Blame view
fs/ext4/mballoc.c
127 KB
c9de560de ext4: Add multi b... |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
/* * Copyright (c) 2003-2006, Cluster File Systems, Inc, info@clusterfs.com * Written by Alex Tomas <alex@clusterfs.com> * * 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 */ |
8f6e39a7a ext4: Move mballo... |
23 |
#include "mballoc.h" |
6ba495e92 ext4: Add configu... |
24 |
#include <linux/debugfs.h> |
5a0e3ad6a include cleanup: ... |
25 |
#include <linux/slab.h> |
9bffad1ed ext4: convert ins... |
26 |
#include <trace/events/ext4.h> |
c9de560de ext4: Add multi b... |
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 |
/* * 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. * |
b713a5ec5 ext4: remove /pro... |
51 52 53 54 55 56 57 58 59 |
* During initialization phase of the allocator we decide to use the * group preallocation or inode preallocation depending on the size of * the 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 than sbi->s_mb_stream_request we * select to use the group preallocation. The default value of * s_mb_stream_request is 16 blocks. This can also be tuned via * /sys/fs/ext4/<partition>/mb_stream_req. The value is represented in * terms of number of blocks. |
c9de560de ext4: Add multi b... |
60 61 |
* * The main motivation for having small file use group preallocation is to |
b713a5ec5 ext4: remove /pro... |
62 |
* ensure that we have small files closer together on the disk. |
c9de560de ext4: Add multi b... |
63 |
* |
b713a5ec5 ext4: remove /pro... |
64 65 66 67 |
* First stage the allocator looks at the inode prealloc list, * ext4_inode_info->i_prealloc_list, which contains list of prealloc * spaces for this particular inode. The inode prealloc space is * represented as: |
c9de560de ext4: Add multi b... |
68 69 70 |
* * pa_lstart -> the logical start block for this prealloc space * pa_pstart -> the physical start block for this prealloc space |
1537a3638 tree-wide: fix 'l... |
71 |
* pa_len -> length for this prealloc space |
c9de560de ext4: Add multi b... |
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 |
* 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 } |
c3a326a65 ext4: cleanup mba... |
106 |
* [ group 0 bitmap][ group 0 buddy] [group 1][ group 1]... |
c9de560de ext4: Add multi b... |
107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 |
* * * 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 |
b713a5ec5 ext4: remove /pro... |
127 |
* sbi->s_mb_group_prealloc. Default value of s_mb_group_prealloc is |
c9de560de ext4: Add multi b... |
128 |
* 512 blocks. This can be tuned via |
b713a5ec5 ext4: remove /pro... |
129 |
* /sys/fs/ext4/<partition/mb_group_prealloc. The value is represented in |
c9de560de ext4: Add multi b... |
130 131 132 133 |
* 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) * |
b713a5ec5 ext4: remove /pro... |
134 |
* The regular allocator(using the buddy cache) supports few tunables. |
c9de560de ext4: Add multi b... |
135 |
* |
b713a5ec5 ext4: remove /pro... |
136 137 138 |
* /sys/fs/ext4/<partition>/mb_min_to_scan * /sys/fs/ext4/<partition>/mb_max_to_scan * /sys/fs/ext4/<partition>/mb_order2_req |
c9de560de ext4: Add multi b... |
139 |
* |
b713a5ec5 ext4: remove /pro... |
140 |
* The regular allocator uses buddy scan only if the request len is power of |
c9de560de ext4: Add multi b... |
141 142 |
* 2 blocks and the order of allocation is >= sbi->s_mb_order2_reqs. The * value of s_mb_order2_reqs can be tuned via |
b713a5ec5 ext4: remove /pro... |
143 |
* /sys/fs/ext4/<partition>/mb_order2_req. If the request len is equal to |
af901ca18 tree-wide: fix as... |
144 |
* stripe size (sbi->s_stripe), we try to search for contiguous block in |
b713a5ec5 ext4: remove /pro... |
145 146 147 |
* stripe size. This should result in better allocation on RAID setups. If * not, we search in the specific group using bitmap for best extents. The * tunable min_to_scan and max_to_scan control the behaviour here. |
c9de560de ext4: Add multi b... |
148 |
* min_to_scan indicate how long the mballoc __must__ look for a best |
b713a5ec5 ext4: remove /pro... |
149 |
* extent and max_to_scan indicates how long the mballoc __can__ look for a |
c9de560de ext4: Add multi b... |
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 |
* 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 * */ |
c3a326a65 ext4: cleanup mba... |
336 337 338 339 340 |
static struct kmem_cache *ext4_pspace_cachep; static struct kmem_cache *ext4_ac_cachep; static struct kmem_cache *ext4_free_ext_cachep; static void ext4_mb_generate_from_pa(struct super_block *sb, void *bitmap, ext4_group_t group); |
7a2fcbf7f ext4: don't use b... |
341 342 |
static void ext4_mb_generate_from_freelist(struct super_block *sb, void *bitmap, ext4_group_t group); |
c3a326a65 ext4: cleanup mba... |
343 |
static void release_blocks_on_commit(journal_t *journal, transaction_t *txn); |
ffad0a44b ext4: ext4_find_n... |
344 345 |
static inline void *mb_correct_addr_and_bit(int *bit, void *addr) { |
c9de560de ext4: Add multi b... |
346 |
#if BITS_PER_LONG == 64 |
ffad0a44b ext4: ext4_find_n... |
347 348 |
*bit += ((unsigned long) addr & 7UL) << 3; addr = (void *) ((unsigned long) addr & ~7UL); |
c9de560de ext4: Add multi b... |
349 |
#elif BITS_PER_LONG == 32 |
ffad0a44b ext4: ext4_find_n... |
350 351 |
*bit += ((unsigned long) addr & 3UL) << 3; addr = (void *) ((unsigned long) addr & ~3UL); |
c9de560de ext4: Add multi b... |
352 353 354 |
#else #error "how many bits you are?!" #endif |
ffad0a44b ext4: ext4_find_n... |
355 356 |
return addr; } |
c9de560de ext4: Add multi b... |
357 358 359 360 361 362 363 |
static inline int mb_test_bit(int bit, void *addr) { /* * ext4_test_bit on architecture like powerpc * needs unsigned long aligned address */ |
ffad0a44b ext4: ext4_find_n... |
364 |
addr = mb_correct_addr_and_bit(&bit, addr); |
c9de560de ext4: Add multi b... |
365 366 367 368 369 |
return ext4_test_bit(bit, addr); } static inline void mb_set_bit(int bit, void *addr) { |
ffad0a44b ext4: ext4_find_n... |
370 |
addr = mb_correct_addr_and_bit(&bit, addr); |
c9de560de ext4: Add multi b... |
371 372 |
ext4_set_bit(bit, addr); } |
c9de560de ext4: Add multi b... |
373 374 |
static inline void mb_clear_bit(int bit, void *addr) { |
ffad0a44b ext4: ext4_find_n... |
375 |
addr = mb_correct_addr_and_bit(&bit, addr); |
c9de560de ext4: Add multi b... |
376 377 |
ext4_clear_bit(bit, addr); } |
ffad0a44b ext4: ext4_find_n... |
378 379 |
static inline int mb_find_next_zero_bit(void *addr, int max, int start) { |
e7dfb2463 ext4: Fix mb_find... |
380 |
int fix = 0, ret, tmpmax; |
ffad0a44b ext4: ext4_find_n... |
381 |
addr = mb_correct_addr_and_bit(&fix, addr); |
e7dfb2463 ext4: Fix mb_find... |
382 |
tmpmax = max + fix; |
ffad0a44b ext4: ext4_find_n... |
383 |
start += fix; |
e7dfb2463 ext4: Fix mb_find... |
384 385 386 387 |
ret = ext4_find_next_zero_bit(addr, tmpmax, start) - fix; if (ret > max) return max; return ret; |
ffad0a44b ext4: ext4_find_n... |
388 389 390 391 |
} static inline int mb_find_next_bit(void *addr, int max, int start) { |
e7dfb2463 ext4: Fix mb_find... |
392 |
int fix = 0, ret, tmpmax; |
ffad0a44b ext4: ext4_find_n... |
393 |
addr = mb_correct_addr_and_bit(&fix, addr); |
e7dfb2463 ext4: Fix mb_find... |
394 |
tmpmax = max + fix; |
ffad0a44b ext4: ext4_find_n... |
395 |
start += fix; |
e7dfb2463 ext4: Fix mb_find... |
396 397 398 399 |
ret = ext4_find_next_bit(addr, tmpmax, start) - fix; if (ret > max) return max; return ret; |
ffad0a44b ext4: ext4_find_n... |
400 |
} |
c9de560de ext4: Add multi b... |
401 402 403 |
static void *mb_find_buddy(struct ext4_buddy *e4b, int order, int *max) { char *bb; |
c9de560de ext4: Add multi b... |
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 |
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; |
bc8e67409 ext4: Fix spinloc... |
432 |
assert_spin_locked(ext4_group_lock_ptr(sb, e4b->bd_group)); |
c9de560de ext4: Add multi b... |
433 434 435 |
for (i = 0; i < count; i++) { if (!mb_test_bit(first + i, e4b->bd_info->bb_bitmap)) { ext4_fsblk_t blocknr; |
5661bd686 ext4: cleanup to ... |
436 437 |
blocknr = ext4_group_first_block_no(sb, e4b->bd_group); |
c9de560de ext4: Add multi b... |
438 |
blocknr += first + i; |
5d1b1b3f4 ext4: fix BUG whe... |
439 |
ext4_grp_locked_error(sb, e4b->bd_group, |
e29136f80 ext4: Enhance ext... |
440 441 442 443 444 |
inode ? inode->i_ino : 0, blocknr, "freeing block already freed " "(bit %u)", first + i); |
c9de560de ext4: Add multi b... |
445 446 447 448 449 450 451 452 453 454 455 |
} 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; |
bc8e67409 ext4: Fix spinloc... |
456 |
assert_spin_locked(ext4_group_lock_ptr(e4b->bd_sb, e4b->bd_group)); |
c9de560de ext4: Add multi b... |
457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 |
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]) { |
a9df9a491 ext4: Make ext4_g... |
472 |
printk(KERN_ERR "corruption in group %u " |
4776004f5 ext4: Add printk ... |
473 474 475 476 |
"at byte %u(%u): %x in copy != %x " "on disk/prealloc ", e4b->bd_group, i, i * 8, b1[i], b2[i]); |
c9de560de ext4: Add multi b... |
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 |
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\" ", \ 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; |
c9de560de ext4: Add multi b... |
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 |
{ 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; |
60bd63d19 ext4: cleanup for... |
602 603 |
pa = list_entry(cur, struct ext4_prealloc_space, pa_group_list); ext4_get_group_no_and_offset(sb, pa->pa_pstart, &groupnr, &k); |
c9de560de ext4: Add multi b... |
604 |
MB_CHECK_ASSERT(groupnr == e4b->bd_group); |
60bd63d19 ext4: cleanup for... |
605 |
for (i = 0; i < pa->pa_len; i++) |
c9de560de ext4: Add multi b... |
606 607 608 609 610 611 |
MB_CHECK_ASSERT(mb_test_bit(k + i, buddy)); } return 0; } #undef MB_CHECK_ASSERT #define mb_check_buddy(e4b) __mb_check_buddy(e4b, \ |
46e665e9d ext4: replace rem... |
612 |
__FILE__, __func__, __LINE__) |
c9de560de ext4: Add multi b... |
613 614 615 616 617 618 |
#else #define mb_check_buddy(e4b) #endif /* FIXME!! need more doc */ static void ext4_mb_mark_free_simple(struct super_block *sb, |
a36b44988 ext4: use ext4_gr... |
619 |
void *buddy, ext4_grpblk_t first, ext4_grpblk_t len, |
c9de560de ext4: Add multi b... |
620 621 622 |
struct ext4_group_info *grp) { struct ext4_sb_info *sbi = EXT4_SB(sb); |
a36b44988 ext4: use ext4_gr... |
623 624 625 |
ext4_grpblk_t min; ext4_grpblk_t max; ext4_grpblk_t chunk; |
c9de560de ext4: Add multi b... |
626 |
unsigned short border; |
b73fce69e ext4: Fix kernel ... |
627 |
BUG_ON(len > EXT4_BLOCKS_PER_GROUP(sb)); |
c9de560de ext4: Add multi b... |
628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 |
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; } } |
8a57d9d61 ext4: check for a... |
652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 |
/* * Cache the order of the largest free extent we have available in this block * group. */ static void mb_set_largest_free_order(struct super_block *sb, struct ext4_group_info *grp) { int i; int bits; grp->bb_largest_free_order = -1; /* uninit */ bits = sb->s_blocksize_bits + 1; for (i = bits; i >= 0; i--) { if (grp->bb_counters[i] > 0) { grp->bb_largest_free_order = i; break; } } } |
089ceecc1 ext4: mark severa... |
672 673 |
static noinline_for_stack void ext4_mb_generate_buddy(struct super_block *sb, |
c9de560de ext4: Add multi b... |
674 675 676 |
void *buddy, void *bitmap, ext4_group_t group) { struct ext4_group_info *grp = ext4_get_group_info(sb, group); |
a36b44988 ext4: use ext4_gr... |
677 678 679 680 |
ext4_grpblk_t max = EXT4_BLOCKS_PER_GROUP(sb); ext4_grpblk_t i = 0; ext4_grpblk_t first; ext4_grpblk_t len; |
c9de560de ext4: Add multi b... |
681 682 683 684 685 686 |
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 */ |
ffad0a44b ext4: ext4_find_n... |
687 |
i = mb_find_next_zero_bit(bitmap, max, 0); |
c9de560de ext4: Add multi b... |
688 689 690 691 |
grp->bb_first_free = i; while (i < max) { fragments++; first = i; |
ffad0a44b ext4: ext4_find_n... |
692 |
i = mb_find_next_bit(bitmap, max, i); |
c9de560de ext4: Add multi b... |
693 694 695 696 697 698 699 |
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) |
ffad0a44b ext4: ext4_find_n... |
700 |
i = mb_find_next_zero_bit(bitmap, max, i); |
c9de560de ext4: Add multi b... |
701 702 703 704 |
} grp->bb_fragments = fragments; if (free != grp->bb_free) { |
e29136f80 ext4: Enhance ext... |
705 706 707 |
ext4_grp_locked_error(sb, group, 0, 0, "%u blocks in bitmap, %u in gd", free, grp->bb_free); |
e56eb6590 ext4: Don't claim... |
708 709 710 711 |
/* * If we intent to continue, we consider group descritor * corrupt and update bb_free using bitmap value */ |
c9de560de ext4: Add multi b... |
712 713 |
grp->bb_free = free; } |
8a57d9d61 ext4: check for a... |
714 |
mb_set_largest_free_order(sb, grp); |
c9de560de ext4: Add multi b... |
715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 |
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 } |
c3a326a65 ext4: cleanup mba... |
732 |
* [ group 0 bitmap][ group 0 buddy] [group 1][ group 1]... |
c9de560de ext4: Add multi b... |
733 734 735 736 737 738 739 |
* * * 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 |
8a57d9d61 ext4: check for a... |
740 741 742 |
* * Locking note: This routine takes the block group lock of all groups * for this page; do not hold this lock when calling this routine! |
c9de560de ext4: Add multi b... |
743 744 745 746 |
*/ static int ext4_mb_init_cache(struct page *page, char *incore) { |
8df9675f8 ext4: Avoid races... |
747 |
ext4_group_t ngroups; |
c9de560de ext4: Add multi b... |
748 749 750 751 752 753 754 755 756 757 758 759 760 |
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; |
6ba495e92 ext4: Add configu... |
761 762 |
mb_debug(1, "init page %lu ", page->index); |
c9de560de ext4: Add multi b... |
763 764 765 |
inode = page->mapping->host; sb = inode->i_sb; |
8df9675f8 ext4: Avoid races... |
766 |
ngroups = ext4_get_groups_count(sb); |
c9de560de ext4: Add multi b... |
767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 |
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; |
8df9675f8 ext4: Avoid races... |
789 |
if (first_group + i >= ngroups) |
c9de560de ext4: Add multi b... |
790 791 792 793 794 795 796 797 798 799 800 |
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; |
2ccb5fb9f ext4: Use new buf... |
801 |
if (bitmap_uptodate(bh[i])) |
c9de560de ext4: Add multi b... |
802 |
continue; |
c806e68f5 ext4: fix initial... |
803 |
lock_buffer(bh[i]); |
2ccb5fb9f ext4: Use new buf... |
804 805 806 807 |
if (bitmap_uptodate(bh[i])) { unlock_buffer(bh[i]); continue; } |
955ce5f5b ext4: Convert ext... |
808 |
ext4_lock_group(sb, first_group + i); |
c9de560de ext4: Add multi b... |
809 810 811 |
if (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) { ext4_init_block_bitmap(sb, bh[i], first_group + i, desc); |
2ccb5fb9f ext4: Use new buf... |
812 |
set_bitmap_uptodate(bh[i]); |
c9de560de ext4: Add multi b... |
813 |
set_buffer_uptodate(bh[i]); |
955ce5f5b ext4: Convert ext... |
814 |
ext4_unlock_group(sb, first_group + i); |
3300beda5 ext4: code cleanup |
815 |
unlock_buffer(bh[i]); |
c9de560de ext4: Add multi b... |
816 817 |
continue; } |
955ce5f5b ext4: Convert ext... |
818 |
ext4_unlock_group(sb, first_group + i); |
2ccb5fb9f ext4: Use new buf... |
819 820 821 822 823 824 825 826 827 |
if (buffer_uptodate(bh[i])) { /* * if not uninit if bh is uptodate, * bitmap is also uptodate */ set_bitmap_uptodate(bh[i]); unlock_buffer(bh[i]); continue; } |
c9de560de ext4: Add multi b... |
828 |
get_bh(bh[i]); |
2ccb5fb9f ext4: Use new buf... |
829 830 831 832 833 834 835 |
/* * submit the buffer_head for read. We can * safely mark the bitmap as uptodate now. * We do it here so the bitmap uptodate bit * get set with buffer lock held. */ set_bitmap_uptodate(bh[i]); |
c9de560de ext4: Add multi b... |
836 837 |
bh[i]->b_end_io = end_buffer_read_sync; submit_bh(READ, bh[i]); |
6ba495e92 ext4: Add configu... |
838 839 |
mb_debug(1, "read bitmap for group %u ", first_group + i); |
c9de560de ext4: Add multi b... |
840 841 842 843 844 845 846 847 848 849 |
} /* 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; |
31b481dc7 ext4: Fix ext4_mb... |
850 |
err = 0; |
c9de560de ext4: Add multi b... |
851 |
first_block = page->index * blocks_per_page; |
29eaf0249 ext4: Init the co... |
852 853 |
/* init the page */ memset(page_address(page), 0xff, PAGE_CACHE_SIZE); |
c9de560de ext4: Add multi b... |
854 855 856 857 858 |
for (i = 0; i < blocks_per_page; i++) { int group; struct ext4_group_info *grinfo; group = (first_block + i) >> 1; |
8df9675f8 ext4: Avoid races... |
859 |
if (group >= ngroups) |
c9de560de ext4: Add multi b... |
860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 |
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); |
6ba495e92 ext4: Add configu... |
878 879 |
mb_debug(1, "put buddy for group %u in page %lu/%x ", |
c9de560de ext4: Add multi b... |
880 |
group, page->index, i * blocksize); |
f307333e1 ext4: Add new tra... |
881 |
trace_ext4_mb_buddy_bitmap_load(sb, group); |
c9de560de ext4: Add multi b... |
882 883 884 |
grinfo = ext4_get_group_info(sb, group); grinfo->bb_fragments = 0; memset(grinfo->bb_counters, 0, |
1927805e6 ext4: use variabl... |
885 886 |
sizeof(*grinfo->bb_counters) * (sb->s_blocksize_bits+2)); |
c9de560de ext4: Add multi b... |
887 888 889 |
/* * incore got set to the group block bitmap below */ |
7a2fcbf7f ext4: don't use b... |
890 |
ext4_lock_group(sb, group); |
c9de560de ext4: Add multi b... |
891 |
ext4_mb_generate_buddy(sb, data, incore, group); |
7a2fcbf7f ext4: don't use b... |
892 |
ext4_unlock_group(sb, group); |
c9de560de ext4: Add multi b... |
893 894 895 896 |
incore = NULL; } else { /* this is block of bitmap */ BUG_ON(incore != NULL); |
6ba495e92 ext4: Add configu... |
897 898 |
mb_debug(1, "put bitmap for group %u in page %lu/%x ", |
c9de560de ext4: Add multi b... |
899 |
group, page->index, i * blocksize); |
f307333e1 ext4: Add new tra... |
900 |
trace_ext4_mb_bitmap_load(sb, group); |
c9de560de ext4: Add multi b... |
901 902 903 904 905 906 907 |
/* 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); |
7a2fcbf7f ext4: don't use b... |
908 |
ext4_mb_generate_from_freelist(sb, data, group); |
c9de560de ext4: Add multi b... |
909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 |
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; } |
8a57d9d61 ext4: check for a... |
928 929 930 931 932 |
/* * Locking note: This routine calls ext4_mb_init_cache(), which takes the * block group lock of all groups for this page; do not hold the BG lock when * calling this routine! */ |
b6a758ec3 ext4: move ext4_m... |
933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 |
static noinline_for_stack int ext4_mb_init_group(struct super_block *sb, ext4_group_t group) { int ret = 0; void *bitmap; int blocks_per_page; int block, pnum, poff; int num_grp_locked = 0; struct ext4_group_info *this_grp; struct ext4_sb_info *sbi = EXT4_SB(sb); struct inode *inode = sbi->s_buddy_cache; struct page *page = NULL, *bitmap_page = NULL; mb_debug(1, "init group %u ", group); blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize; this_grp = ext4_get_group_info(sb, group); /* |
08c3a8133 ext4: Clarify the... |
952 953 954 955 956 |
* This ensures that we don't reinit the buddy cache * page which map to the group from which we are already * allocating. If we are looking at the buddy cache we would * have taken a reference using ext4_mb_load_buddy and that * would have taken the alloc_sem lock. |
b6a758ec3 ext4: move ext4_m... |
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 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 |
*/ num_grp_locked = ext4_mb_get_buddy_cache_lock(sb, group); if (!EXT4_MB_GRP_NEED_INIT(this_grp)) { /* * somebody initialized the group * return without doing anything */ ret = 0; goto err; } /* * 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; page = find_or_create_page(inode->i_mapping, pnum, GFP_NOFS); if (page) { BUG_ON(page->mapping != inode->i_mapping); ret = ext4_mb_init_cache(page, NULL); if (ret) { unlock_page(page); goto err; } unlock_page(page); } if (page == NULL || !PageUptodate(page)) { ret = -EIO; goto err; } mark_page_accessed(page); bitmap_page = page; bitmap = page_address(page) + (poff * sb->s_blocksize); /* init buddy cache */ block++; pnum = block / blocks_per_page; poff = block % blocks_per_page; page = find_or_create_page(inode->i_mapping, pnum, GFP_NOFS); if (page == bitmap_page) { /* * If both the bitmap and buddy are in * the same page we don't need to force * init the buddy */ unlock_page(page); } else if (page) { BUG_ON(page->mapping != inode->i_mapping); ret = ext4_mb_init_cache(page, bitmap); if (ret) { unlock_page(page); goto err; } unlock_page(page); } if (page == NULL || !PageUptodate(page)) { ret = -EIO; goto err; } mark_page_accessed(page); err: ext4_mb_put_buddy_cache_lock(sb, group, num_grp_locked); if (bitmap_page) page_cache_release(bitmap_page); if (page) page_cache_release(page); return ret; } |
8a57d9d61 ext4: check for a... |
1027 1028 1029 1030 1031 |
/* * Locking note: This routine calls ext4_mb_init_cache(), which takes the * block group lock of all groups for this page; do not hold the BG lock when * calling this routine! */ |
4ddfef7b4 ext4: reduce mbal... |
1032 1033 1034 |
static noinline_for_stack int ext4_mb_load_buddy(struct super_block *sb, ext4_group_t group, struct ext4_buddy *e4b) |
c9de560de ext4: Add multi b... |
1035 |
{ |
c9de560de ext4: Add multi b... |
1036 1037 1038 1039 1040 |
int blocks_per_page; int block; int pnum; int poff; struct page *page; |
fdf6c7a76 ext4: add error p... |
1041 |
int ret; |
920313a72 ext4: Use EXT4_GR... |
1042 1043 1044 |
struct ext4_group_info *grp; struct ext4_sb_info *sbi = EXT4_SB(sb); struct inode *inode = sbi->s_buddy_cache; |
c9de560de ext4: Add multi b... |
1045 |
|
6ba495e92 ext4: Add configu... |
1046 1047 |
mb_debug(1, "load group %u ", group); |
c9de560de ext4: Add multi b... |
1048 1049 |
blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize; |
920313a72 ext4: Use EXT4_GR... |
1050 |
grp = ext4_get_group_info(sb, group); |
c9de560de ext4: Add multi b... |
1051 1052 1053 1054 1055 1056 1057 |
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; |
920313a72 ext4: Use EXT4_GR... |
1058 1059 1060 1061 1062 1063 1064 1065 |
e4b->alloc_semp = &grp->alloc_sem; /* Take the read lock on the group alloc * sem. This would make sure a parallel * ext4_mb_init_group happening on other * groups mapped by the page is blocked * till we are done with allocation */ |
f41c07505 ext4: check for n... |
1066 |
repeat_load_buddy: |
920313a72 ext4: Use EXT4_GR... |
1067 |
down_read(e4b->alloc_semp); |
c9de560de ext4: Add multi b... |
1068 |
|
f41c07505 ext4: check for n... |
1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 |
if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) { /* we need to check for group need init flag * with alloc_semp held so that we can be sure * that new blocks didn't get added to the group * when we are loading the buddy cache */ up_read(e4b->alloc_semp); /* * we need full data about the group * to make a good selection */ ret = ext4_mb_init_group(sb, group); if (ret) return ret; goto repeat_load_buddy; } |
c9de560de ext4: Add multi b... |
1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 |
/* * 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) |
920313a72 ext4: Use EXT4_GR... |
1099 1100 1101 1102 1103 1104 1105 1106 |
/* * drop the page reference and try * to get the page with lock. If we * are not uptodate that implies * somebody just created the page but * is yet to initialize the same. So * wait for it to initialize. */ |
c9de560de ext4: Add multi b... |
1107 1108 1109 1110 1111 |
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)) { |
fdf6c7a76 ext4: add error p... |
1112 1113 1114 1115 1116 |
ret = ext4_mb_init_cache(page, NULL); if (ret) { unlock_page(page); goto err; } |
c9de560de ext4: Add multi b... |
1117 1118 1119 1120 1121 1122 |
mb_cmp_bitmaps(e4b, page_address(page) + (poff * sb->s_blocksize)); } unlock_page(page); } } |
fdf6c7a76 ext4: add error p... |
1123 1124 |
if (page == NULL || !PageUptodate(page)) { ret = -EIO; |
c9de560de ext4: Add multi b... |
1125 |
goto err; |
fdf6c7a76 ext4: add error p... |
1126 |
} |
c9de560de ext4: Add multi b... |
1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 |
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); |
fdf6c7a76 ext4: add error p... |
1142 1143 1144 1145 1146 1147 1148 |
if (!PageUptodate(page)) { ret = ext4_mb_init_cache(page, e4b->bd_bitmap); if (ret) { unlock_page(page); goto err; } } |
c9de560de ext4: Add multi b... |
1149 1150 1151 |
unlock_page(page); } } |
fdf6c7a76 ext4: add error p... |
1152 1153 |
if (page == NULL || !PageUptodate(page)) { ret = -EIO; |
c9de560de ext4: Add multi b... |
1154 |
goto err; |
fdf6c7a76 ext4: add error p... |
1155 |
} |
c9de560de ext4: Add multi b... |
1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 |
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; |
920313a72 ext4: Use EXT4_GR... |
1172 1173 1174 |
/* Done with the buddy cache */ up_read(e4b->alloc_semp); |
fdf6c7a76 ext4: add error p... |
1175 |
return ret; |
c9de560de ext4: Add multi b... |
1176 |
} |
e39e07fdf ext4: rename ext4... |
1177 |
static void ext4_mb_unload_buddy(struct ext4_buddy *e4b) |
c9de560de ext4: Add multi b... |
1178 1179 1180 1181 1182 |
{ if (e4b->bd_bitmap_page) page_cache_release(e4b->bd_bitmap_page); if (e4b->bd_buddy_page) page_cache_release(e4b->bd_buddy_page); |
920313a72 ext4: Use EXT4_GR... |
1183 |
/* Done with the buddy cache */ |
8556e8f3b ext4: Don't allow... |
1184 1185 |
if (e4b->alloc_semp) up_read(e4b->alloc_semp); |
c9de560de ext4: Add multi b... |
1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 |
} static int mb_find_order_for_block(struct ext4_buddy *e4b, int block) { int order = 1; void *bb; BUG_ON(EXT4_MB_BITMAP(e4b) == EXT4_MB_BUDDY(e4b)); BUG_ON(block >= (1 << (e4b->bd_blkbits + 3))); bb = EXT4_MB_BUDDY(e4b); while (order <= e4b->bd_blkbits + 1) { block = block >> 1; if (!mb_test_bit(block, bb)) { /* this block is part of buddy of order 'order' */ return order; } bb += 1 << (e4b->bd_blkbits - order); order++; } return 0; } |
955ce5f5b ext4: Convert ext... |
1209 |
static void mb_clear_bits(void *bm, int cur, int len) |
c9de560de ext4: Add multi b... |
1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 |
{ __u32 *addr; len = cur + len; while (cur < len) { if ((cur & 31) == 0 && (len - cur) >= 32) { /* fast path: clear whole word at once */ addr = bm + (cur >> 3); *addr = 0; cur += 32; continue; } |
955ce5f5b ext4: Convert ext... |
1222 |
mb_clear_bit(cur, bm); |
c9de560de ext4: Add multi b... |
1223 1224 1225 |
cur++; } } |
955ce5f5b ext4: Convert ext... |
1226 |
static void mb_set_bits(void *bm, int cur, int len) |
c9de560de ext4: Add multi b... |
1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 |
{ __u32 *addr; len = cur + len; while (cur < len) { if ((cur & 31) == 0 && (len - cur) >= 32) { /* fast path: set whole word at once */ addr = bm + (cur >> 3); *addr = 0xffffffff; cur += 32; continue; } |
955ce5f5b ext4: Convert ext... |
1239 |
mb_set_bit(cur, bm); |
c9de560de ext4: Add multi b... |
1240 1241 1242 |
cur++; } } |
7e5a8cdd8 ext4: fix error p... |
1243 |
static void mb_free_blocks(struct inode *inode, struct ext4_buddy *e4b, |
c9de560de ext4: Add multi b... |
1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 |
int first, int count) { int block = 0; int max = 0; int order; void *buddy; void *buddy2; struct super_block *sb = e4b->bd_sb; BUG_ON(first + count > (sb->s_blocksize << 3)); |
bc8e67409 ext4: Fix spinloc... |
1254 |
assert_spin_locked(ext4_group_lock_ptr(sb, e4b->bd_group)); |
c9de560de ext4: Add multi b... |
1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 |
mb_check_buddy(e4b); mb_free_blocks_double(inode, e4b, first, count); e4b->bd_info->bb_free += count; if (first < e4b->bd_info->bb_first_free) e4b->bd_info->bb_first_free = first; /* let's maintain fragments counter */ if (first != 0) block = !mb_test_bit(first - 1, EXT4_MB_BITMAP(e4b)); if (first + count < EXT4_SB(sb)->s_mb_maxs[0]) max = !mb_test_bit(first + count, EXT4_MB_BITMAP(e4b)); if (block && max) e4b->bd_info->bb_fragments--; else if (!block && !max) e4b->bd_info->bb_fragments++; /* let's maintain buddy itself */ while (count-- > 0) { block = first++; order = 0; if (!mb_test_bit(block, EXT4_MB_BITMAP(e4b))) { ext4_fsblk_t blocknr; |
5661bd686 ext4: cleanup to ... |
1279 1280 |
blocknr = ext4_group_first_block_no(sb, e4b->bd_group); |
c9de560de ext4: Add multi b... |
1281 |
blocknr += block; |
5d1b1b3f4 ext4: fix BUG whe... |
1282 |
ext4_grp_locked_error(sb, e4b->bd_group, |
e29136f80 ext4: Enhance ext... |
1283 1284 1285 1286 |
inode ? inode->i_ino : 0, blocknr, "freeing already freed block " "(bit %u)", block); |
c9de560de ext4: Add multi b... |
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 |
} mb_clear_bit(block, EXT4_MB_BITMAP(e4b)); e4b->bd_info->bb_counters[order]++; /* start of the buddy */ buddy = mb_find_buddy(e4b, order, &max); do { block &= ~1UL; if (mb_test_bit(block, buddy) || mb_test_bit(block + 1, buddy)) break; /* both the buddies are free, try to coalesce them */ buddy2 = mb_find_buddy(e4b, order + 1, &max); if (!buddy2) break; if (order > 0) { /* for special purposes, we don't set * free bits in bitmap */ mb_set_bit(block, buddy); mb_set_bit(block + 1, buddy); } e4b->bd_info->bb_counters[order]--; e4b->bd_info->bb_counters[order]--; block = block >> 1; order++; e4b->bd_info->bb_counters[order]++; mb_clear_bit(block, buddy2); buddy = buddy2; } while (1); } |
8a57d9d61 ext4: check for a... |
1323 |
mb_set_largest_free_order(sb, e4b->bd_info); |
c9de560de ext4: Add multi b... |
1324 |
mb_check_buddy(e4b); |
c9de560de ext4: Add multi b... |
1325 1326 1327 1328 1329 1330 1331 1332 1333 |
} static int mb_find_extent(struct ext4_buddy *e4b, int order, int block, int needed, struct ext4_free_extent *ex) { int next = block; int max; int ord; void *buddy; |
bc8e67409 ext4: Fix spinloc... |
1334 |
assert_spin_locked(ext4_group_lock_ptr(e4b->bd_sb, e4b->bd_group)); |
c9de560de ext4: Add multi b... |
1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 |
BUG_ON(ex == NULL); buddy = mb_find_buddy(e4b, order, &max); BUG_ON(buddy == NULL); BUG_ON(block >= max); if (mb_test_bit(block, buddy)) { ex->fe_len = 0; ex->fe_start = 0; ex->fe_group = 0; return 0; } /* FIXME dorp order completely ? */ if (likely(order == 0)) { /* find actual order */ order = mb_find_order_for_block(e4b, block); block = block >> order; } ex->fe_len = 1 << order; ex->fe_start = block << order; ex->fe_group = e4b->bd_group; /* calc difference from given start */ next = next - ex->fe_start; ex->fe_len -= next; ex->fe_start += next; while (needed > ex->fe_len && (buddy = mb_find_buddy(e4b, order, &max))) { if (block + 1 >= max) break; next = (block + 1) * (1 << order); if (mb_test_bit(next, EXT4_MB_BITMAP(e4b))) break; ord = mb_find_order_for_block(e4b, next); order = ord; block = next >> order; ex->fe_len += 1 << order; } BUG_ON(ex->fe_start + ex->fe_len > (1 << (e4b->bd_blkbits + 3))); return ex->fe_len; } static int mb_mark_used(struct ext4_buddy *e4b, struct ext4_free_extent *ex) { int ord; int mlen = 0; int max = 0; int cur; int start = ex->fe_start; int len = ex->fe_len; unsigned ret = 0; int len0 = len; void *buddy; BUG_ON(start + len > (e4b->bd_sb->s_blocksize << 3)); BUG_ON(e4b->bd_group != ex->fe_group); |
bc8e67409 ext4: Fix spinloc... |
1398 |
assert_spin_locked(ext4_group_lock_ptr(e4b->bd_sb, e4b->bd_group)); |
c9de560de ext4: Add multi b... |
1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 |
mb_check_buddy(e4b); mb_mark_used_double(e4b, start, len); e4b->bd_info->bb_free -= len; if (e4b->bd_info->bb_first_free == start) e4b->bd_info->bb_first_free += len; /* let's maintain fragments counter */ if (start != 0) mlen = !mb_test_bit(start - 1, EXT4_MB_BITMAP(e4b)); if (start + len < EXT4_SB(e4b->bd_sb)->s_mb_maxs[0]) max = !mb_test_bit(start + len, EXT4_MB_BITMAP(e4b)); if (mlen && max) e4b->bd_info->bb_fragments++; else if (!mlen && !max) e4b->bd_info->bb_fragments--; /* let's maintain buddy itself */ while (len) { ord = mb_find_order_for_block(e4b, start); if (((start >> ord) << ord) == start && len >= (1 << ord)) { /* the whole chunk may be allocated at once! */ mlen = 1 << ord; buddy = mb_find_buddy(e4b, ord, &max); BUG_ON((start >> ord) >= max); mb_set_bit(start >> ord, buddy); e4b->bd_info->bb_counters[ord]--; start += mlen; len -= mlen; BUG_ON(len < 0); continue; } /* store for history */ if (ret == 0) ret = len | (ord << 16); /* we have to split large buddy */ BUG_ON(ord <= 0); buddy = mb_find_buddy(e4b, ord, &max); mb_set_bit(start >> ord, buddy); e4b->bd_info->bb_counters[ord]--; ord--; cur = (start >> ord) & ~1U; buddy = mb_find_buddy(e4b, ord, &max); mb_clear_bit(cur, buddy); mb_clear_bit(cur + 1, buddy); e4b->bd_info->bb_counters[ord]++; e4b->bd_info->bb_counters[ord]++; } |
8a57d9d61 ext4: check for a... |
1451 |
mb_set_largest_free_order(e4b->bd_sb, e4b->bd_info); |
c9de560de ext4: Add multi b... |
1452 |
|
955ce5f5b ext4: Convert ext... |
1453 |
mb_set_bits(EXT4_MB_BITMAP(e4b), ex->fe_start, len0); |
c9de560de ext4: Add multi b... |
1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 |
mb_check_buddy(e4b); return ret; } /* * Must be called under group lock! */ static void ext4_mb_use_best_found(struct ext4_allocation_context *ac, struct ext4_buddy *e4b) { struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); int ret; BUG_ON(ac->ac_b_ex.fe_group != e4b->bd_group); BUG_ON(ac->ac_status == AC_STATUS_FOUND); ac->ac_b_ex.fe_len = min(ac->ac_b_ex.fe_len, ac->ac_g_ex.fe_len); ac->ac_b_ex.fe_logical = ac->ac_g_ex.fe_logical; ret = mb_mark_used(e4b, &ac->ac_b_ex); /* preallocation can change ac_b_ex, thus we store actually * allocated blocks for history */ ac->ac_f_ex = ac->ac_b_ex; ac->ac_status = AC_STATUS_FOUND; ac->ac_tail = ret & 0xffff; ac->ac_buddy = ret >> 16; |
c3a326a65 ext4: cleanup mba... |
1482 1483 1484 1485 1486 1487 1488 |
/* * take the page reference. We want the page to be pinned * so that we don't get a ext4_mb_init_cache_call for this * group until we update the bitmap. That would mean we * double allocate blocks. The reference is dropped * in ext4_mb_release_context */ |
c9de560de ext4: Add multi b... |
1489 1490 1491 1492 |
ac->ac_bitmap_page = e4b->bd_bitmap_page; get_page(ac->ac_bitmap_page); ac->ac_buddy_page = e4b->bd_buddy_page; get_page(ac->ac_buddy_page); |
8556e8f3b ext4: Don't allow... |
1493 1494 1495 |
/* on allocation we use ac to track the held semaphore */ ac->alloc_semp = e4b->alloc_semp; e4b->alloc_semp = NULL; |
c9de560de ext4: Add multi b... |
1496 |
/* store last allocated for subsequent stream allocation */ |
4ba74d00a ext4: Fix bugs in... |
1497 |
if (ac->ac_flags & EXT4_MB_STREAM_ALLOC) { |
c9de560de ext4: Add multi b... |
1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 |
spin_lock(&sbi->s_md_lock); sbi->s_mb_last_group = ac->ac_f_ex.fe_group; sbi->s_mb_last_start = ac->ac_f_ex.fe_start; spin_unlock(&sbi->s_md_lock); } } /* * regular allocator, for general purposes allocation */ static void ext4_mb_check_limits(struct ext4_allocation_context *ac, struct ext4_buddy *e4b, int finish_group) { struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); struct ext4_free_extent *bex = &ac->ac_b_ex; struct ext4_free_extent *gex = &ac->ac_g_ex; struct ext4_free_extent ex; int max; |
032115fce ext4: Don't overw... |
1518 1519 |
if (ac->ac_status == AC_STATUS_FOUND) return; |
c9de560de ext4: Add multi b... |
1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 |
/* * We don't want to scan for a whole year */ if (ac->ac_found > sbi->s_mb_max_to_scan && !(ac->ac_flags & EXT4_MB_HINT_FIRST)) { ac->ac_status = AC_STATUS_BREAK; return; } /* * Haven't found good chunk so far, let's continue */ if (bex->fe_len < gex->fe_len) return; if ((finish_group || ac->ac_found > sbi->s_mb_min_to_scan) && bex->fe_group == e4b->bd_group) { /* recheck chunk's availability - we don't know * when it was found (within this lock-unlock * period or not) */ max = mb_find_extent(e4b, 0, bex->fe_start, gex->fe_len, &ex); if (max >= gex->fe_len) { ext4_mb_use_best_found(ac, e4b); return; } } } /* * The routine checks whether found extent is good enough. If it is, * then the extent gets marked used and flag is set to the context * to stop scanning. Otherwise, the extent is compared with the * previous found extent and if new one is better, then it's stored * in the context. Later, the best found extent will be used, if * mballoc can't find good enough extent. * * FIXME: real allocation policy is to be designed yet! */ static void ext4_mb_measure_extent(struct ext4_allocation_context *ac, struct ext4_free_extent *ex, struct ext4_buddy *e4b) { struct ext4_free_extent *bex = &ac->ac_b_ex; struct ext4_free_extent *gex = &ac->ac_g_ex; BUG_ON(ex->fe_len <= 0); |
8d03c7a0c ext4: fix bogus B... |
1566 |
BUG_ON(ex->fe_len > EXT4_BLOCKS_PER_GROUP(ac->ac_sb)); |
c9de560de ext4: Add multi b... |
1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 |
BUG_ON(ex->fe_start >= EXT4_BLOCKS_PER_GROUP(ac->ac_sb)); BUG_ON(ac->ac_status != AC_STATUS_CONTINUE); ac->ac_found++; /* * The special case - take what you catch first */ if (unlikely(ac->ac_flags & EXT4_MB_HINT_FIRST)) { *bex = *ex; ext4_mb_use_best_found(ac, e4b); return; } /* * Let's check whether the chuck is good enough */ if (ex->fe_len == gex->fe_len) { *bex = *ex; ext4_mb_use_best_found(ac, e4b); return; } /* * If this is first found extent, just store it in the context */ if (bex->fe_len == 0) { *bex = *ex; return; } /* * If new found extent is better, store it in the context */ if (bex->fe_len < gex->fe_len) { /* if the request isn't satisfied, any found extent * larger than previous best one is better */ if (ex->fe_len > bex->fe_len) *bex = *ex; } else if (ex->fe_len > gex->fe_len) { /* if the request is satisfied, then we try to find * an extent that still satisfy the request, but is * smaller than previous one */ if (ex->fe_len < bex->fe_len) *bex = *ex; } ext4_mb_check_limits(ac, e4b, 0); } |
089ceecc1 ext4: mark severa... |
1616 1617 |
static noinline_for_stack int ext4_mb_try_best_found(struct ext4_allocation_context *ac, |
c9de560de ext4: Add multi b... |
1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 |
struct ext4_buddy *e4b) { struct ext4_free_extent ex = ac->ac_b_ex; ext4_group_t group = ex.fe_group; int max; int err; BUG_ON(ex.fe_len <= 0); err = ext4_mb_load_buddy(ac->ac_sb, group, e4b); if (err) return err; ext4_lock_group(ac->ac_sb, group); max = mb_find_extent(e4b, 0, ex.fe_start, ex.fe_len, &ex); if (max > 0) { ac->ac_b_ex = ex; ext4_mb_use_best_found(ac, e4b); } ext4_unlock_group(ac->ac_sb, group); |
e39e07fdf ext4: rename ext4... |
1639 |
ext4_mb_unload_buddy(e4b); |
c9de560de ext4: Add multi b... |
1640 1641 1642 |
return 0; } |
089ceecc1 ext4: mark severa... |
1643 1644 |
static noinline_for_stack int ext4_mb_find_by_goal(struct ext4_allocation_context *ac, |
c9de560de ext4: Add multi b... |
1645 1646 1647 1648 1649 1650 |
struct ext4_buddy *e4b) { ext4_group_t group = ac->ac_g_ex.fe_group; int max; int err; struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); |
c9de560de ext4: Add multi b... |
1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 |
struct ext4_free_extent ex; if (!(ac->ac_flags & EXT4_MB_HINT_TRY_GOAL)) return 0; err = ext4_mb_load_buddy(ac->ac_sb, group, e4b); if (err) return err; ext4_lock_group(ac->ac_sb, group); max = mb_find_extent(e4b, 0, ac->ac_g_ex.fe_start, ac->ac_g_ex.fe_len, &ex); if (max >= ac->ac_g_ex.fe_len && ac->ac_g_ex.fe_len == sbi->s_stripe) { ext4_fsblk_t start; |
5661bd686 ext4: cleanup to ... |
1666 1667 |
start = ext4_group_first_block_no(ac->ac_sb, e4b->bd_group) + ex.fe_start; |
c9de560de ext4: Add multi b... |
1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 |
/* use do_div to get remainder (would be 64-bit modulo) */ if (do_div(start, sbi->s_stripe) == 0) { ac->ac_found++; ac->ac_b_ex = ex; ext4_mb_use_best_found(ac, e4b); } } else if (max >= ac->ac_g_ex.fe_len) { BUG_ON(ex.fe_len <= 0); BUG_ON(ex.fe_group != ac->ac_g_ex.fe_group); BUG_ON(ex.fe_start != ac->ac_g_ex.fe_start); ac->ac_found++; ac->ac_b_ex = ex; ext4_mb_use_best_found(ac, e4b); } else if (max > 0 && (ac->ac_flags & EXT4_MB_HINT_MERGE)) { /* Sometimes, caller may want to merge even small * number of blocks to an existing extent */ BUG_ON(ex.fe_len <= 0); BUG_ON(ex.fe_group != ac->ac_g_ex.fe_group); BUG_ON(ex.fe_start != ac->ac_g_ex.fe_start); ac->ac_found++; ac->ac_b_ex = ex; ext4_mb_use_best_found(ac, e4b); } ext4_unlock_group(ac->ac_sb, group); |
e39e07fdf ext4: rename ext4... |
1692 |
ext4_mb_unload_buddy(e4b); |
c9de560de ext4: Add multi b... |
1693 1694 1695 1696 1697 1698 1699 1700 |
return 0; } /* * The routine scans buddy structures (not bitmap!) from given order * to max order and tries to find big enough chunk to satisfy the req */ |
089ceecc1 ext4: mark severa... |
1701 1702 |
static noinline_for_stack void ext4_mb_simple_scan_group(struct ext4_allocation_context *ac, |
c9de560de ext4: Add multi b... |
1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 |
struct ext4_buddy *e4b) { struct super_block *sb = ac->ac_sb; struct ext4_group_info *grp = e4b->bd_info; void *buddy; int i; int k; int max; BUG_ON(ac->ac_2order <= 0); for (i = ac->ac_2order; i <= sb->s_blocksize_bits + 1; i++) { if (grp->bb_counters[i] == 0) continue; buddy = mb_find_buddy(e4b, i, &max); BUG_ON(buddy == NULL); |
ffad0a44b ext4: ext4_find_n... |
1719 |
k = mb_find_next_zero_bit(buddy, max, 0); |
c9de560de ext4: Add multi b... |
1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 |
BUG_ON(k >= max); ac->ac_found++; ac->ac_b_ex.fe_len = 1 << i; ac->ac_b_ex.fe_start = k << i; ac->ac_b_ex.fe_group = e4b->bd_group; ext4_mb_use_best_found(ac, e4b); BUG_ON(ac->ac_b_ex.fe_len != ac->ac_g_ex.fe_len); if (EXT4_SB(sb)->s_mb_stats) atomic_inc(&EXT4_SB(sb)->s_bal_2orders); break; } } /* * The routine scans the group and measures all found extents. * In order to optimize scanning, caller must pass number of * free blocks in the group, so the routine can know upper limit. */ |
089ceecc1 ext4: mark severa... |
1744 1745 |
static noinline_for_stack void ext4_mb_complex_scan_group(struct ext4_allocation_context *ac, |
c9de560de ext4: Add multi b... |
1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 |
struct ext4_buddy *e4b) { struct super_block *sb = ac->ac_sb; void *bitmap = EXT4_MB_BITMAP(e4b); struct ext4_free_extent ex; int i; int free; free = e4b->bd_info->bb_free; BUG_ON(free <= 0); i = e4b->bd_info->bb_first_free; while (free && ac->ac_status == AC_STATUS_CONTINUE) { |
ffad0a44b ext4: ext4_find_n... |
1760 |
i = mb_find_next_zero_bit(bitmap, |
c9de560de ext4: Add multi b... |
1761 1762 |
EXT4_BLOCKS_PER_GROUP(sb), i); if (i >= EXT4_BLOCKS_PER_GROUP(sb)) { |
26346ff68 ext4: Don't panic... |
1763 |
/* |
e56eb6590 ext4: Don't claim... |
1764 |
* IF we have corrupt bitmap, we won't find any |
26346ff68 ext4: Don't panic... |
1765 1766 1767 |
* free blocks even though group info says we * we have free blocks */ |
e29136f80 ext4: Enhance ext... |
1768 1769 |
ext4_grp_locked_error(sb, e4b->bd_group, 0, 0, "%d free blocks as per " |
fde4d95ad ext4: remove extr... |
1770 |
"group info. But bitmap says 0", |
26346ff68 ext4: Don't panic... |
1771 |
free); |
c9de560de ext4: Add multi b... |
1772 1773 1774 1775 1776 |
break; } mb_find_extent(e4b, 0, i, ac->ac_g_ex.fe_len, &ex); BUG_ON(ex.fe_len <= 0); |
26346ff68 ext4: Don't panic... |
1777 |
if (free < ex.fe_len) { |
e29136f80 ext4: Enhance ext... |
1778 1779 |
ext4_grp_locked_error(sb, e4b->bd_group, 0, 0, "%d free blocks as per " |
fde4d95ad ext4: remove extr... |
1780 |
"group info. But got %d blocks", |
26346ff68 ext4: Don't panic... |
1781 |
free, ex.fe_len); |
e56eb6590 ext4: Don't claim... |
1782 1783 1784 1785 1786 1787 |
/* * The number of free blocks differs. This mostly * indicate that the bitmap is corrupt. So exit * without claiming the space. */ break; |
26346ff68 ext4: Don't panic... |
1788 |
} |
c9de560de ext4: Add multi b... |
1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 |
ext4_mb_measure_extent(ac, &ex, e4b); i += ex.fe_len; free -= ex.fe_len; } ext4_mb_check_limits(ac, e4b, 1); } /* * This is a special case for storages like raid5 |
506bf2d82 ext4: allocate st... |
1801 |
* we try to find stripe-aligned chunks for stripe-size-multiple requests |
c9de560de ext4: Add multi b... |
1802 |
*/ |
089ceecc1 ext4: mark severa... |
1803 1804 |
static noinline_for_stack void ext4_mb_scan_aligned(struct ext4_allocation_context *ac, |
c9de560de ext4: Add multi b... |
1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 |
struct ext4_buddy *e4b) { struct super_block *sb = ac->ac_sb; struct ext4_sb_info *sbi = EXT4_SB(sb); void *bitmap = EXT4_MB_BITMAP(e4b); struct ext4_free_extent ex; ext4_fsblk_t first_group_block; ext4_fsblk_t a; ext4_grpblk_t i; int max; BUG_ON(sbi->s_stripe == 0); /* find first stripe-aligned block in group */ |
5661bd686 ext4: cleanup to ... |
1819 |
first_group_block = ext4_group_first_block_no(sb, e4b->bd_group); |
c9de560de ext4: Add multi b... |
1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 |
a = first_group_block + sbi->s_stripe - 1; do_div(a, sbi->s_stripe); i = (a * sbi->s_stripe) - first_group_block; while (i < EXT4_BLOCKS_PER_GROUP(sb)) { if (!mb_test_bit(i, bitmap)) { max = mb_find_extent(e4b, 0, i, sbi->s_stripe, &ex); if (max >= sbi->s_stripe) { ac->ac_found++; ac->ac_b_ex = ex; ext4_mb_use_best_found(ac, e4b); break; } } i += sbi->s_stripe; } } |
8a57d9d61 ext4: check for a... |
1837 |
/* This is now called BEFORE we load the buddy bitmap. */ |
c9de560de ext4: Add multi b... |
1838 1839 1840 1841 |
static int ext4_mb_good_group(struct ext4_allocation_context *ac, ext4_group_t group, int cr) { unsigned free, fragments; |
a4912123b ext4: New inode/b... |
1842 |
int flex_size = ext4_flex_bg_size(EXT4_SB(ac->ac_sb)); |
c9de560de ext4: Add multi b... |
1843 1844 1845 |
struct ext4_group_info *grp = ext4_get_group_info(ac->ac_sb, group); BUG_ON(cr < 0 || cr >= 4); |
8a57d9d61 ext4: check for a... |
1846 1847 1848 1849 1850 1851 1852 |
/* We only do this if the grp has never been initialized */ if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) { int ret = ext4_mb_init_group(ac->ac_sb, group); if (ret) return 0; } |
c9de560de ext4: Add multi b... |
1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 |
free = grp->bb_free; fragments = grp->bb_fragments; if (free == 0) return 0; if (fragments == 0) return 0; switch (cr) { case 0: BUG_ON(ac->ac_2order == 0); |
c9de560de ext4: Add multi b... |
1864 |
|
8a57d9d61 ext4: check for a... |
1865 1866 |
if (grp->bb_largest_free_order < ac->ac_2order) return 0; |
a4912123b ext4: New inode/b... |
1867 1868 1869 1870 1871 |
/* Avoid using the first bg of a flexgroup for data files */ if ((ac->ac_flags & EXT4_MB_HINT_DATA) && (flex_size >= EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME) && ((group % flex_size) == 0)) return 0; |
8a57d9d61 ext4: check for a... |
1872 |
return 1; |
c9de560de ext4: Add multi b... |
1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 |
case 1: if ((free / fragments) >= ac->ac_g_ex.fe_len) return 1; break; case 2: if (free >= ac->ac_g_ex.fe_len) return 1; break; case 3: return 1; default: BUG(); } return 0; } |
920313a72 ext4: Use EXT4_GR... |
1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 |
/* * lock the group_info alloc_sem of all the groups * belonging to the same buddy cache page. This * make sure other parallel operation on the buddy * cache doesn't happen whild holding the buddy cache * lock */ int ext4_mb_get_buddy_cache_lock(struct super_block *sb, ext4_group_t group) { int i; int block, pnum; int blocks_per_page; int groups_per_page; |
8df9675f8 ext4: Avoid races... |
1902 |
ext4_group_t ngroups = ext4_get_groups_count(sb); |
920313a72 ext4: Use EXT4_GR... |
1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 |
ext4_group_t first_group; struct ext4_group_info *grp; blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize; /* * 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; first_group = pnum * blocks_per_page / 2; groups_per_page = blocks_per_page >> 1; if (groups_per_page == 0) groups_per_page = 1; /* read all groups the page covers into the cache */ for (i = 0; i < groups_per_page; i++) { |
8df9675f8 ext4: Avoid races... |
1921 |
if ((first_group + i) >= ngroups) |
920313a72 ext4: Use EXT4_GR... |
1922 1923 1924 1925 1926 1927 1928 |
break; grp = ext4_get_group_info(sb, first_group + i); /* take all groups write allocation * semaphore. This make sure there is * no block allocation going on in any * of that groups */ |
b7be019e8 ext4: Fix lockdep... |
1929 |
down_write_nested(&grp->alloc_sem, i); |
920313a72 ext4: Use EXT4_GR... |
1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 |
} return i; } void ext4_mb_put_buddy_cache_lock(struct super_block *sb, ext4_group_t group, int locked_group) { int i; int block, pnum; int blocks_per_page; ext4_group_t first_group; struct ext4_group_info *grp; blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize; /* * 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; first_group = pnum * blocks_per_page / 2; /* release locks on all the groups */ for (i = 0; i < locked_group; i++) { grp = ext4_get_group_info(sb, first_group + i); /* take all groups write allocation * semaphore. This make sure there is * no block allocation going on in any * of that groups */ up_write(&grp->alloc_sem); } } |
4ddfef7b4 ext4: reduce mbal... |
1965 1966 |
static noinline_for_stack int ext4_mb_regular_allocator(struct ext4_allocation_context *ac) |
c9de560de ext4: Add multi b... |
1967 |
{ |
8df9675f8 ext4: Avoid races... |
1968 |
ext4_group_t ngroups, group, i; |
c9de560de ext4: Add multi b... |
1969 1970 |
int cr; int err = 0; |
c9de560de ext4: Add multi b... |
1971 1972 1973 |
struct ext4_sb_info *sbi; struct super_block *sb; struct ext4_buddy e4b; |
c9de560de ext4: Add multi b... |
1974 1975 1976 |
sb = ac->ac_sb; sbi = EXT4_SB(sb); |
8df9675f8 ext4: Avoid races... |
1977 |
ngroups = ext4_get_groups_count(sb); |
fb0a387dc ext4: limit block... |
1978 |
/* non-extent files are limited to low blocks/groups */ |
12e9b8920 ext4: Use bitops ... |
1979 |
if (!(ext4_test_inode_flag(ac->ac_inode, EXT4_INODE_EXTENTS))) |
fb0a387dc ext4: limit block... |
1980 |
ngroups = sbi->s_blockfile_groups; |
c9de560de ext4: Add multi b... |
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 |
BUG_ON(ac->ac_status == AC_STATUS_FOUND); /* first, try the goal */ err = ext4_mb_find_by_goal(ac, &e4b); if (err || ac->ac_status == AC_STATUS_FOUND) goto out; if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY)) goto out; /* * ac->ac2_order is set only if the fe_len is a power of 2 * if ac2_order is set we also set criteria to 0 so that we * try exact allocation using buddy. */ i = fls(ac->ac_g_ex.fe_len); ac->ac_2order = 0; /* * We search using buddy data only if the order of the request * is greater than equal to the sbi_s_mb_order2_reqs |
b713a5ec5 ext4: remove /pro... |
2001 |
* You can tune it via /sys/fs/ext4/<partition>/mb_order2_req |
c9de560de ext4: Add multi b... |
2002 2003 2004 2005 2006 2007 2008 2009 |
*/ if (i >= sbi->s_mb_order2_reqs) { /* * This should tell if fe_len is exactly power of 2 */ if ((ac->ac_g_ex.fe_len & (~(1 << (i - 1)))) == 0) ac->ac_2order = i - 1; } |
4ba74d00a ext4: Fix bugs in... |
2010 2011 |
/* if stream allocation is enabled, use global goal */ if (ac->ac_flags & EXT4_MB_STREAM_ALLOC) { |
c9de560de ext4: Add multi b... |
2012 2013 2014 2015 2016 2017 |
/* TBD: may be hot point */ spin_lock(&sbi->s_md_lock); ac->ac_g_ex.fe_group = sbi->s_mb_last_group; ac->ac_g_ex.fe_start = sbi->s_mb_last_start; spin_unlock(&sbi->s_md_lock); } |
4ba74d00a ext4: Fix bugs in... |
2018 |
|
c9de560de ext4: Add multi b... |
2019 2020 2021 2022 2023 2024 2025 2026 2027 |
/* Let's just scan groups to find more-less suitable blocks */ cr = ac->ac_2order ? 0 : 1; /* * cr == 0 try to get exact allocation, * cr == 3 try to get anything */ repeat: for (; cr < 4 && ac->ac_status == AC_STATUS_CONTINUE; cr++) { ac->ac_criteria = cr; |
ed8f9c751 ext4: start searc... |
2028 2029 2030 2031 2032 |
/* * searching for the right group start * from the goal value specified */ group = ac->ac_g_ex.fe_group; |
8df9675f8 ext4: Avoid races... |
2033 |
for (i = 0; i < ngroups; group++, i++) { |
8df9675f8 ext4: Avoid races... |
2034 |
if (group == ngroups) |
c9de560de ext4: Add multi b... |
2035 |
group = 0; |
8a57d9d61 ext4: check for a... |
2036 2037 |
/* This now checks without needing the buddy page */ if (!ext4_mb_good_group(ac, group, cr)) |
c9de560de ext4: Add multi b... |
2038 |
continue; |
c9de560de ext4: Add multi b... |
2039 2040 2041 2042 2043 |
err = ext4_mb_load_buddy(sb, group, &e4b); if (err) goto out; ext4_lock_group(sb, group); |
8a57d9d61 ext4: check for a... |
2044 2045 2046 2047 2048 |
/* * We need to check again after locking the * block group */ |
c9de560de ext4: Add multi b... |
2049 |
if (!ext4_mb_good_group(ac, group, cr)) { |
c9de560de ext4: Add multi b... |
2050 |
ext4_unlock_group(sb, group); |
e39e07fdf ext4: rename ext4... |
2051 |
ext4_mb_unload_buddy(&e4b); |
c9de560de ext4: Add multi b... |
2052 2053 2054 2055 |
continue; } ac->ac_groups_scanned++; |
75507efb1 ext4: Don't avoid... |
2056 |
if (cr == 0) |
c9de560de ext4: Add multi b... |
2057 |
ext4_mb_simple_scan_group(ac, &e4b); |
506bf2d82 ext4: allocate st... |
2058 2059 |
else if (cr == 1 && sbi->s_stripe && !(ac->ac_g_ex.fe_len % sbi->s_stripe)) |
c9de560de ext4: Add multi b... |
2060 2061 2062 2063 2064 |
ext4_mb_scan_aligned(ac, &e4b); else ext4_mb_complex_scan_group(ac, &e4b); ext4_unlock_group(sb, group); |
e39e07fdf ext4: rename ext4... |
2065 |
ext4_mb_unload_buddy(&e4b); |
c9de560de ext4: Add multi b... |
2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 |
if (ac->ac_status != AC_STATUS_CONTINUE) break; } } if (ac->ac_b_ex.fe_len > 0 && ac->ac_status != AC_STATUS_FOUND && !(ac->ac_flags & EXT4_MB_HINT_FIRST)) { /* * We've been searching too long. Let's try to allocate * the best chunk we've found so far */ ext4_mb_try_best_found(ac, &e4b); if (ac->ac_status != AC_STATUS_FOUND) { /* * Someone more lucky has already allocated it. * The only thing we can do is just take first * found block(s) printk(KERN_DEBUG "EXT4-fs: someone won our chunk "); */ ac->ac_b_ex.fe_group = 0; ac->ac_b_ex.fe_start = 0; ac->ac_b_ex.fe_len = 0; ac->ac_status = AC_STATUS_CONTINUE; ac->ac_flags |= EXT4_MB_HINT_FIRST; cr = 3; atomic_inc(&sbi->s_mb_lost_chunks); goto repeat; } } out: return err; } |
c9de560de ext4: Add multi b... |
2101 2102 2103 |
static void *ext4_mb_seq_groups_start(struct seq_file *seq, loff_t *pos) { struct super_block *sb = seq->private; |
c9de560de ext4: Add multi b... |
2104 |
ext4_group_t group; |
8df9675f8 ext4: Avoid races... |
2105 |
if (*pos < 0 || *pos >= ext4_get_groups_count(sb)) |
c9de560de ext4: Add multi b... |
2106 |
return NULL; |
c9de560de ext4: Add multi b... |
2107 |
group = *pos + 1; |
a9df9a491 ext4: Make ext4_g... |
2108 |
return (void *) ((unsigned long) group); |
c9de560de ext4: Add multi b... |
2109 2110 2111 2112 2113 |
} static void *ext4_mb_seq_groups_next(struct seq_file *seq, void *v, loff_t *pos) { struct super_block *sb = seq->private; |
c9de560de ext4: Add multi b... |
2114 2115 2116 |
ext4_group_t group; ++*pos; |
8df9675f8 ext4: Avoid races... |
2117 |
if (*pos < 0 || *pos >= ext4_get_groups_count(sb)) |
c9de560de ext4: Add multi b... |
2118 2119 |
return NULL; group = *pos + 1; |
a9df9a491 ext4: Make ext4_g... |
2120 |
return (void *) ((unsigned long) group); |
c9de560de ext4: Add multi b... |
2121 2122 2123 2124 2125 |
} static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v) { struct super_block *sb = seq->private; |
a9df9a491 ext4: Make ext4_g... |
2126 |
ext4_group_t group = (ext4_group_t) ((unsigned long) v); |
c9de560de ext4: Add multi b... |
2127 2128 2129 2130 2131 |
int i; int err; struct ext4_buddy e4b; struct sg { struct ext4_group_info info; |
a36b44988 ext4: use ext4_gr... |
2132 |
ext4_grpblk_t counters[16]; |
c9de560de ext4: Add multi b... |
2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 |
} sg; group--; if (group == 0) seq_printf(seq, "#%-5s: %-5s %-5s %-5s " "[ %-5s %-5s %-5s %-5s %-5s %-5s %-5s " "%-5s %-5s %-5s %-5s %-5s %-5s %-5s ] ", "group", "free", "frags", "first", "2^0", "2^1", "2^2", "2^3", "2^4", "2^5", "2^6", "2^7", "2^8", "2^9", "2^10", "2^11", "2^12", "2^13"); i = (sb->s_blocksize_bits + 2) * sizeof(sg.info.bb_counters[0]) + sizeof(struct ext4_group_info); err = ext4_mb_load_buddy(sb, group, &e4b); if (err) { |
a9df9a491 ext4: Make ext4_g... |
2149 2150 |
seq_printf(seq, "#%-5u: I/O error ", group); |
c9de560de ext4: Add multi b... |
2151 2152 2153 2154 2155 |
return 0; } ext4_lock_group(sb, group); memcpy(&sg, ext4_get_group_info(sb, group), i); ext4_unlock_group(sb, group); |
e39e07fdf ext4: rename ext4... |
2156 |
ext4_mb_unload_buddy(&e4b); |
c9de560de ext4: Add multi b... |
2157 |
|
a9df9a491 ext4: Make ext4_g... |
2158 |
seq_printf(seq, "#%-5u: %-5u %-5u %-5u [", group, sg.info.bb_free, |
c9de560de ext4: Add multi b... |
2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 |
sg.info.bb_fragments, sg.info.bb_first_free); for (i = 0; i <= 13; i++) seq_printf(seq, " %-5u", i <= sb->s_blocksize_bits + 1 ? sg.info.bb_counters[i] : 0); seq_printf(seq, " ] "); return 0; } static void ext4_mb_seq_groups_stop(struct seq_file *seq, void *v) { } |
7f1346a9d ext4: Declare seq... |
2172 |
static const struct seq_operations ext4_mb_seq_groups_ops = { |
c9de560de ext4: Add multi b... |
2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 |
.start = ext4_mb_seq_groups_start, .next = ext4_mb_seq_groups_next, .stop = ext4_mb_seq_groups_stop, .show = ext4_mb_seq_groups_show, }; static int ext4_mb_seq_groups_open(struct inode *inode, struct file *file) { struct super_block *sb = PDE(inode)->data; int rc; rc = seq_open(file, &ext4_mb_seq_groups_ops); if (rc == 0) { |
a271fe852 ext4: Remove unne... |
2186 |
struct seq_file *m = file->private_data; |
c9de560de ext4: Add multi b... |
2187 2188 2189 2190 2191 |
m->private = sb; } return rc; } |
7f1346a9d ext4: Declare seq... |
2192 |
static const struct file_operations ext4_mb_seq_groups_fops = { |
c9de560de ext4: Add multi b... |
2193 2194 2195 2196 2197 2198 |
.owner = THIS_MODULE, .open = ext4_mb_seq_groups_open, .read = seq_read, .llseek = seq_lseek, .release = seq_release, }; |
5f21b0e64 ext4: fix online ... |
2199 2200 |
/* Create and initialize ext4_group_info data for the given group. */ |
920313a72 ext4: Use EXT4_GR... |
2201 |
int ext4_mb_add_groupinfo(struct super_block *sb, ext4_group_t group, |
5f21b0e64 ext4: fix online ... |
2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 |
struct ext4_group_desc *desc) { int i, len; int metalen = 0; struct ext4_sb_info *sbi = EXT4_SB(sb); struct ext4_group_info **meta_group_info; /* * First check if this group is the first of a reserved block. * If it's true, we have to allocate a new table of pointers * to ext4_group_info structures */ if (group % EXT4_DESC_PER_BLOCK(sb) == 0) { metalen = sizeof(*meta_group_info) << EXT4_DESC_PER_BLOCK_BITS(sb); meta_group_info = kmalloc(metalen, GFP_KERNEL); if (meta_group_info == NULL) { printk(KERN_ERR "EXT4-fs: can't allocate mem for a " "buddy group "); goto exit_meta_group_info; } sbi->s_group_info[group >> EXT4_DESC_PER_BLOCK_BITS(sb)] = meta_group_info; } /* * calculate needed size. if change bb_counters size, * don't forget about ext4_mb_generate_buddy() */ len = offsetof(typeof(**meta_group_info), bb_counters[sb->s_blocksize_bits + 2]); meta_group_info = sbi->s_group_info[group >> EXT4_DESC_PER_BLOCK_BITS(sb)]; i = group & (EXT4_DESC_PER_BLOCK(sb) - 1); meta_group_info[i] = kzalloc(len, GFP_KERNEL); if (meta_group_info[i] == NULL) { printk(KERN_ERR "EXT4-fs: can't allocate buddy mem "); goto exit_group_info; } set_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &(meta_group_info[i]->bb_state)); /* * initialize bb_free to be able to skip * empty groups without initialization */ if (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) { meta_group_info[i]->bb_free = ext4_free_blocks_after_init(sb, group, desc); } else { meta_group_info[i]->bb_free = |
560671a0d ext4: Use high 16... |
2257 |
ext4_free_blks_count(sb, desc); |
5f21b0e64 ext4: fix online ... |
2258 2259 2260 |
} INIT_LIST_HEAD(&meta_group_info[i]->bb_prealloc_list); |
920313a72 ext4: Use EXT4_GR... |
2261 |
init_rwsem(&meta_group_info[i]->alloc_sem); |
64e290ec6 ext4: fix up rb_r... |
2262 |
meta_group_info[i]->bb_free_root = RB_ROOT; |
8a57d9d61 ext4: check for a... |
2263 |
meta_group_info[i]->bb_largest_free_order = -1; /* uninit */ |
5f21b0e64 ext4: fix online ... |
2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 |
#ifdef DOUBLE_CHECK { struct buffer_head *bh; meta_group_info[i]->bb_bitmap = kmalloc(sb->s_blocksize, GFP_KERNEL); BUG_ON(meta_group_info[i]->bb_bitmap == NULL); bh = ext4_read_block_bitmap(sb, group); BUG_ON(bh == NULL); memcpy(meta_group_info[i]->bb_bitmap, bh->b_data, sb->s_blocksize); put_bh(bh); } #endif return 0; exit_group_info: /* If a meta_group_info table has been allocated, release it now */ if (group % EXT4_DESC_PER_BLOCK(sb) == 0) kfree(sbi->s_group_info[group >> EXT4_DESC_PER_BLOCK_BITS(sb)]); exit_meta_group_info: return -ENOMEM; } /* ext4_mb_add_groupinfo */ |
c9de560de ext4: Add multi b... |
2288 2289 |
static int ext4_mb_init_backend(struct super_block *sb) { |
8df9675f8 ext4: Avoid races... |
2290 |
ext4_group_t ngroups = ext4_get_groups_count(sb); |
c9de560de ext4: Add multi b... |
2291 |
ext4_group_t i; |
c9de560de ext4: Add multi b... |
2292 |
struct ext4_sb_info *sbi = EXT4_SB(sb); |
5f21b0e64 ext4: fix online ... |
2293 2294 2295 2296 |
struct ext4_super_block *es = sbi->s_es; int num_meta_group_infos; int num_meta_group_infos_max; int array_size; |
5f21b0e64 ext4: fix online ... |
2297 2298 2299 |
struct ext4_group_desc *desc; /* This is the number of blocks used by GDT */ |
8df9675f8 ext4: Avoid races... |
2300 |
num_meta_group_infos = (ngroups + EXT4_DESC_PER_BLOCK(sb) - |
5f21b0e64 ext4: fix online ... |
2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 |
1) >> EXT4_DESC_PER_BLOCK_BITS(sb); /* * This is the total number of blocks used by GDT including * the number of reserved blocks for GDT. * The s_group_info array is allocated with this value * to allow a clean online resize without a complex * manipulation of pointer. * The drawback is the unused memory when no resize * occurs but it's very low in terms of pages * (see comments below) * Need to handle this properly when META_BG resizing is allowed */ num_meta_group_infos_max = num_meta_group_infos + le16_to_cpu(es->s_reserved_gdt_blocks); |
c9de560de ext4: Add multi b... |
2316 |
|
5f21b0e64 ext4: fix online ... |
2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 |
/* * array_size is the size of s_group_info array. We round it * to the next power of two because this approximation is done * internally by kmalloc so we can have some more memory * for free here (e.g. may be used for META_BG resize). */ array_size = 1; while (array_size < sizeof(*sbi->s_group_info) * num_meta_group_infos_max) array_size = array_size << 1; |
c9de560de ext4: Add multi b... |
2327 2328 2329 |
/* An 8TB filesystem with 64-bit pointers requires a 4096 byte * kmalloc. A 128kb malloc should suffice for a 256TB filesystem. * So a two level scheme suffices for now. */ |
5f21b0e64 ext4: fix online ... |
2330 |
sbi->s_group_info = kmalloc(array_size, GFP_KERNEL); |
c9de560de ext4: Add multi b... |
2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 |
if (sbi->s_group_info == NULL) { printk(KERN_ERR "EXT4-fs: can't allocate buddy meta group "); return -ENOMEM; } sbi->s_buddy_cache = new_inode(sb); if (sbi->s_buddy_cache == NULL) { printk(KERN_ERR "EXT4-fs: can't get new inode "); goto err_freesgi; } EXT4_I(sbi->s_buddy_cache)->i_disksize = 0; |
8df9675f8 ext4: Avoid races... |
2343 |
for (i = 0; i < ngroups; i++) { |
c9de560de ext4: Add multi b... |
2344 2345 2346 |
desc = ext4_get_group_desc(sb, i, NULL); if (desc == NULL) { printk(KERN_ERR |
a9df9a491 ext4: Make ext4_g... |
2347 2348 |
"EXT4-fs: can't read descriptor %u ", i); |
c9de560de ext4: Add multi b... |
2349 2350 |
goto err_freebuddy; } |
5f21b0e64 ext4: fix online ... |
2351 2352 |
if (ext4_mb_add_groupinfo(sb, i, desc) != 0) goto err_freebuddy; |
c9de560de ext4: Add multi b... |
2353 2354 2355 2356 2357 |
} return 0; err_freebuddy: |
f1fa3342e ext4: fix hot spi... |
2358 |
while (i-- > 0) |
c9de560de ext4: Add multi b... |
2359 |
kfree(ext4_get_group_info(sb, i)); |
c9de560de ext4: Add multi b... |
2360 |
i = num_meta_group_infos; |
f1fa3342e ext4: fix hot spi... |
2361 |
while (i-- > 0) |
c9de560de ext4: Add multi b... |
2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 |
kfree(sbi->s_group_info[i]); iput(sbi->s_buddy_cache); err_freesgi: kfree(sbi->s_group_info); return -ENOMEM; } int ext4_mb_init(struct super_block *sb, int needs_recovery) { struct ext4_sb_info *sbi = EXT4_SB(sb); |
6be2ded1d ext4: Don't allow... |
2372 |
unsigned i, j; |
c9de560de ext4: Add multi b... |
2373 2374 |
unsigned offset; unsigned max; |
74767c5a2 ext4: miscellaneo... |
2375 |
int ret; |
c9de560de ext4: Add multi b... |
2376 |
|
1927805e6 ext4: use variabl... |
2377 |
i = (sb->s_blocksize_bits + 2) * sizeof(*sbi->s_mb_offsets); |
c9de560de ext4: Add multi b... |
2378 2379 2380 |
sbi->s_mb_offsets = kmalloc(i, GFP_KERNEL); if (sbi->s_mb_offsets == NULL) { |
c9de560de ext4: Add multi b... |
2381 2382 |
return -ENOMEM; } |
ff7ef329b ext4: Widen type ... |
2383 |
|
1927805e6 ext4: use variabl... |
2384 |
i = (sb->s_blocksize_bits + 2) * sizeof(*sbi->s_mb_maxs); |
c9de560de ext4: Add multi b... |
2385 2386 |
sbi->s_mb_maxs = kmalloc(i, GFP_KERNEL); if (sbi->s_mb_maxs == NULL) { |
a7b19448d ext4: fix typo wh... |
2387 |
kfree(sbi->s_mb_offsets); |
c9de560de ext4: Add multi b... |
2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 |
return -ENOMEM; } /* order 0 is regular bitmap */ sbi->s_mb_maxs[0] = sb->s_blocksize << 3; sbi->s_mb_offsets[0] = 0; i = 1; offset = 0; max = sb->s_blocksize << 2; do { sbi->s_mb_offsets[i] = offset; sbi->s_mb_maxs[i] = max; offset += 1 << (sb->s_blocksize_bits - i); max = max >> 1; i++; } while (i <= sb->s_blocksize_bits + 1); /* init file for buddy data */ |
74767c5a2 ext4: miscellaneo... |
2407 2408 |
ret = ext4_mb_init_backend(sb); if (ret != 0) { |
c9de560de ext4: Add multi b... |
2409 2410 |
kfree(sbi->s_mb_offsets); kfree(sbi->s_mb_maxs); |
74767c5a2 ext4: miscellaneo... |
2411 |
return ret; |
c9de560de ext4: Add multi b... |
2412 2413 2414 |
} spin_lock_init(&sbi->s_md_lock); |
c9de560de ext4: Add multi b... |
2415 2416 2417 2418 2419 2420 2421 |
spin_lock_init(&sbi->s_bal_lock); sbi->s_mb_max_to_scan = MB_DEFAULT_MAX_TO_SCAN; sbi->s_mb_min_to_scan = MB_DEFAULT_MIN_TO_SCAN; sbi->s_mb_stats = MB_DEFAULT_STATS; sbi->s_mb_stream_request = MB_DEFAULT_STREAM_THRESHOLD; sbi->s_mb_order2_reqs = MB_DEFAULT_ORDER2_REQS; |
c9de560de ext4: Add multi b... |
2422 |
sbi->s_mb_group_prealloc = MB_DEFAULT_GROUP_PREALLOC; |
730c213c7 ext4: use percpu ... |
2423 |
sbi->s_locality_groups = alloc_percpu(struct ext4_locality_group); |
c9de560de ext4: Add multi b... |
2424 |
if (sbi->s_locality_groups == NULL) { |
c9de560de ext4: Add multi b... |
2425 2426 2427 2428 |
kfree(sbi->s_mb_offsets); kfree(sbi->s_mb_maxs); return -ENOMEM; } |
730c213c7 ext4: use percpu ... |
2429 |
for_each_possible_cpu(i) { |
c9de560de ext4: Add multi b... |
2430 |
struct ext4_locality_group *lg; |
730c213c7 ext4: use percpu ... |
2431 |
lg = per_cpu_ptr(sbi->s_locality_groups, i); |
c9de560de ext4: Add multi b... |
2432 |
mutex_init(&lg->lg_mutex); |
6be2ded1d ext4: Don't allow... |
2433 2434 |
for (j = 0; j < PREALLOC_TB_SIZE; j++) INIT_LIST_HEAD(&lg->lg_prealloc_list[j]); |
c9de560de ext4: Add multi b... |
2435 2436 |
spin_lock_init(&lg->lg_prealloc_lock); } |
296c355cd ext4: Use tracepo... |
2437 2438 2439 |
if (sbi->s_proc) proc_create_data("mb_groups", S_IRUGO, sbi->s_proc, &ext4_mb_seq_groups_fops, sb); |
c9de560de ext4: Add multi b... |
2440 |
|
0390131ba ext4: Allow ext4 ... |
2441 2442 |
if (sbi->s_journal) sbi->s_journal->j_commit_callback = release_blocks_on_commit; |
c9de560de ext4: Add multi b... |
2443 2444 |
return 0; } |
955ce5f5b ext4: Convert ext... |
2445 |
/* need to called with the ext4 group lock held */ |
c9de560de ext4: Add multi b... |
2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 |
static void ext4_mb_cleanup_pa(struct ext4_group_info *grp) { struct ext4_prealloc_space *pa; struct list_head *cur, *tmp; int count = 0; list_for_each_safe(cur, tmp, &grp->bb_prealloc_list) { pa = list_entry(cur, struct ext4_prealloc_space, pa_group_list); list_del(&pa->pa_group_list); count++; |
688f05a01 ext4: Free ext4_p... |
2456 |
kmem_cache_free(ext4_pspace_cachep, pa); |
c9de560de ext4: Add multi b... |
2457 2458 |
} if (count) |
6ba495e92 ext4: Add configu... |
2459 2460 |
mb_debug(1, "mballoc: %u PAs left ", count); |
c9de560de ext4: Add multi b... |
2461 2462 2463 2464 2465 |
} int ext4_mb_release(struct super_block *sb) { |
8df9675f8 ext4: Avoid races... |
2466 |
ext4_group_t ngroups = ext4_get_groups_count(sb); |
c9de560de ext4: Add multi b... |
2467 2468 2469 2470 |
ext4_group_t i; int num_meta_group_infos; struct ext4_group_info *grinfo; struct ext4_sb_info *sbi = EXT4_SB(sb); |
c9de560de ext4: Add multi b... |
2471 |
if (sbi->s_group_info) { |
8df9675f8 ext4: Avoid races... |
2472 |
for (i = 0; i < ngroups; i++) { |
c9de560de ext4: Add multi b... |
2473 2474 2475 2476 2477 2478 2479 2480 2481 |
grinfo = ext4_get_group_info(sb, i); #ifdef DOUBLE_CHECK kfree(grinfo->bb_bitmap); #endif ext4_lock_group(sb, i); ext4_mb_cleanup_pa(grinfo); ext4_unlock_group(sb, i); kfree(grinfo); } |
8df9675f8 ext4: Avoid races... |
2482 |
num_meta_group_infos = (ngroups + |
c9de560de ext4: Add multi b... |
2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 |
EXT4_DESC_PER_BLOCK(sb) - 1) >> EXT4_DESC_PER_BLOCK_BITS(sb); for (i = 0; i < num_meta_group_infos; i++) kfree(sbi->s_group_info[i]); kfree(sbi->s_group_info); } kfree(sbi->s_mb_offsets); kfree(sbi->s_mb_maxs); if (sbi->s_buddy_cache) iput(sbi->s_buddy_cache); if (sbi->s_mb_stats) { printk(KERN_INFO "EXT4-fs: mballoc: %u blocks %u reqs (%u success) ", atomic_read(&sbi->s_bal_allocated), atomic_read(&sbi->s_bal_reqs), atomic_read(&sbi->s_bal_success)); printk(KERN_INFO "EXT4-fs: mballoc: %u extents scanned, %u goal hits, " "%u 2^N hits, %u breaks, %u lost ", atomic_read(&sbi->s_bal_ex_scanned), atomic_read(&sbi->s_bal_goals), atomic_read(&sbi->s_bal_2orders), atomic_read(&sbi->s_bal_breaks), atomic_read(&sbi->s_mb_lost_chunks)); printk(KERN_INFO "EXT4-fs: mballoc: %lu generated and it took %Lu ", sbi->s_mb_buddies_generated++, sbi->s_mb_generation_time); printk(KERN_INFO "EXT4-fs: mballoc: %u preallocated, %u discarded ", atomic_read(&sbi->s_mb_preallocated), atomic_read(&sbi->s_mb_discarded)); } |
730c213c7 ext4: use percpu ... |
2520 |
free_percpu(sbi->s_locality_groups); |
296c355cd ext4: Use tracepo... |
2521 2522 |
if (sbi->s_proc) remove_proc_entry("mb_groups", sbi->s_proc); |
c9de560de ext4: Add multi b... |
2523 2524 2525 |
return 0; } |
5c521830c ext4: Support dis... |
2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 |
static inline void ext4_issue_discard(struct super_block *sb, ext4_group_t block_group, ext4_grpblk_t block, int count) { int ret; ext4_fsblk_t discard_block; discard_block = block + ext4_group_first_block_no(sb, block_group); trace_ext4_discard_blocks(sb, (unsigned long long) discard_block, count); ret = sb_issue_discard(sb, discard_block, count); if (ret == EOPNOTSUPP) { ext4_warning(sb, "discard not supported, disabling"); clear_opt(EXT4_SB(sb)->s_mount_opt, DISCARD); } } |
3e624fc72 ext4: Replace hac... |
2541 2542 2543 2544 2545 |
/* * This function is called by the jbd2 layer once the commit has finished, * so we know we can free the blocks that were released with that commit. */ static void release_blocks_on_commit(journal_t *journal, transaction_t *txn) |
c9de560de ext4: Add multi b... |
2546 |
{ |
3e624fc72 ext4: Replace hac... |
2547 |
struct super_block *sb = journal->j_private; |
c9de560de ext4: Add multi b... |
2548 |
struct ext4_buddy e4b; |
c894058d6 ext4: Use an rbtr... |
2549 |
struct ext4_group_info *db; |
c894058d6 ext4: Use an rbtr... |
2550 2551 |
int err, count = 0, count2 = 0; struct ext4_free_data *entry; |
3e624fc72 ext4: Replace hac... |
2552 |
struct list_head *l, *ltmp; |
c9de560de ext4: Add multi b... |
2553 |
|
3e624fc72 ext4: Replace hac... |
2554 2555 |
list_for_each_safe(l, ltmp, &txn->t_private_list) { entry = list_entry(l, struct ext4_free_data, list); |
c9de560de ext4: Add multi b... |
2556 |
|
6ba495e92 ext4: Add configu... |
2557 |
mb_debug(1, "gonna free %u blocks in group %u (0x%p):", |
3e624fc72 ext4: Replace hac... |
2558 |
entry->count, entry->group, entry); |
c9de560de ext4: Add multi b... |
2559 |
|
5c521830c ext4: Support dis... |
2560 2561 2562 |
if (test_opt(sb, DISCARD)) ext4_issue_discard(sb, entry->group, entry->start_blk, entry->count); |
b90f68701 ext4: Issue the d... |
2563 |
|
c894058d6 ext4: Use an rbtr... |
2564 |
err = ext4_mb_load_buddy(sb, entry->group, &e4b); |
c9de560de ext4: Add multi b... |
2565 2566 |
/* we expect to find existing buddy because it's pinned */ BUG_ON(err != 0); |
c894058d6 ext4: Use an rbtr... |
2567 |
db = e4b.bd_info; |
c9de560de ext4: Add multi b... |
2568 |
/* there are blocks to put in buddy to make them really free */ |
c894058d6 ext4: Use an rbtr... |
2569 |
count += entry->count; |
c9de560de ext4: Add multi b... |
2570 |
count2++; |
c894058d6 ext4: Use an rbtr... |
2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 |
ext4_lock_group(sb, entry->group); /* Take it out of per group rb tree */ rb_erase(&entry->node, &(db->bb_free_root)); mb_free_blocks(NULL, &e4b, entry->start_blk, entry->count); if (!db->bb_free_root.rb_node) { /* No more items in the per group rb tree * balance refcounts from ext4_mb_free_metadata() */ page_cache_release(e4b.bd_buddy_page); page_cache_release(e4b.bd_bitmap_page); |
c9de560de ext4: Add multi b... |
2582 |
} |
c894058d6 ext4: Use an rbtr... |
2583 |
ext4_unlock_group(sb, entry->group); |
c894058d6 ext4: Use an rbtr... |
2584 |
kmem_cache_free(ext4_free_ext_cachep, entry); |
e39e07fdf ext4: rename ext4... |
2585 |
ext4_mb_unload_buddy(&e4b); |
3e624fc72 ext4: Replace hac... |
2586 |
} |
c9de560de ext4: Add multi b... |
2587 |
|
6ba495e92 ext4: Add configu... |
2588 2589 |
mb_debug(1, "freed %u blocks in %u structures ", count, count2); |
c9de560de ext4: Add multi b... |
2590 |
} |
6ba495e92 ext4: Add configu... |
2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 |
#ifdef CONFIG_EXT4_DEBUG u8 mb_enable_debug __read_mostly; static struct dentry *debugfs_dir; static struct dentry *debugfs_debug; static void __init ext4_create_debugfs_entry(void) { debugfs_dir = debugfs_create_dir("ext4", NULL); if (debugfs_dir) debugfs_debug = debugfs_create_u8("mballoc-debug", S_IRUGO | S_IWUSR, debugfs_dir, &mb_enable_debug); } static void ext4_remove_debugfs_entry(void) { debugfs_remove(debugfs_debug); debugfs_remove(debugfs_dir); } #else static void __init ext4_create_debugfs_entry(void) { } static void ext4_remove_debugfs_entry(void) { } #endif |
c9de560de ext4: Add multi b... |
2624 2625 2626 2627 2628 2629 2630 2631 |
int __init init_ext4_mballoc(void) { ext4_pspace_cachep = kmem_cache_create("ext4_prealloc_space", sizeof(struct ext4_prealloc_space), 0, SLAB_RECLAIM_ACCOUNT, NULL); if (ext4_pspace_cachep == NULL) return -ENOMEM; |
256bdb497 ext4: allocate st... |
2632 2633 2634 2635 2636 2637 2638 2639 |
ext4_ac_cachep = kmem_cache_create("ext4_alloc_context", sizeof(struct ext4_allocation_context), 0, SLAB_RECLAIM_ACCOUNT, NULL); if (ext4_ac_cachep == NULL) { kmem_cache_destroy(ext4_pspace_cachep); return -ENOMEM; } |
c894058d6 ext4: Use an rbtr... |
2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 |
ext4_free_ext_cachep = kmem_cache_create("ext4_free_block_extents", sizeof(struct ext4_free_data), 0, SLAB_RECLAIM_ACCOUNT, NULL); if (ext4_free_ext_cachep == NULL) { kmem_cache_destroy(ext4_pspace_cachep); kmem_cache_destroy(ext4_ac_cachep); return -ENOMEM; } |
6ba495e92 ext4: Add configu... |
2650 |
ext4_create_debugfs_entry(); |
c9de560de ext4: Add multi b... |
2651 2652 2653 2654 2655 |
return 0; } void exit_ext4_mballoc(void) { |
60e6679e2 ext4: Drop whites... |
2656 |
/* |
3e03f9ca6 ext4: Use rcu_bar... |
2657 2658 2659 2660 |
* Wait for completion of call_rcu()'s on ext4_pspace_cachep * before destroying the slab cache. */ rcu_barrier(); |
c9de560de ext4: Add multi b... |
2661 |
kmem_cache_destroy(ext4_pspace_cachep); |
256bdb497 ext4: allocate st... |
2662 |
kmem_cache_destroy(ext4_ac_cachep); |
c894058d6 ext4: Use an rbtr... |
2663 |
kmem_cache_destroy(ext4_free_ext_cachep); |
6ba495e92 ext4: Add configu... |
2664 |
ext4_remove_debugfs_entry(); |
c9de560de ext4: Add multi b... |
2665 2666 2667 2668 |
} /* |
73b2c7165 fix comment typo ... |
2669 |
* Check quota and mark chosen space (ac->ac_b_ex) non-free in bitmaps |
c9de560de ext4: Add multi b... |
2670 2671 |
* Returns 0 if success or error code */ |
4ddfef7b4 ext4: reduce mbal... |
2672 2673 |
static noinline_for_stack int ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac, |
498e5f241 ext4: Change unsi... |
2674 |
handle_t *handle, unsigned int reserv_blks) |
c9de560de ext4: Add multi b... |
2675 2676 |
{ struct buffer_head *bitmap_bh = NULL; |
c9de560de ext4: Add multi b... |
2677 2678 2679 2680 2681 |
struct ext4_group_desc *gdp; struct buffer_head *gdp_bh; struct ext4_sb_info *sbi; struct super_block *sb; ext4_fsblk_t block; |
519deca04 ext4: Retry block... |
2682 |
int err, len; |
c9de560de ext4: Add multi b... |
2683 2684 2685 2686 2687 2688 |
BUG_ON(ac->ac_status != AC_STATUS_FOUND); BUG_ON(ac->ac_b_ex.fe_len <= 0); sb = ac->ac_sb; sbi = EXT4_SB(sb); |
c9de560de ext4: Add multi b... |
2689 2690 |
err = -EIO; |
574ca174c ext4: Rename read... |
2691 |
bitmap_bh = ext4_read_block_bitmap(sb, ac->ac_b_ex.fe_group); |
c9de560de ext4: Add multi b... |
2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 |
if (!bitmap_bh) goto out_err; err = ext4_journal_get_write_access(handle, bitmap_bh); if (err) goto out_err; err = -EIO; gdp = ext4_get_group_desc(sb, ac->ac_b_ex.fe_group, &gdp_bh); if (!gdp) goto out_err; |
a9df9a491 ext4: Make ext4_g... |
2703 2704 |
ext4_debug("using block group %u(%d) ", ac->ac_b_ex.fe_group, |
9fd9784c9 ext4: Fix buildin... |
2705 |
ext4_free_blks_count(sb, gdp)); |
03cddb80e ext4: Fix use of ... |
2706 |
|
c9de560de ext4: Add multi b... |
2707 2708 2709 |
err = ext4_journal_get_write_access(handle, gdp_bh); if (err) goto out_err; |
bda00de7e ext4: cleanup to ... |
2710 |
block = ext4_grp_offs_to_block(sb, &ac->ac_b_ex); |
c9de560de ext4: Add multi b... |
2711 |
|
519deca04 ext4: Retry block... |
2712 |
len = ac->ac_b_ex.fe_len; |
6fd058f77 ext4: Add a compr... |
2713 |
if (!ext4_data_block_valid(sbi, block, len)) { |
12062dddd ext4: move __func... |
2714 |
ext4_error(sb, "Allocating blocks %llu-%llu which overlap " |
6fd058f77 ext4: Add a compr... |
2715 2716 |
"fs metadata ", block, block+len); |
519deca04 ext4: Retry block... |
2717 2718 2719 2720 |
/* File system mounted not to panic on error * Fix the bitmap and repeat the block allocation * We leak some of the blocks here. */ |
955ce5f5b ext4: Convert ext... |
2721 2722 2723 2724 |
ext4_lock_group(sb, ac->ac_b_ex.fe_group); mb_set_bits(bitmap_bh->b_data, ac->ac_b_ex.fe_start, ac->ac_b_ex.fe_len); ext4_unlock_group(sb, ac->ac_b_ex.fe_group); |
0390131ba ext4: Allow ext4 ... |
2725 |
err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh); |
519deca04 ext4: Retry block... |
2726 2727 2728 |
if (!err) err = -EAGAIN; goto out_err; |
c9de560de ext4: Add multi b... |
2729 |
} |
955ce5f5b ext4: Convert ext... |
2730 2731 |
ext4_lock_group(sb, ac->ac_b_ex.fe_group); |
c9de560de ext4: Add multi b... |
2732 2733 2734 2735 2736 2737 2738 2739 2740 |
#ifdef AGGRESSIVE_CHECK { int i; for (i = 0; i < ac->ac_b_ex.fe_len; i++) { BUG_ON(mb_test_bit(ac->ac_b_ex.fe_start + i, bitmap_bh->b_data)); } } #endif |
955ce5f5b ext4: Convert ext... |
2741 |
mb_set_bits(bitmap_bh->b_data, ac->ac_b_ex.fe_start,ac->ac_b_ex.fe_len); |
c9de560de ext4: Add multi b... |
2742 2743 |
if (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) { gdp->bg_flags &= cpu_to_le16(~EXT4_BG_BLOCK_UNINIT); |
560671a0d ext4: Use high 16... |
2744 2745 2746 |
ext4_free_blks_set(sb, gdp, ext4_free_blocks_after_init(sb, ac->ac_b_ex.fe_group, gdp)); |
c9de560de ext4: Add multi b... |
2747 |
} |
560671a0d ext4: Use high 16... |
2748 2749 |
len = ext4_free_blks_count(sb, gdp) - ac->ac_b_ex.fe_len; ext4_free_blks_set(sb, gdp, len); |
c9de560de ext4: Add multi b... |
2750 |
gdp->bg_checksum = ext4_group_desc_csum(sbi, ac->ac_b_ex.fe_group, gdp); |
955ce5f5b ext4: Convert ext... |
2751 2752 |
ext4_unlock_group(sb, ac->ac_b_ex.fe_group); |
6bc6e63fc ext4: Add percpu ... |
2753 |
percpu_counter_sub(&sbi->s_freeblocks_counter, ac->ac_b_ex.fe_len); |
d2a176379 ext4: delayed all... |
2754 |
/* |
6bc6e63fc ext4: Add percpu ... |
2755 |
* Now reduce the dirty block count also. Should not go negative |
d2a176379 ext4: delayed all... |
2756 |
*/ |
6bc6e63fc ext4: Add percpu ... |
2757 2758 2759 |
if (!(ac->ac_flags & EXT4_MB_DELALLOC_RESERVED)) /* release all the reserved blocks if non delalloc */ percpu_counter_sub(&sbi->s_dirtyblocks_counter, reserv_blks); |
c9de560de ext4: Add multi b... |
2760 |
|
772cb7c83 ext4: New inode a... |
2761 2762 2763 |
if (sbi->s_log_groups_per_flex) { ext4_group_t flex_group = ext4_flex_group(sbi, ac->ac_b_ex.fe_group); |
9f24e4208 ext4: Use atomic_... |
2764 2765 |
atomic_sub(ac->ac_b_ex.fe_len, &sbi->s_flex_groups[flex_group].free_blocks); |
772cb7c83 ext4: New inode a... |
2766 |
} |
0390131ba ext4: Allow ext4 ... |
2767 |
err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh); |
c9de560de ext4: Add multi b... |
2768 2769 |
if (err) goto out_err; |
0390131ba ext4: Allow ext4 ... |
2770 |
err = ext4_handle_dirty_metadata(handle, NULL, gdp_bh); |
c9de560de ext4: Add multi b... |
2771 2772 |
out_err: |
a0375156c ext4: Clean up s_... |
2773 |
ext4_mark_super_dirty(sb); |
42a10add8 ext4: Fix null bh... |
2774 |
brelse(bitmap_bh); |
c9de560de ext4: Add multi b... |
2775 2776 2777 2778 2779 2780 2781 |
return err; } /* * here we normalize request for locality group * Group request are normalized to s_strip size if we set the same via mount * option. If not we set it to s_mb_group_prealloc which can be configured via |
b713a5ec5 ext4: remove /pro... |
2782 |
* /sys/fs/ext4/<partition>/mb_group_prealloc |
c9de560de ext4: Add multi b... |
2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 |
* * XXX: should we try to preallocate more than the group has now? */ static void ext4_mb_normalize_group_request(struct ext4_allocation_context *ac) { struct super_block *sb = ac->ac_sb; struct ext4_locality_group *lg = ac->ac_lg; BUG_ON(lg == NULL); if (EXT4_SB(sb)->s_stripe) ac->ac_g_ex.fe_len = EXT4_SB(sb)->s_stripe; else ac->ac_g_ex.fe_len = EXT4_SB(sb)->s_mb_group_prealloc; |
6ba495e92 ext4: Add configu... |
2796 2797 |
mb_debug(1, "#%u: goal %u blocks for locality group ", |
c9de560de ext4: Add multi b... |
2798 2799 2800 2801 2802 2803 2804 |
current->pid, ac->ac_g_ex.fe_len); } /* * Normalization means making request better in terms of * size and alignment */ |
4ddfef7b4 ext4: reduce mbal... |
2805 2806 |
static noinline_for_stack void ext4_mb_normalize_request(struct ext4_allocation_context *ac, |
c9de560de ext4: Add multi b... |
2807 2808 2809 2810 |
struct ext4_allocation_request *ar) { int bsbits, max; ext4_lblk_t end; |
c9de560de ext4: Add multi b... |
2811 |
loff_t size, orig_size, start_off; |
5a0790c2c ext4: remove init... |
2812 |
ext4_lblk_t start; |
c9de560de ext4: Add multi b... |
2813 |
struct ext4_inode_info *ei = EXT4_I(ac->ac_inode); |
9a0762c5a ext4: Convert li... |
2814 |
struct ext4_prealloc_space *pa; |
c9de560de ext4: Add multi b... |
2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 |
/* do normalize only data requests, metadata requests do not need preallocation */ if (!(ac->ac_flags & EXT4_MB_HINT_DATA)) return; /* sometime caller may want exact blocks */ if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY)) return; /* caller may indicate that preallocation isn't * required (it's a tail, for example) */ if (ac->ac_flags & EXT4_MB_HINT_NOPREALLOC) return; if (ac->ac_flags & EXT4_MB_HINT_GROUP_ALLOC) { ext4_mb_normalize_group_request(ac); return ; } bsbits = ac->ac_sb->s_blocksize_bits; /* first, let's learn actual file size * given current request is allocated */ size = ac->ac_o_ex.fe_logical + ac->ac_o_ex.fe_len; size = size << bsbits; if (size < i_size_read(ac->ac_inode)) size = i_size_read(ac->ac_inode); |
5a0790c2c ext4: remove init... |
2843 |
orig_size = size; |
c9de560de ext4: Add multi b... |
2844 |
|
1930479c4 ext4: mballoc fix... |
2845 2846 |
/* max size of free chunks */ max = 2 << bsbits; |
c9de560de ext4: Add multi b... |
2847 |
|
1930479c4 ext4: mballoc fix... |
2848 2849 |
#define NRL_CHECK_SIZE(req, size, max, chunk_size) \ (req <= (size) || max <= (chunk_size)) |
c9de560de ext4: Add multi b... |
2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 |
/* first, try to predict filesize */ /* XXX: should this table be tunable? */ start_off = 0; if (size <= 16 * 1024) { size = 16 * 1024; } else if (size <= 32 * 1024) { size = 32 * 1024; } else if (size <= 64 * 1024) { size = 64 * 1024; } else if (size <= 128 * 1024) { size = 128 * 1024; } else if (size <= 256 * 1024) { size = 256 * 1024; } else if (size <= 512 * 1024) { size = 512 * 1024; } else if (size <= 1024 * 1024) { size = 1024 * 1024; |
1930479c4 ext4: mballoc fix... |
2868 |
} else if (NRL_CHECK_SIZE(size, 4 * 1024 * 1024, max, 2 * 1024)) { |
c9de560de ext4: Add multi b... |
2869 |
start_off = ((loff_t)ac->ac_o_ex.fe_logical >> |
1930479c4 ext4: mballoc fix... |
2870 2871 2872 |
(21 - bsbits)) << 21; size = 2 * 1024 * 1024; } else if (NRL_CHECK_SIZE(size, 8 * 1024 * 1024, max, 4 * 1024)) { |
c9de560de ext4: Add multi b... |
2873 2874 2875 2876 |
start_off = ((loff_t)ac->ac_o_ex.fe_logical >> (22 - bsbits)) << 22; size = 4 * 1024 * 1024; } else if (NRL_CHECK_SIZE(ac->ac_o_ex.fe_len, |
1930479c4 ext4: mballoc fix... |
2877 |
(8<<20)>>bsbits, max, 8 * 1024)) { |
c9de560de ext4: Add multi b... |
2878 2879 2880 2881 2882 2883 2884 |
start_off = ((loff_t)ac->ac_o_ex.fe_logical >> (23 - bsbits)) << 23; size = 8 * 1024 * 1024; } else { start_off = (loff_t)ac->ac_o_ex.fe_logical << bsbits; size = ac->ac_o_ex.fe_len << bsbits; } |
5a0790c2c ext4: remove init... |
2885 2886 |
size = size >> bsbits; start = start_off >> bsbits; |
c9de560de ext4: Add multi b... |
2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 |
/* don't cover already allocated blocks in selected range */ if (ar->pleft && start <= ar->lleft) { size -= ar->lleft + 1 - start; start = ar->lleft + 1; } if (ar->pright && start + size - 1 >= ar->lright) size -= start + size - ar->lright; end = start + size; /* check we don't cross already preallocated blocks */ rcu_read_lock(); |
9a0762c5a ext4: Convert li... |
2900 |
list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) { |
498e5f241 ext4: Change unsi... |
2901 |
ext4_lblk_t pa_end; |
c9de560de ext4: Add multi b... |
2902 |
|
c9de560de ext4: Add multi b... |
2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 |
if (pa->pa_deleted) continue; spin_lock(&pa->pa_lock); if (pa->pa_deleted) { spin_unlock(&pa->pa_lock); continue; } pa_end = pa->pa_lstart + pa->pa_len; /* PA must not overlap original request */ BUG_ON(!(ac->ac_o_ex.fe_logical >= pa_end || ac->ac_o_ex.fe_logical < pa->pa_lstart)); |
38877f4e8 simplify some log... |
2916 2917 |
/* skip PAs this normalized request doesn't overlap with */ if (pa->pa_lstart >= end || pa_end <= start) { |
c9de560de ext4: Add multi b... |
2918 2919 2920 2921 |
spin_unlock(&pa->pa_lock); continue; } BUG_ON(pa->pa_lstart <= start && pa_end >= end); |
38877f4e8 simplify some log... |
2922 |
/* adjust start or end to be adjacent to this pa */ |
c9de560de ext4: Add multi b... |
2923 2924 2925 |
if (pa_end <= ac->ac_o_ex.fe_logical) { BUG_ON(pa_end < start); start = pa_end; |
38877f4e8 simplify some log... |
2926 |
} else if (pa->pa_lstart > ac->ac_o_ex.fe_logical) { |
c9de560de ext4: Add multi b... |
2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 |
BUG_ON(pa->pa_lstart > end); end = pa->pa_lstart; } spin_unlock(&pa->pa_lock); } rcu_read_unlock(); size = end - start; /* XXX: extra loop to check we really don't overlap preallocations */ rcu_read_lock(); |
9a0762c5a ext4: Convert li... |
2937 |
list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) { |
498e5f241 ext4: Change unsi... |
2938 |
ext4_lblk_t pa_end; |
c9de560de ext4: Add multi b... |
2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 |
spin_lock(&pa->pa_lock); if (pa->pa_deleted == 0) { pa_end = pa->pa_lstart + pa->pa_len; BUG_ON(!(start >= pa_end || end <= pa->pa_lstart)); } spin_unlock(&pa->pa_lock); } rcu_read_unlock(); if (start + size <= ac->ac_o_ex.fe_logical && start > ac->ac_o_ex.fe_logical) { printk(KERN_ERR "start %lu, size %lu, fe_logical %lu ", (unsigned long) start, (unsigned long) size, (unsigned long) ac->ac_o_ex.fe_logical); } BUG_ON(start + size <= ac->ac_o_ex.fe_logical && start > ac->ac_o_ex.fe_logical); |
8d03c7a0c ext4: fix bogus B... |
2957 |
BUG_ON(size <= 0 || size > EXT4_BLOCKS_PER_GROUP(ac->ac_sb)); |
c9de560de ext4: Add multi b... |
2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 |
/* now prepare goal request */ /* XXX: is it better to align blocks WRT to logical * placement or satisfy big request as is */ ac->ac_g_ex.fe_logical = start; ac->ac_g_ex.fe_len = size; /* define goal start in order to merge */ if (ar->pright && (ar->lright == (start + size))) { /* merge to the right */ ext4_get_group_no_and_offset(ac->ac_sb, ar->pright - size, &ac->ac_f_ex.fe_group, &ac->ac_f_ex.fe_start); ac->ac_flags |= EXT4_MB_HINT_TRY_GOAL; } if (ar->pleft && (ar->lleft + 1 == start)) { /* merge to the left */ ext4_get_group_no_and_offset(ac->ac_sb, ar->pleft + 1, &ac->ac_f_ex.fe_group, &ac->ac_f_ex.fe_start); ac->ac_flags |= EXT4_MB_HINT_TRY_GOAL; } |
6ba495e92 ext4: Add configu... |
2981 2982 |
mb_debug(1, "goal: %u(was %u) blocks at %u ", (unsigned) size, |
c9de560de ext4: Add multi b... |
2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 |
(unsigned) orig_size, (unsigned) start); } static void ext4_mb_collect_stats(struct ext4_allocation_context *ac) { struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); if (sbi->s_mb_stats && ac->ac_g_ex.fe_len > 1) { atomic_inc(&sbi->s_bal_reqs); atomic_add(ac->ac_b_ex.fe_len, &sbi->s_bal_allocated); |
291dae472 ext4: Fix for ext... |
2993 |
if (ac->ac_b_ex.fe_len >= ac->ac_o_ex.fe_len) |
c9de560de ext4: Add multi b... |
2994 2995 2996 2997 2998 2999 3000 3001 |
atomic_inc(&sbi->s_bal_success); atomic_add(ac->ac_found, &sbi->s_bal_ex_scanned); if (ac->ac_g_ex.fe_start == ac->ac_b_ex.fe_start && ac->ac_g_ex.fe_group == ac->ac_b_ex.fe_group) atomic_inc(&sbi->s_bal_goals); if (ac->ac_found > sbi->s_mb_max_to_scan) atomic_inc(&sbi->s_bal_breaks); } |
296c355cd ext4: Use tracepo... |
3002 3003 3004 3005 |
if (ac->ac_op == EXT4_MB_HISTORY_ALLOC) trace_ext4_mballoc_alloc(ac); else trace_ext4_mballoc_prealloc(ac); |
c9de560de ext4: Add multi b... |
3006 3007 3008 |
} /* |
b844167ed ext4: remove bloc... |
3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 |
* Called on failure; free up any blocks from the inode PA for this * context. We don't need this for MB_GROUP_PA because we only change * pa_free in ext4_mb_release_context(), but on failure, we've already * zeroed out ac->ac_b_ex.fe_len, so group_pa->pa_free is not changed. */ static void ext4_discard_allocated_blocks(struct ext4_allocation_context *ac) { struct ext4_prealloc_space *pa = ac->ac_pa; int len; if (pa && pa->pa_type == MB_INODE_PA) { len = ac->ac_b_ex.fe_len; pa->pa_free += len; } } /* |
c9de560de ext4: Add multi b... |
3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 |
* use blocks preallocated to inode */ static void ext4_mb_use_inode_pa(struct ext4_allocation_context *ac, struct ext4_prealloc_space *pa) { ext4_fsblk_t start; ext4_fsblk_t end; int len; /* found preallocated blocks, use them */ start = pa->pa_pstart + (ac->ac_o_ex.fe_logical - pa->pa_lstart); end = min(pa->pa_pstart + pa->pa_len, start + ac->ac_o_ex.fe_len); len = end - start; ext4_get_group_no_and_offset(ac->ac_sb, start, &ac->ac_b_ex.fe_group, &ac->ac_b_ex.fe_start); ac->ac_b_ex.fe_len = len; ac->ac_status = AC_STATUS_FOUND; ac->ac_pa = pa; BUG_ON(start < pa->pa_pstart); BUG_ON(start + len > pa->pa_pstart + pa->pa_len); BUG_ON(pa->pa_free < len); pa->pa_free -= len; |
6ba495e92 ext4: Add configu... |
3050 3051 |
mb_debug(1, "use %llu/%u from inode pa %p ", start, len, pa); |
c9de560de ext4: Add multi b... |
3052 3053 3054 3055 3056 3057 3058 3059 |
} /* * use blocks preallocated to locality group */ static void ext4_mb_use_group_pa(struct ext4_allocation_context *ac, struct ext4_prealloc_space *pa) { |
03cddb80e ext4: Fix use of ... |
3060 |
unsigned int len = ac->ac_o_ex.fe_len; |
6be2ded1d ext4: Don't allow... |
3061 |
|
c9de560de ext4: Add multi b... |
3062 3063 3064 3065 3066 3067 3068 3069 |
ext4_get_group_no_and_offset(ac->ac_sb, pa->pa_pstart, &ac->ac_b_ex.fe_group, &ac->ac_b_ex.fe_start); ac->ac_b_ex.fe_len = len; ac->ac_status = AC_STATUS_FOUND; ac->ac_pa = pa; /* we don't correct pa_pstart or pa_plen here to avoid |
26346ff68 ext4: Don't panic... |
3070 |
* possible race when the group is being loaded concurrently |
c9de560de ext4: Add multi b... |
3071 |
* instead we correct pa later, after blocks are marked |
26346ff68 ext4: Don't panic... |
3072 3073 |
* in on-disk bitmap -- see ext4_mb_release_context() * Other CPUs are prevented from allocating from this pa by lg_mutex |
c9de560de ext4: Add multi b... |
3074 |
*/ |
6ba495e92 ext4: Add configu... |
3075 3076 |
mb_debug(1, "use %u/%u from group pa %p ", pa->pa_lstart-len, len, pa); |
c9de560de ext4: Add multi b... |
3077 3078 3079 |
} /* |
5e745b041 ext4: Fix small f... |
3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 |
* Return the prealloc space that have minimal distance * from the goal block. @cpa is the prealloc * space that is having currently known minimal distance * from the goal block. */ static struct ext4_prealloc_space * ext4_mb_check_group_pa(ext4_fsblk_t goal_block, struct ext4_prealloc_space *pa, struct ext4_prealloc_space *cpa) { ext4_fsblk_t cur_distance, new_distance; if (cpa == NULL) { atomic_inc(&pa->pa_count); return pa; } cur_distance = abs(goal_block - cpa->pa_pstart); new_distance = abs(goal_block - pa->pa_pstart); if (cur_distance < new_distance) return cpa; /* drop the previous reference */ atomic_dec(&cpa->pa_count); atomic_inc(&pa->pa_count); return pa; } /* |
c9de560de ext4: Add multi b... |
3109 3110 |
* search goal blocks in preallocated space */ |
4ddfef7b4 ext4: reduce mbal... |
3111 3112 |
static noinline_for_stack int ext4_mb_use_preallocated(struct ext4_allocation_context *ac) |
c9de560de ext4: Add multi b... |
3113 |
{ |
6be2ded1d ext4: Don't allow... |
3114 |
int order, i; |
c9de560de ext4: Add multi b... |
3115 3116 |
struct ext4_inode_info *ei = EXT4_I(ac->ac_inode); struct ext4_locality_group *lg; |
5e745b041 ext4: Fix small f... |
3117 3118 |
struct ext4_prealloc_space *pa, *cpa = NULL; ext4_fsblk_t goal_block; |
c9de560de ext4: Add multi b... |
3119 3120 3121 3122 3123 3124 3125 |
/* only data can be preallocated */ if (!(ac->ac_flags & EXT4_MB_HINT_DATA)) return 0; /* first, try per-file preallocation */ rcu_read_lock(); |
9a0762c5a ext4: Convert li... |
3126 |
list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) { |
c9de560de ext4: Add multi b... |
3127 3128 3129 3130 3131 3132 |
/* all fields in this condition don't change, * so we can skip locking for them */ if (ac->ac_o_ex.fe_logical < pa->pa_lstart || ac->ac_o_ex.fe_logical >= pa->pa_lstart + pa->pa_len) continue; |
fb0a387dc ext4: limit block... |
3133 |
/* non-extent files can't have physical blocks past 2^32 */ |
12e9b8920 ext4: Use bitops ... |
3134 |
if (!(ext4_test_inode_flag(ac->ac_inode, EXT4_INODE_EXTENTS)) && |
fb0a387dc ext4: limit block... |
3135 3136 |
pa->pa_pstart + pa->pa_len > EXT4_MAX_BLOCK_FILE_PHYS) continue; |
c9de560de ext4: Add multi b... |
3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 |
/* found preallocated blocks, use them */ spin_lock(&pa->pa_lock); if (pa->pa_deleted == 0 && pa->pa_free) { atomic_inc(&pa->pa_count); ext4_mb_use_inode_pa(ac, pa); spin_unlock(&pa->pa_lock); ac->ac_criteria = 10; rcu_read_unlock(); return 1; } spin_unlock(&pa->pa_lock); } rcu_read_unlock(); /* can we use group allocation? */ if (!(ac->ac_flags & EXT4_MB_HINT_GROUP_ALLOC)) return 0; /* inode may have no locality group for some reason */ lg = ac->ac_lg; if (lg == NULL) return 0; |
6be2ded1d ext4: Don't allow... |
3159 3160 3161 3162 |
order = fls(ac->ac_o_ex.fe_len) - 1; if (order > PREALLOC_TB_SIZE - 1) /* The max size of hash table is PREALLOC_TB_SIZE */ order = PREALLOC_TB_SIZE - 1; |
bda00de7e ext4: cleanup to ... |
3163 |
goal_block = ext4_grp_offs_to_block(ac->ac_sb, &ac->ac_g_ex); |
5e745b041 ext4: Fix small f... |
3164 3165 3166 3167 |
/* * search for the prealloc space that is having * minimal distance from the goal block. */ |
6be2ded1d ext4: Don't allow... |
3168 3169 3170 3171 3172 3173 3174 |
for (i = order; i < PREALLOC_TB_SIZE; i++) { rcu_read_lock(); list_for_each_entry_rcu(pa, &lg->lg_prealloc_list[i], pa_inode_list) { spin_lock(&pa->pa_lock); if (pa->pa_deleted == 0 && pa->pa_free >= ac->ac_o_ex.fe_len) { |
5e745b041 ext4: Fix small f... |
3175 3176 3177 |
cpa = ext4_mb_check_group_pa(goal_block, pa, cpa); |
6be2ded1d ext4: Don't allow... |
3178 |
} |
c9de560de ext4: Add multi b... |
3179 |
spin_unlock(&pa->pa_lock); |
c9de560de ext4: Add multi b... |
3180 |
} |
6be2ded1d ext4: Don't allow... |
3181 |
rcu_read_unlock(); |
c9de560de ext4: Add multi b... |
3182 |
} |
5e745b041 ext4: Fix small f... |
3183 3184 3185 3186 3187 |
if (cpa) { ext4_mb_use_group_pa(ac, cpa); ac->ac_criteria = 20; return 1; } |
c9de560de ext4: Add multi b... |
3188 3189 3190 3191 |
return 0; } /* |
7a2fcbf7f ext4: don't use b... |
3192 3193 3194 |
* the function goes through all block freed in the group * but not yet committed and marks them used in in-core bitmap. * buddy must be generated from this bitmap |
955ce5f5b ext4: Convert ext... |
3195 |
* Need to be called with the ext4 group lock held |
7a2fcbf7f ext4: don't use b... |
3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 |
*/ static void ext4_mb_generate_from_freelist(struct super_block *sb, void *bitmap, ext4_group_t group) { struct rb_node *n; struct ext4_group_info *grp; struct ext4_free_data *entry; grp = ext4_get_group_info(sb, group); n = rb_first(&(grp->bb_free_root)); while (n) { entry = rb_entry(n, struct ext4_free_data, node); |
955ce5f5b ext4: Convert ext... |
3209 |
mb_set_bits(bitmap, entry->start_blk, entry->count); |
7a2fcbf7f ext4: don't use b... |
3210 3211 3212 3213 3214 3215 |
n = rb_next(n); } return; } /* |
c9de560de ext4: Add multi b... |
3216 3217 |
* the function goes through all preallocation in this group and marks them * used in in-core bitmap. buddy must be generated from this bitmap |
955ce5f5b ext4: Convert ext... |
3218 |
* Need to be called with ext4 group lock held |
c9de560de ext4: Add multi b... |
3219 |
*/ |
089ceecc1 ext4: mark severa... |
3220 3221 |
static noinline_for_stack void ext4_mb_generate_from_pa(struct super_block *sb, void *bitmap, |
c9de560de ext4: Add multi b... |
3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 |
ext4_group_t group) { struct ext4_group_info *grp = ext4_get_group_info(sb, group); struct ext4_prealloc_space *pa; struct list_head *cur; ext4_group_t groupnr; ext4_grpblk_t start; int preallocated = 0; int count = 0; int len; /* all form of preallocation discards first load group, * so the only competing code is preallocation use. * we don't need any locking here * notice we do NOT ignore preallocations with pa_deleted * otherwise we could leave used blocks available for * allocation in buddy when concurrent ext4_mb_put_pa() * is dropping preallocation */ list_for_each(cur, &grp->bb_prealloc_list) { pa = list_entry(cur, struct ext4_prealloc_space, pa_group_list); spin_lock(&pa->pa_lock); ext4_get_group_no_and_offset(sb, pa->pa_pstart, &groupnr, &start); len = pa->pa_len; spin_unlock(&pa->pa_lock); if (unlikely(len == 0)) continue; BUG_ON(groupnr != group); |
955ce5f5b ext4: Convert ext... |
3251 |
mb_set_bits(bitmap, start, len); |
c9de560de ext4: Add multi b... |
3252 3253 3254 |
preallocated += len; count++; } |
6ba495e92 ext4: Add configu... |
3255 3256 |
mb_debug(1, "prellocated %u for group %u ", preallocated, group); |
c9de560de ext4: Add multi b... |
3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 |
} static void ext4_mb_pa_callback(struct rcu_head *head) { struct ext4_prealloc_space *pa; pa = container_of(head, struct ext4_prealloc_space, u.pa_rcu); kmem_cache_free(ext4_pspace_cachep, pa); } /* * drops a reference to preallocated space descriptor * if this was the last reference and the space is consumed */ static void ext4_mb_put_pa(struct ext4_allocation_context *ac, struct super_block *sb, struct ext4_prealloc_space *pa) { |
a9df9a491 ext4: Make ext4_g... |
3273 |
ext4_group_t grp; |
d33a1976f ext4: fix bb_prea... |
3274 |
ext4_fsblk_t grp_blk; |
c9de560de ext4: Add multi b... |
3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 |
if (!atomic_dec_and_test(&pa->pa_count) || pa->pa_free != 0) return; /* in this short window concurrent discard can set pa_deleted */ spin_lock(&pa->pa_lock); if (pa->pa_deleted == 1) { spin_unlock(&pa->pa_lock); return; } pa->pa_deleted = 1; spin_unlock(&pa->pa_lock); |
d33a1976f ext4: fix bb_prea... |
3288 |
grp_blk = pa->pa_pstart; |
60e6679e2 ext4: Drop whites... |
3289 |
/* |
cc0fb9ad7 ext4: Rename pa_l... |
3290 3291 3292 3293 |
* If doing group-based preallocation, pa_pstart may be in the * next group when pa is used up */ if (pa->pa_type == MB_GROUP_PA) |
d33a1976f ext4: fix bb_prea... |
3294 3295 3296 |
grp_blk--; ext4_get_group_no_and_offset(sb, grp_blk, &grp, NULL); |
c9de560de ext4: Add multi b... |
3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 |
/* * possible race: * * P1 (buddy init) P2 (regular allocation) * find block B in PA * copy on-disk bitmap to buddy * mark B in on-disk bitmap * drop PA from group * mark all PAs in buddy * * thus, P1 initializes buddy with B available. to prevent this * we make "copy" and "mark all PAs" atomic and serialize "drop PA" * against that pair */ ext4_lock_group(sb, grp); list_del(&pa->pa_group_list); ext4_unlock_group(sb, grp); spin_lock(pa->pa_obj_lock); list_del_rcu(&pa->pa_inode_list); spin_unlock(pa->pa_obj_lock); call_rcu(&(pa)->u.pa_rcu, ext4_mb_pa_callback); } /* * creates new preallocated space for given inode */ |
4ddfef7b4 ext4: reduce mbal... |
3326 3327 |
static noinline_for_stack int ext4_mb_new_inode_pa(struct ext4_allocation_context *ac) |
c9de560de ext4: Add multi b... |
3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 |
{ struct super_block *sb = ac->ac_sb; struct ext4_prealloc_space *pa; struct ext4_group_info *grp; struct ext4_inode_info *ei; /* preallocate only when found space is larger then requested */ BUG_ON(ac->ac_o_ex.fe_len >= ac->ac_b_ex.fe_len); BUG_ON(ac->ac_status != AC_STATUS_FOUND); BUG_ON(!S_ISREG(ac->ac_inode->i_mode)); pa = kmem_cache_alloc(ext4_pspace_cachep, GFP_NOFS); if (pa == NULL) return -ENOMEM; if (ac->ac_b_ex.fe_len < ac->ac_g_ex.fe_len) { int winl; int wins; int win; int offs; /* we can't allocate as much as normalizer wants. * so, found space must get proper lstart * to cover original request */ BUG_ON(ac->ac_g_ex.fe_logical > ac->ac_o_ex.fe_logical); BUG_ON(ac->ac_g_ex.fe_len < ac->ac_o_ex.fe_len); /* we're limited by original request in that * logical block must be covered any way * winl is window we can move our chunk within */ winl = ac->ac_o_ex.fe_logical - ac->ac_g_ex.fe_logical; /* also, we should cover whole original request */ wins = ac->ac_b_ex.fe_len - ac->ac_o_ex.fe_len; /* the smallest one defines real window */ win = min(winl, wins); offs = ac->ac_o_ex.fe_logical % ac->ac_b_ex.fe_len; if (offs && offs < win) win = offs; ac->ac_b_ex.fe_logical = ac->ac_o_ex.fe_logical - win; BUG_ON(ac->ac_o_ex.fe_logical < ac->ac_b_ex.fe_logical); BUG_ON(ac->ac_o_ex.fe_len > ac->ac_b_ex.fe_len); } /* preallocation can change ac_b_ex, thus we store actually * allocated blocks for history */ ac->ac_f_ex = ac->ac_b_ex; pa->pa_lstart = ac->ac_b_ex.fe_logical; pa->pa_pstart = ext4_grp_offs_to_block(sb, &ac->ac_b_ex); pa->pa_len = ac->ac_b_ex.fe_len; pa->pa_free = pa->pa_len; atomic_set(&pa->pa_count, 1); spin_lock_init(&pa->pa_lock); |
d794bf8e0 ext4: Initialize ... |
3385 3386 |
INIT_LIST_HEAD(&pa->pa_inode_list); INIT_LIST_HEAD(&pa->pa_group_list); |
c9de560de ext4: Add multi b... |
3387 |
pa->pa_deleted = 0; |
cc0fb9ad7 ext4: Rename pa_l... |
3388 |
pa->pa_type = MB_INODE_PA; |
c9de560de ext4: Add multi b... |
3389 |
|
6ba495e92 ext4: Add configu... |
3390 3391 |
mb_debug(1, "new inode pa %p: %llu/%u for %u ", pa, |
c9de560de ext4: Add multi b... |
3392 |
pa->pa_pstart, pa->pa_len, pa->pa_lstart); |
9bffad1ed ext4: convert ins... |
3393 |
trace_ext4_mb_new_inode_pa(ac, pa); |
c9de560de ext4: Add multi b... |
3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 |
ext4_mb_use_inode_pa(ac, pa); atomic_add(pa->pa_free, &EXT4_SB(sb)->s_mb_preallocated); ei = EXT4_I(ac->ac_inode); grp = ext4_get_group_info(sb, ac->ac_b_ex.fe_group); pa->pa_obj_lock = &ei->i_prealloc_lock; pa->pa_inode = ac->ac_inode; ext4_lock_group(sb, ac->ac_b_ex.fe_group); list_add(&pa->pa_group_list, &grp->bb_prealloc_list); ext4_unlock_group(sb, ac->ac_b_ex.fe_group); spin_lock(pa->pa_obj_lock); list_add_rcu(&pa->pa_inode_list, &ei->i_prealloc_list); spin_unlock(pa->pa_obj_lock); return 0; } /* * creates new preallocated space for locality group inodes belongs to */ |
4ddfef7b4 ext4: reduce mbal... |
3418 3419 |
static noinline_for_stack int ext4_mb_new_group_pa(struct ext4_allocation_context *ac) |
c9de560de ext4: Add multi b... |
3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 |
{ struct super_block *sb = ac->ac_sb; struct ext4_locality_group *lg; struct ext4_prealloc_space *pa; struct ext4_group_info *grp; /* preallocate only when found space is larger then requested */ BUG_ON(ac->ac_o_ex.fe_len >= ac->ac_b_ex.fe_len); BUG_ON(ac->ac_status != AC_STATUS_FOUND); BUG_ON(!S_ISREG(ac->ac_inode->i_mode)); BUG_ON(ext4_pspace_cachep == NULL); pa = kmem_cache_alloc(ext4_pspace_cachep, GFP_NOFS); if (pa == NULL) return -ENOMEM; /* preallocation can change ac_b_ex, thus we store actually * allocated blocks for history */ ac->ac_f_ex = ac->ac_b_ex; pa->pa_pstart = ext4_grp_offs_to_block(sb, &ac->ac_b_ex); pa->pa_lstart = pa->pa_pstart; pa->pa_len = ac->ac_b_ex.fe_len; pa->pa_free = pa->pa_len; atomic_set(&pa->pa_count, 1); spin_lock_init(&pa->pa_lock); |
6be2ded1d ext4: Don't allow... |
3446 |
INIT_LIST_HEAD(&pa->pa_inode_list); |
d794bf8e0 ext4: Initialize ... |
3447 |
INIT_LIST_HEAD(&pa->pa_group_list); |
c9de560de ext4: Add multi b... |
3448 |
pa->pa_deleted = 0; |
cc0fb9ad7 ext4: Rename pa_l... |
3449 |
pa->pa_type = MB_GROUP_PA; |
c9de560de ext4: Add multi b... |
3450 |
|
6ba495e92 ext4: Add configu... |
3451 3452 |
mb_debug(1, "new group pa %p: %llu/%u for %u ", pa, |
9bffad1ed ext4: convert ins... |
3453 3454 |
pa->pa_pstart, pa->pa_len, pa->pa_lstart); trace_ext4_mb_new_group_pa(ac, pa); |
c9de560de ext4: Add multi b... |
3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 |
ext4_mb_use_group_pa(ac, pa); atomic_add(pa->pa_free, &EXT4_SB(sb)->s_mb_preallocated); grp = ext4_get_group_info(sb, ac->ac_b_ex.fe_group); lg = ac->ac_lg; BUG_ON(lg == NULL); pa->pa_obj_lock = &lg->lg_prealloc_lock; pa->pa_inode = NULL; ext4_lock_group(sb, ac->ac_b_ex.fe_group); list_add(&pa->pa_group_list, &grp->bb_prealloc_list); ext4_unlock_group(sb, ac->ac_b_ex.fe_group); |
6be2ded1d ext4: Don't allow... |
3469 3470 3471 3472 |
/* * We will later add the new pa to the right bucket * after updating the pa_free in ext4_mb_release_context */ |
c9de560de ext4: Add multi b... |
3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 |
return 0; } static int ext4_mb_new_preallocation(struct ext4_allocation_context *ac) { int err; if (ac->ac_flags & EXT4_MB_HINT_GROUP_ALLOC) err = ext4_mb_new_group_pa(ac); else err = ext4_mb_new_inode_pa(ac); return err; } /* * finds all unused blocks in on-disk bitmap, frees them in * in-core bitmap and buddy. * @pa must be unlinked from inode and group lists, so that * nobody else can find/use it. * the caller MUST hold group/inode locks. * TODO: optimize the case when there are no in-core structures yet */ |
4ddfef7b4 ext4: reduce mbal... |
3495 3496 |
static noinline_for_stack int ext4_mb_release_inode_pa(struct ext4_buddy *e4b, struct buffer_head *bitmap_bh, |
c83617db7 ext4: Don't do GF... |
3497 3498 |
struct ext4_prealloc_space *pa, struct ext4_allocation_context *ac) |
c9de560de ext4: Add multi b... |
3499 |
{ |
c9de560de ext4: Add multi b... |
3500 3501 |
struct super_block *sb = e4b->bd_sb; struct ext4_sb_info *sbi = EXT4_SB(sb); |
498e5f241 ext4: Change unsi... |
3502 3503 |
unsigned int end; unsigned int next; |
c9de560de ext4: Add multi b... |
3504 3505 |
ext4_group_t group; ext4_grpblk_t bit; |
ba80b1019 ext4: Add markers... |
3506 |
unsigned long long grp_blk_start; |
c9de560de ext4: Add multi b... |
3507 3508 3509 3510 3511 |
int err = 0; int free = 0; BUG_ON(pa->pa_deleted == 0); ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, &bit); |
ba80b1019 ext4: Add markers... |
3512 |
grp_blk_start = pa->pa_pstart - bit; |
c9de560de ext4: Add multi b... |
3513 3514 |
BUG_ON(group != e4b->bd_group && pa->pa_len != 0); end = bit + pa->pa_len; |
256bdb497 ext4: allocate st... |
3515 3516 3517 |
if (ac) { ac->ac_sb = sb; ac->ac_inode = pa->pa_inode; |
256bdb497 ext4: allocate st... |
3518 |
} |
c9de560de ext4: Add multi b... |
3519 3520 |
while (bit < end) { |
ffad0a44b ext4: ext4_find_n... |
3521 |
bit = mb_find_next_zero_bit(bitmap_bh->b_data, end, bit); |
c9de560de ext4: Add multi b... |
3522 3523 |
if (bit >= end) break; |
ffad0a44b ext4: ext4_find_n... |
3524 |
next = mb_find_next_bit(bitmap_bh->b_data, end, bit); |
6ba495e92 ext4: Add configu... |
3525 3526 |
mb_debug(1, " free preallocated %u/%u in group %u ", |
5a0790c2c ext4: remove init... |
3527 3528 |
(unsigned) ext4_group_first_block_no(sb, group) + bit, (unsigned) next - bit, (unsigned) group); |
c9de560de ext4: Add multi b... |
3529 |
free += next - bit; |
256bdb497 ext4: allocate st... |
3530 3531 3532 3533 3534 |
if (ac) { ac->ac_b_ex.fe_group = group; ac->ac_b_ex.fe_start = bit; ac->ac_b_ex.fe_len = next - bit; ac->ac_b_ex.fe_logical = 0; |
296c355cd ext4: Use tracepo... |
3535 |
trace_ext4_mballoc_discard(ac); |
256bdb497 ext4: allocate st... |
3536 |
} |
c9de560de ext4: Add multi b... |
3537 |
|
e5880d76a ext4: fix potenti... |
3538 |
trace_ext4_mb_release_inode_pa(sb, ac, pa, grp_blk_start + bit, |
9bffad1ed ext4: convert ins... |
3539 |
next - bit); |
c9de560de ext4: Add multi b... |
3540 3541 3542 3543 |
mb_free_blocks(pa->pa_inode, e4b, bit, next - bit); bit = next + 1; } if (free != pa->pa_free) { |
26346ff68 ext4: Don't panic... |
3544 3545 |
printk(KERN_CRIT "pa %p: logic %lu, phys. %lu, len %lu ", |
c9de560de ext4: Add multi b... |
3546 3547 3548 |
pa, (unsigned long) pa->pa_lstart, (unsigned long) pa->pa_pstart, (unsigned long) pa->pa_len); |
e29136f80 ext4: Enhance ext... |
3549 |
ext4_grp_locked_error(sb, group, 0, 0, "free %u, pa_free %u", |
5d1b1b3f4 ext4: fix BUG whe... |
3550 |
free, pa->pa_free); |
e56eb6590 ext4: Don't claim... |
3551 3552 3553 3554 |
/* * pa is already deleted so we use the value obtained * from the bitmap and continue. */ |
c9de560de ext4: Add multi b... |
3555 |
} |
c9de560de ext4: Add multi b... |
3556 3557 3558 3559 |
atomic_add(free, &sbi->s_mb_discarded); return err; } |
4ddfef7b4 ext4: reduce mbal... |
3560 3561 |
static noinline_for_stack int ext4_mb_release_group_pa(struct ext4_buddy *e4b, |
c83617db7 ext4: Don't do GF... |
3562 3563 |
struct ext4_prealloc_space *pa, struct ext4_allocation_context *ac) |
c9de560de ext4: Add multi b... |
3564 |
{ |
c9de560de ext4: Add multi b... |
3565 3566 3567 |
struct super_block *sb = e4b->bd_sb; ext4_group_t group; ext4_grpblk_t bit; |
e5880d76a ext4: fix potenti... |
3568 |
trace_ext4_mb_release_group_pa(sb, ac, pa); |
c9de560de ext4: Add multi b... |
3569 3570 3571 3572 3573 |
BUG_ON(pa->pa_deleted == 0); ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, &bit); BUG_ON(group != e4b->bd_group && pa->pa_len != 0); mb_free_blocks(pa->pa_inode, e4b, bit, pa->pa_len); atomic_add(pa->pa_len, &EXT4_SB(sb)->s_mb_discarded); |
256bdb497 ext4: allocate st... |
3574 3575 3576 3577 3578 3579 3580 |
if (ac) { ac->ac_sb = sb; ac->ac_inode = NULL; ac->ac_b_ex.fe_group = group; ac->ac_b_ex.fe_start = bit; ac->ac_b_ex.fe_len = pa->pa_len; ac->ac_b_ex.fe_logical = 0; |
296c355cd ext4: Use tracepo... |
3581 |
trace_ext4_mballoc_discard(ac); |
256bdb497 ext4: allocate st... |
3582 |
} |
c9de560de ext4: Add multi b... |
3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 |
return 0; } /* * releases all preallocations in given group * * first, we need to decide discard policy: * - when do we discard * 1) ENOSPC * - how many do we discard * 1) how many requested */ |
4ddfef7b4 ext4: reduce mbal... |
3596 3597 |
static noinline_for_stack int ext4_mb_discard_group_preallocations(struct super_block *sb, |
c9de560de ext4: Add multi b... |
3598 3599 3600 3601 3602 |
ext4_group_t group, int needed) { struct ext4_group_info *grp = ext4_get_group_info(sb, group); struct buffer_head *bitmap_bh = NULL; struct ext4_prealloc_space *pa, *tmp; |
c83617db7 ext4: Don't do GF... |
3603 |
struct ext4_allocation_context *ac; |
c9de560de ext4: Add multi b... |
3604 3605 3606 3607 3608 |
struct list_head list; struct ext4_buddy e4b; int err; int busy = 0; int free = 0; |
6ba495e92 ext4: Add configu... |
3609 3610 |
mb_debug(1, "discard preallocation for group %u ", group); |
c9de560de ext4: Add multi b... |
3611 3612 3613 |
if (list_empty(&grp->bb_prealloc_list)) return 0; |
574ca174c ext4: Rename read... |
3614 |
bitmap_bh = ext4_read_block_bitmap(sb, group); |
c9de560de ext4: Add multi b... |
3615 |
if (bitmap_bh == NULL) { |
12062dddd ext4: move __func... |
3616 |
ext4_error(sb, "Error reading block bitmap for %u", group); |
ce89f46cb ext4: Improve err... |
3617 |
return 0; |
c9de560de ext4: Add multi b... |
3618 3619 3620 |
} err = ext4_mb_load_buddy(sb, group, &e4b); |
ce89f46cb ext4: Improve err... |
3621 |
if (err) { |
12062dddd ext4: move __func... |
3622 |
ext4_error(sb, "Error loading buddy information for %u", group); |
ce89f46cb ext4: Improve err... |
3623 3624 3625 |
put_bh(bitmap_bh); return 0; } |
c9de560de ext4: Add multi b... |
3626 3627 3628 |
if (needed == 0) needed = EXT4_BLOCKS_PER_GROUP(sb) + 1; |
c9de560de ext4: Add multi b... |
3629 |
INIT_LIST_HEAD(&list); |
c83617db7 ext4: Don't do GF... |
3630 |
ac = kmem_cache_alloc(ext4_ac_cachep, GFP_NOFS); |
9bffad1ed ext4: convert ins... |
3631 3632 |
if (ac) ac->ac_sb = sb; |
c9de560de ext4: Add multi b... |
3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 |
repeat: ext4_lock_group(sb, group); list_for_each_entry_safe(pa, tmp, &grp->bb_prealloc_list, pa_group_list) { spin_lock(&pa->pa_lock); if (atomic_read(&pa->pa_count)) { spin_unlock(&pa->pa_lock); busy = 1; continue; } if (pa->pa_deleted) { spin_unlock(&pa->pa_lock); continue; } /* seems this one can be freed ... */ pa->pa_deleted = 1; /* we can trust pa_free ... */ free += pa->pa_free; spin_unlock(&pa->pa_lock); list_del(&pa->pa_group_list); list_add(&pa->u.pa_tmp_list, &list); } /* if we still need more blocks and some PAs were used, try again */ if (free < needed && busy) { busy = 0; ext4_unlock_group(sb, group); /* * Yield the CPU here so that we don't get soft lockup * in non preempt case. */ yield(); goto repeat; } /* found anything to free? */ if (list_empty(&list)) { BUG_ON(free != 0); goto out; } /* now free all selected PAs */ list_for_each_entry_safe(pa, tmp, &list, u.pa_tmp_list) { /* remove from object (inode or locality group) */ spin_lock(pa->pa_obj_lock); list_del_rcu(&pa->pa_inode_list); spin_unlock(pa->pa_obj_lock); |
cc0fb9ad7 ext4: Rename pa_l... |
3685 |
if (pa->pa_type == MB_GROUP_PA) |
c83617db7 ext4: Don't do GF... |
3686 |
ext4_mb_release_group_pa(&e4b, pa, ac); |
c9de560de ext4: Add multi b... |
3687 |
else |
c83617db7 ext4: Don't do GF... |
3688 |
ext4_mb_release_inode_pa(&e4b, bitmap_bh, pa, ac); |
c9de560de ext4: Add multi b... |
3689 3690 3691 3692 3693 3694 3695 |
list_del(&pa->u.pa_tmp_list); call_rcu(&(pa)->u.pa_rcu, ext4_mb_pa_callback); } out: ext4_unlock_group(sb, group); |
c83617db7 ext4: Don't do GF... |
3696 3697 |
if (ac) kmem_cache_free(ext4_ac_cachep, ac); |
e39e07fdf ext4: rename ext4... |
3698 |
ext4_mb_unload_buddy(&e4b); |
c9de560de ext4: Add multi b... |
3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 |
put_bh(bitmap_bh); return free; } /* * releases all non-used preallocated blocks for given inode * * It's important to discard preallocations under i_data_sem * We don't want another block to be served from the prealloc * space when we are discarding the inode prealloc space. * * FIXME!! Make sure it is valid at all the call sites */ |
c2ea3fde6 ext4: Remove old ... |
3712 |
void ext4_discard_preallocations(struct inode *inode) |
c9de560de ext4: Add multi b... |
3713 3714 3715 3716 3717 |
{ struct ext4_inode_info *ei = EXT4_I(inode); struct super_block *sb = inode->i_sb; struct buffer_head *bitmap_bh = NULL; struct ext4_prealloc_space *pa, *tmp; |
c83617db7 ext4: Don't do GF... |
3718 |
struct ext4_allocation_context *ac; |
c9de560de ext4: Add multi b... |
3719 3720 3721 3722 |
ext4_group_t group = 0; struct list_head list; struct ext4_buddy e4b; int err; |
c2ea3fde6 ext4: Remove old ... |
3723 |
if (!S_ISREG(inode->i_mode)) { |
c9de560de ext4: Add multi b... |
3724 3725 3726 |
/*BUG_ON(!list_empty(&ei->i_prealloc_list));*/ return; } |
6ba495e92 ext4: Add configu... |
3727 3728 |
mb_debug(1, "discard preallocation for inode %lu ", inode->i_ino); |
9bffad1ed ext4: convert ins... |
3729 |
trace_ext4_discard_preallocations(inode); |
c9de560de ext4: Add multi b... |
3730 3731 |
INIT_LIST_HEAD(&list); |
c83617db7 ext4: Don't do GF... |
3732 |
ac = kmem_cache_alloc(ext4_ac_cachep, GFP_NOFS); |
9bffad1ed ext4: convert ins... |
3733 3734 3735 3736 |
if (ac) { ac->ac_sb = sb; ac->ac_inode = inode; } |
c9de560de ext4: Add multi b... |
3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 |
repeat: /* first, collect all pa's in the inode */ spin_lock(&ei->i_prealloc_lock); while (!list_empty(&ei->i_prealloc_list)) { pa = list_entry(ei->i_prealloc_list.next, struct ext4_prealloc_space, pa_inode_list); BUG_ON(pa->pa_obj_lock != &ei->i_prealloc_lock); spin_lock(&pa->pa_lock); if (atomic_read(&pa->pa_count)) { /* this shouldn't happen often - nobody should * use preallocation while we're discarding it */ spin_unlock(&pa->pa_lock); spin_unlock(&ei->i_prealloc_lock); printk(KERN_ERR "uh-oh! used pa while discarding "); WARN_ON(1); schedule_timeout_uninterruptible(HZ); goto repeat; } if (pa->pa_deleted == 0) { pa->pa_deleted = 1; spin_unlock(&pa->pa_lock); list_del_rcu(&pa->pa_inode_list); list_add(&pa->u.pa_tmp_list, &list); continue; } /* someone is deleting pa right now */ spin_unlock(&pa->pa_lock); spin_unlock(&ei->i_prealloc_lock); /* we have to wait here because pa_deleted * doesn't mean pa is already unlinked from * the list. as we might be called from * ->clear_inode() the inode will get freed * and concurrent thread which is unlinking * pa from inode's list may access already * freed memory, bad-bad-bad */ /* XXX: if this happens too often, we can * add a flag to force wait only in case * of ->clear_inode(), but not in case of * regular truncate */ schedule_timeout_uninterruptible(HZ); goto repeat; } spin_unlock(&ei->i_prealloc_lock); list_for_each_entry_safe(pa, tmp, &list, u.pa_tmp_list) { |
cc0fb9ad7 ext4: Rename pa_l... |
3787 |
BUG_ON(pa->pa_type != MB_INODE_PA); |
c9de560de ext4: Add multi b... |
3788 3789 3790 |
ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, NULL); err = ext4_mb_load_buddy(sb, group, &e4b); |
ce89f46cb ext4: Improve err... |
3791 |
if (err) { |
12062dddd ext4: move __func... |
3792 3793 |
ext4_error(sb, "Error loading buddy information for %u", group); |
ce89f46cb ext4: Improve err... |
3794 3795 |
continue; } |
c9de560de ext4: Add multi b... |
3796 |
|
574ca174c ext4: Rename read... |
3797 |
bitmap_bh = ext4_read_block_bitmap(sb, group); |
c9de560de ext4: Add multi b... |
3798 |
if (bitmap_bh == NULL) { |
12062dddd ext4: move __func... |
3799 3800 |
ext4_error(sb, "Error reading block bitmap for %u", group); |
e39e07fdf ext4: rename ext4... |
3801 |
ext4_mb_unload_buddy(&e4b); |
ce89f46cb ext4: Improve err... |
3802 |
continue; |
c9de560de ext4: Add multi b... |
3803 3804 3805 3806 |
} ext4_lock_group(sb, group); list_del(&pa->pa_group_list); |
c83617db7 ext4: Don't do GF... |
3807 |
ext4_mb_release_inode_pa(&e4b, bitmap_bh, pa, ac); |
c9de560de ext4: Add multi b... |
3808 |
ext4_unlock_group(sb, group); |
e39e07fdf ext4: rename ext4... |
3809 |
ext4_mb_unload_buddy(&e4b); |
c9de560de ext4: Add multi b... |
3810 3811 3812 3813 3814 |
put_bh(bitmap_bh); list_del(&pa->u.pa_tmp_list); call_rcu(&(pa)->u.pa_rcu, ext4_mb_pa_callback); } |
c83617db7 ext4: Don't do GF... |
3815 3816 |
if (ac) kmem_cache_free(ext4_ac_cachep, ac); |
c9de560de ext4: Add multi b... |
3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 |
} /* * finds all preallocated spaces and return blocks being freed to them * if preallocated space becomes full (no block is used from the space) * then the function frees space in buddy * XXX: at the moment, truncate (which is the only way to free blocks) * discards all preallocations */ static void ext4_mb_return_to_preallocation(struct inode *inode, struct ext4_buddy *e4b, sector_t block, int count) { BUG_ON(!list_empty(&EXT4_I(inode)->i_prealloc_list)); } |
6ba495e92 ext4: Add configu... |
3832 |
#ifdef CONFIG_EXT4_DEBUG |
c9de560de ext4: Add multi b... |
3833 3834 3835 |
static void ext4_mb_show_ac(struct ext4_allocation_context *ac) { struct super_block *sb = ac->ac_sb; |
8df9675f8 ext4: Avoid races... |
3836 |
ext4_group_t ngroups, i; |
c9de560de ext4: Add multi b... |
3837 |
|
e3570639c ext4: don't print... |
3838 3839 |
if (EXT4_SB(sb)->s_mount_flags & EXT4_MF_FS_ABORTED) return; |
c9de560de ext4: Add multi b... |
3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 |
printk(KERN_ERR "EXT4-fs: Can't allocate:" " Allocation context details: "); printk(KERN_ERR "EXT4-fs: status %d flags %d ", ac->ac_status, ac->ac_flags); printk(KERN_ERR "EXT4-fs: orig %lu/%lu/%lu@%lu, goal %lu/%lu/%lu@%lu, " "best %lu/%lu/%lu@%lu cr %d ", (unsigned long)ac->ac_o_ex.fe_group, (unsigned long)ac->ac_o_ex.fe_start, (unsigned long)ac->ac_o_ex.fe_len, (unsigned long)ac->ac_o_ex.fe_logical, (unsigned long)ac->ac_g_ex.fe_group, (unsigned long)ac->ac_g_ex.fe_start, (unsigned long)ac->ac_g_ex.fe_len, (unsigned long)ac->ac_g_ex.fe_logical, (unsigned long)ac->ac_b_ex.fe_group, (unsigned long)ac->ac_b_ex.fe_start, (unsigned long)ac->ac_b_ex.fe_len, (unsigned long)ac->ac_b_ex.fe_logical, (int)ac->ac_criteria); printk(KERN_ERR "EXT4-fs: %lu scanned, %d found ", ac->ac_ex_scanned, ac->ac_found); printk(KERN_ERR "EXT4-fs: groups: "); |
8df9675f8 ext4: Avoid races... |
3867 3868 |
ngroups = ext4_get_groups_count(sb); for (i = 0; i < ngroups; i++) { |
c9de560de ext4: Add multi b... |
3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880 |
struct ext4_group_info *grp = ext4_get_group_info(sb, i); struct ext4_prealloc_space *pa; ext4_grpblk_t start; struct list_head *cur; ext4_lock_group(sb, i); list_for_each(cur, &grp->bb_prealloc_list) { pa = list_entry(cur, struct ext4_prealloc_space, pa_group_list); spin_lock(&pa->pa_lock); ext4_get_group_no_and_offset(sb, pa->pa_pstart, NULL, &start); spin_unlock(&pa->pa_lock); |
1c7185051 ext4: Fix compile... |
3881 3882 3883 |
printk(KERN_ERR "PA:%u:%d:%u ", i, start, pa->pa_len); |
c9de560de ext4: Add multi b... |
3884 |
} |
60bd63d19 ext4: cleanup for... |
3885 |
ext4_unlock_group(sb, i); |
c9de560de ext4: Add multi b... |
3886 3887 3888 |
if (grp->bb_free == 0) continue; |
1c7185051 ext4: Fix compile... |
3889 3890 |
printk(KERN_ERR "%u: %d/%d ", |
c9de560de ext4: Add multi b... |
3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 |
i, grp->bb_free, grp->bb_fragments); } printk(KERN_ERR " "); } #else static inline void ext4_mb_show_ac(struct ext4_allocation_context *ac) { return; } #endif /* * We use locality group preallocation for small size file. The size of the * file is determined by the current size or the resulting size after * allocation which ever is larger * |
b713a5ec5 ext4: remove /pro... |
3908 |
* One can tune this size via /sys/fs/ext4/<partition>/mb_stream_req |
c9de560de ext4: Add multi b... |
3909 3910 3911 3912 3913 3914 3915 3916 3917 |
*/ static void ext4_mb_group_or_file(struct ext4_allocation_context *ac) { struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); int bsbits = ac->ac_sb->s_blocksize_bits; loff_t size, isize; if (!(ac->ac_flags & EXT4_MB_HINT_DATA)) return; |
4ba74d00a ext4: Fix bugs in... |
3918 3919 |
if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY)) return; |
c9de560de ext4: Add multi b... |
3920 |
size = ac->ac_o_ex.fe_logical + ac->ac_o_ex.fe_len; |
50797481a ext4: Avoid group... |
3921 3922 |
isize = (i_size_read(ac->ac_inode) + ac->ac_sb->s_blocksize - 1) >> bsbits; |
c9de560de ext4: Add multi b... |
3923 |
|
50797481a ext4: Avoid group... |
3924 3925 3926 3927 3928 3929 |
if ((size == isize) && !ext4_fs_is_busy(sbi) && (atomic_read(&ac->ac_inode->i_writecount) == 0)) { ac->ac_flags |= EXT4_MB_HINT_NOPREALLOC; return; } |
c9de560de ext4: Add multi b... |
3930 |
/* don't use group allocation for large files */ |
717805773 ext4: Fix huerist... |
3931 |
size = max(size, isize); |
cc483f102 ext4: Fix fencepo... |
3932 |
if (size > sbi->s_mb_stream_request) { |
4ba74d00a ext4: Fix bugs in... |
3933 |
ac->ac_flags |= EXT4_MB_STREAM_ALLOC; |
c9de560de ext4: Add multi b... |
3934 |
return; |
4ba74d00a ext4: Fix bugs in... |
3935 |
} |
c9de560de ext4: Add multi b... |
3936 3937 3938 3939 3940 3941 3942 |
BUG_ON(ac->ac_lg != NULL); /* * locality group prealloc space are per cpu. The reason for having * per cpu locality group is to reduce the contention between block * request from multiple CPUs. */ |
ca0c9584b this_cpu: Straigh... |
3943 |
ac->ac_lg = __this_cpu_ptr(sbi->s_locality_groups); |
c9de560de ext4: Add multi b... |
3944 3945 3946 3947 3948 3949 3950 |
/* we're going to use group allocation */ ac->ac_flags |= EXT4_MB_HINT_GROUP_ALLOC; /* serialize all allocations in the group */ mutex_lock(&ac->ac_lg->lg_mutex); } |
4ddfef7b4 ext4: reduce mbal... |
3951 3952 |
static noinline_for_stack int ext4_mb_initialize_context(struct ext4_allocation_context *ac, |
c9de560de ext4: Add multi b... |
3953 3954 3955 3956 3957 3958 |
struct ext4_allocation_request *ar) { struct super_block *sb = ar->inode->i_sb; struct ext4_sb_info *sbi = EXT4_SB(sb); struct ext4_super_block *es = sbi->s_es; ext4_group_t group; |
498e5f241 ext4: Change unsi... |
3959 3960 |
unsigned int len; ext4_fsblk_t goal; |
c9de560de ext4: Add multi b... |
3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 |
ext4_grpblk_t block; /* we can't allocate > group size */ len = ar->len; /* just a dirty hack to filter too big requests */ if (len >= EXT4_BLOCKS_PER_GROUP(sb) - 10) len = EXT4_BLOCKS_PER_GROUP(sb) - 10; /* start searching from the goal */ goal = ar->goal; if (goal < le32_to_cpu(es->s_first_data_block) || goal >= ext4_blocks_count(es)) goal = le32_to_cpu(es->s_first_data_block); ext4_get_group_no_and_offset(sb, goal, &group, &block); /* set up allocation goals */ |
833576b36 ext4: Fix ext4_mb... |
3978 |
memset(ac, 0, sizeof(struct ext4_allocation_context)); |
c9de560de ext4: Add multi b... |
3979 |
ac->ac_b_ex.fe_logical = ar->logical; |
c9de560de ext4: Add multi b... |
3980 |
ac->ac_status = AC_STATUS_CONTINUE; |
c9de560de ext4: Add multi b... |
3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 |
ac->ac_sb = sb; ac->ac_inode = ar->inode; ac->ac_o_ex.fe_logical = ar->logical; ac->ac_o_ex.fe_group = group; ac->ac_o_ex.fe_start = block; ac->ac_o_ex.fe_len = len; ac->ac_g_ex.fe_logical = ar->logical; ac->ac_g_ex.fe_group = group; ac->ac_g_ex.fe_start = block; ac->ac_g_ex.fe_len = len; |
c9de560de ext4: Add multi b... |
3991 |
ac->ac_flags = ar->flags; |
c9de560de ext4: Add multi b... |
3992 3993 3994 3995 |
/* we have to define context: we'll we work with a file or * locality group. this is a policy, actually */ ext4_mb_group_or_file(ac); |
6ba495e92 ext4: Add configu... |
3996 |
mb_debug(1, "init ac: %u blocks @ %u, goal %u, flags %x, 2^%d, " |
c9de560de ext4: Add multi b... |
3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 |
"left: %u/%u, right %u/%u to %swritable ", (unsigned) ar->len, (unsigned) ar->logical, (unsigned) ar->goal, ac->ac_flags, ac->ac_2order, (unsigned) ar->lleft, (unsigned) ar->pleft, (unsigned) ar->lright, (unsigned) ar->pright, atomic_read(&ar->inode->i_writecount) ? "" : "non-"); return 0; } |
6be2ded1d ext4: Don't allow... |
4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 |
static noinline_for_stack void ext4_mb_discard_lg_preallocations(struct super_block *sb, struct ext4_locality_group *lg, int order, int total_entries) { ext4_group_t group = 0; struct ext4_buddy e4b; struct list_head discard_list; struct ext4_prealloc_space *pa, *tmp; struct ext4_allocation_context *ac; |
6ba495e92 ext4: Add configu... |
4017 4018 |
mb_debug(1, "discard locality group preallocation "); |
6be2ded1d ext4: Don't allow... |
4019 4020 4021 |
INIT_LIST_HEAD(&discard_list); ac = kmem_cache_alloc(ext4_ac_cachep, GFP_NOFS); |
9bffad1ed ext4: convert ins... |
4022 4023 |
if (ac) ac->ac_sb = sb; |
6be2ded1d ext4: Don't allow... |
4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 |
spin_lock(&lg->lg_prealloc_lock); list_for_each_entry_rcu(pa, &lg->lg_prealloc_list[order], pa_inode_list) { spin_lock(&pa->pa_lock); if (atomic_read(&pa->pa_count)) { /* * This is the pa that we just used * for block allocation. So don't * free that */ spin_unlock(&pa->pa_lock); continue; } if (pa->pa_deleted) { spin_unlock(&pa->pa_lock); continue; } /* only lg prealloc space */ |
cc0fb9ad7 ext4: Rename pa_l... |
4043 |
BUG_ON(pa->pa_type != MB_GROUP_PA); |
6be2ded1d ext4: Don't allow... |
4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064 4065 4066 4067 4068 |
/* seems this one can be freed ... */ pa->pa_deleted = 1; spin_unlock(&pa->pa_lock); list_del_rcu(&pa->pa_inode_list); list_add(&pa->u.pa_tmp_list, &discard_list); total_entries--; if (total_entries <= 5) { /* * we want to keep only 5 entries * allowing it to grow to 8. This * mak sure we don't call discard * soon for this list. */ break; } } spin_unlock(&lg->lg_prealloc_lock); list_for_each_entry_safe(pa, tmp, &discard_list, u.pa_tmp_list) { ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, NULL); if (ext4_mb_load_buddy(sb, group, &e4b)) { |
12062dddd ext4: move __func... |
4069 4070 |
ext4_error(sb, "Error loading buddy information for %u", group); |
6be2ded1d ext4: Don't allow... |
4071 4072 4073 4074 4075 4076 |
continue; } ext4_lock_group(sb, group); list_del(&pa->pa_group_list); ext4_mb_release_group_pa(&e4b, pa, ac); ext4_unlock_group(sb, group); |
e39e07fdf ext4: rename ext4... |
4077 |
ext4_mb_unload_buddy(&e4b); |
6be2ded1d ext4: Don't allow... |
4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 |
list_del(&pa->u.pa_tmp_list); call_rcu(&(pa)->u.pa_rcu, ext4_mb_pa_callback); } if (ac) kmem_cache_free(ext4_ac_cachep, ac); } /* * We have incremented pa_count. So it cannot be freed at this * point. Also we hold lg_mutex. So no parallel allocation is * possible from this lg. That means pa_free cannot be updated. * * A parallel ext4_mb_discard_group_preallocations is possible. * which can cause the lg_prealloc_list to be updated. */ static void ext4_mb_add_n_trim(struct ext4_allocation_context *ac) { int order, added = 0, lg_prealloc_count = 1; struct super_block *sb = ac->ac_sb; struct ext4_locality_group *lg = ac->ac_lg; struct ext4_prealloc_space *tmp_pa, *pa = ac->ac_pa; order = fls(pa->pa_free) - 1; if (order > PREALLOC_TB_SIZE - 1) /* The max size of hash table is PREALLOC_TB_SIZE */ order = PREALLOC_TB_SIZE - 1; /* Add the prealloc space to lg */ rcu_read_lock(); list_for_each_entry_rcu(tmp_pa, &lg->lg_prealloc_list[order], pa_inode_list) { spin_lock(&tmp_pa->pa_lock); if (tmp_pa->pa_deleted) { |
e7c9e3e99 ext4: fix locking... |
4111 |
spin_unlock(&tmp_pa->pa_lock); |
6be2ded1d ext4: Don't allow... |
4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 |
continue; } if (!added && pa->pa_free < tmp_pa->pa_free) { /* Add to the tail of the previous entry */ list_add_tail_rcu(&pa->pa_inode_list, &tmp_pa->pa_inode_list); added = 1; /* * we want to count the total * number of entries in the list */ } spin_unlock(&tmp_pa->pa_lock); lg_prealloc_count++; } if (!added) list_add_tail_rcu(&pa->pa_inode_list, &lg->lg_prealloc_list[order]); rcu_read_unlock(); /* Now trim the list to be not more than 8 elements */ if (lg_prealloc_count > 8) { ext4_mb_discard_lg_preallocations(sb, lg, order, lg_prealloc_count); return; } return ; } |
c9de560de ext4: Add multi b... |
4140 4141 4142 4143 4144 |
/* * release all resource we used in allocation */ static int ext4_mb_release_context(struct ext4_allocation_context *ac) { |
6be2ded1d ext4: Don't allow... |
4145 4146 |
struct ext4_prealloc_space *pa = ac->ac_pa; if (pa) { |
cc0fb9ad7 ext4: Rename pa_l... |
4147 |
if (pa->pa_type == MB_GROUP_PA) { |
c9de560de ext4: Add multi b... |
4148 |
/* see comment in ext4_mb_use_group_pa() */ |
6be2ded1d ext4: Don't allow... |
4149 4150 4151 4152 4153 4154 |
spin_lock(&pa->pa_lock); pa->pa_pstart += ac->ac_b_ex.fe_len; pa->pa_lstart += ac->ac_b_ex.fe_len; pa->pa_free -= ac->ac_b_ex.fe_len; pa->pa_len -= ac->ac_b_ex.fe_len; spin_unlock(&pa->pa_lock); |
c9de560de ext4: Add multi b... |
4155 |
} |
c9de560de ext4: Add multi b... |
4156 |
} |
8556e8f3b ext4: Don't allow... |
4157 4158 |
if (ac->alloc_semp) up_read(ac->alloc_semp); |
ba4439165 ext4: Fix lockdep... |
4159 4160 4161 4162 4163 4164 4165 4166 |
if (pa) { /* * We want to add the pa to the right bucket. * Remove it from the list and while adding * make sure the list to which we are adding * doesn't grow big. We need to release * alloc_semp before calling ext4_mb_add_n_trim() */ |
cc0fb9ad7 ext4: Rename pa_l... |
4167 |
if ((pa->pa_type == MB_GROUP_PA) && likely(pa->pa_free)) { |
ba4439165 ext4: Fix lockdep... |
4168 4169 4170 4171 4172 4173 4174 |
spin_lock(pa->pa_obj_lock); list_del_rcu(&pa->pa_inode_list); spin_unlock(pa->pa_obj_lock); ext4_mb_add_n_trim(ac); } ext4_mb_put_pa(ac, ac->ac_sb, pa); } |
c9de560de ext4: Add multi b... |
4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 |
if (ac->ac_bitmap_page) page_cache_release(ac->ac_bitmap_page); if (ac->ac_buddy_page) page_cache_release(ac->ac_buddy_page); if (ac->ac_flags & EXT4_MB_HINT_GROUP_ALLOC) mutex_unlock(&ac->ac_lg->lg_mutex); ext4_mb_collect_stats(ac); return 0; } static int ext4_mb_discard_preallocations(struct super_block *sb, int needed) { |
8df9675f8 ext4: Avoid races... |
4187 |
ext4_group_t i, ngroups = ext4_get_groups_count(sb); |
c9de560de ext4: Add multi b... |
4188 4189 |
int ret; int freed = 0; |
9bffad1ed ext4: convert ins... |
4190 |
trace_ext4_mb_discard_preallocations(sb, needed); |
8df9675f8 ext4: Avoid races... |
4191 |
for (i = 0; i < ngroups && needed > 0; i++) { |
c9de560de ext4: Add multi b... |
4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 |
ret = ext4_mb_discard_group_preallocations(sb, i, needed); freed += ret; needed -= ret; } return freed; } /* * Main entry point into mballoc to allocate blocks * it tries to use preallocation first, then falls back * to usual allocation */ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle, |
6c7a120ac ext4: Adding erro... |
4206 |
struct ext4_allocation_request *ar, int *errp) |
c9de560de ext4: Add multi b... |
4207 |
{ |
6bc6e63fc ext4: Add percpu ... |
4208 |
int freed; |
256bdb497 ext4: allocate st... |
4209 |
struct ext4_allocation_context *ac = NULL; |
c9de560de ext4: Add multi b... |
4210 4211 4212 |
struct ext4_sb_info *sbi; struct super_block *sb; ext4_fsblk_t block = 0; |
60e58e0f3 ext4: quota reser... |
4213 |
unsigned int inquota = 0; |
498e5f241 ext4: Change unsi... |
4214 |
unsigned int reserv_blks = 0; |
c9de560de ext4: Add multi b... |
4215 4216 4217 |
sb = ar->inode->i_sb; sbi = EXT4_SB(sb); |
9bffad1ed ext4: convert ins... |
4218 |
trace_ext4_request_blocks(ar); |
ba80b1019 ext4: Add markers... |
4219 |
|
60e58e0f3 ext4: quota reser... |
4220 4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 |
/* * For delayed allocation, we could skip the ENOSPC and * EDQUOT check, as blocks and quotas have been already * reserved when data being copied into pagecache. */ if (EXT4_I(ar->inode)->i_delalloc_reserved_flag) ar->flags |= EXT4_MB_DELALLOC_RESERVED; else { /* Without delayed allocation we need to verify * there is enough free blocks to do block allocation * and verify allocation doesn't exceed the quota limits. |
d2a176379 ext4: delayed all... |
4231 |
*/ |
030ba6bc6 ext4: Retry block... |
4232 4233 4234 4235 4236 4237 |
while (ar->len && ext4_claim_free_blocks(sbi, ar->len)) { /* let others to free the space */ yield(); ar->len = ar->len >> 1; } if (!ar->len) { |
a30d542a0 ext4: Make sure a... |
4238 4239 4240 |
*errp = -ENOSPC; return 0; } |
6bc6e63fc ext4: Add percpu ... |
4241 |
reserv_blks = ar->len; |
5dd4056db dquot: cleanup sp... |
4242 |
while (ar->len && dquot_alloc_block(ar->inode, ar->len)) { |
60e58e0f3 ext4: quota reser... |
4243 4244 4245 4246 4247 4248 |
ar->flags |= EXT4_MB_HINT_NOPREALLOC; ar->len--; } inquota = ar->len; if (ar->len == 0) { *errp = -EDQUOT; |
6c7a120ac ext4: Adding erro... |
4249 |
goto out; |
60e58e0f3 ext4: quota reser... |
4250 |
} |
070314310 ext4: mballoc avo... |
4251 |
} |
d2a176379 ext4: delayed all... |
4252 |
|
256bdb497 ext4: allocate st... |
4253 |
ac = kmem_cache_alloc(ext4_ac_cachep, GFP_NOFS); |
833576b36 ext4: Fix ext4_mb... |
4254 |
if (!ac) { |
363d4251d ext4: remove quot... |
4255 |
ar->len = 0; |
256bdb497 ext4: allocate st... |
4256 |
*errp = -ENOMEM; |
6c7a120ac ext4: Adding erro... |
4257 |
goto out; |
256bdb497 ext4: allocate st... |
4258 |
} |
256bdb497 ext4: allocate st... |
4259 |
*errp = ext4_mb_initialize_context(ac, ar); |
c9de560de ext4: Add multi b... |
4260 4261 |
if (*errp) { ar->len = 0; |
6c7a120ac ext4: Adding erro... |
4262 |
goto out; |
c9de560de ext4: Add multi b... |
4263 |
} |
256bdb497 ext4: allocate st... |
4264 4265 |
ac->ac_op = EXT4_MB_HISTORY_PREALLOC; if (!ext4_mb_use_preallocated(ac)) { |
256bdb497 ext4: allocate st... |
4266 4267 |
ac->ac_op = EXT4_MB_HISTORY_ALLOC; ext4_mb_normalize_request(ac, ar); |
c9de560de ext4: Add multi b... |
4268 4269 |
repeat: /* allocate space in core */ |
6c7a120ac ext4: Adding erro... |
4270 4271 4272 |
*errp = ext4_mb_regular_allocator(ac); if (*errp) goto errout; |
c9de560de ext4: Add multi b... |
4273 4274 4275 4276 |
/* as we've just preallocated more space than * user requested orinally, we store allocated * space in a special descriptor */ |
256bdb497 ext4: allocate st... |
4277 4278 4279 |
if (ac->ac_status == AC_STATUS_FOUND && ac->ac_o_ex.fe_len < ac->ac_b_ex.fe_len) ext4_mb_new_preallocation(ac); |
c9de560de ext4: Add multi b... |
4280 |
} |
256bdb497 ext4: allocate st... |
4281 |
if (likely(ac->ac_status == AC_STATUS_FOUND)) { |
6bc6e63fc ext4: Add percpu ... |
4282 |
*errp = ext4_mb_mark_diskspace_used(ac, handle, reserv_blks); |
6c7a120ac ext4: Adding erro... |
4283 |
if (*errp == -EAGAIN) { |
8556e8f3b ext4: Don't allow... |
4284 4285 4286 4287 4288 |
/* * drop the reference that we took * in ext4_mb_use_best_found */ ext4_mb_release_context(ac); |
519deca04 ext4: Retry block... |
4289 4290 4291 4292 4293 |
ac->ac_b_ex.fe_group = 0; ac->ac_b_ex.fe_start = 0; ac->ac_b_ex.fe_len = 0; ac->ac_status = AC_STATUS_CONTINUE; goto repeat; |
6c7a120ac ext4: Adding erro... |
4294 4295 |
} else if (*errp) errout: |
b844167ed ext4: remove bloc... |
4296 |
ext4_discard_allocated_blocks(ac); |
6c7a120ac ext4: Adding erro... |
4297 |
else { |
519deca04 ext4: Retry block... |
4298 4299 4300 |
block = ext4_grp_offs_to_block(sb, &ac->ac_b_ex); ar->len = ac->ac_b_ex.fe_len; } |
c9de560de ext4: Add multi b... |
4301 |
} else { |
256bdb497 ext4: allocate st... |
4302 |
freed = ext4_mb_discard_preallocations(sb, ac->ac_o_ex.fe_len); |
c9de560de ext4: Add multi b... |
4303 4304 4305 |
if (freed) goto repeat; *errp = -ENOSPC; |
6c7a120ac ext4: Adding erro... |
4306 4307 4308 |
} if (*errp) { |
256bdb497 ext4: allocate st... |
4309 |
ac->ac_b_ex.fe_len = 0; |
c9de560de ext4: Add multi b... |
4310 |
ar->len = 0; |
256bdb497 ext4: allocate st... |
4311 |
ext4_mb_show_ac(ac); |
c9de560de ext4: Add multi b... |
4312 |
} |
256bdb497 ext4: allocate st... |
4313 |
ext4_mb_release_context(ac); |
6c7a120ac ext4: Adding erro... |
4314 4315 4316 |
out: if (ac) kmem_cache_free(ext4_ac_cachep, ac); |
60e58e0f3 ext4: quota reser... |
4317 |
if (inquota && ar->len < inquota) |
5dd4056db dquot: cleanup sp... |
4318 |
dquot_free_block(ar->inode, inquota - ar->len); |
0087d9fb3 ext4: Fix s_dirty... |
4319 4320 4321 4322 4323 4324 |
if (!ar->len) { if (!EXT4_I(ar->inode)->i_delalloc_reserved_flag) /* release all the reserved blocks if non delalloc */ percpu_counter_sub(&sbi->s_dirtyblocks_counter, reserv_blks); } |
c9de560de ext4: Add multi b... |
4325 |
|
9bffad1ed ext4: convert ins... |
4326 |
trace_ext4_allocate_blocks(ar, (unsigned long long)block); |
ba80b1019 ext4: Add markers... |
4327 |
|
c9de560de ext4: Add multi b... |
4328 4329 |
return block; } |
c9de560de ext4: Add multi b... |
4330 |
|
c894058d6 ext4: Use an rbtr... |
4331 4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344 |
/* * We can merge two free data extents only if the physical blocks * are contiguous, AND the extents were freed by the same transaction, * AND the blocks are associated with the same group. */ static int can_merge(struct ext4_free_data *entry1, struct ext4_free_data *entry2) { if ((entry1->t_tid == entry2->t_tid) && (entry1->group == entry2->group) && ((entry1->start_blk + entry1->count) == entry2->start_blk)) return 1; return 0; } |
4ddfef7b4 ext4: reduce mbal... |
4345 4346 |
static noinline_for_stack int ext4_mb_free_metadata(handle_t *handle, struct ext4_buddy *e4b, |
7a2fcbf7f ext4: don't use b... |
4347 |
struct ext4_free_data *new_entry) |
c9de560de ext4: Add multi b... |
4348 |
{ |
e29136f80 ext4: Enhance ext... |
4349 |
ext4_group_t group = e4b->bd_group; |
7a2fcbf7f ext4: don't use b... |
4350 4351 |
ext4_grpblk_t block; struct ext4_free_data *entry; |
c9de560de ext4: Add multi b... |
4352 4353 4354 |
struct ext4_group_info *db = e4b->bd_info; struct super_block *sb = e4b->bd_sb; struct ext4_sb_info *sbi = EXT4_SB(sb); |
c894058d6 ext4: Use an rbtr... |
4355 4356 |
struct rb_node **n = &db->bb_free_root.rb_node, *node; struct rb_node *parent = NULL, *new_node; |
0390131ba ext4: Allow ext4 ... |
4357 |
BUG_ON(!ext4_handle_valid(handle)); |
c9de560de ext4: Add multi b... |
4358 4359 |
BUG_ON(e4b->bd_bitmap_page == NULL); BUG_ON(e4b->bd_buddy_page == NULL); |
c894058d6 ext4: Use an rbtr... |
4360 |
new_node = &new_entry->node; |
7a2fcbf7f ext4: don't use b... |
4361 |
block = new_entry->start_blk; |
c894058d6 ext4: Use an rbtr... |
4362 |
|
c894058d6 ext4: Use an rbtr... |
4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 |
if (!*n) { /* first free block exent. We need to protect buddy cache from being freed, * otherwise we'll refresh it from * on-disk bitmap and lose not-yet-available * blocks */ page_cache_get(e4b->bd_buddy_page); page_cache_get(e4b->bd_bitmap_page); } while (*n) { parent = *n; entry = rb_entry(parent, struct ext4_free_data, node); if (block < entry->start_blk) n = &(*n)->rb_left; else if (block >= (entry->start_blk + entry->count)) n = &(*n)->rb_right; else { |
e29136f80 ext4: Enhance ext... |
4380 4381 4382 |
ext4_grp_locked_error(sb, group, 0, ext4_group_first_block_no(sb, group) + block, "Block already on to-be-freed list"); |
c894058d6 ext4: Use an rbtr... |
4383 |
return 0; |
c9de560de ext4: Add multi b... |
4384 |
} |
c894058d6 ext4: Use an rbtr... |
4385 |
} |
c9de560de ext4: Add multi b... |
4386 |
|
c894058d6 ext4: Use an rbtr... |
4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 |
rb_link_node(new_node, parent, n); rb_insert_color(new_node, &db->bb_free_root); /* Now try to see the extent can be merged to left and right */ node = rb_prev(new_node); if (node) { entry = rb_entry(node, struct ext4_free_data, node); if (can_merge(entry, new_entry)) { new_entry->start_blk = entry->start_blk; new_entry->count += entry->count; rb_erase(node, &(db->bb_free_root)); spin_lock(&sbi->s_md_lock); list_del(&entry->list); spin_unlock(&sbi->s_md_lock); kmem_cache_free(ext4_free_ext_cachep, entry); |
c9de560de ext4: Add multi b... |
4402 |
} |
c894058d6 ext4: Use an rbtr... |
4403 |
} |
c9de560de ext4: Add multi b... |
4404 |
|
c894058d6 ext4: Use an rbtr... |
4405 4406 4407 4408 4409 4410 4411 4412 4413 4414 |
node = rb_next(new_node); if (node) { entry = rb_entry(node, struct ext4_free_data, node); if (can_merge(new_entry, entry)) { new_entry->count += entry->count; rb_erase(node, &(db->bb_free_root)); spin_lock(&sbi->s_md_lock); list_del(&entry->list); spin_unlock(&sbi->s_md_lock); kmem_cache_free(ext4_free_ext_cachep, entry); |
c9de560de ext4: Add multi b... |
4415 4416 |
} } |
3e624fc72 ext4: Replace hac... |
4417 |
/* Add the extent to transaction's private list */ |
c894058d6 ext4: Use an rbtr... |
4418 |
spin_lock(&sbi->s_md_lock); |
3e624fc72 ext4: Replace hac... |
4419 |
list_add(&new_entry->list, &handle->h_transaction->t_private_list); |
c894058d6 ext4: Use an rbtr... |
4420 |
spin_unlock(&sbi->s_md_lock); |
c9de560de ext4: Add multi b... |
4421 4422 |
return 0; } |
443387113 ext4: fold ext4_f... |
4423 4424 4425 4426 4427 4428 4429 |
/** * ext4_free_blocks() -- Free given blocks and update quota * @handle: handle for this transaction * @inode: inode * @block: start physical block to free * @count: number of blocks to count * @metadata: Are these metadata blocks |
c9de560de ext4: Add multi b... |
4430 |
*/ |
443387113 ext4: fold ext4_f... |
4431 |
void ext4_free_blocks(handle_t *handle, struct inode *inode, |
e6362609b ext4: call ext4_f... |
4432 4433 |
struct buffer_head *bh, ext4_fsblk_t block, unsigned long count, int flags) |
c9de560de ext4: Add multi b... |
4434 |
{ |
26346ff68 ext4: Don't panic... |
4435 |
struct buffer_head *bitmap_bh = NULL; |
c9de560de ext4: Add multi b... |
4436 |
struct super_block *sb = inode->i_sb; |
256bdb497 ext4: allocate st... |
4437 |
struct ext4_allocation_context *ac = NULL; |
c9de560de ext4: Add multi b... |
4438 |
struct ext4_group_desc *gdp; |
443387113 ext4: fold ext4_f... |
4439 |
unsigned long freed = 0; |
498e5f241 ext4: Change unsi... |
4440 |
unsigned int overflow; |
c9de560de ext4: Add multi b... |
4441 4442 4443 4444 4445 4446 4447 |
ext4_grpblk_t bit; struct buffer_head *gd_bh; ext4_group_t block_group; struct ext4_sb_info *sbi; struct ext4_buddy e4b; int err = 0; int ret; |
e6362609b ext4: call ext4_f... |
4448 4449 4450 4451 4452 4453 |
if (bh) { if (block) BUG_ON(block != bh->b_blocknr); else block = bh->b_blocknr; } |
c9de560de ext4: Add multi b... |
4454 |
|
c9de560de ext4: Add multi b... |
4455 |
sbi = EXT4_SB(sb); |
1f2acb601 ext4: Add block v... |
4456 4457 |
if (!(flags & EXT4_FREE_BLOCKS_VALIDATED) && !ext4_data_block_valid(sbi, block, count)) { |
12062dddd ext4: move __func... |
4458 |
ext4_error(sb, "Freeing blocks not in datazone - " |
1f2acb601 ext4: Add block v... |
4459 |
"block = %llu, count = %lu", block, count); |
c9de560de ext4: Add multi b... |
4460 4461 |
goto error_return; } |
0610b6e99 ext4: Fix 64-bit ... |
4462 4463 |
ext4_debug("freeing block %llu ", block); |
e6362609b ext4: call ext4_f... |
4464 4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 |
trace_ext4_free_blocks(inode, block, count, flags); if (flags & EXT4_FREE_BLOCKS_FORGET) { struct buffer_head *tbh = bh; int i; BUG_ON(bh && (count > 1)); for (i = 0; i < count; i++) { if (!bh) tbh = sb_find_get_block(inode->i_sb, block + i); |
60e6679e2 ext4: Drop whites... |
4476 |
ext4_forget(handle, flags & EXT4_FREE_BLOCKS_METADATA, |
e6362609b ext4: call ext4_f... |
4477 4478 4479 |
inode, tbh, block + i); } } |
60e6679e2 ext4: Drop whites... |
4480 |
/* |
e6362609b ext4: call ext4_f... |
4481 4482 4483 4484 4485 4486 4487 4488 |
* We need to make sure we don't reuse the freed block until * after the transaction is committed, which we can do by * treating the block as metadata, below. We make an * exception if the inode is to be written in writeback mode * since writeback mode has weak data consistency guarantees. */ if (!ext4_should_writeback_data(inode)) flags |= EXT4_FREE_BLOCKS_METADATA; |
c9de560de ext4: Add multi b... |
4489 |
|
256bdb497 ext4: allocate st... |
4490 4491 |
ac = kmem_cache_alloc(ext4_ac_cachep, GFP_NOFS); if (ac) { |
256bdb497 ext4: allocate st... |
4492 4493 4494 |
ac->ac_inode = inode; ac->ac_sb = sb; } |
c9de560de ext4: Add multi b... |
4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 |
do_more: overflow = 0; ext4_get_group_no_and_offset(sb, block, &block_group, &bit); /* * Check to see if we are freeing blocks across a group * boundary. */ if (bit + count > EXT4_BLOCKS_PER_GROUP(sb)) { overflow = bit + count - EXT4_BLOCKS_PER_GROUP(sb); count -= overflow; } |
574ca174c ext4: Rename read... |
4508 |
bitmap_bh = ext4_read_block_bitmap(sb, block_group); |
ce89f46cb ext4: Improve err... |
4509 4510 |
if (!bitmap_bh) { err = -EIO; |
c9de560de ext4: Add multi b... |
4511 |
goto error_return; |
ce89f46cb ext4: Improve err... |
4512 |
} |
c9de560de ext4: Add multi b... |
4513 |
gdp = ext4_get_group_desc(sb, block_group, &gd_bh); |
ce89f46cb ext4: Improve err... |
4514 4515 |
if (!gdp) { err = -EIO; |
c9de560de ext4: Add multi b... |
4516 |
goto error_return; |
ce89f46cb ext4: Improve err... |
4517 |
} |
c9de560de ext4: Add multi b... |
4518 4519 4520 4521 4522 4523 4524 |
if (in_range(ext4_block_bitmap(sb, gdp), block, count) || in_range(ext4_inode_bitmap(sb, gdp), block, count) || in_range(block, ext4_inode_table(sb, gdp), EXT4_SB(sb)->s_itb_per_group) || in_range(block + count - 1, ext4_inode_table(sb, gdp), EXT4_SB(sb)->s_itb_per_group)) { |
12062dddd ext4: move __func... |
4525 |
ext4_error(sb, "Freeing blocks in system zone - " |
0610b6e99 ext4: Fix 64-bit ... |
4526 |
"Block = %llu, count = %lu", block, count); |
519deca04 ext4: Retry block... |
4527 4528 |
/* err = 0. ext4_std_error should be a no op */ goto error_return; |
c9de560de ext4: Add multi b... |
4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 |
} BUFFER_TRACE(bitmap_bh, "getting write access"); err = ext4_journal_get_write_access(handle, bitmap_bh); if (err) goto error_return; /* * We are about to modify some metadata. Call the journal APIs * to unshare ->b_data if a currently-committing transaction is * using it */ BUFFER_TRACE(gd_bh, "get_write_access"); err = ext4_journal_get_write_access(handle, gd_bh); if (err) goto error_return; |
c9de560de ext4: Add multi b... |
4545 4546 4547 4548 4549 4550 4551 |
#ifdef AGGRESSIVE_CHECK { int i; for (i = 0; i < count; i++) BUG_ON(!mb_test_bit(bit + i, bitmap_bh->b_data)); } #endif |
256bdb497 ext4: allocate st... |
4552 4553 4554 4555 |
if (ac) { ac->ac_b_ex.fe_group = block_group; ac->ac_b_ex.fe_start = bit; ac->ac_b_ex.fe_len = count; |
296c355cd ext4: Use tracepo... |
4556 |
trace_ext4_mballoc_free(ac); |
256bdb497 ext4: allocate st... |
4557 |
} |
c9de560de ext4: Add multi b... |
4558 |
|
920313a72 ext4: Use EXT4_GR... |
4559 4560 4561 |
err = ext4_mb_load_buddy(sb, block_group, &e4b); if (err) goto error_return; |
e6362609b ext4: call ext4_f... |
4562 4563 |
if ((flags & EXT4_FREE_BLOCKS_METADATA) && ext4_handle_valid(handle)) { |
7a2fcbf7f ext4: don't use b... |
4564 4565 4566 4567 4568 4569 4570 4571 4572 4573 |
struct ext4_free_data *new_entry; /* * blocks being freed are metadata. these blocks shouldn't * be used until this transaction is committed */ new_entry = kmem_cache_alloc(ext4_free_ext_cachep, GFP_NOFS); new_entry->start_blk = bit; new_entry->group = block_group; new_entry->count = count; new_entry->t_tid = handle->h_transaction->t_tid; |
955ce5f5b ext4: Convert ext... |
4574 |
|
7a2fcbf7f ext4: don't use b... |
4575 |
ext4_lock_group(sb, block_group); |
955ce5f5b ext4: Convert ext... |
4576 |
mb_clear_bits(bitmap_bh->b_data, bit, count); |
7a2fcbf7f ext4: don't use b... |
4577 |
ext4_mb_free_metadata(handle, &e4b, new_entry); |
c9de560de ext4: Add multi b... |
4578 |
} else { |
7a2fcbf7f ext4: don't use b... |
4579 4580 4581 4582 |
/* need to update group_info->bb_free and bitmap * with group lock held. generate_buddy look at * them with group lock_held */ |
955ce5f5b ext4: Convert ext... |
4583 4584 |
ext4_lock_group(sb, block_group); mb_clear_bits(bitmap_bh->b_data, bit, count); |
7e5a8cdd8 ext4: fix error p... |
4585 |
mb_free_blocks(inode, &e4b, bit, count); |
c9de560de ext4: Add multi b... |
4586 |
ext4_mb_return_to_preallocation(inode, &e4b, block, count); |
5c521830c ext4: Support dis... |
4587 4588 |
if (test_opt(sb, DISCARD)) ext4_issue_discard(sb, block_group, bit, count); |
c9de560de ext4: Add multi b... |
4589 |
} |
560671a0d ext4: Use high 16... |
4590 4591 |
ret = ext4_free_blks_count(sb, gdp) + count; ext4_free_blks_set(sb, gdp, ret); |
c9de560de ext4: Add multi b... |
4592 |
gdp->bg_checksum = ext4_group_desc_csum(sbi, block_group, gdp); |
955ce5f5b ext4: Convert ext... |
4593 |
ext4_unlock_group(sb, block_group); |
c9de560de ext4: Add multi b... |
4594 |
percpu_counter_add(&sbi->s_freeblocks_counter, count); |
772cb7c83 ext4: New inode a... |
4595 4596 |
if (sbi->s_log_groups_per_flex) { ext4_group_t flex_group = ext4_flex_group(sbi, block_group); |
9f24e4208 ext4: Use atomic_... |
4597 |
atomic_add(count, &sbi->s_flex_groups[flex_group].free_blocks); |
772cb7c83 ext4: New inode a... |
4598 |
} |
e39e07fdf ext4: rename ext4... |
4599 |
ext4_mb_unload_buddy(&e4b); |
c9de560de ext4: Add multi b... |
4600 |
|
443387113 ext4: fold ext4_f... |
4601 |
freed += count; |
c9de560de ext4: Add multi b... |
4602 |
|
7a2fcbf7f ext4: don't use b... |
4603 4604 4605 |
/* We dirtied the bitmap block */ BUFFER_TRACE(bitmap_bh, "dirtied bitmap block"); err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh); |
c9de560de ext4: Add multi b... |
4606 4607 |
/* And the group descriptor block */ BUFFER_TRACE(gd_bh, "dirtied group descriptor block"); |
0390131ba ext4: Allow ext4 ... |
4608 |
ret = ext4_handle_dirty_metadata(handle, NULL, gd_bh); |
c9de560de ext4: Add multi b... |
4609 4610 4611 4612 4613 4614 4615 4616 4617 |
if (!err) err = ret; if (overflow && !err) { block += count; count = overflow; put_bh(bitmap_bh); goto do_more; } |
a0375156c ext4: Clean up s_... |
4618 |
ext4_mark_super_dirty(sb); |
c9de560de ext4: Add multi b... |
4619 |
error_return: |
443387113 ext4: fold ext4_f... |
4620 |
if (freed) |
5dd4056db dquot: cleanup sp... |
4621 |
dquot_free_block(inode, freed); |
c9de560de ext4: Add multi b... |
4622 4623 |
brelse(bitmap_bh); ext4_std_error(sb, err); |
256bdb497 ext4: allocate st... |
4624 4625 |
if (ac) kmem_cache_free(ext4_ac_cachep, ac); |
c9de560de ext4: Add multi b... |
4626 4627 |
return; } |