1 // SPDX-License-Identifier: GPL-2.0 2 3 #include "bcachefs.h" 4 #include "bkey_methods.h" 5 #include "bkey_sort.h" 6 #include "btree_cache.h" 7 #include "btree_io.h" 8 #include "btree_iter.h" 9 #include "btree_locking.h" 10 #include "btree_update.h" 11 #include "btree_update_interior.h" 12 #include "buckets.h" 13 #include "checksum.h" 14 #include "debug.h" 15 #include "error.h" 16 #include "extents.h" 17 #include "io_write.h" 18 #include "journal_reclaim.h" 19 #include "journal_seq_blacklist.h" 20 #include "recovery.h" 21 #include "super-io.h" 22 #include "trace.h" 23 24 #include <linux/sched/mm.h> 25 26 void bch2_btree_node_io_unlock(struct btree *b) 27 { 28 EBUG_ON(!btree_node_write_in_flight(b)); 29 30 clear_btree_node_write_in_flight_inner(b); 31 clear_btree_node_write_in_flight(b); 32 wake_up_bit(&b->flags, BTREE_NODE_write_in_flight); 33 } 34 35 void bch2_btree_node_io_lock(struct btree *b) 36 { 37 bch2_assert_btree_nodes_not_locked(); 38 39 wait_on_bit_lock_io(&b->flags, BTREE_NODE_write_in_flight, 40 TASK_UNINTERRUPTIBLE); 41 } 42 43 void __bch2_btree_node_wait_on_read(struct btree *b) 44 { 45 wait_on_bit_io(&b->flags, BTREE_NODE_read_in_flight, 46 TASK_UNINTERRUPTIBLE); 47 } 48 49 void __bch2_btree_node_wait_on_write(struct btree *b) 50 { 51 wait_on_bit_io(&b->flags, BTREE_NODE_write_in_flight, 52 TASK_UNINTERRUPTIBLE); 53 } 54 55 void bch2_btree_node_wait_on_read(struct btree *b) 56 { 57 bch2_assert_btree_nodes_not_locked(); 58 59 wait_on_bit_io(&b->flags, BTREE_NODE_read_in_flight, 60 TASK_UNINTERRUPTIBLE); 61 } 62 63 void bch2_btree_node_wait_on_write(struct btree *b) 64 { 65 bch2_assert_btree_nodes_not_locked(); 66 67 wait_on_bit_io(&b->flags, BTREE_NODE_write_in_flight, 68 TASK_UNINTERRUPTIBLE); 69 } 70 71 static void verify_no_dups(struct btree *b, 72 struct bkey_packed *start, 73 struct bkey_packed *end) 74 { 75 #ifdef CONFIG_BCACHEFS_DEBUG 76 struct bkey_packed *k, *p; 77 78 if (start == end) 79 return; 80 81 for (p = start, k = bkey_p_next(start); 82 k != end; 83 p = k, k = bkey_p_next(k)) { 84 struct bkey l = bkey_unpack_key(b, p); 85 struct bkey r = bkey_unpack_key(b, k); 86 87 BUG_ON(bpos_ge(l.p, bkey_start_pos(&r))); 88 } 89 #endif 90 } 91 92 static void set_needs_whiteout(struct bset *i, int v) 93 { 94 struct bkey_packed *k; 95 96 for (k = i->start; k != vstruct_last(i); k = bkey_p_next(k)) 97 k->needs_whiteout = v; 98 } 99 100 static void btree_bounce_free(struct bch_fs *c, size_t size, 101 bool used_mempool, void *p) 102 { 103 if (used_mempool) 104 mempool_free(p, &c->btree_bounce_pool); 105 else 106 kvfree(p); 107 } 108 109 static void *btree_bounce_alloc(struct bch_fs *c, size_t size, 110 bool *used_mempool) 111 { 112 unsigned flags = memalloc_nofs_save(); 113 void *p; 114 115 BUG_ON(size > c->opts.btree_node_size); 116 117 *used_mempool = false; 118 p = kvmalloc(size, __GFP_NOWARN|GFP_NOWAIT); 119 if (!p) { 120 *used_mempool = true; 121 p = mempool_alloc(&c->btree_bounce_pool, GFP_NOFS); 122 } 123 memalloc_nofs_restore(flags); 124 return p; 125 } 126 127 static void sort_bkey_ptrs(const struct btree *bt, 128 struct bkey_packed **ptrs, unsigned nr) 129 { 130 unsigned n = nr, a = nr / 2, b, c, d; 131 132 if (!a) 133 return; 134 135 /* Heap sort: see lib/sort.c: */ 136 while (1) { 137 if (a) 138 a--; 139 else if (--n) 140 swap(ptrs[0], ptrs[n]); 141 else 142 break; 143 144 for (b = a; c = 2 * b + 1, (d = c + 1) < n;) 145 b = bch2_bkey_cmp_packed(bt, 146 ptrs[c], 147 ptrs[d]) >= 0 ? c : d; 148 if (d == n) 149 b = c; 150 151 while (b != a && 152 bch2_bkey_cmp_packed(bt, 153 ptrs[a], 154 ptrs[b]) >= 0) 155 b = (b - 1) / 2; 156 c = b; 157 while (b != a) { 158 b = (b - 1) / 2; 159 swap(ptrs[b], ptrs[c]); 160 } 161 } 162 } 163 164 static void bch2_sort_whiteouts(struct bch_fs *c, struct btree *b) 165 { 166 struct bkey_packed *new_whiteouts, **ptrs, **ptrs_end, *k; 167 bool used_mempool = false; 168 size_t bytes = b->whiteout_u64s * sizeof(u64); 169 170 if (!b->whiteout_u64s) 171 return; 172 173 new_whiteouts = btree_bounce_alloc(c, bytes, &used_mempool); 174 175 ptrs = ptrs_end = ((void *) new_whiteouts + bytes); 176 177 for (k = unwritten_whiteouts_start(b); 178 k != unwritten_whiteouts_end(b); 179 k = bkey_p_next(k)) 180 *--ptrs = k; 181 182 sort_bkey_ptrs(b, ptrs, ptrs_end - ptrs); 183 184 k = new_whiteouts; 185 186 while (ptrs != ptrs_end) { 187 bkey_p_copy(k, *ptrs); 188 k = bkey_p_next(k); 189 ptrs++; 190 } 191 192 verify_no_dups(b, new_whiteouts, 193 (void *) ((u64 *) new_whiteouts + b->whiteout_u64s)); 194 195 memcpy_u64s(unwritten_whiteouts_start(b), 196 new_whiteouts, b->whiteout_u64s); 197 198 btree_bounce_free(c, bytes, used_mempool, new_whiteouts); 199 } 200 201 static bool should_compact_bset(struct btree *b, struct bset_tree *t, 202 bool compacting, enum compact_mode mode) 203 { 204 if (!bset_dead_u64s(b, t)) 205 return false; 206 207 switch (mode) { 208 case COMPACT_LAZY: 209 return should_compact_bset_lazy(b, t) || 210 (compacting && !bset_written(b, bset(b, t))); 211 case COMPACT_ALL: 212 return true; 213 default: 214 BUG(); 215 } 216 } 217 218 static bool bch2_drop_whiteouts(struct btree *b, enum compact_mode mode) 219 { 220 struct bset_tree *t; 221 bool ret = false; 222 223 for_each_bset(b, t) { 224 struct bset *i = bset(b, t); 225 struct bkey_packed *k, *n, *out, *start, *end; 226 struct btree_node_entry *src = NULL, *dst = NULL; 227 228 if (t != b->set && !bset_written(b, i)) { 229 src = container_of(i, struct btree_node_entry, keys); 230 dst = max(write_block(b), 231 (void *) btree_bkey_last(b, t - 1)); 232 } 233 234 if (src != dst) 235 ret = true; 236 237 if (!should_compact_bset(b, t, ret, mode)) { 238 if (src != dst) { 239 memmove(dst, src, sizeof(*src) + 240 le16_to_cpu(src->keys.u64s) * 241 sizeof(u64)); 242 i = &dst->keys; 243 set_btree_bset(b, t, i); 244 } 245 continue; 246 } 247 248 start = btree_bkey_first(b, t); 249 end = btree_bkey_last(b, t); 250 251 if (src != dst) { 252 memmove(dst, src, sizeof(*src)); 253 i = &dst->keys; 254 set_btree_bset(b, t, i); 255 } 256 257 out = i->start; 258 259 for (k = start; k != end; k = n) { 260 n = bkey_p_next(k); 261 262 if (!bkey_deleted(k)) { 263 bkey_p_copy(out, k); 264 out = bkey_p_next(out); 265 } else { 266 BUG_ON(k->needs_whiteout); 267 } 268 } 269 270 i->u64s = cpu_to_le16((u64 *) out - i->_data); 271 set_btree_bset_end(b, t); 272 bch2_bset_set_no_aux_tree(b, t); 273 ret = true; 274 } 275 276 bch2_verify_btree_nr_keys(b); 277 278 bch2_btree_build_aux_trees(b); 279 280 return ret; 281 } 282 283 bool bch2_compact_whiteouts(struct bch_fs *c, struct btree *b, 284 enum compact_mode mode) 285 { 286 return bch2_drop_whiteouts(b, mode); 287 } 288 289 static void btree_node_sort(struct bch_fs *c, struct btree *b, 290 unsigned start_idx, 291 unsigned end_idx, 292 bool filter_whiteouts) 293 { 294 struct btree_node *out; 295 struct sort_iter_stack sort_iter; 296 struct bset_tree *t; 297 struct bset *start_bset = bset(b, &b->set[start_idx]); 298 bool used_mempool = false; 299 u64 start_time, seq = 0; 300 unsigned i, u64s = 0, bytes, shift = end_idx - start_idx - 1; 301 bool sorting_entire_node = start_idx == 0 && 302 end_idx == b->nsets; 303 304 sort_iter_stack_init(&sort_iter, b); 305 306 for (t = b->set + start_idx; 307 t < b->set + end_idx; 308 t++) { 309 u64s += le16_to_cpu(bset(b, t)->u64s); 310 sort_iter_add(&sort_iter.iter, 311 btree_bkey_first(b, t), 312 btree_bkey_last(b, t)); 313 } 314 315 bytes = sorting_entire_node 316 ? btree_buf_bytes(b) 317 : __vstruct_bytes(struct btree_node, u64s); 318 319 out = btree_bounce_alloc(c, bytes, &used_mempool); 320 321 start_time = local_clock(); 322 323 u64s = bch2_sort_keys(out->keys.start, &sort_iter.iter, filter_whiteouts); 324 325 out->keys.u64s = cpu_to_le16(u64s); 326 327 BUG_ON(vstruct_end(&out->keys) > (void *) out + bytes); 328 329 if (sorting_entire_node) 330 bch2_time_stats_update(&c->times[BCH_TIME_btree_node_sort], 331 start_time); 332 333 /* Make sure we preserve bset journal_seq: */ 334 for (t = b->set + start_idx; t < b->set + end_idx; t++) 335 seq = max(seq, le64_to_cpu(bset(b, t)->journal_seq)); 336 start_bset->journal_seq = cpu_to_le64(seq); 337 338 if (sorting_entire_node) { 339 u64s = le16_to_cpu(out->keys.u64s); 340 341 BUG_ON(bytes != btree_buf_bytes(b)); 342 343 /* 344 * Our temporary buffer is the same size as the btree node's 345 * buffer, we can just swap buffers instead of doing a big 346 * memcpy() 347 */ 348 *out = *b->data; 349 out->keys.u64s = cpu_to_le16(u64s); 350 swap(out, b->data); 351 set_btree_bset(b, b->set, &b->data->keys); 352 } else { 353 start_bset->u64s = out->keys.u64s; 354 memcpy_u64s(start_bset->start, 355 out->keys.start, 356 le16_to_cpu(out->keys.u64s)); 357 } 358 359 for (i = start_idx + 1; i < end_idx; i++) 360 b->nr.bset_u64s[start_idx] += 361 b->nr.bset_u64s[i]; 362 363 b->nsets -= shift; 364 365 for (i = start_idx + 1; i < b->nsets; i++) { 366 b->nr.bset_u64s[i] = b->nr.bset_u64s[i + shift]; 367 b->set[i] = b->set[i + shift]; 368 } 369 370 for (i = b->nsets; i < MAX_BSETS; i++) 371 b->nr.bset_u64s[i] = 0; 372 373 set_btree_bset_end(b, &b->set[start_idx]); 374 bch2_bset_set_no_aux_tree(b, &b->set[start_idx]); 375 376 btree_bounce_free(c, bytes, used_mempool, out); 377 378 bch2_verify_btree_nr_keys(b); 379 } 380 381 void bch2_btree_sort_into(struct bch_fs *c, 382 struct btree *dst, 383 struct btree *src) 384 { 385 struct btree_nr_keys nr; 386 struct btree_node_iter src_iter; 387 u64 start_time = local_clock(); 388 389 BUG_ON(dst->nsets != 1); 390 391 bch2_bset_set_no_aux_tree(dst, dst->set); 392 393 bch2_btree_node_iter_init_from_start(&src_iter, src); 394 395 nr = bch2_sort_repack(btree_bset_first(dst), 396 src, &src_iter, 397 &dst->format, 398 true); 399 400 bch2_time_stats_update(&c->times[BCH_TIME_btree_node_sort], 401 start_time); 402 403 set_btree_bset_end(dst, dst->set); 404 405 dst->nr.live_u64s += nr.live_u64s; 406 dst->nr.bset_u64s[0] += nr.bset_u64s[0]; 407 dst->nr.packed_keys += nr.packed_keys; 408 dst->nr.unpacked_keys += nr.unpacked_keys; 409 410 bch2_verify_btree_nr_keys(dst); 411 } 412 413 /* 414 * We're about to add another bset to the btree node, so if there's currently 415 * too many bsets - sort some of them together: 416 */ 417 static bool btree_node_compact(struct bch_fs *c, struct btree *b) 418 { 419 unsigned unwritten_idx; 420 bool ret = false; 421 422 for (unwritten_idx = 0; 423 unwritten_idx < b->nsets; 424 unwritten_idx++) 425 if (!bset_written(b, bset(b, &b->set[unwritten_idx]))) 426 break; 427 428 if (b->nsets - unwritten_idx > 1) { 429 btree_node_sort(c, b, unwritten_idx, 430 b->nsets, false); 431 ret = true; 432 } 433 434 if (unwritten_idx > 1) { 435 btree_node_sort(c, b, 0, unwritten_idx, false); 436 ret = true; 437 } 438 439 return ret; 440 } 441 442 void bch2_btree_build_aux_trees(struct btree *b) 443 { 444 struct bset_tree *t; 445 446 for_each_bset(b, t) 447 bch2_bset_build_aux_tree(b, t, 448 !bset_written(b, bset(b, t)) && 449 t == bset_tree_last(b)); 450 } 451 452 /* 453 * If we have MAX_BSETS (3) bsets, should we sort them all down to just one? 454 * 455 * The first bset is going to be of similar order to the size of the node, the 456 * last bset is bounded by btree_write_set_buffer(), which is set to keep the 457 * memmove on insert from being too expensive: the middle bset should, ideally, 458 * be the geometric mean of the first and the last. 459 * 460 * Returns true if the middle bset is greater than that geometric mean: 461 */ 462 static inline bool should_compact_all(struct bch_fs *c, struct btree *b) 463 { 464 unsigned mid_u64s_bits = 465 (ilog2(btree_max_u64s(c)) + BTREE_WRITE_SET_U64s_BITS) / 2; 466 467 return bset_u64s(&b->set[1]) > 1U << mid_u64s_bits; 468 } 469 470 /* 471 * @bch_btree_init_next - initialize a new (unwritten) bset that can then be 472 * inserted into 473 * 474 * Safe to call if there already is an unwritten bset - will only add a new bset 475 * if @b doesn't already have one. 476 * 477 * Returns true if we sorted (i.e. invalidated iterators 478 */ 479 void bch2_btree_init_next(struct btree_trans *trans, struct btree *b) 480 { 481 struct bch_fs *c = trans->c; 482 struct btree_node_entry *bne; 483 bool reinit_iter = false; 484 485 EBUG_ON(!six_lock_counts(&b->c.lock).n[SIX_LOCK_write]); 486 BUG_ON(bset_written(b, bset(b, &b->set[1]))); 487 BUG_ON(btree_node_just_written(b)); 488 489 if (b->nsets == MAX_BSETS && 490 !btree_node_write_in_flight(b) && 491 should_compact_all(c, b)) { 492 bch2_btree_node_write(c, b, SIX_LOCK_write, 493 BTREE_WRITE_init_next_bset); 494 reinit_iter = true; 495 } 496 497 if (b->nsets == MAX_BSETS && 498 btree_node_compact(c, b)) 499 reinit_iter = true; 500 501 BUG_ON(b->nsets >= MAX_BSETS); 502 503 bne = want_new_bset(c, b); 504 if (bne) 505 bch2_bset_init_next(b, bne); 506 507 bch2_btree_build_aux_trees(b); 508 509 if (reinit_iter) 510 bch2_trans_node_reinit_iter(trans, b); 511 } 512 513 static void btree_err_msg(struct printbuf *out, struct bch_fs *c, 514 struct bch_dev *ca, 515 struct btree *b, struct bset *i, 516 unsigned offset, int write) 517 { 518 prt_printf(out, bch2_log_msg(c, "%s"), 519 write == READ 520 ? "error validating btree node " 521 : "corrupt btree node before write "); 522 if (ca) 523 prt_printf(out, "on %s ", ca->name); 524 prt_printf(out, "at btree "); 525 bch2_btree_pos_to_text(out, c, b); 526 527 prt_printf(out, "\n node offset %u/%u", 528 b->written, btree_ptr_sectors_written(&b->key)); 529 if (i) 530 prt_printf(out, " bset u64s %u", le16_to_cpu(i->u64s)); 531 prt_str(out, ": "); 532 } 533 534 __printf(9, 10) 535 static int __btree_err(int ret, 536 struct bch_fs *c, 537 struct bch_dev *ca, 538 struct btree *b, 539 struct bset *i, 540 int write, 541 bool have_retry, 542 enum bch_sb_error_id err_type, 543 const char *fmt, ...) 544 { 545 struct printbuf out = PRINTBUF; 546 va_list args; 547 548 btree_err_msg(&out, c, ca, b, i, b->written, write); 549 550 va_start(args, fmt); 551 prt_vprintf(&out, fmt, args); 552 va_end(args); 553 554 if (write == WRITE) { 555 bch2_print_string_as_lines(KERN_ERR, out.buf); 556 ret = c->opts.errors == BCH_ON_ERROR_continue 557 ? 0 558 : -BCH_ERR_fsck_errors_not_fixed; 559 goto out; 560 } 561 562 if (!have_retry && ret == -BCH_ERR_btree_node_read_err_want_retry) 563 ret = -BCH_ERR_btree_node_read_err_fixable; 564 if (!have_retry && ret == -BCH_ERR_btree_node_read_err_must_retry) 565 ret = -BCH_ERR_btree_node_read_err_bad_node; 566 567 if (ret != -BCH_ERR_btree_node_read_err_fixable) 568 bch2_sb_error_count(c, err_type); 569 570 switch (ret) { 571 case -BCH_ERR_btree_node_read_err_fixable: 572 ret = bch2_fsck_err(c, FSCK_CAN_FIX, err_type, "%s", out.buf); 573 if (ret != -BCH_ERR_fsck_fix && 574 ret != -BCH_ERR_fsck_ignore) 575 goto fsck_err; 576 ret = -BCH_ERR_fsck_fix; 577 break; 578 case -BCH_ERR_btree_node_read_err_want_retry: 579 case -BCH_ERR_btree_node_read_err_must_retry: 580 bch2_print_string_as_lines(KERN_ERR, out.buf); 581 break; 582 case -BCH_ERR_btree_node_read_err_bad_node: 583 bch2_print_string_as_lines(KERN_ERR, out.buf); 584 ret = bch2_topology_error(c); 585 break; 586 case -BCH_ERR_btree_node_read_err_incompatible: 587 bch2_print_string_as_lines(KERN_ERR, out.buf); 588 ret = -BCH_ERR_fsck_errors_not_fixed; 589 break; 590 default: 591 BUG(); 592 } 593 out: 594 fsck_err: 595 printbuf_exit(&out); 596 return ret; 597 } 598 599 #define btree_err(type, c, ca, b, i, _err_type, msg, ...) \ 600 ({ \ 601 int _ret = __btree_err(type, c, ca, b, i, write, have_retry, \ 602 BCH_FSCK_ERR_##_err_type, \ 603 msg, ##__VA_ARGS__); \ 604 \ 605 if (_ret != -BCH_ERR_fsck_fix) { \ 606 ret = _ret; \ 607 goto fsck_err; \ 608 } \ 609 \ 610 *saw_error = true; \ 611 }) 612 613 #define btree_err_on(cond, ...) ((cond) ? btree_err(__VA_ARGS__) : false) 614 615 /* 616 * When btree topology repair changes the start or end of a node, that might 617 * mean we have to drop keys that are no longer inside the node: 618 */ 619 __cold 620 void bch2_btree_node_drop_keys_outside_node(struct btree *b) 621 { 622 struct bset_tree *t; 623 624 for_each_bset(b, t) { 625 struct bset *i = bset(b, t); 626 struct bkey_packed *k; 627 628 for (k = i->start; k != vstruct_last(i); k = bkey_p_next(k)) 629 if (bkey_cmp_left_packed(b, k, &b->data->min_key) >= 0) 630 break; 631 632 if (k != i->start) { 633 unsigned shift = (u64 *) k - (u64 *) i->start; 634 635 memmove_u64s_down(i->start, k, 636 (u64 *) vstruct_end(i) - (u64 *) k); 637 i->u64s = cpu_to_le16(le16_to_cpu(i->u64s) - shift); 638 set_btree_bset_end(b, t); 639 } 640 641 for (k = i->start; k != vstruct_last(i); k = bkey_p_next(k)) 642 if (bkey_cmp_left_packed(b, k, &b->data->max_key) > 0) 643 break; 644 645 if (k != vstruct_last(i)) { 646 i->u64s = cpu_to_le16((u64 *) k - (u64 *) i->start); 647 set_btree_bset_end(b, t); 648 } 649 } 650 651 /* 652 * Always rebuild search trees: eytzinger search tree nodes directly 653 * depend on the values of min/max key: 654 */ 655 bch2_bset_set_no_aux_tree(b, b->set); 656 bch2_btree_build_aux_trees(b); 657 b->nr = bch2_btree_node_count_keys(b); 658 659 struct bkey_s_c k; 660 struct bkey unpacked; 661 struct btree_node_iter iter; 662 for_each_btree_node_key_unpack(b, k, &iter, &unpacked) { 663 BUG_ON(bpos_lt(k.k->p, b->data->min_key)); 664 BUG_ON(bpos_gt(k.k->p, b->data->max_key)); 665 } 666 } 667 668 static int validate_bset(struct bch_fs *c, struct bch_dev *ca, 669 struct btree *b, struct bset *i, 670 unsigned offset, unsigned sectors, 671 int write, bool have_retry, bool *saw_error) 672 { 673 unsigned version = le16_to_cpu(i->version); 674 struct printbuf buf1 = PRINTBUF; 675 struct printbuf buf2 = PRINTBUF; 676 int ret = 0; 677 678 btree_err_on(!bch2_version_compatible(version), 679 -BCH_ERR_btree_node_read_err_incompatible, 680 c, ca, b, i, 681 btree_node_unsupported_version, 682 "unsupported bset version %u.%u", 683 BCH_VERSION_MAJOR(version), 684 BCH_VERSION_MINOR(version)); 685 686 if (btree_err_on(version < c->sb.version_min, 687 -BCH_ERR_btree_node_read_err_fixable, 688 c, NULL, b, i, 689 btree_node_bset_older_than_sb_min, 690 "bset version %u older than superblock version_min %u", 691 version, c->sb.version_min)) { 692 mutex_lock(&c->sb_lock); 693 c->disk_sb.sb->version_min = cpu_to_le16(version); 694 bch2_write_super(c); 695 mutex_unlock(&c->sb_lock); 696 } 697 698 if (btree_err_on(BCH_VERSION_MAJOR(version) > 699 BCH_VERSION_MAJOR(c->sb.version), 700 -BCH_ERR_btree_node_read_err_fixable, 701 c, NULL, b, i, 702 btree_node_bset_newer_than_sb, 703 "bset version %u newer than superblock version %u", 704 version, c->sb.version)) { 705 mutex_lock(&c->sb_lock); 706 c->disk_sb.sb->version = cpu_to_le16(version); 707 bch2_write_super(c); 708 mutex_unlock(&c->sb_lock); 709 } 710 711 btree_err_on(BSET_SEPARATE_WHITEOUTS(i), 712 -BCH_ERR_btree_node_read_err_incompatible, 713 c, ca, b, i, 714 btree_node_unsupported_version, 715 "BSET_SEPARATE_WHITEOUTS no longer supported"); 716 717 if (btree_err_on(offset + sectors > btree_sectors(c), 718 -BCH_ERR_btree_node_read_err_fixable, 719 c, ca, b, i, 720 bset_past_end_of_btree_node, 721 "bset past end of btree node")) { 722 i->u64s = 0; 723 ret = 0; 724 goto out; 725 } 726 727 btree_err_on(offset && !i->u64s, 728 -BCH_ERR_btree_node_read_err_fixable, 729 c, ca, b, i, 730 bset_empty, 731 "empty bset"); 732 733 btree_err_on(BSET_OFFSET(i) && BSET_OFFSET(i) != offset, 734 -BCH_ERR_btree_node_read_err_want_retry, 735 c, ca, b, i, 736 bset_wrong_sector_offset, 737 "bset at wrong sector offset"); 738 739 if (!offset) { 740 struct btree_node *bn = 741 container_of(i, struct btree_node, keys); 742 /* These indicate that we read the wrong btree node: */ 743 744 if (b->key.k.type == KEY_TYPE_btree_ptr_v2) { 745 struct bch_btree_ptr_v2 *bp = 746 &bkey_i_to_btree_ptr_v2(&b->key)->v; 747 748 /* XXX endianness */ 749 btree_err_on(bp->seq != bn->keys.seq, 750 -BCH_ERR_btree_node_read_err_must_retry, 751 c, ca, b, NULL, 752 bset_bad_seq, 753 "incorrect sequence number (wrong btree node)"); 754 } 755 756 btree_err_on(BTREE_NODE_ID(bn) != b->c.btree_id, 757 -BCH_ERR_btree_node_read_err_must_retry, 758 c, ca, b, i, 759 btree_node_bad_btree, 760 "incorrect btree id"); 761 762 btree_err_on(BTREE_NODE_LEVEL(bn) != b->c.level, 763 -BCH_ERR_btree_node_read_err_must_retry, 764 c, ca, b, i, 765 btree_node_bad_level, 766 "incorrect level"); 767 768 if (!write) 769 compat_btree_node(b->c.level, b->c.btree_id, version, 770 BSET_BIG_ENDIAN(i), write, bn); 771 772 if (b->key.k.type == KEY_TYPE_btree_ptr_v2) { 773 struct bch_btree_ptr_v2 *bp = 774 &bkey_i_to_btree_ptr_v2(&b->key)->v; 775 776 if (BTREE_PTR_RANGE_UPDATED(bp)) { 777 b->data->min_key = bp->min_key; 778 b->data->max_key = b->key.k.p; 779 } 780 781 btree_err_on(!bpos_eq(b->data->min_key, bp->min_key), 782 -BCH_ERR_btree_node_read_err_must_retry, 783 c, ca, b, NULL, 784 btree_node_bad_min_key, 785 "incorrect min_key: got %s should be %s", 786 (printbuf_reset(&buf1), 787 bch2_bpos_to_text(&buf1, bn->min_key), buf1.buf), 788 (printbuf_reset(&buf2), 789 bch2_bpos_to_text(&buf2, bp->min_key), buf2.buf)); 790 } 791 792 btree_err_on(!bpos_eq(bn->max_key, b->key.k.p), 793 -BCH_ERR_btree_node_read_err_must_retry, 794 c, ca, b, i, 795 btree_node_bad_max_key, 796 "incorrect max key %s", 797 (printbuf_reset(&buf1), 798 bch2_bpos_to_text(&buf1, bn->max_key), buf1.buf)); 799 800 if (write) 801 compat_btree_node(b->c.level, b->c.btree_id, version, 802 BSET_BIG_ENDIAN(i), write, bn); 803 804 btree_err_on(bch2_bkey_format_invalid(c, &bn->format, write, &buf1), 805 -BCH_ERR_btree_node_read_err_bad_node, 806 c, ca, b, i, 807 btree_node_bad_format, 808 "invalid bkey format: %s\n %s", buf1.buf, 809 (printbuf_reset(&buf2), 810 bch2_bkey_format_to_text(&buf2, &bn->format), buf2.buf)); 811 printbuf_reset(&buf1); 812 813 compat_bformat(b->c.level, b->c.btree_id, version, 814 BSET_BIG_ENDIAN(i), write, 815 &bn->format); 816 } 817 out: 818 fsck_err: 819 printbuf_exit(&buf2); 820 printbuf_exit(&buf1); 821 return ret; 822 } 823 824 static int bset_key_invalid(struct bch_fs *c, struct btree *b, 825 struct bkey_s_c k, 826 bool updated_range, int rw, 827 struct printbuf *err) 828 { 829 return __bch2_bkey_invalid(c, k, btree_node_type(b), READ, err) ?: 830 (!updated_range ? bch2_bkey_in_btree_node(c, b, k, err) : 0) ?: 831 (rw == WRITE ? bch2_bkey_val_invalid(c, k, READ, err) : 0); 832 } 833 834 static bool bkey_packed_valid(struct bch_fs *c, struct btree *b, 835 struct bset *i, struct bkey_packed *k) 836 { 837 if (bkey_p_next(k) > vstruct_last(i)) 838 return false; 839 840 if (k->format > KEY_FORMAT_CURRENT) 841 return false; 842 843 if (!bkeyp_u64s_valid(&b->format, k)) 844 return false; 845 846 struct printbuf buf = PRINTBUF; 847 struct bkey tmp; 848 struct bkey_s u = __bkey_disassemble(b, k, &tmp); 849 bool ret = __bch2_bkey_invalid(c, u.s_c, btree_node_type(b), READ, &buf); 850 printbuf_exit(&buf); 851 return ret; 852 } 853 854 static int validate_bset_keys(struct bch_fs *c, struct btree *b, 855 struct bset *i, int write, 856 bool have_retry, bool *saw_error) 857 { 858 unsigned version = le16_to_cpu(i->version); 859 struct bkey_packed *k, *prev = NULL; 860 struct printbuf buf = PRINTBUF; 861 bool updated_range = b->key.k.type == KEY_TYPE_btree_ptr_v2 && 862 BTREE_PTR_RANGE_UPDATED(&bkey_i_to_btree_ptr_v2(&b->key)->v); 863 int ret = 0; 864 865 for (k = i->start; 866 k != vstruct_last(i);) { 867 struct bkey_s u; 868 struct bkey tmp; 869 unsigned next_good_key; 870 871 if (btree_err_on(bkey_p_next(k) > vstruct_last(i), 872 -BCH_ERR_btree_node_read_err_fixable, 873 c, NULL, b, i, 874 btree_node_bkey_past_bset_end, 875 "key extends past end of bset")) { 876 i->u64s = cpu_to_le16((u64 *) k - i->_data); 877 break; 878 } 879 880 if (btree_err_on(k->format > KEY_FORMAT_CURRENT, 881 -BCH_ERR_btree_node_read_err_fixable, 882 c, NULL, b, i, 883 btree_node_bkey_bad_format, 884 "invalid bkey format %u", k->format)) 885 goto drop_this_key; 886 887 if (btree_err_on(!bkeyp_u64s_valid(&b->format, k), 888 -BCH_ERR_btree_node_read_err_fixable, 889 c, NULL, b, i, 890 btree_node_bkey_bad_u64s, 891 "bad k->u64s %u (min %u max %lu)", k->u64s, 892 bkeyp_key_u64s(&b->format, k), 893 U8_MAX - BKEY_U64s + bkeyp_key_u64s(&b->format, k))) 894 goto drop_this_key; 895 896 if (!write) 897 bch2_bkey_compat(b->c.level, b->c.btree_id, version, 898 BSET_BIG_ENDIAN(i), write, 899 &b->format, k); 900 901 u = __bkey_disassemble(b, k, &tmp); 902 903 printbuf_reset(&buf); 904 if (bset_key_invalid(c, b, u.s_c, updated_range, write, &buf)) { 905 printbuf_reset(&buf); 906 bset_key_invalid(c, b, u.s_c, updated_range, write, &buf); 907 prt_printf(&buf, "\n "); 908 bch2_bkey_val_to_text(&buf, c, u.s_c); 909 910 btree_err(-BCH_ERR_btree_node_read_err_fixable, 911 c, NULL, b, i, 912 btree_node_bad_bkey, 913 "invalid bkey: %s", buf.buf); 914 goto drop_this_key; 915 } 916 917 if (write) 918 bch2_bkey_compat(b->c.level, b->c.btree_id, version, 919 BSET_BIG_ENDIAN(i), write, 920 &b->format, k); 921 922 if (prev && bkey_iter_cmp(b, prev, k) > 0) { 923 struct bkey up = bkey_unpack_key(b, prev); 924 925 printbuf_reset(&buf); 926 prt_printf(&buf, "keys out of order: "); 927 bch2_bkey_to_text(&buf, &up); 928 prt_printf(&buf, " > "); 929 bch2_bkey_to_text(&buf, u.k); 930 931 if (btree_err(-BCH_ERR_btree_node_read_err_fixable, 932 c, NULL, b, i, 933 btree_node_bkey_out_of_order, 934 "%s", buf.buf)) 935 goto drop_this_key; 936 } 937 938 prev = k; 939 k = bkey_p_next(k); 940 continue; 941 drop_this_key: 942 next_good_key = k->u64s; 943 944 if (!next_good_key || 945 (BSET_BIG_ENDIAN(i) == CPU_BIG_ENDIAN && 946 version >= bcachefs_metadata_version_snapshot)) { 947 /* 948 * only do scanning if bch2_bkey_compat() has nothing to 949 * do 950 */ 951 952 if (!bkey_packed_valid(c, b, i, (void *) ((u64 *) k + next_good_key))) { 953 for (next_good_key = 1; 954 next_good_key < (u64 *) vstruct_last(i) - (u64 *) k; 955 next_good_key++) 956 if (bkey_packed_valid(c, b, i, (void *) ((u64 *) k + next_good_key))) 957 goto got_good_key; 958 } 959 960 /* 961 * didn't find a good key, have to truncate the rest of 962 * the bset 963 */ 964 next_good_key = (u64 *) vstruct_last(i) - (u64 *) k; 965 } 966 got_good_key: 967 le16_add_cpu(&i->u64s, -next_good_key); 968 memmove_u64s_down(k, bkey_p_next(k), (u64 *) vstruct_end(i) - (u64 *) k); 969 } 970 fsck_err: 971 printbuf_exit(&buf); 972 return ret; 973 } 974 975 int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca, 976 struct btree *b, bool have_retry, bool *saw_error) 977 { 978 struct btree_node_entry *bne; 979 struct sort_iter *iter; 980 struct btree_node *sorted; 981 struct bkey_packed *k; 982 struct bset *i; 983 bool used_mempool, blacklisted; 984 bool updated_range = b->key.k.type == KEY_TYPE_btree_ptr_v2 && 985 BTREE_PTR_RANGE_UPDATED(&bkey_i_to_btree_ptr_v2(&b->key)->v); 986 unsigned u64s; 987 unsigned ptr_written = btree_ptr_sectors_written(&b->key); 988 struct printbuf buf = PRINTBUF; 989 int ret = 0, retry_read = 0, write = READ; 990 u64 start_time = local_clock(); 991 992 b->version_ondisk = U16_MAX; 993 /* We might get called multiple times on read retry: */ 994 b->written = 0; 995 996 iter = mempool_alloc(&c->fill_iter, GFP_NOFS); 997 sort_iter_init(iter, b, (btree_blocks(c) + 1) * 2); 998 999 if (bch2_meta_read_fault("btree")) 1000 btree_err(-BCH_ERR_btree_node_read_err_must_retry, 1001 c, ca, b, NULL, 1002 btree_node_fault_injected, 1003 "dynamic fault"); 1004 1005 btree_err_on(le64_to_cpu(b->data->magic) != bset_magic(c), 1006 -BCH_ERR_btree_node_read_err_must_retry, 1007 c, ca, b, NULL, 1008 btree_node_bad_magic, 1009 "bad magic: want %llx, got %llx", 1010 bset_magic(c), le64_to_cpu(b->data->magic)); 1011 1012 if (b->key.k.type == KEY_TYPE_btree_ptr_v2) { 1013 struct bch_btree_ptr_v2 *bp = 1014 &bkey_i_to_btree_ptr_v2(&b->key)->v; 1015 1016 bch2_bpos_to_text(&buf, b->data->min_key); 1017 prt_str(&buf, "-"); 1018 bch2_bpos_to_text(&buf, b->data->max_key); 1019 1020 btree_err_on(b->data->keys.seq != bp->seq, 1021 -BCH_ERR_btree_node_read_err_must_retry, 1022 c, ca, b, NULL, 1023 btree_node_bad_seq, 1024 "got wrong btree node (want %llx got %llx)\n" 1025 "got btree %s level %llu pos %s", 1026 bp->seq, b->data->keys.seq, 1027 bch2_btree_id_str(BTREE_NODE_ID(b->data)), 1028 BTREE_NODE_LEVEL(b->data), 1029 buf.buf); 1030 } else { 1031 btree_err_on(!b->data->keys.seq, 1032 -BCH_ERR_btree_node_read_err_must_retry, 1033 c, ca, b, NULL, 1034 btree_node_bad_seq, 1035 "bad btree header: seq 0"); 1036 } 1037 1038 while (b->written < (ptr_written ?: btree_sectors(c))) { 1039 unsigned sectors; 1040 struct nonce nonce; 1041 bool first = !b->written; 1042 bool csum_bad; 1043 1044 if (!b->written) { 1045 i = &b->data->keys; 1046 1047 btree_err_on(!bch2_checksum_type_valid(c, BSET_CSUM_TYPE(i)), 1048 -BCH_ERR_btree_node_read_err_want_retry, 1049 c, ca, b, i, 1050 bset_unknown_csum, 1051 "unknown checksum type %llu", BSET_CSUM_TYPE(i)); 1052 1053 nonce = btree_nonce(i, b->written << 9); 1054 1055 struct bch_csum csum = csum_vstruct(c, BSET_CSUM_TYPE(i), nonce, b->data); 1056 csum_bad = bch2_crc_cmp(b->data->csum, csum); 1057 if (csum_bad) 1058 bch2_io_error(ca, BCH_MEMBER_ERROR_checksum); 1059 1060 btree_err_on(csum_bad, 1061 -BCH_ERR_btree_node_read_err_want_retry, 1062 c, ca, b, i, 1063 bset_bad_csum, 1064 "%s", 1065 (printbuf_reset(&buf), 1066 bch2_csum_err_msg(&buf, BSET_CSUM_TYPE(i), b->data->csum, csum), 1067 buf.buf)); 1068 1069 ret = bset_encrypt(c, i, b->written << 9); 1070 if (bch2_fs_fatal_err_on(ret, c, 1071 "decrypting btree node: %s", bch2_err_str(ret))) 1072 goto fsck_err; 1073 1074 btree_err_on(btree_node_type_is_extents(btree_node_type(b)) && 1075 !BTREE_NODE_NEW_EXTENT_OVERWRITE(b->data), 1076 -BCH_ERR_btree_node_read_err_incompatible, 1077 c, NULL, b, NULL, 1078 btree_node_unsupported_version, 1079 "btree node does not have NEW_EXTENT_OVERWRITE set"); 1080 1081 sectors = vstruct_sectors(b->data, c->block_bits); 1082 } else { 1083 bne = write_block(b); 1084 i = &bne->keys; 1085 1086 if (i->seq != b->data->keys.seq) 1087 break; 1088 1089 btree_err_on(!bch2_checksum_type_valid(c, BSET_CSUM_TYPE(i)), 1090 -BCH_ERR_btree_node_read_err_want_retry, 1091 c, ca, b, i, 1092 bset_unknown_csum, 1093 "unknown checksum type %llu", BSET_CSUM_TYPE(i)); 1094 1095 nonce = btree_nonce(i, b->written << 9); 1096 struct bch_csum csum = csum_vstruct(c, BSET_CSUM_TYPE(i), nonce, bne); 1097 csum_bad = bch2_crc_cmp(bne->csum, csum); 1098 if (csum_bad) 1099 bch2_io_error(ca, BCH_MEMBER_ERROR_checksum); 1100 1101 btree_err_on(csum_bad, 1102 -BCH_ERR_btree_node_read_err_want_retry, 1103 c, ca, b, i, 1104 bset_bad_csum, 1105 "%s", 1106 (printbuf_reset(&buf), 1107 bch2_csum_err_msg(&buf, BSET_CSUM_TYPE(i), bne->csum, csum), 1108 buf.buf)); 1109 1110 ret = bset_encrypt(c, i, b->written << 9); 1111 if (bch2_fs_fatal_err_on(ret, c, 1112 "decrypting btree node: %s", bch2_err_str(ret))) 1113 goto fsck_err; 1114 1115 sectors = vstruct_sectors(bne, c->block_bits); 1116 } 1117 1118 b->version_ondisk = min(b->version_ondisk, 1119 le16_to_cpu(i->version)); 1120 1121 ret = validate_bset(c, ca, b, i, b->written, sectors, 1122 READ, have_retry, saw_error); 1123 if (ret) 1124 goto fsck_err; 1125 1126 if (!b->written) 1127 btree_node_set_format(b, b->data->format); 1128 1129 ret = validate_bset_keys(c, b, i, READ, have_retry, saw_error); 1130 if (ret) 1131 goto fsck_err; 1132 1133 SET_BSET_BIG_ENDIAN(i, CPU_BIG_ENDIAN); 1134 1135 blacklisted = bch2_journal_seq_is_blacklisted(c, 1136 le64_to_cpu(i->journal_seq), 1137 true); 1138 1139 btree_err_on(blacklisted && first, 1140 -BCH_ERR_btree_node_read_err_fixable, 1141 c, ca, b, i, 1142 bset_blacklisted_journal_seq, 1143 "first btree node bset has blacklisted journal seq (%llu)", 1144 le64_to_cpu(i->journal_seq)); 1145 1146 btree_err_on(blacklisted && ptr_written, 1147 -BCH_ERR_btree_node_read_err_fixable, 1148 c, ca, b, i, 1149 first_bset_blacklisted_journal_seq, 1150 "found blacklisted bset (journal seq %llu) in btree node at offset %u-%u/%u", 1151 le64_to_cpu(i->journal_seq), 1152 b->written, b->written + sectors, ptr_written); 1153 1154 b->written += sectors; 1155 1156 if (blacklisted && !first) 1157 continue; 1158 1159 sort_iter_add(iter, 1160 vstruct_idx(i, 0), 1161 vstruct_last(i)); 1162 } 1163 1164 if (ptr_written) { 1165 btree_err_on(b->written < ptr_written, 1166 -BCH_ERR_btree_node_read_err_want_retry, 1167 c, ca, b, NULL, 1168 btree_node_data_missing, 1169 "btree node data missing: expected %u sectors, found %u", 1170 ptr_written, b->written); 1171 } else { 1172 for (bne = write_block(b); 1173 bset_byte_offset(b, bne) < btree_buf_bytes(b); 1174 bne = (void *) bne + block_bytes(c)) 1175 btree_err_on(bne->keys.seq == b->data->keys.seq && 1176 !bch2_journal_seq_is_blacklisted(c, 1177 le64_to_cpu(bne->keys.journal_seq), 1178 true), 1179 -BCH_ERR_btree_node_read_err_want_retry, 1180 c, ca, b, NULL, 1181 btree_node_bset_after_end, 1182 "found bset signature after last bset"); 1183 } 1184 1185 sorted = btree_bounce_alloc(c, btree_buf_bytes(b), &used_mempool); 1186 sorted->keys.u64s = 0; 1187 1188 set_btree_bset(b, b->set, &b->data->keys); 1189 1190 b->nr = bch2_key_sort_fix_overlapping(c, &sorted->keys, iter); 1191 1192 u64s = le16_to_cpu(sorted->keys.u64s); 1193 *sorted = *b->data; 1194 sorted->keys.u64s = cpu_to_le16(u64s); 1195 swap(sorted, b->data); 1196 set_btree_bset(b, b->set, &b->data->keys); 1197 b->nsets = 1; 1198 1199 BUG_ON(b->nr.live_u64s != u64s); 1200 1201 btree_bounce_free(c, btree_buf_bytes(b), used_mempool, sorted); 1202 1203 if (updated_range) 1204 bch2_btree_node_drop_keys_outside_node(b); 1205 1206 i = &b->data->keys; 1207 for (k = i->start; k != vstruct_last(i);) { 1208 struct bkey tmp; 1209 struct bkey_s u = __bkey_disassemble(b, k, &tmp); 1210 1211 printbuf_reset(&buf); 1212 1213 if (bch2_bkey_val_invalid(c, u.s_c, READ, &buf) || 1214 (bch2_inject_invalid_keys && 1215 !bversion_cmp(u.k->version, MAX_VERSION))) { 1216 printbuf_reset(&buf); 1217 1218 prt_printf(&buf, "invalid bkey: "); 1219 bch2_bkey_val_invalid(c, u.s_c, READ, &buf); 1220 prt_printf(&buf, "\n "); 1221 bch2_bkey_val_to_text(&buf, c, u.s_c); 1222 1223 btree_err(-BCH_ERR_btree_node_read_err_fixable, 1224 c, NULL, b, i, 1225 btree_node_bad_bkey, 1226 "%s", buf.buf); 1227 1228 btree_keys_account_key_drop(&b->nr, 0, k); 1229 1230 i->u64s = cpu_to_le16(le16_to_cpu(i->u64s) - k->u64s); 1231 memmove_u64s_down(k, bkey_p_next(k), 1232 (u64 *) vstruct_end(i) - (u64 *) k); 1233 set_btree_bset_end(b, b->set); 1234 continue; 1235 } 1236 1237 if (u.k->type == KEY_TYPE_btree_ptr_v2) { 1238 struct bkey_s_btree_ptr_v2 bp = bkey_s_to_btree_ptr_v2(u); 1239 1240 bp.v->mem_ptr = 0; 1241 } 1242 1243 k = bkey_p_next(k); 1244 } 1245 1246 bch2_bset_build_aux_tree(b, b->set, false); 1247 1248 set_needs_whiteout(btree_bset_first(b), true); 1249 1250 btree_node_reset_sib_u64s(b); 1251 1252 bkey_for_each_ptr(bch2_bkey_ptrs(bkey_i_to_s(&b->key)), ptr) { 1253 struct bch_dev *ca2 = bch_dev_bkey_exists(c, ptr->dev); 1254 1255 if (ca2->mi.state != BCH_MEMBER_STATE_rw) 1256 set_btree_node_need_rewrite(b); 1257 } 1258 1259 if (!ptr_written) 1260 set_btree_node_need_rewrite(b); 1261 out: 1262 mempool_free(iter, &c->fill_iter); 1263 printbuf_exit(&buf); 1264 bch2_time_stats_update(&c->times[BCH_TIME_btree_node_read_done], start_time); 1265 return retry_read; 1266 fsck_err: 1267 if (ret == -BCH_ERR_btree_node_read_err_want_retry || 1268 ret == -BCH_ERR_btree_node_read_err_must_retry) { 1269 retry_read = 1; 1270 } else { 1271 set_btree_node_read_error(b); 1272 bch2_btree_lost_data(c, b->c.btree_id); 1273 } 1274 goto out; 1275 } 1276 1277 static void btree_node_read_work(struct work_struct *work) 1278 { 1279 struct btree_read_bio *rb = 1280 container_of(work, struct btree_read_bio, work); 1281 struct bch_fs *c = rb->c; 1282 struct btree *b = rb->b; 1283 struct bch_dev *ca = bch_dev_bkey_exists(c, rb->pick.ptr.dev); 1284 struct bio *bio = &rb->bio; 1285 struct bch_io_failures failed = { .nr = 0 }; 1286 struct printbuf buf = PRINTBUF; 1287 bool saw_error = false; 1288 bool retry = false; 1289 bool can_retry; 1290 1291 goto start; 1292 while (1) { 1293 retry = true; 1294 bch_info(c, "retrying read"); 1295 ca = bch_dev_bkey_exists(c, rb->pick.ptr.dev); 1296 rb->have_ioref = bch2_dev_get_ioref(ca, READ); 1297 bio_reset(bio, NULL, REQ_OP_READ|REQ_SYNC|REQ_META); 1298 bio->bi_iter.bi_sector = rb->pick.ptr.offset; 1299 bio->bi_iter.bi_size = btree_buf_bytes(b); 1300 1301 if (rb->have_ioref) { 1302 bio_set_dev(bio, ca->disk_sb.bdev); 1303 submit_bio_wait(bio); 1304 } else { 1305 bio->bi_status = BLK_STS_REMOVED; 1306 } 1307 start: 1308 printbuf_reset(&buf); 1309 bch2_btree_pos_to_text(&buf, c, b); 1310 bch2_dev_io_err_on(bio->bi_status, ca, BCH_MEMBER_ERROR_read, 1311 "btree read error %s for %s", 1312 bch2_blk_status_to_str(bio->bi_status), buf.buf); 1313 if (rb->have_ioref) 1314 percpu_ref_put(&ca->io_ref); 1315 rb->have_ioref = false; 1316 1317 bch2_mark_io_failure(&failed, &rb->pick); 1318 1319 can_retry = bch2_bkey_pick_read_device(c, 1320 bkey_i_to_s_c(&b->key), 1321 &failed, &rb->pick) > 0; 1322 1323 if (!bio->bi_status && 1324 !bch2_btree_node_read_done(c, ca, b, can_retry, &saw_error)) { 1325 if (retry) 1326 bch_info(c, "retry success"); 1327 break; 1328 } 1329 1330 saw_error = true; 1331 1332 if (!can_retry) { 1333 set_btree_node_read_error(b); 1334 bch2_btree_lost_data(c, b->c.btree_id); 1335 break; 1336 } 1337 } 1338 1339 bch2_time_stats_update(&c->times[BCH_TIME_btree_node_read], 1340 rb->start_time); 1341 bio_put(&rb->bio); 1342 1343 if (saw_error && 1344 !btree_node_read_error(b) && 1345 c->curr_recovery_pass != BCH_RECOVERY_PASS_scan_for_btree_nodes) { 1346 printbuf_reset(&buf); 1347 bch2_bpos_to_text(&buf, b->key.k.p); 1348 bch_err_ratelimited(c, "%s: rewriting btree node at btree=%s level=%u %s due to error", 1349 __func__, bch2_btree_id_str(b->c.btree_id), b->c.level, buf.buf); 1350 1351 bch2_btree_node_rewrite_async(c, b); 1352 } 1353 1354 printbuf_exit(&buf); 1355 clear_btree_node_read_in_flight(b); 1356 wake_up_bit(&b->flags, BTREE_NODE_read_in_flight); 1357 } 1358 1359 static void btree_node_read_endio(struct bio *bio) 1360 { 1361 struct btree_read_bio *rb = 1362 container_of(bio, struct btree_read_bio, bio); 1363 struct bch_fs *c = rb->c; 1364 1365 if (rb->have_ioref) { 1366 struct bch_dev *ca = bch_dev_bkey_exists(c, rb->pick.ptr.dev); 1367 1368 bch2_latency_acct(ca, rb->start_time, READ); 1369 } 1370 1371 queue_work(c->io_complete_wq, &rb->work); 1372 } 1373 1374 struct btree_node_read_all { 1375 struct closure cl; 1376 struct bch_fs *c; 1377 struct btree *b; 1378 unsigned nr; 1379 void *buf[BCH_REPLICAS_MAX]; 1380 struct bio *bio[BCH_REPLICAS_MAX]; 1381 blk_status_t err[BCH_REPLICAS_MAX]; 1382 }; 1383 1384 static unsigned btree_node_sectors_written(struct bch_fs *c, void *data) 1385 { 1386 struct btree_node *bn = data; 1387 struct btree_node_entry *bne; 1388 unsigned offset = 0; 1389 1390 if (le64_to_cpu(bn->magic) != bset_magic(c)) 1391 return 0; 1392 1393 while (offset < btree_sectors(c)) { 1394 if (!offset) { 1395 offset += vstruct_sectors(bn, c->block_bits); 1396 } else { 1397 bne = data + (offset << 9); 1398 if (bne->keys.seq != bn->keys.seq) 1399 break; 1400 offset += vstruct_sectors(bne, c->block_bits); 1401 } 1402 } 1403 1404 return offset; 1405 } 1406 1407 static bool btree_node_has_extra_bsets(struct bch_fs *c, unsigned offset, void *data) 1408 { 1409 struct btree_node *bn = data; 1410 struct btree_node_entry *bne; 1411 1412 if (!offset) 1413 return false; 1414 1415 while (offset < btree_sectors(c)) { 1416 bne = data + (offset << 9); 1417 if (bne->keys.seq == bn->keys.seq) 1418 return true; 1419 offset++; 1420 } 1421 1422 return false; 1423 return offset; 1424 } 1425 1426 static CLOSURE_CALLBACK(btree_node_read_all_replicas_done) 1427 { 1428 closure_type(ra, struct btree_node_read_all, cl); 1429 struct bch_fs *c = ra->c; 1430 struct btree *b = ra->b; 1431 struct printbuf buf = PRINTBUF; 1432 bool dump_bset_maps = false; 1433 bool have_retry = false; 1434 int ret = 0, best = -1, write = READ; 1435 unsigned i, written = 0, written2 = 0; 1436 __le64 seq = b->key.k.type == KEY_TYPE_btree_ptr_v2 1437 ? bkey_i_to_btree_ptr_v2(&b->key)->v.seq : 0; 1438 bool _saw_error = false, *saw_error = &_saw_error; 1439 1440 for (i = 0; i < ra->nr; i++) { 1441 struct btree_node *bn = ra->buf[i]; 1442 1443 if (ra->err[i]) 1444 continue; 1445 1446 if (le64_to_cpu(bn->magic) != bset_magic(c) || 1447 (seq && seq != bn->keys.seq)) 1448 continue; 1449 1450 if (best < 0) { 1451 best = i; 1452 written = btree_node_sectors_written(c, bn); 1453 continue; 1454 } 1455 1456 written2 = btree_node_sectors_written(c, ra->buf[i]); 1457 if (btree_err_on(written2 != written, -BCH_ERR_btree_node_read_err_fixable, 1458 c, NULL, b, NULL, 1459 btree_node_replicas_sectors_written_mismatch, 1460 "btree node sectors written mismatch: %u != %u", 1461 written, written2) || 1462 btree_err_on(btree_node_has_extra_bsets(c, written2, ra->buf[i]), 1463 -BCH_ERR_btree_node_read_err_fixable, 1464 c, NULL, b, NULL, 1465 btree_node_bset_after_end, 1466 "found bset signature after last bset") || 1467 btree_err_on(memcmp(ra->buf[best], ra->buf[i], written << 9), 1468 -BCH_ERR_btree_node_read_err_fixable, 1469 c, NULL, b, NULL, 1470 btree_node_replicas_data_mismatch, 1471 "btree node replicas content mismatch")) 1472 dump_bset_maps = true; 1473 1474 if (written2 > written) { 1475 written = written2; 1476 best = i; 1477 } 1478 } 1479 fsck_err: 1480 if (dump_bset_maps) { 1481 for (i = 0; i < ra->nr; i++) { 1482 struct btree_node *bn = ra->buf[i]; 1483 struct btree_node_entry *bne = NULL; 1484 unsigned offset = 0, sectors; 1485 bool gap = false; 1486 1487 if (ra->err[i]) 1488 continue; 1489 1490 printbuf_reset(&buf); 1491 1492 while (offset < btree_sectors(c)) { 1493 if (!offset) { 1494 sectors = vstruct_sectors(bn, c->block_bits); 1495 } else { 1496 bne = ra->buf[i] + (offset << 9); 1497 if (bne->keys.seq != bn->keys.seq) 1498 break; 1499 sectors = vstruct_sectors(bne, c->block_bits); 1500 } 1501 1502 prt_printf(&buf, " %u-%u", offset, offset + sectors); 1503 if (bne && bch2_journal_seq_is_blacklisted(c, 1504 le64_to_cpu(bne->keys.journal_seq), false)) 1505 prt_printf(&buf, "*"); 1506 offset += sectors; 1507 } 1508 1509 while (offset < btree_sectors(c)) { 1510 bne = ra->buf[i] + (offset << 9); 1511 if (bne->keys.seq == bn->keys.seq) { 1512 if (!gap) 1513 prt_printf(&buf, " GAP"); 1514 gap = true; 1515 1516 sectors = vstruct_sectors(bne, c->block_bits); 1517 prt_printf(&buf, " %u-%u", offset, offset + sectors); 1518 if (bch2_journal_seq_is_blacklisted(c, 1519 le64_to_cpu(bne->keys.journal_seq), false)) 1520 prt_printf(&buf, "*"); 1521 } 1522 offset++; 1523 } 1524 1525 bch_err(c, "replica %u:%s", i, buf.buf); 1526 } 1527 } 1528 1529 if (best >= 0) { 1530 memcpy(b->data, ra->buf[best], btree_buf_bytes(b)); 1531 ret = bch2_btree_node_read_done(c, NULL, b, false, saw_error); 1532 } else { 1533 ret = -1; 1534 } 1535 1536 if (ret) { 1537 set_btree_node_read_error(b); 1538 bch2_btree_lost_data(c, b->c.btree_id); 1539 } else if (*saw_error) 1540 bch2_btree_node_rewrite_async(c, b); 1541 1542 for (i = 0; i < ra->nr; i++) { 1543 mempool_free(ra->buf[i], &c->btree_bounce_pool); 1544 bio_put(ra->bio[i]); 1545 } 1546 1547 closure_debug_destroy(&ra->cl); 1548 kfree(ra); 1549 printbuf_exit(&buf); 1550 1551 clear_btree_node_read_in_flight(b); 1552 wake_up_bit(&b->flags, BTREE_NODE_read_in_flight); 1553 } 1554 1555 static void btree_node_read_all_replicas_endio(struct bio *bio) 1556 { 1557 struct btree_read_bio *rb = 1558 container_of(bio, struct btree_read_bio, bio); 1559 struct bch_fs *c = rb->c; 1560 struct btree_node_read_all *ra = rb->ra; 1561 1562 if (rb->have_ioref) { 1563 struct bch_dev *ca = bch_dev_bkey_exists(c, rb->pick.ptr.dev); 1564 1565 bch2_latency_acct(ca, rb->start_time, READ); 1566 } 1567 1568 ra->err[rb->idx] = bio->bi_status; 1569 closure_put(&ra->cl); 1570 } 1571 1572 /* 1573 * XXX This allocates multiple times from the same mempools, and can deadlock 1574 * under sufficient memory pressure (but is only a debug path) 1575 */ 1576 static int btree_node_read_all_replicas(struct bch_fs *c, struct btree *b, bool sync) 1577 { 1578 struct bkey_s_c k = bkey_i_to_s_c(&b->key); 1579 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 1580 const union bch_extent_entry *entry; 1581 struct extent_ptr_decoded pick; 1582 struct btree_node_read_all *ra; 1583 unsigned i; 1584 1585 ra = kzalloc(sizeof(*ra), GFP_NOFS); 1586 if (!ra) 1587 return -BCH_ERR_ENOMEM_btree_node_read_all_replicas; 1588 1589 closure_init(&ra->cl, NULL); 1590 ra->c = c; 1591 ra->b = b; 1592 ra->nr = bch2_bkey_nr_ptrs(k); 1593 1594 for (i = 0; i < ra->nr; i++) { 1595 ra->buf[i] = mempool_alloc(&c->btree_bounce_pool, GFP_NOFS); 1596 ra->bio[i] = bio_alloc_bioset(NULL, 1597 buf_pages(ra->buf[i], btree_buf_bytes(b)), 1598 REQ_OP_READ|REQ_SYNC|REQ_META, 1599 GFP_NOFS, 1600 &c->btree_bio); 1601 } 1602 1603 i = 0; 1604 bkey_for_each_ptr_decode(k.k, ptrs, pick, entry) { 1605 struct bch_dev *ca = bch_dev_bkey_exists(c, pick.ptr.dev); 1606 struct btree_read_bio *rb = 1607 container_of(ra->bio[i], struct btree_read_bio, bio); 1608 rb->c = c; 1609 rb->b = b; 1610 rb->ra = ra; 1611 rb->start_time = local_clock(); 1612 rb->have_ioref = bch2_dev_get_ioref(ca, READ); 1613 rb->idx = i; 1614 rb->pick = pick; 1615 rb->bio.bi_iter.bi_sector = pick.ptr.offset; 1616 rb->bio.bi_end_io = btree_node_read_all_replicas_endio; 1617 bch2_bio_map(&rb->bio, ra->buf[i], btree_buf_bytes(b)); 1618 1619 if (rb->have_ioref) { 1620 this_cpu_add(ca->io_done->sectors[READ][BCH_DATA_btree], 1621 bio_sectors(&rb->bio)); 1622 bio_set_dev(&rb->bio, ca->disk_sb.bdev); 1623 1624 closure_get(&ra->cl); 1625 submit_bio(&rb->bio); 1626 } else { 1627 ra->err[i] = BLK_STS_REMOVED; 1628 } 1629 1630 i++; 1631 } 1632 1633 if (sync) { 1634 closure_sync(&ra->cl); 1635 btree_node_read_all_replicas_done(&ra->cl.work); 1636 } else { 1637 continue_at(&ra->cl, btree_node_read_all_replicas_done, 1638 c->io_complete_wq); 1639 } 1640 1641 return 0; 1642 } 1643 1644 void bch2_btree_node_read(struct btree_trans *trans, struct btree *b, 1645 bool sync) 1646 { 1647 struct bch_fs *c = trans->c; 1648 struct extent_ptr_decoded pick; 1649 struct btree_read_bio *rb; 1650 struct bch_dev *ca; 1651 struct bio *bio; 1652 int ret; 1653 1654 trace_and_count(c, btree_node_read, trans, b); 1655 1656 if (bch2_verify_all_btree_replicas && 1657 !btree_node_read_all_replicas(c, b, sync)) 1658 return; 1659 1660 ret = bch2_bkey_pick_read_device(c, bkey_i_to_s_c(&b->key), 1661 NULL, &pick); 1662 1663 if (ret <= 0) { 1664 struct printbuf buf = PRINTBUF; 1665 1666 prt_str(&buf, "btree node read error: no device to read from\n at "); 1667 bch2_btree_pos_to_text(&buf, c, b); 1668 bch_err_ratelimited(c, "%s", buf.buf); 1669 1670 if (c->recovery_passes_explicit & BIT_ULL(BCH_RECOVERY_PASS_check_topology) && 1671 c->curr_recovery_pass > BCH_RECOVERY_PASS_check_topology) 1672 bch2_fatal_error(c); 1673 1674 set_btree_node_read_error(b); 1675 bch2_btree_lost_data(c, b->c.btree_id); 1676 clear_btree_node_read_in_flight(b); 1677 wake_up_bit(&b->flags, BTREE_NODE_read_in_flight); 1678 printbuf_exit(&buf); 1679 return; 1680 } 1681 1682 ca = bch_dev_bkey_exists(c, pick.ptr.dev); 1683 1684 bio = bio_alloc_bioset(NULL, 1685 buf_pages(b->data, btree_buf_bytes(b)), 1686 REQ_OP_READ|REQ_SYNC|REQ_META, 1687 GFP_NOFS, 1688 &c->btree_bio); 1689 rb = container_of(bio, struct btree_read_bio, bio); 1690 rb->c = c; 1691 rb->b = b; 1692 rb->ra = NULL; 1693 rb->start_time = local_clock(); 1694 rb->have_ioref = bch2_dev_get_ioref(ca, READ); 1695 rb->pick = pick; 1696 INIT_WORK(&rb->work, btree_node_read_work); 1697 bio->bi_iter.bi_sector = pick.ptr.offset; 1698 bio->bi_end_io = btree_node_read_endio; 1699 bch2_bio_map(bio, b->data, btree_buf_bytes(b)); 1700 1701 if (rb->have_ioref) { 1702 this_cpu_add(ca->io_done->sectors[READ][BCH_DATA_btree], 1703 bio_sectors(bio)); 1704 bio_set_dev(bio, ca->disk_sb.bdev); 1705 1706 if (sync) { 1707 submit_bio_wait(bio); 1708 bch2_latency_acct(ca, rb->start_time, READ); 1709 btree_node_read_work(&rb->work); 1710 } else { 1711 submit_bio(bio); 1712 } 1713 } else { 1714 bio->bi_status = BLK_STS_REMOVED; 1715 1716 if (sync) 1717 btree_node_read_work(&rb->work); 1718 else 1719 queue_work(c->io_complete_wq, &rb->work); 1720 } 1721 } 1722 1723 static int __bch2_btree_root_read(struct btree_trans *trans, enum btree_id id, 1724 const struct bkey_i *k, unsigned level) 1725 { 1726 struct bch_fs *c = trans->c; 1727 struct closure cl; 1728 struct btree *b; 1729 int ret; 1730 1731 closure_init_stack(&cl); 1732 1733 do { 1734 ret = bch2_btree_cache_cannibalize_lock(trans, &cl); 1735 closure_sync(&cl); 1736 } while (ret); 1737 1738 b = bch2_btree_node_mem_alloc(trans, level != 0); 1739 bch2_btree_cache_cannibalize_unlock(trans); 1740 1741 BUG_ON(IS_ERR(b)); 1742 1743 bkey_copy(&b->key, k); 1744 BUG_ON(bch2_btree_node_hash_insert(&c->btree_cache, b, level, id)); 1745 1746 set_btree_node_read_in_flight(b); 1747 1748 bch2_btree_node_read(trans, b, true); 1749 1750 if (btree_node_read_error(b)) { 1751 bch2_btree_node_hash_remove(&c->btree_cache, b); 1752 1753 mutex_lock(&c->btree_cache.lock); 1754 list_move(&b->list, &c->btree_cache.freeable); 1755 mutex_unlock(&c->btree_cache.lock); 1756 1757 ret = -BCH_ERR_btree_node_read_error; 1758 goto err; 1759 } 1760 1761 bch2_btree_set_root_for_read(c, b); 1762 err: 1763 six_unlock_write(&b->c.lock); 1764 six_unlock_intent(&b->c.lock); 1765 1766 return ret; 1767 } 1768 1769 int bch2_btree_root_read(struct bch_fs *c, enum btree_id id, 1770 const struct bkey_i *k, unsigned level) 1771 { 1772 return bch2_trans_run(c, __bch2_btree_root_read(trans, id, k, level)); 1773 } 1774 1775 static void bch2_btree_complete_write(struct bch_fs *c, struct btree *b, 1776 struct btree_write *w) 1777 { 1778 unsigned long old, new, v = READ_ONCE(b->will_make_reachable); 1779 1780 do { 1781 old = new = v; 1782 if (!(old & 1)) 1783 break; 1784 1785 new &= ~1UL; 1786 } while ((v = cmpxchg(&b->will_make_reachable, old, new)) != old); 1787 1788 if (old & 1) 1789 closure_put(&((struct btree_update *) new)->cl); 1790 1791 bch2_journal_pin_drop(&c->journal, &w->journal); 1792 } 1793 1794 static void __btree_node_write_done(struct bch_fs *c, struct btree *b) 1795 { 1796 struct btree_write *w = btree_prev_write(b); 1797 unsigned long old, new, v; 1798 unsigned type = 0; 1799 1800 bch2_btree_complete_write(c, b, w); 1801 1802 v = READ_ONCE(b->flags); 1803 do { 1804 old = new = v; 1805 1806 if ((old & (1U << BTREE_NODE_dirty)) && 1807 (old & (1U << BTREE_NODE_need_write)) && 1808 !(old & (1U << BTREE_NODE_never_write)) && 1809 !(old & (1U << BTREE_NODE_write_blocked)) && 1810 !(old & (1U << BTREE_NODE_will_make_reachable))) { 1811 new &= ~(1U << BTREE_NODE_dirty); 1812 new &= ~(1U << BTREE_NODE_need_write); 1813 new |= (1U << BTREE_NODE_write_in_flight); 1814 new |= (1U << BTREE_NODE_write_in_flight_inner); 1815 new |= (1U << BTREE_NODE_just_written); 1816 new ^= (1U << BTREE_NODE_write_idx); 1817 1818 type = new & BTREE_WRITE_TYPE_MASK; 1819 new &= ~BTREE_WRITE_TYPE_MASK; 1820 } else { 1821 new &= ~(1U << BTREE_NODE_write_in_flight); 1822 new &= ~(1U << BTREE_NODE_write_in_flight_inner); 1823 } 1824 } while ((v = cmpxchg(&b->flags, old, new)) != old); 1825 1826 if (new & (1U << BTREE_NODE_write_in_flight)) 1827 __bch2_btree_node_write(c, b, BTREE_WRITE_ALREADY_STARTED|type); 1828 else 1829 wake_up_bit(&b->flags, BTREE_NODE_write_in_flight); 1830 } 1831 1832 static void btree_node_write_done(struct bch_fs *c, struct btree *b) 1833 { 1834 struct btree_trans *trans = bch2_trans_get(c); 1835 1836 btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_read); 1837 __btree_node_write_done(c, b); 1838 six_unlock_read(&b->c.lock); 1839 1840 bch2_trans_put(trans); 1841 } 1842 1843 static void btree_node_write_work(struct work_struct *work) 1844 { 1845 struct btree_write_bio *wbio = 1846 container_of(work, struct btree_write_bio, work); 1847 struct bch_fs *c = wbio->wbio.c; 1848 struct btree *b = wbio->wbio.bio.bi_private; 1849 struct bch_extent_ptr *ptr; 1850 int ret = 0; 1851 1852 btree_bounce_free(c, 1853 wbio->data_bytes, 1854 wbio->wbio.used_mempool, 1855 wbio->data); 1856 1857 bch2_bkey_drop_ptrs(bkey_i_to_s(&wbio->key), ptr, 1858 bch2_dev_list_has_dev(wbio->wbio.failed, ptr->dev)); 1859 1860 if (!bch2_bkey_nr_ptrs(bkey_i_to_s_c(&wbio->key))) { 1861 ret = -BCH_ERR_btree_node_write_all_failed; 1862 goto err; 1863 } 1864 1865 if (wbio->wbio.first_btree_write) { 1866 if (wbio->wbio.failed.nr) { 1867 1868 } 1869 } else { 1870 ret = bch2_trans_do(c, NULL, NULL, 0, 1871 bch2_btree_node_update_key_get_iter(trans, b, &wbio->key, 1872 BCH_WATERMARK_interior_updates| 1873 BCH_TRANS_COMMIT_journal_reclaim| 1874 BCH_TRANS_COMMIT_no_enospc| 1875 BCH_TRANS_COMMIT_no_check_rw, 1876 !wbio->wbio.failed.nr)); 1877 if (ret) 1878 goto err; 1879 } 1880 out: 1881 bio_put(&wbio->wbio.bio); 1882 btree_node_write_done(c, b); 1883 return; 1884 err: 1885 set_btree_node_noevict(b); 1886 bch2_fs_fatal_err_on(!bch2_err_matches(ret, EROFS), c, 1887 "writing btree node: %s", bch2_err_str(ret)); 1888 goto out; 1889 } 1890 1891 static void btree_node_write_endio(struct bio *bio) 1892 { 1893 struct bch_write_bio *wbio = to_wbio(bio); 1894 struct bch_write_bio *parent = wbio->split ? wbio->parent : NULL; 1895 struct bch_write_bio *orig = parent ?: wbio; 1896 struct btree_write_bio *wb = container_of(orig, struct btree_write_bio, wbio); 1897 struct bch_fs *c = wbio->c; 1898 struct btree *b = wbio->bio.bi_private; 1899 struct bch_dev *ca = bch_dev_bkey_exists(c, wbio->dev); 1900 unsigned long flags; 1901 1902 if (wbio->have_ioref) 1903 bch2_latency_acct(ca, wbio->submit_time, WRITE); 1904 1905 if (bch2_dev_io_err_on(bio->bi_status, ca, BCH_MEMBER_ERROR_write, 1906 "btree write error: %s", 1907 bch2_blk_status_to_str(bio->bi_status)) || 1908 bch2_meta_write_fault("btree")) { 1909 spin_lock_irqsave(&c->btree_write_error_lock, flags); 1910 bch2_dev_list_add_dev(&orig->failed, wbio->dev); 1911 spin_unlock_irqrestore(&c->btree_write_error_lock, flags); 1912 } 1913 1914 if (wbio->have_ioref) 1915 percpu_ref_put(&ca->io_ref); 1916 1917 if (parent) { 1918 bio_put(bio); 1919 bio_endio(&parent->bio); 1920 return; 1921 } 1922 1923 clear_btree_node_write_in_flight_inner(b); 1924 wake_up_bit(&b->flags, BTREE_NODE_write_in_flight_inner); 1925 INIT_WORK(&wb->work, btree_node_write_work); 1926 queue_work(c->btree_io_complete_wq, &wb->work); 1927 } 1928 1929 static int validate_bset_for_write(struct bch_fs *c, struct btree *b, 1930 struct bset *i, unsigned sectors) 1931 { 1932 struct printbuf buf = PRINTBUF; 1933 bool saw_error; 1934 int ret; 1935 1936 ret = bch2_bkey_invalid(c, bkey_i_to_s_c(&b->key), 1937 BKEY_TYPE_btree, WRITE, &buf); 1938 1939 if (ret) 1940 bch2_fs_inconsistent(c, "invalid btree node key before write: %s", buf.buf); 1941 printbuf_exit(&buf); 1942 if (ret) 1943 return ret; 1944 1945 ret = validate_bset_keys(c, b, i, WRITE, false, &saw_error) ?: 1946 validate_bset(c, NULL, b, i, b->written, sectors, WRITE, false, &saw_error); 1947 if (ret) { 1948 bch2_inconsistent_error(c); 1949 dump_stack(); 1950 } 1951 1952 return ret; 1953 } 1954 1955 static void btree_write_submit(struct work_struct *work) 1956 { 1957 struct btree_write_bio *wbio = container_of(work, struct btree_write_bio, work); 1958 BKEY_PADDED_ONSTACK(k, BKEY_BTREE_PTR_VAL_U64s_MAX) tmp; 1959 1960 bkey_copy(&tmp.k, &wbio->key); 1961 1962 bkey_for_each_ptr(bch2_bkey_ptrs(bkey_i_to_s(&tmp.k)), ptr) 1963 ptr->offset += wbio->sector_offset; 1964 1965 bch2_submit_wbio_replicas(&wbio->wbio, wbio->wbio.c, BCH_DATA_btree, 1966 &tmp.k, false); 1967 } 1968 1969 void __bch2_btree_node_write(struct bch_fs *c, struct btree *b, unsigned flags) 1970 { 1971 struct btree_write_bio *wbio; 1972 struct bset_tree *t; 1973 struct bset *i; 1974 struct btree_node *bn = NULL; 1975 struct btree_node_entry *bne = NULL; 1976 struct sort_iter_stack sort_iter; 1977 struct nonce nonce; 1978 unsigned bytes_to_write, sectors_to_write, bytes, u64s; 1979 u64 seq = 0; 1980 bool used_mempool; 1981 unsigned long old, new; 1982 bool validate_before_checksum = false; 1983 enum btree_write_type type = flags & BTREE_WRITE_TYPE_MASK; 1984 void *data; 1985 int ret; 1986 1987 if (flags & BTREE_WRITE_ALREADY_STARTED) 1988 goto do_write; 1989 1990 /* 1991 * We may only have a read lock on the btree node - the dirty bit is our 1992 * "lock" against racing with other threads that may be trying to start 1993 * a write, we do a write iff we clear the dirty bit. Since setting the 1994 * dirty bit requires a write lock, we can't race with other threads 1995 * redirtying it: 1996 */ 1997 do { 1998 old = new = READ_ONCE(b->flags); 1999 2000 if (!(old & (1 << BTREE_NODE_dirty))) 2001 return; 2002 2003 if ((flags & BTREE_WRITE_ONLY_IF_NEED) && 2004 !(old & (1 << BTREE_NODE_need_write))) 2005 return; 2006 2007 if (old & 2008 ((1 << BTREE_NODE_never_write)| 2009 (1 << BTREE_NODE_write_blocked))) 2010 return; 2011 2012 if (b->written && 2013 (old & (1 << BTREE_NODE_will_make_reachable))) 2014 return; 2015 2016 if (old & (1 << BTREE_NODE_write_in_flight)) 2017 return; 2018 2019 if (flags & BTREE_WRITE_ONLY_IF_NEED) 2020 type = new & BTREE_WRITE_TYPE_MASK; 2021 new &= ~BTREE_WRITE_TYPE_MASK; 2022 2023 new &= ~(1 << BTREE_NODE_dirty); 2024 new &= ~(1 << BTREE_NODE_need_write); 2025 new |= (1 << BTREE_NODE_write_in_flight); 2026 new |= (1 << BTREE_NODE_write_in_flight_inner); 2027 new |= (1 << BTREE_NODE_just_written); 2028 new ^= (1 << BTREE_NODE_write_idx); 2029 } while (cmpxchg_acquire(&b->flags, old, new) != old); 2030 2031 if (new & (1U << BTREE_NODE_need_write)) 2032 return; 2033 do_write: 2034 BUG_ON((type == BTREE_WRITE_initial) != (b->written == 0)); 2035 2036 atomic_dec(&c->btree_cache.dirty); 2037 2038 BUG_ON(btree_node_fake(b)); 2039 BUG_ON((b->will_make_reachable != 0) != !b->written); 2040 2041 BUG_ON(b->written >= btree_sectors(c)); 2042 BUG_ON(b->written & (block_sectors(c) - 1)); 2043 BUG_ON(bset_written(b, btree_bset_last(b))); 2044 BUG_ON(le64_to_cpu(b->data->magic) != bset_magic(c)); 2045 BUG_ON(memcmp(&b->data->format, &b->format, sizeof(b->format))); 2046 2047 bch2_sort_whiteouts(c, b); 2048 2049 sort_iter_stack_init(&sort_iter, b); 2050 2051 bytes = !b->written 2052 ? sizeof(struct btree_node) 2053 : sizeof(struct btree_node_entry); 2054 2055 bytes += b->whiteout_u64s * sizeof(u64); 2056 2057 for_each_bset(b, t) { 2058 i = bset(b, t); 2059 2060 if (bset_written(b, i)) 2061 continue; 2062 2063 bytes += le16_to_cpu(i->u64s) * sizeof(u64); 2064 sort_iter_add(&sort_iter.iter, 2065 btree_bkey_first(b, t), 2066 btree_bkey_last(b, t)); 2067 seq = max(seq, le64_to_cpu(i->journal_seq)); 2068 } 2069 2070 BUG_ON(b->written && !seq); 2071 2072 /* bch2_varint_decode may read up to 7 bytes past the end of the buffer: */ 2073 bytes += 8; 2074 2075 /* buffer must be a multiple of the block size */ 2076 bytes = round_up(bytes, block_bytes(c)); 2077 2078 data = btree_bounce_alloc(c, bytes, &used_mempool); 2079 2080 if (!b->written) { 2081 bn = data; 2082 *bn = *b->data; 2083 i = &bn->keys; 2084 } else { 2085 bne = data; 2086 bne->keys = b->data->keys; 2087 i = &bne->keys; 2088 } 2089 2090 i->journal_seq = cpu_to_le64(seq); 2091 i->u64s = 0; 2092 2093 sort_iter_add(&sort_iter.iter, 2094 unwritten_whiteouts_start(b), 2095 unwritten_whiteouts_end(b)); 2096 SET_BSET_SEPARATE_WHITEOUTS(i, false); 2097 2098 b->whiteout_u64s = 0; 2099 2100 u64s = bch2_sort_keys(i->start, &sort_iter.iter, false); 2101 le16_add_cpu(&i->u64s, u64s); 2102 2103 BUG_ON(!b->written && i->u64s != b->data->keys.u64s); 2104 2105 set_needs_whiteout(i, false); 2106 2107 /* do we have data to write? */ 2108 if (b->written && !i->u64s) 2109 goto nowrite; 2110 2111 bytes_to_write = vstruct_end(i) - data; 2112 sectors_to_write = round_up(bytes_to_write, block_bytes(c)) >> 9; 2113 2114 if (!b->written && 2115 b->key.k.type == KEY_TYPE_btree_ptr_v2) 2116 BUG_ON(btree_ptr_sectors_written(&b->key) != sectors_to_write); 2117 2118 memset(data + bytes_to_write, 0, 2119 (sectors_to_write << 9) - bytes_to_write); 2120 2121 BUG_ON(b->written + sectors_to_write > btree_sectors(c)); 2122 BUG_ON(BSET_BIG_ENDIAN(i) != CPU_BIG_ENDIAN); 2123 BUG_ON(i->seq != b->data->keys.seq); 2124 2125 i->version = cpu_to_le16(c->sb.version); 2126 SET_BSET_OFFSET(i, b->written); 2127 SET_BSET_CSUM_TYPE(i, bch2_meta_checksum_type(c)); 2128 2129 if (bch2_csum_type_is_encryption(BSET_CSUM_TYPE(i))) 2130 validate_before_checksum = true; 2131 2132 /* validate_bset will be modifying: */ 2133 if (le16_to_cpu(i->version) < bcachefs_metadata_version_current) 2134 validate_before_checksum = true; 2135 2136 /* if we're going to be encrypting, check metadata validity first: */ 2137 if (validate_before_checksum && 2138 validate_bset_for_write(c, b, i, sectors_to_write)) 2139 goto err; 2140 2141 ret = bset_encrypt(c, i, b->written << 9); 2142 if (bch2_fs_fatal_err_on(ret, c, 2143 "encrypting btree node: %s", bch2_err_str(ret))) 2144 goto err; 2145 2146 nonce = btree_nonce(i, b->written << 9); 2147 2148 if (bn) 2149 bn->csum = csum_vstruct(c, BSET_CSUM_TYPE(i), nonce, bn); 2150 else 2151 bne->csum = csum_vstruct(c, BSET_CSUM_TYPE(i), nonce, bne); 2152 2153 /* if we're not encrypting, check metadata after checksumming: */ 2154 if (!validate_before_checksum && 2155 validate_bset_for_write(c, b, i, sectors_to_write)) 2156 goto err; 2157 2158 /* 2159 * We handle btree write errors by immediately halting the journal - 2160 * after we've done that, we can't issue any subsequent btree writes 2161 * because they might have pointers to new nodes that failed to write. 2162 * 2163 * Furthermore, there's no point in doing any more btree writes because 2164 * with the journal stopped, we're never going to update the journal to 2165 * reflect that those writes were done and the data flushed from the 2166 * journal: 2167 * 2168 * Also on journal error, the pending write may have updates that were 2169 * never journalled (interior nodes, see btree_update_nodes_written()) - 2170 * it's critical that we don't do the write in that case otherwise we 2171 * will have updates visible that weren't in the journal: 2172 * 2173 * Make sure to update b->written so bch2_btree_init_next() doesn't 2174 * break: 2175 */ 2176 if (bch2_journal_error(&c->journal) || 2177 c->opts.nochanges) 2178 goto err; 2179 2180 trace_and_count(c, btree_node_write, b, bytes_to_write, sectors_to_write); 2181 2182 wbio = container_of(bio_alloc_bioset(NULL, 2183 buf_pages(data, sectors_to_write << 9), 2184 REQ_OP_WRITE|REQ_META, 2185 GFP_NOFS, 2186 &c->btree_bio), 2187 struct btree_write_bio, wbio.bio); 2188 wbio_init(&wbio->wbio.bio); 2189 wbio->data = data; 2190 wbio->data_bytes = bytes; 2191 wbio->sector_offset = b->written; 2192 wbio->wbio.c = c; 2193 wbio->wbio.used_mempool = used_mempool; 2194 wbio->wbio.first_btree_write = !b->written; 2195 wbio->wbio.bio.bi_end_io = btree_node_write_endio; 2196 wbio->wbio.bio.bi_private = b; 2197 2198 bch2_bio_map(&wbio->wbio.bio, data, sectors_to_write << 9); 2199 2200 bkey_copy(&wbio->key, &b->key); 2201 2202 b->written += sectors_to_write; 2203 2204 if (wbio->key.k.type == KEY_TYPE_btree_ptr_v2) 2205 bkey_i_to_btree_ptr_v2(&wbio->key)->v.sectors_written = 2206 cpu_to_le16(b->written); 2207 2208 atomic64_inc(&c->btree_write_stats[type].nr); 2209 atomic64_add(bytes_to_write, &c->btree_write_stats[type].bytes); 2210 2211 INIT_WORK(&wbio->work, btree_write_submit); 2212 queue_work(c->io_complete_wq, &wbio->work); 2213 return; 2214 err: 2215 set_btree_node_noevict(b); 2216 b->written += sectors_to_write; 2217 nowrite: 2218 btree_bounce_free(c, bytes, used_mempool, data); 2219 __btree_node_write_done(c, b); 2220 } 2221 2222 /* 2223 * Work that must be done with write lock held: 2224 */ 2225 bool bch2_btree_post_write_cleanup(struct bch_fs *c, struct btree *b) 2226 { 2227 bool invalidated_iter = false; 2228 struct btree_node_entry *bne; 2229 struct bset_tree *t; 2230 2231 if (!btree_node_just_written(b)) 2232 return false; 2233 2234 BUG_ON(b->whiteout_u64s); 2235 2236 clear_btree_node_just_written(b); 2237 2238 /* 2239 * Note: immediately after write, bset_written() doesn't work - the 2240 * amount of data we had to write after compaction might have been 2241 * smaller than the offset of the last bset. 2242 * 2243 * However, we know that all bsets have been written here, as long as 2244 * we're still holding the write lock: 2245 */ 2246 2247 /* 2248 * XXX: decide if we really want to unconditionally sort down to a 2249 * single bset: 2250 */ 2251 if (b->nsets > 1) { 2252 btree_node_sort(c, b, 0, b->nsets, true); 2253 invalidated_iter = true; 2254 } else { 2255 invalidated_iter = bch2_drop_whiteouts(b, COMPACT_ALL); 2256 } 2257 2258 for_each_bset(b, t) 2259 set_needs_whiteout(bset(b, t), true); 2260 2261 bch2_btree_verify(c, b); 2262 2263 /* 2264 * If later we don't unconditionally sort down to a single bset, we have 2265 * to ensure this is still true: 2266 */ 2267 BUG_ON((void *) btree_bkey_last(b, bset_tree_last(b)) > write_block(b)); 2268 2269 bne = want_new_bset(c, b); 2270 if (bne) 2271 bch2_bset_init_next(b, bne); 2272 2273 bch2_btree_build_aux_trees(b); 2274 2275 return invalidated_iter; 2276 } 2277 2278 /* 2279 * Use this one if the node is intent locked: 2280 */ 2281 void bch2_btree_node_write(struct bch_fs *c, struct btree *b, 2282 enum six_lock_type lock_type_held, 2283 unsigned flags) 2284 { 2285 if (lock_type_held == SIX_LOCK_intent || 2286 (lock_type_held == SIX_LOCK_read && 2287 six_lock_tryupgrade(&b->c.lock))) { 2288 __bch2_btree_node_write(c, b, flags); 2289 2290 /* don't cycle lock unnecessarily: */ 2291 if (btree_node_just_written(b) && 2292 six_trylock_write(&b->c.lock)) { 2293 bch2_btree_post_write_cleanup(c, b); 2294 six_unlock_write(&b->c.lock); 2295 } 2296 2297 if (lock_type_held == SIX_LOCK_read) 2298 six_lock_downgrade(&b->c.lock); 2299 } else { 2300 __bch2_btree_node_write(c, b, flags); 2301 if (lock_type_held == SIX_LOCK_write && 2302 btree_node_just_written(b)) 2303 bch2_btree_post_write_cleanup(c, b); 2304 } 2305 } 2306 2307 static bool __bch2_btree_flush_all(struct bch_fs *c, unsigned flag) 2308 { 2309 struct bucket_table *tbl; 2310 struct rhash_head *pos; 2311 struct btree *b; 2312 unsigned i; 2313 bool ret = false; 2314 restart: 2315 rcu_read_lock(); 2316 for_each_cached_btree(b, c, tbl, i, pos) 2317 if (test_bit(flag, &b->flags)) { 2318 rcu_read_unlock(); 2319 wait_on_bit_io(&b->flags, flag, TASK_UNINTERRUPTIBLE); 2320 ret = true; 2321 goto restart; 2322 } 2323 rcu_read_unlock(); 2324 2325 return ret; 2326 } 2327 2328 bool bch2_btree_flush_all_reads(struct bch_fs *c) 2329 { 2330 return __bch2_btree_flush_all(c, BTREE_NODE_read_in_flight); 2331 } 2332 2333 bool bch2_btree_flush_all_writes(struct bch_fs *c) 2334 { 2335 return __bch2_btree_flush_all(c, BTREE_NODE_write_in_flight); 2336 } 2337 2338 static const char * const bch2_btree_write_types[] = { 2339 #define x(t, n) [n] = #t, 2340 BCH_BTREE_WRITE_TYPES() 2341 NULL 2342 }; 2343 2344 void bch2_btree_write_stats_to_text(struct printbuf *out, struct bch_fs *c) 2345 { 2346 printbuf_tabstop_push(out, 20); 2347 printbuf_tabstop_push(out, 10); 2348 2349 prt_tab(out); 2350 prt_str(out, "nr"); 2351 prt_tab(out); 2352 prt_str(out, "size"); 2353 prt_newline(out); 2354 2355 for (unsigned i = 0; i < BTREE_WRITE_TYPE_NR; i++) { 2356 u64 nr = atomic64_read(&c->btree_write_stats[i].nr); 2357 u64 bytes = atomic64_read(&c->btree_write_stats[i].bytes); 2358 2359 prt_printf(out, "%s:", bch2_btree_write_types[i]); 2360 prt_tab(out); 2361 prt_u64(out, nr); 2362 prt_tab(out); 2363 prt_human_readable_u64(out, nr ? div64_u64(bytes, nr) : 0); 2364 prt_newline(out); 2365 } 2366 } 2367