forked from torvalds/linux
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathjournal.c
2056 lines (1807 loc) · 57.8 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
// SPDX-License-Identifier: GPL-2.0-only
/*
* This file is part of UBIFS.
*
* Copyright (C) 2006-2008 Nokia Corporation.
*
* Authors: Artem Bityutskiy (Битюцкий Артём)
* Adrian Hunter
*/
/*
* This file implements UBIFS journal.
*
* The journal consists of 2 parts - the log and bud LEBs. The log has fixed
* length and position, while a bud logical eraseblock is any LEB in the main
* area. Buds contain file system data - data nodes, inode nodes, etc. The log
* contains only references to buds and some other stuff like commit
* start node. The idea is that when we commit the journal, we do
* not copy the data, the buds just become indexed. Since after the commit the
* nodes in bud eraseblocks become leaf nodes of the file system index tree, we
* use term "bud". Analogy is obvious, bud eraseblocks contain nodes which will
* become leafs in the future.
*
* The journal is multi-headed because we want to write data to the journal as
* optimally as possible. It is nice to have nodes belonging to the same inode
* in one LEB, so we may write data owned by different inodes to different
* journal heads, although at present only one data head is used.
*
* For recovery reasons, the base head contains all inode nodes, all directory
* entry nodes and all truncate nodes. This means that the other heads contain
* only data nodes.
*
* Bud LEBs may be half-indexed. For example, if the bud was not full at the
* time of commit, the bud is retained to continue to be used in the journal,
* even though the "front" of the LEB is now indexed. In that case, the log
* reference contains the offset where the bud starts for the purposes of the
* journal.
*
* The journal size has to be limited, because the larger is the journal, the
* longer it takes to mount UBIFS (scanning the journal) and the more memory it
* takes (indexing in the TNC).
*
* All the journal write operations like 'ubifs_jnl_update()' here, which write
* multiple UBIFS nodes to the journal at one go, are atomic with respect to
* unclean reboots. Should the unclean reboot happen, the recovery code drops
* all the nodes.
*/
#include "ubifs.h"
/**
* zero_ino_node_unused - zero out unused fields of an on-flash inode node.
* @ino: the inode to zero out
*/
static inline void zero_ino_node_unused(struct ubifs_ino_node *ino)
{
memset(ino->padding1, 0, 4);
memset(ino->padding2, 0, 26);
}
/**
* zero_dent_node_unused - zero out unused fields of an on-flash directory
* entry node.
* @dent: the directory entry to zero out
*/
static inline void zero_dent_node_unused(struct ubifs_dent_node *dent)
{
dent->padding1 = 0;
}
/**
* zero_trun_node_unused - zero out unused fields of an on-flash truncation
* node.
* @trun: the truncation node to zero out
*/
static inline void zero_trun_node_unused(struct ubifs_trun_node *trun)
{
memset(trun->padding, 0, 12);
}
static void ubifs_add_auth_dirt(struct ubifs_info *c, int lnum)
{
if (ubifs_authenticated(c))
ubifs_add_dirt(c, lnum, ubifs_auth_node_sz(c));
}
/**
* reserve_space - reserve space in the journal.
* @c: UBIFS file-system description object
* @jhead: journal head number
* @len: node length
*
* This function reserves space in journal head @head. If the reservation
* succeeded, the journal head stays locked and later has to be unlocked using
* 'release_head()'. Returns zero in case of success, %-EAGAIN if commit has to
* be done, and other negative error codes in case of other failures.
*/
static int reserve_space(struct ubifs_info *c, int jhead, int len)
{
int err = 0, err1, retries = 0, avail, lnum, offs, squeeze;
struct ubifs_wbuf *wbuf = &c->jheads[jhead].wbuf;
/*
* Typically, the base head has smaller nodes written to it, so it is
* better to try to allocate space at the ends of eraseblocks. This is
* what the squeeze parameter does.
*/
ubifs_assert(c, !c->ro_media && !c->ro_mount);
squeeze = (jhead == BASEHD);
again:
mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
if (c->ro_error) {
err = -EROFS;
goto out_unlock;
}
avail = c->leb_size - wbuf->offs - wbuf->used;
if (wbuf->lnum != -1 && avail >= len)
return 0;
/*
* Write buffer wasn't seek'ed or there is no enough space - look for an
* LEB with some empty space.
*/
lnum = ubifs_find_free_space(c, len, &offs, squeeze);
if (lnum >= 0)
goto out;
err = lnum;
if (err != -ENOSPC)
goto out_unlock;
/*
* No free space, we have to run garbage collector to make
* some. But the write-buffer mutex has to be unlocked because
* GC also takes it.
*/
dbg_jnl("no free space in jhead %s, run GC", dbg_jhead(jhead));
mutex_unlock(&wbuf->io_mutex);
lnum = ubifs_garbage_collect(c, 0);
if (lnum < 0) {
err = lnum;
if (err != -ENOSPC)
return err;
/*
* GC could not make a free LEB. But someone else may
* have allocated new bud for this journal head,
* because we dropped @wbuf->io_mutex, so try once
* again.
*/
dbg_jnl("GC couldn't make a free LEB for jhead %s",
dbg_jhead(jhead));
if (retries++ < 2) {
dbg_jnl("retry (%d)", retries);
goto again;
}
dbg_jnl("return -ENOSPC");
return err;
}
mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
dbg_jnl("got LEB %d for jhead %s", lnum, dbg_jhead(jhead));
avail = c->leb_size - wbuf->offs - wbuf->used;
if (wbuf->lnum != -1 && avail >= len) {
/*
* Someone else has switched the journal head and we have
* enough space now. This happens when more than one process is
* trying to write to the same journal head at the same time.
*/
dbg_jnl("return LEB %d back, already have LEB %d:%d",
lnum, wbuf->lnum, wbuf->offs + wbuf->used);
err = ubifs_return_leb(c, lnum);
if (err)
goto out_unlock;
return 0;
}
offs = 0;
out:
/*
* Make sure we synchronize the write-buffer before we add the new bud
* to the log. Otherwise we may have a power cut after the log
* reference node for the last bud (@lnum) is written but before the
* write-buffer data are written to the next-to-last bud
* (@wbuf->lnum). And the effect would be that the recovery would see
* that there is corruption in the next-to-last bud.
*/
err = ubifs_wbuf_sync_nolock(wbuf);
if (err)
goto out_return;
err = ubifs_add_bud_to_log(c, jhead, lnum, offs);
if (err)
goto out_return;
err = ubifs_wbuf_seek_nolock(wbuf, lnum, offs);
if (err)
goto out_unlock;
return 0;
out_unlock:
mutex_unlock(&wbuf->io_mutex);
return err;
out_return:
/* An error occurred and the LEB has to be returned to lprops */
ubifs_assert(c, err < 0);
err1 = ubifs_return_leb(c, lnum);
if (err1 && err == -EAGAIN)
/*
* Return original error code only if it is not %-EAGAIN,
* which is not really an error. Otherwise, return the error
* code of 'ubifs_return_leb()'.
*/
err = err1;
mutex_unlock(&wbuf->io_mutex);
return err;
}
static int ubifs_hash_nodes(struct ubifs_info *c, void *node,
int len, struct shash_desc *hash)
{
int auth_node_size = ubifs_auth_node_sz(c);
int err;
while (1) {
const struct ubifs_ch *ch = node;
int nodelen = le32_to_cpu(ch->len);
ubifs_assert(c, len >= auth_node_size);
if (len == auth_node_size)
break;
ubifs_assert(c, len > nodelen);
ubifs_assert(c, ch->magic == cpu_to_le32(UBIFS_NODE_MAGIC));
err = ubifs_shash_update(c, hash, (void *)node, nodelen);
if (err)
return err;
node += ALIGN(nodelen, 8);
len -= ALIGN(nodelen, 8);
}
return ubifs_prepare_auth_node(c, node, hash);
}
/**
* write_head - write data to a journal head.
* @c: UBIFS file-system description object
* @jhead: journal head
* @buf: buffer to write
* @len: length to write
* @lnum: LEB number written is returned here
* @offs: offset written is returned here
* @sync: non-zero if the write-buffer has to by synchronized
*
* This function writes data to the reserved space of journal head @jhead.
* Returns zero in case of success and a negative error code in case of
* failure.
*/
static int write_head(struct ubifs_info *c, int jhead, void *buf, int len,
int *lnum, int *offs, int sync)
{
int err;
struct ubifs_wbuf *wbuf = &c->jheads[jhead].wbuf;
ubifs_assert(c, jhead != GCHD);
*lnum = c->jheads[jhead].wbuf.lnum;
*offs = c->jheads[jhead].wbuf.offs + c->jheads[jhead].wbuf.used;
dbg_jnl("jhead %s, LEB %d:%d, len %d",
dbg_jhead(jhead), *lnum, *offs, len);
if (ubifs_authenticated(c)) {
err = ubifs_hash_nodes(c, buf, len, c->jheads[jhead].log_hash);
if (err)
return err;
}
err = ubifs_wbuf_write_nolock(wbuf, buf, len);
if (err)
return err;
if (sync)
err = ubifs_wbuf_sync_nolock(wbuf);
return err;
}
/**
* __queue_and_wait - queue a task and wait until the task is waked up.
* @c: UBIFS file-system description object
*
* This function adds current task in queue and waits until the task is waked
* up. This function should be called with @c->reserve_space_wq locked.
*/
static void __queue_and_wait(struct ubifs_info *c)
{
DEFINE_WAIT(wait);
__add_wait_queue_entry_tail_exclusive(&c->reserve_space_wq, &wait);
set_current_state(TASK_UNINTERRUPTIBLE);
spin_unlock(&c->reserve_space_wq.lock);
schedule();
finish_wait(&c->reserve_space_wq, &wait);
}
/**
* wait_for_reservation - try queuing current task to wait until waked up.
* @c: UBIFS file-system description object
*
* This function queues current task to wait until waked up, if queuing is
* started(@c->need_wait_space is not %0). Returns %true if current task is
* added in queue, otherwise %false is returned.
*/
static bool wait_for_reservation(struct ubifs_info *c)
{
if (likely(atomic_read(&c->need_wait_space) == 0))
/* Quick path to check whether queuing is started. */
return false;
spin_lock(&c->reserve_space_wq.lock);
if (atomic_read(&c->need_wait_space) == 0) {
/* Queuing is not started, don't queue current task. */
spin_unlock(&c->reserve_space_wq.lock);
return false;
}
__queue_and_wait(c);
return true;
}
/**
* wake_up_reservation - wake up first task in queue or stop queuing.
* @c: UBIFS file-system description object
*
* This function wakes up the first task in queue if it exists, or stops
* queuing if no tasks in queue.
*/
static void wake_up_reservation(struct ubifs_info *c)
{
spin_lock(&c->reserve_space_wq.lock);
if (waitqueue_active(&c->reserve_space_wq))
wake_up_locked(&c->reserve_space_wq);
else
/*
* Compared with wait_for_reservation(), set @c->need_wait_space
* under the protection of wait queue lock, which can avoid that
* @c->need_wait_space is set to 0 after new task queued.
*/
atomic_set(&c->need_wait_space, 0);
spin_unlock(&c->reserve_space_wq.lock);
}
/**
* add_or_start_queue - add current task in queue or start queuing.
* @c: UBIFS file-system description object
*
* This function starts queuing if queuing is not started, otherwise adds
* current task in queue.
*/
static void add_or_start_queue(struct ubifs_info *c)
{
spin_lock(&c->reserve_space_wq.lock);
if (atomic_cmpxchg(&c->need_wait_space, 0, 1) == 0) {
/* Starts queuing, task can go on directly. */
spin_unlock(&c->reserve_space_wq.lock);
return;
}
/*
* There are at least two tasks have retried more than 32 times
* at certain point, first task has started queuing, just queue
* the left tasks.
*/
__queue_and_wait(c);
}
/**
* make_reservation - reserve journal space.
* @c: UBIFS file-system description object
* @jhead: journal head
* @len: how many bytes to reserve
*
* This function makes space reservation in journal head @jhead. The function
* takes the commit lock and locks the journal head, and the caller has to
* unlock the head and finish the reservation with 'finish_reservation()'.
* Returns zero in case of success and a negative error code in case of
* failure.
*
* Note, the journal head may be unlocked as soon as the data is written, while
* the commit lock has to be released after the data has been added to the
* TNC.
*/
static int make_reservation(struct ubifs_info *c, int jhead, int len)
{
int err, cmt_retries = 0, nospc_retries = 0;
bool blocked = wait_for_reservation(c);
again:
down_read(&c->commit_sem);
err = reserve_space(c, jhead, len);
if (!err) {
/* c->commit_sem will get released via finish_reservation(). */
goto out_wake_up;
}
up_read(&c->commit_sem);
if (err == -ENOSPC) {
/*
* GC could not make any progress. We should try to commit
* because it could make some dirty space and GC would make
* progress, so make the error -EAGAIN so that the below
* will commit and re-try.
*/
nospc_retries++;
dbg_jnl("no space, retry");
err = -EAGAIN;
}
if (err != -EAGAIN)
goto out;
/*
* -EAGAIN means that the journal is full or too large, or the above
* code wants to do one commit. Do this and re-try.
*/
if (cmt_retries > 128) {
/*
* This should not happen unless:
* 1. The journal size limitations are too tough.
* 2. The budgeting is incorrect. We always have to be able to
* write to the media, because all operations are budgeted.
* Deletions are not budgeted, though, but we reserve an
* extra LEB for them.
*/
ubifs_err(c, "stuck in space allocation, nospc_retries %d",
nospc_retries);
err = -ENOSPC;
goto out;
} else if (cmt_retries > 32) {
/*
* It's almost impossible to happen, unless there are many tasks
* making reservation concurrently and someone task has retried
* gc + commit for many times, generated available space during
* this period are grabbed by other tasks.
* But if it happens, start queuing up all tasks that will make
* space reservation, then there is only one task making space
* reservation at any time, and it can always make success under
* the premise of correct budgeting.
*/
ubifs_warn(c, "too many space allocation cmt_retries (%d) "
"nospc_retries (%d), start queuing tasks",
cmt_retries, nospc_retries);
if (!blocked) {
blocked = true;
add_or_start_queue(c);
}
}
dbg_jnl("-EAGAIN, commit and retry (retried %d times)",
cmt_retries);
cmt_retries += 1;
err = ubifs_run_commit(c);
if (err)
goto out_wake_up;
goto again;
out:
ubifs_err(c, "cannot reserve %d bytes in jhead %d, error %d",
len, jhead, err);
if (err == -ENOSPC) {
/* This are some budgeting problems, print useful information */
down_write(&c->commit_sem);
dump_stack();
ubifs_dump_budg(c, &c->bi);
ubifs_dump_lprops(c);
cmt_retries = dbg_check_lprops(c);
up_write(&c->commit_sem);
}
out_wake_up:
if (blocked) {
/*
* Only tasks that have ever started queuing or ever been queued
* can wake up other queued tasks, which can make sure that
* there is only one task waked up to make space reservation.
* For example:
* task A task B task C
* make_reservation make_reservation
* reserve_space // 0
* wake_up_reservation
* atomic_cmpxchg // 0, start queuing
* reserve_space
* wait_for_reservation
* __queue_and_wait
* add_wait_queue
* if (blocked) // false
* // So that task C won't be waked up to race with task B
*/
wake_up_reservation(c);
}
return err;
}
/**
* release_head - release a journal head.
* @c: UBIFS file-system description object
* @jhead: journal head
*
* This function releases journal head @jhead which was locked by
* the 'make_reservation()' function. It has to be called after each successful
* 'make_reservation()' invocation.
*/
static inline void release_head(struct ubifs_info *c, int jhead)
{
mutex_unlock(&c->jheads[jhead].wbuf.io_mutex);
}
/**
* finish_reservation - finish a reservation.
* @c: UBIFS file-system description object
*
* This function finishes journal space reservation. It must be called after
* 'make_reservation()'.
*/
static void finish_reservation(struct ubifs_info *c)
{
up_read(&c->commit_sem);
}
/**
* get_dent_type - translate VFS inode mode to UBIFS directory entry type.
* @mode: inode mode
*/
static int get_dent_type(int mode)
{
switch (mode & S_IFMT) {
case S_IFREG:
return UBIFS_ITYPE_REG;
case S_IFDIR:
return UBIFS_ITYPE_DIR;
case S_IFLNK:
return UBIFS_ITYPE_LNK;
case S_IFBLK:
return UBIFS_ITYPE_BLK;
case S_IFCHR:
return UBIFS_ITYPE_CHR;
case S_IFIFO:
return UBIFS_ITYPE_FIFO;
case S_IFSOCK:
return UBIFS_ITYPE_SOCK;
default:
BUG();
}
return 0;
}
/**
* pack_inode - pack an inode node.
* @c: UBIFS file-system description object
* @ino: buffer in which to pack inode node
* @inode: inode to pack
* @last: indicates the last node of the group
*/
static void pack_inode(struct ubifs_info *c, struct ubifs_ino_node *ino,
const struct inode *inode, int last)
{
int data_len = 0, last_reference = !inode->i_nlink;
struct ubifs_inode *ui = ubifs_inode(inode);
ino->ch.node_type = UBIFS_INO_NODE;
ino_key_init_flash(c, &ino->key, inode->i_ino);
ino->creat_sqnum = cpu_to_le64(ui->creat_sqnum);
ino->atime_sec = cpu_to_le64(inode_get_atime_sec(inode));
ino->atime_nsec = cpu_to_le32(inode_get_atime_nsec(inode));
ino->ctime_sec = cpu_to_le64(inode_get_ctime_sec(inode));
ino->ctime_nsec = cpu_to_le32(inode_get_ctime_nsec(inode));
ino->mtime_sec = cpu_to_le64(inode_get_mtime_sec(inode));
ino->mtime_nsec = cpu_to_le32(inode_get_mtime_nsec(inode));
ino->uid = cpu_to_le32(i_uid_read(inode));
ino->gid = cpu_to_le32(i_gid_read(inode));
ino->mode = cpu_to_le32(inode->i_mode);
ino->flags = cpu_to_le32(ui->flags);
ino->size = cpu_to_le64(ui->ui_size);
ino->nlink = cpu_to_le32(inode->i_nlink);
ino->compr_type = cpu_to_le16(ui->compr_type);
ino->data_len = cpu_to_le32(ui->data_len);
ino->xattr_cnt = cpu_to_le32(ui->xattr_cnt);
ino->xattr_size = cpu_to_le32(ui->xattr_size);
ino->xattr_names = cpu_to_le32(ui->xattr_names);
zero_ino_node_unused(ino);
/*
* Drop the attached data if this is a deletion inode, the data is not
* needed anymore.
*/
if (!last_reference) {
memcpy(ino->data, ui->data, ui->data_len);
data_len = ui->data_len;
}
ubifs_prep_grp_node(c, ino, UBIFS_INO_NODE_SZ + data_len, last);
}
/**
* mark_inode_clean - mark UBIFS inode as clean.
* @c: UBIFS file-system description object
* @ui: UBIFS inode to mark as clean
*
* This helper function marks UBIFS inode @ui as clean by cleaning the
* @ui->dirty flag and releasing its budget. Note, VFS may still treat the
* inode as dirty and try to write it back, but 'ubifs_write_inode()' would
* just do nothing.
*/
static void mark_inode_clean(struct ubifs_info *c, struct ubifs_inode *ui)
{
if (ui->dirty)
ubifs_release_dirty_inode_budget(c, ui);
ui->dirty = 0;
}
static void set_dent_cookie(struct ubifs_info *c, struct ubifs_dent_node *dent)
{
if (c->double_hash)
dent->cookie = (__force __le32) get_random_u32();
else
dent->cookie = 0;
}
/**
* ubifs_jnl_update - update inode.
* @c: UBIFS file-system description object
* @dir: parent inode or host inode in case of extended attributes
* @nm: directory entry name
* @inode: inode to update
* @deletion: indicates a directory entry deletion i.e unlink or rmdir
* @xent: non-zero if the directory entry is an extended attribute entry
* @in_orphan: indicates whether the @inode is in orphan list
*
* This function updates an inode by writing a directory entry (or extended
* attribute entry), the inode itself, and the parent directory inode (or the
* host inode) to the journal.
*
* The function writes the host inode @dir last, which is important in case of
* extended attributes. Indeed, then we guarantee that if the host inode gets
* synchronized (with 'fsync()'), and the write-buffer it sits in gets flushed,
* the extended attribute inode gets flushed too. And this is exactly what the
* user expects - synchronizing the host inode synchronizes its extended
* attributes. Similarly, this guarantees that if @dir is synchronized, its
* directory entry corresponding to @nm gets synchronized too.
*
* If the inode (@inode) or the parent directory (@dir) are synchronous, this
* function synchronizes the write-buffer.
*
* This function marks the @dir and @inode inodes as clean and returns zero on
* success. In case of failure, a negative error code is returned.
*/
int ubifs_jnl_update(struct ubifs_info *c, const struct inode *dir,
const struct fscrypt_name *nm, const struct inode *inode,
int deletion, int xent, int in_orphan)
{
int err, dlen, ilen, len, lnum, ino_offs, dent_offs, orphan_added = 0;
int aligned_dlen, aligned_ilen, sync = IS_DIRSYNC(dir);
int last_reference = !!(deletion && inode->i_nlink == 0);
struct ubifs_inode *ui = ubifs_inode(inode);
struct ubifs_inode *host_ui = ubifs_inode(dir);
struct ubifs_dent_node *dent;
struct ubifs_ino_node *ino;
union ubifs_key dent_key, ino_key;
u8 hash_dent[UBIFS_HASH_ARR_SZ];
u8 hash_ino[UBIFS_HASH_ARR_SZ];
u8 hash_ino_host[UBIFS_HASH_ARR_SZ];
ubifs_assert(c, mutex_is_locked(&host_ui->ui_mutex));
dlen = UBIFS_DENT_NODE_SZ + fname_len(nm) + 1;
ilen = UBIFS_INO_NODE_SZ;
/*
* If the last reference to the inode is being deleted, then there is
* no need to attach and write inode data, it is being deleted anyway.
* And if the inode is being deleted, no need to synchronize
* write-buffer even if the inode is synchronous.
*/
if (!last_reference) {
ilen += ui->data_len;
sync |= IS_SYNC(inode);
}
aligned_dlen = ALIGN(dlen, 8);
aligned_ilen = ALIGN(ilen, 8);
len = aligned_dlen + aligned_ilen + UBIFS_INO_NODE_SZ;
/* Make sure to also account for extended attributes */
if (ubifs_authenticated(c))
len += ALIGN(host_ui->data_len, 8) + ubifs_auth_node_sz(c);
else
len += host_ui->data_len;
dent = kzalloc(len, GFP_NOFS);
if (!dent)
return -ENOMEM;
/* Make reservation before allocating sequence numbers */
err = make_reservation(c, BASEHD, len);
if (err)
goto out_free;
if (!xent) {
dent->ch.node_type = UBIFS_DENT_NODE;
if (fname_name(nm) == NULL)
dent_key_init_hash(c, &dent_key, dir->i_ino, nm->hash);
else
dent_key_init(c, &dent_key, dir->i_ino, nm);
} else {
dent->ch.node_type = UBIFS_XENT_NODE;
xent_key_init(c, &dent_key, dir->i_ino, nm);
}
key_write(c, &dent_key, dent->key);
dent->inum = deletion ? 0 : cpu_to_le64(inode->i_ino);
dent->type = get_dent_type(inode->i_mode);
dent->nlen = cpu_to_le16(fname_len(nm));
memcpy(dent->name, fname_name(nm), fname_len(nm));
dent->name[fname_len(nm)] = '\0';
set_dent_cookie(c, dent);
zero_dent_node_unused(dent);
ubifs_prep_grp_node(c, dent, dlen, 0);
err = ubifs_node_calc_hash(c, dent, hash_dent);
if (err)
goto out_release;
ino = (void *)dent + aligned_dlen;
pack_inode(c, ino, inode, 0);
err = ubifs_node_calc_hash(c, ino, hash_ino);
if (err)
goto out_release;
ino = (void *)ino + aligned_ilen;
pack_inode(c, ino, dir, 1);
err = ubifs_node_calc_hash(c, ino, hash_ino_host);
if (err)
goto out_release;
if (last_reference && !in_orphan) {
err = ubifs_add_orphan(c, inode->i_ino);
if (err) {
release_head(c, BASEHD);
goto out_finish;
}
ui->del_cmtno = c->cmt_no;
orphan_added = 1;
}
err = write_head(c, BASEHD, dent, len, &lnum, &dent_offs, sync);
if (err)
goto out_release;
if (!sync) {
struct ubifs_wbuf *wbuf = &c->jheads[BASEHD].wbuf;
ubifs_wbuf_add_ino_nolock(wbuf, inode->i_ino);
ubifs_wbuf_add_ino_nolock(wbuf, dir->i_ino);
}
release_head(c, BASEHD);
kfree(dent);
ubifs_add_auth_dirt(c, lnum);
if (deletion) {
if (fname_name(nm) == NULL)
err = ubifs_tnc_remove_dh(c, &dent_key, nm->minor_hash);
else
err = ubifs_tnc_remove_nm(c, &dent_key, nm);
if (err)
goto out_ro;
err = ubifs_add_dirt(c, lnum, dlen);
} else
err = ubifs_tnc_add_nm(c, &dent_key, lnum, dent_offs, dlen,
hash_dent, nm);
if (err)
goto out_ro;
/*
* Note, we do not remove the inode from TNC even if the last reference
* to it has just been deleted, because the inode may still be opened.
* Instead, the inode has been added to orphan lists and the orphan
* subsystem will take further care about it.
*/
ino_key_init(c, &ino_key, inode->i_ino);
ino_offs = dent_offs + aligned_dlen;
err = ubifs_tnc_add(c, &ino_key, lnum, ino_offs, ilen, hash_ino);
if (err)
goto out_ro;
ino_key_init(c, &ino_key, dir->i_ino);
ino_offs += aligned_ilen;
err = ubifs_tnc_add(c, &ino_key, lnum, ino_offs,
UBIFS_INO_NODE_SZ + host_ui->data_len, hash_ino_host);
if (err)
goto out_ro;
if (in_orphan && inode->i_nlink)
ubifs_delete_orphan(c, inode->i_ino);
finish_reservation(c);
spin_lock(&ui->ui_lock);
ui->synced_i_size = ui->ui_size;
spin_unlock(&ui->ui_lock);
if (xent) {
spin_lock(&host_ui->ui_lock);
host_ui->synced_i_size = host_ui->ui_size;
spin_unlock(&host_ui->ui_lock);
}
mark_inode_clean(c, ui);
mark_inode_clean(c, host_ui);
return 0;
out_finish:
finish_reservation(c);
out_free:
kfree(dent);
return err;
out_release:
release_head(c, BASEHD);
kfree(dent);
out_ro:
ubifs_ro_mode(c, err);
if (orphan_added)
ubifs_delete_orphan(c, inode->i_ino);
finish_reservation(c);
return err;
}
/**
* ubifs_jnl_write_data - write a data node to the journal.
* @c: UBIFS file-system description object
* @inode: inode the data node belongs to
* @key: node key
* @buf: buffer to write
* @len: data length (must not exceed %UBIFS_BLOCK_SIZE)
*
* This function writes a data node to the journal. Returns %0 if the data node
* was successfully written, and a negative error code in case of failure.
*/
int ubifs_jnl_write_data(struct ubifs_info *c, const struct inode *inode,
const union ubifs_key *key, const void *buf, int len)
{
struct ubifs_data_node *data;
int err, lnum, offs, compr_type, out_len, compr_len, auth_len;
int dlen = COMPRESSED_DATA_NODE_BUF_SZ, allocated = 1;
int write_len;
struct ubifs_inode *ui = ubifs_inode(inode);
bool encrypted = IS_ENCRYPTED(inode);
u8 hash[UBIFS_HASH_ARR_SZ];
dbg_jnlk(key, "ino %lu, blk %u, len %d, key ",
(unsigned long)key_inum(c, key), key_block(c, key), len);
ubifs_assert(c, len <= UBIFS_BLOCK_SIZE);
if (encrypted)
dlen += UBIFS_CIPHER_BLOCK_SIZE;
auth_len = ubifs_auth_node_sz(c);
data = kmalloc(dlen + auth_len, GFP_NOFS | __GFP_NOWARN);
if (!data) {
/*
* Fall-back to the write reserve buffer. Note, we might be
* currently on the memory reclaim path, when the kernel is
* trying to free some memory by writing out dirty pages. The
* write reserve buffer helps us to guarantee that we are
* always able to write the data.
*/
allocated = 0;
mutex_lock(&c->write_reserve_mutex);
data = c->write_reserve_buf;
}
data->ch.node_type = UBIFS_DATA_NODE;
key_write(c, key, &data->key);
data->size = cpu_to_le32(len);
if (!(ui->flags & UBIFS_COMPR_FL))
/* Compression is disabled for this inode */
compr_type = UBIFS_COMPR_NONE;
else
compr_type = ui->compr_type;
out_len = compr_len = dlen - UBIFS_DATA_NODE_SZ;
ubifs_compress(c, buf, len, &data->data, &compr_len, &compr_type);
ubifs_assert(c, compr_len <= UBIFS_BLOCK_SIZE);
if (encrypted) {
err = ubifs_encrypt(inode, data, compr_len, &out_len, key_block(c, key));
if (err)
goto out_free;
} else {
data->compr_size = 0;
out_len = compr_len;
}
dlen = UBIFS_DATA_NODE_SZ + out_len;
if (ubifs_authenticated(c))
write_len = ALIGN(dlen, 8) + auth_len;
else
write_len = dlen;
data->compr_type = cpu_to_le16(compr_type);
/* Make reservation before allocating sequence numbers */
err = make_reservation(c, DATAHD, write_len);
if (err)
goto out_free;
ubifs_prepare_node(c, data, dlen, 0);
err = write_head(c, DATAHD, data, write_len, &lnum, &offs, 0);
if (err)
goto out_release;
err = ubifs_node_calc_hash(c, data, hash);
if (err)
goto out_release;
ubifs_wbuf_add_ino_nolock(&c->jheads[DATAHD].wbuf, key_inum(c, key));
release_head(c, DATAHD);
ubifs_add_auth_dirt(c, lnum);
err = ubifs_tnc_add(c, key, lnum, offs, dlen, hash);
if (err)
goto out_ro;
finish_reservation(c);
if (!allocated)
mutex_unlock(&c->write_reserve_mutex);
else
kfree(data);
return 0;
out_release:
release_head(c, DATAHD);
out_ro:
ubifs_ro_mode(c, err);
finish_reservation(c);
out_free:
if (!allocated)
mutex_unlock(&c->write_reserve_mutex);
else
kfree(data);
return err;
}
/**
* ubifs_jnl_write_inode - flush inode to the journal.
* @c: UBIFS file-system description object
* @inode: inode to flush
*
* This function writes inode @inode to the journal. If the inode is
* synchronous, it also synchronizes the write-buffer. Returns zero in case of
* success and a negative error code in case of failure.
*/
int ubifs_jnl_write_inode(struct ubifs_info *c, const struct inode *inode)
{
int err, lnum, offs;
struct ubifs_ino_node *ino, *ino_start;
struct ubifs_inode *ui = ubifs_inode(inode);
int sync = 0, write_len = 0, ilen = UBIFS_INO_NODE_SZ;
int last_reference = !inode->i_nlink;
int kill_xattrs = ui->xattr_cnt && last_reference;
u8 hash[UBIFS_HASH_ARR_SZ];
dbg_jnl("ino %lu, nlink %u", inode->i_ino, inode->i_nlink);
if (kill_xattrs && ui->xattr_cnt > ubifs_xattr_max_cnt(c)) {
ubifs_err(c, "Cannot delete inode, it has too much xattrs!");
err = -EPERM;
ubifs_ro_mode(c, err);
return err;
}
/*
* If the inode is being deleted, do not write the attached data. No
* need to synchronize the write-buffer either.
*/
if (!last_reference) {
ilen += ui->data_len;
sync = IS_SYNC(inode);
} else if (kill_xattrs) {
write_len += UBIFS_INO_NODE_SZ * ui->xattr_cnt;
}