aboutsummaryrefslogtreecommitdiffstats
path: root/include/linux
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2013-05-06 16:11:19 -0400
committerLinus Torvalds <torvalds@linux-foundation.org>2013-05-06 16:11:19 -0400
commit91f8575685e35f3bd021286bc82d26397458f5a9 (patch)
tree09de8d889758a12071adb9427ed741e27c907aa6 /include/linux
parent2e378f3eebd28feefbb1f9953834a5a19482f053 (diff)
parentb5b09be30cf99f9c699e825629f02e3bce555d44 (diff)
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client
Pull Ceph changes from Alex Elder: "This is a big pull. Most of it is culmination of Alex's work to implement RBD image layering, which is now complete (yay!). There is also some work from Yan to fix i_mutex behavior surrounding writes in cephfs, a sync write fix, a fix for RBD images that get resized while they are mapped, and a few patches from me that resolve annoying auth warnings and fix several bugs in the ceph auth code." * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client: (254 commits) rbd: fix image request leak on parent read libceph: use slab cache for osd client requests libceph: allocate ceph message data with a slab allocator libceph: allocate ceph messages with a slab allocator rbd: allocate image object names with a slab allocator rbd: allocate object requests with a slab allocator rbd: allocate name separate from obj_request rbd: allocate image requests with a slab allocator rbd: use binary search for snapshot lookup rbd: clear EXISTS flag if mapped snapshot disappears rbd: kill off the snapshot list rbd: define rbd_snap_size() and rbd_snap_features() rbd: use snap_id not index to look up snap info rbd: look up snapshot name in names buffer rbd: drop obj_request->version rbd: drop rbd_obj_method_sync() version parameter rbd: more version parameter removal rbd: get rid of some version parameters rbd: stop tracking header object version rbd: snap names are pointer to constant data ...
Diffstat (limited to 'include/linux')
-rw-r--r--include/linux/ceph/auth.h18
-rw-r--r--include/linux/ceph/ceph_features.h2
-rw-r--r--include/linux/ceph/decode.h30
-rw-r--r--include/linux/ceph/libceph.h31
-rw-r--r--include/linux/ceph/messenger.h104
-rw-r--r--include/linux/ceph/msgr.h1
-rw-r--r--include/linux/ceph/osd_client.h204
-rw-r--r--include/linux/ceph/osdmap.h30
8 files changed, 299 insertions, 121 deletions
diff --git a/include/linux/ceph/auth.h b/include/linux/ceph/auth.h
index d4080f309b56..5f3386844134 100644
--- a/include/linux/ceph/auth.h
+++ b/include/linux/ceph/auth.h
@@ -52,6 +52,9 @@ struct ceph_auth_client_ops {
52 */ 52 */
53 int (*create_authorizer)(struct ceph_auth_client *ac, int peer_type, 53 int (*create_authorizer)(struct ceph_auth_client *ac, int peer_type,
54 struct ceph_auth_handshake *auth); 54 struct ceph_auth_handshake *auth);
55 /* ensure that an existing authorizer is up to date */
56 int (*update_authorizer)(struct ceph_auth_client *ac, int peer_type,
57 struct ceph_auth_handshake *auth);
55 int (*verify_authorizer_reply)(struct ceph_auth_client *ac, 58 int (*verify_authorizer_reply)(struct ceph_auth_client *ac,
56 struct ceph_authorizer *a, size_t len); 59 struct ceph_authorizer *a, size_t len);
57 void (*destroy_authorizer)(struct ceph_auth_client *ac, 60 void (*destroy_authorizer)(struct ceph_auth_client *ac,
@@ -75,6 +78,8 @@ struct ceph_auth_client {
75 u64 global_id; /* our unique id in system */ 78 u64 global_id; /* our unique id in system */
76 const struct ceph_crypto_key *key; /* our secret key */ 79 const struct ceph_crypto_key *key; /* our secret key */
77 unsigned want_keys; /* which services we want */ 80 unsigned want_keys; /* which services we want */
81
82 struct mutex mutex;
78}; 83};
79 84
80extern struct ceph_auth_client *ceph_auth_init(const char *name, 85extern struct ceph_auth_client *ceph_auth_init(const char *name,
@@ -94,5 +99,18 @@ extern int ceph_build_auth(struct ceph_auth_client *ac,
94 void *msg_buf, size_t msg_len); 99 void *msg_buf, size_t msg_len);
95 100
96extern int ceph_auth_is_authenticated(struct ceph_auth_client *ac); 101extern int ceph_auth_is_authenticated(struct ceph_auth_client *ac);
102extern int ceph_auth_create_authorizer(struct ceph_auth_client *ac,
103 int peer_type,
104 struct ceph_auth_handshake *auth);
105extern void ceph_auth_destroy_authorizer(struct ceph_auth_client *ac,
106 struct ceph_authorizer *a);
107extern int ceph_auth_update_authorizer(struct ceph_auth_client *ac,
108 int peer_type,
109 struct ceph_auth_handshake *a);
110extern int ceph_auth_verify_authorizer_reply(struct ceph_auth_client *ac,
111 struct ceph_authorizer *a,
112 size_t len);
113extern void ceph_auth_invalidate_authorizer(struct ceph_auth_client *ac,
114 int peer_type);
97 115
98#endif 116#endif
diff --git a/include/linux/ceph/ceph_features.h b/include/linux/ceph/ceph_features.h
index 76554cecaab2..4c42080347af 100644
--- a/include/linux/ceph/ceph_features.h
+++ b/include/linux/ceph/ceph_features.h
@@ -41,6 +41,7 @@
41 */ 41 */
42#define CEPH_FEATURES_SUPPORTED_DEFAULT \ 42#define CEPH_FEATURES_SUPPORTED_DEFAULT \
43 (CEPH_FEATURE_NOSRCADDR | \ 43 (CEPH_FEATURE_NOSRCADDR | \
44 CEPH_FEATURE_RECONNECT_SEQ | \
44 CEPH_FEATURE_PGID64 | \ 45 CEPH_FEATURE_PGID64 | \
45 CEPH_FEATURE_PGPOOL3 | \ 46 CEPH_FEATURE_PGPOOL3 | \
46 CEPH_FEATURE_OSDENC | \ 47 CEPH_FEATURE_OSDENC | \
@@ -51,6 +52,7 @@
51 52
52#define CEPH_FEATURES_REQUIRED_DEFAULT \ 53#define CEPH_FEATURES_REQUIRED_DEFAULT \
53 (CEPH_FEATURE_NOSRCADDR | \ 54 (CEPH_FEATURE_NOSRCADDR | \
55 CEPH_FEATURE_RECONNECT_SEQ | \
54 CEPH_FEATURE_PGID64 | \ 56 CEPH_FEATURE_PGID64 | \
55 CEPH_FEATURE_PGPOOL3 | \ 57 CEPH_FEATURE_PGPOOL3 | \
56 CEPH_FEATURE_OSDENC) 58 CEPH_FEATURE_OSDENC)
diff --git a/include/linux/ceph/decode.h b/include/linux/ceph/decode.h
index 360d9d08ca9e..379f71508995 100644
--- a/include/linux/ceph/decode.h
+++ b/include/linux/ceph/decode.h
@@ -8,6 +8,23 @@
8 8
9#include <linux/ceph/types.h> 9#include <linux/ceph/types.h>
10 10
11/* This seemed to be the easiest place to define these */
12
13#define U8_MAX ((u8)(~0U))
14#define U16_MAX ((u16)(~0U))
15#define U32_MAX ((u32)(~0U))
16#define U64_MAX ((u64)(~0ULL))
17
18#define S8_MAX ((s8)(U8_MAX >> 1))
19#define S16_MAX ((s16)(U16_MAX >> 1))
20#define S32_MAX ((s32)(U32_MAX >> 1))
21#define S64_MAX ((s64)(U64_MAX >> 1LL))
22
23#define S8_MIN ((s8)(-S8_MAX - 1))
24#define S16_MIN ((s16)(-S16_MAX - 1))
25#define S32_MIN ((s32)(-S32_MAX - 1))
26#define S64_MIN ((s64)(-S64_MAX - 1LL))
27
11/* 28/*
12 * in all cases, 29 * in all cases,
13 * void **p pointer to position pointer 30 * void **p pointer to position pointer
@@ -137,14 +154,19 @@ bad:
137static inline void ceph_decode_timespec(struct timespec *ts, 154static inline void ceph_decode_timespec(struct timespec *ts,
138 const struct ceph_timespec *tv) 155 const struct ceph_timespec *tv)
139{ 156{
140 ts->tv_sec = le32_to_cpu(tv->tv_sec); 157 ts->tv_sec = (__kernel_time_t)le32_to_cpu(tv->tv_sec);
141 ts->tv_nsec = le32_to_cpu(tv->tv_nsec); 158 ts->tv_nsec = (long)le32_to_cpu(tv->tv_nsec);
142} 159}
143static inline void ceph_encode_timespec(struct ceph_timespec *tv, 160static inline void ceph_encode_timespec(struct ceph_timespec *tv,
144 const struct timespec *ts) 161 const struct timespec *ts)
145{ 162{
146 tv->tv_sec = cpu_to_le32(ts->tv_sec); 163 BUG_ON(ts->tv_sec < 0);
147 tv->tv_nsec = cpu_to_le32(ts->tv_nsec); 164 BUG_ON(ts->tv_sec > (__kernel_time_t)U32_MAX);
165 BUG_ON(ts->tv_nsec < 0);
166 BUG_ON(ts->tv_nsec > (long)U32_MAX);
167
168 tv->tv_sec = cpu_to_le32((u32)ts->tv_sec);
169 tv->tv_nsec = cpu_to_le32((u32)ts->tv_nsec);
148} 170}
149 171
150/* 172/*
diff --git a/include/linux/ceph/libceph.h b/include/linux/ceph/libceph.h
index 29818fc3fa49..2e3024881a5e 100644
--- a/include/linux/ceph/libceph.h
+++ b/include/linux/ceph/libceph.h
@@ -66,6 +66,7 @@ struct ceph_options {
66#define CEPH_OSD_IDLE_TTL_DEFAULT 60 66#define CEPH_OSD_IDLE_TTL_DEFAULT 60
67 67
68#define CEPH_MSG_MAX_FRONT_LEN (16*1024*1024) 68#define CEPH_MSG_MAX_FRONT_LEN (16*1024*1024)
69#define CEPH_MSG_MAX_MIDDLE_LEN (16*1024*1024)
69#define CEPH_MSG_MAX_DATA_LEN (16*1024*1024) 70#define CEPH_MSG_MAX_DATA_LEN (16*1024*1024)
70 71
71#define CEPH_AUTH_NAME_DEFAULT "guest" 72#define CEPH_AUTH_NAME_DEFAULT "guest"
@@ -156,31 +157,11 @@ struct ceph_snap_context {
156 u64 snaps[]; 157 u64 snaps[];
157}; 158};
158 159
159static inline struct ceph_snap_context * 160extern struct ceph_snap_context *ceph_create_snap_context(u32 snap_count,
160ceph_get_snap_context(struct ceph_snap_context *sc) 161 gfp_t gfp_flags);
161{ 162extern struct ceph_snap_context *ceph_get_snap_context(
162 /* 163 struct ceph_snap_context *sc);
163 printk("get_snap_context %p %d -> %d\n", sc, atomic_read(&sc->nref), 164extern void ceph_put_snap_context(struct ceph_snap_context *sc);
164 atomic_read(&sc->nref)+1);
165 */
166 if (sc)
167 atomic_inc(&sc->nref);
168 return sc;
169}
170
171static inline void ceph_put_snap_context(struct ceph_snap_context *sc)
172{
173 if (!sc)
174 return;
175 /*
176 printk("put_snap_context %p %d -> %d\n", sc, atomic_read(&sc->nref),
177 atomic_read(&sc->nref)-1);
178 */
179 if (atomic_dec_and_test(&sc->nref)) {
180 /*printk(" deleting snap_context %p\n", sc);*/
181 kfree(sc);
182 }
183}
184 165
185/* 166/*
186 * calculate the number of pages a given length and offset map onto, 167 * calculate the number of pages a given length and offset map onto,
diff --git a/include/linux/ceph/messenger.h b/include/linux/ceph/messenger.h
index 60903e0f665c..7c1420bb1dce 100644
--- a/include/linux/ceph/messenger.h
+++ b/include/linux/ceph/messenger.h
@@ -64,6 +64,77 @@ struct ceph_messenger {
64 u32 required_features; 64 u32 required_features;
65}; 65};
66 66
67enum ceph_msg_data_type {
68 CEPH_MSG_DATA_NONE, /* message contains no data payload */
69 CEPH_MSG_DATA_PAGES, /* data source/destination is a page array */
70 CEPH_MSG_DATA_PAGELIST, /* data source/destination is a pagelist */
71#ifdef CONFIG_BLOCK
72 CEPH_MSG_DATA_BIO, /* data source/destination is a bio list */
73#endif /* CONFIG_BLOCK */
74};
75
76static __inline__ bool ceph_msg_data_type_valid(enum ceph_msg_data_type type)
77{
78 switch (type) {
79 case CEPH_MSG_DATA_NONE:
80 case CEPH_MSG_DATA_PAGES:
81 case CEPH_MSG_DATA_PAGELIST:
82#ifdef CONFIG_BLOCK
83 case CEPH_MSG_DATA_BIO:
84#endif /* CONFIG_BLOCK */
85 return true;
86 default:
87 return false;
88 }
89}
90
91struct ceph_msg_data {
92 struct list_head links; /* ceph_msg->data */
93 enum ceph_msg_data_type type;
94 union {
95#ifdef CONFIG_BLOCK
96 struct {
97 struct bio *bio;
98 size_t bio_length;
99 };
100#endif /* CONFIG_BLOCK */
101 struct {
102 struct page **pages; /* NOT OWNER. */
103 size_t length; /* total # bytes */
104 unsigned int alignment; /* first page */
105 };
106 struct ceph_pagelist *pagelist;
107 };
108};
109
110struct ceph_msg_data_cursor {
111 size_t total_resid; /* across all data items */
112 struct list_head *data_head; /* = &ceph_msg->data */
113
114 struct ceph_msg_data *data; /* current data item */
115 size_t resid; /* bytes not yet consumed */
116 bool last_piece; /* current is last piece */
117 bool need_crc; /* crc update needed */
118 union {
119#ifdef CONFIG_BLOCK
120 struct { /* bio */
121 struct bio *bio; /* bio from list */
122 unsigned int vector_index; /* vector from bio */
123 unsigned int vector_offset; /* bytes from vector */
124 };
125#endif /* CONFIG_BLOCK */
126 struct { /* pages */
127 unsigned int page_offset; /* offset in page */
128 unsigned short page_index; /* index in array */
129 unsigned short page_count; /* pages in array */
130 };
131 struct { /* pagelist */
132 struct page *page; /* page from list */
133 size_t offset; /* bytes from list */
134 };
135 };
136};
137
67/* 138/*
68 * a single message. it contains a header (src, dest, message type, etc.), 139 * a single message. it contains a header (src, dest, message type, etc.),
69 * footer (crc values, mainly), a "front" message body, and possibly a 140 * footer (crc values, mainly), a "front" message body, and possibly a
@@ -74,21 +145,15 @@ struct ceph_msg {
74 struct ceph_msg_footer footer; /* footer */ 145 struct ceph_msg_footer footer; /* footer */
75 struct kvec front; /* unaligned blobs of message */ 146 struct kvec front; /* unaligned blobs of message */
76 struct ceph_buffer *middle; 147 struct ceph_buffer *middle;
77 struct page **pages; /* data payload. NOT OWNER. */ 148
78 unsigned nr_pages; /* size of page array */ 149 size_t data_length;
79 unsigned page_alignment; /* io offset in first page */ 150 struct list_head data;
80 struct ceph_pagelist *pagelist; /* instead of pages */ 151 struct ceph_msg_data_cursor cursor;
81 152
82 struct ceph_connection *con; 153 struct ceph_connection *con;
83 struct list_head list_head; 154 struct list_head list_head; /* links for connection lists */
84 155
85 struct kref kref; 156 struct kref kref;
86#ifdef CONFIG_BLOCK
87 struct bio *bio; /* instead of pages/pagelist */
88 struct bio *bio_iter; /* bio iterator */
89 int bio_seg; /* current bio segment */
90#endif /* CONFIG_BLOCK */
91 struct ceph_pagelist *trail; /* the trailing part of the data */
92 bool front_is_vmalloc; 157 bool front_is_vmalloc;
93 bool more_to_follow; 158 bool more_to_follow;
94 bool needs_out_seq; 159 bool needs_out_seq;
@@ -98,12 +163,6 @@ struct ceph_msg {
98 struct ceph_msgpool *pool; 163 struct ceph_msgpool *pool;
99}; 164};
100 165
101struct ceph_msg_pos {
102 int page, page_pos; /* which page; offset in page */
103 int data_pos; /* offset in data payload */
104 bool did_page_crc; /* true if we've calculated crc for current page */
105};
106
107/* ceph connection fault delay defaults, for exponential backoff */ 166/* ceph connection fault delay defaults, for exponential backoff */
108#define BASE_DELAY_INTERVAL (HZ/2) 167#define BASE_DELAY_INTERVAL (HZ/2)
109#define MAX_DELAY_INTERVAL (5 * 60 * HZ) 168#define MAX_DELAY_INTERVAL (5 * 60 * HZ)
@@ -161,7 +220,6 @@ struct ceph_connection {
161 struct ceph_msg *out_msg; /* sending message (== tail of 220 struct ceph_msg *out_msg; /* sending message (== tail of
162 out_sent) */ 221 out_sent) */
163 bool out_msg_done; 222 bool out_msg_done;
164 struct ceph_msg_pos out_msg_pos;
165 223
166 struct kvec out_kvec[8], /* sending header/footer data */ 224 struct kvec out_kvec[8], /* sending header/footer data */
167 *out_kvec_cur; 225 *out_kvec_cur;
@@ -175,7 +233,6 @@ struct ceph_connection {
175 /* message in temps */ 233 /* message in temps */
176 struct ceph_msg_header in_hdr; 234 struct ceph_msg_header in_hdr;
177 struct ceph_msg *in_msg; 235 struct ceph_msg *in_msg;
178 struct ceph_msg_pos in_msg_pos;
179 u32 in_front_crc, in_middle_crc, in_data_crc; /* calculated crc */ 236 u32 in_front_crc, in_middle_crc, in_data_crc; /* calculated crc */
180 237
181 char in_tag; /* protocol control byte */ 238 char in_tag; /* protocol control byte */
@@ -218,6 +275,15 @@ extern void ceph_msg_revoke_incoming(struct ceph_msg *msg);
218 275
219extern void ceph_con_keepalive(struct ceph_connection *con); 276extern void ceph_con_keepalive(struct ceph_connection *con);
220 277
278extern void ceph_msg_data_add_pages(struct ceph_msg *msg, struct page **pages,
279 size_t length, size_t alignment);
280extern void ceph_msg_data_add_pagelist(struct ceph_msg *msg,
281 struct ceph_pagelist *pagelist);
282#ifdef CONFIG_BLOCK
283extern void ceph_msg_data_add_bio(struct ceph_msg *msg, struct bio *bio,
284 size_t length);
285#endif /* CONFIG_BLOCK */
286
221extern struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags, 287extern struct ceph_msg *ceph_msg_new(int type, int front_len, gfp_t flags,
222 bool can_fail); 288 bool can_fail);
223extern void ceph_msg_kfree(struct ceph_msg *m); 289extern void ceph_msg_kfree(struct ceph_msg *m);
diff --git a/include/linux/ceph/msgr.h b/include/linux/ceph/msgr.h
index 680d3d648cac..3d94a73b5f30 100644
--- a/include/linux/ceph/msgr.h
+++ b/include/linux/ceph/msgr.h
@@ -87,6 +87,7 @@ struct ceph_entity_inst {
87#define CEPH_MSGR_TAG_BADPROTOVER 10 /* bad protocol version */ 87#define CEPH_MSGR_TAG_BADPROTOVER 10 /* bad protocol version */
88#define CEPH_MSGR_TAG_BADAUTHORIZER 11 /* bad authorizer */ 88#define CEPH_MSGR_TAG_BADAUTHORIZER 11 /* bad authorizer */
89#define CEPH_MSGR_TAG_FEATURES 12 /* insufficient features */ 89#define CEPH_MSGR_TAG_FEATURES 12 /* insufficient features */
90#define CEPH_MSGR_TAG_SEQ 13 /* 64-bit int follows with seen seq number */
90 91
91 92
92/* 93/*
diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h
index 1dd5d466b6f9..186db0bf4951 100644
--- a/include/linux/ceph/osd_client.h
+++ b/include/linux/ceph/osd_client.h
@@ -29,6 +29,7 @@ struct ceph_authorizer;
29 */ 29 */
30typedef void (*ceph_osdc_callback_t)(struct ceph_osd_request *, 30typedef void (*ceph_osdc_callback_t)(struct ceph_osd_request *,
31 struct ceph_msg *); 31 struct ceph_msg *);
32typedef void (*ceph_osdc_unsafe_callback_t)(struct ceph_osd_request *, bool);
32 33
33/* a given osd we're communicating with */ 34/* a given osd we're communicating with */
34struct ceph_osd { 35struct ceph_osd {
@@ -48,7 +49,67 @@ struct ceph_osd {
48}; 49};
49 50
50 51
51#define CEPH_OSD_MAX_OP 10 52#define CEPH_OSD_MAX_OP 2
53
54enum ceph_osd_data_type {
55 CEPH_OSD_DATA_TYPE_NONE = 0,
56 CEPH_OSD_DATA_TYPE_PAGES,
57 CEPH_OSD_DATA_TYPE_PAGELIST,
58#ifdef CONFIG_BLOCK
59 CEPH_OSD_DATA_TYPE_BIO,
60#endif /* CONFIG_BLOCK */
61};
62
63struct ceph_osd_data {
64 enum ceph_osd_data_type type;
65 union {
66 struct {
67 struct page **pages;
68 u64 length;
69 u32 alignment;
70 bool pages_from_pool;
71 bool own_pages;
72 };
73 struct ceph_pagelist *pagelist;
74#ifdef CONFIG_BLOCK
75 struct {
76 struct bio *bio; /* list of bios */
77 size_t bio_length; /* total in list */
78 };
79#endif /* CONFIG_BLOCK */
80 };
81};
82
83struct ceph_osd_req_op {
84 u16 op; /* CEPH_OSD_OP_* */
85 u32 payload_len;
86 union {
87 struct ceph_osd_data raw_data_in;
88 struct {
89 u64 offset, length;
90 u64 truncate_size;
91 u32 truncate_seq;
92 struct ceph_osd_data osd_data;
93 } extent;
94 struct {
95 const char *class_name;
96 const char *method_name;
97 struct ceph_osd_data request_info;
98 struct ceph_osd_data request_data;
99 struct ceph_osd_data response_data;
100 __u8 class_len;
101 __u8 method_len;
102 __u8 argc;
103 } cls;
104 struct {
105 u64 cookie;
106 u64 ver;
107 u32 prot_ver;
108 u32 timeout;
109 __u8 flag;
110 } watch;
111 };
112};
52 113
53/* an in-flight request */ 114/* an in-flight request */
54struct ceph_osd_request { 115struct ceph_osd_request {
@@ -63,15 +124,14 @@ struct ceph_osd_request {
63 int r_pg_osds[CEPH_PG_MAX_SIZE]; 124 int r_pg_osds[CEPH_PG_MAX_SIZE];
64 int r_num_pg_osds; 125 int r_num_pg_osds;
65 126
66 struct ceph_connection *r_con_filling_msg;
67
68 struct ceph_msg *r_request, *r_reply; 127 struct ceph_msg *r_request, *r_reply;
69 int r_flags; /* any additional flags for the osd */ 128 int r_flags; /* any additional flags for the osd */
70 u32 r_sent; /* >0 if r_request is sending/sent */ 129 u32 r_sent; /* >0 if r_request is sending/sent */
71 int r_num_ops;
72 130
73 /* encoded message content */ 131 /* request osd ops array */
74 struct ceph_osd_op *r_request_ops; 132 unsigned int r_num_ops;
133 struct ceph_osd_req_op r_ops[CEPH_OSD_MAX_OP];
134
75 /* these are updated on each send */ 135 /* these are updated on each send */
76 __le32 *r_request_osdmap_epoch; 136 __le32 *r_request_osdmap_epoch;
77 __le32 *r_request_flags; 137 __le32 *r_request_flags;
@@ -85,12 +145,14 @@ struct ceph_osd_request {
85 s32 r_reply_op_result[CEPH_OSD_MAX_OP]; 145 s32 r_reply_op_result[CEPH_OSD_MAX_OP];
86 int r_got_reply; 146 int r_got_reply;
87 int r_linger; 147 int r_linger;
148 int r_completed;
88 149
89 struct ceph_osd_client *r_osdc; 150 struct ceph_osd_client *r_osdc;
90 struct kref r_kref; 151 struct kref r_kref;
91 bool r_mempool; 152 bool r_mempool;
92 struct completion r_completion, r_safe_completion; 153 struct completion r_completion, r_safe_completion;
93 ceph_osdc_callback_t r_callback, r_safe_callback; 154 ceph_osdc_callback_t r_callback;
155 ceph_osdc_unsafe_callback_t r_unsafe_callback;
94 struct ceph_eversion r_reassert_version; 156 struct ceph_eversion r_reassert_version;
95 struct list_head r_unsafe_item; 157 struct list_head r_unsafe_item;
96 158
@@ -104,16 +166,6 @@ struct ceph_osd_request {
104 166
105 struct ceph_file_layout r_file_layout; 167 struct ceph_file_layout r_file_layout;
106 struct ceph_snap_context *r_snapc; /* snap context for writes */ 168 struct ceph_snap_context *r_snapc; /* snap context for writes */
107 unsigned r_num_pages; /* size of page array (follows) */
108 unsigned r_page_alignment; /* io offset in first page */
109 struct page **r_pages; /* pages for data payload */
110 int r_pages_from_pool;
111 int r_own_pages; /* if true, i own page list */
112#ifdef CONFIG_BLOCK
113 struct bio *r_bio; /* instead of pages */
114#endif
115
116 struct ceph_pagelist r_trail; /* trailing part of the data */
117}; 169};
118 170
119struct ceph_osd_event { 171struct ceph_osd_event {
@@ -172,48 +224,8 @@ struct ceph_osd_client {
172 struct workqueue_struct *notify_wq; 224 struct workqueue_struct *notify_wq;
173}; 225};
174 226
175struct ceph_osd_req_op { 227extern int ceph_osdc_setup(void);
176 u16 op; /* CEPH_OSD_OP_* */ 228extern void ceph_osdc_cleanup(void);
177 u32 payload_len;
178 union {
179 struct {
180 u64 offset, length;
181 u64 truncate_size;
182 u32 truncate_seq;
183 } extent;
184 struct {
185 const char *name;
186 const char *val;
187 u32 name_len;
188 u32 value_len;
189 __u8 cmp_op; /* CEPH_OSD_CMPXATTR_OP_* */
190 __u8 cmp_mode; /* CEPH_OSD_CMPXATTR_MODE_* */
191 } xattr;
192 struct {
193 const char *class_name;
194 const char *method_name;
195 const char *indata;
196 u32 indata_len;
197 __u8 class_len;
198 __u8 method_len;
199 __u8 argc;
200 } cls;
201 struct {
202 u64 cookie;
203 u64 count;
204 } pgls;
205 struct {
206 u64 snapid;
207 } snap;
208 struct {
209 u64 cookie;
210 u64 ver;
211 u32 prot_ver;
212 u32 timeout;
213 __u8 flag;
214 } watch;
215 };
216};
217 229
218extern int ceph_osdc_init(struct ceph_osd_client *osdc, 230extern int ceph_osdc_init(struct ceph_osd_client *osdc,
219 struct ceph_client *client); 231 struct ceph_client *client);
@@ -224,16 +236,71 @@ extern void ceph_osdc_handle_reply(struct ceph_osd_client *osdc,
224extern void ceph_osdc_handle_map(struct ceph_osd_client *osdc, 236extern void ceph_osdc_handle_map(struct ceph_osd_client *osdc,
225 struct ceph_msg *msg); 237 struct ceph_msg *msg);
226 238
239extern void osd_req_op_init(struct ceph_osd_request *osd_req,
240 unsigned int which, u16 opcode);
241
242extern void osd_req_op_raw_data_in_pages(struct ceph_osd_request *,
243 unsigned int which,
244 struct page **pages, u64 length,
245 u32 alignment, bool pages_from_pool,
246 bool own_pages);
247
248extern void osd_req_op_extent_init(struct ceph_osd_request *osd_req,
249 unsigned int which, u16 opcode,
250 u64 offset, u64 length,
251 u64 truncate_size, u32 truncate_seq);
252extern void osd_req_op_extent_update(struct ceph_osd_request *osd_req,
253 unsigned int which, u64 length);
254
255extern struct ceph_osd_data *osd_req_op_extent_osd_data(
256 struct ceph_osd_request *osd_req,
257 unsigned int which);
258extern struct ceph_osd_data *osd_req_op_cls_response_data(
259 struct ceph_osd_request *osd_req,
260 unsigned int which);
261
262extern void osd_req_op_extent_osd_data_pages(struct ceph_osd_request *,
263 unsigned int which,
264 struct page **pages, u64 length,
265 u32 alignment, bool pages_from_pool,
266 bool own_pages);
267extern void osd_req_op_extent_osd_data_pagelist(struct ceph_osd_request *,
268 unsigned int which,
269 struct ceph_pagelist *pagelist);
270#ifdef CONFIG_BLOCK
271extern void osd_req_op_extent_osd_data_bio(struct ceph_osd_request *,
272 unsigned int which,
273 struct bio *bio, size_t bio_length);
274#endif /* CONFIG_BLOCK */
275
276extern void osd_req_op_cls_request_data_pagelist(struct ceph_osd_request *,
277 unsigned int which,
278 struct ceph_pagelist *pagelist);
279extern void osd_req_op_cls_request_data_pages(struct ceph_osd_request *,
280 unsigned int which,
281 struct page **pages, u64 length,
282 u32 alignment, bool pages_from_pool,
283 bool own_pages);
284extern void osd_req_op_cls_response_data_pages(struct ceph_osd_request *,
285 unsigned int which,
286 struct page **pages, u64 length,
287 u32 alignment, bool pages_from_pool,
288 bool own_pages);
289
290extern void osd_req_op_cls_init(struct ceph_osd_request *osd_req,
291 unsigned int which, u16 opcode,
292 const char *class, const char *method);
293extern void osd_req_op_watch_init(struct ceph_osd_request *osd_req,
294 unsigned int which, u16 opcode,
295 u64 cookie, u64 version, int flag);
296
227extern struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, 297extern struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc,
228 struct ceph_snap_context *snapc, 298 struct ceph_snap_context *snapc,
229 unsigned int num_op, 299 unsigned int num_ops,
230 bool use_mempool, 300 bool use_mempool,
231 gfp_t gfp_flags); 301 gfp_t gfp_flags);
232 302
233extern void ceph_osdc_build_request(struct ceph_osd_request *req, 303extern void ceph_osdc_build_request(struct ceph_osd_request *req, u64 off,
234 u64 off, u64 len,
235 unsigned int num_op,
236 struct ceph_osd_req_op *src_ops,
237 struct ceph_snap_context *snapc, 304 struct ceph_snap_context *snapc,
238 u64 snap_id, 305 u64 snap_id,
239 struct timespec *mtime); 306 struct timespec *mtime);
@@ -241,12 +308,11 @@ extern void ceph_osdc_build_request(struct ceph_osd_request *req,
241extern struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *, 308extern struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *,
242 struct ceph_file_layout *layout, 309 struct ceph_file_layout *layout,
243 struct ceph_vino vino, 310 struct ceph_vino vino,
244 u64 offset, u64 *len, int op, int flags, 311 u64 offset, u64 *len,
312 int num_ops, int opcode, int flags,
245 struct ceph_snap_context *snapc, 313 struct ceph_snap_context *snapc,
246 int do_sync, u32 truncate_seq, 314 u32 truncate_seq, u64 truncate_size,
247 u64 truncate_size, 315 bool use_mempool);
248 struct timespec *mtime,
249 bool use_mempool, int page_align);
250 316
251extern void ceph_osdc_set_request_linger(struct ceph_osd_client *osdc, 317extern void ceph_osdc_set_request_linger(struct ceph_osd_client *osdc,
252 struct ceph_osd_request *req); 318 struct ceph_osd_request *req);
diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h
index c819190d1642..d05cc4451af6 100644
--- a/include/linux/ceph/osdmap.h
+++ b/include/linux/ceph/osdmap.h
@@ -3,6 +3,7 @@
3 3
4#include <linux/rbtree.h> 4#include <linux/rbtree.h>
5#include <linux/ceph/types.h> 5#include <linux/ceph/types.h>
6#include <linux/ceph/decode.h>
6#include <linux/ceph/ceph_fs.h> 7#include <linux/ceph/ceph_fs.h>
7#include <linux/crush/crush.h> 8#include <linux/crush/crush.h>
8 9
@@ -119,6 +120,29 @@ static inline struct ceph_entity_addr *ceph_osd_addr(struct ceph_osdmap *map,
119 return &map->osd_addr[osd]; 120 return &map->osd_addr[osd];
120} 121}
121 122
123static inline int ceph_decode_pgid(void **p, void *end, struct ceph_pg *pgid)
124{
125 __u8 version;
126
127 if (!ceph_has_room(p, end, 1 + 8 + 4 + 4)) {
128 pr_warning("incomplete pg encoding");
129
130 return -EINVAL;
131 }
132 version = ceph_decode_8(p);
133 if (version > 1) {
134 pr_warning("do not understand pg encoding %d > 1",
135 (int)version);
136 return -EINVAL;
137 }
138
139 pgid->pool = ceph_decode_64(p);
140 pgid->seed = ceph_decode_32(p);
141 *p += 4; /* skip deprecated preferred value */
142
143 return 0;
144}
145
122extern struct ceph_osdmap *osdmap_decode(void **p, void *end); 146extern struct ceph_osdmap *osdmap_decode(void **p, void *end);
123extern struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end, 147extern struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
124 struct ceph_osdmap *map, 148 struct ceph_osdmap *map,
@@ -131,10 +155,8 @@ extern int ceph_calc_file_object_mapping(struct ceph_file_layout *layout,
131 u64 *bno, u64 *oxoff, u64 *oxlen); 155 u64 *bno, u64 *oxoff, u64 *oxlen);
132 156
133/* calculate mapping of object to a placement group */ 157/* calculate mapping of object to a placement group */
134extern int ceph_calc_object_layout(struct ceph_pg *pg, 158extern int ceph_calc_ceph_pg(struct ceph_pg *pg, const char *oid,
135 const char *oid, 159 struct ceph_osdmap *osdmap, uint64_t pool);
136 struct ceph_file_layout *fl,
137 struct ceph_osdmap *osdmap);
138extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, 160extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap,
139 struct ceph_pg pgid, 161 struct ceph_pg pgid,
140 int *acting); 162 int *acting);
generated by cgit v1.2.2 (git 2.25.0) at 2026-05-15 21:21:03 -0400