aboutsummaryrefslogtreecommitdiffstats
path: root/net
diff options
context:
space:
mode:
authorJeff Garzik <jgarzik@pobox.com>2005-11-05 15:39:24 -0500
committerJeff Garzik <jgarzik@pobox.com>2005-11-05 15:39:24 -0500
commit7211bb9b64f17b23834d91fc3d0c1d78671ee9a8 (patch)
tree541909f86c31fee97cd70d28ec2fe5c23e1ceb02 /net
parentf1e691a24955ea987f021f378324fb5003b1b208 (diff)
parent70d9d825e0a5a78ec1dacaaaf5c72ff5b0206fab (diff)
Merge branch 'master'
Diffstat (limited to 'net')
-rw-r--r--net/bluetooth/af_bluetooth.c6
-rw-r--r--net/bluetooth/hci_core.c4
-rw-r--r--net/bluetooth/hci_sock.c12
-rw-r--r--net/bluetooth/hidp/Kconfig2
-rw-r--r--net/bluetooth/hidp/core.c13
-rw-r--r--net/bluetooth/rfcomm/Makefile2
-rw-r--r--net/bluetooth/rfcomm/core.c43
-rw-r--r--net/bluetooth/rfcomm/crc.c71
-rw-r--r--net/bridge/br_fdb.c12
-rw-r--r--net/bridge/br_input.c2
-rw-r--r--net/bridge/br_stp_if.c9
-rw-r--r--net/core/datagram.c4
-rw-r--r--net/core/dev.c14
-rw-r--r--net/core/ethtool.c53
-rw-r--r--net/core/skbuff.c75
-rw-r--r--net/dccp/dccp.h1
-rw-r--r--net/dccp/output.c34
-rw-r--r--net/ethernet/eth.c17
-rw-r--r--net/ipv4/fib_frontend.c2
-rw-r--r--net/ipv4/igmp.c5
-rw-r--r--net/ipv4/ip_output.c83
-rw-r--r--net/ipv4/netfilter/arp_tables.c201
-rw-r--r--net/ipv4/netfilter/ipt_addrtype.c2
-rw-r--r--net/ipv4/tcp_bic.c2
-rw-r--r--net/ipv6/addrconf.c12
-rw-r--r--net/ipv6/ip6_output.c71
-rw-r--r--net/ipv6/mcast.c25
-rw-r--r--net/ipv6/netfilter/ip6_tables.c298
-rw-r--r--net/ipv6/netfilter/ip6t_MARK.c8
-rw-r--r--net/ipv6/route.c2
-rw-r--r--net/rose/rose_timer.c1
-rw-r--r--net/sched/Kconfig394
-rw-r--r--net/sctp/sm_make_chunk.c2
-rw-r--r--net/sctp/socket.c90
-rw-r--r--net/sctp/ulpevent.c6
-rw-r--r--net/sunrpc/auth.c15
-rw-r--r--net/sunrpc/auth_gss/gss_krb5_crypto.c23
-rw-r--r--net/sunrpc/sunrpc_syms.c2
-rw-r--r--net/sunrpc/sysctl.c7
-rw-r--r--net/sunrpc/xprtsock.c9
40 files changed, 1025 insertions, 609 deletions
diff --git a/net/bluetooth/af_bluetooth.c b/net/bluetooth/af_bluetooth.c
index 12b43345b54f..03532062a46a 100644
--- a/net/bluetooth/af_bluetooth.c
+++ b/net/bluetooth/af_bluetooth.c
@@ -308,12 +308,6 @@ static struct net_proto_family bt_sock_family_ops = {
308 .create = bt_sock_create, 308 .create = bt_sock_create,
309}; 309};
310 310
311extern int hci_sock_init(void);
312extern int hci_sock_cleanup(void);
313
314extern int bt_sysfs_init(void);
315extern int bt_sysfs_cleanup(void);
316
317static int __init bt_init(void) 311static int __init bt_init(void)
318{ 312{
319 BT_INFO("Core ver %s", VERSION); 313 BT_INFO("Core ver %s", VERSION);
diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c
index 55dc42eac92c..cf0df1c8c933 100644
--- a/net/bluetooth/hci_core.c
+++ b/net/bluetooth/hci_core.c
@@ -87,7 +87,7 @@ int hci_unregister_notifier(struct notifier_block *nb)
87 return notifier_chain_unregister(&hci_notifier, nb); 87 return notifier_chain_unregister(&hci_notifier, nb);
88} 88}
89 89
90void hci_notify(struct hci_dev *hdev, int event) 90static void hci_notify(struct hci_dev *hdev, int event)
91{ 91{
92 notifier_call_chain(&hci_notifier, event, hdev); 92 notifier_call_chain(&hci_notifier, event, hdev);
93} 93}
@@ -1347,7 +1347,7 @@ static inline void hci_scodata_packet(struct hci_dev *hdev, struct sk_buff *skb)
1347 kfree_skb(skb); 1347 kfree_skb(skb);
1348} 1348}
1349 1349
1350void hci_rx_task(unsigned long arg) 1350static void hci_rx_task(unsigned long arg)
1351{ 1351{
1352 struct hci_dev *hdev = (struct hci_dev *) arg; 1352 struct hci_dev *hdev = (struct hci_dev *) arg;
1353 struct sk_buff *skb; 1353 struct sk_buff *skb;
diff --git a/net/bluetooth/hci_sock.c b/net/bluetooth/hci_sock.c
index 32ef7975a139..799e448750ad 100644
--- a/net/bluetooth/hci_sock.c
+++ b/net/bluetooth/hci_sock.c
@@ -66,20 +66,20 @@ static struct hci_sec_filter hci_sec_filter = {
66 /* Packet types */ 66 /* Packet types */
67 0x10, 67 0x10,
68 /* Events */ 68 /* Events */
69 { 0x1000d9fe, 0x0000300c }, 69 { 0x1000d9fe, 0x0000b00c },
70 /* Commands */ 70 /* Commands */
71 { 71 {
72 { 0x0 }, 72 { 0x0 },
73 /* OGF_LINK_CTL */ 73 /* OGF_LINK_CTL */
74 { 0xbe000006, 0x00000001, 0x0000, 0x00 }, 74 { 0xbe000006, 0x00000001, 0x000000, 0x00 },
75 /* OGF_LINK_POLICY */ 75 /* OGF_LINK_POLICY */
76 { 0x00005200, 0x00000000, 0x0000, 0x00 }, 76 { 0x00005200, 0x00000000, 0x000000, 0x00 },
77 /* OGF_HOST_CTL */ 77 /* OGF_HOST_CTL */
78 { 0xaab00200, 0x2b402aaa, 0x0154, 0x00 }, 78 { 0xaab00200, 0x2b402aaa, 0x020154, 0x00 },
79 /* OGF_INFO_PARAM */ 79 /* OGF_INFO_PARAM */
80 { 0x000002be, 0x00000000, 0x0000, 0x00 }, 80 { 0x000002be, 0x00000000, 0x000000, 0x00 },
81 /* OGF_STATUS_PARAM */ 81 /* OGF_STATUS_PARAM */
82 { 0x000000ea, 0x00000000, 0x0000, 0x00 } 82 { 0x000000ea, 0x00000000, 0x000000, 0x00 }
83 } 83 }
84}; 84};
85 85
diff --git a/net/bluetooth/hidp/Kconfig b/net/bluetooth/hidp/Kconfig
index 4e958f7d9418..edfea772fb67 100644
--- a/net/bluetooth/hidp/Kconfig
+++ b/net/bluetooth/hidp/Kconfig
@@ -1,6 +1,6 @@
1config BT_HIDP 1config BT_HIDP
2 tristate "HIDP protocol support" 2 tristate "HIDP protocol support"
3 depends on BT && BT_L2CAP 3 depends on BT && BT_L2CAP && (BROKEN || !S390)
4 select INPUT 4 select INPUT
5 help 5 help
6 HIDP (Human Interface Device Protocol) is a transport layer 6 HIDP (Human Interface Device Protocol) is a transport layer
diff --git a/net/bluetooth/hidp/core.c b/net/bluetooth/hidp/core.c
index de8af5f42394..860444a7fc0f 100644
--- a/net/bluetooth/hidp/core.c
+++ b/net/bluetooth/hidp/core.c
@@ -520,7 +520,7 @@ static int hidp_session(void *arg)
520 520
521 if (session->input) { 521 if (session->input) {
522 input_unregister_device(session->input); 522 input_unregister_device(session->input);
523 kfree(session->input); 523 session->input = NULL;
524 } 524 }
525 525
526 up_write(&hidp_session_sem); 526 up_write(&hidp_session_sem);
@@ -536,6 +536,8 @@ static inline void hidp_setup_input(struct hidp_session *session, struct hidp_co
536 536
537 input->private = session; 537 input->private = session;
538 538
539 input->name = "Bluetooth HID Boot Protocol Device";
540
539 input->id.bustype = BUS_BLUETOOTH; 541 input->id.bustype = BUS_BLUETOOTH;
540 input->id.vendor = req->vendor; 542 input->id.vendor = req->vendor;
541 input->id.product = req->product; 543 input->id.product = req->product;
@@ -582,16 +584,15 @@ int hidp_add_connection(struct hidp_connadd_req *req, struct socket *ctrl_sock,
582 return -ENOTUNIQ; 584 return -ENOTUNIQ;
583 585
584 session = kmalloc(sizeof(struct hidp_session), GFP_KERNEL); 586 session = kmalloc(sizeof(struct hidp_session), GFP_KERNEL);
585 if (!session) 587 if (!session)
586 return -ENOMEM; 588 return -ENOMEM;
587 memset(session, 0, sizeof(struct hidp_session)); 589 memset(session, 0, sizeof(struct hidp_session));
588 590
589 session->input = kmalloc(sizeof(struct input_dev), GFP_KERNEL); 591 session->input = input_allocate_device();
590 if (!session->input) { 592 if (!session->input) {
591 kfree(session); 593 kfree(session);
592 return -ENOMEM; 594 return -ENOMEM;
593 } 595 }
594 memset(session->input, 0, sizeof(struct input_dev));
595 596
596 down_write(&hidp_session_sem); 597 down_write(&hidp_session_sem);
597 598
@@ -651,8 +652,10 @@ unlink:
651 652
652 __hidp_unlink_session(session); 653 __hidp_unlink_session(session);
653 654
654 if (session->input) 655 if (session->input) {
655 input_unregister_device(session->input); 656 input_unregister_device(session->input);
657 session->input = NULL; /* don't try to free it here */
658 }
656 659
657failed: 660failed:
658 up_write(&hidp_session_sem); 661 up_write(&hidp_session_sem);
diff --git a/net/bluetooth/rfcomm/Makefile b/net/bluetooth/rfcomm/Makefile
index aecec45ec68d..fe07988a3705 100644
--- a/net/bluetooth/rfcomm/Makefile
+++ b/net/bluetooth/rfcomm/Makefile
@@ -4,5 +4,5 @@
4 4
5obj-$(CONFIG_BT_RFCOMM) += rfcomm.o 5obj-$(CONFIG_BT_RFCOMM) += rfcomm.o
6 6
7rfcomm-y := core.o sock.o crc.o 7rfcomm-y := core.o sock.o
8rfcomm-$(CONFIG_BT_RFCOMM_TTY) += tty.o 8rfcomm-$(CONFIG_BT_RFCOMM_TTY) += tty.o
diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c
index 35adce6482b6..c3d56ead840c 100644
--- a/net/bluetooth/rfcomm/core.c
+++ b/net/bluetooth/rfcomm/core.c
@@ -133,6 +133,49 @@ static inline void rfcomm_session_put(struct rfcomm_session *s)
133 133
134/* ---- RFCOMM FCS computation ---- */ 134/* ---- RFCOMM FCS computation ---- */
135 135
136/* reversed, 8-bit, poly=0x07 */
137static unsigned char rfcomm_crc_table[256] = {
138 0x00, 0x91, 0xe3, 0x72, 0x07, 0x96, 0xe4, 0x75,
139 0x0e, 0x9f, 0xed, 0x7c, 0x09, 0x98, 0xea, 0x7b,
140 0x1c, 0x8d, 0xff, 0x6e, 0x1b, 0x8a, 0xf8, 0x69,
141 0x12, 0x83, 0xf1, 0x60, 0x15, 0x84, 0xf6, 0x67,
142
143 0x38, 0xa9, 0xdb, 0x4a, 0x3f, 0xae, 0xdc, 0x4d,
144 0x36, 0xa7, 0xd5, 0x44, 0x31, 0xa0, 0xd2, 0x43,
145 0x24, 0xb5, 0xc7, 0x56, 0x23, 0xb2, 0xc0, 0x51,
146 0x2a, 0xbb, 0xc9, 0x58, 0x2d, 0xbc, 0xce, 0x5f,
147
148 0x70, 0xe1, 0x93, 0x02, 0x77, 0xe6, 0x94, 0x05,
149 0x7e, 0xef, 0x9d, 0x0c, 0x79, 0xe8, 0x9a, 0x0b,
150 0x6c, 0xfd, 0x8f, 0x1e, 0x6b, 0xfa, 0x88, 0x19,
151 0x62, 0xf3, 0x81, 0x10, 0x65, 0xf4, 0x86, 0x17,
152
153 0x48, 0xd9, 0xab, 0x3a, 0x4f, 0xde, 0xac, 0x3d,
154 0x46, 0xd7, 0xa5, 0x34, 0x41, 0xd0, 0xa2, 0x33,
155 0x54, 0xc5, 0xb7, 0x26, 0x53, 0xc2, 0xb0, 0x21,
156 0x5a, 0xcb, 0xb9, 0x28, 0x5d, 0xcc, 0xbe, 0x2f,
157
158 0xe0, 0x71, 0x03, 0x92, 0xe7, 0x76, 0x04, 0x95,
159 0xee, 0x7f, 0x0d, 0x9c, 0xe9, 0x78, 0x0a, 0x9b,
160 0xfc, 0x6d, 0x1f, 0x8e, 0xfb, 0x6a, 0x18, 0x89,
161 0xf2, 0x63, 0x11, 0x80, 0xf5, 0x64, 0x16, 0x87,
162
163 0xd8, 0x49, 0x3b, 0xaa, 0xdf, 0x4e, 0x3c, 0xad,
164 0xd6, 0x47, 0x35, 0xa4, 0xd1, 0x40, 0x32, 0xa3,
165 0xc4, 0x55, 0x27, 0xb6, 0xc3, 0x52, 0x20, 0xb1,
166 0xca, 0x5b, 0x29, 0xb8, 0xcd, 0x5c, 0x2e, 0xbf,
167
168 0x90, 0x01, 0x73, 0xe2, 0x97, 0x06, 0x74, 0xe5,
169 0x9e, 0x0f, 0x7d, 0xec, 0x99, 0x08, 0x7a, 0xeb,
170 0x8c, 0x1d, 0x6f, 0xfe, 0x8b, 0x1a, 0x68, 0xf9,
171 0x82, 0x13, 0x61, 0xf0, 0x85, 0x14, 0x66, 0xf7,
172
173 0xa8, 0x39, 0x4b, 0xda, 0xaf, 0x3e, 0x4c, 0xdd,
174 0xa6, 0x37, 0x45, 0xd4, 0xa1, 0x30, 0x42, 0xd3,
175 0xb4, 0x25, 0x57, 0xc6, 0xb3, 0x22, 0x50, 0xc1,
176 0xba, 0x2b, 0x59, 0xc8, 0xbd, 0x2c, 0x5e, 0xcf
177};
178
136/* CRC on 2 bytes */ 179/* CRC on 2 bytes */
137#define __crc(data) (rfcomm_crc_table[rfcomm_crc_table[0xff ^ data[0]] ^ data[1]]) 180#define __crc(data) (rfcomm_crc_table[rfcomm_crc_table[0xff ^ data[0]] ^ data[1]])
138 181
diff --git a/net/bluetooth/rfcomm/crc.c b/net/bluetooth/rfcomm/crc.c
deleted file mode 100644
index 1011bc4a8692..000000000000
--- a/net/bluetooth/rfcomm/crc.c
+++ /dev/null
@@ -1,71 +0,0 @@
1/*
2 RFCOMM implementation for Linux Bluetooth stack (BlueZ).
3 Copyright (C) 2002 Maxim Krasnyansky <maxk@qualcomm.com>
4 Copyright (C) 2002 Marcel Holtmann <marcel@holtmann.org>
5
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License version 2 as
8 published by the Free Software Foundation;
9
10 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
11 OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
12 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT OF THIRD PARTY RIGHTS.
13 IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) AND AUTHOR(S) BE LIABLE FOR ANY
14 CLAIM, OR ANY SPECIAL INDIRECT OR CONSEQUENTIAL DAMAGES, OR ANY DAMAGES
15 WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
16 ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
17 OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
18
19 ALL LIABILITY, INCLUDING LIABILITY FOR INFRINGEMENT OF ANY PATENTS,
20 COPYRIGHTS, TRADEMARKS OR OTHER RIGHTS, RELATING TO USE OF THIS
21 SOFTWARE IS DISCLAIMED.
22*/
23
24/*
25 * RFCOMM FCS calculation.
26 *
27 * $Id: crc.c,v 1.2 2002/09/21 09:54:32 holtmann Exp $
28 */
29
30/* reversed, 8-bit, poly=0x07 */
31unsigned char rfcomm_crc_table[256] = {
32 0x00, 0x91, 0xe3, 0x72, 0x07, 0x96, 0xe4, 0x75,
33 0x0e, 0x9f, 0xed, 0x7c, 0x09, 0x98, 0xea, 0x7b,
34 0x1c, 0x8d, 0xff, 0x6e, 0x1b, 0x8a, 0xf8, 0x69,
35 0x12, 0x83, 0xf1, 0x60, 0x15, 0x84, 0xf6, 0x67,
36
37 0x38, 0xa9, 0xdb, 0x4a, 0x3f, 0xae, 0xdc, 0x4d,
38 0x36, 0xa7, 0xd5, 0x44, 0x31, 0xa0, 0xd2, 0x43,
39 0x24, 0xb5, 0xc7, 0x56, 0x23, 0xb2, 0xc0, 0x51,
40 0x2a, 0xbb, 0xc9, 0x58, 0x2d, 0xbc, 0xce, 0x5f,
41
42 0x70, 0xe1, 0x93, 0x02, 0x77, 0xe6, 0x94, 0x05,
43 0x7e, 0xef, 0x9d, 0x0c, 0x79, 0xe8, 0x9a, 0x0b,
44 0x6c, 0xfd, 0x8f, 0x1e, 0x6b, 0xfa, 0x88, 0x19,
45 0x62, 0xf3, 0x81, 0x10, 0x65, 0xf4, 0x86, 0x17,
46
47 0x48, 0xd9, 0xab, 0x3a, 0x4f, 0xde, 0xac, 0x3d,
48 0x46, 0xd7, 0xa5, 0x34, 0x41, 0xd0, 0xa2, 0x33,
49 0x54, 0xc5, 0xb7, 0x26, 0x53, 0xc2, 0xb0, 0x21,
50 0x5a, 0xcb, 0xb9, 0x28, 0x5d, 0xcc, 0xbe, 0x2f,
51
52 0xe0, 0x71, 0x03, 0x92, 0xe7, 0x76, 0x04, 0x95,
53 0xee, 0x7f, 0x0d, 0x9c, 0xe9, 0x78, 0x0a, 0x9b,
54 0xfc, 0x6d, 0x1f, 0x8e, 0xfb, 0x6a, 0x18, 0x89,
55 0xf2, 0x63, 0x11, 0x80, 0xf5, 0x64, 0x16, 0x87,
56
57 0xd8, 0x49, 0x3b, 0xaa, 0xdf, 0x4e, 0x3c, 0xad,
58 0xd6, 0x47, 0x35, 0xa4, 0xd1, 0x40, 0x32, 0xa3,
59 0xc4, 0x55, 0x27, 0xb6, 0xc3, 0x52, 0x20, 0xb1,
60 0xca, 0x5b, 0x29, 0xb8, 0xcd, 0x5c, 0x2e, 0xbf,
61
62 0x90, 0x01, 0x73, 0xe2, 0x97, 0x06, 0x74, 0xe5,
63 0x9e, 0x0f, 0x7d, 0xec, 0x99, 0x08, 0x7a, 0xeb,
64 0x8c, 0x1d, 0x6f, 0xfe, 0x8b, 0x1a, 0x68, 0xf9,
65 0x82, 0x13, 0x61, 0xf0, 0x85, 0x14, 0x66, 0xf7,
66
67 0xa8, 0x39, 0x4b, 0xda, 0xaf, 0x3e, 0x4c, 0xdd,
68 0xa6, 0x37, 0x45, 0xd4, 0xa1, 0x30, 0x42, 0xd3,
69 0xb4, 0x25, 0x57, 0xc6, 0xb3, 0x22, 0x50, 0xc1,
70 0xba, 0x2b, 0x59, 0xc8, 0xbd, 0x2c, 0x5e, 0xcf
71};
diff --git a/net/bridge/br_fdb.c b/net/bridge/br_fdb.c
index 24396b914d11..1f08a59b51ea 100644
--- a/net/bridge/br_fdb.c
+++ b/net/bridge/br_fdb.c
@@ -86,8 +86,8 @@ void br_fdb_changeaddr(struct net_bridge_port *p, const unsigned char *newaddr)
86 struct net_bridge_port *op; 86 struct net_bridge_port *op;
87 list_for_each_entry(op, &br->port_list, list) { 87 list_for_each_entry(op, &br->port_list, list) {
88 if (op != p && 88 if (op != p &&
89 !memcmp(op->dev->dev_addr, 89 !compare_ether_addr(op->dev->dev_addr,
90 f->addr.addr, ETH_ALEN)) { 90 f->addr.addr)) {
91 f->dst = op; 91 f->dst = op;
92 goto insert; 92 goto insert;
93 } 93 }
@@ -151,8 +151,8 @@ void br_fdb_delete_by_port(struct net_bridge *br, struct net_bridge_port *p)
151 struct net_bridge_port *op; 151 struct net_bridge_port *op;
152 list_for_each_entry(op, &br->port_list, list) { 152 list_for_each_entry(op, &br->port_list, list) {
153 if (op != p && 153 if (op != p &&
154 !memcmp(op->dev->dev_addr, 154 !compare_ether_addr(op->dev->dev_addr,
155 f->addr.addr, ETH_ALEN)) { 155 f->addr.addr)) {
156 f->dst = op; 156 f->dst = op;
157 goto skip_delete; 157 goto skip_delete;
158 } 158 }
@@ -174,7 +174,7 @@ struct net_bridge_fdb_entry *__br_fdb_get(struct net_bridge *br,
174 struct net_bridge_fdb_entry *fdb; 174 struct net_bridge_fdb_entry *fdb;
175 175
176 hlist_for_each_entry_rcu(fdb, h, &br->hash[br_mac_hash(addr)], hlist) { 176 hlist_for_each_entry_rcu(fdb, h, &br->hash[br_mac_hash(addr)], hlist) {
177 if (!memcmp(fdb->addr.addr, addr, ETH_ALEN)) { 177 if (!compare_ether_addr(fdb->addr.addr, addr)) {
178 if (unlikely(has_expired(br, fdb))) 178 if (unlikely(has_expired(br, fdb)))
179 break; 179 break;
180 return fdb; 180 return fdb;
@@ -264,7 +264,7 @@ static inline struct net_bridge_fdb_entry *fdb_find(struct hlist_head *head,
264 struct net_bridge_fdb_entry *fdb; 264 struct net_bridge_fdb_entry *fdb;
265 265
266 hlist_for_each_entry_rcu(fdb, h, head, hlist) { 266 hlist_for_each_entry_rcu(fdb, h, head, hlist) {
267 if (!memcmp(fdb->addr.addr, addr, ETH_ALEN)) 267 if (!compare_ether_addr(fdb->addr.addr, addr))
268 return fdb; 268 return fdb;
269 } 269 }
270 return NULL; 270 return NULL;
diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c
index 9a45e6279c57..b88220a64cd8 100644
--- a/net/bridge/br_input.c
+++ b/net/bridge/br_input.c
@@ -128,7 +128,7 @@ int br_handle_frame(struct net_bridge_port *p, struct sk_buff **pskb)
128 dest = eth_hdr(skb)->h_dest; 128 dest = eth_hdr(skb)->h_dest;
129 } 129 }
130 130
131 if (!memcmp(p->br->dev->dev_addr, dest, ETH_ALEN)) 131 if (!compare_ether_addr(p->br->dev->dev_addr, dest))
132 skb->pkt_type = PACKET_HOST; 132 skb->pkt_type = PACKET_HOST;
133 133
134 NF_HOOK(PF_BRIDGE, NF_BR_PRE_ROUTING, skb, skb->dev, NULL, 134 NF_HOOK(PF_BRIDGE, NF_BR_PRE_ROUTING, skb, skb->dev, NULL,
diff --git a/net/bridge/br_stp_if.c b/net/bridge/br_stp_if.c
index 0da11ff05fa3..ac09b6a23523 100644
--- a/net/bridge/br_stp_if.c
+++ b/net/bridge/br_stp_if.c
@@ -15,6 +15,7 @@
15 15
16#include <linux/kernel.h> 16#include <linux/kernel.h>
17#include <linux/smp_lock.h> 17#include <linux/smp_lock.h>
18#include <linux/etherdevice.h>
18 19
19#include "br_private.h" 20#include "br_private.h"
20#include "br_private_stp.h" 21#include "br_private_stp.h"
@@ -133,10 +134,10 @@ static void br_stp_change_bridge_id(struct net_bridge *br,
133 memcpy(br->dev->dev_addr, addr, ETH_ALEN); 134 memcpy(br->dev->dev_addr, addr, ETH_ALEN);
134 135
135 list_for_each_entry(p, &br->port_list, list) { 136 list_for_each_entry(p, &br->port_list, list) {
136 if (!memcmp(p->designated_bridge.addr, oldaddr, ETH_ALEN)) 137 if (!compare_ether_addr(p->designated_bridge.addr, oldaddr))
137 memcpy(p->designated_bridge.addr, addr, ETH_ALEN); 138 memcpy(p->designated_bridge.addr, addr, ETH_ALEN);
138 139
139 if (!memcmp(p->designated_root.addr, oldaddr, ETH_ALEN)) 140 if (!compare_ether_addr(p->designated_root.addr, oldaddr))
140 memcpy(p->designated_root.addr, addr, ETH_ALEN); 141 memcpy(p->designated_root.addr, addr, ETH_ALEN);
141 142
142 } 143 }
@@ -157,12 +158,12 @@ void br_stp_recalculate_bridge_id(struct net_bridge *br)
157 158
158 list_for_each_entry(p, &br->port_list, list) { 159 list_for_each_entry(p, &br->port_list, list) {
159 if (addr == br_mac_zero || 160 if (addr == br_mac_zero ||
160 memcmp(p->dev->dev_addr, addr, ETH_ALEN) < 0) 161 compare_ether_addr(p->dev->dev_addr, addr) < 0)
161 addr = p->dev->dev_addr; 162 addr = p->dev->dev_addr;
162 163
163 } 164 }
164 165
165 if (memcmp(br->bridge_id.addr, addr, ETH_ALEN)) 166 if (compare_ether_addr(br->bridge_id.addr, addr))
166 br_stp_change_bridge_id(br, addr); 167 br_stp_change_bridge_id(br, addr);
167} 168}
168 169
diff --git a/net/core/datagram.c b/net/core/datagram.c
index 81987df536eb..d219435d086c 100644
--- a/net/core/datagram.c
+++ b/net/core/datagram.c
@@ -213,6 +213,10 @@ int skb_copy_datagram_iovec(const struct sk_buff *skb, int offset,
213{ 213{
214 int i, err, fraglen, end = 0; 214 int i, err, fraglen, end = 0;
215 struct sk_buff *next = skb_shinfo(skb)->frag_list; 215 struct sk_buff *next = skb_shinfo(skb)->frag_list;
216
217 if (!len)
218 return 0;
219
216next_skb: 220next_skb:
217 fraglen = skb_headlen(skb); 221 fraglen = skb_headlen(skb);
218 i = -1; 222 i = -1;
diff --git a/net/core/dev.c b/net/core/dev.c
index a44eeef24edf..8d1541595277 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -2717,6 +2717,20 @@ int register_netdevice(struct net_device *dev)
2717 dev->name); 2717 dev->name);
2718 dev->features &= ~NETIF_F_TSO; 2718 dev->features &= ~NETIF_F_TSO;
2719 } 2719 }
2720 if (dev->features & NETIF_F_UFO) {
2721 if (!(dev->features & NETIF_F_HW_CSUM)) {
2722 printk(KERN_ERR "%s: Dropping NETIF_F_UFO since no "
2723 "NETIF_F_HW_CSUM feature.\n",
2724 dev->name);
2725 dev->features &= ~NETIF_F_UFO;
2726 }
2727 if (!(dev->features & NETIF_F_SG)) {
2728 printk(KERN_ERR "%s: Dropping NETIF_F_UFO since no "
2729 "NETIF_F_SG feature.\n",
2730 dev->name);
2731 dev->features &= ~NETIF_F_UFO;
2732 }
2733 }
2720 2734
2721 /* 2735 /*
2722 * nil rebuild_header routine, 2736 * nil rebuild_header routine,
diff --git a/net/core/ethtool.c b/net/core/ethtool.c
index 404b761e82ce..0350586e9195 100644
--- a/net/core/ethtool.c
+++ b/net/core/ethtool.c
@@ -93,6 +93,20 @@ int ethtool_op_get_perm_addr(struct net_device *dev, struct ethtool_perm_addr *a
93} 93}
94 94
95 95
96u32 ethtool_op_get_ufo(struct net_device *dev)
97{
98 return (dev->features & NETIF_F_UFO) != 0;
99}
100
101int ethtool_op_set_ufo(struct net_device *dev, u32 data)
102{
103 if (data)
104 dev->features |= NETIF_F_UFO;
105 else
106 dev->features &= ~NETIF_F_UFO;
107 return 0;
108}
109
96/* Handlers for each ethtool command */ 110/* Handlers for each ethtool command */
97 111
98static int ethtool_get_settings(struct net_device *dev, void __user *useraddr) 112static int ethtool_get_settings(struct net_device *dev, void __user *useraddr)
@@ -483,6 +497,11 @@ static int __ethtool_set_sg(struct net_device *dev, u32 data)
483 return err; 497 return err;
484 } 498 }
485 499
500 if (!data && dev->ethtool_ops->set_ufo) {
501 err = dev->ethtool_ops->set_ufo(dev, 0);
502 if (err)
503 return err;
504 }
486 return dev->ethtool_ops->set_sg(dev, data); 505 return dev->ethtool_ops->set_sg(dev, data);
487} 506}
488 507
@@ -569,6 +588,32 @@ static int ethtool_set_tso(struct net_device *dev, char __user *useraddr)
569 return dev->ethtool_ops->set_tso(dev, edata.data); 588 return dev->ethtool_ops->set_tso(dev, edata.data);
570} 589}
571 590
591static int ethtool_get_ufo(struct net_device *dev, char __user *useraddr)
592{
593 struct ethtool_value edata = { ETHTOOL_GTSO };
594
595 if (!dev->ethtool_ops->get_ufo)
596 return -EOPNOTSUPP;
597 edata.data = dev->ethtool_ops->get_ufo(dev);
598 if (copy_to_user(useraddr, &edata, sizeof(edata)))
599 return -EFAULT;
600 return 0;
601}
602static int ethtool_set_ufo(struct net_device *dev, char __user *useraddr)
603{
604 struct ethtool_value edata;
605
606 if (!dev->ethtool_ops->set_ufo)
607 return -EOPNOTSUPP;
608 if (copy_from_user(&edata, useraddr, sizeof(edata)))
609 return -EFAULT;
610 if (edata.data && !(dev->features & NETIF_F_SG))
611 return -EINVAL;
612 if (edata.data && !(dev->features & NETIF_F_HW_CSUM))
613 return -EINVAL;
614 return dev->ethtool_ops->set_ufo(dev, edata.data);
615}
616
572static int ethtool_self_test(struct net_device *dev, char __user *useraddr) 617static int ethtool_self_test(struct net_device *dev, char __user *useraddr)
573{ 618{
574 struct ethtool_test test; 619 struct ethtool_test test;
@@ -854,6 +899,12 @@ int dev_ethtool(struct ifreq *ifr)
854 case ETHTOOL_GPERMADDR: 899 case ETHTOOL_GPERMADDR:
855 rc = ethtool_get_perm_addr(dev, useraddr); 900 rc = ethtool_get_perm_addr(dev, useraddr);
856 break; 901 break;
902 case ETHTOOL_GUFO:
903 rc = ethtool_get_ufo(dev, useraddr);
904 break;
905 case ETHTOOL_SUFO:
906 rc = ethtool_set_ufo(dev, useraddr);
907 break;
857 default: 908 default:
858 rc = -EOPNOTSUPP; 909 rc = -EOPNOTSUPP;
859 } 910 }
@@ -882,3 +933,5 @@ EXPORT_SYMBOL(ethtool_op_set_sg);
882EXPORT_SYMBOL(ethtool_op_set_tso); 933EXPORT_SYMBOL(ethtool_op_set_tso);
883EXPORT_SYMBOL(ethtool_op_set_tx_csum); 934EXPORT_SYMBOL(ethtool_op_set_tx_csum);
884EXPORT_SYMBOL(ethtool_op_set_tx_hw_csum); 935EXPORT_SYMBOL(ethtool_op_set_tx_hw_csum);
936EXPORT_SYMBOL(ethtool_op_set_ufo);
937EXPORT_SYMBOL(ethtool_op_get_ufo);
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index ef9d46b91eb9..95501e40100e 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -176,6 +176,8 @@ struct sk_buff *__alloc_skb(unsigned int size, gfp_t gfp_mask,
176 skb_shinfo(skb)->tso_size = 0; 176 skb_shinfo(skb)->tso_size = 0;
177 skb_shinfo(skb)->tso_segs = 0; 177 skb_shinfo(skb)->tso_segs = 0;
178 skb_shinfo(skb)->frag_list = NULL; 178 skb_shinfo(skb)->frag_list = NULL;
179 skb_shinfo(skb)->ufo_size = 0;
180 skb_shinfo(skb)->ip6_frag_id = 0;
179out: 181out:
180 return skb; 182 return skb;
181nodata: 183nodata:
@@ -1696,6 +1698,78 @@ unsigned int skb_find_text(struct sk_buff *skb, unsigned int from,
1696 return textsearch_find(config, state); 1698 return textsearch_find(config, state);
1697} 1699}
1698 1700
1701/**
1702 * skb_append_datato_frags: - append the user data to a skb
1703 * @sk: sock structure
1704 * @skb: skb structure to be appened with user data.
1705 * @getfrag: call back function to be used for getting the user data
1706 * @from: pointer to user message iov
1707 * @length: length of the iov message
1708 *
1709 * Description: This procedure append the user data in the fragment part
1710 * of the skb if any page alloc fails user this procedure returns -ENOMEM
1711 */
1712int skb_append_datato_frags(struct sock *sk, struct sk_buff *skb,
1713 int getfrag(void *from, char *to, int offset,
1714 int len, int odd, struct sk_buff *skb),
1715 void *from, int length)
1716{
1717 int frg_cnt = 0;
1718 skb_frag_t *frag = NULL;
1719 struct page *page = NULL;
1720 int copy, left;
1721 int offset = 0;
1722 int ret;
1723
1724 do {
1725 /* Return error if we don't have space for new frag */
1726 frg_cnt = skb_shinfo(skb)->nr_frags;
1727 if (frg_cnt >= MAX_SKB_FRAGS)
1728 return -EFAULT;
1729
1730 /* allocate a new page for next frag */
1731 page = alloc_pages(sk->sk_allocation, 0);
1732
1733 /* If alloc_page fails just return failure and caller will
1734 * free previous allocated pages by doing kfree_skb()
1735 */
1736 if (page == NULL)
1737 return -ENOMEM;
1738
1739 /* initialize the next frag */
1740 sk->sk_sndmsg_page = page;
1741 sk->sk_sndmsg_off = 0;
1742 skb_fill_page_desc(skb, frg_cnt, page, 0, 0);
1743 skb->truesize += PAGE_SIZE;
1744 atomic_add(PAGE_SIZE, &sk->sk_wmem_alloc);
1745
1746 /* get the new initialized frag */
1747 frg_cnt = skb_shinfo(skb)->nr_frags;
1748 frag = &skb_shinfo(skb)->frags[frg_cnt - 1];
1749
1750 /* copy the user data to page */
1751 left = PAGE_SIZE - frag->page_offset;
1752 copy = (length > left)? left : length;
1753
1754 ret = getfrag(from, (page_address(frag->page) +
1755 frag->page_offset + frag->size),
1756 offset, copy, 0, skb);
1757 if (ret < 0)
1758 return -EFAULT;
1759
1760 /* copy was successful so update the size parameters */
1761 sk->sk_sndmsg_off += copy;
1762 frag->size += copy;
1763 skb->len += copy;
1764 skb->data_len += copy;
1765 offset += copy;
1766 length -= copy;
1767
1768 } while (length > 0);
1769
1770 return 0;
1771}
1772
1699void __init skb_init(void) 1773void __init skb_init(void)
1700{ 1774{
1701 skbuff_head_cache = kmem_cache_create("skbuff_head_cache", 1775 skbuff_head_cache = kmem_cache_create("skbuff_head_cache",
@@ -1747,3 +1821,4 @@ EXPORT_SYMBOL(skb_prepare_seq_read);
1747EXPORT_SYMBOL(skb_seq_read); 1821EXPORT_SYMBOL(skb_seq_read);
1748EXPORT_SYMBOL(skb_abort_seq_read); 1822EXPORT_SYMBOL(skb_abort_seq_read);
1749EXPORT_SYMBOL(skb_find_text); 1823EXPORT_SYMBOL(skb_find_text);
1824EXPORT_SYMBOL(skb_append_datato_frags);
diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h
index 5871c027f9dc..f97b85d55ad8 100644
--- a/net/dccp/dccp.h
+++ b/net/dccp/dccp.h
@@ -118,7 +118,6 @@ DECLARE_SNMP_STAT(struct dccp_mib, dccp_statistics);
118#define DCCP_ADD_STATS_USER(field, val) \ 118#define DCCP_ADD_STATS_USER(field, val) \
119 SNMP_ADD_STATS_USER(dccp_statistics, field, val) 119 SNMP_ADD_STATS_USER(dccp_statistics, field, val)
120 120
121extern int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb);
122extern int dccp_retransmit_skb(struct sock *sk, struct sk_buff *skb); 121extern int dccp_retransmit_skb(struct sock *sk, struct sk_buff *skb);
123 122
124extern int dccp_send_response(struct sock *sk); 123extern int dccp_send_response(struct sock *sk);
diff --git a/net/dccp/output.c b/net/dccp/output.c
index d59f86f7ceab..74ff87025878 100644
--- a/net/dccp/output.c
+++ b/net/dccp/output.c
@@ -12,6 +12,7 @@
12 12
13#include <linux/config.h> 13#include <linux/config.h>
14#include <linux/dccp.h> 14#include <linux/dccp.h>
15#include <linux/kernel.h>
15#include <linux/skbuff.h> 16#include <linux/skbuff.h>
16 17
17#include <net/sock.h> 18#include <net/sock.h>
@@ -25,13 +26,20 @@ static inline void dccp_event_ack_sent(struct sock *sk)
25 inet_csk_clear_xmit_timer(sk, ICSK_TIME_DACK); 26 inet_csk_clear_xmit_timer(sk, ICSK_TIME_DACK);
26} 27}
27 28
29static inline void dccp_skb_entail(struct sock *sk, struct sk_buff *skb)
30{
31 skb_set_owner_w(skb, sk);
32 WARN_ON(sk->sk_send_head);
33 sk->sk_send_head = skb;
34}
35
28/* 36/*
29 * All SKB's seen here are completely headerless. It is our 37 * All SKB's seen here are completely headerless. It is our
30 * job to build the DCCP header, and pass the packet down to 38 * job to build the DCCP header, and pass the packet down to
31 * IP so it can do the same plus pass the packet off to the 39 * IP so it can do the same plus pass the packet off to the
32 * device. 40 * device.
33 */ 41 */
34int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb) 42static int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb)
35{ 43{
36 if (likely(skb != NULL)) { 44 if (likely(skb != NULL)) {
37 const struct inet_sock *inet = inet_sk(sk); 45 const struct inet_sock *inet = inet_sk(sk);
@@ -50,10 +58,21 @@ int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb)
50 switch (dcb->dccpd_type) { 58 switch (dcb->dccpd_type) {
51 case DCCP_PKT_DATA: 59 case DCCP_PKT_DATA:
52 set_ack = 0; 60 set_ack = 0;
61 /* fall through */
62 case DCCP_PKT_DATAACK:
53 break; 63 break;
64
54 case DCCP_PKT_SYNC: 65 case DCCP_PKT_SYNC:
55 case DCCP_PKT_SYNCACK: 66 case DCCP_PKT_SYNCACK:
56 ackno = dcb->dccpd_seq; 67 ackno = dcb->dccpd_seq;
68 /* fall through */
69 default:
70 /*
71 * Only data packets should come through with skb->sk
72 * set.
73 */
74 WARN_ON(skb->sk);
75 skb_set_owner_w(skb, sk);
57 break; 76 break;
58 } 77 }
59 78
@@ -63,9 +82,6 @@ int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb)
63 skb->h.raw = skb_push(skb, dccp_header_size); 82 skb->h.raw = skb_push(skb, dccp_header_size);
64 dh = dccp_hdr(skb); 83 dh = dccp_hdr(skb);
65 84
66 if (!skb->sk)
67 skb_set_owner_w(skb, sk);
68
69 /* Build DCCP header and checksum it. */ 85 /* Build DCCP header and checksum it. */
70 memset(dh, 0, dccp_header_size); 86 memset(dh, 0, dccp_header_size);
71 dh->dccph_type = dcb->dccpd_type; 87 dh->dccph_type = dcb->dccpd_type;
@@ -393,10 +409,8 @@ int dccp_connect(struct sock *sk)
393 409
394 DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_REQUEST; 410 DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_REQUEST;
395 skb->csum = 0; 411 skb->csum = 0;
396 skb_set_owner_w(skb, sk);
397 412
398 BUG_TRAP(sk->sk_send_head == NULL); 413 dccp_skb_entail(sk, skb);
399 sk->sk_send_head = skb;
400 dccp_transmit_skb(sk, skb_clone(skb, GFP_KERNEL)); 414 dccp_transmit_skb(sk, skb_clone(skb, GFP_KERNEL));
401 DCCP_INC_STATS(DCCP_MIB_ACTIVEOPENS); 415 DCCP_INC_STATS(DCCP_MIB_ACTIVEOPENS);
402 416
@@ -425,7 +439,6 @@ void dccp_send_ack(struct sock *sk)
425 skb_reserve(skb, MAX_DCCP_HEADER); 439 skb_reserve(skb, MAX_DCCP_HEADER);
426 skb->csum = 0; 440 skb->csum = 0;
427 DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_ACK; 441 DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_ACK;
428 skb_set_owner_w(skb, sk);
429 dccp_transmit_skb(sk, skb); 442 dccp_transmit_skb(sk, skb);
430 } 443 }
431} 444}
@@ -482,7 +495,6 @@ void dccp_send_sync(struct sock *sk, const u64 seq,
482 DCCP_SKB_CB(skb)->dccpd_type = pkt_type; 495 DCCP_SKB_CB(skb)->dccpd_type = pkt_type;
483 DCCP_SKB_CB(skb)->dccpd_seq = seq; 496 DCCP_SKB_CB(skb)->dccpd_seq = seq;
484 497
485 skb_set_owner_w(skb, sk);
486 dccp_transmit_skb(sk, skb); 498 dccp_transmit_skb(sk, skb);
487} 499}
488 500
@@ -507,10 +519,8 @@ void dccp_send_close(struct sock *sk, const int active)
507 DCCP_SKB_CB(skb)->dccpd_type = dp->dccps_role == DCCP_ROLE_CLIENT ? 519 DCCP_SKB_CB(skb)->dccpd_type = dp->dccps_role == DCCP_ROLE_CLIENT ?
508 DCCP_PKT_CLOSE : DCCP_PKT_CLOSEREQ; 520 DCCP_PKT_CLOSE : DCCP_PKT_CLOSEREQ;
509 521
510 skb_set_owner_w(skb, sk);
511 if (active) { 522 if (active) {
512 BUG_TRAP(sk->sk_send_head == NULL); 523 dccp_skb_entail(sk, skb);
513 sk->sk_send_head = skb;
514 dccp_transmit_skb(sk, skb_clone(skb, prio)); 524 dccp_transmit_skb(sk, skb_clone(skb, prio));
515 } else 525 } else
516 dccp_transmit_skb(sk, skb); 526 dccp_transmit_skb(sk, skb);
diff --git a/net/ethernet/eth.c b/net/ethernet/eth.c
index 68a5ca866442..e24577367274 100644
--- a/net/ethernet/eth.c
+++ b/net/ethernet/eth.c
@@ -146,19 +146,6 @@ int eth_rebuild_header(struct sk_buff *skb)
146 return 0; 146 return 0;
147} 147}
148 148
149static inline unsigned int compare_eth_addr(const unsigned char *__a, const unsigned char *__b)
150{
151 const unsigned short *dest = (unsigned short *) __a;
152 const unsigned short *devaddr = (unsigned short *) __b;
153 unsigned int res;
154
155 BUILD_BUG_ON(ETH_ALEN != 6);
156 res = ((dest[0] ^ devaddr[0]) |
157 (dest[1] ^ devaddr[1]) |
158 (dest[2] ^ devaddr[2])) != 0;
159
160 return res;
161}
162 149
163/* 150/*
164 * Determine the packet's protocol ID. The rule here is that we 151 * Determine the packet's protocol ID. The rule here is that we
@@ -176,7 +163,7 @@ __be16 eth_type_trans(struct sk_buff *skb, struct net_device *dev)
176 eth = eth_hdr(skb); 163 eth = eth_hdr(skb);
177 164
178 if (*eth->h_dest&1) { 165 if (*eth->h_dest&1) {
179 if (!compare_eth_addr(eth->h_dest, dev->broadcast)) 166 if (!compare_ether_addr(eth->h_dest, dev->broadcast))
180 skb->pkt_type = PACKET_BROADCAST; 167 skb->pkt_type = PACKET_BROADCAST;
181 else 168 else
182 skb->pkt_type = PACKET_MULTICAST; 169 skb->pkt_type = PACKET_MULTICAST;
@@ -191,7 +178,7 @@ __be16 eth_type_trans(struct sk_buff *skb, struct net_device *dev)
191 */ 178 */
192 179
193 else if(1 /*dev->flags&IFF_PROMISC*/) { 180 else if(1 /*dev->flags&IFF_PROMISC*/) {
194 if (unlikely(compare_eth_addr(eth->h_dest, dev->dev_addr))) 181 if (unlikely(compare_ether_addr(eth->h_dest, dev->dev_addr)))
195 skb->pkt_type = PACKET_OTHERHOST; 182 skb->pkt_type = PACKET_OTHERHOST;
196 } 183 }
197 184
diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c
index e61bc7177eb1..990633c09dfe 100644
--- a/net/ipv4/fib_frontend.c
+++ b/net/ipv4/fib_frontend.c
@@ -591,7 +591,7 @@ static int fib_inetaddr_event(struct notifier_block *this, unsigned long event,
591 break; 591 break;
592 case NETDEV_DOWN: 592 case NETDEV_DOWN:
593 fib_del_ifaddr(ifa); 593 fib_del_ifaddr(ifa);
594 if (ifa->ifa_dev && ifa->ifa_dev->ifa_list == NULL) { 594 if (ifa->ifa_dev->ifa_list == NULL) {
595 /* Last address was deleted from this interface. 595 /* Last address was deleted from this interface.
596 Disable IP. 596 Disable IP.
597 */ 597 */
diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c
index 8b6d3939e1e6..c6247fc84060 100644
--- a/net/ipv4/igmp.c
+++ b/net/ipv4/igmp.c
@@ -1908,8 +1908,11 @@ int ip_mc_msfilter(struct sock *sk, struct ip_msfilter *msf, int ifindex)
1908 sock_kfree_s(sk, newpsl, IP_SFLSIZE(newpsl->sl_max)); 1908 sock_kfree_s(sk, newpsl, IP_SFLSIZE(newpsl->sl_max));
1909 goto done; 1909 goto done;
1910 } 1910 }
1911 } else 1911 } else {
1912 newpsl = NULL; 1912 newpsl = NULL;
1913 (void) ip_mc_add_src(in_dev, &msf->imsf_multiaddr,
1914 msf->imsf_fmode, 0, NULL, 0);
1915 }
1913 psl = pmc->sflist; 1916 psl = pmc->sflist;
1914 if (psl) { 1917 if (psl) {
1915 (void) ip_mc_del_src(in_dev, &msf->imsf_multiaddr, pmc->sfmode, 1918 (void) ip_mc_del_src(in_dev, &msf->imsf_multiaddr, pmc->sfmode,
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index 87e350069abb..17758234a3e3 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -275,7 +275,8 @@ int ip_output(struct sk_buff *skb)
275{ 275{
276 IP_INC_STATS(IPSTATS_MIB_OUTREQUESTS); 276 IP_INC_STATS(IPSTATS_MIB_OUTREQUESTS);
277 277
278 if (skb->len > dst_mtu(skb->dst) && !skb_shinfo(skb)->tso_size) 278 if (skb->len > dst_mtu(skb->dst) &&
279 !(skb_shinfo(skb)->ufo_size || skb_shinfo(skb)->tso_size))
279 return ip_fragment(skb, ip_finish_output); 280 return ip_fragment(skb, ip_finish_output);
280 else 281 else
281 return ip_finish_output(skb); 282 return ip_finish_output(skb);
@@ -688,6 +689,60 @@ csum_page(struct page *page, int offset, int copy)
688 return csum; 689 return csum;
689} 690}
690 691
692inline int ip_ufo_append_data(struct sock *sk,
693 int getfrag(void *from, char *to, int offset, int len,
694 int odd, struct sk_buff *skb),
695 void *from, int length, int hh_len, int fragheaderlen,
696 int transhdrlen, int mtu,unsigned int flags)
697{
698 struct sk_buff *skb;
699 int err;
700
701 /* There is support for UDP fragmentation offload by network
702 * device, so create one single skb packet containing complete
703 * udp datagram
704 */
705 if ((skb = skb_peek_tail(&sk->sk_write_queue)) == NULL) {
706 skb = sock_alloc_send_skb(sk,
707 hh_len + fragheaderlen + transhdrlen + 20,
708 (flags & MSG_DONTWAIT), &err);
709
710 if (skb == NULL)
711 return err;
712
713 /* reserve space for Hardware header */
714 skb_reserve(skb, hh_len);
715
716 /* create space for UDP/IP header */
717 skb_put(skb,fragheaderlen + transhdrlen);
718
719 /* initialize network header pointer */
720 skb->nh.raw = skb->data;
721
722 /* initialize protocol header pointer */
723 skb->h.raw = skb->data + fragheaderlen;
724
725 skb->ip_summed = CHECKSUM_HW;
726 skb->csum = 0;
727 sk->sk_sndmsg_off = 0;
728 }
729
730 err = skb_append_datato_frags(sk,skb, getfrag, from,
731 (length - transhdrlen));
732 if (!err) {
733 /* specify the length of each IP datagram fragment*/
734 skb_shinfo(skb)->ufo_size = (mtu - fragheaderlen);
735 __skb_queue_tail(&sk->sk_write_queue, skb);
736
737 return 0;
738 }
739 /* There is not enough support do UFO ,
740 * so follow normal path
741 */
742 kfree_skb(skb);
743 return err;
744}
745
691/* 746/*
692 * ip_append_data() and ip_append_page() can make one large IP datagram 747 * ip_append_data() and ip_append_page() can make one large IP datagram
693 * from many pieces of data. Each pieces will be holded on the socket 748 * from many pieces of data. Each pieces will be holded on the socket
@@ -777,6 +832,15 @@ int ip_append_data(struct sock *sk,
777 csummode = CHECKSUM_HW; 832 csummode = CHECKSUM_HW;
778 833
779 inet->cork.length += length; 834 inet->cork.length += length;
835 if (((length > mtu) && (sk->sk_protocol == IPPROTO_UDP)) &&
836 (rt->u.dst.dev->features & NETIF_F_UFO)) {
837
838 if(ip_ufo_append_data(sk, getfrag, from, length, hh_len,
839 fragheaderlen, transhdrlen, mtu, flags))
840 goto error;
841
842 return 0;
843 }
780 844
781 /* So, what's going on in the loop below? 845 /* So, what's going on in the loop below?
782 * 846 *
@@ -1008,14 +1072,23 @@ ssize_t ip_append_page(struct sock *sk, struct page *page,
1008 return -EINVAL; 1072 return -EINVAL;
1009 1073
1010 inet->cork.length += size; 1074 inet->cork.length += size;
1075 if ((sk->sk_protocol == IPPROTO_UDP) &&
1076 (rt->u.dst.dev->features & NETIF_F_UFO))
1077 skb_shinfo(skb)->ufo_size = (mtu - fragheaderlen);
1078
1011 1079
1012 while (size > 0) { 1080 while (size > 0) {
1013 int i; 1081 int i;
1014 1082
1015 /* Check if the remaining data fits into current packet. */ 1083 if (skb_shinfo(skb)->ufo_size)
1016 len = mtu - skb->len; 1084 len = size;
1017 if (len < size) 1085 else {
1018 len = maxfraglen - skb->len; 1086
1087 /* Check if the remaining data fits into current packet. */
1088 len = mtu - skb->len;
1089 if (len < size)
1090 len = maxfraglen - skb->len;
1091 }
1019 if (len <= 0) { 1092 if (len <= 0) {
1020 struct sk_buff *skb_prev; 1093 struct sk_buff *skb_prev;
1021 char *data; 1094 char *data;
diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c
index a7969286e6e7..3c2e9639bba6 100644
--- a/net/ipv4/netfilter/arp_tables.c
+++ b/net/ipv4/netfilter/arp_tables.c
@@ -347,58 +347,106 @@ unsigned int arpt_do_table(struct sk_buff **pskb,
347 return verdict; 347 return verdict;
348} 348}
349 349
350static inline void *find_inlist_lock_noload(struct list_head *head, 350/*
351 const char *name, 351 * These are weird, but module loading must not be done with mutex
352 int *error, 352 * held (since they will register), and we have to have a single
353 struct semaphore *mutex) 353 * function to use try_then_request_module().
354 */
355
356/* Find table by name, grabs mutex & ref. Returns ERR_PTR() on error. */
357static inline struct arpt_table *find_table_lock(const char *name)
354{ 358{
355 void *ret; 359 struct arpt_table *t;
356 360
357 *error = down_interruptible(mutex); 361 if (down_interruptible(&arpt_mutex) != 0)
358 if (*error != 0) 362 return ERR_PTR(-EINTR);
359 return NULL;
360 363
361 ret = list_named_find(head, name); 364 list_for_each_entry(t, &arpt_tables, list)
362 if (!ret) { 365 if (strcmp(t->name, name) == 0 && try_module_get(t->me))
363 *error = -ENOENT; 366 return t;
364 up(mutex); 367 up(&arpt_mutex);
365 } 368 return NULL;
366 return ret;
367} 369}
368 370
369#ifndef CONFIG_KMOD 371
370#define find_inlist_lock(h,n,p,e,m) find_inlist_lock_noload((h),(n),(e),(m)) 372/* Find target, grabs ref. Returns ERR_PTR() on error. */
371#else 373static inline struct arpt_target *find_target(const char *name, u8 revision)
372static void *
373find_inlist_lock(struct list_head *head,
374 const char *name,
375 const char *prefix,
376 int *error,
377 struct semaphore *mutex)
378{ 374{
379 void *ret; 375 struct arpt_target *t;
376 int err = 0;
380 377
381 ret = find_inlist_lock_noload(head, name, error, mutex); 378 if (down_interruptible(&arpt_mutex) != 0)
382 if (!ret) { 379 return ERR_PTR(-EINTR);
383 duprintf("find_inlist: loading `%s%s'.\n", prefix, name); 380
384 request_module("%s%s", prefix, name); 381 list_for_each_entry(t, &arpt_target, list) {
385 ret = find_inlist_lock_noload(head, name, error, mutex); 382 if (strcmp(t->name, name) == 0) {
383 if (t->revision == revision) {
384 if (try_module_get(t->me)) {
385 up(&arpt_mutex);
386 return t;
387 }
388 } else
389 err = -EPROTOTYPE; /* Found something. */
390 }
386 } 391 }
392 up(&arpt_mutex);
393 return ERR_PTR(err);
394}
387 395
388 return ret; 396struct arpt_target *arpt_find_target(const char *name, u8 revision)
397{
398 struct arpt_target *target;
399
400 target = try_then_request_module(find_target(name, revision),
401 "arpt_%s", name);
402 if (IS_ERR(target) || !target)
403 return NULL;
404 return target;
389} 405}
390#endif
391 406
392static inline struct arpt_table *arpt_find_table_lock(const char *name, int *error, struct semaphore *mutex) 407static int target_revfn(const char *name, u8 revision, int *bestp)
393{ 408{
394 return find_inlist_lock(&arpt_tables, name, "arptable_", error, mutex); 409 struct arpt_target *t;
410 int have_rev = 0;
411
412 list_for_each_entry(t, &arpt_target, list) {
413 if (strcmp(t->name, name) == 0) {
414 if (t->revision > *bestp)
415 *bestp = t->revision;
416 if (t->revision == revision)
417 have_rev =1;
418 }
419 }
420 return have_rev;
395} 421}
396 422
397static struct arpt_target *arpt_find_target_lock(const char *name, int *error, struct semaphore *mutex) 423/* Returns true or false (if no such extension at all) */
424static inline int find_revision(const char *name, u8 revision,
425 int (*revfn)(const char *, u8, int *),
426 int *err)
398{ 427{
399 return find_inlist_lock(&arpt_target, name, "arpt_", error, mutex); 428 int have_rev, best = -1;
429
430 if (down_interruptible(&arpt_mutex) != 0) {
431 *err = -EINTR;
432 return 1;
433 }
434 have_rev = revfn(name, revision, &best);
435 up(&arpt_mutex);
436
437 /* Nothing at all? Return 0 to try loading module. */
438 if (best == -1) {
439 *err = -ENOENT;
440 return 0;
441 }
442
443 *err = best;
444 if (!have_rev)
445 *err = -EPROTONOSUPPORT;
446 return 1;
400} 447}
401 448
449
402/* All zeroes == unconditional rule. */ 450/* All zeroes == unconditional rule. */
403static inline int unconditional(const struct arpt_arp *arp) 451static inline int unconditional(const struct arpt_arp *arp)
404{ 452{
@@ -544,17 +592,15 @@ static inline int check_entry(struct arpt_entry *e, const char *name, unsigned i
544 } 592 }
545 593
546 t = arpt_get_target(e); 594 t = arpt_get_target(e);
547 target = arpt_find_target_lock(t->u.user.name, &ret, &arpt_mutex); 595 target = try_then_request_module(find_target(t->u.user.name,
548 if (!target) { 596 t->u.user.revision),
597 "arpt_%s", t->u.user.name);
598 if (IS_ERR(target) || !target) {
549 duprintf("check_entry: `%s' not found\n", t->u.user.name); 599 duprintf("check_entry: `%s' not found\n", t->u.user.name);
600 ret = target ? PTR_ERR(target) : -ENOENT;
550 goto out; 601 goto out;
551 } 602 }
552 if (!try_module_get((target->me))) {
553 ret = -ENOENT;
554 goto out_unlock;
555 }
556 t->u.kernel.target = target; 603 t->u.kernel.target = target;
557 up(&arpt_mutex);
558 604
559 if (t->u.kernel.target == &arpt_standard_target) { 605 if (t->u.kernel.target == &arpt_standard_target) {
560 if (!standard_check(t, size)) { 606 if (!standard_check(t, size)) {
@@ -576,8 +622,6 @@ static inline int check_entry(struct arpt_entry *e, const char *name, unsigned i
576 (*i)++; 622 (*i)++;
577 return 0; 623 return 0;
578 624
579out_unlock:
580 up(&arpt_mutex);
581out: 625out:
582 return ret; 626 return ret;
583} 627}
@@ -846,8 +890,8 @@ static int get_entries(const struct arpt_get_entries *entries,
846 int ret; 890 int ret;
847 struct arpt_table *t; 891 struct arpt_table *t;
848 892
849 t = arpt_find_table_lock(entries->name, &ret, &arpt_mutex); 893 t = find_table_lock(entries->name);
850 if (t) { 894 if (t || !IS_ERR(t)) {
851 duprintf("t->private->number = %u\n", 895 duprintf("t->private->number = %u\n",
852 t->private->number); 896 t->private->number);
853 if (entries->size == t->private->size) 897 if (entries->size == t->private->size)
@@ -859,10 +903,10 @@ static int get_entries(const struct arpt_get_entries *entries,
859 entries->size); 903 entries->size);
860 ret = -EINVAL; 904 ret = -EINVAL;
861 } 905 }
906 module_put(t->me);
862 up(&arpt_mutex); 907 up(&arpt_mutex);
863 } else 908 } else
864 duprintf("get_entries: Can't find %s!\n", 909 ret = t ? PTR_ERR(t) : -ENOENT;
865 entries->name);
866 910
867 return ret; 911 return ret;
868} 912}
@@ -913,22 +957,19 @@ static int do_replace(void __user *user, unsigned int len)
913 957
914 duprintf("arp_tables: Translated table\n"); 958 duprintf("arp_tables: Translated table\n");
915 959
916 t = arpt_find_table_lock(tmp.name, &ret, &arpt_mutex); 960 t = try_then_request_module(find_table_lock(tmp.name),
917 if (!t) 961 "arptable_%s", tmp.name);
962 if (!t || IS_ERR(t)) {
963 ret = t ? PTR_ERR(t) : -ENOENT;
918 goto free_newinfo_counters_untrans; 964 goto free_newinfo_counters_untrans;
965 }
919 966
920 /* You lied! */ 967 /* You lied! */
921 if (tmp.valid_hooks != t->valid_hooks) { 968 if (tmp.valid_hooks != t->valid_hooks) {
922 duprintf("Valid hook crap: %08X vs %08X\n", 969 duprintf("Valid hook crap: %08X vs %08X\n",
923 tmp.valid_hooks, t->valid_hooks); 970 tmp.valid_hooks, t->valid_hooks);
924 ret = -EINVAL; 971 ret = -EINVAL;
925 goto free_newinfo_counters_untrans_unlock; 972 goto put_module;
926 }
927
928 /* Get a reference in advance, we're not allowed fail later */
929 if (!try_module_get(t->me)) {
930 ret = -EBUSY;
931 goto free_newinfo_counters_untrans_unlock;
932 } 973 }
933 974
934 oldinfo = replace_table(t, tmp.num_counters, newinfo, &ret); 975 oldinfo = replace_table(t, tmp.num_counters, newinfo, &ret);
@@ -959,7 +1000,6 @@ static int do_replace(void __user *user, unsigned int len)
959 1000
960 put_module: 1001 put_module:
961 module_put(t->me); 1002 module_put(t->me);
962 free_newinfo_counters_untrans_unlock:
963 up(&arpt_mutex); 1003 up(&arpt_mutex);
964 free_newinfo_counters_untrans: 1004 free_newinfo_counters_untrans:
965 ARPT_ENTRY_ITERATE(newinfo->entries, newinfo->size, cleanup_entry, NULL); 1005 ARPT_ENTRY_ITERATE(newinfo->entries, newinfo->size, cleanup_entry, NULL);
@@ -989,7 +1029,7 @@ static int do_add_counters(void __user *user, unsigned int len)
989 unsigned int i; 1029 unsigned int i;
990 struct arpt_counters_info tmp, *paddc; 1030 struct arpt_counters_info tmp, *paddc;
991 struct arpt_table *t; 1031 struct arpt_table *t;
992 int ret; 1032 int ret = 0;
993 1033
994 if (copy_from_user(&tmp, user, sizeof(tmp)) != 0) 1034 if (copy_from_user(&tmp, user, sizeof(tmp)) != 0)
995 return -EFAULT; 1035 return -EFAULT;
@@ -1006,9 +1046,11 @@ static int do_add_counters(void __user *user, unsigned int len)
1006 goto free; 1046 goto free;
1007 } 1047 }
1008 1048
1009 t = arpt_find_table_lock(tmp.name, &ret, &arpt_mutex); 1049 t = find_table_lock(tmp.name);
1010 if (!t) 1050 if (!t || IS_ERR(t)) {
1051 ret = t ? PTR_ERR(t) : -ENOENT;
1011 goto free; 1052 goto free;
1053 }
1012 1054
1013 write_lock_bh(&t->lock); 1055 write_lock_bh(&t->lock);
1014 if (t->private->number != paddc->num_counters) { 1056 if (t->private->number != paddc->num_counters) {
@@ -1025,6 +1067,7 @@ static int do_add_counters(void __user *user, unsigned int len)
1025 unlock_up_free: 1067 unlock_up_free:
1026 write_unlock_bh(&t->lock); 1068 write_unlock_bh(&t->lock);
1027 up(&arpt_mutex); 1069 up(&arpt_mutex);
1070 module_put(t->me);
1028 free: 1071 free:
1029 vfree(paddc); 1072 vfree(paddc);
1030 1073
@@ -1079,8 +1122,10 @@ static int do_arpt_get_ctl(struct sock *sk, int cmd, void __user *user, int *len
1079 break; 1122 break;
1080 } 1123 }
1081 name[ARPT_TABLE_MAXNAMELEN-1] = '\0'; 1124 name[ARPT_TABLE_MAXNAMELEN-1] = '\0';
1082 t = arpt_find_table_lock(name, &ret, &arpt_mutex); 1125
1083 if (t) { 1126 t = try_then_request_module(find_table_lock(name),
1127 "arptable_%s", name);
1128 if (t && !IS_ERR(t)) {
1084 struct arpt_getinfo info; 1129 struct arpt_getinfo info;
1085 1130
1086 info.valid_hooks = t->valid_hooks; 1131 info.valid_hooks = t->valid_hooks;
@@ -1096,9 +1141,10 @@ static int do_arpt_get_ctl(struct sock *sk, int cmd, void __user *user, int *len
1096 ret = -EFAULT; 1141 ret = -EFAULT;
1097 else 1142 else
1098 ret = 0; 1143 ret = 0;
1099
1100 up(&arpt_mutex); 1144 up(&arpt_mutex);
1101 } 1145 module_put(t->me);
1146 } else
1147 ret = t ? PTR_ERR(t) : -ENOENT;
1102 } 1148 }
1103 break; 1149 break;
1104 1150
@@ -1119,6 +1165,24 @@ static int do_arpt_get_ctl(struct sock *sk, int cmd, void __user *user, int *len
1119 break; 1165 break;
1120 } 1166 }
1121 1167
1168 case ARPT_SO_GET_REVISION_TARGET: {
1169 struct arpt_get_revision rev;
1170
1171 if (*len != sizeof(rev)) {
1172 ret = -EINVAL;
1173 break;
1174 }
1175 if (copy_from_user(&rev, user, sizeof(rev)) != 0) {
1176 ret = -EFAULT;
1177 break;
1178 }
1179
1180 try_then_request_module(find_revision(rev.name, rev.revision,
1181 target_revfn, &ret),
1182 "arpt_%s", rev.name);
1183 break;
1184 }
1185
1122 default: 1186 default:
1123 duprintf("do_arpt_get_ctl: unknown request %i\n", cmd); 1187 duprintf("do_arpt_get_ctl: unknown request %i\n", cmd);
1124 ret = -EINVAL; 1188 ret = -EINVAL;
@@ -1136,12 +1200,9 @@ int arpt_register_target(struct arpt_target *target)
1136 if (ret != 0) 1200 if (ret != 0)
1137 return ret; 1201 return ret;
1138 1202
1139 if (!list_named_insert(&arpt_target, target)) { 1203 list_add(&target->list, &arpt_target);
1140 duprintf("arpt_register_target: `%s' already in list!\n",
1141 target->name);
1142 ret = -EINVAL;
1143 }
1144 up(&arpt_mutex); 1204 up(&arpt_mutex);
1205
1145 return ret; 1206 return ret;
1146} 1207}
1147 1208
diff --git a/net/ipv4/netfilter/ipt_addrtype.c b/net/ipv4/netfilter/ipt_addrtype.c
index f5909a4c3fc7..e19c2a52d00c 100644
--- a/net/ipv4/netfilter/ipt_addrtype.c
+++ b/net/ipv4/netfilter/ipt_addrtype.c
@@ -48,7 +48,7 @@ static int checkentry(const char *tablename, const struct ipt_ip *ip,
48 unsigned int hook_mask) 48 unsigned int hook_mask)
49{ 49{
50 if (matchsize != IPT_ALIGN(sizeof(struct ipt_addrtype_info))) { 50 if (matchsize != IPT_ALIGN(sizeof(struct ipt_addrtype_info))) {
51 printk(KERN_ERR "ipt_addrtype: invalid size (%u != %Zu)\n.", 51 printk(KERN_ERR "ipt_addrtype: invalid size (%u != %Zu)\n",
52 matchsize, IPT_ALIGN(sizeof(struct ipt_addrtype_info))); 52 matchsize, IPT_ALIGN(sizeof(struct ipt_addrtype_info)));
53 return 0; 53 return 0;
54 } 54 }
diff --git a/net/ipv4/tcp_bic.c b/net/ipv4/tcp_bic.c
index 6d80e063c187..ae35e0609047 100644
--- a/net/ipv4/tcp_bic.c
+++ b/net/ipv4/tcp_bic.c
@@ -27,7 +27,7 @@
27 */ 27 */
28 28
29static int fast_convergence = 1; 29static int fast_convergence = 1;
30static int max_increment = 32; 30static int max_increment = 16;
31static int low_window = 14; 31static int low_window = 14;
32static int beta = 819; /* = 819/1024 (BICTCP_BETA_SCALE) */ 32static int beta = 819; /* = 819/1024 (BICTCP_BETA_SCALE) */
33static int low_utilization_threshold = 153; 33static int low_utilization_threshold = 153;
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index a970b4727ce8..2c5f57299d63 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -75,7 +75,7 @@
75#ifdef CONFIG_IPV6_PRIVACY 75#ifdef CONFIG_IPV6_PRIVACY
76#include <linux/random.h> 76#include <linux/random.h>
77#include <linux/crypto.h> 77#include <linux/crypto.h>
78#include <asm/scatterlist.h> 78#include <linux/scatterlist.h>
79#endif 79#endif
80 80
81#include <asm/uaccess.h> 81#include <asm/uaccess.h>
@@ -1217,12 +1217,8 @@ static int __ipv6_regen_rndid(struct inet6_dev *idev)
1217 struct net_device *dev; 1217 struct net_device *dev;
1218 struct scatterlist sg[2]; 1218 struct scatterlist sg[2];
1219 1219
1220 sg[0].page = virt_to_page(idev->entropy); 1220 sg_set_buf(&sg[0], idev->entropy, 8);
1221 sg[0].offset = offset_in_page(idev->entropy); 1221 sg_set_buf(&sg[1], idev->work_eui64, 8);
1222 sg[0].length = 8;
1223 sg[1].page = virt_to_page(idev->work_eui64);
1224 sg[1].offset = offset_in_page(idev->work_eui64);
1225 sg[1].length = 8;
1226 1222
1227 dev = idev->dev; 1223 dev = idev->dev;
1228 1224
@@ -2167,7 +2163,7 @@ static int addrconf_ifdown(struct net_device *dev, int how)
2167 2163
2168 /* Step 5: netlink notification of this interface */ 2164 /* Step 5: netlink notification of this interface */
2169 idev->tstamp = jiffies; 2165 idev->tstamp = jiffies;
2170 inet6_ifinfo_notify(RTM_NEWLINK, idev); 2166 inet6_ifinfo_notify(RTM_DELLINK, idev);
2171 2167
2172 /* Shot the device (if unregistered) */ 2168 /* Shot the device (if unregistered) */
2173 2169
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index 563b442ffab8..614296a920c6 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -147,7 +147,8 @@ static int ip6_output2(struct sk_buff *skb)
147 147
148int ip6_output(struct sk_buff *skb) 148int ip6_output(struct sk_buff *skb)
149{ 149{
150 if (skb->len > dst_mtu(skb->dst) || dst_allfrag(skb->dst)) 150 if ((skb->len > dst_mtu(skb->dst) && !skb_shinfo(skb)->ufo_size) ||
151 dst_allfrag(skb->dst))
151 return ip6_fragment(skb, ip6_output2); 152 return ip6_fragment(skb, ip6_output2);
152 else 153 else
153 return ip6_output2(skb); 154 return ip6_output2(skb);
@@ -768,6 +769,65 @@ out_err_release:
768 *dst = NULL; 769 *dst = NULL;
769 return err; 770 return err;
770} 771}
772inline int ip6_ufo_append_data(struct sock *sk,
773 int getfrag(void *from, char *to, int offset, int len,
774 int odd, struct sk_buff *skb),
775 void *from, int length, int hh_len, int fragheaderlen,
776 int transhdrlen, int mtu,unsigned int flags)
777
778{
779 struct sk_buff *skb;
780 int err;
781
782 /* There is support for UDP large send offload by network
783 * device, so create one single skb packet containing complete
784 * udp datagram
785 */
786 if ((skb = skb_peek_tail(&sk->sk_write_queue)) == NULL) {
787 skb = sock_alloc_send_skb(sk,
788 hh_len + fragheaderlen + transhdrlen + 20,
789 (flags & MSG_DONTWAIT), &err);
790 if (skb == NULL)
791 return -ENOMEM;
792
793 /* reserve space for Hardware header */
794 skb_reserve(skb, hh_len);
795
796 /* create space for UDP/IP header */
797 skb_put(skb,fragheaderlen + transhdrlen);
798
799 /* initialize network header pointer */
800 skb->nh.raw = skb->data;
801
802 /* initialize protocol header pointer */
803 skb->h.raw = skb->data + fragheaderlen;
804
805 skb->ip_summed = CHECKSUM_HW;
806 skb->csum = 0;
807 sk->sk_sndmsg_off = 0;
808 }
809
810 err = skb_append_datato_frags(sk,skb, getfrag, from,
811 (length - transhdrlen));
812 if (!err) {
813 struct frag_hdr fhdr;
814
815 /* specify the length of each IP datagram fragment*/
816 skb_shinfo(skb)->ufo_size = (mtu - fragheaderlen) -
817 sizeof(struct frag_hdr);
818 ipv6_select_ident(skb, &fhdr);
819 skb_shinfo(skb)->ip6_frag_id = fhdr.identification;
820 __skb_queue_tail(&sk->sk_write_queue, skb);
821
822 return 0;
823 }
824 /* There is not enough support do UPD LSO,
825 * so follow normal path
826 */
827 kfree_skb(skb);
828
829 return err;
830}
771 831
772int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to, 832int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to,
773 int offset, int len, int odd, struct sk_buff *skb), 833 int offset, int len, int odd, struct sk_buff *skb),
@@ -860,6 +920,15 @@ int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to,
860 */ 920 */
861 921
862 inet->cork.length += length; 922 inet->cork.length += length;
923 if (((length > mtu) && (sk->sk_protocol == IPPROTO_UDP)) &&
924 (rt->u.dst.dev->features & NETIF_F_UFO)) {
925
926 if(ip6_ufo_append_data(sk, getfrag, from, length, hh_len,
927 fragheaderlen, transhdrlen, mtu, flags))
928 goto error;
929
930 return 0;
931 }
863 932
864 if ((skb = skb_peek_tail(&sk->sk_write_queue)) == NULL) 933 if ((skb = skb_peek_tail(&sk->sk_write_queue)) == NULL)
865 goto alloc_new_skb; 934 goto alloc_new_skb;
diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c
index 39a96c768102..f15e04ad026e 100644
--- a/net/ipv6/mcast.c
+++ b/net/ipv6/mcast.c
@@ -164,7 +164,7 @@ static int ip6_mc_leave_src(struct sock *sk, struct ipv6_mc_socklist *iml,
164#define MLDV2_MASK(value, nb) ((nb)>=32 ? (value) : ((1<<(nb))-1) & (value)) 164#define MLDV2_MASK(value, nb) ((nb)>=32 ? (value) : ((1<<(nb))-1) & (value))
165#define MLDV2_EXP(thresh, nbmant, nbexp, value) \ 165#define MLDV2_EXP(thresh, nbmant, nbexp, value) \
166 ((value) < (thresh) ? (value) : \ 166 ((value) < (thresh) ? (value) : \
167 ((MLDV2_MASK(value, nbmant) | (1<<(nbmant+nbexp))) << \ 167 ((MLDV2_MASK(value, nbmant) | (1<<(nbmant))) << \
168 (MLDV2_MASK((value) >> (nbmant), nbexp) + (nbexp)))) 168 (MLDV2_MASK((value) >> (nbmant), nbexp) + (nbexp))))
169 169
170#define MLDV2_QQIC(value) MLDV2_EXP(0x80, 4, 3, value) 170#define MLDV2_QQIC(value) MLDV2_EXP(0x80, 4, 3, value)
@@ -545,8 +545,10 @@ int ip6_mc_msfilter(struct sock *sk, struct group_filter *gsf)
545 sock_kfree_s(sk, newpsl, IP6_SFLSIZE(newpsl->sl_max)); 545 sock_kfree_s(sk, newpsl, IP6_SFLSIZE(newpsl->sl_max));
546 goto done; 546 goto done;
547 } 547 }
548 } else 548 } else {
549 newpsl = NULL; 549 newpsl = NULL;
550 (void) ip6_mc_add_src(idev, group, gsf->gf_fmode, 0, NULL, 0);
551 }
550 psl = pmc->sflist; 552 psl = pmc->sflist;
551 if (psl) { 553 if (psl) {
552 (void) ip6_mc_del_src(idev, group, pmc->sfmode, 554 (void) ip6_mc_del_src(idev, group, pmc->sfmode,
@@ -1087,7 +1089,7 @@ static void mld_marksources(struct ifmcaddr6 *pmc, int nsrcs,
1087 1089
1088int igmp6_event_query(struct sk_buff *skb) 1090int igmp6_event_query(struct sk_buff *skb)
1089{ 1091{
1090 struct mld2_query *mlh2 = (struct mld2_query *) skb->h.raw; 1092 struct mld2_query *mlh2 = NULL;
1091 struct ifmcaddr6 *ma; 1093 struct ifmcaddr6 *ma;
1092 struct in6_addr *group; 1094 struct in6_addr *group;
1093 unsigned long max_delay; 1095 unsigned long max_delay;
@@ -1140,6 +1142,13 @@ int igmp6_event_query(struct sk_buff *skb)
1140 /* clear deleted report items */ 1142 /* clear deleted report items */
1141 mld_clear_delrec(idev); 1143 mld_clear_delrec(idev);
1142 } else if (len >= 28) { 1144 } else if (len >= 28) {
1145 int srcs_offset = sizeof(struct mld2_query) -
1146 sizeof(struct icmp6hdr);
1147 if (!pskb_may_pull(skb, srcs_offset)) {
1148 in6_dev_put(idev);
1149 return -EINVAL;
1150 }
1151 mlh2 = (struct mld2_query *) skb->h.raw;
1143 max_delay = (MLDV2_MRC(ntohs(mlh2->mrc))*HZ)/1000; 1152 max_delay = (MLDV2_MRC(ntohs(mlh2->mrc))*HZ)/1000;
1144 if (!max_delay) 1153 if (!max_delay)
1145 max_delay = 1; 1154 max_delay = 1;
@@ -1156,7 +1165,15 @@ int igmp6_event_query(struct sk_buff *skb)
1156 return 0; 1165 return 0;
1157 } 1166 }
1158 /* mark sources to include, if group & source-specific */ 1167 /* mark sources to include, if group & source-specific */
1159 mark = mlh2->nsrcs != 0; 1168 if (mlh2->nsrcs != 0) {
1169 if (!pskb_may_pull(skb, srcs_offset +
1170 mlh2->nsrcs * sizeof(struct in6_addr))) {
1171 in6_dev_put(idev);
1172 return -EINVAL;
1173 }
1174 mlh2 = (struct mld2_query *) skb->h.raw;
1175 mark = 1;
1176 }
1160 } else { 1177 } else {
1161 in6_dev_put(idev); 1178 in6_dev_put(idev);
1162 return -EINVAL; 1179 return -EINVAL;
diff --git a/net/ipv6/netfilter/ip6_tables.c b/net/ipv6/netfilter/ip6_tables.c
index 21deec25a12b..7d492226c16e 100644
--- a/net/ipv6/netfilter/ip6_tables.c
+++ b/net/ipv6/netfilter/ip6_tables.c
@@ -2,7 +2,7 @@
2 * Packet matching code. 2 * Packet matching code.
3 * 3 *
4 * Copyright (C) 1999 Paul `Rusty' Russell & Michael J. Neuling 4 * Copyright (C) 1999 Paul `Rusty' Russell & Michael J. Neuling
5 * Copyright (C) 2000-2002 Netfilter core team <coreteam@netfilter.org> 5 * Copyright (C) 2000-2005 Netfilter Core Team <coreteam@netfilter.org>
6 * 6 *
7 * This program is free software; you can redistribute it and/or modify 7 * This program is free software; you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License version 2 as 8 * it under the terms of the GNU General Public License version 2 as
@@ -23,7 +23,6 @@
23#include <linux/tcp.h> 23#include <linux/tcp.h>
24#include <linux/udp.h> 24#include <linux/udp.h>
25#include <linux/icmpv6.h> 25#include <linux/icmpv6.h>
26#include <net/ip.h>
27#include <net/ipv6.h> 26#include <net/ipv6.h>
28#include <asm/uaccess.h> 27#include <asm/uaccess.h>
29#include <asm/semaphore.h> 28#include <asm/semaphore.h>
@@ -80,13 +79,12 @@ static DECLARE_MUTEX(ip6t_mutex);
80#define inline 79#define inline
81#endif 80#endif
82 81
83/* Locking is simple: we assume at worst case there will be one packet 82/*
84 in user context and one from bottom halves (or soft irq if Alexey's
85 softnet patch was applied).
86
87 We keep a set of rules for each CPU, so we can avoid write-locking 83 We keep a set of rules for each CPU, so we can avoid write-locking
88 them; doing a readlock_bh() stops packets coming through if we're 84 them in the softirq when updating the counters and therefore
89 in user context. 85 only need to read-lock in the softirq; doing a write_lock_bh() in user
86 context stops packets coming through and allows user context to read
87 the counters or update the rules.
90 88
91 To be cache friendly on SMP, we arrange them like so: 89 To be cache friendly on SMP, we arrange them like so:
92 [ n-entries ] 90 [ n-entries ]
@@ -356,7 +354,7 @@ ip6t_do_table(struct sk_buff **pskb,
356 struct ip6t_table *table, 354 struct ip6t_table *table,
357 void *userdata) 355 void *userdata)
358{ 356{
359 static const char nulldevname[IFNAMSIZ]; 357 static const char nulldevname[IFNAMSIZ] __attribute__((aligned(sizeof(long))));
360 int offset = 0; 358 int offset = 0;
361 unsigned int protoff = 0; 359 unsigned int protoff = 0;
362 int hotdrop = 0; 360 int hotdrop = 0;
@@ -369,7 +367,6 @@ ip6t_do_table(struct sk_buff **pskb,
369 /* Initialization */ 367 /* Initialization */
370 indev = in ? in->name : nulldevname; 368 indev = in ? in->name : nulldevname;
371 outdev = out ? out->name : nulldevname; 369 outdev = out ? out->name : nulldevname;
372
373 /* We handle fragments by dealing with the first fragment as 370 /* We handle fragments by dealing with the first fragment as
374 * if it was a normal packet. All other fragments are treated 371 * if it was a normal packet. All other fragments are treated
375 * normally, except that they will NEVER match rules that ask 372 * normally, except that they will NEVER match rules that ask
@@ -497,75 +494,145 @@ ip6t_do_table(struct sk_buff **pskb,
497#endif 494#endif
498} 495}
499 496
500/* If it succeeds, returns element and locks mutex */ 497/*
501static inline void * 498 * These are weird, but module loading must not be done with mutex
502find_inlist_lock_noload(struct list_head *head, 499 * held (since they will register), and we have to have a single
503 const char *name, 500 * function to use try_then_request_module().
504 int *error, 501 */
505 struct semaphore *mutex) 502
503/* Find table by name, grabs mutex & ref. Returns ERR_PTR() on error. */
504static inline struct ip6t_table *find_table_lock(const char *name)
506{ 505{
507 void *ret; 506 struct ip6t_table *t;
508 507
509#if 1 508 if (down_interruptible(&ip6t_mutex) != 0)
510 duprintf("find_inlist: searching for `%s' in %s.\n", 509 return ERR_PTR(-EINTR);
511 name, head == &ip6t_target ? "ip6t_target"
512 : head == &ip6t_match ? "ip6t_match"
513 : head == &ip6t_tables ? "ip6t_tables" : "UNKNOWN");
514#endif
515 510
516 *error = down_interruptible(mutex); 511 list_for_each_entry(t, &ip6t_tables, list)
517 if (*error != 0) 512 if (strcmp(t->name, name) == 0 && try_module_get(t->me))
518 return NULL; 513 return t;
514 up(&ip6t_mutex);
515 return NULL;
516}
517
518/* Find match, grabs ref. Returns ERR_PTR() on error. */
519static inline struct ip6t_match *find_match(const char *name, u8 revision)
520{
521 struct ip6t_match *m;
522 int err = 0;
519 523
520 ret = list_named_find(head, name); 524 if (down_interruptible(&ip6t_mutex) != 0)
521 if (!ret) { 525 return ERR_PTR(-EINTR);
522 *error = -ENOENT; 526
523 up(mutex); 527 list_for_each_entry(m, &ip6t_match, list) {
528 if (strcmp(m->name, name) == 0) {
529 if (m->revision == revision) {
530 if (try_module_get(m->me)) {
531 up(&ip6t_mutex);
532 return m;
533 }
534 } else
535 err = -EPROTOTYPE; /* Found something. */
536 }
524 } 537 }
525 return ret; 538 up(&ip6t_mutex);
539 return ERR_PTR(err);
526} 540}
527 541
528#ifndef CONFIG_KMOD 542/* Find target, grabs ref. Returns ERR_PTR() on error. */
529#define find_inlist_lock(h,n,p,e,m) find_inlist_lock_noload((h),(n),(e),(m)) 543static inline struct ip6t_target *find_target(const char *name, u8 revision)
530#else
531static void *
532find_inlist_lock(struct list_head *head,
533 const char *name,
534 const char *prefix,
535 int *error,
536 struct semaphore *mutex)
537{ 544{
538 void *ret; 545 struct ip6t_target *t;
546 int err = 0;
539 547
540 ret = find_inlist_lock_noload(head, name, error, mutex); 548 if (down_interruptible(&ip6t_mutex) != 0)
541 if (!ret) { 549 return ERR_PTR(-EINTR);
542 duprintf("find_inlist: loading `%s%s'.\n", prefix, name); 550
543 request_module("%s%s", prefix, name); 551 list_for_each_entry(t, &ip6t_target, list) {
544 ret = find_inlist_lock_noload(head, name, error, mutex); 552 if (strcmp(t->name, name) == 0) {
553 if (t->revision == revision) {
554 if (try_module_get(t->me)) {
555 up(&ip6t_mutex);
556 return t;
557 }
558 } else
559 err = -EPROTOTYPE; /* Found something. */
560 }
545 } 561 }
562 up(&ip6t_mutex);
563 return ERR_PTR(err);
564}
546 565
547 return ret; 566struct ip6t_target *ip6t_find_target(const char *name, u8 revision)
567{
568 struct ip6t_target *target;
569
570 target = try_then_request_module(find_target(name, revision),
571 "ip6t_%s", name);
572 if (IS_ERR(target) || !target)
573 return NULL;
574 return target;
548} 575}
549#endif
550 576
551static inline struct ip6t_table * 577static int match_revfn(const char *name, u8 revision, int *bestp)
552ip6t_find_table_lock(const char *name, int *error, struct semaphore *mutex)
553{ 578{
554 return find_inlist_lock(&ip6t_tables, name, "ip6table_", error, mutex); 579 struct ip6t_match *m;
580 int have_rev = 0;
581
582 list_for_each_entry(m, &ip6t_match, list) {
583 if (strcmp(m->name, name) == 0) {
584 if (m->revision > *bestp)
585 *bestp = m->revision;
586 if (m->revision == revision)
587 have_rev = 1;
588 }
589 }
590 return have_rev;
555} 591}
556 592
557static inline struct ip6t_match * 593static int target_revfn(const char *name, u8 revision, int *bestp)
558find_match_lock(const char *name, int *error, struct semaphore *mutex)
559{ 594{
560 return find_inlist_lock(&ip6t_match, name, "ip6t_", error, mutex); 595 struct ip6t_target *t;
596 int have_rev = 0;
597
598 list_for_each_entry(t, &ip6t_target, list) {
599 if (strcmp(t->name, name) == 0) {
600 if (t->revision > *bestp)
601 *bestp = t->revision;
602 if (t->revision == revision)
603 have_rev = 1;
604 }
605 }
606 return have_rev;
561} 607}
562 608
563static struct ip6t_target * 609/* Returns true or fals (if no such extension at all) */
564ip6t_find_target_lock(const char *name, int *error, struct semaphore *mutex) 610static inline int find_revision(const char *name, u8 revision,
611 int (*revfn)(const char *, u8, int *),
612 int *err)
565{ 613{
566 return find_inlist_lock(&ip6t_target, name, "ip6t_", error, mutex); 614 int have_rev, best = -1;
615
616 if (down_interruptible(&ip6t_mutex) != 0) {
617 *err = -EINTR;
618 return 1;
619 }
620 have_rev = revfn(name, revision, &best);
621 up(&ip6t_mutex);
622
623 /* Nothing at all? Return 0 to try loading module. */
624 if (best == -1) {
625 *err = -ENOENT;
626 return 0;
627 }
628
629 *err = best;
630 if (!have_rev)
631 *err = -EPROTONOSUPPORT;
632 return 1;
567} 633}
568 634
635
569/* All zeroes == unconditional rule. */ 636/* All zeroes == unconditional rule. */
570static inline int 637static inline int
571unconditional(const struct ip6t_ip6 *ipv6) 638unconditional(const struct ip6t_ip6 *ipv6)
@@ -725,20 +792,16 @@ check_match(struct ip6t_entry_match *m,
725 unsigned int hookmask, 792 unsigned int hookmask,
726 unsigned int *i) 793 unsigned int *i)
727{ 794{
728 int ret;
729 struct ip6t_match *match; 795 struct ip6t_match *match;
730 796
731 match = find_match_lock(m->u.user.name, &ret, &ip6t_mutex); 797 match = try_then_request_module(find_match(m->u.user.name,
732 if (!match) { 798 m->u.user.revision),
733 // duprintf("check_match: `%s' not found\n", m->u.name); 799 "ip6t_%s", m->u.user.name);
734 return ret; 800 if (IS_ERR(match) || !match) {
735 } 801 duprintf("check_match: `%s' not found\n", m->u.user.name);
736 if (!try_module_get(match->me)) { 802 return match ? PTR_ERR(match) : -ENOENT;
737 up(&ip6t_mutex);
738 return -ENOENT;
739 } 803 }
740 m->u.kernel.match = match; 804 m->u.kernel.match = match;
741 up(&ip6t_mutex);
742 805
743 if (m->u.kernel.match->checkentry 806 if (m->u.kernel.match->checkentry
744 && !m->u.kernel.match->checkentry(name, ipv6, m->data, 807 && !m->u.kernel.match->checkentry(name, ipv6, m->data,
@@ -776,22 +839,16 @@ check_entry(struct ip6t_entry *e, const char *name, unsigned int size,
776 goto cleanup_matches; 839 goto cleanup_matches;
777 840
778 t = ip6t_get_target(e); 841 t = ip6t_get_target(e);
779 target = ip6t_find_target_lock(t->u.user.name, &ret, &ip6t_mutex); 842 target = try_then_request_module(find_target(t->u.user.name,
780 if (!target) { 843 t->u.user.revision),
844 "ip6t_%s", t->u.user.name);
845 if (IS_ERR(target) || !target) {
781 duprintf("check_entry: `%s' not found\n", t->u.user.name); 846 duprintf("check_entry: `%s' not found\n", t->u.user.name);
782 goto cleanup_matches; 847 ret = target ? PTR_ERR(target) : -ENOENT;
783 }
784 if (!try_module_get(target->me)) {
785 up(&ip6t_mutex);
786 ret = -ENOENT;
787 goto cleanup_matches; 848 goto cleanup_matches;
788 } 849 }
789 t->u.kernel.target = target; 850 t->u.kernel.target = target;
790 up(&ip6t_mutex); 851
791 if (!t->u.kernel.target) {
792 ret = -EBUSY;
793 goto cleanup_matches;
794 }
795 if (t->u.kernel.target == &ip6t_standard_target) { 852 if (t->u.kernel.target == &ip6t_standard_target) {
796 if (!standard_check(t, size)) { 853 if (!standard_check(t, size)) {
797 ret = -EINVAL; 854 ret = -EINVAL;
@@ -1118,8 +1175,8 @@ get_entries(const struct ip6t_get_entries *entries,
1118 int ret; 1175 int ret;
1119 struct ip6t_table *t; 1176 struct ip6t_table *t;
1120 1177
1121 t = ip6t_find_table_lock(entries->name, &ret, &ip6t_mutex); 1178 t = find_table_lock(entries->name);
1122 if (t) { 1179 if (t && !IS_ERR(t)) {
1123 duprintf("t->private->number = %u\n", 1180 duprintf("t->private->number = %u\n",
1124 t->private->number); 1181 t->private->number);
1125 if (entries->size == t->private->size) 1182 if (entries->size == t->private->size)
@@ -1131,10 +1188,10 @@ get_entries(const struct ip6t_get_entries *entries,
1131 entries->size); 1188 entries->size);
1132 ret = -EINVAL; 1189 ret = -EINVAL;
1133 } 1190 }
1191 module_put(t->me);
1134 up(&ip6t_mutex); 1192 up(&ip6t_mutex);
1135 } else 1193 } else
1136 duprintf("get_entries: Can't find %s!\n", 1194 ret = t ? PTR_ERR(t) : -ENOENT;
1137 entries->name);
1138 1195
1139 return ret; 1196 return ret;
1140} 1197}
@@ -1182,22 +1239,19 @@ do_replace(void __user *user, unsigned int len)
1182 1239
1183 duprintf("ip_tables: Translated table\n"); 1240 duprintf("ip_tables: Translated table\n");
1184 1241
1185 t = ip6t_find_table_lock(tmp.name, &ret, &ip6t_mutex); 1242 t = try_then_request_module(find_table_lock(tmp.name),
1186 if (!t) 1243 "ip6table_%s", tmp.name);
1244 if (!t || IS_ERR(t)) {
1245 ret = t ? PTR_ERR(t) : -ENOENT;
1187 goto free_newinfo_counters_untrans; 1246 goto free_newinfo_counters_untrans;
1247 }
1188 1248
1189 /* You lied! */ 1249 /* You lied! */
1190 if (tmp.valid_hooks != t->valid_hooks) { 1250 if (tmp.valid_hooks != t->valid_hooks) {
1191 duprintf("Valid hook crap: %08X vs %08X\n", 1251 duprintf("Valid hook crap: %08X vs %08X\n",
1192 tmp.valid_hooks, t->valid_hooks); 1252 tmp.valid_hooks, t->valid_hooks);
1193 ret = -EINVAL; 1253 ret = -EINVAL;
1194 goto free_newinfo_counters_untrans_unlock; 1254 goto put_module;
1195 }
1196
1197 /* Get a reference in advance, we're not allowed fail later */
1198 if (!try_module_get(t->me)) {
1199 ret = -EBUSY;
1200 goto free_newinfo_counters_untrans_unlock;
1201 } 1255 }
1202 1256
1203 oldinfo = replace_table(t, tmp.num_counters, newinfo, &ret); 1257 oldinfo = replace_table(t, tmp.num_counters, newinfo, &ret);
@@ -1219,7 +1273,6 @@ do_replace(void __user *user, unsigned int len)
1219 /* Decrease module usage counts and free resource */ 1273 /* Decrease module usage counts and free resource */
1220 IP6T_ENTRY_ITERATE(oldinfo->entries, oldinfo->size, cleanup_entry,NULL); 1274 IP6T_ENTRY_ITERATE(oldinfo->entries, oldinfo->size, cleanup_entry,NULL);
1221 vfree(oldinfo); 1275 vfree(oldinfo);
1222 /* Silent error: too late now. */
1223 if (copy_to_user(tmp.counters, counters, 1276 if (copy_to_user(tmp.counters, counters,
1224 sizeof(struct ip6t_counters) * tmp.num_counters) != 0) 1277 sizeof(struct ip6t_counters) * tmp.num_counters) != 0)
1225 ret = -EFAULT; 1278 ret = -EFAULT;
@@ -1229,7 +1282,6 @@ do_replace(void __user *user, unsigned int len)
1229 1282
1230 put_module: 1283 put_module:
1231 module_put(t->me); 1284 module_put(t->me);
1232 free_newinfo_counters_untrans_unlock:
1233 up(&ip6t_mutex); 1285 up(&ip6t_mutex);
1234 free_newinfo_counters_untrans: 1286 free_newinfo_counters_untrans:
1235 IP6T_ENTRY_ITERATE(newinfo->entries, newinfo->size, cleanup_entry,NULL); 1287 IP6T_ENTRY_ITERATE(newinfo->entries, newinfo->size, cleanup_entry,NULL);
@@ -1268,7 +1320,7 @@ do_add_counters(void __user *user, unsigned int len)
1268 unsigned int i; 1320 unsigned int i;
1269 struct ip6t_counters_info tmp, *paddc; 1321 struct ip6t_counters_info tmp, *paddc;
1270 struct ip6t_table *t; 1322 struct ip6t_table *t;
1271 int ret; 1323 int ret = 0;
1272 1324
1273 if (copy_from_user(&tmp, user, sizeof(tmp)) != 0) 1325 if (copy_from_user(&tmp, user, sizeof(tmp)) != 0)
1274 return -EFAULT; 1326 return -EFAULT;
@@ -1285,9 +1337,11 @@ do_add_counters(void __user *user, unsigned int len)
1285 goto free; 1337 goto free;
1286 } 1338 }
1287 1339
1288 t = ip6t_find_table_lock(tmp.name, &ret, &ip6t_mutex); 1340 t = find_table_lock(tmp.name);
1289 if (!t) 1341 if (!t || IS_ERR(t)) {
1342 ret = t ? PTR_ERR(t) : -ENOENT;
1290 goto free; 1343 goto free;
1344 }
1291 1345
1292 write_lock_bh(&t->lock); 1346 write_lock_bh(&t->lock);
1293 if (t->private->number != paddc->num_counters) { 1347 if (t->private->number != paddc->num_counters) {
@@ -1304,6 +1358,7 @@ do_add_counters(void __user *user, unsigned int len)
1304 unlock_up_free: 1358 unlock_up_free:
1305 write_unlock_bh(&t->lock); 1359 write_unlock_bh(&t->lock);
1306 up(&ip6t_mutex); 1360 up(&ip6t_mutex);
1361 module_put(t->me);
1307 free: 1362 free:
1308 vfree(paddc); 1363 vfree(paddc);
1309 1364
@@ -1360,8 +1415,10 @@ do_ip6t_get_ctl(struct sock *sk, int cmd, void __user *user, int *len)
1360 break; 1415 break;
1361 } 1416 }
1362 name[IP6T_TABLE_MAXNAMELEN-1] = '\0'; 1417 name[IP6T_TABLE_MAXNAMELEN-1] = '\0';
1363 t = ip6t_find_table_lock(name, &ret, &ip6t_mutex); 1418
1364 if (t) { 1419 t = try_then_request_module(find_table_lock(name),
1420 "ip6table_%s", name);
1421 if (t && !IS_ERR(t)) {
1365 struct ip6t_getinfo info; 1422 struct ip6t_getinfo info;
1366 1423
1367 info.valid_hooks = t->valid_hooks; 1424 info.valid_hooks = t->valid_hooks;
@@ -1377,9 +1434,10 @@ do_ip6t_get_ctl(struct sock *sk, int cmd, void __user *user, int *len)
1377 ret = -EFAULT; 1434 ret = -EFAULT;
1378 else 1435 else
1379 ret = 0; 1436 ret = 0;
1380
1381 up(&ip6t_mutex); 1437 up(&ip6t_mutex);
1382 } 1438 module_put(t->me);
1439 } else
1440 ret = t ? PTR_ERR(t) : -ENOENT;
1383 } 1441 }
1384 break; 1442 break;
1385 1443
@@ -1400,6 +1458,31 @@ do_ip6t_get_ctl(struct sock *sk, int cmd, void __user *user, int *len)
1400 break; 1458 break;
1401 } 1459 }
1402 1460
1461 case IP6T_SO_GET_REVISION_MATCH:
1462 case IP6T_SO_GET_REVISION_TARGET: {
1463 struct ip6t_get_revision rev;
1464 int (*revfn)(const char *, u8, int *);
1465
1466 if (*len != sizeof(rev)) {
1467 ret = -EINVAL;
1468 break;
1469 }
1470 if (copy_from_user(&rev, user, sizeof(rev)) != 0) {
1471 ret = -EFAULT;
1472 break;
1473 }
1474
1475 if (cmd == IP6T_SO_GET_REVISION_TARGET)
1476 revfn = target_revfn;
1477 else
1478 revfn = match_revfn;
1479
1480 try_then_request_module(find_revision(rev.name, rev.revision,
1481 revfn, &ret),
1482 "ip6t_%s", rev.name);
1483 break;
1484 }
1485
1403 default: 1486 default:
1404 duprintf("do_ip6t_get_ctl: unknown request %i\n", cmd); 1487 duprintf("do_ip6t_get_ctl: unknown request %i\n", cmd);
1405 ret = -EINVAL; 1488 ret = -EINVAL;
@@ -1417,12 +1500,7 @@ ip6t_register_target(struct ip6t_target *target)
1417 ret = down_interruptible(&ip6t_mutex); 1500 ret = down_interruptible(&ip6t_mutex);
1418 if (ret != 0) 1501 if (ret != 0)
1419 return ret; 1502 return ret;
1420 1503 list_add(&target->list, &ip6t_target);
1421 if (!list_named_insert(&ip6t_target, target)) {
1422 duprintf("ip6t_register_target: `%s' already in list!\n",
1423 target->name);
1424 ret = -EINVAL;
1425 }
1426 up(&ip6t_mutex); 1504 up(&ip6t_mutex);
1427 return ret; 1505 return ret;
1428} 1506}
@@ -1444,11 +1522,7 @@ ip6t_register_match(struct ip6t_match *match)
1444 if (ret != 0) 1522 if (ret != 0)
1445 return ret; 1523 return ret;
1446 1524
1447 if (!list_named_insert(&ip6t_match, match)) { 1525 list_add(&match->list, &ip6t_match);
1448 duprintf("ip6t_register_match: `%s' already in list!\n",
1449 match->name);
1450 ret = -EINVAL;
1451 }
1452 up(&ip6t_mutex); 1526 up(&ip6t_mutex);
1453 1527
1454 return ret; 1528 return ret;
diff --git a/net/ipv6/netfilter/ip6t_MARK.c b/net/ipv6/netfilter/ip6t_MARK.c
index 81924fcc5857..0c7584f92172 100644
--- a/net/ipv6/netfilter/ip6t_MARK.c
+++ b/net/ipv6/netfilter/ip6t_MARK.c
@@ -56,8 +56,12 @@ checkentry(const char *tablename,
56 return 1; 56 return 1;
57} 57}
58 58
59static struct ip6t_target ip6t_mark_reg 59static struct ip6t_target ip6t_mark_reg = {
60= { { NULL, NULL }, "MARK", target, checkentry, NULL, THIS_MODULE }; 60 .name = "MARK",
61 .target = target,
62 .checkentry = checkentry,
63 .me = THIS_MODULE
64};
61 65
62static int __init init(void) 66static int __init init(void)
63{ 67{
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 5d5bbb49ec78..227e99ed510c 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -483,7 +483,7 @@ restart:
483 goto out; 483 goto out;
484 } 484 }
485 485
486 rt = rt6_device_match(rt, skb->dev->ifindex, 0); 486 rt = rt6_device_match(rt, skb->dev->ifindex, strict);
487 BACKTRACK(); 487 BACKTRACK();
488 488
489 if (!rt->rt6i_nexthop && !(rt->rt6i_flags & RTF_NONEXTHOP)) { 489 if (!rt->rt6i_nexthop && !(rt->rt6i_flags & RTF_NONEXTHOP)) {
diff --git a/net/rose/rose_timer.c b/net/rose/rose_timer.c
index 50ae0371dab8..b6c8f38cc26c 100644
--- a/net/rose/rose_timer.c
+++ b/net/rose/rose_timer.c
@@ -138,6 +138,7 @@ static void rose_heartbeat_expiry(unsigned long param)
138 is accepted() it isn't 'dead' so doesn't get removed. */ 138 is accepted() it isn't 'dead' so doesn't get removed. */
139 if (sock_flag(sk, SOCK_DESTROY) || 139 if (sock_flag(sk, SOCK_DESTROY) ||
140 (sk->sk_state == TCP_LISTEN && sock_flag(sk, SOCK_DEAD))) { 140 (sk->sk_state == TCP_LISTEN && sock_flag(sk, SOCK_DEAD))) {
141 bh_unlock_sock(sk);
141 rose_destroy_socket(sk); 142 rose_destroy_socket(sk);
142 return; 143 return;
143 } 144 }
diff --git a/net/sched/Kconfig b/net/sched/Kconfig
index 81510da31792..7f34e7fd767c 100644
--- a/net/sched/Kconfig
+++ b/net/sched/Kconfig
@@ -2,13 +2,15 @@
2# Traffic control configuration. 2# Traffic control configuration.
3# 3#
4 4
5menuconfig NET_SCHED 5menu "QoS and/or fair queueing"
6
7config NET_SCHED
6 bool "QoS and/or fair queueing" 8 bool "QoS and/or fair queueing"
7 ---help--- 9 ---help---
8 When the kernel has several packets to send out over a network 10 When the kernel has several packets to send out over a network
9 device, it has to decide which ones to send first, which ones to 11 device, it has to decide which ones to send first, which ones to
10 delay, and which ones to drop. This is the job of the packet 12 delay, and which ones to drop. This is the job of the queueing
11 scheduler, and several different algorithms for how to do this 13 disciplines, several different algorithms for how to do this
12 "fairly" have been proposed. 14 "fairly" have been proposed.
13 15
14 If you say N here, you will get the standard packet scheduler, which 16 If you say N here, you will get the standard packet scheduler, which
@@ -23,13 +25,13 @@ menuconfig NET_SCHED
23 To administer these schedulers, you'll need the user-level utilities 25 To administer these schedulers, you'll need the user-level utilities
24 from the package iproute2+tc at <ftp://ftp.tux.org/pub/net/ip-routing/>. 26 from the package iproute2+tc at <ftp://ftp.tux.org/pub/net/ip-routing/>.
25 That package also contains some documentation; for more, check out 27 That package also contains some documentation; for more, check out
26 <http://snafu.freedom.org/linux2.2/iproute-notes.html>. 28 <http://linux-net.osdl.org/index.php/Iproute2>.
27 29
28 This Quality of Service (QoS) support will enable you to use 30 This Quality of Service (QoS) support will enable you to use
29 Differentiated Services (diffserv) and Resource Reservation Protocol 31 Differentiated Services (diffserv) and Resource Reservation Protocol
30 (RSVP) on your Linux router if you also say Y to "QoS support", 32 (RSVP) on your Linux router if you also say Y to the corresponding
31 "Packet classifier API" and to some classifiers below. Documentation 33 classifiers below. Documentation and software is at
32 and software is at <http://diffserv.sourceforge.net/>. 34 <http://diffserv.sourceforge.net/>.
33 35
34 If you say Y here and to "/proc file system" below, you will be able 36 If you say Y here and to "/proc file system" below, you will be able
35 to read status information about packet schedulers from the file 37 to read status information about packet schedulers from the file
@@ -42,7 +44,7 @@ choice
42 prompt "Packet scheduler clock source" 44 prompt "Packet scheduler clock source"
43 depends on NET_SCHED 45 depends on NET_SCHED
44 default NET_SCH_CLK_JIFFIES 46 default NET_SCH_CLK_JIFFIES
45 help 47 ---help---
46 Packet schedulers need a monotonic clock that increments at a static 48 Packet schedulers need a monotonic clock that increments at a static
47 rate. The kernel provides several suitable interfaces, each with 49 rate. The kernel provides several suitable interfaces, each with
48 different properties: 50 different properties:
@@ -56,7 +58,7 @@ choice
56 58
57config NET_SCH_CLK_JIFFIES 59config NET_SCH_CLK_JIFFIES
58 bool "Timer interrupt" 60 bool "Timer interrupt"
59 help 61 ---help---
60 Say Y here if you want to use the timer interrupt (jiffies) as clock 62 Say Y here if you want to use the timer interrupt (jiffies) as clock
61 source. This clock source is fast, synchronized on all processors and 63 source. This clock source is fast, synchronized on all processors and
62 handles cpu clock frequency changes, but its resolution is too low 64 handles cpu clock frequency changes, but its resolution is too low
@@ -64,7 +66,7 @@ config NET_SCH_CLK_JIFFIES
64 66
65config NET_SCH_CLK_GETTIMEOFDAY 67config NET_SCH_CLK_GETTIMEOFDAY
66 bool "gettimeofday" 68 bool "gettimeofday"
67 help 69 ---help---
68 Say Y here if you want to use gettimeofday as clock source. This clock 70 Say Y here if you want to use gettimeofday as clock source. This clock
69 source has high resolution, is synchronized on all processors and 71 source has high resolution, is synchronized on all processors and
70 handles cpu clock frequency changes, but it is slow. 72 handles cpu clock frequency changes, but it is slow.
@@ -77,7 +79,7 @@ config NET_SCH_CLK_GETTIMEOFDAY
77config NET_SCH_CLK_CPU 79config NET_SCH_CLK_CPU
78 bool "CPU cycle counter" 80 bool "CPU cycle counter"
79 depends on ((X86_TSC || X86_64) && !SMP) || ALPHA || SPARC64 || PPC64 || IA64 81 depends on ((X86_TSC || X86_64) && !SMP) || ALPHA || SPARC64 || PPC64 || IA64
80 help 82 ---help---
81 Say Y here if you want to use the CPU's cycle counter as clock source. 83 Say Y here if you want to use the CPU's cycle counter as clock source.
82 This is a cheap and high resolution clock source, but on some 84 This is a cheap and high resolution clock source, but on some
83 architectures it is not synchronized on all processors and doesn't 85 architectures it is not synchronized on all processors and doesn't
@@ -95,134 +97,129 @@ config NET_SCH_CLK_CPU
95 97
96endchoice 98endchoice
97 99
100comment "Queueing/Scheduling"
101 depends on NET_SCHED
102
98config NET_SCH_CBQ 103config NET_SCH_CBQ
99 tristate "CBQ packet scheduler" 104 tristate "Class Based Queueing (CBQ)"
100 depends on NET_SCHED 105 depends on NET_SCHED
101 ---help--- 106 ---help---
102 Say Y here if you want to use the Class-Based Queueing (CBQ) packet 107 Say Y here if you want to use the Class-Based Queueing (CBQ) packet
103 scheduling algorithm for some of your network devices. This 108 scheduling algorithm. This algorithm classifies the waiting packets
104 algorithm classifies the waiting packets into a tree-like hierarchy 109 into a tree-like hierarchy of classes; the leaves of this tree are
105 of classes; the leaves of this tree are in turn scheduled by 110 in turn scheduled by separate algorithms.
106 separate algorithms (called "disciplines" in this context).
107 111
108 See the top of <file:net/sched/sch_cbq.c> for references about the 112 See the top of <file:net/sched/sch_cbq.c> for more details.
109 CBQ algorithm.
110 113
111 CBQ is a commonly used scheduler, so if you're unsure, you should 114 CBQ is a commonly used scheduler, so if you're unsure, you should
112 say Y here. Then say Y to all the queueing algorithms below that you 115 say Y here. Then say Y to all the queueing algorithms below that you
113 want to use as CBQ disciplines. Then say Y to "Packet classifier 116 want to use as leaf disciplines.
114 API" and say Y to all the classifiers you want to use; a classifier
115 is a routine that allows you to sort your outgoing traffic into
116 classes based on a certain criterion.
117 117
118 To compile this code as a module, choose M here: the 118 To compile this code as a module, choose M here: the
119 module will be called sch_cbq. 119 module will be called sch_cbq.
120 120
121config NET_SCH_HTB 121config NET_SCH_HTB
122 tristate "HTB packet scheduler" 122 tristate "Hierarchical Token Bucket (HTB)"
123 depends on NET_SCHED 123 depends on NET_SCHED
124 ---help--- 124 ---help---
125 Say Y here if you want to use the Hierarchical Token Buckets (HTB) 125 Say Y here if you want to use the Hierarchical Token Buckets (HTB)
126 packet scheduling algorithm for some of your network devices. See 126 packet scheduling algorithm. See
127 <http://luxik.cdi.cz/~devik/qos/htb/> for complete manual and 127 <http://luxik.cdi.cz/~devik/qos/htb/> for complete manual and
128 in-depth articles. 128 in-depth articles.
129 129
130 HTB is very similar to the CBQ regarding its goals however is has 130 HTB is very similar to CBQ regarding its goals however is has
131 different properties and different algorithm. 131 different properties and different algorithm.
132 132
133 To compile this code as a module, choose M here: the 133 To compile this code as a module, choose M here: the
134 module will be called sch_htb. 134 module will be called sch_htb.
135 135
136config NET_SCH_HFSC 136config NET_SCH_HFSC
137 tristate "HFSC packet scheduler" 137 tristate "Hierarchical Fair Service Curve (HFSC)"
138 depends on NET_SCHED 138 depends on NET_SCHED
139 ---help--- 139 ---help---
140 Say Y here if you want to use the Hierarchical Fair Service Curve 140 Say Y here if you want to use the Hierarchical Fair Service Curve
141 (HFSC) packet scheduling algorithm for some of your network devices. 141 (HFSC) packet scheduling algorithm.
142 142
143 To compile this code as a module, choose M here: the 143 To compile this code as a module, choose M here: the
144 module will be called sch_hfsc. 144 module will be called sch_hfsc.
145 145
146#tristate ' H-PFQ packet scheduler' CONFIG_NET_SCH_HPFQ
147config NET_SCH_ATM 146config NET_SCH_ATM
148 tristate "ATM pseudo-scheduler" 147 tristate "ATM Virtual Circuits (ATM)"
149 depends on NET_SCHED && ATM 148 depends on NET_SCHED && ATM
150 ---help--- 149 ---help---
151 Say Y here if you want to use the ATM pseudo-scheduler. This 150 Say Y here if you want to use the ATM pseudo-scheduler. This
152 provides a framework for invoking classifiers (aka "filters"), which 151 provides a framework for invoking classifiers, which in turn
153 in turn select classes of this queuing discipline. Each class maps 152 select classes of this queuing discipline. Each class maps
154 the flow(s) it is handling to a given virtual circuit (see the top of 153 the flow(s) it is handling to a given virtual circuit.
155 <file:net/sched/sch_atm.c>). 154
155 See the top of <file:net/sched/sch_atm.c>) for more details.
156 156
157 To compile this code as a module, choose M here: the 157 To compile this code as a module, choose M here: the
158 module will be called sch_atm. 158 module will be called sch_atm.
159 159
160config NET_SCH_PRIO 160config NET_SCH_PRIO
161 tristate "The simplest PRIO pseudoscheduler" 161 tristate "Multi Band Priority Queueing (PRIO)"
162 depends on NET_SCHED 162 depends on NET_SCHED
163 help 163 ---help---
164 Say Y here if you want to use an n-band priority queue packet 164 Say Y here if you want to use an n-band priority queue packet
165 "scheduler" for some of your network devices or as a leaf discipline 165 scheduler.
166 for the CBQ scheduling algorithm. If unsure, say Y.
167 166
168 To compile this code as a module, choose M here: the 167 To compile this code as a module, choose M here: the
169 module will be called sch_prio. 168 module will be called sch_prio.
170 169
171config NET_SCH_RED 170config NET_SCH_RED
172 tristate "RED queue" 171 tristate "Random Early Detection (RED)"
173 depends on NET_SCHED 172 depends on NET_SCHED
174 help 173 ---help---
175 Say Y here if you want to use the Random Early Detection (RED) 174 Say Y here if you want to use the Random Early Detection (RED)
176 packet scheduling algorithm for some of your network devices (see 175 packet scheduling algorithm.
177 the top of <file:net/sched/sch_red.c> for details and references 176
178 about the algorithm). 177 See the top of <file:net/sched/sch_red.c> for more details.
179 178
180 To compile this code as a module, choose M here: the 179 To compile this code as a module, choose M here: the
181 module will be called sch_red. 180 module will be called sch_red.
182 181
183config NET_SCH_SFQ 182config NET_SCH_SFQ
184 tristate "SFQ queue" 183 tristate "Stochastic Fairness Queueing (SFQ)"
185 depends on NET_SCHED 184 depends on NET_SCHED
186 ---help--- 185 ---help---
187 Say Y here if you want to use the Stochastic Fairness Queueing (SFQ) 186 Say Y here if you want to use the Stochastic Fairness Queueing (SFQ)
188 packet scheduling algorithm for some of your network devices or as a 187 packet scheduling algorithm .
189 leaf discipline for the CBQ scheduling algorithm (see the top of 188
190 <file:net/sched/sch_sfq.c> for details and references about the SFQ 189 See the top of <file:net/sched/sch_sfq.c> for more details.
191 algorithm).
192 190
193 To compile this code as a module, choose M here: the 191 To compile this code as a module, choose M here: the
194 module will be called sch_sfq. 192 module will be called sch_sfq.
195 193
196config NET_SCH_TEQL 194config NET_SCH_TEQL
197 tristate "TEQL queue" 195 tristate "True Link Equalizer (TEQL)"
198 depends on NET_SCHED 196 depends on NET_SCHED
199 ---help--- 197 ---help---
200 Say Y here if you want to use the True Link Equalizer (TLE) packet 198 Say Y here if you want to use the True Link Equalizer (TLE) packet
201 scheduling algorithm for some of your network devices or as a leaf 199 scheduling algorithm. This queueing discipline allows the combination
202 discipline for the CBQ scheduling algorithm. This queueing 200 of several physical devices into one virtual device.
203 discipline allows the combination of several physical devices into 201
204 one virtual device. (see the top of <file:net/sched/sch_teql.c> for 202 See the top of <file:net/sched/sch_teql.c> for more details.
205 details).
206 203
207 To compile this code as a module, choose M here: the 204 To compile this code as a module, choose M here: the
208 module will be called sch_teql. 205 module will be called sch_teql.
209 206
210config NET_SCH_TBF 207config NET_SCH_TBF
211 tristate "TBF queue" 208 tristate "Token Bucket Filter (TBF)"
212 depends on NET_SCHED 209 depends on NET_SCHED
213 help 210 ---help---
214 Say Y here if you want to use the Simple Token Bucket Filter (TBF) 211 Say Y here if you want to use the Token Bucket Filter (TBF) packet
215 packet scheduling algorithm for some of your network devices or as a 212 scheduling algorithm.
216 leaf discipline for the CBQ scheduling algorithm (see the top of 213
217 <file:net/sched/sch_tbf.c> for a description of the TBF algorithm). 214 See the top of <file:net/sched/sch_tbf.c> for more details.
218 215
219 To compile this code as a module, choose M here: the 216 To compile this code as a module, choose M here: the
220 module will be called sch_tbf. 217 module will be called sch_tbf.
221 218
222config NET_SCH_GRED 219config NET_SCH_GRED
223 tristate "GRED queue" 220 tristate "Generic Random Early Detection (GRED)"
224 depends on NET_SCHED 221 depends on NET_SCHED
225 help 222 ---help---
226 Say Y here if you want to use the Generic Random Early Detection 223 Say Y here if you want to use the Generic Random Early Detection
227 (GRED) packet scheduling algorithm for some of your network devices 224 (GRED) packet scheduling algorithm for some of your network devices
228 (see the top of <file:net/sched/sch_red.c> for details and 225 (see the top of <file:net/sched/sch_red.c> for details and
@@ -232,9 +229,9 @@ config NET_SCH_GRED
232 module will be called sch_gred. 229 module will be called sch_gred.
233 230
234config NET_SCH_DSMARK 231config NET_SCH_DSMARK
235 tristate "Diffserv field marker" 232 tristate "Differentiated Services marker (DSMARK)"
236 depends on NET_SCHED 233 depends on NET_SCHED
237 help 234 ---help---
238 Say Y if you want to schedule packets according to the 235 Say Y if you want to schedule packets according to the
239 Differentiated Services architecture proposed in RFC 2475. 236 Differentiated Services architecture proposed in RFC 2475.
240 Technical information on this method, with pointers to associated 237 Technical information on this method, with pointers to associated
@@ -244,9 +241,9 @@ config NET_SCH_DSMARK
244 module will be called sch_dsmark. 241 module will be called sch_dsmark.
245 242
246config NET_SCH_NETEM 243config NET_SCH_NETEM
247 tristate "Network emulator" 244 tristate "Network emulator (NETEM)"
248 depends on NET_SCHED 245 depends on NET_SCHED
249 help 246 ---help---
250 Say Y if you want to emulate network delay, loss, and packet 247 Say Y if you want to emulate network delay, loss, and packet
251 re-ordering. This is often useful to simulate networks when 248 re-ordering. This is often useful to simulate networks when
252 testing applications or protocols. 249 testing applications or protocols.
@@ -259,58 +256,23 @@ config NET_SCH_NETEM
259config NET_SCH_INGRESS 256config NET_SCH_INGRESS
260 tristate "Ingress Qdisc" 257 tristate "Ingress Qdisc"
261 depends on NET_SCHED 258 depends on NET_SCHED
262 help 259 ---help---
263 If you say Y here, you will be able to police incoming bandwidth 260 Say Y here if you want to use classifiers for incoming packets.
264 and drop packets when this bandwidth exceeds your desired rate.
265 If unsure, say Y. 261 If unsure, say Y.
266 262
267 To compile this code as a module, choose M here: the 263 To compile this code as a module, choose M here: the
268 module will be called sch_ingress. 264 module will be called sch_ingress.
269 265
270config NET_QOS 266comment "Classification"
271 bool "QoS support"
272 depends on NET_SCHED 267 depends on NET_SCHED
273 ---help---
274 Say Y here if you want to include Quality Of Service scheduling
275 features, which means that you will be able to request certain
276 rate-of-flow limits for your network devices.
277
278 This Quality of Service (QoS) support will enable you to use
279 Differentiated Services (diffserv) and Resource Reservation Protocol
280 (RSVP) on your Linux router if you also say Y to "Packet classifier
281 API" and to some classifiers below. Documentation and software is at
282 <http://diffserv.sourceforge.net/>.
283
284 Note that the answer to this question won't directly affect the
285 kernel: saying N will just cause the configurator to skip all
286 the questions about QoS support.
287
288config NET_ESTIMATOR
289 bool "Rate estimator"
290 depends on NET_QOS
291 help
292 In order for Quality of Service scheduling to work, the current
293 rate-of-flow for a network device has to be estimated; if you say Y
294 here, the kernel will do just that.
295 268
296config NET_CLS 269config NET_CLS
297 bool "Packet classifier API" 270 boolean
298 depends on NET_SCHED
299 ---help---
300 The CBQ scheduling algorithm requires that network packets which are
301 scheduled to be sent out over a network device be classified
302 according to some criterion. If you say Y here, you will get a
303 choice of several different packet classifiers with the following
304 questions.
305
306 This will enable you to use Differentiated Services (diffserv) and
307 Resource Reservation Protocol (RSVP) on your Linux router.
308 Documentation and software is at
309 <http://diffserv.sourceforge.net/>.
310 271
311config NET_CLS_BASIC 272config NET_CLS_BASIC
312 tristate "Basic classifier" 273 tristate "Elementary classification (BASIC)"
313 depends on NET_CLS 274 depends NET_SCHED
275 select NET_CLS
314 ---help--- 276 ---help---
315 Say Y here if you want to be able to classify packets using 277 Say Y here if you want to be able to classify packets using
316 only extended matches and actions. 278 only extended matches and actions.
@@ -319,24 +281,25 @@ config NET_CLS_BASIC
319 module will be called cls_basic. 281 module will be called cls_basic.
320 282
321config NET_CLS_TCINDEX 283config NET_CLS_TCINDEX
322 tristate "TC index classifier" 284 tristate "Traffic-Control Index (TCINDEX)"
323 depends on NET_CLS 285 depends NET_SCHED
324 help 286 select NET_CLS
325 If you say Y here, you will be able to classify outgoing packets 287 ---help---
326 according to the tc_index field of the skb. You will want this 288 Say Y here if you want to be able to classify packets based on
327 feature if you want to implement Differentiated Services using 289 traffic control indices. You will want this feature if you want
328 sch_dsmark. If unsure, say Y. 290 to implement Differentiated Services together with DSMARK.
329 291
330 To compile this code as a module, choose M here: the 292 To compile this code as a module, choose M here: the
331 module will be called cls_tcindex. 293 module will be called cls_tcindex.
332 294
333config NET_CLS_ROUTE4 295config NET_CLS_ROUTE4
334 tristate "Routing table based classifier" 296 tristate "Routing decision (ROUTE)"
335 depends on NET_CLS 297 depends NET_SCHED
336 select NET_CLS_ROUTE 298 select NET_CLS_ROUTE
337 help 299 select NET_CLS
338 If you say Y here, you will be able to classify outgoing packets 300 ---help---
339 according to the route table entry they matched. If unsure, say Y. 301 If you say Y here, you will be able to classify packets
302 according to the route table entry they matched.
340 303
341 To compile this code as a module, choose M here: the 304 To compile this code as a module, choose M here: the
342 module will be called cls_route. 305 module will be called cls_route.
@@ -346,58 +309,45 @@ config NET_CLS_ROUTE
346 default n 309 default n
347 310
348config NET_CLS_FW 311config NET_CLS_FW
349 tristate "Firewall based classifier" 312 tristate "Netfilter mark (FW)"
350 depends on NET_CLS 313 depends NET_SCHED
351 help 314 select NET_CLS
352 If you say Y here, you will be able to classify outgoing packets 315 ---help---
353 according to firewall criteria you specified. 316 If you say Y here, you will be able to classify packets
317 according to netfilter/firewall marks.
354 318
355 To compile this code as a module, choose M here: the 319 To compile this code as a module, choose M here: the
356 module will be called cls_fw. 320 module will be called cls_fw.
357 321
358config NET_CLS_U32 322config NET_CLS_U32
359 tristate "U32 classifier" 323 tristate "Universal 32bit comparisons w/ hashing (U32)"
360 depends on NET_CLS 324 depends NET_SCHED
361 help 325 select NET_CLS
362 If you say Y here, you will be able to classify outgoing packets 326 ---help---
363 according to their destination address. If unsure, say Y. 327 Say Y here to be able to classify packetes using a universal
328 32bit pieces based comparison scheme.
364 329
365 To compile this code as a module, choose M here: the 330 To compile this code as a module, choose M here: the
366 module will be called cls_u32. 331 module will be called cls_u32.
367 332
368config CLS_U32_PERF 333config CLS_U32_PERF
369 bool "U32 classifier performance counters" 334 bool "Performance counters support"
370 depends on NET_CLS_U32 335 depends on NET_CLS_U32
371 help 336 ---help---
372 gathers stats that could be used to tune u32 classifier performance. 337 Say Y here to make u32 gather additional statistics useful for
373 Requires a new iproute2 338 fine tuning u32 classifiers.
374 You MUST NOT turn this on if you dont have an update iproute2.
375
376config NET_CLS_IND
377 bool "classify input device (slows things u32/fw) "
378 depends on NET_CLS_U32 || NET_CLS_FW
379 help
380 This option will be killed eventually when a
381 metadata action appears because it slows things a little
382 Available only for u32 and fw classifiers.
383 Requires a new iproute2
384 You MUST NOT turn this on if you dont have an update iproute2.
385 339
386config CLS_U32_MARK 340config CLS_U32_MARK
387 bool "Use nfmark as a key in U32 classifier" 341 bool "Netfilter marks support"
388 depends on NET_CLS_U32 && NETFILTER 342 depends on NET_CLS_U32 && NETFILTER
389 help 343 ---help---
390 This allows you to match mark in a u32 filter. 344 Say Y here to be able to use netfilter marks as u32 key.
391 Example:
392 tc filter add dev eth0 protocol ip parent 1:0 prio 5 u32 \
393 match mark 0x0090 0xffff \
394 match ip dst 4.4.4.4 \
395 flowid 1:90
396 You must use a new iproute2 to use this feature.
397 345
398config NET_CLS_RSVP 346config NET_CLS_RSVP
399 tristate "Special RSVP classifier" 347 tristate "IPv4 Resource Reservation Protocol (RSVP)"
400 depends on NET_CLS && NET_QOS 348 depends on NET_SCHED
349 select NET_CLS
350 select NET_ESTIMATOR
401 ---help--- 351 ---help---
402 The Resource Reservation Protocol (RSVP) permits end systems to 352 The Resource Reservation Protocol (RSVP) permits end systems to
403 request a minimum and maximum data flow rate for a connection; this 353 request a minimum and maximum data flow rate for a connection; this
@@ -410,31 +360,33 @@ config NET_CLS_RSVP
410 module will be called cls_rsvp. 360 module will be called cls_rsvp.
411 361
412config NET_CLS_RSVP6 362config NET_CLS_RSVP6
413 tristate "Special RSVP classifier for IPv6" 363 tristate "IPv6 Resource Reservation Protocol (RSVP6)"
414 depends on NET_CLS && NET_QOS 364 depends on NET_SCHED
365 select NET_CLS
366 select NET_ESTIMATOR
415 ---help--- 367 ---help---
416 The Resource Reservation Protocol (RSVP) permits end systems to 368 The Resource Reservation Protocol (RSVP) permits end systems to
417 request a minimum and maximum data flow rate for a connection; this 369 request a minimum and maximum data flow rate for a connection; this
418 is important for real time data such as streaming sound or video. 370 is important for real time data such as streaming sound or video.
419 371
420 Say Y here if you want to be able to classify outgoing packets based 372 Say Y here if you want to be able to classify outgoing packets based
421 on their RSVP requests and you are using the new Internet Protocol 373 on their RSVP requests and you are using the IPv6.
422 IPv6 as opposed to the older and more common IPv4.
423 374
424 To compile this code as a module, choose M here: the 375 To compile this code as a module, choose M here: the
425 module will be called cls_rsvp6. 376 module will be called cls_rsvp6.
426 377
427config NET_EMATCH 378config NET_EMATCH
428 bool "Extended Matches" 379 bool "Extended Matches"
429 depends on NET_CLS 380 depends NET_SCHED
381 select NET_CLS
430 ---help--- 382 ---help---
431 Say Y here if you want to use extended matches on top of classifiers 383 Say Y here if you want to use extended matches on top of classifiers
432 and select the extended matches below. 384 and select the extended matches below.
433 385
434 Extended matches are small classification helpers not worth writing 386 Extended matches are small classification helpers not worth writing
435 a separate classifier. 387 a separate classifier for.
436 388
437 You must have a recent version of the iproute2 tools in order to use 389 A recent version of the iproute2 package is required to use
438 extended matches. 390 extended matches.
439 391
440config NET_EMATCH_STACK 392config NET_EMATCH_STACK
@@ -468,7 +420,7 @@ config NET_EMATCH_NBYTE
468 module will be called em_nbyte. 420 module will be called em_nbyte.
469 421
470config NET_EMATCH_U32 422config NET_EMATCH_U32
471 tristate "U32 hashing key" 423 tristate "U32 key"
472 depends on NET_EMATCH 424 depends on NET_EMATCH
473 ---help--- 425 ---help---
474 Say Y here if you want to be able to classify packets using 426 Say Y here if you want to be able to classify packets using
@@ -496,76 +448,120 @@ config NET_EMATCH_TEXT
496 select TEXTSEARCH_BM 448 select TEXTSEARCH_BM
497 select TEXTSEARCH_FSM 449 select TEXTSEARCH_FSM
498 ---help--- 450 ---help---
499 Say Y here if you want to be ablt to classify packets based on 451 Say Y here if you want to be able to classify packets based on
500 textsearch comparisons. 452 textsearch comparisons.
501 453
502 To compile this code as a module, choose M here: the 454 To compile this code as a module, choose M here: the
503 module will be called em_text. 455 module will be called em_text.
504 456
505config NET_CLS_ACT 457config NET_CLS_ACT
506 bool "Packet ACTION" 458 bool "Actions"
507 depends on EXPERIMENTAL && NET_CLS && NET_QOS 459 depends on EXPERIMENTAL && NET_SCHED
460 select NET_ESTIMATOR
508 ---help--- 461 ---help---
509 This option requires you have a new iproute2. It enables 462 Say Y here if you want to use traffic control actions. Actions
510 tc extensions which can be used with tc classifiers. 463 get attached to classifiers and are invoked after a successful
511 You MUST NOT turn this on if you dont have an update iproute2. 464 classification. They are used to overwrite the classification
465 result, instantly drop or redirect packets, etc.
466
467 A recent version of the iproute2 package is required to use
468 extended matches.
512 469
513config NET_ACT_POLICE 470config NET_ACT_POLICE
514 tristate "Policing Actions" 471 tristate "Traffic Policing"
515 depends on NET_CLS_ACT 472 depends on NET_CLS_ACT
516 ---help--- 473 ---help---
517 If you are using a newer iproute2 select this one, otherwise use one 474 Say Y here if you want to do traffic policing, i.e. strict
518 below to select a policer. 475 bandwidth limiting. This action replaces the existing policing
519 You MUST NOT turn this on if you dont have an update iproute2. 476 module.
477
478 To compile this code as a module, choose M here: the
479 module will be called police.
520 480
521config NET_ACT_GACT 481config NET_ACT_GACT
522 tristate "generic Actions" 482 tristate "Generic actions"
523 depends on NET_CLS_ACT 483 depends on NET_CLS_ACT
524 ---help--- 484 ---help---
525 You must have new iproute2 to use this feature. 485 Say Y here to take generic actions such as dropping and
526 This adds simple filtering actions like drop, accept etc. 486 accepting packets.
487
488 To compile this code as a module, choose M here: the
489 module will be called gact.
527 490
528config GACT_PROB 491config GACT_PROB
529 bool "generic Actions probability" 492 bool "Probability support"
530 depends on NET_ACT_GACT 493 depends on NET_ACT_GACT
531 ---help--- 494 ---help---
532 Allows generic actions to be randomly or deterministically used. 495 Say Y here to use the generic action randomly or deterministically.
533 496
534config NET_ACT_MIRRED 497config NET_ACT_MIRRED
535 tristate "Packet In/Egress redirecton/mirror Actions" 498 tristate "Redirecting and Mirroring"
536 depends on NET_CLS_ACT 499 depends on NET_CLS_ACT
537 ---help--- 500 ---help---
538 requires new iproute2 501 Say Y here to allow packets to be mirrored or redirected to
539 This allows packets to be mirrored or redirected to netdevices 502 other devices.
503
504 To compile this code as a module, choose M here: the
505 module will be called mirred.
540 506
541config NET_ACT_IPT 507config NET_ACT_IPT
542 tristate "iptables Actions" 508 tristate "IPtables targets"
543 depends on NET_CLS_ACT && NETFILTER && IP_NF_IPTABLES 509 depends on NET_CLS_ACT && NETFILTER && IP_NF_IPTABLES
544 ---help--- 510 ---help---
545 requires new iproute2 511 Say Y here to be able to invoke iptables targets after succesful
546 This allows iptables targets to be used by tc filters 512 classification.
513
514 To compile this code as a module, choose M here: the
515 module will be called ipt.
547 516
548config NET_ACT_PEDIT 517config NET_ACT_PEDIT
549 tristate "Generic Packet Editor Actions" 518 tristate "Packet Editing"
550 depends on NET_CLS_ACT 519 depends on NET_CLS_ACT
551 ---help--- 520 ---help---
552 requires new iproute2 521 Say Y here if you want to mangle the content of packets.
553 This allows for packets to be generically edited
554 522
555config NET_CLS_POLICE 523 To compile this code as a module, choose M here: the
556 bool "Traffic policing (needed for in/egress)" 524 module will be called pedit.
557 depends on NET_CLS && NET_QOS && NET_CLS_ACT!=y
558 help
559 Say Y to support traffic policing (bandwidth limits). Needed for
560 ingress and egress rate limiting.
561 525
562config NET_ACT_SIMP 526config NET_ACT_SIMP
563 tristate "Simple action" 527 tristate "Simple Example (Debug)"
564 depends on NET_CLS_ACT 528 depends on NET_CLS_ACT
565 ---help--- 529 ---help---
566 You must have new iproute2 to use this feature. 530 Say Y here to add a simple action for demonstration purposes.
567 This adds a very simple action for demonstration purposes 531 It is meant as an example and for debugging purposes. It will
568 The idea is to give action authors a basic example to look at. 532 print a configured policy string followed by the packet count
569 All this action will do is print on the console the configured 533 to the console for every packet that passes by.
570 policy string followed by _ then packet count. 534
535 If unsure, say N.
536
537 To compile this code as a module, choose M here: the
538 module will be called simple.
539
540config NET_CLS_POLICE
541 bool "Traffic Policing (obsolete)"
542 depends on NET_SCHED && NET_CLS_ACT!=y
543 select NET_ESTIMATOR
544 ---help---
545 Say Y here if you want to do traffic policing, i.e. strict
546 bandwidth limiting. This option is obsoleted by the traffic
547 policer implemented as action, it stays here for compatibility
548 reasons.
549
550config NET_CLS_IND
551 bool "Incoming device classification"
552 depends on NET_SCHED && (NET_CLS_U32 || NET_CLS_FW)
553 ---help---
554 Say Y here to extend the u32 and fw classifier to support
555 classification based on the incoming device. This option is
556 likely to disappear in favour of the metadata ematch.
557
558config NET_ESTIMATOR
559 bool "Rate estimator"
560 depends on NET_SCHED
561 ---help---
562 Say Y here to allow using rate estimators to estimate the current
563 rate-of-flow for network devices, queues, etc. This module is
564 automaticaly selected if needed but can be selected manually for
565 statstical purposes.
571 566
567endmenu
diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
index 10e82ec2ebd3..660c61bdf164 100644
--- a/net/sctp/sm_make_chunk.c
+++ b/net/sctp/sm_make_chunk.c
@@ -554,7 +554,7 @@ struct sctp_chunk *sctp_make_datafrag_empty(struct sctp_association *asoc,
554 dp.ppid = sinfo->sinfo_ppid; 554 dp.ppid = sinfo->sinfo_ppid;
555 555
556 /* Set the flags for an unordered send. */ 556 /* Set the flags for an unordered send. */
557 if (sinfo->sinfo_flags & MSG_UNORDERED) { 557 if (sinfo->sinfo_flags & SCTP_UNORDERED) {
558 flags |= SCTP_DATA_UNORDERED; 558 flags |= SCTP_DATA_UNORDERED;
559 dp.ssn = 0; 559 dp.ssn = 0;
560 } else 560 } else
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index 02e068d3450d..b529af5e6f2a 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -1010,6 +1010,19 @@ static int __sctp_connect(struct sock* sk,
1010 err = -EAGAIN; 1010 err = -EAGAIN;
1011 goto out_free; 1011 goto out_free;
1012 } 1012 }
1013 } else {
1014 /*
1015 * If an unprivileged user inherits a 1-many
1016 * style socket with open associations on a
1017 * privileged port, it MAY be permitted to
1018 * accept new associations, but it SHOULD NOT
1019 * be permitted to open new associations.
1020 */
1021 if (ep->base.bind_addr.port < PROT_SOCK &&
1022 !capable(CAP_NET_BIND_SERVICE)) {
1023 err = -EACCES;
1024 goto out_free;
1025 }
1013 } 1026 }
1014 1027
1015 scope = sctp_scope(&to); 1028 scope = sctp_scope(&to);
@@ -1389,27 +1402,27 @@ SCTP_STATIC int sctp_sendmsg(struct kiocb *iocb, struct sock *sk,
1389 SCTP_DEBUG_PRINTK("msg_len: %zu, sinfo_flags: 0x%x\n", 1402 SCTP_DEBUG_PRINTK("msg_len: %zu, sinfo_flags: 0x%x\n",
1390 msg_len, sinfo_flags); 1403 msg_len, sinfo_flags);
1391 1404
1392 /* MSG_EOF or MSG_ABORT cannot be set on a TCP-style socket. */ 1405 /* SCTP_EOF or SCTP_ABORT cannot be set on a TCP-style socket. */
1393 if (sctp_style(sk, TCP) && (sinfo_flags & (MSG_EOF | MSG_ABORT))) { 1406 if (sctp_style(sk, TCP) && (sinfo_flags & (SCTP_EOF | SCTP_ABORT))) {
1394 err = -EINVAL; 1407 err = -EINVAL;
1395 goto out_nounlock; 1408 goto out_nounlock;
1396 } 1409 }
1397 1410
1398 /* If MSG_EOF is set, no data can be sent. Disallow sending zero 1411 /* If SCTP_EOF is set, no data can be sent. Disallow sending zero
1399 * length messages when MSG_EOF|MSG_ABORT is not set. 1412 * length messages when SCTP_EOF|SCTP_ABORT is not set.
1400 * If MSG_ABORT is set, the message length could be non zero with 1413 * If SCTP_ABORT is set, the message length could be non zero with
1401 * the msg_iov set to the user abort reason. 1414 * the msg_iov set to the user abort reason.
1402 */ 1415 */
1403 if (((sinfo_flags & MSG_EOF) && (msg_len > 0)) || 1416 if (((sinfo_flags & SCTP_EOF) && (msg_len > 0)) ||
1404 (!(sinfo_flags & (MSG_EOF|MSG_ABORT)) && (msg_len == 0))) { 1417 (!(sinfo_flags & (SCTP_EOF|SCTP_ABORT)) && (msg_len == 0))) {
1405 err = -EINVAL; 1418 err = -EINVAL;
1406 goto out_nounlock; 1419 goto out_nounlock;
1407 } 1420 }
1408 1421
1409 /* If MSG_ADDR_OVER is set, there must be an address 1422 /* If SCTP_ADDR_OVER is set, there must be an address
1410 * specified in msg_name. 1423 * specified in msg_name.
1411 */ 1424 */
1412 if ((sinfo_flags & MSG_ADDR_OVER) && (!msg->msg_name)) { 1425 if ((sinfo_flags & SCTP_ADDR_OVER) && (!msg->msg_name)) {
1413 err = -EINVAL; 1426 err = -EINVAL;
1414 goto out_nounlock; 1427 goto out_nounlock;
1415 } 1428 }
@@ -1458,14 +1471,14 @@ SCTP_STATIC int sctp_sendmsg(struct kiocb *iocb, struct sock *sk,
1458 goto out_unlock; 1471 goto out_unlock;
1459 } 1472 }
1460 1473
1461 if (sinfo_flags & MSG_EOF) { 1474 if (sinfo_flags & SCTP_EOF) {
1462 SCTP_DEBUG_PRINTK("Shutting down association: %p\n", 1475 SCTP_DEBUG_PRINTK("Shutting down association: %p\n",
1463 asoc); 1476 asoc);
1464 sctp_primitive_SHUTDOWN(asoc, NULL); 1477 sctp_primitive_SHUTDOWN(asoc, NULL);
1465 err = 0; 1478 err = 0;
1466 goto out_unlock; 1479 goto out_unlock;
1467 } 1480 }
1468 if (sinfo_flags & MSG_ABORT) { 1481 if (sinfo_flags & SCTP_ABORT) {
1469 SCTP_DEBUG_PRINTK("Aborting association: %p\n", asoc); 1482 SCTP_DEBUG_PRINTK("Aborting association: %p\n", asoc);
1470 sctp_primitive_ABORT(asoc, msg); 1483 sctp_primitive_ABORT(asoc, msg);
1471 err = 0; 1484 err = 0;
@@ -1477,7 +1490,7 @@ SCTP_STATIC int sctp_sendmsg(struct kiocb *iocb, struct sock *sk,
1477 if (!asoc) { 1490 if (!asoc) {
1478 SCTP_DEBUG_PRINTK("There is no association yet.\n"); 1491 SCTP_DEBUG_PRINTK("There is no association yet.\n");
1479 1492
1480 if (sinfo_flags & (MSG_EOF | MSG_ABORT)) { 1493 if (sinfo_flags & (SCTP_EOF | SCTP_ABORT)) {
1481 err = -EINVAL; 1494 err = -EINVAL;
1482 goto out_unlock; 1495 goto out_unlock;
1483 } 1496 }
@@ -1515,6 +1528,19 @@ SCTP_STATIC int sctp_sendmsg(struct kiocb *iocb, struct sock *sk,
1515 err = -EAGAIN; 1528 err = -EAGAIN;
1516 goto out_unlock; 1529 goto out_unlock;
1517 } 1530 }
1531 } else {
1532 /*
1533 * If an unprivileged user inherits a one-to-many
1534 * style socket with open associations on a privileged
1535 * port, it MAY be permitted to accept new associations,
1536 * but it SHOULD NOT be permitted to open new
1537 * associations.
1538 */
1539 if (ep->base.bind_addr.port < PROT_SOCK &&
1540 !capable(CAP_NET_BIND_SERVICE)) {
1541 err = -EACCES;
1542 goto out_unlock;
1543 }
1518 } 1544 }
1519 1545
1520 scope = sctp_scope(&to); 1546 scope = sctp_scope(&to);
@@ -1611,10 +1637,10 @@ SCTP_STATIC int sctp_sendmsg(struct kiocb *iocb, struct sock *sk,
1611 1637
1612 /* If an address is passed with the sendto/sendmsg call, it is used 1638 /* If an address is passed with the sendto/sendmsg call, it is used
1613 * to override the primary destination address in the TCP model, or 1639 * to override the primary destination address in the TCP model, or
1614 * when MSG_ADDR_OVER flag is set in the UDP model. 1640 * when SCTP_ADDR_OVER flag is set in the UDP model.
1615 */ 1641 */
1616 if ((sctp_style(sk, TCP) && msg_name) || 1642 if ((sctp_style(sk, TCP) && msg_name) ||
1617 (sinfo_flags & MSG_ADDR_OVER)) { 1643 (sinfo_flags & SCTP_ADDR_OVER)) {
1618 chunk_tp = sctp_assoc_lookup_paddr(asoc, &to); 1644 chunk_tp = sctp_assoc_lookup_paddr(asoc, &to);
1619 if (!chunk_tp) { 1645 if (!chunk_tp) {
1620 err = -EINVAL; 1646 err = -EINVAL;
@@ -2306,16 +2332,14 @@ static int sctp_setsockopt_maxseg(struct sock *sk, char __user *optval, int optl
2306 return -EINVAL; 2332 return -EINVAL;
2307 if (get_user(val, (int __user *)optval)) 2333 if (get_user(val, (int __user *)optval))
2308 return -EFAULT; 2334 return -EFAULT;
2309 if ((val < 8) || (val > SCTP_MAX_CHUNK_LEN)) 2335 if ((val != 0) && ((val < 8) || (val > SCTP_MAX_CHUNK_LEN)))
2310 return -EINVAL; 2336 return -EINVAL;
2311 sp->user_frag = val; 2337 sp->user_frag = val;
2312 2338
2313 if (val) { 2339 /* Update the frag_point of the existing associations. */
2314 /* Update the frag_point of the existing associations. */ 2340 list_for_each(pos, &(sp->ep->asocs)) {
2315 list_for_each(pos, &(sp->ep->asocs)) { 2341 asoc = list_entry(pos, struct sctp_association, asocs);
2316 asoc = list_entry(pos, struct sctp_association, asocs); 2342 asoc->frag_point = sctp_frag_point(sp, asoc->pmtu);
2317 asoc->frag_point = sctp_frag_point(sp, asoc->pmtu);
2318 }
2319 } 2343 }
2320 2344
2321 return 0; 2345 return 0;
@@ -2384,14 +2408,14 @@ static int sctp_setsockopt_peer_primary_addr(struct sock *sk, char __user *optva
2384static int sctp_setsockopt_adaption_layer(struct sock *sk, char __user *optval, 2408static int sctp_setsockopt_adaption_layer(struct sock *sk, char __user *optval,
2385 int optlen) 2409 int optlen)
2386{ 2410{
2387 __u32 val; 2411 struct sctp_setadaption adaption;
2388 2412
2389 if (optlen < sizeof(__u32)) 2413 if (optlen != sizeof(struct sctp_setadaption))
2390 return -EINVAL; 2414 return -EINVAL;
2391 if (copy_from_user(&val, optval, sizeof(__u32))) 2415 if (copy_from_user(&adaption, optval, optlen))
2392 return -EFAULT; 2416 return -EFAULT;
2393 2417
2394 sctp_sk(sk)->adaption_ind = val; 2418 sctp_sk(sk)->adaption_ind = adaption.ssb_adaption_ind;
2395 2419
2396 return 0; 2420 return 0;
2397} 2421}
@@ -3672,17 +3696,15 @@ static int sctp_getsockopt_primary_addr(struct sock *sk, int len,
3672static int sctp_getsockopt_adaption_layer(struct sock *sk, int len, 3696static int sctp_getsockopt_adaption_layer(struct sock *sk, int len,
3673 char __user *optval, int __user *optlen) 3697 char __user *optval, int __user *optlen)
3674{ 3698{
3675 __u32 val; 3699 struct sctp_setadaption adaption;
3676 3700
3677 if (len < sizeof(__u32)) 3701 if (len != sizeof(struct sctp_setadaption))
3678 return -EINVAL; 3702 return -EINVAL;
3679 3703
3680 len = sizeof(__u32); 3704 adaption.ssb_adaption_ind = sctp_sk(sk)->adaption_ind;
3681 val = sctp_sk(sk)->adaption_ind; 3705 if (copy_to_user(optval, &adaption, len))
3682 if (put_user(len, optlen))
3683 return -EFAULT;
3684 if (copy_to_user(optval, &val, len))
3685 return -EFAULT; 3706 return -EFAULT;
3707
3686 return 0; 3708 return 0;
3687} 3709}
3688 3710
@@ -4640,8 +4662,8 @@ SCTP_STATIC int sctp_msghdr_parse(const struct msghdr *msg,
4640 4662
4641 /* Minimally, validate the sinfo_flags. */ 4663 /* Minimally, validate the sinfo_flags. */
4642 if (cmsgs->info->sinfo_flags & 4664 if (cmsgs->info->sinfo_flags &
4643 ~(MSG_UNORDERED | MSG_ADDR_OVER | 4665 ~(SCTP_UNORDERED | SCTP_ADDR_OVER |
4644 MSG_ABORT | MSG_EOF)) 4666 SCTP_ABORT | SCTP_EOF))
4645 return -EINVAL; 4667 return -EINVAL;
4646 break; 4668 break;
4647 4669
diff --git a/net/sctp/ulpevent.c b/net/sctp/ulpevent.c
index 057e7fac3af0..e049f41faa47 100644
--- a/net/sctp/ulpevent.c
+++ b/net/sctp/ulpevent.c
@@ -698,7 +698,7 @@ struct sctp_ulpevent *sctp_ulpevent_make_rcvmsg(struct sctp_association *asoc,
698 event->ssn = ntohs(chunk->subh.data_hdr->ssn); 698 event->ssn = ntohs(chunk->subh.data_hdr->ssn);
699 event->ppid = chunk->subh.data_hdr->ppid; 699 event->ppid = chunk->subh.data_hdr->ppid;
700 if (chunk->chunk_hdr->flags & SCTP_DATA_UNORDERED) { 700 if (chunk->chunk_hdr->flags & SCTP_DATA_UNORDERED) {
701 event->flags |= MSG_UNORDERED; 701 event->flags |= SCTP_UNORDERED;
702 event->cumtsn = sctp_tsnmap_get_ctsn(&asoc->peer.tsn_map); 702 event->cumtsn = sctp_tsnmap_get_ctsn(&asoc->peer.tsn_map);
703 } 703 }
704 event->tsn = ntohl(chunk->subh.data_hdr->tsn); 704 event->tsn = ntohl(chunk->subh.data_hdr->tsn);
@@ -824,7 +824,7 @@ void sctp_ulpevent_read_sndrcvinfo(const struct sctp_ulpevent *event,
824 * 824 *
825 * recvmsg() flags: 825 * recvmsg() flags:
826 * 826 *
827 * MSG_UNORDERED - This flag is present when the message was sent 827 * SCTP_UNORDERED - This flag is present when the message was sent
828 * non-ordered. 828 * non-ordered.
829 */ 829 */
830 sinfo.sinfo_flags = event->flags; 830 sinfo.sinfo_flags = event->flags;
@@ -839,7 +839,7 @@ void sctp_ulpevent_read_sndrcvinfo(const struct sctp_ulpevent *event,
839 * This field will hold the current cumulative TSN as 839 * This field will hold the current cumulative TSN as
840 * known by the underlying SCTP layer. Note this field is 840 * known by the underlying SCTP layer. Note this field is
841 * ignored when sending and only valid for a receive 841 * ignored when sending and only valid for a receive
842 * operation when sinfo_flags are set to MSG_UNORDERED. 842 * operation when sinfo_flags are set to SCTP_UNORDERED.
843 */ 843 */
844 sinfo.sinfo_cumtsn = event->cumtsn; 844 sinfo.sinfo_cumtsn = event->cumtsn;
845 /* sinfo_assoc_id: sizeof (sctp_assoc_t) 845 /* sinfo_assoc_id: sizeof (sctp_assoc_t)
diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c
index a415d99c394d..8c7756036e95 100644
--- a/net/sunrpc/auth.c
+++ b/net/sunrpc/auth.c
@@ -299,11 +299,10 @@ put_rpccred(struct rpc_cred *cred)
299void 299void
300rpcauth_unbindcred(struct rpc_task *task) 300rpcauth_unbindcred(struct rpc_task *task)
301{ 301{
302 struct rpc_auth *auth = task->tk_auth;
303 struct rpc_cred *cred = task->tk_msg.rpc_cred; 302 struct rpc_cred *cred = task->tk_msg.rpc_cred;
304 303
305 dprintk("RPC: %4d releasing %s cred %p\n", 304 dprintk("RPC: %4d releasing %s cred %p\n",
306 task->tk_pid, auth->au_ops->au_name, cred); 305 task->tk_pid, task->tk_auth->au_ops->au_name, cred);
307 306
308 put_rpccred(cred); 307 put_rpccred(cred);
309 task->tk_msg.rpc_cred = NULL; 308 task->tk_msg.rpc_cred = NULL;
@@ -312,22 +311,22 @@ rpcauth_unbindcred(struct rpc_task *task)
312u32 * 311u32 *
313rpcauth_marshcred(struct rpc_task *task, u32 *p) 312rpcauth_marshcred(struct rpc_task *task, u32 *p)
314{ 313{
315 struct rpc_auth *auth = task->tk_auth;
316 struct rpc_cred *cred = task->tk_msg.rpc_cred; 314 struct rpc_cred *cred = task->tk_msg.rpc_cred;
317 315
318 dprintk("RPC: %4d marshaling %s cred %p\n", 316 dprintk("RPC: %4d marshaling %s cred %p\n",
319 task->tk_pid, auth->au_ops->au_name, cred); 317 task->tk_pid, task->tk_auth->au_ops->au_name, cred);
318
320 return cred->cr_ops->crmarshal(task, p); 319 return cred->cr_ops->crmarshal(task, p);
321} 320}
322 321
323u32 * 322u32 *
324rpcauth_checkverf(struct rpc_task *task, u32 *p) 323rpcauth_checkverf(struct rpc_task *task, u32 *p)
325{ 324{
326 struct rpc_auth *auth = task->tk_auth;
327 struct rpc_cred *cred = task->tk_msg.rpc_cred; 325 struct rpc_cred *cred = task->tk_msg.rpc_cred;
328 326
329 dprintk("RPC: %4d validating %s cred %p\n", 327 dprintk("RPC: %4d validating %s cred %p\n",
330 task->tk_pid, auth->au_ops->au_name, cred); 328 task->tk_pid, task->tk_auth->au_ops->au_name, cred);
329
331 return cred->cr_ops->crvalidate(task, p); 330 return cred->cr_ops->crvalidate(task, p);
332} 331}
333 332
@@ -363,12 +362,12 @@ rpcauth_unwrap_resp(struct rpc_task *task, kxdrproc_t decode, void *rqstp,
363int 362int
364rpcauth_refreshcred(struct rpc_task *task) 363rpcauth_refreshcred(struct rpc_task *task)
365{ 364{
366 struct rpc_auth *auth = task->tk_auth;
367 struct rpc_cred *cred = task->tk_msg.rpc_cred; 365 struct rpc_cred *cred = task->tk_msg.rpc_cred;
368 int err; 366 int err;
369 367
370 dprintk("RPC: %4d refreshing %s cred %p\n", 368 dprintk("RPC: %4d refreshing %s cred %p\n",
371 task->tk_pid, auth->au_ops->au_name, cred); 369 task->tk_pid, task->tk_auth->au_ops->au_name, cred);
370
372 err = cred->cr_ops->crrefresh(task); 371 err = cred->cr_ops->crrefresh(task);
373 if (err < 0) 372 if (err < 0)
374 task->tk_status = err; 373 task->tk_status = err;
diff --git a/net/sunrpc/auth_gss/gss_krb5_crypto.c b/net/sunrpc/auth_gss/gss_krb5_crypto.c
index 3f3d5437f02d..97c981fa6b8e 100644
--- a/net/sunrpc/auth_gss/gss_krb5_crypto.c
+++ b/net/sunrpc/auth_gss/gss_krb5_crypto.c
@@ -37,7 +37,7 @@
37#include <linux/types.h> 37#include <linux/types.h>
38#include <linux/mm.h> 38#include <linux/mm.h>
39#include <linux/slab.h> 39#include <linux/slab.h>
40#include <asm/scatterlist.h> 40#include <linux/scatterlist.h>
41#include <linux/crypto.h> 41#include <linux/crypto.h>
42#include <linux/highmem.h> 42#include <linux/highmem.h>
43#include <linux/pagemap.h> 43#include <linux/pagemap.h>
@@ -75,9 +75,7 @@ krb5_encrypt(
75 memcpy(local_iv, iv, crypto_tfm_alg_ivsize(tfm)); 75 memcpy(local_iv, iv, crypto_tfm_alg_ivsize(tfm));
76 76
77 memcpy(out, in, length); 77 memcpy(out, in, length);
78 sg[0].page = virt_to_page(out); 78 sg_set_buf(sg, out, length);
79 sg[0].offset = offset_in_page(out);
80 sg[0].length = length;
81 79
82 ret = crypto_cipher_encrypt_iv(tfm, sg, sg, length, local_iv); 80 ret = crypto_cipher_encrypt_iv(tfm, sg, sg, length, local_iv);
83 81
@@ -117,9 +115,7 @@ krb5_decrypt(
117 memcpy(local_iv,iv, crypto_tfm_alg_ivsize(tfm)); 115 memcpy(local_iv,iv, crypto_tfm_alg_ivsize(tfm));
118 116
119 memcpy(out, in, length); 117 memcpy(out, in, length);
120 sg[0].page = virt_to_page(out); 118 sg_set_buf(sg, out, length);
121 sg[0].offset = offset_in_page(out);
122 sg[0].length = length;
123 119
124 ret = crypto_cipher_decrypt_iv(tfm, sg, sg, length, local_iv); 120 ret = crypto_cipher_decrypt_iv(tfm, sg, sg, length, local_iv);
125 121
@@ -132,13 +128,6 @@ out:
132 128
133EXPORT_SYMBOL(krb5_decrypt); 129EXPORT_SYMBOL(krb5_decrypt);
134 130
135static void
136buf_to_sg(struct scatterlist *sg, char *ptr, int len) {
137 sg->page = virt_to_page(ptr);
138 sg->offset = offset_in_page(ptr);
139 sg->length = len;
140}
141
142static int 131static int
143process_xdr_buf(struct xdr_buf *buf, int offset, int len, 132process_xdr_buf(struct xdr_buf *buf, int offset, int len,
144 int (*actor)(struct scatterlist *, void *), void *data) 133 int (*actor)(struct scatterlist *, void *), void *data)
@@ -152,7 +141,7 @@ process_xdr_buf(struct xdr_buf *buf, int offset, int len,
152 thislen = buf->head[0].iov_len - offset; 141 thislen = buf->head[0].iov_len - offset;
153 if (thislen > len) 142 if (thislen > len)
154 thislen = len; 143 thislen = len;
155 buf_to_sg(sg, buf->head[0].iov_base + offset, thislen); 144 sg_set_buf(sg, buf->head[0].iov_base + offset, thislen);
156 ret = actor(sg, data); 145 ret = actor(sg, data);
157 if (ret) 146 if (ret)
158 goto out; 147 goto out;
@@ -195,7 +184,7 @@ process_xdr_buf(struct xdr_buf *buf, int offset, int len,
195 thislen = buf->tail[0].iov_len - offset; 184 thislen = buf->tail[0].iov_len - offset;
196 if (thislen > len) 185 if (thislen > len)
197 thislen = len; 186 thislen = len;
198 buf_to_sg(sg, buf->tail[0].iov_base + offset, thislen); 187 sg_set_buf(sg, buf->tail[0].iov_base + offset, thislen);
199 ret = actor(sg, data); 188 ret = actor(sg, data);
200 len -= thislen; 189 len -= thislen;
201 } 190 }
@@ -241,7 +230,7 @@ make_checksum(s32 cksumtype, char *header, int hdrlen, struct xdr_buf *body,
241 goto out; 230 goto out;
242 231
243 crypto_digest_init(tfm); 232 crypto_digest_init(tfm);
244 buf_to_sg(sg, header, hdrlen); 233 sg_set_buf(sg, header, hdrlen);
245 crypto_digest_update(tfm, sg, 1); 234 crypto_digest_update(tfm, sg, 1);
246 process_xdr_buf(body, body_offset, body->len - body_offset, 235 process_xdr_buf(body, body_offset, body->len - body_offset,
247 checksummer, tfm); 236 checksummer, tfm);
diff --git a/net/sunrpc/sunrpc_syms.c b/net/sunrpc/sunrpc_syms.c
index 2387e7b823ff..a03d4b600c92 100644
--- a/net/sunrpc/sunrpc_syms.c
+++ b/net/sunrpc/sunrpc_syms.c
@@ -63,8 +63,6 @@ EXPORT_SYMBOL(rpc_mkpipe);
63/* Client transport */ 63/* Client transport */
64EXPORT_SYMBOL(xprt_create_proto); 64EXPORT_SYMBOL(xprt_create_proto);
65EXPORT_SYMBOL(xprt_set_timeout); 65EXPORT_SYMBOL(xprt_set_timeout);
66EXPORT_SYMBOL(xprt_udp_slot_table_entries);
67EXPORT_SYMBOL(xprt_tcp_slot_table_entries);
68 66
69/* Client credential cache */ 67/* Client credential cache */
70EXPORT_SYMBOL(rpcauth_register); 68EXPORT_SYMBOL(rpcauth_register);
diff --git a/net/sunrpc/sysctl.c b/net/sunrpc/sysctl.c
index d0c9f460e411..1065904841fd 100644
--- a/net/sunrpc/sysctl.c
+++ b/net/sunrpc/sysctl.c
@@ -119,13 +119,6 @@ done:
119 return 0; 119 return 0;
120} 120}
121 121
122unsigned int xprt_udp_slot_table_entries = RPC_DEF_SLOT_TABLE;
123unsigned int xprt_tcp_slot_table_entries = RPC_DEF_SLOT_TABLE;
124unsigned int xprt_min_resvport = RPC_DEF_MIN_RESVPORT;
125EXPORT_SYMBOL(xprt_min_resvport);
126unsigned int xprt_max_resvport = RPC_DEF_MAX_RESVPORT;
127EXPORT_SYMBOL(xprt_max_resvport);
128
129 122
130static unsigned int min_slot_table_size = RPC_MIN_SLOT_TABLE; 123static unsigned int min_slot_table_size = RPC_MIN_SLOT_TABLE;
131static unsigned int max_slot_table_size = RPC_MAX_SLOT_TABLE; 124static unsigned int max_slot_table_size = RPC_MAX_SLOT_TABLE;
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index 2e1529217e65..0a51fd46a848 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -36,6 +36,15 @@
36#include <net/tcp.h> 36#include <net/tcp.h>
37 37
38/* 38/*
39 * xprtsock tunables
40 */
41unsigned int xprt_udp_slot_table_entries = RPC_DEF_SLOT_TABLE;
42unsigned int xprt_tcp_slot_table_entries = RPC_DEF_SLOT_TABLE;
43
44unsigned int xprt_min_resvport = RPC_DEF_MIN_RESVPORT;
45unsigned int xprt_max_resvport = RPC_DEF_MAX_RESVPORT;
46
47/*
39 * How many times to try sending a request on a socket before waiting 48 * How many times to try sending a request on a socket before waiting
40 * for the socket buffer to clear. 49 * for the socket buffer to clear.
41 */ 50 */
generated by cgit v1.2.2 (git 2.25.0) at 2024-12-27 10:00:21 -0500