cmd_data(hdev, HCI_OP_WRITE_SSP_MODE); + if (!sent) + return; + + hdev->ssp_mode = *((__u8 *) sent); +} + static void hci_cc_read_local_version(struct hci_dev *hdev, struct sk_buff *skb) { struct hci_rp_read_local_version *rp = (void *) skb->data; @@ -1084,6 +1113,14 @@ static inline void hci_cmd_complete_evt(struct hci_dev *hdev, struct sk_buff *sk hci_cc_host_buffer_size(hdev, skb); break; + case HCI_OP_READ_SSP_MODE: + hci_cc_read_ssp_mode(hdev, skb); + break; + + case HCI_OP_WRITE_SSP_MODE: + hci_cc_write_ssp_mode(hdev, skb); + break; + case HCI_OP_READ_LOCAL_VERSION: hci_cc_read_local_version(hdev, skb); break; -- cgit v1.2.2 From 41a96212b3b7b3cd59e8e8d33e6dabf0e21d9778 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:48 +0200 Subject: [Bluetooth] Track status of remote Simple Pairing mode The Simple Pairing process can only be used if both sides have the support enabled in the host stack. The current Bluetooth specification has three ways to detect this support. If an Extended Inquiry Result has been sent during inquiry then it is safe to assume that Simple Pairing is enabled. It is not allowed to enable Extended Inquiry without Simple Pairing. During the remote name request phase a notification with the remote host supported features will be sent to indicate Simple Pairing support. Also the second page of the remote extended features can indicate support for Simple Pairing. For all three cases the value of remote Simple Pairing mode is stored in the inquiry cache for later use. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_conn.c | 14 +++++++++----- net/bluetooth/hci_event.c | 43 +++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 52 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/bluetooth/hci_conn.c b/net/bluetooth/hci_conn.c index 6175ce841e9e..41351ba692e9 100644 --- a/net/bluetooth/hci_conn.c +++ b/net/bluetooth/hci_conn.c @@ -71,12 +71,16 @@ void hci_acl_connect(struct hci_conn *conn) bacpy(&cp.bdaddr, &conn->dst); cp.pscan_rep_mode = 0x02; - if ((ie = hci_inquiry_cache_lookup(hdev, &conn->dst)) && - inquiry_entry_age(ie) <= INQUIRY_ENTRY_AGE_MAX) { - cp.pscan_rep_mode = ie->data.pscan_rep_mode; - cp.pscan_mode = ie->data.pscan_mode; - cp.clock_offset = ie->data.clock_offset | cpu_to_le16(0x8000); + if ((ie = hci_inquiry_cache_lookup(hdev, &conn->dst))) { + if (inquiry_entry_age(ie) <= INQUIRY_ENTRY_AGE_MAX) { + cp.pscan_rep_mode = ie->data.pscan_rep_mode; + cp.pscan_mode = ie->data.pscan_mode; + cp.clock_offset = ie->data.clock_offset | + cpu_to_le16(0x8000); + } + memcpy(conn->dev_class, ie->data.dev_class, 3); + conn->ssp_mode = ie->data.ssp_mode; } cp.pkt_type = cpu_to_le16(conn->pkt_type); diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index 6077a651aac6..c8fda7dc2986 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -736,6 +736,7 @@ static inline void hci_inquiry_result_evt(struct hci_dev *hdev, struct sk_buff * memcpy(data.dev_class, info->dev_class, 3); data.clock_offset = info->clock_offset; data.rssi = 0x00; + data.ssp_mode = 0x00; info++; hci_inquiry_cache_update(hdev, &data); } @@ -1390,6 +1391,7 @@ static inline void hci_inquiry_result_with_rssi_evt(struct hci_dev *hdev, struct memcpy(data.dev_class, info->dev_class, 3); data.clock_offset = info->clock_offset; data.rssi = info->rssi; + data.ssp_mode = 0x00; info++; hci_inquiry_cache_update(hdev, &data); } @@ -1404,6 +1406,7 @@ static inline void hci_inquiry_result_with_rssi_evt(struct hci_dev *hdev, struct memcpy(data.dev_class, info->dev_class, 3); data.clock_offset = info->clock_offset; data.rssi = info->rssi; + data.ssp_mode = 0x00; info++; hci_inquiry_cache_update(hdev, &data); } @@ -1414,7 +1417,27 @@ static inline void hci_inquiry_result_with_rssi_evt(struct hci_dev *hdev, struct static inline void hci_remote_ext_features_evt(struct hci_dev *hdev, struct sk_buff *skb) { + struct hci_ev_remote_ext_features *ev = (void *) skb->data; + struct hci_conn *conn; + BT_DBG("%s", hdev->name); + + if (ev->status || ev->page != 0x01) + return; + + hci_dev_lock(hdev); + + conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(ev->handle)); + if (conn) { + struct inquiry_entry *ie; + + if ((ie = hci_inquiry_cache_lookup(hdev, &conn->dst))) + ie->data.ssp_mode = (ev->features[0] & 0x01); + + conn->ssp_mode = (ev->features[0] & 0x01); + } + + hci_dev_unlock(hdev); } static inline void hci_sync_conn_complete_evt(struct hci_dev *hdev, struct sk_buff *skb) @@ -1494,6 +1517,7 @@ static inline void hci_extended_inquiry_result_evt(struct hci_dev *hdev, struct memcpy(data.dev_class, info->dev_class, 3); data.clock_offset = info->clock_offset; data.rssi = info->rssi; + data.ssp_mode = 0x01; info++; hci_inquiry_cache_update(hdev, &data); } @@ -1533,6 +1557,21 @@ static inline void hci_simple_pair_complete_evt(struct hci_dev *hdev, struct sk_ hci_dev_unlock(hdev); } +static inline void hci_remote_host_features_evt(struct hci_dev *hdev, struct sk_buff *skb) +{ + struct hci_ev_remote_host_features *ev = (void *) skb->data; + struct inquiry_entry *ie; + + BT_DBG("%s", hdev->name); + + hci_dev_lock(hdev); + + if ((ie = hci_inquiry_cache_lookup(hdev, &ev->bdaddr))) + ie->data.ssp_mode = (ev->features[0] & 0x01); + + hci_dev_unlock(hdev); +} + void hci_event_packet(struct hci_dev *hdev, struct sk_buff *skb) { struct hci_event_hdr *hdr = (void *) skb->data; @@ -1665,6 +1704,10 @@ void hci_event_packet(struct hci_dev *hdev, struct sk_buff *skb) hci_simple_pair_complete_evt(hdev, skb); break; + case HCI_EV_REMOTE_HOST_FEATURES: + hci_remote_host_features_evt(hdev, skb); + break; + default: BT_DBG("%s event 0x%x", hdev->name, event); break; -- cgit v1.2.2 From a8bd28baf21b9ee6b8486666b771283e566c0d31 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:49 +0200 Subject: [Bluetooth] Export remote Simple Pairing mode via sysfs Since the remote Simple Pairing mode is stored together with the inquiry cache, it makes sense to show it together with the other information. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_sysfs.c | 23 +++++++++++++++++++---- 1 file changed, 19 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/bluetooth/hci_sysfs.c b/net/bluetooth/hci_sysfs.c index 84360c117d4e..a18871e01582 100644 --- a/net/bluetooth/hci_sysfs.c +++ b/net/bluetooth/hci_sysfs.c @@ -113,11 +113,13 @@ static ssize_t show_inquiry_cache(struct device *dev, struct device_attribute *a struct inquiry_data *data = &e->data; bdaddr_t bdaddr; baswap(&bdaddr, &data->bdaddr); - n += sprintf(buf + n, "%s %d %d %d 0x%.2x%.2x%.2x 0x%.4x %d %u\n", + n += sprintf(buf + n, "%s %d %d %d 0x%.2x%.2x%.2x 0x%.4x %d %d %u\n", batostr(&bdaddr), - data->pscan_rep_mode, data->pscan_period_mode, data->pscan_mode, - data->dev_class[2], data->dev_class[1], data->dev_class[0], - __le16_to_cpu(data->clock_offset), data->rssi, e->timestamp); + data->pscan_rep_mode, data->pscan_period_mode, + data->pscan_mode, data->dev_class[2], + data->dev_class[1], data->dev_class[0], + __le16_to_cpu(data->clock_offset), + data->rssi, data->ssp_mode, e->timestamp); } hci_dev_unlock_bh(hdev); @@ -249,15 +251,28 @@ static ssize_t show_conn_address(struct device *dev, struct device_attribute *at return sprintf(buf, "%s\n", batostr(&bdaddr)); } +static ssize_t show_conn_features(struct device *dev, struct device_attribute *attr, char *buf) +{ + struct hci_conn *conn = dev_get_drvdata(dev); + + return sprintf(buf, "0x%02x%02x%02x%02x%02x%02x%02x%02x\n", + conn->features[0], conn->features[1], + conn->features[2], conn->features[3], + conn->features[4], conn->features[5], + conn->features[6], conn->features[7]); +} + #define CONN_ATTR(_name,_mode,_show,_store) \ struct device_attribute conn_attr_##_name = __ATTR(_name,_mode,_show,_store) static CONN_ATTR(type, S_IRUGO, show_conn_type, NULL); static CONN_ATTR(address, S_IRUGO, show_conn_address, NULL); +static CONN_ATTR(features, S_IRUGO, show_conn_features, NULL); static struct device_attribute *conn_attrs[] = { &conn_attr_type, &conn_attr_address, + &conn_attr_features, NULL }; -- cgit v1.2.2 From 769be974d0c7b4fe1a52f9cdaad22259b60953f7 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:49 +0200 Subject: [Bluetooth] Use ACL config stage to retrieve remote features The Bluetooth technology introduces new features on a regular basis and for some of them it is important that the hardware on both sides support them. For features like Simple Pairing it is important that the host stacks on both sides have switched this feature on. To make valid decisions, a config stage during ACL link establishment has been introduced that retrieves remote features and if needed also the remote extended features (known as remote host features) before signalling this link as connected. This change introduces full reference counting of incoming and outgoing ACL links and the Bluetooth core will disconnect both if no owner of it is present. To better handle interoperability during the pairing phase the disconnect timeout for incoming connections has been increased to 10 seconds. This is five times more than for outgoing connections. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_conn.c | 2 + net/bluetooth/hci_core.c | 7 ++- net/bluetooth/hci_event.c | 127 ++++++++++++++++++++++++++++++++++++++-------- 3 files changed, 114 insertions(+), 22 deletions(-) (limited to 'net') diff --git a/net/bluetooth/hci_conn.c b/net/bluetooth/hci_conn.c index 41351ba692e9..6f22533e7656 100644 --- a/net/bluetooth/hci_conn.c +++ b/net/bluetooth/hci_conn.c @@ -170,11 +170,13 @@ static void hci_conn_timeout(unsigned long arg) switch (conn->state) { case BT_CONNECT: + case BT_CONNECT2: if (conn->type == ACL_LINK) hci_acl_connect_cancel(conn); else hci_acl_disconn(conn, 0x13); break; + case BT_CONFIG: case BT_CONNECTED: hci_acl_disconn(conn, 0x13); break; diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c index 69b2c1aac08a..f5b21cb93699 100644 --- a/net/bluetooth/hci_core.c +++ b/net/bluetooth/hci_core.c @@ -1283,9 +1283,12 @@ static inline struct hci_conn *hci_low_sent(struct hci_dev *hdev, __u8 type, int struct hci_conn *c; c = list_entry(p, struct hci_conn, list); - if (c->type != type || c->state != BT_CONNECTED - || skb_queue_empty(&c->data_q)) + if (c->type != type || skb_queue_empty(&c->data_q)) continue; + + if (c->state != BT_CONNECTED && c->state != BT_CONFIG) + continue; + num++; if (c->sent < min) { diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index c8fda7dc2986..e3e360c3c536 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -624,6 +624,62 @@ static void hci_cs_remote_name_req(struct hci_dev *hdev, __u8 status) BT_DBG("%s status 0x%x", hdev->name, status); } +static void hci_cs_read_remote_features(struct hci_dev *hdev, __u8 status) +{ + struct hci_cp_read_remote_features *cp; + struct hci_conn *conn; + + BT_DBG("%s status 0x%x", hdev->name, status); + + if (!status) + return; + + cp = hci_sent_cmd_data(hdev, HCI_OP_READ_REMOTE_FEATURES); + if (!cp) + return; + + hci_dev_lock(hdev); + + conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(cp->handle)); + if (conn) { + if (conn->state == BT_CONFIG) { + conn->state = BT_CONNECTED; + hci_proto_connect_cfm(conn, status); + hci_conn_put(conn); + } + } + + hci_dev_unlock(hdev); +} + +static void hci_cs_read_remote_ext_features(struct hci_dev *hdev, __u8 status) +{ + struct hci_cp_read_remote_ext_features *cp; + struct hci_conn *conn; + + BT_DBG("%s status 0x%x", hdev->name, status); + + if (!status) + return; + + cp = hci_sent_cmd_data(hdev, HCI_OP_READ_REMOTE_EXT_FEATURES); + if (!cp) + return; + + hci_dev_lock(hdev); + + conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(cp->handle)); + if (conn) { + if (conn->state == BT_CONFIG) { + conn->state = BT_CONNECTED; + hci_proto_connect_cfm(conn, status); + hci_conn_put(conn); + } + } + + hci_dev_unlock(hdev); +} + static void hci_cs_setup_sync_conn(struct hci_dev *hdev, __u8 status) { struct hci_cp_setup_sync_conn *cp; @@ -759,7 +815,12 @@ static inline void hci_conn_complete_evt(struct hci_dev *hdev, struct sk_buff *s if (!ev->status) { conn->handle = __le16_to_cpu(ev->handle); - conn->state = BT_CONNECTED; + + if (conn->type == ACL_LINK) { + conn->state = BT_CONFIG; + hci_conn_hold(conn); + } else + conn->state = BT_CONNECTED; if (test_bit(HCI_AUTH, &hdev->flags)) conn->link_mode |= HCI_LM_AUTH; @@ -771,7 +832,8 @@ static inline void hci_conn_complete_evt(struct hci_dev *hdev, struct sk_buff *s if (conn->type == ACL_LINK) { struct hci_cp_read_remote_features cp; cp.handle = ev->handle; - hci_send_cmd(hdev, HCI_OP_READ_REMOTE_FEATURES, sizeof(cp), &cp); + hci_send_cmd(hdev, HCI_OP_READ_REMOTE_FEATURES, + sizeof(cp), &cp); } /* Set packet type for incoming connection */ @@ -781,10 +843,6 @@ static inline void hci_conn_complete_evt(struct hci_dev *hdev, struct sk_buff *s cp.pkt_type = cpu_to_le16(conn->pkt_type); hci_send_cmd(hdev, HCI_OP_CHANGE_CONN_PTYPE, sizeof(cp), &cp); - } else { - /* Update disconnect timer */ - hci_conn_hold(conn); - hci_conn_put(conn); } } else conn->state = BT_CLOSED; @@ -804,9 +862,10 @@ static inline void hci_conn_complete_evt(struct hci_dev *hdev, struct sk_buff *s } } - hci_proto_connect_cfm(conn, ev->status); - if (ev->status) + if (ev->status) { + hci_proto_connect_cfm(conn, ev->status); hci_conn_del(conn); + } unlock: hci_dev_unlock(hdev); @@ -1006,14 +1065,29 @@ static inline void hci_remote_features_evt(struct hci_dev *hdev, struct sk_buff BT_DBG("%s status %d", hdev->name, ev->status); - if (ev->status) - return; - hci_dev_lock(hdev); conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(ev->handle)); - if (conn) - memcpy(conn->features, ev->features, 8); + if (conn) { + if (!ev->status) + memcpy(conn->features, ev->features, 8); + + if (conn->state == BT_CONFIG) { + if (!ev->status && lmp_ssp_capable(hdev) && + lmp_ssp_capable(conn)) { + struct hci_cp_read_remote_ext_features cp; + cp.handle = ev->handle; + cp.page = 0x01; + hci_send_cmd(hdev, + HCI_OP_READ_REMOTE_EXT_FEATURES, + sizeof(cp), &cp); + } else { + conn->state = BT_CONNECTED; + hci_proto_connect_cfm(conn, ev->status); + hci_conn_put(conn); + } + } + } hci_dev_unlock(hdev); } @@ -1180,6 +1254,14 @@ static inline void hci_cmd_status_evt(struct hci_dev *hdev, struct sk_buff *skb) hci_cs_remote_name_req(hdev, ev->status); break; + case HCI_OP_READ_REMOTE_FEATURES: + hci_cs_read_remote_features(hdev, ev->status); + break; + + case HCI_OP_READ_REMOTE_EXT_FEATURES: + hci_cs_read_remote_ext_features(hdev, ev->status); + break; + case HCI_OP_SETUP_SYNC_CONN: hci_cs_setup_sync_conn(hdev, ev->status); break; @@ -1422,19 +1504,24 @@ static inline void hci_remote_ext_features_evt(struct hci_dev *hdev, struct sk_b BT_DBG("%s", hdev->name); - if (ev->status || ev->page != 0x01) - return; - hci_dev_lock(hdev); conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(ev->handle)); if (conn) { - struct inquiry_entry *ie; + if (!ev->status && ev->page == 0x01) { + struct inquiry_entry *ie; - if ((ie = hci_inquiry_cache_lookup(hdev, &conn->dst))) - ie->data.ssp_mode = (ev->features[0] & 0x01); + if ((ie = hci_inquiry_cache_lookup(hdev, &conn->dst))) + ie->data.ssp_mode = (ev->features[0] & 0x01); - conn->ssp_mode = (ev->features[0] & 0x01); + conn->ssp_mode = (ev->features[0] & 0x01); + } + + if (conn->state == BT_CONFIG) { + conn->state = BT_CONNECTED; + hci_proto_connect_cfm(conn, ev->status); + hci_conn_put(conn); + } } hci_dev_unlock(hdev); -- cgit v1.2.2 From f8558555f31e177e2644f3c8116801c3e5c29974 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:49 +0200 Subject: [Bluetooth] Initiate authentication during connection establishment With Bluetooth 2.1 and Simple Pairing the requirement is that any new connection needs to be authenticated and that encryption has been switched on before allowing L2CAP to use it. So make sure that all the requirements are fulfilled and otherwise drop the connection with a minimal disconnect timeout of 10 milliseconds. This change only affects Bluetooth 2.1 devices and Simple Pairing needs to be enabled locally and in the remote host stack. The previous changes made sure that these information are discovered before any kind of authentication and encryption is triggered. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_event.c | 114 +++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 103 insertions(+), 11 deletions(-) (limited to 'net') diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index e3e360c3c536..64668e2656a5 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -619,6 +619,60 @@ static void hci_cs_add_sco(struct hci_dev *hdev, __u8 status) hci_dev_unlock(hdev); } +static void hci_cs_auth_requested(struct hci_dev *hdev, __u8 status) +{ + struct hci_cp_auth_requested *cp; + struct hci_conn *conn; + + BT_DBG("%s status 0x%x", hdev->name, status); + + if (!status) + return; + + cp = hci_sent_cmd_data(hdev, HCI_OP_AUTH_REQUESTED); + if (!cp) + return; + + hci_dev_lock(hdev); + + conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(cp->handle)); + if (conn) { + if (conn->state == BT_CONFIG) { + hci_proto_connect_cfm(conn, status); + hci_conn_put(conn); + } + } + + hci_dev_unlock(hdev); +} + +static void hci_cs_set_conn_encrypt(struct hci_dev *hdev, __u8 status) +{ + struct hci_cp_set_conn_encrypt *cp; + struct hci_conn *conn; + + BT_DBG("%s status 0x%x", hdev->name, status); + + if (!status) + return; + + cp = hci_sent_cmd_data(hdev, HCI_OP_SET_CONN_ENCRYPT); + if (!cp) + return; + + hci_dev_lock(hdev); + + conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(cp->handle)); + if (conn) { + if (conn->state == BT_CONFIG) { + hci_proto_connect_cfm(conn, status); + hci_conn_put(conn); + } + } + + hci_dev_unlock(hdev); +} + static void hci_cs_remote_name_req(struct hci_dev *hdev, __u8 status) { BT_DBG("%s status 0x%x", hdev->name, status); @@ -643,7 +697,6 @@ static void hci_cs_read_remote_features(struct hci_dev *hdev, __u8 status) conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(cp->handle)); if (conn) { if (conn->state == BT_CONFIG) { - conn->state = BT_CONNECTED; hci_proto_connect_cfm(conn, status); hci_conn_put(conn); } @@ -671,7 +724,6 @@ static void hci_cs_read_remote_ext_features(struct hci_dev *hdev, __u8 status) conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(cp->handle)); if (conn) { if (conn->state == BT_CONFIG) { - conn->state = BT_CONNECTED; hci_proto_connect_cfm(conn, status); hci_conn_put(conn); } @@ -982,15 +1034,29 @@ static inline void hci_auth_complete_evt(struct hci_dev *hdev, struct sk_buff *s clear_bit(HCI_CONN_AUTH_PEND, &conn->pend); - hci_auth_cfm(conn, ev->status); + if (conn->state == BT_CONFIG) { + if (!ev->status && hdev->ssp_mode > 0 && + conn->ssp_mode > 0) { + struct hci_cp_set_conn_encrypt cp; + cp.handle = ev->handle; + cp.encrypt = 0x01; + hci_send_cmd(hdev, HCI_OP_SET_CONN_ENCRYPT, + sizeof(cp), &cp); + } else { + conn->state = BT_CONNECTED; + hci_proto_connect_cfm(conn, ev->status); + hci_conn_put(conn); + } + } else + hci_auth_cfm(conn, ev->status); if (test_bit(HCI_CONN_ENCRYPT_PEND, &conn->pend)) { if (!ev->status) { struct hci_cp_set_conn_encrypt cp; - cp.handle = cpu_to_le16(conn->handle); - cp.encrypt = 1; - hci_send_cmd(conn->hdev, - HCI_OP_SET_CONN_ENCRYPT, sizeof(cp), &cp); + cp.handle = ev->handle; + cp.encrypt = 0x01; + hci_send_cmd(hdev, HCI_OP_SET_CONN_ENCRYPT, + sizeof(cp), &cp); } else { clear_bit(HCI_CONN_ENCRYPT_PEND, &conn->pend); hci_encrypt_cfm(conn, ev->status, 0x00); @@ -1030,7 +1096,14 @@ static inline void hci_encrypt_change_evt(struct hci_dev *hdev, struct sk_buff * clear_bit(HCI_CONN_ENCRYPT_PEND, &conn->pend); - hci_encrypt_cfm(conn, ev->status, ev->encrypt); + if (conn->state == BT_CONFIG) { + if (!ev->status) + conn->state = BT_CONNECTED; + + hci_proto_connect_cfm(conn, ev->status); + hci_conn_put(conn); + } else + hci_encrypt_cfm(conn, ev->status, ev->encrypt); } hci_dev_unlock(hdev); @@ -1250,6 +1323,14 @@ static inline void hci_cmd_status_evt(struct hci_dev *hdev, struct sk_buff *skb) hci_cs_add_sco(hdev, ev->status); break; + case HCI_OP_AUTH_REQUESTED: + hci_cs_auth_requested(hdev, ev->status); + break; + + case HCI_OP_SET_CONN_ENCRYPT: + hci_cs_set_conn_encrypt(hdev, ev->status); + break; + case HCI_OP_REMOTE_NAME_REQ: hci_cs_remote_name_req(hdev, ev->status); break; @@ -1518,9 +1599,20 @@ static inline void hci_remote_ext_features_evt(struct hci_dev *hdev, struct sk_b } if (conn->state == BT_CONFIG) { - conn->state = BT_CONNECTED; - hci_proto_connect_cfm(conn, ev->status); - hci_conn_put(conn); + if (!ev->status && hdev->ssp_mode > 0 && + conn->ssp_mode > 0) { + if (conn->out) { + struct hci_cp_auth_requested cp; + cp.handle = ev->handle; + hci_send_cmd(hdev, + HCI_OP_AUTH_REQUESTED, + sizeof(cp), &cp); + } + } else { + conn->state = BT_CONNECTED; + hci_proto_connect_cfm(conn, ev->status); + hci_conn_put(conn); + } } } -- cgit v1.2.2 From 40be492fe4fab829951681860c2bb26fa1d5fe4a Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:50 +0200 Subject: [Bluetooth] Export details about authentication requirements With the Simple Pairing support, the authentication requirements are an explicit setting during the bonding process. Track and enforce the requirements and allow higher layers like L2CAP and RFCOMM to increase them if needed. This patch introduces a new IOCTL that allows to query the current authentication requirements. It is also possible to detect Simple Pairing support in the kernel this way. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_conn.c | 38 ++++++++++++++++++++++++++++++++++---- net/bluetooth/hci_sock.c | 18 +++++------------- net/bluetooth/l2cap.c | 14 ++++++++++---- net/bluetooth/rfcomm/core.c | 3 ++- 4 files changed, 51 insertions(+), 22 deletions(-) (limited to 'net') diff --git a/net/bluetooth/hci_conn.c b/net/bluetooth/hci_conn.c index 6f22533e7656..0d4b8aeb8e09 100644 --- a/net/bluetooth/hci_conn.c +++ b/net/bluetooth/hci_conn.c @@ -379,13 +379,21 @@ int hci_conn_auth(struct hci_conn *conn) { BT_DBG("conn %p", conn); + if (conn->ssp_mode > 0 && conn->hdev->ssp_mode > 0) { + if (!(conn->auth_type & 0x01)) { + conn->auth_type = HCI_AT_GENERAL_BONDING_MITM; + conn->link_mode &= ~HCI_LM_AUTH; + } + } + if (conn->link_mode & HCI_LM_AUTH) return 1; if (!test_and_set_bit(HCI_CONN_AUTH_PEND, &conn->pend)) { struct hci_cp_auth_requested cp; cp.handle = cpu_to_le16(conn->handle); - hci_send_cmd(conn->hdev, HCI_OP_AUTH_REQUESTED, sizeof(cp), &cp); + hci_send_cmd(conn->hdev, HCI_OP_AUTH_REQUESTED, + sizeof(cp), &cp); } return 0; } @@ -397,7 +405,7 @@ int hci_conn_encrypt(struct hci_conn *conn) BT_DBG("conn %p", conn); if (conn->link_mode & HCI_LM_ENCRYPT) - return 1; + return hci_conn_auth(conn); if (test_and_set_bit(HCI_CONN_ENCRYPT_PEND, &conn->pend)) return 0; @@ -406,7 +414,8 @@ int hci_conn_encrypt(struct hci_conn *conn) struct hci_cp_set_conn_encrypt cp; cp.handle = cpu_to_le16(conn->handle); cp.encrypt = 1; - hci_send_cmd(conn->hdev, HCI_OP_SET_CONN_ENCRYPT, sizeof(cp), &cp); + hci_send_cmd(conn->hdev, HCI_OP_SET_CONN_ENCRYPT, + sizeof(cp), &cp); } return 0; } @@ -420,7 +429,8 @@ int hci_conn_change_link_key(struct hci_conn *conn) if (!test_and_set_bit(HCI_CONN_AUTH_PEND, &conn->pend)) { struct hci_cp_change_conn_link_key cp; cp.handle = cpu_to_le16(conn->handle); - hci_send_cmd(conn->hdev, HCI_OP_CHANGE_CONN_LINK_KEY, sizeof(cp), &cp); + hci_send_cmd(conn->hdev, HCI_OP_CHANGE_CONN_LINK_KEY, + sizeof(cp), &cp); } return 0; } @@ -624,3 +634,23 @@ int hci_get_conn_info(struct hci_dev *hdev, void __user *arg) return copy_to_user(ptr, &ci, sizeof(ci)) ? -EFAULT : 0; } + +int hci_get_auth_info(struct hci_dev *hdev, void __user *arg) +{ + struct hci_auth_info_req req; + struct hci_conn *conn; + + if (copy_from_user(&req, arg, sizeof(req))) + return -EFAULT; + + hci_dev_lock_bh(hdev); + conn = hci_conn_hash_lookup_ba(hdev, ACL_LINK, &req.bdaddr); + if (conn) + req.type = conn->auth_type; + hci_dev_unlock_bh(hdev); + + if (!conn) + return -ENOENT; + + return copy_to_user(arg, &req, sizeof(req)) ? -EFAULT : 0; +} diff --git a/net/bluetooth/hci_sock.c b/net/bluetooth/hci_sock.c index 747fabd735d2..d62579b67959 100644 --- a/net/bluetooth/hci_sock.c +++ b/net/bluetooth/hci_sock.c @@ -193,19 +193,11 @@ static inline int hci_sock_bound_ioctl(struct sock *sk, unsigned int cmd, unsign return 0; - case HCISETSECMGR: - if (!capable(CAP_NET_ADMIN)) - return -EACCES; - - if (arg) - set_bit(HCI_SECMGR, &hdev->flags); - else - clear_bit(HCI_SECMGR, &hdev->flags); - - return 0; - case HCIGETCONNINFO: - return hci_get_conn_info(hdev, (void __user *)arg); + return hci_get_conn_info(hdev, (void __user *) arg); + + case HCIGETAUTHINFO: + return hci_get_auth_info(hdev, (void __user *) arg); default: if (hdev->ioctl) @@ -217,7 +209,7 @@ static inline int hci_sock_bound_ioctl(struct sock *sk, unsigned int cmd, unsign static int hci_sock_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg) { struct sock *sk = sock->sk; - void __user *argp = (void __user *)arg; + void __user *argp = (void __user *) arg; int err; BT_DBG("cmd %x arg %lx", cmd, arg); diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c index 252264062f59..30ad59b717d5 100644 --- a/net/bluetooth/l2cap.c +++ b/net/bluetooth/l2cap.c @@ -2150,7 +2150,7 @@ static int l2cap_disconn_ind(struct hci_conn *hcon, u8 reason) static int l2cap_auth_cfm(struct hci_conn *hcon, u8 status) { struct l2cap_chan_list *l; - struct l2cap_conn *conn = conn = hcon->l2cap_data; + struct l2cap_conn *conn = hcon->l2cap_data; struct l2cap_conn_rsp rsp; struct sock *sk; int result; @@ -2165,11 +2165,17 @@ static int l2cap_auth_cfm(struct hci_conn *hcon, u8 status) read_lock(&l->lock); for (sk = l->head; sk; sk = l2cap_pi(sk)->next_c) { + struct l2cap_pinfo *pi = l2cap_pi(sk); + bh_lock_sock(sk); - if (sk->sk_state != BT_CONNECT2 || - (l2cap_pi(sk)->link_mode & L2CAP_LM_ENCRYPT) || - (l2cap_pi(sk)->link_mode & L2CAP_LM_SECURE)) { + if (sk->sk_state != BT_CONNECT2) { + bh_unlock_sock(sk); + continue; + } + + if ((pi->link_mode & (L2CAP_LM_ENCRYPT | L2CAP_LM_SECURE)) && + !(hcon->link_mode & HCI_LM_ENCRYPT)) { bh_unlock_sock(sk); continue; } diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c index e7a6a03cea37..e56bcfc35a49 100644 --- a/net/bluetooth/rfcomm/core.c +++ b/net/bluetooth/rfcomm/core.c @@ -1969,7 +1969,8 @@ static void rfcomm_auth_cfm(struct hci_conn *conn, u8 status) list_for_each_safe(p, n, &s->dlcs) { d = list_entry(p, struct rfcomm_dlc, list); - if (d->link_mode & (RFCOMM_LM_ENCRYPT | RFCOMM_LM_SECURE)) + if ((d->link_mode & (RFCOMM_LM_ENCRYPT | RFCOMM_LM_SECURE)) && + !(conn->link_mode & HCI_LM_ENCRYPT) && !status) continue; if (!test_and_clear_bit(RFCOMM_AUTH_PENDING, &d->flags)) -- cgit v1.2.2 From 3241ad820dbb172021e0268b5611031991431626 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:50 +0200 Subject: [Bluetooth] Add timestamp support to L2CAP, RFCOMM and SCO Enable the common timestamp functionality that the network subsystem provides for L2CAP, RFCOMM and SCO sockets. It is possible to either use SO_TIMESTAMP or the IOCTLs to retrieve the timestamp of the current packet. Signed-off-by: Marcel Holtmann --- net/bluetooth/af_bluetooth.c | 28 ++++++++++++++++++++++++++++ net/bluetooth/l2cap.c | 2 +- net/bluetooth/rfcomm/sock.c | 15 +++++++++++---- net/bluetooth/sco.c | 2 +- 4 files changed, 41 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/bluetooth/af_bluetooth.c b/net/bluetooth/af_bluetooth.c index d366423c8392..88afe25003d8 100644 --- a/net/bluetooth/af_bluetooth.c +++ b/net/bluetooth/af_bluetooth.c @@ -36,6 +36,7 @@ #include #include #include +#include #if defined(CONFIG_KMOD) #include @@ -266,6 +267,8 @@ int bt_sock_recvmsg(struct kiocb *iocb, struct socket *sock, skb_reset_transport_header(skb); err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied); + if (err == 0) + sock_recv_timestamp(msg, sk, skb); skb_free_datagram(sk, skb); @@ -329,6 +332,31 @@ unsigned int bt_sock_poll(struct file * file, struct socket *sock, poll_table *w } EXPORT_SYMBOL(bt_sock_poll); +int bt_sock_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg) +{ + struct sock *sk = sock->sk; + int err; + + BT_DBG("sk %p cmd %x arg %lx", sk, cmd, arg); + + switch (cmd) { + case SIOCGSTAMP: + err = sock_get_timestamp(sk, (struct timeval __user *) arg); + break; + + case SIOCGSTAMPNS: + err = sock_get_timestampns(sk, (struct timespec __user *) arg); + break; + + default: + err = -ENOIOCTLCMD; + break; + } + + return err; +} +EXPORT_SYMBOL(bt_sock_ioctl); + int bt_sock_wait_state(struct sock *sk, int state, unsigned long timeo) { DECLARE_WAITQUEUE(wait, current); diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c index 30ad59b717d5..4fcf24af7590 100644 --- a/net/bluetooth/l2cap.c +++ b/net/bluetooth/l2cap.c @@ -2388,9 +2388,9 @@ static const struct proto_ops l2cap_sock_ops = { .sendmsg = l2cap_sock_sendmsg, .recvmsg = bt_sock_recvmsg, .poll = bt_sock_poll, + .ioctl = bt_sock_ioctl, .mmap = sock_no_mmap, .socketpair = sock_no_socketpair, - .ioctl = sock_no_ioctl, .shutdown = l2cap_sock_shutdown, .setsockopt = l2cap_sock_setsockopt, .getsockopt = l2cap_sock_getsockopt diff --git a/net/bluetooth/rfcomm/sock.c b/net/bluetooth/rfcomm/sock.c index cacb1ab51f99..c3ed076481d8 100644 --- a/net/bluetooth/rfcomm/sock.c +++ b/net/bluetooth/rfcomm/sock.c @@ -690,6 +690,8 @@ static int rfcomm_sock_recvmsg(struct kiocb *iocb, struct socket *sock, copied += chunk; size -= chunk; + sock_recv_timestamp(msg, sk, skb); + if (!(flags & MSG_PEEK)) { atomic_sub(chunk, &sk->sk_rmem_alloc); @@ -795,15 +797,20 @@ static int rfcomm_sock_ioctl(struct socket *sock, unsigned int cmd, unsigned lon struct sock *sk = sock->sk; int err; - lock_sock(sk); + BT_DBG("sk %p cmd %x arg %lx", sk, cmd, arg); + err = bt_sock_ioctl(sock, cmd, arg); + + if (err == -ENOIOCTLCMD) { #ifdef CONFIG_BT_RFCOMM_TTY - err = rfcomm_dev_ioctl(sk, cmd, (void __user *)arg); + lock_sock(sk); + err = rfcomm_dev_ioctl(sk, cmd, (void __user *) arg); + release_sock(sk); #else - err = -EOPNOTSUPP; + err = -EOPNOTSUPP; #endif + } - release_sock(sk); return err; } diff --git a/net/bluetooth/sco.c b/net/bluetooth/sco.c index b0d487e2db20..1ad226c9788c 100644 --- a/net/bluetooth/sco.c +++ b/net/bluetooth/sco.c @@ -921,7 +921,7 @@ static const struct proto_ops sco_sock_ops = { .sendmsg = sco_sock_sendmsg, .recvmsg = bt_sock_recvmsg, .poll = bt_sock_poll, - .ioctl = sock_no_ioctl, + .ioctl = bt_sock_ioctl, .mmap = sock_no_mmap, .socketpair = sock_no_socketpair, .shutdown = sock_no_shutdown, -- cgit v1.2.2 From 43cbeee9f9b26300275e4e2d55ed1607f8c5f760 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:51 +0200 Subject: [Bluetooth] Add support for TIOCOUTQ and TIOCINQ ioctls Almost every protocol family supports the TIOCOUTQ and TIOCINQ ioctls and even Bluetooth could make use of them. When implementing audio streaming and integration with GStreamer or PulseAudio they will allow a better timing and synchronization. Signed-off-by: Marcel Holtmann --- net/bluetooth/af_bluetooth.c | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) (limited to 'net') diff --git a/net/bluetooth/af_bluetooth.c b/net/bluetooth/af_bluetooth.c index 88afe25003d8..77fe9b868fee 100644 --- a/net/bluetooth/af_bluetooth.c +++ b/net/bluetooth/af_bluetooth.c @@ -335,11 +335,34 @@ EXPORT_SYMBOL(bt_sock_poll); int bt_sock_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg) { struct sock *sk = sock->sk; + struct sk_buff *skb; + long amount; int err; BT_DBG("sk %p cmd %x arg %lx", sk, cmd, arg); switch (cmd) { + case TIOCOUTQ: + if (sk->sk_state == BT_LISTEN) + return -EINVAL; + + amount = sk->sk_sndbuf - atomic_read(&sk->sk_wmem_alloc); + if (amount < 0) + amount = 0; + err = put_user(amount, (int __user *) arg); + break; + + case TIOCINQ: + if (sk->sk_state == BT_LISTEN) + return -EINVAL; + + lock_sock(sk); + skb = skb_peek(&sk->sk_receive_queue); + amount = skb ? skb->len : 0; + release_sock(sk); + err = put_user(amount, (int __user *) arg); + break; + case SIOCGSTAMP: err = sock_get_timestamp(sk, (struct timeval __user *) arg); break; -- cgit v1.2.2 From 7d0db0a373195385a2e0b19d1f5e4b186fdcffac Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:51 +0200 Subject: [Bluetooth] Use a more unique bus name for connections When attaching Bluetooth low-level connections to the bus, the bus name is constructed from the remote address since at that time the connection handle is not assigned yet. This has worked so far, but also caused a lot of troubles. It is better to postpone the creation of the sysfs entry to the time when the connection actually has been established and then use its connection handle as unique identifier. This also fixes the case where two different adapters try to connect to the same remote device. Signed-off-by: Marcel Holtmann --- net/bluetooth/hci_conn.c | 8 +++++--- net/bluetooth/hci_event.c | 7 +++++++ net/bluetooth/hci_sysfs.c | 8 ++------ 3 files changed, 14 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/bluetooth/hci_conn.c b/net/bluetooth/hci_conn.c index 0d4b8aeb8e09..ca8d05245ca0 100644 --- a/net/bluetooth/hci_conn.c +++ b/net/bluetooth/hci_conn.c @@ -245,8 +245,6 @@ struct hci_conn *hci_conn_add(struct hci_dev *hdev, int type, bdaddr_t *dst) if (hdev->notify) hdev->notify(hdev, HCI_NOTIFY_CONN_ADD); - hci_conn_add_sysfs(conn); - tasklet_enable(&hdev->tx_task); return conn; @@ -278,12 +276,14 @@ int hci_conn_del(struct hci_conn *conn) } tasklet_disable(&hdev->tx_task); + hci_conn_hash_del(hdev, conn); if (hdev->notify) hdev->notify(hdev, HCI_NOTIFY_CONN_DEL); + tasklet_enable(&hdev->tx_task); + skb_queue_purge(&conn->data_q); - hci_conn_del_sysfs(conn); return 0; } @@ -532,6 +532,8 @@ void hci_conn_hash_flush(struct hci_dev *hdev) c->state = BT_CLOSED; + hci_conn_del_sysfs(c); + hci_proto_disconn_ind(c, 0x16); hci_conn_del(c); } diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index 64668e2656a5..0e3db289f4be 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -874,6 +874,8 @@ static inline void hci_conn_complete_evt(struct hci_dev *hdev, struct sk_buff *s } else conn->state = BT_CONNECTED; + hci_conn_add_sysfs(conn); + if (test_bit(HCI_AUTH, &hdev->flags)) conn->link_mode |= HCI_LM_AUTH; @@ -1011,6 +1013,9 @@ static inline void hci_disconn_complete_evt(struct hci_dev *hdev, struct sk_buff conn = hci_conn_hash_lookup_handle(hdev, __le16_to_cpu(ev->handle)); if (conn) { conn->state = BT_CLOSED; + + hci_conn_del_sysfs(conn); + hci_proto_disconn_ind(conn, ev->reason); hci_conn_del(conn); } @@ -1643,6 +1648,8 @@ static inline void hci_sync_conn_complete_evt(struct hci_dev *hdev, struct sk_bu if (!ev->status) { conn->handle = __le16_to_cpu(ev->handle); conn->state = BT_CONNECTED; + + hci_conn_add_sysfs(conn); } else conn->state = BT_CLOSED; diff --git a/net/bluetooth/hci_sysfs.c b/net/bluetooth/hci_sysfs.c index a18871e01582..844ca5f1b2d4 100644 --- a/net/bluetooth/hci_sysfs.c +++ b/net/bluetooth/hci_sysfs.c @@ -311,7 +311,6 @@ static void add_conn(struct work_struct *work) void hci_conn_add_sysfs(struct hci_conn *conn) { struct hci_dev *hdev = conn->hdev; - bdaddr_t *ba = &conn->dst; BT_DBG("conn %p", conn); @@ -320,11 +319,8 @@ void hci_conn_add_sysfs(struct hci_conn *conn) conn->dev.release = bt_release; - snprintf(conn->dev.bus_id, BUS_ID_SIZE, - "%s%2.2X%2.2X%2.2X%2.2X%2.2X%2.2X", - conn->type == ACL_LINK ? "acl" : "sco", - ba->b[5], ba->b[4], ba->b[3], - ba->b[2], ba->b[1], ba->b[0]); + snprintf(conn->dev.bus_id, BUS_ID_SIZE, "%s:%d", + hdev->name, conn->handle); dev_set_drvdata(&conn->dev, conn); -- cgit v1.2.2 From 78c6a1744fafc3a396f85c1970eb78839b55af74 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:51 +0200 Subject: [Bluetooth] Update Bluetooth core version number With all the Bluetooth 2.1 changes and the support for Simple Pairing, it is important to update the Bluetooth core version number. Signed-off-by: Marcel Holtmann --- net/bluetooth/af_bluetooth.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/bluetooth/af_bluetooth.c b/net/bluetooth/af_bluetooth.c index 77fe9b868fee..4e59df5f8e05 100644 --- a/net/bluetooth/af_bluetooth.c +++ b/net/bluetooth/af_bluetooth.c @@ -49,7 +49,7 @@ #define BT_DBG(D...) #endif -#define VERSION "2.11" +#define VERSION "2.12" /* Bluetooth sockets */ #define BT_MAX_PROTO 8 -- cgit v1.2.2 From ca37bdd53b5af06d00e792f2415b93206aa2a541 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:52 +0200 Subject: [Bluetooth] Use non-canonical TTY by default for RFCOMM While the RFCOMM TTY emulation can act like a real serial port, in reality it is not used like this. So to not mess up stupid applications, use the non-canonical mode by default. Signed-off-by: Denis Kenzior Signed-off-by: Marcel Holtmann --- net/bluetooth/rfcomm/tty.c | 1 + 1 file changed, 1 insertion(+) (limited to 'net') diff --git a/net/bluetooth/rfcomm/tty.c b/net/bluetooth/rfcomm/tty.c index c9191871c1e0..248802796e12 100644 --- a/net/bluetooth/rfcomm/tty.c +++ b/net/bluetooth/rfcomm/tty.c @@ -1123,6 +1123,7 @@ int rfcomm_init_ttys(void) rfcomm_tty_driver->flags = TTY_DRIVER_REAL_RAW | TTY_DRIVER_DYNAMIC_DEV; rfcomm_tty_driver->init_termios = tty_std_termios; rfcomm_tty_driver->init_termios.c_cflag = B9600 | CS8 | CREAD | HUPCL | CLOCAL; + rfcomm_tty_driver->init_termios.c_lflag &= ~ICANON; tty_set_operations(rfcomm_tty_driver, &rfcomm_ops); if (tty_register_driver(rfcomm_tty_driver)) { -- cgit v1.2.2 From 8b6b3da765af9600b5edd8e3e84a20523e975884 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:52 +0200 Subject: [Bluetooth] Store remote modem status for RFCOMM TTY When switching a RFCOMM socket to a TTY, the remote modem status might be needed later. Currently it is lost since the original configuration is done via the socket interface. So store the modem status and reply it when the socket has been converted to a TTY. Signed-off-by: Denis Kenzior Signed-off-by: Marcel Holtmann --- net/bluetooth/rfcomm/core.c | 4 ++++ net/bluetooth/rfcomm/tty.c | 3 +++ 2 files changed, 7 insertions(+) (limited to 'net') diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c index e56bcfc35a49..fcd2cafe70c8 100644 --- a/net/bluetooth/rfcomm/core.c +++ b/net/bluetooth/rfcomm/core.c @@ -1463,8 +1463,12 @@ static int rfcomm_recv_msc(struct rfcomm_session *s, int cr, struct sk_buff *skb clear_bit(RFCOMM_TX_THROTTLED, &d->flags); rfcomm_dlc_lock(d); + + d->remote_v24_sig = msc->v24_sig; + if (d->modem_status) d->modem_status(d, msc->v24_sig); + rfcomm_dlc_unlock(d); rfcomm_send_msc(s, 0, dlci, msc->v24_sig); diff --git a/net/bluetooth/rfcomm/tty.c b/net/bluetooth/rfcomm/tty.c index 248802796e12..8fcca08cef8e 100644 --- a/net/bluetooth/rfcomm/tty.c +++ b/net/bluetooth/rfcomm/tty.c @@ -271,6 +271,9 @@ static int rfcomm_dev_add(struct rfcomm_dev_req *req, struct rfcomm_dlc *dlc) dlc->owner = dev; dev->dlc = dlc; + + rfcomm_dev_modem_status(dlc, dlc->remote_v24_sig); + rfcomm_dlc_unlock(dlc); /* It's safe to call __module_get() here because socket already -- cgit v1.2.2 From a0c22f226502be6eab37a1d9bf6fb0fadf551376 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:52 +0200 Subject: [Bluetooth] Move pending packets from RFCOMM socket to TTY When an incoming RFCOMM socket connection gets converted into a TTY, it can happen that packets are lost. This mainly happens with the Handsfree profile where the remote side starts sending data right away. The problem is that these packets are in the socket receive queue. So when creating the TTY make sure to copy all pending packets from the socket receive queue to a private queue inside the TTY. To make this actually work, the flow control on the newly created TTY will be disabled and only enabled again when the TTY is opened by an application. And right before that, the pending packets will be put into the TTY flip buffer. Signed-off-by: Denis Kenzior Signed-off-by: Marcel Holtmann --- net/bluetooth/rfcomm/core.c | 2 +- net/bluetooth/rfcomm/tty.c | 55 ++++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 55 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c index fcd2cafe70c8..b6b3d9b4066f 100644 --- a/net/bluetooth/rfcomm/core.c +++ b/net/bluetooth/rfcomm/core.c @@ -53,7 +53,7 @@ #define BT_DBG(D...) #endif -#define VERSION "1.9" +#define VERSION "1.10" static int disable_cfc = 0; static int channel_mtu = -1; diff --git a/net/bluetooth/rfcomm/tty.c b/net/bluetooth/rfcomm/tty.c index 8fcca08cef8e..ec22ebe0c2c8 100644 --- a/net/bluetooth/rfcomm/tty.c +++ b/net/bluetooth/rfcomm/tty.c @@ -77,6 +77,8 @@ struct rfcomm_dev { struct device *tty_dev; atomic_t wmem_alloc; + + struct sk_buff_head pending; }; static LIST_HEAD(rfcomm_dev_list); @@ -264,7 +266,25 @@ static int rfcomm_dev_add(struct rfcomm_dev_req *req, struct rfcomm_dlc *dlc) init_waitqueue_head(&dev->wait); tasklet_init(&dev->wakeup_task, rfcomm_tty_wakeup, (unsigned long) dev); + skb_queue_head_init(&dev->pending); + rfcomm_dlc_lock(dlc); + + if (req->flags & (1 << RFCOMM_REUSE_DLC)) { + struct sock *sk = dlc->owner; + struct sk_buff *skb; + + BUG_ON(!sk); + + rfcomm_dlc_throttle(dlc); + + while ((skb = skb_dequeue(&sk->sk_receive_queue))) { + skb_orphan(skb); + skb_queue_tail(&dev->pending, skb); + atomic_sub(skb->len, &sk->sk_rmem_alloc); + } + } + dlc->data_ready = rfcomm_dev_data_ready; dlc->state_change = rfcomm_dev_state_change; dlc->modem_status = rfcomm_dev_modem_status; @@ -542,11 +562,16 @@ static void rfcomm_dev_data_ready(struct rfcomm_dlc *dlc, struct sk_buff *skb) struct rfcomm_dev *dev = dlc->owner; struct tty_struct *tty; - if (!dev || !(tty = dev->tty)) { + if (!dev) { kfree_skb(skb); return; } + if (!(tty = dev->tty) || !skb_queue_empty(&dev->pending)) { + skb_queue_tail(&dev->pending, skb); + return; + } + BT_DBG("dlc %p tty %p len %d", dlc, tty, skb->len); tty_insert_flip_string(tty, skb->data, skb->len); @@ -630,6 +655,30 @@ static void rfcomm_tty_wakeup(unsigned long arg) #endif } +static void rfcomm_tty_copy_pending(struct rfcomm_dev *dev) +{ + struct tty_struct *tty = dev->tty; + struct sk_buff *skb; + int inserted = 0; + + if (!tty) + return; + + BT_DBG("dev %p tty %p", dev, tty); + + rfcomm_dlc_lock(dev->dlc); + + while ((skb = skb_dequeue(&dev->pending))) { + inserted += tty_insert_flip_string(tty, skb->data, skb->len); + kfree_skb(skb); + } + + rfcomm_dlc_unlock(dev->dlc); + + if (inserted > 0) + tty_flip_buffer_push(tty); +} + static int rfcomm_tty_open(struct tty_struct *tty, struct file *filp) { DECLARE_WAITQUEUE(wait, current); @@ -694,6 +743,10 @@ static int rfcomm_tty_open(struct tty_struct *tty, struct file *filp) if (err == 0) device_move(dev->tty_dev, rfcomm_get_device(dev)); + rfcomm_tty_copy_pending(dev); + + rfcomm_dlc_unthrottle(dev->dlc); + return err; } -- cgit v1.2.2 From ec8dab36e0738d3059980d144e34f16a26bbda7d Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:53 +0200 Subject: [Bluetooth] Signal user-space for HIDP and BNEP socket errors When using the HIDP or BNEP kernel support, the user-space needs to know if the connection has been terminated for some reasons. Wake up the application if that happens. Otherwise kernel and user-space are no longer on the same page and weird behaviors can happen. Signed-off-by: Marcel Holtmann --- net/bluetooth/bnep/core.c | 5 +++++ net/bluetooth/hidp/core.c | 10 ++++++++++ 2 files changed, 15 insertions(+) (limited to 'net') diff --git a/net/bluetooth/bnep/core.c b/net/bluetooth/bnep/core.c index f85d94643aaf..24e91eb7f649 100644 --- a/net/bluetooth/bnep/core.c +++ b/net/bluetooth/bnep/core.c @@ -507,6 +507,11 @@ static int bnep_session(void *arg) /* Delete network device */ unregister_netdev(dev); + /* Wakeup user-space polling for socket errors */ + s->sock->sk->sk_err = EUNATCH; + + wake_up_interruptible(s->sock->sk->sk_sleep); + /* Release the socket */ fput(s->sock->file); diff --git a/net/bluetooth/hidp/core.c b/net/bluetooth/hidp/core.c index 519cdb920f93..96434d774c84 100644 --- a/net/bluetooth/hidp/core.c +++ b/net/bluetooth/hidp/core.c @@ -581,6 +581,12 @@ static int hidp_session(void *arg) hid_free_device(session->hid); } + /* Wakeup user-space polling for socket errors */ + session->intr_sock->sk->sk_err = EUNATCH; + session->ctrl_sock->sk->sk_err = EUNATCH; + + hidp_schedule(session); + fput(session->intr_sock->file); wait_event_timeout(*(ctrl_sk->sk_sleep), @@ -879,6 +885,10 @@ int hidp_del_connection(struct hidp_conndel_req *req) skb_queue_purge(&session->ctrl_transmit); skb_queue_purge(&session->intr_transmit); + /* Wakeup user-space polling for socket errors */ + session->intr_sock->sk->sk_err = EUNATCH; + session->ctrl_sock->sk->sk_err = EUNATCH; + /* Kill session thread */ atomic_inc(&session->terminate); hidp_schedule(session); -- cgit v1.2.2 From 7cb127d5b0e7af7a0afd23785722ca3edab4ceff Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:53 +0200 Subject: [Bluetooth] Add option to disable eSCO connection creation It has been reported that some eSCO capable headsets are not able to connect properly. The real reason for this is unclear at the moment. So for easier testing add a module parameter to disable eSCO connection creation. Signed-off-by: Marcel Holtmann --- net/bluetooth/sco.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/bluetooth/sco.c b/net/bluetooth/sco.c index 1ad226c9788c..8cda49874868 100644 --- a/net/bluetooth/sco.c +++ b/net/bluetooth/sco.c @@ -53,7 +53,9 @@ #define BT_DBG(D...) #endif -#define VERSION "0.5" +#define VERSION "0.6" + +static int disable_esco = 0; static const struct proto_ops sco_sock_ops; @@ -193,7 +195,10 @@ static int sco_connect(struct sock *sk) err = -ENOMEM; - type = lmp_esco_capable(hdev) ? ESCO_LINK : SCO_LINK; + if (lmp_esco_capable(hdev) && !disable_esco) + type = ESCO_LINK; + else + type = SCO_LINK; hcon = hci_connect(hdev, type, dst); if (!hcon) @@ -994,6 +999,9 @@ static void __exit sco_exit(void) module_init(sco_init); module_exit(sco_exit); +module_param(disable_esco, bool, 0644); +MODULE_PARM_DESC(disable_esco, "Disable eSCO connection creation"); + MODULE_AUTHOR("Maxim Krasnyansky , Marcel Holtmann "); MODULE_DESCRIPTION("Bluetooth SCO ver " VERSION); MODULE_VERSION(VERSION); -- cgit v1.2.2 From b1235d79611e78a07629b4cbe53291c9cffd1834 Mon Sep 17 00:00:00 2001 From: Marcel Holtmann Date: Mon, 14 Jul 2008 20:13:54 +0200 Subject: [Bluetooth] Allow security for outgoing L2CAP connections When requested the L2CAP layer will now enforce authentication and encryption on outgoing connections. The usefulness of this feature is kinda limited since it will not allow proper connection ownership tracking until the authentication procedure has been finished. This is a limitation of Bluetooth 2.0 and before and can only be fixed by using Simple Pairing. Signed-off-by: Marcel Holtmann --- net/bluetooth/l2cap.c | 172 +++++++++++++++++++++++++++++++------------------- 1 file changed, 108 insertions(+), 64 deletions(-) (limited to 'net') diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c index 4fcf24af7590..c1239852834a 100644 --- a/net/bluetooth/l2cap.c +++ b/net/bluetooth/l2cap.c @@ -76,11 +76,21 @@ static struct sk_buff *l2cap_build_cmd(struct l2cap_conn *conn, static void l2cap_sock_timeout(unsigned long arg) { struct sock *sk = (struct sock *) arg; + int reason; BT_DBG("sock %p state %d", sk, sk->sk_state); bh_lock_sock(sk); - __l2cap_sock_close(sk, ETIMEDOUT); + + if (sk->sk_state == BT_CONNECT && + (l2cap_pi(sk)->link_mode & (L2CAP_LM_AUTH | + L2CAP_LM_ENCRYPT | L2CAP_LM_SECURE))) + reason = ECONNREFUSED; + else + reason = ETIMEDOUT; + + __l2cap_sock_close(sk, reason); + bh_unlock_sock(sk); l2cap_sock_kill(sk); @@ -240,7 +250,7 @@ static void l2cap_chan_del(struct sock *sk, int err) hci_conn_put(conn->hcon); } - sk->sk_state = BT_CLOSED; + sk->sk_state = BT_CLOSED; sock_set_flag(sk, SOCK_ZAPPED); if (err) @@ -307,14 +317,16 @@ static void l2cap_do_start(struct sock *sk) struct l2cap_conn *conn = l2cap_pi(sk)->conn; if (conn->info_state & L2CAP_INFO_FEAT_MASK_REQ_SENT) { - struct l2cap_conn_req req; - req.scid = cpu_to_le16(l2cap_pi(sk)->scid); - req.psm = l2cap_pi(sk)->psm; + if (l2cap_check_link_mode(sk)) { + struct l2cap_conn_req req; + req.scid = cpu_to_le16(l2cap_pi(sk)->scid); + req.psm = l2cap_pi(sk)->psm; - l2cap_pi(sk)->ident = l2cap_get_ident(conn); + l2cap_pi(sk)->ident = l2cap_get_ident(conn); - l2cap_send_cmd(conn, l2cap_pi(sk)->ident, + l2cap_send_cmd(conn, l2cap_pi(sk)->ident, L2CAP_CONN_REQ, sizeof(req), &req); + } } else { struct l2cap_info_req req; req.type = cpu_to_le16(L2CAP_IT_FEAT_MASK); @@ -349,14 +361,16 @@ static void l2cap_conn_start(struct l2cap_conn *conn) } if (sk->sk_state == BT_CONNECT) { - struct l2cap_conn_req req; - req.scid = cpu_to_le16(l2cap_pi(sk)->scid); - req.psm = l2cap_pi(sk)->psm; + if (l2cap_check_link_mode(sk)) { + struct l2cap_conn_req req; + req.scid = cpu_to_le16(l2cap_pi(sk)->scid); + req.psm = l2cap_pi(sk)->psm; - l2cap_pi(sk)->ident = l2cap_get_ident(conn); + l2cap_pi(sk)->ident = l2cap_get_ident(conn); - l2cap_send_cmd(conn, l2cap_pi(sk)->ident, + l2cap_send_cmd(conn, l2cap_pi(sk)->ident, L2CAP_CONN_REQ, sizeof(req), &req); + } } else if (sk->sk_state == BT_CONNECT2) { struct l2cap_conn_rsp rsp; rsp.scid = cpu_to_le16(l2cap_pi(sk)->dcid); @@ -455,7 +469,8 @@ static struct l2cap_conn *l2cap_conn_add(struct hci_conn *hcon, u8 status) conn->feat_mask = 0; - setup_timer(&conn->info_timer, l2cap_info_timeout, (unsigned long)conn); + setup_timer(&conn->info_timer, l2cap_info_timeout, + (unsigned long) conn); spin_lock_init(&conn->lock); rwlock_init(&conn->chan_list.lock); @@ -567,7 +582,7 @@ static void l2cap_sock_cleanup_listen(struct sock *parent) while ((sk = bt_accept_dequeue(parent, NULL))) l2cap_sock_close(sk); - parent->sk_state = BT_CLOSED; + parent->sk_state = BT_CLOSED; sock_set_flag(parent, SOCK_ZAPPED); } @@ -610,9 +625,8 @@ static void __l2cap_sock_close(struct sock *sk, int reason) req.scid = cpu_to_le16(l2cap_pi(sk)->scid); l2cap_send_cmd(conn, l2cap_get_ident(conn), L2CAP_DISCONN_REQ, sizeof(req), &req); - } else { + } else l2cap_chan_del(sk, reason); - } break; case BT_CONNECT: @@ -681,9 +695,9 @@ static struct sock *l2cap_sock_alloc(struct net *net, struct socket *sock, int p sock_reset_flag(sk, SOCK_ZAPPED); sk->sk_protocol = proto; - sk->sk_state = BT_OPEN; + sk->sk_state = BT_OPEN; - setup_timer(&sk->sk_timer, l2cap_sock_timeout, (unsigned long)sk); + setup_timer(&sk->sk_timer, l2cap_sock_timeout, (unsigned long) sk); bt_sock_link(&l2cap_sk_list, sk); return sk; @@ -1201,7 +1215,8 @@ static int l2cap_sock_shutdown(struct socket *sock, int how) __l2cap_sock_close(sk, 0); if (sock_flag(sk, SOCK_LINGER) && sk->sk_lingertime) - err = bt_sock_wait_state(sk, BT_CLOSED, sk->sk_lingertime); + err = bt_sock_wait_state(sk, BT_CLOSED, + sk->sk_lingertime); } release_sock(sk); return err; @@ -1245,6 +1260,11 @@ static void l2cap_chan_ready(struct sock *sk) */ parent->sk_data_ready(parent, 0); } + + if (l2cap_pi(sk)->link_mode & L2CAP_LM_SECURE) { + struct l2cap_conn *conn = l2cap_pi(sk)->conn; + hci_conn_change_link_key(conn->hcon); + } } /* Copy frame to all raw sockets on that connection */ @@ -1778,7 +1798,7 @@ static inline int l2cap_config_rsp(struct l2cap_conn *conn, struct l2cap_cmd_hdr default: sk->sk_state = BT_DISCONN; - sk->sk_err = ECONNRESET; + sk->sk_err = ECONNRESET; l2cap_sock_set_timer(sk, HZ * 5); { struct l2cap_disconn_req req; @@ -2151,9 +2171,7 @@ static int l2cap_auth_cfm(struct hci_conn *hcon, u8 status) { struct l2cap_chan_list *l; struct l2cap_conn *conn = hcon->l2cap_data; - struct l2cap_conn_rsp rsp; struct sock *sk; - int result; if (!conn) return 0; @@ -2169,37 +2187,53 @@ static int l2cap_auth_cfm(struct hci_conn *hcon, u8 status) bh_lock_sock(sk); - if (sk->sk_state != BT_CONNECT2) { - bh_unlock_sock(sk); - continue; - } - if ((pi->link_mode & (L2CAP_LM_ENCRYPT | L2CAP_LM_SECURE)) && - !(hcon->link_mode & HCI_LM_ENCRYPT)) { + !(hcon->link_mode & HCI_LM_ENCRYPT) && + !status) { bh_unlock_sock(sk); continue; } - if (!status) { - sk->sk_state = BT_CONFIG; - result = 0; - } else { - sk->sk_state = BT_DISCONN; - l2cap_sock_set_timer(sk, HZ/10); - result = L2CAP_CR_SEC_BLOCK; - } + if (sk->sk_state == BT_CONNECT) { + if (!status) { + struct l2cap_conn_req req; + req.scid = cpu_to_le16(l2cap_pi(sk)->scid); + req.psm = l2cap_pi(sk)->psm; + + l2cap_pi(sk)->ident = l2cap_get_ident(conn); + + l2cap_send_cmd(conn, l2cap_pi(sk)->ident, + L2CAP_CONN_REQ, sizeof(req), &req); + } else { + l2cap_sock_clear_timer(sk); + l2cap_sock_set_timer(sk, HZ / 10); + } + } else if (sk->sk_state == BT_CONNECT2) { + struct l2cap_conn_rsp rsp; + __u16 result; + + if (!status) { + sk->sk_state = BT_CONFIG; + result = L2CAP_CR_SUCCESS; + } else { + sk->sk_state = BT_DISCONN; + l2cap_sock_set_timer(sk, HZ / 10); + result = L2CAP_CR_SEC_BLOCK; + } - rsp.scid = cpu_to_le16(l2cap_pi(sk)->dcid); - rsp.dcid = cpu_to_le16(l2cap_pi(sk)->scid); - rsp.result = cpu_to_le16(result); - rsp.status = cpu_to_le16(0); - l2cap_send_cmd(conn, l2cap_pi(sk)->ident, - L2CAP_CONN_RSP, sizeof(rsp), &rsp); + rsp.scid = cpu_to_le16(l2cap_pi(sk)->dcid); + rsp.dcid = cpu_to_le16(l2cap_pi(sk)->scid); + rsp.result = cpu_to_le16(result); + rsp.status = cpu_to_le16(0); + l2cap_send_cmd(conn, l2cap_pi(sk)->ident, + L2CAP_CONN_RSP, sizeof(rsp), &rsp); + } bh_unlock_sock(sk); } read_unlock(&l->lock); + return 0; } @@ -2207,9 +2241,7 @@ static int l2cap_encrypt_cfm(struct hci_conn *hcon, u8 status, u8 encrypt) { struct l2cap_chan_list *l; struct l2cap_conn *conn = hcon->l2cap_data; - struct l2cap_conn_rsp rsp; struct sock *sk; - int result; if (!conn) return 0; @@ -2234,34 +2266,46 @@ static int l2cap_encrypt_cfm(struct hci_conn *hcon, u8 status, u8 encrypt) continue; } - if (sk->sk_state != BT_CONNECT2) { - bh_unlock_sock(sk); - continue; - } + if (sk->sk_state == BT_CONNECT) { + if (!status) { + struct l2cap_conn_req req; + req.scid = cpu_to_le16(l2cap_pi(sk)->scid); + req.psm = l2cap_pi(sk)->psm; - if (!status) { - sk->sk_state = BT_CONFIG; - result = 0; - } else { - sk->sk_state = BT_DISCONN; - l2cap_sock_set_timer(sk, HZ/10); - result = L2CAP_CR_SEC_BLOCK; - } + l2cap_pi(sk)->ident = l2cap_get_ident(conn); - rsp.scid = cpu_to_le16(l2cap_pi(sk)->dcid); - rsp.dcid = cpu_to_le16(l2cap_pi(sk)->scid); - rsp.result = cpu_to_le16(result); - rsp.status = cpu_to_le16(0); - l2cap_send_cmd(conn, l2cap_pi(sk)->ident, - L2CAP_CONN_RSP, sizeof(rsp), &rsp); + l2cap_send_cmd(conn, l2cap_pi(sk)->ident, + L2CAP_CONN_REQ, sizeof(req), &req); + } else { + l2cap_sock_clear_timer(sk); + l2cap_sock_set_timer(sk, HZ / 10); + } + } else if (sk->sk_state == BT_CONNECT2) { + struct l2cap_conn_rsp rsp; + __u16 result; - if (l2cap_pi(sk)->link_mode & L2CAP_LM_SECURE) - hci_conn_change_link_key(hcon); + if (!status) { + sk->sk_state = BT_CONFIG; + result = L2CAP_CR_SUCCESS; + } else { + sk->sk_state = BT_DISCONN; + l2cap_sock_set_timer(sk, HZ / 10); + result = L2CAP_CR_SEC_BLOCK; + } + + rsp.scid = cpu_to_le16(l2cap_pi(sk)->dcid); + rsp.dcid = cpu_to_le16(l2cap_pi(sk)->scid); + rsp.result = cpu_to_le16(result); + rsp.status = cpu_to_le16(0); + l2cap_send_cmd(conn, l2cap_pi(sk)->ident, + L2CAP_CONN_RSP, sizeof(rsp), &rsp); + } bh_unlock_sock(sk); } read_unlock(&l->lock); + return 0; } -- cgit v1.2.2 From 49292d56352a6ab90d04c3448dd8b6106dfef2d6 Mon Sep 17 00:00:00 2001 From: Samuel Ortiz Date: Fri, 4 Jul 2008 10:49:31 +0200 Subject: mac80211: power management wext hooks This patch implements the power management routines wireless extensions for mac80211. For now we only support switching PS mode between on and off. Signed-off-by: Samuel Ortiz Signed-off-by: John W. Linville --- net/mac80211/wext.c | 43 +++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 41 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c index 736c32e340f2..207971e9ad72 100644 --- a/net/mac80211/wext.c +++ b/net/mac80211/wext.c @@ -1009,6 +1009,45 @@ static int ieee80211_ioctl_giwencode(struct net_device *dev, return 0; } +static int ieee80211_ioctl_siwpower(struct net_device *dev, + struct iw_request_info *info, + struct iw_param *wrq, + char *extra) +{ + struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); + struct ieee80211_conf *conf = &local->hw.conf; + + if (wrq->disabled) { + conf->flags &= ~IEEE80211_CONF_PS; + return ieee80211_hw_config(local); + } + + switch (wrq->flags & IW_POWER_MODE) { + case IW_POWER_ON: /* If not specified */ + case IW_POWER_MODE: /* If set all mask */ + case IW_POWER_ALL_R: /* If explicitely state all */ + conf->flags |= IEEE80211_CONF_PS; + break; + default: /* Otherwise we don't support it */ + return -EINVAL; + } + + return ieee80211_hw_config(local); +} + +static int ieee80211_ioctl_giwpower(struct net_device *dev, + struct iw_request_info *info, + union iwreq_data *wrqu, + char *extra) +{ + struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); + struct ieee80211_conf *conf = &local->hw.conf; + + wrqu->power.disabled = !(conf->flags & IEEE80211_CONF_PS); + + return 0; +} + static int ieee80211_ioctl_siwauth(struct net_device *dev, struct iw_request_info *info, struct iw_param *data, char *extra) @@ -1211,8 +1250,8 @@ static const iw_handler ieee80211_handler[] = (iw_handler) ieee80211_ioctl_giwretry, /* SIOCGIWRETRY */ (iw_handler) ieee80211_ioctl_siwencode, /* SIOCSIWENCODE */ (iw_handler) ieee80211_ioctl_giwencode, /* SIOCGIWENCODE */ - (iw_handler) NULL, /* SIOCSIWPOWER */ - (iw_handler) NULL, /* SIOCGIWPOWER */ + (iw_handler) ieee80211_ioctl_siwpower, /* SIOCSIWPOWER */ + (iw_handler) ieee80211_ioctl_giwpower, /* SIOCGIWPOWER */ (iw_handler) NULL, /* -- hole -- */ (iw_handler) NULL, /* -- hole -- */ (iw_handler) ieee80211_ioctl_siwgenie, /* SIOCSIWGENIE */ -- cgit v1.2.2 From 3e122be089e6fb8d3f322416da4cdbb80ce12927 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Wed, 9 Jul 2008 14:40:34 +0200 Subject: mac80211: make master netdev handling sane Currently, almost every interface type has a 'bss' pointer pointing to BSS information. This BSS information, however, is for a _local_ BSS, not for the BSS we joined, so having it on a STA mode interface makes little sense, but now they have it pointing to the master device, which is an AP mode virtual interface. However, except for some bitrate control data, this pointer is only used in AP/VLAN modes (for power saving stations.) Overall, it is not necessary to even have the master netdev be a valid virtual interface, and it doesn't have to be on the list of interfaces either. This patch changes the master netdev to be special, it now - no longer is on the list of virtual interfaces, which lets me remove a lot of tests for that - no longer has sub_if_data attached, since that isn't used Additionally, this patch changes some vlan/ap mode handling that is related to these 'bss' pointers described above (but in the VLAN case they actually make sense because there they point to the AP they belong to); it also adds some debugging code to IEEE80211_DEV_TO_SUB_IF to validate it is not called on the master netdev any more. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/cfg.c | 2 +- net/mac80211/debugfs_netdev.c | 34 ++++++++++++--- net/mac80211/ieee80211_i.h | 41 ++++++++++-------- net/mac80211/iface.c | 64 ++++++--------------------- net/mac80211/main.c | 96 ++++++++++++----------------------------- net/mac80211/mlme.c | 56 ++++++++---------------- net/mac80211/rc80211_pid_algo.c | 8 ++-- net/mac80211/rx.c | 15 +++---- net/mac80211/sta_info.c | 29 +++++++++---- net/mac80211/tx.c | 11 +++-- net/mac80211/util.c | 4 -- net/mac80211/wext.c | 10 ++--- 12 files changed, 153 insertions(+), 217 deletions(-) (limited to 'net') diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c index 81087281b031..3c95cd9bf8ee 100644 --- a/net/mac80211/cfg.c +++ b/net/mac80211/cfg.c @@ -57,7 +57,7 @@ static int ieee80211_add_iface(struct wiphy *wiphy, char *name, if (itype == IEEE80211_IF_TYPE_INVALID) return -EINVAL; - err = ieee80211_if_add(local->mdev, name, &dev, itype, params); + err = ieee80211_if_add(local, name, &dev, itype, params); if (err || itype != IEEE80211_IF_TYPE_MNTR || !flags) return err; diff --git a/net/mac80211/debugfs_netdev.c b/net/mac80211/debugfs_netdev.c index b2089b2da48a..4aa6621f9970 100644 --- a/net/mac80211/debugfs_netdev.c +++ b/net/mac80211/debugfs_netdev.c @@ -156,6 +156,8 @@ static const struct file_operations name##_ops = { \ /* common attributes */ IEEE80211_IF_FILE(drop_unencrypted, drop_unencrypted, DEC); +IEEE80211_IF_FILE(force_unicast_rateidx, force_unicast_rateidx, DEC); +IEEE80211_IF_FILE(max_ratectrl_rateidx, max_ratectrl_rateidx, DEC); /* STA/IBSS attributes */ IEEE80211_IF_FILE(state, u.sta.state, DEC); @@ -191,8 +193,6 @@ __IEEE80211_IF_FILE(flags); IEEE80211_IF_FILE(num_sta_ps, u.ap.num_sta_ps, ATOMIC); IEEE80211_IF_FILE(dtim_count, u.ap.dtim_count, DEC); IEEE80211_IF_FILE(num_beacons, u.ap.num_beacons, DEC); -IEEE80211_IF_FILE(force_unicast_rateidx, u.ap.force_unicast_rateidx, DEC); -IEEE80211_IF_FILE(max_ratectrl_rateidx, u.ap.max_ratectrl_rateidx, DEC); static ssize_t ieee80211_if_fmt_num_buffered_multicast( const struct ieee80211_sub_if_data *sdata, char *buf, int buflen) @@ -248,6 +248,9 @@ IEEE80211_IF_WFILE(min_discovery_timeout, static void add_sta_files(struct ieee80211_sub_if_data *sdata) { DEBUGFS_ADD(drop_unencrypted, sta); + DEBUGFS_ADD(force_unicast_rateidx, ap); + DEBUGFS_ADD(max_ratectrl_rateidx, ap); + DEBUGFS_ADD(state, sta); DEBUGFS_ADD(bssid, sta); DEBUGFS_ADD(prev_bssid, sta); @@ -268,23 +271,29 @@ static void add_sta_files(struct ieee80211_sub_if_data *sdata) static void add_ap_files(struct ieee80211_sub_if_data *sdata) { DEBUGFS_ADD(drop_unencrypted, ap); + DEBUGFS_ADD(force_unicast_rateidx, ap); + DEBUGFS_ADD(max_ratectrl_rateidx, ap); + DEBUGFS_ADD(num_sta_ps, ap); DEBUGFS_ADD(dtim_count, ap); DEBUGFS_ADD(num_beacons, ap); - DEBUGFS_ADD(force_unicast_rateidx, ap); - DEBUGFS_ADD(max_ratectrl_rateidx, ap); DEBUGFS_ADD(num_buffered_multicast, ap); } static void add_wds_files(struct ieee80211_sub_if_data *sdata) { DEBUGFS_ADD(drop_unencrypted, wds); + DEBUGFS_ADD(force_unicast_rateidx, ap); + DEBUGFS_ADD(max_ratectrl_rateidx, ap); + DEBUGFS_ADD(peer, wds); } static void add_vlan_files(struct ieee80211_sub_if_data *sdata) { DEBUGFS_ADD(drop_unencrypted, vlan); + DEBUGFS_ADD(force_unicast_rateidx, ap); + DEBUGFS_ADD(max_ratectrl_rateidx, ap); } static void add_monitor_files(struct ieee80211_sub_if_data *sdata) @@ -372,6 +381,9 @@ static void add_files(struct ieee80211_sub_if_data *sdata) static void del_sta_files(struct ieee80211_sub_if_data *sdata) { DEBUGFS_DEL(drop_unencrypted, sta); + DEBUGFS_DEL(force_unicast_rateidx, ap); + DEBUGFS_DEL(max_ratectrl_rateidx, ap); + DEBUGFS_DEL(state, sta); DEBUGFS_DEL(bssid, sta); DEBUGFS_DEL(prev_bssid, sta); @@ -392,23 +404,29 @@ static void del_sta_files(struct ieee80211_sub_if_data *sdata) static void del_ap_files(struct ieee80211_sub_if_data *sdata) { DEBUGFS_DEL(drop_unencrypted, ap); + DEBUGFS_DEL(force_unicast_rateidx, ap); + DEBUGFS_DEL(max_ratectrl_rateidx, ap); + DEBUGFS_DEL(num_sta_ps, ap); DEBUGFS_DEL(dtim_count, ap); DEBUGFS_DEL(num_beacons, ap); - DEBUGFS_DEL(force_unicast_rateidx, ap); - DEBUGFS_DEL(max_ratectrl_rateidx, ap); DEBUGFS_DEL(num_buffered_multicast, ap); } static void del_wds_files(struct ieee80211_sub_if_data *sdata) { DEBUGFS_DEL(drop_unencrypted, wds); + DEBUGFS_DEL(force_unicast_rateidx, ap); + DEBUGFS_DEL(max_ratectrl_rateidx, ap); + DEBUGFS_DEL(peer, wds); } static void del_vlan_files(struct ieee80211_sub_if_data *sdata) { DEBUGFS_DEL(drop_unencrypted, vlan); + DEBUGFS_DEL(force_unicast_rateidx, ap); + DEBUGFS_DEL(max_ratectrl_rateidx, ap); } static void del_monitor_files(struct ieee80211_sub_if_data *sdata) @@ -525,7 +543,7 @@ static int netdev_notify(struct notifier_block *nb, { struct net_device *dev = ndev; struct dentry *dir; - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); + struct ieee80211_sub_if_data *sdata; char buf[10+IFNAMSIZ]; if (state != NETDEV_CHANGENAME) @@ -537,6 +555,8 @@ static int netdev_notify(struct notifier_block *nb, if (dev->ieee80211_ptr->wiphy->privid != mac80211_wiphy_privid) return 0; + sdata = IEEE80211_DEV_TO_SUB_IF(dev); + sprintf(buf, "netdev:%s", dev->name); dir = sdata->debugfsdir; if (!debugfs_rename(dir->d_parent, dir, dir->d_parent, buf)) diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 02a8753a4eca..1b1fc53dad36 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -237,8 +237,6 @@ struct ieee80211_if_ap { struct sk_buff_head ps_bc_buf; atomic_t num_sta_ps; /* number of stations in PS mode */ int dtim_count; - int force_unicast_rateidx; /* forced TX rateidx for unicast frames */ - int max_ratectrl_rateidx; /* max TX rateidx for rate control */ int num_beacons; /* number of TXed beacon frames for this BSS */ }; @@ -248,7 +246,6 @@ struct ieee80211_if_wds { }; struct ieee80211_if_vlan { - struct ieee80211_sub_if_data *ap; struct list_head list; }; @@ -432,16 +429,18 @@ struct ieee80211_sub_if_data { struct ieee80211_key *keys[NUM_DEFAULT_KEYS]; struct ieee80211_key *default_key; + /* BSS configuration for this interface. */ + struct ieee80211_bss_conf bss_conf; + /* - * BSS configuration for this interface. - * - * FIXME: I feel bad putting this here when we already have a - * bss pointer, but the bss pointer is just wrong when - * you have multiple virtual STA mode interfaces... - * This needs to be fixed. + * AP this belongs to: self in AP mode and + * corresponding AP in VLAN mode, NULL for + * all others (might be needed later in IBSS) */ - struct ieee80211_bss_conf bss_conf; - struct ieee80211_if_ap *bss; /* BSS that this device belongs to */ + struct ieee80211_if_ap *bss; + + int force_unicast_rateidx; /* forced TX rateidx for unicast frames */ + int max_ratectrl_rateidx; /* max TX rateidx for rate control */ union { struct ieee80211_if_ap ap; @@ -533,8 +532,6 @@ struct ieee80211_sub_if_data *vif_to_sdata(struct ieee80211_vif *p) return container_of(p, struct ieee80211_sub_if_data, vif); } -#define IEEE80211_DEV_TO_SUB_IF(dev) netdev_priv(dev) - enum { IEEE80211_RX_MSG = 1, IEEE80211_TX_STATUS_MSG = 2, @@ -760,6 +757,16 @@ static inline int ieee80211_is_multiqueue(struct ieee80211_local *local) #endif } +static inline struct ieee80211_sub_if_data * +IEEE80211_DEV_TO_SUB_IF(struct net_device *dev) +{ + struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); + + BUG_ON(!local || local->mdev == dev); + + return netdev_priv(dev); +} + /* this struct represents 802.11n's RA/TID combination */ struct ieee80211_ra_tid { u8 ra[ETH_ALEN]; @@ -883,8 +890,8 @@ int ieee80211_sta_scan_results(struct net_device *dev, ieee80211_rx_result ieee80211_sta_rx_scan( struct net_device *dev, struct sk_buff *skb, struct ieee80211_rx_status *rx_status); -void ieee80211_rx_bss_list_init(struct net_device *dev); -void ieee80211_rx_bss_list_deinit(struct net_device *dev); +void ieee80211_rx_bss_list_init(struct ieee80211_local *local); +void ieee80211_rx_bss_list_deinit(struct ieee80211_local *local); int ieee80211_sta_set_extra_ie(struct net_device *dev, char *ie, size_t len); struct sta_info *ieee80211_ibss_add_sta(struct net_device *dev, struct sk_buff *skb, u8 *bssid, @@ -925,8 +932,8 @@ static inline void ieee80211_start_mesh(struct net_device *dev) {} #endif -/* ieee80211_iface.c */ -int ieee80211_if_add(struct net_device *dev, const char *name, +/* interface handling */ +int ieee80211_if_add(struct ieee80211_local *local, const char *name, struct net_device **new_dev, int type, struct vif_params *params); void ieee80211_if_set_type(struct net_device *dev, int type); diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c index eeb16926aa7d..f2aefd4b8637 100644 --- a/net/mac80211/iface.c +++ b/net/mac80211/iface.c @@ -27,6 +27,9 @@ void ieee80211_if_sdata_init(struct ieee80211_sub_if_data *sdata) skb_queue_head_init(&sdata->fragments[i].skb_list); INIT_LIST_HEAD(&sdata->key_list); + + sdata->force_unicast_rateidx = -1; + sdata->max_ratectrl_rateidx = -1; } static void ieee80211_if_sdata_deinit(struct ieee80211_sub_if_data *sdata) @@ -38,12 +41,11 @@ static void ieee80211_if_sdata_deinit(struct ieee80211_sub_if_data *sdata) } /* Must be called with rtnl lock held. */ -int ieee80211_if_add(struct net_device *dev, const char *name, +int ieee80211_if_add(struct ieee80211_local *local, const char *name, struct net_device **new_dev, int type, struct vif_params *params) { struct net_device *ndev; - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct ieee80211_sub_if_data *sdata = NULL; int ret; @@ -67,13 +69,10 @@ int ieee80211_if_add(struct net_device *dev, const char *name, goto fail; memcpy(ndev->dev_addr, local->hw.wiphy->perm_addr, ETH_ALEN); - ndev->base_addr = dev->base_addr; - ndev->irq = dev->irq; - ndev->mem_start = dev->mem_start; - ndev->mem_end = dev->mem_end; SET_NETDEV_DEV(ndev, wiphy_dev(local->hw.wiphy)); - sdata = IEEE80211_DEV_TO_SUB_IF(ndev); + /* don't use IEEE80211_DEV_TO_SUB_IF because it checks too much */ + sdata = netdev_priv(ndev); ndev->ieee80211_ptr = &sdata->wdev; sdata->wdev.wiphy = local->hw.wiphy; sdata->vif.type = IEEE80211_IF_TYPE_AP; @@ -116,14 +115,6 @@ void ieee80211_if_set_type(struct net_device *dev, int type) struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); int oldtype = sdata->vif.type; - /* - * We need to call this function on the master interface - * which already has a hard_start_xmit routine assigned - * which must not be changed. - */ - if (dev != sdata->local->mdev) - dev->hard_start_xmit = ieee80211_subif_start_xmit; - /* * Called even when register_netdevice fails, it would * oops if assigned before initialising the rest. @@ -138,22 +129,16 @@ void ieee80211_if_set_type(struct net_device *dev, int type) switch (type) { case IEEE80211_IF_TYPE_WDS: - /* nothing special */ - break; case IEEE80211_IF_TYPE_VLAN: - sdata->u.vlan.ap = NULL; + /* nothing special */ break; case IEEE80211_IF_TYPE_AP: - sdata->u.ap.force_unicast_rateidx = -1; - sdata->u.ap.max_ratectrl_rateidx = -1; skb_queue_head_init(&sdata->u.ap.ps_bc_buf); - sdata->bss = &sdata->u.ap; INIT_LIST_HEAD(&sdata->u.ap.vlans); break; case IEEE80211_IF_TYPE_MESH_POINT: case IEEE80211_IF_TYPE_STA: case IEEE80211_IF_TYPE_IBSS: { - struct ieee80211_sub_if_data *msdata; struct ieee80211_if_sta *ifsta; ifsta = &sdata->u.sta; @@ -171,9 +156,6 @@ void ieee80211_if_set_type(struct net_device *dev, int type) if (ieee80211_num_regular_queues(&sdata->local->hw) >= 4) ifsta->flags |= IEEE80211_STA_WMM_ENABLED; - msdata = IEEE80211_DEV_TO_SUB_IF(sdata->local->mdev); - sdata->bss = &msdata->u.ap; - if (ieee80211_vif_is_mesh(&sdata->vif)) ieee80211_mesh_init_sdata(sdata); break; @@ -215,27 +197,8 @@ void ieee80211_if_reinit(struct net_device *dev) WARN_ON(1); break; case IEEE80211_IF_TYPE_AP: { - /* Remove all virtual interfaces that use this BSS - * as their sdata->bss */ - struct ieee80211_sub_if_data *tsdata, *n; struct beacon_data *beacon; - list_for_each_entry_safe(tsdata, n, &local->interfaces, list) { - if (tsdata != sdata && tsdata->bss == &sdata->u.ap) { - printk(KERN_DEBUG "%s: removing virtual " - "interface %s because its BSS interface" - " is being removed\n", - sdata->dev->name, tsdata->dev->name); - list_del_rcu(&tsdata->list); - /* - * We have lots of time and can afford - * to sync for each interface - */ - synchronize_rcu(); - __ieee80211_if_del(local, tsdata); - } - } - beacon = sdata->u.ap.beacon; rcu_assign_pointer(sdata->u.ap.beacon, NULL); synchronize_rcu(); @@ -249,6 +212,7 @@ void ieee80211_if_reinit(struct net_device *dev) break; } case IEEE80211_IF_TYPE_WDS: + case IEEE80211_IF_TYPE_VLAN: /* nothing to do */ break; case IEEE80211_IF_TYPE_MESH_POINT: @@ -269,9 +233,6 @@ void ieee80211_if_reinit(struct net_device *dev) case IEEE80211_IF_TYPE_MNTR: dev->type = ARPHRD_ETHER; break; - case IEEE80211_IF_TYPE_VLAN: - sdata->u.vlan.ap = NULL; - break; } flushed = sta_info_flush(local, sdata); @@ -289,8 +250,10 @@ void __ieee80211_if_del(struct ieee80211_local *local, ieee80211_debugfs_remove_netdev(sdata); unregister_netdevice(dev); - /* Except master interface, the net_device will be freed by - * net_device->destructor (i. e. ieee80211_if_free). */ + /* + * The net_device will be freed by its destructor, + * i.e. ieee80211_if_free. + */ } /* Must be called with rtnl lock held. */ @@ -303,8 +266,7 @@ int ieee80211_if_remove(struct net_device *dev, const char *name, int id) list_for_each_entry_safe(sdata, n, &local->interfaces, list) { if ((sdata->vif.type == id || id == -1) && - strcmp(name, sdata->dev->name) == 0 && - sdata->dev != local->mdev) { + strcmp(name, sdata->dev->name) == 0) { list_del_rcu(&sdata->list); synchronize_rcu(); __ieee80211_if_del(local, sdata); diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 1c4d3ba6b878..863923964d21 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -105,7 +105,7 @@ static int ieee80211_master_open(struct net_device *dev) /* we hold the RTNL here so can safely walk the list */ list_for_each_entry(sdata, &local->interfaces, list) { - if (sdata->dev != dev && netif_running(sdata->dev)) { + if (netif_running(sdata->dev)) { res = 0; break; } @@ -126,7 +126,7 @@ static int ieee80211_master_stop(struct net_device *dev) /* we hold the RTNL here so can safely walk the list */ list_for_each_entry(sdata, &local->interfaces, list) - if (sdata->dev != dev && netif_running(sdata->dev)) + if (netif_running(sdata->dev)) dev_close(sdata->dev); return 0; @@ -194,7 +194,7 @@ static int ieee80211_open(struct net_device *dev) list_for_each_entry(nsdata, &local->interfaces, list) { struct net_device *ndev = nsdata->dev; - if (ndev != dev && ndev != local->mdev && netif_running(ndev)) { + if (ndev != dev && netif_running(ndev)) { /* * Allow only a single IBSS interface to be up at any * time. This is restricted because beacon distribution @@ -209,30 +209,6 @@ static int ieee80211_open(struct net_device *dev) nsdata->vif.type == IEEE80211_IF_TYPE_IBSS) return -EBUSY; - /* - * Disallow multiple IBSS/STA mode interfaces. - * - * This is a technical restriction, it is possible although - * most likely not IEEE 802.11 compliant to have multiple - * STAs with just a single hardware (the TSF timer will not - * be adjusted properly.) - * - * However, because mac80211 uses the master device's BSS - * information for each STA/IBSS interface, doing this will - * currently corrupt that BSS information completely, unless, - * a not very useful case, both STAs are associated to the - * same BSS. - * - * To remove this restriction, the BSS information needs to - * be embedded in the STA/IBSS mode sdata instead of using - * the master device's BSS structure. - */ - if ((sdata->vif.type == IEEE80211_IF_TYPE_STA || - sdata->vif.type == IEEE80211_IF_TYPE_IBSS) && - (nsdata->vif.type == IEEE80211_IF_TYPE_STA || - nsdata->vif.type == IEEE80211_IF_TYPE_IBSS)) - return -EBUSY; - /* * The remaining checks are only performed for interfaces * with the same MAC address. @@ -252,7 +228,7 @@ static int ieee80211_open(struct net_device *dev) */ if (sdata->vif.type == IEEE80211_IF_TYPE_VLAN && nsdata->vif.type == IEEE80211_IF_TYPE_AP) - sdata->u.vlan.ap = nsdata; + sdata->bss = &nsdata->u.ap; } } @@ -262,10 +238,13 @@ static int ieee80211_open(struct net_device *dev) return -ENOLINK; break; case IEEE80211_IF_TYPE_VLAN: - if (!sdata->u.vlan.ap) + if (!sdata->bss) return -ENOLINK; + list_add(&sdata->u.vlan.list, &sdata->bss->vlans); break; case IEEE80211_IF_TYPE_AP: + sdata->bss = &sdata->u.ap; + break; case IEEE80211_IF_TYPE_STA: case IEEE80211_IF_TYPE_MNTR: case IEEE80211_IF_TYPE_IBSS: @@ -283,14 +262,13 @@ static int ieee80211_open(struct net_device *dev) if (local->ops->start) res = local->ops->start(local_to_hw(local)); if (res) - return res; + goto err_del_bss; need_hw_reconfig = 1; ieee80211_led_radio(local, local->hw.conf.radio_enabled); } switch (sdata->vif.type) { case IEEE80211_IF_TYPE_VLAN: - list_add(&sdata->u.vlan.list, &sdata->u.vlan.ap->u.ap.vlans); /* no need to tell driver */ break; case IEEE80211_IF_TYPE_MNTR: @@ -404,6 +382,10 @@ static int ieee80211_open(struct net_device *dev) err_stop: if (!local->open_count && local->ops->stop) local->ops->stop(local_to_hw(local)); + err_del_bss: + sdata->bss = NULL; + if (sdata->vif.type == IEEE80211_IF_TYPE_VLAN) + list_del(&sdata->u.vlan.list); return res; } @@ -486,7 +468,6 @@ static int ieee80211_stop(struct net_device *dev) switch (sdata->vif.type) { case IEEE80211_IF_TYPE_VLAN: list_del(&sdata->u.vlan.list); - sdata->u.vlan.ap = NULL; /* no need to tell driver */ break; case IEEE80211_IF_TYPE_MNTR: @@ -549,6 +530,8 @@ static int ieee80211_stop(struct net_device *dev) local->ops->remove_interface(local_to_hw(local), &conf); } + sdata->bss = NULL; + if (local->open_count == 0) { if (netif_running(local->mdev)) dev_close(local->mdev); @@ -1659,7 +1642,7 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) int result; enum ieee80211_band band; struct net_device *mdev; - struct ieee80211_sub_if_data *sdata; + struct wireless_dev *mwdev; /* * generic code guarantees at least one band, @@ -1699,8 +1682,7 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) hw->ampdu_queues = 0; #endif - /* for now, mdev needs sub_if_data :/ */ - mdev = alloc_netdev_mq(sizeof(struct ieee80211_sub_if_data), + mdev = alloc_netdev_mq(sizeof(struct wireless_dev), "wmaster%d", ether_setup, ieee80211_num_queues(hw)); if (!mdev) @@ -1709,13 +1691,13 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) if (ieee80211_num_queues(hw) > 1) mdev->features |= NETIF_F_MULTI_QUEUE; - sdata = IEEE80211_DEV_TO_SUB_IF(mdev); - mdev->ieee80211_ptr = &sdata->wdev; - sdata->wdev.wiphy = local->hw.wiphy; + mwdev = netdev_priv(mdev); + mdev->ieee80211_ptr = mwdev; + mwdev->wiphy = local->hw.wiphy; local->mdev = mdev; - ieee80211_rx_bss_list_init(mdev); + ieee80211_rx_bss_list_init(local); mdev->hard_start_xmit = ieee80211_master_start_xmit; mdev->open = ieee80211_master_open; @@ -1724,16 +1706,6 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) mdev->header_ops = &ieee80211_header_ops; mdev->set_multicast_list = ieee80211_master_set_multicast_list; - sdata->vif.type = IEEE80211_IF_TYPE_AP; - sdata->dev = mdev; - sdata->local = local; - sdata->u.ap.force_unicast_rateidx = -1; - sdata->u.ap.max_ratectrl_rateidx = -1; - ieee80211_if_sdata_init(sdata); - - /* no RCU needed since we're still during init phase */ - list_add_tail(&sdata->list, &local->interfaces); - name = wiphy_dev(local->hw.wiphy)->driver->name; local->hw.workqueue = create_freezeable_workqueue(name); if (!local->hw.workqueue) { @@ -1779,9 +1751,6 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) if (result < 0) goto fail_dev; - ieee80211_debugfs_add_netdev(IEEE80211_DEV_TO_SUB_IF(local->mdev)); - ieee80211_if_set_type(local->mdev, IEEE80211_IF_TYPE_AP); - result = ieee80211_init_rate_ctrl_alg(local, hw->rate_control_algorithm); if (result < 0) { @@ -1801,7 +1770,7 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) ieee80211_install_qdisc(local->mdev); /* add one default STA interface */ - result = ieee80211_if_add(local->mdev, "wlan%d", NULL, + result = ieee80211_if_add(local, "wlan%d", NULL, IEEE80211_IF_TYPE_STA, NULL); if (result) printk(KERN_WARNING "%s: Failed to add default virtual iface\n", @@ -1817,7 +1786,6 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) fail_wep: rate_control_deinitialize(local); fail_rate: - ieee80211_debugfs_remove_netdev(IEEE80211_DEV_TO_SUB_IF(local->mdev)); unregister_netdevice(local->mdev); local->mdev = NULL; fail_dev: @@ -1827,10 +1795,8 @@ fail_sta_info: debugfs_hw_del(local); destroy_workqueue(local->hw.workqueue); fail_workqueue: - if (local->mdev != NULL) { - ieee80211_if_free(local->mdev); - local->mdev = NULL; - } + if (local->mdev) + free_netdev(local->mdev); fail_mdev_alloc: wiphy_unregister(local->hw.wiphy); return result; @@ -1858,24 +1824,19 @@ void ieee80211_unregister_hw(struct ieee80211_hw *hw) */ /* - * First, we remove all non-master interfaces. Do this because they - * may have bss pointer dependency on the master, and when we free - * the master these would be freed as well, breaking our list - * iteration completely. + * First, we remove all virtual interfaces. */ list_for_each_entry_safe(sdata, tmp, &local->interfaces, list) { - if (sdata->dev == local->mdev) - continue; list_del(&sdata->list); __ieee80211_if_del(local, sdata); } /* then, finally, remove the master interface */ - __ieee80211_if_del(local, IEEE80211_DEV_TO_SUB_IF(local->mdev)); + unregister_netdevice(local->mdev); rtnl_unlock(); - ieee80211_rx_bss_list_deinit(local->mdev); + ieee80211_rx_bss_list_deinit(local); ieee80211_clear_tx_pending(local); sta_info_stop(local); rate_control_deinitialize(local); @@ -1892,8 +1853,7 @@ void ieee80211_unregister_hw(struct ieee80211_hw *hw) wiphy_unregister(local->hw.wiphy); ieee80211_wep_free(local); ieee80211_led_exit(local); - ieee80211_if_free(local->mdev); - local->mdev = NULL; + free_netdev(local->mdev); } EXPORT_SYMBOL(ieee80211_unregister_hw); diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index dbc8cf454bc0..64d710a88b86 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -78,7 +78,7 @@ static void ieee80211_send_probe_req(struct net_device *dev, u8 *dst, static struct ieee80211_sta_bss * ieee80211_rx_bss_get(struct net_device *dev, u8 *bssid, int freq, u8 *ssid, u8 ssid_len); -static void ieee80211_rx_bss_put(struct net_device *dev, +static void ieee80211_rx_bss_put(struct ieee80211_local *local, struct ieee80211_sta_bss *bss); static int ieee80211_sta_find_ibss(struct net_device *dev, struct ieee80211_if_sta *ifsta); @@ -554,7 +554,7 @@ static void ieee80211_set_associated(struct net_device *dev, changed |= ieee80211_handle_bss_capability(sdata, bss); - ieee80211_rx_bss_put(dev, bss); + ieee80211_rx_bss_put(local, bss); } if (conf->flags & IEEE80211_CONF_SUPPORT_HT_MODE) { @@ -760,7 +760,7 @@ static void ieee80211_send_assoc(struct net_device *dev, (local->hw.flags & IEEE80211_HW_SPECTRUM_MGMT)) capab |= WLAN_CAPABILITY_SPECTRUM_MGMT; - ieee80211_rx_bss_put(dev, bss); + ieee80211_rx_bss_put(local, bss); } else { rates = ~0; rates_len = sband->n_bitrates; @@ -992,7 +992,7 @@ static int ieee80211_privacy_mismatch(struct net_device *dev, wep_privacy = !!ieee80211_sta_wep_configured(dev); privacy_invoked = !!(ifsta->flags & IEEE80211_STA_PRIVACY_INVOKED); - ieee80211_rx_bss_put(dev, bss); + ieee80211_rx_bss_put(local, bss); if ((bss_privacy == wep_privacy) || (bss_privacy == privacy_invoked)) return 0; @@ -2094,7 +2094,7 @@ static void ieee80211_rx_mgmt_assoc_resp(struct ieee80211_sub_if_data *sdata, sta->last_signal = bss->signal; sta->last_qual = bss->qual; sta->last_noise = bss->noise; - ieee80211_rx_bss_put(dev, bss); + ieee80211_rx_bss_put(local, bss); } err = sta_info_insert(sta); @@ -2212,10 +2212,9 @@ static void __ieee80211_rx_bss_hash_add(struct net_device *dev, /* Caller must hold local->sta_bss_lock */ -static void __ieee80211_rx_bss_hash_del(struct net_device *dev, +static void __ieee80211_rx_bss_hash_del(struct ieee80211_local *local, struct ieee80211_sta_bss *bss) { - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct ieee80211_sta_bss *b, *prev = NULL; b = local->sta_bss_hash[STA_HASH(bss->bssid)]; while (b) { @@ -2367,39 +2366,35 @@ static void ieee80211_rx_bss_free(struct ieee80211_sta_bss *bss) } -static void ieee80211_rx_bss_put(struct net_device *dev, +static void ieee80211_rx_bss_put(struct ieee80211_local *local, struct ieee80211_sta_bss *bss) { - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - local_bh_disable(); if (!atomic_dec_and_lock(&bss->users, &local->sta_bss_lock)) { local_bh_enable(); return; } - __ieee80211_rx_bss_hash_del(dev, bss); + __ieee80211_rx_bss_hash_del(local, bss); list_del(&bss->list); spin_unlock_bh(&local->sta_bss_lock); ieee80211_rx_bss_free(bss); } -void ieee80211_rx_bss_list_init(struct net_device *dev) +void ieee80211_rx_bss_list_init(struct ieee80211_local *local) { - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); spin_lock_init(&local->sta_bss_lock); INIT_LIST_HEAD(&local->sta_bss_list); } -void ieee80211_rx_bss_list_deinit(struct net_device *dev) +void ieee80211_rx_bss_list_deinit(struct ieee80211_local *local) { - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct ieee80211_sta_bss *bss, *tmp; list_for_each_entry_safe(bss, tmp, &local->sta_bss_list, list) - ieee80211_rx_bss_put(dev, bss); + ieee80211_rx_bss_put(local, bss); } @@ -2775,7 +2770,7 @@ static void ieee80211_rx_bss_info(struct net_device *dev, */ if (sdata->vif.type != IEEE80211_IF_TYPE_IBSS && bss->probe_resp && beacon) { - ieee80211_rx_bss_put(dev, bss); + ieee80211_rx_bss_put(local, bss); return; } @@ -2918,7 +2913,7 @@ static void ieee80211_rx_bss_info(struct net_device *dev, } } - ieee80211_rx_bss_put(dev, bss); + ieee80211_rx_bss_put(local, bss); } @@ -3578,7 +3573,7 @@ static int ieee80211_sta_config_auth(struct net_device *dev, selected->ssid_len); ieee80211_sta_set_bssid(dev, selected->bssid); ieee80211_sta_def_wmm_params(dev, selected, 0); - ieee80211_rx_bss_put(dev, selected); + ieee80211_rx_bss_put(local, selected); ifsta->state = IEEE80211_AUTHENTICATE; ieee80211_sta_reset_auth(dev, ifsta); return 0; @@ -3655,7 +3650,7 @@ static int ieee80211_sta_create_ibss(struct net_device *dev, } ret = ieee80211_sta_join_ibss(dev, ifsta, bss); - ieee80211_rx_bss_put(dev, bss); + ieee80211_rx_bss_put(local, bss); return ret; } @@ -3711,7 +3706,7 @@ static int ieee80211_sta_find_ibss(struct net_device *dev, " based on configured SSID\n", dev->name, print_mac(mac, bssid)); ret = ieee80211_sta_join_ibss(dev, ifsta, bss); - ieee80211_rx_bss_put(dev, bss); + ieee80211_rx_bss_put(local, bss); return ret; } #ifdef CONFIG_MAC80211_IBSS_DEBUG @@ -3907,11 +3902,6 @@ void ieee80211_scan_completed(struct ieee80211_hw *hw) rcu_read_lock(); list_for_each_entry_rcu(sdata, &local->interfaces, list) { - - /* No need to wake the master device. */ - if (sdata->dev == local->mdev) - continue; - /* Tell AP we're back */ if (sdata->vif.type == IEEE80211_IF_TYPE_STA && sdata->u.sta.flags & IEEE80211_STA_ASSOCIATED) @@ -4077,12 +4067,6 @@ static int ieee80211_sta_start_scan(struct net_device *dev, rcu_read_lock(); list_for_each_entry_rcu(sdata, &local->interfaces, list) { - - /* Don't stop the master interface, otherwise we can't transmit - * probes! */ - if (sdata->dev == local->mdev) - continue; - netif_stop_queue(sdata->dev); if (sdata->vif.type == IEEE80211_IF_TYPE_STA && (sdata->u.sta.flags & IEEE80211_STA_ASSOCIATED)) @@ -4473,12 +4457,10 @@ void ieee80211_notify_mac(struct ieee80211_hw *hw, case IEEE80211_NOTIFY_RE_ASSOC: rcu_read_lock(); list_for_each_entry_rcu(sdata, &local->interfaces, list) { + if (sdata->vif.type != IEEE80211_IF_TYPE_STA) + continue; - if (sdata->vif.type == IEEE80211_IF_TYPE_STA) { - ieee80211_sta_req_auth(sdata->dev, - &sdata->u.sta); - } - + ieee80211_sta_req_auth(sdata->dev, &sdata->u.sta); } rcu_read_unlock(); break; diff --git a/net/mac80211/rc80211_pid_algo.c b/net/mac80211/rc80211_pid_algo.c index 62388f8e9024..ca636295e628 100644 --- a/net/mac80211/rc80211_pid_algo.c +++ b/net/mac80211/rc80211_pid_algo.c @@ -259,8 +259,8 @@ static void rate_control_pid_tx_status(void *priv, struct net_device *dev, /* Don't update the state if we're not controlling the rate. */ sdata = sta->sdata; - if (sdata->bss && sdata->bss->force_unicast_rateidx > -1) { - sta->txrate_idx = sdata->bss->max_ratectrl_rateidx; + if (sdata->force_unicast_rateidx > -1) { + sta->txrate_idx = sdata->max_ratectrl_rateidx; goto unlock; } @@ -337,8 +337,8 @@ static void rate_control_pid_get_rate(void *priv, struct net_device *dev, /* If a forced rate is in effect, select it. */ sdata = IEEE80211_DEV_TO_SUB_IF(dev); - if (sdata->bss && sdata->bss->force_unicast_rateidx > -1) - sta->txrate_idx = sdata->bss->force_unicast_rateidx; + if (sdata->force_unicast_rateidx > -1) + sta->txrate_idx = sdata->force_unicast_rateidx; rateidx = sta->txrate_idx; diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index fab443d717eb..244ee2d50a58 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -647,8 +647,7 @@ static void ap_sta_ps_start(struct net_device *dev, struct sta_info *sta) sdata = sta->sdata; - if (sdata->bss) - atomic_inc(&sdata->bss->num_sta_ps); + atomic_inc(&sdata->bss->num_sta_ps); set_and_clear_sta_flags(sta, WLAN_STA_PS, WLAN_STA_PSPOLL); #ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG printk(KERN_DEBUG "%s: STA %s aid %d enters power save mode\n", @@ -667,8 +666,7 @@ static int ap_sta_ps_end(struct net_device *dev, struct sta_info *sta) sdata = sta->sdata; - if (sdata->bss) - atomic_dec(&sdata->bss->num_sta_ps); + atomic_dec(&sdata->bss->num_sta_ps); clear_sta_flags(sta, WLAN_STA_PS | WLAN_STA_PSPOLL); @@ -742,7 +740,9 @@ ieee80211_rx_h_sta_process(struct ieee80211_rx_data *rx) sta->last_qual = rx->status->qual; sta->last_noise = rx->status->noise; - if (!ieee80211_has_morefrags(hdr->frame_control)) { + if (!ieee80211_has_morefrags(hdr->frame_control) && + (rx->sdata->vif.type == IEEE80211_IF_TYPE_AP || + rx->sdata->vif.type == IEEE80211_IF_TYPE_VLAN)) { /* Change STA power saving mode only in the end of a frame * exchange sequence */ if (test_sta_flags(sta, WLAN_STA_PS) && @@ -1772,11 +1772,6 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata, return 0; rx->flags &= ~IEEE80211_RX_RA_MATCH; } - if (sdata->dev == sdata->local->mdev && - !(rx->flags & IEEE80211_RX_IN_SCAN)) - /* do not receive anything via - * master device when not scanning */ - return 0; break; case IEEE80211_IF_TYPE_WDS: if (bssid || !ieee80211_is_data(hdr->frame_control)) diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index 47d2c1bbfcaa..f2ba653b9d69 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -320,7 +320,9 @@ int sta_info_insert(struct sta_info *sta) /* notify driver */ if (local->ops->sta_notify) { if (sdata->vif.type == IEEE80211_IF_TYPE_VLAN) - sdata = sdata->u.vlan.ap; + sdata = container_of(sdata->bss, + struct ieee80211_sub_if_data, + u.ap); local->ops->sta_notify(local_to_hw(local), &sdata->vif, STA_NOTIFY_ADD, sta->addr); @@ -375,8 +377,10 @@ static inline void __bss_tim_clear(struct ieee80211_if_ap *bss, u16 aid) static void __sta_info_set_tim_bit(struct ieee80211_if_ap *bss, struct sta_info *sta) { - if (bss) - __bss_tim_set(bss, sta->aid); + BUG_ON(!bss); + + __bss_tim_set(bss, sta->aid); + if (sta->local->ops->set_tim) { sta->local->tim_in_locked_section = true; sta->local->ops->set_tim(local_to_hw(sta->local), sta->aid, 1); @@ -388,6 +392,8 @@ void sta_info_set_tim_bit(struct sta_info *sta) { unsigned long flags; + BUG_ON(!sta->sdata->bss); + spin_lock_irqsave(&sta->local->sta_lock, flags); __sta_info_set_tim_bit(sta->sdata->bss, sta); spin_unlock_irqrestore(&sta->local->sta_lock, flags); @@ -396,8 +402,10 @@ void sta_info_set_tim_bit(struct sta_info *sta) static void __sta_info_clear_tim_bit(struct ieee80211_if_ap *bss, struct sta_info *sta) { - if (bss) - __bss_tim_clear(bss, sta->aid); + BUG_ON(!bss); + + __bss_tim_clear(bss, sta->aid); + if (sta->local->ops->set_tim) { sta->local->tim_in_locked_section = true; sta->local->ops->set_tim(local_to_hw(sta->local), sta->aid, 0); @@ -409,6 +417,8 @@ void sta_info_clear_tim_bit(struct sta_info *sta) { unsigned long flags; + BUG_ON(!sta->sdata->bss); + spin_lock_irqsave(&sta->local->sta_lock, flags); __sta_info_clear_tim_bit(sta->sdata->bss, sta); spin_unlock_irqrestore(&sta->local->sta_lock, flags); @@ -437,8 +447,9 @@ void __sta_info_unlink(struct sta_info **sta) list_del(&(*sta)->list); if (test_and_clear_sta_flags(*sta, WLAN_STA_PS)) { - if (sdata->bss) - atomic_dec(&sdata->bss->num_sta_ps); + BUG_ON(!sdata->bss); + + atomic_dec(&sdata->bss->num_sta_ps); __sta_info_clear_tim_bit(sdata->bss, *sta); } @@ -446,7 +457,9 @@ void __sta_info_unlink(struct sta_info **sta) if (local->ops->sta_notify) { if (sdata->vif.type == IEEE80211_IF_TYPE_VLAN) - sdata = sdata->u.vlan.ap; + sdata = container_of(sdata->bss, + struct ieee80211_sub_if_data, + u.ap); local->ops->sta_notify(local_to_hw(local), &sdata->vif, STA_NOTIFY_REMOVE, (*sta)->addr); diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 9bd9faac3c3c..a757dcc1208d 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -303,8 +303,7 @@ static void purge_old_ps_buffers(struct ieee80211_local *local) list_for_each_entry_rcu(sdata, &local->interfaces, list) { struct ieee80211_if_ap *ap; - if (sdata->dev == local->mdev || - sdata->vif.type != IEEE80211_IF_TYPE_AP) + if (sdata->vif.type != IEEE80211_IF_TYPE_AP) continue; ap = &sdata->u.ap; skb = skb_dequeue(&ap->ps_bc_buf); @@ -346,8 +345,12 @@ ieee80211_tx_h_multicast_ps_buf(struct ieee80211_tx_data *tx) * This is done either by the hardware or us. */ - /* not AP/IBSS or ordered frame */ - if (!tx->sdata->bss || (tx->fc & IEEE80211_FCTL_ORDER)) + /* powersaving STAs only in AP/VLAN mode */ + if (!tx->sdata->bss) + return TX_CONTINUE; + + /* no buffering for ordered frames */ + if (tx->fc & IEEE80211_FCTL_ORDER) return TX_CONTINUE; /* no stations in PS mode */ diff --git a/net/mac80211/util.c b/net/mac80211/util.c index ce62b163b82c..89ce4e07bd84 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -428,8 +428,6 @@ void ieee80211_iterate_active_interfaces( case IEEE80211_IF_TYPE_MESH_POINT: break; } - if (sdata->dev == local->mdev) - continue; if (netif_running(sdata->dev)) iterator(data, sdata->dev->dev_addr, &sdata->vif); @@ -463,8 +461,6 @@ void ieee80211_iterate_active_interfaces_atomic( case IEEE80211_IF_TYPE_MESH_POINT: break; } - if (sdata->dev == local->mdev) - continue; if (netif_running(sdata->dev)) iterator(data, sdata->dev->dev_addr, &sdata->vif); diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c index 207971e9ad72..e8e4a6215e89 100644 --- a/net/mac80211/wext.c +++ b/net/mac80211/wext.c @@ -627,16 +627,14 @@ static int ieee80211_ioctl_siwrate(struct net_device *dev, struct ieee80211_supported_band *sband; sdata = IEEE80211_DEV_TO_SUB_IF(dev); - if (!sdata->bss) - return -ENODEV; sband = local->hw.wiphy->bands[local->hw.conf.channel->band]; /* target_rate = -1, rate->fixed = 0 means auto only, so use all rates * target_rate = X, rate->fixed = 1 means only rate X * target_rate = X, rate->fixed = 0 means all rates <= X */ - sdata->bss->max_ratectrl_rateidx = -1; - sdata->bss->force_unicast_rateidx = -1; + sdata->max_ratectrl_rateidx = -1; + sdata->force_unicast_rateidx = -1; if (rate->value < 0) return 0; @@ -645,9 +643,9 @@ static int ieee80211_ioctl_siwrate(struct net_device *dev, int this_rate = brate->bitrate; if (target_rate == this_rate) { - sdata->bss->max_ratectrl_rateidx = i; + sdata->max_ratectrl_rateidx = i; if (rate->fixed) - sdata->bss->force_unicast_rateidx = i; + sdata->force_unicast_rateidx = i; err = 0; break; } -- cgit v1.2.2 From 75636525fbfa78fa33fd754c89785cfde750acd3 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Wed, 9 Jul 2008 14:40:35 +0200 Subject: mac80211: revamp virtual interface handling This patch revamps the virtual interface handling and makes the code much easier to follow. Fewer functions, better names, less spaghetti code. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/cfg.c | 24 +-- net/mac80211/debugfs.c | 15 +- net/mac80211/debugfs_netdev.c | 14 +- net/mac80211/debugfs_netdev.h | 5 - net/mac80211/ieee80211_i.h | 21 +-- net/mac80211/iface.c | 337 +++++++++++++++++++----------------------- net/mac80211/main.c | 18 +-- net/mac80211/wext.c | 3 +- 8 files changed, 172 insertions(+), 265 deletions(-) (limited to 'net') diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c index 3c95cd9bf8ee..6aa49ad172aa 100644 --- a/net/mac80211/cfg.c +++ b/net/mac80211/cfg.c @@ -50,9 +50,6 @@ static int ieee80211_add_iface(struct wiphy *wiphy, char *name, struct ieee80211_sub_if_data *sdata; int err; - if (unlikely(local->reg_state != IEEE80211_DEV_REGISTERED)) - return -ENODEV; - itype = nl80211_type_to_mac80211_type(type); if (itype == IEEE80211_IF_TYPE_INVALID) return -EINVAL; @@ -68,35 +65,26 @@ static int ieee80211_add_iface(struct wiphy *wiphy, char *name, static int ieee80211_del_iface(struct wiphy *wiphy, int ifindex) { - struct ieee80211_local *local = wiphy_priv(wiphy); struct net_device *dev; - char *name; - - if (unlikely(local->reg_state != IEEE80211_DEV_REGISTERED)) - return -ENODEV; /* we're under RTNL */ dev = __dev_get_by_index(&init_net, ifindex); if (!dev) - return 0; + return -ENODEV; - name = dev->name; + ieee80211_if_remove(dev); - return ieee80211_if_remove(local->mdev, name, -1); + return 0; } static int ieee80211_change_iface(struct wiphy *wiphy, int ifindex, enum nl80211_iftype type, u32 *flags, struct vif_params *params) { - struct ieee80211_local *local = wiphy_priv(wiphy); struct net_device *dev; enum ieee80211_if_types itype; struct ieee80211_sub_if_data *sdata; - if (unlikely(local->reg_state != IEEE80211_DEV_REGISTERED)) - return -ENODEV; - /* we're under RTNL */ dev = __dev_get_by_index(&init_net, ifindex); if (!dev) @@ -111,11 +99,7 @@ static int ieee80211_change_iface(struct wiphy *wiphy, int ifindex, sdata = IEEE80211_DEV_TO_SUB_IF(dev); - if (sdata->vif.type == IEEE80211_IF_TYPE_VLAN) - return -EOPNOTSUPP; - - ieee80211_if_reinit(dev); - ieee80211_if_set_type(dev, itype); + ieee80211_if_change_type(sdata, itype); if (ieee80211_vif_is_mesh(&sdata->vif) && params->mesh_id_len) ieee80211_if_sta_set_mesh_id(&sdata->u.sta, diff --git a/net/mac80211/debugfs.c b/net/mac80211/debugfs.c index d20d90eead1f..ee509f1109e2 100644 --- a/net/mac80211/debugfs.c +++ b/net/mac80211/debugfs.c @@ -70,16 +70,6 @@ DEBUGFS_READONLY_FILE(rate_ctrl_alg, 100, "%s", /* statistics stuff */ -static inline int rtnl_lock_local(struct ieee80211_local *local) -{ - rtnl_lock(); - if (unlikely(local->reg_state != IEEE80211_DEV_REGISTERED)) { - rtnl_unlock(); - return -ENODEV; - } - return 0; -} - #define DEBUGFS_STATS_FILE(name, buflen, fmt, value...) \ DEBUGFS_READONLY_FILE(stats_ ##name, buflen, fmt, ##value) @@ -96,10 +86,7 @@ static ssize_t format_devstat_counter(struct ieee80211_local *local, if (!local->ops->get_stats) return -EOPNOTSUPP; - res = rtnl_lock_local(local); - if (res) - return res; - + rtnl_lock(); res = local->ops->get_stats(local_to_hw(local), &stats); rtnl_unlock(); if (!res) diff --git a/net/mac80211/debugfs_netdev.c b/net/mac80211/debugfs_netdev.c index 4aa6621f9970..475f89a8aee1 100644 --- a/net/mac80211/debugfs_netdev.c +++ b/net/mac80211/debugfs_netdev.c @@ -476,12 +476,12 @@ static void del_mesh_config(struct ieee80211_sub_if_data *sdata) } #endif -static void del_files(struct ieee80211_sub_if_data *sdata, int type) +static void del_files(struct ieee80211_sub_if_data *sdata) { if (!sdata->debugfsdir) return; - switch (type) { + switch (sdata->vif.type) { case IEEE80211_IF_TYPE_MESH_POINT: #ifdef CONFIG_MAC80211_MESH del_mesh_stats(sdata); @@ -521,22 +521,16 @@ void ieee80211_debugfs_add_netdev(struct ieee80211_sub_if_data *sdata) sprintf(buf, "netdev:%s", sdata->dev->name); sdata->debugfsdir = debugfs_create_dir(buf, sdata->local->hw.wiphy->debugfsdir); + add_files(sdata); } void ieee80211_debugfs_remove_netdev(struct ieee80211_sub_if_data *sdata) { - del_files(sdata, sdata->vif.type); + del_files(sdata); debugfs_remove(sdata->debugfsdir); sdata->debugfsdir = NULL; } -void ieee80211_debugfs_change_if_type(struct ieee80211_sub_if_data *sdata, - int oldtype) -{ - del_files(sdata, oldtype); - add_files(sdata); -} - static int netdev_notify(struct notifier_block *nb, unsigned long state, void *ndev) diff --git a/net/mac80211/debugfs_netdev.h b/net/mac80211/debugfs_netdev.h index a690071fde8a..7af731f0b731 100644 --- a/net/mac80211/debugfs_netdev.h +++ b/net/mac80211/debugfs_netdev.h @@ -6,8 +6,6 @@ #ifdef CONFIG_MAC80211_DEBUGFS void ieee80211_debugfs_add_netdev(struct ieee80211_sub_if_data *sdata); void ieee80211_debugfs_remove_netdev(struct ieee80211_sub_if_data *sdata); -void ieee80211_debugfs_change_if_type(struct ieee80211_sub_if_data *sdata, - int oldtype); void ieee80211_debugfs_netdev_init(void); void ieee80211_debugfs_netdev_exit(void); #else @@ -17,9 +15,6 @@ static inline void ieee80211_debugfs_add_netdev( static inline void ieee80211_debugfs_remove_netdev( struct ieee80211_sub_if_data *sdata) {} -static inline void ieee80211_debugfs_change_if_type( - struct ieee80211_sub_if_data *sdata, int oldtype) -{} static inline void ieee80211_debugfs_netdev_init(void) {} diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 1b1fc53dad36..35bcdfef9045 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -558,12 +558,6 @@ struct ieee80211_local { bool tim_in_locked_section; /* see ieee80211_beacon_get() */ int tx_headroom; /* required headroom for hardware/radiotap */ - enum { - IEEE80211_DEV_UNINITIALIZED = 0, - IEEE80211_DEV_REGISTERED, - IEEE80211_DEV_UNREGISTERED, - } reg_state; - /* Tasklet and skb queue to process calls from IRQ mode. All frames * added to skb_queue will be processed, but frames in * skb_queue_unreliable may be dropped if the total length of these @@ -863,7 +857,6 @@ int ieee80211_hw_config(struct ieee80211_local *local); int ieee80211_if_config(struct net_device *dev); int ieee80211_if_config_beacon(struct net_device *dev); void ieee80211_tx_set_protected(struct ieee80211_tx_data *tx); -void ieee80211_if_setup(struct net_device *dev); u32 ieee80211_handle_ht(struct ieee80211_local *local, int enable_ht, struct ieee80211_ht_info *req_ht_cap, struct ieee80211_ht_bss_info *req_bss_cap); @@ -933,16 +926,14 @@ static inline void ieee80211_start_mesh(struct net_device *dev) #endif /* interface handling */ +void ieee80211_if_setup(struct net_device *dev); int ieee80211_if_add(struct ieee80211_local *local, const char *name, - struct net_device **new_dev, int type, + struct net_device **new_dev, enum ieee80211_if_types type, struct vif_params *params); -void ieee80211_if_set_type(struct net_device *dev, int type); -void ieee80211_if_reinit(struct net_device *dev); -void __ieee80211_if_del(struct ieee80211_local *local, - struct ieee80211_sub_if_data *sdata); -int ieee80211_if_remove(struct net_device *dev, const char *name, int id); -void ieee80211_if_free(struct net_device *dev); -void ieee80211_if_sdata_init(struct ieee80211_sub_if_data *sdata); +void ieee80211_if_change_type(struct ieee80211_sub_if_data *sdata, + enum ieee80211_if_types type); +void ieee80211_if_remove(struct net_device *dev); +void ieee80211_remove_interfaces(struct ieee80211_local *local); /* tx handling */ void ieee80211_clear_tx_pending(struct ieee80211_local *local); diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c index f2aefd4b8637..6cf121bebd7a 100644 --- a/net/mac80211/iface.c +++ b/net/mac80211/iface.c @@ -2,6 +2,7 @@ * Copyright 2002-2005, Instant802 Networks, Inc. * Copyright 2005-2006, Devicescape Software, Inc. * Copyright (c) 2006 Jiri Benc + * Copyright 2008, Johannes Berg * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License version 2 as @@ -17,39 +18,149 @@ #include "debugfs_netdev.h" #include "mesh.h" -void ieee80211_if_sdata_init(struct ieee80211_sub_if_data *sdata) +/* + * Called when the netdev is removed or, by the code below, before + * the interface type changes. + */ +static void ieee80211_teardown_sdata(struct net_device *dev) { + struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); + struct ieee80211_local *local = sdata->local; + struct beacon_data *beacon; + struct sk_buff *skb; + int flushed; int i; - /* Default values for sub-interface parameters */ - sdata->drop_unencrypted = 0; + ieee80211_debugfs_remove_netdev(sdata); + + /* free extra data */ + ieee80211_free_keys(sdata); + for (i = 0; i < IEEE80211_FRAGMENT_MAX; i++) - skb_queue_head_init(&sdata->fragments[i].skb_list); + __skb_queue_purge(&sdata->fragments[i].skb_list); + sdata->fragment_next = 0; - INIT_LIST_HEAD(&sdata->key_list); + switch (sdata->vif.type) { + case IEEE80211_IF_TYPE_AP: + beacon = sdata->u.ap.beacon; + rcu_assign_pointer(sdata->u.ap.beacon, NULL); + synchronize_rcu(); + kfree(beacon); - sdata->force_unicast_rateidx = -1; - sdata->max_ratectrl_rateidx = -1; + while ((skb = skb_dequeue(&sdata->u.ap.ps_bc_buf))) { + local->total_ps_buffered--; + dev_kfree_skb(skb); + } + + break; + case IEEE80211_IF_TYPE_MESH_POINT: + /* Allow compiler to elide mesh_rmc_free call. */ + if (ieee80211_vif_is_mesh(&sdata->vif)) + mesh_rmc_free(dev); + /* fall through */ + case IEEE80211_IF_TYPE_STA: + case IEEE80211_IF_TYPE_IBSS: + kfree(sdata->u.sta.extra_ie); + kfree(sdata->u.sta.assocreq_ies); + kfree(sdata->u.sta.assocresp_ies); + kfree_skb(sdata->u.sta.probe_resp); + break; + case IEEE80211_IF_TYPE_WDS: + case IEEE80211_IF_TYPE_VLAN: + case IEEE80211_IF_TYPE_MNTR: + break; + case IEEE80211_IF_TYPE_INVALID: + BUG(); + break; + } + + flushed = sta_info_flush(local, sdata); + WARN_ON(flushed); } -static void ieee80211_if_sdata_deinit(struct ieee80211_sub_if_data *sdata) +/* + * Helper function to initialise an interface to a specific type. + */ +static void ieee80211_setup_sdata(struct ieee80211_sub_if_data *sdata, + enum ieee80211_if_types type) { - int i; + struct ieee80211_if_sta *ifsta; - for (i = 0; i < IEEE80211_FRAGMENT_MAX; i++) - __skb_queue_purge(&sdata->fragments[i].skb_list); + /* clear type-dependent union */ + memset(&sdata->u, 0, sizeof(sdata->u)); + + /* and set some type-dependent values */ + sdata->vif.type = type; + + /* only monitor differs */ + sdata->dev->type = ARPHRD_ETHER; + + switch (type) { + case IEEE80211_IF_TYPE_AP: + skb_queue_head_init(&sdata->u.ap.ps_bc_buf); + INIT_LIST_HEAD(&sdata->u.ap.vlans); + break; + case IEEE80211_IF_TYPE_MESH_POINT: + case IEEE80211_IF_TYPE_STA: + case IEEE80211_IF_TYPE_IBSS: + ifsta = &sdata->u.sta; + INIT_WORK(&ifsta->work, ieee80211_sta_work); + setup_timer(&ifsta->timer, ieee80211_sta_timer, + (unsigned long) sdata); + skb_queue_head_init(&ifsta->skb_queue); + + ifsta->capab = WLAN_CAPABILITY_ESS; + ifsta->auth_algs = IEEE80211_AUTH_ALG_OPEN | + IEEE80211_AUTH_ALG_SHARED_KEY; + ifsta->flags |= IEEE80211_STA_CREATE_IBSS | + IEEE80211_STA_AUTO_BSSID_SEL | + IEEE80211_STA_AUTO_CHANNEL_SEL; + if (ieee80211_num_regular_queues(&sdata->local->hw) >= 4) + ifsta->flags |= IEEE80211_STA_WMM_ENABLED; + + if (ieee80211_vif_is_mesh(&sdata->vif)) + ieee80211_mesh_init_sdata(sdata); + break; + case IEEE80211_IF_TYPE_MNTR: + sdata->dev->type = ARPHRD_IEEE80211_RADIOTAP; + sdata->dev->hard_start_xmit = ieee80211_monitor_start_xmit; + sdata->u.mntr_flags = MONITOR_FLAG_CONTROL | + MONITOR_FLAG_OTHER_BSS; + break; + case IEEE80211_IF_TYPE_WDS: + case IEEE80211_IF_TYPE_VLAN: + break; + case IEEE80211_IF_TYPE_INVALID: + BUG(); + break; + } + + ieee80211_debugfs_add_netdev(sdata); +} + +void ieee80211_if_change_type(struct ieee80211_sub_if_data *sdata, + enum ieee80211_if_types type) +{ + /* Purge and reset type-dependent state. */ + ieee80211_teardown_sdata(sdata->dev); + ieee80211_setup_sdata(sdata, type); + + /* reset some values that shouldn't be kept across type changes */ + sdata->basic_rates = 0; + sdata->drop_unencrypted = 0; + sdata->sequence = 0; } -/* Must be called with rtnl lock held. */ int ieee80211_if_add(struct ieee80211_local *local, const char *name, - struct net_device **new_dev, int type, + struct net_device **new_dev, enum ieee80211_if_types type, struct vif_params *params) { struct net_device *ndev; struct ieee80211_sub_if_data *sdata = NULL; - int ret; + int ret, i; ASSERT_RTNL(); + ndev = alloc_netdev(sizeof(*sdata) + local->hw.vif_data_size, name, ieee80211_if_setup); if (!ndev) @@ -74,18 +185,28 @@ int ieee80211_if_add(struct ieee80211_local *local, const char *name, /* don't use IEEE80211_DEV_TO_SUB_IF because it checks too much */ sdata = netdev_priv(ndev); ndev->ieee80211_ptr = &sdata->wdev; + + /* initialise type-independent data */ sdata->wdev.wiphy = local->hw.wiphy; - sdata->vif.type = IEEE80211_IF_TYPE_AP; - sdata->dev = ndev; sdata->local = local; - ieee80211_if_sdata_init(sdata); + sdata->dev = ndev; + + for (i = 0; i < IEEE80211_FRAGMENT_MAX; i++) + skb_queue_head_init(&sdata->fragments[i].skb_list); + + INIT_LIST_HEAD(&sdata->key_list); + + sdata->force_unicast_rateidx = -1; + sdata->max_ratectrl_rateidx = -1; + + /* setup type-dependent data */ + ieee80211_setup_sdata(sdata, type); ret = register_netdevice(ndev); if (ret) goto fail; - ieee80211_debugfs_add_netdev(sdata); - ieee80211_if_set_type(ndev, type); + ndev->uninit = ieee80211_teardown_sdata; if (ieee80211_vif_is_mesh(&sdata->vif) && params && params->mesh_id_len) @@ -93,11 +214,6 @@ int ieee80211_if_add(struct ieee80211_local *local, const char *name, params->mesh_id_len, params->mesh_id); - /* we're under RTNL so all this is fine */ - if (unlikely(local->reg_state == IEEE80211_DEV_UNREGISTERED)) { - __ieee80211_if_del(local, sdata); - return -ENODEV; - } list_add_tail_rcu(&sdata->list, &local->interfaces); if (new_dev) @@ -105,181 +221,34 @@ int ieee80211_if_add(struct ieee80211_local *local, const char *name, return 0; -fail: + fail: free_netdev(ndev); return ret; } -void ieee80211_if_set_type(struct net_device *dev, int type) +void ieee80211_if_remove(struct net_device *dev) { struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); - int oldtype = sdata->vif.type; - - /* - * Called even when register_netdevice fails, it would - * oops if assigned before initialising the rest. - */ - dev->uninit = ieee80211_if_reinit; - - /* most have no BSS pointer */ - sdata->bss = NULL; - sdata->vif.type = type; - - sdata->basic_rates = 0; - - switch (type) { - case IEEE80211_IF_TYPE_WDS: - case IEEE80211_IF_TYPE_VLAN: - /* nothing special */ - break; - case IEEE80211_IF_TYPE_AP: - skb_queue_head_init(&sdata->u.ap.ps_bc_buf); - INIT_LIST_HEAD(&sdata->u.ap.vlans); - break; - case IEEE80211_IF_TYPE_MESH_POINT: - case IEEE80211_IF_TYPE_STA: - case IEEE80211_IF_TYPE_IBSS: { - struct ieee80211_if_sta *ifsta; - - ifsta = &sdata->u.sta; - INIT_WORK(&ifsta->work, ieee80211_sta_work); - setup_timer(&ifsta->timer, ieee80211_sta_timer, - (unsigned long) sdata); - skb_queue_head_init(&ifsta->skb_queue); - - ifsta->capab = WLAN_CAPABILITY_ESS; - ifsta->auth_algs = IEEE80211_AUTH_ALG_OPEN | - IEEE80211_AUTH_ALG_SHARED_KEY; - ifsta->flags |= IEEE80211_STA_CREATE_IBSS | - IEEE80211_STA_AUTO_BSSID_SEL | - IEEE80211_STA_AUTO_CHANNEL_SEL; - if (ieee80211_num_regular_queues(&sdata->local->hw) >= 4) - ifsta->flags |= IEEE80211_STA_WMM_ENABLED; - - if (ieee80211_vif_is_mesh(&sdata->vif)) - ieee80211_mesh_init_sdata(sdata); - break; - } - case IEEE80211_IF_TYPE_MNTR: - dev->type = ARPHRD_IEEE80211_RADIOTAP; - dev->hard_start_xmit = ieee80211_monitor_start_xmit; - sdata->u.mntr_flags = MONITOR_FLAG_CONTROL | - MONITOR_FLAG_OTHER_BSS; - break; - case IEEE80211_IF_TYPE_INVALID: - BUG(); - break; - } - ieee80211_debugfs_change_if_type(sdata, oldtype); -} - -/* Must be called with rtnl lock held. */ -void ieee80211_if_reinit(struct net_device *dev) -{ - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); - struct sk_buff *skb; - int flushed; ASSERT_RTNL(); - ieee80211_free_keys(sdata); - - ieee80211_if_sdata_deinit(sdata); - - /* Need to handle mesh specially to allow eliding the function call */ - if (ieee80211_vif_is_mesh(&sdata->vif)) - mesh_rmc_free(dev); - - switch (sdata->vif.type) { - case IEEE80211_IF_TYPE_INVALID: - /* cannot happen */ - WARN_ON(1); - break; - case IEEE80211_IF_TYPE_AP: { - struct beacon_data *beacon; - - beacon = sdata->u.ap.beacon; - rcu_assign_pointer(sdata->u.ap.beacon, NULL); - synchronize_rcu(); - kfree(beacon); - - while ((skb = skb_dequeue(&sdata->u.ap.ps_bc_buf))) { - local->total_ps_buffered--; - dev_kfree_skb(skb); - } - - break; - } - case IEEE80211_IF_TYPE_WDS: - case IEEE80211_IF_TYPE_VLAN: - /* nothing to do */ - break; - case IEEE80211_IF_TYPE_MESH_POINT: - case IEEE80211_IF_TYPE_STA: - case IEEE80211_IF_TYPE_IBSS: - kfree(sdata->u.sta.extra_ie); - sdata->u.sta.extra_ie = NULL; - kfree(sdata->u.sta.assocreq_ies); - sdata->u.sta.assocreq_ies = NULL; - kfree(sdata->u.sta.assocresp_ies); - sdata->u.sta.assocresp_ies = NULL; - if (sdata->u.sta.probe_resp) { - dev_kfree_skb(sdata->u.sta.probe_resp); - sdata->u.sta.probe_resp = NULL; - } - - break; - case IEEE80211_IF_TYPE_MNTR: - dev->type = ARPHRD_ETHER; - break; - } - - flushed = sta_info_flush(local, sdata); - WARN_ON(flushed); - - memset(&sdata->u, 0, sizeof(sdata->u)); - ieee80211_if_sdata_init(sdata); -} - -/* Must be called with rtnl lock held. */ -void __ieee80211_if_del(struct ieee80211_local *local, - struct ieee80211_sub_if_data *sdata) -{ - struct net_device *dev = sdata->dev; - - ieee80211_debugfs_remove_netdev(sdata); + list_del_rcu(&sdata->list); + synchronize_rcu(); unregister_netdevice(dev); - /* - * The net_device will be freed by its destructor, - * i.e. ieee80211_if_free. - */ } -/* Must be called with rtnl lock held. */ -int ieee80211_if_remove(struct net_device *dev, const char *name, int id) +/* + * Remove all interfaces, may only be called at hardware unregistration + * time because it doesn't do RCU-safe list removals. + */ +void ieee80211_remove_interfaces(struct ieee80211_local *local) { - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee80211_sub_if_data *sdata, *n; + struct ieee80211_sub_if_data *sdata, *tmp; ASSERT_RTNL(); - list_for_each_entry_safe(sdata, n, &local->interfaces, list) { - if ((sdata->vif.type == id || id == -1) && - strcmp(name, sdata->dev->name) == 0) { - list_del_rcu(&sdata->list); - synchronize_rcu(); - __ieee80211_if_del(local, sdata); - return 0; - } + list_for_each_entry_safe(sdata, tmp, &local->interfaces, list) { + list_del(&sdata->list); + unregister_netdevice(sdata->dev); } - return -ENODEV; -} - -void ieee80211_if_free(struct net_device *dev) -{ - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); - - ieee80211_if_sdata_deinit(sdata); - free_netdev(dev); } diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 863923964d21..0759ab2ca3ff 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -971,7 +971,6 @@ static const struct header_ops ieee80211_header_ops = { .cache_update = eth_header_cache_update, }; -/* Must not be called for mdev */ void ieee80211_if_setup(struct net_device *dev) { ether_setup(dev); @@ -981,7 +980,7 @@ void ieee80211_if_setup(struct net_device *dev) dev->change_mtu = ieee80211_change_mtu; dev->open = ieee80211_open; dev->stop = ieee80211_stop; - dev->destructor = ieee80211_if_free; + dev->destructor = free_netdev; } /* everything else */ @@ -1776,7 +1775,6 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) printk(KERN_WARNING "%s: Failed to add default virtual iface\n", wiphy_name(local->hw.wiphy)); - local->reg_state = IEEE80211_DEV_REGISTERED; rtnl_unlock(); ieee80211_led_init(local); @@ -1806,30 +1804,20 @@ EXPORT_SYMBOL(ieee80211_register_hw); void ieee80211_unregister_hw(struct ieee80211_hw *hw) { struct ieee80211_local *local = hw_to_local(hw); - struct ieee80211_sub_if_data *sdata, *tmp; tasklet_kill(&local->tx_pending_tasklet); tasklet_kill(&local->tasklet); rtnl_lock(); - BUG_ON(local->reg_state != IEEE80211_DEV_REGISTERED); - - local->reg_state = IEEE80211_DEV_UNREGISTERED; - /* * At this point, interface list manipulations are fine * because the driver cannot be handing us frames any * more and the tasklet is killed. */ - /* - * First, we remove all virtual interfaces. - */ - list_for_each_entry_safe(sdata, tmp, &local->interfaces, list) { - list_del(&sdata->list); - __ieee80211_if_del(local, sdata); - } + /* First, we remove all virtual interfaces. */ + ieee80211_remove_interfaces(local); /* then, finally, remove the master interface */ unregister_netdevice(local->mdev); diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c index e8e4a6215e89..f2fdd3342195 100644 --- a/net/mac80211/wext.c +++ b/net/mac80211/wext.c @@ -301,8 +301,7 @@ static int ieee80211_ioctl_siwmode(struct net_device *dev, if (netif_running(dev)) return -EBUSY; - ieee80211_if_reinit(dev); - ieee80211_if_set_type(dev, type); + ieee80211_if_change_type(sdata, type); return 0; } -- cgit v1.2.2 From f3947e2dfa3b18f375b7acd03b7ee2877d0751fc Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Wed, 9 Jul 2008 14:40:36 +0200 Subject: mac80211: push interface checks down This patch pushes the "netif_running()" and "same type as before" checks down into ieee80211_if_change_type() to centralise the logic instead of duplicating it for cfg80211 and wext. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/cfg.c | 8 ++++---- net/mac80211/ieee80211_i.h | 4 ++-- net/mac80211/iface.c | 20 ++++++++++++++++++-- net/mac80211/wext.c | 9 +-------- 4 files changed, 25 insertions(+), 16 deletions(-) (limited to 'net') diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c index 6aa49ad172aa..ea0301025c15 100644 --- a/net/mac80211/cfg.c +++ b/net/mac80211/cfg.c @@ -84,22 +84,22 @@ static int ieee80211_change_iface(struct wiphy *wiphy, int ifindex, struct net_device *dev; enum ieee80211_if_types itype; struct ieee80211_sub_if_data *sdata; + int ret; /* we're under RTNL */ dev = __dev_get_by_index(&init_net, ifindex); if (!dev) return -ENODEV; - if (netif_running(dev)) - return -EBUSY; - itype = nl80211_type_to_mac80211_type(type); if (itype == IEEE80211_IF_TYPE_INVALID) return -EINVAL; sdata = IEEE80211_DEV_TO_SUB_IF(dev); - ieee80211_if_change_type(sdata, itype); + ret = ieee80211_if_change_type(sdata, itype); + if (ret) + return ret; if (ieee80211_vif_is_mesh(&sdata->vif) && params->mesh_id_len) ieee80211_if_sta_set_mesh_id(&sdata->u.sta, diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 35bcdfef9045..2146c0c436d2 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -930,8 +930,8 @@ void ieee80211_if_setup(struct net_device *dev); int ieee80211_if_add(struct ieee80211_local *local, const char *name, struct net_device **new_dev, enum ieee80211_if_types type, struct vif_params *params); -void ieee80211_if_change_type(struct ieee80211_sub_if_data *sdata, - enum ieee80211_if_types type); +int ieee80211_if_change_type(struct ieee80211_sub_if_data *sdata, + enum ieee80211_if_types type); void ieee80211_if_remove(struct net_device *dev); void ieee80211_remove_interfaces(struct ieee80211_local *local); diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c index 6cf121bebd7a..2e3adcb3ce23 100644 --- a/net/mac80211/iface.c +++ b/net/mac80211/iface.c @@ -138,9 +138,23 @@ static void ieee80211_setup_sdata(struct ieee80211_sub_if_data *sdata, ieee80211_debugfs_add_netdev(sdata); } -void ieee80211_if_change_type(struct ieee80211_sub_if_data *sdata, - enum ieee80211_if_types type) +int ieee80211_if_change_type(struct ieee80211_sub_if_data *sdata, + enum ieee80211_if_types type) { + ASSERT_RTNL(); + + if (type == sdata->vif.type) + return 0; + + /* + * We could, here, on changes between IBSS/STA/MESH modes, + * invoke an MLME function instead that disassociates etc. + * and goes into the requested mode. + */ + + if (netif_running(sdata->dev)) + return -EBUSY; + /* Purge and reset type-dependent state. */ ieee80211_teardown_sdata(sdata->dev); ieee80211_setup_sdata(sdata, type); @@ -149,6 +163,8 @@ void ieee80211_if_change_type(struct ieee80211_sub_if_data *sdata, sdata->basic_rates = 0; sdata->drop_unencrypted = 0; sdata->sequence = 0; + + return 0; } int ieee80211_if_add(struct ieee80211_local *local, const char *name, diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c index f2fdd3342195..c041db9556c7 100644 --- a/net/mac80211/wext.c +++ b/net/mac80211/wext.c @@ -296,14 +296,7 @@ static int ieee80211_ioctl_siwmode(struct net_device *dev, return -EINVAL; } - if (type == sdata->vif.type) - return 0; - if (netif_running(dev)) - return -EBUSY; - - ieee80211_if_change_type(sdata, type); - - return 0; + return ieee80211_if_change_type(sdata, type); } -- cgit v1.2.2 From 9d139c810a2aa17365cc548d0cd2a189d8433c65 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Wed, 9 Jul 2008 14:40:37 +0200 Subject: mac80211: revamp beacon configuration This patch changes mac80211's beacon configuration handling to never pass skbs to the driver directly but rather always require the driver to use ieee80211_beacon_get(). Additionally, it introduces "change flags" on the config_interface() call to enable drivers to figure out what is changing. Finally, it removes the beacon_update() driver callback in favour of having IBSS beacon delivered by ieee80211_beacon_get() as well. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/cfg.c | 4 +- net/mac80211/ieee80211_i.h | 3 +- net/mac80211/main.c | 59 ++++++++++-------------- net/mac80211/mlme.c | 112 ++++++++++++++++++--------------------------- net/mac80211/tx.c | 36 +++++++++------ net/mac80211/wext.c | 2 +- 6 files changed, 95 insertions(+), 121 deletions(-) (limited to 'net') diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c index ea0301025c15..8e7ba0e62cf5 100644 --- a/net/mac80211/cfg.c +++ b/net/mac80211/cfg.c @@ -469,7 +469,7 @@ static int ieee80211_config_beacon(struct ieee80211_sub_if_data *sdata, kfree(old); - return ieee80211_if_config_beacon(sdata->dev); + return ieee80211_if_config(sdata, IEEE80211_IFCC_BEACON); } static int ieee80211_add_beacon(struct wiphy *wiphy, struct net_device *dev, @@ -523,7 +523,7 @@ static int ieee80211_del_beacon(struct wiphy *wiphy, struct net_device *dev) synchronize_rcu(); kfree(old); - return ieee80211_if_config_beacon(dev); + return ieee80211_if_config(sdata, IEEE80211_IFCC_BEACON); } /* Layer 2 Update frame (802.2 Type 1 LLC XID Update response) */ diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 2146c0c436d2..934c3ef4f0bc 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -854,8 +854,7 @@ static inline int ieee80211_bssid_match(const u8 *raddr, const u8 *addr) /* ieee80211.c */ int ieee80211_hw_config(struct ieee80211_local *local); -int ieee80211_if_config(struct net_device *dev); -int ieee80211_if_config_beacon(struct net_device *dev); +int ieee80211_if_config(struct ieee80211_sub_if_data *sdata, u32 changed); void ieee80211_tx_set_protected(struct ieee80211_tx_data *tx); u32 ieee80211_handle_ht(struct ieee80211_local *local, int enable_ht, struct ieee80211_ht_info *req_ht_cap, diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 0759ab2ca3ff..36859e794928 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -307,7 +307,8 @@ static int ieee80211_open(struct net_device *dev) if (res) goto err_stop; - ieee80211_if_config(dev); + if (ieee80211_vif_is_mesh(&sdata->vif)) + ieee80211_start_mesh(sdata->dev); changed |= ieee80211_reset_erp_info(dev); ieee80211_bss_info_change_notify(sdata, changed); ieee80211_enable_keys(sdata); @@ -985,57 +986,47 @@ void ieee80211_if_setup(struct net_device *dev) /* everything else */ -static int __ieee80211_if_config(struct net_device *dev, - struct sk_buff *beacon) +int ieee80211_if_config(struct ieee80211_sub_if_data *sdata, u32 changed) { - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); + struct ieee80211_local *local = sdata->local; struct ieee80211_if_conf conf; - if (!local->ops->config_interface || !netif_running(dev)) + if (WARN_ON(!netif_running(sdata->dev))) + return 0; + + if (!local->ops->config_interface) return 0; memset(&conf, 0, sizeof(conf)); - conf.type = sdata->vif.type; + conf.changed = changed; + if (sdata->vif.type == IEEE80211_IF_TYPE_STA || sdata->vif.type == IEEE80211_IF_TYPE_IBSS) { conf.bssid = sdata->u.sta.bssid; conf.ssid = sdata->u.sta.ssid; conf.ssid_len = sdata->u.sta.ssid_len; - } else if (ieee80211_vif_is_mesh(&sdata->vif)) { - conf.beacon = beacon; - ieee80211_start_mesh(dev); } else if (sdata->vif.type == IEEE80211_IF_TYPE_AP) { + conf.bssid = sdata->dev->dev_addr; conf.ssid = sdata->u.ap.ssid; conf.ssid_len = sdata->u.ap.ssid_len; - conf.beacon = beacon; + } else if (ieee80211_vif_is_mesh(&sdata->vif)) { + u8 zero[ETH_ALEN] = { 0 }; + conf.bssid = zero; + conf.ssid = zero; + conf.ssid_len = 0; + } else { + WARN_ON(1); + return -EINVAL; } - return local->ops->config_interface(local_to_hw(local), - &sdata->vif, &conf); -} -int ieee80211_if_config(struct net_device *dev) -{ - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - if (sdata->vif.type == IEEE80211_IF_TYPE_MESH_POINT && - (local->hw.flags & IEEE80211_HW_HOST_GEN_BEACON_TEMPLATE)) - return ieee80211_if_config_beacon(dev); - return __ieee80211_if_config(dev, NULL); -} + if (WARN_ON(!conf.bssid && (changed & IEEE80211_IFCC_BSSID))) + return -EINVAL; -int ieee80211_if_config_beacon(struct net_device *dev) -{ - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); - struct sk_buff *skb; + if (WARN_ON(!conf.ssid && (changed & IEEE80211_IFCC_SSID))) + return -EINVAL; - if (!(local->hw.flags & IEEE80211_HW_HOST_GEN_BEACON_TEMPLATE)) - return 0; - skb = ieee80211_beacon_get(local_to_hw(local), &sdata->vif); - if (!skb) - return -ENOMEM; - return __ieee80211_if_config(dev, skb); + return local->ops->config_interface(local_to_hw(local), + &sdata->vif, &conf); } int ieee80211_hw_config(struct ieee80211_local *local) diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 64d710a88b86..61d7f81bf45e 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -2406,8 +2406,6 @@ static int ieee80211_sta_join_ibss(struct net_device *dev, int res, rates, i, j; struct sk_buff *skb; struct ieee80211_mgmt *mgmt; - struct ieee80211_tx_info *control; - struct rate_selection ratesel; u8 *pos; struct ieee80211_sub_if_data *sdata; struct ieee80211_supported_band *sband; @@ -2425,7 +2423,7 @@ static int ieee80211_sta_join_ibss(struct net_device *dev, local->ops->reset_tsf(local_to_hw(local)); } memcpy(ifsta->bssid, bss->bssid, ETH_ALEN); - res = ieee80211_if_config(dev); + res = ieee80211_if_config(sdata, IEEE80211_IFCC_BSSID); if (res) return res; @@ -2439,19 +2437,16 @@ static int ieee80211_sta_join_ibss(struct net_device *dev, if (res) return res; - /* Set beacon template */ + /* Build IBSS probe response */ skb = dev_alloc_skb(local->hw.extra_tx_headroom + 400); - do { - if (!skb) - break; - + if (skb) { skb_reserve(skb, local->hw.extra_tx_headroom); mgmt = (struct ieee80211_mgmt *) skb_put(skb, 24 + sizeof(mgmt->u.beacon)); memset(mgmt, 0, 24 + sizeof(mgmt->u.beacon)); mgmt->frame_control = IEEE80211_FC(IEEE80211_FTYPE_MGMT, - IEEE80211_STYPE_BEACON); + IEEE80211_STYPE_PROBE_RESP); memset(mgmt->da, 0xff, ETH_ALEN); memcpy(mgmt->sa, dev->dev_addr, ETH_ALEN); memcpy(mgmt->bssid, ifsta->bssid, ETH_ALEN); @@ -2495,61 +2490,22 @@ static int ieee80211_sta_join_ibss(struct net_device *dev, memcpy(pos, &bss->supp_rates[8], rates); } - control = IEEE80211_SKB_CB(skb); - - rate_control_get_rate(dev, sband, skb, &ratesel); - if (ratesel.rate_idx < 0) { - printk(KERN_DEBUG "%s: Failed to determine TX rate " - "for IBSS beacon\n", dev->name); - break; - } - control->control.vif = &sdata->vif; - control->tx_rate_idx = ratesel.rate_idx; - if (sdata->bss_conf.use_short_preamble && - sband->bitrates[ratesel.rate_idx].flags & IEEE80211_RATE_SHORT_PREAMBLE) - control->flags |= IEEE80211_TX_CTL_SHORT_PREAMBLE; - control->antenna_sel_tx = local->hw.conf.antenna_sel_tx; - control->flags |= IEEE80211_TX_CTL_NO_ACK; - control->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; - control->control.retry_limit = 1; - - ifsta->probe_resp = skb_copy(skb, GFP_ATOMIC); - if (ifsta->probe_resp) { - mgmt = (struct ieee80211_mgmt *) - ifsta->probe_resp->data; - mgmt->frame_control = - IEEE80211_FC(IEEE80211_FTYPE_MGMT, - IEEE80211_STYPE_PROBE_RESP); - } else { - printk(KERN_DEBUG "%s: Could not allocate ProbeResp " - "template for IBSS\n", dev->name); - } - - if (local->ops->beacon_update && - local->ops->beacon_update(local_to_hw(local), skb) == 0) { - printk(KERN_DEBUG "%s: Configured IBSS beacon " - "template\n", dev->name); - skb = NULL; - } - - rates = 0; - sband = local->hw.wiphy->bands[local->hw.conf.channel->band]; - for (i = 0; i < bss->supp_rates_len; i++) { - int bitrate = (bss->supp_rates[i] & 0x7f) * 5; - for (j = 0; j < sband->n_bitrates; j++) - if (sband->bitrates[j].bitrate == bitrate) - rates |= BIT(j); - } - ifsta->supp_rates_bits[local->hw.conf.channel->band] = rates; + ifsta->probe_resp = skb; - ieee80211_sta_def_wmm_params(dev, bss, 1); - } while (0); + ieee80211_if_config(sdata, IEEE80211_IFCC_BEACON); + } - if (skb) { - printk(KERN_DEBUG "%s: Failed to configure IBSS beacon " - "template\n", dev->name); - dev_kfree_skb(skb); + rates = 0; + sband = local->hw.wiphy->bands[local->hw.conf.channel->band]; + for (i = 0; i < bss->supp_rates_len; i++) { + int bitrate = (bss->supp_rates[i] & 0x7f) * 5; + for (j = 0; j < sband->n_bitrates; j++) + if (sband->bitrates[j].bitrate == bitrate) + rates |= BIT(j); } + ifsta->supp_rates_bits[local->hw.conf.channel->band] = rates; + + ieee80211_sta_def_wmm_params(dev, bss, 1); ifsta->state = IEEE80211_IBSS_JOINED; mod_timer(&ifsta->timer, jiffies + IEEE80211_IBSS_MERGE_INTERVAL); @@ -3333,7 +3289,7 @@ static void ieee80211_mesh_housekeeping(struct net_device *dev, free_plinks = mesh_plink_availables(sdata); if (free_plinks != sdata->u.sta.accepting_plinks) - ieee80211_if_config_beacon(dev); + ieee80211_if_config(sdata, IEEE80211_IFCC_BEACON); mod_timer(&ifsta->timer, jiffies + IEEE80211_MESH_HOUSEKEEPING_INTERVAL); @@ -3757,28 +3713,45 @@ int ieee80211_sta_set_ssid(struct net_device *dev, char *ssid, size_t len) { struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev); struct ieee80211_if_sta *ifsta; + int res; if (len > IEEE80211_MAX_SSID_LEN) return -EINVAL; ifsta = &sdata->u.sta; - if (ifsta->ssid_len != len || memcmp(ifsta->ssid, ssid, len) != 0) + if (ifsta->ssid_len != len || memcmp(ifsta->ssid, ssid, len) != 0) { + memset(ifsta->ssid, 0, sizeof(ifsta->ssid)); + memcpy(ifsta->ssid, ssid, len); + ifsta->ssid_len = len; ifsta->flags &= ~IEEE80211_STA_PREV_BSSID_SET; - memcpy(ifsta->ssid, ssid, len); - memset(ifsta->ssid + len, 0, IEEE80211_MAX_SSID_LEN - len); - ifsta->ssid_len = len; + + res = 0; + /* + * Hack! MLME code needs to be cleaned up to have different + * entry points for configuration and internal selection change + */ + if (netif_running(sdata->dev)) + res = ieee80211_if_config(sdata, IEEE80211_IFCC_SSID); + if (res) { + printk(KERN_DEBUG "%s: Failed to config new SSID to " + "the low-level driver\n", dev->name); + return res; + } + } if (len) ifsta->flags |= IEEE80211_STA_SSID_SET; else ifsta->flags &= ~IEEE80211_STA_SSID_SET; + if (sdata->vif.type == IEEE80211_IF_TYPE_IBSS && !(ifsta->flags & IEEE80211_STA_BSSID_SET)) { ifsta->ibss_join_req = jiffies; ifsta->state = IEEE80211_IBSS_SEARCH; return ieee80211_sta_find_ibss(dev, ifsta); } + return 0; } @@ -3804,7 +3777,12 @@ int ieee80211_sta_set_bssid(struct net_device *dev, u8 *bssid) if (memcmp(ifsta->bssid, bssid, ETH_ALEN) != 0) { memcpy(ifsta->bssid, bssid, ETH_ALEN); - res = ieee80211_if_config(dev); + res = 0; + /* + * Hack! See also ieee80211_sta_set_ssid. + */ + if (netif_running(sdata->dev)) + res = ieee80211_if_config(sdata, IEEE80211_IFCC_BSSID); if (res) { printk(KERN_DEBUG "%s: Failed to config new BSSID to " "the low-level driver\n", dev->name); diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index a757dcc1208d..8843416e1460 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -1788,17 +1788,17 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, struct ieee80211_vif *vif) { struct ieee80211_local *local = hw_to_local(hw); - struct sk_buff *skb; + struct sk_buff *skb = NULL; struct ieee80211_tx_info *info; struct net_device *bdev; struct ieee80211_sub_if_data *sdata = NULL; struct ieee80211_if_ap *ap = NULL; + struct ieee80211_if_sta *ifsta = NULL; struct rate_selection rsel; struct beacon_data *beacon; struct ieee80211_supported_band *sband; struct ieee80211_mgmt *mgmt; int *num_beacons; - bool err = true; enum ieee80211_band band = local->hw.conf.channel->band; u8 *pos; @@ -1852,9 +1852,24 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, beacon->tail, beacon->tail_len); num_beacons = &ap->num_beacons; + } else + goto out; + } else if (sdata->vif.type == IEEE80211_IF_TYPE_IBSS) { + struct ieee80211_hdr *hdr; + ifsta = &sdata->u.sta; - err = false; - } + if (!ifsta->probe_resp) + goto out; + + skb = skb_copy(ifsta->probe_resp, GFP_ATOMIC); + if (!skb) + goto out; + + hdr = (struct ieee80211_hdr *) skb->data; + hdr->frame_control = IEEE80211_FC(IEEE80211_FTYPE_MGMT, + IEEE80211_STYPE_BEACON); + + num_beacons = &ifsta->num_beacons; } else if (ieee80211_vif_is_mesh(&sdata->vif)) { /* headroom, head length, tail length and maximum TIM length */ skb = dev_alloc_skb(local->tx_headroom + 400); @@ -1881,17 +1896,8 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, mesh_mgmt_ies_add(skb, sdata->dev); num_beacons = &sdata->u.sta.num_beacons; - - err = false; - } - - if (err) { -#ifdef CONFIG_MAC80211_VERBOSE_DEBUG - if (net_ratelimit()) - printk(KERN_DEBUG "no beacon data avail for %s\n", - bdev->name); -#endif /* CONFIG_MAC80211_VERBOSE_DEBUG */ - skb = NULL; + } else { + WARN_ON(1); goto out; } diff --git a/net/mac80211/wext.c b/net/mac80211/wext.c index c041db9556c7..34fa8ed1e784 100644 --- a/net/mac80211/wext.c +++ b/net/mac80211/wext.c @@ -444,7 +444,7 @@ static int ieee80211_ioctl_siwessid(struct net_device *dev, memset(sdata->u.ap.ssid + len, 0, IEEE80211_MAX_SSID_LEN - len); sdata->u.ap.ssid_len = len; - return ieee80211_if_config(dev); + return ieee80211_if_config(sdata, IEEE80211_IFCC_SSID); } return -EOPNOTSUPP; } -- cgit v1.2.2 From 2560b6e2e4b87df211ea39b3b02498959b70b4e8 Mon Sep 17 00:00:00 2001 From: Emmanuel Grumbach Date: Thu, 10 Jul 2008 00:47:19 +0300 Subject: mac80211: Fix ieee80211_rx_reorder_ampdu: ignore QoS null packets This patch fixes the check at the entrance to ieee80211_rx_reorder_ampdu. This check has been broken by 'mac80211: rx.c use new helpers'. Letting QoS NULL packet in ieee80211_rx_reorder_ampdu led to packet loss in RX. Signed-off-by: Emmanuel Grumbach Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/rx.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index 244ee2d50a58..ba332c9dc19a 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -2041,8 +2041,8 @@ static u8 ieee80211_rx_reorder_ampdu(struct ieee80211_local *local, tid_agg_rx = sta->ampdu_mlme.tid_rx[tid]; - /* null data frames are excluded */ - if (unlikely(ieee80211_is_nullfunc(hdr->frame_control))) + /* qos null data frames are excluded */ + if (unlikely(hdr->frame_control & cpu_to_le16(IEEE80211_STYPE_NULLFUNC))) goto end_reorder; /* new un-ordered ampdu frame - process it */ -- cgit v1.2.2 From 1411f9b531f0a910cd1c85a337737c1e6ffbae6a Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 10 Jul 2008 10:11:02 +0200 Subject: mac80211: fix RX sequence number check According to 802.11-2007, we are doing the wrong thing in the sequence number checks when receiving frames. This fixes it. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/rx.c | 19 ++++++++++++------- 1 file changed, 12 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c index ba332c9dc19a..6d9ae67c27ca 100644 --- a/net/mac80211/rx.c +++ b/net/mac80211/rx.c @@ -334,13 +334,18 @@ static void ieee80211_parse_qos(struct ieee80211_rx_data *rx) else rx->flags &= ~IEEE80211_RX_AMSDU; } else { - if (unlikely(ieee80211_is_mgmt(hdr->frame_control))) { - /* Separate TID for management frames */ - tid = NUM_RX_DATA_QUEUES - 1; - } else { - /* no qos control present */ - tid = 0; /* 802.1d - Best Effort */ - } + /* + * IEEE 802.11-2007, 7.1.3.4.1 ("Sequence Number field"): + * + * Sequence numbers for management frames, QoS data + * frames with a broadcast/multicast address in the + * Address 1 field, and all non-QoS data frames sent + * by QoS STAs are assigned using an additional single + * modulo-4096 counter, [...] + * + * We also use that counter for non-QoS STAs. + */ + tid = NUM_RX_DATA_QUEUES - 1; } rx->queue = tid; -- cgit v1.2.2 From 22bb1be4d271961846cd0889b0f8d671db773080 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 10 Jul 2008 11:16:47 +0200 Subject: wext: make sysfs bits optional and deprecate them The /sys/class/net/*/wireless/ direcory is, as far as I know, not used by anyone. Additionally, the same data is available via wext ioctls. Hence the sysfs files are pretty much useless. This patch makes them optional and schedules them for removal. Signed-off-by: Johannes Berg Cc: Jean Tourrilhes Signed-off-by: John W. Linville --- net/core/net-sysfs.c | 4 ++-- net/wireless/Kconfig | 11 +++++++++++ 2 files changed, 13 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index 3f7941319217..c1f4e0d428c0 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -318,7 +318,7 @@ static struct attribute_group netstat_group = { .attrs = netstat_attrs, }; -#ifdef CONFIG_WIRELESS_EXT +#ifdef CONFIG_WIRELESS_EXT_SYSFS /* helper function that does all the locking etc for wireless stats */ static ssize_t wireless_show(struct device *d, char *buf, ssize_t (*format)(const struct iw_statistics *, @@ -459,7 +459,7 @@ int netdev_register_kobject(struct net_device *net) #ifdef CONFIG_SYSFS *groups++ = &netstat_group; -#ifdef CONFIG_WIRELESS_EXT +#ifdef CONFIG_WIRELESS_EXT_SYSFS if (net->wireless_handlers && net->wireless_handlers->get_wireless_stats) *groups++ = &wireless_group; #endif diff --git a/net/wireless/Kconfig b/net/wireless/Kconfig index 79270903bda6..ab015c62d561 100644 --- a/net/wireless/Kconfig +++ b/net/wireless/Kconfig @@ -29,3 +29,14 @@ config WIRELESS_EXT Say N (if you can) unless you know you need wireless extensions for external modules. + +config WIRELESS_EXT_SYSFS + bool "Wireless extensions sysfs files" + default y + depends on WIRELESS_EXT && SYSFS + help + This option enables the deprecated wireless statistics + files in /sys/class/net/*/wireless/. The same information + is available via the ioctls as well. + + Say Y if you have programs using it (we don't know of any). -- cgit v1.2.2 From f591fa5dbbbeaebd95c9c019b3a536a327fb79de Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 10 Jul 2008 11:21:26 +0200 Subject: mac80211: fix TX sequence numbers This patch makes mac80211 assign proper sequence numbers to QoS-data frames. It also removes the old sequence number code because we noticed that only the driver or hardware can assign sequence numbers to non-QoS-data and especially management frames in a race-free manner because beacons aren't passed through mac80211's TX path. This patch also adds temporary code to the rt2x00 drivers to not break them completely, that code will have to be reworked for proper sequence numbers on beacons. It also moves sequence number assignment down in the TX path so no sequence numbers are assigned to frames that are dropped. Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/ieee80211_i.h | 2 -- net/mac80211/iface.c | 1 - net/mac80211/sta_info.h | 1 + net/mac80211/tx.c | 79 ++++++++++++++++++++++++++++++---------------- 4 files changed, 52 insertions(+), 31 deletions(-) (limited to 'net') diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index 934c3ef4f0bc..cbea0154ee3a 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -419,8 +419,6 @@ struct ieee80211_sub_if_data { */ u64 basic_rates; - u16 sequence; - /* Fragment table for host-based reassembly */ struct ieee80211_fragment_entry fragments[IEEE80211_FRAGMENT_MAX]; unsigned int fragment_next; diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c index 2e3adcb3ce23..610ed1d9893a 100644 --- a/net/mac80211/iface.c +++ b/net/mac80211/iface.c @@ -162,7 +162,6 @@ int ieee80211_if_change_type(struct ieee80211_sub_if_data *sdata, /* reset some values that shouldn't be kept across type changes */ sdata->basic_rates = 0; sdata->drop_unencrypted = 0; - sdata->sequence = 0; return 0; } diff --git a/net/mac80211/sta_info.h b/net/mac80211/sta_info.h index 94311dcfe043..109db787ccb7 100644 --- a/net/mac80211/sta_info.h +++ b/net/mac80211/sta_info.h @@ -285,6 +285,7 @@ struct sta_info { unsigned long tx_fragments; int txrate_idx; int last_txrate_idx; + u16 tid_seq[IEEE80211_QOS_CTL_TID_MASK + 1]; #ifdef CONFIG_MAC80211_DEBUG_COUNTERS unsigned int wme_tx_queue[NUM_RX_DATA_QUEUES]; #endif diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 8843416e1460..0fbadd8b983c 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -38,16 +38,6 @@ /* misc utils */ -static inline void ieee80211_include_sequence(struct ieee80211_sub_if_data *sdata, - struct ieee80211_hdr *hdr) -{ - /* Set the sequence number for this frame. */ - hdr->seq_ctrl = cpu_to_le16(sdata->sequence); - - /* Increase the sequence number. */ - sdata->sequence = (sdata->sequence + 0x10) & IEEE80211_SCTL_SEQ; -} - #ifdef CONFIG_MAC80211_LOWTX_FRAME_DUMP static void ieee80211_dump_frame(const char *ifname, const char *title, const struct sk_buff *skb) @@ -274,17 +264,6 @@ ieee80211_tx_h_check_assoc(struct ieee80211_tx_data *tx) return TX_CONTINUE; } -static ieee80211_tx_result debug_noinline -ieee80211_tx_h_sequence(struct ieee80211_tx_data *tx) -{ - struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; - - if (ieee80211_hdrlen(hdr->frame_control) >= 24) - ieee80211_include_sequence(tx->sdata, hdr); - - return TX_CONTINUE; -} - /* This function is called whenever the AP is about to exceed the maximum limit * of buffered frames for power saving STAs. This situation should not really * happen often during normal operation, so dropping the oldest buffered packet @@ -641,6 +620,49 @@ ieee80211_tx_h_misc(struct ieee80211_tx_data *tx) return TX_CONTINUE; } +static ieee80211_tx_result debug_noinline +ieee80211_tx_h_sequence(struct ieee80211_tx_data *tx) +{ + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb); + struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)tx->skb->data; + u16 *seq; + u8 *qc; + int tid; + + /* only for injected frames */ + if (unlikely(ieee80211_is_ctl(hdr->frame_control))) + return TX_CONTINUE; + + if (ieee80211_hdrlen(hdr->frame_control) < 24) + return TX_CONTINUE; + + if (!ieee80211_is_data_qos(hdr->frame_control)) { + info->flags |= IEEE80211_TX_CTL_ASSIGN_SEQ; + return TX_CONTINUE; + } + + /* + * This should be true for injected/management frames only, for + * management frames we have set the IEEE80211_TX_CTL_ASSIGN_SEQ + * above since they are not QoS-data frames. + */ + if (!tx->sta) + return TX_CONTINUE; + + /* include per-STA, per-TID sequence counter */ + + qc = ieee80211_get_qos_ctl(hdr); + tid = *qc & IEEE80211_QOS_CTL_TID_MASK; + seq = &tx->sta->tid_seq[tid]; + + hdr->seq_ctrl = cpu_to_le16(*seq); + + /* Increase the sequence number. */ + *seq = (*seq + 0x10) & IEEE80211_SCTL_SEQ; + + return TX_CONTINUE; +} + static ieee80211_tx_result debug_noinline ieee80211_tx_h_fragment(struct ieee80211_tx_data *tx) { @@ -1110,12 +1132,12 @@ static int invoke_tx_handlers(struct ieee80211_tx_data *tx) goto txh_done; CALL_TXH(ieee80211_tx_h_check_assoc) - CALL_TXH(ieee80211_tx_h_sequence) CALL_TXH(ieee80211_tx_h_ps_buf) CALL_TXH(ieee80211_tx_h_select_key) CALL_TXH(ieee80211_tx_h_michael_mic_add) CALL_TXH(ieee80211_tx_h_rate_ctrl) CALL_TXH(ieee80211_tx_h_misc) + CALL_TXH(ieee80211_tx_h_sequence) CALL_TXH(ieee80211_tx_h_fragment) /* handlers after fragment must be aware of tx info fragmentation! */ CALL_TXH(ieee80211_tx_h_encrypt) @@ -1827,9 +1849,6 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, memcpy(skb_put(skb, beacon->head_len), beacon->head, beacon->head_len); - ieee80211_include_sequence(sdata, - (struct ieee80211_hdr *)skb->data); - /* * Not very nice, but we want to allow the driver to call * ieee80211_beacon_get() as a response to the set_tim() @@ -1919,14 +1938,18 @@ struct sk_buff *ieee80211_beacon_get(struct ieee80211_hw *hw, info->control.vif = vif; info->tx_rate_idx = rsel.rate_idx; + + info->flags |= IEEE80211_TX_CTL_NO_ACK; + info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; + info->flags |= IEEE80211_TX_CTL_CLEAR_PS_FILT; + info->flags |= IEEE80211_TX_CTL_ASSIGN_SEQ; if (sdata->bss_conf.use_short_preamble && sband->bitrates[rsel.rate_idx].flags & IEEE80211_RATE_SHORT_PREAMBLE) info->flags |= IEEE80211_TX_CTL_SHORT_PREAMBLE; + info->antenna_sel_tx = local->hw.conf.antenna_sel_tx; - info->flags |= IEEE80211_TX_CTL_NO_ACK; - info->flags |= IEEE80211_TX_CTL_DO_NOT_ENCRYPT; info->control.retry_limit = 1; - info->flags |= IEEE80211_TX_CTL_CLEAR_PS_FILT; + (*num_beacons)++; out: rcu_read_unlock(); -- cgit v1.2.2 From f434b2d111d9ff84ebdd0f11a7ae42c761453259 Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Thu, 10 Jul 2008 11:22:31 +0200 Subject: mac80211: fix struct ieee80211_tx_queue_params Multiple issues: - there are no "default" values needed - cw_min/cw_max can be larger than documented - restructure to decrease size - use get_unaligned_le16 Signed-off-by: Johannes Berg Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 61d7f81bf45e..a4bbc8d6d0e6 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -345,7 +345,7 @@ static void ieee80211_sta_wmm_params(struct net_device *dev, params.aifs = pos[0] & 0x0f; params.cw_max = ecw2cw((pos[1] & 0xf0) >> 4); params.cw_min = ecw2cw(pos[1] & 0x0f); - params.txop = pos[2] | (pos[3] << 8); + params.txop = get_unaligned_le16(pos + 2); #ifdef CONFIG_MAC80211_VERBOSE_DEBUG printk(KERN_DEBUG "%s: WMM queue=%d aci=%d acm=%d aifs=%d " "cWmin=%d cWmax=%d txop=%d\n", -- cgit v1.2.2 From 1e188637902eb4b62d325d3cc76b076724f3ec55 Mon Sep 17 00:00:00 2001 From: Emmanuel Grumbach Date: Thu, 10 Jul 2008 17:54:14 +0300 Subject: mac80211: dont add a STA which is not in the same IBSS This patch avoids adding STAs that don't belong to our IBSS ieee80211_bssid_match matches also bcast address so also APs were added Signed-off-by: Emmanuel Grumbach Signed-off-by: Tomas Winkler Signed-off-by: John W. Linville --- net/mac80211/mlme.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index a4bbc8d6d0e6..8f51375317dd 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -4360,7 +4360,7 @@ struct sta_info *ieee80211_ibss_add_sta(struct net_device *dev, return NULL; } - if (!ieee80211_bssid_match(bssid, sdata->u.sta.bssid)) + if (compare_ether_addr(bssid, sdata->u.sta.bssid)) return NULL; #ifdef CONFIG_MAC80211_VERBOSE_DEBUG -- cgit v1.2.2 From 4c8894980010536915c4f5513ee180e3614aeca9 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Mon, 14 Jul 2008 20:22:38 -0700 Subject: netfilter: Let nf_ct_kill() callers know if del_timer() returned true. Signed-off-by: David S. Miller --- net/netfilter/nf_conntrack_core.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c index 212a0888408d..28d03e64200b 100644 --- a/net/netfilter/nf_conntrack_core.c +++ b/net/netfilter/nf_conntrack_core.c @@ -848,10 +848,10 @@ acct: } EXPORT_SYMBOL_GPL(__nf_ct_refresh_acct); -void __nf_ct_kill_acct(struct nf_conn *ct, - enum ip_conntrack_info ctinfo, - const struct sk_buff *skb, - int do_acct) +bool __nf_ct_kill_acct(struct nf_conn *ct, + enum ip_conntrack_info ctinfo, + const struct sk_buff *skb, + int do_acct) { #ifdef CONFIG_NF_CT_ACCT if (do_acct) { @@ -862,8 +862,11 @@ void __nf_ct_kill_acct(struct nf_conn *ct, spin_unlock_bh(&nf_conntrack_lock); } #endif - if (del_timer(&ct->timeout)) + if (del_timer(&ct->timeout)) { ct->timeout.function((unsigned long)ct); + return true; + } + return false; } EXPORT_SYMBOL_GPL(__nf_ct_kill_acct); -- cgit v1.2.2 From 72d9794f444734af56ef12833b496326643e2964 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 14 Jul 2008 20:36:32 -0700 Subject: net-sched: cls_flow: add perturbation support Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/sched/cls_flow.c | 52 +++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 43 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/sched/cls_flow.c b/net/sched/cls_flow.c index 971b867e0484..8f63a1a94014 100644 --- a/net/sched/cls_flow.c +++ b/net/sched/cls_flow.c @@ -36,6 +36,8 @@ struct flow_filter { struct list_head list; struct tcf_exts exts; struct tcf_ematch_tree ematches; + struct timer_list perturb_timer; + u32 perturb_period; u32 handle; u32 nkeys; @@ -47,11 +49,9 @@ struct flow_filter { u32 addend; u32 divisor; u32 baseclass; + u32 hashrnd; }; -static u32 flow_hashrnd __read_mostly; -static int flow_hashrnd_initted __read_mostly; - static const struct tcf_ext_map flow_ext_map = { .action = TCA_FLOW_ACT, .police = TCA_FLOW_POLICE, @@ -348,7 +348,7 @@ static int flow_classify(struct sk_buff *skb, struct tcf_proto *tp, } if (f->mode == FLOW_MODE_HASH) - classid = jhash2(keys, f->nkeys, flow_hashrnd); + classid = jhash2(keys, f->nkeys, f->hashrnd); else { classid = keys[0]; classid = (classid & f->mask) ^ f->xor; @@ -369,6 +369,15 @@ static int flow_classify(struct sk_buff *skb, struct tcf_proto *tp, return -1; } +static void flow_perturbation(unsigned long arg) +{ + struct flow_filter *f = (struct flow_filter *)arg; + + get_random_bytes(&f->hashrnd, 4); + if (f->perturb_period) + mod_timer(&f->perturb_timer, jiffies + f->perturb_period); +} + static const struct nla_policy flow_policy[TCA_FLOW_MAX + 1] = { [TCA_FLOW_KEYS] = { .type = NLA_U32 }, [TCA_FLOW_MODE] = { .type = NLA_U32 }, @@ -381,6 +390,7 @@ static const struct nla_policy flow_policy[TCA_FLOW_MAX + 1] = { [TCA_FLOW_ACT] = { .type = NLA_NESTED }, [TCA_FLOW_POLICE] = { .type = NLA_NESTED }, [TCA_FLOW_EMATCHES] = { .type = NLA_NESTED }, + [TCA_FLOW_PERTURB] = { .type = NLA_U32 }, }; static int flow_change(struct tcf_proto *tp, unsigned long base, @@ -394,6 +404,7 @@ static int flow_change(struct tcf_proto *tp, unsigned long base, struct tcf_exts e; struct tcf_ematch_tree t; unsigned int nkeys = 0; + unsigned int perturb_period = 0; u32 baseclass = 0; u32 keymask = 0; u32 mode; @@ -442,6 +453,14 @@ static int flow_change(struct tcf_proto *tp, unsigned long base, mode = nla_get_u32(tb[TCA_FLOW_MODE]); if (mode != FLOW_MODE_HASH && nkeys > 1) goto err2; + + if (mode == FLOW_MODE_HASH) + perturb_period = f->perturb_period; + if (tb[TCA_FLOW_PERTURB]) { + if (mode != FLOW_MODE_HASH) + goto err2; + perturb_period = nla_get_u32(tb[TCA_FLOW_PERTURB]) * HZ; + } } else { err = -EINVAL; if (!handle) @@ -455,6 +474,12 @@ static int flow_change(struct tcf_proto *tp, unsigned long base, if (mode != FLOW_MODE_HASH && nkeys > 1) goto err2; + if (tb[TCA_FLOW_PERTURB]) { + if (mode != FLOW_MODE_HASH) + goto err2; + perturb_period = nla_get_u32(tb[TCA_FLOW_PERTURB]) * HZ; + } + if (TC_H_MAJ(baseclass) == 0) baseclass = TC_H_MAKE(tp->q->handle, baseclass); if (TC_H_MIN(baseclass) == 0) @@ -467,6 +492,11 @@ static int flow_change(struct tcf_proto *tp, unsigned long base, f->handle = handle; f->mask = ~0U; + + get_random_bytes(&f->hashrnd, 4); + f->perturb_timer.function = flow_perturbation; + f->perturb_timer.data = (unsigned long)f; + init_timer_deferrable(&f->perturb_timer); } tcf_exts_change(tp, &f->exts, &e); @@ -495,6 +525,11 @@ static int flow_change(struct tcf_proto *tp, unsigned long base, if (baseclass) f->baseclass = baseclass; + f->perturb_period = perturb_period; + del_timer(&f->perturb_timer); + if (perturb_period) + mod_timer(&f->perturb_timer, jiffies + perturb_period); + if (*arg == 0) list_add_tail(&f->list, &head->filters); @@ -512,6 +547,7 @@ err1: static void flow_destroy_filter(struct tcf_proto *tp, struct flow_filter *f) { + del_timer_sync(&f->perturb_timer); tcf_exts_destroy(tp, &f->exts); tcf_em_tree_destroy(tp, &f->ematches); kfree(f); @@ -532,11 +568,6 @@ static int flow_init(struct tcf_proto *tp) { struct flow_head *head; - if (!flow_hashrnd_initted) { - get_random_bytes(&flow_hashrnd, 4); - flow_hashrnd_initted = 1; - } - head = kzalloc(sizeof(*head), GFP_KERNEL); if (head == NULL) return -ENOBUFS; @@ -605,6 +636,9 @@ static int flow_dump(struct tcf_proto *tp, unsigned long fh, if (f->baseclass) NLA_PUT_U32(skb, TCA_FLOW_BASECLASS, f->baseclass); + if (f->perturb_period) + NLA_PUT_U32(skb, TCA_FLOW_PERTURB, f->perturb_period / HZ); + if (tcf_exts_dump(skb, &f->exts, &flow_ext_map) < 0) goto nla_put_failure; #ifdef CONFIG_NET_EMATCH -- cgit v1.2.2 From 2aeb0b88b3c7a0e3bef55e7ff0efffd5d971aa57 Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Mon, 14 Jul 2008 20:49:46 -0700 Subject: af_packet: Check return of dev_set_promiscuity/allmulti dev_set_promiscuity/allmulti might overflow. Commit: "netdevice: Fix promiscuity and allmulti overflow" in net-next makes dev_set_promiscuity/allmulti return error number if overflow happened. In af_packet, we check all positive increment for promiscuity and allmulti to get error return. Signed-off-by: Wang Chen Acked-by: Patrick McHardy Signed-off-by: David S. Miller --- net/packet/af_packet.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index beca6402f1cf..9f2269166687 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -1173,7 +1173,8 @@ static int packet_getname(struct socket *sock, struct sockaddr *uaddr, return 0; } -static void packet_dev_mc(struct net_device *dev, struct packet_mclist *i, int what) +static int packet_dev_mc(struct net_device *dev, struct packet_mclist *i, + int what) { switch (i->type) { case PACKET_MR_MULTICAST: @@ -1183,13 +1184,14 @@ static void packet_dev_mc(struct net_device *dev, struct packet_mclist *i, int w dev_mc_delete(dev, i->addr, i->alen, 0); break; case PACKET_MR_PROMISC: - dev_set_promiscuity(dev, what); + return dev_set_promiscuity(dev, what); break; case PACKET_MR_ALLMULTI: - dev_set_allmulti(dev, what); + return dev_set_allmulti(dev, what); break; default:; } + return 0; } static void packet_dev_mclist(struct net_device *dev, struct packet_mclist *i, int what) @@ -1243,7 +1245,11 @@ static int packet_mc_add(struct sock *sk, struct packet_mreq_max *mreq) i->count = 1; i->next = po->mclist; po->mclist = i; - packet_dev_mc(dev, i, +1); + err = packet_dev_mc(dev, i, 1); + if (err) { + po->mclist = i->next; + kfree(i); + } done: rtnl_unlock(); -- cgit v1.2.2 From bc3f9076f671f128c82022428992c30be57f22d5 Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Mon, 14 Jul 2008 20:53:13 -0700 Subject: bridge: Check return of dev_set_promiscuity dev_set_promiscuity/allmulti might overflow. Commit: "netdevice: Fix promiscuity and allmulti overflow" in net-next makes dev_set_promiscuity/allmulti return error number if overflow happened. Here, we check the positive increment for promiscuity to get error return. Signed-off-by: Wang Chen Signed-off-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/bridge/br_if.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c index 497df086141a..a072ea5ca6f5 100644 --- a/net/bridge/br_if.c +++ b/net/bridge/br_if.c @@ -373,6 +373,10 @@ int br_add_if(struct net_bridge *br, struct net_device *dev) if (IS_ERR(p)) return PTR_ERR(p); + err = dev_set_promiscuity(dev, 1); + if (err) + goto put_back; + err = kobject_init_and_add(&p->kobj, &brport_ktype, &(dev->dev.kobj), SYSFS_BRIDGE_PORT_ATTR); if (err) @@ -388,7 +392,6 @@ int br_add_if(struct net_bridge *br, struct net_device *dev) rcu_assign_pointer(dev->br_port, p); dev_disable_lro(dev); - dev_set_promiscuity(dev, 1); list_add_rcu(&p->list, &br->port_list); @@ -412,12 +415,12 @@ err2: br_fdb_delete_by_port(br, p, 1); err1: kobject_del(&p->kobj); - goto put_back; err0: kobject_put(&p->kobj); - + dev_set_promiscuity(dev, -1); put_back: dev_put(dev); + kfree(p); return err; } -- cgit v1.2.2 From 5ae7b444137143a4e067b80354171ab128eb1b2b Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Mon, 14 Jul 2008 20:54:23 -0700 Subject: ipv6: Check return of dev_set_allmulti allmulti might overflow. Commit: "netdevice: Fix promiscuity and allmulti overflow" in net-next makes dev_set_promiscuity/allmulti return error number if overflow happened. Here, we check the positive increment for allmulti to get error return. Signed-off-by: Wang Chen Acked-by: Patrick McHardy Signed-off-by: David S. Miller --- net/ipv6/ip6mr.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c index cfac26d674ed..2dd832592a31 100644 --- a/net/ipv6/ip6mr.c +++ b/net/ipv6/ip6mr.c @@ -595,6 +595,7 @@ static int mif6_add(struct mif6ctl *vifc, int mrtsock) int vifi = vifc->mif6c_mifi; struct mif_device *v = &vif6_table[vifi]; struct net_device *dev; + int err; /* Is vif busy ? */ if (MIF_EXISTS(vifi)) @@ -612,6 +613,11 @@ static int mif6_add(struct mif6ctl *vifc, int mrtsock) dev = ip6mr_reg_vif(); if (!dev) return -ENOBUFS; + err = dev_set_allmulti(dev, 1); + if (err) { + unregister_netdevice(dev); + return err; + } break; #endif case 0: @@ -619,13 +625,14 @@ static int mif6_add(struct mif6ctl *vifc, int mrtsock) if (!dev) return -EADDRNOTAVAIL; dev_put(dev); + err = dev_set_allmulti(dev, 1); + if (err) + return err; break; default: return -EINVAL; } - dev_set_allmulti(dev, 1); - /* * Fill in the VIF structures */ -- cgit v1.2.2 From 7af3db78a99f47b9ff40b8cb0bb08160ad6a3d6b Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Mon, 14 Jul 2008 20:54:54 -0700 Subject: ipv6: Fix using after dev_put() Patrick McHardy pointed it out. Signed-off-by: Wang Chen Signed-off-by: David S. Miller --- net/ipv6/ip6mr.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c index 2dd832592a31..0b41aa2675f5 100644 --- a/net/ipv6/ip6mr.c +++ b/net/ipv6/ip6mr.c @@ -443,6 +443,7 @@ static struct net_device *ip6mr_reg_vif(void) if (dev_open(dev)) goto failure; + dev_hold(dev); return dev; failure: @@ -616,6 +617,7 @@ static int mif6_add(struct mif6ctl *vifc, int mrtsock) err = dev_set_allmulti(dev, 1); if (err) { unregister_netdevice(dev); + dev_put(dev); return err; } break; @@ -624,10 +626,11 @@ static int mif6_add(struct mif6ctl *vifc, int mrtsock) dev = dev_get_by_index(&init_net, vifc->mif6c_pifi); if (!dev) return -EADDRNOTAVAIL; - dev_put(dev); err = dev_set_allmulti(dev, 1); - if (err) + if (err) { + dev_put(dev); return err; + } break; default: return -EINVAL; @@ -651,7 +654,6 @@ static int mif6_add(struct mif6ctl *vifc, int mrtsock) /* And finish update writing critical data */ write_lock_bh(&mrt_lock); - dev_hold(dev); v->dev = dev; #ifdef CONFIG_IPV6_PIMSM_V2 if (v->flags & MIFF_REGISTER) -- cgit v1.2.2 From d607032db0ccd7274bee348df3214f6f52b24816 Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Mon, 14 Jul 2008 20:55:26 -0700 Subject: ipv4: Check return of dev_set_allmulti allmulti might overflow. Commit: "netdevice: Fix promiscuity and allmulti overflow" in net-next makes dev_set_promiscuity/allmulti return error number if overflow happened. Here, we check the positive increment for allmulti to get error return. PS: For unwinding tunnel creating, we let ipip->ioctl() to handle it. Signed-off-by: Wang Chen Signed-off-by: David S. Miller --- net/ipv4/ipmr.c | 40 +++++++++++++++++++++++++++++++++++++++- 1 file changed, 39 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index 438fab9c62a0..2f4d8afd067c 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c @@ -118,6 +118,31 @@ static struct timer_list ipmr_expire_timer; /* Service routines creating virtual interfaces: DVMRP tunnels and PIMREG */ +static void ipmr_del_tunnel(struct net_device *dev, struct vifctl *v) +{ + dev_close(dev); + + dev = __dev_get_by_name(&init_net, "tunl0"); + if (dev) { + struct ifreq ifr; + mm_segment_t oldfs; + struct ip_tunnel_parm p; + + memset(&p, 0, sizeof(p)); + p.iph.daddr = v->vifc_rmt_addr.s_addr; + p.iph.saddr = v->vifc_lcl_addr.s_addr; + p.iph.version = 4; + p.iph.ihl = 5; + p.iph.protocol = IPPROTO_IPIP; + sprintf(p.name, "dvmrp%d", v->vifc_vifi); + ifr.ifr_ifru.ifru_data = (__force void __user *)&p; + + oldfs = get_fs(); set_fs(KERNEL_DS); + dev->do_ioctl(dev, &ifr, SIOCDELTUNNEL); + set_fs(oldfs); + } +} + static struct net_device *ipmr_new_tunnel(struct vifctl *v) { @@ -389,6 +414,7 @@ static int vif_add(struct vifctl *vifc, int mrtsock) struct vif_device *v = &vif_table[vifi]; struct net_device *dev; struct in_device *in_dev; + int err; /* Is vif busy ? */ if (VIF_EXISTS(vifi)) @@ -406,18 +432,31 @@ static int vif_add(struct vifctl *vifc, int mrtsock) dev = ipmr_reg_vif(); if (!dev) return -ENOBUFS; + err = dev_set_allmulti(dev, 1); + if (err) { + unregister_netdevice(dev); + return err; + } break; #endif case VIFF_TUNNEL: dev = ipmr_new_tunnel(vifc); if (!dev) return -ENOBUFS; + err = dev_set_allmulti(dev, 1); + if (err) { + ipmr_del_tunnel(dev, vifc); + return err; + } break; case 0: dev = ip_dev_find(&init_net, vifc->vifc_lcl_addr.s_addr); if (!dev) return -EADDRNOTAVAIL; dev_put(dev); + err = dev_set_allmulti(dev, 1); + if (err) + return err; break; default: return -EINVAL; @@ -426,7 +465,6 @@ static int vif_add(struct vifctl *vifc, int mrtsock) if ((in_dev = __in_dev_get_rtnl(dev)) == NULL) return -EADDRNOTAVAIL; IPV4_DEVCONF(in_dev->cnf, MC_FORWARDING)++; - dev_set_allmulti(dev, +1); ip_rt_multicast_event(in_dev); /* -- cgit v1.2.2 From 7dc00c82cbb0119cf4663f65bbaa2cc55f961db2 Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Mon, 14 Jul 2008 20:56:34 -0700 Subject: ipv4: Fix ipmr unregister device oops An oops happens during device unregister. The following oops happened when I add two tunnels, which use a same device, and then delete one tunnel. Obviously deleting tunnel "A" causes device unregister, which send a notification, and after receiving notification, ipmr do unregister again for tunnel "B" which also use same device. That is wrong. After receiving notification, ipmr only needs to decrease reference count and don't do duplicated unregister. Fortunately, IPv6 side doesn't add tunnel in ip6mr, so it's clean. This patch fixs: - unregister device oops - using after dev_put() Here is the oops: === Jul 11 15:39:29 wangchen kernel: ------------[ cut here ]------------ Jul 11 15:39:29 wangchen kernel: kernel BUG at net/core/dev.c:3651! Jul 11 15:39:29 wangchen kernel: invalid opcode: 0000 [#1] Jul 11 15:39:29 wangchen kernel: Modules linked in: ipip tunnel4 nfsd lockd nfs_acl auth_rpcgss sunrpc exportfs ipv6 snd_pcm_oss snd_mixer_oss snd_seq snd_seq_device af_packet binfmt_misc button battery ac loop dm_mod usbhid ff_memless pcmcia firmware_class ohci1394 8139too mii ieee1394 yenta_socket rsrc_nonstatic pcmcia_core ide_cd_mod cdrom snd_intel8x0 snd_ac97_codec ac97_bus snd_pcm i2c_i801 snd_timer snd i2c_core soundcore snd_page_alloc rng_core shpchp ehci_hcd uhci_hcd pci_hotplug intel_agp agpgart usbcore ext3 jbd ata_piix ahci libata dock edd fan thermal processor thermal_sys piix sd_mod scsi_mod ide_disk ide_core [last unloaded: freq_table] Jul 11 15:39:29 wangchen kernel: Jul 11 15:39:29 wangchen kernel: Pid: 4102, comm: mroute Not tainted (2.6.26-rc9-default #69) Jul 11 15:39:29 wangchen kernel: EIP: 0060:[] EFLAGS: 00010202 CPU: 0 Jul 11 15:39:29 wangchen kernel: EIP is at rollback_registered+0x61/0xe3 Jul 11 15:39:29 wangchen kernel: EAX: 00000001 EBX: ecba6000 ECX: 00000000 EDX: ffffffff Jul 11 15:39:29 wangchen kernel: ESI: 00000001 EDI: ecba6000 EBP: c03de2e8 ESP: ed8e7c3c Jul 11 15:39:29 wangchen kernel: DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068 Jul 11 15:39:29 wangchen kernel: Process mroute (pid: 4102, ti=ed8e6000 task=ed41e830 task.ti=ed8e6000) Jul 11 15:39:29 wangchen kernel: Stack: ecba6000 c024641c 00000028 c0284e1a 00000001 c03de2e8 ecba6000 eecff360 Jul 11 15:39:29 wangchen kernel: c0284e4c c03536f4 fffffff8 00000000 c029a819 ecba6000 00000006 ecba6000 Jul 11 15:39:29 wangchen kernel: 00000000 ecba6000 c03de2c0 c012841b ffffffff 00000000 c024639f ecba6000 Jul 11 15:39:29 wangchen kernel: Call Trace: Jul 11 15:39:29 wangchen kernel: [] unregister_netdevice+0x2f/0x51 Jul 11 15:39:29 wangchen kernel: [] vif_delete+0xaf/0xc3 Jul 11 15:39:29 wangchen kernel: [] ipmr_device_event+0x1e/0x30 Jul 11 15:39:29 wangchen kernel: [] notifier_call_chain+0x2a/0x47 Jul 11 15:39:29 wangchen kernel: [] raw_notifier_call_chain+0x9/0xc Jul 11 15:39:29 wangchen kernel: [] rollback_registered+0x95/0xe3 Jul 11 15:39:29 wangchen kernel: [] unregister_netdevice+0x2f/0x51 Jul 11 15:39:29 wangchen kernel: [] vif_delete+0xaf/0xc3 Jul 11 15:39:29 wangchen kernel: [] ip_mroute_setsockopt+0x47a/0x801 Jul 11 15:39:29 wangchen kernel: [] do_get_write_access+0x2df/0x313 [jbd] Jul 11 15:39:29 wangchen kernel: [] __find_get_block_slow+0xda/0xe4 Jul 11 15:39:29 wangchen kernel: [] __find_get_block+0xf8/0x122 Jul 11 15:39:29 wangchen kernel: [] __find_get_block+0xf8/0x122 Jul 11 15:39:29 wangchen kernel: [] journal_cancel_revoke+0xda/0x110 [jbd] Jul 11 15:39:29 wangchen kernel: [] ip_setsockopt+0xa9/0x9ee Jul 11 15:39:29 wangchen kernel: [] journal_cancel_revoke+0xda/0x110 [jbd] Jul 11 15:39:29 wangchen kernel: [] do_get_write_access+0x2df/0x313 [jbd] Jul 11 15:39:29 wangchen kernel: [] __ext3_get_inode_loc+0xcf/0x271 [ext3] Jul 11 15:39:29 wangchen kernel: [] __ext3_journal_dirty_metadata+0x13/0x32 [ext3] Jul 11 15:39:29 wangchen kernel: [] __wake_up+0xf/0x15 Jul 11 15:39:29 wangchen kernel: [] journal_stop+0x1bd/0x1c6 [jbd] Jul 11 15:39:29 wangchen kernel: [] __ext3_journal_stop+0x19/0x34 [ext3] Jul 11 15:39:29 wangchen kernel: [] get_page_from_freelist+0x94/0x369 Jul 11 15:39:29 wangchen kernel: [] filemap_fault+0x1ac/0x2fe Jul 11 15:39:29 wangchen kernel: [] security_sk_alloc+0xd/0xf Jul 11 15:39:29 wangchen kernel: [] sk_prot_alloc+0x36/0x78 Jul 11 15:39:29 wangchen kernel: [] sk_alloc+0x3a/0x40 Jul 11 15:39:29 wangchen kernel: [] raw_hash_sk+0x46/0x4e Jul 11 15:39:29 wangchen kernel: [] d_alloc+0x1b/0x157 Jul 11 15:39:29 wangchen kernel: [] sock_common_setsockopt+0x12/0x16 Jul 11 15:39:29 wangchen kernel: [] sys_setsockopt+0x6f/0x8e Jul 11 15:39:29 wangchen kernel: [] sys_socketcall+0x15c/0x19e Jul 11 15:39:29 wangchen kernel: [] sysenter_past_esp+0x6a/0x99 Jul 11 15:39:29 wangchen kernel: [] unix_poll+0x69/0x78 Jul 11 15:39:29 wangchen kernel: ======================= Jul 11 15:39:29 wangchen kernel: Code: 83 e0 01 00 00 85 c0 75 1f 53 53 68 12 81 31 c0 e8 3c 30 ed ff ba 3f 0e 00 00 b8 b9 7f 31 c0 83 c4 0c 5b e9 f5 26 ed ff 48 74 04 <0f> 0b eb fe 89 d8 e8 21 ff ff ff 89 d8 e8 62 ea ff ff c7 83 e0 Jul 11 15:39:29 wangchen kernel: EIP: [] rollback_registered+0x61/0xe3 SS:ESP 0068:ed8e7c3c Jul 11 15:39:29 wangchen kernel: ---[ end trace c311acf85d169786 ]--- === Signed-off-by: Wang Chen Signed-off-by: David S. Miller --- net/ipv4/ipmr.c | 22 ++++++++++++++-------- 1 file changed, 14 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index 2f4d8afd067c..c9ab47b966b5 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c @@ -184,6 +184,7 @@ struct net_device *ipmr_new_tunnel(struct vifctl *v) if (dev_open(dev)) goto failure; + dev_hold(dev); } } return dev; @@ -250,6 +251,8 @@ static struct net_device *ipmr_reg_vif(void) if (dev_open(dev)) goto failure; + dev_hold(dev); + return dev; failure: @@ -264,9 +267,10 @@ failure: /* * Delete a VIF entry + * @notify: Set to 1, if the caller is a notifier_call */ -static int vif_delete(int vifi) +static int vif_delete(int vifi, int notify) { struct vif_device *v; struct net_device *dev; @@ -309,7 +313,7 @@ static int vif_delete(int vifi) ip_rt_multicast_event(in_dev); } - if (v->flags&(VIFF_TUNNEL|VIFF_REGISTER)) + if (v->flags&(VIFF_TUNNEL|VIFF_REGISTER) && !notify) unregister_netdevice(dev); dev_put(dev); @@ -435,6 +439,7 @@ static int vif_add(struct vifctl *vifc, int mrtsock) err = dev_set_allmulti(dev, 1); if (err) { unregister_netdevice(dev); + dev_put(dev); return err; } break; @@ -446,6 +451,7 @@ static int vif_add(struct vifctl *vifc, int mrtsock) err = dev_set_allmulti(dev, 1); if (err) { ipmr_del_tunnel(dev, vifc); + dev_put(dev); return err; } break; @@ -453,10 +459,11 @@ static int vif_add(struct vifctl *vifc, int mrtsock) dev = ip_dev_find(&init_net, vifc->vifc_lcl_addr.s_addr); if (!dev) return -EADDRNOTAVAIL; - dev_put(dev); err = dev_set_allmulti(dev, 1); - if (err) + if (err) { + dev_put(dev); return err; + } break; default: return -EINVAL; @@ -487,7 +494,6 @@ static int vif_add(struct vifctl *vifc, int mrtsock) /* And finish update writing critical data */ write_lock_bh(&mrt_lock); - dev_hold(dev); v->dev=dev; #ifdef CONFIG_IP_PIMSM if (v->flags&VIFF_REGISTER) @@ -834,7 +840,7 @@ static void mroute_clean_tables(struct sock *sk) */ for (i=0; idev==dev) - vif_delete(ct); + vif_delete(ct, 1); } return NOTIFY_DONE; } -- cgit v1.2.2 From 89146504cbfeb120dd08ec7f9f8314c4986189b8 Mon Sep 17 00:00:00 2001 From: Wang Chen Date: Mon, 14 Jul 2008 20:59:03 -0700 Subject: 8021q: Check return of dev_set_promiscuity/allmulti dev_set_promiscuity/allmulti might overflow. Commit: "netdevice: Fix promiscuity and allmulti overflow" in net-next makes dev_set_promiscuity/allmulti return error number if overflow happened. Here, we check all positive increment for promiscuity and allmulti to get error return. Signed-off-by: Wang Chen Acked-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan_dev.c | 28 ++++++++++++++++++++++------ 1 file changed, 22 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 8efa399823e3..9efd3c67c1db 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -518,19 +518,35 @@ static int vlan_dev_open(struct net_device *dev) if (compare_ether_addr(dev->dev_addr, real_dev->dev_addr)) { err = dev_unicast_add(real_dev, dev->dev_addr, ETH_ALEN); if (err < 0) - return err; + goto out; } - memcpy(vlan->real_dev_addr, real_dev->dev_addr, ETH_ALEN); - if (dev->flags & IFF_ALLMULTI) - dev_set_allmulti(real_dev, 1); - if (dev->flags & IFF_PROMISC) - dev_set_promiscuity(real_dev, 1); + if (dev->flags & IFF_ALLMULTI) { + err = dev_set_allmulti(real_dev, 1); + if (err < 0) + goto del_unicast; + } + if (dev->flags & IFF_PROMISC) { + err = dev_set_promiscuity(real_dev, 1); + if (err < 0) + goto clear_allmulti; + } + + memcpy(vlan->real_dev_addr, real_dev->dev_addr, ETH_ALEN); if (vlan->flags & VLAN_FLAG_GVRP) vlan_gvrp_request_join(dev); return 0; + +clear_allmulti: + if (dev->flags & IFF_ALLMULTI) + dev_set_allmulti(real_dev, -1); +del_unicast: + if (compare_ether_addr(dev->dev_addr, real_dev->dev_addr)) + dev_unicast_delete(real_dev, dev->dev_addr, ETH_ALEN); +out: + return err; } static int vlan_dev_stop(struct net_device *dev) -- cgit v1.2.2 From 7197914c35b31a75cb6e85c7fc2ae93d0027c28e Mon Sep 17 00:00:00 2001 From: Johann Felix Soden Date: Mon, 14 Jul 2008 22:22:29 -0700 Subject: net: Remove references to wan-router.txt in Kconfigs This patch removes references in drivers/net/wan/Kconfig and net/wanrouter/Kconfig to Documentation/networking/wan-router.txt which was removed in commit 99971e70fdc1862e120f3319fc0a4dba8c728acf ("[WANPIPE]: Forgotten bits of Sangoma drivers removal."). Signed-off-by: Johann Felix Soden Signed-off-by: David S. Miller --- net/wanrouter/Kconfig | 2 -- 1 file changed, 2 deletions(-) (limited to 'net') diff --git a/net/wanrouter/Kconfig b/net/wanrouter/Kconfig index 1debe1cb054e..61ceae0b9566 100644 --- a/net/wanrouter/Kconfig +++ b/net/wanrouter/Kconfig @@ -20,8 +20,6 @@ config WAN_ROUTER wish to use your Linux box as a WAN router, say Y here and also to the WAN driver for your card, below. You will then need the wan-tools package which is available from . - Read for more - information. To compile WAN routing support as a module, choose M here: the module will be called wanrouter. -- cgit v1.2.2 From 83aa2e964b9b04effa304aaf3c1090b46812a04b Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Mon, 14 Jul 2008 22:28:25 -0700 Subject: netlabel: return msg overflow error from netlbl_cipsov4_list faster Currently, we are trying to place the information from the kernel to 1, 2, 3 and 4 pages sequentially. These pages are allocated via slab. Though, from the slab point of view steps 3 and 4 are equivalent on most architectures. So, lets skip 3 pages attempt. By the way, should we switch from .doit to .dumpit interface here? The amount of data seems quite big for me. Signed-off-by: Denis V. Lunev Acked-by: Paul Moore Signed-off-by: David S. Miller --- net/netlabel/netlabel_cipso_v4.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/netlabel/netlabel_cipso_v4.c b/net/netlabel/netlabel_cipso_v4.c index 9080c61b71a5..0aec318bf0ef 100644 --- a/net/netlabel/netlabel_cipso_v4.c +++ b/net/netlabel/netlabel_cipso_v4.c @@ -591,7 +591,7 @@ list_retry: if (nlsze_mult < 4) { rcu_read_unlock(); kfree_skb(ans_skb); - nlsze_mult++; + nlsze_mult *= 2; goto list_start; } list_failure_lock: -- cgit v1.2.2 From 0ea522416b658dedfc9d565b331624a55a6260ad Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 14 Jul 2008 22:42:19 -0700 Subject: tipc: Remove unneeded parameter to tipc_createport_raw() This patch eliminates an unneeded parameter when creating a low-level TIPC port object. Instead of returning both the pointer to the port structure and the port's reference ID, it now returns only the pointer since the port structure contains the reference ID as one of its fields. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/port.c | 28 ++++++++++------------------ net/tipc/socket.c | 15 +++++++-------- 2 files changed, 17 insertions(+), 26 deletions(-) (limited to 'net') diff --git a/net/tipc/port.c b/net/tipc/port.c index 2e0cff408ff9..ffba1e7f06d2 100644 --- a/net/tipc/port.c +++ b/net/tipc/port.c @@ -2,7 +2,7 @@ * net/tipc/port.c: TIPC port code * * Copyright (c) 1992-2007, Ericsson AB - * Copyright (c) 2004-2007, Wind River Systems + * Copyright (c) 2004-2008, Wind River Systems * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -213,16 +213,13 @@ exit: /** * tipc_createport_raw - create a generic TIPC port * - * Returns port reference, or 0 if unable to create it - * - * Note: The newly created port is returned in the locked state. + * Returns pointer to (locked) TIPC port, or NULL if unable to create it */ -u32 tipc_createport_raw(void *usr_handle, +struct tipc_port *tipc_createport_raw(void *usr_handle, u32 (*dispatcher)(struct tipc_port *, struct sk_buff *), void (*wakeup)(struct tipc_port *), - const u32 importance, - struct tipc_port **tp_ptr) + const u32 importance) { struct port *p_ptr; struct tipc_msg *msg; @@ -231,13 +228,13 @@ u32 tipc_createport_raw(void *usr_handle, p_ptr = kzalloc(sizeof(*p_ptr), GFP_ATOMIC); if (!p_ptr) { warn("Port creation failed, no memory\n"); - return 0; + return NULL; } ref = tipc_ref_acquire(p_ptr, &p_ptr->publ.lock); if (!ref) { warn("Port creation failed, reference table exhausted\n"); kfree(p_ptr); - return 0; + return NULL; } p_ptr->publ.usr_handle = usr_handle; @@ -260,8 +257,7 @@ u32 tipc_createport_raw(void *usr_handle, INIT_LIST_HEAD(&p_ptr->port_list); list_add_tail(&p_ptr->port_list, &ports); spin_unlock_bh(&tipc_port_list_lock); - *tp_ptr = &p_ptr->publ; - return ref; + return &(p_ptr->publ); } int tipc_deleteport(u32 ref) @@ -1044,21 +1040,18 @@ int tipc_createport(u32 user_ref, { struct user_port *up_ptr; struct port *p_ptr; - struct tipc_port *tp_ptr; - u32 ref; up_ptr = kmalloc(sizeof(*up_ptr), GFP_ATOMIC); if (!up_ptr) { warn("Port creation failed, no memory\n"); return -ENOMEM; } - ref = tipc_createport_raw(NULL, port_dispatcher, port_wakeup, - importance, &tp_ptr); - if (ref == 0) { + p_ptr = (struct port *)tipc_createport_raw(NULL, port_dispatcher, + port_wakeup, importance); + if (!p_ptr) { kfree(up_ptr); return -ENOMEM; } - p_ptr = (struct port *)tp_ptr; p_ptr->user_port = up_ptr; up_ptr->user_ref = user_ref; @@ -1074,7 +1067,6 @@ int tipc_createport(u32 user_ref, INIT_LIST_HEAD(&up_ptr->uport_list); tipc_reg_add_port(up_ptr); *portref = p_ptr->publ.ref; - dbg(" tipc_createport: %x with ref %u\n", p_ptr, p_ptr->publ.ref); tipc_port_unlock(p_ptr); return TIPC_OK; } diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 38f48795b40e..9c362c5759ba 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -2,7 +2,7 @@ * net/tipc/socket.c: TIPC socket API * * Copyright (c) 2001-2007, Ericsson AB - * Copyright (c) 2004-2007, Wind River Systems + * Copyright (c) 2004-2008, Wind River Systems * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -189,7 +189,6 @@ static int tipc_create(struct net *net, struct socket *sock, int protocol) socket_state state; struct sock *sk; struct tipc_port *tp_ptr; - u32 portref; /* Validate arguments */ @@ -225,9 +224,9 @@ static int tipc_create(struct net *net, struct socket *sock, int protocol) /* Allocate TIPC port for socket to use */ - portref = tipc_createport_raw(sk, &dispatch, &wakeupdispatch, - TIPC_LOW_IMPORTANCE, &tp_ptr); - if (unlikely(portref == 0)) { + tp_ptr = tipc_createport_raw(sk, &dispatch, &wakeupdispatch, + TIPC_LOW_IMPORTANCE); + if (unlikely(!tp_ptr)) { sk_free(sk); return -ENOMEM; } @@ -240,14 +239,14 @@ static int tipc_create(struct net *net, struct socket *sock, int protocol) sock_init_data(sock, sk); sk->sk_rcvtimeo = msecs_to_jiffies(CONN_TIMEOUT_DEFAULT); sk->sk_backlog_rcv = backlog_rcv; - tipc_sk(sk)->p = tipc_get_port(portref); + tipc_sk(sk)->p = tp_ptr; spin_unlock_bh(tp_ptr->lock); if (sock->state == SS_READY) { - tipc_set_portunreturnable(portref, 1); + tipc_set_portunreturnable(tp_ptr->ref, 1); if (sock->type == SOCK_DGRAM) - tipc_set_portunreliable(portref, 1); + tipc_set_portunreliable(tp_ptr->ref, 1); } atomic_inc(&tipc_user_count); -- cgit v1.2.2 From 8642bd9e04f51980b2b6293c66acf7e388c9a6e7 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 14 Jul 2008 22:42:51 -0700 Subject: tipc: Optimize pointer dereferencing when receiving stream data This patch eliminates an unnecessary pointer dereference when accessing a stream-based socket's receive queue. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/socket.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 9c362c5759ba..ddcb2a753aa6 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -1133,7 +1133,7 @@ restart: /* Loop around if more data is required */ if ((sz_copied < buf_len) /* didn't get all requested data */ - && (!skb_queue_empty(&sock->sk->sk_receive_queue) || + && (!skb_queue_empty(&sk->sk_receive_queue) || (flags & MSG_WAITALL)) /* ... and more is ready or required */ && (!(flags & MSG_PEEK)) /* ... and aren't just peeking at data */ -- cgit v1.2.2 From 2da59918e26837f305131cfac9c0f1b3b42bb8ae Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 14 Jul 2008 22:43:32 -0700 Subject: tipc: Fix race condition that could cause accept() to fail This patch ensurs that accept() returns successfully even when the newly created socket is immediately disconnected by its peer. Previously, accept() would fail if it was unable to pass back the optional address info for the socket's peer before the socket became disconnected; TIPC now allows accept() to gather peer address information after disconnection. As a bonus, the revised code accesses the socket's port more efficiently, without the overhead incurred by a reference table lookup. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/socket.c | 42 ++++++++++++++++++++++-------------------- 1 file changed, 22 insertions(+), 20 deletions(-) (limited to 'net') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index ddcb2a753aa6..1848693ebb82 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -63,6 +63,7 @@ struct tipc_sock { struct sock sk; struct tipc_port *p; + struct tipc_portid peer_name; }; #define tipc_sk(sk) ((struct tipc_sock *)(sk)) @@ -377,27 +378,29 @@ static int bind(struct socket *sock, struct sockaddr *uaddr, int uaddr_len) * @sock: socket structure * @uaddr: area for returned socket address * @uaddr_len: area for returned length of socket address - * @peer: 0 to obtain socket name, 1 to obtain peer socket name + * @peer: 0 = own ID, 1 = current peer ID, 2 = current/former peer ID * * Returns 0 on success, errno otherwise * - * NOTE: This routine doesn't need to take the socket lock since it doesn't - * access any non-constant socket information. + * NOTE: This routine doesn't need to take the socket lock since it only + * accesses socket information that is unchanging (or which changes in + * a completely predictable manner). */ static int get_name(struct socket *sock, struct sockaddr *uaddr, int *uaddr_len, int peer) { struct sockaddr_tipc *addr = (struct sockaddr_tipc *)uaddr; - u32 portref = tipc_sk_port(sock->sk)->ref; - u32 res; + struct tipc_sock *tsock = tipc_sk(sock->sk); if (peer) { - res = tipc_peer(portref, &addr->addr.id); - if (res) - return res; + if ((sock->state != SS_CONNECTED) && + ((peer != 2) || (sock->state != SS_DISCONNECTING))) + return -ENOTCONN; + addr->addr.id.ref = tsock->peer_name.ref; + addr->addr.id.node = tsock->peer_name.node; } else { - tipc_ownidentity(portref, &addr->addr.id); + tipc_ownidentity(tsock->p->ref, &addr->addr.id); } *uaddr_len = sizeof(*addr); @@ -766,18 +769,17 @@ exit: static int auto_connect(struct socket *sock, struct tipc_msg *msg) { - struct tipc_port *tport = tipc_sk_port(sock->sk); - struct tipc_portid peer; + struct tipc_sock *tsock = tipc_sk(sock->sk); if (msg_errcode(msg)) { sock->state = SS_DISCONNECTING; return -ECONNREFUSED; } - peer.ref = msg_origport(msg); - peer.node = msg_orignode(msg); - tipc_connect2port(tport->ref, &peer); - tipc_set_portimportance(tport->ref, msg_importance(msg)); + tsock->peer_name.ref = msg_origport(msg); + tsock->peer_name.node = msg_orignode(msg); + tipc_connect2port(tsock->p->ref, &tsock->peer_name); + tipc_set_portimportance(tsock->p->ref, msg_importance(msg)); sock->state = SS_CONNECTED; return 0; } @@ -1529,9 +1531,9 @@ static int accept(struct socket *sock, struct socket *new_sock, int flags) res = tipc_create(sock_net(sock->sk), new_sock, 0); if (!res) { struct sock *new_sk = new_sock->sk; - struct tipc_port *new_tport = tipc_sk_port(new_sk); + struct tipc_sock *new_tsock = tipc_sk(new_sk); + struct tipc_port *new_tport = new_tsock->p; u32 new_ref = new_tport->ref; - struct tipc_portid id; struct tipc_msg *msg = buf_msg(buf); lock_sock(new_sk); @@ -1545,9 +1547,9 @@ static int accept(struct socket *sock, struct socket *new_sock, int flags) /* Connect new socket to it's peer */ - id.ref = msg_origport(msg); - id.node = msg_orignode(msg); - tipc_connect2port(new_ref, &id); + new_tsock->peer_name.ref = msg_origport(msg); + new_tsock->peer_name.node = msg_orignode(msg); + tipc_connect2port(new_ref, &new_tsock->peer_name); new_sock->state = SS_CONNECTED; tipc_set_portimportance(new_ref, msg_importance(msg)); -- cgit v1.2.2 From 0e35fd5e5264bb46d1febbe9cd9aa08421c21a96 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 14 Jul 2008 22:44:01 -0700 Subject: tipc: Eliminate improper use of TIPC_OK error code This patch corrects many places where TIPC routines indicated successful completion by returning TIPC_OK instead of 0. (The TIPC_OK symbol has the value 0, but it should only be used in contexts that deal with the error code field of a TIPC message header.) Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/bcast.c | 10 +++++----- net/tipc/bearer.c | 8 ++++---- net/tipc/cluster.c | 2 +- net/tipc/eth_media.c | 6 +++--- net/tipc/link.c | 18 +++++++++--------- net/tipc/net.c | 4 ++-- net/tipc/port.c | 34 +++++++++++++++++----------------- net/tipc/ref.c | 2 +- net/tipc/user_reg.c | 14 +++++++------- 9 files changed, 49 insertions(+), 49 deletions(-) (limited to 'net') diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c index a5883b1452ff..b1ff16aa4bdb 100644 --- a/net/tipc/bcast.c +++ b/net/tipc/bcast.c @@ -611,7 +611,7 @@ swap: bcbearer->bpairs[bp_index].secondary = p; update: if (bcbearer->remains_new.count == 0) - return TIPC_OK; + return 0; bcbearer->remains = bcbearer->remains_new; } @@ -620,7 +620,7 @@ update: bcbearer->bearer.publ.blocked = 1; bcl->stats.bearer_congs++; - return ~TIPC_OK; + return 1; } /** @@ -756,7 +756,7 @@ int tipc_bclink_reset_stats(void) spin_lock_bh(&bc_lock); memset(&bcl->stats, 0, sizeof(bcl->stats)); spin_unlock_bh(&bc_lock); - return TIPC_OK; + return 0; } int tipc_bclink_set_queue_limits(u32 limit) @@ -769,7 +769,7 @@ int tipc_bclink_set_queue_limits(u32 limit) spin_lock_bh(&bc_lock); tipc_link_set_queue_limits(bcl, limit); spin_unlock_bh(&bc_lock); - return TIPC_OK; + return 0; } int tipc_bclink_init(void) @@ -810,7 +810,7 @@ int tipc_bclink_init(void) tipc_printbuf_init(&bcl->print_buf, pb, BCLINK_LOG_BUF_SIZE); } - return TIPC_OK; + return 0; } void tipc_bclink_stop(void) diff --git a/net/tipc/bearer.c b/net/tipc/bearer.c index 271a375b49b7..6a9aba3edd08 100644 --- a/net/tipc/bearer.c +++ b/net/tipc/bearer.c @@ -370,7 +370,7 @@ void tipc_bearer_remove_dest(struct bearer *b_ptr, u32 dest) */ static int bearer_push(struct bearer *b_ptr) { - u32 res = TIPC_OK; + u32 res = 0; struct link *ln, *tln; if (b_ptr->publ.blocked) @@ -607,7 +607,7 @@ int tipc_block_bearer(const char *name) } spin_unlock_bh(&b_ptr->publ.lock); read_unlock_bh(&tipc_net_lock); - return TIPC_OK; + return 0; } /** @@ -645,7 +645,7 @@ static int bearer_disable(const char *name) } spin_unlock_bh(&b_ptr->publ.lock); memset(b_ptr, 0, sizeof(struct bearer)); - return TIPC_OK; + return 0; } int tipc_disable_bearer(const char *name) @@ -668,7 +668,7 @@ int tipc_bearer_init(void) tipc_bearers = kcalloc(MAX_BEARERS, sizeof(struct bearer), GFP_ATOMIC); media_list = kcalloc(MAX_MEDIA, sizeof(struct media), GFP_ATOMIC); if (tipc_bearers && media_list) { - res = TIPC_OK; + res = 0; } else { kfree(tipc_bearers); kfree(media_list); diff --git a/net/tipc/cluster.c b/net/tipc/cluster.c index bc1db474fe01..46ee6c58532d 100644 --- a/net/tipc/cluster.c +++ b/net/tipc/cluster.c @@ -571,6 +571,6 @@ exit: int tipc_cltr_init(void) { tipc_highest_allowed_slave = LOWEST_SLAVE + tipc_max_slaves; - return tipc_cltr_create(tipc_own_addr) ? TIPC_OK : -ENOMEM; + return tipc_cltr_create(tipc_own_addr) ? 0 : -ENOMEM; } diff --git a/net/tipc/eth_media.c b/net/tipc/eth_media.c index 9cd35eec3e7f..bc72fbc4f8b8 100644 --- a/net/tipc/eth_media.c +++ b/net/tipc/eth_media.c @@ -82,7 +82,7 @@ static int send_msg(struct sk_buff *buf, struct tipc_bearer *tb_ptr, dev->dev_addr, clone->len); dev_queue_xmit(clone); } - return TIPC_OK; + return 0; } /** @@ -113,12 +113,12 @@ static int recv_msg(struct sk_buff *buf, struct net_device *dev, if (likely(buf->len == size)) { buf->next = NULL; tipc_recv_msg(buf, eb_ptr->bearer); - return TIPC_OK; + return 0; } } } kfree_skb(buf); - return TIPC_OK; + return 0; } /** diff --git a/net/tipc/link.c b/net/tipc/link.c index 9784a8e963b4..d60113ba4b1b 100644 --- a/net/tipc/link.c +++ b/net/tipc/link.c @@ -1561,7 +1561,7 @@ u32 tipc_link_push_packet(struct link *l_ptr) l_ptr->retransm_queue_head = mod(++r_q_head); l_ptr->retransm_queue_size = --r_q_size; l_ptr->stats.retransmitted++; - return TIPC_OK; + return 0; } else { l_ptr->stats.bearer_congs++; msg_dbg(buf_msg(buf), "|>DEF-RETR>"); @@ -1580,7 +1580,7 @@ u32 tipc_link_push_packet(struct link *l_ptr) l_ptr->unacked_window = 0; buf_discard(buf); l_ptr->proto_msg_queue = NULL; - return TIPC_OK; + return 0; } else { msg_dbg(buf_msg(buf), "|>DEF-PROT>"); l_ptr->stats.bearer_congs++; @@ -1604,7 +1604,7 @@ u32 tipc_link_push_packet(struct link *l_ptr) msg_set_type(msg, CLOSED_MSG); msg_dbg(msg, ">PUSH-DATA>"); l_ptr->next_out = buf->next; - return TIPC_OK; + return 0; } else { msg_dbg(msg, "|PUSH-DATA|"); l_ptr->stats.bearer_congs++; @@ -1628,8 +1628,8 @@ void tipc_link_push_queue(struct link *l_ptr) do { res = tipc_link_push_packet(l_ptr); - } - while (res == TIPC_OK); + } while (!res); + if (res == PUSH_FAILED) tipc_bearer_schedule(l_ptr->b_ptr, l_ptr); } @@ -2998,7 +2998,7 @@ struct sk_buff *tipc_link_cmd_config(const void *req_tlv_area, int req_tlv_space link_set_supervision_props(l_ptr, new_value); tipc_link_send_proto_msg(l_ptr, STATE_MSG, 0, 0, new_value, 0, 0); - res = TIPC_OK; + res = 0; } break; case TIPC_CMD_SET_LINK_PRI: @@ -3007,14 +3007,14 @@ struct sk_buff *tipc_link_cmd_config(const void *req_tlv_area, int req_tlv_space l_ptr->priority = new_value; tipc_link_send_proto_msg(l_ptr, STATE_MSG, 0, 0, 0, new_value, 0); - res = TIPC_OK; + res = 0; } break; case TIPC_CMD_SET_LINK_WINDOW: if ((new_value >= TIPC_MIN_LINK_WIN) && (new_value <= TIPC_MAX_LINK_WIN)) { tipc_link_set_queue_limits(l_ptr, new_value); - res = TIPC_OK; + res = 0; } break; } @@ -3230,7 +3230,7 @@ int link_control(const char *name, u32 op, u32 val) if (op == TIPC_CMD_UNBLOCK_LINK) { l_ptr->blocked = 0; } - res = TIPC_OK; + res = 0; } tipc_node_unlock(node); } diff --git a/net/tipc/net.c b/net/tipc/net.c index cc51fa483672..ec7b04fbdc43 100644 --- a/net/tipc/net.c +++ b/net/tipc/net.c @@ -165,7 +165,7 @@ static int net_init(void) if (!tipc_net.zones) { return -ENOMEM; } - return TIPC_OK; + return 0; } static void net_stop(void) @@ -295,7 +295,7 @@ int tipc_net_start(u32 addr) info("Started in network mode\n"); info("Own node address %s, network identity %u\n", addr_string_fill(addr_string, tipc_own_addr), tipc_net_id); - return TIPC_OK; + return 0; } void tipc_net_stop(void) diff --git a/net/tipc/port.c b/net/tipc/port.c index ffba1e7f06d2..e70d27ea6578 100644 --- a/net/tipc/port.c +++ b/net/tipc/port.c @@ -291,7 +291,7 @@ int tipc_deleteport(u32 ref) kfree(p_ptr); dbg("Deleted port %u\n", ref); tipc_net_route_msg(buf); - return TIPC_OK; + return 0; } /** @@ -336,7 +336,7 @@ int tipc_portunreliable(u32 ref, unsigned int *isunreliable) return -EINVAL; *isunreliable = port_unreliable(p_ptr); tipc_port_unlock(p_ptr); - return TIPC_OK; + return 0; } int tipc_set_portunreliable(u32 ref, unsigned int isunreliable) @@ -348,7 +348,7 @@ int tipc_set_portunreliable(u32 ref, unsigned int isunreliable) return -EINVAL; msg_set_src_droppable(&p_ptr->publ.phdr, (isunreliable != 0)); tipc_port_unlock(p_ptr); - return TIPC_OK; + return 0; } static int port_unreturnable(struct port *p_ptr) @@ -365,7 +365,7 @@ int tipc_portunreturnable(u32 ref, unsigned int *isunrejectable) return -EINVAL; *isunrejectable = port_unreturnable(p_ptr); tipc_port_unlock(p_ptr); - return TIPC_OK; + return 0; } int tipc_set_portunreturnable(u32 ref, unsigned int isunrejectable) @@ -377,7 +377,7 @@ int tipc_set_portunreturnable(u32 ref, unsigned int isunrejectable) return -EINVAL; msg_set_dest_droppable(&p_ptr->publ.phdr, (isunrejectable != 0)); tipc_port_unlock(p_ptr); - return TIPC_OK; + return 0; } /* @@ -963,7 +963,7 @@ static u32 port_dispatcher(struct tipc_port *dummy, struct sk_buff *buf) tipc_k_signal((Handler)port_dispatcher_sigh, 0); } spin_unlock_bh(&queue_lock); - return TIPC_OK; + return 0; } /* @@ -1068,14 +1068,14 @@ int tipc_createport(u32 user_ref, tipc_reg_add_port(up_ptr); *portref = p_ptr->publ.ref; tipc_port_unlock(p_ptr); - return TIPC_OK; + return 0; } int tipc_ownidentity(u32 ref, struct tipc_portid *id) { id->ref = ref; id->node = tipc_own_addr; - return TIPC_OK; + return 0; } int tipc_portimportance(u32 ref, unsigned int *importance) @@ -1087,7 +1087,7 @@ int tipc_portimportance(u32 ref, unsigned int *importance) return -EINVAL; *importance = (unsigned int)msg_importance(&p_ptr->publ.phdr); tipc_port_unlock(p_ptr); - return TIPC_OK; + return 0; } int tipc_set_portimportance(u32 ref, unsigned int imp) @@ -1102,7 +1102,7 @@ int tipc_set_portimportance(u32 ref, unsigned int imp) return -EINVAL; msg_set_importance(&p_ptr->publ.phdr, (u32)imp); tipc_port_unlock(p_ptr); - return TIPC_OK; + return 0; } @@ -1137,7 +1137,7 @@ int tipc_publish(u32 ref, unsigned int scope, struct tipc_name_seq const *seq) list_add(&publ->pport_list, &p_ptr->publications); p_ptr->pub_count++; p_ptr->publ.published = 1; - res = TIPC_OK; + res = 0; } exit: tipc_port_unlock(p_ptr); @@ -1160,7 +1160,7 @@ int tipc_withdraw(u32 ref, unsigned int scope, struct tipc_name_seq const *seq) tipc_nametbl_withdraw(publ->type, publ->lower, publ->ref, publ->key); } - res = TIPC_OK; + res = 0; } else { list_for_each_entry_safe(publ, tpubl, &p_ptr->publications, pport_list) { @@ -1174,7 +1174,7 @@ int tipc_withdraw(u32 ref, unsigned int scope, struct tipc_name_seq const *seq) break; tipc_nametbl_withdraw(publ->type, publ->lower, publ->ref, publ->key); - res = TIPC_OK; + res = 0; break; } } @@ -1218,7 +1218,7 @@ int tipc_connect2port(u32 ref, struct tipc_portid const *peer) tipc_nodesub_subscribe(&p_ptr->subscription,peer->node, (void *)(unsigned long)ref, (net_ev_handler)port_handle_node_down); - res = TIPC_OK; + res = 0; exit: tipc_port_unlock(p_ptr); p_ptr->publ.max_pkt = tipc_link_get_max_pkt(peer->node, ref); @@ -1240,7 +1240,7 @@ int tipc_disconnect_port(struct tipc_port *tp_ptr) /* let timer expire on it's own to avoid deadlock! */ tipc_nodesub_unsubscribe( &((struct port *)tp_ptr)->subscription); - res = TIPC_OK; + res = 0; } else { res = -ENOTCONN; } @@ -1305,7 +1305,7 @@ int tipc_isconnected(u32 ref, int *isconnected) return -EINVAL; *isconnected = p_ptr->publ.connected; tipc_port_unlock(p_ptr); - return TIPC_OK; + return 0; } int tipc_peer(u32 ref, struct tipc_portid *peer) @@ -1319,7 +1319,7 @@ int tipc_peer(u32 ref, struct tipc_portid *peer) if (p_ptr->publ.connected) { peer->ref = port_peerport(p_ptr); peer->node = port_peernode(p_ptr); - res = TIPC_OK; + res = 0; } else res = -ENOTCONN; tipc_port_unlock(p_ptr); diff --git a/net/tipc/ref.c b/net/tipc/ref.c index a101de86824d..414fc34b8bea 100644 --- a/net/tipc/ref.c +++ b/net/tipc/ref.c @@ -123,7 +123,7 @@ int tipc_ref_table_init(u32 requested_size, u32 start) tipc_ref_table.index_mask = actual_size - 1; tipc_ref_table.start_mask = start & ~tipc_ref_table.index_mask; - return TIPC_OK; + return 0; } /** diff --git a/net/tipc/user_reg.c b/net/tipc/user_reg.c index 4146c40cd20b..506928803162 100644 --- a/net/tipc/user_reg.c +++ b/net/tipc/user_reg.c @@ -91,7 +91,7 @@ static int reg_init(void) } } spin_unlock_bh(®_lock); - return users ? TIPC_OK : -ENOMEM; + return users ? 0 : -ENOMEM; } /** @@ -129,7 +129,7 @@ int tipc_reg_start(void) tipc_k_signal((Handler)reg_callback, (unsigned long)&users[u]); } - return TIPC_OK; + return 0; } /** @@ -184,7 +184,7 @@ int tipc_attach(u32 *userid, tipc_mode_event cb, void *usr_handle) if (cb && (tipc_mode != TIPC_NOT_RUNNING)) tipc_k_signal((Handler)reg_callback, (unsigned long)user_ptr); - return TIPC_OK; + return 0; } /** @@ -230,7 +230,7 @@ int tipc_reg_add_port(struct user_port *up_ptr) struct tipc_user *user_ptr; if (up_ptr->user_ref == 0) - return TIPC_OK; + return 0; if (up_ptr->user_ref > MAX_USERID) return -EINVAL; if ((tipc_mode == TIPC_NOT_RUNNING) || !users ) @@ -240,7 +240,7 @@ int tipc_reg_add_port(struct user_port *up_ptr) user_ptr = &users[up_ptr->user_ref]; list_add(&up_ptr->uport_list, &user_ptr->ports); spin_unlock_bh(®_lock); - return TIPC_OK; + return 0; } /** @@ -250,7 +250,7 @@ int tipc_reg_add_port(struct user_port *up_ptr) int tipc_reg_remove_port(struct user_port *up_ptr) { if (up_ptr->user_ref == 0) - return TIPC_OK; + return 0; if (up_ptr->user_ref > MAX_USERID) return -EINVAL; if (!users ) @@ -259,6 +259,6 @@ int tipc_reg_remove_port(struct user_port *up_ptr) spin_lock_bh(®_lock); list_del_init(&up_ptr->uport_list); spin_unlock_bh(®_lock); - return TIPC_OK; + return 0; } -- cgit v1.2.2 From 08d2cf0f74b3ee5e773bb906043a0efe96ded229 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 14 Jul 2008 22:44:32 -0700 Subject: tipc: Fix bug in scope checking for multicast messages This patch ensures that TIPC's multicast message name lookup algorithm does individualized scope checking for each published name it examines. Previously, scope checking was only done for the first name in a name table publication list, which could result in incoming multicast messages being delivered to ports publishing names with "node" scope, or not being delivered to ports publishing names with "cluster" or "zone" scope. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/name_table.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/tipc/name_table.c b/net/tipc/name_table.c index 096f7bd240a0..1e53b0c8e73d 100644 --- a/net/tipc/name_table.c +++ b/net/tipc/name_table.c @@ -710,9 +710,11 @@ int tipc_nametbl_mc_translate(u32 type, u32 lower, u32 upper, u32 limit, if (sseq->lower > upper) break; publ = sseq->cluster_list; - if (publ && (publ->scope <= limit)) + if (publ) do { - if (publ->node == tipc_own_addr) + if (publ->scope > limit) + /* ignore out-of-scope publication */ ; + else if (publ->node == tipc_own_addr) tipc_port_list_add(dports, publ->ref); else res = 1; -- cgit v1.2.2 From 1aad72d6cd518872c5f545320823bf7f4dafb026 Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 14 Jul 2008 22:44:58 -0700 Subject: tipc: Add missing locks when inspecting node list & link list This patch ensures that TIPC configuration commands that display info about neighboring nodes and their links take the spinlocks that protect the node list and link lists from changing while the lists are being traversed. Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/node.c | 29 ++++++++++++++++++++++++----- 1 file changed, 24 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/tipc/node.c b/net/tipc/node.c index 34e9a2bb7c19..ee952ad60218 100644 --- a/net/tipc/node.c +++ b/net/tipc/node.c @@ -600,12 +600,14 @@ u32 tipc_available_nodes(const u32 domain) struct node *n_ptr; u32 cnt = 0; + read_lock_bh(&tipc_net_lock); for (n_ptr = tipc_nodes; n_ptr; n_ptr = n_ptr->next) { if (!in_scope(domain, n_ptr->addr)) continue; if (tipc_node_is_up(n_ptr)) cnt++; } + read_unlock_bh(&tipc_net_lock); return cnt; } @@ -625,19 +627,26 @@ struct sk_buff *tipc_node_get_nodes(const void *req_tlv_area, int req_tlv_space) return tipc_cfg_reply_error_string(TIPC_CFG_INVALID_VALUE " (network address)"); - if (!tipc_nodes) + read_lock_bh(&tipc_net_lock); + if (!tipc_nodes) { + read_unlock_bh(&tipc_net_lock); return tipc_cfg_reply_none(); + } /* For now, get space for all other nodes (will need to modify this when slave nodes are supported */ payload_size = TLV_SPACE(sizeof(node_info)) * (tipc_max_nodes - 1); - if (payload_size > 32768u) + if (payload_size > 32768u) { + read_unlock_bh(&tipc_net_lock); return tipc_cfg_reply_error_string(TIPC_CFG_NOT_SUPPORTED " (too many nodes)"); + } buf = tipc_cfg_reply_alloc(payload_size); - if (!buf) + if (!buf) { + read_unlock_bh(&tipc_net_lock); return NULL; + } /* Add TLVs for all nodes in scope */ @@ -650,6 +659,7 @@ struct sk_buff *tipc_node_get_nodes(const void *req_tlv_area, int req_tlv_space) &node_info, sizeof(node_info)); } + read_unlock_bh(&tipc_net_lock); return buf; } @@ -672,16 +682,22 @@ struct sk_buff *tipc_node_get_links(const void *req_tlv_area, int req_tlv_space) if (tipc_mode != TIPC_NET_MODE) return tipc_cfg_reply_none(); + read_lock_bh(&tipc_net_lock); + /* Get space for all unicast links + multicast link */ payload_size = TLV_SPACE(sizeof(link_info)) * (tipc_net.zones[tipc_zone(tipc_own_addr)]->links + 1); - if (payload_size > 32768u) + if (payload_size > 32768u) { + read_unlock_bh(&tipc_net_lock); return tipc_cfg_reply_error_string(TIPC_CFG_NOT_SUPPORTED " (too many links)"); + } buf = tipc_cfg_reply_alloc(payload_size); - if (!buf) + if (!buf) { + read_unlock_bh(&tipc_net_lock); return NULL; + } /* Add TLV for broadcast link */ @@ -697,6 +713,7 @@ struct sk_buff *tipc_node_get_links(const void *req_tlv_area, int req_tlv_space) if (!in_scope(domain, n_ptr->addr)) continue; + tipc_node_lock(n_ptr); for (i = 0; i < MAX_BEARERS; i++) { if (!n_ptr->links[i]) continue; @@ -706,7 +723,9 @@ struct sk_buff *tipc_node_get_links(const void *req_tlv_area, int req_tlv_space) tipc_cfg_append_tlv(buf, TIPC_TLV_LINK_INFO, &link_info, sizeof(link_info)); } + tipc_node_unlock(n_ptr); } + read_unlock_bh(&tipc_net_lock); return buf; } -- cgit v1.2.2 From 968edbe1c82f1a50d80225ed7e410aba419e55bf Mon Sep 17 00:00:00 2001 From: Allan Stephens Date: Mon, 14 Jul 2008 22:45:33 -0700 Subject: tipc: Optimization to multicast name lookup algorithm This patch simplifies and speeds up TIPC's algorithm for identifying on-node and off-node destinations that overlap a multicast name sequence range. Rather than traversing the list of all known name publications within the cluster, it now traverses the (potentially much shorter) list of name publications made by the node itself, and determines if any off-node destinations exist by comparing the sizes of the two lists. (Since the node list must be a subset of the cluster list, a difference in sizes means that at least one off-node destination must exist.) Signed-off-by: Allan Stephens Signed-off-by: David S. Miller --- net/tipc/name_table.c | 43 ++++++++++++++++++++++++++++++------------- 1 file changed, 30 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/tipc/name_table.c b/net/tipc/name_table.c index 1e53b0c8e73d..cd72e22b132b 100644 --- a/net/tipc/name_table.c +++ b/net/tipc/name_table.c @@ -2,7 +2,7 @@ * net/tipc/name_table.c: TIPC name table code * * Copyright (c) 2000-2006, Ericsson AB - * Copyright (c) 2004-2005, Wind River Systems + * Copyright (c) 2004-2008, Wind River Systems * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -52,9 +52,16 @@ static int tipc_nametbl_size = 1024; /* must be a power of 2 */ * struct sub_seq - container for all published instances of a name sequence * @lower: name sequence lower bound * @upper: name sequence upper bound - * @node_list: circular list of matching publications with >= node scope - * @cluster_list: circular list of matching publications with >= cluster scope - * @zone_list: circular list of matching publications with >= zone scope + * @node_list: circular list of publications made by own node + * @cluster_list: circular list of publications made by own cluster + * @zone_list: circular list of publications made by own zone + * @node_list_size: number of entries in "node_list" + * @cluster_list_size: number of entries in "cluster_list" + * @zone_list_size: number of entries in "zone_list" + * + * Note: The zone list always contains at least one entry, since all + * publications of the associated name sequence belong to it. + * (The cluster and node lists may be empty.) */ struct sub_seq { @@ -63,6 +70,9 @@ struct sub_seq { struct publication *node_list; struct publication *cluster_list; struct publication *zone_list; + u32 node_list_size; + u32 cluster_list_size; + u32 zone_list_size; }; /** @@ -317,6 +327,7 @@ static struct publication *tipc_nameseq_insert_publ(struct name_seq *nseq, dbg("inserting publ %p, node=0x%x publ->node=0x%x, subscr->node=%p\n", publ, node, publ->node, publ->subscr.node); + sseq->zone_list_size++; if (!sseq->zone_list) sseq->zone_list = publ->zone_list_next = publ; else { @@ -325,6 +336,7 @@ static struct publication *tipc_nameseq_insert_publ(struct name_seq *nseq, } if (in_own_cluster(node)) { + sseq->cluster_list_size++; if (!sseq->cluster_list) sseq->cluster_list = publ->cluster_list_next = publ; else { @@ -335,6 +347,7 @@ static struct publication *tipc_nameseq_insert_publ(struct name_seq *nseq, } if (node == tipc_own_addr) { + sseq->node_list_size++; if (!sseq->node_list) sseq->node_list = publ->node_list_next = publ; else { @@ -411,6 +424,7 @@ static struct publication *tipc_nameseq_remove_publ(struct name_seq *nseq, u32 i } else { sseq->zone_list = NULL; } + sseq->zone_list_size--; /* Remove publication from cluster scope list, if present */ @@ -439,6 +453,7 @@ static struct publication *tipc_nameseq_remove_publ(struct name_seq *nseq, u32 i } else { sseq->cluster_list = NULL; } + sseq->cluster_list_size--; } end_cluster: @@ -469,6 +484,7 @@ end_cluster: } else { sseq->node_list = NULL; } + sseq->node_list_size--; } end_node: @@ -709,17 +725,18 @@ int tipc_nametbl_mc_translate(u32 type, u32 lower, u32 upper, u32 limit, if (sseq->lower > upper) break; - publ = sseq->cluster_list; - if (publ) + + publ = sseq->node_list; + if (publ) { do { - if (publ->scope > limit) - /* ignore out-of-scope publication */ ; - else if (publ->node == tipc_own_addr) + if (publ->scope <= limit) tipc_port_list_add(dports, publ->ref); - else - res = 1; - publ = publ->cluster_list_next; - } while (publ != sseq->cluster_list); + publ = publ->node_list_next; + } while (publ != sseq->node_list); + } + + if (sseq->cluster_list_size != sseq->node_list_size) + res = 1; } spin_unlock_bh(&seq->lock); -- cgit v1.2.2 From 6aa895b047720f71ec4eb11452f7c3ce8426941f Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 14 Jul 2008 22:49:06 -0700 Subject: vlan: Don't store VLAN tag in cb Use a real skb member to store the skb to avoid clashes with qdiscs, which are allowed to use the cb area themselves. As currently only real devices that consume the skb set the NETIF_F_HW_VLAN_TX flag, no explicit invalidation is neccessary. The new member fills a hole on 64 bit, the skb layout changes from: __u32 mark; /* 172 4 */ sk_buff_data_t transport_header; /* 176 4 */ sk_buff_data_t network_header; /* 180 4 */ sk_buff_data_t mac_header; /* 184 4 */ sk_buff_data_t tail; /* 188 4 */ /* --- cacheline 3 boundary (192 bytes) --- */ sk_buff_data_t end; /* 192 4 */ /* XXX 4 bytes hole, try to pack */ to __u32 mark; /* 172 4 */ __u16 vlan_tci; /* 176 2 */ /* XXX 2 bytes hole, try to pack */ sk_buff_data_t transport_header; /* 180 4 */ sk_buff_data_t network_header; /* 184 4 */ Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/core/skbuff.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'net') diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 7c571560e9d2..50a853f7cd8e 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -459,6 +459,8 @@ static void __copy_skb_header(struct sk_buff *new, const struct sk_buff *old) new->tc_verd = old->tc_verd; #endif #endif + new->vlan_tci = old->vlan_tci; + skb_copy_secmark(new, old); } @@ -2286,6 +2288,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, int features) skb_copy_queue_mapping(nskb, skb); nskb->priority = skb->priority; nskb->protocol = skb->protocol; + nskb->vlan_tci = skb->vlan_tci; nskb->dst = dst_clone(skb->dst); memcpy(nskb->cb, skb->cb, sizeof(skb->cb)); nskb->pkt_type = skb->pkt_type; -- cgit v1.2.2 From bc1d0411b804ad190cdadabac48a10067f17b9e6 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 14 Jul 2008 22:49:30 -0700 Subject: vlan: deliver packets received with VLAN acceleration to network taps When VLAN header stripping is used, packets currently bypass packet sockets (and other network taps) completely. For locally existing VLANs, they appear directly on the VLAN device, for unknown VLANs they are silently dropped. Add a new function netif_nit_deliver() to deliver incoming packets to all network interface taps and use it in __vlan_hwaccel_rx() to make VLAN packets visible on the underlying device. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan_core.c | 4 ++++ net/core/dev.c | 27 +++++++++++++++++++++++++++ 2 files changed, 31 insertions(+) (limited to 'net') diff --git a/net/8021q/vlan_core.c b/net/8021q/vlan_core.c index 68df12d3664b..916061f681b6 100644 --- a/net/8021q/vlan_core.c +++ b/net/8021q/vlan_core.c @@ -14,6 +14,9 @@ int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp, return NET_RX_DROP; } + skb->vlan_tci = vlan_tci; + netif_nit_deliver(skb); + skb->dev = vlan_group_get_device(grp, vlan_tci & VLAN_VID_MASK); if (skb->dev == NULL) { dev_kfree_skb_any(skb); @@ -22,6 +25,7 @@ int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp, return NET_RX_SUCCESS; } skb->dev->last_rx = jiffies; + skb->vlan_tci = 0; stats = &skb->dev->stats; stats->rx_packets++; diff --git a/net/core/dev.c b/net/core/dev.c index a29a359b15d1..feaab4898a5b 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2068,6 +2068,33 @@ out: } #endif +/* + * netif_nit_deliver - deliver received packets to network taps + * @skb: buffer + * + * This function is used to deliver incoming packets to network + * taps. It should be used when the normal netif_receive_skb path + * is bypassed, for example because of VLAN acceleration. + */ +void netif_nit_deliver(struct sk_buff *skb) +{ + struct packet_type *ptype; + + if (list_empty(&ptype_all)) + return; + + skb_reset_network_header(skb); + skb_reset_transport_header(skb); + skb->mac_len = skb->network_header - skb->mac_header; + + rcu_read_lock(); + list_for_each_entry_rcu(ptype, &ptype_all, list) { + if (!ptype->dev || ptype->dev == skb->dev) + deliver_skb(skb, ptype, skb->dev); + } + rcu_read_unlock(); +} + /** * netif_receive_skb - process receive buffer from network * @skb: buffer to process -- cgit v1.2.2 From bbd6ef87c544d88c30e4b762b1b61ef267a7d279 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 14 Jul 2008 22:50:15 -0700 Subject: packet: support extensible, 64 bit clean mmaped ring structure The tpacket_hdr is not 64 bit clean due to use of an unsigned long and can't be extended because the following struct sockaddr_ll needs to be at a fixed offset. Add support for a version 2 tpacket protocol that removes these limitations. Userspace can query the header size through a new getsockopt option and change the protocol version through a setsockopt option. The changes needed to switch to the new protocol version are: 1. replace struct tpacket_hdr by struct tpacket2_hdr 2. query header len and save 3. set protocol version to 2 - set up ring as usual 4. for getting the sockaddr_ll, use (void *)hdr + TPACKET_ALIGN(hdrlen) instead of (void *)hdr + TPACKET_ALIGN(sizeof(struct tpacket_hdr)) Steps 2 and 4 can be omitted if the struct sockaddr_ll isn't needed. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/packet/af_packet.c | 179 ++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 146 insertions(+), 33 deletions(-) (limited to 'net') diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index 9f2269166687..4f059775d48f 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -186,6 +186,8 @@ struct packet_sock { unsigned int pg_vec_order; unsigned int pg_vec_pages; unsigned int pg_vec_len; + enum tpacket_versions tp_version; + unsigned int tp_hdrlen; #endif }; @@ -201,14 +203,52 @@ struct packet_skb_cb { #ifdef CONFIG_PACKET_MMAP -static inline struct tpacket_hdr *packet_lookup_frame(struct packet_sock *po, unsigned int position) +static void *packet_lookup_frame(struct packet_sock *po, unsigned int position, + int status) { unsigned int pg_vec_pos, frame_offset; + union { + struct tpacket_hdr *h1; + struct tpacket2_hdr *h2; + void *raw; + } h; pg_vec_pos = position / po->frames_per_block; frame_offset = position % po->frames_per_block; - return (struct tpacket_hdr *)(po->pg_vec[pg_vec_pos] + (frame_offset * po->frame_size)); + h.raw = po->pg_vec[pg_vec_pos] + (frame_offset * po->frame_size); + switch (po->tp_version) { + case TPACKET_V1: + if (status != h.h1->tp_status ? TP_STATUS_USER : + TP_STATUS_KERNEL) + return NULL; + break; + case TPACKET_V2: + if (status != h.h2->tp_status ? TP_STATUS_USER : + TP_STATUS_KERNEL) + return NULL; + break; + } + return h.raw; +} + +static void __packet_set_status(struct packet_sock *po, void *frame, int status) +{ + union { + struct tpacket_hdr *h1; + struct tpacket2_hdr *h2; + void *raw; + } h; + + h.raw = frame; + switch (po->tp_version) { + case TPACKET_V1: + h.h1->tp_status = status; + break; + case TPACKET_V2: + h.h2->tp_status = status; + break; + } } #endif @@ -551,14 +591,19 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, struct packe struct sock *sk; struct packet_sock *po; struct sockaddr_ll *sll; - struct tpacket_hdr *h; + union { + struct tpacket_hdr *h1; + struct tpacket2_hdr *h2; + void *raw; + } h; u8 * skb_head = skb->data; int skb_len = skb->len; unsigned int snaplen, res; unsigned long status = TP_STATUS_LOSING|TP_STATUS_USER; - unsigned short macoff, netoff; + unsigned short macoff, netoff, hdrlen; struct sk_buff *copy_skb = NULL; struct timeval tv; + struct timespec ts; if (skb->pkt_type == PACKET_LOOPBACK) goto drop; @@ -590,10 +635,11 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, struct packe snaplen = res; if (sk->sk_type == SOCK_DGRAM) { - macoff = netoff = TPACKET_ALIGN(TPACKET_HDRLEN) + 16; + macoff = netoff = TPACKET_ALIGN(po->tp_hdrlen) + 16; } else { unsigned maclen = skb_network_offset(skb); - netoff = TPACKET_ALIGN(TPACKET_HDRLEN + (maclen < 16 ? 16 : maclen)); + netoff = TPACKET_ALIGN(po->tp_hdrlen + + (maclen < 16 ? 16 : maclen)); macoff = netoff - maclen; } @@ -616,9 +662,8 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, struct packe } spin_lock(&sk->sk_receive_queue.lock); - h = packet_lookup_frame(po, po->head); - - if (h->tp_status) + h.raw = packet_lookup_frame(po, po->head, TP_STATUS_KERNEL); + if (!h.raw) goto ring_is_full; po->head = po->head != po->frame_max ? po->head+1 : 0; po->stats.tp_packets++; @@ -630,20 +675,40 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, struct packe status &= ~TP_STATUS_LOSING; spin_unlock(&sk->sk_receive_queue.lock); - skb_copy_bits(skb, 0, (u8*)h + macoff, snaplen); + skb_copy_bits(skb, 0, h.raw + macoff, snaplen); - h->tp_len = skb->len; - h->tp_snaplen = snaplen; - h->tp_mac = macoff; - h->tp_net = netoff; - if (skb->tstamp.tv64) - tv = ktime_to_timeval(skb->tstamp); - else - do_gettimeofday(&tv); - h->tp_sec = tv.tv_sec; - h->tp_usec = tv.tv_usec; + switch (po->tp_version) { + case TPACKET_V1: + h.h1->tp_len = skb->len; + h.h1->tp_snaplen = snaplen; + h.h1->tp_mac = macoff; + h.h1->tp_net = netoff; + if (skb->tstamp.tv64) + tv = ktime_to_timeval(skb->tstamp); + else + do_gettimeofday(&tv); + h.h1->tp_sec = tv.tv_sec; + h.h1->tp_usec = tv.tv_usec; + hdrlen = sizeof(*h.h1); + break; + case TPACKET_V2: + h.h2->tp_len = skb->len; + h.h2->tp_snaplen = snaplen; + h.h2->tp_mac = macoff; + h.h2->tp_net = netoff; + if (skb->tstamp.tv64) + ts = ktime_to_timespec(skb->tstamp); + else + getnstimeofday(&ts); + h.h2->tp_sec = ts.tv_sec; + h.h2->tp_nsec = ts.tv_nsec; + hdrlen = sizeof(*h.h2); + break; + default: + BUG(); + } - sll = (struct sockaddr_ll*)((u8*)h + TPACKET_ALIGN(sizeof(*h))); + sll = h.raw + TPACKET_ALIGN(hdrlen); sll->sll_halen = dev_parse_header(skb, sll->sll_addr); sll->sll_family = AF_PACKET; sll->sll_hatype = dev->type; @@ -654,14 +719,14 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, struct packe else sll->sll_ifindex = dev->ifindex; - h->tp_status = status; + __packet_set_status(po, h.raw, status); smp_mb(); { struct page *p_start, *p_end; - u8 *h_end = (u8 *)h + macoff + snaplen - 1; + u8 *h_end = h.raw + macoff + snaplen - 1; - p_start = virt_to_page(h); + p_start = virt_to_page(h.raw); p_end = virt_to_page(h_end); while (p_start <= p_end) { flush_dcache_page(p_start); @@ -1362,6 +1427,25 @@ packet_setsockopt(struct socket *sock, int level, int optname, char __user *optv pkt_sk(sk)->copy_thresh = val; return 0; } + case PACKET_VERSION: + { + int val; + + if (optlen != sizeof(val)) + return -EINVAL; + if (po->pg_vec) + return -EBUSY; + if (copy_from_user(&val, optval, sizeof(val))) + return -EFAULT; + switch (val) { + case TPACKET_V1: + case TPACKET_V2: + po->tp_version = val; + return 0; + default: + return -EINVAL; + } + } #endif case PACKET_AUXDATA: { @@ -1437,6 +1521,31 @@ static int packet_getsockopt(struct socket *sock, int level, int optname, data = &val; break; +#ifdef CONFIG_PACKET_MMAP + case PACKET_VERSION: + if (len > sizeof(int)) + len = sizeof(int); + val = po->tp_version; + data = &val; + break; + case PACKET_HDRLEN: + if (len > sizeof(int)) + len = sizeof(int); + if (copy_from_user(&val, optval, len)) + return -EFAULT; + switch (val) { + case TPACKET_V1: + val = sizeof(struct tpacket_hdr); + break; + case TPACKET_V2: + val = sizeof(struct tpacket2_hdr); + break; + default: + return -EINVAL; + } + data = &val; + break; +#endif default: return -ENOPROTOOPT; } @@ -1570,11 +1679,8 @@ static unsigned int packet_poll(struct file * file, struct socket *sock, spin_lock_bh(&sk->sk_receive_queue.lock); if (po->pg_vec) { unsigned last = po->head ? po->head-1 : po->frame_max; - struct tpacket_hdr *h; - - h = packet_lookup_frame(po, last); - if (h->tp_status) + if (packet_lookup_frame(po, last, TP_STATUS_USER)) mask |= POLLIN | POLLRDNORM; } spin_unlock_bh(&sk->sk_receive_queue.lock); @@ -1669,11 +1775,20 @@ static int packet_set_ring(struct sock *sk, struct tpacket_req *req, int closing if (unlikely(po->pg_vec)) return -EBUSY; + switch (po->tp_version) { + case TPACKET_V1: + po->tp_hdrlen = TPACKET_HDRLEN; + break; + case TPACKET_V2: + po->tp_hdrlen = TPACKET2_HDRLEN; + break; + } + if (unlikely((int)req->tp_block_size <= 0)) return -EINVAL; if (unlikely(req->tp_block_size & (PAGE_SIZE - 1))) return -EINVAL; - if (unlikely(req->tp_frame_size < TPACKET_HDRLEN)) + if (unlikely(req->tp_frame_size < po->tp_hdrlen)) return -EINVAL; if (unlikely(req->tp_frame_size & (TPACKET_ALIGNMENT - 1))) return -EINVAL; @@ -1692,13 +1807,11 @@ static int packet_set_ring(struct sock *sk, struct tpacket_req *req, int closing goto out; for (i = 0; i < req->tp_block_nr; i++) { - char *ptr = pg_vec[i]; - struct tpacket_hdr *header; + void *ptr = pg_vec[i]; int k; for (k = 0; k < po->frames_per_block; k++) { - header = (struct tpacket_hdr *) ptr; - header->tp_status = TP_STATUS_KERNEL; + __packet_set_status(po, ptr, TP_STATUS_KERNEL); ptr += req->tp_frame_size; } } -- cgit v1.2.2 From 393e52e33c6c26ec7db290dab803bac1bed962d4 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 14 Jul 2008 22:50:39 -0700 Subject: packet: deliver VLAN TCI to userspace Store the VLAN tag in the auxillary data/tpacket2_hdr so userspace can properly deal with hardware VLAN tagging/stripping. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/packet/af_packet.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'net') diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index 4f059775d48f..db792e02a37f 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -702,6 +702,7 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, struct packe getnstimeofday(&ts); h.h2->tp_sec = ts.tv_sec; h.h2->tp_nsec = ts.tv_nsec; + h.h2->tp_vlan_tci = skb->vlan_tci; hdrlen = sizeof(*h.h2); break; default: @@ -1172,6 +1173,7 @@ static int packet_recvmsg(struct kiocb *iocb, struct socket *sock, aux.tp_snaplen = skb->len; aux.tp_mac = 0; aux.tp_net = skb_network_offset(skb); + aux.tp_vlan_tci = skb->vlan_tci; put_cmsg(msg, SOL_PACKET, PACKET_AUXDATA, sizeof(aux), &aux); } -- cgit v1.2.2 From 19b9a4e256758a0c032c915eebe0a39b370ea133 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 14 Jul 2008 22:51:01 -0700 Subject: vlan: ethtool ->get_flags support Allow to query LRO settings of underlying device when VLAN RX acceleration is used. Suggested by Ben Hutchings . Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan_dev.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'net') diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 9efd3c67c1db..29aa4cc2a260 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -725,9 +725,22 @@ static u32 vlan_ethtool_get_rx_csum(struct net_device *dev) return real_dev->ethtool_ops->get_rx_csum(real_dev); } +static u32 vlan_ethtool_get_flags(struct net_device *dev) +{ + const struct vlan_dev_info *vlan = vlan_dev_info(dev); + struct net_device *real_dev = vlan->real_dev; + + if (!(real_dev->features & NETIF_F_HW_VLAN_RX) || + real_dev->ethtool_ops == NULL || + real_dev->ethtool_ops->get_flags == NULL) + return 0; + return real_dev->ethtool_ops->get_flags(real_dev); +} + static const struct ethtool_ops vlan_ethtool_ops = { .get_link = ethtool_op_get_link, .get_rx_csum = vlan_ethtool_get_rx_csum, + .get_flags = vlan_ethtool_get_flags, }; void vlan_setup(struct net_device *dev) -- cgit v1.2.2 From 1349fe9a6bc580fb80e1a43a93b68c15d674ed0a Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 14 Jul 2008 22:51:19 -0700 Subject: vlan: clean up vlan_dev_hard_header() Remove some debugging and excessive comments, merge the two dev_hard_header calls into one. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan_dev.c | 55 +++++++++------------------------------------------- 1 file changed, 9 insertions(+), 46 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 29aa4cc2a260..c6678605fc03 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -256,43 +256,18 @@ static int vlan_dev_hard_header(struct sk_buff *skb, struct net_device *dev, unsigned int len) { struct vlan_hdr *vhdr; + unsigned int vhdrlen = 0; u16 vlan_tci = 0; - int rc = 0; - int build_vlan_header = 0; - - pr_debug("%s: skb: %p type: %hx len: %u vlan_id: %hx, daddr: %p\n", - __func__, skb, type, len, vlan_dev_info(dev)->vlan_id, - daddr); + int rc; if (WARN_ON(skb_headroom(skb) < dev->hard_header_len)) return -ENOSPC; - /* build vlan header only if re_order_header flag is NOT set. This - * fixes some programs that get confused when they see a VLAN device - * sending a frame that is VLAN encoded (the consensus is that the VLAN - * device should look completely like an Ethernet device when the - * REORDER_HEADER flag is set) The drawback to this is some extra - * header shuffling in the hard_start_xmit. Users can turn off this - * REORDER behaviour with the vconfig tool. - */ - if (!(vlan_dev_info(dev)->flags & VLAN_FLAG_REORDER_HDR)) - build_vlan_header = 1; - - if (build_vlan_header) { + if (!(vlan_dev_info(dev)->flags & VLAN_FLAG_REORDER_HDR)) { vhdr = (struct vlan_hdr *) skb_push(skb, VLAN_HLEN); - /* build the four bytes that make this a VLAN header. */ - - /* Now, construct the second two bytes. This field looks - * something like: - * usr_priority: 3 bits (high bits) - * CFI 1 bit - * VLAN ID 12 bits (low bits) - * - */ vlan_tci = vlan_dev_info(dev)->vlan_id; vlan_tci |= vlan_dev_get_egress_qos_mask(dev, skb); - vhdr->h_vlan_TCI = htons(vlan_tci); /* @@ -300,37 +275,25 @@ static int vlan_dev_hard_header(struct sk_buff *skb, struct net_device *dev, * put the length in here instead. It is up to the 802.2 * layer to carry protocol information. */ - if (type != ETH_P_802_3) vhdr->h_vlan_encapsulated_proto = htons(type); else vhdr->h_vlan_encapsulated_proto = htons(len); skb->protocol = htons(ETH_P_8021Q); + type = ETH_P_8021Q; + vhdrlen = VLAN_HLEN; } /* Before delegating work to the lower layer, enter our MAC-address */ if (saddr == NULL) saddr = dev->dev_addr; + /* Now make the underlying real hard header */ dev = vlan_dev_info(dev)->real_dev; - - if (build_vlan_header) { - /* Now make the underlying real hard header */ - rc = dev_hard_header(skb, dev, ETH_P_8021Q, daddr, saddr, - len + VLAN_HLEN); - if (rc > 0) - rc += VLAN_HLEN; - else if (rc < 0) - rc -= VLAN_HLEN; - } else - /* If here, then we'll just make a normal looking ethernet - * frame, but, the hard_start_xmit method will insert the tag - * (it has to be able to do this for bridged and other skbs - * that don't come down the protocol stack in an orderly manner. - */ - rc = dev_hard_header(skb, dev, type, daddr, saddr, len); - + rc = dev_hard_header(skb, dev, type, daddr, saddr, len + vhdrlen); + if (rc > 0) + rc += vhdrlen; return rc; } -- cgit v1.2.2 From d80aa31bbffc4bc8b5be36c57cbea128f52e1e1f Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 14 Jul 2008 22:51:39 -0700 Subject: vlan: clean up hard_start_xmit functions Remove excessive comments and debugging, use NETDEV_TX codes, remove some empty lines. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan_dev.c | 41 ++++++----------------------------------- 1 file changed, 6 insertions(+), 35 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index c6678605fc03..c3adba9501cb 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -307,53 +307,31 @@ static int vlan_dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev) * NOTE: THIS ASSUMES DIX ETHERNET, SPECIFICALLY NOT SUPPORTING * OTHER THINGS LIKE FDDI/TokenRing/802.3 SNAPs... */ - if (veth->h_vlan_proto != htons(ETH_P_8021Q) || - vlan_dev_info(dev)->flags & VLAN_FLAG_REORDER_HDR) { - int orig_headroom = skb_headroom(skb); + vlan_dev_info(dev)->flags & VLAN_FLAG_REORDER_HDR) { + unsigned int orig_headroom = skb_headroom(skb); u16 vlan_tci; - /* This is not a VLAN frame...but we can fix that! */ vlan_dev_info(dev)->cnt_encap_on_xmit++; - pr_debug("%s: proto to encap: 0x%hx\n", - __func__, ntohs(veth->h_vlan_proto)); - /* Construct the second two bytes. This field looks something - * like: - * usr_priority: 3 bits (high bits) - * CFI 1 bit - * VLAN ID 12 bits (low bits) - */ vlan_tci = vlan_dev_info(dev)->vlan_id; vlan_tci |= vlan_dev_get_egress_qos_mask(dev, skb); - skb = __vlan_put_tag(skb, vlan_tci); if (!skb) { stats->tx_dropped++; - return 0; + return NETDEV_TX_OK; } if (orig_headroom < VLAN_HLEN) vlan_dev_info(dev)->cnt_inc_headroom_on_tx++; } - pr_debug("%s: about to send skb: %p to dev: %s\n", - __func__, skb, skb->dev->name); - pr_debug(" " MAC_FMT " " MAC_FMT " %4hx %4hx %4hx\n", - veth->h_dest[0], veth->h_dest[1], veth->h_dest[2], - veth->h_dest[3], veth->h_dest[4], veth->h_dest[5], - veth->h_source[0], veth->h_source[1], veth->h_source[2], - veth->h_source[3], veth->h_source[4], veth->h_source[5], - veth->h_vlan_proto, veth->h_vlan_TCI, - veth->h_vlan_encapsulated_proto); - - stats->tx_packets++; /* for statics only */ + stats->tx_packets++; stats->tx_bytes += skb->len; skb->dev = vlan_dev_info(dev)->real_dev; dev_queue_xmit(skb); - - return 0; + return NETDEV_TX_OK; } static int vlan_dev_hwaccel_hard_start_xmit(struct sk_buff *skb, @@ -362,12 +340,6 @@ static int vlan_dev_hwaccel_hard_start_xmit(struct sk_buff *skb, struct net_device_stats *stats = &dev->stats; u16 vlan_tci; - /* Construct the second two bytes. This field looks something - * like: - * usr_priority: 3 bits (high bits) - * CFI 1 bit - * VLAN ID 12 bits (low bits) - */ vlan_tci = vlan_dev_info(dev)->vlan_id; vlan_tci |= vlan_dev_get_egress_qos_mask(dev, skb); skb = __vlan_hwaccel_put_tag(skb, vlan_tci); @@ -377,8 +349,7 @@ static int vlan_dev_hwaccel_hard_start_xmit(struct sk_buff *skb, skb->dev = vlan_dev_info(dev)->real_dev; dev_queue_xmit(skb); - - return 0; + return NETDEV_TX_OK; } static int vlan_dev_change_mtu(struct net_device *dev, int new_mtu) -- cgit v1.2.2 From 61362766d769c934a9d12d5516323c544c161908 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Mon, 14 Jul 2008 22:51:55 -0700 Subject: vlan: remove unnecessary include statements Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/8021q/vlan.c | 8 ++------ net/8021q/vlan_dev.c | 7 ------- net/8021q/vlanproc.c | 11 ++--------- 3 files changed, 4 insertions(+), 22 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c index 68bdcf4a795c..9f04cab92f95 100644 --- a/net/8021q/vlan.c +++ b/net/8021q/vlan.c @@ -18,21 +18,17 @@ * 2 of the License, or (at your option) any later version. */ -#include /* for copy_from_user */ #include #include #include #include -#include -#include -#include #include -#include -#include #include #include +#include #include #include +#include #include #include "vlan.h" diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index c3adba9501cb..6b985f23fd9f 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -21,22 +21,15 @@ */ #include -#include -#include -#include -#include /* for copy_from_user */ #include #include #include #include -#include -#include #include #include "vlan.h" #include "vlanproc.h" #include -#include /* * Rebuild the Ethernet MAC header. This is called after an ARP diff --git a/net/8021q/vlanproc.c b/net/8021q/vlanproc.c index 6073a888b6f9..0feefa4e1a4b 100644 --- a/net/8021q/vlanproc.c +++ b/net/8021q/vlanproc.c @@ -18,16 +18,9 @@ *****************************************************************************/ #include -#include /* offsetof(), etc. */ -#include /* return codes */ +#include #include -#include /* kmalloc(), kfree() */ -#include -#include /* inline mem*, str* functions */ -#include /* __initfunc et al. */ -#include /* htons(), etc. */ -#include /* copy_to_user */ -#include +#include #include #include #include -- cgit v1.2.2 From 0388b0042624714e6f8db8cc7994101a0a02d392 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 14 Jul 2008 23:00:43 -0700 Subject: icmp: add struct net argument to icmp_out_count This routine deals with ICMP statistics, but doesn't have a struct net at hands, so add one. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/icmp.c | 2 +- net/ipv4/ip_output.c | 3 ++- net/ipv4/raw.c | 3 ++- 3 files changed, 5 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c index aa7cf46853b7..1ffe7add4920 100644 --- a/net/ipv4/icmp.c +++ b/net/ipv4/icmp.c @@ -296,7 +296,7 @@ out: /* * Maintain the counters used in the SNMP statistics for outgoing ICMP */ -void icmp_out_count(unsigned char type) +void icmp_out_count(struct net *net, unsigned char type) { ICMPMSGOUT_INC_STATS(type); ICMP_INC_STATS(ICMP_MIB_OUTMSGS); diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c index f1278eecf56d..f003186b93b1 100644 --- a/net/ipv4/ip_output.c +++ b/net/ipv4/ip_output.c @@ -1211,6 +1211,7 @@ int ip_push_pending_frames(struct sock *sk) struct sk_buff *skb, *tmp_skb; struct sk_buff **tail_skb; struct inet_sock *inet = inet_sk(sk); + struct net *net = sock_net(sk); struct ip_options *opt = NULL; struct rtable *rt = (struct rtable *)inet->cork.dst; struct iphdr *iph; @@ -1280,7 +1281,7 @@ int ip_push_pending_frames(struct sock *sk) skb->dst = dst_clone(&rt->u.dst); if (iph->protocol == IPPROTO_ICMP) - icmp_out_count(((struct icmphdr *) + icmp_out_count(net, ((struct icmphdr *) skb_transport_header(skb))->type); /* Netfilter gets whole the not fragmented skb. */ diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c index 925fdf18cf92..7f39ea443ec8 100644 --- a/net/ipv4/raw.c +++ b/net/ipv4/raw.c @@ -320,6 +320,7 @@ static int raw_send_hdrinc(struct sock *sk, void *from, size_t length, unsigned int flags) { struct inet_sock *inet = inet_sk(sk); + struct net *net = sock_net(sk); struct iphdr *iph; struct sk_buff *skb; unsigned int iphlen; @@ -368,7 +369,7 @@ static int raw_send_hdrinc(struct sock *sk, void *from, size_t length, iph->check = ip_fast_csum((unsigned char *)iph, iph->ihl); } if (iph->protocol == IPPROTO_ICMP) - icmp_out_count(((struct icmphdr *) + icmp_out_count(net, ((struct icmphdr *) skb_transport_header(skb))->type); err = NF_HOOK(PF_INET, NF_INET_LOCAL_OUT, skb, NULL, rt->u.dst.dev, -- cgit v1.2.2 From fd54d716b1f6a3551ec17a4bb34027727b2db09a Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 14 Jul 2008 23:01:40 -0700 Subject: inet: toss struct net initialization around Some places, that deal with ICMP statistics already have where to get a struct net from, but use it directly, without declaring a separate variable on the stack. Since I will need this net soon, I declare a struct net on the stack and use it in the existing places in a separate patch not to spoil the future ones. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/dccp/ipv4.c | 3 ++- net/ipv4/icmp.c | 4 +--- net/ipv4/tcp_ipv4.c | 3 ++- net/ipv4/udp.c | 3 ++- 4 files changed, 7 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index 37d27bcb361f..cc4f33461f2d 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -205,13 +205,14 @@ static void dccp_v4_err(struct sk_buff *skb, u32 info) struct sock *sk; __u64 seq; int err; + struct net *net = dev_net(skb->dev); if (skb->len < (iph->ihl << 2) + 8) { ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); return; } - sk = inet_lookup(dev_net(skb->dev), &dccp_hashinfo, + sk = inet_lookup(net, &dccp_hashinfo, iph->daddr, dh->dccph_dport, iph->saddr, dh->dccph_sport, inet_iif(skb)); if (sk == NULL) { diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c index 1ffe7add4920..56d6b943345c 100644 --- a/net/ipv4/icmp.c +++ b/net/ipv4/icmp.c @@ -973,6 +973,7 @@ int icmp_rcv(struct sk_buff *skb) { struct icmphdr *icmph; struct rtable *rt = skb->rtable; + struct net *net = dev_net(rt->u.dst.dev); if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) { int nh; @@ -1027,9 +1028,6 @@ int icmp_rcv(struct sk_buff *skb) */ if (rt->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST)) { - struct net *net; - - net = dev_net(rt->u.dst.dev); /* * RFC 1122: 3.2.2.6 An ICMP_ECHO to broadcast MAY be * silently ignored (we let user decide with a sysctl). diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 4300bcf2ceaf..ca41b77f3f32 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -343,13 +343,14 @@ void tcp_v4_err(struct sk_buff *skb, u32 info) struct sock *sk; __u32 seq; int err; + struct net *net = dev_net(skb->dev); if (skb->len < (iph->ihl << 2) + 8) { ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); return; } - sk = inet_lookup(dev_net(skb->dev), &tcp_hashinfo, iph->daddr, th->dest, + sk = inet_lookup(net, &tcp_hashinfo, iph->daddr, th->dest, iph->saddr, th->source, inet_iif(skb)); if (!sk) { ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 7187121e922d..9342cfda3d0a 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -354,8 +354,9 @@ void __udp4_lib_err(struct sk_buff *skb, u32 info, struct hlist_head udptable[]) struct sock *sk; int harderr; int err; + struct net *net = dev_net(skb->dev); - sk = __udp4_lib_lookup(dev_net(skb->dev), iph->daddr, uh->dest, + sk = __udp4_lib_lookup(net, iph->daddr, uh->dest, iph->saddr, uh->source, skb->dev->ifindex, udptable); if (sk == NULL) { ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); -- cgit v1.2.2 From 75c939bb4d6da790f758a2a3dcc7432f6d8778ee Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 14 Jul 2008 23:02:35 -0700 Subject: mib: add struct net to ICMP_INC_STATS Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/icmp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c index 56d6b943345c..e94de411cfa0 100644 --- a/net/ipv4/icmp.c +++ b/net/ipv4/icmp.c @@ -299,7 +299,7 @@ out: void icmp_out_count(struct net *net, unsigned char type) { ICMPMSGOUT_INC_STATS(type); - ICMP_INC_STATS(ICMP_MIB_OUTMSGS); + ICMP_INC_STATS(net, ICMP_MIB_OUTMSGS); } /* -- cgit v1.2.2 From dcfc23cac103b54dbc00a6f52f47656ad5c75844 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 14 Jul 2008 23:03:00 -0700 Subject: mib: add struct net to ICMP_INC_STATS_BH Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/dccp/ipv4.c | 4 ++-- net/ipv4/icmp.c | 10 +++++----- net/ipv4/tcp_ipv4.c | 4 ++-- net/ipv4/udp.c | 2 +- net/sctp/input.c | 5 +++-- 5 files changed, 13 insertions(+), 12 deletions(-) (limited to 'net') diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index cc4f33461f2d..2b03c47cab1b 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -208,7 +208,7 @@ static void dccp_v4_err(struct sk_buff *skb, u32 info) struct net *net = dev_net(skb->dev); if (skb->len < (iph->ihl << 2) + 8) { - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS); return; } @@ -216,7 +216,7 @@ static void dccp_v4_err(struct sk_buff *skb, u32 info) iph->daddr, dh->dccph_dport, iph->saddr, dh->dccph_sport, inet_iif(skb)); if (sk == NULL) { - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS); return; } diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c index e94de411cfa0..56a7bbc7950c 100644 --- a/net/ipv4/icmp.c +++ b/net/ipv4/icmp.c @@ -763,7 +763,7 @@ static void icmp_unreach(struct sk_buff *skb) out: return; out_err: - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS); goto out; } @@ -803,7 +803,7 @@ static void icmp_redirect(struct sk_buff *skb) out: return; out_err: - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + ICMP_INC_STATS_BH(dev_net(skb->dev), ICMP_MIB_INERRORS); goto out; } @@ -874,7 +874,7 @@ static void icmp_timestamp(struct sk_buff *skb) out: return; out_err: - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + ICMP_INC_STATS_BH(dev_net(skb->dst->dev), ICMP_MIB_INERRORS); goto out; } @@ -994,7 +994,7 @@ int icmp_rcv(struct sk_buff *skb) skb_set_network_header(skb, nh); } - ICMP_INC_STATS_BH(ICMP_MIB_INMSGS); + ICMP_INC_STATS_BH(net, ICMP_MIB_INMSGS); switch (skb->ip_summed) { case CHECKSUM_COMPLETE: @@ -1053,7 +1053,7 @@ drop: kfree_skb(skb); return 0; error: - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS); goto drop; } diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index ca41b77f3f32..0fefd4409413 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -346,14 +346,14 @@ void tcp_v4_err(struct sk_buff *skb, u32 info) struct net *net = dev_net(skb->dev); if (skb->len < (iph->ihl << 2) + 8) { - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS); return; } sk = inet_lookup(net, &tcp_hashinfo, iph->daddr, th->dest, iph->saddr, th->source, inet_iif(skb)); if (!sk) { - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS); return; } if (sk->sk_state == TCP_TIME_WAIT) { diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 9342cfda3d0a..fdd4faa1f120 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -359,7 +359,7 @@ void __udp4_lib_err(struct sk_buff *skb, u32 info, struct hlist_head udptable[]) sk = __udp4_lib_lookup(net, iph->daddr, uh->dest, iph->saddr, uh->source, skb->dev->ifindex, udptable); if (sk == NULL) { - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS); return; /* No socket for error */ } diff --git a/net/sctp/input.c b/net/sctp/input.c index d354a23972d4..ed8834e7f143 100644 --- a/net/sctp/input.c +++ b/net/sctp/input.c @@ -61,6 +61,7 @@ #include #include #include +#include /* Forward declarations for internal helpers. */ static int sctp_rcv_ootb(struct sk_buff *); @@ -534,7 +535,7 @@ void sctp_v4_err(struct sk_buff *skb, __u32 info) int err; if (skb->len < ihlen + 8) { - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + ICMP_INC_STATS_BH(&init_net, ICMP_MIB_INERRORS); return; } @@ -548,7 +549,7 @@ void sctp_v4_err(struct sk_buff *skb, __u32 info) skb->network_header = saveip; skb->transport_header = savesctp; if (!sk) { - ICMP_INC_STATS_BH(ICMP_MIB_INERRORS); + ICMP_INC_STATS_BH(&init_net, ICMP_MIB_INERRORS); return; } /* Warning: The sock lock is held. Remember to call -- cgit v1.2.2 From 903fc1964e746b8d8e2971ea20c89b7aeab8bd9a Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 14 Jul 2008 23:03:35 -0700 Subject: mib: add struct net to ICMPMSGOUT_INC_STATS Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/icmp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c index 56a7bbc7950c..6f4c65b898b4 100644 --- a/net/ipv4/icmp.c +++ b/net/ipv4/icmp.c @@ -298,7 +298,7 @@ out: */ void icmp_out_count(struct net *net, unsigned char type) { - ICMPMSGOUT_INC_STATS(type); + ICMPMSGOUT_INC_STATS(net, type); ICMP_INC_STATS(net, ICMP_MIB_OUTMSGS); } -- cgit v1.2.2 From f66ac03d497c162c70cd0ccc802ce1777073cdf3 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 14 Jul 2008 23:04:00 -0700 Subject: mib: add struct net to ICMPMSGIN_INC_STATS_BH Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/icmp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c index 6f4c65b898b4..ea60ad41008c 100644 --- a/net/ipv4/icmp.c +++ b/net/ipv4/icmp.c @@ -1012,7 +1012,7 @@ int icmp_rcv(struct sk_buff *skb) icmph = icmp_hdr(skb); - ICMPMSGIN_INC_STATS_BH(icmph->type); + ICMPMSGIN_INC_STATS_BH(net, icmph->type); /* * 18 is the highest 'known' ICMP type. Anything else is a mystery * -- cgit v1.2.2 From f1f28aa3510ddb84c966bac65611bb866c77a092 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 00:08:33 -0700 Subject: netdev: Add addr_list_lock to struct net_device. This will be used to protect the per-device unicast and multicast address lists, as well as the callbacks into the drivers which configure such state such as ->set_rx_mode() and ->set_multicast_list(). Signed-off-by: David S. Miller --- net/core/dev.c | 1 + 1 file changed, 1 insertion(+) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index feaab4898a5b..d933d1bfa6fa 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3836,6 +3836,7 @@ int register_netdevice(struct net_device *dev) BUG_ON(!dev_net(dev)); net = dev_net(dev); + spin_lock_init(&dev->addr_list_lock); netdev_init_queue_locks(dev); dev->iflink = -1; -- cgit v1.2.2 From e308a5d806c852f56590ffdd3834d0df0cbed8d7 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 00:13:44 -0700 Subject: netdev: Add netdev->addr_list_lock protection. Add netif_addr_{lock,unlock}{,_bh}() helpers. Use them to protect operations that operate on or read the network device unicast and multicast address lists. Also use them in cases where the code simply wants to block calls into the driver's ->set_rx_mode() and ->set_multicast_list() methods. Signed-off-by: David S. Miller --- net/core/dev.c | 14 ++++++++++++++ net/core/dev_mcast.c | 12 ++++++++++++ net/mac80211/main.c | 4 ++++ net/mac80211/mlme.c | 4 ++++ 4 files changed, 34 insertions(+) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index d933d1bfa6fa..ef1502d71f25 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2982,7 +2982,9 @@ void __dev_set_rx_mode(struct net_device *dev) void dev_set_rx_mode(struct net_device *dev) { netif_tx_lock_bh(dev); + netif_addr_lock(dev); __dev_set_rx_mode(dev); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); } @@ -3062,9 +3064,11 @@ int dev_unicast_delete(struct net_device *dev, void *addr, int alen) ASSERT_RTNL(); netif_tx_lock_bh(dev); + netif_addr_lock(dev); err = __dev_addr_delete(&dev->uc_list, &dev->uc_count, addr, alen, 0); if (!err) __dev_set_rx_mode(dev); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return err; } @@ -3088,9 +3092,11 @@ int dev_unicast_add(struct net_device *dev, void *addr, int alen) ASSERT_RTNL(); netif_tx_lock_bh(dev); + netif_addr_lock(dev); err = __dev_addr_add(&dev->uc_list, &dev->uc_count, addr, alen, 0); if (!err) __dev_set_rx_mode(dev); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return err; } @@ -3159,10 +3165,12 @@ int dev_unicast_sync(struct net_device *to, struct net_device *from) int err = 0; netif_tx_lock_bh(to); + netif_addr_lock(to); err = __dev_addr_sync(&to->uc_list, &to->uc_count, &from->uc_list, &from->uc_count); if (!err) __dev_set_rx_mode(to); + netif_addr_unlock(to); netif_tx_unlock_bh(to); return err; } @@ -3180,13 +3188,17 @@ EXPORT_SYMBOL(dev_unicast_sync); void dev_unicast_unsync(struct net_device *to, struct net_device *from) { netif_tx_lock_bh(from); + netif_addr_lock(from); netif_tx_lock_bh(to); + netif_addr_lock(to); __dev_addr_unsync(&to->uc_list, &to->uc_count, &from->uc_list, &from->uc_count); __dev_set_rx_mode(to); + netif_addr_unlock(to); netif_tx_unlock_bh(to); + netif_addr_unlock(from); netif_tx_unlock_bh(from); } EXPORT_SYMBOL(dev_unicast_unsync); @@ -3208,6 +3220,7 @@ static void __dev_addr_discard(struct dev_addr_list **list) static void dev_addr_discard(struct net_device *dev) { netif_tx_lock_bh(dev); + netif_addr_lock(dev); __dev_addr_discard(&dev->uc_list); dev->uc_count = 0; @@ -3215,6 +3228,7 @@ static void dev_addr_discard(struct net_device *dev) __dev_addr_discard(&dev->mc_list); dev->mc_count = 0; + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); } diff --git a/net/core/dev_mcast.c b/net/core/dev_mcast.c index f8a3455f4493..b6b2a129971a 100644 --- a/net/core/dev_mcast.c +++ b/net/core/dev_mcast.c @@ -73,6 +73,7 @@ int dev_mc_delete(struct net_device *dev, void *addr, int alen, int glbl) int err; netif_tx_lock_bh(dev); + netif_addr_lock(dev); err = __dev_addr_delete(&dev->mc_list, &dev->mc_count, addr, alen, glbl); if (!err) { @@ -83,6 +84,7 @@ int dev_mc_delete(struct net_device *dev, void *addr, int alen, int glbl) __dev_set_rx_mode(dev); } + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return err; } @@ -96,9 +98,11 @@ int dev_mc_add(struct net_device *dev, void *addr, int alen, int glbl) int err; netif_tx_lock_bh(dev); + netif_addr_lock(dev); err = __dev_addr_add(&dev->mc_list, &dev->mc_count, addr, alen, glbl); if (!err) __dev_set_rx_mode(dev); + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return err; } @@ -120,10 +124,12 @@ int dev_mc_sync(struct net_device *to, struct net_device *from) int err = 0; netif_tx_lock_bh(to); + netif_addr_lock(to); err = __dev_addr_sync(&to->mc_list, &to->mc_count, &from->mc_list, &from->mc_count); if (!err) __dev_set_rx_mode(to); + netif_addr_unlock(to); netif_tx_unlock_bh(to); return err; @@ -144,13 +150,17 @@ EXPORT_SYMBOL(dev_mc_sync); void dev_mc_unsync(struct net_device *to, struct net_device *from) { netif_tx_lock_bh(from); + netif_addr_lock(from); netif_tx_lock_bh(to); + netif_addr_lock(to); __dev_addr_unsync(&to->mc_list, &to->mc_count, &from->mc_list, &from->mc_count); __dev_set_rx_mode(to); + netif_addr_unlock(to); netif_tx_unlock_bh(to); + netif_addr_unlock(from); netif_tx_unlock_bh(from); } EXPORT_SYMBOL(dev_mc_unsync); @@ -165,6 +175,7 @@ static int dev_mc_seq_show(struct seq_file *seq, void *v) return 0; netif_tx_lock_bh(dev); + netif_addr_lock(dev); for (m = dev->mc_list; m; m = m->next) { int i; @@ -176,6 +187,7 @@ static int dev_mc_seq_show(struct seq_file *seq, void *v) seq_putc(seq, '\n'); } + netif_addr_unlock(dev); netif_tx_unlock_bh(dev); return 0; } diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 36859e794928..095b7d928d64 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -292,7 +292,9 @@ static int ieee80211_open(struct net_device *dev) local->fif_other_bss++; netif_tx_lock_bh(local->mdev); + netif_addr_lock(local->mdev); ieee80211_configure_filter(local); + netif_addr_unlock(local->mdev); netif_tx_unlock_bh(local->mdev); break; case IEEE80211_IF_TYPE_STA: @@ -491,7 +493,9 @@ static int ieee80211_stop(struct net_device *dev) local->fif_other_bss--; netif_tx_lock_bh(local->mdev); + netif_addr_lock(local->mdev); ieee80211_configure_filter(local); + netif_addr_unlock(local->mdev); netif_tx_unlock_bh(local->mdev); break; case IEEE80211_IF_TYPE_MESH_POINT: diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 8f51375317dd..1232ba25e1e9 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -3869,6 +3869,7 @@ void ieee80211_scan_completed(struct ieee80211_hw *hw) netif_tx_lock_bh(local->mdev); + netif_addr_lock(local->mdev); local->filter_flags &= ~FIF_BCN_PRBRESP_PROMISC; local->ops->configure_filter(local_to_hw(local), FIF_BCN_PRBRESP_PROMISC, @@ -3876,6 +3877,7 @@ void ieee80211_scan_completed(struct ieee80211_hw *hw) local->mdev->mc_count, local->mdev->mc_list); + netif_addr_unlock(local->mdev); netif_tx_unlock_bh(local->mdev); rcu_read_lock(); @@ -4063,12 +4065,14 @@ static int ieee80211_sta_start_scan(struct net_device *dev, local->scan_dev = dev; netif_tx_lock_bh(local->mdev); + netif_addr_lock(local->mdev); local->filter_flags |= FIF_BCN_PRBRESP_PROMISC; local->ops->configure_filter(local_to_hw(local), FIF_BCN_PRBRESP_PROMISC, &local->filter_flags, local->mdev->mc_count, local->mdev->mc_list); + netif_addr_unlock(local->mdev); netif_tx_unlock_bh(local->mdev); /* TODO: start scan as soon as all nullfunc frames are ACKed */ -- cgit v1.2.2 From b9e40857682ecfc5bcd0356a23ff409883ffb982 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 00:15:08 -0700 Subject: netdev: Do not use TX lock to protect address lists. Now that we have a specific lock to protect the network device unicast and multicast lists, remove extraneous grabs of the TX lock in cases where the code only needs address list protection. Signed-off-by: David S. Miller --- net/core/dev.c | 38 ++++++++++++-------------------------- net/core/dev_mcast.c | 32 ++++++++++---------------------- net/mac80211/main.c | 12 ++++-------- net/mac80211/mlme.c | 6 ++---- 4 files changed, 28 insertions(+), 60 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index ef1502d71f25..9b49f74a9820 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2981,11 +2981,9 @@ void __dev_set_rx_mode(struct net_device *dev) void dev_set_rx_mode(struct net_device *dev) { - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); __dev_set_rx_mode(dev); - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); } int __dev_addr_delete(struct dev_addr_list **list, int *count, @@ -3063,13 +3061,11 @@ int dev_unicast_delete(struct net_device *dev, void *addr, int alen) ASSERT_RTNL(); - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); err = __dev_addr_delete(&dev->uc_list, &dev->uc_count, addr, alen, 0); if (!err) __dev_set_rx_mode(dev); - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); return err; } EXPORT_SYMBOL(dev_unicast_delete); @@ -3091,13 +3087,11 @@ int dev_unicast_add(struct net_device *dev, void *addr, int alen) ASSERT_RTNL(); - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); err = __dev_addr_add(&dev->uc_list, &dev->uc_count, addr, alen, 0); if (!err) __dev_set_rx_mode(dev); - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); return err; } EXPORT_SYMBOL(dev_unicast_add); @@ -3164,14 +3158,12 @@ int dev_unicast_sync(struct net_device *to, struct net_device *from) { int err = 0; - netif_tx_lock_bh(to); - netif_addr_lock(to); + netif_addr_lock_bh(to); err = __dev_addr_sync(&to->uc_list, &to->uc_count, &from->uc_list, &from->uc_count); if (!err) __dev_set_rx_mode(to); - netif_addr_unlock(to); - netif_tx_unlock_bh(to); + netif_addr_unlock_bh(to); return err; } EXPORT_SYMBOL(dev_unicast_sync); @@ -3187,9 +3179,7 @@ EXPORT_SYMBOL(dev_unicast_sync); */ void dev_unicast_unsync(struct net_device *to, struct net_device *from) { - netif_tx_lock_bh(from); - netif_addr_lock(from); - netif_tx_lock_bh(to); + netif_addr_lock_bh(from); netif_addr_lock(to); __dev_addr_unsync(&to->uc_list, &to->uc_count, @@ -3197,9 +3187,7 @@ void dev_unicast_unsync(struct net_device *to, struct net_device *from) __dev_set_rx_mode(to); netif_addr_unlock(to); - netif_tx_unlock_bh(to); - netif_addr_unlock(from); - netif_tx_unlock_bh(from); + netif_addr_unlock_bh(from); } EXPORT_SYMBOL(dev_unicast_unsync); @@ -3219,8 +3207,7 @@ static void __dev_addr_discard(struct dev_addr_list **list) static void dev_addr_discard(struct net_device *dev) { - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); __dev_addr_discard(&dev->uc_list); dev->uc_count = 0; @@ -3228,8 +3215,7 @@ static void dev_addr_discard(struct net_device *dev) __dev_addr_discard(&dev->mc_list); dev->mc_count = 0; - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); } unsigned dev_get_flags(const struct net_device *dev) diff --git a/net/core/dev_mcast.c b/net/core/dev_mcast.c index b6b2a129971a..5402b3b38e0d 100644 --- a/net/core/dev_mcast.c +++ b/net/core/dev_mcast.c @@ -72,8 +72,7 @@ int dev_mc_delete(struct net_device *dev, void *addr, int alen, int glbl) { int err; - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); err = __dev_addr_delete(&dev->mc_list, &dev->mc_count, addr, alen, glbl); if (!err) { @@ -84,8 +83,7 @@ int dev_mc_delete(struct net_device *dev, void *addr, int alen, int glbl) __dev_set_rx_mode(dev); } - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); return err; } @@ -97,13 +95,11 @@ int dev_mc_add(struct net_device *dev, void *addr, int alen, int glbl) { int err; - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); err = __dev_addr_add(&dev->mc_list, &dev->mc_count, addr, alen, glbl); if (!err) __dev_set_rx_mode(dev); - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); return err; } @@ -123,14 +119,12 @@ int dev_mc_sync(struct net_device *to, struct net_device *from) { int err = 0; - netif_tx_lock_bh(to); - netif_addr_lock(to); + netif_addr_lock_bh(to); err = __dev_addr_sync(&to->mc_list, &to->mc_count, &from->mc_list, &from->mc_count); if (!err) __dev_set_rx_mode(to); - netif_addr_unlock(to); - netif_tx_unlock_bh(to); + netif_addr_unlock_bh(to); return err; } @@ -149,9 +143,7 @@ EXPORT_SYMBOL(dev_mc_sync); */ void dev_mc_unsync(struct net_device *to, struct net_device *from) { - netif_tx_lock_bh(from); - netif_addr_lock(from); - netif_tx_lock_bh(to); + netif_addr_lock_bh(from); netif_addr_lock(to); __dev_addr_unsync(&to->mc_list, &to->mc_count, @@ -159,9 +151,7 @@ void dev_mc_unsync(struct net_device *to, struct net_device *from) __dev_set_rx_mode(to); netif_addr_unlock(to); - netif_tx_unlock_bh(to); - netif_addr_unlock(from); - netif_tx_unlock_bh(from); + netif_addr_unlock_bh(from); } EXPORT_SYMBOL(dev_mc_unsync); @@ -174,8 +164,7 @@ static int dev_mc_seq_show(struct seq_file *seq, void *v) if (v == SEQ_START_TOKEN) return 0; - netif_tx_lock_bh(dev); - netif_addr_lock(dev); + netif_addr_lock_bh(dev); for (m = dev->mc_list; m; m = m->next) { int i; @@ -187,8 +176,7 @@ static int dev_mc_seq_show(struct seq_file *seq, void *v) seq_putc(seq, '\n'); } - netif_addr_unlock(dev); - netif_tx_unlock_bh(dev); + netif_addr_unlock_bh(dev); return 0; } diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 095b7d928d64..af0056e7e5b3 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -291,11 +291,9 @@ static int ieee80211_open(struct net_device *dev) if (sdata->u.mntr_flags & MONITOR_FLAG_OTHER_BSS) local->fif_other_bss++; - netif_tx_lock_bh(local->mdev); - netif_addr_lock(local->mdev); + netif_addr_lock_bh(local->mdev); ieee80211_configure_filter(local); - netif_addr_unlock(local->mdev); - netif_tx_unlock_bh(local->mdev); + netif_addr_unlock_bh(local->mdev); break; case IEEE80211_IF_TYPE_STA: case IEEE80211_IF_TYPE_IBSS: @@ -492,11 +490,9 @@ static int ieee80211_stop(struct net_device *dev) if (sdata->u.mntr_flags & MONITOR_FLAG_OTHER_BSS) local->fif_other_bss--; - netif_tx_lock_bh(local->mdev); - netif_addr_lock(local->mdev); + netif_addr_lock_bh(local->mdev); ieee80211_configure_filter(local); - netif_addr_unlock(local->mdev); - netif_tx_unlock_bh(local->mdev); + netif_addr_unlock_bh(local->mdev); break; case IEEE80211_IF_TYPE_MESH_POINT: case IEEE80211_IF_TYPE_STA: diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 1232ba25e1e9..d7c371e36bf0 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -4064,16 +4064,14 @@ static int ieee80211_sta_start_scan(struct net_device *dev, local->scan_band = IEEE80211_BAND_2GHZ; local->scan_dev = dev; - netif_tx_lock_bh(local->mdev); - netif_addr_lock(local->mdev); + netif_addr_lock_bh(local->mdev); local->filter_flags |= FIF_BCN_PRBRESP_PROMISC; local->ops->configure_filter(local_to_hw(local), FIF_BCN_PRBRESP_PROMISC, &local->filter_flags, local->mdev->mc_count, local->mdev->mc_list); - netif_addr_unlock(local->mdev); - netif_tx_unlock_bh(local->mdev); + netif_addr_unlock_bh(local->mdev); /* TODO: start scan as soon as all nullfunc frames are ACKed */ queue_delayed_work(local->hw.workqueue, &local->scan_work, -- cgit v1.2.2 From 2870c43d1795bcb40b12bad6456f07ad8e64b3de Mon Sep 17 00:00:00 2001 From: Octavian Purdila Date: Tue, 15 Jul 2008 00:49:11 -0700 Subject: net: refactor tcp splice receive path to improve readability - move all of the details on offsets, lengths and buffers into a single function instead of doing these operation from multiple places - use a bottom up approach: try to avoid details in the high level functions, introduce them gradually as we go deeper in the function call stack With helpful feedback from Jarek Poplawski. Signed-off-by: Octavian Purdila Acked-by: Jarek Poplawski Signed-off-by: David S. Miller --- net/core/skbuff.c | 153 ++++++++++++++++++++++-------------------------------- 1 file changed, 61 insertions(+), 92 deletions(-) (limited to 'net') diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 50a853f7cd8e..48a43727d12f 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -1282,114 +1282,83 @@ static inline int spd_fill_page(struct splice_pipe_desc *spd, struct page *page, return 0; } -/* - * Map linear and fragment data from the skb to spd. Returns number of - * pages mapped. - */ -static int __skb_splice_bits(struct sk_buff *skb, unsigned int *offset, - unsigned int *total_len, - struct splice_pipe_desc *spd) -{ - unsigned int nr_pages = spd->nr_pages; - unsigned int poff, plen, len, toff, tlen; - int headlen, seg, error = 0; - - toff = *offset; - tlen = *total_len; - if (!tlen) { - error = 1; - goto err; +static inline void __segment_seek(struct page **page, unsigned int *poff, + unsigned int *plen, unsigned int off) +{ + *poff += off; + *page += *poff / PAGE_SIZE; + *poff = *poff % PAGE_SIZE; + *plen -= off; +} + +static inline int __splice_segment(struct page *page, unsigned int poff, + unsigned int plen, unsigned int *off, + unsigned int *len, struct sk_buff *skb, + struct splice_pipe_desc *spd) +{ + if (!*len) + return 1; + + /* skip this segment if already processed */ + if (*off >= plen) { + *off -= plen; + return 0; } - /* - * if the offset is greater than the linear part, go directly to - * the fragments. - */ - headlen = skb_headlen(skb); - if (toff >= headlen) { - toff -= headlen; - goto map_frag; + /* ignore any bits we already processed */ + if (*off) { + __segment_seek(&page, &poff, &plen, *off); + *off = 0; } - /* - * first map the linear region into the pages/partial map, skipping - * any potential initial offset. - */ - len = 0; - while (len < headlen) { - void *p = skb->data + len; - - poff = (unsigned long) p & (PAGE_SIZE - 1); - plen = min_t(unsigned int, headlen - len, PAGE_SIZE - poff); - len += plen; - - if (toff) { - if (plen <= toff) { - toff -= plen; - continue; - } - plen -= toff; - poff += toff; - toff = 0; - } + do { + unsigned int flen = min(*len, plen); - plen = min(plen, tlen); - if (!plen) - break; + /* the linear region may spread across several pages */ + flen = min_t(unsigned int, flen, PAGE_SIZE - poff); - /* - * just jump directly to update and return, no point - * in going over fragments when the output is full. - */ - error = spd_fill_page(spd, virt_to_page(p), plen, poff, skb); - if (error) - goto done; + if (spd_fill_page(spd, page, flen, poff, skb)) + return 1; - tlen -= plen; - } + __segment_seek(&page, &poff, &plen, flen); + *len -= flen; + + } while (*len && plen); + + return 0; +} + +/* + * Map linear and fragment data from the skb to spd. It reports failure if the + * pipe is full or if we already spliced the requested length. + */ +int __skb_splice_bits(struct sk_buff *skb, unsigned int *offset, + unsigned int *len, + struct splice_pipe_desc *spd) +{ + int seg; + + /* + * map the linear part + */ + if (__splice_segment(virt_to_page(skb->data), + (unsigned long) skb->data & (PAGE_SIZE - 1), + skb_headlen(skb), + offset, len, skb, spd)) + return 1; /* * then map the fragments */ -map_frag: for (seg = 0; seg < skb_shinfo(skb)->nr_frags; seg++) { const skb_frag_t *f = &skb_shinfo(skb)->frags[seg]; - plen = f->size; - poff = f->page_offset; - - if (toff) { - if (plen <= toff) { - toff -= plen; - continue; - } - plen -= toff; - poff += toff; - toff = 0; - } - - plen = min(plen, tlen); - if (!plen) - break; - - error = spd_fill_page(spd, f->page, plen, poff, skb); - if (error) - break; - - tlen -= plen; + if (__splice_segment(f->page, f->page_offset, f->size, + offset, len, skb, spd)) + return 1; } -done: - if (spd->nr_pages - nr_pages) { - *offset = 0; - *total_len = tlen; - return 0; - } -err: - /* update the offset to reflect the linear part skip, if any */ - if (!error) - *offset = toff; - return error; + return 0; } /* -- cgit v1.2.2 From d0236f8f8223ad841f461ea3a635d452d194806b Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Tue, 15 Jul 2008 02:09:53 -0700 Subject: iucv: fix memory leak in cpu hotplug error path. Fix memory leak in error path in CPU_UP_PREPARE notifier. Signed-off-by: Akinobu Mita Signed-off-by: David S. Miller --- net/iucv/iucv.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/iucv/iucv.c b/net/iucv/iucv.c index 531a206ce7a6..d8e0635aace0 100644 --- a/net/iucv/iucv.c +++ b/net/iucv/iucv.c @@ -567,8 +567,11 @@ static int __cpuinit iucv_cpu_notify(struct notifier_block *self, return NOTIFY_BAD; iucv_param[cpu] = kmalloc_node(sizeof(union iucv_param), GFP_KERNEL|GFP_DMA, cpu_to_node(cpu)); - if (!iucv_param[cpu]) + if (!iucv_param[cpu]) { + kfree(iucv_irq_data[cpu]); + iucv_irq_data[cpu] = NULL; return NOTIFY_BAD; + } break; case CPU_UP_CANCELED: case CPU_UP_CANCELED_FROZEN: -- cgit v1.2.2 From d56400504a40a4aa197af629300d76544169e821 Mon Sep 17 00:00:00 2001 From: Sven Wegener Date: Wed, 16 Jul 2008 11:13:35 +0000 Subject: ipvs: Initialize mcast addr at compile time There's no need to do it at runtime, the values are constant. Signed-off-by: Sven Wegener Acked-by: Simon Horman --- net/ipv4/ipvs/ip_vs_sync.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/ipv4/ipvs/ip_vs_sync.c b/net/ipv4/ipvs/ip_vs_sync.c index 2d4a86f73325..8c900dfe832d 100644 --- a/net/ipv4/ipvs/ip_vs_sync.c +++ b/net/ipv4/ipvs/ip_vs_sync.c @@ -139,7 +139,11 @@ char ip_vs_master_mcast_ifn[IP_VS_IFNAME_MAXLEN]; char ip_vs_backup_mcast_ifn[IP_VS_IFNAME_MAXLEN]; /* multicast addr */ -static struct sockaddr_in mcast_addr; +static struct sockaddr_in mcast_addr = { + .sin_family = AF_INET, + .sin_port = __constant_htons(IP_VS_SYNC_PORT), + .sin_addr.s_addr = __constant_htonl(IP_VS_SYNC_GROUP), +}; static inline void sb_queue_tail(struct ip_vs_sync_buff *sb) @@ -862,11 +866,6 @@ static int sync_thread(void *startup) /* set the maximum length of sync message */ set_sync_mesg_maxlen(state); - /* set up multicast address */ - mcast_addr.sin_family = AF_INET; - mcast_addr.sin_port = htons(IP_VS_SYNC_PORT); - mcast_addr.sin_addr.s_addr = htonl(IP_VS_SYNC_GROUP); - add_wait_queue(&sync_wait, &wait); set_sync_pid(state, task_pid_nr(current)); -- cgit v1.2.2 From e6dd731c75cba986a485924f908e6e05b088ea9e Mon Sep 17 00:00:00 2001 From: Sven Wegener Date: Wed, 16 Jul 2008 11:13:43 +0000 Subject: ipvs: Use ERR_PTR for returning errors from make_receive_sock() and make_send_sock() The additional information we now return to the caller is currently not used, but will be used to return errors to user space. Signed-off-by: Sven Wegener Acked-by: Simon Horman --- net/ipv4/ipvs/ip_vs_sync.c | 46 +++++++++++++++++++++++++++------------------- 1 file changed, 27 insertions(+), 19 deletions(-) (limited to 'net') diff --git a/net/ipv4/ipvs/ip_vs_sync.c b/net/ipv4/ipvs/ip_vs_sync.c index 8c900dfe832d..60b96823c9ae 100644 --- a/net/ipv4/ipvs/ip_vs_sync.c +++ b/net/ipv4/ipvs/ip_vs_sync.c @@ -27,6 +27,7 @@ #include #include /* for ip_mc_join_group */ #include +#include #include #include @@ -576,14 +577,17 @@ static int bind_mcastif_addr(struct socket *sock, char *ifname) static struct socket * make_send_sock(void) { struct socket *sock; + int result; /* First create a socket */ - if (sock_create_kern(PF_INET, SOCK_DGRAM, IPPROTO_UDP, &sock) < 0) { + result = sock_create_kern(PF_INET, SOCK_DGRAM, IPPROTO_UDP, &sock); + if (result < 0) { IP_VS_ERR("Error during creation of socket; terminating\n"); - return NULL; + return ERR_PTR(result); } - if (set_mcast_if(sock->sk, ip_vs_master_mcast_ifn) < 0) { + result = set_mcast_if(sock->sk, ip_vs_master_mcast_ifn); + if (result < 0) { IP_VS_ERR("Error setting outbound mcast interface\n"); goto error; } @@ -591,14 +595,15 @@ static struct socket * make_send_sock(void) set_mcast_loop(sock->sk, 0); set_mcast_ttl(sock->sk, 1); - if (bind_mcastif_addr(sock, ip_vs_master_mcast_ifn) < 0) { + result = bind_mcastif_addr(sock, ip_vs_master_mcast_ifn); + if (result < 0) { IP_VS_ERR("Error binding address of the mcast interface\n"); goto error; } - if (sock->ops->connect(sock, - (struct sockaddr*)&mcast_addr, - sizeof(struct sockaddr), 0) < 0) { + result = sock->ops->connect(sock, (struct sockaddr *) &mcast_addr, + sizeof(struct sockaddr), 0); + if (result < 0) { IP_VS_ERR("Error connecting to the multicast addr\n"); goto error; } @@ -607,7 +612,7 @@ static struct socket * make_send_sock(void) error: sock_release(sock); - return NULL; + return ERR_PTR(result); } @@ -617,27 +622,30 @@ static struct socket * make_send_sock(void) static struct socket * make_receive_sock(void) { struct socket *sock; + int result; /* First create a socket */ - if (sock_create_kern(PF_INET, SOCK_DGRAM, IPPROTO_UDP, &sock) < 0) { + result = sock_create_kern(PF_INET, SOCK_DGRAM, IPPROTO_UDP, &sock); + if (result < 0) { IP_VS_ERR("Error during creation of socket; terminating\n"); - return NULL; + return ERR_PTR(result); } /* it is equivalent to the REUSEADDR option in user-space */ sock->sk->sk_reuse = 1; - if (sock->ops->bind(sock, - (struct sockaddr*)&mcast_addr, - sizeof(struct sockaddr)) < 0) { + result = sock->ops->bind(sock, (struct sockaddr *) &mcast_addr, + sizeof(struct sockaddr)); + if (result < 0) { IP_VS_ERR("Error binding to the multicast addr\n"); goto error; } /* join the multicast group */ - if (join_mcast_group(sock->sk, - (struct in_addr*)&mcast_addr.sin_addr, - ip_vs_backup_mcast_ifn) < 0) { + result = join_mcast_group(sock->sk, + (struct in_addr *) &mcast_addr.sin_addr, + ip_vs_backup_mcast_ifn); + if (result < 0) { IP_VS_ERR("Error joining to the multicast group\n"); goto error; } @@ -646,7 +654,7 @@ static struct socket * make_receive_sock(void) error: sock_release(sock); - return NULL; + return ERR_PTR(result); } @@ -719,7 +727,7 @@ static void sync_master_loop(void) /* create the sending multicast socket */ sock = make_send_sock(); - if (!sock) + if (IS_ERR(sock)) return; IP_VS_INFO("sync thread started: state = MASTER, mcast_ifn = %s, " @@ -772,7 +780,7 @@ static void sync_backup_loop(void) /* create the receiving multicast socket */ sock = make_receive_sock(); - if (!sock) + if (IS_ERR(sock)) goto out; IP_VS_INFO("sync thread started: state = BACKUP, mcast_ifn = %s, " -- cgit v1.2.2 From 998e7a76804b7a273a0460c2cdd5a51fa9856717 Mon Sep 17 00:00:00 2001 From: Sven Wegener Date: Wed, 16 Jul 2008 11:13:50 +0000 Subject: ipvs: Use kthread_run() instead of doing a double-fork via kernel_thread() This also moves the setup code out of the daemons, so that we're able to return proper error codes to user space. The current code will return success to user space when the daemon is started with an invald mcast interface. With these changes we get an appropriate "No such device" error. We longer need our own completion to be sure the daemons are actually running, because they no longer contain code that can fail and kthread_run() takes care of the rest. Signed-off-by: Sven Wegener Acked-by: Simon Horman --- net/ipv4/ipvs/ip_vs_sync.c | 369 +++++++++++++++++---------------------------- 1 file changed, 136 insertions(+), 233 deletions(-) (limited to 'net') diff --git a/net/ipv4/ipvs/ip_vs_sync.c b/net/ipv4/ipvs/ip_vs_sync.c index 60b96823c9ae..550563a56607 100644 --- a/net/ipv4/ipvs/ip_vs_sync.c +++ b/net/ipv4/ipvs/ip_vs_sync.c @@ -28,10 +28,10 @@ #include /* for ip_mc_join_group */ #include #include +#include #include #include -#include /* for get_fs and set_fs */ #include @@ -67,8 +67,8 @@ struct ip_vs_sync_conn_options { }; struct ip_vs_sync_thread_data { - struct completion *startup; - int state; + struct socket *sock; + char *buf; }; #define SIMPLE_CONN_SIZE (sizeof(struct ip_vs_sync_conn)) @@ -139,6 +139,10 @@ volatile int ip_vs_backup_syncid = 0; char ip_vs_master_mcast_ifn[IP_VS_IFNAME_MAXLEN]; char ip_vs_backup_mcast_ifn[IP_VS_IFNAME_MAXLEN]; +/* sync daemon tasks */ +static struct task_struct *sync_master_thread; +static struct task_struct *sync_backup_thread; + /* multicast addr */ static struct sockaddr_in mcast_addr = { .sin_family = AF_INET, @@ -147,14 +151,7 @@ static struct sockaddr_in mcast_addr = { }; -static inline void sb_queue_tail(struct ip_vs_sync_buff *sb) -{ - spin_lock(&ip_vs_sync_lock); - list_add_tail(&sb->list, &ip_vs_sync_queue); - spin_unlock(&ip_vs_sync_lock); -} - -static inline struct ip_vs_sync_buff * sb_dequeue(void) +static inline struct ip_vs_sync_buff *sb_dequeue(void) { struct ip_vs_sync_buff *sb; @@ -198,6 +195,16 @@ static inline void ip_vs_sync_buff_release(struct ip_vs_sync_buff *sb) kfree(sb); } +static inline void sb_queue_tail(struct ip_vs_sync_buff *sb) +{ + spin_lock(&ip_vs_sync_lock); + if (ip_vs_sync_state & IP_VS_STATE_MASTER) + list_add_tail(&sb->list, &ip_vs_sync_queue); + else + ip_vs_sync_buff_release(sb); + spin_unlock(&ip_vs_sync_lock); +} + /* * Get the current sync buffer if it has been created for more * than the specified time or the specified time is zero. @@ -712,43 +719,28 @@ ip_vs_receive(struct socket *sock, char *buffer, const size_t buflen) } -static DECLARE_WAIT_QUEUE_HEAD(sync_wait); -static pid_t sync_master_pid = 0; -static pid_t sync_backup_pid = 0; - -static DECLARE_WAIT_QUEUE_HEAD(stop_sync_wait); -static int stop_master_sync = 0; -static int stop_backup_sync = 0; - -static void sync_master_loop(void) +static int sync_thread_master(void *data) { - struct socket *sock; + struct ip_vs_sync_thread_data *tinfo = data; struct ip_vs_sync_buff *sb; - /* create the sending multicast socket */ - sock = make_send_sock(); - if (IS_ERR(sock)) - return; - IP_VS_INFO("sync thread started: state = MASTER, mcast_ifn = %s, " "syncid = %d\n", ip_vs_master_mcast_ifn, ip_vs_master_syncid); - for (;;) { - while ((sb=sb_dequeue())) { - ip_vs_send_sync_msg(sock, sb->mesg); + while (!kthread_should_stop()) { + while ((sb = sb_dequeue())) { + ip_vs_send_sync_msg(tinfo->sock, sb->mesg); ip_vs_sync_buff_release(sb); } /* check if entries stay in curr_sb for 2 seconds */ - if ((sb = get_curr_sync_buff(2*HZ))) { - ip_vs_send_sync_msg(sock, sb->mesg); + sb = get_curr_sync_buff(2 * HZ); + if (sb) { + ip_vs_send_sync_msg(tinfo->sock, sb->mesg); ip_vs_sync_buff_release(sb); } - if (stop_master_sync) - break; - msleep_interruptible(1000); } @@ -763,262 +755,173 @@ static void sync_master_loop(void) } /* release the sending multicast socket */ - sock_release(sock); + sock_release(tinfo->sock); + kfree(tinfo); + + return 0; } -static void sync_backup_loop(void) +static int sync_thread_backup(void *data) { - struct socket *sock; - char *buf; + struct ip_vs_sync_thread_data *tinfo = data; int len; - if (!(buf = kmalloc(sync_recv_mesg_maxlen, GFP_ATOMIC))) { - IP_VS_ERR("sync_backup_loop: kmalloc error\n"); - return; - } - - /* create the receiving multicast socket */ - sock = make_receive_sock(); - if (IS_ERR(sock)) - goto out; - IP_VS_INFO("sync thread started: state = BACKUP, mcast_ifn = %s, " "syncid = %d\n", ip_vs_backup_mcast_ifn, ip_vs_backup_syncid); - for (;;) { - /* do you have data now? */ - while (!skb_queue_empty(&(sock->sk->sk_receive_queue))) { - if ((len = - ip_vs_receive(sock, buf, - sync_recv_mesg_maxlen)) <= 0) { + while (!kthread_should_stop()) { + /* do we have data now? */ + while (!skb_queue_empty(&(tinfo->sock->sk->sk_receive_queue))) { + len = ip_vs_receive(tinfo->sock, tinfo->buf, + sync_recv_mesg_maxlen); + if (len <= 0) { IP_VS_ERR("receiving message error\n"); break; } - /* disable bottom half, because it accessed the data + + /* disable bottom half, because it accesses the data shared by softirq while getting/creating conns */ local_bh_disable(); - ip_vs_process_message(buf, len); + ip_vs_process_message(tinfo->buf, len); local_bh_enable(); } - if (stop_backup_sync) - break; - msleep_interruptible(1000); } /* release the sending multicast socket */ - sock_release(sock); + sock_release(tinfo->sock); + kfree(tinfo->buf); + kfree(tinfo); - out: - kfree(buf); + return 0; } -static void set_sync_pid(int sync_state, pid_t sync_pid) -{ - if (sync_state == IP_VS_STATE_MASTER) - sync_master_pid = sync_pid; - else if (sync_state == IP_VS_STATE_BACKUP) - sync_backup_pid = sync_pid; -} - -static void set_stop_sync(int sync_state, int set) +int start_sync_thread(int state, char *mcast_ifn, __u8 syncid) { - if (sync_state == IP_VS_STATE_MASTER) - stop_master_sync = set; - else if (sync_state == IP_VS_STATE_BACKUP) - stop_backup_sync = set; - else { - stop_master_sync = set; - stop_backup_sync = set; - } -} + struct ip_vs_sync_thread_data *tinfo; + struct task_struct **realtask, *task; + struct socket *sock; + char *name, *buf = NULL; + int (*threadfn)(void *data); + int result = -ENOMEM; -static int sync_thread(void *startup) -{ - DECLARE_WAITQUEUE(wait, current); - mm_segment_t oldmm; - int state; - const char *name; - struct ip_vs_sync_thread_data *tinfo = startup; + IP_VS_DBG(7, "%s: pid %d\n", __func__, task_pid_nr(current)); + IP_VS_DBG(7, "Each ip_vs_sync_conn entry needs %Zd bytes\n", + sizeof(struct ip_vs_sync_conn)); - /* increase the module use count */ - ip_vs_use_count_inc(); + if (state == IP_VS_STATE_MASTER) { + if (sync_master_thread) + return -EEXIST; - if (ip_vs_sync_state & IP_VS_STATE_MASTER && !sync_master_pid) { - state = IP_VS_STATE_MASTER; + strlcpy(ip_vs_master_mcast_ifn, mcast_ifn, + sizeof(ip_vs_master_mcast_ifn)); + ip_vs_master_syncid = syncid; + realtask = &sync_master_thread; name = "ipvs_syncmaster"; - } else if (ip_vs_sync_state & IP_VS_STATE_BACKUP && !sync_backup_pid) { - state = IP_VS_STATE_BACKUP; + threadfn = sync_thread_master; + sock = make_send_sock(); + } else if (state == IP_VS_STATE_BACKUP) { + if (sync_backup_thread) + return -EEXIST; + + strlcpy(ip_vs_backup_mcast_ifn, mcast_ifn, + sizeof(ip_vs_backup_mcast_ifn)); + ip_vs_backup_syncid = syncid; + realtask = &sync_backup_thread; name = "ipvs_syncbackup"; + threadfn = sync_thread_backup; + sock = make_receive_sock(); } else { - IP_VS_BUG(); - ip_vs_use_count_dec(); return -EINVAL; } - daemonize(name); - - oldmm = get_fs(); - set_fs(KERNEL_DS); - - /* Block all signals */ - spin_lock_irq(¤t->sighand->siglock); - siginitsetinv(¤t->blocked, 0); - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); + if (IS_ERR(sock)) { + result = PTR_ERR(sock); + goto out; + } - /* set the maximum length of sync message */ set_sync_mesg_maxlen(state); + if (state == IP_VS_STATE_BACKUP) { + buf = kmalloc(sync_recv_mesg_maxlen, GFP_KERNEL); + if (!buf) + goto outsocket; + } - add_wait_queue(&sync_wait, &wait); - - set_sync_pid(state, task_pid_nr(current)); - complete(tinfo->startup); - - /* - * once we call the completion queue above, we should - * null out that reference, since its allocated on the - * stack of the creating kernel thread - */ - tinfo->startup = NULL; - - /* processing master/backup loop here */ - if (state == IP_VS_STATE_MASTER) - sync_master_loop(); - else if (state == IP_VS_STATE_BACKUP) - sync_backup_loop(); - else IP_VS_BUG(); - - remove_wait_queue(&sync_wait, &wait); - - /* thread exits */ - - /* - * If we weren't explicitly stopped, then we - * exited in error, and should undo our state - */ - if ((!stop_master_sync) && (!stop_backup_sync)) - ip_vs_sync_state -= tinfo->state; + tinfo = kmalloc(sizeof(*tinfo), GFP_KERNEL); + if (!tinfo) + goto outbuf; - set_sync_pid(state, 0); - IP_VS_INFO("sync thread stopped!\n"); + tinfo->sock = sock; + tinfo->buf = buf; - set_fs(oldmm); + task = kthread_run(threadfn, tinfo, name); + if (IS_ERR(task)) { + result = PTR_ERR(task); + goto outtinfo; + } - /* decrease the module use count */ - ip_vs_use_count_dec(); + /* mark as active */ + *realtask = task; + ip_vs_sync_state |= state; - set_stop_sync(state, 0); - wake_up(&stop_sync_wait); + /* increase the module use count */ + ip_vs_use_count_inc(); - /* - * we need to free the structure that was allocated - * for us in start_sync_thread - */ - kfree(tinfo); return 0; -} - - -static int fork_sync_thread(void *startup) -{ - pid_t pid; - - /* fork the sync thread here, then the parent process of the - sync thread is the init process after this thread exits. */ - repeat: - if ((pid = kernel_thread(sync_thread, startup, 0)) < 0) { - IP_VS_ERR("could not create sync_thread due to %d... " - "retrying.\n", pid); - msleep_interruptible(1000); - goto repeat; - } - return 0; +outtinfo: + kfree(tinfo); +outbuf: + kfree(buf); +outsocket: + sock_release(sock); +out: + return result; } -int start_sync_thread(int state, char *mcast_ifn, __u8 syncid) +int stop_sync_thread(int state) { - DECLARE_COMPLETION_ONSTACK(startup); - pid_t pid; - struct ip_vs_sync_thread_data *tinfo; - - if ((state == IP_VS_STATE_MASTER && sync_master_pid) || - (state == IP_VS_STATE_BACKUP && sync_backup_pid)) - return -EEXIST; - - /* - * Note that tinfo will be freed in sync_thread on exit - */ - tinfo = kmalloc(sizeof(struct ip_vs_sync_thread_data), GFP_KERNEL); - if (!tinfo) - return -ENOMEM; - IP_VS_DBG(7, "%s: pid %d\n", __func__, task_pid_nr(current)); - IP_VS_DBG(7, "Each ip_vs_sync_conn entry need %Zd bytes\n", - sizeof(struct ip_vs_sync_conn)); - ip_vs_sync_state |= state; if (state == IP_VS_STATE_MASTER) { - strlcpy(ip_vs_master_mcast_ifn, mcast_ifn, - sizeof(ip_vs_master_mcast_ifn)); - ip_vs_master_syncid = syncid; - } else { - strlcpy(ip_vs_backup_mcast_ifn, mcast_ifn, - sizeof(ip_vs_backup_mcast_ifn)); - ip_vs_backup_syncid = syncid; - } - - tinfo->state = state; - tinfo->startup = &startup; + if (!sync_master_thread) + return -ESRCH; - repeat: - if ((pid = kernel_thread(fork_sync_thread, tinfo, 0)) < 0) { - IP_VS_ERR("could not create fork_sync_thread due to %d... " - "retrying.\n", pid); - msleep_interruptible(1000); - goto repeat; - } + IP_VS_INFO("stopping master sync thread %d ...\n", + task_pid_nr(sync_master_thread)); - wait_for_completion(&startup); - - return 0; -} - - -int stop_sync_thread(int state) -{ - DECLARE_WAITQUEUE(wait, current); + /* + * The lock synchronizes with sb_queue_tail(), so that we don't + * add sync buffers to the queue, when we are already in + * progress of stopping the master sync daemon. + */ - if ((state == IP_VS_STATE_MASTER && !sync_master_pid) || - (state == IP_VS_STATE_BACKUP && !sync_backup_pid)) - return -ESRCH; + spin_lock(&ip_vs_sync_lock); + ip_vs_sync_state &= ~IP_VS_STATE_MASTER; + spin_unlock(&ip_vs_sync_lock); + kthread_stop(sync_master_thread); + sync_master_thread = NULL; + } else if (state == IP_VS_STATE_BACKUP) { + if (!sync_backup_thread) + return -ESRCH; + + IP_VS_INFO("stopping backup sync thread %d ...\n", + task_pid_nr(sync_backup_thread)); + + ip_vs_sync_state &= ~IP_VS_STATE_BACKUP; + kthread_stop(sync_backup_thread); + sync_backup_thread = NULL; + } else { + return -EINVAL; + } - IP_VS_DBG(7, "%s: pid %d\n", __func__, task_pid_nr(current)); - IP_VS_INFO("stopping sync thread %d ...\n", - (state == IP_VS_STATE_MASTER) ? - sync_master_pid : sync_backup_pid); - - __set_current_state(TASK_UNINTERRUPTIBLE); - add_wait_queue(&stop_sync_wait, &wait); - set_stop_sync(state, 1); - ip_vs_sync_state -= state; - wake_up(&sync_wait); - schedule(); - __set_current_state(TASK_RUNNING); - remove_wait_queue(&stop_sync_wait, &wait); - - /* Note: no need to reap the sync thread, because its parent - process is the init process */ - - if ((state == IP_VS_STATE_MASTER && stop_master_sync) || - (state == IP_VS_STATE_BACKUP && stop_backup_sync)) - IP_VS_BUG(); + /* decrease the module use count */ + ip_vs_use_count_dec(); return 0; } -- cgit v1.2.2 From ba6fd85021dec97d58373d9aea4bea8fc24258be Mon Sep 17 00:00:00 2001 From: Sven Wegener Date: Wed, 16 Jul 2008 11:13:56 +0000 Subject: ipvs: Put backup thread on mcast socket wait queue Instead of doing an endless loop with sleeping for one second, we now put the backup thread onto the mcast socket wait queue and it gets woken up as soon as we have data to process. Signed-off-by: Sven Wegener Acked-by: Simon Horman --- net/ipv4/ipvs/ip_vs_sync.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/ipv4/ipvs/ip_vs_sync.c b/net/ipv4/ipvs/ip_vs_sync.c index 550563a56607..cf5ed758ea1f 100644 --- a/net/ipv4/ipvs/ip_vs_sync.c +++ b/net/ipv4/ipvs/ip_vs_sync.c @@ -29,6 +29,7 @@ #include #include #include +#include #include #include @@ -772,6 +773,10 @@ static int sync_thread_backup(void *data) ip_vs_backup_mcast_ifn, ip_vs_backup_syncid); while (!kthread_should_stop()) { + wait_event_interruptible(*tinfo->sock->sk->sk_sleep, + !skb_queue_empty(&tinfo->sock->sk->sk_receive_queue) + || kthread_should_stop()); + /* do we have data now? */ while (!skb_queue_empty(&(tinfo->sock->sk->sk_receive_queue))) { len = ip_vs_receive(tinfo->sock, tinfo->buf, @@ -787,8 +792,6 @@ static int sync_thread_backup(void *data) ip_vs_process_message(tinfo->buf, len); local_bh_enable(); } - - msleep_interruptible(1000); } /* release the sending multicast socket */ -- cgit v1.2.2 From 375c6bbabf210ded3151481ca1ac7f730903512b Mon Sep 17 00:00:00 2001 From: Sven Wegener Date: Wed, 16 Jul 2008 11:14:03 +0000 Subject: ipvs: Use schedule_timeout_interruptible() instead of msleep_interruptible() So that kthread_stop() can wake up the thread and we don't have to wait one second in the worst case for the daemon to actually stop. Signed-off-by: Sven Wegener Acked-by: Simon Horman --- net/ipv4/ipvs/ip_vs_sync.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/ipvs/ip_vs_sync.c b/net/ipv4/ipvs/ip_vs_sync.c index cf5ed758ea1f..45e9bd96c286 100644 --- a/net/ipv4/ipvs/ip_vs_sync.c +++ b/net/ipv4/ipvs/ip_vs_sync.c @@ -742,7 +742,7 @@ static int sync_thread_master(void *data) ip_vs_sync_buff_release(sb); } - msleep_interruptible(1000); + schedule_timeout_interruptible(HZ); } /* clean up the sync_buff queue */ -- cgit v1.2.2 From 9d3a0de7dc3c1c4456db5ceb92c445bef0a47681 Mon Sep 17 00:00:00 2001 From: "Rumen G. Bogdanovski" Date: Wed, 16 Jul 2008 20:04:23 -0700 Subject: ipvs: More reliable synchronization on connection close This patch enhances the synchronization of the closing connections between the master and the backup director. It prevents the closed connections to expire with the 15 min timeout of the ESTABLISHED state on the backup and makes them expire as they would do on the master with much shorter timeouts. Signed-off-by: Rumen G. Bogdanovski Acked-by: Simon Horman Signed-off-by: David S. Miller --- net/ipv4/ipvs/ip_vs_core.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/ipvs/ip_vs_core.c b/net/ipv4/ipvs/ip_vs_core.c index bcf6276ba4b2..a7879eafc3b5 100644 --- a/net/ipv4/ipvs/ip_vs_core.c +++ b/net/ipv4/ipvs/ip_vs_core.c @@ -991,7 +991,8 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb, == sysctl_ip_vs_sync_threshold[0])) || ((cp->protocol == IPPROTO_TCP) && (cp->old_state != cp->state) && ((cp->state == IP_VS_TCP_S_FIN_WAIT) || - (cp->state == IP_VS_TCP_S_CLOSE))))) + (cp->state == IP_VS_TCP_S_CLOSE_WAIT) || + (cp->state == IP_VS_TCP_S_TIME_WAIT))))) ip_vs_sync_conn(cp); cp->old_state = cp->state; -- cgit v1.2.2 From 7b1c65faa27f5ade3915e4bbc9186b6e64d2d6ec Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 16 Jul 2008 20:12:30 -0700 Subject: net: make __skb_splice_bits static net/core/skbuff.c:1335:5: warning: symbol '__skb_splice_bits' was not declared. Should it be static? Signed-off-by: Harvey Harrison Signed-off-by: David S. Miller --- net/core/skbuff.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 48a43727d12f..e4115672b6cf 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -1332,7 +1332,7 @@ static inline int __splice_segment(struct page *page, unsigned int poff, * Map linear and fragment data from the skb to spd. It reports failure if the * pipe is full or if we already spliced the requested length. */ -int __skb_splice_bits(struct sk_buff *skb, unsigned int *offset, +static int __skb_splice_bits(struct sk_buff *skb, unsigned int *offset, unsigned int *len, struct splice_pipe_desc *spd) { -- cgit v1.2.2 From 70efce27fc3d54271519244dc5e47da4ed711dd4 Mon Sep 17 00:00:00 2001 From: Will Newton Date: Wed, 16 Jul 2008 20:13:43 -0700 Subject: net/ipv4/tcp.c: Fix use of PULLHUP instead of POLLHUP in comments. Change PULLHUP to POLLHUP in tcp_poll comments and clean up another comment for grammar and coding style. Signed-off-by: Will Newton Signed-off-by: David S. Miller --- net/ipv4/tcp.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 56a133c61452..631133ebb20a 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -344,8 +344,8 @@ unsigned int tcp_poll(struct file *file, struct socket *sock, poll_table *wait) return inet_csk_listen_poll(sk); /* Socket is not locked. We are protected from async events - by poll logic and correct handling of state changes - made by another threads is impossible in any case. + * by poll logic and correct handling of state changes + * made by other threads is impossible in any case. */ mask = 0; @@ -371,10 +371,10 @@ unsigned int tcp_poll(struct file *file, struct socket *sock, poll_table *wait) * in state CLOSE_WAIT. One solution is evident --- to set POLLHUP * if and only if shutdown has been made in both directions. * Actually, it is interesting to look how Solaris and DUX - * solve this dilemma. I would prefer, if PULLHUP were maskable, + * solve this dilemma. I would prefer, if POLLHUP were maskable, * then we could set it on SND_SHUTDOWN. BTW examples given * in Stevens' books assume exactly this behaviour, it explains - * why PULLHUP is incompatible with POLLOUT. --ANK + * why POLLHUP is incompatible with POLLOUT. --ANK * * NOTE. Check for TCP_CLOSE is added. The goal is to prevent * blocking on fresh not-connected or disconnected socket. --ANK -- cgit v1.2.2 From 84a3aa000eacbaf841d745b07ef3a3280899056b Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:19:08 -0700 Subject: ipv4: prepare net initialization for IP accounting Some places, that deal with IP statistics already have where to get a struct net from, but use it directly, without declaring a separate variable on the stack. So, save this net on the stack for future IP_XXX_STATS macros. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/inet_connection_sock.c | 3 ++- net/ipv4/ip_fragment.c | 6 +++--- net/ipv4/udp.c | 4 +++- 3 files changed, 8 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c index 5bbf00051512..8338e1066654 100644 --- a/net/ipv4/inet_connection_sock.c +++ b/net/ipv4/inet_connection_sock.c @@ -338,9 +338,10 @@ struct dst_entry* inet_csk_route_req(struct sock *sk, .uli_u = { .ports = { .sport = inet_sk(sk)->sport, .dport = ireq->rmt_port } } }; + struct net *net = sock_net(sk); security_req_classify_flow(req, &fl); - if (ip_route_output_flow(sock_net(sk), &rt, &fl, sk, 0)) { + if (ip_route_output_flow(net, &rt, &fl, sk, 0)) { IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES); return NULL; } diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c index fbd5804b5d83..e23be5b36e1d 100644 --- a/net/ipv4/ip_fragment.c +++ b/net/ipv4/ip_fragment.c @@ -187,8 +187,10 @@ static void ip_evictor(struct net *net) static void ip_expire(unsigned long arg) { struct ipq *qp; + struct net *net; qp = container_of((struct inet_frag_queue *) arg, struct ipq, q); + net = container_of(qp->q.net, struct net, ipv4.frags); spin_lock(&qp->q.lock); @@ -202,9 +204,7 @@ static void ip_expire(unsigned long arg) if ((qp->q.last_in & INET_FRAG_FIRST_IN) && qp->q.fragments != NULL) { struct sk_buff *head = qp->q.fragments; - struct net *net; - net = container_of(qp->q.net, struct net, ipv4.frags); /* Send an ICMP "Fragment Reassembly Timeout" message. */ if ((head->dev = dev_get_by_index(net, qp->iif)) != NULL) { icmp_send(head, ICMP_TIME_EXCEEDED, ICMP_EXC_FRAGTIME, 0); @@ -570,9 +570,9 @@ int ip_defrag(struct sk_buff *skb, u32 user) struct ipq *qp; struct net *net; + net = skb->dev ? dev_net(skb->dev) : dev_net(skb->dst->dev); IP_INC_STATS_BH(IPSTATS_MIB_REASMREQDS); - net = skb->dev ? dev_net(skb->dev) : dev_net(skb->dst->dev); /* Start by cleaning up the memory. */ if (atomic_read(&net->ipv4.frags.mem) > net->ipv4.frags.high_thresh) ip_evictor(net); diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index fdd4faa1f120..048ef57edc1d 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -656,8 +656,10 @@ int udp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg, .uli_u = { .ports = { .sport = inet->sport, .dport = dport } } }; + struct net *net = sock_net(sk); + security_sk_classify_flow(sk, &fl); - err = ip_route_output_flow(sock_net(sk), &rt, &fl, sk, 1); + err = ip_route_output_flow(net, &rt, &fl, sk, 1); if (err) { if (err == -ENETUNREACH) IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES); -- cgit v1.2.2 From 5e38e270444f2629de7a706b5a9ca1b333d14517 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:19:49 -0700 Subject: mib: add net to IP_INC_STATS All the callers already have either the net itself, or the place where to get it from. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/ip_forward.c | 2 +- net/ipv4/ip_output.c | 30 +++++++++++++++--------------- net/ipv4/raw.c | 2 +- 3 files changed, 17 insertions(+), 17 deletions(-) (limited to 'net') diff --git a/net/ipv4/ip_forward.c b/net/ipv4/ip_forward.c index da14725916d3..7f78a5a7e1e7 100644 --- a/net/ipv4/ip_forward.c +++ b/net/ipv4/ip_forward.c @@ -88,7 +88,7 @@ int ip_forward(struct sk_buff *skb) if (unlikely(skb->len > dst_mtu(&rt->u.dst) && !skb_is_gso(skb) && (ip_hdr(skb)->frag_off & htons(IP_DF))) && !skb->local_df) { - IP_INC_STATS(IPSTATS_MIB_FRAGFAILS); + IP_INC_STATS(dev_net(rt->u.dst.dev), IPSTATS_MIB_FRAGFAILS); icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED, htonl(dst_mtu(&rt->u.dst))); goto drop; diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c index f003186b93b1..465544f6281a 100644 --- a/net/ipv4/ip_output.c +++ b/net/ipv4/ip_output.c @@ -182,9 +182,9 @@ static inline int ip_finish_output2(struct sk_buff *skb) unsigned int hh_len = LL_RESERVED_SPACE(dev); if (rt->rt_type == RTN_MULTICAST) - IP_INC_STATS(IPSTATS_MIB_OUTMCASTPKTS); + IP_INC_STATS(dev_net(dev), IPSTATS_MIB_OUTMCASTPKTS); else if (rt->rt_type == RTN_BROADCAST) - IP_INC_STATS(IPSTATS_MIB_OUTBCASTPKTS); + IP_INC_STATS(dev_net(dev), IPSTATS_MIB_OUTBCASTPKTS); /* Be paranoid, rather than too clever. */ if (unlikely(skb_headroom(skb) < hh_len && dev->header_ops)) { @@ -244,7 +244,7 @@ int ip_mc_output(struct sk_buff *skb) /* * If the indicated interface is up and running, send the packet. */ - IP_INC_STATS(IPSTATS_MIB_OUTREQUESTS); + IP_INC_STATS(dev_net(dev), IPSTATS_MIB_OUTREQUESTS); skb->dev = dev; skb->protocol = htons(ETH_P_IP); @@ -298,7 +298,7 @@ int ip_output(struct sk_buff *skb) { struct net_device *dev = skb->dst->dev; - IP_INC_STATS(IPSTATS_MIB_OUTREQUESTS); + IP_INC_STATS(dev_net(dev), IPSTATS_MIB_OUTREQUESTS); skb->dev = dev; skb->protocol = htons(ETH_P_IP); @@ -389,7 +389,7 @@ packet_routed: return ip_local_out(skb); no_route: - IP_INC_STATS(IPSTATS_MIB_OUTNOROUTES); + IP_INC_STATS(sock_net(sk), IPSTATS_MIB_OUTNOROUTES); kfree_skb(skb); return -EHOSTUNREACH; } @@ -451,7 +451,7 @@ int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff*)) iph = ip_hdr(skb); if (unlikely((iph->frag_off & htons(IP_DF)) && !skb->local_df)) { - IP_INC_STATS(IPSTATS_MIB_FRAGFAILS); + IP_INC_STATS(dev_net(dev), IPSTATS_MIB_FRAGFAILS); icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED, htonl(ip_skb_dst_mtu(skb))); kfree_skb(skb); @@ -542,7 +542,7 @@ int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff*)) err = output(skb); if (!err) - IP_INC_STATS(IPSTATS_MIB_FRAGCREATES); + IP_INC_STATS(dev_net(dev), IPSTATS_MIB_FRAGCREATES); if (err || !frag) break; @@ -552,7 +552,7 @@ int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff*)) } if (err == 0) { - IP_INC_STATS(IPSTATS_MIB_FRAGOKS); + IP_INC_STATS(dev_net(dev), IPSTATS_MIB_FRAGOKS); return 0; } @@ -561,7 +561,7 @@ int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff*)) kfree_skb(frag); frag = skb; } - IP_INC_STATS(IPSTATS_MIB_FRAGFAILS); + IP_INC_STATS(dev_net(dev), IPSTATS_MIB_FRAGFAILS); return err; } @@ -673,15 +673,15 @@ slow_path: if (err) goto fail; - IP_INC_STATS(IPSTATS_MIB_FRAGCREATES); + IP_INC_STATS(dev_net(dev), IPSTATS_MIB_FRAGCREATES); } kfree_skb(skb); - IP_INC_STATS(IPSTATS_MIB_FRAGOKS); + IP_INC_STATS(dev_net(dev), IPSTATS_MIB_FRAGOKS); return err; fail: kfree_skb(skb); - IP_INC_STATS(IPSTATS_MIB_FRAGFAILS); + IP_INC_STATS(dev_net(dev), IPSTATS_MIB_FRAGFAILS); return err; } @@ -1047,7 +1047,7 @@ alloc_new_skb: error: inet->cork.length -= length; - IP_INC_STATS(IPSTATS_MIB_OUTDISCARDS); + IP_INC_STATS(sock_net(sk), IPSTATS_MIB_OUTDISCARDS); return err; } @@ -1189,7 +1189,7 @@ ssize_t ip_append_page(struct sock *sk, struct page *page, error: inet->cork.length -= size; - IP_INC_STATS(IPSTATS_MIB_OUTDISCARDS); + IP_INC_STATS(sock_net(sk), IPSTATS_MIB_OUTDISCARDS); return err; } @@ -1298,7 +1298,7 @@ out: return err; error: - IP_INC_STATS(IPSTATS_MIB_OUTDISCARDS); + IP_INC_STATS(net, IPSTATS_MIB_OUTDISCARDS); goto out; } diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c index 7f39ea443ec8..cd975743bcd2 100644 --- a/net/ipv4/raw.c +++ b/net/ipv4/raw.c @@ -385,7 +385,7 @@ error_fault: err = -EFAULT; kfree_skb(skb); error: - IP_INC_STATS(IPSTATS_MIB_OUTDISCARDS); + IP_INC_STATS(net, IPSTATS_MIB_OUTDISCARDS); return err; } -- cgit v1.2.2 From 7c73a6faffae0bfae70639113aecf06af666e714 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:20:11 -0700 Subject: mib: add net to IP_INC_STATS_BH Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/dccp/ipv4.c | 2 +- net/ipv4/datagram.c | 2 +- net/ipv4/inet_connection_sock.c | 4 ++-- net/ipv4/ip_forward.c | 4 ++-- net/ipv4/ip_fragment.c | 17 ++++++++++------- net/ipv4/ip_input.c | 30 ++++++++++++++++-------------- net/ipv4/ipmr.c | 4 ++-- net/ipv4/route.c | 3 ++- net/ipv4/tcp_ipv4.c | 2 +- net/ipv4/udp.c | 2 +- net/sctp/output.c | 3 ++- 11 files changed, 40 insertions(+), 33 deletions(-) (limited to 'net') diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index 2b03c47cab1b..9f760a1e312c 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -465,7 +465,7 @@ static struct dst_entry* dccp_v4_route_skb(struct net *net, struct sock *sk, security_skb_classify_flow(skb, &fl); if (ip_route_output_flow(net, &rt, &fl, sk, 0)) { - IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES); + IP_INC_STATS_BH(net, IPSTATS_MIB_OUTNOROUTES); return NULL; } diff --git a/net/ipv4/datagram.c b/net/ipv4/datagram.c index 0c0c73f368ce..5e6c5a0f3fde 100644 --- a/net/ipv4/datagram.c +++ b/net/ipv4/datagram.c @@ -52,7 +52,7 @@ int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) inet->sport, usin->sin_port, sk, 1); if (err) { if (err == -ENETUNREACH) - IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES); + IP_INC_STATS_BH(sock_net(sk), IPSTATS_MIB_OUTNOROUTES); return err; } diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c index 8338e1066654..bb81c958b744 100644 --- a/net/ipv4/inet_connection_sock.c +++ b/net/ipv4/inet_connection_sock.c @@ -342,12 +342,12 @@ struct dst_entry* inet_csk_route_req(struct sock *sk, security_req_classify_flow(req, &fl); if (ip_route_output_flow(net, &rt, &fl, sk, 0)) { - IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES); + IP_INC_STATS_BH(net, IPSTATS_MIB_OUTNOROUTES); return NULL; } if (opt && opt->is_strictroute && rt->rt_dst != rt->rt_gateway) { ip_rt_put(rt); - IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES); + IP_INC_STATS_BH(net, IPSTATS_MIB_OUTNOROUTES); return NULL; } return &rt->u.dst; diff --git a/net/ipv4/ip_forward.c b/net/ipv4/ip_forward.c index 7f78a5a7e1e7..450016b89a18 100644 --- a/net/ipv4/ip_forward.c +++ b/net/ipv4/ip_forward.c @@ -42,7 +42,7 @@ static int ip_forward_finish(struct sk_buff *skb) { struct ip_options * opt = &(IPCB(skb)->opt); - IP_INC_STATS_BH(IPSTATS_MIB_OUTFORWDATAGRAMS); + IP_INC_STATS_BH(dev_net(skb->dst->dev), IPSTATS_MIB_OUTFORWDATAGRAMS); if (unlikely(opt->optlen)) ip_forward_options(skb); @@ -123,7 +123,7 @@ sr_failed: too_many_hops: /* Tell the sender its packet died... */ - IP_INC_STATS_BH(IPSTATS_MIB_INHDRERRORS); + IP_INC_STATS_BH(dev_net(skb->dst->dev), IPSTATS_MIB_INHDRERRORS); icmp_send(skb, ICMP_TIME_EXCEEDED, ICMP_EXC_TTL, 0); drop: kfree_skb(skb); diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c index e23be5b36e1d..65364c06ada5 100644 --- a/net/ipv4/ip_fragment.c +++ b/net/ipv4/ip_fragment.c @@ -199,8 +199,8 @@ static void ip_expire(unsigned long arg) ipq_kill(qp); - IP_INC_STATS_BH(IPSTATS_MIB_REASMTIMEOUT); - IP_INC_STATS_BH(IPSTATS_MIB_REASMFAILS); + IP_INC_STATS_BH(net, IPSTATS_MIB_REASMTIMEOUT); + IP_INC_STATS_BH(net, IPSTATS_MIB_REASMFAILS); if ((qp->q.last_in & INET_FRAG_FIRST_IN) && qp->q.fragments != NULL) { struct sk_buff *head = qp->q.fragments; @@ -261,7 +261,10 @@ static inline int ip_frag_too_far(struct ipq *qp) rc = qp->q.fragments && (end - start) > max; if (rc) { - IP_INC_STATS_BH(IPSTATS_MIB_REASMFAILS); + struct net *net; + + net = container_of(qp->q.net, struct net, ipv4.frags); + IP_INC_STATS_BH(net, IPSTATS_MIB_REASMFAILS); } return rc; @@ -545,7 +548,7 @@ static int ip_frag_reasm(struct ipq *qp, struct sk_buff *prev, iph = ip_hdr(head); iph->frag_off = 0; iph->tot_len = htons(len); - IP_INC_STATS_BH(IPSTATS_MIB_REASMOKS); + IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_REASMOKS); qp->q.fragments = NULL; return 0; @@ -560,7 +563,7 @@ out_oversize: "Oversized IP packet from " NIPQUAD_FMT ".\n", NIPQUAD(qp->saddr)); out_fail: - IP_INC_STATS_BH(IPSTATS_MIB_REASMFAILS); + IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_REASMFAILS); return err; } @@ -571,7 +574,7 @@ int ip_defrag(struct sk_buff *skb, u32 user) struct net *net; net = skb->dev ? dev_net(skb->dev) : dev_net(skb->dst->dev); - IP_INC_STATS_BH(IPSTATS_MIB_REASMREQDS); + IP_INC_STATS_BH(net, IPSTATS_MIB_REASMREQDS); /* Start by cleaning up the memory. */ if (atomic_read(&net->ipv4.frags.mem) > net->ipv4.frags.high_thresh) @@ -590,7 +593,7 @@ int ip_defrag(struct sk_buff *skb, u32 user) return ret; } - IP_INC_STATS_BH(IPSTATS_MIB_REASMFAILS); + IP_INC_STATS_BH(net, IPSTATS_MIB_REASMFAILS); kfree_skb(skb); return -ENOMEM; } diff --git a/net/ipv4/ip_input.c b/net/ipv4/ip_input.c index 7c26428ea67b..043f640df4b7 100644 --- a/net/ipv4/ip_input.c +++ b/net/ipv4/ip_input.c @@ -230,16 +230,16 @@ static int ip_local_deliver_finish(struct sk_buff *skb) protocol = -ret; goto resubmit; } - IP_INC_STATS_BH(IPSTATS_MIB_INDELIVERS); + IP_INC_STATS_BH(net, IPSTATS_MIB_INDELIVERS); } else { if (!raw) { if (xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) { - IP_INC_STATS_BH(IPSTATS_MIB_INUNKNOWNPROTOS); + IP_INC_STATS_BH(net, IPSTATS_MIB_INUNKNOWNPROTOS); icmp_send(skb, ICMP_DEST_UNREACH, ICMP_PROT_UNREACH, 0); } } else - IP_INC_STATS_BH(IPSTATS_MIB_INDELIVERS); + IP_INC_STATS_BH(net, IPSTATS_MIB_INDELIVERS); kfree_skb(skb); } } @@ -281,7 +281,7 @@ static inline int ip_rcv_options(struct sk_buff *skb) --ANK (980813) */ if (skb_cow(skb, skb_headroom(skb))) { - IP_INC_STATS_BH(IPSTATS_MIB_INDISCARDS); + IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS); goto drop; } @@ -290,7 +290,7 @@ static inline int ip_rcv_options(struct sk_buff *skb) opt->optlen = iph->ihl*4 - sizeof(struct iphdr); if (ip_options_compile(dev_net(dev), opt, skb)) { - IP_INC_STATS_BH(IPSTATS_MIB_INHDRERRORS); + IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INHDRERRORS); goto drop; } @@ -334,9 +334,11 @@ static int ip_rcv_finish(struct sk_buff *skb) skb->dev); if (unlikely(err)) { if (err == -EHOSTUNREACH) - IP_INC_STATS_BH(IPSTATS_MIB_INADDRERRORS); + IP_INC_STATS_BH(dev_net(skb->dev), + IPSTATS_MIB_INADDRERRORS); else if (err == -ENETUNREACH) - IP_INC_STATS_BH(IPSTATS_MIB_INNOROUTES); + IP_INC_STATS_BH(dev_net(skb->dev), + IPSTATS_MIB_INNOROUTES); goto drop; } } @@ -357,9 +359,9 @@ static int ip_rcv_finish(struct sk_buff *skb) rt = skb->rtable; if (rt->rt_type == RTN_MULTICAST) - IP_INC_STATS_BH(IPSTATS_MIB_INMCASTPKTS); + IP_INC_STATS_BH(dev_net(rt->u.dst.dev), IPSTATS_MIB_INMCASTPKTS); else if (rt->rt_type == RTN_BROADCAST) - IP_INC_STATS_BH(IPSTATS_MIB_INBCASTPKTS); + IP_INC_STATS_BH(dev_net(rt->u.dst.dev), IPSTATS_MIB_INBCASTPKTS); return dst_input(skb); @@ -382,10 +384,10 @@ int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, if (skb->pkt_type == PACKET_OTHERHOST) goto drop; - IP_INC_STATS_BH(IPSTATS_MIB_INRECEIVES); + IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INRECEIVES); if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL) { - IP_INC_STATS_BH(IPSTATS_MIB_INDISCARDS); + IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS); goto out; } @@ -418,7 +420,7 @@ int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, len = ntohs(iph->tot_len); if (skb->len < len) { - IP_INC_STATS_BH(IPSTATS_MIB_INTRUNCATEDPKTS); + IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INTRUNCATEDPKTS); goto drop; } else if (len < (iph->ihl*4)) goto inhdr_error; @@ -428,7 +430,7 @@ int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, * Note this now means skb->len holds ntohs(iph->tot_len). */ if (pskb_trim_rcsum(skb, len)) { - IP_INC_STATS_BH(IPSTATS_MIB_INDISCARDS); + IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS); goto drop; } @@ -439,7 +441,7 @@ int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, ip_rcv_finish); inhdr_error: - IP_INC_STATS_BH(IPSTATS_MIB_INHDRERRORS); + IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INHDRERRORS); drop: kfree_skb(skb); out: diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index c9ab47b966b5..033c712c3a5d 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c @@ -1178,7 +1178,7 @@ static inline int ipmr_forward_finish(struct sk_buff *skb) { struct ip_options * opt = &(IPCB(skb)->opt); - IP_INC_STATS_BH(IPSTATS_MIB_OUTFORWDATAGRAMS); + IP_INC_STATS_BH(dev_net(skb->dst->dev), IPSTATS_MIB_OUTFORWDATAGRAMS); if (unlikely(opt->optlen)) ip_forward_options(skb); @@ -1241,7 +1241,7 @@ static void ipmr_queue_xmit(struct sk_buff *skb, struct mfc_cache *c, int vifi) to blackhole. */ - IP_INC_STATS_BH(IPSTATS_MIB_FRAGFAILS); + IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_FRAGFAILS); ip_rt_put(rt); goto out_free; } diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 79c1e74263a1..e4ab0ac94f92 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -1429,7 +1429,8 @@ static int ip_error(struct sk_buff *skb) break; case ENETUNREACH: code = ICMP_NET_UNREACH; - IP_INC_STATS_BH(IPSTATS_MIB_INNOROUTES); + IP_INC_STATS_BH(dev_net(rt->u.dst.dev), + IPSTATS_MIB_INNOROUTES); break; case EACCES: code = ICMP_PKT_FILTERED; diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 0fefd4409413..7797d528701b 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -175,7 +175,7 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) inet->sport, usin->sin_port, sk, 1); if (tmp < 0) { if (tmp == -ENETUNREACH) - IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES); + IP_INC_STATS_BH(sock_net(sk), IPSTATS_MIB_OUTNOROUTES); return tmp; } diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 048ef57edc1d..1560d11ba6ef 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -662,7 +662,7 @@ int udp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg, err = ip_route_output_flow(net, &rt, &fl, sk, 1); if (err) { if (err == -ENETUNREACH) - IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES); + IP_INC_STATS_BH(net, IPSTATS_MIB_OUTNOROUTES); goto out; } diff --git a/net/sctp/output.c b/net/sctp/output.c index abcd00dc05eb..9a63a3fb9011 100644 --- a/net/sctp/output.c +++ b/net/sctp/output.c @@ -50,6 +50,7 @@ #include #include #include +#include #ifndef TEST_FRAME #include @@ -595,7 +596,7 @@ out: return err; no_route: kfree_skb(nskb); - IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES); + IP_INC_STATS_BH(&init_net, IPSTATS_MIB_OUTNOROUTES); /* FIXME: Returning the 'err' will effect all the associations * associated with a socket, although only one of the paths of the -- cgit v1.2.2 From c5346fe396f5e22bbfb3ec037c43891c3c57d3e6 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:20:33 -0700 Subject: mib: add net to IP_ADD_STATS_BH Very simple - only ip_evictor (fragments) requires such. This patch ends up the IP_XXX_STATS patching. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/ip_fragment.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c index 65364c06ada5..38d38f058018 100644 --- a/net/ipv4/ip_fragment.c +++ b/net/ipv4/ip_fragment.c @@ -178,7 +178,7 @@ static void ip_evictor(struct net *net) evicted = inet_frag_evictor(&net->ipv4.frags, &ip4_frags); if (evicted) - IP_ADD_STATS_BH(IPSTATS_MIB_REASMFAILS, evicted); + IP_ADD_STATS_BH(net, IPSTATS_MIB_REASMFAILS, evicted); } /* -- cgit v1.2.2 From a86b1e3019455283a677c2485cfeda2dc36df3eb Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:20:58 -0700 Subject: inet: prepare struct net for TCP MIB accounting This is the same as the first patch in the set, but preparing the net for TCP_XXX_STATS - save the struct net on the stack where required and possible. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/tcp_ipv4.c | 10 +++++++--- net/ipv6/tcp_ipv6.c | 3 ++- 2 files changed, 9 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 7797d528701b..db3bf9be076f 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -544,6 +544,7 @@ static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb) #ifdef CONFIG_TCP_MD5SIG struct tcp_md5sig_key *key; #endif + struct net *net; /* Never send a reset in response to a reset. */ if (th->rst) @@ -594,7 +595,8 @@ static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb) sizeof(struct tcphdr), IPPROTO_TCP, 0); arg.csumoffset = offsetof(struct tcphdr, check) / 2; - ip_send_reply(dev_net(skb->dst->dev)->ipv4.tcp_sock, skb, + net = dev_net(skb->dst->dev); + ip_send_reply(net->ipv4.tcp_sock, skb, &arg, arg.iov[0].iov_len); TCP_INC_STATS_BH(TCP_MIB_OUTSEGS); @@ -619,6 +621,7 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack, ]; } rep; struct ip_reply_arg arg; + struct net *net = dev_net(skb->dev); memset(&rep.th, 0, sizeof(struct tcphdr)); memset(&arg, 0, sizeof(arg)); @@ -668,7 +671,7 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack, if (oif) arg.bound_dev_if = oif; - ip_send_reply(dev_net(skb->dev)->ipv4.tcp_sock, skb, + ip_send_reply(net->ipv4.tcp_sock, skb, &arg, arg.iov[0].iov_len); TCP_INC_STATS_BH(TCP_MIB_OUTSEGS); @@ -1505,6 +1508,7 @@ int tcp_v4_rcv(struct sk_buff *skb) struct tcphdr *th; struct sock *sk; int ret; + struct net *net = dev_net(skb->dev); if (skb->pkt_type != PACKET_HOST) goto discard_it; @@ -1539,7 +1543,7 @@ int tcp_v4_rcv(struct sk_buff *skb) TCP_SKB_CB(skb)->flags = iph->tos; TCP_SKB_CB(skb)->sacked = 0; - sk = __inet_lookup(dev_net(skb->dev), &tcp_hashinfo, iph->saddr, + sk = __inet_lookup(net, &tcp_hashinfo, iph->saddr, th->source, iph->daddr, th->dest, inet_iif(skb)); if (!sk) goto no_tcp_socket; diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 30dbab7cc3cc..fc5f7166208a 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1617,6 +1617,7 @@ static int tcp_v6_rcv(struct sk_buff *skb) struct tcphdr *th; struct sock *sk; int ret; + struct net *net = dev_net(skb->dev); if (skb->pkt_type != PACKET_HOST) goto discard_it; @@ -1648,7 +1649,7 @@ static int tcp_v6_rcv(struct sk_buff *skb) TCP_SKB_CB(skb)->flags = ipv6_get_dsfield(ipv6_hdr(skb)); TCP_SKB_CB(skb)->sacked = 0; - sk = __inet6_lookup(dev_net(skb->dev), &tcp_hashinfo, + sk = __inet6_lookup(net, &tcp_hashinfo, &ipv6_hdr(skb)->saddr, th->source, &ipv6_hdr(skb)->daddr, ntohs(th->dest), inet6_iif(skb)); -- cgit v1.2.2 From a9c19329eccdb145a08a4a2e969d7b40c54c9bcc Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:21:42 -0700 Subject: tcp: add net to tcp_mib_init This one sets TCP MIBs after zeroing them, and thus requires the net. The existing single caller can use init_net (temporarily). Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/af_inet.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index dc411335c14f..95a966dd1916 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -1363,7 +1363,7 @@ static int __init init_ipv4_mibs(void) sizeof(struct udp_mib)) < 0) goto err_udplite_mib; - tcp_mib_init(); + tcp_mib_init(&init_net); return 0; -- cgit v1.2.2 From 81cc8a75d944fa39fc333c2c329c8e8b3c62cada Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:22:04 -0700 Subject: mib: add net to TCP_INC_STATS Fortunately (almost) all the TCP code has a sock to get the net from :) Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/tcp.c | 4 ++-- net/ipv4/tcp_output.c | 10 +++++----- 2 files changed, 7 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 631133ebb20a..aa79198fb02e 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -1668,12 +1668,12 @@ void tcp_set_state(struct sock *sk, int state) switch (state) { case TCP_ESTABLISHED: if (oldstate != TCP_ESTABLISHED) - TCP_INC_STATS(TCP_MIB_CURRESTAB); + TCP_INC_STATS(sock_net(sk), TCP_MIB_CURRESTAB); break; case TCP_CLOSE: if (oldstate == TCP_CLOSE_WAIT || oldstate == TCP_ESTABLISHED) - TCP_INC_STATS(TCP_MIB_ESTABRESETS); + TCP_INC_STATS(sock_net(sk), TCP_MIB_ESTABRESETS); sk->sk_prot->unhash(sk); if (inet_csk(sk)->icsk_bind_hash && diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index edef2afe905e..c3b0da826980 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -618,7 +618,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, tcp_event_data_sent(tp, skb, sk); if (after(tcb->end_seq, tp->snd_nxt) || tcb->seq == tcb->end_seq) - TCP_INC_STATS(TCP_MIB_OUTSEGS); + TCP_INC_STATS(sock_net(sk), TCP_MIB_OUTSEGS); err = icsk->icsk_af_ops->queue_xmit(skb, 0); if (likely(err <= 0)) @@ -1910,7 +1910,7 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb) if (err == 0) { /* Update global TCP statistics. */ - TCP_INC_STATS(TCP_MIB_RETRANSSEGS); + TCP_INC_STATS(sock_net(sk), TCP_MIB_RETRANSSEGS); tp->total_retrans++; @@ -2132,7 +2132,7 @@ void tcp_send_active_reset(struct sock *sk, gfp_t priority) if (tcp_transmit_skb(sk, skb, 0, priority)) NET_INC_STATS(LINUX_MIB_TCPABORTFAILED); - TCP_INC_STATS(TCP_MIB_OUTRSTS); + TCP_INC_STATS(sock_net(sk), TCP_MIB_OUTRSTS); } /* WARNING: This routine must only be called when we have already sent @@ -2258,7 +2258,7 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst, ); th->doff = (tcp_header_size >> 2); - TCP_INC_STATS(TCP_MIB_OUTSEGS); + TCP_INC_STATS(sock_net(sk), TCP_MIB_OUTSEGS); #ifdef CONFIG_TCP_MD5SIG /* Okay, we have all we need - do the md5 hash if needed */ @@ -2367,7 +2367,7 @@ int tcp_connect(struct sock *sk) */ tp->snd_nxt = tp->write_seq; tp->pushed_seq = tp->write_seq; - TCP_INC_STATS(TCP_MIB_ACTIVEOPENS); + TCP_INC_STATS(sock_net(sk), TCP_MIB_ACTIVEOPENS); /* Timer for repeating the SYN until an answer. */ inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, -- cgit v1.2.2 From 63231bddf6514778792d3784f63822473d250fc0 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:22:25 -0700 Subject: mib: add net to TCP_INC_STATS_BH Same as before - the sock is always there to get the net from, but there are also some places with the net already saved on the stack. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/tcp.c | 2 +- net/ipv4/tcp_input.c | 6 +++--- net/ipv4/tcp_ipv4.c | 14 +++++++------- net/ipv4/tcp_minisocks.c | 4 ++-- net/ipv6/tcp_ipv6.c | 14 +++++++------- 5 files changed, 20 insertions(+), 20 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index aa79198fb02e..f742f84026b5 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -2663,7 +2663,7 @@ EXPORT_SYMBOL(__tcp_put_md5sig_pool); void tcp_done(struct sock *sk) { if(sk->sk_state == TCP_SYN_SENT || sk->sk_state == TCP_SYN_RECV) - TCP_INC_STATS_BH(TCP_MIB_ATTEMPTFAILS); + TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_ATTEMPTFAILS); tcp_set_state(sk, TCP_CLOSE); tcp_clear_xmit_timers(sk); diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index d6ea970a1513..01e8004911b4 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -4795,7 +4795,7 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb, tcp_data_snd_check(sk); return 0; } else { /* Header too small */ - TCP_INC_STATS_BH(TCP_MIB_INERRS); + TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS); goto discard; } } else { @@ -4934,7 +4934,7 @@ slow_path: tcp_replace_ts_recent(tp, TCP_SKB_CB(skb)->seq); if (th->syn && !before(TCP_SKB_CB(skb)->seq, tp->rcv_nxt)) { - TCP_INC_STATS_BH(TCP_MIB_INERRS); + TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS); NET_INC_STATS_BH(LINUX_MIB_TCPABORTONSYN); tcp_reset(sk); return 1; @@ -4957,7 +4957,7 @@ step5: return 0; csum_error: - TCP_INC_STATS_BH(TCP_MIB_INERRS); + TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS); discard: __kfree_skb(skb); diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index db3bf9be076f..e876312b950a 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -599,8 +599,8 @@ static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb) ip_send_reply(net->ipv4.tcp_sock, skb, &arg, arg.iov[0].iov_len); - TCP_INC_STATS_BH(TCP_MIB_OUTSEGS); - TCP_INC_STATS_BH(TCP_MIB_OUTRSTS); + TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS); + TCP_INC_STATS_BH(net, TCP_MIB_OUTRSTS); } /* The code following below sending ACKs in SYN-RECV and TIME-WAIT states @@ -674,7 +674,7 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack, ip_send_reply(net->ipv4.tcp_sock, skb, &arg, arg.iov[0].iov_len); - TCP_INC_STATS_BH(TCP_MIB_OUTSEGS); + TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS); } static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb) @@ -1494,7 +1494,7 @@ discard: return 0; csum_err: - TCP_INC_STATS_BH(TCP_MIB_INERRS); + TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS); goto discard; } @@ -1514,7 +1514,7 @@ int tcp_v4_rcv(struct sk_buff *skb) goto discard_it; /* Count it even if it's bad */ - TCP_INC_STATS_BH(TCP_MIB_INSEGS); + TCP_INC_STATS_BH(net, TCP_MIB_INSEGS); if (!pskb_may_pull(skb, sizeof(struct tcphdr))) goto discard_it; @@ -1590,7 +1590,7 @@ no_tcp_socket: if (skb->len < (th->doff << 2) || tcp_checksum_complete(skb)) { bad_packet: - TCP_INC_STATS_BH(TCP_MIB_INERRS); + TCP_INC_STATS_BH(net, TCP_MIB_INERRS); } else { tcp_v4_send_reset(NULL, skb); } @@ -1611,7 +1611,7 @@ do_time_wait: } if (skb->len < (th->doff << 2) || tcp_checksum_complete(skb)) { - TCP_INC_STATS_BH(TCP_MIB_INERRS); + TCP_INC_STATS_BH(net, TCP_MIB_INERRS); inet_twsk_put(inet_twsk(sk)); goto discard_it; } diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index ea68a478fad6..8b02b1039968 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -480,7 +480,7 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req, newtp->rx_opt.mss_clamp = req->mss; TCP_ECN_openreq_child(newtp, req); - TCP_INC_STATS_BH(TCP_MIB_PASSIVEOPENS); + TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_PASSIVEOPENS); } return newsk; } @@ -630,7 +630,7 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb, * "fourth, check the SYN bit" */ if (flg & (TCP_FLAG_RST|TCP_FLAG_SYN)) { - TCP_INC_STATS_BH(TCP_MIB_ATTEMPTFAILS); + TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_ATTEMPTFAILS); goto embryonic_reset; } diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index fc5f7166208a..3895d91ea9fa 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1004,8 +1004,8 @@ static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb) if (xfrm_lookup(&buff->dst, &fl, NULL, 0) >= 0) { ip6_xmit(ctl_sk, buff, &fl, NULL, 0); - TCP_INC_STATS_BH(TCP_MIB_OUTSEGS); - TCP_INC_STATS_BH(TCP_MIB_OUTRSTS); + TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS); + TCP_INC_STATS_BH(net, TCP_MIB_OUTRSTS); return; } } @@ -1089,7 +1089,7 @@ static void tcp_v6_send_ack(struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 if (!ip6_dst_lookup(ctl_sk, &buff->dst, &fl)) { if (xfrm_lookup(&buff->dst, &fl, NULL, 0) >= 0) { ip6_xmit(ctl_sk, buff, &fl, NULL, 0); - TCP_INC_STATS_BH(TCP_MIB_OUTSEGS); + TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS); return; } } @@ -1579,7 +1579,7 @@ discard: kfree_skb(skb); return 0; csum_err: - TCP_INC_STATS_BH(TCP_MIB_INERRS); + TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS); goto discard; @@ -1625,7 +1625,7 @@ static int tcp_v6_rcv(struct sk_buff *skb) /* * Count it even if it's bad. */ - TCP_INC_STATS_BH(TCP_MIB_INSEGS); + TCP_INC_STATS_BH(net, TCP_MIB_INSEGS); if (!pskb_may_pull(skb, sizeof(struct tcphdr))) goto discard_it; @@ -1697,7 +1697,7 @@ no_tcp_socket: if (skb->len < (th->doff<<2) || tcp_checksum_complete(skb)) { bad_packet: - TCP_INC_STATS_BH(TCP_MIB_INERRS); + TCP_INC_STATS_BH(net, TCP_MIB_INERRS); } else { tcp_v6_send_reset(NULL, skb); } @@ -1722,7 +1722,7 @@ do_time_wait: } if (skb->len < (th->doff<<2) || tcp_checksum_complete(skb)) { - TCP_INC_STATS_BH(TCP_MIB_INERRS); + TCP_INC_STATS_BH(net, TCP_MIB_INERRS); inet_twsk_put(inet_twsk(sk)); goto discard_it; } -- cgit v1.2.2 From 74688e487a407a33d42879957b478601aca616b8 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:22:46 -0700 Subject: mib: add net to TCP_DEC_STATS Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/tcp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index f742f84026b5..525dcf534153 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -1682,7 +1682,7 @@ void tcp_set_state(struct sock *sk, int state) /* fall through */ default: if (oldstate==TCP_ESTABLISHED) - TCP_DEC_STATS(TCP_MIB_CURRESTAB); + TCP_DEC_STATS(sock_net(sk), TCP_MIB_CURRESTAB); } /* Change state AFTER socket is unhashed to avoid closed -- cgit v1.2.2 From 5c52ba170f8167511bdb65b981f4582100c40675 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:28:10 -0700 Subject: sock: add net to prot->enter_memory_pressure callback The tcp_enter_memory_pressure calls NET_INC_STATS, but doesn't have where to get the net from. I decided to add a sk argument, not the net itself, only to factor all the required sock_net(sk) calls inside the enter_memory_pressure callback itself. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/core/sock.c | 2 +- net/decnet/af_decnet.c | 2 +- net/ipv4/tcp.c | 4 ++-- net/sctp/socket.c | 2 +- 4 files changed, 5 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/core/sock.c b/net/core/sock.c index 2c0ba52e5303..10a64d57078c 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -1442,7 +1442,7 @@ int __sk_mem_schedule(struct sock *sk, int size, int kind) /* Under pressure. */ if (allocated > prot->sysctl_mem[1]) if (prot->enter_memory_pressure) - prot->enter_memory_pressure(); + prot->enter_memory_pressure(sk); /* Over hard limit. */ if (allocated > prot->sysctl_mem[2]) diff --git a/net/decnet/af_decnet.c b/net/decnet/af_decnet.c index 931bdf9cb756..61b7df577ddd 100644 --- a/net/decnet/af_decnet.c +++ b/net/decnet/af_decnet.c @@ -451,7 +451,7 @@ static void dn_destruct(struct sock *sk) static int dn_memory_pressure; -static void dn_enter_memory_pressure(void) +static void dn_enter_memory_pressure(struct sock *sk) { if (!dn_memory_pressure) { dn_memory_pressure = 1; diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 525dcf534153..bc8559a6f7e5 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -316,7 +316,7 @@ int tcp_memory_pressure __read_mostly; EXPORT_SYMBOL(tcp_memory_pressure); -void tcp_enter_memory_pressure(void) +void tcp_enter_memory_pressure(struct sock *sk) { if (!tcp_memory_pressure) { NET_INC_STATS(LINUX_MIB_TCPMEMORYPRESSURES); @@ -649,7 +649,7 @@ struct sk_buff *sk_stream_alloc_skb(struct sock *sk, int size, gfp_t gfp) } __kfree_skb(skb); } else { - sk->sk_prot->enter_memory_pressure(); + sk->sk_prot->enter_memory_pressure(sk); sk_stream_moderate_sndbuf(sk); } return NULL; diff --git a/net/sctp/socket.c b/net/sctp/socket.c index df5572c39f0c..6aba01b0ce4e 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -116,7 +116,7 @@ static int sctp_memory_pressure; static atomic_t sctp_memory_allocated; static atomic_t sctp_sockets_allocated; -static void sctp_enter_memory_pressure(void) +static void sctp_enter_memory_pressure(struct sock *sk) { sctp_memory_pressure = 1; } -- cgit v1.2.2 From ca12a1a443a51298afcca627ad0bcbd8ad1dcddc Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:28:42 -0700 Subject: inet: prepare net on the stack for NET accounting macros Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/dccp/ipv6.c | 3 ++- net/ipv4/arp.c | 3 ++- net/ipv6/tcp_ipv6.c | 3 ++- 3 files changed, 6 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c index eec3c4717890..83cc9bbc6208 100644 --- a/net/dccp/ipv6.c +++ b/net/dccp/ipv6.c @@ -93,8 +93,9 @@ static void dccp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, struct sock *sk; int err; __u64 seq; + struct net *net = dev_net(skb->dev); - sk = inet6_lookup(dev_net(skb->dev), &dccp_hashinfo, + sk = inet6_lookup(net, &dccp_hashinfo, &hdr->daddr, dh->dccph_dport, &hdr->saddr, dh->dccph_sport, inet6_iif(skb)); diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c index 29df75a6bcc7..aab98b8a9945 100644 --- a/net/ipv4/arp.c +++ b/net/ipv4/arp.c @@ -421,8 +421,9 @@ static int arp_filter(__be32 sip, __be32 tip, struct net_device *dev) struct rtable *rt; int flag = 0; /*unsigned long now; */ + struct net *net = dev_net(dev); - if (ip_route_output_key(dev_net(dev), &rt, &fl) < 0) + if (ip_route_output_key(net, &rt, &fl) < 0) return 1; if (rt->u.dst.dev != dev) { NET_INC_STATS_BH(LINUX_MIB_ARPFILTER); diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 3895d91ea9fa..d58b83ac06fb 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -323,8 +323,9 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, int err; struct tcp_sock *tp; __u32 seq; + struct net *net = dev_net(skb->dev); - sk = inet6_lookup(dev_net(skb->dev), &tcp_hashinfo, &hdr->daddr, + sk = inet6_lookup(net, &tcp_hashinfo, &hdr->daddr, th->dest, &hdr->saddr, th->source, skb->dev->ifindex); if (sk == NULL) { -- cgit v1.2.2 From 1ed834655a0d42ecd80ff051e681e2ea44747b6c Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:29:51 -0700 Subject: tcp: replace tcp_sock argument with sock in some places These places have a tcp_sock, but we'd prefer the sock itself to get net from it. Fortunately, tcp_sk macro is just a type cast, so this replace is really cheap. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/tcp_input.c | 31 ++++++++++++++++++------------- 1 file changed, 18 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 01e8004911b4..f50d8433f042 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -1169,10 +1169,11 @@ static void tcp_mark_lost_retrans(struct sock *sk) tp->lost_retrans_low = new_low_seq; } -static int tcp_check_dsack(struct tcp_sock *tp, struct sk_buff *ack_skb, +static int tcp_check_dsack(struct sock *sk, struct sk_buff *ack_skb, struct tcp_sack_block_wire *sp, int num_sacks, u32 prior_snd_una) { + struct tcp_sock *tp = tcp_sk(sk); u32 start_seq_0 = get_unaligned_be32(&sp[0].start_seq); u32 end_seq_0 = get_unaligned_be32(&sp[0].end_seq); int dup_sack = 0; @@ -1434,7 +1435,7 @@ tcp_sacktag_write_queue(struct sock *sk, struct sk_buff *ack_skb, tcp_highest_sack_reset(sk); } - found_dup_sack = tcp_check_dsack(tp, ack_skb, sp_wire, + found_dup_sack = tcp_check_dsack(sk, ack_skb, sp_wire, num_sacks, prior_snd_una); if (found_dup_sack) flag |= FLAG_DSACKING_ACK; @@ -3711,8 +3712,10 @@ static inline int tcp_sack_extend(struct tcp_sack_block *sp, u32 seq, return 0; } -static void tcp_dsack_set(struct tcp_sock *tp, u32 seq, u32 end_seq) +static void tcp_dsack_set(struct sock *sk, u32 seq, u32 end_seq) { + struct tcp_sock *tp = tcp_sk(sk); + if (tcp_is_sack(tp) && sysctl_tcp_dsack) { int mib_idx; @@ -3731,10 +3734,12 @@ static void tcp_dsack_set(struct tcp_sock *tp, u32 seq, u32 end_seq) } } -static void tcp_dsack_extend(struct tcp_sock *tp, u32 seq, u32 end_seq) +static void tcp_dsack_extend(struct sock *sk, u32 seq, u32 end_seq) { + struct tcp_sock *tp = tcp_sk(sk); + if (!tp->rx_opt.dsack) - tcp_dsack_set(tp, seq, end_seq); + tcp_dsack_set(sk, seq, end_seq); else tcp_sack_extend(tp->duplicate_sack, seq, end_seq); } @@ -3753,7 +3758,7 @@ static void tcp_send_dupack(struct sock *sk, struct sk_buff *skb) if (after(TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt)) end_seq = tp->rcv_nxt; - tcp_dsack_set(tp, TCP_SKB_CB(skb)->seq, end_seq); + tcp_dsack_set(sk, TCP_SKB_CB(skb)->seq, end_seq); } } @@ -3906,7 +3911,7 @@ static void tcp_ofo_queue(struct sock *sk) __u32 dsack = dsack_high; if (before(TCP_SKB_CB(skb)->end_seq, dsack_high)) dsack_high = TCP_SKB_CB(skb)->end_seq; - tcp_dsack_extend(tp, TCP_SKB_CB(skb)->seq, dsack); + tcp_dsack_extend(sk, TCP_SKB_CB(skb)->seq, dsack); } if (!after(TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt)) { @@ -4035,7 +4040,7 @@ queue_and_out: if (!after(TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt)) { /* A retransmit, 2nd most common case. Force an immediate ack. */ NET_INC_STATS_BH(LINUX_MIB_DELAYEDACKLOST); - tcp_dsack_set(tp, TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb)->end_seq); + tcp_dsack_set(sk, TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb)->end_seq); out_of_window: tcp_enter_quickack_mode(sk); @@ -4057,7 +4062,7 @@ drop: tp->rcv_nxt, TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb)->end_seq); - tcp_dsack_set(tp, TCP_SKB_CB(skb)->seq, tp->rcv_nxt); + tcp_dsack_set(sk, TCP_SKB_CB(skb)->seq, tp->rcv_nxt); /* If window is closed, drop tail of packet. But after * remembering D-SACK for its head made in previous line. @@ -4122,12 +4127,12 @@ drop: if (!after(end_seq, TCP_SKB_CB(skb1)->end_seq)) { /* All the bits are present. Drop. */ __kfree_skb(skb); - tcp_dsack_set(tp, seq, end_seq); + tcp_dsack_set(sk, seq, end_seq); goto add_sack; } if (after(seq, TCP_SKB_CB(skb1)->seq)) { /* Partial overlap. */ - tcp_dsack_set(tp, seq, + tcp_dsack_set(sk, seq, TCP_SKB_CB(skb1)->end_seq); } else { skb1 = skb1->prev; @@ -4140,12 +4145,12 @@ drop: (struct sk_buff *)&tp->out_of_order_queue && after(end_seq, TCP_SKB_CB(skb1)->seq)) { if (before(end_seq, TCP_SKB_CB(skb1)->end_seq)) { - tcp_dsack_extend(tp, TCP_SKB_CB(skb1)->seq, + tcp_dsack_extend(sk, TCP_SKB_CB(skb1)->seq, end_seq); break; } __skb_unlink(skb1, &tp->out_of_order_queue); - tcp_dsack_extend(tp, TCP_SKB_CB(skb1)->seq, + tcp_dsack_extend(sk, TCP_SKB_CB(skb1)->seq, TCP_SKB_CB(skb1)->end_seq); __kfree_skb(skb1); } -- cgit v1.2.2 From 4e6734447dbc7a0a85e09616821c0782d9fb1141 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:30:14 -0700 Subject: mib: add net to NET_INC_STATS Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/tcp.c | 2 +- net/ipv4/tcp_output.c | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index bc8559a6f7e5..85f08291e928 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -319,7 +319,7 @@ EXPORT_SYMBOL(tcp_memory_pressure); void tcp_enter_memory_pressure(struct sock *sk) { if (!tcp_memory_pressure) { - NET_INC_STATS(LINUX_MIB_TCPMEMORYPRESSURES); + NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMEMORYPRESSURES); tcp_memory_pressure = 1; } } diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index c3b0da826980..176f0702b8ac 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -2119,7 +2119,7 @@ void tcp_send_active_reset(struct sock *sk, gfp_t priority) /* NOTE: No TCP options attached and we never retransmit this. */ skb = alloc_skb(MAX_TCP_HEADER, priority); if (!skb) { - NET_INC_STATS(LINUX_MIB_TCPABORTFAILED); + NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPABORTFAILED); return; } @@ -2130,7 +2130,7 @@ void tcp_send_active_reset(struct sock *sk, gfp_t priority) /* Send it off. */ TCP_SKB_CB(skb)->when = tcp_time_stamp; if (tcp_transmit_skb(sk, skb, 0, priority)) - NET_INC_STATS(LINUX_MIB_TCPABORTFAILED); + NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPABORTFAILED); TCP_INC_STATS(sock_net(sk), TCP_MIB_OUTRSTS); } -- cgit v1.2.2 From de0744af1fe2d0a3d428f6af0f2fe1f6179b1a9c Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:31:16 -0700 Subject: mib: add net to NET_INC_STATS_BH Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/dccp/ipv4.c | 10 +++---- net/dccp/ipv6.c | 8 +++--- net/dccp/timer.c | 4 +-- net/ipv4/arp.c | 2 +- net/ipv4/inet_hashtables.c | 4 +-- net/ipv4/syncookies.c | 6 ++--- net/ipv4/tcp.c | 6 +++-- net/ipv4/tcp_input.c | 65 +++++++++++++++++++++++---------------------- net/ipv4/tcp_ipv4.c | 12 ++++----- net/ipv4/tcp_minisocks.c | 6 ++--- net/ipv4/tcp_output.c | 4 +-- net/ipv4/tcp_timer.c | 12 ++++----- net/ipv6/inet6_hashtables.c | 4 +-- net/ipv6/syncookies.c | 6 ++--- net/ipv6/tcp_ipv6.c | 10 +++---- net/sctp/input.c | 2 +- 16 files changed, 82 insertions(+), 79 deletions(-) (limited to 'net') diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index 9f760a1e312c..2622ace17c46 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -230,7 +230,7 @@ static void dccp_v4_err(struct sk_buff *skb, u32 info) * servers this needs to be solved differently. */ if (sock_owned_by_user(sk)) - NET_INC_STATS_BH(LINUX_MIB_LOCKDROPPEDICMPS); + NET_INC_STATS_BH(net, LINUX_MIB_LOCKDROPPEDICMPS); if (sk->sk_state == DCCP_CLOSED) goto out; @@ -239,7 +239,7 @@ static void dccp_v4_err(struct sk_buff *skb, u32 info) seq = dccp_hdr_seq(dh); if ((1 << sk->sk_state) & ~(DCCPF_REQUESTING | DCCPF_LISTEN) && !between48(seq, dp->dccps_swl, dp->dccps_swh)) { - NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS); + NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS); goto out; } @@ -286,7 +286,7 @@ static void dccp_v4_err(struct sk_buff *skb, u32 info) BUG_TRAP(!req->sk); if (seq != dccp_rsk(req)->dreq_iss) { - NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS); + NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS); goto out; } /* @@ -409,9 +409,9 @@ struct sock *dccp_v4_request_recv_sock(struct sock *sk, struct sk_buff *skb, return newsk; exit_overflow: - NET_INC_STATS_BH(LINUX_MIB_LISTENOVERFLOWS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS); exit: - NET_INC_STATS_BH(LINUX_MIB_LISTENDROPS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS); dst_release(dst); return NULL; } diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c index 83cc9bbc6208..b74e8b2cbe55 100644 --- a/net/dccp/ipv6.c +++ b/net/dccp/ipv6.c @@ -111,7 +111,7 @@ static void dccp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, bh_lock_sock(sk); if (sock_owned_by_user(sk)) - NET_INC_STATS_BH(LINUX_MIB_LOCKDROPPEDICMPS); + NET_INC_STATS_BH(net, LINUX_MIB_LOCKDROPPEDICMPS); if (sk->sk_state == DCCP_CLOSED) goto out; @@ -189,7 +189,7 @@ static void dccp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, BUG_TRAP(req->sk == NULL); if (seq != dccp_rsk(req)->dreq_iss) { - NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS); + NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS); goto out; } @@ -630,9 +630,9 @@ static struct sock *dccp_v6_request_recv_sock(struct sock *sk, return newsk; out_overflow: - NET_INC_STATS_BH(LINUX_MIB_LISTENOVERFLOWS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS); out: - NET_INC_STATS_BH(LINUX_MIB_LISTENDROPS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS); if (opt != NULL && opt != np->opt) sock_kfree_s(sk, opt, opt->tot_len); dst_release(dst); diff --git a/net/dccp/timer.c b/net/dccp/timer.c index 8703a792b560..3608d5342ca2 100644 --- a/net/dccp/timer.c +++ b/net/dccp/timer.c @@ -224,7 +224,7 @@ static void dccp_delack_timer(unsigned long data) if (sock_owned_by_user(sk)) { /* Try again later. */ icsk->icsk_ack.blocked = 1; - NET_INC_STATS_BH(LINUX_MIB_DELAYEDACKLOCKED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_DELAYEDACKLOCKED); sk_reset_timer(sk, &icsk->icsk_delack_timer, jiffies + TCP_DELACK_MIN); goto out; @@ -254,7 +254,7 @@ static void dccp_delack_timer(unsigned long data) icsk->icsk_ack.ato = TCP_ATO_MIN; } dccp_send_ack(sk); - NET_INC_STATS_BH(LINUX_MIB_DELAYEDACKS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_DELAYEDACKS); } out: bh_unlock_sock(sk); diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c index aab98b8a9945..b043eda60b04 100644 --- a/net/ipv4/arp.c +++ b/net/ipv4/arp.c @@ -426,7 +426,7 @@ static int arp_filter(__be32 sip, __be32 tip, struct net_device *dev) if (ip_route_output_key(net, &rt, &fl) < 0) return 1; if (rt->u.dst.dev != dev) { - NET_INC_STATS_BH(LINUX_MIB_ARPFILTER); + NET_INC_STATS_BH(net, LINUX_MIB_ARPFILTER); flag = 1; } ip_rt_put(rt); diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index eca5899729e3..115f53722d20 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -312,11 +312,11 @@ unique: if (twp) { *twp = tw; - NET_INC_STATS_BH(LINUX_MIB_TIMEWAITRECYCLED); + NET_INC_STATS_BH(net, LINUX_MIB_TIMEWAITRECYCLED); } else if (tw) { /* Silly. Should hash-dance instead... */ inet_twsk_deschedule(tw, death_row); - NET_INC_STATS_BH(LINUX_MIB_TIMEWAITRECYCLED); + NET_INC_STATS_BH(net, LINUX_MIB_TIMEWAITRECYCLED); inet_twsk_put(tw); } diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c index fdde2ae07e24..51bc24d3b8a7 100644 --- a/net/ipv4/syncookies.c +++ b/net/ipv4/syncookies.c @@ -173,7 +173,7 @@ __u32 cookie_v4_init_sequence(struct sock *sk, struct sk_buff *skb, __u16 *mssp) ; *mssp = msstab[mssind] + 1; - NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESSENT); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESSENT); return secure_tcp_syn_cookie(iph->saddr, iph->daddr, th->source, th->dest, ntohl(th->seq), @@ -269,11 +269,11 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb, if (time_after(jiffies, tp->last_synq_overflow + TCP_TIMEOUT_INIT) || (mss = cookie_check(skb, cookie)) == 0) { - NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESFAILED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESFAILED); goto out; } - NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESRECV); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESRECV); /* check for timestamp cookie support */ memset(&tcp_opt, 0, sizeof(tcp_opt)); diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 85f08291e928..9e0e45c37806 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -1871,7 +1871,8 @@ adjudge_to_death: if (tp->linger2 < 0) { tcp_set_state(sk, TCP_CLOSE); tcp_send_active_reset(sk, GFP_ATOMIC); - NET_INC_STATS_BH(LINUX_MIB_TCPABORTONLINGER); + NET_INC_STATS_BH(sock_net(sk), + LINUX_MIB_TCPABORTONLINGER); } else { const int tmo = tcp_fin_time(sk); @@ -1893,7 +1894,8 @@ adjudge_to_death: "sockets\n"); tcp_set_state(sk, TCP_CLOSE); tcp_send_active_reset(sk, GFP_ATOMIC); - NET_INC_STATS_BH(LINUX_MIB_TCPABORTONMEMORY); + NET_INC_STATS_BH(sock_net(sk), + LINUX_MIB_TCPABORTONMEMORY); } } diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index f50d8433f042..fac49a6e1611 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -961,7 +961,7 @@ static void tcp_update_reordering(struct sock *sk, const int metric, else mib_idx = LINUX_MIB_TCPSACKREORDER; - NET_INC_STATS_BH(mib_idx); + NET_INC_STATS_BH(sock_net(sk), mib_idx); #if FASTRETRANS_DEBUG > 1 printk(KERN_DEBUG "Disorder%d %d %u f%u s%u rr%d\n", tp->rx_opt.sack_ok, inet_csk(sk)->icsk_ca_state, @@ -1157,7 +1157,7 @@ static void tcp_mark_lost_retrans(struct sock *sk) tp->lost_out += tcp_skb_pcount(skb); TCP_SKB_CB(skb)->sacked |= TCPCB_LOST; } - NET_INC_STATS_BH(LINUX_MIB_TCPLOSTRETRANSMIT); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPLOSTRETRANSMIT); } else { if (before(ack_seq, new_low_seq)) new_low_seq = ack_seq; @@ -1181,7 +1181,7 @@ static int tcp_check_dsack(struct sock *sk, struct sk_buff *ack_skb, if (before(start_seq_0, TCP_SKB_CB(ack_skb)->ack_seq)) { dup_sack = 1; tcp_dsack_seen(tp); - NET_INC_STATS_BH(LINUX_MIB_TCPDSACKRECV); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPDSACKRECV); } else if (num_sacks > 1) { u32 end_seq_1 = get_unaligned_be32(&sp[1].end_seq); u32 start_seq_1 = get_unaligned_be32(&sp[1].start_seq); @@ -1190,7 +1190,8 @@ static int tcp_check_dsack(struct sock *sk, struct sk_buff *ack_skb, !before(start_seq_0, start_seq_1)) { dup_sack = 1; tcp_dsack_seen(tp); - NET_INC_STATS_BH(LINUX_MIB_TCPDSACKOFORECV); + NET_INC_STATS_BH(sock_net(sk), + LINUX_MIB_TCPDSACKOFORECV); } } @@ -1476,7 +1477,7 @@ tcp_sacktag_write_queue(struct sock *sk, struct sk_buff *ack_skb, mib_idx = LINUX_MIB_TCPSACKDISCARD; } - NET_INC_STATS_BH(mib_idx); + NET_INC_STATS_BH(sock_net(sk), mib_idx); if (i == 0) first_sack_index = -1; continue; @@ -1969,7 +1970,7 @@ static int tcp_check_sack_reneging(struct sock *sk, int flag) { if (flag & FLAG_SACK_RENEGING) { struct inet_connection_sock *icsk = inet_csk(sk); - NET_INC_STATS_BH(LINUX_MIB_TCPSACKRENEGING); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPSACKRENEGING); tcp_enter_loss(sk, 1); icsk->icsk_retransmits++; @@ -2401,7 +2402,7 @@ static int tcp_try_undo_recovery(struct sock *sk) else mib_idx = LINUX_MIB_TCPFULLUNDO; - NET_INC_STATS_BH(mib_idx); + NET_INC_STATS_BH(sock_net(sk), mib_idx); tp->undo_marker = 0; } if (tp->snd_una == tp->high_seq && tcp_is_reno(tp)) { @@ -2424,7 +2425,7 @@ static void tcp_try_undo_dsack(struct sock *sk) DBGUNDO(sk, "D-SACK"); tcp_undo_cwr(sk, 1); tp->undo_marker = 0; - NET_INC_STATS_BH(LINUX_MIB_TCPDSACKUNDO); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPDSACKUNDO); } } @@ -2447,7 +2448,7 @@ static int tcp_try_undo_partial(struct sock *sk, int acked) DBGUNDO(sk, "Hoe"); tcp_undo_cwr(sk, 0); - NET_INC_STATS_BH(LINUX_MIB_TCPPARTIALUNDO); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPPARTIALUNDO); /* So... Do not make Hoe's retransmit yet. * If the first packet was delayed, the rest @@ -2476,7 +2477,7 @@ static int tcp_try_undo_loss(struct sock *sk) DBGUNDO(sk, "partial loss"); tp->lost_out = 0; tcp_undo_cwr(sk, 1); - NET_INC_STATS_BH(LINUX_MIB_TCPLOSSUNDO); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPLOSSUNDO); inet_csk(sk)->icsk_retransmits = 0; tp->undo_marker = 0; if (tcp_is_sack(tp)) @@ -2595,7 +2596,7 @@ static void tcp_fastretrans_alert(struct sock *sk, int pkts_acked, int flag) icsk->icsk_ca_state != TCP_CA_Open && tp->fackets_out > tp->reordering) { tcp_mark_head_lost(sk, tp->fackets_out - tp->reordering); - NET_INC_STATS_BH(LINUX_MIB_TCPLOSS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPLOSS); } /* D. Check consistency of the current state. */ @@ -2700,7 +2701,7 @@ static void tcp_fastretrans_alert(struct sock *sk, int pkts_acked, int flag) else mib_idx = LINUX_MIB_TCPSACKRECOVERY; - NET_INC_STATS_BH(mib_idx); + NET_INC_STATS_BH(sock_net(sk), mib_idx); tp->high_seq = tp->snd_nxt; tp->prior_ssthresh = 0; @@ -3211,7 +3212,7 @@ static int tcp_process_frto(struct sock *sk, int flag) } tp->frto_counter = 0; tp->undo_marker = 0; - NET_INC_STATS_BH(LINUX_MIB_TCPSPURIOUSRTOS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPSPURIOUSRTOS); } return 0; } @@ -3264,12 +3265,12 @@ static int tcp_ack(struct sock *sk, struct sk_buff *skb, int flag) tcp_ca_event(sk, CA_EVENT_FAST_ACK); - NET_INC_STATS_BH(LINUX_MIB_TCPHPACKS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPHPACKS); } else { if (ack_seq != TCP_SKB_CB(skb)->end_seq) flag |= FLAG_DATA; else - NET_INC_STATS_BH(LINUX_MIB_TCPPUREACKS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPPUREACKS); flag |= tcp_ack_update_window(sk, skb, ack, ack_seq); @@ -3724,7 +3725,7 @@ static void tcp_dsack_set(struct sock *sk, u32 seq, u32 end_seq) else mib_idx = LINUX_MIB_TCPDSACKOFOSENT; - NET_INC_STATS_BH(mib_idx); + NET_INC_STATS_BH(sock_net(sk), mib_idx); tp->rx_opt.dsack = 1; tp->duplicate_sack[0].start_seq = seq; @@ -3750,7 +3751,7 @@ static void tcp_send_dupack(struct sock *sk, struct sk_buff *skb) if (TCP_SKB_CB(skb)->end_seq != TCP_SKB_CB(skb)->seq && before(TCP_SKB_CB(skb)->seq, tp->rcv_nxt)) { - NET_INC_STATS_BH(LINUX_MIB_DELAYEDACKLOST); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_DELAYEDACKLOST); tcp_enter_quickack_mode(sk); if (tcp_is_sack(tp) && sysctl_tcp_dsack) { @@ -4039,7 +4040,7 @@ queue_and_out: if (!after(TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt)) { /* A retransmit, 2nd most common case. Force an immediate ack. */ - NET_INC_STATS_BH(LINUX_MIB_DELAYEDACKLOST); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_DELAYEDACKLOST); tcp_dsack_set(sk, TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb)->end_seq); out_of_window: @@ -4181,7 +4182,7 @@ tcp_collapse(struct sock *sk, struct sk_buff_head *list, struct sk_buff *next = skb->next; __skb_unlink(skb, list); __kfree_skb(skb); - NET_INC_STATS_BH(LINUX_MIB_TCPRCVCOLLAPSED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPRCVCOLLAPSED); skb = next; continue; } @@ -4249,7 +4250,7 @@ tcp_collapse(struct sock *sk, struct sk_buff_head *list, struct sk_buff *next = skb->next; __skb_unlink(skb, list); __kfree_skb(skb); - NET_INC_STATS_BH(LINUX_MIB_TCPRCVCOLLAPSED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPRCVCOLLAPSED); skb = next; if (skb == tail || tcp_hdr(skb)->syn || @@ -4312,7 +4313,7 @@ static int tcp_prune_ofo_queue(struct sock *sk) int res = 0; if (!skb_queue_empty(&tp->out_of_order_queue)) { - NET_INC_STATS_BH(LINUX_MIB_OFOPRUNED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_OFOPRUNED); __skb_queue_purge(&tp->out_of_order_queue); /* Reset SACK state. A conforming SACK implementation will @@ -4341,7 +4342,7 @@ static int tcp_prune_queue(struct sock *sk) SOCK_DEBUG(sk, "prune_queue: c=%x\n", tp->copied_seq); - NET_INC_STATS_BH(LINUX_MIB_PRUNECALLED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_PRUNECALLED); if (atomic_read(&sk->sk_rmem_alloc) >= sk->sk_rcvbuf) tcp_clamp_window(sk); @@ -4370,7 +4371,7 @@ static int tcp_prune_queue(struct sock *sk) * drop receive data on the floor. It will get retransmitted * and hopefully then we'll have sufficient space. */ - NET_INC_STATS_BH(LINUX_MIB_RCVPRUNED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_RCVPRUNED); /* Massive buffer overcommit. */ tp->pred_flags = 0; @@ -4837,7 +4838,7 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb, __skb_pull(skb, tcp_header_len); tp->rcv_nxt = TCP_SKB_CB(skb)->end_seq; - NET_INC_STATS_BH(LINUX_MIB_TCPHPHITSTOUSER); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPHPHITSTOUSER); } if (copied_early) tcp_cleanup_rbuf(sk, skb->len); @@ -4860,7 +4861,7 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb, if ((int)skb->truesize > sk->sk_forward_alloc) goto step5; - NET_INC_STATS_BH(LINUX_MIB_TCPHPHITS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPHPHITS); /* Bulk data transfer: receiver */ __skb_pull(skb, tcp_header_len); @@ -4904,7 +4905,7 @@ slow_path: if (tcp_fast_parse_options(skb, th, tp) && tp->rx_opt.saw_tstamp && tcp_paws_discard(sk, skb)) { if (!th->rst) { - NET_INC_STATS_BH(LINUX_MIB_PAWSESTABREJECTED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_PAWSESTABREJECTED); tcp_send_dupack(sk, skb); goto discard; } @@ -4940,7 +4941,7 @@ slow_path: if (th->syn && !before(TCP_SKB_CB(skb)->seq, tp->rcv_nxt)) { TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS); - NET_INC_STATS_BH(LINUX_MIB_TCPABORTONSYN); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPABORTONSYN); tcp_reset(sk); return 1; } @@ -4996,7 +4997,7 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb, if (tp->rx_opt.saw_tstamp && tp->rx_opt.rcv_tsecr && !between(tp->rx_opt.rcv_tsecr, tp->retrans_stamp, tcp_time_stamp)) { - NET_INC_STATS_BH(LINUX_MIB_PAWSACTIVEREJECTED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_PAWSACTIVEREJECTED); goto reset_and_undo; } @@ -5280,7 +5281,7 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb, if (tcp_fast_parse_options(skb, th, tp) && tp->rx_opt.saw_tstamp && tcp_paws_discard(sk, skb)) { if (!th->rst) { - NET_INC_STATS_BH(LINUX_MIB_PAWSESTABREJECTED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_PAWSESTABREJECTED); tcp_send_dupack(sk, skb); goto discard; } @@ -5309,7 +5310,7 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb, * Check for a SYN in window. */ if (th->syn && !before(TCP_SKB_CB(skb)->seq, tp->rcv_nxt)) { - NET_INC_STATS_BH(LINUX_MIB_TCPABORTONSYN); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPABORTONSYN); tcp_reset(sk); return 1; } @@ -5391,7 +5392,7 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb, (TCP_SKB_CB(skb)->end_seq != TCP_SKB_CB(skb)->seq && after(TCP_SKB_CB(skb)->end_seq - th->fin, tp->rcv_nxt))) { tcp_done(sk); - NET_INC_STATS_BH(LINUX_MIB_TCPABORTONDATA); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPABORTONDATA); return 1; } @@ -5451,7 +5452,7 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb, if (sk->sk_shutdown & RCV_SHUTDOWN) { if (TCP_SKB_CB(skb)->end_seq != TCP_SKB_CB(skb)->seq && after(TCP_SKB_CB(skb)->end_seq - th->fin, tp->rcv_nxt)) { - NET_INC_STATS_BH(LINUX_MIB_TCPABORTONDATA); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPABORTONDATA); tcp_reset(sk); return 1; } diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index e876312b950a..29adc668ad51 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -366,7 +366,7 @@ void tcp_v4_err(struct sk_buff *skb, u32 info) * servers this needs to be solved differently. */ if (sock_owned_by_user(sk)) - NET_INC_STATS_BH(LINUX_MIB_LOCKDROPPEDICMPS); + NET_INC_STATS_BH(net, LINUX_MIB_LOCKDROPPEDICMPS); if (sk->sk_state == TCP_CLOSE) goto out; @@ -375,7 +375,7 @@ void tcp_v4_err(struct sk_buff *skb, u32 info) seq = ntohl(th->seq); if (sk->sk_state != TCP_LISTEN && !between(seq, tp->snd_una, tp->snd_nxt)) { - NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS); + NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS); goto out; } @@ -422,7 +422,7 @@ void tcp_v4_err(struct sk_buff *skb, u32 info) BUG_TRAP(!req->sk); if (seq != tcp_rsk(req)->snt_isn) { - NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS); + NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS); goto out; } @@ -1251,7 +1251,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb) if (get_seconds() < peer->tcp_ts_stamp + TCP_PAWS_MSL && (s32)(peer->tcp_ts - req->ts_recent) > TCP_PAWS_WINDOW) { - NET_INC_STATS_BH(LINUX_MIB_PAWSPASSIVEREJECTED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_PAWSPASSIVEREJECTED); goto drop_and_release; } } @@ -1365,9 +1365,9 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb, return newsk; exit_overflow: - NET_INC_STATS_BH(LINUX_MIB_LISTENOVERFLOWS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS); exit: - NET_INC_STATS_BH(LINUX_MIB_LISTENDROPS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS); dst_release(dst); return NULL; } diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index 8b02b1039968..204c42162660 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -244,7 +244,7 @@ kill: } if (paws_reject) - NET_INC_STATS_BH(LINUX_MIB_PAWSESTABREJECTED); + NET_INC_STATS_BH(twsk_net(tw), LINUX_MIB_PAWSESTABREJECTED); if (!th->rst) { /* In this case we must reset the TIMEWAIT timer. @@ -611,7 +611,7 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb, if (!(flg & TCP_FLAG_RST)) req->rsk_ops->send_ack(skb, req); if (paws_reject) - NET_INC_STATS_BH(LINUX_MIB_PAWSESTABREJECTED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_PAWSESTABREJECTED); return NULL; } @@ -695,7 +695,7 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb, } embryonic_reset: - NET_INC_STATS_BH(LINUX_MIB_EMBRYONICRSTS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_EMBRYONICRSTS); if (!(flg & TCP_FLAG_RST)) req->rsk_ops->send_reset(sk, skb); diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 176f0702b8ac..36a19707f67f 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1995,7 +1995,7 @@ void tcp_xmit_retransmit_queue(struct sock *sk) mib_idx = LINUX_MIB_TCPFASTRETRANS; else mib_idx = LINUX_MIB_TCPSLOWSTARTRETRANS; - NET_INC_STATS_BH(mib_idx); + NET_INC_STATS_BH(sock_net(sk), mib_idx); if (skb == tcp_write_queue_head(sk)) inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, @@ -2065,7 +2065,7 @@ void tcp_xmit_retransmit_queue(struct sock *sk) inet_csk(sk)->icsk_rto, TCP_RTO_MAX); - NET_INC_STATS_BH(LINUX_MIB_TCPFORWARDRETRANS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPFORWARDRETRANS); } } diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c index 6a480d1fd8f6..328e0cf42b3c 100644 --- a/net/ipv4/tcp_timer.c +++ b/net/ipv4/tcp_timer.c @@ -48,7 +48,7 @@ static void tcp_write_err(struct sock *sk) sk->sk_error_report(sk); tcp_done(sk); - NET_INC_STATS_BH(LINUX_MIB_TCPABORTONTIMEOUT); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPABORTONTIMEOUT); } /* Do not allow orphaned sockets to eat all our resources. @@ -89,7 +89,7 @@ static int tcp_out_of_resources(struct sock *sk, int do_reset) if (do_reset) tcp_send_active_reset(sk, GFP_ATOMIC); tcp_done(sk); - NET_INC_STATS_BH(LINUX_MIB_TCPABORTONMEMORY); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPABORTONMEMORY); return 1; } return 0; @@ -179,7 +179,7 @@ static void tcp_delack_timer(unsigned long data) if (sock_owned_by_user(sk)) { /* Try again later. */ icsk->icsk_ack.blocked = 1; - NET_INC_STATS_BH(LINUX_MIB_DELAYEDACKLOCKED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_DELAYEDACKLOCKED); sk_reset_timer(sk, &icsk->icsk_delack_timer, jiffies + TCP_DELACK_MIN); goto out_unlock; } @@ -198,7 +198,7 @@ static void tcp_delack_timer(unsigned long data) if (!skb_queue_empty(&tp->ucopy.prequeue)) { struct sk_buff *skb; - NET_INC_STATS_BH(LINUX_MIB_TCPSCHEDULERFAILED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPSCHEDULERFAILED); while ((skb = __skb_dequeue(&tp->ucopy.prequeue)) != NULL) sk->sk_backlog_rcv(sk, skb); @@ -218,7 +218,7 @@ static void tcp_delack_timer(unsigned long data) icsk->icsk_ack.ato = TCP_ATO_MIN; } tcp_send_ack(sk); - NET_INC_STATS_BH(LINUX_MIB_DELAYEDACKS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_DELAYEDACKS); } TCP_CHECK_TIMER(sk); @@ -346,7 +346,7 @@ static void tcp_retransmit_timer(struct sock *sk) } else { mib_idx = LINUX_MIB_TCPTIMEOUTS; } - NET_INC_STATS_BH(mib_idx); + NET_INC_STATS_BH(sock_net(sk), mib_idx); } if (tcp_use_frto(sk)) { diff --git a/net/ipv6/inet6_hashtables.c b/net/ipv6/inet6_hashtables.c index a9cc8ab33a49..00a8a5f9380c 100644 --- a/net/ipv6/inet6_hashtables.c +++ b/net/ipv6/inet6_hashtables.c @@ -210,11 +210,11 @@ unique: if (twp != NULL) { *twp = tw; - NET_INC_STATS_BH(LINUX_MIB_TIMEWAITRECYCLED); + NET_INC_STATS_BH(twsk_net(tw), LINUX_MIB_TIMEWAITRECYCLED); } else if (tw != NULL) { /* Silly. Should hash-dance instead... */ inet_twsk_deschedule(tw, death_row); - NET_INC_STATS_BH(LINUX_MIB_TIMEWAITRECYCLED); + NET_INC_STATS_BH(twsk_net(tw), LINUX_MIB_TIMEWAITRECYCLED); inet_twsk_put(tw); } diff --git a/net/ipv6/syncookies.c b/net/ipv6/syncookies.c index 3ecc1157994e..6a68eeb7bbf8 100644 --- a/net/ipv6/syncookies.c +++ b/net/ipv6/syncookies.c @@ -137,7 +137,7 @@ __u32 cookie_v6_init_sequence(struct sock *sk, struct sk_buff *skb, __u16 *mssp) ; *mssp = msstab[mssind] + 1; - NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESSENT); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESSENT); return secure_tcp_syn_cookie(&iph->saddr, &iph->daddr, th->source, th->dest, ntohl(th->seq), @@ -177,11 +177,11 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb) if (time_after(jiffies, tp->last_synq_overflow + TCP_TIMEOUT_INIT) || (mss = cookie_check(skb, cookie)) == 0) { - NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESFAILED); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESFAILED); goto out; } - NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESRECV); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESRECV); /* check for timestamp cookie support */ memset(&tcp_opt, 0, sizeof(tcp_opt)); diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index d58b83ac06fb..ca5b93a5c02a 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -340,7 +340,7 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, bh_lock_sock(sk); if (sock_owned_by_user(sk)) - NET_INC_STATS_BH(LINUX_MIB_LOCKDROPPEDICMPS); + NET_INC_STATS_BH(net, LINUX_MIB_LOCKDROPPEDICMPS); if (sk->sk_state == TCP_CLOSE) goto out; @@ -349,7 +349,7 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, seq = ntohl(th->seq); if (sk->sk_state != TCP_LISTEN && !between(seq, tp->snd_una, tp->snd_nxt)) { - NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS); + NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS); goto out; } @@ -424,7 +424,7 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt, BUG_TRAP(req->sk == NULL); if (seq != tcp_rsk(req)->snt_isn) { - NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS); + NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS); goto out; } @@ -1449,9 +1449,9 @@ static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb, return newsk; out_overflow: - NET_INC_STATS_BH(LINUX_MIB_LISTENOVERFLOWS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS); out: - NET_INC_STATS_BH(LINUX_MIB_LISTENDROPS); + NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS); if (opt && opt != np->opt) sock_kfree_s(sk, opt, opt->tot_len); dst_release(dst); diff --git a/net/sctp/input.c b/net/sctp/input.c index ed8834e7f143..5ed93c05c23e 100644 --- a/net/sctp/input.c +++ b/net/sctp/input.c @@ -486,7 +486,7 @@ struct sock *sctp_err_lookup(int family, struct sk_buff *skb, * servers this needs to be solved differently. */ if (sock_owned_by_user(sk)) - NET_INC_STATS_BH(LINUX_MIB_LOCKDROPPEDICMPS); + NET_INC_STATS_BH(&init_net, LINUX_MIB_LOCKDROPPEDICMPS); *app = asoc; *tpp = transport; -- cgit v1.2.2 From 6f67c817fcfd94f5ca0f14b114b7fa25c0210c8b Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:31:39 -0700 Subject: mib: add net to NET_INC_STATS_USER Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/tcp.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 9e0e45c37806..dec0b8681119 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -1153,7 +1153,7 @@ static void tcp_prequeue_process(struct sock *sk) struct sk_buff *skb; struct tcp_sock *tp = tcp_sk(sk); - NET_INC_STATS_USER(LINUX_MIB_TCPPREQUEUED); + NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPPREQUEUED); /* RX process wants to run with disabled BHs, though it is not * necessary */ @@ -1793,13 +1793,13 @@ void tcp_close(struct sock *sk, long timeout) */ if (data_was_unread) { /* Unread data was tossed, zap the connection. */ - NET_INC_STATS_USER(LINUX_MIB_TCPABORTONCLOSE); + NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPABORTONCLOSE); tcp_set_state(sk, TCP_CLOSE); tcp_send_active_reset(sk, GFP_KERNEL); } else if (sock_flag(sk, SOCK_LINGER) && !sk->sk_lingertime) { /* Check zero linger _after_ checking for unread data. */ sk->sk_prot->disconnect(sk, 0); - NET_INC_STATS_USER(LINUX_MIB_TCPABORTONDATA); + NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPABORTONDATA); } else if (tcp_close_state(sk)) { /* We FIN if the application ate all the data before * zapping the connection. -- cgit v1.2.2 From f2bf415cfed703de5ba94d25cdb160920c01fb00 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:32:25 -0700 Subject: mib: add net to NET_ADD_STATS_BH This one is tricky. The thing is that this macro is only used when killing tw buckets, but since this killer is promiscuous wrt to which net each particular tw belongs to, I have to use it only when NET_NS is off. When the net namespaces are on, I use the INET_INC_STATS_BH for each bucket. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/inet_timewait_sock.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv4/inet_timewait_sock.c b/net/ipv4/inet_timewait_sock.c index 06006a5ac8b9..75c2def8f9a0 100644 --- a/net/ipv4/inet_timewait_sock.c +++ b/net/ipv4/inet_timewait_sock.c @@ -160,6 +160,9 @@ rescan: __inet_twsk_del_dead_node(tw); spin_unlock(&twdr->death_lock); __inet_twsk_kill(tw, twdr->hashinfo); +#ifdef CONFIG_NET_NS + NET_INC_STATS_BH(twsk_net(tw), LINUX_MIB_TIMEWAITED); +#endif inet_twsk_put(tw); killed++; spin_lock(&twdr->death_lock); @@ -178,8 +181,9 @@ rescan: } twdr->tw_count -= killed; - NET_ADD_STATS_BH(LINUX_MIB_TIMEWAITED, killed); - +#ifndef CONFIG_NET_NS + NET_ADD_STATS_BH(&init_net, LINUX_MIB_TIMEWAITED, killed); +#endif return ret; } @@ -372,6 +376,9 @@ void inet_twdr_twcal_tick(unsigned long data) &twdr->twcal_row[slot]) { __inet_twsk_del_dead_node(tw); __inet_twsk_kill(tw, twdr->hashinfo); +#ifdef CONFIG_NET_NS + NET_INC_STATS_BH(twsk_net(tw), LINUX_MIB_TIMEWAITKILLED); +#endif inet_twsk_put(tw); killed++; } @@ -395,7 +402,9 @@ void inet_twdr_twcal_tick(unsigned long data) out: if ((twdr->tw_count -= killed) == 0) del_timer(&twdr->tw_timer); - NET_ADD_STATS_BH(LINUX_MIB_TIMEWAITKILLED, killed); +#ifndef CONFIG_NET_NS + NET_ADD_STATS_BH(&init_net, LINUX_MIB_TIMEWAITKILLED, killed); +#endif spin_unlock(&twdr->death_lock); } -- cgit v1.2.2 From ed88098e25d77bef3b2ad8c9d8e2ebf454d9ccbf Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 16 Jul 2008 20:32:45 -0700 Subject: mib: add net to NET_ADD_STATS_USER Done with NET_XXX_STATS macros :) To be continued... Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/tcp.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index dec0b8681119..eec8cf7c0247 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -1475,7 +1475,7 @@ int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg, /* __ Restore normal policy in scheduler __ */ if ((chunk = len - tp->ucopy.len) != 0) { - NET_ADD_STATS_USER(LINUX_MIB_TCPDIRECTCOPYFROMBACKLOG, chunk); + NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMBACKLOG, chunk); len -= chunk; copied += chunk; } @@ -1486,7 +1486,7 @@ do_prequeue: tcp_prequeue_process(sk); if ((chunk = len - tp->ucopy.len) != 0) { - NET_ADD_STATS_USER(LINUX_MIB_TCPDIRECTCOPYFROMPREQUEUE, chunk); + NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMPREQUEUE, chunk); len -= chunk; copied += chunk; } @@ -1601,7 +1601,7 @@ skip_copy: tcp_prequeue_process(sk); if (copied > 0 && (chunk = len - tp->ucopy.len) != 0) { - NET_ADD_STATS_USER(LINUX_MIB_TCPDIRECTCOPYFROMPREQUEUE, chunk); + NET_ADD_STATS_USER(sock_net(sk), LINUX_MIB_TCPDIRECTCOPYFROMPREQUEUE, chunk); len -= chunk; copied += chunk; } -- cgit v1.2.2 From 9a6d276e85aa3d8f308fc5e8de6892daeb60ae5f Mon Sep 17 00:00:00 2001 From: Neil Horman Date: Wed, 16 Jul 2008 20:50:49 -0700 Subject: core: add stat to track unresolved discards in neighbor cache in __neigh_event_send, if we have a neighbour entry which is in NUD_INCOMPLETE state, we enqueue any outbound frames to that neighbour to the neighbours arp_queue, which is default capped to a length of 3 skbs. If that queue exceeds its set length, it will drop an skb on the queue to enqueue the newly arrived skb. This results in a drop for which we have no statistics incremented. This patch adds an unresolved_discards stat to /proc/net/stat/ndisc_cache to track these lost frames. Signed-off-by: Neil Horman Signed-off-by: David S. Miller --- net/core/neighbour.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/core/neighbour.c b/net/core/neighbour.c index 65f01f71b3f3..f62c8af85d38 100644 --- a/net/core/neighbour.c +++ b/net/core/neighbour.c @@ -930,6 +930,7 @@ int __neigh_event_send(struct neighbour *neigh, struct sk_buff *skb) buff = neigh->arp_queue.next; __skb_unlink(buff, &neigh->arp_queue); kfree_skb(buff); + NEIGH_CACHE_STAT_INC(neigh->tbl, unres_discards); } __skb_queue_tail(&neigh->arp_queue, skb); } @@ -2462,12 +2463,12 @@ static int neigh_stat_seq_show(struct seq_file *seq, void *v) struct neigh_statistics *st = v; if (v == SEQ_START_TOKEN) { - seq_printf(seq, "entries allocs destroys hash_grows lookups hits res_failed rcv_probes_mcast rcv_probes_ucast periodic_gc_runs forced_gc_runs\n"); + seq_printf(seq, "entries allocs destroys hash_grows lookups hits res_failed rcv_probes_mcast rcv_probes_ucast periodic_gc_runs forced_gc_runs unresolved_discards\n"); return 0; } seq_printf(seq, "%08x %08lx %08lx %08lx %08lx %08lx %08lx " - "%08lx %08lx %08lx %08lx\n", + "%08lx %08lx %08lx %08lx %08lx\n", atomic_read(&tbl->entries), st->allocs, @@ -2483,7 +2484,8 @@ static int neigh_stat_seq_show(struct seq_file *seq, void *v) st->rcv_probes_ucast, st->periodic_gc_runs, - st->forced_gc_runs + st->forced_gc_runs, + st->unres_discards ); return 0; -- cgit v1.2.2 From 51ce7ec92187b22ab1598987bb5f9776f6e0ebad Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Wed, 16 Jul 2008 20:51:47 -0700 Subject: garp: retry sending JoinIn messages after allocation failures Increase reliability by retrying to send JoinIn messages after memory allocation failures on each TRANSMIT_PDU event until it succeeds. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/802/garp.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/802/garp.c b/net/802/garp.c index 3b78f7b74fd4..1dcb0660c49d 100644 --- a/net/802/garp.c +++ b/net/802/garp.c @@ -323,7 +323,10 @@ static void garp_attr_event(struct garp_applicant *app, case GARP_ACTION_NONE: break; case GARP_ACTION_S_JOIN_IN: - garp_pdu_append_attr(app, attr, GARP_JOIN_IN); + /* When appending the attribute fails, don't update state in + * order to retry on next TRANSMIT_PDU event. */ + if (garp_pdu_append_attr(app, attr, GARP_JOIN_IN) < 0) + return; break; case GARP_ACTION_S_LEAVE_EMPTY: garp_pdu_append_attr(app, attr, GARP_LEAVE_EMPTY); -- cgit v1.2.2 From e8a0464cc950972824e2e128028ae3db666ec1ed Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 00:34:19 -0700 Subject: netdev: Allocate multiple queues for TX. alloc_netdev_mq() now allocates an array of netdev_queue structures for TX, based upon the queue_count argument. Furthermore, all accesses to the TX queues are now vectored through the netdev_get_tx_queue() and netdev_for_each_tx_queue() interfaces. This makes it easy to grep the tree for all things that want to get to a TX queue of a net device. Problem spots which are not really multiqueue aware yet, and only work with one queue, can easily be spotted by grepping for all netdev_get_tx_queue() calls that pass in a zero index. Signed-off-by: David S. Miller --- net/8021q/vlan_dev.c | 10 +-- net/core/dev.c | 40 ++++++++--- net/core/rtnetlink.c | 2 +- net/mac80211/main.c | 4 +- net/mac80211/wme.c | 12 ++-- net/netrom/af_netrom.c | 6 +- net/rose/af_rose.c | 6 +- net/sched/cls_api.c | 4 +- net/sched/sch_api.c | 32 ++++++--- net/sched/sch_generic.c | 178 +++++++++++++++++++++++++++++++++++------------- net/sched/sch_teql.c | 21 ++++-- 11 files changed, 223 insertions(+), 92 deletions(-) (limited to 'net') diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 6b985f23fd9f..f42bc2b26b85 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -570,16 +570,18 @@ static void vlan_dev_set_rx_mode(struct net_device *vlan_dev) */ static struct lock_class_key vlan_netdev_xmit_lock_key; -static void vlan_dev_set_lockdep_one(struct netdev_queue *txq, - int subclass) +static void vlan_dev_set_lockdep_one(struct net_device *dev, + struct netdev_queue *txq, + void *_subclass) { lockdep_set_class_and_subclass(&txq->_xmit_lock, - &vlan_netdev_xmit_lock_key, subclass); + &vlan_netdev_xmit_lock_key, + *(int *)_subclass); } static void vlan_dev_set_lockdep_class(struct net_device *dev, int subclass) { - vlan_dev_set_lockdep_one(&dev->tx_queue, subclass); + netdev_for_each_tx_queue(dev, vlan_dev_set_lockdep_one, &subclass); } static const struct header_ops vlan_header_ops = { diff --git a/net/core/dev.c b/net/core/dev.c index 9b49f74a9820..69378f250695 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1666,6 +1666,12 @@ out_kfree_skb: * --BLG */ +static struct netdev_queue *dev_pick_tx(struct net_device *dev, + struct sk_buff *skb) +{ + return netdev_get_tx_queue(dev, 0); +} + int dev_queue_xmit(struct sk_buff *skb) { struct net_device *dev = skb->dev; @@ -1702,7 +1708,7 @@ int dev_queue_xmit(struct sk_buff *skb) } gso: - txq = &dev->tx_queue; + txq = dev_pick_tx(dev, skb); spin_lock_prefetch(&txq->lock); /* Disable soft irqs for various locks below. Also @@ -3788,8 +3794,9 @@ static void rollback_registered(struct net_device *dev) dev_put(dev); } -static void __netdev_init_queue_locks_one(struct netdev_queue *dev_queue, - struct net_device *dev) +static void __netdev_init_queue_locks_one(struct net_device *dev, + struct netdev_queue *dev_queue, + void *_unused) { spin_lock_init(&dev_queue->_xmit_lock); netdev_set_lockdep_class(&dev_queue->_xmit_lock, dev->type); @@ -3798,8 +3805,8 @@ static void __netdev_init_queue_locks_one(struct netdev_queue *dev_queue, static void netdev_init_queue_locks(struct net_device *dev) { - __netdev_init_queue_locks_one(&dev->tx_queue, dev); - __netdev_init_queue_locks_one(&dev->rx_queue, dev); + netdev_for_each_tx_queue(dev, __netdev_init_queue_locks_one, NULL); + __netdev_init_queue_locks_one(dev, &dev->rx_queue, NULL); } /** @@ -4119,7 +4126,8 @@ static struct net_device_stats *internal_stats(struct net_device *dev) } static void netdev_init_one_queue(struct net_device *dev, - struct netdev_queue *queue) + struct netdev_queue *queue, + void *_unused) { spin_lock_init(&queue->lock); queue->dev = dev; @@ -4127,8 +4135,8 @@ static void netdev_init_one_queue(struct net_device *dev, static void netdev_init_queues(struct net_device *dev) { - netdev_init_one_queue(dev, &dev->rx_queue); - netdev_init_one_queue(dev, &dev->tx_queue); + netdev_init_one_queue(dev, &dev->rx_queue, NULL); + netdev_for_each_tx_queue(dev, netdev_init_one_queue, NULL); } /** @@ -4145,9 +4153,10 @@ static void netdev_init_queues(struct net_device *dev) struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, void (*setup)(struct net_device *), unsigned int queue_count) { - void *p; + struct netdev_queue *tx; struct net_device *dev; int alloc_size; + void *p; BUG_ON(strlen(name) >= sizeof(dev->name)); @@ -4167,11 +4176,22 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, return NULL; } + tx = kzalloc(sizeof(struct netdev_queue) * queue_count, GFP_KERNEL); + if (!tx) { + printk(KERN_ERR "alloc_netdev: Unable to allocate " + "tx qdiscs.\n"); + kfree(p); + return NULL; + } + dev = (struct net_device *) (((long)p + NETDEV_ALIGN_CONST) & ~NETDEV_ALIGN_CONST); dev->padded = (char *)dev - (char *)p; dev_net_set(dev, &init_net); + dev->_tx = tx; + dev->num_tx_queues = queue_count; + if (sizeof_priv) { dev->priv = ((char *)dev + ((sizeof(struct net_device) + @@ -4205,6 +4225,8 @@ void free_netdev(struct net_device *dev) { release_net(dev_net(dev)); + kfree(dev->_tx); + /* Compatibility with error handling in drivers */ if (dev->reg_state == NETREG_UNINITIALIZED) { kfree((char *)dev - dev->padded); diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 8ef9f1db610e..71edb8b36341 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -636,7 +636,7 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev, if (dev->master) NLA_PUT_U32(skb, IFLA_MASTER, dev->master->ifindex); - txq = &dev->tx_queue; + txq = netdev_get_tx_queue(dev, 0); if (txq->qdisc_sleeping) NLA_PUT_STRING(skb, IFLA_QDISC, txq->qdisc_sleeping->ops->id); diff --git a/net/mac80211/main.c b/net/mac80211/main.c index af0056e7e5b3..b486e634f4fe 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -621,7 +621,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) /* ensure that TX flow won't interrupt us * until the end of the call to requeue function */ - txq = &local->mdev->tx_queue; + txq = netdev_get_tx_queue(local->mdev, 0); spin_lock_bh(&txq->lock); /* create a new queue for this aggregation */ @@ -862,7 +862,7 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) /* avoid ordering issues: we are the only one that can modify * the content of the qdiscs */ - txq = &local->mdev->tx_queue; + txq = netdev_get_tx_queue(local->mdev, 0); spin_lock_bh(&txq->lock); /* remove the queue for this aggregation */ ieee80211_ht_agg_queue_remove(local, sta, tid, 1); diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 6ae43a3c7726..f014cd38c2d0 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -574,7 +574,7 @@ static struct Qdisc_ops wme_qdisc_ops __read_mostly = void ieee80211_install_qdisc(struct net_device *dev) { - struct netdev_queue *txq = &dev->tx_queue; + struct netdev_queue *txq = netdev_get_tx_queue(dev, 0); struct Qdisc *qdisc; qdisc = qdisc_create_dflt(dev, txq, @@ -596,7 +596,7 @@ void ieee80211_install_qdisc(struct net_device *dev) int ieee80211_qdisc_installed(struct net_device *dev) { - struct netdev_queue *txq = &dev->tx_queue; + struct netdev_queue *txq = netdev_get_tx_queue(dev, 0); return txq->qdisc_sleeping->ops == &wme_qdisc_ops; } @@ -617,7 +617,7 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, struct sta_info *sta, u16 tid) { int i; - struct netdev_queue *txq = &local->mdev->tx_queue; + struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, 0); struct ieee80211_sched_data *q = qdisc_priv(txq->qdisc_sleeping); DECLARE_MAC_BUF(mac); @@ -652,14 +652,14 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, } /** - * the caller needs to hold local->mdev->tx_queue.lock + * the caller needs to hold netdev_get_tx_queue(local->mdev, X)->lock */ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, struct sta_info *sta, u16 tid, u8 requeue) { struct ieee80211_hw *hw = &local->hw; - struct netdev_queue *txq = &local->mdev->tx_queue; + struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, 0); struct ieee80211_sched_data *q = qdisc_priv(txq->qdisc_sleeping); int agg_queue = sta->tid_to_tx_q[tid]; @@ -676,7 +676,7 @@ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, void ieee80211_requeue(struct ieee80211_local *local, int queue) { - struct netdev_queue *txq = &local->mdev->tx_queue; + struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, 0); struct Qdisc *root_qd = txq->qdisc_sleeping; struct ieee80211_sched_data *q = qdisc_priv(root_qd); struct Qdisc *qdisc = q->queues[queue]; diff --git a/net/netrom/af_netrom.c b/net/netrom/af_netrom.c index 819afc449e1e..d41be0d66eb0 100644 --- a/net/netrom/af_netrom.c +++ b/net/netrom/af_netrom.c @@ -74,14 +74,16 @@ static const struct proto_ops nr_proto_ops; */ static struct lock_class_key nr_netdev_xmit_lock_key; -static void nr_set_lockdep_one(struct netdev_queue *txq) +static void nr_set_lockdep_one(struct net_device *dev, + struct netdev_queue *txq, + void *_unused) { lockdep_set_class(&txq->_xmit_lock, &nr_netdev_xmit_lock_key); } static void nr_set_lockdep_key(struct net_device *dev) { - nr_set_lockdep_one(&dev->tx_queue); + netdev_for_each_tx_queue(dev, nr_set_lockdep_one, NULL); } /* diff --git a/net/rose/af_rose.c b/net/rose/af_rose.c index 7dbbc0891623..f3a691f34909 100644 --- a/net/rose/af_rose.c +++ b/net/rose/af_rose.c @@ -75,14 +75,16 @@ ax25_address rose_callsign; */ static struct lock_class_key rose_netdev_xmit_lock_key; -static void rose_set_lockdep_one(struct netdev_queue *txq) +static void rose_set_lockdep_one(struct net_device *dev, + struct netdev_queue *txq, + void *_unused) { lockdep_set_class(&txq->_xmit_lock, &rose_netdev_xmit_lock_key); } static void rose_set_lockdep_key(struct net_device *dev) { - rose_set_lockdep_one(&dev->tx_queue); + netdev_for_each_tx_queue(dev, rose_set_lockdep_one, NULL); } /* diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c index b483bbea6118..d0b0a9b14394 100644 --- a/net/sched/cls_api.c +++ b/net/sched/cls_api.c @@ -166,7 +166,7 @@ replay: /* Find qdisc */ if (!parent) { - struct netdev_queue *dev_queue = &dev->tx_queue; + struct netdev_queue *dev_queue = netdev_get_tx_queue(dev, 0); q = dev_queue->qdisc_sleeping; parent = q->handle; } else { @@ -410,7 +410,7 @@ static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb) if ((dev = dev_get_by_index(&init_net, tcm->tcm_ifindex)) == NULL) return skb->len; - dev_queue = &dev->tx_queue; + dev_queue = netdev_get_tx_queue(dev, 0); if (!tcm->tcm_parent) q = dev_queue->qdisc_sleeping; else diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 95873f8dd37c..830ccc544a15 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -183,9 +183,8 @@ EXPORT_SYMBOL(unregister_qdisc); (root qdisc, all its children, children of children etc.) */ -struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) +static struct Qdisc *__qdisc_lookup(struct netdev_queue *dev_queue, u32 handle) { - struct netdev_queue *dev_queue = &dev->tx_queue; struct Qdisc *q; list_for_each_entry(q, &dev_queue->qdisc_list, list) { @@ -195,6 +194,19 @@ struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) return NULL; } +struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) +{ + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + struct Qdisc *q = __qdisc_lookup(txq, handle); + if (q) + return q; + } + return NULL; +} + static struct Qdisc *qdisc_leaf(struct Qdisc *p, u32 classid) { unsigned long cl; @@ -462,7 +474,7 @@ dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) } } else { - dev_queue = &dev->tx_queue; + dev_queue = netdev_get_tx_queue(dev, 0); oqdisc = dev_queue->qdisc_sleeping; /* Prune old scheduler */ @@ -742,7 +754,8 @@ static int tc_get_qdisc(struct sk_buff *skb, struct nlmsghdr *n, void *arg) q = dev->rx_queue.qdisc; } } else { - struct netdev_queue *dev_queue = &dev->tx_queue; + struct netdev_queue *dev_queue; + dev_queue = netdev_get_tx_queue(dev, 0); q = dev_queue->qdisc_sleeping; } if (!q) @@ -817,7 +830,8 @@ replay: q = dev->rx_queue.qdisc; } } else { - struct netdev_queue *dev_queue = &dev->tx_queue; + struct netdev_queue *dev_queue; + dev_queue = netdev_get_tx_queue(dev, 0); q = dev_queue->qdisc_sleeping; } @@ -899,7 +913,7 @@ create_n_graft: tcm->tcm_parent, tcm->tcm_parent, tca, &err); else - q = qdisc_create(dev, &dev->tx_queue, + q = qdisc_create(dev, netdev_get_tx_queue(dev, 0), tcm->tcm_parent, tcm->tcm_handle, tca, &err); if (q == NULL) { @@ -1025,7 +1039,7 @@ static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb) if (idx > s_idx) s_q_idx = 0; q_idx = 0; - dev_queue = &dev->tx_queue; + dev_queue = netdev_get_tx_queue(dev, 0); list_for_each_entry(q, &dev_queue->qdisc_list, list) { if (q_idx < s_q_idx) { q_idx++; @@ -1098,7 +1112,7 @@ static int tc_ctl_tclass(struct sk_buff *skb, struct nlmsghdr *n, void *arg) /* Step 1. Determine qdisc handle X:0 */ - dev_queue = &dev->tx_queue; + dev_queue = netdev_get_tx_queue(dev, 0); if (pid != TC_H_ROOT) { u32 qid1 = TC_H_MAJ(pid); @@ -1275,7 +1289,7 @@ static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) s_t = cb->args[0]; t = 0; - dev_queue = &dev->tx_queue; + dev_queue = netdev_get_tx_queue(dev, 0); list_for_each_entry(q, &dev_queue->qdisc_list, list) { if (t < s_t || !q->ops->cl_ops || (tcm->tcm_parent && diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 243de935b182..4e2b865cbba0 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -40,20 +40,30 @@ */ void qdisc_lock_tree(struct net_device *dev) - __acquires(dev->tx_queue.lock) __acquires(dev->rx_queue.lock) { - spin_lock_bh(&dev->tx_queue.lock); + unsigned int i; + + local_bh_disable(); + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + spin_lock(&txq->lock); + } spin_lock(&dev->rx_queue.lock); } EXPORT_SYMBOL(qdisc_lock_tree); void qdisc_unlock_tree(struct net_device *dev) __releases(dev->rx_queue.lock) - __releases(dev->tx_queue.lock) { + unsigned int i; + spin_unlock(&dev->rx_queue.lock); - spin_unlock_bh(&dev->tx_queue.lock); + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + spin_unlock(&txq->lock); + } + local_bh_enable(); } EXPORT_SYMBOL(qdisc_unlock_tree); @@ -212,22 +222,37 @@ void __qdisc_run(struct netdev_queue *txq) static void dev_watchdog(unsigned long arg) { struct net_device *dev = (struct net_device *)arg; - struct netdev_queue *txq = &dev->tx_queue; netif_tx_lock(dev); - if (txq->qdisc != &noop_qdisc) { + if (!qdisc_tx_is_noop(dev)) { if (netif_device_present(dev) && netif_running(dev) && netif_carrier_ok(dev)) { - if (netif_queue_stopped(dev) && - time_after(jiffies, dev->trans_start + dev->watchdog_timeo)) { + int some_queue_stopped = 0; + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq; + + txq = netdev_get_tx_queue(dev, i); + if (netif_tx_queue_stopped(txq)) { + some_queue_stopped = 1; + break; + } + } - printk(KERN_INFO "NETDEV WATCHDOG: %s: transmit timed out\n", + if (some_queue_stopped && + time_after(jiffies, (dev->trans_start + + dev->watchdog_timeo))) { + printk(KERN_INFO "NETDEV WATCHDOG: %s: " + "transmit timed out\n", dev->name); dev->tx_timeout(dev); WARN_ON_ONCE(1); } - if (!mod_timer(&dev->watchdog_timer, round_jiffies(jiffies + dev->watchdog_timeo))) + if (!mod_timer(&dev->watchdog_timer, + round_jiffies(jiffies + + dev->watchdog_timeo))) dev_hold(dev); } } @@ -542,9 +567,55 @@ void qdisc_destroy(struct Qdisc *qdisc) } EXPORT_SYMBOL(qdisc_destroy); +static bool dev_all_qdisc_sleeping_noop(struct net_device *dev) +{ + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + + if (txq->qdisc_sleeping != &noop_qdisc) + return false; + } + return true; +} + +static void attach_one_default_qdisc(struct net_device *dev, + struct netdev_queue *dev_queue, + void *_unused) +{ + struct Qdisc *qdisc; + + if (dev->tx_queue_len) { + qdisc = qdisc_create_dflt(dev, dev_queue, + &pfifo_fast_ops, TC_H_ROOT); + if (!qdisc) { + printk(KERN_INFO "%s: activation failed\n", dev->name); + return; + } + list_add_tail(&qdisc->list, &dev_queue->qdisc_list); + } else { + qdisc = &noqueue_qdisc; + } + dev_queue->qdisc_sleeping = qdisc; +} + +static void transition_one_qdisc(struct net_device *dev, + struct netdev_queue *dev_queue, + void *_need_watchdog) +{ + int *need_watchdog_p = _need_watchdog; + + spin_lock_bh(&dev_queue->lock); + rcu_assign_pointer(dev_queue->qdisc, dev_queue->qdisc_sleeping); + if (dev_queue->qdisc != &noqueue_qdisc) + *need_watchdog_p = 1; + spin_unlock_bh(&dev_queue->lock); +} + void dev_activate(struct net_device *dev) { - struct netdev_queue *txq = &dev->tx_queue; + int need_watchdog; /* No queueing discipline is attached to device; create default one i.e. pfifo_fast for devices, @@ -552,39 +623,27 @@ void dev_activate(struct net_device *dev) virtual interfaces */ - if (txq->qdisc_sleeping == &noop_qdisc) { - struct Qdisc *qdisc; - if (dev->tx_queue_len) { - qdisc = qdisc_create_dflt(dev, txq, - &pfifo_fast_ops, - TC_H_ROOT); - if (qdisc == NULL) { - printk(KERN_INFO "%s: activation failed\n", dev->name); - return; - } - list_add_tail(&qdisc->list, &txq->qdisc_list); - } else { - qdisc = &noqueue_qdisc; - } - txq->qdisc_sleeping = qdisc; - } + if (dev_all_qdisc_sleeping_noop(dev)) + netdev_for_each_tx_queue(dev, attach_one_default_qdisc, NULL); if (!netif_carrier_ok(dev)) /* Delay activation until next carrier-on event */ return; - spin_lock_bh(&txq->lock); - rcu_assign_pointer(txq->qdisc, txq->qdisc_sleeping); - if (txq->qdisc != &noqueue_qdisc) { + need_watchdog = 0; + netdev_for_each_tx_queue(dev, transition_one_qdisc, &need_watchdog); + + if (need_watchdog) { dev->trans_start = jiffies; dev_watchdog_up(dev); } - spin_unlock_bh(&txq->lock); } -static void dev_deactivate_queue(struct netdev_queue *dev_queue, - struct Qdisc *qdisc_default) +static void dev_deactivate_queue(struct net_device *dev, + struct netdev_queue *dev_queue, + void *_qdisc_default) { + struct Qdisc *qdisc_default = _qdisc_default; struct Qdisc *qdisc; struct sk_buff *skb; @@ -603,12 +662,35 @@ static void dev_deactivate_queue(struct netdev_queue *dev_queue, kfree_skb(skb); } +static bool some_qdisc_is_running(struct net_device *dev, int lock) +{ + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *dev_queue; + int val; + + dev_queue = netdev_get_tx_queue(dev, i); + + if (lock) + spin_lock_bh(&dev_queue->lock); + + val = test_bit(__QUEUE_STATE_QDISC_RUNNING, &dev_queue->state); + + if (lock) + spin_unlock_bh(&dev_queue->lock); + + if (val) + return true; + } + return false; +} + void dev_deactivate(struct net_device *dev) { - struct netdev_queue *dev_queue = &dev->tx_queue; - int running; + bool running; - dev_deactivate_queue(dev_queue, &noop_qdisc); + netdev_for_each_tx_queue(dev, dev_deactivate_queue, &noop_qdisc); dev_watchdog_down(dev); @@ -617,17 +699,14 @@ void dev_deactivate(struct net_device *dev) /* Wait for outstanding qdisc_run calls. */ do { - while (test_bit(__QUEUE_STATE_QDISC_RUNNING, &dev_queue->state)) + while (some_qdisc_is_running(dev, 0)) yield(); /* * Double-check inside queue lock to ensure that all effects * of the queue run are visible when we return. */ - spin_lock_bh(&dev_queue->lock); - running = test_bit(__QUEUE_STATE_QDISC_RUNNING, - &dev_queue->state); - spin_unlock_bh(&dev_queue->lock); + running = some_qdisc_is_running(dev, 1); /* * The running flag should never be set at this point because @@ -642,8 +721,10 @@ void dev_deactivate(struct net_device *dev) static void dev_init_scheduler_queue(struct net_device *dev, struct netdev_queue *dev_queue, - struct Qdisc *qdisc) + void *_qdisc) { + struct Qdisc *qdisc = _qdisc; + dev_queue->qdisc = qdisc; dev_queue->qdisc_sleeping = qdisc; INIT_LIST_HEAD(&dev_queue->qdisc_list); @@ -652,18 +733,19 @@ static void dev_init_scheduler_queue(struct net_device *dev, void dev_init_scheduler(struct net_device *dev) { qdisc_lock_tree(dev); - dev_init_scheduler_queue(dev, &dev->tx_queue, &noop_qdisc); + netdev_for_each_tx_queue(dev, dev_init_scheduler_queue, &noop_qdisc); dev_init_scheduler_queue(dev, &dev->rx_queue, NULL); qdisc_unlock_tree(dev); setup_timer(&dev->watchdog_timer, dev_watchdog, (unsigned long)dev); } -static void dev_shutdown_scheduler_queue(struct net_device *dev, - struct netdev_queue *dev_queue, - struct Qdisc *qdisc_default) +static void shutdown_scheduler_queue(struct net_device *dev, + struct netdev_queue *dev_queue, + void *_qdisc_default) { struct Qdisc *qdisc = dev_queue->qdisc_sleeping; + struct Qdisc *qdisc_default = _qdisc_default; if (qdisc) { dev_queue->qdisc = qdisc_default; @@ -676,8 +758,8 @@ static void dev_shutdown_scheduler_queue(struct net_device *dev, void dev_shutdown(struct net_device *dev) { qdisc_lock_tree(dev); - dev_shutdown_scheduler_queue(dev, &dev->tx_queue, &noop_qdisc); - dev_shutdown_scheduler_queue(dev, &dev->rx_queue, NULL); + netdev_for_each_tx_queue(dev, shutdown_scheduler_queue, &noop_qdisc); + shutdown_scheduler_queue(dev, &dev->rx_queue, NULL); BUG_TRAP(!timer_pending(&dev->watchdog_timer)); qdisc_unlock_tree(dev); } diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index 8ac05981be20..44a2c3451f4d 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -111,7 +111,7 @@ teql_dequeue(struct Qdisc* sch) struct sk_buff *skb; skb = __skb_dequeue(&dat->q); - dat_queue = &dat->m->dev->tx_queue; + dat_queue = netdev_get_tx_queue(dat->m->dev, 0); if (skb == NULL) { struct net_device *m = qdisc_dev(dat_queue->qdisc); if (m) { @@ -155,10 +155,13 @@ teql_destroy(struct Qdisc* sch) if (q == master->slaves) { master->slaves = NEXT_SLAVE(q); if (q == master->slaves) { + struct netdev_queue *txq; + + txq = netdev_get_tx_queue(master->dev, 0); master->slaves = NULL; - spin_lock_bh(&master->dev->tx_queue.lock); - qdisc_reset(master->dev->tx_queue.qdisc); - spin_unlock_bh(&master->dev->tx_queue.lock); + spin_lock_bh(&txq->lock); + qdisc_reset(txq->qdisc); + spin_unlock_bh(&txq->lock); } } skb_queue_purge(&dat->q); @@ -218,7 +221,8 @@ static int teql_qdisc_init(struct Qdisc *sch, struct nlattr *opt) static int __teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device *dev) { - struct teql_sched_data *q = qdisc_priv(dev->tx_queue.qdisc); + struct netdev_queue *dev_queue = netdev_get_tx_queue(dev, 0); + struct teql_sched_data *q = qdisc_priv(dev_queue->qdisc); struct neighbour *mn = skb->dst->neighbour; struct neighbour *n = q->ncache; @@ -254,7 +258,8 @@ __teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device * static inline int teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device *dev) { - if (dev->tx_queue.qdisc == &noop_qdisc) + struct netdev_queue *txq = netdev_get_tx_queue(dev, 0); + if (txq->qdisc == &noop_qdisc) return -ENODEV; if (dev->header_ops == NULL || @@ -285,8 +290,10 @@ restart: do { struct net_device *slave = qdisc_dev(q); + struct netdev_queue *slave_txq; - if (slave->tx_queue.qdisc_sleeping != q) + slave_txq = netdev_get_tx_queue(slave, 0); + if (slave_txq->qdisc_sleeping != q) continue; if (netif_queue_stopped(slave) || __netif_subqueue_stopped(slave, subq) || -- cgit v1.2.2 From 09e83b5d7d1878065e2453239b49b684cd0fe4e5 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 01:52:12 -0700 Subject: netdev: Kill NETIF_F_MULTI_QUEUE. There is no need for a feature bit for something that can be tested by simply checking the TX queue count. Signed-off-by: David S. Miller --- net/mac80211/main.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'net') diff --git a/net/mac80211/main.c b/net/mac80211/main.c index b486e634f4fe..c74607eda1ee 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -1678,9 +1678,6 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) if (!mdev) goto fail_mdev_alloc; - if (ieee80211_num_queues(hw) > 1) - mdev->features |= NETIF_F_MULTI_QUEUE; - mwdev = netdev_priv(mdev); mdev->ieee80211_ptr = mwdev; mwdev->wiphy = local->hw.wiphy; -- cgit v1.2.2 From 1d8ae3fdeb001b8f534a6782c261aba6ec1779f5 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 02:52:19 -0700 Subject: pkt_sched: Remove RR scheduler. This actually fixes a bug added by the RR scheduler changes. The ->bands and ->prio2band parameters were being set outside of the sch_tree_lock() and thus could result in strange behavior and inconsistencies. It might be possible, in the new design (where there will be one qdisc per device TX queue) to allow similar functionality via a TX hash algorithm for RR but I really see no reason to export this aspect of how these multiqueue cards actually implement the scheduling of the the individual DMA TX rings and the single physical MAC/PHY port. Signed-off-by: David S. Miller --- net/sched/sch_prio.c | 136 ++++++--------------------------------------------- 1 file changed, 16 insertions(+), 120 deletions(-) (limited to 'net') diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c index 39157f7bc046..536ca474dc69 100644 --- a/net/sched/sch_prio.c +++ b/net/sched/sch_prio.c @@ -24,11 +24,9 @@ struct prio_sched_data { int bands; - int curband; /* for round-robin */ struct tcf_proto *filter_list; u8 prio2band[TC_PRIO_MAX+1]; struct Qdisc *queues[TCQ_PRIO_BANDS]; - int mq; }; @@ -55,17 +53,14 @@ prio_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr) if (!q->filter_list || err < 0) { if (TC_H_MAJ(band)) band = 0; - band = q->prio2band[band&TC_PRIO_MAX]; - goto out; + return q->queues[q->prio2band[band&TC_PRIO_MAX]]; } band = res.classid; } band = TC_H_MIN(band) - 1; if (band >= q->bands) - band = q->prio2band[0]; -out: - if (q->mq) - skb_set_queue_mapping(skb, band); + return q->queues[q->prio2band[0]]; + return q->queues[band]; } @@ -123,68 +118,23 @@ prio_requeue(struct sk_buff *skb, struct Qdisc* sch) } -static struct sk_buff * -prio_dequeue(struct Qdisc* sch) +static struct sk_buff *prio_dequeue(struct Qdisc* sch) { - struct sk_buff *skb; struct prio_sched_data *q = qdisc_priv(sch); int prio; - struct Qdisc *qdisc; for (prio = 0; prio < q->bands; prio++) { - /* Check if the target subqueue is available before - * pulling an skb. This way we avoid excessive requeues - * for slower queues. - */ - if (!__netif_subqueue_stopped(qdisc_dev(sch), - (q->mq ? prio : 0))) { - qdisc = q->queues[prio]; - skb = qdisc->dequeue(qdisc); - if (skb) { - sch->q.qlen--; - return skb; - } + struct Qdisc *qdisc = q->queues[prio]; + struct sk_buff *skb = qdisc->dequeue(qdisc); + if (skb) { + sch->q.qlen--; + return skb; } } return NULL; } -static struct sk_buff *rr_dequeue(struct Qdisc* sch) -{ - struct sk_buff *skb; - struct prio_sched_data *q = qdisc_priv(sch); - struct Qdisc *qdisc; - int bandcount; - - /* Only take one pass through the queues. If nothing is available, - * return nothing. - */ - for (bandcount = 0; bandcount < q->bands; bandcount++) { - /* Check if the target subqueue is available before - * pulling an skb. This way we avoid excessive requeues - * for slower queues. If the queue is stopped, try the - * next queue. - */ - if (!__netif_subqueue_stopped(qdisc_dev(sch), - (q->mq ? q->curband : 0))) { - qdisc = q->queues[q->curband]; - skb = qdisc->dequeue(qdisc); - if (skb) { - sch->q.qlen--; - q->curband++; - if (q->curband >= q->bands) - q->curband = 0; - return skb; - } - } - q->curband++; - if (q->curband >= q->bands) - q->curband = 0; - } - return NULL; -} - static unsigned int prio_drop(struct Qdisc* sch) { struct prio_sched_data *q = qdisc_priv(sch); @@ -229,45 +179,22 @@ static int prio_tune(struct Qdisc *sch, struct nlattr *opt) { struct prio_sched_data *q = qdisc_priv(sch); struct tc_prio_qopt *qopt; - struct nlattr *tb[TCA_PRIO_MAX + 1]; - int err; int i; - err = nla_parse_nested_compat(tb, TCA_PRIO_MAX, opt, NULL, qopt, - sizeof(*qopt)); - if (err < 0) - return err; - - q->bands = qopt->bands; - /* If we're multiqueue, make sure the number of incoming bands - * matches the number of queues on the device we're associating with. - * If the number of bands requested is zero, then set q->bands to - * dev->egress_subqueue_count. Also, the root qdisc must be the - * only one that is enabled for multiqueue, since it's the only one - * that interacts with the underlying device. - */ - q->mq = nla_get_flag(tb[TCA_PRIO_MQ]); - if (q->mq) { - if (sch->parent != TC_H_ROOT) - return -EINVAL; - if (netif_is_multiqueue(qdisc_dev(sch))) { - if (q->bands == 0) - q->bands = qdisc_dev(sch)->egress_subqueue_count; - else if (q->bands != qdisc_dev(sch)->egress_subqueue_count) - return -EINVAL; - } else - return -EOPNOTSUPP; - } + if (nla_len(opt) < sizeof(*qopt)) + return -EINVAL; + qopt = nla_data(opt); - if (q->bands > TCQ_PRIO_BANDS || q->bands < 2) + if (qopt->bands > TCQ_PRIO_BANDS || qopt->bands < 2) return -EINVAL; for (i=0; i<=TC_PRIO_MAX; i++) { - if (qopt->priomap[i] >= q->bands) + if (qopt->priomap[i] >= qopt->bands) return -EINVAL; } sch_tree_lock(sch); + q->bands = qopt->bands; memcpy(q->prio2band, qopt->priomap, TC_PRIO_MAX+1); for (i=q->bands; imq) { - if (nla_put_flag(skb, TCA_PRIO_MQ) < 0) - goto nla_put_failure; - } nla_nest_compat_end(skb, nest); return skb->len; @@ -509,44 +432,17 @@ static struct Qdisc_ops prio_qdisc_ops __read_mostly = { .owner = THIS_MODULE, }; -static struct Qdisc_ops rr_qdisc_ops __read_mostly = { - .next = NULL, - .cl_ops = &prio_class_ops, - .id = "rr", - .priv_size = sizeof(struct prio_sched_data), - .enqueue = prio_enqueue, - .dequeue = rr_dequeue, - .requeue = prio_requeue, - .drop = prio_drop, - .init = prio_init, - .reset = prio_reset, - .destroy = prio_destroy, - .change = prio_tune, - .dump = prio_dump, - .owner = THIS_MODULE, -}; - static int __init prio_module_init(void) { - int err; - - err = register_qdisc(&prio_qdisc_ops); - if (err < 0) - return err; - err = register_qdisc(&rr_qdisc_ops); - if (err < 0) - unregister_qdisc(&prio_qdisc_ops); - return err; + return register_qdisc(&prio_qdisc_ops); } static void __exit prio_module_exit(void) { unregister_qdisc(&prio_qdisc_ops); - unregister_qdisc(&rr_qdisc_ops); } module_init(prio_module_init) module_exit(prio_module_exit) MODULE_LICENSE("GPL"); -MODULE_ALIAS("sch_rr"); -- cgit v1.2.2 From 24344d2600108b9b79a60c0e4c43b3c499856d14 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 02:53:04 -0700 Subject: mac80211: Temporarily mark QoS support BROKEN. We will undo this after a few changsets. Signed-off-by: David S. Miller --- net/mac80211/Kconfig | 1 + 1 file changed, 1 insertion(+) (limited to 'net') diff --git a/net/mac80211/Kconfig b/net/mac80211/Kconfig index d2038418e2bd..10579def490b 100644 --- a/net/mac80211/Kconfig +++ b/net/mac80211/Kconfig @@ -15,6 +15,7 @@ config MAC80211_QOS def_bool y depends on MAC80211 depends on NET_SCHED + depends on BROKEN comment "QoS/HT support disabled" depends on MAC80211 && !MAC80211_QOS -- cgit v1.2.2 From fd2ea0a79faad824258af5dcec1927aa24d81c16 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 01:56:23 -0700 Subject: net: Use queue aware tests throughout. This effectively "flips the switch" by making the core networking and multiqueue-aware drivers use the new TX multiqueue structures. Non-multiqueue drivers need no changes. The interfaces they use such as netif_stop_queue() degenerate into an operation on TX queue zero. So everything "just works" for them. Code that really wants to do "X" to all TX queues now invokes a routine that does so, such as netif_tx_wake_all_queues(), netif_tx_stop_all_queues(), etc. pktgen and netpoll required a little bit more surgery than the others. In particular the pktgen changes, whilst functional, could be largely improved. The initial check in pktgen_xmit() will sometimes check the wrong queue, which is mostly harmless. The thing to do is probably to invoke fill_packet() earlier. The bulk of the netpoll changes is to make the code operate solely on the TX queue indicated by by the SKB queue mapping. Setting of the SKB queue mapping is entirely confined inside of net/core/dev.c:dev_pick_tx(). If we end up needing any kind of special semantics (drops, for example) it will be implemented here. Finally, we now have a "real_num_tx_queues" which is where the driver indicates how many TX queues are actually active. With IGB changes from Jeff Kirsher. Signed-off-by: David S. Miller --- net/core/dev.c | 28 +++++++++----------- net/core/netpoll.c | 24 ++++++++++------- net/core/pktgen.c | 69 +++++++++++++++++++++++++++++++------------------ net/sched/sch_generic.c | 5 ++-- net/sched/sch_teql.c | 6 ++--- 5 files changed, 74 insertions(+), 58 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 69378f250695..f027a1ac4fbb 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1598,7 +1598,8 @@ static int dev_gso_segment(struct sk_buff *skb) return 0; } -int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev) +int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev, + struct netdev_queue *txq) { if (likely(!skb->next)) { if (!list_empty(&ptype_all)) @@ -1627,9 +1628,7 @@ gso: skb->next = nskb; return rc; } - if (unlikely((netif_queue_stopped(dev) || - netif_subqueue_stopped(dev, skb)) && - skb->next)) + if (unlikely(netif_tx_queue_stopped(txq) && skb->next)) return NETDEV_TX_BUSY; } while (skb->next); @@ -1669,7 +1668,10 @@ out_kfree_skb: static struct netdev_queue *dev_pick_tx(struct net_device *dev, struct sk_buff *skb) { - return netdev_get_tx_queue(dev, 0); + u16 queue_index = 0; + + skb_set_queue_mapping(skb, queue_index); + return netdev_get_tx_queue(dev, queue_index); } int dev_queue_xmit(struct sk_buff *skb) @@ -1737,8 +1739,6 @@ gso: spin_lock(&txq->lock); q = txq->qdisc; if (q->enqueue) { - /* reset queue_mapping to zero */ - skb_set_queue_mapping(skb, 0); rc = q->enqueue(skb, q); qdisc_run(txq); spin_unlock(&txq->lock); @@ -1768,10 +1768,9 @@ gso: HARD_TX_LOCK(dev, txq, cpu); - if (!netif_queue_stopped(dev) && - !netif_subqueue_stopped(dev, skb)) { + if (!netif_tx_queue_stopped(txq)) { rc = 0; - if (!dev_hard_start_xmit(skb, dev)) { + if (!dev_hard_start_xmit(skb, dev, txq)) { HARD_TX_UNLOCK(dev, txq); goto out; } @@ -4160,8 +4159,7 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, BUG_ON(strlen(name) >= sizeof(dev->name)); - alloc_size = sizeof(struct net_device) + - sizeof(struct net_device_subqueue) * (queue_count - 1); + alloc_size = sizeof(struct net_device); if (sizeof_priv) { /* ensure 32-byte alignment of private area */ alloc_size = (alloc_size + NETDEV_ALIGN_CONST) & ~NETDEV_ALIGN_CONST; @@ -4191,16 +4189,14 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name, dev->_tx = tx; dev->num_tx_queues = queue_count; + dev->real_num_tx_queues = queue_count; if (sizeof_priv) { dev->priv = ((char *)dev + - ((sizeof(struct net_device) + - (sizeof(struct net_device_subqueue) * - (queue_count - 1)) + NETDEV_ALIGN_CONST) + ((sizeof(struct net_device) + NETDEV_ALIGN_CONST) & ~NETDEV_ALIGN_CONST)); } - dev->egress_subqueue_count = queue_count; dev->gso_max_size = GSO_MAX_SIZE; netdev_init_queues(dev); diff --git a/net/core/netpoll.c b/net/core/netpoll.c index 8fb134da0346..c12720895ecf 100644 --- a/net/core/netpoll.c +++ b/net/core/netpoll.c @@ -58,25 +58,27 @@ static void queue_process(struct work_struct *work) while ((skb = skb_dequeue(&npinfo->txq))) { struct net_device *dev = skb->dev; + struct netdev_queue *txq; if (!netif_device_present(dev) || !netif_running(dev)) { __kfree_skb(skb); continue; } + txq = netdev_get_tx_queue(dev, skb_get_queue_mapping(skb)); + local_irq_save(flags); - netif_tx_lock(dev); - if ((netif_queue_stopped(dev) || - netif_subqueue_stopped(dev, skb)) || - dev->hard_start_xmit(skb, dev) != NETDEV_TX_OK) { + __netif_tx_lock(txq, smp_processor_id()); + if (netif_tx_queue_stopped(txq) || + dev->hard_start_xmit(skb, dev) != NETDEV_TX_OK) { skb_queue_head(&npinfo->txq, skb); - netif_tx_unlock(dev); + __netif_tx_unlock(txq); local_irq_restore(flags); schedule_delayed_work(&npinfo->tx_work, HZ/10); return; } - netif_tx_unlock(dev); + __netif_tx_unlock(txq); local_irq_restore(flags); } } @@ -278,17 +280,19 @@ static void netpoll_send_skb(struct netpoll *np, struct sk_buff *skb) /* don't get messages out of order, and no recursion */ if (skb_queue_len(&npinfo->txq) == 0 && !netpoll_owner_active(dev)) { + struct netdev_queue *txq; unsigned long flags; + txq = netdev_get_tx_queue(dev, skb_get_queue_mapping(skb)); + local_irq_save(flags); /* try until next clock tick */ for (tries = jiffies_to_usecs(1)/USEC_PER_POLL; tries > 0; --tries) { - if (netif_tx_trylock(dev)) { - if (!netif_queue_stopped(dev) && - !netif_subqueue_stopped(dev, skb)) + if (__netif_tx_trylock(txq)) { + if (!netif_tx_queue_stopped(txq)) status = dev->hard_start_xmit(skb, dev); - netif_tx_unlock(dev); + __netif_tx_unlock(txq); if (status == NETDEV_TX_OK) break; diff --git a/net/core/pktgen.c b/net/core/pktgen.c index fdf537707e51..906802db4ed4 100644 --- a/net/core/pktgen.c +++ b/net/core/pktgen.c @@ -2123,6 +2123,24 @@ static void get_ipsec_sa(struct pktgen_dev *pkt_dev, int flow) } } #endif +static void set_cur_queue_map(struct pktgen_dev *pkt_dev) +{ + if (pkt_dev->queue_map_min < pkt_dev->queue_map_max) { + __u16 t; + if (pkt_dev->flags & F_QUEUE_MAP_RND) { + t = random32() % + (pkt_dev->queue_map_max - + pkt_dev->queue_map_min + 1) + + pkt_dev->queue_map_min; + } else { + t = pkt_dev->cur_queue_map + 1; + if (t > pkt_dev->queue_map_max) + t = pkt_dev->queue_map_min; + } + pkt_dev->cur_queue_map = t; + } +} + /* Increment/randomize headers according to flags and current values * for IP src/dest, UDP src/dst port, MAC-Addr src/dst */ @@ -2325,19 +2343,7 @@ static void mod_cur_headers(struct pktgen_dev *pkt_dev) pkt_dev->cur_pkt_size = t; } - if (pkt_dev->queue_map_min < pkt_dev->queue_map_max) { - __u16 t; - if (pkt_dev->flags & F_QUEUE_MAP_RND) { - t = random32() % - (pkt_dev->queue_map_max - pkt_dev->queue_map_min + 1) - + pkt_dev->queue_map_min; - } else { - t = pkt_dev->cur_queue_map + 1; - if (t > pkt_dev->queue_map_max) - t = pkt_dev->queue_map_min; - } - pkt_dev->cur_queue_map = t; - } + set_cur_queue_map(pkt_dev); pkt_dev->flows[flow].count++; } @@ -2458,7 +2464,7 @@ static struct sk_buff *fill_packet_ipv4(struct net_device *odev, __be16 *vlan_encapsulated_proto = NULL; /* packet type ID field (or len) for VLAN tag */ __be16 *svlan_tci = NULL; /* Encapsulates priority and SVLAN ID */ __be16 *svlan_encapsulated_proto = NULL; /* packet type ID field (or len) for SVLAN tag */ - + u16 queue_map; if (pkt_dev->nr_labels) protocol = htons(ETH_P_MPLS_UC); @@ -2469,6 +2475,7 @@ static struct sk_buff *fill_packet_ipv4(struct net_device *odev, /* Update any of the values, used when we're incrementing various * fields. */ + queue_map = pkt_dev->cur_queue_map; mod_cur_headers(pkt_dev); datalen = (odev->hard_header_len + 16) & ~0xf; @@ -2507,7 +2514,7 @@ static struct sk_buff *fill_packet_ipv4(struct net_device *odev, skb->network_header = skb->tail; skb->transport_header = skb->network_header + sizeof(struct iphdr); skb_put(skb, sizeof(struct iphdr) + sizeof(struct udphdr)); - skb_set_queue_mapping(skb, pkt_dev->cur_queue_map); + skb_set_queue_mapping(skb, queue_map); iph = ip_hdr(skb); udph = udp_hdr(skb); @@ -2797,6 +2804,7 @@ static struct sk_buff *fill_packet_ipv6(struct net_device *odev, __be16 *vlan_encapsulated_proto = NULL; /* packet type ID field (or len) for VLAN tag */ __be16 *svlan_tci = NULL; /* Encapsulates priority and SVLAN ID */ __be16 *svlan_encapsulated_proto = NULL; /* packet type ID field (or len) for SVLAN tag */ + u16 queue_map; if (pkt_dev->nr_labels) protocol = htons(ETH_P_MPLS_UC); @@ -2807,6 +2815,7 @@ static struct sk_buff *fill_packet_ipv6(struct net_device *odev, /* Update any of the values, used when we're incrementing various * fields. */ + queue_map = pkt_dev->cur_queue_map; mod_cur_headers(pkt_dev); skb = alloc_skb(pkt_dev->cur_pkt_size + 64 + 16 + @@ -2844,7 +2853,7 @@ static struct sk_buff *fill_packet_ipv6(struct net_device *odev, skb->network_header = skb->tail; skb->transport_header = skb->network_header + sizeof(struct ipv6hdr); skb_put(skb, sizeof(struct ipv6hdr) + sizeof(struct udphdr)); - skb_set_queue_mapping(skb, pkt_dev->cur_queue_map); + skb_set_queue_mapping(skb, queue_map); iph = ipv6_hdr(skb); udph = udp_hdr(skb); @@ -3263,7 +3272,9 @@ static void pktgen_rem_thread(struct pktgen_thread *t) static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) { struct net_device *odev = NULL; + struct netdev_queue *txq; __u64 idle_start = 0; + u16 queue_map; int ret; odev = pkt_dev->odev; @@ -3285,9 +3296,15 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) } } - if ((netif_queue_stopped(odev) || - (pkt_dev->skb && - netif_subqueue_stopped(odev, pkt_dev->skb))) || + if (!pkt_dev->skb) { + set_cur_queue_map(pkt_dev); + queue_map = pkt_dev->cur_queue_map; + } else { + queue_map = skb_get_queue_mapping(pkt_dev->skb); + } + + txq = netdev_get_tx_queue(odev, queue_map); + if (netif_tx_queue_stopped(txq) || need_resched()) { idle_start = getCurUs(); @@ -3303,8 +3320,7 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) pkt_dev->idle_acc += getCurUs() - idle_start; - if (netif_queue_stopped(odev) || - netif_subqueue_stopped(odev, pkt_dev->skb)) { + if (netif_tx_queue_stopped(txq)) { pkt_dev->next_tx_us = getCurUs(); /* TODO */ pkt_dev->next_tx_ns = 0; goto out; /* Try the next interface */ @@ -3331,9 +3347,12 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) } } - netif_tx_lock_bh(odev); - if (!netif_queue_stopped(odev) && - !netif_subqueue_stopped(odev, pkt_dev->skb)) { + /* fill_packet() might have changed the queue */ + queue_map = skb_get_queue_mapping(pkt_dev->skb); + txq = netdev_get_tx_queue(odev, queue_map); + + __netif_tx_lock_bh(txq); + if (!netif_tx_queue_stopped(txq)) { atomic_inc(&(pkt_dev->skb->users)); retry_now: @@ -3377,7 +3396,7 @@ static __inline__ void pktgen_xmit(struct pktgen_dev *pkt_dev) pkt_dev->next_tx_ns = 0; } - netif_tx_unlock_bh(odev); + __netif_tx_unlock_bh(txq); /* If pkt_dev->count is zero, then run forever */ if ((pkt_dev->count != 0) && (pkt_dev->sofar >= pkt_dev->count)) { diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 4e2b865cbba0..2f575b9017d1 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -166,7 +166,7 @@ static inline int qdisc_restart(struct netdev_queue *txq) HARD_TX_LOCK(dev, txq, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) - ret = dev_hard_start_xmit(skb, dev); + ret = dev_hard_start_xmit(skb, dev, txq); HARD_TX_UNLOCK(dev, txq); spin_lock(&txq->lock); @@ -198,11 +198,10 @@ static inline int qdisc_restart(struct netdev_queue *txq) void __qdisc_run(struct netdev_queue *txq) { - struct net_device *dev = txq->dev; unsigned long start_time = jiffies; while (qdisc_restart(txq)) { - if (netif_queue_stopped(dev)) + if (netif_tx_queue_stopped(txq)) break; /* diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index 44a2c3451f4d..ade3372221c7 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -295,8 +295,7 @@ restart: slave_txq = netdev_get_tx_queue(slave, 0); if (slave_txq->qdisc_sleeping != q) continue; - if (netif_queue_stopped(slave) || - __netif_subqueue_stopped(slave, subq) || + if (__netif_subqueue_stopped(slave, subq) || !netif_running(slave)) { busy = 1; continue; @@ -305,8 +304,7 @@ restart: switch (teql_resolve(skb, skb_res, slave)) { case 0: if (netif_tx_trylock(slave)) { - if (!netif_queue_stopped(slave) && - !__netif_subqueue_stopped(slave, subq) && + if (!__netif_subqueue_stopped(slave, subq) && slave->hard_start_xmit(skb, slave) == 0) { netif_tx_unlock(slave); master->slaves = NEXT_SLAVE(q); -- cgit v1.2.2 From eae792b722fef08dcf3aee88266ee7def9710757 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 03:03:33 -0700 Subject: netdev: Add netdev->select_queue() method. Devices or device layers can set this to control the queue selection performed by dev_pick_tx(). This function runs under RCU protection, which allows overriding functions to have some way of synchronizing with things like dynamic ->real_num_tx_queues adjustments. This makes the spinlock prefetch in dev_queue_xmit() a little bit less effective, but that's the price right now for correctness. Signed-off-by: David S. Miller --- net/core/dev.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index f027a1ac4fbb..7ca9564d2f44 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1670,6 +1670,9 @@ static struct netdev_queue *dev_pick_tx(struct net_device *dev, { u16 queue_index = 0; + if (dev->select_queue) + queue_index = dev->select_queue(dev, skb); + skb_set_queue_mapping(skb, queue_index); return netdev_get_tx_queue(dev, queue_index); } @@ -1710,14 +1713,14 @@ int dev_queue_xmit(struct sk_buff *skb) } gso: - txq = dev_pick_tx(dev, skb); - spin_lock_prefetch(&txq->lock); - /* Disable soft irqs for various locks below. Also * stops preemption for RCU. */ rcu_read_lock_bh(); + txq = dev_pick_tx(dev, skb); + spin_lock_prefetch(&txq->lock); + /* Updates of qdisc are serialized by queue->lock. * The struct Qdisc which is pointed to by qdisc is now a * rcu structure - it may be accessed without acquiring -- cgit v1.2.2 From 51cb6db0f5654f08a4a6bfa3888dc36a51c2df3e Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 03:34:57 -0700 Subject: mac80211: Reimplement WME using ->select_queue(). The only behavior change is that we do not drop packets under any circumstances. If that is absolutely needed, we could easily add it back. With cleanups and help from Johannes Berg. Signed-off-by: David S. Miller --- net/mac80211/Kconfig | 11 - net/mac80211/Makefile | 2 +- net/mac80211/ieee80211_i.h | 15 +- net/mac80211/main.c | 61 ++--- net/mac80211/util.c | 14 +- net/mac80211/wme.c | 602 ++++++++------------------------------------- net/mac80211/wme.h | 35 +-- 7 files changed, 124 insertions(+), 616 deletions(-) (limited to 'net') diff --git a/net/mac80211/Kconfig b/net/mac80211/Kconfig index 10579def490b..80d693392b0f 100644 --- a/net/mac80211/Kconfig +++ b/net/mac80211/Kconfig @@ -11,17 +11,6 @@ config MAC80211 This option enables the hardware independent IEEE 802.11 networking stack. -config MAC80211_QOS - def_bool y - depends on MAC80211 - depends on NET_SCHED - depends on BROKEN - -comment "QoS/HT support disabled" - depends on MAC80211 && !MAC80211_QOS -comment "QoS/HT support needs CONFIG_NET_SCHED" - depends on MAC80211 && !NET_SCHED - menu "Rate control algorithm selection" depends on MAC80211 != n diff --git a/net/mac80211/Makefile b/net/mac80211/Makefile index fa47438e338f..a169b0201d61 100644 --- a/net/mac80211/Makefile +++ b/net/mac80211/Makefile @@ -18,10 +18,10 @@ mac80211-y := \ tx.o \ key.o \ util.o \ + wme.o \ event.o mac80211-$(CONFIG_MAC80211_LEDS) += led.o -mac80211-$(CONFIG_MAC80211_QOS) += wme.o mac80211-$(CONFIG_MAC80211_DEBUGFS) += \ debugfs.o \ debugfs_sta.o \ diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h index cbea0154ee3a..a4f9a832722a 100644 --- a/net/mac80211/ieee80211_i.h +++ b/net/mac80211/ieee80211_i.h @@ -25,6 +25,7 @@ #include #include #include +#include #include "key.h" #include "sta_info.h" @@ -537,6 +538,9 @@ enum { IEEE80211_ADDBA_MSG = 4, }; +/* maximum number of hardware queues we support. */ +#define QD_MAX_QUEUES (IEEE80211_MAX_AMPDU_QUEUES + IEEE80211_MAX_QUEUES) + struct ieee80211_local { /* embed the driver visible part. * don't cast (use the static inlines below), but we keep @@ -545,6 +549,8 @@ struct ieee80211_local { const struct ieee80211_ops *ops; + unsigned long queue_pool[BITS_TO_LONGS(QD_MAX_QUEUES)]; + struct net_device *mdev; /* wmaster# - "master" 802.11 device */ int open_count; int monitors, cooked_mntrs; @@ -740,15 +746,6 @@ struct ieee80211_local { #endif }; -static inline int ieee80211_is_multiqueue(struct ieee80211_local *local) -{ -#ifdef CONFIG_MAC80211_QOS - return netif_is_multiqueue(local->mdev); -#else - return 0; -#endif -} - static inline struct ieee80211_sub_if_data * IEEE80211_DEV_TO_SUB_IF(struct net_device *dev) { diff --git a/net/mac80211/main.c b/net/mac80211/main.c index c74607eda1ee..f1a83d450ea0 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -114,7 +114,7 @@ static int ieee80211_master_open(struct net_device *dev) if (res) return res; - netif_start_queue(local->mdev); + netif_tx_start_all_queues(local->mdev); return 0; } @@ -375,7 +375,7 @@ static int ieee80211_open(struct net_device *dev) queue_work(local->hw.workqueue, &ifsta->work); } - netif_start_queue(dev); + netif_tx_start_all_queues(dev); return 0; err_del_interface: @@ -400,7 +400,7 @@ static int ieee80211_stop(struct net_device *dev) /* * Stop TX on this interface first. */ - netif_stop_queue(dev); + netif_tx_stop_all_queues(dev); /* * Now delete all active aggregation sessions. @@ -554,7 +554,6 @@ static int ieee80211_stop(struct net_device *dev) int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) { struct ieee80211_local *local = hw_to_local(hw); - struct netdev_queue *txq; struct sta_info *sta; struct ieee80211_sub_if_data *sdata; u16 start_seq_num = 0; @@ -619,11 +618,6 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) (unsigned long)&sta->timer_to_tid[tid]; init_timer(&sta->ampdu_mlme.tid_tx[tid]->addba_resp_timer); - /* ensure that TX flow won't interrupt us - * until the end of the call to requeue function */ - txq = netdev_get_tx_queue(local->mdev, 0); - spin_lock_bh(&txq->lock); - /* create a new queue for this aggregation */ ret = ieee80211_ht_agg_queue_add(local, sta, tid); @@ -650,7 +644,7 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) /* No need to requeue the packets in the agg queue, since we * held the tx lock: no packet could be enqueued to the newly * allocated queue */ - ieee80211_ht_agg_queue_remove(local, sta, tid, 0); + ieee80211_ht_agg_queue_remove(local, sta, tid, 0); #ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "BA request denied - HW unavailable for" " tid %d\n", tid); @@ -661,7 +655,6 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) /* Will put all the packets in the new SW queue */ ieee80211_requeue(local, ieee802_1d_to_ac[tid]); - spin_unlock_bh(&txq->lock); spin_unlock_bh(&sta->lock); /* send an addBA request */ @@ -687,7 +680,6 @@ int ieee80211_start_tx_ba_session(struct ieee80211_hw *hw, u8 *ra, u16 tid) err_unlock_queue: kfree(sta->ampdu_mlme.tid_tx[tid]); sta->ampdu_mlme.tid_tx[tid] = NULL; - spin_unlock_bh(&txq->lock); ret = -EBUSY; err_unlock_sta: spin_unlock_bh(&sta->lock); @@ -812,7 +804,6 @@ EXPORT_SYMBOL(ieee80211_start_tx_ba_cb); void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) { struct ieee80211_local *local = hw_to_local(hw); - struct netdev_queue *txq; struct sta_info *sta; u8 *state; int agg_queue; @@ -844,8 +835,9 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) state = &sta->ampdu_mlme.tid_state_tx[tid]; /* NOTE: no need to use sta->lock in this state check, as - * ieee80211_stop_tx_ba_session will let only - * one stop call to pass through per sta/tid */ + * ieee80211_stop_tx_ba_session will let only one stop call to + * pass through per sta/tid + */ if ((*state & HT_AGG_STATE_REQ_STOP_BA_MSK) == 0) { #ifdef CONFIG_MAC80211_HT_DEBUG printk(KERN_DEBUG "unexpected callback to A-MPDU stop\n"); @@ -860,19 +852,14 @@ void ieee80211_stop_tx_ba_cb(struct ieee80211_hw *hw, u8 *ra, u8 tid) agg_queue = sta->tid_to_tx_q[tid]; - /* avoid ordering issues: we are the only one that can modify - * the content of the qdiscs */ - txq = netdev_get_tx_queue(local->mdev, 0); - spin_lock_bh(&txq->lock); - /* remove the queue for this aggregation */ ieee80211_ht_agg_queue_remove(local, sta, tid, 1); - spin_unlock_bh(&txq->lock); - /* we just requeued the all the frames that were in the removed - * queue, and since we might miss a softirq we do netif_schedule_queue. - * ieee80211_wake_queue is not used here as this queue is not - * necessarily stopped */ - netif_schedule_queue(txq); + /* We just requeued the all the frames that were in the + * removed queue, and since we might miss a softirq we do + * netif_schedule_queue. ieee80211_wake_queue is not used + * here as this queue is not necessarily stopped + */ + netif_schedule_queue(netdev_get_tx_queue(local->mdev, agg_queue)); spin_lock_bh(&sta->lock); *state = HT_AGG_STATE_IDLE; sta->ampdu_mlme.addba_req_num[tid] = 0; @@ -1660,17 +1647,12 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) * We use the number of queues for feature tests (QoS, HT) internally * so restrict them appropriately. */ -#ifdef CONFIG_MAC80211_QOS if (hw->queues > IEEE80211_MAX_QUEUES) hw->queues = IEEE80211_MAX_QUEUES; if (hw->ampdu_queues > IEEE80211_MAX_AMPDU_QUEUES) hw->ampdu_queues = IEEE80211_MAX_AMPDU_QUEUES; if (hw->queues < 4) hw->ampdu_queues = 0; -#else - hw->queues = 1; - hw->ampdu_queues = 0; -#endif mdev = alloc_netdev_mq(sizeof(struct wireless_dev), "wmaster%d", ether_setup, @@ -1754,7 +1736,7 @@ int ieee80211_register_hw(struct ieee80211_hw *hw) goto fail_wep; } - ieee80211_install_qdisc(local->mdev); + local->mdev->select_queue = ieee80211_select_queue; /* add one default STA interface */ result = ieee80211_if_add(local, "wlan%d", NULL, @@ -1852,23 +1834,11 @@ static int __init ieee80211_init(void) ret = rc80211_pid_init(); if (ret) - goto out; - - ret = ieee80211_wme_register(); - if (ret) { - printk(KERN_DEBUG "ieee80211_init: failed to " - "initialize WME (err=%d)\n", ret); - goto out_cleanup_pid; - } + return ret; ieee80211_debugfs_netdev_init(); return 0; - - out_cleanup_pid: - rc80211_pid_exit(); - out: - return ret; } static void __exit ieee80211_exit(void) @@ -1884,7 +1854,6 @@ static void __exit ieee80211_exit(void) if (mesh_allocated) ieee80211s_stop(); - ieee80211_wme_unregister(); ieee80211_debugfs_netdev_exit(); } diff --git a/net/mac80211/util.c b/net/mac80211/util.c index 89ce4e07bd84..19f85e1b3695 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -363,12 +363,7 @@ void ieee80211_wake_queue(struct ieee80211_hw *hw, int queue) if (test_bit(queue, local->queues_pending)) { tasklet_schedule(&local->tx_pending_tasklet); } else { - if (ieee80211_is_multiqueue(local)) { - netif_wake_subqueue(local->mdev, queue); - } else { - WARN_ON(queue != 0); - netif_wake_queue(local->mdev); - } + netif_wake_subqueue(local->mdev, queue); } } EXPORT_SYMBOL(ieee80211_wake_queue); @@ -377,12 +372,7 @@ void ieee80211_stop_queue(struct ieee80211_hw *hw, int queue) { struct ieee80211_local *local = hw_to_local(hw); - if (ieee80211_is_multiqueue(local)) { - netif_stop_subqueue(local->mdev, queue); - } else { - WARN_ON(queue != 0); - netif_stop_queue(local->mdev); - } + netif_stop_subqueue(local->mdev, queue); } EXPORT_SYMBOL(ieee80211_stop_queue); diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index f014cd38c2d0..b21cfec4b6ce 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -18,67 +18,42 @@ #include "ieee80211_i.h" #include "wme.h" -/* maximum number of hardware queues we support. */ -#define QD_MAX_QUEUES (IEEE80211_MAX_AMPDU_QUEUES + IEEE80211_MAX_QUEUES) -/* current number of hardware queues we support. */ -#define QD_NUM(hw) ((hw)->queues + (hw)->ampdu_queues) - -/* - * Default mapping in classifier to work with default +/* Default mapping in classifier to work with default * queue setup. */ const int ieee802_1d_to_ac[8] = { 2, 3, 3, 2, 1, 1, 0, 0 }; -struct ieee80211_sched_data -{ - unsigned long qdisc_pool[BITS_TO_LONGS(QD_MAX_QUEUES)]; - struct tcf_proto *filter_list; - struct Qdisc *queues[QD_MAX_QUEUES]; - struct sk_buff_head requeued[QD_MAX_QUEUES]; -}; - static const char llc_ip_hdr[8] = {0xAA, 0xAA, 0x3, 0, 0, 0, 0x08, 0}; -/* given a data frame determine the 802.1p/1d tag to use */ -static inline unsigned classify_1d(struct sk_buff *skb, struct Qdisc *qd) +/* Given a data frame determine the 802.1p/1d tag to use. */ +static unsigned int classify_1d(struct sk_buff *skb) { - struct iphdr *ip; - int dscp; - int offset; - - struct ieee80211_sched_data *q = qdisc_priv(qd); - struct tcf_result res = { -1, 0 }; - - /* if there is a user set filter list, call out to that */ - if (q->filter_list) { - tc_classify(skb, q->filter_list, &res); - if (res.class != -1) - return res.class; - } + unsigned int dscp; /* skb->priority values from 256->263 are magic values to - * directly indicate a specific 802.1d priority. - * This is used to allow 802.1d priority to be passed directly in - * from VLAN tags, etc. */ + * directly indicate a specific 802.1d priority. This is used + * to allow 802.1d priority to be passed directly in from VLAN + * tags, etc. + */ if (skb->priority >= 256 && skb->priority <= 263) return skb->priority - 256; - /* check there is a valid IP header present */ - offset = ieee80211_get_hdrlen_from_skb(skb); - if (skb->len < offset + sizeof(llc_ip_hdr) + sizeof(*ip) || - memcmp(skb->data + offset, llc_ip_hdr, sizeof(llc_ip_hdr))) - return 0; + switch (skb->protocol) { + case __constant_htons(ETH_P_IP): + dscp = ip_hdr(skb)->tos & 0xfc; + break; - ip = (struct iphdr *) (skb->data + offset + sizeof(llc_ip_hdr)); + default: + return 0; + } - dscp = ip->tos & 0xfc; if (dscp & 0x1c) return 0; return dscp >> 5; } -static inline int wme_downgrade_ac(struct sk_buff *skb) +static int wme_downgrade_ac(struct sk_buff *skb) { switch (skb->priority) { case 6: @@ -99,11 +74,10 @@ static inline int wme_downgrade_ac(struct sk_buff *skb) } -/* positive return value indicates which queue to use - * negative return value indicates to drop the frame */ -static int classify80211(struct sk_buff *skb, struct Qdisc *qd) +/* Indicate which queue to use. */ +static u16 classify80211(struct sk_buff *skb, struct net_device *dev) { - struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); + struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; if (!ieee80211_is_data(hdr->frame_control)) { @@ -123,13 +97,15 @@ static int classify80211(struct sk_buff *skb, struct Qdisc *qd) /* use the data classifier to determine what 802.1d tag the * data frame has */ - skb->priority = classify_1d(skb, qd); + skb->priority = classify_1d(skb); /* in case we are a client verify acm is not set for this ac */ while (unlikely(local->wmm_acm & BIT(skb->priority))) { if (wme_downgrade_ac(skb)) { - /* No AC with lower priority has acm=0, drop packet. */ - return -1; + /* The old code would drop the packet in this + * case. + */ + return 0; } } @@ -137,28 +113,29 @@ static int classify80211(struct sk_buff *skb, struct Qdisc *qd) return ieee802_1d_to_ac[skb->priority]; } - -static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) +u16 ieee80211_select_queue(struct net_device *dev, struct sk_buff *skb) { - struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); - struct ieee80211_hw *hw = &local->hw; - struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data; - struct Qdisc *qdisc; + struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); + struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb); struct sta_info *sta; - int err, queue; + u16 queue; u8 tid; + queue = classify80211(skb, dev); + if (unlikely(queue >= local->hw.queues)) + queue = local->hw.queues - 1; + if (info->flags & IEEE80211_TX_CTL_REQUEUE) { - queue = skb_get_queue_mapping(skb); rcu_read_lock(); sta = sta_info_get(local, hdr->addr1); tid = skb->priority & IEEE80211_QOS_CTL_TAG1D_MASK; if (sta) { + struct ieee80211_hw *hw = &local->hw; int ampdu_queue = sta->tid_to_tx_q[tid]; - if ((ampdu_queue < QD_NUM(hw)) && - test_bit(ampdu_queue, q->qdisc_pool)) { + + if ((ampdu_queue < ieee80211_num_queues(hw)) && + test_bit(ampdu_queue, local->queue_pool)) { queue = ampdu_queue; info->flags |= IEEE80211_TX_CTL_AMPDU; } else { @@ -166,17 +143,12 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) } } rcu_read_unlock(); - skb_queue_tail(&q->requeued[queue], skb); - qd->q.qlen++; - return 0; - } - - queue = classify80211(skb, qd); - if (unlikely(queue >= local->hw.queues)) - queue = local->hw.queues - 1; + return queue; + } - /* now we know the 1d priority, fill in the QoS header if there is one + /* Now we know the 1d priority, fill in the QoS header if + * there is one. */ if (ieee80211_is_data_qos(hdr->frame_control)) { u8 *p = ieee80211_get_qos_ctl(hdr); @@ -194,8 +166,10 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) sta = sta_info_get(local, hdr->addr1); if (sta) { int ampdu_queue = sta->tid_to_tx_q[tid]; - if ((ampdu_queue < QD_NUM(hw)) && - test_bit(ampdu_queue, q->qdisc_pool)) { + struct ieee80211_hw *hw = &local->hw; + + if ((ampdu_queue < ieee80211_num_queues(hw)) && + test_bit(ampdu_queue, local->queue_pool)) { queue = ampdu_queue; info->flags |= IEEE80211_TX_CTL_AMPDU; } else { @@ -206,421 +180,13 @@ static int wme_qdiscop_enqueue(struct sk_buff *skb, struct Qdisc* qd) rcu_read_unlock(); } - if (unlikely(queue < 0)) { - kfree_skb(skb); - err = NET_XMIT_DROP; - } else { - skb_set_queue_mapping(skb, queue); - qdisc = q->queues[queue]; - err = qdisc->enqueue(skb, qdisc); - if (err == NET_XMIT_SUCCESS) { - qd->q.qlen++; - qd->bstats.bytes += skb->len; - qd->bstats.packets++; - return NET_XMIT_SUCCESS; - } - } - qd->qstats.drops++; - return err; -} - - -/* TODO: clean up the cases where master_hard_start_xmit - * returns non 0 - it shouldn't ever do that. Once done we - * can remove this function */ -static int wme_qdiscop_requeue(struct sk_buff *skb, struct Qdisc* qd) -{ - struct ieee80211_sched_data *q = qdisc_priv(qd); - struct Qdisc *qdisc; - int err; - - /* we recorded which queue to use earlier! */ - qdisc = q->queues[skb_get_queue_mapping(skb)]; - - if ((err = qdisc->ops->requeue(skb, qdisc)) == 0) { - qd->q.qlen++; - return 0; - } - qd->qstats.drops++; - return err; -} - - -static struct sk_buff *wme_qdiscop_dequeue(struct Qdisc* qd) -{ - struct ieee80211_sched_data *q = qdisc_priv(qd); - struct net_device *dev = qdisc_dev(qd); - struct ieee80211_local *local = wdev_priv(dev->ieee80211_ptr); - struct ieee80211_hw *hw = &local->hw; - struct sk_buff *skb; - struct Qdisc *qdisc; - int queue; - - /* check all the h/w queues in numeric/priority order */ - for (queue = 0; queue < QD_NUM(hw); queue++) { - /* see if there is room in this hardware queue */ - if (__netif_subqueue_stopped(local->mdev, queue) || - !test_bit(queue, q->qdisc_pool)) - continue; - - /* there is space - try and get a frame */ - skb = skb_dequeue(&q->requeued[queue]); - if (skb) { - qd->q.qlen--; - return skb; - } - - qdisc = q->queues[queue]; - skb = qdisc->dequeue(qdisc); - if (skb) { - qd->q.qlen--; - return skb; - } - } - /* returning a NULL here when all the h/w queues are full means we - * never need to call netif_stop_queue in the driver */ - return NULL; -} - - -static void wme_qdiscop_reset(struct Qdisc* qd) -{ - struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); - struct ieee80211_hw *hw = &local->hw; - int queue; - - /* QUESTION: should we have some hardware flush functionality here? */ - - for (queue = 0; queue < QD_NUM(hw); queue++) { - skb_queue_purge(&q->requeued[queue]); - qdisc_reset(q->queues[queue]); - } - qd->q.qlen = 0; -} - - -static void wme_qdiscop_destroy(struct Qdisc* qd) -{ - struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); - struct ieee80211_hw *hw = &local->hw; - int queue; - - tcf_destroy_chain(&q->filter_list); - - for (queue = 0; queue < QD_NUM(hw); queue++) { - skb_queue_purge(&q->requeued[queue]); - qdisc_destroy(q->queues[queue]); - q->queues[queue] = &noop_qdisc; - } -} - - -/* called whenever parameters are updated on existing qdisc */ -static int wme_qdiscop_tune(struct Qdisc *qd, struct nlattr *opt) -{ - return 0; -} - - -/* called during initial creation of qdisc on device */ -static int wme_qdiscop_init(struct Qdisc *qd, struct nlattr *opt) -{ - struct ieee80211_sched_data *q = qdisc_priv(qd); - struct net_device *dev = qdisc_dev(qd); - struct ieee80211_local *local; - struct ieee80211_hw *hw; - int err = 0, i; - - /* check that device is a mac80211 device */ - if (!dev->ieee80211_ptr || - dev->ieee80211_ptr->wiphy->privid != mac80211_wiphy_privid) - return -EINVAL; - - local = wdev_priv(dev->ieee80211_ptr); - hw = &local->hw; - - /* only allow on master dev */ - if (dev != local->mdev) - return -EINVAL; - - /* ensure that we are root qdisc */ - if (qd->parent != TC_H_ROOT) - return -EINVAL; - - if (qd->flags & TCQ_F_INGRESS) - return -EINVAL; - - /* if options were passed in, set them */ - if (opt) - err = wme_qdiscop_tune(qd, opt); - - /* create child queues */ - for (i = 0; i < QD_NUM(hw); i++) { - skb_queue_head_init(&q->requeued[i]); - q->queues[i] = qdisc_create_dflt(qdisc_dev(qd), qd->dev_queue, - &pfifo_qdisc_ops, - qd->handle); - if (!q->queues[i]) { - q->queues[i] = &noop_qdisc; - printk(KERN_ERR "%s child qdisc %i creation failed\n", - dev->name, i); - } - } - - /* non-aggregation queues: reserve/mark as used */ - for (i = 0; i < local->hw.queues; i++) - set_bit(i, q->qdisc_pool); - - return err; -} - -static int wme_qdiscop_dump(struct Qdisc *qd, struct sk_buff *skb) -{ - return -1; -} - - -static int wme_classop_graft(struct Qdisc *qd, unsigned long arg, - struct Qdisc *new, struct Qdisc **old) -{ - struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); - struct ieee80211_hw *hw = &local->hw; - unsigned long queue = arg - 1; - - if (queue >= QD_NUM(hw)) - return -EINVAL; - - if (!new) - new = &noop_qdisc; - - sch_tree_lock(qd); - *old = q->queues[queue]; - q->queues[queue] = new; - qdisc_reset(*old); - sch_tree_unlock(qd); - - return 0; -} - - -static struct Qdisc * -wme_classop_leaf(struct Qdisc *qd, unsigned long arg) -{ - struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); - struct ieee80211_hw *hw = &local->hw; - unsigned long queue = arg - 1; - - if (queue >= QD_NUM(hw)) - return NULL; - - return q->queues[queue]; -} - - -static unsigned long wme_classop_get(struct Qdisc *qd, u32 classid) -{ - struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); - struct ieee80211_hw *hw = &local->hw; - unsigned long queue = TC_H_MIN(classid); - - if (queue - 1 >= QD_NUM(hw)) - return 0; - return queue; } - -static unsigned long wme_classop_bind(struct Qdisc *qd, unsigned long parent, - u32 classid) -{ - return wme_classop_get(qd, classid); -} - - -static void wme_classop_put(struct Qdisc *q, unsigned long cl) -{ -} - - -static int wme_classop_change(struct Qdisc *qd, u32 handle, u32 parent, - struct nlattr **tca, unsigned long *arg) -{ - unsigned long cl = *arg; - struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); - struct ieee80211_hw *hw = &local->hw; - - if (cl - 1 > QD_NUM(hw)) - return -ENOENT; - - /* TODO: put code to program hardware queue parameters here, - * to allow programming from tc command line */ - - return 0; -} - - -/* we don't support deleting hardware queues - * when we add WMM-SA support - TSPECs may be deleted here */ -static int wme_classop_delete(struct Qdisc *qd, unsigned long cl) -{ - struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); - struct ieee80211_hw *hw = &local->hw; - - if (cl - 1 > QD_NUM(hw)) - return -ENOENT; - return 0; -} - - -static int wme_classop_dump_class(struct Qdisc *qd, unsigned long cl, - struct sk_buff *skb, struct tcmsg *tcm) -{ - struct ieee80211_sched_data *q = qdisc_priv(qd); - struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); - struct ieee80211_hw *hw = &local->hw; - - if (cl - 1 > QD_NUM(hw)) - return -ENOENT; - tcm->tcm_handle = TC_H_MIN(cl); - tcm->tcm_parent = qd->handle; - tcm->tcm_info = q->queues[cl-1]->handle; /* do we need this? */ - return 0; -} - - -static void wme_classop_walk(struct Qdisc *qd, struct qdisc_walker *arg) -{ - struct ieee80211_local *local = wdev_priv(qdisc_dev(qd)->ieee80211_ptr); - struct ieee80211_hw *hw = &local->hw; - int queue; - - if (arg->stop) - return; - - for (queue = 0; queue < QD_NUM(hw); queue++) { - if (arg->count < arg->skip) { - arg->count++; - continue; - } - /* we should return classids for our internal queues here - * as well as the external ones */ - if (arg->fn(qd, queue+1, arg) < 0) { - arg->stop = 1; - break; - } - arg->count++; - } -} - - -static struct tcf_proto ** wme_classop_find_tcf(struct Qdisc *qd, - unsigned long cl) -{ - struct ieee80211_sched_data *q = qdisc_priv(qd); - - if (cl) - return NULL; - - return &q->filter_list; -} - - -/* this qdisc is classful (i.e. has classes, some of which may have leaf qdiscs attached) - * - these are the operations on the classes */ -static const struct Qdisc_class_ops class_ops = -{ - .graft = wme_classop_graft, - .leaf = wme_classop_leaf, - - .get = wme_classop_get, - .put = wme_classop_put, - .change = wme_classop_change, - .delete = wme_classop_delete, - .walk = wme_classop_walk, - - .tcf_chain = wme_classop_find_tcf, - .bind_tcf = wme_classop_bind, - .unbind_tcf = wme_classop_put, - - .dump = wme_classop_dump_class, -}; - - -/* queueing discipline operations */ -static struct Qdisc_ops wme_qdisc_ops __read_mostly = -{ - .next = NULL, - .cl_ops = &class_ops, - .id = "ieee80211", - .priv_size = sizeof(struct ieee80211_sched_data), - - .enqueue = wme_qdiscop_enqueue, - .dequeue = wme_qdiscop_dequeue, - .requeue = wme_qdiscop_requeue, - .drop = NULL, /* drop not needed since we are always the root qdisc */ - - .init = wme_qdiscop_init, - .reset = wme_qdiscop_reset, - .destroy = wme_qdiscop_destroy, - .change = wme_qdiscop_tune, - - .dump = wme_qdiscop_dump, -}; - - -void ieee80211_install_qdisc(struct net_device *dev) -{ - struct netdev_queue *txq = netdev_get_tx_queue(dev, 0); - struct Qdisc *qdisc; - - qdisc = qdisc_create_dflt(dev, txq, - &wme_qdisc_ops, TC_H_ROOT); - if (!qdisc) { - printk(KERN_ERR "%s: qdisc installation failed\n", dev->name); - return; - } - - /* same handle as would be allocated by qdisc_alloc_handle() */ - qdisc->handle = 0x80010000; - - qdisc_lock_tree(dev); - list_add_tail(&qdisc->list, &txq->qdisc_list); - txq->qdisc_sleeping = qdisc; - qdisc_unlock_tree(dev); -} - - -int ieee80211_qdisc_installed(struct net_device *dev) -{ - struct netdev_queue *txq = netdev_get_tx_queue(dev, 0); - - return txq->qdisc_sleeping->ops == &wme_qdisc_ops; -} - - -int ieee80211_wme_register(void) -{ - return register_qdisc(&wme_qdisc_ops); -} - - -void ieee80211_wme_unregister(void) -{ - unregister_qdisc(&wme_qdisc_ops); -} - int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, - struct sta_info *sta, u16 tid) + struct sta_info *sta, u16 tid) { int i; - struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, 0); - struct ieee80211_sched_data *q = - qdisc_priv(txq->qdisc_sleeping); - DECLARE_MAC_BUF(mac); /* prepare the filter and save it for the SW queue * matching the received HW queue */ @@ -629,8 +195,8 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, return -EPERM; /* try to get a Qdisc from the pool */ - for (i = local->hw.queues; i < QD_NUM(&local->hw); i++) - if (!test_and_set_bit(i, q->qdisc_pool)) { + for (i = local->hw.queues; i < ieee80211_num_queues(&local->hw); i++) + if (!test_and_set_bit(i, local->queue_pool)) { ieee80211_stop_queue(local_to_hw(local), i); sta->tid_to_tx_q[tid] = i; @@ -639,11 +205,13 @@ int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, * on the previous queue * since HT is strict in order */ #ifdef CONFIG_MAC80211_HT_DEBUG - if (net_ratelimit()) + if (net_ratelimit()) { + DECLARE_MAC_BUF(mac); printk(KERN_DEBUG "allocated aggregation queue" " %d tid %d addr %s pool=0x%lX\n", i, tid, print_mac(mac, sta->addr), - q->qdisc_pool[0]); + local->queue_pool[0]); + } #endif /* CONFIG_MAC80211_HT_DEBUG */ return 0; } @@ -658,40 +226,68 @@ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, struct sta_info *sta, u16 tid, u8 requeue) { - struct ieee80211_hw *hw = &local->hw; - struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, 0); - struct ieee80211_sched_data *q = - qdisc_priv(txq->qdisc_sleeping); int agg_queue = sta->tid_to_tx_q[tid]; + struct ieee80211_hw *hw = &local->hw; /* return the qdisc to the pool */ - clear_bit(agg_queue, q->qdisc_pool); - sta->tid_to_tx_q[tid] = QD_NUM(hw); + clear_bit(agg_queue, local->queue_pool); + sta->tid_to_tx_q[tid] = ieee80211_num_queues(hw); - if (requeue) + if (requeue) { ieee80211_requeue(local, agg_queue); - else - q->queues[agg_queue]->ops->reset(q->queues[agg_queue]); + } else { + struct netdev_queue *txq; + + txq = netdev_get_tx_queue(local->mdev, agg_queue); + + spin_lock_bh(&txq->lock); + qdisc_reset(txq->qdisc); + spin_unlock_bh(&txq->lock); + } } void ieee80211_requeue(struct ieee80211_local *local, int queue) { - struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, 0); - struct Qdisc *root_qd = txq->qdisc_sleeping; - struct ieee80211_sched_data *q = qdisc_priv(root_qd); - struct Qdisc *qdisc = q->queues[queue]; - struct sk_buff *skb = NULL; + struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, queue); + struct sk_buff_head list; + struct Qdisc *qdisc; u32 len; + rcu_read_lock_bh(); + + qdisc = rcu_dereference(txq->qdisc); if (!qdisc || !qdisc->dequeue) - return; + goto out_unlock; + + skb_queue_head_init(&list); + spin_lock(&txq->lock); for (len = qdisc->q.qlen; len > 0; len--) { - skb = qdisc->dequeue(qdisc); - root_qd->q.qlen--; - /* packet will be classified again and */ - /* skb->packet_data->queue will be overridden if needed */ + struct sk_buff *skb = qdisc->dequeue(qdisc); + if (skb) - wme_qdiscop_enqueue(skb, root_qd); + __skb_queue_tail(&list, skb); + } + spin_unlock(&txq->lock); + + for (len = list.qlen; len > 0; len--) { + struct sk_buff *skb = __skb_dequeue(&list); + u16 new_queue; + + BUG_ON(!skb); + new_queue = ieee80211_select_queue(local->mdev, skb); + skb_set_queue_mapping(skb, new_queue); + + txq = netdev_get_tx_queue(local->mdev, new_queue); + + spin_lock(&txq->lock); + + qdisc = rcu_dereference(txq->qdisc); + qdisc->enqueue(skb, qdisc); + + spin_unlock(&txq->lock); } + +out_unlock: + rcu_read_unlock_bh(); } diff --git a/net/mac80211/wme.h b/net/mac80211/wme.h index 1aca609eccfc..04de28c071a6 100644 --- a/net/mac80211/wme.h +++ b/net/mac80211/wme.h @@ -23,45 +23,12 @@ extern const int ieee802_1d_to_ac[8]; -#ifdef CONFIG_MAC80211_QOS -void ieee80211_install_qdisc(struct net_device *dev); -int ieee80211_qdisc_installed(struct net_device *dev); +u16 ieee80211_select_queue(struct net_device *dev, struct sk_buff *skb); int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, struct sta_info *sta, u16 tid); void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, struct sta_info *sta, u16 tid, u8 requeue); void ieee80211_requeue(struct ieee80211_local *local, int queue); -int ieee80211_wme_register(void); -void ieee80211_wme_unregister(void); -#else -static inline void ieee80211_install_qdisc(struct net_device *dev) -{ -} -static inline int ieee80211_qdisc_installed(struct net_device *dev) -{ - return 0; -} -static inline int ieee80211_ht_agg_queue_add(struct ieee80211_local *local, - struct sta_info *sta, u16 tid) -{ - return -EAGAIN; -} -static inline void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, - struct sta_info *sta, u16 tid, - u8 requeue) -{ -} -static inline void ieee80211_requeue(struct ieee80211_local *local, int queue) -{ -} -static inline int ieee80211_wme_register(void) -{ - return 0; -} -static inline void ieee80211_wme_unregister(void) -{ -} -#endif /* CONFIG_NET_SCHED */ #endif /* _WME_H */ -- cgit v1.2.2 From 8f0f2223cc08a5ae9a77f40edfe02e8a9f1abd77 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 03:47:03 -0700 Subject: net: Implement simple sw TX hashing. It just xor hashes over IPv4/IPv6 addresses and ports of transport. The only assumption it makes is that skb_network_header() is set correctly. With bug fixes from Eric Dumazet. Signed-off-by: David S. Miller --- net/core/dev.c | 52 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 52 insertions(+) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 7ca9564d2f44..467bfb325123 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -121,6 +121,9 @@ #include #include #include +#include +#include +#include #include "net-sysfs.h" @@ -1665,6 +1668,53 @@ out_kfree_skb: * --BLG */ +static u16 simple_tx_hash(struct net_device *dev, struct sk_buff *skb) +{ + u32 *addr, *ports, hash, ihl; + u8 ip_proto; + int alen; + + switch (skb->protocol) { + case __constant_htons(ETH_P_IP): + ip_proto = ip_hdr(skb)->protocol; + addr = &ip_hdr(skb)->saddr; + ihl = ip_hdr(skb)->ihl; + alen = 2; + break; + case __constant_htons(ETH_P_IPV6): + ip_proto = ipv6_hdr(skb)->nexthdr; + addr = &ipv6_hdr(skb)->saddr.s6_addr32[0]; + ihl = (40 >> 2); + alen = 8; + break; + default: + return 0; + } + + ports = (u32 *) (skb_network_header(skb) + (ihl * 4)); + + hash = 0; + while (alen--) + hash ^= *addr++; + + switch (ip_proto) { + case IPPROTO_TCP: + case IPPROTO_UDP: + case IPPROTO_DCCP: + case IPPROTO_ESP: + case IPPROTO_AH: + case IPPROTO_SCTP: + case IPPROTO_UDPLITE: + hash ^= *ports; + break; + + default: + break; + } + + return hash % dev->real_num_tx_queues; +} + static struct netdev_queue *dev_pick_tx(struct net_device *dev, struct sk_buff *skb) { @@ -1672,6 +1722,8 @@ static struct netdev_queue *dev_pick_tx(struct net_device *dev, if (dev->select_queue) queue_index = dev->select_queue(dev, skb); + else if (dev->real_num_tx_queues > 1) + queue_index = simple_tx_hash(dev, skb); skb_set_queue_mapping(skb, queue_index); return netdev_get_tx_queue(dev, queue_index); -- cgit v1.2.2 From d3b753db7c4f1f37a98b51974d484fda5d86dab5 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 15 Jul 2008 20:14:35 -0700 Subject: pkt_sched: Move gso_skb into Qdisc. We liberate any dangling gso_skb during qdisc destruction. It really only matters for the root qdisc. But when qdiscs can be shared by multiple netdev_queue objects, we can't have the gso_skb in the netdev_queue any more. Signed-off-by: David S. Miller --- net/sched/sch_generic.c | 19 +++++++++---------- 1 file changed, 9 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 2f575b9017d1..2bd75befa066 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -77,7 +77,7 @@ static inline int dev_requeue_skb(struct sk_buff *skb, struct Qdisc *q) { if (unlikely(skb->next)) - dev_queue->gso_skb = skb; + q->gso_skb = skb; else q->ops->requeue(skb, q); @@ -85,13 +85,12 @@ static inline int dev_requeue_skb(struct sk_buff *skb, return 0; } -static inline struct sk_buff *dequeue_skb(struct netdev_queue *dev_queue, - struct Qdisc *q) +static inline struct sk_buff *dequeue_skb(struct Qdisc *q) { struct sk_buff *skb; - if ((skb = dev_queue->gso_skb)) - dev_queue->gso_skb = NULL; + if ((skb = q->gso_skb)) + q->gso_skb = NULL; else skb = q->dequeue(q); @@ -155,10 +154,9 @@ static inline int qdisc_restart(struct netdev_queue *txq) struct sk_buff *skb; /* Dequeue packet */ - if (unlikely((skb = dequeue_skb(txq, q)) == NULL)) + if (unlikely((skb = dequeue_skb(q)) == NULL)) return 0; - /* And release queue */ spin_unlock(&txq->lock); @@ -643,8 +641,8 @@ static void dev_deactivate_queue(struct net_device *dev, void *_qdisc_default) { struct Qdisc *qdisc_default = _qdisc_default; + struct sk_buff *skb = NULL; struct Qdisc *qdisc; - struct sk_buff *skb; spin_lock_bh(&dev_queue->lock); @@ -652,9 +650,10 @@ static void dev_deactivate_queue(struct net_device *dev, if (qdisc) { dev_queue->qdisc = qdisc_default; qdisc_reset(qdisc); + + skb = qdisc->gso_skb; + qdisc->gso_skb = NULL; } - skb = dev_queue->gso_skb; - dev_queue->gso_skb = NULL; spin_unlock_bh(&dev_queue->lock); -- cgit v1.2.2 From e2627c8c2241bce45e368e150654d076b58a4595 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 16 Jul 2008 00:56:32 -0700 Subject: pkt_sched: Make QDISC_RUNNING a qdisc state. Currently it is associated with a netdev_queue, but when we have qdisc sharing that no longer makes any sense. Signed-off-by: David S. Miller --- net/sched/sch_generic.c | 18 ++++++++++-------- 1 file changed, 10 insertions(+), 8 deletions(-) (limited to 'net') diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 2bd75befa066..ac208c2b2d10 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -130,8 +130,8 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, /* * NOTE: Called under queue->lock with locally disabled BH. * - * __QUEUE_STATE_QDISC_RUNNING guarantees only one CPU can process - * this queue at a time. queue->lock serializes queue accesses for + * __QDISC_STATE_RUNNING guarantees only one CPU can process + * this qdisc at a time. queue->lock serializes queue accesses for * this queue AND txq->qdisc pointer itself. * * netif_tx_lock serializes accesses to device driver. @@ -146,9 +146,9 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * >0 - queue is not empty. * */ -static inline int qdisc_restart(struct netdev_queue *txq) +static inline int qdisc_restart(struct netdev_queue *txq, + struct Qdisc *q) { - struct Qdisc *q = txq->qdisc; int ret = NETDEV_TX_BUSY; struct net_device *dev; struct sk_buff *skb; @@ -168,7 +168,6 @@ static inline int qdisc_restart(struct netdev_queue *txq) HARD_TX_UNLOCK(dev, txq); spin_lock(&txq->lock); - q = txq->qdisc; switch (ret) { case NETDEV_TX_OK: @@ -197,8 +196,9 @@ static inline int qdisc_restart(struct netdev_queue *txq) void __qdisc_run(struct netdev_queue *txq) { unsigned long start_time = jiffies; + struct Qdisc *q = txq->qdisc; - while (qdisc_restart(txq)) { + while (qdisc_restart(txq, q)) { if (netif_tx_queue_stopped(txq)) break; @@ -213,7 +213,7 @@ void __qdisc_run(struct netdev_queue *txq) } } - clear_bit(__QUEUE_STATE_QDISC_RUNNING, &txq->state); + clear_bit(__QDISC_STATE_RUNNING, &q->state); } static void dev_watchdog(unsigned long arg) @@ -666,14 +666,16 @@ static bool some_qdisc_is_running(struct net_device *dev, int lock) for (i = 0; i < dev->num_tx_queues; i++) { struct netdev_queue *dev_queue; + struct Qdisc *q; int val; dev_queue = netdev_get_tx_queue(dev, i); + q = dev_queue->qdisc; if (lock) spin_lock_bh(&dev_queue->lock); - val = test_bit(__QUEUE_STATE_QDISC_RUNNING, &dev_queue->state); + val = test_bit(__QDISC_STATE_RUNNING, &q->state); if (lock) spin_unlock_bh(&dev_queue->lock); -- cgit v1.2.2 From 7698b4fcabcd790efc4f226bada1e7b5870653af Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 16 Jul 2008 01:42:40 -0700 Subject: pkt_sched: Add and use qdisc_root() and qdisc_root_lock(). When code wants to lock the qdisc tree state, the logic operation it's doing is locking the top-level qdisc that sits of the root of the netdev_queue. Add qdisc_root_lock() to represent this and convert the easiest cases. In order for this to work out in all cases, we have to hook up the noop_qdisc to a dummy netdev_queue. Signed-off-by: David S. Miller --- net/sched/sch_api.c | 8 ++++---- net/sched/sch_cbq.c | 9 +++++---- net/sched/sch_generic.c | 21 ++++++++++++++++----- net/sched/sch_hfsc.c | 4 ++-- net/sched/sch_htb.c | 18 ++++++++++-------- net/sched/sch_netem.c | 9 ++++++--- 6 files changed, 43 insertions(+), 26 deletions(-) (limited to 'net') diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 830ccc544a15..19c244a00839 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -633,7 +633,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, if (!ops->init || (err = ops->init(sch, tca[TCA_OPTIONS])) == 0) { if (tca[TCA_RATE]) { err = gen_new_estimator(&sch->bstats, &sch->rate_est, - &sch->dev_queue->lock, + qdisc_root_lock(sch), tca[TCA_RATE]); if (err) { /* @@ -675,7 +675,7 @@ static int qdisc_change(struct Qdisc *sch, struct nlattr **tca) } if (tca[TCA_RATE]) gen_replace_estimator(&sch->bstats, &sch->rate_est, - &sch->dev_queue->lock, tca[TCA_RATE]); + qdisc_root_lock(sch), tca[TCA_RATE]); return 0; } @@ -967,7 +967,7 @@ static int tc_fill_qdisc(struct sk_buff *skb, struct Qdisc *q, u32 clid, q->qstats.qlen = q->q.qlen; if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS, - TCA_XSTATS, &q->dev_queue->lock, &d) < 0) + TCA_XSTATS, qdisc_root_lock(q), &d) < 0) goto nla_put_failure; if (q->ops->dump_stats && q->ops->dump_stats(q, &d) < 0) @@ -1216,7 +1216,7 @@ static int tc_fill_tclass(struct sk_buff *skb, struct Qdisc *q, goto nla_put_failure; if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS, - TCA_XSTATS, &q->dev_queue->lock, &d) < 0) + TCA_XSTATS, qdisc_root_lock(q), &d) < 0) goto nla_put_failure; if (cl_ops->dump_stats && cl_ops->dump_stats(q, cl, &d) < 0) diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 4efc836cbf38..37ae653db683 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -1744,12 +1744,13 @@ static void cbq_put(struct Qdisc *sch, unsigned long arg) if (--cl->refcnt == 0) { #ifdef CONFIG_NET_CLS_ACT + spinlock_t *root_lock = qdisc_root_lock(sch); struct cbq_sched_data *q = qdisc_priv(sch); - spin_lock_bh(&sch->dev_queue->lock); + spin_lock_bh(root_lock); if (q->rx_class == cl) q->rx_class = NULL; - spin_unlock_bh(&sch->dev_queue->lock); + spin_unlock_bh(root_lock); #endif cbq_destroy_class(sch, cl); @@ -1828,7 +1829,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &sch->dev_queue->lock, + qdisc_root_lock(sch), tca[TCA_RATE]); return 0; } @@ -1919,7 +1920,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t if (tca[TCA_RATE]) gen_new_estimator(&cl->bstats, &cl->rate_est, - &sch->dev_queue->lock, tca[TCA_RATE]); + qdisc_root_lock(sch), tca[TCA_RATE]); *arg = (unsigned long)cl; return 0; diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index ac208c2b2d10..739a8711ab30 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -151,14 +151,17 @@ static inline int qdisc_restart(struct netdev_queue *txq, { int ret = NETDEV_TX_BUSY; struct net_device *dev; + spinlock_t *root_lock; struct sk_buff *skb; /* Dequeue packet */ if (unlikely((skb = dequeue_skb(q)) == NULL)) return 0; - /* And release queue */ - spin_unlock(&txq->lock); + root_lock = qdisc_root_lock(q); + + /* And release qdisc */ + spin_unlock(root_lock); dev = txq->dev; @@ -167,7 +170,7 @@ static inline int qdisc_restart(struct netdev_queue *txq, ret = dev_hard_start_xmit(skb, dev, txq); HARD_TX_UNLOCK(dev, txq); - spin_lock(&txq->lock); + spin_lock(root_lock); switch (ret) { case NETDEV_TX_OK: @@ -345,12 +348,18 @@ struct Qdisc_ops noop_qdisc_ops __read_mostly = { .owner = THIS_MODULE, }; +static struct netdev_queue noop_netdev_queue = { + .lock = __SPIN_LOCK_UNLOCKED(noop_netdev_queue.lock), + .qdisc = &noop_qdisc, +}; + struct Qdisc noop_qdisc = { .enqueue = noop_enqueue, .dequeue = noop_dequeue, .flags = TCQ_F_BUILTIN, .ops = &noop_qdisc_ops, .list = LIST_HEAD_INIT(noop_qdisc.list), + .dev_queue = &noop_netdev_queue, }; EXPORT_SYMBOL(noop_qdisc); @@ -666,19 +675,21 @@ static bool some_qdisc_is_running(struct net_device *dev, int lock) for (i = 0; i < dev->num_tx_queues; i++) { struct netdev_queue *dev_queue; + spinlock_t *root_lock; struct Qdisc *q; int val; dev_queue = netdev_get_tx_queue(dev, i); q = dev_queue->qdisc; + root_lock = qdisc_root_lock(q); if (lock) - spin_lock_bh(&dev_queue->lock); + spin_lock_bh(root_lock); val = test_bit(__QDISC_STATE_RUNNING, &q->state); if (lock) - spin_unlock_bh(&dev_queue->lock); + spin_unlock_bh(root_lock); if (val) return true; diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 997d520ca580..5090708ba384 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1045,7 +1045,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &sch->dev_queue->lock, + qdisc_root_lock(sch), tca[TCA_RATE]); return 0; } @@ -1104,7 +1104,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid, if (tca[TCA_RATE]) gen_new_estimator(&cl->bstats, &cl->rate_est, - &sch->dev_queue->lock, tca[TCA_RATE]); + qdisc_root_lock(sch), tca[TCA_RATE]); *arg = (unsigned long)cl; return 0; } diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index c8ca54cc26b0..ee48457eaa4a 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -1039,11 +1039,12 @@ static int htb_init(struct Qdisc *sch, struct nlattr *opt) static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) { + spinlock_t *root_lock = qdisc_root_lock(sch); struct htb_sched *q = qdisc_priv(sch); struct nlattr *nest; struct tc_htb_glob gopt; - spin_lock_bh(&sch->dev_queue->lock); + spin_lock_bh(root_lock); gopt.direct_pkts = q->direct_pkts; gopt.version = HTB_VER; @@ -1057,11 +1058,11 @@ static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) NLA_PUT(skb, TCA_HTB_INIT, sizeof(gopt), &gopt); nla_nest_end(skb, nest); - spin_unlock_bh(&sch->dev_queue->lock); + spin_unlock_bh(root_lock); return skb->len; nla_put_failure: - spin_unlock_bh(&sch->dev_queue->lock); + spin_unlock_bh(root_lock); nla_nest_cancel(skb, nest); return -1; } @@ -1070,10 +1071,11 @@ static int htb_dump_class(struct Qdisc *sch, unsigned long arg, struct sk_buff *skb, struct tcmsg *tcm) { struct htb_class *cl = (struct htb_class *)arg; + spinlock_t *root_lock = qdisc_root_lock(sch); struct nlattr *nest; struct tc_htb_opt opt; - spin_lock_bh(&sch->dev_queue->lock); + spin_lock_bh(root_lock); tcm->tcm_parent = cl->parent ? cl->parent->common.classid : TC_H_ROOT; tcm->tcm_handle = cl->common.classid; if (!cl->level && cl->un.leaf.q) @@ -1095,11 +1097,11 @@ static int htb_dump_class(struct Qdisc *sch, unsigned long arg, NLA_PUT(skb, TCA_HTB_PARMS, sizeof(opt), &opt); nla_nest_end(skb, nest); - spin_unlock_bh(&sch->dev_queue->lock); + spin_unlock_bh(root_lock); return skb->len; nla_put_failure: - spin_unlock_bh(&sch->dev_queue->lock); + spin_unlock_bh(root_lock); nla_nest_cancel(skb, nest); return -1; } @@ -1365,7 +1367,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, goto failure; gen_new_estimator(&cl->bstats, &cl->rate_est, - &sch->dev_queue->lock, + qdisc_root_lock(sch), tca[TCA_RATE] ? : &est.nla); cl->refcnt = 1; cl->children = 0; @@ -1420,7 +1422,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid, } else { if (tca[TCA_RATE]) gen_replace_estimator(&cl->bstats, &cl->rate_est, - &sch->dev_queue->lock, + qdisc_root_lock(sch), tca[TCA_RATE]); sch_tree_lock(sch); } diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index bc585f2089ff..c5ea40c9eb21 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -180,7 +180,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) * skb will be queued. */ if (count > 1 && (skb2 = skb_clone(skb, GFP_ATOMIC)) != NULL) { - struct Qdisc *rootq = sch->dev_queue->qdisc; + struct Qdisc *rootq = qdisc_root(sch); u32 dupsave = q->duplicate; /* prevent duplicating a dup... */ q->duplicate = 0; @@ -319,6 +319,7 @@ static int get_dist_table(struct Qdisc *sch, const struct nlattr *attr) struct netem_sched_data *q = qdisc_priv(sch); unsigned long n = nla_len(attr)/sizeof(__s16); const __s16 *data = nla_data(attr); + spinlock_t *root_lock; struct disttable *d; int i; @@ -333,9 +334,11 @@ static int get_dist_table(struct Qdisc *sch, const struct nlattr *attr) for (i = 0; i < n; i++) d->table[i] = data[i]; - spin_lock_bh(&sch->dev_queue->lock); + root_lock = qdisc_root_lock(sch); + + spin_lock_bh(root_lock); d = xchg(&q->delay_dist, d); - spin_unlock_bh(&sch->dev_queue->lock); + spin_unlock_bh(root_lock); kfree(d); return 0; -- cgit v1.2.2 From 37437bb2e1ae8af470dfcd5b4ff454110894ccaf Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 16 Jul 2008 02:15:04 -0700 Subject: pkt_sched: Schedule qdiscs instead of netdev_queue. When we have shared qdiscs, packets come out of the qdiscs for multiple transmit queues. Therefore it doesn't make any sense to schedule the transmit queue when logically we cannot know ahead of time the TX queue of the SKB that the qdisc->dequeue() will give us. Just for sanity I added a BUG check to make sure we never get into a state where the noop_qdisc is scheduled. Signed-off-by: David S. Miller --- net/core/dev.c | 68 ++++++++++++++++++++----------------------------- net/sched/sch_api.c | 3 +-- net/sched/sch_cbq.c | 2 +- net/sched/sch_generic.c | 30 ++++++++++------------ 4 files changed, 44 insertions(+), 59 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 467bfb325123..0b909b74f698 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1323,18 +1323,18 @@ static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev) } -void __netif_schedule(struct netdev_queue *txq) +void __netif_schedule(struct Qdisc *q) { - struct net_device *dev = txq->dev; + BUG_ON(q == &noop_qdisc); - if (!test_and_set_bit(__LINK_STATE_SCHED, &dev->state)) { + if (!test_and_set_bit(__QDISC_STATE_SCHED, &q->state)) { struct softnet_data *sd; unsigned long flags; local_irq_save(flags); sd = &__get_cpu_var(softnet_data); - txq->next_sched = sd->output_queue; - sd->output_queue = txq; + q->next_sched = sd->output_queue; + sd->output_queue = q; raise_softirq_irqoff(NET_TX_SOFTIRQ); local_irq_restore(flags); } @@ -1771,37 +1771,23 @@ gso: rcu_read_lock_bh(); txq = dev_pick_tx(dev, skb); - spin_lock_prefetch(&txq->lock); - - /* Updates of qdisc are serialized by queue->lock. - * The struct Qdisc which is pointed to by qdisc is now a - * rcu structure - it may be accessed without acquiring - * a lock (but the structure may be stale.) The freeing of the - * qdisc will be deferred until it's known that there are no - * more references to it. - * - * If the qdisc has an enqueue function, we still need to - * hold the queue->lock before calling it, since queue->lock - * also serializes access to the device queue. - */ - q = rcu_dereference(txq->qdisc); + #ifdef CONFIG_NET_CLS_ACT skb->tc_verd = SET_TC_AT(skb->tc_verd,AT_EGRESS); #endif if (q->enqueue) { - /* Grab device queue */ - spin_lock(&txq->lock); - q = txq->qdisc; - if (q->enqueue) { - rc = q->enqueue(skb, q); - qdisc_run(txq); - spin_unlock(&txq->lock); - - rc = rc == NET_XMIT_BYPASS ? NET_XMIT_SUCCESS : rc; - goto out; - } - spin_unlock(&txq->lock); + spinlock_t *root_lock = qdisc_root_lock(q); + + spin_lock(root_lock); + + rc = q->enqueue(skb, q); + qdisc_run(q); + + spin_unlock(root_lock); + + rc = rc == NET_XMIT_BYPASS ? NET_XMIT_SUCCESS : rc; + goto out; } /* The device has no queue. Common case for software devices: @@ -1974,7 +1960,7 @@ static void net_tx_action(struct softirq_action *h) } if (sd->output_queue) { - struct netdev_queue *head; + struct Qdisc *head; local_irq_disable(); head = sd->output_queue; @@ -1982,18 +1968,20 @@ static void net_tx_action(struct softirq_action *h) local_irq_enable(); while (head) { - struct netdev_queue *txq = head; - struct net_device *dev = txq->dev; + struct Qdisc *q = head; + spinlock_t *root_lock; + head = head->next_sched; smp_mb__before_clear_bit(); - clear_bit(__LINK_STATE_SCHED, &dev->state); + clear_bit(__QDISC_STATE_SCHED, &q->state); - if (spin_trylock(&txq->lock)) { - qdisc_run(txq); - spin_unlock(&txq->lock); + root_lock = qdisc_root_lock(q); + if (spin_trylock(root_lock)) { + qdisc_run(q); + spin_unlock(root_lock); } else { - netif_schedule_queue(txq); + __netif_schedule(q); } } } @@ -4459,7 +4447,7 @@ static int dev_cpu_callback(struct notifier_block *nfb, void *ocpu) { struct sk_buff **list_skb; - struct netdev_queue **list_net; + struct Qdisc **list_net; struct sk_buff *skb; unsigned int cpu, oldcpu = (unsigned long)ocpu; struct softnet_data *sd, *oldsd; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 19c244a00839..8e8c5becc348 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -294,11 +294,10 @@ static enum hrtimer_restart qdisc_watchdog(struct hrtimer *timer) { struct qdisc_watchdog *wd = container_of(timer, struct qdisc_watchdog, timer); - struct netdev_queue *txq = wd->qdisc->dev_queue; wd->qdisc->flags &= ~TCQ_F_THROTTLED; smp_wmb(); - netif_schedule_queue(txq); + __netif_schedule(wd->qdisc); return HRTIMER_NORESTART; } diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 37ae653db683..a3953bbe2d79 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -650,7 +650,7 @@ static enum hrtimer_restart cbq_undelay(struct hrtimer *timer) } sch->flags &= ~TCQ_F_THROTTLED; - netif_schedule_queue(sch->dev_queue); + __netif_schedule(sch); return HRTIMER_NORESTART; } diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 739a8711ab30..dd5c4e70abe4 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -72,16 +72,14 @@ static inline int qdisc_qlen(struct Qdisc *q) return q->q.qlen; } -static inline int dev_requeue_skb(struct sk_buff *skb, - struct netdev_queue *dev_queue, - struct Qdisc *q) +static inline int dev_requeue_skb(struct sk_buff *skb, struct Qdisc *q) { if (unlikely(skb->next)) q->gso_skb = skb; else q->ops->requeue(skb, q); - netif_schedule_queue(dev_queue); + __netif_schedule(q); return 0; } @@ -121,7 +119,7 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * some time. */ __get_cpu_var(netdev_rx_stat).cpu_collision++; - ret = dev_requeue_skb(skb, dev_queue, q); + ret = dev_requeue_skb(skb, q); } return ret; @@ -146,9 +144,9 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, * >0 - queue is not empty. * */ -static inline int qdisc_restart(struct netdev_queue *txq, - struct Qdisc *q) +static inline int qdisc_restart(struct Qdisc *q) { + struct netdev_queue *txq; int ret = NETDEV_TX_BUSY; struct net_device *dev; spinlock_t *root_lock; @@ -163,7 +161,8 @@ static inline int qdisc_restart(struct netdev_queue *txq, /* And release qdisc */ spin_unlock(root_lock); - dev = txq->dev; + dev = qdisc_dev(q); + txq = netdev_get_tx_queue(dev, skb_get_queue_mapping(skb)); HARD_TX_LOCK(dev, txq, smp_processor_id()); if (!netif_subqueue_stopped(dev, skb)) @@ -189,29 +188,28 @@ static inline int qdisc_restart(struct netdev_queue *txq, printk(KERN_WARNING "BUG %s code %d qlen %d\n", dev->name, ret, q->q.qlen); - ret = dev_requeue_skb(skb, txq, q); + ret = dev_requeue_skb(skb, q); break; } + if (ret && netif_tx_queue_stopped(txq)) + ret = 0; + return ret; } -void __qdisc_run(struct netdev_queue *txq) +void __qdisc_run(struct Qdisc *q) { unsigned long start_time = jiffies; - struct Qdisc *q = txq->qdisc; - - while (qdisc_restart(txq, q)) { - if (netif_tx_queue_stopped(txq)) - break; + while (qdisc_restart(q)) { /* * Postpone processing if * 1. another process needs the CPU; * 2. we've been doing it for too long. */ if (need_resched() || jiffies != start_time) { - netif_schedule_queue(txq); + __netif_schedule(q); break; } } -- cgit v1.2.2 From 16361127ebed0fb8f9d7cc94c6e137eaf710f676 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 16 Jul 2008 02:23:17 -0700 Subject: pkt_sched: dev_init_scheduler() does not need to lock qdisc tree. We are registering the device, there is no way anyone can get at this object's qdiscs yet in any meaningful way. Signed-off-by: David S. Miller --- net/sched/sch_generic.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'net') diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index dd5c4e70abe4..7e078c593194 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -741,10 +741,8 @@ static void dev_init_scheduler_queue(struct net_device *dev, void dev_init_scheduler(struct net_device *dev) { - qdisc_lock_tree(dev); netdev_for_each_tx_queue(dev, dev_init_scheduler_queue, &noop_qdisc); dev_init_scheduler_queue(dev, &dev->rx_queue, NULL); - qdisc_unlock_tree(dev); setup_timer(&dev->watchdog_timer, dev_watchdog, (unsigned long)dev); } -- cgit v1.2.2 From 8a34c5dc3a7c6431f1cd94c0904be81b296e08ca Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 00:47:45 -0700 Subject: pkt_sched: Perform bulk of qdisc destruction in RCU. This allows less strict control of access to the qdisc attached to a netdev_queue. It is even allowed to enqueue into a qdisc which is in the process of being destroyed. The RCU handler will toss out those packets. We will need this to handle sharing of a qdisc amongst multiple TX queues. In such a setup the lock has to be shared, so will be inside of the qdisc itself. At which point the netdev_queue lock cannot be used to hard synchronize access to the ->qdisc pointer. One operation we have to keep inside of qdisc_destroy() is the list deletion. It is the only piece of state visible after the RCU quiesce period, so we have to undo it early and under the appropriate locking. The operations in the RCU handler do not need any looking because the qdisc tree is no longer visible to anything at that point. Signed-off-by: David S. Miller --- net/sched/sch_generic.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 7e078c593194..082db8abe703 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -545,6 +545,17 @@ EXPORT_SYMBOL(qdisc_reset); static void __qdisc_destroy(struct rcu_head *head) { struct Qdisc *qdisc = container_of(head, struct Qdisc, q_rcu); + const struct Qdisc_ops *ops = qdisc->ops; + + gen_kill_estimator(&qdisc->bstats, &qdisc->rate_est); + if (ops->reset) + ops->reset(qdisc); + if (ops->destroy) + ops->destroy(qdisc); + + module_put(ops->owner); + dev_put(qdisc_dev(qdisc)); + kfree((char *) qdisc - qdisc->padded); } @@ -552,21 +563,12 @@ static void __qdisc_destroy(struct rcu_head *head) void qdisc_destroy(struct Qdisc *qdisc) { - const struct Qdisc_ops *ops = qdisc->ops; - if (qdisc->flags & TCQ_F_BUILTIN || !atomic_dec_and_test(&qdisc->refcnt)) return; list_del(&qdisc->list); - gen_kill_estimator(&qdisc->bstats, &qdisc->rate_est); - if (ops->reset) - ops->reset(qdisc); - if (ops->destroy) - ops->destroy(qdisc); - module_put(ops->owner); - dev_put(qdisc_dev(qdisc)); call_rcu(&qdisc->q_rcu, __qdisc_destroy); } EXPORT_SYMBOL(qdisc_destroy); -- cgit v1.2.2 From 17715e62a5e5c7224e5f906a4b8f9e5084100118 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 16 Jul 2008 02:36:04 -0700 Subject: pkt_sched: Use per-queue locking in shutdown_scheduler_queue. This eliminates another qdisc_lock_tree user. Signed-off-by: David S. Miller --- net/sched/sch_generic.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 082db8abe703..efa418a1b34e 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -757,18 +757,20 @@ static void shutdown_scheduler_queue(struct net_device *dev, struct Qdisc *qdisc_default = _qdisc_default; if (qdisc) { + spinlock_t *root_lock = qdisc_root_lock(qdisc); + dev_queue->qdisc = qdisc_default; dev_queue->qdisc_sleeping = qdisc_default; + spin_lock(root_lock); qdisc_destroy(qdisc); + spin_unlock(root_lock); } } void dev_shutdown(struct net_device *dev) { - qdisc_lock_tree(dev); netdev_for_each_tx_queue(dev, shutdown_scheduler_queue, &noop_qdisc); shutdown_scheduler_queue(dev, &dev->rx_queue, NULL); BUG_TRAP(!timer_pending(&dev->watchdog_timer)); - qdisc_unlock_tree(dev); } -- cgit v1.2.2 From 55dbc640c31db373fa07eb1e3af9b8eadbdf80db Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 16 Jul 2008 02:40:45 -0700 Subject: pkt_sched: Remove qdisc_lock_tree usage in cls_api.c It just wants the qdisc tree for the filter to be synchronized. So just BH lock qdisc_root_lock(q) instead. Signed-off-by: David S. Miller --- net/sched/cls_api.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c index d0b0a9b14394..d2b6f54a6261 100644 --- a/net/sched/cls_api.c +++ b/net/sched/cls_api.c @@ -120,6 +120,7 @@ static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n, void *arg) { struct net *net = sock_net(skb->sk); struct nlattr *tca[TCA_MAX + 1]; + spinlock_t *root_lock; struct tcmsg *t; u32 protocol; u32 prio; @@ -204,6 +205,8 @@ replay: } } + root_lock = qdisc_root_lock(q); + if (tp == NULL) { /* Proto-tcf does not exist, create new one */ @@ -263,10 +266,10 @@ replay: goto errout; } - qdisc_lock_tree(dev); + spin_lock_bh(root_lock); tp->next = *back; *back = tp; - qdisc_unlock_tree(dev); + spin_unlock_bh(root_lock); } else if (tca[TCA_KIND] && nla_strcmp(tca[TCA_KIND], tp->ops->kind)) goto errout; @@ -275,9 +278,9 @@ replay: if (fh == 0) { if (n->nlmsg_type == RTM_DELTFILTER && t->tcm_handle == 0) { - qdisc_lock_tree(dev); + spin_lock_bh(root_lock); *back = tp->next; - qdisc_unlock_tree(dev); + spin_lock_bh(root_lock); tfilter_notify(skb, n, tp, fh, RTM_DELTFILTER); tcf_destroy(tp); -- cgit v1.2.2 From 15b458fa65cbba395724a99ab1b7d3785ca76c1c Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 16 Jul 2008 02:42:51 -0700 Subject: pkt_sched: Kill qdisc_lock_tree usage in cls_route.c It just wants the qdisc tree to be synchronized, so grabbing qdisc_root_lock() is sufficient. Signed-off-by: David S. Miller --- net/sched/cls_route.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/sched/cls_route.c b/net/sched/cls_route.c index 5a16ca28aa3d..481260a4f10f 100644 --- a/net/sched/cls_route.c +++ b/net/sched/cls_route.c @@ -73,11 +73,13 @@ static __inline__ int route4_fastmap_hash(u32 id, int iif) } static inline -void route4_reset_fastmap(struct net_device *dev, struct route4_head *head, u32 id) +void route4_reset_fastmap(struct Qdisc *q, struct route4_head *head, u32 id) { - qdisc_lock_tree(dev); + spinlock_t *root_lock = qdisc_root_lock(q); + + spin_lock_bh(root_lock); memset(head->fastmap, 0, sizeof(head->fastmap)); - qdisc_unlock_tree(dev); + spin_unlock_bh(root_lock); } static inline void @@ -302,7 +304,7 @@ static int route4_delete(struct tcf_proto *tp, unsigned long arg) *fp = f->next; tcf_tree_unlock(tp); - route4_reset_fastmap(qdisc_dev(tp->q), head, f->id); + route4_reset_fastmap(tp->q, head, f->id); route4_delete_filter(tp, f); /* Strip tree */ @@ -500,7 +502,7 @@ reinsert: } tcf_tree_unlock(tp); - route4_reset_fastmap(qdisc_dev(tp->q), head, f->id); + route4_reset_fastmap(tp->q, head, f->id); *arg = (unsigned long)f; return 0; -- cgit v1.2.2 From ead81cc5fc6d996db6afb20f211241612610a07a Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 00:50:32 -0700 Subject: netdevice: Move qdisc_list back into net_device proper. And give it it's own lock. Signed-off-by: David S. Miller --- net/core/dev.c | 2 ++ net/sched/sch_api.c | 31 +++++++------------------------ net/sched/sch_generic.c | 9 +++++++-- 3 files changed, 16 insertions(+), 26 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 0b909b74f698..6741e344ac59 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3886,6 +3886,8 @@ int register_netdevice(struct net_device *dev) net = dev_net(dev); spin_lock_init(&dev->addr_list_lock); + spin_lock_init(&dev->qdisc_list_lock); + INIT_LIST_HEAD(&dev->qdisc_list); netdev_init_queue_locks(dev); dev->iflink = -1; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 8e8c5becc348..6958fe7c9a77 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -183,30 +183,17 @@ EXPORT_SYMBOL(unregister_qdisc); (root qdisc, all its children, children of children etc.) */ -static struct Qdisc *__qdisc_lookup(struct netdev_queue *dev_queue, u32 handle) +struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) { struct Qdisc *q; - list_for_each_entry(q, &dev_queue->qdisc_list, list) { + list_for_each_entry(q, &dev->qdisc_list, list) { if (q->handle == handle) return q; } return NULL; } -struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) -{ - unsigned int i; - - for (i = 0; i < dev->num_tx_queues; i++) { - struct netdev_queue *txq = netdev_get_tx_queue(dev, i); - struct Qdisc *q = __qdisc_lookup(txq, handle); - if (q) - return q; - } - return NULL; -} - static struct Qdisc *qdisc_leaf(struct Qdisc *p, u32 classid) { unsigned long cl; @@ -645,9 +632,9 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, goto err_out3; } } - qdisc_lock_tree(dev); - list_add_tail(&sch->list, &dev_queue->qdisc_list); - qdisc_unlock_tree(dev); + spin_lock_bh(&dev->qdisc_list_lock); + list_add_tail(&sch->list, &dev->qdisc_list); + spin_unlock_bh(&dev->qdisc_list_lock); return sch; } @@ -1032,14 +1019,12 @@ static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb) read_lock(&dev_base_lock); idx = 0; for_each_netdev(&init_net, dev) { - struct netdev_queue *dev_queue; if (idx < s_idx) goto cont; if (idx > s_idx) s_q_idx = 0; q_idx = 0; - dev_queue = netdev_get_tx_queue(dev, 0); - list_for_each_entry(q, &dev_queue->qdisc_list, list) { + list_for_each_entry(q, &dev->qdisc_list, list) { if (q_idx < s_q_idx) { q_idx++; continue; @@ -1269,7 +1254,6 @@ static int qdisc_class_dump(struct Qdisc *q, unsigned long cl, struct qdisc_walk static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) { struct net *net = sock_net(skb->sk); - struct netdev_queue *dev_queue; int t; int s_t; struct net_device *dev; @@ -1288,8 +1272,7 @@ static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) s_t = cb->args[0]; t = 0; - dev_queue = netdev_get_tx_queue(dev, 0); - list_for_each_entry(q, &dev_queue->qdisc_list, list) { + list_for_each_entry(q, &dev->qdisc_list, list) { if (t < s_t || !q->ops->cl_ops || (tcm->tcm_parent && TC_H_MAJ(tcm->tcm_parent) != q->handle)) { diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index efa418a1b34e..8cdf0b4a6a5a 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -563,11 +563,15 @@ static void __qdisc_destroy(struct rcu_head *head) void qdisc_destroy(struct Qdisc *qdisc) { + struct net_device *dev = qdisc_dev(qdisc); + if (qdisc->flags & TCQ_F_BUILTIN || !atomic_dec_and_test(&qdisc->refcnt)) return; + spin_lock_bh(&dev->qdisc_list_lock); list_del(&qdisc->list); + spin_unlock_bh(&dev->qdisc_list_lock); call_rcu(&qdisc->q_rcu, __qdisc_destroy); } @@ -599,7 +603,9 @@ static void attach_one_default_qdisc(struct net_device *dev, printk(KERN_INFO "%s: activation failed\n", dev->name); return; } - list_add_tail(&qdisc->list, &dev_queue->qdisc_list); + spin_lock_bh(&dev->qdisc_list_lock); + list_add_tail(&qdisc->list, &dev->qdisc_list); + spin_unlock_bh(&dev->qdisc_list_lock); } else { qdisc = &noqueue_qdisc; } @@ -738,7 +744,6 @@ static void dev_init_scheduler_queue(struct net_device *dev, dev_queue->qdisc = qdisc; dev_queue->qdisc_sleeping = qdisc; - INIT_LIST_HEAD(&dev_queue->qdisc_list); } void dev_init_scheduler(struct net_device *dev) -- cgit v1.2.2 From 53049978df1d9ae55bf397c9879e6b33218352db Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 16 Jul 2008 03:00:19 -0700 Subject: pkt_sched: Make qdisc grafting locking more specific. Lock the root of the qdisc being operated upon. All explicit references to qdisc_tree_lock() are now gone. The only remaining uses are via the sch_tree_{lock,unlock}() and tcf_tree_{lock,unlock}() macros. Signed-off-by: David S. Miller --- net/sched/sch_api.c | 41 +++++++++++++++++++++++++++++------------ 1 file changed, 29 insertions(+), 12 deletions(-) (limited to 'net') diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 6958fe7c9a77..74924893ef7f 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -441,15 +441,29 @@ static struct Qdisc * dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) { struct netdev_queue *dev_queue; + spinlock_t *root_lock; struct Qdisc *oqdisc; + int ingress; if (dev->flags & IFF_UP) dev_deactivate(dev); - qdisc_lock_tree(dev); - if (qdisc && qdisc->flags&TCQ_F_INGRESS) { + ingress = 0; + if (qdisc && qdisc->flags&TCQ_F_INGRESS) + ingress = 1; + + if (ingress) { dev_queue = &dev->rx_queue; oqdisc = dev_queue->qdisc; + } else { + dev_queue = netdev_get_tx_queue(dev, 0); + oqdisc = dev_queue->qdisc_sleeping; + } + + root_lock = qdisc_root_lock(oqdisc); + spin_lock_bh(root_lock); + + if (ingress) { /* Prune old scheduler */ if (oqdisc && atomic_read(&oqdisc->refcnt) <= 1) { /* delete */ @@ -460,9 +474,6 @@ dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) } } else { - dev_queue = netdev_get_tx_queue(dev, 0); - oqdisc = dev_queue->qdisc_sleeping; - /* Prune old scheduler */ if (oqdisc && atomic_read(&oqdisc->refcnt) <= 1) qdisc_reset(oqdisc); @@ -474,7 +485,7 @@ dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) dev_queue->qdisc = &noop_qdisc; } - qdisc_unlock_tree(dev); + spin_unlock_bh(root_lock); if (dev->flags & IFF_UP) dev_activate(dev); @@ -765,10 +776,12 @@ static int tc_get_qdisc(struct sk_buff *skb, struct nlmsghdr *n, void *arg) if ((err = qdisc_graft(dev, p, clid, NULL, &q)) != 0) return err; if (q) { + spinlock_t *root_lock = qdisc_root_lock(q); + qdisc_notify(skb, n, clid, q, NULL); - qdisc_lock_tree(dev); + spin_unlock_bh(root_lock); qdisc_destroy(q); - qdisc_unlock_tree(dev); + spin_unlock_bh(root_lock); } } else { qdisc_notify(skb, n, clid, NULL, q); @@ -911,20 +924,24 @@ create_n_graft: graft: if (1) { struct Qdisc *old_q = NULL; + spinlock_t *root_lock; + err = qdisc_graft(dev, p, clid, q, &old_q); if (err) { if (q) { - qdisc_lock_tree(dev); + root_lock = qdisc_root_lock(q); + spin_lock_bh(root_lock); qdisc_destroy(q); - qdisc_unlock_tree(dev); + spin_unlock_bh(root_lock); } return err; } qdisc_notify(skb, n, clid, old_q, q); if (old_q) { - qdisc_lock_tree(dev); + root_lock = qdisc_root_lock(old_q); + spin_lock_bh(root_lock); qdisc_destroy(old_q); - qdisc_unlock_tree(dev); + spin_unlock_bh(root_lock); } } return 0; -- cgit v1.2.2 From c7e4f3bbb4ba4e48ab3b529d5016e454cee1ccd6 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 16 Jul 2008 03:22:39 -0700 Subject: pkt_sched: Kill qdisc_lock_tree and qdisc_unlock_tree. No longer used. Signed-off-by: David S. Miller --- net/sched/sch_generic.c | 36 +++--------------------------------- 1 file changed, 3 insertions(+), 33 deletions(-) (limited to 'net') diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 8cdf0b4a6a5a..3d53e92ad9c8 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -29,44 +29,14 @@ /* Main transmission queue. */ /* Modifications to data participating in scheduling must be protected with - * queue->lock spinlock. + * qdisc_root_lock(qdisc) spinlock. * * The idea is the following: - * - enqueue, dequeue are serialized via top level device - * spinlock queue->lock. - * - ingress filtering is serialized via top level device - * spinlock dev->rx_queue.lock. + * - enqueue, dequeue are serialized via qdisc root lock + * - ingress filtering is also serialized via qdisc root lock * - updates to tree and tree walking are only done under the rtnl mutex. */ -void qdisc_lock_tree(struct net_device *dev) - __acquires(dev->rx_queue.lock) -{ - unsigned int i; - - local_bh_disable(); - for (i = 0; i < dev->num_tx_queues; i++) { - struct netdev_queue *txq = netdev_get_tx_queue(dev, i); - spin_lock(&txq->lock); - } - spin_lock(&dev->rx_queue.lock); -} -EXPORT_SYMBOL(qdisc_lock_tree); - -void qdisc_unlock_tree(struct net_device *dev) - __releases(dev->rx_queue.lock) -{ - unsigned int i; - - spin_unlock(&dev->rx_queue.lock); - for (i = 0; i < dev->num_tx_queues; i++) { - struct netdev_queue *txq = netdev_get_tx_queue(dev, i); - spin_unlock(&txq->lock); - } - local_bh_enable(); -} -EXPORT_SYMBOL(qdisc_unlock_tree); - static inline int qdisc_qlen(struct Qdisc *q) { return q->q.qlen; -- cgit v1.2.2 From 83874000929ed63aef30b44083a9f713135ff040 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 00:53:03 -0700 Subject: pkt_sched: Kill netdev_queue lock. We can simply use the qdisc->q.lock for all of the qdisc tree synchronization. Signed-off-by: David S. Miller --- net/core/dev.c | 9 +++++---- net/mac80211/wme.c | 19 ++++++++++++------- net/sched/sch_generic.c | 32 +++++++++++++++----------------- net/sched/sch_teql.c | 7 +++++-- 4 files changed, 37 insertions(+), 30 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 6741e344ac59..32a13772c1cb 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2080,10 +2080,12 @@ static int ing_filter(struct sk_buff *skb) rxq = &dev->rx_queue; - spin_lock(&rxq->lock); - if ((q = rxq->qdisc) != NULL) + q = rxq->qdisc; + if (q) { + spin_lock(qdisc_lock(q)); result = q->enqueue(skb, q); - spin_unlock(&rxq->lock); + spin_unlock(qdisc_lock(q)); + } return result; } @@ -4173,7 +4175,6 @@ static void netdev_init_one_queue(struct net_device *dev, struct netdev_queue *queue, void *_unused) { - spin_lock_init(&queue->lock); queue->dev = dev; } diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index b21cfec4b6ce..6e8099e77043 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -237,12 +237,14 @@ void ieee80211_ht_agg_queue_remove(struct ieee80211_local *local, ieee80211_requeue(local, agg_queue); } else { struct netdev_queue *txq; + spinlock_t *root_lock; txq = netdev_get_tx_queue(local->mdev, agg_queue); + root_lock = qdisc_root_lock(txq->qdisc); - spin_lock_bh(&txq->lock); + spin_lock_bh(root_lock); qdisc_reset(txq->qdisc); - spin_unlock_bh(&txq->lock); + spin_unlock_bh(root_lock); } } @@ -250,6 +252,7 @@ void ieee80211_requeue(struct ieee80211_local *local, int queue) { struct netdev_queue *txq = netdev_get_tx_queue(local->mdev, queue); struct sk_buff_head list; + spinlock_t *root_lock; struct Qdisc *qdisc; u32 len; @@ -261,14 +264,15 @@ void ieee80211_requeue(struct ieee80211_local *local, int queue) skb_queue_head_init(&list); - spin_lock(&txq->lock); + root_lock = qdisc_root_lock(qdisc); + spin_lock(root_lock); for (len = qdisc->q.qlen; len > 0; len--) { struct sk_buff *skb = qdisc->dequeue(qdisc); if (skb) __skb_queue_tail(&list, skb); } - spin_unlock(&txq->lock); + spin_unlock(root_lock); for (len = list.qlen; len > 0; len--) { struct sk_buff *skb = __skb_dequeue(&list); @@ -280,12 +284,13 @@ void ieee80211_requeue(struct ieee80211_local *local, int queue) txq = netdev_get_tx_queue(local->mdev, new_queue); - spin_lock(&txq->lock); qdisc = rcu_dereference(txq->qdisc); - qdisc->enqueue(skb, qdisc); + root_lock = qdisc_root_lock(qdisc); - spin_unlock(&txq->lock); + spin_lock(root_lock); + qdisc->enqueue(skb, qdisc); + spin_unlock(root_lock); } out_unlock: diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 3d53e92ad9c8..8fc580b3e173 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -96,15 +96,15 @@ static inline int handle_dev_cpu_collision(struct sk_buff *skb, } /* - * NOTE: Called under queue->lock with locally disabled BH. + * NOTE: Called under qdisc_lock(q) with locally disabled BH. * * __QDISC_STATE_RUNNING guarantees only one CPU can process - * this qdisc at a time. queue->lock serializes queue accesses for - * this queue AND txq->qdisc pointer itself. + * this qdisc at a time. qdisc_lock(q) serializes queue accesses for + * this queue. * * netif_tx_lock serializes accesses to device driver. * - * queue->lock and netif_tx_lock are mutually exclusive, + * qdisc_lock(q) and netif_tx_lock are mutually exclusive, * if one is grabbed, another must be free. * * Note, that this procedure can be called by a watchdog timer @@ -317,7 +317,6 @@ struct Qdisc_ops noop_qdisc_ops __read_mostly = { }; static struct netdev_queue noop_netdev_queue = { - .lock = __SPIN_LOCK_UNLOCKED(noop_netdev_queue.lock), .qdisc = &noop_qdisc, }; @@ -327,6 +326,7 @@ struct Qdisc noop_qdisc = { .flags = TCQ_F_BUILTIN, .ops = &noop_qdisc_ops, .list = LIST_HEAD_INIT(noop_qdisc.list), + .q.lock = __SPIN_LOCK_UNLOCKED(noop_qdisc.q.lock), .dev_queue = &noop_netdev_queue, }; EXPORT_SYMBOL(noop_qdisc); @@ -498,7 +498,7 @@ errout: } EXPORT_SYMBOL(qdisc_create_dflt); -/* Under queue->lock and BH! */ +/* Under qdisc_root_lock(qdisc) and BH! */ void qdisc_reset(struct Qdisc *qdisc) { @@ -526,10 +526,12 @@ static void __qdisc_destroy(struct rcu_head *head) module_put(ops->owner); dev_put(qdisc_dev(qdisc)); + kfree_skb(qdisc->gso_skb); + kfree((char *) qdisc - qdisc->padded); } -/* Under queue->lock and BH! */ +/* Under qdisc_root_lock(qdisc) and BH! */ void qdisc_destroy(struct Qdisc *qdisc) { @@ -586,13 +588,12 @@ static void transition_one_qdisc(struct net_device *dev, struct netdev_queue *dev_queue, void *_need_watchdog) { + struct Qdisc *new_qdisc = dev_queue->qdisc_sleeping; int *need_watchdog_p = _need_watchdog; - spin_lock_bh(&dev_queue->lock); - rcu_assign_pointer(dev_queue->qdisc, dev_queue->qdisc_sleeping); - if (dev_queue->qdisc != &noqueue_qdisc) + rcu_assign_pointer(dev_queue->qdisc, new_qdisc); + if (new_qdisc != &noqueue_qdisc) *need_watchdog_p = 1; - spin_unlock_bh(&dev_queue->lock); } void dev_activate(struct net_device *dev) @@ -629,19 +630,16 @@ static void dev_deactivate_queue(struct net_device *dev, struct sk_buff *skb = NULL; struct Qdisc *qdisc; - spin_lock_bh(&dev_queue->lock); - qdisc = dev_queue->qdisc; if (qdisc) { + spin_lock_bh(qdisc_lock(qdisc)); + dev_queue->qdisc = qdisc_default; qdisc_reset(qdisc); - skb = qdisc->gso_skb; - qdisc->gso_skb = NULL; + spin_unlock_bh(qdisc_lock(qdisc)); } - spin_unlock_bh(&dev_queue->lock); - kfree_skb(skb); } diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index ade3372221c7..8b0ff345f9da 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -156,12 +156,15 @@ teql_destroy(struct Qdisc* sch) master->slaves = NEXT_SLAVE(q); if (q == master->slaves) { struct netdev_queue *txq; + spinlock_t *root_lock; txq = netdev_get_tx_queue(master->dev, 0); master->slaves = NULL; - spin_lock_bh(&txq->lock); + + root_lock = qdisc_root_lock(txq->qdisc); + spin_lock_bh(root_lock); qdisc_reset(txq->qdisc); - spin_unlock_bh(&txq->lock); + spin_unlock_bh(root_lock); } } skb_queue_purge(&dat->q); -- cgit v1.2.2 From 99194cff398d056e5ee469647c294466c246c88a Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 04:54:10 -0700 Subject: pkt_sched: Add multiqueue handling to qdisc_graft(). Move the destruction of the old queue into qdisc_graft(). When operating on a root qdisc (ie. "parent == NULL"), apply the operation to all queues. The caller has grabbed a single implicit reference for this graft, therefore when we apply the change to more than one queue we must grab additional qdisc references. Otherwise, we are operating on a class of a specific parent qdisc, and therefore no multiqueue handling is necessary. Signed-off-by: David S. Miller --- net/sched/sch_api.c | 101 ++++++++++++++++++++++++++++++---------------------- 1 file changed, 59 insertions(+), 42 deletions(-) (limited to 'net') diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 74924893ef7f..b3ef8307204e 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -435,28 +435,22 @@ static u32 qdisc_alloc_handle(struct net_device *dev) return i>0 ? autohandle : 0; } -/* Attach toplevel qdisc to device dev */ +/* Attach toplevel qdisc to device queue. */ -static struct Qdisc * -dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) +static struct Qdisc *dev_graft_qdisc(struct netdev_queue *dev_queue, + struct Qdisc *qdisc) { - struct netdev_queue *dev_queue; spinlock_t *root_lock; struct Qdisc *oqdisc; int ingress; - if (dev->flags & IFF_UP) - dev_deactivate(dev); - ingress = 0; if (qdisc && qdisc->flags&TCQ_F_INGRESS) ingress = 1; if (ingress) { - dev_queue = &dev->rx_queue; oqdisc = dev_queue->qdisc; } else { - dev_queue = netdev_get_tx_queue(dev, 0); oqdisc = dev_queue->qdisc_sleeping; } @@ -487,9 +481,6 @@ dev_graft_qdisc(struct net_device *dev, struct Qdisc *qdisc) spin_unlock_bh(root_lock); - if (dev->flags & IFF_UP) - dev_activate(dev); - return oqdisc; } @@ -521,26 +512,66 @@ void qdisc_tree_decrease_qlen(struct Qdisc *sch, unsigned int n) } EXPORT_SYMBOL(qdisc_tree_decrease_qlen); -/* Graft qdisc "new" to class "classid" of qdisc "parent" or - to device "dev". +static void notify_and_destroy(struct sk_buff *skb, struct nlmsghdr *n, u32 clid, + struct Qdisc *old, struct Qdisc *new) +{ + if (new || old) + qdisc_notify(skb, n, clid, old, new); - Old qdisc is not destroyed but returned in *old. + if (old) { + spin_lock_bh(&old->q.lock); + qdisc_destroy(old); + spin_unlock_bh(&old->q.lock); + } +} + +/* Graft qdisc "new" to class "classid" of qdisc "parent" or + * to device "dev". + * + * When appropriate send a netlink notification using 'skb' + * and "n". + * + * On success, destroy old qdisc. */ static int qdisc_graft(struct net_device *dev, struct Qdisc *parent, - u32 classid, - struct Qdisc *new, struct Qdisc **old) + struct sk_buff *skb, struct nlmsghdr *n, u32 classid, + struct Qdisc *new, struct Qdisc *old) { + struct Qdisc *q = old; int err = 0; - struct Qdisc *q = *old; - if (parent == NULL) { - if (q && q->flags&TCQ_F_INGRESS) { - *old = dev_graft_qdisc(dev, q); - } else { - *old = dev_graft_qdisc(dev, new); + unsigned int i, num_q, ingress; + + ingress = 0; + num_q = dev->num_tx_queues; + if (q && q->flags & TCQ_F_INGRESS) { + num_q = 1; + ingress = 1; } + + if (dev->flags & IFF_UP) + dev_deactivate(dev); + + for (i = 0; i < num_q; i++) { + struct netdev_queue *dev_queue = &dev->rx_queue; + + if (!ingress) + dev_queue = netdev_get_tx_queue(dev, i); + + if (ingress) { + old = dev_graft_qdisc(dev_queue, q); + } else { + old = dev_graft_qdisc(dev_queue, new); + if (new && i > 0) + atomic_inc(&new->refcnt); + } + notify_and_destroy(skb, n, classid, old, new); + } + + if (dev->flags & IFF_UP) + dev_activate(dev); } else { const struct Qdisc_class_ops *cops = parent->ops->cl_ops; @@ -549,10 +580,12 @@ static int qdisc_graft(struct net_device *dev, struct Qdisc *parent, if (cops) { unsigned long cl = cops->get(parent, classid); if (cl) { - err = cops->graft(parent, cl, new, old); + err = cops->graft(parent, cl, new, &old); cops->put(parent, cl); } } + if (!err) + notify_and_destroy(skb, n, classid, old, new); } return err; } @@ -773,16 +806,8 @@ static int tc_get_qdisc(struct sk_buff *skb, struct nlmsghdr *n, void *arg) return -EINVAL; if (q->handle == 0) return -ENOENT; - if ((err = qdisc_graft(dev, p, clid, NULL, &q)) != 0) + if ((err = qdisc_graft(dev, p, skb, n, clid, NULL, q)) != 0) return err; - if (q) { - spinlock_t *root_lock = qdisc_root_lock(q); - - qdisc_notify(skb, n, clid, q, NULL); - spin_unlock_bh(root_lock); - qdisc_destroy(q); - spin_unlock_bh(root_lock); - } } else { qdisc_notify(skb, n, clid, NULL, q); } @@ -923,10 +948,9 @@ create_n_graft: graft: if (1) { - struct Qdisc *old_q = NULL; spinlock_t *root_lock; - err = qdisc_graft(dev, p, clid, q, &old_q); + err = qdisc_graft(dev, p, skb, n, clid, q, NULL); if (err) { if (q) { root_lock = qdisc_root_lock(q); @@ -936,13 +960,6 @@ graft: } return err; } - qdisc_notify(skb, n, clid, old_q, q); - if (old_q) { - root_lock = qdisc_root_lock(old_q); - spin_lock_bh(root_lock); - qdisc_destroy(old_q); - spin_unlock_bh(root_lock); - } } return 0; } -- cgit v1.2.2 From a0c80b80e0fb48129e4e9d6a9ede914f9ff1850d Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 17 Jul 2008 01:46:06 -0700 Subject: pkt_sched: Make default qdisc nonshared-multiqueue safe. Instead of 'pfifo_fast' we have just plain 'fifo_fast'. No priority queues, just a straight FIFO. This is necessary in order to legally have a seperate qdisc per queue in multi-TX-queue setups, and thus get full parallelization. Signed-off-by: David S. Miller --- net/sched/sch_generic.c | 99 +++++++++++-------------------------------------- 1 file changed, 22 insertions(+), 77 deletions(-) (limited to 'net') diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 8fc580b3e173..e244c462e6bd 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -349,99 +349,44 @@ static struct Qdisc noqueue_qdisc = { }; -static const u8 prio2band[TC_PRIO_MAX+1] = - { 1, 2, 2, 2, 1, 2, 0, 0 , 1, 1, 1, 1, 1, 1, 1, 1 }; - -/* 3-band FIFO queue: old style, but should be a bit faster than - generic prio+fifo combination. - */ - -#define PFIFO_FAST_BANDS 3 - -static inline struct sk_buff_head *prio2list(struct sk_buff *skb, - struct Qdisc *qdisc) -{ - struct sk_buff_head *list = qdisc_priv(qdisc); - return list + prio2band[skb->priority & TC_PRIO_MAX]; -} - -static int pfifo_fast_enqueue(struct sk_buff *skb, struct Qdisc* qdisc) +static int fifo_fast_enqueue(struct sk_buff *skb, struct Qdisc* qdisc) { - struct sk_buff_head *list = prio2list(skb, qdisc); + struct sk_buff_head *list = &qdisc->q; - if (skb_queue_len(list) < qdisc_dev(qdisc)->tx_queue_len) { - qdisc->q.qlen++; + if (skb_queue_len(list) < qdisc_dev(qdisc)->tx_queue_len) return __qdisc_enqueue_tail(skb, qdisc, list); - } return qdisc_drop(skb, qdisc); } -static struct sk_buff *pfifo_fast_dequeue(struct Qdisc* qdisc) +static struct sk_buff *fifo_fast_dequeue(struct Qdisc* qdisc) { - int prio; - struct sk_buff_head *list = qdisc_priv(qdisc); + struct sk_buff_head *list = &qdisc->q; - for (prio = 0; prio < PFIFO_FAST_BANDS; prio++) { - if (!skb_queue_empty(list + prio)) { - qdisc->q.qlen--; - return __qdisc_dequeue_head(qdisc, list + prio); - } - } + if (!skb_queue_empty(list)) + return __qdisc_dequeue_head(qdisc, list); return NULL; } -static int pfifo_fast_requeue(struct sk_buff *skb, struct Qdisc* qdisc) +static int fifo_fast_requeue(struct sk_buff *skb, struct Qdisc* qdisc) { - qdisc->q.qlen++; - return __qdisc_requeue(skb, qdisc, prio2list(skb, qdisc)); + return __qdisc_requeue(skb, qdisc, &qdisc->q); } -static void pfifo_fast_reset(struct Qdisc* qdisc) +static void fifo_fast_reset(struct Qdisc* qdisc) { - int prio; - struct sk_buff_head *list = qdisc_priv(qdisc); - - for (prio = 0; prio < PFIFO_FAST_BANDS; prio++) - __qdisc_reset_queue(qdisc, list + prio); - + __qdisc_reset_queue(qdisc, &qdisc->q); qdisc->qstats.backlog = 0; - qdisc->q.qlen = 0; } -static int pfifo_fast_dump(struct Qdisc *qdisc, struct sk_buff *skb) -{ - struct tc_prio_qopt opt = { .bands = PFIFO_FAST_BANDS }; - - memcpy(&opt.priomap, prio2band, TC_PRIO_MAX+1); - NLA_PUT(skb, TCA_OPTIONS, sizeof(opt), &opt); - return skb->len; - -nla_put_failure: - return -1; -} - -static int pfifo_fast_init(struct Qdisc *qdisc, struct nlattr *opt) -{ - int prio; - struct sk_buff_head *list = qdisc_priv(qdisc); - - for (prio = 0; prio < PFIFO_FAST_BANDS; prio++) - skb_queue_head_init(list + prio); - - return 0; -} - -static struct Qdisc_ops pfifo_fast_ops __read_mostly = { - .id = "pfifo_fast", - .priv_size = PFIFO_FAST_BANDS * sizeof(struct sk_buff_head), - .enqueue = pfifo_fast_enqueue, - .dequeue = pfifo_fast_dequeue, - .requeue = pfifo_fast_requeue, - .init = pfifo_fast_init, - .reset = pfifo_fast_reset, - .dump = pfifo_fast_dump, +static struct Qdisc_ops fifo_fast_ops __read_mostly = { + .id = "fifo_fast", + .priv_size = 0, + .enqueue = fifo_fast_enqueue, + .dequeue = fifo_fast_dequeue, + .requeue = fifo_fast_requeue, + .reset = fifo_fast_reset, .owner = THIS_MODULE, }; @@ -570,7 +515,7 @@ static void attach_one_default_qdisc(struct net_device *dev, if (dev->tx_queue_len) { qdisc = qdisc_create_dflt(dev, dev_queue, - &pfifo_fast_ops, TC_H_ROOT); + &fifo_fast_ops, TC_H_ROOT); if (!qdisc) { printk(KERN_INFO "%s: activation failed\n", dev->name); return; @@ -601,9 +546,9 @@ void dev_activate(struct net_device *dev) int need_watchdog; /* No queueing discipline is attached to device; - create default one i.e. pfifo_fast for devices, - which need queueing and noqueue_qdisc for - virtual interfaces + * create default one i.e. fifo_fast for devices, + * which need queueing and noqueue_qdisc for + * virtual interfaces. */ if (dev_all_qdisc_sleeping_noop(dev)) -- cgit v1.2.2 From 9b4661bd6e5437508e0920608f3213c23212cd1b Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:01:44 -0700 Subject: ipv4: add pernet mib operations These ones are currently empty, but stuff from init_ipv4_mibs will sequentially migrate there. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/af_inet.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) (limited to 'net') diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 95a966dd1916..b4b77aa07953 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -110,6 +110,7 @@ #include #include #include +#include #ifdef CONFIG_IP_MROUTE #include #endif @@ -1339,6 +1340,20 @@ static struct net_protocol icmp_protocol = { .netns_ok = 1, }; +static __net_init int ipv4_mib_init_net(struct net *net) +{ + return 0; +} + +static __net_exit void ipv4_mib_exit_net(struct net *net) +{ +} + +static __net_initdata struct pernet_operations ipv4_mib_ops = { + .init = ipv4_mib_init_net, + .exit = ipv4_mib_exit_net, +}; + static int __init init_ipv4_mibs(void) { if (snmp_mib_init((void **)net_statistics, @@ -1365,8 +1380,13 @@ static int __init init_ipv4_mibs(void) tcp_mib_init(&init_net); + if (register_pernet_subsys(&ipv4_mib_ops)) + goto err_net; + return 0; +err_net: + snmp_mib_free((void **)udplite_statistics); err_udplite_mib: snmp_mib_free((void **)udp_statistics); err_udp_mib: -- cgit v1.2.2 From 57ef42d59d1c1d79be59fc3c6380ae14234e38c3 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:02:08 -0700 Subject: mib: put tcp statistics on struct net Proc temporary uses stats from init_net. BTW, TCP_XXX_STATS are beautiful (w/o do { } while (0) facing) again :) Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/af_inet.c | 16 +++++++++------- net/ipv4/proc.c | 4 ++-- net/ipv4/tcp.c | 3 --- 3 files changed, 11 insertions(+), 12 deletions(-) (limited to 'net') diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index b4b77aa07953..c1a3e986f8b4 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -1342,11 +1342,20 @@ static struct net_protocol icmp_protocol = { static __net_init int ipv4_mib_init_net(struct net *net) { + if (snmp_mib_init((void **)net->mib.tcp_statistics, + sizeof(struct tcp_mib)) < 0) + goto err_tcp_mib; + + tcp_mib_init(net); return 0; + +err_tcp_mib: + return -ENOMEM; } static __net_exit void ipv4_mib_exit_net(struct net *net) { + snmp_mib_free((void **)net->mib.tcp_statistics); } static __net_initdata struct pernet_operations ipv4_mib_ops = { @@ -1368,9 +1377,6 @@ static int __init init_ipv4_mibs(void) if (snmp_mib_init((void **)icmpmsg_statistics, sizeof(struct icmpmsg_mib)) < 0) goto err_icmpmsg_mib; - if (snmp_mib_init((void **)tcp_statistics, - sizeof(struct tcp_mib)) < 0) - goto err_tcp_mib; if (snmp_mib_init((void **)udp_statistics, sizeof(struct udp_mib)) < 0) goto err_udp_mib; @@ -1378,8 +1384,6 @@ static int __init init_ipv4_mibs(void) sizeof(struct udp_mib)) < 0) goto err_udplite_mib; - tcp_mib_init(&init_net); - if (register_pernet_subsys(&ipv4_mib_ops)) goto err_net; @@ -1390,8 +1394,6 @@ err_net: err_udplite_mib: snmp_mib_free((void **)udp_statistics); err_udp_mib: - snmp_mib_free((void **)tcp_statistics); -err_tcp_mib: snmp_mib_free((void **)icmpmsg_statistics); err_icmpmsg_mib: snmp_mib_free((void **)icmp_statistics); diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index eb5cee279c5f..19e1d8e257da 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -359,11 +359,11 @@ static int snmp_seq_show(struct seq_file *seq, void *v) /* MaxConn field is signed, RFC 2012 */ if (snmp4_tcp_list[i].entry == TCP_MIB_MAXCONN) seq_printf(seq, " %ld", - snmp_fold_field((void **)tcp_statistics, + snmp_fold_field((void **)init_net.mib.tcp_statistics, snmp4_tcp_list[i].entry)); else seq_printf(seq, " %lu", - snmp_fold_field((void **)tcp_statistics, + snmp_fold_field((void **)init_net.mib.tcp_statistics, snmp4_tcp_list[i].entry)); } diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index eec8cf7c0247..827e6132af5f 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -277,8 +277,6 @@ int sysctl_tcp_fin_timeout __read_mostly = TCP_FIN_TIMEOUT; -DEFINE_SNMP_STAT(struct tcp_mib, tcp_statistics) __read_mostly; - atomic_t tcp_orphan_count = ATOMIC_INIT(0); EXPORT_SYMBOL_GPL(tcp_orphan_count); @@ -2802,4 +2800,3 @@ EXPORT_SYMBOL(tcp_splice_read); EXPORT_SYMBOL(tcp_sendpage); EXPORT_SYMBOL(tcp_setsockopt); EXPORT_SYMBOL(tcp_shutdown); -EXPORT_SYMBOL(tcp_statistics); -- cgit v1.2.2 From a20f5799ca7ceb24d63c74b6fdad4b0c0ee91f4f Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:02:42 -0700 Subject: mib: put ip statistics on struct net Similar to tcp one. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/af_inet.c | 11 ++++++----- net/ipv4/ip_input.c | 8 -------- net/ipv4/proc.c | 4 ++-- 3 files changed, 8 insertions(+), 15 deletions(-) (limited to 'net') diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index c1a3e986f8b4..3090a9307c46 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -1345,16 +1345,22 @@ static __net_init int ipv4_mib_init_net(struct net *net) if (snmp_mib_init((void **)net->mib.tcp_statistics, sizeof(struct tcp_mib)) < 0) goto err_tcp_mib; + if (snmp_mib_init((void **)net->mib.ip_statistics, + sizeof(struct ipstats_mib)) < 0) + goto err_ip_mib; tcp_mib_init(net); return 0; +err_ip_mib: + snmp_mib_free((void **)net->mib.tcp_statistics); err_tcp_mib: return -ENOMEM; } static __net_exit void ipv4_mib_exit_net(struct net *net) { + snmp_mib_free((void **)net->mib.ip_statistics); snmp_mib_free((void **)net->mib.tcp_statistics); } @@ -1368,9 +1374,6 @@ static int __init init_ipv4_mibs(void) if (snmp_mib_init((void **)net_statistics, sizeof(struct linux_mib)) < 0) goto err_net_mib; - if (snmp_mib_init((void **)ip_statistics, - sizeof(struct ipstats_mib)) < 0) - goto err_ip_mib; if (snmp_mib_init((void **)icmp_statistics, sizeof(struct icmp_mib)) < 0) goto err_icmp_mib; @@ -1398,8 +1401,6 @@ err_udp_mib: err_icmpmsg_mib: snmp_mib_free((void **)icmp_statistics); err_icmp_mib: - snmp_mib_free((void **)ip_statistics); -err_ip_mib: snmp_mib_free((void **)net_statistics); err_net_mib: return -ENOMEM; diff --git a/net/ipv4/ip_input.c b/net/ipv4/ip_input.c index 043f640df4b7..e0bed56c51f1 100644 --- a/net/ipv4/ip_input.c +++ b/net/ipv4/ip_input.c @@ -144,12 +144,6 @@ #include #include -/* - * SNMP management statistics - */ - -DEFINE_SNMP_STAT(struct ipstats_mib, ip_statistics) __read_mostly; - /* * Process Router Attention IP option */ @@ -447,5 +441,3 @@ drop: out: return NET_RX_DROP; } - -EXPORT_SYMBOL(ip_statistics); diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 19e1d8e257da..2698bb2ce98f 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -344,7 +344,7 @@ static int snmp_seq_show(struct seq_file *seq, void *v) for (i = 0; snmp4_ipstats_list[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **)ip_statistics, + snmp_fold_field((void **)init_net.mib.ip_statistics, snmp4_ipstats_list[i].entry)); icmp_put(seq); /* RFC 2011 compatibility */ @@ -431,7 +431,7 @@ static int netstat_seq_show(struct seq_file *seq, void *v) seq_puts(seq, "\nIpExt:"); for (i = 0; snmp4_ipextstats_list[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **)ip_statistics, + snmp_fold_field((void **)init_net.mib.ip_statistics, snmp4_ipextstats_list[i].entry)); seq_putc(seq, '\n'); -- cgit v1.2.2 From 61a7e26028b94805fd686a6dc9dbd9941f8f19b0 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:03:08 -0700 Subject: mib: put net statistics on struct net Similar to ip and tcp ones :) Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/af_inet.c | 14 ++++++-------- net/ipv4/proc.c | 2 +- 2 files changed, 7 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 3090a9307c46..776ed3199b52 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -115,8 +115,6 @@ #include #endif -DEFINE_SNMP_STAT(struct linux_mib, net_statistics) __read_mostly; - extern void ip_mc_drop_socket(struct sock *sk); /* The inetsw table contains everything that inet_create needs to @@ -1348,10 +1346,15 @@ static __net_init int ipv4_mib_init_net(struct net *net) if (snmp_mib_init((void **)net->mib.ip_statistics, sizeof(struct ipstats_mib)) < 0) goto err_ip_mib; + if (snmp_mib_init((void **)net->mib.net_statistics, + sizeof(struct linux_mib)) < 0) + goto err_net_mib; tcp_mib_init(net); return 0; +err_net_mib: + snmp_mib_free((void **)net->mib.ip_statistics); err_ip_mib: snmp_mib_free((void **)net->mib.tcp_statistics); err_tcp_mib: @@ -1360,6 +1363,7 @@ err_tcp_mib: static __net_exit void ipv4_mib_exit_net(struct net *net) { + snmp_mib_free((void **)net->mib.net_statistics); snmp_mib_free((void **)net->mib.ip_statistics); snmp_mib_free((void **)net->mib.tcp_statistics); } @@ -1371,9 +1375,6 @@ static __net_initdata struct pernet_operations ipv4_mib_ops = { static int __init init_ipv4_mibs(void) { - if (snmp_mib_init((void **)net_statistics, - sizeof(struct linux_mib)) < 0) - goto err_net_mib; if (snmp_mib_init((void **)icmp_statistics, sizeof(struct icmp_mib)) < 0) goto err_icmp_mib; @@ -1401,8 +1402,6 @@ err_udp_mib: err_icmpmsg_mib: snmp_mib_free((void **)icmp_statistics); err_icmp_mib: - snmp_mib_free((void **)net_statistics); -err_net_mib: return -ENOMEM; } @@ -1582,5 +1581,4 @@ EXPORT_SYMBOL(inet_sock_destruct); EXPORT_SYMBOL(inet_stream_connect); EXPORT_SYMBOL(inet_stream_ops); EXPORT_SYMBOL(inet_unregister_protosw); -EXPORT_SYMBOL(net_statistics); EXPORT_SYMBOL(sysctl_ip_nonlocal_bind); diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 2698bb2ce98f..ef38b1dccfda 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -421,7 +421,7 @@ static int netstat_seq_show(struct seq_file *seq, void *v) seq_puts(seq, "\nTcpExt:"); for (i = 0; snmp4_net_list[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **)net_statistics, + snmp_fold_field((void **)init_net.mib.net_statistics, snmp4_net_list[i].entry)); seq_puts(seq, "\nIpExt:"); -- cgit v1.2.2 From 2f275f91a438abd8eec5321798d66a4ffe6869fa Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:03:27 -0700 Subject: mib: put udp statistics on struct net Similar to... ouch, I repeat myself. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/af_inet.c | 11 ++++++----- net/ipv4/proc.c | 2 +- net/ipv4/udp.c | 3 --- 3 files changed, 7 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 776ed3199b52..1f418164ebf2 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -1349,10 +1349,15 @@ static __net_init int ipv4_mib_init_net(struct net *net) if (snmp_mib_init((void **)net->mib.net_statistics, sizeof(struct linux_mib)) < 0) goto err_net_mib; + if (snmp_mib_init((void **)net->mib.udp_statistics, + sizeof(struct udp_mib)) < 0) + goto err_udp_mib; tcp_mib_init(net); return 0; +err_udp_mib: + snmp_mib_free((void **)net->mib.net_statistics); err_net_mib: snmp_mib_free((void **)net->mib.ip_statistics); err_ip_mib: @@ -1363,6 +1368,7 @@ err_tcp_mib: static __net_exit void ipv4_mib_exit_net(struct net *net) { + snmp_mib_free((void **)net->mib.udp_statistics); snmp_mib_free((void **)net->mib.net_statistics); snmp_mib_free((void **)net->mib.ip_statistics); snmp_mib_free((void **)net->mib.tcp_statistics); @@ -1381,9 +1387,6 @@ static int __init init_ipv4_mibs(void) if (snmp_mib_init((void **)icmpmsg_statistics, sizeof(struct icmpmsg_mib)) < 0) goto err_icmpmsg_mib; - if (snmp_mib_init((void **)udp_statistics, - sizeof(struct udp_mib)) < 0) - goto err_udp_mib; if (snmp_mib_init((void **)udplite_statistics, sizeof(struct udp_mib)) < 0) goto err_udplite_mib; @@ -1396,8 +1399,6 @@ static int __init init_ipv4_mibs(void) err_net: snmp_mib_free((void **)udplite_statistics); err_udplite_mib: - snmp_mib_free((void **)udp_statistics); -err_udp_mib: snmp_mib_free((void **)icmpmsg_statistics); err_icmpmsg_mib: snmp_mib_free((void **)icmp_statistics); diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index ef38b1dccfda..869085c8e43f 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -374,7 +374,7 @@ static int snmp_seq_show(struct seq_file *seq, void *v) seq_puts(seq, "\nUdp:"); for (i = 0; snmp4_udp_list[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **)udp_statistics, + snmp_fold_field((void **)init_net.mib.udp_statistics, snmp4_udp_list[i].entry)); /* the UDP and UDP-Lite MIBs are the same */ diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 1560d11ba6ef..a751770947a3 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -108,9 +108,6 @@ * Snmp MIB for the UDP layer */ -DEFINE_SNMP_STAT(struct udp_mib, udp_statistics) __read_mostly; -EXPORT_SYMBOL(udp_statistics); - DEFINE_SNMP_STAT(struct udp_mib, udp_stats_in6) __read_mostly; EXPORT_SYMBOL(udp_stats_in6); -- cgit v1.2.2 From 386019d3514b3ed9de8d0b05b67e638a7048375b Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:03:45 -0700 Subject: mib: put udplite statistics on struct net Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/af_inet.c | 11 ++++++----- net/ipv4/proc.c | 2 +- net/ipv4/udplite.c | 1 - 3 files changed, 7 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 1f418164ebf2..bf1f200c6575 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -1352,10 +1352,15 @@ static __net_init int ipv4_mib_init_net(struct net *net) if (snmp_mib_init((void **)net->mib.udp_statistics, sizeof(struct udp_mib)) < 0) goto err_udp_mib; + if (snmp_mib_init((void **)net->mib.udplite_statistics, + sizeof(struct udp_mib)) < 0) + goto err_udplite_mib; tcp_mib_init(net); return 0; +err_udplite_mib: + snmp_mib_free((void **)net->mib.udp_statistics); err_udp_mib: snmp_mib_free((void **)net->mib.net_statistics); err_net_mib: @@ -1368,6 +1373,7 @@ err_tcp_mib: static __net_exit void ipv4_mib_exit_net(struct net *net) { + snmp_mib_free((void **)net->mib.udplite_statistics); snmp_mib_free((void **)net->mib.udp_statistics); snmp_mib_free((void **)net->mib.net_statistics); snmp_mib_free((void **)net->mib.ip_statistics); @@ -1387,9 +1393,6 @@ static int __init init_ipv4_mibs(void) if (snmp_mib_init((void **)icmpmsg_statistics, sizeof(struct icmpmsg_mib)) < 0) goto err_icmpmsg_mib; - if (snmp_mib_init((void **)udplite_statistics, - sizeof(struct udp_mib)) < 0) - goto err_udplite_mib; if (register_pernet_subsys(&ipv4_mib_ops)) goto err_net; @@ -1397,8 +1400,6 @@ static int __init init_ipv4_mibs(void) return 0; err_net: - snmp_mib_free((void **)udplite_statistics); -err_udplite_mib: snmp_mib_free((void **)icmpmsg_statistics); err_icmpmsg_mib: snmp_mib_free((void **)icmp_statistics); diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 869085c8e43f..765418334b3c 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -385,7 +385,7 @@ static int snmp_seq_show(struct seq_file *seq, void *v) seq_puts(seq, "\nUdpLite:"); for (i = 0; snmp4_udp_list[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **)udplite_statistics, + snmp_fold_field((void **)init_net.mib.udplite_statistics, snmp4_udp_list[i].entry)); seq_putc(seq, '\n'); diff --git a/net/ipv4/udplite.c b/net/ipv4/udplite.c index 4ad16b6d5138..3c807964da96 100644 --- a/net/ipv4/udplite.c +++ b/net/ipv4/udplite.c @@ -11,7 +11,6 @@ * 2 of the License, or (at your option) any later version. */ #include "udp_impl.h" -DEFINE_SNMP_STAT(struct udp_mib, udplite_statistics) __read_mostly; struct hlist_head udplite_hash[UDP_HTABLE_SIZE]; -- cgit v1.2.2 From b60538a0d737609213e4b758881913498d3ff0b4 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:04:02 -0700 Subject: mib: put icmp statistics on struct net Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/af_inet.c | 11 ++++++----- net/ipv4/icmp.c | 2 -- net/ipv4/proc.c | 8 ++++---- 3 files changed, 10 insertions(+), 11 deletions(-) (limited to 'net') diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index bf1f200c6575..5c72bb3adadd 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -1355,10 +1355,15 @@ static __net_init int ipv4_mib_init_net(struct net *net) if (snmp_mib_init((void **)net->mib.udplite_statistics, sizeof(struct udp_mib)) < 0) goto err_udplite_mib; + if (snmp_mib_init((void **)net->mib.icmp_statistics, + sizeof(struct icmp_mib)) < 0) + goto err_icmp_mib; tcp_mib_init(net); return 0; +err_icmp_mib: + snmp_mib_free((void **)net->mib.udplite_statistics); err_udplite_mib: snmp_mib_free((void **)net->mib.udp_statistics); err_udp_mib: @@ -1373,6 +1378,7 @@ err_tcp_mib: static __net_exit void ipv4_mib_exit_net(struct net *net) { + snmp_mib_free((void **)net->mib.icmp_statistics); snmp_mib_free((void **)net->mib.udplite_statistics); snmp_mib_free((void **)net->mib.udp_statistics); snmp_mib_free((void **)net->mib.net_statistics); @@ -1387,9 +1393,6 @@ static __net_initdata struct pernet_operations ipv4_mib_ops = { static int __init init_ipv4_mibs(void) { - if (snmp_mib_init((void **)icmp_statistics, - sizeof(struct icmp_mib)) < 0) - goto err_icmp_mib; if (snmp_mib_init((void **)icmpmsg_statistics, sizeof(struct icmpmsg_mib)) < 0) goto err_icmpmsg_mib; @@ -1402,8 +1405,6 @@ static int __init init_ipv4_mibs(void) err_net: snmp_mib_free((void **)icmpmsg_statistics); err_icmpmsg_mib: - snmp_mib_free((void **)icmp_statistics); -err_icmp_mib: return -ENOMEM; } diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c index ea60ad41008c..33f958902d9b 100644 --- a/net/ipv4/icmp.c +++ b/net/ipv4/icmp.c @@ -114,7 +114,6 @@ struct icmp_bxm { /* * Statistics */ -DEFINE_SNMP_STAT(struct icmp_mib, icmp_statistics) __read_mostly; DEFINE_SNMP_STAT(struct icmpmsg_mib, icmpmsg_statistics) __read_mostly; /* An array of errno for error messages from dest unreach. */ @@ -1213,5 +1212,4 @@ int __init icmp_init(void) EXPORT_SYMBOL(icmp_err_convert); EXPORT_SYMBOL(icmp_send); -EXPORT_SYMBOL(icmp_statistics); EXPORT_SYMBOL(xrlim_allow); diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 765418334b3c..6b43cce05587 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -311,15 +311,15 @@ static void icmp_put(struct seq_file *seq) for (i=0; icmpmibmap[i].name != NULL; i++) seq_printf(seq, " Out%s", icmpmibmap[i].name); seq_printf(seq, "\nIcmp: %lu %lu", - snmp_fold_field((void **) icmp_statistics, ICMP_MIB_INMSGS), - snmp_fold_field((void **) icmp_statistics, ICMP_MIB_INERRORS)); + snmp_fold_field((void **) init_net.mib.icmp_statistics, ICMP_MIB_INMSGS), + snmp_fold_field((void **) init_net.mib.icmp_statistics, ICMP_MIB_INERRORS)); for (i=0; icmpmibmap[i].name != NULL; i++) seq_printf(seq, " %lu", snmp_fold_field((void **) icmpmsg_statistics, icmpmibmap[i].index)); seq_printf(seq, " %lu %lu", - snmp_fold_field((void **) icmp_statistics, ICMP_MIB_OUTMSGS), - snmp_fold_field((void **) icmp_statistics, ICMP_MIB_OUTERRORS)); + snmp_fold_field((void **) init_net.mib.icmp_statistics, ICMP_MIB_OUTMSGS), + snmp_fold_field((void **) init_net.mib.icmp_statistics, ICMP_MIB_OUTERRORS)); for (i=0; icmpmibmap[i].name != NULL; i++) seq_printf(seq, " %lu", snmp_fold_field((void **) icmpmsg_statistics, -- cgit v1.2.2 From 923c6586b0dc0a00df07a1608185437145a0c68b Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:04:22 -0700 Subject: mib: put icmpmsg statistics on struct net Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/af_inet.c | 12 ++++++------ net/ipv4/icmp.c | 5 ----- net/ipv4/proc.c | 10 +++++----- 3 files changed, 11 insertions(+), 16 deletions(-) (limited to 'net') diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 5c72bb3adadd..36ff6dc769c6 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -1358,10 +1358,15 @@ static __net_init int ipv4_mib_init_net(struct net *net) if (snmp_mib_init((void **)net->mib.icmp_statistics, sizeof(struct icmp_mib)) < 0) goto err_icmp_mib; + if (snmp_mib_init((void **)net->mib.icmpmsg_statistics, + sizeof(struct icmpmsg_mib)) < 0) + goto err_icmpmsg_mib; tcp_mib_init(net); return 0; +err_icmpmsg_mib: + snmp_mib_free((void **)net->mib.icmp_statistics); err_icmp_mib: snmp_mib_free((void **)net->mib.udplite_statistics); err_udplite_mib: @@ -1378,6 +1383,7 @@ err_tcp_mib: static __net_exit void ipv4_mib_exit_net(struct net *net) { + snmp_mib_free((void **)net->mib.icmpmsg_statistics); snmp_mib_free((void **)net->mib.icmp_statistics); snmp_mib_free((void **)net->mib.udplite_statistics); snmp_mib_free((void **)net->mib.udp_statistics); @@ -1393,18 +1399,12 @@ static __net_initdata struct pernet_operations ipv4_mib_ops = { static int __init init_ipv4_mibs(void) { - if (snmp_mib_init((void **)icmpmsg_statistics, - sizeof(struct icmpmsg_mib)) < 0) - goto err_icmpmsg_mib; - if (register_pernet_subsys(&ipv4_mib_ops)) goto err_net; return 0; err_net: - snmp_mib_free((void **)icmpmsg_statistics); -err_icmpmsg_mib: return -ENOMEM; } diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c index 33f958902d9b..860558633b2c 100644 --- a/net/ipv4/icmp.c +++ b/net/ipv4/icmp.c @@ -111,11 +111,6 @@ struct icmp_bxm { unsigned char optbuf[40]; }; -/* - * Statistics - */ -DEFINE_SNMP_STAT(struct icmpmsg_mib, icmpmsg_statistics) __read_mostly; - /* An array of errno for error messages from dest unreach. */ /* RFC 1122: 3.2.2.1 States that NET_UNREACH, HOST_UNREACH and SR_FAILED MUST be considered 'transient errs'. */ diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 6b43cce05587..e11144b0134a 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -270,7 +270,7 @@ static void icmpmsg_put(struct seq_file *seq) count = 0; for (i = 0; i < ICMPMSG_MIB_MAX; i++) { - if (snmp_fold_field((void **) icmpmsg_statistics, i)) + if (snmp_fold_field((void **) init_net.mib.icmpmsg_statistics, i)) out[count++] = i; if (count < PERLINE) continue; @@ -282,7 +282,7 @@ static void icmpmsg_put(struct seq_file *seq) seq_printf(seq, "\nIcmpMsg: "); for (j = 0; j < PERLINE; ++j) seq_printf(seq, " %lu", - snmp_fold_field((void **) icmpmsg_statistics, + snmp_fold_field((void **) init_net.mib.icmpmsg_statistics, out[j])); seq_putc(seq, '\n'); } @@ -294,7 +294,7 @@ static void icmpmsg_put(struct seq_file *seq) seq_printf(seq, "\nIcmpMsg:"); for (j = 0; j < count; ++j) seq_printf(seq, " %lu", snmp_fold_field((void **) - icmpmsg_statistics, out[j])); + init_net.mib.icmpmsg_statistics, out[j])); } #undef PERLINE @@ -315,14 +315,14 @@ static void icmp_put(struct seq_file *seq) snmp_fold_field((void **) init_net.mib.icmp_statistics, ICMP_MIB_INERRORS)); for (i=0; icmpmibmap[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **) icmpmsg_statistics, + snmp_fold_field((void **) init_net.mib.icmpmsg_statistics, icmpmibmap[i].index)); seq_printf(seq, " %lu %lu", snmp_fold_field((void **) init_net.mib.icmp_statistics, ICMP_MIB_OUTMSGS), snmp_fold_field((void **) init_net.mib.icmp_statistics, ICMP_MIB_OUTERRORS)); for (i=0; icmpmibmap[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **) icmpmsg_statistics, + snmp_fold_field((void **) init_net.mib.icmpmsg_statistics, icmpmibmap[i].index | 0x100)); } -- cgit v1.2.2 From d89cbbb1e69a3bfea38038fb058bd51013902ef3 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:04:51 -0700 Subject: ipv4: clean the init_ipv4_mibs error paths After moving all the stuff outside this function it looks a bit ugly - make it look better. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/af_inet.c | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-) (limited to 'net') diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 36ff6dc769c6..dd919d84285f 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -1399,13 +1399,7 @@ static __net_initdata struct pernet_operations ipv4_mib_ops = { static int __init init_ipv4_mibs(void) { - if (register_pernet_subsys(&ipv4_mib_ops)) - goto err_net; - - return 0; - -err_net: - return -ENOMEM; + return register_pernet_subsys(&ipv4_mib_ops); } static int ipv4_proc_init(void); -- cgit v1.2.2 From 7b7a9dfdf6ccda647f54ea5fa3bd0ac17a189eeb Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:05:17 -0700 Subject: proc: create /proc/net/netstat file in each net Now all the shown in it statistics is netnsizated, time to show it in appropriate net. The appropriate net init/exit ops already exist - they make the sockstat file per net - so just extend them. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/proc.c | 47 ++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 38 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index e11144b0134a..df8c4affde0f 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -413,6 +413,7 @@ static const struct file_operations snmp_seq_fops = { static int netstat_seq_show(struct seq_file *seq, void *v) { int i; + struct net *net = seq->private; seq_puts(seq, "TcpExt:"); for (i = 0; snmp4_net_list[i].name != NULL; i++) @@ -421,7 +422,7 @@ static int netstat_seq_show(struct seq_file *seq, void *v) seq_puts(seq, "\nTcpExt:"); for (i = 0; snmp4_net_list[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **)init_net.mib.net_statistics, + snmp_fold_field((void **)net->mib.net_statistics, snmp4_net_list[i].entry)); seq_puts(seq, "\nIpExt:"); @@ -431,7 +432,7 @@ static int netstat_seq_show(struct seq_file *seq, void *v) seq_puts(seq, "\nIpExt:"); for (i = 0; snmp4_ipextstats_list[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **)init_net.mib.ip_statistics, + snmp_fold_field((void **)net->mib.ip_statistics, snmp4_ipextstats_list[i].entry)); seq_putc(seq, '\n'); @@ -440,7 +441,32 @@ static int netstat_seq_show(struct seq_file *seq, void *v) static int netstat_seq_open(struct inode *inode, struct file *file) { - return single_open(file, netstat_seq_show, NULL); + int err; + struct net *net; + + err = -ENXIO; + net = get_proc_net(inode); + if (net == NULL) + goto err_net; + + err = single_open(file, netstat_seq_show, net); + if (err < 0) + goto err_open; + + return 0; + +err_open: + put_net(net); +err_net: + return err; +} + +static int netstat_seq_release(struct inode *inode, struct file *file) +{ + struct net *net = ((struct seq_file *)file->private_data)->private; + + put_net(net); + return single_release(inode, file); } static const struct file_operations netstat_seq_fops = { @@ -448,18 +474,26 @@ static const struct file_operations netstat_seq_fops = { .open = netstat_seq_open, .read = seq_read, .llseek = seq_lseek, - .release = single_release, + .release = netstat_seq_release, }; static __net_init int ip_proc_init_net(struct net *net) { if (!proc_net_fops_create(net, "sockstat", S_IRUGO, &sockstat_seq_fops)) return -ENOMEM; + if (!proc_net_fops_create(net, "netstat", S_IRUGO, &netstat_seq_fops)) + goto out_netstat; + return 0; + +out_netstat: + proc_net_remove(net, "sockstat"); + return -ENOMEM; } static __net_exit void ip_proc_exit_net(struct net *net) { + proc_net_remove(net, "netstat"); proc_net_remove(net, "sockstat"); } @@ -475,16 +509,11 @@ int __init ip_misc_proc_init(void) if (register_pernet_subsys(&ip_proc_ops)) goto out_pernet; - if (!proc_net_fops_create(&init_net, "netstat", S_IRUGO, &netstat_seq_fops)) - goto out_netstat; - if (!proc_net_fops_create(&init_net, "snmp", S_IRUGO, &snmp_seq_fops)) goto out_snmp; out: return rc; out_snmp: - proc_net_remove(&init_net, "netstat"); -out_netstat: unregister_pernet_subsys(&ip_proc_ops); out_pernet: rc = -ENOMEM; -- cgit v1.2.2 From 229bf0cbaa054f1502ab4ee219f05985d2c838d1 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:06:04 -0700 Subject: proc: create /proc/net/snmp file in each net All the statistics shown in this file have been made per-net already. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/proc.c | 70 ++++++++++++++++++++++++++++++++++++++++----------------- 1 file changed, 50 insertions(+), 20 deletions(-) (limited to 'net') diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index df8c4affde0f..367b81f6f57a 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -266,11 +266,12 @@ static void icmpmsg_put(struct seq_file *seq) int j, i, count; static int out[PERLINE]; + struct net *net = seq->private; count = 0; for (i = 0; i < ICMPMSG_MIB_MAX; i++) { - if (snmp_fold_field((void **) init_net.mib.icmpmsg_statistics, i)) + if (snmp_fold_field((void **) net->mib.icmpmsg_statistics, i)) out[count++] = i; if (count < PERLINE) continue; @@ -282,7 +283,7 @@ static void icmpmsg_put(struct seq_file *seq) seq_printf(seq, "\nIcmpMsg: "); for (j = 0; j < PERLINE; ++j) seq_printf(seq, " %lu", - snmp_fold_field((void **) init_net.mib.icmpmsg_statistics, + snmp_fold_field((void **) net->mib.icmpmsg_statistics, out[j])); seq_putc(seq, '\n'); } @@ -294,7 +295,7 @@ static void icmpmsg_put(struct seq_file *seq) seq_printf(seq, "\nIcmpMsg:"); for (j = 0; j < count; ++j) seq_printf(seq, " %lu", snmp_fold_field((void **) - init_net.mib.icmpmsg_statistics, out[j])); + net->mib.icmpmsg_statistics, out[j])); } #undef PERLINE @@ -303,6 +304,7 @@ static void icmpmsg_put(struct seq_file *seq) static void icmp_put(struct seq_file *seq) { int i; + struct net *net = seq->private; seq_puts(seq, "\nIcmp: InMsgs InErrors"); for (i=0; icmpmibmap[i].name != NULL; i++) @@ -311,18 +313,18 @@ static void icmp_put(struct seq_file *seq) for (i=0; icmpmibmap[i].name != NULL; i++) seq_printf(seq, " Out%s", icmpmibmap[i].name); seq_printf(seq, "\nIcmp: %lu %lu", - snmp_fold_field((void **) init_net.mib.icmp_statistics, ICMP_MIB_INMSGS), - snmp_fold_field((void **) init_net.mib.icmp_statistics, ICMP_MIB_INERRORS)); + snmp_fold_field((void **) net->mib.icmp_statistics, ICMP_MIB_INMSGS), + snmp_fold_field((void **) net->mib.icmp_statistics, ICMP_MIB_INERRORS)); for (i=0; icmpmibmap[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **) init_net.mib.icmpmsg_statistics, + snmp_fold_field((void **) net->mib.icmpmsg_statistics, icmpmibmap[i].index)); seq_printf(seq, " %lu %lu", - snmp_fold_field((void **) init_net.mib.icmp_statistics, ICMP_MIB_OUTMSGS), - snmp_fold_field((void **) init_net.mib.icmp_statistics, ICMP_MIB_OUTERRORS)); + snmp_fold_field((void **) net->mib.icmp_statistics, ICMP_MIB_OUTMSGS), + snmp_fold_field((void **) net->mib.icmp_statistics, ICMP_MIB_OUTERRORS)); for (i=0; icmpmibmap[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **) init_net.mib.icmpmsg_statistics, + snmp_fold_field((void **) net->mib.icmpmsg_statistics, icmpmibmap[i].index | 0x100)); } @@ -332,6 +334,7 @@ static void icmp_put(struct seq_file *seq) static int snmp_seq_show(struct seq_file *seq, void *v) { int i; + struct net *net = seq->private; seq_puts(seq, "Ip: Forwarding DefaultTTL"); @@ -344,7 +347,7 @@ static int snmp_seq_show(struct seq_file *seq, void *v) for (i = 0; snmp4_ipstats_list[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **)init_net.mib.ip_statistics, + snmp_fold_field((void **)net->mib.ip_statistics, snmp4_ipstats_list[i].entry)); icmp_put(seq); /* RFC 2011 compatibility */ @@ -359,11 +362,11 @@ static int snmp_seq_show(struct seq_file *seq, void *v) /* MaxConn field is signed, RFC 2012 */ if (snmp4_tcp_list[i].entry == TCP_MIB_MAXCONN) seq_printf(seq, " %ld", - snmp_fold_field((void **)init_net.mib.tcp_statistics, + snmp_fold_field((void **)net->mib.tcp_statistics, snmp4_tcp_list[i].entry)); else seq_printf(seq, " %lu", - snmp_fold_field((void **)init_net.mib.tcp_statistics, + snmp_fold_field((void **)net->mib.tcp_statistics, snmp4_tcp_list[i].entry)); } @@ -374,7 +377,7 @@ static int snmp_seq_show(struct seq_file *seq, void *v) seq_puts(seq, "\nUdp:"); for (i = 0; snmp4_udp_list[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **)init_net.mib.udp_statistics, + snmp_fold_field((void **)net->mib.udp_statistics, snmp4_udp_list[i].entry)); /* the UDP and UDP-Lite MIBs are the same */ @@ -385,7 +388,7 @@ static int snmp_seq_show(struct seq_file *seq, void *v) seq_puts(seq, "\nUdpLite:"); for (i = 0; snmp4_udp_list[i].name != NULL; i++) seq_printf(seq, " %lu", - snmp_fold_field((void **)init_net.mib.udplite_statistics, + snmp_fold_field((void **)net->mib.udplite_statistics, snmp4_udp_list[i].entry)); seq_putc(seq, '\n'); @@ -394,7 +397,32 @@ static int snmp_seq_show(struct seq_file *seq, void *v) static int snmp_seq_open(struct inode *inode, struct file *file) { - return single_open(file, snmp_seq_show, NULL); + int err; + struct net *net; + + err = -ENXIO; + net = get_proc_net(inode); + if (net == NULL) + goto err_net; + + err = single_open(file, snmp_seq_show, net); + if (err < 0) + goto err_open; + + return 0; + +err_open: + put_net(net); +err_net: + return err; +} + +static int snmp_seq_release(struct inode *inode, struct file *file) +{ + struct net *net = ((struct seq_file *)file->private_data)->private; + + put_net(net); + return single_release(inode, file); } static const struct file_operations snmp_seq_fops = { @@ -402,7 +430,7 @@ static const struct file_operations snmp_seq_fops = { .open = snmp_seq_open, .read = seq_read, .llseek = seq_lseek, - .release = single_release, + .release = snmp_seq_release, }; @@ -483,9 +511,13 @@ static __net_init int ip_proc_init_net(struct net *net) return -ENOMEM; if (!proc_net_fops_create(net, "netstat", S_IRUGO, &netstat_seq_fops)) goto out_netstat; + if (!proc_net_fops_create(net, "snmp", S_IRUGO, &snmp_seq_fops)) + goto out_snmp; return 0; +out_snmp: + proc_net_remove(net, "netstat"); out_netstat: proc_net_remove(net, "sockstat"); return -ENOMEM; @@ -493,6 +525,7 @@ out_netstat: static __net_exit void ip_proc_exit_net(struct net *net) { + proc_net_remove(net, "snmp"); proc_net_remove(net, "netstat"); proc_net_remove(net, "sockstat"); } @@ -509,12 +542,9 @@ int __init ip_misc_proc_init(void) if (register_pernet_subsys(&ip_proc_ops)) goto out_pernet; - if (!proc_net_fops_create(&init_net, "snmp", S_IRUGO, &snmp_seq_fops)) - goto out_snmp; out: return rc; -out_snmp: - unregister_pernet_subsys(&ip_proc_ops); + out_pernet: rc = -ENOMEM; goto out; -- cgit v1.2.2 From 8e3461d01bdbc3dbf993448ed9ad4acaaeb6495d Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:06:26 -0700 Subject: proc: show per-net ip_devconf.forwarding in /proc/net/snmp This one has become per-net long ago, but the appropriate file is per-net only now. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/proc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 367b81f6f57a..120e1f7461da 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -342,7 +342,7 @@ static int snmp_seq_show(struct seq_file *seq, void *v) seq_printf(seq, " %s", snmp4_ipstats_list[i].name); seq_printf(seq, "\nIp: %d %d", - IPV4_DEVCONF_ALL(&init_net, FORWARDING) ? 1 : 2, + IPV4_DEVCONF_ALL(net, FORWARDING) ? 1 : 2, sysctl_ip_default_ttl); for (i = 0; snmp4_ipstats_list[i].name != NULL; i++) -- cgit v1.2.2 From 60bdde95807e982a824be9cfdd35055cc721a88a Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:06:50 -0700 Subject: proc: clean the ip_misc_proc_init and ip_proc_init_net error paths After all this stuff is moved outside, this function can look better. Besides, I tuned the error path in ip_proc_init_net to make it have only 2 exit points, not 3. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/proc.c | 15 +++------------ 1 file changed, 3 insertions(+), 12 deletions(-) (limited to 'net') diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 120e1f7461da..554390431a4e 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -508,7 +508,7 @@ static const struct file_operations netstat_seq_fops = { static __net_init int ip_proc_init_net(struct net *net) { if (!proc_net_fops_create(net, "sockstat", S_IRUGO, &sockstat_seq_fops)) - return -ENOMEM; + goto out_sockstat; if (!proc_net_fops_create(net, "netstat", S_IRUGO, &netstat_seq_fops)) goto out_netstat; if (!proc_net_fops_create(net, "snmp", S_IRUGO, &snmp_seq_fops)) @@ -520,6 +520,7 @@ out_snmp: proc_net_remove(net, "netstat"); out_netstat: proc_net_remove(net, "sockstat"); +out_sockstat: return -ENOMEM; } @@ -537,16 +538,6 @@ static __net_initdata struct pernet_operations ip_proc_ops = { int __init ip_misc_proc_init(void) { - int rc = 0; - - if (register_pernet_subsys(&ip_proc_ops)) - goto out_pernet; - -out: - return rc; - -out_pernet: - rc = -ENOMEM; - goto out; + return register_pernet_subsys(&ip_proc_ops); } -- cgit v1.2.2 From de05c557b24c7dffc6d392e3db120cf11c9f6ae7 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:07:21 -0700 Subject: proc: consolidate per-net single_open callers There are already 7 of them - time to kill some duplicate code. Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/fib_trie.c | 13 +----------- net/ipv4/proc.c | 57 +++-------------------------------------------------- net/ipv6/proc.c | 19 +----------------- net/ipv6/route.c | 26 ++---------------------- 4 files changed, 7 insertions(+), 108 deletions(-) (limited to 'net') diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c index f155a66d6ebf..6009df238ed9 100644 --- a/net/ipv4/fib_trie.c +++ b/net/ipv4/fib_trie.c @@ -2251,18 +2251,7 @@ static int fib_triestat_seq_show(struct seq_file *seq, void *v) static int fib_triestat_seq_open(struct inode *inode, struct file *file) { - int err; - struct net *net; - - net = get_proc_net(inode); - if (net == NULL) - return -ENXIO; - err = single_open(file, fib_triestat_seq_show, net); - if (err < 0) { - put_net(net); - return err; - } - return 0; + return single_open_net(inode, file, fib_triestat_seq_show); } static int fib_triestat_seq_release(struct inode *ino, struct file *f) diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index 554390431a4e..daf5d3c80cef 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -71,24 +71,7 @@ static int sockstat_seq_show(struct seq_file *seq, void *v) static int sockstat_seq_open(struct inode *inode, struct file *file) { - int err; - struct net *net; - - err = -ENXIO; - net = get_proc_net(inode); - if (net == NULL) - goto err_net; - - err = single_open(file, sockstat_seq_show, net); - if (err < 0) - goto err_open; - - return 0; - -err_open: - put_net(net); -err_net: - return err; + return single_open_net(inode, file, sockstat_seq_show); } static int sockstat_seq_release(struct inode *inode, struct file *file) @@ -397,24 +380,7 @@ static int snmp_seq_show(struct seq_file *seq, void *v) static int snmp_seq_open(struct inode *inode, struct file *file) { - int err; - struct net *net; - - err = -ENXIO; - net = get_proc_net(inode); - if (net == NULL) - goto err_net; - - err = single_open(file, snmp_seq_show, net); - if (err < 0) - goto err_open; - - return 0; - -err_open: - put_net(net); -err_net: - return err; + return single_open_net(inode, file, snmp_seq_show); } static int snmp_seq_release(struct inode *inode, struct file *file) @@ -469,24 +435,7 @@ static int netstat_seq_show(struct seq_file *seq, void *v) static int netstat_seq_open(struct inode *inode, struct file *file) { - int err; - struct net *net; - - err = -ENXIO; - net = get_proc_net(inode); - if (net == NULL) - goto err_net; - - err = single_open(file, netstat_seq_show, net); - if (err < 0) - goto err_open; - - return 0; - -err_open: - put_net(net); -err_net: - return err; + return single_open_net(inode, file, netstat_seq_show); } static int netstat_seq_release(struct inode *inode, struct file *file) diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c index cbc7e514d3ec..29c5a79444c2 100644 --- a/net/ipv6/proc.c +++ b/net/ipv6/proc.c @@ -183,24 +183,7 @@ static int snmp6_seq_show(struct seq_file *seq, void *v) static int sockstat6_seq_open(struct inode *inode, struct file *file) { - int err; - struct net *net; - - err = -ENXIO; - net = get_proc_net(inode); - if (net == NULL) - goto err_net; - - err = single_open(file, sockstat6_seq_show, net); - if (err < 0) - goto err_open; - - return 0; - -err_open: - put_net(net); -err_net: - return err; + return single_open_net(inode, file, sockstat6_seq_show); } static int sockstat6_seq_release(struct inode *inode, struct file *file) diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 5d6c166dfbb6..fb7ff8f0c6db 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -2416,18 +2416,7 @@ static int ipv6_route_show(struct seq_file *m, void *v) static int ipv6_route_open(struct inode *inode, struct file *file) { - int err; - struct net *net = get_proc_net(inode); - if (!net) - return -ENXIO; - - err = single_open(file, ipv6_route_show, net); - if (err < 0) { - put_net(net); - return err; - } - - return 0; + return single_open_net(inode, file, ipv6_route_show); } static int ipv6_route_release(struct inode *inode, struct file *file) @@ -2463,18 +2452,7 @@ static int rt6_stats_seq_show(struct seq_file *seq, void *v) static int rt6_stats_seq_open(struct inode *inode, struct file *file) { - int err; - struct net *net = get_proc_net(inode); - if (!net) - return -ENXIO; - - err = single_open(file, rt6_stats_seq_show, net); - if (err < 0) { - put_net(net); - return err; - } - - return 0; + return single_open_net(inode, file, rt6_stats_seq_show); } static int rt6_stats_seq_release(struct inode *inode, struct file *file) -- cgit v1.2.2 From b6fcbdb4f283f7ba67cec3cda6be23da8e959031 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Fri, 18 Jul 2008 04:07:44 -0700 Subject: proc: consolidate per-net single-release callers They are symmetrical to single_open ones :) Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller --- net/ipv4/fib_trie.c | 9 +-------- net/ipv4/proc.c | 30 +++--------------------------- net/ipv6/proc.c | 10 +--------- net/ipv6/route.c | 20 ++------------------ 4 files changed, 7 insertions(+), 62 deletions(-) (limited to 'net') diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c index 6009df238ed9..5cb72786a8af 100644 --- a/net/ipv4/fib_trie.c +++ b/net/ipv4/fib_trie.c @@ -2254,19 +2254,12 @@ static int fib_triestat_seq_open(struct inode *inode, struct file *file) return single_open_net(inode, file, fib_triestat_seq_show); } -static int fib_triestat_seq_release(struct inode *ino, struct file *f) -{ - struct seq_file *seq = f->private_data; - put_net(seq->private); - return single_release(ino, f); -} - static const struct file_operations fib_triestat_fops = { .owner = THIS_MODULE, .open = fib_triestat_seq_open, .read = seq_read, .llseek = seq_lseek, - .release = fib_triestat_seq_release, + .release = single_release_net, }; static struct node *fib_trie_get_idx(struct seq_file *seq, loff_t pos) diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c index daf5d3c80cef..834356ea99df 100644 --- a/net/ipv4/proc.c +++ b/net/ipv4/proc.c @@ -74,20 +74,12 @@ static int sockstat_seq_open(struct inode *inode, struct file *file) return single_open_net(inode, file, sockstat_seq_show); } -static int sockstat_seq_release(struct inode *inode, struct file *file) -{ - struct net *net = ((struct seq_file *)file->private_data)->private; - - put_net(net); - return single_release(inode, file); -} - static const struct file_operations sockstat_seq_fops = { .owner = THIS_MODULE, .open = sockstat_seq_open, .read = seq_read, .llseek = seq_lseek, - .release = sockstat_seq_release, + .release = single_release_net, }; /* snmp items */ @@ -383,20 +375,12 @@ static int snmp_seq_open(struct inode *inode, struct file *file) return single_open_net(inode, file, snmp_seq_show); } -static int snmp_seq_release(struct inode *inode, struct file *file) -{ - struct net *net = ((struct seq_file *)file->private_data)->private; - - put_net(net); - return single_release(inode, file); -} - static const struct file_operations snmp_seq_fops = { .owner = THIS_MODULE, .open = snmp_seq_open, .read = seq_read, .llseek = seq_lseek, - .release = snmp_seq_release, + .release = single_release_net, }; @@ -438,20 +422,12 @@ static int netstat_seq_open(struct inode *inode, struct file *file) return single_open_net(inode, file, netstat_seq_show); } -static int netstat_seq_release(struct inode *inode, struct file *file) -{ - struct net *net = ((struct seq_file *)file->private_data)->private; - - put_net(net); - return single_release(inode, file); -} - static const struct file_operations netstat_seq_fops = { .owner = THIS_MODULE, .open = netstat_seq_open, .read = seq_read, .llseek = seq_lseek, - .release = netstat_seq_release, + .release = single_release_net, }; static __net_init int ip_proc_init_net(struct net *net) diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c index 29c5a79444c2..70940b3654a1 100644 --- a/net/ipv6/proc.c +++ b/net/ipv6/proc.c @@ -186,20 +186,12 @@ static int sockstat6_seq_open(struct inode *inode, struct file *file) return single_open_net(inode, file, sockstat6_seq_show); } -static int sockstat6_seq_release(struct inode *inode, struct file *file) -{ - struct net *net = ((struct seq_file *)file->private_data)->private; - - put_net(net); - return single_release(inode, file); -} - static const struct file_operations sockstat6_seq_fops = { .owner = THIS_MODULE, .open = sockstat6_seq_open, .read = seq_read, .llseek = seq_lseek, - .release = sockstat6_seq_release, + .release = single_release_net, }; static int snmp6_seq_open(struct inode *inode, struct file *file) diff --git a/net/ipv6/route.c b/net/ipv6/route.c index fb7ff8f0c6db..cb8a51271b67 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -2419,20 +2419,12 @@ static int ipv6_route_open(struct inode *inode, struct file *file) return single_open_net(inode, file, ipv6_route_show); } -static int ipv6_route_release(struct inode *inode, struct file *file) -{ - struct seq_file *seq = file->private_data; - struct net *net = seq->private; - put_net(net); - return single_release(inode, file); -} - static const struct file_operations ipv6_route_proc_fops = { .owner = THIS_MODULE, .open = ipv6_route_open, .read = seq_read, .llseek = seq_lseek, - .release = ipv6_route_release, + .release = single_release_net, }; static int rt6_stats_seq_show(struct seq_file *seq, void *v) @@ -2455,20 +2447,12 @@ static int rt6_stats_seq_open(struct inode *inode, struct file *file) return single_open_net(inode, file, rt6_stats_seq_show); } -static int rt6_stats_seq_release(struct inode *inode, struct file *file) -{ - struct seq_file *seq = file->private_data; - struct net *net = (struct net *)seq->private; - put_net(net); - return single_release(inode, file); -} - static const struct file_operations rt6_stats_seq_fops = { .owner = THIS_MODULE, .open = rt6_stats_seq_open, .read = seq_read, .llseek = seq_lseek, - .release = rt6_stats_seq_release, + .release = single_release_net, }; #endif /* CONFIG_PROC_FS */ -- cgit v1.2.2 From 8913336a7e8d56e984109a3137d6c0e3362596a4 Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Fri, 18 Jul 2008 18:05:19 -0700 Subject: packet: add PACKET_RESERVE sockopt Add new sockopt to reserve some headroom in the mmaped ring frames in front of the packet payload. This can be used f.i. when the VLAN header needs to be (re)constructed to avoid moving the entire payload. Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller --- net/packet/af_packet.c | 29 ++++++++++++++++++++++++++--- 1 file changed, 26 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index db792e02a37f..de73bcb5235a 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -188,6 +188,7 @@ struct packet_sock { unsigned int pg_vec_len; enum tpacket_versions tp_version; unsigned int tp_hdrlen; + unsigned int tp_reserve; #endif }; @@ -635,11 +636,13 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, struct packe snaplen = res; if (sk->sk_type == SOCK_DGRAM) { - macoff = netoff = TPACKET_ALIGN(po->tp_hdrlen) + 16; + macoff = netoff = TPACKET_ALIGN(po->tp_hdrlen) + 16 + + po->tp_reserve; } else { unsigned maclen = skb_network_offset(skb); netoff = TPACKET_ALIGN(po->tp_hdrlen + - (maclen < 16 ? 16 : maclen)); + (maclen < 16 ? 16 : maclen)) + + po->tp_reserve; macoff = netoff - maclen; } @@ -1448,6 +1451,19 @@ packet_setsockopt(struct socket *sock, int level, int optname, char __user *optv return -EINVAL; } } + case PACKET_RESERVE: + { + unsigned int val; + + if (optlen != sizeof(val)) + return -EINVAL; + if (po->pg_vec) + return -EBUSY; + if (copy_from_user(&val, optval, sizeof(val))) + return -EFAULT; + po->tp_reserve = val; + return 0; + } #endif case PACKET_AUXDATA: { @@ -1547,6 +1563,12 @@ static int packet_getsockopt(struct socket *sock, int level, int optname, } data = &val; break; + case PACKET_RESERVE: + if (len > sizeof(unsigned int)) + len = sizeof(unsigned int); + val = po->tp_reserve; + data = &val; + break; #endif default: return -ENOPROTOOPT; @@ -1790,7 +1812,8 @@ static int packet_set_ring(struct sock *sk, struct tpacket_req *req, int closing return -EINVAL; if (unlikely(req->tp_block_size & (PAGE_SIZE - 1))) return -EINVAL; - if (unlikely(req->tp_frame_size < po->tp_hdrlen)) + if (unlikely(req->tp_frame_size < po->tp_hdrlen + + po->tp_reserve)) return -EINVAL; if (unlikely(req->tp_frame_size & (TPACKET_ALIGNMENT - 1))) return -EINVAL; -- cgit v1.2.2 From 72b25a913ed9b1ab49c7022adaf3f271a65ea219 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 18 Jul 2008 20:54:17 -0700 Subject: pkt_sched: Get rid of u32_list. The u32_list is just an indirect way of maintaining a reference to a U32 node on a per-qdisc basis. Just add an explicit node pointer for u32 to struct Qdisc an do away with this global list. Signed-off-by: David S. Miller --- net/sched/cls_u32.c | 18 +++--------------- 1 file changed, 3 insertions(+), 15 deletions(-) (limited to 'net') diff --git a/net/sched/cls_u32.c b/net/sched/cls_u32.c index 4d755444c449..527db2559dd2 100644 --- a/net/sched/cls_u32.c +++ b/net/sched/cls_u32.c @@ -75,7 +75,6 @@ struct tc_u_hnode struct tc_u_common { - struct tc_u_common *next; struct tc_u_hnode *hlist; struct Qdisc *q; int refcnt; @@ -87,8 +86,6 @@ static const struct tcf_ext_map u32_ext_map = { .police = TCA_U32_POLICE }; -static struct tc_u_common *u32_list; - static __inline__ unsigned u32_hash_fold(__be32 key, struct tc_u32_sel *sel, u8 fshift) { unsigned h = ntohl(key & sel->hmask)>>fshift; @@ -287,9 +284,7 @@ static int u32_init(struct tcf_proto *tp) struct tc_u_hnode *root_ht; struct tc_u_common *tp_c; - for (tp_c = u32_list; tp_c; tp_c = tp_c->next) - if (tp_c->q == tp->q) - break; + tp_c = tp->q->u32_node; root_ht = kzalloc(sizeof(*root_ht), GFP_KERNEL); if (root_ht == NULL) @@ -307,8 +302,7 @@ static int u32_init(struct tcf_proto *tp) return -ENOBUFS; } tp_c->q = tp->q; - tp_c->next = u32_list; - u32_list = tp_c; + tp->q->u32_node = tp_c; } tp_c->refcnt++; @@ -402,14 +396,8 @@ static void u32_destroy(struct tcf_proto *tp) if (--tp_c->refcnt == 0) { struct tc_u_hnode *ht; - struct tc_u_common **tp_cp; - for (tp_cp = &u32_list; *tp_cp; tp_cp = &(*tp_cp)->next) { - if (*tp_cp == tp_c) { - *tp_cp = tp_c->next; - break; - } - } + tp->q->u32_node = NULL; for (ht = tp_c->hlist; ht; ht = ht->next) { ht->refcnt--; -- cgit v1.2.2 From 3072367300aa8c779e3a14ee8e89de079e90f3ad Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 18 Jul 2008 22:50:15 -0700 Subject: pkt_sched: Manage qdisc list inside of root qdisc. Idea is from Patrick McHardy. Instead of managing the list of qdiscs on the device level, manage it in the root qdisc of a netdev_queue. This solves all kinds of visibility issues during qdisc destruction. The way to iterate over all qdiscs of a netdev_queue is to visit the netdev_queue->qdisc, and then traverse it's list. The only special case is to ignore builting qdiscs at the root when dumping or doing a qdisc_lookup(). That was not needed previously because builtin qdiscs were not added to the device's qdisc_list. Signed-off-by: David S. Miller --- net/core/dev.c | 2 - net/sched/sch_api.c | 175 ++++++++++++++++++++++++++++++++++++------------ net/sched/sch_generic.c | 10 +-- 3 files changed, 133 insertions(+), 54 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index e54acde839da..065b9817e209 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3888,8 +3888,6 @@ int register_netdevice(struct net_device *dev) net = dev_net(dev); spin_lock_init(&dev->addr_list_lock); - spin_lock_init(&dev->qdisc_list_lock); - INIT_LIST_HEAD(&dev->qdisc_list); netdev_init_queue_locks(dev); dev->iflink = -1; diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index b3ef8307204e..fb43731c9860 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -185,11 +185,20 @@ EXPORT_SYMBOL(unregister_qdisc); struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle) { - struct Qdisc *q; + unsigned int i; + + for (i = 0; i < dev->num_tx_queues; i++) { + struct netdev_queue *txq = netdev_get_tx_queue(dev, i); + struct Qdisc *q, *txq_root = txq->qdisc; - list_for_each_entry(q, &dev->qdisc_list, list) { - if (q->handle == handle) - return q; + if (!(txq_root->flags & TCQ_F_BUILTIN) && + txq_root->handle == handle) + return txq_root; + + list_for_each_entry(q, &txq_root->list, list) { + if (q->handle == handle) + return q; + } } return NULL; } @@ -676,9 +685,8 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, goto err_out3; } } - spin_lock_bh(&dev->qdisc_list_lock); - list_add_tail(&sch->list, &dev->qdisc_list); - spin_unlock_bh(&dev->qdisc_list_lock); + if (parent) + list_add_tail(&sch->list, &dev_queue->qdisc->list); return sch; } @@ -1037,13 +1045,57 @@ err_out: return -EINVAL; } +static bool tc_qdisc_dump_ignore(struct Qdisc *q) +{ + return (q->flags & TCQ_F_BUILTIN) ? true : false; +} + +static int tc_dump_qdisc_root(struct Qdisc *root, struct sk_buff *skb, + struct netlink_callback *cb, + int *q_idx_p, int s_q_idx) +{ + int ret = 0, q_idx = *q_idx_p; + struct Qdisc *q; + + if (!root) + return 0; + + q = root; + if (q_idx < s_q_idx) { + q_idx++; + } else { + if (!tc_qdisc_dump_ignore(q) && + tc_fill_qdisc(skb, q, q->parent, NETLINK_CB(cb->skb).pid, + cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWQDISC) <= 0) + goto done; + q_idx++; + } + list_for_each_entry(q, &root->list, list) { + if (q_idx < s_q_idx) { + q_idx++; + continue; + } + if (!tc_qdisc_dump_ignore(q) && + tc_fill_qdisc(skb, q, q->parent, NETLINK_CB(cb->skb).pid, + cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWQDISC) <= 0) + goto done; + q_idx++; + } + +out: + *q_idx_p = q_idx; + return ret; +done: + ret = -1; + goto out; +} + static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb) { struct net *net = sock_net(skb->sk); int idx, q_idx; int s_idx, s_q_idx; struct net_device *dev; - struct Qdisc *q; if (net != &init_net) return 0; @@ -1053,21 +1105,22 @@ static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb) read_lock(&dev_base_lock); idx = 0; for_each_netdev(&init_net, dev) { + struct netdev_queue *dev_queue; + if (idx < s_idx) goto cont; if (idx > s_idx) s_q_idx = 0; q_idx = 0; - list_for_each_entry(q, &dev->qdisc_list, list) { - if (q_idx < s_q_idx) { - q_idx++; - continue; - } - if (tc_fill_qdisc(skb, q, q->parent, NETLINK_CB(cb->skb).pid, - cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWQDISC) <= 0) - goto done; - q_idx++; - } + + dev_queue = netdev_get_tx_queue(dev, 0); + if (tc_dump_qdisc_root(dev_queue->qdisc, skb, cb, &q_idx, s_q_idx) < 0) + goto done; + + dev_queue = &dev->rx_queue; + if (tc_dump_qdisc_root(dev_queue->qdisc, skb, cb, &q_idx, s_q_idx) < 0) + goto done; + cont: idx++; } @@ -1285,15 +1338,62 @@ static int qdisc_class_dump(struct Qdisc *q, unsigned long cl, struct qdisc_walk a->cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWTCLASS); } +static int tc_dump_tclass_qdisc(struct Qdisc *q, struct sk_buff *skb, + struct tcmsg *tcm, struct netlink_callback *cb, + int *t_p, int s_t) +{ + struct qdisc_dump_args arg; + + if (tc_qdisc_dump_ignore(q) || + *t_p < s_t || !q->ops->cl_ops || + (tcm->tcm_parent && + TC_H_MAJ(tcm->tcm_parent) != q->handle)) { + (*t_p)++; + return 0; + } + if (*t_p > s_t) + memset(&cb->args[1], 0, sizeof(cb->args)-sizeof(cb->args[0])); + arg.w.fn = qdisc_class_dump; + arg.skb = skb; + arg.cb = cb; + arg.w.stop = 0; + arg.w.skip = cb->args[1]; + arg.w.count = 0; + q->ops->cl_ops->walk(q, &arg.w); + cb->args[1] = arg.w.count; + if (arg.w.stop) + return -1; + (*t_p)++; + return 0; +} + +static int tc_dump_tclass_root(struct Qdisc *root, struct sk_buff *skb, + struct tcmsg *tcm, struct netlink_callback *cb, + int *t_p, int s_t) +{ + struct Qdisc *q; + + if (!root) + return 0; + + if (tc_dump_tclass_qdisc(root, skb, tcm, cb, t_p, s_t) < 0) + return -1; + + list_for_each_entry(q, &root->list, list) { + if (tc_dump_tclass_qdisc(q, skb, tcm, cb, t_p, s_t) < 0) + return -1; + } + + return 0; +} + static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) { + struct tcmsg *tcm = (struct tcmsg*)NLMSG_DATA(cb->nlh); struct net *net = sock_net(skb->sk); - int t; - int s_t; + struct netdev_queue *dev_queue; struct net_device *dev; - struct Qdisc *q; - struct tcmsg *tcm = (struct tcmsg*)NLMSG_DATA(cb->nlh); - struct qdisc_dump_args arg; + int t, s_t; if (net != &init_net) return 0; @@ -1306,28 +1406,15 @@ static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb) s_t = cb->args[0]; t = 0; - list_for_each_entry(q, &dev->qdisc_list, list) { - if (t < s_t || !q->ops->cl_ops || - (tcm->tcm_parent && - TC_H_MAJ(tcm->tcm_parent) != q->handle)) { - t++; - continue; - } - if (t > s_t) - memset(&cb->args[1], 0, sizeof(cb->args)-sizeof(cb->args[0])); - arg.w.fn = qdisc_class_dump; - arg.skb = skb; - arg.cb = cb; - arg.w.stop = 0; - arg.w.skip = cb->args[1]; - arg.w.count = 0; - q->ops->cl_ops->walk(q, &arg.w); - cb->args[1] = arg.w.count; - if (arg.w.stop) - break; - t++; - } + dev_queue = netdev_get_tx_queue(dev, 0); + if (tc_dump_tclass_root(dev_queue->qdisc, skb, tcm, cb, &t, s_t) < 0) + goto done; + + dev_queue = &dev->rx_queue; + if (tc_dump_tclass_root(dev_queue->qdisc, skb, tcm, cb, &t, s_t) < 0) + goto done; +done: cb->args[0] = t; dev_put(dev); diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index e244c462e6bd..14cc443d0490 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -480,15 +480,12 @@ static void __qdisc_destroy(struct rcu_head *head) void qdisc_destroy(struct Qdisc *qdisc) { - struct net_device *dev = qdisc_dev(qdisc); - if (qdisc->flags & TCQ_F_BUILTIN || !atomic_dec_and_test(&qdisc->refcnt)) return; - spin_lock_bh(&dev->qdisc_list_lock); - list_del(&qdisc->list); - spin_unlock_bh(&dev->qdisc_list_lock); + if (qdisc->parent) + list_del(&qdisc->list); call_rcu(&qdisc->q_rcu, __qdisc_destroy); } @@ -520,9 +517,6 @@ static void attach_one_default_qdisc(struct net_device *dev, printk(KERN_INFO "%s: activation failed\n", dev->name); return; } - spin_lock_bh(&dev->qdisc_list_lock); - list_add_tail(&qdisc->list, &dev->qdisc_list); - spin_unlock_bh(&dev->qdisc_list_lock); } else { qdisc = &noqueue_qdisc; } -- cgit v1.2.2 From 30ee42be00b7a50929a73cb617f70b1d3219eb69 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 18 Jul 2008 23:00:11 -0700 Subject: pkt_sched: Fix noqueue_qdisc initialization. Like noop_qdisc, it needs a dummy backpointer and explicit qdisc->q.lock initialization. Based upon a report by Stephen Hemminger. Signed-off-by: David S. Miller --- net/sched/sch_generic.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'net') diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 14cc443d0490..522a41a9f904 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -340,12 +340,19 @@ static struct Qdisc_ops noqueue_qdisc_ops __read_mostly = { .owner = THIS_MODULE, }; +static struct Qdisc noqueue_qdisc; +static struct netdev_queue noqueue_netdev_queue = { + .qdisc = &noqueue_qdisc, +}; + static struct Qdisc noqueue_qdisc = { .enqueue = NULL, .dequeue = noop_dequeue, .flags = TCQ_F_BUILTIN, .ops = &noqueue_qdisc_ops, .list = LIST_HEAD_INIT(noqueue_qdisc.list), + .q.lock = __SPIN_LOCK_UNLOCKED(noqueue_qdisc.q.lock), + .dev_queue = &noqueue_netdev_queue, }; -- cgit v1.2.2 From c1e20f7c8b9ccbafc9ea78f2b406738728ce6b81 Mon Sep 17 00:00:00 2001 From: Stephen Hemminger Date: Fri, 18 Jul 2008 23:02:15 -0700 Subject: tcp: RTT metrics scaling Some of the metrics (RTT, RTTVAR and RTAX_RTO_MIN) are stored in kernel units (jiffies) and this leaks out through the netlink API to user space where the units for jiffies are unknown. This patches changes the kernel to convert to/from milliseconds. This changes the ABI, but milliseconds seemed like the most natural unit for these parameters. Values available via syscall in /proc/net/rt_cache and netlink will be in milliseconds. Signed-off-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/ipv4/tcp_input.c | 31 ++++++++++++++++++------------- 1 file changed, 18 insertions(+), 13 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index fac49a6e1611..88810bc01370 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -602,7 +602,7 @@ static u32 tcp_rto_min(struct sock *sk) u32 rto_min = TCP_RTO_MIN; if (dst && dst_metric_locked(dst, RTAX_RTO_MIN)) - rto_min = dst_metric(dst, RTAX_RTO_MIN); + rto_min = dst_metric_rtt(dst, RTAX_RTO_MIN); return rto_min; } @@ -729,6 +729,7 @@ void tcp_update_metrics(struct sock *sk) if (dst && (dst->flags & DST_HOST)) { const struct inet_connection_sock *icsk = inet_csk(sk); int m; + unsigned long rtt; if (icsk->icsk_backoff || !tp->srtt) { /* This session failed to estimate rtt. Why? @@ -740,7 +741,8 @@ void tcp_update_metrics(struct sock *sk) return; } - m = dst_metric(dst, RTAX_RTT) - tp->srtt; + rtt = dst_metric_rtt(dst, RTAX_RTT); + m = rtt - tp->srtt; /* If newly calculated rtt larger than stored one, * store new one. Otherwise, use EWMA. Remember, @@ -748,12 +750,13 @@ void tcp_update_metrics(struct sock *sk) */ if (!(dst_metric_locked(dst, RTAX_RTT))) { if (m <= 0) - dst->metrics[RTAX_RTT - 1] = tp->srtt; + set_dst_metric_rtt(dst, RTAX_RTT, tp->srtt); else - dst->metrics[RTAX_RTT - 1] -= (m >> 3); + set_dst_metric_rtt(dst, RTAX_RTT, rtt - (m >> 3)); } if (!(dst_metric_locked(dst, RTAX_RTTVAR))) { + unsigned long var; if (m < 0) m = -m; @@ -762,11 +765,13 @@ void tcp_update_metrics(struct sock *sk) if (m < tp->mdev) m = tp->mdev; - if (m >= dst_metric(dst, RTAX_RTTVAR)) - dst->metrics[RTAX_RTTVAR - 1] = m; + var = dst_metric_rtt(dst, RTAX_RTTVAR); + if (m >= var) + var = m; else - dst->metrics[RTAX_RTTVAR-1] -= - (dst_metric(dst, RTAX_RTTVAR) - m)>>2; + var -= (var - m) >> 2; + + set_dst_metric_rtt(dst, RTAX_RTTVAR, var); } if (tp->snd_ssthresh >= 0xFFFF) { @@ -897,7 +902,7 @@ static void tcp_init_metrics(struct sock *sk) if (dst_metric(dst, RTAX_RTT) == 0) goto reset; - if (!tp->srtt && dst_metric(dst, RTAX_RTT) < (TCP_TIMEOUT_INIT << 3)) + if (!tp->srtt && dst_metric_rtt(dst, RTAX_RTT) < (TCP_TIMEOUT_INIT << 3)) goto reset; /* Initial rtt is determined from SYN,SYN-ACK. @@ -914,12 +919,12 @@ static void tcp_init_metrics(struct sock *sk) * to low value, and then abruptly stops to do it and starts to delay * ACKs, wait for troubles. */ - if (dst_metric(dst, RTAX_RTT) > tp->srtt) { - tp->srtt = dst_metric(dst, RTAX_RTT); + if (dst_metric_rtt(dst, RTAX_RTT) > tp->srtt) { + tp->srtt = dst_metric_rtt(dst, RTAX_RTT); tp->rtt_seq = tp->snd_nxt; } - if (dst_metric(dst, RTAX_RTTVAR) > tp->mdev) { - tp->mdev = dst_metric(dst, RTAX_RTTVAR); + if (dst_metric_rtt(dst, RTAX_RTTVAR) > tp->mdev) { + tp->mdev = dst_metric_rtt(dst, RTAX_RTTVAR); tp->mdev_max = tp->rttvar = max(tp->mdev, tcp_rto_min(sk)); } tcp_set_rto(sk); -- cgit v1.2.2 From c4e85f82edcd6027cfe67331a2e00741b009756b Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Fri, 18 Jul 2008 23:03:44 -0700 Subject: sctp: Don't abort initialization when CONFIG_PROC_FS=n This puts CONFIG_PROC_FS defines around the proc init/exit functions and also avoids compiling proc.c if procfs is not supported. Also make SCTP_DBG_OBJCNT depend on procfs. Signed-off-by: Florian Westphal Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/Kconfig | 4 ++-- net/sctp/Makefile | 4 ++-- net/sctp/protocol.c | 11 ++++++++++- 3 files changed, 14 insertions(+), 5 deletions(-) (limited to 'net') diff --git a/net/sctp/Kconfig b/net/sctp/Kconfig index 0b79f869c4ea..58b3e882a187 100644 --- a/net/sctp/Kconfig +++ b/net/sctp/Kconfig @@ -47,11 +47,11 @@ config SCTP_DBG_MSG config SCTP_DBG_OBJCNT bool "SCTP: Debug object counts" + depends on PROC_FS help If you say Y, this will enable debugging support for counting the type of objects that are currently allocated. This is useful for - identifying memory leaks. If the /proc filesystem is enabled this - debug information can be viewed by + identifying memory leaks. This debug information can be viewed by 'cat /proc/net/sctp/sctp_dbg_objcnt' If unsure, say N diff --git a/net/sctp/Makefile b/net/sctp/Makefile index f5356b9d5ee3..6b794734380a 100644 --- a/net/sctp/Makefile +++ b/net/sctp/Makefile @@ -9,10 +9,10 @@ sctp-y := sm_statetable.o sm_statefuns.o sm_sideeffect.o \ transport.o chunk.o sm_make_chunk.o ulpevent.o \ inqueue.o outqueue.o ulpqueue.o command.o \ tsnmap.o bind_addr.o socket.o primitive.o \ - output.o input.o debug.o ssnmap.o proc.o \ - auth.o + output.o input.o debug.o ssnmap.o auth.o sctp-$(CONFIG_SCTP_DBG_OBJCNT) += objcnt.o +sctp-$(CONFIG_PROC_FS) += proc.o sctp-$(CONFIG_SYSCTL) += sysctl.o sctp-$(subst m,y,$(CONFIG_IPV6)) += ipv6.o diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c index 98c6a882016a..dd811a8456a8 100644 --- a/net/sctp/protocol.c +++ b/net/sctp/protocol.c @@ -64,9 +64,12 @@ /* Global data structures. */ struct sctp_globals sctp_globals __read_mostly; -struct proc_dir_entry *proc_net_sctp; DEFINE_SNMP_STAT(struct sctp_mib, sctp_statistics) __read_mostly; +#ifdef CONFIG_PROC_FS +struct proc_dir_entry *proc_net_sctp; +#endif + struct idr sctp_assocs_id; DEFINE_SPINLOCK(sctp_assocs_id_lock); @@ -97,6 +100,7 @@ struct sock *sctp_get_ctl_sock(void) /* Set up the proc fs entry for the SCTP protocol. */ static __init int sctp_proc_init(void) { +#ifdef CONFIG_PROC_FS if (!proc_net_sctp) { struct proc_dir_entry *ent; ent = proc_mkdir("sctp", init_net.proc_net); @@ -131,6 +135,9 @@ out_snmp_proc_init: } out_nomem: return -ENOMEM; +#else + return 0; +#endif /* CONFIG_PROC_FS */ } /* Clean up the proc fs entry for the SCTP protocol. @@ -139,6 +146,7 @@ out_nomem: */ static void sctp_proc_exit(void) { +#ifdef CONFIG_PROC_FS sctp_snmp_proc_exit(); sctp_eps_proc_exit(); sctp_assocs_proc_exit(); @@ -148,6 +156,7 @@ static void sctp_proc_exit(void) proc_net_sctp = NULL; remove_proc_entry("sctp", init_net.proc_net); } +#endif } /* Private helper to extract ipv4 address and stash them in -- cgit v1.2.2 From 6d0ccbac688207ca0616ab5094932af4db4747b3 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Fri, 18 Jul 2008 23:04:39 -0700 Subject: sctp: Prevent uninitialized memory access valgrind reports uninizialized memory accesses when running sctp inside the network simulation cradle simulator: Conditional jump or move depends on uninitialised value(s) at 0x570E34A: sctp_assoc_sync_pmtu (associola.c:1324) by 0x57427DA: sctp_packet_transmit (output.c:403) by 0x5710EFF: sctp_outq_flush (outqueue.c:824) by 0x5710B88: sctp_outq_uncork (outqueue.c:701) by 0x5745262: sctp_cmd_interpreter (sm_sideeffect.c:1548) by 0x57444B7: sctp_side_effects (sm_sideeffect.c:976) by 0x5744460: sctp_do_sm (sm_sideeffect.c:945) by 0x572157D: sctp_primitive_ASSOCIATE (primitive.c:94) by 0x5725C04: __sctp_connect (socket.c:1094) by 0x57297DC: sctp_connect (socket.c:3297) Conditional jump or move depends on uninitialised value(s) at 0x575D3A5: mod_timer (timer.c:630) by 0x5752B78: sctp_cmd_hb_timers_start (sm_sideeffect.c:555) by 0x5754133: sctp_cmd_interpreter (sm_sideeffect.c:1448) by 0x5753607: sctp_side_effects (sm_sideeffect.c:976) by 0x57535B0: sctp_do_sm (sm_sideeffect.c:945) by 0x571E9AE: sctp_endpoint_bh_rcv (endpointola.c:474) by 0x573347F: sctp_inq_push (inqueue.c:104) by 0x572EF93: sctp_rcv (input.c:256) by 0x5689623: ip_local_deliver_finish (ip_input.c:230) by 0x5689759: ip_local_deliver (ip_input.c:268) by 0x5689CAC: ip_rcv_finish (dst.h:246) #1 is due to "if (t->pmtu_pending)". 8a4794914f9cf2681235ec2311e189fe307c28c7 "[SCTP] Flag a pmtu change request" suggests it should be initialized to 0. #2 is the heartbeat timer 'expires' value, which is uninizialised, but test by mod_timer(). T3_rtx_timer seems to be affected by the same problem, so initialize it, too. Signed-off-by: Florian Westphal Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/associola.c | 1 + net/sctp/transport.c | 3 +++ 2 files changed, 4 insertions(+) (limited to 'net') diff --git a/net/sctp/associola.c b/net/sctp/associola.c index 35b6a023a6d0..ec2a0a33fd78 100644 --- a/net/sctp/associola.c +++ b/net/sctp/associola.c @@ -653,6 +653,7 @@ struct sctp_transport *sctp_assoc_add_peer(struct sctp_association *asoc, SCTP_DEBUG_PRINTK("sctp_assoc_add_peer:association %p PMTU set to " "%d\n", asoc, asoc->pathmtu); + peer->pmtu_pending = 0; asoc->frag_point = sctp_frag_point(sp, asoc->pathmtu); diff --git a/net/sctp/transport.c b/net/sctp/transport.c index 3f34f61221ec..e745c118f239 100644 --- a/net/sctp/transport.c +++ b/net/sctp/transport.c @@ -100,6 +100,9 @@ static struct sctp_transport *sctp_transport_init(struct sctp_transport *peer, INIT_LIST_HEAD(&peer->send_ready); INIT_LIST_HEAD(&peer->transports); + peer->T3_rtx_timer.expires = 0; + peer->hb_timer.expires = 0; + setup_timer(&peer->T3_rtx_timer, sctp_generate_t3_rtx_event, (unsigned long)peer); setup_timer(&peer->hb_timer, sctp_generate_heartbeat_event, -- cgit v1.2.2 From 7dab83de50c7b2b7ceac695a0b56fa6c0f95b0bc Mon Sep 17 00:00:00 2001 From: Vlad Yasevich Date: Fri, 18 Jul 2008 23:05:40 -0700 Subject: sctp: Support ipv6only AF_INET6 sockets. Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/bind_addr.c | 37 +++++++++++++++++++++++++++++++++++++ net/sctp/ipv6.c | 20 ++++++++++++++++---- net/sctp/protocol.c | 7 +++++++ net/sctp/sm_make_chunk.c | 7 ++++++- net/sctp/socket.c | 30 +++++++++++++++++++++++++----- 5 files changed, 91 insertions(+), 10 deletions(-) (limited to 'net') diff --git a/net/sctp/bind_addr.c b/net/sctp/bind_addr.c index 80e6df06967a..f62bc2468935 100644 --- a/net/sctp/bind_addr.c +++ b/net/sctp/bind_addr.c @@ -348,6 +348,43 @@ int sctp_bind_addr_match(struct sctp_bind_addr *bp, return match; } +/* Does the address 'addr' conflict with any addresses in + * the bp. + */ +int sctp_bind_addr_conflict(struct sctp_bind_addr *bp, + const union sctp_addr *addr, + struct sctp_sock *bp_sp, + struct sctp_sock *addr_sp) +{ + struct sctp_sockaddr_entry *laddr; + int conflict = 0; + struct sctp_sock *sp; + + /* Pick the IPv6 socket as the basis of comparison + * since it's usually a superset of the IPv4. + * If there is no IPv6 socket, then default to bind_addr. + */ + if (sctp_opt2sk(bp_sp)->sk_family == AF_INET6) + sp = bp_sp; + else if (sctp_opt2sk(addr_sp)->sk_family == AF_INET6) + sp = addr_sp; + else + sp = bp_sp; + + rcu_read_lock(); + list_for_each_entry_rcu(laddr, &bp->address_list, list) { + if (!laddr->valid) + continue; + + conflict = sp->pf->cmp_addr(&laddr->a, addr, sp); + if (conflict) + break; + } + rcu_read_unlock(); + + return conflict; +} + /* Get the state of the entry in the bind_addr_list */ int sctp_bind_addr_state(const struct sctp_bind_addr *bp, const union sctp_addr *addr) diff --git a/net/sctp/ipv6.c b/net/sctp/ipv6.c index a2f4d4d51593..a238d6834b33 100644 --- a/net/sctp/ipv6.c +++ b/net/sctp/ipv6.c @@ -818,7 +818,7 @@ static int sctp_inet6_af_supported(sa_family_t family, struct sctp_sock *sp) return 1; /* v4-mapped-v6 addresses */ case AF_INET: - if (!__ipv6_only_sock(sctp_opt2sk(sp)) && sp->v4mapped) + if (!__ipv6_only_sock(sctp_opt2sk(sp))) return 1; default: return 0; @@ -840,6 +840,11 @@ static int sctp_inet6_cmp_addr(const union sctp_addr *addr1, if (!af1 || !af2) return 0; + + /* If the socket is IPv6 only, v4 addrs will not match */ + if (__ipv6_only_sock(sctp_opt2sk(opt)) && af1 != af2) + return 0; + /* Today, wildcard AF_INET/AF_INET6. */ if (sctp_is_any(addr1) || sctp_is_any(addr2)) return 1; @@ -876,7 +881,11 @@ static int sctp_inet6_bind_verify(struct sctp_sock *opt, union sctp_addr *addr) return 0; } dev_put(dev); + } else if (type == IPV6_ADDR_MAPPED) { + if (!opt->v4mapped) + return 0; } + af = opt->pf->af; } return af->available(addr, opt); @@ -919,9 +928,12 @@ static int sctp_inet6_send_verify(struct sctp_sock *opt, union sctp_addr *addr) static int sctp_inet6_supported_addrs(const struct sctp_sock *opt, __be16 *types) { - types[0] = SCTP_PARAM_IPV4_ADDRESS; - types[1] = SCTP_PARAM_IPV6_ADDRESS; - return 2; + types[0] = SCTP_PARAM_IPV6_ADDRESS; + if (!opt || !ipv6_only_sock(sctp_opt2sk(opt))) { + types[1] = SCTP_PARAM_IPV4_ADDRESS; + return 2; + } + return 1; } static const struct proto_ops inet6_seqpacket_ops = { diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c index dd811a8456a8..cdd881142f2c 100644 --- a/net/sctp/protocol.c +++ b/net/sctp/protocol.c @@ -381,6 +381,10 @@ static int sctp_v4_addr_valid(union sctp_addr *addr, struct sctp_sock *sp, const struct sk_buff *skb) { + /* IPv4 addresses not allowed */ + if (sp && ipv6_only_sock(sctp_opt2sk(sp))) + return 0; + /* Is this a non-unicast address or a unusable SCTP address? */ if (IS_IPV4_UNUSABLE_ADDRESS(addr->v4.sin_addr.s_addr)) return 0; @@ -404,6 +408,9 @@ static int sctp_v4_available(union sctp_addr *addr, struct sctp_sock *sp) !sysctl_ip_nonlocal_bind) return 0; + if (ipv6_only_sock(sctp_opt2sk(sp))) + return 0; + return 1; } diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c index bbc7107c86cf..e8ca4e54981f 100644 --- a/net/sctp/sm_make_chunk.c +++ b/net/sctp/sm_make_chunk.c @@ -2364,8 +2364,13 @@ static int sctp_process_param(struct sctp_association *asoc, case SCTP_PARAM_IPV6_ADDRESS: if (PF_INET6 != asoc->base.sk->sk_family) break; - /* Fall through. */ + goto do_addr_param; + case SCTP_PARAM_IPV4_ADDRESS: + /* v4 addresses are not allowed on v6-only socket */ + if (ipv6_only_sock(asoc->base.sk)) + break; +do_addr_param: af = sctp_get_af_specific(param_type2af(param.p->type)); af->from_addr_param(&addr, param.addr, htons(asoc->peer.port), 0); scope = sctp_scope(peer_addr); diff --git a/net/sctp/socket.c b/net/sctp/socket.c index 6aba01b0ce4e..a0e879bb202d 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -308,9 +308,16 @@ static struct sctp_af *sctp_sockaddr_af(struct sctp_sock *opt, if (len < sizeof (struct sockaddr)) return NULL; - /* Does this PF support this AF? */ - if (!opt->pf->af_supported(addr->sa.sa_family, opt)) - return NULL; + /* V4 mapped address are really of AF_INET family */ + if (addr->sa.sa_family == AF_INET6 && + ipv6_addr_v4mapped(&addr->v6.sin6_addr)) { + if (!opt->pf->af_supported(AF_INET, opt)) + return NULL; + } else { + /* Does this PF support this AF? */ + if (!opt->pf->af_supported(addr->sa.sa_family, opt)) + return NULL; + } /* If we get this far, af is valid. */ af = sctp_get_af_specific(addr->sa.sa_family); @@ -4395,6 +4402,11 @@ static int sctp_getsockopt_local_addrs_num_old(struct sock *sk, int len, (AF_INET6 == addr->a.sa.sa_family)) continue; + if ((PF_INET6 == sk->sk_family) && + inet_v6_ipv6only(sk) && + (AF_INET == addr->a.sa.sa_family)) + continue; + cnt++; } rcu_read_unlock(); @@ -4435,6 +4447,10 @@ static int sctp_copy_laddrs_old(struct sock *sk, __u16 port, if ((PF_INET == sk->sk_family) && (AF_INET6 == addr->a.sa.sa_family)) continue; + if ((PF_INET6 == sk->sk_family) && + inet_v6_ipv6only(sk) && + (AF_INET == addr->a.sa.sa_family)) + continue; memcpy(&temp, &addr->a, sizeof(temp)); if (!temp.v4.sin_port) temp.v4.sin_port = htons(port); @@ -4470,6 +4486,10 @@ static int sctp_copy_laddrs(struct sock *sk, __u16 port, void *to, if ((PF_INET == sk->sk_family) && (AF_INET6 == addr->a.sa.sa_family)) continue; + if ((PF_INET6 == sk->sk_family) && + inet_v6_ipv6only(sk) && + (AF_INET == addr->a.sa.sa_family)) + continue; memcpy(&temp, &addr->a, sizeof(temp)); if (!temp.v4.sin_port) temp.v4.sin_port = htons(port); @@ -5568,8 +5588,8 @@ pp_found: sk2->sk_state != SCTP_SS_LISTENING) continue; - if (sctp_bind_addr_match(&ep2->base.bind_addr, addr, - sctp_sk(sk))) { + if (sctp_bind_addr_conflict(&ep2->base.bind_addr, addr, + sctp_sk(sk2), sctp_sk(sk))) { ret = (long)sk2; goto fail_unlock; } -- cgit v1.2.2 From 23b29ed80bd7184398317a111dc488605cb66c7f Mon Sep 17 00:00:00 2001 From: Vlad Yasevich Date: Fri, 18 Jul 2008 23:06:07 -0700 Subject: sctp: Do not leak memory on multiple listen() calls SCTP permits multiple listen call and on subsequent calls we leak he memory allocated for the crypto transforms. Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/socket.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/sctp/socket.c b/net/sctp/socket.c index a0e879bb202d..fd7ed9d46a48 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -5773,7 +5773,7 @@ int sctp_inet_listen(struct socket *sock, int backlog) goto out; /* Allocate HMAC for generating cookie. */ - if (sctp_hmac_alg) { + if (!sctp_sk(sk)->hmac && sctp_hmac_alg) { tfm = crypto_alloc_hash(sctp_hmac_alg, 0, CRYPTO_ALG_ASYNC); if (IS_ERR(tfm)) { if (net_ratelimit()) { @@ -5801,7 +5801,8 @@ int sctp_inet_listen(struct socket *sock, int backlog) goto cleanup; /* Store away the transform reference. */ - sctp_sk(sk)->hmac = tfm; + if (!sctp_sk(sk)->hmac) + sctp_sk(sk)->hmac = tfm; out: sctp_release_sock(sk); return err; -- cgit v1.2.2 From 4e54064e0a13b7a7d4a481123c1783f770538e30 Mon Sep 17 00:00:00 2001 From: Vlad Yasevich Date: Fri, 18 Jul 2008 23:06:32 -0700 Subject: sctp: Allow only 1 listening socket with SO_REUSEADDR When multiple socket bind to the same port with SO_REUSEADDR, only 1 can be listining. Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/socket.c | 25 ++++++++++++++++--------- 1 file changed, 16 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/sctp/socket.c b/net/sctp/socket.c index fd7ed9d46a48..79bece16aede 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -377,18 +377,19 @@ SCTP_STATIC int sctp_do_bind(struct sock *sk, union sctp_addr *addr, int len) if (snum && snum < PROT_SOCK && !capable(CAP_NET_BIND_SERVICE)) return -EACCES; + /* See if the address matches any of the addresses we may have + * already bound before checking against other endpoints. + */ + if (sctp_bind_addr_match(bp, addr, sp)) + return -EINVAL; + /* Make sure we are allowed to bind here. * The function sctp_get_port_local() does duplicate address * detection. */ addr->v4.sin_port = htons(snum); if ((ret = sctp_get_port_local(sk, addr))) { - if (ret == (long) sk) { - /* This endpoint has a conflicting address. */ - return -EINVAL; - } else { - return -EADDRINUSE; - } + return -EADDRINUSE; } /* Refresh ephemeral port. */ @@ -5584,8 +5585,9 @@ pp_found: struct sctp_endpoint *ep2; ep2 = sctp_sk(sk2)->ep; - if (reuse && sk2->sk_reuse && - sk2->sk_state != SCTP_SS_LISTENING) + if (sk == sk2 || + (reuse && sk2->sk_reuse && + sk2->sk_state != SCTP_SS_LISTENING)) continue; if (sctp_bind_addr_conflict(&ep2->base.bind_addr, addr, @@ -5702,8 +5704,13 @@ SCTP_STATIC int sctp_seqpacket_listen(struct sock *sk, int backlog) if (!ep->base.bind_addr.port) { if (sctp_autobind(sk)) return -EAGAIN; - } else + } else { + if (sctp_get_port(sk, inet_sk(sk)->num)) { + sk->sk_state = SCTP_SS_CLOSED; + return -EADDRINUSE; + } sctp_sk(sk)->bind_hash->fastreuse = 0; + } sctp_hash_endpoint(ep); return 0; -- cgit v1.2.2 From 336d3262df71fcd2661180bb35d5ea41b4cbca58 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Fri, 18 Jul 2008 23:07:09 -0700 Subject: sctp: remove unnecessary byteshifting, calculate directly in big-endian Signed-off-by: Harvey Harrison Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/ipv4/netfilter/nf_nat_proto_sctp.c | 4 ++-- net/sctp/input.c | 4 ++-- net/sctp/output.c | 4 ++-- 3 files changed, 6 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/ipv4/netfilter/nf_nat_proto_sctp.c b/net/ipv4/netfilter/nf_nat_proto_sctp.c index 82e4c0e286b8..65e470bc6123 100644 --- a/net/ipv4/netfilter/nf_nat_proto_sctp.c +++ b/net/ipv4/netfilter/nf_nat_proto_sctp.c @@ -36,7 +36,7 @@ sctp_manip_pkt(struct sk_buff *skb, sctp_sctphdr_t *hdr; unsigned int hdroff = iphdroff + iph->ihl*4; __be32 oldip, newip; - u32 crc32; + __be32 crc32; if (!skb_make_writable(skb, hdroff + sizeof(*hdr))) return false; @@ -61,7 +61,7 @@ sctp_manip_pkt(struct sk_buff *skb, crc32 = sctp_update_cksum((u8 *)skb->data, skb_headlen(skb), crc32); crc32 = sctp_end_cksum(crc32); - hdr->checksum = htonl(crc32); + hdr->checksum = crc32; return true; } diff --git a/net/sctp/input.c b/net/sctp/input.c index 5ed93c05c23e..a49fa80b57b9 100644 --- a/net/sctp/input.c +++ b/net/sctp/input.c @@ -83,8 +83,8 @@ static inline int sctp_rcv_checksum(struct sk_buff *skb) { struct sk_buff *list = skb_shinfo(skb)->frag_list; struct sctphdr *sh = sctp_hdr(skb); - __u32 cmp = ntohl(sh->checksum); - __u32 val = sctp_start_cksum((__u8 *)sh, skb_headlen(skb)); + __be32 cmp = sh->checksum; + __be32 val = sctp_start_cksum((__u8 *)sh, skb_headlen(skb)); for (; list; list = list->next) val = sctp_update_cksum((__u8 *)list->data, skb_headlen(list), diff --git a/net/sctp/output.c b/net/sctp/output.c index 9a63a3fb9011..45684646b1db 100644 --- a/net/sctp/output.c +++ b/net/sctp/output.c @@ -365,7 +365,7 @@ int sctp_packet_transmit(struct sctp_packet *packet) struct sctp_transport *tp = packet->transport; struct sctp_association *asoc = tp->asoc; struct sctphdr *sh; - __u32 crc32 = 0; + __be32 crc32 = __constant_cpu_to_be32(0); struct sk_buff *nskb; struct sctp_chunk *chunk, *tmp; struct sock *sk; @@ -538,7 +538,7 @@ int sctp_packet_transmit(struct sctp_packet *packet) /* 3) Put the resultant value into the checksum field in the * common header, and leave the rest of the bits unchanged. */ - sh->checksum = htonl(crc32); + sh->checksum = crc32; /* IP layer ECN support * From RFC 2481 -- cgit v1.2.2 From 845525a642c1c9e1335c33a274d4273906ee58eb Mon Sep 17 00:00:00 2001 From: Vlad Yasevich Date: Fri, 18 Jul 2008 23:08:21 -0700 Subject: sctp: Update sctp global memory limit allocations. Update sctp global memory limit allocations to be the same as TCP. Signed-off-by: Vlad Yasevich Signed-off-by: David S. Miller --- net/sctp/protocol.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'net') diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c index cdd881142f2c..ed9acffe8105 100644 --- a/net/sctp/protocol.c +++ b/net/sctp/protocol.c @@ -52,6 +52,8 @@ #include #include #include +#include +#include #include #include #include @@ -1080,6 +1082,7 @@ SCTP_STATIC __init int sctp_init(void) int status = -EINVAL; unsigned long goal; unsigned long limit; + unsigned long nr_pages; int max_share; int order; @@ -1175,8 +1178,9 @@ SCTP_STATIC __init int sctp_init(void) * Note this initalizes the data in sctpv6_prot too * Unabashedly stolen from tcp_init */ - limit = min(num_physpages, 1UL<<(28-PAGE_SHIFT)) >> (20-PAGE_SHIFT); - limit = (limit * (num_physpages >> (20-PAGE_SHIFT))) >> (PAGE_SHIFT-11); + nr_pages = totalram_pages - totalhigh_pages; + limit = min(nr_pages, 1UL<<(28-PAGE_SHIFT)) >> (20-PAGE_SHIFT); + limit = (limit * (nr_pages >> (20-PAGE_SHIFT))) >> (PAGE_SHIFT-11); limit = max(limit, 128UL); sysctl_sctp_mem[0] = limit / 4 * 3; sysctl_sctp_mem[1] = limit; @@ -1186,7 +1190,7 @@ SCTP_STATIC __init int sctp_init(void) limit = (sysctl_sctp_mem[1]) << (PAGE_SHIFT - 7); max_share = min(4UL*1024*1024, limit); - sysctl_sctp_rmem[0] = PAGE_SIZE; /* give each asoc 1 page min */ + sysctl_sctp_rmem[0] = SK_MEM_QUANTUM; /* give each asoc 1 page min */ sysctl_sctp_rmem[1] = (1500 *(sizeof(struct sk_buff) + 1)); sysctl_sctp_rmem[2] = max(sysctl_sctp_rmem[1], max_share); -- cgit v1.2.2 From 49a72dfb8814c2d65bd9f8c9c6daf6395a1ec58d Mon Sep 17 00:00:00 2001 From: Adam Langley Date: Sat, 19 Jul 2008 00:01:42 -0700 Subject: tcp: Fix MD5 signatures for non-linear skbs Currently, the MD5 code assumes that the SKBs are linear and, in the case that they aren't, happily goes off and hashes off the end of the SKB and into random memory. Reported by Stephen Hemminger in [1]. Advice thanks to Stephen and Evgeniy Polyakov. Also includes a couple of missed route_caps from Stephen's patch in [2]. [1] http://marc.info/?l=linux-netdev&m=121445989106145&w=2 [2] http://marc.info/?l=linux-netdev&m=121459157816964&w=2 Signed-off-by: Adam Langley Acked-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/ipv4/tcp.c | 127 ++++++++++++++++++++------------------------- net/ipv4/tcp_ipv4.c | 140 ++++++++++++++++++++++++++++++-------------------- net/ipv4/tcp_output.c | 14 ++--- net/ipv6/tcp_ipv6.c | 127 +++++++++++++++++++++++++++++---------------- 4 files changed, 230 insertions(+), 178 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 827e6132af5f..0b491bf03db4 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -2465,76 +2465,6 @@ static unsigned long tcp_md5sig_users; static struct tcp_md5sig_pool **tcp_md5sig_pool; static DEFINE_SPINLOCK(tcp_md5sig_pool_lock); -int tcp_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, - int bplen, - struct tcphdr *th, unsigned int tcplen, - struct tcp_md5sig_pool *hp) -{ - struct scatterlist sg[4]; - __u16 data_len; - int block = 0; - __sum16 cksum; - struct hash_desc *desc = &hp->md5_desc; - int err; - unsigned int nbytes = 0; - - sg_init_table(sg, 4); - - /* 1. The TCP pseudo-header */ - sg_set_buf(&sg[block++], &hp->md5_blk, bplen); - nbytes += bplen; - - /* 2. The TCP header, excluding options, and assuming a - * checksum of zero - */ - cksum = th->check; - th->check = 0; - sg_set_buf(&sg[block++], th, sizeof(*th)); - nbytes += sizeof(*th); - - /* 3. The TCP segment data (if any) */ - data_len = tcplen - (th->doff << 2); - if (data_len > 0) { - u8 *data = (u8 *)th + (th->doff << 2); - sg_set_buf(&sg[block++], data, data_len); - nbytes += data_len; - } - - /* 4. an independently-specified key or password, known to both - * TCPs and presumably connection-specific - */ - sg_set_buf(&sg[block++], key->key, key->keylen); - nbytes += key->keylen; - - sg_mark_end(&sg[block - 1]); - - /* Now store the hash into the packet */ - err = crypto_hash_init(desc); - if (err) { - if (net_ratelimit()) - printk(KERN_WARNING "%s(): hash_init failed\n", __func__); - return -1; - } - err = crypto_hash_update(desc, sg, nbytes); - if (err) { - if (net_ratelimit()) - printk(KERN_WARNING "%s(): hash_update failed\n", __func__); - return -1; - } - err = crypto_hash_final(desc, md5_hash); - if (err) { - if (net_ratelimit()) - printk(KERN_WARNING "%s(): hash_final failed\n", __func__); - return -1; - } - - /* Reset header */ - th->check = cksum; - - return 0; -} -EXPORT_SYMBOL(tcp_calc_md5_hash); - static void __tcp_free_md5sig_pool(struct tcp_md5sig_pool **pool) { int cpu; @@ -2658,6 +2588,63 @@ void __tcp_put_md5sig_pool(void) } EXPORT_SYMBOL(__tcp_put_md5sig_pool); + +int tcp_md5_hash_header(struct tcp_md5sig_pool *hp, + struct tcphdr *th) +{ + struct scatterlist sg; + int err; + + __sum16 old_checksum = th->check; + th->check = 0; + /* options aren't included in the hash */ + sg_init_one(&sg, th, sizeof(struct tcphdr)); + err = crypto_hash_update(&hp->md5_desc, &sg, sizeof(struct tcphdr)); + th->check = old_checksum; + return err; +} + +EXPORT_SYMBOL(tcp_md5_hash_header); + +int tcp_md5_hash_skb_data(struct tcp_md5sig_pool *hp, + struct sk_buff *skb, unsigned header_len) +{ + struct scatterlist sg; + const struct tcphdr *tp = tcp_hdr(skb); + struct hash_desc *desc = &hp->md5_desc; + unsigned i; + const unsigned head_data_len = skb_headlen(skb) > header_len ? + skb_headlen(skb) - header_len : 0; + const struct skb_shared_info *shi = skb_shinfo(skb); + + sg_init_table(&sg, 1); + + sg_set_buf(&sg, ((u8 *) tp) + header_len, head_data_len); + if (crypto_hash_update(desc, &sg, head_data_len)) + return 1; + + for (i = 0; i < shi->nr_frags; ++i) { + const struct skb_frag_struct *f = &shi->frags[i]; + sg_set_page(&sg, f->page, f->size, f->page_offset); + if (crypto_hash_update(desc, &sg, f->size)) + return 1; + } + + return 0; +} + +EXPORT_SYMBOL(tcp_md5_hash_skb_data); + +int tcp_md5_hash_key(struct tcp_md5sig_pool *hp, struct tcp_md5sig_key *key) +{ + struct scatterlist sg; + + sg_init_one(&sg, key->key, key->keylen); + return crypto_hash_update(&hp->md5_desc, &sg, key->keylen); +} + +EXPORT_SYMBOL(tcp_md5_hash_key); + #endif void tcp_done(struct sock *sk) diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 29adc668ad51..5400d75ff17a 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -87,9 +87,8 @@ int sysctl_tcp_low_latency __read_mostly; #ifdef CONFIG_TCP_MD5SIG static struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr); -static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, - __be32 saddr, __be32 daddr, - struct tcphdr *th, unsigned int tcplen); +static int tcp_v4_md5_hash_hdr(char *md5_hash, struct tcp_md5sig_key *key, + __be32 daddr, __be32 saddr, struct tcphdr *th); #else static inline struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr) @@ -583,11 +582,9 @@ static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb) arg.iov[0].iov_len += TCPOLEN_MD5SIG_ALIGNED; rep.th.doff = arg.iov[0].iov_len / 4; - tcp_v4_do_calc_md5_hash((__u8 *)&rep.opt[1], - key, - ip_hdr(skb)->daddr, - ip_hdr(skb)->saddr, - &rep.th, arg.iov[0].iov_len); + tcp_v4_md5_hash_hdr((__u8 *) &rep.opt[1], + key, ip_hdr(skb)->daddr, + ip_hdr(skb)->saddr, &rep.th); } #endif arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr, @@ -657,11 +654,9 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack, arg.iov[0].iov_len += TCPOLEN_MD5SIG_ALIGNED; rep.th.doff = arg.iov[0].iov_len/4; - tcp_v4_do_calc_md5_hash((__u8 *)&rep.opt[offset], - key, - ip_hdr(skb)->daddr, - ip_hdr(skb)->saddr, - &rep.th, arg.iov[0].iov_len); + tcp_v4_md5_hash_hdr((__u8 *) &rep.opt[offset], + key, ip_hdr(skb)->daddr, + ip_hdr(skb)->saddr, &rep.th); } #endif arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr, @@ -989,28 +984,16 @@ static int tcp_v4_parse_md5_keys(struct sock *sk, char __user *optval, newkey, cmd.tcpm_keylen); } -static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, - __be32 saddr, __be32 daddr, - struct tcphdr *th, - unsigned int tcplen) +static int tcp_v4_md5_hash_pseudoheader(struct tcp_md5sig_pool *hp, + __be32 daddr, __be32 saddr, int nbytes) { - struct tcp_md5sig_pool *hp; struct tcp4_pseudohdr *bp; - int err; - - /* - * Okay, so RFC2385 is turned on for this connection, - * so we need to generate the MD5 hash for the packet now. - */ - - hp = tcp_get_md5sig_pool(); - if (!hp) - goto clear_hash_noput; + struct scatterlist sg; bp = &hp->md5_blk.ip4; /* - * The TCP pseudo-header (in the order: source IP address, + * 1. the TCP pseudo-header (in the order: source IP address, * destination IP address, zero-padded protocol number, and * segment length) */ @@ -1018,48 +1001,95 @@ static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, bp->daddr = daddr; bp->pad = 0; bp->protocol = IPPROTO_TCP; - bp->len = htons(tcplen); + bp->len = cpu_to_be16(nbytes); - err = tcp_calc_md5_hash(md5_hash, key, sizeof(*bp), - th, tcplen, hp); - if (err) + sg_init_one(&sg, bp, sizeof(*bp)); + return crypto_hash_update(&hp->md5_desc, &sg, sizeof(*bp)); +} + +static int tcp_v4_md5_hash_hdr(char *md5_hash, struct tcp_md5sig_key *key, + __be32 daddr, __be32 saddr, struct tcphdr *th) +{ + struct tcp_md5sig_pool *hp; + struct hash_desc *desc; + + hp = tcp_get_md5sig_pool(); + if (!hp) + goto clear_hash_noput; + desc = &hp->md5_desc; + + if (crypto_hash_init(desc)) + goto clear_hash; + if (tcp_v4_md5_hash_pseudoheader(hp, daddr, saddr, th->doff << 2)) + goto clear_hash; + if (tcp_md5_hash_header(hp, th)) + goto clear_hash; + if (tcp_md5_hash_key(hp, key)) + goto clear_hash; + if (crypto_hash_final(desc, md5_hash)) goto clear_hash; - /* Free up the crypto pool */ tcp_put_md5sig_pool(); -out: return 0; + clear_hash: tcp_put_md5sig_pool(); clear_hash_noput: memset(md5_hash, 0, 16); - goto out; + return 1; } -int tcp_v4_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, - struct sock *sk, - struct dst_entry *dst, - struct request_sock *req, - struct tcphdr *th, - unsigned int tcplen) +int tcp_v4_md5_hash_skb(char *md5_hash, struct tcp_md5sig_key *key, + struct sock *sk, struct request_sock *req, + struct sk_buff *skb) { + struct tcp_md5sig_pool *hp; + struct hash_desc *desc; + struct tcphdr *th = tcp_hdr(skb); __be32 saddr, daddr; if (sk) { saddr = inet_sk(sk)->saddr; daddr = inet_sk(sk)->daddr; + } else if (req) { + saddr = inet_rsk(req)->loc_addr; + daddr = inet_rsk(req)->rmt_addr; } else { - struct rtable *rt = (struct rtable *)dst; - BUG_ON(!rt); - saddr = rt->rt_src; - daddr = rt->rt_dst; + const struct iphdr *iph = ip_hdr(skb); + saddr = iph->saddr; + daddr = iph->daddr; } - return tcp_v4_do_calc_md5_hash(md5_hash, key, - saddr, daddr, - th, tcplen); + + hp = tcp_get_md5sig_pool(); + if (!hp) + goto clear_hash_noput; + desc = &hp->md5_desc; + + if (crypto_hash_init(desc)) + goto clear_hash; + + if (tcp_v4_md5_hash_pseudoheader(hp, daddr, saddr, skb->len)) + goto clear_hash; + if (tcp_md5_hash_header(hp, th)) + goto clear_hash; + if (tcp_md5_hash_skb_data(hp, skb, th->doff << 2)) + goto clear_hash; + if (tcp_md5_hash_key(hp, key)) + goto clear_hash; + if (crypto_hash_final(desc, md5_hash)) + goto clear_hash; + + tcp_put_md5sig_pool(); + return 0; + +clear_hash: + tcp_put_md5sig_pool(); +clear_hash_noput: + memset(md5_hash, 0, 16); + return 1; } -EXPORT_SYMBOL(tcp_v4_calc_md5_hash); +EXPORT_SYMBOL(tcp_v4_md5_hash_skb); static int tcp_v4_inbound_md5_hash(struct sock *sk, struct sk_buff *skb) { @@ -1104,10 +1134,9 @@ static int tcp_v4_inbound_md5_hash(struct sock *sk, struct sk_buff *skb) /* Okay, so this is hash_expected and hash_location - * so we need to calculate the checksum. */ - genhash = tcp_v4_do_calc_md5_hash(newhash, - hash_expected, - iph->saddr, iph->daddr, - th, skb->len); + genhash = tcp_v4_md5_hash_skb(newhash, + hash_expected, + NULL, NULL, skb); if (genhash || memcmp(hash_location, newhash, 16) != 0) { if (net_ratelimit()) { @@ -1356,6 +1385,7 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb, if (newkey != NULL) tcp_v4_md5_do_add(newsk, inet_sk(sk)->daddr, newkey, key->keylen); + newsk->sk_route_caps &= ~NETIF_F_GSO_MASK; } #endif @@ -1719,7 +1749,7 @@ struct inet_connection_sock_af_ops ipv4_specific = { #ifdef CONFIG_TCP_MD5SIG static struct tcp_sock_af_ops tcp_sock_ipv4_specific = { .md5_lookup = tcp_v4_md5_lookup, - .calc_md5_hash = tcp_v4_calc_md5_hash, + .calc_md5_hash = tcp_v4_md5_hash_skb, .md5_add = tcp_v4_md5_add_func, .md5_parse = tcp_v4_parse_md5_keys, }; diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 36a19707f67f..958ff486165f 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -540,8 +540,10 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, * room for it. */ md5 = tp->af_specific->md5_lookup(sk, sk); - if (md5) + if (md5) { tcp_header_size += TCPOLEN_MD5SIG_ALIGNED; + sk->sk_route_caps &= ~NETIF_F_GSO_MASK; + } #endif skb_push(skb, tcp_header_size); @@ -602,10 +604,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, /* Calculate the MD5 hash, as we have all we need now */ if (md5) { tp->af_specific->calc_md5_hash(md5_hash_location, - md5, - sk, NULL, NULL, - tcp_hdr(skb), - skb->len); + md5, sk, NULL, skb); } #endif @@ -2264,10 +2263,7 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst, /* Okay, we have all we need - do the md5 hash if needed */ if (md5) { tp->af_specific->calc_md5_hash(md5_hash_location, - md5, - NULL, dst, req, - tcp_hdr(skb), - skb->len); + md5, NULL, req, skb); } #endif diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index ca5b93a5c02a..ae45f9835014 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -736,64 +736,105 @@ static int tcp_v6_parse_md5_keys (struct sock *sk, char __user *optval, return tcp_v6_md5_do_add(sk, &sin6->sin6_addr, newkey, cmd.tcpm_keylen); } -static int tcp_v6_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, - struct in6_addr *saddr, - struct in6_addr *daddr, - struct tcphdr *th, unsigned int tcplen) +static int tcp_v6_md5_hash_pseudoheader(struct tcp_md5sig_pool *hp, + struct in6_addr *daddr, + struct in6_addr *saddr, int nbytes) { - struct tcp_md5sig_pool *hp; struct tcp6_pseudohdr *bp; - int err; - - hp = tcp_get_md5sig_pool(); - if (!hp) { - printk(KERN_WARNING "%s(): hash pool not found...\n", __func__); - goto clear_hash_noput; - } + struct scatterlist sg; bp = &hp->md5_blk.ip6; - /* 1. TCP pseudo-header (RFC2460) */ ipv6_addr_copy(&bp->saddr, saddr); ipv6_addr_copy(&bp->daddr, daddr); - bp->len = htonl(tcplen); - bp->protocol = htonl(IPPROTO_TCP); + bp->protocol = cpu_to_be32(IPPROTO_TCP); + bp->len = cpu_to_be16(nbytes); - err = tcp_calc_md5_hash(md5_hash, key, sizeof(*bp), - th, tcplen, hp); + sg_init_one(&sg, bp, sizeof(*bp)); + return crypto_hash_update(&hp->md5_desc, &sg, sizeof(*bp)); +} - if (err) +static int tcp_v6_md5_hash_hdr(char *md5_hash, struct tcp_md5sig_key *key, + struct in6_addr *daddr, struct in6_addr *saddr, + struct tcphdr *th) +{ + struct tcp_md5sig_pool *hp; + struct hash_desc *desc; + + hp = tcp_get_md5sig_pool(); + if (!hp) + goto clear_hash_noput; + desc = &hp->md5_desc; + + if (crypto_hash_init(desc)) + goto clear_hash; + if (tcp_v6_md5_hash_pseudoheader(hp, daddr, saddr, th->doff << 2)) + goto clear_hash; + if (tcp_md5_hash_header(hp, th)) + goto clear_hash; + if (tcp_md5_hash_key(hp, key)) + goto clear_hash; + if (crypto_hash_final(desc, md5_hash)) goto clear_hash; - /* Free up the crypto pool */ tcp_put_md5sig_pool(); -out: return 0; + clear_hash: tcp_put_md5sig_pool(); clear_hash_noput: memset(md5_hash, 0, 16); - goto out; + return 1; } -static int tcp_v6_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, - struct sock *sk, - struct dst_entry *dst, - struct request_sock *req, - struct tcphdr *th, unsigned int tcplen) +static int tcp_v6_md5_hash_skb(char *md5_hash, struct tcp_md5sig_key *key, + struct sock *sk, struct request_sock *req, + struct sk_buff *skb) { struct in6_addr *saddr, *daddr; + struct tcp_md5sig_pool *hp; + struct hash_desc *desc; + struct tcphdr *th = tcp_hdr(skb); if (sk) { saddr = &inet6_sk(sk)->saddr; daddr = &inet6_sk(sk)->daddr; - } else { + } else if (req) { saddr = &inet6_rsk(req)->loc_addr; daddr = &inet6_rsk(req)->rmt_addr; + } else { + struct ipv6hdr *ip6h = ipv6_hdr(skb); + saddr = &ip6h->saddr; + daddr = &ip6h->daddr; } - return tcp_v6_do_calc_md5_hash(md5_hash, key, - saddr, daddr, - th, tcplen); + + hp = tcp_get_md5sig_pool(); + if (!hp) + goto clear_hash_noput; + desc = &hp->md5_desc; + + if (crypto_hash_init(desc)) + goto clear_hash; + + if (tcp_v6_md5_hash_pseudoheader(hp, daddr, saddr, skb->len)) + goto clear_hash; + if (tcp_md5_hash_header(hp, th)) + goto clear_hash; + if (tcp_md5_hash_skb_data(hp, skb, th->doff << 2)) + goto clear_hash; + if (tcp_md5_hash_key(hp, key)) + goto clear_hash; + if (crypto_hash_final(desc, md5_hash)) + goto clear_hash; + + tcp_put_md5sig_pool(); + return 0; + +clear_hash: + tcp_put_md5sig_pool(); +clear_hash_noput: + memset(md5_hash, 0, 16); + return 1; } static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb) @@ -834,10 +875,10 @@ static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb) } /* check the signature */ - genhash = tcp_v6_do_calc_md5_hash(newhash, - hash_expected, - &ip6h->saddr, &ip6h->daddr, - th, skb->len); + genhash = tcp_v6_md5_hash_skb(newhash, + hash_expected, + NULL, NULL, skb); + if (genhash || memcmp(hash_location, newhash, 16) != 0) { if (net_ratelimit()) { printk(KERN_INFO "MD5 Hash %s for " @@ -974,10 +1015,9 @@ static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb) (TCPOPT_NOP << 16) | (TCPOPT_MD5SIG << 8) | TCPOLEN_MD5SIG); - tcp_v6_do_calc_md5_hash((__u8 *)&opt[1], key, - &ipv6_hdr(skb)->daddr, - &ipv6_hdr(skb)->saddr, - t1, tot_len); + tcp_v6_md5_hash_hdr((__u8 *)&opt[1], key, + &ipv6_hdr(skb)->daddr, + &ipv6_hdr(skb)->saddr, t1); } #endif @@ -1064,10 +1104,9 @@ static void tcp_v6_send_ack(struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 if (key) { *topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) | (TCPOPT_MD5SIG << 8) | TCPOLEN_MD5SIG); - tcp_v6_do_calc_md5_hash((__u8 *)topt, key, - &ipv6_hdr(skb)->daddr, - &ipv6_hdr(skb)->saddr, - t1, tot_len); + tcp_v6_md5_hash_hdr((__u8 *)topt, key, + &ipv6_hdr(skb)->daddr, + &ipv6_hdr(skb)->saddr, t1); } #endif @@ -1783,7 +1822,7 @@ static struct inet_connection_sock_af_ops ipv6_specific = { #ifdef CONFIG_TCP_MD5SIG static struct tcp_sock_af_ops tcp_sock_ipv6_specific = { .md5_lookup = tcp_v6_md5_lookup, - .calc_md5_hash = tcp_v6_calc_md5_hash, + .calc_md5_hash = tcp_v6_md5_hash_skb, .md5_add = tcp_v6_md5_add_func, .md5_parse = tcp_v6_parse_md5_keys, }; @@ -1815,7 +1854,7 @@ static struct inet_connection_sock_af_ops ipv6_mapped = { #ifdef CONFIG_TCP_MD5SIG static struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific = { .md5_lookup = tcp_v4_md5_lookup, - .calc_md5_hash = tcp_v4_calc_md5_hash, + .calc_md5_hash = tcp_v4_md5_hash_skb, .md5_add = tcp_v6_md5_add_func, .md5_parse = tcp_v6_parse_md5_keys, }; -- cgit v1.2.2 From 33ad798c924b4a1afad3593f2796d465040aadd5 Mon Sep 17 00:00:00 2001 From: Adam Langley Date: Sat, 19 Jul 2008 00:04:31 -0700 Subject: tcp: options clean up This should fix the following bugs: * Connections with MD5 signatures produce invalid packets whenever SACK options are included * MD5 signatures are counted twice in the MSS calculations Behaviour changes: * A SYN with MD5 + SACK + TS elicits a SYNACK with MD5 + SACK This is because we can't fit any SACK blocks in a packet with MD5 + TS options. There was discussion about disabling SACK rather than TS in order to fit in better with old, buggy kernels, but that was deemed to be unnecessary. * SYNs with MD5 don't include a TS option See above. Additionally, it removes a bunch of duplicated logic for calculating options, which should help avoid these sort of issues in the future. Signed-off-by: Adam Langley Signed-off-by: David S. Miller --- net/ipv4/tcp_output.c | 432 +++++++++++++++++++++++++++----------------------- 1 file changed, 236 insertions(+), 196 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 958ff486165f..1fa683c0ba9b 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -345,28 +345,82 @@ static void tcp_init_nondata_skb(struct sk_buff *skb, u32 seq, u8 flags) TCP_SKB_CB(skb)->end_seq = seq; } -static void tcp_build_and_update_options(__be32 *ptr, struct tcp_sock *tp, - __u32 tstamp, __u8 **md5_hash) -{ - if (tp->rx_opt.tstamp_ok) { +#define OPTION_SACK_ADVERTISE (1 << 0) +#define OPTION_TS (1 << 1) +#define OPTION_MD5 (1 << 2) + +struct tcp_out_options { + u8 options; /* bit field of OPTION_* */ + u8 ws; /* window scale, 0 to disable */ + u8 num_sack_blocks; /* number of SACK blocks to include */ + u16 mss; /* 0 to disable */ + __u32 tsval, tsecr; /* need to include OPTION_TS */ +}; + +static void tcp_options_write(__be32 *ptr, struct tcp_sock *tp, + const struct tcp_out_options *opts, + __u8 **md5_hash) { + if (unlikely(OPTION_MD5 & opts->options)) { *ptr++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) | - (TCPOPT_TIMESTAMP << 8) | - TCPOLEN_TIMESTAMP); - *ptr++ = htonl(tstamp); - *ptr++ = htonl(tp->rx_opt.ts_recent); + (TCPOPT_MD5SIG << 8) | + TCPOLEN_MD5SIG); + *md5_hash = (__u8 *)ptr; + ptr += 4; + } else { + *md5_hash = NULL; } - if (tp->rx_opt.eff_sacks) { - struct tcp_sack_block *sp = tp->rx_opt.dsack ? tp->duplicate_sack : tp->selective_acks; + + if (likely(OPTION_TS & opts->options)) { + if (unlikely(OPTION_SACK_ADVERTISE & opts->options)) { + *ptr++ = htonl((TCPOPT_SACK_PERM << 24) | + (TCPOLEN_SACK_PERM << 16) | + (TCPOPT_TIMESTAMP << 8) | + TCPOLEN_TIMESTAMP); + } else { + *ptr++ = htonl((TCPOPT_NOP << 24) | + (TCPOPT_NOP << 16) | + (TCPOPT_TIMESTAMP << 8) | + TCPOLEN_TIMESTAMP); + } + *ptr++ = htonl(opts->tsval); + *ptr++ = htonl(opts->tsecr); + } + + if (unlikely(opts->mss)) { + *ptr++ = htonl((TCPOPT_MSS << 24) | + (TCPOLEN_MSS << 16) | + opts->mss); + } + + if (unlikely(OPTION_SACK_ADVERTISE & opts->options && + !(OPTION_TS & opts->options))) { + *ptr++ = htonl((TCPOPT_NOP << 24) | + (TCPOPT_NOP << 16) | + (TCPOPT_SACK_PERM << 8) | + TCPOLEN_SACK_PERM); + } + + if (unlikely(opts->ws)) { + *ptr++ = htonl((TCPOPT_NOP << 24) | + (TCPOPT_WINDOW << 16) | + (TCPOLEN_WINDOW << 8) | + opts->ws); + } + + if (unlikely(opts->num_sack_blocks)) { + struct tcp_sack_block *sp = tp->rx_opt.dsack ? + tp->duplicate_sack : tp->selective_acks; int this_sack; *ptr++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) | (TCPOPT_SACK << 8) | - (TCPOLEN_SACK_BASE + (tp->rx_opt.eff_sacks * + (TCPOLEN_SACK_BASE + (opts->num_sack_blocks * TCPOLEN_SACK_PERBLOCK))); - for (this_sack = 0; this_sack < tp->rx_opt.eff_sacks; this_sack++) { + for (this_sack = 0; this_sack < opts->num_sack_blocks; + ++this_sack) { *ptr++ = htonl(sp[this_sack].start_seq); *ptr++ = htonl(sp[this_sack].end_seq); } @@ -376,81 +430,137 @@ static void tcp_build_and_update_options(__be32 *ptr, struct tcp_sock *tp, tp->rx_opt.eff_sacks--; } } +} + +static unsigned tcp_syn_options(struct sock *sk, struct sk_buff *skb, + struct tcp_out_options *opts, + struct tcp_md5sig_key **md5) { + struct tcp_sock *tp = tcp_sk(sk); + unsigned size = 0; + #ifdef CONFIG_TCP_MD5SIG - if (md5_hash) { - *ptr++ = htonl((TCPOPT_NOP << 24) | - (TCPOPT_NOP << 16) | - (TCPOPT_MD5SIG << 8) | - TCPOLEN_MD5SIG); - *md5_hash = (__u8 *)ptr; + *md5 = tp->af_specific->md5_lookup(sk, sk); + if (*md5) { + opts->options |= OPTION_MD5; + size += TCPOLEN_MD5SIG_ALIGNED; } +#else + *md5 = NULL; #endif + + /* We always get an MSS option. The option bytes which will be seen in + * normal data packets should timestamps be used, must be in the MSS + * advertised. But we subtract them from tp->mss_cache so that + * calculations in tcp_sendmsg are simpler etc. So account for this + * fact here if necessary. If we don't do this correctly, as a + * receiver we won't recognize data packets as being full sized when we + * should, and thus we won't abide by the delayed ACK rules correctly. + * SACKs don't matter, we never delay an ACK when we have any of those + * going out. */ + opts->mss = tcp_advertise_mss(sk); + size += TCPOLEN_MSS_ALIGNED; + + if (likely(sysctl_tcp_timestamps && *md5 == NULL)) { + opts->options |= OPTION_TS; + opts->tsval = TCP_SKB_CB(skb)->when; + opts->tsecr = tp->rx_opt.ts_recent; + size += TCPOLEN_TSTAMP_ALIGNED; + } + if (likely(sysctl_tcp_window_scaling)) { + opts->ws = tp->rx_opt.rcv_wscale; + size += TCPOLEN_WSCALE_ALIGNED; + } + if (likely(sysctl_tcp_sack)) { + opts->options |= OPTION_SACK_ADVERTISE; + if (unlikely(!OPTION_TS & opts->options)) + size += TCPOLEN_SACKPERM_ALIGNED; + } + + return size; } -/* Construct a tcp options header for a SYN or SYN_ACK packet. - * If this is every changed make sure to change the definition of - * MAX_SYN_SIZE to match the new maximum number of options that you - * can generate. - * - * Note - that with the RFC2385 TCP option, we make room for the - * 16 byte MD5 hash. This will be filled in later, so the pointer for the - * location to be filled is passed back up. - */ -static void tcp_syn_build_options(__be32 *ptr, int mss, int ts, int sack, - int offer_wscale, int wscale, __u32 tstamp, - __u32 ts_recent, __u8 **md5_hash) -{ - /* We always get an MSS option. - * The option bytes which will be seen in normal data - * packets should timestamps be used, must be in the MSS - * advertised. But we subtract them from tp->mss_cache so - * that calculations in tcp_sendmsg are simpler etc. - * So account for this fact here if necessary. If we - * don't do this correctly, as a receiver we won't - * recognize data packets as being full sized when we - * should, and thus we won't abide by the delayed ACK - * rules correctly. - * SACKs don't matter, we never delay an ACK when we - * have any of those going out. - */ - *ptr++ = htonl((TCPOPT_MSS << 24) | (TCPOLEN_MSS << 16) | mss); - if (ts) { - if (sack) - *ptr++ = htonl((TCPOPT_SACK_PERM << 24) | - (TCPOLEN_SACK_PERM << 16) | - (TCPOPT_TIMESTAMP << 8) | - TCPOLEN_TIMESTAMP); - else - *ptr++ = htonl((TCPOPT_NOP << 24) | - (TCPOPT_NOP << 16) | - (TCPOPT_TIMESTAMP << 8) | - TCPOLEN_TIMESTAMP); - *ptr++ = htonl(tstamp); /* TSVAL */ - *ptr++ = htonl(ts_recent); /* TSECR */ - } else if (sack) - *ptr++ = htonl((TCPOPT_NOP << 24) | - (TCPOPT_NOP << 16) | - (TCPOPT_SACK_PERM << 8) | - TCPOLEN_SACK_PERM); - if (offer_wscale) - *ptr++ = htonl((TCPOPT_NOP << 24) | - (TCPOPT_WINDOW << 16) | - (TCPOLEN_WINDOW << 8) | - (wscale)); +static unsigned tcp_synack_options(struct sock *sk, + struct request_sock *req, + unsigned mss, struct sk_buff *skb, + struct tcp_out_options *opts, + struct tcp_md5sig_key **md5) { + unsigned size = 0; + struct inet_request_sock *ireq = inet_rsk(req); + char doing_ts; + #ifdef CONFIG_TCP_MD5SIG - /* - * If MD5 is enabled, then we set the option, and include the size - * (always 18). The actual MD5 hash is added just before the - * packet is sent. - */ - if (md5_hash) { - *ptr++ = htonl((TCPOPT_NOP << 24) | - (TCPOPT_NOP << 16) | - (TCPOPT_MD5SIG << 8) | - TCPOLEN_MD5SIG); - *md5_hash = (__u8 *)ptr; + *md5 = tcp_rsk(req)->af_specific->md5_lookup(sk, req); + if (*md5) { + opts->options |= OPTION_MD5; + size += TCPOLEN_MD5SIG_ALIGNED; } +#else + *md5 = NULL; #endif + + /* we can't fit any SACK blocks in a packet with MD5 + TS + options. There was discussion about disabling SACK rather than TS in + order to fit in better with old, buggy kernels, but that was deemed + to be unnecessary. */ + doing_ts = ireq->tstamp_ok && !(*md5 && ireq->sack_ok); + + opts->mss = mss; + size += TCPOLEN_MSS_ALIGNED; + + if (likely(ireq->wscale_ok)) { + opts->ws = ireq->rcv_wscale; + size += TCPOLEN_WSCALE_ALIGNED; + } + if (likely(doing_ts)) { + opts->options |= OPTION_TS; + opts->tsval = TCP_SKB_CB(skb)->when; + opts->tsecr = req->ts_recent; + size += TCPOLEN_TSTAMP_ALIGNED; + } + if (likely(ireq->sack_ok)) { + opts->options |= OPTION_SACK_ADVERTISE; + if (unlikely(!doing_ts)) + size += TCPOLEN_SACKPERM_ALIGNED; + } + + return size; +} + +static unsigned tcp_established_options(struct sock *sk, struct sk_buff *skb, + struct tcp_out_options *opts, + struct tcp_md5sig_key **md5) { + struct tcp_skb_cb *tcb = skb ? TCP_SKB_CB(skb) : NULL; + struct tcp_sock *tp = tcp_sk(sk); + unsigned size = 0; + +#ifdef CONFIG_TCP_MD5SIG + *md5 = tp->af_specific->md5_lookup(sk, sk); + if (unlikely(*md5)) { + opts->options |= OPTION_MD5; + size += TCPOLEN_MD5SIG_ALIGNED; + } +#else + *md5 = NULL; +#endif + + if (likely(tp->rx_opt.tstamp_ok)) { + opts->options |= OPTION_TS; + opts->tsval = tcb ? tcb->when : 0; + opts->tsecr = tp->rx_opt.ts_recent; + size += TCPOLEN_TSTAMP_ALIGNED; + } + + if (unlikely(tp->rx_opt.eff_sacks)) { + const unsigned remaining = MAX_TCP_OPTION_SPACE - size; + opts->num_sack_blocks = + min_t(unsigned, tp->rx_opt.eff_sacks, + (remaining - TCPOLEN_SACK_BASE_ALIGNED) / + TCPOLEN_SACK_PERBLOCK); + size += TCPOLEN_SACK_BASE_ALIGNED + + opts->num_sack_blocks * TCPOLEN_SACK_PERBLOCK; + } + + return size; } /* This routine actually transmits TCP packets queued in by @@ -471,13 +581,11 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, struct inet_sock *inet; struct tcp_sock *tp; struct tcp_skb_cb *tcb; - int tcp_header_size; -#ifdef CONFIG_TCP_MD5SIG + struct tcp_out_options opts; + unsigned tcp_options_size, tcp_header_size; struct tcp_md5sig_key *md5; __u8 *md5_hash_location; -#endif struct tcphdr *th; - int sysctl_flags; int err; BUG_ON(!skb || !tcp_skb_pcount(skb)); @@ -500,52 +608,18 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, inet = inet_sk(sk); tp = tcp_sk(sk); tcb = TCP_SKB_CB(skb); - tcp_header_size = tp->tcp_header_len; - -#define SYSCTL_FLAG_TSTAMPS 0x1 -#define SYSCTL_FLAG_WSCALE 0x2 -#define SYSCTL_FLAG_SACK 0x4 + memset(&opts, 0, sizeof(opts)); - sysctl_flags = 0; - if (unlikely(tcb->flags & TCPCB_FLAG_SYN)) { - tcp_header_size = sizeof(struct tcphdr) + TCPOLEN_MSS; - if (sysctl_tcp_timestamps) { - tcp_header_size += TCPOLEN_TSTAMP_ALIGNED; - sysctl_flags |= SYSCTL_FLAG_TSTAMPS; - } - if (sysctl_tcp_window_scaling) { - tcp_header_size += TCPOLEN_WSCALE_ALIGNED; - sysctl_flags |= SYSCTL_FLAG_WSCALE; - } - if (sysctl_tcp_sack) { - sysctl_flags |= SYSCTL_FLAG_SACK; - if (!(sysctl_flags & SYSCTL_FLAG_TSTAMPS)) - tcp_header_size += TCPOLEN_SACKPERM_ALIGNED; - } - } else if (unlikely(tp->rx_opt.eff_sacks)) { - /* A SACK is 2 pad bytes, a 2 byte header, plus - * 2 32-bit sequence numbers for each SACK block. - */ - tcp_header_size += (TCPOLEN_SACK_BASE_ALIGNED + - (tp->rx_opt.eff_sacks * - TCPOLEN_SACK_PERBLOCK)); - } + if (unlikely(tcb->flags & TCPCB_FLAG_SYN)) + tcp_options_size = tcp_syn_options(sk, skb, &opts, &md5); + else + tcp_options_size = tcp_established_options(sk, skb, &opts, + &md5); + tcp_header_size = tcp_options_size + sizeof(struct tcphdr); if (tcp_packets_in_flight(tp) == 0) tcp_ca_event(sk, CA_EVENT_TX_START); -#ifdef CONFIG_TCP_MD5SIG - /* - * Are we doing MD5 on this segment? If so - make - * room for it. - */ - md5 = tp->af_specific->md5_lookup(sk, sk); - if (md5) { - tcp_header_size += TCPOLEN_MD5SIG_ALIGNED; - sk->sk_route_caps &= ~NETIF_F_GSO_MASK; - } -#endif - skb_push(skb, tcp_header_size); skb_reset_transport_header(skb); skb_set_owner_w(skb, sk); @@ -576,33 +650,14 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, th->urg = 1; } - if (unlikely(tcb->flags & TCPCB_FLAG_SYN)) { - tcp_syn_build_options((__be32 *)(th + 1), - tcp_advertise_mss(sk), - (sysctl_flags & SYSCTL_FLAG_TSTAMPS), - (sysctl_flags & SYSCTL_FLAG_SACK), - (sysctl_flags & SYSCTL_FLAG_WSCALE), - tp->rx_opt.rcv_wscale, - tcb->when, - tp->rx_opt.ts_recent, - -#ifdef CONFIG_TCP_MD5SIG - md5 ? &md5_hash_location : -#endif - NULL); - } else { - tcp_build_and_update_options((__be32 *)(th + 1), - tp, tcb->when, -#ifdef CONFIG_TCP_MD5SIG - md5 ? &md5_hash_location : -#endif - NULL); + tcp_options_write((__be32 *)(th + 1), tp, &opts, &md5_hash_location); + if (likely((tcb->flags & TCPCB_FLAG_SYN) == 0)) TCP_ECN_send(sk, skb, tcp_header_size); - } #ifdef CONFIG_TCP_MD5SIG /* Calculate the MD5 hash, as we have all we need now */ if (md5) { + sk->sk_route_caps &= ~NETIF_F_GSO_MASK; tp->af_specific->calc_md5_hash(md5_hash_location, md5, sk, NULL, skb); } @@ -626,10 +681,6 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, tcp_enter_cwr(sk, 1); return net_xmit_eval(err); - -#undef SYSCTL_FLAG_TSTAMPS -#undef SYSCTL_FLAG_WSCALE -#undef SYSCTL_FLAG_SACK } /* This routine just queue's the buffer @@ -970,6 +1021,9 @@ unsigned int tcp_current_mss(struct sock *sk, int large_allowed) u32 mss_now; u16 xmit_size_goal; int doing_tso = 0; + unsigned header_len; + struct tcp_out_options opts; + struct tcp_md5sig_key *md5; mss_now = tp->mss_cache; @@ -982,14 +1036,16 @@ unsigned int tcp_current_mss(struct sock *sk, int large_allowed) mss_now = tcp_sync_mss(sk, mtu); } - if (tp->rx_opt.eff_sacks) - mss_now -= (TCPOLEN_SACK_BASE_ALIGNED + - (tp->rx_opt.eff_sacks * TCPOLEN_SACK_PERBLOCK)); - -#ifdef CONFIG_TCP_MD5SIG - if (tp->af_specific->md5_lookup(sk, sk)) - mss_now -= TCPOLEN_MD5SIG_ALIGNED; -#endif + header_len = tcp_established_options(sk, NULL, &opts, &md5) + + sizeof(struct tcphdr); + /* The mss_cache is sized based on tp->tcp_header_len, which assumes + * some common options. If this is an odd packet (because we have SACK + * blocks etc) then our calculated header_len will be different, and + * we have to adjust mss_now correspondingly */ + if (header_len != tp->tcp_header_len) { + int delta = (int) header_len - tp->tcp_header_len; + mss_now -= delta; + } xmit_size_goal = mss_now; @@ -2179,11 +2235,10 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst, struct tcp_sock *tp = tcp_sk(sk); struct tcphdr *th; int tcp_header_size; + struct tcp_out_options opts; struct sk_buff *skb; -#ifdef CONFIG_TCP_MD5SIG struct tcp_md5sig_key *md5; __u8 *md5_hash_location; -#endif skb = sock_wmalloc(sk, MAX_TCP_HEADER + 15, 1, GFP_ATOMIC); if (skb == NULL) @@ -2194,18 +2249,27 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst, skb->dst = dst_clone(dst); - tcp_header_size = (sizeof(struct tcphdr) + TCPOLEN_MSS + - (ireq->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0) + - (ireq->wscale_ok ? TCPOLEN_WSCALE_ALIGNED : 0) + - /* SACK_PERM is in the place of NOP NOP of TS */ - ((ireq->sack_ok && !ireq->tstamp_ok) ? TCPOLEN_SACKPERM_ALIGNED : 0)); + if (req->rcv_wnd == 0) { /* ignored for retransmitted syns */ + __u8 rcv_wscale; + /* Set this up on the first call only */ + req->window_clamp = tp->window_clamp ? : dst_metric(dst, RTAX_WINDOW); + /* tcp_full_space because it is guaranteed to be the first packet */ + tcp_select_initial_window(tcp_full_space(sk), + dst_metric(dst, RTAX_ADVMSS) - (ireq->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0), + &req->rcv_wnd, + &req->window_clamp, + ireq->wscale_ok, + &rcv_wscale); + ireq->rcv_wscale = rcv_wscale; + } + + memset(&opts, 0, sizeof(opts)); + TCP_SKB_CB(skb)->when = tcp_time_stamp; + tcp_header_size = tcp_synack_options(sk, req, + dst_metric(dst, RTAX_ADVMSS), + skb, &opts, &md5) + + sizeof(struct tcphdr); -#ifdef CONFIG_TCP_MD5SIG - /* Are we doing MD5 on this segment? If so - make room for it */ - md5 = tcp_rsk(req)->af_specific->md5_lookup(sk, req); - if (md5) - tcp_header_size += TCPOLEN_MD5SIG_ALIGNED; -#endif skb_push(skb, tcp_header_size); skb_reset_transport_header(skb); @@ -2223,19 +2287,6 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst, TCPCB_FLAG_SYN | TCPCB_FLAG_ACK); th->seq = htonl(TCP_SKB_CB(skb)->seq); th->ack_seq = htonl(tcp_rsk(req)->rcv_isn + 1); - if (req->rcv_wnd == 0) { /* ignored for retransmitted syns */ - __u8 rcv_wscale; - /* Set this up on the first call only */ - req->window_clamp = tp->window_clamp ? : dst_metric(dst, RTAX_WINDOW); - /* tcp_full_space because it is guaranteed to be the first packet */ - tcp_select_initial_window(tcp_full_space(sk), - dst_metric(dst, RTAX_ADVMSS) - (ireq->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0), - &req->rcv_wnd, - &req->window_clamp, - ireq->wscale_ok, - &rcv_wscale); - ireq->rcv_wscale = rcv_wscale; - } /* RFC1323: The window in SYN & SYN/ACK segments is never scaled. */ th->window = htons(min(req->rcv_wnd, 65535U)); @@ -2244,18 +2295,7 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst, TCP_SKB_CB(skb)->when = cookie_init_timestamp(req); else #endif - TCP_SKB_CB(skb)->when = tcp_time_stamp; - tcp_syn_build_options((__be32 *)(th + 1), dst_metric(dst, RTAX_ADVMSS), ireq->tstamp_ok, - ireq->sack_ok, ireq->wscale_ok, ireq->rcv_wscale, - TCP_SKB_CB(skb)->when, - req->ts_recent, - ( -#ifdef CONFIG_TCP_MD5SIG - md5 ? &md5_hash_location : -#endif - NULL) - ); - + tcp_options_write((__be32 *)(th + 1), tp, &opts, &md5_hash_location); th->doff = (tcp_header_size >> 2); TCP_INC_STATS(sock_net(sk), TCP_MIB_OUTSEGS); -- cgit v1.2.2 From 4389dded7767d24290463f2a8302ba3253ebdd56 Mon Sep 17 00:00:00 2001 From: Adam Langley Date: Sat, 19 Jul 2008 00:07:02 -0700 Subject: tcp: Remove redundant checks when setting eff_sacks Remove redundant checks when setting eff_sacks and make the number of SACKs a compile time constant. Now that the options code knows how many SACK blocks can fit in the header, we don't need to have the SACK code guessing at it. Signed-off-by: Adam Langley Signed-off-by: David S. Miller --- net/ipv4/tcp_input.c | 25 ++++++++++--------------- 1 file changed, 10 insertions(+), 15 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 88810bc01370..1f5e6049883e 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -1423,10 +1423,10 @@ tcp_sacktag_write_queue(struct sock *sk, struct sk_buff *ack_skb, unsigned char *ptr = (skb_transport_header(ack_skb) + TCP_SKB_CB(ack_skb)->sacked); struct tcp_sack_block_wire *sp_wire = (struct tcp_sack_block_wire *)(ptr+2); - struct tcp_sack_block sp[4]; + struct tcp_sack_block sp[TCP_NUM_SACKS]; struct tcp_sack_block *cache; struct sk_buff *skb; - int num_sacks = (ptr[1] - TCPOLEN_SACK_BASE) >> 3; + int num_sacks = min(TCP_NUM_SACKS, (ptr[1] - TCPOLEN_SACK_BASE) >> 3); int used_sacks; int reord = tp->packets_out; int flag = 0; @@ -3735,8 +3735,7 @@ static void tcp_dsack_set(struct sock *sk, u32 seq, u32 end_seq) tp->rx_opt.dsack = 1; tp->duplicate_sack[0].start_seq = seq; tp->duplicate_sack[0].end_seq = end_seq; - tp->rx_opt.eff_sacks = min(tp->rx_opt.num_sacks + 1, - 4 - tp->rx_opt.tstamp_ok); + tp->rx_opt.eff_sacks = tp->rx_opt.num_sacks + 1; } } @@ -3791,9 +3790,8 @@ static void tcp_sack_maybe_coalesce(struct tcp_sock *tp) * Decrease num_sacks. */ tp->rx_opt.num_sacks--; - tp->rx_opt.eff_sacks = min(tp->rx_opt.num_sacks + - tp->rx_opt.dsack, - 4 - tp->rx_opt.tstamp_ok); + tp->rx_opt.eff_sacks = tp->rx_opt.num_sacks + + tp->rx_opt.dsack; for (i = this_sack; i < tp->rx_opt.num_sacks; i++) sp[i] = sp[i + 1]; continue; @@ -3843,7 +3841,7 @@ static void tcp_sack_new_ofo_skb(struct sock *sk, u32 seq, u32 end_seq) * * If the sack array is full, forget about the last one. */ - if (this_sack >= 4) { + if (this_sack >= TCP_NUM_SACKS) { this_sack--; tp->rx_opt.num_sacks--; sp--; @@ -3856,8 +3854,7 @@ new_sack: sp->start_seq = seq; sp->end_seq = end_seq; tp->rx_opt.num_sacks++; - tp->rx_opt.eff_sacks = min(tp->rx_opt.num_sacks + tp->rx_opt.dsack, - 4 - tp->rx_opt.tstamp_ok); + tp->rx_opt.eff_sacks = tp->rx_opt.num_sacks + tp->rx_opt.dsack; } /* RCV.NXT advances, some SACKs should be eaten. */ @@ -3894,9 +3891,8 @@ static void tcp_sack_remove(struct tcp_sock *tp) } if (num_sacks != tp->rx_opt.num_sacks) { tp->rx_opt.num_sacks = num_sacks; - tp->rx_opt.eff_sacks = min(tp->rx_opt.num_sacks + - tp->rx_opt.dsack, - 4 - tp->rx_opt.tstamp_ok); + tp->rx_opt.eff_sacks = tp->rx_opt.num_sacks + + tp->rx_opt.dsack; } } @@ -3975,8 +3971,7 @@ static void tcp_data_queue(struct sock *sk, struct sk_buff *skb) if (tp->rx_opt.dsack) { tp->rx_opt.dsack = 0; - tp->rx_opt.eff_sacks = min_t(unsigned int, tp->rx_opt.num_sacks, - 4 - tp->rx_opt.tstamp_ok); + tp->rx_opt.eff_sacks = tp->rx_opt.num_sacks; } /* Queue data for delivery to the user. -- cgit v1.2.2 From bdccc4ca13a639d759206c5b21ed73f8a813eaba Mon Sep 17 00:00:00 2001 From: Daniel Lezcano Date: Sat, 19 Jul 2008 00:15:13 -0700 Subject: tcp: fix kernel panic with listening_get_next # BUG: unable to handle kernel NULL pointer dereference at 0000000000000038 IP: [] listening_get_next+0x50/0x1b3 PGD 11e4b9067 PUD 11d16c067 PMD 0 Oops: 0000 [1] SMP last sysfs file: /sys/devices/system/cpu/cpu3/cache/index2/shared_cpu_map CPU 3 Modules linked in: bridge ipv6 button battery ac loop dm_mod tg3 ext3 jbd edd fan thermal processor thermal_sys hwmon sg sata_svw libata dock serverworks sd_mod scsi_mod ide_disk ide_core [last unloaded: freq_table] Pid: 3368, comm: slpd Not tainted 2.6.26-rc2-mm1-lxc4 #1 RIP: 0010:[] [] listening_get_next+0x50/0x1b3 RSP: 0018:ffff81011e1fbe18 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffff8100be0ad3c0 RCX: ffff8100619f50c0 RDX: ffffffff82475be0 RSI: ffff81011d9ae6c0 RDI: ffff8100be0ad508 RBP: ffff81011f4f1240 R08: 00000000ffffffff R09: ffff8101185b6780 R10: 000000000000002d R11: ffffffff820fdbfa R12: ffff8100be0ad3c8 R13: ffff8100be0ad6a0 R14: ffff8100be0ad3c0 R15: ffffffff825b8ce0 FS: 00007f6a0ebd16d0(0000) GS:ffff81011f424540(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 0000000000000038 CR3: 000000011dc20000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process slpd (pid: 3368, threadinfo ffff81011e1fa000, task ffff81011f4b8660) Stack: 00000000000002ee ffff81011f5a57c0 ffff81011f4f1240 ffff81011e1fbe90 0000000000001000 0000000000000000 00007fff16bf2590 ffffffff821ed9c8 ffff81011f5a57c0 ffff81011d9ae6c0 000000000000041a ffffffff820b0abd Call Trace: [] ? tcp_seq_next+0x34/0x7e [] ? seq_read+0x1aa/0x29d [] ? proc_reg_read+0x73/0x8e [] ? vfs_read+0xaa/0x152 [] ? sys_read+0x45/0x6e [] ? system_call_after_swapgs+0x7b/0x80 Code: 31 a9 25 00 e9 b5 00 00 00 ff 45 20 83 7d 0c 01 75 79 4c 8b 75 10 48 8b 0e eb 1d 48 8b 51 20 0f b7 45 08 39 02 75 0e 48 8b 41 28 <4c> 39 78 38 0f 84 93 00 00 00 48 8b 09 48 85 c9 75 de 8b 55 1c RIP [] listening_get_next+0x50/0x1b3 RSP CR2: 0000000000000038 This kernel panic appears with CONFIG_NET_NS=y. How to reproduce ? On the buggy host (host A) * ip addr add 1.2.3.4/24 dev eth0 On a remote host (host B) * ip addr add 1.2.3.5/24 dev eth0 * iptables -A INPUT -p tcp -s 1.2.3.4 -j DROP * ssh 1.2.3.4 On host A: * netstat -ta or cat /proc/net/tcp This bug happens when reading /proc/net/tcp[6] when there is a req_sock at the SYN_RECV state. When a SYN is received the minisock is created and the sk field is set to NULL. In the listening_get_next function, we try to look at the field req->sk->sk_net. When looking at how to fix this bug, I noticed that is useless to do the check for the minisock belonging to the namespace. A minisock belongs to a listen point and this one is per namespace, so when browsing the minisock they are always per namespace. Signed-off-by: Daniel Lezcano Signed-off-by: David S. Miller --- net/ipv4/tcp_ipv4.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'net') diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 5400d75ff17a..a82df6307567 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -1892,8 +1892,7 @@ static void *listening_get_next(struct seq_file *seq, void *cur) req = req->dl_next; while (1) { while (req) { - if (req->rsk_ops->family == st->family && - net_eq(sock_net(req->sk), net)) { + if (req->rsk_ops->family == st->family) { cur = req; goto out; } -- cgit v1.2.2 From 725a8ff04a5dc473cd9d8eb7817ca96fc36c7789 Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Sat, 19 Jul 2008 00:28:58 -0700 Subject: ipv6: remove unused parameter from ip6_ra_control Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- net/ipv6/ipv6_sockglue.c | 7 ++----- net/ipv6/raw.c | 2 +- 2 files changed, 3 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/ipv6/ipv6_sockglue.c b/net/ipv6/ipv6_sockglue.c index 030c0c956f9d..8c6ea07f4d56 100644 --- a/net/ipv6/ipv6_sockglue.c +++ b/net/ipv6/ipv6_sockglue.c @@ -59,7 +59,7 @@ DEFINE_SNMP_STAT(struct ipstats_mib, ipv6_statistics) __read_mostly; struct ip6_ra_chain *ip6_ra_chain; DEFINE_RWLOCK(ip6_ra_lock); -int ip6_ra_control(struct sock *sk, int sel, void (*destructor)(struct sock *)) +int ip6_ra_control(struct sock *sk, int sel) { struct ip6_ra_chain *ra, *new_ra, **rap; @@ -81,8 +81,6 @@ int ip6_ra_control(struct sock *sk, int sel, void (*destructor)(struct sock *)) *rap = ra->next; write_unlock_bh(&ip6_ra_lock); - if (ra->destructor) - ra->destructor(sk); sock_put(sk); kfree(ra); return 0; @@ -94,7 +92,6 @@ int ip6_ra_control(struct sock *sk, int sel, void (*destructor)(struct sock *)) } new_ra->sk = sk; new_ra->sel = sel; - new_ra->destructor = destructor; new_ra->next = ra; *rap = new_ra; sock_hold(sk); @@ -632,7 +629,7 @@ done: case IPV6_ROUTER_ALERT: if (optlen < sizeof(int)) goto e_inval; - retv = ip6_ra_control(sk, val, NULL); + retv = ip6_ra_control(sk, val); break; case IPV6_MTU_DISCOVER: if (optlen < sizeof(int)) diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c index 34cfb3f41c2c..01d47674f7e5 100644 --- a/net/ipv6/raw.c +++ b/net/ipv6/raw.c @@ -1157,7 +1157,7 @@ static int rawv6_ioctl(struct sock *sk, int cmd, unsigned long arg) static void rawv6_close(struct sock *sk, long timeout) { if (inet_sk(sk)->num == IPPROTO_RAW) - ip6_ra_control(sk, -1, NULL); + ip6_ra_control(sk, -1); ip6mr_sk_done(sk); sk_common_release(sk); } -- cgit v1.2.2 From 721499e8931c5732202481ae24f2dfbf9910f129 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Sat, 19 Jul 2008 22:34:43 -0700 Subject: netns: Use net_eq() to compare net-namespaces for optimization. Without CONFIG_NET_NS, namespace is always &init_net. Compiler will be able to omit namespace comparisons with this patch. Signed-off-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller --- net/appletalk/aarp.c | 4 ++-- net/appletalk/ddp.c | 6 +++--- net/atm/clip.c | 2 +- net/atm/mpc.c | 2 +- net/ax25/af_ax25.c | 2 +- net/ax25/ax25_in.c | 2 +- net/bridge/br_notify.c | 2 +- net/bridge/br_stp_bpdu.c | 2 +- net/can/af_can.c | 4 ++-- net/can/bcm.c | 2 +- net/can/raw.c | 2 +- net/core/pktgen.c | 2 +- net/decnet/af_decnet.c | 2 +- net/decnet/dn_route.c | 2 +- net/econet/af_econet.c | 4 ++-- net/ipv4/igmp.c | 26 +++++++++++++------------- net/ipv4/ipconfig.c | 4 ++-- net/ipv4/ipmr.c | 2 +- net/ipv4/netfilter/ip_queue.c | 2 +- net/ipv4/netfilter/ipt_MASQUERADE.c | 2 +- net/ipv6/ip6mr.c | 2 +- net/ipv6/netfilter/ip6_queue.c | 2 +- net/ipv6/proc.c | 2 +- net/ipx/af_ipx.c | 4 ++-- net/irda/irlap_frame.c | 2 +- net/llc/llc_input.c | 2 +- net/netfilter/nf_sockopt.c | 2 +- net/netfilter/nfnetlink_queue.c | 2 +- net/netlabel/netlabel_unlabeled.c | 2 +- net/netrom/af_netrom.c | 2 +- net/packet/af_packet.c | 2 +- net/rose/af_rose.c | 2 +- net/sctp/protocol.c | 2 +- net/tipc/eth_media.c | 4 ++-- net/wireless/wext.c | 2 +- net/x25/af_x25.c | 2 +- net/x25/x25_dev.c | 2 +- net/xfrm/xfrm_policy.c | 2 +- 38 files changed, 58 insertions(+), 58 deletions(-) (limited to 'net') diff --git a/net/appletalk/aarp.c b/net/appletalk/aarp.c index 25aa37ce9430..b25c1e909d14 100644 --- a/net/appletalk/aarp.c +++ b/net/appletalk/aarp.c @@ -333,7 +333,7 @@ static int aarp_device_event(struct notifier_block *this, unsigned long event, struct net_device *dev = ptr; int ct; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event == NETDEV_DOWN) { @@ -716,7 +716,7 @@ static int aarp_rcv(struct sk_buff *skb, struct net_device *dev, struct atalk_addr sa, *ma, da; struct atalk_iface *ifa; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto out0; /* We only do Ethernet SNAP AARP. */ diff --git a/net/appletalk/ddp.c b/net/appletalk/ddp.c index 44cd42f7786b..07b5b82c5eab 100644 --- a/net/appletalk/ddp.c +++ b/net/appletalk/ddp.c @@ -648,7 +648,7 @@ static int ddp_device_event(struct notifier_block *this, unsigned long event, { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event == NETDEV_DOWN) @@ -1405,7 +1405,7 @@ static int atalk_rcv(struct sk_buff *skb, struct net_device *dev, int origlen; __u16 len_hops; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto freeit; /* Don't mangle buffer if shared */ @@ -1493,7 +1493,7 @@ freeit: static int ltalk_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, struct net_device *orig_dev) { - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto freeit; /* Expand any short form frames */ diff --git a/net/atm/clip.c b/net/atm/clip.c index 6f8223ebf551..5b5b96344ce6 100644 --- a/net/atm/clip.c +++ b/net/atm/clip.c @@ -612,7 +612,7 @@ static int clip_device_event(struct notifier_block *this, unsigned long event, { struct net_device *dev = arg; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event == NETDEV_UNREGISTER) { diff --git a/net/atm/mpc.c b/net/atm/mpc.c index 9db332e7a6c0..4fccaa1e07be 100644 --- a/net/atm/mpc.c +++ b/net/atm/mpc.c @@ -964,7 +964,7 @@ static int mpoa_event_listener(struct notifier_block *mpoa_notifier, unsigned lo dev = (struct net_device *)dev_ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (dev->name == NULL || strncmp(dev->name, "lec", 3)) diff --git a/net/ax25/af_ax25.c b/net/ax25/af_ax25.c index 97eaa23ad9ea..01c83e2a4c19 100644 --- a/net/ax25/af_ax25.c +++ b/net/ax25/af_ax25.c @@ -116,7 +116,7 @@ static int ax25_device_event(struct notifier_block *this, unsigned long event, { struct net_device *dev = (struct net_device *)ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* Reject non AX.25 devices */ diff --git a/net/ax25/ax25_in.c b/net/ax25/ax25_in.c index 33790a8efbc8..4a5ba978a804 100644 --- a/net/ax25/ax25_in.c +++ b/net/ax25/ax25_in.c @@ -451,7 +451,7 @@ int ax25_kiss_rcv(struct sk_buff *skb, struct net_device *dev, skb->sk = NULL; /* Initially we don't know who it's for */ skb->destructor = NULL; /* Who initializes this, dammit?! */ - if (dev_net(dev) != &init_net) { + if (!net_eq(dev_net(dev), &init_net)) { kfree_skb(skb); return 0; } diff --git a/net/bridge/br_notify.c b/net/bridge/br_notify.c index 88d8ec7b3142..76340bdd052e 100644 --- a/net/bridge/br_notify.c +++ b/net/bridge/br_notify.c @@ -35,7 +35,7 @@ static int br_device_event(struct notifier_block *unused, unsigned long event, v struct net_bridge_port *p = dev->br_port; struct net_bridge *br; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* not a port of a bridge */ diff --git a/net/bridge/br_stp_bpdu.c b/net/bridge/br_stp_bpdu.c index 996476174517..8b200f96f722 100644 --- a/net/bridge/br_stp_bpdu.c +++ b/net/bridge/br_stp_bpdu.c @@ -140,7 +140,7 @@ void br_stp_rcv(const struct stp_proto *proto, struct sk_buff *skb, struct net_bridge *br; const unsigned char *buf; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto err; if (!p) diff --git a/net/can/af_can.c b/net/can/af_can.c index 484bbf6dd032..8035fbf526ae 100644 --- a/net/can/af_can.c +++ b/net/can/af_can.c @@ -615,7 +615,7 @@ static int can_rcv(struct sk_buff *skb, struct net_device *dev, struct can_frame *cf = (struct can_frame *)skb->data; int matches; - if (dev->type != ARPHRD_CAN || dev_net(dev) != &init_net) { + if (dev->type != ARPHRD_CAN || !net_eq(dev_net(dev), &init_net)) { kfree_skb(skb); return 0; } @@ -728,7 +728,7 @@ static int can_notifier(struct notifier_block *nb, unsigned long msg, struct net_device *dev = (struct net_device *)data; struct dev_rcv_lists *d; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (dev->type != ARPHRD_CAN) diff --git a/net/can/bcm.c b/net/can/bcm.c index 72c2ce904f83..d0dd382001e2 100644 --- a/net/can/bcm.c +++ b/net/can/bcm.c @@ -1303,7 +1303,7 @@ static int bcm_notifier(struct notifier_block *nb, unsigned long msg, struct bcm_op *op; int notify_enodev = 0; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (dev->type != ARPHRD_CAN) diff --git a/net/can/raw.c b/net/can/raw.c index 3e46ee36a1aa..6e0663faaf9f 100644 --- a/net/can/raw.c +++ b/net/can/raw.c @@ -210,7 +210,7 @@ static int raw_notifier(struct notifier_block *nb, struct raw_sock *ro = container_of(nb, struct raw_sock, notifier); struct sock *sk = &ro->sk; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (dev->type != ARPHRD_CAN) diff --git a/net/core/pktgen.c b/net/core/pktgen.c index 906802db4ed4..c7d484f7e1c4 100644 --- a/net/core/pktgen.c +++ b/net/core/pktgen.c @@ -1875,7 +1875,7 @@ static int pktgen_device_event(struct notifier_block *unused, { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* It is OK that we do not hold the group lock right now, diff --git a/net/decnet/af_decnet.c b/net/decnet/af_decnet.c index 61b7df577ddd..3c23ab33dbc0 100644 --- a/net/decnet/af_decnet.c +++ b/net/decnet/af_decnet.c @@ -2089,7 +2089,7 @@ static int dn_device_event(struct notifier_block *this, unsigned long event, { struct net_device *dev = (struct net_device *)ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; switch(event) { diff --git a/net/decnet/dn_route.c b/net/decnet/dn_route.c index f50e88bf2661..821bd1cdec04 100644 --- a/net/decnet/dn_route.c +++ b/net/decnet/dn_route.c @@ -580,7 +580,7 @@ int dn_route_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type struct dn_dev *dn = (struct dn_dev *)dev->dn_ptr; unsigned char padlen = 0; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto dump_it; if (dn == NULL) diff --git a/net/econet/af_econet.c b/net/econet/af_econet.c index d35127bb84e1..8789d2bb1b06 100644 --- a/net/econet/af_econet.c +++ b/net/econet/af_econet.c @@ -1062,7 +1062,7 @@ static int econet_rcv(struct sk_buff *skb, struct net_device *dev, struct packet struct sock *sk; struct ec_device *edev = dev->ec_ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; if (skb->pkt_type == PACKET_OTHERHOST) @@ -1119,7 +1119,7 @@ static int econet_notifier(struct notifier_block *this, unsigned long msg, void struct net_device *dev = (struct net_device *)data; struct ec_device *edev; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; switch (msg) { diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c index 68e84a933e90..6203ece53606 100644 --- a/net/ipv4/igmp.c +++ b/net/ipv4/igmp.c @@ -1196,7 +1196,7 @@ void ip_mc_inc_group(struct in_device *in_dev, __be32 addr) ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; for (im=in_dev->mc_list; im; im=im->next) { @@ -1278,7 +1278,7 @@ void ip_mc_dec_group(struct in_device *in_dev, __be32 addr) ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; for (ip=&in_dev->mc_list; (i=*ip)!=NULL; ip=&i->next) { @@ -1308,7 +1308,7 @@ void ip_mc_down(struct in_device *in_dev) ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; for (i=in_dev->mc_list; i; i=i->next) @@ -1331,7 +1331,7 @@ void ip_mc_init_dev(struct in_device *in_dev) { ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; in_dev->mc_tomb = NULL; @@ -1357,7 +1357,7 @@ void ip_mc_up(struct in_device *in_dev) ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; ip_mc_inc_group(in_dev, IGMP_ALL_HOSTS); @@ -1376,7 +1376,7 @@ void ip_mc_destroy_dev(struct in_device *in_dev) ASSERT_RTNL(); - if (dev_net(in_dev->dev) != &init_net) + if (!net_eq(dev_net(in_dev->dev), &init_net)) return; /* Deactivate timers */ @@ -1760,7 +1760,7 @@ int ip_mc_join_group(struct sock *sk , struct ip_mreqn *imr) if (!ipv4_is_multicast(addr)) return -EINVAL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -1831,7 +1831,7 @@ int ip_mc_leave_group(struct sock *sk, struct ip_mreqn *imr) u32 ifindex; int ret = -EADDRNOTAVAIL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -1879,7 +1879,7 @@ int ip_mc_source(int add, int omode, struct sock *sk, struct if (!ipv4_is_multicast(addr)) return -EINVAL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -2015,7 +2015,7 @@ int ip_mc_msfilter(struct sock *sk, struct ip_msfilter *msf, int ifindex) msf->imsf_fmode != MCAST_EXCLUDE) return -EINVAL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -2098,7 +2098,7 @@ int ip_mc_msfget(struct sock *sk, struct ip_msfilter *msf, if (!ipv4_is_multicast(addr)) return -EINVAL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -2163,7 +2163,7 @@ int ip_mc_gsfget(struct sock *sk, struct group_filter *gsf, if (!ipv4_is_multicast(addr)) return -EINVAL; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -EPROTONOSUPPORT; rtnl_lock(); @@ -2250,7 +2250,7 @@ void ip_mc_drop_socket(struct sock *sk) if (inet->mc_list == NULL) return; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return; rtnl_lock(); diff --git a/net/ipv4/ipconfig.c b/net/ipv4/ipconfig.c index b88aa9afa42e..42065fff46c4 100644 --- a/net/ipv4/ipconfig.c +++ b/net/ipv4/ipconfig.c @@ -432,7 +432,7 @@ ic_rarp_recv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt unsigned char *sha, *tha; /* s for "source", t for "target" */ struct ic_device *d; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL) @@ -852,7 +852,7 @@ static int __init ic_bootp_recv(struct sk_buff *skb, struct net_device *dev, str struct ic_device *d; int len, ext_len; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; /* Perform verifications before taking the lock. */ diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index 033c712c3a5d..c519b8d30eee 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c @@ -1124,7 +1124,7 @@ static int ipmr_device_event(struct notifier_block *this, unsigned long event, v struct vif_device *v; int ct; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event != NETDEV_UNREGISTER) diff --git a/net/ipv4/netfilter/ip_queue.c b/net/ipv4/netfilter/ip_queue.c index aa33a4a7a715..432ce9d1c11c 100644 --- a/net/ipv4/netfilter/ip_queue.c +++ b/net/ipv4/netfilter/ip_queue.c @@ -477,7 +477,7 @@ ipq_rcv_dev_event(struct notifier_block *this, { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* Drop any packets associated with the downed device */ diff --git a/net/ipv4/netfilter/ipt_MASQUERADE.c b/net/ipv4/netfilter/ipt_MASQUERADE.c index 84c26dd27d81..0841aefaa503 100644 --- a/net/ipv4/netfilter/ipt_MASQUERADE.c +++ b/net/ipv4/netfilter/ipt_MASQUERADE.c @@ -120,7 +120,7 @@ static int masq_device_event(struct notifier_block *this, { const struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event == NETDEV_DOWN) { diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c index 0b41aa2675f5..095bc453ff4c 100644 --- a/net/ipv6/ip6mr.c +++ b/net/ipv6/ip6mr.c @@ -935,7 +935,7 @@ static int ip6mr_device_event(struct notifier_block *this, struct mif_device *v; int ct; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event != NETDEV_UNREGISTER) diff --git a/net/ipv6/netfilter/ip6_queue.c b/net/ipv6/netfilter/ip6_queue.c index 1b8815f6153d..5859c046cbc4 100644 --- a/net/ipv6/netfilter/ip6_queue.c +++ b/net/ipv6/netfilter/ip6_queue.c @@ -480,7 +480,7 @@ ipq_rcv_dev_event(struct notifier_block *this, { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* Drop any packets associated with the downed device */ diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c index 70940b3654a1..f82f6074cf85 100644 --- a/net/ipv6/proc.c +++ b/net/ipv6/proc.c @@ -214,7 +214,7 @@ int snmp6_register_dev(struct inet6_dev *idev) if (!idev || !idev->dev) return -EINVAL; - if (dev_net(idev->dev) != &init_net) + if (!net_eq(dev_net(idev->dev), &init_net)) return 0; if (!proc_net_devsnmp6) diff --git a/net/ipx/af_ipx.c b/net/ipx/af_ipx.c index 81ae8735f5e3..b6e70f92e7fb 100644 --- a/net/ipx/af_ipx.c +++ b/net/ipx/af_ipx.c @@ -335,7 +335,7 @@ static int ipxitf_device_event(struct notifier_block *notifier, struct net_device *dev = ptr; struct ipx_interface *i, *tmp; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event != NETDEV_DOWN && event != NETDEV_UP) @@ -1636,7 +1636,7 @@ static int ipx_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_ty u16 ipx_pktsize; int rc = 0; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; /* Not ours */ diff --git a/net/irda/irlap_frame.c b/net/irda/irlap_frame.c index 90894534f3cc..f17b65af9c9b 100644 --- a/net/irda/irlap_frame.c +++ b/net/irda/irlap_frame.c @@ -1326,7 +1326,7 @@ int irlap_driver_rcv(struct sk_buff *skb, struct net_device *dev, int command; __u8 control; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto out; /* FIXME: should we get our own field? */ diff --git a/net/llc/llc_input.c b/net/llc/llc_input.c index 1c45f172991e..57ad974e4d94 100644 --- a/net/llc/llc_input.c +++ b/net/llc/llc_input.c @@ -150,7 +150,7 @@ int llc_rcv(struct sk_buff *skb, struct net_device *dev, int (*rcv)(struct sk_buff *, struct net_device *, struct packet_type *, struct net_device *); - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; /* diff --git a/net/netfilter/nf_sockopt.c b/net/netfilter/nf_sockopt.c index 69d699f95f4c..01489681fa96 100644 --- a/net/netfilter/nf_sockopt.c +++ b/net/netfilter/nf_sockopt.c @@ -65,7 +65,7 @@ static struct nf_sockopt_ops *nf_sockopt_find(struct sock *sk, int pf, { struct nf_sockopt_ops *ops; - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return ERR_PTR(-ENOPROTOOPT); if (mutex_lock_interruptible(&nf_sockopt_mutex) != 0) diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c index 04e9c965f8ca..8c860112ce05 100644 --- a/net/netfilter/nfnetlink_queue.c +++ b/net/netfilter/nfnetlink_queue.c @@ -555,7 +555,7 @@ nfqnl_rcv_dev_event(struct notifier_block *this, { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* Drop any packets associated with the downed device */ diff --git a/net/netlabel/netlabel_unlabeled.c b/net/netlabel/netlabel_unlabeled.c index 56f80872924e..921c118ead89 100644 --- a/net/netlabel/netlabel_unlabeled.c +++ b/net/netlabel/netlabel_unlabeled.c @@ -954,7 +954,7 @@ static int netlbl_unlhsh_netdev_handler(struct notifier_block *this, struct net_device *dev = ptr; struct netlbl_unlhsh_iface *iface = NULL; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; /* XXX - should this be a check for NETDEV_DOWN or _UNREGISTER? */ diff --git a/net/netrom/af_netrom.c b/net/netrom/af_netrom.c index d41be0d66eb0..fccc250f95f5 100644 --- a/net/netrom/af_netrom.c +++ b/net/netrom/af_netrom.c @@ -118,7 +118,7 @@ static int nr_device_event(struct notifier_block *this, unsigned long event, voi { struct net_device *dev = (struct net_device *)ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event != NETDEV_DOWN) diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index de73bcb5235a..d56cae112dc8 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -1677,7 +1677,7 @@ static int packet_ioctl(struct socket *sock, unsigned int cmd, case SIOCGIFDSTADDR: case SIOCSIFDSTADDR: case SIOCSIFFLAGS: - if (sock_net(sk) != &init_net) + if (!net_eq(sock_net(sk), &init_net)) return -ENOIOCTLCMD; return inet_dgram_ops.ioctl(sock, cmd, arg); #endif diff --git a/net/rose/af_rose.c b/net/rose/af_rose.c index f3a691f34909..dbc963b4f5fb 100644 --- a/net/rose/af_rose.c +++ b/net/rose/af_rose.c @@ -209,7 +209,7 @@ static int rose_device_event(struct notifier_block *this, unsigned long event, { struct net_device *dev = (struct net_device *)ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (event != NETDEV_DOWN) diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c index ed9acffe8105..a6e0818bcff5 100644 --- a/net/sctp/protocol.c +++ b/net/sctp/protocol.c @@ -668,7 +668,7 @@ static int sctp_inetaddr_event(struct notifier_block *this, unsigned long ev, struct sctp_sockaddr_entry *temp; int found = 0; - if (dev_net(ifa->ifa_dev->dev) != &init_net) + if (!net_eq(dev_net(ifa->ifa_dev->dev), &init_net)) return NOTIFY_DONE; switch (ev) { diff --git a/net/tipc/eth_media.c b/net/tipc/eth_media.c index bc72fbc4f8b8..fe43ef7dd7e3 100644 --- a/net/tipc/eth_media.c +++ b/net/tipc/eth_media.c @@ -101,7 +101,7 @@ static int recv_msg(struct sk_buff *buf, struct net_device *dev, struct eth_bearer *eb_ptr = (struct eth_bearer *)pt->af_packet_priv; u32 size; - if (dev_net(dev) != &init_net) { + if (!net_eq(dev_net(dev), &init_net)) { kfree_skb(buf); return 0; } @@ -198,7 +198,7 @@ static int recv_notification(struct notifier_block *nb, unsigned long evt, struct eth_bearer *eb_ptr = ð_bearers[0]; struct eth_bearer *stop = ð_bearers[MAX_ETH_BEARERS]; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; while ((eb_ptr->dev != dev)) { diff --git a/net/wireless/wext.c b/net/wireless/wext.c index 273a84359998..df5b3886c36b 100644 --- a/net/wireless/wext.c +++ b/net/wireless/wext.c @@ -1299,7 +1299,7 @@ static void rtmsg_iwinfo(struct net_device *dev, char *event, int event_len) struct sk_buff *skb; int err; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return; skb = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_ATOMIC); diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c index 7b1c6ef04553..9fc5b023d111 100644 --- a/net/x25/af_x25.c +++ b/net/x25/af_x25.c @@ -191,7 +191,7 @@ static int x25_device_event(struct notifier_block *this, unsigned long event, struct net_device *dev = ptr; struct x25_neigh *nb; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; if (dev->type == ARPHRD_X25 diff --git a/net/x25/x25_dev.c b/net/x25/x25_dev.c index 3ff206c0ae94..3e1efe534645 100644 --- a/net/x25/x25_dev.c +++ b/net/x25/x25_dev.c @@ -95,7 +95,7 @@ int x25_lapb_receive_frame(struct sk_buff *skb, struct net_device *dev, struct sk_buff *nskb; struct x25_neigh *nb; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) goto drop; nskb = skb_copy(skb, GFP_ATOMIC); diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c index cae9fd815543..841b32a2e680 100644 --- a/net/xfrm/xfrm_policy.c +++ b/net/xfrm/xfrm_policy.c @@ -2360,7 +2360,7 @@ static int xfrm_dev_event(struct notifier_block *this, unsigned long event, void { struct net_device *dev = ptr; - if (dev_net(dev) != &init_net) + if (!net_eq(dev_net(dev), &init_net)) return NOTIFY_DONE; switch (event) { -- cgit v1.2.2 From 53b7997fd5c62408d10b9aafb38974ce90fd2356 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Sat, 19 Jul 2008 22:35:03 -0700 Subject: ipv6 netns: Make several "global" sysctl variables namespace aware. Signed-off-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller --- net/ipv6/addrconf.c | 8 +++++--- net/ipv6/anycast.c | 2 +- net/ipv6/exthdrs.c | 2 +- net/ipv6/ip6_input.c | 2 +- net/ipv6/ip6_output.c | 4 ++-- net/ipv6/ipv6_sockglue.c | 2 +- net/ipv6/mcast.c | 2 +- net/ipv6/ndisc.c | 11 +++++++---- net/ipv6/route.c | 4 ++-- 9 files changed, 21 insertions(+), 16 deletions(-) (limited to 'net') diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 30184e0dd74c..580ae506c399 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -1863,6 +1863,7 @@ void addrconf_prefix_rcv(struct net_device *dev, u8 *opt, int len) struct inet6_ifaddr * ifp; struct in6_addr addr; int create = 0, update_lft = 0; + struct net *net = dev_net(dev); if (pinfo->prefix_len == 64) { memcpy(&addr, &pinfo->prefix, 8); @@ -1881,7 +1882,7 @@ void addrconf_prefix_rcv(struct net_device *dev, u8 *opt, int len) ok: - ifp = ipv6_get_ifaddr(dev_net(dev), &addr, dev, 1); + ifp = ipv6_get_ifaddr(net, &addr, dev, 1); if (ifp == NULL && valid_lft) { int max_addresses = in6_dev->cnf.max_addresses; @@ -1889,7 +1890,7 @@ ok: #ifdef CONFIG_IPV6_OPTIMISTIC_DAD if (in6_dev->cnf.optimistic_dad && - !ipv6_devconf.forwarding) + !net->ipv6.devconf_all->forwarding) addr_flags = IFA_F_OPTIMISTIC; #endif @@ -2314,11 +2315,12 @@ static void init_loopback(struct net_device *dev) static void addrconf_add_linklocal(struct inet6_dev *idev, struct in6_addr *addr) { struct inet6_ifaddr * ifp; + struct net *net = dev_net(idev->dev); u32 addr_flags = IFA_F_PERMANENT; #ifdef CONFIG_IPV6_OPTIMISTIC_DAD if (idev->cnf.optimistic_dad && - !ipv6_devconf.forwarding) + !net->ipv6.devconf_all->forwarding) addr_flags |= IFA_F_OPTIMISTIC; #endif diff --git a/net/ipv6/anycast.c b/net/ipv6/anycast.c index 4e1b29fabdf0..8336cd81cb4f 100644 --- a/net/ipv6/anycast.c +++ b/net/ipv6/anycast.c @@ -60,7 +60,7 @@ int ipv6_sock_ac_join(struct sock *sk, int ifindex, struct in6_addr *addr) struct inet6_dev *idev; struct ipv6_ac_socklist *pac; struct net *net = sock_net(sk); - int ishost = !ipv6_devconf.forwarding; + int ishost = !net->ipv6.devconf_all->forwarding; int err = 0; if (!capable(CAP_NET_ADMIN)) diff --git a/net/ipv6/exthdrs.c b/net/ipv6/exthdrs.c index 9f1084b4c0e8..837c830d6d8e 100644 --- a/net/ipv6/exthdrs.c +++ b/net/ipv6/exthdrs.c @@ -319,7 +319,7 @@ static int ipv6_rthdr_rcv(struct sk_buff *skb) int n, i; struct ipv6_rt_hdr *hdr; struct rt0_hdr *rthdr; - int accept_source_route = ipv6_devconf.accept_source_route; + int accept_source_route = dev_net(skb->dev)->ipv6.devconf_all->accept_source_route; idev = in6_dev_get(skb->dev); if (idev) { diff --git a/net/ipv6/ip6_input.c b/net/ipv6/ip6_input.c index ea81c614dde2..7e14cccd0561 100644 --- a/net/ipv6/ip6_input.c +++ b/net/ipv6/ip6_input.c @@ -249,7 +249,7 @@ int ip6_mc_input(struct sk_buff *skb) /* * IPv6 multicast router mode is now supported ;) */ - if (ipv6_devconf.mc_forwarding && + if (dev_net(skb->dev)->ipv6.devconf_all->mc_forwarding && likely(!(IP6CB(skb)->flags & IP6SKB_FORWARDED))) { /* * Okay, we try to forward - split and duplicate diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 0981c1ef3057..6407c64ea4a5 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -411,7 +411,7 @@ int ip6_forward(struct sk_buff *skb) struct inet6_skb_parm *opt = IP6CB(skb); struct net *net = dev_net(dst->dev); - if (ipv6_devconf.forwarding == 0) + if (net->ipv6.devconf_all->forwarding == 0) goto error; if (skb_warn_if_lro(skb)) @@ -458,7 +458,7 @@ int ip6_forward(struct sk_buff *skb) } /* XXX: idev->cnf.proxy_ndp? */ - if (ipv6_devconf.proxy_ndp && + if (net->ipv6.devconf_all->proxy_ndp && pneigh_lookup(&nd_tbl, net, &hdr->daddr, skb->dev, 0)) { int proxied = ip6_forward_proxy_check(skb); if (proxied > 0) diff --git a/net/ipv6/ipv6_sockglue.c b/net/ipv6/ipv6_sockglue.c index 8c6ea07f4d56..ea33b26512c2 100644 --- a/net/ipv6/ipv6_sockglue.c +++ b/net/ipv6/ipv6_sockglue.c @@ -1038,7 +1038,7 @@ static int do_ipv6_getsockopt(struct sock *sk, int level, int optname, dst_release(dst); } if (val < 0) - val = ipv6_devconf.hop_limit; + val = sock_net(sk)->ipv6.devconf_all->hop_limit; break; } diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c index bd2fe4cfafa7..1b285371124a 100644 --- a/net/ipv6/mcast.c +++ b/net/ipv6/mcast.c @@ -151,7 +151,7 @@ static int ip6_mc_leave_src(struct sock *sk, struct ipv6_mc_socklist *iml, #define IGMP6_UNSOLICITED_IVAL (10*HZ) #define MLD_QRV_DEFAULT 2 -#define MLD_V1_SEEN(idev) (ipv6_devconf.force_mld_version == 1 || \ +#define MLD_V1_SEEN(idev) (dev_net((idev)->dev)->ipv6.devconf_all->force_mld_version == 1 || \ (idev)->cnf.force_mld_version == 1 || \ ((idev)->mc_v1_seen && \ time_before(jiffies, (idev)->mc_v1_seen))) diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c index 282fdb31f8ed..beb48e3f038a 100644 --- a/net/ipv6/ndisc.c +++ b/net/ipv6/ndisc.c @@ -784,15 +784,17 @@ static void ndisc_recv_ns(struct sk_buff *skb) idev = ifp->idev; } else { + struct net *net = dev_net(dev); + idev = in6_dev_get(dev); if (!idev) { /* XXX: count this drop? */ return; } - if (ipv6_chk_acast_addr(dev_net(dev), dev, &msg->target) || + if (ipv6_chk_acast_addr(net, dev, &msg->target) || (idev->cnf.forwarding && - (ipv6_devconf.proxy_ndp || idev->cnf.proxy_ndp) && + (net->ipv6.devconf_all->proxy_ndp || idev->cnf.proxy_ndp) && (is_router = pndisc_is_router(&msg->target, dev)) >= 0)) { if (!(NEIGH_CB(skb)->flags & LOCALLY_ENQUEUED) && skb->pkt_type != PACKET_HOST && @@ -921,6 +923,7 @@ static void ndisc_recv_na(struct sk_buff *skb) if (neigh) { u8 old_flags = neigh->flags; + struct net *net = dev_net(dev); if (neigh->nud_state & NUD_FAILED) goto out; @@ -931,8 +934,8 @@ static void ndisc_recv_na(struct sk_buff *skb) * has already sent a NA to us. */ if (lladdr && !memcmp(lladdr, dev->dev_addr, dev->addr_len) && - ipv6_devconf.forwarding && ipv6_devconf.proxy_ndp && - pneigh_lookup(&nd_tbl, dev_net(dev), &msg->target, dev, 0)) { + net->ipv6.devconf_all->forwarding && net->ipv6.devconf_all->proxy_ndp && + pneigh_lookup(&nd_tbl, net, &msg->target, dev, 0)) { /* XXX: idev->cnf.prixy_ndp */ goto out; } diff --git a/net/ipv6/route.c b/net/ipv6/route.c index cb8a51271b67..615b328de251 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -676,7 +676,7 @@ static struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table, int strict = 0; int attempts = 3; int err; - int reachable = ipv6_devconf.forwarding ? 0 : RT6_LOOKUP_F_REACHABLE; + int reachable = net->ipv6.devconf_all->forwarding ? 0 : RT6_LOOKUP_F_REACHABLE; strict |= flags & RT6_LOOKUP_F_IFACE; @@ -1058,7 +1058,7 @@ int ip6_dst_hoplimit(struct dst_entry *dst) hoplimit = idev->cnf.hop_limit; in6_dev_put(idev); } else - hoplimit = ipv6_devconf.hop_limit; + hoplimit = dev_net(dev)->ipv6.devconf_all->hop_limit; } return hoplimit; } -- cgit v1.2.2 From 230b183921ecbaa5fedc0d35ad6ba7bb64b6e06a Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Sat, 19 Jul 2008 22:35:47 -0700 Subject: net: Use standard structures for generic socket address structures. Use sockaddr_storage{} for generic socket address storage and ensures proper alignment. Use sockaddr{} for pointers to omit several casts. Signed-off-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller --- net/compat.c | 2 +- net/core/iovec.c | 2 +- net/socket.c | 82 +++++++++++++++++++++++++++++++------------------------- 3 files changed, 48 insertions(+), 38 deletions(-) (limited to 'net') diff --git a/net/compat.c b/net/compat.c index c823f6f290cb..6e1b03b51933 100644 --- a/net/compat.c +++ b/net/compat.c @@ -75,7 +75,7 @@ int get_compat_msghdr(struct msghdr *kmsg, struct compat_msghdr __user *umsg) /* I've named the args so it is easy to tell whose space the pointers are in. */ int verify_compat_iovec(struct msghdr *kern_msg, struct iovec *kern_iov, - char *kern_address, int mode) + struct sockaddr *kern_address, int mode) { int tot_len; diff --git a/net/core/iovec.c b/net/core/iovec.c index 755c37fdaee7..4c9c0121c9da 100644 --- a/net/core/iovec.c +++ b/net/core/iovec.c @@ -36,7 +36,7 @@ * in any case. */ -int verify_iovec(struct msghdr *m, struct iovec *iov, char *address, int mode) +int verify_iovec(struct msghdr *m, struct iovec *iov, struct sockaddr *address, int mode) { int size, err, ct; diff --git a/net/socket.c b/net/socket.c index 81fe82513046..1ba57d888981 100644 --- a/net/socket.c +++ b/net/socket.c @@ -180,9 +180,9 @@ static DEFINE_PER_CPU(int, sockets_in_use) = 0; * invalid addresses -EFAULT is returned. On a success 0 is returned. */ -int move_addr_to_kernel(void __user *uaddr, int ulen, void *kaddr) +int move_addr_to_kernel(void __user *uaddr, int ulen, struct sockaddr *kaddr) { - if (ulen < 0 || ulen > MAX_SOCK_ADDR) + if (ulen < 0 || ulen > sizeof(struct sockaddr_storage)) return -EINVAL; if (ulen == 0) return 0; @@ -208,7 +208,7 @@ int move_addr_to_kernel(void __user *uaddr, int ulen, void *kaddr) * specified. Zero is returned for a success. */ -int move_addr_to_user(void *kaddr, int klen, void __user *uaddr, +int move_addr_to_user(struct sockaddr *kaddr, int klen, void __user *uaddr, int __user *ulen) { int err; @@ -219,7 +219,7 @@ int move_addr_to_user(void *kaddr, int klen, void __user *uaddr, return err; if (len > klen) len = klen; - if (len < 0 || len > MAX_SOCK_ADDR) + if (len < 0 || len > sizeof(struct sockaddr_storage)) return -EINVAL; if (len) { if (audit_sockaddr(klen, kaddr)) @@ -1342,20 +1342,20 @@ out_fd: asmlinkage long sys_bind(int fd, struct sockaddr __user *umyaddr, int addrlen) { struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int err, fput_needed; sock = sockfd_lookup_light(fd, &err, &fput_needed); if (sock) { - err = move_addr_to_kernel(umyaddr, addrlen, address); + err = move_addr_to_kernel(umyaddr, addrlen, (struct sockaddr *)&address); if (err >= 0) { err = security_socket_bind(sock, - (struct sockaddr *)address, + (struct sockaddr *)&address, addrlen); if (!err) err = sock->ops->bind(sock, (struct sockaddr *) - address, addrlen); + &address, addrlen); } fput_light(sock->file, fput_needed); } @@ -1407,7 +1407,7 @@ asmlinkage long sys_accept(int fd, struct sockaddr __user *upeer_sockaddr, struct socket *sock, *newsock; struct file *newfile; int err, len, newfd, fput_needed; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; sock = sockfd_lookup_light(fd, &err, &fput_needed); if (!sock) @@ -1446,13 +1446,13 @@ asmlinkage long sys_accept(int fd, struct sockaddr __user *upeer_sockaddr, goto out_fd; if (upeer_sockaddr) { - if (newsock->ops->getname(newsock, (struct sockaddr *)address, + if (newsock->ops->getname(newsock, (struct sockaddr *)&address, &len, 2) < 0) { err = -ECONNABORTED; goto out_fd; } - err = move_addr_to_user(address, len, upeer_sockaddr, - upeer_addrlen); + err = move_addr_to_user((struct sockaddr *)&address, + len, upeer_sockaddr, upeer_addrlen); if (err < 0) goto out_fd; } @@ -1495,22 +1495,22 @@ asmlinkage long sys_connect(int fd, struct sockaddr __user *uservaddr, int addrlen) { struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int err, fput_needed; sock = sockfd_lookup_light(fd, &err, &fput_needed); if (!sock) goto out; - err = move_addr_to_kernel(uservaddr, addrlen, address); + err = move_addr_to_kernel(uservaddr, addrlen, (struct sockaddr *)&address); if (err < 0) goto out_put; err = - security_socket_connect(sock, (struct sockaddr *)address, addrlen); + security_socket_connect(sock, (struct sockaddr *)&address, addrlen); if (err) goto out_put; - err = sock->ops->connect(sock, (struct sockaddr *)address, addrlen, + err = sock->ops->connect(sock, (struct sockaddr *)&address, addrlen, sock->file->f_flags); out_put: fput_light(sock->file, fput_needed); @@ -1527,7 +1527,7 @@ asmlinkage long sys_getsockname(int fd, struct sockaddr __user *usockaddr, int __user *usockaddr_len) { struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int len, err, fput_needed; sock = sockfd_lookup_light(fd, &err, &fput_needed); @@ -1538,10 +1538,10 @@ asmlinkage long sys_getsockname(int fd, struct sockaddr __user *usockaddr, if (err) goto out_put; - err = sock->ops->getname(sock, (struct sockaddr *)address, &len, 0); + err = sock->ops->getname(sock, (struct sockaddr *)&address, &len, 0); if (err) goto out_put; - err = move_addr_to_user(address, len, usockaddr, usockaddr_len); + err = move_addr_to_user((struct sockaddr *)&address, len, usockaddr, usockaddr_len); out_put: fput_light(sock->file, fput_needed); @@ -1558,7 +1558,7 @@ asmlinkage long sys_getpeername(int fd, struct sockaddr __user *usockaddr, int __user *usockaddr_len) { struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int len, err, fput_needed; sock = sockfd_lookup_light(fd, &err, &fput_needed); @@ -1570,10 +1570,10 @@ asmlinkage long sys_getpeername(int fd, struct sockaddr __user *usockaddr, } err = - sock->ops->getname(sock, (struct sockaddr *)address, &len, + sock->ops->getname(sock, (struct sockaddr *)&address, &len, 1); if (!err) - err = move_addr_to_user(address, len, usockaddr, + err = move_addr_to_user((struct sockaddr *)&address, len, usockaddr, usockaddr_len); fput_light(sock->file, fput_needed); } @@ -1591,7 +1591,7 @@ asmlinkage long sys_sendto(int fd, void __user *buff, size_t len, int addr_len) { struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int err; struct msghdr msg; struct iovec iov; @@ -1610,10 +1610,10 @@ asmlinkage long sys_sendto(int fd, void __user *buff, size_t len, msg.msg_controllen = 0; msg.msg_namelen = 0; if (addr) { - err = move_addr_to_kernel(addr, addr_len, address); + err = move_addr_to_kernel(addr, addr_len, (struct sockaddr *)&address); if (err < 0) goto out_put; - msg.msg_name = address; + msg.msg_name = (struct sockaddr *)&address; msg.msg_namelen = addr_len; } if (sock->file->f_flags & O_NONBLOCK) @@ -1649,7 +1649,7 @@ asmlinkage long sys_recvfrom(int fd, void __user *ubuf, size_t size, struct socket *sock; struct iovec iov; struct msghdr msg; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; int err, err2; int fput_needed; @@ -1663,14 +1663,15 @@ asmlinkage long sys_recvfrom(int fd, void __user *ubuf, size_t size, msg.msg_iov = &iov; iov.iov_len = size; iov.iov_base = ubuf; - msg.msg_name = address; - msg.msg_namelen = MAX_SOCK_ADDR; + msg.msg_name = (struct sockaddr *)&address; + msg.msg_namelen = sizeof(address); if (sock->file->f_flags & O_NONBLOCK) flags |= MSG_DONTWAIT; err = sock_recvmsg(sock, &msg, size, flags); if (err >= 0 && addr != NULL) { - err2 = move_addr_to_user(address, msg.msg_namelen, addr, addr_len); + err2 = move_addr_to_user((struct sockaddr *)&address, + msg.msg_namelen, addr, addr_len); if (err2 < 0) err = err2; } @@ -1790,7 +1791,7 @@ asmlinkage long sys_sendmsg(int fd, struct msghdr __user *msg, unsigned flags) struct compat_msghdr __user *msg_compat = (struct compat_msghdr __user *)msg; struct socket *sock; - char address[MAX_SOCK_ADDR]; + struct sockaddr_storage address; struct iovec iovstack[UIO_FASTIOV], *iov = iovstack; unsigned char ctl[sizeof(struct cmsghdr) + 20] __attribute__ ((aligned(sizeof(__kernel_size_t)))); @@ -1828,9 +1829,13 @@ asmlinkage long sys_sendmsg(int fd, struct msghdr __user *msg, unsigned flags) /* This will also move the address data into kernel space */ if (MSG_CMSG_COMPAT & flags) { - err = verify_compat_iovec(&msg_sys, iov, address, VERIFY_READ); + err = verify_compat_iovec(&msg_sys, iov, + (struct sockaddr *)&address, + VERIFY_READ); } else - err = verify_iovec(&msg_sys, iov, address, VERIFY_READ); + err = verify_iovec(&msg_sys, iov, + (struct sockaddr *)&address, + VERIFY_READ); if (err < 0) goto out_freeiov; total_len = err; @@ -1901,7 +1906,7 @@ asmlinkage long sys_recvmsg(int fd, struct msghdr __user *msg, int fput_needed; /* kernel mode address */ - char addr[MAX_SOCK_ADDR]; + struct sockaddr_storage addr; /* user mode address pointers */ struct sockaddr __user *uaddr; @@ -1939,9 +1944,13 @@ asmlinkage long sys_recvmsg(int fd, struct msghdr __user *msg, uaddr = (__force void __user *)msg_sys.msg_name; uaddr_len = COMPAT_NAMELEN(msg); if (MSG_CMSG_COMPAT & flags) { - err = verify_compat_iovec(&msg_sys, iov, addr, VERIFY_WRITE); + err = verify_compat_iovec(&msg_sys, iov, + (struct sockaddr *)&addr, + VERIFY_WRITE); } else - err = verify_iovec(&msg_sys, iov, addr, VERIFY_WRITE); + err = verify_iovec(&msg_sys, iov, + (struct sockaddr *)&addr, + VERIFY_WRITE); if (err < 0) goto out_freeiov; total_len = err; @@ -1957,7 +1966,8 @@ asmlinkage long sys_recvmsg(int fd, struct msghdr __user *msg, len = err; if (uaddr != NULL) { - err = move_addr_to_user(addr, msg_sys.msg_namelen, uaddr, + err = move_addr_to_user((struct sockaddr *)&addr, + msg_sys.msg_namelen, uaddr, uaddr_len); if (err < 0) goto out_freeiov; -- cgit v1.2.2 From a6ffb404dc03f806a257faaab831a6cb55c0b790 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Sat, 19 Jul 2008 22:36:07 -0700 Subject: ipv6 mcast: Omit redundant address family checks in ip6_mc_source(). The caller has alredy checked for them. Signed-off-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller --- net/ipv6/mcast.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'net') diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c index 1b285371124a..e7c03bcc2788 100644 --- a/net/ipv6/mcast.c +++ b/net/ipv6/mcast.c @@ -367,10 +367,6 @@ int ip6_mc_source(int add, int omode, struct sock *sk, int pmclocked = 0; int err; - if (pgsr->gsr_group.ss_family != AF_INET6 || - pgsr->gsr_source.ss_family != AF_INET6) - return -EINVAL; - source = &((struct sockaddr_in6 *)&pgsr->gsr_source)->sin6_addr; group = &((struct sockaddr_in6 *)&pgsr->gsr_group)->sin6_addr; -- cgit v1.2.2 From 5f86173bdf15981ca49d0434f638b68f70a35644 Mon Sep 17 00:00:00 2001 From: Jussi Kivilinna Date: Sun, 20 Jul 2008 00:08:04 -0700 Subject: net_sched: Add qdisc_enqueue wrapper Signed-off-by: Jussi Kivilinna Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++-- net/mac80211/wme.c | 2 +- net/sched/sch_atm.c | 2 +- net/sched/sch_cbq.c | 5 +++-- net/sched/sch_dsmark.c | 2 +- net/sched/sch_hfsc.c | 2 +- net/sched/sch_htb.c | 3 +-- net/sched/sch_netem.c | 20 ++++++++++++-------- net/sched/sch_prio.c | 3 ++- net/sched/sch_red.c | 2 +- net/sched/sch_tbf.c | 3 ++- 11 files changed, 27 insertions(+), 21 deletions(-) (limited to 'net') diff --git a/net/core/dev.c b/net/core/dev.c index 065b9817e209..2eed17bcb2dd 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1781,7 +1781,7 @@ gso: spin_lock(root_lock); - rc = q->enqueue(skb, q); + rc = qdisc_enqueue_root(skb, q); qdisc_run(q); spin_unlock(root_lock); @@ -2083,7 +2083,7 @@ static int ing_filter(struct sk_buff *skb) q = rxq->qdisc; if (q) { spin_lock(qdisc_lock(q)); - result = q->enqueue(skb, q); + result = qdisc_enqueue_root(skb, q); spin_unlock(qdisc_lock(q)); } diff --git a/net/mac80211/wme.c b/net/mac80211/wme.c index 6e8099e77043..07edda0b8a5c 100644 --- a/net/mac80211/wme.c +++ b/net/mac80211/wme.c @@ -289,7 +289,7 @@ void ieee80211_requeue(struct ieee80211_local *local, int queue) root_lock = qdisc_root_lock(qdisc); spin_lock(root_lock); - qdisc->enqueue(skb, qdisc); + qdisc_enqueue_root(skb, qdisc); spin_unlock(root_lock); } diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c index 0de757e3be4a..68ed35e2a763 100644 --- a/net/sched/sch_atm.c +++ b/net/sched/sch_atm.c @@ -429,7 +429,7 @@ static int atm_tc_enqueue(struct sk_buff *skb, struct Qdisc *sch) #endif } - ret = flow->q->enqueue(skb, flow->q); + ret = qdisc_enqueue(skb, flow->q); if (ret != 0) { drop: __maybe_unused sch->qstats.drops++; diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index a3953bbe2d79..1afe3eece627 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -387,7 +387,8 @@ cbq_enqueue(struct sk_buff *skb, struct Qdisc *sch) #ifdef CONFIG_NET_CLS_ACT cl->q->__parent = sch; #endif - if ((ret = cl->q->enqueue(skb, cl->q)) == NET_XMIT_SUCCESS) { + ret = qdisc_enqueue(skb, cl->q); + if (ret == NET_XMIT_SUCCESS) { sch->q.qlen++; sch->bstats.packets++; sch->bstats.bytes+=len; @@ -671,7 +672,7 @@ static int cbq_reshape_fail(struct sk_buff *skb, struct Qdisc *child) q->rx_class = cl; cl->q->__parent = sch; - if (cl->q->enqueue(skb, cl->q) == 0) { + if (qdisc_enqueue(skb, cl->q) == 0) { sch->q.qlen++; sch->bstats.packets++; sch->bstats.bytes+=len; diff --git a/net/sched/sch_dsmark.c b/net/sched/sch_dsmark.c index 3aafbd17393a..44d347e831cf 100644 --- a/net/sched/sch_dsmark.c +++ b/net/sched/sch_dsmark.c @@ -252,7 +252,7 @@ static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch) } } - err = p->q->enqueue(skb, p->q); + err = qdisc_enqueue(skb, p->q); if (err != NET_XMIT_SUCCESS) { sch->qstats.drops++; return err; diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index 5090708ba384..fd61ed6ee1e7 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -1586,7 +1586,7 @@ hfsc_enqueue(struct sk_buff *skb, struct Qdisc *sch) } len = skb->len; - err = cl->qdisc->enqueue(skb, cl->qdisc); + err = qdisc_enqueue(skb, cl->qdisc); if (unlikely(err != NET_XMIT_SUCCESS)) { cl->qstats.drops++; sch->qstats.drops++; diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index ee48457eaa4a..72b5a946178f 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -572,8 +572,7 @@ static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch) kfree_skb(skb); return ret; #endif - } else if (cl->un.leaf.q->enqueue(skb, cl->un.leaf.q) != - NET_XMIT_SUCCESS) { + } else if (qdisc_enqueue(skb, cl->un.leaf.q) != NET_XMIT_SUCCESS) { sch->qstats.drops++; cl->qstats.drops++; return NET_XMIT_DROP; diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index c5ea40c9eb21..13c4821e42b8 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -82,6 +82,12 @@ struct netem_skb_cb { psched_time_t time_to_send; }; +static inline struct netem_skb_cb *netem_skb_cb(struct sk_buff *skb) +{ + BUILD_BUG_ON(sizeof(skb->cb) < sizeof(struct netem_skb_cb)); + return (struct netem_skb_cb *)skb->cb; +} + /* init_crandom - initialize correlated random number generator * Use entropy source for initial seed. */ @@ -184,7 +190,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) u32 dupsave = q->duplicate; /* prevent duplicating a dup... */ q->duplicate = 0; - rootq->enqueue(skb2, rootq); + qdisc_enqueue_root(skb2, rootq); q->duplicate = dupsave; } @@ -205,7 +211,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) skb->data[net_random() % skb_headlen(skb)] ^= 1<<(net_random() % 8); } - cb = (struct netem_skb_cb *)skb->cb; + cb = netem_skb_cb(skb); if (q->gap == 0 /* not doing reordering */ || q->counter < q->gap /* inside last reordering gap */ || q->reorder < get_crandom(&q->reorder_cor)) { @@ -218,7 +224,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) now = psched_get_time(); cb->time_to_send = now + delay; ++q->counter; - ret = q->qdisc->enqueue(skb, q->qdisc); + ret = qdisc_enqueue(skb, q->qdisc); } else { /* * Do re-ordering by putting one out of N packets at the front @@ -277,8 +283,7 @@ static struct sk_buff *netem_dequeue(struct Qdisc *sch) skb = q->qdisc->dequeue(q->qdisc); if (skb) { - const struct netem_skb_cb *cb - = (const struct netem_skb_cb *)skb->cb; + const struct netem_skb_cb *cb = netem_skb_cb(skb); psched_time_t now = psched_get_time(); /* if more time remaining? */ @@ -457,7 +462,7 @@ static int tfifo_enqueue(struct sk_buff *nskb, struct Qdisc *sch) { struct fifo_sched_data *q = qdisc_priv(sch); struct sk_buff_head *list = &sch->q; - psched_time_t tnext = ((struct netem_skb_cb *)nskb->cb)->time_to_send; + psched_time_t tnext = netem_skb_cb(nskb)->time_to_send; struct sk_buff *skb; if (likely(skb_queue_len(list) < q->limit)) { @@ -468,8 +473,7 @@ static int tfifo_enqueue(struct sk_buff *nskb, struct Qdisc *sch) } skb_queue_reverse_walk(list, skb) { - const struct netem_skb_cb *cb - = (const struct netem_skb_cb *)skb->cb; + const struct netem_skb_cb *cb = netem_skb_cb(skb); if (tnext >= cb->time_to_send) break; diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c index 536ca474dc69..d29c2f87fc0b 100644 --- a/net/sched/sch_prio.c +++ b/net/sched/sch_prio.c @@ -81,7 +81,8 @@ prio_enqueue(struct sk_buff *skb, struct Qdisc *sch) } #endif - if ((ret = qdisc->enqueue(skb, qdisc)) == NET_XMIT_SUCCESS) { + ret = qdisc_enqueue(skb, qdisc); + if (ret == NET_XMIT_SUCCESS) { sch->bstats.bytes += skb->len; sch->bstats.packets++; sch->q.qlen++; diff --git a/net/sched/sch_red.c b/net/sched/sch_red.c index 77098acf0adc..b48a391bc129 100644 --- a/net/sched/sch_red.c +++ b/net/sched/sch_red.c @@ -92,7 +92,7 @@ static int red_enqueue(struct sk_buff *skb, struct Qdisc* sch) break; } - ret = child->enqueue(skb, child); + ret = qdisc_enqueue(skb, child); if (likely(ret == NET_XMIT_SUCCESS)) { sch->bstats.bytes += skb->len; sch->bstats.packets++; diff --git a/net/sched/sch_tbf.c b/net/sched/sch_tbf.c index 444c227fcb6b..7d705b86dae5 100644 --- a/net/sched/sch_tbf.c +++ b/net/sched/sch_tbf.c @@ -133,7 +133,8 @@ static int tbf_enqueue(struct sk_buff *skb, struct Qdisc* sch) return NET_XMIT_DROP; } - if ((ret = q->qdisc->enqueue(skb, q->qdisc)) != 0) { + ret = qdisc_enqueue(skb, q->qdisc); + if (ret != 0) { sch->qstats.drops++; return ret; } -- cgit v1.2.2 From 0abf77e55a2459aa9905be4b226e4729d5b4f0cb Mon Sep 17 00:00:00 2001 From: Jussi Kivilinna Date: Sun, 20 Jul 2008 00:08:27 -0700 Subject: net_sched: Add accessor function for packet length for qdiscs Signed-off-by: Jussi Kivilinna Signed-off-by: David S. Miller --- net/sched/act_gact.c | 2 +- net/sched/act_ipt.c | 2 +- net/sched/act_mirred.c | 4 ++-- net/sched/act_nat.c | 2 +- net/sched/act_pedit.c | 2 +- net/sched/act_police.c | 8 ++++---- net/sched/act_simple.c | 2 +- net/sched/sch_atm.c | 4 ++-- net/sched/sch_cbq.c | 14 ++++++-------- net/sched/sch_dsmark.c | 2 +- net/sched/sch_fifo.c | 2 +- net/sched/sch_gred.c | 12 ++++++------ net/sched/sch_hfsc.c | 14 ++++++-------- net/sched/sch_htb.c | 9 +++++---- net/sched/sch_ingress.c | 2 +- net/sched/sch_netem.c | 6 +++--- net/sched/sch_prio.c | 2 +- net/sched/sch_red.c | 2 +- net/sched/sch_sfq.c | 16 ++++++++-------- net/sched/sch_tbf.c | 6 +++--- net/sched/sch_teql.c | 6 +++--- 21 files changed, 58 insertions(+), 61 deletions(-) (limited to 'net') diff --git a/net/sched/act_gact.c b/net/sched/act_gact.c index 422872c4f14b..ac04289da5d7 100644 --- a/net/sched/act_gact.c +++ b/net/sched/act_gact.c @@ -139,7 +139,7 @@ static int tcf_gact(struct sk_buff *skb, struct tc_action *a, struct tcf_result #else action = gact->tcf_action; #endif - gact->tcf_bstats.bytes += skb->len; + gact->tcf_bstats.bytes += qdisc_pkt_len(skb); gact->tcf_bstats.packets++; if (action == TC_ACT_SHOT) gact->tcf_qstats.drops++; diff --git a/net/sched/act_ipt.c b/net/sched/act_ipt.c index da696fd3e341..d1263b3c96c3 100644 --- a/net/sched/act_ipt.c +++ b/net/sched/act_ipt.c @@ -205,7 +205,7 @@ static int tcf_ipt(struct sk_buff *skb, struct tc_action *a, spin_lock(&ipt->tcf_lock); ipt->tcf_tm.lastuse = jiffies; - ipt->tcf_bstats.bytes += skb->len; + ipt->tcf_bstats.bytes += qdisc_pkt_len(skb); ipt->tcf_bstats.packets++; /* yes, we have to worry about both in and out dev diff --git a/net/sched/act_mirred.c b/net/sched/act_mirred.c index 1aff005d95cd..70341c020b6d 100644 --- a/net/sched/act_mirred.c +++ b/net/sched/act_mirred.c @@ -164,7 +164,7 @@ bad_mirred: if (skb2 != NULL) kfree_skb(skb2); m->tcf_qstats.overlimits++; - m->tcf_bstats.bytes += skb->len; + m->tcf_bstats.bytes += qdisc_pkt_len(skb); m->tcf_bstats.packets++; spin_unlock(&m->tcf_lock); /* should we be asking for packet to be dropped? @@ -184,7 +184,7 @@ bad_mirred: goto bad_mirred; } - m->tcf_bstats.bytes += skb2->len; + m->tcf_bstats.bytes += qdisc_pkt_len(skb2); m->tcf_bstats.packets++; if (!(at & AT_EGRESS)) if (m->tcfm_ok_push) diff --git a/net/sched/act_nat.c b/net/sched/act_nat.c index 0a3c8339767a..7b39ed485bca 100644 --- a/net/sched/act_nat.c +++ b/net/sched/act_nat.c @@ -124,7 +124,7 @@ static int tcf_nat(struct sk_buff *skb, struct tc_action *a, egress = p->flags & TCA_NAT_FLAG_EGRESS; action = p->tcf_action; - p->tcf_bstats.bytes += skb->len; + p->tcf_bstats.bytes += qdisc_pkt_len(skb); p->tcf_bstats.packets++; spin_unlock(&p->tcf_lock); diff --git a/net/sched/act_pedit.c b/net/sched/act_pedit.c index 3cc4cb9e500e..d5f4e3404864 100644 --- a/net/sched/act_pedit.c +++ b/net/sched/act_pedit.c @@ -182,7 +182,7 @@ static int tcf_pedit(struct sk_buff *skb, struct tc_action *a, bad: p->tcf_qstats.overlimits++; done: - p->tcf_bstats.bytes += skb->len; + p->tcf_bstats.bytes += qdisc_pkt_len(skb); p->tcf_bstats.packets++; spin_unlock(&p->tcf_lock); return p->tcf_action; diff --git a/net/sched/act_police.c b/net/sched/act_police.c index 0898120bbcc0..32c3f9d9fb7a 100644 --- a/net/sched/act_police.c +++ b/net/sched/act_police.c @@ -272,7 +272,7 @@ static int tcf_act_police(struct sk_buff *skb, struct tc_action *a, spin_lock(&police->tcf_lock); - police->tcf_bstats.bytes += skb->len; + police->tcf_bstats.bytes += qdisc_pkt_len(skb); police->tcf_bstats.packets++; if (police->tcfp_ewma_rate && @@ -282,7 +282,7 @@ static int tcf_act_police(struct sk_buff *skb, struct tc_action *a, return police->tcf_action; } - if (skb->len <= police->tcfp_mtu) { + if (qdisc_pkt_len(skb) <= police->tcfp_mtu) { if (police->tcfp_R_tab == NULL) { spin_unlock(&police->tcf_lock); return police->tcfp_result; @@ -295,12 +295,12 @@ static int tcf_act_police(struct sk_buff *skb, struct tc_action *a, ptoks = toks + police->tcfp_ptoks; if (ptoks > (long)L2T_P(police, police->tcfp_mtu)) ptoks = (long)L2T_P(police, police->tcfp_mtu); - ptoks -= L2T_P(police, skb->len); + ptoks -= L2T_P(police, qdisc_pkt_len(skb)); } toks += police->tcfp_toks; if (toks > (long)police->tcfp_burst) toks = police->tcfp_burst; - toks -= L2T(police, skb->len); + toks -= L2T(police, qdisc_pkt_len(skb)); if ((toks|ptoks) >= 0) { police->tcfp_t_c = now; police->tcfp_toks = toks; diff --git a/net/sched/act_simple.c b/net/sched/act_simple.c index 1d421d059caf..e7851ce92cfe 100644 --- a/net/sched/act_simple.c +++ b/net/sched/act_simple.c @@ -41,7 +41,7 @@ static int tcf_simp(struct sk_buff *skb, struct tc_action *a, struct tcf_result spin_lock(&d->tcf_lock); d->tcf_tm.lastuse = jiffies; - d->tcf_bstats.bytes += skb->len; + d->tcf_bstats.bytes += qdisc_pkt_len(skb); d->tcf_bstats.packets++; /* print policy string followed by _ then packet count diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c index 68ed35e2a763..04faa835be17 100644 --- a/net/sched/sch_atm.c +++ b/net/sched/sch_atm.c @@ -437,9 +437,9 @@ drop: __maybe_unused flow->qstats.drops++; return ret; } - sch->bstats.bytes += skb->len; + sch->bstats.bytes += qdisc_pkt_len(skb); sch->bstats.packets++; - flow->bstats.bytes += skb->len; + flow->bstats.bytes += qdisc_pkt_len(skb); flow->bstats.packets++; /* * Okay, this may seem weird. We pretend we've dropped the packet if diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c index 1afe3eece627..f1d2f8ec8b4c 100644 --- a/net/sched/sch_cbq.c +++ b/net/sched/sch_cbq.c @@ -370,7 +370,6 @@ static int cbq_enqueue(struct sk_buff *skb, struct Qdisc *sch) { struct cbq_sched_data *q = qdisc_priv(sch); - int len = skb->len; int uninitialized_var(ret); struct cbq_class *cl = cbq_classify(skb, sch, &ret); @@ -391,7 +390,7 @@ cbq_enqueue(struct sk_buff *skb, struct Qdisc *sch) if (ret == NET_XMIT_SUCCESS) { sch->q.qlen++; sch->bstats.packets++; - sch->bstats.bytes+=len; + sch->bstats.bytes += qdisc_pkt_len(skb); cbq_mark_toplevel(q, cl); if (!cl->next_alive) cbq_activate_class(cl); @@ -658,7 +657,6 @@ static enum hrtimer_restart cbq_undelay(struct hrtimer *timer) #ifdef CONFIG_NET_CLS_ACT static int cbq_reshape_fail(struct sk_buff *skb, struct Qdisc *child) { - int len = skb->len; struct Qdisc *sch = child->__parent; struct cbq_sched_data *q = qdisc_priv(sch); struct cbq_class *cl = q->rx_class; @@ -675,7 +673,7 @@ static int cbq_reshape_fail(struct sk_buff *skb, struct Qdisc *child) if (qdisc_enqueue(skb, cl->q) == 0) { sch->q.qlen++; sch->bstats.packets++; - sch->bstats.bytes+=len; + sch->bstats.bytes += qdisc_pkt_len(skb); if (!cl->next_alive) cbq_activate_class(cl); return 0; @@ -881,7 +879,7 @@ cbq_dequeue_prio(struct Qdisc *sch, int prio) if (skb == NULL) goto skip_class; - cl->deficit -= skb->len; + cl->deficit -= qdisc_pkt_len(skb); q->tx_class = cl; q->tx_borrowed = borrow; if (borrow != cl) { @@ -889,11 +887,11 @@ cbq_dequeue_prio(struct Qdisc *sch, int prio) borrow->xstats.borrows++; cl->xstats.borrows++; #else - borrow->xstats.borrows += skb->len; - cl->xstats.borrows += skb->len; + borrow->xstats.borrows += qdisc_pkt_len(skb); + cl->xstats.borrows += qdisc_pkt_len(skb); #endif } - q->tx_len = skb->len; + q->tx_len = qdisc_pkt_len(skb); if (cl->deficit <= 0) { q->active[prio] = cl; diff --git a/net/sched/sch_dsmark.c b/net/sched/sch_dsmark.c index 44d347e831cf..a935676987e2 100644 --- a/net/sched/sch_dsmark.c +++ b/net/sched/sch_dsmark.c @@ -258,7 +258,7 @@ static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch) return err; } - sch->bstats.bytes += skb->len; + sch->bstats.bytes += qdisc_pkt_len(skb); sch->bstats.packets++; sch->q.qlen++; diff --git a/net/sched/sch_fifo.c b/net/sched/sch_fifo.c index 1d97fa42c902..23d258bfe8ac 100644 --- a/net/sched/sch_fifo.c +++ b/net/sched/sch_fifo.c @@ -27,7 +27,7 @@ static int bfifo_enqueue(struct sk_buff *skb, struct Qdisc* sch) { struct fifo_sched_data *q = qdisc_priv(sch); - if (likely(sch->qstats.backlog + skb->len <= q->limit)) + if (likely(sch->qstats.backlog + qdisc_pkt_len(skb) <= q->limit)) return qdisc_enqueue_tail(skb, sch); return qdisc_reshape_fail(skb, sch); diff --git a/net/sched/sch_gred.c b/net/sched/sch_gred.c index 39fa28511f07..c1ad6b8de105 100644 --- a/net/sched/sch_gred.c +++ b/net/sched/sch_gred.c @@ -188,7 +188,7 @@ static int gred_enqueue(struct sk_buff *skb, struct Qdisc* sch) } q->packetsin++; - q->bytesin += skb->len; + q->bytesin += qdisc_pkt_len(skb); if (gred_wred_mode(t)) gred_load_wred_set(t, q); @@ -226,8 +226,8 @@ static int gred_enqueue(struct sk_buff *skb, struct Qdisc* sch) break; } - if (q->backlog + skb->len <= q->limit) { - q->backlog += skb->len; + if (q->backlog + qdisc_pkt_len(skb) <= q->limit) { + q->backlog += qdisc_pkt_len(skb); return qdisc_enqueue_tail(skb, sch); } @@ -254,7 +254,7 @@ static int gred_requeue(struct sk_buff *skb, struct Qdisc* sch) } else { if (red_is_idling(&q->parms)) red_end_of_idle_period(&q->parms); - q->backlog += skb->len; + q->backlog += qdisc_pkt_len(skb); } return qdisc_requeue(skb, sch); @@ -277,7 +277,7 @@ static struct sk_buff *gred_dequeue(struct Qdisc* sch) "VQ 0x%x after dequeue, screwing up " "backlog.\n", tc_index_to_dp(skb)); } else { - q->backlog -= skb->len; + q->backlog -= qdisc_pkt_len(skb); if (!q->backlog && !gred_wred_mode(t)) red_start_of_idle_period(&q->parms); @@ -299,7 +299,7 @@ static unsigned int gred_drop(struct Qdisc* sch) skb = qdisc_dequeue_tail(sch); if (skb) { - unsigned int len = skb->len; + unsigned int len = qdisc_pkt_len(skb); struct gred_sched_data *q; u16 dp = tc_index_to_dp(skb); diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c index fd61ed6ee1e7..0ae7d19dcba8 100644 --- a/net/sched/sch_hfsc.c +++ b/net/sched/sch_hfsc.c @@ -895,7 +895,7 @@ qdisc_peek_len(struct Qdisc *sch) printk("qdisc_peek_len: non work-conserving qdisc ?\n"); return 0; } - len = skb->len; + len = qdisc_pkt_len(skb); if (unlikely(sch->ops->requeue(skb, sch) != NET_XMIT_SUCCESS)) { if (net_ratelimit()) printk("qdisc_peek_len: failed to requeue\n"); @@ -1574,7 +1574,6 @@ static int hfsc_enqueue(struct sk_buff *skb, struct Qdisc *sch) { struct hfsc_class *cl; - unsigned int len; int err; cl = hfsc_classify(skb, sch, &err); @@ -1585,7 +1584,6 @@ hfsc_enqueue(struct sk_buff *skb, struct Qdisc *sch) return err; } - len = skb->len; err = qdisc_enqueue(skb, cl->qdisc); if (unlikely(err != NET_XMIT_SUCCESS)) { cl->qstats.drops++; @@ -1594,12 +1592,12 @@ hfsc_enqueue(struct sk_buff *skb, struct Qdisc *sch) } if (cl->qdisc->q.qlen == 1) - set_active(cl, len); + set_active(cl, qdisc_pkt_len(skb)); cl->bstats.packets++; - cl->bstats.bytes += len; + cl->bstats.bytes += qdisc_pkt_len(skb); sch->bstats.packets++; - sch->bstats.bytes += len; + sch->bstats.bytes += qdisc_pkt_len(skb); sch->q.qlen++; return NET_XMIT_SUCCESS; @@ -1649,9 +1647,9 @@ hfsc_dequeue(struct Qdisc *sch) return NULL; } - update_vf(cl, skb->len, cur_time); + update_vf(cl, qdisc_pkt_len(skb), cur_time); if (realtime) - cl->cl_cumul += skb->len; + cl->cl_cumul += qdisc_pkt_len(skb); if (cl->qdisc->q.qlen != 0) { if (cl->cl_flags & HFSC_RSC) { diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 72b5a946178f..30c999c61b01 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -579,13 +579,13 @@ static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch) } else { cl->bstats.packets += skb_is_gso(skb)?skb_shinfo(skb)->gso_segs:1; - cl->bstats.bytes += skb->len; + cl->bstats.bytes += qdisc_pkt_len(skb); htb_activate(q, cl); } sch->q.qlen++; sch->bstats.packets += skb_is_gso(skb)?skb_shinfo(skb)->gso_segs:1; - sch->bstats.bytes += skb->len; + sch->bstats.bytes += qdisc_pkt_len(skb); return NET_XMIT_SUCCESS; } @@ -642,7 +642,7 @@ static int htb_requeue(struct sk_buff *skb, struct Qdisc *sch) static void htb_charge_class(struct htb_sched *q, struct htb_class *cl, int level, struct sk_buff *skb) { - int bytes = skb->len; + int bytes = qdisc_pkt_len(skb); long toks, diff; enum htb_cmode old_mode; @@ -855,7 +855,8 @@ next: } while (cl != start); if (likely(skb != NULL)) { - if ((cl->un.leaf.deficit[level] -= skb->len) < 0) { + cl->un.leaf.deficit[level] -= qdisc_pkt_len(skb); + if (cl->un.leaf.deficit[level] < 0) { cl->un.leaf.deficit[level] += cl->un.leaf.quantum; htb_next_rb_node((level ? cl->parent->un.inner.ptr : q-> ptr[0]) + prio); diff --git a/net/sched/sch_ingress.c b/net/sched/sch_ingress.c index 956c80ad5965..4a2b77374358 100644 --- a/net/sched/sch_ingress.c +++ b/net/sched/sch_ingress.c @@ -77,7 +77,7 @@ static int ingress_enqueue(struct sk_buff *skb, struct Qdisc *sch) result = tc_classify(skb, p->filter_list, &res); sch->bstats.packets++; - sch->bstats.bytes += skb->len; + sch->bstats.bytes += qdisc_pkt_len(skb); switch (result) { case TC_ACT_SHOT: result = TC_ACT_SHOT; diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index 13c4821e42b8..ae49be00022f 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -237,7 +237,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch) if (likely(ret == NET_XMIT_SUCCESS)) { sch->q.qlen++; - sch->bstats.bytes += skb->len; + sch->bstats.bytes += qdisc_pkt_len(skb); sch->bstats.packets++; } else sch->qstats.drops++; @@ -481,8 +481,8 @@ static int tfifo_enqueue(struct sk_buff *nskb, struct Qdisc *sch) __skb_queue_after(list, skb, nskb); - sch->qstats.backlog += nskb->len; - sch->bstats.bytes += nskb->len; + sch->qstats.backlog += qdisc_pkt_len(nskb); + sch->bstats.bytes += qdisc_pkt_len(nskb); sch->bstats.packets++; return NET_XMIT_SUCCESS; diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c index d29c2f87fc0b..f849243eb095 100644 --- a/net/sched/sch_prio.c +++ b/net/sched/sch_prio.c @@ -83,7 +83,7 @@ prio_enqueue(struct sk_buff *skb, struct Qdisc *sch) ret = qdisc_enqueue(skb, qdisc); if (ret == NET_XMIT_SUCCESS) { - sch->bstats.bytes += skb->len; + sch->bstats.bytes += qdisc_pkt_len(skb); sch->bstats.packets++; sch->q.qlen++; return NET_XMIT_SUCCESS; diff --git a/net/sched/sch_red.c b/net/sched/sch_red.c index b48a391bc129..3f2d1d7f3bbd 100644 --- a/net/sched/sch_red.c +++ b/net/sched/sch_red.c @@ -94,7 +94,7 @@ static int red_enqueue(struct sk_buff *skb, struct Qdisc* sch) ret = qdisc_enqueue(skb, child); if (likely(ret == NET_XMIT_SUCCESS)) { - sch->bstats.bytes += skb->len; + sch->bstats.bytes += qdisc_pkt_len(skb); sch->bstats.packets++; sch->q.qlen++; } else { diff --git a/net/sched/sch_sfq.c b/net/sched/sch_sfq.c index 8458f630fac4..8589da666568 100644 --- a/net/sched/sch_sfq.c +++ b/net/sched/sch_sfq.c @@ -245,7 +245,7 @@ static unsigned int sfq_drop(struct Qdisc *sch) if (d > 1) { sfq_index x = q->dep[d + SFQ_DEPTH].next; skb = q->qs[x].prev; - len = skb->len; + len = qdisc_pkt_len(skb); __skb_unlink(skb, &q->qs[x]); kfree_skb(skb); sfq_dec(q, x); @@ -261,7 +261,7 @@ static unsigned int sfq_drop(struct Qdisc *sch) q->next[q->tail] = q->next[d]; q->allot[q->next[d]] += q->quantum; skb = q->qs[d].prev; - len = skb->len; + len = qdisc_pkt_len(skb); __skb_unlink(skb, &q->qs[d]); kfree_skb(skb); sfq_dec(q, d); @@ -305,7 +305,7 @@ sfq_enqueue(struct sk_buff *skb, struct Qdisc *sch) if (q->qs[x].qlen >= q->limit) return qdisc_drop(skb, sch); - sch->qstats.backlog += skb->len; + sch->qstats.backlog += qdisc_pkt_len(skb); __skb_queue_tail(&q->qs[x], skb); sfq_inc(q, x); if (q->qs[x].qlen == 1) { /* The flow is new */ @@ -320,7 +320,7 @@ sfq_enqueue(struct sk_buff *skb, struct Qdisc *sch) } } if (++sch->q.qlen <= q->limit) { - sch->bstats.bytes += skb->len; + sch->bstats.bytes += qdisc_pkt_len(skb); sch->bstats.packets++; return 0; } @@ -352,7 +352,7 @@ sfq_requeue(struct sk_buff *skb, struct Qdisc *sch) q->hash[x] = hash; } - sch->qstats.backlog += skb->len; + sch->qstats.backlog += qdisc_pkt_len(skb); __skb_queue_head(&q->qs[x], skb); /* If selected queue has length q->limit+1, this means that * all another queues are empty and we do simple tail drop. @@ -363,7 +363,7 @@ sfq_requeue(struct sk_buff *skb, struct Qdisc *sch) skb = q->qs[x].prev; __skb_unlink(skb, &q->qs[x]); sch->qstats.drops++; - sch->qstats.backlog -= skb->len; + sch->qstats.backlog -= qdisc_pkt_len(skb); kfree_skb(skb); return NET_XMIT_CN; } @@ -411,7 +411,7 @@ sfq_dequeue(struct Qdisc *sch) skb = __skb_dequeue(&q->qs[a]); sfq_dec(q, a); sch->q.qlen--; - sch->qstats.backlog -= skb->len; + sch->qstats.backlog -= qdisc_pkt_len(skb); /* Is the slot empty? */ if (q->qs[a].qlen == 0) { @@ -423,7 +423,7 @@ sfq_dequeue(struct Qdisc *sch) } q->next[q->tail] = a; q->allot[a] += q->quantum; - } else if ((q->allot[a] -= skb->len) <= 0) { + } else if ((q->allot[a] -= qdisc_pkt_len(skb)) <= 0) { q->tail = a; a = q->next[a]; q->allot[a] += q->quantum; diff --git a/net/sched/sch_tbf.c b/net/sched/sch_tbf.c index 7d705b86dae5..b296672f7632 100644 --- a/net/sched/sch_tbf.c +++ b/net/sched/sch_tbf.c @@ -123,7 +123,7 @@ static int tbf_enqueue(struct sk_buff *skb, struct Qdisc* sch) struct tbf_sched_data *q = qdisc_priv(sch); int ret; - if (skb->len > q->max_size) { + if (qdisc_pkt_len(skb) > q->max_size) { sch->qstats.drops++; #ifdef CONFIG_NET_CLS_ACT if (sch->reshape_fail == NULL || sch->reshape_fail(skb, sch)) @@ -140,7 +140,7 @@ static int tbf_enqueue(struct sk_buff *skb, struct Qdisc* sch) } sch->q.qlen++; - sch->bstats.bytes += skb->len; + sch->bstats.bytes += qdisc_pkt_len(skb); sch->bstats.packets++; return 0; } @@ -181,7 +181,7 @@ static struct sk_buff *tbf_dequeue(struct Qdisc* sch) psched_time_t now; long toks; long ptoks = 0; - unsigned int len = skb->len; + unsigned int len = qdisc_pkt_len(skb); now = psched_get_time(); toks = psched_tdiff_bounded(now, q->t_c, q->buffer); diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index 8b0ff345f9da..537223642b6e 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -83,7 +83,7 @@ teql_enqueue(struct sk_buff *skb, struct Qdisc* sch) if (q->q.qlen < dev->tx_queue_len) { __skb_queue_tail(&q->q, skb); - sch->bstats.bytes += skb->len; + sch->bstats.bytes += qdisc_pkt_len(skb); sch->bstats.packets++; return 0; } @@ -278,7 +278,6 @@ static int teql_master_xmit(struct sk_buff *skb, struct net_device *dev) struct Qdisc *start, *q; int busy; int nores; - int len = skb->len; int subq = skb_get_queue_mapping(skb); struct sk_buff *skb_res = NULL; @@ -313,7 +312,8 @@ restart: master->slaves = NEXT_SLAVE(q); netif_wake_queue(dev); master->stats.tx_packets++; - master->stats.tx_bytes += len; + master->stats.tx_bytes += + qdisc_pkt_len(skb); return 0; } netif_tx_unlock(slave); -- cgit v1.2.2 From 175f9c1bba9b825d22b142d183c9e175488b260c Mon Sep 17 00:00:00 2001 From: Jussi Kivilinna Date: Sun, 20 Jul 2008 00:08:47 -0700 Subject: net_sched: Add size table for qdiscs Add size table functions for qdiscs and calculate packet size in qdisc_enqueue(). Based on patch by Patrick McHardy http://marc.info/?l=linux-netdev&m=115201979221729&w=2 Signed-off-by: Jussi Kivilinna Signed-off-by: David S. Miller --- net/sched/sch_api.c | 151 +++++++++++++++++++++++++++++++++++++++++++++++- net/sched/sch_generic.c | 1 + net/sched/sch_netem.c | 5 +- 3 files changed, 153 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index fb43731c9860..5219d5f9d754 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -286,6 +286,129 @@ void qdisc_put_rtab(struct qdisc_rate_table *tab) } EXPORT_SYMBOL(qdisc_put_rtab); +static LIST_HEAD(qdisc_stab_list); +static DEFINE_SPINLOCK(qdisc_stab_lock); + +static const struct nla_policy stab_policy[TCA_STAB_MAX + 1] = { + [TCA_STAB_BASE] = { .len = sizeof(struct tc_sizespec) }, + [TCA_STAB_DATA] = { .type = NLA_BINARY }, +}; + +static struct qdisc_size_table *qdisc_get_stab(struct nlattr *opt) +{ + struct nlattr *tb[TCA_STAB_MAX + 1]; + struct qdisc_size_table *stab; + struct tc_sizespec *s; + unsigned int tsize = 0; + u16 *tab = NULL; + int err; + + err = nla_parse_nested(tb, TCA_STAB_MAX, opt, stab_policy); + if (err < 0) + return ERR_PTR(err); + if (!tb[TCA_STAB_BASE]) + return ERR_PTR(-EINVAL); + + s = nla_data(tb[TCA_STAB_BASE]); + + if (s->tsize > 0) { + if (!tb[TCA_STAB_DATA]) + return ERR_PTR(-EINVAL); + tab = nla_data(tb[TCA_STAB_DATA]); + tsize = nla_len(tb[TCA_STAB_DATA]) / sizeof(u16); + } + + if (!s || tsize != s->tsize || (!tab && tsize > 0)) + return ERR_PTR(-EINVAL); + + spin_lock(&qdisc_stab_lock); + + list_for_each_entry(stab, &qdisc_stab_list, list) { + if (memcmp(&stab->szopts, s, sizeof(*s))) + continue; + if (tsize > 0 && memcmp(stab->data, tab, tsize * sizeof(u16))) + continue; + stab->refcnt++; + spin_unlock(&qdisc_stab_lock); + return stab; + } + + spin_unlock(&qdisc_stab_lock); + + stab = kmalloc(sizeof(*stab) + tsize * sizeof(u16), GFP_KERNEL); + if (!stab) + return ERR_PTR(-ENOMEM); + + stab->refcnt = 1; + stab->szopts = *s; + if (tsize > 0) + memcpy(stab->data, tab, tsize * sizeof(u16)); + + spin_lock(&qdisc_stab_lock); + list_add_tail(&stab->list, &qdisc_stab_list); + spin_unlock(&qdisc_stab_lock); + + return stab; +} + +void qdisc_put_stab(struct qdisc_size_table *tab) +{ + if (!tab) + return; + + spin_lock(&qdisc_stab_lock); + + if (--tab->refcnt == 0) { + list_del(&tab->list); + kfree(tab); + } + + spin_unlock(&qdisc_stab_lock); +} +EXPORT_SYMBOL(qdisc_put_stab); + +static int qdisc_dump_stab(struct sk_buff *skb, struct qdisc_size_table *stab) +{ + struct nlattr *nest; + + nest = nla_nest_start(skb, TCA_STAB); + NLA_PUT(skb, TCA_STAB_BASE, sizeof(stab->szopts), &stab->szopts); + nla_nest_end(skb, nest); + + return skb->len; + +nla_put_failure: + return -1; +} + +void qdisc_calculate_pkt_len(struct sk_buff *skb, struct qdisc_size_table *stab) +{ + int pkt_len, slot; + + pkt_len = skb->len + stab->szopts.overhead; + if (unlikely(!stab->szopts.tsize)) + goto out; + + slot = pkt_len + stab->szopts.cell_align; + if (unlikely(slot < 0)) + slot = 0; + + slot >>= stab->szopts.cell_log; + if (likely(slot < stab->szopts.tsize)) + pkt_len = stab->data[slot]; + else + pkt_len = stab->data[stab->szopts.tsize - 1] * + (slot / stab->szopts.tsize) + + stab->data[slot % stab->szopts.tsize]; + + pkt_len <<= stab->szopts.size_log; +out: + if (unlikely(pkt_len < 1)) + pkt_len = 1; + qdisc_skb_cb(skb)->pkt_len = pkt_len; +} +EXPORT_SYMBOL(qdisc_calculate_pkt_len); + static enum hrtimer_restart qdisc_watchdog(struct hrtimer *timer) { struct qdisc_watchdog *wd = container_of(timer, struct qdisc_watchdog, @@ -613,6 +736,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, struct nlattr *kind = tca[TCA_KIND]; struct Qdisc *sch; struct Qdisc_ops *ops; + struct qdisc_size_table *stab; ops = qdisc_lookup_ops(kind); #ifdef CONFIG_KMOD @@ -670,6 +794,14 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, sch->handle = handle; if (!ops->init || (err = ops->init(sch, tca[TCA_OPTIONS])) == 0) { + if (tca[TCA_STAB]) { + stab = qdisc_get_stab(tca[TCA_STAB]); + if (IS_ERR(stab)) { + err = PTR_ERR(stab); + goto err_out3; + } + sch->stab = stab; + } if (tca[TCA_RATE]) { err = gen_new_estimator(&sch->bstats, &sch->rate_est, qdisc_root_lock(sch), @@ -691,6 +823,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, return sch; } err_out3: + qdisc_put_stab(sch->stab); dev_put(dev); kfree((char *) sch - sch->padded); err_out2: @@ -702,15 +835,26 @@ err_out: static int qdisc_change(struct Qdisc *sch, struct nlattr **tca) { - if (tca[TCA_OPTIONS]) { - int err; + struct qdisc_size_table *stab = NULL; + int err = 0; + if (tca[TCA_OPTIONS]) { if (sch->ops->change == NULL) return -EINVAL; err = sch->ops->change(sch, tca[TCA_OPTIONS]); if (err) return err; } + + if (tca[TCA_STAB]) { + stab = qdisc_get_stab(tca[TCA_STAB]); + if (IS_ERR(stab)) + return PTR_ERR(stab); + } + + qdisc_put_stab(sch->stab); + sch->stab = stab; + if (tca[TCA_RATE]) gen_replace_estimator(&sch->bstats, &sch->rate_est, qdisc_root_lock(sch), tca[TCA_RATE]); @@ -994,6 +1138,9 @@ static int tc_fill_qdisc(struct sk_buff *skb, struct Qdisc *q, u32 clid, goto nla_put_failure; q->qstats.qlen = q->q.qlen; + if (q->stab && qdisc_dump_stab(skb, q->stab) < 0) + goto nla_put_failure; + if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS, TCA_XSTATS, qdisc_root_lock(q), &d) < 0) goto nla_put_failure; diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 522a41a9f904..27a51f04db49 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -469,6 +469,7 @@ static void __qdisc_destroy(struct rcu_head *head) struct Qdisc *qdisc = container_of(head, struct Qdisc, q_rcu); const struct Qdisc_ops *ops = qdisc->ops; + qdisc_put_stab(qdisc->stab); gen_kill_estimator(&qdisc->bstats, &qdisc->rate_est); if (ops->reset) ops->reset(qdisc); diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c index ae49be00022f..a59085700678 100644 --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -84,8 +84,9 @@ struct netem_skb_cb { static inline struct netem_skb_cb *netem_skb_cb(struct sk_buff *skb) { - BUILD_BUG_ON(sizeof(skb->cb) < sizeof(struct netem_skb_cb)); - return (struct netem_skb_cb *)skb->cb; + BUILD_BUG_ON(sizeof(skb->cb) < + sizeof(struct qdisc_skb_cb) + sizeof(struct netem_skb_cb)); + return (struct netem_skb_cb *)qdisc_skb_cb(skb)->data; } /* init_crandom - initialize correlated random number generator -- cgit v1.2.2 From fb65a7c091529bfffb1262515252c0d0f6241c5c Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Sun, 20 Jul 2008 10:18:44 -0700 Subject: iucv: Fix bad merging. Noticed by Stephen Rothwell. Signed-off-by: David S. Miller --- net/iucv/iucv.c | 10 ---------- 1 file changed, 10 deletions(-) (limited to 'net') diff --git a/net/iucv/iucv.c b/net/iucv/iucv.c index a598c7384840..265b1b289a32 100644 --- a/net/iucv/iucv.c +++ b/net/iucv/iucv.c @@ -523,12 +523,8 @@ static int iucv_enable(void) rc = -EIO; get_online_cpus(); for_each_online_cpu(cpu) -<<<<<<< HEAD:net/iucv/iucv.c - smp_call_function_single(cpu, iucv_declare_cpu, NULL, 0, 1); -======= smp_call_function_single(cpu, iucv_declare_cpu, NULL, 1); preempt_enable(); ->>>>>>> 5b664cb235e97afbf34db9c4d77f08ebd725335e:net/iucv/iucv.c if (cpus_empty(iucv_buffer_cpumask)) /* No cpu could declare an iucv buffer. */ goto out_path; @@ -551,13 +547,7 @@ out: */ static void iucv_disable(void) { -<<<<<<< HEAD:net/iucv/iucv.c - get_online_cpus(); - on_each_cpu(iucv_retrieve_cpu, NULL, 0, 1); - put_online_cpus(); -======= on_each_cpu(iucv_retrieve_cpu, NULL, 1); ->>>>>>> 5b664cb235e97afbf34db9c4d77f08ebd725335e:net/iucv/iucv.c kfree(iucv_path_table); } -- cgit v1.2.2