?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
[PATCH] user of the jiffies rounding patch: Slab
[linux-2.6/linux-acpi-2.6/ibm-acpi-2.6.git] / fs / ocfs2 / suballoc.c
blob000d71cca6c50dfe919524fb16c7d8e261230701
1 /* -*- mode: c; c-basic-offset: 8; -*-
2 * vim: noexpandtab sw=8 ts=8 sts=0:
3 *
4 * suballoc.c
5 *
6 * metadata alloc and free
7 * Inspired by ext3 block groups.
8 *
9 * Copyright (C) 2002, 2004 Oracle. All rights reserved.
10 *
11 * This program is free software; you can redistribute it and/or
12 * modify it under the terms of the GNU General Public
13 * License as published by the Free Software Foundation; either
14 * version 2 of the License, or (at your option) any later version.
15 *
16 * This program is distributed in the hope that it will be useful,
17 * but WITHOUT ANY WARRANTY; without even the implied warranty of
18 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
19 * General Public License for more details.
20 *
21 * You should have received a copy of the GNU General Public
22 * License along with this program; if not, write to the
23 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
24 * Boston, MA 021110-1307, USA.
25 */
26
27 #include <linux/fs.h>
28 #include <linux/types.h>
29 #include <linux/slab.h>
30 #include <linux/highmem.h>
31
32 #define MLOG_MASK_PREFIX ML_DISK_ALLOC
33 #include <cluster/masklog.h>
34
35 #include "ocfs2.h"
36
37 #include "alloc.h"
38 #include "dlmglue.h"
39 #include "inode.h"
40 #include "journal.h"
41 #include "localalloc.h"
42 #include "suballoc.h"
43 #include "super.h"
44 #include "sysfile.h"
45 #include "uptodate.h"
46
47 #include "buffer_head_io.h"
48
49 static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg);
50 static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe);
51 static inline u16 ocfs2_find_victim_chain(struct ocfs2_chain_list *cl);
52 static int ocfs2_block_group_fill(handle_t *handle,
53 struct inode *alloc_inode,
54 struct buffer_head *bg_bh,
55 u64 group_blkno,
56 u16 my_chain,
57 struct ocfs2_chain_list *cl);
58 static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
59 struct inode *alloc_inode,
60 struct buffer_head *bh);
61
62 static int ocfs2_cluster_group_search(struct inode *inode,
63 struct buffer_head *group_bh,
64 u32 bits_wanted, u32 min_bits,
65 u16 *bit_off, u16 *bits_found);
66 static int ocfs2_block_group_search(struct inode *inode,
67 struct buffer_head *group_bh,
68 u32 bits_wanted, u32 min_bits,
69 u16 *bit_off, u16 *bits_found);
70 static int ocfs2_claim_suballoc_bits(struct ocfs2_super *osb,
71 struct ocfs2_alloc_context *ac,
72 handle_t *handle,
73 u32 bits_wanted,
74 u32 min_bits,
75 u16 *bit_off,
76 unsigned int *num_bits,
77 u64 *bg_blkno);
78 static int ocfs2_test_bg_bit_allocatable(struct buffer_head *bg_bh,
79 int nr);
80 static inline int ocfs2_block_group_set_bits(handle_t *handle,
81 struct inode *alloc_inode,
82 struct ocfs2_group_desc *bg,
83 struct buffer_head *group_bh,
84 unsigned int bit_off,
85 unsigned int num_bits);
86 static inline int ocfs2_block_group_clear_bits(handle_t *handle,
87 struct inode *alloc_inode,
88 struct ocfs2_group_desc *bg,
89 struct buffer_head *group_bh,
90 unsigned int bit_off,
91 unsigned int num_bits);
92
93 static int ocfs2_relink_block_group(handle_t *handle,
94 struct inode *alloc_inode,
95 struct buffer_head *fe_bh,
96 struct buffer_head *bg_bh,
97 struct buffer_head *prev_bg_bh,
98 u16 chain);
99 static inline int ocfs2_block_group_reasonably_empty(struct ocfs2_group_desc *bg,
100 u32 wanted);
101 static int ocfs2_free_suballoc_bits(handle_t *handle,
102 struct inode *alloc_inode,
103 struct buffer_head *alloc_bh,
104 unsigned int start_bit,
105 u64 bg_blkno,
106 unsigned int count);
107 static inline u64 ocfs2_which_suballoc_group(u64 block,
108 unsigned int bit);
109 static inline u32 ocfs2_desc_bitmap_to_cluster_off(struct inode *inode,
110 u64 bg_blkno,
111 u16 bg_bit_off);
112 static inline u64 ocfs2_which_cluster_group(struct inode *inode,
113 u32 cluster);
114 static inline void ocfs2_block_to_cluster_group(struct inode *inode,
115 u64 data_blkno,
116 u64 *bg_blkno,
117 u16 *bg_bit_off);
118
119 void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac)
120 {
121 struct inode *inode = ac->ac_inode;
122
123 if (inode) {
124 if (ac->ac_which != OCFS2_AC_USE_LOCAL)
125 ocfs2_meta_unlock(inode, 1);
126
127 mutex_unlock(&inode->i_mutex);
128
129 iput(inode);
130 }
131 if (ac->ac_bh)
132 brelse(ac->ac_bh);
133 kfree(ac);
134 }
135
136 static u32 ocfs2_bits_per_group(struct ocfs2_chain_list *cl)
137 {
138 return (u32)le16_to_cpu(cl->cl_cpg) * (u32)le16_to_cpu(cl->cl_bpc);
139 }
140
141 /* somewhat more expensive than our other checks, so use sparingly. */
142 static int ocfs2_check_group_descriptor(struct super_block *sb,
143 struct ocfs2_dinode *di,
144 struct ocfs2_group_desc *gd)
145 {
146 unsigned int max_bits;
147
148 if (!OCFS2_IS_VALID_GROUP_DESC(gd)) {
149 OCFS2_RO_ON_INVALID_GROUP_DESC(sb, gd);
150 return -EIO;
151 }
152
153 if (di->i_blkno != gd->bg_parent_dinode) {
154 ocfs2_error(sb, "Group descriptor # %llu has bad parent "
155 "pointer (%llu, expected %llu)",
156 (unsigned long long)le64_to_cpu(gd->bg_blkno),
157 (unsigned long long)le64_to_cpu(gd->bg_parent_dinode),
158 (unsigned long long)le64_to_cpu(di->i_blkno));
159 return -EIO;
160 }
161
162 max_bits = le16_to_cpu(di->id2.i_chain.cl_cpg) * le16_to_cpu(di->id2.i_chain.cl_bpc);
163 if (le16_to_cpu(gd->bg_bits) > max_bits) {
164 ocfs2_error(sb, "Group descriptor # %llu has bit count of %u",
165 (unsigned long long)le64_to_cpu(gd->bg_blkno),
166 le16_to_cpu(gd->bg_bits));
167 return -EIO;
168 }
169
170 if (le16_to_cpu(gd->bg_chain) >=
171 le16_to_cpu(di->id2.i_chain.cl_next_free_rec)) {
172 ocfs2_error(sb, "Group descriptor # %llu has bad chain %u",
173 (unsigned long long)le64_to_cpu(gd->bg_blkno),
174 le16_to_cpu(gd->bg_chain));
175 return -EIO;
176 }
177
178 if (le16_to_cpu(gd->bg_free_bits_count) > le16_to_cpu(gd->bg_bits)) {
179 ocfs2_error(sb, "Group descriptor # %llu has bit count %u but "
180 "claims that %u are free",
181 (unsigned long long)le64_to_cpu(gd->bg_blkno),
182 le16_to_cpu(gd->bg_bits),
183 le16_to_cpu(gd->bg_free_bits_count));
184 return -EIO;
185 }
186
187 if (le16_to_cpu(gd->bg_bits) > (8 * le16_to_cpu(gd->bg_size))) {
188 ocfs2_error(sb, "Group descriptor # %llu has bit count %u but "
189 "max bitmap bits of %u",
190 (unsigned long long)le64_to_cpu(gd->bg_blkno),
191 le16_to_cpu(gd->bg_bits),
192 8 * le16_to_cpu(gd->bg_size));
193 return -EIO;
194 }
195
196 return 0;
197 }
198
199 static int ocfs2_block_group_fill(handle_t *handle,
200 struct inode *alloc_inode,
201 struct buffer_head *bg_bh,
202 u64 group_blkno,
203 u16 my_chain,
204 struct ocfs2_chain_list *cl)
205 {
206 int status = 0;
207 struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
208 struct super_block * sb = alloc_inode->i_sb;
209
210 mlog_entry_void();
211
212 if (((unsigned long long) bg_bh->b_blocknr) != group_blkno) {
213 ocfs2_error(alloc_inode->i_sb, "group block (%llu) != "
214 "b_blocknr (%llu)",
215 (unsigned long long)group_blkno,
216 (unsigned long long) bg_bh->b_blocknr);
217 status = -EIO;
218 goto bail;
219 }
220
221 status = ocfs2_journal_access(handle,
222 alloc_inode,
223 bg_bh,
224 OCFS2_JOURNAL_ACCESS_CREATE);
225 if (status < 0) {
226 mlog_errno(status);
227 goto bail;
228 }
229
230 memset(bg, 0, sb->s_blocksize);
231 strcpy(bg->bg_signature, OCFS2_GROUP_DESC_SIGNATURE);
232 bg->bg_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
233 bg->bg_size = cpu_to_le16(ocfs2_group_bitmap_size(sb));
234 bg->bg_bits = cpu_to_le16(ocfs2_bits_per_group(cl));
235 bg->bg_chain = cpu_to_le16(my_chain);
236 bg->bg_next_group = cl->cl_recs[my_chain].c_blkno;
237 bg->bg_parent_dinode = cpu_to_le64(OCFS2_I(alloc_inode)->ip_blkno);
238 bg->bg_blkno = cpu_to_le64(group_blkno);
239 /* set the 1st bit in the bitmap to account for the descriptor block */
240 ocfs2_set_bit(0, (unsigned long *)bg->bg_bitmap);
241 bg->bg_free_bits_count = cpu_to_le16(le16_to_cpu(bg->bg_bits) - 1);
242
243 status = ocfs2_journal_dirty(handle, bg_bh);
244 if (status < 0)
245 mlog_errno(status);
246
247 /* There is no need to zero out or otherwise initialize the
248 * other blocks in a group - All valid FS metadata in a block
249 * group stores the superblock fs_generation value at
250 * allocation time. */
251
252 bail:
253 mlog_exit(status);
254 return status;
255 }
256
257 static inline u16 ocfs2_find_smallest_chain(struct ocfs2_chain_list *cl)
258 {
259 u16 curr, best;
260
261 best = curr = 0;
262 while (curr < le16_to_cpu(cl->cl_count)) {
263 if (le32_to_cpu(cl->cl_recs[best].c_total) >
264 le32_to_cpu(cl->cl_recs[curr].c_total))
265 best = curr;
266 curr++;
267 }
268 return best;
269 }
270
271 /*
272 * We expect the block group allocator to already be locked.
273 */
274 static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
275 struct inode *alloc_inode,
276 struct buffer_head *bh)
277 {
278 int status, credits;
279 struct ocfs2_dinode *fe = (struct ocfs2_dinode *) bh->b_data;
280 struct ocfs2_chain_list *cl;
281 struct ocfs2_alloc_context *ac = NULL;
282 handle_t *handle = NULL;
283 u32 bit_off, num_bits;
284 u16 alloc_rec;
285 u64 bg_blkno;
286 struct buffer_head *bg_bh = NULL;
287 struct ocfs2_group_desc *bg;
288
289 BUG_ON(ocfs2_is_cluster_bitmap(alloc_inode));
290
291 mlog_entry_void();
292
293 cl = &fe->id2.i_chain;
294 status = ocfs2_reserve_clusters(osb,
295 le16_to_cpu(cl->cl_cpg),
296 &ac);
297 if (status < 0) {
298 if (status != -ENOSPC)
299 mlog_errno(status);
300 goto bail;
301 }
302
303 credits = ocfs2_calc_group_alloc_credits(osb->sb,
304 le16_to_cpu(cl->cl_cpg));
305 handle = ocfs2_start_trans(osb, credits);
306 if (IS_ERR(handle)) {
307 status = PTR_ERR(handle);
308 handle = NULL;
309 mlog_errno(status);
310 goto bail;
311 }
312
313 status = ocfs2_claim_clusters(osb,
314 handle,
315 ac,
316 le16_to_cpu(cl->cl_cpg),
317 &bit_off,
318 &num_bits);
319 if (status < 0) {
320 if (status != -ENOSPC)
321 mlog_errno(status);
322 goto bail;
323 }
324
325 alloc_rec = ocfs2_find_smallest_chain(cl);
326
327 /* setup the group */
328 bg_blkno = ocfs2_clusters_to_blocks(osb->sb, bit_off);
329 mlog(0, "new descriptor, record %u, at block %llu\n",
330 alloc_rec, (unsigned long long)bg_blkno);
331
332 bg_bh = sb_getblk(osb->sb, bg_blkno);
333 if (!bg_bh) {
334 status = -EIO;
335 mlog_errno(status);
336 goto bail;
337 }
338 ocfs2_set_new_buffer_uptodate(alloc_inode, bg_bh);
339
340 status = ocfs2_block_group_fill(handle,
341 alloc_inode,
342 bg_bh,
343 bg_blkno,
344 alloc_rec,
345 cl);
346 if (status < 0) {
347 mlog_errno(status);
348 goto bail;
349 }
350
351 bg = (struct ocfs2_group_desc *) bg_bh->b_data;
352
353 status = ocfs2_journal_access(handle, alloc_inode,
354 bh, OCFS2_JOURNAL_ACCESS_WRITE);
355 if (status < 0) {
356 mlog_errno(status);
357 goto bail;
358 }
359
360 le32_add_cpu(&cl->cl_recs[alloc_rec].c_free,
361 le16_to_cpu(bg->bg_free_bits_count));
362 le32_add_cpu(&cl->cl_recs[alloc_rec].c_total, le16_to_cpu(bg->bg_bits));
363 cl->cl_recs[alloc_rec].c_blkno = cpu_to_le64(bg_blkno);
364 if (le16_to_cpu(cl->cl_next_free_rec) < le16_to_cpu(cl->cl_count))
365 le16_add_cpu(&cl->cl_next_free_rec, 1);
366
367 le32_add_cpu(&fe->id1.bitmap1.i_used, le16_to_cpu(bg->bg_bits) -
368 le16_to_cpu(bg->bg_free_bits_count));
369 le32_add_cpu(&fe->id1.bitmap1.i_total, le16_to_cpu(bg->bg_bits));
370 le32_add_cpu(&fe->i_clusters, le16_to_cpu(cl->cl_cpg));
371
372 status = ocfs2_journal_dirty(handle, bh);
373 if (status < 0) {
374 mlog_errno(status);
375 goto bail;
376 }
377
378 spin_lock(&OCFS2_I(alloc_inode)->ip_lock);
379 OCFS2_I(alloc_inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
380 fe->i_size = cpu_to_le64(ocfs2_clusters_to_bytes(alloc_inode->i_sb,
381 le32_to_cpu(fe->i_clusters)));
382 spin_unlock(&OCFS2_I(alloc_inode)->ip_lock);
383 i_size_write(alloc_inode, le64_to_cpu(fe->i_size));
384 alloc_inode->i_blocks =
385 ocfs2_align_bytes_to_sectors(i_size_read(alloc_inode));
386
387 status = 0;
388 bail:
389 if (handle)
390 ocfs2_commit_trans(osb, handle);
391
392 if (ac)
393 ocfs2_free_alloc_context(ac);
394
395 if (bg_bh)
396 brelse(bg_bh);
397
398 mlog_exit(status);
399 return status;
400 }
401
402 static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb,
403 struct ocfs2_alloc_context *ac,
404 int type,
405 u32 slot)
406 {
407 int status;
408 u32 bits_wanted = ac->ac_bits_wanted;
409 struct inode *alloc_inode;
410 struct buffer_head *bh = NULL;
411 struct ocfs2_dinode *fe;
412 u32 free_bits;
413
414 mlog_entry_void();
415
416 alloc_inode = ocfs2_get_system_file_inode(osb, type, slot);
417 if (!alloc_inode) {
418 mlog_errno(-EINVAL);
419 return -EINVAL;
420 }
421
422 mutex_lock(&alloc_inode->i_mutex);
423
424 status = ocfs2_meta_lock(alloc_inode, &bh, 1);
425 if (status < 0) {
426 mutex_unlock(&alloc_inode->i_mutex);
427 iput(alloc_inode);
428
429 mlog_errno(status);
430 return status;
431 }
432
433 ac->ac_inode = alloc_inode;
434
435 fe = (struct ocfs2_dinode *) bh->b_data;
436 if (!OCFS2_IS_VALID_DINODE(fe)) {
437 OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
438 status = -EIO;
439 goto bail;
440 }
441 if (!(fe->i_flags & cpu_to_le32(OCFS2_CHAIN_FL))) {
442 ocfs2_error(alloc_inode->i_sb, "Invalid chain allocator %llu",
443 (unsigned long long)le64_to_cpu(fe->i_blkno));
444 status = -EIO;
445 goto bail;
446 }
447
448 free_bits = le32_to_cpu(fe->id1.bitmap1.i_total) -
449 le32_to_cpu(fe->id1.bitmap1.i_used);
450
451 if (bits_wanted > free_bits) {
452 /* cluster bitmap never grows */
453 if (ocfs2_is_cluster_bitmap(alloc_inode)) {
454 mlog(0, "Disk Full: wanted=%u, free_bits=%u\n",
455 bits_wanted, free_bits);
456 status = -ENOSPC;
457 goto bail;
458 }
459
460 status = ocfs2_block_group_alloc(osb, alloc_inode, bh);
461 if (status < 0) {
462 if (status != -ENOSPC)
463 mlog_errno(status);
464 goto bail;
465 }
466 atomic_inc(&osb->alloc_stats.bg_extends);
467
468 /* You should never ask for this much metadata */
469 BUG_ON(bits_wanted >
470 (le32_to_cpu(fe->id1.bitmap1.i_total)
471 - le32_to_cpu(fe->id1.bitmap1.i_used)));
472 }
473
474 get_bh(bh);
475 ac->ac_bh = bh;
476 bail:
477 if (bh)
478 brelse(bh);
479
480 mlog_exit(status);
481 return status;
482 }
483
484 int ocfs2_reserve_new_metadata(struct ocfs2_super *osb,
485 struct ocfs2_dinode *fe,
486 struct ocfs2_alloc_context **ac)
487 {
488 int status;
489 u32 slot;
490
491 *ac = kcalloc(1, sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
492 if (!(*ac)) {
493 status = -ENOMEM;
494 mlog_errno(status);
495 goto bail;
496 }
497
498 (*ac)->ac_bits_wanted = ocfs2_extend_meta_needed(fe);
499 (*ac)->ac_which = OCFS2_AC_USE_META;
500
501 #ifndef OCFS2_USE_ALL_METADATA_SUBALLOCATORS
502 slot = 0;
503 #else
504 slot = osb->slot_num;
505 #endif
506
507 (*ac)->ac_group_search = ocfs2_block_group_search;
508
509 status = ocfs2_reserve_suballoc_bits(osb, (*ac),
510 EXTENT_ALLOC_SYSTEM_INODE, slot);
511 if (status < 0) {
512 if (status != -ENOSPC)
513 mlog_errno(status);
514 goto bail;
515 }
516
517 status = 0;
518 bail:
519 if ((status < 0) && *ac) {
520 ocfs2_free_alloc_context(*ac);
521 *ac = NULL;
522 }
523
524 mlog_exit(status);
525 return status;
526 }
527
528 int ocfs2_reserve_new_inode(struct ocfs2_super *osb,
529 struct ocfs2_alloc_context **ac)
530 {
531 int status;
532
533 *ac = kcalloc(1, sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
534 if (!(*ac)) {
535 status = -ENOMEM;
536 mlog_errno(status);
537 goto bail;
538 }
539
540 (*ac)->ac_bits_wanted = 1;
541 (*ac)->ac_which = OCFS2_AC_USE_INODE;
542
543 (*ac)->ac_group_search = ocfs2_block_group_search;
544
545 status = ocfs2_reserve_suballoc_bits(osb, *ac,
546 INODE_ALLOC_SYSTEM_INODE,
547 osb->slot_num);
548 if (status < 0) {
549 if (status != -ENOSPC)
550 mlog_errno(status);
551 goto bail;
552 }
553
554 status = 0;
555 bail:
556 if ((status < 0) && *ac) {
557 ocfs2_free_alloc_context(*ac);
558 *ac = NULL;
559 }
560
561 mlog_exit(status);
562 return status;
563 }
564
565 /* local alloc code has to do the same thing, so rather than do this
566 * twice.. */
567 int ocfs2_reserve_cluster_bitmap_bits(struct ocfs2_super *osb,
568 struct ocfs2_alloc_context *ac)
569 {
570 int status;
571
572 ac->ac_which = OCFS2_AC_USE_MAIN;
573 ac->ac_group_search = ocfs2_cluster_group_search;
574
575 status = ocfs2_reserve_suballoc_bits(osb, ac,
576 GLOBAL_BITMAP_SYSTEM_INODE,
577 OCFS2_INVALID_SLOT);
578 if (status < 0 && status != -ENOSPC) {
579 mlog_errno(status);
580 goto bail;
581 }
582
583 bail:
584 return status;
585 }
586
587 /* Callers don't need to care which bitmap (local alloc or main) to
588 * use so we figure it out for them, but unfortunately this clutters
589 * things a bit. */
590 int ocfs2_reserve_clusters(struct ocfs2_super *osb,
591 u32 bits_wanted,
592 struct ocfs2_alloc_context **ac)
593 {
594 int status;
595
596 mlog_entry_void();
597
598 *ac = kcalloc(1, sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
599 if (!(*ac)) {
600 status = -ENOMEM;
601 mlog_errno(status);
602 goto bail;
603 }
604
605 (*ac)->ac_bits_wanted = bits_wanted;
606
607 status = -ENOSPC;
608 if (ocfs2_alloc_should_use_local(osb, bits_wanted)) {
609 status = ocfs2_reserve_local_alloc_bits(osb,
610 bits_wanted,
611 *ac);
612 if ((status < 0) && (status != -ENOSPC)) {
613 mlog_errno(status);
614 goto bail;
615 } else if (status == -ENOSPC) {
616 /* reserve_local_bits will return enospc with
617 * the local alloc inode still locked, so we
618 * can change this safely here. */
619 mlog(0, "Disabling local alloc\n");
620 /* We set to OCFS2_LA_DISABLED so that umount
621 * can clean up what's left of the local
622 * allocation */
623 osb->local_alloc_state = OCFS2_LA_DISABLED;
624 }
625 }
626
627 if (status == -ENOSPC) {
628 status = ocfs2_reserve_cluster_bitmap_bits(osb, *ac);
629 if (status < 0) {
630 if (status != -ENOSPC)
631 mlog_errno(status);
632 goto bail;
633 }
634 }
635
636 status = 0;
637 bail:
638 if ((status < 0) && *ac) {
639 ocfs2_free_alloc_context(*ac);
640 *ac = NULL;
641 }
642
643 mlog_exit(status);
644 return status;
645 }
646
647 /*
648 * More or less lifted from ext3. I'll leave their description below:
649 *
650 * "For ext3 allocations, we must not reuse any blocks which are
651 * allocated in the bitmap buffer's "last committed data" copy. This
652 * prevents deletes from freeing up the page for reuse until we have
653 * committed the delete transaction.
654 *
655 * If we didn't do this, then deleting something and reallocating it as
656 * data would allow the old block to be overwritten before the
657 * transaction committed (because we force data to disk before commit).
658 * This would lead to corruption if we crashed between overwriting the
659 * data and committing the delete.
660 *
661 * @@@ We may want to make this allocation behaviour conditional on
662 * data-writes at some point, and disable it for metadata allocations or
663 * sync-data inodes."
664 *
665 * Note: OCFS2 already does this differently for metadata vs data
666 * allocations, as those bitmaps are seperate and undo access is never
667 * called on a metadata group descriptor.
668 */
669 static int ocfs2_test_bg_bit_allocatable(struct buffer_head *bg_bh,
670 int nr)
671 {
672 struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
673
674 if (ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap))
675 return 0;
676 if (!buffer_jbd(bg_bh) || !bh2jh(bg_bh)->b_committed_data)
677 return 1;
678
679 bg = (struct ocfs2_group_desc *) bh2jh(bg_bh)->b_committed_data;
680 return !ocfs2_test_bit(nr, (unsigned long *)bg->bg_bitmap);
681 }
682
683 static int ocfs2_block_group_find_clear_bits(struct ocfs2_super *osb,
684 struct buffer_head *bg_bh,
685 unsigned int bits_wanted,
686 unsigned int total_bits,
687 u16 *bit_off,
688 u16 *bits_found)
689 {
690 void *bitmap;
691 u16 best_offset, best_size;
692 int offset, start, found, status = 0;
693 struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
694
695 if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
696 OCFS2_RO_ON_INVALID_GROUP_DESC(osb->sb, bg);
697 return -EIO;
698 }
699
700 found = start = best_offset = best_size = 0;
701 bitmap = bg->bg_bitmap;
702
703 while((offset = ocfs2_find_next_zero_bit(bitmap, total_bits, start)) != -1) {
704 if (offset == total_bits)
705 break;
706
707 if (!ocfs2_test_bg_bit_allocatable(bg_bh, offset)) {
708 /* We found a zero, but we can't use it as it
709 * hasn't been put to disk yet! */
710 found = 0;
711 start = offset + 1;
712 } else if (offset == start) {
713 /* we found a zero */
714 found++;
715 /* move start to the next bit to test */
716 start++;
717 } else {
718 /* got a zero after some ones */
719 found = 1;
720 start = offset + 1;
721 }
722 if (found > best_size) {
723 best_size = found;
724 best_offset = start - found;
725 }
726 /* we got everything we needed */
727 if (found == bits_wanted) {
728 /* mlog(0, "Found it all!\n"); */
729 break;
730 }
731 }
732
733 /* XXX: I think the first clause is equivalent to the second
734 * - jlbec */
735 if (found == bits_wanted) {
736 *bit_off = start - found;
737 *bits_found = found;
738 } else if (best_size) {
739 *bit_off = best_offset;
740 *bits_found = best_size;
741 } else {
742 status = -ENOSPC;
743 /* No error log here -- see the comment above
744 * ocfs2_test_bg_bit_allocatable */
745 }
746
747 return status;
748 }
749
750 static inline int ocfs2_block_group_set_bits(handle_t *handle,
751 struct inode *alloc_inode,
752 struct ocfs2_group_desc *bg,
753 struct buffer_head *group_bh,
754 unsigned int bit_off,
755 unsigned int num_bits)
756 {
757 int status;
758 void *bitmap = bg->bg_bitmap;
759 int journal_type = OCFS2_JOURNAL_ACCESS_WRITE;
760
761 mlog_entry_void();
762
763 if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
764 OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, bg);
765 status = -EIO;
766 goto bail;
767 }
768 BUG_ON(le16_to_cpu(bg->bg_free_bits_count) < num_bits);
769
770 mlog(0, "block_group_set_bits: off = %u, num = %u\n", bit_off,
771 num_bits);
772
773 if (ocfs2_is_cluster_bitmap(alloc_inode))
774 journal_type = OCFS2_JOURNAL_ACCESS_UNDO;
775
776 status = ocfs2_journal_access(handle,
777 alloc_inode,
778 group_bh,
779 journal_type);
780 if (status < 0) {
781 mlog_errno(status);
782 goto bail;
783 }
784
785 le16_add_cpu(&bg->bg_free_bits_count, -num_bits);
786
787 while(num_bits--)
788 ocfs2_set_bit(bit_off++, bitmap);
789
790 status = ocfs2_journal_dirty(handle,
791 group_bh);
792 if (status < 0) {
793 mlog_errno(status);
794 goto bail;
795 }
796
797 bail:
798 mlog_exit(status);
799 return status;
800 }
801
802 /* find the one with the most empty bits */
803 static inline u16 ocfs2_find_victim_chain(struct ocfs2_chain_list *cl)
804 {
805 u16 curr, best;
806
807 BUG_ON(!cl->cl_next_free_rec);
808
809 best = curr = 0;
810 while (curr < le16_to_cpu(cl->cl_next_free_rec)) {
811 if (le32_to_cpu(cl->cl_recs[curr].c_free) >
812 le32_to_cpu(cl->cl_recs[best].c_free))
813 best = curr;
814 curr++;
815 }
816
817 BUG_ON(best >= le16_to_cpu(cl->cl_next_free_rec));
818 return best;
819 }
820
821 static int ocfs2_relink_block_group(handle_t *handle,
822 struct inode *alloc_inode,
823 struct buffer_head *fe_bh,
824 struct buffer_head *bg_bh,
825 struct buffer_head *prev_bg_bh,
826 u16 chain)
827 {
828 int status;
829 /* there is a really tiny chance the journal calls could fail,
830 * but we wouldn't want inconsistent blocks in *any* case. */
831 u64 fe_ptr, bg_ptr, prev_bg_ptr;
832 struct ocfs2_dinode *fe = (struct ocfs2_dinode *) fe_bh->b_data;
833 struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
834 struct ocfs2_group_desc *prev_bg = (struct ocfs2_group_desc *) prev_bg_bh->b_data;
835
836 if (!OCFS2_IS_VALID_DINODE(fe)) {
837 OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
838 status = -EIO;
839 goto out;
840 }
841 if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
842 OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, bg);
843 status = -EIO;
844 goto out;
845 }
846 if (!OCFS2_IS_VALID_GROUP_DESC(prev_bg)) {
847 OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, prev_bg);
848 status = -EIO;
849 goto out;
850 }
851
852 mlog(0, "Suballoc %llu, chain %u, move group %llu to top, prev = %llu\n",
853 (unsigned long long)fe->i_blkno, chain,
854 (unsigned long long)bg->bg_blkno,
855 (unsigned long long)prev_bg->bg_blkno);
856
857 fe_ptr = le64_to_cpu(fe->id2.i_chain.cl_recs[chain].c_blkno);
858 bg_ptr = le64_to_cpu(bg->bg_next_group);
859 prev_bg_ptr = le64_to_cpu(prev_bg->bg_next_group);
860
861 status = ocfs2_journal_access(handle, alloc_inode, prev_bg_bh,
862 OCFS2_JOURNAL_ACCESS_WRITE);
863 if (status < 0) {
864 mlog_errno(status);
865 goto out_rollback;
866 }
867
868 prev_bg->bg_next_group = bg->bg_next_group;
869
870 status = ocfs2_journal_dirty(handle, prev_bg_bh);
871 if (status < 0) {
872 mlog_errno(status);
873 goto out_rollback;
874 }
875
876 status = ocfs2_journal_access(handle, alloc_inode, bg_bh,
877 OCFS2_JOURNAL_ACCESS_WRITE);
878 if (status < 0) {
879 mlog_errno(status);
880 goto out_rollback;
881 }
882
883 bg->bg_next_group = fe->id2.i_chain.cl_recs[chain].c_blkno;
884
885 status = ocfs2_journal_dirty(handle, bg_bh);
886 if (status < 0) {
887 mlog_errno(status);
888 goto out_rollback;
889 }
890
891 status = ocfs2_journal_access(handle, alloc_inode, fe_bh,
892 OCFS2_JOURNAL_ACCESS_WRITE);
893 if (status < 0) {
894 mlog_errno(status);
895 goto out_rollback;
896 }
897
898 fe->id2.i_chain.cl_recs[chain].c_blkno = bg->bg_blkno;
899
900 status = ocfs2_journal_dirty(handle, fe_bh);
901 if (status < 0) {
902 mlog_errno(status);
903 goto out_rollback;
904 }
905
906 status = 0;
907 out_rollback:
908 if (status < 0) {
909 fe->id2.i_chain.cl_recs[chain].c_blkno = cpu_to_le64(fe_ptr);
910 bg->bg_next_group = cpu_to_le64(bg_ptr);
911 prev_bg->bg_next_group = cpu_to_le64(prev_bg_ptr);
912 }
913 out:
914 mlog_exit(status);
915 return status;
916 }
917
918 static inline int ocfs2_block_group_reasonably_empty(struct ocfs2_group_desc *bg,
919 u32 wanted)
920 {
921 return le16_to_cpu(bg->bg_free_bits_count) > wanted;
922 }
923
924 /* return 0 on success, -ENOSPC to keep searching and any other < 0
925 * value on error. */
926 static int ocfs2_cluster_group_search(struct inode *inode,
927 struct buffer_head *group_bh,
928 u32 bits_wanted, u32 min_bits,
929 u16 *bit_off, u16 *bits_found)
930 {
931 int search = -ENOSPC;
932 int ret;
933 struct ocfs2_group_desc *gd = (struct ocfs2_group_desc *) group_bh->b_data;
934 u16 tmp_off, tmp_found;
935 unsigned int max_bits, gd_cluster_off;
936
937 BUG_ON(!ocfs2_is_cluster_bitmap(inode));
938
939 if (gd->bg_free_bits_count) {
940 max_bits = le16_to_cpu(gd->bg_bits);
941
942 /* Tail groups in cluster bitmaps which aren't cpg
943 * aligned are prone to partial extention by a failed
944 * fs resize. If the file system resize never got to
945 * update the dinode cluster count, then we don't want
946 * to trust any clusters past it, regardless of what
947 * the group descriptor says. */
948 gd_cluster_off = ocfs2_blocks_to_clusters(inode->i_sb,
949 le64_to_cpu(gd->bg_blkno));
950 if ((gd_cluster_off + max_bits) >
951 OCFS2_I(inode)->ip_clusters) {
952 max_bits = OCFS2_I(inode)->ip_clusters - gd_cluster_off;
953 mlog(0, "Desc %llu, bg_bits %u, clusters %u, use %u\n",
954 (unsigned long long)le64_to_cpu(gd->bg_blkno),
955 le16_to_cpu(gd->bg_bits),
956 OCFS2_I(inode)->ip_clusters, max_bits);
957 }
958
959 ret = ocfs2_block_group_find_clear_bits(OCFS2_SB(inode->i_sb),
960 group_bh, bits_wanted,
961 max_bits,
962 &tmp_off, &tmp_found);
963 if (ret)
964 return ret;
965
966 /* ocfs2_block_group_find_clear_bits() might
967 * return success, but we still want to return
968 * -ENOSPC unless it found the minimum number
969 * of bits. */
970 if (min_bits <= tmp_found) {
971 *bit_off = tmp_off;
972 *bits_found = tmp_found;
973 search = 0; /* success */
974 }
975 }
976
977 return search;
978 }
979
980 static int ocfs2_block_group_search(struct inode *inode,
981 struct buffer_head *group_bh,
982 u32 bits_wanted, u32 min_bits,
983 u16 *bit_off, u16 *bits_found)
984 {
985 int ret = -ENOSPC;
986 struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) group_bh->b_data;
987
988 BUG_ON(min_bits != 1);
989 BUG_ON(ocfs2_is_cluster_bitmap(inode));
990
991 if (bg->bg_free_bits_count)
992 ret = ocfs2_block_group_find_clear_bits(OCFS2_SB(inode->i_sb),
993 group_bh, bits_wanted,
994 le16_to_cpu(bg->bg_bits),
995 bit_off, bits_found);
996
997 return ret;
998 }
999
1000 static int ocfs2_alloc_dinode_update_counts(struct inode *inode,
1001 handle_t *handle,
1002 struct buffer_head *di_bh,
1003 u32 num_bits,
1004 u16 chain)
1005 {
1006 int ret;
1007 u32 tmp_used;
1008 struct ocfs2_dinode *di = (struct ocfs2_dinode *) di_bh->b_data;
1009 struct ocfs2_chain_list *cl = (struct ocfs2_chain_list *) &di->id2.i_chain;
1010
1011 ret = ocfs2_journal_access(handle, inode, di_bh,
1012 OCFS2_JOURNAL_ACCESS_WRITE);
1013 if (ret < 0) {
1014 mlog_errno(ret);
1015 goto out;
1016 }
1017
1018 tmp_used = le32_to_cpu(di->id1.bitmap1.i_used);
1019 di->id1.bitmap1.i_used = cpu_to_le32(num_bits + tmp_used);
1020 le32_add_cpu(&cl->cl_recs[chain].c_free, -num_bits);
1021
1022 ret = ocfs2_journal_dirty(handle, di_bh);
1023 if (ret < 0)
1024 mlog_errno(ret);
1025
1026 out:
1027 return ret;
1028 }
1029
1030 static int ocfs2_search_one_group(struct ocfs2_alloc_context *ac,
1031 handle_t *handle,
1032 u32 bits_wanted,
1033 u32 min_bits,
1034 u16 *bit_off,
1035 unsigned int *num_bits,
1036 u64 gd_blkno,
1037 u16 *bits_left)
1038 {
1039 int ret;
1040 u16 found;
1041 struct buffer_head *group_bh = NULL;
1042 struct ocfs2_group_desc *gd;
1043 struct inode *alloc_inode = ac->ac_inode;
1044
1045 ret = ocfs2_read_block(OCFS2_SB(alloc_inode->i_sb), gd_blkno,
1046 &group_bh, OCFS2_BH_CACHED, alloc_inode);
1047 if (ret < 0) {
1048 mlog_errno(ret);
1049 return ret;
1050 }
1051
1052 gd = (struct ocfs2_group_desc *) group_bh->b_data;
1053 if (!OCFS2_IS_VALID_GROUP_DESC(gd)) {
1054 OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, gd);
1055 ret = -EIO;
1056 goto out;
1057 }
1058
1059 ret = ac->ac_group_search(alloc_inode, group_bh, bits_wanted, min_bits,
1060 bit_off, &found);
1061 if (ret < 0) {
1062 if (ret != -ENOSPC)
1063 mlog_errno(ret);
1064 goto out;
1065 }
1066
1067 *num_bits = found;
1068
1069 ret = ocfs2_alloc_dinode_update_counts(alloc_inode, handle, ac->ac_bh,
1070 *num_bits,
1071 le16_to_cpu(gd->bg_chain));
1072 if (ret < 0) {
1073 mlog_errno(ret);
1074 goto out;
1075 }
1076
1077 ret = ocfs2_block_group_set_bits(handle, alloc_inode, gd, group_bh,
1078 *bit_off, *num_bits);
1079 if (ret < 0)
1080 mlog_errno(ret);
1081
1082 *bits_left = le16_to_cpu(gd->bg_free_bits_count);
1083
1084 out:
1085 brelse(group_bh);
1086
1087 return ret;
1088 }
1089
1090 static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
1091 handle_t *handle,
1092 u32 bits_wanted,
1093 u32 min_bits,
1094 u16 *bit_off,
1095 unsigned int *num_bits,
1096 u64 *bg_blkno,
1097 u16 *bits_left)
1098 {
1099 int status;
1100 u16 chain, tmp_bits;
1101 u32 tmp_used;
1102 u64 next_group;
1103 struct inode *alloc_inode = ac->ac_inode;
1104 struct buffer_head *group_bh = NULL;
1105 struct buffer_head *prev_group_bh = NULL;
1106 struct ocfs2_dinode *fe = (struct ocfs2_dinode *) ac->ac_bh->b_data;
1107 struct ocfs2_chain_list *cl = (struct ocfs2_chain_list *) &fe->id2.i_chain;
1108 struct ocfs2_group_desc *bg;
1109
1110 chain = ac->ac_chain;
1111 mlog(0, "trying to alloc %u bits from chain %u, inode %llu\n",
1112 bits_wanted, chain,
1113 (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno);
1114
1115 status = ocfs2_read_block(OCFS2_SB(alloc_inode->i_sb),
1116 le64_to_cpu(cl->cl_recs[chain].c_blkno),
1117 &group_bh, OCFS2_BH_CACHED, alloc_inode);
1118 if (status < 0) {
1119 mlog_errno(status);
1120 goto bail;
1121 }
1122 bg = (struct ocfs2_group_desc *) group_bh->b_data;
1123 status = ocfs2_check_group_descriptor(alloc_inode->i_sb, fe, bg);
1124 if (status) {
1125 mlog_errno(status);
1126 goto bail;
1127 }
1128
1129 status = -ENOSPC;
1130 /* for now, the chain search is a bit simplistic. We just use
1131 * the 1st group with any empty bits. */
1132 while ((status = ac->ac_group_search(alloc_inode, group_bh,
1133 bits_wanted, min_bits, bit_off,
1134 &tmp_bits)) == -ENOSPC) {
1135 if (!bg->bg_next_group)
1136 break;
1137
1138 if (prev_group_bh) {
1139 brelse(prev_group_bh);
1140 prev_group_bh = NULL;
1141 }
1142 next_group = le64_to_cpu(bg->bg_next_group);
1143 prev_group_bh = group_bh;
1144 group_bh = NULL;
1145 status = ocfs2_read_block(OCFS2_SB(alloc_inode->i_sb),
1146 next_group, &group_bh,
1147 OCFS2_BH_CACHED, alloc_inode);
1148 if (status < 0) {
1149 mlog_errno(status);
1150 goto bail;
1151 }
1152 bg = (struct ocfs2_group_desc *) group_bh->b_data;
1153 status = ocfs2_check_group_descriptor(alloc_inode->i_sb, fe, bg);
1154 if (status) {
1155 mlog_errno(status);
1156 goto bail;
1157 }
1158 }
1159 if (status < 0) {
1160 if (status != -ENOSPC)
1161 mlog_errno(status);
1162 goto bail;
1163 }
1164
1165 mlog(0, "alloc succeeds: we give %u bits from block group %llu\n",
1166 tmp_bits, (unsigned long long)bg->bg_blkno);
1167
1168 *num_bits = tmp_bits;
1169
1170 BUG_ON(*num_bits == 0);
1171
1172 /*
1173 * Keep track of previous block descriptor read. When
1174 * we find a target, if we have read more than X
1175 * number of descriptors, and the target is reasonably
1176 * empty, relink him to top of his chain.
1177 *
1178 * We've read 0 extra blocks and only send one more to
1179 * the transaction, yet the next guy to search has a
1180 * much easier time.
1181 *
1182 * Do this *after* figuring out how many bits we're taking out
1183 * of our target group.
1184 */
1185 if (ac->ac_allow_chain_relink &&
1186 (prev_group_bh) &&
1187 (ocfs2_block_group_reasonably_empty(bg, *num_bits))) {
1188 status = ocfs2_relink_block_group(handle, alloc_inode,
1189 ac->ac_bh, group_bh,
1190 prev_group_bh, chain);
1191 if (status < 0) {
1192 mlog_errno(status);
1193 goto bail;
1194 }
1195 }
1196
1197 /* Ok, claim our bits now: set the info on dinode, chainlist
1198 * and then the group */
1199 status = ocfs2_journal_access(handle,
1200 alloc_inode,
1201 ac->ac_bh,
1202 OCFS2_JOURNAL_ACCESS_WRITE);
1203 if (status < 0) {
1204 mlog_errno(status);
1205 goto bail;
1206 }
1207
1208 tmp_used = le32_to_cpu(fe->id1.bitmap1.i_used);
1209 fe->id1.bitmap1.i_used = cpu_to_le32(*num_bits + tmp_used);
1210 le32_add_cpu(&cl->cl_recs[chain].c_free, -(*num_bits));
1211
1212 status = ocfs2_journal_dirty(handle,
1213 ac->ac_bh);
1214 if (status < 0) {
1215 mlog_errno(status);
1216 goto bail;
1217 }
1218
1219 status = ocfs2_block_group_set_bits(handle,
1220 alloc_inode,
1221 bg,
1222 group_bh,
1223 *bit_off,
1224 *num_bits);
1225 if (status < 0) {
1226 mlog_errno(status);
1227 goto bail;
1228 }
1229
1230 mlog(0, "Allocated %u bits from suballocator %llu\n", *num_bits,
1231 (unsigned long long)fe->i_blkno);
1232
1233 *bg_blkno = le64_to_cpu(bg->bg_blkno);
1234 *bits_left = le16_to_cpu(bg->bg_free_bits_count);
1235 bail:
1236 if (group_bh)
1237 brelse(group_bh);
1238 if (prev_group_bh)
1239 brelse(prev_group_bh);
1240
1241 mlog_exit(status);
1242 return status;
1243 }
1244
1245 /* will give out up to bits_wanted contiguous bits. */
1246 static int ocfs2_claim_suballoc_bits(struct ocfs2_super *osb,
1247 struct ocfs2_alloc_context *ac,
1248 handle_t *handle,
1249 u32 bits_wanted,
1250 u32 min_bits,
1251 u16 *bit_off,
1252 unsigned int *num_bits,
1253 u64 *bg_blkno)
1254 {
1255 int status;
1256 u16 victim, i;
1257 u16 bits_left = 0;
1258 u64 hint_blkno = ac->ac_last_group;
1259 struct ocfs2_chain_list *cl;
1260 struct ocfs2_dinode *fe;
1261
1262 mlog_entry_void();
1263
1264 BUG_ON(ac->ac_bits_given >= ac->ac_bits_wanted);
1265 BUG_ON(bits_wanted > (ac->ac_bits_wanted - ac->ac_bits_given));
1266 BUG_ON(!ac->ac_bh);
1267
1268 fe = (struct ocfs2_dinode *) ac->ac_bh->b_data;
1269 if (!OCFS2_IS_VALID_DINODE(fe)) {
1270 OCFS2_RO_ON_INVALID_DINODE(osb->sb, fe);
1271 status = -EIO;
1272 goto bail;
1273 }
1274 if (le32_to_cpu(fe->id1.bitmap1.i_used) >=
1275 le32_to_cpu(fe->id1.bitmap1.i_total)) {
1276 ocfs2_error(osb->sb, "Chain allocator dinode %llu has %u used "
1277 "bits but only %u total.",
1278 (unsigned long long)le64_to_cpu(fe->i_blkno),
1279 le32_to_cpu(fe->id1.bitmap1.i_used),
1280 le32_to_cpu(fe->id1.bitmap1.i_total));
1281 status = -EIO;
1282 goto bail;
1283 }
1284
1285 if (hint_blkno) {
1286 /* Attempt to short-circuit the usual search mechanism
1287 * by jumping straight to the most recently used
1288 * allocation group. This helps us mantain some
1289 * contiguousness across allocations. */
1290 status = ocfs2_search_one_group(ac, handle, bits_wanted,
1291 min_bits, bit_off, num_bits,
1292 hint_blkno, &bits_left);
1293 if (!status) {
1294 /* Be careful to update *bg_blkno here as the
1295 * caller is expecting it to be filled in, and
1296 * ocfs2_search_one_group() won't do that for
1297 * us. */
1298 *bg_blkno = hint_blkno;
1299 goto set_hint;
1300 }
1301 if (status < 0 && status != -ENOSPC) {
1302 mlog_errno(status);
1303 goto bail;
1304 }
1305 }
1306
1307 cl = (struct ocfs2_chain_list *) &fe->id2.i_chain;
1308
1309 victim = ocfs2_find_victim_chain(cl);
1310 ac->ac_chain = victim;
1311 ac->ac_allow_chain_relink = 1;
1312
1313 status = ocfs2_search_chain(ac, handle, bits_wanted, min_bits, bit_off,
1314 num_bits, bg_blkno, &bits_left);
1315 if (!status)
1316 goto set_hint;
1317 if (status < 0 && status != -ENOSPC) {
1318 mlog_errno(status);
1319 goto bail;
1320 }
1321
1322 mlog(0, "Search of victim chain %u came up with nothing, "
1323 "trying all chains now.\n", victim);
1324
1325 /* If we didn't pick a good victim, then just default to
1326 * searching each chain in order. Don't allow chain relinking
1327 * because we only calculate enough journal credits for one
1328 * relink per alloc. */
1329 ac->ac_allow_chain_relink = 0;
1330 for (i = 0; i < le16_to_cpu(cl->cl_next_free_rec); i ++) {
1331 if (i == victim)
1332 continue;
1333 if (!cl->cl_recs[i].c_free)
1334 continue;
1335
1336 ac->ac_chain = i;
1337 status = ocfs2_search_chain(ac, handle, bits_wanted, min_bits,
1338 bit_off, num_bits, bg_blkno,
1339 &bits_left);
1340 if (!status)
1341 break;
1342 if (status < 0 && status != -ENOSPC) {
1343 mlog_errno(status);
1344 goto bail;
1345 }
1346 }
1347
1348 set_hint:
1349 if (status != -ENOSPC) {
1350 /* If the next search of this group is not likely to
1351 * yield a suitable extent, then we reset the last
1352 * group hint so as to not waste a disk read */
1353 if (bits_left < min_bits)
1354 ac->ac_last_group = 0;
1355 else
1356 ac->ac_last_group = *bg_blkno;
1357 }
1358
1359 bail:
1360 mlog_exit(status);
1361 return status;
1362 }
1363
1364 int ocfs2_claim_metadata(struct ocfs2_super *osb,
1365 handle_t *handle,
1366 struct ocfs2_alloc_context *ac,
1367 u32 bits_wanted,
1368 u16 *suballoc_bit_start,
1369 unsigned int *num_bits,
1370 u64 *blkno_start)
1371 {
1372 int status;
1373 u64 bg_blkno;
1374
1375 BUG_ON(!ac);
1376 BUG_ON(ac->ac_bits_wanted < (ac->ac_bits_given + bits_wanted));
1377 BUG_ON(ac->ac_which != OCFS2_AC_USE_META);
1378
1379 status = ocfs2_claim_suballoc_bits(osb,
1380 ac,
1381 handle,
1382 bits_wanted,
1383 1,
1384 suballoc_bit_start,
1385 num_bits,
1386 &bg_blkno);
1387 if (status < 0) {
1388 mlog_errno(status);
1389 goto bail;
1390 }
1391 atomic_inc(&osb->alloc_stats.bg_allocs);
1392
1393 *blkno_start = bg_blkno + (u64) *suballoc_bit_start;
1394 ac->ac_bits_given += (*num_bits);
1395 status = 0;
1396 bail:
1397 mlog_exit(status);
1398 return status;
1399 }
1400
1401 int ocfs2_claim_new_inode(struct ocfs2_super *osb,
1402 handle_t *handle,
1403 struct ocfs2_alloc_context *ac,
1404 u16 *suballoc_bit,
1405 u64 *fe_blkno)
1406 {
1407 int status;
1408 unsigned int num_bits;
1409 u64 bg_blkno;
1410
1411 mlog_entry_void();
1412
1413 BUG_ON(!ac);
1414 BUG_ON(ac->ac_bits_given != 0);
1415 BUG_ON(ac->ac_bits_wanted != 1);
1416 BUG_ON(ac->ac_which != OCFS2_AC_USE_INODE);
1417
1418 status = ocfs2_claim_suballoc_bits(osb,
1419 ac,
1420 handle,
1421 1,
1422 1,
1423 suballoc_bit,
1424 &num_bits,
1425 &bg_blkno);
1426 if (status < 0) {
1427 mlog_errno(status);
1428 goto bail;
1429 }
1430 atomic_inc(&osb->alloc_stats.bg_allocs);
1431
1432 BUG_ON(num_bits != 1);
1433
1434 *fe_blkno = bg_blkno + (u64) (*suballoc_bit);
1435 ac->ac_bits_given++;
1436 status = 0;
1437 bail:
1438 mlog_exit(status);
1439 return status;
1440 }
1441
1442 /* translate a group desc. blkno and it's bitmap offset into
1443 * disk cluster offset. */
1444 static inline u32 ocfs2_desc_bitmap_to_cluster_off(struct inode *inode,
1445 u64 bg_blkno,
1446 u16 bg_bit_off)
1447 {
1448 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1449 u32 cluster = 0;
1450
1451 BUG_ON(!ocfs2_is_cluster_bitmap(inode));
1452
1453 if (bg_blkno != osb->first_cluster_group_blkno)
1454 cluster = ocfs2_blocks_to_clusters(inode->i_sb, bg_blkno);
1455 cluster += (u32) bg_bit_off;
1456 return cluster;
1457 }
1458
1459 /* given a cluster offset, calculate which block group it belongs to
1460 * and return that block offset. */
1461 static inline u64 ocfs2_which_cluster_group(struct inode *inode,
1462 u32 cluster)
1463 {
1464 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1465 u32 group_no;
1466
1467 BUG_ON(!ocfs2_is_cluster_bitmap(inode));
1468
1469 group_no = cluster / osb->bitmap_cpg;
1470 if (!group_no)
1471 return osb->first_cluster_group_blkno;
1472 return ocfs2_clusters_to_blocks(inode->i_sb,
1473 group_no * osb->bitmap_cpg);
1474 }
1475
1476 /* given the block number of a cluster start, calculate which cluster
1477 * group and descriptor bitmap offset that corresponds to. */
1478 static inline void ocfs2_block_to_cluster_group(struct inode *inode,
1479 u64 data_blkno,
1480 u64 *bg_blkno,
1481 u16 *bg_bit_off)
1482 {
1483 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1484 u32 data_cluster = ocfs2_blocks_to_clusters(osb->sb, data_blkno);
1485
1486 BUG_ON(!ocfs2_is_cluster_bitmap(inode));
1487
1488 *bg_blkno = ocfs2_which_cluster_group(inode,
1489 data_cluster);
1490
1491 if (*bg_blkno == osb->first_cluster_group_blkno)
1492 *bg_bit_off = (u16) data_cluster;
1493 else
1494 *bg_bit_off = (u16) ocfs2_blocks_to_clusters(osb->sb,
1495 data_blkno - *bg_blkno);
1496 }
1497
1498 /*
1499 * min_bits - minimum contiguous chunk from this total allocation we
1500 * can handle. set to what we asked for originally for a full
1501 * contig. allocation, set to '1' to indicate we can deal with extents
1502 * of any size.
1503 */
1504 int ocfs2_claim_clusters(struct ocfs2_super *osb,
1505 handle_t *handle,
1506 struct ocfs2_alloc_context *ac,
1507 u32 min_clusters,
1508 u32 *cluster_start,
1509 u32 *num_clusters)
1510 {
1511 int status;
1512 unsigned int bits_wanted = ac->ac_bits_wanted - ac->ac_bits_given;
1513 u64 bg_blkno = 0;
1514 u16 bg_bit_off;
1515
1516 mlog_entry_void();
1517
1518 BUG_ON(!ac);
1519 BUG_ON(ac->ac_bits_given >= ac->ac_bits_wanted);
1520
1521 BUG_ON(ac->ac_which != OCFS2_AC_USE_LOCAL
1522 && ac->ac_which != OCFS2_AC_USE_MAIN);
1523
1524 if (ac->ac_which == OCFS2_AC_USE_LOCAL) {
1525 status = ocfs2_claim_local_alloc_bits(osb,
1526 handle,
1527 ac,
1528 bits_wanted,
1529 cluster_start,
1530 num_clusters);
1531 if (!status)
1532 atomic_inc(&osb->alloc_stats.local_data);
1533 } else {
1534 if (min_clusters > (osb->bitmap_cpg - 1)) {
1535 /* The only paths asking for contiguousness
1536 * should know about this already. */
1537 mlog(ML_ERROR, "minimum allocation requested exceeds "
1538 "group bitmap size!");
1539 status = -ENOSPC;
1540 goto bail;
1541 }
1542 /* clamp the current request down to a realistic size. */
1543 if (bits_wanted > (osb->bitmap_cpg - 1))
1544 bits_wanted = osb->bitmap_cpg - 1;
1545
1546 status = ocfs2_claim_suballoc_bits(osb,
1547 ac,
1548 handle,
1549 bits_wanted,
1550 min_clusters,
1551 &bg_bit_off,
1552 num_clusters,
1553 &bg_blkno);
1554 if (!status) {
1555 *cluster_start =
1556 ocfs2_desc_bitmap_to_cluster_off(ac->ac_inode,
1557 bg_blkno,
1558 bg_bit_off);
1559 atomic_inc(&osb->alloc_stats.bitmap_data);
1560 }
1561 }
1562 if (status < 0) {
1563 if (status != -ENOSPC)
1564 mlog_errno(status);
1565 goto bail;
1566 }
1567
1568 ac->ac_bits_given += *num_clusters;
1569
1570 bail:
1571 mlog_exit(status);
1572 return status;
1573 }
1574
1575 static inline int ocfs2_block_group_clear_bits(handle_t *handle,
1576 struct inode *alloc_inode,
1577 struct ocfs2_group_desc *bg,
1578 struct buffer_head *group_bh,
1579 unsigned int bit_off,
1580 unsigned int num_bits)
1581 {
1582 int status;
1583 unsigned int tmp;
1584 int journal_type = OCFS2_JOURNAL_ACCESS_WRITE;
1585 struct ocfs2_group_desc *undo_bg = NULL;
1586
1587 mlog_entry_void();
1588
1589 if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
1590 OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, bg);
1591 status = -EIO;
1592 goto bail;
1593 }
1594
1595 mlog(0, "off = %u, num = %u\n", bit_off, num_bits);
1596
1597 if (ocfs2_is_cluster_bitmap(alloc_inode))
1598 journal_type = OCFS2_JOURNAL_ACCESS_UNDO;
1599
1600 status = ocfs2_journal_access(handle, alloc_inode, group_bh,
1601 journal_type);
1602 if (status < 0) {
1603 mlog_errno(status);
1604 goto bail;
1605 }
1606
1607 if (ocfs2_is_cluster_bitmap(alloc_inode))
1608 undo_bg = (struct ocfs2_group_desc *) bh2jh(group_bh)->b_committed_data;
1609
1610 tmp = num_bits;
1611 while(tmp--) {
1612 ocfs2_clear_bit((bit_off + tmp),
1613 (unsigned long *) bg->bg_bitmap);
1614 if (ocfs2_is_cluster_bitmap(alloc_inode))
1615 ocfs2_set_bit(bit_off + tmp,
1616 (unsigned long *) undo_bg->bg_bitmap);
1617 }
1618 le16_add_cpu(&bg->bg_free_bits_count, num_bits);
1619
1620 status = ocfs2_journal_dirty(handle, group_bh);
1621 if (status < 0)
1622 mlog_errno(status);
1623 bail:
1624 return status;
1625 }
1626
1627 /*
1628 * expects the suballoc inode to already be locked.
1629 */
1630 static int ocfs2_free_suballoc_bits(handle_t *handle,
1631 struct inode *alloc_inode,
1632 struct buffer_head *alloc_bh,
1633 unsigned int start_bit,
1634 u64 bg_blkno,
1635 unsigned int count)
1636 {
1637 int status = 0;
1638 u32 tmp_used;
1639 struct ocfs2_super *osb = OCFS2_SB(alloc_inode->i_sb);
1640 struct ocfs2_dinode *fe = (struct ocfs2_dinode *) alloc_bh->b_data;
1641 struct ocfs2_chain_list *cl = &fe->id2.i_chain;
1642 struct buffer_head *group_bh = NULL;
1643 struct ocfs2_group_desc *group;
1644
1645 mlog_entry_void();
1646
1647 if (!OCFS2_IS_VALID_DINODE(fe)) {
1648 OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
1649 status = -EIO;
1650 goto bail;
1651 }
1652 BUG_ON((count + start_bit) > ocfs2_bits_per_group(cl));
1653
1654 mlog(0, "%llu: freeing %u bits from group %llu, starting at %u\n",
1655 (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno, count,
1656 (unsigned long long)bg_blkno, start_bit);
1657
1658 status = ocfs2_read_block(osb, bg_blkno, &group_bh, OCFS2_BH_CACHED,
1659 alloc_inode);
1660 if (status < 0) {
1661 mlog_errno(status);
1662 goto bail;
1663 }
1664
1665 group = (struct ocfs2_group_desc *) group_bh->b_data;
1666 status = ocfs2_check_group_descriptor(alloc_inode->i_sb, fe, group);
1667 if (status) {
1668 mlog_errno(status);
1669 goto bail;
1670 }
1671 BUG_ON((count + start_bit) > le16_to_cpu(group->bg_bits));
1672
1673 status = ocfs2_block_group_clear_bits(handle, alloc_inode,
1674 group, group_bh,
1675 start_bit, count);
1676 if (status < 0) {
1677 mlog_errno(status);
1678 goto bail;
1679 }
1680
1681 status = ocfs2_journal_access(handle, alloc_inode, alloc_bh,
1682 OCFS2_JOURNAL_ACCESS_WRITE);
1683 if (status < 0) {
1684 mlog_errno(status);
1685 goto bail;
1686 }
1687
1688 le32_add_cpu(&cl->cl_recs[le16_to_cpu(group->bg_chain)].c_free,
1689 count);
1690 tmp_used = le32_to_cpu(fe->id1.bitmap1.i_used);
1691 fe->id1.bitmap1.i_used = cpu_to_le32(tmp_used - count);
1692
1693 status = ocfs2_journal_dirty(handle, alloc_bh);
1694 if (status < 0) {
1695 mlog_errno(status);
1696 goto bail;
1697 }
1698
1699 bail:
1700 if (group_bh)
1701 brelse(group_bh);
1702
1703 mlog_exit(status);
1704 return status;
1705 }
1706
1707 static inline u64 ocfs2_which_suballoc_group(u64 block, unsigned int bit)
1708 {
1709 u64 group = block - (u64) bit;
1710
1711 return group;
1712 }
1713
1714 int ocfs2_free_dinode(handle_t *handle,
1715 struct inode *inode_alloc_inode,
1716 struct buffer_head *inode_alloc_bh,
1717 struct ocfs2_dinode *di)
1718 {
1719 u64 blk = le64_to_cpu(di->i_blkno);
1720 u16 bit = le16_to_cpu(di->i_suballoc_bit);
1721 u64 bg_blkno = ocfs2_which_suballoc_group(blk, bit);
1722
1723 return ocfs2_free_suballoc_bits(handle, inode_alloc_inode,
1724 inode_alloc_bh, bit, bg_blkno, 1);
1725 }
1726
1727 int ocfs2_free_extent_block(handle_t *handle,
1728 struct inode *eb_alloc_inode,
1729 struct buffer_head *eb_alloc_bh,
1730 struct ocfs2_extent_block *eb)
1731 {
1732 u64 blk = le64_to_cpu(eb->h_blkno);
1733 u16 bit = le16_to_cpu(eb->h_suballoc_bit);
1734 u64 bg_blkno = ocfs2_which_suballoc_group(blk, bit);
1735
1736 return ocfs2_free_suballoc_bits(handle, eb_alloc_inode, eb_alloc_bh,
1737 bit, bg_blkno, 1);
1738 }
1739
1740 int ocfs2_free_clusters(handle_t *handle,
1741 struct inode *bitmap_inode,
1742 struct buffer_head *bitmap_bh,
1743 u64 start_blk,
1744 unsigned int num_clusters)
1745 {
1746 int status;
1747 u16 bg_start_bit;
1748 u64 bg_blkno;
1749 struct ocfs2_dinode *fe;
1750
1751 /* You can't ever have a contiguous set of clusters
1752 * bigger than a block group bitmap so we never have to worry
1753 * about looping on them. */
1754
1755 mlog_entry_void();
1756
1757 /* This is expensive. We can safely remove once this stuff has
1758 * gotten tested really well. */
1759 BUG_ON(start_blk != ocfs2_clusters_to_blocks(bitmap_inode->i_sb, ocfs2_blocks_to_clusters(bitmap_inode->i_sb, start_blk)));
1760
1761 fe = (struct ocfs2_dinode *) bitmap_bh->b_data;
1762
1763 ocfs2_block_to_cluster_group(bitmap_inode, start_blk, &bg_blkno,
1764 &bg_start_bit);
1765
1766 mlog(0, "want to free %u clusters starting at block %llu\n",
1767 num_clusters, (unsigned long long)start_blk);
1768 mlog(0, "bg_blkno = %llu, bg_start_bit = %u\n",
1769 (unsigned long long)bg_blkno, bg_start_bit);
1770
1771 status = ocfs2_free_suballoc_bits(handle, bitmap_inode, bitmap_bh,
1772 bg_start_bit, bg_blkno,
1773 num_clusters);
1774 if (status < 0)
1775 mlog_errno(status);
1776
1777 mlog_exit(status);
1778 return status;
1779 }
1780
1781 static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg)
1782 {
1783 printk("Block Group:\n");
1784 printk("bg_signature: %s\n", bg->bg_signature);
1785 printk("bg_size: %u\n", bg->bg_size);
1786 printk("bg_bits: %u\n", bg->bg_bits);
1787 printk("bg_free_bits_count: %u\n", bg->bg_free_bits_count);
1788 printk("bg_chain: %u\n", bg->bg_chain);
1789 printk("bg_generation: %u\n", le32_to_cpu(bg->bg_generation));
1790 printk("bg_next_group: %llu\n",
1791 (unsigned long long)bg->bg_next_group);
1792 printk("bg_parent_dinode: %llu\n",
1793 (unsigned long long)bg->bg_parent_dinode);
1794 printk("bg_blkno: %llu\n",
1795 (unsigned long long)bg->bg_blkno);
1796 }
1797
1798 static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe)
1799 {
1800 int i;
1801
1802 printk("Suballoc Inode %llu:\n", (unsigned long long)fe->i_blkno);
1803 printk("i_signature: %s\n", fe->i_signature);
1804 printk("i_size: %llu\n",
1805 (unsigned long long)fe->i_size);
1806 printk("i_clusters: %u\n", fe->i_clusters);
1807 printk("i_generation: %u\n",
1808 le32_to_cpu(fe->i_generation));
1809 printk("id1.bitmap1.i_used: %u\n",
1810 le32_to_cpu(fe->id1.bitmap1.i_used));
1811 printk("id1.bitmap1.i_total: %u\n",
1812 le32_to_cpu(fe->id1.bitmap1.i_total));
1813 printk("id2.i_chain.cl_cpg: %u\n", fe->id2.i_chain.cl_cpg);
1814 printk("id2.i_chain.cl_bpc: %u\n", fe->id2.i_chain.cl_bpc);
1815 printk("id2.i_chain.cl_count: %u\n", fe->id2.i_chain.cl_count);
1816 printk("id2.i_chain.cl_next_free_rec: %u\n",
1817 fe->id2.i_chain.cl_next_free_rec);
1818 for(i = 0; i < fe->id2.i_chain.cl_next_free_rec; i++) {
1819 printk("fe->id2.i_chain.cl_recs[%d].c_free: %u\n", i,
1820 fe->id2.i_chain.cl_recs[i].c_free);
1821 printk("fe->id2.i_chain.cl_recs[%d].c_total: %u\n", i,
1822 fe->id2.i_chain.cl_recs[i].c_total);
1823 printk("fe->id2.i_chain.cl_recs[%d].c_blkno: %llu\n", i,
1824 (unsigned long long)fe->id2.i_chain.cl_recs[i].c_blkno);
1825 }
1826 }
Linux 2.6 ibm-acpi/thinkpad-acpi driver git tree