| 1 | /* |
| 2 | * Copyright (C) 2008 Red Hat. 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/pagemap.h> |
| 20 | #include <linux/sched.h> |
| 21 | #include <linux/slab.h> |
| 22 | #include <linux/math64.h> |
| 23 | #include <linux/ratelimit.h> |
| 24 | #include "ctree.h" |
| 25 | #include "free-space-cache.h" |
| 26 | #include "transaction.h" |
| 27 | #include "disk-io.h" |
| 28 | #include "extent_io.h" |
| 29 | #include "inode-map.h" |
| 30 | |
| 31 | #define BITS_PER_BITMAP (PAGE_CACHE_SIZE * 8) |
| 32 | #define MAX_CACHE_BYTES_PER_GIG (32 * 1024) |
| 33 | |
| 34 | static int link_free_space(struct btrfs_free_space_ctl *ctl, |
| 35 | struct btrfs_free_space *info); |
| 36 | static void unlink_free_space(struct btrfs_free_space_ctl *ctl, |
| 37 | struct btrfs_free_space *info); |
| 38 | |
| 39 | static struct inode *__lookup_free_space_inode(struct btrfs_root *root, |
| 40 | struct btrfs_path *path, |
| 41 | u64 offset) |
| 42 | { |
| 43 | struct btrfs_key key; |
| 44 | struct btrfs_key location; |
| 45 | struct btrfs_disk_key disk_key; |
| 46 | struct btrfs_free_space_header *header; |
| 47 | struct extent_buffer *leaf; |
| 48 | struct inode *inode = NULL; |
| 49 | int ret; |
| 50 | |
| 51 | key.objectid = BTRFS_FREE_SPACE_OBJECTID; |
| 52 | key.offset = offset; |
| 53 | key.type = 0; |
| 54 | |
| 55 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 56 | if (ret < 0) |
| 57 | return ERR_PTR(ret); |
| 58 | if (ret > 0) { |
| 59 | btrfs_release_path(path); |
| 60 | return ERR_PTR(-ENOENT); |
| 61 | } |
| 62 | |
| 63 | leaf = path->nodes[0]; |
| 64 | header = btrfs_item_ptr(leaf, path->slots[0], |
| 65 | struct btrfs_free_space_header); |
| 66 | btrfs_free_space_key(leaf, header, &disk_key); |
| 67 | btrfs_disk_key_to_cpu(&location, &disk_key); |
| 68 | btrfs_release_path(path); |
| 69 | |
| 70 | inode = btrfs_iget(root->fs_info->sb, &location, root, NULL); |
| 71 | if (!inode) |
| 72 | return ERR_PTR(-ENOENT); |
| 73 | if (IS_ERR(inode)) |
| 74 | return inode; |
| 75 | if (is_bad_inode(inode)) { |
| 76 | iput(inode); |
| 77 | return ERR_PTR(-ENOENT); |
| 78 | } |
| 79 | |
| 80 | mapping_set_gfp_mask(inode->i_mapping, |
| 81 | mapping_gfp_mask(inode->i_mapping) & ~__GFP_FS); |
| 82 | |
| 83 | return inode; |
| 84 | } |
| 85 | |
| 86 | struct inode *lookup_free_space_inode(struct btrfs_root *root, |
| 87 | struct btrfs_block_group_cache |
| 88 | *block_group, struct btrfs_path *path) |
| 89 | { |
| 90 | struct inode *inode = NULL; |
| 91 | u32 flags = BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW; |
| 92 | |
| 93 | spin_lock(&block_group->lock); |
| 94 | if (block_group->inode) |
| 95 | inode = igrab(block_group->inode); |
| 96 | spin_unlock(&block_group->lock); |
| 97 | if (inode) |
| 98 | return inode; |
| 99 | |
| 100 | inode = __lookup_free_space_inode(root, path, |
| 101 | block_group->key.objectid); |
| 102 | if (IS_ERR(inode)) |
| 103 | return inode; |
| 104 | |
| 105 | spin_lock(&block_group->lock); |
| 106 | if (!((BTRFS_I(inode)->flags & flags) == flags)) { |
| 107 | btrfs_info(root->fs_info, |
| 108 | "Old style space inode found, converting."); |
| 109 | BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM | |
| 110 | BTRFS_INODE_NODATACOW; |
| 111 | block_group->disk_cache_state = BTRFS_DC_CLEAR; |
| 112 | } |
| 113 | |
| 114 | if (!block_group->iref) { |
| 115 | block_group->inode = igrab(inode); |
| 116 | block_group->iref = 1; |
| 117 | } |
| 118 | spin_unlock(&block_group->lock); |
| 119 | |
| 120 | return inode; |
| 121 | } |
| 122 | |
| 123 | static int __create_free_space_inode(struct btrfs_root *root, |
| 124 | struct btrfs_trans_handle *trans, |
| 125 | struct btrfs_path *path, |
| 126 | u64 ino, u64 offset) |
| 127 | { |
| 128 | struct btrfs_key key; |
| 129 | struct btrfs_disk_key disk_key; |
| 130 | struct btrfs_free_space_header *header; |
| 131 | struct btrfs_inode_item *inode_item; |
| 132 | struct extent_buffer *leaf; |
| 133 | u64 flags = BTRFS_INODE_NOCOMPRESS | BTRFS_INODE_PREALLOC; |
| 134 | int ret; |
| 135 | |
| 136 | ret = btrfs_insert_empty_inode(trans, root, path, ino); |
| 137 | if (ret) |
| 138 | return ret; |
| 139 | |
| 140 | /* We inline crc's for the free disk space cache */ |
| 141 | if (ino != BTRFS_FREE_INO_OBJECTID) |
| 142 | flags |= BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW; |
| 143 | |
| 144 | leaf = path->nodes[0]; |
| 145 | inode_item = btrfs_item_ptr(leaf, path->slots[0], |
| 146 | struct btrfs_inode_item); |
| 147 | btrfs_item_key(leaf, &disk_key, path->slots[0]); |
| 148 | memset_extent_buffer(leaf, 0, (unsigned long)inode_item, |
| 149 | sizeof(*inode_item)); |
| 150 | btrfs_set_inode_generation(leaf, inode_item, trans->transid); |
| 151 | btrfs_set_inode_size(leaf, inode_item, 0); |
| 152 | btrfs_set_inode_nbytes(leaf, inode_item, 0); |
| 153 | btrfs_set_inode_uid(leaf, inode_item, 0); |
| 154 | btrfs_set_inode_gid(leaf, inode_item, 0); |
| 155 | btrfs_set_inode_mode(leaf, inode_item, S_IFREG | 0600); |
| 156 | btrfs_set_inode_flags(leaf, inode_item, flags); |
| 157 | btrfs_set_inode_nlink(leaf, inode_item, 1); |
| 158 | btrfs_set_inode_transid(leaf, inode_item, trans->transid); |
| 159 | btrfs_set_inode_block_group(leaf, inode_item, offset); |
| 160 | btrfs_mark_buffer_dirty(leaf); |
| 161 | btrfs_release_path(path); |
| 162 | |
| 163 | key.objectid = BTRFS_FREE_SPACE_OBJECTID; |
| 164 | key.offset = offset; |
| 165 | key.type = 0; |
| 166 | |
| 167 | ret = btrfs_insert_empty_item(trans, root, path, &key, |
| 168 | sizeof(struct btrfs_free_space_header)); |
| 169 | if (ret < 0) { |
| 170 | btrfs_release_path(path); |
| 171 | return ret; |
| 172 | } |
| 173 | leaf = path->nodes[0]; |
| 174 | header = btrfs_item_ptr(leaf, path->slots[0], |
| 175 | struct btrfs_free_space_header); |
| 176 | memset_extent_buffer(leaf, 0, (unsigned long)header, sizeof(*header)); |
| 177 | btrfs_set_free_space_key(leaf, header, &disk_key); |
| 178 | btrfs_mark_buffer_dirty(leaf); |
| 179 | btrfs_release_path(path); |
| 180 | |
| 181 | return 0; |
| 182 | } |
| 183 | |
| 184 | int create_free_space_inode(struct btrfs_root *root, |
| 185 | struct btrfs_trans_handle *trans, |
| 186 | struct btrfs_block_group_cache *block_group, |
| 187 | struct btrfs_path *path) |
| 188 | { |
| 189 | int ret; |
| 190 | u64 ino; |
| 191 | |
| 192 | ret = btrfs_find_free_objectid(root, &ino); |
| 193 | if (ret < 0) |
| 194 | return ret; |
| 195 | |
| 196 | return __create_free_space_inode(root, trans, path, ino, |
| 197 | block_group->key.objectid); |
| 198 | } |
| 199 | |
| 200 | int btrfs_check_trunc_cache_free_space(struct btrfs_root *root, |
| 201 | struct btrfs_block_rsv *rsv) |
| 202 | { |
| 203 | u64 needed_bytes; |
| 204 | int ret; |
| 205 | |
| 206 | /* 1 for slack space, 1 for updating the inode */ |
| 207 | needed_bytes = btrfs_calc_trunc_metadata_size(root, 1) + |
| 208 | btrfs_calc_trans_metadata_size(root, 1); |
| 209 | |
| 210 | spin_lock(&rsv->lock); |
| 211 | if (rsv->reserved < needed_bytes) |
| 212 | ret = -ENOSPC; |
| 213 | else |
| 214 | ret = 0; |
| 215 | spin_unlock(&rsv->lock); |
| 216 | return ret; |
| 217 | } |
| 218 | |
| 219 | int btrfs_truncate_free_space_cache(struct btrfs_root *root, |
| 220 | struct btrfs_trans_handle *trans, |
| 221 | struct inode *inode) |
| 222 | { |
| 223 | int ret = 0; |
| 224 | |
| 225 | btrfs_i_size_write(inode, 0); |
| 226 | truncate_pagecache(inode, 0); |
| 227 | |
| 228 | /* |
| 229 | * We don't need an orphan item because truncating the free space cache |
| 230 | * will never be split across transactions. |
| 231 | */ |
| 232 | ret = btrfs_truncate_inode_items(trans, root, inode, |
| 233 | 0, BTRFS_EXTENT_DATA_KEY); |
| 234 | if (ret) { |
| 235 | btrfs_abort_transaction(trans, root, ret); |
| 236 | return ret; |
| 237 | } |
| 238 | |
| 239 | ret = btrfs_update_inode(trans, root, inode); |
| 240 | if (ret) |
| 241 | btrfs_abort_transaction(trans, root, ret); |
| 242 | |
| 243 | return ret; |
| 244 | } |
| 245 | |
| 246 | static int readahead_cache(struct inode *inode) |
| 247 | { |
| 248 | struct file_ra_state *ra; |
| 249 | unsigned long last_index; |
| 250 | |
| 251 | ra = kzalloc(sizeof(*ra), GFP_NOFS); |
| 252 | if (!ra) |
| 253 | return -ENOMEM; |
| 254 | |
| 255 | file_ra_state_init(ra, inode->i_mapping); |
| 256 | last_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT; |
| 257 | |
| 258 | page_cache_sync_readahead(inode->i_mapping, ra, NULL, 0, last_index); |
| 259 | |
| 260 | kfree(ra); |
| 261 | |
| 262 | return 0; |
| 263 | } |
| 264 | |
| 265 | struct io_ctl { |
| 266 | void *cur, *orig; |
| 267 | struct page *page; |
| 268 | struct page **pages; |
| 269 | struct btrfs_root *root; |
| 270 | unsigned long size; |
| 271 | int index; |
| 272 | int num_pages; |
| 273 | unsigned check_crcs:1; |
| 274 | }; |
| 275 | |
| 276 | static int io_ctl_init(struct io_ctl *io_ctl, struct inode *inode, |
| 277 | struct btrfs_root *root) |
| 278 | { |
| 279 | memset(io_ctl, 0, sizeof(struct io_ctl)); |
| 280 | io_ctl->num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> |
| 281 | PAGE_CACHE_SHIFT; |
| 282 | io_ctl->pages = kzalloc(sizeof(struct page *) * io_ctl->num_pages, |
| 283 | GFP_NOFS); |
| 284 | if (!io_ctl->pages) |
| 285 | return -ENOMEM; |
| 286 | io_ctl->root = root; |
| 287 | if (btrfs_ino(inode) != BTRFS_FREE_INO_OBJECTID) |
| 288 | io_ctl->check_crcs = 1; |
| 289 | return 0; |
| 290 | } |
| 291 | |
| 292 | static void io_ctl_free(struct io_ctl *io_ctl) |
| 293 | { |
| 294 | kfree(io_ctl->pages); |
| 295 | } |
| 296 | |
| 297 | static void io_ctl_unmap_page(struct io_ctl *io_ctl) |
| 298 | { |
| 299 | if (io_ctl->cur) { |
| 300 | kunmap(io_ctl->page); |
| 301 | io_ctl->cur = NULL; |
| 302 | io_ctl->orig = NULL; |
| 303 | } |
| 304 | } |
| 305 | |
| 306 | static void io_ctl_map_page(struct io_ctl *io_ctl, int clear) |
| 307 | { |
| 308 | ASSERT(io_ctl->index < io_ctl->num_pages); |
| 309 | io_ctl->page = io_ctl->pages[io_ctl->index++]; |
| 310 | io_ctl->cur = kmap(io_ctl->page); |
| 311 | io_ctl->orig = io_ctl->cur; |
| 312 | io_ctl->size = PAGE_CACHE_SIZE; |
| 313 | if (clear) |
| 314 | memset(io_ctl->cur, 0, PAGE_CACHE_SIZE); |
| 315 | } |
| 316 | |
| 317 | static void io_ctl_drop_pages(struct io_ctl *io_ctl) |
| 318 | { |
| 319 | int i; |
| 320 | |
| 321 | io_ctl_unmap_page(io_ctl); |
| 322 | |
| 323 | for (i = 0; i < io_ctl->num_pages; i++) { |
| 324 | if (io_ctl->pages[i]) { |
| 325 | ClearPageChecked(io_ctl->pages[i]); |
| 326 | unlock_page(io_ctl->pages[i]); |
| 327 | page_cache_release(io_ctl->pages[i]); |
| 328 | } |
| 329 | } |
| 330 | } |
| 331 | |
| 332 | static int io_ctl_prepare_pages(struct io_ctl *io_ctl, struct inode *inode, |
| 333 | int uptodate) |
| 334 | { |
| 335 | struct page *page; |
| 336 | gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping); |
| 337 | int i; |
| 338 | |
| 339 | for (i = 0; i < io_ctl->num_pages; i++) { |
| 340 | page = find_or_create_page(inode->i_mapping, i, mask); |
| 341 | if (!page) { |
| 342 | io_ctl_drop_pages(io_ctl); |
| 343 | return -ENOMEM; |
| 344 | } |
| 345 | io_ctl->pages[i] = page; |
| 346 | if (uptodate && !PageUptodate(page)) { |
| 347 | btrfs_readpage(NULL, page); |
| 348 | lock_page(page); |
| 349 | if (!PageUptodate(page)) { |
| 350 | btrfs_err(BTRFS_I(inode)->root->fs_info, |
| 351 | "error reading free space cache"); |
| 352 | io_ctl_drop_pages(io_ctl); |
| 353 | return -EIO; |
| 354 | } |
| 355 | } |
| 356 | } |
| 357 | |
| 358 | for (i = 0; i < io_ctl->num_pages; i++) { |
| 359 | clear_page_dirty_for_io(io_ctl->pages[i]); |
| 360 | set_page_extent_mapped(io_ctl->pages[i]); |
| 361 | } |
| 362 | |
| 363 | return 0; |
| 364 | } |
| 365 | |
| 366 | static void io_ctl_set_generation(struct io_ctl *io_ctl, u64 generation) |
| 367 | { |
| 368 | __le64 *val; |
| 369 | |
| 370 | io_ctl_map_page(io_ctl, 1); |
| 371 | |
| 372 | /* |
| 373 | * Skip the csum areas. If we don't check crcs then we just have a |
| 374 | * 64bit chunk at the front of the first page. |
| 375 | */ |
| 376 | if (io_ctl->check_crcs) { |
| 377 | io_ctl->cur += (sizeof(u32) * io_ctl->num_pages); |
| 378 | io_ctl->size -= sizeof(u64) + (sizeof(u32) * io_ctl->num_pages); |
| 379 | } else { |
| 380 | io_ctl->cur += sizeof(u64); |
| 381 | io_ctl->size -= sizeof(u64) * 2; |
| 382 | } |
| 383 | |
| 384 | val = io_ctl->cur; |
| 385 | *val = cpu_to_le64(generation); |
| 386 | io_ctl->cur += sizeof(u64); |
| 387 | } |
| 388 | |
| 389 | static int io_ctl_check_generation(struct io_ctl *io_ctl, u64 generation) |
| 390 | { |
| 391 | __le64 *gen; |
| 392 | |
| 393 | /* |
| 394 | * Skip the crc area. If we don't check crcs then we just have a 64bit |
| 395 | * chunk at the front of the first page. |
| 396 | */ |
| 397 | if (io_ctl->check_crcs) { |
| 398 | io_ctl->cur += sizeof(u32) * io_ctl->num_pages; |
| 399 | io_ctl->size -= sizeof(u64) + |
| 400 | (sizeof(u32) * io_ctl->num_pages); |
| 401 | } else { |
| 402 | io_ctl->cur += sizeof(u64); |
| 403 | io_ctl->size -= sizeof(u64) * 2; |
| 404 | } |
| 405 | |
| 406 | gen = io_ctl->cur; |
| 407 | if (le64_to_cpu(*gen) != generation) { |
| 408 | printk_ratelimited(KERN_ERR "BTRFS: space cache generation " |
| 409 | "(%Lu) does not match inode (%Lu)\n", *gen, |
| 410 | generation); |
| 411 | io_ctl_unmap_page(io_ctl); |
| 412 | return -EIO; |
| 413 | } |
| 414 | io_ctl->cur += sizeof(u64); |
| 415 | return 0; |
| 416 | } |
| 417 | |
| 418 | static void io_ctl_set_crc(struct io_ctl *io_ctl, int index) |
| 419 | { |
| 420 | u32 *tmp; |
| 421 | u32 crc = ~(u32)0; |
| 422 | unsigned offset = 0; |
| 423 | |
| 424 | if (!io_ctl->check_crcs) { |
| 425 | io_ctl_unmap_page(io_ctl); |
| 426 | return; |
| 427 | } |
| 428 | |
| 429 | if (index == 0) |
| 430 | offset = sizeof(u32) * io_ctl->num_pages; |
| 431 | |
| 432 | crc = btrfs_csum_data(io_ctl->orig + offset, crc, |
| 433 | PAGE_CACHE_SIZE - offset); |
| 434 | btrfs_csum_final(crc, (char *)&crc); |
| 435 | io_ctl_unmap_page(io_ctl); |
| 436 | tmp = kmap(io_ctl->pages[0]); |
| 437 | tmp += index; |
| 438 | *tmp = crc; |
| 439 | kunmap(io_ctl->pages[0]); |
| 440 | } |
| 441 | |
| 442 | static int io_ctl_check_crc(struct io_ctl *io_ctl, int index) |
| 443 | { |
| 444 | u32 *tmp, val; |
| 445 | u32 crc = ~(u32)0; |
| 446 | unsigned offset = 0; |
| 447 | |
| 448 | if (!io_ctl->check_crcs) { |
| 449 | io_ctl_map_page(io_ctl, 0); |
| 450 | return 0; |
| 451 | } |
| 452 | |
| 453 | if (index == 0) |
| 454 | offset = sizeof(u32) * io_ctl->num_pages; |
| 455 | |
| 456 | tmp = kmap(io_ctl->pages[0]); |
| 457 | tmp += index; |
| 458 | val = *tmp; |
| 459 | kunmap(io_ctl->pages[0]); |
| 460 | |
| 461 | io_ctl_map_page(io_ctl, 0); |
| 462 | crc = btrfs_csum_data(io_ctl->orig + offset, crc, |
| 463 | PAGE_CACHE_SIZE - offset); |
| 464 | btrfs_csum_final(crc, (char *)&crc); |
| 465 | if (val != crc) { |
| 466 | printk_ratelimited(KERN_ERR "BTRFS: csum mismatch on free " |
| 467 | "space cache\n"); |
| 468 | io_ctl_unmap_page(io_ctl); |
| 469 | return -EIO; |
| 470 | } |
| 471 | |
| 472 | return 0; |
| 473 | } |
| 474 | |
| 475 | static int io_ctl_add_entry(struct io_ctl *io_ctl, u64 offset, u64 bytes, |
| 476 | void *bitmap) |
| 477 | { |
| 478 | struct btrfs_free_space_entry *entry; |
| 479 | |
| 480 | if (!io_ctl->cur) |
| 481 | return -ENOSPC; |
| 482 | |
| 483 | entry = io_ctl->cur; |
| 484 | entry->offset = cpu_to_le64(offset); |
| 485 | entry->bytes = cpu_to_le64(bytes); |
| 486 | entry->type = (bitmap) ? BTRFS_FREE_SPACE_BITMAP : |
| 487 | BTRFS_FREE_SPACE_EXTENT; |
| 488 | io_ctl->cur += sizeof(struct btrfs_free_space_entry); |
| 489 | io_ctl->size -= sizeof(struct btrfs_free_space_entry); |
| 490 | |
| 491 | if (io_ctl->size >= sizeof(struct btrfs_free_space_entry)) |
| 492 | return 0; |
| 493 | |
| 494 | io_ctl_set_crc(io_ctl, io_ctl->index - 1); |
| 495 | |
| 496 | /* No more pages to map */ |
| 497 | if (io_ctl->index >= io_ctl->num_pages) |
| 498 | return 0; |
| 499 | |
| 500 | /* map the next page */ |
| 501 | io_ctl_map_page(io_ctl, 1); |
| 502 | return 0; |
| 503 | } |
| 504 | |
| 505 | static int io_ctl_add_bitmap(struct io_ctl *io_ctl, void *bitmap) |
| 506 | { |
| 507 | if (!io_ctl->cur) |
| 508 | return -ENOSPC; |
| 509 | |
| 510 | /* |
| 511 | * If we aren't at the start of the current page, unmap this one and |
| 512 | * map the next one if there is any left. |
| 513 | */ |
| 514 | if (io_ctl->cur != io_ctl->orig) { |
| 515 | io_ctl_set_crc(io_ctl, io_ctl->index - 1); |
| 516 | if (io_ctl->index >= io_ctl->num_pages) |
| 517 | return -ENOSPC; |
| 518 | io_ctl_map_page(io_ctl, 0); |
| 519 | } |
| 520 | |
| 521 | memcpy(io_ctl->cur, bitmap, PAGE_CACHE_SIZE); |
| 522 | io_ctl_set_crc(io_ctl, io_ctl->index - 1); |
| 523 | if (io_ctl->index < io_ctl->num_pages) |
| 524 | io_ctl_map_page(io_ctl, 0); |
| 525 | return 0; |
| 526 | } |
| 527 | |
| 528 | static void io_ctl_zero_remaining_pages(struct io_ctl *io_ctl) |
| 529 | { |
| 530 | /* |
| 531 | * If we're not on the boundary we know we've modified the page and we |
| 532 | * need to crc the page. |
| 533 | */ |
| 534 | if (io_ctl->cur != io_ctl->orig) |
| 535 | io_ctl_set_crc(io_ctl, io_ctl->index - 1); |
| 536 | else |
| 537 | io_ctl_unmap_page(io_ctl); |
| 538 | |
| 539 | while (io_ctl->index < io_ctl->num_pages) { |
| 540 | io_ctl_map_page(io_ctl, 1); |
| 541 | io_ctl_set_crc(io_ctl, io_ctl->index - 1); |
| 542 | } |
| 543 | } |
| 544 | |
| 545 | static int io_ctl_read_entry(struct io_ctl *io_ctl, |
| 546 | struct btrfs_free_space *entry, u8 *type) |
| 547 | { |
| 548 | struct btrfs_free_space_entry *e; |
| 549 | int ret; |
| 550 | |
| 551 | if (!io_ctl->cur) { |
| 552 | ret = io_ctl_check_crc(io_ctl, io_ctl->index); |
| 553 | if (ret) |
| 554 | return ret; |
| 555 | } |
| 556 | |
| 557 | e = io_ctl->cur; |
| 558 | entry->offset = le64_to_cpu(e->offset); |
| 559 | entry->bytes = le64_to_cpu(e->bytes); |
| 560 | *type = e->type; |
| 561 | io_ctl->cur += sizeof(struct btrfs_free_space_entry); |
| 562 | io_ctl->size -= sizeof(struct btrfs_free_space_entry); |
| 563 | |
| 564 | if (io_ctl->size >= sizeof(struct btrfs_free_space_entry)) |
| 565 | return 0; |
| 566 | |
| 567 | io_ctl_unmap_page(io_ctl); |
| 568 | |
| 569 | return 0; |
| 570 | } |
| 571 | |
| 572 | static int io_ctl_read_bitmap(struct io_ctl *io_ctl, |
| 573 | struct btrfs_free_space *entry) |
| 574 | { |
| 575 | int ret; |
| 576 | |
| 577 | ret = io_ctl_check_crc(io_ctl, io_ctl->index); |
| 578 | if (ret) |
| 579 | return ret; |
| 580 | |
| 581 | memcpy(entry->bitmap, io_ctl->cur, PAGE_CACHE_SIZE); |
| 582 | io_ctl_unmap_page(io_ctl); |
| 583 | |
| 584 | return 0; |
| 585 | } |
| 586 | |
| 587 | /* |
| 588 | * Since we attach pinned extents after the fact we can have contiguous sections |
| 589 | * of free space that are split up in entries. This poses a problem with the |
| 590 | * tree logging stuff since it could have allocated across what appears to be 2 |
| 591 | * entries since we would have merged the entries when adding the pinned extents |
| 592 | * back to the free space cache. So run through the space cache that we just |
| 593 | * loaded and merge contiguous entries. This will make the log replay stuff not |
| 594 | * blow up and it will make for nicer allocator behavior. |
| 595 | */ |
| 596 | static void merge_space_tree(struct btrfs_free_space_ctl *ctl) |
| 597 | { |
| 598 | struct btrfs_free_space *e, *prev = NULL; |
| 599 | struct rb_node *n; |
| 600 | |
| 601 | again: |
| 602 | spin_lock(&ctl->tree_lock); |
| 603 | for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) { |
| 604 | e = rb_entry(n, struct btrfs_free_space, offset_index); |
| 605 | if (!prev) |
| 606 | goto next; |
| 607 | if (e->bitmap || prev->bitmap) |
| 608 | goto next; |
| 609 | if (prev->offset + prev->bytes == e->offset) { |
| 610 | unlink_free_space(ctl, prev); |
| 611 | unlink_free_space(ctl, e); |
| 612 | prev->bytes += e->bytes; |
| 613 | kmem_cache_free(btrfs_free_space_cachep, e); |
| 614 | link_free_space(ctl, prev); |
| 615 | prev = NULL; |
| 616 | spin_unlock(&ctl->tree_lock); |
| 617 | goto again; |
| 618 | } |
| 619 | next: |
| 620 | prev = e; |
| 621 | } |
| 622 | spin_unlock(&ctl->tree_lock); |
| 623 | } |
| 624 | |
| 625 | static int __load_free_space_cache(struct btrfs_root *root, struct inode *inode, |
| 626 | struct btrfs_free_space_ctl *ctl, |
| 627 | struct btrfs_path *path, u64 offset) |
| 628 | { |
| 629 | struct btrfs_free_space_header *header; |
| 630 | struct extent_buffer *leaf; |
| 631 | struct io_ctl io_ctl; |
| 632 | struct btrfs_key key; |
| 633 | struct btrfs_free_space *e, *n; |
| 634 | struct list_head bitmaps; |
| 635 | u64 num_entries; |
| 636 | u64 num_bitmaps; |
| 637 | u64 generation; |
| 638 | u8 type; |
| 639 | int ret = 0; |
| 640 | |
| 641 | INIT_LIST_HEAD(&bitmaps); |
| 642 | |
| 643 | /* Nothing in the space cache, goodbye */ |
| 644 | if (!i_size_read(inode)) |
| 645 | return 0; |
| 646 | |
| 647 | key.objectid = BTRFS_FREE_SPACE_OBJECTID; |
| 648 | key.offset = offset; |
| 649 | key.type = 0; |
| 650 | |
| 651 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 652 | if (ret < 0) |
| 653 | return 0; |
| 654 | else if (ret > 0) { |
| 655 | btrfs_release_path(path); |
| 656 | return 0; |
| 657 | } |
| 658 | |
| 659 | ret = -1; |
| 660 | |
| 661 | leaf = path->nodes[0]; |
| 662 | header = btrfs_item_ptr(leaf, path->slots[0], |
| 663 | struct btrfs_free_space_header); |
| 664 | num_entries = btrfs_free_space_entries(leaf, header); |
| 665 | num_bitmaps = btrfs_free_space_bitmaps(leaf, header); |
| 666 | generation = btrfs_free_space_generation(leaf, header); |
| 667 | btrfs_release_path(path); |
| 668 | |
| 669 | if (BTRFS_I(inode)->generation != generation) { |
| 670 | btrfs_err(root->fs_info, |
| 671 | "free space inode generation (%llu) " |
| 672 | "did not match free space cache generation (%llu)", |
| 673 | BTRFS_I(inode)->generation, generation); |
| 674 | return 0; |
| 675 | } |
| 676 | |
| 677 | if (!num_entries) |
| 678 | return 0; |
| 679 | |
| 680 | ret = io_ctl_init(&io_ctl, inode, root); |
| 681 | if (ret) |
| 682 | return ret; |
| 683 | |
| 684 | ret = readahead_cache(inode); |
| 685 | if (ret) |
| 686 | goto out; |
| 687 | |
| 688 | ret = io_ctl_prepare_pages(&io_ctl, inode, 1); |
| 689 | if (ret) |
| 690 | goto out; |
| 691 | |
| 692 | ret = io_ctl_check_crc(&io_ctl, 0); |
| 693 | if (ret) |
| 694 | goto free_cache; |
| 695 | |
| 696 | ret = io_ctl_check_generation(&io_ctl, generation); |
| 697 | if (ret) |
| 698 | goto free_cache; |
| 699 | |
| 700 | while (num_entries) { |
| 701 | e = kmem_cache_zalloc(btrfs_free_space_cachep, |
| 702 | GFP_NOFS); |
| 703 | if (!e) |
| 704 | goto free_cache; |
| 705 | |
| 706 | ret = io_ctl_read_entry(&io_ctl, e, &type); |
| 707 | if (ret) { |
| 708 | kmem_cache_free(btrfs_free_space_cachep, e); |
| 709 | goto free_cache; |
| 710 | } |
| 711 | |
| 712 | if (!e->bytes) { |
| 713 | kmem_cache_free(btrfs_free_space_cachep, e); |
| 714 | goto free_cache; |
| 715 | } |
| 716 | |
| 717 | if (type == BTRFS_FREE_SPACE_EXTENT) { |
| 718 | spin_lock(&ctl->tree_lock); |
| 719 | ret = link_free_space(ctl, e); |
| 720 | spin_unlock(&ctl->tree_lock); |
| 721 | if (ret) { |
| 722 | btrfs_err(root->fs_info, |
| 723 | "Duplicate entries in free space cache, dumping"); |
| 724 | kmem_cache_free(btrfs_free_space_cachep, e); |
| 725 | goto free_cache; |
| 726 | } |
| 727 | } else { |
| 728 | ASSERT(num_bitmaps); |
| 729 | num_bitmaps--; |
| 730 | e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); |
| 731 | if (!e->bitmap) { |
| 732 | kmem_cache_free( |
| 733 | btrfs_free_space_cachep, e); |
| 734 | goto free_cache; |
| 735 | } |
| 736 | spin_lock(&ctl->tree_lock); |
| 737 | ret = link_free_space(ctl, e); |
| 738 | ctl->total_bitmaps++; |
| 739 | ctl->op->recalc_thresholds(ctl); |
| 740 | spin_unlock(&ctl->tree_lock); |
| 741 | if (ret) { |
| 742 | btrfs_err(root->fs_info, |
| 743 | "Duplicate entries in free space cache, dumping"); |
| 744 | kmem_cache_free(btrfs_free_space_cachep, e); |
| 745 | goto free_cache; |
| 746 | } |
| 747 | list_add_tail(&e->list, &bitmaps); |
| 748 | } |
| 749 | |
| 750 | num_entries--; |
| 751 | } |
| 752 | |
| 753 | io_ctl_unmap_page(&io_ctl); |
| 754 | |
| 755 | /* |
| 756 | * We add the bitmaps at the end of the entries in order that |
| 757 | * the bitmap entries are added to the cache. |
| 758 | */ |
| 759 | list_for_each_entry_safe(e, n, &bitmaps, list) { |
| 760 | list_del_init(&e->list); |
| 761 | ret = io_ctl_read_bitmap(&io_ctl, e); |
| 762 | if (ret) |
| 763 | goto free_cache; |
| 764 | } |
| 765 | |
| 766 | io_ctl_drop_pages(&io_ctl); |
| 767 | merge_space_tree(ctl); |
| 768 | ret = 1; |
| 769 | out: |
| 770 | io_ctl_free(&io_ctl); |
| 771 | return ret; |
| 772 | free_cache: |
| 773 | io_ctl_drop_pages(&io_ctl); |
| 774 | __btrfs_remove_free_space_cache(ctl); |
| 775 | goto out; |
| 776 | } |
| 777 | |
| 778 | int load_free_space_cache(struct btrfs_fs_info *fs_info, |
| 779 | struct btrfs_block_group_cache *block_group) |
| 780 | { |
| 781 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 782 | struct btrfs_root *root = fs_info->tree_root; |
| 783 | struct inode *inode; |
| 784 | struct btrfs_path *path; |
| 785 | int ret = 0; |
| 786 | bool matched; |
| 787 | u64 used = btrfs_block_group_used(&block_group->item); |
| 788 | |
| 789 | /* |
| 790 | * If this block group has been marked to be cleared for one reason or |
| 791 | * another then we can't trust the on disk cache, so just return. |
| 792 | */ |
| 793 | spin_lock(&block_group->lock); |
| 794 | if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) { |
| 795 | spin_unlock(&block_group->lock); |
| 796 | return 0; |
| 797 | } |
| 798 | spin_unlock(&block_group->lock); |
| 799 | |
| 800 | path = btrfs_alloc_path(); |
| 801 | if (!path) |
| 802 | return 0; |
| 803 | path->search_commit_root = 1; |
| 804 | path->skip_locking = 1; |
| 805 | |
| 806 | inode = lookup_free_space_inode(root, block_group, path); |
| 807 | if (IS_ERR(inode)) { |
| 808 | btrfs_free_path(path); |
| 809 | return 0; |
| 810 | } |
| 811 | |
| 812 | /* We may have converted the inode and made the cache invalid. */ |
| 813 | spin_lock(&block_group->lock); |
| 814 | if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) { |
| 815 | spin_unlock(&block_group->lock); |
| 816 | btrfs_free_path(path); |
| 817 | goto out; |
| 818 | } |
| 819 | spin_unlock(&block_group->lock); |
| 820 | |
| 821 | ret = __load_free_space_cache(fs_info->tree_root, inode, ctl, |
| 822 | path, block_group->key.objectid); |
| 823 | btrfs_free_path(path); |
| 824 | if (ret <= 0) |
| 825 | goto out; |
| 826 | |
| 827 | spin_lock(&ctl->tree_lock); |
| 828 | matched = (ctl->free_space == (block_group->key.offset - used - |
| 829 | block_group->bytes_super)); |
| 830 | spin_unlock(&ctl->tree_lock); |
| 831 | |
| 832 | if (!matched) { |
| 833 | __btrfs_remove_free_space_cache(ctl); |
| 834 | btrfs_warn(fs_info, "block group %llu has wrong amount of free space", |
| 835 | block_group->key.objectid); |
| 836 | ret = -1; |
| 837 | } |
| 838 | out: |
| 839 | if (ret < 0) { |
| 840 | /* This cache is bogus, make sure it gets cleared */ |
| 841 | spin_lock(&block_group->lock); |
| 842 | block_group->disk_cache_state = BTRFS_DC_CLEAR; |
| 843 | spin_unlock(&block_group->lock); |
| 844 | ret = 0; |
| 845 | |
| 846 | btrfs_warn(fs_info, "failed to load free space cache for block group %llu, rebuild it now", |
| 847 | block_group->key.objectid); |
| 848 | } |
| 849 | |
| 850 | iput(inode); |
| 851 | return ret; |
| 852 | } |
| 853 | |
| 854 | static noinline_for_stack |
| 855 | int write_cache_extent_entries(struct io_ctl *io_ctl, |
| 856 | struct btrfs_free_space_ctl *ctl, |
| 857 | struct btrfs_block_group_cache *block_group, |
| 858 | int *entries, int *bitmaps, |
| 859 | struct list_head *bitmap_list) |
| 860 | { |
| 861 | int ret; |
| 862 | struct btrfs_free_cluster *cluster = NULL; |
| 863 | struct rb_node *node = rb_first(&ctl->free_space_offset); |
| 864 | |
| 865 | /* Get the cluster for this block_group if it exists */ |
| 866 | if (block_group && !list_empty(&block_group->cluster_list)) { |
| 867 | cluster = list_entry(block_group->cluster_list.next, |
| 868 | struct btrfs_free_cluster, |
| 869 | block_group_list); |
| 870 | } |
| 871 | |
| 872 | if (!node && cluster) { |
| 873 | node = rb_first(&cluster->root); |
| 874 | cluster = NULL; |
| 875 | } |
| 876 | |
| 877 | /* Write out the extent entries */ |
| 878 | while (node) { |
| 879 | struct btrfs_free_space *e; |
| 880 | |
| 881 | e = rb_entry(node, struct btrfs_free_space, offset_index); |
| 882 | *entries += 1; |
| 883 | |
| 884 | ret = io_ctl_add_entry(io_ctl, e->offset, e->bytes, |
| 885 | e->bitmap); |
| 886 | if (ret) |
| 887 | goto fail; |
| 888 | |
| 889 | if (e->bitmap) { |
| 890 | list_add_tail(&e->list, bitmap_list); |
| 891 | *bitmaps += 1; |
| 892 | } |
| 893 | node = rb_next(node); |
| 894 | if (!node && cluster) { |
| 895 | node = rb_first(&cluster->root); |
| 896 | cluster = NULL; |
| 897 | } |
| 898 | } |
| 899 | return 0; |
| 900 | fail: |
| 901 | return -ENOSPC; |
| 902 | } |
| 903 | |
| 904 | static noinline_for_stack int |
| 905 | update_cache_item(struct btrfs_trans_handle *trans, |
| 906 | struct btrfs_root *root, |
| 907 | struct inode *inode, |
| 908 | struct btrfs_path *path, u64 offset, |
| 909 | int entries, int bitmaps) |
| 910 | { |
| 911 | struct btrfs_key key; |
| 912 | struct btrfs_free_space_header *header; |
| 913 | struct extent_buffer *leaf; |
| 914 | int ret; |
| 915 | |
| 916 | key.objectid = BTRFS_FREE_SPACE_OBJECTID; |
| 917 | key.offset = offset; |
| 918 | key.type = 0; |
| 919 | |
| 920 | ret = btrfs_search_slot(trans, root, &key, path, 0, 1); |
| 921 | if (ret < 0) { |
| 922 | clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1, |
| 923 | EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL, |
| 924 | GFP_NOFS); |
| 925 | goto fail; |
| 926 | } |
| 927 | leaf = path->nodes[0]; |
| 928 | if (ret > 0) { |
| 929 | struct btrfs_key found_key; |
| 930 | ASSERT(path->slots[0]); |
| 931 | path->slots[0]--; |
| 932 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); |
| 933 | if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID || |
| 934 | found_key.offset != offset) { |
| 935 | clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, |
| 936 | inode->i_size - 1, |
| 937 | EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, |
| 938 | NULL, GFP_NOFS); |
| 939 | btrfs_release_path(path); |
| 940 | goto fail; |
| 941 | } |
| 942 | } |
| 943 | |
| 944 | BTRFS_I(inode)->generation = trans->transid; |
| 945 | header = btrfs_item_ptr(leaf, path->slots[0], |
| 946 | struct btrfs_free_space_header); |
| 947 | btrfs_set_free_space_entries(leaf, header, entries); |
| 948 | btrfs_set_free_space_bitmaps(leaf, header, bitmaps); |
| 949 | btrfs_set_free_space_generation(leaf, header, trans->transid); |
| 950 | btrfs_mark_buffer_dirty(leaf); |
| 951 | btrfs_release_path(path); |
| 952 | |
| 953 | return 0; |
| 954 | |
| 955 | fail: |
| 956 | return -1; |
| 957 | } |
| 958 | |
| 959 | static noinline_for_stack int |
| 960 | add_ioctl_entries(struct btrfs_root *root, |
| 961 | struct inode *inode, |
| 962 | struct btrfs_block_group_cache *block_group, |
| 963 | struct io_ctl *io_ctl, |
| 964 | struct extent_state **cached_state, |
| 965 | struct list_head *bitmap_list, |
| 966 | int *entries) |
| 967 | { |
| 968 | u64 start, extent_start, extent_end, len; |
| 969 | struct list_head *pos, *n; |
| 970 | struct extent_io_tree *unpin = NULL; |
| 971 | int ret; |
| 972 | |
| 973 | /* |
| 974 | * We want to add any pinned extents to our free space cache |
| 975 | * so we don't leak the space |
| 976 | * |
| 977 | * We shouldn't have switched the pinned extents yet so this is the |
| 978 | * right one |
| 979 | */ |
| 980 | unpin = root->fs_info->pinned_extents; |
| 981 | |
| 982 | if (block_group) |
| 983 | start = block_group->key.objectid; |
| 984 | |
| 985 | while (block_group && (start < block_group->key.objectid + |
| 986 | block_group->key.offset)) { |
| 987 | ret = find_first_extent_bit(unpin, start, |
| 988 | &extent_start, &extent_end, |
| 989 | EXTENT_DIRTY, NULL); |
| 990 | if (ret) { |
| 991 | ret = 0; |
| 992 | break; |
| 993 | } |
| 994 | |
| 995 | /* This pinned extent is out of our range */ |
| 996 | if (extent_start >= block_group->key.objectid + |
| 997 | block_group->key.offset) |
| 998 | break; |
| 999 | |
| 1000 | extent_start = max(extent_start, start); |
| 1001 | extent_end = min(block_group->key.objectid + |
| 1002 | block_group->key.offset, extent_end + 1); |
| 1003 | len = extent_end - extent_start; |
| 1004 | |
| 1005 | *entries += 1; |
| 1006 | ret = io_ctl_add_entry(io_ctl, extent_start, len, NULL); |
| 1007 | if (ret) |
| 1008 | goto out_nospc; |
| 1009 | |
| 1010 | start = extent_end; |
| 1011 | } |
| 1012 | |
| 1013 | /* Write out the bitmaps */ |
| 1014 | list_for_each_safe(pos, n, bitmap_list) { |
| 1015 | struct btrfs_free_space *entry = |
| 1016 | list_entry(pos, struct btrfs_free_space, list); |
| 1017 | |
| 1018 | ret = io_ctl_add_bitmap(io_ctl, entry->bitmap); |
| 1019 | if (ret) |
| 1020 | goto out_nospc; |
| 1021 | list_del_init(&entry->list); |
| 1022 | } |
| 1023 | |
| 1024 | /* Zero out the rest of the pages just to make sure */ |
| 1025 | io_ctl_zero_remaining_pages(io_ctl); |
| 1026 | |
| 1027 | ret = btrfs_dirty_pages(root, inode, io_ctl->pages, io_ctl->num_pages, |
| 1028 | 0, i_size_read(inode), cached_state); |
| 1029 | io_ctl_drop_pages(io_ctl); |
| 1030 | unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0, |
| 1031 | i_size_read(inode) - 1, cached_state, GFP_NOFS); |
| 1032 | |
| 1033 | if (ret) |
| 1034 | goto fail; |
| 1035 | |
| 1036 | ret = btrfs_wait_ordered_range(inode, 0, (u64)-1); |
| 1037 | if (ret) { |
| 1038 | clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1, |
| 1039 | EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL, |
| 1040 | GFP_NOFS); |
| 1041 | goto fail; |
| 1042 | } |
| 1043 | return 0; |
| 1044 | |
| 1045 | fail: |
| 1046 | return -1; |
| 1047 | |
| 1048 | out_nospc: |
| 1049 | return -ENOSPC; |
| 1050 | } |
| 1051 | |
| 1052 | static void noinline_for_stack |
| 1053 | cleanup_write_cache_enospc(struct inode *inode, |
| 1054 | struct io_ctl *io_ctl, |
| 1055 | struct extent_state **cached_state, |
| 1056 | struct list_head *bitmap_list) |
| 1057 | { |
| 1058 | struct list_head *pos, *n; |
| 1059 | list_for_each_safe(pos, n, bitmap_list) { |
| 1060 | struct btrfs_free_space *entry = |
| 1061 | list_entry(pos, struct btrfs_free_space, list); |
| 1062 | list_del_init(&entry->list); |
| 1063 | } |
| 1064 | io_ctl_drop_pages(io_ctl); |
| 1065 | unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0, |
| 1066 | i_size_read(inode) - 1, cached_state, |
| 1067 | GFP_NOFS); |
| 1068 | } |
| 1069 | |
| 1070 | /** |
| 1071 | * __btrfs_write_out_cache - write out cached info to an inode |
| 1072 | * @root - the root the inode belongs to |
| 1073 | * @ctl - the free space cache we are going to write out |
| 1074 | * @block_group - the block_group for this cache if it belongs to a block_group |
| 1075 | * @trans - the trans handle |
| 1076 | * @path - the path to use |
| 1077 | * @offset - the offset for the key we'll insert |
| 1078 | * |
| 1079 | * This function writes out a free space cache struct to disk for quick recovery |
| 1080 | * on mount. This will return 0 if it was successfull in writing the cache out, |
| 1081 | * and -1 if it was not. |
| 1082 | */ |
| 1083 | static int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, |
| 1084 | struct btrfs_free_space_ctl *ctl, |
| 1085 | struct btrfs_block_group_cache *block_group, |
| 1086 | struct btrfs_trans_handle *trans, |
| 1087 | struct btrfs_path *path, u64 offset) |
| 1088 | { |
| 1089 | struct extent_state *cached_state = NULL; |
| 1090 | struct io_ctl io_ctl; |
| 1091 | struct list_head bitmap_list; |
| 1092 | int entries = 0; |
| 1093 | int bitmaps = 0; |
| 1094 | int ret; |
| 1095 | int err = -1; |
| 1096 | |
| 1097 | INIT_LIST_HEAD(&bitmap_list); |
| 1098 | |
| 1099 | if (!i_size_read(inode)) |
| 1100 | return -1; |
| 1101 | |
| 1102 | ret = io_ctl_init(&io_ctl, inode, root); |
| 1103 | if (ret) |
| 1104 | return -1; |
| 1105 | |
| 1106 | /* Lock all pages first so we can lock the extent safely. */ |
| 1107 | io_ctl_prepare_pages(&io_ctl, inode, 0); |
| 1108 | |
| 1109 | lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1, |
| 1110 | 0, &cached_state); |
| 1111 | |
| 1112 | |
| 1113 | /* Make sure we can fit our crcs into the first page */ |
| 1114 | if (io_ctl.check_crcs && |
| 1115 | (io_ctl.num_pages * sizeof(u32)) >= PAGE_CACHE_SIZE) |
| 1116 | goto out_nospc; |
| 1117 | |
| 1118 | io_ctl_set_generation(&io_ctl, trans->transid); |
| 1119 | |
| 1120 | ret = write_cache_extent_entries(&io_ctl, ctl, |
| 1121 | block_group, &entries, &bitmaps, |
| 1122 | &bitmap_list); |
| 1123 | if (ret) |
| 1124 | goto out_nospc; |
| 1125 | |
| 1126 | ret = add_ioctl_entries(root, inode, block_group, &io_ctl, |
| 1127 | &cached_state, &bitmap_list, &entries); |
| 1128 | |
| 1129 | if (ret == -ENOSPC) |
| 1130 | goto out_nospc; |
| 1131 | else if (ret) |
| 1132 | goto out; |
| 1133 | |
| 1134 | err = update_cache_item(trans, root, inode, path, offset, |
| 1135 | entries, bitmaps); |
| 1136 | |
| 1137 | out: |
| 1138 | io_ctl_free(&io_ctl); |
| 1139 | if (err) { |
| 1140 | invalidate_inode_pages2(inode->i_mapping); |
| 1141 | BTRFS_I(inode)->generation = 0; |
| 1142 | } |
| 1143 | btrfs_update_inode(trans, root, inode); |
| 1144 | return err; |
| 1145 | |
| 1146 | out_nospc: |
| 1147 | |
| 1148 | cleanup_write_cache_enospc(inode, &io_ctl, &cached_state, &bitmap_list); |
| 1149 | goto out; |
| 1150 | } |
| 1151 | |
| 1152 | int btrfs_write_out_cache(struct btrfs_root *root, |
| 1153 | struct btrfs_trans_handle *trans, |
| 1154 | struct btrfs_block_group_cache *block_group, |
| 1155 | struct btrfs_path *path) |
| 1156 | { |
| 1157 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 1158 | struct inode *inode; |
| 1159 | int ret = 0; |
| 1160 | |
| 1161 | root = root->fs_info->tree_root; |
| 1162 | |
| 1163 | spin_lock(&block_group->lock); |
| 1164 | if (block_group->disk_cache_state < BTRFS_DC_SETUP) { |
| 1165 | spin_unlock(&block_group->lock); |
| 1166 | return 0; |
| 1167 | } |
| 1168 | spin_unlock(&block_group->lock); |
| 1169 | |
| 1170 | inode = lookup_free_space_inode(root, block_group, path); |
| 1171 | if (IS_ERR(inode)) |
| 1172 | return 0; |
| 1173 | |
| 1174 | ret = __btrfs_write_out_cache(root, inode, ctl, block_group, trans, |
| 1175 | path, block_group->key.objectid); |
| 1176 | if (ret) { |
| 1177 | spin_lock(&block_group->lock); |
| 1178 | block_group->disk_cache_state = BTRFS_DC_ERROR; |
| 1179 | spin_unlock(&block_group->lock); |
| 1180 | ret = 0; |
| 1181 | #ifdef DEBUG |
| 1182 | btrfs_err(root->fs_info, |
| 1183 | "failed to write free space cache for block group %llu", |
| 1184 | block_group->key.objectid); |
| 1185 | #endif |
| 1186 | } |
| 1187 | |
| 1188 | iput(inode); |
| 1189 | return ret; |
| 1190 | } |
| 1191 | |
| 1192 | static inline unsigned long offset_to_bit(u64 bitmap_start, u32 unit, |
| 1193 | u64 offset) |
| 1194 | { |
| 1195 | ASSERT(offset >= bitmap_start); |
| 1196 | offset -= bitmap_start; |
| 1197 | return (unsigned long)(div_u64(offset, unit)); |
| 1198 | } |
| 1199 | |
| 1200 | static inline unsigned long bytes_to_bits(u64 bytes, u32 unit) |
| 1201 | { |
| 1202 | return (unsigned long)(div_u64(bytes, unit)); |
| 1203 | } |
| 1204 | |
| 1205 | static inline u64 offset_to_bitmap(struct btrfs_free_space_ctl *ctl, |
| 1206 | u64 offset) |
| 1207 | { |
| 1208 | u64 bitmap_start; |
| 1209 | u64 bytes_per_bitmap; |
| 1210 | |
| 1211 | bytes_per_bitmap = BITS_PER_BITMAP * ctl->unit; |
| 1212 | bitmap_start = offset - ctl->start; |
| 1213 | bitmap_start = div64_u64(bitmap_start, bytes_per_bitmap); |
| 1214 | bitmap_start *= bytes_per_bitmap; |
| 1215 | bitmap_start += ctl->start; |
| 1216 | |
| 1217 | return bitmap_start; |
| 1218 | } |
| 1219 | |
| 1220 | static int tree_insert_offset(struct rb_root *root, u64 offset, |
| 1221 | struct rb_node *node, int bitmap) |
| 1222 | { |
| 1223 | struct rb_node **p = &root->rb_node; |
| 1224 | struct rb_node *parent = NULL; |
| 1225 | struct btrfs_free_space *info; |
| 1226 | |
| 1227 | while (*p) { |
| 1228 | parent = *p; |
| 1229 | info = rb_entry(parent, struct btrfs_free_space, offset_index); |
| 1230 | |
| 1231 | if (offset < info->offset) { |
| 1232 | p = &(*p)->rb_left; |
| 1233 | } else if (offset > info->offset) { |
| 1234 | p = &(*p)->rb_right; |
| 1235 | } else { |
| 1236 | /* |
| 1237 | * we could have a bitmap entry and an extent entry |
| 1238 | * share the same offset. If this is the case, we want |
| 1239 | * the extent entry to always be found first if we do a |
| 1240 | * linear search through the tree, since we want to have |
| 1241 | * the quickest allocation time, and allocating from an |
| 1242 | * extent is faster than allocating from a bitmap. So |
| 1243 | * if we're inserting a bitmap and we find an entry at |
| 1244 | * this offset, we want to go right, or after this entry |
| 1245 | * logically. If we are inserting an extent and we've |
| 1246 | * found a bitmap, we want to go left, or before |
| 1247 | * logically. |
| 1248 | */ |
| 1249 | if (bitmap) { |
| 1250 | if (info->bitmap) { |
| 1251 | WARN_ON_ONCE(1); |
| 1252 | return -EEXIST; |
| 1253 | } |
| 1254 | p = &(*p)->rb_right; |
| 1255 | } else { |
| 1256 | if (!info->bitmap) { |
| 1257 | WARN_ON_ONCE(1); |
| 1258 | return -EEXIST; |
| 1259 | } |
| 1260 | p = &(*p)->rb_left; |
| 1261 | } |
| 1262 | } |
| 1263 | } |
| 1264 | |
| 1265 | rb_link_node(node, parent, p); |
| 1266 | rb_insert_color(node, root); |
| 1267 | |
| 1268 | return 0; |
| 1269 | } |
| 1270 | |
| 1271 | /* |
| 1272 | * searches the tree for the given offset. |
| 1273 | * |
| 1274 | * fuzzy - If this is set, then we are trying to make an allocation, and we just |
| 1275 | * want a section that has at least bytes size and comes at or after the given |
| 1276 | * offset. |
| 1277 | */ |
| 1278 | static struct btrfs_free_space * |
| 1279 | tree_search_offset(struct btrfs_free_space_ctl *ctl, |
| 1280 | u64 offset, int bitmap_only, int fuzzy) |
| 1281 | { |
| 1282 | struct rb_node *n = ctl->free_space_offset.rb_node; |
| 1283 | struct btrfs_free_space *entry, *prev = NULL; |
| 1284 | |
| 1285 | /* find entry that is closest to the 'offset' */ |
| 1286 | while (1) { |
| 1287 | if (!n) { |
| 1288 | entry = NULL; |
| 1289 | break; |
| 1290 | } |
| 1291 | |
| 1292 | entry = rb_entry(n, struct btrfs_free_space, offset_index); |
| 1293 | prev = entry; |
| 1294 | |
| 1295 | if (offset < entry->offset) |
| 1296 | n = n->rb_left; |
| 1297 | else if (offset > entry->offset) |
| 1298 | n = n->rb_right; |
| 1299 | else |
| 1300 | break; |
| 1301 | } |
| 1302 | |
| 1303 | if (bitmap_only) { |
| 1304 | if (!entry) |
| 1305 | return NULL; |
| 1306 | if (entry->bitmap) |
| 1307 | return entry; |
| 1308 | |
| 1309 | /* |
| 1310 | * bitmap entry and extent entry may share same offset, |
| 1311 | * in that case, bitmap entry comes after extent entry. |
| 1312 | */ |
| 1313 | n = rb_next(n); |
| 1314 | if (!n) |
| 1315 | return NULL; |
| 1316 | entry = rb_entry(n, struct btrfs_free_space, offset_index); |
| 1317 | if (entry->offset != offset) |
| 1318 | return NULL; |
| 1319 | |
| 1320 | WARN_ON(!entry->bitmap); |
| 1321 | return entry; |
| 1322 | } else if (entry) { |
| 1323 | if (entry->bitmap) { |
| 1324 | /* |
| 1325 | * if previous extent entry covers the offset, |
| 1326 | * we should return it instead of the bitmap entry |
| 1327 | */ |
| 1328 | n = rb_prev(&entry->offset_index); |
| 1329 | if (n) { |
| 1330 | prev = rb_entry(n, struct btrfs_free_space, |
| 1331 | offset_index); |
| 1332 | if (!prev->bitmap && |
| 1333 | prev->offset + prev->bytes > offset) |
| 1334 | entry = prev; |
| 1335 | } |
| 1336 | } |
| 1337 | return entry; |
| 1338 | } |
| 1339 | |
| 1340 | if (!prev) |
| 1341 | return NULL; |
| 1342 | |
| 1343 | /* find last entry before the 'offset' */ |
| 1344 | entry = prev; |
| 1345 | if (entry->offset > offset) { |
| 1346 | n = rb_prev(&entry->offset_index); |
| 1347 | if (n) { |
| 1348 | entry = rb_entry(n, struct btrfs_free_space, |
| 1349 | offset_index); |
| 1350 | ASSERT(entry->offset <= offset); |
| 1351 | } else { |
| 1352 | if (fuzzy) |
| 1353 | return entry; |
| 1354 | else |
| 1355 | return NULL; |
| 1356 | } |
| 1357 | } |
| 1358 | |
| 1359 | if (entry->bitmap) { |
| 1360 | n = rb_prev(&entry->offset_index); |
| 1361 | if (n) { |
| 1362 | prev = rb_entry(n, struct btrfs_free_space, |
| 1363 | offset_index); |
| 1364 | if (!prev->bitmap && |
| 1365 | prev->offset + prev->bytes > offset) |
| 1366 | return prev; |
| 1367 | } |
| 1368 | if (entry->offset + BITS_PER_BITMAP * ctl->unit > offset) |
| 1369 | return entry; |
| 1370 | } else if (entry->offset + entry->bytes > offset) |
| 1371 | return entry; |
| 1372 | |
| 1373 | if (!fuzzy) |
| 1374 | return NULL; |
| 1375 | |
| 1376 | while (1) { |
| 1377 | if (entry->bitmap) { |
| 1378 | if (entry->offset + BITS_PER_BITMAP * |
| 1379 | ctl->unit > offset) |
| 1380 | break; |
| 1381 | } else { |
| 1382 | if (entry->offset + entry->bytes > offset) |
| 1383 | break; |
| 1384 | } |
| 1385 | |
| 1386 | n = rb_next(&entry->offset_index); |
| 1387 | if (!n) |
| 1388 | return NULL; |
| 1389 | entry = rb_entry(n, struct btrfs_free_space, offset_index); |
| 1390 | } |
| 1391 | return entry; |
| 1392 | } |
| 1393 | |
| 1394 | static inline void |
| 1395 | __unlink_free_space(struct btrfs_free_space_ctl *ctl, |
| 1396 | struct btrfs_free_space *info) |
| 1397 | { |
| 1398 | rb_erase(&info->offset_index, &ctl->free_space_offset); |
| 1399 | ctl->free_extents--; |
| 1400 | } |
| 1401 | |
| 1402 | static void unlink_free_space(struct btrfs_free_space_ctl *ctl, |
| 1403 | struct btrfs_free_space *info) |
| 1404 | { |
| 1405 | __unlink_free_space(ctl, info); |
| 1406 | ctl->free_space -= info->bytes; |
| 1407 | } |
| 1408 | |
| 1409 | static int link_free_space(struct btrfs_free_space_ctl *ctl, |
| 1410 | struct btrfs_free_space *info) |
| 1411 | { |
| 1412 | int ret = 0; |
| 1413 | |
| 1414 | ASSERT(info->bytes || info->bitmap); |
| 1415 | ret = tree_insert_offset(&ctl->free_space_offset, info->offset, |
| 1416 | &info->offset_index, (info->bitmap != NULL)); |
| 1417 | if (ret) |
| 1418 | return ret; |
| 1419 | |
| 1420 | ctl->free_space += info->bytes; |
| 1421 | ctl->free_extents++; |
| 1422 | return ret; |
| 1423 | } |
| 1424 | |
| 1425 | static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl) |
| 1426 | { |
| 1427 | struct btrfs_block_group_cache *block_group = ctl->private; |
| 1428 | u64 max_bytes; |
| 1429 | u64 bitmap_bytes; |
| 1430 | u64 extent_bytes; |
| 1431 | u64 size = block_group->key.offset; |
| 1432 | u64 bytes_per_bg = BITS_PER_BITMAP * ctl->unit; |
| 1433 | int max_bitmaps = div64_u64(size + bytes_per_bg - 1, bytes_per_bg); |
| 1434 | |
| 1435 | max_bitmaps = max(max_bitmaps, 1); |
| 1436 | |
| 1437 | ASSERT(ctl->total_bitmaps <= max_bitmaps); |
| 1438 | |
| 1439 | /* |
| 1440 | * The goal is to keep the total amount of memory used per 1gb of space |
| 1441 | * at or below 32k, so we need to adjust how much memory we allow to be |
| 1442 | * used by extent based free space tracking |
| 1443 | */ |
| 1444 | if (size < 1024 * 1024 * 1024) |
| 1445 | max_bytes = MAX_CACHE_BYTES_PER_GIG; |
| 1446 | else |
| 1447 | max_bytes = MAX_CACHE_BYTES_PER_GIG * |
| 1448 | div64_u64(size, 1024 * 1024 * 1024); |
| 1449 | |
| 1450 | /* |
| 1451 | * we want to account for 1 more bitmap than what we have so we can make |
| 1452 | * sure we don't go over our overall goal of MAX_CACHE_BYTES_PER_GIG as |
| 1453 | * we add more bitmaps. |
| 1454 | */ |
| 1455 | bitmap_bytes = (ctl->total_bitmaps + 1) * PAGE_CACHE_SIZE; |
| 1456 | |
| 1457 | if (bitmap_bytes >= max_bytes) { |
| 1458 | ctl->extents_thresh = 0; |
| 1459 | return; |
| 1460 | } |
| 1461 | |
| 1462 | /* |
| 1463 | * we want the extent entry threshold to always be at most 1/2 the maxw |
| 1464 | * bytes we can have, or whatever is less than that. |
| 1465 | */ |
| 1466 | extent_bytes = max_bytes - bitmap_bytes; |
| 1467 | extent_bytes = min_t(u64, extent_bytes, div64_u64(max_bytes, 2)); |
| 1468 | |
| 1469 | ctl->extents_thresh = |
| 1470 | div64_u64(extent_bytes, (sizeof(struct btrfs_free_space))); |
| 1471 | } |
| 1472 | |
| 1473 | static inline void __bitmap_clear_bits(struct btrfs_free_space_ctl *ctl, |
| 1474 | struct btrfs_free_space *info, |
| 1475 | u64 offset, u64 bytes) |
| 1476 | { |
| 1477 | unsigned long start, count; |
| 1478 | |
| 1479 | start = offset_to_bit(info->offset, ctl->unit, offset); |
| 1480 | count = bytes_to_bits(bytes, ctl->unit); |
| 1481 | ASSERT(start + count <= BITS_PER_BITMAP); |
| 1482 | |
| 1483 | bitmap_clear(info->bitmap, start, count); |
| 1484 | |
| 1485 | info->bytes -= bytes; |
| 1486 | } |
| 1487 | |
| 1488 | static void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl, |
| 1489 | struct btrfs_free_space *info, u64 offset, |
| 1490 | u64 bytes) |
| 1491 | { |
| 1492 | __bitmap_clear_bits(ctl, info, offset, bytes); |
| 1493 | ctl->free_space -= bytes; |
| 1494 | } |
| 1495 | |
| 1496 | static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl, |
| 1497 | struct btrfs_free_space *info, u64 offset, |
| 1498 | u64 bytes) |
| 1499 | { |
| 1500 | unsigned long start, count; |
| 1501 | |
| 1502 | start = offset_to_bit(info->offset, ctl->unit, offset); |
| 1503 | count = bytes_to_bits(bytes, ctl->unit); |
| 1504 | ASSERT(start + count <= BITS_PER_BITMAP); |
| 1505 | |
| 1506 | bitmap_set(info->bitmap, start, count); |
| 1507 | |
| 1508 | info->bytes += bytes; |
| 1509 | ctl->free_space += bytes; |
| 1510 | } |
| 1511 | |
| 1512 | /* |
| 1513 | * If we can not find suitable extent, we will use bytes to record |
| 1514 | * the size of the max extent. |
| 1515 | */ |
| 1516 | static int search_bitmap(struct btrfs_free_space_ctl *ctl, |
| 1517 | struct btrfs_free_space *bitmap_info, u64 *offset, |
| 1518 | u64 *bytes) |
| 1519 | { |
| 1520 | unsigned long found_bits = 0; |
| 1521 | unsigned long max_bits = 0; |
| 1522 | unsigned long bits, i; |
| 1523 | unsigned long next_zero; |
| 1524 | unsigned long extent_bits; |
| 1525 | |
| 1526 | i = offset_to_bit(bitmap_info->offset, ctl->unit, |
| 1527 | max_t(u64, *offset, bitmap_info->offset)); |
| 1528 | bits = bytes_to_bits(*bytes, ctl->unit); |
| 1529 | |
| 1530 | for_each_set_bit_from(i, bitmap_info->bitmap, BITS_PER_BITMAP) { |
| 1531 | next_zero = find_next_zero_bit(bitmap_info->bitmap, |
| 1532 | BITS_PER_BITMAP, i); |
| 1533 | extent_bits = next_zero - i; |
| 1534 | if (extent_bits >= bits) { |
| 1535 | found_bits = extent_bits; |
| 1536 | break; |
| 1537 | } else if (extent_bits > max_bits) { |
| 1538 | max_bits = extent_bits; |
| 1539 | } |
| 1540 | i = next_zero; |
| 1541 | } |
| 1542 | |
| 1543 | if (found_bits) { |
| 1544 | *offset = (u64)(i * ctl->unit) + bitmap_info->offset; |
| 1545 | *bytes = (u64)(found_bits) * ctl->unit; |
| 1546 | return 0; |
| 1547 | } |
| 1548 | |
| 1549 | *bytes = (u64)(max_bits) * ctl->unit; |
| 1550 | return -1; |
| 1551 | } |
| 1552 | |
| 1553 | /* Cache the size of the max extent in bytes */ |
| 1554 | static struct btrfs_free_space * |
| 1555 | find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes, |
| 1556 | unsigned long align, u64 *max_extent_size) |
| 1557 | { |
| 1558 | struct btrfs_free_space *entry; |
| 1559 | struct rb_node *node; |
| 1560 | u64 tmp; |
| 1561 | u64 align_off; |
| 1562 | int ret; |
| 1563 | |
| 1564 | if (!ctl->free_space_offset.rb_node) |
| 1565 | goto out; |
| 1566 | |
| 1567 | entry = tree_search_offset(ctl, offset_to_bitmap(ctl, *offset), 0, 1); |
| 1568 | if (!entry) |
| 1569 | goto out; |
| 1570 | |
| 1571 | for (node = &entry->offset_index; node; node = rb_next(node)) { |
| 1572 | entry = rb_entry(node, struct btrfs_free_space, offset_index); |
| 1573 | if (entry->bytes < *bytes) { |
| 1574 | if (entry->bytes > *max_extent_size) |
| 1575 | *max_extent_size = entry->bytes; |
| 1576 | continue; |
| 1577 | } |
| 1578 | |
| 1579 | /* make sure the space returned is big enough |
| 1580 | * to match our requested alignment |
| 1581 | */ |
| 1582 | if (*bytes >= align) { |
| 1583 | tmp = entry->offset - ctl->start + align - 1; |
| 1584 | do_div(tmp, align); |
| 1585 | tmp = tmp * align + ctl->start; |
| 1586 | align_off = tmp - entry->offset; |
| 1587 | } else { |
| 1588 | align_off = 0; |
| 1589 | tmp = entry->offset; |
| 1590 | } |
| 1591 | |
| 1592 | if (entry->bytes < *bytes + align_off) { |
| 1593 | if (entry->bytes > *max_extent_size) |
| 1594 | *max_extent_size = entry->bytes; |
| 1595 | continue; |
| 1596 | } |
| 1597 | |
| 1598 | if (entry->bitmap) { |
| 1599 | u64 size = *bytes; |
| 1600 | |
| 1601 | ret = search_bitmap(ctl, entry, &tmp, &size); |
| 1602 | if (!ret) { |
| 1603 | *offset = tmp; |
| 1604 | *bytes = size; |
| 1605 | return entry; |
| 1606 | } else if (size > *max_extent_size) { |
| 1607 | *max_extent_size = size; |
| 1608 | } |
| 1609 | continue; |
| 1610 | } |
| 1611 | |
| 1612 | *offset = tmp; |
| 1613 | *bytes = entry->bytes - align_off; |
| 1614 | return entry; |
| 1615 | } |
| 1616 | out: |
| 1617 | return NULL; |
| 1618 | } |
| 1619 | |
| 1620 | static void add_new_bitmap(struct btrfs_free_space_ctl *ctl, |
| 1621 | struct btrfs_free_space *info, u64 offset) |
| 1622 | { |
| 1623 | info->offset = offset_to_bitmap(ctl, offset); |
| 1624 | info->bytes = 0; |
| 1625 | INIT_LIST_HEAD(&info->list); |
| 1626 | link_free_space(ctl, info); |
| 1627 | ctl->total_bitmaps++; |
| 1628 | |
| 1629 | ctl->op->recalc_thresholds(ctl); |
| 1630 | } |
| 1631 | |
| 1632 | static void free_bitmap(struct btrfs_free_space_ctl *ctl, |
| 1633 | struct btrfs_free_space *bitmap_info) |
| 1634 | { |
| 1635 | unlink_free_space(ctl, bitmap_info); |
| 1636 | kfree(bitmap_info->bitmap); |
| 1637 | kmem_cache_free(btrfs_free_space_cachep, bitmap_info); |
| 1638 | ctl->total_bitmaps--; |
| 1639 | ctl->op->recalc_thresholds(ctl); |
| 1640 | } |
| 1641 | |
| 1642 | static noinline int remove_from_bitmap(struct btrfs_free_space_ctl *ctl, |
| 1643 | struct btrfs_free_space *bitmap_info, |
| 1644 | u64 *offset, u64 *bytes) |
| 1645 | { |
| 1646 | u64 end; |
| 1647 | u64 search_start, search_bytes; |
| 1648 | int ret; |
| 1649 | |
| 1650 | again: |
| 1651 | end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit) - 1; |
| 1652 | |
| 1653 | /* |
| 1654 | * We need to search for bits in this bitmap. We could only cover some |
| 1655 | * of the extent in this bitmap thanks to how we add space, so we need |
| 1656 | * to search for as much as it as we can and clear that amount, and then |
| 1657 | * go searching for the next bit. |
| 1658 | */ |
| 1659 | search_start = *offset; |
| 1660 | search_bytes = ctl->unit; |
| 1661 | search_bytes = min(search_bytes, end - search_start + 1); |
| 1662 | ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes); |
| 1663 | if (ret < 0 || search_start != *offset) |
| 1664 | return -EINVAL; |
| 1665 | |
| 1666 | /* We may have found more bits than what we need */ |
| 1667 | search_bytes = min(search_bytes, *bytes); |
| 1668 | |
| 1669 | /* Cannot clear past the end of the bitmap */ |
| 1670 | search_bytes = min(search_bytes, end - search_start + 1); |
| 1671 | |
| 1672 | bitmap_clear_bits(ctl, bitmap_info, search_start, search_bytes); |
| 1673 | *offset += search_bytes; |
| 1674 | *bytes -= search_bytes; |
| 1675 | |
| 1676 | if (*bytes) { |
| 1677 | struct rb_node *next = rb_next(&bitmap_info->offset_index); |
| 1678 | if (!bitmap_info->bytes) |
| 1679 | free_bitmap(ctl, bitmap_info); |
| 1680 | |
| 1681 | /* |
| 1682 | * no entry after this bitmap, but we still have bytes to |
| 1683 | * remove, so something has gone wrong. |
| 1684 | */ |
| 1685 | if (!next) |
| 1686 | return -EINVAL; |
| 1687 | |
| 1688 | bitmap_info = rb_entry(next, struct btrfs_free_space, |
| 1689 | offset_index); |
| 1690 | |
| 1691 | /* |
| 1692 | * if the next entry isn't a bitmap we need to return to let the |
| 1693 | * extent stuff do its work. |
| 1694 | */ |
| 1695 | if (!bitmap_info->bitmap) |
| 1696 | return -EAGAIN; |
| 1697 | |
| 1698 | /* |
| 1699 | * Ok the next item is a bitmap, but it may not actually hold |
| 1700 | * the information for the rest of this free space stuff, so |
| 1701 | * look for it, and if we don't find it return so we can try |
| 1702 | * everything over again. |
| 1703 | */ |
| 1704 | search_start = *offset; |
| 1705 | search_bytes = ctl->unit; |
| 1706 | ret = search_bitmap(ctl, bitmap_info, &search_start, |
| 1707 | &search_bytes); |
| 1708 | if (ret < 0 || search_start != *offset) |
| 1709 | return -EAGAIN; |
| 1710 | |
| 1711 | goto again; |
| 1712 | } else if (!bitmap_info->bytes) |
| 1713 | free_bitmap(ctl, bitmap_info); |
| 1714 | |
| 1715 | return 0; |
| 1716 | } |
| 1717 | |
| 1718 | static u64 add_bytes_to_bitmap(struct btrfs_free_space_ctl *ctl, |
| 1719 | struct btrfs_free_space *info, u64 offset, |
| 1720 | u64 bytes) |
| 1721 | { |
| 1722 | u64 bytes_to_set = 0; |
| 1723 | u64 end; |
| 1724 | |
| 1725 | end = info->offset + (u64)(BITS_PER_BITMAP * ctl->unit); |
| 1726 | |
| 1727 | bytes_to_set = min(end - offset, bytes); |
| 1728 | |
| 1729 | bitmap_set_bits(ctl, info, offset, bytes_to_set); |
| 1730 | |
| 1731 | return bytes_to_set; |
| 1732 | |
| 1733 | } |
| 1734 | |
| 1735 | static bool use_bitmap(struct btrfs_free_space_ctl *ctl, |
| 1736 | struct btrfs_free_space *info) |
| 1737 | { |
| 1738 | struct btrfs_block_group_cache *block_group = ctl->private; |
| 1739 | |
| 1740 | /* |
| 1741 | * If we are below the extents threshold then we can add this as an |
| 1742 | * extent, and don't have to deal with the bitmap |
| 1743 | */ |
| 1744 | if (ctl->free_extents < ctl->extents_thresh) { |
| 1745 | /* |
| 1746 | * If this block group has some small extents we don't want to |
| 1747 | * use up all of our free slots in the cache with them, we want |
| 1748 | * to reserve them to larger extents, however if we have plent |
| 1749 | * of cache left then go ahead an dadd them, no sense in adding |
| 1750 | * the overhead of a bitmap if we don't have to. |
| 1751 | */ |
| 1752 | if (info->bytes <= block_group->sectorsize * 4) { |
| 1753 | if (ctl->free_extents * 2 <= ctl->extents_thresh) |
| 1754 | return false; |
| 1755 | } else { |
| 1756 | return false; |
| 1757 | } |
| 1758 | } |
| 1759 | |
| 1760 | /* |
| 1761 | * The original block groups from mkfs can be really small, like 8 |
| 1762 | * megabytes, so don't bother with a bitmap for those entries. However |
| 1763 | * some block groups can be smaller than what a bitmap would cover but |
| 1764 | * are still large enough that they could overflow the 32k memory limit, |
| 1765 | * so allow those block groups to still be allowed to have a bitmap |
| 1766 | * entry. |
| 1767 | */ |
| 1768 | if (((BITS_PER_BITMAP * ctl->unit) >> 1) > block_group->key.offset) |
| 1769 | return false; |
| 1770 | |
| 1771 | return true; |
| 1772 | } |
| 1773 | |
| 1774 | static struct btrfs_free_space_op free_space_op = { |
| 1775 | .recalc_thresholds = recalculate_thresholds, |
| 1776 | .use_bitmap = use_bitmap, |
| 1777 | }; |
| 1778 | |
| 1779 | static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl, |
| 1780 | struct btrfs_free_space *info) |
| 1781 | { |
| 1782 | struct btrfs_free_space *bitmap_info; |
| 1783 | struct btrfs_block_group_cache *block_group = NULL; |
| 1784 | int added = 0; |
| 1785 | u64 bytes, offset, bytes_added; |
| 1786 | int ret; |
| 1787 | |
| 1788 | bytes = info->bytes; |
| 1789 | offset = info->offset; |
| 1790 | |
| 1791 | if (!ctl->op->use_bitmap(ctl, info)) |
| 1792 | return 0; |
| 1793 | |
| 1794 | if (ctl->op == &free_space_op) |
| 1795 | block_group = ctl->private; |
| 1796 | again: |
| 1797 | /* |
| 1798 | * Since we link bitmaps right into the cluster we need to see if we |
| 1799 | * have a cluster here, and if so and it has our bitmap we need to add |
| 1800 | * the free space to that bitmap. |
| 1801 | */ |
| 1802 | if (block_group && !list_empty(&block_group->cluster_list)) { |
| 1803 | struct btrfs_free_cluster *cluster; |
| 1804 | struct rb_node *node; |
| 1805 | struct btrfs_free_space *entry; |
| 1806 | |
| 1807 | cluster = list_entry(block_group->cluster_list.next, |
| 1808 | struct btrfs_free_cluster, |
| 1809 | block_group_list); |
| 1810 | spin_lock(&cluster->lock); |
| 1811 | node = rb_first(&cluster->root); |
| 1812 | if (!node) { |
| 1813 | spin_unlock(&cluster->lock); |
| 1814 | goto no_cluster_bitmap; |
| 1815 | } |
| 1816 | |
| 1817 | entry = rb_entry(node, struct btrfs_free_space, offset_index); |
| 1818 | if (!entry->bitmap) { |
| 1819 | spin_unlock(&cluster->lock); |
| 1820 | goto no_cluster_bitmap; |
| 1821 | } |
| 1822 | |
| 1823 | if (entry->offset == offset_to_bitmap(ctl, offset)) { |
| 1824 | bytes_added = add_bytes_to_bitmap(ctl, entry, |
| 1825 | offset, bytes); |
| 1826 | bytes -= bytes_added; |
| 1827 | offset += bytes_added; |
| 1828 | } |
| 1829 | spin_unlock(&cluster->lock); |
| 1830 | if (!bytes) { |
| 1831 | ret = 1; |
| 1832 | goto out; |
| 1833 | } |
| 1834 | } |
| 1835 | |
| 1836 | no_cluster_bitmap: |
| 1837 | bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), |
| 1838 | 1, 0); |
| 1839 | if (!bitmap_info) { |
| 1840 | ASSERT(added == 0); |
| 1841 | goto new_bitmap; |
| 1842 | } |
| 1843 | |
| 1844 | bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes); |
| 1845 | bytes -= bytes_added; |
| 1846 | offset += bytes_added; |
| 1847 | added = 0; |
| 1848 | |
| 1849 | if (!bytes) { |
| 1850 | ret = 1; |
| 1851 | goto out; |
| 1852 | } else |
| 1853 | goto again; |
| 1854 | |
| 1855 | new_bitmap: |
| 1856 | if (info && info->bitmap) { |
| 1857 | add_new_bitmap(ctl, info, offset); |
| 1858 | added = 1; |
| 1859 | info = NULL; |
| 1860 | goto again; |
| 1861 | } else { |
| 1862 | spin_unlock(&ctl->tree_lock); |
| 1863 | |
| 1864 | /* no pre-allocated info, allocate a new one */ |
| 1865 | if (!info) { |
| 1866 | info = kmem_cache_zalloc(btrfs_free_space_cachep, |
| 1867 | GFP_NOFS); |
| 1868 | if (!info) { |
| 1869 | spin_lock(&ctl->tree_lock); |
| 1870 | ret = -ENOMEM; |
| 1871 | goto out; |
| 1872 | } |
| 1873 | } |
| 1874 | |
| 1875 | /* allocate the bitmap */ |
| 1876 | info->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); |
| 1877 | spin_lock(&ctl->tree_lock); |
| 1878 | if (!info->bitmap) { |
| 1879 | ret = -ENOMEM; |
| 1880 | goto out; |
| 1881 | } |
| 1882 | goto again; |
| 1883 | } |
| 1884 | |
| 1885 | out: |
| 1886 | if (info) { |
| 1887 | if (info->bitmap) |
| 1888 | kfree(info->bitmap); |
| 1889 | kmem_cache_free(btrfs_free_space_cachep, info); |
| 1890 | } |
| 1891 | |
| 1892 | return ret; |
| 1893 | } |
| 1894 | |
| 1895 | static bool try_merge_free_space(struct btrfs_free_space_ctl *ctl, |
| 1896 | struct btrfs_free_space *info, bool update_stat) |
| 1897 | { |
| 1898 | struct btrfs_free_space *left_info; |
| 1899 | struct btrfs_free_space *right_info; |
| 1900 | bool merged = false; |
| 1901 | u64 offset = info->offset; |
| 1902 | u64 bytes = info->bytes; |
| 1903 | |
| 1904 | /* |
| 1905 | * first we want to see if there is free space adjacent to the range we |
| 1906 | * are adding, if there is remove that struct and add a new one to |
| 1907 | * cover the entire range |
| 1908 | */ |
| 1909 | right_info = tree_search_offset(ctl, offset + bytes, 0, 0); |
| 1910 | if (right_info && rb_prev(&right_info->offset_index)) |
| 1911 | left_info = rb_entry(rb_prev(&right_info->offset_index), |
| 1912 | struct btrfs_free_space, offset_index); |
| 1913 | else |
| 1914 | left_info = tree_search_offset(ctl, offset - 1, 0, 0); |
| 1915 | |
| 1916 | if (right_info && !right_info->bitmap) { |
| 1917 | if (update_stat) |
| 1918 | unlink_free_space(ctl, right_info); |
| 1919 | else |
| 1920 | __unlink_free_space(ctl, right_info); |
| 1921 | info->bytes += right_info->bytes; |
| 1922 | kmem_cache_free(btrfs_free_space_cachep, right_info); |
| 1923 | merged = true; |
| 1924 | } |
| 1925 | |
| 1926 | if (left_info && !left_info->bitmap && |
| 1927 | left_info->offset + left_info->bytes == offset) { |
| 1928 | if (update_stat) |
| 1929 | unlink_free_space(ctl, left_info); |
| 1930 | else |
| 1931 | __unlink_free_space(ctl, left_info); |
| 1932 | info->offset = left_info->offset; |
| 1933 | info->bytes += left_info->bytes; |
| 1934 | kmem_cache_free(btrfs_free_space_cachep, left_info); |
| 1935 | merged = true; |
| 1936 | } |
| 1937 | |
| 1938 | return merged; |
| 1939 | } |
| 1940 | |
| 1941 | int __btrfs_add_free_space(struct btrfs_free_space_ctl *ctl, |
| 1942 | u64 offset, u64 bytes) |
| 1943 | { |
| 1944 | struct btrfs_free_space *info; |
| 1945 | int ret = 0; |
| 1946 | |
| 1947 | info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS); |
| 1948 | if (!info) |
| 1949 | return -ENOMEM; |
| 1950 | |
| 1951 | info->offset = offset; |
| 1952 | info->bytes = bytes; |
| 1953 | |
| 1954 | spin_lock(&ctl->tree_lock); |
| 1955 | |
| 1956 | if (try_merge_free_space(ctl, info, true)) |
| 1957 | goto link; |
| 1958 | |
| 1959 | /* |
| 1960 | * There was no extent directly to the left or right of this new |
| 1961 | * extent then we know we're going to have to allocate a new extent, so |
| 1962 | * before we do that see if we need to drop this into a bitmap |
| 1963 | */ |
| 1964 | ret = insert_into_bitmap(ctl, info); |
| 1965 | if (ret < 0) { |
| 1966 | goto out; |
| 1967 | } else if (ret) { |
| 1968 | ret = 0; |
| 1969 | goto out; |
| 1970 | } |
| 1971 | link: |
| 1972 | ret = link_free_space(ctl, info); |
| 1973 | if (ret) |
| 1974 | kmem_cache_free(btrfs_free_space_cachep, info); |
| 1975 | out: |
| 1976 | spin_unlock(&ctl->tree_lock); |
| 1977 | |
| 1978 | if (ret) { |
| 1979 | printk(KERN_CRIT "BTRFS: unable to add free space :%d\n", ret); |
| 1980 | ASSERT(ret != -EEXIST); |
| 1981 | } |
| 1982 | |
| 1983 | return ret; |
| 1984 | } |
| 1985 | |
| 1986 | int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group, |
| 1987 | u64 offset, u64 bytes) |
| 1988 | { |
| 1989 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 1990 | struct btrfs_free_space *info; |
| 1991 | int ret; |
| 1992 | bool re_search = false; |
| 1993 | |
| 1994 | spin_lock(&ctl->tree_lock); |
| 1995 | |
| 1996 | again: |
| 1997 | ret = 0; |
| 1998 | if (!bytes) |
| 1999 | goto out_lock; |
| 2000 | |
| 2001 | info = tree_search_offset(ctl, offset, 0, 0); |
| 2002 | if (!info) { |
| 2003 | /* |
| 2004 | * oops didn't find an extent that matched the space we wanted |
| 2005 | * to remove, look for a bitmap instead |
| 2006 | */ |
| 2007 | info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), |
| 2008 | 1, 0); |
| 2009 | if (!info) { |
| 2010 | /* |
| 2011 | * If we found a partial bit of our free space in a |
| 2012 | * bitmap but then couldn't find the other part this may |
| 2013 | * be a problem, so WARN about it. |
| 2014 | */ |
| 2015 | WARN_ON(re_search); |
| 2016 | goto out_lock; |
| 2017 | } |
| 2018 | } |
| 2019 | |
| 2020 | re_search = false; |
| 2021 | if (!info->bitmap) { |
| 2022 | unlink_free_space(ctl, info); |
| 2023 | if (offset == info->offset) { |
| 2024 | u64 to_free = min(bytes, info->bytes); |
| 2025 | |
| 2026 | info->bytes -= to_free; |
| 2027 | info->offset += to_free; |
| 2028 | if (info->bytes) { |
| 2029 | ret = link_free_space(ctl, info); |
| 2030 | WARN_ON(ret); |
| 2031 | } else { |
| 2032 | kmem_cache_free(btrfs_free_space_cachep, info); |
| 2033 | } |
| 2034 | |
| 2035 | offset += to_free; |
| 2036 | bytes -= to_free; |
| 2037 | goto again; |
| 2038 | } else { |
| 2039 | u64 old_end = info->bytes + info->offset; |
| 2040 | |
| 2041 | info->bytes = offset - info->offset; |
| 2042 | ret = link_free_space(ctl, info); |
| 2043 | WARN_ON(ret); |
| 2044 | if (ret) |
| 2045 | goto out_lock; |
| 2046 | |
| 2047 | /* Not enough bytes in this entry to satisfy us */ |
| 2048 | if (old_end < offset + bytes) { |
| 2049 | bytes -= old_end - offset; |
| 2050 | offset = old_end; |
| 2051 | goto again; |
| 2052 | } else if (old_end == offset + bytes) { |
| 2053 | /* all done */ |
| 2054 | goto out_lock; |
| 2055 | } |
| 2056 | spin_unlock(&ctl->tree_lock); |
| 2057 | |
| 2058 | ret = btrfs_add_free_space(block_group, offset + bytes, |
| 2059 | old_end - (offset + bytes)); |
| 2060 | WARN_ON(ret); |
| 2061 | goto out; |
| 2062 | } |
| 2063 | } |
| 2064 | |
| 2065 | ret = remove_from_bitmap(ctl, info, &offset, &bytes); |
| 2066 | if (ret == -EAGAIN) { |
| 2067 | re_search = true; |
| 2068 | goto again; |
| 2069 | } |
| 2070 | out_lock: |
| 2071 | spin_unlock(&ctl->tree_lock); |
| 2072 | out: |
| 2073 | return ret; |
| 2074 | } |
| 2075 | |
| 2076 | void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group, |
| 2077 | u64 bytes) |
| 2078 | { |
| 2079 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2080 | struct btrfs_free_space *info; |
| 2081 | struct rb_node *n; |
| 2082 | int count = 0; |
| 2083 | |
| 2084 | for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) { |
| 2085 | info = rb_entry(n, struct btrfs_free_space, offset_index); |
| 2086 | if (info->bytes >= bytes && !block_group->ro) |
| 2087 | count++; |
| 2088 | btrfs_crit(block_group->fs_info, |
| 2089 | "entry offset %llu, bytes %llu, bitmap %s", |
| 2090 | info->offset, info->bytes, |
| 2091 | (info->bitmap) ? "yes" : "no"); |
| 2092 | } |
| 2093 | btrfs_info(block_group->fs_info, "block group has cluster?: %s", |
| 2094 | list_empty(&block_group->cluster_list) ? "no" : "yes"); |
| 2095 | btrfs_info(block_group->fs_info, |
| 2096 | "%d blocks of free space at or bigger than bytes is", count); |
| 2097 | } |
| 2098 | |
| 2099 | void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group) |
| 2100 | { |
| 2101 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2102 | |
| 2103 | spin_lock_init(&ctl->tree_lock); |
| 2104 | ctl->unit = block_group->sectorsize; |
| 2105 | ctl->start = block_group->key.objectid; |
| 2106 | ctl->private = block_group; |
| 2107 | ctl->op = &free_space_op; |
| 2108 | |
| 2109 | /* |
| 2110 | * we only want to have 32k of ram per block group for keeping |
| 2111 | * track of free space, and if we pass 1/2 of that we want to |
| 2112 | * start converting things over to using bitmaps |
| 2113 | */ |
| 2114 | ctl->extents_thresh = ((1024 * 32) / 2) / |
| 2115 | sizeof(struct btrfs_free_space); |
| 2116 | } |
| 2117 | |
| 2118 | /* |
| 2119 | * for a given cluster, put all of its extents back into the free |
| 2120 | * space cache. If the block group passed doesn't match the block group |
| 2121 | * pointed to by the cluster, someone else raced in and freed the |
| 2122 | * cluster already. In that case, we just return without changing anything |
| 2123 | */ |
| 2124 | static int |
| 2125 | __btrfs_return_cluster_to_free_space( |
| 2126 | struct btrfs_block_group_cache *block_group, |
| 2127 | struct btrfs_free_cluster *cluster) |
| 2128 | { |
| 2129 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2130 | struct btrfs_free_space *entry; |
| 2131 | struct rb_node *node; |
| 2132 | |
| 2133 | spin_lock(&cluster->lock); |
| 2134 | if (cluster->block_group != block_group) |
| 2135 | goto out; |
| 2136 | |
| 2137 | cluster->block_group = NULL; |
| 2138 | cluster->window_start = 0; |
| 2139 | list_del_init(&cluster->block_group_list); |
| 2140 | |
| 2141 | node = rb_first(&cluster->root); |
| 2142 | while (node) { |
| 2143 | bool bitmap; |
| 2144 | |
| 2145 | entry = rb_entry(node, struct btrfs_free_space, offset_index); |
| 2146 | node = rb_next(&entry->offset_index); |
| 2147 | rb_erase(&entry->offset_index, &cluster->root); |
| 2148 | |
| 2149 | bitmap = (entry->bitmap != NULL); |
| 2150 | if (!bitmap) |
| 2151 | try_merge_free_space(ctl, entry, false); |
| 2152 | tree_insert_offset(&ctl->free_space_offset, |
| 2153 | entry->offset, &entry->offset_index, bitmap); |
| 2154 | } |
| 2155 | cluster->root = RB_ROOT; |
| 2156 | |
| 2157 | out: |
| 2158 | spin_unlock(&cluster->lock); |
| 2159 | btrfs_put_block_group(block_group); |
| 2160 | return 0; |
| 2161 | } |
| 2162 | |
| 2163 | static void __btrfs_remove_free_space_cache_locked( |
| 2164 | struct btrfs_free_space_ctl *ctl) |
| 2165 | { |
| 2166 | struct btrfs_free_space *info; |
| 2167 | struct rb_node *node; |
| 2168 | |
| 2169 | while ((node = rb_last(&ctl->free_space_offset)) != NULL) { |
| 2170 | info = rb_entry(node, struct btrfs_free_space, offset_index); |
| 2171 | if (!info->bitmap) { |
| 2172 | unlink_free_space(ctl, info); |
| 2173 | kmem_cache_free(btrfs_free_space_cachep, info); |
| 2174 | } else { |
| 2175 | free_bitmap(ctl, info); |
| 2176 | } |
| 2177 | if (need_resched()) { |
| 2178 | spin_unlock(&ctl->tree_lock); |
| 2179 | cond_resched(); |
| 2180 | spin_lock(&ctl->tree_lock); |
| 2181 | } |
| 2182 | } |
| 2183 | } |
| 2184 | |
| 2185 | void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl) |
| 2186 | { |
| 2187 | spin_lock(&ctl->tree_lock); |
| 2188 | __btrfs_remove_free_space_cache_locked(ctl); |
| 2189 | spin_unlock(&ctl->tree_lock); |
| 2190 | } |
| 2191 | |
| 2192 | void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group) |
| 2193 | { |
| 2194 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2195 | struct btrfs_free_cluster *cluster; |
| 2196 | struct list_head *head; |
| 2197 | |
| 2198 | spin_lock(&ctl->tree_lock); |
| 2199 | while ((head = block_group->cluster_list.next) != |
| 2200 | &block_group->cluster_list) { |
| 2201 | cluster = list_entry(head, struct btrfs_free_cluster, |
| 2202 | block_group_list); |
| 2203 | |
| 2204 | WARN_ON(cluster->block_group != block_group); |
| 2205 | __btrfs_return_cluster_to_free_space(block_group, cluster); |
| 2206 | if (need_resched()) { |
| 2207 | spin_unlock(&ctl->tree_lock); |
| 2208 | cond_resched(); |
| 2209 | spin_lock(&ctl->tree_lock); |
| 2210 | } |
| 2211 | } |
| 2212 | __btrfs_remove_free_space_cache_locked(ctl); |
| 2213 | spin_unlock(&ctl->tree_lock); |
| 2214 | |
| 2215 | } |
| 2216 | |
| 2217 | u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group, |
| 2218 | u64 offset, u64 bytes, u64 empty_size, |
| 2219 | u64 *max_extent_size) |
| 2220 | { |
| 2221 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2222 | struct btrfs_free_space *entry = NULL; |
| 2223 | u64 bytes_search = bytes + empty_size; |
| 2224 | u64 ret = 0; |
| 2225 | u64 align_gap = 0; |
| 2226 | u64 align_gap_len = 0; |
| 2227 | |
| 2228 | spin_lock(&ctl->tree_lock); |
| 2229 | entry = find_free_space(ctl, &offset, &bytes_search, |
| 2230 | block_group->full_stripe_len, max_extent_size); |
| 2231 | if (!entry) |
| 2232 | goto out; |
| 2233 | |
| 2234 | ret = offset; |
| 2235 | if (entry->bitmap) { |
| 2236 | bitmap_clear_bits(ctl, entry, offset, bytes); |
| 2237 | if (!entry->bytes) |
| 2238 | free_bitmap(ctl, entry); |
| 2239 | } else { |
| 2240 | unlink_free_space(ctl, entry); |
| 2241 | align_gap_len = offset - entry->offset; |
| 2242 | align_gap = entry->offset; |
| 2243 | |
| 2244 | entry->offset = offset + bytes; |
| 2245 | WARN_ON(entry->bytes < bytes + align_gap_len); |
| 2246 | |
| 2247 | entry->bytes -= bytes + align_gap_len; |
| 2248 | if (!entry->bytes) |
| 2249 | kmem_cache_free(btrfs_free_space_cachep, entry); |
| 2250 | else |
| 2251 | link_free_space(ctl, entry); |
| 2252 | } |
| 2253 | out: |
| 2254 | spin_unlock(&ctl->tree_lock); |
| 2255 | |
| 2256 | if (align_gap_len) |
| 2257 | __btrfs_add_free_space(ctl, align_gap, align_gap_len); |
| 2258 | return ret; |
| 2259 | } |
| 2260 | |
| 2261 | /* |
| 2262 | * given a cluster, put all of its extents back into the free space |
| 2263 | * cache. If a block group is passed, this function will only free |
| 2264 | * a cluster that belongs to the passed block group. |
| 2265 | * |
| 2266 | * Otherwise, it'll get a reference on the block group pointed to by the |
| 2267 | * cluster and remove the cluster from it. |
| 2268 | */ |
| 2269 | int btrfs_return_cluster_to_free_space( |
| 2270 | struct btrfs_block_group_cache *block_group, |
| 2271 | struct btrfs_free_cluster *cluster) |
| 2272 | { |
| 2273 | struct btrfs_free_space_ctl *ctl; |
| 2274 | int ret; |
| 2275 | |
| 2276 | /* first, get a safe pointer to the block group */ |
| 2277 | spin_lock(&cluster->lock); |
| 2278 | if (!block_group) { |
| 2279 | block_group = cluster->block_group; |
| 2280 | if (!block_group) { |
| 2281 | spin_unlock(&cluster->lock); |
| 2282 | return 0; |
| 2283 | } |
| 2284 | } else if (cluster->block_group != block_group) { |
| 2285 | /* someone else has already freed it don't redo their work */ |
| 2286 | spin_unlock(&cluster->lock); |
| 2287 | return 0; |
| 2288 | } |
| 2289 | atomic_inc(&block_group->count); |
| 2290 | spin_unlock(&cluster->lock); |
| 2291 | |
| 2292 | ctl = block_group->free_space_ctl; |
| 2293 | |
| 2294 | /* now return any extents the cluster had on it */ |
| 2295 | spin_lock(&ctl->tree_lock); |
| 2296 | ret = __btrfs_return_cluster_to_free_space(block_group, cluster); |
| 2297 | spin_unlock(&ctl->tree_lock); |
| 2298 | |
| 2299 | /* finally drop our ref */ |
| 2300 | btrfs_put_block_group(block_group); |
| 2301 | return ret; |
| 2302 | } |
| 2303 | |
| 2304 | static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group, |
| 2305 | struct btrfs_free_cluster *cluster, |
| 2306 | struct btrfs_free_space *entry, |
| 2307 | u64 bytes, u64 min_start, |
| 2308 | u64 *max_extent_size) |
| 2309 | { |
| 2310 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2311 | int err; |
| 2312 | u64 search_start = cluster->window_start; |
| 2313 | u64 search_bytes = bytes; |
| 2314 | u64 ret = 0; |
| 2315 | |
| 2316 | search_start = min_start; |
| 2317 | search_bytes = bytes; |
| 2318 | |
| 2319 | err = search_bitmap(ctl, entry, &search_start, &search_bytes); |
| 2320 | if (err) { |
| 2321 | if (search_bytes > *max_extent_size) |
| 2322 | *max_extent_size = search_bytes; |
| 2323 | return 0; |
| 2324 | } |
| 2325 | |
| 2326 | ret = search_start; |
| 2327 | __bitmap_clear_bits(ctl, entry, ret, bytes); |
| 2328 | |
| 2329 | return ret; |
| 2330 | } |
| 2331 | |
| 2332 | /* |
| 2333 | * given a cluster, try to allocate 'bytes' from it, returns 0 |
| 2334 | * if it couldn't find anything suitably large, or a logical disk offset |
| 2335 | * if things worked out |
| 2336 | */ |
| 2337 | u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, |
| 2338 | struct btrfs_free_cluster *cluster, u64 bytes, |
| 2339 | u64 min_start, u64 *max_extent_size) |
| 2340 | { |
| 2341 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2342 | struct btrfs_free_space *entry = NULL; |
| 2343 | struct rb_node *node; |
| 2344 | u64 ret = 0; |
| 2345 | |
| 2346 | spin_lock(&cluster->lock); |
| 2347 | if (bytes > cluster->max_size) |
| 2348 | goto out; |
| 2349 | |
| 2350 | if (cluster->block_group != block_group) |
| 2351 | goto out; |
| 2352 | |
| 2353 | node = rb_first(&cluster->root); |
| 2354 | if (!node) |
| 2355 | goto out; |
| 2356 | |
| 2357 | entry = rb_entry(node, struct btrfs_free_space, offset_index); |
| 2358 | while (1) { |
| 2359 | if (entry->bytes < bytes && entry->bytes > *max_extent_size) |
| 2360 | *max_extent_size = entry->bytes; |
| 2361 | |
| 2362 | if (entry->bytes < bytes || |
| 2363 | (!entry->bitmap && entry->offset < min_start)) { |
| 2364 | node = rb_next(&entry->offset_index); |
| 2365 | if (!node) |
| 2366 | break; |
| 2367 | entry = rb_entry(node, struct btrfs_free_space, |
| 2368 | offset_index); |
| 2369 | continue; |
| 2370 | } |
| 2371 | |
| 2372 | if (entry->bitmap) { |
| 2373 | ret = btrfs_alloc_from_bitmap(block_group, |
| 2374 | cluster, entry, bytes, |
| 2375 | cluster->window_start, |
| 2376 | max_extent_size); |
| 2377 | if (ret == 0) { |
| 2378 | node = rb_next(&entry->offset_index); |
| 2379 | if (!node) |
| 2380 | break; |
| 2381 | entry = rb_entry(node, struct btrfs_free_space, |
| 2382 | offset_index); |
| 2383 | continue; |
| 2384 | } |
| 2385 | cluster->window_start += bytes; |
| 2386 | } else { |
| 2387 | ret = entry->offset; |
| 2388 | |
| 2389 | entry->offset += bytes; |
| 2390 | entry->bytes -= bytes; |
| 2391 | } |
| 2392 | |
| 2393 | if (entry->bytes == 0) |
| 2394 | rb_erase(&entry->offset_index, &cluster->root); |
| 2395 | break; |
| 2396 | } |
| 2397 | out: |
| 2398 | spin_unlock(&cluster->lock); |
| 2399 | |
| 2400 | if (!ret) |
| 2401 | return 0; |
| 2402 | |
| 2403 | spin_lock(&ctl->tree_lock); |
| 2404 | |
| 2405 | ctl->free_space -= bytes; |
| 2406 | if (entry->bytes == 0) { |
| 2407 | ctl->free_extents--; |
| 2408 | if (entry->bitmap) { |
| 2409 | kfree(entry->bitmap); |
| 2410 | ctl->total_bitmaps--; |
| 2411 | ctl->op->recalc_thresholds(ctl); |
| 2412 | } |
| 2413 | kmem_cache_free(btrfs_free_space_cachep, entry); |
| 2414 | } |
| 2415 | |
| 2416 | spin_unlock(&ctl->tree_lock); |
| 2417 | |
| 2418 | return ret; |
| 2419 | } |
| 2420 | |
| 2421 | static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group, |
| 2422 | struct btrfs_free_space *entry, |
| 2423 | struct btrfs_free_cluster *cluster, |
| 2424 | u64 offset, u64 bytes, |
| 2425 | u64 cont1_bytes, u64 min_bytes) |
| 2426 | { |
| 2427 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2428 | unsigned long next_zero; |
| 2429 | unsigned long i; |
| 2430 | unsigned long want_bits; |
| 2431 | unsigned long min_bits; |
| 2432 | unsigned long found_bits; |
| 2433 | unsigned long start = 0; |
| 2434 | unsigned long total_found = 0; |
| 2435 | int ret; |
| 2436 | |
| 2437 | i = offset_to_bit(entry->offset, ctl->unit, |
| 2438 | max_t(u64, offset, entry->offset)); |
| 2439 | want_bits = bytes_to_bits(bytes, ctl->unit); |
| 2440 | min_bits = bytes_to_bits(min_bytes, ctl->unit); |
| 2441 | |
| 2442 | again: |
| 2443 | found_bits = 0; |
| 2444 | for_each_set_bit_from(i, entry->bitmap, BITS_PER_BITMAP) { |
| 2445 | next_zero = find_next_zero_bit(entry->bitmap, |
| 2446 | BITS_PER_BITMAP, i); |
| 2447 | if (next_zero - i >= min_bits) { |
| 2448 | found_bits = next_zero - i; |
| 2449 | break; |
| 2450 | } |
| 2451 | i = next_zero; |
| 2452 | } |
| 2453 | |
| 2454 | if (!found_bits) |
| 2455 | return -ENOSPC; |
| 2456 | |
| 2457 | if (!total_found) { |
| 2458 | start = i; |
| 2459 | cluster->max_size = 0; |
| 2460 | } |
| 2461 | |
| 2462 | total_found += found_bits; |
| 2463 | |
| 2464 | if (cluster->max_size < found_bits * ctl->unit) |
| 2465 | cluster->max_size = found_bits * ctl->unit; |
| 2466 | |
| 2467 | if (total_found < want_bits || cluster->max_size < cont1_bytes) { |
| 2468 | i = next_zero + 1; |
| 2469 | goto again; |
| 2470 | } |
| 2471 | |
| 2472 | cluster->window_start = start * ctl->unit + entry->offset; |
| 2473 | rb_erase(&entry->offset_index, &ctl->free_space_offset); |
| 2474 | ret = tree_insert_offset(&cluster->root, entry->offset, |
| 2475 | &entry->offset_index, 1); |
| 2476 | ASSERT(!ret); /* -EEXIST; Logic error */ |
| 2477 | |
| 2478 | trace_btrfs_setup_cluster(block_group, cluster, |
| 2479 | total_found * ctl->unit, 1); |
| 2480 | return 0; |
| 2481 | } |
| 2482 | |
| 2483 | /* |
| 2484 | * This searches the block group for just extents to fill the cluster with. |
| 2485 | * Try to find a cluster with at least bytes total bytes, at least one |
| 2486 | * extent of cont1_bytes, and other clusters of at least min_bytes. |
| 2487 | */ |
| 2488 | static noinline int |
| 2489 | setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, |
| 2490 | struct btrfs_free_cluster *cluster, |
| 2491 | struct list_head *bitmaps, u64 offset, u64 bytes, |
| 2492 | u64 cont1_bytes, u64 min_bytes) |
| 2493 | { |
| 2494 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2495 | struct btrfs_free_space *first = NULL; |
| 2496 | struct btrfs_free_space *entry = NULL; |
| 2497 | struct btrfs_free_space *last; |
| 2498 | struct rb_node *node; |
| 2499 | u64 window_free; |
| 2500 | u64 max_extent; |
| 2501 | u64 total_size = 0; |
| 2502 | |
| 2503 | entry = tree_search_offset(ctl, offset, 0, 1); |
| 2504 | if (!entry) |
| 2505 | return -ENOSPC; |
| 2506 | |
| 2507 | /* |
| 2508 | * We don't want bitmaps, so just move along until we find a normal |
| 2509 | * extent entry. |
| 2510 | */ |
| 2511 | while (entry->bitmap || entry->bytes < min_bytes) { |
| 2512 | if (entry->bitmap && list_empty(&entry->list)) |
| 2513 | list_add_tail(&entry->list, bitmaps); |
| 2514 | node = rb_next(&entry->offset_index); |
| 2515 | if (!node) |
| 2516 | return -ENOSPC; |
| 2517 | entry = rb_entry(node, struct btrfs_free_space, offset_index); |
| 2518 | } |
| 2519 | |
| 2520 | window_free = entry->bytes; |
| 2521 | max_extent = entry->bytes; |
| 2522 | first = entry; |
| 2523 | last = entry; |
| 2524 | |
| 2525 | for (node = rb_next(&entry->offset_index); node; |
| 2526 | node = rb_next(&entry->offset_index)) { |
| 2527 | entry = rb_entry(node, struct btrfs_free_space, offset_index); |
| 2528 | |
| 2529 | if (entry->bitmap) { |
| 2530 | if (list_empty(&entry->list)) |
| 2531 | list_add_tail(&entry->list, bitmaps); |
| 2532 | continue; |
| 2533 | } |
| 2534 | |
| 2535 | if (entry->bytes < min_bytes) |
| 2536 | continue; |
| 2537 | |
| 2538 | last = entry; |
| 2539 | window_free += entry->bytes; |
| 2540 | if (entry->bytes > max_extent) |
| 2541 | max_extent = entry->bytes; |
| 2542 | } |
| 2543 | |
| 2544 | if (window_free < bytes || max_extent < cont1_bytes) |
| 2545 | return -ENOSPC; |
| 2546 | |
| 2547 | cluster->window_start = first->offset; |
| 2548 | |
| 2549 | node = &first->offset_index; |
| 2550 | |
| 2551 | /* |
| 2552 | * now we've found our entries, pull them out of the free space |
| 2553 | * cache and put them into the cluster rbtree |
| 2554 | */ |
| 2555 | do { |
| 2556 | int ret; |
| 2557 | |
| 2558 | entry = rb_entry(node, struct btrfs_free_space, offset_index); |
| 2559 | node = rb_next(&entry->offset_index); |
| 2560 | if (entry->bitmap || entry->bytes < min_bytes) |
| 2561 | continue; |
| 2562 | |
| 2563 | rb_erase(&entry->offset_index, &ctl->free_space_offset); |
| 2564 | ret = tree_insert_offset(&cluster->root, entry->offset, |
| 2565 | &entry->offset_index, 0); |
| 2566 | total_size += entry->bytes; |
| 2567 | ASSERT(!ret); /* -EEXIST; Logic error */ |
| 2568 | } while (node && entry != last); |
| 2569 | |
| 2570 | cluster->max_size = max_extent; |
| 2571 | trace_btrfs_setup_cluster(block_group, cluster, total_size, 0); |
| 2572 | return 0; |
| 2573 | } |
| 2574 | |
| 2575 | /* |
| 2576 | * This specifically looks for bitmaps that may work in the cluster, we assume |
| 2577 | * that we have already failed to find extents that will work. |
| 2578 | */ |
| 2579 | static noinline int |
| 2580 | setup_cluster_bitmap(struct btrfs_block_group_cache *block_group, |
| 2581 | struct btrfs_free_cluster *cluster, |
| 2582 | struct list_head *bitmaps, u64 offset, u64 bytes, |
| 2583 | u64 cont1_bytes, u64 min_bytes) |
| 2584 | { |
| 2585 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2586 | struct btrfs_free_space *entry; |
| 2587 | int ret = -ENOSPC; |
| 2588 | u64 bitmap_offset = offset_to_bitmap(ctl, offset); |
| 2589 | |
| 2590 | if (ctl->total_bitmaps == 0) |
| 2591 | return -ENOSPC; |
| 2592 | |
| 2593 | /* |
| 2594 | * The bitmap that covers offset won't be in the list unless offset |
| 2595 | * is just its start offset. |
| 2596 | */ |
| 2597 | entry = list_first_entry(bitmaps, struct btrfs_free_space, list); |
| 2598 | if (entry->offset != bitmap_offset) { |
| 2599 | entry = tree_search_offset(ctl, bitmap_offset, 1, 0); |
| 2600 | if (entry && list_empty(&entry->list)) |
| 2601 | list_add(&entry->list, bitmaps); |
| 2602 | } |
| 2603 | |
| 2604 | list_for_each_entry(entry, bitmaps, list) { |
| 2605 | if (entry->bytes < bytes) |
| 2606 | continue; |
| 2607 | ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset, |
| 2608 | bytes, cont1_bytes, min_bytes); |
| 2609 | if (!ret) |
| 2610 | return 0; |
| 2611 | } |
| 2612 | |
| 2613 | /* |
| 2614 | * The bitmaps list has all the bitmaps that record free space |
| 2615 | * starting after offset, so no more search is required. |
| 2616 | */ |
| 2617 | return -ENOSPC; |
| 2618 | } |
| 2619 | |
| 2620 | /* |
| 2621 | * here we try to find a cluster of blocks in a block group. The goal |
| 2622 | * is to find at least bytes+empty_size. |
| 2623 | * We might not find them all in one contiguous area. |
| 2624 | * |
| 2625 | * returns zero and sets up cluster if things worked out, otherwise |
| 2626 | * it returns -enospc |
| 2627 | */ |
| 2628 | int btrfs_find_space_cluster(struct btrfs_root *root, |
| 2629 | struct btrfs_block_group_cache *block_group, |
| 2630 | struct btrfs_free_cluster *cluster, |
| 2631 | u64 offset, u64 bytes, u64 empty_size) |
| 2632 | { |
| 2633 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2634 | struct btrfs_free_space *entry, *tmp; |
| 2635 | LIST_HEAD(bitmaps); |
| 2636 | u64 min_bytes; |
| 2637 | u64 cont1_bytes; |
| 2638 | int ret; |
| 2639 | |
| 2640 | /* |
| 2641 | * Choose the minimum extent size we'll require for this |
| 2642 | * cluster. For SSD_SPREAD, don't allow any fragmentation. |
| 2643 | * For metadata, allow allocates with smaller extents. For |
| 2644 | * data, keep it dense. |
| 2645 | */ |
| 2646 | if (btrfs_test_opt(root, SSD_SPREAD)) { |
| 2647 | cont1_bytes = min_bytes = bytes + empty_size; |
| 2648 | } else if (block_group->flags & BTRFS_BLOCK_GROUP_METADATA) { |
| 2649 | cont1_bytes = bytes; |
| 2650 | min_bytes = block_group->sectorsize; |
| 2651 | } else { |
| 2652 | cont1_bytes = max(bytes, (bytes + empty_size) >> 2); |
| 2653 | min_bytes = block_group->sectorsize; |
| 2654 | } |
| 2655 | |
| 2656 | spin_lock(&ctl->tree_lock); |
| 2657 | |
| 2658 | /* |
| 2659 | * If we know we don't have enough space to make a cluster don't even |
| 2660 | * bother doing all the work to try and find one. |
| 2661 | */ |
| 2662 | if (ctl->free_space < bytes) { |
| 2663 | spin_unlock(&ctl->tree_lock); |
| 2664 | return -ENOSPC; |
| 2665 | } |
| 2666 | |
| 2667 | spin_lock(&cluster->lock); |
| 2668 | |
| 2669 | /* someone already found a cluster, hooray */ |
| 2670 | if (cluster->block_group) { |
| 2671 | ret = 0; |
| 2672 | goto out; |
| 2673 | } |
| 2674 | |
| 2675 | trace_btrfs_find_cluster(block_group, offset, bytes, empty_size, |
| 2676 | min_bytes); |
| 2677 | |
| 2678 | INIT_LIST_HEAD(&bitmaps); |
| 2679 | ret = setup_cluster_no_bitmap(block_group, cluster, &bitmaps, offset, |
| 2680 | bytes + empty_size, |
| 2681 | cont1_bytes, min_bytes); |
| 2682 | if (ret) |
| 2683 | ret = setup_cluster_bitmap(block_group, cluster, &bitmaps, |
| 2684 | offset, bytes + empty_size, |
| 2685 | cont1_bytes, min_bytes); |
| 2686 | |
| 2687 | /* Clear our temporary list */ |
| 2688 | list_for_each_entry_safe(entry, tmp, &bitmaps, list) |
| 2689 | list_del_init(&entry->list); |
| 2690 | |
| 2691 | if (!ret) { |
| 2692 | atomic_inc(&block_group->count); |
| 2693 | list_add_tail(&cluster->block_group_list, |
| 2694 | &block_group->cluster_list); |
| 2695 | cluster->block_group = block_group; |
| 2696 | } else { |
| 2697 | trace_btrfs_failed_cluster_setup(block_group); |
| 2698 | } |
| 2699 | out: |
| 2700 | spin_unlock(&cluster->lock); |
| 2701 | spin_unlock(&ctl->tree_lock); |
| 2702 | |
| 2703 | return ret; |
| 2704 | } |
| 2705 | |
| 2706 | /* |
| 2707 | * simple code to zero out a cluster |
| 2708 | */ |
| 2709 | void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster) |
| 2710 | { |
| 2711 | spin_lock_init(&cluster->lock); |
| 2712 | spin_lock_init(&cluster->refill_lock); |
| 2713 | cluster->root = RB_ROOT; |
| 2714 | cluster->max_size = 0; |
| 2715 | INIT_LIST_HEAD(&cluster->block_group_list); |
| 2716 | cluster->block_group = NULL; |
| 2717 | } |
| 2718 | |
| 2719 | static int do_trimming(struct btrfs_block_group_cache *block_group, |
| 2720 | u64 *total_trimmed, u64 start, u64 bytes, |
| 2721 | u64 reserved_start, u64 reserved_bytes) |
| 2722 | { |
| 2723 | struct btrfs_space_info *space_info = block_group->space_info; |
| 2724 | struct btrfs_fs_info *fs_info = block_group->fs_info; |
| 2725 | int ret; |
| 2726 | int update = 0; |
| 2727 | u64 trimmed = 0; |
| 2728 | |
| 2729 | spin_lock(&space_info->lock); |
| 2730 | spin_lock(&block_group->lock); |
| 2731 | if (!block_group->ro) { |
| 2732 | block_group->reserved += reserved_bytes; |
| 2733 | space_info->bytes_reserved += reserved_bytes; |
| 2734 | update = 1; |
| 2735 | } |
| 2736 | spin_unlock(&block_group->lock); |
| 2737 | spin_unlock(&space_info->lock); |
| 2738 | |
| 2739 | ret = btrfs_error_discard_extent(fs_info->extent_root, |
| 2740 | start, bytes, &trimmed); |
| 2741 | if (!ret) |
| 2742 | *total_trimmed += trimmed; |
| 2743 | |
| 2744 | btrfs_add_free_space(block_group, reserved_start, reserved_bytes); |
| 2745 | |
| 2746 | if (update) { |
| 2747 | spin_lock(&space_info->lock); |
| 2748 | spin_lock(&block_group->lock); |
| 2749 | if (block_group->ro) |
| 2750 | space_info->bytes_readonly += reserved_bytes; |
| 2751 | block_group->reserved -= reserved_bytes; |
| 2752 | space_info->bytes_reserved -= reserved_bytes; |
| 2753 | spin_unlock(&space_info->lock); |
| 2754 | spin_unlock(&block_group->lock); |
| 2755 | } |
| 2756 | |
| 2757 | return ret; |
| 2758 | } |
| 2759 | |
| 2760 | static int trim_no_bitmap(struct btrfs_block_group_cache *block_group, |
| 2761 | u64 *total_trimmed, u64 start, u64 end, u64 minlen) |
| 2762 | { |
| 2763 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2764 | struct btrfs_free_space *entry; |
| 2765 | struct rb_node *node; |
| 2766 | int ret = 0; |
| 2767 | u64 extent_start; |
| 2768 | u64 extent_bytes; |
| 2769 | u64 bytes; |
| 2770 | |
| 2771 | while (start < end) { |
| 2772 | spin_lock(&ctl->tree_lock); |
| 2773 | |
| 2774 | if (ctl->free_space < minlen) { |
| 2775 | spin_unlock(&ctl->tree_lock); |
| 2776 | break; |
| 2777 | } |
| 2778 | |
| 2779 | entry = tree_search_offset(ctl, start, 0, 1); |
| 2780 | if (!entry) { |
| 2781 | spin_unlock(&ctl->tree_lock); |
| 2782 | break; |
| 2783 | } |
| 2784 | |
| 2785 | /* skip bitmaps */ |
| 2786 | while (entry->bitmap) { |
| 2787 | node = rb_next(&entry->offset_index); |
| 2788 | if (!node) { |
| 2789 | spin_unlock(&ctl->tree_lock); |
| 2790 | goto out; |
| 2791 | } |
| 2792 | entry = rb_entry(node, struct btrfs_free_space, |
| 2793 | offset_index); |
| 2794 | } |
| 2795 | |
| 2796 | if (entry->offset >= end) { |
| 2797 | spin_unlock(&ctl->tree_lock); |
| 2798 | break; |
| 2799 | } |
| 2800 | |
| 2801 | extent_start = entry->offset; |
| 2802 | extent_bytes = entry->bytes; |
| 2803 | start = max(start, extent_start); |
| 2804 | bytes = min(extent_start + extent_bytes, end) - start; |
| 2805 | if (bytes < minlen) { |
| 2806 | spin_unlock(&ctl->tree_lock); |
| 2807 | goto next; |
| 2808 | } |
| 2809 | |
| 2810 | unlink_free_space(ctl, entry); |
| 2811 | kmem_cache_free(btrfs_free_space_cachep, entry); |
| 2812 | |
| 2813 | spin_unlock(&ctl->tree_lock); |
| 2814 | |
| 2815 | ret = do_trimming(block_group, total_trimmed, start, bytes, |
| 2816 | extent_start, extent_bytes); |
| 2817 | if (ret) |
| 2818 | break; |
| 2819 | next: |
| 2820 | start += bytes; |
| 2821 | |
| 2822 | if (fatal_signal_pending(current)) { |
| 2823 | ret = -ERESTARTSYS; |
| 2824 | break; |
| 2825 | } |
| 2826 | |
| 2827 | cond_resched(); |
| 2828 | } |
| 2829 | out: |
| 2830 | return ret; |
| 2831 | } |
| 2832 | |
| 2833 | static int trim_bitmaps(struct btrfs_block_group_cache *block_group, |
| 2834 | u64 *total_trimmed, u64 start, u64 end, u64 minlen) |
| 2835 | { |
| 2836 | struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; |
| 2837 | struct btrfs_free_space *entry; |
| 2838 | int ret = 0; |
| 2839 | int ret2; |
| 2840 | u64 bytes; |
| 2841 | u64 offset = offset_to_bitmap(ctl, start); |
| 2842 | |
| 2843 | while (offset < end) { |
| 2844 | bool next_bitmap = false; |
| 2845 | |
| 2846 | spin_lock(&ctl->tree_lock); |
| 2847 | |
| 2848 | if (ctl->free_space < minlen) { |
| 2849 | spin_unlock(&ctl->tree_lock); |
| 2850 | break; |
| 2851 | } |
| 2852 | |
| 2853 | entry = tree_search_offset(ctl, offset, 1, 0); |
| 2854 | if (!entry) { |
| 2855 | spin_unlock(&ctl->tree_lock); |
| 2856 | next_bitmap = true; |
| 2857 | goto next; |
| 2858 | } |
| 2859 | |
| 2860 | bytes = minlen; |
| 2861 | ret2 = search_bitmap(ctl, entry, &start, &bytes); |
| 2862 | if (ret2 || start >= end) { |
| 2863 | spin_unlock(&ctl->tree_lock); |
| 2864 | next_bitmap = true; |
| 2865 | goto next; |
| 2866 | } |
| 2867 | |
| 2868 | bytes = min(bytes, end - start); |
| 2869 | if (bytes < minlen) { |
| 2870 | spin_unlock(&ctl->tree_lock); |
| 2871 | goto next; |
| 2872 | } |
| 2873 | |
| 2874 | bitmap_clear_bits(ctl, entry, start, bytes); |
| 2875 | if (entry->bytes == 0) |
| 2876 | free_bitmap(ctl, entry); |
| 2877 | |
| 2878 | spin_unlock(&ctl->tree_lock); |
| 2879 | |
| 2880 | ret = do_trimming(block_group, total_trimmed, start, bytes, |
| 2881 | start, bytes); |
| 2882 | if (ret) |
| 2883 | break; |
| 2884 | next: |
| 2885 | if (next_bitmap) { |
| 2886 | offset += BITS_PER_BITMAP * ctl->unit; |
| 2887 | } else { |
| 2888 | start += bytes; |
| 2889 | if (start >= offset + BITS_PER_BITMAP * ctl->unit) |
| 2890 | offset += BITS_PER_BITMAP * ctl->unit; |
| 2891 | } |
| 2892 | |
| 2893 | if (fatal_signal_pending(current)) { |
| 2894 | ret = -ERESTARTSYS; |
| 2895 | break; |
| 2896 | } |
| 2897 | |
| 2898 | cond_resched(); |
| 2899 | } |
| 2900 | |
| 2901 | return ret; |
| 2902 | } |
| 2903 | |
| 2904 | int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, |
| 2905 | u64 *trimmed, u64 start, u64 end, u64 minlen) |
| 2906 | { |
| 2907 | int ret; |
| 2908 | |
| 2909 | *trimmed = 0; |
| 2910 | |
| 2911 | ret = trim_no_bitmap(block_group, trimmed, start, end, minlen); |
| 2912 | if (ret) |
| 2913 | return ret; |
| 2914 | |
| 2915 | ret = trim_bitmaps(block_group, trimmed, start, end, minlen); |
| 2916 | |
| 2917 | return ret; |
| 2918 | } |
| 2919 | |
| 2920 | /* |
| 2921 | * Find the left-most item in the cache tree, and then return the |
| 2922 | * smallest inode number in the item. |
| 2923 | * |
| 2924 | * Note: the returned inode number may not be the smallest one in |
| 2925 | * the tree, if the left-most item is a bitmap. |
| 2926 | */ |
| 2927 | u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root) |
| 2928 | { |
| 2929 | struct btrfs_free_space_ctl *ctl = fs_root->free_ino_ctl; |
| 2930 | struct btrfs_free_space *entry = NULL; |
| 2931 | u64 ino = 0; |
| 2932 | |
| 2933 | spin_lock(&ctl->tree_lock); |
| 2934 | |
| 2935 | if (RB_EMPTY_ROOT(&ctl->free_space_offset)) |
| 2936 | goto out; |
| 2937 | |
| 2938 | entry = rb_entry(rb_first(&ctl->free_space_offset), |
| 2939 | struct btrfs_free_space, offset_index); |
| 2940 | |
| 2941 | if (!entry->bitmap) { |
| 2942 | ino = entry->offset; |
| 2943 | |
| 2944 | unlink_free_space(ctl, entry); |
| 2945 | entry->offset++; |
| 2946 | entry->bytes--; |
| 2947 | if (!entry->bytes) |
| 2948 | kmem_cache_free(btrfs_free_space_cachep, entry); |
| 2949 | else |
| 2950 | link_free_space(ctl, entry); |
| 2951 | } else { |
| 2952 | u64 offset = 0; |
| 2953 | u64 count = 1; |
| 2954 | int ret; |
| 2955 | |
| 2956 | ret = search_bitmap(ctl, entry, &offset, &count); |
| 2957 | /* Logic error; Should be empty if it can't find anything */ |
| 2958 | ASSERT(!ret); |
| 2959 | |
| 2960 | ino = offset; |
| 2961 | bitmap_clear_bits(ctl, entry, offset, 1); |
| 2962 | if (entry->bytes == 0) |
| 2963 | free_bitmap(ctl, entry); |
| 2964 | } |
| 2965 | out: |
| 2966 | spin_unlock(&ctl->tree_lock); |
| 2967 | |
| 2968 | return ino; |
| 2969 | } |
| 2970 | |
| 2971 | struct inode *lookup_free_ino_inode(struct btrfs_root *root, |
| 2972 | struct btrfs_path *path) |
| 2973 | { |
| 2974 | struct inode *inode = NULL; |
| 2975 | |
| 2976 | spin_lock(&root->cache_lock); |
| 2977 | if (root->cache_inode) |
| 2978 | inode = igrab(root->cache_inode); |
| 2979 | spin_unlock(&root->cache_lock); |
| 2980 | if (inode) |
| 2981 | return inode; |
| 2982 | |
| 2983 | inode = __lookup_free_space_inode(root, path, 0); |
| 2984 | if (IS_ERR(inode)) |
| 2985 | return inode; |
| 2986 | |
| 2987 | spin_lock(&root->cache_lock); |
| 2988 | if (!btrfs_fs_closing(root->fs_info)) |
| 2989 | root->cache_inode = igrab(inode); |
| 2990 | spin_unlock(&root->cache_lock); |
| 2991 | |
| 2992 | return inode; |
| 2993 | } |
| 2994 | |
| 2995 | int create_free_ino_inode(struct btrfs_root *root, |
| 2996 | struct btrfs_trans_handle *trans, |
| 2997 | struct btrfs_path *path) |
| 2998 | { |
| 2999 | return __create_free_space_inode(root, trans, path, |
| 3000 | BTRFS_FREE_INO_OBJECTID, 0); |
| 3001 | } |
| 3002 | |
| 3003 | int load_free_ino_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root) |
| 3004 | { |
| 3005 | struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; |
| 3006 | struct btrfs_path *path; |
| 3007 | struct inode *inode; |
| 3008 | int ret = 0; |
| 3009 | u64 root_gen = btrfs_root_generation(&root->root_item); |
| 3010 | |
| 3011 | if (!btrfs_test_opt(root, INODE_MAP_CACHE)) |
| 3012 | return 0; |
| 3013 | |
| 3014 | /* |
| 3015 | * If we're unmounting then just return, since this does a search on the |
| 3016 | * normal root and not the commit root and we could deadlock. |
| 3017 | */ |
| 3018 | if (btrfs_fs_closing(fs_info)) |
| 3019 | return 0; |
| 3020 | |
| 3021 | path = btrfs_alloc_path(); |
| 3022 | if (!path) |
| 3023 | return 0; |
| 3024 | |
| 3025 | inode = lookup_free_ino_inode(root, path); |
| 3026 | if (IS_ERR(inode)) |
| 3027 | goto out; |
| 3028 | |
| 3029 | if (root_gen != BTRFS_I(inode)->generation) |
| 3030 | goto out_put; |
| 3031 | |
| 3032 | ret = __load_free_space_cache(root, inode, ctl, path, 0); |
| 3033 | |
| 3034 | if (ret < 0) |
| 3035 | btrfs_err(fs_info, |
| 3036 | "failed to load free ino cache for root %llu", |
| 3037 | root->root_key.objectid); |
| 3038 | out_put: |
| 3039 | iput(inode); |
| 3040 | out: |
| 3041 | btrfs_free_path(path); |
| 3042 | return ret; |
| 3043 | } |
| 3044 | |
| 3045 | int btrfs_write_out_ino_cache(struct btrfs_root *root, |
| 3046 | struct btrfs_trans_handle *trans, |
| 3047 | struct btrfs_path *path, |
| 3048 | struct inode *inode) |
| 3049 | { |
| 3050 | struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; |
| 3051 | int ret; |
| 3052 | |
| 3053 | if (!btrfs_test_opt(root, INODE_MAP_CACHE)) |
| 3054 | return 0; |
| 3055 | |
| 3056 | ret = __btrfs_write_out_cache(root, inode, ctl, NULL, trans, path, 0); |
| 3057 | if (ret) { |
| 3058 | btrfs_delalloc_release_metadata(inode, inode->i_size); |
| 3059 | #ifdef DEBUG |
| 3060 | btrfs_err(root->fs_info, |
| 3061 | "failed to write free ino cache for root %llu", |
| 3062 | root->root_key.objectid); |
| 3063 | #endif |
| 3064 | } |
| 3065 | |
| 3066 | return ret; |
| 3067 | } |
| 3068 | |
| 3069 | #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS |
| 3070 | /* |
| 3071 | * Use this if you need to make a bitmap or extent entry specifically, it |
| 3072 | * doesn't do any of the merging that add_free_space does, this acts a lot like |
| 3073 | * how the free space cache loading stuff works, so you can get really weird |
| 3074 | * configurations. |
| 3075 | */ |
| 3076 | int test_add_free_space_entry(struct btrfs_block_group_cache *cache, |
| 3077 | u64 offset, u64 bytes, bool bitmap) |
| 3078 | { |
| 3079 | struct btrfs_free_space_ctl *ctl = cache->free_space_ctl; |
| 3080 | struct btrfs_free_space *info = NULL, *bitmap_info; |
| 3081 | void *map = NULL; |
| 3082 | u64 bytes_added; |
| 3083 | int ret; |
| 3084 | |
| 3085 | again: |
| 3086 | if (!info) { |
| 3087 | info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS); |
| 3088 | if (!info) |
| 3089 | return -ENOMEM; |
| 3090 | } |
| 3091 | |
| 3092 | if (!bitmap) { |
| 3093 | spin_lock(&ctl->tree_lock); |
| 3094 | info->offset = offset; |
| 3095 | info->bytes = bytes; |
| 3096 | ret = link_free_space(ctl, info); |
| 3097 | spin_unlock(&ctl->tree_lock); |
| 3098 | if (ret) |
| 3099 | kmem_cache_free(btrfs_free_space_cachep, info); |
| 3100 | return ret; |
| 3101 | } |
| 3102 | |
| 3103 | if (!map) { |
| 3104 | map = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); |
| 3105 | if (!map) { |
| 3106 | kmem_cache_free(btrfs_free_space_cachep, info); |
| 3107 | return -ENOMEM; |
| 3108 | } |
| 3109 | } |
| 3110 | |
| 3111 | spin_lock(&ctl->tree_lock); |
| 3112 | bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), |
| 3113 | 1, 0); |
| 3114 | if (!bitmap_info) { |
| 3115 | info->bitmap = map; |
| 3116 | map = NULL; |
| 3117 | add_new_bitmap(ctl, info, offset); |
| 3118 | bitmap_info = info; |
| 3119 | } |
| 3120 | |
| 3121 | bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes); |
| 3122 | bytes -= bytes_added; |
| 3123 | offset += bytes_added; |
| 3124 | spin_unlock(&ctl->tree_lock); |
| 3125 | |
| 3126 | if (bytes) |
| 3127 | goto again; |
| 3128 | |
| 3129 | if (map) |
| 3130 | kfree(map); |
| 3131 | return 0; |
| 3132 | } |
| 3133 | |
| 3134 | /* |
| 3135 | * Checks to see if the given range is in the free space cache. This is really |
| 3136 | * just used to check the absence of space, so if there is free space in the |
| 3137 | * range at all we will return 1. |
| 3138 | */ |
| 3139 | int test_check_exists(struct btrfs_block_group_cache *cache, |
| 3140 | u64 offset, u64 bytes) |
| 3141 | { |
| 3142 | struct btrfs_free_space_ctl *ctl = cache->free_space_ctl; |
| 3143 | struct btrfs_free_space *info; |
| 3144 | int ret = 0; |
| 3145 | |
| 3146 | spin_lock(&ctl->tree_lock); |
| 3147 | info = tree_search_offset(ctl, offset, 0, 0); |
| 3148 | if (!info) { |
| 3149 | info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), |
| 3150 | 1, 0); |
| 3151 | if (!info) |
| 3152 | goto out; |
| 3153 | } |
| 3154 | |
| 3155 | have_info: |
| 3156 | if (info->bitmap) { |
| 3157 | u64 bit_off, bit_bytes; |
| 3158 | struct rb_node *n; |
| 3159 | struct btrfs_free_space *tmp; |
| 3160 | |
| 3161 | bit_off = offset; |
| 3162 | bit_bytes = ctl->unit; |
| 3163 | ret = search_bitmap(ctl, info, &bit_off, &bit_bytes); |
| 3164 | if (!ret) { |
| 3165 | if (bit_off == offset) { |
| 3166 | ret = 1; |
| 3167 | goto out; |
| 3168 | } else if (bit_off > offset && |
| 3169 | offset + bytes > bit_off) { |
| 3170 | ret = 1; |
| 3171 | goto out; |
| 3172 | } |
| 3173 | } |
| 3174 | |
| 3175 | n = rb_prev(&info->offset_index); |
| 3176 | while (n) { |
| 3177 | tmp = rb_entry(n, struct btrfs_free_space, |
| 3178 | offset_index); |
| 3179 | if (tmp->offset + tmp->bytes < offset) |
| 3180 | break; |
| 3181 | if (offset + bytes < tmp->offset) { |
| 3182 | n = rb_prev(&info->offset_index); |
| 3183 | continue; |
| 3184 | } |
| 3185 | info = tmp; |
| 3186 | goto have_info; |
| 3187 | } |
| 3188 | |
| 3189 | n = rb_next(&info->offset_index); |
| 3190 | while (n) { |
| 3191 | tmp = rb_entry(n, struct btrfs_free_space, |
| 3192 | offset_index); |
| 3193 | if (offset + bytes < tmp->offset) |
| 3194 | break; |
| 3195 | if (tmp->offset + tmp->bytes < offset) { |
| 3196 | n = rb_next(&info->offset_index); |
| 3197 | continue; |
| 3198 | } |
| 3199 | info = tmp; |
| 3200 | goto have_info; |
| 3201 | } |
| 3202 | |
| 3203 | goto out; |
| 3204 | } |
| 3205 | |
| 3206 | if (info->offset == offset) { |
| 3207 | ret = 1; |
| 3208 | goto out; |
| 3209 | } |
| 3210 | |
| 3211 | if (offset > info->offset && offset < info->offset + info->bytes) |
| 3212 | ret = 1; |
| 3213 | out: |
| 3214 | spin_unlock(&ctl->tree_lock); |
| 3215 | return ret; |
| 3216 | } |
| 3217 | #endif /* CONFIG_BTRFS_FS_RUN_SANITY_TESTS */ |