Blame view
fs/jbd/checkpoint.c
21.3 KB
1da177e4c Linux-2.6.12-rc2 |
1 |
/* |
588626996 fix file specific... |
2 |
* linux/fs/jbd/checkpoint.c |
ae6ddcc5f [PATCH] ext3 and ... |
3 |
* |
1da177e4c Linux-2.6.12-rc2 |
4 5 6 7 8 9 10 11 |
* Written by Stephen C. Tweedie <sct@redhat.com>, 1999 * * Copyright 1999 Red Hat Software --- All Rights Reserved * * This file is part of the Linux kernel and is made available under * the terms of the GNU General Public License, version 2, or at your * option, any later version, incorporated herein by reference. * |
ae6ddcc5f [PATCH] ext3 and ... |
12 13 |
* Checkpoint routines for the generic filesystem journaling code. * Part of the ext2fs journaling system. |
1da177e4c Linux-2.6.12-rc2 |
14 15 16 17 18 19 20 21 22 23 24 |
* * Checkpointing is the process of ensuring that a section of the log is * committed fully to disk, so that that portion of the log can be * reused. */ #include <linux/time.h> #include <linux/fs.h> #include <linux/jbd.h> #include <linux/errno.h> #include <linux/slab.h> |
a212d1a71 jbd: Use WRITE_SY... |
25 |
#include <linux/blkdev.h> |
99cb1a318 jbd: Add fixed tr... |
26 |
#include <trace/events/jbd.h> |
1da177e4c Linux-2.6.12-rc2 |
27 28 |
/* |
78ce89c92 [PATCH] JBD: spli... |
29 |
* Unlink a buffer from a transaction checkpoint list. |
1da177e4c Linux-2.6.12-rc2 |
30 31 32 |
* * Called with j_list_lock held. */ |
78ce89c92 [PATCH] JBD: spli... |
33 |
static inline void __buffer_unlink_first(struct journal_head *jh) |
1da177e4c Linux-2.6.12-rc2 |
34 |
{ |
78ce89c92 [PATCH] JBD: spli... |
35 |
transaction_t *transaction = jh->b_cp_transaction; |
1da177e4c Linux-2.6.12-rc2 |
36 37 38 |
jh->b_cpnext->b_cpprev = jh->b_cpprev; jh->b_cpprev->b_cpnext = jh->b_cpnext; |
78ce89c92 [PATCH] JBD: spli... |
39 |
if (transaction->t_checkpoint_list == jh) { |
1da177e4c Linux-2.6.12-rc2 |
40 |
transaction->t_checkpoint_list = jh->b_cpnext; |
78ce89c92 [PATCH] JBD: spli... |
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 |
if (transaction->t_checkpoint_list == jh) transaction->t_checkpoint_list = NULL; } } /* * Unlink a buffer from a transaction checkpoint(io) list. * * Called with j_list_lock held. */ static inline void __buffer_unlink(struct journal_head *jh) { transaction_t *transaction = jh->b_cp_transaction; __buffer_unlink_first(jh); if (transaction->t_checkpoint_io_list == jh) { transaction->t_checkpoint_io_list = jh->b_cpnext; if (transaction->t_checkpoint_io_list == jh) transaction->t_checkpoint_io_list = NULL; } } /* * Move a buffer from the checkpoint list to the checkpoint io list * * Called with j_list_lock held */ static inline void __buffer_relink_io(struct journal_head *jh) { transaction_t *transaction = jh->b_cp_transaction; __buffer_unlink_first(jh); if (!transaction->t_checkpoint_io_list) { jh->b_cpnext = jh->b_cpprev = jh; } else { jh->b_cpnext = transaction->t_checkpoint_io_list; jh->b_cpprev = transaction->t_checkpoint_io_list->b_cpprev; jh->b_cpprev->b_cpnext = jh; jh->b_cpnext->b_cpprev = jh; } transaction->t_checkpoint_io_list = jh; |
1da177e4c Linux-2.6.12-rc2 |
83 84 85 86 |
} /* * Try to release a checkpointed buffer from its transaction. |
78ce89c92 [PATCH] JBD: spli... |
87 88 89 |
* Returns 1 if we released it and 2 if we also released the * whole transaction. * |
1da177e4c Linux-2.6.12-rc2 |
90 91 92 93 94 95 96 |
* Requires j_list_lock * Called under jbd_lock_bh_state(jh2bh(jh)), and drops it */ static int __try_to_free_cp_buf(struct journal_head *jh) { int ret = 0; struct buffer_head *bh = jh2bh(jh); |
4afe97853 jbd: fix error ha... |
97 |
if (jh->b_jlist == BJ_None && !buffer_locked(bh) && |
9f818b4ac jbd: test BH_Writ... |
98 |
!buffer_dirty(bh) && !buffer_write_io_error(bh)) { |
bb189247f jbd: Fix oops in ... |
99 100 101 102 103 |
/* * Get our reference so that bh cannot be freed before * we unlock it */ get_bh(bh); |
1da177e4c Linux-2.6.12-rc2 |
104 |
JBUFFER_TRACE(jh, "remove from checkpoint list"); |
78ce89c92 [PATCH] JBD: spli... |
105 |
ret = __journal_remove_checkpoint(jh) + 1; |
1da177e4c Linux-2.6.12-rc2 |
106 |
jbd_unlock_bh_state(bh); |
1da177e4c Linux-2.6.12-rc2 |
107 108 |
BUFFER_TRACE(bh, "release"); __brelse(bh); |
1da177e4c Linux-2.6.12-rc2 |
109 110 111 112 113 114 115 116 117 118 119 120 121 122 |
} else { jbd_unlock_bh_state(bh); } return ret; } /* * __log_wait_for_space: wait until there is space in the journal. * * Called under j-state_lock *only*. It will be unlocked if we have to wait * for a checkpoint to free up some space in the log. */ void __log_wait_for_space(journal_t *journal) { |
e219cca08 jbd: don't give u... |
123 |
int nblocks, space_left; |
1da177e4c Linux-2.6.12-rc2 |
124 125 126 127 128 129 130 |
assert_spin_locked(&journal->j_state_lock); nblocks = jbd_space_needed(journal); while (__log_space_left(journal) < nblocks) { if (journal->j_flags & JFS_ABORT) return; spin_unlock(&journal->j_state_lock); |
2c68ee754 [PATCH] sem2mutex... |
131 |
mutex_lock(&journal->j_checkpoint_mutex); |
1da177e4c Linux-2.6.12-rc2 |
132 133 134 |
/* * Test again, another process may have checkpointed while we |
be07c4ed4 jbd: abort instea... |
135 |
* were waiting for the checkpoint lock. If there are no |
e219cca08 jbd: don't give u... |
136 137 138 139 140 141 142 |
* transactions ready to be checkpointed, try to recover * journal space by calling cleanup_journal_tail(), and if * that doesn't work, by waiting for the currently committing * transaction to complete. If there is absolutely no way * to make progress, this is either a BUG or corrupted * filesystem, so abort the journal and leave a stack * trace for forensic evidence. |
1da177e4c Linux-2.6.12-rc2 |
143 144 |
*/ spin_lock(&journal->j_state_lock); |
be07c4ed4 jbd: abort instea... |
145 |
spin_lock(&journal->j_list_lock); |
1da177e4c Linux-2.6.12-rc2 |
146 |
nblocks = jbd_space_needed(journal); |
e219cca08 jbd: don't give u... |
147 148 |
space_left = __log_space_left(journal); if (space_left < nblocks) { |
be07c4ed4 jbd: abort instea... |
149 |
int chkpt = journal->j_checkpoint_transactions != NULL; |
e219cca08 jbd: don't give u... |
150 |
tid_t tid = 0; |
be07c4ed4 jbd: abort instea... |
151 |
|
e219cca08 jbd: don't give u... |
152 153 |
if (journal->j_committing_transaction) tid = journal->j_committing_transaction->t_tid; |
be07c4ed4 jbd: abort instea... |
154 |
spin_unlock(&journal->j_list_lock); |
1da177e4c Linux-2.6.12-rc2 |
155 |
spin_unlock(&journal->j_state_lock); |
be07c4ed4 jbd: abort instea... |
156 157 |
if (chkpt) { log_do_checkpoint(journal); |
e219cca08 jbd: don't give u... |
158 159 160 161 162 |
} else if (cleanup_journal_tail(journal) == 0) { /* We were able to recover space; yay! */ ; } else if (tid) { log_wait_commit(journal, tid); |
be07c4ed4 jbd: abort instea... |
163 |
} else { |
e219cca08 jbd: don't give u... |
164 165 166 167 168 169 170 171 |
printk(KERN_ERR "%s: needed %d blocks and " "only had %d space available ", __func__, nblocks, space_left); printk(KERN_ERR "%s: no way to get more " "journal space ", __func__); WARN_ON(1); |
be07c4ed4 jbd: abort instea... |
172 173 |
journal_abort(journal, 0); } |
1da177e4c Linux-2.6.12-rc2 |
174 |
spin_lock(&journal->j_state_lock); |
be07c4ed4 jbd: abort instea... |
175 176 |
} else { spin_unlock(&journal->j_list_lock); |
1da177e4c Linux-2.6.12-rc2 |
177 |
} |
2c68ee754 [PATCH] sem2mutex... |
178 |
mutex_unlock(&journal->j_checkpoint_mutex); |
1da177e4c Linux-2.6.12-rc2 |
179 180 181 182 183 184 185 186 187 |
} } /* * We were unable to perform jbd_trylock_bh_state() inside j_list_lock. * The caller must restart a list walk. Wait for someone else to run * jbd_unlock_bh_state(). */ static void jbd_sync_bh(journal_t *journal, struct buffer_head *bh) |
e7ab8d650 [PATCH] jbd: add ... |
188 |
__releases(journal->j_list_lock) |
1da177e4c Linux-2.6.12-rc2 |
189 190 191 192 193 194 195 196 197 |
{ get_bh(bh); spin_unlock(&journal->j_list_lock); jbd_lock_bh_state(bh); jbd_unlock_bh_state(bh); put_bh(bh); } /* |
78ce89c92 [PATCH] JBD: spli... |
198 199 200 201 |
* Clean up transaction's list of buffers submitted for io. * We wait for any pending IO to complete and remove any clean * buffers. Note that we take the buffers in the opposite ordering * from the one in which they were submitted for IO. |
1da177e4c Linux-2.6.12-rc2 |
202 |
* |
4afe97853 jbd: fix error ha... |
203 204 205 |
* Return 0 on success, and return <0 if some buffers have failed * to be written out. * |
1da177e4c Linux-2.6.12-rc2 |
206 207 |
* Called with j_list_lock held. */ |
4afe97853 jbd: fix error ha... |
208 |
static int __wait_cp_io(journal_t *journal, transaction_t *transaction) |
1da177e4c Linux-2.6.12-rc2 |
209 |
{ |
78ce89c92 [PATCH] JBD: spli... |
210 |
struct journal_head *jh; |
1da177e4c Linux-2.6.12-rc2 |
211 |
struct buffer_head *bh; |
78ce89c92 [PATCH] JBD: spli... |
212 213 |
tid_t this_tid; int released = 0; |
4afe97853 jbd: fix error ha... |
214 |
int ret = 0; |
78ce89c92 [PATCH] JBD: spli... |
215 216 217 218 219 220 |
this_tid = transaction->t_tid; restart: /* Did somebody clean up the transaction in the meanwhile? */ if (journal->j_checkpoint_transactions != transaction || transaction->t_tid != this_tid) |
4afe97853 jbd: fix error ha... |
221 |
return ret; |
78ce89c92 [PATCH] JBD: spli... |
222 223 |
while (!released && transaction->t_checkpoint_io_list) { jh = transaction->t_checkpoint_io_list; |
1da177e4c Linux-2.6.12-rc2 |
224 |
bh = jh2bh(jh); |
78ce89c92 [PATCH] JBD: spli... |
225 226 227 228 229 |
if (!jbd_trylock_bh_state(bh)) { jbd_sync_bh(journal, bh); spin_lock(&journal->j_list_lock); goto restart; } |
bb189247f jbd: Fix oops in ... |
230 |
get_bh(bh); |
1da177e4c Linux-2.6.12-rc2 |
231 |
if (buffer_locked(bh)) { |
1da177e4c Linux-2.6.12-rc2 |
232 |
spin_unlock(&journal->j_list_lock); |
78ce89c92 [PATCH] JBD: spli... |
233 |
jbd_unlock_bh_state(bh); |
1da177e4c Linux-2.6.12-rc2 |
234 235 236 237 |
wait_on_buffer(bh); /* the journal_head may have gone by now */ BUFFER_TRACE(bh, "brelse"); __brelse(bh); |
78ce89c92 [PATCH] JBD: spli... |
238 239 |
spin_lock(&journal->j_list_lock); goto restart; |
1da177e4c Linux-2.6.12-rc2 |
240 |
} |
9f818b4ac jbd: test BH_Writ... |
241 |
if (unlikely(buffer_write_io_error(bh))) |
4afe97853 jbd: fix error ha... |
242 |
ret = -EIO; |
1da177e4c Linux-2.6.12-rc2 |
243 |
/* |
78ce89c92 [PATCH] JBD: spli... |
244 245 |
* Now in whatever state the buffer currently is, we know that * it has been written out and so we can drop it from the list |
1da177e4c Linux-2.6.12-rc2 |
246 |
*/ |
78ce89c92 [PATCH] JBD: spli... |
247 248 |
released = __journal_remove_checkpoint(jh); jbd_unlock_bh_state(bh); |
78ce89c92 [PATCH] JBD: spli... |
249 250 |
__brelse(bh); } |
4afe97853 jbd: fix error ha... |
251 252 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
253 254 255 256 257 258 259 260 |
} #define NR_BATCH 64 static void __flush_batch(journal_t *journal, struct buffer_head **bhs, int *batch_count) { int i; |
a212d1a71 jbd: Use WRITE_SY... |
261 |
struct blk_plug plug; |
1da177e4c Linux-2.6.12-rc2 |
262 |
|
a212d1a71 jbd: Use WRITE_SY... |
263 |
blk_start_plug(&plug); |
9cb569d60 remove SWRITE* I/... |
264 |
for (i = 0; i < *batch_count; i++) |
a212d1a71 jbd: Use WRITE_SY... |
265 266 |
write_dirty_buffer(bhs[i], WRITE_SYNC); blk_finish_plug(&plug); |
9cb569d60 remove SWRITE* I/... |
267 |
|
1da177e4c Linux-2.6.12-rc2 |
268 269 270 271 272 273 274 275 276 277 278 279 280 |
for (i = 0; i < *batch_count; i++) { struct buffer_head *bh = bhs[i]; clear_buffer_jwrite(bh); BUFFER_TRACE(bh, "brelse"); __brelse(bh); } *batch_count = 0; } /* * Try to flush one buffer from the checkpoint list to disk. * * Return 1 if something happened which requires us to abort the current |
4afe97853 jbd: fix error ha... |
281 282 |
* scan of the checkpoint list. Return <0 if the buffer has failed to * be written out. |
1da177e4c Linux-2.6.12-rc2 |
283 |
* |
78ce89c92 [PATCH] JBD: spli... |
284 |
* Called with j_list_lock held and drops it if 1 is returned |
1da177e4c Linux-2.6.12-rc2 |
285 286 |
* Called under jbd_lock_bh_state(jh2bh(jh)), and drops it */ |
78ce89c92 [PATCH] JBD: spli... |
287 288 |
static int __process_buffer(journal_t *journal, struct journal_head *jh, struct buffer_head **bhs, int *batch_count) |
1da177e4c Linux-2.6.12-rc2 |
289 290 291 |
{ struct buffer_head *bh = jh2bh(jh); int ret = 0; |
78ce89c92 [PATCH] JBD: spli... |
292 |
if (buffer_locked(bh)) { |
e4d5e3a49 jbd: Convert atom... |
293 |
get_bh(bh); |
78ce89c92 [PATCH] JBD: spli... |
294 295 296 297 298 299 300 301 302 303 |
spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); wait_on_buffer(bh); /* the journal_head may have gone by now */ BUFFER_TRACE(bh, "brelse"); __brelse(bh); ret = 1; } else if (jh->b_transaction != NULL) { transaction_t *t = jh->b_transaction; tid_t tid = t->t_tid; |
1da177e4c Linux-2.6.12-rc2 |
304 |
|
78ce89c92 [PATCH] JBD: spli... |
305 306 307 308 309 310 |
spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); log_start_commit(journal, tid); log_wait_commit(journal, tid); ret = 1; } else if (!buffer_dirty(bh)) { |
4afe97853 jbd: fix error ha... |
311 |
ret = 1; |
9f818b4ac jbd: test BH_Writ... |
312 |
if (unlikely(buffer_write_io_error(bh))) |
4afe97853 jbd: fix error ha... |
313 |
ret = -EIO; |
bb189247f jbd: Fix oops in ... |
314 |
get_bh(bh); |
78ce89c92 [PATCH] JBD: spli... |
315 316 317 318 319 |
J_ASSERT_JH(jh, !buffer_jbddirty(bh)); BUFFER_TRACE(bh, "remove from checkpoint"); __journal_remove_checkpoint(jh); spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); |
78ce89c92 [PATCH] JBD: spli... |
320 |
__brelse(bh); |
78ce89c92 [PATCH] JBD: spli... |
321 |
} else { |
1da177e4c Linux-2.6.12-rc2 |
322 323 324 325 326 |
/* * Important: we are about to write the buffer, and * possibly block, while still holding the journal lock. * We cannot afford to let the transaction logic start * messing around with this buffer before we write it to |
ae6ddcc5f [PATCH] ext3 and ... |
327 |
* disk, as that would break recoverability. |
1da177e4c Linux-2.6.12-rc2 |
328 329 330 331 332 333 |
*/ BUFFER_TRACE(bh, "queue"); get_bh(bh); J_ASSERT_BH(bh, !buffer_jwrite(bh)); set_buffer_jwrite(bh); bhs[*batch_count] = bh; |
78ce89c92 [PATCH] JBD: spli... |
334 |
__buffer_relink_io(jh); |
1da177e4c Linux-2.6.12-rc2 |
335 336 337 |
jbd_unlock_bh_state(bh); (*batch_count)++; if (*batch_count == NR_BATCH) { |
78ce89c92 [PATCH] JBD: spli... |
338 |
spin_unlock(&journal->j_list_lock); |
1da177e4c Linux-2.6.12-rc2 |
339 340 341 |
__flush_batch(journal, bhs, batch_count); ret = 1; } |
1da177e4c Linux-2.6.12-rc2 |
342 343 344 345 346 |
} return ret; } /* |
78ce89c92 [PATCH] JBD: spli... |
347 348 349 |
* Perform an actual checkpoint. We take the first transaction on the * list of transactions to be checkpointed and send all its buffers * to disk. We submit larger chunks of data at once. |
ae6ddcc5f [PATCH] ext3 and ... |
350 |
* |
1da177e4c Linux-2.6.12-rc2 |
351 |
* The journal should be locked before calling this function. |
4afe97853 jbd: fix error ha... |
352 |
* Called with j_checkpoint_mutex held. |
1da177e4c Linux-2.6.12-rc2 |
353 354 355 |
*/ int log_do_checkpoint(journal_t *journal) { |
78ce89c92 [PATCH] JBD: spli... |
356 357 |
transaction_t *transaction; tid_t this_tid; |
1da177e4c Linux-2.6.12-rc2 |
358 |
int result; |
1da177e4c Linux-2.6.12-rc2 |
359 360 361 |
jbd_debug(1, "Start checkpoint "); |
ae6ddcc5f [PATCH] ext3 and ... |
362 |
/* |
1da177e4c Linux-2.6.12-rc2 |
363 364 |
* First thing: if there are any transactions in the log which * don't need checkpointing, just eliminate them from the |
ae6ddcc5f [PATCH] ext3 and ... |
365 |
* journal straight away. |
1da177e4c Linux-2.6.12-rc2 |
366 367 |
*/ result = cleanup_journal_tail(journal); |
99cb1a318 jbd: Add fixed tr... |
368 |
trace_jbd_checkpoint(journal, result); |
1da177e4c Linux-2.6.12-rc2 |
369 370 371 372 373 374 |
jbd_debug(1, "cleanup_journal_tail returned %d ", result); if (result <= 0) return result; /* |
78ce89c92 [PATCH] JBD: spli... |
375 376 |
* OK, we need to start writing disk blocks. Take one transaction * and write it. |
1da177e4c Linux-2.6.12-rc2 |
377 |
*/ |
4afe97853 jbd: fix error ha... |
378 |
result = 0; |
78ce89c92 [PATCH] JBD: spli... |
379 380 381 382 383 384 |
spin_lock(&journal->j_list_lock); if (!journal->j_checkpoint_transactions) goto out; transaction = journal->j_checkpoint_transactions; this_tid = transaction->t_tid; restart: |
1da177e4c Linux-2.6.12-rc2 |
385 |
/* |
78ce89c92 [PATCH] JBD: spli... |
386 387 388 |
* If someone cleaned up this transaction while we slept, we're * done (maybe it's a new transaction, but it fell at the same * address). |
1da177e4c Linux-2.6.12-rc2 |
389 |
*/ |
78ce89c92 [PATCH] JBD: spli... |
390 391 392 393 394 |
if (journal->j_checkpoint_transactions == transaction && transaction->t_tid == this_tid) { int batch_count = 0; struct buffer_head *bhs[NR_BATCH]; struct journal_head *jh; |
4afe97853 jbd: fix error ha... |
395 |
int retry = 0, err; |
78ce89c92 [PATCH] JBD: spli... |
396 397 |
while (!retry && transaction->t_checkpoint_list) { |
1da177e4c Linux-2.6.12-rc2 |
398 |
struct buffer_head *bh; |
78ce89c92 [PATCH] JBD: spli... |
399 |
jh = transaction->t_checkpoint_list; |
1da177e4c Linux-2.6.12-rc2 |
400 401 402 |
bh = jh2bh(jh); if (!jbd_trylock_bh_state(bh)) { jbd_sync_bh(journal, bh); |
1da177e4c Linux-2.6.12-rc2 |
403 404 405 |
retry = 1; break; } |
78ce89c92 [PATCH] JBD: spli... |
406 |
retry = __process_buffer(journal, jh, bhs,&batch_count); |
4afe97853 jbd: fix error ha... |
407 408 |
if (retry < 0 && !result) result = retry; |
95c354fe9 spinlock: lockbre... |
409 410 |
if (!retry && (need_resched() || spin_needbreak(&journal->j_list_lock))) { |
78ce89c92 [PATCH] JBD: spli... |
411 |
spin_unlock(&journal->j_list_lock); |
1da177e4c Linux-2.6.12-rc2 |
412 413 414 |
retry = 1; break; } |
78ce89c92 [PATCH] JBD: spli... |
415 |
} |
1da177e4c Linux-2.6.12-rc2 |
416 |
|
00ea81459 [PATCH] ext3: fix... |
417 |
if (batch_count) { |
78ce89c92 [PATCH] JBD: spli... |
418 419 420 421 |
if (!retry) { spin_unlock(&journal->j_list_lock); retry = 1; } |
1da177e4c Linux-2.6.12-rc2 |
422 |
__flush_batch(journal, bhs, &batch_count); |
00ea81459 [PATCH] ext3: fix... |
423 |
} |
1da177e4c Linux-2.6.12-rc2 |
424 |
|
78ce89c92 [PATCH] JBD: spli... |
425 426 427 428 |
if (retry) { spin_lock(&journal->j_list_lock); goto restart; } |
1da177e4c Linux-2.6.12-rc2 |
429 |
/* |
78ce89c92 [PATCH] JBD: spli... |
430 431 |
* Now we have cleaned up the first transaction's checkpoint * list. Let's clean up the second one |
7c8903f63 [PATCH] jbd: reve... |
432 |
*/ |
4afe97853 jbd: fix error ha... |
433 434 435 |
err = __wait_cp_io(journal, transaction); if (!result) result = err; |
1da177e4c Linux-2.6.12-rc2 |
436 |
} |
78ce89c92 [PATCH] JBD: spli... |
437 |
out: |
1da177e4c Linux-2.6.12-rc2 |
438 |
spin_unlock(&journal->j_list_lock); |
1da177e4c Linux-2.6.12-rc2 |
439 |
if (result < 0) |
4afe97853 jbd: fix error ha... |
440 441 442 443 444 |
journal_abort(journal, result); else result = cleanup_journal_tail(journal); return (result < 0) ? result : 0; |
1da177e4c Linux-2.6.12-rc2 |
445 446 447 448 449 450 451 |
} /* * Check the list of checkpoint transactions for the journal to see if * we have already got rid of any since the last update of the log tail * in the journal superblock. If so, we can instantly roll the * superblock forward to remove those transactions from the log. |
ae6ddcc5f [PATCH] ext3 and ... |
452 |
* |
1da177e4c Linux-2.6.12-rc2 |
453 |
* Return <0 on error, 0 on success, 1 if there was nothing to clean up. |
ae6ddcc5f [PATCH] ext3 and ... |
454 |
* |
1da177e4c Linux-2.6.12-rc2 |
455 456 457 458 459 |
* Called with the journal lock held. * * This is the only part of the journaling code which really needs to be * aware of transaction aborts. Checkpointing involves writing to the * main filesystem area rather than to the journal, so it can proceed |
4afe97853 jbd: fix error ha... |
460 461 462 |
* even in abort state, but we must not update the super block if * checkpointing may have failed. Otherwise, we would lose some metadata * buffers which should be written-back to the filesystem. |
1da177e4c Linux-2.6.12-rc2 |
463 464 465 466 467 468 |
*/ int cleanup_journal_tail(journal_t *journal) { transaction_t * transaction; tid_t first_tid; |
9c28cbcce jbd: Journal bloc... |
469 |
unsigned int blocknr, freed; |
1da177e4c Linux-2.6.12-rc2 |
470 |
|
4afe97853 jbd: fix error ha... |
471 472 |
if (is_journal_aborted(journal)) return 1; |
1da177e4c Linux-2.6.12-rc2 |
473 |
/* OK, work out the oldest transaction remaining in the log, and |
ae6ddcc5f [PATCH] ext3 and ... |
474 475 |
* the log block it starts at. * |
1da177e4c Linux-2.6.12-rc2 |
476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 |
* If the log is now empty, we need to work out which is the * next transaction ID we will write, and where it will * start. */ spin_lock(&journal->j_state_lock); spin_lock(&journal->j_list_lock); transaction = journal->j_checkpoint_transactions; if (transaction) { first_tid = transaction->t_tid; blocknr = transaction->t_log_start; } else if ((transaction = journal->j_committing_transaction) != NULL) { first_tid = transaction->t_tid; blocknr = transaction->t_log_start; } else if ((transaction = journal->j_running_transaction) != NULL) { first_tid = transaction->t_tid; blocknr = journal->j_head; } else { first_tid = journal->j_transaction_sequence; blocknr = journal->j_head; } spin_unlock(&journal->j_list_lock); J_ASSERT(blocknr != 0); /* If the oldest pinned transaction is at the tail of the log already then there's not much we can do right now. */ if (journal->j_tail_sequence == first_tid) { spin_unlock(&journal->j_state_lock); return 1; } /* OK, update the superblock to recover the freed space. * Physical blocks come first: have we wrapped beyond the end of * the log? */ freed = blocknr - journal->j_tail; if (blocknr < journal->j_tail) freed = freed + journal->j_last - journal->j_first; |
99cb1a318 jbd: Add fixed tr... |
512 |
trace_jbd_cleanup_journal_tail(journal, first_tid, blocknr, freed); |
1da177e4c Linux-2.6.12-rc2 |
513 |
jbd_debug(1, |
9c28cbcce jbd: Journal bloc... |
514 515 516 |
"Cleaning journal tail from %d to %d (offset %u), " "freeing %u ", |
1da177e4c Linux-2.6.12-rc2 |
517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 |
journal->j_tail_sequence, first_tid, blocknr, freed); journal->j_free += freed; journal->j_tail_sequence = first_tid; journal->j_tail = blocknr; spin_unlock(&journal->j_state_lock); if (!(journal->j_flags & JFS_ABORT)) journal_update_superblock(journal, 1); return 0; } /* Checkpoint list management */ /* |
78ce89c92 [PATCH] JBD: spli... |
532 533 |
* journal_clean_one_cp_list * |
bb189247f jbd: Fix oops in ... |
534 535 |
* Find all the written-back checkpoint buffers in the given list and release * them. |
78ce89c92 [PATCH] JBD: spli... |
536 |
* |
78ce89c92 [PATCH] JBD: spli... |
537 |
* Called with j_list_lock held. |
90802ed9c treewide: Fix com... |
538 |
* Returns number of buffers reaped (for debug) |
78ce89c92 [PATCH] JBD: spli... |
539 540 541 542 543 544 545 546 547 548 549 |
*/ static int journal_clean_one_cp_list(struct journal_head *jh, int *released) { struct journal_head *last_jh; struct journal_head *next_jh = jh; int ret, freed = 0; *released = 0; if (!jh) return 0; |
e9ad5620b [PATCH] ext3: Mor... |
550 |
last_jh = jh->b_cpprev; |
78ce89c92 [PATCH] JBD: spli... |
551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 |
do { jh = next_jh; next_jh = jh->b_cpnext; /* Use trylock because of the ranking */ if (jbd_trylock_bh_state(jh2bh(jh))) { ret = __try_to_free_cp_buf(jh); if (ret) { freed++; if (ret == 2) { *released = 1; return freed; } } } /* * This function only frees up some memory * if possible so we dont have an obligation * to finish processing. Bail out if preemption * requested: */ if (need_resched()) return freed; } while (jh != last_jh); return freed; } /* |
1da177e4c Linux-2.6.12-rc2 |
579 580 581 582 583 584 |
* journal_clean_checkpoint_list * * Find all the written-back checkpoint buffers in the journal and release them. * * Called with the journal locked. * Called with j_list_lock held. |
78ce89c92 [PATCH] JBD: spli... |
585 |
* Returns number of buffers reaped (for debug) |
1da177e4c Linux-2.6.12-rc2 |
586 587 588 589 590 |
*/ int __journal_clean_checkpoint_list(journal_t *journal) { transaction_t *transaction, *last_transaction, *next_transaction; |
7c8903f63 [PATCH] jbd: reve... |
591 |
int ret = 0; |
78ce89c92 [PATCH] JBD: spli... |
592 |
int released; |
1da177e4c Linux-2.6.12-rc2 |
593 594 |
transaction = journal->j_checkpoint_transactions; |
78ce89c92 [PATCH] JBD: spli... |
595 |
if (!transaction) |
1da177e4c Linux-2.6.12-rc2 |
596 597 598 599 600 |
goto out; last_transaction = transaction->t_cpprev; next_transaction = transaction; do { |
1da177e4c Linux-2.6.12-rc2 |
601 602 |
transaction = next_transaction; next_transaction = transaction->t_cpnext; |
78ce89c92 [PATCH] JBD: spli... |
603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 |
ret += journal_clean_one_cp_list(transaction-> t_checkpoint_list, &released); /* * This function only frees up some memory if possible so we * dont have an obligation to finish processing. Bail out if * preemption requested: */ if (need_resched()) goto out; if (released) continue; /* * It is essential that we are as careful as in the case of * t_checkpoint_list with removing the buffer from the list as * we can possibly see not yet submitted buffers on io_list */ ret += journal_clean_one_cp_list(transaction-> t_checkpoint_io_list, &released); if (need_resched()) goto out; |
1da177e4c Linux-2.6.12-rc2 |
623 624 625 626 |
} while (transaction != last_transaction); out: return ret; } |
ae6ddcc5f [PATCH] ext3 and ... |
627 |
/* |
1da177e4c Linux-2.6.12-rc2 |
628 629 630 631 632 633 634 635 |
* journal_remove_checkpoint: called after a buffer has been committed * to disk (either by being write-back flushed to disk, or being * committed to the log). * * We cannot safely clean a transaction out of the log until all of the * buffer updates committed in that transaction have safely been stored * elsewhere on disk. To achieve this, all of the buffers in a * transaction need to be maintained on the transaction's checkpoint |
78ce89c92 [PATCH] JBD: spli... |
636 |
* lists until they have been rewritten, at which point this function is |
1da177e4c Linux-2.6.12-rc2 |
637 |
* called to remove the buffer from the existing transaction's |
78ce89c92 [PATCH] JBD: spli... |
638 639 640 |
* checkpoint lists. * * The function returns 1 if it frees the transaction, 0 otherwise. |
bb189247f jbd: Fix oops in ... |
641 |
* The function can free jh and bh. |
1da177e4c Linux-2.6.12-rc2 |
642 |
* |
1da177e4c Linux-2.6.12-rc2 |
643 |
* This function is called with j_list_lock held. |
78ce89c92 [PATCH] JBD: spli... |
644 |
* This function is called with jbd_lock_bh_state(jh2bh(jh)) |
1da177e4c Linux-2.6.12-rc2 |
645 |
*/ |
78ce89c92 [PATCH] JBD: spli... |
646 |
int __journal_remove_checkpoint(struct journal_head *jh) |
1da177e4c Linux-2.6.12-rc2 |
647 648 649 |
{ transaction_t *transaction; journal_t *journal; |
78ce89c92 [PATCH] JBD: spli... |
650 |
int ret = 0; |
1da177e4c Linux-2.6.12-rc2 |
651 652 653 654 655 656 657 658 |
JBUFFER_TRACE(jh, "entry"); if ((transaction = jh->b_cp_transaction) == NULL) { JBUFFER_TRACE(jh, "not on transaction"); goto out; } journal = transaction->t_journal; |
bb189247f jbd: Fix oops in ... |
659 |
JBUFFER_TRACE(jh, "removing from transaction"); |
1da177e4c Linux-2.6.12-rc2 |
660 |
__buffer_unlink(jh); |
78ce89c92 [PATCH] JBD: spli... |
661 |
jh->b_cp_transaction = NULL; |
bb189247f jbd: Fix oops in ... |
662 |
journal_put_journal_head(jh); |
1da177e4c Linux-2.6.12-rc2 |
663 |
|
78ce89c92 [PATCH] JBD: spli... |
664 665 |
if (transaction->t_checkpoint_list != NULL || transaction->t_checkpoint_io_list != NULL) |
1da177e4c Linux-2.6.12-rc2 |
666 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
667 668 669 |
/* * There is one special case to worry about: if we have just pulled the |
d4beaf4ab jbd: Fix assertio... |
670 671 672 |
* buffer off a running or committing transaction's checkpoing list, * then even if the checkpoint list is empty, the transaction obviously * cannot be dropped! |
1da177e4c Linux-2.6.12-rc2 |
673 |
* |
d4beaf4ab jbd: Fix assertio... |
674 |
* The locking here around t_state is a bit sleazy. |
1da177e4c Linux-2.6.12-rc2 |
675 676 |
* See the comment at the end of journal_commit_transaction(). */ |
bb189247f jbd: Fix oops in ... |
677 |
if (transaction->t_state != T_FINISHED) |
1da177e4c Linux-2.6.12-rc2 |
678 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
679 680 681 682 683 684 685 686 687 |
/* OK, that was the last buffer for the transaction: we can now safely remove this transaction from the log */ __journal_drop_transaction(journal, transaction); /* Just in case anybody was waiting for more transactions to be checkpointed... */ wake_up(&journal->j_wait_logspace); |
78ce89c92 [PATCH] JBD: spli... |
688 |
ret = 1; |
1da177e4c Linux-2.6.12-rc2 |
689 |
out: |
78ce89c92 [PATCH] JBD: spli... |
690 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
691 692 693 694 695 696 697 698 699 700 |
} /* * journal_insert_checkpoint: put a committed buffer onto a checkpoint * list so that we know when it is safe to clean the transaction out of * the log. * * Called with the journal locked. * Called with j_list_lock held. */ |
ae6ddcc5f [PATCH] ext3 and ... |
701 |
void __journal_insert_checkpoint(struct journal_head *jh, |
1da177e4c Linux-2.6.12-rc2 |
702 703 704 705 706 |
transaction_t *transaction) { JBUFFER_TRACE(jh, "entry"); J_ASSERT_JH(jh, buffer_dirty(jh2bh(jh)) || buffer_jbddirty(jh2bh(jh))); J_ASSERT_JH(jh, jh->b_cp_transaction == NULL); |
bb189247f jbd: Fix oops in ... |
707 708 |
/* Get reference for checkpointing transaction */ journal_grab_journal_head(jh2bh(jh)); |
1da177e4c Linux-2.6.12-rc2 |
709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 |
jh->b_cp_transaction = transaction; if (!transaction->t_checkpoint_list) { jh->b_cpnext = jh->b_cpprev = jh; } else { jh->b_cpnext = transaction->t_checkpoint_list; jh->b_cpprev = transaction->t_checkpoint_list->b_cpprev; jh->b_cpprev->b_cpnext = jh; jh->b_cpnext->b_cpprev = jh; } transaction->t_checkpoint_list = jh; } /* * We've finished with this transaction structure: adios... |
ae6ddcc5f [PATCH] ext3 and ... |
724 |
* |
1da177e4c Linux-2.6.12-rc2 |
725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 |
* The transaction must have no links except for the checkpoint by this * point. * * Called with the journal locked. * Called with j_list_lock held. */ void __journal_drop_transaction(journal_t *journal, transaction_t *transaction) { assert_spin_locked(&journal->j_list_lock); if (transaction->t_cpnext) { transaction->t_cpnext->t_cpprev = transaction->t_cpprev; transaction->t_cpprev->t_cpnext = transaction->t_cpnext; if (journal->j_checkpoint_transactions == transaction) journal->j_checkpoint_transactions = transaction->t_cpnext; if (journal->j_checkpoint_transactions == transaction) journal->j_checkpoint_transactions = NULL; } J_ASSERT(transaction->t_state == T_FINISHED); J_ASSERT(transaction->t_buffers == NULL); J_ASSERT(transaction->t_sync_datalist == NULL); J_ASSERT(transaction->t_forget == NULL); J_ASSERT(transaction->t_iobuf_list == NULL); J_ASSERT(transaction->t_shadow_list == NULL); J_ASSERT(transaction->t_log_list == NULL); J_ASSERT(transaction->t_checkpoint_list == NULL); |
78ce89c92 [PATCH] JBD: spli... |
753 |
J_ASSERT(transaction->t_checkpoint_io_list == NULL); |
1da177e4c Linux-2.6.12-rc2 |
754 755 756 |
J_ASSERT(transaction->t_updates == 0); J_ASSERT(journal->j_committing_transaction != transaction); J_ASSERT(journal->j_running_transaction != transaction); |
99cb1a318 jbd: Add fixed tr... |
757 |
trace_jbd_drop_transaction(journal, transaction); |
1da177e4c Linux-2.6.12-rc2 |
758 759 760 761 |
jbd_debug(1, "Dropping transaction %d, all done ", transaction->t_tid); kfree(transaction); } |