Blame view
fs/xfs/xfs_log.c
106 KB
0b61f8a40 xfs: convert to S... |
1 |
// SPDX-License-Identifier: GPL-2.0 |
1da177e4c Linux-2.6.12-rc2 |
2 |
/* |
7b7187698 [XFS] Update lice... |
3 4 |
* Copyright (c) 2000-2005 Silicon Graphics, Inc. * All Rights Reserved. |
1da177e4c Linux-2.6.12-rc2 |
5 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
6 |
#include "xfs.h" |
a844f4510 [XFS] Remove xfs_... |
7 |
#include "xfs_fs.h" |
70a9883c5 xfs: create a sha... |
8 |
#include "xfs_shared.h" |
a4fbe6ab1 xfs: decouple ino... |
9 |
#include "xfs_format.h" |
239880ef6 xfs: decouple log... |
10 11 |
#include "xfs_log_format.h" #include "xfs_trans_resv.h" |
1da177e4c Linux-2.6.12-rc2 |
12 |
#include "xfs_mount.h" |
e9e899a2a xfs: move error i... |
13 |
#include "xfs_errortag.h" |
1da177e4c Linux-2.6.12-rc2 |
14 |
#include "xfs_error.h" |
239880ef6 xfs: decouple log... |
15 16 17 |
#include "xfs_trans.h" #include "xfs_trans_priv.h" #include "xfs_log.h" |
1da177e4c Linux-2.6.12-rc2 |
18 |
#include "xfs_log_priv.h" |
0b1b213fc xfs: event tracin... |
19 |
#include "xfs_trace.h" |
baff4e44b xfs: add xlog sys... |
20 |
#include "xfs_sysfs.h" |
61e63ecb5 xfs: consolidate ... |
21 |
#include "xfs_sb.h" |
39353ff6e xfs: replace the ... |
22 |
#include "xfs_health.h" |
1da177e4c Linux-2.6.12-rc2 |
23 |
|
eb01c9cd8 [XFS] Remove the ... |
24 |
kmem_zone_t *xfs_log_ticket_zone; |
1da177e4c Linux-2.6.12-rc2 |
25 |
|
1da177e4c Linux-2.6.12-rc2 |
26 |
/* Local miscellaneous function prototypes */ |
9a8d2fdbb xfs: remove xlog_... |
27 28 29 30 31 32 |
STATIC struct xlog * xlog_alloc_log( struct xfs_mount *mp, struct xfs_buftarg *log_target, xfs_daddr_t blk_offset, int num_bblks); |
ad223e603 xfs: rename log s... |
33 34 35 36 |
STATIC int xlog_space_left( struct xlog *log, atomic64_t *head); |
9a8d2fdbb xfs: remove xlog_... |
37 38 39 |
STATIC void xlog_dealloc_log( struct xlog *log); |
1da177e4c Linux-2.6.12-rc2 |
40 41 |
/* local state machine functions */ |
d15cbf2f3 xfs: stop using X... |
42 |
STATIC void xlog_state_done_syncing( |
12e6a0f44 xfs: remove the a... |
43 |
struct xlog_in_core *iclog); |
9a8d2fdbb xfs: remove xlog_... |
44 45 46 47 48 49 50 51 |
STATIC int xlog_state_get_iclog_space( struct xlog *log, int len, struct xlog_in_core **iclog, struct xlog_ticket *ticket, int *continued_write, int *logoffsetp); |
9a8d2fdbb xfs: remove xlog_... |
52 53 54 55 56 57 |
STATIC void xlog_state_switch_iclogs( struct xlog *log, struct xlog_in_core *iclog, int eventual_size); STATIC void |
ad223e603 xfs: rename log s... |
58 |
xlog_grant_push_ail( |
9a8d2fdbb xfs: remove xlog_... |
59 60 61 |
struct xlog *log, int need_bytes); STATIC void |
df732b29c xfs: call xlog_st... |
62 63 64 |
xlog_sync( struct xlog *log, struct xlog_in_core *iclog); |
cfcbbbd08 [XFS] Remove old,... |
65 |
#if defined(DEBUG) |
9a8d2fdbb xfs: remove xlog_... |
66 67 68 |
STATIC void xlog_verify_dest_ptr( struct xlog *log, |
5809d5e08 xfs: use void poi... |
69 |
void *ptr); |
ad223e603 xfs: rename log s... |
70 71 |
STATIC void xlog_verify_grant_tail( |
9a8d2fdbb xfs: remove xlog_... |
72 73 74 75 76 |
struct xlog *log); STATIC void xlog_verify_iclog( struct xlog *log, struct xlog_in_core *iclog, |
abca1f33f xfs: remove the s... |
77 |
int count); |
9a8d2fdbb xfs: remove xlog_... |
78 79 80 81 82 |
STATIC void xlog_verify_tail_lsn( struct xlog *log, struct xlog_in_core *iclog, xfs_lsn_t tail_lsn); |
1da177e4c Linux-2.6.12-rc2 |
83 84 |
#else #define xlog_verify_dest_ptr(a,b) |
3f336c6fa xfs: fact out com... |
85 |
#define xlog_verify_grant_tail(a) |
abca1f33f xfs: remove the s... |
86 |
#define xlog_verify_iclog(a,b,c) |
1da177e4c Linux-2.6.12-rc2 |
87 88 |
#define xlog_verify_tail_lsn(a,b,c) #endif |
9a8d2fdbb xfs: remove xlog_... |
89 90 91 |
STATIC int xlog_iclogs_empty( struct xlog *log); |
1da177e4c Linux-2.6.12-rc2 |
92 |
|
dd954c69d [XFS] turn xlog h... |
93 |
static void |
663e496a7 xfs: rework log g... |
94 |
xlog_grant_sub_space( |
ad223e603 xfs: rename log s... |
95 96 97 |
struct xlog *log, atomic64_t *head, int bytes) |
dd954c69d [XFS] turn xlog h... |
98 |
{ |
d0eb2f38b xfs: convert gran... |
99 100 |
int64_t head_val = atomic64_read(head); int64_t new, old; |
a69ed03c2 xfs: combine gran... |
101 |
|
d0eb2f38b xfs: convert gran... |
102 103 |
do { int cycle, space; |
a69ed03c2 xfs: combine gran... |
104 |
|
d0eb2f38b xfs: convert gran... |
105 |
xlog_crack_grant_head_val(head_val, &cycle, &space); |
a69ed03c2 xfs: combine gran... |
106 |
|
d0eb2f38b xfs: convert gran... |
107 108 109 110 111 112 113 114 115 116 |
space -= bytes; if (space < 0) { space += log->l_logsize; cycle--; } old = head_val; new = xlog_assign_grant_head_val(cycle, space); head_val = atomic64_cmpxchg(head, old, new); } while (head_val != old); |
dd954c69d [XFS] turn xlog h... |
117 118 119 |
} static void |
663e496a7 xfs: rework log g... |
120 |
xlog_grant_add_space( |
ad223e603 xfs: rename log s... |
121 122 123 |
struct xlog *log, atomic64_t *head, int bytes) |
dd954c69d [XFS] turn xlog h... |
124 |
{ |
d0eb2f38b xfs: convert gran... |
125 126 |
int64_t head_val = atomic64_read(head); int64_t new, old; |
a69ed03c2 xfs: combine gran... |
127 |
|
d0eb2f38b xfs: convert gran... |
128 129 130 |
do { int tmp; int cycle, space; |
a69ed03c2 xfs: combine gran... |
131 |
|
d0eb2f38b xfs: convert gran... |
132 |
xlog_crack_grant_head_val(head_val, &cycle, &space); |
a69ed03c2 xfs: combine gran... |
133 |
|
d0eb2f38b xfs: convert gran... |
134 135 136 137 138 139 140 141 142 143 144 145 |
tmp = log->l_logsize - space; if (tmp > bytes) space += bytes; else { space = bytes - tmp; cycle++; } old = head_val; new = xlog_assign_grant_head_val(cycle, space); head_val = atomic64_cmpxchg(head, old, new); } while (head_val != old); |
dd954c69d [XFS] turn xlog h... |
146 |
} |
a69ed03c2 xfs: combine gran... |
147 |
|
c303c5b8c xfs: add xlog_gra... |
148 149 150 151 152 153 154 155 |
STATIC void xlog_grant_head_init( struct xlog_grant_head *head) { xlog_assign_grant_head(&head->grant, 1, 0); INIT_LIST_HEAD(&head->waiters); spin_lock_init(&head->lock); } |
a79bf2d75 xfs: add xlog_gra... |
156 157 158 159 160 161 162 163 164 165 166 |
STATIC void xlog_grant_head_wake_all( struct xlog_grant_head *head) { struct xlog_ticket *tic; spin_lock(&head->lock); list_for_each_entry(tic, &head->waiters, t_queue) wake_up_process(tic->t_task); spin_unlock(&head->lock); } |
e179840d7 xfs: share code f... |
167 168 |
static inline int xlog_ticket_reservation( |
ad223e603 xfs: rename log s... |
169 |
struct xlog *log, |
e179840d7 xfs: share code f... |
170 171 |
struct xlog_grant_head *head, struct xlog_ticket *tic) |
9f9c19ec1 xfs: fix the logs... |
172 |
{ |
e179840d7 xfs: share code f... |
173 174 175 176 |
if (head == &log->l_write_head) { ASSERT(tic->t_flags & XLOG_TIC_PERM_RESERV); return tic->t_unit_res; } else { |
9f9c19ec1 xfs: fix the logs... |
177 |
if (tic->t_flags & XLOG_TIC_PERM_RESERV) |
e179840d7 xfs: share code f... |
178 |
return tic->t_unit_res * tic->t_cnt; |
9f9c19ec1 xfs: fix the logs... |
179 |
else |
e179840d7 xfs: share code f... |
180 |
return tic->t_unit_res; |
9f9c19ec1 xfs: fix the logs... |
181 |
} |
9f9c19ec1 xfs: fix the logs... |
182 183 184 |
} STATIC bool |
e179840d7 xfs: share code f... |
185 |
xlog_grant_head_wake( |
ad223e603 xfs: rename log s... |
186 |
struct xlog *log, |
e179840d7 xfs: share code f... |
187 |
struct xlog_grant_head *head, |
9f9c19ec1 xfs: fix the logs... |
188 189 190 191 |
int *free_bytes) { struct xlog_ticket *tic; int need_bytes; |
7c107afb8 xfs: push the AIL... |
192 |
bool woken_task = false; |
9f9c19ec1 xfs: fix the logs... |
193 |
|
e179840d7 xfs: share code f... |
194 |
list_for_each_entry(tic, &head->waiters, t_queue) { |
7c107afb8 xfs: push the AIL... |
195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 |
/* * There is a chance that the size of the CIL checkpoints in * progress at the last AIL push target calculation resulted in * limiting the target to the log head (l_last_sync_lsn) at the * time. This may not reflect where the log head is now as the * CIL checkpoints may have completed. * * Hence when we are woken here, it may be that the head of the * log that has moved rather than the tail. As the tail didn't * move, there still won't be space available for the * reservation we require. However, if the AIL has already * pushed to the target defined by the old log head location, we * will hang here waiting for something else to update the AIL * push target. * * Therefore, if there isn't space to wake the first waiter on * the grant head, we need to push the AIL again to ensure the * target reflects both the current log tail and log head * position before we wait for the tail to move again. */ |
e179840d7 xfs: share code f... |
216 |
need_bytes = xlog_ticket_reservation(log, head, tic); |
7c107afb8 xfs: push the AIL... |
217 218 219 |
if (*free_bytes < need_bytes) { if (!woken_task) xlog_grant_push_ail(log, need_bytes); |
9f9c19ec1 xfs: fix the logs... |
220 |
return false; |
7c107afb8 xfs: push the AIL... |
221 |
} |
9f9c19ec1 xfs: fix the logs... |
222 |
|
e179840d7 xfs: share code f... |
223 224 |
*free_bytes -= need_bytes; trace_xfs_log_grant_wake_up(log, tic); |
14a7235fb xfs: remove log s... |
225 |
wake_up_process(tic->t_task); |
7c107afb8 xfs: push the AIL... |
226 |
woken_task = true; |
9f9c19ec1 xfs: fix the logs... |
227 228 229 230 231 232 |
} return true; } STATIC int |
23ee3df34 xfs: share code f... |
233 |
xlog_grant_head_wait( |
ad223e603 xfs: rename log s... |
234 |
struct xlog *log, |
23ee3df34 xfs: share code f... |
235 |
struct xlog_grant_head *head, |
9f9c19ec1 xfs: fix the logs... |
236 |
struct xlog_ticket *tic, |
a30b03679 xfs: fix some min... |
237 238 |
int need_bytes) __releases(&head->lock) __acquires(&head->lock) |
9f9c19ec1 xfs: fix the logs... |
239 |
{ |
23ee3df34 xfs: share code f... |
240 |
list_add_tail(&tic->t_queue, &head->waiters); |
9f9c19ec1 xfs: fix the logs... |
241 242 243 244 245 |
do { if (XLOG_FORCED_SHUTDOWN(log)) goto shutdown; xlog_grant_push_ail(log, need_bytes); |
14a7235fb xfs: remove log s... |
246 |
__set_current_state(TASK_UNINTERRUPTIBLE); |
23ee3df34 xfs: share code f... |
247 |
spin_unlock(&head->lock); |
14a7235fb xfs: remove log s... |
248 |
|
ff6d6af23 xfs: per-filesyst... |
249 |
XFS_STATS_INC(log->l_mp, xs_sleep_logspace); |
9f9c19ec1 xfs: fix the logs... |
250 |
|
14a7235fb xfs: remove log s... |
251 252 |
trace_xfs_log_grant_sleep(log, tic); schedule(); |
9f9c19ec1 xfs: fix the logs... |
253 |
trace_xfs_log_grant_wake(log, tic); |
23ee3df34 xfs: share code f... |
254 |
spin_lock(&head->lock); |
9f9c19ec1 xfs: fix the logs... |
255 256 |
if (XLOG_FORCED_SHUTDOWN(log)) goto shutdown; |
23ee3df34 xfs: share code f... |
257 |
} while (xlog_space_left(log, &head->grant) < need_bytes); |
9f9c19ec1 xfs: fix the logs... |
258 259 260 261 262 |
list_del_init(&tic->t_queue); return 0; shutdown: list_del_init(&tic->t_queue); |
2451337dd xfs: global error... |
263 |
return -EIO; |
9f9c19ec1 xfs: fix the logs... |
264 |
} |
42ceedb3c xfs: share code f... |
265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 |
/* * Atomically get the log space required for a log ticket. * * Once a ticket gets put onto head->waiters, it will only return after the * needed reservation is satisfied. * * This function is structured so that it has a lock free fast path. This is * necessary because every new transaction reservation will come through this * path. Hence any lock will be globally hot if we take it unconditionally on * every pass. * * As tickets are only ever moved on and off head->waiters under head->lock, we * only need to take that lock if we are going to add the ticket to the queue * and sleep. We can avoid taking the lock if the ticket was never added to * head->waiters because the t_queue list head will be empty and we hold the * only reference to it so it can safely be checked unlocked. */ STATIC int xlog_grant_head_check( |
ad223e603 xfs: rename log s... |
284 |
struct xlog *log, |
42ceedb3c xfs: share code f... |
285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 |
struct xlog_grant_head *head, struct xlog_ticket *tic, int *need_bytes) { int free_bytes; int error = 0; ASSERT(!(log->l_flags & XLOG_ACTIVE_RECOVERY)); /* * If there are other waiters on the queue then give them a chance at * logspace before us. Wake up the first waiters, if we do not wake * up all the waiters then go to sleep waiting for more free space, * otherwise try to get some space for this transaction. */ *need_bytes = xlog_ticket_reservation(log, head, tic); free_bytes = xlog_space_left(log, &head->grant); if (!list_empty_careful(&head->waiters)) { spin_lock(&head->lock); if (!xlog_grant_head_wake(log, head, &free_bytes) || free_bytes < *need_bytes) { error = xlog_grant_head_wait(log, head, tic, *need_bytes); } spin_unlock(&head->lock); } else if (free_bytes < *need_bytes) { spin_lock(&head->lock); error = xlog_grant_head_wait(log, head, tic, *need_bytes); spin_unlock(&head->lock); } return error; } |
0adba5363 [XFS] replace som... |
318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 |
static void xlog_tic_reset_res(xlog_ticket_t *tic) { tic->t_res_num = 0; tic->t_res_arr_sum = 0; tic->t_res_num_ophdrs = 0; } static void xlog_tic_add_region(xlog_ticket_t *tic, uint len, uint type) { if (tic->t_res_num == XLOG_TIC_LEN_MAX) { /* add to overflow and start again */ tic->t_res_o_flow += tic->t_res_arr_sum; tic->t_res_num = 0; tic->t_res_arr_sum = 0; } tic->t_res_arr[tic->t_res_num].r_len = len; tic->t_res_arr[tic->t_res_num].r_type = type; tic->t_res_arr_sum += len; tic->t_res_num++; } |
dd954c69d [XFS] turn xlog h... |
341 |
|
1da177e4c Linux-2.6.12-rc2 |
342 |
/* |
9006fb91c xfs: split and cl... |
343 344 345 346 347 348 349 |
* Replenish the byte reservation required by moving the grant write head. */ int xfs_log_regrant( struct xfs_mount *mp, struct xlog_ticket *tic) { |
ad223e603 xfs: rename log s... |
350 |
struct xlog *log = mp->m_log; |
9006fb91c xfs: split and cl... |
351 352 353 354 |
int need_bytes; int error = 0; if (XLOG_FORCED_SHUTDOWN(log)) |
2451337dd xfs: global error... |
355 |
return -EIO; |
9006fb91c xfs: split and cl... |
356 |
|
ff6d6af23 xfs: per-filesyst... |
357 |
XFS_STATS_INC(mp, xs_try_logspace); |
9006fb91c xfs: split and cl... |
358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 |
/* * This is a new transaction on the ticket, so we need to change the * transaction ID so that the next transaction has a different TID in * the log. Just add one to the existing tid so that we can see chains * of rolling transactions in the log easily. */ tic->t_tid++; xlog_grant_push_ail(log, tic->t_unit_res); tic->t_curr_res = tic->t_unit_res; xlog_tic_reset_res(tic); if (tic->t_cnt > 0) return 0; trace_xfs_log_regrant(log, tic); error = xlog_grant_head_check(log, &log->l_write_head, tic, &need_bytes); if (error) goto out_error; xlog_grant_add_space(log, &log->l_write_head.grant, need_bytes); trace_xfs_log_regrant_exit(log, tic); xlog_verify_grant_tail(log); return 0; out_error: /* * If we are failing, make sure the ticket doesn't have any current * reservations. We don't want to add this back when the ticket/ * transaction gets cancelled. */ tic->t_curr_res = 0; tic->t_cnt = 0; /* ungrant will give back unit_res * t_cnt. */ return error; } /* |
a0e336ba3 xfs: fix a commen... |
399 |
* Reserve log space and return a ticket corresponding to the reservation. |
9006fb91c xfs: split and cl... |
400 401 402 403 404 405 406 407 408 409 410 411 |
* * Each reservation is going to reserve extra space for a log record header. * When writes happen to the on-disk log, we don't subtract the length of the * log record header from any reservation. By wasting space in each * reservation, we prevent over allocation problems. */ int xfs_log_reserve( struct xfs_mount *mp, int unit_bytes, int cnt, struct xlog_ticket **ticp, |
c8ce540db xfs: remove doubl... |
412 |
uint8_t client, |
710b1e2c2 xfs: remove trans... |
413 |
bool permanent) |
9006fb91c xfs: split and cl... |
414 |
{ |
ad223e603 xfs: rename log s... |
415 |
struct xlog *log = mp->m_log; |
9006fb91c xfs: split and cl... |
416 417 418 419 420 421 422 |
struct xlog_ticket *tic; int need_bytes; int error = 0; ASSERT(client == XFS_TRANSACTION || client == XFS_LOG); if (XLOG_FORCED_SHUTDOWN(log)) |
2451337dd xfs: global error... |
423 |
return -EIO; |
9006fb91c xfs: split and cl... |
424 |
|
ff6d6af23 xfs: per-filesyst... |
425 |
XFS_STATS_INC(mp, xs_try_logspace); |
9006fb91c xfs: split and cl... |
426 427 |
ASSERT(*ticp == NULL); |
ca4f25899 xfs: Modify xlog_... |
428 |
tic = xlog_ticket_alloc(log, unit_bytes, cnt, client, permanent); |
9006fb91c xfs: split and cl... |
429 |
*ticp = tic; |
437a255aa xfs: fix direct I... |
430 431 |
xlog_grant_push_ail(log, tic->t_cnt ? tic->t_unit_res * tic->t_cnt : tic->t_unit_res); |
9006fb91c xfs: split and cl... |
432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 |
trace_xfs_log_reserve(log, tic); error = xlog_grant_head_check(log, &log->l_reserve_head, tic, &need_bytes); if (error) goto out_error; xlog_grant_add_space(log, &log->l_reserve_head.grant, need_bytes); xlog_grant_add_space(log, &log->l_write_head.grant, need_bytes); trace_xfs_log_reserve_exit(log, tic); xlog_verify_grant_tail(log); return 0; out_error: /* * If we are failing, make sure the ticket doesn't have any current * reservations. We don't want to add this back when the ticket/ * transaction gets cancelled. */ tic->t_curr_res = 0; tic->t_cnt = 0; /* ungrant will give back unit_res * t_cnt. */ return error; } |
df732b29c xfs: call xlog_st... |
456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 |
static bool __xlog_state_release_iclog( struct xlog *log, struct xlog_in_core *iclog) { lockdep_assert_held(&log->l_icloglock); if (iclog->ic_state == XLOG_STATE_WANT_SYNC) { /* update tail before writing to iclog */ xfs_lsn_t tail_lsn = xlog_assign_tail_lsn(log->l_mp); iclog->ic_state = XLOG_STATE_SYNCING; iclog->ic_header.h_tail_lsn = cpu_to_be64(tail_lsn); xlog_verify_tail_lsn(log, iclog, tail_lsn); /* cycle incremented when incrementing curr_block */ return true; } ASSERT(iclog->ic_state == XLOG_STATE_ACTIVE); return false; } /* * Flush iclog to disk if this is the last reference to the given iclog and the * it is in the WANT_SYNC state. */ static int xlog_state_release_iclog( struct xlog *log, struct xlog_in_core *iclog) { lockdep_assert_held(&log->l_icloglock); |
1858bb0be xfs: turn ic_stat... |
488 |
if (iclog->ic_state == XLOG_STATE_IOERROR) |
df732b29c xfs: call xlog_st... |
489 490 491 492 493 494 495 496 497 498 499 |
return -EIO; if (atomic_dec_and_test(&iclog->ic_refcnt) && __xlog_state_release_iclog(log, iclog)) { spin_unlock(&log->l_icloglock); xlog_sync(log, iclog); spin_lock(&log->l_icloglock); } return 0; } |
f97a43e43 xfs: simplify the... |
500 |
void |
35a8a72f0 xfs: stop passing... |
501 |
xfs_log_release_iclog( |
35a8a72f0 xfs: stop passing... |
502 |
struct xlog_in_core *iclog) |
1da177e4c Linux-2.6.12-rc2 |
503 |
{ |
f97a43e43 xfs: simplify the... |
504 |
struct xlog *log = iclog->ic_log; |
a582f32fa xfs: simplify log... |
505 |
bool sync = false; |
1da177e4c Linux-2.6.12-rc2 |
506 |
|
df732b29c xfs: call xlog_st... |
507 |
if (atomic_dec_and_lock(&iclog->ic_refcnt, &log->l_icloglock)) { |
a582f32fa xfs: simplify log... |
508 509 |
if (iclog->ic_state != XLOG_STATE_IOERROR) sync = __xlog_state_release_iclog(log, iclog); |
df732b29c xfs: call xlog_st... |
510 |
spin_unlock(&log->l_icloglock); |
df732b29c xfs: call xlog_st... |
511 |
} |
a582f32fa xfs: simplify log... |
512 513 514 |
if (sync) xlog_sync(log, iclog); |
1da177e4c Linux-2.6.12-rc2 |
515 516 517 |
} /* |
1da177e4c Linux-2.6.12-rc2 |
518 519 520 521 522 523 524 525 526 527 |
* Mount a log filesystem * * mp - ubiquitous xfs mount point structure * log_target - buftarg of on-disk log device * blk_offset - Start block # where block size is 512 bytes (BBSIZE) * num_bblocks - Number of BBSIZE blocks in on-disk log * * Return error or zero. */ int |
249a8c112 [XFS] Move AIL pu... |
528 529 530 531 532 |
xfs_log_mount( xfs_mount_t *mp, xfs_buftarg_t *log_target, xfs_daddr_t blk_offset, int num_bblks) |
1da177e4c Linux-2.6.12-rc2 |
533 |
{ |
9c92ee208 xfs: validate sb_... |
534 |
bool fatal = xfs_sb_version_hascrc(&mp->m_sb); |
3e7b91cf8 xfs: Validate log... |
535 536 |
int error = 0; int min_logfsbs; |
249a8c112 [XFS] Move AIL pu... |
537 |
|
c99d609a1 xfs: fully suppor... |
538 539 540 541 |
if (!(mp->m_flags & XFS_MOUNT_NORECOVERY)) { xfs_notice(mp, "Mounting V%d Filesystem", XFS_SB_VERSION_NUM(&mp->m_sb)); } else { |
a0fa2b679 xfs: Convert xlog... |
542 |
xfs_notice(mp, |
c99d609a1 xfs: fully suppor... |
543 544 |
"Mounting V%d filesystem in no-recovery mode. Filesystem will be inconsistent.", XFS_SB_VERSION_NUM(&mp->m_sb)); |
bd186aa90 [XFS] kill the vf... |
545 |
ASSERT(mp->m_flags & XFS_MOUNT_RDONLY); |
1da177e4c Linux-2.6.12-rc2 |
546 547 548 |
} mp->m_log = xlog_alloc_log(mp, log_target, blk_offset, num_bblks); |
a6cb767e2 xfs: validate log... |
549 |
if (IS_ERR(mp->m_log)) { |
2451337dd xfs: global error... |
550 |
error = PTR_ERR(mp->m_log); |
644c3567d [XFS] handle memo... |
551 552 |
goto out; } |
1da177e4c Linux-2.6.12-rc2 |
553 |
|
1da177e4c Linux-2.6.12-rc2 |
554 |
/* |
3e7b91cf8 xfs: Validate log... |
555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 |
* Validate the given log space and drop a critical message via syslog * if the log size is too small that would lead to some unexpected * situations in transaction log space reservation stage. * * Note: we can't just reject the mount if the validation fails. This * would mean that people would have to downgrade their kernel just to * remedy the situation as there is no way to grow the log (short of * black magic surgery with xfs_db). * * We can, however, reject mounts for CRC format filesystems, as the * mkfs binary being used to make the filesystem should never create a * filesystem with a log that is too small. */ min_logfsbs = xfs_log_calc_minimum_size(mp); if (mp->m_sb.sb_logblocks < min_logfsbs) { xfs_warn(mp, "Log size %d blocks too small, minimum size is %d blocks", mp->m_sb.sb_logblocks, min_logfsbs); |
2451337dd xfs: global error... |
574 |
error = -EINVAL; |
3e7b91cf8 xfs: Validate log... |
575 576 577 578 |
} else if (mp->m_sb.sb_logblocks > XFS_MAX_LOG_BLOCKS) { xfs_warn(mp, "Log size %d blocks too large, maximum size is %lld blocks", mp->m_sb.sb_logblocks, XFS_MAX_LOG_BLOCKS); |
2451337dd xfs: global error... |
579 |
error = -EINVAL; |
3e7b91cf8 xfs: Validate log... |
580 581 582 583 584 |
} else if (XFS_FSB_TO_B(mp, mp->m_sb.sb_logblocks) > XFS_MAX_LOG_BYTES) { xfs_warn(mp, "log size %lld bytes too large, maximum size is %lld bytes", XFS_FSB_TO_B(mp, mp->m_sb.sb_logblocks), XFS_MAX_LOG_BYTES); |
2451337dd xfs: global error... |
585 |
error = -EINVAL; |
9c92ee208 xfs: validate sb_... |
586 587 588 589 590 591 592 |
} else if (mp->m_sb.sb_logsunit > 1 && mp->m_sb.sb_logsunit % mp->m_sb.sb_blocksize) { xfs_warn(mp, "log stripe unit %u bytes must be a multiple of block size", mp->m_sb.sb_logsunit); error = -EINVAL; fatal = true; |
3e7b91cf8 xfs: Validate log... |
593 594 |
} if (error) { |
9c92ee208 xfs: validate sb_... |
595 596 597 598 599 |
/* * Log check errors are always fatal on v5; or whenever bad * metadata leads to a crash. */ if (fatal) { |
3e7b91cf8 xfs: Validate log... |
600 601 602 603 |
xfs_crit(mp, "AAIEEE! Log failed size checks. Abort!"); ASSERT(0); goto out_free_log; } |
f41febd2e xfs: Use consiste... |
604 |
xfs_crit(mp, "Log size out of supported range."); |
3e7b91cf8 xfs: Validate log... |
605 |
xfs_crit(mp, |
f41febd2e xfs: Use consiste... |
606 |
"Continuing onwards, but if log hangs are experienced then please report this message in the bug report."); |
3e7b91cf8 xfs: Validate log... |
607 608 609 |
} /* |
249a8c112 [XFS] Move AIL pu... |
610 611 |
* Initialize the AIL now we have a log. */ |
249a8c112 [XFS] Move AIL pu... |
612 613 |
error = xfs_trans_ail_init(mp); if (error) { |
a0fa2b679 xfs: Convert xlog... |
614 |
xfs_warn(mp, "AIL initialisation failed: error %d", error); |
264307520 xfs: fix error ha... |
615 |
goto out_free_log; |
249a8c112 [XFS] Move AIL pu... |
616 |
} |
a9c21c1b9 [XFS] Given the l... |
617 |
mp->m_log->l_ailp = mp->m_ail; |
249a8c112 [XFS] Move AIL pu... |
618 619 |
/* |
1da177e4c Linux-2.6.12-rc2 |
620 621 622 623 |
* skip log recovery on a norecovery mount. pretend it all * just worked. */ if (!(mp->m_flags & XFS_MOUNT_NORECOVERY)) { |
249a8c112 [XFS] Move AIL pu... |
624 |
int readonly = (mp->m_flags & XFS_MOUNT_RDONLY); |
1da177e4c Linux-2.6.12-rc2 |
625 626 |
if (readonly) |
bd186aa90 [XFS] kill the vf... |
627 |
mp->m_flags &= ~XFS_MOUNT_RDONLY; |
1da177e4c Linux-2.6.12-rc2 |
628 |
|
65be60541 [XFS] remove unus... |
629 |
error = xlog_recover(mp->m_log); |
1da177e4c Linux-2.6.12-rc2 |
630 631 |
if (readonly) |
bd186aa90 [XFS] kill the vf... |
632 |
mp->m_flags |= XFS_MOUNT_RDONLY; |
1da177e4c Linux-2.6.12-rc2 |
633 |
if (error) { |
a0fa2b679 xfs: Convert xlog... |
634 635 |
xfs_warn(mp, "log mount/recovery failed: error %d", error); |
f0b2efad1 xfs: don't leave ... |
636 |
xlog_recover_cancel(mp->m_log); |
264307520 xfs: fix error ha... |
637 |
goto out_destroy_ail; |
1da177e4c Linux-2.6.12-rc2 |
638 639 |
} } |
baff4e44b xfs: add xlog sys... |
640 641 642 643 |
error = xfs_sysfs_init(&mp->m_log->l_kobj, &xfs_log_ktype, &mp->m_kobj, "log"); if (error) goto out_destroy_ail; |
1da177e4c Linux-2.6.12-rc2 |
644 645 |
/* Normal transactions can now occur */ mp->m_log->l_flags &= ~XLOG_ACTIVE_RECOVERY; |
71e330b59 xfs: Introduce de... |
646 647 648 649 650 651 |
/* * Now the log has been fully initialised and we know were our * space grant counters are, we can initialise the permanent ticket * needed for delayed logging to work. */ xlog_cil_init_post_recovery(mp->m_log); |
1da177e4c Linux-2.6.12-rc2 |
652 |
return 0; |
264307520 xfs: fix error ha... |
653 654 655 656 657 |
out_destroy_ail: xfs_trans_ail_destroy(mp); out_free_log: xlog_dealloc_log(mp->m_log); |
644c3567d [XFS] handle memo... |
658 |
out: |
249a8c112 [XFS] Move AIL pu... |
659 |
return error; |
264307520 xfs: fix error ha... |
660 |
} |
1da177e4c Linux-2.6.12-rc2 |
661 662 |
/* |
f661f1e0b xfs: sync work is... |
663 664 665 666 |
* Finish the recovery of the file system. This is separate from the * xfs_log_mount() call, because it depends on the code in xfs_mountfs() to read * in the root and real-time bitmap inodes between calling xfs_log_mount() and * here. |
1da177e4c Linux-2.6.12-rc2 |
667 |
* |
f661f1e0b xfs: sync work is... |
668 669 670 |
* If we finish recovery successfully, start the background log work. If we are * not doing recovery, then we have a RO filesystem and we don't need to start * it. |
1da177e4c Linux-2.6.12-rc2 |
671 672 |
*/ int |
f0b2efad1 xfs: don't leave ... |
673 674 |
xfs_log_mount_finish( struct xfs_mount *mp) |
1da177e4c Linux-2.6.12-rc2 |
675 |
{ |
f661f1e0b xfs: sync work is... |
676 |
int error = 0; |
6f4a1eefd xfs: toggle reado... |
677 |
bool readonly = (mp->m_flags & XFS_MOUNT_RDONLY); |
f1b92bbc2 xfs: drain the bu... |
678 |
bool recovered = mp->m_log->l_flags & XLOG_RECOVERY_NEEDED; |
1da177e4c Linux-2.6.12-rc2 |
679 |
|
f0b2efad1 xfs: don't leave ... |
680 |
if (mp->m_flags & XFS_MOUNT_NORECOVERY) { |
bd186aa90 [XFS] kill the vf... |
681 |
ASSERT(mp->m_flags & XFS_MOUNT_RDONLY); |
f0b2efad1 xfs: don't leave ... |
682 |
return 0; |
6f4a1eefd xfs: toggle reado... |
683 684 685 |
} else if (readonly) { /* Allow unlinked processing to proceed */ mp->m_flags &= ~XFS_MOUNT_RDONLY; |
1da177e4c Linux-2.6.12-rc2 |
686 |
} |
8204f8dda xfs: clear MS_ACT... |
687 688 689 690 691 692 693 694 |
/* * During the second phase of log recovery, we need iget and * iput to behave like they do for an active filesystem. * xfs_fs_drop_inode needs to be able to prevent the deletion * of inodes before we're done replaying log items on those * inodes. Turn it off immediately after recovery finishes * so that we don't leak the quota inodes if subsequent mount * activities fail. |
799ea9e9c xfs: evict all in... |
695 696 697 698 699 700 |
* * We let all inodes involved in redo item processing end up on * the LRU instead of being evicted immediately so that if we do * something to an unlinked inode, the irele won't cause * premature truncation and freeing of the inode, which results * in log recovery failure. We have to evict the unreferenced |
1751e8a6c Rename superblock... |
701 |
* lru inodes after clearing SB_ACTIVE because we don't |
799ea9e9c xfs: evict all in... |
702 703 704 705 |
* otherwise clean up the lru if there's a subsequent failure in * xfs_mountfs, which leads to us leaking the inodes if nothing * else (e.g. quotacheck) references the inodes before the * mount failure occurs. |
8204f8dda xfs: clear MS_ACT... |
706 |
*/ |
1751e8a6c Rename superblock... |
707 |
mp->m_super->s_flags |= SB_ACTIVE; |
f0b2efad1 xfs: don't leave ... |
708 709 710 |
error = xlog_recover_finish(mp->m_log); if (!error) xfs_log_work_queue(mp); |
1751e8a6c Rename superblock... |
711 |
mp->m_super->s_flags &= ~SB_ACTIVE; |
799ea9e9c xfs: evict all in... |
712 |
evict_inodes(mp->m_super); |
f0b2efad1 xfs: don't leave ... |
713 |
|
f1b92bbc2 xfs: drain the bu... |
714 715 716 717 718 719 720 721 722 723 724 725 726 727 |
/* * Drain the buffer LRU after log recovery. This is required for v4 * filesystems to avoid leaving around buffers with NULL verifier ops, * but we do it unconditionally to make sure we're always in a clean * cache state after mount. * * Don't push in the error case because the AIL may have pending intents * that aren't removed until recovery is cancelled. */ if (!error && recovered) { xfs_log_force(mp, XFS_LOG_SYNC); xfs_ail_push_all_sync(mp->m_ail); } xfs_wait_buftarg(mp->m_ddev_targp); |
6f4a1eefd xfs: toggle reado... |
728 729 |
if (readonly) mp->m_flags |= XFS_MOUNT_RDONLY; |
f0b2efad1 xfs: don't leave ... |
730 731 732 733 734 735 736 |
return error; } /* * The mount has failed. Cancel the recovery if it hasn't completed and destroy * the log. */ |
a7a9250e1 fs: xfs: xfs_log:... |
737 |
void |
f0b2efad1 xfs: don't leave ... |
738 739 740 |
xfs_log_mount_cancel( struct xfs_mount *mp) { |
a7a9250e1 fs: xfs: xfs_log:... |
741 |
xlog_recover_cancel(mp->m_log); |
f0b2efad1 xfs: don't leave ... |
742 |
xfs_log_unmount(mp); |
1da177e4c Linux-2.6.12-rc2 |
743 744 745 |
} /* |
81e5b50a8 xfs: factor out a... |
746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 |
* Wait for the iclog to be written disk, or return an error if the log has been * shut down. */ static int xlog_wait_on_iclog( struct xlog_in_core *iclog) __releases(iclog->ic_log->l_icloglock) { struct xlog *log = iclog->ic_log; if (!XLOG_FORCED_SHUTDOWN(log) && iclog->ic_state != XLOG_STATE_ACTIVE && iclog->ic_state != XLOG_STATE_DIRTY) { XFS_STATS_INC(log->l_mp, xs_log_force_sleep); xlog_wait(&iclog->ic_force_wait, &log->l_icloglock); } else { spin_unlock(&log->l_icloglock); } if (XLOG_FORCED_SHUTDOWN(log)) return -EIO; return 0; } /* |
3c702f959 xfs: refactor unm... |
771 772 773 |
* Write out an unmount record using the ticket provided. We have to account for * the data space used in the unmount ticket as this write is not done from a * transaction context that has already done the accounting for us. |
1da177e4c Linux-2.6.12-rc2 |
774 |
*/ |
3c702f959 xfs: refactor unm... |
775 776 777 778 779 780 |
static int xlog_write_unmount_record( struct xlog *log, struct xlog_ticket *ticket, xfs_lsn_t *lsn, uint flags) |
53235f221 xfs: refactor unm... |
781 |
{ |
3c702f959 xfs: refactor unm... |
782 |
struct xfs_unmount_log_format ulf = { |
53235f221 xfs: refactor unm... |
783 784 785 |
.magic = XLOG_UNMOUNT_TYPE, }; struct xfs_log_iovec reg = { |
3c702f959 xfs: refactor unm... |
786 787 |
.i_addr = &ulf, .i_len = sizeof(ulf), |
53235f221 xfs: refactor unm... |
788 789 790 791 792 793 |
.i_type = XLOG_REG_TYPE_UNMOUNT, }; struct xfs_log_vec vec = { .lv_niovecs = 1, .lv_iovecp = ®, }; |
3c702f959 xfs: refactor unm... |
794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 |
/* account for space used by record data */ ticket->t_curr_res -= sizeof(ulf); return xlog_write(log, &vec, ticket, lsn, NULL, flags, false); } /* * Mark the filesystem clean by writing an unmount record to the head of the * log. */ static void xlog_unmount_write( struct xlog *log) { struct xfs_mount *mp = log->l_mp; |
53235f221 xfs: refactor unm... |
809 810 811 |
struct xlog_in_core *iclog; struct xlog_ticket *tic = NULL; xfs_lsn_t lsn; |
f467cad95 xfs: force summar... |
812 |
uint flags = XLOG_UNMOUNT_TRANS; |
53235f221 xfs: refactor unm... |
813 814 815 816 817 |
int error; error = xfs_log_reserve(mp, 600, 1, &tic, XFS_LOG, 0); if (error) goto out_err; |
3c702f959 xfs: refactor unm... |
818 |
error = xlog_write_unmount_record(log, tic, &lsn, flags); |
53235f221 xfs: refactor unm... |
819 820 821 822 823 824 825 826 827 828 829 |
/* * At this point, we're umounting anyway, so there's no point in * transitioning log state to IOERROR. Just continue... */ out_err: if (error) xfs_alert(mp, "%s: unmount record failed", __func__); spin_lock(&log->l_icloglock); iclog = log->l_iclog; atomic_inc(&iclog->ic_refcnt); |
693639994 xfs: remove xlog_... |
830 831 832 833 834 |
if (iclog->ic_state == XLOG_STATE_ACTIVE) xlog_state_switch_iclogs(log, iclog, 0); else ASSERT(iclog->ic_state == XLOG_STATE_WANT_SYNC || iclog->ic_state == XLOG_STATE_IOERROR); |
53235f221 xfs: refactor unm... |
835 |
error = xlog_state_release_iclog(log, iclog); |
81e5b50a8 xfs: factor out a... |
836 |
xlog_wait_on_iclog(iclog); |
53235f221 xfs: refactor unm... |
837 838 839 |
if (tic) { trace_xfs_log_umount_write(log, tic); |
8b41e3f98 xfs: split xlog_t... |
840 |
xfs_log_ticket_ungrant(log, tic); |
53235f221 xfs: refactor unm... |
841 842 |
} } |
13859c984 xfs: cleanup xfs_... |
843 844 845 846 847 848 849 850 851 852 853 |
static void xfs_log_unmount_verify_iclog( struct xlog *log) { struct xlog_in_core *iclog = log->l_iclog; do { ASSERT(iclog->ic_state == XLOG_STATE_ACTIVE); ASSERT(iclog->ic_offset == 0); } while ((iclog = iclog->ic_next) != log->l_iclog); } |
1da177e4c Linux-2.6.12-rc2 |
854 855 856 857 |
/* * Unmount record used to have a string "Unmount filesystem--" in the * data section where the "Un" was really a magic number (XLOG_UNMOUNT_TYPE). * We just write the magic number now since that particular field isn't |
8e159e72e xfs: fix the comm... |
858 |
* currently architecture converted and "Unmount" is a bit foo. |
1da177e4c Linux-2.6.12-rc2 |
859 860 |
* As far as I know, there weren't any dependencies on the old behaviour. */ |
550319e9d xfs: remove the u... |
861 |
static void |
13859c984 xfs: cleanup xfs_... |
862 863 |
xfs_log_unmount_write( struct xfs_mount *mp) |
1da177e4c Linux-2.6.12-rc2 |
864 |
{ |
13859c984 xfs: cleanup xfs_... |
865 |
struct xlog *log = mp->m_log; |
1da177e4c Linux-2.6.12-rc2 |
866 |
|
1da177e4c Linux-2.6.12-rc2 |
867 |
/* |
757a69ef6 xfs: write unmoun... |
868 |
* Don't write out unmount record on norecovery mounts or ro devices. |
1da177e4c Linux-2.6.12-rc2 |
869 870 |
* Or, if we are doing a forced umount (typically because of IO errors). */ |
757a69ef6 xfs: write unmoun... |
871 |
if (mp->m_flags & XFS_MOUNT_NORECOVERY || |
2d15d2c0e xfs: make use of ... |
872 |
xfs_readonly_buftarg(log->l_targ)) { |
757a69ef6 xfs: write unmoun... |
873 |
ASSERT(mp->m_flags & XFS_MOUNT_RDONLY); |
550319e9d xfs: remove the u... |
874 |
return; |
757a69ef6 xfs: write unmoun... |
875 |
} |
1da177e4c Linux-2.6.12-rc2 |
876 |
|
550319e9d xfs: remove the u... |
877 |
xfs_log_force(mp, XFS_LOG_SYNC); |
1da177e4c Linux-2.6.12-rc2 |
878 |
|
6178d1040 xfs: remove dead ... |
879 880 |
if (XLOG_FORCED_SHUTDOWN(log)) return; |
5cc3c006e xfs: don't write ... |
881 882 883 884 885 886 887 888 889 890 891 892 893 |
/* * If we think the summary counters are bad, avoid writing the unmount * record to force log recovery at next mount, after which the summary * counters will be recalculated. Refer to xlog_check_unmount_rec for * more details. */ if (XFS_TEST_ERROR(xfs_fs_has_sickness(mp, XFS_SICK_FS_COUNTERS), mp, XFS_ERRTAG_FORCE_SUMMARY_RECALC)) { xfs_alert(mp, "%s: will fix summary counters at next mount", __func__); return; } |
13859c984 xfs: cleanup xfs_... |
894 |
xfs_log_unmount_verify_iclog(log); |
3c702f959 xfs: refactor unm... |
895 |
xlog_unmount_write(log); |
550319e9d xfs: remove the u... |
896 |
} |
1da177e4c Linux-2.6.12-rc2 |
897 898 |
/* |
c75921a72 xfs: xfs_quiesce_... |
899 |
* Empty the log for unmount/freeze. |
cf2931db2 xfs: Bring some s... |
900 901 902 903 904 |
* * To do this, we first need to shut down the background log work so it is not * trying to cover the log as we clean up. We then need to unpin all objects in * the log so we can then flush them out. Once they have completed their IO and * run the callbacks removing themselves from the AIL, we can write the unmount |
c75921a72 xfs: xfs_quiesce_... |
905 |
* record. |
1da177e4c Linux-2.6.12-rc2 |
906 907 |
*/ void |
c75921a72 xfs: xfs_quiesce_... |
908 909 |
xfs_log_quiesce( struct xfs_mount *mp) |
1da177e4c Linux-2.6.12-rc2 |
910 |
{ |
f661f1e0b xfs: sync work is... |
911 |
cancel_delayed_work_sync(&mp->m_log->l_work); |
cf2931db2 xfs: Bring some s... |
912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 |
xfs_log_force(mp, XFS_LOG_SYNC); /* * The superblock buffer is uncached and while xfs_ail_push_all_sync() * will push it, xfs_wait_buftarg() will not wait for it. Further, * xfs_buf_iowait() cannot be used because it was pushed with the * XBF_ASYNC flag set, so we need to use a lock/unlock pair to wait for * the IO to complete. */ xfs_ail_push_all_sync(mp->m_ail); xfs_wait_buftarg(mp->m_ddev_targp); xfs_buf_lock(mp->m_sb_bp); xfs_buf_unlock(mp->m_sb_bp); xfs_log_unmount_write(mp); |
c75921a72 xfs: xfs_quiesce_... |
927 928 929 930 931 932 933 934 935 936 937 938 939 940 |
} /* * Shut down and release the AIL and Log. * * During unmount, we need to ensure we flush all the dirty metadata objects * from the AIL so that the log is empty before we write the unmount record to * the log. Once this is done, we can tear down the AIL and the log. */ void xfs_log_unmount( struct xfs_mount *mp) { xfs_log_quiesce(mp); |
cf2931db2 xfs: Bring some s... |
941 |
|
249a8c112 [XFS] Move AIL pu... |
942 |
xfs_trans_ail_destroy(mp); |
baff4e44b xfs: add xlog sys... |
943 944 |
xfs_sysfs_del(&mp->m_log->l_kobj); |
c41564b5a [XFS] We really s... |
945 |
xlog_dealloc_log(mp->m_log); |
1da177e4c Linux-2.6.12-rc2 |
946 |
} |
43f5efc5b xfs: factor log i... |
947 948 949 950 951 |
void xfs_log_item_init( struct xfs_mount *mp, struct xfs_log_item *item, int type, |
272e42b21 xfs: constify xfs... |
952 |
const struct xfs_item_ops *ops) |
43f5efc5b xfs: factor log i... |
953 954 955 956 957 |
{ item->li_mountp = mp; item->li_ailp = mp->m_ail; item->li_type = type; item->li_ops = ops; |
71e330b59 xfs: Introduce de... |
958 959 960 961 |
item->li_lv = NULL; INIT_LIST_HEAD(&item->li_ail); INIT_LIST_HEAD(&item->li_cil); |
643c8c05e Use list_head inf... |
962 |
INIT_LIST_HEAD(&item->li_bio_list); |
e6631f855 xfs: get rid of t... |
963 |
INIT_LIST_HEAD(&item->li_trans); |
43f5efc5b xfs: factor log i... |
964 |
} |
09a423a3d xfs: split tail_l... |
965 966 |
/* * Wake up processes waiting for log space after we have moved the log tail. |
09a423a3d xfs: split tail_l... |
967 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
968 |
void |
09a423a3d xfs: split tail_l... |
969 |
xfs_log_space_wake( |
cfb7cdca0 xfs: cleanup xfs_... |
970 |
struct xfs_mount *mp) |
1da177e4c Linux-2.6.12-rc2 |
971 |
{ |
ad223e603 xfs: rename log s... |
972 |
struct xlog *log = mp->m_log; |
cfb7cdca0 xfs: cleanup xfs_... |
973 |
int free_bytes; |
1da177e4c Linux-2.6.12-rc2 |
974 |
|
1da177e4c Linux-2.6.12-rc2 |
975 976 |
if (XLOG_FORCED_SHUTDOWN(log)) return; |
1da177e4c Linux-2.6.12-rc2 |
977 |
|
28496968a xfs: add the xlog... |
978 |
if (!list_empty_careful(&log->l_write_head.waiters)) { |
09a423a3d xfs: split tail_l... |
979 |
ASSERT(!(log->l_flags & XLOG_ACTIVE_RECOVERY)); |
28496968a xfs: add the xlog... |
980 981 |
spin_lock(&log->l_write_head.lock); free_bytes = xlog_space_left(log, &log->l_write_head.grant); |
e179840d7 xfs: share code f... |
982 |
xlog_grant_head_wake(log, &log->l_write_head, &free_bytes); |
28496968a xfs: add the xlog... |
983 |
spin_unlock(&log->l_write_head.lock); |
1da177e4c Linux-2.6.12-rc2 |
984 |
} |
105479419 xfs: convert log ... |
985 |
|
28496968a xfs: add the xlog... |
986 |
if (!list_empty_careful(&log->l_reserve_head.waiters)) { |
09a423a3d xfs: split tail_l... |
987 |
ASSERT(!(log->l_flags & XLOG_ACTIVE_RECOVERY)); |
28496968a xfs: add the xlog... |
988 989 |
spin_lock(&log->l_reserve_head.lock); free_bytes = xlog_space_left(log, &log->l_reserve_head.grant); |
e179840d7 xfs: share code f... |
990 |
xlog_grant_head_wake(log, &log->l_reserve_head, &free_bytes); |
28496968a xfs: add the xlog... |
991 |
spin_unlock(&log->l_reserve_head.lock); |
1da177e4c Linux-2.6.12-rc2 |
992 |
} |
3f16b9850 xfs: introduce ne... |
993 |
} |
1da177e4c Linux-2.6.12-rc2 |
994 995 |
/* |
2c6e24ce1 xfs: prevent dead... |
996 997 998 999 |
* Determine if we have a transaction that has gone to disk that needs to be * covered. To begin the transition to the idle state firstly the log needs to * be idle. That means the CIL, the AIL and the iclogs needs to be empty before * we start attempting to cover the log. |
b6f8dd49d xfs: ensure that ... |
1000 |
* |
2c6e24ce1 xfs: prevent dead... |
1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 |
* Only if we are then in a state where covering is needed, the caller is * informed that dummy transactions are required to move the log into the idle * state. * * If there are any items in the AIl or CIL, then we do not want to attempt to * cover the log as we may be in a situation where there isn't log space * available to run a dummy transaction and this can lead to deadlocks when the * tail of the log is pinned by an item that is modified in the CIL. Hence * there's no point in running a dummy transaction at this point because we * can't start trying to idle the log until both the CIL and AIL are empty. |
1da177e4c Linux-2.6.12-rc2 |
1011 |
*/ |
0d5a75e9e xfs: make several... |
1012 |
static int |
1da177e4c Linux-2.6.12-rc2 |
1013 1014 |
xfs_log_need_covered(xfs_mount_t *mp) { |
9a8d2fdbb xfs: remove xlog_... |
1015 |
struct xlog *log = mp->m_log; |
2c6e24ce1 xfs: prevent dead... |
1016 |
int needed = 0; |
1da177e4c Linux-2.6.12-rc2 |
1017 |
|
91ee575f2 xfs: allow lazy s... |
1018 |
if (!xfs_fs_writable(mp, SB_FREEZE_WRITE)) |
1da177e4c Linux-2.6.12-rc2 |
1019 |
return 0; |
2c6e24ce1 xfs: prevent dead... |
1020 1021 |
if (!xlog_cil_empty(log)) return 0; |
b22cd72c9 [XFS] Unwrap LOG_... |
1022 |
spin_lock(&log->l_icloglock); |
b6f8dd49d xfs: ensure that ... |
1023 1024 1025 1026 1027 1028 1029 |
switch (log->l_covered_state) { case XLOG_STATE_COVER_DONE: case XLOG_STATE_COVER_DONE2: case XLOG_STATE_COVER_IDLE: break; case XLOG_STATE_COVER_NEED: case XLOG_STATE_COVER_NEED2: |
2c6e24ce1 xfs: prevent dead... |
1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 |
if (xfs_ail_min_lsn(log->l_ailp)) break; if (!xlog_iclogs_empty(log)) break; needed = 1; if (log->l_covered_state == XLOG_STATE_COVER_NEED) log->l_covered_state = XLOG_STATE_COVER_DONE; else log->l_covered_state = XLOG_STATE_COVER_DONE2; break; |
b6f8dd49d xfs: ensure that ... |
1041 |
default: |
1da177e4c Linux-2.6.12-rc2 |
1042 |
needed = 1; |
b6f8dd49d xfs: ensure that ... |
1043 |
break; |
1da177e4c Linux-2.6.12-rc2 |
1044 |
} |
b22cd72c9 [XFS] Unwrap LOG_... |
1045 |
spin_unlock(&log->l_icloglock); |
014c2544e return statement ... |
1046 |
return needed; |
1da177e4c Linux-2.6.12-rc2 |
1047 |
} |
09a423a3d xfs: split tail_l... |
1048 |
/* |
1da177e4c Linux-2.6.12-rc2 |
1049 1050 1051 |
* We may be holding the log iclog lock upon entering this routine. */ xfs_lsn_t |
1c3046254 xfs: allow assign... |
1052 |
xlog_assign_tail_lsn_locked( |
1c3cb9ec0 xfs: convert l_ta... |
1053 |
struct xfs_mount *mp) |
1da177e4c Linux-2.6.12-rc2 |
1054 |
{ |
ad223e603 xfs: rename log s... |
1055 |
struct xlog *log = mp->m_log; |
1c3046254 xfs: allow assign... |
1056 1057 |
struct xfs_log_item *lip; xfs_lsn_t tail_lsn; |
57e809561 xfs: Rename xa_ e... |
1058 |
assert_spin_locked(&mp->m_ail->ail_lock); |
1da177e4c Linux-2.6.12-rc2 |
1059 |
|
09a423a3d xfs: split tail_l... |
1060 1061 1062 |
/* * To make sure we always have a valid LSN for the log tail we keep * track of the last LSN which was committed in log->l_last_sync_lsn, |
1c3046254 xfs: allow assign... |
1063 |
* and use that when the AIL was empty. |
09a423a3d xfs: split tail_l... |
1064 |
*/ |
1c3046254 xfs: allow assign... |
1065 1066 1067 1068 |
lip = xfs_ail_min(mp->m_ail); if (lip) tail_lsn = lip->li_lsn; else |
84f3c683c xfs: convert l_la... |
1069 |
tail_lsn = atomic64_read(&log->l_last_sync_lsn); |
750b9c906 xfs: trace AIL ma... |
1070 |
trace_xfs_log_assign_tail_lsn(log, tail_lsn); |
1c3cb9ec0 xfs: convert l_ta... |
1071 |
atomic64_set(&log->l_tail_lsn, tail_lsn); |
1da177e4c Linux-2.6.12-rc2 |
1072 |
return tail_lsn; |
1c3cb9ec0 xfs: convert l_ta... |
1073 |
} |
1da177e4c Linux-2.6.12-rc2 |
1074 |
|
1c3046254 xfs: allow assign... |
1075 1076 1077 1078 1079 |
xfs_lsn_t xlog_assign_tail_lsn( struct xfs_mount *mp) { xfs_lsn_t tail_lsn; |
57e809561 xfs: Rename xa_ e... |
1080 |
spin_lock(&mp->m_ail->ail_lock); |
1c3046254 xfs: allow assign... |
1081 |
tail_lsn = xlog_assign_tail_lsn_locked(mp); |
57e809561 xfs: Rename xa_ e... |
1082 |
spin_unlock(&mp->m_ail->ail_lock); |
1c3046254 xfs: allow assign... |
1083 1084 1085 |
return tail_lsn; } |
1da177e4c Linux-2.6.12-rc2 |
1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 |
/* * Return the space in the log between the tail and the head. The head * is passed in the cycle/bytes formal parms. In the special case where * the reserve head has wrapped passed the tail, this calculation is no * longer valid. In this case, just return 0 which means there is no space * in the log. This works for all places where this function is called * with the reserve head. Of course, if the write head were to ever * wrap the tail, we should blow up. Rather than catch this case here, * we depend on other ASSERTions in other parts of the code. XXXmiken * * This code also handles the case where the reservation head is behind * the tail. The details of this case are described below, but the end * result is that we return the size of the log as the amount of space left. */ |
a8272ce0c [XFS] Fix up spar... |
1100 |
STATIC int |
a69ed03c2 xfs: combine gran... |
1101 |
xlog_space_left( |
ad223e603 xfs: rename log s... |
1102 |
struct xlog *log, |
c8a09ff8c xfs: convert log ... |
1103 |
atomic64_t *head) |
1da177e4c Linux-2.6.12-rc2 |
1104 |
{ |
a69ed03c2 xfs: combine gran... |
1105 1106 1107 1108 1109 |
int free_bytes; int tail_bytes; int tail_cycle; int head_cycle; int head_bytes; |
1da177e4c Linux-2.6.12-rc2 |
1110 |
|
a69ed03c2 xfs: combine gran... |
1111 |
xlog_crack_grant_head(head, &head_cycle, &head_bytes); |
1c3cb9ec0 xfs: convert l_ta... |
1112 1113 |
xlog_crack_atomic_lsn(&log->l_tail_lsn, &tail_cycle, &tail_bytes); tail_bytes = BBTOB(tail_bytes); |
a69ed03c2 xfs: combine gran... |
1114 1115 1116 |
if (tail_cycle == head_cycle && head_bytes >= tail_bytes) free_bytes = log->l_logsize - (head_bytes - tail_bytes); else if (tail_cycle + 1 < head_cycle) |
1da177e4c Linux-2.6.12-rc2 |
1117 |
return 0; |
a69ed03c2 xfs: combine gran... |
1118 1119 1120 |
else if (tail_cycle < head_cycle) { ASSERT(tail_cycle == (head_cycle - 1)); free_bytes = tail_bytes - head_bytes; |
1da177e4c Linux-2.6.12-rc2 |
1121 1122 1123 1124 1125 1126 |
} else { /* * The reservation head is behind the tail. * In this case we just want to return the size of the * log as the amount of space left. */ |
f41febd2e xfs: Use consiste... |
1127 |
xfs_alert(log->l_mp, "xlog_space_left: head behind tail"); |
a0fa2b679 xfs: Convert xlog... |
1128 |
xfs_alert(log->l_mp, |
f41febd2e xfs: Use consiste... |
1129 1130 1131 1132 1133 |
" tail_cycle = %d, tail_bytes = %d", tail_cycle, tail_bytes); xfs_alert(log->l_mp, " GH cycle = %d, GH bytes = %d", head_cycle, head_bytes); |
1da177e4c Linux-2.6.12-rc2 |
1134 1135 1136 1137 |
ASSERT(0); free_bytes = log->l_logsize; } return free_bytes; |
a69ed03c2 xfs: combine gran... |
1138 |
} |
1da177e4c Linux-2.6.12-rc2 |
1139 |
|
0d5a75e9e xfs: make several... |
1140 |
static void |
79b54d9bf xfs: use bios dir... |
1141 1142 |
xlog_ioend_work( struct work_struct *work) |
1da177e4c Linux-2.6.12-rc2 |
1143 |
{ |
79b54d9bf xfs: use bios dir... |
1144 1145 1146 |
struct xlog_in_core *iclog = container_of(work, struct xlog_in_core, ic_end_io_work); struct xlog *log = iclog->ic_log; |
79b54d9bf xfs: use bios dir... |
1147 |
int error; |
1da177e4c Linux-2.6.12-rc2 |
1148 |
|
79b54d9bf xfs: use bios dir... |
1149 |
error = blk_status_to_errno(iclog->ic_bio.bi_status); |
366fc4b89 xfs: remove XLOG_... |
1150 1151 1152 |
#ifdef DEBUG /* treat writes with injected CRC errors as failed */ if (iclog->ic_fail_crc) |
79b54d9bf xfs: use bios dir... |
1153 |
error = -EIO; |
366fc4b89 xfs: remove XLOG_... |
1154 |
#endif |
1da177e4c Linux-2.6.12-rc2 |
1155 |
/* |
366fc4b89 xfs: remove XLOG_... |
1156 |
* Race to shutdown the filesystem if we see an error. |
1da177e4c Linux-2.6.12-rc2 |
1157 |
*/ |
79b54d9bf xfs: use bios dir... |
1158 1159 1160 |
if (XFS_TEST_ERROR(error, log->l_mp, XFS_ERRTAG_IODONE_IOERR)) { xfs_alert(log->l_mp, "log I/O error %d", error); xfs_force_shutdown(log->l_mp, SHUTDOWN_LOG_IO_ERROR); |
1da177e4c Linux-2.6.12-rc2 |
1161 |
} |
3db296f34 [XFS] Fix use-aft... |
1162 |
|
12e6a0f44 xfs: remove the a... |
1163 |
xlog_state_done_syncing(iclog); |
79b54d9bf xfs: use bios dir... |
1164 |
bio_uninit(&iclog->ic_bio); |
9c23eccc1 xfs: unmount does... |
1165 |
|
3db296f34 [XFS] Fix use-aft... |
1166 |
/* |
79b54d9bf xfs: use bios dir... |
1167 1168 1169 1170 |
* Drop the lock to signal that we are done. Nothing references the * iclog after this, so an unmount waiting on this lock can now tear it * down safely. As such, it is unsafe to reference the iclog after the * unlock as we could race with it being freed. |
3db296f34 [XFS] Fix use-aft... |
1171 |
*/ |
79b54d9bf xfs: use bios dir... |
1172 |
up(&iclog->ic_sema); |
c3f8fc73a xfs: make buffer ... |
1173 |
} |
1da177e4c Linux-2.6.12-rc2 |
1174 1175 |
/* |
1da177e4c Linux-2.6.12-rc2 |
1176 1177 |
* Return size of each in-core log record buffer. * |
9da096fd1 xfs: fix various ... |
1178 |
* All machines get 8 x 32kB buffers by default, unless tuned otherwise. |
1da177e4c Linux-2.6.12-rc2 |
1179 1180 1181 1182 |
* * If the filesystem blocksize is too large, we may need to choose a * larger size since the directory code currently logs entire blocks. */ |
1da177e4c Linux-2.6.12-rc2 |
1183 |
STATIC void |
9a8d2fdbb xfs: remove xlog_... |
1184 1185 1186 |
xlog_get_iclog_buffer_size( struct xfs_mount *mp, struct xlog *log) |
1da177e4c Linux-2.6.12-rc2 |
1187 |
{ |
1cb512587 [XFS] choose sing... |
1188 |
if (mp->m_logbufs <= 0) |
4f62282a3 xfs: cleanup xlog... |
1189 1190 1191 1192 1193 1194 |
mp->m_logbufs = XLOG_MAX_ICLOGS; if (mp->m_logbsize <= 0) mp->m_logbsize = XLOG_BIG_RECORD_BSIZE; log->l_iclog_bufs = mp->m_logbufs; log->l_iclog_size = mp->m_logbsize; |
1da177e4c Linux-2.6.12-rc2 |
1195 1196 |
/* |
4f62282a3 xfs: cleanup xlog... |
1197 |
* # headers = size / 32k - one header holds cycles from 32k of data. |
1da177e4c Linux-2.6.12-rc2 |
1198 |
*/ |
4f62282a3 xfs: cleanup xlog... |
1199 1200 1201 1202 |
log->l_iclog_heads = DIV_ROUND_UP(mp->m_logbsize, XLOG_HEADER_CYCLE_SIZE); log->l_iclog_hsize = log->l_iclog_heads << BBSHIFT; } |
1da177e4c Linux-2.6.12-rc2 |
1203 |
|
f661f1e0b xfs: sync work is... |
1204 1205 1206 1207 |
void xfs_log_work_queue( struct xfs_mount *mp) { |
696a56207 xfs: use dedicate... |
1208 |
queue_delayed_work(mp->m_sync_workqueue, &mp->m_log->l_work, |
f661f1e0b xfs: sync work is... |
1209 1210 1211 1212 1213 1214 1215 1216 |
msecs_to_jiffies(xfs_syncd_centisecs * 10)); } /* * Every sync period we need to unpin all items in the AIL and push them to * disk. If there is nothing dirty, then we might need to cover the log to * indicate that the filesystem is idle. */ |
0d5a75e9e xfs: make several... |
1217 |
static void |
f661f1e0b xfs: sync work is... |
1218 1219 1220 1221 1222 1223 1224 1225 |
xfs_log_worker( struct work_struct *work) { struct xlog *log = container_of(to_delayed_work(work), struct xlog, l_work); struct xfs_mount *mp = log->l_mp; /* dgc: errors ignored - not fatal and nowhere to report them */ |
61e63ecb5 xfs: consolidate ... |
1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 |
if (xfs_log_need_covered(mp)) { /* * Dump a transaction into the log that contains no real change. * This is needed to stamp the current tail LSN into the log * during the covering operation. * * We cannot use an inode here for this - that will push dirty * state back up into the VFS and then periodic inode flushing * will prevent log covering from making progress. Hence we * synchronously log the superblock instead to ensure the * superblock is immediately unpinned and can be written back. */ xfs_sync_sb(mp, true); } else |
f661f1e0b xfs: sync work is... |
1240 1241 1242 1243 1244 1245 1246 1247 |
xfs_log_force(mp, 0); /* start pushing all the metadata that is currently dirty */ xfs_ail_push_all(mp->m_ail); /* queue us up again */ xfs_log_work_queue(mp); } |
1da177e4c Linux-2.6.12-rc2 |
1248 1249 1250 1251 1252 |
/* * This routine initializes some of the log structure for a given mount point. * Its primary purpose is to fill in enough, so recovery can occur. However, * some other stuff may be filled in too. */ |
9a8d2fdbb xfs: remove xlog_... |
1253 1254 1255 1256 1257 1258 |
STATIC struct xlog * xlog_alloc_log( struct xfs_mount *mp, struct xfs_buftarg *log_target, xfs_daddr_t blk_offset, int num_bblks) |
1da177e4c Linux-2.6.12-rc2 |
1259 |
{ |
9a8d2fdbb xfs: remove xlog_... |
1260 |
struct xlog *log; |
1da177e4c Linux-2.6.12-rc2 |
1261 1262 1263 |
xlog_rec_header_t *head; xlog_in_core_t **iclogp; xlog_in_core_t *iclog, *prev_iclog=NULL; |
1da177e4c Linux-2.6.12-rc2 |
1264 |
int i; |
2451337dd xfs: global error... |
1265 |
int error = -ENOMEM; |
69ce58f08 xfs: record log s... |
1266 |
uint log2_size = 0; |
1da177e4c Linux-2.6.12-rc2 |
1267 |
|
9a8d2fdbb xfs: remove xlog_... |
1268 |
log = kmem_zalloc(sizeof(struct xlog), KM_MAYFAIL); |
a6cb767e2 xfs: validate log... |
1269 |
if (!log) { |
a0fa2b679 xfs: Convert xlog... |
1270 |
xfs_warn(mp, "Log allocation failed: No memory!"); |
a6cb767e2 xfs: validate log... |
1271 1272 |
goto out; } |
1da177e4c Linux-2.6.12-rc2 |
1273 1274 1275 1276 1277 1278 1279 1280 |
log->l_mp = mp; log->l_targ = log_target; log->l_logsize = BBTOB(num_bblks); log->l_logBBstart = blk_offset; log->l_logBBsize = num_bblks; log->l_covered_state = XLOG_STATE_COVER_IDLE; log->l_flags |= XLOG_ACTIVE_RECOVERY; |
f661f1e0b xfs: sync work is... |
1281 |
INIT_DELAYED_WORK(&log->l_work, xfs_log_worker); |
1da177e4c Linux-2.6.12-rc2 |
1282 1283 |
log->l_prev_block = -1; |
1da177e4c Linux-2.6.12-rc2 |
1284 |
/* log->l_tail_lsn = 0x100000000LL; cycle = 1; current block = 0 */ |
1c3cb9ec0 xfs: convert l_ta... |
1285 1286 |
xlog_assign_atomic_lsn(&log->l_tail_lsn, 1, 0); xlog_assign_atomic_lsn(&log->l_last_sync_lsn, 1, 0); |
1da177e4c Linux-2.6.12-rc2 |
1287 |
log->l_curr_cycle = 1; /* 0 is bad since this is initial value */ |
c303c5b8c xfs: add xlog_gra... |
1288 1289 1290 |
xlog_grant_head_init(&log->l_reserve_head); xlog_grant_head_init(&log->l_write_head); |
1da177e4c Linux-2.6.12-rc2 |
1291 |
|
2451337dd xfs: global error... |
1292 |
error = -EFSCORRUPTED; |
621187099 [XFS] remove shou... |
1293 |
if (xfs_sb_version_hassector(&mp->m_sb)) { |
69ce58f08 xfs: record log s... |
1294 1295 |
log2_size = mp->m_sb.sb_logsectlog; if (log2_size < BBSHIFT) { |
a0fa2b679 xfs: Convert xlog... |
1296 1297 |
xfs_warn(mp, "Log sector size too small (0x%x < 0x%x)", log2_size, BBSHIFT); |
a6cb767e2 xfs: validate log... |
1298 1299 |
goto out_free_log; } |
69ce58f08 xfs: record log s... |
1300 1301 |
log2_size -= BBSHIFT; if (log2_size > mp->m_sectbb_log) { |
a0fa2b679 xfs: Convert xlog... |
1302 1303 |
xfs_warn(mp, "Log sector size too large (0x%x > 0x%x)", log2_size, mp->m_sectbb_log); |
a6cb767e2 xfs: validate log... |
1304 1305 |
goto out_free_log; } |
69ce58f08 xfs: record log s... |
1306 1307 1308 1309 |
/* for larger sector sizes, must have v2 or external log */ if (log2_size && log->l_logBBstart > 0 && !xfs_sb_version_haslogv2(&mp->m_sb)) { |
a0fa2b679 xfs: Convert xlog... |
1310 1311 1312 |
xfs_warn(mp, "log sector size (0x%x) invalid for configuration.", log2_size); |
a6cb767e2 xfs: validate log... |
1313 1314 |
goto out_free_log; } |
1da177e4c Linux-2.6.12-rc2 |
1315 |
} |
69ce58f08 xfs: record log s... |
1316 |
log->l_sectBBsize = 1 << log2_size; |
1da177e4c Linux-2.6.12-rc2 |
1317 1318 |
xlog_get_iclog_buffer_size(mp, log); |
007c61c68 [XFS] Remove spin.h |
1319 |
spin_lock_init(&log->l_icloglock); |
eb40a8750 xfs: use wait que... |
1320 |
init_waitqueue_head(&log->l_flush_wait); |
1da177e4c Linux-2.6.12-rc2 |
1321 |
|
1da177e4c Linux-2.6.12-rc2 |
1322 1323 1324 1325 1326 1327 1328 1329 |
iclogp = &log->l_iclog; /* * The amount of memory to allocate for the iclog structure is * rather funky due to the way the structure is defined. It is * done this way so that we can use different sizes for machines * with different amounts of memory. See the definition of * xlog_in_core_t in xfs_log_priv.h for details. */ |
1da177e4c Linux-2.6.12-rc2 |
1330 |
ASSERT(log->l_iclog_size >= 4096); |
79b54d9bf xfs: use bios dir... |
1331 |
for (i = 0; i < log->l_iclog_bufs; i++) { |
f8f9ee479 xfs: add kmem_all... |
1332 |
int align_mask = xfs_buftarg_dma_alignment(mp->m_logdev_targp); |
89b171acb xfs: fix iclog al... |
1333 1334 |
size_t bvec_size = howmany(log->l_iclog_size, PAGE_SIZE) * sizeof(struct bio_vec); |
79b54d9bf xfs: use bios dir... |
1335 1336 1337 |
iclog = kmem_zalloc(sizeof(*iclog) + bvec_size, KM_MAYFAIL); if (!iclog) |
644c3567d [XFS] handle memo... |
1338 |
goto out_free_iclog; |
79b54d9bf xfs: use bios dir... |
1339 |
*iclogp = iclog; |
1da177e4c Linux-2.6.12-rc2 |
1340 1341 |
iclog->ic_prev = prev_iclog; prev_iclog = iclog; |
1fa40b01a [XFS] Only use re... |
1342 |
|
f8f9ee479 xfs: add kmem_all... |
1343 |
iclog->ic_data = kmem_alloc_io(log->l_iclog_size, align_mask, |
3219e8cf0 xfs: assure zeroe... |
1344 |
KM_MAYFAIL | KM_ZERO); |
79b54d9bf xfs: use bios dir... |
1345 |
if (!iclog->ic_data) |
644c3567d [XFS] handle memo... |
1346 |
goto out_free_iclog; |
4679b2d36 [XFS] Reorganise ... |
1347 |
#ifdef DEBUG |
5809d5e08 xfs: use void poi... |
1348 |
log->l_iclog_bak[i] = &iclog->ic_header; |
4679b2d36 [XFS] Reorganise ... |
1349 |
#endif |
1da177e4c Linux-2.6.12-rc2 |
1350 1351 |
head = &iclog->ic_header; memset(head, 0, sizeof(xlog_rec_header_t)); |
b53e675dc [XFS] xlog_rec_he... |
1352 1353 |
head->h_magicno = cpu_to_be32(XLOG_HEADER_MAGIC_NUM); head->h_version = cpu_to_be32( |
621187099 [XFS] remove shou... |
1354 |
xfs_sb_version_haslogv2(&log->l_mp->m_sb) ? 2 : 1); |
b53e675dc [XFS] xlog_rec_he... |
1355 |
head->h_size = cpu_to_be32(log->l_iclog_size); |
1da177e4c Linux-2.6.12-rc2 |
1356 |
/* new fields */ |
b53e675dc [XFS] xlog_rec_he... |
1357 |
head->h_fmt = cpu_to_be32(XLOG_FMT); |
1da177e4c Linux-2.6.12-rc2 |
1358 |
memcpy(&head->h_fs_uuid, &mp->m_sb.sb_uuid, sizeof(uuid_t)); |
79b54d9bf xfs: use bios dir... |
1359 |
iclog->ic_size = log->l_iclog_size - log->l_iclog_hsize; |
1da177e4c Linux-2.6.12-rc2 |
1360 1361 |
iclog->ic_state = XLOG_STATE_ACTIVE; iclog->ic_log = log; |
114d23aae [XFS] Per iclog c... |
1362 1363 |
atomic_set(&iclog->ic_refcnt, 0); spin_lock_init(&iclog->ic_callback_lock); |
89ae379d5 xfs: use a list_h... |
1364 |
INIT_LIST_HEAD(&iclog->ic_callbacks); |
b28708d6a [XFS] sanitize xl... |
1365 |
iclog->ic_datap = (char *)iclog->ic_data + log->l_iclog_hsize; |
1da177e4c Linux-2.6.12-rc2 |
1366 |
|
eb40a8750 xfs: use wait que... |
1367 1368 |
init_waitqueue_head(&iclog->ic_force_wait); init_waitqueue_head(&iclog->ic_write_wait); |
79b54d9bf xfs: use bios dir... |
1369 1370 |
INIT_WORK(&iclog->ic_end_io_work, xlog_ioend_work); sema_init(&iclog->ic_sema, 1); |
1da177e4c Linux-2.6.12-rc2 |
1371 1372 1373 1374 1375 |
iclogp = &iclog->ic_next; } *iclogp = log->l_iclog; /* complete ring */ log->l_iclog->ic_prev = prev_iclog; /* re-write 1st prev ptr */ |
1058d0f5e xfs: move the log... |
1376 1377 |
log->l_ioend_workqueue = alloc_workqueue("xfs-log/%s", WQ_MEM_RECLAIM | WQ_FREEZABLE | WQ_HIGHPRI, 0, |
e1d3d2188 xfs: use super s_... |
1378 |
mp->m_super->s_id); |
1058d0f5e xfs: move the log... |
1379 1380 |
if (!log->l_ioend_workqueue) goto out_free_iclog; |
71e330b59 xfs: Introduce de... |
1381 1382 |
error = xlog_cil_init(log); if (error) |
1058d0f5e xfs: move the log... |
1383 |
goto out_destroy_workqueue; |
1da177e4c Linux-2.6.12-rc2 |
1384 |
return log; |
644c3567d [XFS] handle memo... |
1385 |
|
1058d0f5e xfs: move the log... |
1386 1387 |
out_destroy_workqueue: destroy_workqueue(log->l_ioend_workqueue); |
644c3567d [XFS] handle memo... |
1388 1389 1390 |
out_free_iclog: for (iclog = log->l_iclog; iclog; iclog = prev_iclog) { prev_iclog = iclog->ic_next; |
79b54d9bf xfs: use bios dir... |
1391 |
kmem_free(iclog->ic_data); |
644c3567d [XFS] handle memo... |
1392 |
kmem_free(iclog); |
798a9cada xfs: fix mount fa... |
1393 1394 |
if (prev_iclog == log->l_iclog) break; |
644c3567d [XFS] handle memo... |
1395 |
} |
644c3567d [XFS] handle memo... |
1396 1397 |
out_free_log: kmem_free(log); |
a6cb767e2 xfs: validate log... |
1398 |
out: |
2451337dd xfs: global error... |
1399 |
return ERR_PTR(error); |
1da177e4c Linux-2.6.12-rc2 |
1400 |
} /* xlog_alloc_log */ |
1da177e4c Linux-2.6.12-rc2 |
1401 1402 |
/* * Write out the commit record of a transaction associated with the given |
f10e925de xfs: merge xlog_c... |
1403 |
* ticket to close off a running log write. Return the lsn of the commit record. |
1da177e4c Linux-2.6.12-rc2 |
1404 |
*/ |
f10e925de xfs: merge xlog_c... |
1405 |
int |
55b66332d xfs: introduce ne... |
1406 |
xlog_commit_record( |
ad223e603 xfs: rename log s... |
1407 |
struct xlog *log, |
55b66332d xfs: introduce ne... |
1408 1409 |
struct xlog_ticket *ticket, struct xlog_in_core **iclog, |
f10e925de xfs: merge xlog_c... |
1410 |
xfs_lsn_t *lsn) |
1da177e4c Linux-2.6.12-rc2 |
1411 |
{ |
55b66332d xfs: introduce ne... |
1412 1413 1414 1415 1416 1417 1418 1419 1420 |
struct xfs_log_iovec reg = { .i_addr = NULL, .i_len = 0, .i_type = XLOG_REG_TYPE_COMMIT, }; struct xfs_log_vec vec = { .lv_niovecs = 1, .lv_iovecp = ®, }; |
f10e925de xfs: merge xlog_c... |
1421 |
int error; |
1da177e4c Linux-2.6.12-rc2 |
1422 |
|
f10e925de xfs: merge xlog_c... |
1423 1424 1425 1426 1427 |
if (XLOG_FORCED_SHUTDOWN(log)) return -EIO; error = xlog_write(log, &vec, ticket, lsn, iclog, XLOG_COMMIT_TRANS, false); |
55b66332d xfs: introduce ne... |
1428 |
if (error) |
f10e925de xfs: merge xlog_c... |
1429 |
xfs_force_shutdown(log->l_mp, SHUTDOWN_LOG_IO_ERROR); |
014c2544e return statement ... |
1430 |
return error; |
55b66332d xfs: introduce ne... |
1431 |
} |
1da177e4c Linux-2.6.12-rc2 |
1432 1433 |
/* |
ed1575daf xfs: expose the l... |
1434 1435 1436 1437 1438 |
* Compute the LSN that we'd need to push the log tail towards in order to have * (a) enough on-disk log space to log the number of bytes specified, (b) at * least 25% of the log space free, and (c) at least 256 blocks free. If the * log free space already meets all three thresholds, this function returns * NULLCOMMITLSN. |
1da177e4c Linux-2.6.12-rc2 |
1439 |
*/ |
ed1575daf xfs: expose the l... |
1440 1441 |
xfs_lsn_t xlog_grant_push_threshold( |
ad223e603 xfs: rename log s... |
1442 |
struct xlog *log, |
2ced19cba xfs: make AIL tai... |
1443 |
int need_bytes) |
1da177e4c Linux-2.6.12-rc2 |
1444 |
{ |
2ced19cba xfs: make AIL tai... |
1445 |
xfs_lsn_t threshold_lsn = 0; |
84f3c683c xfs: convert l_la... |
1446 |
xfs_lsn_t last_sync_lsn; |
2ced19cba xfs: make AIL tai... |
1447 1448 1449 1450 1451 1452 1453 |
int free_blocks; int free_bytes; int threshold_block; int threshold_cycle; int free_threshold; ASSERT(BTOBB(need_bytes) < log->l_logBBsize); |
28496968a xfs: add the xlog... |
1454 |
free_bytes = xlog_space_left(log, &log->l_reserve_head.grant); |
2ced19cba xfs: make AIL tai... |
1455 1456 1457 1458 1459 1460 1461 1462 |
free_blocks = BTOBBT(free_bytes); /* * Set the threshold for the minimum number of free blocks in the * log to the maximum of what the caller needs, one quarter of the * log, and 256 blocks. */ free_threshold = BTOBB(need_bytes); |
9bb54cb56 xfs: clean up MIN... |
1463 1464 |
free_threshold = max(free_threshold, (log->l_logBBsize >> 2)); free_threshold = max(free_threshold, 256); |
2ced19cba xfs: make AIL tai... |
1465 |
if (free_blocks >= free_threshold) |
ed1575daf xfs: expose the l... |
1466 |
return NULLCOMMITLSN; |
2ced19cba xfs: make AIL tai... |
1467 |
|
1c3cb9ec0 xfs: convert l_ta... |
1468 1469 1470 |
xlog_crack_atomic_lsn(&log->l_tail_lsn, &threshold_cycle, &threshold_block); threshold_block += free_threshold; |
1da177e4c Linux-2.6.12-rc2 |
1471 |
if (threshold_block >= log->l_logBBsize) { |
2ced19cba xfs: make AIL tai... |
1472 1473 |
threshold_block -= log->l_logBBsize; threshold_cycle += 1; |
1da177e4c Linux-2.6.12-rc2 |
1474 |
} |
2ced19cba xfs: make AIL tai... |
1475 1476 1477 1478 |
threshold_lsn = xlog_assign_lsn(threshold_cycle, threshold_block); /* * Don't pass in an lsn greater than the lsn of the last |
84f3c683c xfs: convert l_la... |
1479 1480 |
* log record known to be on disk. Use a snapshot of the last sync lsn * so that it doesn't change between the compare and the set. |
1da177e4c Linux-2.6.12-rc2 |
1481 |
*/ |
84f3c683c xfs: convert l_la... |
1482 1483 1484 |
last_sync_lsn = atomic64_read(&log->l_last_sync_lsn); if (XFS_LSN_CMP(threshold_lsn, last_sync_lsn) > 0) threshold_lsn = last_sync_lsn; |
2ced19cba xfs: make AIL tai... |
1485 |
|
ed1575daf xfs: expose the l... |
1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 |
return threshold_lsn; } /* * Push the tail of the log if we need to do so to maintain the free log space * thresholds set out by xlog_grant_push_threshold. We may need to adopt a * policy which pushes on an lsn which is further along in the log once we * reach the high water mark. In this manner, we would be creating a low water * mark. */ STATIC void xlog_grant_push_ail( struct xlog *log, int need_bytes) { xfs_lsn_t threshold_lsn; threshold_lsn = xlog_grant_push_threshold(log, need_bytes); if (threshold_lsn == NULLCOMMITLSN || XLOG_FORCED_SHUTDOWN(log)) return; |
2ced19cba xfs: make AIL tai... |
1506 1507 1508 1509 1510 |
/* * Get the transaction layer to kick the dirty buffers out to * disk asynchronously. No point in trying to do this if * the filesystem is shutting down. */ |
ed1575daf xfs: expose the l... |
1511 |
xfs_ail_push(log->l_ailp, threshold_lsn); |
2ced19cba xfs: make AIL tai... |
1512 |
} |
1da177e4c Linux-2.6.12-rc2 |
1513 |
|
873ff5501 xfs: clean up log... |
1514 |
/* |
0e446be44 xfs: add CRC chec... |
1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 |
* Stamp cycle number in every block */ STATIC void xlog_pack_data( struct xlog *log, struct xlog_in_core *iclog, int roundoff) { int i, j, k; int size = iclog->ic_offset + roundoff; __be32 cycle_lsn; |
b2a922cd6 xfs: remove xfs_c... |
1526 |
char *dp; |
0e446be44 xfs: add CRC chec... |
1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 |
cycle_lsn = CYCLE_LSN_DISK(iclog->ic_header.h_lsn); dp = iclog->ic_datap; for (i = 0; i < BTOBB(size); i++) { if (i >= (XLOG_HEADER_CYCLE_SIZE / BBSIZE)) break; iclog->ic_header.h_cycle_data[i] = *(__be32 *)dp; *(__be32 *)dp = cycle_lsn; dp += BBSIZE; } if (xfs_sb_version_haslogv2(&log->l_mp->m_sb)) { xlog_in_core_2_t *xhdr = iclog->ic_data; for ( ; i < BTOBB(size); i++) { j = i / (XLOG_HEADER_CYCLE_SIZE / BBSIZE); k = i % (XLOG_HEADER_CYCLE_SIZE / BBSIZE); xhdr[j].hic_xheader.xh_cycle_data[k] = *(__be32 *)dp; *(__be32 *)dp = cycle_lsn; dp += BBSIZE; } for (i = 1; i < log->l_iclog_heads; i++) xhdr[i].hic_xheader.xh_cycle = cycle_lsn; } } /* * Calculate the checksum for a log buffer. * * This is a little more complicated than it should be because the various * headers and the actual data are non-contiguous. */ |
f9668a09e xfs: fix sparse r... |
1561 |
__le32 |
0e446be44 xfs: add CRC chec... |
1562 1563 1564 1565 1566 1567 |
xlog_cksum( struct xlog *log, struct xlog_rec_header *rhead, char *dp, int size) { |
c8ce540db xfs: remove doubl... |
1568 |
uint32_t crc; |
0e446be44 xfs: add CRC chec... |
1569 1570 |
/* first generate the crc for the record header ... */ |
cae028df5 xfs: optimise CRC... |
1571 |
crc = xfs_start_cksum_update((char *)rhead, |
0e446be44 xfs: add CRC chec... |
1572 1573 1574 1575 1576 1577 1578 |
sizeof(struct xlog_rec_header), offsetof(struct xlog_rec_header, h_crc)); /* ... then for additional cycle data for v2 logs ... */ if (xfs_sb_version_haslogv2(&log->l_mp->m_sb)) { union xlog_in_core2 *xhdr = (union xlog_in_core2 *)rhead; int i; |
a3f200146 xfs: checksum log... |
1579 |
int xheads; |
0e446be44 xfs: add CRC chec... |
1580 |
|
0c771b99d xfs: clean up cal... |
1581 |
xheads = DIV_ROUND_UP(size, XLOG_HEADER_CYCLE_SIZE); |
0e446be44 xfs: add CRC chec... |
1582 |
|
a3f200146 xfs: checksum log... |
1583 |
for (i = 1; i < xheads; i++) { |
0e446be44 xfs: add CRC chec... |
1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 |
crc = crc32c(crc, &xhdr[i].hic_xheader, sizeof(struct xlog_rec_ext_header)); } } /* ... and finally for the payload */ crc = crc32c(crc, dp, size); return xfs_end_cksum(crc); } |
79b54d9bf xfs: use bios dir... |
1594 1595 1596 1597 1598 |
static void xlog_bio_end_io( struct bio *bio) { struct xlog_in_core *iclog = bio->bi_private; |
1058d0f5e xfs: move the log... |
1599 |
queue_work(iclog->ic_log->l_ioend_workqueue, |
79b54d9bf xfs: use bios dir... |
1600 1601 |
&iclog->ic_end_io_work); } |
842a42d12 xfs: shutdown on ... |
1602 |
static int |
79b54d9bf xfs: use bios dir... |
1603 1604 1605 1606 1607 1608 1609 1610 1611 |
xlog_map_iclog_data( struct bio *bio, void *data, size_t count) { do { struct page *page = kmem_to_page(data); unsigned int off = offset_in_page(data); size_t len = min_t(size_t, count, PAGE_SIZE - off); |
842a42d12 xfs: shutdown on ... |
1612 1613 |
if (bio_add_page(bio, page, len, off) != len) return -EIO; |
79b54d9bf xfs: use bios dir... |
1614 1615 1616 1617 |
data += len; count -= len; } while (count); |
842a42d12 xfs: shutdown on ... |
1618 1619 |
return 0; |
79b54d9bf xfs: use bios dir... |
1620 |
} |
94860a301 xfs: factor out l... |
1621 1622 1623 1624 |
STATIC void xlog_write_iclog( struct xlog *log, struct xlog_in_core *iclog, |
94860a301 xfs: factor out l... |
1625 |
uint64_t bno, |
79b54d9bf xfs: use bios dir... |
1626 |
unsigned int count, |
94860a301 xfs: factor out l... |
1627 |
bool need_flush) |
873ff5501 xfs: clean up log... |
1628 |
{ |
94860a301 xfs: factor out l... |
1629 |
ASSERT(bno < log->l_logBBsize); |
94860a301 xfs: factor out l... |
1630 1631 1632 1633 1634 1635 1636 1637 1638 |
/* * We lock the iclogbufs here so that we can serialise against I/O * completion during unmount. We might be processing a shutdown * triggered during unmount, and that can occur asynchronously to the * unmount thread, and hence we need to ensure that completes before * tearing down the iclogbufs. Hence we need to hold the buffer lock * across the log IO to archieve that. */ |
79b54d9bf xfs: use bios dir... |
1639 |
down(&iclog->ic_sema); |
1858bb0be xfs: turn ic_stat... |
1640 |
if (unlikely(iclog->ic_state == XLOG_STATE_IOERROR)) { |
873ff5501 xfs: clean up log... |
1641 1642 1643 |
/* * It would seem logical to return EIO here, but we rely on * the log state machine to propagate I/O errors instead of |
79b54d9bf xfs: use bios dir... |
1644 1645 1646 |
* doing it here. We kick of the state machine and unlock * the buffer manually, the code needs to be kept in sync * with the I/O completion path. |
873ff5501 xfs: clean up log... |
1647 |
*/ |
12e6a0f44 xfs: remove the a... |
1648 |
xlog_state_done_syncing(iclog); |
79b54d9bf xfs: use bios dir... |
1649 |
up(&iclog->ic_sema); |
94860a301 xfs: factor out l... |
1650 |
return; |
873ff5501 xfs: clean up log... |
1651 |
} |
79b54d9bf xfs: use bios dir... |
1652 1653 1654 1655 1656 |
bio_init(&iclog->ic_bio, iclog->ic_bvec, howmany(count, PAGE_SIZE)); bio_set_dev(&iclog->ic_bio, log->l_targ->bt_bdev); iclog->ic_bio.bi_iter.bi_sector = log->l_logBBstart + bno; iclog->ic_bio.bi_end_io = xlog_bio_end_io; iclog->ic_bio.bi_private = iclog; |
2def2845c xfs: don't allow ... |
1657 1658 1659 1660 1661 1662 1663 1664 1665 |
/* * We use REQ_SYNC | REQ_IDLE here to tell the block layer the are more * IOs coming immediately after this one. This prevents the block layer * writeback throttle from throttling log writes behind background * metadata writeback and causing priority inversions. */ iclog->ic_bio.bi_opf = REQ_OP_WRITE | REQ_META | REQ_SYNC | REQ_IDLE | REQ_FUA; |
79b54d9bf xfs: use bios dir... |
1666 1667 |
if (need_flush) iclog->ic_bio.bi_opf |= REQ_PREFLUSH; |
842a42d12 xfs: shutdown on ... |
1668 1669 1670 1671 |
if (xlog_map_iclog_data(&iclog->ic_bio, iclog->ic_data, count)) { xfs_force_shutdown(log->l_mp, SHUTDOWN_LOG_IO_ERROR); return; } |
79b54d9bf xfs: use bios dir... |
1672 |
if (is_vmalloc_addr(iclog->ic_data)) |
2c68a1dfb xfs: remove the u... |
1673 |
flush_kernel_vmap_range(iclog->ic_data, count); |
79b54d9bf xfs: use bios dir... |
1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 |
/* * If this log buffer would straddle the end of the log we will have * to split it up into two bios, so that we can continue at the start. */ if (bno + BTOBB(count) > log->l_logBBsize) { struct bio *split; split = bio_split(&iclog->ic_bio, log->l_logBBsize - bno, GFP_NOIO, &fs_bio_set); bio_chain(split, &iclog->ic_bio); submit_bio(split); /* restart at logical offset zero for the remainder */ iclog->ic_bio.bi_iter.bi_sector = log->l_logBBstart; } submit_bio(&iclog->ic_bio); |
873ff5501 xfs: clean up log... |
1692 |
} |
1da177e4c Linux-2.6.12-rc2 |
1693 1694 |
/* |
569338480 xfs: factor out s... |
1695 1696 1697 1698 |
* We need to bump cycle number for the part of the iclog that is * written to the start of the log. Watch out for the header magic * number case, though. */ |
79b54d9bf xfs: use bios dir... |
1699 |
static void |
569338480 xfs: factor out s... |
1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 |
xlog_split_iclog( struct xlog *log, void *data, uint64_t bno, unsigned int count) { unsigned int split_offset = BBTOB(log->l_logBBsize - bno); unsigned int i; for (i = split_offset; i < count; i += BBSIZE) { uint32_t cycle = get_unaligned_be32(data + i); if (++cycle == XLOG_HEADER_MAGIC_NUM) cycle++; put_unaligned_be32(cycle, data + i); } |
569338480 xfs: factor out s... |
1716 |
} |
db0a6faf9 xfs: factor out i... |
1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 |
static int xlog_calc_iclog_size( struct xlog *log, struct xlog_in_core *iclog, uint32_t *roundoff) { uint32_t count_init, count; bool use_lsunit; use_lsunit = xfs_sb_version_haslogv2(&log->l_mp->m_sb) && log->l_mp->m_sb.sb_logsunit > 1; /* Add for LR header */ count_init = log->l_iclog_hsize + iclog->ic_offset; /* Round out the log write size */ if (use_lsunit) { /* we have a v2 stripe unit to use */ count = XLOG_LSUNITTOB(log, XLOG_BTOLSUNIT(log, count_init)); } else { count = BBTOB(BTOBB(count_init)); } ASSERT(count >= count_init); *roundoff = count - count_init; if (use_lsunit) ASSERT(*roundoff < log->l_mp->m_sb.sb_logsunit); else ASSERT(*roundoff < BBTOB(1)); return count; } |
569338480 xfs: factor out s... |
1749 |
/* |
1da177e4c Linux-2.6.12-rc2 |
1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 |
* Flush out the in-core log (iclog) to the on-disk log in an asynchronous * fashion. Previously, we should have moved the current iclog * ptr in the log to point to the next available iclog. This allows further * write to continue while this code syncs out an iclog ready to go. * Before an in-core log can be written out, the data section must be scanned * to save away the 1st word of each BBSIZE block into the header. We replace * it with the current cycle count. Each BBSIZE block is tagged with the * cycle count because there in an implicit assumption that drives will * guarantee that entire 512 byte blocks get written at once. In other words, * we can't have part of a 512 byte block written and part not written. By * tagging each block, we will know which blocks are valid when recovering * after an unclean shutdown. * * This routine is single threaded on the iclog. No other thread can be in * this routine with the same iclog. Changing contents of iclog can there- * fore be done without grabbing the state machine lock. Updating the global * log will require grabbing the lock though. * * The entire log manager uses a logical block numbering scheme. Only |
94860a301 xfs: factor out l... |
1769 1770 |
* xlog_write_iclog knows about the fact that the log may not start with * block zero on a given device. |
1da177e4c Linux-2.6.12-rc2 |
1771 |
*/ |
94860a301 xfs: factor out l... |
1772 |
STATIC void |
9a8d2fdbb xfs: remove xlog_... |
1773 1774 1775 |
xlog_sync( struct xlog *log, struct xlog_in_core *iclog) |
1da177e4c Linux-2.6.12-rc2 |
1776 |
{ |
db0a6faf9 xfs: factor out i... |
1777 1778 1779 |
unsigned int count; /* byte count of bwrite */ unsigned int roundoff; /* roundoff to BB or stripe */ uint64_t bno; |
db0a6faf9 xfs: factor out i... |
1780 |
unsigned int size; |
79b54d9bf xfs: use bios dir... |
1781 |
bool need_flush = true, split = false; |
1da177e4c Linux-2.6.12-rc2 |
1782 |
|
155cc6b78 [XFS] Use atomics... |
1783 |
ASSERT(atomic_read(&iclog->ic_refcnt) == 0); |
1da177e4c Linux-2.6.12-rc2 |
1784 |
|
db0a6faf9 xfs: factor out i... |
1785 |
count = xlog_calc_iclog_size(log, iclog, &roundoff); |
1da177e4c Linux-2.6.12-rc2 |
1786 1787 |
/* move grant heads by roundoff in sync */ |
28496968a xfs: add the xlog... |
1788 1789 |
xlog_grant_add_space(log, &log->l_reserve_head.grant, roundoff); xlog_grant_add_space(log, &log->l_write_head.grant, roundoff); |
1da177e4c Linux-2.6.12-rc2 |
1790 1791 1792 1793 1794 |
/* put cycle number in every block */ xlog_pack_data(log, iclog, roundoff); /* real byte length */ |
0e446be44 xfs: add CRC chec... |
1795 |
size = iclog->ic_offset; |
db0a6faf9 xfs: factor out i... |
1796 |
if (xfs_sb_version_haslogv2(&log->l_mp->m_sb)) |
0e446be44 xfs: add CRC chec... |
1797 1798 |
size += roundoff; iclog->ic_header.h_len = cpu_to_be32(size); |
1da177e4c Linux-2.6.12-rc2 |
1799 |
|
9b0489c1d xfs: update both ... |
1800 |
XFS_STATS_INC(log->l_mp, xs_log_writes); |
ff6d6af23 xfs: per-filesyst... |
1801 |
XFS_STATS_ADD(log->l_mp, xs_log_blocks, BTOBB(count)); |
1da177e4c Linux-2.6.12-rc2 |
1802 |
|
94860a301 xfs: factor out l... |
1803 |
bno = BLOCK_LSN(be64_to_cpu(iclog->ic_header.h_lsn)); |
1da177e4c Linux-2.6.12-rc2 |
1804 |
/* Do we need to split this write into 2 parts? */ |
79b54d9bf xfs: use bios dir... |
1805 1806 1807 1808 |
if (bno + BTOBB(count) > log->l_logBBsize) { xlog_split_iclog(log, &iclog->ic_header, bno, count); split = true; } |
0e446be44 xfs: add CRC chec... |
1809 1810 1811 1812 |
/* calculcate the checksum */ iclog->ic_header.h_crc = xlog_cksum(log, &iclog->ic_header, iclog->ic_datap, size); |
609adfc2e xfs: debug mode l... |
1813 1814 1815 1816 1817 1818 1819 |
/* * Intentionally corrupt the log record CRC based on the error injection * frequency, if defined. This facilitates testing log recovery in the * event of torn writes. Hence, set the IOABORT state to abort the log * write on I/O completion and shutdown the fs. The subsequent mount * detects the bad CRC and attempts to recover. */ |
366fc4b89 xfs: remove XLOG_... |
1820 |
#ifdef DEBUG |
3e88a0078 xfs: replace log_... |
1821 |
if (XFS_TEST_ERROR(false, log->l_mp, XFS_ERRTAG_LOG_BAD_CRC)) { |
e2a641922 xfs: corruption n... |
1822 |
iclog->ic_header.h_crc &= cpu_to_le32(0xAAAAAAAA); |
366fc4b89 xfs: remove XLOG_... |
1823 |
iclog->ic_fail_crc = true; |
609adfc2e xfs: debug mode l... |
1824 1825 1826 1827 |
xfs_warn(log->l_mp, "Intentionally corrupted log record at LSN 0x%llx. Shutdown imminent.", be64_to_cpu(iclog->ic_header.h_lsn)); } |
366fc4b89 xfs: remove XLOG_... |
1828 |
#endif |
0e446be44 xfs: add CRC chec... |
1829 |
|
2291dab2c xfs: Always flush... |
1830 1831 1832 1833 1834 1835 1836 1837 |
/* * Flush the data device before flushing the log to make sure all meta * data written back from the AIL actually made it to disk before * stamping the new log tail LSN into the log buffer. For an external * log we need to issue the flush explicitly, and unfortunately * synchronously here; for an internal log we can simply use the block * layer state machine for preflushes. */ |
2d15d2c0e xfs: make use of ... |
1838 |
if (log->l_targ != log->l_mp->m_ddev_targp || split) { |
2291dab2c xfs: Always flush... |
1839 |
xfs_blkdev_issue_flush(log->l_mp->m_ddev_targp); |
94860a301 xfs: factor out l... |
1840 1841 |
need_flush = false; } |
1da177e4c Linux-2.6.12-rc2 |
1842 |
|
abca1f33f xfs: remove the s... |
1843 |
xlog_verify_iclog(log, iclog, count); |
79b54d9bf xfs: use bios dir... |
1844 |
xlog_write_iclog(log, iclog, bno, count, need_flush); |
94860a301 xfs: factor out l... |
1845 |
} |
1da177e4c Linux-2.6.12-rc2 |
1846 |
|
1da177e4c Linux-2.6.12-rc2 |
1847 |
/* |
c41564b5a [XFS] We really s... |
1848 |
* Deallocate a log structure |
1da177e4c Linux-2.6.12-rc2 |
1849 |
*/ |
a8272ce0c [XFS] Fix up spar... |
1850 |
STATIC void |
9a8d2fdbb xfs: remove xlog_... |
1851 1852 |
xlog_dealloc_log( struct xlog *log) |
1da177e4c Linux-2.6.12-rc2 |
1853 1854 |
{ xlog_in_core_t *iclog, *next_iclog; |
1da177e4c Linux-2.6.12-rc2 |
1855 |
int i; |
71e330b59 xfs: Introduce de... |
1856 |
xlog_cil_destroy(log); |
44396476a xfs: reset buffer... |
1857 |
/* |
9c23eccc1 xfs: unmount does... |
1858 1859 1860 1861 1862 |
* Cycle all the iclogbuf locks to make sure all log IO completion * is done before we tear down these buffers. */ iclog = log->l_iclog; for (i = 0; i < log->l_iclog_bufs; i++) { |
79b54d9bf xfs: use bios dir... |
1863 1864 |
down(&iclog->ic_sema); up(&iclog->ic_sema); |
9c23eccc1 xfs: unmount does... |
1865 1866 |
iclog = iclog->ic_next; } |
1da177e4c Linux-2.6.12-rc2 |
1867 |
iclog = log->l_iclog; |
9c23eccc1 xfs: unmount does... |
1868 |
for (i = 0; i < log->l_iclog_bufs; i++) { |
1da177e4c Linux-2.6.12-rc2 |
1869 |
next_iclog = iclog->ic_next; |
79b54d9bf xfs: use bios dir... |
1870 |
kmem_free(iclog->ic_data); |
f0e2d93c2 [XFS] Remove unus... |
1871 |
kmem_free(iclog); |
1da177e4c Linux-2.6.12-rc2 |
1872 1873 |
iclog = next_iclog; } |
1da177e4c Linux-2.6.12-rc2 |
1874 |
|
1da177e4c Linux-2.6.12-rc2 |
1875 |
log->l_mp->m_log = NULL; |
1058d0f5e xfs: move the log... |
1876 |
destroy_workqueue(log->l_ioend_workqueue); |
f0e2d93c2 [XFS] Remove unus... |
1877 |
kmem_free(log); |
b843299ba xfs: remove some ... |
1878 |
} |
1da177e4c Linux-2.6.12-rc2 |
1879 1880 1881 1882 |
/* * Update counters atomically now that memcpy is done. */ |
1da177e4c Linux-2.6.12-rc2 |
1883 |
static inline void |
9a8d2fdbb xfs: remove xlog_... |
1884 1885 1886 1887 1888 |
xlog_state_finish_copy( struct xlog *log, struct xlog_in_core *iclog, int record_cnt, int copy_bytes) |
1da177e4c Linux-2.6.12-rc2 |
1889 |
{ |
390aab0a1 xfs: move the loc... |
1890 |
lockdep_assert_held(&log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
1891 |
|
413d57c99 xfs: convert beX_... |
1892 |
be32_add_cpu(&iclog->ic_header.h_num_logops, record_cnt); |
1da177e4c Linux-2.6.12-rc2 |
1893 |
iclog->ic_offset += copy_bytes; |
390aab0a1 xfs: move the loc... |
1894 |
} |
1da177e4c Linux-2.6.12-rc2 |
1895 1896 |
/* |
7e9c63961 [XFS] 929956 add ... |
1897 1898 1899 |
* print out info relating to regions written which consume * the reservation */ |
71e330b59 xfs: Introduce de... |
1900 1901 1902 1903 |
void xlog_print_tic_res( struct xfs_mount *mp, struct xlog_ticket *ticket) |
7e9c63961 [XFS] 929956 add ... |
1904 1905 1906 1907 1908 |
{ uint i; uint ophdr_spc = ticket->t_res_num_ophdrs * (uint)sizeof(xlog_op_header_t); /* match with XLOG_REG_TYPE_* in xfs_log.h */ |
5110cd82c xfs: use named ar... |
1909 |
#define REG_TYPE_STR(type, str) [XLOG_REG_TYPE_##type] = str |
d31d71852 xfs: fix broken l... |
1910 |
static char *res_type_str[] = { |
5110cd82c xfs: use named ar... |
1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 |
REG_TYPE_STR(BFORMAT, "bformat"), REG_TYPE_STR(BCHUNK, "bchunk"), REG_TYPE_STR(EFI_FORMAT, "efi_format"), REG_TYPE_STR(EFD_FORMAT, "efd_format"), REG_TYPE_STR(IFORMAT, "iformat"), REG_TYPE_STR(ICORE, "icore"), REG_TYPE_STR(IEXT, "iext"), REG_TYPE_STR(IBROOT, "ibroot"), REG_TYPE_STR(ILOCAL, "ilocal"), REG_TYPE_STR(IATTR_EXT, "iattr_ext"), REG_TYPE_STR(IATTR_BROOT, "iattr_broot"), REG_TYPE_STR(IATTR_LOCAL, "iattr_local"), REG_TYPE_STR(QFORMAT, "qformat"), REG_TYPE_STR(DQUOT, "dquot"), REG_TYPE_STR(QUOTAOFF, "quotaoff"), REG_TYPE_STR(LRHEADER, "LR header"), REG_TYPE_STR(UNMOUNT, "unmount"), REG_TYPE_STR(COMMIT, "commit"), REG_TYPE_STR(TRANSHDR, "trans header"), |
d31d71852 xfs: fix broken l... |
1930 1931 1932 1933 1934 1935 1936 |
REG_TYPE_STR(ICREATE, "inode create"), REG_TYPE_STR(RUI_FORMAT, "rui_format"), REG_TYPE_STR(RUD_FORMAT, "rud_format"), REG_TYPE_STR(CUI_FORMAT, "cui_format"), REG_TYPE_STR(CUD_FORMAT, "cud_format"), REG_TYPE_STR(BUI_FORMAT, "bui_format"), REG_TYPE_STR(BUD_FORMAT, "bud_format"), |
7e9c63961 [XFS] 929956 add ... |
1937 |
}; |
d31d71852 xfs: fix broken l... |
1938 |
BUILD_BUG_ON(ARRAY_SIZE(res_type_str) != XLOG_REG_TYPE_MAX + 1); |
5110cd82c xfs: use named ar... |
1939 |
#undef REG_TYPE_STR |
7e9c63961 [XFS] 929956 add ... |
1940 |
|
7d2d56534 xfs: separate shu... |
1941 |
xfs_warn(mp, "ticket reservation summary:"); |
f41febd2e xfs: Use consiste... |
1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 |
xfs_warn(mp, " unit res = %d bytes", ticket->t_unit_res); xfs_warn(mp, " current res = %d bytes", ticket->t_curr_res); xfs_warn(mp, " total reg = %u bytes (o/flow = %u bytes)", ticket->t_res_arr_sum, ticket->t_res_o_flow); xfs_warn(mp, " ophdrs = %u (ophdr space = %u bytes)", ticket->t_res_num_ophdrs, ophdr_spc); xfs_warn(mp, " ophdr + reg = %u bytes", ticket->t_res_arr_sum + ticket->t_res_o_flow + ophdr_spc); xfs_warn(mp, " num regions = %u", ticket->t_res_num); |
7e9c63961 [XFS] 929956 add ... |
1954 1955 |
for (i = 0; i < ticket->t_res_num; i++) { |
a0fa2b679 xfs: Convert xlog... |
1956 |
uint r_type = ticket->t_res_arr[i].r_type; |
08e96e1a3 xfs: remove newli... |
1957 |
xfs_warn(mp, "region[%u]: %s - %u bytes", i, |
7e9c63961 [XFS] 929956 add ... |
1958 |
((r_type <= 0 || r_type > XLOG_REG_TYPE_MAX) ? |
5110cd82c xfs: use named ar... |
1959 |
"bad-rtype" : res_type_str[r_type]), |
7e9c63961 [XFS] 929956 add ... |
1960 1961 1962 |
ticket->t_res_arr[i].r_len); } } |
7e9c63961 [XFS] 929956 add ... |
1963 1964 |
/* |
d4ca1d550 xfs: dump transac... |
1965 1966 1967 1968 |
* Print a summary of the transaction. */ void xlog_print_trans( |
e6631f855 xfs: get rid of t... |
1969 |
struct xfs_trans *tp) |
d4ca1d550 xfs: dump transac... |
1970 |
{ |
e6631f855 xfs: get rid of t... |
1971 1972 |
struct xfs_mount *mp = tp->t_mountp; struct xfs_log_item *lip; |
d4ca1d550 xfs: dump transac... |
1973 1974 1975 |
/* dump core transaction and ticket info */ xfs_warn(mp, "transaction summary:"); |
2c8f62653 xfs: print transa... |
1976 1977 1978 |
xfs_warn(mp, " log res = %d", tp->t_log_res); xfs_warn(mp, " log count = %d", tp->t_log_count); xfs_warn(mp, " flags = 0x%x", tp->t_flags); |
d4ca1d550 xfs: dump transac... |
1979 1980 1981 1982 |
xlog_print_tic_res(mp, tp->t_ticket); /* dump each log item */ |
e6631f855 xfs: get rid of t... |
1983 |
list_for_each_entry(lip, &tp->t_items, li_trans) { |
d4ca1d550 xfs: dump transac... |
1984 1985 1986 1987 1988 1989 |
struct xfs_log_vec *lv = lip->li_lv; struct xfs_log_iovec *vec; int i; xfs_warn(mp, "log item: "); xfs_warn(mp, " type = 0x%x", lip->li_type); |
22525c17e xfs: log item fla... |
1990 |
xfs_warn(mp, " flags = 0x%lx", lip->li_flags); |
d4ca1d550 xfs: dump transac... |
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 |
if (!lv) continue; xfs_warn(mp, " niovecs = %d", lv->lv_niovecs); xfs_warn(mp, " size = %d", lv->lv_size); xfs_warn(mp, " bytes = %d", lv->lv_bytes); xfs_warn(mp, " buf len = %d", lv->lv_buf_len); /* dump each iovec for the log item */ vec = lv->lv_iovecp; for (i = 0; i < lv->lv_niovecs; i++) { int dumplen = min(vec->i_len, 32); xfs_warn(mp, " iovec[%d]", i); xfs_warn(mp, " type = 0x%x", vec->i_type); xfs_warn(mp, " len = %d", vec->i_len); xfs_warn(mp, " first %d bytes of iovec[%d]:", dumplen, i); |
244e3dea5 xfs: fix semicolo... |
2007 |
xfs_hex_dump(vec->i_addr, dumplen); |
d4ca1d550 xfs: dump transac... |
2008 2009 2010 2011 2012 2013 2014 |
vec++; } } } /* |
7ec949212 xfs: don't try to... |
2015 2016 2017 |
* Calculate the potential space needed by the log vector. We may need a start * record, and each region gets its own struct xlog_op_header and may need to be * double word aligned. |
b5203cd0a xfs: factor xlog_... |
2018 2019 2020 2021 |
*/ static int xlog_write_calc_vec_length( struct xlog_ticket *ticket, |
7ec949212 xfs: don't try to... |
2022 2023 |
struct xfs_log_vec *log_vector, bool need_start_rec) |
b5203cd0a xfs: factor xlog_... |
2024 |
{ |
55b66332d xfs: introduce ne... |
2025 |
struct xfs_log_vec *lv; |
7ec949212 xfs: don't try to... |
2026 |
int headers = need_start_rec ? 1 : 0; |
b5203cd0a xfs: factor xlog_... |
2027 2028 |
int len = 0; int i; |
55b66332d xfs: introduce ne... |
2029 |
for (lv = log_vector; lv; lv = lv->lv_next) { |
fd63875cc xfs: Introduce or... |
2030 2031 2032 |
/* we don't write ordered log vectors */ if (lv->lv_buf_len == XFS_LOG_VEC_ORDERED) continue; |
55b66332d xfs: introduce ne... |
2033 2034 2035 2036 |
headers += lv->lv_niovecs; for (i = 0; i < lv->lv_niovecs; i++) { struct xfs_log_iovec *vecp = &lv->lv_iovecp[i]; |
b5203cd0a xfs: factor xlog_... |
2037 |
|
55b66332d xfs: introduce ne... |
2038 2039 2040 |
len += vecp->i_len; xlog_tic_add_region(ticket, vecp->i_len, vecp->i_type); } |
b5203cd0a xfs: factor xlog_... |
2041 2042 2043 2044 2045 2046 2047 |
} ticket->t_res_num_ophdrs += headers; len += headers * sizeof(struct xlog_op_header); return len; } |
7ec949212 xfs: don't try to... |
2048 |
static void |
b5203cd0a xfs: factor xlog_... |
2049 |
xlog_write_start_rec( |
e6b1f2737 xfs: clean up xlo... |
2050 |
struct xlog_op_header *ophdr, |
b5203cd0a xfs: factor xlog_... |
2051 2052 |
struct xlog_ticket *ticket) { |
b5203cd0a xfs: factor xlog_... |
2053 2054 2055 2056 2057 |
ophdr->oh_tid = cpu_to_be32(ticket->t_tid); ophdr->oh_clientid = ticket->t_clientid; ophdr->oh_len = 0; ophdr->oh_flags = XLOG_START_TRANS; ophdr->oh_res2 = 0; |
b5203cd0a xfs: factor xlog_... |
2058 2059 2060 2061 |
} static xlog_op_header_t * xlog_write_setup_ophdr( |
ad223e603 xfs: rename log s... |
2062 |
struct xlog *log, |
e6b1f2737 xfs: clean up xlo... |
2063 |
struct xlog_op_header *ophdr, |
b5203cd0a xfs: factor xlog_... |
2064 2065 2066 |
struct xlog_ticket *ticket, uint flags) { |
b5203cd0a xfs: factor xlog_... |
2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 |
ophdr->oh_tid = cpu_to_be32(ticket->t_tid); ophdr->oh_clientid = ticket->t_clientid; ophdr->oh_res2 = 0; /* are we copying a commit or unmount record? */ ophdr->oh_flags = flags; /* * We've seen logs corrupted with bad transaction client ids. This * makes sure that XFS doesn't generate them on. Turn this into an EIO * and shut down the filesystem. */ switch (ophdr->oh_clientid) { case XFS_TRANSACTION: case XFS_VOLUME: case XFS_LOG: break; default: |
a0fa2b679 xfs: Convert xlog... |
2085 |
xfs_warn(log->l_mp, |
c96900435 xfs: use %px for ... |
2086 |
"Bad XFS transaction clientid 0x%x in ticket "PTR_FMT, |
b5203cd0a xfs: factor xlog_... |
2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 |
ophdr->oh_clientid, ticket); return NULL; } return ophdr; } /* * Set up the parameters of the region copy into the log. This has * to handle region write split across multiple log buffers - this * state is kept external to this function so that this code can |
ac0e300fa xfs: fix the comm... |
2098 |
* be written in an obvious, self documenting manner. |
b5203cd0a xfs: factor xlog_... |
2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 |
*/ static int xlog_write_setup_copy( struct xlog_ticket *ticket, struct xlog_op_header *ophdr, int space_available, int space_required, int *copy_off, int *copy_len, int *last_was_partial_copy, int *bytes_consumed) { int still_to_copy; still_to_copy = space_required - *bytes_consumed; *copy_off = *bytes_consumed; if (still_to_copy <= space_available) { /* write of region completes here */ *copy_len = still_to_copy; ophdr->oh_len = cpu_to_be32(*copy_len); if (*last_was_partial_copy) ophdr->oh_flags |= (XLOG_END_TRANS|XLOG_WAS_CONT_TRANS); *last_was_partial_copy = 0; *bytes_consumed = 0; return 0; } /* partial write of region, needs extra log op header reservation */ *copy_len = space_available; ophdr->oh_len = cpu_to_be32(*copy_len); ophdr->oh_flags |= XLOG_CONTINUE_TRANS; if (*last_was_partial_copy) ophdr->oh_flags |= XLOG_WAS_CONT_TRANS; *bytes_consumed += *copy_len; (*last_was_partial_copy)++; /* account for new log op header */ ticket->t_curr_res -= sizeof(struct xlog_op_header); ticket->t_res_num_ophdrs++; return sizeof(struct xlog_op_header); } static int xlog_write_copy_finish( |
ad223e603 xfs: rename log s... |
2145 |
struct xlog *log, |
b5203cd0a xfs: factor xlog_... |
2146 2147 2148 2149 2150 2151 2152 2153 2154 |
struct xlog_in_core *iclog, uint flags, int *record_cnt, int *data_cnt, int *partial_copy, int *partial_copy_len, int log_offset, struct xlog_in_core **commit_iclog) { |
df732b29c xfs: call xlog_st... |
2155 |
int error; |
b5203cd0a xfs: factor xlog_... |
2156 2157 2158 2159 2160 |
if (*partial_copy) { /* * This iclog has already been marked WANT_SYNC by * xlog_state_get_iclog_space. */ |
390aab0a1 xfs: move the loc... |
2161 |
spin_lock(&log->l_icloglock); |
b5203cd0a xfs: factor xlog_... |
2162 2163 2164 |
xlog_state_finish_copy(log, iclog, *record_cnt, *data_cnt); *record_cnt = 0; *data_cnt = 0; |
df732b29c xfs: call xlog_st... |
2165 |
goto release_iclog; |
b5203cd0a xfs: factor xlog_... |
2166 2167 2168 2169 2170 2171 2172 |
} *partial_copy = 0; *partial_copy_len = 0; if (iclog->ic_size - log_offset <= sizeof(xlog_op_header_t)) { /* no more space in this iclog - push it. */ |
390aab0a1 xfs: move the loc... |
2173 |
spin_lock(&log->l_icloglock); |
b5203cd0a xfs: factor xlog_... |
2174 2175 2176 |
xlog_state_finish_copy(log, iclog, *record_cnt, *data_cnt); *record_cnt = 0; *data_cnt = 0; |
693639994 xfs: remove xlog_... |
2177 2178 2179 2180 2181 |
if (iclog->ic_state == XLOG_STATE_ACTIVE) xlog_state_switch_iclogs(log, iclog, 0); else ASSERT(iclog->ic_state == XLOG_STATE_WANT_SYNC || iclog->ic_state == XLOG_STATE_IOERROR); |
b5203cd0a xfs: factor xlog_... |
2182 |
if (!commit_iclog) |
df732b29c xfs: call xlog_st... |
2183 2184 |
goto release_iclog; spin_unlock(&log->l_icloglock); |
b5203cd0a xfs: factor xlog_... |
2185 2186 2187 2188 2189 |
ASSERT(flags & XLOG_COMMIT_TRANS); *commit_iclog = iclog; } return 0; |
df732b29c xfs: call xlog_st... |
2190 2191 2192 2193 2194 |
release_iclog: error = xlog_state_release_iclog(log, iclog); spin_unlock(&log->l_icloglock); return error; |
b5203cd0a xfs: factor xlog_... |
2195 2196 2197 |
} /* |
1da177e4c Linux-2.6.12-rc2 |
2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 |
* Write some region out to in-core log * * This will be called when writing externally provided regions or when * writing out a commit record for a given transaction. * * General algorithm: * 1. Find total length of this write. This may include adding to the * lengths passed in. * 2. Check whether we violate the tickets reservation. * 3. While writing to this iclog * A. Reserve as much space in this iclog as can get * B. If this is first write, save away start lsn * C. While writing this region: * 1. If first write of transaction, write start record * 2. Write log operation header (header per region) * 3. Find out if we can fit entire region into this iclog * 4. Potentially, verify destination memcpy ptr * 5. Memcpy (partial) region * 6. If partial copy, release iclog; otherwise, continue * copying more regions into current iclog * 4. Mark want sync bit (in simulation mode) * 5. Release iclog for potential flush to on-disk log. * * ERRORS: * 1. Panic if reservation is overrun. This should never happen since * reservation amounts are generated internal to the filesystem. * NOTES: * 1. Tickets are single threaded data structures. * 2. The XLOG_END_TRANS & XLOG_CONTINUE_TRANS flags are passed down to the * syncing routine. When a single log_write region needs to span * multiple in-core logs, the XLOG_CONTINUE_TRANS bit should be set * on all log operation writes which don't contain the end of the * region. The XLOG_END_TRANS bit is used for the in-core log * operation which contains the end of the continued log_write region. * 3. When xlog_state_get_iclog_space() grabs the rest of the current iclog, * we don't really know exactly how much space will be used. As a result, * we don't update ic_offset until the end when we know exactly how many * bytes have been written out. */ |
71e330b59 xfs: Introduce de... |
2237 |
int |
35a8a72f0 xfs: stop passing... |
2238 |
xlog_write( |
ad223e603 xfs: rename log s... |
2239 |
struct xlog *log, |
55b66332d xfs: introduce ne... |
2240 |
struct xfs_log_vec *log_vector, |
35a8a72f0 xfs: stop passing... |
2241 2242 2243 |
struct xlog_ticket *ticket, xfs_lsn_t *start_lsn, struct xlog_in_core **commit_iclog, |
7ec949212 xfs: don't try to... |
2244 2245 |
uint flags, bool need_start_rec) |
1da177e4c Linux-2.6.12-rc2 |
2246 |
{ |
99428ad0f xfs: reindent xlo... |
2247 |
struct xlog_in_core *iclog = NULL; |
9590e9c68 xfs: re-order ini... |
2248 2249 2250 |
struct xfs_log_vec *lv = log_vector; struct xfs_log_iovec *vecp = lv->lv_iovecp; int index = 0; |
99428ad0f xfs: reindent xlo... |
2251 |
int len; |
99428ad0f xfs: reindent xlo... |
2252 2253 2254 2255 2256 |
int partial_copy = 0; int partial_copy_len = 0; int contwr = 0; int record_cnt = 0; int data_cnt = 0; |
df732b29c xfs: call xlog_st... |
2257 |
int error = 0; |
99428ad0f xfs: reindent xlo... |
2258 |
|
93b8a5854 xfs: remove the d... |
2259 |
/* |
9590e9c68 xfs: re-order ini... |
2260 2261 2262 2263 |
* If this is a commit or unmount transaction, we don't need a start * record to be written. We do, however, have to account for the * commit or unmount header that gets written. Hence we always have * to account for an extra xlog_op_header here. |
93b8a5854 xfs: remove the d... |
2264 |
*/ |
9590e9c68 xfs: re-order ini... |
2265 |
ticket->t_curr_res -= sizeof(struct xlog_op_header); |
7d2d56534 xfs: separate shu... |
2266 2267 2268 |
if (ticket->t_curr_res < 0) { xfs_alert_tag(log->l_mp, XFS_PTAG_LOGRES, "ctx ticket reservation ran out. Need to up reservation"); |
55b66332d xfs: introduce ne... |
2269 |
xlog_print_tic_res(log->l_mp, ticket); |
7d2d56534 xfs: separate shu... |
2270 2271 |
xfs_force_shutdown(log->l_mp, SHUTDOWN_LOG_IO_ERROR); } |
1da177e4c Linux-2.6.12-rc2 |
2272 |
|
7ec949212 xfs: don't try to... |
2273 |
len = xlog_write_calc_vec_length(ticket, log_vector, need_start_rec); |
9590e9c68 xfs: re-order ini... |
2274 |
*start_lsn = 0; |
fd63875cc xfs: Introduce or... |
2275 |
while (lv && (!lv->lv_niovecs || index < lv->lv_niovecs)) { |
e6b1f2737 xfs: clean up xlo... |
2276 |
void *ptr; |
99428ad0f xfs: reindent xlo... |
2277 |
int log_offset; |
1da177e4c Linux-2.6.12-rc2 |
2278 |
|
99428ad0f xfs: reindent xlo... |
2279 2280 2281 2282 |
error = xlog_state_get_iclog_space(log, len, &iclog, ticket, &contwr, &log_offset); if (error) return error; |
1da177e4c Linux-2.6.12-rc2 |
2283 |
|
99428ad0f xfs: reindent xlo... |
2284 |
ASSERT(log_offset <= iclog->ic_size - 1); |
e6b1f2737 xfs: clean up xlo... |
2285 |
ptr = iclog->ic_datap + log_offset; |
1da177e4c Linux-2.6.12-rc2 |
2286 |
|
99428ad0f xfs: reindent xlo... |
2287 2288 2289 |
/* start_lsn is the first lsn written to. That's all we need. */ if (!*start_lsn) *start_lsn = be64_to_cpu(iclog->ic_header.h_lsn); |
b5203cd0a xfs: factor xlog_... |
2290 |
|
99428ad0f xfs: reindent xlo... |
2291 2292 2293 2294 |
/* * This loop writes out as many regions as can fit in the amount * of space which was allocated by xlog_state_get_iclog_space(). */ |
fd63875cc xfs: Introduce or... |
2295 2296 |
while (lv && (!lv->lv_niovecs || index < lv->lv_niovecs)) { struct xfs_log_iovec *reg; |
99428ad0f xfs: reindent xlo... |
2297 |
struct xlog_op_header *ophdr; |
99428ad0f xfs: reindent xlo... |
2298 2299 |
int copy_len; int copy_off; |
fd63875cc xfs: Introduce or... |
2300 2301 2302 2303 2304 2305 2306 2307 |
bool ordered = false; /* ordered log vectors have no regions to write */ if (lv->lv_buf_len == XFS_LOG_VEC_ORDERED) { ASSERT(lv->lv_niovecs == 0); ordered = true; goto next_lv; } |
99428ad0f xfs: reindent xlo... |
2308 |
|
fd63875cc xfs: Introduce or... |
2309 |
reg = &vecp[index]; |
c8ce540db xfs: remove doubl... |
2310 2311 |
ASSERT(reg->i_len % sizeof(int32_t) == 0); ASSERT((unsigned long)ptr % sizeof(int32_t) == 0); |
99428ad0f xfs: reindent xlo... |
2312 |
|
7ec949212 xfs: don't try to... |
2313 2314 2315 2316 2317 2318 2319 |
/* * Before we start formatting log vectors, we need to * write a start record. Only do this for the first * iclog we write to. */ if (need_start_rec) { xlog_write_start_rec(ptr, ticket); |
e6b1f2737 xfs: clean up xlo... |
2320 |
xlog_write_adv_cnt(&ptr, &len, &log_offset, |
7ec949212 xfs: don't try to... |
2321 |
sizeof(struct xlog_op_header)); |
99428ad0f xfs: reindent xlo... |
2322 |
} |
b5203cd0a xfs: factor xlog_... |
2323 |
|
99428ad0f xfs: reindent xlo... |
2324 2325 |
ophdr = xlog_write_setup_ophdr(log, ptr, ticket, flags); if (!ophdr) |
2451337dd xfs: global error... |
2326 |
return -EIO; |
99428ad0f xfs: reindent xlo... |
2327 |
|
e6b1f2737 xfs: clean up xlo... |
2328 |
xlog_write_adv_cnt(&ptr, &len, &log_offset, |
99428ad0f xfs: reindent xlo... |
2329 2330 2331 2332 |
sizeof(struct xlog_op_header)); len += xlog_write_setup_copy(ticket, ophdr, iclog->ic_size-log_offset, |
55b66332d xfs: introduce ne... |
2333 |
reg->i_len, |
99428ad0f xfs: reindent xlo... |
2334 2335 2336 2337 |
©_off, ©_len, &partial_copy, &partial_copy_len); xlog_verify_dest_ptr(log, ptr); |
91f9f5fe1 xfs: avoid null *... |
2338 2339 2340 2341 2342 2343 2344 2345 |
/* * Copy region. * * Unmount records just log an opheader, so can have * empty payloads with no data region to copy. Hence we * only copy the payload if the vector says it has data * to copy. */ |
99428ad0f xfs: reindent xlo... |
2346 |
ASSERT(copy_len >= 0); |
91f9f5fe1 xfs: avoid null *... |
2347 2348 2349 2350 2351 |
if (copy_len > 0) { memcpy(ptr, reg->i_addr + copy_off, copy_len); xlog_write_adv_cnt(&ptr, &len, &log_offset, copy_len); } |
7ec949212 xfs: don't try to... |
2352 |
copy_len += sizeof(struct xlog_op_header); |
99428ad0f xfs: reindent xlo... |
2353 |
record_cnt++; |
7ec949212 xfs: don't try to... |
2354 2355 2356 2357 2358 |
if (need_start_rec) { copy_len += sizeof(struct xlog_op_header); record_cnt++; need_start_rec = false; } |
99428ad0f xfs: reindent xlo... |
2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 |
data_cnt += contwr ? copy_len : 0; error = xlog_write_copy_finish(log, iclog, flags, &record_cnt, &data_cnt, &partial_copy, &partial_copy_len, log_offset, commit_iclog); if (error) return error; /* * if we had a partial copy, we need to get more iclog * space but we don't want to increment the region * index because there is still more is this region to * write. * * If we completed writing this region, and we flushed * the iclog (indicated by resetting of the record * count), then we also need to get more log space. If * this was the last record, though, we are done and * can just return. */ if (partial_copy) break; |
55b66332d xfs: introduce ne... |
2384 |
if (++index == lv->lv_niovecs) { |
fd63875cc xfs: Introduce or... |
2385 |
next_lv: |
55b66332d xfs: introduce ne... |
2386 2387 2388 2389 2390 |
lv = lv->lv_next; index = 0; if (lv) vecp = lv->lv_iovecp; } |
749f24f33 xfs: Fix bool ini... |
2391 |
if (record_cnt == 0 && !ordered) { |
55b66332d xfs: introduce ne... |
2392 |
if (!lv) |
99428ad0f xfs: reindent xlo... |
2393 2394 2395 2396 2397 2398 2399 |
return 0; break; } } } ASSERT(len == 0); |
390aab0a1 xfs: move the loc... |
2400 |
spin_lock(&log->l_icloglock); |
99428ad0f xfs: reindent xlo... |
2401 |
xlog_state_finish_copy(log, iclog, record_cnt, data_cnt); |
df732b29c xfs: call xlog_st... |
2402 2403 2404 2405 2406 2407 |
if (commit_iclog) { ASSERT(flags & XLOG_COMMIT_TRANS); *commit_iclog = iclog; } else { error = xlog_state_release_iclog(log, iclog); } |
390aab0a1 xfs: move the loc... |
2408 |
spin_unlock(&log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
2409 |
|
df732b29c xfs: call xlog_st... |
2410 |
return error; |
99428ad0f xfs: reindent xlo... |
2411 |
} |
1da177e4c Linux-2.6.12-rc2 |
2412 |
|
c814b4f24 xfs: refactor xlo... |
2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 |
static void xlog_state_activate_iclog( struct xlog_in_core *iclog, int *iclogs_changed) { ASSERT(list_empty_careful(&iclog->ic_callbacks)); /* * If the number of ops in this iclog indicate it just contains the * dummy transaction, we can change state into IDLE (the second time * around). Otherwise we should change the state into NEED a dummy. * We don't need to cover the dummy. */ if (*iclogs_changed == 0 && iclog->ic_header.h_num_logops == cpu_to_be32(XLOG_COVER_OPS)) { *iclogs_changed = 1; } else { /* * We have two dirty iclogs so start over. This could also be * num of ops indicating this is not the dummy going out. */ *iclogs_changed = 2; } iclog->ic_state = XLOG_STATE_ACTIVE; iclog->ic_offset = 0; iclog->ic_header.h_num_logops = 0; memset(iclog->ic_header.h_cycle_data, 0, sizeof(iclog->ic_header.h_cycle_data)); iclog->ic_header.h_lsn = 0; } |
0383f543d xfs: push iclog s... |
2444 |
/* |
c814b4f24 xfs: refactor xlo... |
2445 2446 |
* Loop through all iclogs and mark all iclogs currently marked DIRTY as * ACTIVE after iclog I/O has completed. |
1da177e4c Linux-2.6.12-rc2 |
2447 |
*/ |
c814b4f24 xfs: refactor xlo... |
2448 2449 |
static void xlog_state_activate_iclogs( |
0383f543d xfs: push iclog s... |
2450 |
struct xlog *log, |
c814b4f24 xfs: refactor xlo... |
2451 |
int *iclogs_changed) |
1da177e4c Linux-2.6.12-rc2 |
2452 |
{ |
c814b4f24 xfs: refactor xlo... |
2453 |
struct xlog_in_core *iclog = log->l_iclog; |
1da177e4c Linux-2.6.12-rc2 |
2454 |
|
1da177e4c Linux-2.6.12-rc2 |
2455 |
do { |
c814b4f24 xfs: refactor xlo... |
2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 |
if (iclog->ic_state == XLOG_STATE_DIRTY) xlog_state_activate_iclog(iclog, iclogs_changed); /* * The ordering of marking iclogs ACTIVE must be maintained, so * an iclog doesn't become ACTIVE beyond one that is SYNCING. */ else if (iclog->ic_state != XLOG_STATE_ACTIVE) break; } while ((iclog = iclog->ic_next) != log->l_iclog); } |
0383f543d xfs: push iclog s... |
2466 |
|
c814b4f24 xfs: refactor xlo... |
2467 2468 2469 2470 2471 |
static int xlog_covered_state( int prev_state, int iclogs_changed) { |
0383f543d xfs: push iclog s... |
2472 |
/* |
c814b4f24 xfs: refactor xlo... |
2473 2474 2475 |
* We usually go to NEED. But we go to NEED2 if the changed indicates we * are done writing the dummy record. If we are done with the second * dummy recored (DONE2), then we go to IDLE. |
0383f543d xfs: push iclog s... |
2476 |
*/ |
c814b4f24 xfs: refactor xlo... |
2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 |
switch (prev_state) { case XLOG_STATE_COVER_IDLE: case XLOG_STATE_COVER_NEED: case XLOG_STATE_COVER_NEED2: break; case XLOG_STATE_COVER_DONE: if (iclogs_changed == 1) return XLOG_STATE_COVER_NEED2; break; case XLOG_STATE_COVER_DONE2: if (iclogs_changed == 1) return XLOG_STATE_COVER_IDLE; break; default: ASSERT(0); } |
0383f543d xfs: push iclog s... |
2493 |
|
c814b4f24 xfs: refactor xlo... |
2494 2495 |
return XLOG_STATE_COVER_NEED; } |
1da177e4c Linux-2.6.12-rc2 |
2496 |
|
c814b4f24 xfs: refactor xlo... |
2497 2498 2499 2500 2501 2502 |
STATIC void xlog_state_clean_iclog( struct xlog *log, struct xlog_in_core *dirty_iclog) { int iclogs_changed = 0; |
1da177e4c Linux-2.6.12-rc2 |
2503 |
|
5781464bd xfs: move the ioe... |
2504 |
dirty_iclog->ic_state = XLOG_STATE_DIRTY; |
1da177e4c Linux-2.6.12-rc2 |
2505 |
|
c814b4f24 xfs: refactor xlo... |
2506 2507 2508 2509 2510 2511 |
xlog_state_activate_iclogs(log, &iclogs_changed); wake_up_all(&dirty_iclog->ic_force_wait); if (iclogs_changed) { log->l_covered_state = xlog_covered_state(log->l_covered_state, iclogs_changed); |
1da177e4c Linux-2.6.12-rc2 |
2512 |
} |
0383f543d xfs: push iclog s... |
2513 |
} |
1da177e4c Linux-2.6.12-rc2 |
2514 2515 2516 |
STATIC xfs_lsn_t xlog_get_lowest_lsn( |
9bff31325 xfs: reformat xlo... |
2517 |
struct xlog *log) |
1da177e4c Linux-2.6.12-rc2 |
2518 |
{ |
9bff31325 xfs: reformat xlo... |
2519 2520 |
struct xlog_in_core *iclog = log->l_iclog; xfs_lsn_t lowest_lsn = 0, lsn; |
1da177e4c Linux-2.6.12-rc2 |
2521 |
|
1da177e4c Linux-2.6.12-rc2 |
2522 |
do { |
1858bb0be xfs: turn ic_stat... |
2523 2524 |
if (iclog->ic_state == XLOG_STATE_ACTIVE || iclog->ic_state == XLOG_STATE_DIRTY) |
9bff31325 xfs: reformat xlo... |
2525 2526 2527 2528 |
continue; lsn = be64_to_cpu(iclog->ic_header.h_lsn); if ((lsn && !lowest_lsn) || XFS_LSN_CMP(lsn, lowest_lsn) < 0) |
1da177e4c Linux-2.6.12-rc2 |
2529 |
lowest_lsn = lsn; |
9bff31325 xfs: reformat xlo... |
2530 |
} while ((iclog = iclog->ic_next) != log->l_iclog); |
014c2544e return statement ... |
2531 |
return lowest_lsn; |
1da177e4c Linux-2.6.12-rc2 |
2532 |
} |
6546818c8 xfs: factor callb... |
2533 |
/* |
14e15f1bc xfs: push the gra... |
2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 |
* Completion of a iclog IO does not imply that a transaction has completed, as * transactions can be large enough to span many iclogs. We cannot change the * tail of the log half way through a transaction as this may be the only * transaction in the log and moving the tail to point to the middle of it * will prevent recovery from finding the start of the transaction. Hence we * should only update the last_sync_lsn if this iclog contains transaction * completion callbacks on it. * * We have to do this before we drop the icloglock to ensure we are the only one * that can update it. * * If we are moving the last_sync_lsn forwards, we also need to ensure we kick * the reservation grant head pushing. This is due to the fact that the push * target is bound by the current last_sync_lsn value. Hence if we have a large * amount of log space bound up in this committing transaction then the * last_sync_lsn value may be the limiting factor preventing tail pushing from * freeing space in the log. Hence once we've updated the last_sync_lsn we * should push the AIL to ensure the push target (and hence the grant head) is * no longer bound by the old log head location and can move forwards and make * progress again. */ static void xlog_state_set_callback( struct xlog *log, struct xlog_in_core *iclog, xfs_lsn_t header_lsn) { iclog->ic_state = XLOG_STATE_CALLBACK; ASSERT(XFS_LSN_CMP(atomic64_read(&log->l_last_sync_lsn), header_lsn) <= 0); if (list_empty_careful(&iclog->ic_callbacks)) return; atomic64_set(&log->l_last_sync_lsn, header_lsn); xlog_grant_push_ail(log, 0); } /* |
5e96fa8d2 xfs: factor iclog... |
2574 2575 2576 2577 2578 2579 2580 2581 |
* Return true if we need to stop processing, false to continue to the next * iclog. The caller will need to run callbacks if the iclog is returned in the * XLOG_STATE_CALLBACK state. */ static bool xlog_state_iodone_process_iclog( struct xlog *log, struct xlog_in_core *iclog, |
5e96fa8d2 xfs: factor iclog... |
2582 2583 2584 |
bool *ioerror) { xfs_lsn_t lowest_lsn; |
14e15f1bc xfs: push the gra... |
2585 |
xfs_lsn_t header_lsn; |
5e96fa8d2 xfs: factor iclog... |
2586 |
|
1858bb0be xfs: turn ic_stat... |
2587 2588 2589 2590 2591 2592 |
switch (iclog->ic_state) { case XLOG_STATE_ACTIVE: case XLOG_STATE_DIRTY: /* * Skip all iclogs in the ACTIVE & DIRTY states: */ |
5e96fa8d2 xfs: factor iclog... |
2593 |
return false; |
1858bb0be xfs: turn ic_stat... |
2594 2595 2596 2597 2598 |
case XLOG_STATE_IOERROR: /* * Between marking a filesystem SHUTDOWN and stopping the log, * we do flush all iclogs to disk (if there wasn't a log I/O * error). So, we do want things to go smoothly in case of just |
4b29ab04a xfs: remove the X... |
2599 |
* a SHUTDOWN w/o a LOG_IO_ERROR. |
1858bb0be xfs: turn ic_stat... |
2600 |
*/ |
5e96fa8d2 xfs: factor iclog... |
2601 2602 |
*ioerror = true; return false; |
1858bb0be xfs: turn ic_stat... |
2603 |
case XLOG_STATE_DONE_SYNC: |
1858bb0be xfs: turn ic_stat... |
2604 |
/* |
4b29ab04a xfs: remove the X... |
2605 2606 2607 2608 |
* Now that we have an iclog that is in the DONE_SYNC state, do * one more check here to see if we have chased our tail around. * If this is not the lowest lsn iclog, then we will leave it * for another completion to process. |
1858bb0be xfs: turn ic_stat... |
2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 |
*/ header_lsn = be64_to_cpu(iclog->ic_header.h_lsn); lowest_lsn = xlog_get_lowest_lsn(log); if (lowest_lsn && XFS_LSN_CMP(lowest_lsn, header_lsn) < 0) return false; xlog_state_set_callback(log, iclog, header_lsn); return false; default: /* * Can only perform callbacks in order. Since this iclog is not |
4b29ab04a xfs: remove the X... |
2619 2620 |
* in the DONE_SYNC state, we skip the rest and just try to * clean up. |
1858bb0be xfs: turn ic_stat... |
2621 |
*/ |
5e96fa8d2 xfs: factor iclog... |
2622 2623 |
return true; } |
5e96fa8d2 xfs: factor iclog... |
2624 2625 2626 |
} /* |
6546818c8 xfs: factor callb... |
2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 |
* Keep processing entries in the iclog callback list until we come around and * it is empty. We need to atomically see that the list is empty and change the * state to DIRTY so that we don't miss any more callbacks being added. * * This function is called with the icloglock held and returns with it held. We * drop it while running callbacks, however, as holding it over thousands of * callbacks is unnecessary and causes excessive contention if we do. */ static void xlog_state_do_iclog_callbacks( struct xlog *log, |
12e6a0f44 xfs: remove the a... |
2638 |
struct xlog_in_core *iclog) |
f75597935 xfs: annotate fun... |
2639 2640 |
__releases(&log->l_icloglock) __acquires(&log->l_icloglock) |
6546818c8 xfs: factor callb... |
2641 2642 2643 2644 2645 2646 2647 2648 2649 |
{ spin_unlock(&log->l_icloglock); spin_lock(&iclog->ic_callback_lock); while (!list_empty(&iclog->ic_callbacks)) { LIST_HEAD(tmp); list_splice_init(&iclog->ic_callbacks, &tmp); spin_unlock(&iclog->ic_callback_lock); |
12e6a0f44 xfs: remove the a... |
2650 |
xlog_cil_process_committed(&tmp); |
6546818c8 xfs: factor callb... |
2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 |
spin_lock(&iclog->ic_callback_lock); } /* * Pick up the icloglock while still holding the callback lock so we * serialise against anyone trying to add more callbacks to this iclog * now we've finished processing. */ spin_lock(&log->l_icloglock); spin_unlock(&iclog->ic_callback_lock); } |
1da177e4c Linux-2.6.12-rc2 |
2662 2663 |
STATIC void xlog_state_do_callback( |
12e6a0f44 xfs: remove the a... |
2664 |
struct xlog *log) |
1da177e4c Linux-2.6.12-rc2 |
2665 |
{ |
5e96fa8d2 xfs: factor iclog... |
2666 2667 |
struct xlog_in_core *iclog; struct xlog_in_core *first_iclog; |
5e96fa8d2 xfs: factor iclog... |
2668 2669 2670 2671 |
bool cycled_icloglock; bool ioerror; int flushcnt = 0; int repeats = 0; |
1da177e4c Linux-2.6.12-rc2 |
2672 |
|
b22cd72c9 [XFS] Unwrap LOG_... |
2673 |
spin_lock(&log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 |
do { /* * Scan all iclogs starting with the one pointed to by the * log. Reset this starting point each time the log is * unlocked (during callbacks). * * Keep looping through iclogs until one full pass is made * without running any callbacks. */ first_iclog = log->l_iclog; iclog = log->l_iclog; |
6546818c8 xfs: factor callb... |
2685 |
cycled_icloglock = false; |
5e96fa8d2 xfs: factor iclog... |
2686 |
ioerror = false; |
1da177e4c Linux-2.6.12-rc2 |
2687 2688 2689 |
repeats++; do { |
5e96fa8d2 xfs: factor iclog... |
2690 |
if (xlog_state_iodone_process_iclog(log, iclog, |
4b29ab04a xfs: remove the X... |
2691 |
&ioerror)) |
5e96fa8d2 xfs: factor iclog... |
2692 |
break; |
1da177e4c Linux-2.6.12-rc2 |
2693 |
|
1858bb0be xfs: turn ic_stat... |
2694 2695 |
if (iclog->ic_state != XLOG_STATE_CALLBACK && iclog->ic_state != XLOG_STATE_IOERROR) { |
1da177e4c Linux-2.6.12-rc2 |
2696 2697 2698 2699 2700 |
iclog = iclog->ic_next; continue; } /* |
6546818c8 xfs: factor callb... |
2701 2702 |
* Running callbacks will drop the icloglock which means * we'll have to run at least one more complete loop. |
114d23aae [XFS] Per iclog c... |
2703 |
*/ |
6546818c8 xfs: factor callb... |
2704 |
cycled_icloglock = true; |
12e6a0f44 xfs: remove the a... |
2705 |
xlog_state_do_iclog_callbacks(log, iclog); |
5781464bd xfs: move the ioe... |
2706 2707 2708 2709 |
if (XLOG_FORCED_SHUTDOWN(log)) wake_up_all(&iclog->ic_force_wait); else xlog_state_clean_iclog(log, iclog); |
1da177e4c Linux-2.6.12-rc2 |
2710 2711 |
iclog = iclog->ic_next; } while (first_iclog != iclog); |
a3c6685ea [XFS] Ensure xlog... |
2712 2713 2714 2715 |
if (repeats > 5000) { flushcnt += repeats; repeats = 0; |
a0fa2b679 xfs: Convert xlog... |
2716 |
xfs_warn(log->l_mp, |
a3c6685ea [XFS] Ensure xlog... |
2717 |
"%s: possible infinite loop (%d iterations)", |
34a622b2e [XFS] replace rem... |
2718 |
__func__, flushcnt); |
1da177e4c Linux-2.6.12-rc2 |
2719 |
} |
5e96fa8d2 xfs: factor iclog... |
2720 |
} while (!ioerror && cycled_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
2721 |
|
1858bb0be xfs: turn ic_stat... |
2722 2723 |
if (log->l_iclog->ic_state == XLOG_STATE_ACTIVE || log->l_iclog->ic_state == XLOG_STATE_IOERROR) |
eb40a8750 xfs: use wait que... |
2724 |
wake_up_all(&log->l_flush_wait); |
cdea5459c xfs: fix missed w... |
2725 2726 |
spin_unlock(&log->l_icloglock); |
d748c6236 [XFS] Convert l_f... |
2727 |
} |
1da177e4c Linux-2.6.12-rc2 |
2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 |
/* * Finish transitioning this iclog to the dirty state. * * Make sure that we completely execute this routine only when this is * the last call to the iclog. There is a good chance that iclog flushes, * when we reach the end of the physical log, get turned into 2 separate * calls to bwrite. Hence, one iclog flush could generate two calls to this * routine. By using the reference count bwritecnt, we guarantee that only * the second completion goes through. * * Callbacks could take time, so they are done outside the scope of the |
12017faf3 [XFS] clean up st... |
2741 |
* global state machine log lock. |
1da177e4c Linux-2.6.12-rc2 |
2742 |
*/ |
a8272ce0c [XFS] Fix up spar... |
2743 |
STATIC void |
1da177e4c Linux-2.6.12-rc2 |
2744 |
xlog_state_done_syncing( |
12e6a0f44 xfs: remove the a... |
2745 |
struct xlog_in_core *iclog) |
1da177e4c Linux-2.6.12-rc2 |
2746 |
{ |
d15cbf2f3 xfs: stop using X... |
2747 |
struct xlog *log = iclog->ic_log; |
1da177e4c Linux-2.6.12-rc2 |
2748 |
|
b22cd72c9 [XFS] Unwrap LOG_... |
2749 |
spin_lock(&log->l_icloglock); |
155cc6b78 [XFS] Use atomics... |
2750 |
ASSERT(atomic_read(&iclog->ic_refcnt) == 0); |
1da177e4c Linux-2.6.12-rc2 |
2751 2752 2753 |
/* * If we got an error, either on the first buffer, or in the case of |
12e6a0f44 xfs: remove the a... |
2754 2755 |
* split log writes, on the second, we shut down the file system and * no iclogs should ever be attempted to be written to disk again. |
1da177e4c Linux-2.6.12-rc2 |
2756 |
*/ |
12e6a0f44 xfs: remove the a... |
2757 2758 |
if (!XLOG_FORCED_SHUTDOWN(log)) { ASSERT(iclog->ic_state == XLOG_STATE_SYNCING); |
1da177e4c Linux-2.6.12-rc2 |
2759 |
iclog->ic_state = XLOG_STATE_DONE_SYNC; |
12e6a0f44 xfs: remove the a... |
2760 |
} |
1da177e4c Linux-2.6.12-rc2 |
2761 2762 2763 2764 2765 2766 |
/* * Someone could be sleeping prior to writing out the next * iclog buffer, we wake them all, one will get to do the * I/O, the others get to wait for the result. */ |
eb40a8750 xfs: use wait que... |
2767 |
wake_up_all(&iclog->ic_write_wait); |
b22cd72c9 [XFS] Unwrap LOG_... |
2768 |
spin_unlock(&log->l_icloglock); |
b843299ba xfs: remove some ... |
2769 |
xlog_state_do_callback(log); |
12e6a0f44 xfs: remove the a... |
2770 |
} |
1da177e4c Linux-2.6.12-rc2 |
2771 2772 2773 |
/* * If the head of the in-core log ring is not (ACTIVE or DIRTY), then we must |
12017faf3 [XFS] clean up st... |
2774 2775 2776 |
* sleep. We wait on the flush queue on the head iclog as that should be * the first iclog to complete flushing. Hence if all iclogs are syncing, * we will wait here and all new writes will sleep until a sync completes. |
1da177e4c Linux-2.6.12-rc2 |
2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 |
* * The in-core logs are used in a circular fashion. They are not used * out-of-order even when an iclog past the head is free. * * return: * * log_offset where xlog_write() can start writing into the in-core * log's data space. * * in-core log pointer to which xlog_write() should write. * * boolean indicating this is a continued write to an in-core log. * If this is the last write, then the in-core log's offset field * needs to be incremented, depending on the amount of data which * is copied. */ |
a8272ce0c [XFS] Fix up spar... |
2790 |
STATIC int |
9a8d2fdbb xfs: remove xlog_... |
2791 2792 2793 2794 2795 2796 2797 |
xlog_state_get_iclog_space( struct xlog *log, int len, struct xlog_in_core **iclogp, struct xlog_ticket *ticket, int *continued_write, int *logoffsetp) |
1da177e4c Linux-2.6.12-rc2 |
2798 |
{ |
1da177e4c Linux-2.6.12-rc2 |
2799 2800 2801 |
int log_offset; xlog_rec_header_t *head; xlog_in_core_t *iclog; |
1da177e4c Linux-2.6.12-rc2 |
2802 2803 |
restart: |
b22cd72c9 [XFS] Unwrap LOG_... |
2804 |
spin_lock(&log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
2805 |
if (XLOG_FORCED_SHUTDOWN(log)) { |
b22cd72c9 [XFS] Unwrap LOG_... |
2806 |
spin_unlock(&log->l_icloglock); |
2451337dd xfs: global error... |
2807 |
return -EIO; |
1da177e4c Linux-2.6.12-rc2 |
2808 2809 2810 |
} iclog = log->l_iclog; |
d748c6236 [XFS] Convert l_f... |
2811 |
if (iclog->ic_state != XLOG_STATE_ACTIVE) { |
ff6d6af23 xfs: per-filesyst... |
2812 |
XFS_STATS_INC(log->l_mp, xs_log_noiclogs); |
d748c6236 [XFS] Convert l_f... |
2813 2814 |
/* Wait for log writes to have flushed */ |
eb40a8750 xfs: use wait que... |
2815 |
xlog_wait(&log->l_flush_wait, &log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
2816 2817 |
goto restart; } |
d748c6236 [XFS] Convert l_f... |
2818 |
|
1da177e4c Linux-2.6.12-rc2 |
2819 |
head = &iclog->ic_header; |
155cc6b78 [XFS] Use atomics... |
2820 |
atomic_inc(&iclog->ic_refcnt); /* prevents sync */ |
1da177e4c Linux-2.6.12-rc2 |
2821 2822 2823 2824 2825 2826 2827 2828 2829 |
log_offset = iclog->ic_offset; /* On the 1st write to an iclog, figure out lsn. This works * if iclogs marked XLOG_STATE_WANT_SYNC always write out what they are * committing to. If the offset is set, that's how many blocks * must be written. */ if (log_offset == 0) { ticket->t_curr_res -= log->l_iclog_hsize; |
0adba5363 [XFS] replace som... |
2830 |
xlog_tic_add_region(ticket, |
7e9c63961 [XFS] 929956 add ... |
2831 2832 |
log->l_iclog_hsize, XLOG_REG_TYPE_LRHEADER); |
b53e675dc [XFS] xlog_rec_he... |
2833 2834 |
head->h_cycle = cpu_to_be32(log->l_curr_cycle); head->h_lsn = cpu_to_be64( |
03bea6fe6 [XFS] clean up so... |
2835 |
xlog_assign_lsn(log->l_curr_cycle, log->l_curr_block)); |
1da177e4c Linux-2.6.12-rc2 |
2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 |
ASSERT(log->l_curr_block >= 0); } /* If there is enough room to write everything, then do it. Otherwise, * claim the rest of the region and make sure the XLOG_STATE_WANT_SYNC * bit is on, so this will get flushed out. Don't update ic_offset * until you know exactly how many bytes get copied. Therefore, wait * until later to update ic_offset. * * xlog_write() algorithm assumes that at least 2 xlog_op_header_t's * can fit into remaining data section. */ if (iclog->ic_size - iclog->ic_offset < 2*sizeof(xlog_op_header_t)) { |
df732b29c xfs: call xlog_st... |
2849 |
int error = 0; |
1da177e4c Linux-2.6.12-rc2 |
2850 |
xlog_state_switch_iclogs(log, iclog, iclog->ic_size); |
49641f1ac Fix reference cou... |
2851 |
/* |
df732b29c xfs: call xlog_st... |
2852 2853 2854 |
* If we are the only one writing to this iclog, sync it to * disk. We need to do an atomic compare and decrement here to * avoid racing with concurrent atomic_dec_and_lock() calls in |
49641f1ac Fix reference cou... |
2855 2856 2857 |
* xlog_state_release_iclog() when there is more than one * reference to the iclog. */ |
df732b29c xfs: call xlog_st... |
2858 |
if (!atomic_add_unless(&iclog->ic_refcnt, -1, 1)) |
49641f1ac Fix reference cou... |
2859 |
error = xlog_state_release_iclog(log, iclog); |
df732b29c xfs: call xlog_st... |
2860 2861 2862 |
spin_unlock(&log->l_icloglock); if (error) return error; |
1da177e4c Linux-2.6.12-rc2 |
2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 |
goto restart; } /* Do we have enough room to write the full amount in the remainder * of this iclog? Or must we continue a write on the next iclog and * mark this iclog as completely taken? In the case where we switch * iclogs (to mark it taken), this particular iclog will release/sync * to disk in xlog_write(). */ if (len <= iclog->ic_size - iclog->ic_offset) { *continued_write = 0; iclog->ic_offset += len; } else { *continued_write = 1; xlog_state_switch_iclogs(log, iclog, iclog->ic_size); } *iclogp = iclog; ASSERT(iclog->ic_offset <= iclog->ic_size); |
b22cd72c9 [XFS] Unwrap LOG_... |
2882 |
spin_unlock(&log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
2883 2884 2885 |
*logoffsetp = log_offset; return 0; |
b843299ba xfs: remove some ... |
2886 |
} |
1da177e4c Linux-2.6.12-rc2 |
2887 |
|
8b41e3f98 xfs: split xlog_t... |
2888 |
/* |
b843299ba xfs: remove some ... |
2889 2890 2891 2892 2893 |
* The first cnt-1 times a ticket goes through here we don't need to move the * grant write head because the permanent reservation has reserved cnt times the * unit amount. Release part of current permanent unit reservation and reset * current reservation to be one units worth. Also move grant reservation head * forward. |
1da177e4c Linux-2.6.12-rc2 |
2894 |
*/ |
8b41e3f98 xfs: split xlog_t... |
2895 2896 |
void xfs_log_ticket_regrant( |
9a8d2fdbb xfs: remove xlog_... |
2897 2898 |
struct xlog *log, struct xlog_ticket *ticket) |
1da177e4c Linux-2.6.12-rc2 |
2899 |
{ |
8b41e3f98 xfs: split xlog_t... |
2900 |
trace_xfs_log_ticket_regrant(log, ticket); |
0b1b213fc xfs: event tracin... |
2901 |
|
1da177e4c Linux-2.6.12-rc2 |
2902 2903 |
if (ticket->t_cnt > 0) ticket->t_cnt--; |
28496968a xfs: add the xlog... |
2904 |
xlog_grant_sub_space(log, &log->l_reserve_head.grant, |
a69ed03c2 xfs: combine gran... |
2905 |
ticket->t_curr_res); |
28496968a xfs: add the xlog... |
2906 |
xlog_grant_sub_space(log, &log->l_write_head.grant, |
a69ed03c2 xfs: combine gran... |
2907 |
ticket->t_curr_res); |
1da177e4c Linux-2.6.12-rc2 |
2908 |
ticket->t_curr_res = ticket->t_unit_res; |
0adba5363 [XFS] replace som... |
2909 |
xlog_tic_reset_res(ticket); |
0b1b213fc xfs: event tracin... |
2910 |
|
8b41e3f98 xfs: split xlog_t... |
2911 |
trace_xfs_log_ticket_regrant_sub(log, ticket); |
0b1b213fc xfs: event tracin... |
2912 |
|
1da177e4c Linux-2.6.12-rc2 |
2913 |
/* just return if we still have some of the pre-reserved space */ |
8b41e3f98 xfs: split xlog_t... |
2914 2915 2916 2917 |
if (!ticket->t_cnt) { xlog_grant_add_space(log, &log->l_reserve_head.grant, ticket->t_unit_res); trace_xfs_log_ticket_regrant_exit(log, ticket); |
1da177e4c Linux-2.6.12-rc2 |
2918 |
|
8b41e3f98 xfs: split xlog_t... |
2919 2920 2921 |
ticket->t_curr_res = ticket->t_unit_res; xlog_tic_reset_res(ticket); } |
1da177e4c Linux-2.6.12-rc2 |
2922 |
|
8b41e3f98 xfs: split xlog_t... |
2923 2924 |
xfs_log_ticket_put(ticket); } |
1da177e4c Linux-2.6.12-rc2 |
2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 |
/* * Give back the space left from a reservation. * * All the information we need to make a correct determination of space left * is present. For non-permanent reservations, things are quite easy. The * count should have been decremented to zero. We only need to deal with the * space remaining in the current reservation part of the ticket. If the * ticket contains a permanent reservation, there may be left over space which * needs to be released. A count of N means that N-1 refills of the current * reservation can be done before we need to ask for more space. The first * one goes to fill up the first current reservation. Once we run out of * space, the count will stay at zero and the only space remaining will be * in the current reservation field. */ |
8b41e3f98 xfs: split xlog_t... |
2940 2941 |
void xfs_log_ticket_ungrant( |
9a8d2fdbb xfs: remove xlog_... |
2942 2943 |
struct xlog *log, struct xlog_ticket *ticket) |
1da177e4c Linux-2.6.12-rc2 |
2944 |
{ |
8b41e3f98 xfs: split xlog_t... |
2945 2946 2947 |
int bytes; trace_xfs_log_ticket_ungrant(log, ticket); |
663e496a7 xfs: rework log g... |
2948 |
|
1da177e4c Linux-2.6.12-rc2 |
2949 2950 |
if (ticket->t_cnt > 0) ticket->t_cnt--; |
8b41e3f98 xfs: split xlog_t... |
2951 |
trace_xfs_log_ticket_ungrant_sub(log, ticket); |
1da177e4c Linux-2.6.12-rc2 |
2952 |
|
663e496a7 xfs: rework log g... |
2953 2954 |
/* * If this is a permanent reservation ticket, we may be able to free |
1da177e4c Linux-2.6.12-rc2 |
2955 2956 |
* up more space based on the remaining count. */ |
663e496a7 xfs: rework log g... |
2957 |
bytes = ticket->t_curr_res; |
1da177e4c Linux-2.6.12-rc2 |
2958 2959 |
if (ticket->t_cnt > 0) { ASSERT(ticket->t_flags & XLOG_TIC_PERM_RESERV); |
663e496a7 xfs: rework log g... |
2960 |
bytes += ticket->t_unit_res*ticket->t_cnt; |
1da177e4c Linux-2.6.12-rc2 |
2961 |
} |
28496968a xfs: add the xlog... |
2962 2963 |
xlog_grant_sub_space(log, &log->l_reserve_head.grant, bytes); xlog_grant_sub_space(log, &log->l_write_head.grant, bytes); |
663e496a7 xfs: rework log g... |
2964 |
|
8b41e3f98 xfs: split xlog_t... |
2965 |
trace_xfs_log_ticket_ungrant_exit(log, ticket); |
0b1b213fc xfs: event tracin... |
2966 |
|
cfb7cdca0 xfs: cleanup xfs_... |
2967 |
xfs_log_space_wake(log->l_mp); |
8b41e3f98 xfs: split xlog_t... |
2968 |
xfs_log_ticket_put(ticket); |
09a423a3d xfs: split tail_l... |
2969 |
} |
1da177e4c Linux-2.6.12-rc2 |
2970 2971 |
/* |
b843299ba xfs: remove some ... |
2972 2973 |
* This routine will mark the current iclog in the ring as WANT_SYNC and move * the current iclog pointer to the next iclog in the ring. |
1da177e4c Linux-2.6.12-rc2 |
2974 2975 |
*/ STATIC void |
9a8d2fdbb xfs: remove xlog_... |
2976 2977 2978 2979 |
xlog_state_switch_iclogs( struct xlog *log, struct xlog_in_core *iclog, int eventual_size) |
1da177e4c Linux-2.6.12-rc2 |
2980 2981 |
{ ASSERT(iclog->ic_state == XLOG_STATE_ACTIVE); |
693639994 xfs: remove xlog_... |
2982 |
assert_spin_locked(&log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
2983 2984 2985 |
if (!eventual_size) eventual_size = iclog->ic_offset; iclog->ic_state = XLOG_STATE_WANT_SYNC; |
b53e675dc [XFS] xlog_rec_he... |
2986 |
iclog->ic_header.h_prev_block = cpu_to_be32(log->l_prev_block); |
1da177e4c Linux-2.6.12-rc2 |
2987 2988 2989 2990 2991 2992 2993 |
log->l_prev_block = log->l_curr_block; log->l_prev_cycle = log->l_curr_cycle; /* roll log?: ic_offset changed later */ log->l_curr_block += BTOBB(eventual_size)+BTOBB(log->l_iclog_hsize); /* Round up to next log-sunit */ |
621187099 [XFS] remove shou... |
2994 |
if (xfs_sb_version_haslogv2(&log->l_mp->m_sb) && |
1da177e4c Linux-2.6.12-rc2 |
2995 |
log->l_mp->m_sb.sb_logsunit > 1) { |
c8ce540db xfs: remove doubl... |
2996 |
uint32_t sunit_bb = BTOBB(log->l_mp->m_sb.sb_logsunit); |
1da177e4c Linux-2.6.12-rc2 |
2997 2998 2999 3000 |
log->l_curr_block = roundup(log->l_curr_block, sunit_bb); } if (log->l_curr_block >= log->l_logBBsize) { |
a45086e27 xfs: validate met... |
3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 |
/* * Rewind the current block before the cycle is bumped to make * sure that the combined LSN never transiently moves forward * when the log wraps to the next cycle. This is to support the * unlocked sample of these fields from xlog_valid_lsn(). Most * other cases should acquire l_icloglock. */ log->l_curr_block -= log->l_logBBsize; ASSERT(log->l_curr_block >= 0); smp_wmb(); |
1da177e4c Linux-2.6.12-rc2 |
3011 3012 3013 |
log->l_curr_cycle++; if (log->l_curr_cycle == XLOG_HEADER_MAGIC_NUM) log->l_curr_cycle++; |
1da177e4c Linux-2.6.12-rc2 |
3014 3015 3016 |
} ASSERT(iclog == log->l_iclog); log->l_iclog = iclog->ic_next; |
b843299ba xfs: remove some ... |
3017 |
} |
1da177e4c Linux-2.6.12-rc2 |
3018 |
|
1da177e4c Linux-2.6.12-rc2 |
3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 |
/* * Write out all data in the in-core log as of this exact moment in time. * * Data may be written to the in-core log during this call. However, * we don't guarantee this data will be written out. A change from past * implementation means this routine will *not* write out zero length LRs. * * Basically, we try and perform an intelligent scan of the in-core logs. * If we determine there is no flushable data, we just return. There is no * flushable data if: * * 1. the current iclog is active and has no data; the previous iclog * is in the active or dirty state. * 2. the current iclog is drity, and the previous iclog is in the * active or dirty state. * |
12017faf3 [XFS] clean up st... |
3035 |
* We may sleep if: |
1da177e4c Linux-2.6.12-rc2 |
3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 |
* * 1. the current iclog is not in the active nor dirty state. * 2. the current iclog dirty, and the previous iclog is not in the * active nor dirty state. * 3. the current iclog is active, and there is another thread writing * to this particular iclog. * 4. a) the current iclog is active and has no other writers * b) when we return from flushing out this iclog, it is still * not in the active nor dirty state. */ |
a14a348bf xfs: cleanup up x... |
3046 |
int |
60e5bb784 xfs: merge _xfs_l... |
3047 |
xfs_log_force( |
a14a348bf xfs: cleanup up x... |
3048 |
struct xfs_mount *mp, |
60e5bb784 xfs: merge _xfs_l... |
3049 |
uint flags) |
1da177e4c Linux-2.6.12-rc2 |
3050 |
{ |
ad223e603 xfs: rename log s... |
3051 |
struct xlog *log = mp->m_log; |
a14a348bf xfs: cleanup up x... |
3052 3053 |
struct xlog_in_core *iclog; xfs_lsn_t lsn; |
ff6d6af23 xfs: per-filesyst... |
3054 |
XFS_STATS_INC(mp, xs_log_force); |
60e5bb784 xfs: merge _xfs_l... |
3055 |
trace_xfs_log_force(mp, 0, _RET_IP_); |
1da177e4c Linux-2.6.12-rc2 |
3056 |
|
93b8a5854 xfs: remove the d... |
3057 |
xlog_cil_force(log); |
71e330b59 xfs: Introduce de... |
3058 |
|
b22cd72c9 [XFS] Unwrap LOG_... |
3059 |
spin_lock(&log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
3060 |
iclog = log->l_iclog; |
1858bb0be xfs: turn ic_stat... |
3061 |
if (iclog->ic_state == XLOG_STATE_IOERROR) |
e6b965705 xfs: refactor xfs... |
3062 |
goto out_error; |
1da177e4c Linux-2.6.12-rc2 |
3063 |
|
e6b965705 xfs: refactor xfs... |
3064 3065 3066 |
if (iclog->ic_state == XLOG_STATE_DIRTY || (iclog->ic_state == XLOG_STATE_ACTIVE && atomic_read(&iclog->ic_refcnt) == 0 && iclog->ic_offset == 0)) { |
1da177e4c Linux-2.6.12-rc2 |
3067 |
/* |
e6b965705 xfs: refactor xfs... |
3068 3069 3070 3071 3072 |
* If the head is dirty or (active and empty), then we need to * look at the previous iclog. * * If the previous iclog is active or dirty we are done. There * is nothing to sync out. Otherwise, we attach ourselves to the |
1da177e4c Linux-2.6.12-rc2 |
3073 3074 |
* previous iclog and go to sleep. */ |
e6b965705 xfs: refactor xfs... |
3075 |
iclog = iclog->ic_prev; |
e6b965705 xfs: refactor xfs... |
3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 |
} else if (iclog->ic_state == XLOG_STATE_ACTIVE) { if (atomic_read(&iclog->ic_refcnt) == 0) { /* * We are the only one with access to this iclog. * * Flush it out now. There should be a roundoff of zero * to show that someone has already taken care of the * roundoff from the previous sync. */ atomic_inc(&iclog->ic_refcnt); lsn = be64_to_cpu(iclog->ic_header.h_lsn); xlog_state_switch_iclogs(log, iclog, 0); |
e6b965705 xfs: refactor xfs... |
3088 |
if (xlog_state_release_iclog(log, iclog)) |
df732b29c xfs: call xlog_st... |
3089 |
goto out_error; |
1da177e4c Linux-2.6.12-rc2 |
3090 |
|
81e5b50a8 xfs: factor out a... |
3091 |
if (be64_to_cpu(iclog->ic_header.h_lsn) != lsn) |
e6b965705 xfs: refactor xfs... |
3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 |
goto out_unlock; } else { /* * Someone else is writing to this iclog. * * Use its call to flush out the data. However, the * other thread may not force out this LR, so we mark * it WANT_SYNC. */ xlog_state_switch_iclogs(log, iclog, 0); |
1da177e4c Linux-2.6.12-rc2 |
3102 |
} |
e6b965705 xfs: refactor xfs... |
3103 |
} else { |
1da177e4c Linux-2.6.12-rc2 |
3104 |
/* |
e6b965705 xfs: refactor xfs... |
3105 3106 |
* If the head iclog is not active nor dirty, we just attach * ourselves to the head and go to sleep if necessary. |
1da177e4c Linux-2.6.12-rc2 |
3107 |
*/ |
e6b965705 xfs: refactor xfs... |
3108 |
; |
1da177e4c Linux-2.6.12-rc2 |
3109 |
} |
e6b965705 xfs: refactor xfs... |
3110 |
|
81e5b50a8 xfs: factor out a... |
3111 3112 |
if (flags & XFS_LOG_SYNC) return xlog_wait_on_iclog(iclog); |
e6b965705 xfs: refactor xfs... |
3113 3114 3115 3116 3117 3118 |
out_unlock: spin_unlock(&log->l_icloglock); return 0; out_error: spin_unlock(&log->l_icloglock); return -EIO; |
a14a348bf xfs: cleanup up x... |
3119 |
} |
1da177e4c Linux-2.6.12-rc2 |
3120 |
|
3e4da466b xfs: unwind the t... |
3121 3122 |
static int __xfs_log_force_lsn( |
a14a348bf xfs: cleanup up x... |
3123 3124 3125 |
struct xfs_mount *mp, xfs_lsn_t lsn, uint flags, |
3e4da466b xfs: unwind the t... |
3126 3127 |
int *log_flushed, bool already_slept) |
1da177e4c Linux-2.6.12-rc2 |
3128 |
{ |
ad223e603 xfs: rename log s... |
3129 |
struct xlog *log = mp->m_log; |
a14a348bf xfs: cleanup up x... |
3130 |
struct xlog_in_core *iclog; |
71e330b59 xfs: Introduce de... |
3131 |
|
a14a348bf xfs: cleanup up x... |
3132 3133 |
spin_lock(&log->l_icloglock); iclog = log->l_iclog; |
1858bb0be xfs: turn ic_stat... |
3134 |
if (iclog->ic_state == XLOG_STATE_IOERROR) |
93806299b xfs: refactor xfs... |
3135 |
goto out_error; |
1da177e4c Linux-2.6.12-rc2 |
3136 |
|
93806299b xfs: refactor xfs... |
3137 3138 3139 3140 3141 |
while (be64_to_cpu(iclog->ic_header.h_lsn) != lsn) { iclog = iclog->ic_next; if (iclog == log->l_iclog) goto out_unlock; } |
a14a348bf xfs: cleanup up x... |
3142 |
|
93806299b xfs: refactor xfs... |
3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 |
if (iclog->ic_state == XLOG_STATE_ACTIVE) { /* * We sleep here if we haven't already slept (e.g. this is the * first time we've looked at the correct iclog buf) and the * buffer before us is going to be sync'ed. The reason for this * is that if we are doing sync transactions here, by waiting * for the previous I/O to complete, we can allow a few more * transactions into this iclog before we close it down. * * Otherwise, we mark the buffer WANT_SYNC, and bump up the * refcnt so we can release the log (which drops the ref count). * The state switch keeps new transaction commits from using * this buffer. When the current commits finish writing into * the buffer, the refcount will drop to zero and the buffer * will go out then. */ if (!already_slept && |
1858bb0be xfs: turn ic_stat... |
3160 3161 |
(iclog->ic_prev->ic_state == XLOG_STATE_WANT_SYNC || iclog->ic_prev->ic_state == XLOG_STATE_SYNCING)) { |
93806299b xfs: refactor xfs... |
3162 |
XFS_STATS_INC(mp, xs_log_force_sleep); |
a14a348bf xfs: cleanup up x... |
3163 |
|
93806299b xfs: refactor xfs... |
3164 3165 |
xlog_wait(&iclog->ic_prev->ic_write_wait, &log->l_icloglock); |
3e4da466b xfs: unwind the t... |
3166 |
return -EAGAIN; |
1da177e4c Linux-2.6.12-rc2 |
3167 |
} |
93806299b xfs: refactor xfs... |
3168 3169 |
atomic_inc(&iclog->ic_refcnt); xlog_state_switch_iclogs(log, iclog, 0); |
93806299b xfs: refactor xfs... |
3170 |
if (xlog_state_release_iclog(log, iclog)) |
df732b29c xfs: call xlog_st... |
3171 |
goto out_error; |
93806299b xfs: refactor xfs... |
3172 3173 |
if (log_flushed) *log_flushed = 1; |
93806299b xfs: refactor xfs... |
3174 |
} |
1da177e4c Linux-2.6.12-rc2 |
3175 |
|
81e5b50a8 xfs: factor out a... |
3176 3177 |
if (flags & XFS_LOG_SYNC) return xlog_wait_on_iclog(iclog); |
93806299b xfs: refactor xfs... |
3178 |
out_unlock: |
a14a348bf xfs: cleanup up x... |
3179 3180 |
spin_unlock(&log->l_icloglock); return 0; |
93806299b xfs: refactor xfs... |
3181 3182 3183 |
out_error: spin_unlock(&log->l_icloglock); return -EIO; |
a14a348bf xfs: cleanup up x... |
3184 3185 3186 |
} /* |
3e4da466b xfs: unwind the t... |
3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 |
* Force the in-core log to disk for a specific LSN. * * Find in-core log with lsn. * If it is in the DIRTY state, just return. * If it is in the ACTIVE state, move the in-core log into the WANT_SYNC * state and go to sleep or return. * If it is in any other state, go to sleep or return. * * Synchronous forces are implemented with a wait queue. All callers trying * to force a given lsn to disk must wait on the queue attached to the * specific in-core log. When given in-core log finally completes its write * to disk, that thread will wake up all threads waiting on the queue. */ int xfs_log_force_lsn( struct xfs_mount *mp, xfs_lsn_t lsn, uint flags, int *log_flushed) { int ret; ASSERT(lsn != 0); XFS_STATS_INC(mp, xs_log_force); trace_xfs_log_force(mp, lsn, _RET_IP_); lsn = xlog_cil_force_lsn(mp->m_log, lsn); if (lsn == NULLCOMMITLSN) return 0; ret = __xfs_log_force_lsn(mp, lsn, flags, log_flushed, false); if (ret == -EAGAIN) ret = __xfs_log_force_lsn(mp, lsn, flags, log_flushed, true); return ret; } |
1da177e4c Linux-2.6.12-rc2 |
3222 |
/* |
9da096fd1 xfs: fix various ... |
3223 |
* Free a used ticket when its refcount falls to zero. |
1da177e4c Linux-2.6.12-rc2 |
3224 |
*/ |
cc09c0dc5 [XFS] Fix double ... |
3225 3226 3227 |
void xfs_log_ticket_put( xlog_ticket_t *ticket) |
1da177e4c Linux-2.6.12-rc2 |
3228 |
{ |
cc09c0dc5 [XFS] Fix double ... |
3229 |
ASSERT(atomic_read(&ticket->t_ref) > 0); |
eb40a8750 xfs: use wait que... |
3230 |
if (atomic_dec_and_test(&ticket->t_ref)) |
377bcd5f3 xfs: Remove kmem_... |
3231 |
kmem_cache_free(xfs_log_ticket_zone, ticket); |
cc09c0dc5 [XFS] Fix double ... |
3232 |
} |
1da177e4c Linux-2.6.12-rc2 |
3233 |
|
cc09c0dc5 [XFS] Fix double ... |
3234 3235 3236 3237 3238 3239 3240 3241 |
xlog_ticket_t * xfs_log_ticket_get( xlog_ticket_t *ticket) { ASSERT(atomic_read(&ticket->t_ref) > 0); atomic_inc(&ticket->t_ref); return ticket; } |
1da177e4c Linux-2.6.12-rc2 |
3242 3243 |
/* |
e773fc934 xfs: Refactor xfs... |
3244 3245 |
* Figure out the total log space unit (in bytes) that would be * required for a log ticket. |
1da177e4c Linux-2.6.12-rc2 |
3246 |
*/ |
e773fc934 xfs: Refactor xfs... |
3247 3248 3249 3250 |
int xfs_log_calc_unit_res( struct xfs_mount *mp, int unit_bytes) |
1da177e4c Linux-2.6.12-rc2 |
3251 |
{ |
e773fc934 xfs: Refactor xfs... |
3252 3253 3254 |
struct xlog *log = mp->m_log; int iclog_space; uint num_headers; |
1da177e4c Linux-2.6.12-rc2 |
3255 3256 3257 3258 3259 3260 3261 3262 |
/* * Permanent reservations have up to 'cnt'-1 active log operations * in the log. A unit in this case is the amount of space for one * of these log operations. Normal reservations have a cnt of 1 * and their unit amount is the total amount of space required. * * The following lines of code account for non-transaction data |
32fb9b57a [XFS] Fix up the ... |
3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 |
* which occupy space in the on-disk log. * * Normal form of a transaction is: * <oph><trans-hdr><start-oph><reg1-oph><reg1><reg2-oph>...<commit-oph> * and then there are LR hdrs, split-recs and roundoff at end of syncs. * * We need to account for all the leadup data and trailer data * around the transaction data. * And then we need to account for the worst case in terms of using * more space. * The worst case will happen if: * - the placement of the transaction happens to be such that the * roundoff is at its maximum * - the transaction data is synced before the commit record is synced * i.e. <transaction-data><roundoff> | <commit-rec><roundoff> * Therefore the commit record is in its own Log Record. * This can happen as the commit record is called with its * own region to xlog_write(). * This then means that in the worst case, roundoff can happen for * the commit-rec as well. * The commit-rec is smaller than padding in this scenario and so it is * not added separately. |
1da177e4c Linux-2.6.12-rc2 |
3285 |
*/ |
32fb9b57a [XFS] Fix up the ... |
3286 3287 3288 |
/* for trans header */ unit_bytes += sizeof(xlog_op_header_t); unit_bytes += sizeof(xfs_trans_header_t); |
1da177e4c Linux-2.6.12-rc2 |
3289 |
/* for start-rec */ |
32fb9b57a [XFS] Fix up the ... |
3290 |
unit_bytes += sizeof(xlog_op_header_t); |
9b9fc2b76 xfs: log ticket r... |
3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 |
/* * for LR headers - the space for data in an iclog is the size minus * the space used for the headers. If we use the iclog size, then we * undercalculate the number of headers required. * * Furthermore - the addition of op headers for split-recs might * increase the space required enough to require more log and op * headers, so take that into account too. * * IMPORTANT: This reservation makes the assumption that if this * transaction is the first in an iclog and hence has the LR headers * accounted to it, then the remaining space in the iclog is * exclusively for this transaction. i.e. if the transaction is larger * than the iclog, it will be the only thing in that iclog. * Fundamentally, this means we must pass the entire log vector to * xlog_write to guarantee this. */ iclog_space = log->l_iclog_size - log->l_iclog_hsize; num_headers = howmany(unit_bytes, iclog_space); /* for split-recs - ophdrs added when data split over LRs */ unit_bytes += sizeof(xlog_op_header_t) * num_headers; /* add extra header reservations if we overrun */ while (!num_headers || howmany(unit_bytes, iclog_space) > num_headers) { unit_bytes += sizeof(xlog_op_header_t); num_headers++; } |
32fb9b57a [XFS] Fix up the ... |
3320 |
unit_bytes += log->l_iclog_hsize * num_headers; |
1da177e4c Linux-2.6.12-rc2 |
3321 |
|
32fb9b57a [XFS] Fix up the ... |
3322 3323 |
/* for commit-rec LR header - note: padding will subsume the ophdr */ unit_bytes += log->l_iclog_hsize; |
32fb9b57a [XFS] Fix up the ... |
3324 |
/* for roundoff padding for transaction data and one for commit record */ |
e773fc934 xfs: Refactor xfs... |
3325 |
if (xfs_sb_version_haslogv2(&mp->m_sb) && mp->m_sb.sb_logsunit > 1) { |
1da177e4c Linux-2.6.12-rc2 |
3326 |
/* log su roundoff */ |
e773fc934 xfs: Refactor xfs... |
3327 |
unit_bytes += 2 * mp->m_sb.sb_logsunit; |
1da177e4c Linux-2.6.12-rc2 |
3328 3329 |
} else { /* BB roundoff */ |
e773fc934 xfs: Refactor xfs... |
3330 |
unit_bytes += 2 * BBSIZE; |
1da177e4c Linux-2.6.12-rc2 |
3331 |
} |
e773fc934 xfs: Refactor xfs... |
3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 |
return unit_bytes; } /* * Allocate and initialise a new log ticket. */ struct xlog_ticket * xlog_ticket_alloc( struct xlog *log, int unit_bytes, int cnt, char client, |
ca4f25899 xfs: Modify xlog_... |
3344 |
bool permanent) |
e773fc934 xfs: Refactor xfs... |
3345 3346 3347 |
{ struct xlog_ticket *tic; int unit_res; |
ca4f25899 xfs: Modify xlog_... |
3348 |
tic = kmem_cache_zalloc(xfs_log_ticket_zone, GFP_NOFS | __GFP_NOFAIL); |
e773fc934 xfs: Refactor xfs... |
3349 3350 |
unit_res = xfs_log_calc_unit_res(log->l_mp, unit_bytes); |
cc09c0dc5 [XFS] Fix double ... |
3351 |
atomic_set(&tic->t_ref, 1); |
14a7235fb xfs: remove log s... |
3352 |
tic->t_task = current; |
105479419 xfs: convert log ... |
3353 |
INIT_LIST_HEAD(&tic->t_queue); |
e773fc934 xfs: Refactor xfs... |
3354 3355 |
tic->t_unit_res = unit_res; tic->t_curr_res = unit_res; |
1da177e4c Linux-2.6.12-rc2 |
3356 3357 |
tic->t_cnt = cnt; tic->t_ocnt = cnt; |
ecb3403de xfs: rename rando... |
3358 |
tic->t_tid = prandom_u32(); |
1da177e4c Linux-2.6.12-rc2 |
3359 |
tic->t_clientid = client; |
9006fb91c xfs: split and cl... |
3360 |
if (permanent) |
1da177e4c Linux-2.6.12-rc2 |
3361 |
tic->t_flags |= XLOG_TIC_PERM_RESERV; |
1da177e4c Linux-2.6.12-rc2 |
3362 |
|
0adba5363 [XFS] replace som... |
3363 |
xlog_tic_reset_res(tic); |
7e9c63961 [XFS] 929956 add ... |
3364 |
|
1da177e4c Linux-2.6.12-rc2 |
3365 |
return tic; |
cc09c0dc5 [XFS] Fix double ... |
3366 |
} |
1da177e4c Linux-2.6.12-rc2 |
3367 |
|
cfcbbbd08 [XFS] Remove old,... |
3368 |
#if defined(DEBUG) |
1da177e4c Linux-2.6.12-rc2 |
3369 3370 3371 3372 3373 |
/* * Make sure that the destination ptr is within the valid data region of * one of the iclogs. This uses backup pointers stored in a different * part of the log in case we trash the log structure. */ |
181fdfe66 xfs: mark xlog_ve... |
3374 |
STATIC void |
e6b1f2737 xfs: clean up xlo... |
3375 |
xlog_verify_dest_ptr( |
ad223e603 xfs: rename log s... |
3376 |
struct xlog *log, |
5809d5e08 xfs: use void poi... |
3377 |
void *ptr) |
1da177e4c Linux-2.6.12-rc2 |
3378 3379 3380 |
{ int i; int good_ptr = 0; |
e6b1f2737 xfs: clean up xlo... |
3381 3382 3383 |
for (i = 0; i < log->l_iclog_bufs; i++) { if (ptr >= log->l_iclog_bak[i] && ptr <= log->l_iclog_bak[i] + log->l_iclog_size) |
1da177e4c Linux-2.6.12-rc2 |
3384 3385 |
good_ptr++; } |
e6b1f2737 xfs: clean up xlo... |
3386 3387 |
if (!good_ptr) |
a0fa2b679 xfs: Convert xlog... |
3388 |
xfs_emerg(log->l_mp, "%s: invalid ptr", __func__); |
e6b1f2737 xfs: clean up xlo... |
3389 |
} |
1da177e4c Linux-2.6.12-rc2 |
3390 |
|
da8a1a4a4 xfs: convert log ... |
3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 |
/* * Check to make sure the grant write head didn't just over lap the tail. If * the cycles are the same, we can't be overlapping. Otherwise, make sure that * the cycles differ by exactly one and check the byte count. * * This check is run unlocked, so can give false positives. Rather than assert * on failures, use a warn-once flag and a panic tag to allow the admin to * determine if they want to panic the machine when such an error occurs. For * debug kernels this will have the same effect as using an assert but, unlinke * an assert, it can be turned off at runtime. */ |
1da177e4c Linux-2.6.12-rc2 |
3402 |
STATIC void |
3f336c6fa xfs: fact out com... |
3403 |
xlog_verify_grant_tail( |
ad223e603 xfs: rename log s... |
3404 |
struct xlog *log) |
3f336c6fa xfs: fact out com... |
3405 |
{ |
1c3cb9ec0 xfs: convert l_ta... |
3406 |
int tail_cycle, tail_blocks; |
a69ed03c2 xfs: combine gran... |
3407 |
int cycle, space; |
3f336c6fa xfs: fact out com... |
3408 |
|
28496968a xfs: add the xlog... |
3409 |
xlog_crack_grant_head(&log->l_write_head.grant, &cycle, &space); |
1c3cb9ec0 xfs: convert l_ta... |
3410 3411 |
xlog_crack_atomic_lsn(&log->l_tail_lsn, &tail_cycle, &tail_blocks); if (tail_cycle != cycle) { |
da8a1a4a4 xfs: convert log ... |
3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 |
if (cycle - 1 != tail_cycle && !(log->l_flags & XLOG_TAIL_WARN)) { xfs_alert_tag(log->l_mp, XFS_PTAG_LOGRES, "%s: cycle - 1 != tail_cycle", __func__); log->l_flags |= XLOG_TAIL_WARN; } if (space > BBTOB(tail_blocks) && !(log->l_flags & XLOG_TAIL_WARN)) { xfs_alert_tag(log->l_mp, XFS_PTAG_LOGRES, "%s: space > BBTOB(tail_blocks)", __func__); log->l_flags |= XLOG_TAIL_WARN; } |
3f336c6fa xfs: fact out com... |
3425 3426 |
} } |
1da177e4c Linux-2.6.12-rc2 |
3427 3428 |
/* check if it will fit */ STATIC void |
9a8d2fdbb xfs: remove xlog_... |
3429 3430 3431 3432 |
xlog_verify_tail_lsn( struct xlog *log, struct xlog_in_core *iclog, xfs_lsn_t tail_lsn) |
1da177e4c Linux-2.6.12-rc2 |
3433 3434 3435 3436 3437 3438 3439 |
{ int blocks; if (CYCLE_LSN(tail_lsn) == log->l_prev_cycle) { blocks = log->l_logBBsize - (log->l_prev_block - BLOCK_LSN(tail_lsn)); if (blocks < BTOBB(iclog->ic_offset)+BTOBB(log->l_iclog_hsize)) |
a0fa2b679 xfs: Convert xlog... |
3440 |
xfs_emerg(log->l_mp, "%s: ran out of log space", __func__); |
1da177e4c Linux-2.6.12-rc2 |
3441 3442 3443 3444 |
} else { ASSERT(CYCLE_LSN(tail_lsn)+1 == log->l_prev_cycle); if (BLOCK_LSN(tail_lsn) == log->l_prev_block) |
a0fa2b679 xfs: Convert xlog... |
3445 |
xfs_emerg(log->l_mp, "%s: tail wrapped", __func__); |
1da177e4c Linux-2.6.12-rc2 |
3446 3447 3448 |
blocks = BLOCK_LSN(tail_lsn) - log->l_prev_block; if (blocks < BTOBB(iclog->ic_offset) + 1) |
a0fa2b679 xfs: Convert xlog... |
3449 |
xfs_emerg(log->l_mp, "%s: ran out of log space", __func__); |
1da177e4c Linux-2.6.12-rc2 |
3450 |
} |
b843299ba xfs: remove some ... |
3451 |
} |
1da177e4c Linux-2.6.12-rc2 |
3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 |
/* * Perform a number of checks on the iclog before writing to disk. * * 1. Make sure the iclogs are still circular * 2. Make sure we have a good magic number * 3. Make sure we don't have magic numbers in the data * 4. Check fields of each log operation header for: * A. Valid client identifier * B. tid ptr value falls in valid ptr space (user space code) * C. Length in log record header is correct according to the * individual operation headers within record. * 5. When a bwrite will occur within 5 blocks of the front of the physical * log, check the preceding blocks of the physical log to make sure all * the cycle numbers agree with the current cycle number. */ STATIC void |
9a8d2fdbb xfs: remove xlog_... |
3469 3470 3471 |
xlog_verify_iclog( struct xlog *log, struct xlog_in_core *iclog, |
abca1f33f xfs: remove the s... |
3472 |
int count) |
1da177e4c Linux-2.6.12-rc2 |
3473 3474 3475 3476 |
{ xlog_op_header_t *ophead; xlog_in_core_t *icptr; xlog_in_core_2_t *xhdr; |
5809d5e08 xfs: use void poi... |
3477 |
void *base_ptr, *ptr, *p; |
db9d67d6b xfs: remove __psi... |
3478 |
ptrdiff_t field_offset; |
c8ce540db xfs: remove doubl... |
3479 |
uint8_t clientid; |
1da177e4c Linux-2.6.12-rc2 |
3480 3481 |
int len, i, j, k, op_len; int idx; |
1da177e4c Linux-2.6.12-rc2 |
3482 3483 |
/* check validity of iclog pointers */ |
b22cd72c9 [XFS] Unwrap LOG_... |
3484 |
spin_lock(&log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
3485 |
icptr = log->l_iclog; |
643f7c4e5 xfs: fix possible... |
3486 3487 |
for (i = 0; i < log->l_iclog_bufs; i++, icptr = icptr->ic_next) ASSERT(icptr); |
1da177e4c Linux-2.6.12-rc2 |
3488 |
if (icptr != log->l_iclog) |
a0fa2b679 xfs: Convert xlog... |
3489 |
xfs_emerg(log->l_mp, "%s: corrupt iclog ring", __func__); |
b22cd72c9 [XFS] Unwrap LOG_... |
3490 |
spin_unlock(&log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
3491 3492 |
/* check log magic numbers */ |
69ef921b5 xfs: byteswap con... |
3493 |
if (iclog->ic_header.h_magicno != cpu_to_be32(XLOG_HEADER_MAGIC_NUM)) |
a0fa2b679 xfs: Convert xlog... |
3494 |
xfs_emerg(log->l_mp, "%s: invalid magic num", __func__); |
1da177e4c Linux-2.6.12-rc2 |
3495 |
|
5809d5e08 xfs: use void poi... |
3496 3497 3498 |
base_ptr = ptr = &iclog->ic_header; p = &iclog->ic_header; for (ptr += BBSIZE; ptr < base_ptr + count; ptr += BBSIZE) { |
69ef921b5 xfs: byteswap con... |
3499 |
if (*(__be32 *)ptr == cpu_to_be32(XLOG_HEADER_MAGIC_NUM)) |
a0fa2b679 xfs: Convert xlog... |
3500 3501 |
xfs_emerg(log->l_mp, "%s: unexpected magic num", __func__); |
1da177e4c Linux-2.6.12-rc2 |
3502 3503 3504 |
} /* check fields */ |
b53e675dc [XFS] xlog_rec_he... |
3505 |
len = be32_to_cpu(iclog->ic_header.h_num_logops); |
5809d5e08 xfs: use void poi... |
3506 3507 |
base_ptr = ptr = iclog->ic_datap; ophead = ptr; |
b28708d6a [XFS] sanitize xl... |
3508 |
xhdr = iclog->ic_data; |
1da177e4c Linux-2.6.12-rc2 |
3509 |
for (i = 0; i < len; i++) { |
5809d5e08 xfs: use void poi... |
3510 |
ophead = ptr; |
1da177e4c Linux-2.6.12-rc2 |
3511 3512 |
/* clientid is only 1 byte */ |
5809d5e08 xfs: use void poi... |
3513 3514 |
p = &ophead->oh_clientid; field_offset = p - base_ptr; |
abca1f33f xfs: remove the s... |
3515 |
if (field_offset & 0x1ff) { |
1da177e4c Linux-2.6.12-rc2 |
3516 3517 |
clientid = ophead->oh_clientid; } else { |
b2a922cd6 xfs: remove xfs_c... |
3518 |
idx = BTOBBT((char *)&ophead->oh_clientid - iclog->ic_datap); |
1da177e4c Linux-2.6.12-rc2 |
3519 3520 3521 |
if (idx >= (XLOG_HEADER_CYCLE_SIZE / BBSIZE)) { j = idx / (XLOG_HEADER_CYCLE_SIZE / BBSIZE); k = idx % (XLOG_HEADER_CYCLE_SIZE / BBSIZE); |
03bea6fe6 [XFS] clean up so... |
3522 3523 |
clientid = xlog_get_client_id( xhdr[j].hic_xheader.xh_cycle_data[k]); |
1da177e4c Linux-2.6.12-rc2 |
3524 |
} else { |
03bea6fe6 [XFS] clean up so... |
3525 3526 |
clientid = xlog_get_client_id( iclog->ic_header.h_cycle_data[idx]); |
1da177e4c Linux-2.6.12-rc2 |
3527 3528 3529 |
} } if (clientid != XFS_TRANSACTION && clientid != XFS_LOG) |
a0fa2b679 xfs: Convert xlog... |
3530 |
xfs_warn(log->l_mp, |
c96900435 xfs: use %px for ... |
3531 |
"%s: invalid clientid %d op "PTR_FMT" offset 0x%lx", |
a0fa2b679 xfs: Convert xlog... |
3532 3533 |
__func__, clientid, ophead, (unsigned long)field_offset); |
1da177e4c Linux-2.6.12-rc2 |
3534 3535 |
/* check length */ |
5809d5e08 xfs: use void poi... |
3536 3537 |
p = &ophead->oh_len; field_offset = p - base_ptr; |
abca1f33f xfs: remove the s... |
3538 |
if (field_offset & 0x1ff) { |
67fcb7bfb [XFS] clean up so... |
3539 |
op_len = be32_to_cpu(ophead->oh_len); |
1da177e4c Linux-2.6.12-rc2 |
3540 |
} else { |
db9d67d6b xfs: remove __psi... |
3541 3542 |
idx = BTOBBT((uintptr_t)&ophead->oh_len - (uintptr_t)iclog->ic_datap); |
1da177e4c Linux-2.6.12-rc2 |
3543 3544 3545 |
if (idx >= (XLOG_HEADER_CYCLE_SIZE / BBSIZE)) { j = idx / (XLOG_HEADER_CYCLE_SIZE / BBSIZE); k = idx % (XLOG_HEADER_CYCLE_SIZE / BBSIZE); |
b53e675dc [XFS] xlog_rec_he... |
3546 |
op_len = be32_to_cpu(xhdr[j].hic_xheader.xh_cycle_data[k]); |
1da177e4c Linux-2.6.12-rc2 |
3547 |
} else { |
b53e675dc [XFS] xlog_rec_he... |
3548 |
op_len = be32_to_cpu(iclog->ic_header.h_cycle_data[idx]); |
1da177e4c Linux-2.6.12-rc2 |
3549 3550 3551 3552 |
} } ptr += sizeof(xlog_op_header_t) + op_len; } |
b843299ba xfs: remove some ... |
3553 |
} |
cfcbbbd08 [XFS] Remove old,... |
3554 |
#endif |
1da177e4c Linux-2.6.12-rc2 |
3555 3556 |
/* |
b22cd72c9 [XFS] Unwrap LOG_... |
3557 |
* Mark all iclogs IOERROR. l_icloglock is held by the caller. |
1da177e4c Linux-2.6.12-rc2 |
3558 3559 3560 |
*/ STATIC int xlog_state_ioerror( |
9a8d2fdbb xfs: remove xlog_... |
3561 |
struct xlog *log) |
1da177e4c Linux-2.6.12-rc2 |
3562 3563 3564 3565 |
{ xlog_in_core_t *iclog, *ic; iclog = log->l_iclog; |
1858bb0be xfs: turn ic_stat... |
3566 |
if (iclog->ic_state != XLOG_STATE_IOERROR) { |
1da177e4c Linux-2.6.12-rc2 |
3567 3568 3569 3570 3571 3572 3573 3574 3575 |
/* * Mark all the incore logs IOERROR. * From now on, no log flushes will result. */ ic = iclog; do { ic->ic_state = XLOG_STATE_IOERROR; ic = ic->ic_next; } while (ic != iclog); |
014c2544e return statement ... |
3576 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
3577 3578 3579 3580 |
} /* * Return non-zero, if state transition has already happened. */ |
014c2544e return statement ... |
3581 |
return 1; |
1da177e4c Linux-2.6.12-rc2 |
3582 3583 3584 3585 3586 3587 |
} /* * This is called from xfs_force_shutdown, when we're forcibly * shutting down the filesystem, typically because of an IO error. * Our main objectives here are to make sure that: |
a870fe6df xfs: force the lo... |
3588 3589 3590 |
* a. if !logerror, flush the logs to disk. Anything modified * after this is ignored. * b. the filesystem gets marked 'SHUTDOWN' for all interested |
1da177e4c Linux-2.6.12-rc2 |
3591 |
* parties to find out, 'atomically'. |
a870fe6df xfs: force the lo... |
3592 |
* c. those who're sleeping on log reservations, pinned objects and |
1da177e4c Linux-2.6.12-rc2 |
3593 |
* other resources get woken up, and be told the bad news. |
a870fe6df xfs: force the lo... |
3594 |
* d. nothing new gets queued up after (b) and (c) are done. |
9da1ab181 xfs: forced unmou... |
3595 |
* |
a870fe6df xfs: force the lo... |
3596 3597 3598 |
* Note: for the !logerror case we need to flush the regions held in memory out * to disk first. This needs to be done before the log is marked as shutdown, * otherwise the iclog writes will fail. |
1da177e4c Linux-2.6.12-rc2 |
3599 3600 3601 3602 3603 3604 |
*/ int xfs_log_force_umount( struct xfs_mount *mp, int logerror) { |
9a8d2fdbb xfs: remove xlog_... |
3605 |
struct xlog *log; |
1da177e4c Linux-2.6.12-rc2 |
3606 |
int retval; |
1da177e4c Linux-2.6.12-rc2 |
3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 |
log = mp->m_log; /* * If this happens during log recovery, don't worry about * locking; the log isn't open for business yet. */ if (!log || log->l_flags & XLOG_ACTIVE_RECOVERY) { mp->m_flags |= XFS_MOUNT_FS_SHUTDOWN; |
bac8dca9f [XFS] fix NULL po... |
3617 |
if (mp->m_sb_bp) |
b0388bf10 xfs: remove XBF_D... |
3618 |
mp->m_sb_bp->b_flags |= XBF_DONE; |
014c2544e return statement ... |
3619 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
3620 3621 3622 3623 3624 3625 |
} /* * Somebody could've already done the hard work for us. * No need to get locks for this. */ |
1858bb0be xfs: turn ic_stat... |
3626 |
if (logerror && log->l_iclog->ic_state == XLOG_STATE_IOERROR) { |
1da177e4c Linux-2.6.12-rc2 |
3627 |
ASSERT(XLOG_FORCED_SHUTDOWN(log)); |
014c2544e return statement ... |
3628 |
return 1; |
1da177e4c Linux-2.6.12-rc2 |
3629 |
} |
9da1ab181 xfs: forced unmou... |
3630 3631 |
/* |
a870fe6df xfs: force the lo... |
3632 3633 3634 3635 3636 |
* Flush all the completed transactions to disk before marking the log * being shut down. We need to do it in this order to ensure that * completed operations are safely on disk before we shut down, and that * we don't have to issue any buffer IO after the shutdown flags are set * to guarantee this. |
9da1ab181 xfs: forced unmou... |
3637 |
*/ |
93b8a5854 xfs: remove the d... |
3638 |
if (!logerror) |
60e5bb784 xfs: merge _xfs_l... |
3639 |
xfs_log_force(mp, XFS_LOG_SYNC); |
9da1ab181 xfs: forced unmou... |
3640 |
|
1da177e4c Linux-2.6.12-rc2 |
3641 |
/* |
3f16b9850 xfs: introduce ne... |
3642 3643 |
* mark the filesystem and the as in a shutdown state and wake * everybody up to tell them the bad news. |
1da177e4c Linux-2.6.12-rc2 |
3644 |
*/ |
b22cd72c9 [XFS] Unwrap LOG_... |
3645 |
spin_lock(&log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
3646 |
mp->m_flags |= XFS_MOUNT_FS_SHUTDOWN; |
bac8dca9f [XFS] fix NULL po... |
3647 |
if (mp->m_sb_bp) |
b0388bf10 xfs: remove XBF_D... |
3648 |
mp->m_sb_bp->b_flags |= XBF_DONE; |
bac8dca9f [XFS] fix NULL po... |
3649 |
|
1da177e4c Linux-2.6.12-rc2 |
3650 |
/* |
a870fe6df xfs: force the lo... |
3651 3652 |
* Mark the log and the iclogs with IO error flags to prevent any * further log IO from being issued or completed. |
1da177e4c Linux-2.6.12-rc2 |
3653 3654 |
*/ log->l_flags |= XLOG_IO_ERROR; |
a870fe6df xfs: force the lo... |
3655 |
retval = xlog_state_ioerror(log); |
b22cd72c9 [XFS] Unwrap LOG_... |
3656 |
spin_unlock(&log->l_icloglock); |
1da177e4c Linux-2.6.12-rc2 |
3657 3658 |
/* |
105479419 xfs: convert log ... |
3659 3660 3661 3662 |
* We don't want anybody waiting for log reservations after this. That * means we have to wake up everybody queued up on reserveq as well as * writeq. In addition, we make sure in xlog_{re}grant_log_space that * we don't enqueue anything once the SHUTDOWN flag is set, and this |
3f16b9850 xfs: introduce ne... |
3663 |
* action is protected by the grant locks. |
1da177e4c Linux-2.6.12-rc2 |
3664 |
*/ |
a79bf2d75 xfs: add xlog_gra... |
3665 3666 |
xlog_grant_head_wake_all(&log->l_reserve_head); xlog_grant_head_wake_all(&log->l_write_head); |
1da177e4c Linux-2.6.12-rc2 |
3667 |
|
1da177e4c Linux-2.6.12-rc2 |
3668 |
/* |
ac983517e xfs: don't sleep ... |
3669 3670 3671 3672 |
* Wake up everybody waiting on xfs_log_force. Wake the CIL push first * as if the log writes were completed. The abort handling in the log * item committed callback functions will do this again under lock to * avoid races. |
1da177e4c Linux-2.6.12-rc2 |
3673 |
*/ |
cdea5459c xfs: fix missed w... |
3674 |
spin_lock(&log->l_cilp->xc_push_lock); |
ac983517e xfs: don't sleep ... |
3675 |
wake_up_all(&log->l_cilp->xc_commit_wait); |
cdea5459c xfs: fix missed w... |
3676 |
spin_unlock(&log->l_cilp->xc_push_lock); |
12e6a0f44 xfs: remove the a... |
3677 |
xlog_state_do_callback(log); |
1da177e4c Linux-2.6.12-rc2 |
3678 |
|
1da177e4c Linux-2.6.12-rc2 |
3679 |
/* return non-zero if log IOERROR transition had already happened */ |
014c2544e return statement ... |
3680 |
return retval; |
1da177e4c Linux-2.6.12-rc2 |
3681 |
} |
ba0f32d46 [XFS] mark variou... |
3682 |
STATIC int |
9a8d2fdbb xfs: remove xlog_... |
3683 3684 |
xlog_iclogs_empty( struct xlog *log) |
1da177e4c Linux-2.6.12-rc2 |
3685 3686 3687 3688 3689 3690 3691 3692 3693 |
{ xlog_in_core_t *iclog; iclog = log->l_iclog; do { /* endianness does not matter here, zero is zero in * any language. */ if (iclog->ic_header.h_num_logops) |
014c2544e return statement ... |
3694 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
3695 3696 |
iclog = iclog->ic_next; } while (iclog != log->l_iclog); |
014c2544e return statement ... |
3697 |
return 1; |
1da177e4c Linux-2.6.12-rc2 |
3698 |
} |
f661f1e0b xfs: sync work is... |
3699 |
|
a45086e27 xfs: validate met... |
3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 |
/* * Verify that an LSN stamped into a piece of metadata is valid. This is * intended for use in read verifiers on v5 superblocks. */ bool xfs_log_check_lsn( struct xfs_mount *mp, xfs_lsn_t lsn) { struct xlog *log = mp->m_log; bool valid; /* * norecovery mode skips mount-time log processing and unconditionally * resets the in-core LSN. We can't validate in this mode, but * modifications are not allowed anyways so just return true. */ if (mp->m_flags & XFS_MOUNT_NORECOVERY) return true; /* * Some metadata LSNs are initialized to NULL (e.g., the agfl). This is * handled by recovery and thus safe to ignore here. */ if (lsn == NULLCOMMITLSN) return true; valid = xlog_valid_lsn(mp->m_log, lsn); /* warn the user about what's gone wrong before verifier failure */ if (!valid) { spin_lock(&log->l_icloglock); xfs_warn(mp, "Corruption warning: Metadata has LSN (%d:%d) ahead of current LSN (%d:%d). " "Please unmount and run xfs_repair (>= v4.3) to resolve.", CYCLE_LSN(lsn), BLOCK_LSN(lsn), log->l_curr_cycle, log->l_curr_block); spin_unlock(&log->l_icloglock); } return valid; } |
0c60d3aa0 xfs: refactor log... |
3742 3743 3744 3745 3746 3747 3748 3749 3750 |
bool xfs_log_in_recovery( struct xfs_mount *mp) { struct xlog *log = mp->m_log; return log->l_flags & XLOG_ACTIVE_RECOVERY; } |