forked from torvalds/linux
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathinode.c
1458 lines (1230 loc) · 39.4 KB
/
inode.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:
*
* inode.c
*
* vfs' aops, fops, dops and iops
*
* Copyright (C) 2002, 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/pagemap.h>
#include <linux/quotaops.h>
#include <asm/byteorder.h>
#define MLOG_MASK_PREFIX ML_INODE
#include <cluster/masklog.h>
#include "ocfs2.h"
#include "alloc.h"
#include "dir.h"
#include "blockcheck.h"
#include "dlmglue.h"
#include "extent_map.h"
#include "file.h"
#include "heartbeat.h"
#include "inode.h"
#include "journal.h"
#include "namei.h"
#include "suballoc.h"
#include "super.h"
#include "symlink.h"
#include "sysfile.h"
#include "uptodate.h"
#include "xattr.h"
#include "refcounttree.h"
#include "buffer_head_io.h"
struct ocfs2_find_inode_args
{
u64 fi_blkno;
unsigned long fi_ino;
unsigned int fi_flags;
unsigned int fi_sysfile_type;
};
static struct lock_class_key ocfs2_sysfile_lock_key[NUM_SYSTEM_INODES];
static int ocfs2_read_locked_inode(struct inode *inode,
struct ocfs2_find_inode_args *args);
static int ocfs2_init_locked_inode(struct inode *inode, void *opaque);
static int ocfs2_find_actor(struct inode *inode, void *opaque);
static int ocfs2_truncate_for_delete(struct ocfs2_super *osb,
struct inode *inode,
struct buffer_head *fe_bh);
void ocfs2_set_inode_flags(struct inode *inode)
{
unsigned int flags = OCFS2_I(inode)->ip_attr;
inode->i_flags &= ~(S_IMMUTABLE |
S_SYNC | S_APPEND | S_NOATIME | S_DIRSYNC);
if (flags & OCFS2_IMMUTABLE_FL)
inode->i_flags |= S_IMMUTABLE;
if (flags & OCFS2_SYNC_FL)
inode->i_flags |= S_SYNC;
if (flags & OCFS2_APPEND_FL)
inode->i_flags |= S_APPEND;
if (flags & OCFS2_NOATIME_FL)
inode->i_flags |= S_NOATIME;
if (flags & OCFS2_DIRSYNC_FL)
inode->i_flags |= S_DIRSYNC;
}
/* Propagate flags from i_flags to OCFS2_I(inode)->ip_attr */
void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi)
{
unsigned int flags = oi->vfs_inode.i_flags;
oi->ip_attr &= ~(OCFS2_SYNC_FL|OCFS2_APPEND_FL|
OCFS2_IMMUTABLE_FL|OCFS2_NOATIME_FL|OCFS2_DIRSYNC_FL);
if (flags & S_SYNC)
oi->ip_attr |= OCFS2_SYNC_FL;
if (flags & S_APPEND)
oi->ip_attr |= OCFS2_APPEND_FL;
if (flags & S_IMMUTABLE)
oi->ip_attr |= OCFS2_IMMUTABLE_FL;
if (flags & S_NOATIME)
oi->ip_attr |= OCFS2_NOATIME_FL;
if (flags & S_DIRSYNC)
oi->ip_attr |= OCFS2_DIRSYNC_FL;
}
struct inode *ocfs2_ilookup(struct super_block *sb, u64 blkno)
{
struct ocfs2_find_inode_args args;
args.fi_blkno = blkno;
args.fi_flags = 0;
args.fi_ino = ino_from_blkno(sb, blkno);
args.fi_sysfile_type = 0;
return ilookup5(sb, blkno, ocfs2_find_actor, &args);
}
struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags,
int sysfile_type)
{
struct inode *inode = NULL;
struct super_block *sb = osb->sb;
struct ocfs2_find_inode_args args;
mlog_entry("(blkno = %llu)\n", (unsigned long long)blkno);
/* Ok. By now we've either got the offsets passed to us by the
* caller, or we just pulled them off the bh. Lets do some
* sanity checks to make sure they're OK. */
if (blkno == 0) {
inode = ERR_PTR(-EINVAL);
mlog_errno(PTR_ERR(inode));
goto bail;
}
args.fi_blkno = blkno;
args.fi_flags = flags;
args.fi_ino = ino_from_blkno(sb, blkno);
args.fi_sysfile_type = sysfile_type;
inode = iget5_locked(sb, args.fi_ino, ocfs2_find_actor,
ocfs2_init_locked_inode, &args);
/* inode was *not* in the inode cache. 2.6.x requires
* us to do our own read_inode call and unlock it
* afterwards. */
if (inode && inode->i_state & I_NEW) {
mlog(0, "Inode was not in inode cache, reading it.\n");
ocfs2_read_locked_inode(inode, &args);
unlock_new_inode(inode);
}
if (inode == NULL) {
inode = ERR_PTR(-ENOMEM);
mlog_errno(PTR_ERR(inode));
goto bail;
}
if (is_bad_inode(inode)) {
iput(inode);
inode = ERR_PTR(-ESTALE);
goto bail;
}
bail:
if (!IS_ERR(inode)) {
mlog(0, "returning inode with number %llu\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno);
mlog_exit_ptr(inode);
}
return inode;
}
/*
* here's how inodes get read from disk:
* iget5_locked -> find_actor -> OCFS2_FIND_ACTOR
* found? : return the in-memory inode
* not found? : get_new_inode -> OCFS2_INIT_LOCKED_INODE
*/
static int ocfs2_find_actor(struct inode *inode, void *opaque)
{
struct ocfs2_find_inode_args *args = NULL;
struct ocfs2_inode_info *oi = OCFS2_I(inode);
int ret = 0;
mlog_entry("(0x%p, %lu, 0x%p)\n", inode, inode->i_ino, opaque);
args = opaque;
mlog_bug_on_msg(!inode, "No inode in find actor!\n");
if (oi->ip_blkno != args->fi_blkno)
goto bail;
ret = 1;
bail:
mlog_exit(ret);
return ret;
}
/*
* initialize the new inode, but don't do anything that would cause
* us to sleep.
* return 0 on success, 1 on failure
*/
static int ocfs2_init_locked_inode(struct inode *inode, void *opaque)
{
struct ocfs2_find_inode_args *args = opaque;
static struct lock_class_key ocfs2_quota_ip_alloc_sem_key,
ocfs2_file_ip_alloc_sem_key;
mlog_entry("inode = %p, opaque = %p\n", inode, opaque);
inode->i_ino = args->fi_ino;
OCFS2_I(inode)->ip_blkno = args->fi_blkno;
if (args->fi_sysfile_type != 0)
lockdep_set_class(&inode->i_mutex,
&ocfs2_sysfile_lock_key[args->fi_sysfile_type]);
if (args->fi_sysfile_type == USER_QUOTA_SYSTEM_INODE ||
args->fi_sysfile_type == GROUP_QUOTA_SYSTEM_INODE ||
args->fi_sysfile_type == LOCAL_USER_QUOTA_SYSTEM_INODE ||
args->fi_sysfile_type == LOCAL_GROUP_QUOTA_SYSTEM_INODE)
lockdep_set_class(&OCFS2_I(inode)->ip_alloc_sem,
&ocfs2_quota_ip_alloc_sem_key);
else
lockdep_set_class(&OCFS2_I(inode)->ip_alloc_sem,
&ocfs2_file_ip_alloc_sem_key);
mlog_exit(0);
return 0;
}
void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
int create_ino)
{
struct super_block *sb;
struct ocfs2_super *osb;
int use_plocks = 1;
mlog_entry("(0x%p, size:%llu)\n", inode,
(unsigned long long)le64_to_cpu(fe->i_size));
sb = inode->i_sb;
osb = OCFS2_SB(sb);
if ((osb->s_mount_opt & OCFS2_MOUNT_LOCALFLOCKS) ||
ocfs2_mount_local(osb) || !ocfs2_stack_supports_plocks())
use_plocks = 0;
/*
* These have all been checked by ocfs2_read_inode_block() or set
* by ocfs2_mknod_locked(), so a failure is a code bug.
*/
BUG_ON(!OCFS2_IS_VALID_DINODE(fe)); /* This means that read_inode
cannot create a superblock
inode today. change if
that is needed. */
BUG_ON(!(fe->i_flags & cpu_to_le32(OCFS2_VALID_FL)));
BUG_ON(le32_to_cpu(fe->i_fs_generation) != osb->fs_generation);
OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
OCFS2_I(inode)->ip_attr = le32_to_cpu(fe->i_attr);
OCFS2_I(inode)->ip_dyn_features = le16_to_cpu(fe->i_dyn_features);
inode->i_version = 1;
inode->i_generation = le32_to_cpu(fe->i_generation);
inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
inode->i_mode = le16_to_cpu(fe->i_mode);
inode->i_uid = le32_to_cpu(fe->i_uid);
inode->i_gid = le32_to_cpu(fe->i_gid);
/* Fast symlinks will have i_size but no allocated clusters. */
if (S_ISLNK(inode->i_mode) && !fe->i_clusters)
inode->i_blocks = 0;
else
inode->i_blocks = ocfs2_inode_sector_count(inode);
inode->i_mapping->a_ops = &ocfs2_aops;
inode->i_atime.tv_sec = le64_to_cpu(fe->i_atime);
inode->i_atime.tv_nsec = le32_to_cpu(fe->i_atime_nsec);
inode->i_mtime.tv_sec = le64_to_cpu(fe->i_mtime);
inode->i_mtime.tv_nsec = le32_to_cpu(fe->i_mtime_nsec);
inode->i_ctime.tv_sec = le64_to_cpu(fe->i_ctime);
inode->i_ctime.tv_nsec = le32_to_cpu(fe->i_ctime_nsec);
if (OCFS2_I(inode)->ip_blkno != le64_to_cpu(fe->i_blkno))
mlog(ML_ERROR,
"ip_blkno %llu != i_blkno %llu!\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno,
(unsigned long long)le64_to_cpu(fe->i_blkno));
inode->i_nlink = ocfs2_read_links_count(fe);
if (fe->i_flags & cpu_to_le32(OCFS2_SYSTEM_FL)) {
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_SYSTEM_FILE;
inode->i_flags |= S_NOQUOTA;
}
if (fe->i_flags & cpu_to_le32(OCFS2_LOCAL_ALLOC_FL)) {
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_BITMAP;
mlog(0, "local alloc inode: i_ino=%lu\n", inode->i_ino);
} else if (fe->i_flags & cpu_to_le32(OCFS2_BITMAP_FL)) {
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_BITMAP;
} else if (fe->i_flags & cpu_to_le32(OCFS2_QUOTA_FL)) {
inode->i_flags |= S_NOQUOTA;
} else if (fe->i_flags & cpu_to_le32(OCFS2_SUPER_BLOCK_FL)) {
mlog(0, "superblock inode: i_ino=%lu\n", inode->i_ino);
/* we can't actually hit this as read_inode can't
* handle superblocks today ;-) */
BUG();
}
switch (inode->i_mode & S_IFMT) {
case S_IFREG:
if (use_plocks)
inode->i_fop = &ocfs2_fops;
else
inode->i_fop = &ocfs2_fops_no_plocks;
inode->i_op = &ocfs2_file_iops;
i_size_write(inode, le64_to_cpu(fe->i_size));
break;
case S_IFDIR:
inode->i_op = &ocfs2_dir_iops;
if (use_plocks)
inode->i_fop = &ocfs2_dops;
else
inode->i_fop = &ocfs2_dops_no_plocks;
i_size_write(inode, le64_to_cpu(fe->i_size));
break;
case S_IFLNK:
if (ocfs2_inode_is_fast_symlink(inode))
inode->i_op = &ocfs2_fast_symlink_inode_operations;
else
inode->i_op = &ocfs2_symlink_inode_operations;
i_size_write(inode, le64_to_cpu(fe->i_size));
break;
default:
inode->i_op = &ocfs2_special_file_iops;
init_special_inode(inode, inode->i_mode,
inode->i_rdev);
break;
}
if (create_ino) {
inode->i_ino = ino_from_blkno(inode->i_sb,
le64_to_cpu(fe->i_blkno));
/*
* If we ever want to create system files from kernel,
* the generation argument to
* ocfs2_inode_lock_res_init() will have to change.
*/
BUG_ON(le32_to_cpu(fe->i_flags) & OCFS2_SYSTEM_FL);
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_inode_lockres,
OCFS2_LOCK_TYPE_META, 0, inode);
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_open_lockres,
OCFS2_LOCK_TYPE_OPEN, 0, inode);
}
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_rw_lockres,
OCFS2_LOCK_TYPE_RW, inode->i_generation,
inode);
ocfs2_set_inode_flags(inode);
OCFS2_I(inode)->ip_last_used_slot = 0;
OCFS2_I(inode)->ip_last_used_group = 0;
mlog_exit_void();
}
static int ocfs2_read_locked_inode(struct inode *inode,
struct ocfs2_find_inode_args *args)
{
struct super_block *sb;
struct ocfs2_super *osb;
struct ocfs2_dinode *fe;
struct buffer_head *bh = NULL;
int status, can_lock;
u32 generation = 0;
mlog_entry("(0x%p, 0x%p)\n", inode, args);
status = -EINVAL;
if (inode == NULL || inode->i_sb == NULL) {
mlog(ML_ERROR, "bad inode\n");
return status;
}
sb = inode->i_sb;
osb = OCFS2_SB(sb);
if (!args) {
mlog(ML_ERROR, "bad inode args\n");
make_bad_inode(inode);
return status;
}
/*
* To improve performance of cold-cache inode stats, we take
* the cluster lock here if possible.
*
* Generally, OCFS2 never trusts the contents of an inode
* unless it's holding a cluster lock, so taking it here isn't
* a correctness issue as much as it is a performance
* improvement.
*
* There are three times when taking the lock is not a good idea:
*
* 1) During startup, before we have initialized the DLM.
*
* 2) If we are reading certain system files which never get
* cluster locks (local alloc, truncate log).
*
* 3) If the process doing the iget() is responsible for
* orphan dir recovery. We're holding the orphan dir lock and
* can get into a deadlock with another process on another
* node in ->delete_inode().
*
* #1 and #2 can be simply solved by never taking the lock
* here for system files (which are the only type we read
* during mount). It's a heavier approach, but our main
* concern is user-accesible files anyway.
*
* #3 works itself out because we'll eventually take the
* cluster lock before trusting anything anyway.
*/
can_lock = !(args->fi_flags & OCFS2_FI_FLAG_SYSFILE)
&& !(args->fi_flags & OCFS2_FI_FLAG_ORPHAN_RECOVERY)
&& !ocfs2_mount_local(osb);
/*
* To maintain backwards compatibility with older versions of
* ocfs2-tools, we still store the generation value for system
* files. The only ones that actually matter to userspace are
* the journals, but it's easier and inexpensive to just flag
* all system files similarly.
*/
if (args->fi_flags & OCFS2_FI_FLAG_SYSFILE)
generation = osb->fs_generation;
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_inode_lockres,
OCFS2_LOCK_TYPE_META,
generation, inode);
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_open_lockres,
OCFS2_LOCK_TYPE_OPEN,
0, inode);
if (can_lock) {
status = ocfs2_open_lock(inode);
if (status) {
make_bad_inode(inode);
mlog_errno(status);
return status;
}
status = ocfs2_inode_lock(inode, NULL, 0);
if (status) {
make_bad_inode(inode);
mlog_errno(status);
return status;
}
}
if (args->fi_flags & OCFS2_FI_FLAG_ORPHAN_RECOVERY) {
status = ocfs2_try_open_lock(inode, 0);
if (status) {
make_bad_inode(inode);
return status;
}
}
if (can_lock) {
status = ocfs2_read_inode_block_full(inode, &bh,
OCFS2_BH_IGNORE_CACHE);
} else {
status = ocfs2_read_blocks_sync(osb, args->fi_blkno, 1, &bh);
if (!status)
status = ocfs2_validate_inode_block(osb->sb, bh);
}
if (status < 0) {
mlog_errno(status);
goto bail;
}
status = -EINVAL;
fe = (struct ocfs2_dinode *) bh->b_data;
/*
* This is a code bug. Right now the caller needs to
* understand whether it is asking for a system file inode or
* not so the proper lock names can be built.
*/
mlog_bug_on_msg(!!(fe->i_flags & cpu_to_le32(OCFS2_SYSTEM_FL)) !=
!!(args->fi_flags & OCFS2_FI_FLAG_SYSFILE),
"Inode %llu: system file state is ambigous\n",
(unsigned long long)args->fi_blkno);
if (S_ISCHR(le16_to_cpu(fe->i_mode)) ||
S_ISBLK(le16_to_cpu(fe->i_mode)))
inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
ocfs2_populate_inode(inode, fe, 0);
BUG_ON(args->fi_blkno != le64_to_cpu(fe->i_blkno));
status = 0;
bail:
if (can_lock)
ocfs2_inode_unlock(inode, 0);
if (status < 0)
make_bad_inode(inode);
if (args && bh)
brelse(bh);
mlog_exit(status);
return status;
}
void ocfs2_sync_blockdev(struct super_block *sb)
{
sync_blockdev(sb->s_bdev);
}
static int ocfs2_truncate_for_delete(struct ocfs2_super *osb,
struct inode *inode,
struct buffer_head *fe_bh)
{
int status = 0;
struct ocfs2_truncate_context *tc = NULL;
struct ocfs2_dinode *fe;
handle_t *handle = NULL;
mlog_entry_void();
fe = (struct ocfs2_dinode *) fe_bh->b_data;
/*
* This check will also skip truncate of inodes with inline
* data and fast symlinks.
*/
if (fe->i_clusters) {
if (ocfs2_should_order_data(inode))
ocfs2_begin_ordered_truncate(inode, 0);
handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
if (IS_ERR(handle)) {
status = PTR_ERR(handle);
mlog_errno(status);
goto out;
}
status = ocfs2_journal_access_di(handle, INODE_CACHE(inode),
fe_bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
mlog_errno(status);
goto out;
}
i_size_write(inode, 0);
status = ocfs2_mark_inode_dirty(handle, inode, fe_bh);
if (status < 0) {
mlog_errno(status);
goto out;
}
ocfs2_commit_trans(osb, handle);
handle = NULL;
status = ocfs2_prepare_truncate(osb, inode, fe_bh, &tc);
if (status < 0) {
mlog_errno(status);
goto out;
}
status = ocfs2_commit_truncate(osb, inode, fe_bh, tc);
if (status < 0) {
mlog_errno(status);
goto out;
}
}
out:
if (handle)
ocfs2_commit_trans(osb, handle);
mlog_exit(status);
return status;
}
static int ocfs2_remove_inode(struct inode *inode,
struct buffer_head *di_bh,
struct inode *orphan_dir_inode,
struct buffer_head *orphan_dir_bh)
{
int status;
struct inode *inode_alloc_inode = NULL;
struct buffer_head *inode_alloc_bh = NULL;
handle_t *handle;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct ocfs2_dinode *di = (struct ocfs2_dinode *) di_bh->b_data;
inode_alloc_inode =
ocfs2_get_system_file_inode(osb, INODE_ALLOC_SYSTEM_INODE,
le16_to_cpu(di->i_suballoc_slot));
if (!inode_alloc_inode) {
status = -EEXIST;
mlog_errno(status);
goto bail;
}
mutex_lock(&inode_alloc_inode->i_mutex);
status = ocfs2_inode_lock(inode_alloc_inode, &inode_alloc_bh, 1);
if (status < 0) {
mutex_unlock(&inode_alloc_inode->i_mutex);
mlog_errno(status);
goto bail;
}
handle = ocfs2_start_trans(osb, OCFS2_DELETE_INODE_CREDITS +
ocfs2_quota_trans_credits(inode->i_sb));
if (IS_ERR(handle)) {
status = PTR_ERR(handle);
mlog_errno(status);
goto bail_unlock;
}
status = ocfs2_orphan_del(osb, handle, orphan_dir_inode, inode,
orphan_dir_bh);
if (status < 0) {
mlog_errno(status);
goto bail_commit;
}
/* set the inodes dtime */
status = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
mlog_errno(status);
goto bail_commit;
}
di->i_dtime = cpu_to_le64(CURRENT_TIME.tv_sec);
di->i_flags &= cpu_to_le32(~(OCFS2_VALID_FL | OCFS2_ORPHANED_FL));
status = ocfs2_journal_dirty(handle, di_bh);
if (status < 0) {
mlog_errno(status);
goto bail_commit;
}
ocfs2_remove_from_cache(INODE_CACHE(inode), di_bh);
dquot_free_inode(inode);
status = ocfs2_free_dinode(handle, inode_alloc_inode,
inode_alloc_bh, di);
if (status < 0)
mlog_errno(status);
bail_commit:
ocfs2_commit_trans(osb, handle);
bail_unlock:
ocfs2_inode_unlock(inode_alloc_inode, 1);
mutex_unlock(&inode_alloc_inode->i_mutex);
brelse(inode_alloc_bh);
bail:
iput(inode_alloc_inode);
return status;
}
/*
* Serialize with orphan dir recovery. If the process doing
* recovery on this orphan dir does an iget() with the dir
* i_mutex held, we'll deadlock here. Instead we detect this
* and exit early - recovery will wipe this inode for us.
*/
static int ocfs2_check_orphan_recovery_state(struct ocfs2_super *osb,
int slot)
{
int ret = 0;
spin_lock(&osb->osb_lock);
if (ocfs2_node_map_test_bit(osb, &osb->osb_recovering_orphan_dirs, slot)) {
mlog(0, "Recovery is happening on orphan dir %d, will skip "
"this inode\n", slot);
ret = -EDEADLK;
goto out;
}
/* This signals to the orphan recovery process that it should
* wait for us to handle the wipe. */
osb->osb_orphan_wipes[slot]++;
out:
spin_unlock(&osb->osb_lock);
return ret;
}
static void ocfs2_signal_wipe_completion(struct ocfs2_super *osb,
int slot)
{
spin_lock(&osb->osb_lock);
osb->osb_orphan_wipes[slot]--;
spin_unlock(&osb->osb_lock);
wake_up(&osb->osb_wipe_event);
}
static int ocfs2_wipe_inode(struct inode *inode,
struct buffer_head *di_bh)
{
int status, orphaned_slot;
struct inode *orphan_dir_inode = NULL;
struct buffer_head *orphan_dir_bh = NULL;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct ocfs2_dinode *di;
di = (struct ocfs2_dinode *) di_bh->b_data;
orphaned_slot = le16_to_cpu(di->i_orphaned_slot);
status = ocfs2_check_orphan_recovery_state(osb, orphaned_slot);
if (status)
return status;
orphan_dir_inode = ocfs2_get_system_file_inode(osb,
ORPHAN_DIR_SYSTEM_INODE,
orphaned_slot);
if (!orphan_dir_inode) {
status = -EEXIST;
mlog_errno(status);
goto bail;
}
/* Lock the orphan dir. The lock will be held for the entire
* delete_inode operation. We do this now to avoid races with
* recovery completion on other nodes. */
mutex_lock(&orphan_dir_inode->i_mutex);
status = ocfs2_inode_lock(orphan_dir_inode, &orphan_dir_bh, 1);
if (status < 0) {
mutex_unlock(&orphan_dir_inode->i_mutex);
mlog_errno(status);
goto bail;
}
/* we do this while holding the orphan dir lock because we
* don't want recovery being run from another node to try an
* inode delete underneath us -- this will result in two nodes
* truncating the same file! */
status = ocfs2_truncate_for_delete(osb, inode, di_bh);
if (status < 0) {
mlog_errno(status);
goto bail_unlock_dir;
}
/* Remove any dir index tree */
if (S_ISDIR(inode->i_mode)) {
status = ocfs2_dx_dir_truncate(inode, di_bh);
if (status) {
mlog_errno(status);
goto bail_unlock_dir;
}
}
/*Free extended attribute resources associated with this inode.*/
status = ocfs2_xattr_remove(inode, di_bh);
if (status < 0) {
mlog_errno(status);
goto bail_unlock_dir;
}
status = ocfs2_remove_refcount_tree(inode, di_bh);
if (status < 0) {
mlog_errno(status);
goto bail_unlock_dir;
}
status = ocfs2_remove_inode(inode, di_bh, orphan_dir_inode,
orphan_dir_bh);
if (status < 0)
mlog_errno(status);
bail_unlock_dir:
ocfs2_inode_unlock(orphan_dir_inode, 1);
mutex_unlock(&orphan_dir_inode->i_mutex);
brelse(orphan_dir_bh);
bail:
iput(orphan_dir_inode);
ocfs2_signal_wipe_completion(osb, orphaned_slot);
return status;
}
/* There is a series of simple checks that should be done before a
* trylock is even considered. Encapsulate those in this function. */
static int ocfs2_inode_is_valid_to_delete(struct inode *inode)
{
int ret = 0;
struct ocfs2_inode_info *oi = OCFS2_I(inode);
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
/* We shouldn't be getting here for the root directory
* inode.. */
if (inode == osb->root_inode) {
mlog(ML_ERROR, "Skipping delete of root inode.\n");
goto bail;
}
/* If we're coming from downconvert_thread we can't go into our own
* voting [hello, deadlock city!], so unforuntately we just
* have to skip deleting this guy. That's OK though because
* the node who's doing the actual deleting should handle it
* anyway. */
if (current == osb->dc_task) {
mlog(0, "Skipping delete of %lu because we're currently "
"in downconvert\n", inode->i_ino);
goto bail;
}
spin_lock(&oi->ip_lock);
/* OCFS2 *never* deletes system files. This should technically
* never get here as system file inodes should always have a
* positive link count. */
if (oi->ip_flags & OCFS2_INODE_SYSTEM_FILE) {
mlog(ML_ERROR, "Skipping delete of system file %llu\n",
(unsigned long long)oi->ip_blkno);
goto bail_unlock;
}
/* If we have allowd wipe of this inode for another node, it
* will be marked here so we can safely skip it. Recovery will
* cleanup any inodes we might inadvertantly skip here. */
if (oi->ip_flags & OCFS2_INODE_SKIP_DELETE) {
mlog(0, "Skipping delete of %lu because another node "
"has done this for us.\n", inode->i_ino);
goto bail_unlock;
}
ret = 1;
bail_unlock:
spin_unlock(&oi->ip_lock);
bail:
return ret;
}
/* Query the cluster to determine whether we should wipe an inode from
* disk or not.
*
* Requires the inode to have the cluster lock. */
static int ocfs2_query_inode_wipe(struct inode *inode,
struct buffer_head *di_bh,
int *wipe)
{
int status = 0;
struct ocfs2_inode_info *oi = OCFS2_I(inode);
struct ocfs2_dinode *di;
*wipe = 0;
/* While we were waiting for the cluster lock in
* ocfs2_delete_inode, another node might have asked to delete
* the inode. Recheck our flags to catch this. */
if (!ocfs2_inode_is_valid_to_delete(inode)) {
mlog(0, "Skipping delete of %llu because flags changed\n",
(unsigned long long)oi->ip_blkno);
goto bail;
}
/* Now that we have an up to date inode, we can double check
* the link count. */
if (inode->i_nlink) {
mlog(0, "Skipping delete of %llu because nlink = %u\n",
(unsigned long long)oi->ip_blkno, inode->i_nlink);
goto bail;
}
/* Do some basic inode verification... */
di = (struct ocfs2_dinode *) di_bh->b_data;
if (!(di->i_flags & cpu_to_le32(OCFS2_ORPHANED_FL))) {
/* for lack of a better error? */
status = -EEXIST;
mlog(ML_ERROR,
"Inode %llu (on-disk %llu) not orphaned! "
"Disk flags 0x%x, inode flags 0x%x\n",
(unsigned long long)oi->ip_blkno,
(unsigned long long)le64_to_cpu(di->i_blkno),
le32_to_cpu(di->i_flags), oi->ip_flags);
goto bail;
}
/* has someone already deleted us?! baaad... */
if (di->i_dtime) {
status = -EEXIST;
mlog_errno(status);
goto bail;
}
/*
* This is how ocfs2 determines whether an inode is still live
* within the cluster. Every node takes a shared read lock on
* the inode open lock in ocfs2_read_locked_inode(). When we
* get to ->delete_inode(), each node tries to convert it's
* lock to an exclusive. Trylocks are serialized by the inode
* meta data lock. If the upconvert suceeds, we know the inode
* is no longer live and can be deleted.
*
* Though we call this with the meta data lock held, the
* trylock keeps us from ABBA deadlock.
*/
status = ocfs2_try_open_lock(inode, 1);
if (status == -EAGAIN) {
status = 0;
mlog(0, "Skipping delete of %llu because it is in use on "
"other nodes\n", (unsigned long long)oi->ip_blkno);
goto bail;
}
if (status < 0) {
mlog_errno(status);
goto bail;
}
*wipe = 1;
mlog(0, "Inode %llu is ok to wipe from orphan dir %u\n",
(unsigned long long)oi->ip_blkno,
le16_to_cpu(di->i_orphaned_slot));
bail:
return status;
}
/* Support function for ocfs2_delete_inode. Will help us keep the
* inode data in a consistent state for clear_inode. Always truncates
* pages, optionally sync's them first. */
static void ocfs2_cleanup_delete_inode(struct inode *inode,
int sync_data)
{
mlog(0, "Cleanup inode %llu, sync = %d\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno, sync_data);
if (sync_data)
write_inode_now(inode, 1);
truncate_inode_pages(&inode->i_data, 0);
}
void ocfs2_delete_inode(struct inode *inode)
{
int wipe, status;
sigset_t blocked, oldset;
struct buffer_head *di_bh = NULL;
mlog_entry("(inode->i_ino = %lu)\n", inode->i_ino);
/* When we fail in read_inode() we mark inode as bad. The second test
* catches the case when inode allocation fails before allocating
* a block for inode. */
if (is_bad_inode(inode) || !OCFS2_I(inode)->ip_blkno) {
mlog(0, "Skipping delete of bad inode\n");
goto bail;
}
if (!ocfs2_inode_is_valid_to_delete(inode)) {
/* It's probably not necessary to truncate_inode_pages
* here but we do it for safety anyway (it will most
* likely be a no-op anyway) */
ocfs2_cleanup_delete_inode(inode, 0);
goto bail;
}
/* We want to block signals in delete_inode as the lock and
* messaging paths may return us -ERESTARTSYS. Which would
* cause us to exit early, resulting in inodes being orphaned
* forever. */
sigfillset(&blocked);
status = sigprocmask(SIG_BLOCK, &blocked, &oldset);
if (status < 0) {
mlog_errno(status);
ocfs2_cleanup_delete_inode(inode, 1);
goto bail;
}
/*
* Synchronize us against ocfs2_get_dentry. We take this in
* shared mode so that all nodes can still concurrently
* process deletes.
*/
status = ocfs2_nfs_sync_lock(OCFS2_SB(inode->i_sb), 0);
if (status < 0) {