1 // SPDX-License-Identifier: GPL-2.0 2 3 #include "misc.h" 4 #include "ctree.h" 5 #include "space-info.h" 6 #include "sysfs.h" 7 #include "volumes.h" 8 #include "free-space-cache.h" 9 #include "ordered-data.h" 10 #include "transaction.h" 11 #include "block-group.h" 12 #include "zoned.h" 13 #include "fs.h" 14 #include "accessors.h" 15 #include "extent-tree.h" 16 17 /* 18 * HOW DOES SPACE RESERVATION WORK 19 * 20 * If you want to know about delalloc specifically, there is a separate comment 21 * for that with the delalloc code. This comment is about how the whole system 22 * works generally. 23 * 24 * BASIC CONCEPTS 25 * 26 * 1) space_info. This is the ultimate arbiter of how much space we can use. 27 * There's a description of the bytes_ fields with the struct declaration, 28 * refer to that for specifics on each field. Suffice it to say that for 29 * reservations we care about total_bytes - SUM(space_info->bytes_) when 30 * determining if there is space to make an allocation. There is a space_info 31 * for METADATA, SYSTEM, and DATA areas. 32 * 33 * 2) block_rsv's. These are basically buckets for every different type of 34 * metadata reservation we have. You can see the comment in the block_rsv 35 * code on the rules for each type, but generally block_rsv->reserved is how 36 * much space is accounted for in space_info->bytes_may_use. 37 * 38 * 3) btrfs_calc*_size. These are the worst case calculations we used based 39 * on the number of items we will want to modify. We have one for changing 40 * items, and one for inserting new items. Generally we use these helpers to 41 * determine the size of the block reserves, and then use the actual bytes 42 * values to adjust the space_info counters. 43 * 44 * MAKING RESERVATIONS, THE NORMAL CASE 45 * 46 * We call into either btrfs_reserve_data_bytes() or 47 * btrfs_reserve_metadata_bytes(), depending on which we're looking for, with 48 * num_bytes we want to reserve. 49 * 50 * ->reserve 51 * space_info->bytes_may_reserve += num_bytes 52 * 53 * ->extent allocation 54 * Call btrfs_add_reserved_bytes() which does 55 * space_info->bytes_may_reserve -= num_bytes 56 * space_info->bytes_reserved += extent_bytes 57 * 58 * ->insert reference 59 * Call btrfs_update_block_group() which does 60 * space_info->bytes_reserved -= extent_bytes 61 * space_info->bytes_used += extent_bytes 62 * 63 * MAKING RESERVATIONS, FLUSHING NORMALLY (non-priority) 64 * 65 * Assume we are unable to simply make the reservation because we do not have 66 * enough space 67 * 68 * -> __reserve_bytes 69 * create a reserve_ticket with ->bytes set to our reservation, add it to 70 * the tail of space_info->tickets, kick async flush thread 71 * 72 * ->handle_reserve_ticket 73 * wait on ticket->wait for ->bytes to be reduced to 0, or ->error to be set 74 * on the ticket. 75 * 76 * -> btrfs_async_reclaim_metadata_space/btrfs_async_reclaim_data_space 77 * Flushes various things attempting to free up space. 78 * 79 * -> btrfs_try_granting_tickets() 80 * This is called by anything that either subtracts space from 81 * space_info->bytes_may_use, ->bytes_pinned, etc, or adds to the 82 * space_info->total_bytes. This loops through the ->priority_tickets and 83 * then the ->tickets list checking to see if the reservation can be 84 * completed. If it can the space is added to space_info->bytes_may_use and 85 * the ticket is woken up. 86 * 87 * -> ticket wakeup 88 * Check if ->bytes == 0, if it does we got our reservation and we can carry 89 * on, if not return the appropriate error (ENOSPC, but can be EINTR if we 90 * were interrupted.) 91 * 92 * MAKING RESERVATIONS, FLUSHING HIGH PRIORITY 93 * 94 * Same as the above, except we add ourselves to the 95 * space_info->priority_tickets, and we do not use ticket->wait, we simply 96 * call flush_space() ourselves for the states that are safe for us to call 97 * without deadlocking and hope for the best. 98 * 99 * THE FLUSHING STATES 100 * 101 * Generally speaking we will have two cases for each state, a "nice" state 102 * and a "ALL THE THINGS" state. In btrfs we delay a lot of work in order to 103 * reduce the locking over head on the various trees, and even to keep from 104 * doing any work at all in the case of delayed refs. Each of these delayed 105 * things however hold reservations, and so letting them run allows us to 106 * reclaim space so we can make new reservations. 107 * 108 * FLUSH_DELAYED_ITEMS 109 * Every inode has a delayed item to update the inode. Take a simple write 110 * for example, we would update the inode item at write time to update the 111 * mtime, and then again at finish_ordered_io() time in order to update the 112 * isize or bytes. We keep these delayed items to coalesce these operations 113 * into a single operation done on demand. These are an easy way to reclaim 114 * metadata space. 115 * 116 * FLUSH_DELALLOC 117 * Look at the delalloc comment to get an idea of how much space is reserved 118 * for delayed allocation. We can reclaim some of this space simply by 119 * running delalloc, but usually we need to wait for ordered extents to 120 * reclaim the bulk of this space. 121 * 122 * FLUSH_DELAYED_REFS 123 * We have a block reserve for the outstanding delayed refs space, and every 124 * delayed ref operation holds a reservation. Running these is a quick way 125 * to reclaim space, but we want to hold this until the end because COW can 126 * churn a lot and we can avoid making some extent tree modifications if we 127 * are able to delay for as long as possible. 128 * 129 * ALLOC_CHUNK 130 * We will skip this the first time through space reservation, because of 131 * overcommit and we don't want to have a lot of useless metadata space when 132 * our worst case reservations will likely never come true. 133 * 134 * RUN_DELAYED_IPUTS 135 * If we're freeing inodes we're likely freeing checksums, file extent 136 * items, and extent tree items. Loads of space could be freed up by these 137 * operations, however they won't be usable until the transaction commits. 138 * 139 * COMMIT_TRANS 140 * This will commit the transaction. Historically we had a lot of logic 141 * surrounding whether or not we'd commit the transaction, but this waits born 142 * out of a pre-tickets era where we could end up committing the transaction 143 * thousands of times in a row without making progress. Now thanks to our 144 * ticketing system we know if we're not making progress and can error 145 * everybody out after a few commits rather than burning the disk hoping for 146 * a different answer. 147 * 148 * OVERCOMMIT 149 * 150 * Because we hold so many reservations for metadata we will allow you to 151 * reserve more space than is currently free in the currently allocate 152 * metadata space. This only happens with metadata, data does not allow 153 * overcommitting. 154 * 155 * You can see the current logic for when we allow overcommit in 156 * btrfs_can_overcommit(), but it only applies to unallocated space. If there 157 * is no unallocated space to be had, all reservations are kept within the 158 * free space in the allocated metadata chunks. 159 * 160 * Because of overcommitting, you generally want to use the 161 * btrfs_can_overcommit() logic for metadata allocations, as it does the right 162 * thing with or without extra unallocated space. 163 */ 164 165 u64 __pure btrfs_space_info_used(struct btrfs_space_info *s_info, 166 bool may_use_included) 167 { 168 ASSERT(s_info); 169 return s_info->bytes_used + s_info->bytes_reserved + 170 s_info->bytes_pinned + s_info->bytes_readonly + 171 s_info->bytes_zone_unusable + 172 (may_use_included ? s_info->bytes_may_use : 0); 173 } 174 175 /* 176 * after adding space to the filesystem, we need to clear the full flags 177 * on all the space infos. 178 */ 179 void btrfs_clear_space_info_full(struct btrfs_fs_info *info) 180 { 181 struct list_head *head = &info->space_info; 182 struct btrfs_space_info *found; 183 184 list_for_each_entry(found, head, list) 185 found->full = 0; 186 } 187 188 /* 189 * Block groups with more than this value (percents) of unusable space will be 190 * scheduled for background reclaim. 191 */ 192 #define BTRFS_DEFAULT_ZONED_RECLAIM_THRESH (75) 193 194 /* 195 * Calculate chunk size depending on volume type (regular or zoned). 196 */ 197 static u64 calc_chunk_size(const struct btrfs_fs_info *fs_info, u64 flags) 198 { 199 if (btrfs_is_zoned(fs_info)) 200 return fs_info->zone_size; 201 202 ASSERT(flags & BTRFS_BLOCK_GROUP_TYPE_MASK); 203 204 if (flags & BTRFS_BLOCK_GROUP_DATA) 205 return BTRFS_MAX_DATA_CHUNK_SIZE; 206 else if (flags & BTRFS_BLOCK_GROUP_SYSTEM) 207 return SZ_32M; 208 209 /* Handle BTRFS_BLOCK_GROUP_METADATA */ 210 if (fs_info->fs_devices->total_rw_bytes > 50ULL * SZ_1G) 211 return SZ_1G; 212 213 return SZ_256M; 214 } 215 216 /* 217 * Update default chunk size. 218 */ 219 void btrfs_update_space_info_chunk_size(struct btrfs_space_info *space_info, 220 u64 chunk_size) 221 { 222 WRITE_ONCE(space_info->chunk_size, chunk_size); 223 } 224 225 static int create_space_info(struct btrfs_fs_info *info, u64 flags) 226 { 227 228 struct btrfs_space_info *space_info; 229 int i; 230 int ret; 231 232 space_info = kzalloc(sizeof(*space_info), GFP_NOFS); 233 if (!space_info) 234 return -ENOMEM; 235 236 for (i = 0; i < BTRFS_NR_RAID_TYPES; i++) 237 INIT_LIST_HEAD(&space_info->block_groups[i]); 238 init_rwsem(&space_info->groups_sem); 239 spin_lock_init(&space_info->lock); 240 space_info->flags = flags & BTRFS_BLOCK_GROUP_TYPE_MASK; 241 space_info->force_alloc = CHUNK_ALLOC_NO_FORCE; 242 INIT_LIST_HEAD(&space_info->ro_bgs); 243 INIT_LIST_HEAD(&space_info->tickets); 244 INIT_LIST_HEAD(&space_info->priority_tickets); 245 space_info->clamp = 1; 246 btrfs_update_space_info_chunk_size(space_info, calc_chunk_size(info, flags)); 247 248 if (btrfs_is_zoned(info)) 249 space_info->bg_reclaim_threshold = BTRFS_DEFAULT_ZONED_RECLAIM_THRESH; 250 251 ret = btrfs_sysfs_add_space_info_type(info, space_info); 252 if (ret) 253 return ret; 254 255 list_add(&space_info->list, &info->space_info); 256 if (flags & BTRFS_BLOCK_GROUP_DATA) 257 info->data_sinfo = space_info; 258 259 return ret; 260 } 261 262 int btrfs_init_space_info(struct btrfs_fs_info *fs_info) 263 { 264 struct btrfs_super_block *disk_super; 265 u64 features; 266 u64 flags; 267 int mixed = 0; 268 int ret; 269 270 disk_super = fs_info->super_copy; 271 if (!btrfs_super_root(disk_super)) 272 return -EINVAL; 273 274 features = btrfs_super_incompat_flags(disk_super); 275 if (features & BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS) 276 mixed = 1; 277 278 flags = BTRFS_BLOCK_GROUP_SYSTEM; 279 ret = create_space_info(fs_info, flags); 280 if (ret) 281 goto out; 282 283 if (mixed) { 284 flags = BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_DATA; 285 ret = create_space_info(fs_info, flags); 286 } else { 287 flags = BTRFS_BLOCK_GROUP_METADATA; 288 ret = create_space_info(fs_info, flags); 289 if (ret) 290 goto out; 291 292 flags = BTRFS_BLOCK_GROUP_DATA; 293 ret = create_space_info(fs_info, flags); 294 } 295 out: 296 return ret; 297 } 298 299 void btrfs_add_bg_to_space_info(struct btrfs_fs_info *info, 300 struct btrfs_block_group *block_group) 301 { 302 struct btrfs_space_info *found; 303 int factor, index; 304 305 factor = btrfs_bg_type_to_factor(block_group->flags); 306 307 found = btrfs_find_space_info(info, block_group->flags); 308 ASSERT(found); 309 spin_lock(&found->lock); 310 found->total_bytes += block_group->length; 311 found->disk_total += block_group->length * factor; 312 found->bytes_used += block_group->used; 313 found->disk_used += block_group->used * factor; 314 found->bytes_readonly += block_group->bytes_super; 315 found->bytes_zone_unusable += block_group->zone_unusable; 316 if (block_group->length > 0) 317 found->full = 0; 318 btrfs_try_granting_tickets(info, found); 319 spin_unlock(&found->lock); 320 321 block_group->space_info = found; 322 323 index = btrfs_bg_flags_to_raid_index(block_group->flags); 324 down_write(&found->groups_sem); 325 list_add_tail(&block_group->list, &found->block_groups[index]); 326 up_write(&found->groups_sem); 327 } 328 329 struct btrfs_space_info *btrfs_find_space_info(struct btrfs_fs_info *info, 330 u64 flags) 331 { 332 struct list_head *head = &info->space_info; 333 struct btrfs_space_info *found; 334 335 flags &= BTRFS_BLOCK_GROUP_TYPE_MASK; 336 337 list_for_each_entry(found, head, list) { 338 if (found->flags & flags) 339 return found; 340 } 341 return NULL; 342 } 343 344 static u64 calc_available_free_space(struct btrfs_fs_info *fs_info, 345 struct btrfs_space_info *space_info, 346 enum btrfs_reserve_flush_enum flush) 347 { 348 u64 profile; 349 u64 avail; 350 int factor; 351 352 if (space_info->flags & BTRFS_BLOCK_GROUP_SYSTEM) 353 profile = btrfs_system_alloc_profile(fs_info); 354 else 355 profile = btrfs_metadata_alloc_profile(fs_info); 356 357 avail = atomic64_read(&fs_info->free_chunk_space); 358 359 /* 360 * If we have dup, raid1 or raid10 then only half of the free 361 * space is actually usable. For raid56, the space info used 362 * doesn't include the parity drive, so we don't have to 363 * change the math 364 */ 365 factor = btrfs_bg_type_to_factor(profile); 366 avail = div_u64(avail, factor); 367 368 /* 369 * If we aren't flushing all things, let us overcommit up to 370 * 1/2th of the space. If we can flush, don't let us overcommit 371 * too much, let it overcommit up to 1/8 of the space. 372 */ 373 if (flush == BTRFS_RESERVE_FLUSH_ALL) 374 avail >>= 3; 375 else 376 avail >>= 1; 377 return avail; 378 } 379 380 int btrfs_can_overcommit(struct btrfs_fs_info *fs_info, 381 struct btrfs_space_info *space_info, u64 bytes, 382 enum btrfs_reserve_flush_enum flush) 383 { 384 u64 avail; 385 u64 used; 386 387 /* Don't overcommit when in mixed mode */ 388 if (space_info->flags & BTRFS_BLOCK_GROUP_DATA) 389 return 0; 390 391 used = btrfs_space_info_used(space_info, true); 392 avail = calc_available_free_space(fs_info, space_info, flush); 393 394 if (used + bytes < space_info->total_bytes + avail) 395 return 1; 396 return 0; 397 } 398 399 static void remove_ticket(struct btrfs_space_info *space_info, 400 struct reserve_ticket *ticket) 401 { 402 if (!list_empty(&ticket->list)) { 403 list_del_init(&ticket->list); 404 ASSERT(space_info->reclaim_size >= ticket->bytes); 405 space_info->reclaim_size -= ticket->bytes; 406 } 407 } 408 409 /* 410 * This is for space we already have accounted in space_info->bytes_may_use, so 411 * basically when we're returning space from block_rsv's. 412 */ 413 void btrfs_try_granting_tickets(struct btrfs_fs_info *fs_info, 414 struct btrfs_space_info *space_info) 415 { 416 struct list_head *head; 417 enum btrfs_reserve_flush_enum flush = BTRFS_RESERVE_NO_FLUSH; 418 419 lockdep_assert_held(&space_info->lock); 420 421 head = &space_info->priority_tickets; 422 again: 423 while (!list_empty(head)) { 424 struct reserve_ticket *ticket; 425 u64 used = btrfs_space_info_used(space_info, true); 426 427 ticket = list_first_entry(head, struct reserve_ticket, list); 428 429 /* Check and see if our ticket can be satisfied now. */ 430 if ((used + ticket->bytes <= space_info->total_bytes) || 431 btrfs_can_overcommit(fs_info, space_info, ticket->bytes, 432 flush)) { 433 btrfs_space_info_update_bytes_may_use(fs_info, 434 space_info, 435 ticket->bytes); 436 remove_ticket(space_info, ticket); 437 ticket->bytes = 0; 438 space_info->tickets_id++; 439 wake_up(&ticket->wait); 440 } else { 441 break; 442 } 443 } 444 445 if (head == &space_info->priority_tickets) { 446 head = &space_info->tickets; 447 flush = BTRFS_RESERVE_FLUSH_ALL; 448 goto again; 449 } 450 } 451 452 #define DUMP_BLOCK_RSV(fs_info, rsv_name) \ 453 do { \ 454 struct btrfs_block_rsv *__rsv = &(fs_info)->rsv_name; \ 455 spin_lock(&__rsv->lock); \ 456 btrfs_info(fs_info, #rsv_name ": size %llu reserved %llu", \ 457 __rsv->size, __rsv->reserved); \ 458 spin_unlock(&__rsv->lock); \ 459 } while (0) 460 461 static const char *space_info_flag_to_str(const struct btrfs_space_info *space_info) 462 { 463 switch (space_info->flags) { 464 case BTRFS_BLOCK_GROUP_SYSTEM: 465 return "SYSTEM"; 466 case BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_DATA: 467 return "DATA+METADATA"; 468 case BTRFS_BLOCK_GROUP_DATA: 469 return "DATA"; 470 case BTRFS_BLOCK_GROUP_METADATA: 471 return "METADATA"; 472 default: 473 return "UNKNOWN"; 474 } 475 } 476 477 static void dump_global_block_rsv(struct btrfs_fs_info *fs_info) 478 { 479 DUMP_BLOCK_RSV(fs_info, global_block_rsv); 480 DUMP_BLOCK_RSV(fs_info, trans_block_rsv); 481 DUMP_BLOCK_RSV(fs_info, chunk_block_rsv); 482 DUMP_BLOCK_RSV(fs_info, delayed_block_rsv); 483 DUMP_BLOCK_RSV(fs_info, delayed_refs_rsv); 484 } 485 486 static void __btrfs_dump_space_info(struct btrfs_fs_info *fs_info, 487 struct btrfs_space_info *info) 488 { 489 const char *flag_str = space_info_flag_to_str(info); 490 lockdep_assert_held(&info->lock); 491 492 /* The free space could be negative in case of overcommit */ 493 btrfs_info(fs_info, "space_info %s has %lld free, is %sfull", 494 flag_str, 495 (s64)(info->total_bytes - btrfs_space_info_used(info, true)), 496 info->full ? "" : "not "); 497 btrfs_info(fs_info, 498 "space_info total=%llu, used=%llu, pinned=%llu, reserved=%llu, may_use=%llu, readonly=%llu zone_unusable=%llu", 499 info->total_bytes, info->bytes_used, info->bytes_pinned, 500 info->bytes_reserved, info->bytes_may_use, 501 info->bytes_readonly, info->bytes_zone_unusable); 502 } 503 504 void btrfs_dump_space_info(struct btrfs_fs_info *fs_info, 505 struct btrfs_space_info *info, u64 bytes, 506 int dump_block_groups) 507 { 508 struct btrfs_block_group *cache; 509 u64 total_avail = 0; 510 int index = 0; 511 512 spin_lock(&info->lock); 513 __btrfs_dump_space_info(fs_info, info); 514 dump_global_block_rsv(fs_info); 515 spin_unlock(&info->lock); 516 517 if (!dump_block_groups) 518 return; 519 520 down_read(&info->groups_sem); 521 again: 522 list_for_each_entry(cache, &info->block_groups[index], list) { 523 u64 avail; 524 525 spin_lock(&cache->lock); 526 avail = cache->length - cache->used - cache->pinned - 527 cache->reserved - cache->delalloc_bytes - 528 cache->bytes_super - cache->zone_unusable; 529 btrfs_info(fs_info, 530 "block group %llu has %llu bytes, %llu used %llu pinned %llu reserved %llu delalloc %llu super %llu zone_unusable (%llu bytes available) %s", 531 cache->start, cache->length, cache->used, cache->pinned, 532 cache->reserved, cache->delalloc_bytes, 533 cache->bytes_super, cache->zone_unusable, 534 avail, cache->ro ? "[readonly]" : ""); 535 spin_unlock(&cache->lock); 536 btrfs_dump_free_space(cache, bytes); 537 total_avail += avail; 538 } 539 if (++index < BTRFS_NR_RAID_TYPES) 540 goto again; 541 up_read(&info->groups_sem); 542 543 btrfs_info(fs_info, "%llu bytes available across all block groups", total_avail); 544 } 545 546 static inline u64 calc_reclaim_items_nr(const struct btrfs_fs_info *fs_info, 547 u64 to_reclaim) 548 { 549 u64 bytes; 550 u64 nr; 551 552 bytes = btrfs_calc_insert_metadata_size(fs_info, 1); 553 nr = div64_u64(to_reclaim, bytes); 554 if (!nr) 555 nr = 1; 556 return nr; 557 } 558 559 static inline u64 calc_delayed_refs_nr(const struct btrfs_fs_info *fs_info, 560 u64 to_reclaim) 561 { 562 const u64 bytes = btrfs_calc_delayed_ref_bytes(fs_info, 1); 563 u64 nr; 564 565 nr = div64_u64(to_reclaim, bytes); 566 if (!nr) 567 nr = 1; 568 return nr; 569 } 570 571 #define EXTENT_SIZE_PER_ITEM SZ_256K 572 573 /* 574 * shrink metadata reservation for delalloc 575 */ 576 static void shrink_delalloc(struct btrfs_fs_info *fs_info, 577 struct btrfs_space_info *space_info, 578 u64 to_reclaim, bool wait_ordered, 579 bool for_preempt) 580 { 581 struct btrfs_trans_handle *trans; 582 u64 delalloc_bytes; 583 u64 ordered_bytes; 584 u64 items; 585 long time_left; 586 int loops; 587 588 delalloc_bytes = percpu_counter_sum_positive(&fs_info->delalloc_bytes); 589 ordered_bytes = percpu_counter_sum_positive(&fs_info->ordered_bytes); 590 if (delalloc_bytes == 0 && ordered_bytes == 0) 591 return; 592 593 /* Calc the number of the pages we need flush for space reservation */ 594 if (to_reclaim == U64_MAX) { 595 items = U64_MAX; 596 } else { 597 /* 598 * to_reclaim is set to however much metadata we need to 599 * reclaim, but reclaiming that much data doesn't really track 600 * exactly. What we really want to do is reclaim full inode's 601 * worth of reservations, however that's not available to us 602 * here. We will take a fraction of the delalloc bytes for our 603 * flushing loops and hope for the best. Delalloc will expand 604 * the amount we write to cover an entire dirty extent, which 605 * will reclaim the metadata reservation for that range. If 606 * it's not enough subsequent flush stages will be more 607 * aggressive. 608 */ 609 to_reclaim = max(to_reclaim, delalloc_bytes >> 3); 610 items = calc_reclaim_items_nr(fs_info, to_reclaim) * 2; 611 } 612 613 trans = current->journal_info; 614 615 /* 616 * If we are doing more ordered than delalloc we need to just wait on 617 * ordered extents, otherwise we'll waste time trying to flush delalloc 618 * that likely won't give us the space back we need. 619 */ 620 if (ordered_bytes > delalloc_bytes && !for_preempt) 621 wait_ordered = true; 622 623 loops = 0; 624 while ((delalloc_bytes || ordered_bytes) && loops < 3) { 625 u64 temp = min(delalloc_bytes, to_reclaim) >> PAGE_SHIFT; 626 long nr_pages = min_t(u64, temp, LONG_MAX); 627 int async_pages; 628 629 btrfs_start_delalloc_roots(fs_info, nr_pages, true); 630 631 /* 632 * We need to make sure any outstanding async pages are now 633 * processed before we continue. This is because things like 634 * sync_inode() try to be smart and skip writing if the inode is 635 * marked clean. We don't use filemap_fwrite for flushing 636 * because we want to control how many pages we write out at a 637 * time, thus this is the only safe way to make sure we've 638 * waited for outstanding compressed workers to have started 639 * their jobs and thus have ordered extents set up properly. 640 * 641 * This exists because we do not want to wait for each 642 * individual inode to finish its async work, we simply want to 643 * start the IO on everybody, and then come back here and wait 644 * for all of the async work to catch up. Once we're done with 645 * that we know we'll have ordered extents for everything and we 646 * can decide if we wait for that or not. 647 * 648 * If we choose to replace this in the future, make absolutely 649 * sure that the proper waiting is being done in the async case, 650 * as there have been bugs in that area before. 651 */ 652 async_pages = atomic_read(&fs_info->async_delalloc_pages); 653 if (!async_pages) 654 goto skip_async; 655 656 /* 657 * We don't want to wait forever, if we wrote less pages in this 658 * loop than we have outstanding, only wait for that number of 659 * pages, otherwise we can wait for all async pages to finish 660 * before continuing. 661 */ 662 if (async_pages > nr_pages) 663 async_pages -= nr_pages; 664 else 665 async_pages = 0; 666 wait_event(fs_info->async_submit_wait, 667 atomic_read(&fs_info->async_delalloc_pages) <= 668 async_pages); 669 skip_async: 670 loops++; 671 if (wait_ordered && !trans) { 672 btrfs_wait_ordered_roots(fs_info, items, 0, (u64)-1); 673 } else { 674 time_left = schedule_timeout_killable(1); 675 if (time_left) 676 break; 677 } 678 679 /* 680 * If we are for preemption we just want a one-shot of delalloc 681 * flushing so we can stop flushing if we decide we don't need 682 * to anymore. 683 */ 684 if (for_preempt) 685 break; 686 687 spin_lock(&space_info->lock); 688 if (list_empty(&space_info->tickets) && 689 list_empty(&space_info->priority_tickets)) { 690 spin_unlock(&space_info->lock); 691 break; 692 } 693 spin_unlock(&space_info->lock); 694 695 delalloc_bytes = percpu_counter_sum_positive( 696 &fs_info->delalloc_bytes); 697 ordered_bytes = percpu_counter_sum_positive( 698 &fs_info->ordered_bytes); 699 } 700 } 701 702 /* 703 * Try to flush some data based on policy set by @state. This is only advisory 704 * and may fail for various reasons. The caller is supposed to examine the 705 * state of @space_info to detect the outcome. 706 */ 707 static void flush_space(struct btrfs_fs_info *fs_info, 708 struct btrfs_space_info *space_info, u64 num_bytes, 709 enum btrfs_flush_state state, bool for_preempt) 710 { 711 struct btrfs_root *root = fs_info->tree_root; 712 struct btrfs_trans_handle *trans; 713 int nr; 714 int ret = 0; 715 716 switch (state) { 717 case FLUSH_DELAYED_ITEMS_NR: 718 case FLUSH_DELAYED_ITEMS: 719 if (state == FLUSH_DELAYED_ITEMS_NR) 720 nr = calc_reclaim_items_nr(fs_info, num_bytes) * 2; 721 else 722 nr = -1; 723 724 trans = btrfs_join_transaction_nostart(root); 725 if (IS_ERR(trans)) { 726 ret = PTR_ERR(trans); 727 if (ret == -ENOENT) 728 ret = 0; 729 break; 730 } 731 ret = btrfs_run_delayed_items_nr(trans, nr); 732 btrfs_end_transaction(trans); 733 break; 734 case FLUSH_DELALLOC: 735 case FLUSH_DELALLOC_WAIT: 736 case FLUSH_DELALLOC_FULL: 737 if (state == FLUSH_DELALLOC_FULL) 738 num_bytes = U64_MAX; 739 shrink_delalloc(fs_info, space_info, num_bytes, 740 state != FLUSH_DELALLOC, for_preempt); 741 break; 742 case FLUSH_DELAYED_REFS_NR: 743 case FLUSH_DELAYED_REFS: 744 trans = btrfs_join_transaction_nostart(root); 745 if (IS_ERR(trans)) { 746 ret = PTR_ERR(trans); 747 if (ret == -ENOENT) 748 ret = 0; 749 break; 750 } 751 if (state == FLUSH_DELAYED_REFS_NR) 752 nr = calc_delayed_refs_nr(fs_info, num_bytes); 753 else 754 nr = 0; 755 btrfs_run_delayed_refs(trans, nr); 756 btrfs_end_transaction(trans); 757 break; 758 case ALLOC_CHUNK: 759 case ALLOC_CHUNK_FORCE: 760 trans = btrfs_join_transaction(root); 761 if (IS_ERR(trans)) { 762 ret = PTR_ERR(trans); 763 break; 764 } 765 ret = btrfs_chunk_alloc(trans, 766 btrfs_get_alloc_profile(fs_info, space_info->flags), 767 (state == ALLOC_CHUNK) ? CHUNK_ALLOC_NO_FORCE : 768 CHUNK_ALLOC_FORCE); 769 btrfs_end_transaction(trans); 770 771 if (ret > 0 || ret == -ENOSPC) 772 ret = 0; 773 break; 774 case RUN_DELAYED_IPUTS: 775 /* 776 * If we have pending delayed iputs then we could free up a 777 * bunch of pinned space, so make sure we run the iputs before 778 * we do our pinned bytes check below. 779 */ 780 btrfs_run_delayed_iputs(fs_info); 781 btrfs_wait_on_delayed_iputs(fs_info); 782 break; 783 case COMMIT_TRANS: 784 ASSERT(current->journal_info == NULL); 785 /* 786 * We don't want to start a new transaction, just attach to the 787 * current one or wait it fully commits in case its commit is 788 * happening at the moment. Note: we don't use a nostart join 789 * because that does not wait for a transaction to fully commit 790 * (only for it to be unblocked, state TRANS_STATE_UNBLOCKED). 791 */ 792 trans = btrfs_attach_transaction_barrier(root); 793 if (IS_ERR(trans)) { 794 ret = PTR_ERR(trans); 795 if (ret == -ENOENT) 796 ret = 0; 797 break; 798 } 799 ret = btrfs_commit_transaction(trans); 800 break; 801 default: 802 ret = -ENOSPC; 803 break; 804 } 805 806 trace_btrfs_flush_space(fs_info, space_info->flags, num_bytes, state, 807 ret, for_preempt); 808 return; 809 } 810 811 static inline u64 812 btrfs_calc_reclaim_metadata_size(struct btrfs_fs_info *fs_info, 813 struct btrfs_space_info *space_info) 814 { 815 u64 used; 816 u64 avail; 817 u64 to_reclaim = space_info->reclaim_size; 818 819 lockdep_assert_held(&space_info->lock); 820 821 avail = calc_available_free_space(fs_info, space_info, 822 BTRFS_RESERVE_FLUSH_ALL); 823 used = btrfs_space_info_used(space_info, true); 824 825 /* 826 * We may be flushing because suddenly we have less space than we had 827 * before, and now we're well over-committed based on our current free 828 * space. If that's the case add in our overage so we make sure to put 829 * appropriate pressure on the flushing state machine. 830 */ 831 if (space_info->total_bytes + avail < used) 832 to_reclaim += used - (space_info->total_bytes + avail); 833 834 return to_reclaim; 835 } 836 837 static bool need_preemptive_reclaim(struct btrfs_fs_info *fs_info, 838 struct btrfs_space_info *space_info) 839 { 840 u64 global_rsv_size = fs_info->global_block_rsv.reserved; 841 u64 ordered, delalloc; 842 u64 thresh; 843 u64 used; 844 845 thresh = mult_perc(space_info->total_bytes, 90); 846 847 lockdep_assert_held(&space_info->lock); 848 849 /* If we're just plain full then async reclaim just slows us down. */ 850 if ((space_info->bytes_used + space_info->bytes_reserved + 851 global_rsv_size) >= thresh) 852 return false; 853 854 used = space_info->bytes_may_use + space_info->bytes_pinned; 855 856 /* The total flushable belongs to the global rsv, don't flush. */ 857 if (global_rsv_size >= used) 858 return false; 859 860 /* 861 * 128MiB is 1/4 of the maximum global rsv size. If we have less than 862 * that devoted to other reservations then there's no sense in flushing, 863 * we don't have a lot of things that need flushing. 864 */ 865 if (used - global_rsv_size <= SZ_128M) 866 return false; 867 868 /* 869 * We have tickets queued, bail so we don't compete with the async 870 * flushers. 871 */ 872 if (space_info->reclaim_size) 873 return false; 874 875 /* 876 * If we have over half of the free space occupied by reservations or 877 * pinned then we want to start flushing. 878 * 879 * We do not do the traditional thing here, which is to say 880 * 881 * if (used >= ((total_bytes + avail) / 2)) 882 * return 1; 883 * 884 * because this doesn't quite work how we want. If we had more than 50% 885 * of the space_info used by bytes_used and we had 0 available we'd just 886 * constantly run the background flusher. Instead we want it to kick in 887 * if our reclaimable space exceeds our clamped free space. 888 * 889 * Our clamping range is 2^1 -> 2^8. Practically speaking that means 890 * the following: 891 * 892 * Amount of RAM Minimum threshold Maximum threshold 893 * 894 * 256GiB 1GiB 128GiB 895 * 128GiB 512MiB 64GiB 896 * 64GiB 256MiB 32GiB 897 * 32GiB 128MiB 16GiB 898 * 16GiB 64MiB 8GiB 899 * 900 * These are the range our thresholds will fall in, corresponding to how 901 * much delalloc we need for the background flusher to kick in. 902 */ 903 904 thresh = calc_available_free_space(fs_info, space_info, 905 BTRFS_RESERVE_FLUSH_ALL); 906 used = space_info->bytes_used + space_info->bytes_reserved + 907 space_info->bytes_readonly + global_rsv_size; 908 if (used < space_info->total_bytes) 909 thresh += space_info->total_bytes - used; 910 thresh >>= space_info->clamp; 911 912 used = space_info->bytes_pinned; 913 914 /* 915 * If we have more ordered bytes than delalloc bytes then we're either 916 * doing a lot of DIO, or we simply don't have a lot of delalloc waiting 917 * around. Preemptive flushing is only useful in that it can free up 918 * space before tickets need to wait for things to finish. In the case 919 * of ordered extents, preemptively waiting on ordered extents gets us 920 * nothing, if our reservations are tied up in ordered extents we'll 921 * simply have to slow down writers by forcing them to wait on ordered 922 * extents. 923 * 924 * In the case that ordered is larger than delalloc, only include the 925 * block reserves that we would actually be able to directly reclaim 926 * from. In this case if we're heavy on metadata operations this will 927 * clearly be heavy enough to warrant preemptive flushing. In the case 928 * of heavy DIO or ordered reservations, preemptive flushing will just 929 * waste time and cause us to slow down. 930 * 931 * We want to make sure we truly are maxed out on ordered however, so 932 * cut ordered in half, and if it's still higher than delalloc then we 933 * can keep flushing. This is to avoid the case where we start 934 * flushing, and now delalloc == ordered and we stop preemptively 935 * flushing when we could still have several gigs of delalloc to flush. 936 */ 937 ordered = percpu_counter_read_positive(&fs_info->ordered_bytes) >> 1; 938 delalloc = percpu_counter_read_positive(&fs_info->delalloc_bytes); 939 if (ordered >= delalloc) 940 used += fs_info->delayed_refs_rsv.reserved + 941 fs_info->delayed_block_rsv.reserved; 942 else 943 used += space_info->bytes_may_use - global_rsv_size; 944 945 return (used >= thresh && !btrfs_fs_closing(fs_info) && 946 !test_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state)); 947 } 948 949 static bool steal_from_global_rsv(struct btrfs_fs_info *fs_info, 950 struct btrfs_space_info *space_info, 951 struct reserve_ticket *ticket) 952 { 953 struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv; 954 u64 min_bytes; 955 956 if (!ticket->steal) 957 return false; 958 959 if (global_rsv->space_info != space_info) 960 return false; 961 962 spin_lock(&global_rsv->lock); 963 min_bytes = mult_perc(global_rsv->size, 10); 964 if (global_rsv->reserved < min_bytes + ticket->bytes) { 965 spin_unlock(&global_rsv->lock); 966 return false; 967 } 968 global_rsv->reserved -= ticket->bytes; 969 remove_ticket(space_info, ticket); 970 ticket->bytes = 0; 971 wake_up(&ticket->wait); 972 space_info->tickets_id++; 973 if (global_rsv->reserved < global_rsv->size) 974 global_rsv->full = 0; 975 spin_unlock(&global_rsv->lock); 976 977 return true; 978 } 979 980 /* 981 * maybe_fail_all_tickets - we've exhausted our flushing, start failing tickets 982 * @fs_info - fs_info for this fs 983 * @space_info - the space info we were flushing 984 * 985 * We call this when we've exhausted our flushing ability and haven't made 986 * progress in satisfying tickets. The reservation code handles tickets in 987 * order, so if there is a large ticket first and then smaller ones we could 988 * very well satisfy the smaller tickets. This will attempt to wake up any 989 * tickets in the list to catch this case. 990 * 991 * This function returns true if it was able to make progress by clearing out 992 * other tickets, or if it stumbles across a ticket that was smaller than the 993 * first ticket. 994 */ 995 static bool maybe_fail_all_tickets(struct btrfs_fs_info *fs_info, 996 struct btrfs_space_info *space_info) 997 { 998 struct reserve_ticket *ticket; 999 u64 tickets_id = space_info->tickets_id; 1000 const bool aborted = BTRFS_FS_ERROR(fs_info); 1001 1002 trace_btrfs_fail_all_tickets(fs_info, space_info); 1003 1004 if (btrfs_test_opt(fs_info, ENOSPC_DEBUG)) { 1005 btrfs_info(fs_info, "cannot satisfy tickets, dumping space info"); 1006 __btrfs_dump_space_info(fs_info, space_info); 1007 } 1008 1009 while (!list_empty(&space_info->tickets) && 1010 tickets_id == space_info->tickets_id) { 1011 ticket = list_first_entry(&space_info->tickets, 1012 struct reserve_ticket, list); 1013 1014 if (!aborted && steal_from_global_rsv(fs_info, space_info, ticket)) 1015 return true; 1016 1017 if (!aborted && btrfs_test_opt(fs_info, ENOSPC_DEBUG)) 1018 btrfs_info(fs_info, "failing ticket with %llu bytes", 1019 ticket->bytes); 1020 1021 remove_ticket(space_info, ticket); 1022 if (aborted) 1023 ticket->error = -EIO; 1024 else 1025 ticket->error = -ENOSPC; 1026 wake_up(&ticket->wait); 1027 1028 /* 1029 * We're just throwing tickets away, so more flushing may not 1030 * trip over btrfs_try_granting_tickets, so we need to call it 1031 * here to see if we can make progress with the next ticket in 1032 * the list. 1033 */ 1034 if (!aborted) 1035 btrfs_try_granting_tickets(fs_info, space_info); 1036 } 1037 return (tickets_id != space_info->tickets_id); 1038 } 1039 1040 /* 1041 * This is for normal flushers, we can wait all goddamned day if we want to. We 1042 * will loop and continuously try to flush as long as we are making progress. 1043 * We count progress as clearing off tickets each time we have to loop. 1044 */ 1045 static void btrfs_async_reclaim_metadata_space(struct work_struct *work) 1046 { 1047 struct btrfs_fs_info *fs_info; 1048 struct btrfs_space_info *space_info; 1049 u64 to_reclaim; 1050 enum btrfs_flush_state flush_state; 1051 int commit_cycles = 0; 1052 u64 last_tickets_id; 1053 1054 fs_info = container_of(work, struct btrfs_fs_info, async_reclaim_work); 1055 space_info = btrfs_find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); 1056 1057 spin_lock(&space_info->lock); 1058 to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info, space_info); 1059 if (!to_reclaim) { 1060 space_info->flush = 0; 1061 spin_unlock(&space_info->lock); 1062 return; 1063 } 1064 last_tickets_id = space_info->tickets_id; 1065 spin_unlock(&space_info->lock); 1066 1067 flush_state = FLUSH_DELAYED_ITEMS_NR; 1068 do { 1069 flush_space(fs_info, space_info, to_reclaim, flush_state, false); 1070 spin_lock(&space_info->lock); 1071 if (list_empty(&space_info->tickets)) { 1072 space_info->flush = 0; 1073 spin_unlock(&space_info->lock); 1074 return; 1075 } 1076 to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info, 1077 space_info); 1078 if (last_tickets_id == space_info->tickets_id) { 1079 flush_state++; 1080 } else { 1081 last_tickets_id = space_info->tickets_id; 1082 flush_state = FLUSH_DELAYED_ITEMS_NR; 1083 if (commit_cycles) 1084 commit_cycles--; 1085 } 1086 1087 /* 1088 * We do not want to empty the system of delalloc unless we're 1089 * under heavy pressure, so allow one trip through the flushing 1090 * logic before we start doing a FLUSH_DELALLOC_FULL. 1091 */ 1092 if (flush_state == FLUSH_DELALLOC_FULL && !commit_cycles) 1093 flush_state++; 1094 1095 /* 1096 * We don't want to force a chunk allocation until we've tried 1097 * pretty hard to reclaim space. Think of the case where we 1098 * freed up a bunch of space and so have a lot of pinned space 1099 * to reclaim. We would rather use that than possibly create a 1100 * underutilized metadata chunk. So if this is our first run 1101 * through the flushing state machine skip ALLOC_CHUNK_FORCE and 1102 * commit the transaction. If nothing has changed the next go 1103 * around then we can force a chunk allocation. 1104 */ 1105 if (flush_state == ALLOC_CHUNK_FORCE && !commit_cycles) 1106 flush_state++; 1107 1108 if (flush_state > COMMIT_TRANS) { 1109 commit_cycles++; 1110 if (commit_cycles > 2) { 1111 if (maybe_fail_all_tickets(fs_info, space_info)) { 1112 flush_state = FLUSH_DELAYED_ITEMS_NR; 1113 commit_cycles--; 1114 } else { 1115 space_info->flush = 0; 1116 } 1117 } else { 1118 flush_state = FLUSH_DELAYED_ITEMS_NR; 1119 } 1120 } 1121 spin_unlock(&space_info->lock); 1122 } while (flush_state <= COMMIT_TRANS); 1123 } 1124 1125 /* 1126 * This handles pre-flushing of metadata space before we get to the point that 1127 * we need to start blocking threads on tickets. The logic here is different 1128 * from the other flush paths because it doesn't rely on tickets to tell us how 1129 * much we need to flush, instead it attempts to keep us below the 80% full 1130 * watermark of space by flushing whichever reservation pool is currently the 1131 * largest. 1132 */ 1133 static void btrfs_preempt_reclaim_metadata_space(struct work_struct *work) 1134 { 1135 struct btrfs_fs_info *fs_info; 1136 struct btrfs_space_info *space_info; 1137 struct btrfs_block_rsv *delayed_block_rsv; 1138 struct btrfs_block_rsv *delayed_refs_rsv; 1139 struct btrfs_block_rsv *global_rsv; 1140 struct btrfs_block_rsv *trans_rsv; 1141 int loops = 0; 1142 1143 fs_info = container_of(work, struct btrfs_fs_info, 1144 preempt_reclaim_work); 1145 space_info = btrfs_find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); 1146 delayed_block_rsv = &fs_info->delayed_block_rsv; 1147 delayed_refs_rsv = &fs_info->delayed_refs_rsv; 1148 global_rsv = &fs_info->global_block_rsv; 1149 trans_rsv = &fs_info->trans_block_rsv; 1150 1151 spin_lock(&space_info->lock); 1152 while (need_preemptive_reclaim(fs_info, space_info)) { 1153 enum btrfs_flush_state flush; 1154 u64 delalloc_size = 0; 1155 u64 to_reclaim, block_rsv_size; 1156 u64 global_rsv_size = global_rsv->reserved; 1157 1158 loops++; 1159 1160 /* 1161 * We don't have a precise counter for the metadata being 1162 * reserved for delalloc, so we'll approximate it by subtracting 1163 * out the block rsv's space from the bytes_may_use. If that 1164 * amount is higher than the individual reserves, then we can 1165 * assume it's tied up in delalloc reservations. 1166 */ 1167 block_rsv_size = global_rsv_size + 1168 delayed_block_rsv->reserved + 1169 delayed_refs_rsv->reserved + 1170 trans_rsv->reserved; 1171 if (block_rsv_size < space_info->bytes_may_use) 1172 delalloc_size = space_info->bytes_may_use - block_rsv_size; 1173 1174 /* 1175 * We don't want to include the global_rsv in our calculation, 1176 * because that's space we can't touch. Subtract it from the 1177 * block_rsv_size for the next checks. 1178 */ 1179 block_rsv_size -= global_rsv_size; 1180 1181 /* 1182 * We really want to avoid flushing delalloc too much, as it 1183 * could result in poor allocation patterns, so only flush it if 1184 * it's larger than the rest of the pools combined. 1185 */ 1186 if (delalloc_size > block_rsv_size) { 1187 to_reclaim = delalloc_size; 1188 flush = FLUSH_DELALLOC; 1189 } else if (space_info->bytes_pinned > 1190 (delayed_block_rsv->reserved + 1191 delayed_refs_rsv->reserved)) { 1192 to_reclaim = space_info->bytes_pinned; 1193 flush = COMMIT_TRANS; 1194 } else if (delayed_block_rsv->reserved > 1195 delayed_refs_rsv->reserved) { 1196 to_reclaim = delayed_block_rsv->reserved; 1197 flush = FLUSH_DELAYED_ITEMS_NR; 1198 } else { 1199 to_reclaim = delayed_refs_rsv->reserved; 1200 flush = FLUSH_DELAYED_REFS_NR; 1201 } 1202 1203 spin_unlock(&space_info->lock); 1204 1205 /* 1206 * We don't want to reclaim everything, just a portion, so scale 1207 * down the to_reclaim by 1/4. If it takes us down to 0, 1208 * reclaim 1 items worth. 1209 */ 1210 to_reclaim >>= 2; 1211 if (!to_reclaim) 1212 to_reclaim = btrfs_calc_insert_metadata_size(fs_info, 1); 1213 flush_space(fs_info, space_info, to_reclaim, flush, true); 1214 cond_resched(); 1215 spin_lock(&space_info->lock); 1216 } 1217 1218 /* We only went through once, back off our clamping. */ 1219 if (loops == 1 && !space_info->reclaim_size) 1220 space_info->clamp = max(1, space_info->clamp - 1); 1221 trace_btrfs_done_preemptive_reclaim(fs_info, space_info); 1222 spin_unlock(&space_info->lock); 1223 } 1224 1225 /* 1226 * FLUSH_DELALLOC_WAIT: 1227 * Space is freed from flushing delalloc in one of two ways. 1228 * 1229 * 1) compression is on and we allocate less space than we reserved 1230 * 2) we are overwriting existing space 1231 * 1232 * For #1 that extra space is reclaimed as soon as the delalloc pages are 1233 * COWed, by way of btrfs_add_reserved_bytes() which adds the actual extent 1234 * length to ->bytes_reserved, and subtracts the reserved space from 1235 * ->bytes_may_use. 1236 * 1237 * For #2 this is trickier. Once the ordered extent runs we will drop the 1238 * extent in the range we are overwriting, which creates a delayed ref for 1239 * that freed extent. This however is not reclaimed until the transaction 1240 * commits, thus the next stages. 1241 * 1242 * RUN_DELAYED_IPUTS 1243 * If we are freeing inodes, we want to make sure all delayed iputs have 1244 * completed, because they could have been on an inode with i_nlink == 0, and 1245 * thus have been truncated and freed up space. But again this space is not 1246 * immediately re-usable, it comes in the form of a delayed ref, which must be 1247 * run and then the transaction must be committed. 1248 * 1249 * COMMIT_TRANS 1250 * This is where we reclaim all of the pinned space generated by running the 1251 * iputs 1252 * 1253 * ALLOC_CHUNK_FORCE 1254 * For data we start with alloc chunk force, however we could have been full 1255 * before, and then the transaction commit could have freed new block groups, 1256 * so if we now have space to allocate do the force chunk allocation. 1257 */ 1258 static const enum btrfs_flush_state data_flush_states[] = { 1259 FLUSH_DELALLOC_FULL, 1260 RUN_DELAYED_IPUTS, 1261 COMMIT_TRANS, 1262 ALLOC_CHUNK_FORCE, 1263 }; 1264 1265 static void btrfs_async_reclaim_data_space(struct work_struct *work) 1266 { 1267 struct btrfs_fs_info *fs_info; 1268 struct btrfs_space_info *space_info; 1269 u64 last_tickets_id; 1270 enum btrfs_flush_state flush_state = 0; 1271 1272 fs_info = container_of(work, struct btrfs_fs_info, async_data_reclaim_work); 1273 space_info = fs_info->data_sinfo; 1274 1275 spin_lock(&space_info->lock); 1276 if (list_empty(&space_info->tickets)) { 1277 space_info->flush = 0; 1278 spin_unlock(&space_info->lock); 1279 return; 1280 } 1281 last_tickets_id = space_info->tickets_id; 1282 spin_unlock(&space_info->lock); 1283 1284 while (!space_info->full) { 1285 flush_space(fs_info, space_info, U64_MAX, ALLOC_CHUNK_FORCE, false); 1286 spin_lock(&space_info->lock); 1287 if (list_empty(&space_info->tickets)) { 1288 space_info->flush = 0; 1289 spin_unlock(&space_info->lock); 1290 return; 1291 } 1292 1293 /* Something happened, fail everything and bail. */ 1294 if (BTRFS_FS_ERROR(fs_info)) 1295 goto aborted_fs; 1296 last_tickets_id = space_info->tickets_id; 1297 spin_unlock(&space_info->lock); 1298 } 1299 1300 while (flush_state < ARRAY_SIZE(data_flush_states)) { 1301 flush_space(fs_info, space_info, U64_MAX, 1302 data_flush_states[flush_state], false); 1303 spin_lock(&space_info->lock); 1304 if (list_empty(&space_info->tickets)) { 1305 space_info->flush = 0; 1306 spin_unlock(&space_info->lock); 1307 return; 1308 } 1309 1310 if (last_tickets_id == space_info->tickets_id) { 1311 flush_state++; 1312 } else { 1313 last_tickets_id = space_info->tickets_id; 1314 flush_state = 0; 1315 } 1316 1317 if (flush_state >= ARRAY_SIZE(data_flush_states)) { 1318 if (space_info->full) { 1319 if (maybe_fail_all_tickets(fs_info, space_info)) 1320 flush_state = 0; 1321 else 1322 space_info->flush = 0; 1323 } else { 1324 flush_state = 0; 1325 } 1326 1327 /* Something happened, fail everything and bail. */ 1328 if (BTRFS_FS_ERROR(fs_info)) 1329 goto aborted_fs; 1330 1331 } 1332 spin_unlock(&space_info->lock); 1333 } 1334 return; 1335 1336 aborted_fs: 1337 maybe_fail_all_tickets(fs_info, space_info); 1338 space_info->flush = 0; 1339 spin_unlock(&space_info->lock); 1340 } 1341 1342 void btrfs_init_async_reclaim_work(struct btrfs_fs_info *fs_info) 1343 { 1344 INIT_WORK(&fs_info->async_reclaim_work, btrfs_async_reclaim_metadata_space); 1345 INIT_WORK(&fs_info->async_data_reclaim_work, btrfs_async_reclaim_data_space); 1346 INIT_WORK(&fs_info->preempt_reclaim_work, 1347 btrfs_preempt_reclaim_metadata_space); 1348 } 1349 1350 static const enum btrfs_flush_state priority_flush_states[] = { 1351 FLUSH_DELAYED_ITEMS_NR, 1352 FLUSH_DELAYED_ITEMS, 1353 ALLOC_CHUNK, 1354 }; 1355 1356 static const enum btrfs_flush_state evict_flush_states[] = { 1357 FLUSH_DELAYED_ITEMS_NR, 1358 FLUSH_DELAYED_ITEMS, 1359 FLUSH_DELAYED_REFS_NR, 1360 FLUSH_DELAYED_REFS, 1361 FLUSH_DELALLOC, 1362 FLUSH_DELALLOC_WAIT, 1363 FLUSH_DELALLOC_FULL, 1364 ALLOC_CHUNK, 1365 COMMIT_TRANS, 1366 }; 1367 1368 static void priority_reclaim_metadata_space(struct btrfs_fs_info *fs_info, 1369 struct btrfs_space_info *space_info, 1370 struct reserve_ticket *ticket, 1371 const enum btrfs_flush_state *states, 1372 int states_nr) 1373 { 1374 u64 to_reclaim; 1375 int flush_state = 0; 1376 1377 spin_lock(&space_info->lock); 1378 to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info, space_info); 1379 /* 1380 * This is the priority reclaim path, so to_reclaim could be >0 still 1381 * because we may have only satisfied the priority tickets and still 1382 * left non priority tickets on the list. We would then have 1383 * to_reclaim but ->bytes == 0. 1384 */ 1385 if (ticket->bytes == 0) { 1386 spin_unlock(&space_info->lock); 1387 return; 1388 } 1389 1390 while (flush_state < states_nr) { 1391 spin_unlock(&space_info->lock); 1392 flush_space(fs_info, space_info, to_reclaim, states[flush_state], 1393 false); 1394 flush_state++; 1395 spin_lock(&space_info->lock); 1396 if (ticket->bytes == 0) { 1397 spin_unlock(&space_info->lock); 1398 return; 1399 } 1400 } 1401 1402 /* 1403 * Attempt to steal from the global rsv if we can, except if the fs was 1404 * turned into error mode due to a transaction abort when flushing space 1405 * above, in that case fail with the abort error instead of returning 1406 * success to the caller if we can steal from the global rsv - this is 1407 * just to have caller fail immeditelly instead of later when trying to 1408 * modify the fs, making it easier to debug -ENOSPC problems. 1409 */ 1410 if (BTRFS_FS_ERROR(fs_info)) { 1411 ticket->error = BTRFS_FS_ERROR(fs_info); 1412 remove_ticket(space_info, ticket); 1413 } else if (!steal_from_global_rsv(fs_info, space_info, ticket)) { 1414 ticket->error = -ENOSPC; 1415 remove_ticket(space_info, ticket); 1416 } 1417 1418 /* 1419 * We must run try_granting_tickets here because we could be a large 1420 * ticket in front of a smaller ticket that can now be satisfied with 1421 * the available space. 1422 */ 1423 btrfs_try_granting_tickets(fs_info, space_info); 1424 spin_unlock(&space_info->lock); 1425 } 1426 1427 static void priority_reclaim_data_space(struct btrfs_fs_info *fs_info, 1428 struct btrfs_space_info *space_info, 1429 struct reserve_ticket *ticket) 1430 { 1431 spin_lock(&space_info->lock); 1432 1433 /* We could have been granted before we got here. */ 1434 if (ticket->bytes == 0) { 1435 spin_unlock(&space_info->lock); 1436 return; 1437 } 1438 1439 while (!space_info->full) { 1440 spin_unlock(&space_info->lock); 1441 flush_space(fs_info, space_info, U64_MAX, ALLOC_CHUNK_FORCE, false); 1442 spin_lock(&space_info->lock); 1443 if (ticket->bytes == 0) { 1444 spin_unlock(&space_info->lock); 1445 return; 1446 } 1447 } 1448 1449 ticket->error = -ENOSPC; 1450 remove_ticket(space_info, ticket); 1451 btrfs_try_granting_tickets(fs_info, space_info); 1452 spin_unlock(&space_info->lock); 1453 } 1454 1455 static void wait_reserve_ticket(struct btrfs_fs_info *fs_info, 1456 struct btrfs_space_info *space_info, 1457 struct reserve_ticket *ticket) 1458 1459 { 1460 DEFINE_WAIT(wait); 1461 int ret = 0; 1462 1463 spin_lock(&space_info->lock); 1464 while (ticket->bytes > 0 && ticket->error == 0) { 1465 ret = prepare_to_wait_event(&ticket->wait, &wait, TASK_KILLABLE); 1466 if (ret) { 1467 /* 1468 * Delete us from the list. After we unlock the space 1469 * info, we don't want the async reclaim job to reserve 1470 * space for this ticket. If that would happen, then the 1471 * ticket's task would not known that space was reserved 1472 * despite getting an error, resulting in a space leak 1473 * (bytes_may_use counter of our space_info). 1474 */ 1475 remove_ticket(space_info, ticket); 1476 ticket->error = -EINTR; 1477 break; 1478 } 1479 spin_unlock(&space_info->lock); 1480 1481 schedule(); 1482 1483 finish_wait(&ticket->wait, &wait); 1484 spin_lock(&space_info->lock); 1485 } 1486 spin_unlock(&space_info->lock); 1487 } 1488 1489 /* 1490 * Do the appropriate flushing and waiting for a ticket. 1491 * 1492 * @fs_info: the filesystem 1493 * @space_info: space info for the reservation 1494 * @ticket: ticket for the reservation 1495 * @start_ns: timestamp when the reservation started 1496 * @orig_bytes: amount of bytes originally reserved 1497 * @flush: how much we can flush 1498 * 1499 * This does the work of figuring out how to flush for the ticket, waiting for 1500 * the reservation, and returning the appropriate error if there is one. 1501 */ 1502 static int handle_reserve_ticket(struct btrfs_fs_info *fs_info, 1503 struct btrfs_space_info *space_info, 1504 struct reserve_ticket *ticket, 1505 u64 start_ns, u64 orig_bytes, 1506 enum btrfs_reserve_flush_enum flush) 1507 { 1508 int ret; 1509 1510 switch (flush) { 1511 case BTRFS_RESERVE_FLUSH_DATA: 1512 case BTRFS_RESERVE_FLUSH_ALL: 1513 case BTRFS_RESERVE_FLUSH_ALL_STEAL: 1514 wait_reserve_ticket(fs_info, space_info, ticket); 1515 break; 1516 case BTRFS_RESERVE_FLUSH_LIMIT: 1517 priority_reclaim_metadata_space(fs_info, space_info, ticket, 1518 priority_flush_states, 1519 ARRAY_SIZE(priority_flush_states)); 1520 break; 1521 case BTRFS_RESERVE_FLUSH_EVICT: 1522 priority_reclaim_metadata_space(fs_info, space_info, ticket, 1523 evict_flush_states, 1524 ARRAY_SIZE(evict_flush_states)); 1525 break; 1526 case BTRFS_RESERVE_FLUSH_FREE_SPACE_INODE: 1527 priority_reclaim_data_space(fs_info, space_info, ticket); 1528 break; 1529 default: 1530 ASSERT(0); 1531 break; 1532 } 1533 1534 ret = ticket->error; 1535 ASSERT(list_empty(&ticket->list)); 1536 /* 1537 * Check that we can't have an error set if the reservation succeeded, 1538 * as that would confuse tasks and lead them to error out without 1539 * releasing reserved space (if an error happens the expectation is that 1540 * space wasn't reserved at all). 1541 */ 1542 ASSERT(!(ticket->bytes == 0 && ticket->error)); 1543 trace_btrfs_reserve_ticket(fs_info, space_info->flags, orig_bytes, 1544 start_ns, flush, ticket->error); 1545 return ret; 1546 } 1547 1548 /* 1549 * This returns true if this flush state will go through the ordinary flushing 1550 * code. 1551 */ 1552 static inline bool is_normal_flushing(enum btrfs_reserve_flush_enum flush) 1553 { 1554 return (flush == BTRFS_RESERVE_FLUSH_ALL) || 1555 (flush == BTRFS_RESERVE_FLUSH_ALL_STEAL); 1556 } 1557 1558 static inline void maybe_clamp_preempt(struct btrfs_fs_info *fs_info, 1559 struct btrfs_space_info *space_info) 1560 { 1561 u64 ordered = percpu_counter_sum_positive(&fs_info->ordered_bytes); 1562 u64 delalloc = percpu_counter_sum_positive(&fs_info->delalloc_bytes); 1563 1564 /* 1565 * If we're heavy on ordered operations then clamping won't help us. We 1566 * need to clamp specifically to keep up with dirty'ing buffered 1567 * writers, because there's not a 1:1 correlation of writing delalloc 1568 * and freeing space, like there is with flushing delayed refs or 1569 * delayed nodes. If we're already more ordered than delalloc then 1570 * we're keeping up, otherwise we aren't and should probably clamp. 1571 */ 1572 if (ordered < delalloc) 1573 space_info->clamp = min(space_info->clamp + 1, 8); 1574 } 1575 1576 static inline bool can_steal(enum btrfs_reserve_flush_enum flush) 1577 { 1578 return (flush == BTRFS_RESERVE_FLUSH_ALL_STEAL || 1579 flush == BTRFS_RESERVE_FLUSH_EVICT); 1580 } 1581 1582 /* 1583 * NO_FLUSH and FLUSH_EMERGENCY don't want to create a ticket, they just want to 1584 * fail as quickly as possible. 1585 */ 1586 static inline bool can_ticket(enum btrfs_reserve_flush_enum flush) 1587 { 1588 return (flush != BTRFS_RESERVE_NO_FLUSH && 1589 flush != BTRFS_RESERVE_FLUSH_EMERGENCY); 1590 } 1591 1592 /* 1593 * Try to reserve bytes from the block_rsv's space. 1594 * 1595 * @fs_info: the filesystem 1596 * @space_info: space info we want to allocate from 1597 * @orig_bytes: number of bytes we want 1598 * @flush: whether or not we can flush to make our reservation 1599 * 1600 * This will reserve orig_bytes number of bytes from the space info associated 1601 * with the block_rsv. If there is not enough space it will make an attempt to 1602 * flush out space to make room. It will do this by flushing delalloc if 1603 * possible or committing the transaction. If flush is 0 then no attempts to 1604 * regain reservations will be made and this will fail if there is not enough 1605 * space already. 1606 */ 1607 static int __reserve_bytes(struct btrfs_fs_info *fs_info, 1608 struct btrfs_space_info *space_info, u64 orig_bytes, 1609 enum btrfs_reserve_flush_enum flush) 1610 { 1611 struct work_struct *async_work; 1612 struct reserve_ticket ticket; 1613 u64 start_ns = 0; 1614 u64 used; 1615 int ret = -ENOSPC; 1616 bool pending_tickets; 1617 1618 ASSERT(orig_bytes); 1619 /* 1620 * If have a transaction handle (current->journal_info != NULL), then 1621 * the flush method can not be neither BTRFS_RESERVE_FLUSH_ALL* nor 1622 * BTRFS_RESERVE_FLUSH_EVICT, as we could deadlock because those 1623 * flushing methods can trigger transaction commits. 1624 */ 1625 if (current->journal_info) { 1626 /* One assert per line for easier debugging. */ 1627 ASSERT(flush != BTRFS_RESERVE_FLUSH_ALL); 1628 ASSERT(flush != BTRFS_RESERVE_FLUSH_ALL_STEAL); 1629 ASSERT(flush != BTRFS_RESERVE_FLUSH_EVICT); 1630 } 1631 1632 if (flush == BTRFS_RESERVE_FLUSH_DATA) 1633 async_work = &fs_info->async_data_reclaim_work; 1634 else 1635 async_work = &fs_info->async_reclaim_work; 1636 1637 spin_lock(&space_info->lock); 1638 used = btrfs_space_info_used(space_info, true); 1639 1640 /* 1641 * We don't want NO_FLUSH allocations to jump everybody, they can 1642 * generally handle ENOSPC in a different way, so treat them the same as 1643 * normal flushers when it comes to skipping pending tickets. 1644 */ 1645 if (is_normal_flushing(flush) || (flush == BTRFS_RESERVE_NO_FLUSH)) 1646 pending_tickets = !list_empty(&space_info->tickets) || 1647 !list_empty(&space_info->priority_tickets); 1648 else 1649 pending_tickets = !list_empty(&space_info->priority_tickets); 1650 1651 /* 1652 * Carry on if we have enough space (short-circuit) OR call 1653 * can_overcommit() to ensure we can overcommit to continue. 1654 */ 1655 if (!pending_tickets && 1656 ((used + orig_bytes <= space_info->total_bytes) || 1657 btrfs_can_overcommit(fs_info, space_info, orig_bytes, flush))) { 1658 btrfs_space_info_update_bytes_may_use(fs_info, space_info, 1659 orig_bytes); 1660 ret = 0; 1661 } 1662 1663 /* 1664 * Things are dire, we need to make a reservation so we don't abort. We 1665 * will let this reservation go through as long as we have actual space 1666 * left to allocate for the block. 1667 */ 1668 if (ret && unlikely(flush == BTRFS_RESERVE_FLUSH_EMERGENCY)) { 1669 used = btrfs_space_info_used(space_info, false); 1670 if (used + orig_bytes <= space_info->total_bytes) { 1671 btrfs_space_info_update_bytes_may_use(fs_info, space_info, 1672 orig_bytes); 1673 ret = 0; 1674 } 1675 } 1676 1677 /* 1678 * If we couldn't make a reservation then setup our reservation ticket 1679 * and kick the async worker if it's not already running. 1680 * 1681 * If we are a priority flusher then we just need to add our ticket to 1682 * the list and we will do our own flushing further down. 1683 */ 1684 if (ret && can_ticket(flush)) { 1685 ticket.bytes = orig_bytes; 1686 ticket.error = 0; 1687 space_info->reclaim_size += ticket.bytes; 1688 init_waitqueue_head(&ticket.wait); 1689 ticket.steal = can_steal(flush); 1690 if (trace_btrfs_reserve_ticket_enabled()) 1691 start_ns = ktime_get_ns(); 1692 1693 if (flush == BTRFS_RESERVE_FLUSH_ALL || 1694 flush == BTRFS_RESERVE_FLUSH_ALL_STEAL || 1695 flush == BTRFS_RESERVE_FLUSH_DATA) { 1696 list_add_tail(&ticket.list, &space_info->tickets); 1697 if (!space_info->flush) { 1698 /* 1699 * We were forced to add a reserve ticket, so 1700 * our preemptive flushing is unable to keep 1701 * up. Clamp down on the threshold for the 1702 * preemptive flushing in order to keep up with 1703 * the workload. 1704 */ 1705 maybe_clamp_preempt(fs_info, space_info); 1706 1707 space_info->flush = 1; 1708 trace_btrfs_trigger_flush(fs_info, 1709 space_info->flags, 1710 orig_bytes, flush, 1711 "enospc"); 1712 queue_work(system_unbound_wq, async_work); 1713 } 1714 } else { 1715 list_add_tail(&ticket.list, 1716 &space_info->priority_tickets); 1717 } 1718 } else if (!ret && space_info->flags & BTRFS_BLOCK_GROUP_METADATA) { 1719 /* 1720 * We will do the space reservation dance during log replay, 1721 * which means we won't have fs_info->fs_root set, so don't do 1722 * the async reclaim as we will panic. 1723 */ 1724 if (!test_bit(BTRFS_FS_LOG_RECOVERING, &fs_info->flags) && 1725 !work_busy(&fs_info->preempt_reclaim_work) && 1726 need_preemptive_reclaim(fs_info, space_info)) { 1727 trace_btrfs_trigger_flush(fs_info, space_info->flags, 1728 orig_bytes, flush, "preempt"); 1729 queue_work(system_unbound_wq, 1730 &fs_info->preempt_reclaim_work); 1731 } 1732 } 1733 spin_unlock(&space_info->lock); 1734 if (!ret || !can_ticket(flush)) 1735 return ret; 1736 1737 return handle_reserve_ticket(fs_info, space_info, &ticket, start_ns, 1738 orig_bytes, flush); 1739 } 1740 1741 /* 1742 * Try to reserve metadata bytes from the block_rsv's space. 1743 * 1744 * @fs_info: the filesystem 1745 * @block_rsv: block_rsv we're allocating for 1746 * @orig_bytes: number of bytes we want 1747 * @flush: whether or not we can flush to make our reservation 1748 * 1749 * This will reserve orig_bytes number of bytes from the space info associated 1750 * with the block_rsv. If there is not enough space it will make an attempt to 1751 * flush out space to make room. It will do this by flushing delalloc if 1752 * possible or committing the transaction. If flush is 0 then no attempts to 1753 * regain reservations will be made and this will fail if there is not enough 1754 * space already. 1755 */ 1756 int btrfs_reserve_metadata_bytes(struct btrfs_fs_info *fs_info, 1757 struct btrfs_block_rsv *block_rsv, 1758 u64 orig_bytes, 1759 enum btrfs_reserve_flush_enum flush) 1760 { 1761 int ret; 1762 1763 ret = __reserve_bytes(fs_info, block_rsv->space_info, orig_bytes, flush); 1764 if (ret == -ENOSPC) { 1765 trace_btrfs_space_reservation(fs_info, "space_info:enospc", 1766 block_rsv->space_info->flags, 1767 orig_bytes, 1); 1768 1769 if (btrfs_test_opt(fs_info, ENOSPC_DEBUG)) 1770 btrfs_dump_space_info(fs_info, block_rsv->space_info, 1771 orig_bytes, 0); 1772 } 1773 return ret; 1774 } 1775 1776 /* 1777 * Try to reserve data bytes for an allocation. 1778 * 1779 * @fs_info: the filesystem 1780 * @bytes: number of bytes we need 1781 * @flush: how we are allowed to flush 1782 * 1783 * This will reserve bytes from the data space info. If there is not enough 1784 * space then we will attempt to flush space as specified by flush. 1785 */ 1786 int btrfs_reserve_data_bytes(struct btrfs_fs_info *fs_info, u64 bytes, 1787 enum btrfs_reserve_flush_enum flush) 1788 { 1789 struct btrfs_space_info *data_sinfo = fs_info->data_sinfo; 1790 int ret; 1791 1792 ASSERT(flush == BTRFS_RESERVE_FLUSH_DATA || 1793 flush == BTRFS_RESERVE_FLUSH_FREE_SPACE_INODE || 1794 flush == BTRFS_RESERVE_NO_FLUSH); 1795 ASSERT(!current->journal_info || flush != BTRFS_RESERVE_FLUSH_DATA); 1796 1797 ret = __reserve_bytes(fs_info, data_sinfo, bytes, flush); 1798 if (ret == -ENOSPC) { 1799 trace_btrfs_space_reservation(fs_info, "space_info:enospc", 1800 data_sinfo->flags, bytes, 1); 1801 if (btrfs_test_opt(fs_info, ENOSPC_DEBUG)) 1802 btrfs_dump_space_info(fs_info, data_sinfo, bytes, 0); 1803 } 1804 return ret; 1805 } 1806 1807 /* Dump all the space infos when we abort a transaction due to ENOSPC. */ 1808 __cold void btrfs_dump_space_info_for_trans_abort(struct btrfs_fs_info *fs_info) 1809 { 1810 struct btrfs_space_info *space_info; 1811 1812 btrfs_info(fs_info, "dumping space info:"); 1813 list_for_each_entry(space_info, &fs_info->space_info, list) { 1814 spin_lock(&space_info->lock); 1815 __btrfs_dump_space_info(fs_info, space_info); 1816 spin_unlock(&space_info->lock); 1817 } 1818 dump_global_block_rsv(fs_info); 1819 } 1820 1821 /* 1822 * Account the unused space of all the readonly block group in the space_info. 1823 * takes mirrors into account. 1824 */ 1825 u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo) 1826 { 1827 struct btrfs_block_group *block_group; 1828 u64 free_bytes = 0; 1829 int factor; 1830 1831 /* It's df, we don't care if it's racy */ 1832 if (list_empty(&sinfo->ro_bgs)) 1833 return 0; 1834 1835 spin_lock(&sinfo->lock); 1836 list_for_each_entry(block_group, &sinfo->ro_bgs, ro_list) { 1837 spin_lock(&block_group->lock); 1838 1839 if (!block_group->ro) { 1840 spin_unlock(&block_group->lock); 1841 continue; 1842 } 1843 1844 factor = btrfs_bg_type_to_factor(block_group->flags); 1845 free_bytes += (block_group->length - 1846 block_group->used) * factor; 1847 1848 spin_unlock(&block_group->lock); 1849 } 1850 spin_unlock(&sinfo->lock); 1851 1852 return free_bytes; 1853 } 1854