diff options
Diffstat (limited to 'drivers/md/bcache')
32 files changed, 18453 insertions, 0 deletions
diff --git a/drivers/md/bcache/Kconfig b/drivers/md/bcache/Kconfig new file mode 100644 index 000000000..d1ca4d059 --- /dev/null +++ b/drivers/md/bcache/Kconfig @@ -0,0 +1,37 @@ +# SPDX-License-Identifier: GPL-2.0-only + +config BCACHE + tristate "Block device as cache" + select CRC64 + help + Allows a block device to be used as cache for other devices; uses + a btree for indexing and the layout is optimized for SSDs. + + See Documentation/admin-guide/bcache.rst for details. + +config BCACHE_DEBUG + bool "Bcache debugging" + depends on BCACHE + help + Don't select this option unless you're a developer + + Enables extra debugging tools, allows expensive runtime checks to be + turned on. + +config BCACHE_CLOSURES_DEBUG + bool "Debug closures" + depends on BCACHE + select DEBUG_FS + help + Keeps all active closures in a linked list and provides a debugfs + interface to list them, which makes it possible to see asynchronous + operations that get stuck. + +config BCACHE_ASYNC_REGISTRATION + bool "Asynchronous device registration (EXPERIMENTAL)" + depends on BCACHE + help + Add a sysfs file /sys/fs/bcache/register_async. Writing registering + device path into this file will returns immediately and the real + registration work is handled in kernel work queue in asynchronous + way. diff --git a/drivers/md/bcache/Makefile b/drivers/md/bcache/Makefile new file mode 100644 index 000000000..5b87e5967 --- /dev/null +++ b/drivers/md/bcache/Makefile @@ -0,0 +1,7 @@ +# SPDX-License-Identifier: GPL-2.0 + +obj-$(CONFIG_BCACHE) += bcache.o + +bcache-y := alloc.o bset.o btree.o closure.o debug.o extents.o\ + io.o journal.o movinggc.o request.o stats.o super.o sysfs.o trace.o\ + util.o writeback.o features.o diff --git a/drivers/md/bcache/alloc.c b/drivers/md/bcache/alloc.c new file mode 100644 index 000000000..8c371d5ee --- /dev/null +++ b/drivers/md/bcache/alloc.c @@ -0,0 +1,737 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Primary bucket allocation code + * + * Copyright 2012 Google, Inc. + * + * Allocation in bcache is done in terms of buckets: + * + * Each bucket has associated an 8 bit gen; this gen corresponds to the gen in + * btree pointers - they must match for the pointer to be considered valid. + * + * Thus (assuming a bucket has no dirty data or metadata in it) we can reuse a + * bucket simply by incrementing its gen. + * + * The gens (along with the priorities; it's really the gens are important but + * the code is named as if it's the priorities) are written in an arbitrary list + * of buckets on disk, with a pointer to them in the journal header. + * + * When we invalidate a bucket, we have to write its new gen to disk and wait + * for that write to complete before we use it - otherwise after a crash we + * could have pointers that appeared to be good but pointed to data that had + * been overwritten. + * + * Since the gens and priorities are all stored contiguously on disk, we can + * batch this up: We fill up the free_inc list with freshly invalidated buckets, + * call prio_write(), and when prio_write() finishes we pull buckets off the + * free_inc list and optionally discard them. + * + * free_inc isn't the only freelist - if it was, we'd often to sleep while + * priorities and gens were being written before we could allocate. c->free is a + * smaller freelist, and buckets on that list are always ready to be used. + * + * If we've got discards enabled, that happens when a bucket moves from the + * free_inc list to the free list. + * + * There is another freelist, because sometimes we have buckets that we know + * have nothing pointing into them - these we can reuse without waiting for + * priorities to be rewritten. These come from freed btree nodes and buckets + * that garbage collection discovered no longer had valid keys pointing into + * them (because they were overwritten). That's the unused list - buckets on the + * unused list move to the free list, optionally being discarded in the process. + * + * It's also important to ensure that gens don't wrap around - with respect to + * either the oldest gen in the btree or the gen on disk. This is quite + * difficult to do in practice, but we explicitly guard against it anyways - if + * a bucket is in danger of wrapping around we simply skip invalidating it that + * time around, and we garbage collect or rewrite the priorities sooner than we + * would have otherwise. + * + * bch_bucket_alloc() allocates a single bucket from a specific cache. + * + * bch_bucket_alloc_set() allocates one bucket from different caches + * out of a cache set. + * + * free_some_buckets() drives all the processes described above. It's called + * from bch_bucket_alloc() and a few other places that need to make sure free + * buckets are ready. + * + * invalidate_buckets_(lru|fifo)() find buckets that are available to be + * invalidated, and then invalidate them and stick them on the free_inc list - + * in either lru or fifo order. + */ + +#include "bcache.h" +#include "btree.h" + +#include <linux/blkdev.h> +#include <linux/kthread.h> +#include <linux/random.h> +#include <trace/events/bcache.h> + +#define MAX_OPEN_BUCKETS 128 + +/* Bucket heap / gen */ + +uint8_t bch_inc_gen(struct cache *ca, struct bucket *b) +{ + uint8_t ret = ++b->gen; + + ca->set->need_gc = max(ca->set->need_gc, bucket_gc_gen(b)); + WARN_ON_ONCE(ca->set->need_gc > BUCKET_GC_GEN_MAX); + + return ret; +} + +void bch_rescale_priorities(struct cache_set *c, int sectors) +{ + struct cache *ca; + struct bucket *b; + unsigned long next = c->nbuckets * c->cache->sb.bucket_size / 1024; + int r; + + atomic_sub(sectors, &c->rescale); + + do { + r = atomic_read(&c->rescale); + + if (r >= 0) + return; + } while (atomic_cmpxchg(&c->rescale, r, r + next) != r); + + mutex_lock(&c->bucket_lock); + + c->min_prio = USHRT_MAX; + + ca = c->cache; + for_each_bucket(b, ca) + if (b->prio && + b->prio != BTREE_PRIO && + !atomic_read(&b->pin)) { + b->prio--; + c->min_prio = min(c->min_prio, b->prio); + } + + mutex_unlock(&c->bucket_lock); +} + +/* + * Background allocation thread: scans for buckets to be invalidated, + * invalidates them, rewrites prios/gens (marking them as invalidated on disk), + * then optionally issues discard commands to the newly free buckets, then puts + * them on the various freelists. + */ + +static inline bool can_inc_bucket_gen(struct bucket *b) +{ + return bucket_gc_gen(b) < BUCKET_GC_GEN_MAX; +} + +bool bch_can_invalidate_bucket(struct cache *ca, struct bucket *b) +{ + BUG_ON(!ca->set->gc_mark_valid); + + return (!GC_MARK(b) || + GC_MARK(b) == GC_MARK_RECLAIMABLE) && + !atomic_read(&b->pin) && + can_inc_bucket_gen(b); +} + +void __bch_invalidate_one_bucket(struct cache *ca, struct bucket *b) +{ + lockdep_assert_held(&ca->set->bucket_lock); + BUG_ON(GC_MARK(b) && GC_MARK(b) != GC_MARK_RECLAIMABLE); + + if (GC_SECTORS_USED(b)) + trace_bcache_invalidate(ca, b - ca->buckets); + + bch_inc_gen(ca, b); + b->prio = INITIAL_PRIO; + atomic_inc(&b->pin); +} + +static void bch_invalidate_one_bucket(struct cache *ca, struct bucket *b) +{ + __bch_invalidate_one_bucket(ca, b); + + fifo_push(&ca->free_inc, b - ca->buckets); +} + +/* + * Determines what order we're going to reuse buckets, smallest bucket_prio() + * first: we also take into account the number of sectors of live data in that + * bucket, and in order for that multiply to make sense we have to scale bucket + * + * Thus, we scale the bucket priorities so that the bucket with the smallest + * prio is worth 1/8th of what INITIAL_PRIO is worth. + */ + +#define bucket_prio(b) \ +({ \ + unsigned int min_prio = (INITIAL_PRIO - ca->set->min_prio) / 8; \ + \ + (b->prio - ca->set->min_prio + min_prio) * GC_SECTORS_USED(b); \ +}) + +#define bucket_max_cmp(l, r) (bucket_prio(l) < bucket_prio(r)) +#define bucket_min_cmp(l, r) (bucket_prio(l) > bucket_prio(r)) + +static void invalidate_buckets_lru(struct cache *ca) +{ + struct bucket *b; + ssize_t i; + + ca->heap.used = 0; + + for_each_bucket(b, ca) { + if (!bch_can_invalidate_bucket(ca, b)) + continue; + + if (!heap_full(&ca->heap)) + heap_add(&ca->heap, b, bucket_max_cmp); + else if (bucket_max_cmp(b, heap_peek(&ca->heap))) { + ca->heap.data[0] = b; + heap_sift(&ca->heap, 0, bucket_max_cmp); + } + } + + for (i = ca->heap.used / 2 - 1; i >= 0; --i) + heap_sift(&ca->heap, i, bucket_min_cmp); + + while (!fifo_full(&ca->free_inc)) { + if (!heap_pop(&ca->heap, b, bucket_min_cmp)) { + /* + * We don't want to be calling invalidate_buckets() + * multiple times when it can't do anything + */ + ca->invalidate_needs_gc = 1; + wake_up_gc(ca->set); + return; + } + + bch_invalidate_one_bucket(ca, b); + } +} + +static void invalidate_buckets_fifo(struct cache *ca) +{ + struct bucket *b; + size_t checked = 0; + + while (!fifo_full(&ca->free_inc)) { + if (ca->fifo_last_bucket < ca->sb.first_bucket || + ca->fifo_last_bucket >= ca->sb.nbuckets) + ca->fifo_last_bucket = ca->sb.first_bucket; + + b = ca->buckets + ca->fifo_last_bucket++; + + if (bch_can_invalidate_bucket(ca, b)) + bch_invalidate_one_bucket(ca, b); + + if (++checked >= ca->sb.nbuckets) { + ca->invalidate_needs_gc = 1; + wake_up_gc(ca->set); + return; + } + } +} + +static void invalidate_buckets_random(struct cache *ca) +{ + struct bucket *b; + size_t checked = 0; + + while (!fifo_full(&ca->free_inc)) { + size_t n; + + get_random_bytes(&n, sizeof(n)); + + n %= (size_t) (ca->sb.nbuckets - ca->sb.first_bucket); + n += ca->sb.first_bucket; + + b = ca->buckets + n; + + if (bch_can_invalidate_bucket(ca, b)) + bch_invalidate_one_bucket(ca, b); + + if (++checked >= ca->sb.nbuckets / 2) { + ca->invalidate_needs_gc = 1; + wake_up_gc(ca->set); + return; + } + } +} + +static void invalidate_buckets(struct cache *ca) +{ + BUG_ON(ca->invalidate_needs_gc); + + switch (CACHE_REPLACEMENT(&ca->sb)) { + case CACHE_REPLACEMENT_LRU: + invalidate_buckets_lru(ca); + break; + case CACHE_REPLACEMENT_FIFO: + invalidate_buckets_fifo(ca); + break; + case CACHE_REPLACEMENT_RANDOM: + invalidate_buckets_random(ca); + break; + } +} + +#define allocator_wait(ca, cond) \ +do { \ + while (1) { \ + set_current_state(TASK_INTERRUPTIBLE); \ + if (cond) \ + break; \ + \ + mutex_unlock(&(ca)->set->bucket_lock); \ + if (kthread_should_stop() || \ + test_bit(CACHE_SET_IO_DISABLE, &ca->set->flags)) { \ + set_current_state(TASK_RUNNING); \ + goto out; \ + } \ + \ + schedule(); \ + mutex_lock(&(ca)->set->bucket_lock); \ + } \ + __set_current_state(TASK_RUNNING); \ +} while (0) + +static int bch_allocator_push(struct cache *ca, long bucket) +{ + unsigned int i; + + /* Prios/gens are actually the most important reserve */ + if (fifo_push(&ca->free[RESERVE_PRIO], bucket)) + return true; + + for (i = 0; i < RESERVE_NR; i++) + if (fifo_push(&ca->free[i], bucket)) + return true; + + return false; +} + +static int bch_allocator_thread(void *arg) +{ + struct cache *ca = arg; + + mutex_lock(&ca->set->bucket_lock); + + while (1) { + /* + * First, we pull buckets off of the unused and free_inc lists, + * possibly issue discards to them, then we add the bucket to + * the free list: + */ + while (1) { + long bucket; + + if (!fifo_pop(&ca->free_inc, bucket)) + break; + + if (ca->discard) { + mutex_unlock(&ca->set->bucket_lock); + blkdev_issue_discard(ca->bdev, + bucket_to_sector(ca->set, bucket), + ca->sb.bucket_size, GFP_KERNEL, 0); + mutex_lock(&ca->set->bucket_lock); + } + + allocator_wait(ca, bch_allocator_push(ca, bucket)); + wake_up(&ca->set->btree_cache_wait); + wake_up(&ca->set->bucket_wait); + } + + /* + * We've run out of free buckets, we need to find some buckets + * we can invalidate. First, invalidate them in memory and add + * them to the free_inc list: + */ + +retry_invalidate: + allocator_wait(ca, ca->set->gc_mark_valid && + !ca->invalidate_needs_gc); + invalidate_buckets(ca); + + /* + * Now, we write their new gens to disk so we can start writing + * new stuff to them: + */ + allocator_wait(ca, !atomic_read(&ca->set->prio_blocked)); + if (CACHE_SYNC(&ca->sb)) { + /* + * This could deadlock if an allocation with a btree + * node locked ever blocked - having the btree node + * locked would block garbage collection, but here we're + * waiting on garbage collection before we invalidate + * and free anything. + * + * But this should be safe since the btree code always + * uses btree_check_reserve() before allocating now, and + * if it fails it blocks without btree nodes locked. + */ + if (!fifo_full(&ca->free_inc)) + goto retry_invalidate; + + if (bch_prio_write(ca, false) < 0) { + ca->invalidate_needs_gc = 1; + wake_up_gc(ca->set); + } + } + } +out: + wait_for_kthread_stop(); + return 0; +} + +/* Allocation */ + +long bch_bucket_alloc(struct cache *ca, unsigned int reserve, bool wait) +{ + DEFINE_WAIT(w); + struct bucket *b; + long r; + + + /* No allocation if CACHE_SET_IO_DISABLE bit is set */ + if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &ca->set->flags))) + return -1; + + /* fastpath */ + if (fifo_pop(&ca->free[RESERVE_NONE], r) || + fifo_pop(&ca->free[reserve], r)) + goto out; + + if (!wait) { + trace_bcache_alloc_fail(ca, reserve); + return -1; + } + + do { + prepare_to_wait(&ca->set->bucket_wait, &w, + TASK_UNINTERRUPTIBLE); + + mutex_unlock(&ca->set->bucket_lock); + schedule(); + mutex_lock(&ca->set->bucket_lock); + } while (!fifo_pop(&ca->free[RESERVE_NONE], r) && + !fifo_pop(&ca->free[reserve], r)); + + finish_wait(&ca->set->bucket_wait, &w); +out: + if (ca->alloc_thread) + wake_up_process(ca->alloc_thread); + + trace_bcache_alloc(ca, reserve); + + if (expensive_debug_checks(ca->set)) { + size_t iter; + long i; + unsigned int j; + + for (iter = 0; iter < prio_buckets(ca) * 2; iter++) + BUG_ON(ca->prio_buckets[iter] == (uint64_t) r); + + for (j = 0; j < RESERVE_NR; j++) + fifo_for_each(i, &ca->free[j], iter) + BUG_ON(i == r); + fifo_for_each(i, &ca->free_inc, iter) + BUG_ON(i == r); + } + + b = ca->buckets + r; + + BUG_ON(atomic_read(&b->pin) != 1); + + SET_GC_SECTORS_USED(b, ca->sb.bucket_size); + + if (reserve <= RESERVE_PRIO) { + SET_GC_MARK(b, GC_MARK_METADATA); + SET_GC_MOVE(b, 0); + b->prio = BTREE_PRIO; + } else { + SET_GC_MARK(b, GC_MARK_RECLAIMABLE); + SET_GC_MOVE(b, 0); + b->prio = INITIAL_PRIO; + } + + if (ca->set->avail_nbuckets > 0) { + ca->set->avail_nbuckets--; + bch_update_bucket_in_use(ca->set, &ca->set->gc_stats); + } + + return r; +} + +void __bch_bucket_free(struct cache *ca, struct bucket *b) +{ + SET_GC_MARK(b, 0); + SET_GC_SECTORS_USED(b, 0); + + if (ca->set->avail_nbuckets < ca->set->nbuckets) { + ca->set->avail_nbuckets++; + bch_update_bucket_in_use(ca->set, &ca->set->gc_stats); + } +} + +void bch_bucket_free(struct cache_set *c, struct bkey *k) +{ + unsigned int i; + + for (i = 0; i < KEY_PTRS(k); i++) + __bch_bucket_free(PTR_CACHE(c, k, i), + PTR_BUCKET(c, k, i)); +} + +int __bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve, + struct bkey *k, bool wait) +{ + struct cache *ca; + long b; + + /* No allocation if CACHE_SET_IO_DISABLE bit is set */ + if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &c->flags))) + return -1; + + lockdep_assert_held(&c->bucket_lock); + + bkey_init(k); + + ca = c->cache; + b = bch_bucket_alloc(ca, reserve, wait); + if (b == -1) + goto err; + + k->ptr[0] = MAKE_PTR(ca->buckets[b].gen, + bucket_to_sector(c, b), + ca->sb.nr_this_dev); + + SET_KEY_PTRS(k, 1); + + return 0; +err: + bch_bucket_free(c, k); + bkey_put(c, k); + return -1; +} + +int bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve, + struct bkey *k, bool wait) +{ + int ret; + + mutex_lock(&c->bucket_lock); + ret = __bch_bucket_alloc_set(c, reserve, k, wait); + mutex_unlock(&c->bucket_lock); + return ret; +} + +/* Sector allocator */ + +struct open_bucket { + struct list_head list; + unsigned int last_write_point; + unsigned int sectors_free; + BKEY_PADDED(key); +}; + +/* + * We keep multiple buckets open for writes, and try to segregate different + * write streams for better cache utilization: first we try to segregate flash + * only volume write streams from cached devices, secondly we look for a bucket + * where the last write to it was sequential with the current write, and + * failing that we look for a bucket that was last used by the same task. + * + * The ideas is if you've got multiple tasks pulling data into the cache at the + * same time, you'll get better cache utilization if you try to segregate their + * data and preserve locality. + * + * For example, dirty sectors of flash only volume is not reclaimable, if their + * dirty sectors mixed with dirty sectors of cached device, such buckets will + * be marked as dirty and won't be reclaimed, though the dirty data of cached + * device have been written back to backend device. + * + * And say you've starting Firefox at the same time you're copying a + * bunch of files. Firefox will likely end up being fairly hot and stay in the + * cache awhile, but the data you copied might not be; if you wrote all that + * data to the same buckets it'd get invalidated at the same time. + * + * Both of those tasks will be doing fairly random IO so we can't rely on + * detecting sequential IO to segregate their data, but going off of the task + * should be a sane heuristic. + */ +static struct open_bucket *pick_data_bucket(struct cache_set *c, + const struct bkey *search, + unsigned int write_point, + struct bkey *alloc) +{ + struct open_bucket *ret, *ret_task = NULL; + + list_for_each_entry_reverse(ret, &c->data_buckets, list) + if (UUID_FLASH_ONLY(&c->uuids[KEY_INODE(&ret->key)]) != + UUID_FLASH_ONLY(&c->uuids[KEY_INODE(search)])) + continue; + else if (!bkey_cmp(&ret->key, search)) + goto found; + else if (ret->last_write_point == write_point) + ret_task = ret; + + ret = ret_task ?: list_first_entry(&c->data_buckets, + struct open_bucket, list); +found: + if (!ret->sectors_free && KEY_PTRS(alloc)) { + ret->sectors_free = c->cache->sb.bucket_size; + bkey_copy(&ret->key, alloc); + bkey_init(alloc); + } + + if (!ret->sectors_free) + ret = NULL; + + return ret; +} + +/* + * Allocates some space in the cache to write to, and k to point to the newly + * allocated space, and updates KEY_SIZE(k) and KEY_OFFSET(k) (to point to the + * end of the newly allocated space). + * + * May allocate fewer sectors than @sectors, KEY_SIZE(k) indicates how many + * sectors were actually allocated. + * + * If s->writeback is true, will not fail. + */ +bool bch_alloc_sectors(struct cache_set *c, + struct bkey *k, + unsigned int sectors, + unsigned int write_point, + unsigned int write_prio, + bool wait) +{ + struct open_bucket *b; + BKEY_PADDED(key) alloc; + unsigned int i; + + /* + * We might have to allocate a new bucket, which we can't do with a + * spinlock held. So if we have to allocate, we drop the lock, allocate + * and then retry. KEY_PTRS() indicates whether alloc points to + * allocated bucket(s). + */ + + bkey_init(&alloc.key); + spin_lock(&c->data_bucket_lock); + + while (!(b = pick_data_bucket(c, k, write_point, &alloc.key))) { + unsigned int watermark = write_prio + ? RESERVE_MOVINGGC + : RESERVE_NONE; + + spin_unlock(&c->data_bucket_lock); + + if (bch_bucket_alloc_set(c, watermark, &alloc.key, wait)) + return false; + + spin_lock(&c->data_bucket_lock); + } + + /* + * If we had to allocate, we might race and not need to allocate the + * second time we call pick_data_bucket(). If we allocated a bucket but + * didn't use it, drop the refcount bch_bucket_alloc_set() took: + */ + if (KEY_PTRS(&alloc.key)) + bkey_put(c, &alloc.key); + + for (i = 0; i < KEY_PTRS(&b->key); i++) + EBUG_ON(ptr_stale(c, &b->key, i)); + + /* Set up the pointer to the space we're allocating: */ + + for (i = 0; i < KEY_PTRS(&b->key); i++) + k->ptr[i] = b->key.ptr[i]; + + sectors = min(sectors, b->sectors_free); + + SET_KEY_OFFSET(k, KEY_OFFSET(k) + sectors); + SET_KEY_SIZE(k, sectors); + SET_KEY_PTRS(k, KEY_PTRS(&b->key)); + + /* + * Move b to the end of the lru, and keep track of what this bucket was + * last used for: + */ + list_move_tail(&b->list, &c->data_buckets); + bkey_copy_key(&b->key, k); + b->last_write_point = write_point; + + b->sectors_free -= sectors; + + for (i = 0; i < KEY_PTRS(&b->key); i++) { + SET_PTR_OFFSET(&b->key, i, PTR_OFFSET(&b->key, i) + sectors); + + atomic_long_add(sectors, + &PTR_CACHE(c, &b->key, i)->sectors_written); + } + + if (b->sectors_free < c->cache->sb.block_size) + b->sectors_free = 0; + + /* + * k takes refcounts on the buckets it points to until it's inserted + * into the btree, but if we're done with this bucket we just transfer + * get_data_bucket()'s refcount. + */ + if (b->sectors_free) + for (i = 0; i < KEY_PTRS(&b->key); i++) + atomic_inc(&PTR_BUCKET(c, &b->key, i)->pin); + + spin_unlock(&c->data_bucket_lock); + return true; +} + +/* Init */ + +void bch_open_buckets_free(struct cache_set *c) +{ + struct open_bucket *b; + + while (!list_empty(&c->data_buckets)) { + b = list_first_entry(&c->data_buckets, + struct open_bucket, list); + list_del(&b->list); + kfree(b); + } +} + +int bch_open_buckets_alloc(struct cache_set *c) +{ + int i; + + spin_lock_init(&c->data_bucket_lock); + + for (i = 0; i < MAX_OPEN_BUCKETS; i++) { + struct open_bucket *b = kzalloc(sizeof(*b), GFP_KERNEL); + + if (!b) + return -ENOMEM; + + list_add(&b->list, &c->data_buckets); + } + + return 0; +} + +int bch_cache_allocator_start(struct cache *ca) +{ + struct task_struct *k = kthread_run(bch_allocator_thread, + ca, "bcache_allocator"); + if (IS_ERR(k)) + return PTR_ERR(k); + + ca->alloc_thread = k; + return 0; +} diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h new file mode 100644 index 000000000..5804a8e6e --- /dev/null +++ b/drivers/md/bcache/bcache.h @@ -0,0 +1,1050 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _BCACHE_H +#define _BCACHE_H + +/* + * SOME HIGH LEVEL CODE DOCUMENTATION: + * + * Bcache mostly works with cache sets, cache devices, and backing devices. + * + * Support for multiple cache devices hasn't quite been finished off yet, but + * it's about 95% plumbed through. A cache set and its cache devices is sort of + * like a md raid array and its component devices. Most of the code doesn't care + * about individual cache devices, the main abstraction is the cache set. + * + * Multiple cache devices is intended to give us the ability to mirror dirty + * cached data and metadata, without mirroring clean cached data. + * + * Backing devices are different, in that they have a lifetime independent of a + * cache set. When you register a newly formatted backing device it'll come up + * in passthrough mode, and then you can attach and detach a backing device from + * a cache set at runtime - while it's mounted and in use. Detaching implicitly + * invalidates any cached data for that backing device. + * + * A cache set can have multiple (many) backing devices attached to it. + * + * There's also flash only volumes - this is the reason for the distinction + * between struct cached_dev and struct bcache_device. A flash only volume + * works much like a bcache device that has a backing device, except the + * "cached" data is always dirty. The end result is that we get thin + * provisioning with very little additional code. + * + * Flash only volumes work but they're not production ready because the moving + * garbage collector needs more work. More on that later. + * + * BUCKETS/ALLOCATION: + * + * Bcache is primarily designed for caching, which means that in normal + * operation all of our available space will be allocated. Thus, we need an + * efficient way of deleting things from the cache so we can write new things to + * it. + * + * To do this, we first divide the cache device up into buckets. A bucket is the + * unit of allocation; they're typically around 1 mb - anywhere from 128k to 2M+ + * works efficiently. + * + * Each bucket has a 16 bit priority, and an 8 bit generation associated with + * it. The gens and priorities for all the buckets are stored contiguously and + * packed on disk (in a linked list of buckets - aside from the superblock, all + * of bcache's metadata is stored in buckets). + * + * The priority is used to implement an LRU. We reset a bucket's priority when + * we allocate it or on cache it, and every so often we decrement the priority + * of each bucket. It could be used to implement something more sophisticated, + * if anyone ever gets around to it. + * + * The generation is used for invalidating buckets. Each pointer also has an 8 + * bit generation embedded in it; for a pointer to be considered valid, its gen + * must match the gen of the bucket it points into. Thus, to reuse a bucket all + * we have to do is increment its gen (and write its new gen to disk; we batch + * this up). + * + * Bcache is entirely COW - we never write twice to a bucket, even buckets that + * contain metadata (including btree nodes). + * + * THE BTREE: + * + * Bcache is in large part design around the btree. + * + * At a high level, the btree is just an index of key -> ptr tuples. + * + * Keys represent extents, and thus have a size field. Keys also have a variable + * number of pointers attached to them (potentially zero, which is handy for + * invalidating the cache). + * + * The key itself is an inode:offset pair. The inode number corresponds to a + * backing device or a flash only volume. The offset is the ending offset of the + * extent within the inode - not the starting offset; this makes lookups + * slightly more convenient. + * + * Pointers contain the cache device id, the offset on that device, and an 8 bit + * generation number. More on the gen later. + * + * Index lookups are not fully abstracted - cache lookups in particular are + * still somewhat mixed in with the btree code, but things are headed in that + * direction. + * + * Updates are fairly well abstracted, though. There are two different ways of + * updating the btree; insert and replace. + * + * BTREE_INSERT will just take a list of keys and insert them into the btree - + * overwriting (possibly only partially) any extents they overlap with. This is + * used to update the index after a write. + * + * BTREE_REPLACE is really cmpxchg(); it inserts a key into the btree iff it is + * overwriting a key that matches another given key. This is used for inserting + * data into the cache after a cache miss, and for background writeback, and for + * the moving garbage collector. + * + * There is no "delete" operation; deleting things from the index is + * accomplished by either by invalidating pointers (by incrementing a bucket's + * gen) or by inserting a key with 0 pointers - which will overwrite anything + * previously present at that location in the index. + * + * This means that there are always stale/invalid keys in the btree. They're + * filtered out by the code that iterates through a btree node, and removed when + * a btree node is rewritten. + * + * BTREE NODES: + * + * Our unit of allocation is a bucket, and we we can't arbitrarily allocate and + * free smaller than a bucket - so, that's how big our btree nodes are. + * + * (If buckets are really big we'll only use part of the bucket for a btree node + * - no less than 1/4th - but a bucket still contains no more than a single + * btree node. I'd actually like to change this, but for now we rely on the + * bucket's gen for deleting btree nodes when we rewrite/split a node.) + * + * Anyways, btree nodes are big - big enough to be inefficient with a textbook + * btree implementation. + * + * The way this is solved is that btree nodes are internally log structured; we + * can append new keys to an existing btree node without rewriting it. This + * means each set of keys we write is sorted, but the node is not. + * + * We maintain this log structure in memory - keeping 1Mb of keys sorted would + * be expensive, and we have to distinguish between the keys we have written and + * the keys we haven't. So to do a lookup in a btree node, we have to search + * each sorted set. But we do merge written sets together lazily, so the cost of + * these extra searches is quite low (normally most of the keys in a btree node + * will be in one big set, and then there'll be one or two sets that are much + * smaller). + * + * This log structure makes bcache's btree more of a hybrid between a + * conventional btree and a compacting data structure, with some of the + * advantages of both. + * + * GARBAGE COLLECTION: + * + * We can't just invalidate any bucket - it might contain dirty data or + * metadata. If it once contained dirty data, other writes might overwrite it + * later, leaving no valid pointers into that bucket in the index. + * + * Thus, the primary purpose of garbage collection is to find buckets to reuse. + * It also counts how much valid data it each bucket currently contains, so that + * allocation can reuse buckets sooner when they've been mostly overwritten. + * + * It also does some things that are really internal to the btree + * implementation. If a btree node contains pointers that are stale by more than + * some threshold, it rewrites the btree node to avoid the bucket's generation + * wrapping around. It also merges adjacent btree nodes if they're empty enough. + * + * THE JOURNAL: + * + * Bcache's journal is not necessary for consistency; we always strictly + * order metadata writes so that the btree and everything else is consistent on + * disk in the event of an unclean shutdown, and in fact bcache had writeback + * caching (with recovery from unclean shutdown) before journalling was + * implemented. + * + * Rather, the journal is purely a performance optimization; we can't complete a + * write until we've updated the index on disk, otherwise the cache would be + * inconsistent in the event of an unclean shutdown. This means that without the + * journal, on random write workloads we constantly have to update all the leaf + * nodes in the btree, and those writes will be mostly empty (appending at most + * a few keys each) - highly inefficient in terms of amount of metadata writes, + * and it puts more strain on the various btree resorting/compacting code. + * + * The journal is just a log of keys we've inserted; on startup we just reinsert + * all the keys in the open journal entries. That means that when we're updating + * a node in the btree, we can wait until a 4k block of keys fills up before + * writing them out. + * + * For simplicity, we only journal updates to leaf nodes; updates to parent + * nodes are rare enough (since our leaf nodes are huge) that it wasn't worth + * the complexity to deal with journalling them (in particular, journal replay) + * - updates to non leaf nodes just happen synchronously (see btree_split()). + */ + +#define pr_fmt(fmt) "bcache: %s() " fmt, __func__ + +#include <linux/bcache.h> +#include <linux/bio.h> +#include <linux/kobject.h> +#include <linux/list.h> +#include <linux/mutex.h> +#include <linux/rbtree.h> +#include <linux/rwsem.h> +#include <linux/refcount.h> +#include <linux/types.h> +#include <linux/workqueue.h> +#include <linux/kthread.h> + +#include "bset.h" +#include "util.h" +#include "closure.h" + +struct bucket { + atomic_t pin; + uint16_t prio; + uint8_t gen; + uint8_t last_gc; /* Most out of date gen in the btree */ + uint16_t gc_mark; /* Bitfield used by GC. See below for field */ +}; + +/* + * I'd use bitfields for these, but I don't trust the compiler not to screw me + * as multiple threads touch struct bucket without locking + */ + +BITMASK(GC_MARK, struct bucket, gc_mark, 0, 2); +#define GC_MARK_RECLAIMABLE 1 +#define GC_MARK_DIRTY 2 +#define GC_MARK_METADATA 3 +#define GC_SECTORS_USED_SIZE 13 +#define MAX_GC_SECTORS_USED (~(~0ULL << GC_SECTORS_USED_SIZE)) +BITMASK(GC_SECTORS_USED, struct bucket, gc_mark, 2, GC_SECTORS_USED_SIZE); +BITMASK(GC_MOVE, struct bucket, gc_mark, 15, 1); + +#include "journal.h" +#include "stats.h" +struct search; +struct btree; +struct keybuf; + +struct keybuf_key { + struct rb_node node; + BKEY_PADDED(key); + void *private; +}; + +struct keybuf { + struct bkey last_scanned; + spinlock_t lock; + + /* + * Beginning and end of range in rb tree - so that we can skip taking + * lock and checking the rb tree when we need to check for overlapping + * keys. + */ + struct bkey start; + struct bkey end; + + struct rb_root keys; + +#define KEYBUF_NR 500 + DECLARE_ARRAY_ALLOCATOR(struct keybuf_key, freelist, KEYBUF_NR); +}; + +struct bcache_device { + struct closure cl; + + struct kobject kobj; + + struct cache_set *c; + unsigned int id; +#define BCACHEDEVNAME_SIZE 12 + char name[BCACHEDEVNAME_SIZE]; + + struct gendisk *disk; + + unsigned long flags; +#define BCACHE_DEV_CLOSING 0 +#define BCACHE_DEV_DETACHING 1 +#define BCACHE_DEV_UNLINK_DONE 2 +#define BCACHE_DEV_WB_RUNNING 3 +#define BCACHE_DEV_RATE_DW_RUNNING 4 + int nr_stripes; +#define BCH_MIN_STRIPE_SZ ((4 << 20) >> SECTOR_SHIFT) + unsigned int stripe_size; + atomic_t *stripe_sectors_dirty; + unsigned long *full_dirty_stripes; + + struct bio_set bio_split; + + unsigned int data_csum:1; + + int (*cache_miss)(struct btree *b, struct search *s, + struct bio *bio, unsigned int sectors); + int (*ioctl)(struct bcache_device *d, fmode_t mode, + unsigned int cmd, unsigned long arg); +}; + +struct io { + /* Used to track sequential IO so it can be skipped */ + struct hlist_node hash; + struct list_head lru; + + unsigned long jiffies; + unsigned int sequential; + sector_t last; +}; + +enum stop_on_failure { + BCH_CACHED_DEV_STOP_AUTO = 0, + BCH_CACHED_DEV_STOP_ALWAYS, + BCH_CACHED_DEV_STOP_MODE_MAX, +}; + +struct cached_dev { + struct list_head list; + struct bcache_device disk; + struct block_device *bdev; + + struct cache_sb sb; + struct cache_sb_disk *sb_disk; + struct bio sb_bio; + struct bio_vec sb_bv[1]; + struct closure sb_write; + struct semaphore sb_write_mutex; + + /* Refcount on the cache set. Always nonzero when we're caching. */ + refcount_t count; + struct work_struct detach; + + /* + * Device might not be running if it's dirty and the cache set hasn't + * showed up yet. + */ + atomic_t running; + + /* + * Writes take a shared lock from start to finish; scanning for dirty + * data to refill the rb tree requires an exclusive lock. + */ + struct rw_semaphore writeback_lock; + + /* + * Nonzero, and writeback has a refcount (d->count), iff there is dirty + * data in the cache. Protected by writeback_lock; must have an + * shared lock to set and exclusive lock to clear. + */ + atomic_t has_dirty; + +#define BCH_CACHE_READA_ALL 0 +#define BCH_CACHE_READA_META_ONLY 1 + unsigned int cache_readahead_policy; + struct bch_ratelimit writeback_rate; + struct delayed_work writeback_rate_update; + + /* Limit number of writeback bios in flight */ + struct semaphore in_flight; + struct task_struct *writeback_thread; + struct workqueue_struct *writeback_write_wq; + + struct keybuf writeback_keys; + + struct task_struct *status_update_thread; + /* + * Order the write-half of writeback operations strongly in dispatch + * order. (Maintain LBA order; don't allow reads completing out of + * order to re-order the writes...) + */ + struct closure_waitlist writeback_ordering_wait; + atomic_t writeback_sequence_next; + + /* For tracking sequential IO */ +#define RECENT_IO_BITS 7 +#define RECENT_IO (1 << RECENT_IO_BITS) + struct io io[RECENT_IO]; + struct hlist_head io_hash[RECENT_IO + 1]; + struct list_head io_lru; + spinlock_t io_lock; + + struct cache_accounting accounting; + + /* The rest of this all shows up in sysfs */ + unsigned int sequential_cutoff; + unsigned int readahead; + + unsigned int io_disable:1; + unsigned int verify:1; + unsigned int bypass_torture_test:1; + + unsigned int partial_stripes_expensive:1; + unsigned int writeback_metadata:1; + unsigned int writeback_running:1; + unsigned char writeback_percent; + unsigned int writeback_delay; + + uint64_t writeback_rate_target; + int64_t writeback_rate_proportional; + int64_t writeback_rate_integral; + int64_t writeback_rate_integral_scaled; + int32_t writeback_rate_change; + + unsigned int writeback_rate_update_seconds; + unsigned int writeback_rate_i_term_inverse; + unsigned int writeback_rate_p_term_inverse; + unsigned int writeback_rate_minimum; + + enum stop_on_failure stop_when_cache_set_failed; +#define DEFAULT_CACHED_DEV_ERROR_LIMIT 64 + atomic_t io_errors; + unsigned int error_limit; + unsigned int offline_seconds; + + char backing_dev_name[BDEVNAME_SIZE]; +}; + +enum alloc_reserve { + RESERVE_BTREE, + RESERVE_PRIO, + RESERVE_MOVINGGC, + RESERVE_NONE, + RESERVE_NR, +}; + +struct cache { + struct cache_set *set; + struct cache_sb sb; + struct cache_sb_disk *sb_disk; + struct bio sb_bio; + struct bio_vec sb_bv[1]; + + struct kobject kobj; + struct block_device *bdev; + + struct task_struct *alloc_thread; + + struct closure prio; + struct prio_set *disk_buckets; + + /* + * When allocating new buckets, prio_write() gets first dibs - since we + * may not be allocate at all without writing priorities and gens. + * prio_last_buckets[] contains the last buckets we wrote priorities to + * (so gc can mark them as metadata), prio_buckets[] contains the + * buckets allocated for the next prio write. + */ + uint64_t *prio_buckets; + uint64_t *prio_last_buckets; + + /* + * free: Buckets that are ready to be used + * + * free_inc: Incoming buckets - these are buckets that currently have + * cached data in them, and we can't reuse them until after we write + * their new gen to disk. After prio_write() finishes writing the new + * gens/prios, they'll be moved to the free list (and possibly discarded + * in the process) + */ + DECLARE_FIFO(long, free)[RESERVE_NR]; + DECLARE_FIFO(long, free_inc); + + size_t fifo_last_bucket; + + /* Allocation stuff: */ + struct bucket *buckets; + + DECLARE_HEAP(struct bucket *, heap); + + /* + * If nonzero, we know we aren't going to find any buckets to invalidate + * until a gc finishes - otherwise we could pointlessly burn a ton of + * cpu + */ + unsigned int invalidate_needs_gc; + + bool discard; /* Get rid of? */ + + struct journal_device journal; + + /* The rest of this all shows up in sysfs */ +#define IO_ERROR_SHIFT 20 + atomic_t io_errors; + atomic_t io_count; + + atomic_long_t meta_sectors_written; + atomic_long_t btree_sectors_written; + atomic_long_t sectors_written; + + char cache_dev_name[BDEVNAME_SIZE]; +}; + +struct gc_stat { + size_t nodes; + size_t nodes_pre; + size_t key_bytes; + + size_t nkeys; + uint64_t data; /* sectors */ + unsigned int in_use; /* percent */ +}; + +/* + * Flag bits, for how the cache set is shutting down, and what phase it's at: + * + * CACHE_SET_UNREGISTERING means we're not just shutting down, we're detaching + * all the backing devices first (their cached data gets invalidated, and they + * won't automatically reattach). + * + * CACHE_SET_STOPPING always gets set first when we're closing down a cache set; + * we'll continue to run normally for awhile with CACHE_SET_STOPPING set (i.e. + * flushing dirty data). + * + * CACHE_SET_RUNNING means all cache devices have been registered and journal + * replay is complete. + * + * CACHE_SET_IO_DISABLE is set when bcache is stopping the whold cache set, all + * external and internal I/O should be denied when this flag is set. + * + */ +#define CACHE_SET_UNREGISTERING 0 +#define CACHE_SET_STOPPING 1 +#define CACHE_SET_RUNNING 2 +#define CACHE_SET_IO_DISABLE 3 + +struct cache_set { + struct closure cl; + + struct list_head list; + struct kobject kobj; + struct kobject internal; + struct dentry *debug; + struct cache_accounting accounting; + + unsigned long flags; + atomic_t idle_counter; + atomic_t at_max_writeback_rate; + + struct cache *cache; + + struct bcache_device **devices; + unsigned int devices_max_used; + atomic_t attached_dev_nr; + struct list_head cached_devs; + uint64_t cached_dev_sectors; + atomic_long_t flash_dev_dirty_sectors; + struct closure caching; + + struct closure sb_write; + struct semaphore sb_write_mutex; + + mempool_t search; + mempool_t bio_meta; + struct bio_set bio_split; + + /* For the btree cache */ + struct shrinker shrink; + + /* For the btree cache and anything allocation related */ + struct mutex bucket_lock; + + /* log2(bucket_size), in sectors */ + unsigned short bucket_bits; + + /* log2(block_size), in sectors */ + unsigned short block_bits; + + /* + * Default number of pages for a new btree node - may be less than a + * full bucket + */ + unsigned int btree_pages; + + /* + * Lists of struct btrees; lru is the list for structs that have memory + * allocated for actual btree node, freed is for structs that do not. + * + * We never free a struct btree, except on shutdown - we just put it on + * the btree_cache_freed list and reuse it later. This simplifies the + * code, and it doesn't cost us much memory as the memory usage is + * dominated by buffers that hold the actual btree node data and those + * can be freed - and the number of struct btrees allocated is + * effectively bounded. + * + * btree_cache_freeable effectively is a small cache - we use it because + * high order page allocations can be rather expensive, and it's quite + * common to delete and allocate btree nodes in quick succession. It + * should never grow past ~2-3 nodes in practice. + */ + struct list_head btree_cache; + struct list_head btree_cache_freeable; + struct list_head btree_cache_freed; + + /* Number of elements in btree_cache + btree_cache_freeable lists */ + unsigned int btree_cache_used; + + /* + * If we need to allocate memory for a new btree node and that + * allocation fails, we can cannibalize another node in the btree cache + * to satisfy the allocation - lock to guarantee only one thread does + * this at a time: + */ + wait_queue_head_t btree_cache_wait; + struct task_struct *btree_cache_alloc_lock; + spinlock_t btree_cannibalize_lock; + + /* + * When we free a btree node, we increment the gen of the bucket the + * node is in - but we can't rewrite the prios and gens until we + * finished whatever it is we were doing, otherwise after a crash the + * btree node would be freed but for say a split, we might not have the + * pointers to the new nodes inserted into the btree yet. + * + * This is a refcount that blocks prio_write() until the new keys are + * written. + */ + atomic_t prio_blocked; + wait_queue_head_t bucket_wait; + + /* + * For any bio we don't skip we subtract the number of sectors from + * rescale; when it hits 0 we rescale all the bucket priorities. + */ + atomic_t rescale; + /* + * used for GC, identify if any front side I/Os is inflight + */ + atomic_t search_inflight; + /* + * When we invalidate buckets, we use both the priority and the amount + * of good data to determine which buckets to reuse first - to weight + * those together consistently we keep track of the smallest nonzero + * priority of any bucket. + */ + uint16_t min_prio; + + /* + * max(gen - last_gc) for all buckets. When it gets too big we have to + * gc to keep gens from wrapping around. + */ + uint8_t need_gc; + struct gc_stat gc_stats; + size_t nbuckets; + size_t avail_nbuckets; + + struct task_struct *gc_thread; + /* Where in the btree gc currently is */ + struct bkey gc_done; + + /* + * For automatical garbage collection after writeback completed, this + * varialbe is used as bit fields, + * - 0000 0001b (BCH_ENABLE_AUTO_GC): enable gc after writeback + * - 0000 0010b (BCH_DO_AUTO_GC): do gc after writeback + * This is an optimization for following write request after writeback + * finished, but read hit rate dropped due to clean data on cache is + * discarded. Unless user explicitly sets it via sysfs, it won't be + * enabled. + */ +#define BCH_ENABLE_AUTO_GC 1 +#define BCH_DO_AUTO_GC 2 + uint8_t gc_after_writeback; + + /* + * The allocation code needs gc_mark in struct bucket to be correct, but + * it's not while a gc is in progress. Protected by bucket_lock. + */ + int gc_mark_valid; + + /* Counts how many sectors bio_insert has added to the cache */ + atomic_t sectors_to_gc; + wait_queue_head_t gc_wait; + + struct keybuf moving_gc_keys; + /* Number of moving GC bios in flight */ + struct semaphore moving_in_flight; + + struct workqueue_struct *moving_gc_wq; + + struct btree *root; + +#ifdef CONFIG_BCACHE_DEBUG + struct btree *verify_data; + struct bset *verify_ondisk; + struct mutex verify_lock; +#endif + + uint8_t set_uuid[16]; + unsigned int nr_uuids; + struct uuid_entry *uuids; + BKEY_PADDED(uuid_bucket); + struct closure uuid_write; + struct semaphore uuid_write_mutex; + + /* + * A btree node on disk could have too many bsets for an iterator to fit + * on the stack - have to dynamically allocate them. + * bch_cache_set_alloc() will make sure the pool can allocate iterators + * equipped with enough room that can host + * (sb.bucket_size / sb.block_size) + * btree_iter_sets, which is more than static MAX_BSETS. + */ + mempool_t fill_iter; + + struct bset_sort_state sort; + + /* List of buckets we're currently writing data to */ + struct list_head data_buckets; + spinlock_t data_bucket_lock; + + struct journal journal; + +#define CONGESTED_MAX 1024 + unsigned int congested_last_us; + atomic_t congested; + + /* The rest of this all shows up in sysfs */ + unsigned int congested_read_threshold_us; + unsigned int congested_write_threshold_us; + + struct time_stats btree_gc_time; + struct time_stats btree_split_time; + struct time_stats btree_read_time; + + atomic_long_t cache_read_races; + atomic_long_t writeback_keys_done; + atomic_long_t writeback_keys_failed; + + atomic_long_t reclaim; + atomic_long_t reclaimed_journal_buckets; + atomic_long_t flush_write; + + enum { + ON_ERROR_UNREGISTER, + ON_ERROR_PANIC, + } on_error; +#define DEFAULT_IO_ERROR_LIMIT 8 + unsigned int error_limit; + unsigned int error_decay; + + unsigned short journal_delay_ms; + bool expensive_debug_checks; + unsigned int verify:1; + unsigned int key_merging_disabled:1; + unsigned int gc_always_rewrite:1; + unsigned int shrinker_disabled:1; + unsigned int copy_gc_enabled:1; + unsigned int idle_max_writeback_rate_enabled:1; + +#define BUCKET_HASH_BITS 12 + struct hlist_head bucket_hash[1 << BUCKET_HASH_BITS]; +}; + +struct bbio { + unsigned int submit_time_us; + union { + struct bkey key; + uint64_t _pad[3]; + /* + * We only need pad = 3 here because we only ever carry around a + * single pointer - i.e. the pointer we're doing io to/from. + */ + }; + struct bio bio; +}; + +#define BTREE_PRIO USHRT_MAX +#define INITIAL_PRIO 32768U + +#define btree_bytes(c) ((c)->btree_pages * PAGE_SIZE) +#define btree_blocks(b) \ + ((unsigned int) (KEY_SIZE(&b->key) >> (b)->c->block_bits)) + +#define btree_default_blocks(c) \ + ((unsigned int) ((PAGE_SECTORS * (c)->btree_pages) >> (c)->block_bits)) + +#define bucket_bytes(ca) ((ca)->sb.bucket_size << 9) +#define block_bytes(ca) ((ca)->sb.block_size << 9) + +static inline unsigned int meta_bucket_pages(struct cache_sb *sb) +{ + unsigned int n, max_pages; + + max_pages = min_t(unsigned int, + __rounddown_pow_of_two(USHRT_MAX) / PAGE_SECTORS, + MAX_ORDER_NR_PAGES); + + n = sb->bucket_size / PAGE_SECTORS; + if (n > max_pages) + n = max_pages; + + return n; +} + +static inline unsigned int meta_bucket_bytes(struct cache_sb *sb) +{ + return meta_bucket_pages(sb) << PAGE_SHIFT; +} + +#define prios_per_bucket(ca) \ + ((meta_bucket_bytes(&(ca)->sb) - sizeof(struct prio_set)) / \ + sizeof(struct bucket_disk)) + +#define prio_buckets(ca) \ + DIV_ROUND_UP((size_t) (ca)->sb.nbuckets, prios_per_bucket(ca)) + +static inline size_t sector_to_bucket(struct cache_set *c, sector_t s) +{ + return s >> c->bucket_bits; +} + +static inline sector_t bucket_to_sector(struct cache_set *c, size_t b) +{ + return ((sector_t) b) << c->bucket_bits; +} + +static inline sector_t bucket_remainder(struct cache_set *c, sector_t s) +{ + return s & (c->cache->sb.bucket_size - 1); +} + +static inline struct cache *PTR_CACHE(struct cache_set *c, + const struct bkey *k, + unsigned int ptr) +{ + return c->cache; +} + +static inline size_t PTR_BUCKET_NR(struct cache_set *c, + const struct bkey *k, + unsigned int ptr) +{ + return sector_to_bucket(c, PTR_OFFSET(k, ptr)); +} + +static inline struct bucket *PTR_BUCKET(struct cache_set *c, + const struct bkey *k, + unsigned int ptr) +{ + return PTR_CACHE(c, k, ptr)->buckets + PTR_BUCKET_NR(c, k, ptr); +} + +static inline uint8_t gen_after(uint8_t a, uint8_t b) +{ + uint8_t r = a - b; + + return r > 128U ? 0 : r; +} + +static inline uint8_t ptr_stale(struct cache_set *c, const struct bkey *k, + unsigned int i) +{ + return gen_after(PTR_BUCKET(c, k, i)->gen, PTR_GEN(k, i)); +} + +static inline bool ptr_available(struct cache_set *c, const struct bkey *k, + unsigned int i) +{ + return (PTR_DEV(k, i) < MAX_CACHES_PER_SET) && PTR_CACHE(c, k, i); +} + +/* Btree key macros */ + +/* + * This is used for various on disk data structures - cache_sb, prio_set, bset, + * jset: The checksum is _always_ the first 8 bytes of these structs + */ +#define csum_set(i) \ + bch_crc64(((void *) (i)) + sizeof(uint64_t), \ + ((void *) bset_bkey_last(i)) - \ + (((void *) (i)) + sizeof(uint64_t))) + +/* Error handling macros */ + +#define btree_bug(b, ...) \ +do { \ + if (bch_cache_set_error((b)->c, __VA_ARGS__)) \ + dump_stack(); \ +} while (0) + +#define cache_bug(c, ...) \ +do { \ + if (bch_cache_set_error(c, __VA_ARGS__)) \ + dump_stack(); \ +} while (0) + +#define btree_bug_on(cond, b, ...) \ +do { \ + if (cond) \ + btree_bug(b, __VA_ARGS__); \ +} while (0) + +#define cache_bug_on(cond, c, ...) \ +do { \ + if (cond) \ + cache_bug(c, __VA_ARGS__); \ +} while (0) + +#define cache_set_err_on(cond, c, ...) \ +do { \ + if (cond) \ + bch_cache_set_error(c, __VA_ARGS__); \ +} while (0) + +/* Looping macros */ + +#define for_each_bucket(b, ca) \ + for (b = (ca)->buckets + (ca)->sb.first_bucket; \ + b < (ca)->buckets + (ca)->sb.nbuckets; b++) + +static inline void cached_dev_put(struct cached_dev *dc) +{ + if (refcount_dec_and_test(&dc->count)) + schedule_work(&dc->detach); +} + +static inline bool cached_dev_get(struct cached_dev *dc) +{ + if (!refcount_inc_not_zero(&dc->count)) + return false; + + /* Paired with the mb in cached_dev_attach */ + smp_mb__after_atomic(); + return true; +} + +/* + * bucket_gc_gen() returns the difference between the bucket's current gen and + * the oldest gen of any pointer into that bucket in the btree (last_gc). + */ + +static inline uint8_t bucket_gc_gen(struct bucket *b) +{ + return b->gen - b->last_gc; +} + +#define BUCKET_GC_GEN_MAX 96U + +#define kobj_attribute_write(n, fn) \ + static struct kobj_attribute ksysfs_##n = __ATTR(n, 0200, NULL, fn) + +#define kobj_attribute_rw(n, show, store) \ + static struct kobj_attribute ksysfs_##n = \ + __ATTR(n, 0600, show, store) + +static inline void wake_up_allocators(struct cache_set *c) +{ + struct cache *ca = c->cache; + + wake_up_process(ca->alloc_thread); +} + +static inline void closure_bio_submit(struct cache_set *c, + struct bio *bio, + struct closure *cl) +{ + closure_get(cl); + if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &c->flags))) { + bio->bi_status = BLK_STS_IOERR; + bio_endio(bio); + return; + } + submit_bio_noacct(bio); +} + +/* + * Prevent the kthread exits directly, and make sure when kthread_stop() + * is called to stop a kthread, it is still alive. If a kthread might be + * stopped by CACHE_SET_IO_DISABLE bit set, wait_for_kthread_stop() is + * necessary before the kthread returns. + */ +static inline void wait_for_kthread_stop(void) +{ + while (!kthread_should_stop()) { + set_current_state(TASK_INTERRUPTIBLE); + schedule(); + } +} + +/* Forward declarations */ + +void bch_count_backing_io_errors(struct cached_dev *dc, struct bio *bio); +void bch_count_io_errors(struct cache *ca, blk_status_t error, + int is_read, const char *m); +void bch_bbio_count_io_errors(struct cache_set *c, struct bio *bio, + blk_status_t error, const char *m); +void bch_bbio_endio(struct cache_set *c, struct bio *bio, + blk_status_t error, const char *m); +void bch_bbio_free(struct bio *bio, struct cache_set *c); +struct bio *bch_bbio_alloc(struct cache_set *c); + +void __bch_submit_bbio(struct bio *bio, struct cache_set *c); +void bch_submit_bbio(struct bio *bio, struct cache_set *c, + struct bkey *k, unsigned int ptr); + +uint8_t bch_inc_gen(struct cache *ca, struct bucket *b); +void bch_rescale_priorities(struct cache_set *c, int sectors); + +bool bch_can_invalidate_bucket(struct cache *ca, struct bucket *b); +void __bch_invalidate_one_bucket(struct cache *ca, struct bucket *b); + +void __bch_bucket_free(struct cache *ca, struct bucket *b); +void bch_bucket_free(struct cache_set *c, struct bkey *k); + +long bch_bucket_alloc(struct cache *ca, unsigned int reserve, bool wait); +int __bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve, + struct bkey *k, bool wait); +int bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve, + struct bkey *k, bool wait); +bool bch_alloc_sectors(struct cache_set *c, struct bkey *k, + unsigned int sectors, unsigned int write_point, + unsigned int write_prio, bool wait); +bool bch_cached_dev_error(struct cached_dev *dc); + +__printf(2, 3) +bool bch_cache_set_error(struct cache_set *c, const char *fmt, ...); + +int bch_prio_write(struct cache *ca, bool wait); +void bch_write_bdev_super(struct cached_dev *dc, struct closure *parent); + +extern struct workqueue_struct *bcache_wq; +extern struct workqueue_struct *bch_journal_wq; +extern struct workqueue_struct *bch_flush_wq; +extern struct mutex bch_register_lock; +extern struct list_head bch_cache_sets; + +extern struct kobj_type bch_cached_dev_ktype; +extern struct kobj_type bch_flash_dev_ktype; +extern struct kobj_type bch_cache_set_ktype; +extern struct kobj_type bch_cache_set_internal_ktype; +extern struct kobj_type bch_cache_ktype; + +void bch_cached_dev_release(struct kobject *kobj); +void bch_flash_dev_release(struct kobject *kobj); +void bch_cache_set_release(struct kobject *kobj); +void bch_cache_release(struct kobject *kobj); + +int bch_uuid_write(struct cache_set *c); +void bcache_write_super(struct cache_set *c); + +int bch_flash_dev_create(struct cache_set *c, uint64_t size); + +int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c, + uint8_t *set_uuid); +void bch_cached_dev_detach(struct cached_dev *dc); +int bch_cached_dev_run(struct cached_dev *dc); +void bcache_device_stop(struct bcache_device *d); + +void bch_cache_set_unregister(struct cache_set *c); +void bch_cache_set_stop(struct cache_set *c); + +struct cache_set *bch_cache_set_alloc(struct cache_sb *sb); +void bch_btree_cache_free(struct cache_set *c); +int bch_btree_cache_alloc(struct cache_set *c); +void bch_moving_init_cache_set(struct cache_set *c); +int bch_open_buckets_alloc(struct cache_set *c); +void bch_open_buckets_free(struct cache_set *c); + +int bch_cache_allocator_start(struct cache *ca); + +void bch_debug_exit(void); +void bch_debug_init(void); +void bch_request_exit(void); +int bch_request_init(void); +void bch_btree_exit(void); +int bch_btree_init(void); + +#endif /* _BCACHE_H */ diff --git a/drivers/md/bcache/bset.c b/drivers/md/bcache/bset.c new file mode 100644 index 000000000..67a2c47f4 --- /dev/null +++ b/drivers/md/bcache/bset.c @@ -0,0 +1,1386 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Code for working with individual keys, and sorted sets of keys with in a + * btree node + * + * Copyright 2012 Google, Inc. + */ + +#define pr_fmt(fmt) "bcache: %s() " fmt, __func__ + +#include "util.h" +#include "bset.h" + +#include <linux/console.h> +#include <linux/sched/clock.h> +#include <linux/random.h> +#include <linux/prefetch.h> + +#ifdef CONFIG_BCACHE_DEBUG + +void bch_dump_bset(struct btree_keys *b, struct bset *i, unsigned int set) +{ + struct bkey *k, *next; + + for (k = i->start; k < bset_bkey_last(i); k = next) { + next = bkey_next(k); + + pr_err("block %u key %u/%u: ", set, + (unsigned int) ((u64 *) k - i->d), i->keys); + + if (b->ops->key_dump) + b->ops->key_dump(b, k); + else + pr_cont("%llu:%llu\n", KEY_INODE(k), KEY_OFFSET(k)); + + if (next < bset_bkey_last(i) && + bkey_cmp(k, b->ops->is_extents ? + &START_KEY(next) : next) > 0) + pr_err("Key skipped backwards\n"); + } +} + +void bch_dump_bucket(struct btree_keys *b) +{ + unsigned int i; + + console_lock(); + for (i = 0; i <= b->nsets; i++) + bch_dump_bset(b, b->set[i].data, + bset_sector_offset(b, b->set[i].data)); + console_unlock(); +} + +int __bch_count_data(struct btree_keys *b) +{ + unsigned int ret = 0; + struct btree_iter iter; + struct bkey *k; + + if (b->ops->is_extents) + for_each_key(b, k, &iter) + ret += KEY_SIZE(k); + return ret; +} + +void __bch_check_keys(struct btree_keys *b, const char *fmt, ...) +{ + va_list args; + struct bkey *k, *p = NULL; + struct btree_iter iter; + const char *err; + + for_each_key(b, k, &iter) { + if (b->ops->is_extents) { + err = "Keys out of order"; + if (p && bkey_cmp(&START_KEY(p), &START_KEY(k)) > 0) + goto bug; + + if (bch_ptr_invalid(b, k)) + continue; + + err = "Overlapping keys"; + if (p && bkey_cmp(p, &START_KEY(k)) > 0) + goto bug; + } else { + if (bch_ptr_bad(b, k)) + continue; + + err = "Duplicate keys"; + if (p && !bkey_cmp(p, k)) + goto bug; + } + p = k; + } +#if 0 + err = "Key larger than btree node key"; + if (p && bkey_cmp(p, &b->key) > 0) + goto bug; +#endif + return; +bug: + bch_dump_bucket(b); + + va_start(args, fmt); + vprintk(fmt, args); + va_end(args); + + panic("bch_check_keys error: %s:\n", err); +} + +static void bch_btree_iter_next_check(struct btree_iter *iter) +{ + struct bkey *k = iter->data->k, *next = bkey_next(k); + + if (next < iter->data->end && + bkey_cmp(k, iter->b->ops->is_extents ? + &START_KEY(next) : next) > 0) { + bch_dump_bucket(iter->b); + panic("Key skipped backwards\n"); + } +} + +#else + +static inline void bch_btree_iter_next_check(struct btree_iter *iter) {} + +#endif + +/* Keylists */ + +int __bch_keylist_realloc(struct keylist *l, unsigned int u64s) +{ + size_t oldsize = bch_keylist_nkeys(l); + size_t newsize = oldsize + u64s; + uint64_t *old_keys = l->keys_p == l->inline_keys ? NULL : l->keys_p; + uint64_t *new_keys; + + newsize = roundup_pow_of_two(newsize); + + if (newsize <= KEYLIST_INLINE || + roundup_pow_of_two(oldsize) == newsize) + return 0; + + new_keys = krealloc(old_keys, sizeof(uint64_t) * newsize, GFP_NOIO); + + if (!new_keys) + return -ENOMEM; + + if (!old_keys) + memcpy(new_keys, l->inline_keys, sizeof(uint64_t) * oldsize); + + l->keys_p = new_keys; + l->top_p = new_keys + oldsize; + + return 0; +} + +/* Pop the top key of keylist by pointing l->top to its previous key */ +struct bkey *bch_keylist_pop(struct keylist *l) +{ + struct bkey *k = l->keys; + + if (k == l->top) + return NULL; + + while (bkey_next(k) != l->top) + k = bkey_next(k); + + return l->top = k; +} + +/* Pop the bottom key of keylist and update l->top_p */ +void bch_keylist_pop_front(struct keylist *l) +{ + l->top_p -= bkey_u64s(l->keys); + + memmove(l->keys, + bkey_next(l->keys), + bch_keylist_bytes(l)); +} + +/* Key/pointer manipulation */ + +void bch_bkey_copy_single_ptr(struct bkey *dest, const struct bkey *src, + unsigned int i) +{ + BUG_ON(i > KEY_PTRS(src)); + + /* Only copy the header, key, and one pointer. */ + memcpy(dest, src, 2 * sizeof(uint64_t)); + dest->ptr[0] = src->ptr[i]; + SET_KEY_PTRS(dest, 1); + /* We didn't copy the checksum so clear that bit. */ + SET_KEY_CSUM(dest, 0); +} + +bool __bch_cut_front(const struct bkey *where, struct bkey *k) +{ + unsigned int i, len = 0; + + if (bkey_cmp(where, &START_KEY(k)) <= 0) + return false; + + if (bkey_cmp(where, k) < 0) + len = KEY_OFFSET(k) - KEY_OFFSET(where); + else + bkey_copy_key(k, where); + + for (i = 0; i < KEY_PTRS(k); i++) + SET_PTR_OFFSET(k, i, PTR_OFFSET(k, i) + KEY_SIZE(k) - len); + + BUG_ON(len > KEY_SIZE(k)); + SET_KEY_SIZE(k, len); + return true; +} + +bool __bch_cut_back(const struct bkey *where, struct bkey *k) +{ + unsigned int len = 0; + + if (bkey_cmp(where, k) >= 0) + return false; + + BUG_ON(KEY_INODE(where) != KEY_INODE(k)); + + if (bkey_cmp(where, &START_KEY(k)) > 0) + len = KEY_OFFSET(where) - KEY_START(k); + + bkey_copy_key(k, where); + + BUG_ON(len > KEY_SIZE(k)); + SET_KEY_SIZE(k, len); + return true; +} + +/* Auxiliary search trees */ + +/* 32 bits total: */ +#define BKEY_MID_BITS 3 +#define BKEY_EXPONENT_BITS 7 +#define BKEY_MANTISSA_BITS (32 - BKEY_MID_BITS - BKEY_EXPONENT_BITS) +#define BKEY_MANTISSA_MASK ((1 << BKEY_MANTISSA_BITS) - 1) + +struct bkey_float { + unsigned int exponent:BKEY_EXPONENT_BITS; + unsigned int m:BKEY_MID_BITS; + unsigned int mantissa:BKEY_MANTISSA_BITS; +} __packed; + +/* + * BSET_CACHELINE was originally intended to match the hardware cacheline size - + * it used to be 64, but I realized the lookup code would touch slightly less + * memory if it was 128. + * + * It definites the number of bytes (in struct bset) per struct bkey_float in + * the auxiliar search tree - when we're done searching the bset_float tree we + * have this many bytes left that we do a linear search over. + * + * Since (after level 5) every level of the bset_tree is on a new cacheline, + * we're touching one fewer cacheline in the bset tree in exchange for one more + * cacheline in the linear search - but the linear search might stop before it + * gets to the second cacheline. + */ + +#define BSET_CACHELINE 128 + +/* Space required for the btree node keys */ +static inline size_t btree_keys_bytes(struct btree_keys *b) +{ + return PAGE_SIZE << b->page_order; +} + +static inline size_t btree_keys_cachelines(struct btree_keys *b) +{ + return btree_keys_bytes(b) / BSET_CACHELINE; +} + +/* Space required for the auxiliary search trees */ +static inline size_t bset_tree_bytes(struct btree_keys *b) +{ + return btree_keys_cachelines(b) * sizeof(struct bkey_float); +} + +/* Space required for the prev pointers */ +static inline size_t bset_prev_bytes(struct btree_keys *b) +{ + return btree_keys_cachelines(b) * sizeof(uint8_t); +} + +/* Memory allocation */ + +void bch_btree_keys_free(struct btree_keys *b) +{ + struct bset_tree *t = b->set; + + if (bset_prev_bytes(b) < PAGE_SIZE) + kfree(t->prev); + else + free_pages((unsigned long) t->prev, + get_order(bset_prev_bytes(b))); + + if (bset_tree_bytes(b) < PAGE_SIZE) + kfree(t->tree); + else + free_pages((unsigned long) t->tree, + get_order(bset_tree_bytes(b))); + + free_pages((unsigned long) t->data, b->page_order); + + t->prev = NULL; + t->tree = NULL; + t->data = NULL; +} + +int bch_btree_keys_alloc(struct btree_keys *b, + unsigned int page_order, + gfp_t gfp) +{ + struct bset_tree *t = b->set; + + BUG_ON(t->data); + + b->page_order = page_order; + + t->data = (void *) __get_free_pages(__GFP_COMP|gfp, b->page_order); + if (!t->data) + goto err; + + t->tree = bset_tree_bytes(b) < PAGE_SIZE + ? kmalloc(bset_tree_bytes(b), gfp) + : (void *) __get_free_pages(gfp, get_order(bset_tree_bytes(b))); + if (!t->tree) + goto err; + + t->prev = bset_prev_bytes(b) < PAGE_SIZE + ? kmalloc(bset_prev_bytes(b), gfp) + : (void *) __get_free_pages(gfp, get_order(bset_prev_bytes(b))); + if (!t->prev) + goto err; + + return 0; +err: + bch_btree_keys_free(b); + return -ENOMEM; +} + +void bch_btree_keys_init(struct btree_keys *b, const struct btree_keys_ops *ops, + bool *expensive_debug_checks) +{ + b->ops = ops; + b->expensive_debug_checks = expensive_debug_checks; + b->nsets = 0; + b->last_set_unwritten = 0; + + /* + * struct btree_keys in embedded in struct btree, and struct + * bset_tree is embedded into struct btree_keys. They are all + * initialized as 0 by kzalloc() in mca_bucket_alloc(), and + * b->set[0].data is allocated in bch_btree_keys_alloc(), so we + * don't have to initiate b->set[].size and b->set[].data here + * any more. + */ +} + +/* Binary tree stuff for auxiliary search trees */ + +/* + * return array index next to j when does in-order traverse + * of a binary tree which is stored in a linear array + */ +static unsigned int inorder_next(unsigned int j, unsigned int size) +{ + if (j * 2 + 1 < size) { + j = j * 2 + 1; + + while (j * 2 < size) + j *= 2; + } else + j >>= ffz(j) + 1; + + return j; +} + +/* + * return array index previous to j when does in-order traverse + * of a binary tree which is stored in a linear array + */ +static unsigned int inorder_prev(unsigned int j, unsigned int size) +{ + if (j * 2 < size) { + j = j * 2; + + while (j * 2 + 1 < size) + j = j * 2 + 1; + } else + j >>= ffs(j); + + return j; +} + +/* + * I have no idea why this code works... and I'm the one who wrote it + * + * However, I do know what it does: + * Given a binary tree constructed in an array (i.e. how you normally implement + * a heap), it converts a node in the tree - referenced by array index - to the + * index it would have if you did an inorder traversal. + * + * Also tested for every j, size up to size somewhere around 6 million. + * + * The binary tree starts at array index 1, not 0 + * extra is a function of size: + * extra = (size - rounddown_pow_of_two(size - 1)) << 1; + */ +static unsigned int __to_inorder(unsigned int j, + unsigned int size, + unsigned int extra) +{ + unsigned int b = fls(j); + unsigned int shift = fls(size - 1) - b; + + j ^= 1U << (b - 1); + j <<= 1; + j |= 1; + j <<= shift; + + if (j > extra) + j -= (j - extra) >> 1; + + return j; +} + +/* + * Return the cacheline index in bset_tree->data, where j is index + * from a linear array which stores the auxiliar binary tree + */ +static unsigned int to_inorder(unsigned int j, struct bset_tree *t) +{ + return __to_inorder(j, t->size, t->extra); +} + +static unsigned int __inorder_to_tree(unsigned int j, + unsigned int size, + unsigned int extra) +{ + unsigned int shift; + + if (j > extra) + j += j - extra; + + shift = ffs(j); + + j >>= shift; + j |= roundup_pow_of_two(size) >> shift; + + return j; +} + +/* + * Return an index from a linear array which stores the auxiliar binary + * tree, j is the cacheline index of t->data. + */ +static unsigned int inorder_to_tree(unsigned int j, struct bset_tree *t) +{ + return __inorder_to_tree(j, t->size, t->extra); +} + +#if 0 +void inorder_test(void) +{ + unsigned long done = 0; + ktime_t start = ktime_get(); + + for (unsigned int size = 2; + size < 65536000; + size++) { + unsigned int extra = + (size - rounddown_pow_of_two(size - 1)) << 1; + unsigned int i = 1, j = rounddown_pow_of_two(size - 1); + + if (!(size % 4096)) + pr_notice("loop %u, %llu per us\n", size, + done / ktime_us_delta(ktime_get(), start)); + + while (1) { + if (__inorder_to_tree(i, size, extra) != j) + panic("size %10u j %10u i %10u", size, j, i); + + if (__to_inorder(j, size, extra) != i) + panic("size %10u j %10u i %10u", size, j, i); + + if (j == rounddown_pow_of_two(size) - 1) + break; + + BUG_ON(inorder_prev(inorder_next(j, size), size) != j); + + j = inorder_next(j, size); + i++; + } + + done += size - 1; + } +} +#endif + +/* + * Cacheline/offset <-> bkey pointer arithmetic: + * + * t->tree is a binary search tree in an array; each node corresponds to a key + * in one cacheline in t->set (BSET_CACHELINE bytes). + * + * This means we don't have to store the full index of the key that a node in + * the binary tree points to; to_inorder() gives us the cacheline, and then + * bkey_float->m gives us the offset within that cacheline, in units of 8 bytes. + * + * cacheline_to_bkey() and friends abstract out all the pointer arithmetic to + * make this work. + * + * To construct the bfloat for an arbitrary key we need to know what the key + * immediately preceding it is: we have to check if the two keys differ in the + * bits we're going to store in bkey_float->mantissa. t->prev[j] stores the size + * of the previous key so we can walk backwards to it from t->tree[j]'s key. + */ + +static struct bkey *cacheline_to_bkey(struct bset_tree *t, + unsigned int cacheline, + unsigned int offset) +{ + return ((void *) t->data) + cacheline * BSET_CACHELINE + offset * 8; +} + +static unsigned int bkey_to_cacheline(struct bset_tree *t, struct bkey *k) +{ + return ((void *) k - (void *) t->data) / BSET_CACHELINE; +} + +static unsigned int bkey_to_cacheline_offset(struct bset_tree *t, + unsigned int cacheline, + struct bkey *k) +{ + return (u64 *) k - (u64 *) cacheline_to_bkey(t, cacheline, 0); +} + +static struct bkey *tree_to_bkey(struct bset_tree *t, unsigned int j) +{ + return cacheline_to_bkey(t, to_inorder(j, t), t->tree[j].m); +} + +static struct bkey *tree_to_prev_bkey(struct bset_tree *t, unsigned int j) +{ + return (void *) (((uint64_t *) tree_to_bkey(t, j)) - t->prev[j]); +} + +/* + * For the write set - the one we're currently inserting keys into - we don't + * maintain a full search tree, we just keep a simple lookup table in t->prev. + */ +static struct bkey *table_to_bkey(struct bset_tree *t, unsigned int cacheline) +{ + return cacheline_to_bkey(t, cacheline, t->prev[cacheline]); +} + +static inline uint64_t shrd128(uint64_t high, uint64_t low, uint8_t shift) +{ + low >>= shift; + low |= (high << 1) << (63U - shift); + return low; +} + +/* + * Calculate mantissa value for struct bkey_float. + * If most significant bit of f->exponent is not set, then + * - f->exponent >> 6 is 0 + * - p[0] points to bkey->low + * - p[-1] borrows bits from KEY_INODE() of bkey->high + * if most isgnificant bits of f->exponent is set, then + * - f->exponent >> 6 is 1 + * - p[0] points to bits from KEY_INODE() of bkey->high + * - p[-1] points to other bits from KEY_INODE() of + * bkey->high too. + * See make_bfloat() to check when most significant bit of f->exponent + * is set or not. + */ +static inline unsigned int bfloat_mantissa(const struct bkey *k, + struct bkey_float *f) +{ + const uint64_t *p = &k->low - (f->exponent >> 6); + + return shrd128(p[-1], p[0], f->exponent & 63) & BKEY_MANTISSA_MASK; +} + +static void make_bfloat(struct bset_tree *t, unsigned int j) +{ + struct bkey_float *f = &t->tree[j]; + struct bkey *m = tree_to_bkey(t, j); + struct bkey *p = tree_to_prev_bkey(t, j); + + struct bkey *l = is_power_of_2(j) + ? t->data->start + : tree_to_prev_bkey(t, j >> ffs(j)); + + struct bkey *r = is_power_of_2(j + 1) + ? bset_bkey_idx(t->data, t->data->keys - bkey_u64s(&t->end)) + : tree_to_bkey(t, j >> (ffz(j) + 1)); + + BUG_ON(m < l || m > r); + BUG_ON(bkey_next(p) != m); + + /* + * If l and r have different KEY_INODE values (different backing + * device), f->exponent records how many least significant bits + * are different in KEY_INODE values and sets most significant + * bits to 1 (by +64). + * If l and r have same KEY_INODE value, f->exponent records + * how many different bits in least significant bits of bkey->low. + * See bfloat_mantiss() how the most significant bit of + * f->exponent is used to calculate bfloat mantissa value. + */ + if (KEY_INODE(l) != KEY_INODE(r)) + f->exponent = fls64(KEY_INODE(r) ^ KEY_INODE(l)) + 64; + else + f->exponent = fls64(r->low ^ l->low); + + f->exponent = max_t(int, f->exponent - BKEY_MANTISSA_BITS, 0); + + /* + * Setting f->exponent = 127 flags this node as failed, and causes the + * lookup code to fall back to comparing against the original key. + */ + + if (bfloat_mantissa(m, f) != bfloat_mantissa(p, f)) + f->mantissa = bfloat_mantissa(m, f) - 1; + else + f->exponent = 127; +} + +static void bset_alloc_tree(struct btree_keys *b, struct bset_tree *t) +{ + if (t != b->set) { + unsigned int j = roundup(t[-1].size, + 64 / sizeof(struct bkey_float)); + + t->tree = t[-1].tree + j; + t->prev = t[-1].prev + j; + } + + while (t < b->set + MAX_BSETS) + t++->size = 0; +} + +static void bch_bset_build_unwritten_tree(struct btree_keys *b) +{ + struct bset_tree *t = bset_tree_last(b); + + BUG_ON(b->last_set_unwritten); + b->last_set_unwritten = 1; + + bset_alloc_tree(b, t); + + if (t->tree != b->set->tree + btree_keys_cachelines(b)) { + t->prev[0] = bkey_to_cacheline_offset(t, 0, t->data->start); + t->size = 1; + } +} + +void bch_bset_init_next(struct btree_keys *b, struct bset *i, uint64_t magic) +{ + if (i != b->set->data) { + b->set[++b->nsets].data = i; + i->seq = b->set->data->seq; + } else + get_random_bytes(&i->seq, sizeof(uint64_t)); + + i->magic = magic; + i->version = 0; + i->keys = 0; + + bch_bset_build_unwritten_tree(b); +} + +/* + * Build auxiliary binary tree 'struct bset_tree *t', this tree is used to + * accelerate bkey search in a btree node (pointed by bset_tree->data in + * memory). After search in the auxiliar tree by calling bset_search_tree(), + * a struct bset_search_iter is returned which indicates range [l, r] from + * bset_tree->data where the searching bkey might be inside. Then a followed + * linear comparison does the exact search, see __bch_bset_search() for how + * the auxiliary tree is used. + */ +void bch_bset_build_written_tree(struct btree_keys *b) +{ + struct bset_tree *t = bset_tree_last(b); + struct bkey *prev = NULL, *k = t->data->start; + unsigned int j, cacheline = 1; + + b->last_set_unwritten = 0; + + bset_alloc_tree(b, t); + + t->size = min_t(unsigned int, + bkey_to_cacheline(t, bset_bkey_last(t->data)), + b->set->tree + btree_keys_cachelines(b) - t->tree); + + if (t->size < 2) { + t->size = 0; + return; + } + + t->extra = (t->size - rounddown_pow_of_two(t->size - 1)) << 1; + + /* First we figure out where the first key in each cacheline is */ + for (j = inorder_next(0, t->size); + j; + j = inorder_next(j, t->size)) { + while (bkey_to_cacheline(t, k) < cacheline) + prev = k, k = bkey_next(k); + + t->prev[j] = bkey_u64s(prev); + t->tree[j].m = bkey_to_cacheline_offset(t, cacheline++, k); + } + + while (bkey_next(k) != bset_bkey_last(t->data)) + k = bkey_next(k); + + t->end = *k; + + /* Then we build the tree */ + for (j = inorder_next(0, t->size); + j; + j = inorder_next(j, t->size)) + make_bfloat(t, j); +} + +/* Insert */ + +void bch_bset_fix_invalidated_key(struct btree_keys *b, struct bkey *k) +{ + struct bset_tree *t; + unsigned int inorder, j = 1; + + for (t = b->set; t <= bset_tree_last(b); t++) + if (k < bset_bkey_last(t->data)) + goto found_set; + + BUG(); +found_set: + if (!t->size || !bset_written(b, t)) + return; + + inorder = bkey_to_cacheline(t, k); + + if (k == t->data->start) + goto fix_left; + + if (bkey_next(k) == bset_bkey_last(t->data)) { + t->end = *k; + goto fix_right; + } + + j = inorder_to_tree(inorder, t); + + if (j && + j < t->size && + k == tree_to_bkey(t, j)) +fix_left: do { + make_bfloat(t, j); + j = j * 2; + } while (j < t->size); + + j = inorder_to_tree(inorder + 1, t); + + if (j && + j < t->size && + k == tree_to_prev_bkey(t, j)) +fix_right: do { + make_bfloat(t, j); + j = j * 2 + 1; + } while (j < t->size); +} + +static void bch_bset_fix_lookup_table(struct btree_keys *b, + struct bset_tree *t, + struct bkey *k) +{ + unsigned int shift = bkey_u64s(k); + unsigned int j = bkey_to_cacheline(t, k); + + /* We're getting called from btree_split() or btree_gc, just bail out */ + if (!t->size) + return; + + /* + * k is the key we just inserted; we need to find the entry in the + * lookup table for the first key that is strictly greater than k: + * it's either k's cacheline or the next one + */ + while (j < t->size && + table_to_bkey(t, j) <= k) + j++; + + /* + * Adjust all the lookup table entries, and find a new key for any that + * have gotten too big + */ + for (; j < t->size; j++) { + t->prev[j] += shift; + + if (t->prev[j] > 7) { + k = table_to_bkey(t, j - 1); + + while (k < cacheline_to_bkey(t, j, 0)) + k = bkey_next(k); + + t->prev[j] = bkey_to_cacheline_offset(t, j, k); + } + } + + if (t->size == b->set->tree + btree_keys_cachelines(b) - t->tree) + return; + + /* Possibly add a new entry to the end of the lookup table */ + + for (k = table_to_bkey(t, t->size - 1); + k != bset_bkey_last(t->data); + k = bkey_next(k)) + if (t->size == bkey_to_cacheline(t, k)) { + t->prev[t->size] = + bkey_to_cacheline_offset(t, t->size, k); + t->size++; + } +} + +/* + * Tries to merge l and r: l should be lower than r + * Returns true if we were able to merge. If we did merge, l will be the merged + * key, r will be untouched. + */ +bool bch_bkey_try_merge(struct btree_keys *b, struct bkey *l, struct bkey *r) +{ + if (!b->ops->key_merge) + return false; + + /* + * Generic header checks + * Assumes left and right are in order + * Left and right must be exactly aligned + */ + if (!bch_bkey_equal_header(l, r) || + bkey_cmp(l, &START_KEY(r))) + return false; + + return b->ops->key_merge(b, l, r); +} + +void bch_bset_insert(struct btree_keys *b, struct bkey *where, + struct bkey *insert) +{ + struct bset_tree *t = bset_tree_last(b); + + BUG_ON(!b->last_set_unwritten); + BUG_ON(bset_byte_offset(b, t->data) + + __set_bytes(t->data, t->data->keys + bkey_u64s(insert)) > + PAGE_SIZE << b->page_order); + + memmove((uint64_t *) where + bkey_u64s(insert), + where, + (void *) bset_bkey_last(t->data) - (void *) where); + + t->data->keys += bkey_u64s(insert); + bkey_copy(where, insert); + bch_bset_fix_lookup_table(b, t, where); +} + +unsigned int bch_btree_insert_key(struct btree_keys *b, struct bkey *k, + struct bkey *replace_key) +{ + unsigned int status = BTREE_INSERT_STATUS_NO_INSERT; + struct bset *i = bset_tree_last(b)->data; + struct bkey *m, *prev = NULL; + struct btree_iter iter; + struct bkey preceding_key_on_stack = ZERO_KEY; + struct bkey *preceding_key_p = &preceding_key_on_stack; + + BUG_ON(b->ops->is_extents && !KEY_SIZE(k)); + + /* + * If k has preceding key, preceding_key_p will be set to address + * of k's preceding key; otherwise preceding_key_p will be set + * to NULL inside preceding_key(). + */ + if (b->ops->is_extents) + preceding_key(&START_KEY(k), &preceding_key_p); + else + preceding_key(k, &preceding_key_p); + + m = bch_btree_iter_init(b, &iter, preceding_key_p); + + if (b->ops->insert_fixup(b, k, &iter, replace_key)) + return status; + + status = BTREE_INSERT_STATUS_INSERT; + + while (m != bset_bkey_last(i) && + bkey_cmp(k, b->ops->is_extents ? &START_KEY(m) : m) > 0) + prev = m, m = bkey_next(m); + + /* prev is in the tree, if we merge we're done */ + status = BTREE_INSERT_STATUS_BACK_MERGE; + if (prev && + bch_bkey_try_merge(b, prev, k)) + goto merged; +#if 0 + status = BTREE_INSERT_STATUS_OVERWROTE; + if (m != bset_bkey_last(i) && + KEY_PTRS(m) == KEY_PTRS(k) && !KEY_SIZE(m)) + goto copy; +#endif + status = BTREE_INSERT_STATUS_FRONT_MERGE; + if (m != bset_bkey_last(i) && + bch_bkey_try_merge(b, k, m)) + goto copy; + + bch_bset_insert(b, m, k); +copy: bkey_copy(m, k); +merged: + return status; +} + +/* Lookup */ + +struct bset_search_iter { + struct bkey *l, *r; +}; + +static struct bset_search_iter bset_search_write_set(struct bset_tree *t, + const struct bkey *search) +{ + unsigned int li = 0, ri = t->size; + + while (li + 1 != ri) { + unsigned int m = (li + ri) >> 1; + + if (bkey_cmp(table_to_bkey(t, m), search) > 0) + ri = m; + else + li = m; + } + + return (struct bset_search_iter) { + table_to_bkey(t, li), + ri < t->size ? table_to_bkey(t, ri) : bset_bkey_last(t->data) + }; +} + +static struct bset_search_iter bset_search_tree(struct bset_tree *t, + const struct bkey *search) +{ + struct bkey *l, *r; + struct bkey_float *f; + unsigned int inorder, j, n = 1; + + do { + unsigned int p = n << 4; + + if (p < t->size) + prefetch(&t->tree[p]); + + j = n; + f = &t->tree[j]; + + if (likely(f->exponent != 127)) { + if (f->mantissa >= bfloat_mantissa(search, f)) + n = j * 2; + else + n = j * 2 + 1; + } else { + if (bkey_cmp(tree_to_bkey(t, j), search) > 0) + n = j * 2; + else + n = j * 2 + 1; + } + } while (n < t->size); + + inorder = to_inorder(j, t); + + /* + * n would have been the node we recursed to - the low bit tells us if + * we recursed left or recursed right. + */ + if (n & 1) { + l = cacheline_to_bkey(t, inorder, f->m); + + if (++inorder != t->size) { + f = &t->tree[inorder_next(j, t->size)]; + r = cacheline_to_bkey(t, inorder, f->m); + } else + r = bset_bkey_last(t->data); + } else { + r = cacheline_to_bkey(t, inorder, f->m); + + if (--inorder) { + f = &t->tree[inorder_prev(j, t->size)]; + l = cacheline_to_bkey(t, inorder, f->m); + } else + l = t->data->start; + } + + return (struct bset_search_iter) {l, r}; +} + +struct bkey *__bch_bset_search(struct btree_keys *b, struct bset_tree *t, + const struct bkey *search) +{ + struct bset_search_iter i; + + /* + * First, we search for a cacheline, then lastly we do a linear search + * within that cacheline. + * + * To search for the cacheline, there's three different possibilities: + * * The set is too small to have a search tree, so we just do a linear + * search over the whole set. + * * The set is the one we're currently inserting into; keeping a full + * auxiliary search tree up to date would be too expensive, so we + * use a much simpler lookup table to do a binary search - + * bset_search_write_set(). + * * Or we use the auxiliary search tree we constructed earlier - + * bset_search_tree() + */ + + if (unlikely(!t->size)) { + i.l = t->data->start; + i.r = bset_bkey_last(t->data); + } else if (bset_written(b, t)) { + /* + * Each node in the auxiliary search tree covers a certain range + * of bits, and keys above and below the set it covers might + * differ outside those bits - so we have to special case the + * start and end - handle that here: + */ + + if (unlikely(bkey_cmp(search, &t->end) >= 0)) + return bset_bkey_last(t->data); + + if (unlikely(bkey_cmp(search, t->data->start) < 0)) + return t->data->start; + + i = bset_search_tree(t, search); + } else { + BUG_ON(!b->nsets && + t->size < bkey_to_cacheline(t, bset_bkey_last(t->data))); + + i = bset_search_write_set(t, search); + } + + if (btree_keys_expensive_checks(b)) { + BUG_ON(bset_written(b, t) && + i.l != t->data->start && + bkey_cmp(tree_to_prev_bkey(t, + inorder_to_tree(bkey_to_cacheline(t, i.l), t)), + search) > 0); + + BUG_ON(i.r != bset_bkey_last(t->data) && + bkey_cmp(i.r, search) <= 0); + } + + while (likely(i.l != i.r) && + bkey_cmp(i.l, search) <= 0) + i.l = bkey_next(i.l); + + return i.l; +} + +/* Btree iterator */ + +typedef bool (btree_iter_cmp_fn)(struct btree_iter_set, + struct btree_iter_set); + +static inline bool btree_iter_cmp(struct btree_iter_set l, + struct btree_iter_set r) +{ + return bkey_cmp(l.k, r.k) > 0; +} + +static inline bool btree_iter_end(struct btree_iter *iter) +{ + return !iter->used; +} + +void bch_btree_iter_push(struct btree_iter *iter, struct bkey *k, + struct bkey *end) +{ + if (k != end) + BUG_ON(!heap_add(iter, + ((struct btree_iter_set) { k, end }), + btree_iter_cmp)); +} + +static struct bkey *__bch_btree_iter_init(struct btree_keys *b, + struct btree_iter *iter, + struct bkey *search, + struct bset_tree *start) +{ + struct bkey *ret = NULL; + + iter->size = ARRAY_SIZE(iter->data); + iter->used = 0; + +#ifdef CONFIG_BCACHE_DEBUG + iter->b = b; +#endif + + for (; start <= bset_tree_last(b); start++) { + ret = bch_bset_search(b, start, search); + bch_btree_iter_push(iter, ret, bset_bkey_last(start->data)); + } + + return ret; +} + +struct bkey *bch_btree_iter_init(struct btree_keys *b, + struct btree_iter *iter, + struct bkey *search) +{ + return __bch_btree_iter_init(b, iter, search, b->set); +} + +static inline struct bkey *__bch_btree_iter_next(struct btree_iter *iter, + btree_iter_cmp_fn *cmp) +{ + struct btree_iter_set b __maybe_unused; + struct bkey *ret = NULL; + + if (!btree_iter_end(iter)) { + bch_btree_iter_next_check(iter); + + ret = iter->data->k; + iter->data->k = bkey_next(iter->data->k); + + if (iter->data->k > iter->data->end) { + WARN_ONCE(1, "bset was corrupt!\n"); + iter->data->k = iter->data->end; + } + + if (iter->data->k == iter->data->end) + heap_pop(iter, b, cmp); + else + heap_sift(iter, 0, cmp); + } + + return ret; +} + +struct bkey *bch_btree_iter_next(struct btree_iter *iter) +{ + return __bch_btree_iter_next(iter, btree_iter_cmp); + +} + +struct bkey *bch_btree_iter_next_filter(struct btree_iter *iter, + struct btree_keys *b, ptr_filter_fn fn) +{ + struct bkey *ret; + + do { + ret = bch_btree_iter_next(iter); + } while (ret && fn(b, ret)); + + return ret; +} + +/* Mergesort */ + +void bch_bset_sort_state_free(struct bset_sort_state *state) +{ + mempool_exit(&state->pool); +} + +int bch_bset_sort_state_init(struct bset_sort_state *state, + unsigned int page_order) +{ + spin_lock_init(&state->time.lock); + + state->page_order = page_order; + state->crit_factor = int_sqrt(1 << page_order); + + return mempool_init_page_pool(&state->pool, 1, page_order); +} + +static void btree_mergesort(struct btree_keys *b, struct bset *out, + struct btree_iter *iter, + bool fixup, bool remove_stale) +{ + int i; + struct bkey *k, *last = NULL; + BKEY_PADDED(k) tmp; + bool (*bad)(struct btree_keys *, const struct bkey *) = remove_stale + ? bch_ptr_bad + : bch_ptr_invalid; + + /* Heapify the iterator, using our comparison function */ + for (i = iter->used / 2 - 1; i >= 0; --i) + heap_sift(iter, i, b->ops->sort_cmp); + + while (!btree_iter_end(iter)) { + if (b->ops->sort_fixup && fixup) + k = b->ops->sort_fixup(iter, &tmp.k); + else + k = NULL; + + if (!k) + k = __bch_btree_iter_next(iter, b->ops->sort_cmp); + + if (bad(b, k)) + continue; + + if (!last) { + last = out->start; + bkey_copy(last, k); + } else if (!bch_bkey_try_merge(b, last, k)) { + last = bkey_next(last); + bkey_copy(last, k); + } + } + + out->keys = last ? (uint64_t *) bkey_next(last) - out->d : 0; + + pr_debug("sorted %i keys\n", out->keys); +} + +static void __btree_sort(struct btree_keys *b, struct btree_iter *iter, + unsigned int start, unsigned int order, bool fixup, + struct bset_sort_state *state) +{ + uint64_t start_time; + bool used_mempool = false; + struct bset *out = (void *) __get_free_pages(__GFP_NOWARN|GFP_NOWAIT, + order); + if (!out) { + struct page *outp; + + BUG_ON(order > state->page_order); + + outp = mempool_alloc(&state->pool, GFP_NOIO); + out = page_address(outp); + used_mempool = true; + order = state->page_order; + } + + start_time = local_clock(); + + btree_mergesort(b, out, iter, fixup, false); + b->nsets = start; + + if (!start && order == b->page_order) { + /* + * Our temporary buffer is the same size as the btree node's + * buffer, we can just swap buffers instead of doing a big + * memcpy() + * + * Don't worry event 'out' is allocated from mempool, it can + * still be swapped here. Because state->pool is a page mempool + * creaated by by mempool_init_page_pool(), which allocates + * pages by alloc_pages() indeed. + */ + + out->magic = b->set->data->magic; + out->seq = b->set->data->seq; + out->version = b->set->data->version; + swap(out, b->set->data); + } else { + b->set[start].data->keys = out->keys; + memcpy(b->set[start].data->start, out->start, + (void *) bset_bkey_last(out) - (void *) out->start); + } + + if (used_mempool) + mempool_free(virt_to_page(out), &state->pool); + else + free_pages((unsigned long) out, order); + + bch_bset_build_written_tree(b); + + if (!start) + bch_time_stats_update(&state->time, start_time); +} + +void bch_btree_sort_partial(struct btree_keys *b, unsigned int start, + struct bset_sort_state *state) +{ + size_t order = b->page_order, keys = 0; + struct btree_iter iter; + int oldsize = bch_count_data(b); + + __bch_btree_iter_init(b, &iter, NULL, &b->set[start]); + + if (start) { + unsigned int i; + + for (i = start; i <= b->nsets; i++) + keys += b->set[i].data->keys; + + order = get_order(__set_bytes(b->set->data, keys)); + } + + __btree_sort(b, &iter, start, order, false, state); + + EBUG_ON(oldsize >= 0 && bch_count_data(b) != oldsize); +} + +void bch_btree_sort_and_fix_extents(struct btree_keys *b, + struct btree_iter *iter, + struct bset_sort_state *state) +{ + __btree_sort(b, iter, 0, b->page_order, true, state); +} + +void bch_btree_sort_into(struct btree_keys *b, struct btree_keys *new, + struct bset_sort_state *state) +{ + uint64_t start_time = local_clock(); + struct btree_iter iter; + + bch_btree_iter_init(b, &iter, NULL); + + btree_mergesort(b, new->set->data, &iter, false, true); + + bch_time_stats_update(&state->time, start_time); + + new->set->size = 0; // XXX: why? +} + +#define SORT_CRIT (4096 / sizeof(uint64_t)) + +void bch_btree_sort_lazy(struct btree_keys *b, struct bset_sort_state *state) +{ + unsigned int crit = SORT_CRIT; + int i; + + /* Don't sort if nothing to do */ + if (!b->nsets) + goto out; + + for (i = b->nsets - 1; i >= 0; --i) { + crit *= state->crit_factor; + + if (b->set[i].data->keys < crit) { + bch_btree_sort_partial(b, i, state); + return; + } + } + + /* Sort if we'd overflow */ + if (b->nsets + 1 == MAX_BSETS) { + bch_btree_sort(b, state); + return; + } + +out: + bch_bset_build_written_tree(b); +} + +void bch_btree_keys_stats(struct btree_keys *b, struct bset_stats *stats) +{ + unsigned int i; + + for (i = 0; i <= b->nsets; i++) { + struct bset_tree *t = &b->set[i]; + size_t bytes = t->data->keys * sizeof(uint64_t); + size_t j; + + if (bset_written(b, t)) { + stats->sets_written++; + stats->bytes_written += bytes; + + stats->floats += t->size - 1; + + for (j = 1; j < t->size; j++) + if (t->tree[j].exponent == 127) + stats->failed++; + } else { + stats->sets_unwritten++; + stats->bytes_unwritten += bytes; + } + } +} diff --git a/drivers/md/bcache/bset.h b/drivers/md/bcache/bset.h new file mode 100644 index 000000000..a50dcfda6 --- /dev/null +++ b/drivers/md/bcache/bset.h @@ -0,0 +1,593 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _BCACHE_BSET_H +#define _BCACHE_BSET_H + +#include <linux/bcache.h> +#include <linux/kernel.h> +#include <linux/types.h> + +#include "util.h" /* for time_stats */ + +/* + * BKEYS: + * + * A bkey contains a key, a size field, a variable number of pointers, and some + * ancillary flag bits. + * + * We use two different functions for validating bkeys, bch_ptr_invalid and + * bch_ptr_bad(). + * + * bch_ptr_invalid() primarily filters out keys and pointers that would be + * invalid due to some sort of bug, whereas bch_ptr_bad() filters out keys and + * pointer that occur in normal practice but don't point to real data. + * + * The one exception to the rule that ptr_invalid() filters out invalid keys is + * that it also filters out keys of size 0 - these are keys that have been + * completely overwritten. It'd be safe to delete these in memory while leaving + * them on disk, just unnecessary work - so we filter them out when resorting + * instead. + * + * We can't filter out stale keys when we're resorting, because garbage + * collection needs to find them to ensure bucket gens don't wrap around - + * unless we're rewriting the btree node those stale keys still exist on disk. + * + * We also implement functions here for removing some number of sectors from the + * front or the back of a bkey - this is mainly used for fixing overlapping + * extents, by removing the overlapping sectors from the older key. + * + * BSETS: + * + * A bset is an array of bkeys laid out contiguously in memory in sorted order, + * along with a header. A btree node is made up of a number of these, written at + * different times. + * + * There could be many of them on disk, but we never allow there to be more than + * 4 in memory - we lazily resort as needed. + * + * We implement code here for creating and maintaining auxiliary search trees + * (described below) for searching an individial bset, and on top of that we + * implement a btree iterator. + * + * BTREE ITERATOR: + * + * Most of the code in bcache doesn't care about an individual bset - it needs + * to search entire btree nodes and iterate over them in sorted order. + * + * The btree iterator code serves both functions; it iterates through the keys + * in a btree node in sorted order, starting from either keys after a specific + * point (if you pass it a search key) or the start of the btree node. + * + * AUXILIARY SEARCH TREES: + * + * Since keys are variable length, we can't use a binary search on a bset - we + * wouldn't be able to find the start of the next key. But binary searches are + * slow anyways, due to terrible cache behaviour; bcache originally used binary + * searches and that code topped out at under 50k lookups/second. + * + * So we need to construct some sort of lookup table. Since we only insert keys + * into the last (unwritten) set, most of the keys within a given btree node are + * usually in sets that are mostly constant. We use two different types of + * lookup tables to take advantage of this. + * + * Both lookup tables share in common that they don't index every key in the + * set; they index one key every BSET_CACHELINE bytes, and then a linear search + * is used for the rest. + * + * For sets that have been written to disk and are no longer being inserted + * into, we construct a binary search tree in an array - traversing a binary + * search tree in an array gives excellent locality of reference and is very + * fast, since both children of any node are adjacent to each other in memory + * (and their grandchildren, and great grandchildren...) - this means + * prefetching can be used to great effect. + * + * It's quite useful performance wise to keep these nodes small - not just + * because they're more likely to be in L2, but also because we can prefetch + * more nodes on a single cacheline and thus prefetch more iterations in advance + * when traversing this tree. + * + * Nodes in the auxiliary search tree must contain both a key to compare against + * (we don't want to fetch the key from the set, that would defeat the purpose), + * and a pointer to the key. We use a few tricks to compress both of these. + * + * To compress the pointer, we take advantage of the fact that one node in the + * search tree corresponds to precisely BSET_CACHELINE bytes in the set. We have + * a function (to_inorder()) that takes the index of a node in a binary tree and + * returns what its index would be in an inorder traversal, so we only have to + * store the low bits of the offset. + * + * The key is 84 bits (KEY_DEV + key->key, the offset on the device). To + * compress that, we take advantage of the fact that when we're traversing the + * search tree at every iteration we know that both our search key and the key + * we're looking for lie within some range - bounded by our previous + * comparisons. (We special case the start of a search so that this is true even + * at the root of the tree). + * + * So we know the key we're looking for is between a and b, and a and b don't + * differ higher than bit 50, we don't need to check anything higher than bit + * 50. + * + * We don't usually need the rest of the bits, either; we only need enough bits + * to partition the key range we're currently checking. Consider key n - the + * key our auxiliary search tree node corresponds to, and key p, the key + * immediately preceding n. The lowest bit we need to store in the auxiliary + * search tree is the highest bit that differs between n and p. + * + * Note that this could be bit 0 - we might sometimes need all 80 bits to do the + * comparison. But we'd really like our nodes in the auxiliary search tree to be + * of fixed size. + * + * The solution is to make them fixed size, and when we're constructing a node + * check if p and n differed in the bits we needed them to. If they don't we + * flag that node, and when doing lookups we fallback to comparing against the + * real key. As long as this doesn't happen to often (and it seems to reliably + * happen a bit less than 1% of the time), we win - even on failures, that key + * is then more likely to be in cache than if we were doing binary searches all + * the way, since we're touching so much less memory. + * + * The keys in the auxiliary search tree are stored in (software) floating + * point, with an exponent and a mantissa. The exponent needs to be big enough + * to address all the bits in the original key, but the number of bits in the + * mantissa is somewhat arbitrary; more bits just gets us fewer failures. + * + * We need 7 bits for the exponent and 3 bits for the key's offset (since keys + * are 8 byte aligned); using 22 bits for the mantissa means a node is 4 bytes. + * We need one node per 128 bytes in the btree node, which means the auxiliary + * search trees take up 3% as much memory as the btree itself. + * + * Constructing these auxiliary search trees is moderately expensive, and we + * don't want to be constantly rebuilding the search tree for the last set + * whenever we insert another key into it. For the unwritten set, we use a much + * simpler lookup table - it's just a flat array, so index i in the lookup table + * corresponds to the i range of BSET_CACHELINE bytes in the set. Indexing + * within each byte range works the same as with the auxiliary search trees. + * + * These are much easier to keep up to date when we insert a key - we do it + * somewhat lazily; when we shift a key up we usually just increment the pointer + * to it, only when it would overflow do we go to the trouble of finding the + * first key in that range of bytes again. + */ + +struct btree_keys; +struct btree_iter; +struct btree_iter_set; +struct bkey_float; + +#define MAX_BSETS 4U + +struct bset_tree { + /* + * We construct a binary tree in an array as if the array + * started at 1, so that things line up on the same cachelines + * better: see comments in bset.c at cacheline_to_bkey() for + * details + */ + + /* size of the binary tree and prev array */ + unsigned int size; + + /* function of size - precalculated for to_inorder() */ + unsigned int extra; + + /* copy of the last key in the set */ + struct bkey end; + struct bkey_float *tree; + + /* + * The nodes in the bset tree point to specific keys - this + * array holds the sizes of the previous key. + * + * Conceptually it's a member of struct bkey_float, but we want + * to keep bkey_float to 4 bytes and prev isn't used in the fast + * path. + */ + uint8_t *prev; + + /* The actual btree node, with pointers to each sorted set */ + struct bset *data; +}; + +struct btree_keys_ops { + bool (*sort_cmp)(struct btree_iter_set l, + struct btree_iter_set r); + struct bkey *(*sort_fixup)(struct btree_iter *iter, + struct bkey *tmp); + bool (*insert_fixup)(struct btree_keys *b, + struct bkey *insert, + struct btree_iter *iter, + struct bkey *replace_key); + bool (*key_invalid)(struct btree_keys *bk, + const struct bkey *k); + bool (*key_bad)(struct btree_keys *bk, + const struct bkey *k); + bool (*key_merge)(struct btree_keys *bk, + struct bkey *l, struct bkey *r); + void (*key_to_text)(char *buf, + size_t size, + const struct bkey *k); + void (*key_dump)(struct btree_keys *keys, + const struct bkey *k); + + /* + * Only used for deciding whether to use START_KEY(k) or just the key + * itself in a couple places + */ + bool is_extents; +}; + +struct btree_keys { + const struct btree_keys_ops *ops; + uint8_t page_order; + uint8_t nsets; + unsigned int last_set_unwritten:1; + bool *expensive_debug_checks; + + /* + * Sets of sorted keys - the real btree node - plus a binary search tree + * + * set[0] is special; set[0]->tree, set[0]->prev and set[0]->data point + * to the memory we have allocated for this btree node. Additionally, + * set[0]->data points to the entire btree node as it exists on disk. + */ + struct bset_tree set[MAX_BSETS]; +}; + +static inline struct bset_tree *bset_tree_last(struct btree_keys *b) +{ + return b->set + b->nsets; +} + +static inline bool bset_written(struct btree_keys *b, struct bset_tree *t) +{ + return t <= b->set + b->nsets - b->last_set_unwritten; +} + +static inline bool bkey_written(struct btree_keys *b, struct bkey *k) +{ + return !b->last_set_unwritten || k < b->set[b->nsets].data->start; +} + +static inline unsigned int bset_byte_offset(struct btree_keys *b, + struct bset *i) +{ + return ((size_t) i) - ((size_t) b->set->data); +} + +static inline unsigned int bset_sector_offset(struct btree_keys *b, + struct bset *i) +{ + return bset_byte_offset(b, i) >> 9; +} + +#define __set_bytes(i, k) (sizeof(*(i)) + (k) * sizeof(uint64_t)) +#define set_bytes(i) __set_bytes(i, i->keys) + +#define __set_blocks(i, k, block_bytes) \ + DIV_ROUND_UP(__set_bytes(i, k), block_bytes) +#define set_blocks(i, block_bytes) \ + __set_blocks(i, (i)->keys, block_bytes) + +static inline size_t bch_btree_keys_u64s_remaining(struct btree_keys *b) +{ + struct bset_tree *t = bset_tree_last(b); + + BUG_ON((PAGE_SIZE << b->page_order) < + (bset_byte_offset(b, t->data) + set_bytes(t->data))); + + if (!b->last_set_unwritten) + return 0; + + return ((PAGE_SIZE << b->page_order) - + (bset_byte_offset(b, t->data) + set_bytes(t->data))) / + sizeof(u64); +} + +static inline struct bset *bset_next_set(struct btree_keys *b, + unsigned int block_bytes) +{ + struct bset *i = bset_tree_last(b)->data; + + return ((void *) i) + roundup(set_bytes(i), block_bytes); +} + +void bch_btree_keys_free(struct btree_keys *b); +int bch_btree_keys_alloc(struct btree_keys *b, unsigned int page_order, + gfp_t gfp); +void bch_btree_keys_init(struct btree_keys *b, const struct btree_keys_ops *ops, + bool *expensive_debug_checks); + +void bch_bset_init_next(struct btree_keys *b, struct bset *i, uint64_t magic); +void bch_bset_build_written_tree(struct btree_keys *b); +void bch_bset_fix_invalidated_key(struct btree_keys *b, struct bkey *k); +bool bch_bkey_try_merge(struct btree_keys *b, struct bkey *l, struct bkey *r); +void bch_bset_insert(struct btree_keys *b, struct bkey *where, + struct bkey *insert); +unsigned int bch_btree_insert_key(struct btree_keys *b, struct bkey *k, + struct bkey *replace_key); + +enum { + BTREE_INSERT_STATUS_NO_INSERT = 0, + BTREE_INSERT_STATUS_INSERT, + BTREE_INSERT_STATUS_BACK_MERGE, + BTREE_INSERT_STATUS_OVERWROTE, + BTREE_INSERT_STATUS_FRONT_MERGE, +}; + +/* Btree key iteration */ + +struct btree_iter { + size_t size, used; +#ifdef CONFIG_BCACHE_DEBUG + struct btree_keys *b; +#endif + struct btree_iter_set { + struct bkey *k, *end; + } data[MAX_BSETS]; +}; + +typedef bool (*ptr_filter_fn)(struct btree_keys *b, const struct bkey *k); + +struct bkey *bch_btree_iter_next(struct btree_iter *iter); +struct bkey *bch_btree_iter_next_filter(struct btree_iter *iter, + struct btree_keys *b, + ptr_filter_fn fn); + +void bch_btree_iter_push(struct btree_iter *iter, struct bkey *k, + struct bkey *end); +struct bkey *bch_btree_iter_init(struct btree_keys *b, + struct btree_iter *iter, + struct bkey *search); + +struct bkey *__bch_bset_search(struct btree_keys *b, struct bset_tree *t, + const struct bkey *search); + +/* + * Returns the first key that is strictly greater than search + */ +static inline struct bkey *bch_bset_search(struct btree_keys *b, + struct bset_tree *t, + const struct bkey *search) +{ + return search ? __bch_bset_search(b, t, search) : t->data->start; +} + +#define for_each_key_filter(b, k, iter, filter) \ + for (bch_btree_iter_init((b), (iter), NULL); \ + ((k) = bch_btree_iter_next_filter((iter), (b), filter));) + +#define for_each_key(b, k, iter) \ + for (bch_btree_iter_init((b), (iter), NULL); \ + ((k) = bch_btree_iter_next(iter));) + +/* Sorting */ + +struct bset_sort_state { + mempool_t pool; + + unsigned int page_order; + unsigned int crit_factor; + + struct time_stats time; +}; + +void bch_bset_sort_state_free(struct bset_sort_state *state); +int bch_bset_sort_state_init(struct bset_sort_state *state, + unsigned int page_order); +void bch_btree_sort_lazy(struct btree_keys *b, struct bset_sort_state *state); +void bch_btree_sort_into(struct btree_keys *b, struct btree_keys *new, + struct bset_sort_state *state); +void bch_btree_sort_and_fix_extents(struct btree_keys *b, + struct btree_iter *iter, + struct bset_sort_state *state); +void bch_btree_sort_partial(struct btree_keys *b, unsigned int start, + struct bset_sort_state *state); + +static inline void bch_btree_sort(struct btree_keys *b, + struct bset_sort_state *state) +{ + bch_btree_sort_partial(b, 0, state); +} + +struct bset_stats { + size_t sets_written, sets_unwritten; + size_t bytes_written, bytes_unwritten; + size_t floats, failed; +}; + +void bch_btree_keys_stats(struct btree_keys *b, struct bset_stats *state); + +/* Bkey utility code */ + +#define bset_bkey_last(i) bkey_idx((struct bkey *) (i)->d, \ + (unsigned int)(i)->keys) + +static inline struct bkey *bset_bkey_idx(struct bset *i, unsigned int idx) +{ + return bkey_idx(i->start, idx); +} + +static inline void bkey_init(struct bkey *k) +{ + *k = ZERO_KEY; +} + +static __always_inline int64_t bkey_cmp(const struct bkey *l, + const struct bkey *r) +{ + return unlikely(KEY_INODE(l) != KEY_INODE(r)) + ? (int64_t) KEY_INODE(l) - (int64_t) KEY_INODE(r) + : (int64_t) KEY_OFFSET(l) - (int64_t) KEY_OFFSET(r); +} + +void bch_bkey_copy_single_ptr(struct bkey *dest, const struct bkey *src, + unsigned int i); +bool __bch_cut_front(const struct bkey *where, struct bkey *k); +bool __bch_cut_back(const struct bkey *where, struct bkey *k); + +static inline bool bch_cut_front(const struct bkey *where, struct bkey *k) +{ + BUG_ON(bkey_cmp(where, k) > 0); + return __bch_cut_front(where, k); +} + +static inline bool bch_cut_back(const struct bkey *where, struct bkey *k) +{ + BUG_ON(bkey_cmp(where, &START_KEY(k)) < 0); + return __bch_cut_back(where, k); +} + +/* + * Pointer '*preceding_key_p' points to a memory object to store preceding + * key of k. If the preceding key does not exist, set '*preceding_key_p' to + * NULL. So the caller of preceding_key() needs to take care of memory + * which '*preceding_key_p' pointed to before calling preceding_key(). + * Currently the only caller of preceding_key() is bch_btree_insert_key(), + * and it points to an on-stack variable, so the memory release is handled + * by stackframe itself. + */ +static inline void preceding_key(struct bkey *k, struct bkey **preceding_key_p) +{ + if (KEY_INODE(k) || KEY_OFFSET(k)) { + (**preceding_key_p) = KEY(KEY_INODE(k), KEY_OFFSET(k), 0); + if (!(*preceding_key_p)->low) + (*preceding_key_p)->high--; + (*preceding_key_p)->low--; + } else { + (*preceding_key_p) = NULL; + } +} + +static inline bool bch_ptr_invalid(struct btree_keys *b, const struct bkey *k) +{ + return b->ops->key_invalid(b, k); +} + +static inline bool bch_ptr_bad(struct btree_keys *b, const struct bkey *k) +{ + return b->ops->key_bad(b, k); +} + +static inline void bch_bkey_to_text(struct btree_keys *b, char *buf, + size_t size, const struct bkey *k) +{ + return b->ops->key_to_text(buf, size, k); +} + +static inline bool bch_bkey_equal_header(const struct bkey *l, + const struct bkey *r) +{ + return (KEY_DIRTY(l) == KEY_DIRTY(r) && + KEY_PTRS(l) == KEY_PTRS(r) && + KEY_CSUM(l) == KEY_CSUM(r)); +} + +/* Keylists */ + +struct keylist { + union { + struct bkey *keys; + uint64_t *keys_p; + }; + union { + struct bkey *top; + uint64_t *top_p; + }; + + /* Enough room for btree_split's keys without realloc */ +#define KEYLIST_INLINE 16 + uint64_t inline_keys[KEYLIST_INLINE]; +}; + +static inline void bch_keylist_init(struct keylist *l) +{ + l->top_p = l->keys_p = l->inline_keys; +} + +static inline void bch_keylist_init_single(struct keylist *l, struct bkey *k) +{ + l->keys = k; + l->top = bkey_next(k); +} + +static inline void bch_keylist_push(struct keylist *l) +{ + l->top = bkey_next(l->top); +} + +static inline void bch_keylist_add(struct keylist *l, struct bkey *k) +{ + bkey_copy(l->top, k); + bch_keylist_push(l); +} + +static inline bool bch_keylist_empty(struct keylist *l) +{ + return l->top == l->keys; +} + +static inline void bch_keylist_reset(struct keylist *l) +{ + l->top = l->keys; +} + +static inline void bch_keylist_free(struct keylist *l) +{ + if (l->keys_p != l->inline_keys) + kfree(l->keys_p); +} + +static inline size_t bch_keylist_nkeys(struct keylist *l) +{ + return l->top_p - l->keys_p; +} + +static inline size_t bch_keylist_bytes(struct keylist *l) +{ + return bch_keylist_nkeys(l) * sizeof(uint64_t); +} + +struct bkey *bch_keylist_pop(struct keylist *l); +void bch_keylist_pop_front(struct keylist *l); +int __bch_keylist_realloc(struct keylist *l, unsigned int u64s); + +/* Debug stuff */ + +#ifdef CONFIG_BCACHE_DEBUG + +int __bch_count_data(struct btree_keys *b); +void __printf(2, 3) __bch_check_keys(struct btree_keys *b, + const char *fmt, + ...); +void bch_dump_bset(struct btree_keys *b, struct bset *i, unsigned int set); +void bch_dump_bucket(struct btree_keys *b); + +#else + +static inline int __bch_count_data(struct btree_keys *b) { return -1; } +static inline void __printf(2, 3) + __bch_check_keys(struct btree_keys *b, const char *fmt, ...) {} +static inline void bch_dump_bucket(struct btree_keys *b) {} +void bch_dump_bset(struct btree_keys *b, struct bset *i, unsigned int set); + +#endif + +static inline bool btree_keys_expensive_checks(struct btree_keys *b) +{ +#ifdef CONFIG_BCACHE_DEBUG + return *b->expensive_debug_checks; +#else + return false; +#endif +} + +static inline int bch_count_data(struct btree_keys *b) +{ + return btree_keys_expensive_checks(b) ? __bch_count_data(b) : -1; +} + +#define bch_check_keys(b, ...) \ +do { \ + if (btree_keys_expensive_checks(b)) \ + __bch_check_keys(b, __VA_ARGS__); \ +} while (0) + +#endif diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c new file mode 100644 index 000000000..1a1a95544 --- /dev/null +++ b/drivers/md/bcache/btree.c @@ -0,0 +1,2800 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2010 Kent Overstreet <kent.overstreet@gmail.com> + * + * Uses a block device as cache for other block devices; optimized for SSDs. + * All allocation is done in buckets, which should match the erase block size + * of the device. + * + * Buckets containing cached data are kept on a heap sorted by priority; + * bucket priority is increased on cache hit, and periodically all the buckets + * on the heap have their priority scaled down. This currently is just used as + * an LRU but in the future should allow for more intelligent heuristics. + * + * Buckets have an 8 bit counter; freeing is accomplished by incrementing the + * counter. Garbage collection is used to remove stale pointers. + * + * Indexing is done via a btree; nodes are not necessarily fully sorted, rather + * as keys are inserted we only sort the pages that have not yet been written. + * When garbage collection is run, we resort the entire node. + * + * All configuration is done via sysfs; see Documentation/admin-guide/bcache.rst. + */ + +#include "bcache.h" +#include "btree.h" +#include "debug.h" +#include "extents.h" + +#include <linux/slab.h> +#include <linux/bitops.h> +#include <linux/hash.h> +#include <linux/kthread.h> +#include <linux/prefetch.h> +#include <linux/random.h> +#include <linux/rcupdate.h> +#include <linux/sched/clock.h> +#include <linux/rculist.h> +#include <linux/delay.h> +#include <trace/events/bcache.h> + +/* + * Todo: + * register_bcache: Return errors out to userspace correctly + * + * Writeback: don't undirty key until after a cache flush + * + * Create an iterator for key pointers + * + * On btree write error, mark bucket such that it won't be freed from the cache + * + * Journalling: + * Check for bad keys in replay + * Propagate barriers + * Refcount journal entries in journal_replay + * + * Garbage collection: + * Finish incremental gc + * Gc should free old UUIDs, data for invalid UUIDs + * + * Provide a way to list backing device UUIDs we have data cached for, and + * probably how long it's been since we've seen them, and a way to invalidate + * dirty data for devices that will never be attached again + * + * Keep 1 min/5 min/15 min statistics of how busy a block device has been, so + * that based on that and how much dirty data we have we can keep writeback + * from being starved + * + * Add a tracepoint or somesuch to watch for writeback starvation + * + * When btree depth > 1 and splitting an interior node, we have to make sure + * alloc_bucket() cannot fail. This should be true but is not completely + * obvious. + * + * Plugging? + * + * If data write is less than hard sector size of ssd, round up offset in open + * bucket to the next whole sector + * + * Superblock needs to be fleshed out for multiple cache devices + * + * Add a sysfs tunable for the number of writeback IOs in flight + * + * Add a sysfs tunable for the number of open data buckets + * + * IO tracking: Can we track when one process is doing io on behalf of another? + * IO tracking: Don't use just an average, weigh more recent stuff higher + * + * Test module load/unload + */ + +#define MAX_NEED_GC 64 +#define MAX_SAVE_PRIO 72 +#define MAX_GC_TIMES 100 +#define MIN_GC_NODES 100 +#define GC_SLEEP_MS 100 + +#define PTR_DIRTY_BIT (((uint64_t) 1 << 36)) + +#define PTR_HASH(c, k) \ + (((k)->ptr[0] >> c->bucket_bits) | PTR_GEN(k, 0)) + +static struct workqueue_struct *btree_io_wq; + +#define insert_lock(s, b) ((b)->level <= (s)->lock) + + +static inline struct bset *write_block(struct btree *b) +{ + return ((void *) btree_bset_first(b)) + b->written * block_bytes(b->c->cache); +} + +static void bch_btree_init_next(struct btree *b) +{ + /* If not a leaf node, always sort */ + if (b->level && b->keys.nsets) + bch_btree_sort(&b->keys, &b->c->sort); + else + bch_btree_sort_lazy(&b->keys, &b->c->sort); + + if (b->written < btree_blocks(b)) + bch_bset_init_next(&b->keys, write_block(b), + bset_magic(&b->c->cache->sb)); + +} + +/* Btree key manipulation */ + +void bkey_put(struct cache_set *c, struct bkey *k) +{ + unsigned int i; + + for (i = 0; i < KEY_PTRS(k); i++) + if (ptr_available(c, k, i)) + atomic_dec_bug(&PTR_BUCKET(c, k, i)->pin); +} + +/* Btree IO */ + +static uint64_t btree_csum_set(struct btree *b, struct bset *i) +{ + uint64_t crc = b->key.ptr[0]; + void *data = (void *) i + 8, *end = bset_bkey_last(i); + + crc = bch_crc64_update(crc, data, end - data); + return crc ^ 0xffffffffffffffffULL; +} + +void bch_btree_node_read_done(struct btree *b) +{ + const char *err = "bad btree header"; + struct bset *i = btree_bset_first(b); + struct btree_iter *iter; + + /* + * c->fill_iter can allocate an iterator with more memory space + * than static MAX_BSETS. + * See the comment arount cache_set->fill_iter. + */ + iter = mempool_alloc(&b->c->fill_iter, GFP_NOIO); + iter->size = b->c->cache->sb.bucket_size / b->c->cache->sb.block_size; + iter->used = 0; + +#ifdef CONFIG_BCACHE_DEBUG + iter->b = &b->keys; +#endif + + if (!i->seq) + goto err; + + for (; + b->written < btree_blocks(b) && i->seq == b->keys.set[0].data->seq; + i = write_block(b)) { + err = "unsupported bset version"; + if (i->version > BCACHE_BSET_VERSION) + goto err; + + err = "bad btree header"; + if (b->written + set_blocks(i, block_bytes(b->c->cache)) > + btree_blocks(b)) + goto err; + + err = "bad magic"; + if (i->magic != bset_magic(&b->c->cache->sb)) + goto err; + + err = "bad checksum"; + switch (i->version) { + case 0: + if (i->csum != csum_set(i)) + goto err; + break; + case BCACHE_BSET_VERSION: + if (i->csum != btree_csum_set(b, i)) + goto err; + break; + } + + err = "empty set"; + if (i != b->keys.set[0].data && !i->keys) + goto err; + + bch_btree_iter_push(iter, i->start, bset_bkey_last(i)); + + b->written += set_blocks(i, block_bytes(b->c->cache)); + } + + err = "corrupted btree"; + for (i = write_block(b); + bset_sector_offset(&b->keys, i) < KEY_SIZE(&b->key); + i = ((void *) i) + block_bytes(b->c->cache)) + if (i->seq == b->keys.set[0].data->seq) + goto err; + + bch_btree_sort_and_fix_extents(&b->keys, iter, &b->c->sort); + + i = b->keys.set[0].data; + err = "short btree key"; + if (b->keys.set[0].size && + bkey_cmp(&b->key, &b->keys.set[0].end) < 0) + goto err; + + if (b->written < btree_blocks(b)) + bch_bset_init_next(&b->keys, write_block(b), + bset_magic(&b->c->cache->sb)); +out: + mempool_free(iter, &b->c->fill_iter); + return; +err: + set_btree_node_io_error(b); + bch_cache_set_error(b->c, "%s at bucket %zu, block %u, %u keys", + err, PTR_BUCKET_NR(b->c, &b->key, 0), + bset_block_offset(b, i), i->keys); + goto out; +} + +static void btree_node_read_endio(struct bio *bio) +{ + struct closure *cl = bio->bi_private; + + closure_put(cl); +} + +static void bch_btree_node_read(struct btree *b) +{ + uint64_t start_time = local_clock(); + struct closure cl; + struct bio *bio; + + trace_bcache_btree_read(b); + + closure_init_stack(&cl); + + bio = bch_bbio_alloc(b->c); + bio->bi_iter.bi_size = KEY_SIZE(&b->key) << 9; + bio->bi_end_io = btree_node_read_endio; + bio->bi_private = &cl; + bio->bi_opf = REQ_OP_READ | REQ_META; + + bch_bio_map(bio, b->keys.set[0].data); + + bch_submit_bbio(bio, b->c, &b->key, 0); + closure_sync(&cl); + + if (bio->bi_status) + set_btree_node_io_error(b); + + bch_bbio_free(bio, b->c); + + if (btree_node_io_error(b)) + goto err; + + bch_btree_node_read_done(b); + bch_time_stats_update(&b->c->btree_read_time, start_time); + + return; +err: + bch_cache_set_error(b->c, "io error reading bucket %zu", + PTR_BUCKET_NR(b->c, &b->key, 0)); +} + +static void btree_complete_write(struct btree *b, struct btree_write *w) +{ + if (w->prio_blocked && + !atomic_sub_return(w->prio_blocked, &b->c->prio_blocked)) + wake_up_allocators(b->c); + + if (w->journal) { + atomic_dec_bug(w->journal); + __closure_wake_up(&b->c->journal.wait); + } + + w->prio_blocked = 0; + w->journal = NULL; +} + +static void btree_node_write_unlock(struct closure *cl) +{ + struct btree *b = container_of(cl, struct btree, io); + + up(&b->io_mutex); +} + +static void __btree_node_write_done(struct closure *cl) +{ + struct btree *b = container_of(cl, struct btree, io); + struct btree_write *w = btree_prev_write(b); + + bch_bbio_free(b->bio, b->c); + b->bio = NULL; + btree_complete_write(b, w); + + if (btree_node_dirty(b)) + queue_delayed_work(btree_io_wq, &b->work, 30 * HZ); + + closure_return_with_destructor(cl, btree_node_write_unlock); +} + +static void btree_node_write_done(struct closure *cl) +{ + struct btree *b = container_of(cl, struct btree, io); + + bio_free_pages(b->bio); + __btree_node_write_done(cl); +} + +static void btree_node_write_endio(struct bio *bio) +{ + struct closure *cl = bio->bi_private; + struct btree *b = container_of(cl, struct btree, io); + + if (bio->bi_status) + set_btree_node_io_error(b); + + bch_bbio_count_io_errors(b->c, bio, bio->bi_status, "writing btree"); + closure_put(cl); +} + +static void do_btree_node_write(struct btree *b) +{ + struct closure *cl = &b->io; + struct bset *i = btree_bset_last(b); + BKEY_PADDED(key) k; + + i->version = BCACHE_BSET_VERSION; + i->csum = btree_csum_set(b, i); + + BUG_ON(b->bio); + b->bio = bch_bbio_alloc(b->c); + + b->bio->bi_end_io = btree_node_write_endio; + b->bio->bi_private = cl; + b->bio->bi_iter.bi_size = roundup(set_bytes(i), block_bytes(b->c->cache)); + b->bio->bi_opf = REQ_OP_WRITE | REQ_META | REQ_FUA; + bch_bio_map(b->bio, i); + + /* + * If we're appending to a leaf node, we don't technically need FUA - + * this write just needs to be persisted before the next journal write, + * which will be marked FLUSH|FUA. + * + * Similarly if we're writing a new btree root - the pointer is going to + * be in the next journal entry. + * + * But if we're writing a new btree node (that isn't a root) or + * appending to a non leaf btree node, we need either FUA or a flush + * when we write the parent with the new pointer. FUA is cheaper than a + * flush, and writes appending to leaf nodes aren't blocking anything so + * just make all btree node writes FUA to keep things sane. + */ + + bkey_copy(&k.key, &b->key); + SET_PTR_OFFSET(&k.key, 0, PTR_OFFSET(&k.key, 0) + + bset_sector_offset(&b->keys, i)); + + if (!bch_bio_alloc_pages(b->bio, __GFP_NOWARN|GFP_NOWAIT)) { + struct bio_vec *bv; + void *addr = (void *) ((unsigned long) i & ~(PAGE_SIZE - 1)); + struct bvec_iter_all iter_all; + + bio_for_each_segment_all(bv, b->bio, iter_all) { + memcpy(page_address(bv->bv_page), addr, PAGE_SIZE); + addr += PAGE_SIZE; + } + + bch_submit_bbio(b->bio, b->c, &k.key, 0); + + continue_at(cl, btree_node_write_done, NULL); + } else { + /* + * No problem for multipage bvec since the bio is + * just allocated + */ + b->bio->bi_vcnt = 0; + bch_bio_map(b->bio, i); + + bch_submit_bbio(b->bio, b->c, &k.key, 0); + + closure_sync(cl); + continue_at_nobarrier(cl, __btree_node_write_done, NULL); + } +} + +void __bch_btree_node_write(struct btree *b, struct closure *parent) +{ + struct bset *i = btree_bset_last(b); + + lockdep_assert_held(&b->write_lock); + + trace_bcache_btree_write(b); + + BUG_ON(current->bio_list); + BUG_ON(b->written >= btree_blocks(b)); + BUG_ON(b->written && !i->keys); + BUG_ON(btree_bset_first(b)->seq != i->seq); + bch_check_keys(&b->keys, "writing"); + + cancel_delayed_work(&b->work); + + /* If caller isn't waiting for write, parent refcount is cache set */ + down(&b->io_mutex); + closure_init(&b->io, parent ?: &b->c->cl); + + clear_bit(BTREE_NODE_dirty, &b->flags); + change_bit(BTREE_NODE_write_idx, &b->flags); + + do_btree_node_write(b); + + atomic_long_add(set_blocks(i, block_bytes(b->c->cache)) * b->c->cache->sb.block_size, + &PTR_CACHE(b->c, &b->key, 0)->btree_sectors_written); + + b->written += set_blocks(i, block_bytes(b->c->cache)); +} + +void bch_btree_node_write(struct btree *b, struct closure *parent) +{ + unsigned int nsets = b->keys.nsets; + + lockdep_assert_held(&b->lock); + + __bch_btree_node_write(b, parent); + + /* + * do verify if there was more than one set initially (i.e. we did a + * sort) and we sorted down to a single set: + */ + if (nsets && !b->keys.nsets) + bch_btree_verify(b); + + bch_btree_init_next(b); +} + +static void bch_btree_node_write_sync(struct btree *b) +{ + struct closure cl; + + closure_init_stack(&cl); + + mutex_lock(&b->write_lock); + bch_btree_node_write(b, &cl); + mutex_unlock(&b->write_lock); + + closure_sync(&cl); +} + +static void btree_node_write_work(struct work_struct *w) +{ + struct btree *b = container_of(to_delayed_work(w), struct btree, work); + + mutex_lock(&b->write_lock); + if (btree_node_dirty(b)) + __bch_btree_node_write(b, NULL); + mutex_unlock(&b->write_lock); +} + +static void bch_btree_leaf_dirty(struct btree *b, atomic_t *journal_ref) +{ + struct bset *i = btree_bset_last(b); + struct btree_write *w = btree_current_write(b); + + lockdep_assert_held(&b->write_lock); + + BUG_ON(!b->written); + BUG_ON(!i->keys); + + if (!btree_node_dirty(b)) + queue_delayed_work(btree_io_wq, &b->work, 30 * HZ); + + set_btree_node_dirty(b); + + /* + * w->journal is always the oldest journal pin of all bkeys + * in the leaf node, to make sure the oldest jset seq won't + * be increased before this btree node is flushed. + */ + if (journal_ref) { + if (w->journal && + journal_pin_cmp(b->c, w->journal, journal_ref)) { + atomic_dec_bug(w->journal); + w->journal = NULL; + } + + if (!w->journal) { + w->journal = journal_ref; + atomic_inc(w->journal); + } + } + + /* Force write if set is too big */ + if (set_bytes(i) > PAGE_SIZE - 48 && + !current->bio_list) + bch_btree_node_write(b, NULL); +} + +/* + * Btree in memory cache - allocation/freeing + * mca -> memory cache + */ + +#define mca_reserve(c) (((!IS_ERR_OR_NULL(c->root) && c->root->level) \ + ? c->root->level : 1) * 8 + 16) +#define mca_can_free(c) \ + max_t(int, 0, c->btree_cache_used - mca_reserve(c)) + +static void mca_data_free(struct btree *b) +{ + BUG_ON(b->io_mutex.count != 1); + + bch_btree_keys_free(&b->keys); + + b->c->btree_cache_used--; + list_move(&b->list, &b->c->btree_cache_freed); +} + +static void mca_bucket_free(struct btree *b) +{ + BUG_ON(btree_node_dirty(b)); + + b->key.ptr[0] = 0; + hlist_del_init_rcu(&b->hash); + list_move(&b->list, &b->c->btree_cache_freeable); +} + +static unsigned int btree_order(struct bkey *k) +{ + return ilog2(KEY_SIZE(k) / PAGE_SECTORS ?: 1); +} + +static void mca_data_alloc(struct btree *b, struct bkey *k, gfp_t gfp) +{ + if (!bch_btree_keys_alloc(&b->keys, + max_t(unsigned int, + ilog2(b->c->btree_pages), + btree_order(k)), + gfp)) { + b->c->btree_cache_used++; + list_move(&b->list, &b->c->btree_cache); + } else { + list_move(&b->list, &b->c->btree_cache_freed); + } +} + +static struct btree *mca_bucket_alloc(struct cache_set *c, + struct bkey *k, gfp_t gfp) +{ + /* + * kzalloc() is necessary here for initialization, + * see code comments in bch_btree_keys_init(). + */ + struct btree *b = kzalloc(sizeof(struct btree), gfp); + + if (!b) + return NULL; + + init_rwsem(&b->lock); + lockdep_set_novalidate_class(&b->lock); + mutex_init(&b->write_lock); + lockdep_set_novalidate_class(&b->write_lock); + INIT_LIST_HEAD(&b->list); + INIT_DELAYED_WORK(&b->work, btree_node_write_work); + b->c = c; + sema_init(&b->io_mutex, 1); + + mca_data_alloc(b, k, gfp); + return b; +} + +static int mca_reap(struct btree *b, unsigned int min_order, bool flush) +{ + struct closure cl; + + closure_init_stack(&cl); + lockdep_assert_held(&b->c->bucket_lock); + + if (!down_write_trylock(&b->lock)) + return -ENOMEM; + + BUG_ON(btree_node_dirty(b) && !b->keys.set[0].data); + + if (b->keys.page_order < min_order) + goto out_unlock; + + if (!flush) { + if (btree_node_dirty(b)) + goto out_unlock; + + if (down_trylock(&b->io_mutex)) + goto out_unlock; + up(&b->io_mutex); + } + +retry: + /* + * BTREE_NODE_dirty might be cleared in btree_flush_btree() by + * __bch_btree_node_write(). To avoid an extra flush, acquire + * b->write_lock before checking BTREE_NODE_dirty bit. + */ + mutex_lock(&b->write_lock); + /* + * If this btree node is selected in btree_flush_write() by journal + * code, delay and retry until the node is flushed by journal code + * and BTREE_NODE_journal_flush bit cleared by btree_flush_write(). + */ + if (btree_node_journal_flush(b)) { + pr_debug("bnode %p is flushing by journal, retry\n", b); + mutex_unlock(&b->write_lock); + udelay(1); + goto retry; + } + + if (btree_node_dirty(b)) + __bch_btree_node_write(b, &cl); + mutex_unlock(&b->write_lock); + + closure_sync(&cl); + + /* wait for any in flight btree write */ + down(&b->io_mutex); + up(&b->io_mutex); + + return 0; +out_unlock: + rw_unlock(true, b); + return -ENOMEM; +} + +static unsigned long bch_mca_scan(struct shrinker *shrink, + struct shrink_control *sc) +{ + struct cache_set *c = container_of(shrink, struct cache_set, shrink); + struct btree *b, *t; + unsigned long i, nr = sc->nr_to_scan; + unsigned long freed = 0; + unsigned int btree_cache_used; + + if (c->shrinker_disabled) + return SHRINK_STOP; + + if (c->btree_cache_alloc_lock) + return SHRINK_STOP; + + /* Return -1 if we can't do anything right now */ + if (sc->gfp_mask & __GFP_IO) + mutex_lock(&c->bucket_lock); + else if (!mutex_trylock(&c->bucket_lock)) + return -1; + + /* + * It's _really_ critical that we don't free too many btree nodes - we + * have to always leave ourselves a reserve. The reserve is how we + * guarantee that allocating memory for a new btree node can always + * succeed, so that inserting keys into the btree can always succeed and + * IO can always make forward progress: + */ + nr /= c->btree_pages; + if (nr == 0) + nr = 1; + nr = min_t(unsigned long, nr, mca_can_free(c)); + + i = 0; + btree_cache_used = c->btree_cache_used; + list_for_each_entry_safe_reverse(b, t, &c->btree_cache_freeable, list) { + if (nr <= 0) + goto out; + + if (!mca_reap(b, 0, false)) { + mca_data_free(b); + rw_unlock(true, b); + freed++; + } + nr--; + i++; + } + + list_for_each_entry_safe_reverse(b, t, &c->btree_cache, list) { + if (nr <= 0 || i >= btree_cache_used) + goto out; + + if (!mca_reap(b, 0, false)) { + mca_bucket_free(b); + mca_data_free(b); + rw_unlock(true, b); + freed++; + } + + nr--; + i++; + } +out: + mutex_unlock(&c->bucket_lock); + return freed * c->btree_pages; +} + +static unsigned long bch_mca_count(struct shrinker *shrink, + struct shrink_control *sc) +{ + struct cache_set *c = container_of(shrink, struct cache_set, shrink); + + if (c->shrinker_disabled) + return 0; + + if (c->btree_cache_alloc_lock) + return 0; + + return mca_can_free(c) * c->btree_pages; +} + +void bch_btree_cache_free(struct cache_set *c) +{ + struct btree *b; + struct closure cl; + + closure_init_stack(&cl); + + if (c->shrink.list.next) + unregister_shrinker(&c->shrink); + + mutex_lock(&c->bucket_lock); + +#ifdef CONFIG_BCACHE_DEBUG + if (c->verify_data) + list_move(&c->verify_data->list, &c->btree_cache); + + free_pages((unsigned long) c->verify_ondisk, ilog2(meta_bucket_pages(&c->cache->sb))); +#endif + + list_splice(&c->btree_cache_freeable, + &c->btree_cache); + + while (!list_empty(&c->btree_cache)) { + b = list_first_entry(&c->btree_cache, struct btree, list); + + /* + * This function is called by cache_set_free(), no I/O + * request on cache now, it is unnecessary to acquire + * b->write_lock before clearing BTREE_NODE_dirty anymore. + */ + if (btree_node_dirty(b)) { + btree_complete_write(b, btree_current_write(b)); + clear_bit(BTREE_NODE_dirty, &b->flags); + } + mca_data_free(b); + } + + while (!list_empty(&c->btree_cache_freed)) { + b = list_first_entry(&c->btree_cache_freed, + struct btree, list); + list_del(&b->list); + cancel_delayed_work_sync(&b->work); + kfree(b); + } + + mutex_unlock(&c->bucket_lock); +} + +int bch_btree_cache_alloc(struct cache_set *c) +{ + unsigned int i; + + for (i = 0; i < mca_reserve(c); i++) + if (!mca_bucket_alloc(c, &ZERO_KEY, GFP_KERNEL)) + return -ENOMEM; + + list_splice_init(&c->btree_cache, + &c->btree_cache_freeable); + +#ifdef CONFIG_BCACHE_DEBUG + mutex_init(&c->verify_lock); + + c->verify_ondisk = (void *) + __get_free_pages(GFP_KERNEL|__GFP_COMP, + ilog2(meta_bucket_pages(&c->cache->sb))); + if (!c->verify_ondisk) { + /* + * Don't worry about the mca_rereserve buckets + * allocated in previous for-loop, they will be + * handled properly in bch_cache_set_unregister(). + */ + return -ENOMEM; + } + + c->verify_data = mca_bucket_alloc(c, &ZERO_KEY, GFP_KERNEL); + + if (c->verify_data && + c->verify_data->keys.set->data) + list_del_init(&c->verify_data->list); + else + c->verify_data = NULL; +#endif + + c->shrink.count_objects = bch_mca_count; + c->shrink.scan_objects = bch_mca_scan; + c->shrink.seeks = 4; + c->shrink.batch = c->btree_pages * 2; + + if (register_shrinker(&c->shrink)) + pr_warn("bcache: %s: could not register shrinker\n", + __func__); + + return 0; +} + +/* Btree in memory cache - hash table */ + +static struct hlist_head *mca_hash(struct cache_set *c, struct bkey *k) +{ + return &c->bucket_hash[hash_32(PTR_HASH(c, k), BUCKET_HASH_BITS)]; +} + +static struct btree *mca_find(struct cache_set *c, struct bkey *k) +{ + struct btree *b; + + rcu_read_lock(); + hlist_for_each_entry_rcu(b, mca_hash(c, k), hash) + if (PTR_HASH(c, &b->key) == PTR_HASH(c, k)) + goto out; + b = NULL; +out: + rcu_read_unlock(); + return b; +} + +static int mca_cannibalize_lock(struct cache_set *c, struct btree_op *op) +{ + spin_lock(&c->btree_cannibalize_lock); + if (likely(c->btree_cache_alloc_lock == NULL)) { + c->btree_cache_alloc_lock = current; + } else if (c->btree_cache_alloc_lock != current) { + if (op) + prepare_to_wait(&c->btree_cache_wait, &op->wait, + TASK_UNINTERRUPTIBLE); + spin_unlock(&c->btree_cannibalize_lock); + return -EINTR; + } + spin_unlock(&c->btree_cannibalize_lock); + + return 0; +} + +static struct btree *mca_cannibalize(struct cache_set *c, struct btree_op *op, + struct bkey *k) +{ + struct btree *b; + + trace_bcache_btree_cache_cannibalize(c); + + if (mca_cannibalize_lock(c, op)) + return ERR_PTR(-EINTR); + + list_for_each_entry_reverse(b, &c->btree_cache, list) + if (!mca_reap(b, btree_order(k), false)) + return b; + + list_for_each_entry_reverse(b, &c->btree_cache, list) + if (!mca_reap(b, btree_order(k), true)) + return b; + + WARN(1, "btree cache cannibalize failed\n"); + return ERR_PTR(-ENOMEM); +} + +/* + * We can only have one thread cannibalizing other cached btree nodes at a time, + * or we'll deadlock. We use an open coded mutex to ensure that, which a + * cannibalize_bucket() will take. This means every time we unlock the root of + * the btree, we need to release this lock if we have it held. + */ +void bch_cannibalize_unlock(struct cache_set *c) +{ + spin_lock(&c->btree_cannibalize_lock); + if (c->btree_cache_alloc_lock == current) { + c->btree_cache_alloc_lock = NULL; + wake_up(&c->btree_cache_wait); + } + spin_unlock(&c->btree_cannibalize_lock); +} + +static struct btree *mca_alloc(struct cache_set *c, struct btree_op *op, + struct bkey *k, int level) +{ + struct btree *b; + + BUG_ON(current->bio_list); + + lockdep_assert_held(&c->bucket_lock); + + if (mca_find(c, k)) + return NULL; + + /* btree_free() doesn't free memory; it sticks the node on the end of + * the list. Check if there's any freed nodes there: + */ + list_for_each_entry(b, &c->btree_cache_freeable, list) + if (!mca_reap(b, btree_order(k), false)) + goto out; + + /* We never free struct btree itself, just the memory that holds the on + * disk node. Check the freed list before allocating a new one: + */ + list_for_each_entry(b, &c->btree_cache_freed, list) + if (!mca_reap(b, 0, false)) { + mca_data_alloc(b, k, __GFP_NOWARN|GFP_NOIO); + if (!b->keys.set[0].data) + goto err; + else + goto out; + } + + b = mca_bucket_alloc(c, k, __GFP_NOWARN|GFP_NOIO); + if (!b) + goto err; + + BUG_ON(!down_write_trylock(&b->lock)); + if (!b->keys.set->data) + goto err; +out: + BUG_ON(b->io_mutex.count != 1); + + bkey_copy(&b->key, k); + list_move(&b->list, &c->btree_cache); + hlist_del_init_rcu(&b->hash); + hlist_add_head_rcu(&b->hash, mca_hash(c, k)); + + lock_set_subclass(&b->lock.dep_map, level + 1, _THIS_IP_); + b->parent = (void *) ~0UL; + b->flags = 0; + b->written = 0; + b->level = level; + + if (!b->level) + bch_btree_keys_init(&b->keys, &bch_extent_keys_ops, + &b->c->expensive_debug_checks); + else + bch_btree_keys_init(&b->keys, &bch_btree_keys_ops, + &b->c->expensive_debug_checks); + + return b; +err: + if (b) + rw_unlock(true, b); + + b = mca_cannibalize(c, op, k); + if (!IS_ERR(b)) + goto out; + + return b; +} + +/* + * bch_btree_node_get - find a btree node in the cache and lock it, reading it + * in from disk if necessary. + * + * If IO is necessary and running under submit_bio_noacct, returns -EAGAIN. + * + * The btree node will have either a read or a write lock held, depending on + * level and op->lock. + * + * Note: Only error code or btree pointer will be returned, it is unncessary + * for callers to check NULL pointer. + */ +struct btree *bch_btree_node_get(struct cache_set *c, struct btree_op *op, + struct bkey *k, int level, bool write, + struct btree *parent) +{ + int i = 0; + struct btree *b; + + BUG_ON(level < 0); +retry: + b = mca_find(c, k); + + if (!b) { + if (current->bio_list) + return ERR_PTR(-EAGAIN); + + mutex_lock(&c->bucket_lock); + b = mca_alloc(c, op, k, level); + mutex_unlock(&c->bucket_lock); + + if (!b) + goto retry; + if (IS_ERR(b)) + return b; + + bch_btree_node_read(b); + + if (!write) + downgrade_write(&b->lock); + } else { + rw_lock(write, b, level); + if (PTR_HASH(c, &b->key) != PTR_HASH(c, k)) { + rw_unlock(write, b); + goto retry; + } + BUG_ON(b->level != level); + } + + if (btree_node_io_error(b)) { + rw_unlock(write, b); + return ERR_PTR(-EIO); + } + + BUG_ON(!b->written); + + b->parent = parent; + + for (; i <= b->keys.nsets && b->keys.set[i].size; i++) { + prefetch(b->keys.set[i].tree); + prefetch(b->keys.set[i].data); + } + + for (; i <= b->keys.nsets; i++) + prefetch(b->keys.set[i].data); + + return b; +} + +static void btree_node_prefetch(struct btree *parent, struct bkey *k) +{ + struct btree *b; + + mutex_lock(&parent->c->bucket_lock); + b = mca_alloc(parent->c, NULL, k, parent->level - 1); + mutex_unlock(&parent->c->bucket_lock); + + if (!IS_ERR_OR_NULL(b)) { + b->parent = parent; + bch_btree_node_read(b); + rw_unlock(true, b); + } +} + +/* Btree alloc */ + +static void btree_node_free(struct btree *b) +{ + trace_bcache_btree_node_free(b); + + BUG_ON(b == b->c->root); + +retry: + mutex_lock(&b->write_lock); + /* + * If the btree node is selected and flushing in btree_flush_write(), + * delay and retry until the BTREE_NODE_journal_flush bit cleared, + * then it is safe to free the btree node here. Otherwise this btree + * node will be in race condition. + */ + if (btree_node_journal_flush(b)) { + mutex_unlock(&b->write_lock); + pr_debug("bnode %p journal_flush set, retry\n", b); + udelay(1); + goto retry; + } + + if (btree_node_dirty(b)) { + btree_complete_write(b, btree_current_write(b)); + clear_bit(BTREE_NODE_dirty, &b->flags); + } + + mutex_unlock(&b->write_lock); + + cancel_delayed_work(&b->work); + + mutex_lock(&b->c->bucket_lock); + bch_bucket_free(b->c, &b->key); + mca_bucket_free(b); + mutex_unlock(&b->c->bucket_lock); +} + +/* + * Only error code or btree pointer will be returned, it is unncessary for + * callers to check NULL pointer. + */ +struct btree *__bch_btree_node_alloc(struct cache_set *c, struct btree_op *op, + int level, bool wait, + struct btree *parent) +{ + BKEY_PADDED(key) k; + struct btree *b; + + mutex_lock(&c->bucket_lock); +retry: + /* return ERR_PTR(-EAGAIN) when it fails */ + b = ERR_PTR(-EAGAIN); + if (__bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, wait)) + goto err; + + bkey_put(c, &k.key); + SET_KEY_SIZE(&k.key, c->btree_pages * PAGE_SECTORS); + + b = mca_alloc(c, op, &k.key, level); + if (IS_ERR(b)) + goto err_free; + + if (!b) { + cache_bug(c, + "Tried to allocate bucket that was in btree cache"); + goto retry; + } + + b->parent = parent; + bch_bset_init_next(&b->keys, b->keys.set->data, bset_magic(&b->c->cache->sb)); + + mutex_unlock(&c->bucket_lock); + + trace_bcache_btree_node_alloc(b); + return b; +err_free: + bch_bucket_free(c, &k.key); +err: + mutex_unlock(&c->bucket_lock); + + trace_bcache_btree_node_alloc_fail(c); + return b; +} + +static struct btree *bch_btree_node_alloc(struct cache_set *c, + struct btree_op *op, int level, + struct btree *parent) +{ + return __bch_btree_node_alloc(c, op, level, op != NULL, parent); +} + +static struct btree *btree_node_alloc_replacement(struct btree *b, + struct btree_op *op) +{ + struct btree *n = bch_btree_node_alloc(b->c, op, b->level, b->parent); + + if (!IS_ERR(n)) { + mutex_lock(&n->write_lock); + bch_btree_sort_into(&b->keys, &n->keys, &b->c->sort); + bkey_copy_key(&n->key, &b->key); + mutex_unlock(&n->write_lock); + } + + return n; +} + +static void make_btree_freeing_key(struct btree *b, struct bkey *k) +{ + unsigned int i; + + mutex_lock(&b->c->bucket_lock); + + atomic_inc(&b->c->prio_blocked); + + bkey_copy(k, &b->key); + bkey_copy_key(k, &ZERO_KEY); + + for (i = 0; i < KEY_PTRS(k); i++) + SET_PTR_GEN(k, i, + bch_inc_gen(PTR_CACHE(b->c, &b->key, i), + PTR_BUCKET(b->c, &b->key, i))); + + mutex_unlock(&b->c->bucket_lock); +} + +static int btree_check_reserve(struct btree *b, struct btree_op *op) +{ + struct cache_set *c = b->c; + struct cache *ca = c->cache; + unsigned int reserve = (c->root->level - b->level) * 2 + 1; + + mutex_lock(&c->bucket_lock); + + if (fifo_used(&ca->free[RESERVE_BTREE]) < reserve) { + if (op) + prepare_to_wait(&c->btree_cache_wait, &op->wait, + TASK_UNINTERRUPTIBLE); + mutex_unlock(&c->bucket_lock); + return -EINTR; + } + + mutex_unlock(&c->bucket_lock); + + return mca_cannibalize_lock(b->c, op); +} + +/* Garbage collection */ + +static uint8_t __bch_btree_mark_key(struct cache_set *c, int level, + struct bkey *k) +{ + uint8_t stale = 0; + unsigned int i; + struct bucket *g; + + /* + * ptr_invalid() can't return true for the keys that mark btree nodes as + * freed, but since ptr_bad() returns true we'll never actually use them + * for anything and thus we don't want mark their pointers here + */ + if (!bkey_cmp(k, &ZERO_KEY)) + return stale; + + for (i = 0; i < KEY_PTRS(k); i++) { + if (!ptr_available(c, k, i)) + continue; + + g = PTR_BUCKET(c, k, i); + + if (gen_after(g->last_gc, PTR_GEN(k, i))) + g->last_gc = PTR_GEN(k, i); + + if (ptr_stale(c, k, i)) { + stale = max(stale, ptr_stale(c, k, i)); + continue; + } + + cache_bug_on(GC_MARK(g) && + (GC_MARK(g) == GC_MARK_METADATA) != (level != 0), + c, "inconsistent ptrs: mark = %llu, level = %i", + GC_MARK(g), level); + + if (level) + SET_GC_MARK(g, GC_MARK_METADATA); + else if (KEY_DIRTY(k)) + SET_GC_MARK(g, GC_MARK_DIRTY); + else if (!GC_MARK(g)) + SET_GC_MARK(g, GC_MARK_RECLAIMABLE); + + /* guard against overflow */ + SET_GC_SECTORS_USED(g, min_t(unsigned int, + GC_SECTORS_USED(g) + KEY_SIZE(k), + MAX_GC_SECTORS_USED)); + + BUG_ON(!GC_SECTORS_USED(g)); + } + + return stale; +} + +#define btree_mark_key(b, k) __bch_btree_mark_key(b->c, b->level, k) + +void bch_initial_mark_key(struct cache_set *c, int level, struct bkey *k) +{ + unsigned int i; + + for (i = 0; i < KEY_PTRS(k); i++) + if (ptr_available(c, k, i) && + !ptr_stale(c, k, i)) { + struct bucket *b = PTR_BUCKET(c, k, i); + + b->gen = PTR_GEN(k, i); + + if (level && bkey_cmp(k, &ZERO_KEY)) + b->prio = BTREE_PRIO; + else if (!level && b->prio == BTREE_PRIO) + b->prio = INITIAL_PRIO; + } + + __bch_btree_mark_key(c, level, k); +} + +void bch_update_bucket_in_use(struct cache_set *c, struct gc_stat *stats) +{ + stats->in_use = (c->nbuckets - c->avail_nbuckets) * 100 / c->nbuckets; +} + +static bool btree_gc_mark_node(struct btree *b, struct gc_stat *gc) +{ + uint8_t stale = 0; + unsigned int keys = 0, good_keys = 0; + struct bkey *k; + struct btree_iter iter; + struct bset_tree *t; + + gc->nodes++; + + for_each_key_filter(&b->keys, k, &iter, bch_ptr_invalid) { + stale = max(stale, btree_mark_key(b, k)); + keys++; + + if (bch_ptr_bad(&b->keys, k)) + continue; + + gc->key_bytes += bkey_u64s(k); + gc->nkeys++; + good_keys++; + + gc->data += KEY_SIZE(k); + } + + for (t = b->keys.set; t <= &b->keys.set[b->keys.nsets]; t++) + btree_bug_on(t->size && + bset_written(&b->keys, t) && + bkey_cmp(&b->key, &t->end) < 0, + b, "found short btree key in gc"); + + if (b->c->gc_always_rewrite) + return true; + + if (stale > 10) + return true; + + if ((keys - good_keys) * 2 > keys) + return true; + + return false; +} + +#define GC_MERGE_NODES 4U + +struct gc_merge_info { + struct btree *b; + unsigned int keys; +}; + +static int bch_btree_insert_node(struct btree *b, struct btree_op *op, + struct keylist *insert_keys, + atomic_t *journal_ref, + struct bkey *replace_key); + +static int btree_gc_coalesce(struct btree *b, struct btree_op *op, + struct gc_stat *gc, struct gc_merge_info *r) +{ + unsigned int i, nodes = 0, keys = 0, blocks; + struct btree *new_nodes[GC_MERGE_NODES]; + struct keylist keylist; + struct closure cl; + struct bkey *k; + + bch_keylist_init(&keylist); + + if (btree_check_reserve(b, NULL)) + return 0; + + memset(new_nodes, 0, sizeof(new_nodes)); + closure_init_stack(&cl); + + while (nodes < GC_MERGE_NODES && !IS_ERR_OR_NULL(r[nodes].b)) + keys += r[nodes++].keys; + + blocks = btree_default_blocks(b->c) * 2 / 3; + + if (nodes < 2 || + __set_blocks(b->keys.set[0].data, keys, + block_bytes(b->c->cache)) > blocks * (nodes - 1)) + return 0; + + for (i = 0; i < nodes; i++) { + new_nodes[i] = btree_node_alloc_replacement(r[i].b, NULL); + if (IS_ERR(new_nodes[i])) + goto out_nocoalesce; + } + + /* + * We have to check the reserve here, after we've allocated our new + * nodes, to make sure the insert below will succeed - we also check + * before as an optimization to potentially avoid a bunch of expensive + * allocs/sorts + */ + if (btree_check_reserve(b, NULL)) + goto out_nocoalesce; + + for (i = 0; i < nodes; i++) + mutex_lock(&new_nodes[i]->write_lock); + + for (i = nodes - 1; i > 0; --i) { + struct bset *n1 = btree_bset_first(new_nodes[i]); + struct bset *n2 = btree_bset_first(new_nodes[i - 1]); + struct bkey *k, *last = NULL; + + keys = 0; + + if (i > 1) { + for (k = n2->start; + k < bset_bkey_last(n2); + k = bkey_next(k)) { + if (__set_blocks(n1, n1->keys + keys + + bkey_u64s(k), + block_bytes(b->c->cache)) > blocks) + break; + + last = k; + keys += bkey_u64s(k); + } + } else { + /* + * Last node we're not getting rid of - we're getting + * rid of the node at r[0]. Have to try and fit all of + * the remaining keys into this node; we can't ensure + * they will always fit due to rounding and variable + * length keys (shouldn't be possible in practice, + * though) + */ + if (__set_blocks(n1, n1->keys + n2->keys, + block_bytes(b->c->cache)) > + btree_blocks(new_nodes[i])) + goto out_unlock_nocoalesce; + + keys = n2->keys; + /* Take the key of the node we're getting rid of */ + last = &r->b->key; + } + + BUG_ON(__set_blocks(n1, n1->keys + keys, block_bytes(b->c->cache)) > + btree_blocks(new_nodes[i])); + + if (last) + bkey_copy_key(&new_nodes[i]->key, last); + + memcpy(bset_bkey_last(n1), + n2->start, + (void *) bset_bkey_idx(n2, keys) - (void *) n2->start); + + n1->keys += keys; + r[i].keys = n1->keys; + + memmove(n2->start, + bset_bkey_idx(n2, keys), + (void *) bset_bkey_last(n2) - + (void *) bset_bkey_idx(n2, keys)); + + n2->keys -= keys; + + if (__bch_keylist_realloc(&keylist, + bkey_u64s(&new_nodes[i]->key))) + goto out_unlock_nocoalesce; + + bch_btree_node_write(new_nodes[i], &cl); + bch_keylist_add(&keylist, &new_nodes[i]->key); + } + + for (i = 0; i < nodes; i++) + mutex_unlock(&new_nodes[i]->write_lock); + + closure_sync(&cl); + + /* We emptied out this node */ + BUG_ON(btree_bset_first(new_nodes[0])->keys); + btree_node_free(new_nodes[0]); + rw_unlock(true, new_nodes[0]); + new_nodes[0] = NULL; + + for (i = 0; i < nodes; i++) { + if (__bch_keylist_realloc(&keylist, bkey_u64s(&r[i].b->key))) + goto out_nocoalesce; + + make_btree_freeing_key(r[i].b, keylist.top); + bch_keylist_push(&keylist); + } + + bch_btree_insert_node(b, op, &keylist, NULL, NULL); + BUG_ON(!bch_keylist_empty(&keylist)); + + for (i = 0; i < nodes; i++) { + btree_node_free(r[i].b); + rw_unlock(true, r[i].b); + + r[i].b = new_nodes[i]; + } + + memmove(r, r + 1, sizeof(r[0]) * (nodes - 1)); + r[nodes - 1].b = ERR_PTR(-EINTR); + + trace_bcache_btree_gc_coalesce(nodes); + gc->nodes--; + + bch_keylist_free(&keylist); + + /* Invalidated our iterator */ + return -EINTR; + +out_unlock_nocoalesce: + for (i = 0; i < nodes; i++) + mutex_unlock(&new_nodes[i]->write_lock); + +out_nocoalesce: + closure_sync(&cl); + + while ((k = bch_keylist_pop(&keylist))) + if (!bkey_cmp(k, &ZERO_KEY)) + atomic_dec(&b->c->prio_blocked); + bch_keylist_free(&keylist); + + for (i = 0; i < nodes; i++) + if (!IS_ERR_OR_NULL(new_nodes[i])) { + btree_node_free(new_nodes[i]); + rw_unlock(true, new_nodes[i]); + } + return 0; +} + +static int btree_gc_rewrite_node(struct btree *b, struct btree_op *op, + struct btree *replace) +{ + struct keylist keys; + struct btree *n; + + if (btree_check_reserve(b, NULL)) + return 0; + + n = btree_node_alloc_replacement(replace, NULL); + if (IS_ERR(n)) + return 0; + + /* recheck reserve after allocating replacement node */ + if (btree_check_reserve(b, NULL)) { + btree_node_free(n); + rw_unlock(true, n); + return 0; + } + + bch_btree_node_write_sync(n); + + bch_keylist_init(&keys); + bch_keylist_add(&keys, &n->key); + + make_btree_freeing_key(replace, keys.top); + bch_keylist_push(&keys); + + bch_btree_insert_node(b, op, &keys, NULL, NULL); + BUG_ON(!bch_keylist_empty(&keys)); + + btree_node_free(replace); + rw_unlock(true, n); + + /* Invalidated our iterator */ + return -EINTR; +} + +static unsigned int btree_gc_count_keys(struct btree *b) +{ + struct bkey *k; + struct btree_iter iter; + unsigned int ret = 0; + + for_each_key_filter(&b->keys, k, &iter, bch_ptr_bad) + ret += bkey_u64s(k); + + return ret; +} + +static size_t btree_gc_min_nodes(struct cache_set *c) +{ + size_t min_nodes; + + /* + * Since incremental GC would stop 100ms when front + * side I/O comes, so when there are many btree nodes, + * if GC only processes constant (100) nodes each time, + * GC would last a long time, and the front side I/Os + * would run out of the buckets (since no new bucket + * can be allocated during GC), and be blocked again. + * So GC should not process constant nodes, but varied + * nodes according to the number of btree nodes, which + * realized by dividing GC into constant(100) times, + * so when there are many btree nodes, GC can process + * more nodes each time, otherwise, GC will process less + * nodes each time (but no less than MIN_GC_NODES) + */ + min_nodes = c->gc_stats.nodes / MAX_GC_TIMES; + if (min_nodes < MIN_GC_NODES) + min_nodes = MIN_GC_NODES; + + return min_nodes; +} + + +static int btree_gc_recurse(struct btree *b, struct btree_op *op, + struct closure *writes, struct gc_stat *gc) +{ + int ret = 0; + bool should_rewrite; + struct bkey *k; + struct btree_iter iter; + struct gc_merge_info r[GC_MERGE_NODES]; + struct gc_merge_info *i, *last = r + ARRAY_SIZE(r) - 1; + + bch_btree_iter_init(&b->keys, &iter, &b->c->gc_done); + + for (i = r; i < r + ARRAY_SIZE(r); i++) + i->b = ERR_PTR(-EINTR); + + while (1) { + k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad); + if (k) { + r->b = bch_btree_node_get(b->c, op, k, b->level - 1, + true, b); + if (IS_ERR(r->b)) { + ret = PTR_ERR(r->b); + break; + } + + r->keys = btree_gc_count_keys(r->b); + + ret = btree_gc_coalesce(b, op, gc, r); + if (ret) + break; + } + + if (!last->b) + break; + + if (!IS_ERR(last->b)) { + should_rewrite = btree_gc_mark_node(last->b, gc); + if (should_rewrite) { + ret = btree_gc_rewrite_node(b, op, last->b); + if (ret) + break; + } + + if (last->b->level) { + ret = btree_gc_recurse(last->b, op, writes, gc); + if (ret) + break; + } + + bkey_copy_key(&b->c->gc_done, &last->b->key); + + /* + * Must flush leaf nodes before gc ends, since replace + * operations aren't journalled + */ + mutex_lock(&last->b->write_lock); + if (btree_node_dirty(last->b)) + bch_btree_node_write(last->b, writes); + mutex_unlock(&last->b->write_lock); + rw_unlock(true, last->b); + } + + memmove(r + 1, r, sizeof(r[0]) * (GC_MERGE_NODES - 1)); + r->b = NULL; + + if (atomic_read(&b->c->search_inflight) && + gc->nodes >= gc->nodes_pre + btree_gc_min_nodes(b->c)) { + gc->nodes_pre = gc->nodes; + ret = -EAGAIN; + break; + } + + if (need_resched()) { + ret = -EAGAIN; + break; + } + } + + for (i = r; i < r + ARRAY_SIZE(r); i++) + if (!IS_ERR_OR_NULL(i->b)) { + mutex_lock(&i->b->write_lock); + if (btree_node_dirty(i->b)) + bch_btree_node_write(i->b, writes); + mutex_unlock(&i->b->write_lock); + rw_unlock(true, i->b); + } + + return ret; +} + +static int bch_btree_gc_root(struct btree *b, struct btree_op *op, + struct closure *writes, struct gc_stat *gc) +{ + struct btree *n = NULL; + int ret = 0; + bool should_rewrite; + + should_rewrite = btree_gc_mark_node(b, gc); + if (should_rewrite) { + n = btree_node_alloc_replacement(b, NULL); + + if (!IS_ERR(n)) { + bch_btree_node_write_sync(n); + + bch_btree_set_root(n); + btree_node_free(b); + rw_unlock(true, n); + + return -EINTR; + } + } + + __bch_btree_mark_key(b->c, b->level + 1, &b->key); + + if (b->level) { + ret = btree_gc_recurse(b, op, writes, gc); + if (ret) + return ret; + } + + bkey_copy_key(&b->c->gc_done, &b->key); + + return ret; +} + +static void btree_gc_start(struct cache_set *c) +{ + struct cache *ca; + struct bucket *b; + + if (!c->gc_mark_valid) + return; + + mutex_lock(&c->bucket_lock); + + c->gc_mark_valid = 0; + c->gc_done = ZERO_KEY; + + ca = c->cache; + for_each_bucket(b, ca) { + b->last_gc = b->gen; + if (!atomic_read(&b->pin)) { + SET_GC_MARK(b, 0); + SET_GC_SECTORS_USED(b, 0); + } + } + + mutex_unlock(&c->bucket_lock); +} + +static void bch_btree_gc_finish(struct cache_set *c) +{ + struct bucket *b; + struct cache *ca; + unsigned int i, j; + uint64_t *k; + + mutex_lock(&c->bucket_lock); + + set_gc_sectors(c); + c->gc_mark_valid = 1; + c->need_gc = 0; + + for (i = 0; i < KEY_PTRS(&c->uuid_bucket); i++) + SET_GC_MARK(PTR_BUCKET(c, &c->uuid_bucket, i), + GC_MARK_METADATA); + + /* don't reclaim buckets to which writeback keys point */ + rcu_read_lock(); + for (i = 0; i < c->devices_max_used; i++) { + struct bcache_device *d = c->devices[i]; + struct cached_dev *dc; + struct keybuf_key *w, *n; + + if (!d || UUID_FLASH_ONLY(&c->uuids[i])) + continue; + dc = container_of(d, struct cached_dev, disk); + + spin_lock(&dc->writeback_keys.lock); + rbtree_postorder_for_each_entry_safe(w, n, + &dc->writeback_keys.keys, node) + for (j = 0; j < KEY_PTRS(&w->key); j++) + SET_GC_MARK(PTR_BUCKET(c, &w->key, j), + GC_MARK_DIRTY); + spin_unlock(&dc->writeback_keys.lock); + } + rcu_read_unlock(); + + c->avail_nbuckets = 0; + + ca = c->cache; + ca->invalidate_needs_gc = 0; + + for (k = ca->sb.d; k < ca->sb.d + ca->sb.keys; k++) + SET_GC_MARK(ca->buckets + *k, GC_MARK_METADATA); + + for (k = ca->prio_buckets; + k < ca->prio_buckets + prio_buckets(ca) * 2; k++) + SET_GC_MARK(ca->buckets + *k, GC_MARK_METADATA); + + for_each_bucket(b, ca) { + c->need_gc = max(c->need_gc, bucket_gc_gen(b)); + + if (atomic_read(&b->pin)) + continue; + + BUG_ON(!GC_MARK(b) && GC_SECTORS_USED(b)); + + if (!GC_MARK(b) || GC_MARK(b) == GC_MARK_RECLAIMABLE) + c->avail_nbuckets++; + } + + mutex_unlock(&c->bucket_lock); +} + +static void bch_btree_gc(struct cache_set *c) +{ + int ret; + struct gc_stat stats; + struct closure writes; + struct btree_op op; + uint64_t start_time = local_clock(); + + trace_bcache_gc_start(c); + + memset(&stats, 0, sizeof(struct gc_stat)); + closure_init_stack(&writes); + bch_btree_op_init(&op, SHRT_MAX); + + btree_gc_start(c); + + /* if CACHE_SET_IO_DISABLE set, gc thread should stop too */ + do { + ret = bcache_btree_root(gc_root, c, &op, &writes, &stats); + closure_sync(&writes); + cond_resched(); + + if (ret == -EAGAIN) + schedule_timeout_interruptible(msecs_to_jiffies + (GC_SLEEP_MS)); + else if (ret) + pr_warn("gc failed!\n"); + } while (ret && !test_bit(CACHE_SET_IO_DISABLE, &c->flags)); + + bch_btree_gc_finish(c); + wake_up_allocators(c); + + bch_time_stats_update(&c->btree_gc_time, start_time); + + stats.key_bytes *= sizeof(uint64_t); + stats.data <<= 9; + bch_update_bucket_in_use(c, &stats); + memcpy(&c->gc_stats, &stats, sizeof(struct gc_stat)); + + trace_bcache_gc_end(c); + + bch_moving_gc(c); +} + +static bool gc_should_run(struct cache_set *c) +{ + struct cache *ca = c->cache; + + if (ca->invalidate_needs_gc) + return true; + + if (atomic_read(&c->sectors_to_gc) < 0) + return true; + + return false; +} + +static int bch_gc_thread(void *arg) +{ + struct cache_set *c = arg; + + while (1) { + wait_event_interruptible(c->gc_wait, + kthread_should_stop() || + test_bit(CACHE_SET_IO_DISABLE, &c->flags) || + gc_should_run(c)); + + if (kthread_should_stop() || + test_bit(CACHE_SET_IO_DISABLE, &c->flags)) + break; + + set_gc_sectors(c); + bch_btree_gc(c); + } + + wait_for_kthread_stop(); + return 0; +} + +int bch_gc_thread_start(struct cache_set *c) +{ + c->gc_thread = kthread_run(bch_gc_thread, c, "bcache_gc"); + return PTR_ERR_OR_ZERO(c->gc_thread); +} + +/* Initial partial gc */ + +static int bch_btree_check_recurse(struct btree *b, struct btree_op *op) +{ + int ret = 0; + struct bkey *k, *p = NULL; + struct btree_iter iter; + + for_each_key_filter(&b->keys, k, &iter, bch_ptr_invalid) + bch_initial_mark_key(b->c, b->level, k); + + bch_initial_mark_key(b->c, b->level + 1, &b->key); + + if (b->level) { + bch_btree_iter_init(&b->keys, &iter, NULL); + + do { + k = bch_btree_iter_next_filter(&iter, &b->keys, + bch_ptr_bad); + if (k) { + btree_node_prefetch(b, k); + /* + * initiallize c->gc_stats.nodes + * for incremental GC + */ + b->c->gc_stats.nodes++; + } + + if (p) + ret = bcache_btree(check_recurse, p, b, op); + + p = k; + } while (p && !ret); + } + + return ret; +} + + +static int bch_btree_check_thread(void *arg) +{ + int ret; + struct btree_check_info *info = arg; + struct btree_check_state *check_state = info->state; + struct cache_set *c = check_state->c; + struct btree_iter iter; + struct bkey *k, *p; + int cur_idx, prev_idx, skip_nr; + + k = p = NULL; + cur_idx = prev_idx = 0; + ret = 0; + + /* root node keys are checked before thread created */ + bch_btree_iter_init(&c->root->keys, &iter, NULL); + k = bch_btree_iter_next_filter(&iter, &c->root->keys, bch_ptr_bad); + BUG_ON(!k); + + p = k; + while (k) { + /* + * Fetch a root node key index, skip the keys which + * should be fetched by other threads, then check the + * sub-tree indexed by the fetched key. + */ + spin_lock(&check_state->idx_lock); + cur_idx = check_state->key_idx; + check_state->key_idx++; + spin_unlock(&check_state->idx_lock); + + skip_nr = cur_idx - prev_idx; + + while (skip_nr) { + k = bch_btree_iter_next_filter(&iter, + &c->root->keys, + bch_ptr_bad); + if (k) + p = k; + else { + /* + * No more keys to check in root node, + * current checking threads are enough, + * stop creating more. + */ + atomic_set(&check_state->enough, 1); + /* Update check_state->enough earlier */ + smp_mb__after_atomic(); + goto out; + } + skip_nr--; + cond_resched(); + } + + if (p) { + struct btree_op op; + + btree_node_prefetch(c->root, p); + c->gc_stats.nodes++; + bch_btree_op_init(&op, 0); + ret = bcache_btree(check_recurse, p, c->root, &op); + /* + * The op may be added to cache_set's btree_cache_wait + * in mca_cannibalize(), must ensure it is removed from + * the list and release btree_cache_alloc_lock before + * free op memory. + * Otherwise, the btree_cache_wait will be damaged. + */ + bch_cannibalize_unlock(c); + finish_wait(&c->btree_cache_wait, &(&op)->wait); + if (ret) + goto out; + } + p = NULL; + prev_idx = cur_idx; + cond_resched(); + } + +out: + info->result = ret; + /* update check_state->started among all CPUs */ + smp_mb__before_atomic(); + if (atomic_dec_and_test(&check_state->started)) + wake_up(&check_state->wait); + + return ret; +} + + + +static int bch_btree_chkthread_nr(void) +{ + int n = num_online_cpus()/2; + + if (n == 0) + n = 1; + else if (n > BCH_BTR_CHKTHREAD_MAX) + n = BCH_BTR_CHKTHREAD_MAX; + + return n; +} + +int bch_btree_check(struct cache_set *c) +{ + int ret = 0; + int i; + struct bkey *k = NULL; + struct btree_iter iter; + struct btree_check_state check_state; + + /* check and mark root node keys */ + for_each_key_filter(&c->root->keys, k, &iter, bch_ptr_invalid) + bch_initial_mark_key(c, c->root->level, k); + + bch_initial_mark_key(c, c->root->level + 1, &c->root->key); + + if (c->root->level == 0) + return 0; + + memset(&check_state, 0, sizeof(struct btree_check_state)); + check_state.c = c; + check_state.total_threads = bch_btree_chkthread_nr(); + check_state.key_idx = 0; + spin_lock_init(&check_state.idx_lock); + atomic_set(&check_state.started, 0); + atomic_set(&check_state.enough, 0); + init_waitqueue_head(&check_state.wait); + + rw_lock(0, c->root, c->root->level); + /* + * Run multiple threads to check btree nodes in parallel, + * if check_state.enough is non-zero, it means current + * running check threads are enough, unncessary to create + * more. + */ + for (i = 0; i < check_state.total_threads; i++) { + /* fetch latest check_state.enough earlier */ + smp_mb__before_atomic(); + if (atomic_read(&check_state.enough)) + break; + + check_state.infos[i].result = 0; + check_state.infos[i].state = &check_state; + + check_state.infos[i].thread = + kthread_run(bch_btree_check_thread, + &check_state.infos[i], + "bch_btrchk[%d]", i); + if (IS_ERR(check_state.infos[i].thread)) { + pr_err("fails to run thread bch_btrchk[%d]\n", i); + for (--i; i >= 0; i--) + kthread_stop(check_state.infos[i].thread); + ret = -ENOMEM; + goto out; + } + atomic_inc(&check_state.started); + } + + /* + * Must wait for all threads to stop. + */ + wait_event(check_state.wait, atomic_read(&check_state.started) == 0); + + for (i = 0; i < check_state.total_threads; i++) { + if (check_state.infos[i].result) { + ret = check_state.infos[i].result; + goto out; + } + } + +out: + rw_unlock(0, c->root); + return ret; +} + +void bch_initial_gc_finish(struct cache_set *c) +{ + struct cache *ca = c->cache; + struct bucket *b; + + bch_btree_gc_finish(c); + + mutex_lock(&c->bucket_lock); + + /* + * We need to put some unused buckets directly on the prio freelist in + * order to get the allocator thread started - it needs freed buckets in + * order to rewrite the prios and gens, and it needs to rewrite prios + * and gens in order to free buckets. + * + * This is only safe for buckets that have no live data in them, which + * there should always be some of. + */ + for_each_bucket(b, ca) { + if (fifo_full(&ca->free[RESERVE_PRIO]) && + fifo_full(&ca->free[RESERVE_BTREE])) + break; + + if (bch_can_invalidate_bucket(ca, b) && + !GC_MARK(b)) { + __bch_invalidate_one_bucket(ca, b); + if (!fifo_push(&ca->free[RESERVE_PRIO], + b - ca->buckets)) + fifo_push(&ca->free[RESERVE_BTREE], + b - ca->buckets); + } + } + + mutex_unlock(&c->bucket_lock); +} + +/* Btree insertion */ + +static bool btree_insert_key(struct btree *b, struct bkey *k, + struct bkey *replace_key) +{ + unsigned int status; + + BUG_ON(bkey_cmp(k, &b->key) > 0); + + status = bch_btree_insert_key(&b->keys, k, replace_key); + if (status != BTREE_INSERT_STATUS_NO_INSERT) { + bch_check_keys(&b->keys, "%u for %s", status, + replace_key ? "replace" : "insert"); + + trace_bcache_btree_insert_key(b, k, replace_key != NULL, + status); + return true; + } else + return false; +} + +static size_t insert_u64s_remaining(struct btree *b) +{ + long ret = bch_btree_keys_u64s_remaining(&b->keys); + + /* + * Might land in the middle of an existing extent and have to split it + */ + if (b->keys.ops->is_extents) + ret -= KEY_MAX_U64S; + + return max(ret, 0L); +} + +static bool bch_btree_insert_keys(struct btree *b, struct btree_op *op, + struct keylist *insert_keys, + struct bkey *replace_key) +{ + bool ret = false; + int oldsize = bch_count_data(&b->keys); + + while (!bch_keylist_empty(insert_keys)) { + struct bkey *k = insert_keys->keys; + + if (bkey_u64s(k) > insert_u64s_remaining(b)) + break; + + if (bkey_cmp(k, &b->key) <= 0) { + if (!b->level) + bkey_put(b->c, k); + + ret |= btree_insert_key(b, k, replace_key); + bch_keylist_pop_front(insert_keys); + } else if (bkey_cmp(&START_KEY(k), &b->key) < 0) { + BKEY_PADDED(key) temp; + bkey_copy(&temp.key, insert_keys->keys); + + bch_cut_back(&b->key, &temp.key); + bch_cut_front(&b->key, insert_keys->keys); + + ret |= btree_insert_key(b, &temp.key, replace_key); + break; + } else { + break; + } + } + + if (!ret) + op->insert_collision = true; + + BUG_ON(!bch_keylist_empty(insert_keys) && b->level); + + BUG_ON(bch_count_data(&b->keys) < oldsize); + return ret; +} + +static int btree_split(struct btree *b, struct btree_op *op, + struct keylist *insert_keys, + struct bkey *replace_key) +{ + bool split; + struct btree *n1, *n2 = NULL, *n3 = NULL; + uint64_t start_time = local_clock(); + struct closure cl; + struct keylist parent_keys; + + closure_init_stack(&cl); + bch_keylist_init(&parent_keys); + + if (btree_check_reserve(b, op)) { + if (!b->level) + return -EINTR; + else + WARN(1, "insufficient reserve for split\n"); + } + + n1 = btree_node_alloc_replacement(b, op); + if (IS_ERR(n1)) + goto err; + + split = set_blocks(btree_bset_first(n1), + block_bytes(n1->c->cache)) > (btree_blocks(b) * 4) / 5; + + if (split) { + unsigned int keys = 0; + + trace_bcache_btree_node_split(b, btree_bset_first(n1)->keys); + + n2 = bch_btree_node_alloc(b->c, op, b->level, b->parent); + if (IS_ERR(n2)) + goto err_free1; + + if (!b->parent) { + n3 = bch_btree_node_alloc(b->c, op, b->level + 1, NULL); + if (IS_ERR(n3)) + goto err_free2; + } + + mutex_lock(&n1->write_lock); + mutex_lock(&n2->write_lock); + + bch_btree_insert_keys(n1, op, insert_keys, replace_key); + + /* + * Has to be a linear search because we don't have an auxiliary + * search tree yet + */ + + while (keys < (btree_bset_first(n1)->keys * 3) / 5) + keys += bkey_u64s(bset_bkey_idx(btree_bset_first(n1), + keys)); + + bkey_copy_key(&n1->key, + bset_bkey_idx(btree_bset_first(n1), keys)); + keys += bkey_u64s(bset_bkey_idx(btree_bset_first(n1), keys)); + + btree_bset_first(n2)->keys = btree_bset_first(n1)->keys - keys; + btree_bset_first(n1)->keys = keys; + + memcpy(btree_bset_first(n2)->start, + bset_bkey_last(btree_bset_first(n1)), + btree_bset_first(n2)->keys * sizeof(uint64_t)); + + bkey_copy_key(&n2->key, &b->key); + + bch_keylist_add(&parent_keys, &n2->key); + bch_btree_node_write(n2, &cl); + mutex_unlock(&n2->write_lock); + rw_unlock(true, n2); + } else { + trace_bcache_btree_node_compact(b, btree_bset_first(n1)->keys); + + mutex_lock(&n1->write_lock); + bch_btree_insert_keys(n1, op, insert_keys, replace_key); + } + + bch_keylist_add(&parent_keys, &n1->key); + bch_btree_node_write(n1, &cl); + mutex_unlock(&n1->write_lock); + + if (n3) { + /* Depth increases, make a new root */ + mutex_lock(&n3->write_lock); + bkey_copy_key(&n3->key, &MAX_KEY); + bch_btree_insert_keys(n3, op, &parent_keys, NULL); + bch_btree_node_write(n3, &cl); + mutex_unlock(&n3->write_lock); + + closure_sync(&cl); + bch_btree_set_root(n3); + rw_unlock(true, n3); + } else if (!b->parent) { + /* Root filled up but didn't need to be split */ + closure_sync(&cl); + bch_btree_set_root(n1); + } else { + /* Split a non root node */ + closure_sync(&cl); + make_btree_freeing_key(b, parent_keys.top); + bch_keylist_push(&parent_keys); + + bch_btree_insert_node(b->parent, op, &parent_keys, NULL, NULL); + BUG_ON(!bch_keylist_empty(&parent_keys)); + } + + btree_node_free(b); + rw_unlock(true, n1); + + bch_time_stats_update(&b->c->btree_split_time, start_time); + + return 0; +err_free2: + bkey_put(b->c, &n2->key); + btree_node_free(n2); + rw_unlock(true, n2); +err_free1: + bkey_put(b->c, &n1->key); + btree_node_free(n1); + rw_unlock(true, n1); +err: + WARN(1, "bcache: btree split failed (level %u)", b->level); + + if (n3 == ERR_PTR(-EAGAIN) || + n2 == ERR_PTR(-EAGAIN) || + n1 == ERR_PTR(-EAGAIN)) + return -EAGAIN; + + return -ENOMEM; +} + +static int bch_btree_insert_node(struct btree *b, struct btree_op *op, + struct keylist *insert_keys, + atomic_t *journal_ref, + struct bkey *replace_key) +{ + struct closure cl; + + BUG_ON(b->level && replace_key); + + closure_init_stack(&cl); + + mutex_lock(&b->write_lock); + + if (write_block(b) != btree_bset_last(b) && + b->keys.last_set_unwritten) + bch_btree_init_next(b); /* just wrote a set */ + + if (bch_keylist_nkeys(insert_keys) > insert_u64s_remaining(b)) { + mutex_unlock(&b->write_lock); + goto split; + } + + BUG_ON(write_block(b) != btree_bset_last(b)); + + if (bch_btree_insert_keys(b, op, insert_keys, replace_key)) { + if (!b->level) + bch_btree_leaf_dirty(b, journal_ref); + else + bch_btree_node_write(b, &cl); + } + + mutex_unlock(&b->write_lock); + + /* wait for btree node write if necessary, after unlock */ + closure_sync(&cl); + + return 0; +split: + if (current->bio_list) { + op->lock = b->c->root->level + 1; + return -EAGAIN; + } else if (op->lock <= b->c->root->level) { + op->lock = b->c->root->level + 1; + return -EINTR; + } else { + /* Invalidated all iterators */ + int ret = btree_split(b, op, insert_keys, replace_key); + + if (bch_keylist_empty(insert_keys)) + return 0; + else if (!ret) + return -EINTR; + return ret; + } +} + +int bch_btree_insert_check_key(struct btree *b, struct btree_op *op, + struct bkey *check_key) +{ + int ret = -EINTR; + uint64_t btree_ptr = b->key.ptr[0]; + unsigned long seq = b->seq; + struct keylist insert; + bool upgrade = op->lock == -1; + + bch_keylist_init(&insert); + + if (upgrade) { + rw_unlock(false, b); + rw_lock(true, b, b->level); + + if (b->key.ptr[0] != btree_ptr || + b->seq != seq + 1) { + op->lock = b->level; + goto out; + } + } + + SET_KEY_PTRS(check_key, 1); + get_random_bytes(&check_key->ptr[0], sizeof(uint64_t)); + + SET_PTR_DEV(check_key, 0, PTR_CHECK_DEV); + + bch_keylist_add(&insert, check_key); + + ret = bch_btree_insert_node(b, op, &insert, NULL, NULL); + + BUG_ON(!ret && !bch_keylist_empty(&insert)); +out: + if (upgrade) + downgrade_write(&b->lock); + return ret; +} + +struct btree_insert_op { + struct btree_op op; + struct keylist *keys; + atomic_t *journal_ref; + struct bkey *replace_key; +}; + +static int btree_insert_fn(struct btree_op *b_op, struct btree *b) +{ + struct btree_insert_op *op = container_of(b_op, + struct btree_insert_op, op); + + int ret = bch_btree_insert_node(b, &op->op, op->keys, + op->journal_ref, op->replace_key); + if (ret && !bch_keylist_empty(op->keys)) + return ret; + else + return MAP_DONE; +} + +int bch_btree_insert(struct cache_set *c, struct keylist *keys, + atomic_t *journal_ref, struct bkey *replace_key) +{ + struct btree_insert_op op; + int ret = 0; + + BUG_ON(current->bio_list); + BUG_ON(bch_keylist_empty(keys)); + + bch_btree_op_init(&op.op, 0); + op.keys = keys; + op.journal_ref = journal_ref; + op.replace_key = replace_key; + + while (!ret && !bch_keylist_empty(keys)) { + op.op.lock = 0; + ret = bch_btree_map_leaf_nodes(&op.op, c, + &START_KEY(keys->keys), + btree_insert_fn); + } + + if (ret) { + struct bkey *k; + + pr_err("error %i\n", ret); + + while ((k = bch_keylist_pop(keys))) + bkey_put(c, k); + } else if (op.op.insert_collision) + ret = -ESRCH; + + return ret; +} + +void bch_btree_set_root(struct btree *b) +{ + unsigned int i; + struct closure cl; + + closure_init_stack(&cl); + + trace_bcache_btree_set_root(b); + + BUG_ON(!b->written); + + for (i = 0; i < KEY_PTRS(&b->key); i++) + BUG_ON(PTR_BUCKET(b->c, &b->key, i)->prio != BTREE_PRIO); + + mutex_lock(&b->c->bucket_lock); + list_del_init(&b->list); + mutex_unlock(&b->c->bucket_lock); + + b->c->root = b; + + bch_journal_meta(b->c, &cl); + closure_sync(&cl); +} + +/* Map across nodes or keys */ + +static int bch_btree_map_nodes_recurse(struct btree *b, struct btree_op *op, + struct bkey *from, + btree_map_nodes_fn *fn, int flags) +{ + int ret = MAP_CONTINUE; + + if (b->level) { + struct bkey *k; + struct btree_iter iter; + + bch_btree_iter_init(&b->keys, &iter, from); + + while ((k = bch_btree_iter_next_filter(&iter, &b->keys, + bch_ptr_bad))) { + ret = bcache_btree(map_nodes_recurse, k, b, + op, from, fn, flags); + from = NULL; + + if (ret != MAP_CONTINUE) + return ret; + } + } + + if (!b->level || flags == MAP_ALL_NODES) + ret = fn(op, b); + + return ret; +} + +int __bch_btree_map_nodes(struct btree_op *op, struct cache_set *c, + struct bkey *from, btree_map_nodes_fn *fn, int flags) +{ + return bcache_btree_root(map_nodes_recurse, c, op, from, fn, flags); +} + +int bch_btree_map_keys_recurse(struct btree *b, struct btree_op *op, + struct bkey *from, btree_map_keys_fn *fn, + int flags) +{ + int ret = MAP_CONTINUE; + struct bkey *k; + struct btree_iter iter; + + bch_btree_iter_init(&b->keys, &iter, from); + + while ((k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad))) { + ret = !b->level + ? fn(op, b, k) + : bcache_btree(map_keys_recurse, k, + b, op, from, fn, flags); + from = NULL; + + if (ret != MAP_CONTINUE) + return ret; + } + + if (!b->level && (flags & MAP_END_KEY)) + ret = fn(op, b, &KEY(KEY_INODE(&b->key), + KEY_OFFSET(&b->key), 0)); + + return ret; +} + +int bch_btree_map_keys(struct btree_op *op, struct cache_set *c, + struct bkey *from, btree_map_keys_fn *fn, int flags) +{ + return bcache_btree_root(map_keys_recurse, c, op, from, fn, flags); +} + +/* Keybuf code */ + +static inline int keybuf_cmp(struct keybuf_key *l, struct keybuf_key *r) +{ + /* Overlapping keys compare equal */ + if (bkey_cmp(&l->key, &START_KEY(&r->key)) <= 0) + return -1; + if (bkey_cmp(&START_KEY(&l->key), &r->key) >= 0) + return 1; + return 0; +} + +static inline int keybuf_nonoverlapping_cmp(struct keybuf_key *l, + struct keybuf_key *r) +{ + return clamp_t(int64_t, bkey_cmp(&l->key, &r->key), -1, 1); +} + +struct refill { + struct btree_op op; + unsigned int nr_found; + struct keybuf *buf; + struct bkey *end; + keybuf_pred_fn *pred; +}; + +static int refill_keybuf_fn(struct btree_op *op, struct btree *b, + struct bkey *k) +{ + struct refill *refill = container_of(op, struct refill, op); + struct keybuf *buf = refill->buf; + int ret = MAP_CONTINUE; + + if (bkey_cmp(k, refill->end) > 0) { + ret = MAP_DONE; + goto out; + } + + if (!KEY_SIZE(k)) /* end key */ + goto out; + + if (refill->pred(buf, k)) { + struct keybuf_key *w; + + spin_lock(&buf->lock); + + w = array_alloc(&buf->freelist); + if (!w) { + spin_unlock(&buf->lock); + return MAP_DONE; + } + + w->private = NULL; + bkey_copy(&w->key, k); + + if (RB_INSERT(&buf->keys, w, node, keybuf_cmp)) + array_free(&buf->freelist, w); + else + refill->nr_found++; + + if (array_freelist_empty(&buf->freelist)) + ret = MAP_DONE; + + spin_unlock(&buf->lock); + } +out: + buf->last_scanned = *k; + return ret; +} + +void bch_refill_keybuf(struct cache_set *c, struct keybuf *buf, + struct bkey *end, keybuf_pred_fn *pred) +{ + struct bkey start = buf->last_scanned; + struct refill refill; + + cond_resched(); + + bch_btree_op_init(&refill.op, -1); + refill.nr_found = 0; + refill.buf = buf; + refill.end = end; + refill.pred = pred; + + bch_btree_map_keys(&refill.op, c, &buf->last_scanned, + refill_keybuf_fn, MAP_END_KEY); + + trace_bcache_keyscan(refill.nr_found, + KEY_INODE(&start), KEY_OFFSET(&start), + KEY_INODE(&buf->last_scanned), + KEY_OFFSET(&buf->last_scanned)); + + spin_lock(&buf->lock); + + if (!RB_EMPTY_ROOT(&buf->keys)) { + struct keybuf_key *w; + + w = RB_FIRST(&buf->keys, struct keybuf_key, node); + buf->start = START_KEY(&w->key); + + w = RB_LAST(&buf->keys, struct keybuf_key, node); + buf->end = w->key; + } else { + buf->start = MAX_KEY; + buf->end = MAX_KEY; + } + + spin_unlock(&buf->lock); +} + +static void __bch_keybuf_del(struct keybuf *buf, struct keybuf_key *w) +{ + rb_erase(&w->node, &buf->keys); + array_free(&buf->freelist, w); +} + +void bch_keybuf_del(struct keybuf *buf, struct keybuf_key *w) +{ + spin_lock(&buf->lock); + __bch_keybuf_del(buf, w); + spin_unlock(&buf->lock); +} + +bool bch_keybuf_check_overlapping(struct keybuf *buf, struct bkey *start, + struct bkey *end) +{ + bool ret = false; + struct keybuf_key *p, *w, s; + + s.key = *start; + + if (bkey_cmp(end, &buf->start) <= 0 || + bkey_cmp(start, &buf->end) >= 0) + return false; + + spin_lock(&buf->lock); + w = RB_GREATER(&buf->keys, s, node, keybuf_nonoverlapping_cmp); + + while (w && bkey_cmp(&START_KEY(&w->key), end) < 0) { + p = w; + w = RB_NEXT(w, node); + + if (p->private) + ret = true; + else + __bch_keybuf_del(buf, p); + } + + spin_unlock(&buf->lock); + return ret; +} + +struct keybuf_key *bch_keybuf_next(struct keybuf *buf) +{ + struct keybuf_key *w; + + spin_lock(&buf->lock); + + w = RB_FIRST(&buf->keys, struct keybuf_key, node); + + while (w && w->private) + w = RB_NEXT(w, node); + + if (w) + w->private = ERR_PTR(-EINTR); + + spin_unlock(&buf->lock); + return w; +} + +struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *c, + struct keybuf *buf, + struct bkey *end, + keybuf_pred_fn *pred) +{ + struct keybuf_key *ret; + + while (1) { + ret = bch_keybuf_next(buf); + if (ret) + break; + + if (bkey_cmp(&buf->last_scanned, end) >= 0) { + pr_debug("scan finished\n"); + break; + } + + bch_refill_keybuf(c, buf, end, pred); + } + + return ret; +} + +void bch_keybuf_init(struct keybuf *buf) +{ + buf->last_scanned = MAX_KEY; + buf->keys = RB_ROOT; + + spin_lock_init(&buf->lock); + array_allocator_init(&buf->freelist); +} + +void bch_btree_exit(void) +{ + if (btree_io_wq) + destroy_workqueue(btree_io_wq); +} + +int __init bch_btree_init(void) +{ + btree_io_wq = alloc_workqueue("bch_btree_io", WQ_MEM_RECLAIM, 0); + if (!btree_io_wq) + return -ENOMEM; + + return 0; +} diff --git a/drivers/md/bcache/btree.h b/drivers/md/bcache/btree.h new file mode 100644 index 000000000..a2920bbfc --- /dev/null +++ b/drivers/md/bcache/btree.h @@ -0,0 +1,417 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _BCACHE_BTREE_H +#define _BCACHE_BTREE_H + +/* + * THE BTREE: + * + * At a high level, bcache's btree is relatively standard b+ tree. All keys and + * pointers are in the leaves; interior nodes only have pointers to the child + * nodes. + * + * In the interior nodes, a struct bkey always points to a child btree node, and + * the key is the highest key in the child node - except that the highest key in + * an interior node is always MAX_KEY. The size field refers to the size on disk + * of the child node - this would allow us to have variable sized btree nodes + * (handy for keeping the depth of the btree 1 by expanding just the root). + * + * Btree nodes are themselves log structured, but this is hidden fairly + * thoroughly. Btree nodes on disk will in practice have extents that overlap + * (because they were written at different times), but in memory we never have + * overlapping extents - when we read in a btree node from disk, the first thing + * we do is resort all the sets of keys with a mergesort, and in the same pass + * we check for overlapping extents and adjust them appropriately. + * + * struct btree_op is a central interface to the btree code. It's used for + * specifying read vs. write locking, and the embedded closure is used for + * waiting on IO or reserve memory. + * + * BTREE CACHE: + * + * Btree nodes are cached in memory; traversing the btree might require reading + * in btree nodes which is handled mostly transparently. + * + * bch_btree_node_get() looks up a btree node in the cache and reads it in from + * disk if necessary. This function is almost never called directly though - the + * btree() macro is used to get a btree node, call some function on it, and + * unlock the node after the function returns. + * + * The root is special cased - it's taken out of the cache's lru (thus pinning + * it in memory), so we can find the root of the btree by just dereferencing a + * pointer instead of looking it up in the cache. This makes locking a bit + * tricky, since the root pointer is protected by the lock in the btree node it + * points to - the btree_root() macro handles this. + * + * In various places we must be able to allocate memory for multiple btree nodes + * in order to make forward progress. To do this we use the btree cache itself + * as a reserve; if __get_free_pages() fails, we'll find a node in the btree + * cache we can reuse. We can't allow more than one thread to be doing this at a + * time, so there's a lock, implemented by a pointer to the btree_op closure - + * this allows the btree_root() macro to implicitly release this lock. + * + * BTREE IO: + * + * Btree nodes never have to be explicitly read in; bch_btree_node_get() handles + * this. + * + * For writing, we have two btree_write structs embeddded in struct btree - one + * write in flight, and one being set up, and we toggle between them. + * + * Writing is done with a single function - bch_btree_write() really serves two + * different purposes and should be broken up into two different functions. When + * passing now = false, it merely indicates that the node is now dirty - calling + * it ensures that the dirty keys will be written at some point in the future. + * + * When passing now = true, bch_btree_write() causes a write to happen + * "immediately" (if there was already a write in flight, it'll cause the write + * to happen as soon as the previous write completes). It returns immediately + * though - but it takes a refcount on the closure in struct btree_op you passed + * to it, so a closure_sync() later can be used to wait for the write to + * complete. + * + * This is handy because btree_split() and garbage collection can issue writes + * in parallel, reducing the amount of time they have to hold write locks. + * + * LOCKING: + * + * When traversing the btree, we may need write locks starting at some level - + * inserting a key into the btree will typically only require a write lock on + * the leaf node. + * + * This is specified with the lock field in struct btree_op; lock = 0 means we + * take write locks at level <= 0, i.e. only leaf nodes. bch_btree_node_get() + * checks this field and returns the node with the appropriate lock held. + * + * If, after traversing the btree, the insertion code discovers it has to split + * then it must restart from the root and take new locks - to do this it changes + * the lock field and returns -EINTR, which causes the btree_root() macro to + * loop. + * + * Handling cache misses require a different mechanism for upgrading to a write + * lock. We do cache lookups with only a read lock held, but if we get a cache + * miss and we wish to insert this data into the cache, we have to insert a + * placeholder key to detect races - otherwise, we could race with a write and + * overwrite the data that was just written to the cache with stale data from + * the backing device. + * + * For this we use a sequence number that write locks and unlocks increment - to + * insert the check key it unlocks the btree node and then takes a write lock, + * and fails if the sequence number doesn't match. + */ + +#include "bset.h" +#include "debug.h" + +struct btree_write { + atomic_t *journal; + + /* If btree_split() frees a btree node, it writes a new pointer to that + * btree node indicating it was freed; it takes a refcount on + * c->prio_blocked because we can't write the gens until the new + * pointer is on disk. This allows btree_write_endio() to release the + * refcount that btree_split() took. + */ + int prio_blocked; +}; + +struct btree { + /* Hottest entries first */ + struct hlist_node hash; + + /* Key/pointer for this btree node */ + BKEY_PADDED(key); + + unsigned long seq; + struct rw_semaphore lock; + struct cache_set *c; + struct btree *parent; + + struct mutex write_lock; + + unsigned long flags; + uint16_t written; /* would be nice to kill */ + uint8_t level; + + struct btree_keys keys; + + /* For outstanding btree writes, used as a lock - protects write_idx */ + struct closure io; + struct semaphore io_mutex; + + struct list_head list; + struct delayed_work work; + + struct btree_write writes[2]; + struct bio *bio; +}; + + + + +#define BTREE_FLAG(flag) \ +static inline bool btree_node_ ## flag(struct btree *b) \ +{ return test_bit(BTREE_NODE_ ## flag, &b->flags); } \ + \ +static inline void set_btree_node_ ## flag(struct btree *b) \ +{ set_bit(BTREE_NODE_ ## flag, &b->flags); } + +enum btree_flags { + BTREE_NODE_io_error, + BTREE_NODE_dirty, + BTREE_NODE_write_idx, + BTREE_NODE_journal_flush, +}; + +BTREE_FLAG(io_error); +BTREE_FLAG(dirty); +BTREE_FLAG(write_idx); +BTREE_FLAG(journal_flush); + +static inline struct btree_write *btree_current_write(struct btree *b) +{ + return b->writes + btree_node_write_idx(b); +} + +static inline struct btree_write *btree_prev_write(struct btree *b) +{ + return b->writes + (btree_node_write_idx(b) ^ 1); +} + +static inline struct bset *btree_bset_first(struct btree *b) +{ + return b->keys.set->data; +} + +static inline struct bset *btree_bset_last(struct btree *b) +{ + return bset_tree_last(&b->keys)->data; +} + +static inline unsigned int bset_block_offset(struct btree *b, struct bset *i) +{ + return bset_sector_offset(&b->keys, i) >> b->c->block_bits; +} + +static inline void set_gc_sectors(struct cache_set *c) +{ + atomic_set(&c->sectors_to_gc, c->cache->sb.bucket_size * c->nbuckets / 16); +} + +void bkey_put(struct cache_set *c, struct bkey *k); + +/* Looping macros */ + +#define for_each_cached_btree(b, c, iter) \ + for (iter = 0; \ + iter < ARRAY_SIZE((c)->bucket_hash); \ + iter++) \ + hlist_for_each_entry_rcu((b), (c)->bucket_hash + iter, hash) + +/* Recursing down the btree */ + +struct btree_op { + /* for waiting on btree reserve in btree_split() */ + wait_queue_entry_t wait; + + /* Btree level at which we start taking write locks */ + short lock; + + unsigned int insert_collision:1; +}; + +struct btree_check_state; +struct btree_check_info { + struct btree_check_state *state; + struct task_struct *thread; + int result; +}; + +#define BCH_BTR_CHKTHREAD_MAX 12 +struct btree_check_state { + struct cache_set *c; + int total_threads; + int key_idx; + spinlock_t idx_lock; + atomic_t started; + atomic_t enough; + wait_queue_head_t wait; + struct btree_check_info infos[BCH_BTR_CHKTHREAD_MAX]; +}; + +static inline void bch_btree_op_init(struct btree_op *op, int write_lock_level) +{ + memset(op, 0, sizeof(struct btree_op)); + init_wait(&op->wait); + op->lock = write_lock_level; +} + +static inline void rw_lock(bool w, struct btree *b, int level) +{ + w ? down_write_nested(&b->lock, level + 1) + : down_read_nested(&b->lock, level + 1); + if (w) + b->seq++; +} + +static inline void rw_unlock(bool w, struct btree *b) +{ + if (w) + b->seq++; + (w ? up_write : up_read)(&b->lock); +} + +void bch_btree_node_read_done(struct btree *b); +void __bch_btree_node_write(struct btree *b, struct closure *parent); +void bch_btree_node_write(struct btree *b, struct closure *parent); + +void bch_btree_set_root(struct btree *b); +struct btree *__bch_btree_node_alloc(struct cache_set *c, struct btree_op *op, + int level, bool wait, + struct btree *parent); +struct btree *bch_btree_node_get(struct cache_set *c, struct btree_op *op, + struct bkey *k, int level, bool write, + struct btree *parent); + +int bch_btree_insert_check_key(struct btree *b, struct btree_op *op, + struct bkey *check_key); +int bch_btree_insert(struct cache_set *c, struct keylist *keys, + atomic_t *journal_ref, struct bkey *replace_key); + +int bch_gc_thread_start(struct cache_set *c); +void bch_initial_gc_finish(struct cache_set *c); +void bch_moving_gc(struct cache_set *c); +int bch_btree_check(struct cache_set *c); +void bch_initial_mark_key(struct cache_set *c, int level, struct bkey *k); +void bch_cannibalize_unlock(struct cache_set *c); + +static inline void wake_up_gc(struct cache_set *c) +{ + wake_up(&c->gc_wait); +} + +static inline void force_wake_up_gc(struct cache_set *c) +{ + /* + * Garbage collection thread only works when sectors_to_gc < 0, + * calling wake_up_gc() won't start gc thread if sectors_to_gc is + * not a nagetive value. + * Therefore sectors_to_gc is set to -1 here, before waking up + * gc thread by calling wake_up_gc(). Then gc_should_run() will + * give a chance to permit gc thread to run. "Give a chance" means + * before going into gc_should_run(), there is still possibility + * that c->sectors_to_gc being set to other positive value. So + * this routine won't 100% make sure gc thread will be woken up + * to run. + */ + atomic_set(&c->sectors_to_gc, -1); + wake_up_gc(c); +} + +/* + * These macros are for recursing down the btree - they handle the details of + * locking and looking up nodes in the cache for you. They're best treated as + * mere syntax when reading code that uses them. + * + * op->lock determines whether we take a read or a write lock at a given depth. + * If you've got a read lock and find that you need a write lock (i.e. you're + * going to have to split), set op->lock and return -EINTR; btree_root() will + * call you again and you'll have the correct lock. + */ + +/** + * btree - recurse down the btree on a specified key + * @fn: function to call, which will be passed the child node + * @key: key to recurse on + * @b: parent btree node + * @op: pointer to struct btree_op + */ +#define bcache_btree(fn, key, b, op, ...) \ +({ \ + int _r, l = (b)->level - 1; \ + bool _w = l <= (op)->lock; \ + struct btree *_child = bch_btree_node_get((b)->c, op, key, l, \ + _w, b); \ + if (!IS_ERR(_child)) { \ + _r = bch_btree_ ## fn(_child, op, ##__VA_ARGS__); \ + rw_unlock(_w, _child); \ + } else \ + _r = PTR_ERR(_child); \ + _r; \ +}) + +/** + * btree_root - call a function on the root of the btree + * @fn: function to call, which will be passed the child node + * @c: cache set + * @op: pointer to struct btree_op + */ +#define bcache_btree_root(fn, c, op, ...) \ +({ \ + int _r = -EINTR; \ + do { \ + struct btree *_b = (c)->root; \ + bool _w = insert_lock(op, _b); \ + rw_lock(_w, _b, _b->level); \ + if (_b == (c)->root && \ + _w == insert_lock(op, _b)) { \ + _r = bch_btree_ ## fn(_b, op, ##__VA_ARGS__); \ + } \ + rw_unlock(_w, _b); \ + bch_cannibalize_unlock(c); \ + if (_r == -EINTR) \ + schedule(); \ + } while (_r == -EINTR); \ + \ + finish_wait(&(c)->btree_cache_wait, &(op)->wait); \ + _r; \ +}) + +#define MAP_DONE 0 +#define MAP_CONTINUE 1 + +#define MAP_ALL_NODES 0 +#define MAP_LEAF_NODES 1 + +#define MAP_END_KEY 1 + +typedef int (btree_map_nodes_fn)(struct btree_op *b_op, struct btree *b); +int __bch_btree_map_nodes(struct btree_op *op, struct cache_set *c, + struct bkey *from, btree_map_nodes_fn *fn, int flags); + +static inline int bch_btree_map_nodes(struct btree_op *op, struct cache_set *c, + struct bkey *from, btree_map_nodes_fn *fn) +{ + return __bch_btree_map_nodes(op, c, from, fn, MAP_ALL_NODES); +} + +static inline int bch_btree_map_leaf_nodes(struct btree_op *op, + struct cache_set *c, + struct bkey *from, + btree_map_nodes_fn *fn) +{ + return __bch_btree_map_nodes(op, c, from, fn, MAP_LEAF_NODES); +} + +typedef int (btree_map_keys_fn)(struct btree_op *op, struct btree *b, + struct bkey *k); +int bch_btree_map_keys(struct btree_op *op, struct cache_set *c, + struct bkey *from, btree_map_keys_fn *fn, int flags); +int bch_btree_map_keys_recurse(struct btree *b, struct btree_op *op, + struct bkey *from, btree_map_keys_fn *fn, + int flags); + +typedef bool (keybuf_pred_fn)(struct keybuf *buf, struct bkey *k); + +void bch_keybuf_init(struct keybuf *buf); +void bch_refill_keybuf(struct cache_set *c, struct keybuf *buf, + struct bkey *end, keybuf_pred_fn *pred); +bool bch_keybuf_check_overlapping(struct keybuf *buf, struct bkey *start, + struct bkey *end); +void bch_keybuf_del(struct keybuf *buf, struct keybuf_key *w); +struct keybuf_key *bch_keybuf_next(struct keybuf *buf); +struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *c, + struct keybuf *buf, + struct bkey *end, + keybuf_pred_fn *pred); +void bch_update_bucket_in_use(struct cache_set *c, struct gc_stat *stats); +#endif diff --git a/drivers/md/bcache/closure.c b/drivers/md/bcache/closure.c new file mode 100644 index 000000000..d8d9394a6 --- /dev/null +++ b/drivers/md/bcache/closure.c @@ -0,0 +1,207 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Asynchronous refcounty things + * + * Copyright 2010, 2011 Kent Overstreet <kent.overstreet@gmail.com> + * Copyright 2012 Google, Inc. + */ + +#include <linux/debugfs.h> +#include <linux/module.h> +#include <linux/seq_file.h> +#include <linux/sched/debug.h> + +#include "closure.h" + +static inline void closure_put_after_sub(struct closure *cl, int flags) +{ + int r = flags & CLOSURE_REMAINING_MASK; + + BUG_ON(flags & CLOSURE_GUARD_MASK); + BUG_ON(!r && (flags & ~CLOSURE_DESTRUCTOR)); + + if (!r) { + if (cl->fn && !(flags & CLOSURE_DESTRUCTOR)) { + atomic_set(&cl->remaining, + CLOSURE_REMAINING_INITIALIZER); + closure_queue(cl); + } else { + struct closure *parent = cl->parent; + closure_fn *destructor = cl->fn; + + closure_debug_destroy(cl); + + if (destructor) + destructor(cl); + + if (parent) + closure_put(parent); + } + } +} + +/* For clearing flags with the same atomic op as a put */ +void closure_sub(struct closure *cl, int v) +{ + closure_put_after_sub(cl, atomic_sub_return(v, &cl->remaining)); +} + +/* + * closure_put - decrement a closure's refcount + */ +void closure_put(struct closure *cl) +{ + closure_put_after_sub(cl, atomic_dec_return(&cl->remaining)); +} + +/* + * closure_wake_up - wake up all closures on a wait list, without memory barrier + */ +void __closure_wake_up(struct closure_waitlist *wait_list) +{ + struct llist_node *list; + struct closure *cl, *t; + struct llist_node *reverse = NULL; + + list = llist_del_all(&wait_list->list); + + /* We first reverse the list to preserve FIFO ordering and fairness */ + reverse = llist_reverse_order(list); + + /* Then do the wakeups */ + llist_for_each_entry_safe(cl, t, reverse, list) { + closure_set_waiting(cl, 0); + closure_sub(cl, CLOSURE_WAITING + 1); + } +} + +/** + * closure_wait - add a closure to a waitlist + * @waitlist: will own a ref on @cl, which will be released when + * closure_wake_up() is called on @waitlist. + * @cl: closure pointer. + * + */ +bool closure_wait(struct closure_waitlist *waitlist, struct closure *cl) +{ + if (atomic_read(&cl->remaining) & CLOSURE_WAITING) + return false; + + closure_set_waiting(cl, _RET_IP_); + atomic_add(CLOSURE_WAITING + 1, &cl->remaining); + llist_add(&cl->list, &waitlist->list); + + return true; +} + +struct closure_syncer { + struct task_struct *task; + int done; +}; + +static void closure_sync_fn(struct closure *cl) +{ + struct closure_syncer *s = cl->s; + struct task_struct *p; + + rcu_read_lock(); + p = READ_ONCE(s->task); + s->done = 1; + wake_up_process(p); + rcu_read_unlock(); +} + +void __sched __closure_sync(struct closure *cl) +{ + struct closure_syncer s = { .task = current }; + + cl->s = &s; + continue_at(cl, closure_sync_fn, NULL); + + while (1) { + set_current_state(TASK_UNINTERRUPTIBLE); + if (s.done) + break; + schedule(); + } + + __set_current_state(TASK_RUNNING); +} + +#ifdef CONFIG_BCACHE_CLOSURES_DEBUG + +static LIST_HEAD(closure_list); +static DEFINE_SPINLOCK(closure_list_lock); + +void closure_debug_create(struct closure *cl) +{ + unsigned long flags; + + BUG_ON(cl->magic == CLOSURE_MAGIC_ALIVE); + cl->magic = CLOSURE_MAGIC_ALIVE; + + spin_lock_irqsave(&closure_list_lock, flags); + list_add(&cl->all, &closure_list); + spin_unlock_irqrestore(&closure_list_lock, flags); +} + +void closure_debug_destroy(struct closure *cl) +{ + unsigned long flags; + + BUG_ON(cl->magic != CLOSURE_MAGIC_ALIVE); + cl->magic = CLOSURE_MAGIC_DEAD; + + spin_lock_irqsave(&closure_list_lock, flags); + list_del(&cl->all); + spin_unlock_irqrestore(&closure_list_lock, flags); +} + +static struct dentry *closure_debug; + +static int debug_show(struct seq_file *f, void *data) +{ + struct closure *cl; + + spin_lock_irq(&closure_list_lock); + + list_for_each_entry(cl, &closure_list, all) { + int r = atomic_read(&cl->remaining); + + seq_printf(f, "%p: %pS -> %pS p %p r %i ", + cl, (void *) cl->ip, cl->fn, cl->parent, + r & CLOSURE_REMAINING_MASK); + + seq_printf(f, "%s%s\n", + test_bit(WORK_STRUCT_PENDING_BIT, + work_data_bits(&cl->work)) ? "Q" : "", + r & CLOSURE_RUNNING ? "R" : ""); + + if (r & CLOSURE_WAITING) + seq_printf(f, " W %pS\n", + (void *) cl->waiting_on); + + seq_printf(f, "\n"); + } + + spin_unlock_irq(&closure_list_lock); + return 0; +} + +DEFINE_SHOW_ATTRIBUTE(debug); + +void __init closure_debug_init(void) +{ + if (!IS_ERR_OR_NULL(bcache_debug)) + /* + * it is unnecessary to check return value of + * debugfs_create_file(), we should not care + * about this. + */ + closure_debug = debugfs_create_file( + "closures", 0400, bcache_debug, NULL, &debug_fops); +} +#endif + +MODULE_AUTHOR("Kent Overstreet <koverstreet@google.com>"); +MODULE_LICENSE("GPL"); diff --git a/drivers/md/bcache/closure.h b/drivers/md/bcache/closure.h new file mode 100644 index 000000000..c88cdc4ae --- /dev/null +++ b/drivers/md/bcache/closure.h @@ -0,0 +1,378 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _LINUX_CLOSURE_H +#define _LINUX_CLOSURE_H + +#include <linux/llist.h> +#include <linux/sched.h> +#include <linux/sched/task_stack.h> +#include <linux/workqueue.h> + +/* + * Closure is perhaps the most overused and abused term in computer science, but + * since I've been unable to come up with anything better you're stuck with it + * again. + * + * What are closures? + * + * They embed a refcount. The basic idea is they count "things that are in + * progress" - in flight bios, some other thread that's doing something else - + * anything you might want to wait on. + * + * The refcount may be manipulated with closure_get() and closure_put(). + * closure_put() is where many of the interesting things happen, when it causes + * the refcount to go to 0. + * + * Closures can be used to wait on things both synchronously and asynchronously, + * and synchronous and asynchronous use can be mixed without restriction. To + * wait synchronously, use closure_sync() - you will sleep until your closure's + * refcount hits 1. + * + * To wait asynchronously, use + * continue_at(cl, next_function, workqueue); + * + * passing it, as you might expect, the function to run when nothing is pending + * and the workqueue to run that function out of. + * + * continue_at() also, critically, requires a 'return' immediately following the + * location where this macro is referenced, to return to the calling function. + * There's good reason for this. + * + * To use safely closures asynchronously, they must always have a refcount while + * they are running owned by the thread that is running them. Otherwise, suppose + * you submit some bios and wish to have a function run when they all complete: + * + * foo_endio(struct bio *bio) + * { + * closure_put(cl); + * } + * + * closure_init(cl); + * + * do_stuff(); + * closure_get(cl); + * bio1->bi_endio = foo_endio; + * bio_submit(bio1); + * + * do_more_stuff(); + * closure_get(cl); + * bio2->bi_endio = foo_endio; + * bio_submit(bio2); + * + * continue_at(cl, complete_some_read, system_wq); + * + * If closure's refcount started at 0, complete_some_read() could run before the + * second bio was submitted - which is almost always not what you want! More + * importantly, it wouldn't be possible to say whether the original thread or + * complete_some_read()'s thread owned the closure - and whatever state it was + * associated with! + * + * So, closure_init() initializes a closure's refcount to 1 - and when a + * closure_fn is run, the refcount will be reset to 1 first. + * + * Then, the rule is - if you got the refcount with closure_get(), release it + * with closure_put() (i.e, in a bio->bi_endio function). If you have a refcount + * on a closure because you called closure_init() or you were run out of a + * closure - _always_ use continue_at(). Doing so consistently will help + * eliminate an entire class of particularly pernicious races. + * + * Lastly, you might have a wait list dedicated to a specific event, and have no + * need for specifying the condition - you just want to wait until someone runs + * closure_wake_up() on the appropriate wait list. In that case, just use + * closure_wait(). It will return either true or false, depending on whether the + * closure was already on a wait list or not - a closure can only be on one wait + * list at a time. + * + * Parents: + * + * closure_init() takes two arguments - it takes the closure to initialize, and + * a (possibly null) parent. + * + * If parent is non null, the new closure will have a refcount for its lifetime; + * a closure is considered to be "finished" when its refcount hits 0 and the + * function to run is null. Hence + * + * continue_at(cl, NULL, NULL); + * + * returns up the (spaghetti) stack of closures, precisely like normal return + * returns up the C stack. continue_at() with non null fn is better thought of + * as doing a tail call. + * + * All this implies that a closure should typically be embedded in a particular + * struct (which its refcount will normally control the lifetime of), and that + * struct can very much be thought of as a stack frame. + */ + +struct closure; +struct closure_syncer; +typedef void (closure_fn) (struct closure *); +extern struct dentry *bcache_debug; + +struct closure_waitlist { + struct llist_head list; +}; + +enum closure_state { + /* + * CLOSURE_WAITING: Set iff the closure is on a waitlist. Must be set by + * the thread that owns the closure, and cleared by the thread that's + * waking up the closure. + * + * The rest are for debugging and don't affect behaviour: + * + * CLOSURE_RUNNING: Set when a closure is running (i.e. by + * closure_init() and when closure_put() runs then next function), and + * must be cleared before remaining hits 0. Primarily to help guard + * against incorrect usage and accidentally transferring references. + * continue_at() and closure_return() clear it for you, if you're doing + * something unusual you can use closure_set_dead() which also helps + * annotate where references are being transferred. + */ + + CLOSURE_BITS_START = (1U << 26), + CLOSURE_DESTRUCTOR = (1U << 26), + CLOSURE_WAITING = (1U << 28), + CLOSURE_RUNNING = (1U << 30), +}; + +#define CLOSURE_GUARD_MASK \ + ((CLOSURE_DESTRUCTOR|CLOSURE_WAITING|CLOSURE_RUNNING) << 1) + +#define CLOSURE_REMAINING_MASK (CLOSURE_BITS_START - 1) +#define CLOSURE_REMAINING_INITIALIZER (1|CLOSURE_RUNNING) + +struct closure { + union { + struct { + struct workqueue_struct *wq; + struct closure_syncer *s; + struct llist_node list; + closure_fn *fn; + }; + struct work_struct work; + }; + + struct closure *parent; + + atomic_t remaining; + +#ifdef CONFIG_BCACHE_CLOSURES_DEBUG +#define CLOSURE_MAGIC_DEAD 0xc054dead +#define CLOSURE_MAGIC_ALIVE 0xc054a11e + + unsigned int magic; + struct list_head all; + unsigned long ip; + unsigned long waiting_on; +#endif +}; + +void closure_sub(struct closure *cl, int v); +void closure_put(struct closure *cl); +void __closure_wake_up(struct closure_waitlist *list); +bool closure_wait(struct closure_waitlist *list, struct closure *cl); +void __closure_sync(struct closure *cl); + +/** + * closure_sync - sleep until a closure a closure has nothing left to wait on + * + * Sleeps until the refcount hits 1 - the thread that's running the closure owns + * the last refcount. + */ +static inline void closure_sync(struct closure *cl) +{ + if ((atomic_read(&cl->remaining) & CLOSURE_REMAINING_MASK) != 1) + __closure_sync(cl); +} + +#ifdef CONFIG_BCACHE_CLOSURES_DEBUG + +void closure_debug_init(void); +void closure_debug_create(struct closure *cl); +void closure_debug_destroy(struct closure *cl); + +#else + +static inline void closure_debug_init(void) {} +static inline void closure_debug_create(struct closure *cl) {} +static inline void closure_debug_destroy(struct closure *cl) {} + +#endif + +static inline void closure_set_ip(struct closure *cl) +{ +#ifdef CONFIG_BCACHE_CLOSURES_DEBUG + cl->ip = _THIS_IP_; +#endif +} + +static inline void closure_set_ret_ip(struct closure *cl) +{ +#ifdef CONFIG_BCACHE_CLOSURES_DEBUG + cl->ip = _RET_IP_; +#endif +} + +static inline void closure_set_waiting(struct closure *cl, unsigned long f) +{ +#ifdef CONFIG_BCACHE_CLOSURES_DEBUG + cl->waiting_on = f; +#endif +} + +static inline void closure_set_stopped(struct closure *cl) +{ + atomic_sub(CLOSURE_RUNNING, &cl->remaining); +} + +static inline void set_closure_fn(struct closure *cl, closure_fn *fn, + struct workqueue_struct *wq) +{ + closure_set_ip(cl); + cl->fn = fn; + cl->wq = wq; + /* between atomic_dec() in closure_put() */ + smp_mb__before_atomic(); +} + +static inline void closure_queue(struct closure *cl) +{ + struct workqueue_struct *wq = cl->wq; + /** + * Changes made to closure, work_struct, or a couple of other structs + * may cause work.func not pointing to the right location. + */ + BUILD_BUG_ON(offsetof(struct closure, fn) + != offsetof(struct work_struct, func)); + if (wq) { + INIT_WORK(&cl->work, cl->work.func); + BUG_ON(!queue_work(wq, &cl->work)); + } else + cl->fn(cl); +} + +/** + * closure_get - increment a closure's refcount + */ +static inline void closure_get(struct closure *cl) +{ +#ifdef CONFIG_BCACHE_CLOSURES_DEBUG + BUG_ON((atomic_inc_return(&cl->remaining) & + CLOSURE_REMAINING_MASK) <= 1); +#else + atomic_inc(&cl->remaining); +#endif +} + +/** + * closure_init - Initialize a closure, setting the refcount to 1 + * @cl: closure to initialize + * @parent: parent of the new closure. cl will take a refcount on it for its + * lifetime; may be NULL. + */ +static inline void closure_init(struct closure *cl, struct closure *parent) +{ + memset(cl, 0, sizeof(struct closure)); + cl->parent = parent; + if (parent) + closure_get(parent); + + atomic_set(&cl->remaining, CLOSURE_REMAINING_INITIALIZER); + + closure_debug_create(cl); + closure_set_ip(cl); +} + +static inline void closure_init_stack(struct closure *cl) +{ + memset(cl, 0, sizeof(struct closure)); + atomic_set(&cl->remaining, CLOSURE_REMAINING_INITIALIZER); +} + +/** + * closure_wake_up - wake up all closures on a wait list, + * with memory barrier + */ +static inline void closure_wake_up(struct closure_waitlist *list) +{ + /* Memory barrier for the wait list */ + smp_mb(); + __closure_wake_up(list); +} + +/** + * continue_at - jump to another function with barrier + * + * After @cl is no longer waiting on anything (i.e. all outstanding refs have + * been dropped with closure_put()), it will resume execution at @fn running out + * of @wq (or, if @wq is NULL, @fn will be called by closure_put() directly). + * + * This is because after calling continue_at() you no longer have a ref on @cl, + * and whatever @cl owns may be freed out from under you - a running closure fn + * has a ref on its own closure which continue_at() drops. + * + * Note you are expected to immediately return after using this macro. + */ +#define continue_at(_cl, _fn, _wq) \ +do { \ + set_closure_fn(_cl, _fn, _wq); \ + closure_sub(_cl, CLOSURE_RUNNING + 1); \ +} while (0) + +/** + * closure_return - finish execution of a closure + * + * This is used to indicate that @cl is finished: when all outstanding refs on + * @cl have been dropped @cl's ref on its parent closure (as passed to + * closure_init()) will be dropped, if one was specified - thus this can be + * thought of as returning to the parent closure. + */ +#define closure_return(_cl) continue_at((_cl), NULL, NULL) + +/** + * continue_at_nobarrier - jump to another function without barrier + * + * Causes @fn to be executed out of @cl, in @wq context (or called directly if + * @wq is NULL). + * + * The ref the caller of continue_at_nobarrier() had on @cl is now owned by @fn, + * thus it's not safe to touch anything protected by @cl after a + * continue_at_nobarrier(). + */ +#define continue_at_nobarrier(_cl, _fn, _wq) \ +do { \ + set_closure_fn(_cl, _fn, _wq); \ + closure_queue(_cl); \ +} while (0) + +/** + * closure_return_with_destructor - finish execution of a closure, + * with destructor + * + * Works like closure_return(), except @destructor will be called when all + * outstanding refs on @cl have been dropped; @destructor may be used to safely + * free the memory occupied by @cl, and it is called with the ref on the parent + * closure still held - so @destructor could safely return an item to a + * freelist protected by @cl's parent. + */ +#define closure_return_with_destructor(_cl, _destructor) \ +do { \ + set_closure_fn(_cl, _destructor, NULL); \ + closure_sub(_cl, CLOSURE_RUNNING - CLOSURE_DESTRUCTOR + 1); \ +} while (0) + +/** + * closure_call - execute @fn out of a new, uninitialized closure + * + * Typically used when running out of one closure, and we want to run @fn + * asynchronously out of a new closure - @parent will then wait for @cl to + * finish. + */ +static inline void closure_call(struct closure *cl, closure_fn fn, + struct workqueue_struct *wq, + struct closure *parent) +{ + closure_init(cl, parent); + continue_at_nobarrier(cl, fn, wq); +} + +#endif /* _LINUX_CLOSURE_H */ diff --git a/drivers/md/bcache/debug.c b/drivers/md/bcache/debug.c new file mode 100644 index 000000000..b00fd08d6 --- /dev/null +++ b/drivers/md/bcache/debug.c @@ -0,0 +1,262 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Assorted bcache debug code + * + * Copyright 2010, 2011 Kent Overstreet <kent.overstreet@gmail.com> + * Copyright 2012 Google, Inc. + */ + +#include "bcache.h" +#include "btree.h" +#include "debug.h" +#include "extents.h" + +#include <linux/console.h> +#include <linux/debugfs.h> +#include <linux/module.h> +#include <linux/random.h> +#include <linux/seq_file.h> + +struct dentry *bcache_debug; + +#ifdef CONFIG_BCACHE_DEBUG + +#define for_each_written_bset(b, start, i) \ + for (i = (start); \ + (void *) i < (void *) (start) + (KEY_SIZE(&b->key) << 9) &&\ + i->seq == (start)->seq; \ + i = (void *) i + set_blocks(i, block_bytes(b->c->cache)) * \ + block_bytes(b->c->cache)) + +void bch_btree_verify(struct btree *b) +{ + struct btree *v = b->c->verify_data; + struct bset *ondisk, *sorted, *inmemory; + struct bio *bio; + + if (!b->c->verify || !b->c->verify_ondisk) + return; + + down(&b->io_mutex); + mutex_lock(&b->c->verify_lock); + + ondisk = b->c->verify_ondisk; + sorted = b->c->verify_data->keys.set->data; + inmemory = b->keys.set->data; + + bkey_copy(&v->key, &b->key); + v->written = 0; + v->level = b->level; + v->keys.ops = b->keys.ops; + + bio = bch_bbio_alloc(b->c); + bio_set_dev(bio, PTR_CACHE(b->c, &b->key, 0)->bdev); + bio->bi_iter.bi_sector = PTR_OFFSET(&b->key, 0); + bio->bi_iter.bi_size = KEY_SIZE(&v->key) << 9; + bio->bi_opf = REQ_OP_READ | REQ_META; + bch_bio_map(bio, sorted); + + submit_bio_wait(bio); + bch_bbio_free(bio, b->c); + + memcpy(ondisk, sorted, KEY_SIZE(&v->key) << 9); + + bch_btree_node_read_done(v); + sorted = v->keys.set->data; + + if (inmemory->keys != sorted->keys || + memcmp(inmemory->start, + sorted->start, + (void *) bset_bkey_last(inmemory) - + (void *) inmemory->start)) { + struct bset *i; + unsigned int j; + + console_lock(); + + pr_err("*** in memory:\n"); + bch_dump_bset(&b->keys, inmemory, 0); + + pr_err("*** read back in:\n"); + bch_dump_bset(&v->keys, sorted, 0); + + for_each_written_bset(b, ondisk, i) { + unsigned int block = ((void *) i - (void *) ondisk) / + block_bytes(b->c->cache); + + pr_err("*** on disk block %u:\n", block); + bch_dump_bset(&b->keys, i, block); + } + + pr_err("*** block %zu not written\n", + ((void *) i - (void *) ondisk) / block_bytes(b->c->cache)); + + for (j = 0; j < inmemory->keys; j++) + if (inmemory->d[j] != sorted->d[j]) + break; + + pr_err("b->written %u\n", b->written); + + console_unlock(); + panic("verify failed at %u\n", j); + } + + mutex_unlock(&b->c->verify_lock); + up(&b->io_mutex); +} + +void bch_data_verify(struct cached_dev *dc, struct bio *bio) +{ + struct bio *check; + struct bio_vec bv, cbv; + struct bvec_iter iter, citer = { 0 }; + + check = bio_kmalloc(GFP_NOIO, bio_segments(bio)); + if (!check) + return; + check->bi_disk = bio->bi_disk; + check->bi_opf = REQ_OP_READ; + check->bi_iter.bi_sector = bio->bi_iter.bi_sector; + check->bi_iter.bi_size = bio->bi_iter.bi_size; + + bch_bio_map(check, NULL); + if (bch_bio_alloc_pages(check, GFP_NOIO)) + goto out_put; + + submit_bio_wait(check); + + citer.bi_size = UINT_MAX; + bio_for_each_segment(bv, bio, iter) { + void *p1 = kmap_atomic(bv.bv_page); + void *p2; + + cbv = bio_iter_iovec(check, citer); + p2 = page_address(cbv.bv_page); + + cache_set_err_on(memcmp(p1 + bv.bv_offset, + p2 + bv.bv_offset, + bv.bv_len), + dc->disk.c, + "verify failed at dev %s sector %llu", + dc->backing_dev_name, + (uint64_t) bio->bi_iter.bi_sector); + + kunmap_atomic(p1); + bio_advance_iter(check, &citer, bv.bv_len); + } + + bio_free_pages(check); +out_put: + bio_put(check); +} + +#endif + +#ifdef CONFIG_DEBUG_FS + +/* XXX: cache set refcounting */ + +struct dump_iterator { + char buf[PAGE_SIZE]; + size_t bytes; + struct cache_set *c; + struct keybuf keys; +}; + +static bool dump_pred(struct keybuf *buf, struct bkey *k) +{ + return true; +} + +static ssize_t bch_dump_read(struct file *file, char __user *buf, + size_t size, loff_t *ppos) +{ + struct dump_iterator *i = file->private_data; + ssize_t ret = 0; + char kbuf[80]; + + while (size) { + struct keybuf_key *w; + unsigned int bytes = min(i->bytes, size); + + if (copy_to_user(buf, i->buf, bytes)) + return -EFAULT; + + ret += bytes; + buf += bytes; + size -= bytes; + i->bytes -= bytes; + memmove(i->buf, i->buf + bytes, i->bytes); + + if (i->bytes) + break; + + w = bch_keybuf_next_rescan(i->c, &i->keys, &MAX_KEY, dump_pred); + if (!w) + break; + + bch_extent_to_text(kbuf, sizeof(kbuf), &w->key); + i->bytes = snprintf(i->buf, PAGE_SIZE, "%s\n", kbuf); + bch_keybuf_del(&i->keys, w); + } + + return ret; +} + +static int bch_dump_open(struct inode *inode, struct file *file) +{ + struct cache_set *c = inode->i_private; + struct dump_iterator *i; + + i = kzalloc(sizeof(struct dump_iterator), GFP_KERNEL); + if (!i) + return -ENOMEM; + + file->private_data = i; + i->c = c; + bch_keybuf_init(&i->keys); + i->keys.last_scanned = KEY(0, 0, 0); + + return 0; +} + +static int bch_dump_release(struct inode *inode, struct file *file) +{ + kfree(file->private_data); + return 0; +} + +static const struct file_operations cache_set_debug_ops = { + .owner = THIS_MODULE, + .open = bch_dump_open, + .read = bch_dump_read, + .release = bch_dump_release +}; + +void bch_debug_init_cache_set(struct cache_set *c) +{ + if (!IS_ERR_OR_NULL(bcache_debug)) { + char name[50]; + + snprintf(name, 50, "bcache-%pU", c->set_uuid); + c->debug = debugfs_create_file(name, 0400, bcache_debug, c, + &cache_set_debug_ops); + } +} + +#endif + +void bch_debug_exit(void) +{ + debugfs_remove_recursive(bcache_debug); +} + +void __init bch_debug_init(void) +{ + /* + * it is unnecessary to check return value of + * debugfs_create_file(), we should not care + * about this. + */ + bcache_debug = debugfs_create_dir("bcache", NULL); +} diff --git a/drivers/md/bcache/debug.h b/drivers/md/bcache/debug.h new file mode 100644 index 000000000..fb3d4dff4 --- /dev/null +++ b/drivers/md/bcache/debug.h @@ -0,0 +1,35 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _BCACHE_DEBUG_H +#define _BCACHE_DEBUG_H + +struct bio; +struct cached_dev; +struct cache_set; + +#ifdef CONFIG_BCACHE_DEBUG + +void bch_btree_verify(struct btree *b); +void bch_data_verify(struct cached_dev *dc, struct bio *bio); + +#define expensive_debug_checks(c) ((c)->expensive_debug_checks) +#define key_merging_disabled(c) ((c)->key_merging_disabled) +#define bypass_torture_test(d) ((d)->bypass_torture_test) + +#else /* DEBUG */ + +static inline void bch_btree_verify(struct btree *b) {} +static inline void bch_data_verify(struct cached_dev *dc, struct bio *bio) {} + +#define expensive_debug_checks(c) 0 +#define key_merging_disabled(c) 0 +#define bypass_torture_test(d) 0 + +#endif + +#ifdef CONFIG_DEBUG_FS +void bch_debug_init_cache_set(struct cache_set *c); +#else +static inline void bch_debug_init_cache_set(struct cache_set *c) {} +#endif + +#endif diff --git a/drivers/md/bcache/extents.c b/drivers/md/bcache/extents.c new file mode 100644 index 000000000..f4658a1f3 --- /dev/null +++ b/drivers/md/bcache/extents.c @@ -0,0 +1,630 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2010 Kent Overstreet <kent.overstreet@gmail.com> + * + * Uses a block device as cache for other block devices; optimized for SSDs. + * All allocation is done in buckets, which should match the erase block size + * of the device. + * + * Buckets containing cached data are kept on a heap sorted by priority; + * bucket priority is increased on cache hit, and periodically all the buckets + * on the heap have their priority scaled down. This currently is just used as + * an LRU but in the future should allow for more intelligent heuristics. + * + * Buckets have an 8 bit counter; freeing is accomplished by incrementing the + * counter. Garbage collection is used to remove stale pointers. + * + * Indexing is done via a btree; nodes are not necessarily fully sorted, rather + * as keys are inserted we only sort the pages that have not yet been written. + * When garbage collection is run, we resort the entire node. + * + * All configuration is done via sysfs; see Documentation/admin-guide/bcache.rst. + */ + +#include "bcache.h" +#include "btree.h" +#include "debug.h" +#include "extents.h" +#include "writeback.h" + +static void sort_key_next(struct btree_iter *iter, + struct btree_iter_set *i) +{ + i->k = bkey_next(i->k); + + if (i->k == i->end) + *i = iter->data[--iter->used]; +} + +static bool bch_key_sort_cmp(struct btree_iter_set l, + struct btree_iter_set r) +{ + int64_t c = bkey_cmp(l.k, r.k); + + return c ? c > 0 : l.k < r.k; +} + +static bool __ptr_invalid(struct cache_set *c, const struct bkey *k) +{ + unsigned int i; + + for (i = 0; i < KEY_PTRS(k); i++) + if (ptr_available(c, k, i)) { + struct cache *ca = PTR_CACHE(c, k, i); + size_t bucket = PTR_BUCKET_NR(c, k, i); + size_t r = bucket_remainder(c, PTR_OFFSET(k, i)); + + if (KEY_SIZE(k) + r > c->cache->sb.bucket_size || + bucket < ca->sb.first_bucket || + bucket >= ca->sb.nbuckets) + return true; + } + + return false; +} + +/* Common among btree and extent ptrs */ + +static const char *bch_ptr_status(struct cache_set *c, const struct bkey *k) +{ + unsigned int i; + + for (i = 0; i < KEY_PTRS(k); i++) + if (ptr_available(c, k, i)) { + struct cache *ca = PTR_CACHE(c, k, i); + size_t bucket = PTR_BUCKET_NR(c, k, i); + size_t r = bucket_remainder(c, PTR_OFFSET(k, i)); + + if (KEY_SIZE(k) + r > c->cache->sb.bucket_size) + return "bad, length too big"; + if (bucket < ca->sb.first_bucket) + return "bad, short offset"; + if (bucket >= ca->sb.nbuckets) + return "bad, offset past end of device"; + if (ptr_stale(c, k, i)) + return "stale"; + } + + if (!bkey_cmp(k, &ZERO_KEY)) + return "bad, null key"; + if (!KEY_PTRS(k)) + return "bad, no pointers"; + if (!KEY_SIZE(k)) + return "zeroed key"; + return ""; +} + +void bch_extent_to_text(char *buf, size_t size, const struct bkey *k) +{ + unsigned int i = 0; + char *out = buf, *end = buf + size; + +#define p(...) (out += scnprintf(out, end - out, __VA_ARGS__)) + + p("%llu:%llu len %llu -> [", KEY_INODE(k), KEY_START(k), KEY_SIZE(k)); + + for (i = 0; i < KEY_PTRS(k); i++) { + if (i) + p(", "); + + if (PTR_DEV(k, i) == PTR_CHECK_DEV) + p("check dev"); + else + p("%llu:%llu gen %llu", PTR_DEV(k, i), + PTR_OFFSET(k, i), PTR_GEN(k, i)); + } + + p("]"); + + if (KEY_DIRTY(k)) + p(" dirty"); + if (KEY_CSUM(k)) + p(" cs%llu %llx", KEY_CSUM(k), k->ptr[1]); +#undef p +} + +static void bch_bkey_dump(struct btree_keys *keys, const struct bkey *k) +{ + struct btree *b = container_of(keys, struct btree, keys); + unsigned int j; + char buf[80]; + + bch_extent_to_text(buf, sizeof(buf), k); + pr_cont(" %s", buf); + + for (j = 0; j < KEY_PTRS(k); j++) { + size_t n = PTR_BUCKET_NR(b->c, k, j); + + pr_cont(" bucket %zu", n); + if (n >= b->c->cache->sb.first_bucket && n < b->c->cache->sb.nbuckets) + pr_cont(" prio %i", + PTR_BUCKET(b->c, k, j)->prio); + } + + pr_cont(" %s\n", bch_ptr_status(b->c, k)); +} + +/* Btree ptrs */ + +bool __bch_btree_ptr_invalid(struct cache_set *c, const struct bkey *k) +{ + char buf[80]; + + if (!KEY_PTRS(k) || !KEY_SIZE(k) || KEY_DIRTY(k)) + goto bad; + + if (__ptr_invalid(c, k)) + goto bad; + + return false; +bad: + bch_extent_to_text(buf, sizeof(buf), k); + cache_bug(c, "spotted btree ptr %s: %s", buf, bch_ptr_status(c, k)); + return true; +} + +static bool bch_btree_ptr_invalid(struct btree_keys *bk, const struct bkey *k) +{ + struct btree *b = container_of(bk, struct btree, keys); + + return __bch_btree_ptr_invalid(b->c, k); +} + +static bool btree_ptr_bad_expensive(struct btree *b, const struct bkey *k) +{ + unsigned int i; + char buf[80]; + struct bucket *g; + + if (mutex_trylock(&b->c->bucket_lock)) { + for (i = 0; i < KEY_PTRS(k); i++) + if (ptr_available(b->c, k, i)) { + g = PTR_BUCKET(b->c, k, i); + + if (KEY_DIRTY(k) || + g->prio != BTREE_PRIO || + (b->c->gc_mark_valid && + GC_MARK(g) != GC_MARK_METADATA)) + goto err; + } + + mutex_unlock(&b->c->bucket_lock); + } + + return false; +err: + mutex_unlock(&b->c->bucket_lock); + bch_extent_to_text(buf, sizeof(buf), k); + btree_bug(b, +"inconsistent btree pointer %s: bucket %zi pin %i prio %i gen %i last_gc %i mark %llu", + buf, PTR_BUCKET_NR(b->c, k, i), atomic_read(&g->pin), + g->prio, g->gen, g->last_gc, GC_MARK(g)); + return true; +} + +static bool bch_btree_ptr_bad(struct btree_keys *bk, const struct bkey *k) +{ + struct btree *b = container_of(bk, struct btree, keys); + unsigned int i; + + if (!bkey_cmp(k, &ZERO_KEY) || + !KEY_PTRS(k) || + bch_ptr_invalid(bk, k)) + return true; + + for (i = 0; i < KEY_PTRS(k); i++) + if (!ptr_available(b->c, k, i) || + ptr_stale(b->c, k, i)) + return true; + + if (expensive_debug_checks(b->c) && + btree_ptr_bad_expensive(b, k)) + return true; + + return false; +} + +static bool bch_btree_ptr_insert_fixup(struct btree_keys *bk, + struct bkey *insert, + struct btree_iter *iter, + struct bkey *replace_key) +{ + struct btree *b = container_of(bk, struct btree, keys); + + if (!KEY_OFFSET(insert)) + btree_current_write(b)->prio_blocked++; + + return false; +} + +const struct btree_keys_ops bch_btree_keys_ops = { + .sort_cmp = bch_key_sort_cmp, + .insert_fixup = bch_btree_ptr_insert_fixup, + .key_invalid = bch_btree_ptr_invalid, + .key_bad = bch_btree_ptr_bad, + .key_to_text = bch_extent_to_text, + .key_dump = bch_bkey_dump, +}; + +/* Extents */ + +/* + * Returns true if l > r - unless l == r, in which case returns true if l is + * older than r. + * + * Necessary for btree_sort_fixup() - if there are multiple keys that compare + * equal in different sets, we have to process them newest to oldest. + */ +static bool bch_extent_sort_cmp(struct btree_iter_set l, + struct btree_iter_set r) +{ + int64_t c = bkey_cmp(&START_KEY(l.k), &START_KEY(r.k)); + + return c ? c > 0 : l.k < r.k; +} + +static struct bkey *bch_extent_sort_fixup(struct btree_iter *iter, + struct bkey *tmp) +{ + while (iter->used > 1) { + struct btree_iter_set *top = iter->data, *i = top + 1; + + if (iter->used > 2 && + bch_extent_sort_cmp(i[0], i[1])) + i++; + + if (bkey_cmp(top->k, &START_KEY(i->k)) <= 0) + break; + + if (!KEY_SIZE(i->k)) { + sort_key_next(iter, i); + heap_sift(iter, i - top, bch_extent_sort_cmp); + continue; + } + + if (top->k > i->k) { + if (bkey_cmp(top->k, i->k) >= 0) + sort_key_next(iter, i); + else + bch_cut_front(top->k, i->k); + + heap_sift(iter, i - top, bch_extent_sort_cmp); + } else { + /* can't happen because of comparison func */ + BUG_ON(!bkey_cmp(&START_KEY(top->k), &START_KEY(i->k))); + + if (bkey_cmp(i->k, top->k) < 0) { + bkey_copy(tmp, top->k); + + bch_cut_back(&START_KEY(i->k), tmp); + bch_cut_front(i->k, top->k); + heap_sift(iter, 0, bch_extent_sort_cmp); + + return tmp; + } else { + bch_cut_back(&START_KEY(i->k), top->k); + } + } + } + + return NULL; +} + +static void bch_subtract_dirty(struct bkey *k, + struct cache_set *c, + uint64_t offset, + int sectors) +{ + if (KEY_DIRTY(k)) + bcache_dev_sectors_dirty_add(c, KEY_INODE(k), + offset, -sectors); +} + +static bool bch_extent_insert_fixup(struct btree_keys *b, + struct bkey *insert, + struct btree_iter *iter, + struct bkey *replace_key) +{ + struct cache_set *c = container_of(b, struct btree, keys)->c; + + uint64_t old_offset; + unsigned int old_size, sectors_found = 0; + + BUG_ON(!KEY_OFFSET(insert)); + BUG_ON(!KEY_SIZE(insert)); + + while (1) { + struct bkey *k = bch_btree_iter_next(iter); + + if (!k) + break; + + if (bkey_cmp(&START_KEY(k), insert) >= 0) { + if (KEY_SIZE(k)) + break; + else + continue; + } + + if (bkey_cmp(k, &START_KEY(insert)) <= 0) + continue; + + old_offset = KEY_START(k); + old_size = KEY_SIZE(k); + + /* + * We might overlap with 0 size extents; we can't skip these + * because if they're in the set we're inserting to we have to + * adjust them so they don't overlap with the key we're + * inserting. But we don't want to check them for replace + * operations. + */ + + if (replace_key && KEY_SIZE(k)) { + /* + * k might have been split since we inserted/found the + * key we're replacing + */ + unsigned int i; + uint64_t offset = KEY_START(k) - + KEY_START(replace_key); + + /* But it must be a subset of the replace key */ + if (KEY_START(k) < KEY_START(replace_key) || + KEY_OFFSET(k) > KEY_OFFSET(replace_key)) + goto check_failed; + + /* We didn't find a key that we were supposed to */ + if (KEY_START(k) > KEY_START(insert) + sectors_found) + goto check_failed; + + if (!bch_bkey_equal_header(k, replace_key)) + goto check_failed; + + /* skip past gen */ + offset <<= 8; + + BUG_ON(!KEY_PTRS(replace_key)); + + for (i = 0; i < KEY_PTRS(replace_key); i++) + if (k->ptr[i] != replace_key->ptr[i] + offset) + goto check_failed; + + sectors_found = KEY_OFFSET(k) - KEY_START(insert); + } + + if (bkey_cmp(insert, k) < 0 && + bkey_cmp(&START_KEY(insert), &START_KEY(k)) > 0) { + /* + * We overlapped in the middle of an existing key: that + * means we have to split the old key. But we have to do + * slightly different things depending on whether the + * old key has been written out yet. + */ + + struct bkey *top; + + bch_subtract_dirty(k, c, KEY_START(insert), + KEY_SIZE(insert)); + + if (bkey_written(b, k)) { + /* + * We insert a new key to cover the top of the + * old key, and the old key is modified in place + * to represent the bottom split. + * + * It's completely arbitrary whether the new key + * is the top or the bottom, but it has to match + * up with what btree_sort_fixup() does - it + * doesn't check for this kind of overlap, it + * depends on us inserting a new key for the top + * here. + */ + top = bch_bset_search(b, bset_tree_last(b), + insert); + bch_bset_insert(b, top, k); + } else { + BKEY_PADDED(key) temp; + bkey_copy(&temp.key, k); + bch_bset_insert(b, k, &temp.key); + top = bkey_next(k); + } + + bch_cut_front(insert, top); + bch_cut_back(&START_KEY(insert), k); + bch_bset_fix_invalidated_key(b, k); + goto out; + } + + if (bkey_cmp(insert, k) < 0) { + bch_cut_front(insert, k); + } else { + if (bkey_cmp(&START_KEY(insert), &START_KEY(k)) > 0) + old_offset = KEY_START(insert); + + if (bkey_written(b, k) && + bkey_cmp(&START_KEY(insert), &START_KEY(k)) <= 0) { + /* + * Completely overwrote, so we don't have to + * invalidate the binary search tree + */ + bch_cut_front(k, k); + } else { + __bch_cut_back(&START_KEY(insert), k); + bch_bset_fix_invalidated_key(b, k); + } + } + + bch_subtract_dirty(k, c, old_offset, old_size - KEY_SIZE(k)); + } + +check_failed: + if (replace_key) { + if (!sectors_found) { + return true; + } else if (sectors_found < KEY_SIZE(insert)) { + SET_KEY_OFFSET(insert, KEY_OFFSET(insert) - + (KEY_SIZE(insert) - sectors_found)); + SET_KEY_SIZE(insert, sectors_found); + } + } +out: + if (KEY_DIRTY(insert)) + bcache_dev_sectors_dirty_add(c, KEY_INODE(insert), + KEY_START(insert), + KEY_SIZE(insert)); + + return false; +} + +bool __bch_extent_invalid(struct cache_set *c, const struct bkey *k) +{ + char buf[80]; + + if (!KEY_SIZE(k)) + return true; + + if (KEY_SIZE(k) > KEY_OFFSET(k)) + goto bad; + + if (__ptr_invalid(c, k)) + goto bad; + + return false; +bad: + bch_extent_to_text(buf, sizeof(buf), k); + cache_bug(c, "spotted extent %s: %s", buf, bch_ptr_status(c, k)); + return true; +} + +static bool bch_extent_invalid(struct btree_keys *bk, const struct bkey *k) +{ + struct btree *b = container_of(bk, struct btree, keys); + + return __bch_extent_invalid(b->c, k); +} + +static bool bch_extent_bad_expensive(struct btree *b, const struct bkey *k, + unsigned int ptr) +{ + struct bucket *g = PTR_BUCKET(b->c, k, ptr); + char buf[80]; + + if (mutex_trylock(&b->c->bucket_lock)) { + if (b->c->gc_mark_valid && + (!GC_MARK(g) || + GC_MARK(g) == GC_MARK_METADATA || + (GC_MARK(g) != GC_MARK_DIRTY && KEY_DIRTY(k)))) + goto err; + + if (g->prio == BTREE_PRIO) + goto err; + + mutex_unlock(&b->c->bucket_lock); + } + + return false; +err: + mutex_unlock(&b->c->bucket_lock); + bch_extent_to_text(buf, sizeof(buf), k); + btree_bug(b, +"inconsistent extent pointer %s:\nbucket %zu pin %i prio %i gen %i last_gc %i mark %llu", + buf, PTR_BUCKET_NR(b->c, k, ptr), atomic_read(&g->pin), + g->prio, g->gen, g->last_gc, GC_MARK(g)); + return true; +} + +static bool bch_extent_bad(struct btree_keys *bk, const struct bkey *k) +{ + struct btree *b = container_of(bk, struct btree, keys); + unsigned int i, stale; + char buf[80]; + + if (!KEY_PTRS(k) || + bch_extent_invalid(bk, k)) + return true; + + for (i = 0; i < KEY_PTRS(k); i++) + if (!ptr_available(b->c, k, i)) + return true; + + for (i = 0; i < KEY_PTRS(k); i++) { + stale = ptr_stale(b->c, k, i); + + if (stale && KEY_DIRTY(k)) { + bch_extent_to_text(buf, sizeof(buf), k); + pr_info("stale dirty pointer, stale %u, key: %s\n", + stale, buf); + } + + btree_bug_on(stale > BUCKET_GC_GEN_MAX, b, + "key too stale: %i, need_gc %u", + stale, b->c->need_gc); + + if (stale) + return true; + + if (expensive_debug_checks(b->c) && + bch_extent_bad_expensive(b, k, i)) + return true; + } + + return false; +} + +static uint64_t merge_chksums(struct bkey *l, struct bkey *r) +{ + return (l->ptr[KEY_PTRS(l)] + r->ptr[KEY_PTRS(r)]) & + ~((uint64_t)1 << 63); +} + +static bool bch_extent_merge(struct btree_keys *bk, + struct bkey *l, + struct bkey *r) +{ + struct btree *b = container_of(bk, struct btree, keys); + unsigned int i; + + if (key_merging_disabled(b->c)) + return false; + + for (i = 0; i < KEY_PTRS(l); i++) + if (l->ptr[i] + MAKE_PTR(0, KEY_SIZE(l), 0) != r->ptr[i] || + PTR_BUCKET_NR(b->c, l, i) != PTR_BUCKET_NR(b->c, r, i)) + return false; + + /* Keys with no pointers aren't restricted to one bucket and could + * overflow KEY_SIZE + */ + if (KEY_SIZE(l) + KEY_SIZE(r) > USHRT_MAX) { + SET_KEY_OFFSET(l, KEY_OFFSET(l) + USHRT_MAX - KEY_SIZE(l)); + SET_KEY_SIZE(l, USHRT_MAX); + + bch_cut_front(l, r); + return false; + } + + if (KEY_CSUM(l)) { + if (KEY_CSUM(r)) + l->ptr[KEY_PTRS(l)] = merge_chksums(l, r); + else + SET_KEY_CSUM(l, 0); + } + + SET_KEY_OFFSET(l, KEY_OFFSET(l) + KEY_SIZE(r)); + SET_KEY_SIZE(l, KEY_SIZE(l) + KEY_SIZE(r)); + + return true; +} + +const struct btree_keys_ops bch_extent_keys_ops = { + .sort_cmp = bch_extent_sort_cmp, + .sort_fixup = bch_extent_sort_fixup, + .insert_fixup = bch_extent_insert_fixup, + .key_invalid = bch_extent_invalid, + .key_bad = bch_extent_bad, + .key_merge = bch_extent_merge, + .key_to_text = bch_extent_to_text, + .key_dump = bch_bkey_dump, + .is_extents = true, +}; diff --git a/drivers/md/bcache/extents.h b/drivers/md/bcache/extents.h new file mode 100644 index 000000000..4d667e05b --- /dev/null +++ b/drivers/md/bcache/extents.h @@ -0,0 +1,15 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _BCACHE_EXTENTS_H +#define _BCACHE_EXTENTS_H + +extern const struct btree_keys_ops bch_btree_keys_ops; +extern const struct btree_keys_ops bch_extent_keys_ops; + +struct bkey; +struct cache_set; + +void bch_extent_to_text(char *buf, size_t size, const struct bkey *k); +bool __bch_btree_ptr_invalid(struct cache_set *c, const struct bkey *k); +bool __bch_extent_invalid(struct cache_set *c, const struct bkey *k); + +#endif /* _BCACHE_EXTENTS_H */ diff --git a/drivers/md/bcache/features.c b/drivers/md/bcache/features.c new file mode 100644 index 000000000..d636b7b2d --- /dev/null +++ b/drivers/md/bcache/features.c @@ -0,0 +1,75 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Feature set bits and string conversion. + * Inspired by ext4's features compat/incompat/ro_compat related code. + * + * Copyright 2020 Coly Li <colyli@suse.de> + * + */ +#include <linux/bcache.h> +#include "bcache.h" +#include "features.h" + +struct feature { + int compat; + unsigned int mask; + const char *string; +}; + +static struct feature feature_list[] = { + {BCH_FEATURE_INCOMPAT, BCH_FEATURE_INCOMPAT_LOG_LARGE_BUCKET_SIZE, + "large_bucket"}, + {0, 0, 0 }, +}; + +#define compose_feature_string(type) \ +({ \ + struct feature *f; \ + bool first = true; \ + \ + for (f = &feature_list[0]; f->compat != 0; f++) { \ + if (f->compat != BCH_FEATURE_ ## type) \ + continue; \ + if (BCH_HAS_ ## type ## _FEATURE(&c->cache->sb, f->mask)) { \ + if (first) { \ + out += snprintf(out, buf + size - out, \ + "["); \ + } else { \ + out += snprintf(out, buf + size - out, \ + " ["); \ + } \ + } else if (!first) { \ + out += snprintf(out, buf + size - out, " "); \ + } \ + \ + out += snprintf(out, buf + size - out, "%s", f->string);\ + \ + if (BCH_HAS_ ## type ## _FEATURE(&c->cache->sb, f->mask)) \ + out += snprintf(out, buf + size - out, "]"); \ + \ + first = false; \ + } \ + if (!first) \ + out += snprintf(out, buf + size - out, "\n"); \ +}) + +int bch_print_cache_set_feature_compat(struct cache_set *c, char *buf, int size) +{ + char *out = buf; + compose_feature_string(COMPAT); + return out - buf; +} + +int bch_print_cache_set_feature_ro_compat(struct cache_set *c, char *buf, int size) +{ + char *out = buf; + compose_feature_string(RO_COMPAT); + return out - buf; +} + +int bch_print_cache_set_feature_incompat(struct cache_set *c, char *buf, int size) +{ + char *out = buf; + compose_feature_string(INCOMPAT); + return out - buf; +} diff --git a/drivers/md/bcache/features.h b/drivers/md/bcache/features.h new file mode 100644 index 000000000..d1c8fd397 --- /dev/null +++ b/drivers/md/bcache/features.h @@ -0,0 +1,112 @@ +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */ +#ifndef _BCACHE_FEATURES_H +#define _BCACHE_FEATURES_H + +#include <linux/bcache.h> +#include <linux/kernel.h> +#include <linux/types.h> + +#define BCH_FEATURE_COMPAT 0 +#define BCH_FEATURE_RO_COMPAT 1 +#define BCH_FEATURE_INCOMPAT 2 +#define BCH_FEATURE_TYPE_MASK 0x03 + +/* Feature set definition */ +/* Incompat feature set */ +/* 32bit bucket size, obsoleted */ +#define BCH_FEATURE_INCOMPAT_OBSO_LARGE_BUCKET 0x0001 +/* real bucket size is (1 << bucket_size) */ +#define BCH_FEATURE_INCOMPAT_LOG_LARGE_BUCKET_SIZE 0x0002 + +#define BCH_FEATURE_COMPAT_SUPP 0 +#define BCH_FEATURE_RO_COMPAT_SUPP 0 +#define BCH_FEATURE_INCOMPAT_SUPP (BCH_FEATURE_INCOMPAT_OBSO_LARGE_BUCKET| \ + BCH_FEATURE_INCOMPAT_LOG_LARGE_BUCKET_SIZE) + +#define BCH_HAS_COMPAT_FEATURE(sb, mask) \ + ((sb)->feature_compat & (mask)) +#define BCH_HAS_RO_COMPAT_FEATURE(sb, mask) \ + ((sb)->feature_ro_compat & (mask)) +#define BCH_HAS_INCOMPAT_FEATURE(sb, mask) \ + ((sb)->feature_incompat & (mask)) + +#define BCH_FEATURE_COMPAT_FUNCS(name, flagname) \ +static inline int bch_has_feature_##name(struct cache_sb *sb) \ +{ \ + if (sb->version < BCACHE_SB_VERSION_CDEV_WITH_FEATURES) \ + return 0; \ + return (((sb)->feature_compat & \ + BCH##_FEATURE_COMPAT_##flagname) != 0); \ +} \ +static inline void bch_set_feature_##name(struct cache_sb *sb) \ +{ \ + (sb)->feature_compat |= \ + BCH##_FEATURE_COMPAT_##flagname; \ +} \ +static inline void bch_clear_feature_##name(struct cache_sb *sb) \ +{ \ + (sb)->feature_compat &= \ + ~BCH##_FEATURE_COMPAT_##flagname; \ +} + +#define BCH_FEATURE_RO_COMPAT_FUNCS(name, flagname) \ +static inline int bch_has_feature_##name(struct cache_sb *sb) \ +{ \ + if (sb->version < BCACHE_SB_VERSION_CDEV_WITH_FEATURES) \ + return 0; \ + return (((sb)->feature_ro_compat & \ + BCH##_FEATURE_RO_COMPAT_##flagname) != 0); \ +} \ +static inline void bch_set_feature_##name(struct cache_sb *sb) \ +{ \ + (sb)->feature_ro_compat |= \ + BCH##_FEATURE_RO_COMPAT_##flagname; \ +} \ +static inline void bch_clear_feature_##name(struct cache_sb *sb) \ +{ \ + (sb)->feature_ro_compat &= \ + ~BCH##_FEATURE_RO_COMPAT_##flagname; \ +} + +#define BCH_FEATURE_INCOMPAT_FUNCS(name, flagname) \ +static inline int bch_has_feature_##name(struct cache_sb *sb) \ +{ \ + if (sb->version < BCACHE_SB_VERSION_CDEV_WITH_FEATURES) \ + return 0; \ + return (((sb)->feature_incompat & \ + BCH##_FEATURE_INCOMPAT_##flagname) != 0); \ +} \ +static inline void bch_set_feature_##name(struct cache_sb *sb) \ +{ \ + (sb)->feature_incompat |= \ + BCH##_FEATURE_INCOMPAT_##flagname; \ +} \ +static inline void bch_clear_feature_##name(struct cache_sb *sb) \ +{ \ + (sb)->feature_incompat &= \ + ~BCH##_FEATURE_INCOMPAT_##flagname; \ +} + +BCH_FEATURE_INCOMPAT_FUNCS(obso_large_bucket, OBSO_LARGE_BUCKET); +BCH_FEATURE_INCOMPAT_FUNCS(large_bucket, LOG_LARGE_BUCKET_SIZE); + +static inline bool bch_has_unknown_compat_features(struct cache_sb *sb) +{ + return ((sb->feature_compat & ~BCH_FEATURE_COMPAT_SUPP) != 0); +} + +static inline bool bch_has_unknown_ro_compat_features(struct cache_sb *sb) +{ + return ((sb->feature_ro_compat & ~BCH_FEATURE_RO_COMPAT_SUPP) != 0); +} + +static inline bool bch_has_unknown_incompat_features(struct cache_sb *sb) +{ + return ((sb->feature_incompat & ~BCH_FEATURE_INCOMPAT_SUPP) != 0); +} + +int bch_print_cache_set_feature_compat(struct cache_set *c, char *buf, int size); +int bch_print_cache_set_feature_ro_compat(struct cache_set *c, char *buf, int size); +int bch_print_cache_set_feature_incompat(struct cache_set *c, char *buf, int size); + +#endif diff --git a/drivers/md/bcache/io.c b/drivers/md/bcache/io.c new file mode 100644 index 000000000..dad71a6b7 --- /dev/null +++ b/drivers/md/bcache/io.c @@ -0,0 +1,174 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Some low level IO code, and hacks for various block layer limitations + * + * Copyright 2010, 2011 Kent Overstreet <kent.overstreet@gmail.com> + * Copyright 2012 Google, Inc. + */ + +#include "bcache.h" +#include "bset.h" +#include "debug.h" + +#include <linux/blkdev.h> + +/* Bios with headers */ + +void bch_bbio_free(struct bio *bio, struct cache_set *c) +{ + struct bbio *b = container_of(bio, struct bbio, bio); + + mempool_free(b, &c->bio_meta); +} + +struct bio *bch_bbio_alloc(struct cache_set *c) +{ + struct bbio *b = mempool_alloc(&c->bio_meta, GFP_NOIO); + struct bio *bio = &b->bio; + + bio_init(bio, bio->bi_inline_vecs, meta_bucket_pages(&c->cache->sb)); + + return bio; +} + +void __bch_submit_bbio(struct bio *bio, struct cache_set *c) +{ + struct bbio *b = container_of(bio, struct bbio, bio); + + bio->bi_iter.bi_sector = PTR_OFFSET(&b->key, 0); + bio_set_dev(bio, PTR_CACHE(c, &b->key, 0)->bdev); + + b->submit_time_us = local_clock_us(); + closure_bio_submit(c, bio, bio->bi_private); +} + +void bch_submit_bbio(struct bio *bio, struct cache_set *c, + struct bkey *k, unsigned int ptr) +{ + struct bbio *b = container_of(bio, struct bbio, bio); + + bch_bkey_copy_single_ptr(&b->key, k, ptr); + __bch_submit_bbio(bio, c); +} + +/* IO errors */ +void bch_count_backing_io_errors(struct cached_dev *dc, struct bio *bio) +{ + unsigned int errors; + + WARN_ONCE(!dc, "NULL pointer of struct cached_dev"); + + /* + * Read-ahead requests on a degrading and recovering md raid + * (e.g. raid6) device might be failured immediately by md + * raid code, which is not a real hardware media failure. So + * we shouldn't count failed REQ_RAHEAD bio to dc->io_errors. + */ + if (bio->bi_opf & REQ_RAHEAD) { + pr_warn_ratelimited("%s: Read-ahead I/O failed on backing device, ignore\n", + dc->backing_dev_name); + return; + } + + errors = atomic_add_return(1, &dc->io_errors); + if (errors < dc->error_limit) + pr_err("%s: IO error on backing device, unrecoverable\n", + dc->backing_dev_name); + else + bch_cached_dev_error(dc); +} + +void bch_count_io_errors(struct cache *ca, + blk_status_t error, + int is_read, + const char *m) +{ + /* + * The halflife of an error is: + * log2(1/2)/log2(127/128) * refresh ~= 88 * refresh + */ + + if (ca->set->error_decay) { + unsigned int count = atomic_inc_return(&ca->io_count); + + while (count > ca->set->error_decay) { + unsigned int errors; + unsigned int old = count; + unsigned int new = count - ca->set->error_decay; + + /* + * First we subtract refresh from count; each time we + * successfully do so, we rescale the errors once: + */ + + count = atomic_cmpxchg(&ca->io_count, old, new); + + if (count == old) { + count = new; + + errors = atomic_read(&ca->io_errors); + do { + old = errors; + new = ((uint64_t) errors * 127) / 128; + errors = atomic_cmpxchg(&ca->io_errors, + old, new); + } while (old != errors); + } + } + } + + if (error) { + unsigned int errors = atomic_add_return(1 << IO_ERROR_SHIFT, + &ca->io_errors); + errors >>= IO_ERROR_SHIFT; + + if (errors < ca->set->error_limit) + pr_err("%s: IO error on %s%s\n", + ca->cache_dev_name, m, + is_read ? ", recovering." : "."); + else + bch_cache_set_error(ca->set, + "%s: too many IO errors %s\n", + ca->cache_dev_name, m); + } +} + +void bch_bbio_count_io_errors(struct cache_set *c, struct bio *bio, + blk_status_t error, const char *m) +{ + struct bbio *b = container_of(bio, struct bbio, bio); + struct cache *ca = PTR_CACHE(c, &b->key, 0); + int is_read = (bio_data_dir(bio) == READ ? 1 : 0); + + unsigned int threshold = op_is_write(bio_op(bio)) + ? c->congested_write_threshold_us + : c->congested_read_threshold_us; + + if (threshold) { + unsigned int t = local_clock_us(); + int us = t - b->submit_time_us; + int congested = atomic_read(&c->congested); + + if (us > (int) threshold) { + int ms = us / 1024; + + c->congested_last_us = t; + + ms = min(ms, CONGESTED_MAX + congested); + atomic_sub(ms, &c->congested); + } else if (congested < 0) + atomic_inc(&c->congested); + } + + bch_count_io_errors(ca, error, is_read, m); +} + +void bch_bbio_endio(struct cache_set *c, struct bio *bio, + blk_status_t error, const char *m) +{ + struct closure *cl = bio->bi_private; + + bch_bbio_count_io_errors(c, bio, error, m); + bio_put(bio); + closure_put(cl); +} diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c new file mode 100644 index 000000000..aea4833f6 --- /dev/null +++ b/drivers/md/bcache/journal.c @@ -0,0 +1,1005 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * bcache journalling code, for btree insertions + * + * Copyright 2012 Google, Inc. + */ + +#include "bcache.h" +#include "btree.h" +#include "debug.h" +#include "extents.h" + +#include <trace/events/bcache.h> + +/* + * Journal replay/recovery: + * + * This code is all driven from run_cache_set(); we first read the journal + * entries, do some other stuff, then we mark all the keys in the journal + * entries (same as garbage collection would), then we replay them - reinserting + * them into the cache in precisely the same order as they appear in the + * journal. + * + * We only journal keys that go in leaf nodes, which simplifies things quite a + * bit. + */ + +static void journal_read_endio(struct bio *bio) +{ + struct closure *cl = bio->bi_private; + + closure_put(cl); +} + +static int journal_read_bucket(struct cache *ca, struct list_head *list, + unsigned int bucket_index) +{ + struct journal_device *ja = &ca->journal; + struct bio *bio = &ja->bio; + + struct journal_replay *i; + struct jset *j, *data = ca->set->journal.w[0].data; + struct closure cl; + unsigned int len, left, offset = 0; + int ret = 0; + sector_t bucket = bucket_to_sector(ca->set, ca->sb.d[bucket_index]); + + closure_init_stack(&cl); + + pr_debug("reading %u\n", bucket_index); + + while (offset < ca->sb.bucket_size) { +reread: left = ca->sb.bucket_size - offset; + len = min_t(unsigned int, left, PAGE_SECTORS << JSET_BITS); + + bio_reset(bio); + bio->bi_iter.bi_sector = bucket + offset; + bio_set_dev(bio, ca->bdev); + bio->bi_iter.bi_size = len << 9; + + bio->bi_end_io = journal_read_endio; + bio->bi_private = &cl; + bio_set_op_attrs(bio, REQ_OP_READ, 0); + bch_bio_map(bio, data); + + closure_bio_submit(ca->set, bio, &cl); + closure_sync(&cl); + + /* This function could be simpler now since we no longer write + * journal entries that overlap bucket boundaries; this means + * the start of a bucket will always have a valid journal entry + * if it has any journal entries at all. + */ + + j = data; + while (len) { + struct list_head *where; + size_t blocks, bytes = set_bytes(j); + + if (j->magic != jset_magic(&ca->sb)) { + pr_debug("%u: bad magic\n", bucket_index); + return ret; + } + + if (bytes > left << 9 || + bytes > PAGE_SIZE << JSET_BITS) { + pr_info("%u: too big, %zu bytes, offset %u\n", + bucket_index, bytes, offset); + return ret; + } + + if (bytes > len << 9) + goto reread; + + if (j->csum != csum_set(j)) { + pr_info("%u: bad csum, %zu bytes, offset %u\n", + bucket_index, bytes, offset); + return ret; + } + + blocks = set_blocks(j, block_bytes(ca)); + + /* + * Nodes in 'list' are in linear increasing order of + * i->j.seq, the node on head has the smallest (oldest) + * journal seq, the node on tail has the biggest + * (latest) journal seq. + */ + + /* + * Check from the oldest jset for last_seq. If + * i->j.seq < j->last_seq, it means the oldest jset + * in list is expired and useless, remove it from + * this list. Otherwise, j is a condidate jset for + * further following checks. + */ + while (!list_empty(list)) { + i = list_first_entry(list, + struct journal_replay, list); + if (i->j.seq >= j->last_seq) + break; + list_del(&i->list); + kfree(i); + } + + /* iterate list in reverse order (from latest jset) */ + list_for_each_entry_reverse(i, list, list) { + if (j->seq == i->j.seq) + goto next_set; + + /* + * if j->seq is less than any i->j.last_seq + * in list, j is an expired and useless jset. + */ + if (j->seq < i->j.last_seq) + goto next_set; + + /* + * 'where' points to first jset in list which + * is elder then j. + */ + if (j->seq > i->j.seq) { + where = &i->list; + goto add; + } + } + + where = list; +add: + i = kmalloc(offsetof(struct journal_replay, j) + + bytes, GFP_KERNEL); + if (!i) + return -ENOMEM; + memcpy(&i->j, j, bytes); + /* Add to the location after 'where' points to */ + list_add(&i->list, where); + ret = 1; + + if (j->seq > ja->seq[bucket_index]) + ja->seq[bucket_index] = j->seq; +next_set: + offset += blocks * ca->sb.block_size; + len -= blocks * ca->sb.block_size; + j = ((void *) j) + blocks * block_bytes(ca); + } + } + + return ret; +} + +int bch_journal_read(struct cache_set *c, struct list_head *list) +{ +#define read_bucket(b) \ + ({ \ + ret = journal_read_bucket(ca, list, b); \ + __set_bit(b, bitmap); \ + if (ret < 0) \ + return ret; \ + ret; \ + }) + + struct cache *ca = c->cache; + int ret = 0; + struct journal_device *ja = &ca->journal; + DECLARE_BITMAP(bitmap, SB_JOURNAL_BUCKETS); + unsigned int i, l, r, m; + uint64_t seq; + + bitmap_zero(bitmap, SB_JOURNAL_BUCKETS); + pr_debug("%u journal buckets\n", ca->sb.njournal_buckets); + + /* + * Read journal buckets ordered by golden ratio hash to quickly + * find a sequence of buckets with valid journal entries + */ + for (i = 0; i < ca->sb.njournal_buckets; i++) { + /* + * We must try the index l with ZERO first for + * correctness due to the scenario that the journal + * bucket is circular buffer which might have wrapped + */ + l = (i * 2654435769U) % ca->sb.njournal_buckets; + + if (test_bit(l, bitmap)) + break; + + if (read_bucket(l)) + goto bsearch; + } + + /* + * If that fails, check all the buckets we haven't checked + * already + */ + pr_debug("falling back to linear search\n"); + + for_each_clear_bit(l, bitmap, ca->sb.njournal_buckets) + if (read_bucket(l)) + goto bsearch; + + /* no journal entries on this device? */ + if (l == ca->sb.njournal_buckets) + goto out; +bsearch: + BUG_ON(list_empty(list)); + + /* Binary search */ + m = l; + r = find_next_bit(bitmap, ca->sb.njournal_buckets, l + 1); + pr_debug("starting binary search, l %u r %u\n", l, r); + + while (l + 1 < r) { + seq = list_entry(list->prev, struct journal_replay, + list)->j.seq; + + m = (l + r) >> 1; + read_bucket(m); + + if (seq != list_entry(list->prev, struct journal_replay, + list)->j.seq) + l = m; + else + r = m; + } + + /* + * Read buckets in reverse order until we stop finding more + * journal entries + */ + pr_debug("finishing up: m %u njournal_buckets %u\n", + m, ca->sb.njournal_buckets); + l = m; + + while (1) { + if (!l--) + l = ca->sb.njournal_buckets - 1; + + if (l == m) + break; + + if (test_bit(l, bitmap)) + continue; + + if (!read_bucket(l)) + break; + } + + seq = 0; + + for (i = 0; i < ca->sb.njournal_buckets; i++) + if (ja->seq[i] > seq) { + seq = ja->seq[i]; + /* + * When journal_reclaim() goes to allocate for + * the first time, it'll use the bucket after + * ja->cur_idx + */ + ja->cur_idx = i; + ja->last_idx = ja->discard_idx = (i + 1) % + ca->sb.njournal_buckets; + + } + +out: + if (!list_empty(list)) + c->journal.seq = list_entry(list->prev, + struct journal_replay, + list)->j.seq; + + return 0; +#undef read_bucket +} + +void bch_journal_mark(struct cache_set *c, struct list_head *list) +{ + atomic_t p = { 0 }; + struct bkey *k; + struct journal_replay *i; + struct journal *j = &c->journal; + uint64_t last = j->seq; + + /* + * journal.pin should never fill up - we never write a journal + * entry when it would fill up. But if for some reason it does, we + * iterate over the list in reverse order so that we can just skip that + * refcount instead of bugging. + */ + + list_for_each_entry_reverse(i, list, list) { + BUG_ON(last < i->j.seq); + i->pin = NULL; + + while (last-- != i->j.seq) + if (fifo_free(&j->pin) > 1) { + fifo_push_front(&j->pin, p); + atomic_set(&fifo_front(&j->pin), 0); + } + + if (fifo_free(&j->pin) > 1) { + fifo_push_front(&j->pin, p); + i->pin = &fifo_front(&j->pin); + atomic_set(i->pin, 1); + } + + for (k = i->j.start; + k < bset_bkey_last(&i->j); + k = bkey_next(k)) + if (!__bch_extent_invalid(c, k)) { + unsigned int j; + + for (j = 0; j < KEY_PTRS(k); j++) + if (ptr_available(c, k, j)) + atomic_inc(&PTR_BUCKET(c, k, j)->pin); + + bch_initial_mark_key(c, 0, k); + } + } +} + +static bool is_discard_enabled(struct cache_set *s) +{ + struct cache *ca = s->cache; + + if (ca->discard) + return true; + + return false; +} + +int bch_journal_replay(struct cache_set *s, struct list_head *list) +{ + int ret = 0, keys = 0, entries = 0; + struct bkey *k; + struct journal_replay *i = + list_entry(list->prev, struct journal_replay, list); + + uint64_t start = i->j.last_seq, end = i->j.seq, n = start; + struct keylist keylist; + + list_for_each_entry(i, list, list) { + BUG_ON(i->pin && atomic_read(i->pin) != 1); + + if (n != i->j.seq) { + if (n == start && is_discard_enabled(s)) + pr_info("journal entries %llu-%llu may be discarded! (replaying %llu-%llu)\n", + n, i->j.seq - 1, start, end); + else { + pr_err("journal entries %llu-%llu missing! (replaying %llu-%llu)\n", + n, i->j.seq - 1, start, end); + ret = -EIO; + goto err; + } + } + + for (k = i->j.start; + k < bset_bkey_last(&i->j); + k = bkey_next(k)) { + trace_bcache_journal_replay_key(k); + + bch_keylist_init_single(&keylist, k); + + ret = bch_btree_insert(s, &keylist, i->pin, NULL); + if (ret) + goto err; + + BUG_ON(!bch_keylist_empty(&keylist)); + keys++; + + cond_resched(); + } + + if (i->pin) + atomic_dec(i->pin); + n = i->j.seq + 1; + entries++; + } + + pr_info("journal replay done, %i keys in %i entries, seq %llu\n", + keys, entries, end); +err: + while (!list_empty(list)) { + i = list_first_entry(list, struct journal_replay, list); + list_del(&i->list); + kfree(i); + } + + return ret; +} + +void bch_journal_space_reserve(struct journal *j) +{ + j->do_reserve = true; +} + +/* Journalling */ + +static void btree_flush_write(struct cache_set *c) +{ + struct btree *b, *t, *btree_nodes[BTREE_FLUSH_NR]; + unsigned int i, nr; + int ref_nr; + atomic_t *fifo_front_p, *now_fifo_front_p; + size_t mask; + + if (c->journal.btree_flushing) + return; + + spin_lock(&c->journal.flush_write_lock); + if (c->journal.btree_flushing) { + spin_unlock(&c->journal.flush_write_lock); + return; + } + c->journal.btree_flushing = true; + spin_unlock(&c->journal.flush_write_lock); + + /* get the oldest journal entry and check its refcount */ + spin_lock(&c->journal.lock); + fifo_front_p = &fifo_front(&c->journal.pin); + ref_nr = atomic_read(fifo_front_p); + if (ref_nr <= 0) { + /* + * do nothing if no btree node references + * the oldest journal entry + */ + spin_unlock(&c->journal.lock); + goto out; + } + spin_unlock(&c->journal.lock); + + mask = c->journal.pin.mask; + nr = 0; + atomic_long_inc(&c->flush_write); + memset(btree_nodes, 0, sizeof(btree_nodes)); + + mutex_lock(&c->bucket_lock); + list_for_each_entry_safe_reverse(b, t, &c->btree_cache, list) { + /* + * It is safe to get now_fifo_front_p without holding + * c->journal.lock here, because we don't need to know + * the exactly accurate value, just check whether the + * front pointer of c->journal.pin is changed. + */ + now_fifo_front_p = &fifo_front(&c->journal.pin); + /* + * If the oldest journal entry is reclaimed and front + * pointer of c->journal.pin changes, it is unnecessary + * to scan c->btree_cache anymore, just quit the loop and + * flush out what we have already. + */ + if (now_fifo_front_p != fifo_front_p) + break; + /* + * quit this loop if all matching btree nodes are + * scanned and record in btree_nodes[] already. + */ + ref_nr = atomic_read(fifo_front_p); + if (nr >= ref_nr) + break; + + if (btree_node_journal_flush(b)) + pr_err("BUG: flush_write bit should not be set here!\n"); + + mutex_lock(&b->write_lock); + + if (!btree_node_dirty(b)) { + mutex_unlock(&b->write_lock); + continue; + } + + if (!btree_current_write(b)->journal) { + mutex_unlock(&b->write_lock); + continue; + } + + /* + * Only select the btree node which exactly references + * the oldest journal entry. + * + * If the journal entry pointed by fifo_front_p is + * reclaimed in parallel, don't worry: + * - the list_for_each_xxx loop will quit when checking + * next now_fifo_front_p. + * - If there are matched nodes recorded in btree_nodes[], + * they are clean now (this is why and how the oldest + * journal entry can be reclaimed). These selected nodes + * will be ignored and skipped in the folowing for-loop. + */ + if (((btree_current_write(b)->journal - fifo_front_p) & + mask) != 0) { + mutex_unlock(&b->write_lock); + continue; + } + + set_btree_node_journal_flush(b); + + mutex_unlock(&b->write_lock); + + btree_nodes[nr++] = b; + /* + * To avoid holding c->bucket_lock too long time, + * only scan for BTREE_FLUSH_NR matched btree nodes + * at most. If there are more btree nodes reference + * the oldest journal entry, try to flush them next + * time when btree_flush_write() is called. + */ + if (nr == BTREE_FLUSH_NR) + break; + } + mutex_unlock(&c->bucket_lock); + + for (i = 0; i < nr; i++) { + b = btree_nodes[i]; + if (!b) { + pr_err("BUG: btree_nodes[%d] is NULL\n", i); + continue; + } + + /* safe to check without holding b->write_lock */ + if (!btree_node_journal_flush(b)) { + pr_err("BUG: bnode %p: journal_flush bit cleaned\n", b); + continue; + } + + mutex_lock(&b->write_lock); + if (!btree_current_write(b)->journal) { + clear_bit(BTREE_NODE_journal_flush, &b->flags); + mutex_unlock(&b->write_lock); + pr_debug("bnode %p: written by others\n", b); + continue; + } + + if (!btree_node_dirty(b)) { + clear_bit(BTREE_NODE_journal_flush, &b->flags); + mutex_unlock(&b->write_lock); + pr_debug("bnode %p: dirty bit cleaned by others\n", b); + continue; + } + + __bch_btree_node_write(b, NULL); + clear_bit(BTREE_NODE_journal_flush, &b->flags); + mutex_unlock(&b->write_lock); + } + +out: + spin_lock(&c->journal.flush_write_lock); + c->journal.btree_flushing = false; + spin_unlock(&c->journal.flush_write_lock); +} + +#define last_seq(j) ((j)->seq - fifo_used(&(j)->pin) + 1) + +static void journal_discard_endio(struct bio *bio) +{ + struct journal_device *ja = + container_of(bio, struct journal_device, discard_bio); + struct cache *ca = container_of(ja, struct cache, journal); + + atomic_set(&ja->discard_in_flight, DISCARD_DONE); + + closure_wake_up(&ca->set->journal.wait); + closure_put(&ca->set->cl); +} + +static void journal_discard_work(struct work_struct *work) +{ + struct journal_device *ja = + container_of(work, struct journal_device, discard_work); + + submit_bio(&ja->discard_bio); +} + +static void do_journal_discard(struct cache *ca) +{ + struct journal_device *ja = &ca->journal; + struct bio *bio = &ja->discard_bio; + + if (!ca->discard) { + ja->discard_idx = ja->last_idx; + return; + } + + switch (atomic_read(&ja->discard_in_flight)) { + case DISCARD_IN_FLIGHT: + return; + + case DISCARD_DONE: + ja->discard_idx = (ja->discard_idx + 1) % + ca->sb.njournal_buckets; + + atomic_set(&ja->discard_in_flight, DISCARD_READY); + fallthrough; + + case DISCARD_READY: + if (ja->discard_idx == ja->last_idx) + return; + + atomic_set(&ja->discard_in_flight, DISCARD_IN_FLIGHT); + + bio_init(bio, bio->bi_inline_vecs, 1); + bio_set_op_attrs(bio, REQ_OP_DISCARD, 0); + bio->bi_iter.bi_sector = bucket_to_sector(ca->set, + ca->sb.d[ja->discard_idx]); + bio_set_dev(bio, ca->bdev); + bio->bi_iter.bi_size = bucket_bytes(ca); + bio->bi_end_io = journal_discard_endio; + + closure_get(&ca->set->cl); + INIT_WORK(&ja->discard_work, journal_discard_work); + queue_work(bch_journal_wq, &ja->discard_work); + } +} + +static unsigned int free_journal_buckets(struct cache_set *c) +{ + struct journal *j = &c->journal; + struct cache *ca = c->cache; + struct journal_device *ja = &c->cache->journal; + unsigned int n; + + /* In case njournal_buckets is not power of 2 */ + if (ja->cur_idx >= ja->discard_idx) + n = ca->sb.njournal_buckets + ja->discard_idx - ja->cur_idx; + else + n = ja->discard_idx - ja->cur_idx; + + if (n > (1 + j->do_reserve)) + return n - (1 + j->do_reserve); + + return 0; +} + +static void journal_reclaim(struct cache_set *c) +{ + struct bkey *k = &c->journal.key; + struct cache *ca = c->cache; + uint64_t last_seq; + struct journal_device *ja = &ca->journal; + atomic_t p __maybe_unused; + + atomic_long_inc(&c->reclaim); + + while (!atomic_read(&fifo_front(&c->journal.pin))) + fifo_pop(&c->journal.pin, p); + + last_seq = last_seq(&c->journal); + + /* Update last_idx */ + + while (ja->last_idx != ja->cur_idx && + ja->seq[ja->last_idx] < last_seq) + ja->last_idx = (ja->last_idx + 1) % + ca->sb.njournal_buckets; + + do_journal_discard(ca); + + if (c->journal.blocks_free) + goto out; + + if (!free_journal_buckets(c)) + goto out; + + ja->cur_idx = (ja->cur_idx + 1) % ca->sb.njournal_buckets; + k->ptr[0] = MAKE_PTR(0, + bucket_to_sector(c, ca->sb.d[ja->cur_idx]), + ca->sb.nr_this_dev); + atomic_long_inc(&c->reclaimed_journal_buckets); + + bkey_init(k); + SET_KEY_PTRS(k, 1); + c->journal.blocks_free = ca->sb.bucket_size >> c->block_bits; + +out: + if (!journal_full(&c->journal)) + __closure_wake_up(&c->journal.wait); +} + +void bch_journal_next(struct journal *j) +{ + atomic_t p = { 1 }; + + j->cur = (j->cur == j->w) + ? &j->w[1] + : &j->w[0]; + + /* + * The fifo_push() needs to happen at the same time as j->seq is + * incremented for last_seq() to be calculated correctly + */ + BUG_ON(!fifo_push(&j->pin, p)); + atomic_set(&fifo_back(&j->pin), 1); + + j->cur->data->seq = ++j->seq; + j->cur->dirty = false; + j->cur->need_write = false; + j->cur->data->keys = 0; + + if (fifo_full(&j->pin)) + pr_debug("journal_pin full (%zu)\n", fifo_used(&j->pin)); +} + +static void journal_write_endio(struct bio *bio) +{ + struct journal_write *w = bio->bi_private; + + cache_set_err_on(bio->bi_status, w->c, "journal io error"); + closure_put(&w->c->journal.io); +} + +static void journal_write(struct closure *cl); + +static void journal_write_done(struct closure *cl) +{ + struct journal *j = container_of(cl, struct journal, io); + struct journal_write *w = (j->cur == j->w) + ? &j->w[1] + : &j->w[0]; + + __closure_wake_up(&w->wait); + continue_at_nobarrier(cl, journal_write, bch_journal_wq); +} + +static void journal_write_unlock(struct closure *cl) + __releases(&c->journal.lock) +{ + struct cache_set *c = container_of(cl, struct cache_set, journal.io); + + c->journal.io_in_flight = 0; + spin_unlock(&c->journal.lock); +} + +static void journal_write_unlocked(struct closure *cl) + __releases(c->journal.lock) +{ + struct cache_set *c = container_of(cl, struct cache_set, journal.io); + struct cache *ca = c->cache; + struct journal_write *w = c->journal.cur; + struct bkey *k = &c->journal.key; + unsigned int i, sectors = set_blocks(w->data, block_bytes(ca)) * + ca->sb.block_size; + + struct bio *bio; + struct bio_list list; + + bio_list_init(&list); + + if (!w->need_write) { + closure_return_with_destructor(cl, journal_write_unlock); + return; + } else if (journal_full(&c->journal)) { + journal_reclaim(c); + spin_unlock(&c->journal.lock); + + btree_flush_write(c); + continue_at(cl, journal_write, bch_journal_wq); + return; + } + + c->journal.blocks_free -= set_blocks(w->data, block_bytes(ca)); + + w->data->btree_level = c->root->level; + + bkey_copy(&w->data->btree_root, &c->root->key); + bkey_copy(&w->data->uuid_bucket, &c->uuid_bucket); + + w->data->prio_bucket[ca->sb.nr_this_dev] = ca->prio_buckets[0]; + w->data->magic = jset_magic(&ca->sb); + w->data->version = BCACHE_JSET_VERSION; + w->data->last_seq = last_seq(&c->journal); + w->data->csum = csum_set(w->data); + + for (i = 0; i < KEY_PTRS(k); i++) { + ca = PTR_CACHE(c, k, i); + bio = &ca->journal.bio; + + atomic_long_add(sectors, &ca->meta_sectors_written); + + bio_reset(bio); + bio->bi_iter.bi_sector = PTR_OFFSET(k, i); + bio_set_dev(bio, ca->bdev); + bio->bi_iter.bi_size = sectors << 9; + + bio->bi_end_io = journal_write_endio; + bio->bi_private = w; + bio_set_op_attrs(bio, REQ_OP_WRITE, + REQ_SYNC|REQ_META|REQ_PREFLUSH|REQ_FUA); + bch_bio_map(bio, w->data); + + trace_bcache_journal_write(bio, w->data->keys); + bio_list_add(&list, bio); + + SET_PTR_OFFSET(k, i, PTR_OFFSET(k, i) + sectors); + + ca->journal.seq[ca->journal.cur_idx] = w->data->seq; + } + + /* If KEY_PTRS(k) == 0, this jset gets lost in air */ + BUG_ON(i == 0); + + atomic_dec_bug(&fifo_back(&c->journal.pin)); + bch_journal_next(&c->journal); + journal_reclaim(c); + + spin_unlock(&c->journal.lock); + + while ((bio = bio_list_pop(&list))) + closure_bio_submit(c, bio, cl); + + continue_at(cl, journal_write_done, NULL); +} + +static void journal_write(struct closure *cl) +{ + struct cache_set *c = container_of(cl, struct cache_set, journal.io); + + spin_lock(&c->journal.lock); + journal_write_unlocked(cl); +} + +static void journal_try_write(struct cache_set *c) + __releases(c->journal.lock) +{ + struct closure *cl = &c->journal.io; + struct journal_write *w = c->journal.cur; + + w->need_write = true; + + if (!c->journal.io_in_flight) { + c->journal.io_in_flight = 1; + closure_call(cl, journal_write_unlocked, NULL, &c->cl); + } else { + spin_unlock(&c->journal.lock); + } +} + +static struct journal_write *journal_wait_for_write(struct cache_set *c, + unsigned int nkeys) + __acquires(&c->journal.lock) +{ + size_t sectors; + struct closure cl; + bool wait = false; + struct cache *ca = c->cache; + + closure_init_stack(&cl); + + spin_lock(&c->journal.lock); + + while (1) { + struct journal_write *w = c->journal.cur; + + sectors = __set_blocks(w->data, w->data->keys + nkeys, + block_bytes(ca)) * ca->sb.block_size; + + if (sectors <= min_t(size_t, + c->journal.blocks_free * ca->sb.block_size, + PAGE_SECTORS << JSET_BITS)) + return w; + + if (wait) + closure_wait(&c->journal.wait, &cl); + + if (!journal_full(&c->journal)) { + if (wait) + trace_bcache_journal_entry_full(c); + + /* + * XXX: If we were inserting so many keys that they + * won't fit in an _empty_ journal write, we'll + * deadlock. For now, handle this in + * bch_keylist_realloc() - but something to think about. + */ + BUG_ON(!w->data->keys); + + journal_try_write(c); /* unlocks */ + } else { + if (wait) + trace_bcache_journal_full(c); + + journal_reclaim(c); + spin_unlock(&c->journal.lock); + + btree_flush_write(c); + } + + closure_sync(&cl); + spin_lock(&c->journal.lock); + wait = true; + } +} + +static void journal_write_work(struct work_struct *work) +{ + struct cache_set *c = container_of(to_delayed_work(work), + struct cache_set, + journal.work); + spin_lock(&c->journal.lock); + if (c->journal.cur->dirty) + journal_try_write(c); + else + spin_unlock(&c->journal.lock); +} + +/* + * Entry point to the journalling code - bio_insert() and btree_invalidate() + * pass bch_journal() a list of keys to be journalled, and then + * bch_journal() hands those same keys off to btree_insert_async() + */ + +atomic_t *bch_journal(struct cache_set *c, + struct keylist *keys, + struct closure *parent) +{ + struct journal_write *w; + atomic_t *ret; + + /* No journaling if CACHE_SET_IO_DISABLE set already */ + if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &c->flags))) + return NULL; + + if (!CACHE_SYNC(&c->cache->sb)) + return NULL; + + w = journal_wait_for_write(c, bch_keylist_nkeys(keys)); + + memcpy(bset_bkey_last(w->data), keys->keys, bch_keylist_bytes(keys)); + w->data->keys += bch_keylist_nkeys(keys); + + ret = &fifo_back(&c->journal.pin); + atomic_inc(ret); + + if (parent) { + closure_wait(&w->wait, parent); + journal_try_write(c); + } else if (!w->dirty) { + w->dirty = true; + queue_delayed_work(bch_flush_wq, &c->journal.work, + msecs_to_jiffies(c->journal_delay_ms)); + spin_unlock(&c->journal.lock); + } else { + spin_unlock(&c->journal.lock); + } + + + return ret; +} + +void bch_journal_meta(struct cache_set *c, struct closure *cl) +{ + struct keylist keys; + atomic_t *ref; + + bch_keylist_init(&keys); + + ref = bch_journal(c, &keys, cl); + if (ref) + atomic_dec_bug(ref); +} + +void bch_journal_free(struct cache_set *c) +{ + free_pages((unsigned long) c->journal.w[1].data, JSET_BITS); + free_pages((unsigned long) c->journal.w[0].data, JSET_BITS); + free_fifo(&c->journal.pin); +} + +int bch_journal_alloc(struct cache_set *c) +{ + struct journal *j = &c->journal; + + spin_lock_init(&j->lock); + spin_lock_init(&j->flush_write_lock); + INIT_DELAYED_WORK(&j->work, journal_write_work); + + c->journal_delay_ms = 100; + + j->w[0].c = c; + j->w[1].c = c; + + if (!(init_fifo(&j->pin, JOURNAL_PIN, GFP_KERNEL)) || + !(j->w[0].data = (void *) __get_free_pages(GFP_KERNEL|__GFP_COMP, JSET_BITS)) || + !(j->w[1].data = (void *) __get_free_pages(GFP_KERNEL|__GFP_COMP, JSET_BITS))) + return -ENOMEM; + + return 0; +} diff --git a/drivers/md/bcache/journal.h b/drivers/md/bcache/journal.h new file mode 100644 index 000000000..cd316b4a1 --- /dev/null +++ b/drivers/md/bcache/journal.h @@ -0,0 +1,188 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _BCACHE_JOURNAL_H +#define _BCACHE_JOURNAL_H + +/* + * THE JOURNAL: + * + * The journal is treated as a circular buffer of buckets - a journal entry + * never spans two buckets. This means (not implemented yet) we can resize the + * journal at runtime, and will be needed for bcache on raw flash support. + * + * Journal entries contain a list of keys, ordered by the time they were + * inserted; thus journal replay just has to reinsert the keys. + * + * We also keep some things in the journal header that are logically part of the + * superblock - all the things that are frequently updated. This is for future + * bcache on raw flash support; the superblock (which will become another + * journal) can't be moved or wear leveled, so it contains just enough + * information to find the main journal, and the superblock only has to be + * rewritten when we want to move/wear level the main journal. + * + * Currently, we don't journal BTREE_REPLACE operations - this will hopefully be + * fixed eventually. This isn't a bug - BTREE_REPLACE is used for insertions + * from cache misses, which don't have to be journaled, and for writeback and + * moving gc we work around it by flushing the btree to disk before updating the + * gc information. But it is a potential issue with incremental garbage + * collection, and it's fragile. + * + * OPEN JOURNAL ENTRIES: + * + * Each journal entry contains, in the header, the sequence number of the last + * journal entry still open - i.e. that has keys that haven't been flushed to + * disk in the btree. + * + * We track this by maintaining a refcount for every open journal entry, in a + * fifo; each entry in the fifo corresponds to a particular journal + * entry/sequence number. When the refcount at the tail of the fifo goes to + * zero, we pop it off - thus, the size of the fifo tells us the number of open + * journal entries + * + * We take a refcount on a journal entry when we add some keys to a journal + * entry that we're going to insert (held by struct btree_op), and then when we + * insert those keys into the btree the btree write we're setting up takes a + * copy of that refcount (held by struct btree_write). That refcount is dropped + * when the btree write completes. + * + * A struct btree_write can only hold a refcount on a single journal entry, but + * might contain keys for many journal entries - we handle this by making sure + * it always has a refcount on the _oldest_ journal entry of all the journal + * entries it has keys for. + * + * JOURNAL RECLAIM: + * + * As mentioned previously, our fifo of refcounts tells us the number of open + * journal entries; from that and the current journal sequence number we compute + * last_seq - the oldest journal entry we still need. We write last_seq in each + * journal entry, and we also have to keep track of where it exists on disk so + * we don't overwrite it when we loop around the journal. + * + * To do that we track, for each journal bucket, the sequence number of the + * newest journal entry it contains - if we don't need that journal entry we + * don't need anything in that bucket anymore. From that we track the last + * journal bucket we still need; all this is tracked in struct journal_device + * and updated by journal_reclaim(). + * + * JOURNAL FILLING UP: + * + * There are two ways the journal could fill up; either we could run out of + * space to write to, or we could have too many open journal entries and run out + * of room in the fifo of refcounts. Since those refcounts are decremented + * without any locking we can't safely resize that fifo, so we handle it the + * same way. + * + * If the journal fills up, we start flushing dirty btree nodes until we can + * allocate space for a journal write again - preferentially flushing btree + * nodes that are pinning the oldest journal entries first. + */ + +/* + * Only used for holding the journal entries we read in btree_journal_read() + * during cache_registration + */ +struct journal_replay { + struct list_head list; + atomic_t *pin; + struct jset j; +}; + +/* + * We put two of these in struct journal; we used them for writes to the + * journal that are being staged or in flight. + */ +struct journal_write { + struct jset *data; +#define JSET_BITS 3 + + struct cache_set *c; + struct closure_waitlist wait; + bool dirty; + bool need_write; +}; + +/* Embedded in struct cache_set */ +struct journal { + spinlock_t lock; + spinlock_t flush_write_lock; + bool btree_flushing; + bool do_reserve; + /* used when waiting because the journal was full */ + struct closure_waitlist wait; + struct closure io; + int io_in_flight; + struct delayed_work work; + + /* Number of blocks free in the bucket(s) we're currently writing to */ + unsigned int blocks_free; + uint64_t seq; + DECLARE_FIFO(atomic_t, pin); + + BKEY_PADDED(key); + + struct journal_write w[2], *cur; +}; + +/* + * Embedded in struct cache. First three fields refer to the array of journal + * buckets, in cache_sb. + */ +struct journal_device { + /* + * For each journal bucket, contains the max sequence number of the + * journal writes it contains - so we know when a bucket can be reused. + */ + uint64_t seq[SB_JOURNAL_BUCKETS]; + + /* Journal bucket we're currently writing to */ + unsigned int cur_idx; + + /* Last journal bucket that still contains an open journal entry */ + unsigned int last_idx; + + /* Next journal bucket to be discarded */ + unsigned int discard_idx; + +#define DISCARD_READY 0 +#define DISCARD_IN_FLIGHT 1 +#define DISCARD_DONE 2 + /* 1 - discard in flight, -1 - discard completed */ + atomic_t discard_in_flight; + + struct work_struct discard_work; + struct bio discard_bio; + struct bio_vec discard_bv; + + /* Bio for journal reads/writes to this device */ + struct bio bio; + struct bio_vec bv[8]; +}; + +#define BTREE_FLUSH_NR 8 + +#define journal_pin_cmp(c, l, r) \ + (fifo_idx(&(c)->journal.pin, (l)) > fifo_idx(&(c)->journal.pin, (r))) + +#define JOURNAL_PIN 20000 + +#define journal_full(j) \ + (!(j)->blocks_free || fifo_free(&(j)->pin) <= 1) + +struct closure; +struct cache_set; +struct btree_op; +struct keylist; + +atomic_t *bch_journal(struct cache_set *c, + struct keylist *keys, + struct closure *parent); +void bch_journal_next(struct journal *j); +void bch_journal_mark(struct cache_set *c, struct list_head *list); +void bch_journal_meta(struct cache_set *c, struct closure *cl); +int bch_journal_read(struct cache_set *c, struct list_head *list); +int bch_journal_replay(struct cache_set *c, struct list_head *list); + +void bch_journal_free(struct cache_set *c); +int bch_journal_alloc(struct cache_set *c); +void bch_journal_space_reserve(struct journal *j); + +#endif /* _BCACHE_JOURNAL_H */ diff --git a/drivers/md/bcache/movinggc.c b/drivers/md/bcache/movinggc.c new file mode 100644 index 000000000..b9c3d27ec --- /dev/null +++ b/drivers/md/bcache/movinggc.c @@ -0,0 +1,252 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Moving/copying garbage collector + * + * Copyright 2012 Google, Inc. + */ + +#include "bcache.h" +#include "btree.h" +#include "debug.h" +#include "request.h" + +#include <trace/events/bcache.h> + +struct moving_io { + struct closure cl; + struct keybuf_key *w; + struct data_insert_op op; + struct bbio bio; +}; + +static bool moving_pred(struct keybuf *buf, struct bkey *k) +{ + struct cache_set *c = container_of(buf, struct cache_set, + moving_gc_keys); + unsigned int i; + + for (i = 0; i < KEY_PTRS(k); i++) + if (ptr_available(c, k, i) && + GC_MOVE(PTR_BUCKET(c, k, i))) + return true; + + return false; +} + +/* Moving GC - IO loop */ + +static void moving_io_destructor(struct closure *cl) +{ + struct moving_io *io = container_of(cl, struct moving_io, cl); + + kfree(io); +} + +static void write_moving_finish(struct closure *cl) +{ + struct moving_io *io = container_of(cl, struct moving_io, cl); + struct bio *bio = &io->bio.bio; + + bio_free_pages(bio); + + if (io->op.replace_collision) + trace_bcache_gc_copy_collision(&io->w->key); + + bch_keybuf_del(&io->op.c->moving_gc_keys, io->w); + + up(&io->op.c->moving_in_flight); + + closure_return_with_destructor(cl, moving_io_destructor); +} + +static void read_moving_endio(struct bio *bio) +{ + struct bbio *b = container_of(bio, struct bbio, bio); + struct moving_io *io = container_of(bio->bi_private, + struct moving_io, cl); + + if (bio->bi_status) + io->op.status = bio->bi_status; + else if (!KEY_DIRTY(&b->key) && + ptr_stale(io->op.c, &b->key, 0)) { + io->op.status = BLK_STS_IOERR; + } + + bch_bbio_endio(io->op.c, bio, bio->bi_status, "reading data to move"); +} + +static void moving_init(struct moving_io *io) +{ + struct bio *bio = &io->bio.bio; + + bio_init(bio, bio->bi_inline_vecs, + DIV_ROUND_UP(KEY_SIZE(&io->w->key), PAGE_SECTORS)); + bio_get(bio); + bio_set_prio(bio, IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0)); + + bio->bi_iter.bi_size = KEY_SIZE(&io->w->key) << 9; + bio->bi_private = &io->cl; + bch_bio_map(bio, NULL); +} + +static void write_moving(struct closure *cl) +{ + struct moving_io *io = container_of(cl, struct moving_io, cl); + struct data_insert_op *op = &io->op; + + if (!op->status) { + moving_init(io); + + io->bio.bio.bi_iter.bi_sector = KEY_START(&io->w->key); + op->write_prio = 1; + op->bio = &io->bio.bio; + + op->writeback = KEY_DIRTY(&io->w->key); + op->csum = KEY_CSUM(&io->w->key); + + bkey_copy(&op->replace_key, &io->w->key); + op->replace = true; + + closure_call(&op->cl, bch_data_insert, NULL, cl); + } + + continue_at(cl, write_moving_finish, op->wq); +} + +static void read_moving_submit(struct closure *cl) +{ + struct moving_io *io = container_of(cl, struct moving_io, cl); + struct bio *bio = &io->bio.bio; + + bch_submit_bbio(bio, io->op.c, &io->w->key, 0); + + continue_at(cl, write_moving, io->op.wq); +} + +static void read_moving(struct cache_set *c) +{ + struct keybuf_key *w; + struct moving_io *io; + struct bio *bio; + struct closure cl; + + closure_init_stack(&cl); + + /* XXX: if we error, background writeback could stall indefinitely */ + + while (!test_bit(CACHE_SET_STOPPING, &c->flags)) { + w = bch_keybuf_next_rescan(c, &c->moving_gc_keys, + &MAX_KEY, moving_pred); + if (!w) + break; + + if (ptr_stale(c, &w->key, 0)) { + bch_keybuf_del(&c->moving_gc_keys, w); + continue; + } + + io = kzalloc(struct_size(io, bio.bio.bi_inline_vecs, + DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS)), + GFP_KERNEL); + if (!io) + goto err; + + w->private = io; + io->w = w; + io->op.inode = KEY_INODE(&w->key); + io->op.c = c; + io->op.wq = c->moving_gc_wq; + + moving_init(io); + bio = &io->bio.bio; + + bio_set_op_attrs(bio, REQ_OP_READ, 0); + bio->bi_end_io = read_moving_endio; + + if (bch_bio_alloc_pages(bio, GFP_KERNEL)) + goto err; + + trace_bcache_gc_copy(&w->key); + + down(&c->moving_in_flight); + closure_call(&io->cl, read_moving_submit, NULL, &cl); + } + + if (0) { +err: if (!IS_ERR_OR_NULL(w->private)) + kfree(w->private); + + bch_keybuf_del(&c->moving_gc_keys, w); + } + + closure_sync(&cl); +} + +static bool bucket_cmp(struct bucket *l, struct bucket *r) +{ + return GC_SECTORS_USED(l) < GC_SECTORS_USED(r); +} + +static unsigned int bucket_heap_top(struct cache *ca) +{ + struct bucket *b; + + return (b = heap_peek(&ca->heap)) ? GC_SECTORS_USED(b) : 0; +} + +void bch_moving_gc(struct cache_set *c) +{ + struct cache *ca = c->cache; + struct bucket *b; + unsigned long sectors_to_move, reserve_sectors; + + if (!c->copy_gc_enabled) + return; + + mutex_lock(&c->bucket_lock); + + sectors_to_move = 0; + reserve_sectors = ca->sb.bucket_size * + fifo_used(&ca->free[RESERVE_MOVINGGC]); + + ca->heap.used = 0; + + for_each_bucket(b, ca) { + if (GC_MARK(b) == GC_MARK_METADATA || + !GC_SECTORS_USED(b) || + GC_SECTORS_USED(b) == ca->sb.bucket_size || + atomic_read(&b->pin)) + continue; + + if (!heap_full(&ca->heap)) { + sectors_to_move += GC_SECTORS_USED(b); + heap_add(&ca->heap, b, bucket_cmp); + } else if (bucket_cmp(b, heap_peek(&ca->heap))) { + sectors_to_move -= bucket_heap_top(ca); + sectors_to_move += GC_SECTORS_USED(b); + + ca->heap.data[0] = b; + heap_sift(&ca->heap, 0, bucket_cmp); + } + } + + while (sectors_to_move > reserve_sectors) { + heap_pop(&ca->heap, b, bucket_cmp); + sectors_to_move -= GC_SECTORS_USED(b); + } + + while (heap_pop(&ca->heap, b, bucket_cmp)) + SET_GC_MOVE(b, 1); + + mutex_unlock(&c->bucket_lock); + + c->moving_gc_keys.last_scanned = ZERO_KEY; + + read_moving(c); +} + +void bch_moving_init_cache_set(struct cache_set *c) +{ + bch_keybuf_init(&c->moving_gc_keys); + sema_init(&c->moving_in_flight, 64); +} diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c new file mode 100644 index 000000000..97895262f --- /dev/null +++ b/drivers/md/bcache/request.c @@ -0,0 +1,1346 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Main bcache entry point - handle a read or a write request and decide what to + * do with it; the make_request functions are called by the block layer. + * + * Copyright 2010, 2011 Kent Overstreet <kent.overstreet@gmail.com> + * Copyright 2012 Google, Inc. + */ + +#include "bcache.h" +#include "btree.h" +#include "debug.h" +#include "request.h" +#include "writeback.h" + +#include <linux/module.h> +#include <linux/hash.h> +#include <linux/random.h> +#include <linux/backing-dev.h> + +#include <trace/events/bcache.h> + +#define CUTOFF_CACHE_ADD 95 +#define CUTOFF_CACHE_READA 90 + +struct kmem_cache *bch_search_cache; + +static void bch_data_insert_start(struct closure *cl); + +static unsigned int cache_mode(struct cached_dev *dc) +{ + return BDEV_CACHE_MODE(&dc->sb); +} + +static bool verify(struct cached_dev *dc) +{ + return dc->verify; +} + +static void bio_csum(struct bio *bio, struct bkey *k) +{ + struct bio_vec bv; + struct bvec_iter iter; + uint64_t csum = 0; + + bio_for_each_segment(bv, bio, iter) { + void *d = kmap(bv.bv_page) + bv.bv_offset; + + csum = bch_crc64_update(csum, d, bv.bv_len); + kunmap(bv.bv_page); + } + + k->ptr[KEY_PTRS(k)] = csum & (~0ULL >> 1); +} + +/* Insert data into cache */ + +static void bch_data_insert_keys(struct closure *cl) +{ + struct data_insert_op *op = container_of(cl, struct data_insert_op, cl); + atomic_t *journal_ref = NULL; + struct bkey *replace_key = op->replace ? &op->replace_key : NULL; + int ret; + + if (!op->replace) + journal_ref = bch_journal(op->c, &op->insert_keys, + op->flush_journal ? cl : NULL); + + ret = bch_btree_insert(op->c, &op->insert_keys, + journal_ref, replace_key); + if (ret == -ESRCH) { + op->replace_collision = true; + } else if (ret) { + op->status = BLK_STS_RESOURCE; + op->insert_data_done = true; + } + + if (journal_ref) + atomic_dec_bug(journal_ref); + + if (!op->insert_data_done) { + continue_at(cl, bch_data_insert_start, op->wq); + return; + } + + bch_keylist_free(&op->insert_keys); + closure_return(cl); +} + +static int bch_keylist_realloc(struct keylist *l, unsigned int u64s, + struct cache_set *c) +{ + size_t oldsize = bch_keylist_nkeys(l); + size_t newsize = oldsize + u64s; + + /* + * The journalling code doesn't handle the case where the keys to insert + * is bigger than an empty write: If we just return -ENOMEM here, + * bch_data_insert_keys() will insert the keys created so far + * and finish the rest when the keylist is empty. + */ + if (newsize * sizeof(uint64_t) > block_bytes(c->cache) - sizeof(struct jset)) + return -ENOMEM; + + return __bch_keylist_realloc(l, u64s); +} + +static void bch_data_invalidate(struct closure *cl) +{ + struct data_insert_op *op = container_of(cl, struct data_insert_op, cl); + struct bio *bio = op->bio; + + pr_debug("invalidating %i sectors from %llu\n", + bio_sectors(bio), (uint64_t) bio->bi_iter.bi_sector); + + while (bio_sectors(bio)) { + unsigned int sectors = min(bio_sectors(bio), + 1U << (KEY_SIZE_BITS - 1)); + + if (bch_keylist_realloc(&op->insert_keys, 2, op->c)) + goto out; + + bio->bi_iter.bi_sector += sectors; + bio->bi_iter.bi_size -= sectors << 9; + + bch_keylist_add(&op->insert_keys, + &KEY(op->inode, + bio->bi_iter.bi_sector, + sectors)); + } + + op->insert_data_done = true; + /* get in bch_data_insert() */ + bio_put(bio); +out: + continue_at(cl, bch_data_insert_keys, op->wq); +} + +static void bch_data_insert_error(struct closure *cl) +{ + struct data_insert_op *op = container_of(cl, struct data_insert_op, cl); + + /* + * Our data write just errored, which means we've got a bunch of keys to + * insert that point to data that wasn't successfully written. + * + * We don't have to insert those keys but we still have to invalidate + * that region of the cache - so, if we just strip off all the pointers + * from the keys we'll accomplish just that. + */ + + struct bkey *src = op->insert_keys.keys, *dst = op->insert_keys.keys; + + while (src != op->insert_keys.top) { + struct bkey *n = bkey_next(src); + + SET_KEY_PTRS(src, 0); + memmove(dst, src, bkey_bytes(src)); + + dst = bkey_next(dst); + src = n; + } + + op->insert_keys.top = dst; + + bch_data_insert_keys(cl); +} + +static void bch_data_insert_endio(struct bio *bio) +{ + struct closure *cl = bio->bi_private; + struct data_insert_op *op = container_of(cl, struct data_insert_op, cl); + + if (bio->bi_status) { + /* TODO: We could try to recover from this. */ + if (op->writeback) + op->status = bio->bi_status; + else if (!op->replace) + set_closure_fn(cl, bch_data_insert_error, op->wq); + else + set_closure_fn(cl, NULL, NULL); + } + + bch_bbio_endio(op->c, bio, bio->bi_status, "writing data to cache"); +} + +static void bch_data_insert_start(struct closure *cl) +{ + struct data_insert_op *op = container_of(cl, struct data_insert_op, cl); + struct bio *bio = op->bio, *n; + + if (op->bypass) + return bch_data_invalidate(cl); + + if (atomic_sub_return(bio_sectors(bio), &op->c->sectors_to_gc) < 0) + wake_up_gc(op->c); + + /* + * Journal writes are marked REQ_PREFLUSH; if the original write was a + * flush, it'll wait on the journal write. + */ + bio->bi_opf &= ~(REQ_PREFLUSH|REQ_FUA); + + do { + unsigned int i; + struct bkey *k; + struct bio_set *split = &op->c->bio_split; + + /* 1 for the device pointer and 1 for the chksum */ + if (bch_keylist_realloc(&op->insert_keys, + 3 + (op->csum ? 1 : 0), + op->c)) { + continue_at(cl, bch_data_insert_keys, op->wq); + return; + } + + k = op->insert_keys.top; + bkey_init(k); + SET_KEY_INODE(k, op->inode); + SET_KEY_OFFSET(k, bio->bi_iter.bi_sector); + + if (!bch_alloc_sectors(op->c, k, bio_sectors(bio), + op->write_point, op->write_prio, + op->writeback)) + goto err; + + n = bio_next_split(bio, KEY_SIZE(k), GFP_NOIO, split); + + n->bi_end_io = bch_data_insert_endio; + n->bi_private = cl; + + if (op->writeback) { + SET_KEY_DIRTY(k, true); + + for (i = 0; i < KEY_PTRS(k); i++) + SET_GC_MARK(PTR_BUCKET(op->c, k, i), + GC_MARK_DIRTY); + } + + SET_KEY_CSUM(k, op->csum); + if (KEY_CSUM(k)) + bio_csum(n, k); + + trace_bcache_cache_insert(k); + bch_keylist_push(&op->insert_keys); + + bio_set_op_attrs(n, REQ_OP_WRITE, 0); + bch_submit_bbio(n, op->c, k, 0); + } while (n != bio); + + op->insert_data_done = true; + continue_at(cl, bch_data_insert_keys, op->wq); + return; +err: + /* bch_alloc_sectors() blocks if s->writeback = true */ + BUG_ON(op->writeback); + + /* + * But if it's not a writeback write we'd rather just bail out if + * there aren't any buckets ready to write to - it might take awhile and + * we might be starving btree writes for gc or something. + */ + + if (!op->replace) { + /* + * Writethrough write: We can't complete the write until we've + * updated the index. But we don't want to delay the write while + * we wait for buckets to be freed up, so just invalidate the + * rest of the write. + */ + op->bypass = true; + return bch_data_invalidate(cl); + } else { + /* + * From a cache miss, we can just insert the keys for the data + * we have written or bail out if we didn't do anything. + */ + op->insert_data_done = true; + bio_put(bio); + + if (!bch_keylist_empty(&op->insert_keys)) + continue_at(cl, bch_data_insert_keys, op->wq); + else + closure_return(cl); + } +} + +/** + * bch_data_insert - stick some data in the cache + * @cl: closure pointer. + * + * This is the starting point for any data to end up in a cache device; it could + * be from a normal write, or a writeback write, or a write to a flash only + * volume - it's also used by the moving garbage collector to compact data in + * mostly empty buckets. + * + * It first writes the data to the cache, creating a list of keys to be inserted + * (if the data had to be fragmented there will be multiple keys); after the + * data is written it calls bch_journal, and after the keys have been added to + * the next journal write they're inserted into the btree. + * + * It inserts the data in op->bio; bi_sector is used for the key offset, + * and op->inode is used for the key inode. + * + * If op->bypass is true, instead of inserting the data it invalidates the + * region of the cache represented by op->bio and op->inode. + */ +void bch_data_insert(struct closure *cl) +{ + struct data_insert_op *op = container_of(cl, struct data_insert_op, cl); + + trace_bcache_write(op->c, op->inode, op->bio, + op->writeback, op->bypass); + + bch_keylist_init(&op->insert_keys); + bio_get(op->bio); + bch_data_insert_start(cl); +} + +/* + * Congested? Return 0 (not congested) or the limit (in sectors) + * beyond which we should bypass the cache due to congestion. + */ +unsigned int bch_get_congested(const struct cache_set *c) +{ + int i; + + if (!c->congested_read_threshold_us && + !c->congested_write_threshold_us) + return 0; + + i = (local_clock_us() - c->congested_last_us) / 1024; + if (i < 0) + return 0; + + i += atomic_read(&c->congested); + if (i >= 0) + return 0; + + i += CONGESTED_MAX; + + if (i > 0) + i = fract_exp_two(i, 6); + + i -= hweight32(get_random_u32()); + + return i > 0 ? i : 1; +} + +static void add_sequential(struct task_struct *t) +{ + ewma_add(t->sequential_io_avg, + t->sequential_io, 8, 0); + + t->sequential_io = 0; +} + +static struct hlist_head *iohash(struct cached_dev *dc, uint64_t k) +{ + return &dc->io_hash[hash_64(k, RECENT_IO_BITS)]; +} + +static bool check_should_bypass(struct cached_dev *dc, struct bio *bio) +{ + struct cache_set *c = dc->disk.c; + unsigned int mode = cache_mode(dc); + unsigned int sectors, congested; + struct task_struct *task = current; + struct io *i; + + if (test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags) || + c->gc_stats.in_use > CUTOFF_CACHE_ADD || + (bio_op(bio) == REQ_OP_DISCARD)) + goto skip; + + if (mode == CACHE_MODE_NONE || + (mode == CACHE_MODE_WRITEAROUND && + op_is_write(bio_op(bio)))) + goto skip; + + /* + * If the bio is for read-ahead or background IO, bypass it or + * not depends on the following situations, + * - If the IO is for meta data, always cache it and no bypass + * - If the IO is not meta data, check dc->cache_reada_policy, + * BCH_CACHE_READA_ALL: cache it and not bypass + * BCH_CACHE_READA_META_ONLY: not cache it and bypass + * That is, read-ahead request for metadata always get cached + * (eg, for gfs2 or xfs). + */ + if ((bio->bi_opf & (REQ_RAHEAD|REQ_BACKGROUND))) { + if (!(bio->bi_opf & (REQ_META|REQ_PRIO)) && + (dc->cache_readahead_policy != BCH_CACHE_READA_ALL)) + goto skip; + } + + if (bio->bi_iter.bi_sector & (c->cache->sb.block_size - 1) || + bio_sectors(bio) & (c->cache->sb.block_size - 1)) { + pr_debug("skipping unaligned io\n"); + goto skip; + } + + if (bypass_torture_test(dc)) { + if ((get_random_int() & 3) == 3) + goto skip; + else + goto rescale; + } + + congested = bch_get_congested(c); + if (!congested && !dc->sequential_cutoff) + goto rescale; + + spin_lock(&dc->io_lock); + + hlist_for_each_entry(i, iohash(dc, bio->bi_iter.bi_sector), hash) + if (i->last == bio->bi_iter.bi_sector && + time_before(jiffies, i->jiffies)) + goto found; + + i = list_first_entry(&dc->io_lru, struct io, lru); + + add_sequential(task); + i->sequential = 0; +found: + if (i->sequential + bio->bi_iter.bi_size > i->sequential) + i->sequential += bio->bi_iter.bi_size; + + i->last = bio_end_sector(bio); + i->jiffies = jiffies + msecs_to_jiffies(5000); + task->sequential_io = i->sequential; + + hlist_del(&i->hash); + hlist_add_head(&i->hash, iohash(dc, i->last)); + list_move_tail(&i->lru, &dc->io_lru); + + spin_unlock(&dc->io_lock); + + sectors = max(task->sequential_io, + task->sequential_io_avg) >> 9; + + if (dc->sequential_cutoff && + sectors >= dc->sequential_cutoff >> 9) { + trace_bcache_bypass_sequential(bio); + goto skip; + } + + if (congested && sectors >= congested) { + trace_bcache_bypass_congested(bio); + goto skip; + } + +rescale: + bch_rescale_priorities(c, bio_sectors(bio)); + return false; +skip: + bch_mark_sectors_bypassed(c, dc, bio_sectors(bio)); + return true; +} + +/* Cache lookup */ + +struct search { + /* Stack frame for bio_complete */ + struct closure cl; + + struct bbio bio; + struct bio *orig_bio; + struct bio *cache_miss; + struct bcache_device *d; + + unsigned int insert_bio_sectors; + unsigned int recoverable:1; + unsigned int write:1; + unsigned int read_dirty_data:1; + unsigned int cache_missed:1; + + struct hd_struct *part; + unsigned long start_time; + + struct btree_op op; + struct data_insert_op iop; +}; + +static void bch_cache_read_endio(struct bio *bio) +{ + struct bbio *b = container_of(bio, struct bbio, bio); + struct closure *cl = bio->bi_private; + struct search *s = container_of(cl, struct search, cl); + + /* + * If the bucket was reused while our bio was in flight, we might have + * read the wrong data. Set s->error but not error so it doesn't get + * counted against the cache device, but we'll still reread the data + * from the backing device. + */ + + if (bio->bi_status) + s->iop.status = bio->bi_status; + else if (!KEY_DIRTY(&b->key) && + ptr_stale(s->iop.c, &b->key, 0)) { + atomic_long_inc(&s->iop.c->cache_read_races); + s->iop.status = BLK_STS_IOERR; + } + + bch_bbio_endio(s->iop.c, bio, bio->bi_status, "reading from cache"); +} + +/* + * Read from a single key, handling the initial cache miss if the key starts in + * the middle of the bio + */ +static int cache_lookup_fn(struct btree_op *op, struct btree *b, struct bkey *k) +{ + struct search *s = container_of(op, struct search, op); + struct bio *n, *bio = &s->bio.bio; + struct bkey *bio_key; + unsigned int ptr; + + if (bkey_cmp(k, &KEY(s->iop.inode, bio->bi_iter.bi_sector, 0)) <= 0) + return MAP_CONTINUE; + + if (KEY_INODE(k) != s->iop.inode || + KEY_START(k) > bio->bi_iter.bi_sector) { + unsigned int bio_sectors = bio_sectors(bio); + unsigned int sectors = KEY_INODE(k) == s->iop.inode + ? min_t(uint64_t, INT_MAX, + KEY_START(k) - bio->bi_iter.bi_sector) + : INT_MAX; + int ret = s->d->cache_miss(b, s, bio, sectors); + + if (ret != MAP_CONTINUE) + return ret; + + /* if this was a complete miss we shouldn't get here */ + BUG_ON(bio_sectors <= sectors); + } + + if (!KEY_SIZE(k)) + return MAP_CONTINUE; + + /* XXX: figure out best pointer - for multiple cache devices */ + ptr = 0; + + PTR_BUCKET(b->c, k, ptr)->prio = INITIAL_PRIO; + + if (KEY_DIRTY(k)) + s->read_dirty_data = true; + + n = bio_next_split(bio, min_t(uint64_t, INT_MAX, + KEY_OFFSET(k) - bio->bi_iter.bi_sector), + GFP_NOIO, &s->d->bio_split); + + bio_key = &container_of(n, struct bbio, bio)->key; + bch_bkey_copy_single_ptr(bio_key, k, ptr); + + bch_cut_front(&KEY(s->iop.inode, n->bi_iter.bi_sector, 0), bio_key); + bch_cut_back(&KEY(s->iop.inode, bio_end_sector(n), 0), bio_key); + + n->bi_end_io = bch_cache_read_endio; + n->bi_private = &s->cl; + + /* + * The bucket we're reading from might be reused while our bio + * is in flight, and we could then end up reading the wrong + * data. + * + * We guard against this by checking (in cache_read_endio()) if + * the pointer is stale again; if so, we treat it as an error + * and reread from the backing device (but we don't pass that + * error up anywhere). + */ + + __bch_submit_bbio(n, b->c); + return n == bio ? MAP_DONE : MAP_CONTINUE; +} + +static void cache_lookup(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, iop.cl); + struct bio *bio = &s->bio.bio; + struct cached_dev *dc; + int ret; + + bch_btree_op_init(&s->op, -1); + + ret = bch_btree_map_keys(&s->op, s->iop.c, + &KEY(s->iop.inode, bio->bi_iter.bi_sector, 0), + cache_lookup_fn, MAP_END_KEY); + if (ret == -EAGAIN) { + continue_at(cl, cache_lookup, bcache_wq); + return; + } + + /* + * We might meet err when searching the btree, If that happens, we will + * get negative ret, in this scenario we should not recover data from + * backing device (when cache device is dirty) because we don't know + * whether bkeys the read request covered are all clean. + * + * And after that happened, s->iop.status is still its initial value + * before we submit s->bio.bio + */ + if (ret < 0) { + BUG_ON(ret == -EINTR); + if (s->d && s->d->c && + !UUID_FLASH_ONLY(&s->d->c->uuids[s->d->id])) { + dc = container_of(s->d, struct cached_dev, disk); + if (dc && atomic_read(&dc->has_dirty)) + s->recoverable = false; + } + if (!s->iop.status) + s->iop.status = BLK_STS_IOERR; + } + + closure_return(cl); +} + +/* Common code for the make_request functions */ + +static void request_endio(struct bio *bio) +{ + struct closure *cl = bio->bi_private; + + if (bio->bi_status) { + struct search *s = container_of(cl, struct search, cl); + + s->iop.status = bio->bi_status; + /* Only cache read errors are recoverable */ + s->recoverable = false; + } + + bio_put(bio); + closure_put(cl); +} + +static void backing_request_endio(struct bio *bio) +{ + struct closure *cl = bio->bi_private; + + if (bio->bi_status) { + struct search *s = container_of(cl, struct search, cl); + struct cached_dev *dc = container_of(s->d, + struct cached_dev, disk); + /* + * If a bio has REQ_PREFLUSH for writeback mode, it is + * speically assembled in cached_dev_write() for a non-zero + * write request which has REQ_PREFLUSH. we don't set + * s->iop.status by this failure, the status will be decided + * by result of bch_data_insert() operation. + */ + if (unlikely(s->iop.writeback && + bio->bi_opf & REQ_PREFLUSH)) { + pr_err("Can't flush %s: returned bi_status %i\n", + dc->backing_dev_name, bio->bi_status); + } else { + /* set to orig_bio->bi_status in bio_complete() */ + s->iop.status = bio->bi_status; + } + s->recoverable = false; + /* should count I/O error for backing device here */ + bch_count_backing_io_errors(dc, bio); + } + + bio_put(bio); + closure_put(cl); +} + +static void bio_complete(struct search *s) +{ + if (s->orig_bio) { + /* Count on bcache device */ + part_end_io_acct(s->part, s->orig_bio, s->start_time); + + trace_bcache_request_end(s->d, s->orig_bio); + s->orig_bio->bi_status = s->iop.status; + bio_endio(s->orig_bio); + s->orig_bio = NULL; + } +} + +static void do_bio_hook(struct search *s, + struct bio *orig_bio, + bio_end_io_t *end_io_fn) +{ + struct bio *bio = &s->bio.bio; + + bio_init(bio, NULL, 0); + __bio_clone_fast(bio, orig_bio); + /* + * bi_end_io can be set separately somewhere else, e.g. the + * variants in, + * - cache_bio->bi_end_io from cached_dev_cache_miss() + * - n->bi_end_io from cache_lookup_fn() + */ + bio->bi_end_io = end_io_fn; + bio->bi_private = &s->cl; + + bio_cnt_set(bio, 3); +} + +static void search_free(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, cl); + + atomic_dec(&s->iop.c->search_inflight); + + if (s->iop.bio) + bio_put(s->iop.bio); + + bio_complete(s); + closure_debug_destroy(cl); + mempool_free(s, &s->iop.c->search); +} + +static inline struct search *search_alloc(struct bio *bio, + struct bcache_device *d) +{ + struct search *s; + + s = mempool_alloc(&d->c->search, GFP_NOIO); + + closure_init(&s->cl, NULL); + do_bio_hook(s, bio, request_endio); + atomic_inc(&d->c->search_inflight); + + s->orig_bio = bio; + s->cache_miss = NULL; + s->cache_missed = 0; + s->d = d; + s->recoverable = 1; + s->write = op_is_write(bio_op(bio)); + s->read_dirty_data = 0; + /* Count on the bcache device */ + s->start_time = part_start_io_acct(d->disk, &s->part, bio); + s->iop.c = d->c; + s->iop.bio = NULL; + s->iop.inode = d->id; + s->iop.write_point = hash_long((unsigned long) current, 16); + s->iop.write_prio = 0; + s->iop.status = 0; + s->iop.flags = 0; + s->iop.flush_journal = op_is_flush(bio->bi_opf); + s->iop.wq = bcache_wq; + + return s; +} + +/* Cached devices */ + +static void cached_dev_bio_complete(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, cl); + struct cached_dev *dc = container_of(s->d, struct cached_dev, disk); + + cached_dev_put(dc); + search_free(cl); +} + +/* Process reads */ + +static void cached_dev_read_error_done(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, cl); + + if (s->iop.replace_collision) + bch_mark_cache_miss_collision(s->iop.c, s->d); + + if (s->iop.bio) + bio_free_pages(s->iop.bio); + + cached_dev_bio_complete(cl); +} + +static void cached_dev_read_error(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, cl); + struct bio *bio = &s->bio.bio; + + /* + * If read request hit dirty data (s->read_dirty_data is true), + * then recovery a failed read request from cached device may + * get a stale data back. So read failure recovery is only + * permitted when read request hit clean data in cache device, + * or when cache read race happened. + */ + if (s->recoverable && !s->read_dirty_data) { + /* Retry from the backing device: */ + trace_bcache_read_retry(s->orig_bio); + + s->iop.status = 0; + do_bio_hook(s, s->orig_bio, backing_request_endio); + + /* XXX: invalidate cache */ + + /* I/O request sent to backing device */ + closure_bio_submit(s->iop.c, bio, cl); + } + + continue_at(cl, cached_dev_read_error_done, NULL); +} + +static void cached_dev_cache_miss_done(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, cl); + struct bcache_device *d = s->d; + + if (s->iop.replace_collision) + bch_mark_cache_miss_collision(s->iop.c, s->d); + + if (s->iop.bio) + bio_free_pages(s->iop.bio); + + cached_dev_bio_complete(cl); + closure_put(&d->cl); +} + +static void cached_dev_read_done(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, cl); + struct cached_dev *dc = container_of(s->d, struct cached_dev, disk); + + /* + * We had a cache miss; cache_bio now contains data ready to be inserted + * into the cache. + * + * First, we copy the data we just read from cache_bio's bounce buffers + * to the buffers the original bio pointed to: + */ + + if (s->iop.bio) { + bio_reset(s->iop.bio); + s->iop.bio->bi_iter.bi_sector = + s->cache_miss->bi_iter.bi_sector; + bio_copy_dev(s->iop.bio, s->cache_miss); + s->iop.bio->bi_iter.bi_size = s->insert_bio_sectors << 9; + bch_bio_map(s->iop.bio, NULL); + + bio_copy_data(s->cache_miss, s->iop.bio); + + bio_put(s->cache_miss); + s->cache_miss = NULL; + } + + if (verify(dc) && s->recoverable && !s->read_dirty_data) + bch_data_verify(dc, s->orig_bio); + + closure_get(&dc->disk.cl); + bio_complete(s); + + if (s->iop.bio && + !test_bit(CACHE_SET_STOPPING, &s->iop.c->flags)) { + BUG_ON(!s->iop.replace); + closure_call(&s->iop.cl, bch_data_insert, NULL, cl); + } + + continue_at(cl, cached_dev_cache_miss_done, NULL); +} + +static void cached_dev_read_done_bh(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, cl); + struct cached_dev *dc = container_of(s->d, struct cached_dev, disk); + + bch_mark_cache_accounting(s->iop.c, s->d, + !s->cache_missed, s->iop.bypass); + trace_bcache_read(s->orig_bio, !s->cache_missed, s->iop.bypass); + + if (s->iop.status) + continue_at_nobarrier(cl, cached_dev_read_error, bcache_wq); + else if (s->iop.bio || verify(dc)) + continue_at_nobarrier(cl, cached_dev_read_done, bcache_wq); + else + continue_at_nobarrier(cl, cached_dev_bio_complete, NULL); +} + +static int cached_dev_cache_miss(struct btree *b, struct search *s, + struct bio *bio, unsigned int sectors) +{ + int ret = MAP_CONTINUE; + unsigned int reada = 0; + struct cached_dev *dc = container_of(s->d, struct cached_dev, disk); + struct bio *miss, *cache_bio; + + s->cache_missed = 1; + + if (s->cache_miss || s->iop.bypass) { + miss = bio_next_split(bio, sectors, GFP_NOIO, &s->d->bio_split); + ret = miss == bio ? MAP_DONE : MAP_CONTINUE; + goto out_submit; + } + + if (!(bio->bi_opf & REQ_RAHEAD) && + !(bio->bi_opf & (REQ_META|REQ_PRIO)) && + s->iop.c->gc_stats.in_use < CUTOFF_CACHE_READA) + reada = min_t(sector_t, dc->readahead >> 9, + get_capacity(bio->bi_disk) - bio_end_sector(bio)); + + s->insert_bio_sectors = min(sectors, bio_sectors(bio) + reada); + + s->iop.replace_key = KEY(s->iop.inode, + bio->bi_iter.bi_sector + s->insert_bio_sectors, + s->insert_bio_sectors); + + ret = bch_btree_insert_check_key(b, &s->op, &s->iop.replace_key); + if (ret) + return ret; + + s->iop.replace = true; + + miss = bio_next_split(bio, sectors, GFP_NOIO, &s->d->bio_split); + + /* btree_search_recurse()'s btree iterator is no good anymore */ + ret = miss == bio ? MAP_DONE : -EINTR; + + cache_bio = bio_alloc_bioset(GFP_NOWAIT, + DIV_ROUND_UP(s->insert_bio_sectors, PAGE_SECTORS), + &dc->disk.bio_split); + if (!cache_bio) + goto out_submit; + + cache_bio->bi_iter.bi_sector = miss->bi_iter.bi_sector; + bio_copy_dev(cache_bio, miss); + cache_bio->bi_iter.bi_size = s->insert_bio_sectors << 9; + + cache_bio->bi_end_io = backing_request_endio; + cache_bio->bi_private = &s->cl; + + bch_bio_map(cache_bio, NULL); + if (bch_bio_alloc_pages(cache_bio, __GFP_NOWARN|GFP_NOIO)) + goto out_put; + + if (reada) + bch_mark_cache_readahead(s->iop.c, s->d); + + s->cache_miss = miss; + s->iop.bio = cache_bio; + bio_get(cache_bio); + /* I/O request sent to backing device */ + closure_bio_submit(s->iop.c, cache_bio, &s->cl); + + return ret; +out_put: + bio_put(cache_bio); +out_submit: + miss->bi_end_io = backing_request_endio; + miss->bi_private = &s->cl; + /* I/O request sent to backing device */ + closure_bio_submit(s->iop.c, miss, &s->cl); + return ret; +} + +static void cached_dev_read(struct cached_dev *dc, struct search *s) +{ + struct closure *cl = &s->cl; + + closure_call(&s->iop.cl, cache_lookup, NULL, cl); + continue_at(cl, cached_dev_read_done_bh, NULL); +} + +/* Process writes */ + +static void cached_dev_write_complete(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, cl); + struct cached_dev *dc = container_of(s->d, struct cached_dev, disk); + + up_read_non_owner(&dc->writeback_lock); + cached_dev_bio_complete(cl); +} + +static void cached_dev_write(struct cached_dev *dc, struct search *s) +{ + struct closure *cl = &s->cl; + struct bio *bio = &s->bio.bio; + struct bkey start = KEY(dc->disk.id, bio->bi_iter.bi_sector, 0); + struct bkey end = KEY(dc->disk.id, bio_end_sector(bio), 0); + + bch_keybuf_check_overlapping(&s->iop.c->moving_gc_keys, &start, &end); + + down_read_non_owner(&dc->writeback_lock); + if (bch_keybuf_check_overlapping(&dc->writeback_keys, &start, &end)) { + /* + * We overlap with some dirty data undergoing background + * writeback, force this write to writeback + */ + s->iop.bypass = false; + s->iop.writeback = true; + } + + /* + * Discards aren't _required_ to do anything, so skipping if + * check_overlapping returned true is ok + * + * But check_overlapping drops dirty keys for which io hasn't started, + * so we still want to call it. + */ + if (bio_op(bio) == REQ_OP_DISCARD) + s->iop.bypass = true; + + if (should_writeback(dc, s->orig_bio, + cache_mode(dc), + s->iop.bypass)) { + s->iop.bypass = false; + s->iop.writeback = true; + } + + if (s->iop.bypass) { + s->iop.bio = s->orig_bio; + bio_get(s->iop.bio); + + if (bio_op(bio) == REQ_OP_DISCARD && + !blk_queue_discard(bdev_get_queue(dc->bdev))) + goto insert_data; + + /* I/O request sent to backing device */ + bio->bi_end_io = backing_request_endio; + closure_bio_submit(s->iop.c, bio, cl); + + } else if (s->iop.writeback) { + bch_writeback_add(dc); + s->iop.bio = bio; + + if (bio->bi_opf & REQ_PREFLUSH) { + /* + * Also need to send a flush to the backing + * device. + */ + struct bio *flush; + + flush = bio_alloc_bioset(GFP_NOIO, 0, + &dc->disk.bio_split); + if (!flush) { + s->iop.status = BLK_STS_RESOURCE; + goto insert_data; + } + bio_copy_dev(flush, bio); + flush->bi_end_io = backing_request_endio; + flush->bi_private = cl; + flush->bi_opf = REQ_OP_WRITE | REQ_PREFLUSH; + /* I/O request sent to backing device */ + closure_bio_submit(s->iop.c, flush, cl); + } + } else { + s->iop.bio = bio_clone_fast(bio, GFP_NOIO, &dc->disk.bio_split); + /* I/O request sent to backing device */ + bio->bi_end_io = backing_request_endio; + closure_bio_submit(s->iop.c, bio, cl); + } + +insert_data: + closure_call(&s->iop.cl, bch_data_insert, NULL, cl); + continue_at(cl, cached_dev_write_complete, NULL); +} + +static void cached_dev_nodata(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, cl); + struct bio *bio = &s->bio.bio; + + if (s->iop.flush_journal) + bch_journal_meta(s->iop.c, cl); + + /* If it's a flush, we send the flush to the backing device too */ + bio->bi_end_io = backing_request_endio; + closure_bio_submit(s->iop.c, bio, cl); + + continue_at(cl, cached_dev_bio_complete, NULL); +} + +struct detached_dev_io_private { + struct bcache_device *d; + unsigned long start_time; + bio_end_io_t *bi_end_io; + void *bi_private; + struct hd_struct *part; +}; + +static void detached_dev_end_io(struct bio *bio) +{ + struct detached_dev_io_private *ddip; + + ddip = bio->bi_private; + bio->bi_end_io = ddip->bi_end_io; + bio->bi_private = ddip->bi_private; + + /* Count on the bcache device */ + part_end_io_acct(ddip->part, bio, ddip->start_time); + + if (bio->bi_status) { + struct cached_dev *dc = container_of(ddip->d, + struct cached_dev, disk); + /* should count I/O error for backing device here */ + bch_count_backing_io_errors(dc, bio); + } + + kfree(ddip); + bio->bi_end_io(bio); +} + +static void detached_dev_do_request(struct bcache_device *d, struct bio *bio) +{ + struct detached_dev_io_private *ddip; + struct cached_dev *dc = container_of(d, struct cached_dev, disk); + + /* + * no need to call closure_get(&dc->disk.cl), + * because upper layer had already opened bcache device, + * which would call closure_get(&dc->disk.cl) + */ + ddip = kzalloc(sizeof(struct detached_dev_io_private), GFP_NOIO); + if (!ddip) { + bio->bi_status = BLK_STS_RESOURCE; + bio->bi_end_io(bio); + return; + } + + ddip->d = d; + /* Count on the bcache device */ + ddip->start_time = part_start_io_acct(d->disk, &ddip->part, bio); + ddip->bi_end_io = bio->bi_end_io; + ddip->bi_private = bio->bi_private; + bio->bi_end_io = detached_dev_end_io; + bio->bi_private = ddip; + + if ((bio_op(bio) == REQ_OP_DISCARD) && + !blk_queue_discard(bdev_get_queue(dc->bdev))) + bio->bi_end_io(bio); + else + submit_bio_noacct(bio); +} + +static void quit_max_writeback_rate(struct cache_set *c, + struct cached_dev *this_dc) +{ + int i; + struct bcache_device *d; + struct cached_dev *dc; + + /* + * mutex bch_register_lock may compete with other parallel requesters, + * or attach/detach operations on other backing device. Waiting to + * the mutex lock may increase I/O request latency for seconds or more. + * To avoid such situation, if mutext_trylock() failed, only writeback + * rate of current cached device is set to 1, and __update_write_back() + * will decide writeback rate of other cached devices (remember now + * c->idle_counter is 0 already). + */ + if (mutex_trylock(&bch_register_lock)) { + for (i = 0; i < c->devices_max_used; i++) { + if (!c->devices[i]) + continue; + + if (UUID_FLASH_ONLY(&c->uuids[i])) + continue; + + d = c->devices[i]; + dc = container_of(d, struct cached_dev, disk); + /* + * set writeback rate to default minimum value, + * then let update_writeback_rate() to decide the + * upcoming rate. + */ + atomic_long_set(&dc->writeback_rate.rate, 1); + } + mutex_unlock(&bch_register_lock); + } else + atomic_long_set(&this_dc->writeback_rate.rate, 1); +} + +/* Cached devices - read & write stuff */ + +blk_qc_t cached_dev_submit_bio(struct bio *bio) +{ + struct search *s; + struct bcache_device *d = bio->bi_disk->private_data; + struct cached_dev *dc = container_of(d, struct cached_dev, disk); + int rw = bio_data_dir(bio); + + if (unlikely((d->c && test_bit(CACHE_SET_IO_DISABLE, &d->c->flags)) || + dc->io_disable)) { + bio->bi_status = BLK_STS_IOERR; + bio_endio(bio); + return BLK_QC_T_NONE; + } + + if (likely(d->c)) { + if (atomic_read(&d->c->idle_counter)) + atomic_set(&d->c->idle_counter, 0); + /* + * If at_max_writeback_rate of cache set is true and new I/O + * comes, quit max writeback rate of all cached devices + * attached to this cache set, and set at_max_writeback_rate + * to false. + */ + if (unlikely(atomic_read(&d->c->at_max_writeback_rate) == 1)) { + atomic_set(&d->c->at_max_writeback_rate, 0); + quit_max_writeback_rate(d->c, dc); + } + } + + bio_set_dev(bio, dc->bdev); + bio->bi_iter.bi_sector += dc->sb.data_offset; + + if (cached_dev_get(dc)) { + s = search_alloc(bio, d); + trace_bcache_request_start(s->d, bio); + + if (!bio->bi_iter.bi_size) { + /* + * can't call bch_journal_meta from under + * submit_bio_noacct + */ + continue_at_nobarrier(&s->cl, + cached_dev_nodata, + bcache_wq); + } else { + s->iop.bypass = check_should_bypass(dc, bio); + + if (rw) + cached_dev_write(dc, s); + else + cached_dev_read(dc, s); + } + } else + /* I/O request sent to backing device */ + detached_dev_do_request(d, bio); + + return BLK_QC_T_NONE; +} + +static int cached_dev_ioctl(struct bcache_device *d, fmode_t mode, + unsigned int cmd, unsigned long arg) +{ + struct cached_dev *dc = container_of(d, struct cached_dev, disk); + + if (dc->io_disable) + return -EIO; + + return __blkdev_driver_ioctl(dc->bdev, mode, cmd, arg); +} + +void bch_cached_dev_request_init(struct cached_dev *dc) +{ + dc->disk.cache_miss = cached_dev_cache_miss; + dc->disk.ioctl = cached_dev_ioctl; +} + +/* Flash backed devices */ + +static int flash_dev_cache_miss(struct btree *b, struct search *s, + struct bio *bio, unsigned int sectors) +{ + unsigned int bytes = min(sectors, bio_sectors(bio)) << 9; + + swap(bio->bi_iter.bi_size, bytes); + zero_fill_bio(bio); + swap(bio->bi_iter.bi_size, bytes); + + bio_advance(bio, bytes); + + if (!bio->bi_iter.bi_size) + return MAP_DONE; + + return MAP_CONTINUE; +} + +static void flash_dev_nodata(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, cl); + + if (s->iop.flush_journal) + bch_journal_meta(s->iop.c, cl); + + continue_at(cl, search_free, NULL); +} + +blk_qc_t flash_dev_submit_bio(struct bio *bio) +{ + struct search *s; + struct closure *cl; + struct bcache_device *d = bio->bi_disk->private_data; + + if (unlikely(d->c && test_bit(CACHE_SET_IO_DISABLE, &d->c->flags))) { + bio->bi_status = BLK_STS_IOERR; + bio_endio(bio); + return BLK_QC_T_NONE; + } + + s = search_alloc(bio, d); + cl = &s->cl; + bio = &s->bio.bio; + + trace_bcache_request_start(s->d, bio); + + if (!bio->bi_iter.bi_size) { + /* + * can't call bch_journal_meta from under submit_bio_noacct + */ + continue_at_nobarrier(&s->cl, + flash_dev_nodata, + bcache_wq); + return BLK_QC_T_NONE; + } else if (bio_data_dir(bio)) { + bch_keybuf_check_overlapping(&s->iop.c->moving_gc_keys, + &KEY(d->id, bio->bi_iter.bi_sector, 0), + &KEY(d->id, bio_end_sector(bio), 0)); + + s->iop.bypass = (bio_op(bio) == REQ_OP_DISCARD) != 0; + s->iop.writeback = true; + s->iop.bio = bio; + + closure_call(&s->iop.cl, bch_data_insert, NULL, cl); + } else { + closure_call(&s->iop.cl, cache_lookup, NULL, cl); + } + + continue_at(cl, search_free, NULL); + return BLK_QC_T_NONE; +} + +static int flash_dev_ioctl(struct bcache_device *d, fmode_t mode, + unsigned int cmd, unsigned long arg) +{ + return -ENOTTY; +} + +void bch_flash_dev_request_init(struct bcache_device *d) +{ + d->cache_miss = flash_dev_cache_miss; + d->ioctl = flash_dev_ioctl; +} + +void bch_request_exit(void) +{ + kmem_cache_destroy(bch_search_cache); +} + +int __init bch_request_init(void) +{ + bch_search_cache = KMEM_CACHE(search, 0); + if (!bch_search_cache) + return -ENOMEM; + + return 0; +} diff --git a/drivers/md/bcache/request.h b/drivers/md/bcache/request.h new file mode 100644 index 000000000..82b38366a --- /dev/null +++ b/drivers/md/bcache/request.h @@ -0,0 +1,47 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _BCACHE_REQUEST_H_ +#define _BCACHE_REQUEST_H_ + +struct data_insert_op { + struct closure cl; + struct cache_set *c; + struct bio *bio; + struct workqueue_struct *wq; + + unsigned int inode; + uint16_t write_point; + uint16_t write_prio; + blk_status_t status; + + union { + uint16_t flags; + + struct { + unsigned int bypass:1; + unsigned int writeback:1; + unsigned int flush_journal:1; + unsigned int csum:1; + + unsigned int replace:1; + unsigned int replace_collision:1; + + unsigned int insert_data_done:1; + }; + }; + + struct keylist insert_keys; + BKEY_PADDED(replace_key); +}; + +unsigned int bch_get_congested(const struct cache_set *c); +void bch_data_insert(struct closure *cl); + +void bch_cached_dev_request_init(struct cached_dev *dc); +blk_qc_t cached_dev_submit_bio(struct bio *bio); + +void bch_flash_dev_request_init(struct bcache_device *d); +blk_qc_t flash_dev_submit_bio(struct bio *bio); + +extern struct kmem_cache *bch_search_cache; + +#endif /* _BCACHE_REQUEST_H_ */ diff --git a/drivers/md/bcache/stats.c b/drivers/md/bcache/stats.c new file mode 100644 index 000000000..503aafe18 --- /dev/null +++ b/drivers/md/bcache/stats.c @@ -0,0 +1,247 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * bcache stats code + * + * Copyright 2012 Google, Inc. + */ + +#include "bcache.h" +#include "stats.h" +#include "btree.h" +#include "sysfs.h" + +/* + * We keep absolute totals of various statistics, and addionally a set of three + * rolling averages. + * + * Every so often, a timer goes off and rescales the rolling averages. + * accounting_rescale[] is how many times the timer has to go off before we + * rescale each set of numbers; that gets us half lives of 5 minutes, one hour, + * and one day. + * + * accounting_delay is how often the timer goes off - 22 times in 5 minutes, + * and accounting_weight is what we use to rescale: + * + * pow(31 / 32, 22) ~= 1/2 + * + * So that we don't have to increment each set of numbers every time we (say) + * get a cache hit, we increment a single atomic_t in acc->collector, and when + * the rescale function runs it resets the atomic counter to 0 and adds its + * old value to each of the exported numbers. + * + * To reduce rounding error, the numbers in struct cache_stats are all + * stored left shifted by 16, and scaled back in the sysfs show() function. + */ + +static const unsigned int DAY_RESCALE = 288; +static const unsigned int HOUR_RESCALE = 12; +static const unsigned int FIVE_MINUTE_RESCALE = 1; +static const unsigned int accounting_delay = (HZ * 300) / 22; +static const unsigned int accounting_weight = 32; + +/* sysfs reading/writing */ + +read_attribute(cache_hits); +read_attribute(cache_misses); +read_attribute(cache_bypass_hits); +read_attribute(cache_bypass_misses); +read_attribute(cache_hit_ratio); +read_attribute(cache_readaheads); +read_attribute(cache_miss_collisions); +read_attribute(bypassed); + +SHOW(bch_stats) +{ + struct cache_stats *s = + container_of(kobj, struct cache_stats, kobj); +#define var(stat) (s->stat >> 16) + var_print(cache_hits); + var_print(cache_misses); + var_print(cache_bypass_hits); + var_print(cache_bypass_misses); + + sysfs_print(cache_hit_ratio, + DIV_SAFE(var(cache_hits) * 100, + var(cache_hits) + var(cache_misses))); + + var_print(cache_readaheads); + var_print(cache_miss_collisions); + sysfs_hprint(bypassed, var(sectors_bypassed) << 9); +#undef var + return 0; +} + +STORE(bch_stats) +{ + return size; +} + +static void bch_stats_release(struct kobject *k) +{ +} + +static struct attribute *bch_stats_files[] = { + &sysfs_cache_hits, + &sysfs_cache_misses, + &sysfs_cache_bypass_hits, + &sysfs_cache_bypass_misses, + &sysfs_cache_hit_ratio, + &sysfs_cache_readaheads, + &sysfs_cache_miss_collisions, + &sysfs_bypassed, + NULL +}; +static KTYPE(bch_stats); + +int bch_cache_accounting_add_kobjs(struct cache_accounting *acc, + struct kobject *parent) +{ + int ret = kobject_add(&acc->total.kobj, parent, + "stats_total"); + ret = ret ?: kobject_add(&acc->five_minute.kobj, parent, + "stats_five_minute"); + ret = ret ?: kobject_add(&acc->hour.kobj, parent, + "stats_hour"); + ret = ret ?: kobject_add(&acc->day.kobj, parent, + "stats_day"); + return ret; +} + +void bch_cache_accounting_clear(struct cache_accounting *acc) +{ + acc->total.cache_hits = 0; + acc->total.cache_misses = 0; + acc->total.cache_bypass_hits = 0; + acc->total.cache_bypass_misses = 0; + acc->total.cache_readaheads = 0; + acc->total.cache_miss_collisions = 0; + acc->total.sectors_bypassed = 0; +} + +void bch_cache_accounting_destroy(struct cache_accounting *acc) +{ + kobject_put(&acc->total.kobj); + kobject_put(&acc->five_minute.kobj); + kobject_put(&acc->hour.kobj); + kobject_put(&acc->day.kobj); + + atomic_set(&acc->closing, 1); + if (del_timer_sync(&acc->timer)) + closure_return(&acc->cl); +} + +/* EWMA scaling */ + +static void scale_stat(unsigned long *stat) +{ + *stat = ewma_add(*stat, 0, accounting_weight, 0); +} + +static void scale_stats(struct cache_stats *stats, unsigned long rescale_at) +{ + if (++stats->rescale == rescale_at) { + stats->rescale = 0; + scale_stat(&stats->cache_hits); + scale_stat(&stats->cache_misses); + scale_stat(&stats->cache_bypass_hits); + scale_stat(&stats->cache_bypass_misses); + scale_stat(&stats->cache_readaheads); + scale_stat(&stats->cache_miss_collisions); + scale_stat(&stats->sectors_bypassed); + } +} + +static void scale_accounting(struct timer_list *t) +{ + struct cache_accounting *acc = from_timer(acc, t, timer); + +#define move_stat(name) do { \ + unsigned int t = atomic_xchg(&acc->collector.name, 0); \ + t <<= 16; \ + acc->five_minute.name += t; \ + acc->hour.name += t; \ + acc->day.name += t; \ + acc->total.name += t; \ +} while (0) + + move_stat(cache_hits); + move_stat(cache_misses); + move_stat(cache_bypass_hits); + move_stat(cache_bypass_misses); + move_stat(cache_readaheads); + move_stat(cache_miss_collisions); + move_stat(sectors_bypassed); + + scale_stats(&acc->total, 0); + scale_stats(&acc->day, DAY_RESCALE); + scale_stats(&acc->hour, HOUR_RESCALE); + scale_stats(&acc->five_minute, FIVE_MINUTE_RESCALE); + + acc->timer.expires += accounting_delay; + + if (!atomic_read(&acc->closing)) + add_timer(&acc->timer); + else + closure_return(&acc->cl); +} + +static void mark_cache_stats(struct cache_stat_collector *stats, + bool hit, bool bypass) +{ + if (!bypass) + if (hit) + atomic_inc(&stats->cache_hits); + else + atomic_inc(&stats->cache_misses); + else + if (hit) + atomic_inc(&stats->cache_bypass_hits); + else + atomic_inc(&stats->cache_bypass_misses); +} + +void bch_mark_cache_accounting(struct cache_set *c, struct bcache_device *d, + bool hit, bool bypass) +{ + struct cached_dev *dc = container_of(d, struct cached_dev, disk); + + mark_cache_stats(&dc->accounting.collector, hit, bypass); + mark_cache_stats(&c->accounting.collector, hit, bypass); +} + +void bch_mark_cache_readahead(struct cache_set *c, struct bcache_device *d) +{ + struct cached_dev *dc = container_of(d, struct cached_dev, disk); + + atomic_inc(&dc->accounting.collector.cache_readaheads); + atomic_inc(&c->accounting.collector.cache_readaheads); +} + +void bch_mark_cache_miss_collision(struct cache_set *c, struct bcache_device *d) +{ + struct cached_dev *dc = container_of(d, struct cached_dev, disk); + + atomic_inc(&dc->accounting.collector.cache_miss_collisions); + atomic_inc(&c->accounting.collector.cache_miss_collisions); +} + +void bch_mark_sectors_bypassed(struct cache_set *c, struct cached_dev *dc, + int sectors) +{ + atomic_add(sectors, &dc->accounting.collector.sectors_bypassed); + atomic_add(sectors, &c->accounting.collector.sectors_bypassed); +} + +void bch_cache_accounting_init(struct cache_accounting *acc, + struct closure *parent) +{ + kobject_init(&acc->total.kobj, &bch_stats_ktype); + kobject_init(&acc->five_minute.kobj, &bch_stats_ktype); + kobject_init(&acc->hour.kobj, &bch_stats_ktype); + kobject_init(&acc->day.kobj, &bch_stats_ktype); + + closure_init(&acc->cl, parent); + timer_setup(&acc->timer, scale_accounting, 0); + acc->timer.expires = jiffies + accounting_delay; + add_timer(&acc->timer); +} diff --git a/drivers/md/bcache/stats.h b/drivers/md/bcache/stats.h new file mode 100644 index 000000000..abfaabf7e --- /dev/null +++ b/drivers/md/bcache/stats.h @@ -0,0 +1,65 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _BCACHE_STATS_H_ +#define _BCACHE_STATS_H_ + +struct cache_stat_collector { + atomic_t cache_hits; + atomic_t cache_misses; + atomic_t cache_bypass_hits; + atomic_t cache_bypass_misses; + atomic_t cache_readaheads; + atomic_t cache_miss_collisions; + atomic_t sectors_bypassed; +}; + +struct cache_stats { + struct kobject kobj; + + unsigned long cache_hits; + unsigned long cache_misses; + unsigned long cache_bypass_hits; + unsigned long cache_bypass_misses; + unsigned long cache_readaheads; + unsigned long cache_miss_collisions; + unsigned long sectors_bypassed; + + unsigned int rescale; +}; + +struct cache_accounting { + struct closure cl; + struct timer_list timer; + atomic_t closing; + + struct cache_stat_collector collector; + + struct cache_stats total; + struct cache_stats five_minute; + struct cache_stats hour; + struct cache_stats day; +}; + +struct cache_set; +struct cached_dev; +struct bcache_device; + +void bch_cache_accounting_init(struct cache_accounting *acc, + struct closure *parent); + +int bch_cache_accounting_add_kobjs(struct cache_accounting *acc, + struct kobject *parent); + +void bch_cache_accounting_clear(struct cache_accounting *acc); + +void bch_cache_accounting_destroy(struct cache_accounting *acc); + +void bch_mark_cache_accounting(struct cache_set *c, struct bcache_device *d, + bool hit, bool bypass); +void bch_mark_cache_readahead(struct cache_set *c, struct bcache_device *d); +void bch_mark_cache_miss_collision(struct cache_set *c, + struct bcache_device *d); +void bch_mark_sectors_bypassed(struct cache_set *c, + struct cached_dev *dc, + int sectors); + +#endif /* _BCACHE_STATS_H_ */ diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c new file mode 100644 index 000000000..04ddaa4bb --- /dev/null +++ b/drivers/md/bcache/super.c @@ -0,0 +1,2961 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * bcache setup/teardown code, and some metadata io - read a superblock and + * figure out what to do with it. + * + * Copyright 2010, 2011 Kent Overstreet <kent.overstreet@gmail.com> + * Copyright 2012 Google, Inc. + */ + +#include "bcache.h" +#include "btree.h" +#include "debug.h" +#include "extents.h" +#include "request.h" +#include "writeback.h" +#include "features.h" + +#include <linux/blkdev.h> +#include <linux/debugfs.h> +#include <linux/genhd.h> +#include <linux/idr.h> +#include <linux/kthread.h> +#include <linux/workqueue.h> +#include <linux/module.h> +#include <linux/random.h> +#include <linux/reboot.h> +#include <linux/sysfs.h> + +unsigned int bch_cutoff_writeback; +unsigned int bch_cutoff_writeback_sync; + +static const char bcache_magic[] = { + 0xc6, 0x85, 0x73, 0xf6, 0x4e, 0x1a, 0x45, 0xca, + 0x82, 0x65, 0xf5, 0x7f, 0x48, 0xba, 0x6d, 0x81 +}; + +static const char invalid_uuid[] = { + 0xa0, 0x3e, 0xf8, 0xed, 0x3e, 0xe1, 0xb8, 0x78, + 0xc8, 0x50, 0xfc, 0x5e, 0xcb, 0x16, 0xcd, 0x99 +}; + +static struct kobject *bcache_kobj; +struct mutex bch_register_lock; +bool bcache_is_reboot; +LIST_HEAD(bch_cache_sets); +static LIST_HEAD(uncached_devices); + +static int bcache_major; +static DEFINE_IDA(bcache_device_idx); +static wait_queue_head_t unregister_wait; +struct workqueue_struct *bcache_wq; +struct workqueue_struct *bch_flush_wq; +struct workqueue_struct *bch_journal_wq; + + +#define BTREE_MAX_PAGES (256 * 1024 / PAGE_SIZE) +/* limitation of partitions number on single bcache device */ +#define BCACHE_MINORS 128 +/* limitation of bcache devices number on single system */ +#define BCACHE_DEVICE_IDX_MAX ((1U << MINORBITS)/BCACHE_MINORS) + +/* Superblock */ + +static unsigned int get_bucket_size(struct cache_sb *sb, struct cache_sb_disk *s) +{ + unsigned int bucket_size = le16_to_cpu(s->bucket_size); + + if (sb->version >= BCACHE_SB_VERSION_CDEV_WITH_FEATURES) { + if (bch_has_feature_large_bucket(sb)) { + unsigned int max, order; + + max = sizeof(unsigned int) * BITS_PER_BYTE - 1; + order = le16_to_cpu(s->bucket_size); + /* + * bcache tool will make sure the overflow won't + * happen, an error message here is enough. + */ + if (order > max) + pr_err("Bucket size (1 << %u) overflows\n", + order); + bucket_size = 1 << order; + } else if (bch_has_feature_obso_large_bucket(sb)) { + bucket_size += + le16_to_cpu(s->obso_bucket_size_hi) << 16; + } + } + + return bucket_size; +} + +static const char *read_super_common(struct cache_sb *sb, struct block_device *bdev, + struct cache_sb_disk *s) +{ + const char *err; + unsigned int i; + + sb->first_bucket= le16_to_cpu(s->first_bucket); + sb->nbuckets = le64_to_cpu(s->nbuckets); + sb->bucket_size = get_bucket_size(sb, s); + + sb->nr_in_set = le16_to_cpu(s->nr_in_set); + sb->nr_this_dev = le16_to_cpu(s->nr_this_dev); + + err = "Too many journal buckets"; + if (sb->keys > SB_JOURNAL_BUCKETS) + goto err; + + err = "Too many buckets"; + if (sb->nbuckets > LONG_MAX) + goto err; + + err = "Not enough buckets"; + if (sb->nbuckets < 1 << 7) + goto err; + + err = "Bad block size (not power of 2)"; + if (!is_power_of_2(sb->block_size)) + goto err; + + err = "Bad block size (larger than page size)"; + if (sb->block_size > PAGE_SECTORS) + goto err; + + err = "Bad bucket size (not power of 2)"; + if (!is_power_of_2(sb->bucket_size)) + goto err; + + err = "Bad bucket size (smaller than page size)"; + if (sb->bucket_size < PAGE_SECTORS) + goto err; + + err = "Invalid superblock: device too small"; + if (get_capacity(bdev->bd_disk) < + sb->bucket_size * sb->nbuckets) + goto err; + + err = "Bad UUID"; + if (bch_is_zero(sb->set_uuid, 16)) + goto err; + + err = "Bad cache device number in set"; + if (!sb->nr_in_set || + sb->nr_in_set <= sb->nr_this_dev || + sb->nr_in_set > MAX_CACHES_PER_SET) + goto err; + + err = "Journal buckets not sequential"; + for (i = 0; i < sb->keys; i++) + if (sb->d[i] != sb->first_bucket + i) + goto err; + + err = "Too many journal buckets"; + if (sb->first_bucket + sb->keys > sb->nbuckets) + goto err; + + err = "Invalid superblock: first bucket comes before end of super"; + if (sb->first_bucket * sb->bucket_size < 16) + goto err; + + err = NULL; +err: + return err; +} + + +static const char *read_super(struct cache_sb *sb, struct block_device *bdev, + struct cache_sb_disk **res) +{ + const char *err; + struct cache_sb_disk *s; + struct page *page; + unsigned int i; + + page = read_cache_page_gfp(bdev->bd_inode->i_mapping, + SB_OFFSET >> PAGE_SHIFT, GFP_KERNEL); + if (IS_ERR(page)) + return "IO error"; + s = page_address(page) + offset_in_page(SB_OFFSET); + + sb->offset = le64_to_cpu(s->offset); + sb->version = le64_to_cpu(s->version); + + memcpy(sb->magic, s->magic, 16); + memcpy(sb->uuid, s->uuid, 16); + memcpy(sb->set_uuid, s->set_uuid, 16); + memcpy(sb->label, s->label, SB_LABEL_SIZE); + + sb->flags = le64_to_cpu(s->flags); + sb->seq = le64_to_cpu(s->seq); + sb->last_mount = le32_to_cpu(s->last_mount); + sb->keys = le16_to_cpu(s->keys); + + for (i = 0; i < SB_JOURNAL_BUCKETS; i++) + sb->d[i] = le64_to_cpu(s->d[i]); + + pr_debug("read sb version %llu, flags %llu, seq %llu, journal size %u\n", + sb->version, sb->flags, sb->seq, sb->keys); + + err = "Not a bcache superblock (bad offset)"; + if (sb->offset != SB_SECTOR) + goto err; + + err = "Not a bcache superblock (bad magic)"; + if (memcmp(sb->magic, bcache_magic, 16)) + goto err; + + err = "Bad checksum"; + if (s->csum != csum_set(s)) + goto err; + + err = "Bad UUID"; + if (bch_is_zero(sb->uuid, 16)) + goto err; + + sb->block_size = le16_to_cpu(s->block_size); + + err = "Superblock block size smaller than device block size"; + if (sb->block_size << 9 < bdev_logical_block_size(bdev)) + goto err; + + switch (sb->version) { + case BCACHE_SB_VERSION_BDEV: + sb->data_offset = BDEV_DATA_START_DEFAULT; + break; + case BCACHE_SB_VERSION_BDEV_WITH_OFFSET: + case BCACHE_SB_VERSION_BDEV_WITH_FEATURES: + sb->data_offset = le64_to_cpu(s->data_offset); + + err = "Bad data offset"; + if (sb->data_offset < BDEV_DATA_START_DEFAULT) + goto err; + + break; + case BCACHE_SB_VERSION_CDEV: + case BCACHE_SB_VERSION_CDEV_WITH_UUID: + err = read_super_common(sb, bdev, s); + if (err) + goto err; + break; + case BCACHE_SB_VERSION_CDEV_WITH_FEATURES: + /* + * Feature bits are needed in read_super_common(), + * convert them firstly. + */ + sb->feature_compat = le64_to_cpu(s->feature_compat); + sb->feature_incompat = le64_to_cpu(s->feature_incompat); + sb->feature_ro_compat = le64_to_cpu(s->feature_ro_compat); + + /* Check incompatible features */ + err = "Unsupported compatible feature found"; + if (bch_has_unknown_compat_features(sb)) + goto err; + + err = "Unsupported read-only compatible feature found"; + if (bch_has_unknown_ro_compat_features(sb)) + goto err; + + err = "Unsupported incompatible feature found"; + if (bch_has_unknown_incompat_features(sb)) + goto err; + + err = read_super_common(sb, bdev, s); + if (err) + goto err; + break; + default: + err = "Unsupported superblock version"; + goto err; + } + + sb->last_mount = (u32)ktime_get_real_seconds(); + *res = s; + return NULL; +err: + put_page(page); + return err; +} + +static void write_bdev_super_endio(struct bio *bio) +{ + struct cached_dev *dc = bio->bi_private; + + if (bio->bi_status) + bch_count_backing_io_errors(dc, bio); + + closure_put(&dc->sb_write); +} + +static void __write_super(struct cache_sb *sb, struct cache_sb_disk *out, + struct bio *bio) +{ + unsigned int i; + + bio->bi_opf = REQ_OP_WRITE | REQ_SYNC | REQ_META; + bio->bi_iter.bi_sector = SB_SECTOR; + __bio_add_page(bio, virt_to_page(out), SB_SIZE, + offset_in_page(out)); + + out->offset = cpu_to_le64(sb->offset); + + memcpy(out->uuid, sb->uuid, 16); + memcpy(out->set_uuid, sb->set_uuid, 16); + memcpy(out->label, sb->label, SB_LABEL_SIZE); + + out->flags = cpu_to_le64(sb->flags); + out->seq = cpu_to_le64(sb->seq); + + out->last_mount = cpu_to_le32(sb->last_mount); + out->first_bucket = cpu_to_le16(sb->first_bucket); + out->keys = cpu_to_le16(sb->keys); + + for (i = 0; i < sb->keys; i++) + out->d[i] = cpu_to_le64(sb->d[i]); + + if (sb->version >= BCACHE_SB_VERSION_CDEV_WITH_FEATURES) { + out->feature_compat = cpu_to_le64(sb->feature_compat); + out->feature_incompat = cpu_to_le64(sb->feature_incompat); + out->feature_ro_compat = cpu_to_le64(sb->feature_ro_compat); + } + + out->version = cpu_to_le64(sb->version); + out->csum = csum_set(out); + + pr_debug("ver %llu, flags %llu, seq %llu\n", + sb->version, sb->flags, sb->seq); + + submit_bio(bio); +} + +static void bch_write_bdev_super_unlock(struct closure *cl) +{ + struct cached_dev *dc = container_of(cl, struct cached_dev, sb_write); + + up(&dc->sb_write_mutex); +} + +void bch_write_bdev_super(struct cached_dev *dc, struct closure *parent) +{ + struct closure *cl = &dc->sb_write; + struct bio *bio = &dc->sb_bio; + + down(&dc->sb_write_mutex); + closure_init(cl, parent); + + bio_init(bio, dc->sb_bv, 1); + bio_set_dev(bio, dc->bdev); + bio->bi_end_io = write_bdev_super_endio; + bio->bi_private = dc; + + closure_get(cl); + /* I/O request sent to backing device */ + __write_super(&dc->sb, dc->sb_disk, bio); + + closure_return_with_destructor(cl, bch_write_bdev_super_unlock); +} + +static void write_super_endio(struct bio *bio) +{ + struct cache *ca = bio->bi_private; + + /* is_read = 0 */ + bch_count_io_errors(ca, bio->bi_status, 0, + "writing superblock"); + closure_put(&ca->set->sb_write); +} + +static void bcache_write_super_unlock(struct closure *cl) +{ + struct cache_set *c = container_of(cl, struct cache_set, sb_write); + + up(&c->sb_write_mutex); +} + +void bcache_write_super(struct cache_set *c) +{ + struct closure *cl = &c->sb_write; + struct cache *ca = c->cache; + struct bio *bio = &ca->sb_bio; + unsigned int version = BCACHE_SB_VERSION_CDEV_WITH_UUID; + + down(&c->sb_write_mutex); + closure_init(cl, &c->cl); + + ca->sb.seq++; + + if (ca->sb.version < version) + ca->sb.version = version; + + bio_init(bio, ca->sb_bv, 1); + bio_set_dev(bio, ca->bdev); + bio->bi_end_io = write_super_endio; + bio->bi_private = ca; + + closure_get(cl); + __write_super(&ca->sb, ca->sb_disk, bio); + + closure_return_with_destructor(cl, bcache_write_super_unlock); +} + +/* UUID io */ + +static void uuid_endio(struct bio *bio) +{ + struct closure *cl = bio->bi_private; + struct cache_set *c = container_of(cl, struct cache_set, uuid_write); + + cache_set_err_on(bio->bi_status, c, "accessing uuids"); + bch_bbio_free(bio, c); + closure_put(cl); +} + +static void uuid_io_unlock(struct closure *cl) +{ + struct cache_set *c = container_of(cl, struct cache_set, uuid_write); + + up(&c->uuid_write_mutex); +} + +static void uuid_io(struct cache_set *c, int op, unsigned long op_flags, + struct bkey *k, struct closure *parent) +{ + struct closure *cl = &c->uuid_write; + struct uuid_entry *u; + unsigned int i; + char buf[80]; + + BUG_ON(!parent); + down(&c->uuid_write_mutex); + closure_init(cl, parent); + + for (i = 0; i < KEY_PTRS(k); i++) { + struct bio *bio = bch_bbio_alloc(c); + + bio->bi_opf = REQ_SYNC | REQ_META | op_flags; + bio->bi_iter.bi_size = KEY_SIZE(k) << 9; + + bio->bi_end_io = uuid_endio; + bio->bi_private = cl; + bio_set_op_attrs(bio, op, REQ_SYNC|REQ_META|op_flags); + bch_bio_map(bio, c->uuids); + + bch_submit_bbio(bio, c, k, i); + + if (op != REQ_OP_WRITE) + break; + } + + bch_extent_to_text(buf, sizeof(buf), k); + pr_debug("%s UUIDs at %s\n", op == REQ_OP_WRITE ? "wrote" : "read", buf); + + for (u = c->uuids; u < c->uuids + c->nr_uuids; u++) + if (!bch_is_zero(u->uuid, 16)) + pr_debug("Slot %zi: %pU: %s: 1st: %u last: %u inv: %u\n", + u - c->uuids, u->uuid, u->label, + u->first_reg, u->last_reg, u->invalidated); + + closure_return_with_destructor(cl, uuid_io_unlock); +} + +static char *uuid_read(struct cache_set *c, struct jset *j, struct closure *cl) +{ + struct bkey *k = &j->uuid_bucket; + + if (__bch_btree_ptr_invalid(c, k)) + return "bad uuid pointer"; + + bkey_copy(&c->uuid_bucket, k); + uuid_io(c, REQ_OP_READ, 0, k, cl); + + if (j->version < BCACHE_JSET_VERSION_UUIDv1) { + struct uuid_entry_v0 *u0 = (void *) c->uuids; + struct uuid_entry *u1 = (void *) c->uuids; + int i; + + closure_sync(cl); + + /* + * Since the new uuid entry is bigger than the old, we have to + * convert starting at the highest memory address and work down + * in order to do it in place + */ + + for (i = c->nr_uuids - 1; + i >= 0; + --i) { + memcpy(u1[i].uuid, u0[i].uuid, 16); + memcpy(u1[i].label, u0[i].label, 32); + + u1[i].first_reg = u0[i].first_reg; + u1[i].last_reg = u0[i].last_reg; + u1[i].invalidated = u0[i].invalidated; + + u1[i].flags = 0; + u1[i].sectors = 0; + } + } + + return NULL; +} + +static int __uuid_write(struct cache_set *c) +{ + BKEY_PADDED(key) k; + struct closure cl; + struct cache *ca = c->cache; + unsigned int size; + + closure_init_stack(&cl); + lockdep_assert_held(&bch_register_lock); + + if (bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, true)) + return 1; + + size = meta_bucket_pages(&ca->sb) * PAGE_SECTORS; + SET_KEY_SIZE(&k.key, size); + uuid_io(c, REQ_OP_WRITE, 0, &k.key, &cl); + closure_sync(&cl); + + /* Only one bucket used for uuid write */ + atomic_long_add(ca->sb.bucket_size, &ca->meta_sectors_written); + + bkey_copy(&c->uuid_bucket, &k.key); + bkey_put(c, &k.key); + return 0; +} + +int bch_uuid_write(struct cache_set *c) +{ + int ret = __uuid_write(c); + + if (!ret) + bch_journal_meta(c, NULL); + + return ret; +} + +static struct uuid_entry *uuid_find(struct cache_set *c, const char *uuid) +{ + struct uuid_entry *u; + + for (u = c->uuids; + u < c->uuids + c->nr_uuids; u++) + if (!memcmp(u->uuid, uuid, 16)) + return u; + + return NULL; +} + +static struct uuid_entry *uuid_find_empty(struct cache_set *c) +{ + static const char zero_uuid[16] = "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"; + + return uuid_find(c, zero_uuid); +} + +/* + * Bucket priorities/gens: + * + * For each bucket, we store on disk its + * 8 bit gen + * 16 bit priority + * + * See alloc.c for an explanation of the gen. The priority is used to implement + * lru (and in the future other) cache replacement policies; for most purposes + * it's just an opaque integer. + * + * The gens and the priorities don't have a whole lot to do with each other, and + * it's actually the gens that must be written out at specific times - it's no + * big deal if the priorities don't get written, if we lose them we just reuse + * buckets in suboptimal order. + * + * On disk they're stored in a packed array, and in as many buckets are required + * to fit them all. The buckets we use to store them form a list; the journal + * header points to the first bucket, the first bucket points to the second + * bucket, et cetera. + * + * This code is used by the allocation code; periodically (whenever it runs out + * of buckets to allocate from) the allocation code will invalidate some + * buckets, but it can't use those buckets until their new gens are safely on + * disk. + */ + +static void prio_endio(struct bio *bio) +{ + struct cache *ca = bio->bi_private; + + cache_set_err_on(bio->bi_status, ca->set, "accessing priorities"); + bch_bbio_free(bio, ca->set); + closure_put(&ca->prio); +} + +static void prio_io(struct cache *ca, uint64_t bucket, int op, + unsigned long op_flags) +{ + struct closure *cl = &ca->prio; + struct bio *bio = bch_bbio_alloc(ca->set); + + closure_init_stack(cl); + + bio->bi_iter.bi_sector = bucket * ca->sb.bucket_size; + bio_set_dev(bio, ca->bdev); + bio->bi_iter.bi_size = meta_bucket_bytes(&ca->sb); + + bio->bi_end_io = prio_endio; + bio->bi_private = ca; + bio_set_op_attrs(bio, op, REQ_SYNC|REQ_META|op_flags); + bch_bio_map(bio, ca->disk_buckets); + + closure_bio_submit(ca->set, bio, &ca->prio); + closure_sync(cl); +} + +int bch_prio_write(struct cache *ca, bool wait) +{ + int i; + struct bucket *b; + struct closure cl; + + pr_debug("free_prio=%zu, free_none=%zu, free_inc=%zu\n", + fifo_used(&ca->free[RESERVE_PRIO]), + fifo_used(&ca->free[RESERVE_NONE]), + fifo_used(&ca->free_inc)); + + /* + * Pre-check if there are enough free buckets. In the non-blocking + * scenario it's better to fail early rather than starting to allocate + * buckets and do a cleanup later in case of failure. + */ + if (!wait) { + size_t avail = fifo_used(&ca->free[RESERVE_PRIO]) + + fifo_used(&ca->free[RESERVE_NONE]); + if (prio_buckets(ca) > avail) + return -ENOMEM; + } + + closure_init_stack(&cl); + + lockdep_assert_held(&ca->set->bucket_lock); + + ca->disk_buckets->seq++; + + atomic_long_add(ca->sb.bucket_size * prio_buckets(ca), + &ca->meta_sectors_written); + + for (i = prio_buckets(ca) - 1; i >= 0; --i) { + long bucket; + struct prio_set *p = ca->disk_buckets; + struct bucket_disk *d = p->data; + struct bucket_disk *end = d + prios_per_bucket(ca); + + for (b = ca->buckets + i * prios_per_bucket(ca); + b < ca->buckets + ca->sb.nbuckets && d < end; + b++, d++) { + d->prio = cpu_to_le16(b->prio); + d->gen = b->gen; + } + + p->next_bucket = ca->prio_buckets[i + 1]; + p->magic = pset_magic(&ca->sb); + p->csum = bch_crc64(&p->magic, meta_bucket_bytes(&ca->sb) - 8); + + bucket = bch_bucket_alloc(ca, RESERVE_PRIO, wait); + BUG_ON(bucket == -1); + + mutex_unlock(&ca->set->bucket_lock); + prio_io(ca, bucket, REQ_OP_WRITE, 0); + mutex_lock(&ca->set->bucket_lock); + + ca->prio_buckets[i] = bucket; + atomic_dec_bug(&ca->buckets[bucket].pin); + } + + mutex_unlock(&ca->set->bucket_lock); + + bch_journal_meta(ca->set, &cl); + closure_sync(&cl); + + mutex_lock(&ca->set->bucket_lock); + + /* + * Don't want the old priorities to get garbage collected until after we + * finish writing the new ones, and they're journalled + */ + for (i = 0; i < prio_buckets(ca); i++) { + if (ca->prio_last_buckets[i]) + __bch_bucket_free(ca, + &ca->buckets[ca->prio_last_buckets[i]]); + + ca->prio_last_buckets[i] = ca->prio_buckets[i]; + } + return 0; +} + +static int prio_read(struct cache *ca, uint64_t bucket) +{ + struct prio_set *p = ca->disk_buckets; + struct bucket_disk *d = p->data + prios_per_bucket(ca), *end = d; + struct bucket *b; + unsigned int bucket_nr = 0; + int ret = -EIO; + + for (b = ca->buckets; + b < ca->buckets + ca->sb.nbuckets; + b++, d++) { + if (d == end) { + ca->prio_buckets[bucket_nr] = bucket; + ca->prio_last_buckets[bucket_nr] = bucket; + bucket_nr++; + + prio_io(ca, bucket, REQ_OP_READ, 0); + + if (p->csum != + bch_crc64(&p->magic, meta_bucket_bytes(&ca->sb) - 8)) { + pr_warn("bad csum reading priorities\n"); + goto out; + } + + if (p->magic != pset_magic(&ca->sb)) { + pr_warn("bad magic reading priorities\n"); + goto out; + } + + bucket = p->next_bucket; + d = p->data; + } + + b->prio = le16_to_cpu(d->prio); + b->gen = b->last_gc = d->gen; + } + + ret = 0; +out: + return ret; +} + +/* Bcache device */ + +static int open_dev(struct block_device *b, fmode_t mode) +{ + struct bcache_device *d = b->bd_disk->private_data; + + if (test_bit(BCACHE_DEV_CLOSING, &d->flags)) + return -ENXIO; + + closure_get(&d->cl); + return 0; +} + +static void release_dev(struct gendisk *b, fmode_t mode) +{ + struct bcache_device *d = b->private_data; + + closure_put(&d->cl); +} + +static int ioctl_dev(struct block_device *b, fmode_t mode, + unsigned int cmd, unsigned long arg) +{ + struct bcache_device *d = b->bd_disk->private_data; + + return d->ioctl(d, mode, cmd, arg); +} + +static const struct block_device_operations bcache_cached_ops = { + .submit_bio = cached_dev_submit_bio, + .open = open_dev, + .release = release_dev, + .ioctl = ioctl_dev, + .owner = THIS_MODULE, +}; + +static const struct block_device_operations bcache_flash_ops = { + .submit_bio = flash_dev_submit_bio, + .open = open_dev, + .release = release_dev, + .ioctl = ioctl_dev, + .owner = THIS_MODULE, +}; + +void bcache_device_stop(struct bcache_device *d) +{ + if (!test_and_set_bit(BCACHE_DEV_CLOSING, &d->flags)) + /* + * closure_fn set to + * - cached device: cached_dev_flush() + * - flash dev: flash_dev_flush() + */ + closure_queue(&d->cl); +} + +static void bcache_device_unlink(struct bcache_device *d) +{ + lockdep_assert_held(&bch_register_lock); + + if (d->c && !test_and_set_bit(BCACHE_DEV_UNLINK_DONE, &d->flags)) { + struct cache *ca = d->c->cache; + + sysfs_remove_link(&d->c->kobj, d->name); + sysfs_remove_link(&d->kobj, "cache"); + + bd_unlink_disk_holder(ca->bdev, d->disk); + } +} + +static void bcache_device_link(struct bcache_device *d, struct cache_set *c, + const char *name) +{ + struct cache *ca = c->cache; + int ret; + + bd_link_disk_holder(ca->bdev, d->disk); + + snprintf(d->name, BCACHEDEVNAME_SIZE, + "%s%u", name, d->id); + + ret = sysfs_create_link(&d->kobj, &c->kobj, "cache"); + if (ret < 0) + pr_err("Couldn't create device -> cache set symlink\n"); + + ret = sysfs_create_link(&c->kobj, &d->kobj, d->name); + if (ret < 0) + pr_err("Couldn't create cache set -> device symlink\n"); + + clear_bit(BCACHE_DEV_UNLINK_DONE, &d->flags); +} + +static void bcache_device_detach(struct bcache_device *d) +{ + lockdep_assert_held(&bch_register_lock); + + atomic_dec(&d->c->attached_dev_nr); + + if (test_bit(BCACHE_DEV_DETACHING, &d->flags)) { + struct uuid_entry *u = d->c->uuids + d->id; + + SET_UUID_FLASH_ONLY(u, 0); + memcpy(u->uuid, invalid_uuid, 16); + u->invalidated = cpu_to_le32((u32)ktime_get_real_seconds()); + bch_uuid_write(d->c); + } + + bcache_device_unlink(d); + + d->c->devices[d->id] = NULL; + closure_put(&d->c->caching); + d->c = NULL; +} + +static void bcache_device_attach(struct bcache_device *d, struct cache_set *c, + unsigned int id) +{ + d->id = id; + d->c = c; + c->devices[id] = d; + + if (id >= c->devices_max_used) + c->devices_max_used = id + 1; + + closure_get(&c->caching); +} + +static inline int first_minor_to_idx(int first_minor) +{ + return (first_minor/BCACHE_MINORS); +} + +static inline int idx_to_first_minor(int idx) +{ + return (idx * BCACHE_MINORS); +} + +static void bcache_device_free(struct bcache_device *d) +{ + struct gendisk *disk = d->disk; + + lockdep_assert_held(&bch_register_lock); + + if (disk) + pr_info("%s stopped\n", disk->disk_name); + else + pr_err("bcache device (NULL gendisk) stopped\n"); + + if (d->c) + bcache_device_detach(d); + + if (disk) { + bool disk_added = (disk->flags & GENHD_FL_UP) != 0; + + if (disk_added) + del_gendisk(disk); + + if (disk->queue) + blk_cleanup_queue(disk->queue); + + ida_simple_remove(&bcache_device_idx, + first_minor_to_idx(disk->first_minor)); + if (disk_added) + put_disk(disk); + } + + bioset_exit(&d->bio_split); + kvfree(d->full_dirty_stripes); + kvfree(d->stripe_sectors_dirty); + + closure_debug_destroy(&d->cl); +} + +static int bcache_device_init(struct bcache_device *d, unsigned int block_size, + sector_t sectors, struct block_device *cached_bdev, + const struct block_device_operations *ops) +{ + struct request_queue *q; + const size_t max_stripes = min_t(size_t, INT_MAX, + SIZE_MAX / sizeof(atomic_t)); + uint64_t n; + int idx; + + if (!d->stripe_size) + d->stripe_size = 1 << 31; + else if (d->stripe_size < BCH_MIN_STRIPE_SZ) + d->stripe_size = roundup(BCH_MIN_STRIPE_SZ, d->stripe_size); + + n = DIV_ROUND_UP_ULL(sectors, d->stripe_size); + if (!n || n > max_stripes) { + pr_err("nr_stripes too large or invalid: %llu (start sector beyond end of disk?)\n", + n); + return -ENOMEM; + } + d->nr_stripes = n; + + n = d->nr_stripes * sizeof(atomic_t); + d->stripe_sectors_dirty = kvzalloc(n, GFP_KERNEL); + if (!d->stripe_sectors_dirty) + return -ENOMEM; + + n = BITS_TO_LONGS(d->nr_stripes) * sizeof(unsigned long); + d->full_dirty_stripes = kvzalloc(n, GFP_KERNEL); + if (!d->full_dirty_stripes) + goto out_free_stripe_sectors_dirty; + + idx = ida_simple_get(&bcache_device_idx, 0, + BCACHE_DEVICE_IDX_MAX, GFP_KERNEL); + if (idx < 0) + goto out_free_full_dirty_stripes; + + if (bioset_init(&d->bio_split, 4, offsetof(struct bbio, bio), + BIOSET_NEED_BVECS|BIOSET_NEED_RESCUER)) + goto out_ida_remove; + + d->disk = alloc_disk(BCACHE_MINORS); + if (!d->disk) + goto out_bioset_exit; + + set_capacity(d->disk, sectors); + snprintf(d->disk->disk_name, DISK_NAME_LEN, "bcache%i", idx); + + d->disk->major = bcache_major; + d->disk->first_minor = idx_to_first_minor(idx); + d->disk->fops = ops; + d->disk->private_data = d; + + q = blk_alloc_queue(NUMA_NO_NODE); + if (!q) + return -ENOMEM; + + d->disk->queue = q; + q->limits.max_hw_sectors = UINT_MAX; + q->limits.max_sectors = UINT_MAX; + q->limits.max_segment_size = UINT_MAX; + q->limits.max_segments = BIO_MAX_PAGES; + blk_queue_max_discard_sectors(q, UINT_MAX); + q->limits.discard_granularity = 512; + q->limits.io_min = block_size; + q->limits.logical_block_size = block_size; + q->limits.physical_block_size = block_size; + + if (q->limits.logical_block_size > PAGE_SIZE && cached_bdev) { + /* + * This should only happen with BCACHE_SB_VERSION_BDEV. + * Block/page size is checked for BCACHE_SB_VERSION_CDEV. + */ + pr_info("%s: sb/logical block size (%u) greater than page size (%lu) falling back to device logical block size (%u)\n", + d->disk->disk_name, q->limits.logical_block_size, + PAGE_SIZE, bdev_logical_block_size(cached_bdev)); + + /* This also adjusts physical block size/min io size if needed */ + blk_queue_logical_block_size(q, bdev_logical_block_size(cached_bdev)); + } + + blk_queue_flag_set(QUEUE_FLAG_NONROT, d->disk->queue); + blk_queue_flag_clear(QUEUE_FLAG_ADD_RANDOM, d->disk->queue); + blk_queue_flag_set(QUEUE_FLAG_DISCARD, d->disk->queue); + + blk_queue_write_cache(q, true, true); + + return 0; + +out_bioset_exit: + bioset_exit(&d->bio_split); +out_ida_remove: + ida_simple_remove(&bcache_device_idx, idx); +out_free_full_dirty_stripes: + kvfree(d->full_dirty_stripes); +out_free_stripe_sectors_dirty: + kvfree(d->stripe_sectors_dirty); + return -ENOMEM; + +} + +/* Cached device */ + +static void calc_cached_dev_sectors(struct cache_set *c) +{ + uint64_t sectors = 0; + struct cached_dev *dc; + + list_for_each_entry(dc, &c->cached_devs, list) + sectors += bdev_sectors(dc->bdev); + + c->cached_dev_sectors = sectors; +} + +#define BACKING_DEV_OFFLINE_TIMEOUT 5 +static int cached_dev_status_update(void *arg) +{ + struct cached_dev *dc = arg; + struct request_queue *q; + + /* + * If this delayed worker is stopping outside, directly quit here. + * dc->io_disable might be set via sysfs interface, so check it + * here too. + */ + while (!kthread_should_stop() && !dc->io_disable) { + q = bdev_get_queue(dc->bdev); + if (blk_queue_dying(q)) + dc->offline_seconds++; + else + dc->offline_seconds = 0; + + if (dc->offline_seconds >= BACKING_DEV_OFFLINE_TIMEOUT) { + pr_err("%s: device offline for %d seconds\n", + dc->backing_dev_name, + BACKING_DEV_OFFLINE_TIMEOUT); + pr_err("%s: disable I/O request due to backing device offline\n", + dc->disk.name); + dc->io_disable = true; + /* let others know earlier that io_disable is true */ + smp_mb(); + bcache_device_stop(&dc->disk); + break; + } + schedule_timeout_interruptible(HZ); + } + + wait_for_kthread_stop(); + return 0; +} + + +int bch_cached_dev_run(struct cached_dev *dc) +{ + struct bcache_device *d = &dc->disk; + char *buf = kmemdup_nul(dc->sb.label, SB_LABEL_SIZE, GFP_KERNEL); + char *env[] = { + "DRIVER=bcache", + kasprintf(GFP_KERNEL, "CACHED_UUID=%pU", dc->sb.uuid), + kasprintf(GFP_KERNEL, "CACHED_LABEL=%s", buf ? : ""), + NULL, + }; + + if (dc->io_disable) { + pr_err("I/O disabled on cached dev %s\n", + dc->backing_dev_name); + kfree(env[1]); + kfree(env[2]); + kfree(buf); + return -EIO; + } + + if (atomic_xchg(&dc->running, 1)) { + kfree(env[1]); + kfree(env[2]); + kfree(buf); + pr_info("cached dev %s is running already\n", + dc->backing_dev_name); + return -EBUSY; + } + + if (!d->c && + BDEV_STATE(&dc->sb) != BDEV_STATE_NONE) { + struct closure cl; + + closure_init_stack(&cl); + + SET_BDEV_STATE(&dc->sb, BDEV_STATE_STALE); + bch_write_bdev_super(dc, &cl); + closure_sync(&cl); + } + + add_disk(d->disk); + bd_link_disk_holder(dc->bdev, dc->disk.disk); + /* + * won't show up in the uevent file, use udevadm monitor -e instead + * only class / kset properties are persistent + */ + kobject_uevent_env(&disk_to_dev(d->disk)->kobj, KOBJ_CHANGE, env); + kfree(env[1]); + kfree(env[2]); + kfree(buf); + + if (sysfs_create_link(&d->kobj, &disk_to_dev(d->disk)->kobj, "dev") || + sysfs_create_link(&disk_to_dev(d->disk)->kobj, + &d->kobj, "bcache")) { + pr_err("Couldn't create bcache dev <-> disk sysfs symlinks\n"); + return -ENOMEM; + } + + dc->status_update_thread = kthread_run(cached_dev_status_update, + dc, "bcache_status_update"); + if (IS_ERR(dc->status_update_thread)) { + pr_warn("failed to create bcache_status_update kthread, continue to run without monitoring backing device status\n"); + } + + return 0; +} + +/* + * If BCACHE_DEV_RATE_DW_RUNNING is set, it means routine of the delayed + * work dc->writeback_rate_update is running. Wait until the routine + * quits (BCACHE_DEV_RATE_DW_RUNNING is clear), then continue to + * cancel it. If BCACHE_DEV_RATE_DW_RUNNING is not clear after time_out + * seconds, give up waiting here and continue to cancel it too. + */ +static void cancel_writeback_rate_update_dwork(struct cached_dev *dc) +{ + int time_out = WRITEBACK_RATE_UPDATE_SECS_MAX * HZ; + + do { + if (!test_bit(BCACHE_DEV_RATE_DW_RUNNING, + &dc->disk.flags)) + break; + time_out--; + schedule_timeout_interruptible(1); + } while (time_out > 0); + + if (time_out == 0) + pr_warn("give up waiting for dc->writeback_write_update to quit\n"); + + cancel_delayed_work_sync(&dc->writeback_rate_update); +} + +static void cached_dev_detach_finish(struct work_struct *w) +{ + struct cached_dev *dc = container_of(w, struct cached_dev, detach); + struct closure cl; + + closure_init_stack(&cl); + + BUG_ON(!test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags)); + BUG_ON(refcount_read(&dc->count)); + + + if (test_and_clear_bit(BCACHE_DEV_WB_RUNNING, &dc->disk.flags)) + cancel_writeback_rate_update_dwork(dc); + + if (!IS_ERR_OR_NULL(dc->writeback_thread)) { + kthread_stop(dc->writeback_thread); + dc->writeback_thread = NULL; + } + + memset(&dc->sb.set_uuid, 0, 16); + SET_BDEV_STATE(&dc->sb, BDEV_STATE_NONE); + + bch_write_bdev_super(dc, &cl); + closure_sync(&cl); + + mutex_lock(&bch_register_lock); + + calc_cached_dev_sectors(dc->disk.c); + bcache_device_detach(&dc->disk); + list_move(&dc->list, &uncached_devices); + + clear_bit(BCACHE_DEV_DETACHING, &dc->disk.flags); + clear_bit(BCACHE_DEV_UNLINK_DONE, &dc->disk.flags); + + mutex_unlock(&bch_register_lock); + + pr_info("Caching disabled for %s\n", dc->backing_dev_name); + + /* Drop ref we took in cached_dev_detach() */ + closure_put(&dc->disk.cl); +} + +void bch_cached_dev_detach(struct cached_dev *dc) +{ + lockdep_assert_held(&bch_register_lock); + + if (test_bit(BCACHE_DEV_CLOSING, &dc->disk.flags)) + return; + + if (test_and_set_bit(BCACHE_DEV_DETACHING, &dc->disk.flags)) + return; + + /* + * Block the device from being closed and freed until we're finished + * detaching + */ + closure_get(&dc->disk.cl); + + bch_writeback_queue(dc); + + cached_dev_put(dc); +} + +int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c, + uint8_t *set_uuid) +{ + uint32_t rtime = cpu_to_le32((u32)ktime_get_real_seconds()); + struct uuid_entry *u; + struct cached_dev *exist_dc, *t; + int ret = 0; + + if ((set_uuid && memcmp(set_uuid, c->set_uuid, 16)) || + (!set_uuid && memcmp(dc->sb.set_uuid, c->set_uuid, 16))) + return -ENOENT; + + if (dc->disk.c) { + pr_err("Can't attach %s: already attached\n", + dc->backing_dev_name); + return -EINVAL; + } + + if (test_bit(CACHE_SET_STOPPING, &c->flags)) { + pr_err("Can't attach %s: shutting down\n", + dc->backing_dev_name); + return -EINVAL; + } + + if (dc->sb.block_size < c->cache->sb.block_size) { + /* Will die */ + pr_err("Couldn't attach %s: block size less than set's block size\n", + dc->backing_dev_name); + return -EINVAL; + } + + /* Check whether already attached */ + list_for_each_entry_safe(exist_dc, t, &c->cached_devs, list) { + if (!memcmp(dc->sb.uuid, exist_dc->sb.uuid, 16)) { + pr_err("Tried to attach %s but duplicate UUID already attached\n", + dc->backing_dev_name); + + return -EINVAL; + } + } + + u = uuid_find(c, dc->sb.uuid); + + if (u && + (BDEV_STATE(&dc->sb) == BDEV_STATE_STALE || + BDEV_STATE(&dc->sb) == BDEV_STATE_NONE)) { + memcpy(u->uuid, invalid_uuid, 16); + u->invalidated = cpu_to_le32((u32)ktime_get_real_seconds()); + u = NULL; + } + + if (!u) { + if (BDEV_STATE(&dc->sb) == BDEV_STATE_DIRTY) { + pr_err("Couldn't find uuid for %s in set\n", + dc->backing_dev_name); + return -ENOENT; + } + + u = uuid_find_empty(c); + if (!u) { + pr_err("Not caching %s, no room for UUID\n", + dc->backing_dev_name); + return -EINVAL; + } + } + + /* + * Deadlocks since we're called via sysfs... + * sysfs_remove_file(&dc->kobj, &sysfs_attach); + */ + + if (bch_is_zero(u->uuid, 16)) { + struct closure cl; + + closure_init_stack(&cl); + + memcpy(u->uuid, dc->sb.uuid, 16); + memcpy(u->label, dc->sb.label, SB_LABEL_SIZE); + u->first_reg = u->last_reg = rtime; + bch_uuid_write(c); + + memcpy(dc->sb.set_uuid, c->set_uuid, 16); + SET_BDEV_STATE(&dc->sb, BDEV_STATE_CLEAN); + + bch_write_bdev_super(dc, &cl); + closure_sync(&cl); + } else { + u->last_reg = rtime; + bch_uuid_write(c); + } + + bcache_device_attach(&dc->disk, c, u - c->uuids); + list_move(&dc->list, &c->cached_devs); + calc_cached_dev_sectors(c); + + /* + * dc->c must be set before dc->count != 0 - paired with the mb in + * cached_dev_get() + */ + smp_wmb(); + refcount_set(&dc->count, 1); + + /* Block writeback thread, but spawn it */ + down_write(&dc->writeback_lock); + if (bch_cached_dev_writeback_start(dc)) { + up_write(&dc->writeback_lock); + pr_err("Couldn't start writeback facilities for %s\n", + dc->disk.disk->disk_name); + return -ENOMEM; + } + + if (BDEV_STATE(&dc->sb) == BDEV_STATE_DIRTY) { + atomic_set(&dc->has_dirty, 1); + bch_writeback_queue(dc); + } + + bch_sectors_dirty_init(&dc->disk); + + ret = bch_cached_dev_run(dc); + if (ret && (ret != -EBUSY)) { + up_write(&dc->writeback_lock); + /* + * bch_register_lock is held, bcache_device_stop() is not + * able to be directly called. The kthread and kworker + * created previously in bch_cached_dev_writeback_start() + * have to be stopped manually here. + */ + kthread_stop(dc->writeback_thread); + cancel_writeback_rate_update_dwork(dc); + pr_err("Couldn't run cached device %s\n", + dc->backing_dev_name); + return ret; + } + + bcache_device_link(&dc->disk, c, "bdev"); + atomic_inc(&c->attached_dev_nr); + + if (bch_has_feature_obso_large_bucket(&(c->cache->sb))) { + pr_err("The obsoleted large bucket layout is unsupported, set the bcache device into read-only\n"); + pr_err("Please update to the latest bcache-tools to create the cache device\n"); + set_disk_ro(dc->disk.disk, 1); + } + + /* Allow the writeback thread to proceed */ + up_write(&dc->writeback_lock); + + pr_info("Caching %s as %s on set %pU\n", + dc->backing_dev_name, + dc->disk.disk->disk_name, + dc->disk.c->set_uuid); + return 0; +} + +/* when dc->disk.kobj released */ +void bch_cached_dev_release(struct kobject *kobj) +{ + struct cached_dev *dc = container_of(kobj, struct cached_dev, + disk.kobj); + kfree(dc); + module_put(THIS_MODULE); +} + +static void cached_dev_free(struct closure *cl) +{ + struct cached_dev *dc = container_of(cl, struct cached_dev, disk.cl); + + if (test_and_clear_bit(BCACHE_DEV_WB_RUNNING, &dc->disk.flags)) + cancel_writeback_rate_update_dwork(dc); + + if (!IS_ERR_OR_NULL(dc->writeback_thread)) + kthread_stop(dc->writeback_thread); + if (!IS_ERR_OR_NULL(dc->status_update_thread)) + kthread_stop(dc->status_update_thread); + + mutex_lock(&bch_register_lock); + + if (atomic_read(&dc->running)) + bd_unlink_disk_holder(dc->bdev, dc->disk.disk); + bcache_device_free(&dc->disk); + list_del(&dc->list); + + mutex_unlock(&bch_register_lock); + + if (dc->sb_disk) + put_page(virt_to_page(dc->sb_disk)); + + if (!IS_ERR_OR_NULL(dc->bdev)) + blkdev_put(dc->bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL); + + wake_up(&unregister_wait); + + kobject_put(&dc->disk.kobj); +} + +static void cached_dev_flush(struct closure *cl) +{ + struct cached_dev *dc = container_of(cl, struct cached_dev, disk.cl); + struct bcache_device *d = &dc->disk; + + mutex_lock(&bch_register_lock); + bcache_device_unlink(d); + mutex_unlock(&bch_register_lock); + + bch_cache_accounting_destroy(&dc->accounting); + kobject_del(&d->kobj); + + continue_at(cl, cached_dev_free, system_wq); +} + +static int cached_dev_init(struct cached_dev *dc, unsigned int block_size) +{ + int ret; + struct io *io; + struct request_queue *q = bdev_get_queue(dc->bdev); + + __module_get(THIS_MODULE); + INIT_LIST_HEAD(&dc->list); + closure_init(&dc->disk.cl, NULL); + set_closure_fn(&dc->disk.cl, cached_dev_flush, system_wq); + kobject_init(&dc->disk.kobj, &bch_cached_dev_ktype); + INIT_WORK(&dc->detach, cached_dev_detach_finish); + sema_init(&dc->sb_write_mutex, 1); + INIT_LIST_HEAD(&dc->io_lru); + spin_lock_init(&dc->io_lock); + bch_cache_accounting_init(&dc->accounting, &dc->disk.cl); + + dc->sequential_cutoff = 4 << 20; + + for (io = dc->io; io < dc->io + RECENT_IO; io++) { + list_add(&io->lru, &dc->io_lru); + hlist_add_head(&io->hash, dc->io_hash + RECENT_IO); + } + + dc->disk.stripe_size = q->limits.io_opt >> 9; + + if (dc->disk.stripe_size) + dc->partial_stripes_expensive = + q->limits.raid_partial_stripes_expensive; + + ret = bcache_device_init(&dc->disk, block_size, + dc->bdev->bd_part->nr_sects - dc->sb.data_offset, + dc->bdev, &bcache_cached_ops); + if (ret) + return ret; + + blk_queue_io_opt(dc->disk.disk->queue, + max(queue_io_opt(dc->disk.disk->queue), queue_io_opt(q))); + + atomic_set(&dc->io_errors, 0); + dc->io_disable = false; + dc->error_limit = DEFAULT_CACHED_DEV_ERROR_LIMIT; + /* default to auto */ + dc->stop_when_cache_set_failed = BCH_CACHED_DEV_STOP_AUTO; + + bch_cached_dev_request_init(dc); + bch_cached_dev_writeback_init(dc); + return 0; +} + +/* Cached device - bcache superblock */ + +static int register_bdev(struct cache_sb *sb, struct cache_sb_disk *sb_disk, + struct block_device *bdev, + struct cached_dev *dc) +{ + const char *err = "cannot allocate memory"; + struct cache_set *c; + int ret = -ENOMEM; + + bdevname(bdev, dc->backing_dev_name); + memcpy(&dc->sb, sb, sizeof(struct cache_sb)); + dc->bdev = bdev; + dc->bdev->bd_holder = dc; + dc->sb_disk = sb_disk; + + if (cached_dev_init(dc, sb->block_size << 9)) + goto err; + + err = "error creating kobject"; + if (kobject_add(&dc->disk.kobj, &part_to_dev(bdev->bd_part)->kobj, + "bcache")) + goto err; + if (bch_cache_accounting_add_kobjs(&dc->accounting, &dc->disk.kobj)) + goto err; + + pr_info("registered backing device %s\n", dc->backing_dev_name); + + list_add(&dc->list, &uncached_devices); + /* attach to a matched cache set if it exists */ + list_for_each_entry(c, &bch_cache_sets, list) + bch_cached_dev_attach(dc, c, NULL); + + if (BDEV_STATE(&dc->sb) == BDEV_STATE_NONE || + BDEV_STATE(&dc->sb) == BDEV_STATE_STALE) { + err = "failed to run cached device"; + ret = bch_cached_dev_run(dc); + if (ret) + goto err; + } + + return 0; +err: + pr_notice("error %s: %s\n", dc->backing_dev_name, err); + bcache_device_stop(&dc->disk); + return ret; +} + +/* Flash only volumes */ + +/* When d->kobj released */ +void bch_flash_dev_release(struct kobject *kobj) +{ + struct bcache_device *d = container_of(kobj, struct bcache_device, + kobj); + kfree(d); +} + +static void flash_dev_free(struct closure *cl) +{ + struct bcache_device *d = container_of(cl, struct bcache_device, cl); + + mutex_lock(&bch_register_lock); + atomic_long_sub(bcache_dev_sectors_dirty(d), + &d->c->flash_dev_dirty_sectors); + bcache_device_free(d); + mutex_unlock(&bch_register_lock); + kobject_put(&d->kobj); +} + +static void flash_dev_flush(struct closure *cl) +{ + struct bcache_device *d = container_of(cl, struct bcache_device, cl); + + mutex_lock(&bch_register_lock); + bcache_device_unlink(d); + mutex_unlock(&bch_register_lock); + kobject_del(&d->kobj); + continue_at(cl, flash_dev_free, system_wq); +} + +static int flash_dev_run(struct cache_set *c, struct uuid_entry *u) +{ + struct bcache_device *d = kzalloc(sizeof(struct bcache_device), + GFP_KERNEL); + if (!d) + return -ENOMEM; + + closure_init(&d->cl, NULL); + set_closure_fn(&d->cl, flash_dev_flush, system_wq); + + kobject_init(&d->kobj, &bch_flash_dev_ktype); + + if (bcache_device_init(d, block_bytes(c->cache), u->sectors, + NULL, &bcache_flash_ops)) + goto err; + + bcache_device_attach(d, c, u - c->uuids); + bch_sectors_dirty_init(d); + bch_flash_dev_request_init(d); + add_disk(d->disk); + + if (kobject_add(&d->kobj, &disk_to_dev(d->disk)->kobj, "bcache")) + goto err; + + bcache_device_link(d, c, "volume"); + + if (bch_has_feature_obso_large_bucket(&c->cache->sb)) { + pr_err("The obsoleted large bucket layout is unsupported, set the bcache device into read-only\n"); + pr_err("Please update to the latest bcache-tools to create the cache device\n"); + set_disk_ro(d->disk, 1); + } + + return 0; +err: + kobject_put(&d->kobj); + return -ENOMEM; +} + +static int flash_devs_run(struct cache_set *c) +{ + int ret = 0; + struct uuid_entry *u; + + for (u = c->uuids; + u < c->uuids + c->nr_uuids && !ret; + u++) + if (UUID_FLASH_ONLY(u)) + ret = flash_dev_run(c, u); + + return ret; +} + +int bch_flash_dev_create(struct cache_set *c, uint64_t size) +{ + struct uuid_entry *u; + + if (test_bit(CACHE_SET_STOPPING, &c->flags)) + return -EINTR; + + if (!test_bit(CACHE_SET_RUNNING, &c->flags)) + return -EPERM; + + u = uuid_find_empty(c); + if (!u) { + pr_err("Can't create volume, no room for UUID\n"); + return -EINVAL; + } + + get_random_bytes(u->uuid, 16); + memset(u->label, 0, 32); + u->first_reg = u->last_reg = cpu_to_le32((u32)ktime_get_real_seconds()); + + SET_UUID_FLASH_ONLY(u, 1); + u->sectors = size >> 9; + + bch_uuid_write(c); + + return flash_dev_run(c, u); +} + +bool bch_cached_dev_error(struct cached_dev *dc) +{ + if (!dc || test_bit(BCACHE_DEV_CLOSING, &dc->disk.flags)) + return false; + + dc->io_disable = true; + /* make others know io_disable is true earlier */ + smp_mb(); + + pr_err("stop %s: too many IO errors on backing device %s\n", + dc->disk.disk->disk_name, dc->backing_dev_name); + + bcache_device_stop(&dc->disk); + return true; +} + +/* Cache set */ + +__printf(2, 3) +bool bch_cache_set_error(struct cache_set *c, const char *fmt, ...) +{ + struct va_format vaf; + va_list args; + + if (c->on_error != ON_ERROR_PANIC && + test_bit(CACHE_SET_STOPPING, &c->flags)) + return false; + + if (test_and_set_bit(CACHE_SET_IO_DISABLE, &c->flags)) + pr_info("CACHE_SET_IO_DISABLE already set\n"); + + /* + * XXX: we can be called from atomic context + * acquire_console_sem(); + */ + + va_start(args, fmt); + + vaf.fmt = fmt; + vaf.va = &args; + + pr_err("error on %pU: %pV, disabling caching\n", + c->set_uuid, &vaf); + + va_end(args); + + if (c->on_error == ON_ERROR_PANIC) + panic("panic forced after error\n"); + + bch_cache_set_unregister(c); + return true; +} + +/* When c->kobj released */ +void bch_cache_set_release(struct kobject *kobj) +{ + struct cache_set *c = container_of(kobj, struct cache_set, kobj); + + kfree(c); + module_put(THIS_MODULE); +} + +static void cache_set_free(struct closure *cl) +{ + struct cache_set *c = container_of(cl, struct cache_set, cl); + struct cache *ca; + + debugfs_remove(c->debug); + + bch_open_buckets_free(c); + bch_btree_cache_free(c); + bch_journal_free(c); + + mutex_lock(&bch_register_lock); + bch_bset_sort_state_free(&c->sort); + free_pages((unsigned long) c->uuids, ilog2(meta_bucket_pages(&c->cache->sb))); + + ca = c->cache; + if (ca) { + ca->set = NULL; + c->cache = NULL; + kobject_put(&ca->kobj); + } + + + if (c->moving_gc_wq) + destroy_workqueue(c->moving_gc_wq); + bioset_exit(&c->bio_split); + mempool_exit(&c->fill_iter); + mempool_exit(&c->bio_meta); + mempool_exit(&c->search); + kfree(c->devices); + + list_del(&c->list); + mutex_unlock(&bch_register_lock); + + pr_info("Cache set %pU unregistered\n", c->set_uuid); + wake_up(&unregister_wait); + + closure_debug_destroy(&c->cl); + kobject_put(&c->kobj); +} + +static void cache_set_flush(struct closure *cl) +{ + struct cache_set *c = container_of(cl, struct cache_set, caching); + struct cache *ca = c->cache; + struct btree *b; + + bch_cache_accounting_destroy(&c->accounting); + + kobject_put(&c->internal); + kobject_del(&c->kobj); + + if (!IS_ERR_OR_NULL(c->gc_thread)) + kthread_stop(c->gc_thread); + + if (!IS_ERR(c->root)) + list_add(&c->root->list, &c->btree_cache); + + /* + * Avoid flushing cached nodes if cache set is retiring + * due to too many I/O errors detected. + */ + if (!test_bit(CACHE_SET_IO_DISABLE, &c->flags)) + list_for_each_entry(b, &c->btree_cache, list) { + mutex_lock(&b->write_lock); + if (btree_node_dirty(b)) + __bch_btree_node_write(b, NULL); + mutex_unlock(&b->write_lock); + } + + if (ca->alloc_thread) + kthread_stop(ca->alloc_thread); + + if (c->journal.cur) { + cancel_delayed_work_sync(&c->journal.work); + /* flush last journal entry if needed */ + c->journal.work.work.func(&c->journal.work.work); + } + + closure_return(cl); +} + +/* + * This function is only called when CACHE_SET_IO_DISABLE is set, which means + * cache set is unregistering due to too many I/O errors. In this condition, + * the bcache device might be stopped, it depends on stop_when_cache_set_failed + * value and whether the broken cache has dirty data: + * + * dc->stop_when_cache_set_failed dc->has_dirty stop bcache device + * BCH_CACHED_STOP_AUTO 0 NO + * BCH_CACHED_STOP_AUTO 1 YES + * BCH_CACHED_DEV_STOP_ALWAYS 0 YES + * BCH_CACHED_DEV_STOP_ALWAYS 1 YES + * + * The expected behavior is, if stop_when_cache_set_failed is configured to + * "auto" via sysfs interface, the bcache device will not be stopped if the + * backing device is clean on the broken cache device. + */ +static void conditional_stop_bcache_device(struct cache_set *c, + struct bcache_device *d, + struct cached_dev *dc) +{ + if (dc->stop_when_cache_set_failed == BCH_CACHED_DEV_STOP_ALWAYS) { + pr_warn("stop_when_cache_set_failed of %s is \"always\", stop it for failed cache set %pU.\n", + d->disk->disk_name, c->set_uuid); + bcache_device_stop(d); + } else if (atomic_read(&dc->has_dirty)) { + /* + * dc->stop_when_cache_set_failed == BCH_CACHED_STOP_AUTO + * and dc->has_dirty == 1 + */ + pr_warn("stop_when_cache_set_failed of %s is \"auto\" and cache is dirty, stop it to avoid potential data corruption.\n", + d->disk->disk_name); + /* + * There might be a small time gap that cache set is + * released but bcache device is not. Inside this time + * gap, regular I/O requests will directly go into + * backing device as no cache set attached to. This + * behavior may also introduce potential inconsistence + * data in writeback mode while cache is dirty. + * Therefore before calling bcache_device_stop() due + * to a broken cache device, dc->io_disable should be + * explicitly set to true. + */ + dc->io_disable = true; + /* make others know io_disable is true earlier */ + smp_mb(); + bcache_device_stop(d); + } else { + /* + * dc->stop_when_cache_set_failed == BCH_CACHED_STOP_AUTO + * and dc->has_dirty == 0 + */ + pr_warn("stop_when_cache_set_failed of %s is \"auto\" and cache is clean, keep it alive.\n", + d->disk->disk_name); + } +} + +static void __cache_set_unregister(struct closure *cl) +{ + struct cache_set *c = container_of(cl, struct cache_set, caching); + struct cached_dev *dc; + struct bcache_device *d; + size_t i; + + mutex_lock(&bch_register_lock); + + for (i = 0; i < c->devices_max_used; i++) { + d = c->devices[i]; + if (!d) + continue; + + if (!UUID_FLASH_ONLY(&c->uuids[i]) && + test_bit(CACHE_SET_UNREGISTERING, &c->flags)) { + dc = container_of(d, struct cached_dev, disk); + bch_cached_dev_detach(dc); + if (test_bit(CACHE_SET_IO_DISABLE, &c->flags)) + conditional_stop_bcache_device(c, d, dc); + } else { + bcache_device_stop(d); + } + } + + mutex_unlock(&bch_register_lock); + + continue_at(cl, cache_set_flush, system_wq); +} + +void bch_cache_set_stop(struct cache_set *c) +{ + if (!test_and_set_bit(CACHE_SET_STOPPING, &c->flags)) + /* closure_fn set to __cache_set_unregister() */ + closure_queue(&c->caching); +} + +void bch_cache_set_unregister(struct cache_set *c) +{ + set_bit(CACHE_SET_UNREGISTERING, &c->flags); + bch_cache_set_stop(c); +} + +#define alloc_meta_bucket_pages(gfp, sb) \ + ((void *) __get_free_pages(__GFP_ZERO|__GFP_COMP|gfp, ilog2(meta_bucket_pages(sb)))) + +struct cache_set *bch_cache_set_alloc(struct cache_sb *sb) +{ + int iter_size; + struct cache *ca = container_of(sb, struct cache, sb); + struct cache_set *c = kzalloc(sizeof(struct cache_set), GFP_KERNEL); + + if (!c) + return NULL; + + __module_get(THIS_MODULE); + closure_init(&c->cl, NULL); + set_closure_fn(&c->cl, cache_set_free, system_wq); + + closure_init(&c->caching, &c->cl); + set_closure_fn(&c->caching, __cache_set_unregister, system_wq); + + /* Maybe create continue_at_noreturn() and use it here? */ + closure_set_stopped(&c->cl); + closure_put(&c->cl); + + kobject_init(&c->kobj, &bch_cache_set_ktype); + kobject_init(&c->internal, &bch_cache_set_internal_ktype); + + bch_cache_accounting_init(&c->accounting, &c->cl); + + memcpy(c->set_uuid, sb->set_uuid, 16); + + c->cache = ca; + c->cache->set = c; + c->bucket_bits = ilog2(sb->bucket_size); + c->block_bits = ilog2(sb->block_size); + c->nr_uuids = meta_bucket_bytes(sb) / sizeof(struct uuid_entry); + c->devices_max_used = 0; + atomic_set(&c->attached_dev_nr, 0); + c->btree_pages = meta_bucket_pages(sb); + if (c->btree_pages > BTREE_MAX_PAGES) + c->btree_pages = max_t(int, c->btree_pages / 4, + BTREE_MAX_PAGES); + + sema_init(&c->sb_write_mutex, 1); + mutex_init(&c->bucket_lock); + init_waitqueue_head(&c->btree_cache_wait); + spin_lock_init(&c->btree_cannibalize_lock); + init_waitqueue_head(&c->bucket_wait); + init_waitqueue_head(&c->gc_wait); + sema_init(&c->uuid_write_mutex, 1); + + spin_lock_init(&c->btree_gc_time.lock); + spin_lock_init(&c->btree_split_time.lock); + spin_lock_init(&c->btree_read_time.lock); + + bch_moving_init_cache_set(c); + + INIT_LIST_HEAD(&c->list); + INIT_LIST_HEAD(&c->cached_devs); + INIT_LIST_HEAD(&c->btree_cache); + INIT_LIST_HEAD(&c->btree_cache_freeable); + INIT_LIST_HEAD(&c->btree_cache_freed); + INIT_LIST_HEAD(&c->data_buckets); + + iter_size = ((meta_bucket_pages(sb) * PAGE_SECTORS) / sb->block_size + 1) * + sizeof(struct btree_iter_set); + + c->devices = kcalloc(c->nr_uuids, sizeof(void *), GFP_KERNEL); + if (!c->devices) + goto err; + + if (mempool_init_slab_pool(&c->search, 32, bch_search_cache)) + goto err; + + if (mempool_init_kmalloc_pool(&c->bio_meta, 2, + sizeof(struct bbio) + + sizeof(struct bio_vec) * meta_bucket_pages(sb))) + goto err; + + if (mempool_init_kmalloc_pool(&c->fill_iter, 1, iter_size)) + goto err; + + if (bioset_init(&c->bio_split, 4, offsetof(struct bbio, bio), + BIOSET_NEED_BVECS|BIOSET_NEED_RESCUER)) + goto err; + + c->uuids = alloc_meta_bucket_pages(GFP_KERNEL, sb); + if (!c->uuids) + goto err; + + c->moving_gc_wq = alloc_workqueue("bcache_gc", WQ_MEM_RECLAIM, 0); + if (!c->moving_gc_wq) + goto err; + + if (bch_journal_alloc(c)) + goto err; + + if (bch_btree_cache_alloc(c)) + goto err; + + if (bch_open_buckets_alloc(c)) + goto err; + + if (bch_bset_sort_state_init(&c->sort, ilog2(c->btree_pages))) + goto err; + + c->congested_read_threshold_us = 2000; + c->congested_write_threshold_us = 20000; + c->error_limit = DEFAULT_IO_ERROR_LIMIT; + c->idle_max_writeback_rate_enabled = 1; + WARN_ON(test_and_clear_bit(CACHE_SET_IO_DISABLE, &c->flags)); + + return c; +err: + bch_cache_set_unregister(c); + return NULL; +} + +static int run_cache_set(struct cache_set *c) +{ + const char *err = "cannot allocate memory"; + struct cached_dev *dc, *t; + struct cache *ca = c->cache; + struct closure cl; + LIST_HEAD(journal); + struct journal_replay *l; + + closure_init_stack(&cl); + + c->nbuckets = ca->sb.nbuckets; + set_gc_sectors(c); + + if (CACHE_SYNC(&c->cache->sb)) { + struct bkey *k; + struct jset *j; + + err = "cannot allocate memory for journal"; + if (bch_journal_read(c, &journal)) + goto err; + + pr_debug("btree_journal_read() done\n"); + + err = "no journal entries found"; + if (list_empty(&journal)) + goto err; + + j = &list_entry(journal.prev, struct journal_replay, list)->j; + + err = "IO error reading priorities"; + if (prio_read(ca, j->prio_bucket[ca->sb.nr_this_dev])) + goto err; + + /* + * If prio_read() fails it'll call cache_set_error and we'll + * tear everything down right away, but if we perhaps checked + * sooner we could avoid journal replay. + */ + + k = &j->btree_root; + + err = "bad btree root"; + if (__bch_btree_ptr_invalid(c, k)) + goto err; + + err = "error reading btree root"; + c->root = bch_btree_node_get(c, NULL, k, + j->btree_level, + true, NULL); + if (IS_ERR(c->root)) + goto err; + + list_del_init(&c->root->list); + rw_unlock(true, c->root); + + err = uuid_read(c, j, &cl); + if (err) + goto err; + + err = "error in recovery"; + if (bch_btree_check(c)) + goto err; + + bch_journal_mark(c, &journal); + bch_initial_gc_finish(c); + pr_debug("btree_check() done\n"); + + /* + * bcache_journal_next() can't happen sooner, or + * btree_gc_finish() will give spurious errors about last_gc > + * gc_gen - this is a hack but oh well. + */ + bch_journal_next(&c->journal); + + err = "error starting allocator thread"; + if (bch_cache_allocator_start(ca)) + goto err; + + /* + * First place it's safe to allocate: btree_check() and + * btree_gc_finish() have to run before we have buckets to + * allocate, and bch_bucket_alloc_set() might cause a journal + * entry to be written so bcache_journal_next() has to be called + * first. + * + * If the uuids were in the old format we have to rewrite them + * before the next journal entry is written: + */ + if (j->version < BCACHE_JSET_VERSION_UUID) + __uuid_write(c); + + err = "bcache: replay journal failed"; + if (bch_journal_replay(c, &journal)) + goto err; + } else { + unsigned int j; + + pr_notice("invalidating existing data\n"); + ca->sb.keys = clamp_t(int, ca->sb.nbuckets >> 7, + 2, SB_JOURNAL_BUCKETS); + + for (j = 0; j < ca->sb.keys; j++) + ca->sb.d[j] = ca->sb.first_bucket + j; + + bch_initial_gc_finish(c); + + err = "error starting allocator thread"; + if (bch_cache_allocator_start(ca)) + goto err; + + mutex_lock(&c->bucket_lock); + bch_prio_write(ca, true); + mutex_unlock(&c->bucket_lock); + + err = "cannot allocate new UUID bucket"; + if (__uuid_write(c)) + goto err; + + err = "cannot allocate new btree root"; + c->root = __bch_btree_node_alloc(c, NULL, 0, true, NULL); + if (IS_ERR(c->root)) + goto err; + + mutex_lock(&c->root->write_lock); + bkey_copy_key(&c->root->key, &MAX_KEY); + bch_btree_node_write(c->root, &cl); + mutex_unlock(&c->root->write_lock); + + bch_btree_set_root(c->root); + rw_unlock(true, c->root); + + /* + * We don't want to write the first journal entry until + * everything is set up - fortunately journal entries won't be + * written until the SET_CACHE_SYNC() here: + */ + SET_CACHE_SYNC(&c->cache->sb, true); + + bch_journal_next(&c->journal); + bch_journal_meta(c, &cl); + } + + err = "error starting gc thread"; + if (bch_gc_thread_start(c)) + goto err; + + closure_sync(&cl); + c->cache->sb.last_mount = (u32)ktime_get_real_seconds(); + bcache_write_super(c); + + if (bch_has_feature_obso_large_bucket(&c->cache->sb)) + pr_err("Detect obsoleted large bucket layout, all attached bcache device will be read-only\n"); + + list_for_each_entry_safe(dc, t, &uncached_devices, list) + bch_cached_dev_attach(dc, c, NULL); + + flash_devs_run(c); + + bch_journal_space_reserve(&c->journal); + set_bit(CACHE_SET_RUNNING, &c->flags); + return 0; +err: + while (!list_empty(&journal)) { + l = list_first_entry(&journal, struct journal_replay, list); + list_del(&l->list); + kfree(l); + } + + closure_sync(&cl); + + bch_cache_set_error(c, "%s", err); + + return -EIO; +} + +static const char *register_cache_set(struct cache *ca) +{ + char buf[12]; + const char *err = "cannot allocate memory"; + struct cache_set *c; + + list_for_each_entry(c, &bch_cache_sets, list) + if (!memcmp(c->set_uuid, ca->sb.set_uuid, 16)) { + if (c->cache) + return "duplicate cache set member"; + + goto found; + } + + c = bch_cache_set_alloc(&ca->sb); + if (!c) + return err; + + err = "error creating kobject"; + if (kobject_add(&c->kobj, bcache_kobj, "%pU", c->set_uuid) || + kobject_add(&c->internal, &c->kobj, "internal")) + goto err; + + if (bch_cache_accounting_add_kobjs(&c->accounting, &c->kobj)) + goto err; + + bch_debug_init_cache_set(c); + + list_add(&c->list, &bch_cache_sets); +found: + sprintf(buf, "cache%i", ca->sb.nr_this_dev); + if (sysfs_create_link(&ca->kobj, &c->kobj, "set") || + sysfs_create_link(&c->kobj, &ca->kobj, buf)) + goto err; + + kobject_get(&ca->kobj); + ca->set = c; + ca->set->cache = ca; + + err = "failed to run cache set"; + if (run_cache_set(c) < 0) + goto err; + + return NULL; +err: + bch_cache_set_unregister(c); + return err; +} + +/* Cache device */ + +/* When ca->kobj released */ +void bch_cache_release(struct kobject *kobj) +{ + struct cache *ca = container_of(kobj, struct cache, kobj); + unsigned int i; + + if (ca->set) { + BUG_ON(ca->set->cache != ca); + ca->set->cache = NULL; + } + + free_pages((unsigned long) ca->disk_buckets, ilog2(meta_bucket_pages(&ca->sb))); + kfree(ca->prio_buckets); + vfree(ca->buckets); + + free_heap(&ca->heap); + free_fifo(&ca->free_inc); + + for (i = 0; i < RESERVE_NR; i++) + free_fifo(&ca->free[i]); + + if (ca->sb_disk) + put_page(virt_to_page(ca->sb_disk)); + + if (!IS_ERR_OR_NULL(ca->bdev)) + blkdev_put(ca->bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL); + + kfree(ca); + module_put(THIS_MODULE); +} + +static int cache_alloc(struct cache *ca) +{ + size_t free; + size_t btree_buckets; + struct bucket *b; + int ret = -ENOMEM; + const char *err = NULL; + + __module_get(THIS_MODULE); + kobject_init(&ca->kobj, &bch_cache_ktype); + + bio_init(&ca->journal.bio, ca->journal.bio.bi_inline_vecs, 8); + + /* + * when ca->sb.njournal_buckets is not zero, journal exists, + * and in bch_journal_replay(), tree node may split, + * so bucket of RESERVE_BTREE type is needed, + * the worst situation is all journal buckets are valid journal, + * and all the keys need to replay, + * so the number of RESERVE_BTREE type buckets should be as much + * as journal buckets + */ + btree_buckets = ca->sb.njournal_buckets ?: 8; + free = roundup_pow_of_two(ca->sb.nbuckets) >> 10; + if (!free) { + ret = -EPERM; + err = "ca->sb.nbuckets is too small"; + goto err_free; + } + + if (!init_fifo(&ca->free[RESERVE_BTREE], btree_buckets, + GFP_KERNEL)) { + err = "ca->free[RESERVE_BTREE] alloc failed"; + goto err_btree_alloc; + } + + if (!init_fifo_exact(&ca->free[RESERVE_PRIO], prio_buckets(ca), + GFP_KERNEL)) { + err = "ca->free[RESERVE_PRIO] alloc failed"; + goto err_prio_alloc; + } + + if (!init_fifo(&ca->free[RESERVE_MOVINGGC], free, GFP_KERNEL)) { + err = "ca->free[RESERVE_MOVINGGC] alloc failed"; + goto err_movinggc_alloc; + } + + if (!init_fifo(&ca->free[RESERVE_NONE], free, GFP_KERNEL)) { + err = "ca->free[RESERVE_NONE] alloc failed"; + goto err_none_alloc; + } + + if (!init_fifo(&ca->free_inc, free << 2, GFP_KERNEL)) { + err = "ca->free_inc alloc failed"; + goto err_free_inc_alloc; + } + + if (!init_heap(&ca->heap, free << 3, GFP_KERNEL)) { + err = "ca->heap alloc failed"; + goto err_heap_alloc; + } + + ca->buckets = vzalloc(array_size(sizeof(struct bucket), + ca->sb.nbuckets)); + if (!ca->buckets) { + err = "ca->buckets alloc failed"; + goto err_buckets_alloc; + } + + ca->prio_buckets = kzalloc(array3_size(sizeof(uint64_t), + prio_buckets(ca), 2), + GFP_KERNEL); + if (!ca->prio_buckets) { + err = "ca->prio_buckets alloc failed"; + goto err_prio_buckets_alloc; + } + + ca->disk_buckets = alloc_meta_bucket_pages(GFP_KERNEL, &ca->sb); + if (!ca->disk_buckets) { + err = "ca->disk_buckets alloc failed"; + goto err_disk_buckets_alloc; + } + + ca->prio_last_buckets = ca->prio_buckets + prio_buckets(ca); + + for_each_bucket(b, ca) + atomic_set(&b->pin, 0); + return 0; + +err_disk_buckets_alloc: + kfree(ca->prio_buckets); +err_prio_buckets_alloc: + vfree(ca->buckets); +err_buckets_alloc: + free_heap(&ca->heap); +err_heap_alloc: + free_fifo(&ca->free_inc); +err_free_inc_alloc: + free_fifo(&ca->free[RESERVE_NONE]); +err_none_alloc: + free_fifo(&ca->free[RESERVE_MOVINGGC]); +err_movinggc_alloc: + free_fifo(&ca->free[RESERVE_PRIO]); +err_prio_alloc: + free_fifo(&ca->free[RESERVE_BTREE]); +err_btree_alloc: +err_free: + module_put(THIS_MODULE); + if (err) + pr_notice("error %s: %s\n", ca->cache_dev_name, err); + return ret; +} + +static int register_cache(struct cache_sb *sb, struct cache_sb_disk *sb_disk, + struct block_device *bdev, struct cache *ca) +{ + const char *err = NULL; /* must be set for any error case */ + int ret = 0; + + bdevname(bdev, ca->cache_dev_name); + memcpy(&ca->sb, sb, sizeof(struct cache_sb)); + ca->bdev = bdev; + ca->bdev->bd_holder = ca; + ca->sb_disk = sb_disk; + + if (blk_queue_discard(bdev_get_queue(bdev))) + ca->discard = CACHE_DISCARD(&ca->sb); + + ret = cache_alloc(ca); + if (ret != 0) { + /* + * If we failed here, it means ca->kobj is not initialized yet, + * kobject_put() won't be called and there is no chance to + * call blkdev_put() to bdev in bch_cache_release(). So we + * explicitly call blkdev_put() here. + */ + blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL); + if (ret == -ENOMEM) + err = "cache_alloc(): -ENOMEM"; + else if (ret == -EPERM) + err = "cache_alloc(): cache device is too small"; + else + err = "cache_alloc(): unknown error"; + goto err; + } + + if (kobject_add(&ca->kobj, + &part_to_dev(bdev->bd_part)->kobj, + "bcache")) { + err = "error calling kobject_add"; + ret = -ENOMEM; + goto out; + } + + mutex_lock(&bch_register_lock); + err = register_cache_set(ca); + mutex_unlock(&bch_register_lock); + + if (err) { + ret = -ENODEV; + goto out; + } + + pr_info("registered cache device %s\n", ca->cache_dev_name); + +out: + kobject_put(&ca->kobj); + +err: + if (err) + pr_notice("error %s: %s\n", ca->cache_dev_name, err); + + return ret; +} + +/* Global interfaces/init */ + +static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr, + const char *buffer, size_t size); +static ssize_t bch_pending_bdevs_cleanup(struct kobject *k, + struct kobj_attribute *attr, + const char *buffer, size_t size); + +kobj_attribute_write(register, register_bcache); +kobj_attribute_write(register_quiet, register_bcache); +kobj_attribute_write(pendings_cleanup, bch_pending_bdevs_cleanup); + +static bool bch_is_open_backing(struct block_device *bdev) +{ + struct cache_set *c, *tc; + struct cached_dev *dc, *t; + + list_for_each_entry_safe(c, tc, &bch_cache_sets, list) + list_for_each_entry_safe(dc, t, &c->cached_devs, list) + if (dc->bdev == bdev) + return true; + list_for_each_entry_safe(dc, t, &uncached_devices, list) + if (dc->bdev == bdev) + return true; + return false; +} + +static bool bch_is_open_cache(struct block_device *bdev) +{ + struct cache_set *c, *tc; + + list_for_each_entry_safe(c, tc, &bch_cache_sets, list) { + struct cache *ca = c->cache; + + if (ca->bdev == bdev) + return true; + } + + return false; +} + +static bool bch_is_open(struct block_device *bdev) +{ + return bch_is_open_cache(bdev) || bch_is_open_backing(bdev); +} + +struct async_reg_args { + struct delayed_work reg_work; + char *path; + struct cache_sb *sb; + struct cache_sb_disk *sb_disk; + struct block_device *bdev; +}; + +static void register_bdev_worker(struct work_struct *work) +{ + int fail = false; + struct async_reg_args *args = + container_of(work, struct async_reg_args, reg_work.work); + struct cached_dev *dc; + + dc = kzalloc(sizeof(*dc), GFP_KERNEL); + if (!dc) { + fail = true; + put_page(virt_to_page(args->sb_disk)); + blkdev_put(args->bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL); + goto out; + } + + mutex_lock(&bch_register_lock); + if (register_bdev(args->sb, args->sb_disk, args->bdev, dc) < 0) + fail = true; + mutex_unlock(&bch_register_lock); + +out: + if (fail) + pr_info("error %s: fail to register backing device\n", + args->path); + kfree(args->sb); + kfree(args->path); + kfree(args); + module_put(THIS_MODULE); +} + +static void register_cache_worker(struct work_struct *work) +{ + int fail = false; + struct async_reg_args *args = + container_of(work, struct async_reg_args, reg_work.work); + struct cache *ca; + + ca = kzalloc(sizeof(*ca), GFP_KERNEL); + if (!ca) { + fail = true; + put_page(virt_to_page(args->sb_disk)); + blkdev_put(args->bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL); + goto out; + } + + /* blkdev_put() will be called in bch_cache_release() */ + if (register_cache(args->sb, args->sb_disk, args->bdev, ca) != 0) + fail = true; + +out: + if (fail) + pr_info("error %s: fail to register cache device\n", + args->path); + kfree(args->sb); + kfree(args->path); + kfree(args); + module_put(THIS_MODULE); +} + +static void register_device_aync(struct async_reg_args *args) +{ + if (SB_IS_BDEV(args->sb)) + INIT_DELAYED_WORK(&args->reg_work, register_bdev_worker); + else + INIT_DELAYED_WORK(&args->reg_work, register_cache_worker); + + /* 10 jiffies is enough for a delay */ + queue_delayed_work(system_wq, &args->reg_work, 10); +} + +static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr, + const char *buffer, size_t size) +{ + const char *err; + char *path = NULL; + struct cache_sb *sb; + struct cache_sb_disk *sb_disk; + struct block_device *bdev; + ssize_t ret; + bool async_registration = false; + +#ifdef CONFIG_BCACHE_ASYNC_REGISTRATION + async_registration = true; +#endif + + ret = -EBUSY; + err = "failed to reference bcache module"; + if (!try_module_get(THIS_MODULE)) + goto out; + + /* For latest state of bcache_is_reboot */ + smp_mb(); + err = "bcache is in reboot"; + if (bcache_is_reboot) + goto out_module_put; + + ret = -ENOMEM; + err = "cannot allocate memory"; + path = kstrndup(buffer, size, GFP_KERNEL); + if (!path) + goto out_module_put; + + sb = kmalloc(sizeof(struct cache_sb), GFP_KERNEL); + if (!sb) + goto out_free_path; + + ret = -EINVAL; + err = "failed to open device"; + bdev = blkdev_get_by_path(strim(path), + FMODE_READ|FMODE_WRITE|FMODE_EXCL, + sb); + if (IS_ERR(bdev)) { + if (bdev == ERR_PTR(-EBUSY)) { + bdev = lookup_bdev(strim(path)); + mutex_lock(&bch_register_lock); + if (!IS_ERR(bdev) && bch_is_open(bdev)) + err = "device already registered"; + else + err = "device busy"; + mutex_unlock(&bch_register_lock); + if (!IS_ERR(bdev)) + bdput(bdev); + if (attr == &ksysfs_register_quiet) + goto done; + } + goto out_free_sb; + } + + err = "failed to set blocksize"; + if (set_blocksize(bdev, 4096)) + goto out_blkdev_put; + + err = read_super(sb, bdev, &sb_disk); + if (err) + goto out_blkdev_put; + + err = "failed to register device"; + + if (async_registration) { + /* register in asynchronous way */ + struct async_reg_args *args = + kzalloc(sizeof(struct async_reg_args), GFP_KERNEL); + + if (!args) { + ret = -ENOMEM; + err = "cannot allocate memory"; + goto out_put_sb_page; + } + + args->path = path; + args->sb = sb; + args->sb_disk = sb_disk; + args->bdev = bdev; + register_device_aync(args); + /* No wait and returns to user space */ + goto async_done; + } + + if (SB_IS_BDEV(sb)) { + struct cached_dev *dc = kzalloc(sizeof(*dc), GFP_KERNEL); + + if (!dc) + goto out_put_sb_page; + + mutex_lock(&bch_register_lock); + ret = register_bdev(sb, sb_disk, bdev, dc); + mutex_unlock(&bch_register_lock); + /* blkdev_put() will be called in cached_dev_free() */ + if (ret < 0) + goto out_free_sb; + } else { + struct cache *ca = kzalloc(sizeof(*ca), GFP_KERNEL); + + if (!ca) + goto out_put_sb_page; + + /* blkdev_put() will be called in bch_cache_release() */ + if (register_cache(sb, sb_disk, bdev, ca) != 0) + goto out_free_sb; + } + +done: + kfree(sb); + kfree(path); + module_put(THIS_MODULE); +async_done: + return size; + +out_put_sb_page: + put_page(virt_to_page(sb_disk)); +out_blkdev_put: + blkdev_put(bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL); +out_free_sb: + kfree(sb); +out_free_path: + kfree(path); + path = NULL; +out_module_put: + module_put(THIS_MODULE); +out: + pr_info("error %s: %s\n", path?path:"", err); + return ret; +} + + +struct pdev { + struct list_head list; + struct cached_dev *dc; +}; + +static ssize_t bch_pending_bdevs_cleanup(struct kobject *k, + struct kobj_attribute *attr, + const char *buffer, + size_t size) +{ + LIST_HEAD(pending_devs); + ssize_t ret = size; + struct cached_dev *dc, *tdc; + struct pdev *pdev, *tpdev; + struct cache_set *c, *tc; + + mutex_lock(&bch_register_lock); + list_for_each_entry_safe(dc, tdc, &uncached_devices, list) { + pdev = kmalloc(sizeof(struct pdev), GFP_KERNEL); + if (!pdev) + break; + pdev->dc = dc; + list_add(&pdev->list, &pending_devs); + } + + list_for_each_entry_safe(pdev, tpdev, &pending_devs, list) { + list_for_each_entry_safe(c, tc, &bch_cache_sets, list) { + char *pdev_set_uuid = pdev->dc->sb.set_uuid; + char *set_uuid = c->set_uuid; + + if (!memcmp(pdev_set_uuid, set_uuid, 16)) { + list_del(&pdev->list); + kfree(pdev); + break; + } + } + } + mutex_unlock(&bch_register_lock); + + list_for_each_entry_safe(pdev, tpdev, &pending_devs, list) { + pr_info("delete pdev %p\n", pdev); + list_del(&pdev->list); + bcache_device_stop(&pdev->dc->disk); + kfree(pdev); + } + + return ret; +} + +static int bcache_reboot(struct notifier_block *n, unsigned long code, void *x) +{ + if (bcache_is_reboot) + return NOTIFY_DONE; + + if (code == SYS_DOWN || + code == SYS_HALT || + code == SYS_POWER_OFF) { + DEFINE_WAIT(wait); + unsigned long start = jiffies; + bool stopped = false; + + struct cache_set *c, *tc; + struct cached_dev *dc, *tdc; + + mutex_lock(&bch_register_lock); + + if (bcache_is_reboot) + goto out; + + /* New registration is rejected since now */ + bcache_is_reboot = true; + /* + * Make registering caller (if there is) on other CPU + * core know bcache_is_reboot set to true earlier + */ + smp_mb(); + + if (list_empty(&bch_cache_sets) && + list_empty(&uncached_devices)) + goto out; + + mutex_unlock(&bch_register_lock); + + pr_info("Stopping all devices:\n"); + + /* + * The reason bch_register_lock is not held to call + * bch_cache_set_stop() and bcache_device_stop() is to + * avoid potential deadlock during reboot, because cache + * set or bcache device stopping process will acqurie + * bch_register_lock too. + * + * We are safe here because bcache_is_reboot sets to + * true already, register_bcache() will reject new + * registration now. bcache_is_reboot also makes sure + * bcache_reboot() won't be re-entered on by other thread, + * so there is no race in following list iteration by + * list_for_each_entry_safe(). + */ + list_for_each_entry_safe(c, tc, &bch_cache_sets, list) + bch_cache_set_stop(c); + + list_for_each_entry_safe(dc, tdc, &uncached_devices, list) + bcache_device_stop(&dc->disk); + + + /* + * Give an early chance for other kthreads and + * kworkers to stop themselves + */ + schedule(); + + /* What's a condition variable? */ + while (1) { + long timeout = start + 10 * HZ - jiffies; + + mutex_lock(&bch_register_lock); + stopped = list_empty(&bch_cache_sets) && + list_empty(&uncached_devices); + + if (timeout < 0 || stopped) + break; + + prepare_to_wait(&unregister_wait, &wait, + TASK_UNINTERRUPTIBLE); + + mutex_unlock(&bch_register_lock); + schedule_timeout(timeout); + } + + finish_wait(&unregister_wait, &wait); + + if (stopped) + pr_info("All devices stopped\n"); + else + pr_notice("Timeout waiting for devices to be closed\n"); +out: + mutex_unlock(&bch_register_lock); + } + + return NOTIFY_DONE; +} + +static struct notifier_block reboot = { + .notifier_call = bcache_reboot, + .priority = INT_MAX, /* before any real devices */ +}; + +static void bcache_exit(void) +{ + bch_debug_exit(); + bch_request_exit(); + if (bcache_kobj) + kobject_put(bcache_kobj); + if (bcache_wq) + destroy_workqueue(bcache_wq); + if (bch_journal_wq) + destroy_workqueue(bch_journal_wq); + if (bch_flush_wq) + destroy_workqueue(bch_flush_wq); + bch_btree_exit(); + + if (bcache_major) + unregister_blkdev(bcache_major, "bcache"); + unregister_reboot_notifier(&reboot); + mutex_destroy(&bch_register_lock); +} + +/* Check and fixup module parameters */ +static void check_module_parameters(void) +{ + if (bch_cutoff_writeback_sync == 0) + bch_cutoff_writeback_sync = CUTOFF_WRITEBACK_SYNC; + else if (bch_cutoff_writeback_sync > CUTOFF_WRITEBACK_SYNC_MAX) { + pr_warn("set bch_cutoff_writeback_sync (%u) to max value %u\n", + bch_cutoff_writeback_sync, CUTOFF_WRITEBACK_SYNC_MAX); + bch_cutoff_writeback_sync = CUTOFF_WRITEBACK_SYNC_MAX; + } + + if (bch_cutoff_writeback == 0) + bch_cutoff_writeback = CUTOFF_WRITEBACK; + else if (bch_cutoff_writeback > CUTOFF_WRITEBACK_MAX) { + pr_warn("set bch_cutoff_writeback (%u) to max value %u\n", + bch_cutoff_writeback, CUTOFF_WRITEBACK_MAX); + bch_cutoff_writeback = CUTOFF_WRITEBACK_MAX; + } + + if (bch_cutoff_writeback > bch_cutoff_writeback_sync) { + pr_warn("set bch_cutoff_writeback (%u) to %u\n", + bch_cutoff_writeback, bch_cutoff_writeback_sync); + bch_cutoff_writeback = bch_cutoff_writeback_sync; + } +} + +static int __init bcache_init(void) +{ + static const struct attribute *files[] = { + &ksysfs_register.attr, + &ksysfs_register_quiet.attr, + &ksysfs_pendings_cleanup.attr, + NULL + }; + + check_module_parameters(); + + mutex_init(&bch_register_lock); + init_waitqueue_head(&unregister_wait); + register_reboot_notifier(&reboot); + + bcache_major = register_blkdev(0, "bcache"); + if (bcache_major < 0) { + unregister_reboot_notifier(&reboot); + mutex_destroy(&bch_register_lock); + return bcache_major; + } + + if (bch_btree_init()) + goto err; + + bcache_wq = alloc_workqueue("bcache", WQ_MEM_RECLAIM, 0); + if (!bcache_wq) + goto err; + + /* + * Let's not make this `WQ_MEM_RECLAIM` for the following reasons: + * + * 1. It used `system_wq` before which also does no memory reclaim. + * 2. With `WQ_MEM_RECLAIM` desktop stalls, increased boot times, and + * reduced throughput can be observed. + * + * We still want to user our own queue to not congest the `system_wq`. + */ + bch_flush_wq = alloc_workqueue("bch_flush", 0, 0); + if (!bch_flush_wq) + goto err; + + bch_journal_wq = alloc_workqueue("bch_journal", WQ_MEM_RECLAIM, 0); + if (!bch_journal_wq) + goto err; + + bcache_kobj = kobject_create_and_add("bcache", fs_kobj); + if (!bcache_kobj) + goto err; + + if (bch_request_init() || + sysfs_create_files(bcache_kobj, files)) + goto err; + + bch_debug_init(); + closure_debug_init(); + + bcache_is_reboot = false; + + return 0; +err: + bcache_exit(); + return -ENOMEM; +} + +/* + * Module hooks + */ +module_exit(bcache_exit); +module_init(bcache_init); + +module_param(bch_cutoff_writeback, uint, 0); +MODULE_PARM_DESC(bch_cutoff_writeback, "threshold to cutoff writeback"); + +module_param(bch_cutoff_writeback_sync, uint, 0); +MODULE_PARM_DESC(bch_cutoff_writeback_sync, "hard threshold to cutoff writeback"); + +MODULE_DESCRIPTION("Bcache: a Linux block layer cache"); +MODULE_AUTHOR("Kent Overstreet <kent.overstreet@gmail.com>"); +MODULE_LICENSE("GPL"); diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c new file mode 100644 index 000000000..ca3e2f000 --- /dev/null +++ b/drivers/md/bcache/sysfs.c @@ -0,0 +1,1178 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * bcache sysfs interfaces + * + * Copyright 2010, 2011 Kent Overstreet <kent.overstreet@gmail.com> + * Copyright 2012 Google, Inc. + */ + +#include "bcache.h" +#include "sysfs.h" +#include "btree.h" +#include "request.h" +#include "writeback.h" +#include "features.h" + +#include <linux/blkdev.h> +#include <linux/sort.h> +#include <linux/sched/clock.h> + +extern bool bcache_is_reboot; + +/* Default is 0 ("writethrough") */ +static const char * const bch_cache_modes[] = { + "writethrough", + "writeback", + "writearound", + "none", + NULL +}; + +static const char * const bch_reada_cache_policies[] = { + "all", + "meta-only", + NULL +}; + +/* Default is 0 ("auto") */ +static const char * const bch_stop_on_failure_modes[] = { + "auto", + "always", + NULL +}; + +static const char * const cache_replacement_policies[] = { + "lru", + "fifo", + "random", + NULL +}; + +static const char * const error_actions[] = { + "unregister", + "panic", + NULL +}; + +write_attribute(attach); +write_attribute(detach); +write_attribute(unregister); +write_attribute(stop); +write_attribute(clear_stats); +write_attribute(trigger_gc); +write_attribute(prune_cache); +write_attribute(flash_vol_create); + +read_attribute(bucket_size); +read_attribute(block_size); +read_attribute(nbuckets); +read_attribute(tree_depth); +read_attribute(root_usage_percent); +read_attribute(priority_stats); +read_attribute(btree_cache_size); +read_attribute(btree_cache_max_chain); +read_attribute(cache_available_percent); +read_attribute(written); +read_attribute(btree_written); +read_attribute(metadata_written); +read_attribute(active_journal_entries); +read_attribute(backing_dev_name); +read_attribute(backing_dev_uuid); + +sysfs_time_stats_attribute(btree_gc, sec, ms); +sysfs_time_stats_attribute(btree_split, sec, us); +sysfs_time_stats_attribute(btree_sort, ms, us); +sysfs_time_stats_attribute(btree_read, ms, us); + +read_attribute(btree_nodes); +read_attribute(btree_used_percent); +read_attribute(average_key_size); +read_attribute(dirty_data); +read_attribute(bset_tree_stats); +read_attribute(feature_compat); +read_attribute(feature_ro_compat); +read_attribute(feature_incompat); + +read_attribute(state); +read_attribute(cache_read_races); +read_attribute(reclaim); +read_attribute(reclaimed_journal_buckets); +read_attribute(flush_write); +read_attribute(writeback_keys_done); +read_attribute(writeback_keys_failed); +read_attribute(io_errors); +read_attribute(congested); +read_attribute(cutoff_writeback); +read_attribute(cutoff_writeback_sync); +rw_attribute(congested_read_threshold_us); +rw_attribute(congested_write_threshold_us); + +rw_attribute(sequential_cutoff); +rw_attribute(data_csum); +rw_attribute(cache_mode); +rw_attribute(readahead_cache_policy); +rw_attribute(stop_when_cache_set_failed); +rw_attribute(writeback_metadata); +rw_attribute(writeback_running); +rw_attribute(writeback_percent); +rw_attribute(writeback_delay); +rw_attribute(writeback_rate); + +rw_attribute(writeback_rate_update_seconds); +rw_attribute(writeback_rate_i_term_inverse); +rw_attribute(writeback_rate_p_term_inverse); +rw_attribute(writeback_rate_minimum); +read_attribute(writeback_rate_debug); + +read_attribute(stripe_size); +read_attribute(partial_stripes_expensive); + +rw_attribute(synchronous); +rw_attribute(journal_delay_ms); +rw_attribute(io_disable); +rw_attribute(discard); +rw_attribute(running); +rw_attribute(label); +rw_attribute(readahead); +rw_attribute(errors); +rw_attribute(io_error_limit); +rw_attribute(io_error_halflife); +rw_attribute(verify); +rw_attribute(bypass_torture_test); +rw_attribute(key_merging_disabled); +rw_attribute(gc_always_rewrite); +rw_attribute(expensive_debug_checks); +rw_attribute(cache_replacement_policy); +rw_attribute(btree_shrinker_disabled); +rw_attribute(copy_gc_enabled); +rw_attribute(idle_max_writeback_rate); +rw_attribute(gc_after_writeback); +rw_attribute(size); + +static ssize_t bch_snprint_string_list(char *buf, + size_t size, + const char * const list[], + size_t selected) +{ + char *out = buf; + size_t i; + + for (i = 0; list[i]; i++) + out += scnprintf(out, buf + size - out, + i == selected ? "[%s] " : "%s ", list[i]); + + out[-1] = '\n'; + return out - buf; +} + +SHOW(__bch_cached_dev) +{ + struct cached_dev *dc = container_of(kobj, struct cached_dev, + disk.kobj); + char const *states[] = { "no cache", "clean", "dirty", "inconsistent" }; + int wb = dc->writeback_running; + +#define var(stat) (dc->stat) + + if (attr == &sysfs_cache_mode) + return bch_snprint_string_list(buf, PAGE_SIZE, + bch_cache_modes, + BDEV_CACHE_MODE(&dc->sb)); + + if (attr == &sysfs_readahead_cache_policy) + return bch_snprint_string_list(buf, PAGE_SIZE, + bch_reada_cache_policies, + dc->cache_readahead_policy); + + if (attr == &sysfs_stop_when_cache_set_failed) + return bch_snprint_string_list(buf, PAGE_SIZE, + bch_stop_on_failure_modes, + dc->stop_when_cache_set_failed); + + + sysfs_printf(data_csum, "%i", dc->disk.data_csum); + var_printf(verify, "%i"); + var_printf(bypass_torture_test, "%i"); + var_printf(writeback_metadata, "%i"); + var_printf(writeback_running, "%i"); + var_print(writeback_delay); + var_print(writeback_percent); + sysfs_hprint(writeback_rate, + wb ? atomic_long_read(&dc->writeback_rate.rate) << 9 : 0); + sysfs_printf(io_errors, "%i", atomic_read(&dc->io_errors)); + sysfs_printf(io_error_limit, "%i", dc->error_limit); + sysfs_printf(io_disable, "%i", dc->io_disable); + var_print(writeback_rate_update_seconds); + var_print(writeback_rate_i_term_inverse); + var_print(writeback_rate_p_term_inverse); + var_print(writeback_rate_minimum); + + if (attr == &sysfs_writeback_rate_debug) { + char rate[20]; + char dirty[20]; + char target[20]; + char proportional[20]; + char integral[20]; + char change[20]; + s64 next_io; + + /* + * Except for dirty and target, other values should + * be 0 if writeback is not running. + */ + bch_hprint(rate, + wb ? atomic_long_read(&dc->writeback_rate.rate) << 9 + : 0); + bch_hprint(dirty, bcache_dev_sectors_dirty(&dc->disk) << 9); + bch_hprint(target, dc->writeback_rate_target << 9); + bch_hprint(proportional, + wb ? dc->writeback_rate_proportional << 9 : 0); + bch_hprint(integral, + wb ? dc->writeback_rate_integral_scaled << 9 : 0); + bch_hprint(change, wb ? dc->writeback_rate_change << 9 : 0); + next_io = wb ? div64_s64(dc->writeback_rate.next-local_clock(), + NSEC_PER_MSEC) : 0; + + return sprintf(buf, + "rate:\t\t%s/sec\n" + "dirty:\t\t%s\n" + "target:\t\t%s\n" + "proportional:\t%s\n" + "integral:\t%s\n" + "change:\t\t%s/sec\n" + "next io:\t%llims\n", + rate, dirty, target, proportional, + integral, change, next_io); + } + + sysfs_hprint(dirty_data, + bcache_dev_sectors_dirty(&dc->disk) << 9); + + sysfs_hprint(stripe_size, ((uint64_t)dc->disk.stripe_size) << 9); + var_printf(partial_stripes_expensive, "%u"); + + var_hprint(sequential_cutoff); + var_hprint(readahead); + + sysfs_print(running, atomic_read(&dc->running)); + sysfs_print(state, states[BDEV_STATE(&dc->sb)]); + + if (attr == &sysfs_label) { + memcpy(buf, dc->sb.label, SB_LABEL_SIZE); + buf[SB_LABEL_SIZE + 1] = '\0'; + strcat(buf, "\n"); + return strlen(buf); + } + + if (attr == &sysfs_backing_dev_name) { + snprintf(buf, BDEVNAME_SIZE + 1, "%s", dc->backing_dev_name); + strcat(buf, "\n"); + return strlen(buf); + } + + if (attr == &sysfs_backing_dev_uuid) { + /* convert binary uuid into 36-byte string plus '\0' */ + snprintf(buf, 36+1, "%pU", dc->sb.uuid); + strcat(buf, "\n"); + return strlen(buf); + } + +#undef var + return 0; +} +SHOW_LOCKED(bch_cached_dev) + +STORE(__cached_dev) +{ + struct cached_dev *dc = container_of(kobj, struct cached_dev, + disk.kobj); + ssize_t v; + struct cache_set *c; + struct kobj_uevent_env *env; + + /* no user space access if system is rebooting */ + if (bcache_is_reboot) + return -EBUSY; + +#define d_strtoul(var) sysfs_strtoul(var, dc->var) +#define d_strtoul_nonzero(var) sysfs_strtoul_clamp(var, dc->var, 1, INT_MAX) +#define d_strtoi_h(var) sysfs_hatoi(var, dc->var) + + sysfs_strtoul(data_csum, dc->disk.data_csum); + d_strtoul(verify); + sysfs_strtoul_bool(bypass_torture_test, dc->bypass_torture_test); + sysfs_strtoul_bool(writeback_metadata, dc->writeback_metadata); + sysfs_strtoul_bool(writeback_running, dc->writeback_running); + sysfs_strtoul_clamp(writeback_delay, dc->writeback_delay, 0, UINT_MAX); + + sysfs_strtoul_clamp(writeback_percent, dc->writeback_percent, + 0, bch_cutoff_writeback); + + if (attr == &sysfs_writeback_rate) { + ssize_t ret; + long int v = atomic_long_read(&dc->writeback_rate.rate); + + ret = strtoul_safe_clamp(buf, v, 1, INT_MAX); + + if (!ret) { + atomic_long_set(&dc->writeback_rate.rate, v); + ret = size; + } + + return ret; + } + + sysfs_strtoul_clamp(writeback_rate_update_seconds, + dc->writeback_rate_update_seconds, + 1, WRITEBACK_RATE_UPDATE_SECS_MAX); + sysfs_strtoul_clamp(writeback_rate_i_term_inverse, + dc->writeback_rate_i_term_inverse, + 1, UINT_MAX); + sysfs_strtoul_clamp(writeback_rate_p_term_inverse, + dc->writeback_rate_p_term_inverse, + 1, UINT_MAX); + sysfs_strtoul_clamp(writeback_rate_minimum, + dc->writeback_rate_minimum, + 1, UINT_MAX); + + sysfs_strtoul_clamp(io_error_limit, dc->error_limit, 0, INT_MAX); + + if (attr == &sysfs_io_disable) { + int v = strtoul_or_return(buf); + + dc->io_disable = v ? 1 : 0; + } + + sysfs_strtoul_clamp(sequential_cutoff, + dc->sequential_cutoff, + 0, UINT_MAX); + d_strtoi_h(readahead); + + if (attr == &sysfs_clear_stats) + bch_cache_accounting_clear(&dc->accounting); + + if (attr == &sysfs_running && + strtoul_or_return(buf)) { + v = bch_cached_dev_run(dc); + if (v) + return v; + } + + if (attr == &sysfs_cache_mode) { + v = __sysfs_match_string(bch_cache_modes, -1, buf); + if (v < 0) + return v; + + if ((unsigned int) v != BDEV_CACHE_MODE(&dc->sb)) { + SET_BDEV_CACHE_MODE(&dc->sb, v); + bch_write_bdev_super(dc, NULL); + } + } + + if (attr == &sysfs_readahead_cache_policy) { + v = __sysfs_match_string(bch_reada_cache_policies, -1, buf); + if (v < 0) + return v; + + if ((unsigned int) v != dc->cache_readahead_policy) + dc->cache_readahead_policy = v; + } + + if (attr == &sysfs_stop_when_cache_set_failed) { + v = __sysfs_match_string(bch_stop_on_failure_modes, -1, buf); + if (v < 0) + return v; + + dc->stop_when_cache_set_failed = v; + } + + if (attr == &sysfs_label) { + if (size > SB_LABEL_SIZE) + return -EINVAL; + memcpy(dc->sb.label, buf, size); + if (size < SB_LABEL_SIZE) + dc->sb.label[size] = '\0'; + if (size && dc->sb.label[size - 1] == '\n') + dc->sb.label[size - 1] = '\0'; + bch_write_bdev_super(dc, NULL); + if (dc->disk.c) { + memcpy(dc->disk.c->uuids[dc->disk.id].label, + buf, SB_LABEL_SIZE); + bch_uuid_write(dc->disk.c); + } + env = kzalloc(sizeof(struct kobj_uevent_env), GFP_KERNEL); + if (!env) + return -ENOMEM; + add_uevent_var(env, "DRIVER=bcache"); + add_uevent_var(env, "CACHED_UUID=%pU", dc->sb.uuid), + add_uevent_var(env, "CACHED_LABEL=%s", buf); + kobject_uevent_env(&disk_to_dev(dc->disk.disk)->kobj, + KOBJ_CHANGE, + env->envp); + kfree(env); + } + + if (attr == &sysfs_attach) { + uint8_t set_uuid[16]; + + if (bch_parse_uuid(buf, set_uuid) < 16) + return -EINVAL; + + v = -ENOENT; + list_for_each_entry(c, &bch_cache_sets, list) { + v = bch_cached_dev_attach(dc, c, set_uuid); + if (!v) + return size; + } + if (v == -ENOENT) + pr_err("Can't attach %s: cache set not found\n", buf); + return v; + } + + if (attr == &sysfs_detach && dc->disk.c) + bch_cached_dev_detach(dc); + + if (attr == &sysfs_stop) + bcache_device_stop(&dc->disk); + + return size; +} + +STORE(bch_cached_dev) +{ + struct cached_dev *dc = container_of(kobj, struct cached_dev, + disk.kobj); + + /* no user space access if system is rebooting */ + if (bcache_is_reboot) + return -EBUSY; + + mutex_lock(&bch_register_lock); + size = __cached_dev_store(kobj, attr, buf, size); + + if (attr == &sysfs_writeback_running) { + /* dc->writeback_running changed in __cached_dev_store() */ + if (IS_ERR_OR_NULL(dc->writeback_thread)) { + /* + * reject setting it to 1 via sysfs if writeback + * kthread is not created yet. + */ + if (dc->writeback_running) { + dc->writeback_running = false; + pr_err("%s: failed to run non-existent writeback thread\n", + dc->disk.disk->disk_name); + } + } else + /* + * writeback kthread will check if dc->writeback_running + * is true or false. + */ + bch_writeback_queue(dc); + } + + /* + * Only set BCACHE_DEV_WB_RUNNING when cached device attached to + * a cache set, otherwise it doesn't make sense. + */ + if (attr == &sysfs_writeback_percent) + if ((dc->disk.c != NULL) && + (!test_and_set_bit(BCACHE_DEV_WB_RUNNING, &dc->disk.flags))) + schedule_delayed_work(&dc->writeback_rate_update, + dc->writeback_rate_update_seconds * HZ); + + mutex_unlock(&bch_register_lock); + return size; +} + +static struct attribute *bch_cached_dev_files[] = { + &sysfs_attach, + &sysfs_detach, + &sysfs_stop, +#if 0 + &sysfs_data_csum, +#endif + &sysfs_cache_mode, + &sysfs_readahead_cache_policy, + &sysfs_stop_when_cache_set_failed, + &sysfs_writeback_metadata, + &sysfs_writeback_running, + &sysfs_writeback_delay, + &sysfs_writeback_percent, + &sysfs_writeback_rate, + &sysfs_writeback_rate_update_seconds, + &sysfs_writeback_rate_i_term_inverse, + &sysfs_writeback_rate_p_term_inverse, + &sysfs_writeback_rate_minimum, + &sysfs_writeback_rate_debug, + &sysfs_io_errors, + &sysfs_io_error_limit, + &sysfs_io_disable, + &sysfs_dirty_data, + &sysfs_stripe_size, + &sysfs_partial_stripes_expensive, + &sysfs_sequential_cutoff, + &sysfs_clear_stats, + &sysfs_running, + &sysfs_state, + &sysfs_label, + &sysfs_readahead, +#ifdef CONFIG_BCACHE_DEBUG + &sysfs_verify, + &sysfs_bypass_torture_test, +#endif + &sysfs_backing_dev_name, + &sysfs_backing_dev_uuid, + NULL +}; +KTYPE(bch_cached_dev); + +SHOW(bch_flash_dev) +{ + struct bcache_device *d = container_of(kobj, struct bcache_device, + kobj); + struct uuid_entry *u = &d->c->uuids[d->id]; + + sysfs_printf(data_csum, "%i", d->data_csum); + sysfs_hprint(size, u->sectors << 9); + + if (attr == &sysfs_label) { + memcpy(buf, u->label, SB_LABEL_SIZE); + buf[SB_LABEL_SIZE + 1] = '\0'; + strcat(buf, "\n"); + return strlen(buf); + } + + return 0; +} + +STORE(__bch_flash_dev) +{ + struct bcache_device *d = container_of(kobj, struct bcache_device, + kobj); + struct uuid_entry *u = &d->c->uuids[d->id]; + + /* no user space access if system is rebooting */ + if (bcache_is_reboot) + return -EBUSY; + + sysfs_strtoul(data_csum, d->data_csum); + + if (attr == &sysfs_size) { + uint64_t v; + + strtoi_h_or_return(buf, v); + + u->sectors = v >> 9; + bch_uuid_write(d->c); + set_capacity(d->disk, u->sectors); + } + + if (attr == &sysfs_label) { + memcpy(u->label, buf, SB_LABEL_SIZE); + bch_uuid_write(d->c); + } + + if (attr == &sysfs_unregister) { + set_bit(BCACHE_DEV_DETACHING, &d->flags); + bcache_device_stop(d); + } + + return size; +} +STORE_LOCKED(bch_flash_dev) + +static struct attribute *bch_flash_dev_files[] = { + &sysfs_unregister, +#if 0 + &sysfs_data_csum, +#endif + &sysfs_label, + &sysfs_size, + NULL +}; +KTYPE(bch_flash_dev); + +struct bset_stats_op { + struct btree_op op; + size_t nodes; + struct bset_stats stats; +}; + +static int bch_btree_bset_stats(struct btree_op *b_op, struct btree *b) +{ + struct bset_stats_op *op = container_of(b_op, struct bset_stats_op, op); + + op->nodes++; + bch_btree_keys_stats(&b->keys, &op->stats); + + return MAP_CONTINUE; +} + +static int bch_bset_print_stats(struct cache_set *c, char *buf) +{ + struct bset_stats_op op; + int ret; + + memset(&op, 0, sizeof(op)); + bch_btree_op_init(&op.op, -1); + + ret = bch_btree_map_nodes(&op.op, c, &ZERO_KEY, bch_btree_bset_stats); + if (ret < 0) + return ret; + + return snprintf(buf, PAGE_SIZE, + "btree nodes: %zu\n" + "written sets: %zu\n" + "unwritten sets: %zu\n" + "written key bytes: %zu\n" + "unwritten key bytes: %zu\n" + "floats: %zu\n" + "failed: %zu\n", + op.nodes, + op.stats.sets_written, op.stats.sets_unwritten, + op.stats.bytes_written, op.stats.bytes_unwritten, + op.stats.floats, op.stats.failed); +} + +static unsigned int bch_root_usage(struct cache_set *c) +{ + unsigned int bytes = 0; + struct bkey *k; + struct btree *b; + struct btree_iter iter; + + goto lock_root; + + do { + rw_unlock(false, b); +lock_root: + b = c->root; + rw_lock(false, b, b->level); + } while (b != c->root); + + for_each_key_filter(&b->keys, k, &iter, bch_ptr_bad) + bytes += bkey_bytes(k); + + rw_unlock(false, b); + + return (bytes * 100) / btree_bytes(c); +} + +static size_t bch_cache_size(struct cache_set *c) +{ + size_t ret = 0; + struct btree *b; + + mutex_lock(&c->bucket_lock); + list_for_each_entry(b, &c->btree_cache, list) + ret += 1 << (b->keys.page_order + PAGE_SHIFT); + + mutex_unlock(&c->bucket_lock); + return ret; +} + +static unsigned int bch_cache_max_chain(struct cache_set *c) +{ + unsigned int ret = 0; + struct hlist_head *h; + + mutex_lock(&c->bucket_lock); + + for (h = c->bucket_hash; + h < c->bucket_hash + (1 << BUCKET_HASH_BITS); + h++) { + unsigned int i = 0; + struct hlist_node *p; + + hlist_for_each(p, h) + i++; + + ret = max(ret, i); + } + + mutex_unlock(&c->bucket_lock); + return ret; +} + +static unsigned int bch_btree_used(struct cache_set *c) +{ + return div64_u64(c->gc_stats.key_bytes * 100, + (c->gc_stats.nodes ?: 1) * btree_bytes(c)); +} + +static unsigned int bch_average_key_size(struct cache_set *c) +{ + return c->gc_stats.nkeys + ? div64_u64(c->gc_stats.data, c->gc_stats.nkeys) + : 0; +} + +SHOW(__bch_cache_set) +{ + struct cache_set *c = container_of(kobj, struct cache_set, kobj); + + sysfs_print(synchronous, CACHE_SYNC(&c->cache->sb)); + sysfs_print(journal_delay_ms, c->journal_delay_ms); + sysfs_hprint(bucket_size, bucket_bytes(c->cache)); + sysfs_hprint(block_size, block_bytes(c->cache)); + sysfs_print(tree_depth, c->root->level); + sysfs_print(root_usage_percent, bch_root_usage(c)); + + sysfs_hprint(btree_cache_size, bch_cache_size(c)); + sysfs_print(btree_cache_max_chain, bch_cache_max_chain(c)); + sysfs_print(cache_available_percent, 100 - c->gc_stats.in_use); + + sysfs_print_time_stats(&c->btree_gc_time, btree_gc, sec, ms); + sysfs_print_time_stats(&c->btree_split_time, btree_split, sec, us); + sysfs_print_time_stats(&c->sort.time, btree_sort, ms, us); + sysfs_print_time_stats(&c->btree_read_time, btree_read, ms, us); + + sysfs_print(btree_used_percent, bch_btree_used(c)); + sysfs_print(btree_nodes, c->gc_stats.nodes); + sysfs_hprint(average_key_size, bch_average_key_size(c)); + + sysfs_print(cache_read_races, + atomic_long_read(&c->cache_read_races)); + + sysfs_print(reclaim, + atomic_long_read(&c->reclaim)); + + sysfs_print(reclaimed_journal_buckets, + atomic_long_read(&c->reclaimed_journal_buckets)); + + sysfs_print(flush_write, + atomic_long_read(&c->flush_write)); + + sysfs_print(writeback_keys_done, + atomic_long_read(&c->writeback_keys_done)); + sysfs_print(writeback_keys_failed, + atomic_long_read(&c->writeback_keys_failed)); + + if (attr == &sysfs_errors) + return bch_snprint_string_list(buf, PAGE_SIZE, error_actions, + c->on_error); + + /* See count_io_errors for why 88 */ + sysfs_print(io_error_halflife, c->error_decay * 88); + sysfs_print(io_error_limit, c->error_limit); + + sysfs_hprint(congested, + ((uint64_t) bch_get_congested(c)) << 9); + sysfs_print(congested_read_threshold_us, + c->congested_read_threshold_us); + sysfs_print(congested_write_threshold_us, + c->congested_write_threshold_us); + + sysfs_print(cutoff_writeback, bch_cutoff_writeback); + sysfs_print(cutoff_writeback_sync, bch_cutoff_writeback_sync); + + sysfs_print(active_journal_entries, fifo_used(&c->journal.pin)); + sysfs_printf(verify, "%i", c->verify); + sysfs_printf(key_merging_disabled, "%i", c->key_merging_disabled); + sysfs_printf(expensive_debug_checks, + "%i", c->expensive_debug_checks); + sysfs_printf(gc_always_rewrite, "%i", c->gc_always_rewrite); + sysfs_printf(btree_shrinker_disabled, "%i", c->shrinker_disabled); + sysfs_printf(copy_gc_enabled, "%i", c->copy_gc_enabled); + sysfs_printf(idle_max_writeback_rate, "%i", + c->idle_max_writeback_rate_enabled); + sysfs_printf(gc_after_writeback, "%i", c->gc_after_writeback); + sysfs_printf(io_disable, "%i", + test_bit(CACHE_SET_IO_DISABLE, &c->flags)); + + if (attr == &sysfs_bset_tree_stats) + return bch_bset_print_stats(c, buf); + + if (attr == &sysfs_feature_compat) + return bch_print_cache_set_feature_compat(c, buf, PAGE_SIZE); + if (attr == &sysfs_feature_ro_compat) + return bch_print_cache_set_feature_ro_compat(c, buf, PAGE_SIZE); + if (attr == &sysfs_feature_incompat) + return bch_print_cache_set_feature_incompat(c, buf, PAGE_SIZE); + + return 0; +} +SHOW_LOCKED(bch_cache_set) + +STORE(__bch_cache_set) +{ + struct cache_set *c = container_of(kobj, struct cache_set, kobj); + ssize_t v; + + /* no user space access if system is rebooting */ + if (bcache_is_reboot) + return -EBUSY; + + if (attr == &sysfs_unregister) + bch_cache_set_unregister(c); + + if (attr == &sysfs_stop) + bch_cache_set_stop(c); + + if (attr == &sysfs_synchronous) { + bool sync = strtoul_or_return(buf); + + if (sync != CACHE_SYNC(&c->cache->sb)) { + SET_CACHE_SYNC(&c->cache->sb, sync); + bcache_write_super(c); + } + } + + if (attr == &sysfs_flash_vol_create) { + int r; + uint64_t v; + + strtoi_h_or_return(buf, v); + + r = bch_flash_dev_create(c, v); + if (r) + return r; + } + + if (attr == &sysfs_clear_stats) { + atomic_long_set(&c->writeback_keys_done, 0); + atomic_long_set(&c->writeback_keys_failed, 0); + + memset(&c->gc_stats, 0, sizeof(struct gc_stat)); + bch_cache_accounting_clear(&c->accounting); + } + + if (attr == &sysfs_trigger_gc) + force_wake_up_gc(c); + + if (attr == &sysfs_prune_cache) { + struct shrink_control sc; + + sc.gfp_mask = GFP_KERNEL; + sc.nr_to_scan = strtoul_or_return(buf); + c->shrink.scan_objects(&c->shrink, &sc); + } + + sysfs_strtoul_clamp(congested_read_threshold_us, + c->congested_read_threshold_us, + 0, UINT_MAX); + sysfs_strtoul_clamp(congested_write_threshold_us, + c->congested_write_threshold_us, + 0, UINT_MAX); + + if (attr == &sysfs_errors) { + v = __sysfs_match_string(error_actions, -1, buf); + if (v < 0) + return v; + + c->on_error = v; + } + + sysfs_strtoul_clamp(io_error_limit, c->error_limit, 0, UINT_MAX); + + /* See count_io_errors() for why 88 */ + if (attr == &sysfs_io_error_halflife) { + unsigned long v = 0; + ssize_t ret; + + ret = strtoul_safe_clamp(buf, v, 0, UINT_MAX); + if (!ret) { + c->error_decay = v / 88; + return size; + } + return ret; + } + + if (attr == &sysfs_io_disable) { + v = strtoul_or_return(buf); + if (v) { + if (test_and_set_bit(CACHE_SET_IO_DISABLE, + &c->flags)) + pr_warn("CACHE_SET_IO_DISABLE already set\n"); + } else { + if (!test_and_clear_bit(CACHE_SET_IO_DISABLE, + &c->flags)) + pr_warn("CACHE_SET_IO_DISABLE already cleared\n"); + } + } + + sysfs_strtoul_clamp(journal_delay_ms, + c->journal_delay_ms, + 0, USHRT_MAX); + sysfs_strtoul_bool(verify, c->verify); + sysfs_strtoul_bool(key_merging_disabled, c->key_merging_disabled); + sysfs_strtoul(expensive_debug_checks, c->expensive_debug_checks); + sysfs_strtoul_bool(gc_always_rewrite, c->gc_always_rewrite); + sysfs_strtoul_bool(btree_shrinker_disabled, c->shrinker_disabled); + sysfs_strtoul_bool(copy_gc_enabled, c->copy_gc_enabled); + sysfs_strtoul_bool(idle_max_writeback_rate, + c->idle_max_writeback_rate_enabled); + + /* + * write gc_after_writeback here may overwrite an already set + * BCH_DO_AUTO_GC, it doesn't matter because this flag will be + * set in next chance. + */ + sysfs_strtoul_clamp(gc_after_writeback, c->gc_after_writeback, 0, 1); + + return size; +} +STORE_LOCKED(bch_cache_set) + +SHOW(bch_cache_set_internal) +{ + struct cache_set *c = container_of(kobj, struct cache_set, internal); + + return bch_cache_set_show(&c->kobj, attr, buf); +} + +STORE(bch_cache_set_internal) +{ + struct cache_set *c = container_of(kobj, struct cache_set, internal); + + /* no user space access if system is rebooting */ + if (bcache_is_reboot) + return -EBUSY; + + return bch_cache_set_store(&c->kobj, attr, buf, size); +} + +static void bch_cache_set_internal_release(struct kobject *k) +{ +} + +static struct attribute *bch_cache_set_files[] = { + &sysfs_unregister, + &sysfs_stop, + &sysfs_synchronous, + &sysfs_journal_delay_ms, + &sysfs_flash_vol_create, + + &sysfs_bucket_size, + &sysfs_block_size, + &sysfs_tree_depth, + &sysfs_root_usage_percent, + &sysfs_btree_cache_size, + &sysfs_cache_available_percent, + + &sysfs_average_key_size, + + &sysfs_errors, + &sysfs_io_error_limit, + &sysfs_io_error_halflife, + &sysfs_congested, + &sysfs_congested_read_threshold_us, + &sysfs_congested_write_threshold_us, + &sysfs_clear_stats, + NULL +}; +KTYPE(bch_cache_set); + +static struct attribute *bch_cache_set_internal_files[] = { + &sysfs_active_journal_entries, + + sysfs_time_stats_attribute_list(btree_gc, sec, ms) + sysfs_time_stats_attribute_list(btree_split, sec, us) + sysfs_time_stats_attribute_list(btree_sort, ms, us) + sysfs_time_stats_attribute_list(btree_read, ms, us) + + &sysfs_btree_nodes, + &sysfs_btree_used_percent, + &sysfs_btree_cache_max_chain, + + &sysfs_bset_tree_stats, + &sysfs_cache_read_races, + &sysfs_reclaim, + &sysfs_reclaimed_journal_buckets, + &sysfs_flush_write, + &sysfs_writeback_keys_done, + &sysfs_writeback_keys_failed, + + &sysfs_trigger_gc, + &sysfs_prune_cache, +#ifdef CONFIG_BCACHE_DEBUG + &sysfs_verify, + &sysfs_key_merging_disabled, + &sysfs_expensive_debug_checks, +#endif + &sysfs_gc_always_rewrite, + &sysfs_btree_shrinker_disabled, + &sysfs_copy_gc_enabled, + &sysfs_idle_max_writeback_rate, + &sysfs_gc_after_writeback, + &sysfs_io_disable, + &sysfs_cutoff_writeback, + &sysfs_cutoff_writeback_sync, + &sysfs_feature_compat, + &sysfs_feature_ro_compat, + &sysfs_feature_incompat, + NULL +}; +KTYPE(bch_cache_set_internal); + +static int __bch_cache_cmp(const void *l, const void *r) +{ + cond_resched(); + return *((uint16_t *)r) - *((uint16_t *)l); +} + +SHOW(__bch_cache) +{ + struct cache *ca = container_of(kobj, struct cache, kobj); + + sysfs_hprint(bucket_size, bucket_bytes(ca)); + sysfs_hprint(block_size, block_bytes(ca)); + sysfs_print(nbuckets, ca->sb.nbuckets); + sysfs_print(discard, ca->discard); + sysfs_hprint(written, atomic_long_read(&ca->sectors_written) << 9); + sysfs_hprint(btree_written, + atomic_long_read(&ca->btree_sectors_written) << 9); + sysfs_hprint(metadata_written, + (atomic_long_read(&ca->meta_sectors_written) + + atomic_long_read(&ca->btree_sectors_written)) << 9); + + sysfs_print(io_errors, + atomic_read(&ca->io_errors) >> IO_ERROR_SHIFT); + + if (attr == &sysfs_cache_replacement_policy) + return bch_snprint_string_list(buf, PAGE_SIZE, + cache_replacement_policies, + CACHE_REPLACEMENT(&ca->sb)); + + if (attr == &sysfs_priority_stats) { + struct bucket *b; + size_t n = ca->sb.nbuckets, i; + size_t unused = 0, available = 0, dirty = 0, meta = 0; + uint64_t sum = 0; + /* Compute 31 quantiles */ + uint16_t q[31], *p, *cached; + ssize_t ret; + + cached = p = vmalloc(array_size(sizeof(uint16_t), + ca->sb.nbuckets)); + if (!p) + return -ENOMEM; + + mutex_lock(&ca->set->bucket_lock); + for_each_bucket(b, ca) { + if (!GC_SECTORS_USED(b)) + unused++; + if (GC_MARK(b) == GC_MARK_RECLAIMABLE) + available++; + if (GC_MARK(b) == GC_MARK_DIRTY) + dirty++; + if (GC_MARK(b) == GC_MARK_METADATA) + meta++; + } + + for (i = ca->sb.first_bucket; i < n; i++) + p[i] = ca->buckets[i].prio; + mutex_unlock(&ca->set->bucket_lock); + + sort(p, n, sizeof(uint16_t), __bch_cache_cmp, NULL); + + while (n && + !cached[n - 1]) + --n; + + while (cached < p + n && + *cached == BTREE_PRIO) + cached++, n--; + + for (i = 0; i < n; i++) + sum += INITIAL_PRIO - cached[i]; + + if (n) + sum = div64_u64(sum, n); + + for (i = 0; i < ARRAY_SIZE(q); i++) + q[i] = INITIAL_PRIO - cached[n * (i + 1) / + (ARRAY_SIZE(q) + 1)]; + + vfree(p); + + ret = scnprintf(buf, PAGE_SIZE, + "Unused: %zu%%\n" + "Clean: %zu%%\n" + "Dirty: %zu%%\n" + "Metadata: %zu%%\n" + "Average: %llu\n" + "Sectors per Q: %zu\n" + "Quantiles: [", + unused * 100 / (size_t) ca->sb.nbuckets, + available * 100 / (size_t) ca->sb.nbuckets, + dirty * 100 / (size_t) ca->sb.nbuckets, + meta * 100 / (size_t) ca->sb.nbuckets, sum, + n * ca->sb.bucket_size / (ARRAY_SIZE(q) + 1)); + + for (i = 0; i < ARRAY_SIZE(q); i++) + ret += scnprintf(buf + ret, PAGE_SIZE - ret, + "%u ", q[i]); + ret--; + + ret += scnprintf(buf + ret, PAGE_SIZE - ret, "]\n"); + + return ret; + } + + return 0; +} +SHOW_LOCKED(bch_cache) + +STORE(__bch_cache) +{ + struct cache *ca = container_of(kobj, struct cache, kobj); + ssize_t v; + + /* no user space access if system is rebooting */ + if (bcache_is_reboot) + return -EBUSY; + + if (attr == &sysfs_discard) { + bool v = strtoul_or_return(buf); + + if (blk_queue_discard(bdev_get_queue(ca->bdev))) + ca->discard = v; + + if (v != CACHE_DISCARD(&ca->sb)) { + SET_CACHE_DISCARD(&ca->sb, v); + bcache_write_super(ca->set); + } + } + + if (attr == &sysfs_cache_replacement_policy) { + v = __sysfs_match_string(cache_replacement_policies, -1, buf); + if (v < 0) + return v; + + if ((unsigned int) v != CACHE_REPLACEMENT(&ca->sb)) { + mutex_lock(&ca->set->bucket_lock); + SET_CACHE_REPLACEMENT(&ca->sb, v); + mutex_unlock(&ca->set->bucket_lock); + + bcache_write_super(ca->set); + } + } + + if (attr == &sysfs_clear_stats) { + atomic_long_set(&ca->sectors_written, 0); + atomic_long_set(&ca->btree_sectors_written, 0); + atomic_long_set(&ca->meta_sectors_written, 0); + atomic_set(&ca->io_count, 0); + atomic_set(&ca->io_errors, 0); + } + + return size; +} +STORE_LOCKED(bch_cache) + +static struct attribute *bch_cache_files[] = { + &sysfs_bucket_size, + &sysfs_block_size, + &sysfs_nbuckets, + &sysfs_priority_stats, + &sysfs_discard, + &sysfs_written, + &sysfs_btree_written, + &sysfs_metadata_written, + &sysfs_io_errors, + &sysfs_clear_stats, + &sysfs_cache_replacement_policy, + NULL +}; +KTYPE(bch_cache); diff --git a/drivers/md/bcache/sysfs.h b/drivers/md/bcache/sysfs.h new file mode 100644 index 000000000..215df32f5 --- /dev/null +++ b/drivers/md/bcache/sysfs.h @@ -0,0 +1,128 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _BCACHE_SYSFS_H_ +#define _BCACHE_SYSFS_H_ + +#define KTYPE(type) \ +struct kobj_type type ## _ktype = { \ + .release = type ## _release, \ + .sysfs_ops = &((const struct sysfs_ops) { \ + .show = type ## _show, \ + .store = type ## _store \ + }), \ + .default_attrs = type ## _files \ +} + +#define SHOW(fn) \ +static ssize_t fn ## _show(struct kobject *kobj, struct attribute *attr,\ + char *buf) \ + +#define STORE(fn) \ +static ssize_t fn ## _store(struct kobject *kobj, struct attribute *attr,\ + const char *buf, size_t size) \ + +#define SHOW_LOCKED(fn) \ +SHOW(fn) \ +{ \ + ssize_t ret; \ + mutex_lock(&bch_register_lock); \ + ret = __ ## fn ## _show(kobj, attr, buf); \ + mutex_unlock(&bch_register_lock); \ + return ret; \ +} + +#define STORE_LOCKED(fn) \ +STORE(fn) \ +{ \ + ssize_t ret; \ + mutex_lock(&bch_register_lock); \ + ret = __ ## fn ## _store(kobj, attr, buf, size); \ + mutex_unlock(&bch_register_lock); \ + return ret; \ +} + +#define __sysfs_attribute(_name, _mode) \ + static struct attribute sysfs_##_name = \ + { .name = #_name, .mode = _mode } + +#define write_attribute(n) __sysfs_attribute(n, 0200) +#define read_attribute(n) __sysfs_attribute(n, 0444) +#define rw_attribute(n) __sysfs_attribute(n, 0644) + +#define sysfs_printf(file, fmt, ...) \ +do { \ + if (attr == &sysfs_ ## file) \ + return snprintf(buf, PAGE_SIZE, fmt "\n", __VA_ARGS__); \ +} while (0) + +#define sysfs_print(file, var) \ +do { \ + if (attr == &sysfs_ ## file) \ + return snprint(buf, PAGE_SIZE, var); \ +} while (0) + +#define sysfs_hprint(file, val) \ +do { \ + if (attr == &sysfs_ ## file) { \ + ssize_t ret = bch_hprint(buf, val); \ + strcat(buf, "\n"); \ + return ret + 1; \ + } \ +} while (0) + +#define var_printf(_var, fmt) sysfs_printf(_var, fmt, var(_var)) +#define var_print(_var) sysfs_print(_var, var(_var)) +#define var_hprint(_var) sysfs_hprint(_var, var(_var)) + +#define sysfs_strtoul(file, var) \ +do { \ + if (attr == &sysfs_ ## file) \ + return strtoul_safe(buf, var) ?: (ssize_t) size; \ +} while (0) + +#define sysfs_strtoul_bool(file, var) \ +do { \ + if (attr == &sysfs_ ## file) { \ + unsigned long v = strtoul_or_return(buf); \ + \ + var = v ? 1 : 0; \ + return size; \ + } \ +} while (0) + +#define sysfs_strtoul_clamp(file, var, min, max) \ +do { \ + if (attr == &sysfs_ ## file) { \ + unsigned long v = 0; \ + ssize_t ret; \ + ret = strtoul_safe_clamp(buf, v, min, max); \ + if (!ret) { \ + var = v; \ + return size; \ + } \ + return ret; \ + } \ +} while (0) + +#define strtoul_or_return(cp) \ +({ \ + unsigned long _v; \ + int _r = kstrtoul(cp, 10, &_v); \ + if (_r) \ + return _r; \ + _v; \ +}) + +#define strtoi_h_or_return(cp, v) \ +do { \ + int _r = strtoi_h(cp, &v); \ + if (_r) \ + return _r; \ +} while (0) + +#define sysfs_hatoi(file, var) \ +do { \ + if (attr == &sysfs_ ## file) \ + return strtoi_h(buf, &var) ?: (ssize_t) size; \ +} while (0) + +#endif /* _BCACHE_SYSFS_H_ */ diff --git a/drivers/md/bcache/trace.c b/drivers/md/bcache/trace.c new file mode 100644 index 000000000..a9a73f560 --- /dev/null +++ b/drivers/md/bcache/trace.c @@ -0,0 +1,53 @@ +// SPDX-License-Identifier: GPL-2.0 +#include "bcache.h" +#include "btree.h" + +#include <linux/blktrace_api.h> +#include <linux/module.h> + +#define CREATE_TRACE_POINTS +#include <trace/events/bcache.h> + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_request_start); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_request_end); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_bypass_sequential); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_bypass_congested); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_read); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_write); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_read_retry); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_cache_insert); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_journal_replay_key); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_journal_write); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_journal_full); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_journal_entry_full); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_cache_cannibalize); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_read); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_write); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_node_alloc); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_node_alloc_fail); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_node_free); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_gc_coalesce); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_gc_start); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_gc_end); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_gc_copy); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_gc_copy_collision); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_insert_key); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_node_split); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_node_compact); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_set_root); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_invalidate); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_alloc_fail); + +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_writeback); +EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_writeback_collision); diff --git a/drivers/md/bcache/util.c b/drivers/md/bcache/util.c new file mode 100644 index 000000000..ae380bc39 --- /dev/null +++ b/drivers/md/bcache/util.c @@ -0,0 +1,287 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * random utiility code, for bcache but in theory not specific to bcache + * + * Copyright 2010, 2011 Kent Overstreet <kent.overstreet@gmail.com> + * Copyright 2012 Google, Inc. + */ + +#include <linux/bio.h> +#include <linux/blkdev.h> +#include <linux/ctype.h> +#include <linux/debugfs.h> +#include <linux/module.h> +#include <linux/seq_file.h> +#include <linux/types.h> +#include <linux/sched/clock.h> + +#include "util.h" + +#define simple_strtoint(c, end, base) simple_strtol(c, end, base) +#define simple_strtouint(c, end, base) simple_strtoul(c, end, base) + +#define STRTO_H(name, type) \ +int bch_ ## name ## _h(const char *cp, type *res) \ +{ \ + int u = 0; \ + char *e; \ + type i = simple_ ## name(cp, &e, 10); \ + \ + switch (tolower(*e)) { \ + default: \ + return -EINVAL; \ + case 'y': \ + case 'z': \ + u++; \ + fallthrough; \ + case 'e': \ + u++; \ + fallthrough; \ + case 'p': \ + u++; \ + fallthrough; \ + case 't': \ + u++; \ + fallthrough; \ + case 'g': \ + u++; \ + fallthrough; \ + case 'm': \ + u++; \ + fallthrough; \ + case 'k': \ + u++; \ + if (e++ == cp) \ + return -EINVAL; \ + fallthrough; \ + case '\n': \ + case '\0': \ + if (*e == '\n') \ + e++; \ + } \ + \ + if (*e) \ + return -EINVAL; \ + \ + while (u--) { \ + if ((type) ~0 > 0 && \ + (type) ~0 / 1024 <= i) \ + return -EINVAL; \ + if ((i > 0 && ANYSINT_MAX(type) / 1024 < i) || \ + (i < 0 && -ANYSINT_MAX(type) / 1024 > i)) \ + return -EINVAL; \ + i *= 1024; \ + } \ + \ + *res = i; \ + return 0; \ +} \ + +STRTO_H(strtoint, int) +STRTO_H(strtouint, unsigned int) +STRTO_H(strtoll, long long) +STRTO_H(strtoull, unsigned long long) + +/** + * bch_hprint - formats @v to human readable string for sysfs. + * @buf: the (at least 8 byte) buffer to format the result into. + * @v: signed 64 bit integer + * + * Returns the number of bytes used by format. + */ +ssize_t bch_hprint(char *buf, int64_t v) +{ + static const char units[] = "?kMGTPEZY"; + int u = 0, t; + + uint64_t q; + + if (v < 0) + q = -v; + else + q = v; + + /* For as long as the number is more than 3 digits, but at least + * once, shift right / divide by 1024. Keep the remainder for + * a digit after the decimal point. + */ + do { + u++; + + t = q & ~(~0 << 10); + q >>= 10; + } while (q >= 1000); + + if (v < 0) + /* '-', up to 3 digits, '.', 1 digit, 1 character, null; + * yields 8 bytes. + */ + return sprintf(buf, "-%llu.%i%c", q, t * 10 / 1024, units[u]); + else + return sprintf(buf, "%llu.%i%c", q, t * 10 / 1024, units[u]); +} + +bool bch_is_zero(const char *p, size_t n) +{ + size_t i; + + for (i = 0; i < n; i++) + if (p[i]) + return false; + return true; +} + +int bch_parse_uuid(const char *s, char *uuid) +{ + size_t i, j, x; + + memset(uuid, 0, 16); + + for (i = 0, j = 0; + i < strspn(s, "-0123456789:ABCDEFabcdef") && j < 32; + i++) { + x = s[i] | 32; + + switch (x) { + case '0'...'9': + x -= '0'; + break; + case 'a'...'f': + x -= 'a' - 10; + break; + default: + continue; + } + + if (!(j & 1)) + x <<= 4; + uuid[j++ >> 1] |= x; + } + return i; +} + +void bch_time_stats_update(struct time_stats *stats, uint64_t start_time) +{ + uint64_t now, duration, last; + + spin_lock(&stats->lock); + + now = local_clock(); + duration = time_after64(now, start_time) + ? now - start_time : 0; + last = time_after64(now, stats->last) + ? now - stats->last : 0; + + stats->max_duration = max(stats->max_duration, duration); + + if (stats->last) { + ewma_add(stats->average_duration, duration, 8, 8); + + if (stats->average_frequency) + ewma_add(stats->average_frequency, last, 8, 8); + else + stats->average_frequency = last << 8; + } else { + stats->average_duration = duration << 8; + } + + stats->last = now ?: 1; + + spin_unlock(&stats->lock); +} + +/** + * bch_next_delay() - update ratelimiting statistics and calculate next delay + * @d: the struct bch_ratelimit to update + * @done: the amount of work done, in arbitrary units + * + * Increment @d by the amount of work done, and return how long to delay in + * jiffies until the next time to do some work. + */ +uint64_t bch_next_delay(struct bch_ratelimit *d, uint64_t done) +{ + uint64_t now = local_clock(); + + d->next += div_u64(done * NSEC_PER_SEC, atomic_long_read(&d->rate)); + + /* Bound the time. Don't let us fall further than 2 seconds behind + * (this prevents unnecessary backlog that would make it impossible + * to catch up). If we're ahead of the desired writeback rate, + * don't let us sleep more than 2.5 seconds (so we can notice/respond + * if the control system tells us to speed up!). + */ + if (time_before64(now + NSEC_PER_SEC * 5LLU / 2LLU, d->next)) + d->next = now + NSEC_PER_SEC * 5LLU / 2LLU; + + if (time_after64(now - NSEC_PER_SEC * 2, d->next)) + d->next = now - NSEC_PER_SEC * 2; + + return time_after64(d->next, now) + ? div_u64(d->next - now, NSEC_PER_SEC / HZ) + : 0; +} + +/* + * Generally it isn't good to access .bi_io_vec and .bi_vcnt directly, + * the preferred way is bio_add_page, but in this case, bch_bio_map() + * supposes that the bvec table is empty, so it is safe to access + * .bi_vcnt & .bi_io_vec in this way even after multipage bvec is + * supported. + */ +void bch_bio_map(struct bio *bio, void *base) +{ + size_t size = bio->bi_iter.bi_size; + struct bio_vec *bv = bio->bi_io_vec; + + BUG_ON(!bio->bi_iter.bi_size); + BUG_ON(bio->bi_vcnt); + + bv->bv_offset = base ? offset_in_page(base) : 0; + goto start; + + for (; size; bio->bi_vcnt++, bv++) { + bv->bv_offset = 0; +start: bv->bv_len = min_t(size_t, PAGE_SIZE - bv->bv_offset, + size); + if (base) { + bv->bv_page = is_vmalloc_addr(base) + ? vmalloc_to_page(base) + : virt_to_page(base); + + base += bv->bv_len; + } + + size -= bv->bv_len; + } +} + +/** + * bch_bio_alloc_pages - allocates a single page for each bvec in a bio + * @bio: bio to allocate pages for + * @gfp_mask: flags for allocation + * + * Allocates pages up to @bio->bi_vcnt. + * + * Returns 0 on success, -ENOMEM on failure. On failure, any allocated pages are + * freed. + */ +int bch_bio_alloc_pages(struct bio *bio, gfp_t gfp_mask) +{ + int i; + struct bio_vec *bv; + + /* + * This is called on freshly new bio, so it is safe to access the + * bvec table directly. + */ + for (i = 0, bv = bio->bi_io_vec; i < bio->bi_vcnt; bv++, i++) { + bv->bv_page = alloc_page(gfp_mask); + if (!bv->bv_page) { + while (--bv >= bio->bi_io_vec) + __free_page(bv->bv_page); + return -ENOMEM; + } + } + + return 0; +} diff --git a/drivers/md/bcache/util.h b/drivers/md/bcache/util.h new file mode 100644 index 000000000..c029f7443 --- /dev/null +++ b/drivers/md/bcache/util.h @@ -0,0 +1,593 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +#ifndef _BCACHE_UTIL_H +#define _BCACHE_UTIL_H + +#include <linux/blkdev.h> +#include <linux/errno.h> +#include <linux/kernel.h> +#include <linux/sched/clock.h> +#include <linux/llist.h> +#include <linux/ratelimit.h> +#include <linux/vmalloc.h> +#include <linux/workqueue.h> +#include <linux/crc64.h> + +#include "closure.h" + +#define PAGE_SECTORS (PAGE_SIZE / 512) + +struct closure; + +#ifdef CONFIG_BCACHE_DEBUG + +#define EBUG_ON(cond) BUG_ON(cond) +#define atomic_dec_bug(v) BUG_ON(atomic_dec_return(v) < 0) +#define atomic_inc_bug(v, i) BUG_ON(atomic_inc_return(v) <= i) + +#else /* DEBUG */ + +#define EBUG_ON(cond) do { if (cond); } while (0) +#define atomic_dec_bug(v) atomic_dec(v) +#define atomic_inc_bug(v, i) atomic_inc(v) + +#endif + +#define DECLARE_HEAP(type, name) \ + struct { \ + size_t size, used; \ + type *data; \ + } name + +#define init_heap(heap, _size, gfp) \ +({ \ + size_t _bytes; \ + (heap)->used = 0; \ + (heap)->size = (_size); \ + _bytes = (heap)->size * sizeof(*(heap)->data); \ + (heap)->data = kvmalloc(_bytes, (gfp) & GFP_KERNEL); \ + (heap)->data; \ +}) + +#define free_heap(heap) \ +do { \ + kvfree((heap)->data); \ + (heap)->data = NULL; \ +} while (0) + +#define heap_swap(h, i, j) swap((h)->data[i], (h)->data[j]) + +#define heap_sift(h, i, cmp) \ +do { \ + size_t _r, _j = i; \ + \ + for (; _j * 2 + 1 < (h)->used; _j = _r) { \ + _r = _j * 2 + 1; \ + if (_r + 1 < (h)->used && \ + cmp((h)->data[_r], (h)->data[_r + 1])) \ + _r++; \ + \ + if (cmp((h)->data[_r], (h)->data[_j])) \ + break; \ + heap_swap(h, _r, _j); \ + } \ +} while (0) + +#define heap_sift_down(h, i, cmp) \ +do { \ + while (i) { \ + size_t p = (i - 1) / 2; \ + if (cmp((h)->data[i], (h)->data[p])) \ + break; \ + heap_swap(h, i, p); \ + i = p; \ + } \ +} while (0) + +#define heap_add(h, d, cmp) \ +({ \ + bool _r = !heap_full(h); \ + if (_r) { \ + size_t _i = (h)->used++; \ + (h)->data[_i] = d; \ + \ + heap_sift_down(h, _i, cmp); \ + heap_sift(h, _i, cmp); \ + } \ + _r; \ +}) + +#define heap_pop(h, d, cmp) \ +({ \ + bool _r = (h)->used; \ + if (_r) { \ + (d) = (h)->data[0]; \ + (h)->used--; \ + heap_swap(h, 0, (h)->used); \ + heap_sift(h, 0, cmp); \ + } \ + _r; \ +}) + +#define heap_peek(h) ((h)->used ? (h)->data[0] : NULL) + +#define heap_full(h) ((h)->used == (h)->size) + +#define DECLARE_FIFO(type, name) \ + struct { \ + size_t front, back, size, mask; \ + type *data; \ + } name + +#define fifo_for_each(c, fifo, iter) \ + for (iter = (fifo)->front; \ + c = (fifo)->data[iter], iter != (fifo)->back; \ + iter = (iter + 1) & (fifo)->mask) + +#define __init_fifo(fifo, gfp) \ +({ \ + size_t _allocated_size, _bytes; \ + BUG_ON(!(fifo)->size); \ + \ + _allocated_size = roundup_pow_of_two((fifo)->size + 1); \ + _bytes = _allocated_size * sizeof(*(fifo)->data); \ + \ + (fifo)->mask = _allocated_size - 1; \ + (fifo)->front = (fifo)->back = 0; \ + \ + (fifo)->data = kvmalloc(_bytes, (gfp) & GFP_KERNEL); \ + (fifo)->data; \ +}) + +#define init_fifo_exact(fifo, _size, gfp) \ +({ \ + (fifo)->size = (_size); \ + __init_fifo(fifo, gfp); \ +}) + +#define init_fifo(fifo, _size, gfp) \ +({ \ + (fifo)->size = (_size); \ + if ((fifo)->size > 4) \ + (fifo)->size = roundup_pow_of_two((fifo)->size) - 1; \ + __init_fifo(fifo, gfp); \ +}) + +#define free_fifo(fifo) \ +do { \ + kvfree((fifo)->data); \ + (fifo)->data = NULL; \ +} while (0) + +#define fifo_used(fifo) (((fifo)->back - (fifo)->front) & (fifo)->mask) +#define fifo_free(fifo) ((fifo)->size - fifo_used(fifo)) + +#define fifo_empty(fifo) (!fifo_used(fifo)) +#define fifo_full(fifo) (!fifo_free(fifo)) + +#define fifo_front(fifo) ((fifo)->data[(fifo)->front]) +#define fifo_back(fifo) \ + ((fifo)->data[((fifo)->back - 1) & (fifo)->mask]) + +#define fifo_idx(fifo, p) (((p) - &fifo_front(fifo)) & (fifo)->mask) + +#define fifo_push_back(fifo, i) \ +({ \ + bool _r = !fifo_full((fifo)); \ + if (_r) { \ + (fifo)->data[(fifo)->back++] = (i); \ + (fifo)->back &= (fifo)->mask; \ + } \ + _r; \ +}) + +#define fifo_pop_front(fifo, i) \ +({ \ + bool _r = !fifo_empty((fifo)); \ + if (_r) { \ + (i) = (fifo)->data[(fifo)->front++]; \ + (fifo)->front &= (fifo)->mask; \ + } \ + _r; \ +}) + +#define fifo_push_front(fifo, i) \ +({ \ + bool _r = !fifo_full((fifo)); \ + if (_r) { \ + --(fifo)->front; \ + (fifo)->front &= (fifo)->mask; \ + (fifo)->data[(fifo)->front] = (i); \ + } \ + _r; \ +}) + +#define fifo_pop_back(fifo, i) \ +({ \ + bool _r = !fifo_empty((fifo)); \ + if (_r) { \ + --(fifo)->back; \ + (fifo)->back &= (fifo)->mask; \ + (i) = (fifo)->data[(fifo)->back] \ + } \ + _r; \ +}) + +#define fifo_push(fifo, i) fifo_push_back(fifo, (i)) +#define fifo_pop(fifo, i) fifo_pop_front(fifo, (i)) + +#define fifo_swap(l, r) \ +do { \ + swap((l)->front, (r)->front); \ + swap((l)->back, (r)->back); \ + swap((l)->size, (r)->size); \ + swap((l)->mask, (r)->mask); \ + swap((l)->data, (r)->data); \ +} while (0) + +#define fifo_move(dest, src) \ +do { \ + typeof(*((dest)->data)) _t; \ + while (!fifo_full(dest) && \ + fifo_pop(src, _t)) \ + fifo_push(dest, _t); \ +} while (0) + +/* + * Simple array based allocator - preallocates a number of elements and you can + * never allocate more than that, also has no locking. + * + * Handy because if you know you only need a fixed number of elements you don't + * have to worry about memory allocation failure, and sometimes a mempool isn't + * what you want. + * + * We treat the free elements as entries in a singly linked list, and the + * freelist as a stack - allocating and freeing push and pop off the freelist. + */ + +#define DECLARE_ARRAY_ALLOCATOR(type, name, size) \ + struct { \ + type *freelist; \ + type data[size]; \ + } name + +#define array_alloc(array) \ +({ \ + typeof((array)->freelist) _ret = (array)->freelist; \ + \ + if (_ret) \ + (array)->freelist = *((typeof((array)->freelist) *) _ret);\ + \ + _ret; \ +}) + +#define array_free(array, ptr) \ +do { \ + typeof((array)->freelist) _ptr = ptr; \ + \ + *((typeof((array)->freelist) *) _ptr) = (array)->freelist; \ + (array)->freelist = _ptr; \ +} while (0) + +#define array_allocator_init(array) \ +do { \ + typeof((array)->freelist) _i; \ + \ + BUILD_BUG_ON(sizeof((array)->data[0]) < sizeof(void *)); \ + (array)->freelist = NULL; \ + \ + for (_i = (array)->data; \ + _i < (array)->data + ARRAY_SIZE((array)->data); \ + _i++) \ + array_free(array, _i); \ +} while (0) + +#define array_freelist_empty(array) ((array)->freelist == NULL) + +#define ANYSINT_MAX(t) \ + ((((t) 1 << (sizeof(t) * 8 - 2)) - (t) 1) * (t) 2 + (t) 1) + +int bch_strtoint_h(const char *cp, int *res); +int bch_strtouint_h(const char *cp, unsigned int *res); +int bch_strtoll_h(const char *cp, long long *res); +int bch_strtoull_h(const char *cp, unsigned long long *res); + +static inline int bch_strtol_h(const char *cp, long *res) +{ +#if BITS_PER_LONG == 32 + return bch_strtoint_h(cp, (int *) res); +#else + return bch_strtoll_h(cp, (long long *) res); +#endif +} + +static inline int bch_strtoul_h(const char *cp, long *res) +{ +#if BITS_PER_LONG == 32 + return bch_strtouint_h(cp, (unsigned int *) res); +#else + return bch_strtoull_h(cp, (unsigned long long *) res); +#endif +} + +#define strtoi_h(cp, res) \ + (__builtin_types_compatible_p(typeof(*res), int) \ + ? bch_strtoint_h(cp, (void *) res) \ + : __builtin_types_compatible_p(typeof(*res), long) \ + ? bch_strtol_h(cp, (void *) res) \ + : __builtin_types_compatible_p(typeof(*res), long long) \ + ? bch_strtoll_h(cp, (void *) res) \ + : __builtin_types_compatible_p(typeof(*res), unsigned int) \ + ? bch_strtouint_h(cp, (void *) res) \ + : __builtin_types_compatible_p(typeof(*res), unsigned long) \ + ? bch_strtoul_h(cp, (void *) res) \ + : __builtin_types_compatible_p(typeof(*res), unsigned long long)\ + ? bch_strtoull_h(cp, (void *) res) : -EINVAL) + +#define strtoul_safe(cp, var) \ +({ \ + unsigned long _v; \ + int _r = kstrtoul(cp, 10, &_v); \ + if (!_r) \ + var = _v; \ + _r; \ +}) + +#define strtoul_safe_clamp(cp, var, min, max) \ +({ \ + unsigned long _v; \ + int _r = kstrtoul(cp, 10, &_v); \ + if (!_r) \ + var = clamp_t(typeof(var), _v, min, max); \ + _r; \ +}) + +#define snprint(buf, size, var) \ + snprintf(buf, size, \ + __builtin_types_compatible_p(typeof(var), int) \ + ? "%i\n" : \ + __builtin_types_compatible_p(typeof(var), unsigned int) \ + ? "%u\n" : \ + __builtin_types_compatible_p(typeof(var), long) \ + ? "%li\n" : \ + __builtin_types_compatible_p(typeof(var), unsigned long)\ + ? "%lu\n" : \ + __builtin_types_compatible_p(typeof(var), int64_t) \ + ? "%lli\n" : \ + __builtin_types_compatible_p(typeof(var), uint64_t) \ + ? "%llu\n" : \ + __builtin_types_compatible_p(typeof(var), const char *) \ + ? "%s\n" : "%i\n", var) + +ssize_t bch_hprint(char *buf, int64_t v); + +bool bch_is_zero(const char *p, size_t n); +int bch_parse_uuid(const char *s, char *uuid); + +struct time_stats { + spinlock_t lock; + /* + * all fields are in nanoseconds, averages are ewmas stored left shifted + * by 8 + */ + uint64_t max_duration; + uint64_t average_duration; + uint64_t average_frequency; + uint64_t last; +}; + +void bch_time_stats_update(struct time_stats *stats, uint64_t time); + +static inline unsigned int local_clock_us(void) +{ + return local_clock() >> 10; +} + +#define NSEC_PER_ns 1L +#define NSEC_PER_us NSEC_PER_USEC +#define NSEC_PER_ms NSEC_PER_MSEC +#define NSEC_PER_sec NSEC_PER_SEC + +#define __print_time_stat(stats, name, stat, units) \ + sysfs_print(name ## _ ## stat ## _ ## units, \ + div_u64((stats)->stat >> 8, NSEC_PER_ ## units)) + +#define sysfs_print_time_stats(stats, name, \ + frequency_units, \ + duration_units) \ +do { \ + __print_time_stat(stats, name, \ + average_frequency, frequency_units); \ + __print_time_stat(stats, name, \ + average_duration, duration_units); \ + sysfs_print(name ## _ ##max_duration ## _ ## duration_units, \ + div_u64((stats)->max_duration, \ + NSEC_PER_ ## duration_units)); \ + \ + sysfs_print(name ## _last_ ## frequency_units, (stats)->last \ + ? div_s64(local_clock() - (stats)->last, \ + NSEC_PER_ ## frequency_units) \ + : -1LL); \ +} while (0) + +#define sysfs_time_stats_attribute(name, \ + frequency_units, \ + duration_units) \ +read_attribute(name ## _average_frequency_ ## frequency_units); \ +read_attribute(name ## _average_duration_ ## duration_units); \ +read_attribute(name ## _max_duration_ ## duration_units); \ +read_attribute(name ## _last_ ## frequency_units) + +#define sysfs_time_stats_attribute_list(name, \ + frequency_units, \ + duration_units) \ +&sysfs_ ## name ## _average_frequency_ ## frequency_units, \ +&sysfs_ ## name ## _average_duration_ ## duration_units, \ +&sysfs_ ## name ## _max_duration_ ## duration_units, \ +&sysfs_ ## name ## _last_ ## frequency_units, + +#define ewma_add(ewma, val, weight, factor) \ +({ \ + (ewma) *= (weight) - 1; \ + (ewma) += (val) << factor; \ + (ewma) /= (weight); \ + (ewma) >> factor; \ +}) + +struct bch_ratelimit { + /* Next time we want to do some work, in nanoseconds */ + uint64_t next; + + /* + * Rate at which we want to do work, in units per second + * The units here correspond to the units passed to bch_next_delay() + */ + atomic_long_t rate; +}; + +static inline void bch_ratelimit_reset(struct bch_ratelimit *d) +{ + d->next = local_clock(); +} + +uint64_t bch_next_delay(struct bch_ratelimit *d, uint64_t done); + +#define __DIV_SAFE(n, d, zero) \ +({ \ + typeof(n) _n = (n); \ + typeof(d) _d = (d); \ + _d ? _n / _d : zero; \ +}) + +#define DIV_SAFE(n, d) __DIV_SAFE(n, d, 0) + +#define container_of_or_null(ptr, type, member) \ +({ \ + typeof(ptr) _ptr = ptr; \ + _ptr ? container_of(_ptr, type, member) : NULL; \ +}) + +#define RB_INSERT(root, new, member, cmp) \ +({ \ + __label__ dup; \ + struct rb_node **n = &(root)->rb_node, *parent = NULL; \ + typeof(new) this; \ + int res, ret = -1; \ + \ + while (*n) { \ + parent = *n; \ + this = container_of(*n, typeof(*(new)), member); \ + res = cmp(new, this); \ + if (!res) \ + goto dup; \ + n = res < 0 \ + ? &(*n)->rb_left \ + : &(*n)->rb_right; \ + } \ + \ + rb_link_node(&(new)->member, parent, n); \ + rb_insert_color(&(new)->member, root); \ + ret = 0; \ +dup: \ + ret; \ +}) + +#define RB_SEARCH(root, search, member, cmp) \ +({ \ + struct rb_node *n = (root)->rb_node; \ + typeof(&(search)) this, ret = NULL; \ + int res; \ + \ + while (n) { \ + this = container_of(n, typeof(search), member); \ + res = cmp(&(search), this); \ + if (!res) { \ + ret = this; \ + break; \ + } \ + n = res < 0 \ + ? n->rb_left \ + : n->rb_right; \ + } \ + ret; \ +}) + +#define RB_GREATER(root, search, member, cmp) \ +({ \ + struct rb_node *n = (root)->rb_node; \ + typeof(&(search)) this, ret = NULL; \ + int res; \ + \ + while (n) { \ + this = container_of(n, typeof(search), member); \ + res = cmp(&(search), this); \ + if (res < 0) { \ + ret = this; \ + n = n->rb_left; \ + } else \ + n = n->rb_right; \ + } \ + ret; \ +}) + +#define RB_FIRST(root, type, member) \ + container_of_or_null(rb_first(root), type, member) + +#define RB_LAST(root, type, member) \ + container_of_or_null(rb_last(root), type, member) + +#define RB_NEXT(ptr, member) \ + container_of_or_null(rb_next(&(ptr)->member), typeof(*ptr), member) + +#define RB_PREV(ptr, member) \ + container_of_or_null(rb_prev(&(ptr)->member), typeof(*ptr), member) + +static inline uint64_t bch_crc64(const void *p, size_t len) +{ + uint64_t crc = 0xffffffffffffffffULL; + + crc = crc64_be(crc, p, len); + return crc ^ 0xffffffffffffffffULL; +} + +static inline uint64_t bch_crc64_update(uint64_t crc, + const void *p, + size_t len) +{ + crc = crc64_be(crc, p, len); + return crc; +} + +/* + * A stepwise-linear pseudo-exponential. This returns 1 << (x >> + * frac_bits), with the less-significant bits filled in by linear + * interpolation. + * + * This can also be interpreted as a floating-point number format, + * where the low frac_bits are the mantissa (with implicit leading + * 1 bit), and the more significant bits are the exponent. + * The return value is 1.mantissa * 2^exponent. + * + * The way this is used, fract_bits is 6 and the largest possible + * input is CONGESTED_MAX-1 = 1023 (exponent 16, mantissa 0x1.fc), + * so the maximum output is 0x1fc00. + */ +static inline unsigned int fract_exp_two(unsigned int x, + unsigned int fract_bits) +{ + unsigned int mantissa = 1 << fract_bits; /* Implicit bit */ + + mantissa += x & (mantissa - 1); + x >>= fract_bits; /* The exponent */ + /* Largest intermediate value 0x7f0000 */ + return mantissa << x >> fract_bits; +} + +void bch_bio_map(struct bio *bio, void *base); +int bch_bio_alloc_pages(struct bio *bio, gfp_t gfp_mask); + +static inline sector_t bdev_sectors(struct block_device *bdev) +{ + return bdev->bd_inode->i_size >> 9; +} +#endif /* _BCACHE_UTIL_H */ diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c new file mode 100644 index 000000000..8e3f5f004 --- /dev/null +++ b/drivers/md/bcache/writeback.c @@ -0,0 +1,1037 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * background writeback - scan btree for dirty data and write it to the backing + * device + * + * Copyright 2010, 2011 Kent Overstreet <kent.overstreet@gmail.com> + * Copyright 2012 Google, Inc. + */ + +#include "bcache.h" +#include "btree.h" +#include "debug.h" +#include "writeback.h" + +#include <linux/delay.h> +#include <linux/kthread.h> +#include <linux/sched/clock.h> +#include <trace/events/bcache.h> + +static void update_gc_after_writeback(struct cache_set *c) +{ + if (c->gc_after_writeback != (BCH_ENABLE_AUTO_GC) || + c->gc_stats.in_use < BCH_AUTO_GC_DIRTY_THRESHOLD) + return; + + c->gc_after_writeback |= BCH_DO_AUTO_GC; +} + +/* Rate limiting */ +static uint64_t __calc_target_rate(struct cached_dev *dc) +{ + struct cache_set *c = dc->disk.c; + + /* + * This is the size of the cache, minus the amount used for + * flash-only devices + */ + uint64_t cache_sectors = c->nbuckets * c->cache->sb.bucket_size - + atomic_long_read(&c->flash_dev_dirty_sectors); + + /* + * Unfortunately there is no control of global dirty data. If the + * user states that they want 10% dirty data in the cache, and has, + * e.g., 5 backing volumes of equal size, we try and ensure each + * backing volume uses about 2% of the cache for dirty data. + */ + uint32_t bdev_share = + div64_u64(bdev_sectors(dc->bdev) << WRITEBACK_SHARE_SHIFT, + c->cached_dev_sectors); + + uint64_t cache_dirty_target = + div_u64(cache_sectors * dc->writeback_percent, 100); + + /* Ensure each backing dev gets at least one dirty share */ + if (bdev_share < 1) + bdev_share = 1; + + return (cache_dirty_target * bdev_share) >> WRITEBACK_SHARE_SHIFT; +} + +static void __update_writeback_rate(struct cached_dev *dc) +{ + /* + * PI controller: + * Figures out the amount that should be written per second. + * + * First, the error (number of sectors that are dirty beyond our + * target) is calculated. The error is accumulated (numerically + * integrated). + * + * Then, the proportional value and integral value are scaled + * based on configured values. These are stored as inverses to + * avoid fixed point math and to make configuration easy-- e.g. + * the default value of 40 for writeback_rate_p_term_inverse + * attempts to write at a rate that would retire all the dirty + * blocks in 40 seconds. + * + * The writeback_rate_i_inverse value of 10000 means that 1/10000th + * of the error is accumulated in the integral term per second. + * This acts as a slow, long-term average that is not subject to + * variations in usage like the p term. + */ + int64_t target = __calc_target_rate(dc); + int64_t dirty = bcache_dev_sectors_dirty(&dc->disk); + int64_t error = dirty - target; + int64_t proportional_scaled = + div_s64(error, dc->writeback_rate_p_term_inverse); + int64_t integral_scaled; + uint32_t new_rate; + + if ((error < 0 && dc->writeback_rate_integral > 0) || + (error > 0 && time_before64(local_clock(), + dc->writeback_rate.next + NSEC_PER_MSEC))) { + /* + * Only decrease the integral term if it's more than + * zero. Only increase the integral term if the device + * is keeping up. (Don't wind up the integral + * ineffectively in either case). + * + * It's necessary to scale this by + * writeback_rate_update_seconds to keep the integral + * term dimensioned properly. + */ + dc->writeback_rate_integral += error * + dc->writeback_rate_update_seconds; + } + + integral_scaled = div_s64(dc->writeback_rate_integral, + dc->writeback_rate_i_term_inverse); + + new_rate = clamp_t(int32_t, (proportional_scaled + integral_scaled), + dc->writeback_rate_minimum, NSEC_PER_SEC); + + dc->writeback_rate_proportional = proportional_scaled; + dc->writeback_rate_integral_scaled = integral_scaled; + dc->writeback_rate_change = new_rate - + atomic_long_read(&dc->writeback_rate.rate); + atomic_long_set(&dc->writeback_rate.rate, new_rate); + dc->writeback_rate_target = target; +} + +static bool idle_counter_exceeded(struct cache_set *c) +{ + int counter, dev_nr; + + /* + * If c->idle_counter is overflow (idel for really long time), + * reset as 0 and not set maximum rate this time for code + * simplicity. + */ + counter = atomic_inc_return(&c->idle_counter); + if (counter <= 0) { + atomic_set(&c->idle_counter, 0); + return false; + } + + dev_nr = atomic_read(&c->attached_dev_nr); + if (dev_nr == 0) + return false; + + /* + * c->idle_counter is increased by writeback thread of all + * attached backing devices, in order to represent a rough + * time period, counter should be divided by dev_nr. + * Otherwise the idle time cannot be larger with more backing + * device attached. + * The following calculation equals to checking + * (counter / dev_nr) < (dev_nr * 6) + */ + if (counter < (dev_nr * dev_nr * 6)) + return false; + + return true; +} + +/* + * Idle_counter is increased every time when update_writeback_rate() is + * called. If all backing devices attached to the same cache set have + * identical dc->writeback_rate_update_seconds values, it is about 6 + * rounds of update_writeback_rate() on each backing device before + * c->at_max_writeback_rate is set to 1, and then max wrteback rate set + * to each dc->writeback_rate.rate. + * In order to avoid extra locking cost for counting exact dirty cached + * devices number, c->attached_dev_nr is used to calculate the idle + * throushold. It might be bigger if not all cached device are in write- + * back mode, but it still works well with limited extra rounds of + * update_writeback_rate(). + */ +static bool set_at_max_writeback_rate(struct cache_set *c, + struct cached_dev *dc) +{ + /* Don't sst max writeback rate if it is disabled */ + if (!c->idle_max_writeback_rate_enabled) + return false; + + /* Don't set max writeback rate if gc is running */ + if (!c->gc_mark_valid) + return false; + + if (!idle_counter_exceeded(c)) + return false; + + if (atomic_read(&c->at_max_writeback_rate) != 1) + atomic_set(&c->at_max_writeback_rate, 1); + + atomic_long_set(&dc->writeback_rate.rate, INT_MAX); + + /* keep writeback_rate_target as existing value */ + dc->writeback_rate_proportional = 0; + dc->writeback_rate_integral_scaled = 0; + dc->writeback_rate_change = 0; + + /* + * In case new I/O arrives during before + * set_at_max_writeback_rate() returns. + */ + if (!idle_counter_exceeded(c) || + !atomic_read(&c->at_max_writeback_rate)) + return false; + + return true; +} + +static void update_writeback_rate(struct work_struct *work) +{ + struct cached_dev *dc = container_of(to_delayed_work(work), + struct cached_dev, + writeback_rate_update); + struct cache_set *c = dc->disk.c; + + /* + * should check BCACHE_DEV_RATE_DW_RUNNING before calling + * cancel_delayed_work_sync(). + */ + set_bit(BCACHE_DEV_RATE_DW_RUNNING, &dc->disk.flags); + /* paired with where BCACHE_DEV_RATE_DW_RUNNING is tested */ + smp_mb__after_atomic(); + + /* + * CACHE_SET_IO_DISABLE might be set via sysfs interface, + * check it here too. + */ + if (!test_bit(BCACHE_DEV_WB_RUNNING, &dc->disk.flags) || + test_bit(CACHE_SET_IO_DISABLE, &c->flags)) { + clear_bit(BCACHE_DEV_RATE_DW_RUNNING, &dc->disk.flags); + /* paired with where BCACHE_DEV_RATE_DW_RUNNING is tested */ + smp_mb__after_atomic(); + return; + } + + if (atomic_read(&dc->has_dirty) && dc->writeback_percent) { + /* + * If the whole cache set is idle, set_at_max_writeback_rate() + * will set writeback rate to a max number. Then it is + * unncessary to update writeback rate for an idle cache set + * in maximum writeback rate number(s). + */ + if (!set_at_max_writeback_rate(c, dc)) { + down_read(&dc->writeback_lock); + __update_writeback_rate(dc); + update_gc_after_writeback(c); + up_read(&dc->writeback_lock); + } + } + + + /* + * CACHE_SET_IO_DISABLE might be set via sysfs interface, + * check it here too. + */ + if (test_bit(BCACHE_DEV_WB_RUNNING, &dc->disk.flags) && + !test_bit(CACHE_SET_IO_DISABLE, &c->flags)) { + schedule_delayed_work(&dc->writeback_rate_update, + dc->writeback_rate_update_seconds * HZ); + } + + /* + * should check BCACHE_DEV_RATE_DW_RUNNING before calling + * cancel_delayed_work_sync(). + */ + clear_bit(BCACHE_DEV_RATE_DW_RUNNING, &dc->disk.flags); + /* paired with where BCACHE_DEV_RATE_DW_RUNNING is tested */ + smp_mb__after_atomic(); +} + +static unsigned int writeback_delay(struct cached_dev *dc, + unsigned int sectors) +{ + if (test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags) || + !dc->writeback_percent) + return 0; + + return bch_next_delay(&dc->writeback_rate, sectors); +} + +struct dirty_io { + struct closure cl; + struct cached_dev *dc; + uint16_t sequence; + struct bio bio; +}; + +static void dirty_init(struct keybuf_key *w) +{ + struct dirty_io *io = w->private; + struct bio *bio = &io->bio; + + bio_init(bio, bio->bi_inline_vecs, + DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS)); + if (!io->dc->writeback_percent) + bio_set_prio(bio, IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0)); + + bio->bi_iter.bi_size = KEY_SIZE(&w->key) << 9; + bio->bi_private = w; + bch_bio_map(bio, NULL); +} + +static void dirty_io_destructor(struct closure *cl) +{ + struct dirty_io *io = container_of(cl, struct dirty_io, cl); + + kfree(io); +} + +static void write_dirty_finish(struct closure *cl) +{ + struct dirty_io *io = container_of(cl, struct dirty_io, cl); + struct keybuf_key *w = io->bio.bi_private; + struct cached_dev *dc = io->dc; + + bio_free_pages(&io->bio); + + /* This is kind of a dumb way of signalling errors. */ + if (KEY_DIRTY(&w->key)) { + int ret; + unsigned int i; + struct keylist keys; + + bch_keylist_init(&keys); + + bkey_copy(keys.top, &w->key); + SET_KEY_DIRTY(keys.top, false); + bch_keylist_push(&keys); + + for (i = 0; i < KEY_PTRS(&w->key); i++) + atomic_inc(&PTR_BUCKET(dc->disk.c, &w->key, i)->pin); + + ret = bch_btree_insert(dc->disk.c, &keys, NULL, &w->key); + + if (ret) + trace_bcache_writeback_collision(&w->key); + + atomic_long_inc(ret + ? &dc->disk.c->writeback_keys_failed + : &dc->disk.c->writeback_keys_done); + } + + bch_keybuf_del(&dc->writeback_keys, w); + up(&dc->in_flight); + + closure_return_with_destructor(cl, dirty_io_destructor); +} + +static void dirty_endio(struct bio *bio) +{ + struct keybuf_key *w = bio->bi_private; + struct dirty_io *io = w->private; + + if (bio->bi_status) { + SET_KEY_DIRTY(&w->key, false); + bch_count_backing_io_errors(io->dc, bio); + } + + closure_put(&io->cl); +} + +static void write_dirty(struct closure *cl) +{ + struct dirty_io *io = container_of(cl, struct dirty_io, cl); + struct keybuf_key *w = io->bio.bi_private; + struct cached_dev *dc = io->dc; + + uint16_t next_sequence; + + if (atomic_read(&dc->writeback_sequence_next) != io->sequence) { + /* Not our turn to write; wait for a write to complete */ + closure_wait(&dc->writeback_ordering_wait, cl); + + if (atomic_read(&dc->writeback_sequence_next) == io->sequence) { + /* + * Edge case-- it happened in indeterminate order + * relative to when we were added to wait list.. + */ + closure_wake_up(&dc->writeback_ordering_wait); + } + + continue_at(cl, write_dirty, io->dc->writeback_write_wq); + return; + } + + next_sequence = io->sequence + 1; + + /* + * IO errors are signalled using the dirty bit on the key. + * If we failed to read, we should not attempt to write to the + * backing device. Instead, immediately go to write_dirty_finish + * to clean up. + */ + if (KEY_DIRTY(&w->key)) { + dirty_init(w); + bio_set_op_attrs(&io->bio, REQ_OP_WRITE, 0); + io->bio.bi_iter.bi_sector = KEY_START(&w->key); + bio_set_dev(&io->bio, io->dc->bdev); + io->bio.bi_end_io = dirty_endio; + + /* I/O request sent to backing device */ + closure_bio_submit(io->dc->disk.c, &io->bio, cl); + } + + atomic_set(&dc->writeback_sequence_next, next_sequence); + closure_wake_up(&dc->writeback_ordering_wait); + + continue_at(cl, write_dirty_finish, io->dc->writeback_write_wq); +} + +static void read_dirty_endio(struct bio *bio) +{ + struct keybuf_key *w = bio->bi_private; + struct dirty_io *io = w->private; + + /* is_read = 1 */ + bch_count_io_errors(PTR_CACHE(io->dc->disk.c, &w->key, 0), + bio->bi_status, 1, + "reading dirty data from cache"); + + dirty_endio(bio); +} + +static void read_dirty_submit(struct closure *cl) +{ + struct dirty_io *io = container_of(cl, struct dirty_io, cl); + + closure_bio_submit(io->dc->disk.c, &io->bio, cl); + + continue_at(cl, write_dirty, io->dc->writeback_write_wq); +} + +static void read_dirty(struct cached_dev *dc) +{ + unsigned int delay = 0; + struct keybuf_key *next, *keys[MAX_WRITEBACKS_IN_PASS], *w; + size_t size; + int nk, i; + struct dirty_io *io; + struct closure cl; + uint16_t sequence = 0; + + BUG_ON(!llist_empty(&dc->writeback_ordering_wait.list)); + atomic_set(&dc->writeback_sequence_next, sequence); + closure_init_stack(&cl); + + /* + * XXX: if we error, background writeback just spins. Should use some + * mempools. + */ + + next = bch_keybuf_next(&dc->writeback_keys); + + while (!kthread_should_stop() && + !test_bit(CACHE_SET_IO_DISABLE, &dc->disk.c->flags) && + next) { + size = 0; + nk = 0; + + do { + BUG_ON(ptr_stale(dc->disk.c, &next->key, 0)); + + /* + * Don't combine too many operations, even if they + * are all small. + */ + if (nk >= MAX_WRITEBACKS_IN_PASS) + break; + + /* + * If the current operation is very large, don't + * further combine operations. + */ + if (size >= MAX_WRITESIZE_IN_PASS) + break; + + /* + * Operations are only eligible to be combined + * if they are contiguous. + * + * TODO: add a heuristic willing to fire a + * certain amount of non-contiguous IO per pass, + * so that we can benefit from backing device + * command queueing. + */ + if ((nk != 0) && bkey_cmp(&keys[nk-1]->key, + &START_KEY(&next->key))) + break; + + size += KEY_SIZE(&next->key); + keys[nk++] = next; + } while ((next = bch_keybuf_next(&dc->writeback_keys))); + + /* Now we have gathered a set of 1..5 keys to write back. */ + for (i = 0; i < nk; i++) { + w = keys[i]; + + io = kzalloc(struct_size(io, bio.bi_inline_vecs, + DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS)), + GFP_KERNEL); + if (!io) + goto err; + + w->private = io; + io->dc = dc; + io->sequence = sequence++; + + dirty_init(w); + bio_set_op_attrs(&io->bio, REQ_OP_READ, 0); + io->bio.bi_iter.bi_sector = PTR_OFFSET(&w->key, 0); + bio_set_dev(&io->bio, + PTR_CACHE(dc->disk.c, &w->key, 0)->bdev); + io->bio.bi_end_io = read_dirty_endio; + + if (bch_bio_alloc_pages(&io->bio, GFP_KERNEL)) + goto err_free; + + trace_bcache_writeback(&w->key); + + down(&dc->in_flight); + + /* + * We've acquired a semaphore for the maximum + * simultaneous number of writebacks; from here + * everything happens asynchronously. + */ + closure_call(&io->cl, read_dirty_submit, NULL, &cl); + } + + delay = writeback_delay(dc, size); + + while (!kthread_should_stop() && + !test_bit(CACHE_SET_IO_DISABLE, &dc->disk.c->flags) && + delay) { + schedule_timeout_interruptible(delay); + delay = writeback_delay(dc, 0); + } + } + + if (0) { +err_free: + kfree(w->private); +err: + bch_keybuf_del(&dc->writeback_keys, w); + } + + /* + * Wait for outstanding writeback IOs to finish (and keybuf slots to be + * freed) before refilling again + */ + closure_sync(&cl); +} + +/* Scan for dirty data */ + +void bcache_dev_sectors_dirty_add(struct cache_set *c, unsigned int inode, + uint64_t offset, int nr_sectors) +{ + struct bcache_device *d = c->devices[inode]; + unsigned int stripe_offset, sectors_dirty; + int stripe; + + if (!d) + return; + + stripe = offset_to_stripe(d, offset); + if (stripe < 0) + return; + + if (UUID_FLASH_ONLY(&c->uuids[inode])) + atomic_long_add(nr_sectors, &c->flash_dev_dirty_sectors); + + stripe_offset = offset & (d->stripe_size - 1); + + while (nr_sectors) { + int s = min_t(unsigned int, abs(nr_sectors), + d->stripe_size - stripe_offset); + + if (nr_sectors < 0) + s = -s; + + if (stripe >= d->nr_stripes) + return; + + sectors_dirty = atomic_add_return(s, + d->stripe_sectors_dirty + stripe); + if (sectors_dirty == d->stripe_size) + set_bit(stripe, d->full_dirty_stripes); + else + clear_bit(stripe, d->full_dirty_stripes); + + nr_sectors -= s; + stripe_offset = 0; + stripe++; + } +} + +static bool dirty_pred(struct keybuf *buf, struct bkey *k) +{ + struct cached_dev *dc = container_of(buf, + struct cached_dev, + writeback_keys); + + BUG_ON(KEY_INODE(k) != dc->disk.id); + + return KEY_DIRTY(k); +} + +static void refill_full_stripes(struct cached_dev *dc) +{ + struct keybuf *buf = &dc->writeback_keys; + unsigned int start_stripe, next_stripe; + int stripe; + bool wrapped = false; + + stripe = offset_to_stripe(&dc->disk, KEY_OFFSET(&buf->last_scanned)); + if (stripe < 0) + stripe = 0; + + start_stripe = stripe; + + while (1) { + stripe = find_next_bit(dc->disk.full_dirty_stripes, + dc->disk.nr_stripes, stripe); + + if (stripe == dc->disk.nr_stripes) + goto next; + + next_stripe = find_next_zero_bit(dc->disk.full_dirty_stripes, + dc->disk.nr_stripes, stripe); + + buf->last_scanned = KEY(dc->disk.id, + stripe * dc->disk.stripe_size, 0); + + bch_refill_keybuf(dc->disk.c, buf, + &KEY(dc->disk.id, + next_stripe * dc->disk.stripe_size, 0), + dirty_pred); + + if (array_freelist_empty(&buf->freelist)) + return; + + stripe = next_stripe; +next: + if (wrapped && stripe > start_stripe) + return; + + if (stripe == dc->disk.nr_stripes) { + stripe = 0; + wrapped = true; + } + } +} + +/* + * Returns true if we scanned the entire disk + */ +static bool refill_dirty(struct cached_dev *dc) +{ + struct keybuf *buf = &dc->writeback_keys; + struct bkey start = KEY(dc->disk.id, 0, 0); + struct bkey end = KEY(dc->disk.id, MAX_KEY_OFFSET, 0); + struct bkey start_pos; + + /* + * make sure keybuf pos is inside the range for this disk - at bringup + * we might not be attached yet so this disk's inode nr isn't + * initialized then + */ + if (bkey_cmp(&buf->last_scanned, &start) < 0 || + bkey_cmp(&buf->last_scanned, &end) > 0) + buf->last_scanned = start; + + if (dc->partial_stripes_expensive) { + refill_full_stripes(dc); + if (array_freelist_empty(&buf->freelist)) + return false; + } + + start_pos = buf->last_scanned; + bch_refill_keybuf(dc->disk.c, buf, &end, dirty_pred); + + if (bkey_cmp(&buf->last_scanned, &end) < 0) + return false; + + /* + * If we get to the end start scanning again from the beginning, and + * only scan up to where we initially started scanning from: + */ + buf->last_scanned = start; + bch_refill_keybuf(dc->disk.c, buf, &start_pos, dirty_pred); + + return bkey_cmp(&buf->last_scanned, &start_pos) >= 0; +} + +static int bch_writeback_thread(void *arg) +{ + struct cached_dev *dc = arg; + struct cache_set *c = dc->disk.c; + bool searched_full_index; + + bch_ratelimit_reset(&dc->writeback_rate); + + while (!kthread_should_stop() && + !test_bit(CACHE_SET_IO_DISABLE, &c->flags)) { + down_write(&dc->writeback_lock); + set_current_state(TASK_INTERRUPTIBLE); + /* + * If the bache device is detaching, skip here and continue + * to perform writeback. Otherwise, if no dirty data on cache, + * or there is dirty data on cache but writeback is disabled, + * the writeback thread should sleep here and wait for others + * to wake up it. + */ + if (!test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags) && + (!atomic_read(&dc->has_dirty) || !dc->writeback_running)) { + up_write(&dc->writeback_lock); + + if (kthread_should_stop() || + test_bit(CACHE_SET_IO_DISABLE, &c->flags)) { + set_current_state(TASK_RUNNING); + break; + } + + schedule(); + continue; + } + set_current_state(TASK_RUNNING); + + searched_full_index = refill_dirty(dc); + + if (searched_full_index && + RB_EMPTY_ROOT(&dc->writeback_keys.keys)) { + atomic_set(&dc->has_dirty, 0); + SET_BDEV_STATE(&dc->sb, BDEV_STATE_CLEAN); + bch_write_bdev_super(dc, NULL); + /* + * If bcache device is detaching via sysfs interface, + * writeback thread should stop after there is no dirty + * data on cache. BCACHE_DEV_DETACHING flag is set in + * bch_cached_dev_detach(). + */ + if (test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags)) { + up_write(&dc->writeback_lock); + break; + } + + /* + * When dirty data rate is high (e.g. 50%+), there might + * be heavy buckets fragmentation after writeback + * finished, which hurts following write performance. + * If users really care about write performance they + * may set BCH_ENABLE_AUTO_GC via sysfs, then when + * BCH_DO_AUTO_GC is set, garbage collection thread + * will be wake up here. After moving gc, the shrunk + * btree and discarded free buckets SSD space may be + * helpful for following write requests. + */ + if (c->gc_after_writeback == + (BCH_ENABLE_AUTO_GC|BCH_DO_AUTO_GC)) { + c->gc_after_writeback &= ~BCH_DO_AUTO_GC; + force_wake_up_gc(c); + } + } + + up_write(&dc->writeback_lock); + + read_dirty(dc); + + if (searched_full_index) { + unsigned int delay = dc->writeback_delay * HZ; + + while (delay && + !kthread_should_stop() && + !test_bit(CACHE_SET_IO_DISABLE, &c->flags) && + !test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags)) + delay = schedule_timeout_interruptible(delay); + + bch_ratelimit_reset(&dc->writeback_rate); + } + } + + if (dc->writeback_write_wq) { + flush_workqueue(dc->writeback_write_wq); + destroy_workqueue(dc->writeback_write_wq); + } + cached_dev_put(dc); + wait_for_kthread_stop(); + + return 0; +} + +/* Init */ +#define INIT_KEYS_EACH_TIME 500000 + +struct sectors_dirty_init { + struct btree_op op; + unsigned int inode; + size_t count; +}; + +static int sectors_dirty_init_fn(struct btree_op *_op, struct btree *b, + struct bkey *k) +{ + struct sectors_dirty_init *op = container_of(_op, + struct sectors_dirty_init, op); + if (KEY_INODE(k) > op->inode) + return MAP_DONE; + + if (KEY_DIRTY(k)) + bcache_dev_sectors_dirty_add(b->c, KEY_INODE(k), + KEY_START(k), KEY_SIZE(k)); + + op->count++; + if (!(op->count % INIT_KEYS_EACH_TIME)) + cond_resched(); + + return MAP_CONTINUE; +} + +static int bch_root_node_dirty_init(struct cache_set *c, + struct bcache_device *d, + struct bkey *k) +{ + struct sectors_dirty_init op; + int ret; + + bch_btree_op_init(&op.op, -1); + op.inode = d->id; + op.count = 0; + + ret = bcache_btree(map_keys_recurse, + k, + c->root, + &op.op, + &KEY(op.inode, 0, 0), + sectors_dirty_init_fn, + 0); + if (ret < 0) + pr_warn("sectors dirty init failed, ret=%d!\n", ret); + + /* + * The op may be added to cache_set's btree_cache_wait + * in mca_cannibalize(), must ensure it is removed from + * the list and release btree_cache_alloc_lock before + * free op memory. + * Otherwise, the btree_cache_wait will be damaged. + */ + bch_cannibalize_unlock(c); + finish_wait(&c->btree_cache_wait, &(&op.op)->wait); + + return ret; +} + +static int bch_dirty_init_thread(void *arg) +{ + struct dirty_init_thrd_info *info = arg; + struct bch_dirty_init_state *state = info->state; + struct cache_set *c = state->c; + struct btree_iter iter; + struct bkey *k, *p; + int cur_idx, prev_idx, skip_nr; + + k = p = NULL; + prev_idx = 0; + + bch_btree_iter_init(&c->root->keys, &iter, NULL); + k = bch_btree_iter_next_filter(&iter, &c->root->keys, bch_ptr_bad); + BUG_ON(!k); + + p = k; + + while (k) { + spin_lock(&state->idx_lock); + cur_idx = state->key_idx; + state->key_idx++; + spin_unlock(&state->idx_lock); + + skip_nr = cur_idx - prev_idx; + + while (skip_nr) { + k = bch_btree_iter_next_filter(&iter, + &c->root->keys, + bch_ptr_bad); + if (k) + p = k; + else { + atomic_set(&state->enough, 1); + /* Update state->enough earlier */ + smp_mb__after_atomic(); + goto out; + } + skip_nr--; + } + + if (p) { + if (bch_root_node_dirty_init(c, state->d, p) < 0) + goto out; + } + + p = NULL; + prev_idx = cur_idx; + } + +out: + /* In order to wake up state->wait in time */ + smp_mb__before_atomic(); + if (atomic_dec_and_test(&state->started)) + wake_up(&state->wait); + + return 0; +} + +static int bch_btre_dirty_init_thread_nr(void) +{ + int n = num_online_cpus()/2; + + if (n == 0) + n = 1; + else if (n > BCH_DIRTY_INIT_THRD_MAX) + n = BCH_DIRTY_INIT_THRD_MAX; + + return n; +} + +void bch_sectors_dirty_init(struct bcache_device *d) +{ + int i; + struct btree *b = NULL; + struct bkey *k = NULL; + struct btree_iter iter; + struct sectors_dirty_init op; + struct cache_set *c = d->c; + struct bch_dirty_init_state state; + +retry_lock: + b = c->root; + rw_lock(0, b, b->level); + if (b != c->root) { + rw_unlock(0, b); + goto retry_lock; + } + + /* Just count root keys if no leaf node */ + if (c->root->level == 0) { + bch_btree_op_init(&op.op, -1); + op.inode = d->id; + op.count = 0; + + for_each_key_filter(&c->root->keys, + k, &iter, bch_ptr_invalid) { + if (KEY_INODE(k) != op.inode) + continue; + sectors_dirty_init_fn(&op.op, c->root, k); + } + + rw_unlock(0, b); + return; + } + + memset(&state, 0, sizeof(struct bch_dirty_init_state)); + state.c = c; + state.d = d; + state.total_threads = bch_btre_dirty_init_thread_nr(); + state.key_idx = 0; + spin_lock_init(&state.idx_lock); + atomic_set(&state.started, 0); + atomic_set(&state.enough, 0); + init_waitqueue_head(&state.wait); + + for (i = 0; i < state.total_threads; i++) { + /* Fetch latest state.enough earlier */ + smp_mb__before_atomic(); + if (atomic_read(&state.enough)) + break; + + atomic_inc(&state.started); + state.infos[i].state = &state; + state.infos[i].thread = + kthread_run(bch_dirty_init_thread, &state.infos[i], + "bch_dirtcnt[%d]", i); + if (IS_ERR(state.infos[i].thread)) { + pr_err("fails to run thread bch_dirty_init[%d]\n", i); + atomic_dec(&state.started); + for (--i; i >= 0; i--) + kthread_stop(state.infos[i].thread); + goto out; + } + } + +out: + /* Must wait for all threads to stop. */ + wait_event(state.wait, atomic_read(&state.started) == 0); + rw_unlock(0, b); +} + +void bch_cached_dev_writeback_init(struct cached_dev *dc) +{ + sema_init(&dc->in_flight, 64); + init_rwsem(&dc->writeback_lock); + bch_keybuf_init(&dc->writeback_keys); + + dc->writeback_metadata = true; + dc->writeback_running = false; + dc->writeback_percent = 10; + dc->writeback_delay = 30; + atomic_long_set(&dc->writeback_rate.rate, 1024); + dc->writeback_rate_minimum = 8; + + dc->writeback_rate_update_seconds = WRITEBACK_RATE_UPDATE_SECS_DEFAULT; + dc->writeback_rate_p_term_inverse = 40; + dc->writeback_rate_i_term_inverse = 10000; + + WARN_ON(test_and_clear_bit(BCACHE_DEV_WB_RUNNING, &dc->disk.flags)); + INIT_DELAYED_WORK(&dc->writeback_rate_update, update_writeback_rate); +} + +int bch_cached_dev_writeback_start(struct cached_dev *dc) +{ + dc->writeback_write_wq = alloc_workqueue("bcache_writeback_wq", + WQ_MEM_RECLAIM, 0); + if (!dc->writeback_write_wq) + return -ENOMEM; + + cached_dev_get(dc); + dc->writeback_thread = kthread_create(bch_writeback_thread, dc, + "bcache_writeback"); + if (IS_ERR(dc->writeback_thread)) { + cached_dev_put(dc); + destroy_workqueue(dc->writeback_write_wq); + return PTR_ERR(dc->writeback_thread); + } + dc->writeback_running = true; + + WARN_ON(test_and_set_bit(BCACHE_DEV_WB_RUNNING, &dc->disk.flags)); + schedule_delayed_work(&dc->writeback_rate_update, + dc->writeback_rate_update_seconds * HZ); + + bch_writeback_queue(dc); + + return 0; +} diff --git a/drivers/md/bcache/writeback.h b/drivers/md/bcache/writeback.h new file mode 100644 index 000000000..0f1d96920 --- /dev/null +++ b/drivers/md/bcache/writeback.h @@ -0,0 +1,151 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _BCACHE_WRITEBACK_H +#define _BCACHE_WRITEBACK_H + +#define CUTOFF_WRITEBACK 40 +#define CUTOFF_WRITEBACK_SYNC 70 + +#define CUTOFF_WRITEBACK_MAX 70 +#define CUTOFF_WRITEBACK_SYNC_MAX 90 + +#define MAX_WRITEBACKS_IN_PASS 5 +#define MAX_WRITESIZE_IN_PASS 5000 /* *512b */ + +#define WRITEBACK_RATE_UPDATE_SECS_MAX 60 +#define WRITEBACK_RATE_UPDATE_SECS_DEFAULT 5 + +#define BCH_AUTO_GC_DIRTY_THRESHOLD 50 + +#define BCH_DIRTY_INIT_THRD_MAX 12 +/* + * 14 (16384ths) is chosen here as something that each backing device + * should be a reasonable fraction of the share, and not to blow up + * until individual backing devices are a petabyte. + */ +#define WRITEBACK_SHARE_SHIFT 14 + +struct bch_dirty_init_state; +struct dirty_init_thrd_info { + struct bch_dirty_init_state *state; + struct task_struct *thread; +}; + +struct bch_dirty_init_state { + struct cache_set *c; + struct bcache_device *d; + int total_threads; + int key_idx; + spinlock_t idx_lock; + atomic_t started; + atomic_t enough; + wait_queue_head_t wait; + struct dirty_init_thrd_info infos[BCH_DIRTY_INIT_THRD_MAX]; +}; + +static inline uint64_t bcache_dev_sectors_dirty(struct bcache_device *d) +{ + uint64_t i, ret = 0; + + for (i = 0; i < d->nr_stripes; i++) + ret += atomic_read(d->stripe_sectors_dirty + i); + + return ret; +} + +static inline int offset_to_stripe(struct bcache_device *d, + uint64_t offset) +{ + do_div(offset, d->stripe_size); + + /* d->nr_stripes is in range [1, INT_MAX] */ + if (unlikely(offset >= d->nr_stripes)) { + pr_err("Invalid stripe %llu (>= nr_stripes %d).\n", + offset, d->nr_stripes); + return -EINVAL; + } + + /* + * Here offset is definitly smaller than INT_MAX, + * return it as int will never overflow. + */ + return offset; +} + +static inline bool bcache_dev_stripe_dirty(struct cached_dev *dc, + uint64_t offset, + unsigned int nr_sectors) +{ + int stripe = offset_to_stripe(&dc->disk, offset); + + if (stripe < 0) + return false; + + while (1) { + if (atomic_read(dc->disk.stripe_sectors_dirty + stripe)) + return true; + + if (nr_sectors <= dc->disk.stripe_size) + return false; + + nr_sectors -= dc->disk.stripe_size; + stripe++; + } +} + +extern unsigned int bch_cutoff_writeback; +extern unsigned int bch_cutoff_writeback_sync; + +static inline bool should_writeback(struct cached_dev *dc, struct bio *bio, + unsigned int cache_mode, bool would_skip) +{ + unsigned int in_use = dc->disk.c->gc_stats.in_use; + + if (cache_mode != CACHE_MODE_WRITEBACK || + test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags) || + in_use > bch_cutoff_writeback_sync) + return false; + + if (bio_op(bio) == REQ_OP_DISCARD) + return false; + + if (dc->partial_stripes_expensive && + bcache_dev_stripe_dirty(dc, bio->bi_iter.bi_sector, + bio_sectors(bio))) + return true; + + if (would_skip) + return false; + + return (op_is_sync(bio->bi_opf) || + bio->bi_opf & (REQ_META|REQ_PRIO) || + in_use <= bch_cutoff_writeback); +} + +static inline void bch_writeback_queue(struct cached_dev *dc) +{ + if (!IS_ERR_OR_NULL(dc->writeback_thread)) + wake_up_process(dc->writeback_thread); +} + +static inline void bch_writeback_add(struct cached_dev *dc) +{ + if (!atomic_read(&dc->has_dirty) && + !atomic_xchg(&dc->has_dirty, 1)) { + if (BDEV_STATE(&dc->sb) != BDEV_STATE_DIRTY) { + SET_BDEV_STATE(&dc->sb, BDEV_STATE_DIRTY); + /* XXX: should do this synchronously */ + bch_write_bdev_super(dc, NULL); + } + + bch_writeback_queue(dc); + } +} + +void bcache_dev_sectors_dirty_add(struct cache_set *c, unsigned int inode, + uint64_t offset, int nr_sectors); + +void bch_sectors_dirty_init(struct bcache_device *d); +void bch_cached_dev_writeback_init(struct cached_dev *dc); +int bch_cached_dev_writeback_start(struct cached_dev *dc); + +#endif |