| 1 | /* |
| 2 | * Copyright (C) 2015 Facebook. All rights reserved. |
| 3 | * |
| 4 | * This program is free software; you can redistribute it and/or |
| 5 | * modify it under the terms of the GNU General Public |
| 6 | * License v2 as published by the Free Software Foundation. |
| 7 | * |
| 8 | * This program is distributed in the hope that it will be useful, |
| 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 11 | * General Public License for more details. |
| 12 | * |
| 13 | * You should have received a copy of the GNU General Public |
| 14 | * License along with this program; if not, write to the |
| 15 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, |
| 16 | * Boston, MA 021110-1307, USA. |
| 17 | */ |
| 18 | |
| 19 | #include <linux/kernel.h> |
| 20 | #include <linux/vmalloc.h> |
| 21 | #include "ctree.h" |
| 22 | #include "disk-io.h" |
| 23 | #include "locking.h" |
| 24 | #include "free-space-tree.h" |
| 25 | #include "transaction.h" |
| 26 | |
| 27 | static int __add_block_group_free_space(struct btrfs_trans_handle *trans, |
| 28 | struct btrfs_fs_info *fs_info, |
| 29 | struct btrfs_block_group_cache *block_group, |
| 30 | struct btrfs_path *path); |
| 31 | |
| 32 | void set_free_space_tree_thresholds(struct btrfs_block_group_cache *cache) |
| 33 | { |
| 34 | u32 bitmap_range; |
| 35 | size_t bitmap_size; |
| 36 | u64 num_bitmaps, total_bitmap_size; |
| 37 | |
| 38 | /* |
| 39 | * We convert to bitmaps when the disk space required for using extents |
| 40 | * exceeds that required for using bitmaps. |
| 41 | */ |
| 42 | bitmap_range = cache->sectorsize * BTRFS_FREE_SPACE_BITMAP_BITS; |
| 43 | num_bitmaps = div_u64(cache->key.offset + bitmap_range - 1, |
| 44 | bitmap_range); |
| 45 | bitmap_size = sizeof(struct btrfs_item) + BTRFS_FREE_SPACE_BITMAP_SIZE; |
| 46 | total_bitmap_size = num_bitmaps * bitmap_size; |
| 47 | cache->bitmap_high_thresh = div_u64(total_bitmap_size, |
| 48 | sizeof(struct btrfs_item)); |
| 49 | |
| 50 | /* |
| 51 | * We allow for a small buffer between the high threshold and low |
| 52 | * threshold to avoid thrashing back and forth between the two formats. |
| 53 | */ |
| 54 | if (cache->bitmap_high_thresh > 100) |
| 55 | cache->bitmap_low_thresh = cache->bitmap_high_thresh - 100; |
| 56 | else |
| 57 | cache->bitmap_low_thresh = 0; |
| 58 | } |
| 59 | |
| 60 | static int add_new_free_space_info(struct btrfs_trans_handle *trans, |
| 61 | struct btrfs_fs_info *fs_info, |
| 62 | struct btrfs_block_group_cache *block_group, |
| 63 | struct btrfs_path *path) |
| 64 | { |
| 65 | struct btrfs_root *root = fs_info->free_space_root; |
| 66 | struct btrfs_free_space_info *info; |
| 67 | struct btrfs_key key; |
| 68 | struct extent_buffer *leaf; |
| 69 | int ret; |
| 70 | |
| 71 | key.objectid = block_group->key.objectid; |
| 72 | key.type = BTRFS_FREE_SPACE_INFO_KEY; |
| 73 | key.offset = block_group->key.offset; |
| 74 | |
| 75 | ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(*info)); |
| 76 | if (ret) |
| 77 | goto out; |
| 78 | |
| 79 | leaf = path->nodes[0]; |
| 80 | info = btrfs_item_ptr(leaf, path->slots[0], |
| 81 | struct btrfs_free_space_info); |
| 82 | btrfs_set_free_space_extent_count(leaf, info, 0); |
| 83 | btrfs_set_free_space_flags(leaf, info, 0); |
| 84 | btrfs_mark_buffer_dirty(leaf); |
| 85 | |
| 86 | ret = 0; |
| 87 | out: |
| 88 | btrfs_release_path(path); |
| 89 | return ret; |
| 90 | } |
| 91 | |
| 92 | struct btrfs_free_space_info * |
| 93 | search_free_space_info(struct btrfs_trans_handle *trans, |
| 94 | struct btrfs_fs_info *fs_info, |
| 95 | struct btrfs_block_group_cache *block_group, |
| 96 | struct btrfs_path *path, int cow) |
| 97 | { |
| 98 | struct btrfs_root *root = fs_info->free_space_root; |
| 99 | struct btrfs_key key; |
| 100 | int ret; |
| 101 | |
| 102 | key.objectid = block_group->key.objectid; |
| 103 | key.type = BTRFS_FREE_SPACE_INFO_KEY; |
| 104 | key.offset = block_group->key.offset; |
| 105 | |
| 106 | ret = btrfs_search_slot(trans, root, &key, path, 0, cow); |
| 107 | if (ret < 0) |
| 108 | return ERR_PTR(ret); |
| 109 | if (ret != 0) { |
| 110 | btrfs_warn(fs_info, "missing free space info for %llu\n", |
| 111 | block_group->key.objectid); |
| 112 | ASSERT(0); |
| 113 | return ERR_PTR(-ENOENT); |
| 114 | } |
| 115 | |
| 116 | return btrfs_item_ptr(path->nodes[0], path->slots[0], |
| 117 | struct btrfs_free_space_info); |
| 118 | } |
| 119 | |
| 120 | /* |
| 121 | * btrfs_search_slot() but we're looking for the greatest key less than the |
| 122 | * passed key. |
| 123 | */ |
| 124 | static int btrfs_search_prev_slot(struct btrfs_trans_handle *trans, |
| 125 | struct btrfs_root *root, |
| 126 | struct btrfs_key *key, struct btrfs_path *p, |
| 127 | int ins_len, int cow) |
| 128 | { |
| 129 | int ret; |
| 130 | |
| 131 | ret = btrfs_search_slot(trans, root, key, p, ins_len, cow); |
| 132 | if (ret < 0) |
| 133 | return ret; |
| 134 | |
| 135 | if (ret == 0) { |
| 136 | ASSERT(0); |
| 137 | return -EIO; |
| 138 | } |
| 139 | |
| 140 | if (p->slots[0] == 0) { |
| 141 | ASSERT(0); |
| 142 | return -EIO; |
| 143 | } |
| 144 | p->slots[0]--; |
| 145 | |
| 146 | return 0; |
| 147 | } |
| 148 | |
| 149 | static inline u32 free_space_bitmap_size(u64 size, u32 sectorsize) |
| 150 | { |
| 151 | return DIV_ROUND_UP((u32)div_u64(size, sectorsize), BITS_PER_BYTE); |
| 152 | } |
| 153 | |
| 154 | static unsigned long *alloc_bitmap(u32 bitmap_size) |
| 155 | { |
| 156 | void *mem; |
| 157 | |
| 158 | /* |
| 159 | * The allocation size varies, observed numbers were < 4K up to 16K. |
| 160 | * Using vmalloc unconditionally would be too heavy, we'll try |
| 161 | * contiguous allocations first. |
| 162 | */ |
| 163 | if (bitmap_size <= PAGE_SIZE) |
| 164 | return kzalloc(bitmap_size, GFP_NOFS); |
| 165 | |
| 166 | mem = kzalloc(bitmap_size, GFP_NOFS | __GFP_NOWARN); |
| 167 | if (mem) |
| 168 | return mem; |
| 169 | |
| 170 | return __vmalloc(bitmap_size, GFP_NOFS | __GFP_HIGHMEM | __GFP_ZERO, |
| 171 | PAGE_KERNEL); |
| 172 | } |
| 173 | |
| 174 | int convert_free_space_to_bitmaps(struct btrfs_trans_handle *trans, |
| 175 | struct btrfs_fs_info *fs_info, |
| 176 | struct btrfs_block_group_cache *block_group, |
| 177 | struct btrfs_path *path) |
| 178 | { |
| 179 | struct btrfs_root *root = fs_info->free_space_root; |
| 180 | struct btrfs_free_space_info *info; |
| 181 | struct btrfs_key key, found_key; |
| 182 | struct extent_buffer *leaf; |
| 183 | unsigned long *bitmap; |
| 184 | char *bitmap_cursor; |
| 185 | u64 start, end; |
| 186 | u64 bitmap_range, i; |
| 187 | u32 bitmap_size, flags, expected_extent_count; |
| 188 | u32 extent_count = 0; |
| 189 | int done = 0, nr; |
| 190 | int ret; |
| 191 | |
| 192 | bitmap_size = free_space_bitmap_size(block_group->key.offset, |
| 193 | block_group->sectorsize); |
| 194 | bitmap = alloc_bitmap(bitmap_size); |
| 195 | if (!bitmap) { |
| 196 | ret = -ENOMEM; |
| 197 | goto out; |
| 198 | } |
| 199 | |
| 200 | start = block_group->key.objectid; |
| 201 | end = block_group->key.objectid + block_group->key.offset; |
| 202 | |
| 203 | key.objectid = end - 1; |
| 204 | key.type = (u8)-1; |
| 205 | key.offset = (u64)-1; |
| 206 | |
| 207 | while (!done) { |
| 208 | ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1); |
| 209 | if (ret) |
| 210 | goto out; |
| 211 | |
| 212 | leaf = path->nodes[0]; |
| 213 | nr = 0; |
| 214 | path->slots[0]++; |
| 215 | while (path->slots[0] > 0) { |
| 216 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1); |
| 217 | |
| 218 | if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) { |
| 219 | ASSERT(found_key.objectid == block_group->key.objectid); |
| 220 | ASSERT(found_key.offset == block_group->key.offset); |
| 221 | done = 1; |
| 222 | break; |
| 223 | } else if (found_key.type == BTRFS_FREE_SPACE_EXTENT_KEY) { |
| 224 | u64 first, last; |
| 225 | |
| 226 | ASSERT(found_key.objectid >= start); |
| 227 | ASSERT(found_key.objectid < end); |
| 228 | ASSERT(found_key.objectid + found_key.offset <= end); |
| 229 | |
| 230 | first = div_u64(found_key.objectid - start, |
| 231 | block_group->sectorsize); |
| 232 | last = div_u64(found_key.objectid + found_key.offset - start, |
| 233 | block_group->sectorsize); |
| 234 | bitmap_set(bitmap, first, last - first); |
| 235 | |
| 236 | extent_count++; |
| 237 | nr++; |
| 238 | path->slots[0]--; |
| 239 | } else { |
| 240 | ASSERT(0); |
| 241 | } |
| 242 | } |
| 243 | |
| 244 | ret = btrfs_del_items(trans, root, path, path->slots[0], nr); |
| 245 | if (ret) |
| 246 | goto out; |
| 247 | btrfs_release_path(path); |
| 248 | } |
| 249 | |
| 250 | info = search_free_space_info(trans, fs_info, block_group, path, 1); |
| 251 | if (IS_ERR(info)) { |
| 252 | ret = PTR_ERR(info); |
| 253 | goto out; |
| 254 | } |
| 255 | leaf = path->nodes[0]; |
| 256 | flags = btrfs_free_space_flags(leaf, info); |
| 257 | flags |= BTRFS_FREE_SPACE_USING_BITMAPS; |
| 258 | btrfs_set_free_space_flags(leaf, info, flags); |
| 259 | expected_extent_count = btrfs_free_space_extent_count(leaf, info); |
| 260 | btrfs_mark_buffer_dirty(leaf); |
| 261 | btrfs_release_path(path); |
| 262 | |
| 263 | if (extent_count != expected_extent_count) { |
| 264 | btrfs_err(fs_info, "incorrect extent count for %llu; counted %u, expected %u", |
| 265 | block_group->key.objectid, extent_count, |
| 266 | expected_extent_count); |
| 267 | ASSERT(0); |
| 268 | ret = -EIO; |
| 269 | goto out; |
| 270 | } |
| 271 | |
| 272 | bitmap_cursor = (char *)bitmap; |
| 273 | bitmap_range = block_group->sectorsize * BTRFS_FREE_SPACE_BITMAP_BITS; |
| 274 | i = start; |
| 275 | while (i < end) { |
| 276 | unsigned long ptr; |
| 277 | u64 extent_size; |
| 278 | u32 data_size; |
| 279 | |
| 280 | extent_size = min(end - i, bitmap_range); |
| 281 | data_size = free_space_bitmap_size(extent_size, |
| 282 | block_group->sectorsize); |
| 283 | |
| 284 | key.objectid = i; |
| 285 | key.type = BTRFS_FREE_SPACE_BITMAP_KEY; |
| 286 | key.offset = extent_size; |
| 287 | |
| 288 | ret = btrfs_insert_empty_item(trans, root, path, &key, |
| 289 | data_size); |
| 290 | if (ret) |
| 291 | goto out; |
| 292 | |
| 293 | leaf = path->nodes[0]; |
| 294 | ptr = btrfs_item_ptr_offset(leaf, path->slots[0]); |
| 295 | write_extent_buffer(leaf, bitmap_cursor, ptr, |
| 296 | data_size); |
| 297 | btrfs_mark_buffer_dirty(leaf); |
| 298 | btrfs_release_path(path); |
| 299 | |
| 300 | i += extent_size; |
| 301 | bitmap_cursor += data_size; |
| 302 | } |
| 303 | |
| 304 | ret = 0; |
| 305 | out: |
| 306 | kvfree(bitmap); |
| 307 | if (ret) |
| 308 | btrfs_abort_transaction(trans, root, ret); |
| 309 | return ret; |
| 310 | } |
| 311 | |
| 312 | int convert_free_space_to_extents(struct btrfs_trans_handle *trans, |
| 313 | struct btrfs_fs_info *fs_info, |
| 314 | struct btrfs_block_group_cache *block_group, |
| 315 | struct btrfs_path *path) |
| 316 | { |
| 317 | struct btrfs_root *root = fs_info->free_space_root; |
| 318 | struct btrfs_free_space_info *info; |
| 319 | struct btrfs_key key, found_key; |
| 320 | struct extent_buffer *leaf; |
| 321 | unsigned long *bitmap; |
| 322 | u64 start, end; |
| 323 | /* Initialize to silence GCC. */ |
| 324 | u64 extent_start = 0; |
| 325 | u64 offset; |
| 326 | u32 bitmap_size, flags, expected_extent_count; |
| 327 | int prev_bit = 0, bit, bitnr; |
| 328 | u32 extent_count = 0; |
| 329 | int done = 0, nr; |
| 330 | int ret; |
| 331 | |
| 332 | bitmap_size = free_space_bitmap_size(block_group->key.offset, |
| 333 | block_group->sectorsize); |
| 334 | bitmap = alloc_bitmap(bitmap_size); |
| 335 | if (!bitmap) { |
| 336 | ret = -ENOMEM; |
| 337 | goto out; |
| 338 | } |
| 339 | |
| 340 | start = block_group->key.objectid; |
| 341 | end = block_group->key.objectid + block_group->key.offset; |
| 342 | |
| 343 | key.objectid = end - 1; |
| 344 | key.type = (u8)-1; |
| 345 | key.offset = (u64)-1; |
| 346 | |
| 347 | while (!done) { |
| 348 | ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1); |
| 349 | if (ret) |
| 350 | goto out; |
| 351 | |
| 352 | leaf = path->nodes[0]; |
| 353 | nr = 0; |
| 354 | path->slots[0]++; |
| 355 | while (path->slots[0] > 0) { |
| 356 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1); |
| 357 | |
| 358 | if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) { |
| 359 | ASSERT(found_key.objectid == block_group->key.objectid); |
| 360 | ASSERT(found_key.offset == block_group->key.offset); |
| 361 | done = 1; |
| 362 | break; |
| 363 | } else if (found_key.type == BTRFS_FREE_SPACE_BITMAP_KEY) { |
| 364 | unsigned long ptr; |
| 365 | char *bitmap_cursor; |
| 366 | u32 bitmap_pos, data_size; |
| 367 | |
| 368 | ASSERT(found_key.objectid >= start); |
| 369 | ASSERT(found_key.objectid < end); |
| 370 | ASSERT(found_key.objectid + found_key.offset <= end); |
| 371 | |
| 372 | bitmap_pos = div_u64(found_key.objectid - start, |
| 373 | block_group->sectorsize * |
| 374 | BITS_PER_BYTE); |
| 375 | bitmap_cursor = ((char *)bitmap) + bitmap_pos; |
| 376 | data_size = free_space_bitmap_size(found_key.offset, |
| 377 | block_group->sectorsize); |
| 378 | |
| 379 | ptr = btrfs_item_ptr_offset(leaf, path->slots[0] - 1); |
| 380 | read_extent_buffer(leaf, bitmap_cursor, ptr, |
| 381 | data_size); |
| 382 | |
| 383 | nr++; |
| 384 | path->slots[0]--; |
| 385 | } else { |
| 386 | ASSERT(0); |
| 387 | } |
| 388 | } |
| 389 | |
| 390 | ret = btrfs_del_items(trans, root, path, path->slots[0], nr); |
| 391 | if (ret) |
| 392 | goto out; |
| 393 | btrfs_release_path(path); |
| 394 | } |
| 395 | |
| 396 | info = search_free_space_info(trans, fs_info, block_group, path, 1); |
| 397 | if (IS_ERR(info)) { |
| 398 | ret = PTR_ERR(info); |
| 399 | goto out; |
| 400 | } |
| 401 | leaf = path->nodes[0]; |
| 402 | flags = btrfs_free_space_flags(leaf, info); |
| 403 | flags &= ~BTRFS_FREE_SPACE_USING_BITMAPS; |
| 404 | btrfs_set_free_space_flags(leaf, info, flags); |
| 405 | expected_extent_count = btrfs_free_space_extent_count(leaf, info); |
| 406 | btrfs_mark_buffer_dirty(leaf); |
| 407 | btrfs_release_path(path); |
| 408 | |
| 409 | offset = start; |
| 410 | bitnr = 0; |
| 411 | while (offset < end) { |
| 412 | bit = !!test_bit(bitnr, bitmap); |
| 413 | if (prev_bit == 0 && bit == 1) { |
| 414 | extent_start = offset; |
| 415 | } else if (prev_bit == 1 && bit == 0) { |
| 416 | key.objectid = extent_start; |
| 417 | key.type = BTRFS_FREE_SPACE_EXTENT_KEY; |
| 418 | key.offset = offset - extent_start; |
| 419 | |
| 420 | ret = btrfs_insert_empty_item(trans, root, path, &key, 0); |
| 421 | if (ret) |
| 422 | goto out; |
| 423 | btrfs_release_path(path); |
| 424 | |
| 425 | extent_count++; |
| 426 | } |
| 427 | prev_bit = bit; |
| 428 | offset += block_group->sectorsize; |
| 429 | bitnr++; |
| 430 | } |
| 431 | if (prev_bit == 1) { |
| 432 | key.objectid = extent_start; |
| 433 | key.type = BTRFS_FREE_SPACE_EXTENT_KEY; |
| 434 | key.offset = end - extent_start; |
| 435 | |
| 436 | ret = btrfs_insert_empty_item(trans, root, path, &key, 0); |
| 437 | if (ret) |
| 438 | goto out; |
| 439 | btrfs_release_path(path); |
| 440 | |
| 441 | extent_count++; |
| 442 | } |
| 443 | |
| 444 | if (extent_count != expected_extent_count) { |
| 445 | btrfs_err(fs_info, "incorrect extent count for %llu; counted %u, expected %u", |
| 446 | block_group->key.objectid, extent_count, |
| 447 | expected_extent_count); |
| 448 | ASSERT(0); |
| 449 | ret = -EIO; |
| 450 | goto out; |
| 451 | } |
| 452 | |
| 453 | ret = 0; |
| 454 | out: |
| 455 | kvfree(bitmap); |
| 456 | if (ret) |
| 457 | btrfs_abort_transaction(trans, root, ret); |
| 458 | return ret; |
| 459 | } |
| 460 | |
| 461 | static int update_free_space_extent_count(struct btrfs_trans_handle *trans, |
| 462 | struct btrfs_fs_info *fs_info, |
| 463 | struct btrfs_block_group_cache *block_group, |
| 464 | struct btrfs_path *path, |
| 465 | int new_extents) |
| 466 | { |
| 467 | struct btrfs_free_space_info *info; |
| 468 | u32 flags; |
| 469 | u32 extent_count; |
| 470 | int ret = 0; |
| 471 | |
| 472 | if (new_extents == 0) |
| 473 | return 0; |
| 474 | |
| 475 | info = search_free_space_info(trans, fs_info, block_group, path, 1); |
| 476 | if (IS_ERR(info)) { |
| 477 | ret = PTR_ERR(info); |
| 478 | goto out; |
| 479 | } |
| 480 | flags = btrfs_free_space_flags(path->nodes[0], info); |
| 481 | extent_count = btrfs_free_space_extent_count(path->nodes[0], info); |
| 482 | |
| 483 | extent_count += new_extents; |
| 484 | btrfs_set_free_space_extent_count(path->nodes[0], info, extent_count); |
| 485 | btrfs_mark_buffer_dirty(path->nodes[0]); |
| 486 | btrfs_release_path(path); |
| 487 | |
| 488 | if (!(flags & BTRFS_FREE_SPACE_USING_BITMAPS) && |
| 489 | extent_count > block_group->bitmap_high_thresh) { |
| 490 | ret = convert_free_space_to_bitmaps(trans, fs_info, block_group, |
| 491 | path); |
| 492 | } else if ((flags & BTRFS_FREE_SPACE_USING_BITMAPS) && |
| 493 | extent_count < block_group->bitmap_low_thresh) { |
| 494 | ret = convert_free_space_to_extents(trans, fs_info, block_group, |
| 495 | path); |
| 496 | } |
| 497 | |
| 498 | out: |
| 499 | return ret; |
| 500 | } |
| 501 | |
| 502 | int free_space_test_bit(struct btrfs_block_group_cache *block_group, |
| 503 | struct btrfs_path *path, u64 offset) |
| 504 | { |
| 505 | struct extent_buffer *leaf; |
| 506 | struct btrfs_key key; |
| 507 | u64 found_start, found_end; |
| 508 | unsigned long ptr, i; |
| 509 | |
| 510 | leaf = path->nodes[0]; |
| 511 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 512 | ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY); |
| 513 | |
| 514 | found_start = key.objectid; |
| 515 | found_end = key.objectid + key.offset; |
| 516 | ASSERT(offset >= found_start && offset < found_end); |
| 517 | |
| 518 | ptr = btrfs_item_ptr_offset(leaf, path->slots[0]); |
| 519 | i = div_u64(offset - found_start, block_group->sectorsize); |
| 520 | return !!extent_buffer_test_bit(leaf, ptr, i); |
| 521 | } |
| 522 | |
| 523 | static void free_space_set_bits(struct btrfs_block_group_cache *block_group, |
| 524 | struct btrfs_path *path, u64 *start, u64 *size, |
| 525 | int bit) |
| 526 | { |
| 527 | struct extent_buffer *leaf; |
| 528 | struct btrfs_key key; |
| 529 | u64 end = *start + *size; |
| 530 | u64 found_start, found_end; |
| 531 | unsigned long ptr, first, last; |
| 532 | |
| 533 | leaf = path->nodes[0]; |
| 534 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 535 | ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY); |
| 536 | |
| 537 | found_start = key.objectid; |
| 538 | found_end = key.objectid + key.offset; |
| 539 | ASSERT(*start >= found_start && *start < found_end); |
| 540 | ASSERT(end > found_start); |
| 541 | |
| 542 | if (end > found_end) |
| 543 | end = found_end; |
| 544 | |
| 545 | ptr = btrfs_item_ptr_offset(leaf, path->slots[0]); |
| 546 | first = div_u64(*start - found_start, block_group->sectorsize); |
| 547 | last = div_u64(end - found_start, block_group->sectorsize); |
| 548 | if (bit) |
| 549 | extent_buffer_bitmap_set(leaf, ptr, first, last - first); |
| 550 | else |
| 551 | extent_buffer_bitmap_clear(leaf, ptr, first, last - first); |
| 552 | btrfs_mark_buffer_dirty(leaf); |
| 553 | |
| 554 | *size -= end - *start; |
| 555 | *start = end; |
| 556 | } |
| 557 | |
| 558 | /* |
| 559 | * We can't use btrfs_next_item() in modify_free_space_bitmap() because |
| 560 | * btrfs_next_leaf() doesn't get the path for writing. We can forgo the fancy |
| 561 | * tree walking in btrfs_next_leaf() anyways because we know exactly what we're |
| 562 | * looking for. |
| 563 | */ |
| 564 | static int free_space_next_bitmap(struct btrfs_trans_handle *trans, |
| 565 | struct btrfs_root *root, struct btrfs_path *p) |
| 566 | { |
| 567 | struct btrfs_key key; |
| 568 | |
| 569 | if (p->slots[0] + 1 < btrfs_header_nritems(p->nodes[0])) { |
| 570 | p->slots[0]++; |
| 571 | return 0; |
| 572 | } |
| 573 | |
| 574 | btrfs_item_key_to_cpu(p->nodes[0], &key, p->slots[0]); |
| 575 | btrfs_release_path(p); |
| 576 | |
| 577 | key.objectid += key.offset; |
| 578 | key.type = (u8)-1; |
| 579 | key.offset = (u64)-1; |
| 580 | |
| 581 | return btrfs_search_prev_slot(trans, root, &key, p, 0, 1); |
| 582 | } |
| 583 | |
| 584 | /* |
| 585 | * If remove is 1, then we are removing free space, thus clearing bits in the |
| 586 | * bitmap. If remove is 0, then we are adding free space, thus setting bits in |
| 587 | * the bitmap. |
| 588 | */ |
| 589 | static int modify_free_space_bitmap(struct btrfs_trans_handle *trans, |
| 590 | struct btrfs_fs_info *fs_info, |
| 591 | struct btrfs_block_group_cache *block_group, |
| 592 | struct btrfs_path *path, |
| 593 | u64 start, u64 size, int remove) |
| 594 | { |
| 595 | struct btrfs_root *root = fs_info->free_space_root; |
| 596 | struct btrfs_key key; |
| 597 | u64 end = start + size; |
| 598 | u64 cur_start, cur_size; |
| 599 | int prev_bit, next_bit; |
| 600 | int new_extents; |
| 601 | int ret; |
| 602 | |
| 603 | /* |
| 604 | * Read the bit for the block immediately before the extent of space if |
| 605 | * that block is within the block group. |
| 606 | */ |
| 607 | if (start > block_group->key.objectid) { |
| 608 | u64 prev_block = start - block_group->sectorsize; |
| 609 | |
| 610 | key.objectid = prev_block; |
| 611 | key.type = (u8)-1; |
| 612 | key.offset = (u64)-1; |
| 613 | |
| 614 | ret = btrfs_search_prev_slot(trans, root, &key, path, 0, 1); |
| 615 | if (ret) |
| 616 | goto out; |
| 617 | |
| 618 | prev_bit = free_space_test_bit(block_group, path, prev_block); |
| 619 | |
| 620 | /* The previous block may have been in the previous bitmap. */ |
| 621 | btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); |
| 622 | if (start >= key.objectid + key.offset) { |
| 623 | ret = free_space_next_bitmap(trans, root, path); |
| 624 | if (ret) |
| 625 | goto out; |
| 626 | } |
| 627 | } else { |
| 628 | key.objectid = start; |
| 629 | key.type = (u8)-1; |
| 630 | key.offset = (u64)-1; |
| 631 | |
| 632 | ret = btrfs_search_prev_slot(trans, root, &key, path, 0, 1); |
| 633 | if (ret) |
| 634 | goto out; |
| 635 | |
| 636 | prev_bit = -1; |
| 637 | } |
| 638 | |
| 639 | /* |
| 640 | * Iterate over all of the bitmaps overlapped by the extent of space, |
| 641 | * clearing/setting bits as required. |
| 642 | */ |
| 643 | cur_start = start; |
| 644 | cur_size = size; |
| 645 | while (1) { |
| 646 | free_space_set_bits(block_group, path, &cur_start, &cur_size, |
| 647 | !remove); |
| 648 | if (cur_size == 0) |
| 649 | break; |
| 650 | ret = free_space_next_bitmap(trans, root, path); |
| 651 | if (ret) |
| 652 | goto out; |
| 653 | } |
| 654 | |
| 655 | /* |
| 656 | * Read the bit for the block immediately after the extent of space if |
| 657 | * that block is within the block group. |
| 658 | */ |
| 659 | if (end < block_group->key.objectid + block_group->key.offset) { |
| 660 | /* The next block may be in the next bitmap. */ |
| 661 | btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); |
| 662 | if (end >= key.objectid + key.offset) { |
| 663 | ret = free_space_next_bitmap(trans, root, path); |
| 664 | if (ret) |
| 665 | goto out; |
| 666 | } |
| 667 | |
| 668 | next_bit = free_space_test_bit(block_group, path, end); |
| 669 | } else { |
| 670 | next_bit = -1; |
| 671 | } |
| 672 | |
| 673 | if (remove) { |
| 674 | new_extents = -1; |
| 675 | if (prev_bit == 1) { |
| 676 | /* Leftover on the left. */ |
| 677 | new_extents++; |
| 678 | } |
| 679 | if (next_bit == 1) { |
| 680 | /* Leftover on the right. */ |
| 681 | new_extents++; |
| 682 | } |
| 683 | } else { |
| 684 | new_extents = 1; |
| 685 | if (prev_bit == 1) { |
| 686 | /* Merging with neighbor on the left. */ |
| 687 | new_extents--; |
| 688 | } |
| 689 | if (next_bit == 1) { |
| 690 | /* Merging with neighbor on the right. */ |
| 691 | new_extents--; |
| 692 | } |
| 693 | } |
| 694 | |
| 695 | btrfs_release_path(path); |
| 696 | ret = update_free_space_extent_count(trans, fs_info, block_group, path, |
| 697 | new_extents); |
| 698 | |
| 699 | out: |
| 700 | return ret; |
| 701 | } |
| 702 | |
| 703 | static int remove_free_space_extent(struct btrfs_trans_handle *trans, |
| 704 | struct btrfs_fs_info *fs_info, |
| 705 | struct btrfs_block_group_cache *block_group, |
| 706 | struct btrfs_path *path, |
| 707 | u64 start, u64 size) |
| 708 | { |
| 709 | struct btrfs_root *root = fs_info->free_space_root; |
| 710 | struct btrfs_key key; |
| 711 | u64 found_start, found_end; |
| 712 | u64 end = start + size; |
| 713 | int new_extents = -1; |
| 714 | int ret; |
| 715 | |
| 716 | key.objectid = start; |
| 717 | key.type = (u8)-1; |
| 718 | key.offset = (u64)-1; |
| 719 | |
| 720 | ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1); |
| 721 | if (ret) |
| 722 | goto out; |
| 723 | |
| 724 | btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); |
| 725 | |
| 726 | ASSERT(key.type == BTRFS_FREE_SPACE_EXTENT_KEY); |
| 727 | |
| 728 | found_start = key.objectid; |
| 729 | found_end = key.objectid + key.offset; |
| 730 | ASSERT(start >= found_start && end <= found_end); |
| 731 | |
| 732 | /* |
| 733 | * Okay, now that we've found the free space extent which contains the |
| 734 | * free space that we are removing, there are four cases: |
| 735 | * |
| 736 | * 1. We're using the whole extent: delete the key we found and |
| 737 | * decrement the free space extent count. |
| 738 | * 2. We are using part of the extent starting at the beginning: delete |
| 739 | * the key we found and insert a new key representing the leftover at |
| 740 | * the end. There is no net change in the number of extents. |
| 741 | * 3. We are using part of the extent ending at the end: delete the key |
| 742 | * we found and insert a new key representing the leftover at the |
| 743 | * beginning. There is no net change in the number of extents. |
| 744 | * 4. We are using part of the extent in the middle: delete the key we |
| 745 | * found and insert two new keys representing the leftovers on each |
| 746 | * side. Where we used to have one extent, we now have two, so increment |
| 747 | * the extent count. We may need to convert the block group to bitmaps |
| 748 | * as a result. |
| 749 | */ |
| 750 | |
| 751 | /* Delete the existing key (cases 1-4). */ |
| 752 | ret = btrfs_del_item(trans, root, path); |
| 753 | if (ret) |
| 754 | goto out; |
| 755 | |
| 756 | /* Add a key for leftovers at the beginning (cases 3 and 4). */ |
| 757 | if (start > found_start) { |
| 758 | key.objectid = found_start; |
| 759 | key.type = BTRFS_FREE_SPACE_EXTENT_KEY; |
| 760 | key.offset = start - found_start; |
| 761 | |
| 762 | btrfs_release_path(path); |
| 763 | ret = btrfs_insert_empty_item(trans, root, path, &key, 0); |
| 764 | if (ret) |
| 765 | goto out; |
| 766 | new_extents++; |
| 767 | } |
| 768 | |
| 769 | /* Add a key for leftovers at the end (cases 2 and 4). */ |
| 770 | if (end < found_end) { |
| 771 | key.objectid = end; |
| 772 | key.type = BTRFS_FREE_SPACE_EXTENT_KEY; |
| 773 | key.offset = found_end - end; |
| 774 | |
| 775 | btrfs_release_path(path); |
| 776 | ret = btrfs_insert_empty_item(trans, root, path, &key, 0); |
| 777 | if (ret) |
| 778 | goto out; |
| 779 | new_extents++; |
| 780 | } |
| 781 | |
| 782 | btrfs_release_path(path); |
| 783 | ret = update_free_space_extent_count(trans, fs_info, block_group, path, |
| 784 | new_extents); |
| 785 | |
| 786 | out: |
| 787 | return ret; |
| 788 | } |
| 789 | |
| 790 | int __remove_from_free_space_tree(struct btrfs_trans_handle *trans, |
| 791 | struct btrfs_fs_info *fs_info, |
| 792 | struct btrfs_block_group_cache *block_group, |
| 793 | struct btrfs_path *path, u64 start, u64 size) |
| 794 | { |
| 795 | struct btrfs_free_space_info *info; |
| 796 | u32 flags; |
| 797 | int ret; |
| 798 | |
| 799 | if (block_group->needs_free_space) { |
| 800 | ret = __add_block_group_free_space(trans, fs_info, block_group, |
| 801 | path); |
| 802 | if (ret) |
| 803 | return ret; |
| 804 | } |
| 805 | |
| 806 | info = search_free_space_info(NULL, fs_info, block_group, path, 0); |
| 807 | if (IS_ERR(info)) |
| 808 | return PTR_ERR(info); |
| 809 | flags = btrfs_free_space_flags(path->nodes[0], info); |
| 810 | btrfs_release_path(path); |
| 811 | |
| 812 | if (flags & BTRFS_FREE_SPACE_USING_BITMAPS) { |
| 813 | return modify_free_space_bitmap(trans, fs_info, block_group, |
| 814 | path, start, size, 1); |
| 815 | } else { |
| 816 | return remove_free_space_extent(trans, fs_info, block_group, |
| 817 | path, start, size); |
| 818 | } |
| 819 | } |
| 820 | |
| 821 | int remove_from_free_space_tree(struct btrfs_trans_handle *trans, |
| 822 | struct btrfs_fs_info *fs_info, |
| 823 | u64 start, u64 size) |
| 824 | { |
| 825 | struct btrfs_block_group_cache *block_group; |
| 826 | struct btrfs_path *path; |
| 827 | int ret; |
| 828 | |
| 829 | if (!btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE)) |
| 830 | return 0; |
| 831 | |
| 832 | path = btrfs_alloc_path(); |
| 833 | if (!path) { |
| 834 | ret = -ENOMEM; |
| 835 | goto out; |
| 836 | } |
| 837 | |
| 838 | block_group = btrfs_lookup_block_group(fs_info, start); |
| 839 | if (!block_group) { |
| 840 | ASSERT(0); |
| 841 | ret = -ENOENT; |
| 842 | goto out; |
| 843 | } |
| 844 | |
| 845 | mutex_lock(&block_group->free_space_lock); |
| 846 | ret = __remove_from_free_space_tree(trans, fs_info, block_group, path, |
| 847 | start, size); |
| 848 | mutex_unlock(&block_group->free_space_lock); |
| 849 | |
| 850 | btrfs_put_block_group(block_group); |
| 851 | out: |
| 852 | btrfs_free_path(path); |
| 853 | if (ret) |
| 854 | btrfs_abort_transaction(trans, fs_info->free_space_root, ret); |
| 855 | return ret; |
| 856 | } |
| 857 | |
| 858 | static int add_free_space_extent(struct btrfs_trans_handle *trans, |
| 859 | struct btrfs_fs_info *fs_info, |
| 860 | struct btrfs_block_group_cache *block_group, |
| 861 | struct btrfs_path *path, |
| 862 | u64 start, u64 size) |
| 863 | { |
| 864 | struct btrfs_root *root = fs_info->free_space_root; |
| 865 | struct btrfs_key key, new_key; |
| 866 | u64 found_start, found_end; |
| 867 | u64 end = start + size; |
| 868 | int new_extents = 1; |
| 869 | int ret; |
| 870 | |
| 871 | /* |
| 872 | * We are adding a new extent of free space, but we need to merge |
| 873 | * extents. There are four cases here: |
| 874 | * |
| 875 | * 1. The new extent does not have any immediate neighbors to merge |
| 876 | * with: add the new key and increment the free space extent count. We |
| 877 | * may need to convert the block group to bitmaps as a result. |
| 878 | * 2. The new extent has an immediate neighbor before it: remove the |
| 879 | * previous key and insert a new key combining both of them. There is no |
| 880 | * net change in the number of extents. |
| 881 | * 3. The new extent has an immediate neighbor after it: remove the next |
| 882 | * key and insert a new key combining both of them. There is no net |
| 883 | * change in the number of extents. |
| 884 | * 4. The new extent has immediate neighbors on both sides: remove both |
| 885 | * of the keys and insert a new key combining all of them. Where we used |
| 886 | * to have two extents, we now have one, so decrement the extent count. |
| 887 | */ |
| 888 | |
| 889 | new_key.objectid = start; |
| 890 | new_key.type = BTRFS_FREE_SPACE_EXTENT_KEY; |
| 891 | new_key.offset = size; |
| 892 | |
| 893 | /* Search for a neighbor on the left. */ |
| 894 | if (start == block_group->key.objectid) |
| 895 | goto right; |
| 896 | key.objectid = start - 1; |
| 897 | key.type = (u8)-1; |
| 898 | key.offset = (u64)-1; |
| 899 | |
| 900 | ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1); |
| 901 | if (ret) |
| 902 | goto out; |
| 903 | |
| 904 | btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); |
| 905 | |
| 906 | if (key.type != BTRFS_FREE_SPACE_EXTENT_KEY) { |
| 907 | ASSERT(key.type == BTRFS_FREE_SPACE_INFO_KEY); |
| 908 | btrfs_release_path(path); |
| 909 | goto right; |
| 910 | } |
| 911 | |
| 912 | found_start = key.objectid; |
| 913 | found_end = key.objectid + key.offset; |
| 914 | ASSERT(found_start >= block_group->key.objectid && |
| 915 | found_end > block_group->key.objectid); |
| 916 | ASSERT(found_start < start && found_end <= start); |
| 917 | |
| 918 | /* |
| 919 | * Delete the neighbor on the left and absorb it into the new key (cases |
| 920 | * 2 and 4). |
| 921 | */ |
| 922 | if (found_end == start) { |
| 923 | ret = btrfs_del_item(trans, root, path); |
| 924 | if (ret) |
| 925 | goto out; |
| 926 | new_key.objectid = found_start; |
| 927 | new_key.offset += key.offset; |
| 928 | new_extents--; |
| 929 | } |
| 930 | btrfs_release_path(path); |
| 931 | |
| 932 | right: |
| 933 | /* Search for a neighbor on the right. */ |
| 934 | if (end == block_group->key.objectid + block_group->key.offset) |
| 935 | goto insert; |
| 936 | key.objectid = end; |
| 937 | key.type = (u8)-1; |
| 938 | key.offset = (u64)-1; |
| 939 | |
| 940 | ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1); |
| 941 | if (ret) |
| 942 | goto out; |
| 943 | |
| 944 | btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); |
| 945 | |
| 946 | if (key.type != BTRFS_FREE_SPACE_EXTENT_KEY) { |
| 947 | ASSERT(key.type == BTRFS_FREE_SPACE_INFO_KEY); |
| 948 | btrfs_release_path(path); |
| 949 | goto insert; |
| 950 | } |
| 951 | |
| 952 | found_start = key.objectid; |
| 953 | found_end = key.objectid + key.offset; |
| 954 | ASSERT(found_start >= block_group->key.objectid && |
| 955 | found_end > block_group->key.objectid); |
| 956 | ASSERT((found_start < start && found_end <= start) || |
| 957 | (found_start >= end && found_end > end)); |
| 958 | |
| 959 | /* |
| 960 | * Delete the neighbor on the right and absorb it into the new key |
| 961 | * (cases 3 and 4). |
| 962 | */ |
| 963 | if (found_start == end) { |
| 964 | ret = btrfs_del_item(trans, root, path); |
| 965 | if (ret) |
| 966 | goto out; |
| 967 | new_key.offset += key.offset; |
| 968 | new_extents--; |
| 969 | } |
| 970 | btrfs_release_path(path); |
| 971 | |
| 972 | insert: |
| 973 | /* Insert the new key (cases 1-4). */ |
| 974 | ret = btrfs_insert_empty_item(trans, root, path, &new_key, 0); |
| 975 | if (ret) |
| 976 | goto out; |
| 977 | |
| 978 | btrfs_release_path(path); |
| 979 | ret = update_free_space_extent_count(trans, fs_info, block_group, path, |
| 980 | new_extents); |
| 981 | |
| 982 | out: |
| 983 | return ret; |
| 984 | } |
| 985 | |
| 986 | int __add_to_free_space_tree(struct btrfs_trans_handle *trans, |
| 987 | struct btrfs_fs_info *fs_info, |
| 988 | struct btrfs_block_group_cache *block_group, |
| 989 | struct btrfs_path *path, u64 start, u64 size) |
| 990 | { |
| 991 | struct btrfs_free_space_info *info; |
| 992 | u32 flags; |
| 993 | int ret; |
| 994 | |
| 995 | if (block_group->needs_free_space) { |
| 996 | ret = __add_block_group_free_space(trans, fs_info, block_group, |
| 997 | path); |
| 998 | if (ret) |
| 999 | return ret; |
| 1000 | } |
| 1001 | |
| 1002 | info = search_free_space_info(NULL, fs_info, block_group, path, 0); |
| 1003 | if (IS_ERR(info)) |
| 1004 | return PTR_ERR(info); |
| 1005 | flags = btrfs_free_space_flags(path->nodes[0], info); |
| 1006 | btrfs_release_path(path); |
| 1007 | |
| 1008 | if (flags & BTRFS_FREE_SPACE_USING_BITMAPS) { |
| 1009 | return modify_free_space_bitmap(trans, fs_info, block_group, |
| 1010 | path, start, size, 0); |
| 1011 | } else { |
| 1012 | return add_free_space_extent(trans, fs_info, block_group, path, |
| 1013 | start, size); |
| 1014 | } |
| 1015 | } |
| 1016 | |
| 1017 | int add_to_free_space_tree(struct btrfs_trans_handle *trans, |
| 1018 | struct btrfs_fs_info *fs_info, |
| 1019 | u64 start, u64 size) |
| 1020 | { |
| 1021 | struct btrfs_block_group_cache *block_group; |
| 1022 | struct btrfs_path *path; |
| 1023 | int ret; |
| 1024 | |
| 1025 | if (!btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE)) |
| 1026 | return 0; |
| 1027 | |
| 1028 | path = btrfs_alloc_path(); |
| 1029 | if (!path) { |
| 1030 | ret = -ENOMEM; |
| 1031 | goto out; |
| 1032 | } |
| 1033 | |
| 1034 | block_group = btrfs_lookup_block_group(fs_info, start); |
| 1035 | if (!block_group) { |
| 1036 | ASSERT(0); |
| 1037 | ret = -ENOENT; |
| 1038 | goto out; |
| 1039 | } |
| 1040 | |
| 1041 | mutex_lock(&block_group->free_space_lock); |
| 1042 | ret = __add_to_free_space_tree(trans, fs_info, block_group, path, start, |
| 1043 | size); |
| 1044 | mutex_unlock(&block_group->free_space_lock); |
| 1045 | |
| 1046 | btrfs_put_block_group(block_group); |
| 1047 | out: |
| 1048 | btrfs_free_path(path); |
| 1049 | if (ret) |
| 1050 | btrfs_abort_transaction(trans, fs_info->free_space_root, ret); |
| 1051 | return ret; |
| 1052 | } |
| 1053 | |
| 1054 | /* |
| 1055 | * Populate the free space tree by walking the extent tree. Operations on the |
| 1056 | * extent tree that happen as a result of writes to the free space tree will go |
| 1057 | * through the normal add/remove hooks. |
| 1058 | */ |
| 1059 | static int populate_free_space_tree(struct btrfs_trans_handle *trans, |
| 1060 | struct btrfs_fs_info *fs_info, |
| 1061 | struct btrfs_block_group_cache *block_group) |
| 1062 | { |
| 1063 | struct btrfs_root *extent_root = fs_info->extent_root; |
| 1064 | struct btrfs_path *path, *path2; |
| 1065 | struct btrfs_key key; |
| 1066 | u64 start, end; |
| 1067 | int ret; |
| 1068 | |
| 1069 | path = btrfs_alloc_path(); |
| 1070 | if (!path) |
| 1071 | return -ENOMEM; |
| 1072 | path->reada = 1; |
| 1073 | |
| 1074 | path2 = btrfs_alloc_path(); |
| 1075 | if (!path2) { |
| 1076 | btrfs_free_path(path); |
| 1077 | return -ENOMEM; |
| 1078 | } |
| 1079 | |
| 1080 | ret = add_new_free_space_info(trans, fs_info, block_group, path2); |
| 1081 | if (ret) |
| 1082 | goto out; |
| 1083 | |
| 1084 | mutex_lock(&block_group->free_space_lock); |
| 1085 | |
| 1086 | /* |
| 1087 | * Iterate through all of the extent and metadata items in this block |
| 1088 | * group, adding the free space between them and the free space at the |
| 1089 | * end. Note that EXTENT_ITEM and METADATA_ITEM are less than |
| 1090 | * BLOCK_GROUP_ITEM, so an extent may precede the block group that it's |
| 1091 | * contained in. |
| 1092 | */ |
| 1093 | key.objectid = block_group->key.objectid; |
| 1094 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 1095 | key.offset = 0; |
| 1096 | |
| 1097 | ret = btrfs_search_slot_for_read(extent_root, &key, path, 1, 0); |
| 1098 | if (ret < 0) |
| 1099 | goto out_locked; |
| 1100 | ASSERT(ret == 0); |
| 1101 | |
| 1102 | start = block_group->key.objectid; |
| 1103 | end = block_group->key.objectid + block_group->key.offset; |
| 1104 | while (1) { |
| 1105 | btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); |
| 1106 | |
| 1107 | if (key.type == BTRFS_EXTENT_ITEM_KEY || |
| 1108 | key.type == BTRFS_METADATA_ITEM_KEY) { |
| 1109 | if (key.objectid >= end) |
| 1110 | break; |
| 1111 | |
| 1112 | if (start < key.objectid) { |
| 1113 | ret = __add_to_free_space_tree(trans, fs_info, |
| 1114 | block_group, |
| 1115 | path2, start, |
| 1116 | key.objectid - |
| 1117 | start); |
| 1118 | if (ret) |
| 1119 | goto out_locked; |
| 1120 | } |
| 1121 | start = key.objectid; |
| 1122 | if (key.type == BTRFS_METADATA_ITEM_KEY) |
| 1123 | start += fs_info->tree_root->nodesize; |
| 1124 | else |
| 1125 | start += key.offset; |
| 1126 | } else if (key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) { |
| 1127 | if (key.objectid != block_group->key.objectid) |
| 1128 | break; |
| 1129 | } |
| 1130 | |
| 1131 | ret = btrfs_next_item(extent_root, path); |
| 1132 | if (ret < 0) |
| 1133 | goto out_locked; |
| 1134 | if (ret) |
| 1135 | break; |
| 1136 | } |
| 1137 | if (start < end) { |
| 1138 | ret = __add_to_free_space_tree(trans, fs_info, block_group, |
| 1139 | path2, start, end - start); |
| 1140 | if (ret) |
| 1141 | goto out_locked; |
| 1142 | } |
| 1143 | |
| 1144 | ret = 0; |
| 1145 | out_locked: |
| 1146 | mutex_unlock(&block_group->free_space_lock); |
| 1147 | out: |
| 1148 | btrfs_free_path(path2); |
| 1149 | btrfs_free_path(path); |
| 1150 | return ret; |
| 1151 | } |
| 1152 | |
| 1153 | int btrfs_create_free_space_tree(struct btrfs_fs_info *fs_info) |
| 1154 | { |
| 1155 | struct btrfs_trans_handle *trans; |
| 1156 | struct btrfs_root *tree_root = fs_info->tree_root; |
| 1157 | struct btrfs_root *free_space_root; |
| 1158 | struct btrfs_block_group_cache *block_group; |
| 1159 | struct rb_node *node; |
| 1160 | int ret; |
| 1161 | |
| 1162 | trans = btrfs_start_transaction(tree_root, 0); |
| 1163 | if (IS_ERR(trans)) |
| 1164 | return PTR_ERR(trans); |
| 1165 | |
| 1166 | fs_info->creating_free_space_tree = 1; |
| 1167 | free_space_root = btrfs_create_tree(trans, fs_info, |
| 1168 | BTRFS_FREE_SPACE_TREE_OBJECTID); |
| 1169 | if (IS_ERR(free_space_root)) { |
| 1170 | ret = PTR_ERR(free_space_root); |
| 1171 | goto abort; |
| 1172 | } |
| 1173 | fs_info->free_space_root = free_space_root; |
| 1174 | |
| 1175 | node = rb_first(&fs_info->block_group_cache_tree); |
| 1176 | while (node) { |
| 1177 | block_group = rb_entry(node, struct btrfs_block_group_cache, |
| 1178 | cache_node); |
| 1179 | ret = populate_free_space_tree(trans, fs_info, block_group); |
| 1180 | if (ret) |
| 1181 | goto abort; |
| 1182 | node = rb_next(node); |
| 1183 | } |
| 1184 | |
| 1185 | btrfs_set_fs_compat_ro(fs_info, FREE_SPACE_TREE); |
| 1186 | fs_info->creating_free_space_tree = 0; |
| 1187 | |
| 1188 | ret = btrfs_commit_transaction(trans, tree_root); |
| 1189 | if (ret) |
| 1190 | return ret; |
| 1191 | |
| 1192 | return 0; |
| 1193 | |
| 1194 | abort: |
| 1195 | fs_info->creating_free_space_tree = 0; |
| 1196 | btrfs_abort_transaction(trans, tree_root, ret); |
| 1197 | btrfs_end_transaction(trans, tree_root); |
| 1198 | return ret; |
| 1199 | } |
| 1200 | |
| 1201 | static int clear_free_space_tree(struct btrfs_trans_handle *trans, |
| 1202 | struct btrfs_root *root) |
| 1203 | { |
| 1204 | struct btrfs_path *path; |
| 1205 | struct btrfs_key key; |
| 1206 | int nr; |
| 1207 | int ret; |
| 1208 | |
| 1209 | path = btrfs_alloc_path(); |
| 1210 | if (!path) |
| 1211 | return -ENOMEM; |
| 1212 | |
| 1213 | path->leave_spinning = 1; |
| 1214 | |
| 1215 | key.objectid = 0; |
| 1216 | key.type = 0; |
| 1217 | key.offset = 0; |
| 1218 | |
| 1219 | while (1) { |
| 1220 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); |
| 1221 | if (ret < 0) |
| 1222 | goto out; |
| 1223 | |
| 1224 | nr = btrfs_header_nritems(path->nodes[0]); |
| 1225 | if (!nr) |
| 1226 | break; |
| 1227 | |
| 1228 | path->slots[0] = 0; |
| 1229 | ret = btrfs_del_items(trans, root, path, 0, nr); |
| 1230 | if (ret) |
| 1231 | goto out; |
| 1232 | |
| 1233 | btrfs_release_path(path); |
| 1234 | } |
| 1235 | |
| 1236 | ret = 0; |
| 1237 | out: |
| 1238 | btrfs_free_path(path); |
| 1239 | return ret; |
| 1240 | } |
| 1241 | |
| 1242 | int btrfs_clear_free_space_tree(struct btrfs_fs_info *fs_info) |
| 1243 | { |
| 1244 | struct btrfs_trans_handle *trans; |
| 1245 | struct btrfs_root *tree_root = fs_info->tree_root; |
| 1246 | struct btrfs_root *free_space_root = fs_info->free_space_root; |
| 1247 | int ret; |
| 1248 | |
| 1249 | trans = btrfs_start_transaction(tree_root, 0); |
| 1250 | if (IS_ERR(trans)) |
| 1251 | return PTR_ERR(trans); |
| 1252 | |
| 1253 | btrfs_clear_fs_compat_ro(fs_info, FREE_SPACE_TREE); |
| 1254 | fs_info->free_space_root = NULL; |
| 1255 | |
| 1256 | ret = clear_free_space_tree(trans, free_space_root); |
| 1257 | if (ret) |
| 1258 | goto abort; |
| 1259 | |
| 1260 | ret = btrfs_del_root(trans, tree_root, &free_space_root->root_key); |
| 1261 | if (ret) |
| 1262 | goto abort; |
| 1263 | |
| 1264 | list_del(&free_space_root->dirty_list); |
| 1265 | |
| 1266 | btrfs_tree_lock(free_space_root->node); |
| 1267 | clean_tree_block(trans, tree_root->fs_info, free_space_root->node); |
| 1268 | btrfs_tree_unlock(free_space_root->node); |
| 1269 | btrfs_free_tree_block(trans, free_space_root, free_space_root->node, |
| 1270 | 0, 1); |
| 1271 | |
| 1272 | free_extent_buffer(free_space_root->node); |
| 1273 | free_extent_buffer(free_space_root->commit_root); |
| 1274 | kfree(free_space_root); |
| 1275 | |
| 1276 | ret = btrfs_commit_transaction(trans, tree_root); |
| 1277 | if (ret) |
| 1278 | return ret; |
| 1279 | |
| 1280 | return 0; |
| 1281 | |
| 1282 | abort: |
| 1283 | btrfs_abort_transaction(trans, tree_root, ret); |
| 1284 | btrfs_end_transaction(trans, tree_root); |
| 1285 | return ret; |
| 1286 | } |
| 1287 | |
| 1288 | static int __add_block_group_free_space(struct btrfs_trans_handle *trans, |
| 1289 | struct btrfs_fs_info *fs_info, |
| 1290 | struct btrfs_block_group_cache *block_group, |
| 1291 | struct btrfs_path *path) |
| 1292 | { |
| 1293 | u64 start, end; |
| 1294 | int ret; |
| 1295 | |
| 1296 | start = block_group->key.objectid; |
| 1297 | end = block_group->key.objectid + block_group->key.offset; |
| 1298 | |
| 1299 | block_group->needs_free_space = 0; |
| 1300 | |
| 1301 | ret = add_new_free_space_info(trans, fs_info, block_group, path); |
| 1302 | if (ret) |
| 1303 | return ret; |
| 1304 | |
| 1305 | return __add_to_free_space_tree(trans, fs_info, block_group, path, |
| 1306 | block_group->key.objectid, |
| 1307 | block_group->key.offset); |
| 1308 | } |
| 1309 | |
| 1310 | int add_block_group_free_space(struct btrfs_trans_handle *trans, |
| 1311 | struct btrfs_fs_info *fs_info, |
| 1312 | struct btrfs_block_group_cache *block_group) |
| 1313 | { |
| 1314 | struct btrfs_path *path = NULL; |
| 1315 | int ret = 0; |
| 1316 | |
| 1317 | if (!btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE)) |
| 1318 | return 0; |
| 1319 | |
| 1320 | mutex_lock(&block_group->free_space_lock); |
| 1321 | if (!block_group->needs_free_space) |
| 1322 | goto out; |
| 1323 | |
| 1324 | path = btrfs_alloc_path(); |
| 1325 | if (!path) { |
| 1326 | ret = -ENOMEM; |
| 1327 | goto out; |
| 1328 | } |
| 1329 | |
| 1330 | ret = __add_block_group_free_space(trans, fs_info, block_group, path); |
| 1331 | |
| 1332 | out: |
| 1333 | btrfs_free_path(path); |
| 1334 | mutex_unlock(&block_group->free_space_lock); |
| 1335 | if (ret) |
| 1336 | btrfs_abort_transaction(trans, fs_info->free_space_root, ret); |
| 1337 | return ret; |
| 1338 | } |
| 1339 | |
| 1340 | int remove_block_group_free_space(struct btrfs_trans_handle *trans, |
| 1341 | struct btrfs_fs_info *fs_info, |
| 1342 | struct btrfs_block_group_cache *block_group) |
| 1343 | { |
| 1344 | struct btrfs_root *root = fs_info->free_space_root; |
| 1345 | struct btrfs_path *path; |
| 1346 | struct btrfs_key key, found_key; |
| 1347 | struct extent_buffer *leaf; |
| 1348 | u64 start, end; |
| 1349 | int done = 0, nr; |
| 1350 | int ret; |
| 1351 | |
| 1352 | if (!btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE)) |
| 1353 | return 0; |
| 1354 | |
| 1355 | if (block_group->needs_free_space) { |
| 1356 | /* We never added this block group to the free space tree. */ |
| 1357 | return 0; |
| 1358 | } |
| 1359 | |
| 1360 | path = btrfs_alloc_path(); |
| 1361 | if (!path) { |
| 1362 | ret = -ENOMEM; |
| 1363 | goto out; |
| 1364 | } |
| 1365 | |
| 1366 | start = block_group->key.objectid; |
| 1367 | end = block_group->key.objectid + block_group->key.offset; |
| 1368 | |
| 1369 | key.objectid = end - 1; |
| 1370 | key.type = (u8)-1; |
| 1371 | key.offset = (u64)-1; |
| 1372 | |
| 1373 | while (!done) { |
| 1374 | ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1); |
| 1375 | if (ret) |
| 1376 | goto out; |
| 1377 | |
| 1378 | leaf = path->nodes[0]; |
| 1379 | nr = 0; |
| 1380 | path->slots[0]++; |
| 1381 | while (path->slots[0] > 0) { |
| 1382 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1); |
| 1383 | |
| 1384 | if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) { |
| 1385 | ASSERT(found_key.objectid == block_group->key.objectid); |
| 1386 | ASSERT(found_key.offset == block_group->key.offset); |
| 1387 | done = 1; |
| 1388 | nr++; |
| 1389 | path->slots[0]--; |
| 1390 | break; |
| 1391 | } else if (found_key.type == BTRFS_FREE_SPACE_EXTENT_KEY || |
| 1392 | found_key.type == BTRFS_FREE_SPACE_BITMAP_KEY) { |
| 1393 | ASSERT(found_key.objectid >= start); |
| 1394 | ASSERT(found_key.objectid < end); |
| 1395 | ASSERT(found_key.objectid + found_key.offset <= end); |
| 1396 | nr++; |
| 1397 | path->slots[0]--; |
| 1398 | } else { |
| 1399 | ASSERT(0); |
| 1400 | } |
| 1401 | } |
| 1402 | |
| 1403 | ret = btrfs_del_items(trans, root, path, path->slots[0], nr); |
| 1404 | if (ret) |
| 1405 | goto out; |
| 1406 | btrfs_release_path(path); |
| 1407 | } |
| 1408 | |
| 1409 | ret = 0; |
| 1410 | out: |
| 1411 | btrfs_free_path(path); |
| 1412 | if (ret) |
| 1413 | btrfs_abort_transaction(trans, root, ret); |
| 1414 | return ret; |
| 1415 | } |
| 1416 | |
| 1417 | static int load_free_space_bitmaps(struct btrfs_caching_control *caching_ctl, |
| 1418 | struct btrfs_path *path, |
| 1419 | u32 expected_extent_count) |
| 1420 | { |
| 1421 | struct btrfs_block_group_cache *block_group; |
| 1422 | struct btrfs_fs_info *fs_info; |
| 1423 | struct btrfs_root *root; |
| 1424 | struct btrfs_key key; |
| 1425 | int prev_bit = 0, bit; |
| 1426 | /* Initialize to silence GCC. */ |
| 1427 | u64 extent_start = 0; |
| 1428 | u64 end, offset; |
| 1429 | u64 total_found = 0; |
| 1430 | u32 extent_count = 0; |
| 1431 | int ret; |
| 1432 | |
| 1433 | block_group = caching_ctl->block_group; |
| 1434 | fs_info = block_group->fs_info; |
| 1435 | root = fs_info->free_space_root; |
| 1436 | |
| 1437 | end = block_group->key.objectid + block_group->key.offset; |
| 1438 | |
| 1439 | while (1) { |
| 1440 | ret = btrfs_next_item(root, path); |
| 1441 | if (ret < 0) |
| 1442 | goto out; |
| 1443 | if (ret) |
| 1444 | break; |
| 1445 | |
| 1446 | btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); |
| 1447 | |
| 1448 | if (key.type == BTRFS_FREE_SPACE_INFO_KEY) |
| 1449 | break; |
| 1450 | |
| 1451 | ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY); |
| 1452 | ASSERT(key.objectid < end && key.objectid + key.offset <= end); |
| 1453 | |
| 1454 | caching_ctl->progress = key.objectid; |
| 1455 | |
| 1456 | offset = key.objectid; |
| 1457 | while (offset < key.objectid + key.offset) { |
| 1458 | bit = free_space_test_bit(block_group, path, offset); |
| 1459 | if (prev_bit == 0 && bit == 1) { |
| 1460 | extent_start = offset; |
| 1461 | } else if (prev_bit == 1 && bit == 0) { |
| 1462 | total_found += add_new_free_space(block_group, |
| 1463 | fs_info, |
| 1464 | extent_start, |
| 1465 | offset); |
| 1466 | if (total_found > CACHING_CTL_WAKE_UP) { |
| 1467 | total_found = 0; |
| 1468 | wake_up(&caching_ctl->wait); |
| 1469 | } |
| 1470 | extent_count++; |
| 1471 | } |
| 1472 | prev_bit = bit; |
| 1473 | offset += block_group->sectorsize; |
| 1474 | } |
| 1475 | } |
| 1476 | if (prev_bit == 1) { |
| 1477 | total_found += add_new_free_space(block_group, fs_info, |
| 1478 | extent_start, end); |
| 1479 | extent_count++; |
| 1480 | } |
| 1481 | |
| 1482 | if (extent_count != expected_extent_count) { |
| 1483 | btrfs_err(fs_info, "incorrect extent count for %llu; counted %u, expected %u", |
| 1484 | block_group->key.objectid, extent_count, |
| 1485 | expected_extent_count); |
| 1486 | ASSERT(0); |
| 1487 | ret = -EIO; |
| 1488 | goto out; |
| 1489 | } |
| 1490 | |
| 1491 | caching_ctl->progress = (u64)-1; |
| 1492 | |
| 1493 | ret = 0; |
| 1494 | out: |
| 1495 | return ret; |
| 1496 | } |
| 1497 | |
| 1498 | static int load_free_space_extents(struct btrfs_caching_control *caching_ctl, |
| 1499 | struct btrfs_path *path, |
| 1500 | u32 expected_extent_count) |
| 1501 | { |
| 1502 | struct btrfs_block_group_cache *block_group; |
| 1503 | struct btrfs_fs_info *fs_info; |
| 1504 | struct btrfs_root *root; |
| 1505 | struct btrfs_key key; |
| 1506 | u64 end; |
| 1507 | u64 total_found = 0; |
| 1508 | u32 extent_count = 0; |
| 1509 | int ret; |
| 1510 | |
| 1511 | block_group = caching_ctl->block_group; |
| 1512 | fs_info = block_group->fs_info; |
| 1513 | root = fs_info->free_space_root; |
| 1514 | |
| 1515 | end = block_group->key.objectid + block_group->key.offset; |
| 1516 | |
| 1517 | while (1) { |
| 1518 | ret = btrfs_next_item(root, path); |
| 1519 | if (ret < 0) |
| 1520 | goto out; |
| 1521 | if (ret) |
| 1522 | break; |
| 1523 | |
| 1524 | btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); |
| 1525 | |
| 1526 | if (key.type == BTRFS_FREE_SPACE_INFO_KEY) |
| 1527 | break; |
| 1528 | |
| 1529 | ASSERT(key.type == BTRFS_FREE_SPACE_EXTENT_KEY); |
| 1530 | ASSERT(key.objectid < end && key.objectid + key.offset <= end); |
| 1531 | |
| 1532 | caching_ctl->progress = key.objectid; |
| 1533 | |
| 1534 | total_found += add_new_free_space(block_group, fs_info, |
| 1535 | key.objectid, |
| 1536 | key.objectid + key.offset); |
| 1537 | if (total_found > CACHING_CTL_WAKE_UP) { |
| 1538 | total_found = 0; |
| 1539 | wake_up(&caching_ctl->wait); |
| 1540 | } |
| 1541 | extent_count++; |
| 1542 | } |
| 1543 | |
| 1544 | if (extent_count != expected_extent_count) { |
| 1545 | btrfs_err(fs_info, "incorrect extent count for %llu; counted %u, expected %u", |
| 1546 | block_group->key.objectid, extent_count, |
| 1547 | expected_extent_count); |
| 1548 | ASSERT(0); |
| 1549 | ret = -EIO; |
| 1550 | goto out; |
| 1551 | } |
| 1552 | |
| 1553 | caching_ctl->progress = (u64)-1; |
| 1554 | |
| 1555 | ret = 0; |
| 1556 | out: |
| 1557 | return ret; |
| 1558 | } |
| 1559 | |
| 1560 | int load_free_space_tree(struct btrfs_caching_control *caching_ctl) |
| 1561 | { |
| 1562 | struct btrfs_block_group_cache *block_group; |
| 1563 | struct btrfs_fs_info *fs_info; |
| 1564 | struct btrfs_free_space_info *info; |
| 1565 | struct btrfs_path *path; |
| 1566 | u32 extent_count, flags; |
| 1567 | int ret; |
| 1568 | |
| 1569 | block_group = caching_ctl->block_group; |
| 1570 | fs_info = block_group->fs_info; |
| 1571 | |
| 1572 | path = btrfs_alloc_path(); |
| 1573 | if (!path) |
| 1574 | return -ENOMEM; |
| 1575 | |
| 1576 | /* |
| 1577 | * Just like caching_thread() doesn't want to deadlock on the extent |
| 1578 | * tree, we don't want to deadlock on the free space tree. |
| 1579 | */ |
| 1580 | path->skip_locking = 1; |
| 1581 | path->search_commit_root = 1; |
| 1582 | path->reada = 1; |
| 1583 | |
| 1584 | info = search_free_space_info(NULL, fs_info, block_group, path, 0); |
| 1585 | if (IS_ERR(info)) { |
| 1586 | ret = PTR_ERR(info); |
| 1587 | goto out; |
| 1588 | } |
| 1589 | extent_count = btrfs_free_space_extent_count(path->nodes[0], info); |
| 1590 | flags = btrfs_free_space_flags(path->nodes[0], info); |
| 1591 | |
| 1592 | /* |
| 1593 | * We left path pointing to the free space info item, so now |
| 1594 | * load_free_space_foo can just iterate through the free space tree from |
| 1595 | * there. |
| 1596 | */ |
| 1597 | if (flags & BTRFS_FREE_SPACE_USING_BITMAPS) |
| 1598 | ret = load_free_space_bitmaps(caching_ctl, path, extent_count); |
| 1599 | else |
| 1600 | ret = load_free_space_extents(caching_ctl, path, extent_count); |
| 1601 | |
| 1602 | out: |
| 1603 | btrfs_free_path(path); |
| 1604 | return ret; |
| 1605 | } |