1 /* SPDX-License-Identifier: GPL-2.0 */ 2 #ifndef _BCACHE_H 3 #define _BCACHE_H 4 5 /* 6 * SOME HIGH LEVEL CODE DOCUMENTATION: 7 * 8 * Bcache mostly works with cache sets, cache devices, and backing devices. 9 * 10 * Support for multiple cache devices hasn't quite been finished off yet, but 11 * it's about 95% plumbed through. A cache set and its cache devices is sort of 12 * like a md raid array and its component devices. Most of the code doesn't care 13 * about individual cache devices, the main abstraction is the cache set. 14 * 15 * Multiple cache devices is intended to give us the ability to mirror dirty 16 * cached data and metadata, without mirroring clean cached data. 17 * 18 * Backing devices are different, in that they have a lifetime independent of a 19 * cache set. When you register a newly formatted backing device it'll come up 20 * in passthrough mode, and then you can attach and detach a backing device from 21 * a cache set at runtime - while it's mounted and in use. Detaching implicitly 22 * invalidates any cached data for that backing device. 23 * 24 * A cache set can have multiple (many) backing devices attached to it. 25 * 26 * There's also flash only volumes - this is the reason for the distinction 27 * between struct cached_dev and struct bcache_device. A flash only volume 28 * works much like a bcache device that has a backing device, except the 29 * "cached" data is always dirty. The end result is that we get thin 30 * provisioning with very little additional code. 31 * 32 * Flash only volumes work but they're not production ready because the moving 33 * garbage collector needs more work. More on that later. 34 * 35 * BUCKETS/ALLOCATION: 36 * 37 * Bcache is primarily designed for caching, which means that in normal 38 * operation all of our available space will be allocated. Thus, we need an 39 * efficient way of deleting things from the cache so we can write new things to 40 * it. 41 * 42 * To do this, we first divide the cache device up into buckets. A bucket is the 43 * unit of allocation; they're typically around 1 mb - anywhere from 128k to 2M+ 44 * works efficiently. 45 * 46 * Each bucket has a 16 bit priority, and an 8 bit generation associated with 47 * it. The gens and priorities for all the buckets are stored contiguously and 48 * packed on disk (in a linked list of buckets - aside from the superblock, all 49 * of bcache's metadata is stored in buckets). 50 * 51 * The priority is used to implement an LRU. We reset a bucket's priority when 52 * we allocate it or on cache it, and every so often we decrement the priority 53 * of each bucket. It could be used to implement something more sophisticated, 54 * if anyone ever gets around to it. 55 * 56 * The generation is used for invalidating buckets. Each pointer also has an 8 57 * bit generation embedded in it; for a pointer to be considered valid, its gen 58 * must match the gen of the bucket it points into. Thus, to reuse a bucket all 59 * we have to do is increment its gen (and write its new gen to disk; we batch 60 * this up). 61 * 62 * Bcache is entirely COW - we never write twice to a bucket, even buckets that 63 * contain metadata (including btree nodes). 64 * 65 * THE BTREE: 66 * 67 * Bcache is in large part design around the btree. 68 * 69 * At a high level, the btree is just an index of key -> ptr tuples. 70 * 71 * Keys represent extents, and thus have a size field. Keys also have a variable 72 * number of pointers attached to them (potentially zero, which is handy for 73 * invalidating the cache). 74 * 75 * The key itself is an inode:offset pair. The inode number corresponds to a 76 * backing device or a flash only volume. The offset is the ending offset of the 77 * extent within the inode - not the starting offset; this makes lookups 78 * slightly more convenient. 79 * 80 * Pointers contain the cache device id, the offset on that device, and an 8 bit 81 * generation number. More on the gen later. 82 * 83 * Index lookups are not fully abstracted - cache lookups in particular are 84 * still somewhat mixed in with the btree code, but things are headed in that 85 * direction. 86 * 87 * Updates are fairly well abstracted, though. There are two different ways of 88 * updating the btree; insert and replace. 89 * 90 * BTREE_INSERT will just take a list of keys and insert them into the btree - 91 * overwriting (possibly only partially) any extents they overlap with. This is 92 * used to update the index after a write. 93 * 94 * BTREE_REPLACE is really cmpxchg(); it inserts a key into the btree iff it is 95 * overwriting a key that matches another given key. This is used for inserting 96 * data into the cache after a cache miss, and for background writeback, and for 97 * the moving garbage collector. 98 * 99 * There is no "delete" operation; deleting things from the index is 100 * accomplished by either by invalidating pointers (by incrementing a bucket's 101 * gen) or by inserting a key with 0 pointers - which will overwrite anything 102 * previously present at that location in the index. 103 * 104 * This means that there are always stale/invalid keys in the btree. They're 105 * filtered out by the code that iterates through a btree node, and removed when 106 * a btree node is rewritten. 107 * 108 * BTREE NODES: 109 * 110 * Our unit of allocation is a bucket, and we can't arbitrarily allocate and 111 * free smaller than a bucket - so, that's how big our btree nodes are. 112 * 113 * (If buckets are really big we'll only use part of the bucket for a btree node 114 * - no less than 1/4th - but a bucket still contains no more than a single 115 * btree node. I'd actually like to change this, but for now we rely on the 116 * bucket's gen for deleting btree nodes when we rewrite/split a node.) 117 * 118 * Anyways, btree nodes are big - big enough to be inefficient with a textbook 119 * btree implementation. 120 * 121 * The way this is solved is that btree nodes are internally log structured; we 122 * can append new keys to an existing btree node without rewriting it. This 123 * means each set of keys we write is sorted, but the node is not. 124 * 125 * We maintain this log structure in memory - keeping 1Mb of keys sorted would 126 * be expensive, and we have to distinguish between the keys we have written and 127 * the keys we haven't. So to do a lookup in a btree node, we have to search 128 * each sorted set. But we do merge written sets together lazily, so the cost of 129 * these extra searches is quite low (normally most of the keys in a btree node 130 * will be in one big set, and then there'll be one or two sets that are much 131 * smaller). 132 * 133 * This log structure makes bcache's btree more of a hybrid between a 134 * conventional btree and a compacting data structure, with some of the 135 * advantages of both. 136 * 137 * GARBAGE COLLECTION: 138 * 139 * We can't just invalidate any bucket - it might contain dirty data or 140 * metadata. If it once contained dirty data, other writes might overwrite it 141 * later, leaving no valid pointers into that bucket in the index. 142 * 143 * Thus, the primary purpose of garbage collection is to find buckets to reuse. 144 * It also counts how much valid data it each bucket currently contains, so that 145 * allocation can reuse buckets sooner when they've been mostly overwritten. 146 * 147 * It also does some things that are really internal to the btree 148 * implementation. If a btree node contains pointers that are stale by more than 149 * some threshold, it rewrites the btree node to avoid the bucket's generation 150 * wrapping around. It also merges adjacent btree nodes if they're empty enough. 151 * 152 * THE JOURNAL: 153 * 154 * Bcache's journal is not necessary for consistency; we always strictly 155 * order metadata writes so that the btree and everything else is consistent on 156 * disk in the event of an unclean shutdown, and in fact bcache had writeback 157 * caching (with recovery from unclean shutdown) before journalling was 158 * implemented. 159 * 160 * Rather, the journal is purely a performance optimization; we can't complete a 161 * write until we've updated the index on disk, otherwise the cache would be 162 * inconsistent in the event of an unclean shutdown. This means that without the 163 * journal, on random write workloads we constantly have to update all the leaf 164 * nodes in the btree, and those writes will be mostly empty (appending at most 165 * a few keys each) - highly inefficient in terms of amount of metadata writes, 166 * and it puts more strain on the various btree resorting/compacting code. 167 * 168 * The journal is just a log of keys we've inserted; on startup we just reinsert 169 * all the keys in the open journal entries. That means that when we're updating 170 * a node in the btree, we can wait until a 4k block of keys fills up before 171 * writing them out. 172 * 173 * For simplicity, we only journal updates to leaf nodes; updates to parent 174 * nodes are rare enough (since our leaf nodes are huge) that it wasn't worth 175 * the complexity to deal with journalling them (in particular, journal replay) 176 * - updates to non leaf nodes just happen synchronously (see btree_split()). 177 */ 178 179 #define pr_fmt(fmt) "bcache: %s() " fmt, __func__ 180 181 #include <linux/bio.h> 182 #include <linux/closure.h> 183 #include <linux/kobject.h> 184 #include <linux/list.h> 185 #include <linux/mutex.h> 186 #include <linux/rbtree.h> 187 #include <linux/rwsem.h> 188 #include <linux/refcount.h> 189 #include <linux/types.h> 190 #include <linux/workqueue.h> 191 #include <linux/kthread.h> 192 193 #include "bcache_ondisk.h" 194 #include "bset.h" 195 #include "util.h" 196 197 struct bucket { 198 atomic_t pin; 199 uint16_t prio; 200 uint8_t gen; 201 uint8_t last_gc; /* Most out of date gen in the btree */ 202 uint16_t gc_mark; /* Bitfield used by GC. See below for field */ 203 uint16_t reclaimable_in_gc:1; 204 }; 205 206 /* 207 * I'd use bitfields for these, but I don't trust the compiler not to screw me 208 * as multiple threads touch struct bucket without locking 209 */ 210 211 BITMASK(GC_MARK, struct bucket, gc_mark, 0, 2); 212 #define GC_MARK_RECLAIMABLE 1 213 #define GC_MARK_DIRTY 2 214 #define GC_MARK_METADATA 3 215 #define GC_SECTORS_USED_SIZE 13 216 #define MAX_GC_SECTORS_USED (~(~0ULL << GC_SECTORS_USED_SIZE)) 217 BITMASK(GC_SECTORS_USED, struct bucket, gc_mark, 2, GC_SECTORS_USED_SIZE); 218 BITMASK(GC_MOVE, struct bucket, gc_mark, 15, 1); 219 220 #include "journal.h" 221 #include "stats.h" 222 struct search; 223 struct btree; 224 struct keybuf; 225 226 struct keybuf_key { 227 struct rb_node node; 228 BKEY_PADDED(key); 229 void *private; 230 }; 231 232 struct keybuf { 233 struct bkey last_scanned; 234 spinlock_t lock; 235 236 /* 237 * Beginning and end of range in rb tree - so that we can skip taking 238 * lock and checking the rb tree when we need to check for overlapping 239 * keys. 240 */ 241 struct bkey start; 242 struct bkey end; 243 244 struct rb_root keys; 245 246 #define KEYBUF_NR 500 247 DECLARE_ARRAY_ALLOCATOR(struct keybuf_key, freelist, KEYBUF_NR); 248 }; 249 250 struct bcache_device { 251 struct closure cl; 252 253 struct kobject kobj; 254 255 struct cache_set *c; 256 unsigned int id; 257 #define BCACHEDEVNAME_SIZE 12 258 char name[BCACHEDEVNAME_SIZE]; 259 260 struct gendisk *disk; 261 262 unsigned long flags; 263 #define BCACHE_DEV_CLOSING 0 264 #define BCACHE_DEV_DETACHING 1 265 #define BCACHE_DEV_UNLINK_DONE 2 266 #define BCACHE_DEV_WB_RUNNING 3 267 #define BCACHE_DEV_RATE_DW_RUNNING 4 268 int nr_stripes; 269 #define BCH_MIN_STRIPE_SZ ((4 << 20) >> SECTOR_SHIFT) 270 unsigned int stripe_size; 271 atomic_t *stripe_sectors_dirty; 272 unsigned long *full_dirty_stripes; 273 274 struct bio_set bio_split; 275 276 unsigned int data_csum:1; 277 278 int (*cache_miss)(struct btree *b, struct search *s, 279 struct bio *bio, unsigned int sectors); 280 int (*ioctl)(struct bcache_device *d, blk_mode_t mode, 281 unsigned int cmd, unsigned long arg); 282 }; 283 284 struct io { 285 /* Used to track sequential IO so it can be skipped */ 286 struct hlist_node hash; 287 struct list_head lru; 288 289 unsigned long jiffies; 290 unsigned int sequential; 291 sector_t last; 292 }; 293 294 enum stop_on_failure { 295 BCH_CACHED_DEV_STOP_AUTO = 0, 296 BCH_CACHED_DEV_STOP_ALWAYS, 297 BCH_CACHED_DEV_STOP_MODE_MAX, 298 }; 299 300 struct cached_dev { 301 struct list_head list; 302 struct bcache_device disk; 303 struct block_device *bdev; 304 struct file *bdev_file; 305 306 struct cache_sb sb; 307 struct cache_sb_disk *sb_disk; 308 struct bio sb_bio; 309 struct bio_vec sb_bv[1]; 310 struct closure sb_write; 311 struct semaphore sb_write_mutex; 312 313 /* Refcount on the cache set. Always nonzero when we're caching. */ 314 refcount_t count; 315 struct work_struct detach; 316 317 /* 318 * Device might not be running if it's dirty and the cache set hasn't 319 * showed up yet. 320 */ 321 atomic_t running; 322 323 /* 324 * Writes take a shared lock from start to finish; scanning for dirty 325 * data to refill the rb tree requires an exclusive lock. 326 */ 327 struct rw_semaphore writeback_lock; 328 329 /* 330 * Nonzero, and writeback has a refcount (d->count), iff there is dirty 331 * data in the cache. Protected by writeback_lock; must have an 332 * shared lock to set and exclusive lock to clear. 333 */ 334 atomic_t has_dirty; 335 336 #define BCH_CACHE_READA_ALL 0 337 #define BCH_CACHE_READA_META_ONLY 1 338 unsigned int cache_readahead_policy; 339 struct bch_ratelimit writeback_rate; 340 struct delayed_work writeback_rate_update; 341 342 /* Limit number of writeback bios in flight */ 343 struct semaphore in_flight; 344 struct task_struct *writeback_thread; 345 struct workqueue_struct *writeback_write_wq; 346 347 struct keybuf writeback_keys; 348 349 struct task_struct *status_update_thread; 350 /* 351 * Order the write-half of writeback operations strongly in dispatch 352 * order. (Maintain LBA order; don't allow reads completing out of 353 * order to re-order the writes...) 354 */ 355 struct closure_waitlist writeback_ordering_wait; 356 atomic_t writeback_sequence_next; 357 358 /* For tracking sequential IO */ 359 #define RECENT_IO_BITS 7 360 #define RECENT_IO (1 << RECENT_IO_BITS) 361 struct io io[RECENT_IO]; 362 struct hlist_head io_hash[RECENT_IO + 1]; 363 struct list_head io_lru; 364 spinlock_t io_lock; 365 366 struct cache_accounting accounting; 367 368 /* The rest of this all shows up in sysfs */ 369 unsigned int sequential_cutoff; 370 371 unsigned int io_disable:1; 372 unsigned int verify:1; 373 unsigned int bypass_torture_test:1; 374 375 unsigned int partial_stripes_expensive:1; 376 unsigned int writeback_metadata:1; 377 unsigned int writeback_running:1; 378 unsigned int writeback_consider_fragment:1; 379 unsigned char writeback_percent; 380 unsigned int writeback_delay; 381 382 uint64_t writeback_rate_target; 383 int64_t writeback_rate_proportional; 384 int64_t writeback_rate_integral; 385 int64_t writeback_rate_integral_scaled; 386 int32_t writeback_rate_change; 387 388 unsigned int writeback_rate_update_seconds; 389 unsigned int writeback_rate_i_term_inverse; 390 unsigned int writeback_rate_p_term_inverse; 391 unsigned int writeback_rate_fp_term_low; 392 unsigned int writeback_rate_fp_term_mid; 393 unsigned int writeback_rate_fp_term_high; 394 unsigned int writeback_rate_minimum; 395 396 enum stop_on_failure stop_when_cache_set_failed; 397 #define DEFAULT_CACHED_DEV_ERROR_LIMIT 64 398 atomic_t io_errors; 399 unsigned int error_limit; 400 unsigned int offline_seconds; 401 402 /* 403 * Retry to update writeback_rate if contention happens for 404 * down_read(dc->writeback_lock) in update_writeback_rate() 405 */ 406 #define BCH_WBRATE_UPDATE_MAX_SKIPS 15 407 unsigned int rate_update_retry; 408 }; 409 410 enum alloc_reserve { 411 RESERVE_BTREE, 412 RESERVE_PRIO, 413 RESERVE_MOVINGGC, 414 RESERVE_NONE, 415 RESERVE_NR, 416 }; 417 418 struct cache { 419 struct cache_set *set; 420 struct cache_sb sb; 421 struct cache_sb_disk *sb_disk; 422 struct bio sb_bio; 423 struct bio_vec sb_bv[1]; 424 425 struct kobject kobj; 426 struct block_device *bdev; 427 struct file *bdev_file; 428 429 struct task_struct *alloc_thread; 430 431 struct closure prio; 432 struct prio_set *disk_buckets; 433 434 /* 435 * When allocating new buckets, prio_write() gets first dibs - since we 436 * may not be allocate at all without writing priorities and gens. 437 * prio_last_buckets[] contains the last buckets we wrote priorities to 438 * (so gc can mark them as metadata), prio_buckets[] contains the 439 * buckets allocated for the next prio write. 440 */ 441 uint64_t *prio_buckets; 442 uint64_t *prio_last_buckets; 443 444 /* 445 * free: Buckets that are ready to be used 446 * 447 * free_inc: Incoming buckets - these are buckets that currently have 448 * cached data in them, and we can't reuse them until after we write 449 * their new gen to disk. After prio_write() finishes writing the new 450 * gens/prios, they'll be moved to the free list (and possibly discarded 451 * in the process) 452 */ 453 DECLARE_FIFO(long, free)[RESERVE_NR]; 454 DECLARE_FIFO(long, free_inc); 455 456 size_t fifo_last_bucket; 457 458 /* Allocation stuff: */ 459 struct bucket *buckets; 460 461 DEFINE_MIN_HEAP(struct bucket *, cache_heap) heap; 462 463 /* 464 * If nonzero, we know we aren't going to find any buckets to invalidate 465 * until a gc finishes - otherwise we could pointlessly burn a ton of 466 * cpu 467 */ 468 unsigned int invalidate_needs_gc; 469 470 bool discard; /* Get rid of? */ 471 472 struct journal_device journal; 473 474 /* The rest of this all shows up in sysfs */ 475 #define IO_ERROR_SHIFT 20 476 atomic_t io_errors; 477 atomic_t io_count; 478 479 atomic_long_t meta_sectors_written; 480 atomic_long_t btree_sectors_written; 481 atomic_long_t sectors_written; 482 }; 483 484 struct gc_stat { 485 size_t nodes; 486 size_t nodes_pre; 487 size_t key_bytes; 488 489 size_t nkeys; 490 uint64_t data; /* sectors */ 491 unsigned int in_use; /* percent */ 492 }; 493 494 /* 495 * Flag bits, for how the cache set is shutting down, and what phase it's at: 496 * 497 * CACHE_SET_UNREGISTERING means we're not just shutting down, we're detaching 498 * all the backing devices first (their cached data gets invalidated, and they 499 * won't automatically reattach). 500 * 501 * CACHE_SET_STOPPING always gets set first when we're closing down a cache set; 502 * we'll continue to run normally for awhile with CACHE_SET_STOPPING set (i.e. 503 * flushing dirty data). 504 * 505 * CACHE_SET_RUNNING means all cache devices have been registered and journal 506 * replay is complete. 507 * 508 * CACHE_SET_IO_DISABLE is set when bcache is stopping the whold cache set, all 509 * external and internal I/O should be denied when this flag is set. 510 * 511 */ 512 #define CACHE_SET_UNREGISTERING 0 513 #define CACHE_SET_STOPPING 1 514 #define CACHE_SET_RUNNING 2 515 #define CACHE_SET_IO_DISABLE 3 516 517 struct cache_set { 518 struct closure cl; 519 520 struct list_head list; 521 struct kobject kobj; 522 struct kobject internal; 523 struct dentry *debug; 524 struct cache_accounting accounting; 525 526 unsigned long flags; 527 atomic_t idle_counter; 528 atomic_t at_max_writeback_rate; 529 530 struct cache *cache; 531 532 struct bcache_device **devices; 533 unsigned int devices_max_used; 534 atomic_t attached_dev_nr; 535 struct list_head cached_devs; 536 uint64_t cached_dev_sectors; 537 atomic_long_t flash_dev_dirty_sectors; 538 struct closure caching; 539 540 struct closure sb_write; 541 struct semaphore sb_write_mutex; 542 543 mempool_t search; 544 mempool_t bio_meta; 545 struct bio_set bio_split; 546 547 /* For the btree cache */ 548 struct shrinker *shrink; 549 550 /* For the btree cache and anything allocation related */ 551 struct mutex bucket_lock; 552 553 /* log2(bucket_size), in sectors */ 554 unsigned short bucket_bits; 555 556 /* log2(block_size), in sectors */ 557 unsigned short block_bits; 558 559 /* 560 * Default number of pages for a new btree node - may be less than a 561 * full bucket 562 */ 563 unsigned int btree_pages; 564 565 /* 566 * Lists of struct btrees; lru is the list for structs that have memory 567 * allocated for actual btree node, freed is for structs that do not. 568 * 569 * We never free a struct btree, except on shutdown - we just put it on 570 * the btree_cache_freed list and reuse it later. This simplifies the 571 * code, and it doesn't cost us much memory as the memory usage is 572 * dominated by buffers that hold the actual btree node data and those 573 * can be freed - and the number of struct btrees allocated is 574 * effectively bounded. 575 * 576 * btree_cache_freeable effectively is a small cache - we use it because 577 * high order page allocations can be rather expensive, and it's quite 578 * common to delete and allocate btree nodes in quick succession. It 579 * should never grow past ~2-3 nodes in practice. 580 */ 581 struct list_head btree_cache; 582 struct list_head btree_cache_freeable; 583 struct list_head btree_cache_freed; 584 585 /* Number of elements in btree_cache + btree_cache_freeable lists */ 586 unsigned int btree_cache_used; 587 588 /* 589 * If we need to allocate memory for a new btree node and that 590 * allocation fails, we can cannibalize another node in the btree cache 591 * to satisfy the allocation - lock to guarantee only one thread does 592 * this at a time: 593 */ 594 wait_queue_head_t btree_cache_wait; 595 struct task_struct *btree_cache_alloc_lock; 596 spinlock_t btree_cannibalize_lock; 597 598 /* 599 * When we free a btree node, we increment the gen of the bucket the 600 * node is in - but we can't rewrite the prios and gens until we 601 * finished whatever it is we were doing, otherwise after a crash the 602 * btree node would be freed but for say a split, we might not have the 603 * pointers to the new nodes inserted into the btree yet. 604 * 605 * This is a refcount that blocks prio_write() until the new keys are 606 * written. 607 */ 608 atomic_t prio_blocked; 609 wait_queue_head_t bucket_wait; 610 611 /* 612 * For any bio we don't skip we subtract the number of sectors from 613 * rescale; when it hits 0 we rescale all the bucket priorities. 614 */ 615 atomic_t rescale; 616 /* 617 * used for GC, identify if any front side I/Os is inflight 618 */ 619 atomic_t search_inflight; 620 /* 621 * When we invalidate buckets, we use both the priority and the amount 622 * of good data to determine which buckets to reuse first - to weight 623 * those together consistently we keep track of the smallest nonzero 624 * priority of any bucket. 625 */ 626 uint16_t min_prio; 627 628 /* 629 * max(gen - last_gc) for all buckets. When it gets too big we have to 630 * gc to keep gens from wrapping around. 631 */ 632 uint8_t need_gc; 633 struct gc_stat gc_stats; 634 size_t nbuckets; 635 size_t avail_nbuckets; 636 637 struct task_struct *gc_thread; 638 /* Where in the btree gc currently is */ 639 struct bkey gc_done; 640 641 /* 642 * For automatical garbage collection after writeback completed, this 643 * varialbe is used as bit fields, 644 * - 0000 0001b (BCH_ENABLE_AUTO_GC): enable gc after writeback 645 * - 0000 0010b (BCH_DO_AUTO_GC): do gc after writeback 646 * This is an optimization for following write request after writeback 647 * finished, but read hit rate dropped due to clean data on cache is 648 * discarded. Unless user explicitly sets it via sysfs, it won't be 649 * enabled. 650 */ 651 #define BCH_ENABLE_AUTO_GC 1 652 #define BCH_DO_AUTO_GC 2 653 uint8_t gc_after_writeback; 654 655 /* 656 * The allocation code needs gc_mark in struct bucket to be correct, but 657 * it's not while a gc is in progress. Protected by bucket_lock. 658 */ 659 int gc_mark_valid; 660 661 /* Counts how many sectors bio_insert has added to the cache */ 662 atomic_t sectors_to_gc; 663 wait_queue_head_t gc_wait; 664 665 struct keybuf moving_gc_keys; 666 /* Number of moving GC bios in flight */ 667 struct semaphore moving_in_flight; 668 669 struct workqueue_struct *moving_gc_wq; 670 671 struct btree *root; 672 673 #ifdef CONFIG_BCACHE_DEBUG 674 struct btree *verify_data; 675 struct bset *verify_ondisk; 676 struct mutex verify_lock; 677 #endif 678 679 uint8_t set_uuid[16]; 680 unsigned int nr_uuids; 681 struct uuid_entry *uuids; 682 BKEY_PADDED(uuid_bucket); 683 struct closure uuid_write; 684 struct semaphore uuid_write_mutex; 685 686 /* 687 * A btree node on disk could have too many bsets for an iterator to fit 688 * on the stack - have to dynamically allocate them. 689 * bch_cache_set_alloc() will make sure the pool can allocate iterators 690 * equipped with enough room that can host 691 * (sb.bucket_size / sb.block_size) 692 * btree_iter_sets, which is more than static MAX_BSETS. 693 */ 694 mempool_t fill_iter; 695 696 struct bset_sort_state sort; 697 698 /* List of buckets we're currently writing data to */ 699 struct list_head data_buckets; 700 spinlock_t data_bucket_lock; 701 702 struct journal journal; 703 704 #define CONGESTED_MAX 1024 705 unsigned int congested_last_us; 706 atomic_t congested; 707 708 /* The rest of this all shows up in sysfs */ 709 unsigned int congested_read_threshold_us; 710 unsigned int congested_write_threshold_us; 711 712 struct time_stats btree_gc_time; 713 struct time_stats btree_split_time; 714 struct time_stats btree_read_time; 715 716 atomic_long_t cache_read_races; 717 atomic_long_t writeback_keys_done; 718 atomic_long_t writeback_keys_failed; 719 720 atomic_long_t reclaim; 721 atomic_long_t reclaimed_journal_buckets; 722 atomic_long_t flush_write; 723 724 enum { 725 ON_ERROR_UNREGISTER, 726 ON_ERROR_PANIC, 727 } on_error; 728 #define DEFAULT_IO_ERROR_LIMIT 8 729 unsigned int error_limit; 730 unsigned int error_decay; 731 732 unsigned short journal_delay_ms; 733 bool expensive_debug_checks; 734 unsigned int verify:1; 735 unsigned int key_merging_disabled:1; 736 unsigned int gc_always_rewrite:1; 737 unsigned int shrinker_disabled:1; 738 unsigned int copy_gc_enabled:1; 739 unsigned int idle_max_writeback_rate_enabled:1; 740 741 #define BUCKET_HASH_BITS 12 742 struct hlist_head bucket_hash[1 << BUCKET_HASH_BITS]; 743 }; 744 745 struct bbio { 746 unsigned int submit_time_us; 747 union { 748 struct bkey key; 749 uint64_t _pad[3]; 750 /* 751 * We only need pad = 3 here because we only ever carry around a 752 * single pointer - i.e. the pointer we're doing io to/from. 753 */ 754 }; 755 struct bio bio; 756 }; 757 758 #define BTREE_PRIO USHRT_MAX 759 #define INITIAL_PRIO 32768U 760 761 #define btree_bytes(c) ((c)->btree_pages * PAGE_SIZE) 762 #define btree_blocks(b) \ 763 ((unsigned int) (KEY_SIZE(&b->key) >> (b)->c->block_bits)) 764 765 #define btree_default_blocks(c) \ 766 ((unsigned int) ((PAGE_SECTORS * (c)->btree_pages) >> (c)->block_bits)) 767 768 #define bucket_bytes(ca) ((ca)->sb.bucket_size << 9) 769 #define block_bytes(ca) ((ca)->sb.block_size << 9) 770 meta_bucket_pages(struct cache_sb * sb)771 static inline unsigned int meta_bucket_pages(struct cache_sb *sb) 772 { 773 unsigned int n, max_pages; 774 775 max_pages = min_t(unsigned int, 776 __rounddown_pow_of_two(USHRT_MAX) / PAGE_SECTORS, 777 MAX_ORDER_NR_PAGES); 778 779 n = sb->bucket_size / PAGE_SECTORS; 780 if (n > max_pages) 781 n = max_pages; 782 783 return n; 784 } 785 meta_bucket_bytes(struct cache_sb * sb)786 static inline unsigned int meta_bucket_bytes(struct cache_sb *sb) 787 { 788 return meta_bucket_pages(sb) << PAGE_SHIFT; 789 } 790 791 #define prios_per_bucket(ca) \ 792 ((meta_bucket_bytes(&(ca)->sb) - sizeof(struct prio_set)) / \ 793 sizeof(struct bucket_disk)) 794 795 #define prio_buckets(ca) \ 796 DIV_ROUND_UP((size_t) (ca)->sb.nbuckets, prios_per_bucket(ca)) 797 sector_to_bucket(struct cache_set * c,sector_t s)798 static inline size_t sector_to_bucket(struct cache_set *c, sector_t s) 799 { 800 return s >> c->bucket_bits; 801 } 802 bucket_to_sector(struct cache_set * c,size_t b)803 static inline sector_t bucket_to_sector(struct cache_set *c, size_t b) 804 { 805 return ((sector_t) b) << c->bucket_bits; 806 } 807 bucket_remainder(struct cache_set * c,sector_t s)808 static inline sector_t bucket_remainder(struct cache_set *c, sector_t s) 809 { 810 return s & (c->cache->sb.bucket_size - 1); 811 } 812 PTR_BUCKET_NR(struct cache_set * c,const struct bkey * k,unsigned int ptr)813 static inline size_t PTR_BUCKET_NR(struct cache_set *c, 814 const struct bkey *k, 815 unsigned int ptr) 816 { 817 return sector_to_bucket(c, PTR_OFFSET(k, ptr)); 818 } 819 PTR_BUCKET(struct cache_set * c,const struct bkey * k,unsigned int ptr)820 static inline struct bucket *PTR_BUCKET(struct cache_set *c, 821 const struct bkey *k, 822 unsigned int ptr) 823 { 824 return c->cache->buckets + PTR_BUCKET_NR(c, k, ptr); 825 } 826 gen_after(uint8_t a,uint8_t b)827 static inline uint8_t gen_after(uint8_t a, uint8_t b) 828 { 829 uint8_t r = a - b; 830 831 return r > 128U ? 0 : r; 832 } 833 ptr_stale(struct cache_set * c,const struct bkey * k,unsigned int i)834 static inline uint8_t ptr_stale(struct cache_set *c, const struct bkey *k, 835 unsigned int i) 836 { 837 return gen_after(PTR_BUCKET(c, k, i)->gen, PTR_GEN(k, i)); 838 } 839 ptr_available(struct cache_set * c,const struct bkey * k,unsigned int i)840 static inline bool ptr_available(struct cache_set *c, const struct bkey *k, 841 unsigned int i) 842 { 843 return (PTR_DEV(k, i) < MAX_CACHES_PER_SET) && c->cache; 844 } 845 846 /* Btree key macros */ 847 848 /* 849 * This is used for various on disk data structures - cache_sb, prio_set, bset, 850 * jset: The checksum is _always_ the first 8 bytes of these structs 851 */ 852 #define csum_set(i) \ 853 bch_crc64(((void *) (i)) + sizeof(uint64_t), \ 854 ((void *) bset_bkey_last(i)) - \ 855 (((void *) (i)) + sizeof(uint64_t))) 856 857 /* Error handling macros */ 858 859 #define btree_bug(b, ...) \ 860 do { \ 861 if (bch_cache_set_error((b)->c, __VA_ARGS__)) \ 862 dump_stack(); \ 863 } while (0) 864 865 #define cache_bug(c, ...) \ 866 do { \ 867 if (bch_cache_set_error(c, __VA_ARGS__)) \ 868 dump_stack(); \ 869 } while (0) 870 871 #define btree_bug_on(cond, b, ...) \ 872 do { \ 873 if (cond) \ 874 btree_bug(b, __VA_ARGS__); \ 875 } while (0) 876 877 #define cache_bug_on(cond, c, ...) \ 878 do { \ 879 if (cond) \ 880 cache_bug(c, __VA_ARGS__); \ 881 } while (0) 882 883 #define cache_set_err_on(cond, c, ...) \ 884 do { \ 885 if (cond) \ 886 bch_cache_set_error(c, __VA_ARGS__); \ 887 } while (0) 888 889 /* Looping macros */ 890 891 #define for_each_bucket(b, ca) \ 892 for (b = (ca)->buckets + (ca)->sb.first_bucket; \ 893 b < (ca)->buckets + (ca)->sb.nbuckets; b++) 894 cached_dev_put(struct cached_dev * dc)895 static inline void cached_dev_put(struct cached_dev *dc) 896 { 897 if (refcount_dec_and_test(&dc->count)) 898 schedule_work(&dc->detach); 899 } 900 cached_dev_get(struct cached_dev * dc)901 static inline bool cached_dev_get(struct cached_dev *dc) 902 { 903 if (!refcount_inc_not_zero(&dc->count)) 904 return false; 905 906 /* Paired with the mb in cached_dev_attach */ 907 smp_mb__after_atomic(); 908 return true; 909 } 910 911 /* 912 * bucket_gc_gen() returns the difference between the bucket's current gen and 913 * the oldest gen of any pointer into that bucket in the btree (last_gc). 914 */ 915 bucket_gc_gen(struct bucket * b)916 static inline uint8_t bucket_gc_gen(struct bucket *b) 917 { 918 return b->gen - b->last_gc; 919 } 920 921 #define BUCKET_GC_GEN_MAX 96U 922 923 #define kobj_attribute_write(n, fn) \ 924 static struct kobj_attribute ksysfs_##n = __ATTR(n, 0200, NULL, fn) 925 926 #define kobj_attribute_rw(n, show, store) \ 927 static struct kobj_attribute ksysfs_##n = \ 928 __ATTR(n, 0600, show, store) 929 wake_up_allocators(struct cache_set * c)930 static inline void wake_up_allocators(struct cache_set *c) 931 { 932 struct cache *ca = c->cache; 933 934 wake_up_process(ca->alloc_thread); 935 } 936 closure_bio_submit(struct cache_set * c,struct bio * bio,struct closure * cl)937 static inline void closure_bio_submit(struct cache_set *c, 938 struct bio *bio, 939 struct closure *cl) 940 { 941 closure_get(cl); 942 if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &c->flags))) { 943 bio->bi_status = BLK_STS_IOERR; 944 bio_endio(bio); 945 return; 946 } 947 submit_bio_noacct(bio); 948 } 949 950 /* 951 * Prevent the kthread exits directly, and make sure when kthread_stop() 952 * is called to stop a kthread, it is still alive. If a kthread might be 953 * stopped by CACHE_SET_IO_DISABLE bit set, wait_for_kthread_stop() is 954 * necessary before the kthread returns. 955 */ wait_for_kthread_stop(void)956 static inline void wait_for_kthread_stop(void) 957 { 958 while (!kthread_should_stop()) { 959 set_current_state(TASK_INTERRUPTIBLE); 960 schedule(); 961 } 962 } 963 964 /* Forward declarations */ 965 966 void bch_count_backing_io_errors(struct cached_dev *dc, struct bio *bio); 967 void bch_count_io_errors(struct cache *ca, blk_status_t error, 968 int is_read, const char *m); 969 void bch_bbio_count_io_errors(struct cache_set *c, struct bio *bio, 970 blk_status_t error, const char *m); 971 void bch_bbio_endio(struct cache_set *c, struct bio *bio, 972 blk_status_t error, const char *m); 973 void bch_bbio_free(struct bio *bio, struct cache_set *c); 974 struct bio *bch_bbio_alloc(struct cache_set *c); 975 976 void __bch_submit_bbio(struct bio *bio, struct cache_set *c); 977 void bch_submit_bbio(struct bio *bio, struct cache_set *c, 978 struct bkey *k, unsigned int ptr); 979 980 uint8_t bch_inc_gen(struct cache *ca, struct bucket *b); 981 void bch_rescale_priorities(struct cache_set *c, int sectors); 982 983 bool bch_can_invalidate_bucket(struct cache *ca, struct bucket *b); 984 void __bch_invalidate_one_bucket(struct cache *ca, struct bucket *b); 985 986 void __bch_bucket_free(struct cache *ca, struct bucket *b); 987 void bch_bucket_free(struct cache_set *c, struct bkey *k); 988 989 long bch_bucket_alloc(struct cache *ca, unsigned int reserve, bool wait); 990 int __bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve, 991 struct bkey *k, bool wait); 992 int bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve, 993 struct bkey *k, bool wait); 994 bool bch_alloc_sectors(struct cache_set *c, struct bkey *k, 995 unsigned int sectors, unsigned int write_point, 996 unsigned int write_prio, bool wait); 997 bool bch_cached_dev_error(struct cached_dev *dc); 998 999 __printf(2, 3) 1000 bool bch_cache_set_error(struct cache_set *c, const char *fmt, ...); 1001 1002 int bch_prio_write(struct cache *ca, bool wait); 1003 void bch_write_bdev_super(struct cached_dev *dc, struct closure *parent); 1004 1005 extern struct workqueue_struct *bcache_wq; 1006 extern struct workqueue_struct *bch_journal_wq; 1007 extern struct workqueue_struct *bch_flush_wq; 1008 extern struct mutex bch_register_lock; 1009 extern struct list_head bch_cache_sets; 1010 1011 extern const struct kobj_type bch_cached_dev_ktype; 1012 extern const struct kobj_type bch_flash_dev_ktype; 1013 extern const struct kobj_type bch_cache_set_ktype; 1014 extern const struct kobj_type bch_cache_set_internal_ktype; 1015 extern const struct kobj_type bch_cache_ktype; 1016 1017 void bch_cached_dev_release(struct kobject *kobj); 1018 void bch_flash_dev_release(struct kobject *kobj); 1019 void bch_cache_set_release(struct kobject *kobj); 1020 void bch_cache_release(struct kobject *kobj); 1021 1022 int bch_uuid_write(struct cache_set *c); 1023 void bcache_write_super(struct cache_set *c); 1024 1025 int bch_flash_dev_create(struct cache_set *c, uint64_t size); 1026 1027 int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c, 1028 uint8_t *set_uuid); 1029 void bch_cached_dev_detach(struct cached_dev *dc); 1030 int bch_cached_dev_run(struct cached_dev *dc); 1031 void bcache_device_stop(struct bcache_device *d); 1032 1033 void bch_cache_set_unregister(struct cache_set *c); 1034 void bch_cache_set_stop(struct cache_set *c); 1035 1036 struct cache_set *bch_cache_set_alloc(struct cache_sb *sb); 1037 void bch_btree_cache_free(struct cache_set *c); 1038 int bch_btree_cache_alloc(struct cache_set *c); 1039 void bch_moving_init_cache_set(struct cache_set *c); 1040 int bch_open_buckets_alloc(struct cache_set *c); 1041 void bch_open_buckets_free(struct cache_set *c); 1042 1043 int bch_cache_allocator_start(struct cache *ca); 1044 1045 void bch_debug_exit(void); 1046 void bch_debug_init(void); 1047 void bch_request_exit(void); 1048 int bch_request_init(void); 1049 void bch_btree_exit(void); 1050 int bch_btree_init(void); 1051 1052 #endif /* _BCACHE_H */ 1053