1 // SPDX-License-Identifier: GPL-2.0 2 /* 3 * Copyright (c) 2003-2006, Cluster File Systems, Inc, info@clusterfs.com 4 * Written by Alex Tomas <alex@clusterfs.com> 5 */ 6 7 8 /* 9 * mballoc.c contains the multiblocks allocation routines 10 */ 11 12 #include "ext4_jbd2.h" 13 #include "mballoc.h" 14 #include <linux/log2.h> 15 #include <linux/module.h> 16 #include <linux/slab.h> 17 #include <linux/nospec.h> 18 #include <linux/backing-dev.h> 19 #include <trace/events/ext4.h> 20 21 /* 22 * MUSTDO: 23 * - test ext4_ext_search_left() and ext4_ext_search_right() 24 * - search for metadata in few groups 25 * 26 * TODO v4: 27 * - normalization should take into account whether file is still open 28 * - discard preallocations if no free space left (policy?) 29 * - don't normalize tails 30 * - quota 31 * - reservation for superuser 32 * 33 * TODO v3: 34 * - bitmap read-ahead (proposed by Oleg Drokin aka green) 35 * - track min/max extents in each group for better group selection 36 * - mb_mark_used() may allocate chunk right after splitting buddy 37 * - tree of groups sorted by number of free blocks 38 * - error handling 39 */ 40 41 /* 42 * The allocation request involve request for multiple number of blocks 43 * near to the goal(block) value specified. 44 * 45 * During initialization phase of the allocator we decide to use the 46 * group preallocation or inode preallocation depending on the size of 47 * the file. The size of the file could be the resulting file size we 48 * would have after allocation, or the current file size, which ever 49 * is larger. If the size is less than sbi->s_mb_stream_request we 50 * select to use the group preallocation. The default value of 51 * s_mb_stream_request is 16 blocks. This can also be tuned via 52 * /sys/fs/ext4/<partition>/mb_stream_req. The value is represented in 53 * terms of number of blocks. 54 * 55 * The main motivation for having small file use group preallocation is to 56 * ensure that we have small files closer together on the disk. 57 * 58 * First stage the allocator looks at the inode prealloc list, 59 * ext4_inode_info->i_prealloc_list, which contains list of prealloc 60 * spaces for this particular inode. The inode prealloc space is 61 * represented as: 62 * 63 * pa_lstart -> the logical start block for this prealloc space 64 * pa_pstart -> the physical start block for this prealloc space 65 * pa_len -> length for this prealloc space (in clusters) 66 * pa_free -> free space available in this prealloc space (in clusters) 67 * 68 * The inode preallocation space is used looking at the _logical_ start 69 * block. If only the logical file block falls within the range of prealloc 70 * space we will consume the particular prealloc space. This makes sure that 71 * we have contiguous physical blocks representing the file blocks 72 * 73 * The important thing to be noted in case of inode prealloc space is that 74 * we don't modify the values associated to inode prealloc space except 75 * pa_free. 76 * 77 * If we are not able to find blocks in the inode prealloc space and if we 78 * have the group allocation flag set then we look at the locality group 79 * prealloc space. These are per CPU prealloc list represented as 80 * 81 * ext4_sb_info.s_locality_groups[smp_processor_id()] 82 * 83 * The reason for having a per cpu locality group is to reduce the contention 84 * between CPUs. It is possible to get scheduled at this point. 85 * 86 * The locality group prealloc space is used looking at whether we have 87 * enough free space (pa_free) within the prealloc space. 88 * 89 * If we can't allocate blocks via inode prealloc or/and locality group 90 * prealloc then we look at the buddy cache. The buddy cache is represented 91 * by ext4_sb_info.s_buddy_cache (struct inode) whose file offset gets 92 * mapped to the buddy and bitmap information regarding different 93 * groups. The buddy information is attached to buddy cache inode so that 94 * we can access them through the page cache. The information regarding 95 * each group is loaded via ext4_mb_load_buddy. The information involve 96 * block bitmap and buddy information. The information are stored in the 97 * inode as: 98 * 99 * { page } 100 * [ group 0 bitmap][ group 0 buddy] [group 1][ group 1]... 101 * 102 * 103 * one block each for bitmap and buddy information. So for each group we 104 * take up 2 blocks. A page can contain blocks_per_page (PAGE_SIZE / 105 * blocksize) blocks. So it can have information regarding groups_per_page 106 * which is blocks_per_page/2 107 * 108 * The buddy cache inode is not stored on disk. The inode is thrown 109 * away when the filesystem is unmounted. 110 * 111 * We look for count number of blocks in the buddy cache. If we were able 112 * to locate that many free blocks we return with additional information 113 * regarding rest of the contiguous physical block available 114 * 115 * Before allocating blocks via buddy cache we normalize the request 116 * blocks. This ensure we ask for more blocks that we needed. The extra 117 * blocks that we get after allocation is added to the respective prealloc 118 * list. In case of inode preallocation we follow a list of heuristics 119 * based on file size. This can be found in ext4_mb_normalize_request. If 120 * we are doing a group prealloc we try to normalize the request to 121 * sbi->s_mb_group_prealloc. The default value of s_mb_group_prealloc is 122 * dependent on the cluster size; for non-bigalloc file systems, it is 123 * 512 blocks. This can be tuned via 124 * /sys/fs/ext4/<partition>/mb_group_prealloc. The value is represented in 125 * terms of number of blocks. If we have mounted the file system with -O 126 * stripe=<value> option the group prealloc request is normalized to the 127 * smallest multiple of the stripe value (sbi->s_stripe) which is 128 * greater than the default mb_group_prealloc. 129 * 130 * If "mb_optimize_scan" mount option is set, we maintain in memory group info 131 * structures in two data structures: 132 * 133 * 1) Array of largest free order lists (sbi->s_mb_largest_free_orders) 134 * 135 * Locking: sbi->s_mb_largest_free_orders_locks(array of rw locks) 136 * 137 * This is an array of lists where the index in the array represents the 138 * largest free order in the buddy bitmap of the participating group infos of 139 * that list. So, there are exactly MB_NUM_ORDERS(sb) (which means total 140 * number of buddy bitmap orders possible) number of lists. Group-infos are 141 * placed in appropriate lists. 142 * 143 * 2) Average fragment size rb tree (sbi->s_mb_avg_fragment_size_root) 144 * 145 * Locking: sbi->s_mb_rb_lock (rwlock) 146 * 147 * This is a red black tree consisting of group infos and the tree is sorted 148 * by average fragment sizes (which is calculated as ext4_group_info->bb_free 149 * / ext4_group_info->bb_fragments). 150 * 151 * When "mb_optimize_scan" mount option is set, mballoc consults the above data 152 * structures to decide the order in which groups are to be traversed for 153 * fulfilling an allocation request. 154 * 155 * At CR = 0, we look for groups which have the largest_free_order >= the order 156 * of the request. We directly look at the largest free order list in the data 157 * structure (1) above where largest_free_order = order of the request. If that 158 * list is empty, we look at remaining list in the increasing order of 159 * largest_free_order. This allows us to perform CR = 0 lookup in O(1) time. 160 * 161 * At CR = 1, we only consider groups where average fragment size > request 162 * size. So, we lookup a group which has average fragment size just above or 163 * equal to request size using our rb tree (data structure 2) in O(log N) time. 164 * 165 * If "mb_optimize_scan" mount option is not set, mballoc traverses groups in 166 * linear order which requires O(N) search time for each CR 0 and CR 1 phase. 167 * 168 * The regular allocator (using the buddy cache) supports a few tunables. 169 * 170 * /sys/fs/ext4/<partition>/mb_min_to_scan 171 * /sys/fs/ext4/<partition>/mb_max_to_scan 172 * /sys/fs/ext4/<partition>/mb_order2_req 173 * /sys/fs/ext4/<partition>/mb_linear_limit 174 * 175 * The regular allocator uses buddy scan only if the request len is power of 176 * 2 blocks and the order of allocation is >= sbi->s_mb_order2_reqs. The 177 * value of s_mb_order2_reqs can be tuned via 178 * /sys/fs/ext4/<partition>/mb_order2_req. If the request len is equal to 179 * stripe size (sbi->s_stripe), we try to search for contiguous block in 180 * stripe size. This should result in better allocation on RAID setups. If 181 * not, we search in the specific group using bitmap for best extents. The 182 * tunable min_to_scan and max_to_scan control the behaviour here. 183 * min_to_scan indicate how long the mballoc __must__ look for a best 184 * extent and max_to_scan indicates how long the mballoc __can__ look for a 185 * best extent in the found extents. Searching for the blocks starts with 186 * the group specified as the goal value in allocation context via 187 * ac_g_ex. Each group is first checked based on the criteria whether it 188 * can be used for allocation. ext4_mb_good_group explains how the groups are 189 * checked. 190 * 191 * When "mb_optimize_scan" is turned on, as mentioned above, the groups may not 192 * get traversed linearly. That may result in subsequent allocations being not 193 * close to each other. And so, the underlying device may get filled up in a 194 * non-linear fashion. While that may not matter on non-rotational devices, for 195 * rotational devices that may result in higher seek times. "mb_linear_limit" 196 * tells mballoc how many groups mballoc should search linearly before 197 * performing consulting above data structures for more efficient lookups. For 198 * non rotational devices, this value defaults to 0 and for rotational devices 199 * this is set to MB_DEFAULT_LINEAR_LIMIT. 200 * 201 * Both the prealloc space are getting populated as above. So for the first 202 * request we will hit the buddy cache which will result in this prealloc 203 * space getting filled. The prealloc space is then later used for the 204 * subsequent request. 205 */ 206 207 /* 208 * mballoc operates on the following data: 209 * - on-disk bitmap 210 * - in-core buddy (actually includes buddy and bitmap) 211 * - preallocation descriptors (PAs) 212 * 213 * there are two types of preallocations: 214 * - inode 215 * assiged to specific inode and can be used for this inode only. 216 * it describes part of inode's space preallocated to specific 217 * physical blocks. any block from that preallocated can be used 218 * independent. the descriptor just tracks number of blocks left 219 * unused. so, before taking some block from descriptor, one must 220 * make sure corresponded logical block isn't allocated yet. this 221 * also means that freeing any block within descriptor's range 222 * must discard all preallocated blocks. 223 * - locality group 224 * assigned to specific locality group which does not translate to 225 * permanent set of inodes: inode can join and leave group. space 226 * from this type of preallocation can be used for any inode. thus 227 * it's consumed from the beginning to the end. 228 * 229 * relation between them can be expressed as: 230 * in-core buddy = on-disk bitmap + preallocation descriptors 231 * 232 * this mean blocks mballoc considers used are: 233 * - allocated blocks (persistent) 234 * - preallocated blocks (non-persistent) 235 * 236 * consistency in mballoc world means that at any time a block is either 237 * free or used in ALL structures. notice: "any time" should not be read 238 * literally -- time is discrete and delimited by locks. 239 * 240 * to keep it simple, we don't use block numbers, instead we count number of 241 * blocks: how many blocks marked used/free in on-disk bitmap, buddy and PA. 242 * 243 * all operations can be expressed as: 244 * - init buddy: buddy = on-disk + PAs 245 * - new PA: buddy += N; PA = N 246 * - use inode PA: on-disk += N; PA -= N 247 * - discard inode PA buddy -= on-disk - PA; PA = 0 248 * - use locality group PA on-disk += N; PA -= N 249 * - discard locality group PA buddy -= PA; PA = 0 250 * note: 'buddy -= on-disk - PA' is used to show that on-disk bitmap 251 * is used in real operation because we can't know actual used 252 * bits from PA, only from on-disk bitmap 253 * 254 * if we follow this strict logic, then all operations above should be atomic. 255 * given some of them can block, we'd have to use something like semaphores 256 * killing performance on high-end SMP hardware. let's try to relax it using 257 * the following knowledge: 258 * 1) if buddy is referenced, it's already initialized 259 * 2) while block is used in buddy and the buddy is referenced, 260 * nobody can re-allocate that block 261 * 3) we work on bitmaps and '+' actually means 'set bits'. if on-disk has 262 * bit set and PA claims same block, it's OK. IOW, one can set bit in 263 * on-disk bitmap if buddy has same bit set or/and PA covers corresponded 264 * block 265 * 266 * so, now we're building a concurrency table: 267 * - init buddy vs. 268 * - new PA 269 * blocks for PA are allocated in the buddy, buddy must be referenced 270 * until PA is linked to allocation group to avoid concurrent buddy init 271 * - use inode PA 272 * we need to make sure that either on-disk bitmap or PA has uptodate data 273 * given (3) we care that PA-=N operation doesn't interfere with init 274 * - discard inode PA 275 * the simplest way would be to have buddy initialized by the discard 276 * - use locality group PA 277 * again PA-=N must be serialized with init 278 * - discard locality group PA 279 * the simplest way would be to have buddy initialized by the discard 280 * - new PA vs. 281 * - use inode PA 282 * i_data_sem serializes them 283 * - discard inode PA 284 * discard process must wait until PA isn't used by another process 285 * - use locality group PA 286 * some mutex should serialize them 287 * - discard locality group PA 288 * discard process must wait until PA isn't used by another process 289 * - use inode PA 290 * - use inode PA 291 * i_data_sem or another mutex should serializes them 292 * - discard inode PA 293 * discard process must wait until PA isn't used by another process 294 * - use locality group PA 295 * nothing wrong here -- they're different PAs covering different blocks 296 * - discard locality group PA 297 * discard process must wait until PA isn't used by another process 298 * 299 * now we're ready to make few consequences: 300 * - PA is referenced and while it is no discard is possible 301 * - PA is referenced until block isn't marked in on-disk bitmap 302 * - PA changes only after on-disk bitmap 303 * - discard must not compete with init. either init is done before 304 * any discard or they're serialized somehow 305 * - buddy init as sum of on-disk bitmap and PAs is done atomically 306 * 307 * a special case when we've used PA to emptiness. no need to modify buddy 308 * in this case, but we should care about concurrent init 309 * 310 */ 311 312 /* 313 * Logic in few words: 314 * 315 * - allocation: 316 * load group 317 * find blocks 318 * mark bits in on-disk bitmap 319 * release group 320 * 321 * - use preallocation: 322 * find proper PA (per-inode or group) 323 * load group 324 * mark bits in on-disk bitmap 325 * release group 326 * release PA 327 * 328 * - free: 329 * load group 330 * mark bits in on-disk bitmap 331 * release group 332 * 333 * - discard preallocations in group: 334 * mark PAs deleted 335 * move them onto local list 336 * load on-disk bitmap 337 * load group 338 * remove PA from object (inode or locality group) 339 * mark free blocks in-core 340 * 341 * - discard inode's preallocations: 342 */ 343 344 /* 345 * Locking rules 346 * 347 * Locks: 348 * - bitlock on a group (group) 349 * - object (inode/locality) (object) 350 * - per-pa lock (pa) 351 * - cr0 lists lock (cr0) 352 * - cr1 tree lock (cr1) 353 * 354 * Paths: 355 * - new pa 356 * object 357 * group 358 * 359 * - find and use pa: 360 * pa 361 * 362 * - release consumed pa: 363 * pa 364 * group 365 * object 366 * 367 * - generate in-core bitmap: 368 * group 369 * pa 370 * 371 * - discard all for given object (inode, locality group): 372 * object 373 * pa 374 * group 375 * 376 * - discard all for given group: 377 * group 378 * pa 379 * group 380 * object 381 * 382 * - allocation path (ext4_mb_regular_allocator) 383 * group 384 * cr0/cr1 385 */ 386 static struct kmem_cache *ext4_pspace_cachep; 387 static struct kmem_cache *ext4_ac_cachep; 388 static struct kmem_cache *ext4_free_data_cachep; 389 390 /* We create slab caches for groupinfo data structures based on the 391 * superblock block size. There will be one per mounted filesystem for 392 * each unique s_blocksize_bits */ 393 #define NR_GRPINFO_CACHES 8 394 static struct kmem_cache *ext4_groupinfo_caches[NR_GRPINFO_CACHES]; 395 396 static const char * const ext4_groupinfo_slab_names[NR_GRPINFO_CACHES] = { 397 "ext4_groupinfo_1k", "ext4_groupinfo_2k", "ext4_groupinfo_4k", 398 "ext4_groupinfo_8k", "ext4_groupinfo_16k", "ext4_groupinfo_32k", 399 "ext4_groupinfo_64k", "ext4_groupinfo_128k" 400 }; 401 402 static void ext4_mb_generate_from_pa(struct super_block *sb, void *bitmap, 403 ext4_group_t group); 404 static void ext4_mb_generate_from_freelist(struct super_block *sb, void *bitmap, 405 ext4_group_t group); 406 static void ext4_mb_new_preallocation(struct ext4_allocation_context *ac); 407 408 static bool ext4_mb_good_group(struct ext4_allocation_context *ac, 409 ext4_group_t group, int cr); 410 411 static int ext4_try_to_trim_range(struct super_block *sb, 412 struct ext4_buddy *e4b, ext4_grpblk_t start, 413 ext4_grpblk_t max, ext4_grpblk_t minblocks); 414 415 /* 416 * The algorithm using this percpu seq counter goes below: 417 * 1. We sample the percpu discard_pa_seq counter before trying for block 418 * allocation in ext4_mb_new_blocks(). 419 * 2. We increment this percpu discard_pa_seq counter when we either allocate 420 * or free these blocks i.e. while marking those blocks as used/free in 421 * mb_mark_used()/mb_free_blocks(). 422 * 3. We also increment this percpu seq counter when we successfully identify 423 * that the bb_prealloc_list is not empty and hence proceed for discarding 424 * of those PAs inside ext4_mb_discard_group_preallocations(). 425 * 426 * Now to make sure that the regular fast path of block allocation is not 427 * affected, as a small optimization we only sample the percpu seq counter 428 * on that cpu. Only when the block allocation fails and when freed blocks 429 * found were 0, that is when we sample percpu seq counter for all cpus using 430 * below function ext4_get_discard_pa_seq_sum(). This happens after making 431 * sure that all the PAs on grp->bb_prealloc_list got freed or if it's empty. 432 */ 433 static DEFINE_PER_CPU(u64, discard_pa_seq); 434 static inline u64 ext4_get_discard_pa_seq_sum(void) 435 { 436 int __cpu; 437 u64 __seq = 0; 438 439 for_each_possible_cpu(__cpu) 440 __seq += per_cpu(discard_pa_seq, __cpu); 441 return __seq; 442 } 443 444 static inline void *mb_correct_addr_and_bit(int *bit, void *addr) 445 { 446 #if BITS_PER_LONG == 64 447 *bit += ((unsigned long) addr & 7UL) << 3; 448 addr = (void *) ((unsigned long) addr & ~7UL); 449 #elif BITS_PER_LONG == 32 450 *bit += ((unsigned long) addr & 3UL) << 3; 451 addr = (void *) ((unsigned long) addr & ~3UL); 452 #else 453 #error "how many bits you are?!" 454 #endif 455 return addr; 456 } 457 458 static inline int mb_test_bit(int bit, void *addr) 459 { 460 /* 461 * ext4_test_bit on architecture like powerpc 462 * needs unsigned long aligned address 463 */ 464 addr = mb_correct_addr_and_bit(&bit, addr); 465 return ext4_test_bit(bit, addr); 466 } 467 468 static inline void mb_set_bit(int bit, void *addr) 469 { 470 addr = mb_correct_addr_and_bit(&bit, addr); 471 ext4_set_bit(bit, addr); 472 } 473 474 static inline void mb_clear_bit(int bit, void *addr) 475 { 476 addr = mb_correct_addr_and_bit(&bit, addr); 477 ext4_clear_bit(bit, addr); 478 } 479 480 static inline int mb_test_and_clear_bit(int bit, void *addr) 481 { 482 addr = mb_correct_addr_and_bit(&bit, addr); 483 return ext4_test_and_clear_bit(bit, addr); 484 } 485 486 static inline int mb_find_next_zero_bit(void *addr, int max, int start) 487 { 488 int fix = 0, ret, tmpmax; 489 addr = mb_correct_addr_and_bit(&fix, addr); 490 tmpmax = max + fix; 491 start += fix; 492 493 ret = ext4_find_next_zero_bit(addr, tmpmax, start) - fix; 494 if (ret > max) 495 return max; 496 return ret; 497 } 498 499 static inline int mb_find_next_bit(void *addr, int max, int start) 500 { 501 int fix = 0, ret, tmpmax; 502 addr = mb_correct_addr_and_bit(&fix, addr); 503 tmpmax = max + fix; 504 start += fix; 505 506 ret = ext4_find_next_bit(addr, tmpmax, start) - fix; 507 if (ret > max) 508 return max; 509 return ret; 510 } 511 512 static void *mb_find_buddy(struct ext4_buddy *e4b, int order, int *max) 513 { 514 char *bb; 515 516 BUG_ON(e4b->bd_bitmap == e4b->bd_buddy); 517 BUG_ON(max == NULL); 518 519 if (order > e4b->bd_blkbits + 1) { 520 *max = 0; 521 return NULL; 522 } 523 524 /* at order 0 we see each particular block */ 525 if (order == 0) { 526 *max = 1 << (e4b->bd_blkbits + 3); 527 return e4b->bd_bitmap; 528 } 529 530 bb = e4b->bd_buddy + EXT4_SB(e4b->bd_sb)->s_mb_offsets[order]; 531 *max = EXT4_SB(e4b->bd_sb)->s_mb_maxs[order]; 532 533 return bb; 534 } 535 536 #ifdef DOUBLE_CHECK 537 static void mb_free_blocks_double(struct inode *inode, struct ext4_buddy *e4b, 538 int first, int count) 539 { 540 int i; 541 struct super_block *sb = e4b->bd_sb; 542 543 if (unlikely(e4b->bd_info->bb_bitmap == NULL)) 544 return; 545 assert_spin_locked(ext4_group_lock_ptr(sb, e4b->bd_group)); 546 for (i = 0; i < count; i++) { 547 if (!mb_test_bit(first + i, e4b->bd_info->bb_bitmap)) { 548 ext4_fsblk_t blocknr; 549 550 blocknr = ext4_group_first_block_no(sb, e4b->bd_group); 551 blocknr += EXT4_C2B(EXT4_SB(sb), first + i); 552 ext4_grp_locked_error(sb, e4b->bd_group, 553 inode ? inode->i_ino : 0, 554 blocknr, 555 "freeing block already freed " 556 "(bit %u)", 557 first + i); 558 ext4_mark_group_bitmap_corrupted(sb, e4b->bd_group, 559 EXT4_GROUP_INFO_BBITMAP_CORRUPT); 560 } 561 mb_clear_bit(first + i, e4b->bd_info->bb_bitmap); 562 } 563 } 564 565 static void mb_mark_used_double(struct ext4_buddy *e4b, int first, int count) 566 { 567 int i; 568 569 if (unlikely(e4b->bd_info->bb_bitmap == NULL)) 570 return; 571 assert_spin_locked(ext4_group_lock_ptr(e4b->bd_sb, e4b->bd_group)); 572 for (i = 0; i < count; i++) { 573 BUG_ON(mb_test_bit(first + i, e4b->bd_info->bb_bitmap)); 574 mb_set_bit(first + i, e4b->bd_info->bb_bitmap); 575 } 576 } 577 578 static void mb_cmp_bitmaps(struct ext4_buddy *e4b, void *bitmap) 579 { 580 if (unlikely(e4b->bd_info->bb_bitmap == NULL)) 581 return; 582 if (memcmp(e4b->bd_info->bb_bitmap, bitmap, e4b->bd_sb->s_blocksize)) { 583 unsigned char *b1, *b2; 584 int i; 585 b1 = (unsigned char *) e4b->bd_info->bb_bitmap; 586 b2 = (unsigned char *) bitmap; 587 for (i = 0; i < e4b->bd_sb->s_blocksize; i++) { 588 if (b1[i] != b2[i]) { 589 ext4_msg(e4b->bd_sb, KERN_ERR, 590 "corruption in group %u " 591 "at byte %u(%u): %x in copy != %x " 592 "on disk/prealloc", 593 e4b->bd_group, i, i * 8, b1[i], b2[i]); 594 BUG(); 595 } 596 } 597 } 598 } 599 600 static void mb_group_bb_bitmap_alloc(struct super_block *sb, 601 struct ext4_group_info *grp, ext4_group_t group) 602 { 603 struct buffer_head *bh; 604 605 grp->bb_bitmap = kmalloc(sb->s_blocksize, GFP_NOFS); 606 if (!grp->bb_bitmap) 607 return; 608 609 bh = ext4_read_block_bitmap(sb, group); 610 if (IS_ERR_OR_NULL(bh)) { 611 kfree(grp->bb_bitmap); 612 grp->bb_bitmap = NULL; 613 return; 614 } 615 616 memcpy(grp->bb_bitmap, bh->b_data, sb->s_blocksize); 617 put_bh(bh); 618 } 619 620 static void mb_group_bb_bitmap_free(struct ext4_group_info *grp) 621 { 622 kfree(grp->bb_bitmap); 623 } 624 625 #else 626 static inline void mb_free_blocks_double(struct inode *inode, 627 struct ext4_buddy *e4b, int first, int count) 628 { 629 return; 630 } 631 static inline void mb_mark_used_double(struct ext4_buddy *e4b, 632 int first, int count) 633 { 634 return; 635 } 636 static inline void mb_cmp_bitmaps(struct ext4_buddy *e4b, void *bitmap) 637 { 638 return; 639 } 640 641 static inline void mb_group_bb_bitmap_alloc(struct super_block *sb, 642 struct ext4_group_info *grp, ext4_group_t group) 643 { 644 return; 645 } 646 647 static inline void mb_group_bb_bitmap_free(struct ext4_group_info *grp) 648 { 649 return; 650 } 651 #endif 652 653 #ifdef AGGRESSIVE_CHECK 654 655 #define MB_CHECK_ASSERT(assert) \ 656 do { \ 657 if (!(assert)) { \ 658 printk(KERN_EMERG \ 659 "Assertion failure in %s() at %s:%d: \"%s\"\n", \ 660 function, file, line, # assert); \ 661 BUG(); \ 662 } \ 663 } while (0) 664 665 static int __mb_check_buddy(struct ext4_buddy *e4b, char *file, 666 const char *function, int line) 667 { 668 struct super_block *sb = e4b->bd_sb; 669 int order = e4b->bd_blkbits + 1; 670 int max; 671 int max2; 672 int i; 673 int j; 674 int k; 675 int count; 676 struct ext4_group_info *grp; 677 int fragments = 0; 678 int fstart; 679 struct list_head *cur; 680 void *buddy; 681 void *buddy2; 682 683 if (e4b->bd_info->bb_check_counter++ % 10) 684 return 0; 685 686 while (order > 1) { 687 buddy = mb_find_buddy(e4b, order, &max); 688 MB_CHECK_ASSERT(buddy); 689 buddy2 = mb_find_buddy(e4b, order - 1, &max2); 690 MB_CHECK_ASSERT(buddy2); 691 MB_CHECK_ASSERT(buddy != buddy2); 692 MB_CHECK_ASSERT(max * 2 == max2); 693 694 count = 0; 695 for (i = 0; i < max; i++) { 696 697 if (mb_test_bit(i, buddy)) { 698 /* only single bit in buddy2 may be 0 */ 699 if (!mb_test_bit(i << 1, buddy2)) { 700 MB_CHECK_ASSERT( 701 mb_test_bit((i<<1)+1, buddy2)); 702 } 703 continue; 704 } 705 706 /* both bits in buddy2 must be 1 */ 707 MB_CHECK_ASSERT(mb_test_bit(i << 1, buddy2)); 708 MB_CHECK_ASSERT(mb_test_bit((i << 1) + 1, buddy2)); 709 710 for (j = 0; j < (1 << order); j++) { 711 k = (i * (1 << order)) + j; 712 MB_CHECK_ASSERT( 713 !mb_test_bit(k, e4b->bd_bitmap)); 714 } 715 count++; 716 } 717 MB_CHECK_ASSERT(e4b->bd_info->bb_counters[order] == count); 718 order--; 719 } 720 721 fstart = -1; 722 buddy = mb_find_buddy(e4b, 0, &max); 723 for (i = 0; i < max; i++) { 724 if (!mb_test_bit(i, buddy)) { 725 MB_CHECK_ASSERT(i >= e4b->bd_info->bb_first_free); 726 if (fstart == -1) { 727 fragments++; 728 fstart = i; 729 } 730 continue; 731 } 732 fstart = -1; 733 /* check used bits only */ 734 for (j = 0; j < e4b->bd_blkbits + 1; j++) { 735 buddy2 = mb_find_buddy(e4b, j, &max2); 736 k = i >> j; 737 MB_CHECK_ASSERT(k < max2); 738 MB_CHECK_ASSERT(mb_test_bit(k, buddy2)); 739 } 740 } 741 MB_CHECK_ASSERT(!EXT4_MB_GRP_NEED_INIT(e4b->bd_info)); 742 MB_CHECK_ASSERT(e4b->bd_info->bb_fragments == fragments); 743 744 grp = ext4_get_group_info(sb, e4b->bd_group); 745 list_for_each(cur, &grp->bb_prealloc_list) { 746 ext4_group_t groupnr; 747 struct ext4_prealloc_space *pa; 748 pa = list_entry(cur, struct ext4_prealloc_space, pa_group_list); 749 ext4_get_group_no_and_offset(sb, pa->pa_pstart, &groupnr, &k); 750 MB_CHECK_ASSERT(groupnr == e4b->bd_group); 751 for (i = 0; i < pa->pa_len; i++) 752 MB_CHECK_ASSERT(mb_test_bit(k + i, buddy)); 753 } 754 return 0; 755 } 756 #undef MB_CHECK_ASSERT 757 #define mb_check_buddy(e4b) __mb_check_buddy(e4b, \ 758 __FILE__, __func__, __LINE__) 759 #else 760 #define mb_check_buddy(e4b) 761 #endif 762 763 /* 764 * Divide blocks started from @first with length @len into 765 * smaller chunks with power of 2 blocks. 766 * Clear the bits in bitmap which the blocks of the chunk(s) covered, 767 * then increase bb_counters[] for corresponded chunk size. 768 */ 769 static void ext4_mb_mark_free_simple(struct super_block *sb, 770 void *buddy, ext4_grpblk_t first, ext4_grpblk_t len, 771 struct ext4_group_info *grp) 772 { 773 struct ext4_sb_info *sbi = EXT4_SB(sb); 774 ext4_grpblk_t min; 775 ext4_grpblk_t max; 776 ext4_grpblk_t chunk; 777 unsigned int border; 778 779 BUG_ON(len > EXT4_CLUSTERS_PER_GROUP(sb)); 780 781 border = 2 << sb->s_blocksize_bits; 782 783 while (len > 0) { 784 /* find how many blocks can be covered since this position */ 785 max = ffs(first | border) - 1; 786 787 /* find how many blocks of power 2 we need to mark */ 788 min = fls(len) - 1; 789 790 if (max < min) 791 min = max; 792 chunk = 1 << min; 793 794 /* mark multiblock chunks only */ 795 grp->bb_counters[min]++; 796 if (min > 0) 797 mb_clear_bit(first >> min, 798 buddy + sbi->s_mb_offsets[min]); 799 800 len -= chunk; 801 first += chunk; 802 } 803 } 804 805 static void ext4_mb_rb_insert(struct rb_root *root, struct rb_node *new, 806 int (*cmp)(struct rb_node *, struct rb_node *)) 807 { 808 struct rb_node **iter = &root->rb_node, *parent = NULL; 809 810 while (*iter) { 811 parent = *iter; 812 if (cmp(new, *iter) > 0) 813 iter = &((*iter)->rb_left); 814 else 815 iter = &((*iter)->rb_right); 816 } 817 818 rb_link_node(new, parent, iter); 819 rb_insert_color(new, root); 820 } 821 822 static int 823 ext4_mb_avg_fragment_size_cmp(struct rb_node *rb1, struct rb_node *rb2) 824 { 825 struct ext4_group_info *grp1 = rb_entry(rb1, 826 struct ext4_group_info, 827 bb_avg_fragment_size_rb); 828 struct ext4_group_info *grp2 = rb_entry(rb2, 829 struct ext4_group_info, 830 bb_avg_fragment_size_rb); 831 int num_frags_1, num_frags_2; 832 833 num_frags_1 = grp1->bb_fragments ? 834 grp1->bb_free / grp1->bb_fragments : 0; 835 num_frags_2 = grp2->bb_fragments ? 836 grp2->bb_free / grp2->bb_fragments : 0; 837 838 return (num_frags_2 - num_frags_1); 839 } 840 841 /* 842 * Reinsert grpinfo into the avg_fragment_size tree with new average 843 * fragment size. 844 */ 845 static void 846 mb_update_avg_fragment_size(struct super_block *sb, struct ext4_group_info *grp) 847 { 848 struct ext4_sb_info *sbi = EXT4_SB(sb); 849 850 if (!test_opt2(sb, MB_OPTIMIZE_SCAN) || grp->bb_free == 0) 851 return; 852 853 write_lock(&sbi->s_mb_rb_lock); 854 if (!RB_EMPTY_NODE(&grp->bb_avg_fragment_size_rb)) { 855 rb_erase(&grp->bb_avg_fragment_size_rb, 856 &sbi->s_mb_avg_fragment_size_root); 857 RB_CLEAR_NODE(&grp->bb_avg_fragment_size_rb); 858 } 859 860 ext4_mb_rb_insert(&sbi->s_mb_avg_fragment_size_root, 861 &grp->bb_avg_fragment_size_rb, 862 ext4_mb_avg_fragment_size_cmp); 863 write_unlock(&sbi->s_mb_rb_lock); 864 } 865 866 /* 867 * Choose next group by traversing largest_free_order lists. Updates *new_cr if 868 * cr level needs an update. 869 */ 870 static void ext4_mb_choose_next_group_cr0(struct ext4_allocation_context *ac, 871 int *new_cr, ext4_group_t *group, ext4_group_t ngroups) 872 { 873 struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); 874 struct ext4_group_info *iter, *grp; 875 int i; 876 877 if (ac->ac_status == AC_STATUS_FOUND) 878 return; 879 880 if (unlikely(sbi->s_mb_stats && ac->ac_flags & EXT4_MB_CR0_OPTIMIZED)) 881 atomic_inc(&sbi->s_bal_cr0_bad_suggestions); 882 883 grp = NULL; 884 for (i = ac->ac_2order; i < MB_NUM_ORDERS(ac->ac_sb); i++) { 885 if (list_empty(&sbi->s_mb_largest_free_orders[i])) 886 continue; 887 read_lock(&sbi->s_mb_largest_free_orders_locks[i]); 888 if (list_empty(&sbi->s_mb_largest_free_orders[i])) { 889 read_unlock(&sbi->s_mb_largest_free_orders_locks[i]); 890 continue; 891 } 892 grp = NULL; 893 list_for_each_entry(iter, &sbi->s_mb_largest_free_orders[i], 894 bb_largest_free_order_node) { 895 if (sbi->s_mb_stats) 896 atomic64_inc(&sbi->s_bal_cX_groups_considered[0]); 897 if (likely(ext4_mb_good_group(ac, iter->bb_group, 0))) { 898 grp = iter; 899 break; 900 } 901 } 902 read_unlock(&sbi->s_mb_largest_free_orders_locks[i]); 903 if (grp) 904 break; 905 } 906 907 if (!grp) { 908 /* Increment cr and search again */ 909 *new_cr = 1; 910 } else { 911 *group = grp->bb_group; 912 ac->ac_last_optimal_group = *group; 913 ac->ac_flags |= EXT4_MB_CR0_OPTIMIZED; 914 } 915 } 916 917 /* 918 * Choose next group by traversing average fragment size tree. Updates *new_cr 919 * if cr lvel needs an update. Sets EXT4_MB_SEARCH_NEXT_LINEAR to indicate that 920 * the linear search should continue for one iteration since there's lock 921 * contention on the rb tree lock. 922 */ 923 static void ext4_mb_choose_next_group_cr1(struct ext4_allocation_context *ac, 924 int *new_cr, ext4_group_t *group, ext4_group_t ngroups) 925 { 926 struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); 927 int avg_fragment_size, best_so_far; 928 struct rb_node *node, *found; 929 struct ext4_group_info *grp; 930 931 /* 932 * If there is contention on the lock, instead of waiting for the lock 933 * to become available, just continue searching lineraly. We'll resume 934 * our rb tree search later starting at ac->ac_last_optimal_group. 935 */ 936 if (!read_trylock(&sbi->s_mb_rb_lock)) { 937 ac->ac_flags |= EXT4_MB_SEARCH_NEXT_LINEAR; 938 return; 939 } 940 941 if (unlikely(ac->ac_flags & EXT4_MB_CR1_OPTIMIZED)) { 942 if (sbi->s_mb_stats) 943 atomic_inc(&sbi->s_bal_cr1_bad_suggestions); 944 /* We have found something at CR 1 in the past */ 945 grp = ext4_get_group_info(ac->ac_sb, ac->ac_last_optimal_group); 946 for (found = rb_next(&grp->bb_avg_fragment_size_rb); found != NULL; 947 found = rb_next(found)) { 948 grp = rb_entry(found, struct ext4_group_info, 949 bb_avg_fragment_size_rb); 950 if (sbi->s_mb_stats) 951 atomic64_inc(&sbi->s_bal_cX_groups_considered[1]); 952 if (likely(ext4_mb_good_group(ac, grp->bb_group, 1))) 953 break; 954 } 955 goto done; 956 } 957 958 node = sbi->s_mb_avg_fragment_size_root.rb_node; 959 best_so_far = 0; 960 found = NULL; 961 962 while (node) { 963 grp = rb_entry(node, struct ext4_group_info, 964 bb_avg_fragment_size_rb); 965 avg_fragment_size = 0; 966 if (ext4_mb_good_group(ac, grp->bb_group, 1)) { 967 avg_fragment_size = grp->bb_fragments ? 968 grp->bb_free / grp->bb_fragments : 0; 969 if (!best_so_far || avg_fragment_size < best_so_far) { 970 best_so_far = avg_fragment_size; 971 found = node; 972 } 973 } 974 if (avg_fragment_size > ac->ac_g_ex.fe_len) 975 node = node->rb_right; 976 else 977 node = node->rb_left; 978 } 979 980 done: 981 if (found) { 982 grp = rb_entry(found, struct ext4_group_info, 983 bb_avg_fragment_size_rb); 984 *group = grp->bb_group; 985 ac->ac_flags |= EXT4_MB_CR1_OPTIMIZED; 986 } else { 987 *new_cr = 2; 988 } 989 990 read_unlock(&sbi->s_mb_rb_lock); 991 ac->ac_last_optimal_group = *group; 992 } 993 994 static inline int should_optimize_scan(struct ext4_allocation_context *ac) 995 { 996 if (unlikely(!test_opt2(ac->ac_sb, MB_OPTIMIZE_SCAN))) 997 return 0; 998 if (ac->ac_criteria >= 2) 999 return 0; 1000 if (!ext4_test_inode_flag(ac->ac_inode, EXT4_INODE_EXTENTS)) 1001 return 0; 1002 return 1; 1003 } 1004 1005 /* 1006 * Return next linear group for allocation. If linear traversal should not be 1007 * performed, this function just returns the same group 1008 */ 1009 static int 1010 next_linear_group(struct ext4_allocation_context *ac, int group, int ngroups) 1011 { 1012 if (!should_optimize_scan(ac)) 1013 goto inc_and_return; 1014 1015 if (ac->ac_groups_linear_remaining) { 1016 ac->ac_groups_linear_remaining--; 1017 goto inc_and_return; 1018 } 1019 1020 if (ac->ac_flags & EXT4_MB_SEARCH_NEXT_LINEAR) { 1021 ac->ac_flags &= ~EXT4_MB_SEARCH_NEXT_LINEAR; 1022 goto inc_and_return; 1023 } 1024 1025 return group; 1026 inc_and_return: 1027 /* 1028 * Artificially restricted ngroups for non-extent 1029 * files makes group > ngroups possible on first loop. 1030 */ 1031 return group + 1 >= ngroups ? 0 : group + 1; 1032 } 1033 1034 /* 1035 * ext4_mb_choose_next_group: choose next group for allocation. 1036 * 1037 * @ac Allocation Context 1038 * @new_cr This is an output parameter. If the there is no good group 1039 * available at current CR level, this field is updated to indicate 1040 * the new cr level that should be used. 1041 * @group This is an input / output parameter. As an input it indicates the 1042 * next group that the allocator intends to use for allocation. As 1043 * output, this field indicates the next group that should be used as 1044 * determined by the optimization functions. 1045 * @ngroups Total number of groups 1046 */ 1047 static void ext4_mb_choose_next_group(struct ext4_allocation_context *ac, 1048 int *new_cr, ext4_group_t *group, ext4_group_t ngroups) 1049 { 1050 *new_cr = ac->ac_criteria; 1051 1052 if (!should_optimize_scan(ac) || ac->ac_groups_linear_remaining) 1053 return; 1054 1055 if (*new_cr == 0) { 1056 ext4_mb_choose_next_group_cr0(ac, new_cr, group, ngroups); 1057 } else if (*new_cr == 1) { 1058 ext4_mb_choose_next_group_cr1(ac, new_cr, group, ngroups); 1059 } else { 1060 /* 1061 * TODO: For CR=2, we can arrange groups in an rb tree sorted by 1062 * bb_free. But until that happens, we should never come here. 1063 */ 1064 WARN_ON(1); 1065 } 1066 } 1067 1068 /* 1069 * Cache the order of the largest free extent we have available in this block 1070 * group. 1071 */ 1072 static void 1073 mb_set_largest_free_order(struct super_block *sb, struct ext4_group_info *grp) 1074 { 1075 struct ext4_sb_info *sbi = EXT4_SB(sb); 1076 int i; 1077 1078 if (test_opt2(sb, MB_OPTIMIZE_SCAN) && grp->bb_largest_free_order >= 0) { 1079 write_lock(&sbi->s_mb_largest_free_orders_locks[ 1080 grp->bb_largest_free_order]); 1081 list_del_init(&grp->bb_largest_free_order_node); 1082 write_unlock(&sbi->s_mb_largest_free_orders_locks[ 1083 grp->bb_largest_free_order]); 1084 } 1085 grp->bb_largest_free_order = -1; /* uninit */ 1086 1087 for (i = MB_NUM_ORDERS(sb) - 1; i >= 0; i--) { 1088 if (grp->bb_counters[i] > 0) { 1089 grp->bb_largest_free_order = i; 1090 break; 1091 } 1092 } 1093 if (test_opt2(sb, MB_OPTIMIZE_SCAN) && 1094 grp->bb_largest_free_order >= 0 && grp->bb_free) { 1095 write_lock(&sbi->s_mb_largest_free_orders_locks[ 1096 grp->bb_largest_free_order]); 1097 list_add_tail(&grp->bb_largest_free_order_node, 1098 &sbi->s_mb_largest_free_orders[grp->bb_largest_free_order]); 1099 write_unlock(&sbi->s_mb_largest_free_orders_locks[ 1100 grp->bb_largest_free_order]); 1101 } 1102 } 1103 1104 static noinline_for_stack 1105 void ext4_mb_generate_buddy(struct super_block *sb, 1106 void *buddy, void *bitmap, ext4_group_t group) 1107 { 1108 struct ext4_group_info *grp = ext4_get_group_info(sb, group); 1109 struct ext4_sb_info *sbi = EXT4_SB(sb); 1110 ext4_grpblk_t max = EXT4_CLUSTERS_PER_GROUP(sb); 1111 ext4_grpblk_t i = 0; 1112 ext4_grpblk_t first; 1113 ext4_grpblk_t len; 1114 unsigned free = 0; 1115 unsigned fragments = 0; 1116 unsigned long long period = get_cycles(); 1117 1118 /* initialize buddy from bitmap which is aggregation 1119 * of on-disk bitmap and preallocations */ 1120 i = mb_find_next_zero_bit(bitmap, max, 0); 1121 grp->bb_first_free = i; 1122 while (i < max) { 1123 fragments++; 1124 first = i; 1125 i = mb_find_next_bit(bitmap, max, i); 1126 len = i - first; 1127 free += len; 1128 if (len > 1) 1129 ext4_mb_mark_free_simple(sb, buddy, first, len, grp); 1130 else 1131 grp->bb_counters[0]++; 1132 if (i < max) 1133 i = mb_find_next_zero_bit(bitmap, max, i); 1134 } 1135 grp->bb_fragments = fragments; 1136 1137 if (free != grp->bb_free) { 1138 ext4_grp_locked_error(sb, group, 0, 0, 1139 "block bitmap and bg descriptor " 1140 "inconsistent: %u vs %u free clusters", 1141 free, grp->bb_free); 1142 /* 1143 * If we intend to continue, we consider group descriptor 1144 * corrupt and update bb_free using bitmap value 1145 */ 1146 grp->bb_free = free; 1147 ext4_mark_group_bitmap_corrupted(sb, group, 1148 EXT4_GROUP_INFO_BBITMAP_CORRUPT); 1149 } 1150 mb_set_largest_free_order(sb, grp); 1151 1152 clear_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &(grp->bb_state)); 1153 1154 period = get_cycles() - period; 1155 atomic_inc(&sbi->s_mb_buddies_generated); 1156 atomic64_add(period, &sbi->s_mb_generation_time); 1157 mb_update_avg_fragment_size(sb, grp); 1158 } 1159 1160 /* The buddy information is attached the buddy cache inode 1161 * for convenience. The information regarding each group 1162 * is loaded via ext4_mb_load_buddy. The information involve 1163 * block bitmap and buddy information. The information are 1164 * stored in the inode as 1165 * 1166 * { page } 1167 * [ group 0 bitmap][ group 0 buddy] [group 1][ group 1]... 1168 * 1169 * 1170 * one block each for bitmap and buddy information. 1171 * So for each group we take up 2 blocks. A page can 1172 * contain blocks_per_page (PAGE_SIZE / blocksize) blocks. 1173 * So it can have information regarding groups_per_page which 1174 * is blocks_per_page/2 1175 * 1176 * Locking note: This routine takes the block group lock of all groups 1177 * for this page; do not hold this lock when calling this routine! 1178 */ 1179 1180 static int ext4_mb_init_cache(struct page *page, char *incore, gfp_t gfp) 1181 { 1182 ext4_group_t ngroups; 1183 int blocksize; 1184 int blocks_per_page; 1185 int groups_per_page; 1186 int err = 0; 1187 int i; 1188 ext4_group_t first_group, group; 1189 int first_block; 1190 struct super_block *sb; 1191 struct buffer_head *bhs; 1192 struct buffer_head **bh = NULL; 1193 struct inode *inode; 1194 char *data; 1195 char *bitmap; 1196 struct ext4_group_info *grinfo; 1197 1198 inode = page->mapping->host; 1199 sb = inode->i_sb; 1200 ngroups = ext4_get_groups_count(sb); 1201 blocksize = i_blocksize(inode); 1202 blocks_per_page = PAGE_SIZE / blocksize; 1203 1204 mb_debug(sb, "init page %lu\n", page->index); 1205 1206 groups_per_page = blocks_per_page >> 1; 1207 if (groups_per_page == 0) 1208 groups_per_page = 1; 1209 1210 /* allocate buffer_heads to read bitmaps */ 1211 if (groups_per_page > 1) { 1212 i = sizeof(struct buffer_head *) * groups_per_page; 1213 bh = kzalloc(i, gfp); 1214 if (bh == NULL) { 1215 err = -ENOMEM; 1216 goto out; 1217 } 1218 } else 1219 bh = &bhs; 1220 1221 first_group = page->index * blocks_per_page / 2; 1222 1223 /* read all groups the page covers into the cache */ 1224 for (i = 0, group = first_group; i < groups_per_page; i++, group++) { 1225 if (group >= ngroups) 1226 break; 1227 1228 grinfo = ext4_get_group_info(sb, group); 1229 /* 1230 * If page is uptodate then we came here after online resize 1231 * which added some new uninitialized group info structs, so 1232 * we must skip all initialized uptodate buddies on the page, 1233 * which may be currently in use by an allocating task. 1234 */ 1235 if (PageUptodate(page) && !EXT4_MB_GRP_NEED_INIT(grinfo)) { 1236 bh[i] = NULL; 1237 continue; 1238 } 1239 bh[i] = ext4_read_block_bitmap_nowait(sb, group, false); 1240 if (IS_ERR(bh[i])) { 1241 err = PTR_ERR(bh[i]); 1242 bh[i] = NULL; 1243 goto out; 1244 } 1245 mb_debug(sb, "read bitmap for group %u\n", group); 1246 } 1247 1248 /* wait for I/O completion */ 1249 for (i = 0, group = first_group; i < groups_per_page; i++, group++) { 1250 int err2; 1251 1252 if (!bh[i]) 1253 continue; 1254 err2 = ext4_wait_block_bitmap(sb, group, bh[i]); 1255 if (!err) 1256 err = err2; 1257 } 1258 1259 first_block = page->index * blocks_per_page; 1260 for (i = 0; i < blocks_per_page; i++) { 1261 group = (first_block + i) >> 1; 1262 if (group >= ngroups) 1263 break; 1264 1265 if (!bh[group - first_group]) 1266 /* skip initialized uptodate buddy */ 1267 continue; 1268 1269 if (!buffer_verified(bh[group - first_group])) 1270 /* Skip faulty bitmaps */ 1271 continue; 1272 err = 0; 1273 1274 /* 1275 * data carry information regarding this 1276 * particular group in the format specified 1277 * above 1278 * 1279 */ 1280 data = page_address(page) + (i * blocksize); 1281 bitmap = bh[group - first_group]->b_data; 1282 1283 /* 1284 * We place the buddy block and bitmap block 1285 * close together 1286 */ 1287 if ((first_block + i) & 1) { 1288 /* this is block of buddy */ 1289 BUG_ON(incore == NULL); 1290 mb_debug(sb, "put buddy for group %u in page %lu/%x\n", 1291 group, page->index, i * blocksize); 1292 trace_ext4_mb_buddy_bitmap_load(sb, group); 1293 grinfo = ext4_get_group_info(sb, group); 1294 grinfo->bb_fragments = 0; 1295 memset(grinfo->bb_counters, 0, 1296 sizeof(*grinfo->bb_counters) * 1297 (MB_NUM_ORDERS(sb))); 1298 /* 1299 * incore got set to the group block bitmap below 1300 */ 1301 ext4_lock_group(sb, group); 1302 /* init the buddy */ 1303 memset(data, 0xff, blocksize); 1304 ext4_mb_generate_buddy(sb, data, incore, group); 1305 ext4_unlock_group(sb, group); 1306 incore = NULL; 1307 } else { 1308 /* this is block of bitmap */ 1309 BUG_ON(incore != NULL); 1310 mb_debug(sb, "put bitmap for group %u in page %lu/%x\n", 1311 group, page->index, i * blocksize); 1312 trace_ext4_mb_bitmap_load(sb, group); 1313 1314 /* see comments in ext4_mb_put_pa() */ 1315 ext4_lock_group(sb, group); 1316 memcpy(data, bitmap, blocksize); 1317 1318 /* mark all preallocated blks used in in-core bitmap */ 1319 ext4_mb_generate_from_pa(sb, data, group); 1320 ext4_mb_generate_from_freelist(sb, data, group); 1321 ext4_unlock_group(sb, group); 1322 1323 /* set incore so that the buddy information can be 1324 * generated using this 1325 */ 1326 incore = data; 1327 } 1328 } 1329 SetPageUptodate(page); 1330 1331 out: 1332 if (bh) { 1333 for (i = 0; i < groups_per_page; i++) 1334 brelse(bh[i]); 1335 if (bh != &bhs) 1336 kfree(bh); 1337 } 1338 return err; 1339 } 1340 1341 /* 1342 * Lock the buddy and bitmap pages. This make sure other parallel init_group 1343 * on the same buddy page doesn't happen whild holding the buddy page lock. 1344 * Return locked buddy and bitmap pages on e4b struct. If buddy and bitmap 1345 * are on the same page e4b->bd_buddy_page is NULL and return value is 0. 1346 */ 1347 static int ext4_mb_get_buddy_page_lock(struct super_block *sb, 1348 ext4_group_t group, struct ext4_buddy *e4b, gfp_t gfp) 1349 { 1350 struct inode *inode = EXT4_SB(sb)->s_buddy_cache; 1351 int block, pnum, poff; 1352 int blocks_per_page; 1353 struct page *page; 1354 1355 e4b->bd_buddy_page = NULL; 1356 e4b->bd_bitmap_page = NULL; 1357 1358 blocks_per_page = PAGE_SIZE / sb->s_blocksize; 1359 /* 1360 * the buddy cache inode stores the block bitmap 1361 * and buddy information in consecutive blocks. 1362 * So for each group we need two blocks. 1363 */ 1364 block = group * 2; 1365 pnum = block / blocks_per_page; 1366 poff = block % blocks_per_page; 1367 page = find_or_create_page(inode->i_mapping, pnum, gfp); 1368 if (!page) 1369 return -ENOMEM; 1370 BUG_ON(page->mapping != inode->i_mapping); 1371 e4b->bd_bitmap_page = page; 1372 e4b->bd_bitmap = page_address(page) + (poff * sb->s_blocksize); 1373 1374 if (blocks_per_page >= 2) { 1375 /* buddy and bitmap are on the same page */ 1376 return 0; 1377 } 1378 1379 block++; 1380 pnum = block / blocks_per_page; 1381 page = find_or_create_page(inode->i_mapping, pnum, gfp); 1382 if (!page) 1383 return -ENOMEM; 1384 BUG_ON(page->mapping != inode->i_mapping); 1385 e4b->bd_buddy_page = page; 1386 return 0; 1387 } 1388 1389 static void ext4_mb_put_buddy_page_lock(struct ext4_buddy *e4b) 1390 { 1391 if (e4b->bd_bitmap_page) { 1392 unlock_page(e4b->bd_bitmap_page); 1393 put_page(e4b->bd_bitmap_page); 1394 } 1395 if (e4b->bd_buddy_page) { 1396 unlock_page(e4b->bd_buddy_page); 1397 put_page(e4b->bd_buddy_page); 1398 } 1399 } 1400 1401 /* 1402 * Locking note: This routine calls ext4_mb_init_cache(), which takes the 1403 * block group lock of all groups for this page; do not hold the BG lock when 1404 * calling this routine! 1405 */ 1406 static noinline_for_stack 1407 int ext4_mb_init_group(struct super_block *sb, ext4_group_t group, gfp_t gfp) 1408 { 1409 1410 struct ext4_group_info *this_grp; 1411 struct ext4_buddy e4b; 1412 struct page *page; 1413 int ret = 0; 1414 1415 might_sleep(); 1416 mb_debug(sb, "init group %u\n", group); 1417 this_grp = ext4_get_group_info(sb, group); 1418 /* 1419 * This ensures that we don't reinit the buddy cache 1420 * page which map to the group from which we are already 1421 * allocating. If we are looking at the buddy cache we would 1422 * have taken a reference using ext4_mb_load_buddy and that 1423 * would have pinned buddy page to page cache. 1424 * The call to ext4_mb_get_buddy_page_lock will mark the 1425 * page accessed. 1426 */ 1427 ret = ext4_mb_get_buddy_page_lock(sb, group, &e4b, gfp); 1428 if (ret || !EXT4_MB_GRP_NEED_INIT(this_grp)) { 1429 /* 1430 * somebody initialized the group 1431 * return without doing anything 1432 */ 1433 goto err; 1434 } 1435 1436 page = e4b.bd_bitmap_page; 1437 ret = ext4_mb_init_cache(page, NULL, gfp); 1438 if (ret) 1439 goto err; 1440 if (!PageUptodate(page)) { 1441 ret = -EIO; 1442 goto err; 1443 } 1444 1445 if (e4b.bd_buddy_page == NULL) { 1446 /* 1447 * If both the bitmap and buddy are in 1448 * the same page we don't need to force 1449 * init the buddy 1450 */ 1451 ret = 0; 1452 goto err; 1453 } 1454 /* init buddy cache */ 1455 page = e4b.bd_buddy_page; 1456 ret = ext4_mb_init_cache(page, e4b.bd_bitmap, gfp); 1457 if (ret) 1458 goto err; 1459 if (!PageUptodate(page)) { 1460 ret = -EIO; 1461 goto err; 1462 } 1463 err: 1464 ext4_mb_put_buddy_page_lock(&e4b); 1465 return ret; 1466 } 1467 1468 /* 1469 * Locking note: This routine calls ext4_mb_init_cache(), which takes the 1470 * block group lock of all groups for this page; do not hold the BG lock when 1471 * calling this routine! 1472 */ 1473 static noinline_for_stack int 1474 ext4_mb_load_buddy_gfp(struct super_block *sb, ext4_group_t group, 1475 struct ext4_buddy *e4b, gfp_t gfp) 1476 { 1477 int blocks_per_page; 1478 int block; 1479 int pnum; 1480 int poff; 1481 struct page *page; 1482 int ret; 1483 struct ext4_group_info *grp; 1484 struct ext4_sb_info *sbi = EXT4_SB(sb); 1485 struct inode *inode = sbi->s_buddy_cache; 1486 1487 might_sleep(); 1488 mb_debug(sb, "load group %u\n", group); 1489 1490 blocks_per_page = PAGE_SIZE / sb->s_blocksize; 1491 grp = ext4_get_group_info(sb, group); 1492 1493 e4b->bd_blkbits = sb->s_blocksize_bits; 1494 e4b->bd_info = grp; 1495 e4b->bd_sb = sb; 1496 e4b->bd_group = group; 1497 e4b->bd_buddy_page = NULL; 1498 e4b->bd_bitmap_page = NULL; 1499 1500 if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) { 1501 /* 1502 * we need full data about the group 1503 * to make a good selection 1504 */ 1505 ret = ext4_mb_init_group(sb, group, gfp); 1506 if (ret) 1507 return ret; 1508 } 1509 1510 /* 1511 * the buddy cache inode stores the block bitmap 1512 * and buddy information in consecutive blocks. 1513 * So for each group we need two blocks. 1514 */ 1515 block = group * 2; 1516 pnum = block / blocks_per_page; 1517 poff = block % blocks_per_page; 1518 1519 /* we could use find_or_create_page(), but it locks page 1520 * what we'd like to avoid in fast path ... */ 1521 page = find_get_page_flags(inode->i_mapping, pnum, FGP_ACCESSED); 1522 if (page == NULL || !PageUptodate(page)) { 1523 if (page) 1524 /* 1525 * drop the page reference and try 1526 * to get the page with lock. If we 1527 * are not uptodate that implies 1528 * somebody just created the page but 1529 * is yet to initialize the same. So 1530 * wait for it to initialize. 1531 */ 1532 put_page(page); 1533 page = find_or_create_page(inode->i_mapping, pnum, gfp); 1534 if (page) { 1535 BUG_ON(page->mapping != inode->i_mapping); 1536 if (!PageUptodate(page)) { 1537 ret = ext4_mb_init_cache(page, NULL, gfp); 1538 if (ret) { 1539 unlock_page(page); 1540 goto err; 1541 } 1542 mb_cmp_bitmaps(e4b, page_address(page) + 1543 (poff * sb->s_blocksize)); 1544 } 1545 unlock_page(page); 1546 } 1547 } 1548 if (page == NULL) { 1549 ret = -ENOMEM; 1550 goto err; 1551 } 1552 if (!PageUptodate(page)) { 1553 ret = -EIO; 1554 goto err; 1555 } 1556 1557 /* Pages marked accessed already */ 1558 e4b->bd_bitmap_page = page; 1559 e4b->bd_bitmap = page_address(page) + (poff * sb->s_blocksize); 1560 1561 block++; 1562 pnum = block / blocks_per_page; 1563 poff = block % blocks_per_page; 1564 1565 page = find_get_page_flags(inode->i_mapping, pnum, FGP_ACCESSED); 1566 if (page == NULL || !PageUptodate(page)) { 1567 if (page) 1568 put_page(page); 1569 page = find_or_create_page(inode->i_mapping, pnum, gfp); 1570 if (page) { 1571 BUG_ON(page->mapping != inode->i_mapping); 1572 if (!PageUptodate(page)) { 1573 ret = ext4_mb_init_cache(page, e4b->bd_bitmap, 1574 gfp); 1575 if (ret) { 1576 unlock_page(page); 1577 goto err; 1578 } 1579 } 1580 unlock_page(page); 1581 } 1582 } 1583 if (page == NULL) { 1584 ret = -ENOMEM; 1585 goto err; 1586 } 1587 if (!PageUptodate(page)) { 1588 ret = -EIO; 1589 goto err; 1590 } 1591 1592 /* Pages marked accessed already */ 1593 e4b->bd_buddy_page = page; 1594 e4b->bd_buddy = page_address(page) + (poff * sb->s_blocksize); 1595 1596 return 0; 1597 1598 err: 1599 if (page) 1600 put_page(page); 1601 if (e4b->bd_bitmap_page) 1602 put_page(e4b->bd_bitmap_page); 1603 if (e4b->bd_buddy_page) 1604 put_page(e4b->bd_buddy_page); 1605 e4b->bd_buddy = NULL; 1606 e4b->bd_bitmap = NULL; 1607 return ret; 1608 } 1609 1610 static int ext4_mb_load_buddy(struct super_block *sb, ext4_group_t group, 1611 struct ext4_buddy *e4b) 1612 { 1613 return ext4_mb_load_buddy_gfp(sb, group, e4b, GFP_NOFS); 1614 } 1615 1616 static void ext4_mb_unload_buddy(struct ext4_buddy *e4b) 1617 { 1618 if (e4b->bd_bitmap_page) 1619 put_page(e4b->bd_bitmap_page); 1620 if (e4b->bd_buddy_page) 1621 put_page(e4b->bd_buddy_page); 1622 } 1623 1624 1625 static int mb_find_order_for_block(struct ext4_buddy *e4b, int block) 1626 { 1627 int order = 1, max; 1628 void *bb; 1629 1630 BUG_ON(e4b->bd_bitmap == e4b->bd_buddy); 1631 BUG_ON(block >= (1 << (e4b->bd_blkbits + 3))); 1632 1633 while (order <= e4b->bd_blkbits + 1) { 1634 bb = mb_find_buddy(e4b, order, &max); 1635 if (!mb_test_bit(block >> order, bb)) { 1636 /* this block is part of buddy of order 'order' */ 1637 return order; 1638 } 1639 order++; 1640 } 1641 return 0; 1642 } 1643 1644 static void mb_clear_bits(void *bm, int cur, int len) 1645 { 1646 __u32 *addr; 1647 1648 len = cur + len; 1649 while (cur < len) { 1650 if ((cur & 31) == 0 && (len - cur) >= 32) { 1651 /* fast path: clear whole word at once */ 1652 addr = bm + (cur >> 3); 1653 *addr = 0; 1654 cur += 32; 1655 continue; 1656 } 1657 mb_clear_bit(cur, bm); 1658 cur++; 1659 } 1660 } 1661 1662 /* clear bits in given range 1663 * will return first found zero bit if any, -1 otherwise 1664 */ 1665 static int mb_test_and_clear_bits(void *bm, int cur, int len) 1666 { 1667 __u32 *addr; 1668 int zero_bit = -1; 1669 1670 len = cur + len; 1671 while (cur < len) { 1672 if ((cur & 31) == 0 && (len - cur) >= 32) { 1673 /* fast path: clear whole word at once */ 1674 addr = bm + (cur >> 3); 1675 if (*addr != (__u32)(-1) && zero_bit == -1) 1676 zero_bit = cur + mb_find_next_zero_bit(addr, 32, 0); 1677 *addr = 0; 1678 cur += 32; 1679 continue; 1680 } 1681 if (!mb_test_and_clear_bit(cur, bm) && zero_bit == -1) 1682 zero_bit = cur; 1683 cur++; 1684 } 1685 1686 return zero_bit; 1687 } 1688 1689 void mb_set_bits(void *bm, int cur, int len) 1690 { 1691 __u32 *addr; 1692 1693 len = cur + len; 1694 while (cur < len) { 1695 if ((cur & 31) == 0 && (len - cur) >= 32) { 1696 /* fast path: set whole word at once */ 1697 addr = bm + (cur >> 3); 1698 *addr = 0xffffffff; 1699 cur += 32; 1700 continue; 1701 } 1702 mb_set_bit(cur, bm); 1703 cur++; 1704 } 1705 } 1706 1707 static inline int mb_buddy_adjust_border(int* bit, void* bitmap, int side) 1708 { 1709 if (mb_test_bit(*bit + side, bitmap)) { 1710 mb_clear_bit(*bit, bitmap); 1711 (*bit) -= side; 1712 return 1; 1713 } 1714 else { 1715 (*bit) += side; 1716 mb_set_bit(*bit, bitmap); 1717 return -1; 1718 } 1719 } 1720 1721 static void mb_buddy_mark_free(struct ext4_buddy *e4b, int first, int last) 1722 { 1723 int max; 1724 int order = 1; 1725 void *buddy = mb_find_buddy(e4b, order, &max); 1726 1727 while (buddy) { 1728 void *buddy2; 1729 1730 /* Bits in range [first; last] are known to be set since 1731 * corresponding blocks were allocated. Bits in range 1732 * (first; last) will stay set because they form buddies on 1733 * upper layer. We just deal with borders if they don't 1734 * align with upper layer and then go up. 1735 * Releasing entire group is all about clearing 1736 * single bit of highest order buddy. 1737 */ 1738 1739 /* Example: 1740 * --------------------------------- 1741 * | 1 | 1 | 1 | 1 | 1742 * --------------------------------- 1743 * | 0 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1744 * --------------------------------- 1745 * 0 1 2 3 4 5 6 7 1746 * \_____________________/ 1747 * 1748 * Neither [1] nor [6] is aligned to above layer. 1749 * Left neighbour [0] is free, so mark it busy, 1750 * decrease bb_counters and extend range to 1751 * [0; 6] 1752 * Right neighbour [7] is busy. It can't be coaleasced with [6], so 1753 * mark [6] free, increase bb_counters and shrink range to 1754 * [0; 5]. 1755 * Then shift range to [0; 2], go up and do the same. 1756 */ 1757 1758 1759 if (first & 1) 1760 e4b->bd_info->bb_counters[order] += mb_buddy_adjust_border(&first, buddy, -1); 1761 if (!(last & 1)) 1762 e4b->bd_info->bb_counters[order] += mb_buddy_adjust_border(&last, buddy, 1); 1763 if (first > last) 1764 break; 1765 order++; 1766 1767 if (first == last || !(buddy2 = mb_find_buddy(e4b, order, &max))) { 1768 mb_clear_bits(buddy, first, last - first + 1); 1769 e4b->bd_info->bb_counters[order - 1] += last - first + 1; 1770 break; 1771 } 1772 first >>= 1; 1773 last >>= 1; 1774 buddy = buddy2; 1775 } 1776 } 1777 1778 static void mb_free_blocks(struct inode *inode, struct ext4_buddy *e4b, 1779 int first, int count) 1780 { 1781 int left_is_free = 0; 1782 int right_is_free = 0; 1783 int block; 1784 int last = first + count - 1; 1785 struct super_block *sb = e4b->bd_sb; 1786 1787 if (WARN_ON(count == 0)) 1788 return; 1789 BUG_ON(last >= (sb->s_blocksize << 3)); 1790 assert_spin_locked(ext4_group_lock_ptr(sb, e4b->bd_group)); 1791 /* Don't bother if the block group is corrupt. */ 1792 if (unlikely(EXT4_MB_GRP_BBITMAP_CORRUPT(e4b->bd_info))) 1793 return; 1794 1795 mb_check_buddy(e4b); 1796 mb_free_blocks_double(inode, e4b, first, count); 1797 1798 this_cpu_inc(discard_pa_seq); 1799 e4b->bd_info->bb_free += count; 1800 if (first < e4b->bd_info->bb_first_free) 1801 e4b->bd_info->bb_first_free = first; 1802 1803 /* access memory sequentially: check left neighbour, 1804 * clear range and then check right neighbour 1805 */ 1806 if (first != 0) 1807 left_is_free = !mb_test_bit(first - 1, e4b->bd_bitmap); 1808 block = mb_test_and_clear_bits(e4b->bd_bitmap, first, count); 1809 if (last + 1 < EXT4_SB(sb)->s_mb_maxs[0]) 1810 right_is_free = !mb_test_bit(last + 1, e4b->bd_bitmap); 1811 1812 if (unlikely(block != -1)) { 1813 struct ext4_sb_info *sbi = EXT4_SB(sb); 1814 ext4_fsblk_t blocknr; 1815 1816 blocknr = ext4_group_first_block_no(sb, e4b->bd_group); 1817 blocknr += EXT4_C2B(sbi, block); 1818 if (!(sbi->s_mount_state & EXT4_FC_REPLAY)) { 1819 ext4_grp_locked_error(sb, e4b->bd_group, 1820 inode ? inode->i_ino : 0, 1821 blocknr, 1822 "freeing already freed block (bit %u); block bitmap corrupt.", 1823 block); 1824 ext4_mark_group_bitmap_corrupted( 1825 sb, e4b->bd_group, 1826 EXT4_GROUP_INFO_BBITMAP_CORRUPT); 1827 } 1828 goto done; 1829 } 1830 1831 /* let's maintain fragments counter */ 1832 if (left_is_free && right_is_free) 1833 e4b->bd_info->bb_fragments--; 1834 else if (!left_is_free && !right_is_free) 1835 e4b->bd_info->bb_fragments++; 1836 1837 /* buddy[0] == bd_bitmap is a special case, so handle 1838 * it right away and let mb_buddy_mark_free stay free of 1839 * zero order checks. 1840 * Check if neighbours are to be coaleasced, 1841 * adjust bitmap bb_counters and borders appropriately. 1842 */ 1843 if (first & 1) { 1844 first += !left_is_free; 1845 e4b->bd_info->bb_counters[0] += left_is_free ? -1 : 1; 1846 } 1847 if (!(last & 1)) { 1848 last -= !right_is_free; 1849 e4b->bd_info->bb_counters[0] += right_is_free ? -1 : 1; 1850 } 1851 1852 if (first <= last) 1853 mb_buddy_mark_free(e4b, first >> 1, last >> 1); 1854 1855 done: 1856 mb_set_largest_free_order(sb, e4b->bd_info); 1857 mb_update_avg_fragment_size(sb, e4b->bd_info); 1858 mb_check_buddy(e4b); 1859 } 1860 1861 static int mb_find_extent(struct ext4_buddy *e4b, int block, 1862 int needed, struct ext4_free_extent *ex) 1863 { 1864 int next = block; 1865 int max, order; 1866 void *buddy; 1867 1868 assert_spin_locked(ext4_group_lock_ptr(e4b->bd_sb, e4b->bd_group)); 1869 BUG_ON(ex == NULL); 1870 1871 buddy = mb_find_buddy(e4b, 0, &max); 1872 BUG_ON(buddy == NULL); 1873 BUG_ON(block >= max); 1874 if (mb_test_bit(block, buddy)) { 1875 ex->fe_len = 0; 1876 ex->fe_start = 0; 1877 ex->fe_group = 0; 1878 return 0; 1879 } 1880 1881 /* find actual order */ 1882 order = mb_find_order_for_block(e4b, block); 1883 block = block >> order; 1884 1885 ex->fe_len = 1 << order; 1886 ex->fe_start = block << order; 1887 ex->fe_group = e4b->bd_group; 1888 1889 /* calc difference from given start */ 1890 next = next - ex->fe_start; 1891 ex->fe_len -= next; 1892 ex->fe_start += next; 1893 1894 while (needed > ex->fe_len && 1895 mb_find_buddy(e4b, order, &max)) { 1896 1897 if (block + 1 >= max) 1898 break; 1899 1900 next = (block + 1) * (1 << order); 1901 if (mb_test_bit(next, e4b->bd_bitmap)) 1902 break; 1903 1904 order = mb_find_order_for_block(e4b, next); 1905 1906 block = next >> order; 1907 ex->fe_len += 1 << order; 1908 } 1909 1910 if (ex->fe_start + ex->fe_len > EXT4_CLUSTERS_PER_GROUP(e4b->bd_sb)) { 1911 /* Should never happen! (but apparently sometimes does?!?) */ 1912 WARN_ON(1); 1913 ext4_grp_locked_error(e4b->bd_sb, e4b->bd_group, 0, 0, 1914 "corruption or bug in mb_find_extent " 1915 "block=%d, order=%d needed=%d ex=%u/%d/%d@%u", 1916 block, order, needed, ex->fe_group, ex->fe_start, 1917 ex->fe_len, ex->fe_logical); 1918 ex->fe_len = 0; 1919 ex->fe_start = 0; 1920 ex->fe_group = 0; 1921 } 1922 return ex->fe_len; 1923 } 1924 1925 static int mb_mark_used(struct ext4_buddy *e4b, struct ext4_free_extent *ex) 1926 { 1927 int ord; 1928 int mlen = 0; 1929 int max = 0; 1930 int cur; 1931 int start = ex->fe_start; 1932 int len = ex->fe_len; 1933 unsigned ret = 0; 1934 int len0 = len; 1935 void *buddy; 1936 1937 BUG_ON(start + len > (e4b->bd_sb->s_blocksize << 3)); 1938 BUG_ON(e4b->bd_group != ex->fe_group); 1939 assert_spin_locked(ext4_group_lock_ptr(e4b->bd_sb, e4b->bd_group)); 1940 mb_check_buddy(e4b); 1941 mb_mark_used_double(e4b, start, len); 1942 1943 this_cpu_inc(discard_pa_seq); 1944 e4b->bd_info->bb_free -= len; 1945 if (e4b->bd_info->bb_first_free == start) 1946 e4b->bd_info->bb_first_free += len; 1947 1948 /* let's maintain fragments counter */ 1949 if (start != 0) 1950 mlen = !mb_test_bit(start - 1, e4b->bd_bitmap); 1951 if (start + len < EXT4_SB(e4b->bd_sb)->s_mb_maxs[0]) 1952 max = !mb_test_bit(start + len, e4b->bd_bitmap); 1953 if (mlen && max) 1954 e4b->bd_info->bb_fragments++; 1955 else if (!mlen && !max) 1956 e4b->bd_info->bb_fragments--; 1957 1958 /* let's maintain buddy itself */ 1959 while (len) { 1960 ord = mb_find_order_for_block(e4b, start); 1961 1962 if (((start >> ord) << ord) == start && len >= (1 << ord)) { 1963 /* the whole chunk may be allocated at once! */ 1964 mlen = 1 << ord; 1965 buddy = mb_find_buddy(e4b, ord, &max); 1966 BUG_ON((start >> ord) >= max); 1967 mb_set_bit(start >> ord, buddy); 1968 e4b->bd_info->bb_counters[ord]--; 1969 start += mlen; 1970 len -= mlen; 1971 BUG_ON(len < 0); 1972 continue; 1973 } 1974 1975 /* store for history */ 1976 if (ret == 0) 1977 ret = len | (ord << 16); 1978 1979 /* we have to split large buddy */ 1980 BUG_ON(ord <= 0); 1981 buddy = mb_find_buddy(e4b, ord, &max); 1982 mb_set_bit(start >> ord, buddy); 1983 e4b->bd_info->bb_counters[ord]--; 1984 1985 ord--; 1986 cur = (start >> ord) & ~1U; 1987 buddy = mb_find_buddy(e4b, ord, &max); 1988 mb_clear_bit(cur, buddy); 1989 mb_clear_bit(cur + 1, buddy); 1990 e4b->bd_info->bb_counters[ord]++; 1991 e4b->bd_info->bb_counters[ord]++; 1992 } 1993 mb_set_largest_free_order(e4b->bd_sb, e4b->bd_info); 1994 1995 mb_update_avg_fragment_size(e4b->bd_sb, e4b->bd_info); 1996 mb_set_bits(e4b->bd_bitmap, ex->fe_start, len0); 1997 mb_check_buddy(e4b); 1998 1999 return ret; 2000 } 2001 2002 /* 2003 * Must be called under group lock! 2004 */ 2005 static void ext4_mb_use_best_found(struct ext4_allocation_context *ac, 2006 struct ext4_buddy *e4b) 2007 { 2008 struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); 2009 int ret; 2010 2011 BUG_ON(ac->ac_b_ex.fe_group != e4b->bd_group); 2012 BUG_ON(ac->ac_status == AC_STATUS_FOUND); 2013 2014 ac->ac_b_ex.fe_len = min(ac->ac_b_ex.fe_len, ac->ac_g_ex.fe_len); 2015 ac->ac_b_ex.fe_logical = ac->ac_g_ex.fe_logical; 2016 ret = mb_mark_used(e4b, &ac->ac_b_ex); 2017 2018 /* preallocation can change ac_b_ex, thus we store actually 2019 * allocated blocks for history */ 2020 ac->ac_f_ex = ac->ac_b_ex; 2021 2022 ac->ac_status = AC_STATUS_FOUND; 2023 ac->ac_tail = ret & 0xffff; 2024 ac->ac_buddy = ret >> 16; 2025 2026 /* 2027 * take the page reference. We want the page to be pinned 2028 * so that we don't get a ext4_mb_init_cache_call for this 2029 * group until we update the bitmap. That would mean we 2030 * double allocate blocks. The reference is dropped 2031 * in ext4_mb_release_context 2032 */ 2033 ac->ac_bitmap_page = e4b->bd_bitmap_page; 2034 get_page(ac->ac_bitmap_page); 2035 ac->ac_buddy_page = e4b->bd_buddy_page; 2036 get_page(ac->ac_buddy_page); 2037 /* store last allocated for subsequent stream allocation */ 2038 if (ac->ac_flags & EXT4_MB_STREAM_ALLOC) { 2039 spin_lock(&sbi->s_md_lock); 2040 sbi->s_mb_last_group = ac->ac_f_ex.fe_group; 2041 sbi->s_mb_last_start = ac->ac_f_ex.fe_start; 2042 spin_unlock(&sbi->s_md_lock); 2043 } 2044 /* 2045 * As we've just preallocated more space than 2046 * user requested originally, we store allocated 2047 * space in a special descriptor. 2048 */ 2049 if (ac->ac_o_ex.fe_len < ac->ac_b_ex.fe_len) 2050 ext4_mb_new_preallocation(ac); 2051 2052 } 2053 2054 static void ext4_mb_check_limits(struct ext4_allocation_context *ac, 2055 struct ext4_buddy *e4b, 2056 int finish_group) 2057 { 2058 struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); 2059 struct ext4_free_extent *bex = &ac->ac_b_ex; 2060 struct ext4_free_extent *gex = &ac->ac_g_ex; 2061 struct ext4_free_extent ex; 2062 int max; 2063 2064 if (ac->ac_status == AC_STATUS_FOUND) 2065 return; 2066 /* 2067 * We don't want to scan for a whole year 2068 */ 2069 if (ac->ac_found > sbi->s_mb_max_to_scan && 2070 !(ac->ac_flags & EXT4_MB_HINT_FIRST)) { 2071 ac->ac_status = AC_STATUS_BREAK; 2072 return; 2073 } 2074 2075 /* 2076 * Haven't found good chunk so far, let's continue 2077 */ 2078 if (bex->fe_len < gex->fe_len) 2079 return; 2080 2081 if ((finish_group || ac->ac_found > sbi->s_mb_min_to_scan) 2082 && bex->fe_group == e4b->bd_group) { 2083 /* recheck chunk's availability - we don't know 2084 * when it was found (within this lock-unlock 2085 * period or not) */ 2086 max = mb_find_extent(e4b, bex->fe_start, gex->fe_len, &ex); 2087 if (max >= gex->fe_len) { 2088 ext4_mb_use_best_found(ac, e4b); 2089 return; 2090 } 2091 } 2092 } 2093 2094 /* 2095 * The routine checks whether found extent is good enough. If it is, 2096 * then the extent gets marked used and flag is set to the context 2097 * to stop scanning. Otherwise, the extent is compared with the 2098 * previous found extent and if new one is better, then it's stored 2099 * in the context. Later, the best found extent will be used, if 2100 * mballoc can't find good enough extent. 2101 * 2102 * FIXME: real allocation policy is to be designed yet! 2103 */ 2104 static void ext4_mb_measure_extent(struct ext4_allocation_context *ac, 2105 struct ext4_free_extent *ex, 2106 struct ext4_buddy *e4b) 2107 { 2108 struct ext4_free_extent *bex = &ac->ac_b_ex; 2109 struct ext4_free_extent *gex = &ac->ac_g_ex; 2110 2111 BUG_ON(ex->fe_len <= 0); 2112 BUG_ON(ex->fe_len > EXT4_CLUSTERS_PER_GROUP(ac->ac_sb)); 2113 BUG_ON(ex->fe_start >= EXT4_CLUSTERS_PER_GROUP(ac->ac_sb)); 2114 BUG_ON(ac->ac_status != AC_STATUS_CONTINUE); 2115 2116 ac->ac_found++; 2117 2118 /* 2119 * The special case - take what you catch first 2120 */ 2121 if (unlikely(ac->ac_flags & EXT4_MB_HINT_FIRST)) { 2122 *bex = *ex; 2123 ext4_mb_use_best_found(ac, e4b); 2124 return; 2125 } 2126 2127 /* 2128 * Let's check whether the chuck is good enough 2129 */ 2130 if (ex->fe_len == gex->fe_len) { 2131 *bex = *ex; 2132 ext4_mb_use_best_found(ac, e4b); 2133 return; 2134 } 2135 2136 /* 2137 * If this is first found extent, just store it in the context 2138 */ 2139 if (bex->fe_len == 0) { 2140 *bex = *ex; 2141 return; 2142 } 2143 2144 /* 2145 * If new found extent is better, store it in the context 2146 */ 2147 if (bex->fe_len < gex->fe_len) { 2148 /* if the request isn't satisfied, any found extent 2149 * larger than previous best one is better */ 2150 if (ex->fe_len > bex->fe_len) 2151 *bex = *ex; 2152 } else if (ex->fe_len > gex->fe_len) { 2153 /* if the request is satisfied, then we try to find 2154 * an extent that still satisfy the request, but is 2155 * smaller than previous one */ 2156 if (ex->fe_len < bex->fe_len) 2157 *bex = *ex; 2158 } 2159 2160 ext4_mb_check_limits(ac, e4b, 0); 2161 } 2162 2163 static noinline_for_stack 2164 int ext4_mb_try_best_found(struct ext4_allocation_context *ac, 2165 struct ext4_buddy *e4b) 2166 { 2167 struct ext4_free_extent ex = ac->ac_b_ex; 2168 ext4_group_t group = ex.fe_group; 2169 int max; 2170 int err; 2171 2172 BUG_ON(ex.fe_len <= 0); 2173 err = ext4_mb_load_buddy(ac->ac_sb, group, e4b); 2174 if (err) 2175 return err; 2176 2177 ext4_lock_group(ac->ac_sb, group); 2178 max = mb_find_extent(e4b, ex.fe_start, ex.fe_len, &ex); 2179 2180 if (max > 0) { 2181 ac->ac_b_ex = ex; 2182 ext4_mb_use_best_found(ac, e4b); 2183 } 2184 2185 ext4_unlock_group(ac->ac_sb, group); 2186 ext4_mb_unload_buddy(e4b); 2187 2188 return 0; 2189 } 2190 2191 static noinline_for_stack 2192 int ext4_mb_find_by_goal(struct ext4_allocation_context *ac, 2193 struct ext4_buddy *e4b) 2194 { 2195 ext4_group_t group = ac->ac_g_ex.fe_group; 2196 int max; 2197 int err; 2198 struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); 2199 struct ext4_group_info *grp = ext4_get_group_info(ac->ac_sb, group); 2200 struct ext4_free_extent ex; 2201 2202 if (!(ac->ac_flags & EXT4_MB_HINT_TRY_GOAL)) 2203 return 0; 2204 if (grp->bb_free == 0) 2205 return 0; 2206 2207 err = ext4_mb_load_buddy(ac->ac_sb, group, e4b); 2208 if (err) 2209 return err; 2210 2211 if (unlikely(EXT4_MB_GRP_BBITMAP_CORRUPT(e4b->bd_info))) { 2212 ext4_mb_unload_buddy(e4b); 2213 return 0; 2214 } 2215 2216 ext4_lock_group(ac->ac_sb, group); 2217 max = mb_find_extent(e4b, ac->ac_g_ex.fe_start, 2218 ac->ac_g_ex.fe_len, &ex); 2219 ex.fe_logical = 0xDEADFA11; /* debug value */ 2220 2221 if (max >= ac->ac_g_ex.fe_len && ac->ac_g_ex.fe_len == sbi->s_stripe) { 2222 ext4_fsblk_t start; 2223 2224 start = ext4_group_first_block_no(ac->ac_sb, e4b->bd_group) + 2225 ex.fe_start; 2226 /* use do_div to get remainder (would be 64-bit modulo) */ 2227 if (do_div(start, sbi->s_stripe) == 0) { 2228 ac->ac_found++; 2229 ac->ac_b_ex = ex; 2230 ext4_mb_use_best_found(ac, e4b); 2231 } 2232 } else if (max >= ac->ac_g_ex.fe_len) { 2233 BUG_ON(ex.fe_len <= 0); 2234 BUG_ON(ex.fe_group != ac->ac_g_ex.fe_group); 2235 BUG_ON(ex.fe_start != ac->ac_g_ex.fe_start); 2236 ac->ac_found++; 2237 ac->ac_b_ex = ex; 2238 ext4_mb_use_best_found(ac, e4b); 2239 } else if (max > 0 && (ac->ac_flags & EXT4_MB_HINT_MERGE)) { 2240 /* Sometimes, caller may want to merge even small 2241 * number of blocks to an existing extent */ 2242 BUG_ON(ex.fe_len <= 0); 2243 BUG_ON(ex.fe_group != ac->ac_g_ex.fe_group); 2244 BUG_ON(ex.fe_start != ac->ac_g_ex.fe_start); 2245 ac->ac_found++; 2246 ac->ac_b_ex = ex; 2247 ext4_mb_use_best_found(ac, e4b); 2248 } 2249 ext4_unlock_group(ac->ac_sb, group); 2250 ext4_mb_unload_buddy(e4b); 2251 2252 return 0; 2253 } 2254 2255 /* 2256 * The routine scans buddy structures (not bitmap!) from given order 2257 * to max order and tries to find big enough chunk to satisfy the req 2258 */ 2259 static noinline_for_stack 2260 void ext4_mb_simple_scan_group(struct ext4_allocation_context *ac, 2261 struct ext4_buddy *e4b) 2262 { 2263 struct super_block *sb = ac->ac_sb; 2264 struct ext4_group_info *grp = e4b->bd_info; 2265 void *buddy; 2266 int i; 2267 int k; 2268 int max; 2269 2270 BUG_ON(ac->ac_2order <= 0); 2271 for (i = ac->ac_2order; i < MB_NUM_ORDERS(sb); i++) { 2272 if (grp->bb_counters[i] == 0) 2273 continue; 2274 2275 buddy = mb_find_buddy(e4b, i, &max); 2276 BUG_ON(buddy == NULL); 2277 2278 k = mb_find_next_zero_bit(buddy, max, 0); 2279 if (k >= max) { 2280 ext4_grp_locked_error(ac->ac_sb, e4b->bd_group, 0, 0, 2281 "%d free clusters of order %d. But found 0", 2282 grp->bb_counters[i], i); 2283 ext4_mark_group_bitmap_corrupted(ac->ac_sb, 2284 e4b->bd_group, 2285 EXT4_GROUP_INFO_BBITMAP_CORRUPT); 2286 break; 2287 } 2288 ac->ac_found++; 2289 2290 ac->ac_b_ex.fe_len = 1 << i; 2291 ac->ac_b_ex.fe_start = k << i; 2292 ac->ac_b_ex.fe_group = e4b->bd_group; 2293 2294 ext4_mb_use_best_found(ac, e4b); 2295 2296 BUG_ON(ac->ac_f_ex.fe_len != ac->ac_g_ex.fe_len); 2297 2298 if (EXT4_SB(sb)->s_mb_stats) 2299 atomic_inc(&EXT4_SB(sb)->s_bal_2orders); 2300 2301 break; 2302 } 2303 } 2304 2305 /* 2306 * The routine scans the group and measures all found extents. 2307 * In order to optimize scanning, caller must pass number of 2308 * free blocks in the group, so the routine can know upper limit. 2309 */ 2310 static noinline_for_stack 2311 void ext4_mb_complex_scan_group(struct ext4_allocation_context *ac, 2312 struct ext4_buddy *e4b) 2313 { 2314 struct super_block *sb = ac->ac_sb; 2315 void *bitmap = e4b->bd_bitmap; 2316 struct ext4_free_extent ex; 2317 int i; 2318 int free; 2319 2320 free = e4b->bd_info->bb_free; 2321 if (WARN_ON(free <= 0)) 2322 return; 2323 2324 i = e4b->bd_info->bb_first_free; 2325 2326 while (free && ac->ac_status == AC_STATUS_CONTINUE) { 2327 i = mb_find_next_zero_bit(bitmap, 2328 EXT4_CLUSTERS_PER_GROUP(sb), i); 2329 if (i >= EXT4_CLUSTERS_PER_GROUP(sb)) { 2330 /* 2331 * IF we have corrupt bitmap, we won't find any 2332 * free blocks even though group info says we 2333 * have free blocks 2334 */ 2335 ext4_grp_locked_error(sb, e4b->bd_group, 0, 0, 2336 "%d free clusters as per " 2337 "group info. But bitmap says 0", 2338 free); 2339 ext4_mark_group_bitmap_corrupted(sb, e4b->bd_group, 2340 EXT4_GROUP_INFO_BBITMAP_CORRUPT); 2341 break; 2342 } 2343 2344 mb_find_extent(e4b, i, ac->ac_g_ex.fe_len, &ex); 2345 if (WARN_ON(ex.fe_len <= 0)) 2346 break; 2347 if (free < ex.fe_len) { 2348 ext4_grp_locked_error(sb, e4b->bd_group, 0, 0, 2349 "%d free clusters as per " 2350 "group info. But got %d blocks", 2351 free, ex.fe_len); 2352 ext4_mark_group_bitmap_corrupted(sb, e4b->bd_group, 2353 EXT4_GROUP_INFO_BBITMAP_CORRUPT); 2354 /* 2355 * The number of free blocks differs. This mostly 2356 * indicate that the bitmap is corrupt. So exit 2357 * without claiming the space. 2358 */ 2359 break; 2360 } 2361 ex.fe_logical = 0xDEADC0DE; /* debug value */ 2362 ext4_mb_measure_extent(ac, &ex, e4b); 2363 2364 i += ex.fe_len; 2365 free -= ex.fe_len; 2366 } 2367 2368 ext4_mb_check_limits(ac, e4b, 1); 2369 } 2370 2371 /* 2372 * This is a special case for storages like raid5 2373 * we try to find stripe-aligned chunks for stripe-size-multiple requests 2374 */ 2375 static noinline_for_stack 2376 void ext4_mb_scan_aligned(struct ext4_allocation_context *ac, 2377 struct ext4_buddy *e4b) 2378 { 2379 struct super_block *sb = ac->ac_sb; 2380 struct ext4_sb_info *sbi = EXT4_SB(sb); 2381 void *bitmap = e4b->bd_bitmap; 2382 struct ext4_free_extent ex; 2383 ext4_fsblk_t first_group_block; 2384 ext4_fsblk_t a; 2385 ext4_grpblk_t i; 2386 int max; 2387 2388 BUG_ON(sbi->s_stripe == 0); 2389 2390 /* find first stripe-aligned block in group */ 2391 first_group_block = ext4_group_first_block_no(sb, e4b->bd_group); 2392 2393 a = first_group_block + sbi->s_stripe - 1; 2394 do_div(a, sbi->s_stripe); 2395 i = (a * sbi->s_stripe) - first_group_block; 2396 2397 while (i < EXT4_CLUSTERS_PER_GROUP(sb)) { 2398 if (!mb_test_bit(i, bitmap)) { 2399 max = mb_find_extent(e4b, i, sbi->s_stripe, &ex); 2400 if (max >= sbi->s_stripe) { 2401 ac->ac_found++; 2402 ex.fe_logical = 0xDEADF00D; /* debug value */ 2403 ac->ac_b_ex = ex; 2404 ext4_mb_use_best_found(ac, e4b); 2405 break; 2406 } 2407 } 2408 i += sbi->s_stripe; 2409 } 2410 } 2411 2412 /* 2413 * This is also called BEFORE we load the buddy bitmap. 2414 * Returns either 1 or 0 indicating that the group is either suitable 2415 * for the allocation or not. 2416 */ 2417 static bool ext4_mb_good_group(struct ext4_allocation_context *ac, 2418 ext4_group_t group, int cr) 2419 { 2420 ext4_grpblk_t free, fragments; 2421 int flex_size = ext4_flex_bg_size(EXT4_SB(ac->ac_sb)); 2422 struct ext4_group_info *grp = ext4_get_group_info(ac->ac_sb, group); 2423 2424 BUG_ON(cr < 0 || cr >= 4); 2425 2426 if (unlikely(EXT4_MB_GRP_BBITMAP_CORRUPT(grp))) 2427 return false; 2428 2429 free = grp->bb_free; 2430 if (free == 0) 2431 return false; 2432 2433 fragments = grp->bb_fragments; 2434 if (fragments == 0) 2435 return false; 2436 2437 switch (cr) { 2438 case 0: 2439 BUG_ON(ac->ac_2order == 0); 2440 2441 /* Avoid using the first bg of a flexgroup for data files */ 2442 if ((ac->ac_flags & EXT4_MB_HINT_DATA) && 2443 (flex_size >= EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME) && 2444 ((group % flex_size) == 0)) 2445 return false; 2446 2447 if (free < ac->ac_g_ex.fe_len) 2448 return false; 2449 2450 if (ac->ac_2order >= MB_NUM_ORDERS(ac->ac_sb)) 2451 return true; 2452 2453 if (grp->bb_largest_free_order < ac->ac_2order) 2454 return false; 2455 2456 return true; 2457 case 1: 2458 if ((free / fragments) >= ac->ac_g_ex.fe_len) 2459 return true; 2460 break; 2461 case 2: 2462 if (free >= ac->ac_g_ex.fe_len) 2463 return true; 2464 break; 2465 case 3: 2466 return true; 2467 default: 2468 BUG(); 2469 } 2470 2471 return false; 2472 } 2473 2474 /* 2475 * This could return negative error code if something goes wrong 2476 * during ext4_mb_init_group(). This should not be called with 2477 * ext4_lock_group() held. 2478 * 2479 * Note: because we are conditionally operating with the group lock in 2480 * the EXT4_MB_STRICT_CHECK case, we need to fake out sparse in this 2481 * function using __acquire and __release. This means we need to be 2482 * super careful before messing with the error path handling via "goto 2483 * out"! 2484 */ 2485 static int ext4_mb_good_group_nolock(struct ext4_allocation_context *ac, 2486 ext4_group_t group, int cr) 2487 { 2488 struct ext4_group_info *grp = ext4_get_group_info(ac->ac_sb, group); 2489 struct super_block *sb = ac->ac_sb; 2490 struct ext4_sb_info *sbi = EXT4_SB(sb); 2491 bool should_lock = ac->ac_flags & EXT4_MB_STRICT_CHECK; 2492 ext4_grpblk_t free; 2493 int ret = 0; 2494 2495 if (sbi->s_mb_stats) 2496 atomic64_inc(&sbi->s_bal_cX_groups_considered[ac->ac_criteria]); 2497 if (should_lock) { 2498 ext4_lock_group(sb, group); 2499 __release(ext4_group_lock_ptr(sb, group)); 2500 } 2501 free = grp->bb_free; 2502 if (free == 0) 2503 goto out; 2504 if (cr <= 2 && free < ac->ac_g_ex.fe_len) 2505 goto out; 2506 if (unlikely(EXT4_MB_GRP_BBITMAP_CORRUPT(grp))) 2507 goto out; 2508 if (should_lock) { 2509 __acquire(ext4_group_lock_ptr(sb, group)); 2510 ext4_unlock_group(sb, group); 2511 } 2512 2513 /* We only do this if the grp has never been initialized */ 2514 if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) { 2515 struct ext4_group_desc *gdp = 2516 ext4_get_group_desc(sb, group, NULL); 2517 int ret; 2518 2519 /* cr=0/1 is a very optimistic search to find large 2520 * good chunks almost for free. If buddy data is not 2521 * ready, then this optimization makes no sense. But 2522 * we never skip the first block group in a flex_bg, 2523 * since this gets used for metadata block allocation, 2524 * and we want to make sure we locate metadata blocks 2525 * in the first block group in the flex_bg if possible. 2526 */ 2527 if (cr < 2 && 2528 (!sbi->s_log_groups_per_flex || 2529 ((group & ((1 << sbi->s_log_groups_per_flex) - 1)) != 0)) && 2530 !(ext4_has_group_desc_csum(sb) && 2531 (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)))) 2532 return 0; 2533 ret = ext4_mb_init_group(sb, group, GFP_NOFS); 2534 if (ret) 2535 return ret; 2536 } 2537 2538 if (should_lock) { 2539 ext4_lock_group(sb, group); 2540 __release(ext4_group_lock_ptr(sb, group)); 2541 } 2542 ret = ext4_mb_good_group(ac, group, cr); 2543 out: 2544 if (should_lock) { 2545 __acquire(ext4_group_lock_ptr(sb, group)); 2546 ext4_unlock_group(sb, group); 2547 } 2548 return ret; 2549 } 2550 2551 /* 2552 * Start prefetching @nr block bitmaps starting at @group. 2553 * Return the next group which needs to be prefetched. 2554 */ 2555 ext4_group_t ext4_mb_prefetch(struct super_block *sb, ext4_group_t group, 2556 unsigned int nr, int *cnt) 2557 { 2558 ext4_group_t ngroups = ext4_get_groups_count(sb); 2559 struct buffer_head *bh; 2560 struct blk_plug plug; 2561 2562 blk_start_plug(&plug); 2563 while (nr-- > 0) { 2564 struct ext4_group_desc *gdp = ext4_get_group_desc(sb, group, 2565 NULL); 2566 struct ext4_group_info *grp = ext4_get_group_info(sb, group); 2567 2568 /* 2569 * Prefetch block groups with free blocks; but don't 2570 * bother if it is marked uninitialized on disk, since 2571 * it won't require I/O to read. Also only try to 2572 * prefetch once, so we avoid getblk() call, which can 2573 * be expensive. 2574 */ 2575 if (!EXT4_MB_GRP_TEST_AND_SET_READ(grp) && 2576 EXT4_MB_GRP_NEED_INIT(grp) && 2577 ext4_free_group_clusters(sb, gdp) > 0 && 2578 !(ext4_has_group_desc_csum(sb) && 2579 (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)))) { 2580 bh = ext4_read_block_bitmap_nowait(sb, group, true); 2581 if (bh && !IS_ERR(bh)) { 2582 if (!buffer_uptodate(bh) && cnt) 2583 (*cnt)++; 2584 brelse(bh); 2585 } 2586 } 2587 if (++group >= ngroups) 2588 group = 0; 2589 } 2590 blk_finish_plug(&plug); 2591 return group; 2592 } 2593 2594 /* 2595 * Prefetching reads the block bitmap into the buffer cache; but we 2596 * need to make sure that the buddy bitmap in the page cache has been 2597 * initialized. Note that ext4_mb_init_group() will block if the I/O 2598 * is not yet completed, or indeed if it was not initiated by 2599 * ext4_mb_prefetch did not start the I/O. 2600 * 2601 * TODO: We should actually kick off the buddy bitmap setup in a work 2602 * queue when the buffer I/O is completed, so that we don't block 2603 * waiting for the block allocation bitmap read to finish when 2604 * ext4_mb_prefetch_fini is called from ext4_mb_regular_allocator(). 2605 */ 2606 void ext4_mb_prefetch_fini(struct super_block *sb, ext4_group_t group, 2607 unsigned int nr) 2608 { 2609 while (nr-- > 0) { 2610 struct ext4_group_desc *gdp = ext4_get_group_desc(sb, group, 2611 NULL); 2612 struct ext4_group_info *grp = ext4_get_group_info(sb, group); 2613 2614 if (!group) 2615 group = ext4_get_groups_count(sb); 2616 group--; 2617 grp = ext4_get_group_info(sb, group); 2618 2619 if (EXT4_MB_GRP_NEED_INIT(grp) && 2620 ext4_free_group_clusters(sb, gdp) > 0 && 2621 !(ext4_has_group_desc_csum(sb) && 2622 (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)))) { 2623 if (ext4_mb_init_group(sb, group, GFP_NOFS)) 2624 break; 2625 } 2626 } 2627 } 2628 2629 static noinline_for_stack int 2630 ext4_mb_regular_allocator(struct ext4_allocation_context *ac) 2631 { 2632 ext4_group_t prefetch_grp = 0, ngroups, group, i; 2633 int cr = -1; 2634 int err = 0, first_err = 0; 2635 unsigned int nr = 0, prefetch_ios = 0; 2636 struct ext4_sb_info *sbi; 2637 struct super_block *sb; 2638 struct ext4_buddy e4b; 2639 int lost; 2640 2641 sb = ac->ac_sb; 2642 sbi = EXT4_SB(sb); 2643 ngroups = ext4_get_groups_count(sb); 2644 /* non-extent files are limited to low blocks/groups */ 2645 if (!(ext4_test_inode_flag(ac->ac_inode, EXT4_INODE_EXTENTS))) 2646 ngroups = sbi->s_blockfile_groups; 2647 2648 BUG_ON(ac->ac_status == AC_STATUS_FOUND); 2649 2650 /* first, try the goal */ 2651 err = ext4_mb_find_by_goal(ac, &e4b); 2652 if (err || ac->ac_status == AC_STATUS_FOUND) 2653 goto out; 2654 2655 if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY)) 2656 goto out; 2657 2658 /* 2659 * ac->ac_2order is set only if the fe_len is a power of 2 2660 * if ac->ac_2order is set we also set criteria to 0 so that we 2661 * try exact allocation using buddy. 2662 */ 2663 i = fls(ac->ac_g_ex.fe_len); 2664 ac->ac_2order = 0; 2665 /* 2666 * We search using buddy data only if the order of the request 2667 * is greater than equal to the sbi_s_mb_order2_reqs 2668 * You can tune it via /sys/fs/ext4/<partition>/mb_order2_req 2669 * We also support searching for power-of-two requests only for 2670 * requests upto maximum buddy size we have constructed. 2671 */ 2672 if (i >= sbi->s_mb_order2_reqs && i <= MB_NUM_ORDERS(sb)) { 2673 /* 2674 * This should tell if fe_len is exactly power of 2 2675 */ 2676 if ((ac->ac_g_ex.fe_len & (~(1 << (i - 1)))) == 0) 2677 ac->ac_2order = array_index_nospec(i - 1, 2678 MB_NUM_ORDERS(sb)); 2679 } 2680 2681 /* if stream allocation is enabled, use global goal */ 2682 if (ac->ac_flags & EXT4_MB_STREAM_ALLOC) { 2683 /* TBD: may be hot point */ 2684 spin_lock(&sbi->s_md_lock); 2685 ac->ac_g_ex.fe_group = sbi->s_mb_last_group; 2686 ac->ac_g_ex.fe_start = sbi->s_mb_last_start; 2687 spin_unlock(&sbi->s_md_lock); 2688 } 2689 2690 /* Let's just scan groups to find more-less suitable blocks */ 2691 cr = ac->ac_2order ? 0 : 1; 2692 /* 2693 * cr == 0 try to get exact allocation, 2694 * cr == 3 try to get anything 2695 */ 2696 repeat: 2697 for (; cr < 4 && ac->ac_status == AC_STATUS_CONTINUE; cr++) { 2698 ac->ac_criteria = cr; 2699 /* 2700 * searching for the right group start 2701 * from the goal value specified 2702 */ 2703 group = ac->ac_g_ex.fe_group; 2704 ac->ac_last_optimal_group = group; 2705 ac->ac_groups_linear_remaining = sbi->s_mb_max_linear_groups; 2706 prefetch_grp = group; 2707 2708 for (i = 0; i < ngroups; group = next_linear_group(ac, group, ngroups), 2709 i++) { 2710 int ret = 0, new_cr; 2711 2712 cond_resched(); 2713 2714 ext4_mb_choose_next_group(ac, &new_cr, &group, ngroups); 2715 if (new_cr != cr) { 2716 cr = new_cr; 2717 goto repeat; 2718 } 2719 2720 /* 2721 * Batch reads of the block allocation bitmaps 2722 * to get multiple READs in flight; limit 2723 * prefetching at cr=0/1, otherwise mballoc can 2724 * spend a lot of time loading imperfect groups 2725 */ 2726 if ((prefetch_grp == group) && 2727 (cr > 1 || 2728 prefetch_ios < sbi->s_mb_prefetch_limit)) { 2729 unsigned int curr_ios = prefetch_ios; 2730 2731 nr = sbi->s_mb_prefetch; 2732 if (ext4_has_feature_flex_bg(sb)) { 2733 nr = 1 << sbi->s_log_groups_per_flex; 2734 nr -= group & (nr - 1); 2735 nr = min(nr, sbi->s_mb_prefetch); 2736 } 2737 prefetch_grp = ext4_mb_prefetch(sb, group, 2738 nr, &prefetch_ios); 2739 if (prefetch_ios == curr_ios) 2740 nr = 0; 2741 } 2742 2743 /* This now checks without needing the buddy page */ 2744 ret = ext4_mb_good_group_nolock(ac, group, cr); 2745 if (ret <= 0) { 2746 if (!first_err) 2747 first_err = ret; 2748 continue; 2749 } 2750 2751 err = ext4_mb_load_buddy(sb, group, &e4b); 2752 if (err) 2753 goto out; 2754 2755 ext4_lock_group(sb, group); 2756 2757 /* 2758 * We need to check again after locking the 2759 * block group 2760 */ 2761 ret = ext4_mb_good_group(ac, group, cr); 2762 if (ret == 0) { 2763 ext4_unlock_group(sb, group); 2764 ext4_mb_unload_buddy(&e4b); 2765 continue; 2766 } 2767 2768 ac->ac_groups_scanned++; 2769 if (cr == 0) 2770 ext4_mb_simple_scan_group(ac, &e4b); 2771 else if (cr == 1 && sbi->s_stripe && 2772 !(ac->ac_g_ex.fe_len % sbi->s_stripe)) 2773 ext4_mb_scan_aligned(ac, &e4b); 2774 else 2775 ext4_mb_complex_scan_group(ac, &e4b); 2776 2777 ext4_unlock_group(sb, group); 2778 ext4_mb_unload_buddy(&e4b); 2779 2780 if (ac->ac_status != AC_STATUS_CONTINUE) 2781 break; 2782 } 2783 /* Processed all groups and haven't found blocks */ 2784 if (sbi->s_mb_stats && i == ngroups) 2785 atomic64_inc(&sbi->s_bal_cX_failed[cr]); 2786 } 2787 2788 if (ac->ac_b_ex.fe_len > 0 && ac->ac_status != AC_STATUS_FOUND && 2789 !(ac->ac_flags & EXT4_MB_HINT_FIRST)) { 2790 /* 2791 * We've been searching too long. Let's try to allocate 2792 * the best chunk we've found so far 2793 */ 2794 ext4_mb_try_best_found(ac, &e4b); 2795 if (ac->ac_status != AC_STATUS_FOUND) { 2796 /* 2797 * Someone more lucky has already allocated it. 2798 * The only thing we can do is just take first 2799 * found block(s) 2800 */ 2801 lost = atomic_inc_return(&sbi->s_mb_lost_chunks); 2802 mb_debug(sb, "lost chunk, group: %u, start: %d, len: %d, lost: %d\n", 2803 ac->ac_b_ex.fe_group, ac->ac_b_ex.fe_start, 2804 ac->ac_b_ex.fe_len, lost); 2805 2806 ac->ac_b_ex.fe_group = 0; 2807 ac->ac_b_ex.fe_start = 0; 2808 ac->ac_b_ex.fe_len = 0; 2809 ac->ac_status = AC_STATUS_CONTINUE; 2810 ac->ac_flags |= EXT4_MB_HINT_FIRST; 2811 cr = 3; 2812 goto repeat; 2813 } 2814 } 2815 2816 if (sbi->s_mb_stats && ac->ac_status == AC_STATUS_FOUND) 2817 atomic64_inc(&sbi->s_bal_cX_hits[ac->ac_criteria]); 2818 out: 2819 if (!err && ac->ac_status != AC_STATUS_FOUND && first_err) 2820 err = first_err; 2821 2822 mb_debug(sb, "Best len %d, origin len %d, ac_status %u, ac_flags 0x%x, cr %d ret %d\n", 2823 ac->ac_b_ex.fe_len, ac->ac_o_ex.fe_len, ac->ac_status, 2824 ac->ac_flags, cr, err); 2825 2826 if (nr) 2827 ext4_mb_prefetch_fini(sb, prefetch_grp, nr); 2828 2829 return err; 2830 } 2831 2832 static void *ext4_mb_seq_groups_start(struct seq_file *seq, loff_t *pos) 2833 { 2834 struct super_block *sb = pde_data(file_inode(seq->file)); 2835 ext4_group_t group; 2836 2837 if (*pos < 0 || *pos >= ext4_get_groups_count(sb)) 2838 return NULL; 2839 group = *pos + 1; 2840 return (void *) ((unsigned long) group); 2841 } 2842 2843 static void *ext4_mb_seq_groups_next(struct seq_file *seq, void *v, loff_t *pos) 2844 { 2845 struct super_block *sb = pde_data(file_inode(seq->file)); 2846 ext4_group_t group; 2847 2848 ++*pos; 2849 if (*pos < 0 || *pos >= ext4_get_groups_count(sb)) 2850 return NULL; 2851 group = *pos + 1; 2852 return (void *) ((unsigned long) group); 2853 } 2854 2855 static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v) 2856 { 2857 struct super_block *sb = pde_data(file_inode(seq->file)); 2858 ext4_group_t group = (ext4_group_t) ((unsigned long) v); 2859 int i; 2860 int err, buddy_loaded = 0; 2861 struct ext4_buddy e4b; 2862 struct ext4_group_info *grinfo; 2863 unsigned char blocksize_bits = min_t(unsigned char, 2864 sb->s_blocksize_bits, 2865 EXT4_MAX_BLOCK_LOG_SIZE); 2866 struct sg { 2867 struct ext4_group_info info; 2868 ext4_grpblk_t counters[EXT4_MAX_BLOCK_LOG_SIZE + 2]; 2869 } sg; 2870 2871 group--; 2872 if (group == 0) 2873 seq_puts(seq, "#group: free frags first [" 2874 " 2^0 2^1 2^2 2^3 2^4 2^5 2^6 " 2875 " 2^7 2^8 2^9 2^10 2^11 2^12 2^13 ]\n"); 2876 2877 i = (blocksize_bits + 2) * sizeof(sg.info.bb_counters[0]) + 2878 sizeof(struct ext4_group_info); 2879 2880 grinfo = ext4_get_group_info(sb, group); 2881 /* Load the group info in memory only if not already loaded. */ 2882 if (unlikely(EXT4_MB_GRP_NEED_INIT(grinfo))) { 2883 err = ext4_mb_load_buddy(sb, group, &e4b); 2884 if (err) { 2885 seq_printf(seq, "#%-5u: I/O error\n", group); 2886 return 0; 2887 } 2888 buddy_loaded = 1; 2889 } 2890 2891 memcpy(&sg, ext4_get_group_info(sb, group), i); 2892 2893 if (buddy_loaded) 2894 ext4_mb_unload_buddy(&e4b); 2895 2896 seq_printf(seq, "#%-5u: %-5u %-5u %-5u [", group, sg.info.bb_free, 2897 sg.info.bb_fragments, sg.info.bb_first_free); 2898 for (i = 0; i <= 13; i++) 2899 seq_printf(seq, " %-5u", i <= blocksize_bits + 1 ? 2900 sg.info.bb_counters[i] : 0); 2901 seq_puts(seq, " ]\n"); 2902 2903 return 0; 2904 } 2905 2906 static void ext4_mb_seq_groups_stop(struct seq_file *seq, void *v) 2907 { 2908 } 2909 2910 const struct seq_operations ext4_mb_seq_groups_ops = { 2911 .start = ext4_mb_seq_groups_start, 2912 .next = ext4_mb_seq_groups_next, 2913 .stop = ext4_mb_seq_groups_stop, 2914 .show = ext4_mb_seq_groups_show, 2915 }; 2916 2917 int ext4_seq_mb_stats_show(struct seq_file *seq, void *offset) 2918 { 2919 struct super_block *sb = seq->private; 2920 struct ext4_sb_info *sbi = EXT4_SB(sb); 2921 2922 seq_puts(seq, "mballoc:\n"); 2923 if (!sbi->s_mb_stats) { 2924 seq_puts(seq, "\tmb stats collection turned off.\n"); 2925 seq_puts(seq, "\tTo enable, please write \"1\" to sysfs file mb_stats.\n"); 2926 return 0; 2927 } 2928 seq_printf(seq, "\treqs: %u\n", atomic_read(&sbi->s_bal_reqs)); 2929 seq_printf(seq, "\tsuccess: %u\n", atomic_read(&sbi->s_bal_success)); 2930 2931 seq_printf(seq, "\tgroups_scanned: %u\n", atomic_read(&sbi->s_bal_groups_scanned)); 2932 2933 seq_puts(seq, "\tcr0_stats:\n"); 2934 seq_printf(seq, "\t\thits: %llu\n", atomic64_read(&sbi->s_bal_cX_hits[0])); 2935 seq_printf(seq, "\t\tgroups_considered: %llu\n", 2936 atomic64_read(&sbi->s_bal_cX_groups_considered[0])); 2937 seq_printf(seq, "\t\tuseless_loops: %llu\n", 2938 atomic64_read(&sbi->s_bal_cX_failed[0])); 2939 seq_printf(seq, "\t\tbad_suggestions: %u\n", 2940 atomic_read(&sbi->s_bal_cr0_bad_suggestions)); 2941 2942 seq_puts(seq, "\tcr1_stats:\n"); 2943 seq_printf(seq, "\t\thits: %llu\n", atomic64_read(&sbi->s_bal_cX_hits[1])); 2944 seq_printf(seq, "\t\tgroups_considered: %llu\n", 2945 atomic64_read(&sbi->s_bal_cX_groups_considered[1])); 2946 seq_printf(seq, "\t\tuseless_loops: %llu\n", 2947 atomic64_read(&sbi->s_bal_cX_failed[1])); 2948 seq_printf(seq, "\t\tbad_suggestions: %u\n", 2949 atomic_read(&sbi->s_bal_cr1_bad_suggestions)); 2950 2951 seq_puts(seq, "\tcr2_stats:\n"); 2952 seq_printf(seq, "\t\thits: %llu\n", atomic64_read(&sbi->s_bal_cX_hits[2])); 2953 seq_printf(seq, "\t\tgroups_considered: %llu\n", 2954 atomic64_read(&sbi->s_bal_cX_groups_considered[2])); 2955 seq_printf(seq, "\t\tuseless_loops: %llu\n", 2956 atomic64_read(&sbi->s_bal_cX_failed[2])); 2957 2958 seq_puts(seq, "\tcr3_stats:\n"); 2959 seq_printf(seq, "\t\thits: %llu\n", atomic64_read(&sbi->s_bal_cX_hits[3])); 2960 seq_printf(seq, "\t\tgroups_considered: %llu\n", 2961 atomic64_read(&sbi->s_bal_cX_groups_considered[3])); 2962 seq_printf(seq, "\t\tuseless_loops: %llu\n", 2963 atomic64_read(&sbi->s_bal_cX_failed[3])); 2964 seq_printf(seq, "\textents_scanned: %u\n", atomic_read(&sbi->s_bal_ex_scanned)); 2965 seq_printf(seq, "\t\tgoal_hits: %u\n", atomic_read(&sbi->s_bal_goals)); 2966 seq_printf(seq, "\t\t2^n_hits: %u\n", atomic_read(&sbi->s_bal_2orders)); 2967 seq_printf(seq, "\t\tbreaks: %u\n", atomic_read(&sbi->s_bal_breaks)); 2968 seq_printf(seq, "\t\tlost: %u\n", atomic_read(&sbi->s_mb_lost_chunks)); 2969 2970 seq_printf(seq, "\tbuddies_generated: %u/%u\n", 2971 atomic_read(&sbi->s_mb_buddies_generated), 2972 ext4_get_groups_count(sb)); 2973 seq_printf(seq, "\tbuddies_time_used: %llu\n", 2974 atomic64_read(&sbi->s_mb_generation_time)); 2975 seq_printf(seq, "\tpreallocated: %u\n", 2976 atomic_read(&sbi->s_mb_preallocated)); 2977 seq_printf(seq, "\tdiscarded: %u\n", 2978 atomic_read(&sbi->s_mb_discarded)); 2979 return 0; 2980 } 2981 2982 static void *ext4_mb_seq_structs_summary_start(struct seq_file *seq, loff_t *pos) 2983 __acquires(&EXT4_SB(sb)->s_mb_rb_lock) 2984 { 2985 struct super_block *sb = pde_data(file_inode(seq->file)); 2986 unsigned long position; 2987 2988 read_lock(&EXT4_SB(sb)->s_mb_rb_lock); 2989 2990 if (*pos < 0 || *pos >= MB_NUM_ORDERS(sb) + 1) 2991 return NULL; 2992 position = *pos + 1; 2993 return (void *) ((unsigned long) position); 2994 } 2995 2996 static void *ext4_mb_seq_structs_summary_next(struct seq_file *seq, void *v, loff_t *pos) 2997 { 2998 struct super_block *sb = pde_data(file_inode(seq->file)); 2999 unsigned long position; 3000 3001 ++*pos; 3002 if (*pos < 0 || *pos >= MB_NUM_ORDERS(sb) + 1) 3003 return NULL; 3004 position = *pos + 1; 3005 return (void *) ((unsigned long) position); 3006 } 3007 3008 static int ext4_mb_seq_structs_summary_show(struct seq_file *seq, void *v) 3009 { 3010 struct super_block *sb = pde_data(file_inode(seq->file)); 3011 struct ext4_sb_info *sbi = EXT4_SB(sb); 3012 unsigned long position = ((unsigned long) v); 3013 struct ext4_group_info *grp; 3014 struct rb_node *n; 3015 unsigned int count, min, max; 3016 3017 position--; 3018 if (position >= MB_NUM_ORDERS(sb)) { 3019 seq_puts(seq, "fragment_size_tree:\n"); 3020 n = rb_first(&sbi->s_mb_avg_fragment_size_root); 3021 if (!n) { 3022 seq_puts(seq, "\ttree_min: 0\n\ttree_max: 0\n\ttree_nodes: 0\n"); 3023 return 0; 3024 } 3025 grp = rb_entry(n, struct ext4_group_info, bb_avg_fragment_size_rb); 3026 min = grp->bb_fragments ? grp->bb_free / grp->bb_fragments : 0; 3027 count = 1; 3028 while (rb_next(n)) { 3029 count++; 3030 n = rb_next(n); 3031 } 3032 grp = rb_entry(n, struct ext4_group_info, bb_avg_fragment_size_rb); 3033 max = grp->bb_fragments ? grp->bb_free / grp->bb_fragments : 0; 3034 3035 seq_printf(seq, "\ttree_min: %u\n\ttree_max: %u\n\ttree_nodes: %u\n", 3036 min, max, count); 3037 return 0; 3038 } 3039 3040 if (position == 0) { 3041 seq_printf(seq, "optimize_scan: %d\n", 3042 test_opt2(sb, MB_OPTIMIZE_SCAN) ? 1 : 0); 3043 seq_puts(seq, "max_free_order_lists:\n"); 3044 } 3045 count = 0; 3046 list_for_each_entry(grp, &sbi->s_mb_largest_free_orders[position], 3047 bb_largest_free_order_node) 3048 count++; 3049 seq_printf(seq, "\tlist_order_%u_groups: %u\n", 3050 (unsigned int)position, count); 3051 3052 return 0; 3053 } 3054 3055 static void ext4_mb_seq_structs_summary_stop(struct seq_file *seq, void *v) 3056 __releases(&EXT4_SB(sb)->s_mb_rb_lock) 3057 { 3058 struct super_block *sb = pde_data(file_inode(seq->file)); 3059 3060 read_unlock(&EXT4_SB(sb)->s_mb_rb_lock); 3061 } 3062 3063 const struct seq_operations ext4_mb_seq_structs_summary_ops = { 3064 .start = ext4_mb_seq_structs_summary_start, 3065 .next = ext4_mb_seq_structs_summary_next, 3066 .stop = ext4_mb_seq_structs_summary_stop, 3067 .show = ext4_mb_seq_structs_summary_show, 3068 }; 3069 3070 static struct kmem_cache *get_groupinfo_cache(int blocksize_bits) 3071 { 3072 int cache_index = blocksize_bits - EXT4_MIN_BLOCK_LOG_SIZE; 3073 struct kmem_cache *cachep = ext4_groupinfo_caches[cache_index]; 3074 3075 BUG_ON(!cachep); 3076 return cachep; 3077 } 3078 3079 /* 3080 * Allocate the top-level s_group_info array for the specified number 3081 * of groups 3082 */ 3083 int ext4_mb_alloc_groupinfo(struct super_block *sb, ext4_group_t ngroups) 3084 { 3085 struct ext4_sb_info *sbi = EXT4_SB(sb); 3086 unsigned size; 3087 struct ext4_group_info ***old_groupinfo, ***new_groupinfo; 3088 3089 size = (ngroups + EXT4_DESC_PER_BLOCK(sb) - 1) >> 3090 EXT4_DESC_PER_BLOCK_BITS(sb); 3091 if (size <= sbi->s_group_info_size) 3092 return 0; 3093 3094 size = roundup_pow_of_two(sizeof(*sbi->s_group_info) * size); 3095 new_groupinfo = kvzalloc(size, GFP_KERNEL); 3096 if (!new_groupinfo) { 3097 ext4_msg(sb, KERN_ERR, "can't allocate buddy meta group"); 3098 return -ENOMEM; 3099 } 3100 rcu_read_lock(); 3101 old_groupinfo = rcu_dereference(sbi->s_group_info); 3102 if (old_groupinfo) 3103 memcpy(new_groupinfo, old_groupinfo, 3104 sbi->s_group_info_size * sizeof(*sbi->s_group_info)); 3105 rcu_read_unlock(); 3106 rcu_assign_pointer(sbi->s_group_info, new_groupinfo); 3107 sbi->s_group_info_size = size / sizeof(*sbi->s_group_info); 3108 if (old_groupinfo) 3109 ext4_kvfree_array_rcu(old_groupinfo); 3110 ext4_debug("allocated s_groupinfo array for %d meta_bg's\n", 3111 sbi->s_group_info_size); 3112 return 0; 3113 } 3114 3115 /* Create and initialize ext4_group_info data for the given group. */ 3116 int ext4_mb_add_groupinfo(struct super_block *sb, ext4_group_t group, 3117 struct ext4_group_desc *desc) 3118 { 3119 int i; 3120 int metalen = 0; 3121 int idx = group >> EXT4_DESC_PER_BLOCK_BITS(sb); 3122 struct ext4_sb_info *sbi = EXT4_SB(sb); 3123 struct ext4_group_info **meta_group_info; 3124 struct kmem_cache *cachep = get_groupinfo_cache(sb->s_blocksize_bits); 3125 3126 /* 3127 * First check if this group is the first of a reserved block. 3128 * If it's true, we have to allocate a new table of pointers 3129 * to ext4_group_info structures 3130 */ 3131 if (group % EXT4_DESC_PER_BLOCK(sb) == 0) { 3132 metalen = sizeof(*meta_group_info) << 3133 EXT4_DESC_PER_BLOCK_BITS(sb); 3134 meta_group_info = kmalloc(metalen, GFP_NOFS); 3135 if (meta_group_info == NULL) { 3136 ext4_msg(sb, KERN_ERR, "can't allocate mem " 3137 "for a buddy group"); 3138 goto exit_meta_group_info; 3139 } 3140 rcu_read_lock(); 3141 rcu_dereference(sbi->s_group_info)[idx] = meta_group_info; 3142 rcu_read_unlock(); 3143 } 3144 3145 meta_group_info = sbi_array_rcu_deref(sbi, s_group_info, idx); 3146 i = group & (EXT4_DESC_PER_BLOCK(sb) - 1); 3147 3148 meta_group_info[i] = kmem_cache_zalloc(cachep, GFP_NOFS); 3149 if (meta_group_info[i] == NULL) { 3150 ext4_msg(sb, KERN_ERR, "can't allocate buddy mem"); 3151 goto exit_group_info; 3152 } 3153 set_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, 3154 &(meta_group_info[i]->bb_state)); 3155 3156 /* 3157 * initialize bb_free to be able to skip 3158 * empty groups without initialization 3159 */ 3160 if (ext4_has_group_desc_csum(sb) && 3161 (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT))) { 3162 meta_group_info[i]->bb_free = 3163 ext4_free_clusters_after_init(sb, group, desc); 3164 } else { 3165 meta_group_info[i]->bb_free = 3166 ext4_free_group_clusters(sb, desc); 3167 } 3168 3169 INIT_LIST_HEAD(&meta_group_info[i]->bb_prealloc_list); 3170 init_rwsem(&meta_group_info[i]->alloc_sem); 3171 meta_group_info[i]->bb_free_root = RB_ROOT; 3172 INIT_LIST_HEAD(&meta_group_info[i]->bb_largest_free_order_node); 3173 RB_CLEAR_NODE(&meta_group_info[i]->bb_avg_fragment_size_rb); 3174 meta_group_info[i]->bb_largest_free_order = -1; /* uninit */ 3175 meta_group_info[i]->bb_group = group; 3176 3177 mb_group_bb_bitmap_alloc(sb, meta_group_info[i], group); 3178 return 0; 3179 3180 exit_group_info: 3181 /* If a meta_group_info table has been allocated, release it now */ 3182 if (group % EXT4_DESC_PER_BLOCK(sb) == 0) { 3183 struct ext4_group_info ***group_info; 3184 3185 rcu_read_lock(); 3186 group_info = rcu_dereference(sbi->s_group_info); 3187 kfree(group_info[idx]); 3188 group_info[idx] = NULL; 3189 rcu_read_unlock(); 3190 } 3191 exit_meta_group_info: 3192 return -ENOMEM; 3193 } /* ext4_mb_add_groupinfo */ 3194 3195 static int ext4_mb_init_backend(struct super_block *sb) 3196 { 3197 ext4_group_t ngroups = ext4_get_groups_count(sb); 3198 ext4_group_t i; 3199 struct ext4_sb_info *sbi = EXT4_SB(sb); 3200 int err; 3201 struct ext4_group_desc *desc; 3202 struct ext4_group_info ***group_info; 3203 struct kmem_cache *cachep; 3204 3205 err = ext4_mb_alloc_groupinfo(sb, ngroups); 3206 if (err) 3207 return err; 3208 3209 sbi->s_buddy_cache = new_inode(sb); 3210 if (sbi->s_buddy_cache == NULL) { 3211 ext4_msg(sb, KERN_ERR, "can't get new inode"); 3212 goto err_freesgi; 3213 } 3214 /* To avoid potentially colliding with an valid on-disk inode number, 3215 * use EXT4_BAD_INO for the buddy cache inode number. This inode is 3216 * not in the inode hash, so it should never be found by iget(), but 3217 * this will avoid confusion if it ever shows up during debugging. */ 3218 sbi->s_buddy_cache->i_ino = EXT4_BAD_INO; 3219 EXT4_I(sbi->s_buddy_cache)->i_disksize = 0; 3220 for (i = 0; i < ngroups; i++) { 3221 cond_resched(); 3222 desc = ext4_get_group_desc(sb, i, NULL); 3223 if (desc == NULL) { 3224 ext4_msg(sb, KERN_ERR, "can't read descriptor %u", i); 3225 goto err_freebuddy; 3226 } 3227 if (ext4_mb_add_groupinfo(sb, i, desc) != 0) 3228 goto err_freebuddy; 3229 } 3230 3231 if (ext4_has_feature_flex_bg(sb)) { 3232 /* a single flex group is supposed to be read by a single IO. 3233 * 2 ^ s_log_groups_per_flex != UINT_MAX as s_mb_prefetch is 3234 * unsigned integer, so the maximum shift is 32. 3235 */ 3236 if (sbi->s_es->s_log_groups_per_flex >= 32) { 3237 ext4_msg(sb, KERN_ERR, "too many log groups per flexible block group"); 3238 goto err_freebuddy; 3239 } 3240 sbi->s_mb_prefetch = min_t(uint, 1 << sbi->s_es->s_log_groups_per_flex, 3241 BLK_MAX_SEGMENT_SIZE >> (sb->s_blocksize_bits - 9)); 3242 sbi->s_mb_prefetch *= 8; /* 8 prefetch IOs in flight at most */ 3243 } else { 3244 sbi->s_mb_prefetch = 32; 3245 } 3246 if (sbi->s_mb_prefetch > ext4_get_groups_count(sb)) 3247 sbi->s_mb_prefetch = ext4_get_groups_count(sb); 3248 /* now many real IOs to prefetch within a single allocation at cr=0 3249 * given cr=0 is an CPU-related optimization we shouldn't try to 3250 * load too many groups, at some point we should start to use what 3251 * we've got in memory. 3252 * with an average random access time 5ms, it'd take a second to get 3253 * 200 groups (* N with flex_bg), so let's make this limit 4 3254 */ 3255 sbi->s_mb_prefetch_limit = sbi->s_mb_prefetch * 4; 3256 if (sbi->s_mb_prefetch_limit > ext4_get_groups_count(sb)) 3257 sbi->s_mb_prefetch_limit = ext4_get_groups_count(sb); 3258 3259 return 0; 3260 3261 err_freebuddy: 3262 cachep = get_groupinfo_cache(sb->s_blocksize_bits); 3263 while (i-- > 0) 3264 kmem_cache_free(cachep, ext4_get_group_info(sb, i)); 3265 i = sbi->s_group_info_size; 3266 rcu_read_lock(); 3267 group_info = rcu_dereference(sbi->s_group_info); 3268 while (i-- > 0) 3269 kfree(group_info[i]); 3270 rcu_read_unlock(); 3271 iput(sbi->s_buddy_cache); 3272 err_freesgi: 3273 rcu_read_lock(); 3274 kvfree(rcu_dereference(sbi->s_group_info)); 3275 rcu_read_unlock(); 3276 return -ENOMEM; 3277 } 3278 3279 static void ext4_groupinfo_destroy_slabs(void) 3280 { 3281 int i; 3282 3283 for (i = 0; i < NR_GRPINFO_CACHES; i++) { 3284 kmem_cache_destroy(ext4_groupinfo_caches[i]); 3285 ext4_groupinfo_caches[i] = NULL; 3286 } 3287 } 3288 3289 static int ext4_groupinfo_create_slab(size_t size) 3290 { 3291 static DEFINE_MUTEX(ext4_grpinfo_slab_create_mutex); 3292 int slab_size; 3293 int blocksize_bits = order_base_2(size); 3294 int cache_index = blocksize_bits - EXT4_MIN_BLOCK_LOG_SIZE; 3295 struct kmem_cache *cachep; 3296 3297 if (cache_index >= NR_GRPINFO_CACHES) 3298 return -EINVAL; 3299 3300 if (unlikely(cache_index < 0)) 3301 cache_index = 0; 3302 3303 mutex_lock(&ext4_grpinfo_slab_create_mutex); 3304 if (ext4_groupinfo_caches[cache_index]) { 3305 mutex_unlock(&ext4_grpinfo_slab_create_mutex); 3306 return 0; /* Already created */ 3307 } 3308 3309 slab_size = offsetof(struct ext4_group_info, 3310 bb_counters[blocksize_bits + 2]); 3311 3312 cachep = kmem_cache_create(ext4_groupinfo_slab_names[cache_index], 3313 slab_size, 0, SLAB_RECLAIM_ACCOUNT, 3314 NULL); 3315 3316 ext4_groupinfo_caches[cache_index] = cachep; 3317 3318 mutex_unlock(&ext4_grpinfo_slab_create_mutex); 3319 if (!cachep) { 3320 printk(KERN_EMERG 3321 "EXT4-fs: no memory for groupinfo slab cache\n"); 3322 return -ENOMEM; 3323 } 3324 3325 return 0; 3326 } 3327 3328 static void ext4_discard_work(struct work_struct *work) 3329 { 3330 struct ext4_sb_info *sbi = container_of(work, 3331 struct ext4_sb_info, s_discard_work); 3332 struct super_block *sb = sbi->s_sb; 3333 struct ext4_free_data *fd, *nfd; 3334 struct ext4_buddy e4b; 3335 struct list_head discard_list; 3336 ext4_group_t grp, load_grp; 3337 int err = 0; 3338 3339 INIT_LIST_HEAD(&discard_list); 3340 spin_lock(&sbi->s_md_lock); 3341 list_splice_init(&sbi->s_discard_list, &discard_list); 3342 spin_unlock(&sbi->s_md_lock); 3343 3344 load_grp = UINT_MAX; 3345 list_for_each_entry_safe(fd, nfd, &discard_list, efd_list) { 3346 /* 3347 * If filesystem is umounting or no memory or suffering 3348 * from no space, give up the discard 3349 */ 3350 if ((sb->s_flags & SB_ACTIVE) && !err && 3351 !atomic_read(&sbi->s_retry_alloc_pending)) { 3352 grp = fd->efd_group; 3353 if (grp != load_grp) { 3354 if (load_grp != UINT_MAX) 3355 ext4_mb_unload_buddy(&e4b); 3356 3357 err = ext4_mb_load_buddy(sb, grp, &e4b); 3358 if (err) { 3359 kmem_cache_free(ext4_free_data_cachep, fd); 3360 load_grp = UINT_MAX; 3361 continue; 3362 } else { 3363 load_grp = grp; 3364 } 3365 } 3366 3367 ext4_lock_group(sb, grp); 3368 ext4_try_to_trim_range(sb, &e4b, fd->efd_start_cluster, 3369 fd->efd_start_cluster + fd->efd_count - 1, 1); 3370 ext4_unlock_group(sb, grp); 3371 } 3372 kmem_cache_free(ext4_free_data_cachep, fd); 3373 } 3374 3375 if (load_grp != UINT_MAX) 3376 ext4_mb_unload_buddy(&e4b); 3377 } 3378 3379 int ext4_mb_init(struct super_block *sb) 3380 { 3381 struct ext4_sb_info *sbi = EXT4_SB(sb); 3382 unsigned i, j; 3383 unsigned offset, offset_incr; 3384 unsigned max; 3385 int ret; 3386 3387 i = MB_NUM_ORDERS(sb) * sizeof(*sbi->s_mb_offsets); 3388 3389 sbi->s_mb_offsets = kmalloc(i, GFP_KERNEL); 3390 if (sbi->s_mb_offsets == NULL) { 3391 ret = -ENOMEM; 3392 goto out; 3393 } 3394 3395 i = MB_NUM_ORDERS(sb) * sizeof(*sbi->s_mb_maxs); 3396 sbi->s_mb_maxs = kmalloc(i, GFP_KERNEL); 3397 if (sbi->s_mb_maxs == NULL) { 3398 ret = -ENOMEM; 3399 goto out; 3400 } 3401 3402 ret = ext4_groupinfo_create_slab(sb->s_blocksize); 3403 if (ret < 0) 3404 goto out; 3405 3406 /* order 0 is regular bitmap */ 3407 sbi->s_mb_maxs[0] = sb->s_blocksize << 3; 3408 sbi->s_mb_offsets[0] = 0; 3409 3410 i = 1; 3411 offset = 0; 3412 offset_incr = 1 << (sb->s_blocksize_bits - 1); 3413 max = sb->s_blocksize << 2; 3414 do { 3415 sbi->s_mb_offsets[i] = offset; 3416 sbi->s_mb_maxs[i] = max; 3417 offset += offset_incr; 3418 offset_incr = offset_incr >> 1; 3419 max = max >> 1; 3420 i++; 3421 } while (i < MB_NUM_ORDERS(sb)); 3422 3423 sbi->s_mb_avg_fragment_size_root = RB_ROOT; 3424 sbi->s_mb_largest_free_orders = 3425 kmalloc_array(MB_NUM_ORDERS(sb), sizeof(struct list_head), 3426 GFP_KERNEL); 3427 if (!sbi->s_mb_largest_free_orders) { 3428 ret = -ENOMEM; 3429 goto out; 3430 } 3431 sbi->s_mb_largest_free_orders_locks = 3432 kmalloc_array(MB_NUM_ORDERS(sb), sizeof(rwlock_t), 3433 GFP_KERNEL); 3434 if (!sbi->s_mb_largest_free_orders_locks) { 3435 ret = -ENOMEM; 3436 goto out; 3437 } 3438 for (i = 0; i < MB_NUM_ORDERS(sb); i++) { 3439 INIT_LIST_HEAD(&sbi->s_mb_largest_free_orders[i]); 3440 rwlock_init(&sbi->s_mb_largest_free_orders_locks[i]); 3441 } 3442 rwlock_init(&sbi->s_mb_rb_lock); 3443 3444 spin_lock_init(&sbi->s_md_lock); 3445 sbi->s_mb_free_pending = 0; 3446 INIT_LIST_HEAD(&sbi->s_freed_data_list); 3447 INIT_LIST_HEAD(&sbi->s_discard_list); 3448 INIT_WORK(&sbi->s_discard_work, ext4_discard_work); 3449 atomic_set(&sbi->s_retry_alloc_pending, 0); 3450 3451 sbi->s_mb_max_to_scan = MB_DEFAULT_MAX_TO_SCAN; 3452 sbi->s_mb_min_to_scan = MB_DEFAULT_MIN_TO_SCAN; 3453 sbi->s_mb_stats = MB_DEFAULT_STATS; 3454 sbi->s_mb_stream_request = MB_DEFAULT_STREAM_THRESHOLD; 3455 sbi->s_mb_order2_reqs = MB_DEFAULT_ORDER2_REQS; 3456 sbi->s_mb_max_inode_prealloc = MB_DEFAULT_MAX_INODE_PREALLOC; 3457 /* 3458 * The default group preallocation is 512, which for 4k block 3459 * sizes translates to 2 megabytes. However for bigalloc file 3460 * systems, this is probably too big (i.e, if the cluster size 3461 * is 1 megabyte, then group preallocation size becomes half a 3462 * gigabyte!). As a default, we will keep a two megabyte 3463 * group pralloc size for cluster sizes up to 64k, and after 3464 * that, we will force a minimum group preallocation size of 3465 * 32 clusters. This translates to 8 megs when the cluster 3466 * size is 256k, and 32 megs when the cluster size is 1 meg, 3467 * which seems reasonable as a default. 3468 */ 3469 sbi->s_mb_group_prealloc = max(MB_DEFAULT_GROUP_PREALLOC >> 3470 sbi->s_cluster_bits, 32); 3471 /* 3472 * If there is a s_stripe > 1, then we set the s_mb_group_prealloc 3473 * to the lowest multiple of s_stripe which is bigger than 3474 * the s_mb_group_prealloc as determined above. We want 3475 * the preallocation size to be an exact multiple of the 3476 * RAID stripe size so that preallocations don't fragment 3477 * the stripes. 3478 */ 3479 if (sbi->s_stripe > 1) { 3480 sbi->s_mb_group_prealloc = roundup( 3481 sbi->s_mb_group_prealloc, sbi->s_stripe); 3482 } 3483 3484 sbi->s_locality_groups = alloc_percpu(struct ext4_locality_group); 3485 if (sbi->s_locality_groups == NULL) { 3486 ret = -ENOMEM; 3487 goto out; 3488 } 3489 for_each_possible_cpu(i) { 3490 struct ext4_locality_group *lg; 3491 lg = per_cpu_ptr(sbi->s_locality_groups, i); 3492 mutex_init(&lg->lg_mutex); 3493 for (j = 0; j < PREALLOC_TB_SIZE; j++) 3494 INIT_LIST_HEAD(&lg->lg_prealloc_list[j]); 3495 spin_lock_init(&lg->lg_prealloc_lock); 3496 } 3497 3498 if (bdev_nonrot(sb->s_bdev)) 3499 sbi->s_mb_max_linear_groups = 0; 3500 else 3501 sbi->s_mb_max_linear_groups = MB_DEFAULT_LINEAR_LIMIT; 3502 /* init file for buddy data */ 3503 ret = ext4_mb_init_backend(sb); 3504 if (ret != 0) 3505 goto out_free_locality_groups; 3506 3507 return 0; 3508 3509 out_free_locality_groups: 3510 free_percpu(sbi->s_locality_groups); 3511 sbi->s_locality_groups = NULL; 3512 out: 3513 kfree(sbi->s_mb_largest_free_orders); 3514 kfree(sbi->s_mb_largest_free_orders_locks); 3515 kfree(sbi->s_mb_offsets); 3516 sbi->s_mb_offsets = NULL; 3517 kfree(sbi->s_mb_maxs); 3518 sbi->s_mb_maxs = NULL; 3519 return ret; 3520 } 3521 3522 /* need to called with the ext4 group lock held */ 3523 static int ext4_mb_cleanup_pa(struct ext4_group_info *grp) 3524 { 3525 struct ext4_prealloc_space *pa; 3526 struct list_head *cur, *tmp; 3527 int count = 0; 3528 3529 list_for_each_safe(cur, tmp, &grp->bb_prealloc_list) { 3530 pa = list_entry(cur, struct ext4_prealloc_space, pa_group_list); 3531 list_del(&pa->pa_group_list); 3532 count++; 3533 kmem_cache_free(ext4_pspace_cachep, pa); 3534 } 3535 return count; 3536 } 3537 3538 int ext4_mb_release(struct super_block *sb) 3539 { 3540 ext4_group_t ngroups = ext4_get_groups_count(sb); 3541 ext4_group_t i; 3542 int num_meta_group_infos; 3543 struct ext4_group_info *grinfo, ***group_info; 3544 struct ext4_sb_info *sbi = EXT4_SB(sb); 3545 struct kmem_cache *cachep = get_groupinfo_cache(sb->s_blocksize_bits); 3546 int count; 3547 3548 if (test_opt(sb, DISCARD)) { 3549 /* 3550 * wait the discard work to drain all of ext4_free_data 3551 */ 3552 flush_work(&sbi->s_discard_work); 3553 WARN_ON_ONCE(!list_empty(&sbi->s_discard_list)); 3554 } 3555 3556 if (sbi->s_group_info) { 3557 for (i = 0; i < ngroups; i++) { 3558 cond_resched(); 3559 grinfo = ext4_get_group_info(sb, i); 3560 mb_group_bb_bitmap_free(grinfo); 3561 ext4_lock_group(sb, i); 3562 count = ext4_mb_cleanup_pa(grinfo); 3563 if (count) 3564 mb_debug(sb, "mballoc: %d PAs left\n", 3565 count); 3566 ext4_unlock_group(sb, i); 3567 kmem_cache_free(cachep, grinfo); 3568 } 3569 num_meta_group_infos = (ngroups + 3570 EXT4_DESC_PER_BLOCK(sb) - 1) >> 3571 EXT4_DESC_PER_BLOCK_BITS(sb); 3572 rcu_read_lock(); 3573 group_info = rcu_dereference(sbi->s_group_info); 3574 for (i = 0; i < num_meta_group_infos; i++) 3575 kfree(group_info[i]); 3576 kvfree(group_info); 3577 rcu_read_unlock(); 3578 } 3579 kfree(sbi->s_mb_largest_free_orders); 3580 kfree(sbi->s_mb_largest_free_orders_locks); 3581 kfree(sbi->s_mb_offsets); 3582 kfree(sbi->s_mb_maxs); 3583 iput(sbi->s_buddy_cache); 3584 if (sbi->s_mb_stats) { 3585 ext4_msg(sb, KERN_INFO, 3586 "mballoc: %u blocks %u reqs (%u success)", 3587 atomic_read(&sbi->s_bal_allocated), 3588 atomic_read(&sbi->s_bal_reqs), 3589 atomic_read(&sbi->s_bal_success)); 3590 ext4_msg(sb, KERN_INFO, 3591 "mballoc: %u extents scanned, %u groups scanned, %u goal hits, " 3592 "%u 2^N hits, %u breaks, %u lost", 3593 atomic_read(&sbi->s_bal_ex_scanned), 3594 atomic_read(&sbi->s_bal_groups_scanned), 3595 atomic_read(&sbi->s_bal_goals), 3596 atomic_read(&sbi->s_bal_2orders), 3597 atomic_read(&sbi->s_bal_breaks), 3598 atomic_read(&sbi->s_mb_lost_chunks)); 3599 ext4_msg(sb, KERN_INFO, 3600 "mballoc: %u generated and it took %llu", 3601 atomic_read(&sbi->s_mb_buddies_generated), 3602 atomic64_read(&sbi->s_mb_generation_time)); 3603 ext4_msg(sb, KERN_INFO, 3604 "mballoc: %u preallocated, %u discarded", 3605 atomic_read(&sbi->s_mb_preallocated), 3606 atomic_read(&sbi->s_mb_discarded)); 3607 } 3608 3609 free_percpu(sbi->s_locality_groups); 3610 3611 return 0; 3612 } 3613 3614 static inline int ext4_issue_discard(struct super_block *sb, 3615 ext4_group_t block_group, ext4_grpblk_t cluster, int count, 3616 struct bio **biop) 3617 { 3618 ext4_fsblk_t discard_block; 3619 3620 discard_block = (EXT4_C2B(EXT4_SB(sb), cluster) + 3621 ext4_group_first_block_no(sb, block_group)); 3622 count = EXT4_C2B(EXT4_SB(sb), count); 3623 trace_ext4_discard_blocks(sb, 3624 (unsigned long long) discard_block, count); 3625 if (biop) { 3626 return __blkdev_issue_discard(sb->s_bdev, 3627 (sector_t)discard_block << (sb->s_blocksize_bits - 9), 3628 (sector_t)count << (sb->s_blocksize_bits - 9), 3629 GFP_NOFS, biop); 3630 } else 3631 return sb_issue_discard(sb, discard_block, count, GFP_NOFS, 0); 3632 } 3633 3634 static void ext4_free_data_in_buddy(struct super_block *sb, 3635 struct ext4_free_data *entry) 3636 { 3637 struct ext4_buddy e4b; 3638 struct ext4_group_info *db; 3639 int err, count = 0, count2 = 0; 3640 3641 mb_debug(sb, "gonna free %u blocks in group %u (0x%p):", 3642 entry->efd_count, entry->efd_group, entry); 3643 3644 err = ext4_mb_load_buddy(sb, entry->efd_group, &e4b); 3645 /* we expect to find existing buddy because it's pinned */ 3646 BUG_ON(err != 0); 3647 3648 spin_lock(&EXT4_SB(sb)->s_md_lock); 3649 EXT4_SB(sb)->s_mb_free_pending -= entry->efd_count; 3650 spin_unlock(&EXT4_SB(sb)->s_md_lock); 3651 3652 db = e4b.bd_info; 3653 /* there are blocks to put in buddy to make them really free */ 3654 count += entry->efd_count; 3655 count2++; 3656 ext4_lock_group(sb, entry->efd_group); 3657 /* Take it out of per group rb tree */ 3658 rb_erase(&entry->efd_node, &(db->bb_free_root)); 3659 mb_free_blocks(NULL, &e4b, entry->efd_start_cluster, entry->efd_count); 3660 3661 /* 3662 * Clear the trimmed flag for the group so that the next 3663 * ext4_trim_fs can trim it. 3664 * If the volume is mounted with -o discard, online discard 3665 * is supported and the free blocks will be trimmed online. 3666 */ 3667 if (!test_opt(sb, DISCARD)) 3668 EXT4_MB_GRP_CLEAR_TRIMMED(db); 3669 3670 if (!db->bb_free_root.rb_node) { 3671 /* No more items in the per group rb tree 3672 * balance refcounts from ext4_mb_free_metadata() 3673 */ 3674 put_page(e4b.bd_buddy_page); 3675 put_page(e4b.bd_bitmap_page); 3676 } 3677 ext4_unlock_group(sb, entry->efd_group); 3678 ext4_mb_unload_buddy(&e4b); 3679 3680 mb_debug(sb, "freed %d blocks in %d structures\n", count, 3681 count2); 3682 } 3683 3684 /* 3685 * This function is called by the jbd2 layer once the commit has finished, 3686 * so we know we can free the blocks that were released with that commit. 3687 */ 3688 void ext4_process_freed_data(struct super_block *sb, tid_t commit_tid) 3689 { 3690 struct ext4_sb_info *sbi = EXT4_SB(sb); 3691 struct ext4_free_data *entry, *tmp; 3692 struct list_head freed_data_list; 3693 struct list_head *cut_pos = NULL; 3694 bool wake; 3695 3696 INIT_LIST_HEAD(&freed_data_list); 3697 3698 spin_lock(&sbi->s_md_lock); 3699 list_for_each_entry(entry, &sbi->s_freed_data_list, efd_list) { 3700 if (entry->efd_tid != commit_tid) 3701 break; 3702 cut_pos = &entry->efd_list; 3703 } 3704 if (cut_pos) 3705 list_cut_position(&freed_data_list, &sbi->s_freed_data_list, 3706 cut_pos); 3707 spin_unlock(&sbi->s_md_lock); 3708 3709 list_for_each_entry(entry, &freed_data_list, efd_list) 3710 ext4_free_data_in_buddy(sb, entry); 3711 3712 if (test_opt(sb, DISCARD)) { 3713 spin_lock(&sbi->s_md_lock); 3714 wake = list_empty(&sbi->s_discard_list); 3715 list_splice_tail(&freed_data_list, &sbi->s_discard_list); 3716 spin_unlock(&sbi->s_md_lock); 3717 if (wake) 3718 queue_work(system_unbound_wq, &sbi->s_discard_work); 3719 } else { 3720 list_for_each_entry_safe(entry, tmp, &freed_data_list, efd_list) 3721 kmem_cache_free(ext4_free_data_cachep, entry); 3722 } 3723 } 3724 3725 int __init ext4_init_mballoc(void) 3726 { 3727 ext4_pspace_cachep = KMEM_CACHE(ext4_prealloc_space, 3728 SLAB_RECLAIM_ACCOUNT); 3729 if (ext4_pspace_cachep == NULL) 3730 goto out; 3731 3732 ext4_ac_cachep = KMEM_CACHE(ext4_allocation_context, 3733 SLAB_RECLAIM_ACCOUNT); 3734 if (ext4_ac_cachep == NULL) 3735 goto out_pa_free; 3736 3737 ext4_free_data_cachep = KMEM_CACHE(ext4_free_data, 3738 SLAB_RECLAIM_ACCOUNT); 3739 if (ext4_free_data_cachep == NULL) 3740 goto out_ac_free; 3741 3742 return 0; 3743 3744 out_ac_free: 3745 kmem_cache_destroy(ext4_ac_cachep); 3746 out_pa_free: 3747 kmem_cache_destroy(ext4_pspace_cachep); 3748 out: 3749 return -ENOMEM; 3750 } 3751 3752 void ext4_exit_mballoc(void) 3753 { 3754 /* 3755 * Wait for completion of call_rcu()'s on ext4_pspace_cachep 3756 * before destroying the slab cache. 3757 */ 3758 rcu_barrier(); 3759 kmem_cache_destroy(ext4_pspace_cachep); 3760 kmem_cache_destroy(ext4_ac_cachep); 3761 kmem_cache_destroy(ext4_free_data_cachep); 3762 ext4_groupinfo_destroy_slabs(); 3763 } 3764 3765 3766 /* 3767 * Check quota and mark chosen space (ac->ac_b_ex) non-free in bitmaps 3768 * Returns 0 if success or error code 3769 */ 3770 static noinline_for_stack int 3771 ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac, 3772 handle_t *handle, unsigned int reserv_clstrs) 3773 { 3774 struct buffer_head *bitmap_bh = NULL; 3775 struct ext4_group_desc *gdp; 3776 struct buffer_head *gdp_bh; 3777 struct ext4_sb_info *sbi; 3778 struct super_block *sb; 3779 ext4_fsblk_t block; 3780 int err, len; 3781 3782 BUG_ON(ac->ac_status != AC_STATUS_FOUND); 3783 BUG_ON(ac->ac_b_ex.fe_len <= 0); 3784 3785 sb = ac->ac_sb; 3786 sbi = EXT4_SB(sb); 3787 3788 bitmap_bh = ext4_read_block_bitmap(sb, ac->ac_b_ex.fe_group); 3789 if (IS_ERR(bitmap_bh)) { 3790 err = PTR_ERR(bitmap_bh); 3791 bitmap_bh = NULL; 3792 goto out_err; 3793 } 3794 3795 BUFFER_TRACE(bitmap_bh, "getting write access"); 3796 err = ext4_journal_get_write_access(handle, sb, bitmap_bh, 3797 EXT4_JTR_NONE); 3798 if (err) 3799 goto out_err; 3800 3801 err = -EIO; 3802 gdp = ext4_get_group_desc(sb, ac->ac_b_ex.fe_group, &gdp_bh); 3803 if (!gdp) 3804 goto out_err; 3805 3806 ext4_debug("using block group %u(%d)\n", ac->ac_b_ex.fe_group, 3807 ext4_free_group_clusters(sb, gdp)); 3808 3809 BUFFER_TRACE(gdp_bh, "get_write_access"); 3810 err = ext4_journal_get_write_access(handle, sb, gdp_bh, EXT4_JTR_NONE); 3811 if (err) 3812 goto out_err; 3813 3814 block = ext4_grp_offs_to_block(sb, &ac->ac_b_ex); 3815 3816 len = EXT4_C2B(sbi, ac->ac_b_ex.fe_len); 3817 if (!ext4_inode_block_valid(ac->ac_inode, block, len)) { 3818 ext4_error(sb, "Allocating blocks %llu-%llu which overlap " 3819 "fs metadata", block, block+len); 3820 /* File system mounted not to panic on error 3821 * Fix the bitmap and return EFSCORRUPTED 3822 * We leak some of the blocks here. 3823 */ 3824 ext4_lock_group(sb, ac->ac_b_ex.fe_group); 3825 mb_set_bits(bitmap_bh->b_data, ac->ac_b_ex.fe_start, 3826 ac->ac_b_ex.fe_len); 3827 ext4_unlock_group(sb, ac->ac_b_ex.fe_group); 3828 err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh); 3829 if (!err) 3830 err = -EFSCORRUPTED; 3831 goto out_err; 3832 } 3833 3834 ext4_lock_group(sb, ac->ac_b_ex.fe_group); 3835 #ifdef AGGRESSIVE_CHECK 3836 { 3837 int i; 3838 for (i = 0; i < ac->ac_b_ex.fe_len; i++) { 3839 BUG_ON(mb_test_bit(ac->ac_b_ex.fe_start + i, 3840 bitmap_bh->b_data)); 3841 } 3842 } 3843 #endif 3844 mb_set_bits(bitmap_bh->b_data, ac->ac_b_ex.fe_start, 3845 ac->ac_b_ex.fe_len); 3846 if (ext4_has_group_desc_csum(sb) && 3847 (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT))) { 3848 gdp->bg_flags &= cpu_to_le16(~EXT4_BG_BLOCK_UNINIT); 3849 ext4_free_group_clusters_set(sb, gdp, 3850 ext4_free_clusters_after_init(sb, 3851 ac->ac_b_ex.fe_group, gdp)); 3852 } 3853 len = ext4_free_group_clusters(sb, gdp) - ac->ac_b_ex.fe_len; 3854 ext4_free_group_clusters_set(sb, gdp, len); 3855 ext4_block_bitmap_csum_set(sb, ac->ac_b_ex.fe_group, gdp, bitmap_bh); 3856 ext4_group_desc_csum_set(sb, ac->ac_b_ex.fe_group, gdp); 3857 3858 ext4_unlock_group(sb, ac->ac_b_ex.fe_group); 3859 percpu_counter_sub(&sbi->s_freeclusters_counter, ac->ac_b_ex.fe_len); 3860 /* 3861 * Now reduce the dirty block count also. Should not go negative 3862 */ 3863 if (!(ac->ac_flags & EXT4_MB_DELALLOC_RESERVED)) 3864 /* release all the reserved blocks if non delalloc */ 3865 percpu_counter_sub(&sbi->s_dirtyclusters_counter, 3866 reserv_clstrs); 3867 3868 if (sbi->s_log_groups_per_flex) { 3869 ext4_group_t flex_group = ext4_flex_group(sbi, 3870 ac->ac_b_ex.fe_group); 3871 atomic64_sub(ac->ac_b_ex.fe_len, 3872 &sbi_array_rcu_deref(sbi, s_flex_groups, 3873 flex_group)->free_clusters); 3874 } 3875 3876 err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh); 3877 if (err) 3878 goto out_err; 3879 err = ext4_handle_dirty_metadata(handle, NULL, gdp_bh); 3880 3881 out_err: 3882 brelse(bitmap_bh); 3883 return err; 3884 } 3885 3886 /* 3887 * Idempotent helper for Ext4 fast commit replay path to set the state of 3888 * blocks in bitmaps and update counters. 3889 */ 3890 void ext4_mb_mark_bb(struct super_block *sb, ext4_fsblk_t block, 3891 int len, int state) 3892 { 3893 struct buffer_head *bitmap_bh = NULL; 3894 struct ext4_group_desc *gdp; 3895 struct buffer_head *gdp_bh; 3896 struct ext4_sb_info *sbi = EXT4_SB(sb); 3897 ext4_group_t group; 3898 ext4_grpblk_t blkoff; 3899 int i, err; 3900 int already; 3901 unsigned int clen, clen_changed, thisgrp_len; 3902 3903 while (len > 0) { 3904 ext4_get_group_no_and_offset(sb, block, &group, &blkoff); 3905 3906 /* 3907 * Check to see if we are freeing blocks across a group 3908 * boundary. 3909 * In case of flex_bg, this can happen that (block, len) may 3910 * span across more than one group. In that case we need to 3911 * get the corresponding group metadata to work with. 3912 * For this we have goto again loop. 3913 */ 3914 thisgrp_len = min_t(unsigned int, (unsigned int)len, 3915 EXT4_BLOCKS_PER_GROUP(sb) - EXT4_C2B(sbi, blkoff)); 3916 clen = EXT4_NUM_B2C(sbi, thisgrp_len); 3917 3918 if (!ext4_sb_block_valid(sb, NULL, block, thisgrp_len)) { 3919 ext4_error(sb, "Marking blocks in system zone - " 3920 "Block = %llu, len = %u", 3921 block, thisgrp_len); 3922 bitmap_bh = NULL; 3923 break; 3924 } 3925 3926 bitmap_bh = ext4_read_block_bitmap(sb, group); 3927 if (IS_ERR(bitmap_bh)) { 3928 err = PTR_ERR(bitmap_bh); 3929 bitmap_bh = NULL; 3930 break; 3931 } 3932 3933 err = -EIO; 3934 gdp = ext4_get_group_desc(sb, group, &gdp_bh); 3935 if (!gdp) 3936 break; 3937 3938 ext4_lock_group(sb, group); 3939 already = 0; 3940 for (i = 0; i < clen; i++) 3941 if (!mb_test_bit(blkoff + i, bitmap_bh->b_data) == 3942 !state) 3943 already++; 3944 3945 clen_changed = clen - already; 3946 if (state) 3947 mb_set_bits(bitmap_bh->b_data, blkoff, clen); 3948 else 3949 mb_clear_bits(bitmap_bh->b_data, blkoff, clen); 3950 if (ext4_has_group_desc_csum(sb) && 3951 (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT))) { 3952 gdp->bg_flags &= cpu_to_le16(~EXT4_BG_BLOCK_UNINIT); 3953 ext4_free_group_clusters_set(sb, gdp, 3954 ext4_free_clusters_after_init(sb, group, gdp)); 3955 } 3956 if (state) 3957 clen = ext4_free_group_clusters(sb, gdp) - clen_changed; 3958 else 3959 clen = ext4_free_group_clusters(sb, gdp) + clen_changed; 3960 3961 ext4_free_group_clusters_set(sb, gdp, clen); 3962 ext4_block_bitmap_csum_set(sb, group, gdp, bitmap_bh); 3963 ext4_group_desc_csum_set(sb, group, gdp); 3964 3965 ext4_unlock_group(sb, group); 3966 3967 if (sbi->s_log_groups_per_flex) { 3968 ext4_group_t flex_group = ext4_flex_group(sbi, group); 3969 struct flex_groups *fg = sbi_array_rcu_deref(sbi, 3970 s_flex_groups, flex_group); 3971 3972 if (state) 3973 atomic64_sub(clen_changed, &fg->free_clusters); 3974 else 3975 atomic64_add(clen_changed, &fg->free_clusters); 3976 3977 } 3978 3979 err = ext4_handle_dirty_metadata(NULL, NULL, bitmap_bh); 3980 if (err) 3981 break; 3982 sync_dirty_buffer(bitmap_bh); 3983 err = ext4_handle_dirty_metadata(NULL, NULL, gdp_bh); 3984 sync_dirty_buffer(gdp_bh); 3985 if (err) 3986 break; 3987 3988 block += thisgrp_len; 3989 len -= thisgrp_len; 3990 brelse(bitmap_bh); 3991 BUG_ON(len < 0); 3992 } 3993 3994 if (err) 3995 brelse(bitmap_bh); 3996 } 3997 3998 /* 3999 * here we normalize request for locality group 4000 * Group request are normalized to s_mb_group_prealloc, which goes to 4001 * s_strip if we set the same via mount option. 4002 * s_mb_group_prealloc can be configured via 4003 * /sys/fs/ext4/<partition>/mb_group_prealloc 4004 * 4005 * XXX: should we try to preallocate more than the group has now? 4006 */ 4007 static void ext4_mb_normalize_group_request(struct ext4_allocation_context *ac) 4008 { 4009 struct super_block *sb = ac->ac_sb; 4010 struct ext4_locality_group *lg = ac->ac_lg; 4011 4012 BUG_ON(lg == NULL); 4013 ac->ac_g_ex.fe_len = EXT4_SB(sb)->s_mb_group_prealloc; 4014 mb_debug(sb, "goal %u blocks for locality group\n", ac->ac_g_ex.fe_len); 4015 } 4016 4017 /* 4018 * Normalization means making request better in terms of 4019 * size and alignment 4020 */ 4021 static noinline_for_stack void 4022 ext4_mb_normalize_request(struct ext4_allocation_context *ac, 4023 struct ext4_allocation_request *ar) 4024 { 4025 struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); 4026 int bsbits, max; 4027 ext4_lblk_t end; 4028 loff_t size, start_off; 4029 loff_t orig_size __maybe_unused; 4030 ext4_lblk_t start; 4031 struct ext4_inode_info *ei = EXT4_I(ac->ac_inode); 4032 struct ext4_prealloc_space *pa; 4033 4034 /* do normalize only data requests, metadata requests 4035 do not need preallocation */ 4036 if (!(ac->ac_flags & EXT4_MB_HINT_DATA)) 4037 return; 4038 4039 /* sometime caller may want exact blocks */ 4040 if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY)) 4041 return; 4042 4043 /* caller may indicate that preallocation isn't 4044 * required (it's a tail, for example) */ 4045 if (ac->ac_flags & EXT4_MB_HINT_NOPREALLOC) 4046 return; 4047 4048 if (ac->ac_flags & EXT4_MB_HINT_GROUP_ALLOC) { 4049 ext4_mb_normalize_group_request(ac); 4050 return ; 4051 } 4052 4053 bsbits = ac->ac_sb->s_blocksize_bits; 4054 4055 /* first, let's learn actual file size 4056 * given current request is allocated */ 4057 size = ac->ac_o_ex.fe_logical + EXT4_C2B(sbi, ac->ac_o_ex.fe_len); 4058 size = size << bsbits; 4059 if (size < i_size_read(ac->ac_inode)) 4060 size = i_size_read(ac->ac_inode); 4061 orig_size = size; 4062 4063 /* max size of free chunks */ 4064 max = 2 << bsbits; 4065 4066 #define NRL_CHECK_SIZE(req, size, max, chunk_size) \ 4067 (req <= (size) || max <= (chunk_size)) 4068 4069 /* first, try to predict filesize */ 4070 /* XXX: should this table be tunable? */ 4071 start_off = 0; 4072 if (size <= 16 * 1024) { 4073 size = 16 * 1024; 4074 } else if (size <= 32 * 1024) { 4075 size = 32 * 1024; 4076 } else if (size <= 64 * 1024) { 4077 size = 64 * 1024; 4078 } else if (size <= 128 * 1024) { 4079 size = 128 * 1024; 4080 } else if (size <= 256 * 1024) { 4081 size = 256 * 1024; 4082 } else if (size <= 512 * 1024) { 4083 size = 512 * 1024; 4084 } else if (size <= 1024 * 1024) { 4085 size = 1024 * 1024; 4086 } else if (NRL_CHECK_SIZE(size, 4 * 1024 * 1024, max, 2 * 1024)) { 4087 start_off = ((loff_t)ac->ac_o_ex.fe_logical >> 4088 (21 - bsbits)) << 21; 4089 size = 2 * 1024 * 1024; 4090 } else if (NRL_CHECK_SIZE(size, 8 * 1024 * 1024, max, 4 * 1024)) { 4091 start_off = ((loff_t)ac->ac_o_ex.fe_logical >> 4092 (22 - bsbits)) << 22; 4093 size = 4 * 1024 * 1024; 4094 } else if (NRL_CHECK_SIZE(ac->ac_o_ex.fe_len, 4095 (8<<20)>>bsbits, max, 8 * 1024)) { 4096 start_off = ((loff_t)ac->ac_o_ex.fe_logical >> 4097 (23 - bsbits)) << 23; 4098 size = 8 * 1024 * 1024; 4099 } else { 4100 start_off = (loff_t) ac->ac_o_ex.fe_logical << bsbits; 4101 size = (loff_t) EXT4_C2B(EXT4_SB(ac->ac_sb), 4102 ac->ac_o_ex.fe_len) << bsbits; 4103 } 4104 size = size >> bsbits; 4105 start = start_off >> bsbits; 4106 4107 /* 4108 * For tiny groups (smaller than 8MB) the chosen allocation 4109 * alignment may be larger than group size. Make sure the 4110 * alignment does not move allocation to a different group which 4111 * makes mballoc fail assertions later. 4112 */ 4113 start = max(start, rounddown(ac->ac_o_ex.fe_logical, 4114 (ext4_lblk_t)EXT4_BLOCKS_PER_GROUP(ac->ac_sb))); 4115 4116 /* don't cover already allocated blocks in selected range */ 4117 if (ar->pleft && start <= ar->lleft) { 4118 size -= ar->lleft + 1 - start; 4119 start = ar->lleft + 1; 4120 } 4121 if (ar->pright && start + size - 1 >= ar->lright) 4122 size -= start + size - ar->lright; 4123 4124 /* 4125 * Trim allocation request for filesystems with artificially small 4126 * groups. 4127 */ 4128 if (size > EXT4_BLOCKS_PER_GROUP(ac->ac_sb)) 4129 size = EXT4_BLOCKS_PER_GROUP(ac->ac_sb); 4130 4131 end = start + size; 4132 4133 /* check we don't cross already preallocated blocks */ 4134 rcu_read_lock(); 4135 list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) { 4136 ext4_lblk_t pa_end; 4137 4138 if (pa->pa_deleted) 4139 continue; 4140 spin_lock(&pa->pa_lock); 4141 if (pa->pa_deleted) { 4142 spin_unlock(&pa->pa_lock); 4143 continue; 4144 } 4145 4146 pa_end = pa->pa_lstart + EXT4_C2B(EXT4_SB(ac->ac_sb), 4147 pa->pa_len); 4148 4149 /* PA must not overlap original request */ 4150 BUG_ON(!(ac->ac_o_ex.fe_logical >= pa_end || 4151 ac->ac_o_ex.fe_logical < pa->pa_lstart)); 4152 4153 /* skip PAs this normalized request doesn't overlap with */ 4154 if (pa->pa_lstart >= end || pa_end <= start) { 4155 spin_unlock(&pa->pa_lock); 4156 continue; 4157 } 4158 BUG_ON(pa->pa_lstart <= start && pa_end >= end); 4159 4160 /* adjust start or end to be adjacent to this pa */ 4161 if (pa_end <= ac->ac_o_ex.fe_logical) { 4162 BUG_ON(pa_end < start); 4163 start = pa_end; 4164 } else if (pa->pa_lstart > ac->ac_o_ex.fe_logical) { 4165 BUG_ON(pa->pa_lstart > end); 4166 end = pa->pa_lstart; 4167 } 4168 spin_unlock(&pa->pa_lock); 4169 } 4170 rcu_read_unlock(); 4171 size = end - start; 4172 4173 /* XXX: extra loop to check we really don't overlap preallocations */ 4174 rcu_read_lock(); 4175 list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) { 4176 ext4_lblk_t pa_end; 4177 4178 spin_lock(&pa->pa_lock); 4179 if (pa->pa_deleted == 0) { 4180 pa_end = pa->pa_lstart + EXT4_C2B(EXT4_SB(ac->ac_sb), 4181 pa->pa_len); 4182 BUG_ON(!(start >= pa_end || end <= pa->pa_lstart)); 4183 } 4184 spin_unlock(&pa->pa_lock); 4185 } 4186 rcu_read_unlock(); 4187 4188 /* 4189 * In this function "start" and "size" are normalized for better 4190 * alignment and length such that we could preallocate more blocks. 4191 * This normalization is done such that original request of 4192 * ac->ac_o_ex.fe_logical & fe_len should always lie within "start" and 4193 * "size" boundaries. 4194 * (Note fe_len can be relaxed since FS block allocation API does not 4195 * provide gurantee on number of contiguous blocks allocation since that 4196 * depends upon free space left, etc). 4197 * In case of inode pa, later we use the allocated blocks 4198 * [pa_start + fe_logical - pa_lstart, fe_len/size] from the preallocated 4199 * range of goal/best blocks [start, size] to put it at the 4200 * ac_o_ex.fe_logical extent of this inode. 4201 * (See ext4_mb_use_inode_pa() for more details) 4202 */ 4203 if (start + size <= ac->ac_o_ex.fe_logical || 4204 start > ac->ac_o_ex.fe_logical) { 4205 ext4_msg(ac->ac_sb, KERN_ERR, 4206 "start %lu, size %lu, fe_logical %lu", 4207 (unsigned long) start, (unsigned long) size, 4208 (unsigned long) ac->ac_o_ex.fe_logical); 4209 BUG(); 4210 } 4211 BUG_ON(size <= 0 || size > EXT4_BLOCKS_PER_GROUP(ac->ac_sb)); 4212 4213 /* now prepare goal request */ 4214 4215 /* XXX: is it better to align blocks WRT to logical 4216 * placement or satisfy big request as is */ 4217 ac->ac_g_ex.fe_logical = start; 4218 ac->ac_g_ex.fe_len = EXT4_NUM_B2C(sbi, size); 4219 4220 /* define goal start in order to merge */ 4221 if (ar->pright && (ar->lright == (start + size))) { 4222 /* merge to the right */ 4223 ext4_get_group_no_and_offset(ac->ac_sb, ar->pright - size, 4224 &ac->ac_f_ex.fe_group, 4225 &ac->ac_f_ex.fe_start); 4226 ac->ac_flags |= EXT4_MB_HINT_TRY_GOAL; 4227 } 4228 if (ar->pleft && (ar->lleft + 1 == start)) { 4229 /* merge to the left */ 4230 ext4_get_group_no_and_offset(ac->ac_sb, ar->pleft + 1, 4231 &ac->ac_f_ex.fe_group, 4232 &ac->ac_f_ex.fe_start); 4233 ac->ac_flags |= EXT4_MB_HINT_TRY_GOAL; 4234 } 4235 4236 mb_debug(ac->ac_sb, "goal: %lld(was %lld) blocks at %u\n", size, 4237 orig_size, start); 4238 } 4239 4240 static void ext4_mb_collect_stats(struct ext4_allocation_context *ac) 4241 { 4242 struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); 4243 4244 if (sbi->s_mb_stats && ac->ac_g_ex.fe_len >= 1) { 4245 atomic_inc(&sbi->s_bal_reqs); 4246 atomic_add(ac->ac_b_ex.fe_len, &sbi->s_bal_allocated); 4247 if (ac->ac_b_ex.fe_len >= ac->ac_o_ex.fe_len) 4248 atomic_inc(&sbi->s_bal_success); 4249 atomic_add(ac->ac_found, &sbi->s_bal_ex_scanned); 4250 atomic_add(ac->ac_groups_scanned, &sbi->s_bal_groups_scanned); 4251 if (ac->ac_g_ex.fe_start == ac->ac_b_ex.fe_start && 4252 ac->ac_g_ex.fe_group == ac->ac_b_ex.fe_group) 4253 atomic_inc(&sbi->s_bal_goals); 4254 if (ac->ac_found > sbi->s_mb_max_to_scan) 4255 atomic_inc(&sbi->s_bal_breaks); 4256 } 4257 4258 if (ac->ac_op == EXT4_MB_HISTORY_ALLOC) 4259 trace_ext4_mballoc_alloc(ac); 4260 else 4261 trace_ext4_mballoc_prealloc(ac); 4262 } 4263 4264 /* 4265 * Called on failure; free up any blocks from the inode PA for this 4266 * context. We don't need this for MB_GROUP_PA because we only change 4267 * pa_free in ext4_mb_release_context(), but on failure, we've already 4268 * zeroed out ac->ac_b_ex.fe_len, so group_pa->pa_free is not changed. 4269 */ 4270 static void ext4_discard_allocated_blocks(struct ext4_allocation_context *ac) 4271 { 4272 struct ext4_prealloc_space *pa = ac->ac_pa; 4273 struct ext4_buddy e4b; 4274 int err; 4275 4276 if (pa == NULL) { 4277 if (ac->ac_f_ex.fe_len == 0) 4278 return; 4279 err = ext4_mb_load_buddy(ac->ac_sb, ac->ac_f_ex.fe_group, &e4b); 4280 if (err) { 4281 /* 4282 * This should never happen since we pin the 4283 * pages in the ext4_allocation_context so 4284 * ext4_mb_load_buddy() should never fail. 4285 */ 4286 WARN(1, "mb_load_buddy failed (%d)", err); 4287 return; 4288 } 4289 ext4_lock_group(ac->ac_sb, ac->ac_f_ex.fe_group); 4290 mb_free_blocks(ac->ac_inode, &e4b, ac->ac_f_ex.fe_start, 4291 ac->ac_f_ex.fe_len); 4292 ext4_unlock_group(ac->ac_sb, ac->ac_f_ex.fe_group); 4293 ext4_mb_unload_buddy(&e4b); 4294 return; 4295 } 4296 if (pa->pa_type == MB_INODE_PA) 4297 pa->pa_free += ac->ac_b_ex.fe_len; 4298 } 4299 4300 /* 4301 * use blocks preallocated to inode 4302 */ 4303 static void ext4_mb_use_inode_pa(struct ext4_allocation_context *ac, 4304 struct ext4_prealloc_space *pa) 4305 { 4306 struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); 4307 ext4_fsblk_t start; 4308 ext4_fsblk_t end; 4309 int len; 4310 4311 /* found preallocated blocks, use them */ 4312 start = pa->pa_pstart + (ac->ac_o_ex.fe_logical - pa->pa_lstart); 4313 end = min(pa->pa_pstart + EXT4_C2B(sbi, pa->pa_len), 4314 start + EXT4_C2B(sbi, ac->ac_o_ex.fe_len)); 4315 len = EXT4_NUM_B2C(sbi, end - start); 4316 ext4_get_group_no_and_offset(ac->ac_sb, start, &ac->ac_b_ex.fe_group, 4317 &ac->ac_b_ex.fe_start); 4318 ac->ac_b_ex.fe_len = len; 4319 ac->ac_status = AC_STATUS_FOUND; 4320 ac->ac_pa = pa; 4321 4322 BUG_ON(start < pa->pa_pstart); 4323 BUG_ON(end > pa->pa_pstart + EXT4_C2B(sbi, pa->pa_len)); 4324 BUG_ON(pa->pa_free < len); 4325 pa->pa_free -= len; 4326 4327 mb_debug(ac->ac_sb, "use %llu/%d from inode pa %p\n", start, len, pa); 4328 } 4329 4330 /* 4331 * use blocks preallocated to locality group 4332 */ 4333 static void ext4_mb_use_group_pa(struct ext4_allocation_context *ac, 4334 struct ext4_prealloc_space *pa) 4335 { 4336 unsigned int len = ac->ac_o_ex.fe_len; 4337 4338 ext4_get_group_no_and_offset(ac->ac_sb, pa->pa_pstart, 4339 &ac->ac_b_ex.fe_group, 4340 &ac->ac_b_ex.fe_start); 4341 ac->ac_b_ex.fe_len = len; 4342 ac->ac_status = AC_STATUS_FOUND; 4343 ac->ac_pa = pa; 4344 4345 /* we don't correct pa_pstart or pa_plen here to avoid 4346 * possible race when the group is being loaded concurrently 4347 * instead we correct pa later, after blocks are marked 4348 * in on-disk bitmap -- see ext4_mb_release_context() 4349 * Other CPUs are prevented from allocating from this pa by lg_mutex 4350 */ 4351 mb_debug(ac->ac_sb, "use %u/%u from group pa %p\n", 4352 pa->pa_lstart-len, len, pa); 4353 } 4354 4355 /* 4356 * Return the prealloc space that have minimal distance 4357 * from the goal block. @cpa is the prealloc 4358 * space that is having currently known minimal distance 4359 * from the goal block. 4360 */ 4361 static struct ext4_prealloc_space * 4362 ext4_mb_check_group_pa(ext4_fsblk_t goal_block, 4363 struct ext4_prealloc_space *pa, 4364 struct ext4_prealloc_space *cpa) 4365 { 4366 ext4_fsblk_t cur_distance, new_distance; 4367 4368 if (cpa == NULL) { 4369 atomic_inc(&pa->pa_count); 4370 return pa; 4371 } 4372 cur_distance = abs(goal_block - cpa->pa_pstart); 4373 new_distance = abs(goal_block - pa->pa_pstart); 4374 4375 if (cur_distance <= new_distance) 4376 return cpa; 4377 4378 /* drop the previous reference */ 4379 atomic_dec(&cpa->pa_count); 4380 atomic_inc(&pa->pa_count); 4381 return pa; 4382 } 4383 4384 /* 4385 * search goal blocks in preallocated space 4386 */ 4387 static noinline_for_stack bool 4388 ext4_mb_use_preallocated(struct ext4_allocation_context *ac) 4389 { 4390 struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); 4391 int order, i; 4392 struct ext4_inode_info *ei = EXT4_I(ac->ac_inode); 4393 struct ext4_locality_group *lg; 4394 struct ext4_prealloc_space *pa, *cpa = NULL; 4395 ext4_fsblk_t goal_block; 4396 4397 /* only data can be preallocated */ 4398 if (!(ac->ac_flags & EXT4_MB_HINT_DATA)) 4399 return false; 4400 4401 /* first, try per-file preallocation */ 4402 rcu_read_lock(); 4403 list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) { 4404 4405 /* all fields in this condition don't change, 4406 * so we can skip locking for them */ 4407 if (ac->ac_o_ex.fe_logical < pa->pa_lstart || 4408 ac->ac_o_ex.fe_logical >= (pa->pa_lstart + 4409 EXT4_C2B(sbi, pa->pa_len))) 4410 continue; 4411 4412 /* non-extent files can't have physical blocks past 2^32 */ 4413 if (!(ext4_test_inode_flag(ac->ac_inode, EXT4_INODE_EXTENTS)) && 4414 (pa->pa_pstart + EXT4_C2B(sbi, pa->pa_len) > 4415 EXT4_MAX_BLOCK_FILE_PHYS)) 4416 continue; 4417 4418 /* found preallocated blocks, use them */ 4419 spin_lock(&pa->pa_lock); 4420 if (pa->pa_deleted == 0 && pa->pa_free) { 4421 atomic_inc(&pa->pa_count); 4422 ext4_mb_use_inode_pa(ac, pa); 4423 spin_unlock(&pa->pa_lock); 4424 ac->ac_criteria = 10; 4425 rcu_read_unlock(); 4426 return true; 4427 } 4428 spin_unlock(&pa->pa_lock); 4429 } 4430 rcu_read_unlock(); 4431 4432 /* can we use group allocation? */ 4433 if (!(ac->ac_flags & EXT4_MB_HINT_GROUP_ALLOC)) 4434 return false; 4435 4436 /* inode may have no locality group for some reason */ 4437 lg = ac->ac_lg; 4438 if (lg == NULL) 4439 return false; 4440 order = fls(ac->ac_o_ex.fe_len) - 1; 4441 if (order > PREALLOC_TB_SIZE - 1) 4442 /* The max size of hash table is PREALLOC_TB_SIZE */ 4443 order = PREALLOC_TB_SIZE - 1; 4444 4445 goal_block = ext4_grp_offs_to_block(ac->ac_sb, &ac->ac_g_ex); 4446 /* 4447 * search for the prealloc space that is having 4448 * minimal distance from the goal block. 4449 */ 4450 for (i = order; i < PREALLOC_TB_SIZE; i++) { 4451 rcu_read_lock(); 4452 list_for_each_entry_rcu(pa, &lg->lg_prealloc_list[i], 4453 pa_inode_list) { 4454 spin_lock(&pa->pa_lock); 4455 if (pa->pa_deleted == 0 && 4456 pa->pa_free >= ac->ac_o_ex.fe_len) { 4457 4458 cpa = ext4_mb_check_group_pa(goal_block, 4459 pa, cpa); 4460 } 4461 spin_unlock(&pa->pa_lock); 4462 } 4463 rcu_read_unlock(); 4464 } 4465 if (cpa) { 4466 ext4_mb_use_group_pa(ac, cpa); 4467 ac->ac_criteria = 20; 4468 return true; 4469 } 4470 return false; 4471 } 4472 4473 /* 4474 * the function goes through all block freed in the group 4475 * but not yet committed and marks them used in in-core bitmap. 4476 * buddy must be generated from this bitmap 4477 * Need to be called with the ext4 group lock held 4478 */ 4479 static void ext4_mb_generate_from_freelist(struct super_block *sb, void *bitmap, 4480 ext4_group_t group) 4481 { 4482 struct rb_node *n; 4483 struct ext4_group_info *grp; 4484 struct ext4_free_data *entry; 4485 4486 grp = ext4_get_group_info(sb, group); 4487 n = rb_first(&(grp->bb_free_root)); 4488 4489 while (n) { 4490 entry = rb_entry(n, struct ext4_free_data, efd_node); 4491 mb_set_bits(bitmap, entry->efd_start_cluster, entry->efd_count); 4492 n = rb_next(n); 4493 } 4494 return; 4495 } 4496 4497 /* 4498 * the function goes through all preallocation in this group and marks them 4499 * used in in-core bitmap. buddy must be generated from this bitmap 4500 * Need to be called with ext4 group lock held 4501 */ 4502 static noinline_for_stack 4503 void ext4_mb_generate_from_pa(struct super_block *sb, void *bitmap, 4504 ext4_group_t group) 4505 { 4506 struct ext4_group_info *grp = ext4_get_group_info(sb, group); 4507 struct ext4_prealloc_space *pa; 4508 struct list_head *cur; 4509 ext4_group_t groupnr; 4510 ext4_grpblk_t start; 4511 int preallocated = 0; 4512 int len; 4513 4514 /* all form of preallocation discards first load group, 4515 * so the only competing code is preallocation use. 4516 * we don't need any locking here 4517 * notice we do NOT ignore preallocations with pa_deleted 4518 * otherwise we could leave used blocks available for 4519 * allocation in buddy when concurrent ext4_mb_put_pa() 4520 * is dropping preallocation 4521 */ 4522 list_for_each(cur, &grp->bb_prealloc_list) { 4523 pa = list_entry(cur, struct ext4_prealloc_space, pa_group_list); 4524 spin_lock(&pa->pa_lock); 4525 ext4_get_group_no_and_offset(sb, pa->pa_pstart, 4526 &groupnr, &start); 4527 len = pa->pa_len; 4528 spin_unlock(&pa->pa_lock); 4529 if (unlikely(len == 0)) 4530 continue; 4531 BUG_ON(groupnr != group); 4532 mb_set_bits(bitmap, start, len); 4533 preallocated += len; 4534 } 4535 mb_debug(sb, "preallocated %d for group %u\n", preallocated, group); 4536 } 4537 4538 static void ext4_mb_mark_pa_deleted(struct super_block *sb, 4539 struct ext4_prealloc_space *pa) 4540 { 4541 struct ext4_inode_info *ei; 4542 4543 if (pa->pa_deleted) { 4544 ext4_warning(sb, "deleted pa, type:%d, pblk:%llu, lblk:%u, len:%d\n", 4545 pa->pa_type, pa->pa_pstart, pa->pa_lstart, 4546 pa->pa_len); 4547 return; 4548 } 4549 4550 pa->pa_deleted = 1; 4551 4552 if (pa->pa_type == MB_INODE_PA) { 4553 ei = EXT4_I(pa->pa_inode); 4554 atomic_dec(&ei->i_prealloc_active); 4555 } 4556 } 4557 4558 static void ext4_mb_pa_callback(struct rcu_head *head) 4559 { 4560 struct ext4_prealloc_space *pa; 4561 pa = container_of(head, struct ext4_prealloc_space, u.pa_rcu); 4562 4563 BUG_ON(atomic_read(&pa->pa_count)); 4564 BUG_ON(pa->pa_deleted == 0); 4565 kmem_cache_free(ext4_pspace_cachep, pa); 4566 } 4567 4568 /* 4569 * drops a reference to preallocated space descriptor 4570 * if this was the last reference and the space is consumed 4571 */ 4572 static void ext4_mb_put_pa(struct ext4_allocation_context *ac, 4573 struct super_block *sb, struct ext4_prealloc_space *pa) 4574 { 4575 ext4_group_t grp; 4576 ext4_fsblk_t grp_blk; 4577 4578 /* in this short window concurrent discard can set pa_deleted */ 4579 spin_lock(&pa->pa_lock); 4580 if (!atomic_dec_and_test(&pa->pa_count) || pa->pa_free != 0) { 4581 spin_unlock(&pa->pa_lock); 4582 return; 4583 } 4584 4585 if (pa->pa_deleted == 1) { 4586 spin_unlock(&pa->pa_lock); 4587 return; 4588 } 4589 4590 ext4_mb_mark_pa_deleted(sb, pa); 4591 spin_unlock(&pa->pa_lock); 4592 4593 grp_blk = pa->pa_pstart; 4594 /* 4595 * If doing group-based preallocation, pa_pstart may be in the 4596 * next group when pa is used up 4597 */ 4598 if (pa->pa_type == MB_GROUP_PA) 4599 grp_blk--; 4600 4601 grp = ext4_get_group_number(sb, grp_blk); 4602 4603 /* 4604 * possible race: 4605 * 4606 * P1 (buddy init) P2 (regular allocation) 4607 * find block B in PA 4608 * copy on-disk bitmap to buddy 4609 * mark B in on-disk bitmap 4610 * drop PA from group 4611 * mark all PAs in buddy 4612 * 4613 * thus, P1 initializes buddy with B available. to prevent this 4614 * we make "copy" and "mark all PAs" atomic and serialize "drop PA" 4615 * against that pair 4616 */ 4617 ext4_lock_group(sb, grp); 4618 list_del(&pa->pa_group_list); 4619 ext4_unlock_group(sb, grp); 4620 4621 spin_lock(pa->pa_obj_lock); 4622 list_del_rcu(&pa->pa_inode_list); 4623 spin_unlock(pa->pa_obj_lock); 4624 4625 call_rcu(&(pa)->u.pa_rcu, ext4_mb_pa_callback); 4626 } 4627 4628 /* 4629 * creates new preallocated space for given inode 4630 */ 4631 static noinline_for_stack void 4632 ext4_mb_new_inode_pa(struct ext4_allocation_context *ac) 4633 { 4634 struct super_block *sb = ac->ac_sb; 4635 struct ext4_sb_info *sbi = EXT4_SB(sb); 4636 struct ext4_prealloc_space *pa; 4637 struct ext4_group_info *grp; 4638 struct ext4_inode_info *ei; 4639 4640 /* preallocate only when found space is larger then requested */ 4641 BUG_ON(ac->ac_o_ex.fe_len >= ac->ac_b_ex.fe_len); 4642 BUG_ON(ac->ac_status != AC_STATUS_FOUND); 4643 BUG_ON(!S_ISREG(ac->ac_inode->i_mode)); 4644 BUG_ON(ac->ac_pa == NULL); 4645 4646 pa = ac->ac_pa; 4647 4648 if (ac->ac_b_ex.fe_len < ac->ac_g_ex.fe_len) { 4649 int winl; 4650 int wins; 4651 int win; 4652 int offs; 4653 4654 /* we can't allocate as much as normalizer wants. 4655 * so, found space must get proper lstart 4656 * to cover original request */ 4657 BUG_ON(ac->ac_g_ex.fe_logical > ac->ac_o_ex.fe_logical); 4658 BUG_ON(ac->ac_g_ex.fe_len < ac->ac_o_ex.fe_len); 4659 4660 /* we're limited by original request in that 4661 * logical block must be covered any way 4662 * winl is window we can move our chunk within */ 4663 winl = ac->ac_o_ex.fe_logical - ac->ac_g_ex.fe_logical; 4664 4665 /* also, we should cover whole original request */ 4666 wins = EXT4_C2B(sbi, ac->ac_b_ex.fe_len - ac->ac_o_ex.fe_len); 4667 4668 /* the smallest one defines real window */ 4669 win = min(winl, wins); 4670 4671 offs = ac->ac_o_ex.fe_logical % 4672 EXT4_C2B(sbi, ac->ac_b_ex.fe_len); 4673 if (offs && offs < win) 4674 win = offs; 4675 4676 ac->ac_b_ex.fe_logical = ac->ac_o_ex.fe_logical - 4677 EXT4_NUM_B2C(sbi, win); 4678 BUG_ON(ac->ac_o_ex.fe_logical < ac->ac_b_ex.fe_logical); 4679 BUG_ON(ac->ac_o_ex.fe_len > ac->ac_b_ex.fe_len); 4680 } 4681 4682 /* preallocation can change ac_b_ex, thus we store actually 4683 * allocated blocks for history */ 4684 ac->ac_f_ex = ac->ac_b_ex; 4685 4686 pa->pa_lstart = ac->ac_b_ex.fe_logical; 4687 pa->pa_pstart = ext4_grp_offs_to_block(sb, &ac->ac_b_ex); 4688 pa->pa_len = ac->ac_b_ex.fe_len; 4689 pa->pa_free = pa->pa_len; 4690 spin_lock_init(&pa->pa_lock); 4691 INIT_LIST_HEAD(&pa->pa_inode_list); 4692 INIT_LIST_HEAD(&pa->pa_group_list); 4693 pa->pa_deleted = 0; 4694 pa->pa_type = MB_INODE_PA; 4695 4696 mb_debug(sb, "new inode pa %p: %llu/%d for %u\n", pa, pa->pa_pstart, 4697 pa->pa_len, pa->pa_lstart); 4698 trace_ext4_mb_new_inode_pa(ac, pa); 4699 4700 ext4_mb_use_inode_pa(ac, pa); 4701 atomic_add(pa->pa_free, &sbi->s_mb_preallocated); 4702 4703 ei = EXT4_I(ac->ac_inode); 4704 grp = ext4_get_group_info(sb, ac->ac_b_ex.fe_group); 4705 4706 pa->pa_obj_lock = &ei->i_prealloc_lock; 4707 pa->pa_inode = ac->ac_inode; 4708 4709 list_add(&pa->pa_group_list, &grp->bb_prealloc_list); 4710 4711 spin_lock(pa->pa_obj_lock); 4712 list_add_rcu(&pa->pa_inode_list, &ei->i_prealloc_list); 4713 spin_unlock(pa->pa_obj_lock); 4714 atomic_inc(&ei->i_prealloc_active); 4715 } 4716 4717 /* 4718 * creates new preallocated space for locality group inodes belongs to 4719 */ 4720 static noinline_for_stack void 4721 ext4_mb_new_group_pa(struct ext4_allocation_context *ac) 4722 { 4723 struct super_block *sb = ac->ac_sb; 4724 struct ext4_locality_group *lg; 4725 struct ext4_prealloc_space *pa; 4726 struct ext4_group_info *grp; 4727 4728 /* preallocate only when found space is larger then requested */ 4729 BUG_ON(ac->ac_o_ex.fe_len >= ac->ac_b_ex.fe_len); 4730 BUG_ON(ac->ac_status != AC_STATUS_FOUND); 4731 BUG_ON(!S_ISREG(ac->ac_inode->i_mode)); 4732 BUG_ON(ac->ac_pa == NULL); 4733 4734 pa = ac->ac_pa; 4735 4736 /* preallocation can change ac_b_ex, thus we store actually 4737 * allocated blocks for history */ 4738 ac->ac_f_ex = ac->ac_b_ex; 4739 4740 pa->pa_pstart = ext4_grp_offs_to_block(sb, &ac->ac_b_ex); 4741 pa->pa_lstart = pa->pa_pstart; 4742 pa->pa_len = ac->ac_b_ex.fe_len; 4743 pa->pa_free = pa->pa_len; 4744 spin_lock_init(&pa->pa_lock); 4745 INIT_LIST_HEAD(&pa->pa_inode_list); 4746 INIT_LIST_HEAD(&pa->pa_group_list); 4747 pa->pa_deleted = 0; 4748 pa->pa_type = MB_GROUP_PA; 4749 4750 mb_debug(sb, "new group pa %p: %llu/%d for %u\n", pa, pa->pa_pstart, 4751 pa->pa_len, pa->pa_lstart); 4752 trace_ext4_mb_new_group_pa(ac, pa); 4753 4754 ext4_mb_use_group_pa(ac, pa); 4755 atomic_add(pa->pa_free, &EXT4_SB(sb)->s_mb_preallocated); 4756 4757 grp = ext4_get_group_info(sb, ac->ac_b_ex.fe_group); 4758 lg = ac->ac_lg; 4759 BUG_ON(lg == NULL); 4760 4761 pa->pa_obj_lock = &lg->lg_prealloc_lock; 4762 pa->pa_inode = NULL; 4763 4764 list_add(&pa->pa_group_list, &grp->bb_prealloc_list); 4765 4766 /* 4767 * We will later add the new pa to the right bucket 4768 * after updating the pa_free in ext4_mb_release_context 4769 */ 4770 } 4771 4772 static void ext4_mb_new_preallocation(struct ext4_allocation_context *ac) 4773 { 4774 if (ac->ac_flags & EXT4_MB_HINT_GROUP_ALLOC) 4775 ext4_mb_new_group_pa(ac); 4776 else 4777 ext4_mb_new_inode_pa(ac); 4778 } 4779 4780 /* 4781 * finds all unused blocks in on-disk bitmap, frees them in 4782 * in-core bitmap and buddy. 4783 * @pa must be unlinked from inode and group lists, so that 4784 * nobody else can find/use it. 4785 * the caller MUST hold group/inode locks. 4786 * TODO: optimize the case when there are no in-core structures yet 4787 */ 4788 static noinline_for_stack int 4789 ext4_mb_release_inode_pa(struct ext4_buddy *e4b, struct buffer_head *bitmap_bh, 4790 struct ext4_prealloc_space *pa) 4791 { 4792 struct super_block *sb = e4b->bd_sb; 4793 struct ext4_sb_info *sbi = EXT4_SB(sb); 4794 unsigned int end; 4795 unsigned int next; 4796 ext4_group_t group; 4797 ext4_grpblk_t bit; 4798 unsigned long long grp_blk_start; 4799 int free = 0; 4800 4801 BUG_ON(pa->pa_deleted == 0); 4802 ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, &bit); 4803 grp_blk_start = pa->pa_pstart - EXT4_C2B(sbi, bit); 4804 BUG_ON(group != e4b->bd_group && pa->pa_len != 0); 4805 end = bit + pa->pa_len; 4806 4807 while (bit < end) { 4808 bit = mb_find_next_zero_bit(bitmap_bh->b_data, end, bit); 4809 if (bit >= end) 4810 break; 4811 next = mb_find_next_bit(bitmap_bh->b_data, end, bit); 4812 mb_debug(sb, "free preallocated %u/%u in group %u\n", 4813 (unsigned) ext4_group_first_block_no(sb, group) + bit, 4814 (unsigned) next - bit, (unsigned) group); 4815 free += next - bit; 4816 4817 trace_ext4_mballoc_discard(sb, NULL, group, bit, next - bit); 4818 trace_ext4_mb_release_inode_pa(pa, (grp_blk_start + 4819 EXT4_C2B(sbi, bit)), 4820 next - bit); 4821 mb_free_blocks(pa->pa_inode, e4b, bit, next - bit); 4822 bit = next + 1; 4823 } 4824 if (free != pa->pa_free) { 4825 ext4_msg(e4b->bd_sb, KERN_CRIT, 4826 "pa %p: logic %lu, phys. %lu, len %d", 4827 pa, (unsigned long) pa->pa_lstart, 4828 (unsigned long) pa->pa_pstart, 4829 pa->pa_len); 4830 ext4_grp_locked_error(sb, group, 0, 0, "free %u, pa_free %u", 4831 free, pa->pa_free); 4832 /* 4833 * pa is already deleted so we use the value obtained 4834 * from the bitmap and continue. 4835 */ 4836 } 4837 atomic_add(free, &sbi->s_mb_discarded); 4838 4839 return 0; 4840 } 4841 4842 static noinline_for_stack int 4843 ext4_mb_release_group_pa(struct ext4_buddy *e4b, 4844 struct ext4_prealloc_space *pa) 4845 { 4846 struct super_block *sb = e4b->bd_sb; 4847 ext4_group_t group; 4848 ext4_grpblk_t bit; 4849 4850 trace_ext4_mb_release_group_pa(sb, pa); 4851 BUG_ON(pa->pa_deleted == 0); 4852 ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, &bit); 4853 BUG_ON(group != e4b->bd_group && pa->pa_len != 0); 4854 mb_free_blocks(pa->pa_inode, e4b, bit, pa->pa_len); 4855 atomic_add(pa->pa_len, &EXT4_SB(sb)->s_mb_discarded); 4856 trace_ext4_mballoc_discard(sb, NULL, group, bit, pa->pa_len); 4857 4858 return 0; 4859 } 4860 4861 /* 4862 * releases all preallocations in given group 4863 * 4864 * first, we need to decide discard policy: 4865 * - when do we discard 4866 * 1) ENOSPC 4867 * - how many do we discard 4868 * 1) how many requested 4869 */ 4870 static noinline_for_stack int 4871 ext4_mb_discard_group_preallocations(struct super_block *sb, 4872 ext4_group_t group, int *busy) 4873 { 4874 struct ext4_group_info *grp = ext4_get_group_info(sb, group); 4875 struct buffer_head *bitmap_bh = NULL; 4876 struct ext4_prealloc_space *pa, *tmp; 4877 struct list_head list; 4878 struct ext4_buddy e4b; 4879 int err; 4880 int free = 0; 4881 4882 mb_debug(sb, "discard preallocation for group %u\n", group); 4883 if (list_empty(&grp->bb_prealloc_list)) 4884 goto out_dbg; 4885 4886 bitmap_bh = ext4_read_block_bitmap(sb, group); 4887 if (IS_ERR(bitmap_bh)) { 4888 err = PTR_ERR(bitmap_bh); 4889 ext4_error_err(sb, -err, 4890 "Error %d reading block bitmap for %u", 4891 err, group); 4892 goto out_dbg; 4893 } 4894 4895 err = ext4_mb_load_buddy(sb, group, &e4b); 4896 if (err) { 4897 ext4_warning(sb, "Error %d loading buddy information for %u", 4898 err, group); 4899 put_bh(bitmap_bh); 4900 goto out_dbg; 4901 } 4902 4903 INIT_LIST_HEAD(&list); 4904 ext4_lock_group(sb, group); 4905 list_for_each_entry_safe(pa, tmp, 4906 &grp->bb_prealloc_list, pa_group_list) { 4907 spin_lock(&pa->pa_lock); 4908 if (atomic_read(&pa->pa_count)) { 4909 spin_unlock(&pa->pa_lock); 4910 *busy = 1; 4911 continue; 4912 } 4913 if (pa->pa_deleted) { 4914 spin_unlock(&pa->pa_lock); 4915 continue; 4916 } 4917 4918 /* seems this one can be freed ... */ 4919 ext4_mb_mark_pa_deleted(sb, pa); 4920 4921 if (!free) 4922 this_cpu_inc(discard_pa_seq); 4923 4924 /* we can trust pa_free ... */ 4925 free += pa->pa_free; 4926 4927 spin_unlock(&pa->pa_lock); 4928 4929 list_del(&pa->pa_group_list); 4930 list_add(&pa->u.pa_tmp_list, &list); 4931 } 4932 4933 /* now free all selected PAs */ 4934 list_for_each_entry_safe(pa, tmp, &list, u.pa_tmp_list) { 4935 4936 /* remove from object (inode or locality group) */ 4937 spin_lock(pa->pa_obj_lock); 4938 list_del_rcu(&pa->pa_inode_list); 4939 spin_unlock(pa->pa_obj_lock); 4940 4941 if (pa->pa_type == MB_GROUP_PA) 4942 ext4_mb_release_group_pa(&e4b, pa); 4943 else 4944 ext4_mb_release_inode_pa(&e4b, bitmap_bh, pa); 4945 4946 list_del(&pa->u.pa_tmp_list); 4947 call_rcu(&(pa)->u.pa_rcu, ext4_mb_pa_callback); 4948 } 4949 4950 ext4_unlock_group(sb, group); 4951 ext4_mb_unload_buddy(&e4b); 4952 put_bh(bitmap_bh); 4953 out_dbg: 4954 mb_debug(sb, "discarded (%d) blocks preallocated for group %u bb_free (%d)\n", 4955 free, group, grp->bb_free); 4956 return free; 4957 } 4958 4959 /* 4960 * releases all non-used preallocated blocks for given inode 4961 * 4962 * It's important to discard preallocations under i_data_sem 4963 * We don't want another block to be served from the prealloc 4964 * space when we are discarding the inode prealloc space. 4965 * 4966 * FIXME!! Make sure it is valid at all the call sites 4967 */ 4968 void ext4_discard_preallocations(struct inode *inode, unsigned int needed) 4969 { 4970 struct ext4_inode_info *ei = EXT4_I(inode); 4971 struct super_block *sb = inode->i_sb; 4972 struct buffer_head *bitmap_bh = NULL; 4973 struct ext4_prealloc_space *pa, *tmp; 4974 ext4_group_t group = 0; 4975 struct list_head list; 4976 struct ext4_buddy e4b; 4977 int err; 4978 4979 if (!S_ISREG(inode->i_mode)) { 4980 /*BUG_ON(!list_empty(&ei->i_prealloc_list));*/ 4981 return; 4982 } 4983 4984 if (EXT4_SB(sb)->s_mount_state & EXT4_FC_REPLAY) 4985 return; 4986 4987 mb_debug(sb, "discard preallocation for inode %lu\n", 4988 inode->i_ino); 4989 trace_ext4_discard_preallocations(inode, 4990 atomic_read(&ei->i_prealloc_active), needed); 4991 4992 INIT_LIST_HEAD(&list); 4993 4994 if (needed == 0) 4995 needed = UINT_MAX; 4996 4997 repeat: 4998 /* first, collect all pa's in the inode */ 4999 spin_lock(&ei->i_prealloc_lock); 5000 while (!list_empty(&ei->i_prealloc_list) && needed) { 5001 pa = list_entry(ei->i_prealloc_list.prev, 5002 struct ext4_prealloc_space, pa_inode_list); 5003 BUG_ON(pa->pa_obj_lock != &ei->i_prealloc_lock); 5004 spin_lock(&pa->pa_lock); 5005 if (atomic_read(&pa->pa_count)) { 5006 /* this shouldn't happen often - nobody should 5007 * use preallocation while we're discarding it */ 5008 spin_unlock(&pa->pa_lock); 5009 spin_unlock(&ei->i_prealloc_lock); 5010 ext4_msg(sb, KERN_ERR, 5011 "uh-oh! used pa while discarding"); 5012 WARN_ON(1); 5013 schedule_timeout_uninterruptible(HZ); 5014 goto repeat; 5015 5016 } 5017 if (pa->pa_deleted == 0) { 5018 ext4_mb_mark_pa_deleted(sb, pa); 5019 spin_unlock(&pa->pa_lock); 5020 list_del_rcu(&pa->pa_inode_list); 5021 list_add(&pa->u.pa_tmp_list, &list); 5022 needed--; 5023 continue; 5024 } 5025 5026 /* someone is deleting pa right now */ 5027 spin_unlock(&pa->pa_lock); 5028 spin_unlock(&ei->i_prealloc_lock); 5029 5030 /* we have to wait here because pa_deleted 5031 * doesn't mean pa is already unlinked from 5032 * the list. as we might be called from 5033 * ->clear_inode() the inode will get freed 5034 * and concurrent thread which is unlinking 5035 * pa from inode's list may access already 5036 * freed memory, bad-bad-bad */ 5037 5038 /* XXX: if this happens too often, we can 5039 * add a flag to force wait only in case 5040 * of ->clear_inode(), but not in case of 5041 * regular truncate */ 5042 schedule_timeout_uninterruptible(HZ); 5043 goto repeat; 5044 } 5045 spin_unlock(&ei->i_prealloc_lock); 5046 5047 list_for_each_entry_safe(pa, tmp, &list, u.pa_tmp_list) { 5048 BUG_ON(pa->pa_type != MB_INODE_PA); 5049 group = ext4_get_group_number(sb, pa->pa_pstart); 5050 5051 err = ext4_mb_load_buddy_gfp(sb, group, &e4b, 5052 GFP_NOFS|__GFP_NOFAIL); 5053 if (err) { 5054 ext4_error_err(sb, -err, "Error %d loading buddy information for %u", 5055 err, group); 5056 continue; 5057 } 5058 5059 bitmap_bh = ext4_read_block_bitmap(sb, group); 5060 if (IS_ERR(bitmap_bh)) { 5061 err = PTR_ERR(bitmap_bh); 5062 ext4_error_err(sb, -err, "Error %d reading block bitmap for %u", 5063 err, group); 5064 ext4_mb_unload_buddy(&e4b); 5065 continue; 5066 } 5067 5068 ext4_lock_group(sb, group); 5069 list_del(&pa->pa_group_list); 5070 ext4_mb_release_inode_pa(&e4b, bitmap_bh, pa); 5071 ext4_unlock_group(sb, group); 5072 5073 ext4_mb_unload_buddy(&e4b); 5074 put_bh(bitmap_bh); 5075 5076 list_del(&pa->u.pa_tmp_list); 5077 call_rcu(&(pa)->u.pa_rcu, ext4_mb_pa_callback); 5078 } 5079 } 5080 5081 static int ext4_mb_pa_alloc(struct ext4_allocation_context *ac) 5082 { 5083 struct ext4_prealloc_space *pa; 5084 5085 BUG_ON(ext4_pspace_cachep == NULL); 5086 pa = kmem_cache_zalloc(ext4_pspace_cachep, GFP_NOFS); 5087 if (!pa) 5088 return -ENOMEM; 5089 atomic_set(&pa->pa_count, 1); 5090 ac->ac_pa = pa; 5091 return 0; 5092 } 5093 5094 static void ext4_mb_pa_free(struct ext4_allocation_context *ac) 5095 { 5096 struct ext4_prealloc_space *pa = ac->ac_pa; 5097 5098 BUG_ON(!pa); 5099 ac->ac_pa = NULL; 5100 WARN_ON(!atomic_dec_and_test(&pa->pa_count)); 5101 kmem_cache_free(ext4_pspace_cachep, pa); 5102 } 5103 5104 #ifdef CONFIG_EXT4_DEBUG 5105 static inline void ext4_mb_show_pa(struct super_block *sb) 5106 { 5107 ext4_group_t i, ngroups; 5108 5109 if (ext4_test_mount_flag(sb, EXT4_MF_FS_ABORTED)) 5110 return; 5111 5112 ngroups = ext4_get_groups_count(sb); 5113 mb_debug(sb, "groups: "); 5114 for (i = 0; i < ngroups; i++) { 5115 struct ext4_group_info *grp = ext4_get_group_info(sb, i); 5116 struct ext4_prealloc_space *pa; 5117 ext4_grpblk_t start; 5118 struct list_head *cur; 5119 ext4_lock_group(sb, i); 5120 list_for_each(cur, &grp->bb_prealloc_list) { 5121 pa = list_entry(cur, struct ext4_prealloc_space, 5122 pa_group_list); 5123 spin_lock(&pa->pa_lock); 5124 ext4_get_group_no_and_offset(sb, pa->pa_pstart, 5125 NULL, &start); 5126 spin_unlock(&pa->pa_lock); 5127 mb_debug(sb, "PA:%u:%d:%d\n", i, start, 5128 pa->pa_len); 5129 } 5130 ext4_unlock_group(sb, i); 5131 mb_debug(sb, "%u: %d/%d\n", i, grp->bb_free, 5132 grp->bb_fragments); 5133 } 5134 } 5135 5136 static void ext4_mb_show_ac(struct ext4_allocation_context *ac) 5137 { 5138 struct super_block *sb = ac->ac_sb; 5139 5140 if (ext4_test_mount_flag(sb, EXT4_MF_FS_ABORTED)) 5141 return; 5142 5143 mb_debug(sb, "Can't allocate:" 5144 " Allocation context details:"); 5145 mb_debug(sb, "status %u flags 0x%x", 5146 ac->ac_status, ac->ac_flags); 5147 mb_debug(sb, "orig %lu/%lu/%lu@%lu, " 5148 "goal %lu/%lu/%lu@%lu, " 5149 "best %lu/%lu/%lu@%lu cr %d", 5150 (unsigned long)ac->ac_o_ex.fe_group, 5151 (unsigned long)ac->ac_o_ex.fe_start, 5152 (unsigned long)ac->ac_o_ex.fe_len, 5153 (unsigned long)ac->ac_o_ex.fe_logical, 5154 (unsigned long)ac->ac_g_ex.fe_group, 5155 (unsigned long)ac->ac_g_ex.fe_start, 5156 (unsigned long)ac->ac_g_ex.fe_len, 5157 (unsigned long)ac->ac_g_ex.fe_logical, 5158 (unsigned long)ac->ac_b_ex.fe_group, 5159 (unsigned long)ac->ac_b_ex.fe_start, 5160 (unsigned long)ac->ac_b_ex.fe_len, 5161 (unsigned long)ac->ac_b_ex.fe_logical, 5162 (int)ac->ac_criteria); 5163 mb_debug(sb, "%u found", ac->ac_found); 5164 ext4_mb_show_pa(sb); 5165 } 5166 #else 5167 static inline void ext4_mb_show_pa(struct super_block *sb) 5168 { 5169 return; 5170 } 5171 static inline void ext4_mb_show_ac(struct ext4_allocation_context *ac) 5172 { 5173 ext4_mb_show_pa(ac->ac_sb); 5174 return; 5175 } 5176 #endif 5177 5178 /* 5179 * We use locality group preallocation for small size file. The size of the 5180 * file is determined by the current size or the resulting size after 5181 * allocation which ever is larger 5182 * 5183 * One can tune this size via /sys/fs/ext4/<partition>/mb_stream_req 5184 */ 5185 static void ext4_mb_group_or_file(struct ext4_allocation_context *ac) 5186 { 5187 struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); 5188 int bsbits = ac->ac_sb->s_blocksize_bits; 5189 loff_t size, isize; 5190 5191 if (!(ac->ac_flags & EXT4_MB_HINT_DATA)) 5192 return; 5193 5194 if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY)) 5195 return; 5196 5197 size = ac->ac_o_ex.fe_logical + EXT4_C2B(sbi, ac->ac_o_ex.fe_len); 5198 isize = (i_size_read(ac->ac_inode) + ac->ac_sb->s_blocksize - 1) 5199 >> bsbits; 5200 5201 if ((size == isize) && !ext4_fs_is_busy(sbi) && 5202 !inode_is_open_for_write(ac->ac_inode)) { 5203 ac->ac_flags |= EXT4_MB_HINT_NOPREALLOC; 5204 return; 5205 } 5206 5207 if (sbi->s_mb_group_prealloc <= 0) { 5208 ac->ac_flags |= EXT4_MB_STREAM_ALLOC; 5209 return; 5210 } 5211 5212 /* don't use group allocation for large files */ 5213 size = max(size, isize); 5214 if (size > sbi->s_mb_stream_request) { 5215 ac->ac_flags |= EXT4_MB_STREAM_ALLOC; 5216 return; 5217 } 5218 5219 BUG_ON(ac->ac_lg != NULL); 5220 /* 5221 * locality group prealloc space are per cpu. The reason for having 5222 * per cpu locality group is to reduce the contention between block 5223 * request from multiple CPUs. 5224 */ 5225 ac->ac_lg = raw_cpu_ptr(sbi->s_locality_groups); 5226 5227 /* we're going to use group allocation */ 5228 ac->ac_flags |= EXT4_MB_HINT_GROUP_ALLOC; 5229 5230 /* serialize all allocations in the group */ 5231 mutex_lock(&ac->ac_lg->lg_mutex); 5232 } 5233 5234 static noinline_for_stack int 5235 ext4_mb_initialize_context(struct ext4_allocation_context *ac, 5236 struct ext4_allocation_request *ar) 5237 { 5238 struct super_block *sb = ar->inode->i_sb; 5239 struct ext4_sb_info *sbi = EXT4_SB(sb); 5240 struct ext4_super_block *es = sbi->s_es; 5241 ext4_group_t group; 5242 unsigned int len; 5243 ext4_fsblk_t goal; 5244 ext4_grpblk_t block; 5245 5246 /* we can't allocate > group size */ 5247 len = ar->len; 5248 5249 /* just a dirty hack to filter too big requests */ 5250 if (len >= EXT4_CLUSTERS_PER_GROUP(sb)) 5251 len = EXT4_CLUSTERS_PER_GROUP(sb); 5252 5253 /* start searching from the goal */ 5254 goal = ar->goal; 5255 if (goal < le32_to_cpu(es->s_first_data_block) || 5256 goal >= ext4_blocks_count(es)) 5257 goal = le32_to_cpu(es->s_first_data_block); 5258 ext4_get_group_no_and_offset(sb, goal, &group, &block); 5259 5260 /* set up allocation goals */ 5261 ac->ac_b_ex.fe_logical = EXT4_LBLK_CMASK(sbi, ar->logical); 5262 ac->ac_status = AC_STATUS_CONTINUE; 5263 ac->ac_sb = sb; 5264 ac->ac_inode = ar->inode; 5265 ac->ac_o_ex.fe_logical = ac->ac_b_ex.fe_logical; 5266 ac->ac_o_ex.fe_group = group; 5267 ac->ac_o_ex.fe_start = block; 5268 ac->ac_o_ex.fe_len = len; 5269 ac->ac_g_ex = ac->ac_o_ex; 5270 ac->ac_flags = ar->flags; 5271 5272 /* we have to define context: we'll work with a file or 5273 * locality group. this is a policy, actually */ 5274 ext4_mb_group_or_file(ac); 5275 5276 mb_debug(sb, "init ac: %u blocks @ %u, goal %u, flags 0x%x, 2^%d, " 5277 "left: %u/%u, right %u/%u to %swritable\n", 5278 (unsigned) ar->len, (unsigned) ar->logical, 5279 (unsigned) ar->goal, ac->ac_flags, ac->ac_2order, 5280 (unsigned) ar->lleft, (unsigned) ar->pleft, 5281 (unsigned) ar->lright, (unsigned) ar->pright, 5282 inode_is_open_for_write(ar->inode) ? "" : "non-"); 5283 return 0; 5284 5285 } 5286 5287 static noinline_for_stack void 5288 ext4_mb_discard_lg_preallocations(struct super_block *sb, 5289 struct ext4_locality_group *lg, 5290 int order, int total_entries) 5291 { 5292 ext4_group_t group = 0; 5293 struct ext4_buddy e4b; 5294 struct list_head discard_list; 5295 struct ext4_prealloc_space *pa, *tmp; 5296 5297 mb_debug(sb, "discard locality group preallocation\n"); 5298 5299 INIT_LIST_HEAD(&discard_list); 5300 5301 spin_lock(&lg->lg_prealloc_lock); 5302 list_for_each_entry_rcu(pa, &lg->lg_prealloc_list[order], 5303 pa_inode_list, 5304 lockdep_is_held(&lg->lg_prealloc_lock)) { 5305 spin_lock(&pa->pa_lock); 5306 if (atomic_read(&pa->pa_count)) { 5307 /* 5308 * This is the pa that we just used 5309 * for block allocation. So don't 5310 * free that 5311 */ 5312 spin_unlock(&pa->pa_lock); 5313 continue; 5314 } 5315 if (pa->pa_deleted) { 5316 spin_unlock(&pa->pa_lock); 5317 continue; 5318 } 5319 /* only lg prealloc space */ 5320 BUG_ON(pa->pa_type != MB_GROUP_PA); 5321 5322 /* seems this one can be freed ... */ 5323 ext4_mb_mark_pa_deleted(sb, pa); 5324 spin_unlock(&pa->pa_lock); 5325 5326 list_del_rcu(&pa->pa_inode_list); 5327 list_add(&pa->u.pa_tmp_list, &discard_list); 5328 5329 total_entries--; 5330 if (total_entries <= 5) { 5331 /* 5332 * we want to keep only 5 entries 5333 * allowing it to grow to 8. This 5334 * mak sure we don't call discard 5335 * soon for this list. 5336 */ 5337 break; 5338 } 5339 } 5340 spin_unlock(&lg->lg_prealloc_lock); 5341 5342 list_for_each_entry_safe(pa, tmp, &discard_list, u.pa_tmp_list) { 5343 int err; 5344 5345 group = ext4_get_group_number(sb, pa->pa_pstart); 5346 err = ext4_mb_load_buddy_gfp(sb, group, &e4b, 5347 GFP_NOFS|__GFP_NOFAIL); 5348 if (err) { 5349 ext4_error_err(sb, -err, "Error %d loading buddy information for %u", 5350 err, group); 5351 continue; 5352 } 5353 ext4_lock_group(sb, group); 5354 list_del(&pa->pa_group_list); 5355 ext4_mb_release_group_pa(&e4b, pa); 5356 ext4_unlock_group(sb, group); 5357 5358 ext4_mb_unload_buddy(&e4b); 5359 list_del(&pa->u.pa_tmp_list); 5360 call_rcu(&(pa)->u.pa_rcu, ext4_mb_pa_callback); 5361 } 5362 } 5363 5364 /* 5365 * We have incremented pa_count. So it cannot be freed at this 5366 * point. Also we hold lg_mutex. So no parallel allocation is 5367 * possible from this lg. That means pa_free cannot be updated. 5368 * 5369 * A parallel ext4_mb_discard_group_preallocations is possible. 5370 * which can cause the lg_prealloc_list to be updated. 5371 */ 5372 5373 static void ext4_mb_add_n_trim(struct ext4_allocation_context *ac) 5374 { 5375 int order, added = 0, lg_prealloc_count = 1; 5376 struct super_block *sb = ac->ac_sb; 5377 struct ext4_locality_group *lg = ac->ac_lg; 5378 struct ext4_prealloc_space *tmp_pa, *pa = ac->ac_pa; 5379 5380 order = fls(pa->pa_free) - 1; 5381 if (order > PREALLOC_TB_SIZE - 1) 5382 /* The max size of hash table is PREALLOC_TB_SIZE */ 5383 order = PREALLOC_TB_SIZE - 1; 5384 /* Add the prealloc space to lg */ 5385 spin_lock(&lg->lg_prealloc_lock); 5386 list_for_each_entry_rcu(tmp_pa, &lg->lg_prealloc_list[order], 5387 pa_inode_list, 5388 lockdep_is_held(&lg->lg_prealloc_lock)) { 5389 spin_lock(&tmp_pa->pa_lock); 5390 if (tmp_pa->pa_deleted) { 5391 spin_unlock(&tmp_pa->pa_lock); 5392 continue; 5393 } 5394 if (!added && pa->pa_free < tmp_pa->pa_free) { 5395 /* Add to the tail of the previous entry */ 5396 list_add_tail_rcu(&pa->pa_inode_list, 5397 &tmp_pa->pa_inode_list); 5398 added = 1; 5399 /* 5400 * we want to count the total 5401 * number of entries in the list 5402 */ 5403 } 5404 spin_unlock(&tmp_pa->pa_lock); 5405 lg_prealloc_count++; 5406 } 5407 if (!added) 5408 list_add_tail_rcu(&pa->pa_inode_list, 5409 &lg->lg_prealloc_list[order]); 5410 spin_unlock(&lg->lg_prealloc_lock); 5411 5412 /* Now trim the list to be not more than 8 elements */ 5413 if (lg_prealloc_count > 8) { 5414 ext4_mb_discard_lg_preallocations(sb, lg, 5415 order, lg_prealloc_count); 5416 return; 5417 } 5418 return ; 5419 } 5420 5421 /* 5422 * if per-inode prealloc list is too long, trim some PA 5423 */ 5424 static void ext4_mb_trim_inode_pa(struct inode *inode) 5425 { 5426 struct ext4_inode_info *ei = EXT4_I(inode); 5427 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); 5428 int count, delta; 5429 5430 count = atomic_read(&ei->i_prealloc_active); 5431 delta = (sbi->s_mb_max_inode_prealloc >> 2) + 1; 5432 if (count > sbi->s_mb_max_inode_prealloc + delta) { 5433 count -= sbi->s_mb_max_inode_prealloc; 5434 ext4_discard_preallocations(inode, count); 5435 } 5436 } 5437 5438 /* 5439 * release all resource we used in allocation 5440 */ 5441 static int ext4_mb_release_context(struct ext4_allocation_context *ac) 5442 { 5443 struct inode *inode = ac->ac_inode; 5444 struct ext4_inode_info *ei = EXT4_I(inode); 5445 struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb); 5446 struct ext4_prealloc_space *pa = ac->ac_pa; 5447 if (pa) { 5448 if (pa->pa_type == MB_GROUP_PA) { 5449 /* see comment in ext4_mb_use_group_pa() */ 5450 spin_lock(&pa->pa_lock); 5451 pa->pa_pstart += EXT4_C2B(sbi, ac->ac_b_ex.fe_len); 5452 pa->pa_lstart += EXT4_C2B(sbi, ac->ac_b_ex.fe_len); 5453 pa->pa_free -= ac->ac_b_ex.fe_len; 5454 pa->pa_len -= ac->ac_b_ex.fe_len; 5455 spin_unlock(&pa->pa_lock); 5456 5457 /* 5458 * We want to add the pa to the right bucket. 5459 * Remove it from the list and while adding 5460 * make sure the list to which we are adding 5461 * doesn't grow big. 5462 */ 5463 if (likely(pa->pa_free)) { 5464 spin_lock(pa->pa_obj_lock); 5465 list_del_rcu(&pa->pa_inode_list); 5466 spin_unlock(pa->pa_obj_lock); 5467 ext4_mb_add_n_trim(ac); 5468 } 5469 } 5470 5471 if (pa->pa_type == MB_INODE_PA) { 5472 /* 5473 * treat per-inode prealloc list as a lru list, then try 5474 * to trim the least recently used PA. 5475 */ 5476 spin_lock(pa->pa_obj_lock); 5477 list_move(&pa->pa_inode_list, &ei->i_prealloc_list); 5478 spin_unlock(pa->pa_obj_lock); 5479 } 5480 5481 ext4_mb_put_pa(ac, ac->ac_sb, pa); 5482 } 5483 if (ac->ac_bitmap_page) 5484 put_page(ac->ac_bitmap_page); 5485 if (ac->ac_buddy_page) 5486 put_page(ac->ac_buddy_page); 5487 if (ac->ac_flags & EXT4_MB_HINT_GROUP_ALLOC) 5488 mutex_unlock(&ac->ac_lg->lg_mutex); 5489 ext4_mb_collect_stats(ac); 5490 ext4_mb_trim_inode_pa(inode); 5491 return 0; 5492 } 5493 5494 static int ext4_mb_discard_preallocations(struct super_block *sb, int needed) 5495 { 5496 ext4_group_t i, ngroups = ext4_get_groups_count(sb); 5497 int ret; 5498 int freed = 0, busy = 0; 5499 int retry = 0; 5500 5501 trace_ext4_mb_discard_preallocations(sb, needed); 5502 5503 if (needed == 0) 5504 needed = EXT4_CLUSTERS_PER_GROUP(sb) + 1; 5505 repeat: 5506 for (i = 0; i < ngroups && needed > 0; i++) { 5507 ret = ext4_mb_discard_group_preallocations(sb, i, &busy); 5508 freed += ret; 5509 needed -= ret; 5510 cond_resched(); 5511 } 5512 5513 if (needed > 0 && busy && ++retry < 3) { 5514 busy = 0; 5515 goto repeat; 5516 } 5517 5518 return freed; 5519 } 5520 5521 static bool ext4_mb_discard_preallocations_should_retry(struct super_block *sb, 5522 struct ext4_allocation_context *ac, u64 *seq) 5523 { 5524 int freed; 5525 u64 seq_retry = 0; 5526 bool ret = false; 5527 5528 freed = ext4_mb_discard_preallocations(sb, ac->ac_o_ex.fe_len); 5529 if (freed) { 5530 ret = true; 5531 goto out_dbg; 5532 } 5533 seq_retry = ext4_get_discard_pa_seq_sum(); 5534 if (!(ac->ac_flags & EXT4_MB_STRICT_CHECK) || seq_retry != *seq) { 5535 ac->ac_flags |= EXT4_MB_STRICT_CHECK; 5536 *seq = seq_retry; 5537 ret = true; 5538 } 5539 5540 out_dbg: 5541 mb_debug(sb, "freed %d, retry ? %s\n", freed, ret ? "yes" : "no"); 5542 return ret; 5543 } 5544 5545 static ext4_fsblk_t ext4_mb_new_blocks_simple(handle_t *handle, 5546 struct ext4_allocation_request *ar, int *errp); 5547 5548 /* 5549 * Main entry point into mballoc to allocate blocks 5550 * it tries to use preallocation first, then falls back 5551 * to usual allocation 5552 */ 5553 ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle, 5554 struct ext4_allocation_request *ar, int *errp) 5555 { 5556 struct ext4_allocation_context *ac = NULL; 5557 struct ext4_sb_info *sbi; 5558 struct super_block *sb; 5559 ext4_fsblk_t block = 0; 5560 unsigned int inquota = 0; 5561 unsigned int reserv_clstrs = 0; 5562 u64 seq; 5563 5564 might_sleep(); 5565 sb = ar->inode->i_sb; 5566 sbi = EXT4_SB(sb); 5567 5568 trace_ext4_request_blocks(ar); 5569 if (sbi->s_mount_state & EXT4_FC_REPLAY) 5570 return ext4_mb_new_blocks_simple(handle, ar, errp); 5571 5572 /* Allow to use superuser reservation for quota file */ 5573 if (ext4_is_quota_file(ar->inode)) 5574 ar->flags |= EXT4_MB_USE_ROOT_BLOCKS; 5575 5576 if ((ar->flags & EXT4_MB_DELALLOC_RESERVED) == 0) { 5577 /* Without delayed allocation we need to verify 5578 * there is enough free blocks to do block allocation 5579 * and verify allocation doesn't exceed the quota limits. 5580 */ 5581 while (ar->len && 5582 ext4_claim_free_clusters(sbi, ar->len, ar->flags)) { 5583 5584 /* let others to free the space */ 5585 cond_resched(); 5586 ar->len = ar->len >> 1; 5587 } 5588 if (!ar->len) { 5589 ext4_mb_show_pa(sb); 5590 *errp = -ENOSPC; 5591 return 0; 5592 } 5593 reserv_clstrs = ar->len; 5594 if (ar->flags & EXT4_MB_USE_ROOT_BLOCKS) { 5595 dquot_alloc_block_nofail(ar->inode, 5596 EXT4_C2B(sbi, ar->len)); 5597 } else { 5598 while (ar->len && 5599 dquot_alloc_block(ar->inode, 5600 EXT4_C2B(sbi, ar->len))) { 5601 5602 ar->flags |= EXT4_MB_HINT_NOPREALLOC; 5603 ar->len--; 5604 } 5605 } 5606 inquota = ar->len; 5607 if (ar->len == 0) { 5608 *errp = -EDQUOT; 5609 goto out; 5610 } 5611 } 5612 5613 ac = kmem_cache_zalloc(ext4_ac_cachep, GFP_NOFS); 5614 if (!ac) { 5615 ar->len = 0; 5616 *errp = -ENOMEM; 5617 goto out; 5618 } 5619 5620 *errp = ext4_mb_initialize_context(ac, ar); 5621 if (*errp) { 5622 ar->len = 0; 5623 goto out; 5624 } 5625 5626 ac->ac_op = EXT4_MB_HISTORY_PREALLOC; 5627 seq = this_cpu_read(discard_pa_seq); 5628 if (!ext4_mb_use_preallocated(ac)) { 5629 ac->ac_op = EXT4_MB_HISTORY_ALLOC; 5630 ext4_mb_normalize_request(ac, ar); 5631 5632 *errp = ext4_mb_pa_alloc(ac); 5633 if (*errp) 5634 goto errout; 5635 repeat: 5636 /* allocate space in core */ 5637 *errp = ext4_mb_regular_allocator(ac); 5638 /* 5639 * pa allocated above is added to grp->bb_prealloc_list only 5640 * when we were able to allocate some block i.e. when 5641 * ac->ac_status == AC_STATUS_FOUND. 5642 * And error from above mean ac->ac_status != AC_STATUS_FOUND 5643 * So we have to free this pa here itself. 5644 */ 5645 if (*errp) { 5646 ext4_mb_pa_free(ac); 5647 ext4_discard_allocated_blocks(ac); 5648 goto errout; 5649 } 5650 if (ac->ac_status == AC_STATUS_FOUND && 5651 ac->ac_o_ex.fe_len >= ac->ac_f_ex.fe_len) 5652 ext4_mb_pa_free(ac); 5653 } 5654 if (likely(ac->ac_status == AC_STATUS_FOUND)) { 5655 *errp = ext4_mb_mark_diskspace_used(ac, handle, reserv_clstrs); 5656 if (*errp) { 5657 ext4_discard_allocated_blocks(ac); 5658 goto errout; 5659 } else { 5660 block = ext4_grp_offs_to_block(sb, &ac->ac_b_ex); 5661 ar->len = ac->ac_b_ex.fe_len; 5662 } 5663 } else { 5664 if (ext4_mb_discard_preallocations_should_retry(sb, ac, &seq)) 5665 goto repeat; 5666 /* 5667 * If block allocation fails then the pa allocated above 5668 * needs to be freed here itself. 5669 */ 5670 ext4_mb_pa_free(ac); 5671 *errp = -ENOSPC; 5672 } 5673 5674 errout: 5675 if (*errp) { 5676 ac->ac_b_ex.fe_len = 0; 5677 ar->len = 0; 5678 ext4_mb_show_ac(ac); 5679 } 5680 ext4_mb_release_context(ac); 5681 out: 5682 if (ac) 5683 kmem_cache_free(ext4_ac_cachep, ac); 5684 if (inquota && ar->len < inquota) 5685 dquot_free_block(ar->inode, EXT4_C2B(sbi, inquota - ar->len)); 5686 if (!ar->len) { 5687 if ((ar->flags & EXT4_MB_DELALLOC_RESERVED) == 0) 5688 /* release all the reserved blocks if non delalloc */ 5689 percpu_counter_sub(&sbi->s_dirtyclusters_counter, 5690 reserv_clstrs); 5691 } 5692 5693 trace_ext4_allocate_blocks(ar, (unsigned long long)block); 5694 5695 return block; 5696 } 5697 5698 /* 5699 * We can merge two free data extents only if the physical blocks 5700 * are contiguous, AND the extents were freed by the same transaction, 5701 * AND the blocks are associated with the same group. 5702 */ 5703 static void ext4_try_merge_freed_extent(struct ext4_sb_info *sbi, 5704 struct ext4_free_data *entry, 5705 struct ext4_free_data *new_entry, 5706 struct rb_root *entry_rb_root) 5707 { 5708 if ((entry->efd_tid != new_entry->efd_tid) || 5709 (entry->efd_group != new_entry->efd_group)) 5710 return; 5711 if (entry->efd_start_cluster + entry->efd_count == 5712 new_entry->efd_start_cluster) { 5713 new_entry->efd_start_cluster = entry->efd_start_cluster; 5714 new_entry->efd_count += entry->efd_count; 5715 } else if (new_entry->efd_start_cluster + new_entry->efd_count == 5716 entry->efd_start_cluster) { 5717 new_entry->efd_count += entry->efd_count; 5718 } else 5719 return; 5720 spin_lock(&sbi->s_md_lock); 5721 list_del(&entry->efd_list); 5722 spin_unlock(&sbi->s_md_lock); 5723 rb_erase(&entry->efd_node, entry_rb_root); 5724 kmem_cache_free(ext4_free_data_cachep, entry); 5725 } 5726 5727 static noinline_for_stack int 5728 ext4_mb_free_metadata(handle_t *handle, struct ext4_buddy *e4b, 5729 struct ext4_free_data *new_entry) 5730 { 5731 ext4_group_t group = e4b->bd_group; 5732 ext4_grpblk_t cluster; 5733 ext4_grpblk_t clusters = new_entry->efd_count; 5734 struct ext4_free_data *entry; 5735 struct ext4_group_info *db = e4b->bd_info; 5736 struct super_block *sb = e4b->bd_sb; 5737 struct ext4_sb_info *sbi = EXT4_SB(sb); 5738 struct rb_node **n = &db->bb_free_root.rb_node, *node; 5739 struct rb_node *parent = NULL, *new_node; 5740 5741 BUG_ON(!ext4_handle_valid(handle)); 5742 BUG_ON(e4b->bd_bitmap_page == NULL); 5743 BUG_ON(e4b->bd_buddy_page == NULL); 5744 5745 new_node = &new_entry->efd_node; 5746 cluster = new_entry->efd_start_cluster; 5747 5748 if (!*n) { 5749 /* first free block exent. We need to 5750 protect buddy cache from being freed, 5751 * otherwise we'll refresh it from 5752 * on-disk bitmap and lose not-yet-available 5753 * blocks */ 5754 get_page(e4b->bd_buddy_page); 5755 get_page(e4b->bd_bitmap_page); 5756 } 5757 while (*n) { 5758 parent = *n; 5759 entry = rb_entry(parent, struct ext4_free_data, efd_node); 5760 if (cluster < entry->efd_start_cluster) 5761 n = &(*n)->rb_left; 5762 else if (cluster >= (entry->efd_start_cluster + entry->efd_count)) 5763 n = &(*n)->rb_right; 5764 else { 5765 ext4_grp_locked_error(sb, group, 0, 5766 ext4_group_first_block_no(sb, group) + 5767 EXT4_C2B(sbi, cluster), 5768 "Block already on to-be-freed list"); 5769 kmem_cache_free(ext4_free_data_cachep, new_entry); 5770 return 0; 5771 } 5772 } 5773 5774 rb_link_node(new_node, parent, n); 5775 rb_insert_color(new_node, &db->bb_free_root); 5776 5777 /* Now try to see the extent can be merged to left and right */ 5778 node = rb_prev(new_node); 5779 if (node) { 5780 entry = rb_entry(node, struct ext4_free_data, efd_node); 5781 ext4_try_merge_freed_extent(sbi, entry, new_entry, 5782 &(db->bb_free_root)); 5783 } 5784 5785 node = rb_next(new_node); 5786 if (node) { 5787 entry = rb_entry(node, struct ext4_free_data, efd_node); 5788 ext4_try_merge_freed_extent(sbi, entry, new_entry, 5789 &(db->bb_free_root)); 5790 } 5791 5792 spin_lock(&sbi->s_md_lock); 5793 list_add_tail(&new_entry->efd_list, &sbi->s_freed_data_list); 5794 sbi->s_mb_free_pending += clusters; 5795 spin_unlock(&sbi->s_md_lock); 5796 return 0; 5797 } 5798 5799 /* 5800 * Simple allocator for Ext4 fast commit replay path. It searches for blocks 5801 * linearly starting at the goal block and also excludes the blocks which 5802 * are going to be in use after fast commit replay. 5803 */ 5804 static ext4_fsblk_t ext4_mb_new_blocks_simple(handle_t *handle, 5805 struct ext4_allocation_request *ar, int *errp) 5806 { 5807 struct buffer_head *bitmap_bh; 5808 struct super_block *sb = ar->inode->i_sb; 5809 ext4_group_t group; 5810 ext4_grpblk_t blkoff; 5811 ext4_grpblk_t max = EXT4_CLUSTERS_PER_GROUP(sb); 5812 ext4_grpblk_t i = 0; 5813 ext4_fsblk_t goal, block; 5814 struct ext4_super_block *es = EXT4_SB(sb)->s_es; 5815 5816 goal = ar->goal; 5817 if (goal < le32_to_cpu(es->s_first_data_block) || 5818 goal >= ext4_blocks_count(es)) 5819 goal = le32_to_cpu(es->s_first_data_block); 5820 5821 ar->len = 0; 5822 ext4_get_group_no_and_offset(sb, goal, &group, &blkoff); 5823 for (; group < ext4_get_groups_count(sb); group++) { 5824 bitmap_bh = ext4_read_block_bitmap(sb, group); 5825 if (IS_ERR(bitmap_bh)) { 5826 *errp = PTR_ERR(bitmap_bh); 5827 pr_warn("Failed to read block bitmap\n"); 5828 return 0; 5829 } 5830 5831 ext4_get_group_no_and_offset(sb, 5832 max(ext4_group_first_block_no(sb, group), goal), 5833 NULL, &blkoff); 5834 while (1) { 5835 i = mb_find_next_zero_bit(bitmap_bh->b_data, max, 5836 blkoff); 5837 if (i >= max) 5838 break; 5839 if (ext4_fc_replay_check_excluded(sb, 5840 ext4_group_first_block_no(sb, group) + i)) { 5841 blkoff = i + 1; 5842 } else 5843 break; 5844 } 5845 brelse(bitmap_bh); 5846 if (i < max) 5847 break; 5848 } 5849 5850 if (group >= ext4_get_groups_count(sb) || i >= max) { 5851 *errp = -ENOSPC; 5852 return 0; 5853 } 5854 5855 block = ext4_group_first_block_no(sb, group) + i; 5856 ext4_mb_mark_bb(sb, block, 1, 1); 5857 ar->len = 1; 5858 5859 return block; 5860 } 5861 5862 static void ext4_free_blocks_simple(struct inode *inode, ext4_fsblk_t block, 5863 unsigned long count) 5864 { 5865 struct buffer_head *bitmap_bh; 5866 struct super_block *sb = inode->i_sb; 5867 struct ext4_group_desc *gdp; 5868 struct buffer_head *gdp_bh; 5869 ext4_group_t group; 5870 ext4_grpblk_t blkoff; 5871 int already_freed = 0, err, i; 5872 5873 ext4_get_group_no_and_offset(sb, block, &group, &blkoff); 5874 bitmap_bh = ext4_read_block_bitmap(sb, group); 5875 if (IS_ERR(bitmap_bh)) { 5876 err = PTR_ERR(bitmap_bh); 5877 pr_warn("Failed to read block bitmap\n"); 5878 return; 5879 } 5880 gdp = ext4_get_group_desc(sb, group, &gdp_bh); 5881 if (!gdp) 5882 return; 5883 5884 for (i = 0; i < count; i++) { 5885 if (!mb_test_bit(blkoff + i, bitmap_bh->b_data)) 5886 already_freed++; 5887 } 5888 mb_clear_bits(bitmap_bh->b_data, blkoff, count); 5889 err = ext4_handle_dirty_metadata(NULL, NULL, bitmap_bh); 5890 if (err) 5891 return; 5892 ext4_free_group_clusters_set( 5893 sb, gdp, ext4_free_group_clusters(sb, gdp) + 5894 count - already_freed); 5895 ext4_block_bitmap_csum_set(sb, group, gdp, bitmap_bh); 5896 ext4_group_desc_csum_set(sb, group, gdp); 5897 ext4_handle_dirty_metadata(NULL, NULL, gdp_bh); 5898 sync_dirty_buffer(bitmap_bh); 5899 sync_dirty_buffer(gdp_bh); 5900 brelse(bitmap_bh); 5901 } 5902 5903 /** 5904 * ext4_mb_clear_bb() -- helper function for freeing blocks. 5905 * Used by ext4_free_blocks() 5906 * @handle: handle for this transaction 5907 * @inode: inode 5908 * @block: starting physical block to be freed 5909 * @count: number of blocks to be freed 5910 * @flags: flags used by ext4_free_blocks 5911 */ 5912 static void ext4_mb_clear_bb(handle_t *handle, struct inode *inode, 5913 ext4_fsblk_t block, unsigned long count, 5914 int flags) 5915 { 5916 struct buffer_head *bitmap_bh = NULL; 5917 struct super_block *sb = inode->i_sb; 5918 struct ext4_group_desc *gdp; 5919 unsigned int overflow; 5920 ext4_grpblk_t bit; 5921 struct buffer_head *gd_bh; 5922 ext4_group_t block_group; 5923 struct ext4_sb_info *sbi; 5924 struct ext4_buddy e4b; 5925 unsigned int count_clusters; 5926 int err = 0; 5927 int ret; 5928 5929 sbi = EXT4_SB(sb); 5930 5931 do_more: 5932 overflow = 0; 5933 ext4_get_group_no_and_offset(sb, block, &block_group, &bit); 5934 5935 if (unlikely(EXT4_MB_GRP_BBITMAP_CORRUPT( 5936 ext4_get_group_info(sb, block_group)))) 5937 return; 5938 5939 /* 5940 * Check to see if we are freeing blocks across a group 5941 * boundary. 5942 */ 5943 if (EXT4_C2B(sbi, bit) + count > EXT4_BLOCKS_PER_GROUP(sb)) { 5944 overflow = EXT4_C2B(sbi, bit) + count - 5945 EXT4_BLOCKS_PER_GROUP(sb); 5946 count -= overflow; 5947 } 5948 count_clusters = EXT4_NUM_B2C(sbi, count); 5949 bitmap_bh = ext4_read_block_bitmap(sb, block_group); 5950 if (IS_ERR(bitmap_bh)) { 5951 err = PTR_ERR(bitmap_bh); 5952 bitmap_bh = NULL; 5953 goto error_return; 5954 } 5955 gdp = ext4_get_group_desc(sb, block_group, &gd_bh); 5956 if (!gdp) { 5957 err = -EIO; 5958 goto error_return; 5959 } 5960 5961 if (!ext4_inode_block_valid(inode, block, count)) { 5962 ext4_error(sb, "Freeing blocks in system zone - " 5963 "Block = %llu, count = %lu", block, count); 5964 /* err = 0. ext4_std_error should be a no op */ 5965 goto error_return; 5966 } 5967 5968 BUFFER_TRACE(bitmap_bh, "getting write access"); 5969 err = ext4_journal_get_write_access(handle, sb, bitmap_bh, 5970 EXT4_JTR_NONE); 5971 if (err) 5972 goto error_return; 5973 5974 /* 5975 * We are about to modify some metadata. Call the journal APIs 5976 * to unshare ->b_data if a currently-committing transaction is 5977 * using it 5978 */ 5979 BUFFER_TRACE(gd_bh, "get_write_access"); 5980 err = ext4_journal_get_write_access(handle, sb, gd_bh, EXT4_JTR_NONE); 5981 if (err) 5982 goto error_return; 5983 #ifdef AGGRESSIVE_CHECK 5984 { 5985 int i; 5986 for (i = 0; i < count_clusters; i++) 5987 BUG_ON(!mb_test_bit(bit + i, bitmap_bh->b_data)); 5988 } 5989 #endif 5990 trace_ext4_mballoc_free(sb, inode, block_group, bit, count_clusters); 5991 5992 /* __GFP_NOFAIL: retry infinitely, ignore TIF_MEMDIE and memcg limit. */ 5993 err = ext4_mb_load_buddy_gfp(sb, block_group, &e4b, 5994 GFP_NOFS|__GFP_NOFAIL); 5995 if (err) 5996 goto error_return; 5997 5998 /* 5999 * We need to make sure we don't reuse the freed block until after the 6000 * transaction is committed. We make an exception if the inode is to be 6001 * written in writeback mode since writeback mode has weak data 6002 * consistency guarantees. 6003 */ 6004 if (ext4_handle_valid(handle) && 6005 ((flags & EXT4_FREE_BLOCKS_METADATA) || 6006 !ext4_should_writeback_data(inode))) { 6007 struct ext4_free_data *new_entry; 6008 /* 6009 * We use __GFP_NOFAIL because ext4_free_blocks() is not allowed 6010 * to fail. 6011 */ 6012 new_entry = kmem_cache_alloc(ext4_free_data_cachep, 6013 GFP_NOFS|__GFP_NOFAIL); 6014 new_entry->efd_start_cluster = bit; 6015 new_entry->efd_group = block_group; 6016 new_entry->efd_count = count_clusters; 6017 new_entry->efd_tid = handle->h_transaction->t_tid; 6018 6019 ext4_lock_group(sb, block_group); 6020 mb_clear_bits(bitmap_bh->b_data, bit, count_clusters); 6021 ext4_mb_free_metadata(handle, &e4b, new_entry); 6022 } else { 6023 /* need to update group_info->bb_free and bitmap 6024 * with group lock held. generate_buddy look at 6025 * them with group lock_held 6026 */ 6027 if (test_opt(sb, DISCARD)) { 6028 err = ext4_issue_discard(sb, block_group, bit, count, 6029 NULL); 6030 if (err && err != -EOPNOTSUPP) 6031 ext4_msg(sb, KERN_WARNING, "discard request in" 6032 " group:%u block:%d count:%lu failed" 6033 " with %d", block_group, bit, count, 6034 err); 6035 } else 6036 EXT4_MB_GRP_CLEAR_TRIMMED(e4b.bd_info); 6037 6038 ext4_lock_group(sb, block_group); 6039 mb_clear_bits(bitmap_bh->b_data, bit, count_clusters); 6040 mb_free_blocks(inode, &e4b, bit, count_clusters); 6041 } 6042 6043 ret = ext4_free_group_clusters(sb, gdp) + count_clusters; 6044 ext4_free_group_clusters_set(sb, gdp, ret); 6045 ext4_block_bitmap_csum_set(sb, block_group, gdp, bitmap_bh); 6046 ext4_group_desc_csum_set(sb, block_group, gdp); 6047 ext4_unlock_group(sb, block_group); 6048 6049 if (sbi->s_log_groups_per_flex) { 6050 ext4_group_t flex_group = ext4_flex_group(sbi, block_group); 6051 atomic64_add(count_clusters, 6052 &sbi_array_rcu_deref(sbi, s_flex_groups, 6053 flex_group)->free_clusters); 6054 } 6055 6056 /* 6057 * on a bigalloc file system, defer the s_freeclusters_counter 6058 * update to the caller (ext4_remove_space and friends) so they 6059 * can determine if a cluster freed here should be rereserved 6060 */ 6061 if (!(flags & EXT4_FREE_BLOCKS_RERESERVE_CLUSTER)) { 6062 if (!(flags & EXT4_FREE_BLOCKS_NO_QUOT_UPDATE)) 6063 dquot_free_block(inode, EXT4_C2B(sbi, count_clusters)); 6064 percpu_counter_add(&sbi->s_freeclusters_counter, 6065 count_clusters); 6066 } 6067 6068 ext4_mb_unload_buddy(&e4b); 6069 6070 /* We dirtied the bitmap block */ 6071 BUFFER_TRACE(bitmap_bh, "dirtied bitmap block"); 6072 err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh); 6073 6074 /* And the group descriptor block */ 6075 BUFFER_TRACE(gd_bh, "dirtied group descriptor block"); 6076 ret = ext4_handle_dirty_metadata(handle, NULL, gd_bh); 6077 if (!err) 6078 err = ret; 6079 6080 if (overflow && !err) { 6081 block += count; 6082 count = overflow; 6083 put_bh(bitmap_bh); 6084 goto do_more; 6085 } 6086 error_return: 6087 brelse(bitmap_bh); 6088 ext4_std_error(sb, err); 6089 return; 6090 } 6091 6092 /** 6093 * ext4_free_blocks() -- Free given blocks and update quota 6094 * @handle: handle for this transaction 6095 * @inode: inode 6096 * @bh: optional buffer of the block to be freed 6097 * @block: starting physical block to be freed 6098 * @count: number of blocks to be freed 6099 * @flags: flags used by ext4_free_blocks 6100 */ 6101 void ext4_free_blocks(handle_t *handle, struct inode *inode, 6102 struct buffer_head *bh, ext4_fsblk_t block, 6103 unsigned long count, int flags) 6104 { 6105 struct super_block *sb = inode->i_sb; 6106 unsigned int overflow; 6107 struct ext4_sb_info *sbi; 6108 6109 sbi = EXT4_SB(sb); 6110 6111 if (sbi->s_mount_state & EXT4_FC_REPLAY) { 6112 ext4_free_blocks_simple(inode, block, count); 6113 return; 6114 } 6115 6116 might_sleep(); 6117 if (bh) { 6118 if (block) 6119 BUG_ON(block != bh->b_blocknr); 6120 else 6121 block = bh->b_blocknr; 6122 } 6123 6124 if (!(flags & EXT4_FREE_BLOCKS_VALIDATED) && 6125 !ext4_inode_block_valid(inode, block, count)) { 6126 ext4_error(sb, "Freeing blocks not in datazone - " 6127 "block = %llu, count = %lu", block, count); 6128 return; 6129 } 6130 6131 ext4_debug("freeing block %llu\n", block); 6132 trace_ext4_free_blocks(inode, block, count, flags); 6133 6134 if (bh && (flags & EXT4_FREE_BLOCKS_FORGET)) { 6135 BUG_ON(count > 1); 6136 6137 ext4_forget(handle, flags & EXT4_FREE_BLOCKS_METADATA, 6138 inode, bh, block); 6139 } 6140 6141 /* 6142 * If the extent to be freed does not begin on a cluster 6143 * boundary, we need to deal with partial clusters at the 6144 * beginning and end of the extent. Normally we will free 6145 * blocks at the beginning or the end unless we are explicitly 6146 * requested to avoid doing so. 6147 */ 6148 overflow = EXT4_PBLK_COFF(sbi, block); 6149 if (overflow) { 6150 if (flags & EXT4_FREE_BLOCKS_NOFREE_FIRST_CLUSTER) { 6151 overflow = sbi->s_cluster_ratio - overflow; 6152 block += overflow; 6153 if (count > overflow) 6154 count -= overflow; 6155 else 6156 return; 6157 } else { 6158 block -= overflow; 6159 count += overflow; 6160 } 6161 } 6162 overflow = EXT4_LBLK_COFF(sbi, count); 6163 if (overflow) { 6164 if (flags & EXT4_FREE_BLOCKS_NOFREE_LAST_CLUSTER) { 6165 if (count > overflow) 6166 count -= overflow; 6167 else 6168 return; 6169 } else 6170 count += sbi->s_cluster_ratio - overflow; 6171 } 6172 6173 if (!bh && (flags & EXT4_FREE_BLOCKS_FORGET)) { 6174 int i; 6175 int is_metadata = flags & EXT4_FREE_BLOCKS_METADATA; 6176 6177 for (i = 0; i < count; i++) { 6178 cond_resched(); 6179 if (is_metadata) 6180 bh = sb_find_get_block(inode->i_sb, block + i); 6181 ext4_forget(handle, is_metadata, inode, bh, block + i); 6182 } 6183 } 6184 6185 ext4_mb_clear_bb(handle, inode, block, count, flags); 6186 return; 6187 } 6188 6189 /** 6190 * ext4_group_add_blocks() -- Add given blocks to an existing group 6191 * @handle: handle to this transaction 6192 * @sb: super block 6193 * @block: start physical block to add to the block group 6194 * @count: number of blocks to free 6195 * 6196 * This marks the blocks as free in the bitmap and buddy. 6197 */ 6198 int ext4_group_add_blocks(handle_t *handle, struct super_block *sb, 6199 ext4_fsblk_t block, unsigned long count) 6200 { 6201 struct buffer_head *bitmap_bh = NULL; 6202 struct buffer_head *gd_bh; 6203 ext4_group_t block_group; 6204 ext4_grpblk_t bit; 6205 unsigned int i; 6206 struct ext4_group_desc *desc; 6207 struct ext4_sb_info *sbi = EXT4_SB(sb); 6208 struct ext4_buddy e4b; 6209 int err = 0, ret, free_clusters_count; 6210 ext4_grpblk_t clusters_freed; 6211 ext4_fsblk_t first_cluster = EXT4_B2C(sbi, block); 6212 ext4_fsblk_t last_cluster = EXT4_B2C(sbi, block + count - 1); 6213 unsigned long cluster_count = last_cluster - first_cluster + 1; 6214 6215 ext4_debug("Adding block(s) %llu-%llu\n", block, block + count - 1); 6216 6217 if (count == 0) 6218 return 0; 6219 6220 ext4_get_group_no_and_offset(sb, block, &block_group, &bit); 6221 /* 6222 * Check to see if we are freeing blocks across a group 6223 * boundary. 6224 */ 6225 if (bit + cluster_count > EXT4_CLUSTERS_PER_GROUP(sb)) { 6226 ext4_warning(sb, "too many blocks added to group %u", 6227 block_group); 6228 err = -EINVAL; 6229 goto error_return; 6230 } 6231 6232 bitmap_bh = ext4_read_block_bitmap(sb, block_group); 6233 if (IS_ERR(bitmap_bh)) { 6234 err = PTR_ERR(bitmap_bh); 6235 bitmap_bh = NULL; 6236 goto error_return; 6237 } 6238 6239 desc = ext4_get_group_desc(sb, block_group, &gd_bh); 6240 if (!desc) { 6241 err = -EIO; 6242 goto error_return; 6243 } 6244 6245 if (!ext4_sb_block_valid(sb, NULL, block, count)) { 6246 ext4_error(sb, "Adding blocks in system zones - " 6247 "Block = %llu, count = %lu", 6248 block, count); 6249 err = -EINVAL; 6250 goto error_return; 6251 } 6252 6253 BUFFER_TRACE(bitmap_bh, "getting write access"); 6254 err = ext4_journal_get_write_access(handle, sb, bitmap_bh, 6255 EXT4_JTR_NONE); 6256 if (err) 6257 goto error_return; 6258 6259 /* 6260 * We are about to modify some metadata. Call the journal APIs 6261 * to unshare ->b_data if a currently-committing transaction is 6262 * using it 6263 */ 6264 BUFFER_TRACE(gd_bh, "get_write_access"); 6265 err = ext4_journal_get_write_access(handle, sb, gd_bh, EXT4_JTR_NONE); 6266 if (err) 6267 goto error_return; 6268 6269 for (i = 0, clusters_freed = 0; i < cluster_count; i++) { 6270 BUFFER_TRACE(bitmap_bh, "clear bit"); 6271 if (!mb_test_bit(bit + i, bitmap_bh->b_data)) { 6272 ext4_error(sb, "bit already cleared for block %llu", 6273 (ext4_fsblk_t)(block + i)); 6274 BUFFER_TRACE(bitmap_bh, "bit already cleared"); 6275 } else { 6276 clusters_freed++; 6277 } 6278 } 6279 6280 err = ext4_mb_load_buddy(sb, block_group, &e4b); 6281 if (err) 6282 goto error_return; 6283 6284 /* 6285 * need to update group_info->bb_free and bitmap 6286 * with group lock held. generate_buddy look at 6287 * them with group lock_held 6288 */ 6289 ext4_lock_group(sb, block_group); 6290 mb_clear_bits(bitmap_bh->b_data, bit, cluster_count); 6291 mb_free_blocks(NULL, &e4b, bit, cluster_count); 6292 free_clusters_count = clusters_freed + 6293 ext4_free_group_clusters(sb, desc); 6294 ext4_free_group_clusters_set(sb, desc, free_clusters_count); 6295 ext4_block_bitmap_csum_set(sb, block_group, desc, bitmap_bh); 6296 ext4_group_desc_csum_set(sb, block_group, desc); 6297 ext4_unlock_group(sb, block_group); 6298 percpu_counter_add(&sbi->s_freeclusters_counter, 6299 clusters_freed); 6300 6301 if (sbi->s_log_groups_per_flex) { 6302 ext4_group_t flex_group = ext4_flex_group(sbi, block_group); 6303 atomic64_add(clusters_freed, 6304 &sbi_array_rcu_deref(sbi, s_flex_groups, 6305 flex_group)->free_clusters); 6306 } 6307 6308 ext4_mb_unload_buddy(&e4b); 6309 6310 /* We dirtied the bitmap block */ 6311 BUFFER_TRACE(bitmap_bh, "dirtied bitmap block"); 6312 err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh); 6313 6314 /* And the group descriptor block */ 6315 BUFFER_TRACE(gd_bh, "dirtied group descriptor block"); 6316 ret = ext4_handle_dirty_metadata(handle, NULL, gd_bh); 6317 if (!err) 6318 err = ret; 6319 6320 error_return: 6321 brelse(bitmap_bh); 6322 ext4_std_error(sb, err); 6323 return err; 6324 } 6325 6326 /** 6327 * ext4_trim_extent -- function to TRIM one single free extent in the group 6328 * @sb: super block for the file system 6329 * @start: starting block of the free extent in the alloc. group 6330 * @count: number of blocks to TRIM 6331 * @e4b: ext4 buddy for the group 6332 * 6333 * Trim "count" blocks starting at "start" in the "group". To assure that no 6334 * one will allocate those blocks, mark it as used in buddy bitmap. This must 6335 * be called with under the group lock. 6336 */ 6337 static int ext4_trim_extent(struct super_block *sb, 6338 int start, int count, struct ext4_buddy *e4b) 6339 __releases(bitlock) 6340 __acquires(bitlock) 6341 { 6342 struct ext4_free_extent ex; 6343 ext4_group_t group = e4b->bd_group; 6344 int ret = 0; 6345 6346 trace_ext4_trim_extent(sb, group, start, count); 6347 6348 assert_spin_locked(ext4_group_lock_ptr(sb, group)); 6349 6350 ex.fe_start = start; 6351 ex.fe_group = group; 6352 ex.fe_len = count; 6353 6354 /* 6355 * Mark blocks used, so no one can reuse them while 6356 * being trimmed. 6357 */ 6358 mb_mark_used(e4b, &ex); 6359 ext4_unlock_group(sb, group); 6360 ret = ext4_issue_discard(sb, group, start, count, NULL); 6361 ext4_lock_group(sb, group); 6362 mb_free_blocks(NULL, e4b, start, ex.fe_len); 6363 return ret; 6364 } 6365 6366 static int ext4_try_to_trim_range(struct super_block *sb, 6367 struct ext4_buddy *e4b, ext4_grpblk_t start, 6368 ext4_grpblk_t max, ext4_grpblk_t minblocks) 6369 __acquires(ext4_group_lock_ptr(sb, e4b->bd_group)) 6370 __releases(ext4_group_lock_ptr(sb, e4b->bd_group)) 6371 { 6372 ext4_grpblk_t next, count, free_count; 6373 void *bitmap; 6374 6375 bitmap = e4b->bd_bitmap; 6376 start = (e4b->bd_info->bb_first_free > start) ? 6377 e4b->bd_info->bb_first_free : start; 6378 count = 0; 6379 free_count = 0; 6380 6381 while (start <= max) { 6382 start = mb_find_next_zero_bit(bitmap, max + 1, start); 6383 if (start > max) 6384 break; 6385 next = mb_find_next_bit(bitmap, max + 1, start); 6386 6387 if ((next - start) >= minblocks) { 6388 int ret = ext4_trim_extent(sb, start, next - start, e4b); 6389 6390 if (ret && ret != -EOPNOTSUPP) 6391 break; 6392 count += next - start; 6393 } 6394 free_count += next - start; 6395 start = next + 1; 6396 6397 if (fatal_signal_pending(current)) { 6398 count = -ERESTARTSYS; 6399 break; 6400 } 6401 6402 if (need_resched()) { 6403 ext4_unlock_group(sb, e4b->bd_group); 6404 cond_resched(); 6405 ext4_lock_group(sb, e4b->bd_group); 6406 } 6407 6408 if ((e4b->bd_info->bb_free - free_count) < minblocks) 6409 break; 6410 } 6411 6412 return count; 6413 } 6414 6415 /** 6416 * ext4_trim_all_free -- function to trim all free space in alloc. group 6417 * @sb: super block for file system 6418 * @group: group to be trimmed 6419 * @start: first group block to examine 6420 * @max: last group block to examine 6421 * @minblocks: minimum extent block count 6422 * @set_trimmed: set the trimmed flag if at least one block is trimmed 6423 * 6424 * ext4_trim_all_free walks through group's block bitmap searching for free 6425 * extents. When the free extent is found, mark it as used in group buddy 6426 * bitmap. Then issue a TRIM command on this extent and free the extent in 6427 * the group buddy bitmap. 6428 */ 6429 static ext4_grpblk_t 6430 ext4_trim_all_free(struct super_block *sb, ext4_group_t group, 6431 ext4_grpblk_t start, ext4_grpblk_t max, 6432 ext4_grpblk_t minblocks, bool set_trimmed) 6433 { 6434 struct ext4_buddy e4b; 6435 int ret; 6436 6437 trace_ext4_trim_all_free(sb, group, start, max); 6438 6439 ret = ext4_mb_load_buddy(sb, group, &e4b); 6440 if (ret) { 6441 ext4_warning(sb, "Error %d loading buddy information for %u", 6442 ret, group); 6443 return ret; 6444 } 6445 6446 ext4_lock_group(sb, group); 6447 6448 if (!EXT4_MB_GRP_WAS_TRIMMED(e4b.bd_info) || 6449 minblocks < EXT4_SB(sb)->s_last_trim_minblks) { 6450 ret = ext4_try_to_trim_range(sb, &e4b, start, max, minblocks); 6451 if (ret >= 0 && set_trimmed) 6452 EXT4_MB_GRP_SET_TRIMMED(e4b.bd_info); 6453 } else { 6454 ret = 0; 6455 } 6456 6457 ext4_unlock_group(sb, group); 6458 ext4_mb_unload_buddy(&e4b); 6459 6460 ext4_debug("trimmed %d blocks in the group %d\n", 6461 ret, group); 6462 6463 return ret; 6464 } 6465 6466 /** 6467 * ext4_trim_fs() -- trim ioctl handle function 6468 * @sb: superblock for filesystem 6469 * @range: fstrim_range structure 6470 * 6471 * start: First Byte to trim 6472 * len: number of Bytes to trim from start 6473 * minlen: minimum extent length in Bytes 6474 * ext4_trim_fs goes through all allocation groups containing Bytes from 6475 * start to start+len. For each such a group ext4_trim_all_free function 6476 * is invoked to trim all free space. 6477 */ 6478 int ext4_trim_fs(struct super_block *sb, struct fstrim_range *range) 6479 { 6480 unsigned int discard_granularity = bdev_discard_granularity(sb->s_bdev); 6481 struct ext4_group_info *grp; 6482 ext4_group_t group, first_group, last_group; 6483 ext4_grpblk_t cnt = 0, first_cluster, last_cluster; 6484 uint64_t start, end, minlen, trimmed = 0; 6485 ext4_fsblk_t first_data_blk = 6486 le32_to_cpu(EXT4_SB(sb)->s_es->s_first_data_block); 6487 ext4_fsblk_t max_blks = ext4_blocks_count(EXT4_SB(sb)->s_es); 6488 bool whole_group, eof = false; 6489 int ret = 0; 6490 6491 start = range->start >> sb->s_blocksize_bits; 6492 end = start + (range->len >> sb->s_blocksize_bits) - 1; 6493 minlen = EXT4_NUM_B2C(EXT4_SB(sb), 6494 range->minlen >> sb->s_blocksize_bits); 6495 6496 if (minlen > EXT4_CLUSTERS_PER_GROUP(sb) || 6497 start >= max_blks || 6498 range->len < sb->s_blocksize) 6499 return -EINVAL; 6500 /* No point to try to trim less than discard granularity */ 6501 if (range->minlen < discard_granularity) { 6502 minlen = EXT4_NUM_B2C(EXT4_SB(sb), 6503 discard_granularity >> sb->s_blocksize_bits); 6504 if (minlen > EXT4_CLUSTERS_PER_GROUP(sb)) 6505 goto out; 6506 } 6507 if (end >= max_blks - 1) { 6508 end = max_blks - 1; 6509 eof = true; 6510 } 6511 if (end <= first_data_blk) 6512 goto out; 6513 if (start < first_data_blk) 6514 start = first_data_blk; 6515 6516 /* Determine first and last group to examine based on start and end */ 6517 ext4_get_group_no_and_offset(sb, (ext4_fsblk_t) start, 6518 &first_group, &first_cluster); 6519 ext4_get_group_no_and_offset(sb, (ext4_fsblk_t) end, 6520 &last_group, &last_cluster); 6521 6522 /* end now represents the last cluster to discard in this group */ 6523 end = EXT4_CLUSTERS_PER_GROUP(sb) - 1; 6524 whole_group = true; 6525 6526 for (group = first_group; group <= last_group; group++) { 6527 grp = ext4_get_group_info(sb, group); 6528 /* We only do this if the grp has never been initialized */ 6529 if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) { 6530 ret = ext4_mb_init_group(sb, group, GFP_NOFS); 6531 if (ret) 6532 break; 6533 } 6534 6535 /* 6536 * For all the groups except the last one, last cluster will 6537 * always be EXT4_CLUSTERS_PER_GROUP(sb)-1, so we only need to 6538 * change it for the last group, note that last_cluster is 6539 * already computed earlier by ext4_get_group_no_and_offset() 6540 */ 6541 if (group == last_group) { 6542 end = last_cluster; 6543 whole_group = eof ? true : end == EXT4_CLUSTERS_PER_GROUP(sb) - 1; 6544 } 6545 if (grp->bb_free >= minlen) { 6546 cnt = ext4_trim_all_free(sb, group, first_cluster, 6547 end, minlen, whole_group); 6548 if (cnt < 0) { 6549 ret = cnt; 6550 break; 6551 } 6552 trimmed += cnt; 6553 } 6554 6555 /* 6556 * For every group except the first one, we are sure 6557 * that the first cluster to discard will be cluster #0. 6558 */ 6559 first_cluster = 0; 6560 } 6561 6562 if (!ret) 6563 EXT4_SB(sb)->s_last_trim_minblks = minlen; 6564 6565 out: 6566 range->len = EXT4_C2B(EXT4_SB(sb), trimmed) << sb->s_blocksize_bits; 6567 return ret; 6568 } 6569 6570 /* Iterate all the free extents in the group. */ 6571 int 6572 ext4_mballoc_query_range( 6573 struct super_block *sb, 6574 ext4_group_t group, 6575 ext4_grpblk_t start, 6576 ext4_grpblk_t end, 6577 ext4_mballoc_query_range_fn formatter, 6578 void *priv) 6579 { 6580 void *bitmap; 6581 ext4_grpblk_t next; 6582 struct ext4_buddy e4b; 6583 int error; 6584 6585 error = ext4_mb_load_buddy(sb, group, &e4b); 6586 if (error) 6587 return error; 6588 bitmap = e4b.bd_bitmap; 6589 6590 ext4_lock_group(sb, group); 6591 6592 start = (e4b.bd_info->bb_first_free > start) ? 6593 e4b.bd_info->bb_first_free : start; 6594 if (end >= EXT4_CLUSTERS_PER_GROUP(sb)) 6595 end = EXT4_CLUSTERS_PER_GROUP(sb) - 1; 6596 6597 while (start <= end) { 6598 start = mb_find_next_zero_bit(bitmap, end + 1, start); 6599 if (start > end) 6600 break; 6601 next = mb_find_next_bit(bitmap, end + 1, start); 6602 6603 ext4_unlock_group(sb, group); 6604 error = formatter(sb, group, start, next - start, priv); 6605 if (error) 6606 goto out_unload; 6607 ext4_lock_group(sb, group); 6608 6609 start = next + 1; 6610 } 6611 6612 ext4_unlock_group(sb, group); 6613 out_unload: 6614 ext4_mb_unload_buddy(&e4b); 6615 6616 return error; 6617 } 6618