1 // SPDX-License-Identifier: GPL-2.0 2 /* 3 * Copyright (C) 2010 Kent Overstreet <kent.overstreet@gmail.com> 4 * 5 * Code for managing the extent btree and dynamically updating the writeback 6 * dirty sector count. 7 */ 8 9 #include "bcachefs.h" 10 #include "bkey_methods.h" 11 #include "btree_gc.h" 12 #include "btree_io.h" 13 #include "btree_iter.h" 14 #include "buckets.h" 15 #include "checksum.h" 16 #include "debug.h" 17 #include "disk_groups.h" 18 #include "error.h" 19 #include "extents.h" 20 #include "inode.h" 21 #include "journal.h" 22 #include "replicas.h" 23 #include "super.h" 24 #include "super-io.h" 25 #include "trace.h" 26 #include "util.h" 27 28 static unsigned bch2_crc_field_size_max[] = { 29 [BCH_EXTENT_ENTRY_crc32] = CRC32_SIZE_MAX, 30 [BCH_EXTENT_ENTRY_crc64] = CRC64_SIZE_MAX, 31 [BCH_EXTENT_ENTRY_crc128] = CRC128_SIZE_MAX, 32 }; 33 34 static void bch2_extent_crc_pack(union bch_extent_crc *, 35 struct bch_extent_crc_unpacked, 36 enum bch_extent_entry_type); 37 38 static struct bch_dev_io_failures *dev_io_failures(struct bch_io_failures *f, 39 unsigned dev) 40 { 41 struct bch_dev_io_failures *i; 42 43 for (i = f->devs; i < f->devs + f->nr; i++) 44 if (i->dev == dev) 45 return i; 46 47 return NULL; 48 } 49 50 void bch2_mark_io_failure(struct bch_io_failures *failed, 51 struct extent_ptr_decoded *p) 52 { 53 struct bch_dev_io_failures *f = dev_io_failures(failed, p->ptr.dev); 54 55 if (!f) { 56 BUG_ON(failed->nr >= ARRAY_SIZE(failed->devs)); 57 58 f = &failed->devs[failed->nr++]; 59 f->dev = p->ptr.dev; 60 f->idx = p->idx; 61 f->nr_failed = 1; 62 f->nr_retries = 0; 63 } else if (p->idx != f->idx) { 64 f->idx = p->idx; 65 f->nr_failed = 1; 66 f->nr_retries = 0; 67 } else { 68 f->nr_failed++; 69 } 70 } 71 72 /* 73 * returns true if p1 is better than p2: 74 */ 75 static inline bool ptr_better(struct bch_fs *c, 76 const struct extent_ptr_decoded p1, 77 const struct extent_ptr_decoded p2) 78 { 79 if (likely(!p1.idx && !p2.idx)) { 80 struct bch_dev *dev1 = bch_dev_bkey_exists(c, p1.ptr.dev); 81 struct bch_dev *dev2 = bch_dev_bkey_exists(c, p2.ptr.dev); 82 83 u64 l1 = atomic64_read(&dev1->cur_latency[READ]); 84 u64 l2 = atomic64_read(&dev2->cur_latency[READ]); 85 86 /* Pick at random, biased in favor of the faster device: */ 87 88 return bch2_rand_range(l1 + l2) > l1; 89 } 90 91 if (bch2_force_reconstruct_read) 92 return p1.idx > p2.idx; 93 94 return p1.idx < p2.idx; 95 } 96 97 /* 98 * This picks a non-stale pointer, preferably from a device other than @avoid. 99 * Avoid can be NULL, meaning pick any. If there are no non-stale pointers to 100 * other devices, it will still pick a pointer from avoid. 101 */ 102 int bch2_bkey_pick_read_device(struct bch_fs *c, struct bkey_s_c k, 103 struct bch_io_failures *failed, 104 struct extent_ptr_decoded *pick) 105 { 106 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 107 const union bch_extent_entry *entry; 108 struct extent_ptr_decoded p; 109 struct bch_dev_io_failures *f; 110 struct bch_dev *ca; 111 int ret = 0; 112 113 if (k.k->type == KEY_TYPE_error) 114 return -EIO; 115 116 bkey_for_each_ptr_decode(k.k, ptrs, p, entry) { 117 /* 118 * Unwritten extent: no need to actually read, treat it as a 119 * hole and return 0s: 120 */ 121 if (p.ptr.unwritten) 122 return 0; 123 124 ca = bch_dev_bkey_exists(c, p.ptr.dev); 125 126 /* 127 * If there are any dirty pointers it's an error if we can't 128 * read: 129 */ 130 if (!ret && !p.ptr.cached) 131 ret = -EIO; 132 133 if (p.ptr.cached && ptr_stale(ca, &p.ptr)) 134 continue; 135 136 f = failed ? dev_io_failures(failed, p.ptr.dev) : NULL; 137 if (f) 138 p.idx = f->nr_failed < f->nr_retries 139 ? f->idx 140 : f->idx + 1; 141 142 if (!p.idx && 143 !bch2_dev_is_readable(ca)) 144 p.idx++; 145 146 if (bch2_force_reconstruct_read && 147 !p.idx && p.has_ec) 148 p.idx++; 149 150 if (p.idx >= (unsigned) p.has_ec + 1) 151 continue; 152 153 if (ret > 0 && !ptr_better(c, p, *pick)) 154 continue; 155 156 *pick = p; 157 ret = 1; 158 } 159 160 return ret; 161 } 162 163 /* KEY_TYPE_btree_ptr: */ 164 165 int bch2_btree_ptr_invalid(const struct bch_fs *c, struct bkey_s_c k, 166 enum bkey_invalid_flags flags, 167 struct printbuf *err) 168 { 169 if (bkey_val_u64s(k.k) > BCH_REPLICAS_MAX) { 170 prt_printf(err, "value too big (%zu > %u)", 171 bkey_val_u64s(k.k), BCH_REPLICAS_MAX); 172 return -BCH_ERR_invalid_bkey; 173 } 174 175 return bch2_bkey_ptrs_invalid(c, k, flags, err); 176 } 177 178 void bch2_btree_ptr_to_text(struct printbuf *out, struct bch_fs *c, 179 struct bkey_s_c k) 180 { 181 bch2_bkey_ptrs_to_text(out, c, k); 182 } 183 184 int bch2_btree_ptr_v2_invalid(const struct bch_fs *c, struct bkey_s_c k, 185 enum bkey_invalid_flags flags, 186 struct printbuf *err) 187 { 188 if (bkey_val_u64s(k.k) > BKEY_BTREE_PTR_VAL_U64s_MAX) { 189 prt_printf(err, "value too big (%zu > %zu)", 190 bkey_val_u64s(k.k), BKEY_BTREE_PTR_VAL_U64s_MAX); 191 return -BCH_ERR_invalid_bkey; 192 } 193 194 return bch2_bkey_ptrs_invalid(c, k, flags, err); 195 } 196 197 void bch2_btree_ptr_v2_to_text(struct printbuf *out, struct bch_fs *c, 198 struct bkey_s_c k) 199 { 200 struct bkey_s_c_btree_ptr_v2 bp = bkey_s_c_to_btree_ptr_v2(k); 201 202 prt_printf(out, "seq %llx written %u min_key %s", 203 le64_to_cpu(bp.v->seq), 204 le16_to_cpu(bp.v->sectors_written), 205 BTREE_PTR_RANGE_UPDATED(bp.v) ? "R " : ""); 206 207 bch2_bpos_to_text(out, bp.v->min_key); 208 prt_printf(out, " "); 209 bch2_bkey_ptrs_to_text(out, c, k); 210 } 211 212 void bch2_btree_ptr_v2_compat(enum btree_id btree_id, unsigned version, 213 unsigned big_endian, int write, 214 struct bkey_s k) 215 { 216 struct bkey_s_btree_ptr_v2 bp = bkey_s_to_btree_ptr_v2(k); 217 218 compat_bpos(0, btree_id, version, big_endian, write, &bp.v->min_key); 219 220 if (version < bcachefs_metadata_version_inode_btree_change && 221 btree_id_is_extents(btree_id) && 222 !bkey_eq(bp.v->min_key, POS_MIN)) 223 bp.v->min_key = write 224 ? bpos_nosnap_predecessor(bp.v->min_key) 225 : bpos_nosnap_successor(bp.v->min_key); 226 } 227 228 /* KEY_TYPE_extent: */ 229 230 bool bch2_extent_merge(struct bch_fs *c, struct bkey_s l, struct bkey_s_c r) 231 { 232 struct bkey_ptrs l_ptrs = bch2_bkey_ptrs(l); 233 struct bkey_ptrs_c r_ptrs = bch2_bkey_ptrs_c(r); 234 union bch_extent_entry *en_l; 235 const union bch_extent_entry *en_r; 236 struct extent_ptr_decoded lp, rp; 237 bool use_right_ptr; 238 struct bch_dev *ca; 239 240 en_l = l_ptrs.start; 241 en_r = r_ptrs.start; 242 while (en_l < l_ptrs.end && en_r < r_ptrs.end) { 243 if (extent_entry_type(en_l) != extent_entry_type(en_r)) 244 return false; 245 246 en_l = extent_entry_next(en_l); 247 en_r = extent_entry_next(en_r); 248 } 249 250 if (en_l < l_ptrs.end || en_r < r_ptrs.end) 251 return false; 252 253 en_l = l_ptrs.start; 254 en_r = r_ptrs.start; 255 lp.crc = bch2_extent_crc_unpack(l.k, NULL); 256 rp.crc = bch2_extent_crc_unpack(r.k, NULL); 257 258 while (__bkey_ptr_next_decode(l.k, l_ptrs.end, lp, en_l) && 259 __bkey_ptr_next_decode(r.k, r_ptrs.end, rp, en_r)) { 260 if (lp.ptr.offset + lp.crc.offset + lp.crc.live_size != 261 rp.ptr.offset + rp.crc.offset || 262 lp.ptr.dev != rp.ptr.dev || 263 lp.ptr.gen != rp.ptr.gen || 264 lp.ptr.unwritten != rp.ptr.unwritten || 265 lp.has_ec != rp.has_ec) 266 return false; 267 268 /* Extents may not straddle buckets: */ 269 ca = bch_dev_bkey_exists(c, lp.ptr.dev); 270 if (PTR_BUCKET_NR(ca, &lp.ptr) != PTR_BUCKET_NR(ca, &rp.ptr)) 271 return false; 272 273 if (lp.has_ec != rp.has_ec || 274 (lp.has_ec && 275 (lp.ec.block != rp.ec.block || 276 lp.ec.redundancy != rp.ec.redundancy || 277 lp.ec.idx != rp.ec.idx))) 278 return false; 279 280 if (lp.crc.compression_type != rp.crc.compression_type || 281 lp.crc.nonce != rp.crc.nonce) 282 return false; 283 284 if (lp.crc.offset + lp.crc.live_size + rp.crc.live_size <= 285 lp.crc.uncompressed_size) { 286 /* can use left extent's crc entry */ 287 } else if (lp.crc.live_size <= rp.crc.offset) { 288 /* can use right extent's crc entry */ 289 } else { 290 /* check if checksums can be merged: */ 291 if (lp.crc.csum_type != rp.crc.csum_type || 292 lp.crc.nonce != rp.crc.nonce || 293 crc_is_compressed(lp.crc) || 294 !bch2_checksum_mergeable(lp.crc.csum_type)) 295 return false; 296 297 if (lp.crc.offset + lp.crc.live_size != lp.crc.compressed_size || 298 rp.crc.offset) 299 return false; 300 301 if (lp.crc.csum_type && 302 lp.crc.uncompressed_size + 303 rp.crc.uncompressed_size > (c->opts.encoded_extent_max >> 9)) 304 return false; 305 } 306 307 en_l = extent_entry_next(en_l); 308 en_r = extent_entry_next(en_r); 309 } 310 311 en_l = l_ptrs.start; 312 en_r = r_ptrs.start; 313 while (en_l < l_ptrs.end && en_r < r_ptrs.end) { 314 if (extent_entry_is_crc(en_l)) { 315 struct bch_extent_crc_unpacked crc_l = bch2_extent_crc_unpack(l.k, entry_to_crc(en_l)); 316 struct bch_extent_crc_unpacked crc_r = bch2_extent_crc_unpack(r.k, entry_to_crc(en_r)); 317 318 if (crc_l.uncompressed_size + crc_r.uncompressed_size > 319 bch2_crc_field_size_max[extent_entry_type(en_l)]) 320 return false; 321 } 322 323 en_l = extent_entry_next(en_l); 324 en_r = extent_entry_next(en_r); 325 } 326 327 use_right_ptr = false; 328 en_l = l_ptrs.start; 329 en_r = r_ptrs.start; 330 while (en_l < l_ptrs.end) { 331 if (extent_entry_type(en_l) == BCH_EXTENT_ENTRY_ptr && 332 use_right_ptr) 333 en_l->ptr = en_r->ptr; 334 335 if (extent_entry_is_crc(en_l)) { 336 struct bch_extent_crc_unpacked crc_l = 337 bch2_extent_crc_unpack(l.k, entry_to_crc(en_l)); 338 struct bch_extent_crc_unpacked crc_r = 339 bch2_extent_crc_unpack(r.k, entry_to_crc(en_r)); 340 341 use_right_ptr = false; 342 343 if (crc_l.offset + crc_l.live_size + crc_r.live_size <= 344 crc_l.uncompressed_size) { 345 /* can use left extent's crc entry */ 346 } else if (crc_l.live_size <= crc_r.offset) { 347 /* can use right extent's crc entry */ 348 crc_r.offset -= crc_l.live_size; 349 bch2_extent_crc_pack(entry_to_crc(en_l), crc_r, 350 extent_entry_type(en_l)); 351 use_right_ptr = true; 352 } else { 353 crc_l.csum = bch2_checksum_merge(crc_l.csum_type, 354 crc_l.csum, 355 crc_r.csum, 356 crc_r.uncompressed_size << 9); 357 358 crc_l.uncompressed_size += crc_r.uncompressed_size; 359 crc_l.compressed_size += crc_r.compressed_size; 360 bch2_extent_crc_pack(entry_to_crc(en_l), crc_l, 361 extent_entry_type(en_l)); 362 } 363 } 364 365 en_l = extent_entry_next(en_l); 366 en_r = extent_entry_next(en_r); 367 } 368 369 bch2_key_resize(l.k, l.k->size + r.k->size); 370 return true; 371 } 372 373 /* KEY_TYPE_reservation: */ 374 375 int bch2_reservation_invalid(const struct bch_fs *c, struct bkey_s_c k, 376 enum bkey_invalid_flags flags, 377 struct printbuf *err) 378 { 379 struct bkey_s_c_reservation r = bkey_s_c_to_reservation(k); 380 381 if (!r.v->nr_replicas || r.v->nr_replicas > BCH_REPLICAS_MAX) { 382 prt_printf(err, "invalid nr_replicas (%u)", 383 r.v->nr_replicas); 384 return -BCH_ERR_invalid_bkey; 385 } 386 387 return 0; 388 } 389 390 void bch2_reservation_to_text(struct printbuf *out, struct bch_fs *c, 391 struct bkey_s_c k) 392 { 393 struct bkey_s_c_reservation r = bkey_s_c_to_reservation(k); 394 395 prt_printf(out, "generation %u replicas %u", 396 le32_to_cpu(r.v->generation), 397 r.v->nr_replicas); 398 } 399 400 bool bch2_reservation_merge(struct bch_fs *c, struct bkey_s _l, struct bkey_s_c _r) 401 { 402 struct bkey_s_reservation l = bkey_s_to_reservation(_l); 403 struct bkey_s_c_reservation r = bkey_s_c_to_reservation(_r); 404 405 if (l.v->generation != r.v->generation || 406 l.v->nr_replicas != r.v->nr_replicas) 407 return false; 408 409 bch2_key_resize(l.k, l.k->size + r.k->size); 410 return true; 411 } 412 413 /* Extent checksum entries: */ 414 415 /* returns true if not equal */ 416 static inline bool bch2_crc_unpacked_cmp(struct bch_extent_crc_unpacked l, 417 struct bch_extent_crc_unpacked r) 418 { 419 return (l.csum_type != r.csum_type || 420 l.compression_type != r.compression_type || 421 l.compressed_size != r.compressed_size || 422 l.uncompressed_size != r.uncompressed_size || 423 l.offset != r.offset || 424 l.live_size != r.live_size || 425 l.nonce != r.nonce || 426 bch2_crc_cmp(l.csum, r.csum)); 427 } 428 429 static inline bool can_narrow_crc(struct bch_extent_crc_unpacked u, 430 struct bch_extent_crc_unpacked n) 431 { 432 return !crc_is_compressed(u) && 433 u.csum_type && 434 u.uncompressed_size > u.live_size && 435 bch2_csum_type_is_encryption(u.csum_type) == 436 bch2_csum_type_is_encryption(n.csum_type); 437 } 438 439 bool bch2_can_narrow_extent_crcs(struct bkey_s_c k, 440 struct bch_extent_crc_unpacked n) 441 { 442 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 443 struct bch_extent_crc_unpacked crc; 444 const union bch_extent_entry *i; 445 446 if (!n.csum_type) 447 return false; 448 449 bkey_for_each_crc(k.k, ptrs, crc, i) 450 if (can_narrow_crc(crc, n)) 451 return true; 452 453 return false; 454 } 455 456 /* 457 * We're writing another replica for this extent, so while we've got the data in 458 * memory we'll be computing a new checksum for the currently live data. 459 * 460 * If there are other replicas we aren't moving, and they are checksummed but 461 * not compressed, we can modify them to point to only the data that is 462 * currently live (so that readers won't have to bounce) while we've got the 463 * checksum we need: 464 */ 465 bool bch2_bkey_narrow_crcs(struct bkey_i *k, struct bch_extent_crc_unpacked n) 466 { 467 struct bkey_ptrs ptrs = bch2_bkey_ptrs(bkey_i_to_s(k)); 468 struct bch_extent_crc_unpacked u; 469 struct extent_ptr_decoded p; 470 union bch_extent_entry *i; 471 bool ret = false; 472 473 /* Find a checksum entry that covers only live data: */ 474 if (!n.csum_type) { 475 bkey_for_each_crc(&k->k, ptrs, u, i) 476 if (!crc_is_compressed(u) && 477 u.csum_type && 478 u.live_size == u.uncompressed_size) { 479 n = u; 480 goto found; 481 } 482 return false; 483 } 484 found: 485 BUG_ON(crc_is_compressed(n)); 486 BUG_ON(n.offset); 487 BUG_ON(n.live_size != k->k.size); 488 489 restart_narrow_pointers: 490 ptrs = bch2_bkey_ptrs(bkey_i_to_s(k)); 491 492 bkey_for_each_ptr_decode(&k->k, ptrs, p, i) 493 if (can_narrow_crc(p.crc, n)) { 494 bch2_bkey_drop_ptr_noerror(bkey_i_to_s(k), &i->ptr); 495 p.ptr.offset += p.crc.offset; 496 p.crc = n; 497 bch2_extent_ptr_decoded_append(k, &p); 498 ret = true; 499 goto restart_narrow_pointers; 500 } 501 502 return ret; 503 } 504 505 static void bch2_extent_crc_pack(union bch_extent_crc *dst, 506 struct bch_extent_crc_unpacked src, 507 enum bch_extent_entry_type type) 508 { 509 #define set_common_fields(_dst, _src) \ 510 _dst.type = 1 << type; \ 511 _dst.csum_type = _src.csum_type, \ 512 _dst.compression_type = _src.compression_type, \ 513 _dst._compressed_size = _src.compressed_size - 1, \ 514 _dst._uncompressed_size = _src.uncompressed_size - 1, \ 515 _dst.offset = _src.offset 516 517 switch (type) { 518 case BCH_EXTENT_ENTRY_crc32: 519 set_common_fields(dst->crc32, src); 520 dst->crc32.csum = (u32 __force) *((__le32 *) &src.csum.lo); 521 break; 522 case BCH_EXTENT_ENTRY_crc64: 523 set_common_fields(dst->crc64, src); 524 dst->crc64.nonce = src.nonce; 525 dst->crc64.csum_lo = (u64 __force) src.csum.lo; 526 dst->crc64.csum_hi = (u64 __force) *((__le16 *) &src.csum.hi); 527 break; 528 case BCH_EXTENT_ENTRY_crc128: 529 set_common_fields(dst->crc128, src); 530 dst->crc128.nonce = src.nonce; 531 dst->crc128.csum = src.csum; 532 break; 533 default: 534 BUG(); 535 } 536 #undef set_common_fields 537 } 538 539 void bch2_extent_crc_append(struct bkey_i *k, 540 struct bch_extent_crc_unpacked new) 541 { 542 struct bkey_ptrs ptrs = bch2_bkey_ptrs(bkey_i_to_s(k)); 543 union bch_extent_crc *crc = (void *) ptrs.end; 544 enum bch_extent_entry_type type; 545 546 if (bch_crc_bytes[new.csum_type] <= 4 && 547 new.uncompressed_size <= CRC32_SIZE_MAX && 548 new.nonce <= CRC32_NONCE_MAX) 549 type = BCH_EXTENT_ENTRY_crc32; 550 else if (bch_crc_bytes[new.csum_type] <= 10 && 551 new.uncompressed_size <= CRC64_SIZE_MAX && 552 new.nonce <= CRC64_NONCE_MAX) 553 type = BCH_EXTENT_ENTRY_crc64; 554 else if (bch_crc_bytes[new.csum_type] <= 16 && 555 new.uncompressed_size <= CRC128_SIZE_MAX && 556 new.nonce <= CRC128_NONCE_MAX) 557 type = BCH_EXTENT_ENTRY_crc128; 558 else 559 BUG(); 560 561 bch2_extent_crc_pack(crc, new, type); 562 563 k->k.u64s += extent_entry_u64s(ptrs.end); 564 565 EBUG_ON(bkey_val_u64s(&k->k) > BKEY_EXTENT_VAL_U64s_MAX); 566 } 567 568 /* Generic code for keys with pointers: */ 569 570 unsigned bch2_bkey_nr_ptrs(struct bkey_s_c k) 571 { 572 return bch2_bkey_devs(k).nr; 573 } 574 575 unsigned bch2_bkey_nr_ptrs_allocated(struct bkey_s_c k) 576 { 577 return k.k->type == KEY_TYPE_reservation 578 ? bkey_s_c_to_reservation(k).v->nr_replicas 579 : bch2_bkey_dirty_devs(k).nr; 580 } 581 582 unsigned bch2_bkey_nr_ptrs_fully_allocated(struct bkey_s_c k) 583 { 584 unsigned ret = 0; 585 586 if (k.k->type == KEY_TYPE_reservation) { 587 ret = bkey_s_c_to_reservation(k).v->nr_replicas; 588 } else { 589 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 590 const union bch_extent_entry *entry; 591 struct extent_ptr_decoded p; 592 593 bkey_for_each_ptr_decode(k.k, ptrs, p, entry) 594 ret += !p.ptr.cached && !crc_is_compressed(p.crc); 595 } 596 597 return ret; 598 } 599 600 unsigned bch2_bkey_sectors_compressed(struct bkey_s_c k) 601 { 602 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 603 const union bch_extent_entry *entry; 604 struct extent_ptr_decoded p; 605 unsigned ret = 0; 606 607 bkey_for_each_ptr_decode(k.k, ptrs, p, entry) 608 if (!p.ptr.cached && crc_is_compressed(p.crc)) 609 ret += p.crc.compressed_size; 610 611 return ret; 612 } 613 614 bool bch2_bkey_is_incompressible(struct bkey_s_c k) 615 { 616 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 617 const union bch_extent_entry *entry; 618 struct bch_extent_crc_unpacked crc; 619 620 bkey_for_each_crc(k.k, ptrs, crc, entry) 621 if (crc.compression_type == BCH_COMPRESSION_TYPE_incompressible) 622 return true; 623 return false; 624 } 625 626 unsigned bch2_bkey_replicas(struct bch_fs *c, struct bkey_s_c k) 627 { 628 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 629 const union bch_extent_entry *entry; 630 struct extent_ptr_decoded p = { 0 }; 631 unsigned replicas = 0; 632 633 bkey_for_each_ptr_decode(k.k, ptrs, p, entry) { 634 if (p.ptr.cached) 635 continue; 636 637 if (p.has_ec) 638 replicas += p.ec.redundancy; 639 640 replicas++; 641 642 } 643 644 return replicas; 645 } 646 647 unsigned bch2_extent_ptr_desired_durability(struct bch_fs *c, struct extent_ptr_decoded *p) 648 { 649 struct bch_dev *ca; 650 651 if (p->ptr.cached) 652 return 0; 653 654 ca = bch_dev_bkey_exists(c, p->ptr.dev); 655 656 return ca->mi.durability + 657 (p->has_ec 658 ? p->ec.redundancy 659 : 0); 660 } 661 662 unsigned bch2_extent_ptr_durability(struct bch_fs *c, struct extent_ptr_decoded *p) 663 { 664 struct bch_dev *ca; 665 666 if (p->ptr.cached) 667 return 0; 668 669 ca = bch_dev_bkey_exists(c, p->ptr.dev); 670 671 if (ca->mi.state == BCH_MEMBER_STATE_failed) 672 return 0; 673 674 return ca->mi.durability + 675 (p->has_ec 676 ? p->ec.redundancy 677 : 0); 678 } 679 680 unsigned bch2_bkey_durability(struct bch_fs *c, struct bkey_s_c k) 681 { 682 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 683 const union bch_extent_entry *entry; 684 struct extent_ptr_decoded p; 685 unsigned durability = 0; 686 687 bkey_for_each_ptr_decode(k.k, ptrs, p, entry) 688 durability += bch2_extent_ptr_durability(c, &p); 689 690 return durability; 691 } 692 693 static unsigned bch2_bkey_durability_safe(struct bch_fs *c, struct bkey_s_c k) 694 { 695 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 696 const union bch_extent_entry *entry; 697 struct extent_ptr_decoded p; 698 unsigned durability = 0; 699 700 bkey_for_each_ptr_decode(k.k, ptrs, p, entry) 701 if (p.ptr.dev < c->sb.nr_devices && c->devs[p.ptr.dev]) 702 durability += bch2_extent_ptr_durability(c, &p); 703 704 return durability; 705 } 706 707 void bch2_bkey_extent_entry_drop(struct bkey_i *k, union bch_extent_entry *entry) 708 { 709 union bch_extent_entry *end = bkey_val_end(bkey_i_to_s(k)); 710 union bch_extent_entry *next = extent_entry_next(entry); 711 712 memmove_u64s(entry, next, (u64 *) end - (u64 *) next); 713 k->k.u64s -= extent_entry_u64s(entry); 714 } 715 716 void bch2_extent_ptr_decoded_append(struct bkey_i *k, 717 struct extent_ptr_decoded *p) 718 { 719 struct bkey_ptrs ptrs = bch2_bkey_ptrs(bkey_i_to_s(k)); 720 struct bch_extent_crc_unpacked crc = 721 bch2_extent_crc_unpack(&k->k, NULL); 722 union bch_extent_entry *pos; 723 724 if (!bch2_crc_unpacked_cmp(crc, p->crc)) { 725 pos = ptrs.start; 726 goto found; 727 } 728 729 bkey_for_each_crc(&k->k, ptrs, crc, pos) 730 if (!bch2_crc_unpacked_cmp(crc, p->crc)) { 731 pos = extent_entry_next(pos); 732 goto found; 733 } 734 735 bch2_extent_crc_append(k, p->crc); 736 pos = bkey_val_end(bkey_i_to_s(k)); 737 found: 738 p->ptr.type = 1 << BCH_EXTENT_ENTRY_ptr; 739 __extent_entry_insert(k, pos, to_entry(&p->ptr)); 740 741 if (p->has_ec) { 742 p->ec.type = 1 << BCH_EXTENT_ENTRY_stripe_ptr; 743 __extent_entry_insert(k, pos, to_entry(&p->ec)); 744 } 745 } 746 747 static union bch_extent_entry *extent_entry_prev(struct bkey_ptrs ptrs, 748 union bch_extent_entry *entry) 749 { 750 union bch_extent_entry *i = ptrs.start; 751 752 if (i == entry) 753 return NULL; 754 755 while (extent_entry_next(i) != entry) 756 i = extent_entry_next(i); 757 return i; 758 } 759 760 static void extent_entry_drop(struct bkey_s k, union bch_extent_entry *entry) 761 { 762 union bch_extent_entry *next = extent_entry_next(entry); 763 764 /* stripes have ptrs, but their layout doesn't work with this code */ 765 BUG_ON(k.k->type == KEY_TYPE_stripe); 766 767 memmove_u64s_down(entry, next, 768 (u64 *) bkey_val_end(k) - (u64 *) next); 769 k.k->u64s -= (u64 *) next - (u64 *) entry; 770 } 771 772 /* 773 * Returns pointer to the next entry after the one being dropped: 774 */ 775 union bch_extent_entry *bch2_bkey_drop_ptr_noerror(struct bkey_s k, 776 struct bch_extent_ptr *ptr) 777 { 778 struct bkey_ptrs ptrs = bch2_bkey_ptrs(k); 779 union bch_extent_entry *entry = to_entry(ptr), *next; 780 union bch_extent_entry *ret = entry; 781 bool drop_crc = true; 782 783 EBUG_ON(ptr < &ptrs.start->ptr || 784 ptr >= &ptrs.end->ptr); 785 EBUG_ON(ptr->type != 1 << BCH_EXTENT_ENTRY_ptr); 786 787 for (next = extent_entry_next(entry); 788 next != ptrs.end; 789 next = extent_entry_next(next)) { 790 if (extent_entry_is_crc(next)) { 791 break; 792 } else if (extent_entry_is_ptr(next)) { 793 drop_crc = false; 794 break; 795 } 796 } 797 798 extent_entry_drop(k, entry); 799 800 while ((entry = extent_entry_prev(ptrs, entry))) { 801 if (extent_entry_is_ptr(entry)) 802 break; 803 804 if ((extent_entry_is_crc(entry) && drop_crc) || 805 extent_entry_is_stripe_ptr(entry)) { 806 ret = (void *) ret - extent_entry_bytes(entry); 807 extent_entry_drop(k, entry); 808 } 809 } 810 811 return ret; 812 } 813 814 union bch_extent_entry *bch2_bkey_drop_ptr(struct bkey_s k, 815 struct bch_extent_ptr *ptr) 816 { 817 bool have_dirty = bch2_bkey_dirty_devs(k.s_c).nr; 818 union bch_extent_entry *ret = 819 bch2_bkey_drop_ptr_noerror(k, ptr); 820 821 /* 822 * If we deleted all the dirty pointers and there's still cached 823 * pointers, we could set the cached pointers to dirty if they're not 824 * stale - but to do that correctly we'd need to grab an open_bucket 825 * reference so that we don't race with bucket reuse: 826 */ 827 if (have_dirty && 828 !bch2_bkey_dirty_devs(k.s_c).nr) { 829 k.k->type = KEY_TYPE_error; 830 set_bkey_val_u64s(k.k, 0); 831 ret = NULL; 832 } else if (!bch2_bkey_nr_ptrs(k.s_c)) { 833 k.k->type = KEY_TYPE_deleted; 834 set_bkey_val_u64s(k.k, 0); 835 ret = NULL; 836 } 837 838 return ret; 839 } 840 841 void bch2_bkey_drop_device(struct bkey_s k, unsigned dev) 842 { 843 struct bch_extent_ptr *ptr; 844 845 bch2_bkey_drop_ptrs(k, ptr, ptr->dev == dev); 846 } 847 848 void bch2_bkey_drop_device_noerror(struct bkey_s k, unsigned dev) 849 { 850 struct bch_extent_ptr *ptr = bch2_bkey_has_device(k, dev); 851 852 if (ptr) 853 bch2_bkey_drop_ptr_noerror(k, ptr); 854 } 855 856 const struct bch_extent_ptr *bch2_bkey_has_device_c(struct bkey_s_c k, unsigned dev) 857 { 858 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 859 const struct bch_extent_ptr *ptr; 860 861 bkey_for_each_ptr(ptrs, ptr) 862 if (ptr->dev == dev) 863 return ptr; 864 865 return NULL; 866 } 867 868 bool bch2_bkey_has_target(struct bch_fs *c, struct bkey_s_c k, unsigned target) 869 { 870 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 871 const struct bch_extent_ptr *ptr; 872 873 bkey_for_each_ptr(ptrs, ptr) 874 if (bch2_dev_in_target(c, ptr->dev, target) && 875 (!ptr->cached || 876 !ptr_stale(bch_dev_bkey_exists(c, ptr->dev), ptr))) 877 return true; 878 879 return false; 880 } 881 882 bool bch2_bkey_matches_ptr(struct bch_fs *c, struct bkey_s_c k, 883 struct bch_extent_ptr m, u64 offset) 884 { 885 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 886 const union bch_extent_entry *entry; 887 struct extent_ptr_decoded p; 888 889 bkey_for_each_ptr_decode(k.k, ptrs, p, entry) 890 if (p.ptr.dev == m.dev && 891 p.ptr.gen == m.gen && 892 (s64) p.ptr.offset + p.crc.offset - bkey_start_offset(k.k) == 893 (s64) m.offset - offset) 894 return true; 895 896 return false; 897 } 898 899 /* 900 * Returns true if two extents refer to the same data: 901 */ 902 bool bch2_extents_match(struct bkey_s_c k1, struct bkey_s_c k2) 903 { 904 if (k1.k->type != k2.k->type) 905 return false; 906 907 if (bkey_extent_is_direct_data(k1.k)) { 908 struct bkey_ptrs_c ptrs1 = bch2_bkey_ptrs_c(k1); 909 struct bkey_ptrs_c ptrs2 = bch2_bkey_ptrs_c(k2); 910 const union bch_extent_entry *entry1, *entry2; 911 struct extent_ptr_decoded p1, p2; 912 913 if (bkey_extent_is_unwritten(k1) != bkey_extent_is_unwritten(k2)) 914 return false; 915 916 bkey_for_each_ptr_decode(k1.k, ptrs1, p1, entry1) 917 bkey_for_each_ptr_decode(k2.k, ptrs2, p2, entry2) 918 if (p1.ptr.dev == p2.ptr.dev && 919 p1.ptr.gen == p2.ptr.gen && 920 (s64) p1.ptr.offset + p1.crc.offset - bkey_start_offset(k1.k) == 921 (s64) p2.ptr.offset + p2.crc.offset - bkey_start_offset(k2.k)) 922 return true; 923 924 return false; 925 } else { 926 /* KEY_TYPE_deleted, etc. */ 927 return true; 928 } 929 } 930 931 struct bch_extent_ptr * 932 bch2_extent_has_ptr(struct bkey_s_c k1, struct extent_ptr_decoded p1, struct bkey_s k2) 933 { 934 struct bkey_ptrs ptrs2 = bch2_bkey_ptrs(k2); 935 union bch_extent_entry *entry2; 936 struct extent_ptr_decoded p2; 937 938 bkey_for_each_ptr_decode(k2.k, ptrs2, p2, entry2) 939 if (p1.ptr.dev == p2.ptr.dev && 940 p1.ptr.gen == p2.ptr.gen && 941 (s64) p1.ptr.offset + p1.crc.offset - bkey_start_offset(k1.k) == 942 (s64) p2.ptr.offset + p2.crc.offset - bkey_start_offset(k2.k)) 943 return &entry2->ptr; 944 945 return NULL; 946 } 947 948 void bch2_extent_ptr_set_cached(struct bkey_s k, struct bch_extent_ptr *ptr) 949 { 950 struct bkey_ptrs ptrs = bch2_bkey_ptrs(k); 951 union bch_extent_entry *entry; 952 union bch_extent_entry *ec = NULL; 953 954 bkey_extent_entry_for_each(ptrs, entry) { 955 if (&entry->ptr == ptr) { 956 ptr->cached = true; 957 if (ec) 958 extent_entry_drop(k, ec); 959 return; 960 } 961 962 if (extent_entry_is_stripe_ptr(entry)) 963 ec = entry; 964 else if (extent_entry_is_ptr(entry)) 965 ec = NULL; 966 } 967 968 BUG(); 969 } 970 971 /* 972 * bch_extent_normalize - clean up an extent, dropping stale pointers etc. 973 * 974 * Returns true if @k should be dropped entirely 975 * 976 * For existing keys, only called when btree nodes are being rewritten, not when 977 * they're merely being compacted/resorted in memory. 978 */ 979 bool bch2_extent_normalize(struct bch_fs *c, struct bkey_s k) 980 { 981 struct bch_extent_ptr *ptr; 982 983 bch2_bkey_drop_ptrs(k, ptr, 984 ptr->cached && 985 ptr_stale(bch_dev_bkey_exists(c, ptr->dev), ptr)); 986 987 return bkey_deleted(k.k); 988 } 989 990 void bch2_bkey_ptrs_to_text(struct printbuf *out, struct bch_fs *c, 991 struct bkey_s_c k) 992 { 993 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 994 const union bch_extent_entry *entry; 995 struct bch_extent_crc_unpacked crc; 996 const struct bch_extent_ptr *ptr; 997 const struct bch_extent_stripe_ptr *ec; 998 struct bch_dev *ca; 999 bool first = true; 1000 1001 if (c) 1002 prt_printf(out, "durability: %u ", bch2_bkey_durability_safe(c, k)); 1003 1004 bkey_extent_entry_for_each(ptrs, entry) { 1005 if (!first) 1006 prt_printf(out, " "); 1007 1008 switch (__extent_entry_type(entry)) { 1009 case BCH_EXTENT_ENTRY_ptr: 1010 ptr = entry_to_ptr(entry); 1011 ca = c && ptr->dev < c->sb.nr_devices && c->devs[ptr->dev] 1012 ? bch_dev_bkey_exists(c, ptr->dev) 1013 : NULL; 1014 1015 if (!ca) { 1016 prt_printf(out, "ptr: %u:%llu gen %u%s", ptr->dev, 1017 (u64) ptr->offset, ptr->gen, 1018 ptr->cached ? " cached" : ""); 1019 } else { 1020 u32 offset; 1021 u64 b = sector_to_bucket_and_offset(ca, ptr->offset, &offset); 1022 1023 prt_printf(out, "ptr: %u:%llu:%u gen %u", 1024 ptr->dev, b, offset, ptr->gen); 1025 if (ptr->cached) 1026 prt_str(out, " cached"); 1027 if (ptr->unwritten) 1028 prt_str(out, " unwritten"); 1029 if (ca && ptr_stale(ca, ptr)) 1030 prt_printf(out, " stale"); 1031 } 1032 break; 1033 case BCH_EXTENT_ENTRY_crc32: 1034 case BCH_EXTENT_ENTRY_crc64: 1035 case BCH_EXTENT_ENTRY_crc128: 1036 crc = bch2_extent_crc_unpack(k.k, entry_to_crc(entry)); 1037 1038 prt_printf(out, "crc: c_size %u size %u offset %u nonce %u csum %s compress %s", 1039 crc.compressed_size, 1040 crc.uncompressed_size, 1041 crc.offset, crc.nonce, 1042 bch2_csum_types[crc.csum_type], 1043 bch2_compression_types[crc.compression_type]); 1044 break; 1045 case BCH_EXTENT_ENTRY_stripe_ptr: 1046 ec = &entry->stripe_ptr; 1047 1048 prt_printf(out, "ec: idx %llu block %u", 1049 (u64) ec->idx, ec->block); 1050 break; 1051 default: 1052 prt_printf(out, "(invalid extent entry %.16llx)", *((u64 *) entry)); 1053 return; 1054 } 1055 1056 first = false; 1057 } 1058 } 1059 1060 static int extent_ptr_invalid(const struct bch_fs *c, 1061 struct bkey_s_c k, 1062 enum bkey_invalid_flags flags, 1063 const struct bch_extent_ptr *ptr, 1064 unsigned size_ondisk, 1065 bool metadata, 1066 struct printbuf *err) 1067 { 1068 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 1069 const struct bch_extent_ptr *ptr2; 1070 u64 bucket; 1071 u32 bucket_offset; 1072 struct bch_dev *ca; 1073 1074 if (!bch2_dev_exists2(c, ptr->dev)) { 1075 /* 1076 * If we're in the write path this key might have already been 1077 * overwritten, and we could be seeing a device that doesn't 1078 * exist anymore due to racing with device removal: 1079 */ 1080 if (flags & BKEY_INVALID_WRITE) 1081 return 0; 1082 1083 prt_printf(err, "pointer to invalid device (%u)", ptr->dev); 1084 return -BCH_ERR_invalid_bkey; 1085 } 1086 1087 ca = bch_dev_bkey_exists(c, ptr->dev); 1088 bkey_for_each_ptr(ptrs, ptr2) 1089 if (ptr != ptr2 && ptr->dev == ptr2->dev) { 1090 prt_printf(err, "multiple pointers to same device (%u)", ptr->dev); 1091 return -BCH_ERR_invalid_bkey; 1092 } 1093 1094 bucket = sector_to_bucket_and_offset(ca, ptr->offset, &bucket_offset); 1095 1096 if (bucket >= ca->mi.nbuckets) { 1097 prt_printf(err, "pointer past last bucket (%llu > %llu)", 1098 bucket, ca->mi.nbuckets); 1099 return -BCH_ERR_invalid_bkey; 1100 } 1101 1102 if (ptr->offset < bucket_to_sector(ca, ca->mi.first_bucket)) { 1103 prt_printf(err, "pointer before first bucket (%llu < %u)", 1104 bucket, ca->mi.first_bucket); 1105 return -BCH_ERR_invalid_bkey; 1106 } 1107 1108 if (bucket_offset + size_ondisk > ca->mi.bucket_size) { 1109 prt_printf(err, "pointer spans multiple buckets (%u + %u > %u)", 1110 bucket_offset, size_ondisk, ca->mi.bucket_size); 1111 return -BCH_ERR_invalid_bkey; 1112 } 1113 1114 return 0; 1115 } 1116 1117 int bch2_bkey_ptrs_invalid(const struct bch_fs *c, struct bkey_s_c k, 1118 enum bkey_invalid_flags flags, 1119 struct printbuf *err) 1120 { 1121 struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k); 1122 const union bch_extent_entry *entry; 1123 struct bch_extent_crc_unpacked crc; 1124 unsigned size_ondisk = k.k->size; 1125 unsigned nonce = UINT_MAX; 1126 unsigned nr_ptrs = 0; 1127 bool unwritten = false, have_ec = false, crc_since_last_ptr = false; 1128 int ret; 1129 1130 if (bkey_is_btree_ptr(k.k)) 1131 size_ondisk = btree_sectors(c); 1132 1133 bkey_extent_entry_for_each(ptrs, entry) { 1134 if (__extent_entry_type(entry) >= BCH_EXTENT_ENTRY_MAX) { 1135 prt_printf(err, "invalid extent entry type (got %u, max %u)", 1136 __extent_entry_type(entry), BCH_EXTENT_ENTRY_MAX); 1137 return -BCH_ERR_invalid_bkey; 1138 } 1139 1140 if (bkey_is_btree_ptr(k.k) && 1141 !extent_entry_is_ptr(entry)) { 1142 prt_printf(err, "has non ptr field"); 1143 return -BCH_ERR_invalid_bkey; 1144 } 1145 1146 switch (extent_entry_type(entry)) { 1147 case BCH_EXTENT_ENTRY_ptr: 1148 ret = extent_ptr_invalid(c, k, flags, &entry->ptr, 1149 size_ondisk, false, err); 1150 if (ret) 1151 return ret; 1152 1153 if (nr_ptrs && unwritten != entry->ptr.unwritten) { 1154 prt_printf(err, "extent with unwritten and written ptrs"); 1155 return -BCH_ERR_invalid_bkey; 1156 } 1157 1158 if (k.k->type != KEY_TYPE_extent && entry->ptr.unwritten) { 1159 prt_printf(err, "has unwritten ptrs"); 1160 return -BCH_ERR_invalid_bkey; 1161 } 1162 1163 if (entry->ptr.cached && have_ec) { 1164 prt_printf(err, "cached, erasure coded ptr"); 1165 return -BCH_ERR_invalid_bkey; 1166 } 1167 1168 unwritten = entry->ptr.unwritten; 1169 have_ec = false; 1170 crc_since_last_ptr = false; 1171 nr_ptrs++; 1172 break; 1173 case BCH_EXTENT_ENTRY_crc32: 1174 case BCH_EXTENT_ENTRY_crc64: 1175 case BCH_EXTENT_ENTRY_crc128: 1176 crc = bch2_extent_crc_unpack(k.k, entry_to_crc(entry)); 1177 1178 if (crc.offset + crc.live_size > 1179 crc.uncompressed_size) { 1180 prt_printf(err, "checksum offset + key size > uncompressed size"); 1181 return -BCH_ERR_invalid_bkey; 1182 } 1183 1184 size_ondisk = crc.compressed_size; 1185 1186 if (!bch2_checksum_type_valid(c, crc.csum_type)) { 1187 prt_printf(err, "invalid checksum type"); 1188 return -BCH_ERR_invalid_bkey; 1189 } 1190 1191 if (crc.compression_type >= BCH_COMPRESSION_TYPE_NR) { 1192 prt_printf(err, "invalid compression type"); 1193 return -BCH_ERR_invalid_bkey; 1194 } 1195 1196 if (bch2_csum_type_is_encryption(crc.csum_type)) { 1197 if (nonce == UINT_MAX) 1198 nonce = crc.offset + crc.nonce; 1199 else if (nonce != crc.offset + crc.nonce) { 1200 prt_printf(err, "incorrect nonce"); 1201 return -BCH_ERR_invalid_bkey; 1202 } 1203 } 1204 1205 if (crc_since_last_ptr) { 1206 prt_printf(err, "redundant crc entry"); 1207 return -BCH_ERR_invalid_bkey; 1208 } 1209 crc_since_last_ptr = true; 1210 break; 1211 case BCH_EXTENT_ENTRY_stripe_ptr: 1212 if (have_ec) { 1213 prt_printf(err, "redundant stripe entry"); 1214 return -BCH_ERR_invalid_bkey; 1215 } 1216 have_ec = true; 1217 break; 1218 case BCH_EXTENT_ENTRY_rebalance: 1219 break; 1220 } 1221 } 1222 1223 if (!nr_ptrs) { 1224 prt_str(err, "no ptrs"); 1225 return -BCH_ERR_invalid_bkey; 1226 } 1227 1228 if (nr_ptrs >= BCH_BKEY_PTRS_MAX) { 1229 prt_str(err, "too many ptrs"); 1230 return -BCH_ERR_invalid_bkey; 1231 } 1232 1233 if (crc_since_last_ptr) { 1234 prt_printf(err, "redundant crc entry"); 1235 return -BCH_ERR_invalid_bkey; 1236 } 1237 1238 if (have_ec) { 1239 prt_printf(err, "redundant stripe entry"); 1240 return -BCH_ERR_invalid_bkey; 1241 } 1242 1243 return 0; 1244 } 1245 1246 void bch2_ptr_swab(struct bkey_s k) 1247 { 1248 struct bkey_ptrs ptrs = bch2_bkey_ptrs(k); 1249 union bch_extent_entry *entry; 1250 u64 *d; 1251 1252 for (d = (u64 *) ptrs.start; 1253 d != (u64 *) ptrs.end; 1254 d++) 1255 *d = swab64(*d); 1256 1257 for (entry = ptrs.start; 1258 entry < ptrs.end; 1259 entry = extent_entry_next(entry)) { 1260 switch (extent_entry_type(entry)) { 1261 case BCH_EXTENT_ENTRY_ptr: 1262 break; 1263 case BCH_EXTENT_ENTRY_crc32: 1264 entry->crc32.csum = swab32(entry->crc32.csum); 1265 break; 1266 case BCH_EXTENT_ENTRY_crc64: 1267 entry->crc64.csum_hi = swab16(entry->crc64.csum_hi); 1268 entry->crc64.csum_lo = swab64(entry->crc64.csum_lo); 1269 break; 1270 case BCH_EXTENT_ENTRY_crc128: 1271 entry->crc128.csum.hi = (__force __le64) 1272 swab64((__force u64) entry->crc128.csum.hi); 1273 entry->crc128.csum.lo = (__force __le64) 1274 swab64((__force u64) entry->crc128.csum.lo); 1275 break; 1276 case BCH_EXTENT_ENTRY_stripe_ptr: 1277 break; 1278 case BCH_EXTENT_ENTRY_rebalance: 1279 break; 1280 } 1281 } 1282 } 1283 1284 /* Generic extent code: */ 1285 1286 int bch2_cut_front_s(struct bpos where, struct bkey_s k) 1287 { 1288 unsigned new_val_u64s = bkey_val_u64s(k.k); 1289 int val_u64s_delta; 1290 u64 sub; 1291 1292 if (bkey_le(where, bkey_start_pos(k.k))) 1293 return 0; 1294 1295 EBUG_ON(bkey_gt(where, k.k->p)); 1296 1297 sub = where.offset - bkey_start_offset(k.k); 1298 1299 k.k->size -= sub; 1300 1301 if (!k.k->size) { 1302 k.k->type = KEY_TYPE_deleted; 1303 new_val_u64s = 0; 1304 } 1305 1306 switch (k.k->type) { 1307 case KEY_TYPE_extent: 1308 case KEY_TYPE_reflink_v: { 1309 struct bkey_ptrs ptrs = bch2_bkey_ptrs(k); 1310 union bch_extent_entry *entry; 1311 bool seen_crc = false; 1312 1313 bkey_extent_entry_for_each(ptrs, entry) { 1314 switch (extent_entry_type(entry)) { 1315 case BCH_EXTENT_ENTRY_ptr: 1316 if (!seen_crc) 1317 entry->ptr.offset += sub; 1318 break; 1319 case BCH_EXTENT_ENTRY_crc32: 1320 entry->crc32.offset += sub; 1321 break; 1322 case BCH_EXTENT_ENTRY_crc64: 1323 entry->crc64.offset += sub; 1324 break; 1325 case BCH_EXTENT_ENTRY_crc128: 1326 entry->crc128.offset += sub; 1327 break; 1328 case BCH_EXTENT_ENTRY_stripe_ptr: 1329 break; 1330 case BCH_EXTENT_ENTRY_rebalance: 1331 break; 1332 } 1333 1334 if (extent_entry_is_crc(entry)) 1335 seen_crc = true; 1336 } 1337 1338 break; 1339 } 1340 case KEY_TYPE_reflink_p: { 1341 struct bkey_s_reflink_p p = bkey_s_to_reflink_p(k); 1342 1343 le64_add_cpu(&p.v->idx, sub); 1344 break; 1345 } 1346 case KEY_TYPE_inline_data: 1347 case KEY_TYPE_indirect_inline_data: { 1348 void *p = bkey_inline_data_p(k); 1349 unsigned bytes = bkey_inline_data_bytes(k.k); 1350 1351 sub = min_t(u64, sub << 9, bytes); 1352 1353 memmove(p, p + sub, bytes - sub); 1354 1355 new_val_u64s -= sub >> 3; 1356 break; 1357 } 1358 } 1359 1360 val_u64s_delta = bkey_val_u64s(k.k) - new_val_u64s; 1361 BUG_ON(val_u64s_delta < 0); 1362 1363 set_bkey_val_u64s(k.k, new_val_u64s); 1364 memset(bkey_val_end(k), 0, val_u64s_delta * sizeof(u64)); 1365 return -val_u64s_delta; 1366 } 1367 1368 int bch2_cut_back_s(struct bpos where, struct bkey_s k) 1369 { 1370 unsigned new_val_u64s = bkey_val_u64s(k.k); 1371 int val_u64s_delta; 1372 u64 len = 0; 1373 1374 if (bkey_ge(where, k.k->p)) 1375 return 0; 1376 1377 EBUG_ON(bkey_lt(where, bkey_start_pos(k.k))); 1378 1379 len = where.offset - bkey_start_offset(k.k); 1380 1381 k.k->p.offset = where.offset; 1382 k.k->size = len; 1383 1384 if (!len) { 1385 k.k->type = KEY_TYPE_deleted; 1386 new_val_u64s = 0; 1387 } 1388 1389 switch (k.k->type) { 1390 case KEY_TYPE_inline_data: 1391 case KEY_TYPE_indirect_inline_data: 1392 new_val_u64s = (bkey_inline_data_offset(k.k) + 1393 min(bkey_inline_data_bytes(k.k), k.k->size << 9)) >> 3; 1394 break; 1395 } 1396 1397 val_u64s_delta = bkey_val_u64s(k.k) - new_val_u64s; 1398 BUG_ON(val_u64s_delta < 0); 1399 1400 set_bkey_val_u64s(k.k, new_val_u64s); 1401 memset(bkey_val_end(k), 0, val_u64s_delta * sizeof(u64)); 1402 return -val_u64s_delta; 1403 } 1404