1 // SPDX-License-Identifier: GPL-2.0 2 3 #include "misc.h" 4 #include "ctree.h" 5 #include "block-rsv.h" 6 #include "space-info.h" 7 #include "transaction.h" 8 9 /* 10 * HOW DO BLOCK RESERVES WORK 11 * 12 * Think of block_rsv's as buckets for logically grouped metadata 13 * reservations. Each block_rsv has a ->size and a ->reserved. ->size is 14 * how large we want our block rsv to be, ->reserved is how much space is 15 * currently reserved for this block reserve. 16 * 17 * ->failfast exists for the truncate case, and is described below. 18 * 19 * NORMAL OPERATION 20 * 21 * -> Reserve 22 * Entrance: btrfs_block_rsv_add, btrfs_block_rsv_refill 23 * 24 * We call into btrfs_reserve_metadata_bytes() with our bytes, which is 25 * accounted for in space_info->bytes_may_use, and then add the bytes to 26 * ->reserved, and ->size in the case of btrfs_block_rsv_add. 27 * 28 * ->size is an over-estimation of how much we may use for a particular 29 * operation. 30 * 31 * -> Use 32 * Entrance: btrfs_use_block_rsv 33 * 34 * When we do a btrfs_alloc_tree_block() we call into btrfs_use_block_rsv() 35 * to determine the appropriate block_rsv to use, and then verify that 36 * ->reserved has enough space for our tree block allocation. Once 37 * successful we subtract fs_info->nodesize from ->reserved. 38 * 39 * -> Finish 40 * Entrance: btrfs_block_rsv_release 41 * 42 * We are finished with our operation, subtract our individual reservation 43 * from ->size, and then subtract ->size from ->reserved and free up the 44 * excess if there is any. 45 * 46 * There is some logic here to refill the delayed refs rsv or the global rsv 47 * as needed, otherwise the excess is subtracted from 48 * space_info->bytes_may_use. 49 * 50 * TYPES OF BLOCK RESERVES 51 * 52 * BLOCK_RSV_TRANS, BLOCK_RSV_DELOPS, BLOCK_RSV_CHUNK 53 * These behave normally, as described above, just within the confines of the 54 * lifetime of their particular operation (transaction for the whole trans 55 * handle lifetime, for example). 56 * 57 * BLOCK_RSV_GLOBAL 58 * It is impossible to properly account for all the space that may be required 59 * to make our extent tree updates. This block reserve acts as an overflow 60 * buffer in case our delayed refs reserve does not reserve enough space to 61 * update the extent tree. 62 * 63 * We can steal from this in some cases as well, notably on evict() or 64 * truncate() in order to help users recover from ENOSPC conditions. 65 * 66 * BLOCK_RSV_DELALLOC 67 * The individual item sizes are determined by the per-inode size 68 * calculations, which are described with the delalloc code. This is pretty 69 * straightforward, it's just the calculation of ->size encodes a lot of 70 * different items, and thus it gets used when updating inodes, inserting file 71 * extents, and inserting checksums. 72 * 73 * BLOCK_RSV_DELREFS 74 * We keep a running tally of how many delayed refs we have on the system. 75 * We assume each one of these delayed refs are going to use a full 76 * reservation. We use the transaction items and pre-reserve space for every 77 * operation, and use this reservation to refill any gap between ->size and 78 * ->reserved that may exist. 79 * 80 * From there it's straightforward, removing a delayed ref means we remove its 81 * count from ->size and free up reservations as necessary. Since this is 82 * the most dynamic block reserve in the system, we will try to refill this 83 * block reserve first with any excess returned by any other block reserve. 84 * 85 * BLOCK_RSV_EMPTY 86 * This is the fallback block reserve to make us try to reserve space if we 87 * don't have a specific bucket for this allocation. It is mostly used for 88 * updating the device tree and such, since that is a separate pool we're 89 * content to just reserve space from the space_info on demand. 90 * 91 * BLOCK_RSV_TEMP 92 * This is used by things like truncate and iput. We will temporarily 93 * allocate a block reserve, set it to some size, and then truncate bytes 94 * until we have no space left. With ->failfast set we'll simply return 95 * ENOSPC from btrfs_use_block_rsv() to signal that we need to unwind and try 96 * to make a new reservation. This is because these operations are 97 * unbounded, so we want to do as much work as we can, and then back off and 98 * re-reserve. 99 */ 100 101 static u64 block_rsv_release_bytes(struct btrfs_fs_info *fs_info, 102 struct btrfs_block_rsv *block_rsv, 103 struct btrfs_block_rsv *dest, u64 num_bytes, 104 u64 *qgroup_to_release_ret) 105 { 106 struct btrfs_space_info *space_info = block_rsv->space_info; 107 u64 qgroup_to_release = 0; 108 u64 ret; 109 110 spin_lock(&block_rsv->lock); 111 if (num_bytes == (u64)-1) { 112 num_bytes = block_rsv->size; 113 qgroup_to_release = block_rsv->qgroup_rsv_size; 114 } 115 block_rsv->size -= num_bytes; 116 if (block_rsv->reserved >= block_rsv->size) { 117 num_bytes = block_rsv->reserved - block_rsv->size; 118 block_rsv->reserved = block_rsv->size; 119 block_rsv->full = 1; 120 } else { 121 num_bytes = 0; 122 } 123 if (block_rsv->qgroup_rsv_reserved >= block_rsv->qgroup_rsv_size) { 124 qgroup_to_release = block_rsv->qgroup_rsv_reserved - 125 block_rsv->qgroup_rsv_size; 126 block_rsv->qgroup_rsv_reserved = block_rsv->qgroup_rsv_size; 127 } else { 128 qgroup_to_release = 0; 129 } 130 spin_unlock(&block_rsv->lock); 131 132 ret = num_bytes; 133 if (num_bytes > 0) { 134 if (dest) { 135 spin_lock(&dest->lock); 136 if (!dest->full) { 137 u64 bytes_to_add; 138 139 bytes_to_add = dest->size - dest->reserved; 140 bytes_to_add = min(num_bytes, bytes_to_add); 141 dest->reserved += bytes_to_add; 142 if (dest->reserved >= dest->size) 143 dest->full = 1; 144 num_bytes -= bytes_to_add; 145 } 146 spin_unlock(&dest->lock); 147 } 148 if (num_bytes) 149 btrfs_space_info_free_bytes_may_use(fs_info, 150 space_info, 151 num_bytes); 152 } 153 if (qgroup_to_release_ret) 154 *qgroup_to_release_ret = qgroup_to_release; 155 return ret; 156 } 157 158 int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src, 159 struct btrfs_block_rsv *dst, u64 num_bytes, 160 bool update_size) 161 { 162 int ret; 163 164 ret = btrfs_block_rsv_use_bytes(src, num_bytes); 165 if (ret) 166 return ret; 167 168 btrfs_block_rsv_add_bytes(dst, num_bytes, update_size); 169 return 0; 170 } 171 172 void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv, unsigned short type) 173 { 174 memset(rsv, 0, sizeof(*rsv)); 175 spin_lock_init(&rsv->lock); 176 rsv->type = type; 177 } 178 179 void btrfs_init_metadata_block_rsv(struct btrfs_fs_info *fs_info, 180 struct btrfs_block_rsv *rsv, 181 unsigned short type) 182 { 183 btrfs_init_block_rsv(rsv, type); 184 rsv->space_info = btrfs_find_space_info(fs_info, 185 BTRFS_BLOCK_GROUP_METADATA); 186 } 187 188 struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_fs_info *fs_info, 189 unsigned short type) 190 { 191 struct btrfs_block_rsv *block_rsv; 192 193 block_rsv = kmalloc(sizeof(*block_rsv), GFP_NOFS); 194 if (!block_rsv) 195 return NULL; 196 197 btrfs_init_metadata_block_rsv(fs_info, block_rsv, type); 198 return block_rsv; 199 } 200 201 void btrfs_free_block_rsv(struct btrfs_fs_info *fs_info, 202 struct btrfs_block_rsv *rsv) 203 { 204 if (!rsv) 205 return; 206 btrfs_block_rsv_release(fs_info, rsv, (u64)-1, NULL); 207 kfree(rsv); 208 } 209 210 int btrfs_block_rsv_add(struct btrfs_root *root, 211 struct btrfs_block_rsv *block_rsv, u64 num_bytes, 212 enum btrfs_reserve_flush_enum flush) 213 { 214 int ret; 215 216 if (num_bytes == 0) 217 return 0; 218 219 ret = btrfs_reserve_metadata_bytes(root, block_rsv, num_bytes, flush); 220 if (!ret) 221 btrfs_block_rsv_add_bytes(block_rsv, num_bytes, true); 222 223 return ret; 224 } 225 226 int btrfs_block_rsv_check(struct btrfs_block_rsv *block_rsv, int min_factor) 227 { 228 u64 num_bytes = 0; 229 int ret = -ENOSPC; 230 231 if (!block_rsv) 232 return 0; 233 234 spin_lock(&block_rsv->lock); 235 num_bytes = div_factor(block_rsv->size, min_factor); 236 if (block_rsv->reserved >= num_bytes) 237 ret = 0; 238 spin_unlock(&block_rsv->lock); 239 240 return ret; 241 } 242 243 int btrfs_block_rsv_refill(struct btrfs_root *root, 244 struct btrfs_block_rsv *block_rsv, u64 min_reserved, 245 enum btrfs_reserve_flush_enum flush) 246 { 247 u64 num_bytes = 0; 248 int ret = -ENOSPC; 249 250 if (!block_rsv) 251 return 0; 252 253 spin_lock(&block_rsv->lock); 254 num_bytes = min_reserved; 255 if (block_rsv->reserved >= num_bytes) 256 ret = 0; 257 else 258 num_bytes -= block_rsv->reserved; 259 spin_unlock(&block_rsv->lock); 260 261 if (!ret) 262 return 0; 263 264 ret = btrfs_reserve_metadata_bytes(root, block_rsv, num_bytes, flush); 265 if (!ret) { 266 btrfs_block_rsv_add_bytes(block_rsv, num_bytes, false); 267 return 0; 268 } 269 270 return ret; 271 } 272 273 u64 btrfs_block_rsv_release(struct btrfs_fs_info *fs_info, 274 struct btrfs_block_rsv *block_rsv, u64 num_bytes, 275 u64 *qgroup_to_release) 276 { 277 struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv; 278 struct btrfs_block_rsv *delayed_rsv = &fs_info->delayed_refs_rsv; 279 struct btrfs_block_rsv *target = NULL; 280 281 /* 282 * If we are the delayed_rsv then push to the global rsv, otherwise dump 283 * into the delayed rsv if it is not full. 284 */ 285 if (block_rsv == delayed_rsv) 286 target = global_rsv; 287 else if (block_rsv != global_rsv && !delayed_rsv->full) 288 target = delayed_rsv; 289 290 if (target && block_rsv->space_info != target->space_info) 291 target = NULL; 292 293 return block_rsv_release_bytes(fs_info, block_rsv, target, num_bytes, 294 qgroup_to_release); 295 } 296 297 int btrfs_block_rsv_use_bytes(struct btrfs_block_rsv *block_rsv, u64 num_bytes) 298 { 299 int ret = -ENOSPC; 300 301 spin_lock(&block_rsv->lock); 302 if (block_rsv->reserved >= num_bytes) { 303 block_rsv->reserved -= num_bytes; 304 if (block_rsv->reserved < block_rsv->size) 305 block_rsv->full = 0; 306 ret = 0; 307 } 308 spin_unlock(&block_rsv->lock); 309 return ret; 310 } 311 312 void btrfs_block_rsv_add_bytes(struct btrfs_block_rsv *block_rsv, 313 u64 num_bytes, bool update_size) 314 { 315 spin_lock(&block_rsv->lock); 316 block_rsv->reserved += num_bytes; 317 if (update_size) 318 block_rsv->size += num_bytes; 319 else if (block_rsv->reserved >= block_rsv->size) 320 block_rsv->full = 1; 321 spin_unlock(&block_rsv->lock); 322 } 323 324 int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info, 325 struct btrfs_block_rsv *dest, u64 num_bytes, 326 int min_factor) 327 { 328 struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv; 329 u64 min_bytes; 330 331 if (global_rsv->space_info != dest->space_info) 332 return -ENOSPC; 333 334 spin_lock(&global_rsv->lock); 335 min_bytes = div_factor(global_rsv->size, min_factor); 336 if (global_rsv->reserved < min_bytes + num_bytes) { 337 spin_unlock(&global_rsv->lock); 338 return -ENOSPC; 339 } 340 global_rsv->reserved -= num_bytes; 341 if (global_rsv->reserved < global_rsv->size) 342 global_rsv->full = 0; 343 spin_unlock(&global_rsv->lock); 344 345 btrfs_block_rsv_add_bytes(dest, num_bytes, true); 346 return 0; 347 } 348 349 void btrfs_update_global_block_rsv(struct btrfs_fs_info *fs_info) 350 { 351 struct btrfs_block_rsv *block_rsv = &fs_info->global_block_rsv; 352 struct btrfs_space_info *sinfo = block_rsv->space_info; 353 u64 num_bytes; 354 unsigned min_items; 355 356 /* 357 * The global block rsv is based on the size of the extent tree, the 358 * checksum tree and the root tree. If the fs is empty we want to set 359 * it to a minimal amount for safety. 360 */ 361 num_bytes = btrfs_root_used(&fs_info->extent_root->root_item) + 362 btrfs_root_used(&fs_info->csum_root->root_item) + 363 btrfs_root_used(&fs_info->tree_root->root_item); 364 365 /* 366 * We at a minimum are going to modify the csum root, the tree root, and 367 * the extent root. 368 */ 369 min_items = 3; 370 371 /* 372 * But we also want to reserve enough space so we can do the fallback 373 * global reserve for an unlink, which is an additional 5 items (see the 374 * comment in __unlink_start_trans for what we're modifying.) 375 * 376 * But we also need space for the delayed ref updates from the unlink, 377 * so its 10, 5 for the actual operation, and 5 for the delayed ref 378 * updates. 379 */ 380 min_items += 10; 381 382 num_bytes = max_t(u64, num_bytes, 383 btrfs_calc_insert_metadata_size(fs_info, min_items)); 384 385 spin_lock(&sinfo->lock); 386 spin_lock(&block_rsv->lock); 387 388 block_rsv->size = min_t(u64, num_bytes, SZ_512M); 389 390 if (block_rsv->reserved < block_rsv->size) { 391 num_bytes = block_rsv->size - block_rsv->reserved; 392 btrfs_space_info_update_bytes_may_use(fs_info, sinfo, 393 num_bytes); 394 block_rsv->reserved = block_rsv->size; 395 } else if (block_rsv->reserved > block_rsv->size) { 396 num_bytes = block_rsv->reserved - block_rsv->size; 397 btrfs_space_info_update_bytes_may_use(fs_info, sinfo, 398 -num_bytes); 399 block_rsv->reserved = block_rsv->size; 400 btrfs_try_granting_tickets(fs_info, sinfo); 401 } 402 403 if (block_rsv->reserved == block_rsv->size) 404 block_rsv->full = 1; 405 else 406 block_rsv->full = 0; 407 408 spin_unlock(&block_rsv->lock); 409 spin_unlock(&sinfo->lock); 410 } 411 412 void btrfs_init_global_block_rsv(struct btrfs_fs_info *fs_info) 413 { 414 struct btrfs_space_info *space_info; 415 416 space_info = btrfs_find_space_info(fs_info, BTRFS_BLOCK_GROUP_SYSTEM); 417 fs_info->chunk_block_rsv.space_info = space_info; 418 419 space_info = btrfs_find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); 420 fs_info->global_block_rsv.space_info = space_info; 421 fs_info->trans_block_rsv.space_info = space_info; 422 fs_info->empty_block_rsv.space_info = space_info; 423 fs_info->delayed_block_rsv.space_info = space_info; 424 fs_info->delayed_refs_rsv.space_info = space_info; 425 426 fs_info->extent_root->block_rsv = &fs_info->delayed_refs_rsv; 427 fs_info->csum_root->block_rsv = &fs_info->delayed_refs_rsv; 428 fs_info->dev_root->block_rsv = &fs_info->global_block_rsv; 429 fs_info->tree_root->block_rsv = &fs_info->global_block_rsv; 430 if (fs_info->quota_root) 431 fs_info->quota_root->block_rsv = &fs_info->global_block_rsv; 432 fs_info->chunk_root->block_rsv = &fs_info->chunk_block_rsv; 433 434 btrfs_update_global_block_rsv(fs_info); 435 } 436 437 void btrfs_release_global_block_rsv(struct btrfs_fs_info *fs_info) 438 { 439 btrfs_block_rsv_release(fs_info, &fs_info->global_block_rsv, (u64)-1, 440 NULL); 441 WARN_ON(fs_info->trans_block_rsv.size > 0); 442 WARN_ON(fs_info->trans_block_rsv.reserved > 0); 443 WARN_ON(fs_info->chunk_block_rsv.size > 0); 444 WARN_ON(fs_info->chunk_block_rsv.reserved > 0); 445 WARN_ON(fs_info->delayed_block_rsv.size > 0); 446 WARN_ON(fs_info->delayed_block_rsv.reserved > 0); 447 WARN_ON(fs_info->delayed_refs_rsv.reserved > 0); 448 WARN_ON(fs_info->delayed_refs_rsv.size > 0); 449 } 450 451 static struct btrfs_block_rsv *get_block_rsv( 452 const struct btrfs_trans_handle *trans, 453 const struct btrfs_root *root) 454 { 455 struct btrfs_fs_info *fs_info = root->fs_info; 456 struct btrfs_block_rsv *block_rsv = NULL; 457 458 if (test_bit(BTRFS_ROOT_REF_COWS, &root->state) || 459 (root == fs_info->csum_root && trans->adding_csums) || 460 (root == fs_info->uuid_root)) 461 block_rsv = trans->block_rsv; 462 463 if (!block_rsv) 464 block_rsv = root->block_rsv; 465 466 if (!block_rsv) 467 block_rsv = &fs_info->empty_block_rsv; 468 469 return block_rsv; 470 } 471 472 struct btrfs_block_rsv *btrfs_use_block_rsv(struct btrfs_trans_handle *trans, 473 struct btrfs_root *root, 474 u32 blocksize) 475 { 476 struct btrfs_fs_info *fs_info = root->fs_info; 477 struct btrfs_block_rsv *block_rsv; 478 struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv; 479 int ret; 480 bool global_updated = false; 481 482 block_rsv = get_block_rsv(trans, root); 483 484 if (unlikely(block_rsv->size == 0)) 485 goto try_reserve; 486 again: 487 ret = btrfs_block_rsv_use_bytes(block_rsv, blocksize); 488 if (!ret) 489 return block_rsv; 490 491 if (block_rsv->failfast) 492 return ERR_PTR(ret); 493 494 if (block_rsv->type == BTRFS_BLOCK_RSV_GLOBAL && !global_updated) { 495 global_updated = true; 496 btrfs_update_global_block_rsv(fs_info); 497 goto again; 498 } 499 500 /* 501 * The global reserve still exists to save us from ourselves, so don't 502 * warn_on if we are short on our delayed refs reserve. 503 */ 504 if (block_rsv->type != BTRFS_BLOCK_RSV_DELREFS && 505 btrfs_test_opt(fs_info, ENOSPC_DEBUG)) { 506 static DEFINE_RATELIMIT_STATE(_rs, 507 DEFAULT_RATELIMIT_INTERVAL * 10, 508 /*DEFAULT_RATELIMIT_BURST*/ 1); 509 if (__ratelimit(&_rs)) 510 WARN(1, KERN_DEBUG 511 "BTRFS: block rsv returned %d\n", ret); 512 } 513 try_reserve: 514 ret = btrfs_reserve_metadata_bytes(root, block_rsv, blocksize, 515 BTRFS_RESERVE_NO_FLUSH); 516 if (!ret) 517 return block_rsv; 518 /* 519 * If we couldn't reserve metadata bytes try and use some from 520 * the global reserve if its space type is the same as the global 521 * reservation. 522 */ 523 if (block_rsv->type != BTRFS_BLOCK_RSV_GLOBAL && 524 block_rsv->space_info == global_rsv->space_info) { 525 ret = btrfs_block_rsv_use_bytes(global_rsv, blocksize); 526 if (!ret) 527 return global_rsv; 528 } 529 return ERR_PTR(ret); 530 } 531