Blame view
fs/ceph/caps.c
121 KB
b24413180 License cleanup: ... |
1 |
// SPDX-License-Identifier: GPL-2.0 |
3d14c5d2b ceph: factor out ... |
2 |
#include <linux/ceph/ceph_debug.h> |
a8599bd82 ceph: capability ... |
3 4 5 |
#include <linux/fs.h> #include <linux/kernel.h> |
174cd4b1e sched/headers: Pr... |
6 |
#include <linux/sched/signal.h> |
5a0e3ad6a include cleanup: ... |
7 |
#include <linux/slab.h> |
a8599bd82 ceph: capability ... |
8 9 |
#include <linux/vmalloc.h> #include <linux/wait.h> |
f1a3d5721 ceph: update for ... |
10 |
#include <linux/writeback.h> |
176c77c9c ceph: handle chan... |
11 |
#include <linux/iversion.h> |
a8599bd82 ceph: capability ... |
12 13 |
#include "super.h" |
3d14c5d2b ceph: factor out ... |
14 |
#include "mds_client.h" |
99ccbd229 ceph: use fscache... |
15 |
#include "cache.h" |
3d14c5d2b ceph: factor out ... |
16 17 |
#include <linux/ceph/decode.h> #include <linux/ceph/messenger.h> |
a8599bd82 ceph: capability ... |
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 |
/* * Capability management * * The Ceph metadata servers control client access to inode metadata * and file data by issuing capabilities, granting clients permission * to read and/or write both inode field and file data to OSDs * (storage nodes). Each capability consists of a set of bits * indicating which operations are allowed. * * If the client holds a *_SHARED cap, the client has a coherent value * that can be safely read from the cached inode. * * In the case of a *_EXCL (exclusive) or FILE_WR capabilities, the * client is allowed to change inode attributes (e.g., file size, * mtime), note its dirty state in the ceph_cap, and asynchronously * flush that metadata change to the MDS. * * In the event of a conflicting operation (perhaps by another * client), the MDS will revoke the conflicting client capabilities. * * In order for a client to cache an inode, it must hold a capability * with at least one MDS server. When inodes are released, release * notifications are batched and periodically sent en masse to the MDS * cluster to release server state. */ |
0e2943878 ceph: unify cap f... |
44 |
static u64 __get_oldest_flush_tid(struct ceph_mds_client *mdsc); |
7bc00fddb ceph: kick cap fl... |
45 46 47 48 |
static void __kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session, struct ceph_inode_info *ci, u64 oldest_flush_tid); |
a8599bd82 ceph: capability ... |
49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 |
/* * Generate readable cap strings for debugging output. */ #define MAX_CAP_STR 20 static char cap_str[MAX_CAP_STR][40]; static DEFINE_SPINLOCK(cap_str_lock); static int last_cap_str; static char *gcap_string(char *s, int c) { if (c & CEPH_CAP_GSHARED) *s++ = 's'; if (c & CEPH_CAP_GEXCL) *s++ = 'x'; if (c & CEPH_CAP_GCACHE) *s++ = 'c'; if (c & CEPH_CAP_GRD) *s++ = 'r'; if (c & CEPH_CAP_GWR) *s++ = 'w'; if (c & CEPH_CAP_GBUFFER) *s++ = 'b'; |
49a9f4f67 ceph: always get ... |
72 73 |
if (c & CEPH_CAP_GWREXTEND) *s++ = 'a'; |
a8599bd82 ceph: capability ... |
74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 |
if (c & CEPH_CAP_GLAZYIO) *s++ = 'l'; return s; } const char *ceph_cap_string(int caps) { int i; char *s; int c; spin_lock(&cap_str_lock); i = last_cap_str++; if (last_cap_str == MAX_CAP_STR) last_cap_str = 0; spin_unlock(&cap_str_lock); s = cap_str[i]; if (caps & CEPH_CAP_PIN) *s++ = 'p'; c = (caps >> CEPH_CAP_SAUTH) & 3; if (c) { *s++ = 'A'; s = gcap_string(s, c); } c = (caps >> CEPH_CAP_SLINK) & 3; if (c) { *s++ = 'L'; s = gcap_string(s, c); } c = (caps >> CEPH_CAP_SXATTR) & 3; if (c) { *s++ = 'X'; s = gcap_string(s, c); } c = caps >> CEPH_CAP_SFILE; if (c) { *s++ = 'F'; s = gcap_string(s, c); } if (s == cap_str[i]) *s++ = '-'; *s = 0; return cap_str[i]; } |
37151668b ceph: do caps acc... |
125 |
void ceph_caps_init(struct ceph_mds_client *mdsc) |
a8599bd82 ceph: capability ... |
126 |
{ |
37151668b ceph: do caps acc... |
127 128 |
INIT_LIST_HEAD(&mdsc->caps_list); spin_lock_init(&mdsc->caps_list_lock); |
a8599bd82 ceph: capability ... |
129 |
} |
37151668b ceph: do caps acc... |
130 |
void ceph_caps_finalize(struct ceph_mds_client *mdsc) |
a8599bd82 ceph: capability ... |
131 132 |
{ struct ceph_cap *cap; |
37151668b ceph: do caps acc... |
133 134 135 136 |
spin_lock(&mdsc->caps_list_lock); while (!list_empty(&mdsc->caps_list)) { cap = list_first_entry(&mdsc->caps_list, struct ceph_cap, caps_item); |
a8599bd82 ceph: capability ... |
137 138 139 |
list_del(&cap->caps_item); kmem_cache_free(ceph_cap_cachep, cap); } |
37151668b ceph: do caps acc... |
140 141 142 143 144 145 |
mdsc->caps_total_count = 0; mdsc->caps_avail_count = 0; mdsc->caps_use_count = 0; mdsc->caps_reserve_count = 0; mdsc->caps_min_count = 0; spin_unlock(&mdsc->caps_list_lock); |
85ccce43a ceph: clean up re... |
146 |
} |
fe33032da ceph: add mount o... |
147 148 |
void ceph_adjust_caps_max_min(struct ceph_mds_client *mdsc, struct ceph_mount_options *fsopt) |
85ccce43a ceph: clean up re... |
149 |
{ |
37151668b ceph: do caps acc... |
150 |
spin_lock(&mdsc->caps_list_lock); |
fe33032da ceph: add mount o... |
151 152 153 154 155 156 157 |
mdsc->caps_min_count = fsopt->max_readdir; if (mdsc->caps_min_count < 1024) mdsc->caps_min_count = 1024; mdsc->caps_use_max = fsopt->caps_max; if (mdsc->caps_use_max > 0 && mdsc->caps_use_max < mdsc->caps_min_count) mdsc->caps_use_max = mdsc->caps_min_count; |
37151668b ceph: do caps acc... |
158 |
spin_unlock(&mdsc->caps_list_lock); |
a8599bd82 ceph: capability ... |
159 |
} |
7bf8f736c ceph: refactor ce... |
160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 |
static void __ceph_unreserve_caps(struct ceph_mds_client *mdsc, int nr_caps) { struct ceph_cap *cap; int i; if (nr_caps) { BUG_ON(mdsc->caps_reserve_count < nr_caps); mdsc->caps_reserve_count -= nr_caps; if (mdsc->caps_avail_count >= mdsc->caps_reserve_count + mdsc->caps_min_count) { mdsc->caps_total_count -= nr_caps; for (i = 0; i < nr_caps; i++) { cap = list_first_entry(&mdsc->caps_list, struct ceph_cap, caps_item); list_del(&cap->caps_item); kmem_cache_free(ceph_cap_cachep, cap); } } else { mdsc->caps_avail_count += nr_caps; } dout("%s: caps %d = %d used + %d resv + %d avail ", __func__, mdsc->caps_total_count, mdsc->caps_use_count, mdsc->caps_reserve_count, mdsc->caps_avail_count); BUG_ON(mdsc->caps_total_count != mdsc->caps_use_count + mdsc->caps_reserve_count + mdsc->caps_avail_count); } } |
e30ee5812 ceph: try to allo... |
191 192 193 194 |
/* * Called under mdsc->mutex. */ int ceph_reserve_caps(struct ceph_mds_client *mdsc, |
37151668b ceph: do caps acc... |
195 |
struct ceph_cap_reservation *ctx, int need) |
a8599bd82 ceph: capability ... |
196 |
{ |
e30ee5812 ceph: try to allo... |
197 |
int i, j; |
a8599bd82 ceph: capability ... |
198 199 200 |
struct ceph_cap *cap; int have; int alloc = 0; |
e30ee5812 ceph: try to allo... |
201 |
int max_caps; |
e5bc08d09 ceph: refactor er... |
202 |
int err = 0; |
e30ee5812 ceph: try to allo... |
203 204 |
bool trimmed = false; struct ceph_mds_session *s; |
a8599bd82 ceph: capability ... |
205 |
LIST_HEAD(newcaps); |
a8599bd82 ceph: capability ... |
206 207 208 209 210 |
dout("reserve caps ctx=%p need=%d ", ctx, need); /* first reserve any caps that are already allocated */ |
37151668b ceph: do caps acc... |
211 212 |
spin_lock(&mdsc->caps_list_lock); if (mdsc->caps_avail_count >= need) |
a8599bd82 ceph: capability ... |
213 214 |
have = need; else |
37151668b ceph: do caps acc... |
215 216 217 218 219 220 221 |
have = mdsc->caps_avail_count; mdsc->caps_avail_count -= have; mdsc->caps_reserve_count += have; BUG_ON(mdsc->caps_total_count != mdsc->caps_use_count + mdsc->caps_reserve_count + mdsc->caps_avail_count); spin_unlock(&mdsc->caps_list_lock); |
a8599bd82 ceph: capability ... |
222 |
|
79cd674ae ceph: optimizing ... |
223 |
for (i = have; i < need; ) { |
a8599bd82 ceph: capability ... |
224 |
cap = kmem_cache_alloc(ceph_cap_cachep, GFP_NOFS); |
79cd674ae ceph: optimizing ... |
225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 |
if (cap) { list_add(&cap->caps_item, &newcaps); alloc++; i++; continue; } if (!trimmed) { for (j = 0; j < mdsc->max_sessions; j++) { s = __ceph_lookup_mds_session(mdsc, j); if (!s) continue; mutex_unlock(&mdsc->mutex); mutex_lock(&s->s_mutex); max_caps = s->s_nr_caps - (need - i); ceph_trim_caps(mdsc, s, max_caps); mutex_unlock(&s->s_mutex); ceph_put_mds_session(s); mutex_lock(&mdsc->mutex); |
e30ee5812 ceph: try to allo... |
246 |
} |
79cd674ae ceph: optimizing ... |
247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 |
trimmed = true; spin_lock(&mdsc->caps_list_lock); if (mdsc->caps_avail_count) { int more_have; if (mdsc->caps_avail_count >= need - i) more_have = need - i; else more_have = mdsc->caps_avail_count; i += more_have; have += more_have; mdsc->caps_avail_count -= more_have; mdsc->caps_reserve_count += more_have; } spin_unlock(&mdsc->caps_list_lock); continue; |
e30ee5812 ceph: try to allo... |
266 |
} |
79cd674ae ceph: optimizing ... |
267 268 269 270 |
pr_warn("reserve caps ctx=%p ENOMEM need=%d got=%d ", ctx, need, have + alloc); |
e5bc08d09 ceph: refactor er... |
271 272 273 274 275 276 277 |
err = -ENOMEM; break; } if (!err) { BUG_ON(have + alloc != need); ctx->count = need; |
fe33032da ceph: add mount o... |
278 |
ctx->used = 0; |
a8599bd82 ceph: capability ... |
279 |
} |
a8599bd82 ceph: capability ... |
280 |
|
37151668b ceph: do caps acc... |
281 282 283 284 |
spin_lock(&mdsc->caps_list_lock); mdsc->caps_total_count += alloc; mdsc->caps_reserve_count += alloc; list_splice(&newcaps, &mdsc->caps_list); |
a8599bd82 ceph: capability ... |
285 |
|
37151668b ceph: do caps acc... |
286 287 288 |
BUG_ON(mdsc->caps_total_count != mdsc->caps_use_count + mdsc->caps_reserve_count + mdsc->caps_avail_count); |
e5bc08d09 ceph: refactor er... |
289 290 291 |
if (err) __ceph_unreserve_caps(mdsc, have + alloc); |
37151668b ceph: do caps acc... |
292 |
spin_unlock(&mdsc->caps_list_lock); |
a8599bd82 ceph: capability ... |
293 |
|
a8599bd82 ceph: capability ... |
294 295 |
dout("reserve caps ctx=%p %d = %d used + %d resv + %d avail ", |
37151668b ceph: do caps acc... |
296 297 |
ctx, mdsc->caps_total_count, mdsc->caps_use_count, mdsc->caps_reserve_count, mdsc->caps_avail_count); |
e5bc08d09 ceph: refactor er... |
298 |
return err; |
a8599bd82 ceph: capability ... |
299 |
} |
7bf8f736c ceph: refactor ce... |
300 |
void ceph_unreserve_caps(struct ceph_mds_client *mdsc, |
fe33032da ceph: add mount o... |
301 |
struct ceph_cap_reservation *ctx) |
a8599bd82 ceph: capability ... |
302 |
{ |
fe33032da ceph: add mount o... |
303 304 305 |
bool reclaim = false; if (!ctx->count) return; |
a8599bd82 ceph: capability ... |
306 307 |
dout("unreserve caps ctx=%p count=%d ", ctx, ctx->count); |
7bf8f736c ceph: refactor ce... |
308 309 310 |
spin_lock(&mdsc->caps_list_lock); __ceph_unreserve_caps(mdsc, ctx->count); ctx->count = 0; |
fe33032da ceph: add mount o... |
311 312 313 314 |
if (mdsc->caps_use_max > 0 && mdsc->caps_use_count > mdsc->caps_use_max) reclaim = true; |
7bf8f736c ceph: refactor ce... |
315 |
spin_unlock(&mdsc->caps_list_lock); |
fe33032da ceph: add mount o... |
316 317 318 |
if (reclaim) ceph_reclaim_caps_nr(mdsc, ctx->used); |
a8599bd82 ceph: capability ... |
319 |
} |
d9df27835 ceph: pre-allocat... |
320 321 |
struct ceph_cap *ceph_get_cap(struct ceph_mds_client *mdsc, struct ceph_cap_reservation *ctx) |
a8599bd82 ceph: capability ... |
322 323 324 325 |
{ struct ceph_cap *cap = NULL; /* temporary, until we do something about cap import/export */ |
443b3760a ceph: fix caps us... |
326 327 328 |
if (!ctx) { cap = kmem_cache_alloc(ceph_cap_cachep, GFP_NOFS); if (cap) { |
4d1d0534f ceph: Hold caps_l... |
329 |
spin_lock(&mdsc->caps_list_lock); |
37151668b ceph: do caps acc... |
330 331 |
mdsc->caps_use_count++; mdsc->caps_total_count++; |
4d1d0534f ceph: Hold caps_l... |
332 |
spin_unlock(&mdsc->caps_list_lock); |
e327ce068 ceph: optimizing ... |
333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 |
} else { spin_lock(&mdsc->caps_list_lock); if (mdsc->caps_avail_count) { BUG_ON(list_empty(&mdsc->caps_list)); mdsc->caps_avail_count--; mdsc->caps_use_count++; cap = list_first_entry(&mdsc->caps_list, struct ceph_cap, caps_item); list_del(&cap->caps_item); BUG_ON(mdsc->caps_total_count != mdsc->caps_use_count + mdsc->caps_reserve_count + mdsc->caps_avail_count); } spin_unlock(&mdsc->caps_list_lock); |
443b3760a ceph: fix caps us... |
348 |
} |
e327ce068 ceph: optimizing ... |
349 |
|
443b3760a ceph: fix caps us... |
350 351 |
return cap; } |
a8599bd82 ceph: capability ... |
352 |
|
37151668b ceph: do caps acc... |
353 |
spin_lock(&mdsc->caps_list_lock); |
a8599bd82 ceph: capability ... |
354 355 |
dout("get_cap ctx=%p (%d) %d = %d used + %d resv + %d avail ", |
37151668b ceph: do caps acc... |
356 357 |
ctx, ctx->count, mdsc->caps_total_count, mdsc->caps_use_count, mdsc->caps_reserve_count, mdsc->caps_avail_count); |
a8599bd82 ceph: capability ... |
358 |
BUG_ON(!ctx->count); |
37151668b ceph: do caps acc... |
359 360 |
BUG_ON(ctx->count > mdsc->caps_reserve_count); BUG_ON(list_empty(&mdsc->caps_list)); |
a8599bd82 ceph: capability ... |
361 362 |
ctx->count--; |
fe33032da ceph: add mount o... |
363 |
ctx->used++; |
37151668b ceph: do caps acc... |
364 365 |
mdsc->caps_reserve_count--; mdsc->caps_use_count++; |
a8599bd82 ceph: capability ... |
366 |
|
37151668b ceph: do caps acc... |
367 |
cap = list_first_entry(&mdsc->caps_list, struct ceph_cap, caps_item); |
a8599bd82 ceph: capability ... |
368 |
list_del(&cap->caps_item); |
37151668b ceph: do caps acc... |
369 370 371 |
BUG_ON(mdsc->caps_total_count != mdsc->caps_use_count + mdsc->caps_reserve_count + mdsc->caps_avail_count); spin_unlock(&mdsc->caps_list_lock); |
a8599bd82 ceph: capability ... |
372 373 |
return cap; } |
37151668b ceph: do caps acc... |
374 |
void ceph_put_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap) |
a8599bd82 ceph: capability ... |
375 |
{ |
37151668b ceph: do caps acc... |
376 |
spin_lock(&mdsc->caps_list_lock); |
7c1332b8c ceph: fix iterate... |
377 378 |
dout("put_cap %p %d = %d used + %d resv + %d avail ", |
37151668b ceph: do caps acc... |
379 380 381 |
cap, mdsc->caps_total_count, mdsc->caps_use_count, mdsc->caps_reserve_count, mdsc->caps_avail_count); mdsc->caps_use_count--; |
a8599bd82 ceph: capability ... |
382 |
/* |
85ccce43a ceph: clean up re... |
383 384 |
* Keep some preallocated caps around (ceph_min_count), to * avoid lots of free/alloc churn. |
a8599bd82 ceph: capability ... |
385 |
*/ |
37151668b ceph: do caps acc... |
386 387 388 |
if (mdsc->caps_avail_count >= mdsc->caps_reserve_count + mdsc->caps_min_count) { mdsc->caps_total_count--; |
a8599bd82 ceph: capability ... |
389 390 |
kmem_cache_free(ceph_cap_cachep, cap); } else { |
37151668b ceph: do caps acc... |
391 392 |
mdsc->caps_avail_count++; list_add(&cap->caps_item, &mdsc->caps_list); |
a8599bd82 ceph: capability ... |
393 |
} |
37151668b ceph: do caps acc... |
394 395 396 |
BUG_ON(mdsc->caps_total_count != mdsc->caps_use_count + mdsc->caps_reserve_count + mdsc->caps_avail_count); spin_unlock(&mdsc->caps_list_lock); |
a8599bd82 ceph: capability ... |
397 |
} |
3d14c5d2b ceph: factor out ... |
398 |
void ceph_reservation_status(struct ceph_fs_client *fsc, |
85ccce43a ceph: clean up re... |
399 400 |
int *total, int *avail, int *used, int *reserved, int *min) |
a8599bd82 ceph: capability ... |
401 |
{ |
3d14c5d2b ceph: factor out ... |
402 |
struct ceph_mds_client *mdsc = fsc->mdsc; |
37151668b ceph: do caps acc... |
403 |
|
b884014a9 ceph: adding prot... |
404 |
spin_lock(&mdsc->caps_list_lock); |
a8599bd82 ceph: capability ... |
405 |
if (total) |
37151668b ceph: do caps acc... |
406 |
*total = mdsc->caps_total_count; |
a8599bd82 ceph: capability ... |
407 |
if (avail) |
37151668b ceph: do caps acc... |
408 |
*avail = mdsc->caps_avail_count; |
a8599bd82 ceph: capability ... |
409 |
if (used) |
37151668b ceph: do caps acc... |
410 |
*used = mdsc->caps_use_count; |
a8599bd82 ceph: capability ... |
411 |
if (reserved) |
37151668b ceph: do caps acc... |
412 |
*reserved = mdsc->caps_reserve_count; |
85ccce43a ceph: clean up re... |
413 |
if (min) |
37151668b ceph: do caps acc... |
414 |
*min = mdsc->caps_min_count; |
b884014a9 ceph: adding prot... |
415 416 |
spin_unlock(&mdsc->caps_list_lock); |
a8599bd82 ceph: capability ... |
417 418 419 420 421 |
} /* * Find ceph_cap for given mds, if any. * |
be655596b ceph: use i_ceph_... |
422 |
* Called with i_ceph_lock held. |
a8599bd82 ceph: capability ... |
423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 |
*/ static struct ceph_cap *__get_cap_for_mds(struct ceph_inode_info *ci, int mds) { struct ceph_cap *cap; struct rb_node *n = ci->i_caps.rb_node; while (n) { cap = rb_entry(n, struct ceph_cap, ci_node); if (mds < cap->mds) n = n->rb_left; else if (mds > cap->mds) n = n->rb_right; else return cap; } return NULL; } |
2bc50259f ceph: add ceph_ge... |
440 441 442 |
struct ceph_cap *ceph_get_cap_for_mds(struct ceph_inode_info *ci, int mds) { struct ceph_cap *cap; |
be655596b ceph: use i_ceph_... |
443 |
spin_lock(&ci->i_ceph_lock); |
2bc50259f ceph: add ceph_ge... |
444 |
cap = __get_cap_for_mds(ci, mds); |
be655596b ceph: use i_ceph_... |
445 |
spin_unlock(&ci->i_ceph_lock); |
2bc50259f ceph: add ceph_ge... |
446 447 |
return cap; } |
a8599bd82 ceph: capability ... |
448 |
/* |
be655596b ceph: use i_ceph_... |
449 |
* Called under i_ceph_lock. |
a8599bd82 ceph: capability ... |
450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 |
*/ static void __insert_cap_node(struct ceph_inode_info *ci, struct ceph_cap *new) { struct rb_node **p = &ci->i_caps.rb_node; struct rb_node *parent = NULL; struct ceph_cap *cap = NULL; while (*p) { parent = *p; cap = rb_entry(parent, struct ceph_cap, ci_node); if (new->mds < cap->mds) p = &(*p)->rb_left; else if (new->mds > cap->mds) p = &(*p)->rb_right; else BUG(); } rb_link_node(&new->ci_node, parent, p); rb_insert_color(&new->ci_node, &ci->i_caps); } /* * (re)set cap hold timeouts, which control the delayed release * of unused caps back to the MDS. Should be called on cap use. */ static void __cap_set_timeouts(struct ceph_mds_client *mdsc, struct ceph_inode_info *ci) { |
fe33032da ceph: add mount o... |
480 |
struct ceph_mount_options *opt = mdsc->fsc->mount_options; |
a8599bd82 ceph: capability ... |
481 |
ci->i_hold_caps_max = round_jiffies(jiffies + |
fe33032da ceph: add mount o... |
482 |
opt->caps_wanted_delay_max * HZ); |
a0d93e327 ceph: remove dela... |
483 484 485 |
dout("__cap_set_timeouts %p %lu ", &ci->vfs_inode, ci->i_hold_caps_max - jiffies); |
a8599bd82 ceph: capability ... |
486 487 488 489 490 491 492 |
} /* * (Re)queue cap at the end of the delayed cap release list. * * If I_FLUSH is set, leave the inode at the front of the list. * |
be655596b ceph: use i_ceph_... |
493 |
* Caller holds i_ceph_lock |
a8599bd82 ceph: capability ... |
494 495 496 |
* -> we take mdsc->cap_delay_lock */ static void __cap_delay_requeue(struct ceph_mds_client *mdsc, |
a0d93e327 ceph: remove dela... |
497 |
struct ceph_inode_info *ci) |
a8599bd82 ceph: capability ... |
498 |
{ |
891f3f5a6 ceph: add infrast... |
499 500 |
dout("__cap_delay_requeue %p flags 0x%lx at %lu ", &ci->vfs_inode, |
a8599bd82 ceph: capability ... |
501 502 503 504 505 506 507 508 |
ci->i_ceph_flags, ci->i_hold_caps_max); if (!mdsc->stopping) { spin_lock(&mdsc->cap_delay_lock); if (!list_empty(&ci->i_cap_delay_list)) { if (ci->i_ceph_flags & CEPH_I_FLUSH) goto no_change; list_del_init(&ci->i_cap_delay_list); } |
a0d93e327 ceph: remove dela... |
509 |
__cap_set_timeouts(mdsc, ci); |
a8599bd82 ceph: capability ... |
510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 |
list_add_tail(&ci->i_cap_delay_list, &mdsc->cap_delay_list); no_change: spin_unlock(&mdsc->cap_delay_lock); } } /* * Queue an inode for immediate writeback. Mark inode with I_FLUSH, * indicating we should send a cap message to flush dirty metadata * asap, and move to the front of the delayed cap list. */ static void __cap_delay_requeue_front(struct ceph_mds_client *mdsc, struct ceph_inode_info *ci) { dout("__cap_delay_requeue_front %p ", &ci->vfs_inode); spin_lock(&mdsc->cap_delay_lock); ci->i_ceph_flags |= CEPH_I_FLUSH; if (!list_empty(&ci->i_cap_delay_list)) list_del_init(&ci->i_cap_delay_list); list_add(&ci->i_cap_delay_list, &mdsc->cap_delay_list); spin_unlock(&mdsc->cap_delay_lock); } /* * Cancel delayed work on cap. * |
be655596b ceph: use i_ceph_... |
537 |
* Caller must hold i_ceph_lock. |
a8599bd82 ceph: capability ... |
538 539 540 541 542 543 544 545 546 547 548 549 |
*/ static void __cap_delay_cancel(struct ceph_mds_client *mdsc, struct ceph_inode_info *ci) { dout("__cap_delay_cancel %p ", &ci->vfs_inode); if (list_empty(&ci->i_cap_delay_list)) return; spin_lock(&mdsc->cap_delay_lock); list_del_init(&ci->i_cap_delay_list); spin_unlock(&mdsc->cap_delay_lock); } |
785892fe8 ceph: cache layou... |
550 |
/* Common issue checks for add_cap, handle_cap_grant. */ |
a8599bd82 ceph: capability ... |
551 552 553 554 |
static void __check_cap_issue(struct ceph_inode_info *ci, struct ceph_cap *cap, unsigned issued) { unsigned had = __ceph_caps_issued(ci, NULL); |
785892fe8 ceph: cache layou... |
555 |
lockdep_assert_held(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
556 557 558 559 |
/* * Each time we receive FILE_CACHE anew, we increment * i_rdcache_gen. */ |
525d15e8e ceph: check inode... |
560 561 |
if (S_ISREG(ci->vfs_inode.i_mode) && (issued & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) && |
99ccbd229 ceph: use fscache... |
562 |
(had & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) == 0) { |
a8599bd82 ceph: capability ... |
563 |
ci->i_rdcache_gen++; |
99ccbd229 ceph: use fscache... |
564 |
} |
a8599bd82 ceph: capability ... |
565 566 |
/* |
15b51bd6b ceph: stop on-goi... |
567 568 569 570 |
* If FILE_SHARED is newly issued, mark dir not complete. We don't * know what happened to this directory while we didn't have the cap. * If FILE_SHARED is being revoked, also mark dir not complete. It * stops on-going cached readdir. |
a8599bd82 ceph: capability ... |
571 |
*/ |
15b51bd6b ceph: stop on-goi... |
572 573 |
if ((issued & CEPH_CAP_FILE_SHARED) != (had & CEPH_CAP_FILE_SHARED)) { if (issued & CEPH_CAP_FILE_SHARED) |
97aeb6bf9 ceph: use atomic_... |
574 |
atomic_inc(&ci->i_shared_gen); |
a8673d61a ceph: use I_COMPL... |
575 576 577 |
if (S_ISDIR(ci->vfs_inode.i_mode)) { dout(" marking %p NOT complete ", &ci->vfs_inode); |
2f276c511 ceph: use i_relea... |
578 |
__ceph_dir_clear_complete(ci); |
a8673d61a ceph: use I_COMPL... |
579 |
} |
a8599bd82 ceph: capability ... |
580 |
} |
785892fe8 ceph: cache layou... |
581 582 583 584 585 586 587 |
/* Wipe saved layout if we're losing DIR_CREATE caps */ if (S_ISDIR(ci->vfs_inode.i_mode) && (had & CEPH_CAP_DIR_CREATE) && !(issued & CEPH_CAP_DIR_CREATE)) { ceph_put_string(rcu_dereference_raw(ci->i_cached_layout.pool_ns)); memset(&ci->i_cached_layout, 0, sizeof(ci->i_cached_layout)); } |
a8599bd82 ceph: capability ... |
588 |
} |
1cf03a68e ceph: convert mds... |
589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 |
/** * change_auth_cap_ses - move inode to appropriate lists when auth caps change * @ci: inode to be moved * @session: new auth caps session */ static void change_auth_cap_ses(struct ceph_inode_info *ci, struct ceph_mds_session *session) { lockdep_assert_held(&ci->i_ceph_lock); if (list_empty(&ci->i_dirty_item) && list_empty(&ci->i_flushing_item)) return; spin_lock(&session->s_mdsc->cap_dirty_lock); if (!list_empty(&ci->i_dirty_item)) list_move(&ci->i_dirty_item, &session->s_cap_dirty); if (!list_empty(&ci->i_flushing_item)) list_move_tail(&ci->i_flushing_item, &session->s_cap_flushing); spin_unlock(&session->s_mdsc->cap_dirty_lock); } |
a8599bd82 ceph: capability ... |
609 610 611 |
/* * Add a capability under the given MDS session. * |
354c63a00 ceph: fix comment... |
612 |
* Caller should hold session snap_rwsem (read) and ci->i_ceph_lock |
a8599bd82 ceph: capability ... |
613 614 615 616 617 |
* * @fmode is the open file mode, if we are opening a file, otherwise * it is < 0. (This is so we can atomically add the cap and add an * open file reference to it.) */ |
d9df27835 ceph: pre-allocat... |
618 619 |
void ceph_add_cap(struct inode *inode, struct ceph_mds_session *session, u64 cap_id, |
135e671e5 ceph: simplify ca... |
620 |
unsigned issued, unsigned wanted, |
d9df27835 ceph: pre-allocat... |
621 622 |
unsigned seq, unsigned mseq, u64 realmino, int flags, struct ceph_cap **new_cap) |
a8599bd82 ceph: capability ... |
623 |
{ |
3d14c5d2b ceph: factor out ... |
624 |
struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; |
a8599bd82 ceph: capability ... |
625 |
struct ceph_inode_info *ci = ceph_inode(inode); |
a8599bd82 ceph: capability ... |
626 627 628 |
struct ceph_cap *cap; int mds = session->s_mds; int actual_wanted; |
606d10232 ceph: fetch cap_g... |
629 |
u32 gen; |
a8599bd82 ceph: capability ... |
630 |
|
354c63a00 ceph: fix comment... |
631 |
lockdep_assert_held(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
632 633 634 |
dout("add_cap %p mds%d cap %llx %s seq %d ", inode, session->s_mds, cap_id, ceph_cap_string(issued), seq); |
606d10232 ceph: fetch cap_g... |
635 636 637 |
spin_lock(&session->s_gen_ttl_lock); gen = session->s_cap_gen; spin_unlock(&session->s_gen_ttl_lock); |
a8599bd82 ceph: capability ... |
638 639 |
cap = __get_cap_for_mds(ci, mds); if (!cap) { |
d9df27835 ceph: pre-allocat... |
640 641 |
cap = *new_cap; *new_cap = NULL; |
a8599bd82 ceph: capability ... |
642 643 644 645 646 |
cap->issued = 0; cap->implemented = 0; cap->mds = mds; cap->mds_wanted = 0; |
964266cce ceph: set mds_wan... |
647 |
cap->mseq = 0; |
a8599bd82 ceph: capability ... |
648 649 650 |
cap->ci = ci; __insert_cap_node(ci, cap); |
a8599bd82 ceph: capability ... |
651 652 653 654 655 |
/* add to session cap list */ cap->session = session; spin_lock(&session->s_cap_lock); list_add_tail(&cap->session_caps, &session->s_caps); session->s_nr_caps++; |
4f1d756de ceph: add global ... |
656 |
atomic64_inc(&mdsc->metric.total_caps); |
a8599bd82 ceph: capability ... |
657 |
spin_unlock(&session->s_cap_lock); |
11df2dfb6 ceph: add importe... |
658 |
} else { |
32f6511a6 ceph: touch exist... |
659 660 661 |
spin_lock(&session->s_cap_lock); list_move_tail(&cap->session_caps, &session->s_caps); spin_unlock(&session->s_cap_lock); |
606d10232 ceph: fetch cap_g... |
662 |
if (cap->cap_gen < gen) |
d2f8bb27c ceph: update want... |
663 |
cap->issued = cap->implemented = CEPH_CAP_PIN; |
11df2dfb6 ceph: add importe... |
664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 |
/* * auth mds of the inode changed. we received the cap export * message, but still haven't received the cap import message. * handle_cap_export() updated the new auth MDS' cap. * * "ceph_seq_cmp(seq, cap->seq) <= 0" means we are processing * a message that was send before the cap import message. So * don't remove caps. */ if (ceph_seq_cmp(seq, cap->seq) <= 0) { WARN_ON(cap != ci->i_auth_cap); WARN_ON(cap->cap_id != cap_id); seq = cap->seq; mseq = cap->mseq; issued |= cap->issued; flags |= CEPH_CAP_FLAG_AUTH; } } |
a8599bd82 ceph: capability ... |
682 |
|
7d9c9193b ceph: fix incorre... |
683 684 685 |
if (!ci->i_snap_realm || ((flags & CEPH_CAP_FLAG_AUTH) && realmino != (u64)-1 && ci->i_snap_realm->ino != realmino)) { |
a8599bd82 ceph: capability ... |
686 687 688 689 690 691 |
/* * add this inode to the appropriate snap realm */ struct ceph_snap_realm *realm = ceph_lookup_snap_realm(mdsc, realmino); if (realm) { |
7d9c9193b ceph: fix incorre... |
692 693 694 695 696 697 |
struct ceph_snap_realm *oldrealm = ci->i_snap_realm; if (oldrealm) { spin_lock(&oldrealm->inodes_with_caps_lock); list_del_init(&ci->i_snap_realm_item); spin_unlock(&oldrealm->inodes_with_caps_lock); } |
a8599bd82 ceph: capability ... |
698 |
spin_lock(&realm->inodes_with_caps_lock); |
a8599bd82 ceph: capability ... |
699 700 |
list_add(&ci->i_snap_realm_item, &realm->inodes_with_caps); |
e3161f17d ceph: quota: cach... |
701 702 703 |
ci->i_snap_realm = realm; if (realm->ino == ci->i_vino.ino) realm->inode = inode; |
a8599bd82 ceph: capability ... |
704 |
spin_unlock(&realm->inodes_with_caps_lock); |
7d9c9193b ceph: fix incorre... |
705 706 707 |
if (oldrealm) ceph_put_snap_realm(mdsc, oldrealm); |
a8599bd82 ceph: capability ... |
708 709 710 711 |
} else { pr_err("ceph_add_cap: couldn't find snap realm %llx ", realmino); |
b8cd07e78 ceph: warn on mis... |
712 |
WARN_ON(!realm); |
a8599bd82 ceph: capability ... |
713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 |
} } __check_cap_issue(ci, cap, issued); /* * If we are issued caps we don't want, or the mds' wanted * value appears to be off, queue a check so we'll release * later and/or update the mds wanted value. */ actual_wanted = __ceph_caps_wanted(ci); if ((wanted & ~actual_wanted) || (issued & ~actual_wanted & CEPH_CAP_ANY_WR)) { dout(" issued %s, mds wanted %s, actual %s, queueing ", ceph_cap_string(issued), ceph_cap_string(wanted), ceph_cap_string(actual_wanted)); |
a0d93e327 ceph: remove dela... |
730 |
__cap_delay_requeue(mdsc, ci); |
a8599bd82 ceph: capability ... |
731 |
} |
b8c2f3ae2 ceph: check migra... |
732 |
if (flags & CEPH_CAP_FLAG_AUTH) { |
d37b1d994 ceph: adjust 36 c... |
733 |
if (!ci->i_auth_cap || |
d9ffc4f77 ceph: set mds_wan... |
734 |
ceph_seq_cmp(ci->i_auth_cap->mseq, mseq) < 0) { |
1cf03a68e ceph: convert mds... |
735 736 737 |
if (ci->i_auth_cap && ci->i_auth_cap->session != cap->session) change_auth_cap_ses(ci, cap->session); |
b8c2f3ae2 ceph: check migra... |
738 |
ci->i_auth_cap = cap; |
d9ffc4f77 ceph: set mds_wan... |
739 740 |
cap->mds_wanted = wanted; } |
11df2dfb6 ceph: add importe... |
741 742 |
} else { WARN_ON(ci->i_auth_cap == cap); |
8a92a119b ceph: move dirty ... |
743 |
} |
a8599bd82 ceph: capability ... |
744 745 746 747 748 749 750 751 |
dout("add_cap inode %p (%llx.%llx) cap %p %s now %s seq %d mds%d ", inode, ceph_vinop(inode), cap, ceph_cap_string(issued), ceph_cap_string(issued|cap->issued), seq, mds); cap->cap_id = cap_id; cap->issued = issued; cap->implemented |= issued; |
d1b87809f ceph: use ceph_se... |
752 |
if (ceph_seq_cmp(mseq, cap->mseq) > 0) |
964266cce ceph: set mds_wan... |
753 754 755 |
cap->mds_wanted = wanted; else cap->mds_wanted |= wanted; |
a8599bd82 ceph: capability ... |
756 757 758 |
cap->seq = seq; cap->issue_seq = seq; cap->mseq = mseq; |
606d10232 ceph: fetch cap_g... |
759 |
cap->cap_gen = gen; |
a8599bd82 ceph: capability ... |
760 761 762 763 764 765 766 767 768 769 |
} /* * Return true if cap has not timed out and belongs to the current * generation of the MDS session (i.e. has not gone 'stale' due to * us losing touch with the mds). */ static int __cap_is_valid(struct ceph_cap *cap) { unsigned long ttl; |
cdac83031 ceph: remove reco... |
770 |
u32 gen; |
a8599bd82 ceph: capability ... |
771 |
|
d8fb02abd ceph: create a ne... |
772 |
spin_lock(&cap->session->s_gen_ttl_lock); |
a8599bd82 ceph: capability ... |
773 774 |
gen = cap->session->s_cap_gen; ttl = cap->session->s_cap_ttl; |
d8fb02abd ceph: create a ne... |
775 |
spin_unlock(&cap->session->s_gen_ttl_lock); |
a8599bd82 ceph: capability ... |
776 |
|
685f9a5d1 ceph: do not conf... |
777 |
if (cap->cap_gen < gen || time_after_eq(jiffies, ttl)) { |
a8599bd82 ceph: capability ... |
778 779 780 |
dout("__cap_is_valid %p cap %p issued %s " "but STALE (gen %u vs %u) ", &cap->ci->vfs_inode, |
685f9a5d1 ceph: do not conf... |
781 |
cap, ceph_cap_string(cap->issued), cap->cap_gen, gen); |
a8599bd82 ceph: capability ... |
782 783 784 785 786 787 788 789 790 791 792 793 794 |
return 0; } return 1; } /* * Return set of valid cap bits issued to us. Note that caps time * out, and may be invalidated in bulk if the client session times out * and session->s_cap_gen is bumped. */ int __ceph_caps_issued(struct ceph_inode_info *ci, int *implemented) { |
d9df27835 ceph: pre-allocat... |
795 |
int have = ci->i_snap_caps; |
a8599bd82 ceph: capability ... |
796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 |
struct ceph_cap *cap; struct rb_node *p; if (implemented) *implemented = 0; for (p = rb_first(&ci->i_caps); p; p = rb_next(p)) { cap = rb_entry(p, struct ceph_cap, ci_node); if (!__cap_is_valid(cap)) continue; dout("__ceph_caps_issued %p cap %p issued %s ", &ci->vfs_inode, cap, ceph_cap_string(cap->issued)); have |= cap->issued; if (implemented) *implemented |= cap->implemented; } |
b1530f570 ceph: fix cap rev... |
812 813 814 815 816 817 818 819 820 |
/* * exclude caps issued by non-auth MDS, but are been revoking * by the auth MDS. The non-auth MDS should be revoking/exporting * these caps, but the message is delayed. */ if (ci->i_auth_cap) { cap = ci->i_auth_cap; have &= ~cap->implemented | cap->issued; } |
a8599bd82 ceph: capability ... |
821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 |
return have; } /* * Get cap bits issued by caps other than @ocap */ int __ceph_caps_issued_other(struct ceph_inode_info *ci, struct ceph_cap *ocap) { int have = ci->i_snap_caps; struct ceph_cap *cap; struct rb_node *p; for (p = rb_first(&ci->i_caps); p; p = rb_next(p)) { cap = rb_entry(p, struct ceph_cap, ci_node); if (cap == ocap) continue; if (!__cap_is_valid(cap)) continue; have |= cap->issued; } return have; } /* * Move a cap to the end of the LRU (oldest caps at list head, newest * at list tail). */ static void __touch_cap(struct ceph_cap *cap) { struct ceph_mds_session *s = cap->session; |
a8599bd82 ceph: capability ... |
851 |
spin_lock(&s->s_cap_lock); |
d37b1d994 ceph: adjust 36 c... |
852 |
if (!s->s_cap_iterator) { |
5dacf0912 ceph: do not touc... |
853 854 855 856 857 858 859 860 861 |
dout("__touch_cap %p cap %p mds%d ", &cap->ci->vfs_inode, cap, s->s_mds); list_move_tail(&cap->session_caps, &s->s_caps); } else { dout("__touch_cap %p cap %p mds%d NOP, iterating over caps ", &cap->ci->vfs_inode, cap, s->s_mds); } |
a8599bd82 ceph: capability ... |
862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 |
spin_unlock(&s->s_cap_lock); } /* * Check if we hold the given mask. If so, move the cap(s) to the * front of their respective LRUs. (This is the preferred way for * callers to check for caps they want.) */ int __ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int touch) { struct ceph_cap *cap; struct rb_node *p; int have = ci->i_snap_caps; if ((have & mask) == mask) { |
ebce3eb2f ceph: fix inode n... |
877 878 879 |
dout("__ceph_caps_issued_mask ino 0x%llx snap issued %s" " (mask %s) ", ceph_ino(&ci->vfs_inode), |
a8599bd82 ceph: capability ... |
880 881 882 883 884 885 886 887 888 889 |
ceph_cap_string(have), ceph_cap_string(mask)); return 1; } for (p = rb_first(&ci->i_caps); p; p = rb_next(p)) { cap = rb_entry(p, struct ceph_cap, ci_node); if (!__cap_is_valid(cap)) continue; if ((cap->issued & mask) == mask) { |
ebce3eb2f ceph: fix inode n... |
890 891 892 |
dout("__ceph_caps_issued_mask ino 0x%llx cap %p issued %s" " (mask %s) ", ceph_ino(&ci->vfs_inode), cap, |
a8599bd82 ceph: capability ... |
893 894 895 896 897 898 899 900 901 902 |
ceph_cap_string(cap->issued), ceph_cap_string(mask)); if (touch) __touch_cap(cap); return 1; } /* does a combination of caps satisfy mask? */ have |= cap->issued; if ((have & mask) == mask) { |
ebce3eb2f ceph: fix inode n... |
903 904 905 |
dout("__ceph_caps_issued_mask ino 0x%llx combo issued %s" " (mask %s) ", ceph_ino(&ci->vfs_inode), |
a8599bd82 ceph: capability ... |
906 907 908 909 |
ceph_cap_string(cap->issued), ceph_cap_string(mask)); if (touch) { struct rb_node *q; |
25985edce Fix common misspe... |
910 |
/* touch this + preceding caps */ |
a8599bd82 ceph: capability ... |
911 912 913 914 915 916 917 |
__touch_cap(cap); for (q = rb_first(&ci->i_caps); q != p; q = rb_next(q)) { cap = rb_entry(q, struct ceph_cap, ci_node); if (!__cap_is_valid(cap)) continue; |
9f8b72b3a ceph: only touch ... |
918 919 |
if (cap->issued & mask) __touch_cap(cap); |
a8599bd82 ceph: capability ... |
920 921 922 923 924 925 926 927 |
} } return 1; } } return 0; } |
1af16d547 ceph: add caps pe... |
928 929 930 931 932 933 934 935 936 937 938 939 940 |
int __ceph_caps_issued_mask_metric(struct ceph_inode_info *ci, int mask, int touch) { struct ceph_fs_client *fsc = ceph_sb_to_client(ci->vfs_inode.i_sb); int r; r = __ceph_caps_issued_mask(ci, mask, touch); if (r) ceph_update_cap_hit(&fsc->mdsc->metric); else ceph_update_cap_mis(&fsc->mdsc->metric); return r; } |
a8599bd82 ceph: capability ... |
941 942 943 |
/* * Return true if mask caps are currently being revoked by an MDS. */ |
6ee6b9537 ceph: fix race be... |
944 945 |
int __ceph_caps_revoking_other(struct ceph_inode_info *ci, struct ceph_cap *ocap, int mask) |
a8599bd82 ceph: capability ... |
946 |
{ |
a8599bd82 ceph: capability ... |
947 948 |
struct ceph_cap *cap; struct rb_node *p; |
a8599bd82 ceph: capability ... |
949 |
|
a8599bd82 ceph: capability ... |
950 951 |
for (p = rb_first(&ci->i_caps); p; p = rb_next(p)) { cap = rb_entry(p, struct ceph_cap, ci_node); |
9563f88c1 ceph: fix cache r... |
952 |
if (cap != ocap && |
6ee6b9537 ceph: fix race be... |
953 954 |
(cap->implemented & ~cap->issued & mask)) return 1; |
a8599bd82 ceph: capability ... |
955 |
} |
6ee6b9537 ceph: fix race be... |
956 957 958 959 960 961 962 963 964 965 |
return 0; } int ceph_caps_revoking(struct ceph_inode_info *ci, int mask) { struct inode *inode = &ci->vfs_inode; int ret; spin_lock(&ci->i_ceph_lock); ret = __ceph_caps_revoking_other(ci, NULL, mask); |
be655596b ceph: use i_ceph_... |
966 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
967 968 969 970 971 972 973 974 975 976 977 978 979 |
dout("ceph_caps_revoking %p %s = %d ", inode, ceph_cap_string(mask), ret); return ret; } int __ceph_caps_used(struct ceph_inode_info *ci) { int used = 0; if (ci->i_pin_ref) used |= CEPH_CAP_PIN; if (ci->i_rd_ref) used |= CEPH_CAP_FILE_RD; |
fdd4e1583 ceph: rework dcac... |
980 |
if (ci->i_rdcache_ref || |
525d15e8e ceph: check inode... |
981 |
(S_ISREG(ci->vfs_inode.i_mode) && |
fdd4e1583 ceph: rework dcac... |
982 |
ci->vfs_inode.i_data.nrpages)) |
a8599bd82 ceph: capability ... |
983 984 985 |
used |= CEPH_CAP_FILE_CACHE; if (ci->i_wr_ref) used |= CEPH_CAP_FILE_WR; |
d3d0720d4 ceph: do not use ... |
986 |
if (ci->i_wb_ref || ci->i_wrbuffer_ref) |
a8599bd82 ceph: capability ... |
987 |
used |= CEPH_CAP_FILE_BUFFER; |
f85122afe ceph: add refcoun... |
988 989 |
if (ci->i_fx_ref) used |= CEPH_CAP_FILE_EXCL; |
a8599bd82 ceph: capability ... |
990 991 |
return used; } |
719a2514e ceph: consider in... |
992 |
#define FMODE_WAIT_BIAS 1000 |
a8599bd82 ceph: capability ... |
993 994 995 996 997 |
/* * wanted, by virtue of open file modes */ int __ceph_caps_file_wanted(struct ceph_inode_info *ci) { |
719a2514e ceph: consider in... |
998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 |
const int PIN_SHIFT = ffs(CEPH_FILE_MODE_PIN); const int RD_SHIFT = ffs(CEPH_FILE_MODE_RD); const int WR_SHIFT = ffs(CEPH_FILE_MODE_WR); const int LAZY_SHIFT = ffs(CEPH_FILE_MODE_LAZY); struct ceph_mount_options *opt = ceph_inode_to_client(&ci->vfs_inode)->mount_options; unsigned long used_cutoff = jiffies - opt->caps_wanted_delay_max * HZ; unsigned long idle_cutoff = jiffies - opt->caps_wanted_delay_min * HZ; if (S_ISDIR(ci->vfs_inode.i_mode)) { int want = 0; /* use used_cutoff here, to keep dir's wanted caps longer */ if (ci->i_nr_by_mode[RD_SHIFT] > 0 || time_after(ci->i_last_rd, used_cutoff)) want |= CEPH_CAP_ANY_SHARED; if (ci->i_nr_by_mode[WR_SHIFT] > 0 || time_after(ci->i_last_wr, used_cutoff)) { want |= CEPH_CAP_ANY_SHARED | CEPH_CAP_FILE_EXCL; if (opt->flags & CEPH_MOUNT_OPT_ASYNC_DIROPS) want |= CEPH_CAP_ANY_DIR_OPS; } if (want || ci->i_nr_by_mode[PIN_SHIFT] > 0) want |= CEPH_CAP_PIN; return want; } else { int bits = 0; if (ci->i_nr_by_mode[RD_SHIFT] > 0) { if (ci->i_nr_by_mode[RD_SHIFT] >= FMODE_WAIT_BIAS || time_after(ci->i_last_rd, used_cutoff)) bits |= 1 << RD_SHIFT; } else if (time_after(ci->i_last_rd, idle_cutoff)) { bits |= 1 << RD_SHIFT; } if (ci->i_nr_by_mode[WR_SHIFT] > 0) { if (ci->i_nr_by_mode[WR_SHIFT] >= FMODE_WAIT_BIAS || time_after(ci->i_last_wr, used_cutoff)) bits |= 1 << WR_SHIFT; } else if (time_after(ci->i_last_wr, idle_cutoff)) { bits |= 1 << WR_SHIFT; } /* check lazyio only when read/write is wanted */ if ((bits & (CEPH_FILE_MODE_RDWR << 1)) && ci->i_nr_by_mode[LAZY_SHIFT] > 0) bits |= 1 << LAZY_SHIFT; return bits ? ceph_caps_for_mode(bits >> 1) : 0; |
774a6a118 ceph: reduce i_nr... |
1051 |
} |
a8599bd82 ceph: capability ... |
1052 1053 1054 |
} /* |
525d15e8e ceph: check inode... |
1055 1056 1057 1058 1059 |
* wanted, by virtue of open file modes AND cap refs (buffered/cached data) */ int __ceph_caps_wanted(struct ceph_inode_info *ci) { int w = __ceph_caps_file_wanted(ci) | __ceph_caps_used(ci); |
a25949b99 ceph: cap trackin... |
1060 1061 1062 1063 1064 |
if (S_ISDIR(ci->vfs_inode.i_mode)) { /* we want EXCL if holding caps of dir ops */ if (w & CEPH_CAP_ANY_DIR_OPS) w |= CEPH_CAP_FILE_EXCL; } else { |
525d15e8e ceph: check inode... |
1065 1066 1067 1068 1069 1070 1071 1072 |
/* we want EXCL if dirty data */ if (w & CEPH_CAP_FILE_BUFFER) w |= CEPH_CAP_FILE_EXCL; } return w; } /* |
a8599bd82 ceph: capability ... |
1073 1074 |
* Return caps we have registered with the MDS(s) as 'wanted'. */ |
c1944fedd ceph: avoid calli... |
1075 |
int __ceph_caps_mds_wanted(struct ceph_inode_info *ci, bool check) |
a8599bd82 ceph: capability ... |
1076 1077 1078 1079 1080 1081 1082 |
{ struct ceph_cap *cap; struct rb_node *p; int mds_wanted = 0; for (p = rb_first(&ci->i_caps); p; p = rb_next(p)) { cap = rb_entry(p, struct ceph_cap, ci_node); |
c1944fedd ceph: avoid calli... |
1083 |
if (check && !__cap_is_valid(cap)) |
a8599bd82 ceph: capability ... |
1084 |
continue; |
a25506045 ceph: make sure w... |
1085 1086 1087 1088 |
if (cap == ci->i_auth_cap) mds_wanted |= cap->mds_wanted; else mds_wanted |= (cap->mds_wanted & ~CEPH_CAP_ANY_FILE_WR); |
a8599bd82 ceph: capability ... |
1089 1090 1091 |
} return mds_wanted; } |
9215aeea6 ceph: check inode... |
1092 1093 1094 1095 1096 1097 |
int ceph_is_any_caps(struct inode *inode) { struct ceph_inode_info *ci = ceph_inode(inode); int ret; spin_lock(&ci->i_ceph_lock); |
bd84fbcb3 ceph: switch to g... |
1098 |
ret = __ceph_is_any_real_caps(ci); |
9215aeea6 ceph: check inode... |
1099 1100 1101 1102 |
spin_unlock(&ci->i_ceph_lock); return ret; } |
db40cc170 ceph: keep i_snap... |
1103 1104 1105 1106 1107 1108 1109 |
static void drop_inode_snap_realm(struct ceph_inode_info *ci) { struct ceph_snap_realm *realm = ci->i_snap_realm; spin_lock(&realm->inodes_with_caps_lock); list_del_init(&ci->i_snap_realm_item); ci->i_snap_realm_counter++; ci->i_snap_realm = NULL; |
d95e674c0 ceph: clear inode... |
1110 1111 |
if (realm->ino == ci->i_vino.ino) realm->inode = NULL; |
db40cc170 ceph: keep i_snap... |
1112 1113 1114 1115 |
spin_unlock(&realm->inodes_with_caps_lock); ceph_put_snap_realm(ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc, realm); } |
a8599bd82 ceph: capability ... |
1116 |
/* |
f818a7367 ceph: fix cap rem... |
1117 1118 |
* Remove a cap. Take steps to deal with a racing iterate_session_caps. * |
be655596b ceph: use i_ceph_... |
1119 |
* caller should hold i_ceph_lock. |
a6369741c ceph: fix comment... |
1120 |
* caller will not hold session s_mutex if called from destroy_inode. |
a8599bd82 ceph: capability ... |
1121 |
*/ |
a096b09ae ceph: queue cap r... |
1122 |
void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release) |
a8599bd82 ceph: capability ... |
1123 1124 1125 |
{ struct ceph_mds_session *session = cap->session; struct ceph_inode_info *ci = cap->ci; |
db1c6b8a2 ceph: fix race in... |
1126 |
struct ceph_mds_client *mdsc; |
f818a7367 ceph: fix cap rem... |
1127 |
int removed = 0; |
a8599bd82 ceph: capability ... |
1128 |
|
db1c6b8a2 ceph: fix race in... |
1129 1130 1131 1132 1133 1134 |
/* 'ci' being NULL means the remove have already occurred */ if (!ci) { dout("%s: cap inode is NULL ", __func__); return; } |
a8599bd82 ceph: capability ... |
1135 1136 |
dout("__ceph_remove_cap %p from %p ", cap, &ci->vfs_inode); |
db1c6b8a2 ceph: fix race in... |
1137 |
mdsc = ceph_inode_to_client(&ci->vfs_inode)->mdsc; |
ea60ed6fc ceph: fix use-aft... |
1138 1139 |
/* remove from inode's cap rbtree, and clear auth cap */ rb_erase(&cap->ci_node, &ci->i_caps); |
1cf03a68e ceph: convert mds... |
1140 1141 |
if (ci->i_auth_cap == cap) { WARN_ON_ONCE(!list_empty(&ci->i_dirty_item)); |
ea60ed6fc ceph: fix use-aft... |
1142 |
ci->i_auth_cap = NULL; |
1cf03a68e ceph: convert mds... |
1143 |
} |
ea60ed6fc ceph: fix use-aft... |
1144 |
|
7c1332b8c ceph: fix iterate... |
1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 |
/* remove from session list */ spin_lock(&session->s_cap_lock); if (session->s_cap_iterator == cap) { /* not yet, we are iterating over this very cap */ dout("__ceph_remove_cap delaying %p removal from session %p ", cap, cap->session); } else { list_del_init(&cap->session_caps); session->s_nr_caps--; |
4f1d756de ceph: add global ... |
1155 |
atomic64_dec(&mdsc->metric.total_caps); |
7c1332b8c ceph: fix iterate... |
1156 |
cap->session = NULL; |
f818a7367 ceph: fix cap rem... |
1157 |
removed = 1; |
7c1332b8c ceph: fix iterate... |
1158 |
} |
f818a7367 ceph: fix cap rem... |
1159 1160 |
/* protect backpointer with s_cap_lock: see iterate_session_caps */ cap->ci = NULL; |
745a8e3bc ceph: don't pre-a... |
1161 1162 1163 1164 1165 1166 1167 1168 1169 |
/* * s_cap_reconnect is protected by s_cap_lock. no one changes * s_cap_gen while session is in the reconnect state. */ if (queue_release && (!session->s_cap_reconnect || cap->cap_gen == session->s_cap_gen)) { cap->queue_release = 1; if (removed) { |
e3ec8d689 ceph: send cap re... |
1170 |
__ceph_queue_cap_release(session, cap); |
745a8e3bc ceph: don't pre-a... |
1171 1172 1173 1174 1175 1176 |
removed = 0; } } else { cap->queue_release = 0; } cap->cap_ino = ci->i_vino.ino; |
7c1332b8c ceph: fix iterate... |
1177 |
spin_unlock(&session->s_cap_lock); |
f818a7367 ceph: fix cap rem... |
1178 |
if (removed) |
37151668b ceph: do caps acc... |
1179 |
ceph_put_cap(mdsc, cap); |
a8599bd82 ceph: capability ... |
1180 |
|
bd84fbcb3 ceph: switch to g... |
1181 1182 1183 1184 1185 1186 1187 |
if (!__ceph_is_any_real_caps(ci)) { /* when reconnect denied, we remove session caps forcibly, * i_wr_ref can be non-zero. If there are ongoing write, * keep i_snap_realm. */ if (ci->i_wr_ref == 0 && ci->i_snap_realm) drop_inode_snap_realm(ci); |
db40cc170 ceph: keep i_snap... |
1188 |
|
a8599bd82 ceph: capability ... |
1189 |
__cap_delay_cancel(mdsc, ci); |
bd84fbcb3 ceph: switch to g... |
1190 |
} |
a8599bd82 ceph: capability ... |
1191 |
} |
0ff8bfb39 ceph: define new ... |
1192 1193 1194 1195 1196 |
struct cap_msg_args { struct ceph_mds_session *session; u64 ino, cid, follows; u64 flush_tid, oldest_flush_tid, size, max_size; u64 xattr_version; |
176c77c9c ceph: handle chan... |
1197 |
u64 change_attr; |
0ff8bfb39 ceph: define new ... |
1198 |
struct ceph_buffer *xattr_buf; |
0a454bdd5 ceph: reorganize ... |
1199 |
struct ceph_buffer *old_xattr_buf; |
ec62b894d ceph: handle btim... |
1200 |
struct timespec64 atime, mtime, ctime, btime; |
0ff8bfb39 ceph: define new ... |
1201 1202 |
int op, caps, wanted, dirty; u32 seq, issue_seq, mseq, time_warp_seq; |
1e4ef0c63 ceph: add flags p... |
1203 |
u32 flags; |
0ff8bfb39 ceph: define new ... |
1204 1205 1206 1207 |
kuid_t uid; kgid_t gid; umode_t mode; bool inline_data; |
0a454bdd5 ceph: reorganize ... |
1208 |
bool wake; |
0ff8bfb39 ceph: define new ... |
1209 |
}; |
a8599bd82 ceph: capability ... |
1210 |
/* |
16d68903f ceph: break up se... |
1211 1212 |
* cap struct size + flock buffer size + inline version + inline data size + * osd_epoch_barrier + oldest_flush_tid |
a8599bd82 ceph: capability ... |
1213 |
*/ |
16d68903f ceph: break up se... |
1214 1215 1216 1217 1218 |
#define CAP_MSG_SIZE (sizeof(struct ceph_mds_caps) + \ 4 + 8 + 4 + 4 + 8 + 4 + 4 + 4 + 8 + 8 + 4) /* Marshal up the cap msg to the MDS */ static void encode_cap_msg(struct ceph_msg *msg, struct cap_msg_args *arg) |
a8599bd82 ceph: capability ... |
1219 1220 |
{ struct ceph_mds_caps *fc; |
e20d258d7 ceph: flush inlin... |
1221 |
void *p; |
92475f05b ceph: handle epoc... |
1222 |
struct ceph_osd_client *osdc = &arg->session->s_mdsc->fsc->client->osdc; |
a8599bd82 ceph: capability ... |
1223 |
|
16d68903f ceph: break up se... |
1224 1225 1226 1227 1228 1229 1230 |
dout("%s %s %llx %llx caps %s wanted %s dirty %s seq %u/%u tid %llu/%llu mseq %u follows %lld size %llu/%llu xattr_ver %llu xattr_len %d ", __func__, ceph_cap_op_name(arg->op), arg->cid, arg->ino, ceph_cap_string(arg->caps), ceph_cap_string(arg->wanted), ceph_cap_string(arg->dirty), arg->seq, arg->issue_seq, arg->flush_tid, arg->oldest_flush_tid, arg->mseq, arg->follows, arg->size, arg->max_size, arg->xattr_version, |
0ff8bfb39 ceph: define new ... |
1231 |
arg->xattr_buf ? (int)arg->xattr_buf->vec.iov_len : 0); |
a8599bd82 ceph: capability ... |
1232 |
|
43b296733 ceph: update cap ... |
1233 |
msg->hdr.version = cpu_to_le16(10); |
0ff8bfb39 ceph: define new ... |
1234 |
msg->hdr.tid = cpu_to_le64(arg->flush_tid); |
a8599bd82 ceph: capability ... |
1235 |
|
6df058c02 ceph: include tra... |
1236 |
fc = msg->front.iov_base; |
a8599bd82 ceph: capability ... |
1237 |
memset(fc, 0, sizeof(*fc)); |
0ff8bfb39 ceph: define new ... |
1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 |
fc->cap_id = cpu_to_le64(arg->cid); fc->op = cpu_to_le32(arg->op); fc->seq = cpu_to_le32(arg->seq); fc->issue_seq = cpu_to_le32(arg->issue_seq); fc->migrate_seq = cpu_to_le32(arg->mseq); fc->caps = cpu_to_le32(arg->caps); fc->wanted = cpu_to_le32(arg->wanted); fc->dirty = cpu_to_le32(arg->dirty); fc->ino = cpu_to_le64(arg->ino); fc->snap_follows = cpu_to_le64(arg->follows); fc->size = cpu_to_le64(arg->size); fc->max_size = cpu_to_le64(arg->max_size); |
9bbeab41c ceph: use timespe... |
1251 1252 1253 |
ceph_encode_timespec64(&fc->mtime, &arg->mtime); ceph_encode_timespec64(&fc->atime, &arg->atime); ceph_encode_timespec64(&fc->ctime, &arg->ctime); |
0ff8bfb39 ceph: define new ... |
1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 |
fc->time_warp_seq = cpu_to_le32(arg->time_warp_seq); fc->uid = cpu_to_le32(from_kuid(&init_user_ns, arg->uid)); fc->gid = cpu_to_le32(from_kgid(&init_user_ns, arg->gid)); fc->mode = cpu_to_le32(arg->mode); fc->xattr_version = cpu_to_le64(arg->xattr_version); if (arg->xattr_buf) { msg->middle = ceph_buffer_get(arg->xattr_buf); fc->xattr_len = cpu_to_le32(arg->xattr_buf->vec.iov_len); msg->hdr.middle_len = cpu_to_le32(arg->xattr_buf->vec.iov_len); |
9670079f5 ceph: move xattr ... |
1265 |
} |
e20d258d7 ceph: flush inlin... |
1266 |
p = fc + 1; |
43b296733 ceph: update cap ... |
1267 |
/* flock buffer size (version 2) */ |
e20d258d7 ceph: flush inlin... |
1268 |
ceph_encode_32(&p, 0); |
43b296733 ceph: update cap ... |
1269 |
/* inline version (version 4) */ |
0ff8bfb39 ceph: define new ... |
1270 |
ceph_encode_64(&p, arg->inline_data ? 0 : CEPH_INLINE_NONE); |
e20d258d7 ceph: flush inlin... |
1271 1272 |
/* inline data size */ ceph_encode_32(&p, 0); |
92475f05b ceph: handle epoc... |
1273 1274 1275 1276 1277 1278 |
/* * osd_epoch_barrier (version 5) * The epoch_barrier is protected osdc->lock, so READ_ONCE here in * case it was recently changed */ ceph_encode_32(&p, READ_ONCE(osdc->epoch_barrier)); |
43b296733 ceph: update cap ... |
1279 |
/* oldest_flush_tid (version 6) */ |
0ff8bfb39 ceph: define new ... |
1280 |
ceph_encode_64(&p, arg->oldest_flush_tid); |
e20d258d7 ceph: flush inlin... |
1281 |
|
43b296733 ceph: update cap ... |
1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 |
/* * caller_uid/caller_gid (version 7) * * Currently, we don't properly track which caller dirtied the caps * last, and force a flush of them when there is a conflict. For now, * just set this to 0:0, to emulate how the MDS has worked up to now. */ ceph_encode_32(&p, 0); ceph_encode_32(&p, 0); /* pool namespace (version 8) (mds always ignores this) */ ceph_encode_32(&p, 0); |
176c77c9c ceph: handle chan... |
1294 |
/* btime and change_attr (version 9) */ |
ec62b894d ceph: handle btim... |
1295 |
ceph_encode_timespec64(p, &arg->btime); |
43b296733 ceph: update cap ... |
1296 |
p += sizeof(struct ceph_timespec); |
176c77c9c ceph: handle chan... |
1297 |
ceph_encode_64(&p, arg->change_attr); |
43b296733 ceph: update cap ... |
1298 1299 |
/* Advisory flags (version 10) */ |
1e4ef0c63 ceph: add flags p... |
1300 |
ceph_encode_32(&p, arg->flags); |
a8599bd82 ceph: capability ... |
1301 1302 1303 |
} /* |
d6e478197 ceph: hold i_ceph... |
1304 |
* Queue cap releases when an inode is dropped from our cache. |
a8599bd82 ceph: capability ... |
1305 |
*/ |
d6e478197 ceph: hold i_ceph... |
1306 |
void __ceph_remove_caps(struct ceph_inode_info *ci) |
a8599bd82 ceph: capability ... |
1307 |
{ |
a8599bd82 ceph: capability ... |
1308 |
struct rb_node *p; |
d6e478197 ceph: hold i_ceph... |
1309 1310 1311 |
/* lock i_ceph_lock, because ceph_d_revalidate(..., LOOKUP_RCU) * may call __ceph_caps_issued_mask() on a freeing inode. */ spin_lock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
1312 1313 1314 |
p = rb_first(&ci->i_caps); while (p) { struct ceph_cap *cap = rb_entry(p, struct ceph_cap, ci_node); |
a8599bd82 ceph: capability ... |
1315 |
p = rb_next(p); |
a096b09ae ceph: queue cap r... |
1316 |
__ceph_remove_cap(cap, true); |
a8599bd82 ceph: capability ... |
1317 |
} |
d6e478197 ceph: hold i_ceph... |
1318 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
1319 1320 1321 |
} /* |
0a454bdd5 ceph: reorganize ... |
1322 1323 1324 |
* Prepare to send a cap message to an MDS. Update the cap state, and populate * the arg struct with the parameters that will need to be sent. This should * be done under the i_ceph_lock to guard against changes to cap state. |
a8599bd82 ceph: capability ... |
1325 1326 1327 |
* * Make note of max_size reported/requested from mds, revoked caps * that have now been implemented. |
a8599bd82 ceph: capability ... |
1328 |
*/ |
0a454bdd5 ceph: reorganize ... |
1329 1330 1331 |
static void __prep_cap(struct cap_msg_args *arg, struct ceph_cap *cap, int op, int flags, int used, int want, int retain, int flushing, u64 flush_tid, u64 oldest_flush_tid) |
a8599bd82 ceph: capability ... |
1332 1333 1334 |
{ struct ceph_inode_info *ci = cap->ci; struct inode *inode = &ci->vfs_inode; |
bb0581f01 ceph: remove unus... |
1335 |
int held, revoking; |
a8599bd82 ceph: capability ... |
1336 |
|
0a454bdd5 ceph: reorganize ... |
1337 |
lockdep_assert_held(&ci->i_ceph_lock); |
891f3f5a6 ceph: add infrast... |
1338 |
|
68c283236 ceph: do not reta... |
1339 1340 1341 |
held = cap->issued | cap->implemented; revoking = cap->implemented & ~cap->issued; retain &= ~revoking; |
68c283236 ceph: do not reta... |
1342 |
|
0a454bdd5 ceph: reorganize ... |
1343 1344 1345 |
dout("%s %p cap %p session %p %s -> %s (revoking %s) ", __func__, inode, cap, cap->session, |
a8599bd82 ceph: capability ... |
1346 1347 1348 |
ceph_cap_string(held), ceph_cap_string(held & retain), ceph_cap_string(revoking)); BUG_ON((retain & CEPH_CAP_PIN) == 0); |
a0d93e327 ceph: remove dela... |
1349 |
ci->i_ceph_flags &= ~CEPH_I_FLUSH; |
a8599bd82 ceph: capability ... |
1350 1351 |
cap->issued &= retain; /* drop bits we don't want */ |
0a454bdd5 ceph: reorganize ... |
1352 1353 1354 1355 1356 1357 |
/* * Wake up any waiters on wanted -> needed transition. This is due to * the weird transition from buffered to sync IO... we need to flush * dirty pages _before_ allowing sync writes to avoid reordering. */ arg->wake = cap->implemented & ~cap->issued; |
a8599bd82 ceph: capability ... |
1358 1359 |
cap->implemented &= cap->issued | used; cap->mds_wanted = want; |
0a454bdd5 ceph: reorganize ... |
1360 1361 1362 1363 1364 1365 |
arg->session = cap->session; arg->ino = ceph_vino(inode).ino; arg->cid = cap->cap_id; arg->follows = flushing ? ci->i_head_snapc->seq : 0; arg->flush_tid = flush_tid; arg->oldest_flush_tid = oldest_flush_tid; |
0ff8bfb39 ceph: define new ... |
1366 |
|
0a454bdd5 ceph: reorganize ... |
1367 1368 1369 |
arg->size = inode->i_size; ci->i_reported_size = arg->size; arg->max_size = ci->i_wanted_max_size; |
6f05b30ea ceph: reset i_req... |
1370 1371 1372 1373 1374 1375 |
if (cap == ci->i_auth_cap) { if (want & CEPH_CAP_ANY_FILE_WR) ci->i_requested_max_size = arg->max_size; else ci->i_requested_max_size = 0; } |
a8599bd82 ceph: capability ... |
1376 |
|
082afec92 ceph: fix xattr c... |
1377 |
if (flushing & CEPH_CAP_XATTR_EXCL) { |
0a454bdd5 ceph: reorganize ... |
1378 1379 1380 |
arg->old_xattr_buf = __ceph_build_xattrs_blob(ci); arg->xattr_version = ci->i_xattrs.version; arg->xattr_buf = ci->i_xattrs.blob; |
0ff8bfb39 ceph: define new ... |
1381 |
} else { |
0a454bdd5 ceph: reorganize ... |
1382 1383 |
arg->xattr_buf = NULL; arg->old_xattr_buf = NULL; |
a8599bd82 ceph: capability ... |
1384 |
} |
0a454bdd5 ceph: reorganize ... |
1385 1386 1387 1388 1389 |
arg->mtime = inode->i_mtime; arg->atime = inode->i_atime; arg->ctime = inode->i_ctime; arg->btime = ci->i_btime; arg->change_attr = inode_peek_iversion_raw(inode); |
0ff8bfb39 ceph: define new ... |
1390 |
|
0a454bdd5 ceph: reorganize ... |
1391 1392 1393 1394 |
arg->op = op; arg->caps = cap->implemented; arg->wanted = want; arg->dirty = flushing; |
0ff8bfb39 ceph: define new ... |
1395 |
|
0a454bdd5 ceph: reorganize ... |
1396 1397 1398 1399 |
arg->seq = cap->seq; arg->issue_seq = cap->issue_seq; arg->mseq = cap->mseq; arg->time_warp_seq = ci->i_time_warp_seq; |
0ff8bfb39 ceph: define new ... |
1400 |
|
0a454bdd5 ceph: reorganize ... |
1401 1402 1403 |
arg->uid = inode->i_uid; arg->gid = inode->i_gid; arg->mode = inode->i_mode; |
0ff8bfb39 ceph: define new ... |
1404 |
|
0a454bdd5 ceph: reorganize ... |
1405 |
arg->inline_data = ci->i_inline_version != CEPH_INLINE_NONE; |
49ada6e8d ceph: more precis... |
1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 |
if (!(flags & CEPH_CLIENT_CAPS_PENDING_CAPSNAP) && !list_empty(&ci->i_cap_snaps)) { struct ceph_cap_snap *capsnap; list_for_each_entry_reverse(capsnap, &ci->i_cap_snaps, ci_item) { if (capsnap->cap_flush.tid) break; if (capsnap->need_flush) { flags |= CEPH_CLIENT_CAPS_PENDING_CAPSNAP; break; } } } |
0a454bdd5 ceph: reorganize ... |
1418 1419 |
arg->flags = flags; } |
a8599bd82 ceph: capability ... |
1420 |
|
0a454bdd5 ceph: reorganize ... |
1421 1422 1423 1424 1425 |
/* * Send a cap msg on the given inode. * * Caller should hold snap_rwsem (read), s_mutex. */ |
523119808 ceph: drop separa... |
1426 |
static void __send_cap(struct cap_msg_args *arg, struct ceph_inode_info *ci) |
0a454bdd5 ceph: reorganize ... |
1427 |
{ |
16d68903f ceph: break up se... |
1428 |
struct ceph_msg *msg; |
0a454bdd5 ceph: reorganize ... |
1429 |
struct inode *inode = &ci->vfs_inode; |
12fe3dda7 ceph: fix buffer ... |
1430 |
|
16d68903f ceph: break up se... |
1431 1432 1433 1434 |
msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPS, CAP_MSG_SIZE, GFP_NOFS, false); if (!msg) { pr_err("error allocating cap msg: ino (%llx.%llx) flushing %s tid %llu, requeuing cap. ", |
0a454bdd5 ceph: reorganize ... |
1435 1436 |
ceph_vinop(inode), ceph_cap_string(arg->dirty), arg->flush_tid); |
a0d93e327 ceph: remove dela... |
1437 |
spin_lock(&ci->i_ceph_lock); |
523119808 ceph: drop separa... |
1438 |
__cap_delay_requeue(arg->session->s_mdsc, ci); |
a0d93e327 ceph: remove dela... |
1439 |
spin_unlock(&ci->i_ceph_lock); |
16d68903f ceph: break up se... |
1440 |
return; |
a8599bd82 ceph: capability ... |
1441 |
} |
16d68903f ceph: break up se... |
1442 1443 |
encode_cap_msg(msg, arg); ceph_con_send(&arg->session->s_con, msg); |
0a454bdd5 ceph: reorganize ... |
1444 |
ceph_buffer_put(arg->old_xattr_buf); |
0a454bdd5 ceph: reorganize ... |
1445 1446 |
if (arg->wake) wake_up_all(&ci->i_cap_wq); |
a8599bd82 ceph: capability ... |
1447 |
} |
0e2943878 ceph: unify cap f... |
1448 1449 1450 1451 1452 |
static inline int __send_flush_snap(struct inode *inode, struct ceph_mds_session *session, struct ceph_cap_snap *capsnap, u32 mseq, u64 oldest_flush_tid) { |
0ff8bfb39 ceph: define new ... |
1453 |
struct cap_msg_args arg; |
16d68903f ceph: break up se... |
1454 1455 1456 1457 1458 |
struct ceph_msg *msg; msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPS, CAP_MSG_SIZE, GFP_NOFS, false); if (!msg) return -ENOMEM; |
0ff8bfb39 ceph: define new ... |
1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 |
arg.session = session; arg.ino = ceph_vino(inode).ino; arg.cid = 0; arg.follows = capsnap->follows; arg.flush_tid = capsnap->cap_flush.tid; arg.oldest_flush_tid = oldest_flush_tid; arg.size = capsnap->size; arg.max_size = 0; arg.xattr_version = capsnap->xattr_version; arg.xattr_buf = capsnap->xattr_blob; |
0a454bdd5 ceph: reorganize ... |
1471 |
arg.old_xattr_buf = NULL; |
0ff8bfb39 ceph: define new ... |
1472 1473 1474 1475 |
arg.atime = capsnap->atime; arg.mtime = capsnap->mtime; arg.ctime = capsnap->ctime; |
ec62b894d ceph: handle btim... |
1476 |
arg.btime = capsnap->btime; |
176c77c9c ceph: handle chan... |
1477 |
arg.change_attr = capsnap->change_attr; |
0ff8bfb39 ceph: define new ... |
1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 |
arg.op = CEPH_CAP_OP_FLUSHSNAP; arg.caps = capsnap->issued; arg.wanted = 0; arg.dirty = capsnap->dirty; arg.seq = 0; arg.issue_seq = 0; arg.mseq = mseq; arg.time_warp_seq = capsnap->time_warp_seq; arg.uid = capsnap->uid; arg.gid = capsnap->gid; arg.mode = capsnap->mode; arg.inline_data = capsnap->inline_data; |
1e4ef0c63 ceph: add flags p... |
1494 |
arg.flags = 0; |
0a454bdd5 ceph: reorganize ... |
1495 |
arg.wake = false; |
0ff8bfb39 ceph: define new ... |
1496 |
|
16d68903f ceph: break up se... |
1497 1498 1499 |
encode_cap_msg(msg, &arg); ceph_con_send(&arg.session->s_con, msg); return 0; |
0e2943878 ceph: unify cap f... |
1500 |
} |
a8599bd82 ceph: capability ... |
1501 1502 1503 1504 1505 1506 1507 |
/* * When a snapshot is taken, clients accumulate dirty metadata on * inodes with capabilities in ceph_cap_snaps to describe the file * state at the time the snapshot was taken. This must be flushed * asynchronously back to the MDS once sync writes complete and dirty * data is written out. * |
be655596b ceph: use i_ceph_... |
1508 |
* Called under i_ceph_lock. Takes s_mutex as needed. |
a8599bd82 ceph: capability ... |
1509 |
*/ |
ed9b430c9 ceph: cleanup cep... |
1510 1511 |
static void __ceph_flush_snaps(struct ceph_inode_info *ci, struct ceph_mds_session *session) |
be655596b ceph: use i_ceph_... |
1512 1513 |
__releases(ci->i_ceph_lock) __acquires(ci->i_ceph_lock) |
a8599bd82 ceph: capability ... |
1514 1515 |
{ struct inode *inode = &ci->vfs_inode; |
ed9b430c9 ceph: cleanup cep... |
1516 |
struct ceph_mds_client *mdsc = session->s_mdsc; |
a8599bd82 ceph: capability ... |
1517 |
struct ceph_cap_snap *capsnap; |
ed9b430c9 ceph: cleanup cep... |
1518 1519 |
u64 oldest_flush_tid = 0; u64 first_tid = 1, last_tid = 0; |
a8599bd82 ceph: capability ... |
1520 |
|
ed9b430c9 ceph: cleanup cep... |
1521 1522 |
dout("__flush_snaps %p session %p ", inode, session); |
a8599bd82 ceph: capability ... |
1523 |
|
a8599bd82 ceph: capability ... |
1524 |
list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { |
a8599bd82 ceph: capability ... |
1525 1526 1527 1528 1529 |
/* * we need to wait for sync writes to complete and for dirty * pages to be written out. */ if (capsnap->dirty_pages || capsnap->writing) |
cfc0bf664 ceph: stop sendin... |
1530 |
break; |
a8599bd82 ceph: capability ... |
1531 |
|
860560904 ceph: avoid sendi... |
1532 1533 |
/* should be removed by ceph_try_drop_cap_snap() */ BUG_ON(!capsnap->need_flush); |
819ccbfa4 ceph: fix leaked ... |
1534 |
|
e835124c2 ceph: only send o... |
1535 |
/* only flush each capsnap once */ |
0e2943878 ceph: unify cap f... |
1536 |
if (capsnap->cap_flush.tid > 0) { |
ed9b430c9 ceph: cleanup cep... |
1537 1538 |
dout(" already flushed %p, skipping ", capsnap); |
e835124c2 ceph: only send o... |
1539 1540 |
continue; } |
553adfd94 ceph: track pendi... |
1541 |
spin_lock(&mdsc->cap_dirty_lock); |
0e2943878 ceph: unify cap f... |
1542 1543 1544 |
capsnap->cap_flush.tid = ++mdsc->last_cap_flush_tid; list_add_tail(&capsnap->cap_flush.g_list, &mdsc->cap_flush_list); |
ed9b430c9 ceph: cleanup cep... |
1545 1546 |
if (oldest_flush_tid == 0) oldest_flush_tid = __get_oldest_flush_tid(mdsc); |
0e2943878 ceph: unify cap f... |
1547 1548 1549 1550 |
if (list_empty(&ci->i_flushing_item)) { list_add_tail(&ci->i_flushing_item, &session->s_cap_flushing); } |
553adfd94 ceph: track pendi... |
1551 |
spin_unlock(&mdsc->cap_dirty_lock); |
0e2943878 ceph: unify cap f... |
1552 1553 |
list_add_tail(&capsnap->cap_flush.i_list, &ci->i_cap_flush_list); |
ed9b430c9 ceph: cleanup cep... |
1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 |
if (first_tid == 1) first_tid = capsnap->cap_flush.tid; last_tid = capsnap->cap_flush.tid; } ci->i_ceph_flags &= ~CEPH_I_FLUSH_SNAPS; while (first_tid <= last_tid) { struct ceph_cap *cap = ci->i_auth_cap; struct ceph_cap_flush *cf; int ret; if (!(cap && cap->session == session)) { dout("__flush_snaps %p auth cap %p not mds%d, " "stop ", inode, cap, session->s_mds); break; } ret = -ENOENT; list_for_each_entry(cf, &ci->i_cap_flush_list, i_list) { if (cf->tid >= first_tid) { ret = 0; break; } } if (ret < 0) break; first_tid = cf->tid + 1; capsnap = container_of(cf, struct ceph_cap_snap, cap_flush); |
805692d0e ceph: convert cep... |
1586 |
refcount_inc(&capsnap->nref); |
be655596b ceph: use i_ceph_... |
1587 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
1588 |
|
ed9b430c9 ceph: cleanup cep... |
1589 1590 1591 |
dout("__flush_snaps %p capsnap %p tid %llu %s ", inode, capsnap, cf->tid, ceph_cap_string(capsnap->dirty)); |
a8599bd82 ceph: capability ... |
1592 |
|
ed9b430c9 ceph: cleanup cep... |
1593 1594 1595 1596 1597 1598 1599 1600 |
ret = __send_flush_snap(inode, session, capsnap, cap->mseq, oldest_flush_tid); if (ret < 0) { pr_err("__flush_snaps: error sending cap flushsnap, " "ino (%llx.%llx) tid %llu follows %llu ", ceph_vinop(inode), cf->tid, capsnap->follows); } |
a8599bd82 ceph: capability ... |
1601 |
|
ed9b430c9 ceph: cleanup cep... |
1602 |
ceph_put_cap_snap(capsnap); |
be655596b ceph: use i_ceph_... |
1603 |
spin_lock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
1604 |
} |
ed9b430c9 ceph: cleanup cep... |
1605 |
} |
a8599bd82 ceph: capability ... |
1606 |
|
ed9b430c9 ceph: cleanup cep... |
1607 1608 1609 1610 1611 |
void ceph_flush_snaps(struct ceph_inode_info *ci, struct ceph_mds_session **psession) { struct inode *inode = &ci->vfs_inode; struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; |
e4d2b16a4 ceph: fix null po... |
1612 |
struct ceph_mds_session *session = NULL; |
ed9b430c9 ceph: cleanup cep... |
1613 |
int mds; |
e4d2b16a4 ceph: fix null po... |
1614 |
|
ed9b430c9 ceph: cleanup cep... |
1615 1616 |
dout("ceph_flush_snaps %p ", inode); |
e4d2b16a4 ceph: fix null po... |
1617 1618 |
if (psession) session = *psession; |
ed9b430c9 ceph: cleanup cep... |
1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 |
retry: spin_lock(&ci->i_ceph_lock); if (!(ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS)) { dout(" no capsnap needs flush, doing nothing "); goto out; } if (!ci->i_auth_cap) { dout(" no auth cap (migrating?), doing nothing "); goto out; } |
a8599bd82 ceph: capability ... |
1631 |
|
ed9b430c9 ceph: cleanup cep... |
1632 1633 1634 1635 |
mds = ci->i_auth_cap->session->s_mds; if (session && session->s_mds != mds) { dout(" oops, wrong session %p mutex ", session); |
a8599bd82 ceph: capability ... |
1636 1637 |
mutex_unlock(&session->s_mutex); ceph_put_mds_session(session); |
ed9b430c9 ceph: cleanup cep... |
1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 |
session = NULL; } if (!session) { spin_unlock(&ci->i_ceph_lock); mutex_lock(&mdsc->mutex); session = __ceph_lookup_mds_session(mdsc, mds); mutex_unlock(&mdsc->mutex); if (session) { dout(" inverting session/ino locks on %p ", session); mutex_lock(&session->s_mutex); } goto retry; |
a8599bd82 ceph: capability ... |
1651 |
} |
a8599bd82 ceph: capability ... |
1652 |
|
24d063acc ceph: make sure f... |
1653 |
// make sure flushsnap messages are sent in proper order. |
054f8d41a ceph: clear CEPH_... |
1654 |
if (ci->i_ceph_flags & CEPH_I_KICK_FLUSH) |
24d063acc ceph: make sure f... |
1655 |
__kick_flushing_caps(mdsc, session, ci, 0); |
24d063acc ceph: make sure f... |
1656 |
|
ed9b430c9 ceph: cleanup cep... |
1657 1658 |
__ceph_flush_snaps(ci, session); out: |
be655596b ceph: use i_ceph_... |
1659 |
spin_unlock(&ci->i_ceph_lock); |
ed9b430c9 ceph: cleanup cep... |
1660 1661 1662 |
if (psession) { *psession = session; |
c858a0709 ceph: fix NULL po... |
1663 |
} else if (session) { |
ed9b430c9 ceph: cleanup cep... |
1664 1665 1666 1667 1668 1669 1670 |
mutex_unlock(&session->s_mutex); ceph_put_mds_session(session); } /* we flushed them all; remove this inode from the queue */ spin_lock(&mdsc->snap_flush_lock); list_del_init(&ci->i_snap_flush_item); spin_unlock(&mdsc->snap_flush_lock); |
a8599bd82 ceph: capability ... |
1671 1672 1673 |
} /* |
fca65b4ad ceph: do not call... |
1674 1675 1676 |
* Mark caps dirty. If inode is newly dirty, return the dirty flags. * Caller is then responsible for calling __mark_inode_dirty with the * returned flags value. |
76e3b390d ceph: move dirty ... |
1677 |
*/ |
f66fd9f09 ceph: pre-allocat... |
1678 1679 |
int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask, struct ceph_cap_flush **pcf) |
76e3b390d ceph: move dirty ... |
1680 |
{ |
640ef79d2 ceph: use ceph_sb... |
1681 |
struct ceph_mds_client *mdsc = |
3d14c5d2b ceph: factor out ... |
1682 |
ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc; |
76e3b390d ceph: move dirty ... |
1683 1684 1685 |
struct inode *inode = &ci->vfs_inode; int was = ci->i_dirty_caps; int dirty = 0; |
c7e4f85ce ceph: more caps.c... |
1686 |
lockdep_assert_held(&ci->i_ceph_lock); |
571ade336 ceph: don't mark ... |
1687 1688 1689 1690 1691 1692 1693 |
if (!ci->i_auth_cap) { pr_warn("__mark_dirty_caps %p %llx mask %s, " "but no auth cap (session was closed?) ", inode, ceph_ino(inode), ceph_cap_string(mask)); return 0; } |
76e3b390d ceph: move dirty ... |
1694 1695 1696 1697 1698 1699 |
dout("__mark_dirty_caps %p %s dirty %s -> %s ", &ci->vfs_inode, ceph_cap_string(mask), ceph_cap_string(was), ceph_cap_string(was | mask)); ci->i_dirty_caps |= mask; if (was == 0) { |
1cf03a68e ceph: convert mds... |
1700 |
struct ceph_mds_session *session = ci->i_auth_cap->session; |
f66fd9f09 ceph: pre-allocat... |
1701 1702 |
WARN_ON_ONCE(ci->i_prealloc_cap_flush); swap(ci->i_prealloc_cap_flush, *pcf); |
604d1b024 ceph: take snap_r... |
1703 1704 |
if (!ci->i_head_snapc) { WARN_ON_ONCE(!rwsem_is_locked(&mdsc->snap_rwsem)); |
7d8cb26d7 ceph: maintain i_... |
1705 1706 |
ci->i_head_snapc = ceph_get_snap_context( ci->i_snap_realm->cached_context); |
604d1b024 ceph: take snap_r... |
1707 |
} |
0685235ff ceph: Don't add d... |
1708 1709 1710 |
dout(" inode %p now dirty snapc %p auth cap %p ", &ci->vfs_inode, ci->i_head_snapc, ci->i_auth_cap); |
76e3b390d ceph: move dirty ... |
1711 1712 |
BUG_ON(!list_empty(&ci->i_dirty_item)); spin_lock(&mdsc->cap_dirty_lock); |
1cf03a68e ceph: convert mds... |
1713 |
list_add(&ci->i_dirty_item, &session->s_cap_dirty); |
76e3b390d ceph: move dirty ... |
1714 1715 |
spin_unlock(&mdsc->cap_dirty_lock); if (ci->i_flushing_caps == 0) { |
3772d26d8 ceph: use ihold()... |
1716 |
ihold(inode); |
76e3b390d ceph: move dirty ... |
1717 1718 |
dirty |= I_DIRTY_SYNC; } |
f66fd9f09 ceph: pre-allocat... |
1719 1720 |
} else { WARN_ON_ONCE(!ci->i_prealloc_cap_flush); |
76e3b390d ceph: move dirty ... |
1721 1722 1723 1724 1725 |
} BUG_ON(list_empty(&ci->i_dirty_item)); if (((was | ci->i_flushing_caps) & CEPH_CAP_FILE_BUFFER) && (mask & CEPH_CAP_FILE_BUFFER)) dirty |= I_DIRTY_DATASYNC; |
a0d93e327 ceph: remove dela... |
1726 |
__cap_delay_requeue(mdsc, ci); |
fca65b4ad ceph: do not call... |
1727 |
return dirty; |
76e3b390d ceph: move dirty ... |
1728 |
} |
f66fd9f09 ceph: pre-allocat... |
1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 |
struct ceph_cap_flush *ceph_alloc_cap_flush(void) { return kmem_cache_alloc(ceph_cap_flush_cachep, GFP_KERNEL); } void ceph_free_cap_flush(struct ceph_cap_flush *cf) { if (cf) kmem_cache_free(ceph_cap_flush_cachep, cf); } |
a2971c8cc ceph: send TID of... |
1739 1740 |
static u64 __get_oldest_flush_tid(struct ceph_mds_client *mdsc) { |
e4500b5e3 ceph: use list in... |
1741 |
if (!list_empty(&mdsc->cap_flush_list)) { |
a2971c8cc ceph: send TID of... |
1742 |
struct ceph_cap_flush *cf = |
e4500b5e3 ceph: use list in... |
1743 1744 |
list_first_entry(&mdsc->cap_flush_list, struct ceph_cap_flush, g_list); |
a2971c8cc ceph: send TID of... |
1745 1746 1747 1748 |
return cf->tid; } return 0; } |
76e3b390d ceph: move dirty ... |
1749 |
/* |
c8799fc46 ceph: optimize ca... |
1750 1751 1752 |
* Remove cap_flush from the mdsc's or inode's flushing cap list. * Return true if caller needs to wake up flush waiters. */ |
681ac6348 ceph: split up __... |
1753 1754 |
static bool __detach_cap_flush_from_mdsc(struct ceph_mds_client *mdsc, struct ceph_cap_flush *cf) |
c8799fc46 ceph: optimize ca... |
1755 1756 1757 |
{ struct ceph_cap_flush *prev; bool wake = cf->wake; |
681ac6348 ceph: split up __... |
1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 |
if (wake && cf->g_list.prev != &mdsc->cap_flush_list) { prev = list_prev_entry(cf, g_list); prev->wake = true; wake = false; } list_del(&cf->g_list); return wake; } static bool __detach_cap_flush_from_ci(struct ceph_inode_info *ci, struct ceph_cap_flush *cf) { struct ceph_cap_flush *prev; bool wake = cf->wake; if (wake && cf->i_list.prev != &ci->i_cap_flush_list) { prev = list_prev_entry(cf, i_list); prev->wake = true; wake = false; |
c8799fc46 ceph: optimize ca... |
1778 |
} |
681ac6348 ceph: split up __... |
1779 |
list_del(&cf->i_list); |
c8799fc46 ceph: optimize ca... |
1780 1781 1782 1783 |
return wake; } /* |
a8599bd82 ceph: capability ... |
1784 1785 |
* Add dirty inode to the flushing list. Assigned a seq number so we * can wait for caps to flush without starving. |
cdc35f962 ceph: move generi... |
1786 |
* |
9f3345d8e ceph: have __mark... |
1787 |
* Called under i_ceph_lock. Returns the flush tid. |
a8599bd82 ceph: capability ... |
1788 |
*/ |
9f3345d8e ceph: have __mark... |
1789 |
static u64 __mark_caps_flushing(struct inode *inode, |
c8799fc46 ceph: optimize ca... |
1790 |
struct ceph_mds_session *session, bool wake, |
9f3345d8e ceph: have __mark... |
1791 |
u64 *oldest_flush_tid) |
a8599bd82 ceph: capability ... |
1792 |
{ |
3d14c5d2b ceph: factor out ... |
1793 |
struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; |
a8599bd82 ceph: capability ... |
1794 |
struct ceph_inode_info *ci = ceph_inode(inode); |
f66fd9f09 ceph: pre-allocat... |
1795 |
struct ceph_cap_flush *cf = NULL; |
cdc35f962 ceph: move generi... |
1796 |
int flushing; |
50b885b96 ceph: whitespace ... |
1797 |
|
c7e4f85ce ceph: more caps.c... |
1798 |
lockdep_assert_held(&ci->i_ceph_lock); |
cdc35f962 ceph: move generi... |
1799 |
BUG_ON(ci->i_dirty_caps == 0); |
a8599bd82 ceph: capability ... |
1800 |
BUG_ON(list_empty(&ci->i_dirty_item)); |
f66fd9f09 ceph: pre-allocat... |
1801 |
BUG_ON(!ci->i_prealloc_cap_flush); |
cdc35f962 ceph: move generi... |
1802 1803 1804 1805 1806 1807 1808 1809 1810 |
flushing = ci->i_dirty_caps; dout("__mark_caps_flushing flushing %s, flushing_caps %s -> %s ", ceph_cap_string(flushing), ceph_cap_string(ci->i_flushing_caps), ceph_cap_string(ci->i_flushing_caps | flushing)); ci->i_flushing_caps |= flushing; ci->i_dirty_caps = 0; |
afcdaea3f ceph: flush dirty... |
1811 1812 |
dout(" inode %p now !dirty ", inode); |
cdc35f962 ceph: move generi... |
1813 |
|
f66fd9f09 ceph: pre-allocat... |
1814 |
swap(cf, ci->i_prealloc_cap_flush); |
553adfd94 ceph: track pendi... |
1815 |
cf->caps = flushing; |
c8799fc46 ceph: optimize ca... |
1816 |
cf->wake = wake; |
553adfd94 ceph: track pendi... |
1817 |
|
a8599bd82 ceph: capability ... |
1818 |
spin_lock(&mdsc->cap_dirty_lock); |
afcdaea3f ceph: flush dirty... |
1819 |
list_del_init(&ci->i_dirty_item); |
553adfd94 ceph: track pendi... |
1820 |
cf->tid = ++mdsc->last_cap_flush_tid; |
e4500b5e3 ceph: use list in... |
1821 |
list_add_tail(&cf->g_list, &mdsc->cap_flush_list); |
a2971c8cc ceph: send TID of... |
1822 |
*oldest_flush_tid = __get_oldest_flush_tid(mdsc); |
553adfd94 ceph: track pendi... |
1823 |
|
a8599bd82 ceph: capability ... |
1824 1825 1826 |
if (list_empty(&ci->i_flushing_item)) { list_add_tail(&ci->i_flushing_item, &session->s_cap_flushing); mdsc->num_cap_flushing++; |
a8599bd82 ceph: capability ... |
1827 1828 |
} spin_unlock(&mdsc->cap_dirty_lock); |
cdc35f962 ceph: move generi... |
1829 |
|
e4500b5e3 ceph: use list in... |
1830 |
list_add_tail(&cf->i_list, &ci->i_cap_flush_list); |
553adfd94 ceph: track pendi... |
1831 |
|
9f3345d8e ceph: have __mark... |
1832 |
return cf->tid; |
a8599bd82 ceph: capability ... |
1833 1834 1835 |
} /* |
5ecad6fd7 ceph: fix check f... |
1836 1837 |
* try to invalidate mapping pages without blocking. */ |
5ecad6fd7 ceph: fix check f... |
1838 1839 1840 1841 |
static int try_nonblocking_invalidate(struct inode *inode) { struct ceph_inode_info *ci = ceph_inode(inode); u32 invalidating_gen = ci->i_rdcache_gen; |
be655596b ceph: use i_ceph_... |
1842 |
spin_unlock(&ci->i_ceph_lock); |
5ecad6fd7 ceph: fix check f... |
1843 |
invalidate_mapping_pages(&inode->i_data, 0, -1); |
be655596b ceph: use i_ceph_... |
1844 |
spin_lock(&ci->i_ceph_lock); |
5ecad6fd7 ceph: fix check f... |
1845 |
|
18a38193e ceph: use mapping... |
1846 |
if (inode->i_data.nrpages == 0 && |
5ecad6fd7 ceph: fix check f... |
1847 1848 1849 1850 |
invalidating_gen == ci->i_rdcache_gen) { /* success. */ dout("try_nonblocking_invalidate %p success ", inode); |
cd045cb42 ceph: fix rdcache... |
1851 1852 |
/* save any racing async invalidate some trouble */ ci->i_rdcache_revoking = ci->i_rdcache_gen - 1; |
5ecad6fd7 ceph: fix check f... |
1853 1854 1855 1856 1857 1858 |
return 0; } dout("try_nonblocking_invalidate %p failed ", inode); return -1; } |
efb0ca765 ceph: update the ... |
1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 |
bool __ceph_should_report_size(struct ceph_inode_info *ci) { loff_t size = ci->vfs_inode.i_size; /* mds will adjust max size according to the reported size */ if (ci->i_flushing_caps & CEPH_CAP_FILE_WR) return false; if (size >= ci->i_max_size) return true; /* half of previous max_size increment has been used */ if (ci->i_max_size > ci->i_reported_size && (size << 1) >= ci->i_max_size + ci->i_reported_size) return true; return false; } |
5ecad6fd7 ceph: fix check f... |
1873 |
/* |
a8599bd82 ceph: capability ... |
1874 1875 1876 1877 |
* Swiss army knife function to examine currently used and wanted * versus held caps. Release, flush, ack revoked caps to mds as * appropriate. * |
a8599bd82 ceph: capability ... |
1878 1879 1880 1881 1882 1883 1884 |
* CHECK_CAPS_AUTHONLY - we should only check the auth cap * CHECK_CAPS_FLUSH - we should flush any dirty caps immediately, without * further delay. */ void ceph_check_caps(struct ceph_inode_info *ci, int flags, struct ceph_mds_session *session) { |
a8599bd82 ceph: capability ... |
1885 |
struct inode *inode = &ci->vfs_inode; |
2678da88f ceph: add ceph_sb... |
1886 |
struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb); |
a8599bd82 ceph: capability ... |
1887 |
struct ceph_cap *cap; |
a2971c8cc ceph: send TID of... |
1888 |
u64 flush_tid, oldest_flush_tid; |
395c312b9 ceph: allow revok... |
1889 |
int file_wanted, used, cap_used; |
a8599bd82 ceph: capability ... |
1890 |
int took_snap_rwsem = 0; /* true if mdsc->snap_rwsem held */ |
cbd036359 ceph: cap revocat... |
1891 |
int issued, implemented, want, retain, revoking, flushing = 0; |
a8599bd82 ceph: capability ... |
1892 1893 1894 |
int mds = -1; /* keep track of how far we've gone through i_caps list to avoid an infinite loop on retry */ struct rb_node *p; |
3609404f8 ceph: update type... |
1895 |
bool queue_invalidate = false; |
3609404f8 ceph: update type... |
1896 |
bool tried_invalidate = false; |
a8599bd82 ceph: capability ... |
1897 |
|
be655596b ceph: use i_ceph_... |
1898 |
spin_lock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
1899 1900 |
if (ci->i_ceph_flags & CEPH_I_FLUSH) flags |= CHECK_CAPS_FLUSH; |
a8599bd82 ceph: capability ... |
1901 1902 |
goto retry_locked; retry: |
be655596b ceph: use i_ceph_... |
1903 |
spin_lock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
1904 |
retry_locked: |
c74d79af9 ceph: comment cle... |
1905 |
/* Caps wanted by virtue of active open files. */ |
a8599bd82 ceph: capability ... |
1906 |
file_wanted = __ceph_caps_file_wanted(ci); |
c74d79af9 ceph: comment cle... |
1907 1908 |
/* Caps which have active references against them */ |
a8599bd82 ceph: capability ... |
1909 |
used = __ceph_caps_used(ci); |
c74d79af9 ceph: comment cle... |
1910 1911 1912 1913 1914 1915 1916 |
/* * "issued" represents the current caps that the MDS wants us to have. * "implemented" is the set that we have been granted, and includes the * ones that have not yet been returned to the MDS (the "revoking" set, * usually because they have outstanding references). */ |
cbd036359 ceph: cap revocat... |
1917 1918 |
issued = __ceph_caps_issued(ci, &implemented); revoking = implemented & ~issued; |
a8599bd82 ceph: capability ... |
1919 |
|
41445999a ceph: don't inclu... |
1920 |
want = file_wanted; |
c74d79af9 ceph: comment cle... |
1921 1922 |
/* The ones we currently want to retain (may be adjusted below) */ |
41445999a ceph: don't inclu... |
1923 |
retain = file_wanted | used | CEPH_CAP_PIN; |
a8599bd82 ceph: capability ... |
1924 |
if (!mdsc->stopping && inode->i_nlink > 0) { |
41445999a ceph: don't inclu... |
1925 |
if (file_wanted) { |
a8599bd82 ceph: capability ... |
1926 |
retain |= CEPH_CAP_ANY; /* be greedy */ |
32ec43977 ceph: hold on to ... |
1927 1928 |
} else if (S_ISDIR(inode->i_mode) && (issued & CEPH_CAP_FILE_SHARED) && |
8a2ac3a8e ceph: don't reque... |
1929 |
__ceph_dir_is_complete(ci)) { |
32ec43977 ceph: hold on to ... |
1930 1931 1932 1933 1934 1935 |
/* * If a directory is complete, we want to keep * the exclusive cap. So that MDS does not end up * revoking the shared cap on every create/unlink * operation. */ |
a25949b99 ceph: cap trackin... |
1936 |
if (IS_RDONLY(inode)) { |
8a2ac3a8e ceph: don't reque... |
1937 |
want = CEPH_CAP_ANY_SHARED; |
a25949b99 ceph: cap trackin... |
1938 |
} else { |
719a2514e ceph: consider in... |
1939 |
want |= CEPH_CAP_ANY_SHARED | CEPH_CAP_FILE_EXCL; |
a25949b99 ceph: cap trackin... |
1940 |
} |
32ec43977 ceph: hold on to ... |
1941 |
retain |= want; |
a8599bd82 ceph: capability ... |
1942 |
} else { |
32ec43977 ceph: hold on to ... |
1943 |
|
a8599bd82 ceph: capability ... |
1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 |
retain |= CEPH_CAP_ANY_SHARED; /* * keep RD only if we didn't have the file open RW, * because then the mds would revoke it anyway to * journal max_size=0. */ if (ci->i_max_size == 0) retain |= CEPH_CAP_ANY_RD; } } dout("check_caps %p file_want %s used %s dirty %s flushing %s" |
a0d93e327 ceph: remove dela... |
1956 1957 |
" issued %s revoking %s retain %s %s%s ", inode, |
a8599bd82 ceph: capability ... |
1958 1959 1960 |
ceph_cap_string(file_wanted), ceph_cap_string(used), ceph_cap_string(ci->i_dirty_caps), ceph_cap_string(ci->i_flushing_caps), |
cbd036359 ceph: cap revocat... |
1961 |
ceph_cap_string(issued), ceph_cap_string(revoking), |
a8599bd82 ceph: capability ... |
1962 1963 |
ceph_cap_string(retain), (flags & CHECK_CAPS_AUTHONLY) ? " AUTHONLY" : "", |
a8599bd82 ceph: capability ... |
1964 1965 1966 1967 1968 1969 1970 |
(flags & CHECK_CAPS_FLUSH) ? " FLUSH" : ""); /* * If we no longer need to hold onto old our caps, and we may * have cached pages, but don't want them, then try to invalidate. * If we fail, it's because pages are locked.... try again later. */ |
a0d93e327 ceph: remove dela... |
1971 |
if ((!(flags & CHECK_CAPS_NOINVAL) || mdsc->stopping) && |
525d15e8e ceph: check inode... |
1972 |
S_ISREG(inode->i_mode) && |
9abd4db71 ceph: don't use t... |
1973 |
!(ci->i_wb_ref || ci->i_wrbuffer_ref) && /* no dirty pages... */ |
fdd4e1583 ceph: rework dcac... |
1974 |
inode->i_data.nrpages && /* have cached pages */ |
5e804ac48 ceph: don't inval... |
1975 1976 |
(revoking & (CEPH_CAP_FILE_CACHE| CEPH_CAP_FILE_LAZYIO)) && /* or revoking cache */ |
a8599bd82 ceph: capability ... |
1977 |
!tried_invalidate) { |
a8599bd82 ceph: capability ... |
1978 1979 |
dout("check_caps trying to invalidate on %p ", inode); |
5ecad6fd7 ceph: fix check f... |
1980 |
if (try_nonblocking_invalidate(inode) < 0) { |
ee612d954 ceph: delete unre... |
1981 1982 1983 1984 |
dout("check_caps queuing invalidate "); queue_invalidate = true; ci->i_rdcache_revoking = ci->i_rdcache_gen; |
a8599bd82 ceph: capability ... |
1985 |
} |
3609404f8 ceph: update type... |
1986 |
tried_invalidate = true; |
a8599bd82 ceph: capability ... |
1987 1988 |
goto retry_locked; } |
a8599bd82 ceph: capability ... |
1989 |
for (p = rb_first(&ci->i_caps); p; p = rb_next(p)) { |
d67c72e6c ceph: request exp... |
1990 |
int mflags = 0; |
0a454bdd5 ceph: reorganize ... |
1991 |
struct cap_msg_args arg; |
a8599bd82 ceph: capability ... |
1992 |
cap = rb_entry(p, struct ceph_cap, ci_node); |
a8599bd82 ceph: capability ... |
1993 1994 1995 1996 1997 1998 1999 |
/* avoid looping forever */ if (mds >= cap->mds || ((flags & CHECK_CAPS_AUTHONLY) && cap != ci->i_auth_cap)) continue; /* NOTE: no side-effects allowed, until we take s_mutex */ |
c74d79af9 ceph: comment cle... |
2000 2001 2002 2003 |
/* * If we have an auth cap, we don't need to consider any * overlapping caps as used. */ |
395c312b9 ceph: allow revok... |
2004 2005 2006 |
cap_used = used; if (ci->i_auth_cap && cap != ci->i_auth_cap) cap_used &= ~ci->i_auth_cap->issued; |
a8599bd82 ceph: capability ... |
2007 |
revoking = cap->implemented & ~cap->issued; |
395c312b9 ceph: allow revok... |
2008 2009 |
dout(" mds%d cap %p used %s issued %s implemented %s revoking %s ", |
9abd4db71 ceph: don't use t... |
2010 2011 |
cap->mds, cap, ceph_cap_string(cap_used), ceph_cap_string(cap->issued), |
088b3f5e9 ceph: fix flushin... |
2012 2013 |
ceph_cap_string(cap->implemented), ceph_cap_string(revoking)); |
a8599bd82 ceph: capability ... |
2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 |
if (cap == ci->i_auth_cap && (cap->issued & CEPH_CAP_FILE_WR)) { /* request larger max_size from MDS? */ if (ci->i_wanted_max_size > ci->i_max_size && ci->i_wanted_max_size > ci->i_requested_max_size) { dout("requesting new max_size "); goto ack; } /* approaching file_max? */ |
efb0ca765 ceph: update the ... |
2026 |
if (__ceph_should_report_size(ci)) { |
a8599bd82 ceph: capability ... |
2027 2028 2029 2030 2031 2032 |
dout("i_size approaching max_size "); goto ack; } } /* flush anything dirty? */ |
7bc00fddb ceph: kick cap fl... |
2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 |
if (cap == ci->i_auth_cap) { if ((flags & CHECK_CAPS_FLUSH) && ci->i_dirty_caps) { dout("flushing dirty caps "); goto ack; } if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) { dout("flushing snap caps "); goto ack; } |
a8599bd82 ceph: capability ... |
2044 2045 2046 |
} /* completed revocation? going down and there are no caps? */ |
395c312b9 ceph: allow revok... |
2047 |
if (revoking && (revoking & cap_used) == 0) { |
a8599bd82 ceph: capability ... |
2048 2049 2050 2051 2052 2053 2054 |
dout("completed revocation of %s ", ceph_cap_string(cap->implemented & ~cap->issued)); goto ack; } /* want more caps from mds? */ |
0aa971b6f ceph: don't skip ... |
2055 2056 2057 2058 2059 2060 |
if (want & ~cap->mds_wanted) { if (want & ~(cap->mds_wanted | cap->issued)) goto ack; if (!__cap_is_valid(cap)) goto ack; } |
a8599bd82 ceph: capability ... |
2061 2062 |
/* things we might delay */ |
fdac94fab ceph: skip updati... |
2063 |
if ((cap->issued & ~retain) == 0) |
a8599bd82 ceph: capability ... |
2064 |
continue; /* nope, all good */ |
a8599bd82 ceph: capability ... |
2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 |
ack: if (session && session != cap->session) { dout("oops, wrong session %p mutex ", session); mutex_unlock(&session->s_mutex); session = NULL; } if (!session) { session = cap->session; if (mutex_trylock(&session->s_mutex) == 0) { dout("inverting session/ino locks on %p ", session); |
dc3da0461 ceph: fix potenti... |
2078 |
session = ceph_get_mds_session(session); |
be655596b ceph: use i_ceph_... |
2079 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2080 2081 2082 2083 |
if (took_snap_rwsem) { up_read(&mdsc->snap_rwsem); took_snap_rwsem = 0; } |
dc3da0461 ceph: fix potenti... |
2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 |
if (session) { mutex_lock(&session->s_mutex); ceph_put_mds_session(session); } else { /* * Because we take the reference while * holding the i_ceph_lock, it should * never be NULL. Throw a warning if it * ever is. */ WARN_ON_ONCE(true); } |
a8599bd82 ceph: capability ... |
2096 2097 2098 |
goto retry; } } |
7bc00fddb ceph: kick cap fl... |
2099 2100 2101 2102 2103 2104 |
/* kick flushing and flush snaps before sending normal * cap message */ if (cap == ci->i_auth_cap && (ci->i_ceph_flags & (CEPH_I_KICK_FLUSH | CEPH_I_FLUSH_SNAPS))) { |
054f8d41a ceph: clear CEPH_... |
2105 |
if (ci->i_ceph_flags & CEPH_I_KICK_FLUSH) |
24d063acc ceph: make sure f... |
2106 |
__kick_flushing_caps(mdsc, session, ci, 0); |
ed9b430c9 ceph: cleanup cep... |
2107 2108 |
if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) __ceph_flush_snaps(ci, session); |
7bc00fddb ceph: kick cap fl... |
2109 2110 |
goto retry_locked; } |
a8599bd82 ceph: capability ... |
2111 2112 2113 2114 2115 2116 |
/* take snap_rwsem after session mutex */ if (!took_snap_rwsem) { if (down_read_trylock(&mdsc->snap_rwsem) == 0) { dout("inverting snap/in locks on %p ", inode); |
be655596b ceph: use i_ceph_... |
2117 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2118 2119 2120 2121 2122 2123 |
down_read(&mdsc->snap_rwsem); took_snap_rwsem = 1; goto retry; } took_snap_rwsem = 1; } |
553adfd94 ceph: track pendi... |
2124 |
if (cap == ci->i_auth_cap && ci->i_dirty_caps) { |
9f3345d8e ceph: have __mark... |
2125 2126 2127 |
flushing = ci->i_dirty_caps; flush_tid = __mark_caps_flushing(inode, session, false, &oldest_flush_tid); |
d67c72e6c ceph: request exp... |
2128 2129 2130 |
if (flags & CHECK_CAPS_FLUSH && list_empty(&session->s_cap_dirty)) mflags |= CEPH_CLIENT_CAPS_SYNC; |
553adfd94 ceph: track pendi... |
2131 |
} else { |
24be0c481 ceph: fix erroneo... |
2132 |
flushing = 0; |
553adfd94 ceph: track pendi... |
2133 |
flush_tid = 0; |
a2971c8cc ceph: send TID of... |
2134 2135 2136 |
spin_lock(&mdsc->cap_dirty_lock); oldest_flush_tid = __get_oldest_flush_tid(mdsc); spin_unlock(&mdsc->cap_dirty_lock); |
553adfd94 ceph: track pendi... |
2137 |
} |
a8599bd82 ceph: capability ... |
2138 2139 |
mds = cap->mds; /* remember mds, so we don't repeat */ |
a8599bd82 ceph: capability ... |
2140 |
|
d67c72e6c ceph: request exp... |
2141 2142 |
__prep_cap(&arg, cap, CEPH_CAP_OP_UPDATE, mflags, cap_used, want, retain, flushing, flush_tid, oldest_flush_tid); |
0a454bdd5 ceph: reorganize ... |
2143 |
spin_unlock(&ci->i_ceph_lock); |
523119808 ceph: drop separa... |
2144 |
__send_cap(&arg, ci); |
0a454bdd5 ceph: reorganize ... |
2145 |
|
be655596b ceph: use i_ceph_... |
2146 |
goto retry; /* retake i_ceph_lock and restart our cap scan. */ |
a8599bd82 ceph: capability ... |
2147 |
} |
a0d93e327 ceph: remove dela... |
2148 2149 2150 2151 2152 2153 |
/* periodically re-calculate caps wanted by open files */ if (__ceph_is_any_real_caps(ci) && list_empty(&ci->i_cap_delay_list) && (file_wanted & ~CEPH_CAP_PIN) && !(used & (CEPH_CAP_FILE_RD | CEPH_CAP_ANY_FILE_WR))) { __cap_delay_requeue(mdsc, ci); |
719a2514e ceph: consider in... |
2154 |
} |
a8599bd82 ceph: capability ... |
2155 |
|
be655596b ceph: use i_ceph_... |
2156 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2157 |
|
cbd036359 ceph: cap revocat... |
2158 |
if (queue_invalidate) |
3c6f6b79a ceph: cleanup asy... |
2159 |
ceph_queue_invalidate(inode); |
cbd036359 ceph: cap revocat... |
2160 |
|
cdc2ce056 ceph: fix session... |
2161 |
if (session) |
a8599bd82 ceph: capability ... |
2162 2163 2164 2165 2166 2167 |
mutex_unlock(&session->s_mutex); if (took_snap_rwsem) up_read(&mdsc->snap_rwsem); } /* |
a8599bd82 ceph: capability ... |
2168 2169 |
* Try to flush dirty caps back to the auth mds. */ |
553adfd94 ceph: track pendi... |
2170 |
static int try_flush_caps(struct inode *inode, u64 *ptid) |
a8599bd82 ceph: capability ... |
2171 |
{ |
3d14c5d2b ceph: factor out ... |
2172 |
struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; |
a8599bd82 ceph: capability ... |
2173 |
struct ceph_inode_info *ci = ceph_inode(inode); |
4fe59789a ceph: handle cap ... |
2174 |
struct ceph_mds_session *session = NULL; |
89b52fe14 ceph: fix flushin... |
2175 |
int flushing = 0; |
a2971c8cc ceph: send TID of... |
2176 |
u64 flush_tid = 0, oldest_flush_tid = 0; |
a8599bd82 ceph: capability ... |
2177 2178 |
retry: |
be655596b ceph: use i_ceph_... |
2179 |
spin_lock(&ci->i_ceph_lock); |
d6cee9dbd ceph: kick flushi... |
2180 |
retry_locked: |
a8599bd82 ceph: capability ... |
2181 2182 |
if (ci->i_dirty_caps && ci->i_auth_cap) { struct ceph_cap *cap = ci->i_auth_cap; |
0a454bdd5 ceph: reorganize ... |
2183 |
struct cap_msg_args arg; |
a8599bd82 ceph: capability ... |
2184 |
|
27b0a3920 ceph: remove unne... |
2185 |
if (session != cap->session) { |
be655596b ceph: use i_ceph_... |
2186 |
spin_unlock(&ci->i_ceph_lock); |
4fe59789a ceph: handle cap ... |
2187 2188 |
if (session) mutex_unlock(&session->s_mutex); |
a8599bd82 ceph: capability ... |
2189 2190 2191 2192 |
session = cap->session; mutex_lock(&session->s_mutex); goto retry; } |
6c2838fbd ceph: unlock dang... |
2193 2194 |
if (cap->session->s_state < CEPH_MDS_SESSION_OPEN) { spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2195 |
goto out; |
6c2838fbd ceph: unlock dang... |
2196 |
} |
a8599bd82 ceph: capability ... |
2197 |
|
d6cee9dbd ceph: kick flushi... |
2198 2199 2200 2201 2202 2203 2204 2205 |
if (ci->i_ceph_flags & (CEPH_I_KICK_FLUSH | CEPH_I_FLUSH_SNAPS)) { if (ci->i_ceph_flags & CEPH_I_KICK_FLUSH) __kick_flushing_caps(mdsc, session, ci, 0); if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) __ceph_flush_snaps(ci, session); goto retry_locked; } |
9f3345d8e ceph: have __mark... |
2206 2207 2208 |
flushing = ci->i_dirty_caps; flush_tid = __mark_caps_flushing(inode, session, true, &oldest_flush_tid); |
a8599bd82 ceph: capability ... |
2209 |
|
0a454bdd5 ceph: reorganize ... |
2210 |
__prep_cap(&arg, cap, CEPH_CAP_OP_FLUSH, CEPH_CLIENT_CAPS_SYNC, |
a0d93e327 ceph: remove dela... |
2211 2212 2213 |
__ceph_caps_used(ci), __ceph_caps_wanted(ci), (cap->issued | cap->implemented), flushing, flush_tid, oldest_flush_tid); |
0a454bdd5 ceph: reorganize ... |
2214 |
spin_unlock(&ci->i_ceph_lock); |
523119808 ceph: drop separa... |
2215 |
__send_cap(&arg, ci); |
553adfd94 ceph: track pendi... |
2216 |
} else { |
e4500b5e3 ceph: use list in... |
2217 |
if (!list_empty(&ci->i_cap_flush_list)) { |
553adfd94 ceph: track pendi... |
2218 |
struct ceph_cap_flush *cf = |
e4500b5e3 ceph: use list in... |
2219 |
list_last_entry(&ci->i_cap_flush_list, |
c8799fc46 ceph: optimize ca... |
2220 2221 |
struct ceph_cap_flush, i_list); cf->wake = true; |
553adfd94 ceph: track pendi... |
2222 2223 2224 2225 |
flush_tid = cf->tid; } flushing = ci->i_flushing_caps; spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2226 2227 |
} out: |
4fe59789a ceph: handle cap ... |
2228 |
if (session) |
a8599bd82 ceph: capability ... |
2229 |
mutex_unlock(&session->s_mutex); |
553adfd94 ceph: track pendi... |
2230 2231 |
*ptid = flush_tid; |
a8599bd82 ceph: capability ... |
2232 2233 2234 2235 2236 2237 |
return flushing; } /* * Return true if we've flushed caps through the given flush_tid. */ |
553adfd94 ceph: track pendi... |
2238 |
static int caps_are_flushed(struct inode *inode, u64 flush_tid) |
a8599bd82 ceph: capability ... |
2239 2240 |
{ struct ceph_inode_info *ci = ceph_inode(inode); |
553adfd94 ceph: track pendi... |
2241 |
int ret = 1; |
a8599bd82 ceph: capability ... |
2242 |
|
be655596b ceph: use i_ceph_... |
2243 |
spin_lock(&ci->i_ceph_lock); |
e4500b5e3 ceph: use list in... |
2244 2245 2246 2247 |
if (!list_empty(&ci->i_cap_flush_list)) { struct ceph_cap_flush * cf = list_first_entry(&ci->i_cap_flush_list, struct ceph_cap_flush, i_list); |
553adfd94 ceph: track pendi... |
2248 |
if (cf->tid <= flush_tid) |
a8599bd82 ceph: capability ... |
2249 |
ret = 0; |
89b52fe14 ceph: fix flushin... |
2250 |
} |
be655596b ceph: use i_ceph_... |
2251 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2252 2253 2254 2255 |
return ret; } /* |
68cd5b4b7 ceph: make fsync(... |
2256 |
* wait for any unsafe requests to complete. |
da819c815 ceph: fix directo... |
2257 |
*/ |
68cd5b4b7 ceph: make fsync(... |
2258 |
static int unsafe_request_wait(struct inode *inode) |
da819c815 ceph: fix directo... |
2259 2260 |
{ struct ceph_inode_info *ci = ceph_inode(inode); |
68cd5b4b7 ceph: make fsync(... |
2261 2262 |
struct ceph_mds_request *req1 = NULL, *req2 = NULL; int ret, err = 0; |
da819c815 ceph: fix directo... |
2263 2264 |
spin_lock(&ci->i_unsafe_lock); |
68cd5b4b7 ceph: make fsync(... |
2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 |
if (S_ISDIR(inode->i_mode) && !list_empty(&ci->i_unsafe_dirops)) { req1 = list_last_entry(&ci->i_unsafe_dirops, struct ceph_mds_request, r_unsafe_dir_item); ceph_mdsc_get_request(req1); } if (!list_empty(&ci->i_unsafe_iops)) { req2 = list_last_entry(&ci->i_unsafe_iops, struct ceph_mds_request, r_unsafe_target_item); ceph_mdsc_get_request(req2); } spin_unlock(&ci->i_unsafe_lock); |
da819c815 ceph: fix directo... |
2278 |
|
4945a0847 ceph: fix minor t... |
2279 2280 |
dout("unsafe_request_wait %p wait on tid %llu %llu ", |
68cd5b4b7 ceph: make fsync(... |
2281 2282 2283 2284 |
inode, req1 ? req1->r_tid : 0ULL, req2 ? req2->r_tid : 0ULL); if (req1) { ret = !wait_for_completion_timeout(&req1->r_safe_completion, ceph_timeout_jiffies(req1->r_timeout)); |
da819c815 ceph: fix directo... |
2285 |
if (ret) |
68cd5b4b7 ceph: make fsync(... |
2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 |
err = -EIO; ceph_mdsc_put_request(req1); } if (req2) { ret = !wait_for_completion_timeout(&req2->r_safe_completion, ceph_timeout_jiffies(req2->r_timeout)); if (ret) err = -EIO; ceph_mdsc_put_request(req2); } return err; |
da819c815 ceph: fix directo... |
2297 |
} |
02c24a821 fs: push i_mutex ... |
2298 |
int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync) |
a8599bd82 ceph: capability ... |
2299 |
{ |
f4b978662 ceph: track and r... |
2300 |
struct ceph_file_info *fi = file->private_data; |
7ea808591 drop unused dentr... |
2301 |
struct inode *inode = file->f_mapping->host; |
a8599bd82 ceph: capability ... |
2302 |
struct ceph_inode_info *ci = ceph_inode(inode); |
553adfd94 ceph: track pendi... |
2303 |
u64 flush_tid; |
f4b978662 ceph: track and r... |
2304 |
int ret, err; |
a8599bd82 ceph: capability ... |
2305 2306 2307 2308 |
int dirty; dout("fsync %p%s ", inode, datasync ? " datasync" : ""); |
9a5530c63 ceph: wait unsafe... |
2309 |
|
b74fceae7 ceph: use errseq_... |
2310 |
ret = file_write_and_wait_range(file, start, end); |
da819c815 ceph: fix directo... |
2311 2312 |
if (datasync) goto out; |
891f3f5a6 ceph: add infrast... |
2313 2314 2315 |
ret = ceph_wait_on_async_create(inode); if (ret) goto out; |
553adfd94 ceph: track pendi... |
2316 |
dirty = try_flush_caps(inode, &flush_tid); |
a8599bd82 ceph: capability ... |
2317 2318 |
dout("fsync dirty caps are %s ", ceph_cap_string(dirty)); |
f4b978662 ceph: track and r... |
2319 |
err = unsafe_request_wait(inode); |
da819c815 ceph: fix directo... |
2320 |
|
a8599bd82 ceph: capability ... |
2321 2322 2323 2324 2325 |
/* * only wait on non-file metadata writeback (the mds * can recover size and mtime, so we don't need to * wait for that) */ |
f4b978662 ceph: track and r... |
2326 2327 |
if (!err && (dirty & ~CEPH_CAP_ANY_FILE_WR)) { err = wait_event_interruptible(ci->i_cap_wq, |
da819c815 ceph: fix directo... |
2328 |
caps_are_flushed(inode, flush_tid)); |
a8599bd82 ceph: capability ... |
2329 |
} |
f4b978662 ceph: track and r... |
2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 |
if (err < 0) ret = err; if (errseq_check(&ci->i_meta_err, READ_ONCE(fi->meta_err))) { spin_lock(&file->f_lock); err = errseq_check_and_advance(&ci->i_meta_err, &fi->meta_err); spin_unlock(&file->f_lock); if (err < 0) ret = err; } |
da819c815 ceph: fix directo... |
2342 2343 2344 |
out: dout("fsync %p%s result=%d ", inode, datasync ? " datasync" : "", ret); |
a8599bd82 ceph: capability ... |
2345 2346 2347 2348 2349 2350 2351 2352 2353 |
return ret; } /* * Flush any dirty caps back to the mds. If we aren't asked to wait, * queue inode for flush but don't do so immediately, because we can * get by with fewer MDS messages if we wait for data writeback to * complete first. */ |
f1a3d5721 ceph: update for ... |
2354 |
int ceph_write_inode(struct inode *inode, struct writeback_control *wbc) |
a8599bd82 ceph: capability ... |
2355 2356 |
{ struct ceph_inode_info *ci = ceph_inode(inode); |
553adfd94 ceph: track pendi... |
2357 |
u64 flush_tid; |
a8599bd82 ceph: capability ... |
2358 2359 |
int err = 0; int dirty; |
16515a6d5 ceph: improving e... |
2360 |
int wait = (wbc->sync_mode == WB_SYNC_ALL && !wbc->for_sync); |
a8599bd82 ceph: capability ... |
2361 2362 2363 2364 |
dout("write_inode %p wait=%d ", inode, wait); if (wait) { |
553adfd94 ceph: track pendi... |
2365 |
dirty = try_flush_caps(inode, &flush_tid); |
a8599bd82 ceph: capability ... |
2366 2367 2368 2369 |
if (dirty) err = wait_event_interruptible(ci->i_cap_wq, caps_are_flushed(inode, flush_tid)); } else { |
640ef79d2 ceph: use ceph_sb... |
2370 |
struct ceph_mds_client *mdsc = |
3d14c5d2b ceph: factor out ... |
2371 |
ceph_sb_to_client(inode->i_sb)->mdsc; |
a8599bd82 ceph: capability ... |
2372 |
|
be655596b ceph: use i_ceph_... |
2373 |
spin_lock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2374 2375 |
if (__ceph_caps_dirty(ci)) __cap_delay_requeue_front(mdsc, ci); |
be655596b ceph: use i_ceph_... |
2376 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2377 2378 2379 |
} return err; } |
0e2943878 ceph: unify cap f... |
2380 2381 2382 2383 2384 2385 |
static void __kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session, struct ceph_inode_info *ci, u64 oldest_flush_tid) __releases(ci->i_ceph_lock) __acquires(ci->i_ceph_lock) |
553adfd94 ceph: track pendi... |
2386 2387 2388 2389 |
{ struct inode *inode = &ci->vfs_inode; struct ceph_cap *cap; struct ceph_cap_flush *cf; |
0e2943878 ceph: unify cap f... |
2390 |
int ret; |
553adfd94 ceph: track pendi... |
2391 |
u64 first_tid = 0; |
49ada6e8d ceph: more precis... |
2392 |
u64 last_snap_flush = 0; |
553adfd94 ceph: track pendi... |
2393 |
|
054f8d41a ceph: clear CEPH_... |
2394 |
ci->i_ceph_flags &= ~CEPH_I_KICK_FLUSH; |
49ada6e8d ceph: more precis... |
2395 2396 2397 2398 2399 2400 |
list_for_each_entry_reverse(cf, &ci->i_cap_flush_list, i_list) { if (!cf->caps) { last_snap_flush = cf->tid; break; } } |
e4500b5e3 ceph: use list in... |
2401 2402 2403 |
list_for_each_entry(cf, &ci->i_cap_flush_list, i_list) { if (cf->tid < first_tid) continue; |
553adfd94 ceph: track pendi... |
2404 2405 |
cap = ci->i_auth_cap; if (!(cap && cap->session == session)) { |
0e2943878 ceph: unify cap f... |
2406 2407 2408 |
pr_err("%p auth cap %p not mds%d ??? ", inode, cap, session->s_mds); |
553adfd94 ceph: track pendi... |
2409 2410 |
break; } |
553adfd94 ceph: track pendi... |
2411 |
first_tid = cf->tid + 1; |
0e2943878 ceph: unify cap f... |
2412 |
if (cf->caps) { |
0a454bdd5 ceph: reorganize ... |
2413 |
struct cap_msg_args arg; |
0e2943878 ceph: unify cap f... |
2414 2415 2416 |
dout("kick_flushing_caps %p cap %p tid %llu %s ", inode, cap, cf->tid, ceph_cap_string(cf->caps)); |
0a454bdd5 ceph: reorganize ... |
2417 |
__prep_cap(&arg, cap, CEPH_CAP_OP_FLUSH, |
49ada6e8d ceph: more precis... |
2418 2419 2420 |
(cf->tid < last_snap_flush ? CEPH_CLIENT_CAPS_PENDING_CAPSNAP : 0), __ceph_caps_used(ci), |
0e2943878 ceph: unify cap f... |
2421 |
__ceph_caps_wanted(ci), |
49ada6e8d ceph: more precis... |
2422 |
(cap->issued | cap->implemented), |
0e2943878 ceph: unify cap f... |
2423 |
cf->caps, cf->tid, oldest_flush_tid); |
0a454bdd5 ceph: reorganize ... |
2424 |
spin_unlock(&ci->i_ceph_lock); |
523119808 ceph: drop separa... |
2425 |
__send_cap(&arg, ci); |
0e2943878 ceph: unify cap f... |
2426 2427 2428 2429 2430 2431 2432 2433 |
} else { struct ceph_cap_snap *capsnap = container_of(cf, struct ceph_cap_snap, cap_flush); dout("kick_flushing_caps %p capsnap %p tid %llu %s ", inode, capsnap, cf->tid, ceph_cap_string(capsnap->dirty)); |
805692d0e ceph: convert cep... |
2434 |
refcount_inc(&capsnap->nref); |
0e2943878 ceph: unify cap f... |
2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 |
spin_unlock(&ci->i_ceph_lock); ret = __send_flush_snap(inode, session, capsnap, cap->mseq, oldest_flush_tid); if (ret < 0) { pr_err("kick_flushing_caps: error sending " "cap flushsnap, ino (%llx.%llx) " "tid %llu follows %llu ", ceph_vinop(inode), cf->tid, capsnap->follows); } ceph_put_cap_snap(capsnap); } |
e4500b5e3 ceph: use list in... |
2450 2451 |
spin_lock(&ci->i_ceph_lock); |
553adfd94 ceph: track pendi... |
2452 |
} |
553adfd94 ceph: track pendi... |
2453 |
} |
e548e9b93 ceph: re-send flu... |
2454 2455 2456 2457 2458 |
void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session) { struct ceph_inode_info *ci; struct ceph_cap *cap; |
0e2943878 ceph: unify cap f... |
2459 |
u64 oldest_flush_tid; |
e548e9b93 ceph: re-send flu... |
2460 2461 2462 |
dout("early_kick_flushing_caps mds%d ", session->s_mds); |
0e2943878 ceph: unify cap f... |
2463 2464 2465 2466 |
spin_lock(&mdsc->cap_dirty_lock); oldest_flush_tid = __get_oldest_flush_tid(mdsc); spin_unlock(&mdsc->cap_dirty_lock); |
e548e9b93 ceph: re-send flu... |
2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 |
list_for_each_entry(ci, &session->s_cap_flushing, i_flushing_item) { spin_lock(&ci->i_ceph_lock); cap = ci->i_auth_cap; if (!(cap && cap->session == session)) { pr_err("%p auth cap %p not mds%d ??? ", &ci->vfs_inode, cap, session->s_mds); spin_unlock(&ci->i_ceph_lock); continue; } /* * if flushing caps were revoked, we re-send the cap flush * in client reconnect stage. This guarantees MDS * processes * the cap flush message before issuing the flushing caps to * other client. */ if ((cap->issued & ci->i_flushing_caps) != ci->i_flushing_caps) { |
81c5a1487 ceph: split large... |
2487 2488 2489 2490 2491 2492 |
/* encode_caps_cb() also will reset these sequence * numbers. make sure sequence numbers in cap flush * message match later reconnect message */ cap->seq = 0; cap->issue_seq = 0; cap->mseq = 0; |
0e2943878 ceph: unify cap f... |
2493 2494 |
__kick_flushing_caps(mdsc, session, ci, oldest_flush_tid); |
13c2b57d8 ceph: avoid sendi... |
2495 2496 |
} else { ci->i_ceph_flags |= CEPH_I_KICK_FLUSH; |
e548e9b93 ceph: re-send flu... |
2497 |
} |
e548e9b93 ceph: re-send flu... |
2498 2499 2500 |
spin_unlock(&ci->i_ceph_lock); } } |
a8599bd82 ceph: capability ... |
2501 2502 2503 2504 |
void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session) { struct ceph_inode_info *ci; |
13c2b57d8 ceph: avoid sendi... |
2505 |
struct ceph_cap *cap; |
0e2943878 ceph: unify cap f... |
2506 |
u64 oldest_flush_tid; |
a8599bd82 ceph: capability ... |
2507 |
|
829ad4db9 ceph: ceph_kick_f... |
2508 |
lockdep_assert_held(&session->s_mutex); |
a8599bd82 ceph: capability ... |
2509 2510 |
dout("kick_flushing_caps mds%d ", session->s_mds); |
0e2943878 ceph: unify cap f... |
2511 2512 2513 2514 |
spin_lock(&mdsc->cap_dirty_lock); oldest_flush_tid = __get_oldest_flush_tid(mdsc); spin_unlock(&mdsc->cap_dirty_lock); |
a8599bd82 ceph: capability ... |
2515 |
list_for_each_entry(ci, &session->s_cap_flushing, i_flushing_item) { |
0e2943878 ceph: unify cap f... |
2516 |
spin_lock(&ci->i_ceph_lock); |
13c2b57d8 ceph: avoid sendi... |
2517 2518 2519 2520 2521 2522 2523 2524 2525 |
cap = ci->i_auth_cap; if (!(cap && cap->session == session)) { pr_err("%p auth cap %p not mds%d ??? ", &ci->vfs_inode, cap, session->s_mds); spin_unlock(&ci->i_ceph_lock); continue; } if (ci->i_ceph_flags & CEPH_I_KICK_FLUSH) { |
13c2b57d8 ceph: avoid sendi... |
2526 2527 2528 |
__kick_flushing_caps(mdsc, session, ci, oldest_flush_tid); } |
0e2943878 ceph: unify cap f... |
2529 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2530 2531 |
} } |
e8a4d2677 ceph: clean up ki... |
2532 2533 |
void ceph_kick_flushing_inode_caps(struct ceph_mds_session *session, struct ceph_inode_info *ci) |
088b3f5e9 ceph: fix flushin... |
2534 |
{ |
e8a4d2677 ceph: clean up ki... |
2535 2536 2537 2538 |
struct ceph_mds_client *mdsc = session->s_mdsc; struct ceph_cap *cap = ci->i_auth_cap; lockdep_assert_held(&ci->i_ceph_lock); |
088b3f5e9 ceph: fix flushin... |
2539 |
|
e8a4d2677 ceph: clean up ki... |
2540 2541 |
dout("%s %p flushing %s ", __func__, &ci->vfs_inode, |
8310b0891 ceph: track pendi... |
2542 |
ceph_cap_string(ci->i_flushing_caps)); |
005c46970 ceph: move inode ... |
2543 |
|
0e2943878 ceph: unify cap f... |
2544 2545 |
if (!list_empty(&ci->i_cap_flush_list)) { u64 oldest_flush_tid; |
005c46970 ceph: move inode ... |
2546 2547 2548 |
spin_lock(&mdsc->cap_dirty_lock); list_move_tail(&ci->i_flushing_item, &cap->session->s_cap_flushing); |
0e2943878 ceph: unify cap f... |
2549 |
oldest_flush_tid = __get_oldest_flush_tid(mdsc); |
005c46970 ceph: move inode ... |
2550 |
spin_unlock(&mdsc->cap_dirty_lock); |
0e2943878 ceph: unify cap f... |
2551 |
__kick_flushing_caps(mdsc, session, ci, oldest_flush_tid); |
088b3f5e9 ceph: fix flushin... |
2552 2553 |
} } |
a8599bd82 ceph: capability ... |
2554 2555 2556 2557 |
/* * Take references to capabilities we hold, so that we don't release * them to the MDS prematurely. |
a8599bd82 ceph: capability ... |
2558 |
*/ |
40dcf75e8 ceph: make __take... |
2559 |
void ceph_take_cap_refs(struct ceph_inode_info *ci, int got, |
5dda377cf ceph: set i_head_... |
2560 |
bool snap_rwsem_locked) |
a8599bd82 ceph: capability ... |
2561 |
{ |
40dcf75e8 ceph: make __take... |
2562 |
lockdep_assert_held(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2563 2564 2565 2566 2567 2568 |
if (got & CEPH_CAP_PIN) ci->i_pin_ref++; if (got & CEPH_CAP_FILE_RD) ci->i_rd_ref++; if (got & CEPH_CAP_FILE_CACHE) ci->i_rdcache_ref++; |
f85122afe ceph: add refcoun... |
2569 2570 |
if (got & CEPH_CAP_FILE_EXCL) ci->i_fx_ref++; |
5dda377cf ceph: set i_head_... |
2571 2572 2573 2574 2575 2576 |
if (got & CEPH_CAP_FILE_WR) { if (ci->i_wr_ref == 0 && !ci->i_head_snapc) { BUG_ON(!snap_rwsem_locked); ci->i_head_snapc = ceph_get_snap_context( ci->i_snap_realm->cached_context); } |
a8599bd82 ceph: capability ... |
2577 |
ci->i_wr_ref++; |
5dda377cf ceph: set i_head_... |
2578 |
} |
a8599bd82 ceph: capability ... |
2579 |
if (got & CEPH_CAP_FILE_BUFFER) { |
d3d0720d4 ceph: do not use ... |
2580 |
if (ci->i_wb_ref == 0) |
3772d26d8 ceph: use ihold()... |
2581 |
ihold(&ci->vfs_inode); |
d3d0720d4 ceph: do not use ... |
2582 |
ci->i_wb_ref++; |
40dcf75e8 ceph: make __take... |
2583 2584 |
dout("%s %p wb %d -> %d (?) ", __func__, |
d3d0720d4 ceph: do not use ... |
2585 |
&ci->vfs_inode, ci->i_wb_ref-1, ci->i_wb_ref); |
a8599bd82 ceph: capability ... |
2586 2587 2588 2589 2590 2591 2592 2593 2594 |
} } /* * Try to grab cap references. Specify those refs we @want, and the * minimal set we @need. Also include the larger offset we are writing * to (when applicable), and check against max_size here as well. * Note that caller is responsible for ensuring max_size increases are * requested from the MDS. |
1199d7da2 ceph: simplify ar... |
2595 |
* |
546d40208 ceph: cleanup ret... |
2596 2597 2598 2599 2600 |
* Returns 0 if caps were not able to be acquired (yet), 1 if succeed, * or a negative error code. There are 3 speical error codes: * -EAGAIN: need to sleep but non-blocking is specified * -EFBIG: ask caller to call check_max_size() and try again. * -ESTALE: ask caller to call ceph_renew_caps() and try again. |
a8599bd82 ceph: capability ... |
2601 |
*/ |
ff5d913df ceph: return -EIO... |
2602 |
enum { |
719a2514e ceph: consider in... |
2603 2604 2605 |
/* first 8 bits are reserved for CEPH_FILE_MODE_FOO */ NON_BLOCKING = (1 << 8), CHECK_FILELOCK = (1 << 9), |
ff5d913df ceph: return -EIO... |
2606 |
}; |
5e3ded1bb ceph: pass filp t... |
2607 |
static int try_get_cap_refs(struct inode *inode, int need, int want, |
ff5d913df ceph: return -EIO... |
2608 |
loff_t endoff, int flags, int *got) |
a8599bd82 ceph: capability ... |
2609 |
{ |
5e3ded1bb ceph: pass filp t... |
2610 |
struct ceph_inode_info *ci = ceph_inode(inode); |
5dda377cf ceph: set i_head_... |
2611 |
struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; |
a8599bd82 ceph: capability ... |
2612 |
int ret = 0; |
c4d4a582c ceph: avoid block... |
2613 |
int have, implemented; |
5dda377cf ceph: set i_head_... |
2614 |
bool snap_rwsem_locked = false; |
a8599bd82 ceph: capability ... |
2615 2616 2617 2618 |
dout("get_cap_refs %p need %s want %s ", inode, ceph_cap_string(need), ceph_cap_string(want)); |
c4d4a582c ceph: avoid block... |
2619 |
|
5dda377cf ceph: set i_head_... |
2620 |
again: |
be655596b ceph: use i_ceph_... |
2621 |
spin_lock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2622 |
|
ff5d913df ceph: return -EIO... |
2623 2624 2625 2626 2627 2628 2629 |
if ((flags & CHECK_FILELOCK) && (ci->i_ceph_flags & CEPH_I_ERROR_FILELOCK)) { dout("try_get_cap_refs %p error filelock ", inode); ret = -EIO; goto out_unlock; } |
37505d576 ceph: take i_mute... |
2630 2631 2632 |
/* finish pending truncate */ while (ci->i_truncate_pending) { spin_unlock(&ci->i_ceph_lock); |
5dda377cf ceph: set i_head_... |
2633 2634 2635 2636 |
if (snap_rwsem_locked) { up_read(&mdsc->snap_rwsem); snap_rwsem_locked = false; } |
b415bf4f9 ceph: fix pending... |
2637 |
__ceph_do_pending_vmtruncate(inode); |
37505d576 ceph: take i_mute... |
2638 2639 |
spin_lock(&ci->i_ceph_lock); } |
3871cbb9a ceph: fix request... |
2640 2641 2642 |
have = __ceph_caps_issued(ci, &implemented); if (have & need & CEPH_CAP_FILE_WR) { |
a8599bd82 ceph: capability ... |
2643 2644 2645 2646 |
if (endoff >= 0 && endoff > (loff_t)ci->i_max_size) { dout("get_cap_refs %p endoff %llu > maxsize %llu ", inode, endoff, ci->i_max_size); |
1199d7da2 ceph: simplify ar... |
2647 |
if (endoff > ci->i_requested_max_size) |
42d70f8e3 ceph: request new... |
2648 |
ret = ci->i_auth_cap ? -EFBIG : -ESTALE; |
3738daa68 ceph: fetch inlin... |
2649 |
goto out_unlock; |
a8599bd82 ceph: capability ... |
2650 2651 2652 2653 2654 2655 2656 2657 |
} /* * If a sync write is in progress, we must wait, so that we * can get a final snapshot value for size+mtime. */ if (__ceph_have_pending_cap_snap(ci)) { dout("get_cap_refs %p cap_snap_pending ", inode); |
3738daa68 ceph: fetch inlin... |
2658 |
goto out_unlock; |
a8599bd82 ceph: capability ... |
2659 2660 |
} } |
a8599bd82 ceph: capability ... |
2661 |
|
a8599bd82 ceph: capability ... |
2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 |
if ((have & need) == need) { /* * Look at (implemented & ~have & not) so that we keep waiting * on transition from wanted -> needed caps. This is needed * for WRBUFFER|WR -> WR to avoid a new WR sync write from * going before a prior buffered writeback happens. */ int not = want & ~(have & need); int revoking = implemented & ~have; dout("get_cap_refs %p have %s but not %s (revoking %s) ", inode, ceph_cap_string(have), ceph_cap_string(not), ceph_cap_string(revoking)); if ((revoking & not) == 0) { |
5dda377cf ceph: set i_head_... |
2676 2677 2678 2679 2680 2681 2682 2683 |
if (!snap_rwsem_locked && !ci->i_head_snapc && (need & CEPH_CAP_FILE_WR)) { if (!down_read_trylock(&mdsc->snap_rwsem)) { /* * we can not call down_read() when * task isn't in TASK_RUNNING state */ |
ff5d913df ceph: return -EIO... |
2684 |
if (flags & NON_BLOCKING) { |
1199d7da2 ceph: simplify ar... |
2685 |
ret = -EAGAIN; |
5dda377cf ceph: set i_head_... |
2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 |
goto out_unlock; } spin_unlock(&ci->i_ceph_lock); down_read(&mdsc->snap_rwsem); snap_rwsem_locked = true; goto again; } snap_rwsem_locked = true; } |
173e70e8a ceph: don't take ... |
2696 2697 2698 2699 |
if ((have & want) == want) *got = need | want; else *got = need; |
525d15e8e ceph: check inode... |
2700 2701 |
if (S_ISREG(inode->i_mode) && (need & CEPH_CAP_FILE_RD) && |
f7f7e7a06 ceph: improve fsc... |
2702 2703 |
!(*got & CEPH_CAP_FILE_CACHE)) ceph_disable_fscache_readpage(ci); |
40dcf75e8 ceph: make __take... |
2704 |
ceph_take_cap_refs(ci, *got, true); |
a8599bd82 ceph: capability ... |
2705 2706 2707 |
ret = 1; } } else { |
03f4fcb02 ceph: handle SESS... |
2708 |
int session_readonly = false; |
c0e385b10 ceph: always rene... |
2709 |
int mds_wanted; |
525d15e8e ceph: check inode... |
2710 2711 |
if (ci->i_auth_cap && (need & (CEPH_CAP_FILE_WR | CEPH_CAP_FILE_EXCL))) { |
03f4fcb02 ceph: handle SESS... |
2712 2713 2714 2715 2716 2717 |
struct ceph_mds_session *s = ci->i_auth_cap->session; spin_lock(&s->s_cap_lock); session_readonly = s->s_readonly; spin_unlock(&s->s_cap_lock); } if (session_readonly) { |
c0e385b10 ceph: always rene... |
2718 2719 |
dout("get_cap_refs %p need %s but mds%d readonly ", |
03f4fcb02 ceph: handle SESS... |
2720 |
inode, ceph_cap_string(need), ci->i_auth_cap->mds); |
1199d7da2 ceph: simplify ar... |
2721 |
ret = -EROFS; |
03f4fcb02 ceph: handle SESS... |
2722 2723 |
goto out_unlock; } |
c0e385b10 ceph: always rene... |
2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 |
if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) { dout("get_cap_refs %p forced umount ", inode); ret = -EIO; goto out_unlock; } mds_wanted = __ceph_caps_mds_wanted(ci, false); if (need & ~mds_wanted) { dout("get_cap_refs %p need %s > mds_wanted %s ", inode, ceph_cap_string(need), ceph_cap_string(mds_wanted)); ret = -ESTALE; goto out_unlock; |
48fec5d0a ceph: EIO all ope... |
2738 |
} |
c0e385b10 ceph: always rene... |
2739 2740 |
dout("get_cap_refs %p have %s need %s ", inode, |
a8599bd82 ceph: capability ... |
2741 2742 |
ceph_cap_string(have), ceph_cap_string(need)); } |
3738daa68 ceph: fetch inlin... |
2743 |
out_unlock: |
719a2514e ceph: consider in... |
2744 2745 |
__ceph_touch_fmode(ci, mdsc, flags); |
be655596b ceph: use i_ceph_... |
2746 |
spin_unlock(&ci->i_ceph_lock); |
5dda377cf ceph: set i_head_... |
2747 2748 |
if (snap_rwsem_locked) up_read(&mdsc->snap_rwsem); |
3738daa68 ceph: fetch inlin... |
2749 |
|
1af16d547 ceph: add caps pe... |
2750 2751 2752 2753 |
if (!ret) ceph_update_cap_mis(&mdsc->metric); else if (ret == 1) ceph_update_cap_hit(&mdsc->metric); |
a8599bd82 ceph: capability ... |
2754 2755 |
dout("get_cap_refs %p ret %d got %s ", inode, |
c4d4a582c ceph: avoid block... |
2756 |
ret, ceph_cap_string(*got)); |
a8599bd82 ceph: capability ... |
2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 |
return ret; } /* * Check the offset we are writing up to against our current * max_size. If necessary, tell the MDS we want to write to * a larger offset. */ static void check_max_size(struct inode *inode, loff_t endoff) { struct ceph_inode_info *ci = ceph_inode(inode); int check = 0; /* do we need to explicitly request a larger max_size? */ |
be655596b ceph: use i_ceph_... |
2771 |
spin_lock(&ci->i_ceph_lock); |
3871cbb9a ceph: fix request... |
2772 |
if (endoff >= ci->i_max_size && endoff > ci->i_wanted_max_size) { |
a8599bd82 ceph: capability ... |
2773 2774 2775 2776 |
dout("write %p at large endoff %llu, req max_size ", inode, endoff); ci->i_wanted_max_size = endoff; |
a8599bd82 ceph: capability ... |
2777 |
} |
3871cbb9a ceph: fix request... |
2778 2779 2780 2781 2782 2783 |
/* duplicate ceph_check_caps()'s logic */ if (ci->i_auth_cap && (ci->i_auth_cap->issued & CEPH_CAP_FILE_WR) && ci->i_wanted_max_size > ci->i_max_size && ci->i_wanted_max_size > ci->i_requested_max_size) check = 1; |
be655596b ceph: use i_ceph_... |
2784 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2785 2786 2787 |
if (check) ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL); } |
719a2514e ceph: consider in... |
2788 2789 2790 2791 2792 2793 2794 2795 2796 |
static inline int get_used_fmode(int caps) { int fmode = 0; if (caps & CEPH_CAP_FILE_RD) fmode |= CEPH_FILE_MODE_RD; if (caps & CEPH_CAP_FILE_WR) fmode |= CEPH_FILE_MODE_WR; return fmode; } |
5e3ded1bb ceph: pass filp t... |
2797 |
int ceph_try_get_caps(struct inode *inode, int need, int want, |
2ee9dd958 ceph: add non-blo... |
2798 |
bool nonblock, int *got) |
2b1ac852e ceph: try getting... |
2799 |
{ |
719a2514e ceph: consider in... |
2800 |
int ret, flags; |
2b1ac852e ceph: try getting... |
2801 2802 |
BUG_ON(need & ~CEPH_CAP_FILE_RD); |
a25949b99 ceph: cap trackin... |
2803 2804 2805 2806 2807 2808 2809 2810 |
BUG_ON(want & ~(CEPH_CAP_FILE_CACHE | CEPH_CAP_FILE_LAZYIO | CEPH_CAP_FILE_SHARED | CEPH_CAP_FILE_EXCL | CEPH_CAP_ANY_DIR_OPS)); if (need) { ret = ceph_pool_perm_check(inode, need); if (ret < 0) return ret; } |
2b1ac852e ceph: try getting... |
2811 |
|
719a2514e ceph: consider in... |
2812 2813 2814 2815 2816 |
flags = get_used_fmode(need | want); if (nonblock) flags |= NON_BLOCKING; ret = try_get_cap_refs(inode, need, want, 0, flags, got); |
546d40208 ceph: cleanup ret... |
2817 |
/* three special error codes */ |
7d8976afa ceph: fix special... |
2818 |
if (ret == -EAGAIN || ret == -EFBIG || ret == -ESTALE) |
546d40208 ceph: cleanup ret... |
2819 2820 |
ret = 0; return ret; |
2b1ac852e ceph: try getting... |
2821 |
} |
a8599bd82 ceph: capability ... |
2822 2823 2824 2825 2826 |
/* * Wait for caps, and take cap references. If we can't get a WR cap * due to a small max_size, make sure we check_max_size (and possibly * ask the mds) so we don't get hung up indefinitely. */ |
5e3ded1bb ceph: pass filp t... |
2827 |
int ceph_get_caps(struct file *filp, int need, int want, |
3738daa68 ceph: fetch inlin... |
2828 |
loff_t endoff, int *got, struct page **pinned_page) |
a8599bd82 ceph: capability ... |
2829 |
{ |
ff5d913df ceph: return -EIO... |
2830 |
struct ceph_file_info *fi = filp->private_data; |
5e3ded1bb ceph: pass filp t... |
2831 2832 |
struct inode *inode = file_inode(filp); struct ceph_inode_info *ci = ceph_inode(inode); |
81f148a91 ceph: invalidate ... |
2833 |
struct ceph_fs_client *fsc = ceph_inode_to_client(inode); |
ff5d913df ceph: return -EIO... |
2834 |
int ret, _got, flags; |
a8599bd82 ceph: capability ... |
2835 |
|
5e3ded1bb ceph: pass filp t... |
2836 |
ret = ceph_pool_perm_check(inode, need); |
10183a695 ceph: check OSD c... |
2837 2838 |
if (ret < 0) return ret; |
81f148a91 ceph: invalidate ... |
2839 2840 2841 |
if ((fi->fmode & CEPH_FILE_MODE_WR) && fi->filp_gen != READ_ONCE(fsc->filp_gen)) return -EBADF; |
719a2514e ceph: consider in... |
2842 |
flags = get_used_fmode(need | want); |
5dda377cf ceph: set i_head_... |
2843 |
while (true) { |
719a2514e ceph: consider in... |
2844 2845 2846 |
flags &= CEPH_FILE_MODE_MASK; if (atomic_read(&fi->num_locks)) flags |= CHECK_FILELOCK; |
5dda377cf ceph: set i_head_... |
2847 |
_got = 0; |
5e3ded1bb ceph: pass filp t... |
2848 |
ret = try_get_cap_refs(inode, need, want, endoff, |
ff5d913df ceph: return -EIO... |
2849 |
flags, &_got); |
546d40208 ceph: cleanup ret... |
2850 |
WARN_ON_ONCE(ret == -EAGAIN); |
7b2f936fc ceph: fix error h... |
2851 |
if (!ret) { |
3a3430aff ceph: show tasks ... |
2852 2853 |
struct ceph_mds_client *mdsc = fsc->mdsc; struct cap_wait cw; |
5c341ee32 ceph: fix schedul... |
2854 |
DEFINE_WAIT_FUNC(wait, woken_wake_function); |
3a3430aff ceph: show tasks ... |
2855 |
|
ebce3eb2f ceph: fix inode n... |
2856 |
cw.ino = ceph_ino(inode); |
3a3430aff ceph: show tasks ... |
2857 2858 2859 2860 2861 2862 2863 |
cw.tgid = current->tgid; cw.need = need; cw.want = want; spin_lock(&mdsc->caps_list_lock); list_add(&cw.list, &mdsc->cap_wait_list); spin_unlock(&mdsc->caps_list_lock); |
719a2514e ceph: consider in... |
2864 2865 |
/* make sure used fmode not timeout */ ceph_get_fmode(ci, flags, FMODE_WAIT_BIAS); |
5c341ee32 ceph: fix schedul... |
2866 |
add_wait_queue(&ci->i_cap_wq, &wait); |
ff5d913df ceph: return -EIO... |
2867 |
flags |= NON_BLOCKING; |
5e3ded1bb ceph: pass filp t... |
2868 |
while (!(ret = try_get_cap_refs(inode, need, want, |
ff5d913df ceph: return -EIO... |
2869 |
endoff, flags, &_got))) { |
6e09d0fb6 ceph: fix ceph_ge... |
2870 2871 2872 2873 |
if (signal_pending(current)) { ret = -ERESTARTSYS; break; } |
5c341ee32 ceph: fix schedul... |
2874 |
wait_woken(&wait, TASK_INTERRUPTIBLE, MAX_SCHEDULE_TIMEOUT); |
6e09d0fb6 ceph: fix ceph_ge... |
2875 |
} |
5c341ee32 ceph: fix schedul... |
2876 2877 |
remove_wait_queue(&ci->i_cap_wq, &wait); |
719a2514e ceph: consider in... |
2878 |
ceph_put_fmode(ci, flags, FMODE_WAIT_BIAS); |
3a3430aff ceph: show tasks ... |
2879 2880 2881 2882 |
spin_lock(&mdsc->caps_list_lock); list_del(&cw.list); spin_unlock(&mdsc->caps_list_lock); |
7b2f936fc ceph: fix error h... |
2883 |
if (ret == -EAGAIN) |
5dda377cf ceph: set i_head_... |
2884 |
continue; |
77310320c ceph: renew caps ... |
2885 |
} |
81f148a91 ceph: invalidate ... |
2886 2887 2888 2889 2890 2891 2892 |
if ((fi->fmode & CEPH_FILE_MODE_WR) && fi->filp_gen != READ_ONCE(fsc->filp_gen)) { if (ret >= 0 && _got) ceph_put_cap_refs(ci, _got); return -EBADF; } |
7b2f936fc ceph: fix error h... |
2893 |
if (ret < 0) { |
9bccb7657 ceph: wait for as... |
2894 2895 2896 2897 2898 |
if (ret == -EFBIG || ret == -ESTALE) { int ret2 = ceph_wait_on_async_create(inode); if (ret2 < 0) return ret2; } |
546d40208 ceph: cleanup ret... |
2899 2900 2901 2902 |
if (ret == -EFBIG) { check_max_size(inode, endoff); continue; } |
7b2f936fc ceph: fix error h... |
2903 2904 |
if (ret == -ESTALE) { /* session was killed, try renew caps */ |
719a2514e ceph: consider in... |
2905 |
ret = ceph_renew_caps(inode, flags); |
7b2f936fc ceph: fix error h... |
2906 2907 2908 |
if (ret == 0) continue; } |
77310320c ceph: renew caps ... |
2909 |
return ret; |
5dda377cf ceph: set i_head_... |
2910 |
} |
c4d4a582c ceph: avoid block... |
2911 |
|
525d15e8e ceph: check inode... |
2912 2913 |
if (S_ISREG(ci->vfs_inode.i_mode) && ci->i_inline_version != CEPH_INLINE_NONE && |
5dda377cf ceph: set i_head_... |
2914 |
(_got & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)) && |
5e3ded1bb ceph: pass filp t... |
2915 |
i_size_read(inode) > 0) { |
5dda377cf ceph: set i_head_... |
2916 |
struct page *page = |
5e3ded1bb ceph: pass filp t... |
2917 |
find_get_page(inode->i_mapping, 0); |
5dda377cf ceph: set i_head_... |
2918 2919 2920 2921 2922 |
if (page) { if (PageUptodate(page)) { *pinned_page = page; break; } |
09cbfeaf1 mm, fs: get rid o... |
2923 |
put_page(page); |
c4d4a582c ceph: avoid block... |
2924 |
} |
5dda377cf ceph: set i_head_... |
2925 2926 2927 2928 2929 2930 |
/* * drop cap refs first because getattr while * holding * caps refs can cause deadlock. */ ceph_put_cap_refs(ci, _got); _got = 0; |
c4d4a582c ceph: avoid block... |
2931 |
|
5dda377cf ceph: set i_head_... |
2932 2933 2934 2935 |
/* * getattr request will bring inline data into * page cache */ |
5e3ded1bb ceph: pass filp t... |
2936 |
ret = __ceph_do_getattr(inode, NULL, |
5dda377cf ceph: set i_head_... |
2937 2938 2939 2940 2941 2942 2943 |
CEPH_STAT_CAP_INLINE_DATA, true); if (ret < 0) return ret; continue; } break; |
c4d4a582c ceph: avoid block... |
2944 |
} |
5dda377cf ceph: set i_head_... |
2945 |
|
525d15e8e ceph: check inode... |
2946 2947 |
if (S_ISREG(ci->vfs_inode.i_mode) && (_got & CEPH_CAP_FILE_RD) && (_got & CEPH_CAP_FILE_CACHE)) |
f7f7e7a06 ceph: improve fsc... |
2948 |
ceph_fscache_revalidate_cookie(ci); |
c4d4a582c ceph: avoid block... |
2949 2950 |
*got = _got; return 0; |
a8599bd82 ceph: capability ... |
2951 2952 2953 2954 2955 2956 2957 2958 |
} /* * Take cap refs. Caller must already know we hold at least one ref * on the caps in question or we don't know this is safe. */ void ceph_get_cap_refs(struct ceph_inode_info *ci, int caps) { |
be655596b ceph: use i_ceph_... |
2959 |
spin_lock(&ci->i_ceph_lock); |
40dcf75e8 ceph: make __take... |
2960 |
ceph_take_cap_refs(ci, caps, false); |
be655596b ceph: use i_ceph_... |
2961 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
2962 |
} |
860560904 ceph: avoid sendi... |
2963 2964 2965 2966 2967 |
/* * drop cap_snap that is not associated with any snapshot. * we don't need to send FLUSHSNAP message for it. */ |
70220ac8c ceph: introduce a... |
2968 2969 |
static int ceph_try_drop_cap_snap(struct ceph_inode_info *ci, struct ceph_cap_snap *capsnap) |
860560904 ceph: avoid sendi... |
2970 2971 2972 |
{ if (!capsnap->need_flush && !capsnap->writing && !capsnap->dirty_pages) { |
860560904 ceph: avoid sendi... |
2973 2974 2975 |
dout("dropping cap_snap %p follows %llu ", capsnap, capsnap->follows); |
0e2943878 ceph: unify cap f... |
2976 |
BUG_ON(capsnap->cap_flush.tid > 0); |
860560904 ceph: avoid sendi... |
2977 |
ceph_put_snap_context(capsnap->context); |
70220ac8c ceph: introduce a... |
2978 2979 |
if (!list_is_last(&capsnap->ci_item, &ci->i_cap_snaps)) ci->i_ceph_flags |= CEPH_I_FLUSH_SNAPS; |
860560904 ceph: avoid sendi... |
2980 |
list_del(&capsnap->ci_item); |
860560904 ceph: avoid sendi... |
2981 2982 2983 2984 2985 |
ceph_put_cap_snap(capsnap); return 1; } return 0; } |
a8599bd82 ceph: capability ... |
2986 2987 2988 2989 2990 2991 2992 2993 2994 |
/* * Release cap refs. * * If we released the last ref on any given cap, call ceph_check_caps * to release (or schedule a release). * * If we are releasing a WR cap (from a sync write), finalize any affected * cap_snap, and wake up any waiters. */ |
e64f44a88 ceph: skip checki... |
2995 2996 |
static void __ceph_put_cap_refs(struct ceph_inode_info *ci, int had, bool skip_checking_caps) |
a8599bd82 ceph: capability ... |
2997 2998 2999 |
{ struct inode *inode = &ci->vfs_inode; int last = 0, put = 0, flushsnaps = 0, wake = 0; |
a8599bd82 ceph: capability ... |
3000 |
|
be655596b ceph: use i_ceph_... |
3001 |
spin_lock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
3002 3003 3004 3005 3006 3007 3008 3009 |
if (had & CEPH_CAP_PIN) --ci->i_pin_ref; if (had & CEPH_CAP_FILE_RD) if (--ci->i_rd_ref == 0) last++; if (had & CEPH_CAP_FILE_CACHE) if (--ci->i_rdcache_ref == 0) last++; |
f85122afe ceph: add refcoun... |
3010 3011 3012 |
if (had & CEPH_CAP_FILE_EXCL) if (--ci->i_fx_ref == 0) last++; |
a8599bd82 ceph: capability ... |
3013 |
if (had & CEPH_CAP_FILE_BUFFER) { |
d3d0720d4 ceph: do not use ... |
3014 |
if (--ci->i_wb_ref == 0) { |
a8599bd82 ceph: capability ... |
3015 3016 3017 |
last++; put++; } |
d3d0720d4 ceph: do not use ... |
3018 3019 3020 |
dout("put_cap_refs %p wb %d -> %d (?) ", inode, ci->i_wb_ref+1, ci->i_wb_ref); |
a8599bd82 ceph: capability ... |
3021 3022 3023 3024 |
} if (had & CEPH_CAP_FILE_WR) if (--ci->i_wr_ref == 0) { last++; |
860560904 ceph: avoid sendi... |
3025 3026 3027 3028 3029 3030 |
if (__ceph_have_pending_cap_snap(ci)) { struct ceph_cap_snap *capsnap = list_last_entry(&ci->i_cap_snaps, struct ceph_cap_snap, ci_item); capsnap->writing = 0; |
70220ac8c ceph: introduce a... |
3031 |
if (ceph_try_drop_cap_snap(ci, capsnap)) |
860560904 ceph: avoid sendi... |
3032 3033 3034 3035 |
put++; else if (__ceph_finish_cap_snap(ci, capsnap)) flushsnaps = 1; wake = 1; |
a8599bd82 ceph: capability ... |
3036 |
} |
5dda377cf ceph: set i_head_... |
3037 3038 3039 3040 3041 3042 3043 |
if (ci->i_wrbuffer_ref_head == 0 && ci->i_dirty_caps == 0 && ci->i_flushing_caps == 0) { BUG_ON(!ci->i_head_snapc); ceph_put_snap_context(ci->i_head_snapc); ci->i_head_snapc = NULL; } |
db40cc170 ceph: keep i_snap... |
3044 |
/* see comment in __ceph_remove_cap() */ |
bd84fbcb3 ceph: switch to g... |
3045 |
if (!__ceph_is_any_real_caps(ci) && ci->i_snap_realm) |
db40cc170 ceph: keep i_snap... |
3046 |
drop_inode_snap_realm(ci); |
a8599bd82 ceph: capability ... |
3047 |
} |
be655596b ceph: use i_ceph_... |
3048 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
3049 |
|
819ccbfa4 ceph: fix leaked ... |
3050 3051 3052 |
dout("put_cap_refs %p had %s%s%s ", inode, ceph_cap_string(had), last ? " last" : "", put ? " put" : ""); |
a8599bd82 ceph: capability ... |
3053 |
|
e64f44a88 ceph: skip checki... |
3054 |
if (last && !skip_checking_caps) |
a8599bd82 ceph: capability ... |
3055 3056 |
ceph_check_caps(ci, 0, NULL); else if (flushsnaps) |
ed9b430c9 ceph: cleanup cep... |
3057 |
ceph_flush_snaps(ci, NULL); |
a8599bd82 ceph: capability ... |
3058 |
if (wake) |
03066f234 ceph: use complet... |
3059 |
wake_up_all(&ci->i_cap_wq); |
860560904 ceph: avoid sendi... |
3060 |
while (put-- > 0) |
a8599bd82 ceph: capability ... |
3061 3062 |
iput(inode); } |
e64f44a88 ceph: skip checki... |
3063 3064 3065 3066 3067 3068 3069 3070 3071 |
void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) { __ceph_put_cap_refs(ci, had, false); } void ceph_put_cap_refs_no_check_caps(struct ceph_inode_info *ci, int had) { __ceph_put_cap_refs(ci, had, true); } |
a8599bd82 ceph: capability ... |
3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 |
/* * Release @nr WRBUFFER refs on dirty pages for the given @snapc snap * context. Adjust per-snap dirty page accounting as appropriate. * Once all dirty data for a cap_snap is flushed, flush snapped file * metadata back to the MDS. If we dropped the last ref, call * ceph_check_caps. */ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, struct ceph_snap_context *snapc) { struct inode *inode = &ci->vfs_inode; |
a8599bd82 ceph: capability ... |
3083 |
struct ceph_cap_snap *capsnap = NULL; |
70220ac8c ceph: introduce a... |
3084 3085 3086 3087 3088 |
int put = 0; bool last = false; bool found = false; bool flush_snaps = false; bool complete_capsnap = false; |
a8599bd82 ceph: capability ... |
3089 |
|
be655596b ceph: use i_ceph_... |
3090 |
spin_lock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
3091 |
ci->i_wrbuffer_ref -= nr; |
70220ac8c ceph: introduce a... |
3092 3093 3094 3095 |
if (ci->i_wrbuffer_ref == 0) { last = true; put++; } |
a8599bd82 ceph: capability ... |
3096 3097 3098 |
if (ci->i_head_snapc == snapc) { ci->i_wrbuffer_ref_head -= nr; |
7d8cb26d7 ceph: maintain i_... |
3099 |
if (ci->i_wrbuffer_ref_head == 0 && |
5dda377cf ceph: set i_head_... |
3100 3101 3102 |
ci->i_wr_ref == 0 && ci->i_dirty_caps == 0 && ci->i_flushing_caps == 0) { |
7d8cb26d7 ceph: maintain i_... |
3103 |
BUG_ON(!ci->i_head_snapc); |
a8599bd82 ceph: capability ... |
3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 |
ceph_put_snap_context(ci->i_head_snapc); ci->i_head_snapc = NULL; } dout("put_wrbuffer_cap_refs on %p head %d/%d -> %d/%d %s ", inode, ci->i_wrbuffer_ref+nr, ci->i_wrbuffer_ref_head+nr, ci->i_wrbuffer_ref, ci->i_wrbuffer_ref_head, last ? " LAST" : ""); } else { list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { if (capsnap->context == snapc) { |
70220ac8c ceph: introduce a... |
3116 |
found = true; |
a8599bd82 ceph: capability ... |
3117 3118 3119 3120 |
break; } } BUG_ON(!found); |
819ccbfa4 ceph: fix leaked ... |
3121 3122 |
capsnap->dirty_pages -= nr; if (capsnap->dirty_pages == 0) { |
70220ac8c ceph: introduce a... |
3123 3124 3125 3126 3127 3128 3129 3130 3131 |
complete_capsnap = true; if (!capsnap->writing) { if (ceph_try_drop_cap_snap(ci, capsnap)) { put++; } else { ci->i_ceph_flags |= CEPH_I_FLUSH_SNAPS; flush_snaps = true; } } |
819ccbfa4 ceph: fix leaked ... |
3132 |
} |
a8599bd82 ceph: capability ... |
3133 |
dout("put_wrbuffer_cap_refs on %p cap_snap %p " |
860560904 ceph: avoid sendi... |
3134 3135 |
" snap %lld %d/%d -> %d/%d %s%s ", |
a8599bd82 ceph: capability ... |
3136 3137 3138 3139 |
inode, capsnap, capsnap->context->seq, ci->i_wrbuffer_ref+nr, capsnap->dirty_pages + nr, ci->i_wrbuffer_ref, capsnap->dirty_pages, last ? " (wrbuffer last)" : "", |
860560904 ceph: avoid sendi... |
3140 |
complete_capsnap ? " (complete capsnap)" : ""); |
a8599bd82 ceph: capability ... |
3141 |
} |
be655596b ceph: use i_ceph_... |
3142 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
3143 3144 |
if (last) { |
bf73c62e7 ceph: check all m... |
3145 |
ceph_check_caps(ci, 0, NULL); |
70220ac8c ceph: introduce a... |
3146 |
} else if (flush_snaps) { |
ed9b430c9 ceph: cleanup cep... |
3147 |
ceph_flush_snaps(ci, NULL); |
a8599bd82 ceph: capability ... |
3148 |
} |
70220ac8c ceph: introduce a... |
3149 3150 |
if (complete_capsnap) wake_up_all(&ci->i_cap_wq); |
3e1d0452e ceph: avoid iput_... |
3151 3152 3153 3154 |
while (put-- > 0) { /* avoid calling iput_final() in osd dispatch threads */ ceph_async_iput(inode); } |
a8599bd82 ceph: capability ... |
3155 3156 3157 |
} /* |
ca20c9919 ceph: trim delete... |
3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 |
* Invalidate unlinked inode's aliases, so we can drop the inode ASAP. */ static void invalidate_aliases(struct inode *inode) { struct dentry *dn, *prev = NULL; dout("invalidate_aliases inode %p ", inode); d_prune_aliases(inode); /* * For non-directory inode, d_find_alias() only returns |
fc12c80aa ceph: trivial com... |
3169 3170 |
* hashed dentry. After calling d_invalidate(), the * dentry becomes unhashed. |
ca20c9919 ceph: trim delete... |
3171 |
* |
a8d436f01 ceph: use d_inval... |
3172 |
* For directory inode, d_find_alias() can return |
fc12c80aa ceph: trivial com... |
3173 |
* unhashed dentry. But directory inode should have |
ca20c9919 ceph: trim delete... |
3174 3175 3176 3177 3178 3179 3180 |
* one alias at most. */ while ((dn = d_find_alias(inode))) { if (dn == prev) { dput(dn); break; } |
a8d436f01 ceph: use d_inval... |
3181 |
d_invalidate(dn); |
ca20c9919 ceph: trim delete... |
3182 3183 3184 3185 3186 3187 3188 |
if (prev) dput(prev); prev = dn; } if (prev) dput(prev); } |
a1c6b8358 ceph: define argu... |
3189 3190 3191 3192 3193 3194 |
struct cap_extra_info { struct ceph_string *pool_ns; /* inline data */ u64 inline_version; void *inline_data; u32 inline_len; |
4985d6f9e ceph: handle the ... |
3195 3196 3197 3198 |
/* dirstat */ bool dirstat_valid; u64 nfiles; u64 nsubdirs; |
176c77c9c ceph: handle chan... |
3199 |
u64 change_attr; |
a1c6b8358 ceph: define argu... |
3200 3201 |
/* currently issued */ int issued; |
ec62b894d ceph: handle btim... |
3202 |
struct timespec64 btime; |
a1c6b8358 ceph: define argu... |
3203 |
}; |
ca20c9919 ceph: trim delete... |
3204 |
/* |
a8599bd82 ceph: capability ... |
3205 3206 3207 |
* Handle a cap GRANT message from the MDS. (Note that a GRANT may * actually be a revocation if it specifies a smaller cap set.) * |
be655596b ceph: use i_ceph_... |
3208 |
* caller holds s_mutex and i_ceph_lock, we drop both. |
a8599bd82 ceph: capability ... |
3209 |
*/ |
a1c6b8358 ceph: define argu... |
3210 |
static void handle_cap_grant(struct inode *inode, |
15637c8b1 ceph: clean up ha... |
3211 |
struct ceph_mds_session *session, |
a1c6b8358 ceph: define argu... |
3212 3213 3214 3215 |
struct ceph_cap *cap, struct ceph_mds_caps *grant, struct ceph_buffer *xattr_buf, struct cap_extra_info *extra_info) |
2cd698be9 ceph: handle cap ... |
3216 |
__releases(ci->i_ceph_lock) |
a1c6b8358 ceph: define argu... |
3217 |
__releases(session->s_mdsc->snap_rwsem) |
a8599bd82 ceph: capability ... |
3218 3219 |
{ struct ceph_inode_info *ci = ceph_inode(inode); |
2f56f56ad Revert "ceph: upd... |
3220 |
int seq = le32_to_cpu(grant->seq); |
a8599bd82 ceph: capability ... |
3221 |
int newcaps = le32_to_cpu(grant->caps); |
2cd698be9 ceph: handle cap ... |
3222 |
int used, wanted, dirty; |
a8599bd82 ceph: capability ... |
3223 3224 |
u64 size = le64_to_cpu(grant->size); u64 max_size = le64_to_cpu(grant->max_size); |
fdac94fab ceph: skip updati... |
3225 3226 |
unsigned char check_caps = 0; bool was_stale = cap->cap_gen < session->s_cap_gen; |
ab6c2c3eb ceph: fix bool as... |
3227 3228 3229 3230 |
bool wake = false; bool writeback = false; bool queue_trunc = false; bool queue_invalidate = false; |
ab6c2c3eb ceph: fix bool as... |
3231 |
bool deleted_inode = false; |
31c542a19 ceph: add inline ... |
3232 |
bool fill_inline = false; |
a8599bd82 ceph: capability ... |
3233 |
|
2f56f56ad Revert "ceph: upd... |
3234 3235 |
dout("handle_cap_grant inode %p cap %p mds%d seq %d %s ", |
a1c6b8358 ceph: define argu... |
3236 |
inode, cap, session->s_mds, seq, ceph_cap_string(newcaps)); |
a8599bd82 ceph: capability ... |
3237 3238 3239 |
dout(" size %llu max_size %llu, i_size %llu ", size, max_size, inode->i_size); |
11df2dfb6 ceph: add importe... |
3240 3241 |
/* |
a8599bd82 ceph: capability ... |
3242 3243 3244 3245 |
* If CACHE is being revoked, and we have no dirty buffers, * try to invalidate (once). (If there are dirty buffers, we * will invalidate _after_ writeback.) */ |
525d15e8e ceph: check inode... |
3246 |
if (S_ISREG(inode->i_mode) && /* don't invalidate readdir cache */ |
fdd4e1583 ceph: rework dcac... |
3247 |
((cap->issued & ~newcaps) & CEPH_CAP_FILE_CACHE) && |
3b454c494 ceph: simplify ca... |
3248 |
(newcaps & CEPH_CAP_FILE_LAZYIO) == 0 && |
9abd4db71 ceph: don't use t... |
3249 |
!(ci->i_wrbuffer_ref || ci->i_wb_ref)) { |
e90757432 ceph: remove usel... |
3250 |
if (try_nonblocking_invalidate(inode)) { |
a8599bd82 ceph: capability ... |
3251 3252 3253 |
/* there were locked pages.. invalidate later in a separate thread. */ if (ci->i_rdcache_revoking != ci->i_rdcache_gen) { |
ab6c2c3eb ceph: fix bool as... |
3254 |
queue_invalidate = true; |
a8599bd82 ceph: capability ... |
3255 3256 |
ci->i_rdcache_revoking = ci->i_rdcache_gen; } |
a8599bd82 ceph: capability ... |
3257 |
} |
a8599bd82 ceph: capability ... |
3258 |
} |
d2f8bb27c ceph: update want... |
3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 |
if (was_stale) cap->issued = cap->implemented = CEPH_CAP_PIN; /* * auth mds of the inode changed. we received the cap export message, * but still haven't received the cap import message. handle_cap_export * updated the new auth MDS' cap. * * "ceph_seq_cmp(seq, cap->seq) <= 0" means we are processing a message * that was sent before the cap import message. So don't remove caps. */ if (ceph_seq_cmp(seq, cap->seq) <= 0) { WARN_ON(cap != ci->i_auth_cap); WARN_ON(cap->cap_id != le64_to_cpu(grant->cap_id)); seq = cap->seq; newcaps |= cap->issued; } |
a8599bd82 ceph: capability ... |
3276 |
/* side effects now are allowed */ |
685f9a5d1 ceph: do not conf... |
3277 |
cap->cap_gen = session->s_cap_gen; |
11df2dfb6 ceph: add importe... |
3278 |
cap->seq = seq; |
a8599bd82 ceph: capability ... |
3279 3280 |
__check_cap_issue(ci, cap, newcaps); |
176c77c9c ceph: handle chan... |
3281 |
inode_set_max_iversion_raw(inode, extra_info->change_attr); |
f98a128a5 ceph: update inod... |
3282 |
if ((newcaps & CEPH_CAP_AUTH_SHARED) && |
a1c6b8358 ceph: define argu... |
3283 |
(extra_info->issued & CEPH_CAP_AUTH_EXCL) == 0) { |
a8599bd82 ceph: capability ... |
3284 |
inode->i_mode = le32_to_cpu(grant->mode); |
05cb11c17 ceph: Translate b... |
3285 3286 |
inode->i_uid = make_kuid(&init_user_ns, le32_to_cpu(grant->uid)); inode->i_gid = make_kgid(&init_user_ns, le32_to_cpu(grant->gid)); |
ec62b894d ceph: handle btim... |
3287 |
ci->i_btime = extra_info->btime; |
a8599bd82 ceph: capability ... |
3288 3289 |
dout("%p mode 0%o uid.gid %d.%d ", inode, inode->i_mode, |
bd2bae6a6 ceph: Convert kui... |
3290 3291 |
from_kuid(&init_user_ns, inode->i_uid), from_kgid(&init_user_ns, inode->i_gid)); |
a8599bd82 ceph: capability ... |
3292 |
} |
fa466743a ceph: fix wrong c... |
3293 |
if ((newcaps & CEPH_CAP_LINK_SHARED) && |
a1c6b8358 ceph: define argu... |
3294 |
(extra_info->issued & CEPH_CAP_LINK_EXCL) == 0) { |
bfe868486 filesystems: add ... |
3295 |
set_nlink(inode, le32_to_cpu(grant->nlink)); |
ca20c9919 ceph: trim delete... |
3296 3297 |
if (inode->i_nlink == 0 && (newcaps & (CEPH_CAP_LINK_SHARED | CEPH_CAP_LINK_EXCL))) |
ab6c2c3eb ceph: fix bool as... |
3298 |
deleted_inode = true; |
ca20c9919 ceph: trim delete... |
3299 |
} |
a8599bd82 ceph: capability ... |
3300 |
|
a1c6b8358 ceph: define argu... |
3301 3302 |
if ((extra_info->issued & CEPH_CAP_XATTR_EXCL) == 0 && grant->xattr_len) { |
a8599bd82 ceph: capability ... |
3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 |
int len = le32_to_cpu(grant->xattr_len); u64 version = le64_to_cpu(grant->xattr_version); if (version > ci->i_xattrs.version) { dout(" got new xattrs v%llu on %p len %d ", version, inode, len); if (ci->i_xattrs.blob) ceph_buffer_put(ci->i_xattrs.blob); ci->i_xattrs.blob = ceph_buffer_get(xattr_buf); ci->i_xattrs.version = version; |
7221fe4c2 ceph: add acl for... |
3314 |
ceph_forget_all_cached_acls(inode); |
ac6713ccb ceph: add selinux... |
3315 |
ceph_security_invalidate_secctx(inode); |
a8599bd82 ceph: capability ... |
3316 3317 |
} } |
f98a128a5 ceph: update inod... |
3318 |
if (newcaps & CEPH_CAP_ANY_RD) { |
9bbeab41c ceph: use timespe... |
3319 |
struct timespec64 mtime, atime, ctime; |
f98a128a5 ceph: update inod... |
3320 |
/* ctime/mtime/atime? */ |
9bbeab41c ceph: use timespe... |
3321 3322 3323 |
ceph_decode_timespec64(&mtime, &grant->mtime); ceph_decode_timespec64(&atime, &grant->atime); ceph_decode_timespec64(&ctime, &grant->ctime); |
a1c6b8358 ceph: define argu... |
3324 |
ceph_fill_file_time(inode, extra_info->issued, |
f98a128a5 ceph: update inod... |
3325 3326 3327 |
le32_to_cpu(grant->time_warp_seq), &ctime, &mtime, &atime); } |
4985d6f9e ceph: handle the ... |
3328 3329 3330 3331 |
if ((newcaps & CEPH_CAP_FILE_SHARED) && extra_info->dirstat_valid) { ci->i_files = extra_info->nfiles; ci->i_subdirs = extra_info->nsubdirs; } |
f98a128a5 ceph: update inod... |
3332 3333 |
if (newcaps & (CEPH_CAP_ANY_FILE_RD | CEPH_CAP_ANY_FILE_WR)) { /* file layout may have changed */ |
7627151ea libceph: define n... |
3334 |
s64 old_pool = ci->i_layout.pool_id; |
779fe0fb8 ceph: rados pool ... |
3335 |
struct ceph_string *old_ns; |
7627151ea libceph: define n... |
3336 |
ceph_file_layout_from_legacy(&ci->i_layout, &grant->layout); |
779fe0fb8 ceph: rados pool ... |
3337 3338 |
old_ns = rcu_dereference_protected(ci->i_layout.pool_ns, lockdep_is_held(&ci->i_ceph_lock)); |
a1c6b8358 ceph: define argu... |
3339 |
rcu_assign_pointer(ci->i_layout.pool_ns, extra_info->pool_ns); |
779fe0fb8 ceph: rados pool ... |
3340 |
|
a1c6b8358 ceph: define argu... |
3341 3342 |
if (ci->i_layout.pool_id != old_pool || extra_info->pool_ns != old_ns) |
7627151ea libceph: define n... |
3343 |
ci->i_ceph_flags &= ~CEPH_I_POOL_PERM; |
5ea5c5e0a ceph: initial CEP... |
3344 |
|
a1c6b8358 ceph: define argu... |
3345 |
extra_info->pool_ns = old_ns; |
779fe0fb8 ceph: rados pool ... |
3346 |
|
f98a128a5 ceph: update inod... |
3347 |
/* size/truncate_seq? */ |
a1c6b8358 ceph: define argu... |
3348 |
queue_trunc = ceph_fill_file_size(inode, extra_info->issued, |
f98a128a5 ceph: update inod... |
3349 3350 3351 |
le32_to_cpu(grant->truncate_seq), le64_to_cpu(grant->truncate_size), size); |
84eea8c79 ceph: re-request ... |
3352 3353 3354 3355 |
} if (ci->i_auth_cap == cap && (newcaps & CEPH_CAP_ANY_FILE_WR)) { if (max_size != ci->i_max_size) { |
f98a128a5 ceph: update inod... |
3356 3357 3358 3359 3360 3361 3362 3363 |
dout("max_size %lld -> %llu ", ci->i_max_size, max_size); ci->i_max_size = max_size; if (max_size >= ci->i_wanted_max_size) { ci->i_wanted_max_size = 0; /* reset */ ci->i_requested_max_size = 0; } |
ab6c2c3eb ceph: fix bool as... |
3364 |
wake = true; |
a8599bd82 ceph: capability ... |
3365 |
} |
a8599bd82 ceph: capability ... |
3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 |
} /* check cap bits */ wanted = __ceph_caps_wanted(ci); used = __ceph_caps_used(ci); dirty = __ceph_caps_dirty(ci); dout(" my wanted = %s, used = %s, dirty %s ", ceph_cap_string(wanted), ceph_cap_string(used), ceph_cap_string(dirty)); |
fdac94fab ceph: skip updati... |
3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 |
if ((was_stale || le32_to_cpu(grant->op) == CEPH_CAP_OP_IMPORT) && (wanted & ~(cap->mds_wanted | newcaps))) { /* * If mds is importing cap, prior cap messages that update * 'wanted' may get dropped by mds (migrate seq mismatch). * * We don't send cap message to update 'wanted' if what we * want are already issued. If mds revokes caps, cap message * that releases caps also tells mds what we want. But if * caps got revoked by mds forcedly (session stale). We may * haven't told mds what we want. */ check_caps = 1; |
a8599bd82 ceph: capability ... |
3391 |
} |
a8599bd82 ceph: capability ... |
3392 3393 |
/* revocation, grant, or no-op? */ if (cap->issued & ~newcaps) { |
3b454c494 ceph: simplify ca... |
3394 3395 3396 3397 3398 3399 3400 |
int revoking = cap->issued & ~newcaps; dout("revocation: %s -> %s (revoking %s) ", ceph_cap_string(cap->issued), ceph_cap_string(newcaps), ceph_cap_string(revoking)); |
525d15e8e ceph: check inode... |
3401 3402 |
if (S_ISREG(inode->i_mode) && (revoking & used & CEPH_CAP_FILE_BUFFER)) |
ab6c2c3eb ceph: fix bool as... |
3403 |
writeback = true; /* initiate writeback; will delay ack */ |
525d15e8e ceph: check inode... |
3404 3405 3406 |
else if (queue_invalidate && revoking == CEPH_CAP_FILE_CACHE && (newcaps & CEPH_CAP_FILE_LAZYIO) == 0) |
3b454c494 ceph: simplify ca... |
3407 3408 3409 3410 3411 |
; /* do nothing yet, invalidation will be queued */ else if (cap == ci->i_auth_cap) check_caps = 1; /* check auth cap only */ else check_caps = 2; /* check all caps */ |
a8599bd82 ceph: capability ... |
3412 |
cap->issued = newcaps; |
978097c90 ceph: implemented... |
3413 |
cap->implemented |= newcaps; |
a8599bd82 ceph: capability ... |
3414 3415 3416 3417 3418 3419 3420 3421 |
} else if (cap->issued == newcaps) { dout("caps unchanged: %s -> %s ", ceph_cap_string(cap->issued), ceph_cap_string(newcaps)); } else { dout("grant: %s -> %s ", ceph_cap_string(cap->issued), ceph_cap_string(newcaps)); |
6ee6b9537 ceph: fix race be... |
3422 3423 3424 3425 |
/* non-auth MDS is revoking the newly grant caps ? */ if (cap == ci->i_auth_cap && __ceph_caps_revoking_other(ci, cap, newcaps)) check_caps = 2; |
a8599bd82 ceph: capability ... |
3426 3427 3428 3429 |
cap->issued = newcaps; cap->implemented |= newcaps; /* add bits only, to * avoid stepping on a * pending revocation */ |
ab6c2c3eb ceph: fix bool as... |
3430 |
wake = true; |
a8599bd82 ceph: capability ... |
3431 |
} |
978097c90 ceph: implemented... |
3432 |
BUG_ON(cap->issued & ~cap->implemented); |
a8599bd82 ceph: capability ... |
3433 |
|
a1c6b8358 ceph: define argu... |
3434 3435 3436 |
if (extra_info->inline_version > 0 && extra_info->inline_version >= ci->i_inline_version) { ci->i_inline_version = extra_info->inline_version; |
31c542a19 ceph: add inline ... |
3437 3438 3439 3440 |
if (ci->i_inline_version != CEPH_INLINE_NONE && (newcaps & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO))) fill_inline = true; } |
6f05b30ea ceph: reset i_req... |
3441 3442 |
if (ci->i_auth_cap == cap && le32_to_cpu(grant->op) == CEPH_CAP_OP_IMPORT) { |
a1c6b8358 ceph: define argu... |
3443 |
if (newcaps & ~extra_info->issued) |
ab6c2c3eb ceph: fix bool as... |
3444 |
wake = true; |
6f05b30ea ceph: reset i_req... |
3445 3446 3447 3448 3449 3450 3451 |
if (ci->i_requested_max_size > max_size || !(le32_to_cpu(grant->wanted) & CEPH_CAP_ANY_FILE_WR)) { /* re-request max_size if necessary */ ci->i_requested_max_size = 0; wake = true; } |
e8a4d2677 ceph: clean up ki... |
3452 3453 |
ceph_kick_flushing_inode_caps(session, ci); spin_unlock(&ci->i_ceph_lock); |
a1c6b8358 ceph: define argu... |
3454 |
up_read(&session->s_mdsc->snap_rwsem); |
0e2943878 ceph: unify cap f... |
3455 3456 |
} else { spin_unlock(&ci->i_ceph_lock); |
2cd698be9 ceph: handle cap ... |
3457 |
} |
31c542a19 ceph: add inline ... |
3458 |
if (fill_inline) |
a1c6b8358 ceph: define argu... |
3459 3460 |
ceph_fill_inline_data(inode, NULL, extra_info->inline_data, extra_info->inline_len); |
31c542a19 ceph: add inline ... |
3461 |
|
146497581 ceph: avoid unnec... |
3462 |
if (queue_trunc) |
c6bcda6f5 ceph: queue vmtru... |
3463 |
ceph_queue_vmtruncate(inode); |
c6bcda6f5 ceph: queue vmtru... |
3464 |
|
3c6f6b79a ceph: cleanup asy... |
3465 |
if (writeback) |
a8599bd82 ceph: capability ... |
3466 3467 3468 3469 3470 |
/* * queue inode for writeback: we can't actually call * filemap_write_and_wait, etc. from message handler * context. */ |
3c6f6b79a ceph: cleanup asy... |
3471 3472 3473 |
ceph_queue_writeback(inode); if (queue_invalidate) ceph_queue_invalidate(inode); |
ca20c9919 ceph: trim delete... |
3474 3475 |
if (deleted_inode) invalidate_aliases(inode); |
a8599bd82 ceph: capability ... |
3476 |
if (wake) |
03066f234 ceph: use complet... |
3477 |
wake_up_all(&ci->i_cap_wq); |
15637c8b1 ceph: clean up ha... |
3478 3479 |
if (check_caps == 1) |
a0d93e327 ceph: remove dela... |
3480 |
ceph_check_caps(ci, CHECK_CAPS_AUTHONLY | CHECK_CAPS_NOINVAL, |
15637c8b1 ceph: clean up ha... |
3481 3482 |
session); else if (check_caps == 2) |
a0d93e327 ceph: remove dela... |
3483 |
ceph_check_caps(ci, CHECK_CAPS_NOINVAL, session); |
15637c8b1 ceph: clean up ha... |
3484 3485 |
else mutex_unlock(&session->s_mutex); |
a8599bd82 ceph: capability ... |
3486 3487 3488 3489 3490 3491 |
} /* * Handle FLUSH_ACK from MDS, indicating that metadata we sent to the * MDS has been safely committed. */ |
6df058c02 ceph: include tra... |
3492 |
static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, |
a8599bd82 ceph: capability ... |
3493 3494 3495 |
struct ceph_mds_caps *m, struct ceph_mds_session *session, struct ceph_cap *cap) |
be655596b ceph: use i_ceph_... |
3496 |
__releases(ci->i_ceph_lock) |
a8599bd82 ceph: capability ... |
3497 3498 |
{ struct ceph_inode_info *ci = ceph_inode(inode); |
3d14c5d2b ceph: factor out ... |
3499 |
struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; |
e4500b5e3 ceph: use list in... |
3500 |
struct ceph_cap_flush *cf, *tmp_cf; |
553adfd94 ceph: track pendi... |
3501 |
LIST_HEAD(to_remove); |
a8599bd82 ceph: capability ... |
3502 3503 3504 |
unsigned seq = le32_to_cpu(m->seq); int dirty = le32_to_cpu(m->dirty); int cleaned = 0; |
c8799fc46 ceph: optimize ca... |
3505 |
bool drop = false; |
7271efa79 ceph: fix bool in... |
3506 3507 |
bool wake_ci = false; bool wake_mdsc = false; |
a8599bd82 ceph: capability ... |
3508 |
|
e4500b5e3 ceph: use list in... |
3509 |
list_for_each_entry_safe(cf, tmp_cf, &ci->i_cap_flush_list, i_list) { |
d7dbfb4f2 ceph: add comment... |
3510 |
/* Is this the one that was flushed? */ |
553adfd94 ceph: track pendi... |
3511 3512 |
if (cf->tid == flush_tid) cleaned = cf->caps; |
d7dbfb4f2 ceph: add comment... |
3513 3514 3515 |
/* Is this a capsnap? */ if (cf->caps == 0) |
0e2943878 ceph: unify cap f... |
3516 |
continue; |
d7dbfb4f2 ceph: add comment... |
3517 |
|
553adfd94 ceph: track pendi... |
3518 |
if (cf->tid <= flush_tid) { |
d7dbfb4f2 ceph: add comment... |
3519 3520 3521 3522 |
/* * An earlier or current tid. The FLUSH_ACK should * represent a superset of this flush's caps. */ |
681ac6348 ceph: split up __... |
3523 |
wake_ci |= __detach_cap_flush_from_ci(ci, cf); |
e4500b5e3 ceph: use list in... |
3524 |
list_add_tail(&cf->i_list, &to_remove); |
553adfd94 ceph: track pendi... |
3525 |
} else { |
d7dbfb4f2 ceph: add comment... |
3526 3527 3528 3529 |
/* * This is a later one. Any caps in it are still dirty * so don't count them as cleaned. */ |
553adfd94 ceph: track pendi... |
3530 3531 3532 3533 3534 |
cleaned &= ~cf->caps; if (!cleaned) break; } } |
a8599bd82 ceph: capability ... |
3535 3536 3537 3538 3539 3540 3541 |
dout("handle_cap_flush_ack inode %p mds%d seq %d on %s cleaned %s," " flushing %s -> %s ", inode, session->s_mds, seq, ceph_cap_string(dirty), ceph_cap_string(cleaned), ceph_cap_string(ci->i_flushing_caps), ceph_cap_string(ci->i_flushing_caps & ~cleaned)); |
8310b0891 ceph: track pendi... |
3542 |
if (list_empty(&to_remove) && !cleaned) |
a8599bd82 ceph: capability ... |
3543 |
goto out; |
a8599bd82 ceph: capability ... |
3544 |
ci->i_flushing_caps &= ~cleaned; |
a8599bd82 ceph: capability ... |
3545 3546 |
spin_lock(&mdsc->cap_dirty_lock); |
8310b0891 ceph: track pendi... |
3547 |
|
681ac6348 ceph: split up __... |
3548 3549 |
list_for_each_entry(cf, &to_remove, i_list) wake_mdsc |= __detach_cap_flush_from_mdsc(mdsc, cf); |
8310b0891 ceph: track pendi... |
3550 |
|
a8599bd82 ceph: capability ... |
3551 |
if (ci->i_flushing_caps == 0) { |
0e2943878 ceph: unify cap f... |
3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 |
if (list_empty(&ci->i_cap_flush_list)) { list_del_init(&ci->i_flushing_item); if (!list_empty(&session->s_cap_flushing)) { dout(" mds%d still flushing cap on %p ", session->s_mds, &list_first_entry(&session->s_cap_flushing, struct ceph_inode_info, i_flushing_item)->vfs_inode); } } |
a8599bd82 ceph: capability ... |
3563 |
mdsc->num_cap_flushing--; |
a8599bd82 ceph: capability ... |
3564 3565 |
dout(" inode %p now !flushing ", inode); |
afcdaea3f ceph: flush dirty... |
3566 3567 3568 3569 3570 |
if (ci->i_dirty_caps == 0) { dout(" inode %p now clean ", inode); BUG_ON(!list_empty(&ci->i_dirty_item)); |
c8799fc46 ceph: optimize ca... |
3571 |
drop = true; |
5dda377cf ceph: set i_head_... |
3572 3573 |
if (ci->i_wr_ref == 0 && ci->i_wrbuffer_ref_head == 0) { |
7d8cb26d7 ceph: maintain i_... |
3574 3575 3576 3577 |
BUG_ON(!ci->i_head_snapc); ceph_put_snap_context(ci->i_head_snapc); ci->i_head_snapc = NULL; } |
76e3b390d ceph: move dirty ... |
3578 3579 |
} else { BUG_ON(list_empty(&ci->i_dirty_item)); |
afcdaea3f ceph: flush dirty... |
3580 |
} |
a8599bd82 ceph: capability ... |
3581 3582 |
} spin_unlock(&mdsc->cap_dirty_lock); |
a8599bd82 ceph: capability ... |
3583 3584 |
out: |
be655596b ceph: use i_ceph_... |
3585 |
spin_unlock(&ci->i_ceph_lock); |
553adfd94 ceph: track pendi... |
3586 3587 3588 |
while (!list_empty(&to_remove)) { cf = list_first_entry(&to_remove, |
e4500b5e3 ceph: use list in... |
3589 3590 |
struct ceph_cap_flush, i_list); list_del(&cf->i_list); |
f66fd9f09 ceph: pre-allocat... |
3591 |
ceph_free_cap_flush(cf); |
553adfd94 ceph: track pendi... |
3592 |
} |
c8799fc46 ceph: optimize ca... |
3593 3594 3595 3596 3597 |
if (wake_ci) wake_up_all(&ci->i_cap_wq); if (wake_mdsc) wake_up_all(&mdsc->cap_flushing_wq); |
afcdaea3f ceph: flush dirty... |
3598 |
if (drop) |
a8599bd82 ceph: capability ... |
3599 3600 3601 3602 3603 3604 3605 3606 3607 |
iput(inode); } /* * Handle FLUSHSNAP_ACK. MDS has flushed snap data to disk and we can * throw away our cap_snap. * * Caller hold s_mutex. */ |
6df058c02 ceph: include tra... |
3608 |
static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, |
a8599bd82 ceph: capability ... |
3609 3610 3611 3612 |
struct ceph_mds_caps *m, struct ceph_mds_session *session) { struct ceph_inode_info *ci = ceph_inode(inode); |
affbc19a6 ceph: make sure s... |
3613 |
struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; |
a8599bd82 ceph: capability ... |
3614 |
u64 follows = le64_to_cpu(m->snap_follows); |
a8599bd82 ceph: capability ... |
3615 |
struct ceph_cap_snap *capsnap; |
c8799fc46 ceph: optimize ca... |
3616 3617 3618 |
bool flushed = false; bool wake_ci = false; bool wake_mdsc = false; |
a8599bd82 ceph: capability ... |
3619 3620 3621 3622 |
dout("handle_cap_flushsnap_ack inode %p ci %p mds%d follows %lld ", inode, ci, session->s_mds, follows); |
be655596b ceph: use i_ceph_... |
3623 |
spin_lock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
3624 3625 |
list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { if (capsnap->follows == follows) { |
0e2943878 ceph: unify cap f... |
3626 |
if (capsnap->cap_flush.tid != flush_tid) { |
a8599bd82 ceph: capability ... |
3627 3628 3629 |
dout(" cap_snap %p follows %lld tid %lld !=" " %lld ", capsnap, follows, |
0e2943878 ceph: unify cap f... |
3630 |
flush_tid, capsnap->cap_flush.tid); |
a8599bd82 ceph: capability ... |
3631 3632 |
break; } |
c8799fc46 ceph: optimize ca... |
3633 |
flushed = true; |
a8599bd82 ceph: capability ... |
3634 3635 3636 3637 3638 3639 3640 |
break; } else { dout(" skipping cap_snap %p follows %lld ", capsnap, capsnap->follows); } } |
0e2943878 ceph: unify cap f... |
3641 |
if (flushed) { |
0e2943878 ceph: unify cap f... |
3642 3643 3644 3645 3646 |
WARN_ON(capsnap->dirty_pages || capsnap->writing); dout(" removing %p cap_snap %p follows %lld ", inode, capsnap, follows); list_del(&capsnap->ci_item); |
681ac6348 ceph: split up __... |
3647 |
wake_ci |= __detach_cap_flush_from_ci(ci, &capsnap->cap_flush); |
0e2943878 ceph: unify cap f... |
3648 3649 3650 3651 3652 |
spin_lock(&mdsc->cap_dirty_lock); if (list_empty(&ci->i_cap_flush_list)) list_del_init(&ci->i_flushing_item); |
681ac6348 ceph: split up __... |
3653 3654 |
wake_mdsc |= __detach_cap_flush_from_mdsc(mdsc, &capsnap->cap_flush); |
0e2943878 ceph: unify cap f... |
3655 |
spin_unlock(&mdsc->cap_dirty_lock); |
0e2943878 ceph: unify cap f... |
3656 |
} |
be655596b ceph: use i_ceph_... |
3657 |
spin_unlock(&ci->i_ceph_lock); |
0e2943878 ceph: unify cap f... |
3658 3659 3660 |
if (flushed) { ceph_put_snap_context(capsnap->context); ceph_put_cap_snap(capsnap); |
c8799fc46 ceph: optimize ca... |
3661 3662 3663 3664 |
if (wake_ci) wake_up_all(&ci->i_cap_wq); if (wake_mdsc) wake_up_all(&mdsc->cap_flushing_wq); |
a8599bd82 ceph: capability ... |
3665 |
iput(inode); |
0e2943878 ceph: unify cap f... |
3666 |
} |
a8599bd82 ceph: capability ... |
3667 3668 3669 3670 3671 3672 3673 |
} /* * Handle TRUNC from MDS, indicating file truncation. * * caller hold s_mutex. */ |
7391fba26 ceph: don't relea... |
3674 |
static bool handle_cap_trunc(struct inode *inode, |
a8599bd82 ceph: capability ... |
3675 3676 |
struct ceph_mds_caps *trunc, struct ceph_mds_session *session) |
a8599bd82 ceph: capability ... |
3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 |
{ struct ceph_inode_info *ci = ceph_inode(inode); int mds = session->s_mds; int seq = le32_to_cpu(trunc->seq); u32 truncate_seq = le32_to_cpu(trunc->truncate_seq); u64 truncate_size = le64_to_cpu(trunc->truncate_size); u64 size = le64_to_cpu(trunc->size); int implemented = 0; int dirty = __ceph_caps_dirty(ci); int issued = __ceph_caps_issued(ceph_inode(inode), &implemented); |
7391fba26 ceph: don't relea... |
3687 3688 3689 |
bool queue_trunc = false; lockdep_assert_held(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
3690 3691 3692 3693 3694 3695 3696 3697 |
issued |= implemented | dirty; dout("handle_cap_trunc inode %p mds%d seq %d to %lld seq %d ", inode, mds, seq, truncate_size, truncate_seq); queue_trunc = ceph_fill_file_size(inode, issued, truncate_seq, truncate_size, size); |
7391fba26 ceph: don't relea... |
3698 |
return queue_trunc; |
a8599bd82 ceph: capability ... |
3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 |
} /* * Handle EXPORT from MDS. Cap is being migrated _from_ this mds to a * different one. If we are the most recent migration we've seen (as * indicated by mseq), make note of the migrating cap bits for the * duration (until we see the corresponding IMPORT). * * caller holds s_mutex */ static void handle_cap_export(struct inode *inode, struct ceph_mds_caps *ex, |
11df2dfb6 ceph: add importe... |
3710 3711 |
struct ceph_mds_cap_peer *ph, struct ceph_mds_session *session) |
a8599bd82 ceph: capability ... |
3712 |
{ |
db3540522 ceph: fix cap flu... |
3713 |
struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; |
11df2dfb6 ceph: add importe... |
3714 |
struct ceph_mds_session *tsession = NULL; |
d9df27835 ceph: pre-allocat... |
3715 |
struct ceph_cap *cap, *tcap, *new_cap = NULL; |
a8599bd82 ceph: capability ... |
3716 |
struct ceph_inode_info *ci = ceph_inode(inode); |
11df2dfb6 ceph: add importe... |
3717 |
u64 t_cap_id; |
a8599bd82 ceph: capability ... |
3718 |
unsigned mseq = le32_to_cpu(ex->migrate_seq); |
11df2dfb6 ceph: add importe... |
3719 3720 3721 |
unsigned t_seq, t_mseq; int target, issued; int mds = session->s_mds; |
a8599bd82 ceph: capability ... |
3722 |
|
11df2dfb6 ceph: add importe... |
3723 3724 3725 3726 3727 3728 3729 3730 3731 |
if (ph) { t_cap_id = le64_to_cpu(ph->cap_id); t_seq = le32_to_cpu(ph->seq); t_mseq = le32_to_cpu(ph->mseq); target = le32_to_cpu(ph->mds); } else { t_cap_id = t_seq = t_mseq = 0; target = -1; } |
a8599bd82 ceph: capability ... |
3732 |
|
11df2dfb6 ceph: add importe... |
3733 3734 3735 3736 |
dout("handle_cap_export inode %p ci %p mds%d mseq %d target %d ", inode, ci, mds, mseq, target); retry: |
be655596b ceph: use i_ceph_... |
3737 |
spin_lock(&ci->i_ceph_lock); |
11df2dfb6 ceph: add importe... |
3738 |
cap = __get_cap_for_mds(ci, mds); |
ca665e028 mds: check cap ID... |
3739 |
if (!cap || cap->cap_id != le64_to_cpu(ex->cap_id)) |
11df2dfb6 ceph: add importe... |
3740 |
goto out_unlock; |
a8599bd82 ceph: capability ... |
3741 |
|
11df2dfb6 ceph: add importe... |
3742 |
if (target < 0) { |
d2f8bb27c ceph: update want... |
3743 |
__ceph_remove_cap(cap, false); |
11df2dfb6 ceph: add importe... |
3744 |
goto out_unlock; |
a8599bd82 ceph: capability ... |
3745 |
} |
11df2dfb6 ceph: add importe... |
3746 3747 3748 3749 |
/* * now we know we haven't received the cap import message yet * because the exported cap still exist. */ |
db3540522 ceph: fix cap flu... |
3750 |
|
11df2dfb6 ceph: add importe... |
3751 |
issued = cap->issued; |
d84b37f9f ceph: limit rate ... |
3752 3753 3754 3755 3756 3757 3758 3759 |
if (issued != cap->implemented) pr_err_ratelimited("handle_cap_export: issued != implemented: " "ino (%llx.%llx) mds%d seq %d mseq %d " "issued %s implemented %s ", ceph_vinop(inode), mds, cap->seq, cap->mseq, ceph_cap_string(issued), ceph_cap_string(cap->implemented)); |
11df2dfb6 ceph: add importe... |
3760 3761 3762 3763 |
tcap = __get_cap_for_mds(ci, target); if (tcap) { /* already have caps from the target */ |
fa0aa3b83 ceph: fix message... |
3764 |
if (tcap->cap_id == t_cap_id && |
11df2dfb6 ceph: add importe... |
3765 3766 3767 3768 3769 3770 |
ceph_seq_cmp(tcap->seq, t_seq) < 0) { dout(" updating import cap %p mds%d ", tcap, target); tcap->cap_id = t_cap_id; tcap->seq = t_seq - 1; tcap->issue_seq = t_seq - 1; |
11df2dfb6 ceph: add importe... |
3771 3772 |
tcap->issued |= issued; tcap->implemented |= issued; |
1cf03a68e ceph: convert mds... |
3773 |
if (cap == ci->i_auth_cap) { |
11df2dfb6 ceph: add importe... |
3774 |
ci->i_auth_cap = tcap; |
1cf03a68e ceph: convert mds... |
3775 |
change_auth_cap_ses(ci, tcap->session); |
db3540522 ceph: fix cap flu... |
3776 |
} |
a8599bd82 ceph: capability ... |
3777 |
} |
a096b09ae ceph: queue cap r... |
3778 |
__ceph_remove_cap(cap, false); |
11df2dfb6 ceph: add importe... |
3779 |
goto out_unlock; |
d9df27835 ceph: pre-allocat... |
3780 |
} else if (tsession) { |
11df2dfb6 ceph: add importe... |
3781 |
/* add placeholder for the export tagert */ |
d9df27835 ceph: pre-allocat... |
3782 |
int flag = (cap == ci->i_auth_cap) ? CEPH_CAP_FLAG_AUTH : 0; |
00f06cba5 ceph: make sure f... |
3783 |
tcap = new_cap; |
135e671e5 ceph: simplify ca... |
3784 |
ceph_add_cap(inode, tsession, t_cap_id, issued, 0, |
d9df27835 ceph: pre-allocat... |
3785 |
t_seq - 1, t_mseq, (u64)-1, flag, &new_cap); |
00f06cba5 ceph: make sure f... |
3786 3787 3788 3789 3790 3791 3792 |
if (!list_empty(&ci->i_cap_flush_list) && ci->i_auth_cap == tcap) { spin_lock(&mdsc->cap_dirty_lock); list_move_tail(&ci->i_flushing_item, &tcap->session->s_cap_flushing); spin_unlock(&mdsc->cap_dirty_lock); } |
d9df27835 ceph: pre-allocat... |
3793 3794 |
__ceph_remove_cap(cap, false); goto out_unlock; |
a8599bd82 ceph: capability ... |
3795 |
} |
be655596b ceph: use i_ceph_... |
3796 |
spin_unlock(&ci->i_ceph_lock); |
11df2dfb6 ceph: add importe... |
3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 |
mutex_unlock(&session->s_mutex); /* open target session */ tsession = ceph_mdsc_open_export_target_session(mdsc, target); if (!IS_ERR(tsession)) { if (mds > target) { mutex_lock(&session->s_mutex); mutex_lock_nested(&tsession->s_mutex, SINGLE_DEPTH_NESTING); } else { mutex_lock(&tsession->s_mutex); mutex_lock_nested(&session->s_mutex, SINGLE_DEPTH_NESTING); } |
d9df27835 ceph: pre-allocat... |
3811 |
new_cap = ceph_get_cap(mdsc, NULL); |
11df2dfb6 ceph: add importe... |
3812 3813 3814 3815 |
} else { WARN_ON(1); tsession = NULL; target = -1; |
4d8e28ff3 ceph: fix double ... |
3816 |
mutex_lock(&session->s_mutex); |
11df2dfb6 ceph: add importe... |
3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 |
} goto retry; out_unlock: spin_unlock(&ci->i_ceph_lock); mutex_unlock(&session->s_mutex); if (tsession) { mutex_unlock(&tsession->s_mutex); ceph_put_mds_session(tsession); } |
d9df27835 ceph: pre-allocat... |
3827 3828 |
if (new_cap) ceph_put_cap(mdsc, new_cap); |
a8599bd82 ceph: capability ... |
3829 3830 3831 |
} /* |
2cd698be9 ceph: handle cap ... |
3832 |
* Handle cap IMPORT. |
a8599bd82 ceph: capability ... |
3833 |
* |
2cd698be9 ceph: handle cap ... |
3834 |
* caller holds s_mutex. acquires i_ceph_lock |
a8599bd82 ceph: capability ... |
3835 3836 3837 |
*/ static void handle_cap_import(struct ceph_mds_client *mdsc, struct inode *inode, struct ceph_mds_caps *im, |
4ee6a914e ceph: remove expo... |
3838 |
struct ceph_mds_cap_peer *ph, |
a8599bd82 ceph: capability ... |
3839 |
struct ceph_mds_session *session, |
2cd698be9 ceph: handle cap ... |
3840 |
struct ceph_cap **target_cap, int *old_issued) |
a8599bd82 ceph: capability ... |
3841 3842 |
{ struct ceph_inode_info *ci = ceph_inode(inode); |
2cd698be9 ceph: handle cap ... |
3843 |
struct ceph_cap *cap, *ocap, *new_cap = NULL; |
a8599bd82 ceph: capability ... |
3844 |
int mds = session->s_mds; |
2cd698be9 ceph: handle cap ... |
3845 3846 |
int issued; unsigned caps = le32_to_cpu(im->caps); |
a8599bd82 ceph: capability ... |
3847 3848 3849 3850 3851 |
unsigned wanted = le32_to_cpu(im->wanted); unsigned seq = le32_to_cpu(im->seq); unsigned mseq = le32_to_cpu(im->migrate_seq); u64 realmino = le64_to_cpu(im->realm); u64 cap_id = le64_to_cpu(im->cap_id); |
4ee6a914e ceph: remove expo... |
3852 3853 |
u64 p_cap_id; int peer; |
a8599bd82 ceph: capability ... |
3854 |
|
4ee6a914e ceph: remove expo... |
3855 3856 3857 3858 3859 3860 3861 |
if (ph) { p_cap_id = le64_to_cpu(ph->cap_id); peer = le32_to_cpu(ph->mds); } else { p_cap_id = 0; peer = -1; } |
db3540522 ceph: fix cap flu... |
3862 |
|
4ee6a914e ceph: remove expo... |
3863 3864 3865 |
dout("handle_cap_import inode %p ci %p mds%d mseq %d peer %d ", inode, ci, mds, mseq, peer); |
d9df27835 ceph: pre-allocat... |
3866 |
retry: |
d9df27835 ceph: pre-allocat... |
3867 3868 3869 3870 3871 |
cap = __get_cap_for_mds(ci, mds); if (!cap) { if (!new_cap) { spin_unlock(&ci->i_ceph_lock); new_cap = ceph_get_cap(mdsc, NULL); |
783332336 ceph: don't take ... |
3872 |
spin_lock(&ci->i_ceph_lock); |
d9df27835 ceph: pre-allocat... |
3873 3874 |
goto retry; } |
2cd698be9 ceph: handle cap ... |
3875 3876 3877 3878 3879 3880 |
cap = new_cap; } else { if (new_cap) { ceph_put_cap(mdsc, new_cap); new_cap = NULL; } |
d9df27835 ceph: pre-allocat... |
3881 |
} |
2cd698be9 ceph: handle cap ... |
3882 3883 |
__ceph_caps_issued(ci, &issued); issued |= __ceph_caps_dirty(ci); |
135e671e5 ceph: simplify ca... |
3884 |
ceph_add_cap(inode, session, cap_id, caps, wanted, seq, mseq, |
d9df27835 ceph: pre-allocat... |
3885 |
realmino, CEPH_CAP_FLAG_AUTH, &new_cap); |
2cd698be9 ceph: handle cap ... |
3886 3887 |
ocap = peer >= 0 ? __get_cap_for_mds(ci, peer) : NULL; if (ocap && ocap->cap_id == p_cap_id) { |
4ee6a914e ceph: remove expo... |
3888 3889 |
dout(" remove export cap %p mds%d flags %d ", |
2cd698be9 ceph: handle cap ... |
3890 |
ocap, peer, ph->flags); |
4ee6a914e ceph: remove expo... |
3891 |
if ((ph->flags & CEPH_CAP_FLAG_AUTH) && |
2cd698be9 ceph: handle cap ... |
3892 3893 |
(ocap->seq != le32_to_cpu(ph->seq) || ocap->mseq != le32_to_cpu(ph->mseq))) { |
d84b37f9f ceph: limit rate ... |
3894 3895 3896 3897 3898 3899 3900 3901 |
pr_err_ratelimited("handle_cap_import: " "mismatched seq/mseq: ino (%llx.%llx) " "mds%d seq %d mseq %d importer mds%d " "has peer seq %d mseq %d ", ceph_vinop(inode), peer, ocap->seq, ocap->mseq, mds, le32_to_cpu(ph->seq), le32_to_cpu(ph->mseq)); |
db3540522 ceph: fix cap flu... |
3902 |
} |
2cd698be9 ceph: handle cap ... |
3903 |
__ceph_remove_cap(ocap, (ph->flags & CEPH_CAP_FLAG_RELEASE)); |
a8599bd82 ceph: capability ... |
3904 |
} |
2cd698be9 ceph: handle cap ... |
3905 3906 |
*old_issued = issued; *target_cap = cap; |
a8599bd82 ceph: capability ... |
3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 |
} /* * Handle a caps message from the MDS. * * Identify the appropriate session, inode, and call the right handler * based on the cap op. */ void ceph_handle_caps(struct ceph_mds_session *session, struct ceph_msg *msg) { struct ceph_mds_client *mdsc = session->s_mdsc; |
a8599bd82 ceph: capability ... |
3919 |
struct inode *inode; |
be655596b ceph: use i_ceph_... |
3920 |
struct ceph_inode_info *ci; |
a8599bd82 ceph: capability ... |
3921 3922 |
struct ceph_cap *cap; struct ceph_mds_caps *h; |
4ee6a914e ceph: remove expo... |
3923 |
struct ceph_mds_cap_peer *peer = NULL; |
779fe0fb8 ceph: rados pool ... |
3924 |
struct ceph_snap_realm *realm = NULL; |
a1c6b8358 ceph: define argu... |
3925 |
int op; |
4985d6f9e ceph: handle the ... |
3926 |
int msg_version = le16_to_cpu(msg->hdr.version); |
3d7ded4d8 ceph: release cap... |
3927 |
u32 seq, mseq; |
a8599bd82 ceph: capability ... |
3928 |
struct ceph_vino vino; |
70edb55bd ceph: fix snaptra... |
3929 |
void *snaptrace; |
ce1fbc8dd ceph: support v2 ... |
3930 |
size_t snaptrace_len; |
fb01d1f8b ceph: parse inlin... |
3931 |
void *p, *end; |
a1c6b8358 ceph: define argu... |
3932 |
struct cap_extra_info extra_info = {}; |
7391fba26 ceph: don't relea... |
3933 |
bool queue_trunc; |
a8599bd82 ceph: capability ... |
3934 |
|
a1c6b8358 ceph: define argu... |
3935 3936 |
dout("handle_caps from mds%d ", session->s_mds); |
a8599bd82 ceph: capability ... |
3937 3938 |
/* decode */ |
4ee6a914e ceph: remove expo... |
3939 |
end = msg->front.iov_base + msg->front.iov_len; |
a8599bd82 ceph: capability ... |
3940 3941 3942 3943 3944 3945 |
if (msg->front.iov_len < sizeof(*h)) goto bad; h = msg->front.iov_base; op = le32_to_cpu(h->op); vino.ino = le64_to_cpu(h->ino); vino.snap = CEPH_NOSNAP; |
a8599bd82 ceph: capability ... |
3946 |
seq = le32_to_cpu(h->seq); |
3d7ded4d8 ceph: release cap... |
3947 |
mseq = le32_to_cpu(h->migrate_seq); |
a8599bd82 ceph: capability ... |
3948 |
|
ce1fbc8dd ceph: support v2 ... |
3949 3950 |
snaptrace = h + 1; snaptrace_len = le32_to_cpu(h->snap_trace_len); |
fb01d1f8b ceph: parse inlin... |
3951 |
p = snaptrace + snaptrace_len; |
ce1fbc8dd ceph: support v2 ... |
3952 |
|
4985d6f9e ceph: handle the ... |
3953 |
if (msg_version >= 2) { |
fb01d1f8b ceph: parse inlin... |
3954 |
u32 flock_len; |
ce1fbc8dd ceph: support v2 ... |
3955 |
ceph_decode_32_safe(&p, end, flock_len, bad); |
4ee6a914e ceph: remove expo... |
3956 3957 |
if (p + flock_len > end) goto bad; |
fb01d1f8b ceph: parse inlin... |
3958 |
p += flock_len; |
ce1fbc8dd ceph: support v2 ... |
3959 |
} |
4985d6f9e ceph: handle the ... |
3960 |
if (msg_version >= 3) { |
4ee6a914e ceph: remove expo... |
3961 |
if (op == CEPH_CAP_OP_IMPORT) { |
4ee6a914e ceph: remove expo... |
3962 3963 3964 |
if (p + sizeof(*peer) > end) goto bad; peer = p; |
fb01d1f8b ceph: parse inlin... |
3965 |
p += sizeof(*peer); |
11df2dfb6 ceph: add importe... |
3966 3967 3968 |
} else if (op == CEPH_CAP_OP_EXPORT) { /* recorded in unused fields */ peer = (void *)&h->size; |
4ee6a914e ceph: remove expo... |
3969 3970 |
} } |
4985d6f9e ceph: handle the ... |
3971 |
if (msg_version >= 4) { |
a1c6b8358 ceph: define argu... |
3972 3973 3974 |
ceph_decode_64_safe(&p, end, extra_info.inline_version, bad); ceph_decode_32_safe(&p, end, extra_info.inline_len, bad); if (p + extra_info.inline_len > end) |
fb01d1f8b ceph: parse inlin... |
3975 |
goto bad; |
a1c6b8358 ceph: define argu... |
3976 3977 |
extra_info.inline_data = p; p += extra_info.inline_len; |
fb01d1f8b ceph: parse inlin... |
3978 |
} |
4985d6f9e ceph: handle the ... |
3979 |
if (msg_version >= 5) { |
92475f05b ceph: handle epoc... |
3980 3981 3982 3983 3984 3985 |
struct ceph_osd_client *osdc = &mdsc->fsc->client->osdc; u32 epoch_barrier; ceph_decode_32_safe(&p, end, epoch_barrier, bad); ceph_osdc_update_epoch_barrier(osdc, epoch_barrier); } |
4985d6f9e ceph: handle the ... |
3986 |
if (msg_version >= 8) { |
5ea5c5e0a ceph: initial CEP... |
3987 3988 |
u64 flush_tid; u32 caller_uid, caller_gid; |
779fe0fb8 ceph: rados pool ... |
3989 |
u32 pool_ns_len; |
92475f05b ceph: handle epoc... |
3990 |
|
5ea5c5e0a ceph: initial CEP... |
3991 3992 3993 3994 3995 3996 3997 |
/* version >= 6 */ ceph_decode_64_safe(&p, end, flush_tid, bad); /* version >= 7 */ ceph_decode_32_safe(&p, end, caller_uid, bad); ceph_decode_32_safe(&p, end, caller_gid, bad); /* version >= 8 */ ceph_decode_32_safe(&p, end, pool_ns_len, bad); |
779fe0fb8 ceph: rados pool ... |
3998 3999 |
if (pool_ns_len > 0) { ceph_decode_need(&p, end, pool_ns_len, bad); |
a1c6b8358 ceph: define argu... |
4000 4001 |
extra_info.pool_ns = ceph_find_or_create_string(p, pool_ns_len); |
779fe0fb8 ceph: rados pool ... |
4002 4003 |
p += pool_ns_len; } |
5ea5c5e0a ceph: initial CEP... |
4004 |
} |
ec62b894d ceph: handle btim... |
4005 |
if (msg_version >= 9) { |
4985d6f9e ceph: handle the ... |
4006 |
struct ceph_timespec *btime; |
4985d6f9e ceph: handle the ... |
4007 |
|
4985d6f9e ceph: handle the ... |
4008 4009 4010 |
if (p + sizeof(*btime) > end) goto bad; btime = p; |
ec62b894d ceph: handle btim... |
4011 |
ceph_decode_timespec64(&extra_info.btime, btime); |
4985d6f9e ceph: handle the ... |
4012 |
p += sizeof(*btime); |
176c77c9c ceph: handle chan... |
4013 |
ceph_decode_64_safe(&p, end, extra_info.change_attr, bad); |
ec62b894d ceph: handle btim... |
4014 4015 4016 4017 |
} if (msg_version >= 11) { u32 flags; |
4985d6f9e ceph: handle the ... |
4018 4019 4020 4021 4022 4023 4024 |
/* version >= 10 */ ceph_decode_32_safe(&p, end, flags, bad); /* version >= 11 */ extra_info.dirstat_valid = true; ceph_decode_64_safe(&p, end, extra_info.nfiles, bad); ceph_decode_64_safe(&p, end, extra_info.nsubdirs, bad); } |
6cd3bcad0 ceph: move ceph_f... |
4025 |
/* lookup ino */ |
a1c6b8358 ceph: define argu... |
4026 |
inode = ceph_find_inode(mdsc->fsc->sb, vino); |
6cd3bcad0 ceph: move ceph_f... |
4027 4028 4029 4030 |
ci = ceph_inode(inode); dout(" op %s ino %llx.%llx inode %p ", ceph_cap_op_name(op), vino.ino, vino.snap, inode); |
a8599bd82 ceph: capability ... |
4031 |
mutex_lock(&session->s_mutex); |
62575e270 ceph: check sessi... |
4032 |
inc_session_sequence(session); |
a8599bd82 ceph: capability ... |
4033 4034 4035 |
dout(" mds%d seq %lld cap seq %u ", session->s_mds, session->s_seq, (unsigned)seq); |
a8599bd82 ceph: capability ... |
4036 4037 4038 |
if (!inode) { dout(" i don't have ino %llx ", vino.ino); |
3d7ded4d8 ceph: release cap... |
4039 |
|
a096b09ae ceph: queue cap r... |
4040 |
if (op == CEPH_CAP_OP_IMPORT) { |
745a8e3bc ceph: don't pre-a... |
4041 4042 4043 |
cap = ceph_get_cap(mdsc, NULL); cap->cap_ino = vino.ino; cap->queue_release = 1; |
779fe0fb8 ceph: rados pool ... |
4044 |
cap->cap_id = le64_to_cpu(h->cap_id); |
745a8e3bc ceph: don't pre-a... |
4045 4046 |
cap->mseq = mseq; cap->seq = seq; |
dc24de82d ceph: properly se... |
4047 |
cap->issue_seq = seq; |
a096b09ae ceph: queue cap r... |
4048 |
spin_lock(&session->s_cap_lock); |
e3ec8d689 ceph: send cap re... |
4049 |
__ceph_queue_cap_release(session, cap); |
a096b09ae ceph: queue cap r... |
4050 4051 |
spin_unlock(&session->s_cap_lock); } |
fb33c114d ceph: flush relea... |
4052 |
goto flush_cap_releases; |
a8599bd82 ceph: capability ... |
4053 4054 4055 4056 4057 |
} /* these will work even if we don't have a cap yet */ switch (op) { case CEPH_CAP_OP_FLUSHSNAP_ACK: |
a1c6b8358 ceph: define argu... |
4058 4059 |
handle_cap_flushsnap_ack(inode, le64_to_cpu(msg->hdr.tid), h, session); |
a8599bd82 ceph: capability ... |
4060 4061 4062 |
goto done; case CEPH_CAP_OP_EXPORT: |
11df2dfb6 ceph: add importe... |
4063 4064 |
handle_cap_export(inode, h, peer, session); goto done_unlocked; |
a8599bd82 ceph: capability ... |
4065 4066 |
case CEPH_CAP_OP_IMPORT: |
982d6011b ceph: improve ref... |
4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 |
realm = NULL; if (snaptrace_len) { down_write(&mdsc->snap_rwsem); ceph_update_snap_trace(mdsc, snaptrace, snaptrace + snaptrace_len, false, &realm); downgrade_write(&mdsc->snap_rwsem); } else { down_read(&mdsc->snap_rwsem); } |
783332336 ceph: don't take ... |
4077 |
spin_lock(&ci->i_ceph_lock); |
4ee6a914e ceph: remove expo... |
4078 |
handle_cap_import(mdsc, inode, h, peer, session, |
a1c6b8358 ceph: define argu... |
4079 4080 4081 |
&cap, &extra_info.issued); handle_cap_grant(inode, session, cap, h, msg->middle, &extra_info); |
982d6011b ceph: improve ref... |
4082 4083 |
if (realm) ceph_put_snap_realm(mdsc, realm); |
2cd698be9 ceph: handle cap ... |
4084 |
goto done_unlocked; |
a8599bd82 ceph: capability ... |
4085 4086 4087 |
} /* the rest require a cap */ |
be655596b ceph: use i_ceph_... |
4088 |
spin_lock(&ci->i_ceph_lock); |
a1c6b8358 ceph: define argu... |
4089 |
cap = __get_cap_for_mds(ceph_inode(inode), session->s_mds); |
a8599bd82 ceph: capability ... |
4090 |
if (!cap) { |
9dbd412f5 ceph: fix mislead... |
4091 4092 |
dout(" no cap on %p ino %llx.%llx from mds%d ", |
a1c6b8358 ceph: define argu... |
4093 4094 |
inode, ceph_ino(inode), ceph_snap(inode), session->s_mds); |
be655596b ceph: use i_ceph_... |
4095 |
spin_unlock(&ci->i_ceph_lock); |
21b559de5 ceph: send cap re... |
4096 |
goto flush_cap_releases; |
a8599bd82 ceph: capability ... |
4097 |
} |
be655596b ceph: use i_ceph_... |
4098 |
/* note that each of these drops i_ceph_lock for us */ |
a8599bd82 ceph: capability ... |
4099 4100 4101 |
switch (op) { case CEPH_CAP_OP_REVOKE: case CEPH_CAP_OP_GRANT: |
a1c6b8358 ceph: define argu... |
4102 4103 4104 4105 |
__ceph_caps_issued(ci, &extra_info.issued); extra_info.issued |= __ceph_caps_dirty(ci); handle_cap_grant(inode, session, cap, h, msg->middle, &extra_info); |
15637c8b1 ceph: clean up ha... |
4106 |
goto done_unlocked; |
a8599bd82 ceph: capability ... |
4107 4108 |
case CEPH_CAP_OP_FLUSH_ACK: |
a1c6b8358 ceph: define argu... |
4109 4110 |
handle_cap_flush_ack(inode, le64_to_cpu(msg->hdr.tid), h, session, cap); |
a8599bd82 ceph: capability ... |
4111 4112 4113 |
break; case CEPH_CAP_OP_TRUNC: |
7391fba26 ceph: don't relea... |
4114 4115 4116 4117 |
queue_trunc = handle_cap_trunc(inode, h, session); spin_unlock(&ci->i_ceph_lock); if (queue_trunc) ceph_queue_vmtruncate(inode); |
a8599bd82 ceph: capability ... |
4118 4119 4120 |
break; default: |
be655596b ceph: use i_ceph_... |
4121 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
4122 4123 4124 4125 |
pr_err("ceph_handle_caps: unknown cap op %d %s ", op, ceph_cap_op_name(op)); } |
e3ec8d689 ceph: send cap re... |
4126 4127 4128 |
done: mutex_unlock(&session->s_mutex); done_unlocked: |
e3ec8d689 ceph: send cap re... |
4129 |
ceph_put_string(extra_info.pool_ns); |
3e1d0452e ceph: avoid iput_... |
4130 4131 |
/* avoid calling iput_final() in mds dispatch threads */ ceph_async_iput(inode); |
e3ec8d689 ceph: send cap re... |
4132 |
return; |
21b559de5 ceph: send cap re... |
4133 4134 4135 |
flush_cap_releases: /* |
745a8e3bc ceph: don't pre-a... |
4136 |
* send any cap release message to try to move things |
21b559de5 ceph: send cap re... |
4137 4138 4139 |
* along for the mds (who clearly thinks we still have this * cap). */ |
e3ec8d689 ceph: send cap re... |
4140 4141 |
ceph_flush_cap_releases(mdsc, session); goto done; |
a8599bd82 ceph: capability ... |
4142 4143 4144 4145 |
bad: pr_err("ceph_handle_caps: corrupt message "); |
9ec7cab14 ceph: hex dump co... |
4146 |
ceph_msg_dump(msg); |
a8599bd82 ceph: capability ... |
4147 4148 4149 4150 4151 4152 |
return; } /* * Delayed work handler to process end of delayed cap release LRU list. */ |
afcdaea3f ceph: flush dirty... |
4153 |
void ceph_check_delayed_caps(struct ceph_mds_client *mdsc) |
a8599bd82 ceph: capability ... |
4154 |
{ |
4b9f2042f ceph: avoid acces... |
4155 |
struct inode *inode; |
a8599bd82 ceph: capability ... |
4156 |
struct ceph_inode_info *ci; |
a8599bd82 ceph: capability ... |
4157 |
|
a8599bd82 ceph: capability ... |
4158 4159 |
dout("check_delayed_caps "); |
585d72f33 ceph: clean up an... |
4160 4161 |
spin_lock(&mdsc->cap_delay_lock); while (!list_empty(&mdsc->cap_delay_list)) { |
a8599bd82 ceph: capability ... |
4162 4163 4164 4165 4166 4167 4168 |
ci = list_first_entry(&mdsc->cap_delay_list, struct ceph_inode_info, i_cap_delay_list); if ((ci->i_ceph_flags & CEPH_I_FLUSH) == 0 && time_before(jiffies, ci->i_hold_caps_max)) break; list_del_init(&ci->i_cap_delay_list); |
4b9f2042f ceph: avoid acces... |
4169 4170 |
inode = igrab(&ci->vfs_inode); |
4b9f2042f ceph: avoid acces... |
4171 |
if (inode) { |
585d72f33 ceph: clean up an... |
4172 |
spin_unlock(&mdsc->cap_delay_lock); |
4b9f2042f ceph: avoid acces... |
4173 4174 |
dout("check_delayed_caps on %p ", inode); |
a0d93e327 ceph: remove dela... |
4175 |
ceph_check_caps(ci, 0, NULL); |
3e1d0452e ceph: avoid iput_... |
4176 4177 |
/* avoid calling iput_final() in tick thread */ ceph_async_iput(inode); |
585d72f33 ceph: clean up an... |
4178 |
spin_lock(&mdsc->cap_delay_lock); |
4b9f2042f ceph: avoid acces... |
4179 |
} |
a8599bd82 ceph: capability ... |
4180 4181 4182 4183 4184 |
} spin_unlock(&mdsc->cap_delay_lock); } /* |
afcdaea3f ceph: flush dirty... |
4185 4186 |
* Flush all dirty caps to the mds */ |
1cf03a68e ceph: convert mds... |
4187 |
static void flush_dirty_session_caps(struct ceph_mds_session *s) |
afcdaea3f ceph: flush dirty... |
4188 |
{ |
1cf03a68e ceph: convert mds... |
4189 |
struct ceph_mds_client *mdsc = s->s_mdsc; |
db3540522 ceph: fix cap flu... |
4190 4191 |
struct ceph_inode_info *ci; struct inode *inode; |
afcdaea3f ceph: flush dirty... |
4192 4193 4194 4195 |
dout("flush_dirty_caps "); spin_lock(&mdsc->cap_dirty_lock); |
1cf03a68e ceph: convert mds... |
4196 4197 |
while (!list_empty(&s->s_cap_dirty)) { ci = list_first_entry(&s->s_cap_dirty, struct ceph_inode_info, |
db3540522 ceph: fix cap flu... |
4198 |
i_dirty_item); |
70b666c3b ceph: use ihold w... |
4199 4200 |
inode = &ci->vfs_inode; ihold(inode); |
db3540522 ceph: fix cap flu... |
4201 4202 |
dout("flush_dirty_caps %p ", inode); |
afcdaea3f ceph: flush dirty... |
4203 |
spin_unlock(&mdsc->cap_dirty_lock); |
a0d93e327 ceph: remove dela... |
4204 |
ceph_check_caps(ci, CHECK_CAPS_FLUSH, NULL); |
70b666c3b ceph: use ihold w... |
4205 |
iput(inode); |
afcdaea3f ceph: flush dirty... |
4206 4207 4208 |
spin_lock(&mdsc->cap_dirty_lock); } spin_unlock(&mdsc->cap_dirty_lock); |
db3540522 ceph: fix cap flu... |
4209 4210 |
dout("flush_dirty_caps done "); |
afcdaea3f ceph: flush dirty... |
4211 |
} |
1cf03a68e ceph: convert mds... |
4212 4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 4231 4232 4233 4234 4235 4236 4237 4238 4239 |
static void iterate_sessions(struct ceph_mds_client *mdsc, void (*cb)(struct ceph_mds_session *)) { int mds; mutex_lock(&mdsc->mutex); for (mds = 0; mds < mdsc->max_sessions; ++mds) { struct ceph_mds_session *s; if (!mdsc->sessions[mds]) continue; s = ceph_get_mds_session(mdsc->sessions[mds]); if (!s) continue; mutex_unlock(&mdsc->mutex); cb(s); ceph_put_mds_session(s); mutex_lock(&mdsc->mutex); } mutex_unlock(&mdsc->mutex); } void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc) { iterate_sessions(mdsc, flush_dirty_session_caps); } |
719a2514e ceph: consider in... |
4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 |
void __ceph_touch_fmode(struct ceph_inode_info *ci, struct ceph_mds_client *mdsc, int fmode) { unsigned long now = jiffies; if (fmode & CEPH_FILE_MODE_RD) ci->i_last_rd = now; if (fmode & CEPH_FILE_MODE_WR) ci->i_last_wr = now; /* queue periodic check */ if (fmode && __ceph_is_any_real_caps(ci) && list_empty(&ci->i_cap_delay_list)) |
a0d93e327 ceph: remove dela... |
4252 |
__cap_delay_requeue(mdsc, ci); |
719a2514e ceph: consider in... |
4253 4254 4255 4256 |
} void ceph_get_fmode(struct ceph_inode_info *ci, int fmode, int count) { |
1dd8d4708 ceph: metrics for... |
4257 |
struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(ci->vfs_inode.i_sb); |
719a2514e ceph: consider in... |
4258 |
int bits = (fmode << 1) | 1; |
1dd8d4708 ceph: metrics for... |
4259 4260 4261 4262 4263 |
bool is_opened = false; int i; if (count == 1) atomic64_inc(&mdsc->metric.opened_files); |
719a2514e ceph: consider in... |
4264 4265 4266 4267 |
spin_lock(&ci->i_ceph_lock); for (i = 0; i < CEPH_FILE_MODE_BITS; i++) { if (bits & (1 << i)) ci->i_nr_by_mode[i] += count; |
1dd8d4708 ceph: metrics for... |
4268 4269 4270 4271 4272 4273 4274 4275 |
/* * If any of the mode ref is larger than 1, * that means it has been already opened by * others. Just skip checking the PIN ref. */ if (i && ci->i_nr_by_mode[i] > 1) is_opened = true; |
719a2514e ceph: consider in... |
4276 |
} |
1dd8d4708 ceph: metrics for... |
4277 4278 4279 |
if (!is_opened) percpu_counter_inc(&mdsc->metric.opened_inodes); |
719a2514e ceph: consider in... |
4280 4281 |
spin_unlock(&ci->i_ceph_lock); } |
afcdaea3f ceph: flush dirty... |
4282 |
/* |
a8599bd82 ceph: capability ... |
4283 4284 4285 4286 |
* Drop open file reference. If we were the last open file, * we may need to release capabilities to the MDS (or schedule * their delayed release). */ |
719a2514e ceph: consider in... |
4287 |
void ceph_put_fmode(struct ceph_inode_info *ci, int fmode, int count) |
a8599bd82 ceph: capability ... |
4288 |
{ |
1dd8d4708 ceph: metrics for... |
4289 |
struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(ci->vfs_inode.i_sb); |
774a6a118 ceph: reduce i_nr... |
4290 |
int bits = (fmode << 1) | 1; |
1dd8d4708 ceph: metrics for... |
4291 4292 4293 4294 4295 |
bool is_closed = true; int i; if (count == 1) atomic64_dec(&mdsc->metric.opened_files); |
be655596b ceph: use i_ceph_... |
4296 |
spin_lock(&ci->i_ceph_lock); |
774a6a118 ceph: reduce i_nr... |
4297 4298 |
for (i = 0; i < CEPH_FILE_MODE_BITS; i++) { if (bits & (1 << i)) { |
719a2514e ceph: consider in... |
4299 4300 |
BUG_ON(ci->i_nr_by_mode[i] < count); ci->i_nr_by_mode[i] -= count; |
774a6a118 ceph: reduce i_nr... |
4301 |
} |
1dd8d4708 ceph: metrics for... |
4302 4303 4304 4305 4306 4307 4308 4309 |
/* * If any of the mode ref is not 0 after * decreased, that means it is still opened * by others. Just skip checking the PIN ref. */ if (i && ci->i_nr_by_mode[i]) is_closed = false; |
774a6a118 ceph: reduce i_nr... |
4310 |
} |
1dd8d4708 ceph: metrics for... |
4311 4312 4313 |
if (is_closed) percpu_counter_dec(&mdsc->metric.opened_inodes); |
be655596b ceph: use i_ceph_... |
4314 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
4315 4316 4317 |
} /* |
a452bc063 ceph: fix comment... |
4318 |
* For a soon-to-be unlinked file, drop the LINK caps. If it |
6ef0bc6dd ceph: flush dirty... |
4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 |
* looks like the link count will hit 0, drop any other caps (other * than PIN) we don't specifically want (due to the file still being * open). */ int ceph_drop_caps_for_unlink(struct inode *inode) { struct ceph_inode_info *ci = ceph_inode(inode); int drop = CEPH_CAP_LINK_SHARED | CEPH_CAP_LINK_EXCL; spin_lock(&ci->i_ceph_lock); if (inode->i_nlink == 1) { drop |= ~(__ceph_caps_wanted(ci) | CEPH_CAP_PIN); |
6ef0bc6dd ceph: flush dirty... |
4331 4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 |
if (__ceph_caps_dirty(ci)) { struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; __cap_delay_requeue_front(mdsc, ci); } } spin_unlock(&ci->i_ceph_lock); return drop; } /* |
a8599bd82 ceph: capability ... |
4342 4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 |
* Helpers for embedding cap and dentry lease releases into mds * requests. * * @force is used by dentry_release (below) to force inclusion of a * record for the directory inode, even when there aren't any caps to * drop. */ int ceph_encode_inode_release(void **p, struct inode *inode, int mds, int drop, int unless, int force) { struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_cap *cap; struct ceph_mds_request_release *rel = *p; |
ec97f88ba ceph: only releas... |
4355 |
int used, dirty; |
a8599bd82 ceph: capability ... |
4356 |
int ret = 0; |
a8599bd82 ceph: capability ... |
4357 |
|
be655596b ceph: use i_ceph_... |
4358 |
spin_lock(&ci->i_ceph_lock); |
916623da1 ceph: only releas... |
4359 |
used = __ceph_caps_used(ci); |
ec97f88ba ceph: only releas... |
4360 |
dirty = __ceph_caps_dirty(ci); |
916623da1 ceph: only releas... |
4361 |
|
ec97f88ba ceph: only releas... |
4362 4363 4364 |
dout("encode_inode_release %p mds%d used|dirty %s drop %s unless %s ", inode, mds, ceph_cap_string(used|dirty), ceph_cap_string(drop), |
916623da1 ceph: only releas... |
4365 |
ceph_cap_string(unless)); |
ec97f88ba ceph: only releas... |
4366 4367 |
/* only drop unused, clean caps */ drop &= ~(used | dirty); |
916623da1 ceph: only releas... |
4368 |
|
a8599bd82 ceph: capability ... |
4369 4370 |
cap = __get_cap_for_mds(ci, mds); if (cap && __cap_is_valid(cap)) { |
222b7f90b ceph: voluntarily... |
4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384 |
unless &= cap->issued; if (unless) { if (unless & CEPH_CAP_AUTH_EXCL) drop &= ~CEPH_CAP_AUTH_SHARED; if (unless & CEPH_CAP_LINK_EXCL) drop &= ~CEPH_CAP_LINK_SHARED; if (unless & CEPH_CAP_XATTR_EXCL) drop &= ~CEPH_CAP_XATTR_SHARED; if (unless & CEPH_CAP_FILE_EXCL) drop &= ~CEPH_CAP_FILE_SHARED; } if (force || (cap->issued & drop)) { if (cap->issued & drop) { |
bb137f84d ceph: fix cap rel... |
4385 |
int wanted = __ceph_caps_wanted(ci); |
bb137f84d ceph: fix cap rel... |
4386 4387 4388 |
dout("encode_inode_release %p cap %p " "%s -> %s, wanted %s -> %s ", inode, cap, |
a8599bd82 ceph: capability ... |
4389 |
ceph_cap_string(cap->issued), |
bb137f84d ceph: fix cap rel... |
4390 4391 4392 |
ceph_cap_string(cap->issued & ~drop), ceph_cap_string(cap->mds_wanted), ceph_cap_string(wanted)); |
a8599bd82 ceph: capability ... |
4393 4394 |
cap->issued &= ~drop; cap->implemented &= ~drop; |
bb137f84d ceph: fix cap rel... |
4395 |
cap->mds_wanted = wanted; |
6f05b30ea ceph: reset i_req... |
4396 4397 4398 |
if (cap == ci->i_auth_cap && !(wanted & CEPH_CAP_ANY_FILE_WR)) ci->i_requested_max_size = 0; |
a8599bd82 ceph: capability ... |
4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 |
} else { dout("encode_inode_release %p cap %p %s" " (force) ", inode, cap, ceph_cap_string(cap->issued)); } rel->ino = cpu_to_le64(ceph_ino(inode)); rel->cap_id = cpu_to_le64(cap->cap_id); rel->seq = cpu_to_le32(cap->seq); |
08a0f24e4 ceph: replace com... |
4409 |
rel->issue_seq = cpu_to_le32(cap->issue_seq); |
a8599bd82 ceph: capability ... |
4410 |
rel->mseq = cpu_to_le32(cap->mseq); |
fd7b95cd1 ceph: avoid relea... |
4411 |
rel->caps = cpu_to_le32(cap->implemented); |
a8599bd82 ceph: capability ... |
4412 4413 4414 4415 4416 4417 |
rel->wanted = cpu_to_le32(cap->mds_wanted); rel->dname_len = 0; rel->dname_seq = 0; *p += sizeof(*rel); ret = 1; } else { |
222b7f90b ceph: voluntarily... |
4418 4419 |
dout("encode_inode_release %p cap %p %s (noop) ", |
a8599bd82 ceph: capability ... |
4420 4421 4422 |
inode, cap, ceph_cap_string(cap->issued)); } } |
be655596b ceph: use i_ceph_... |
4423 |
spin_unlock(&ci->i_ceph_lock); |
a8599bd82 ceph: capability ... |
4424 4425 4426 4427 |
return ret; } int ceph_encode_dentry_release(void **p, struct dentry *dentry, |
ca6c8ae0f ceph: pass parent... |
4428 |
struct inode *dir, |
a8599bd82 ceph: capability ... |
4429 4430 |
int mds, int drop, int unless) { |
ca6c8ae0f ceph: pass parent... |
4431 |
struct dentry *parent = NULL; |
a8599bd82 ceph: capability ... |
4432 4433 4434 4435 4436 4437 4438 |
struct ceph_mds_request_release *rel = *p; struct ceph_dentry_info *di = ceph_dentry(dentry); int force = 0; int ret; /* * force an record for the directory caps if we have a dentry lease. |
be655596b ceph: use i_ceph_... |
4439 |
* this is racy (can't take i_ceph_lock and d_lock together), but it |
a8599bd82 ceph: capability ... |
4440 4441 4442 4443 4444 4445 |
* doesn't have to be perfect; the mds will revoke anything we don't * release. */ spin_lock(&dentry->d_lock); if (di->lease_session && di->lease_session->s_mds == mds) force = 1; |
ca6c8ae0f ceph: pass parent... |
4446 4447 4448 4449 |
if (!dir) { parent = dget(dentry->d_parent); dir = d_inode(parent); } |
a8599bd82 ceph: capability ... |
4450 |
spin_unlock(&dentry->d_lock); |
ca6c8ae0f ceph: pass parent... |
4451 |
ret = ceph_encode_inode_release(p, dir, mds, drop, unless, force); |
adf0d6870 ceph: fix unsafe ... |
4452 |
dput(parent); |
a8599bd82 ceph: capability ... |
4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 |
spin_lock(&dentry->d_lock); if (ret && di->lease_session && di->lease_session->s_mds == mds) { dout("encode_dentry_release %p mds%d seq %d ", dentry, mds, (int)di->lease_seq); rel->dname_len = cpu_to_le32(dentry->d_name.len); memcpy(*p, dentry->d_name.name, dentry->d_name.len); *p += dentry->d_name.len; rel->dname_seq = cpu_to_le32(di->lease_seq); |
1dadcce35 ceph: fix dentry ... |
4463 |
__ceph_mdsc_drop_dentry_lease(dentry); |
a8599bd82 ceph: capability ... |
4464 4465 4466 4467 |
} spin_unlock(&dentry->d_lock); return ret; } |