1 /* 2 * Copyright (C) 2007 Oracle. 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/slab.h> 20 #include <linux/blkdev.h> 21 #include <linux/writeback.h> 22 #include <linux/pagevec.h> 23 #include "ctree.h" 24 #include "transaction.h" 25 #include "btrfs_inode.h" 26 #include "extent_io.h" 27 #include "disk-io.h" 28 29 static struct kmem_cache *btrfs_ordered_extent_cache; 30 31 static u64 entry_end(struct btrfs_ordered_extent *entry) 32 { 33 if (entry->file_offset + entry->len < entry->file_offset) 34 return (u64)-1; 35 return entry->file_offset + entry->len; 36 } 37 38 /* returns NULL if the insertion worked, or it returns the node it did find 39 * in the tree 40 */ 41 static struct rb_node *tree_insert(struct rb_root *root, u64 file_offset, 42 struct rb_node *node) 43 { 44 struct rb_node **p = &root->rb_node; 45 struct rb_node *parent = NULL; 46 struct btrfs_ordered_extent *entry; 47 48 while (*p) { 49 parent = *p; 50 entry = rb_entry(parent, struct btrfs_ordered_extent, rb_node); 51 52 if (file_offset < entry->file_offset) 53 p = &(*p)->rb_left; 54 else if (file_offset >= entry_end(entry)) 55 p = &(*p)->rb_right; 56 else 57 return parent; 58 } 59 60 rb_link_node(node, parent, p); 61 rb_insert_color(node, root); 62 return NULL; 63 } 64 65 static void ordered_data_tree_panic(struct inode *inode, int errno, 66 u64 offset) 67 { 68 struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); 69 btrfs_panic(fs_info, errno, "Inconsistency in ordered tree at offset " 70 "%llu", offset); 71 } 72 73 /* 74 * look for a given offset in the tree, and if it can't be found return the 75 * first lesser offset 76 */ 77 static struct rb_node *__tree_search(struct rb_root *root, u64 file_offset, 78 struct rb_node **prev_ret) 79 { 80 struct rb_node *n = root->rb_node; 81 struct rb_node *prev = NULL; 82 struct rb_node *test; 83 struct btrfs_ordered_extent *entry; 84 struct btrfs_ordered_extent *prev_entry = NULL; 85 86 while (n) { 87 entry = rb_entry(n, struct btrfs_ordered_extent, rb_node); 88 prev = n; 89 prev_entry = entry; 90 91 if (file_offset < entry->file_offset) 92 n = n->rb_left; 93 else if (file_offset >= entry_end(entry)) 94 n = n->rb_right; 95 else 96 return n; 97 } 98 if (!prev_ret) 99 return NULL; 100 101 while (prev && file_offset >= entry_end(prev_entry)) { 102 test = rb_next(prev); 103 if (!test) 104 break; 105 prev_entry = rb_entry(test, struct btrfs_ordered_extent, 106 rb_node); 107 if (file_offset < entry_end(prev_entry)) 108 break; 109 110 prev = test; 111 } 112 if (prev) 113 prev_entry = rb_entry(prev, struct btrfs_ordered_extent, 114 rb_node); 115 while (prev && file_offset < entry_end(prev_entry)) { 116 test = rb_prev(prev); 117 if (!test) 118 break; 119 prev_entry = rb_entry(test, struct btrfs_ordered_extent, 120 rb_node); 121 prev = test; 122 } 123 *prev_ret = prev; 124 return NULL; 125 } 126 127 /* 128 * helper to check if a given offset is inside a given entry 129 */ 130 static int offset_in_entry(struct btrfs_ordered_extent *entry, u64 file_offset) 131 { 132 if (file_offset < entry->file_offset || 133 entry->file_offset + entry->len <= file_offset) 134 return 0; 135 return 1; 136 } 137 138 static int range_overlaps(struct btrfs_ordered_extent *entry, u64 file_offset, 139 u64 len) 140 { 141 if (file_offset + len <= entry->file_offset || 142 entry->file_offset + entry->len <= file_offset) 143 return 0; 144 return 1; 145 } 146 147 /* 148 * look find the first ordered struct that has this offset, otherwise 149 * the first one less than this offset 150 */ 151 static inline struct rb_node *tree_search(struct btrfs_ordered_inode_tree *tree, 152 u64 file_offset) 153 { 154 struct rb_root *root = &tree->tree; 155 struct rb_node *prev = NULL; 156 struct rb_node *ret; 157 struct btrfs_ordered_extent *entry; 158 159 if (tree->last) { 160 entry = rb_entry(tree->last, struct btrfs_ordered_extent, 161 rb_node); 162 if (offset_in_entry(entry, file_offset)) 163 return tree->last; 164 } 165 ret = __tree_search(root, file_offset, &prev); 166 if (!ret) 167 ret = prev; 168 if (ret) 169 tree->last = ret; 170 return ret; 171 } 172 173 /* allocate and add a new ordered_extent into the per-inode tree. 174 * file_offset is the logical offset in the file 175 * 176 * start is the disk block number of an extent already reserved in the 177 * extent allocation tree 178 * 179 * len is the length of the extent 180 * 181 * The tree is given a single reference on the ordered extent that was 182 * inserted. 183 */ 184 static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset, 185 u64 start, u64 len, u64 disk_len, 186 int type, int dio, int compress_type) 187 { 188 struct btrfs_root *root = BTRFS_I(inode)->root; 189 struct btrfs_ordered_inode_tree *tree; 190 struct rb_node *node; 191 struct btrfs_ordered_extent *entry; 192 193 tree = &BTRFS_I(inode)->ordered_tree; 194 entry = kmem_cache_zalloc(btrfs_ordered_extent_cache, GFP_NOFS); 195 if (!entry) 196 return -ENOMEM; 197 198 entry->file_offset = file_offset; 199 entry->start = start; 200 entry->len = len; 201 if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM) && 202 !(type == BTRFS_ORDERED_NOCOW)) 203 entry->csum_bytes_left = disk_len; 204 entry->disk_len = disk_len; 205 entry->bytes_left = len; 206 entry->inode = igrab(inode); 207 entry->compress_type = compress_type; 208 entry->truncated_len = (u64)-1; 209 if (type != BTRFS_ORDERED_IO_DONE && type != BTRFS_ORDERED_COMPLETE) 210 set_bit(type, &entry->flags); 211 212 if (dio) 213 set_bit(BTRFS_ORDERED_DIRECT, &entry->flags); 214 215 /* one ref for the tree */ 216 atomic_set(&entry->refs, 1); 217 init_waitqueue_head(&entry->wait); 218 INIT_LIST_HEAD(&entry->list); 219 INIT_LIST_HEAD(&entry->root_extent_list); 220 INIT_LIST_HEAD(&entry->work_list); 221 init_completion(&entry->completion); 222 INIT_LIST_HEAD(&entry->log_list); 223 INIT_LIST_HEAD(&entry->trans_list); 224 225 trace_btrfs_ordered_extent_add(inode, entry); 226 227 spin_lock_irq(&tree->lock); 228 node = tree_insert(&tree->tree, file_offset, 229 &entry->rb_node); 230 if (node) 231 ordered_data_tree_panic(inode, -EEXIST, file_offset); 232 spin_unlock_irq(&tree->lock); 233 234 spin_lock(&root->ordered_extent_lock); 235 list_add_tail(&entry->root_extent_list, 236 &root->ordered_extents); 237 root->nr_ordered_extents++; 238 if (root->nr_ordered_extents == 1) { 239 spin_lock(&root->fs_info->ordered_root_lock); 240 BUG_ON(!list_empty(&root->ordered_root)); 241 list_add_tail(&root->ordered_root, 242 &root->fs_info->ordered_roots); 243 spin_unlock(&root->fs_info->ordered_root_lock); 244 } 245 spin_unlock(&root->ordered_extent_lock); 246 247 return 0; 248 } 249 250 int btrfs_add_ordered_extent(struct inode *inode, u64 file_offset, 251 u64 start, u64 len, u64 disk_len, int type) 252 { 253 return __btrfs_add_ordered_extent(inode, file_offset, start, len, 254 disk_len, type, 0, 255 BTRFS_COMPRESS_NONE); 256 } 257 258 int btrfs_add_ordered_extent_dio(struct inode *inode, u64 file_offset, 259 u64 start, u64 len, u64 disk_len, int type) 260 { 261 return __btrfs_add_ordered_extent(inode, file_offset, start, len, 262 disk_len, type, 1, 263 BTRFS_COMPRESS_NONE); 264 } 265 266 int btrfs_add_ordered_extent_compress(struct inode *inode, u64 file_offset, 267 u64 start, u64 len, u64 disk_len, 268 int type, int compress_type) 269 { 270 return __btrfs_add_ordered_extent(inode, file_offset, start, len, 271 disk_len, type, 0, 272 compress_type); 273 } 274 275 /* 276 * Add a struct btrfs_ordered_sum into the list of checksums to be inserted 277 * when an ordered extent is finished. If the list covers more than one 278 * ordered extent, it is split across multiples. 279 */ 280 void btrfs_add_ordered_sum(struct inode *inode, 281 struct btrfs_ordered_extent *entry, 282 struct btrfs_ordered_sum *sum) 283 { 284 struct btrfs_ordered_inode_tree *tree; 285 286 tree = &BTRFS_I(inode)->ordered_tree; 287 spin_lock_irq(&tree->lock); 288 list_add_tail(&sum->list, &entry->list); 289 WARN_ON(entry->csum_bytes_left < sum->len); 290 entry->csum_bytes_left -= sum->len; 291 if (entry->csum_bytes_left == 0) 292 wake_up(&entry->wait); 293 spin_unlock_irq(&tree->lock); 294 } 295 296 /* 297 * this is used to account for finished IO across a given range 298 * of the file. The IO may span ordered extents. If 299 * a given ordered_extent is completely done, 1 is returned, otherwise 300 * 0. 301 * 302 * test_and_set_bit on a flag in the struct btrfs_ordered_extent is used 303 * to make sure this function only returns 1 once for a given ordered extent. 304 * 305 * file_offset is updated to one byte past the range that is recorded as 306 * complete. This allows you to walk forward in the file. 307 */ 308 int btrfs_dec_test_first_ordered_pending(struct inode *inode, 309 struct btrfs_ordered_extent **cached, 310 u64 *file_offset, u64 io_size, int uptodate) 311 { 312 struct btrfs_ordered_inode_tree *tree; 313 struct rb_node *node; 314 struct btrfs_ordered_extent *entry = NULL; 315 int ret; 316 unsigned long flags; 317 u64 dec_end; 318 u64 dec_start; 319 u64 to_dec; 320 321 tree = &BTRFS_I(inode)->ordered_tree; 322 spin_lock_irqsave(&tree->lock, flags); 323 node = tree_search(tree, *file_offset); 324 if (!node) { 325 ret = 1; 326 goto out; 327 } 328 329 entry = rb_entry(node, struct btrfs_ordered_extent, rb_node); 330 if (!offset_in_entry(entry, *file_offset)) { 331 ret = 1; 332 goto out; 333 } 334 335 dec_start = max(*file_offset, entry->file_offset); 336 dec_end = min(*file_offset + io_size, entry->file_offset + 337 entry->len); 338 *file_offset = dec_end; 339 if (dec_start > dec_end) { 340 btrfs_crit(BTRFS_I(inode)->root->fs_info, 341 "bad ordering dec_start %llu end %llu", dec_start, dec_end); 342 } 343 to_dec = dec_end - dec_start; 344 if (to_dec > entry->bytes_left) { 345 btrfs_crit(BTRFS_I(inode)->root->fs_info, 346 "bad ordered accounting left %llu size %llu", 347 entry->bytes_left, to_dec); 348 } 349 entry->bytes_left -= to_dec; 350 if (!uptodate) 351 set_bit(BTRFS_ORDERED_IOERR, &entry->flags); 352 353 if (entry->bytes_left == 0) { 354 ret = test_and_set_bit(BTRFS_ORDERED_IO_DONE, &entry->flags); 355 if (waitqueue_active(&entry->wait)) 356 wake_up(&entry->wait); 357 } else { 358 ret = 1; 359 } 360 out: 361 if (!ret && cached && entry) { 362 *cached = entry; 363 atomic_inc(&entry->refs); 364 } 365 spin_unlock_irqrestore(&tree->lock, flags); 366 return ret == 0; 367 } 368 369 /* 370 * this is used to account for finished IO across a given range 371 * of the file. The IO should not span ordered extents. If 372 * a given ordered_extent is completely done, 1 is returned, otherwise 373 * 0. 374 * 375 * test_and_set_bit on a flag in the struct btrfs_ordered_extent is used 376 * to make sure this function only returns 1 once for a given ordered extent. 377 */ 378 int btrfs_dec_test_ordered_pending(struct inode *inode, 379 struct btrfs_ordered_extent **cached, 380 u64 file_offset, u64 io_size, int uptodate) 381 { 382 struct btrfs_ordered_inode_tree *tree; 383 struct rb_node *node; 384 struct btrfs_ordered_extent *entry = NULL; 385 unsigned long flags; 386 int ret; 387 388 tree = &BTRFS_I(inode)->ordered_tree; 389 spin_lock_irqsave(&tree->lock, flags); 390 if (cached && *cached) { 391 entry = *cached; 392 goto have_entry; 393 } 394 395 node = tree_search(tree, file_offset); 396 if (!node) { 397 ret = 1; 398 goto out; 399 } 400 401 entry = rb_entry(node, struct btrfs_ordered_extent, rb_node); 402 have_entry: 403 if (!offset_in_entry(entry, file_offset)) { 404 ret = 1; 405 goto out; 406 } 407 408 if (io_size > entry->bytes_left) { 409 btrfs_crit(BTRFS_I(inode)->root->fs_info, 410 "bad ordered accounting left %llu size %llu", 411 entry->bytes_left, io_size); 412 } 413 entry->bytes_left -= io_size; 414 if (!uptodate) 415 set_bit(BTRFS_ORDERED_IOERR, &entry->flags); 416 417 if (entry->bytes_left == 0) { 418 ret = test_and_set_bit(BTRFS_ORDERED_IO_DONE, &entry->flags); 419 if (waitqueue_active(&entry->wait)) 420 wake_up(&entry->wait); 421 } else { 422 ret = 1; 423 } 424 out: 425 if (!ret && cached && entry) { 426 *cached = entry; 427 atomic_inc(&entry->refs); 428 } 429 spin_unlock_irqrestore(&tree->lock, flags); 430 return ret == 0; 431 } 432 433 /* Needs to either be called under a log transaction or the log_mutex */ 434 void btrfs_get_logged_extents(struct inode *inode, 435 struct list_head *logged_list, 436 const loff_t start, 437 const loff_t end) 438 { 439 struct btrfs_ordered_inode_tree *tree; 440 struct btrfs_ordered_extent *ordered; 441 struct rb_node *n; 442 struct rb_node *prev; 443 444 tree = &BTRFS_I(inode)->ordered_tree; 445 spin_lock_irq(&tree->lock); 446 n = __tree_search(&tree->tree, end, &prev); 447 if (!n) 448 n = prev; 449 for (; n; n = rb_prev(n)) { 450 ordered = rb_entry(n, struct btrfs_ordered_extent, rb_node); 451 if (ordered->file_offset > end) 452 continue; 453 if (entry_end(ordered) <= start) 454 break; 455 if (!list_empty(&ordered->log_list)) 456 continue; 457 if (test_bit(BTRFS_ORDERED_LOGGED, &ordered->flags)) 458 continue; 459 list_add(&ordered->log_list, logged_list); 460 atomic_inc(&ordered->refs); 461 } 462 spin_unlock_irq(&tree->lock); 463 } 464 465 void btrfs_put_logged_extents(struct list_head *logged_list) 466 { 467 struct btrfs_ordered_extent *ordered; 468 469 while (!list_empty(logged_list)) { 470 ordered = list_first_entry(logged_list, 471 struct btrfs_ordered_extent, 472 log_list); 473 list_del_init(&ordered->log_list); 474 btrfs_put_ordered_extent(ordered); 475 } 476 } 477 478 void btrfs_submit_logged_extents(struct list_head *logged_list, 479 struct btrfs_root *log) 480 { 481 int index = log->log_transid % 2; 482 483 spin_lock_irq(&log->log_extents_lock[index]); 484 list_splice_tail(logged_list, &log->logged_list[index]); 485 spin_unlock_irq(&log->log_extents_lock[index]); 486 } 487 488 void btrfs_wait_logged_extents(struct btrfs_trans_handle *trans, 489 struct btrfs_root *log, u64 transid) 490 { 491 struct btrfs_ordered_extent *ordered; 492 int index = transid % 2; 493 494 spin_lock_irq(&log->log_extents_lock[index]); 495 while (!list_empty(&log->logged_list[index])) { 496 ordered = list_first_entry(&log->logged_list[index], 497 struct btrfs_ordered_extent, 498 log_list); 499 list_del_init(&ordered->log_list); 500 spin_unlock_irq(&log->log_extents_lock[index]); 501 502 if (!test_bit(BTRFS_ORDERED_IO_DONE, &ordered->flags) && 503 !test_bit(BTRFS_ORDERED_DIRECT, &ordered->flags)) { 504 struct inode *inode = ordered->inode; 505 u64 start = ordered->file_offset; 506 u64 end = ordered->file_offset + ordered->len - 1; 507 508 WARN_ON(!inode); 509 filemap_fdatawrite_range(inode->i_mapping, start, end); 510 } 511 wait_event(ordered->wait, test_bit(BTRFS_ORDERED_IO_DONE, 512 &ordered->flags)); 513 514 if (!test_and_set_bit(BTRFS_ORDERED_LOGGED, &ordered->flags)) 515 list_add_tail(&ordered->trans_list, &trans->ordered); 516 spin_lock_irq(&log->log_extents_lock[index]); 517 } 518 spin_unlock_irq(&log->log_extents_lock[index]); 519 } 520 521 void btrfs_free_logged_extents(struct btrfs_root *log, u64 transid) 522 { 523 struct btrfs_ordered_extent *ordered; 524 int index = transid % 2; 525 526 spin_lock_irq(&log->log_extents_lock[index]); 527 while (!list_empty(&log->logged_list[index])) { 528 ordered = list_first_entry(&log->logged_list[index], 529 struct btrfs_ordered_extent, 530 log_list); 531 list_del_init(&ordered->log_list); 532 spin_unlock_irq(&log->log_extents_lock[index]); 533 btrfs_put_ordered_extent(ordered); 534 spin_lock_irq(&log->log_extents_lock[index]); 535 } 536 spin_unlock_irq(&log->log_extents_lock[index]); 537 } 538 539 /* 540 * used to drop a reference on an ordered extent. This will free 541 * the extent if the last reference is dropped 542 */ 543 void btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry) 544 { 545 struct list_head *cur; 546 struct btrfs_ordered_sum *sum; 547 548 trace_btrfs_ordered_extent_put(entry->inode, entry); 549 550 if (atomic_dec_and_test(&entry->refs)) { 551 if (entry->inode) 552 btrfs_add_delayed_iput(entry->inode); 553 while (!list_empty(&entry->list)) { 554 cur = entry->list.next; 555 sum = list_entry(cur, struct btrfs_ordered_sum, list); 556 list_del(&sum->list); 557 kfree(sum); 558 } 559 kmem_cache_free(btrfs_ordered_extent_cache, entry); 560 } 561 } 562 563 /* 564 * remove an ordered extent from the tree. No references are dropped 565 * and waiters are woken up. 566 */ 567 void btrfs_remove_ordered_extent(struct inode *inode, 568 struct btrfs_ordered_extent *entry) 569 { 570 struct btrfs_ordered_inode_tree *tree; 571 struct btrfs_root *root = BTRFS_I(inode)->root; 572 struct rb_node *node; 573 574 tree = &BTRFS_I(inode)->ordered_tree; 575 spin_lock_irq(&tree->lock); 576 node = &entry->rb_node; 577 rb_erase(node, &tree->tree); 578 if (tree->last == node) 579 tree->last = NULL; 580 set_bit(BTRFS_ORDERED_COMPLETE, &entry->flags); 581 spin_unlock_irq(&tree->lock); 582 583 spin_lock(&root->ordered_extent_lock); 584 list_del_init(&entry->root_extent_list); 585 root->nr_ordered_extents--; 586 587 trace_btrfs_ordered_extent_remove(inode, entry); 588 589 if (!root->nr_ordered_extents) { 590 spin_lock(&root->fs_info->ordered_root_lock); 591 BUG_ON(list_empty(&root->ordered_root)); 592 list_del_init(&root->ordered_root); 593 spin_unlock(&root->fs_info->ordered_root_lock); 594 } 595 spin_unlock(&root->ordered_extent_lock); 596 wake_up(&entry->wait); 597 } 598 599 static void btrfs_run_ordered_extent_work(struct btrfs_work *work) 600 { 601 struct btrfs_ordered_extent *ordered; 602 603 ordered = container_of(work, struct btrfs_ordered_extent, flush_work); 604 btrfs_start_ordered_extent(ordered->inode, ordered, 1); 605 complete(&ordered->completion); 606 } 607 608 /* 609 * wait for all the ordered extents in a root. This is done when balancing 610 * space between drives. 611 */ 612 int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr) 613 { 614 struct list_head splice, works; 615 struct btrfs_ordered_extent *ordered, *next; 616 int count = 0; 617 618 INIT_LIST_HEAD(&splice); 619 INIT_LIST_HEAD(&works); 620 621 mutex_lock(&root->ordered_extent_mutex); 622 spin_lock(&root->ordered_extent_lock); 623 list_splice_init(&root->ordered_extents, &splice); 624 while (!list_empty(&splice) && nr) { 625 ordered = list_first_entry(&splice, struct btrfs_ordered_extent, 626 root_extent_list); 627 list_move_tail(&ordered->root_extent_list, 628 &root->ordered_extents); 629 atomic_inc(&ordered->refs); 630 spin_unlock(&root->ordered_extent_lock); 631 632 btrfs_init_work(&ordered->flush_work, 633 btrfs_flush_delalloc_helper, 634 btrfs_run_ordered_extent_work, NULL, NULL); 635 list_add_tail(&ordered->work_list, &works); 636 btrfs_queue_work(root->fs_info->flush_workers, 637 &ordered->flush_work); 638 639 cond_resched(); 640 spin_lock(&root->ordered_extent_lock); 641 if (nr != -1) 642 nr--; 643 count++; 644 } 645 list_splice_tail(&splice, &root->ordered_extents); 646 spin_unlock(&root->ordered_extent_lock); 647 648 list_for_each_entry_safe(ordered, next, &works, work_list) { 649 list_del_init(&ordered->work_list); 650 wait_for_completion(&ordered->completion); 651 btrfs_put_ordered_extent(ordered); 652 cond_resched(); 653 } 654 mutex_unlock(&root->ordered_extent_mutex); 655 656 return count; 657 } 658 659 void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr) 660 { 661 struct btrfs_root *root; 662 struct list_head splice; 663 int done; 664 665 INIT_LIST_HEAD(&splice); 666 667 mutex_lock(&fs_info->ordered_operations_mutex); 668 spin_lock(&fs_info->ordered_root_lock); 669 list_splice_init(&fs_info->ordered_roots, &splice); 670 while (!list_empty(&splice) && nr) { 671 root = list_first_entry(&splice, struct btrfs_root, 672 ordered_root); 673 root = btrfs_grab_fs_root(root); 674 BUG_ON(!root); 675 list_move_tail(&root->ordered_root, 676 &fs_info->ordered_roots); 677 spin_unlock(&fs_info->ordered_root_lock); 678 679 done = btrfs_wait_ordered_extents(root, nr); 680 btrfs_put_fs_root(root); 681 682 spin_lock(&fs_info->ordered_root_lock); 683 if (nr != -1) { 684 nr -= done; 685 WARN_ON(nr < 0); 686 } 687 } 688 list_splice_tail(&splice, &fs_info->ordered_roots); 689 spin_unlock(&fs_info->ordered_root_lock); 690 mutex_unlock(&fs_info->ordered_operations_mutex); 691 } 692 693 /* 694 * Used to start IO or wait for a given ordered extent to finish. 695 * 696 * If wait is one, this effectively waits on page writeback for all the pages 697 * in the extent, and it waits on the io completion code to insert 698 * metadata into the btree corresponding to the extent 699 */ 700 void btrfs_start_ordered_extent(struct inode *inode, 701 struct btrfs_ordered_extent *entry, 702 int wait) 703 { 704 u64 start = entry->file_offset; 705 u64 end = start + entry->len - 1; 706 707 trace_btrfs_ordered_extent_start(inode, entry); 708 709 /* 710 * pages in the range can be dirty, clean or writeback. We 711 * start IO on any dirty ones so the wait doesn't stall waiting 712 * for the flusher thread to find them 713 */ 714 if (!test_bit(BTRFS_ORDERED_DIRECT, &entry->flags)) 715 filemap_fdatawrite_range(inode->i_mapping, start, end); 716 if (wait) { 717 wait_event(entry->wait, test_bit(BTRFS_ORDERED_COMPLETE, 718 &entry->flags)); 719 } 720 } 721 722 /* 723 * Used to wait on ordered extents across a large range of bytes. 724 */ 725 int btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len) 726 { 727 int ret = 0; 728 u64 end; 729 u64 orig_end; 730 struct btrfs_ordered_extent *ordered; 731 732 if (start + len < start) { 733 orig_end = INT_LIMIT(loff_t); 734 } else { 735 orig_end = start + len - 1; 736 if (orig_end > INT_LIMIT(loff_t)) 737 orig_end = INT_LIMIT(loff_t); 738 } 739 740 /* start IO across the range first to instantiate any delalloc 741 * extents 742 */ 743 ret = btrfs_fdatawrite_range(inode, start, orig_end); 744 if (ret) 745 return ret; 746 747 ret = filemap_fdatawait_range(inode->i_mapping, start, orig_end); 748 if (ret) 749 return ret; 750 751 end = orig_end; 752 while (1) { 753 ordered = btrfs_lookup_first_ordered_extent(inode, end); 754 if (!ordered) 755 break; 756 if (ordered->file_offset > orig_end) { 757 btrfs_put_ordered_extent(ordered); 758 break; 759 } 760 if (ordered->file_offset + ordered->len <= start) { 761 btrfs_put_ordered_extent(ordered); 762 break; 763 } 764 btrfs_start_ordered_extent(inode, ordered, 1); 765 end = ordered->file_offset; 766 if (test_bit(BTRFS_ORDERED_IOERR, &ordered->flags)) 767 ret = -EIO; 768 btrfs_put_ordered_extent(ordered); 769 if (ret || end == 0 || end == start) 770 break; 771 end--; 772 } 773 return ret; 774 } 775 776 /* 777 * find an ordered extent corresponding to file_offset. return NULL if 778 * nothing is found, otherwise take a reference on the extent and return it 779 */ 780 struct btrfs_ordered_extent *btrfs_lookup_ordered_extent(struct inode *inode, 781 u64 file_offset) 782 { 783 struct btrfs_ordered_inode_tree *tree; 784 struct rb_node *node; 785 struct btrfs_ordered_extent *entry = NULL; 786 787 tree = &BTRFS_I(inode)->ordered_tree; 788 spin_lock_irq(&tree->lock); 789 node = tree_search(tree, file_offset); 790 if (!node) 791 goto out; 792 793 entry = rb_entry(node, struct btrfs_ordered_extent, rb_node); 794 if (!offset_in_entry(entry, file_offset)) 795 entry = NULL; 796 if (entry) 797 atomic_inc(&entry->refs); 798 out: 799 spin_unlock_irq(&tree->lock); 800 return entry; 801 } 802 803 /* Since the DIO code tries to lock a wide area we need to look for any ordered 804 * extents that exist in the range, rather than just the start of the range. 805 */ 806 struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode, 807 u64 file_offset, 808 u64 len) 809 { 810 struct btrfs_ordered_inode_tree *tree; 811 struct rb_node *node; 812 struct btrfs_ordered_extent *entry = NULL; 813 814 tree = &BTRFS_I(inode)->ordered_tree; 815 spin_lock_irq(&tree->lock); 816 node = tree_search(tree, file_offset); 817 if (!node) { 818 node = tree_search(tree, file_offset + len); 819 if (!node) 820 goto out; 821 } 822 823 while (1) { 824 entry = rb_entry(node, struct btrfs_ordered_extent, rb_node); 825 if (range_overlaps(entry, file_offset, len)) 826 break; 827 828 if (entry->file_offset >= file_offset + len) { 829 entry = NULL; 830 break; 831 } 832 entry = NULL; 833 node = rb_next(node); 834 if (!node) 835 break; 836 } 837 out: 838 if (entry) 839 atomic_inc(&entry->refs); 840 spin_unlock_irq(&tree->lock); 841 return entry; 842 } 843 844 /* 845 * lookup and return any extent before 'file_offset'. NULL is returned 846 * if none is found 847 */ 848 struct btrfs_ordered_extent * 849 btrfs_lookup_first_ordered_extent(struct inode *inode, u64 file_offset) 850 { 851 struct btrfs_ordered_inode_tree *tree; 852 struct rb_node *node; 853 struct btrfs_ordered_extent *entry = NULL; 854 855 tree = &BTRFS_I(inode)->ordered_tree; 856 spin_lock_irq(&tree->lock); 857 node = tree_search(tree, file_offset); 858 if (!node) 859 goto out; 860 861 entry = rb_entry(node, struct btrfs_ordered_extent, rb_node); 862 atomic_inc(&entry->refs); 863 out: 864 spin_unlock_irq(&tree->lock); 865 return entry; 866 } 867 868 /* 869 * After an extent is done, call this to conditionally update the on disk 870 * i_size. i_size is updated to cover any fully written part of the file. 871 */ 872 int btrfs_ordered_update_i_size(struct inode *inode, u64 offset, 873 struct btrfs_ordered_extent *ordered) 874 { 875 struct btrfs_ordered_inode_tree *tree = &BTRFS_I(inode)->ordered_tree; 876 u64 disk_i_size; 877 u64 new_i_size; 878 u64 i_size = i_size_read(inode); 879 struct rb_node *node; 880 struct rb_node *prev = NULL; 881 struct btrfs_ordered_extent *test; 882 int ret = 1; 883 884 spin_lock_irq(&tree->lock); 885 if (ordered) { 886 offset = entry_end(ordered); 887 if (test_bit(BTRFS_ORDERED_TRUNCATED, &ordered->flags)) 888 offset = min(offset, 889 ordered->file_offset + 890 ordered->truncated_len); 891 } else { 892 offset = ALIGN(offset, BTRFS_I(inode)->root->sectorsize); 893 } 894 disk_i_size = BTRFS_I(inode)->disk_i_size; 895 896 /* truncate file */ 897 if (disk_i_size > i_size) { 898 BTRFS_I(inode)->disk_i_size = i_size; 899 ret = 0; 900 goto out; 901 } 902 903 /* 904 * if the disk i_size is already at the inode->i_size, or 905 * this ordered extent is inside the disk i_size, we're done 906 */ 907 if (disk_i_size == i_size) 908 goto out; 909 910 /* 911 * We still need to update disk_i_size if outstanding_isize is greater 912 * than disk_i_size. 913 */ 914 if (offset <= disk_i_size && 915 (!ordered || ordered->outstanding_isize <= disk_i_size)) 916 goto out; 917 918 /* 919 * walk backward from this ordered extent to disk_i_size. 920 * if we find an ordered extent then we can't update disk i_size 921 * yet 922 */ 923 if (ordered) { 924 node = rb_prev(&ordered->rb_node); 925 } else { 926 prev = tree_search(tree, offset); 927 /* 928 * we insert file extents without involving ordered struct, 929 * so there should be no ordered struct cover this offset 930 */ 931 if (prev) { 932 test = rb_entry(prev, struct btrfs_ordered_extent, 933 rb_node); 934 BUG_ON(offset_in_entry(test, offset)); 935 } 936 node = prev; 937 } 938 for (; node; node = rb_prev(node)) { 939 test = rb_entry(node, struct btrfs_ordered_extent, rb_node); 940 941 /* We treat this entry as if it doesnt exist */ 942 if (test_bit(BTRFS_ORDERED_UPDATED_ISIZE, &test->flags)) 943 continue; 944 if (test->file_offset + test->len <= disk_i_size) 945 break; 946 if (test->file_offset >= i_size) 947 break; 948 if (entry_end(test) > disk_i_size) { 949 /* 950 * we don't update disk_i_size now, so record this 951 * undealt i_size. Or we will not know the real 952 * i_size. 953 */ 954 if (test->outstanding_isize < offset) 955 test->outstanding_isize = offset; 956 if (ordered && 957 ordered->outstanding_isize > 958 test->outstanding_isize) 959 test->outstanding_isize = 960 ordered->outstanding_isize; 961 goto out; 962 } 963 } 964 new_i_size = min_t(u64, offset, i_size); 965 966 /* 967 * Some ordered extents may completed before the current one, and 968 * we hold the real i_size in ->outstanding_isize. 969 */ 970 if (ordered && ordered->outstanding_isize > new_i_size) 971 new_i_size = min_t(u64, ordered->outstanding_isize, i_size); 972 BTRFS_I(inode)->disk_i_size = new_i_size; 973 ret = 0; 974 out: 975 /* 976 * We need to do this because we can't remove ordered extents until 977 * after the i_disk_size has been updated and then the inode has been 978 * updated to reflect the change, so we need to tell anybody who finds 979 * this ordered extent that we've already done all the real work, we 980 * just haven't completed all the other work. 981 */ 982 if (ordered) 983 set_bit(BTRFS_ORDERED_UPDATED_ISIZE, &ordered->flags); 984 spin_unlock_irq(&tree->lock); 985 return ret; 986 } 987 988 /* 989 * search the ordered extents for one corresponding to 'offset' and 990 * try to find a checksum. This is used because we allow pages to 991 * be reclaimed before their checksum is actually put into the btree 992 */ 993 int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr, 994 u32 *sum, int len) 995 { 996 struct btrfs_ordered_sum *ordered_sum; 997 struct btrfs_ordered_extent *ordered; 998 struct btrfs_ordered_inode_tree *tree = &BTRFS_I(inode)->ordered_tree; 999 unsigned long num_sectors; 1000 unsigned long i; 1001 u32 sectorsize = BTRFS_I(inode)->root->sectorsize; 1002 int index = 0; 1003 1004 ordered = btrfs_lookup_ordered_extent(inode, offset); 1005 if (!ordered) 1006 return 0; 1007 1008 spin_lock_irq(&tree->lock); 1009 list_for_each_entry_reverse(ordered_sum, &ordered->list, list) { 1010 if (disk_bytenr >= ordered_sum->bytenr && 1011 disk_bytenr < ordered_sum->bytenr + ordered_sum->len) { 1012 i = (disk_bytenr - ordered_sum->bytenr) >> 1013 inode->i_sb->s_blocksize_bits; 1014 num_sectors = ordered_sum->len >> 1015 inode->i_sb->s_blocksize_bits; 1016 num_sectors = min_t(int, len - index, num_sectors - i); 1017 memcpy(sum + index, ordered_sum->sums + i, 1018 num_sectors); 1019 1020 index += (int)num_sectors; 1021 if (index == len) 1022 goto out; 1023 disk_bytenr += num_sectors * sectorsize; 1024 } 1025 } 1026 out: 1027 spin_unlock_irq(&tree->lock); 1028 btrfs_put_ordered_extent(ordered); 1029 return index; 1030 } 1031 1032 int __init ordered_data_init(void) 1033 { 1034 btrfs_ordered_extent_cache = kmem_cache_create("btrfs_ordered_extent", 1035 sizeof(struct btrfs_ordered_extent), 0, 1036 SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, 1037 NULL); 1038 if (!btrfs_ordered_extent_cache) 1039 return -ENOMEM; 1040 1041 return 0; 1042 } 1043 1044 void ordered_data_exit(void) 1045 { 1046 if (btrfs_ordered_extent_cache) 1047 kmem_cache_destroy(btrfs_ordered_extent_cache); 1048 } 1049