forked from grate-driver/linux
-
Notifications
You must be signed in to change notification settings - Fork 0
/
journal.c
1509 lines (1242 loc) · 37.5 KB
/
journal.c
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
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
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
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
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
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
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
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
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
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
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
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
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/* -*- mode: c; c-basic-offset: 8; -*-
* vim: noexpandtab sw=8 ts=8 sts=0:
*
* journal.c
*
* Defines functions of journalling api
*
* Copyright (C) 2003, 2004 Oracle. All rights reserved.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public
* License along with this program; if not, write to the
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
* Boston, MA 021110-1307, USA.
*/
#include <linux/fs.h>
#include <linux/types.h>
#include <linux/slab.h>
#include <linux/highmem.h>
#include <linux/kthread.h>
#define MLOG_MASK_PREFIX ML_JOURNAL
#include <cluster/masklog.h>
#include "ocfs2.h"
#include "alloc.h"
#include "dlmglue.h"
#include "extent_map.h"
#include "heartbeat.h"
#include "inode.h"
#include "journal.h"
#include "localalloc.h"
#include "namei.h"
#include "slot_map.h"
#include "super.h"
#include "vote.h"
#include "sysfile.h"
#include "buffer_head_io.h"
DEFINE_SPINLOCK(trans_inc_lock);
static int ocfs2_force_read_journal(struct inode *inode);
static int ocfs2_recover_node(struct ocfs2_super *osb,
int node_num);
static int __ocfs2_recovery_thread(void *arg);
static int ocfs2_commit_cache(struct ocfs2_super *osb);
static int ocfs2_wait_on_mount(struct ocfs2_super *osb);
static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
int dirty);
static int ocfs2_trylock_journal(struct ocfs2_super *osb,
int slot_num);
static int ocfs2_recover_orphans(struct ocfs2_super *osb,
int slot);
static int ocfs2_commit_thread(void *arg);
static int ocfs2_commit_cache(struct ocfs2_super *osb)
{
int status = 0;
unsigned int flushed;
unsigned long old_id;
struct ocfs2_journal *journal = NULL;
mlog_entry_void();
journal = osb->journal;
/* Flush all pending commits and checkpoint the journal. */
down_write(&journal->j_trans_barrier);
if (atomic_read(&journal->j_num_trans) == 0) {
up_write(&journal->j_trans_barrier);
mlog(0, "No transactions for me to flush!\n");
goto finally;
}
journal_lock_updates(journal->j_journal);
status = journal_flush(journal->j_journal);
journal_unlock_updates(journal->j_journal);
if (status < 0) {
up_write(&journal->j_trans_barrier);
mlog_errno(status);
goto finally;
}
old_id = ocfs2_inc_trans_id(journal);
flushed = atomic_read(&journal->j_num_trans);
atomic_set(&journal->j_num_trans, 0);
up_write(&journal->j_trans_barrier);
mlog(0, "commit_thread: flushed transaction %lu (%u handles)\n",
journal->j_trans_id, flushed);
ocfs2_kick_vote_thread(osb);
wake_up(&journal->j_checkpointed);
finally:
mlog_exit(status);
return status;
}
/* pass it NULL and it will allocate a new handle object for you. If
* you pass it a handle however, it may still return error, in which
* case it has free'd the passed handle for you. */
handle_t *ocfs2_start_trans(struct ocfs2_super *osb, int max_buffs)
{
journal_t *journal = osb->journal->j_journal;
handle_t *handle;
BUG_ON(!osb || !osb->journal->j_journal);
if (ocfs2_is_hard_readonly(osb))
return ERR_PTR(-EROFS);
BUG_ON(osb->journal->j_state == OCFS2_JOURNAL_FREE);
BUG_ON(max_buffs <= 0);
/* JBD might support this, but our journalling code doesn't yet. */
if (journal_current_handle()) {
mlog(ML_ERROR, "Recursive transaction attempted!\n");
BUG();
}
down_read(&osb->journal->j_trans_barrier);
handle = journal_start(journal, max_buffs);
if (IS_ERR(handle)) {
up_read(&osb->journal->j_trans_barrier);
mlog_errno(PTR_ERR(handle));
if (is_journal_aborted(journal)) {
ocfs2_abort(osb->sb, "Detected aborted journal");
handle = ERR_PTR(-EROFS);
}
} else
atomic_inc(&(osb->journal->j_num_trans));
return handle;
}
int ocfs2_commit_trans(struct ocfs2_super *osb,
handle_t *handle)
{
int ret;
struct ocfs2_journal *journal = osb->journal;
BUG_ON(!handle);
ret = journal_stop(handle);
if (ret < 0)
mlog_errno(ret);
up_read(&journal->j_trans_barrier);
return ret;
}
/*
* 'nblocks' is what you want to add to the current
* transaction. extend_trans will either extend the current handle by
* nblocks, or commit it and start a new one with nblocks credits.
*
* WARNING: This will not release any semaphores or disk locks taken
* during the transaction, so make sure they were taken *before*
* start_trans or we'll have ordering deadlocks.
*
* WARNING2: Note that we do *not* drop j_trans_barrier here. This is
* good because transaction ids haven't yet been recorded on the
* cluster locks associated with this handle.
*/
int ocfs2_extend_trans(handle_t *handle, int nblocks)
{
int status;
BUG_ON(!handle);
BUG_ON(!nblocks);
mlog_entry_void();
mlog(0, "Trying to extend transaction by %d blocks\n", nblocks);
status = journal_extend(handle, nblocks);
if (status < 0) {
mlog_errno(status);
goto bail;
}
if (status > 0) {
mlog(0, "journal_extend failed, trying journal_restart\n");
status = journal_restart(handle, nblocks);
if (status < 0) {
mlog_errno(status);
goto bail;
}
}
status = 0;
bail:
mlog_exit(status);
return status;
}
int ocfs2_journal_access(handle_t *handle,
struct inode *inode,
struct buffer_head *bh,
int type)
{
int status;
BUG_ON(!inode);
BUG_ON(!handle);
BUG_ON(!bh);
mlog_entry("bh->b_blocknr=%llu, type=%d (\"%s\"), bh->b_size = %zu\n",
(unsigned long long)bh->b_blocknr, type,
(type == OCFS2_JOURNAL_ACCESS_CREATE) ?
"OCFS2_JOURNAL_ACCESS_CREATE" :
"OCFS2_JOURNAL_ACCESS_WRITE",
bh->b_size);
/* we can safely remove this assertion after testing. */
if (!buffer_uptodate(bh)) {
mlog(ML_ERROR, "giving me a buffer that's not uptodate!\n");
mlog(ML_ERROR, "b_blocknr=%llu\n",
(unsigned long long)bh->b_blocknr);
BUG();
}
/* Set the current transaction information on the inode so
* that the locking code knows whether it can drop it's locks
* on this inode or not. We're protected from the commit
* thread updating the current transaction id until
* ocfs2_commit_trans() because ocfs2_start_trans() took
* j_trans_barrier for us. */
ocfs2_set_inode_lock_trans(OCFS2_SB(inode->i_sb)->journal, inode);
mutex_lock(&OCFS2_I(inode)->ip_io_mutex);
switch (type) {
case OCFS2_JOURNAL_ACCESS_CREATE:
case OCFS2_JOURNAL_ACCESS_WRITE:
status = journal_get_write_access(handle, bh);
break;
case OCFS2_JOURNAL_ACCESS_UNDO:
status = journal_get_undo_access(handle, bh);
break;
default:
status = -EINVAL;
mlog(ML_ERROR, "Uknown access type!\n");
}
mutex_unlock(&OCFS2_I(inode)->ip_io_mutex);
if (status < 0)
mlog(ML_ERROR, "Error %d getting %d access to buffer!\n",
status, type);
mlog_exit(status);
return status;
}
int ocfs2_journal_dirty(handle_t *handle,
struct buffer_head *bh)
{
int status;
mlog_entry("(bh->b_blocknr=%llu)\n",
(unsigned long long)bh->b_blocknr);
status = journal_dirty_metadata(handle, bh);
if (status < 0)
mlog(ML_ERROR, "Could not dirty metadata buffer. "
"(bh->b_blocknr=%llu)\n",
(unsigned long long)bh->b_blocknr);
mlog_exit(status);
return status;
}
int ocfs2_journal_dirty_data(handle_t *handle,
struct buffer_head *bh)
{
int err = journal_dirty_data(handle, bh);
if (err)
mlog_errno(err);
/* TODO: When we can handle it, abort the handle and go RO on
* error here. */
return err;
}
#define OCFS2_DEFAULT_COMMIT_INTERVAL (HZ * 5)
void ocfs2_set_journal_params(struct ocfs2_super *osb)
{
journal_t *journal = osb->journal->j_journal;
spin_lock(&journal->j_state_lock);
journal->j_commit_interval = OCFS2_DEFAULT_COMMIT_INTERVAL;
if (osb->s_mount_opt & OCFS2_MOUNT_BARRIER)
journal->j_flags |= JFS_BARRIER;
else
journal->j_flags &= ~JFS_BARRIER;
spin_unlock(&journal->j_state_lock);
}
int ocfs2_journal_init(struct ocfs2_journal *journal, int *dirty)
{
int status = -1;
struct inode *inode = NULL; /* the journal inode */
journal_t *j_journal = NULL;
struct ocfs2_dinode *di = NULL;
struct buffer_head *bh = NULL;
struct ocfs2_super *osb;
int meta_lock = 0;
mlog_entry_void();
BUG_ON(!journal);
osb = journal->j_osb;
/* already have the inode for our journal */
inode = ocfs2_get_system_file_inode(osb, JOURNAL_SYSTEM_INODE,
osb->slot_num);
if (inode == NULL) {
status = -EACCES;
mlog_errno(status);
goto done;
}
if (is_bad_inode(inode)) {
mlog(ML_ERROR, "access error (bad inode)\n");
iput(inode);
inode = NULL;
status = -EACCES;
goto done;
}
SET_INODE_JOURNAL(inode);
OCFS2_I(inode)->ip_open_count++;
/* Skip recovery waits here - journal inode metadata never
* changes in a live cluster so it can be considered an
* exception to the rule. */
status = ocfs2_meta_lock_full(inode, &bh, 1, OCFS2_META_LOCK_RECOVERY);
if (status < 0) {
if (status != -ERESTARTSYS)
mlog(ML_ERROR, "Could not get lock on journal!\n");
goto done;
}
meta_lock = 1;
di = (struct ocfs2_dinode *)bh->b_data;
if (inode->i_size < OCFS2_MIN_JOURNAL_SIZE) {
mlog(ML_ERROR, "Journal file size (%lld) is too small!\n",
inode->i_size);
status = -EINVAL;
goto done;
}
mlog(0, "inode->i_size = %lld\n", inode->i_size);
mlog(0, "inode->i_blocks = %llu\n",
(unsigned long long)inode->i_blocks);
mlog(0, "inode->ip_clusters = %u\n", OCFS2_I(inode)->ip_clusters);
/* call the kernels journal init function now */
j_journal = journal_init_inode(inode);
if (j_journal == NULL) {
mlog(ML_ERROR, "Linux journal layer error\n");
status = -EINVAL;
goto done;
}
mlog(0, "Returned from journal_init_inode\n");
mlog(0, "j_journal->j_maxlen = %u\n", j_journal->j_maxlen);
*dirty = (le32_to_cpu(di->id1.journal1.ij_flags) &
OCFS2_JOURNAL_DIRTY_FL);
journal->j_journal = j_journal;
journal->j_inode = inode;
journal->j_bh = bh;
ocfs2_set_journal_params(osb);
journal->j_state = OCFS2_JOURNAL_LOADED;
status = 0;
done:
if (status < 0) {
if (meta_lock)
ocfs2_meta_unlock(inode, 1);
if (bh != NULL)
brelse(bh);
if (inode) {
OCFS2_I(inode)->ip_open_count--;
iput(inode);
}
}
mlog_exit(status);
return status;
}
static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
int dirty)
{
int status;
unsigned int flags;
struct ocfs2_journal *journal = osb->journal;
struct buffer_head *bh = journal->j_bh;
struct ocfs2_dinode *fe;
mlog_entry_void();
fe = (struct ocfs2_dinode *)bh->b_data;
if (!OCFS2_IS_VALID_DINODE(fe)) {
/* This is called from startup/shutdown which will
* handle the errors in a specific manner, so no need
* to call ocfs2_error() here. */
mlog(ML_ERROR, "Journal dinode %llu has invalid "
"signature: %.*s", (unsigned long long)fe->i_blkno, 7,
fe->i_signature);
status = -EIO;
goto out;
}
flags = le32_to_cpu(fe->id1.journal1.ij_flags);
if (dirty)
flags |= OCFS2_JOURNAL_DIRTY_FL;
else
flags &= ~OCFS2_JOURNAL_DIRTY_FL;
fe->id1.journal1.ij_flags = cpu_to_le32(flags);
status = ocfs2_write_block(osb, bh, journal->j_inode);
if (status < 0)
mlog_errno(status);
out:
mlog_exit(status);
return status;
}
/*
* If the journal has been kmalloc'd it needs to be freed after this
* call.
*/
void ocfs2_journal_shutdown(struct ocfs2_super *osb)
{
struct ocfs2_journal *journal = NULL;
int status = 0;
struct inode *inode = NULL;
int num_running_trans = 0;
mlog_entry_void();
BUG_ON(!osb);
journal = osb->journal;
if (!journal)
goto done;
inode = journal->j_inode;
if (journal->j_state != OCFS2_JOURNAL_LOADED)
goto done;
/* need to inc inode use count as journal_destroy will iput. */
if (!igrab(inode))
BUG();
num_running_trans = atomic_read(&(osb->journal->j_num_trans));
if (num_running_trans > 0)
mlog(0, "Shutting down journal: must wait on %d "
"running transactions!\n",
num_running_trans);
/* Do a commit_cache here. It will flush our journal, *and*
* release any locks that are still held.
* set the SHUTDOWN flag and release the trans lock.
* the commit thread will take the trans lock for us below. */
journal->j_state = OCFS2_JOURNAL_IN_SHUTDOWN;
/* The OCFS2_JOURNAL_IN_SHUTDOWN will signal to commit_cache to not
* drop the trans_lock (which we want to hold until we
* completely destroy the journal. */
if (osb->commit_task) {
/* Wait for the commit thread */
mlog(0, "Waiting for ocfs2commit to exit....\n");
kthread_stop(osb->commit_task);
osb->commit_task = NULL;
}
BUG_ON(atomic_read(&(osb->journal->j_num_trans)) != 0);
status = ocfs2_journal_toggle_dirty(osb, 0);
if (status < 0)
mlog_errno(status);
/* Shutdown the kernel journal system */
journal_destroy(journal->j_journal);
OCFS2_I(inode)->ip_open_count--;
/* unlock our journal */
ocfs2_meta_unlock(inode, 1);
brelse(journal->j_bh);
journal->j_bh = NULL;
journal->j_state = OCFS2_JOURNAL_FREE;
// up_write(&journal->j_trans_barrier);
done:
if (inode)
iput(inode);
mlog_exit_void();
}
static void ocfs2_clear_journal_error(struct super_block *sb,
journal_t *journal,
int slot)
{
int olderr;
olderr = journal_errno(journal);
if (olderr) {
mlog(ML_ERROR, "File system error %d recorded in "
"journal %u.\n", olderr, slot);
mlog(ML_ERROR, "File system on device %s needs checking.\n",
sb->s_id);
journal_ack_err(journal);
journal_clear_err(journal);
}
}
int ocfs2_journal_load(struct ocfs2_journal *journal)
{
int status = 0;
struct ocfs2_super *osb;
mlog_entry_void();
if (!journal)
BUG();
osb = journal->j_osb;
status = journal_load(journal->j_journal);
if (status < 0) {
mlog(ML_ERROR, "Failed to load journal!\n");
goto done;
}
ocfs2_clear_journal_error(osb->sb, journal->j_journal, osb->slot_num);
status = ocfs2_journal_toggle_dirty(osb, 1);
if (status < 0) {
mlog_errno(status);
goto done;
}
/* Launch the commit thread */
osb->commit_task = kthread_run(ocfs2_commit_thread, osb, "ocfs2cmt");
if (IS_ERR(osb->commit_task)) {
status = PTR_ERR(osb->commit_task);
osb->commit_task = NULL;
mlog(ML_ERROR, "unable to launch ocfs2commit thread, error=%d",
status);
goto done;
}
done:
mlog_exit(status);
return status;
}
/* 'full' flag tells us whether we clear out all blocks or if we just
* mark the journal clean */
int ocfs2_journal_wipe(struct ocfs2_journal *journal, int full)
{
int status;
mlog_entry_void();
BUG_ON(!journal);
status = journal_wipe(journal->j_journal, full);
if (status < 0) {
mlog_errno(status);
goto bail;
}
status = ocfs2_journal_toggle_dirty(journal->j_osb, 0);
if (status < 0)
mlog_errno(status);
bail:
mlog_exit(status);
return status;
}
/*
* JBD Might read a cached version of another nodes journal file. We
* don't want this as this file changes often and we get no
* notification on those changes. The only way to be sure that we've
* got the most up to date version of those blocks then is to force
* read them off disk. Just searching through the buffer cache won't
* work as there may be pages backing this file which are still marked
* up to date. We know things can't change on this file underneath us
* as we have the lock by now :)
*/
static int ocfs2_force_read_journal(struct inode *inode)
{
int status = 0;
int i, p_blocks;
u64 v_blkno, p_blkno;
#define CONCURRENT_JOURNAL_FILL 32
struct buffer_head *bhs[CONCURRENT_JOURNAL_FILL];
mlog_entry_void();
BUG_ON(inode->i_blocks !=
ocfs2_align_bytes_to_sectors(i_size_read(inode)));
memset(bhs, 0, sizeof(struct buffer_head *) * CONCURRENT_JOURNAL_FILL);
mlog(0, "Force reading %llu blocks\n",
(unsigned long long)(inode->i_blocks >>
(inode->i_sb->s_blocksize_bits - 9)));
v_blkno = 0;
while (v_blkno <
(inode->i_blocks >> (inode->i_sb->s_blocksize_bits - 9))) {
status = ocfs2_extent_map_get_blocks(inode, v_blkno,
1, &p_blkno,
&p_blocks);
if (status < 0) {
mlog_errno(status);
goto bail;
}
if (p_blocks > CONCURRENT_JOURNAL_FILL)
p_blocks = CONCURRENT_JOURNAL_FILL;
/* We are reading journal data which should not
* be put in the uptodate cache */
status = ocfs2_read_blocks(OCFS2_SB(inode->i_sb),
p_blkno, p_blocks, bhs, 0,
NULL);
if (status < 0) {
mlog_errno(status);
goto bail;
}
for(i = 0; i < p_blocks; i++) {
brelse(bhs[i]);
bhs[i] = NULL;
}
v_blkno += p_blocks;
}
bail:
for(i = 0; i < CONCURRENT_JOURNAL_FILL; i++)
if (bhs[i])
brelse(bhs[i]);
mlog_exit(status);
return status;
}
struct ocfs2_la_recovery_item {
struct list_head lri_list;
int lri_slot;
struct ocfs2_dinode *lri_la_dinode;
struct ocfs2_dinode *lri_tl_dinode;
};
/* Does the second half of the recovery process. By this point, the
* node is marked clean and can actually be considered recovered,
* hence it's no longer in the recovery map, but there's still some
* cleanup we can do which shouldn't happen within the recovery thread
* as locking in that context becomes very difficult if we are to take
* recovering nodes into account.
*
* NOTE: This function can and will sleep on recovery of other nodes
* during cluster locking, just like any other ocfs2 process.
*/
void ocfs2_complete_recovery(struct work_struct *work)
{
int ret;
struct ocfs2_journal *journal =
container_of(work, struct ocfs2_journal, j_recovery_work);
struct ocfs2_super *osb = journal->j_osb;
struct ocfs2_dinode *la_dinode, *tl_dinode;
struct ocfs2_la_recovery_item *item;
struct list_head *p, *n;
LIST_HEAD(tmp_la_list);
mlog_entry_void();
mlog(0, "completing recovery from keventd\n");
spin_lock(&journal->j_lock);
list_splice_init(&journal->j_la_cleanups, &tmp_la_list);
spin_unlock(&journal->j_lock);
list_for_each_safe(p, n, &tmp_la_list) {
item = list_entry(p, struct ocfs2_la_recovery_item, lri_list);
list_del_init(&item->lri_list);
mlog(0, "Complete recovery for slot %d\n", item->lri_slot);
la_dinode = item->lri_la_dinode;
if (la_dinode) {
mlog(0, "Clean up local alloc %llu\n",
(unsigned long long)la_dinode->i_blkno);
ret = ocfs2_complete_local_alloc_recovery(osb,
la_dinode);
if (ret < 0)
mlog_errno(ret);
kfree(la_dinode);
}
tl_dinode = item->lri_tl_dinode;
if (tl_dinode) {
mlog(0, "Clean up truncate log %llu\n",
(unsigned long long)tl_dinode->i_blkno);
ret = ocfs2_complete_truncate_log_recovery(osb,
tl_dinode);
if (ret < 0)
mlog_errno(ret);
kfree(tl_dinode);
}
ret = ocfs2_recover_orphans(osb, item->lri_slot);
if (ret < 0)
mlog_errno(ret);
kfree(item);
}
mlog(0, "Recovery completion\n");
mlog_exit_void();
}
/* NOTE: This function always eats your references to la_dinode and
* tl_dinode, either manually on error, or by passing them to
* ocfs2_complete_recovery */
static void ocfs2_queue_recovery_completion(struct ocfs2_journal *journal,
int slot_num,
struct ocfs2_dinode *la_dinode,
struct ocfs2_dinode *tl_dinode)
{
struct ocfs2_la_recovery_item *item;
item = kmalloc(sizeof(struct ocfs2_la_recovery_item), GFP_NOFS);
if (!item) {
/* Though we wish to avoid it, we are in fact safe in
* skipping local alloc cleanup as fsck.ocfs2 is more
* than capable of reclaiming unused space. */
if (la_dinode)
kfree(la_dinode);
if (tl_dinode)
kfree(tl_dinode);
mlog_errno(-ENOMEM);
return;
}
INIT_LIST_HEAD(&item->lri_list);
item->lri_la_dinode = la_dinode;
item->lri_slot = slot_num;
item->lri_tl_dinode = tl_dinode;
spin_lock(&journal->j_lock);
list_add_tail(&item->lri_list, &journal->j_la_cleanups);
queue_work(ocfs2_wq, &journal->j_recovery_work);
spin_unlock(&journal->j_lock);
}
/* Called by the mount code to queue recovery the last part of
* recovery for it's own slot. */
void ocfs2_complete_mount_recovery(struct ocfs2_super *osb)
{
struct ocfs2_journal *journal = osb->journal;
if (osb->dirty) {
/* No need to queue up our truncate_log as regular
* cleanup will catch that. */
ocfs2_queue_recovery_completion(journal,
osb->slot_num,
osb->local_alloc_copy,
NULL);
ocfs2_schedule_truncate_log_flush(osb, 0);
osb->local_alloc_copy = NULL;
osb->dirty = 0;
}
}
static int __ocfs2_recovery_thread(void *arg)
{
int status, node_num;
struct ocfs2_super *osb = arg;
mlog_entry_void();
status = ocfs2_wait_on_mount(osb);
if (status < 0) {
goto bail;
}
restart:
status = ocfs2_super_lock(osb, 1);
if (status < 0) {
mlog_errno(status);
goto bail;
}
while(!ocfs2_node_map_is_empty(osb, &osb->recovery_map)) {
node_num = ocfs2_node_map_first_set_bit(osb,
&osb->recovery_map);
if (node_num == O2NM_INVALID_NODE_NUM) {
mlog(0, "Out of nodes to recover.\n");
break;
}
status = ocfs2_recover_node(osb, node_num);
if (status < 0) {
mlog(ML_ERROR,
"Error %d recovering node %d on device (%u,%u)!\n",
status, node_num,
MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev));
mlog(ML_ERROR, "Volume requires unmount.\n");
continue;
}
ocfs2_recovery_map_clear(osb, node_num);
}
ocfs2_super_unlock(osb, 1);
/* We always run recovery on our own orphan dir - the dead
* node(s) may have voted "no" on an inode delete earlier. A
* revote is therefore required. */
ocfs2_queue_recovery_completion(osb->journal, osb->slot_num, NULL,
NULL);
bail:
mutex_lock(&osb->recovery_lock);
if (!status &&
!ocfs2_node_map_is_empty(osb, &osb->recovery_map)) {
mutex_unlock(&osb->recovery_lock);
goto restart;
}
osb->recovery_thread_task = NULL;
mb(); /* sync with ocfs2_recovery_thread_running */
wake_up(&osb->recovery_event);
mutex_unlock(&osb->recovery_lock);
mlog_exit(status);
/* no one is callint kthread_stop() for us so the kthread() api
* requires that we call do_exit(). And it isn't exported, but
* complete_and_exit() seems to be a minimal wrapper around it. */
complete_and_exit(NULL, status);
return status;
}
void ocfs2_recovery_thread(struct ocfs2_super *osb, int node_num)
{
mlog_entry("(node_num=%d, osb->node_num = %d)\n",
node_num, osb->node_num);
mutex_lock(&osb->recovery_lock);
if (osb->disable_recovery)
goto out;
/* People waiting on recovery will wait on
* the recovery map to empty. */
if (!ocfs2_recovery_map_set(osb, node_num))
mlog(0, "node %d already be in recovery.\n", node_num);
mlog(0, "starting recovery thread...\n");
if (osb->recovery_thread_task)
goto out;
osb->recovery_thread_task = kthread_run(__ocfs2_recovery_thread, osb,
"ocfs2rec");
if (IS_ERR(osb->recovery_thread_task)) {
mlog_errno((int)PTR_ERR(osb->recovery_thread_task));
osb->recovery_thread_task = NULL;
}
out:
mutex_unlock(&osb->recovery_lock);
wake_up(&osb->recovery_event);
mlog_exit_void();
}
/* Does the actual journal replay and marks the journal inode as
* clean. Will only replay if the journal inode is marked dirty. */
static int ocfs2_replay_journal(struct ocfs2_super *osb,
int node_num,
int slot_num)
{
int status;
int got_lock = 0;
unsigned int flags;
struct inode *inode = NULL;
struct ocfs2_dinode *fe;
journal_t *journal = NULL;
struct buffer_head *bh = NULL;
inode = ocfs2_get_system_file_inode(osb, JOURNAL_SYSTEM_INODE,
slot_num);
if (inode == NULL) {
status = -EACCES;
mlog_errno(status);
goto done;
}
if (is_bad_inode(inode)) {
status = -EACCES;
iput(inode);
inode = NULL;
mlog_errno(status);
goto done;
}
SET_INODE_JOURNAL(inode);
status = ocfs2_meta_lock_full(inode, &bh, 1, OCFS2_META_LOCK_RECOVERY);
if (status < 0) {
mlog(0, "status returned from ocfs2_meta_lock=%d\n", status);
if (status != -ERESTARTSYS)
mlog(ML_ERROR, "Could not lock journal!\n");
goto done;
}
got_lock = 1;
fe = (struct ocfs2_dinode *) bh->b_data;
flags = le32_to_cpu(fe->id1.journal1.ij_flags);
if (!(flags & OCFS2_JOURNAL_DIRTY_FL)) {
mlog(0, "No recovery required for node %d\n", node_num);
goto done;
}
mlog(ML_NOTICE, "Recovering node %d from slot %d on device (%u,%u)\n",
node_num, slot_num,
MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev));
OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
status = ocfs2_force_read_journal(inode);
if (status < 0) {
mlog_errno(status);
goto done;
}
mlog(0, "calling journal_init_inode\n");
journal = journal_init_inode(inode);
if (journal == NULL) {
mlog(ML_ERROR, "Linux journal layer error\n");
status = -EIO;
goto done;
}
status = journal_load(journal);
if (status < 0) {
mlog_errno(status);
if (!igrab(inode))
BUG();
journal_destroy(journal);
goto done;