forked from torvalds/linux
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmballoc.c
4657 lines (4095 loc) · 126 KB
/
mballoc.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
/*
* Copyright (c) 2003-2006, Cluster File Systems, Inc, [email protected]
* Written by Alex Tomas <[email protected]>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*
* 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 Licens
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-
*/
/*
* mballoc.c contains the multiblocks allocation routines
*/
#include "mballoc.h"
/*
* MUSTDO:
* - test ext4_ext_search_left() and ext4_ext_search_right()
* - search for metadata in few groups
*
* TODO v4:
* - normalization should take into account whether file is still open
* - discard preallocations if no free space left (policy?)
* - don't normalize tails
* - quota
* - reservation for superuser
*
* TODO v3:
* - bitmap read-ahead (proposed by Oleg Drokin aka green)
* - track min/max extents in each group for better group selection
* - mb_mark_used() may allocate chunk right after splitting buddy
* - tree of groups sorted by number of free blocks
* - error handling
*/
/*
* The allocation request involve request for multiple number of blocks
* near to the goal(block) value specified.
*
* During initialization phase of the allocator we decide to use the group
* preallocation or inode preallocation depending on the size file. The
* size of the file could be the resulting file size we would have after
* allocation or the current file size which ever is larger. If the size is
* less that sbi->s_mb_stream_request we select the group
* preallocation. The default value of s_mb_stream_request is 16
* blocks. This can also be tuned via
* /proc/fs/ext4/<partition>/stream_req. The value is represented in terms
* of number of blocks.
*
* The main motivation for having small file use group preallocation is to
* ensure that we have small file closer in the disk.
*
* First stage the allocator looks at the inode prealloc list
* ext4_inode_info->i_prealloc_list contain list of prealloc spaces for
* this particular inode. The inode prealloc space is represented as:
*
* pa_lstart -> the logical start block for this prealloc space
* pa_pstart -> the physical start block for this prealloc space
* pa_len -> lenght for this prealloc space
* pa_free -> free space available in this prealloc space
*
* The inode preallocation space is used looking at the _logical_ start
* block. If only the logical file block falls within the range of prealloc
* space we will consume the particular prealloc space. This make sure that
* that the we have contiguous physical blocks representing the file blocks
*
* The important thing to be noted in case of inode prealloc space is that
* we don't modify the values associated to inode prealloc space except
* pa_free.
*
* If we are not able to find blocks in the inode prealloc space and if we
* have the group allocation flag set then we look at the locality group
* prealloc space. These are per CPU prealloc list repreasented as
*
* ext4_sb_info.s_locality_groups[smp_processor_id()]
*
* The reason for having a per cpu locality group is to reduce the contention
* between CPUs. It is possible to get scheduled at this point.
*
* The locality group prealloc space is used looking at whether we have
* enough free space (pa_free) withing the prealloc space.
*
* If we can't allocate blocks via inode prealloc or/and locality group
* prealloc then we look at the buddy cache. The buddy cache is represented
* by ext4_sb_info.s_buddy_cache (struct inode) whose file offset gets
* mapped to the buddy and bitmap information regarding different
* groups. The buddy information is attached to buddy cache inode so that
* we can access them through the page cache. The information regarding
* each group is loaded via ext4_mb_load_buddy. The information involve
* block bitmap and buddy information. The information are stored in the
* inode as:
*
* { page }
* [ group 0 buddy][ group 0 bitmap] [group 1][ group 1]...
*
*
* one block each for bitmap and buddy information. So for each group we
* take up 2 blocks. A page can contain blocks_per_page (PAGE_CACHE_SIZE /
* blocksize) blocks. So it can have information regarding groups_per_page
* which is blocks_per_page/2
*
* The buddy cache inode is not stored on disk. The inode is thrown
* away when the filesystem is unmounted.
*
* We look for count number of blocks in the buddy cache. If we were able
* to locate that many free blocks we return with additional information
* regarding rest of the contiguous physical block available
*
* Before allocating blocks via buddy cache we normalize the request
* blocks. This ensure we ask for more blocks that we needed. The extra
* blocks that we get after allocation is added to the respective prealloc
* list. In case of inode preallocation we follow a list of heuristics
* based on file size. This can be found in ext4_mb_normalize_request. If
* we are doing a group prealloc we try to normalize the request to
* sbi->s_mb_group_prealloc. Default value of s_mb_group_prealloc is set to
* 512 blocks. This can be tuned via
* /proc/fs/ext4/<partition/group_prealloc. The value is represented in
* terms of number of blocks. If we have mounted the file system with -O
* stripe=<value> option the group prealloc request is normalized to the
* stripe value (sbi->s_stripe)
*
* The regular allocator(using the buddy cache) support few tunables.
*
* /proc/fs/ext4/<partition>/min_to_scan
* /proc/fs/ext4/<partition>/max_to_scan
* /proc/fs/ext4/<partition>/order2_req
*
* The regular allocator use buddy scan only if the request len is power of
* 2 blocks and the order of allocation is >= sbi->s_mb_order2_reqs. The
* value of s_mb_order2_reqs can be tuned via
* /proc/fs/ext4/<partition>/order2_req. If the request len is equal to
* stripe size (sbi->s_stripe), we try to search for contigous block in
* stripe size. This should result in better allocation on RAID setup. If
* not we search in the specific group using bitmap for best extents. The
* tunable min_to_scan and max_to_scan controll the behaviour here.
* min_to_scan indicate how long the mballoc __must__ look for a best
* extent and max_to_scanindicate how long the mballoc __can__ look for a
* best extent in the found extents. Searching for the blocks starts with
* the group specified as the goal value in allocation context via
* ac_g_ex. Each group is first checked based on the criteria whether it
* can used for allocation. ext4_mb_good_group explains how the groups are
* checked.
*
* Both the prealloc space are getting populated as above. So for the first
* request we will hit the buddy cache which will result in this prealloc
* space getting filled. The prealloc space is then later used for the
* subsequent request.
*/
/*
* mballoc operates on the following data:
* - on-disk bitmap
* - in-core buddy (actually includes buddy and bitmap)
* - preallocation descriptors (PAs)
*
* there are two types of preallocations:
* - inode
* assiged to specific inode and can be used for this inode only.
* it describes part of inode's space preallocated to specific
* physical blocks. any block from that preallocated can be used
* independent. the descriptor just tracks number of blocks left
* unused. so, before taking some block from descriptor, one must
* make sure corresponded logical block isn't allocated yet. this
* also means that freeing any block within descriptor's range
* must discard all preallocated blocks.
* - locality group
* assigned to specific locality group which does not translate to
* permanent set of inodes: inode can join and leave group. space
* from this type of preallocation can be used for any inode. thus
* it's consumed from the beginning to the end.
*
* relation between them can be expressed as:
* in-core buddy = on-disk bitmap + preallocation descriptors
*
* this mean blocks mballoc considers used are:
* - allocated blocks (persistent)
* - preallocated blocks (non-persistent)
*
* consistency in mballoc world means that at any time a block is either
* free or used in ALL structures. notice: "any time" should not be read
* literally -- time is discrete and delimited by locks.
*
* to keep it simple, we don't use block numbers, instead we count number of
* blocks: how many blocks marked used/free in on-disk bitmap, buddy and PA.
*
* all operations can be expressed as:
* - init buddy: buddy = on-disk + PAs
* - new PA: buddy += N; PA = N
* - use inode PA: on-disk += N; PA -= N
* - discard inode PA buddy -= on-disk - PA; PA = 0
* - use locality group PA on-disk += N; PA -= N
* - discard locality group PA buddy -= PA; PA = 0
* note: 'buddy -= on-disk - PA' is used to show that on-disk bitmap
* is used in real operation because we can't know actual used
* bits from PA, only from on-disk bitmap
*
* if we follow this strict logic, then all operations above should be atomic.
* given some of them can block, we'd have to use something like semaphores
* killing performance on high-end SMP hardware. let's try to relax it using
* the following knowledge:
* 1) if buddy is referenced, it's already initialized
* 2) while block is used in buddy and the buddy is referenced,
* nobody can re-allocate that block
* 3) we work on bitmaps and '+' actually means 'set bits'. if on-disk has
* bit set and PA claims same block, it's OK. IOW, one can set bit in
* on-disk bitmap if buddy has same bit set or/and PA covers corresponded
* block
*
* so, now we're building a concurrency table:
* - init buddy vs.
* - new PA
* blocks for PA are allocated in the buddy, buddy must be referenced
* until PA is linked to allocation group to avoid concurrent buddy init
* - use inode PA
* we need to make sure that either on-disk bitmap or PA has uptodate data
* given (3) we care that PA-=N operation doesn't interfere with init
* - discard inode PA
* the simplest way would be to have buddy initialized by the discard
* - use locality group PA
* again PA-=N must be serialized with init
* - discard locality group PA
* the simplest way would be to have buddy initialized by the discard
* - new PA vs.
* - use inode PA
* i_data_sem serializes them
* - discard inode PA
* discard process must wait until PA isn't used by another process
* - use locality group PA
* some mutex should serialize them
* - discard locality group PA
* discard process must wait until PA isn't used by another process
* - use inode PA
* - use inode PA
* i_data_sem or another mutex should serializes them
* - discard inode PA
* discard process must wait until PA isn't used by another process
* - use locality group PA
* nothing wrong here -- they're different PAs covering different blocks
* - discard locality group PA
* discard process must wait until PA isn't used by another process
*
* now we're ready to make few consequences:
* - PA is referenced and while it is no discard is possible
* - PA is referenced until block isn't marked in on-disk bitmap
* - PA changes only after on-disk bitmap
* - discard must not compete with init. either init is done before
* any discard or they're serialized somehow
* - buddy init as sum of on-disk bitmap and PAs is done atomically
*
* a special case when we've used PA to emptiness. no need to modify buddy
* in this case, but we should care about concurrent init
*
*/
/*
* Logic in few words:
*
* - allocation:
* load group
* find blocks
* mark bits in on-disk bitmap
* release group
*
* - use preallocation:
* find proper PA (per-inode or group)
* load group
* mark bits in on-disk bitmap
* release group
* release PA
*
* - free:
* load group
* mark bits in on-disk bitmap
* release group
*
* - discard preallocations in group:
* mark PAs deleted
* move them onto local list
* load on-disk bitmap
* load group
* remove PA from object (inode or locality group)
* mark free blocks in-core
*
* - discard inode's preallocations:
*/
/*
* Locking rules
*
* Locks:
* - bitlock on a group (group)
* - object (inode/locality) (object)
* - per-pa lock (pa)
*
* Paths:
* - new pa
* object
* group
*
* - find and use pa:
* pa
*
* - release consumed pa:
* pa
* group
* object
*
* - generate in-core bitmap:
* group
* pa
*
* - discard all for given object (inode, locality group):
* object
* pa
* group
*
* - discard all for given group:
* group
* pa
* group
* object
*
*/
static inline void *mb_correct_addr_and_bit(int *bit, void *addr)
{
#if BITS_PER_LONG == 64
*bit += ((unsigned long) addr & 7UL) << 3;
addr = (void *) ((unsigned long) addr & ~7UL);
#elif BITS_PER_LONG == 32
*bit += ((unsigned long) addr & 3UL) << 3;
addr = (void *) ((unsigned long) addr & ~3UL);
#else
#error "how many bits you are?!"
#endif
return addr;
}
static inline int mb_test_bit(int bit, void *addr)
{
/*
* ext4_test_bit on architecture like powerpc
* needs unsigned long aligned address
*/
addr = mb_correct_addr_and_bit(&bit, addr);
return ext4_test_bit(bit, addr);
}
static inline void mb_set_bit(int bit, void *addr)
{
addr = mb_correct_addr_and_bit(&bit, addr);
ext4_set_bit(bit, addr);
}
static inline void mb_set_bit_atomic(spinlock_t *lock, int bit, void *addr)
{
addr = mb_correct_addr_and_bit(&bit, addr);
ext4_set_bit_atomic(lock, bit, addr);
}
static inline void mb_clear_bit(int bit, void *addr)
{
addr = mb_correct_addr_and_bit(&bit, addr);
ext4_clear_bit(bit, addr);
}
static inline void mb_clear_bit_atomic(spinlock_t *lock, int bit, void *addr)
{
addr = mb_correct_addr_and_bit(&bit, addr);
ext4_clear_bit_atomic(lock, bit, addr);
}
static inline int mb_find_next_zero_bit(void *addr, int max, int start)
{
int fix = 0, ret, tmpmax;
addr = mb_correct_addr_and_bit(&fix, addr);
tmpmax = max + fix;
start += fix;
ret = ext4_find_next_zero_bit(addr, tmpmax, start) - fix;
if (ret > max)
return max;
return ret;
}
static inline int mb_find_next_bit(void *addr, int max, int start)
{
int fix = 0, ret, tmpmax;
addr = mb_correct_addr_and_bit(&fix, addr);
tmpmax = max + fix;
start += fix;
ret = ext4_find_next_bit(addr, tmpmax, start) - fix;
if (ret > max)
return max;
return ret;
}
static void *mb_find_buddy(struct ext4_buddy *e4b, int order, int *max)
{
char *bb;
BUG_ON(EXT4_MB_BITMAP(e4b) == EXT4_MB_BUDDY(e4b));
BUG_ON(max == NULL);
if (order > e4b->bd_blkbits + 1) {
*max = 0;
return NULL;
}
/* at order 0 we see each particular block */
*max = 1 << (e4b->bd_blkbits + 3);
if (order == 0)
return EXT4_MB_BITMAP(e4b);
bb = EXT4_MB_BUDDY(e4b) + EXT4_SB(e4b->bd_sb)->s_mb_offsets[order];
*max = EXT4_SB(e4b->bd_sb)->s_mb_maxs[order];
return bb;
}
#ifdef DOUBLE_CHECK
static void mb_free_blocks_double(struct inode *inode, struct ext4_buddy *e4b,
int first, int count)
{
int i;
struct super_block *sb = e4b->bd_sb;
if (unlikely(e4b->bd_info->bb_bitmap == NULL))
return;
BUG_ON(!ext4_is_group_locked(sb, e4b->bd_group));
for (i = 0; i < count; i++) {
if (!mb_test_bit(first + i, e4b->bd_info->bb_bitmap)) {
ext4_fsblk_t blocknr;
blocknr = e4b->bd_group * EXT4_BLOCKS_PER_GROUP(sb);
blocknr += first + i;
blocknr +=
le32_to_cpu(EXT4_SB(sb)->s_es->s_first_data_block);
ext4_error(sb, __func__, "double-free of inode"
" %lu's block %llu(bit %u in group %lu)\n",
inode ? inode->i_ino : 0, blocknr,
first + i, e4b->bd_group);
}
mb_clear_bit(first + i, e4b->bd_info->bb_bitmap);
}
}
static void mb_mark_used_double(struct ext4_buddy *e4b, int first, int count)
{
int i;
if (unlikely(e4b->bd_info->bb_bitmap == NULL))
return;
BUG_ON(!ext4_is_group_locked(e4b->bd_sb, e4b->bd_group));
for (i = 0; i < count; i++) {
BUG_ON(mb_test_bit(first + i, e4b->bd_info->bb_bitmap));
mb_set_bit(first + i, e4b->bd_info->bb_bitmap);
}
}
static void mb_cmp_bitmaps(struct ext4_buddy *e4b, void *bitmap)
{
if (memcmp(e4b->bd_info->bb_bitmap, bitmap, e4b->bd_sb->s_blocksize)) {
unsigned char *b1, *b2;
int i;
b1 = (unsigned char *) e4b->bd_info->bb_bitmap;
b2 = (unsigned char *) bitmap;
for (i = 0; i < e4b->bd_sb->s_blocksize; i++) {
if (b1[i] != b2[i]) {
printk(KERN_ERR "corruption in group %lu "
"at byte %u(%u): %x in copy != %x "
"on disk/prealloc\n",
e4b->bd_group, i, i * 8, b1[i], b2[i]);
BUG();
}
}
}
}
#else
static inline void mb_free_blocks_double(struct inode *inode,
struct ext4_buddy *e4b, int first, int count)
{
return;
}
static inline void mb_mark_used_double(struct ext4_buddy *e4b,
int first, int count)
{
return;
}
static inline void mb_cmp_bitmaps(struct ext4_buddy *e4b, void *bitmap)
{
return;
}
#endif
#ifdef AGGRESSIVE_CHECK
#define MB_CHECK_ASSERT(assert) \
do { \
if (!(assert)) { \
printk(KERN_EMERG \
"Assertion failure in %s() at %s:%d: \"%s\"\n", \
function, file, line, # assert); \
BUG(); \
} \
} while (0)
static int __mb_check_buddy(struct ext4_buddy *e4b, char *file,
const char *function, int line)
{
struct super_block *sb = e4b->bd_sb;
int order = e4b->bd_blkbits + 1;
int max;
int max2;
int i;
int j;
int k;
int count;
struct ext4_group_info *grp;
int fragments = 0;
int fstart;
struct list_head *cur;
void *buddy;
void *buddy2;
{
static int mb_check_counter;
if (mb_check_counter++ % 100 != 0)
return 0;
}
while (order > 1) {
buddy = mb_find_buddy(e4b, order, &max);
MB_CHECK_ASSERT(buddy);
buddy2 = mb_find_buddy(e4b, order - 1, &max2);
MB_CHECK_ASSERT(buddy2);
MB_CHECK_ASSERT(buddy != buddy2);
MB_CHECK_ASSERT(max * 2 == max2);
count = 0;
for (i = 0; i < max; i++) {
if (mb_test_bit(i, buddy)) {
/* only single bit in buddy2 may be 1 */
if (!mb_test_bit(i << 1, buddy2)) {
MB_CHECK_ASSERT(
mb_test_bit((i<<1)+1, buddy2));
} else if (!mb_test_bit((i << 1) + 1, buddy2)) {
MB_CHECK_ASSERT(
mb_test_bit(i << 1, buddy2));
}
continue;
}
/* both bits in buddy2 must be 0 */
MB_CHECK_ASSERT(mb_test_bit(i << 1, buddy2));
MB_CHECK_ASSERT(mb_test_bit((i << 1) + 1, buddy2));
for (j = 0; j < (1 << order); j++) {
k = (i * (1 << order)) + j;
MB_CHECK_ASSERT(
!mb_test_bit(k, EXT4_MB_BITMAP(e4b)));
}
count++;
}
MB_CHECK_ASSERT(e4b->bd_info->bb_counters[order] == count);
order--;
}
fstart = -1;
buddy = mb_find_buddy(e4b, 0, &max);
for (i = 0; i < max; i++) {
if (!mb_test_bit(i, buddy)) {
MB_CHECK_ASSERT(i >= e4b->bd_info->bb_first_free);
if (fstart == -1) {
fragments++;
fstart = i;
}
continue;
}
fstart = -1;
/* check used bits only */
for (j = 0; j < e4b->bd_blkbits + 1; j++) {
buddy2 = mb_find_buddy(e4b, j, &max2);
k = i >> j;
MB_CHECK_ASSERT(k < max2);
MB_CHECK_ASSERT(mb_test_bit(k, buddy2));
}
}
MB_CHECK_ASSERT(!EXT4_MB_GRP_NEED_INIT(e4b->bd_info));
MB_CHECK_ASSERT(e4b->bd_info->bb_fragments == fragments);
grp = ext4_get_group_info(sb, e4b->bd_group);
buddy = mb_find_buddy(e4b, 0, &max);
list_for_each(cur, &grp->bb_prealloc_list) {
ext4_group_t groupnr;
struct ext4_prealloc_space *pa;
pa = list_entry(cur, struct ext4_prealloc_space, pa_group_list);
ext4_get_group_no_and_offset(sb, pa->pa_pstart, &groupnr, &k);
MB_CHECK_ASSERT(groupnr == e4b->bd_group);
for (i = 0; i < pa->pa_len; i++)
MB_CHECK_ASSERT(mb_test_bit(k + i, buddy));
}
return 0;
}
#undef MB_CHECK_ASSERT
#define mb_check_buddy(e4b) __mb_check_buddy(e4b, \
__FILE__, __func__, __LINE__)
#else
#define mb_check_buddy(e4b)
#endif
/* FIXME!! need more doc */
static void ext4_mb_mark_free_simple(struct super_block *sb,
void *buddy, unsigned first, int len,
struct ext4_group_info *grp)
{
struct ext4_sb_info *sbi = EXT4_SB(sb);
unsigned short min;
unsigned short max;
unsigned short chunk;
unsigned short border;
BUG_ON(len > EXT4_BLOCKS_PER_GROUP(sb));
border = 2 << sb->s_blocksize_bits;
while (len > 0) {
/* find how many blocks can be covered since this position */
max = ffs(first | border) - 1;
/* find how many blocks of power 2 we need to mark */
min = fls(len) - 1;
if (max < min)
min = max;
chunk = 1 << min;
/* mark multiblock chunks only */
grp->bb_counters[min]++;
if (min > 0)
mb_clear_bit(first >> min,
buddy + sbi->s_mb_offsets[min]);
len -= chunk;
first += chunk;
}
}
static void ext4_mb_generate_buddy(struct super_block *sb,
void *buddy, void *bitmap, ext4_group_t group)
{
struct ext4_group_info *grp = ext4_get_group_info(sb, group);
unsigned short max = EXT4_BLOCKS_PER_GROUP(sb);
unsigned short i = 0;
unsigned short first;
unsigned short len;
unsigned free = 0;
unsigned fragments = 0;
unsigned long long period = get_cycles();
/* initialize buddy from bitmap which is aggregation
* of on-disk bitmap and preallocations */
i = mb_find_next_zero_bit(bitmap, max, 0);
grp->bb_first_free = i;
while (i < max) {
fragments++;
first = i;
i = mb_find_next_bit(bitmap, max, i);
len = i - first;
free += len;
if (len > 1)
ext4_mb_mark_free_simple(sb, buddy, first, len, grp);
else
grp->bb_counters[0]++;
if (i < max)
i = mb_find_next_zero_bit(bitmap, max, i);
}
grp->bb_fragments = fragments;
if (free != grp->bb_free) {
ext4_error(sb, __func__,
"EXT4-fs: group %lu: %u blocks in bitmap, %u in gd\n",
group, free, grp->bb_free);
/*
* If we intent to continue, we consider group descritor
* corrupt and update bb_free using bitmap value
*/
grp->bb_free = free;
}
clear_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &(grp->bb_state));
period = get_cycles() - period;
spin_lock(&EXT4_SB(sb)->s_bal_lock);
EXT4_SB(sb)->s_mb_buddies_generated++;
EXT4_SB(sb)->s_mb_generation_time += period;
spin_unlock(&EXT4_SB(sb)->s_bal_lock);
}
/* The buddy information is attached the buddy cache inode
* for convenience. The information regarding each group
* is loaded via ext4_mb_load_buddy. The information involve
* block bitmap and buddy information. The information are
* stored in the inode as
*
* { page }
* [ group 0 buddy][ group 0 bitmap] [group 1][ group 1]...
*
*
* one block each for bitmap and buddy information.
* So for each group we take up 2 blocks. A page can
* contain blocks_per_page (PAGE_CACHE_SIZE / blocksize) blocks.
* So it can have information regarding groups_per_page which
* is blocks_per_page/2
*/
static int ext4_mb_init_cache(struct page *page, char *incore)
{
int blocksize;
int blocks_per_page;
int groups_per_page;
int err = 0;
int i;
ext4_group_t first_group;
int first_block;
struct super_block *sb;
struct buffer_head *bhs;
struct buffer_head **bh;
struct inode *inode;
char *data;
char *bitmap;
mb_debug("init page %lu\n", page->index);
inode = page->mapping->host;
sb = inode->i_sb;
blocksize = 1 << inode->i_blkbits;
blocks_per_page = PAGE_CACHE_SIZE / blocksize;
groups_per_page = blocks_per_page >> 1;
if (groups_per_page == 0)
groups_per_page = 1;
/* allocate buffer_heads to read bitmaps */
if (groups_per_page > 1) {
err = -ENOMEM;
i = sizeof(struct buffer_head *) * groups_per_page;
bh = kzalloc(i, GFP_NOFS);
if (bh == NULL)
goto out;
} else
bh = &bhs;
first_group = page->index * blocks_per_page / 2;
/* read all groups the page covers into the cache */
for (i = 0; i < groups_per_page; i++) {
struct ext4_group_desc *desc;
if (first_group + i >= EXT4_SB(sb)->s_groups_count)
break;
err = -EIO;
desc = ext4_get_group_desc(sb, first_group + i, NULL);
if (desc == NULL)
goto out;
err = -ENOMEM;
bh[i] = sb_getblk(sb, ext4_block_bitmap(sb, desc));
if (bh[i] == NULL)
goto out;
if (buffer_uptodate(bh[i]) &&
!(desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)))
continue;
lock_buffer(bh[i]);
spin_lock(sb_bgl_lock(EXT4_SB(sb), first_group + i));
if (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
ext4_init_block_bitmap(sb, bh[i],
first_group + i, desc);
set_buffer_uptodate(bh[i]);
unlock_buffer(bh[i]);
spin_unlock(sb_bgl_lock(EXT4_SB(sb), first_group + i));
continue;
}
spin_unlock(sb_bgl_lock(EXT4_SB(sb), first_group + i));
get_bh(bh[i]);
bh[i]->b_end_io = end_buffer_read_sync;
submit_bh(READ, bh[i]);
mb_debug("read bitmap for group %lu\n", first_group + i);
}
/* wait for I/O completion */
for (i = 0; i < groups_per_page && bh[i]; i++)
wait_on_buffer(bh[i]);
err = -EIO;
for (i = 0; i < groups_per_page && bh[i]; i++)
if (!buffer_uptodate(bh[i]))
goto out;
err = 0;
first_block = page->index * blocks_per_page;
for (i = 0; i < blocks_per_page; i++) {
int group;
struct ext4_group_info *grinfo;
group = (first_block + i) >> 1;
if (group >= EXT4_SB(sb)->s_groups_count)
break;
/*
* data carry information regarding this
* particular group in the format specified
* above
*
*/
data = page_address(page) + (i * blocksize);
bitmap = bh[group - first_group]->b_data;
/*
* We place the buddy block and bitmap block
* close together
*/
if ((first_block + i) & 1) {
/* this is block of buddy */
BUG_ON(incore == NULL);
mb_debug("put buddy for group %u in page %lu/%x\n",
group, page->index, i * blocksize);
memset(data, 0xff, blocksize);
grinfo = ext4_get_group_info(sb, group);
grinfo->bb_fragments = 0;
memset(grinfo->bb_counters, 0,
sizeof(unsigned short)*(sb->s_blocksize_bits+2));
/*
* incore got set to the group block bitmap below
*/
ext4_mb_generate_buddy(sb, data, incore, group);
incore = NULL;
} else {
/* this is block of bitmap */
BUG_ON(incore != NULL);
mb_debug("put bitmap for group %u in page %lu/%x\n",
group, page->index, i * blocksize);
/* see comments in ext4_mb_put_pa() */
ext4_lock_group(sb, group);
memcpy(data, bitmap, blocksize);
/* mark all preallocated blks used in in-core bitmap */
ext4_mb_generate_from_pa(sb, data, group);
ext4_unlock_group(sb, group);
/* set incore so that the buddy information can be
* generated using this
*/
incore = data;
}
}
SetPageUptodate(page);
out:
if (bh) {
for (i = 0; i < groups_per_page && bh[i]; i++)
brelse(bh[i]);
if (bh != &bhs)
kfree(bh);
}
return err;
}
static noinline_for_stack int
ext4_mb_load_buddy(struct super_block *sb, ext4_group_t group,
struct ext4_buddy *e4b)
{
struct ext4_sb_info *sbi = EXT4_SB(sb);
struct inode *inode = sbi->s_buddy_cache;
int blocks_per_page;
int block;
int pnum;
int poff;
struct page *page;
int ret;
mb_debug("load group %lu\n", group);
blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize;
e4b->bd_blkbits = sb->s_blocksize_bits;
e4b->bd_info = ext4_get_group_info(sb, group);
e4b->bd_sb = sb;
e4b->bd_group = group;
e4b->bd_buddy_page = NULL;
e4b->bd_bitmap_page = NULL;
/*
* the buddy cache inode stores the block bitmap
* and buddy information in consecutive blocks.
* So for each group we need two blocks.
*/
block = group * 2;
pnum = block / blocks_per_page;
poff = block % blocks_per_page;
/* we could use find_or_create_page(), but it locks page
* what we'd like to avoid in fast path ... */
page = find_get_page(inode->i_mapping, pnum);
if (page == NULL || !PageUptodate(page)) {
if (page)
page_cache_release(page);
page = find_or_create_page(inode->i_mapping, pnum, GFP_NOFS);
if (page) {
BUG_ON(page->mapping != inode->i_mapping);
if (!PageUptodate(page)) {
ret = ext4_mb_init_cache(page, NULL);
if (ret) {
unlock_page(page);
goto err;
}
mb_cmp_bitmaps(e4b, page_address(page) +
(poff * sb->s_blocksize));
}
unlock_page(page);
}
}
if (page == NULL || !PageUptodate(page)) {
ret = -EIO;
goto err;
}
e4b->bd_bitmap_page = page;
e4b->bd_bitmap = page_address(page) + (poff * sb->s_blocksize);
mark_page_accessed(page);
block++;
pnum = block / blocks_per_page;
poff = block % blocks_per_page;
page = find_get_page(inode->i_mapping, pnum);
if (page == NULL || !PageUptodate(page)) {
if (page)
page_cache_release(page);
page = find_or_create_page(inode->i_mapping, pnum, GFP_NOFS);
if (page) {
BUG_ON(page->mapping != inode->i_mapping);
if (!PageUptodate(page)) {
ret = ext4_mb_init_cache(page, e4b->bd_bitmap);
if (ret) {
unlock_page(page);
goto err;
}
}
unlock_page(page);
}
}
if (page == NULL || !PageUptodate(page)) {
ret = -EIO;
goto err;
}
e4b->bd_buddy_page = page;
e4b->bd_buddy = page_address(page) + (poff * sb->s_blocksize);
mark_page_accessed(page);
BUG_ON(e4b->bd_bitmap_page == NULL);
BUG_ON(e4b->bd_buddy_page == NULL);
return 0;
err:
if (e4b->bd_bitmap_page)
page_cache_release(e4b->bd_bitmap_page);
if (e4b->bd_buddy_page)
page_cache_release(e4b->bd_buddy_page);
e4b->bd_buddy = NULL;
e4b->bd_bitmap = NULL;
return ret;
}
static void ext4_mb_release_desc(struct ext4_buddy *e4b)
{
if (e4b->bd_bitmap_page)
page_cache_release(e4b->bd_bitmap_page);
if (e4b->bd_buddy_page)
page_cache_release(e4b->bd_buddy_page);
}
static int mb_find_order_for_block(struct ext4_buddy *e4b, int block)