diff options
Diffstat (limited to 'drivers/md')
102 files changed, 15097 insertions, 8579 deletions
diff --git a/drivers/md/Kconfig b/drivers/md/Kconfig index 3bfc8f1da9f..5bdedf6df15 100644 --- a/drivers/md/Kconfig +++ b/drivers/md/Kconfig @@ -176,8 +176,12 @@ config MD_FAULTY source "drivers/md/bcache/Kconfig" +config BLK_DEV_DM_BUILTIN + boolean + config BLK_DEV_DM tristate "Device mapper support" + select BLK_DEV_DM_BUILTIN ---help--- Device-mapper is a low level volume manager. It works by allowing people to specify mappings for ranges of logical sectors. Various @@ -238,6 +242,7 @@ config DM_CRYPT config DM_SNAPSHOT tristate "Snapshot target" depends on BLK_DEV_DM + select DM_BUFIO ---help--- Allow volume managers to take writable snapshots of a device. @@ -249,16 +254,6 @@ config DM_THIN_PROVISIONING ---help--- Provides thin provisioning and snapshots that share a data store. -config DM_DEBUG_BLOCK_STACK_TRACING - boolean "Keep stack trace of thin provisioning block lock holders" - depends on STACKTRACE_SUPPORT && DM_THIN_PROVISIONING - select STACKTRACE - ---help--- - Enable this for messages that may help debug problems with the - block manager locking used by thin provisioning. - - If unsure, say N. - config DM_CACHE tristate "Cache target (EXPERIMENTAL)" depends on BLK_DEV_DM @@ -290,6 +285,17 @@ config DM_CACHE_CLEANER A simple cache policy that writes back all data to the origin. Used when decommissioning a dm-cache. +config DM_ERA + tristate "Era target (EXPERIMENTAL)" + depends on BLK_DEV_DM + default n + select DM_PERSISTENT_DATA + select DM_BIO_PRISON + ---help--- + dm-era tracks which parts of a block device are written to + over time. Useful for maintaining cache coherency when using + vendor snapshots. + config DM_MIRROR tristate "Mirror target" depends on BLK_DEV_DM @@ -297,6 +303,17 @@ config DM_MIRROR Allow volume managers to mirror logical volumes, also needed for live data migration tools such as 'pvmove'. +config DM_LOG_USERSPACE + tristate "Mirror userspace logging" + depends on DM_MIRROR && NET + select CONNECTOR + ---help--- + The userspace logging module provides a mechanism for + relaying the dm-dirty-log API to userspace. Log designs + which are more suited to userspace implementation (e.g. + shared storage logs) or experimental logs can be implemented + by leveraging this framework. + config DM_RAID tristate "RAID 1/4/5/6/10 target" depends on BLK_DEV_DM @@ -323,17 +340,6 @@ config DM_RAID RAID-5, RAID-6 distributes the syndromes across the drives in one of the available parity distribution methods. -config DM_LOG_USERSPACE - tristate "Mirror userspace logging" - depends on DM_MIRROR && NET - select CONNECTOR - ---help--- - The userspace logging module provides a mechanism for - relaying the dm-dirty-log API to userspace. Log designs - which are more suited to userspace implementation (e.g. - shared storage logs) or experimental logs can be implemented - by leveraging this framework. - config DM_ZERO tristate "Zero target" depends on BLK_DEV_DM @@ -412,4 +418,18 @@ config DM_VERITY If unsure, say N. +config DM_SWITCH + tristate "Switch target support (EXPERIMENTAL)" + depends on BLK_DEV_DM + ---help--- + This device-mapper target creates a device that supports an arbitrary + mapping of fixed-size regions of I/O across a fixed set of paths. + The path used for any specific region can be switched dynamically + by sending the target a message. + + To compile this code as a module, choose M here: the module will + be called dm-switch. + + If unsure, say N. + endif # MD diff --git a/drivers/md/Makefile b/drivers/md/Makefile index 1439fd4ad9b..a2da532b1c2 100644 --- a/drivers/md/Makefile +++ b/drivers/md/Makefile @@ -3,7 +3,7 @@ # dm-mod-y += dm.o dm-table.o dm-target.o dm-linear.o dm-stripe.o \ - dm-ioctl.o dm-io.o dm-kcopyd.o dm-sysfs.o + dm-ioctl.o dm-io.o dm-kcopyd.o dm-sysfs.o dm-stats.o dm-multipath-y += dm-path-selector.o dm-mpath.o dm-snapshot-y += dm-snap.o dm-exception-store.o dm-snap-transient.o \ dm-snap-persistent.o @@ -14,6 +14,7 @@ dm-thin-pool-y += dm-thin.o dm-thin-metadata.o dm-cache-y += dm-cache-target.o dm-cache-metadata.o dm-cache-policy.o dm-cache-mq-y += dm-cache-policy-mq.o dm-cache-cleaner-y += dm-cache-policy-cleaner.o +dm-era-y += dm-era-target.o md-mod-y += md.o bitmap.o raid456-y += raid5.o @@ -32,6 +33,7 @@ obj-$(CONFIG_MD_FAULTY) += faulty.o obj-$(CONFIG_BCACHE) += bcache/ obj-$(CONFIG_BLK_DEV_MD) += md-mod.o obj-$(CONFIG_BLK_DEV_DM) += dm-mod.o +obj-$(CONFIG_BLK_DEV_DM_BUILTIN) += dm-builtin.o obj-$(CONFIG_DM_BUFIO) += dm-bufio.o obj-$(CONFIG_DM_BIO_PRISON) += dm-bio-prison.o obj-$(CONFIG_DM_CRYPT) += dm-crypt.o @@ -40,6 +42,7 @@ obj-$(CONFIG_DM_FLAKEY) += dm-flakey.o obj-$(CONFIG_DM_MULTIPATH) += dm-multipath.o dm-round-robin.o obj-$(CONFIG_DM_MULTIPATH_QL) += dm-queue-length.o obj-$(CONFIG_DM_MULTIPATH_ST) += dm-service-time.o +obj-$(CONFIG_DM_SWITCH) += dm-switch.o obj-$(CONFIG_DM_SNAPSHOT) += dm-snapshot.o obj-$(CONFIG_DM_PERSISTENT_DATA) += persistent-data/ obj-$(CONFIG_DM_MIRROR) += dm-mirror.o dm-log.o dm-region-hash.o @@ -51,6 +54,7 @@ obj-$(CONFIG_DM_VERITY) += dm-verity.o obj-$(CONFIG_DM_CACHE) += dm-cache.o obj-$(CONFIG_DM_CACHE_MQ) += dm-cache-mq.o obj-$(CONFIG_DM_CACHE_CLEANER) += dm-cache-cleaner.o +obj-$(CONFIG_DM_ERA) += dm-era.o ifeq ($(CONFIG_DM_UEVENT),y) dm-mod-objs += dm-uevent.o diff --git a/drivers/md/bcache/Kconfig b/drivers/md/bcache/Kconfig index f950c9d29f3..4d200883c50 100644 --- a/drivers/md/bcache/Kconfig +++ b/drivers/md/bcache/Kconfig @@ -13,15 +13,8 @@ config BCACHE_DEBUG ---help--- Don't select this option unless you're a developer - Enables extra debugging tools (primarily a fuzz tester) - -config BCACHE_EDEBUG - bool "Extended runtime checks" - depends on BCACHE - ---help--- - Don't select this option unless you're a developer - - Enables extra runtime checks which significantly affect performance + Enables extra debugging tools, allows expensive runtime checks to be + turned on. config BCACHE_CLOSURES_DEBUG bool "Debug closures" @@ -31,11 +24,3 @@ config BCACHE_CLOSURES_DEBUG 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. - -# cgroup code needs to be updated: -# -#config CGROUP_BCACHE -# bool "Cgroup controls for bcache" -# depends on BCACHE && BLK_CGROUP -# ---help--- -# TODO diff --git a/drivers/md/bcache/Makefile b/drivers/md/bcache/Makefile index 0e9c82523be..c488b846f83 100644 --- a/drivers/md/bcache/Makefile +++ b/drivers/md/bcache/Makefile @@ -1,7 +1,8 @@ obj-$(CONFIG_BCACHE) += bcache.o -bcache-y := alloc.o btree.o bset.o io.o journal.o writeback.o\ - movinggc.o request.o super.o sysfs.o debug.o util.o trace.o stats.o closure.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 CFLAGS_request.o += -Iblock diff --git a/drivers/md/bcache/alloc.c b/drivers/md/bcache/alloc.c index 048f2947e08..443d03fbac4 100644 --- a/drivers/md/bcache/alloc.c +++ b/drivers/md/bcache/alloc.c @@ -63,9 +63,11 @@ #include "bcache.h" #include "btree.h" +#include <linux/blkdev.h> +#include <linux/freezer.h> +#include <linux/kthread.h> #include <linux/random.h> - -#define MAX_IN_FLIGHT_DISCARDS 8U +#include <trace/events/bcache.h> /* Bucket heap / gen */ @@ -76,12 +78,6 @@ uint8_t bch_inc_gen(struct cache *ca, struct bucket *b) ca->set->need_gc = max(ca->set->need_gc, bucket_gc_gen(b)); WARN_ON_ONCE(ca->set->need_gc > BUCKET_GC_GEN_MAX); - if (CACHE_SYNC(&ca->set->sb)) { - ca->need_save_prio = max(ca->need_save_prio, - bucket_disk_gen(b)); - WARN_ON_ONCE(ca->need_save_prio > BUCKET_DISK_GEN_MAX); - } - return ret; } @@ -118,119 +114,63 @@ void bch_rescale_priorities(struct cache_set *c, int sectors) mutex_unlock(&c->bucket_lock); } -/* Discard/TRIM */ - -struct discard { - struct list_head list; - struct work_struct work; - struct cache *ca; - long bucket; - - struct bio bio; - struct bio_vec bv; -}; - -static void discard_finish(struct work_struct *w) -{ - struct discard *d = container_of(w, struct discard, work); - struct cache *ca = d->ca; - char buf[BDEVNAME_SIZE]; - - if (!test_bit(BIO_UPTODATE, &d->bio.bi_flags)) { - pr_notice("discard error on %s, disabling", - bdevname(ca->bdev, buf)); - d->ca->discard = 0; - } - - mutex_lock(&ca->set->bucket_lock); - - fifo_push(&ca->free, d->bucket); - list_add(&d->list, &ca->discards); - atomic_dec(&ca->discards_in_flight); - - mutex_unlock(&ca->set->bucket_lock); - - closure_wake_up(&ca->set->bucket_wait); - wake_up(&ca->set->alloc_wait); - - closure_put(&ca->set->cl); -} - -static void discard_endio(struct bio *bio, int error) -{ - struct discard *d = container_of(bio, struct discard, bio); - schedule_work(&d->work); -} - -static void do_discard(struct cache *ca, long bucket) -{ - struct discard *d = list_first_entry(&ca->discards, - struct discard, list); - - list_del(&d->list); - d->bucket = bucket; - - atomic_inc(&ca->discards_in_flight); - closure_get(&ca->set->cl); - - bio_init(&d->bio); - - d->bio.bi_sector = bucket_to_sector(ca->set, d->bucket); - d->bio.bi_bdev = ca->bdev; - d->bio.bi_rw = REQ_WRITE|REQ_DISCARD; - d->bio.bi_max_vecs = 1; - d->bio.bi_io_vec = d->bio.bi_inline_vecs; - d->bio.bi_size = bucket_bytes(ca); - d->bio.bi_end_io = discard_endio; - bio_set_prio(&d->bio, IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0)); - - submit_bio(0, &d->bio); -} - -/* Allocation */ +/* + * 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 && - bucket_disk_gen(b) < BUCKET_DISK_GEN_MAX; + return bucket_gc_gen(b) < BUCKET_GC_GEN_MAX; } -bool bch_bucket_add_unused(struct cache *ca, struct bucket *b) +bool bch_can_invalidate_bucket(struct cache *ca, struct bucket *b) { - BUG_ON(GC_MARK(b) || GC_SECTORS_USED(b)); - - if (fifo_used(&ca->free) > ca->watermark[WATERMARK_MOVINGGC] && - CACHE_REPLACEMENT(&ca->sb) == CACHE_REPLACEMENT_FIFO) - return false; - - b->prio = 0; - - if (can_inc_bucket_gen(b) && - fifo_push(&ca->unused, b - ca->buckets)) { - atomic_inc(&b->pin); - return true; - } - - return false; -} + BUG_ON(!ca->set->gc_mark_valid); -static bool can_invalidate_bucket(struct cache *ca, struct bucket *b) -{ - return GC_MARK(b) == GC_MARK_RECLAIMABLE && + return (!GC_MARK(b) || + GC_MARK(b) == GC_MARK_RECLAIMABLE) && !atomic_read(&b->pin) && can_inc_bucket_gen(b); } -static void invalidate_one_bucket(struct cache *ca, struct bucket *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); } -#define bucket_prio(b) \ - (((unsigned) (b->prio - ca->set->min_prio)) * GC_SECTORS_USED(b)) +/* + * 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 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)) @@ -243,20 +183,7 @@ static void invalidate_buckets_lru(struct cache *ca) ca->heap.used = 0; for_each_bucket(b, ca) { - /* - * If we fill up the unused list, if we then return before - * adding anything to the free_inc list we'll skip writing - * prios/gens and just go back to allocating from the unused - * list: - */ - if (fifo_full(&ca->unused)) - return; - - if (!can_invalidate_bucket(ca, b)) - continue; - - if (!GC_SECTORS_USED(b) && - bch_bucket_add_unused(ca, b)) + if (!bch_can_invalidate_bucket(ca, b)) continue; if (!heap_full(&ca->heap)) @@ -277,11 +204,11 @@ static void invalidate_buckets_lru(struct cache *ca) * multiple times when it can't do anything */ ca->invalidate_needs_gc = 1; - bch_queue_gc(ca->set); + wake_up_gc(ca->set); return; } - invalidate_one_bucket(ca, b); + bch_invalidate_one_bucket(ca, b); } } @@ -297,12 +224,12 @@ static void invalidate_buckets_fifo(struct cache *ca) b = ca->buckets + ca->fifo_last_bucket++; - if (can_invalidate_bucket(ca, b)) - invalidate_one_bucket(ca, b); + if (bch_can_invalidate_bucket(ca, b)) + bch_invalidate_one_bucket(ca, b); if (++checked >= ca->sb.nbuckets) { ca->invalidate_needs_gc = 1; - bch_queue_gc(ca->set); + wake_up_gc(ca->set); return; } } @@ -322,12 +249,12 @@ static void invalidate_buckets_random(struct cache *ca) b = ca->buckets + n; - if (can_invalidate_bucket(ca, b)) - invalidate_one_bucket(ca, b); + if (bch_can_invalidate_bucket(ca, b)) + bch_invalidate_one_bucket(ca, b); if (++checked >= ca->sb.nbuckets / 2) { ca->invalidate_needs_gc = 1; - bch_queue_gc(ca->set); + wake_up_gc(ca->set); return; } } @@ -335,8 +262,7 @@ static void invalidate_buckets_random(struct cache *ca) static void invalidate_buckets(struct cache *ca) { - if (ca->invalidate_needs_gc) - return; + BUG_ON(ca->invalidate_needs_gc); switch (CACHE_REPLACEMENT(&ca->sb)) { case CACHE_REPLACEMENT_LRU: @@ -349,39 +275,44 @@ static void invalidate_buckets(struct cache *ca) invalidate_buckets_random(ca); break; } - - pr_debug("free %zu/%zu free_inc %zu/%zu unused %zu/%zu", - fifo_used(&ca->free), ca->free.size, - fifo_used(&ca->free_inc), ca->free_inc.size, - fifo_used(&ca->unused), ca->unused.size); } #define allocator_wait(ca, cond) \ do { \ - DEFINE_WAIT(__wait); \ - \ while (1) { \ - prepare_to_wait(&ca->set->alloc_wait, \ - &__wait, TASK_INTERRUPTIBLE); \ + set_current_state(TASK_INTERRUPTIBLE); \ if (cond) \ break; \ \ mutex_unlock(&(ca)->set->bucket_lock); \ - if (test_bit(CACHE_SET_STOPPING_2, &ca->set->flags)) { \ - finish_wait(&ca->set->alloc_wait, &__wait); \ - closure_return(cl); \ - } \ + if (kthread_should_stop()) \ + return 0; \ \ + try_to_freeze(); \ schedule(); \ mutex_lock(&(ca)->set->bucket_lock); \ } \ - \ - finish_wait(&ca->set->alloc_wait, &__wait); \ + __set_current_state(TASK_RUNNING); \ } while (0) -void bch_allocator_thread(struct closure *cl) +static int bch_allocator_push(struct cache *ca, long bucket) +{ + unsigned 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 = container_of(cl, struct cache, alloc); + struct cache *ca = arg; mutex_lock(&ca->set->bucket_lock); @@ -391,28 +322,22 @@ void bch_allocator_thread(struct closure *cl) * possibly issue discards to them, then we add the bucket to * the free list: */ - while (1) { + while (!fifo_empty(&ca->free_inc)) { long bucket; - if ((!atomic_read(&ca->set->prio_blocked) || - !CACHE_SYNC(&ca->set->sb)) && - !fifo_empty(&ca->unused)) - fifo_pop(&ca->unused, bucket); - else if (!fifo_empty(&ca->free_inc)) - fifo_pop(&ca->free_inc, bucket); - else - break; - - allocator_wait(ca, (int) fifo_free(&ca->free) > - atomic_read(&ca->discards_in_flight)); + fifo_pop(&ca->free_inc, bucket); if (ca->discard) { - allocator_wait(ca, !list_empty(&ca->discards)); - do_discard(ca, bucket); - } else { - fifo_push(&ca->free, bucket); - closure_wake_up(&ca->set->bucket_wait); + mutex_unlock(&ca->set->bucket_lock); + blkdev_issue_discard(ca->bdev, + bucket_to_sector(ca->set, bucket), + ca->sb.block_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); } /* @@ -421,9 +346,9 @@ void bch_allocator_thread(struct closure *cl) * them to the free_inc list: */ +retry_invalidate: allocator_wait(ca, ca->set->gc_mark_valid && - (ca->need_save_prio > 64 || - !ca->invalidate_needs_gc)); + !ca->invalidate_needs_gc); invalidate_buckets(ca); /* @@ -431,84 +356,111 @@ void bch_allocator_thread(struct closure *cl) * new stuff to them: */ allocator_wait(ca, !atomic_read(&ca->set->prio_blocked)); - if (CACHE_SYNC(&ca->set->sb) && - (!fifo_empty(&ca->free_inc) || - ca->need_save_prio > 64)) + if (CACHE_SYNC(&ca->set->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; + bch_prio_write(ca); + } } } -long bch_bucket_alloc(struct cache *ca, unsigned watermark, struct closure *cl) +/* Allocation */ + +long bch_bucket_alloc(struct cache *ca, unsigned reserve, bool wait) { - long r = -1; -again: - wake_up(&ca->set->alloc_wait); - - if (fifo_used(&ca->free) > ca->watermark[watermark] && - fifo_pop(&ca->free, r)) { - struct bucket *b = ca->buckets + r; -#ifdef CONFIG_BCACHE_EDEBUG + DEFINE_WAIT(w); + struct bucket *b; + long r; + + /* 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: + wake_up_process(ca->alloc_thread); + + trace_bcache_alloc(ca, reserve); + + if (expensive_debug_checks(ca->set)) { size_t iter; long i; + unsigned j; for (iter = 0; iter < prio_buckets(ca) * 2; iter++) BUG_ON(ca->prio_buckets[iter] == (uint64_t) r); - fifo_for_each(i, &ca->free, iter) - BUG_ON(i == 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); - fifo_for_each(i, &ca->unused, iter) - BUG_ON(i == r); -#endif - BUG_ON(atomic_read(&b->pin) != 1); - - SET_GC_SECTORS_USED(b, ca->sb.bucket_size); - - if (watermark <= WATERMARK_METADATA) { - SET_GC_MARK(b, GC_MARK_METADATA); - b->prio = BTREE_PRIO; - } else { - SET_GC_MARK(b, GC_MARK_RECLAIMABLE); - b->prio = INITIAL_PRIO; - } - - return r; } - pr_debug("alloc failure: blocked %i free %zu free_inc %zu unused %zu", - atomic_read(&ca->set->prio_blocked), fifo_used(&ca->free), - fifo_used(&ca->free_inc), fifo_used(&ca->unused)); + b = ca->buckets + r; - if (cl) { - closure_wait(&ca->set->bucket_wait, cl); + BUG_ON(atomic_read(&b->pin) != 1); - if (closure_blocking(cl)) { - mutex_unlock(&ca->set->bucket_lock); - closure_sync(cl); - mutex_lock(&ca->set->bucket_lock); - goto again; - } + 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; } - return -1; + return r; +} + +void __bch_bucket_free(struct cache *ca, struct bucket *b) +{ + SET_GC_MARK(b, 0); + SET_GC_SECTORS_USED(b, 0); } void bch_bucket_free(struct cache_set *c, struct bkey *k) { unsigned i; - for (i = 0; i < KEY_PTRS(k); i++) { - struct bucket *b = PTR_BUCKET(c, k, i); - - SET_GC_MARK(b, GC_MARK_RECLAIMABLE); - SET_GC_SECTORS_USED(b, 0); - bch_bucket_add_unused(PTR_CACHE(c, k, i), b); - } + 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 watermark, - struct bkey *k, int n, struct closure *cl) +int __bch_bucket_alloc_set(struct cache_set *c, unsigned reserve, + struct bkey *k, int n, bool wait) { int i; @@ -521,7 +473,7 @@ int __bch_bucket_alloc_set(struct cache_set *c, unsigned watermark, for (i = 0; i < n; i++) { struct cache *ca = c->cache_by_alloc[i]; - long b = bch_bucket_alloc(ca, watermark, cl); + long b = bch_bucket_alloc(ca, reserve, wait); if (b == -1) goto err; @@ -536,64 +488,209 @@ int __bch_bucket_alloc_set(struct cache_set *c, unsigned watermark, return 0; err: bch_bucket_free(c, k); - __bkey_put(c, k); + bkey_put(c, k); return -1; } -int bch_bucket_alloc_set(struct cache_set *c, unsigned watermark, - struct bkey *k, int n, struct closure *cl) +int bch_bucket_alloc_set(struct cache_set *c, unsigned reserve, + struct bkey *k, int n, bool wait) { int ret; mutex_lock(&c->bucket_lock); - ret = __bch_bucket_alloc_set(c, watermark, k, n, cl); + ret = __bch_bucket_alloc_set(c, reserve, k, n, wait); mutex_unlock(&c->bucket_lock); return ret; } -/* Init */ +/* Sector allocator */ -void bch_cache_allocator_exit(struct cache *ca) -{ - struct discard *d; +struct open_bucket { + struct list_head list; + unsigned last_write_point; + unsigned sectors_free; + BKEY_PADDED(key); +}; - while (!list_empty(&ca->discards)) { - d = list_first_entry(&ca->discards, struct discard, list); - cancel_work_sync(&d->work); - list_del(&d->list); - kfree(d); +/* + * We keep multiple buckets open for writes, and try to segregate different + * write streams for better cache utilization: first 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, 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 write_point, + struct bkey *alloc) +{ + struct open_bucket *ret, *ret_task = NULL; + + list_for_each_entry_reverse(ret, &c->data_buckets, list) + 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->sb.bucket_size; + bkey_copy(&ret->key, alloc); + bkey_init(alloc); } + + if (!ret->sectors_free) + ret = NULL; + + return ret; } -int bch_cache_allocator_init(struct cache *ca) +/* + * 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 sectors, + unsigned write_point, unsigned write_prio, bool wait) { + struct open_bucket *b; + BKEY_PADDED(key) alloc; unsigned i; /* - * Reserve: - * Prio/gen writes first - * Then 8 for btree allocations - * Then half for the moving garbage collector + * 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 watermark = write_prio + ? RESERVE_MOVINGGC + : RESERVE_NONE; + + spin_unlock(&c->data_bucket_lock); + + if (bch_bucket_alloc_set(c, watermark, &alloc.key, 1, 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 find_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->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 */ - ca->watermark[WATERMARK_PRIO] = 0; +void bch_open_buckets_free(struct cache_set *c) +{ + struct open_bucket *b; - ca->watermark[WATERMARK_METADATA] = prio_buckets(ca); + while (!list_empty(&c->data_buckets)) { + b = list_first_entry(&c->data_buckets, + struct open_bucket, list); + list_del(&b->list); + kfree(b); + } +} - ca->watermark[WATERMARK_MOVINGGC] = 8 + - ca->watermark[WATERMARK_METADATA]; +int bch_open_buckets_alloc(struct cache_set *c) +{ + int i; - ca->watermark[WATERMARK_NONE] = ca->free.size / 2 + - ca->watermark[WATERMARK_MOVINGGC]; + spin_lock_init(&c->data_bucket_lock); - for (i = 0; i < MAX_IN_FLIGHT_DISCARDS; i++) { - struct discard *d = kzalloc(sizeof(*d), GFP_KERNEL); - if (!d) + for (i = 0; i < 6; i++) { + struct open_bucket *b = kzalloc(sizeof(*b), GFP_KERNEL); + if (!b) return -ENOMEM; - d->ca = ca; - INIT_WORK(&d->work, discard_finish); - list_add(&d->list, &ca->discards); + 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 index d3e15b42a4a..d2ebcf32309 100644 --- a/drivers/md/bcache/bcache.h +++ b/drivers/md/bcache/bcache.h @@ -177,8 +177,8 @@ #define pr_fmt(fmt) "bcache: %s() " fmt "\n", __func__ +#include <linux/bcache.h> #include <linux/bio.h> -#include <linux/blktrace_api.h> #include <linux/kobject.h> #include <linux/list.h> #include <linux/mutex.h> @@ -187,6 +187,7 @@ #include <linux/types.h> #include <linux/workqueue.h> +#include "bset.h" #include "util.h" #include "closure.h" @@ -194,10 +195,8 @@ struct bucket { atomic_t pin; uint16_t prio; uint8_t gen; - uint8_t disk_gen; uint8_t last_gc; /* Most out of date gen in the btree */ - uint8_t gc_gen; - uint16_t gc_mark; + uint16_t gc_mark; /* Bitfield used by GC. See below for field */ }; /* @@ -206,172 +205,13 @@ struct bucket { */ BITMASK(GC_MARK, struct bucket, gc_mark, 0, 2); -#define GC_MARK_RECLAIMABLE 0 -#define GC_MARK_DIRTY 1 -#define GC_MARK_METADATA 2 -BITMASK(GC_SECTORS_USED, struct bucket, gc_mark, 2, 14); - -struct bkey { - uint64_t high; - uint64_t low; - uint64_t ptr[]; -}; - -/* Enough for a key with 6 pointers */ -#define BKEY_PAD 8 - -#define BKEY_PADDED(key) \ - union { struct bkey key; uint64_t key ## _pad[BKEY_PAD]; } - -/* Version 0: Cache device - * Version 1: Backing device - * Version 2: Seed pointer into btree node checksum - * Version 3: Cache device with new UUID format - * Version 4: Backing device with data offset - */ -#define BCACHE_SB_VERSION_CDEV 0 -#define BCACHE_SB_VERSION_BDEV 1 -#define BCACHE_SB_VERSION_CDEV_WITH_UUID 3 -#define BCACHE_SB_VERSION_BDEV_WITH_OFFSET 4 -#define BCACHE_SB_MAX_VERSION 4 - -#define SB_SECTOR 8 -#define SB_SIZE 4096 -#define SB_LABEL_SIZE 32 -#define SB_JOURNAL_BUCKETS 256U -/* SB_JOURNAL_BUCKETS must be divisible by BITS_PER_LONG */ -#define MAX_CACHES_PER_SET 8 - -#define BDEV_DATA_START_DEFAULT 16 /* sectors */ - -struct cache_sb { - uint64_t csum; - uint64_t offset; /* sector where this sb was written */ - uint64_t version; - - uint8_t magic[16]; - - uint8_t uuid[16]; - union { - uint8_t set_uuid[16]; - uint64_t set_magic; - }; - uint8_t label[SB_LABEL_SIZE]; - - uint64_t flags; - uint64_t seq; - uint64_t pad[8]; - - union { - struct { - /* Cache devices */ - uint64_t nbuckets; /* device size */ - - uint16_t block_size; /* sectors */ - uint16_t bucket_size; /* sectors */ - - uint16_t nr_in_set; - uint16_t nr_this_dev; - }; - struct { - /* Backing devices */ - uint64_t data_offset; - - /* - * block_size from the cache device section is still used by - * backing devices, so don't add anything here until we fix - * things to not need it for backing devices anymore - */ - }; - }; - - uint32_t last_mount; /* time_t */ - - uint16_t first_bucket; - union { - uint16_t njournal_buckets; - uint16_t keys; - }; - uint64_t d[SB_JOURNAL_BUCKETS]; /* journal buckets */ -}; - -BITMASK(CACHE_SYNC, struct cache_sb, flags, 0, 1); -BITMASK(CACHE_DISCARD, struct cache_sb, flags, 1, 1); -BITMASK(CACHE_REPLACEMENT, struct cache_sb, flags, 2, 3); -#define CACHE_REPLACEMENT_LRU 0U -#define CACHE_REPLACEMENT_FIFO 1U -#define CACHE_REPLACEMENT_RANDOM 2U - -BITMASK(BDEV_CACHE_MODE, struct cache_sb, flags, 0, 4); -#define CACHE_MODE_WRITETHROUGH 0U -#define CACHE_MODE_WRITEBACK 1U -#define CACHE_MODE_WRITEAROUND 2U -#define CACHE_MODE_NONE 3U -BITMASK(BDEV_STATE, struct cache_sb, flags, 61, 2); -#define BDEV_STATE_NONE 0U -#define BDEV_STATE_CLEAN 1U -#define BDEV_STATE_DIRTY 2U -#define BDEV_STATE_STALE 3U - -/* Version 1: Seed pointer into btree node checksum - */ -#define BCACHE_BSET_VERSION 1 - -/* - * This is the on disk format for btree nodes - a btree node on disk is a list - * of these; within each set the keys are sorted - */ -struct bset { - uint64_t csum; - uint64_t magic; - uint64_t seq; - uint32_t version; - uint32_t keys; - - union { - struct bkey start[0]; - uint64_t d[0]; - }; -}; - -/* - * On disk format for priorities and gens - see super.c near prio_write() for - * more. - */ -struct prio_set { - uint64_t csum; - uint64_t magic; - uint64_t seq; - uint32_t version; - uint32_t pad; - - uint64_t next_bucket; - - struct bucket_disk { - uint16_t prio; - uint8_t gen; - } __attribute((packed)) data[]; -}; - -struct uuid_entry { - union { - struct { - uint8_t uuid[16]; - uint8_t label[32]; - uint32_t first_reg; - uint32_t last_reg; - uint32_t invalidated; - - uint32_t flags; - /* Size of flash only volumes */ - uint64_t sectors; - }; - - uint8_t pad[128]; - }; -}; - -BITMASK(UUID_FLASH_ONLY, struct uuid_entry, flags, 0, 1); +#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" @@ -385,11 +225,7 @@ struct keybuf_key { void *private; }; -typedef bool (keybuf_pred_fn)(struct keybuf *, struct bkey *); - struct keybuf { - keybuf_pred_fn *key_predicate; - struct bkey last_scanned; spinlock_t lock; @@ -403,7 +239,7 @@ struct keybuf { struct rb_root keys; -#define KEYBUF_NR 100 +#define KEYBUF_NR 500 DECLARE_ARRAY_ALLOCATOR(struct keybuf_key, freelist, KEYBUF_NR); }; @@ -432,18 +268,19 @@ struct bcache_device { struct gendisk *disk; - /* If nonzero, we're closing */ - atomic_t closing; + unsigned long flags; +#define BCACHE_DEV_CLOSING 0 +#define BCACHE_DEV_DETACHING 1 +#define BCACHE_DEV_UNLINK_DONE 2 - /* If nonzero, we're detaching/unregistering from cache set */ - atomic_t detaching; + unsigned nr_stripes; + unsigned stripe_size; + atomic_t *stripe_sectors_dirty; + unsigned long *full_dirty_stripes; - atomic_long_t sectors_dirty; - unsigned long sectors_dirty_gc; unsigned long sectors_dirty_last; long sectors_dirty_derivative; - mempool_t *unaligned_bvec; struct bio_set *bio_split; unsigned data_csum:1; @@ -473,7 +310,8 @@ struct cached_dev { struct cache_sb sb; struct bio sb_bio; struct bio_vec sb_bv[1]; - struct closure_with_waitlist sb_write; + struct closure sb_write; + struct semaphore sb_write_mutex; /* Refcount on the cache set. Always nonzero when we're caching. */ atomic_t count; @@ -498,7 +336,7 @@ struct cached_dev { */ atomic_t has_dirty; - struct ratelimit writeback_rate; + struct bch_ratelimit writeback_rate; struct delayed_work writeback_rate_update; /* @@ -507,10 +345,9 @@ struct cached_dev { */ sector_t last_read; - /* Number of writeback bios in flight */ - atomic_t in_flight; - struct closure_with_timer writeback; - struct closure_waitlist writeback_wait; + /* Limit number of writeback bios in flight */ + struct semaphore in_flight; + struct task_struct *writeback_thread; struct keybuf writeback_keys; @@ -528,30 +365,31 @@ struct cached_dev { unsigned sequential_cutoff; unsigned readahead; - unsigned sequential_merge:1; unsigned verify:1; + unsigned bypass_torture_test:1; + unsigned partial_stripes_expensive:1; unsigned writeback_metadata:1; unsigned writeback_running:1; unsigned char writeback_percent; unsigned writeback_delay; - int writeback_rate_change; - int64_t writeback_rate_derivative; uint64_t writeback_rate_target; + int64_t writeback_rate_proportional; + int64_t writeback_rate_derivative; + int64_t writeback_rate_change; unsigned writeback_rate_update_seconds; unsigned writeback_rate_d_term; unsigned writeback_rate_p_term_inverse; - unsigned writeback_rate_d_smooth; }; -enum alloc_watermarks { - WATERMARK_PRIO, - WATERMARK_METADATA, - WATERMARK_MOVINGGC, - WATERMARK_NONE, - WATERMARK_MAX +enum alloc_reserve { + RESERVE_BTREE, + RESERVE_PRIO, + RESERVE_MOVINGGC, + RESERVE_NONE, + RESERVE_NR, }; struct cache { @@ -563,10 +401,7 @@ struct cache { struct kobject kobj; struct block_device *bdev; - unsigned watermark[WATERMARK_MAX]; - - struct closure alloc; - struct workqueue_struct *alloc_workqueue; + struct task_struct *alloc_thread; struct closure prio; struct prio_set *disk_buckets; @@ -589,14 +424,9 @@ struct cache { * 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) - * - * unused: GC found nothing pointing into these buckets (possibly - * because all the data they contained was overwritten), so we only - * need to discard them before they can be moved to the free list. */ - DECLARE_FIFO(long, free); + DECLARE_FIFO(long, free)[RESERVE_NR]; DECLARE_FIFO(long, free_inc); - DECLARE_FIFO(long, unused); size_t fifo_last_bucket; @@ -606,13 +436,6 @@ struct cache { DECLARE_HEAP(struct bucket *, heap); /* - * max(gen - disk_gen) for all buckets. When it gets too big we have to - * call prio_write() to keep gens from wrapping. - */ - uint8_t need_save_prio; - unsigned gc_move_threshold; - - /* * 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 @@ -621,15 +444,6 @@ struct cache { bool discard; /* Get rid of? */ - /* - * We preallocate structs for issuing discards to buckets, and keep them - * on this list when they're not in use; do_discard() issues discards - * whenever there's work to do and is called by free_some_buckets() and - * when a discard finishes. - */ - atomic_t discards_in_flight; - struct list_head discards; - struct journal_device journal; /* The rest of this all shows up in sysfs */ @@ -650,7 +464,6 @@ struct gc_stat { size_t nkeys; uint64_t data; /* sectors */ - uint64_t dirty; /* sectors */ unsigned in_use; /* percent */ }; @@ -664,13 +477,9 @@ struct gc_stat { * 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_STOPPING_2 gets set at the last phase, when it's time to shut down - * the allocation thread. */ #define CACHE_SET_UNREGISTERING 0 #define CACHE_SET_STOPPING 1 -#define CACHE_SET_STOPPING_2 2 struct cache_set { struct closure cl; @@ -694,7 +503,8 @@ struct cache_set { uint64_t cached_dev_sectors; struct closure caching; - struct closure_with_waitlist sb_write; + struct closure sb_write; + struct semaphore sb_write_mutex; mempool_t *search; mempool_t *bio_meta; @@ -703,9 +513,6 @@ struct cache_set { /* For the btree cache */ struct shrinker shrink; - /* For the allocator itself */ - wait_queue_head_t alloc_wait; - /* For the btree cache and anything allocation related */ struct mutex bucket_lock; @@ -742,19 +549,16 @@ struct cache_set { struct list_head btree_cache_freed; /* Number of elements in btree_cache + btree_cache_freeable lists */ - unsigned bucket_cache_used; + unsigned 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. However, only one thread can be doing this - * at a time, for obvious reasons - try_harder and try_wait are - * basically a lock for this that we can wait on asynchronously. The - * btree_root() macro releases the lock when it returns. + * to satisfy the allocation - lock to guarantee only one thread does + * this at a time: */ - struct closure *try_harder; - struct closure_waitlist try_wait; - uint64_t try_harder_start; + wait_queue_head_t btree_cache_wait; + struct task_struct *btree_cache_alloc_lock; /* * When we free a btree node, we increment the gen of the bucket the @@ -767,7 +571,7 @@ struct cache_set { * written. */ atomic_t prio_blocked; - struct closure_waitlist bucket_wait; + wait_queue_head_t bucket_wait; /* * For any bio we don't skip we subtract the number of sectors from @@ -783,14 +587,14 @@ struct cache_set { uint16_t min_prio; /* - * max(gen - gc_gen) for all buckets. When it gets too big we have to gc + * 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; - struct closure_with_waitlist gc; + struct task_struct *gc_thread; /* Where in the btree gc currently is */ struct bkey gc_done; @@ -803,37 +607,34 @@ struct cache_set { /* Counts how many sectors bio_insert has added to the cache */ atomic_t sectors_to_gc; - struct closure moving_gc; - struct closure_waitlist moving_gc_wait; + wait_queue_head_t moving_gc_wait; struct keybuf moving_gc_keys; /* Number of moving GC bios in flight */ - atomic_t 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 unsigned nr_uuids; struct uuid_entry *uuids; BKEY_PADDED(uuid_bucket); - struct closure_with_waitlist uuid_write; + 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 - this is a single element mempool for btree_read_work() + * on the stack - have to dynamically allocate them */ - struct mutex fill_lock; - struct btree_iter *fill_iter; + mempool_t *fill_iter; - /* - * btree_sort() is a merge sort and requires temporary space - single - * element mempool - */ - struct mutex sort_lock; - struct bset *sort; + struct bset_sort_state sort; /* List of buckets we're currently writing data to */ struct list_head data_buckets; @@ -849,20 +650,23 @@ struct cache_set { unsigned congested_read_threshold_us; unsigned congested_write_threshold_us; - spinlock_t sort_time_lock; - struct time_stats sort_time; struct time_stats btree_gc_time; struct time_stats btree_split_time; - spinlock_t btree_read_time_lock; struct time_stats btree_read_time; - struct time_stats try_harder_time; atomic_long_t cache_read_races; atomic_long_t writeback_keys_done; atomic_long_t writeback_keys_failed; + + enum { + ON_ERROR_UNREGISTER, + ON_ERROR_PANIC, + } on_error; unsigned error_limit; unsigned error_decay; + unsigned short journal_delay_ms; + bool expensive_debug_checks; unsigned verify:1; unsigned key_merging_disabled:1; unsigned gc_always_rewrite:1; @@ -873,21 +677,6 @@ struct cache_set { struct hlist_head bucket_hash[1 << BUCKET_HASH_BITS]; }; -static inline bool key_merging_disabled(struct cache_set *c) -{ -#ifdef CONFIG_BCACHE_DEBUG - return c->key_merging_disabled; -#else - return 0; -#endif -} - -static inline bool SB_IS_BDEV(const struct cache_sb *sb) -{ - return sb->version == BCACHE_SB_VERSION_BDEV - || sb->version == BCACHE_SB_VERSION_BDEV_WITH_OFFSET; -} - struct bbio { unsigned submit_time_us; union { @@ -901,15 +690,8 @@ struct bbio { struct bio bio; }; -static inline unsigned local_clock_us(void) -{ - return local_clock() >> 10; -} - -#define MAX_BSETS 4U - #define BTREE_PRIO USHRT_MAX -#define INITIAL_PRIO 32768 +#define INITIAL_PRIO 32768U #define btree_bytes(c) ((c)->btree_pages * PAGE_SIZE) #define btree_blocks(b) \ @@ -922,80 +704,12 @@ static inline unsigned local_clock_us(void) #define bucket_bytes(c) ((c)->sb.bucket_size << 9) #define block_bytes(c) ((c)->sb.block_size << 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, c) DIV_ROUND_UP(__set_bytes(i, k), block_bytes(c)) -#define set_blocks(i, c) __set_blocks(i, (i)->keys, c) - -#define node(i, j) ((struct bkey *) ((i)->d + (j))) -#define end(i) node(i, (i)->keys) - -#define index(i, b) \ - ((size_t) (((void *) i - (void *) (b)->sets[0].data) / \ - block_bytes(b->c))) - -#define btree_data_space(b) (PAGE_SIZE << (b)->page_order) - #define prios_per_bucket(c) \ ((bucket_bytes(c) - sizeof(struct prio_set)) / \ sizeof(struct bucket_disk)) #define prio_buckets(c) \ DIV_ROUND_UP((size_t) (c)->sb.nbuckets, prios_per_bucket(c)) -#define JSET_MAGIC 0x245235c1a3625032ULL -#define PSET_MAGIC 0x6750e15f87337f91ULL -#define BSET_MAGIC 0x90135c78b99e07f5ULL - -#define jset_magic(c) ((c)->sb.set_magic ^ JSET_MAGIC) -#define pset_magic(c) ((c)->sb.set_magic ^ PSET_MAGIC) -#define bset_magic(c) ((c)->sb.set_magic ^ BSET_MAGIC) - -/* Bkey fields: all units are in sectors */ - -#define KEY_FIELD(name, field, offset, size) \ - BITMASK(name, struct bkey, field, offset, size) - -#define PTR_FIELD(name, offset, size) \ - static inline uint64_t name(const struct bkey *k, unsigned i) \ - { return (k->ptr[i] >> offset) & ~(((uint64_t) ~0) << size); } \ - \ - static inline void SET_##name(struct bkey *k, unsigned i, uint64_t v)\ - { \ - k->ptr[i] &= ~(~((uint64_t) ~0 << size) << offset); \ - k->ptr[i] |= v << offset; \ - } - -KEY_FIELD(KEY_PTRS, high, 60, 3) -KEY_FIELD(HEADER_SIZE, high, 58, 2) -KEY_FIELD(KEY_CSUM, high, 56, 2) -KEY_FIELD(KEY_PINNED, high, 55, 1) -KEY_FIELD(KEY_DIRTY, high, 36, 1) - -KEY_FIELD(KEY_SIZE, high, 20, 16) -KEY_FIELD(KEY_INODE, high, 0, 20) - -/* Next time I change the on disk format, KEY_OFFSET() won't be 64 bits */ - -static inline uint64_t KEY_OFFSET(const struct bkey *k) -{ - return k->low; -} - -static inline void SET_KEY_OFFSET(struct bkey *k, uint64_t v) -{ - k->low = v; -} - -PTR_FIELD(PTR_DEV, 51, 12) -PTR_FIELD(PTR_OFFSET, 8, 43) -PTR_FIELD(PTR_GEN, 0, 8) - -#define PTR_CHECK_DEV ((1 << 12) - 1) - -#define PTR(gen, offset, dev) \ - ((((uint64_t) dev) << 51) | ((uint64_t) offset) << 8 | gen) - static inline size_t sector_to_bucket(struct cache_set *c, sector_t s) { return s >> c->bucket_bits; @@ -1032,28 +746,25 @@ static inline struct bucket *PTR_BUCKET(struct cache_set *c, return PTR_CACHE(c, k, ptr)->buckets + PTR_BUCKET_NR(c, k, ptr); } -/* Btree key macros */ +static inline uint8_t gen_after(uint8_t a, uint8_t b) +{ + uint8_t r = a - b; + return r > 128U ? 0 : r; +} -/* - * The high bit being set is a relic from when we used it to do binary - * searches - it told you where a key started. It's not used anymore, - * and can probably be safely dropped. - */ -#define KEY(dev, sector, len) \ -((struct bkey) { \ - .high = (1ULL << 63) | ((uint64_t) (len) << 20) | (dev), \ - .low = (sector) \ -}) +static inline uint8_t ptr_stale(struct cache_set *c, const struct bkey *k, + unsigned i) +{ + return gen_after(PTR_BUCKET(c, k, i)->gen, PTR_GEN(k, i)); +} -static inline void bkey_init(struct bkey *k) +static inline bool ptr_available(struct cache_set *c, const struct bkey *k, + unsigned i) { - *k = KEY(0, 0, 0); + return (PTR_DEV(k, i) < MAX_CACHES_PER_SET) && PTR_CACHE(c, k, i); } -#define KEY_START(k) (KEY_OFFSET(k) - KEY_SIZE(k)) -#define START_KEY(k) KEY(KEY_INODE(k), KEY_START(k), 0) -#define MAX_KEY KEY(~(~0 << 20), ((uint64_t) ~0) >> 1, 0) -#define ZERO_KEY KEY(0, 0, 0) +/* Btree key macros */ /* * This is used for various on disk data structures - cache_sb, prio_set, bset, @@ -1061,7 +772,8 @@ static inline void bkey_init(struct bkey *k) */ #define csum_set(i) \ bch_crc64(((void *) (i)) + sizeof(uint64_t), \ - ((void *) end(i)) - (((void *) (i)) + sizeof(uint64_t))) + ((void *) bset_bkey_last(i)) - \ + (((void *) (i)) + sizeof(uint64_t))) /* Error handling macros */ @@ -1104,31 +816,6 @@ do { \ for (b = (ca)->buckets + (ca)->sb.first_bucket; \ b < (ca)->buckets + (ca)->sb.nbuckets; b++) -static inline void __bkey_put(struct cache_set *c, struct bkey *k) -{ - unsigned i; - - for (i = 0; i < KEY_PTRS(k); i++) - atomic_dec_bug(&PTR_BUCKET(c, k, i)->pin); -} - -/* Blktrace macros */ - -#define blktrace_msg(c, fmt, ...) \ -do { \ - struct request_queue *q = bdev_get_queue(c->bdev); \ - if (q) \ - blk_add_trace_msg(q, fmt, ##__VA_ARGS__); \ -} while (0) - -#define blktrace_msg_all(s, fmt, ...) \ -do { \ - struct cache *_c; \ - unsigned i; \ - for_each_cache(_c, (s), i) \ - blktrace_msg(_c, fmt, ##__VA_ARGS__); \ -} while (0) - static inline void cached_dev_put(struct cached_dev *dc) { if (atomic_dec_and_test(&dc->count)) @@ -1141,16 +828,13 @@ static inline bool cached_dev_get(struct cached_dev *dc) return false; /* Paired with the mb in cached_dev_attach */ - smp_mb__after_atomic_inc(); + 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). - * - * bucket_disk_gen() returns the difference between the current gen and the gen - * on disk; they're both used to make sure gens don't wrap around. */ static inline uint8_t bucket_gc_gen(struct bucket *b) @@ -1158,13 +842,7 @@ static inline uint8_t bucket_gc_gen(struct bucket *b) return b->gen - b->last_gc; } -static inline uint8_t bucket_disk_gen(struct bucket *b) -{ - return b->gen - b->disk_gen; -} - #define BUCKET_GC_GEN_MAX 96U -#define BUCKET_DISK_GEN_MAX 64U #define kobj_attribute_write(n, fn) \ static struct kobj_attribute ksysfs_##n = __ATTR(n, S_IWUSR, NULL, fn) @@ -1173,10 +851,16 @@ static inline uint8_t bucket_disk_gen(struct bucket *b) static struct kobj_attribute ksysfs_##n = \ __ATTR(n, S_IWUSR|S_IRUSR, show, store) -/* Forward declarations */ +static inline void wake_up_allocators(struct cache_set *c) +{ + struct cache *ca; + unsigned i; -void bch_writeback_queue(struct cached_dev *); -void bch_writeback_add(struct cached_dev *, unsigned); + for_each_cache(ca, c, i) + wake_up_process(ca->alloc_thread); +} + +/* Forward declarations */ void bch_count_io_errors(struct cache *, int, const char *); void bch_bbio_count_io_errors(struct cache_set *, struct bio *, @@ -1185,23 +869,26 @@ void bch_bbio_endio(struct cache_set *, struct bio *, int, const char *); void bch_bbio_free(struct bio *, struct cache_set *); struct bio *bch_bbio_alloc(struct cache_set *); -struct bio *bch_bio_split(struct bio *, int, gfp_t, struct bio_set *); void bch_generic_make_request(struct bio *, struct bio_split_pool *); void __bch_submit_bbio(struct bio *, struct cache_set *); void bch_submit_bbio(struct bio *, struct cache_set *, struct bkey *, unsigned); uint8_t bch_inc_gen(struct cache *, struct bucket *); void bch_rescale_priorities(struct cache_set *, int); -bool bch_bucket_add_unused(struct cache *, struct bucket *); -void bch_allocator_thread(struct closure *); -long bch_bucket_alloc(struct cache *, unsigned, struct closure *); +bool bch_can_invalidate_bucket(struct cache *, struct bucket *); +void __bch_invalidate_one_bucket(struct cache *, struct bucket *); + +void __bch_bucket_free(struct cache *, struct bucket *); void bch_bucket_free(struct cache_set *, struct bkey *); +long bch_bucket_alloc(struct cache *, unsigned, bool); int __bch_bucket_alloc_set(struct cache_set *, unsigned, - struct bkey *, int, struct closure *); + struct bkey *, int, bool); int bch_bucket_alloc_set(struct cache_set *, unsigned, - struct bkey *, int, struct closure *); + struct bkey *, int, bool); +bool bch_alloc_sectors(struct cache_set *, struct bkey *, unsigned, + unsigned, unsigned, bool); __printf(2, 3) bool bch_cache_set_error(struct cache_set *, const char *, ...); @@ -1209,7 +896,7 @@ bool bch_cache_set_error(struct cache_set *, const char *, ...); void bch_prio_write(struct cache *); void bch_write_bdev_super(struct cached_dev *, struct closure *); -extern struct workqueue_struct *bcache_wq, *bch_gc_wq; +extern struct workqueue_struct *bcache_wq; extern const char * const bch_cache_modes[]; extern struct mutex bch_register_lock; extern struct list_head bch_cache_sets; @@ -1241,19 +928,15 @@ void bch_cache_set_stop(struct cache_set *); struct cache_set *bch_cache_set_alloc(struct cache_sb *); void bch_btree_cache_free(struct cache_set *); int bch_btree_cache_alloc(struct cache_set *); -void bch_cached_dev_writeback_init(struct cached_dev *); void bch_moving_init_cache_set(struct cache_set *); +int bch_open_buckets_alloc(struct cache_set *); +void bch_open_buckets_free(struct cache_set *); -void bch_cache_allocator_exit(struct cache *ca); -int bch_cache_allocator_init(struct cache *ca); +int bch_cache_allocator_start(struct cache *ca); void bch_debug_exit(void); int bch_debug_init(struct kobject *); -void bch_writeback_exit(void); -int bch_writeback_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 index cb4578a327b..54541641530 100644 --- a/drivers/md/bcache/bset.c +++ b/drivers/md/bcache/bset.c @@ -5,169 +5,174 @@ * Copyright 2012 Google, Inc. */ -#include "bcache.h" -#include "btree.h" -#include "debug.h" +#define pr_fmt(fmt) "bcache: %s() " fmt "\n", __func__ +#include "util.h" +#include "bset.h" + +#include <linux/console.h> #include <linux/random.h> #include <linux/prefetch.h> -/* Keylists */ +#ifdef CONFIG_BCACHE_DEBUG -void bch_keylist_copy(struct keylist *dest, struct keylist *src) +void bch_dump_bset(struct btree_keys *b, struct bset *i, unsigned set) { - *dest = *src; + struct bkey *k, *next; + + for (k = i->start; k < bset_bkey_last(i); k = next) { + next = bkey_next(k); + + printk(KERN_ERR "block %u key %u/%u: ", set, + (unsigned) ((u64 *) k - i->d), i->keys); - if (src->list == src->d) { - size_t n = (uint64_t *) src->top - src->d; - dest->top = (struct bkey *) &dest->d[n]; - dest->list = dest->d; + if (b->ops->key_dump) + b->ops->key_dump(b, k); + else + printk("%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) + printk(KERN_ERR "Key skipped backwards\n"); } } -int bch_keylist_realloc(struct keylist *l, int nptrs, struct cache_set *c) +void bch_dump_bucket(struct btree_keys *b) { - unsigned oldsize = (uint64_t *) l->top - l->list; - unsigned newsize = oldsize + 2 + nptrs; - uint64_t *new; - - /* 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, - * bio_insert() and bio_invalidate() will insert the keys created so far - * and finish the rest when the keylist is empty. - */ - if (newsize * sizeof(uint64_t) > block_bytes(c) - sizeof(struct jset)) - return -ENOMEM; - - newsize = roundup_pow_of_two(newsize); + unsigned i; - if (newsize <= KEYLIST_INLINE || - roundup_pow_of_two(oldsize) == newsize) - return 0; + 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(); +} - new = krealloc(l->list == l->d ? NULL : l->list, - sizeof(uint64_t) * newsize, GFP_NOIO); +int __bch_count_data(struct btree_keys *b) +{ + unsigned ret = 0; + struct btree_iter iter; + struct bkey *k; - if (!new) - return -ENOMEM; + if (b->ops->is_extents) + for_each_key(b, k, &iter) + ret += KEY_SIZE(k); + return ret; +} - if (l->list == l->d) - memcpy(new, l->list, sizeof(uint64_t) * KEYLIST_INLINE); +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; - l->list = new; - l->top = (struct bkey *) (&l->list[oldsize]); + 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; - return 0; -} + if (bch_ptr_invalid(b, k)) + continue; -struct bkey *bch_keylist_pop(struct keylist *l) -{ - struct bkey *k = l->bottom; + err = "Overlapping keys"; + if (p && bkey_cmp(p, &START_KEY(k)) > 0) + goto bug; + } else { + if (bch_ptr_bad(b, k)) + continue; - if (k == l->top) - return NULL; + 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); - while (bkey_next(k) != l->top) - k = bkey_next(k); + va_start(args, fmt); + vprintk(fmt, args); + va_end(args); - return l->top = k; + panic("bch_check_keys error: %s:\n", err); } -/* Pointer validation */ - -bool __bch_ptr_invalid(struct cache_set *c, int level, const struct bkey *k) +static void bch_btree_iter_next_check(struct btree_iter *iter) { - unsigned i; + struct bkey *k = iter->data->k, *next = bkey_next(k); - if (level && (!KEY_PTRS(k) || !KEY_SIZE(k) || KEY_DIRTY(k))) - goto bad; + 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"); + } +} - if (!level && KEY_SIZE(k) > KEY_OFFSET(k)) - goto bad; +#else - if (!KEY_SIZE(k)) - return true; +static inline void bch_btree_iter_next_check(struct btree_iter *iter) {} - 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->sb.bucket_size || - bucket < ca->sb.first_bucket || - bucket >= ca->sb.nbuckets) - goto bad; - } +#endif - return false; -bad: - cache_bug(c, "spotted bad key %s: %s", pkey(k), bch_ptr_status(c, k)); - return true; -} +/* Keylists */ -bool bch_ptr_bad(struct btree *b, const struct bkey *k) +int __bch_keylist_realloc(struct keylist *l, unsigned u64s) { - struct bucket *g; - unsigned i, stale; + 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; - if (!bkey_cmp(k, &ZERO_KEY) || - !KEY_PTRS(k) || - bch_ptr_invalid(b, k)) - return true; + newsize = roundup_pow_of_two(newsize); - if (KEY_PTRS(k) && PTR_DEV(k, 0) == PTR_CHECK_DEV) - return true; + if (newsize <= KEYLIST_INLINE || + roundup_pow_of_two(oldsize) == newsize) + return 0; - for (i = 0; i < KEY_PTRS(k); i++) - if (ptr_available(b->c, k, i)) { - g = PTR_BUCKET(b->c, k, i); - stale = ptr_stale(b->c, k, i); + new_keys = krealloc(old_keys, sizeof(uint64_t) * newsize, GFP_NOIO); - btree_bug_on(stale > 96, b, - "key too stale: %i, need_gc %u", - stale, b->c->need_gc); + if (!new_keys) + return -ENOMEM; - btree_bug_on(stale && KEY_DIRTY(k) && KEY_SIZE(k), - b, "stale dirty pointer"); + if (!old_keys) + memcpy(new_keys, l->inline_keys, sizeof(uint64_t) * oldsize); - if (stale) - return true; + l->keys_p = new_keys; + l->top_p = new_keys + oldsize; -#ifdef CONFIG_BCACHE_EDEBUG - if (!mutex_trylock(&b->c->bucket_lock)) - continue; + return 0; +} - if (b->level) { - if (KEY_DIRTY(k) || - g->prio != BTREE_PRIO || - (b->c->gc_mark_valid && - GC_MARK(g) != GC_MARK_METADATA)) - goto bug; - - } else { - if (g->prio == BTREE_PRIO) - goto bug; - - if (KEY_DIRTY(k) && - b->c->gc_mark_valid && - GC_MARK(g) != GC_MARK_DIRTY) - goto bug; - } - mutex_unlock(&b->c->bucket_lock); -#endif - } +struct bkey *bch_keylist_pop(struct keylist *l) +{ + struct bkey *k = l->keys; - return false; -#ifdef CONFIG_BCACHE_EDEBUG -bug: - mutex_unlock(&b->c->bucket_lock); - btree_bug(b, -"inconsistent pointer %s: bucket %zu pin %i prio %i gen %i last_gc %i mark %llu gc_gen %i", - pkey(k), PTR_BUCKET_NR(b->c, k, i), atomic_read(&g->pin), - g->prio, g->gen, g->last_gc, GC_MARK(g), g->gc_gen); - return true; -#endif + if (k == l->top) + return NULL; + + while (bkey_next(k) != l->top) + k = bkey_next(k); + + return l->top = k; +} + +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 */ @@ -224,56 +229,138 @@ bool __bch_cut_back(const struct bkey *where, struct bkey *k) return true; } -static uint64_t merge_chksums(struct bkey *l, struct bkey *r) +/* 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 exponent:BKEY_EXPONENT_BITS; + unsigned m:BKEY_MID_BITS; + unsigned 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 (l->ptr[KEY_PTRS(l)] + r->ptr[KEY_PTRS(r)]) & - ~((uint64_t)1 << 63); + return btree_keys_bytes(b) / BSET_CACHELINE; } -/* 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 *b, struct bkey *l, struct bkey *r) +/* Space required for the auxiliary search trees */ +static inline size_t bset_tree_bytes(struct btree_keys *b) { - unsigned i; + return btree_keys_cachelines(b) * sizeof(struct bkey_float); +} - if (key_merging_disabled(b->c)) - return false; +/* 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); +} - if (KEY_PTRS(l) != KEY_PTRS(r) || - KEY_DIRTY(l) != KEY_DIRTY(r) || - bkey_cmp(l, &START_KEY(r))) - return false; +/* Memory allocation */ - for (i = 0; i < KEY_PTRS(l); i++) - if (l->ptr[i] + 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; +void bch_btree_keys_free(struct btree_keys *b) +{ + struct bset_tree *t = b->set; - /* 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); + if (bset_prev_bytes(b) < PAGE_SIZE) + kfree(t->prev); + else + free_pages((unsigned long) t->prev, + get_order(bset_prev_bytes(b))); - bch_cut_front(l, r); - return false; - } + if (bset_tree_bytes(b) < PAGE_SIZE) + kfree(t->tree); + else + free_pages((unsigned long) t->tree, + get_order(bset_tree_bytes(b))); - if (KEY_CSUM(l)) { - if (KEY_CSUM(r)) - l->ptr[KEY_PTRS(l)] = merge_chksums(l, r); - else - SET_KEY_CSUM(l, 0); - } + free_pages((unsigned long) t->data, b->page_order); - SET_KEY_OFFSET(l, KEY_OFFSET(l) + KEY_SIZE(r)); - SET_KEY_SIZE(l, KEY_SIZE(l) + KEY_SIZE(r)); + t->prev = NULL; + t->tree = NULL; + t->data = NULL; +} +EXPORT_SYMBOL(bch_btree_keys_free); - return true; +int bch_btree_keys_alloc(struct btree_keys *b, unsigned 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, 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; } +EXPORT_SYMBOL(bch_btree_keys_alloc); + +void bch_btree_keys_init(struct btree_keys *b, const struct btree_keys_ops *ops, + bool *expensive_debug_checks) +{ + unsigned i; + + b->ops = ops; + b->expensive_debug_checks = expensive_debug_checks; + b->nsets = 0; + b->last_set_unwritten = 0; + + /* XXX: shouldn't be needed */ + for (i = 0; i < MAX_BSETS; i++) + b->set[i].size = 0; + /* + * Second loop starts at 1 because b->keys[0]->data is the memory we + * allocated + */ + for (i = 1; i < MAX_BSETS; i++) + b->set[i].data = NULL; +} +EXPORT_SYMBOL(bch_btree_keys_init); /* Binary tree stuff for auxiliary search trees */ @@ -395,7 +482,7 @@ void inorder_test(void) #endif /* - * Cacheline/offset <-> bkey pointer arithmatic: + * 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). @@ -404,7 +491,7 @@ void inorder_test(void) * 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 arithmatic to + * 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 @@ -424,9 +511,11 @@ static unsigned bkey_to_cacheline(struct bset_tree *t, struct bkey *k) return ((void *) k - (void *) t->data) / BSET_CACHELINE; } -static unsigned bkey_to_cacheline_offset(struct bkey *k) +static unsigned bkey_to_cacheline_offset(struct bset_tree *t, + unsigned cacheline, + struct bkey *k) { - return ((size_t) k & (BSET_CACHELINE - 1)) / sizeof(uint64_t); + return (u64 *) k - (u64 *) cacheline_to_bkey(t, cacheline, 0); } static struct bkey *tree_to_bkey(struct bset_tree *t, unsigned j) @@ -450,16 +539,8 @@ static struct bkey *table_to_bkey(struct bset_tree *t, unsigned cacheline) static inline uint64_t shrd128(uint64_t high, uint64_t low, uint8_t shift) { -#ifdef CONFIG_X86_64 - asm("shrd %[shift],%[high],%[low]" - : [low] "+Rm" (low) - : [high] "R" (high), - [shift] "ci" (shift) - : "cc"); -#else low >>= shift; low |= (high << 1) << (63U - shift); -#endif return low; } @@ -481,7 +562,7 @@ static void make_bfloat(struct bset_tree *t, unsigned j) : tree_to_prev_bkey(t, j >> ffs(j)); struct bkey *r = is_power_of_2(j + 1) - ? node(t->data, t->data->keys - bkey_u64s(&t->end)) + ? 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); @@ -505,9 +586,9 @@ static void make_bfloat(struct bset_tree *t, unsigned j) f->exponent = 127; } -static void bset_alloc_tree(struct btree *b, struct bset_tree *t) +static void bset_alloc_tree(struct btree_keys *b, struct bset_tree *t) { - if (t != b->sets) { + if (t != b->set) { unsigned j = roundup(t[-1].size, 64 / sizeof(struct bkey_float)); @@ -515,33 +596,54 @@ static void bset_alloc_tree(struct btree *b, struct bset_tree *t) t->prev = t[-1].prev + j; } - while (t < b->sets + MAX_BSETS) + while (t < b->set + MAX_BSETS) t++->size = 0; } -static void bset_build_unwritten_tree(struct btree *b) +static void bch_bset_build_unwritten_tree(struct btree_keys *b) { - struct bset_tree *t = b->sets + b->nsets; + 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->sets->tree + bset_tree_space(b)) { - t->prev[0] = bkey_to_cacheline_offset(t->data->start); + 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; } } -static void bset_build_written_tree(struct btree *b) +void bch_bset_init_next(struct btree_keys *b, struct bset *i, uint64_t magic) { - struct bset_tree *t = b->sets + b->nsets; - struct bkey *k = t->data->start; + 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); +} +EXPORT_SYMBOL(bch_bset_init_next); + +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 j, cacheline = 1; + b->last_set_unwritten = 0; + bset_alloc_tree(b, t); t->size = min_t(unsigned, - bkey_to_cacheline(t, end(t->data)), - b->sets->tree + bset_tree_space(b) - t->tree); + 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; @@ -554,16 +656,14 @@ static void bset_build_written_tree(struct btree *b) for (j = inorder_next(0, t->size); j; j = inorder_next(j, t->size)) { - while (bkey_to_cacheline(t, k) != cacheline) - k = bkey_next(k); + while (bkey_to_cacheline(t, k) < cacheline) + prev = k, k = bkey_next(k); - t->prev[j] = bkey_u64s(k); - k = bkey_next(k); - cacheline++; - t->tree[j].m = bkey_to_cacheline_offset(k); + t->prev[j] = bkey_u64s(prev); + t->tree[j].m = bkey_to_cacheline_offset(t, cacheline++, k); } - while (bkey_next(k) != end(t->data)) + while (bkey_next(k) != bset_bkey_last(t->data)) k = bkey_next(k); t->end = *k; @@ -574,14 +674,17 @@ static void bset_build_written_tree(struct btree *b) j = inorder_next(j, t->size)) make_bfloat(t, j); } +EXPORT_SYMBOL(bch_bset_build_written_tree); + +/* Insert */ -void bch_bset_fix_invalidated_key(struct btree *b, struct bkey *k) +void bch_bset_fix_invalidated_key(struct btree_keys *b, struct bkey *k) { struct bset_tree *t; unsigned inorder, j = 1; - for (t = b->sets; t <= &b->sets[b->nsets]; t++) - if (k < end(t->data)) + for (t = b->set; t <= bset_tree_last(b); t++) + if (k < bset_bkey_last(t->data)) goto found_set; BUG(); @@ -594,7 +697,7 @@ found_set: if (k == t->data->start) goto fix_left; - if (bkey_next(k) == end(t->data)) { + if (bkey_next(k) == bset_bkey_last(t->data)) { t->end = *k; goto fix_right; } @@ -619,10 +722,12 @@ fix_right: do { j = j * 2 + 1; } while (j < t->size); } +EXPORT_SYMBOL(bch_bset_fix_invalidated_key); -void bch_bset_fix_lookup_table(struct btree *b, struct bkey *k) +static void bch_bset_fix_lookup_table(struct btree_keys *b, + struct bset_tree *t, + struct bkey *k) { - struct bset_tree *t = &b->sets[b->nsets]; unsigned shift = bkey_u64s(k); unsigned j = bkey_to_cacheline(t, k); @@ -634,8 +739,8 @@ void bch_bset_fix_lookup_table(struct btree *b, struct bkey *k) * lookup table for the first key that is strictly greater than k: * it's either k's cacheline or the next one */ - if (j < t->size && - table_to_bkey(t, j) <= k) + 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 @@ -650,54 +755,124 @@ void bch_bset_fix_lookup_table(struct btree *b, struct bkey *k) while (k < cacheline_to_bkey(t, j, 0)) k = bkey_next(k); - t->prev[j] = bkey_to_cacheline_offset(k); + t->prev[j] = bkey_to_cacheline_offset(t, j, k); } } - if (t->size == b->sets->tree + bset_tree_space(b) - t->tree) + 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 != end(t->data); + 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(k); + t->prev[t->size] = bkey_to_cacheline_offset(t, t->size, k); t->size++; } } -void bch_bset_init_next(struct btree *b) +/* + * 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); +} +EXPORT_SYMBOL(bch_bkey_try_merge); + +void bch_bset_insert(struct btree_keys *b, struct bkey *where, + struct bkey *insert) { - struct bset *i = write_block(b); + struct bset_tree *t = bset_tree_last(b); - if (i != b->sets[0].data) { - b->sets[++b->nsets].data = i; - i->seq = b->sets[0].data->seq; - } else - get_random_bytes(&i->seq, sizeof(uint64_t)); + 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); - i->magic = bset_magic(b->c); - i->version = 0; - i->keys = 0; + memmove((uint64_t *) where + bkey_u64s(insert), + where, + (void *) bset_bkey_last(t->data) - (void *) where); - bset_build_unwritten_tree(b); + t->data->keys += bkey_u64s(insert); + bkey_copy(where, insert); + bch_bset_fix_lookup_table(b, t, where); } +EXPORT_SYMBOL(bch_bset_insert); + +unsigned bch_btree_insert_key(struct btree_keys *b, struct bkey *k, + struct bkey *replace_key) +{ + unsigned status = BTREE_INSERT_STATUS_NO_INSERT; + struct bset *i = bset_tree_last(b)->data; + struct bkey *m, *prev = NULL; + struct btree_iter iter; + + BUG_ON(b->ops->is_extents && !KEY_SIZE(k)); + + m = bch_btree_iter_init(b, &iter, b->ops->is_extents + ? PRECEDING_KEY(&START_KEY(k)) + : PRECEDING_KEY(k)); + + 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; +} +EXPORT_SYMBOL(bch_btree_insert_key); + +/* Lookup */ struct bset_search_iter { struct bkey *l, *r; }; -static struct bset_search_iter bset_search_write_set(struct btree *b, - struct bset_tree *t, +static struct bset_search_iter bset_search_write_set(struct bset_tree *t, const struct bkey *search) { unsigned li = 0, ri = t->size; - BUG_ON(!b->nsets && - t->size < bkey_to_cacheline(t, end(t->data))); - while (li + 1 != ri) { unsigned m = (li + ri) >> 1; @@ -709,12 +884,11 @@ static struct bset_search_iter bset_search_write_set(struct btree *b, return (struct bset_search_iter) { table_to_bkey(t, li), - ri < t->size ? table_to_bkey(t, ri) : end(t->data) + ri < t->size ? table_to_bkey(t, ri) : bset_bkey_last(t->data) }; } -static struct bset_search_iter bset_search_tree(struct btree *b, - struct bset_tree *t, +static struct bset_search_iter bset_search_tree(struct bset_tree *t, const struct bkey *search) { struct bkey *l, *r; @@ -761,7 +935,7 @@ static struct bset_search_iter bset_search_tree(struct btree *b, f = &t->tree[inorder_next(j, t->size)]; r = cacheline_to_bkey(t, inorder, f->m); } else - r = end(t->data); + r = bset_bkey_last(t->data); } else { r = cacheline_to_bkey(t, inorder, f->m); @@ -775,7 +949,7 @@ static struct bset_search_iter bset_search_tree(struct btree *b, return (struct bset_search_iter) {l, r}; } -struct bkey *__bch_bset_search(struct btree *b, struct bset_tree *t, +struct bkey *__bch_bset_search(struct btree_keys *b, struct bset_tree *t, const struct bkey *search) { struct bset_search_iter i; @@ -797,7 +971,7 @@ struct bkey *__bch_bset_search(struct btree *b, struct bset_tree *t, if (unlikely(!t->size)) { i.l = t->data->start; - i.r = end(t->data); + i.r = bset_bkey_last(t->data); } else if (bset_written(b, t)) { /* * Each node in the auxiliary search tree covers a certain range @@ -807,25 +981,29 @@ struct bkey *__bch_bset_search(struct btree *b, struct bset_tree *t, */ if (unlikely(bkey_cmp(search, &t->end) >= 0)) - return end(t->data); + return bset_bkey_last(t->data); if (unlikely(bkey_cmp(search, t->data->start) < 0)) return t->data->start; - i = bset_search_tree(b, t, search); - } else - i = bset_search_write_set(b, t, search); + 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); + } -#ifdef CONFIG_BCACHE_EDEBUG - 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); + 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 != end(t->data) && - bkey_cmp(i.r, search) <= 0); -#endif + 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) @@ -833,15 +1011,17 @@ struct bkey *__bch_bset_search(struct btree *b, struct bset_tree *t, return i.l; } +EXPORT_SYMBOL(__bch_bset_search); /* 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) { - int64_t c = bkey_cmp(&START_KEY(l.k), &START_KEY(r.k)); - - return c ? c > 0 : l.k < r.k; + return bkey_cmp(l.k, r.k) > 0; } static inline bool btree_iter_end(struct btree_iter *iter) @@ -858,27 +1038,44 @@ void bch_btree_iter_push(struct btree_iter *iter, struct bkey *k, btree_iter_cmp)); } -struct bkey *__bch_btree_iter_init(struct btree *b, struct btree_iter *iter, - struct bkey *search, struct bset_tree *start) +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; - for (; start <= &b->sets[b->nsets]; start++) { +#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, end(start->data)); + bch_btree_iter_push(iter, ret, bset_bkey_last(start->data)); } return ret; } -struct bkey *bch_btree_iter_next(struct btree_iter *iter) +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); +} +EXPORT_SYMBOL(bch_btree_iter_init); + +static inline struct bkey *__bch_btree_iter_next(struct btree_iter *iter, + btree_iter_cmp_fn *cmp) { struct btree_iter_set 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); @@ -888,16 +1085,23 @@ struct bkey *bch_btree_iter_next(struct btree_iter *iter) } if (iter->data->k == iter->data->end) - heap_pop(iter, unused, btree_iter_cmp); + heap_pop(iter, unused, cmp); else - heap_sift(iter, 0, btree_iter_cmp); + 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); + +} +EXPORT_SYMBOL(bch_btree_iter_next); + struct bkey *bch_btree_iter_next_filter(struct btree_iter *iter, - struct btree *b, ptr_filter_fn fn) + struct btree_keys *b, ptr_filter_fn fn) { struct bkey *ret; @@ -908,63 +1112,60 @@ struct bkey *bch_btree_iter_next_filter(struct btree_iter *iter, return ret; } -struct bkey *bch_next_recurse_key(struct btree *b, struct bkey *search) -{ - struct btree_iter iter; - - bch_btree_iter_init(b, &iter, search); - return bch_btree_iter_next_filter(&iter, b, bch_ptr_bad); -} - /* Mergesort */ -static void btree_sort_fixup(struct btree_iter *iter) +void bch_bset_sort_state_free(struct bset_sort_state *state) { - while (iter->used > 1) { - struct btree_iter_set *top = iter->data, *i = top + 1; - struct bkey *k; + if (state->pool) + mempool_destroy(state->pool); +} - if (iter->used > 2 && - btree_iter_cmp(i[0], i[1])) - i++; +int bch_bset_sort_state_init(struct bset_sort_state *state, unsigned page_order) +{ + spin_lock_init(&state->time.lock); - for (k = i->k; - k != i->end && bkey_cmp(top->k, &START_KEY(k)) > 0; - k = bkey_next(k)) - if (top->k > i->k) - __bch_cut_front(top->k, k); - else if (KEY_SIZE(k)) - bch_cut_back(&START_KEY(k), top->k); + state->page_order = page_order; + state->crit_factor = int_sqrt(1 << page_order); - if (top->k < i->k || k == i->k) - break; + state->pool = mempool_create_page_pool(1, page_order); + if (!state->pool) + return -ENOMEM; - heap_sift(iter, i - top, btree_iter_cmp); - } + return 0; } +EXPORT_SYMBOL(bch_bset_sort_state_init); -static void btree_mergesort(struct btree *b, struct bset *out, +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; - bool (*bad)(struct btree *, const struct bkey *) = remove_stale + 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 (fixup && !b->level) - btree_sort_fixup(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); - k = bch_btree_iter_next(iter); if (bad(b, k)) continue; if (!last) { last = out->start; bkey_copy(last, k); - } else if (b->level || - !bch_bkey_try_merge(b, last, k)) { + } else if (!bch_bkey_try_merge(b, last, k)) { last = bkey_next(last); bkey_copy(last, k); } @@ -973,30 +1174,32 @@ static void btree_mergesort(struct btree *b, struct bset *out, out->keys = last ? (uint64_t *) bkey_next(last) - out->d : 0; pr_debug("sorted %i keys", out->keys); - bch_check_key_order(b, out); } -static void __btree_sort(struct btree *b, struct btree_iter *iter, - unsigned start, unsigned order, bool fixup) +static void __btree_sort(struct btree_keys *b, struct btree_iter *iter, + unsigned start, unsigned order, bool fixup, + struct bset_sort_state *state) { uint64_t start_time; - bool remove_stale = !b->written; + bool used_mempool = false; struct bset *out = (void *) __get_free_pages(__GFP_NOWARN|GFP_NOIO, order); if (!out) { - mutex_lock(&b->c->sort_lock); - out = b->c->sort; - order = ilog2(bucket_pages(b->c)); + 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, remove_stale); + btree_mergesort(b, out, iter, fixup, false); b->nsets = start; - if (!fixup && !start && b->written) - bch_btree_verify(b, out); - if (!start && order == b->page_order) { /* * Our temporary buffer is the same size as the btree node's @@ -1004,135 +1207,110 @@ static void __btree_sort(struct btree *b, struct btree_iter *iter, * memcpy() */ - out->magic = bset_magic(b->c); - out->seq = b->sets[0].data->seq; - out->version = b->sets[0].data->version; - swap(out, b->sets[0].data); - - if (b->c->sort == b->sets[0].data) - b->c->sort = out; + 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->sets[start].data->keys = out->keys; - memcpy(b->sets[start].data->start, out->start, - (void *) end(out) - (void *) out->start); + b->set[start].data->keys = out->keys; + memcpy(b->set[start].data->start, out->start, + (void *) bset_bkey_last(out) - (void *) out->start); } - if (out == b->c->sort) - mutex_unlock(&b->c->sort_lock); + if (used_mempool) + mempool_free(virt_to_page(out), state->pool); else free_pages((unsigned long) out, order); - if (b->written) - bset_build_written_tree(b); + bch_bset_build_written_tree(b); - if (!start) { - spin_lock(&b->c->sort_time_lock); - bch_time_stats_update(&b->c->sort_time, start_time); - spin_unlock(&b->c->sort_time_lock); - } + if (!start) + bch_time_stats_update(&state->time, start_time); } -void bch_btree_sort_partial(struct btree *b, unsigned start) +void bch_btree_sort_partial(struct btree_keys *b, unsigned start, + struct bset_sort_state *state) { - size_t oldsize = 0, order = b->page_order, keys = 0; + size_t order = b->page_order, keys = 0; struct btree_iter iter; - __bch_btree_iter_init(b, &iter, NULL, &b->sets[start]); - - BUG_ON(b->sets[b->nsets].data == write_block(b) && - (b->sets[b->nsets].size || b->nsets)); + int oldsize = bch_count_data(b); - if (b->written) - oldsize = bch_count_data(b); + __bch_btree_iter_init(b, &iter, NULL, &b->set[start]); if (start) { unsigned i; for (i = start; i <= b->nsets; i++) - keys += b->sets[i].data->keys; + keys += b->set[i].data->keys; - order = roundup_pow_of_two(__set_bytes(b->sets->data, - keys)) / PAGE_SIZE; - if (order) - order = ilog2(order); + order = get_order(__set_bytes(b->set->data, keys)); } - __btree_sort(b, &iter, start, order, false); + __btree_sort(b, &iter, start, order, false, state); - EBUG_ON(b->written && bch_count_data(b) != oldsize); + EBUG_ON(oldsize >= 0 && bch_count_data(b) != oldsize); } +EXPORT_SYMBOL(bch_btree_sort_partial); -void bch_btree_sort_and_fix_extents(struct btree *b, struct btree_iter *iter) +void bch_btree_sort_and_fix_extents(struct btree_keys *b, + struct btree_iter *iter, + struct bset_sort_state *state) { - BUG_ON(!b->written); - __btree_sort(b, iter, 0, b->page_order, true); + __btree_sort(b, iter, 0, b->page_order, true, state); } -void bch_btree_sort_into(struct btree *b, struct btree *new) +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->sets->data, &iter, false, true); + btree_mergesort(b, new->set->data, &iter, false, true); - spin_lock(&b->c->sort_time_lock); - bch_time_stats_update(&b->c->sort_time, start_time); - spin_unlock(&b->c->sort_time_lock); + bch_time_stats_update(&state->time, start_time); - bkey_copy_key(&new->key, &b->key); - new->sets->size = 0; + new->set->size = 0; // XXX: why? } -void bch_btree_sort_lazy(struct btree *b) -{ - if (b->nsets) { - unsigned i, j, keys = 0, total; - - for (i = 0; i <= b->nsets; i++) - keys += b->sets[i].data->keys; +#define SORT_CRIT (4096 / sizeof(uint64_t)) - total = keys; +void bch_btree_sort_lazy(struct btree_keys *b, struct bset_sort_state *state) +{ + unsigned crit = SORT_CRIT; + int i; - for (j = 0; j < b->nsets; j++) { - if (keys * 2 < total || - keys < 1000) { - bch_btree_sort_partial(b, j); - return; - } + /* Don't sort if nothing to do */ + if (!b->nsets) + goto out; - keys -= b->sets[j].data->keys; - } + for (i = b->nsets - 1; i >= 0; --i) { + crit *= state->crit_factor; - /* Must sort if b->nsets == 3 or we'll overflow */ - if (b->nsets >= (MAX_BSETS - 1) - b->level) { - bch_btree_sort(b); + if (b->set[i].data->keys < crit) { + bch_btree_sort_partial(b, i, state); return; } } - bset_build_written_tree(b); -} - -/* Sysfs stuff */ + /* Sort if we'd overflow */ + if (b->nsets + 1 == MAX_BSETS) { + bch_btree_sort(b, state); + return; + } -struct bset_stats { - size_t nodes; - size_t sets_written, sets_unwritten; - size_t bytes_written, bytes_unwritten; - size_t floats, failed; -}; +out: + bch_bset_build_written_tree(b); +} +EXPORT_SYMBOL(bch_btree_sort_lazy); -static int bch_btree_bset_stats(struct btree *b, struct btree_op *op, - struct bset_stats *stats) +void bch_btree_keys_stats(struct btree_keys *b, struct bset_stats *stats) { - struct bkey *k; unsigned i; - stats->nodes++; - for (i = 0; i <= b->nsets; i++) { - struct bset_tree *t = &b->sets[i]; + struct bset_tree *t = &b->set[i]; size_t bytes = t->data->keys * sizeof(uint64_t); size_t j; @@ -1150,43 +1328,4 @@ static int bch_btree_bset_stats(struct btree *b, struct btree_op *op, stats->bytes_unwritten += bytes; } } - - if (b->level) { - struct btree_iter iter; - - for_each_key_filter(b, k, &iter, bch_ptr_bad) { - int ret = btree(bset_stats, k, b, op, stats); - if (ret) - return ret; - } - } - - return 0; -} - -int bch_bset_print_stats(struct cache_set *c, char *buf) -{ - struct btree_op op; - struct bset_stats t; - int ret; - - bch_btree_op_init_stack(&op); - memset(&t, 0, sizeof(struct bset_stats)); - - ret = btree_root(bset_stats, c, &op, &t); - if (ret) - 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", - t.nodes, - t.sets_written, t.sets_unwritten, - t.bytes_written, t.bytes_unwritten, - t.floats, t.failed); } diff --git a/drivers/md/bcache/bset.h b/drivers/md/bcache/bset.h index 57a9cff4154..5f6728d5d4d 100644 --- a/drivers/md/bcache/bset.h +++ b/drivers/md/bcache/bset.h @@ -1,6 +1,12 @@ #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: * @@ -140,14 +146,12 @@ * first key in that range of bytes again. */ -/* Btree key comparison/iteration */ +struct btree_keys; +struct btree_iter; +struct btree_iter_set; +struct bkey_float; -struct btree_iter { - size_t size, used; - struct btree_iter_set { - struct bkey *k, *end; - } data[MAX_BSETS]; -}; +#define MAX_BSETS 4U struct bset_tree { /* @@ -158,14 +162,14 @@ struct bset_tree { */ /* size of the binary tree and prev array */ - unsigned size; + unsigned size; /* function of size - precalculated for to_inorder() */ - unsigned extra; + unsigned extra; /* copy of the last key in the set */ - struct bkey end; - struct bkey_float *tree; + struct bkey end; + struct bkey_float *tree; /* * The nodes in the bset tree point to specific keys - this @@ -175,96 +179,227 @@ struct bset_tree { * to keep bkey_float to 4 bytes and prev isn't used in the fast * path. */ - uint8_t *prev; + uint8_t *prev; /* The actual btree node, with pointers to each sorted set */ - struct bset *data; + struct bset *data; }; -static __always_inline int64_t bkey_cmp(const struct bkey *l, - const struct bkey *r) +struct btree_keys_ops { + bool (*sort_cmp)(struct btree_iter_set, + struct btree_iter_set); + struct bkey *(*sort_fixup)(struct btree_iter *, struct bkey *); + bool (*insert_fixup)(struct btree_keys *, struct bkey *, + struct btree_iter *, struct bkey *); + bool (*key_invalid)(struct btree_keys *, + const struct bkey *); + bool (*key_bad)(struct btree_keys *, const struct bkey *); + bool (*key_merge)(struct btree_keys *, + struct bkey *, struct bkey *); + void (*key_to_text)(char *, size_t, const struct bkey *); + void (*key_dump)(struct btree_keys *, const struct bkey *); + + /* + * 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 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 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); + return b->set + b->nsets; } -static inline size_t bkey_u64s(const struct bkey *k) +static inline bool bset_written(struct btree_keys *b, struct bset_tree *t) { - BUG_ON(KEY_CSUM(k) > 1); - return 2 + KEY_PTRS(k) + (KEY_CSUM(k) ? 1 : 0); + return t <= b->set + b->nsets - b->last_set_unwritten; } -static inline size_t bkey_bytes(const struct bkey *k) +static inline bool bkey_written(struct btree_keys *b, struct bkey *k) { - return bkey_u64s(k) * sizeof(uint64_t); + return !b->last_set_unwritten || k < b->set[b->nsets].data->start; } -static inline void bkey_copy(struct bkey *dest, const struct bkey *src) +static inline unsigned bset_byte_offset(struct btree_keys *b, struct bset *i) { - memcpy(dest, src, bkey_bytes(src)); + return ((size_t) i) - ((size_t) b->set->data); } -static inline void bkey_copy_key(struct bkey *dest, const struct bkey *src) +static inline unsigned bset_sector_offset(struct btree_keys *b, struct bset *i) { - if (!src) - src = &KEY(0, 0, 0); + 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) - SET_KEY_INODE(dest, KEY_INODE(src)); - SET_KEY_OFFSET(dest, KEY_OFFSET(src)); +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 bkey *bkey_next(const struct bkey *k) +static inline struct bset *bset_next_set(struct btree_keys *b, + unsigned block_bytes) { - uint64_t *d = (void *) k; - return (struct bkey *) (d + bkey_u64s(k)); + struct bset *i = bset_tree_last(b)->data; + + return ((void *) i) + roundup(set_bytes(i), block_bytes); } -/* Keylists */ +void bch_btree_keys_free(struct btree_keys *); +int bch_btree_keys_alloc(struct btree_keys *, unsigned, gfp_t); +void bch_btree_keys_init(struct btree_keys *, const struct btree_keys_ops *, + bool *); + +void bch_bset_init_next(struct btree_keys *, struct bset *, uint64_t); +void bch_bset_build_written_tree(struct btree_keys *); +void bch_bset_fix_invalidated_key(struct btree_keys *, struct bkey *); +bool bch_bkey_try_merge(struct btree_keys *, struct bkey *, struct bkey *); +void bch_bset_insert(struct btree_keys *, struct bkey *, struct bkey *); +unsigned bch_btree_insert_key(struct btree_keys *, struct bkey *, + struct bkey *); + +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, +}; -struct keylist { - struct bkey *top; - union { - uint64_t *list; - struct bkey *bottom; - }; +/* Btree key iteration */ - /* Enough room for btree_split's keys without realloc */ -#define KEYLIST_INLINE 16 - uint64_t d[KEYLIST_INLINE]; +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]; }; -static inline void bch_keylist_init(struct keylist *l) +typedef bool (*ptr_filter_fn)(struct btree_keys *, const struct bkey *); + +struct bkey *bch_btree_iter_next(struct btree_iter *); +struct bkey *bch_btree_iter_next_filter(struct btree_iter *, + struct btree_keys *, ptr_filter_fn); + +void bch_btree_iter_push(struct btree_iter *, struct bkey *, struct bkey *); +struct bkey *bch_btree_iter_init(struct btree_keys *, struct btree_iter *, + struct bkey *); + +struct bkey *__bch_bset_search(struct btree_keys *, struct bset_tree *, + const struct bkey *); + +/* + * 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) { - l->top = (void *) (l->list = l->d); + return search ? __bch_bset_search(b, t, search) : t->data->start; } -static inline void bch_keylist_push(struct keylist *l) +#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 page_order; + unsigned crit_factor; + + struct time_stats time; +}; + +void bch_bset_sort_state_free(struct bset_sort_state *); +int bch_bset_sort_state_init(struct bset_sort_state *, unsigned); +void bch_btree_sort_lazy(struct btree_keys *, struct bset_sort_state *); +void bch_btree_sort_into(struct btree_keys *, struct btree_keys *, + struct bset_sort_state *); +void bch_btree_sort_and_fix_extents(struct btree_keys *, struct btree_iter *, + struct bset_sort_state *); +void bch_btree_sort_partial(struct btree_keys *, unsigned, + struct bset_sort_state *); + +static inline void bch_btree_sort(struct btree_keys *b, + struct bset_sort_state *state) { - l->top = bkey_next(l->top); + bch_btree_sort_partial(b, 0, state); } -static inline void bch_keylist_add(struct keylist *l, struct bkey *k) +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 *, struct bset_stats *); + +/* Bkey utility code */ + +#define bset_bkey_last(i) bkey_idx((struct bkey *) (i)->d, (i)->keys) + +static inline struct bkey *bset_bkey_idx(struct bset *i, unsigned idx) { - bkey_copy(l->top, k); - bch_keylist_push(l); + return bkey_idx(i->start, idx); } -static inline bool bch_keylist_empty(struct keylist *l) +static inline void bkey_init(struct bkey *k) { - return l->top == (void *) l->list; + *k = ZERO_KEY; } -static inline void bch_keylist_free(struct keylist *l) +static __always_inline int64_t bkey_cmp(const struct bkey *l, + const struct bkey *r) { - if (l->list != l->d) - kfree(l->list); + 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_keylist_copy(struct keylist *, struct keylist *); -struct bkey *bch_keylist_pop(struct keylist *); -int bch_keylist_realloc(struct keylist *, int, struct cache_set *); - void bch_bkey_copy_single_ptr(struct bkey *, const struct bkey *, unsigned); bool __bch_cut_front(const struct bkey *, struct bkey *); @@ -282,98 +417,150 @@ static inline bool bch_cut_back(const struct bkey *where, struct bkey *k) return __bch_cut_back(where, k); } -const char *bch_ptr_status(struct cache_set *, const struct bkey *); -bool __bch_ptr_invalid(struct cache_set *, int level, const struct bkey *); -bool bch_ptr_bad(struct btree *, const struct bkey *); - -static inline uint8_t gen_after(uint8_t a, uint8_t b) +#define PRECEDING_KEY(_k) \ +({ \ + struct bkey *_ret = NULL; \ + \ + if (KEY_INODE(_k) || KEY_OFFSET(_k)) { \ + _ret = &KEY(KEY_INODE(_k), KEY_OFFSET(_k), 0); \ + \ + if (!_ret->low) \ + _ret->high--; \ + _ret->low--; \ + } \ + \ + _ret; \ +}) + +static inline bool bch_ptr_invalid(struct btree_keys *b, const struct bkey *k) { - uint8_t r = a - b; - return r > 128U ? 0 : r; + return b->ops->key_invalid(b, k); } -static inline uint8_t ptr_stale(struct cache_set *c, const struct bkey *k, - unsigned i) +static inline bool bch_ptr_bad(struct btree_keys *b, const struct bkey *k) { - return gen_after(PTR_BUCKET(c, k, i)->gen, PTR_GEN(k, i)); + return b->ops->key_bad(b, k); } -static inline bool ptr_available(struct cache_set *c, const struct bkey *k, - unsigned i) +static inline void bch_bkey_to_text(struct btree_keys *b, char *buf, + size_t size, const struct bkey *k) { - return (PTR_DEV(k, i) < MAX_CACHES_PER_SET) && PTR_CACHE(c, k, i); + 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(l)); +} -typedef bool (*ptr_filter_fn)(struct btree *, const struct bkey *); +/* Keylists */ -struct bkey *bch_next_recurse_key(struct btree *, struct bkey *); -struct bkey *bch_btree_iter_next(struct btree_iter *); -struct bkey *bch_btree_iter_next_filter(struct btree_iter *, - struct btree *, ptr_filter_fn); +struct keylist { + union { + struct bkey *keys; + uint64_t *keys_p; + }; + union { + struct bkey *top; + uint64_t *top_p; + }; -void bch_btree_iter_push(struct btree_iter *, struct bkey *, struct bkey *); -struct bkey *__bch_btree_iter_init(struct btree *, struct btree_iter *, - struct bkey *, struct bset_tree *); + /* Enough room for btree_split's keys without realloc */ +#define KEYLIST_INLINE 16 + uint64_t inline_keys[KEYLIST_INLINE]; +}; -/* 32 bits total: */ -#define BKEY_MID_BITS 3 -#define BKEY_EXPONENT_BITS 7 -#define BKEY_MANTISSA_BITS 22 -#define BKEY_MANTISSA_MASK ((1 << BKEY_MANTISSA_BITS) - 1) +static inline void bch_keylist_init(struct keylist *l) +{ + l->top_p = l->keys_p = l->inline_keys; +} -struct bkey_float { - unsigned exponent:BKEY_EXPONENT_BITS; - unsigned m:BKEY_MID_BITS; - unsigned mantissa:BKEY_MANTISSA_BITS; -} __packed; +static inline void bch_keylist_init_single(struct keylist *l, struct bkey *k) +{ + l->keys = k; + l->top = bkey_next(k); +} -/* - * 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. - */ +static inline void bch_keylist_push(struct keylist *l) +{ + l->top = bkey_next(l->top); +} -#define BSET_CACHELINE 128 -#define bset_tree_space(b) (btree_data_space(b) / BSET_CACHELINE) +static inline void bch_keylist_add(struct keylist *l, struct bkey *k) +{ + bkey_copy(l->top, k); + bch_keylist_push(l); +} -#define bset_tree_bytes(b) (bset_tree_space(b) * sizeof(struct bkey_float)) -#define bset_prev_bytes(b) (bset_tree_space(b) * sizeof(uint8_t)) +static inline bool bch_keylist_empty(struct keylist *l) +{ + return l->top == l->keys; +} -void bch_bset_init_next(struct btree *); +static inline void bch_keylist_reset(struct keylist *l) +{ + l->top = l->keys; +} -void bch_bset_fix_invalidated_key(struct btree *, struct bkey *); -void bch_bset_fix_lookup_table(struct btree *, struct bkey *); +static inline void bch_keylist_free(struct keylist *l) +{ + if (l->keys_p != l->inline_keys) + kfree(l->keys_p); +} -struct bkey *__bch_bset_search(struct btree *, struct bset_tree *, - const struct bkey *); +static inline size_t bch_keylist_nkeys(struct keylist *l) +{ + return l->top_p - l->keys_p; +} -static inline struct bkey *bch_bset_search(struct btree *b, struct bset_tree *t, - const struct bkey *search) +static inline size_t bch_keylist_bytes(struct keylist *l) { - return search ? __bch_bset_search(b, t, search) : t->data->start; + return bch_keylist_nkeys(l) * sizeof(uint64_t); } -bool bch_bkey_try_merge(struct btree *, struct bkey *, struct bkey *); -void bch_btree_sort_lazy(struct btree *); -void bch_btree_sort_into(struct btree *, struct btree *); -void bch_btree_sort_and_fix_extents(struct btree *, struct btree_iter *); -void bch_btree_sort_partial(struct btree *, unsigned); +struct bkey *bch_keylist_pop(struct keylist *); +void bch_keylist_pop_front(struct keylist *); +int __bch_keylist_realloc(struct keylist *, unsigned); + +/* Debug stuff */ + +#ifdef CONFIG_BCACHE_DEBUG + +int __bch_count_data(struct btree_keys *); +void __bch_check_keys(struct btree_keys *, const char *, ...); +void bch_dump_bset(struct btree_keys *, struct bset *, unsigned); +void bch_dump_bucket(struct btree_keys *); + +#else + +static inline int __bch_count_data(struct btree_keys *b) { return -1; } +static inline void __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 *, struct bset *, unsigned); + +#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 void bch_btree_sort(struct btree *b) +static inline int bch_count_data(struct btree_keys *b) { - bch_btree_sort_partial(b, 0); + return btree_keys_expensive_checks(b) ? __bch_count_data(b) : -1; } -int bch_bset_print_stats(struct cache_set *, char *); +#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 index 7a5658f04e6..7347b610096 100644 --- a/drivers/md/bcache/btree.c +++ b/drivers/md/bcache/btree.c @@ -23,11 +23,13 @@ #include "bcache.h" #include "btree.h" #include "debug.h" -#include "request.h" +#include "extents.h" #include <linux/slab.h> #include <linux/bitops.h> +#include <linux/freezer.h> #include <linux/hash.h> +#include <linux/kthread.h> #include <linux/prefetch.h> #include <linux/random.h> #include <linux/rcupdate.h> @@ -66,15 +68,11 @@ * alloc_bucket() cannot fail. This should be true but is not completely * obvious. * - * Make sure all allocations get charged to the root cgroup - * * Plugging? * * If data write is less than hard sector size of ssd, round up offset in open * bucket to the next whole sector * - * Also lookup by cgroup in get_open_bucket() - * * Superblock needs to be fleshed out for multiple cache devices * * Add a sysfs tunable for the number of writeback IOs in flight @@ -87,15 +85,6 @@ * Test module load/unload */ -static const char * const op_types[] = { - "insert", "replace" -}; - -static const char *op_type(struct btree_op *op) -{ - return op_types[op->type]; -} - #define MAX_NEED_GC 64 #define MAX_SAVE_PRIO 72 @@ -104,23 +93,96 @@ static const char *op_type(struct btree_op *op) #define PTR_HASH(c, k) \ (((k)->ptr[0] >> c->bucket_bits) | PTR_GEN(k, 0)) -struct workqueue_struct *bch_gc_wq; -static struct workqueue_struct *btree_io_wq; +#define insert_lock(s, b) ((b)->level <= (s)->lock) + +/* + * 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 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);\ + if (!IS_ERR(_child)) { \ + _child->parent = (b); \ + _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 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)) { \ + _b->parent = NULL; \ + _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; \ +}) + +static inline struct bset *write_block(struct btree *b) +{ + return ((void *) btree_bset_first(b)) + b->written * block_bytes(b->c); +} + +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->sb)); -void bch_btree_op_init_stack(struct btree_op *op) -{ - memset(op, 0, sizeof(struct btree_op)); - closure_init_stack(&op->cl); - op->lock = -1; - bch_keylist_init(&op->keys); } /* Btree key manipulation */ -static void bkey_put(struct cache_set *c, struct bkey *k, int level) +void bkey_put(struct cache_set *c, struct bkey *k) { - if ((level && KEY_OFFSET(k)) || !level) - __bkey_put(c, k); + unsigned 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 */ @@ -128,65 +190,43 @@ static void bkey_put(struct cache_set *c, struct bkey *k, int level) 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 = end(i); + void *data = (void *) i + 8, *end = bset_bkey_last(i); crc = bch_crc64_update(crc, data, end - data); return crc ^ 0xffffffffffffffffULL; } -static void btree_bio_endio(struct bio *bio, int error) +void bch_btree_node_read_done(struct btree *b) { - struct closure *cl = bio->bi_private; - struct btree *b = container_of(cl, struct btree, io.cl); - - if (error) - set_btree_node_io_error(b); - - bch_bbio_count_io_errors(b->c, bio, error, (bio->bi_rw & WRITE) - ? "writing btree" : "reading btree"); - closure_put(cl); -} - -static void btree_bio_init(struct btree *b) -{ - BUG_ON(b->bio); - b->bio = bch_bbio_alloc(b->c); - - b->bio->bi_end_io = btree_bio_endio; - b->bio->bi_private = &b->io.cl; -} - -void bch_btree_read_done(struct closure *cl) -{ - struct btree *b = container_of(cl, struct btree, io.cl); - struct bset *i = b->sets[0].data; - struct btree_iter *iter = b->c->fill_iter; const char *err = "bad btree header"; - BUG_ON(b->nsets || b->written); + struct bset *i = btree_bset_first(b); + struct btree_iter *iter; - bch_bbio_free(b->bio, b->c); - b->bio = NULL; - - mutex_lock(&b->c->fill_lock); + iter = mempool_alloc(b->c->fill_iter, GFP_NOWAIT); + iter->size = b->c->sb.bucket_size / b->c->sb.block_size; iter->used = 0; - if (btree_node_io_error(b) || - !i->seq) +#ifdef CONFIG_BCACHE_DEBUG + iter->b = &b->keys; +#endif + + if (!i->seq) goto err; for (; - b->written < btree_blocks(b) && i->seq == b->sets[0].data->seq; + 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, b->c) > btree_blocks(b)) + if (b->written + set_blocks(i, block_bytes(b->c)) > + btree_blocks(b)) goto err; err = "bad magic"; - if (i->magic != bset_magic(b->c)) + if (i->magic != bset_magic(&b->c->sb)) goto err; err = "bad checksum"; @@ -202,95 +242,112 @@ void bch_btree_read_done(struct closure *cl) } err = "empty set"; - if (i != b->sets[0].data && !i->keys) + if (i != b->keys.set[0].data && !i->keys) goto err; - bch_btree_iter_push(iter, i->start, end(i)); + bch_btree_iter_push(iter, i->start, bset_bkey_last(i)); - b->written += set_blocks(i, b->c); + b->written += set_blocks(i, block_bytes(b->c)); } err = "corrupted btree"; for (i = write_block(b); - index(i, b) < btree_blocks(b); + bset_sector_offset(&b->keys, i) < KEY_SIZE(&b->key); i = ((void *) i) + block_bytes(b->c)) - if (i->seq == b->sets[0].data->seq) + if (i->seq == b->keys.set[0].data->seq) goto err; - bch_btree_sort_and_fix_extents(b, iter); + bch_btree_sort_and_fix_extents(&b->keys, iter, &b->c->sort); - i = b->sets[0].data; + i = b->keys.set[0].data; err = "short btree key"; - if (b->sets[0].size && - bkey_cmp(&b->key, &b->sets[0].end) < 0) + 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); + bch_bset_init_next(&b->keys, write_block(b), + bset_magic(&b->c->sb)); out: - - mutex_unlock(&b->c->fill_lock); - - spin_lock(&b->c->btree_read_time_lock); - bch_time_stats_update(&b->c->btree_read_time, b->io_start_time); - spin_unlock(&b->c->btree_read_time_lock); - - smp_wmb(); /* read_done is our write lock */ - set_btree_node_read_done(b); - - closure_return(cl); + 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 %zu, %u keys", + bch_cache_set_error(b->c, "%s at bucket %zu, block %u, %u keys", err, PTR_BUCKET_NR(b->c, &b->key, 0), - index(i, b), i->keys); + bset_block_offset(b, i), i->keys); goto out; } -void bch_btree_read(struct btree *b) +static void btree_node_read_endio(struct bio *bio, int error) +{ + struct closure *cl = bio->bi_private; + closure_put(cl); +} + +static void bch_btree_node_read(struct btree *b) { - BUG_ON(b->nsets || b->written); + uint64_t start_time = local_clock(); + struct closure cl; + struct bio *bio; + + trace_bcache_btree_read(b); - if (!closure_trylock(&b->io.cl, &b->c->cl)) - BUG(); + closure_init_stack(&cl); + + bio = bch_bbio_alloc(b->c); + bio->bi_rw = REQ_META|READ_SYNC; + bio->bi_iter.bi_size = KEY_SIZE(&b->key) << 9; + bio->bi_end_io = btree_node_read_endio; + bio->bi_private = &cl; - b->io_start_time = local_clock(); + bch_bio_map(bio, b->keys.set[0].data); - btree_bio_init(b); - b->bio->bi_rw = REQ_META|READ_SYNC; - b->bio->bi_size = KEY_SIZE(&b->key) << 9; + bch_submit_bbio(bio, b->c, &b->key, 0); + closure_sync(&cl); - bch_bio_map(b->bio, b->sets[0].data); + if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) + set_btree_node_io_error(b); - pr_debug("%s", pbtree(b)); - trace_bcache_btree_read(b->bio); - bch_submit_bbio(b->bio, b->c, &b->key, 0); + bch_bbio_free(bio, b->c); - continue_at(&b->io.cl, bch_btree_read_done, system_wq); + 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(&b->c->alloc_wait); + wake_up_allocators(b->c); if (w->journal) { atomic_dec_bug(w->journal); __closure_wake_up(&b->c->journal.wait); } - if (w->owner) - closure_put(w->owner); - w->prio_blocked = 0; w->journal = NULL; - w->owner = NULL; } -static void __btree_write_done(struct closure *cl) +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.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); @@ -298,155 +355,199 @@ static void __btree_write_done(struct closure *cl) btree_complete_write(b, w); if (btree_node_dirty(b)) - queue_delayed_work(btree_io_wq, &b->work, - msecs_to_jiffies(30000)); + schedule_delayed_work(&b->work, 30 * HZ); - closure_return(cl); + closure_return_with_destructor(cl, btree_node_write_unlock); } -static void btree_write_done(struct closure *cl) +static void btree_node_write_done(struct closure *cl) { - struct btree *b = container_of(cl, struct btree, io.cl); + struct btree *b = container_of(cl, struct btree, io); struct bio_vec *bv; int n; - __bio_for_each_segment(bv, b->bio, n, 0) + bio_for_each_segment_all(bv, b->bio, n) __free_page(bv->bv_page); - __btree_write_done(cl); + __btree_node_write_done(cl); } -static void do_btree_write(struct btree *b) +static void btree_node_write_endio(struct bio *bio, int error) { - struct closure *cl = &b->io.cl; - struct bset *i = b->sets[b->nsets].data; + struct closure *cl = bio->bi_private; + struct btree *b = container_of(cl, struct btree, io); + + if (error) + set_btree_node_io_error(b); + + bch_bbio_count_io_errors(b->c, bio, error, "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); - btree_bio_init(b); - b->bio->bi_rw = REQ_META|WRITE_SYNC; - b->bio->bi_size = set_blocks(i, b->c) * block_bytes(b->c); + 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_rw = REQ_META|WRITE_SYNC|REQ_FUA; + b->bio->bi_iter.bi_size = roundup(set_bytes(i), block_bytes(b->c)); 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_offset(b, i)); + 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_NOIO)) { + if (!bio_alloc_pages(b->bio, GFP_NOIO)) { int j; struct bio_vec *bv; void *base = (void *) ((unsigned long) i & ~(PAGE_SIZE - 1)); - bio_for_each_segment(bv, b->bio, j) + bio_for_each_segment_all(bv, b->bio, j) memcpy(page_address(bv->bv_page), base + j * PAGE_SIZE, PAGE_SIZE); - trace_bcache_btree_write(b->bio); bch_submit_bbio(b->bio, b->c, &k.key, 0); - continue_at(cl, btree_write_done, NULL); + continue_at(cl, btree_node_write_done, NULL); } else { b->bio->bi_vcnt = 0; bch_bio_map(b->bio, i); - trace_bcache_btree_write(b->bio); bch_submit_bbio(b->bio, b->c, &k.key, 0); closure_sync(cl); - __btree_write_done(cl); + continue_at_nobarrier(cl, __btree_node_write_done, NULL); } } -static void __btree_write(struct btree *b) +void __bch_btree_node_write(struct btree *b, struct closure *parent) { - struct bset *i = b->sets[b->nsets].data; + 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"); - closure_lock(&b->io, &b->c->cl); 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); - bch_check_key_order(b, i); - BUG_ON(b->written && !i->keys); + do_btree_node_write(b); - do_btree_write(b); + atomic_long_add(set_blocks(i, block_bytes(b->c)) * b->c->sb.block_size, + &PTR_CACHE(b->c, &b->key, 0)->btree_sectors_written); - pr_debug("%s block %i keys %i", pbtree(b), b->written, i->keys); + b->written += set_blocks(i, block_bytes(b->c)); +} - b->written += set_blocks(i, b->c); - atomic_long_add(set_blocks(i, b->c) * b->c->sb.block_size, - &PTR_CACHE(b->c, &b->key, 0)->btree_sectors_written); +void bch_btree_node_write(struct btree *b, struct closure *parent) +{ + unsigned nsets = b->keys.nsets; - bch_btree_sort_lazy(b); + lockdep_assert_held(&b->lock); - if (b->written < btree_blocks(b)) - bch_bset_init_next(b); + __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 btree_write_work(struct work_struct *w) +static void bch_btree_node_write_sync(struct btree *b) { - struct btree *b = container_of(to_delayed_work(w), struct btree, work); + struct closure cl; + + closure_init_stack(&cl); - down_write(&b->lock); + 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)) - __btree_write(b); - up_write(&b->lock); + __bch_btree_node_write(b, NULL); + mutex_unlock(&b->write_lock); } -void bch_btree_write(struct btree *b, bool now, struct btree_op *op) +static void bch_btree_leaf_dirty(struct btree *b, atomic_t *journal_ref) { - struct bset *i = b->sets[b->nsets].data; + struct bset *i = btree_bset_last(b); struct btree_write *w = btree_current_write(b); - BUG_ON(b->written && - (b->written >= btree_blocks(b) || - i->seq != b->sets[0].data->seq || - !i->keys)); + lockdep_assert_held(&b->write_lock); - if (!btree_node_dirty(b)) { - set_btree_node_dirty(b); - queue_delayed_work(btree_io_wq, &b->work, - msecs_to_jiffies(30000)); - } + BUG_ON(!b->written); + BUG_ON(!i->keys); + + if (!btree_node_dirty(b)) + schedule_delayed_work(&b->work, 30 * HZ); - w->prio_blocked += b->prio_blocked; - b->prio_blocked = 0; + set_btree_node_dirty(b); - if (op && op->journal && !b->level) { + if (journal_ref) { if (w->journal && - journal_pin_cmp(b->c, w, op)) { + journal_pin_cmp(b->c, w->journal, journal_ref)) { atomic_dec_bug(w->journal); w->journal = NULL; } if (!w->journal) { - w->journal = op->journal; + w->journal = journal_ref; atomic_inc(w->journal); } } - if (current->bio_list) - return; - /* Force write if set is too big */ - if (now || - b->level || - set_bytes(i) > PAGE_SIZE - 48) { - if (op && now) { - /* Must wait on multiple writes */ - BUG_ON(w->owner); - w->owner = &op->cl; - closure_get(&op->cl); - } - - __btree_write(b); - } - BUG_ON(!b->written); + if (set_bytes(i) > PAGE_SIZE - 48 && + !current->bio_list) + bch_btree_node_write(b, NULL); } /* @@ -454,53 +555,19 @@ void bch_btree_write(struct btree *b, bool now, struct btree_op *op) * mca -> memory cache */ -static void mca_reinit(struct btree *b) -{ - unsigned i; - - b->flags = 0; - b->written = 0; - b->nsets = 0; - - for (i = 0; i < MAX_BSETS; i++) - b->sets[i].size = 0; - /* - * Second loop starts at 1 because b->sets[0]->data is the memory we - * allocated - */ - for (i = 1; i < MAX_BSETS; i++) - b->sets[i].data = NULL; -} - #define mca_reserve(c) (((c->root && c->root->level) \ ? c->root->level : 1) * 8 + 16) #define mca_can_free(c) \ - max_t(int, 0, c->bucket_cache_used - mca_reserve(c)) + max_t(int, 0, c->btree_cache_used - mca_reserve(c)) static void mca_data_free(struct btree *b) { - struct bset_tree *t = b->sets; - BUG_ON(!closure_is_unlocked(&b->io.cl)); + BUG_ON(b->io_mutex.count != 1); - if (bset_prev_bytes(b) < PAGE_SIZE) - kfree(t->prev); - else - free_pages((unsigned long) t->prev, - get_order(bset_prev_bytes(b))); + bch_btree_keys_free(&b->keys); - 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; + b->c->btree_cache_used--; list_move(&b->list, &b->c->btree_cache_freed); - b->c->bucket_cache_used--; } static void mca_bucket_free(struct btree *b) @@ -519,34 +586,16 @@ static unsigned btree_order(struct bkey *k) static void mca_data_alloc(struct btree *b, struct bkey *k, gfp_t gfp) { - struct bset_tree *t = b->sets; - BUG_ON(t->data); - - b->page_order = max_t(unsigned, - ilog2(b->c->btree_pages), - btree_order(k)); - - t->data = (void *) __get_free_pages(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; - - list_move(&b->list, &b->c->btree_cache); - b->c->bucket_cache_used++; - return; -err: - mca_data_free(b); + if (!bch_btree_keys_alloc(&b->keys, + max_t(unsigned, + 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, @@ -558,111 +607,134 @@ static struct btree *mca_bucket_alloc(struct cache_set *c, 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_write_work); + INIT_DELAYED_WORK(&b->work, btree_node_write_work); b->c = c; - closure_init_unlocked(&b->io); + sema_init(&b->io_mutex, 1); mca_data_alloc(b, k, gfp); return b; } -static int mca_reap(struct btree *b, struct closure *cl, unsigned min_order) +static int mca_reap(struct btree *b, unsigned 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; - if (b->page_order < min_order) { - rw_unlock(true, b); - return -ENOMEM; - } - - BUG_ON(btree_node_dirty(b) && !b->sets[0].data); + BUG_ON(btree_node_dirty(b) && !b->keys.set[0].data); - if (cl && btree_node_dirty(b)) - bch_btree_write(b, true, NULL); + if (b->keys.page_order < min_order) + goto out_unlock; - if (cl) - closure_wait_event_async(&b->io.wait, cl, - atomic_read(&b->io.cl.remaining) == -1); + if (!flush) { + if (btree_node_dirty(b)) + goto out_unlock; - if (btree_node_dirty(b) || - !closure_is_unlocked(&b->io.cl) || - work_pending(&b->work.work)) { - rw_unlock(true, b); - return -EAGAIN; + if (down_trylock(&b->io_mutex)) + goto out_unlock; + up(&b->io_mutex); } + mutex_lock(&b->write_lock); + 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 int bch_mca_shrink(struct shrinker *shrink, struct shrink_control *sc) +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; if (c->shrinker_disabled) - return 0; + return SHRINK_STOP; - if (c->try_harder) - return 0; - - /* - * If nr == 0, we're supposed to return the number of items we have - * cached. Not allowed to return -1. - */ - if (!nr) - return mca_can_free(c) * c->btree_pages; + if (c->btree_cache_alloc_lock) + return SHRINK_STOP; /* Return -1 if we can't do anything right now */ - if (sc->gfp_mask & __GFP_WAIT) + 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; nr = min_t(unsigned long, nr, mca_can_free(c)); i = 0; list_for_each_entry_safe(b, t, &c->btree_cache_freeable, list) { - if (!nr) + if (freed >= nr) break; if (++i > 3 && - !mca_reap(b, NULL, 0)) { + !mca_reap(b, 0, false)) { mca_data_free(b); rw_unlock(true, b); - --nr; + freed++; } } - /* - * Can happen right when we first start up, before we've read in any - * btree nodes - */ - if (list_empty(&c->btree_cache)) - goto out; + for (i = 0; (nr--) && i < c->btree_cache_used; i++) { + if (list_empty(&c->btree_cache)) + goto out; - for (i = 0; nr && i < c->bucket_cache_used; i++) { b = list_first_entry(&c->btree_cache, struct btree, list); list_rotate_left(&c->btree_cache); if (!b->accessed && - !mca_reap(b, NULL, 0)) { + !mca_reap(b, 0, false)) { mca_bucket_free(b); mca_data_free(b); rw_unlock(true, b); - --nr; + freed++; } else b->accessed = 0; } out: - nr = mca_can_free(c) * c->btree_pages; mutex_unlock(&c->bucket_lock); - return nr; + return freed; +} + +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) @@ -679,6 +751,8 @@ void bch_btree_cache_free(struct cache_set *c) #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(bucket_pages(c))); #endif list_splice(&c->btree_cache_freeable, @@ -709,12 +783,9 @@ int bch_btree_cache_alloc(struct cache_set *c) { unsigned i; - /* XXX: doesn't check for errors */ - - closure_init_unlocked(&c->gc); - for (i = 0; i < mca_reserve(c); i++) - mca_bucket_alloc(c, &ZERO_KEY, GFP_KERNEL); + if (!mca_bucket_alloc(c, &ZERO_KEY, GFP_KERNEL)) + return -ENOMEM; list_splice_init(&c->btree_cache, &c->btree_cache_freeable); @@ -722,16 +793,20 @@ int bch_btree_cache_alloc(struct cache_set *c) #ifdef CONFIG_BCACHE_DEBUG mutex_init(&c->verify_lock); + c->verify_ondisk = (void *) + __get_free_pages(GFP_KERNEL, ilog2(bucket_pages(c))); + c->verify_data = mca_bucket_alloc(c, &ZERO_KEY, GFP_KERNEL); if (c->verify_data && - c->verify_data->sets[0].data) + c->verify_data->keys.set->data) list_del_init(&c->verify_data->list); else c->verify_data = NULL; #endif - c->shrink.shrink = bch_mca_shrink; + 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; register_shrinker(&c->shrink); @@ -760,51 +835,41 @@ out: return b; } -static struct btree *mca_cannibalize(struct cache_set *c, struct bkey *k, - int level, struct closure *cl) +static int mca_cannibalize_lock(struct cache_set *c, struct btree_op *op) { - int ret = -ENOMEM; - struct btree *i; + struct task_struct *old; - if (!cl) - return ERR_PTR(-ENOMEM); + old = cmpxchg(&c->btree_cache_alloc_lock, NULL, current); + if (old && old != current) { + if (op) + prepare_to_wait(&c->btree_cache_wait, &op->wait, + TASK_UNINTERRUPTIBLE); + return -EINTR; + } - /* - * Trying to free up some memory - i.e. reuse some btree nodes - may - * require initiating IO to flush the dirty part of the node. If we're - * running under generic_make_request(), that IO will never finish and - * we would deadlock. Returning -EAGAIN causes the cache lookup code to - * punt to workqueue and retry. - */ - if (current->bio_list) - return ERR_PTR(-EAGAIN); + return 0; +} - if (c->try_harder && c->try_harder != cl) { - closure_wait_event_async(&c->try_wait, cl, !c->try_harder); - return ERR_PTR(-EAGAIN); - } +static struct btree *mca_cannibalize(struct cache_set *c, struct btree_op *op, + struct bkey *k) +{ + struct btree *b; - /* XXX: tracepoint */ - c->try_harder = cl; - c->try_harder_start = local_clock(); -retry: - list_for_each_entry_reverse(i, &c->btree_cache, list) { - int r = mca_reap(i, cl, btree_order(k)); - if (!r) - return i; - if (r != -ENOMEM) - ret = r; - } + trace_bcache_btree_cache_cannibalize(c); - if (ret == -EAGAIN && - closure_blocking(cl)) { - mutex_unlock(&c->bucket_lock); - closure_sync(cl); - mutex_lock(&c->bucket_lock); - goto retry; - } + if (mca_cannibalize_lock(c, op)) + return ERR_PTR(-EINTR); - return ERR_PTR(ret); + 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); } /* @@ -813,20 +878,21 @@ retry: * 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, struct closure *cl) +static void bch_cannibalize_unlock(struct cache_set *c) { - if (c->try_harder == cl) { - bch_time_stats_update(&c->try_harder_time, c->try_harder_start); - c->try_harder = NULL; - __closure_wake_up(&c->try_wait); + if (c->btree_cache_alloc_lock == current) { + c->btree_cache_alloc_lock = NULL; + wake_up(&c->btree_cache_wait); } } -static struct btree *mca_alloc(struct cache_set *c, struct bkey *k, - int level, struct closure *cl) +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)) @@ -836,16 +902,16 @@ static struct btree *mca_alloc(struct cache_set *c, struct bkey *k, * the list. Check if there's any freed nodes there: */ list_for_each_entry(b, &c->btree_cache_freeable, list) - if (!mca_reap(b, NULL, btree_order(k))) + 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, NULL, 0)) { + if (!mca_reap(b, 0, false)) { mca_data_alloc(b, k, __GFP_NOWARN|GFP_NOIO); - if (!b->sets[0].data) + if (!b->keys.set[0].data) goto err; else goto out; @@ -856,10 +922,10 @@ static struct btree *mca_alloc(struct cache_set *c, struct bkey *k, goto err; BUG_ON(!down_write_trylock(&b->lock)); - if (!b->sets->data) + if (!b->keys.set->data) goto err; out: - BUG_ON(!closure_is_unlocked(&b->io.cl)); + BUG_ON(b->io_mutex.count != 1); bkey_copy(&b->key, k); list_move(&b->list, &c->btree_cache); @@ -867,16 +933,24 @@ out: 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; - mca_reinit(b); + 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, k, level, cl); + b = mca_cannibalize(c, op, k); if (!IS_ERR(b)) goto out; @@ -887,17 +961,15 @@ err: * 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, it uses the closure embedded in struct btree_op to wait; - * if that closure is in non blocking mode, will return -EAGAIN. + * If IO is necessary and running under generic_make_request, returns -EAGAIN. * * The btree node will have either a read or a write lock held, depending on * level and op->lock. */ -struct btree *bch_btree_node_get(struct cache_set *c, struct bkey *k, - int level, struct btree_op *op) +struct btree *bch_btree_node_get(struct cache_set *c, struct btree_op *op, + struct bkey *k, int level, bool write) { int i = 0; - bool write = level <= op->lock; struct btree *b; BUG_ON(level < 0); @@ -905,8 +977,11 @@ 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, k, level, &op->cl); + b = mca_alloc(c, op, k, level); mutex_unlock(&c->bucket_lock); if (!b) @@ -914,7 +989,7 @@ retry: if (IS_ERR(b)) return b; - bch_btree_read(b); + bch_btree_node_read(b); if (!write) downgrade_write(&b->lock); @@ -929,23 +1004,20 @@ retry: b->accessed = 1; - for (; i <= b->nsets && b->sets[i].size; i++) { - prefetch(b->sets[i].tree); - prefetch(b->sets[i].data); + 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->nsets; i++) - prefetch(b->sets[i].data); + for (; i <= b->keys.nsets; i++) + prefetch(b->keys.set[i].data); - if (!closure_wait_event(&b->io.wait, &op->cl, - btree_node_read_done(b))) { + if (btree_node_io_error(b)) { rw_unlock(write, b); - b = ERR_PTR(-EAGAIN); - } else if (btree_node_io_error(b)) { - rw_unlock(write, b); - b = ERR_PTR(-EIO); - } else - BUG_ON(!b->written); + return ERR_PTR(-EIO); + } + + BUG_ON(!b->written); return b; } @@ -955,106 +1027,138 @@ static void btree_node_prefetch(struct cache_set *c, struct bkey *k, int level) struct btree *b; mutex_lock(&c->bucket_lock); - b = mca_alloc(c, k, level, NULL); + b = mca_alloc(c, NULL, k, level); mutex_unlock(&c->bucket_lock); if (!IS_ERR_OR_NULL(b)) { - bch_btree_read(b); + bch_btree_node_read(b); rw_unlock(true, b); } } /* Btree alloc */ -static void btree_node_free(struct btree *b, struct btree_op *op) +static void btree_node_free(struct btree *b) { - unsigned i; + trace_bcache_btree_node_free(b); - /* - * The BUG_ON() in btree_node_get() implies that we must have a write - * lock on parent to free or even invalidate a node - */ - BUG_ON(op->lock <= b->level); BUG_ON(b == b->c->root); - pr_debug("bucket %s", pbtree(b)); + + mutex_lock(&b->write_lock); if (btree_node_dirty(b)) btree_complete_write(b, btree_current_write(b)); clear_bit(BTREE_NODE_dirty, &b->flags); - if (b->prio_blocked && - !atomic_sub_return(b->prio_blocked, &b->c->prio_blocked)) - wake_up(&b->c->alloc_wait); - - b->prio_blocked = 0; + mutex_unlock(&b->write_lock); cancel_delayed_work(&b->work); mutex_lock(&b->c->bucket_lock); - - for (i = 0; i < KEY_PTRS(&b->key); i++) { - BUG_ON(atomic_read(&PTR_BUCKET(b->c, &b->key, i)->pin)); - - bch_inc_gen(PTR_CACHE(b->c, &b->key, i), - PTR_BUCKET(b->c, &b->key, i)); - } - bch_bucket_free(b->c, &b->key); mca_bucket_free(b); mutex_unlock(&b->c->bucket_lock); } -struct btree *bch_btree_node_alloc(struct cache_set *c, int level, - struct closure *cl) +struct btree *bch_btree_node_alloc(struct cache_set *c, struct btree_op *op, + int level) { BKEY_PADDED(key) k; struct btree *b = ERR_PTR(-EAGAIN); mutex_lock(&c->bucket_lock); retry: - if (__bch_bucket_alloc_set(c, WATERMARK_METADATA, &k.key, 1, cl)) + if (__bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, 1, op != NULL)) goto err; + bkey_put(c, &k.key); SET_KEY_SIZE(&k.key, c->btree_pages * PAGE_SECTORS); - b = mca_alloc(c, &k.key, level, cl); + 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"); - __bkey_put(c, &k.key); goto retry; } - set_btree_node_read_done(b); b->accessed = 1; - bch_bset_init_next(b); + bch_bset_init_next(&b->keys, b->keys.set->data, bset_magic(&b->c->sb)); mutex_unlock(&c->bucket_lock); + + trace_bcache_btree_node_alloc(b); return b; err_free: bch_bucket_free(c, &k.key); - __bkey_put(c, &k.key); err: mutex_unlock(&c->bucket_lock); + + trace_bcache_btree_node_alloc_fail(b); return b; } static struct btree *btree_node_alloc_replacement(struct btree *b, - struct closure *cl) + struct btree_op *op) { - struct btree *n = bch_btree_node_alloc(b->c, b->level, cl); - if (!IS_ERR_OR_NULL(n)) - bch_btree_sort_into(b, n); + struct btree *n = bch_btree_node_alloc(b->c, op, b->level); + if (!IS_ERR_OR_NULL(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 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; + unsigned i, reserve = (c->root->level - b->level) * 2 + 1; + + mutex_lock(&c->bucket_lock); + + for_each_cache(ca, c, i) + 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 */ -uint8_t __bch_btree_mark_key(struct cache_set *c, int level, struct bkey *k) +static uint8_t __bch_btree_mark_key(struct cache_set *c, int level, + struct bkey *k) { uint8_t stale = 0; unsigned i; @@ -1074,8 +1178,8 @@ uint8_t __bch_btree_mark_key(struct cache_set *c, int level, struct bkey *k) g = PTR_BUCKET(c, k, i); - if (gen_after(g->gc_gen, PTR_GEN(k, i))) - g->gc_gen = PTR_GEN(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)); @@ -1091,11 +1195,13 @@ uint8_t __bch_btree_mark_key(struct cache_set *c, int level, struct bkey *k) 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, GC_SECTORS_USED(g) + KEY_SIZE(k), - (1 << 14) - 1)); + MAX_GC_SECTORS_USED)); BUG_ON(!GC_SECTORS_USED(g)); } @@ -1105,126 +1211,143 @@ uint8_t __bch_btree_mark_key(struct cache_set *c, int level, struct bkey *k) #define btree_mark_key(b, k) __bch_btree_mark_key(b->c, b->level, k) -static int btree_gc_mark_node(struct btree *b, unsigned *keys, - struct gc_stat *gc) +void bch_initial_mark_key(struct cache_set *c, int level, struct bkey *k) +{ + unsigned 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); +} + +static bool btree_gc_mark_node(struct btree *b, struct gc_stat *gc) { uint8_t stale = 0; - unsigned last_dev = -1; - struct bcache_device *d = NULL; + unsigned keys = 0, good_keys = 0; struct bkey *k; struct btree_iter iter; struct bset_tree *t; gc->nodes++; - for_each_key_filter(b, k, &iter, bch_ptr_invalid) { - if (last_dev != KEY_INODE(k)) { - last_dev = KEY_INODE(k); - - d = KEY_INODE(k) < b->c->nr_uuids - ? b->c->devices[last_dev] - : NULL; - } - + 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, k)) + if (bch_ptr_bad(&b->keys, k)) continue; - *keys += bkey_u64s(k); - gc->key_bytes += bkey_u64s(k); gc->nkeys++; + good_keys++; gc->data += KEY_SIZE(k); - if (KEY_DIRTY(k)) { - gc->dirty += KEY_SIZE(k); - if (d) - d->sectors_dirty_gc += KEY_SIZE(k); - } } - for (t = b->sets; t <= &b->sets[b->nsets]; t++) + for (t = b->keys.set; t <= &b->keys.set[b->keys.nsets]; t++) btree_bug_on(t->size && - bset_written(b, t) && + bset_written(&b->keys, t) && bkey_cmp(&b->key, &t->end) < 0, b, "found short btree key in gc"); - return stale; -} - -static struct btree *btree_gc_alloc(struct btree *b, struct bkey *k, - struct btree_op *op) -{ - /* - * We block priorities from being written for the duration of garbage - * collection, so we can't sleep in btree_alloc() -> - * bch_bucket_alloc_set(), or we'd risk deadlock - so we don't pass it - * our closure. - */ - struct btree *n = btree_node_alloc_replacement(b, NULL); + if (b->c->gc_always_rewrite) + return true; - if (!IS_ERR_OR_NULL(n)) { - swap(b, n); - - memcpy(k->ptr, b->key.ptr, - sizeof(uint64_t) * KEY_PTRS(&b->key)); + if (stale > 10) + return true; - __bkey_put(b->c, &b->key); - atomic_inc(&b->c->prio_blocked); - b->prio_blocked++; + if ((keys - good_keys) * 2 > keys) + return true; - btree_node_free(n, op); - up_write(&n->lock); - } - - return b; + return false; } -/* - * Leaving this at 2 until we've got incremental garbage collection done; it - * could be higher (and has been tested with 4) except that garbage collection - * could take much longer, adversely affecting latency. - */ -#define GC_MERGE_NODES 2U +#define GC_MERGE_NODES 4U struct gc_merge_info { struct btree *b; - struct bkey *k; unsigned keys; }; -static void btree_gc_coalesce(struct btree *b, struct btree_op *op, - struct gc_stat *gc, struct gc_merge_info *r) +static int bch_btree_insert_node(struct btree *, struct btree_op *, + struct keylist *, atomic_t *, struct bkey *); + +static int btree_gc_coalesce(struct btree *b, struct btree_op *op, + struct gc_stat *gc, struct gc_merge_info *r) { - unsigned nodes = 0, keys = 0, blocks; - int i; + unsigned 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; - while (nodes < GC_MERGE_NODES && r[nodes].b) + 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->sets[0].data, keys, b->c) > blocks * (nodes - 1)) - return; - - for (i = nodes - 1; i >= 0; --i) { - if (r[i].b->written) - r[i].b = btree_gc_alloc(r[i].b, r[i].k, op); + __set_blocks(b->keys.set[0].data, keys, + block_bytes(b->c)) > blocks * (nodes - 1)) + return 0; - if (r[i].b->written) - return; + for (i = 0; i < nodes; i++) { + new_nodes[i] = btree_node_alloc_replacement(r[i].b, NULL); + if (IS_ERR_OR_NULL(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 = r[i].b->sets->data; - struct bset *n2 = r[i - 1].b->sets->data; + 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) { + 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)) > 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 @@ -1233,138 +1356,226 @@ static void btree_gc_coalesce(struct btree *b, struct btree_op *op, * length keys (shouldn't be possible in practice, * though) */ - if (__set_blocks(n1, n1->keys + r->keys, - b->c) > btree_blocks(r[i].b)) - return; + if (__set_blocks(n1, n1->keys + n2->keys, + block_bytes(b->c)) > + btree_blocks(new_nodes[i])) + goto out_nocoalesce; keys = n2->keys; + /* Take the key of the node we're getting rid of */ last = &r->b->key; - } else - for (k = n2->start; - k < end(n2); - k = bkey_next(k)) { - if (__set_blocks(n1, n1->keys + keys + - bkey_u64s(k), b->c) > blocks) - break; - - last = k; - keys += bkey_u64s(k); - } + } - BUG_ON(__set_blocks(n1, n1->keys + keys, - b->c) > btree_blocks(r[i].b)); + BUG_ON(__set_blocks(n1, n1->keys + keys, block_bytes(b->c)) > + btree_blocks(new_nodes[i])); - if (last) { - bkey_copy_key(&r[i].b->key, last); - bkey_copy_key(r[i].k, last); - } + if (last) + bkey_copy_key(&new_nodes[i]->key, last); - memcpy(end(n1), + memcpy(bset_bkey_last(n1), n2->start, - (void *) node(n2, keys) - (void *) n2->start); + (void *) bset_bkey_idx(n2, keys) - (void *) n2->start); n1->keys += keys; + r[i].keys = n1->keys; memmove(n2->start, - node(n2, keys), - (void *) end(n2) - (void *) node(n2, keys)); + bset_bkey_idx(n2, keys), + (void *) bset_bkey_last(n2) - + (void *) bset_bkey_idx(n2, keys)); n2->keys -= keys; - r[i].keys = n1->keys; - r[i - 1].keys = n2->keys; + if (__bch_keylist_realloc(&keylist, + bkey_u64s(&new_nodes[i]->key))) + goto out_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]); + + 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); } - btree_node_free(r->b, op); - up_write(&r->b->lock); + bch_btree_insert_node(b, op, &keylist, NULL, NULL); + BUG_ON(!bch_keylist_empty(&keylist)); - pr_debug("coalesced %u nodes", nodes); + 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--; - nodes--; - memmove(&r[0], &r[1], sizeof(struct gc_merge_info) * nodes); - memset(&r[nodes], 0, sizeof(struct gc_merge_info)); + bch_keylist_free(&keylist); + + /* Invalidated our iterator */ + return -EINTR; + +out_nocoalesce: + closure_sync(&cl); + bch_keylist_free(&keylist); + + while ((k = bch_keylist_pop(&keylist))) + if (!bkey_cmp(k, &ZERO_KEY)) + atomic_dec(&b->c->prio_blocked); + + 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_recurse(struct btree *b, struct btree_op *op, - struct closure *writes, struct gc_stat *gc) +static int btree_gc_rewrite_node(struct btree *b, struct btree_op *op, + struct btree *replace) { - void write(struct btree *r) - { - if (!r->written) - bch_btree_write(r, true, op); - else if (btree_node_dirty(r)) { - BUG_ON(btree_current_write(r)->owner); - btree_current_write(r)->owner = writes; - closure_get(writes); + struct keylist keys; + struct btree *n; - bch_btree_write(r, true, NULL); - } + if (btree_check_reserve(b, NULL)) + return 0; + + n = btree_node_alloc_replacement(replace, NULL); - up_write(&r->lock); + /* recheck reserve after allocating replacement node */ + if (btree_check_reserve(b, NULL)) { + btree_node_free(n); + rw_unlock(true, n); + return 0; } - int ret = 0, stale; - unsigned i; - struct gc_merge_info r[GC_MERGE_NODES]; + bch_btree_node_write_sync(n); - memset(r, 0, sizeof(r)); + bch_keylist_init(&keys); + bch_keylist_add(&keys, &n->key); - while ((r->k = bch_next_recurse_key(b, &b->c->gc_done))) { - r->b = bch_btree_node_get(b->c, r->k, b->level - 1, op); + make_btree_freeing_key(replace, keys.top); + bch_keylist_push(&keys); - if (IS_ERR(r->b)) { - ret = PTR_ERR(r->b); - break; - } + 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; +} - r->keys = 0; - stale = btree_gc_mark_node(r->b, &r->keys, gc); +static unsigned btree_gc_count_keys(struct btree *b) +{ + struct bkey *k; + struct btree_iter iter; + unsigned ret = 0; - if (!b->written && - (r->b->level || stale > 10 || - b->c->gc_always_rewrite)) - r->b = btree_gc_alloc(r->b, r->k, op); + for_each_key_filter(&b->keys, k, &iter, bch_ptr_bad) + ret += bkey_u64s(k); - if (r->b->level) - ret = btree_gc_recurse(r->b, op, writes, gc); + return ret; +} - if (ret) { - write(r->b); - break; +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); + 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; } - bkey_copy_key(&b->c->gc_done, r->k); + if (!last->b) + break; - if (!b->written) - btree_gc_coalesce(b, op, gc, r); + 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 (r[GC_MERGE_NODES - 1].b) - write(r[GC_MERGE_NODES - 1].b); + if (last->b->level) { + ret = btree_gc_recurse(last->b, op, writes, gc); + if (ret) + break; + } - memmove(&r[1], &r[0], - sizeof(struct gc_merge_info) * (GC_MERGE_NODES - 1)); + 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; - /* When we've got incremental GC working, we'll want to do - * if (should_resched()) - * return -EAGAIN; - */ - cond_resched(); -#if 0 if (need_resched()) { ret = -EAGAIN; break; } -#endif } - for (i = 1; i < GC_MERGE_NODES && r[i].b; i++) - write(r[i].b); - - /* Might have freed some children, must remove their keys */ - if (!b->written) - bch_btree_sort(b); + 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; } @@ -1373,31 +1584,34 @@ static int bch_btree_gc_root(struct btree *b, struct btree_op *op, struct closure *writes, struct gc_stat *gc) { struct btree *n = NULL; - unsigned keys = 0; - int ret = 0, stale = btree_gc_mark_node(b, &keys, gc); + int ret = 0; + bool should_rewrite; - if (b->level || stale > 10) + should_rewrite = btree_gc_mark_node(b, gc); + if (should_rewrite) { n = btree_node_alloc_replacement(b, NULL); - if (!IS_ERR_OR_NULL(n)) - swap(b, n); + if (!IS_ERR_OR_NULL(n)) { + bch_btree_node_write_sync(n); - if (b->level) - ret = btree_gc_recurse(b, op, writes, gc); + bch_btree_set_root(n); + btree_node_free(b); + rw_unlock(true, n); - if (!b->written || btree_node_dirty(b)) { - atomic_inc(&b->c->prio_blocked); - b->prio_blocked++; - bch_btree_write(b, true, n ? op : NULL); + return -EINTR; + } } - if (!IS_ERR_OR_NULL(n)) { - closure_sync(&op->cl); - bch_btree_set_root(b); - btree_node_free(n, op); - rw_unlock(true, b); + __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; } @@ -1405,7 +1619,6 @@ static void btree_gc_start(struct cache_set *c) { struct cache *ca; struct bucket *b; - struct bcache_device **d; unsigned i; if (!c->gc_mark_valid) @@ -1418,26 +1631,21 @@ static void btree_gc_start(struct cache_set *c) for_each_cache(ca, c, i) for_each_bucket(b, ca) { - b->gc_gen = b->gen; - if (!atomic_read(&b->pin)) - SET_GC_MARK(b, GC_MARK_RECLAIMABLE); + b->last_gc = b->gen; + if (!atomic_read(&b->pin)) { + SET_GC_MARK(b, 0); + SET_GC_SECTORS_USED(b, 0); + } } - for (d = c->devices; - d < c->devices + c->nr_uuids; - d++) - if (*d) - (*d)->sectors_dirty_gc = 0; - mutex_unlock(&c->bucket_lock); } -size_t bch_btree_gc_finish(struct cache_set *c) +static size_t bch_btree_gc_finish(struct cache_set *c) { size_t available = 0; struct bucket *b; struct cache *ca; - struct bcache_device **d; unsigned i; mutex_lock(&c->bucket_lock); @@ -1446,15 +1654,32 @@ size_t bch_btree_gc_finish(struct cache_set *c) c->gc_mark_valid = 1; c->need_gc = 0; - if (c->root) - for (i = 0; i < KEY_PTRS(&c->root->key); i++) - SET_GC_MARK(PTR_BUCKET(c, &c->root->key, i), - GC_MARK_METADATA); - 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->nr_uuids; i++) { + struct bcache_device *d = c->devices[i]; + struct cached_dev *dc; + struct keybuf_key *w, *n; + unsigned j; + + 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(); + for_each_cache(ca, c, i) { uint64_t *i; @@ -1468,678 +1693,545 @@ size_t bch_btree_gc_finish(struct cache_set *c) SET_GC_MARK(ca->buckets + *i, GC_MARK_METADATA); for_each_bucket(b, ca) { - b->last_gc = b->gc_gen; c->need_gc = max(c->need_gc, bucket_gc_gen(b)); - if (!atomic_read(&b->pin) && - GC_MARK(b) == GC_MARK_RECLAIMABLE) { - available++; - if (!GC_SECTORS_USED(b)) - bch_bucket_add_unused(ca, b); - } - } - } - - for (d = c->devices; - d < c->devices + c->nr_uuids; - d++) - if (*d) { - unsigned long last = - atomic_long_read(&((*d)->sectors_dirty)); - long difference = (*d)->sectors_dirty_gc - last; - - pr_debug("sectors dirty off by %li", difference); + if (atomic_read(&b->pin)) + continue; - (*d)->sectors_dirty_last += difference; + BUG_ON(!GC_MARK(b) && GC_SECTORS_USED(b)); - atomic_long_set(&((*d)->sectors_dirty), - (*d)->sectors_dirty_gc); + if (!GC_MARK(b) || GC_MARK(b) == GC_MARK_RECLAIMABLE) + available++; } + } mutex_unlock(&c->bucket_lock); return available; } -static void bch_btree_gc(struct closure *cl) +static void bch_btree_gc(struct cache_set *c) { - struct cache_set *c = container_of(cl, struct cache_set, gc.cl); int ret; unsigned long available; struct gc_stat stats; struct closure writes; struct btree_op op; - uint64_t start_time = local_clock(); - trace_bcache_gc_start(c->sb.set_uuid); - blktrace_msg_all(c, "Starting gc"); + + trace_bcache_gc_start(c); memset(&stats, 0, sizeof(struct gc_stat)); closure_init_stack(&writes); - bch_btree_op_init_stack(&op); - op.lock = SHRT_MAX; + bch_btree_op_init(&op, SHRT_MAX); btree_gc_start(c); - ret = btree_root(gc_root, c, &op, &writes, &stats); - closure_sync(&op.cl); - closure_sync(&writes); - - if (ret) { - blktrace_msg_all(c, "Stopped gc"); - pr_warn("gc failed!"); - - continue_at(cl, bch_btree_gc, bch_gc_wq); - } + do { + ret = btree_root(gc_root, c, &op, &writes, &stats); + closure_sync(&writes); - /* Possibly wait for new UUIDs or whatever to hit disk */ - bch_journal_meta(c, &op.cl); - closure_sync(&op.cl); + if (ret && ret != -EAGAIN) + pr_warn("gc failed!"); + } while (ret); available = 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.dirty <<= 9; stats.data <<= 9; stats.in_use = (c->nbuckets - available) * 100 / c->nbuckets; memcpy(&c->gc_stats, &stats, sizeof(struct gc_stat)); - blktrace_msg_all(c, "Finished gc"); - trace_bcache_gc_end(c->sb.set_uuid); - wake_up(&c->alloc_wait); + trace_bcache_gc_end(c); - continue_at(cl, bch_moving_gc, bch_gc_wq); + bch_moving_gc(c); } -void bch_queue_gc(struct cache_set *c) +static int bch_gc_thread(void *arg) { - closure_trylock_call(&c->gc.cl, bch_btree_gc, bch_gc_wq, &c->cl); -} - -/* Initial partial gc */ - -static int bch_btree_check_recurse(struct btree *b, struct btree_op *op, - unsigned long **seen) -{ - int ret; + struct cache_set *c = arg; + struct cache *ca; unsigned i; - struct bkey *k; - struct bucket *g; - struct btree_iter iter; - for_each_key_filter(b, k, &iter, bch_ptr_invalid) { - for (i = 0; i < KEY_PTRS(k); i++) { - if (!ptr_available(b->c, k, i)) - continue; + while (1) { +again: + bch_btree_gc(c); - g = PTR_BUCKET(b->c, k, i); + set_current_state(TASK_INTERRUPTIBLE); + if (kthread_should_stop()) + break; - if (!__test_and_set_bit(PTR_BUCKET_NR(b->c, k, i), - seen[PTR_DEV(k, i)]) || - !ptr_stale(b->c, k, i)) { - g->gen = PTR_GEN(k, i); + mutex_lock(&c->bucket_lock); - if (b->level) - g->prio = BTREE_PRIO; - else if (g->prio == BTREE_PRIO) - g->prio = INITIAL_PRIO; + for_each_cache(ca, c, i) + if (ca->invalidate_needs_gc) { + mutex_unlock(&c->bucket_lock); + set_current_state(TASK_RUNNING); + goto again; } - } - btree_mark_key(b, k); - } - - if (b->level) { - k = bch_next_recurse_key(b, &ZERO_KEY); - - while (k) { - struct bkey *p = bch_next_recurse_key(b, k); - if (p) - btree_node_prefetch(b->c, p, b->level - 1); - - ret = btree(check_recurse, k, b, op, seen); - if (ret) - return ret; + mutex_unlock(&c->bucket_lock); - k = p; - } + try_to_freeze(); + schedule(); } return 0; } -int bch_btree_check(struct cache_set *c, struct btree_op *op) +int bch_gc_thread_start(struct cache_set *c) { - int ret = -ENOMEM; - unsigned i; - unsigned long *seen[MAX_CACHES_PER_SET]; - - memset(seen, 0, sizeof(seen)); - - for (i = 0; c->cache[i]; i++) { - size_t n = DIV_ROUND_UP(c->cache[i]->sb.nbuckets, 8); - seen[i] = kmalloc(n, GFP_KERNEL); - if (!seen[i]) - goto err; + c->gc_thread = kthread_create(bch_gc_thread, c, "bcache_gc"); + if (IS_ERR(c->gc_thread)) + return PTR_ERR(c->gc_thread); - /* Disables the seen array until prio_read() uses it too */ - memset(seen[i], 0xFF, n); - } - - ret = btree_root(check_recurse, c, op, seen); -err: - for (i = 0; i < MAX_CACHES_PER_SET; i++) - kfree(seen[i]); - return ret; + set_task_state(c->gc_thread, TASK_INTERRUPTIBLE); + return 0; } -/* Btree insertion */ - -static void shift_keys(struct btree *b, struct bkey *where, struct bkey *insert) -{ - struct bset *i = b->sets[b->nsets].data; - - memmove((uint64_t *) where + bkey_u64s(insert), - where, - (void *) end(i) - (void *) where); - - i->keys += bkey_u64s(insert); - bkey_copy(where, insert); - bch_bset_fix_lookup_table(b, where); -} +/* Initial partial gc */ -static bool fix_overlapping_extents(struct btree *b, - struct bkey *insert, - struct btree_iter *iter, - struct btree_op *op) +static int bch_btree_check_recurse(struct btree *b, struct btree_op *op) { - void subtract_dirty(struct bkey *k, int sectors) - { - struct bcache_device *d = b->c->devices[KEY_INODE(k)]; - - if (KEY_DIRTY(k) && d) - atomic_long_sub(sectors, &d->sectors_dirty); - } - - unsigned old_size, sectors_found = 0; - - while (1) { - struct bkey *k = bch_btree_iter_next(iter); - if (!k || - bkey_cmp(&START_KEY(k), insert) >= 0) - break; + int ret = 0; + struct bkey *k, *p = NULL; + struct btree_iter iter; - if (bkey_cmp(k, &START_KEY(insert)) <= 0) - continue; + for_each_key_filter(&b->keys, k, &iter, bch_ptr_invalid) + bch_initial_mark_key(b->c, b->level, k); - old_size = KEY_SIZE(k); + bch_initial_mark_key(b->c, b->level + 1, &b->key); - /* - * 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 BTREE_REPLACE - * operations. - */ + if (b->level) { + bch_btree_iter_init(&b->keys, &iter, NULL); - if (op->type == BTREE_REPLACE && - KEY_SIZE(k)) { - /* - * k might have been split since we inserted/found the - * key we're replacing - */ - unsigned i; - uint64_t offset = KEY_START(k) - - KEY_START(&op->replace); + do { + k = bch_btree_iter_next_filter(&iter, &b->keys, + bch_ptr_bad); + if (k) + btree_node_prefetch(b->c, k, b->level - 1); - /* But it must be a subset of the replace key */ - if (KEY_START(k) < KEY_START(&op->replace) || - KEY_OFFSET(k) > KEY_OFFSET(&op->replace)) - goto check_failed; + if (p) + ret = btree(check_recurse, p, b, op); - /* We didn't find a key that we were supposed to */ - if (KEY_START(k) > KEY_START(insert) + sectors_found) - goto check_failed; + p = k; + } while (p && !ret); + } - if (KEY_PTRS(&op->replace) != KEY_PTRS(k)) - goto check_failed; + return ret; +} - /* skip past gen */ - offset <<= 8; +int bch_btree_check(struct cache_set *c) +{ + struct btree_op op; - BUG_ON(!KEY_PTRS(&op->replace)); + bch_btree_op_init(&op, SHRT_MAX); - for (i = 0; i < KEY_PTRS(&op->replace); i++) - if (k->ptr[i] != op->replace.ptr[i] + offset) - goto check_failed; + return btree_root(check_recurse, c, &op); +} - sectors_found = KEY_OFFSET(k) - KEY_START(insert); - } +void bch_initial_gc_finish(struct cache_set *c) +{ + struct cache *ca; + struct bucket *b; + unsigned i; - 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. - */ + bch_btree_gc_finish(c); - struct bkey *top; - - subtract_dirty(k, 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, &b->sets[b->nsets], - insert); - shift_keys(b, top, k); - } else { - BKEY_PADDED(key) temp; - bkey_copy(&temp.key, k); - shift_keys(b, k, &temp.key); - top = bkey_next(k); - } + mutex_lock(&c->bucket_lock); - bch_cut_front(insert, top); - bch_cut_back(&START_KEY(insert), k); - bch_bset_fix_invalidated_key(b, k); - return false; - } + /* + * 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_cache(ca, c, i) { + for_each_bucket(b, ca) { + if (fifo_full(&ca->free[RESERVE_PRIO])) + break; - if (bkey_cmp(insert, k) < 0) { - bch_cut_front(insert, k); - } else { - 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); + if (bch_can_invalidate_bucket(ca, b) && + !GC_MARK(b)) { + __bch_invalidate_one_bucket(ca, b); + fifo_push(&ca->free[RESERVE_PRIO], + b - ca->buckets); } } - - subtract_dirty(k, old_size - KEY_SIZE(k)); - } - -check_failed: - if (op->type == BTREE_REPLACE) { - if (!sectors_found) { - op->insert_collision = true; - 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); - } } - return false; + mutex_unlock(&c->bucket_lock); } -static bool btree_insert_key(struct btree *b, struct btree_op *op, - struct bkey *k) +/* Btree insertion */ + +static bool btree_insert_key(struct btree *b, struct bkey *k, + struct bkey *replace_key) { - struct bset *i = b->sets[b->nsets].data; - struct bkey *m, *prev; - const char *status = "insert"; + unsigned status; BUG_ON(bkey_cmp(k, &b->key) > 0); - BUG_ON(b->level && !KEY_PTRS(k)); - BUG_ON(!b->level && !KEY_OFFSET(k)); - if (!b->level) { - struct btree_iter iter; - struct bkey search = KEY(KEY_INODE(k), KEY_START(k), 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"); - /* - * bset_search() returns the first key that is strictly greater - * than the search key - but for back merging, we want to find - * the first key that is greater than or equal to KEY_START(k) - - * unless KEY_START(k) is 0. - */ - if (KEY_OFFSET(&search)) - SET_KEY_OFFSET(&search, KEY_OFFSET(&search) - 1); - - prev = NULL; - m = bch_btree_iter_init(b, &iter, &search); - - if (fix_overlapping_extents(b, k, &iter, op)) - return false; - - while (m != end(i) && - bkey_cmp(k, &START_KEY(m)) > 0) - prev = m, m = bkey_next(m); - - if (key_merging_disabled(b->c)) - goto insert; - - /* prev is in the tree, if we merge we're done */ - status = "back merging"; - if (prev && - bch_bkey_try_merge(b, prev, k)) - goto merged; - - status = "overwrote front"; - if (m != end(i) && - KEY_PTRS(m) == KEY_PTRS(k) && !KEY_SIZE(m)) - goto copy; - - status = "front merge"; - if (m != end(i) && - bch_bkey_try_merge(b, k, m)) - goto copy; + trace_bcache_btree_insert_key(b, k, replace_key != NULL, + status); + return true; } else - m = bch_bset_search(b, &b->sets[b->nsets], k); - -insert: shift_keys(b, m, k); -copy: bkey_copy(m, k); -merged: - bch_check_keys(b, "%s for %s at %s: %s", status, - op_type(op), pbtree(b), pkey(k)); - bch_check_key_order_msg(b, i, "%s for %s at %s: %s", status, - op_type(op), pbtree(b), pkey(k)); - - if (b->level && !KEY_OFFSET(k)) - b->prio_blocked++; - - pr_debug("%s for %s at %s: %s", status, - op_type(op), pbtree(b), pkey(k)); - - return true; + return false; } -bool bch_btree_insert_keys(struct btree *b, struct btree_op *op) +static size_t insert_u64s_remaining(struct btree *b) { - bool ret = false; - struct bkey *k; - unsigned oldsize = bch_count_data(b); + long ret = bch_btree_keys_u64s_remaining(&b->keys); - while ((k = bch_keylist_pop(&op->keys))) { - bkey_put(b->c, k, b->level); - ret |= btree_insert_key(b, op, k); - } + /* + * Might land in the middle of an existing extent and have to split it + */ + if (b->keys.ops->is_extents) + ret -= KEY_MAX_U64S; - BUG_ON(bch_count_data(b) < oldsize); - return ret; + return max(ret, 0L); } -bool bch_btree_insert_check_key(struct btree *b, struct btree_op *op, - struct bio *bio) +static bool bch_btree_insert_keys(struct btree *b, struct btree_op *op, + struct keylist *insert_keys, + struct bkey *replace_key) { bool ret = false; - uint64_t btree_ptr = b->key.ptr[0]; - unsigned long seq = b->seq; - BKEY_PADDED(k) tmp; + int oldsize = bch_count_data(&b->keys); - rw_unlock(false, b); - rw_lock(true, b, b->level); + while (!bch_keylist_empty(insert_keys)) { + struct bkey *k = insert_keys->keys; - if (b->key.ptr[0] != btree_ptr || - b->seq != seq + 1 || - should_split(b)) - goto out; + if (bkey_u64s(k) > insert_u64s_remaining(b)) + break; - op->replace = KEY(op->inode, bio_end(bio), bio_sectors(bio)); + if (bkey_cmp(k, &b->key) <= 0) { + if (!b->level) + bkey_put(b->c, k); - SET_KEY_PTRS(&op->replace, 1); - get_random_bytes(&op->replace.ptr[0], sizeof(uint64_t)); + 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); - SET_PTR_DEV(&op->replace, 0, PTR_CHECK_DEV); + bch_cut_back(&b->key, &temp.key); + bch_cut_front(&b->key, insert_keys->keys); - bkey_copy(&tmp.k, &op->replace); + ret |= btree_insert_key(b, &temp.key, replace_key); + break; + } else { + break; + } + } - BUG_ON(op->type != BTREE_INSERT); - BUG_ON(!btree_insert_key(b, op, &tmp.k)); - bch_btree_write(b, false, NULL); - ret = true; -out: - downgrade_write(&b->lock); + 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) +static int btree_split(struct btree *b, struct btree_op *op, + struct keylist *insert_keys, + struct bkey *replace_key) { - bool split, root = b == b->c->root; + 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 (b->level) - set_closure_blocking(&op->cl); + 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->cl); + n1 = btree_node_alloc_replacement(b, op); if (IS_ERR(n1)) goto err; - split = set_blocks(n1->sets[0].data, n1->c) > (btree_blocks(b) * 4) / 5; - - pr_debug("%ssplitting at %s keys %i", split ? "" : "not ", - pbtree(b), n1->sets[0].data->keys); + split = set_blocks(btree_bset_first(n1), + block_bytes(n1->c)) > (btree_blocks(b) * 4) / 5; if (split) { unsigned keys = 0; - n2 = bch_btree_node_alloc(b->c, b->level, &op->cl); + trace_bcache_btree_node_split(b, btree_bset_first(n1)->keys); + + n2 = bch_btree_node_alloc(b->c, op, b->level); if (IS_ERR(n2)) goto err_free1; - if (root) { - n3 = bch_btree_node_alloc(b->c, b->level + 1, &op->cl); + if (!b->parent) { + n3 = bch_btree_node_alloc(b->c, op, b->level + 1); if (IS_ERR(n3)) goto err_free2; } - bch_btree_insert_keys(n1, op); + 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 + /* + * Has to be a linear search because we don't have an auxiliary * search tree yet */ - while (keys < (n1->sets[0].data->keys * 3) / 5) - keys += bkey_u64s(node(n1->sets[0].data, keys)); + 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, node(n1->sets[0].data, keys)); - keys += bkey_u64s(node(n1->sets[0].data, 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)); - n2->sets[0].data->keys = n1->sets[0].data->keys - keys; - n1->sets[0].data->keys = keys; + btree_bset_first(n2)->keys = btree_bset_first(n1)->keys - keys; + btree_bset_first(n1)->keys = keys; - memcpy(n2->sets[0].data->start, - end(n1->sets[0].data), - n2->sets[0].data->keys * sizeof(uint64_t)); + 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(&op->keys, &n2->key); - bch_btree_write(n2, true, op); + bch_keylist_add(&parent_keys, &n2->key); + bch_btree_node_write(n2, &cl); + mutex_unlock(&n2->write_lock); rw_unlock(true, n2); - } else - bch_btree_insert_keys(n1, op); + } else { + trace_bcache_btree_node_compact(b, btree_bset_first(n1)->keys); - bch_keylist_add(&op->keys, &n1->key); - bch_btree_write(n1, true, op); + 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); - bch_btree_write(n3, true, op); + bch_btree_insert_keys(n3, op, &parent_keys, NULL); + bch_btree_node_write(n3, &cl); + mutex_unlock(&n3->write_lock); - closure_sync(&op->cl); + closure_sync(&cl); bch_btree_set_root(n3); rw_unlock(true, n3); - } else if (root) { - op->keys.top = op->keys.bottom; - closure_sync(&op->cl); + } else if (!b->parent) { + /* Root filled up but didn't need to be split */ + closure_sync(&cl); bch_btree_set_root(n1); } else { - unsigned i; - - bkey_copy(op->keys.top, &b->key); - bkey_copy_key(op->keys.top, &ZERO_KEY); - - for (i = 0; i < KEY_PTRS(&b->key); i++) { - uint8_t g = PTR_BUCKET(b->c, &b->key, i)->gen + 1; - - SET_PTR_GEN(op->keys.top, i, g); - } + /* Split a non root node */ + closure_sync(&cl); + make_btree_freeing_key(b, parent_keys.top); + bch_keylist_push(&parent_keys); - bch_keylist_push(&op->keys); - closure_sync(&op->cl); - atomic_inc(&b->c->prio_blocked); + 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); - btree_node_free(b, op); bch_time_stats_update(&b->c->btree_split_time, start_time); return 0; err_free2: - __bkey_put(n2->c, &n2->key); - btree_node_free(n2, op); + bkey_put(b->c, &n2->key); + btree_node_free(n2); rw_unlock(true, n2); err_free1: - __bkey_put(n1->c, &n1->key); - btree_node_free(n1, op); + 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; - pr_warn("couldn't split"); return -ENOMEM; } -static int bch_btree_insert_recurse(struct btree *b, struct btree_op *op, - struct keylist *stack_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) { - if (b->level) { - int ret; - struct bkey *insert = op->keys.bottom; - struct bkey *k = bch_next_recurse_key(b, &START_KEY(insert)); + struct closure cl; - if (!k) { - btree_bug(b, "no key to recurse on at level %i/%i", - b->level, b->c->root->level); + BUG_ON(b->level && replace_key); - op->keys.top = op->keys.bottom; - return -EIO; - } + closure_init_stack(&cl); - if (bkey_cmp(insert, k) > 0) { - unsigned i; + mutex_lock(&b->write_lock); - if (op->type == BTREE_REPLACE) { - __bkey_put(b->c, insert); - op->keys.top = op->keys.bottom; - op->insert_collision = true; - return 0; - } + if (write_block(b) != btree_bset_last(b) && + b->keys.last_set_unwritten) + bch_btree_init_next(b); /* just wrote a set */ - for (i = 0; i < KEY_PTRS(insert); i++) - atomic_inc(&PTR_BUCKET(b->c, insert, i)->pin); + if (bch_keylist_nkeys(insert_keys) > insert_u64s_remaining(b)) { + mutex_unlock(&b->write_lock); + goto split; + } - bkey_copy(stack_keys->top, insert); + BUG_ON(write_block(b) != btree_bset_last(b)); - bch_cut_back(k, insert); - bch_cut_front(k, stack_keys->top); + 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); + } - bch_keylist_push(stack_keys); - } + mutex_unlock(&b->write_lock); - ret = btree(insert_recurse, k, b, op, stack_keys); - if (ret) - return ret; + /* 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; } +} - if (!bch_keylist_empty(&op->keys)) { - if (should_split(b)) { - if (op->lock <= b->c->root->level) { - BUG_ON(b->level); - op->lock = b->c->root->level + 1; - return -EINTR; - } - return btree_split(b, op); - } +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; - BUG_ON(write_block(b) != b->sets[b->nsets].data); + bch_keylist_init(&insert); - if (bch_btree_insert_keys(b, op)) - bch_btree_write(b, false, op); + if (upgrade) { + rw_unlock(false, b); + rw_lock(true, b, b->level); + + if (b->key.ptr[0] != btree_ptr || + b->seq != seq + 1) + goto out; } - return 0; + 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; } -int bch_btree_insert(struct btree_op *op, struct cache_set *c) +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) { - int ret = 0; - struct keylist stack_keys; + struct btree_insert_op *op = container_of(b_op, + struct btree_insert_op, op); - /* - * Don't want to block with the btree locked unless we have to, - * otherwise we get deadlocks with try_harder and between split/gc - */ - clear_closure_blocking(&op->cl); - - BUG_ON(bch_keylist_empty(&op->keys)); - bch_keylist_copy(&stack_keys, &op->keys); - bch_keylist_init(&op->keys); - - while (!bch_keylist_empty(&stack_keys) || - !bch_keylist_empty(&op->keys)) { - if (bch_keylist_empty(&op->keys)) { - bch_keylist_add(&op->keys, - bch_keylist_pop(&stack_keys)); - op->lock = 0; - } + 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; +} - ret = btree_root(insert_recurse, c, op, &stack_keys); +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; - if (ret == -EAGAIN) { - ret = 0; - closure_sync(&op->cl); - } else if (ret) { - struct bkey *k; + 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); + } - pr_err("error %i trying to insert key for %s", - ret, op_type(op)); + if (ret) { + struct bkey *k; - while ((k = bch_keylist_pop(&stack_keys) ?: - bch_keylist_pop(&op->keys))) - bkey_put(c, k, 0); - } - } + pr_err("error %i", ret); - bch_keylist_free(&stack_keys); + while ((k = bch_keylist_pop(keys))) + bkey_put(c, k); + } else if (op.op.insert_collision) + ret = -ESRCH; - if (op->journal) - atomic_dec_bug(op->journal); - op->journal = NULL; return ret; } void bch_btree_set_root(struct btree *b) { unsigned i; + struct closure cl; + + closure_init_stack(&cl); + + trace_bcache_btree_set_root(b); BUG_ON(!b->written); @@ -2151,139 +2243,81 @@ void bch_btree_set_root(struct btree *b) mutex_unlock(&b->c->bucket_lock); b->c->root = b; - __bkey_put(b->c, &b->key); - bch_journal_meta(b->c, NULL); - pr_debug("%s for %pf", pbtree(b), __builtin_return_address(0)); + bch_journal_meta(b->c, &cl); + closure_sync(&cl); } -/* Cache lookup */ +/* Map across nodes or keys */ -static int submit_partial_cache_miss(struct btree *b, struct btree_op *op, - struct bkey *k) +static int bch_btree_map_nodes_recurse(struct btree *b, struct btree_op *op, + struct bkey *from, + btree_map_nodes_fn *fn, int flags) { - struct search *s = container_of(op, struct search, op); - struct bio *bio = &s->bio.bio; - int ret = 0; + int ret = MAP_CONTINUE; - while (!ret && - !op->lookup_done) { - unsigned sectors = INT_MAX; + if (b->level) { + struct bkey *k; + struct btree_iter iter; - if (KEY_INODE(k) == op->inode) { - if (KEY_START(k) <= bio->bi_sector) - break; + bch_btree_iter_init(&b->keys, &iter, from); - sectors = min_t(uint64_t, sectors, - KEY_START(k) - bio->bi_sector); - } + while ((k = bch_btree_iter_next_filter(&iter, &b->keys, + bch_ptr_bad))) { + ret = btree(map_nodes_recurse, k, b, + op, from, fn, flags); + from = NULL; - ret = s->d->cache_miss(b, s, bio, sectors); + if (ret != MAP_CONTINUE) + return ret; + } } + if (!b->level || flags == MAP_ALL_NODES) + ret = fn(op, b); + return ret; } -/* - * Read from a single key, handling the initial cache miss if the key starts in - * the middle of the bio - */ -static int submit_partial_cache_hit(struct btree *b, struct btree_op *op, - struct bkey *k) +int __bch_btree_map_nodes(struct btree_op *op, struct cache_set *c, + struct bkey *from, btree_map_nodes_fn *fn, int flags) { - struct search *s = container_of(op, struct search, op); - struct bio *bio = &s->bio.bio; - unsigned ptr; - struct bio *n; - - int ret = submit_partial_cache_miss(b, op, k); - if (ret || op->lookup_done) - return ret; - - /* XXX: figure out best pointer - for multiple cache devices */ - ptr = 0; - - PTR_BUCKET(b->c, k, ptr)->prio = INITIAL_PRIO; - - while (!op->lookup_done && - KEY_INODE(k) == op->inode && - bio->bi_sector < KEY_OFFSET(k)) { - struct bkey *bio_key; - sector_t sector = PTR_OFFSET(k, ptr) + - (bio->bi_sector - KEY_START(k)); - unsigned sectors = min_t(uint64_t, INT_MAX, - KEY_OFFSET(k) - bio->bi_sector); - - n = bch_bio_split(bio, sectors, GFP_NOIO, s->d->bio_split); - if (!n) - return -EAGAIN; - - if (n == bio) - op->lookup_done = true; - - bio_key = &container_of(n, struct bbio, bio)->key; - - /* - * 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_bkey_copy_single_ptr(bio_key, k, ptr); - SET_PTR_OFFSET(bio_key, 0, sector); - - n->bi_end_io = bch_cache_read_endio; - n->bi_private = &s->cl; - - trace_bcache_cache_hit(n); - __bch_submit_bbio(n, b->c); - } - - return 0; + return btree_root(map_nodes_recurse, c, op, from, fn, flags); } -int bch_btree_search_recurse(struct btree *b, struct btree_op *op) +static int bch_btree_map_keys_recurse(struct btree *b, struct btree_op *op, + struct bkey *from, btree_map_keys_fn *fn, + int flags) { - struct search *s = container_of(op, struct search, op); - struct bio *bio = &s->bio.bio; - - int ret = 0; + int ret = MAP_CONTINUE; struct bkey *k; struct btree_iter iter; - bch_btree_iter_init(b, &iter, &KEY(op->inode, bio->bi_sector, 0)); - pr_debug("at %s searching for %u:%llu", pbtree(b), op->inode, - (uint64_t) bio->bi_sector); + bch_btree_iter_init(&b->keys, &iter, from); - do { - k = bch_btree_iter_next_filter(&iter, b, bch_ptr_bad); - if (!k) { - /* - * b->key would be exactly what we want, except that - * pointers to btree nodes have nonzero size - we - * wouldn't go far enough - */ + while ((k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad))) { + ret = !b->level + ? fn(op, b, k) + : btree(map_keys_recurse, k, b, op, from, fn, flags); + from = NULL; - ret = submit_partial_cache_miss(b, op, - &KEY(KEY_INODE(&b->key), - KEY_OFFSET(&b->key), 0)); - break; - } + if (ret != MAP_CONTINUE) + return ret; + } - ret = b->level - ? btree(search_recurse, k, b, op) - : submit_partial_cache_hit(b, op, k); - } while (!ret && - !op->lookup_done); + 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 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) @@ -2302,81 +2336,79 @@ static inline int keybuf_nonoverlapping_cmp(struct keybuf_key *l, return clamp_t(int64_t, bkey_cmp(&l->key, &r->key), -1, 1); } -static int bch_btree_refill_keybuf(struct btree *b, struct btree_op *op, - struct keybuf *buf, struct bkey *end) -{ - struct btree_iter iter; - bch_btree_iter_init(b, &iter, &buf->last_scanned); - - while (!array_freelist_empty(&buf->freelist)) { - struct bkey *k = bch_btree_iter_next_filter(&iter, b, - bch_ptr_bad); - - if (!b->level) { - if (!k) { - buf->last_scanned = b->key; - break; - } - - buf->last_scanned = *k; - if (bkey_cmp(&buf->last_scanned, end) >= 0) - break; +struct refill { + struct btree_op op; + unsigned nr_found; + struct keybuf *buf; + struct bkey *end; + keybuf_pred_fn *pred; +}; - if (buf->key_predicate(buf, k)) { - struct keybuf_key *w; +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; - pr_debug("%s", pkey(k)); + if (bkey_cmp(k, refill->end) >= 0) { + ret = MAP_DONE; + goto out; + } - spin_lock(&buf->lock); + if (!KEY_SIZE(k)) /* end key */ + goto out; - w = array_alloc(&buf->freelist); + if (refill->pred(buf, k)) { + struct keybuf_key *w; - w->private = NULL; - bkey_copy(&w->key, k); + spin_lock(&buf->lock); - if (RB_INSERT(&buf->keys, w, node, keybuf_cmp)) - array_free(&buf->freelist, w); + w = array_alloc(&buf->freelist); + if (!w) { + spin_unlock(&buf->lock); + return MAP_DONE; + } - spin_unlock(&buf->lock); - } - } else { - if (!k) - break; + w->private = NULL; + bkey_copy(&w->key, k); - btree(refill_keybuf, k, b, op, buf, end); - /* - * Might get an error here, but can't really do anything - * and it'll get logged elsewhere. Just read what we - * can. - */ + if (RB_INSERT(&buf->keys, w, node, keybuf_cmp)) + array_free(&buf->freelist, w); + else + refill->nr_found++; - if (bkey_cmp(&buf->last_scanned, end) >= 0) - break; + if (array_freelist_empty(&buf->freelist)) + ret = MAP_DONE; - cond_resched(); - } + spin_unlock(&buf->lock); } - - return 0; +out: + buf->last_scanned = *k; + return ret; } void bch_refill_keybuf(struct cache_set *c, struct keybuf *buf, - struct bkey *end) + struct bkey *end, keybuf_pred_fn *pred) { struct bkey start = buf->last_scanned; - struct btree_op op; - bch_btree_op_init_stack(&op); + struct refill refill; cond_resched(); - btree_root(refill_keybuf, c, &op, buf, end); - closure_sync(&op.cl); + bch_btree_op_init(&refill.op, -1); + refill.nr_found = 0; + refill.buf = buf; + refill.end = end; + refill.pred = pred; - pr_debug("found %s keys from %llu:%llu to %llu:%llu", - RB_EMPTY_ROOT(&buf->keys) ? "no" : - array_freelist_empty(&buf->freelist) ? "some" : "a few", - KEY_INODE(&start), KEY_OFFSET(&start), - KEY_INODE(&buf->last_scanned), KEY_OFFSET(&buf->last_scanned)); + 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); @@ -2454,8 +2486,9 @@ 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) + struct keybuf *buf, + struct bkey *end, + keybuf_pred_fn *pred) { struct keybuf_key *ret; @@ -2469,35 +2502,17 @@ struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *c, break; } - bch_refill_keybuf(c, buf, end); + bch_refill_keybuf(c, buf, end, pred); } return ret; } -void bch_keybuf_init(struct keybuf *buf, keybuf_pred_fn *fn) +void bch_keybuf_init(struct keybuf *buf) { - buf->key_predicate = fn; 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); - if (bch_gc_wq) - destroy_workqueue(bch_gc_wq); -} - -int __init bch_btree_init(void) -{ - if (!(bch_gc_wq = create_singlethread_workqueue("bch_btree_gc")) || - !(btree_io_wq = create_singlethread_workqueue("bch_btree_io"))) - return -ENOMEM; - - return 0; -} diff --git a/drivers/md/bcache/btree.h b/drivers/md/bcache/btree.h index af4a7092a28..91dfa5e6968 100644 --- a/drivers/md/bcache/btree.h +++ b/drivers/md/bcache/btree.h @@ -102,7 +102,6 @@ #include "debug.h" struct btree_write { - struct closure *owner; atomic_t *journal; /* If btree_split() frees a btree node, it writes a new pointer to that @@ -126,32 +125,23 @@ struct btree { 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; - uint8_t nsets; - uint8_t page_order; - - /* - * Set of sorted keys - the real btree node - plus a binary search tree - * - * sets[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 sets[MAX_BSETS]; - /* Used to refcount bio splits, also protects b->bio */ - struct closure_with_waitlist io; + struct btree_keys keys; - /* Gets transferred to w->prio_blocked - see the comment there */ - int prio_blocked; + /* 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; - uint64_t io_start_time; struct btree_write writes[2]; struct bio *bio; }; @@ -164,13 +154,11 @@ static inline void set_btree_node_ ## flag(struct btree *b) \ { set_bit(BTREE_NODE_ ## flag, &b->flags); } \ enum btree_flags { - BTREE_NODE_read_done, BTREE_NODE_io_error, BTREE_NODE_dirty, BTREE_NODE_write_idx, }; -BTREE_FLAG(read_done); BTREE_FLAG(io_error); BTREE_FLAG(dirty); BTREE_FLAG(write_idx); @@ -185,42 +173,27 @@ static inline struct btree_write *btree_prev_write(struct btree *b) return b->writes + (btree_node_write_idx(b) ^ 1); } -static inline unsigned bset_offset(struct btree *b, struct bset *i) -{ - return (((size_t) i) - ((size_t) b->sets->data)) >> 9; -} - -static inline struct bset *write_block(struct btree *b) +static inline struct bset *btree_bset_first(struct btree *b) { - return ((void *) b->sets[0].data) + b->written * block_bytes(b->c); + return b->keys.set->data; } -static inline bool bset_written(struct btree *b, struct bset_tree *t) +static inline struct bset *btree_bset_last(struct btree *b) { - return t->data < write_block(b); + return bset_tree_last(&b->keys)->data; } -static inline bool bkey_written(struct btree *b, struct bkey *k) +static inline unsigned bset_block_offset(struct btree *b, struct bset *i) { - return k < write_block(b)->start; + 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->sb.bucket_size * c->nbuckets / 8); + atomic_set(&c->sectors_to_gc, c->sb.bucket_size * c->nbuckets / 16); } -static inline bool bch_ptr_invalid(struct btree *b, const struct bkey *k) -{ - return __bch_ptr_invalid(b->c, b->level, k); -} - -static inline struct bkey *bch_btree_iter_init(struct btree *b, - struct btree_iter *iter, - struct bkey *search) -{ - return __bch_btree_iter_init(b, iter, search, b->sets); -} +void bkey_put(struct cache_set *c, struct bkey *k); /* Looping macros */ @@ -230,55 +203,24 @@ static inline struct bkey *bch_btree_iter_init(struct btree *b, iter++) \ hlist_for_each_entry_rcu((b), (c)->bucket_hash + iter, hash) -#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));) - /* Recursing down the btree */ struct btree_op { - struct closure cl; - struct cache_set *c; - - /* Journal entry we have a refcount on */ - atomic_t *journal; - - /* Bio to be inserted into the cache */ - struct bio *cache_bio; - - unsigned inode; - - uint16_t write_prio; + /* for waiting on btree reserve in btree_split() */ + wait_queue_t wait; /* Btree level at which we start taking write locks */ short lock; - /* Btree insertion type */ - enum { - BTREE_INSERT, - BTREE_REPLACE - } type:8; - - unsigned csum:1; - unsigned skip:1; - unsigned flush_journal:1; - - unsigned insert_data_done:1; - unsigned lookup_done:1; unsigned insert_collision:1; - - /* Anything after this point won't get zeroed in do_bio_hook() */ - - /* Keys to be inserted */ - struct keylist keys; - BKEY_PADDED(replace); }; -void bch_btree_op_init_stack(struct btree_op *); +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) { @@ -290,116 +232,78 @@ static inline void rw_lock(bool w, struct btree *b, int level) static inline void rw_unlock(bool w, struct btree *b) { -#ifdef CONFIG_BCACHE_EDEBUG - unsigned i; - - if (w && - b->key.ptr[0] && - btree_node_read_done(b)) - for (i = 0; i <= b->nsets; i++) - bch_check_key_order(b, b->sets[i].data); -#endif - if (w) b->seq++; (w ? up_write : up_read)(&b->lock); } -#define insert_lock(s, b) ((b)->level <= (s)->lock) - -/* - * 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 btree(fn, key, b, op, ...) \ -({ \ - int _r, l = (b)->level - 1; \ - bool _w = l <= (op)->lock; \ - struct btree *_b = bch_btree_node_get((b)->c, key, l, op); \ - if (!IS_ERR(_b)) { \ - _r = bch_btree_ ## fn(_b, op, ##__VA_ARGS__); \ - rw_unlock(_w, _b); \ - } else \ - _r = PTR_ERR(_b); \ - _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 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, &(op)->cl); \ - } while (_r == -EINTR); \ - \ - _r; \ -}) +void bch_btree_node_read_done(struct btree *); +void __bch_btree_node_write(struct btree *, struct closure *); +void bch_btree_node_write(struct btree *, struct closure *); -static inline bool should_split(struct btree *b) +void bch_btree_set_root(struct btree *); +struct btree *bch_btree_node_alloc(struct cache_set *, struct btree_op *, int); +struct btree *bch_btree_node_get(struct cache_set *, struct btree_op *, + struct bkey *, int, bool); + +int bch_btree_insert_check_key(struct btree *, struct btree_op *, + struct bkey *); +int bch_btree_insert(struct cache_set *, struct keylist *, + atomic_t *, struct bkey *); + +int bch_gc_thread_start(struct cache_set *); +void bch_initial_gc_finish(struct cache_set *); +void bch_moving_gc(struct cache_set *); +int bch_btree_check(struct cache_set *); +void bch_initial_mark_key(struct cache_set *, int, struct bkey *); + +static inline void wake_up_gc(struct cache_set *c) { - struct bset *i = write_block(b); - return b->written >= btree_blocks(b) || - (i->seq == b->sets[0].data->seq && - b->written + __set_blocks(i, i->keys + 15, b->c) - > btree_blocks(b)); + if (c->gc_thread) + wake_up_process(c->gc_thread); } -void bch_btree_read_done(struct closure *); -void bch_btree_read(struct btree *); -void bch_btree_write(struct btree *b, bool now, struct btree_op *op); +#define MAP_DONE 0 +#define MAP_CONTINUE 1 -void bch_cannibalize_unlock(struct cache_set *, struct closure *); -void bch_btree_set_root(struct btree *); -struct btree *bch_btree_node_alloc(struct cache_set *, int, struct closure *); -struct btree *bch_btree_node_get(struct cache_set *, struct bkey *, - int, struct btree_op *); +#define MAP_ALL_NODES 0 +#define MAP_LEAF_NODES 1 + +#define MAP_END_KEY 1 + +typedef int (btree_map_nodes_fn)(struct btree_op *, struct btree *); +int __bch_btree_map_nodes(struct btree_op *, struct cache_set *, + struct bkey *, btree_map_nodes_fn *, int); -bool bch_btree_insert_keys(struct btree *, struct btree_op *); -bool bch_btree_insert_check_key(struct btree *, struct btree_op *, - struct bio *); -int bch_btree_insert(struct btree_op *, struct cache_set *); +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); +} -int bch_btree_search_recurse(struct btree *, struct btree_op *); +typedef int (btree_map_keys_fn)(struct btree_op *, struct btree *, + struct bkey *); +int bch_btree_map_keys(struct btree_op *, struct cache_set *, + struct bkey *, btree_map_keys_fn *, int); -void bch_queue_gc(struct cache_set *); -size_t bch_btree_gc_finish(struct cache_set *); -void bch_moving_gc(struct closure *); -int bch_btree_check(struct cache_set *, struct btree_op *); -uint8_t __bch_btree_mark_key(struct cache_set *, int, struct bkey *); +typedef bool (keybuf_pred_fn)(struct keybuf *, struct bkey *); -void bch_keybuf_init(struct keybuf *, keybuf_pred_fn *); -void bch_refill_keybuf(struct cache_set *, struct keybuf *, struct bkey *); +void bch_keybuf_init(struct keybuf *); +void bch_refill_keybuf(struct cache_set *, struct keybuf *, + struct bkey *, keybuf_pred_fn *); bool bch_keybuf_check_overlapping(struct keybuf *, struct bkey *, struct bkey *); void bch_keybuf_del(struct keybuf *, struct keybuf_key *); struct keybuf_key *bch_keybuf_next(struct keybuf *); -struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *, - struct keybuf *, struct bkey *); +struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *, struct keybuf *, + struct bkey *, keybuf_pred_fn *); #endif diff --git a/drivers/md/bcache/closure.c b/drivers/md/bcache/closure.c index bd05a9a8c7c..7a228de95fd 100644 --- a/drivers/md/bcache/closure.c +++ b/drivers/md/bcache/closure.c @@ -11,47 +11,12 @@ #include "closure.h" -void closure_queue(struct closure *cl) -{ - struct workqueue_struct *wq = cl->wq; - if (wq) { - INIT_WORK(&cl->work, cl->work.func); - BUG_ON(!queue_work(wq, &cl->work)); - } else - cl->fn(cl); -} -EXPORT_SYMBOL_GPL(closure_queue); - -#define CL_FIELD(type, field) \ - case TYPE_ ## type: \ - return &container_of(cl, struct type, cl)->field - -static struct closure_waitlist *closure_waitlist(struct closure *cl) -{ - switch (cl->type) { - CL_FIELD(closure_with_waitlist, wait); - CL_FIELD(closure_with_waitlist_and_timer, wait); - default: - return NULL; - } -} - -static struct timer_list *closure_timer(struct closure *cl) -{ - switch (cl->type) { - CL_FIELD(closure_with_timer, timer); - CL_FIELD(closure_with_waitlist_and_timer, timer); - default: - return NULL; - } -} - 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|CLOSURE_BLOCKING))); + BUG_ON(!r && (flags & ~CLOSURE_DESTRUCTOR)); /* Must deliver precisely one wakeup */ if (r == 1 && (flags & CLOSURE_SLEEPING)) @@ -59,23 +24,17 @@ static inline void closure_put_after_sub(struct closure *cl, int flags) if (!r) { if (cl->fn && !(flags & CLOSURE_DESTRUCTOR)) { - /* CLOSURE_BLOCKING might be set - clear it */ atomic_set(&cl->remaining, CLOSURE_REMAINING_INITIALIZER); closure_queue(cl); } else { struct closure *parent = cl->parent; - struct closure_waitlist *wait = closure_waitlist(cl); + closure_fn *destructor = cl->fn; closure_debug_destroy(cl); - atomic_set(&cl->remaining, -1); - - if (wait) - closure_wake_up(wait); - - if (cl->fn) - cl->fn(cl); + if (destructor) + destructor(cl); if (parent) closure_put(parent); @@ -88,21 +47,20 @@ void closure_sub(struct closure *cl, int v) { closure_put_after_sub(cl, atomic_sub_return(v, &cl->remaining)); } -EXPORT_SYMBOL_GPL(closure_sub); +EXPORT_SYMBOL(closure_sub); +/** + * closure_put - decrement a closure's refcount + */ void closure_put(struct closure *cl) { closure_put_after_sub(cl, atomic_dec_return(&cl->remaining)); } -EXPORT_SYMBOL_GPL(closure_put); - -static void set_waiting(struct closure *cl, unsigned long f) -{ -#ifdef CONFIG_BCACHE_CLOSURES_DEBUG - cl->waiting_on = f; -#endif -} +EXPORT_SYMBOL(closure_put); +/** + * 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; @@ -127,27 +85,34 @@ void __closure_wake_up(struct closure_waitlist *wait_list) cl = container_of(reverse, struct closure, list); reverse = llist_next(reverse); - set_waiting(cl, 0); + closure_set_waiting(cl, 0); closure_sub(cl, CLOSURE_WAITING + 1); } } -EXPORT_SYMBOL_GPL(__closure_wake_up); +EXPORT_SYMBOL(__closure_wake_up); -bool closure_wait(struct closure_waitlist *list, struct closure *cl) +/** + * 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. + * + */ +bool closure_wait(struct closure_waitlist *waitlist, struct closure *cl) { if (atomic_read(&cl->remaining) & CLOSURE_WAITING) return false; - set_waiting(cl, _RET_IP_); + closure_set_waiting(cl, _RET_IP_); atomic_add(CLOSURE_WAITING + 1, &cl->remaining); - llist_add(&cl->list, &list->list); + llist_add(&cl->list, &waitlist->list); return true; } -EXPORT_SYMBOL_GPL(closure_wait); +EXPORT_SYMBOL(closure_wait); /** - * closure_sync() - sleep until a closure a closure has nothing left to wait on + * 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. @@ -167,93 +132,7 @@ void closure_sync(struct closure *cl) __closure_end_sleep(cl); } -EXPORT_SYMBOL_GPL(closure_sync); - -/** - * closure_trylock() - try to acquire the closure, without waiting - * @cl: closure to lock - * - * Returns true if the closure was succesfully locked. - */ -bool closure_trylock(struct closure *cl, struct closure *parent) -{ - if (atomic_cmpxchg(&cl->remaining, -1, - CLOSURE_REMAINING_INITIALIZER) != -1) - return false; - - closure_set_ret_ip(cl); - - smp_mb(); - cl->parent = parent; - if (parent) - closure_get(parent); - - closure_debug_create(cl); - return true; -} -EXPORT_SYMBOL_GPL(closure_trylock); - -void __closure_lock(struct closure *cl, struct closure *parent, - struct closure_waitlist *wait_list) -{ - struct closure wait; - closure_init_stack(&wait); - - while (1) { - if (closure_trylock(cl, parent)) - return; - - closure_wait_event_sync(wait_list, &wait, - atomic_read(&cl->remaining) == -1); - } -} -EXPORT_SYMBOL_GPL(__closure_lock); - -static void closure_delay_timer_fn(unsigned long data) -{ - struct closure *cl = (struct closure *) data; - closure_sub(cl, CLOSURE_TIMER + 1); -} - -void do_closure_timer_init(struct closure *cl) -{ - struct timer_list *timer = closure_timer(cl); - - init_timer(timer); - timer->data = (unsigned long) cl; - timer->function = closure_delay_timer_fn; -} -EXPORT_SYMBOL_GPL(do_closure_timer_init); - -bool __closure_delay(struct closure *cl, unsigned long delay, - struct timer_list *timer) -{ - if (atomic_read(&cl->remaining) & CLOSURE_TIMER) - return false; - - BUG_ON(timer_pending(timer)); - - timer->expires = jiffies + delay; - - atomic_add(CLOSURE_TIMER + 1, &cl->remaining); - add_timer(timer); - return true; -} -EXPORT_SYMBOL_GPL(__closure_delay); - -void __closure_flush(struct closure *cl, struct timer_list *timer) -{ - if (del_timer(timer)) - closure_sub(cl, CLOSURE_TIMER + 1); -} -EXPORT_SYMBOL_GPL(__closure_flush); - -void __closure_flush_sync(struct closure *cl, struct timer_list *timer) -{ - if (del_timer_sync(timer)) - closure_sub(cl, CLOSURE_TIMER + 1); -} -EXPORT_SYMBOL_GPL(__closure_flush_sync); +EXPORT_SYMBOL(closure_sync); #ifdef CONFIG_BCACHE_CLOSURES_DEBUG @@ -271,7 +150,7 @@ void closure_debug_create(struct closure *cl) list_add(&cl->all, &closure_list); spin_unlock_irqrestore(&closure_list_lock, flags); } -EXPORT_SYMBOL_GPL(closure_debug_create); +EXPORT_SYMBOL(closure_debug_create); void closure_debug_destroy(struct closure *cl) { @@ -284,7 +163,7 @@ void closure_debug_destroy(struct closure *cl) list_del(&cl->all); spin_unlock_irqrestore(&closure_list_lock, flags); } -EXPORT_SYMBOL_GPL(closure_debug_destroy); +EXPORT_SYMBOL(closure_debug_destroy); static struct dentry *debug; @@ -302,14 +181,12 @@ static int debug_seq_show(struct seq_file *f, void *data) cl, (void *) cl->ip, cl->fn, cl->parent, r & CLOSURE_REMAINING_MASK); - seq_printf(f, "%s%s%s%s%s%s\n", + seq_printf(f, "%s%s%s%s\n", test_bit(WORK_STRUCT_PENDING, work_data_bits(&cl->work)) ? "Q" : "", r & CLOSURE_RUNNING ? "R" : "", - r & CLOSURE_BLOCKING ? "B" : "", r & CLOSURE_STACK ? "S" : "", - r & CLOSURE_SLEEPING ? "Sl" : "", - r & CLOSURE_TIMER ? "T" : ""); + r & CLOSURE_SLEEPING ? "Sl" : ""); if (r & CLOSURE_WAITING) seq_printf(f, " W %pF\n", diff --git a/drivers/md/bcache/closure.h b/drivers/md/bcache/closure.h index 00039924ea9..a08e3eeac3c 100644 --- a/drivers/md/bcache/closure.h +++ b/drivers/md/bcache/closure.h @@ -72,30 +72,6 @@ * closure - _always_ use continue_at(). Doing so consistently will help * eliminate an entire class of particularly pernicious races. * - * For a closure to wait on an arbitrary event, we need to introduce waitlists: - * - * struct closure_waitlist list; - * closure_wait_event(list, cl, condition); - * closure_wake_up(wait_list); - * - * These work analagously to wait_event() and wake_up() - except that instead of - * operating on the current thread (for wait_event()) and lists of threads, they - * operate on an explicit closure and lists of closures. - * - * Because it's a closure we can now wait either synchronously or - * asynchronously. closure_wait_event() returns the current value of the - * condition, and if it returned false continue_at() or closure_sync() can be - * used to wait for it to become true. - * - * It's useful for waiting on things when you can't sleep in the context in - * which you must check the condition (perhaps a spinlock held, or you might be - * beneath generic_make_request() - in which case you can't sleep on IO). - * - * closure_wait_event() will wait either synchronously or asynchronously, - * depending on whether the closure is in blocking mode or not. You can pick a - * mode explicitly with closure_wait_event_sync() and - * closure_wait_event_async(), which do just what you might expect. - * * 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 @@ -121,55 +97,6 @@ * 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. - * - * Locking: - * - * Closures are based on work items but they can be thought of as more like - * threads - in that like threads and unlike work items they have a well - * defined lifetime; they are created (with closure_init()) and eventually - * complete after a continue_at(cl, NULL, NULL). - * - * Suppose you've got some larger structure with a closure embedded in it that's - * used for periodically doing garbage collection. You only want one garbage - * collection happening at a time, so the natural thing to do is protect it with - * a lock. However, it's difficult to use a lock protecting a closure correctly - * because the unlock should come after the last continue_to() (additionally, if - * you're using the closure asynchronously a mutex won't work since a mutex has - * to be unlocked by the same process that locked it). - * - * So to make it less error prone and more efficient, we also have the ability - * to use closures as locks: - * - * closure_init_unlocked(); - * closure_trylock(); - * - * That's all we need for trylock() - the last closure_put() implicitly unlocks - * it for you. But for closure_lock(), we also need a wait list: - * - * struct closure_with_waitlist frobnicator_cl; - * - * closure_init_unlocked(&frobnicator_cl); - * closure_lock(&frobnicator_cl); - * - * A closure_with_waitlist embeds a closure and a wait list - much like struct - * delayed_work embeds a work item and a timer_list. The important thing is, use - * it exactly like you would a regular closure and closure_put() will magically - * handle everything for you. - * - * We've got closures that embed timers, too. They're called, appropriately - * enough: - * struct closure_with_timer; - * - * This gives you access to closure_delay(). It takes a refcount for a specified - * number of jiffies - you could then call closure_sync() (for a slightly - * convoluted version of msleep()) or continue_at() - which gives you the same - * effect as using a delayed work item, except you can reuse the work_struct - * already embedded in struct closure. - * - * Lastly, there's struct closure_with_waitlist_and_timer. It does what you - * probably expect, if you happen to need the features of both. (You don't - * really want to know how all this is implemented, but if I've done my job - * right you shouldn't have to care). */ struct closure; @@ -179,19 +106,8 @@ struct closure_waitlist { struct llist_head list; }; -enum closure_type { - TYPE_closure = 0, - TYPE_closure_with_waitlist = 1, - TYPE_closure_with_timer = 2, - TYPE_closure_with_waitlist_and_timer = 3, - MAX_CLOSURE_TYPE = 3, -}; - enum closure_state { /* - * CLOSURE_BLOCKING: Causes closure_wait_event() to block, instead of - * waiting asynchronously - * * 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. @@ -200,10 +116,6 @@ enum closure_state { * - indicates that cl->task is valid and closure_put() may wake it up. * Only set or cleared by the thread that owns the closure. * - * CLOSURE_TIMER: Analagous to CLOSURE_WAITING, indicates that a closure - * has an outstanding timer. Must be set by the thread that owns the - * closure, and cleared by the timer function when the timer goes off. - * * The rest are for debugging and don't affect behaviour: * * CLOSURE_RUNNING: Set when a closure is running (i.e. by @@ -218,19 +130,17 @@ enum closure_state { * closure with this flag set */ - CLOSURE_BITS_START = (1 << 19), - CLOSURE_DESTRUCTOR = (1 << 19), - CLOSURE_BLOCKING = (1 << 21), - CLOSURE_WAITING = (1 << 23), - CLOSURE_SLEEPING = (1 << 25), - CLOSURE_TIMER = (1 << 27), + CLOSURE_BITS_START = (1 << 23), + CLOSURE_DESTRUCTOR = (1 << 23), + CLOSURE_WAITING = (1 << 25), + CLOSURE_SLEEPING = (1 << 27), CLOSURE_RUNNING = (1 << 29), CLOSURE_STACK = (1 << 31), }; #define CLOSURE_GUARD_MASK \ - ((CLOSURE_DESTRUCTOR|CLOSURE_BLOCKING|CLOSURE_WAITING| \ - CLOSURE_SLEEPING|CLOSURE_TIMER|CLOSURE_RUNNING|CLOSURE_STACK) << 1) + ((CLOSURE_DESTRUCTOR|CLOSURE_WAITING|CLOSURE_SLEEPING| \ + CLOSURE_RUNNING|CLOSURE_STACK) << 1) #define CLOSURE_REMAINING_MASK (CLOSURE_BITS_START - 1) #define CLOSURE_REMAINING_INITIALIZER (1|CLOSURE_RUNNING) @@ -250,8 +160,6 @@ struct closure { atomic_t remaining; - enum closure_type type; - #ifdef CONFIG_BCACHE_CLOSURES_DEBUG #define CLOSURE_MAGIC_DEAD 0xc054dead #define CLOSURE_MAGIC_ALIVE 0xc054a11e @@ -263,54 +171,12 @@ struct closure { #endif }; -struct closure_with_waitlist { - struct closure cl; - struct closure_waitlist wait; -}; - -struct closure_with_timer { - struct closure cl; - struct timer_list timer; -}; - -struct closure_with_waitlist_and_timer { - struct closure cl; - struct closure_waitlist wait; - struct timer_list timer; -}; - -extern unsigned invalid_closure_type(void); - -#define __CLOSURE_TYPE(cl, _t) \ - __builtin_types_compatible_p(typeof(cl), struct _t) \ - ? TYPE_ ## _t : \ - -#define __closure_type(cl) \ -( \ - __CLOSURE_TYPE(cl, closure) \ - __CLOSURE_TYPE(cl, closure_with_waitlist) \ - __CLOSURE_TYPE(cl, closure_with_timer) \ - __CLOSURE_TYPE(cl, closure_with_waitlist_and_timer) \ - invalid_closure_type() \ -) - void closure_sub(struct closure *cl, int v); void closure_put(struct closure *cl); -void closure_queue(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); -bool closure_trylock(struct closure *cl, struct closure *parent); -void __closure_lock(struct closure *cl, struct closure *parent, - struct closure_waitlist *wait_list); - -void do_closure_timer_init(struct closure *cl); -bool __closure_delay(struct closure *cl, unsigned long delay, - struct timer_list *timer); -void __closure_flush(struct closure *cl, struct timer_list *timer); -void __closure_flush_sync(struct closure *cl, struct timer_list *timer); - #ifdef CONFIG_BCACHE_CLOSURES_DEBUG void closure_debug_init(void); @@ -339,200 +205,97 @@ static inline void closure_set_ret_ip(struct closure *cl) #endif } -static inline void closure_get(struct closure *cl) +static inline void closure_set_waiting(struct closure *cl, unsigned long f) { #ifdef CONFIG_BCACHE_CLOSURES_DEBUG - BUG_ON((atomic_inc_return(&cl->remaining) & - CLOSURE_REMAINING_MASK) <= 1); -#else - atomic_inc(&cl->remaining); + cl->waiting_on = f; #endif } -static inline void closure_set_stopped(struct closure *cl) +static inline void __closure_end_sleep(struct closure *cl) { - atomic_sub(CLOSURE_RUNNING, &cl->remaining); + __set_current_state(TASK_RUNNING); + + if (atomic_read(&cl->remaining) & CLOSURE_SLEEPING) + atomic_sub(CLOSURE_SLEEPING, &cl->remaining); } -static inline bool closure_is_stopped(struct closure *cl) +static inline void __closure_start_sleep(struct closure *cl) { - return !(atomic_read(&cl->remaining) & CLOSURE_RUNNING); + closure_set_ip(cl); + cl->task = current; + set_current_state(TASK_UNINTERRUPTIBLE); + + if (!(atomic_read(&cl->remaining) & CLOSURE_SLEEPING)) + atomic_add(CLOSURE_SLEEPING, &cl->remaining); } -static inline bool closure_is_unlocked(struct closure *cl) +static inline void closure_set_stopped(struct closure *cl) { - return atomic_read(&cl->remaining) == -1; + atomic_sub(CLOSURE_RUNNING, &cl->remaining); } -static inline void do_closure_init(struct closure *cl, struct closure *parent, - bool running) +static inline void set_closure_fn(struct closure *cl, closure_fn *fn, + struct workqueue_struct *wq) { - switch (cl->type) { - case TYPE_closure_with_timer: - case TYPE_closure_with_waitlist_and_timer: - do_closure_timer_init(cl); - default: - break; - } - - cl->parent = parent; - if (parent) - closure_get(parent); - - if (running) { - closure_debug_create(cl); - atomic_set(&cl->remaining, CLOSURE_REMAINING_INITIALIZER); - } else - atomic_set(&cl->remaining, -1); - + BUG_ON(object_is_on_stack(cl)); closure_set_ip(cl); + cl->fn = fn; + cl->wq = wq; + /* between atomic_dec() in closure_put() */ + smp_mb__before_atomic(); } -/* - * Hack to get at the embedded closure if there is one, by doing an unsafe cast: - * the result of __closure_type() is thrown away, it's used merely for type - * checking. - */ -#define __to_internal_closure(cl) \ -({ \ - BUILD_BUG_ON(__closure_type(*cl) > MAX_CLOSURE_TYPE); \ - (struct closure *) cl; \ -}) - -#define closure_init_type(cl, parent, running) \ -do { \ - struct closure *_cl = __to_internal_closure(cl); \ - _cl->type = __closure_type(*(cl)); \ - do_closure_init(_cl, parent, running); \ -} while (0) +static inline void closure_queue(struct closure *cl) +{ + struct workqueue_struct *wq = cl->wq; + if (wq) { + INIT_WORK(&cl->work, cl->work.func); + BUG_ON(!queue_work(wq, &cl->work)); + } else + cl->fn(cl); +} /** - * __closure_init() - Initialize a closure, skipping the memset() - * - * May be used instead of closure_init() when memory has already been zeroed. + * closure_get - increment a closure's refcount */ -#define __closure_init(cl, parent) \ - closure_init_type(cl, parent, true) +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 + * 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. */ -#define closure_init(cl, parent) \ -do { \ - memset((cl), 0, sizeof(*(cl))); \ - __closure_init(cl, parent); \ -} while (0) - -static inline void closure_init_stack(struct closure *cl) +static inline void closure_init(struct closure *cl, struct closure *parent) { memset(cl, 0, sizeof(struct closure)); - atomic_set(&cl->remaining, CLOSURE_REMAINING_INITIALIZER| - CLOSURE_BLOCKING|CLOSURE_STACK); -} - -/** - * closure_init_unlocked() - Initialize a closure but leave it unlocked. - * @cl: closure to initialize - * - * For when the closure will be used as a lock. The closure may not be used - * until after a closure_lock() or closure_trylock(). - */ -#define closure_init_unlocked(cl) \ -do { \ - memset((cl), 0, sizeof(*(cl))); \ - closure_init_type(cl, NULL, false); \ -} while (0) - -/** - * closure_lock() - lock and initialize a closure. - * @cl: the closure to lock - * @parent: the new parent for this closure - * - * The closure must be of one of the types that has a waitlist (otherwise we - * wouldn't be able to sleep on contention). - * - * @parent has exactly the same meaning as in closure_init(); if non null, the - * closure will take a reference on @parent which will be released when it is - * unlocked. - */ -#define closure_lock(cl, parent) \ - __closure_lock(__to_internal_closure(cl), parent, &(cl)->wait) - -/** - * closure_delay() - delay some number of jiffies - * @cl: the closure that will sleep - * @delay: the delay in jiffies - * - * Takes a refcount on @cl which will be released after @delay jiffies; this may - * be used to have a function run after a delay with continue_at(), or - * closure_sync() may be used for a convoluted version of msleep(). - */ -#define closure_delay(cl, delay) \ - __closure_delay(__to_internal_closure(cl), delay, &(cl)->timer) - -#define closure_flush(cl) \ - __closure_flush(__to_internal_closure(cl), &(cl)->timer) - -#define closure_flush_sync(cl) \ - __closure_flush_sync(__to_internal_closure(cl), &(cl)->timer) - -static inline void __closure_end_sleep(struct closure *cl) -{ - __set_current_state(TASK_RUNNING); + cl->parent = parent; + if (parent) + closure_get(parent); - if (atomic_read(&cl->remaining) & CLOSURE_SLEEPING) - atomic_sub(CLOSURE_SLEEPING, &cl->remaining); -} + atomic_set(&cl->remaining, CLOSURE_REMAINING_INITIALIZER); -static inline void __closure_start_sleep(struct closure *cl) -{ + closure_debug_create(cl); closure_set_ip(cl); - cl->task = current; - set_current_state(TASK_UNINTERRUPTIBLE); - - if (!(atomic_read(&cl->remaining) & CLOSURE_SLEEPING)) - atomic_add(CLOSURE_SLEEPING, &cl->remaining); -} - -/** - * closure_blocking() - returns true if the closure is in blocking mode. - * - * If a closure is in blocking mode, closure_wait_event() will sleep until the - * condition is true instead of waiting asynchronously. - */ -static inline bool closure_blocking(struct closure *cl) -{ - return atomic_read(&cl->remaining) & CLOSURE_BLOCKING; -} - -/** - * set_closure_blocking() - put a closure in blocking mode. - * - * If a closure is in blocking mode, closure_wait_event() will sleep until the - * condition is true instead of waiting asynchronously. - * - * Not thread safe - can only be called by the thread running the closure. - */ -static inline void set_closure_blocking(struct closure *cl) -{ - if (!closure_blocking(cl)) - atomic_add(CLOSURE_BLOCKING, &cl->remaining); } -/* - * Not thread safe - can only be called by the thread running the closure. - */ -static inline void clear_closure_blocking(struct closure *cl) +static inline void closure_init_stack(struct closure *cl) { - if (closure_blocking(cl)) - atomic_sub(CLOSURE_BLOCKING, &cl->remaining); + memset(cl, 0, sizeof(struct closure)); + atomic_set(&cl->remaining, CLOSURE_REMAINING_INITIALIZER|CLOSURE_STACK); } /** - * closure_wake_up() - wake up all closures on a wait list. + * closure_wake_up - wake up all closures on a wait list. */ static inline void closure_wake_up(struct closure_waitlist *list) { @@ -540,96 +303,19 @@ static inline void closure_wake_up(struct closure_waitlist *list) __closure_wake_up(list); } -/* - * Wait on an event, synchronously or asynchronously - analogous to wait_event() - * but for closures. - * - * The loop is oddly structured so as to avoid a race; we must check the - * condition again after we've added ourself to the waitlist. We know if we were - * already on the waitlist because closure_wait() returns false; thus, we only - * schedule or break if closure_wait() returns false. If it returns true, we - * just loop again - rechecking the condition. - * - * The __closure_wake_up() is necessary because we may race with the event - * becoming true; i.e. we see event false -> wait -> recheck condition, but the - * thread that made the event true may have called closure_wake_up() before we - * added ourself to the wait list. - * - * We have to call closure_sync() at the end instead of just - * __closure_end_sleep() because a different thread might've called - * closure_wake_up() before us and gotten preempted before they dropped the - * refcount on our closure. If this was a stack allocated closure, that would be - * bad. - */ -#define __closure_wait_event(list, cl, condition, _block) \ -({ \ - bool block = _block; \ - typeof(condition) ret; \ - \ - while (1) { \ - ret = (condition); \ - if (ret) { \ - __closure_wake_up(list); \ - if (block) \ - closure_sync(cl); \ - \ - break; \ - } \ - \ - if (block) \ - __closure_start_sleep(cl); \ - \ - if (!closure_wait(list, cl)) { \ - if (!block) \ - break; \ - \ - schedule(); \ - } \ - } \ - \ - ret; \ -}) - /** - * closure_wait_event() - wait on a condition, synchronously or asynchronously. - * @list: the wait list to wait on - * @cl: the closure that is doing the waiting - * @condition: a C expression for the event to wait for - * - * If the closure is in blocking mode, sleeps until the @condition evaluates to - * true - exactly like wait_event(). + * continue_at - jump to another function with barrier * - * If the closure is not in blocking mode, waits asynchronously; if the - * condition is currently false the @cl is put onto @list and returns. @list - * owns a refcount on @cl; closure_sync() or continue_at() may be used later to - * wait for another thread to wake up @list, which drops the refcount on @cl. + * 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). * - * Returns the value of @condition; @cl will be on @list iff @condition was - * false. + * NOTE: This macro expands to a return in the calling function! * - * closure_wake_up(@list) must be called after changing any variable that could - * cause @condition to become true. + * 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. */ -#define closure_wait_event(list, cl, condition) \ - __closure_wait_event(list, cl, condition, closure_blocking(cl)) - -#define closure_wait_event_async(list, cl, condition) \ - __closure_wait_event(list, cl, condition, false) - -#define closure_wait_event_sync(list, cl, condition) \ - __closure_wait_event(list, cl, condition, true) - -static inline void set_closure_fn(struct closure *cl, closure_fn *fn, - struct workqueue_struct *wq) -{ - BUG_ON(object_is_on_stack(cl)); - closure_set_ip(cl); - cl->fn = fn; - cl->wq = wq; - /* between atomic_dec() in closure_put() */ - smp_mb__before_atomic_dec(); -} - #define continue_at(_cl, _fn, _wq) \ do { \ set_closure_fn(_cl, _fn, _wq); \ @@ -637,15 +323,44 @@ do { \ return; \ } 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). + * + * NOTE: like continue_at(), this macro expands to a return in the caller! + * + * 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); \ + closure_queue(_cl); \ return; \ } while (0) +/** + * closure_return - 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); \ @@ -653,6 +368,13 @@ do { \ return; \ } 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) @@ -661,12 +383,4 @@ static inline void closure_call(struct closure *cl, closure_fn fn, continue_at_nobarrier(cl, fn, wq); } -static inline void closure_trylock_call(struct closure *cl, closure_fn fn, - struct workqueue_struct *wq, - struct closure *parent) -{ - if (closure_trylock(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 index 89fd5204924..8b1f1d5c181 100644 --- a/drivers/md/bcache/debug.c +++ b/drivers/md/bcache/debug.c @@ -8,7 +8,7 @@ #include "bcache.h" #include "btree.h" #include "debug.h" -#include "request.h" +#include "extents.h" #include <linux/console.h> #include <linux/debugfs.h> @@ -18,304 +18,130 @@ static struct dentry *debug; -const char *bch_ptr_status(struct cache_set *c, const struct bkey *k) -{ - unsigned 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->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 ""; -} - -struct keyprint_hack bch_pkey(const struct bkey *k) -{ - unsigned i = 0; - struct keyprint_hack r; - char *out = r.s, *end = r.s + KEYHACK_SIZE; - -#define p(...) (out += scnprintf(out, end - out, __VA_ARGS__)) - - p("%llu:%llu len %llu -> [", KEY_INODE(k), KEY_OFFSET(k), KEY_SIZE(k)); - - if (KEY_PTRS(k)) - while (1) { - p("%llu:%llu gen %llu", - PTR_DEV(k, i), PTR_OFFSET(k, i), PTR_GEN(k, i)); - - if (++i == KEY_PTRS(k)) - break; - - p(", "); - } - - p("]"); - - if (KEY_DIRTY(k)) - p(" dirty"); - if (KEY_CSUM(k)) - p(" cs%llu %llx", KEY_CSUM(k), k->ptr[1]); -#undef p - return r; -} - -struct keyprint_hack bch_pbtree(const struct btree *b) -{ - struct keyprint_hack r; - - snprintf(r.s, 40, "%zu level %i/%i", PTR_BUCKET_NR(b->c, &b->key, 0), - b->level, b->c->root ? b->c->root->level : -1); - return r; -} - -#if defined(CONFIG_BCACHE_DEBUG) || defined(CONFIG_BCACHE_EDEBUG) - -static bool skipped_backwards(struct btree *b, struct bkey *k) -{ - return bkey_cmp(k, (!b->level) - ? &START_KEY(bkey_next(k)) - : bkey_next(k)) > 0; -} - -static void dump_bset(struct btree *b, struct bset *i) -{ - struct bkey *k; - unsigned j; - - for (k = i->start; k < end(i); k = bkey_next(k)) { - printk(KERN_ERR "block %zu key %zi/%u: %s", index(i, b), - (uint64_t *) k - i->d, i->keys, pkey(k)); - - for (j = 0; j < KEY_PTRS(k); j++) { - size_t n = PTR_BUCKET_NR(b->c, k, j); - printk(" bucket %zu", n); - - if (n >= b->c->sb.first_bucket && n < b->c->sb.nbuckets) - printk(" prio %i", - PTR_BUCKET(b->c, k, j)->prio); - } - - printk(" %s\n", bch_ptr_status(b->c, k)); - - if (bkey_next(k) < end(i) && - skipped_backwards(b, k)) - printk(KERN_ERR "Key skipped backwards\n"); - } -} - -#endif - #ifdef CONFIG_BCACHE_DEBUG -void bch_btree_verify(struct btree *b, struct bset *new) +#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)) * \ + block_bytes(b->c)) + +void bch_btree_verify(struct btree *b) { struct btree *v = b->c->verify_data; - struct closure cl; - closure_init_stack(&cl); + struct bset *ondisk, *sorted, *inmemory; + struct bio *bio; - if (!b->c->verify) + if (!b->c->verify || !b->c->verify_ondisk) return; - closure_wait_event(&b->io.wait, &cl, - atomic_read(&b->io.cl.remaining) == -1); - + 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; - bch_btree_read(v); - closure_wait_event(&v->io.wait, &cl, - atomic_read(&b->io.cl.remaining) == -1); + bio = bch_bbio_alloc(b->c); + bio->bi_bdev = 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; + bch_bio_map(bio, sorted); - if (new->keys != v->sets[0].data->keys || - memcmp(new->start, - v->sets[0].data->start, - (void *) end(new) - (void *) new->start)) { - unsigned i, j; + submit_bio_wait(REQ_META|READ_SYNC, 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 j; console_lock(); - printk(KERN_ERR "*** original memory node:\n"); - for (i = 0; i <= b->nsets; i++) - dump_bset(b, b->sets[i].data); + printk(KERN_ERR "*** in memory:\n"); + bch_dump_bset(&b->keys, inmemory, 0); - printk(KERN_ERR "*** sorted memory node:\n"); - dump_bset(b, new); + printk(KERN_ERR "*** read back in:\n"); + bch_dump_bset(&v->keys, sorted, 0); - printk(KERN_ERR "*** on disk node:\n"); - dump_bset(v, v->sets[0].data); + for_each_written_bset(b, ondisk, i) { + unsigned block = ((void *) i - (void *) ondisk) / + block_bytes(b->c); - for (j = 0; j < new->keys; j++) - if (new->d[j] != v->sets[0].data->d[j]) + printk(KERN_ERR "*** on disk block %u:\n", block); + bch_dump_bset(&b->keys, i, block); + } + + printk(KERN_ERR "*** block %zu not written\n", + ((void *) i - (void *) ondisk) / block_bytes(b->c)); + + for (j = 0; j < inmemory->keys; j++) + if (inmemory->d[j] != sorted->d[j]) break; + printk(KERN_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); } -static void data_verify_endio(struct bio *bio, int error) -{ - struct closure *cl = bio->bi_private; - closure_put(cl); -} - -void bch_data_verify(struct search *s) +void bch_data_verify(struct cached_dev *dc, struct bio *bio) { char name[BDEVNAME_SIZE]; - struct cached_dev *dc = container_of(s->d, struct cached_dev, disk); - struct closure *cl = &s->cl; struct bio *check; - struct bio_vec *bv; + struct bio_vec bv, *bv2; + struct bvec_iter iter; int i; - if (!s->unaligned_bvec) - bio_for_each_segment(bv, s->orig_bio, i) - bv->bv_offset = 0, bv->bv_len = PAGE_SIZE; - - check = bio_clone(s->orig_bio, GFP_NOIO); + check = bio_clone(bio, GFP_NOIO); if (!check) return; - if (bch_bio_alloc_pages(check, GFP_NOIO)) + if (bio_alloc_pages(check, GFP_NOIO)) goto out_put; - check->bi_rw = READ_SYNC; - check->bi_private = cl; - check->bi_end_io = data_verify_endio; + submit_bio_wait(READ_SYNC, check); - closure_bio_submit(check, cl, &dc->disk); - closure_sync(cl); + bio_for_each_segment(bv, bio, iter) { + void *p1 = kmap_atomic(bv.bv_page); + void *p2 = page_address(check->bi_io_vec[iter.bi_idx].bv_page); - bio_for_each_segment(bv, s->orig_bio, i) { - void *p1 = kmap(bv->bv_page); - void *p2 = kmap(check->bi_io_vec[i].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", + bdevname(dc->bdev, name), + (uint64_t) bio->bi_iter.bi_sector); - if (memcmp(p1 + bv->bv_offset, - p2 + bv->bv_offset, - bv->bv_len)) - printk(KERN_ERR - "bcache (%s): verify failed at sector %llu\n", - bdevname(dc->bdev, name), - (uint64_t) s->orig_bio->bi_sector); - - kunmap(bv->bv_page); - kunmap(check->bi_io_vec[i].bv_page); + kunmap_atomic(p1); } - __bio_for_each_segment(bv, check, i, 0) - __free_page(bv->bv_page); + bio_for_each_segment_all(bv2, check, i) + __free_page(bv2->bv_page); out_put: bio_put(check); } #endif -#ifdef CONFIG_BCACHE_EDEBUG - -unsigned bch_count_data(struct btree *b) -{ - unsigned ret = 0; - struct btree_iter iter; - struct bkey *k; - - if (!b->level) - for_each_key(b, k, &iter) - ret += KEY_SIZE(k); - return ret; -} - -static void vdump_bucket_and_panic(struct btree *b, const char *fmt, - va_list args) -{ - unsigned i; - - console_lock(); - - for (i = 0; i <= b->nsets; i++) - dump_bset(b, b->sets[i].data); - - vprintk(fmt, args); - - console_unlock(); - - panic("at %s\n", pbtree(b)); -} - -void bch_check_key_order_msg(struct btree *b, struct bset *i, - const char *fmt, ...) -{ - struct bkey *k; - - if (!i->keys) - return; - - for (k = i->start; bkey_next(k) < end(i); k = bkey_next(k)) - if (skipped_backwards(b, k)) { - va_list args; - va_start(args, fmt); - - vdump_bucket_and_panic(b, fmt, args); - va_end(args); - } -} - -void bch_check_keys(struct btree *b, const char *fmt, ...) -{ - va_list args; - struct bkey *k, *p = NULL; - struct btree_iter iter; - - if (b->level) - return; - - for_each_key(b, k, &iter) { - if (p && bkey_cmp(&START_KEY(p), &START_KEY(k)) > 0) { - printk(KERN_ERR "Keys out of order:\n"); - goto bug; - } - - if (bch_ptr_invalid(b, k)) - continue; - - if (p && bkey_cmp(p, &START_KEY(k)) > 0) { - printk(KERN_ERR "Overlapping keys:\n"); - goto bug; - } - p = k; - } - return; -bug: - va_start(args, fmt); - vdump_bucket_and_panic(b, fmt, args); - va_end(args); -} - -#endif - #ifdef CONFIG_DEBUG_FS /* XXX: cache set refcounting */ @@ -337,6 +163,7 @@ static ssize_t bch_dump_read(struct file *file, char __user *buf, { struct dump_iterator *i = file->private_data; ssize_t ret = 0; + char kbuf[80]; while (size) { struct keybuf_key *w; @@ -355,11 +182,12 @@ static ssize_t bch_dump_read(struct file *file, char __user *buf, if (i->bytes) break; - w = bch_keybuf_next_rescan(i->c, &i->keys, &MAX_KEY); + w = bch_keybuf_next_rescan(i->c, &i->keys, &MAX_KEY, dump_pred); if (!w) break; - i->bytes = snprintf(i->buf, PAGE_SIZE, "%s\n", pkey(&w->key)); + 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); } @@ -377,7 +205,7 @@ static int bch_dump_open(struct inode *inode, struct file *file) file->private_data = i; i->c = c; - bch_keybuf_init(&i->keys, dump_pred); + bch_keybuf_init(&i->keys); i->keys.last_scanned = KEY(0, 0, 0); return 0; @@ -409,142 +237,6 @@ void bch_debug_init_cache_set(struct cache_set *c) #endif -/* Fuzz tester has rotted: */ -#if 0 - -static ssize_t btree_fuzz(struct kobject *k, struct kobj_attribute *a, - const char *buffer, size_t size) -{ - void dump(struct btree *b) - { - struct bset *i; - - for (i = b->sets[0].data; - index(i, b) < btree_blocks(b) && - i->seq == b->sets[0].data->seq; - i = ((void *) i) + set_blocks(i, b->c) * block_bytes(b->c)) - dump_bset(b, i); - } - - struct cache_sb *sb; - struct cache_set *c; - struct btree *all[3], *b, *fill, *orig; - int j; - - struct btree_op op; - bch_btree_op_init_stack(&op); - - sb = kzalloc(sizeof(struct cache_sb), GFP_KERNEL); - if (!sb) - return -ENOMEM; - - sb->bucket_size = 128; - sb->block_size = 4; - - c = bch_cache_set_alloc(sb); - if (!c) - return -ENOMEM; - - for (j = 0; j < 3; j++) { - BUG_ON(list_empty(&c->btree_cache)); - all[j] = list_first_entry(&c->btree_cache, struct btree, list); - list_del_init(&all[j]->list); - - all[j]->key = KEY(0, 0, c->sb.bucket_size); - bkey_copy_key(&all[j]->key, &MAX_KEY); - } - - b = all[0]; - fill = all[1]; - orig = all[2]; - - while (1) { - for (j = 0; j < 3; j++) - all[j]->written = all[j]->nsets = 0; - - bch_bset_init_next(b); - - while (1) { - struct bset *i = write_block(b); - struct bkey *k = op.keys.top; - unsigned rand; - - bkey_init(k); - rand = get_random_int(); - - op.type = rand & 1 - ? BTREE_INSERT - : BTREE_REPLACE; - rand >>= 1; - - SET_KEY_SIZE(k, bucket_remainder(c, rand)); - rand >>= c->bucket_bits; - rand &= 1024 * 512 - 1; - rand += c->sb.bucket_size; - SET_KEY_OFFSET(k, rand); -#if 0 - SET_KEY_PTRS(k, 1); -#endif - bch_keylist_push(&op.keys); - bch_btree_insert_keys(b, &op); - - if (should_split(b) || - set_blocks(i, b->c) != - __set_blocks(i, i->keys + 15, b->c)) { - i->csum = csum_set(i); - - memcpy(write_block(fill), - i, set_bytes(i)); - - b->written += set_blocks(i, b->c); - fill->written = b->written; - if (b->written == btree_blocks(b)) - break; - - bch_btree_sort_lazy(b); - bch_bset_init_next(b); - } - } - - memcpy(orig->sets[0].data, - fill->sets[0].data, - btree_bytes(c)); - - bch_btree_sort(b); - fill->written = 0; - bch_btree_read_done(&fill->io.cl); - - if (b->sets[0].data->keys != fill->sets[0].data->keys || - memcmp(b->sets[0].data->start, - fill->sets[0].data->start, - b->sets[0].data->keys * sizeof(uint64_t))) { - struct bset *i = b->sets[0].data; - struct bkey *k, *l; - - for (k = i->start, - l = fill->sets[0].data->start; - k < end(i); - k = bkey_next(k), l = bkey_next(l)) - if (bkey_cmp(k, l) || - KEY_SIZE(k) != KEY_SIZE(l)) - pr_err("key %zi differs: %s != %s", - (uint64_t *) k - i->d, - pkey(k), pkey(l)); - - for (j = 0; j < 3; j++) { - pr_err("**** Set %i ****", j); - dump(all[j]); - } - panic("\n"); - } - - pr_info("fuzz complete: %i keys", b->sets[0].data->keys); - } -} - -kobj_attribute_write(fuzz, btree_fuzz); -#endif - void bch_debug_exit(void) { if (!IS_ERR_OR_NULL(debug)) @@ -554,11 +246,6 @@ void bch_debug_exit(void) int __init bch_debug_init(struct kobject *kobj) { int ret = 0; -#if 0 - ret = sysfs_create_file(kobj, &ksysfs_fuzz.attr); - if (ret) - return ret; -#endif debug = debugfs_create_dir("bcache", NULL); return ret; diff --git a/drivers/md/bcache/debug.h b/drivers/md/bcache/debug.h index f9378a21814..1f63c195d24 100644 --- a/drivers/md/bcache/debug.h +++ b/drivers/md/bcache/debug.h @@ -1,47 +1,27 @@ #ifndef _BCACHE_DEBUG_H #define _BCACHE_DEBUG_H -/* Btree/bkey debug printing */ - -#define KEYHACK_SIZE 80 -struct keyprint_hack { - char s[KEYHACK_SIZE]; -}; - -struct keyprint_hack bch_pkey(const struct bkey *k); -struct keyprint_hack bch_pbtree(const struct btree *b); -#define pkey(k) (&bch_pkey(k).s[0]) -#define pbtree(b) (&bch_pbtree(b).s[0]) - -#ifdef CONFIG_BCACHE_EDEBUG - -unsigned bch_count_data(struct btree *); -void bch_check_key_order_msg(struct btree *, struct bset *, const char *, ...); -void bch_check_keys(struct btree *, const char *, ...); - -#define bch_check_key_order(b, i) \ - bch_check_key_order_msg(b, i, "keys out of order") -#define EBUG_ON(cond) BUG_ON(cond) - -#else /* EDEBUG */ - -#define bch_count_data(b) 0 -#define bch_check_key_order(b, i) do {} while (0) -#define bch_check_key_order_msg(b, i, ...) do {} while (0) -#define bch_check_keys(b, ...) do {} while (0) -#define EBUG_ON(cond) do {} while (0) - -#endif +struct bio; +struct cached_dev; +struct cache_set; #ifdef CONFIG_BCACHE_DEBUG -void bch_btree_verify(struct btree *, struct bset *); -void bch_data_verify(struct search *); +void bch_btree_verify(struct btree *); +void bch_data_verify(struct cached_dev *, struct 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, struct bset *i) {} -static inline void bch_data_verify(struct search *s) {}; +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 diff --git a/drivers/md/bcache/extents.c b/drivers/md/bcache/extents.c new file mode 100644 index 00000000000..3a0de4cf977 --- /dev/null +++ b/drivers/md/bcache/extents.c @@ -0,0 +1,620 @@ +/* + * 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/bcache.txt. + */ + +#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 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->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 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->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 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 j; + char buf[80]; + + bch_extent_to_text(buf, sizeof(buf), k); + printk(" %s", buf); + + for (j = 0; j < KEY_PTRS(k); j++) { + size_t n = PTR_BUCKET_NR(b->c, k, j); + printk(" bucket %zu", n); + + if (n >= b->c->sb.first_bucket && n < b->c->sb.nbuckets) + printk(" prio %i", + PTR_BUCKET(b->c, k, j)->prio); + } + + printk(" %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 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 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 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 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; +} + +static bool bch_extent_invalid(struct btree_keys *bk, const struct bkey *k) +{ + struct btree *b = container_of(bk, struct btree, keys); + char buf[80]; + + if (!KEY_SIZE(k)) + return true; + + if (KEY_SIZE(k) > KEY_OFFSET(k)) + goto bad; + + if (__ptr_invalid(b->c, k)) + goto bad; + + return false; +bad: + bch_extent_to_text(buf, sizeof(buf), k); + cache_bug(b->c, "spotted extent %s: %s", buf, bch_ptr_status(b->c, k)); + return true; +} + +static bool bch_extent_bad_expensive(struct btree *b, const struct bkey *k, + unsigned 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); + struct bucket *g; + unsigned i, stale; + + 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; + + if (!expensive_debug_checks(b->c) && KEY_DIRTY(k)) + return false; + + for (i = 0; i < KEY_PTRS(k); i++) { + g = PTR_BUCKET(b->c, k, i); + stale = ptr_stale(b->c, k, i); + + btree_bug_on(stale > 96, b, + "key too stale: %i, need_gc %u", + stale, b->c->need_gc); + + btree_bug_on(stale && KEY_DIRTY(k) && KEY_SIZE(k), + b, "stale dirty pointer"); + + 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 i; + + if (key_merging_disabled(b->c)) + return false; + + for (i = 0; i < KEY_PTRS(l); i++) + if (l->ptr[i] + 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 00000000000..e4e23409782 --- /dev/null +++ b/drivers/md/bcache/extents.h @@ -0,0 +1,13 @@ +#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 *, size_t, const struct bkey *); +bool __bch_btree_ptr_invalid(struct cache_set *, const struct bkey *); + +#endif /* _BCACHE_EXTENTS_H */ diff --git a/drivers/md/bcache/io.c b/drivers/md/bcache/io.c index 48efd4dea64..fa028fa82df 100644 --- a/drivers/md/bcache/io.c +++ b/drivers/md/bcache/io.c @@ -9,187 +9,42 @@ #include "bset.h" #include "debug.h" -static void bch_bi_idx_hack_endio(struct bio *bio, int error) -{ - struct bio *p = bio->bi_private; - - bio_endio(p, error); - bio_put(bio); -} - -static void bch_generic_make_request_hack(struct bio *bio) -{ - if (bio->bi_idx) { - struct bio *clone = bio_alloc(GFP_NOIO, bio_segments(bio)); - - memcpy(clone->bi_io_vec, - bio_iovec(bio), - bio_segments(bio) * sizeof(struct bio_vec)); - - clone->bi_sector = bio->bi_sector; - clone->bi_bdev = bio->bi_bdev; - clone->bi_rw = bio->bi_rw; - clone->bi_vcnt = bio_segments(bio); - clone->bi_size = bio->bi_size; - - clone->bi_private = bio; - clone->bi_end_io = bch_bi_idx_hack_endio; - - bio = clone; - } - - /* - * Hack, since drivers that clone bios clone up to bi_max_vecs, but our - * bios might have had more than that (before we split them per device - * limitations). - * - * To be taken out once immutable bvec stuff is in. - */ - bio->bi_max_vecs = bio->bi_vcnt; - - generic_make_request(bio); -} - -/** - * bch_bio_split - split a bio - * @bio: bio to split - * @sectors: number of sectors to split from the front of @bio - * @gfp: gfp mask - * @bs: bio set to allocate from - * - * Allocates and returns a new bio which represents @sectors from the start of - * @bio, and updates @bio to represent the remaining sectors. - * - * If bio_sectors(@bio) was less than or equal to @sectors, returns @bio - * unchanged. - * - * The newly allocated bio will point to @bio's bi_io_vec, if the split was on a - * bvec boundry; it is the caller's responsibility to ensure that @bio is not - * freed before the split. - * - * If bch_bio_split() is running under generic_make_request(), it's not safe to - * allocate more than one bio from the same bio set. Therefore, if it is running - * under generic_make_request() it masks out __GFP_WAIT when doing the - * allocation. The caller must check for failure if there's any possibility of - * it being called from under generic_make_request(); it is then the caller's - * responsibility to retry from a safe context (by e.g. punting to workqueue). - */ -struct bio *bch_bio_split(struct bio *bio, int sectors, - gfp_t gfp, struct bio_set *bs) -{ - unsigned idx = bio->bi_idx, vcnt = 0, nbytes = sectors << 9; - struct bio_vec *bv; - struct bio *ret = NULL; - - BUG_ON(sectors <= 0); - - /* - * If we're being called from underneath generic_make_request() and we - * already allocated any bios from this bio set, we risk deadlock if we - * use the mempool. So instead, we possibly fail and let the caller punt - * to workqueue or somesuch and retry in a safe context. - */ - if (current->bio_list) - gfp &= ~__GFP_WAIT; - - if (sectors >= bio_sectors(bio)) - return bio; - - if (bio->bi_rw & REQ_DISCARD) { - ret = bio_alloc_bioset(gfp, 1, bs); - idx = 0; - goto out; - } - - bio_for_each_segment(bv, bio, idx) { - vcnt = idx - bio->bi_idx; - - if (!nbytes) { - ret = bio_alloc_bioset(gfp, vcnt, bs); - if (!ret) - return NULL; - - memcpy(ret->bi_io_vec, bio_iovec(bio), - sizeof(struct bio_vec) * vcnt); - - break; - } else if (nbytes < bv->bv_len) { - ret = bio_alloc_bioset(gfp, ++vcnt, bs); - if (!ret) - return NULL; - - memcpy(ret->bi_io_vec, bio_iovec(bio), - sizeof(struct bio_vec) * vcnt); - - ret->bi_io_vec[vcnt - 1].bv_len = nbytes; - bv->bv_offset += nbytes; - bv->bv_len -= nbytes; - break; - } - - nbytes -= bv->bv_len; - } -out: - ret->bi_bdev = bio->bi_bdev; - ret->bi_sector = bio->bi_sector; - ret->bi_size = sectors << 9; - ret->bi_rw = bio->bi_rw; - ret->bi_vcnt = vcnt; - ret->bi_max_vecs = vcnt; - - bio->bi_sector += sectors; - bio->bi_size -= sectors << 9; - bio->bi_idx = idx; - - if (bio_integrity(bio)) { - if (bio_integrity_clone(ret, bio, gfp)) { - bio_put(ret); - return NULL; - } - - bio_integrity_trim(ret, 0, bio_sectors(ret)); - bio_integrity_trim(bio, bio_sectors(ret), bio_sectors(bio)); - } - - return ret; -} +#include <linux/blkdev.h> static unsigned bch_bio_max_sectors(struct bio *bio) { - unsigned ret = bio_sectors(bio); struct request_queue *q = bdev_get_queue(bio->bi_bdev); - unsigned max_segments = min_t(unsigned, BIO_MAX_PAGES, - queue_max_segments(q)); - struct bio_vec *bv, *end = bio_iovec(bio) + - min_t(int, bio_segments(bio), max_segments); + struct bio_vec bv; + struct bvec_iter iter; + unsigned ret = 0, seg = 0; if (bio->bi_rw & REQ_DISCARD) - return min(ret, q->limits.max_discard_sectors); - - if (bio_segments(bio) > max_segments || - q->merge_bvec_fn) { - ret = 0; - - for (bv = bio_iovec(bio); bv < end; bv++) { - struct bvec_merge_data bvm = { - .bi_bdev = bio->bi_bdev, - .bi_sector = bio->bi_sector, - .bi_size = ret << 9, - .bi_rw = bio->bi_rw, - }; + return min(bio_sectors(bio), q->limits.max_discard_sectors); + + bio_for_each_segment(bv, bio, iter) { + struct bvec_merge_data bvm = { + .bi_bdev = bio->bi_bdev, + .bi_sector = bio->bi_iter.bi_sector, + .bi_size = ret << 9, + .bi_rw = bio->bi_rw, + }; + + if (seg == min_t(unsigned, BIO_MAX_PAGES, + queue_max_segments(q))) + break; - if (q->merge_bvec_fn && - q->merge_bvec_fn(q, &bvm, bv) < (int) bv->bv_len) - break; + if (q->merge_bvec_fn && + q->merge_bvec_fn(q, &bvm, &bv) < (int) bv.bv_len) + break; - ret += bv->bv_len >> 9; - } + seg++; + ret += bv.bv_len >> 9; } ret = min(ret, queue_max_sectors(q)); WARN_ON(!ret); - ret = max_t(int, ret, bio_iovec(bio)->bv_len >> 9); + ret = max_t(int, ret, bio_iovec(bio).bv_len >> 9); return ret; } @@ -200,7 +55,7 @@ static void bch_bio_submit_split_done(struct closure *cl) s->bio->bi_end_io = s->bi_end_io; s->bio->bi_private = s->bi_private; - bio_endio(s->bio, 0); + bio_endio_nodec(s->bio, 0); closure_debug_destroy(&s->cl); mempool_free(s, s->p->bio_split_hook); @@ -218,30 +73,10 @@ static void bch_bio_submit_split_endio(struct bio *bio, int error) closure_put(cl); } -static void __bch_bio_submit_split(struct closure *cl) -{ - struct bio_split_hook *s = container_of(cl, struct bio_split_hook, cl); - struct bio *bio = s->bio, *n; - - do { - n = bch_bio_split(bio, bch_bio_max_sectors(bio), - GFP_NOIO, s->p->bio_split); - if (!n) - continue_at(cl, __bch_bio_submit_split, system_wq); - - n->bi_end_io = bch_bio_submit_split_endio; - n->bi_private = cl; - - closure_get(cl); - bch_generic_make_request_hack(n); - } while (n != bio); - - continue_at(cl, bch_bio_submit_split_done, NULL); -} - void bch_generic_make_request(struct bio *bio, struct bio_split_pool *p) { struct bio_split_hook *s; + struct bio *n; if (!bio_has_data(bio) && !(bio->bi_rw & REQ_DISCARD)) goto submit; @@ -250,6 +85,7 @@ void bch_generic_make_request(struct bio *bio, struct bio_split_pool *p) goto submit; s = mempool_alloc(p->bio_split_hook, GFP_NOIO); + closure_init(&s->cl, NULL); s->bio = bio; s->p = p; @@ -257,10 +93,20 @@ void bch_generic_make_request(struct bio *bio, struct bio_split_pool *p) s->bi_private = bio->bi_private; bio_get(bio); - closure_call(&s->cl, __bch_bio_submit_split, NULL, NULL); - return; + do { + n = bio_next_split(bio, bch_bio_max_sectors(bio), + GFP_NOIO, s->p->bio_split); + + n->bi_end_io = bch_bio_submit_split_endio; + n->bi_private = &s->cl; + + closure_get(&s->cl); + generic_make_request(n); + } while (n != bio); + + continue_at(&s->cl, bch_bio_submit_split_done, NULL); submit: - bch_generic_make_request_hack(bio); + generic_make_request(bio); } /* Bios with headers */ @@ -288,8 +134,8 @@ void __bch_submit_bbio(struct bio *bio, struct cache_set *c) { struct bbio *b = container_of(bio, struct bbio, bio); - bio->bi_sector = PTR_OFFSET(&b->key, 0); - bio->bi_bdev = PTR_CACHE(c, &b->key, 0)->bdev; + bio->bi_iter.bi_sector = PTR_OFFSET(&b->key, 0); + bio->bi_bdev = PTR_CACHE(c, &b->key, 0)->bdev; b->submit_time_us = local_clock_us(); closure_bio_submit(bio, bio->bi_private, PTR_CACHE(c, &b->key, 0)); diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c index 8c8dfdcd9d4..59e82021b5b 100644 --- a/drivers/md/bcache/journal.c +++ b/drivers/md/bcache/journal.c @@ -7,7 +7,8 @@ #include "bcache.h" #include "btree.h" #include "debug.h" -#include "request.h" + +#include <trace/events/bcache.h> /* * Journal replay/recovery: @@ -29,35 +30,38 @@ static void journal_read_endio(struct bio *bio, int error) } static int journal_read_bucket(struct cache *ca, struct list_head *list, - struct btree_op *op, unsigned bucket_index) + unsigned 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 len, left, offset = 0; int ret = 0; sector_t bucket = bucket_to_sector(ca->set, ca->sb.d[bucket_index]); - pr_debug("reading %llu", (uint64_t) bucket); + closure_init_stack(&cl); + + pr_debug("reading %u", bucket_index); while (offset < ca->sb.bucket_size) { reread: left = ca->sb.bucket_size - offset; - len = min_t(unsigned, left, PAGE_SECTORS * 8); + len = min_t(unsigned, left, PAGE_SECTORS << JSET_BITS); bio_reset(bio); - bio->bi_sector = bucket + offset; + bio->bi_iter.bi_sector = bucket + offset; bio->bi_bdev = ca->bdev; bio->bi_rw = READ; - bio->bi_size = len << 9; + bio->bi_iter.bi_size = len << 9; bio->bi_end_io = journal_read_endio; - bio->bi_private = &op->cl; + bio->bi_private = &cl; bch_bio_map(bio, data); - closure_bio_submit(bio, &op->cl, ca); - closure_sync(&op->cl); + closure_bio_submit(bio, &cl, ca); + closure_sync(&cl); /* This function could be simpler now since we no longer write * journal entries that overlap bucket boundaries; this means @@ -70,19 +74,28 @@ reread: left = ca->sb.bucket_size - offset; struct list_head *where; size_t blocks, bytes = set_bytes(j); - if (j->magic != jset_magic(ca->set)) + if (j->magic != jset_magic(&ca->sb)) { + pr_debug("%u: bad magic", bucket_index); return ret; + } - if (bytes > left << 9) + if (bytes > left << 9 || + bytes > PAGE_SIZE << JSET_BITS) { + pr_info("%u: too big, %zu bytes, offset %u", + bucket_index, bytes, offset); return ret; + } if (bytes > len << 9) goto reread; - if (j->csum != csum_set(j)) + if (j->csum != csum_set(j)) { + pr_info("%u: bad csum, %zu bytes, offset %u", + bucket_index, bytes, offset); return ret; + } - blocks = set_blocks(j, ca->set); + blocks = set_blocks(j, block_bytes(ca->set)); while (!list_empty(list)) { i = list_first_entry(list, @@ -127,12 +140,11 @@ next_set: return ret; } -int bch_journal_read(struct cache_set *c, struct list_head *list, - struct btree_op *op) +int bch_journal_read(struct cache_set *c, struct list_head *list) { #define read_bucket(b) \ ({ \ - int ret = journal_read_bucket(ca, list, op, b); \ + int ret = journal_read_bucket(ca, list, b); \ __set_bit(b, bitmap); \ if (ret < 0) \ return ret; \ @@ -151,7 +163,8 @@ int bch_journal_read(struct cache_set *c, struct list_head *list, bitmap_zero(bitmap, SB_JOURNAL_BUCKETS); pr_debug("%u journal buckets", ca->sb.njournal_buckets); - /* Read journal buckets ordered by golden ratio hash to quickly + /* + * 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++) { @@ -164,36 +177,45 @@ int bch_journal_read(struct cache_set *c, struct list_head *list, goto bsearch; } - /* If that fails, check all the buckets we haven't checked + /* + * If that fails, check all the buckets we haven't checked * already */ pr_debug("falling back to linear search"); - for (l = 0; l < ca->sb.njournal_buckets; l++) { - if (test_bit(l, bitmap)) - continue; - + for (l = find_first_zero_bit(bitmap, ca->sb.njournal_buckets); + l < ca->sb.njournal_buckets; + l = find_next_zero_bit(bitmap, ca->sb.njournal_buckets, l + 1)) if (read_bucket(l)) goto bsearch; - } + + if (list_empty(list)) + continue; bsearch: /* Binary search */ m = r = find_next_bit(bitmap, ca->sb.njournal_buckets, l + 1); pr_debug("starting binary search, l %u r %u", 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 (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 + /* + * Read buckets in reverse order until we stop finding more * journal entries */ - pr_debug("finishing up"); + pr_debug("finishing up: m %u njournal_buckets %u", + m, ca->sb.njournal_buckets); l = m; while (1) { @@ -215,15 +237,22 @@ bsearch: for (i = 0; i < ca->sb.njournal_buckets; i++) if (ja->seq[i] > seq) { seq = ja->seq[i]; - ja->cur_idx = ja->discard_idx = - ja->last_idx = 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; } } - c->journal.seq = list_entry(list->prev, - struct journal_replay, - list)->j.seq; + if (!list_empty(list)) + c->journal.seq = list_entry(list->prev, + struct journal_replay, + list)->j.seq; return 0; #undef read_bucket @@ -261,26 +290,20 @@ void bch_journal_mark(struct cache_set *c, struct list_head *list) } for (k = i->j.start; - k < end(&i->j); + k < bset_bkey_last(&i->j); k = bkey_next(k)) { unsigned j; - for (j = 0; j < KEY_PTRS(k); j++) { - struct bucket *g = PTR_BUCKET(c, k, j); - atomic_inc(&g->pin); + for (j = 0; j < KEY_PTRS(k); j++) + if (ptr_available(c, k, j)) + atomic_inc(&PTR_BUCKET(c, k, j)->pin); - if (g->prio == BTREE_PRIO && - !ptr_stale(c, k, j)) - g->prio = INITIAL_PRIO; - } - - __bch_btree_mark_key(c, 0, k); + bch_initial_mark_key(c, 0, k); } } } -int bch_journal_replay(struct cache_set *s, struct list_head *list, - struct btree_op *op) +int bch_journal_replay(struct cache_set *s, struct list_head *list) { int ret = 0, keys = 0, entries = 0; struct bkey *k; @@ -288,30 +311,27 @@ int bch_journal_replay(struct cache_set *s, struct list_head *list, 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) - pr_err( - "journal entries %llu-%llu missing! (replaying %llu-%llu)\n", - n, i->j.seq - 1, start, end); + cache_set_err_on(n != i->j.seq, s, +"bcache: journal entries %llu-%llu missing! (replaying %llu-%llu)", + n, i->j.seq - 1, start, end); for (k = i->j.start; - k < end(&i->j); + k < bset_bkey_last(&i->j); k = bkey_next(k)) { - pr_debug("%s", pkey(k)); - bkey_copy(op->keys.top, k); - bch_keylist_push(&op->keys); + trace_bcache_journal_replay_key(k); - op->journal = i->pin; - atomic_inc(op->journal); + bch_keylist_init_single(&keylist, k); - ret = bch_btree_insert(op, s); + ret = bch_btree_insert(s, &keylist, i->pin, NULL); if (ret) goto err; - BUG_ON(!bch_keylist_empty(&op->keys)); + BUG_ON(!bch_keylist_empty(&keylist)); keys++; cond_resched(); @@ -325,14 +345,13 @@ int bch_journal_replay(struct cache_set *s, struct list_head *list, pr_info("journal replay done, %i keys in %i entries, seq %llu", keys, entries, end); - +err: while (!list_empty(list)) { i = list_first_entry(list, struct journal_replay, list); list_del(&i->list); kfree(i); } -err: - closure_sync(&op->cl); + return ret; } @@ -344,48 +363,34 @@ static void btree_flush_write(struct cache_set *c) * Try to find the btree node with that references the oldest journal * entry, best is our current candidate and is locked if non NULL: */ - struct btree *b, *best = NULL; - unsigned iter; - - for_each_cached_btree(b, c, iter) { - if (!down_write_trylock(&b->lock)) - continue; - - if (!btree_node_dirty(b) || - !btree_current_write(b)->journal) { - rw_unlock(true, b); - continue; + struct btree *b, *best; + unsigned i; +retry: + best = NULL; + + for_each_cached_btree(b, c, i) + if (btree_current_write(b)->journal) { + if (!best) + best = b; + else if (journal_pin_cmp(c, + btree_current_write(best)->journal, + btree_current_write(b)->journal)) { + best = b; + } } - if (!best) - best = b; - else if (journal_pin_cmp(c, - btree_current_write(best), - btree_current_write(b))) { - rw_unlock(true, best); - best = b; - } else - rw_unlock(true, b); - } - - if (best) - goto out; - - /* We can't find the best btree node, just pick the first */ - list_for_each_entry(b, &c->btree_cache, list) - if (!b->level && btree_node_dirty(b)) { - best = b; - rw_lock(true, best, best->level); - goto found; + b = best; + if (b) { + mutex_lock(&b->write_lock); + if (!btree_current_write(b)->journal) { + mutex_unlock(&b->write_lock); + /* We raced */ + goto retry; } -out: - if (!best) - return; -found: - if (btree_node_dirty(best)) - bch_btree_write(best, true, NULL); - rw_unlock(true, best); + __bch_btree_node_write(b, NULL); + mutex_unlock(&b->write_lock); + } } #define last_seq(j) ((j)->seq - fifo_used(&(j)->pin) + 1) @@ -420,7 +425,7 @@ static void do_journal_discard(struct cache *ca) return; } - switch (atomic_read(&ja->discard_in_flight) == DISCARD_IN_FLIGHT) { + switch (atomic_read(&ja->discard_in_flight)) { case DISCARD_IN_FLIGHT: return; @@ -438,13 +443,13 @@ static void do_journal_discard(struct cache *ca) atomic_set(&ja->discard_in_flight, DISCARD_IN_FLIGHT); bio_init(bio); - bio->bi_sector = bucket_to_sector(ca->set, + bio->bi_iter.bi_sector = bucket_to_sector(ca->set, ca->sb.d[ja->discard_idx]); bio->bi_bdev = ca->bdev; bio->bi_rw = REQ_WRITE|REQ_DISCARD; bio->bi_max_vecs = 1; bio->bi_io_vec = bio->bi_inline_vecs; - bio->bi_size = bucket_bytes(ca); + bio->bi_iter.bi_size = bucket_bytes(ca); bio->bi_end_io = journal_discard_endio; closure_get(&ca->set->cl); @@ -481,7 +486,7 @@ static void journal_reclaim(struct cache_set *c) do_journal_discard(ca); if (c->journal.blocks_free) - return; + goto out; /* * Allocate: @@ -507,7 +512,7 @@ static void journal_reclaim(struct cache_set *c) if (n) c->journal.blocks_free = c->sb.bucket_size >> c->block_bits; - +out: if (!journal_full(&c->journal)) __closure_wake_up(&c->journal.wait); } @@ -528,6 +533,7 @@ void bch_journal_next(struct journal *j) 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; @@ -540,51 +546,46 @@ static void journal_write_endio(struct bio *bio, int error) struct journal_write *w = bio->bi_private; cache_set_err_on(error, w->c, "journal io error"); - closure_put(&w->c->journal.io.cl); + closure_put(&w->c->journal.io); } static void journal_write(struct closure *); static void journal_write_done(struct closure *cl) { - struct journal *j = container_of(cl, struct journal, io.cl); - struct cache_set *c = container_of(j, struct cache_set, journal); - + 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, system_wq); +} - if (c->journal_delay_ms) - closure_delay(&j->io, msecs_to_jiffies(c->journal_delay_ms)); +static void journal_write_unlock(struct closure *cl) +{ + struct cache_set *c = container_of(cl, struct cache_set, journal.io); - continue_at(cl, journal_write, system_wq); + 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.cl); + struct cache_set *c = container_of(cl, struct cache_set, journal.io); struct cache *ca; struct journal_write *w = c->journal.cur; struct bkey *k = &c->journal.key; - unsigned i, sectors = set_blocks(w->data, c) * c->sb.block_size; + unsigned i, sectors = set_blocks(w->data, block_bytes(c)) * + c->sb.block_size; struct bio *bio; struct bio_list list; bio_list_init(&list); if (!w->need_write) { - /* - * XXX: have to unlock closure before we unlock journal lock, - * else we race with bch_journal(). But this way we race - * against cache set unregister. Doh. - */ - set_closure_fn(cl, NULL, NULL); - closure_sub(cl, CLOSURE_RUNNING + 1); - spin_unlock(&c->journal.lock); - return; + closure_return_with_destructor(cl, journal_write_unlock); } else if (journal_full(&c->journal)) { journal_reclaim(c); spin_unlock(&c->journal.lock); @@ -593,7 +594,7 @@ static void journal_write_unlocked(struct closure *cl) continue_at(cl, journal_write, system_wq); } - c->journal.blocks_free -= set_blocks(w->data, c); + c->journal.blocks_free -= set_blocks(w->data, block_bytes(c)); w->data->btree_level = c->root->level; @@ -603,7 +604,7 @@ static void journal_write_unlocked(struct closure *cl) for_each_cache(ca, c, i) w->data->prio_bucket[ca->sb.nr_this_dev] = ca->prio_buckets[0]; - w->data->magic = jset_magic(c); + w->data->magic = jset_magic(&c->sb); w->data->version = BCACHE_JSET_VERSION; w->data->last_seq = last_seq(&c->journal); w->data->csum = csum_set(w->data); @@ -615,10 +616,10 @@ static void journal_write_unlocked(struct closure *cl) atomic_long_add(sectors, &ca->meta_sectors_written); bio_reset(bio); - bio->bi_sector = PTR_OFFSET(k, i); + bio->bi_iter.bi_sector = PTR_OFFSET(k, i); bio->bi_bdev = ca->bdev; - bio->bi_rw = REQ_WRITE|REQ_SYNC|REQ_META|REQ_FLUSH; - bio->bi_size = sectors << 9; + bio->bi_rw = REQ_WRITE|REQ_SYNC|REQ_META|REQ_FLUSH|REQ_FUA; + bio->bi_iter.bi_size = sectors << 9; bio->bi_end_io = journal_write_endio; bio->bi_private = w; @@ -646,117 +647,144 @@ static void journal_write_unlocked(struct closure *cl) static void journal_write(struct closure *cl) { - struct cache_set *c = container_of(cl, struct cache_set, journal.io.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, bool noflush) +static void journal_try_write(struct cache_set *c) __releases(c->journal.lock) { - struct closure *cl = &c->journal.io.cl; + struct closure *cl = &c->journal.io; + struct journal_write *w = c->journal.cur; - if (!closure_trylock(cl, &c->cl)) - spin_unlock(&c->journal.lock); - else if (noflush && journal_full(&c->journal)) { + 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); - continue_at(cl, journal_write, system_wq); - } else - journal_write_unlocked(cl); + } } -#define journal_try_write(c) __journal_try_write(c, false) - -void bch_journal_meta(struct cache_set *c, struct closure *cl) +static struct journal_write *journal_wait_for_write(struct cache_set *c, + unsigned nkeys) { - struct journal_write *w; + size_t sectors; + struct closure cl; + bool wait = false; - if (CACHE_SYNC(&c->sb)) { - spin_lock(&c->journal.lock); + 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(c)) * c->sb.block_size; - w = c->journal.cur; - w->need_write = true; + if (sectors <= min_t(size_t, + c->journal.blocks_free * c->sb.block_size, + PAGE_SECTORS << JSET_BITS)) + return w; - if (cl) - BUG_ON(!closure_wait(&w->wait, cl)); + if (wait) + closure_wait(&c->journal.wait, &cl); - __journal_try_write(c, true); + 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() */ -void bch_journal(struct closure *cl) +atomic_t *bch_journal(struct cache_set *c, + struct keylist *keys, + struct closure *parent) { - struct btree_op *op = container_of(cl, struct btree_op, cl); - struct cache_set *c = op->c; struct journal_write *w; - size_t b, n = ((uint64_t *) op->keys.top) - op->keys.list; + atomic_t *ret; - if (op->type != BTREE_INSERT || - !CACHE_SYNC(&c->sb)) - goto out; + if (!CACHE_SYNC(&c->sb)) + return NULL; - /* - * If we're looping because we errored, might already be waiting on - * another journal write: - */ - while (atomic_read(&cl->parent->remaining) & CLOSURE_WAITING) - closure_sync(cl->parent); + w = journal_wait_for_write(c, bch_keylist_nkeys(keys)); - spin_lock(&c->journal.lock); + memcpy(bset_bkey_last(w->data), keys->keys, bch_keylist_bytes(keys)); + w->data->keys += bch_keylist_nkeys(keys); - if (journal_full(&c->journal)) { - /* XXX: tracepoint */ - closure_wait(&c->journal.wait, cl); + ret = &fifo_back(&c->journal.pin); + atomic_inc(ret); - journal_reclaim(c); + if (parent) { + closure_wait(&w->wait, parent); + journal_try_write(c); + } else if (!w->dirty) { + w->dirty = true; + schedule_delayed_work(&c->journal.work, + msecs_to_jiffies(c->journal_delay_ms)); + spin_unlock(&c->journal.lock); + } else { spin_unlock(&c->journal.lock); - - btree_flush_write(c); - continue_at(cl, bch_journal, bcache_wq); } - w = c->journal.cur; - w->need_write = true; - b = __set_blocks(w->data, w->data->keys + n, c); - - if (b * c->sb.block_size > PAGE_SECTORS << JSET_BITS || - b > c->journal.blocks_free) { - /* 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); - - /* XXX: tracepoint */ - BUG_ON(!closure_wait(&w->wait, cl)); - - closure_flush(&c->journal.io); - - journal_try_write(c); - continue_at(cl, bch_journal, bcache_wq); - } - memcpy(end(w->data), op->keys.list, n * sizeof(uint64_t)); - w->data->keys += n; + return ret; +} - op->journal = &fifo_back(&c->journal.pin); - atomic_inc(op->journal); +void bch_journal_meta(struct cache_set *c, struct closure *cl) +{ + struct keylist keys; + atomic_t *ref; - if (op->flush_journal) { - closure_flush(&c->journal.io); - closure_wait(&w->wait, cl->parent); - } + bch_keylist_init(&keys); - journal_try_write(c); -out: - bch_btree_insert_async(cl); + ref = bch_journal(c, &keys, cl); + if (ref) + atomic_dec_bug(ref); } void bch_journal_free(struct cache_set *c) @@ -770,8 +798,8 @@ int bch_journal_alloc(struct cache_set *c) { struct journal *j = &c->journal; - closure_init_unlocked(&j->io); spin_lock_init(&j->lock); + INIT_DELAYED_WORK(&j->work, journal_write_work); c->journal_delay_ms = 100; diff --git a/drivers/md/bcache/journal.h b/drivers/md/bcache/journal.h index 3d7851274b0..e3c39457afb 100644 --- a/drivers/md/bcache/journal.h +++ b/drivers/md/bcache/journal.h @@ -75,43 +75,6 @@ * nodes that are pinning the oldest journal entries first. */ -#define BCACHE_JSET_VERSION_UUIDv1 1 -/* Always latest UUID format */ -#define BCACHE_JSET_VERSION_UUID 1 -#define BCACHE_JSET_VERSION 1 - -/* - * On disk format for a journal entry: - * seq is monotonically increasing; every journal entry has its own unique - * sequence number. - * - * last_seq is the oldest journal entry that still has keys the btree hasn't - * flushed to disk yet. - * - * version is for on disk format changes. - */ -struct jset { - uint64_t csum; - uint64_t magic; - uint64_t seq; - uint32_t version; - uint32_t keys; - - uint64_t last_seq; - - BKEY_PADDED(uuid_bucket); - BKEY_PADDED(btree_root); - uint16_t btree_level; - uint16_t pad[3]; - - uint64_t prio_bucket[MAX_CACHES_PER_SET]; - - union { - struct bkey start[0]; - uint64_t d[0]; - }; -}; - /* * Only used for holding the journal entries we read in btree_journal_read() * during cache_registration @@ -132,6 +95,7 @@ struct journal_write { struct cache_set *c; struct closure_waitlist wait; + bool dirty; bool need_write; }; @@ -140,7 +104,9 @@ struct journal { spinlock_t lock; /* used when waiting because the journal was full */ struct closure_waitlist wait; - struct closure_with_timer io; + 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 blocks_free; @@ -188,8 +154,7 @@ struct journal_device { }; #define journal_pin_cmp(c, l, r) \ - (fifo_idx(&(c)->journal.pin, (l)->journal) > \ - fifo_idx(&(c)->journal.pin, (r)->journal)) + (fifo_idx(&(c)->journal.pin, (l)) > fifo_idx(&(c)->journal.pin, (r))) #define JOURNAL_PIN 20000 @@ -199,15 +164,14 @@ struct journal_device { struct closure; struct cache_set; struct btree_op; +struct keylist; -void bch_journal(struct closure *); +atomic_t *bch_journal(struct cache_set *, struct keylist *, struct closure *); void bch_journal_next(struct journal *); void bch_journal_mark(struct cache_set *, struct list_head *); void bch_journal_meta(struct cache_set *, struct closure *); -int bch_journal_read(struct cache_set *, struct list_head *, - struct btree_op *); -int bch_journal_replay(struct cache_set *, struct list_head *, - struct btree_op *); +int bch_journal_read(struct cache_set *, struct list_head *); +int bch_journal_replay(struct cache_set *, struct list_head *); void bch_journal_free(struct cache_set *); int bch_journal_alloc(struct cache_set *); diff --git a/drivers/md/bcache/movinggc.c b/drivers/md/bcache/movinggc.c index 8589512c972..cd7490311e5 100644 --- a/drivers/md/bcache/movinggc.c +++ b/drivers/md/bcache/movinggc.c @@ -9,9 +9,12 @@ #include "debug.h" #include "request.h" +#include <trace/events/bcache.h> + struct moving_io { + struct closure cl; struct keybuf_key *w; - struct search s; + struct data_insert_op op; struct bbio bio; }; @@ -21,13 +24,10 @@ static bool moving_pred(struct keybuf *buf, struct bkey *k) moving_gc_keys); unsigned i; - for (i = 0; i < KEY_PTRS(k); i++) { - struct cache *ca = PTR_CACHE(c, k, i); - struct bucket *g = PTR_BUCKET(c, k, i); - - if (GC_SECTORS_USED(g) < ca->gc_move_threshold) + 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; } @@ -36,40 +36,44 @@ static bool moving_pred(struct keybuf *buf, struct bkey *k) static void moving_io_destructor(struct closure *cl) { - struct moving_io *io = container_of(cl, struct moving_io, s.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, s.cl); + struct moving_io *io = container_of(cl, struct moving_io, cl); struct bio *bio = &io->bio.bio; - struct bio_vec *bv = bio_iovec_idx(bio, bio->bi_vcnt); + struct bio_vec *bv; + int i; - while (bv-- != bio->bi_io_vec) + bio_for_each_segment_all(bv, bio, i) __free_page(bv->bv_page); - pr_debug("%s %s", io->s.op.insert_collision - ? "collision moving" : "moved", - pkey(&io->w->key)); + if (io->op.replace_collision) + trace_bcache_gc_copy_collision(&io->w->key); - bch_keybuf_del(&io->s.op.c->moving_gc_keys, io->w); + bch_keybuf_del(&io->op.c->moving_gc_keys, io->w); - atomic_dec_bug(&io->s.op.c->in_flight); - closure_wake_up(&io->s.op.c->moving_gc_wait); + up(&io->op.c->moving_in_flight); closure_return_with_destructor(cl, moving_io_destructor); } static void read_moving_endio(struct bio *bio, int error) { + struct bbio *b = container_of(bio, struct bbio, bio); struct moving_io *io = container_of(bio->bi_private, - struct moving_io, s.cl); + struct moving_io, cl); if (error) - io->s.error = error; + io->op.error = error; + else if (!KEY_DIRTY(&b->key) && + ptr_stale(io->op.c, &b->key, 0)) { + io->op.error = -EINTR; + } - bch_bbio_endio(io->s.op.c, bio, error, "reading data to move"); + bch_bbio_endio(io->op.c, bio, error, "reading data to move"); } static void moving_init(struct moving_io *io) @@ -80,68 +84,70 @@ static void moving_init(struct moving_io *io) bio_get(bio); bio_set_prio(bio, IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0)); - bio->bi_size = KEY_SIZE(&io->w->key) << 9; + bio->bi_iter.bi_size = KEY_SIZE(&io->w->key) << 9; bio->bi_max_vecs = DIV_ROUND_UP(KEY_SIZE(&io->w->key), PAGE_SECTORS); - bio->bi_private = &io->s.cl; + bio->bi_private = &io->cl; bio->bi_io_vec = bio->bi_inline_vecs; bch_bio_map(bio, NULL); } static void write_moving(struct closure *cl) { - struct search *s = container_of(cl, struct search, cl); - struct moving_io *io = container_of(s, struct moving_io, s); - - if (!s->error) { - trace_bcache_write_moving(&io->bio.bio); + struct moving_io *io = container_of(cl, struct moving_io, cl); + struct data_insert_op *op = &io->op; + if (!op->error) { moving_init(io); - io->bio.bio.bi_sector = KEY_START(&io->w->key); - s->op.lock = -1; - s->op.write_prio = 1; - s->op.cache_bio = &io->bio.bio; + io->bio.bio.bi_iter.bi_sector = KEY_START(&io->w->key); + op->write_prio = 1; + op->bio = &io->bio.bio; - s->writeback = KEY_DIRTY(&io->w->key); - s->op.csum = KEY_CSUM(&io->w->key); + op->writeback = KEY_DIRTY(&io->w->key); + op->csum = KEY_CSUM(&io->w->key); - s->op.type = BTREE_REPLACE; - bkey_copy(&s->op.replace, &io->w->key); + bkey_copy(&op->replace_key, &io->w->key); + op->replace = true; - closure_init(&s->op.cl, cl); - bch_insert_data(&s->op.cl); + closure_call(&op->cl, bch_data_insert, NULL, cl); } - continue_at(cl, write_moving_finish, NULL); + continue_at(cl, write_moving_finish, op->wq); } static void read_moving_submit(struct closure *cl) { - struct search *s = container_of(cl, struct search, cl); - struct moving_io *io = container_of(s, struct moving_io, s); + struct moving_io *io = container_of(cl, struct moving_io, cl); struct bio *bio = &io->bio.bio; - trace_bcache_read_moving(bio); - bch_submit_bbio(bio, s->op.c, &io->w->key, 0); + bch_submit_bbio(bio, io->op.c, &io->w->key, 0); - continue_at(cl, write_moving, bch_gc_wq); + continue_at(cl, write_moving, io->op.wq); } -static void read_moving(struct closure *cl) +static void read_moving(struct cache_set *c) { - struct cache_set *c = container_of(cl, struct cache_set, moving_gc); 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); + 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(sizeof(struct moving_io) + sizeof(struct bio_vec) * DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS), GFP_KERNEL); @@ -150,8 +156,9 @@ static void read_moving(struct closure *cl) w->private = io; io->w = w; - io->s.op.inode = KEY_INODE(&w->key); - io->s.op.c = c; + 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; @@ -159,18 +166,13 @@ static void read_moving(struct closure *cl) bio->bi_rw = READ; bio->bi_end_io = read_moving_endio; - if (bch_bio_alloc_pages(bio, GFP_KERNEL)) + if (bio_alloc_pages(bio, GFP_KERNEL)) goto err; - pr_debug("%s", pkey(&w->key)); - - closure_call(&io->s.cl, read_moving_submit, NULL, &c->gc.cl); + trace_bcache_gc_copy(&w->key); - if (atomic_inc_return(&c->in_flight) >= 64) { - closure_wait_event(&c->moving_gc_wait, cl, - atomic_read(&c->in_flight) < 64); - continue_at(cl, read_moving, bch_gc_wq); - } + down(&c->moving_in_flight); + closure_call(&io->cl, read_moving_submit, NULL, &cl); } if (0) { @@ -180,7 +182,7 @@ err: if (!IS_ERR_OR_NULL(w->private)) bch_keybuf_del(&c->moving_gc_keys, w); } - closure_return(cl); + closure_sync(&cl); } static bool bucket_cmp(struct bucket *l, struct bucket *r) @@ -190,30 +192,33 @@ static bool bucket_cmp(struct bucket *l, struct bucket *r) static unsigned bucket_heap_top(struct cache *ca) { - return GC_SECTORS_USED(heap_peek(&ca->heap)); + struct bucket *b; + return (b = heap_peek(&ca->heap)) ? GC_SECTORS_USED(b) : 0; } -void bch_moving_gc(struct closure *cl) +void bch_moving_gc(struct cache_set *c) { - struct cache_set *c = container_of(cl, struct cache_set, gc.cl); struct cache *ca; struct bucket *b; unsigned i; if (!c->copy_gc_enabled) - closure_return(cl); + return; mutex_lock(&c->bucket_lock); for_each_cache(ca, c, i) { unsigned sectors_to_move = 0; unsigned reserve_sectors = ca->sb.bucket_size * - min(fifo_used(&ca->free), ca->free.size / 2); + fifo_used(&ca->free[RESERVE_MOVINGGC]); ca->heap.used = 0; for_each_bucket(b, ca) { - if (!GC_SECTORS_USED(b)) + 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)) { @@ -233,22 +238,19 @@ void bch_moving_gc(struct closure *cl) sectors_to_move -= GC_SECTORS_USED(b); } - ca->gc_move_threshold = bucket_heap_top(ca); - - pr_debug("threshold %u", ca->gc_move_threshold); + 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; - closure_init(&c->moving_gc, cl); - read_moving(&c->moving_gc); - - closure_return(cl); + read_moving(c); } void bch_moving_init_cache_set(struct cache_set *c) { - bch_keybuf_init(&c->moving_gc_keys, moving_pred); + 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 index e5ff12e52d5..15fff4f68a7 100644 --- a/drivers/md/bcache/request.c +++ b/drivers/md/bcache/request.c @@ -10,203 +10,41 @@ #include "btree.h" #include "debug.h" #include "request.h" +#include "writeback.h" -#include <linux/cgroup.h> #include <linux/module.h> #include <linux/hash.h> #include <linux/random.h> -#include "blk-cgroup.h" #include <trace/events/bcache.h> #define CUTOFF_CACHE_ADD 95 #define CUTOFF_CACHE_READA 90 -#define CUTOFF_WRITEBACK 50 -#define CUTOFF_WRITEBACK_SYNC 75 struct kmem_cache *bch_search_cache; -static void check_should_skip(struct cached_dev *, struct search *); - -/* Cgroup interface */ - -#ifdef CONFIG_CGROUP_BCACHE -static struct bch_cgroup bcache_default_cgroup = { .cache_mode = -1 }; - -static struct bch_cgroup *cgroup_to_bcache(struct cgroup *cgroup) -{ - struct cgroup_subsys_state *css; - return cgroup && - (css = cgroup_subsys_state(cgroup, bcache_subsys_id)) - ? container_of(css, struct bch_cgroup, css) - : &bcache_default_cgroup; -} - -struct bch_cgroup *bch_bio_to_cgroup(struct bio *bio) -{ - struct cgroup_subsys_state *css = bio->bi_css - ? cgroup_subsys_state(bio->bi_css->cgroup, bcache_subsys_id) - : task_subsys_state(current, bcache_subsys_id); - - return css - ? container_of(css, struct bch_cgroup, css) - : &bcache_default_cgroup; -} - -static ssize_t cache_mode_read(struct cgroup *cgrp, struct cftype *cft, - struct file *file, - char __user *buf, size_t nbytes, loff_t *ppos) -{ - char tmp[1024]; - int len = bch_snprint_string_list(tmp, PAGE_SIZE, bch_cache_modes, - cgroup_to_bcache(cgrp)->cache_mode + 1); - - if (len < 0) - return len; - - return simple_read_from_buffer(buf, nbytes, ppos, tmp, len); -} - -static int cache_mode_write(struct cgroup *cgrp, struct cftype *cft, - const char *buf) -{ - int v = bch_read_string_list(buf, bch_cache_modes); - if (v < 0) - return v; - - cgroup_to_bcache(cgrp)->cache_mode = v - 1; - return 0; -} - -static u64 bch_verify_read(struct cgroup *cgrp, struct cftype *cft) -{ - return cgroup_to_bcache(cgrp)->verify; -} - -static int bch_verify_write(struct cgroup *cgrp, struct cftype *cft, u64 val) -{ - cgroup_to_bcache(cgrp)->verify = val; - return 0; -} - -static u64 bch_cache_hits_read(struct cgroup *cgrp, struct cftype *cft) -{ - struct bch_cgroup *bcachecg = cgroup_to_bcache(cgrp); - return atomic_read(&bcachecg->stats.cache_hits); -} - -static u64 bch_cache_misses_read(struct cgroup *cgrp, struct cftype *cft) -{ - struct bch_cgroup *bcachecg = cgroup_to_bcache(cgrp); - return atomic_read(&bcachecg->stats.cache_misses); -} - -static u64 bch_cache_bypass_hits_read(struct cgroup *cgrp, - struct cftype *cft) -{ - struct bch_cgroup *bcachecg = cgroup_to_bcache(cgrp); - return atomic_read(&bcachecg->stats.cache_bypass_hits); -} - -static u64 bch_cache_bypass_misses_read(struct cgroup *cgrp, - struct cftype *cft) -{ - struct bch_cgroup *bcachecg = cgroup_to_bcache(cgrp); - return atomic_read(&bcachecg->stats.cache_bypass_misses); -} - -static struct cftype bch_files[] = { - { - .name = "cache_mode", - .read = cache_mode_read, - .write_string = cache_mode_write, - }, - { - .name = "verify", - .read_u64 = bch_verify_read, - .write_u64 = bch_verify_write, - }, - { - .name = "cache_hits", - .read_u64 = bch_cache_hits_read, - }, - { - .name = "cache_misses", - .read_u64 = bch_cache_misses_read, - }, - { - .name = "cache_bypass_hits", - .read_u64 = bch_cache_bypass_hits_read, - }, - { - .name = "cache_bypass_misses", - .read_u64 = bch_cache_bypass_misses_read, - }, - { } /* terminate */ -}; - -static void init_bch_cgroup(struct bch_cgroup *cg) -{ - cg->cache_mode = -1; -} - -static struct cgroup_subsys_state *bcachecg_create(struct cgroup *cgroup) -{ - struct bch_cgroup *cg; - - cg = kzalloc(sizeof(*cg), GFP_KERNEL); - if (!cg) - return ERR_PTR(-ENOMEM); - init_bch_cgroup(cg); - return &cg->css; -} - -static void bcachecg_destroy(struct cgroup *cgroup) -{ - struct bch_cgroup *cg = cgroup_to_bcache(cgroup); - free_css_id(&bcache_subsys, &cg->css); - kfree(cg); -} - -struct cgroup_subsys bcache_subsys = { - .create = bcachecg_create, - .destroy = bcachecg_destroy, - .subsys_id = bcache_subsys_id, - .name = "bcache", - .module = THIS_MODULE, -}; -EXPORT_SYMBOL_GPL(bcache_subsys); -#endif +static void bch_data_insert_start(struct closure *); static unsigned cache_mode(struct cached_dev *dc, struct bio *bio) { -#ifdef CONFIG_CGROUP_BCACHE - int r = bch_bio_to_cgroup(bio)->cache_mode; - if (r >= 0) - return r; -#endif return BDEV_CACHE_MODE(&dc->sb); } static bool verify(struct cached_dev *dc, struct bio *bio) { -#ifdef CONFIG_CGROUP_BCACHE - if (bch_bio_to_cgroup(bio)->verify) - return true; -#endif return dc->verify; } static void bio_csum(struct bio *bio, struct bkey *k) { - struct bio_vec *bv; + struct bio_vec bv; + struct bvec_iter iter; uint64_t csum = 0; - int i; - bio_for_each_segment(bv, bio, i) { - void *d = kmap(bv->bv_page) + bv->bv_offset; - csum = bch_crc64_update(csum, d, bv->bv_len); - kunmap(bv->bv_page); + 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); @@ -214,221 +52,97 @@ static void bio_csum(struct bio *bio, struct bkey *k) /* Insert data into cache */ -static void bio_invalidate(struct closure *cl) +static void bch_data_insert_keys(struct closure *cl) { - struct btree_op *op = container_of(cl, struct btree_op, cl); - struct bio *bio = op->cache_bio; + 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; - pr_debug("invalidating %i sectors from %llu", - bio_sectors(bio), (uint64_t) bio->bi_sector); - - while (bio_sectors(bio)) { - unsigned len = min(bio_sectors(bio), 1U << 14); - - if (bch_keylist_realloc(&op->keys, 0, op->c)) - goto out; - - bio->bi_sector += len; - bio->bi_size -= len << 9; - - bch_keylist_add(&op->keys, - &KEY(op->inode, bio->bi_sector, len)); - } - - op->insert_data_done = true; - bio_put(bio); -out: - continue_at(cl, bch_journal, bcache_wq); -} - -struct open_bucket { - struct list_head list; - struct task_struct *last; - unsigned sectors_free; - BKEY_PADDED(key); -}; - -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); + /* + * If we're looping, might already be waiting on + * another journal write - can't wait on more than one journal write at + * a time + * + * XXX: this looks wrong + */ +#if 0 + while (atomic_read(&s->cl.remaining) & CLOSURE_WAITING) + closure_sync(&s->cl); +#endif - for (i = 0; i < 6; i++) { - struct open_bucket *b = kzalloc(sizeof(*b), GFP_KERNEL); - if (!b) - return -ENOMEM; + if (!op->replace) + journal_ref = bch_journal(op->c, &op->insert_keys, + op->flush_journal ? cl : NULL); - list_add(&b->list, &c->data_buckets); + ret = bch_btree_insert(op->c, &op->insert_keys, + journal_ref, replace_key); + if (ret == -ESRCH) { + op->replace_collision = true; + } else if (ret) { + op->error = -ENOMEM; + op->insert_data_done = true; } - return 0; -} - -/* - * We keep multiple buckets open for writes, and try to segregate different - * write streams for better cache utilization: first 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, 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, - struct task_struct *task, - struct bkey *alloc) -{ - struct open_bucket *ret, *ret_task = NULL; - - list_for_each_entry_reverse(ret, &c->data_buckets, list) - if (!bkey_cmp(&ret->key, search)) - goto found; - else if (ret->last == task) - 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->sb.bucket_size; - bkey_copy(&ret->key, alloc); - bkey_init(alloc); - } + if (journal_ref) + atomic_dec_bug(journal_ref); - if (!ret->sectors_free) - ret = NULL; + if (!op->insert_data_done) + continue_at(cl, bch_data_insert_start, op->wq); - return ret; + bch_keylist_free(&op->insert_keys); + closure_return(cl); } -/* - * 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. - */ -static bool bch_alloc_sectors(struct bkey *k, unsigned sectors, - struct search *s) +static int bch_keylist_realloc(struct keylist *l, unsigned u64s, + struct cache_set *c) { - struct cache_set *c = s->op.c; - struct open_bucket *b; - BKEY_PADDED(key) alloc; - struct closure cl, *w = NULL; - unsigned i; - - if (s->writeback) { - closure_init_stack(&cl); - w = &cl; - } - - /* - * 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, s->task, &alloc.key))) { - unsigned watermark = s->op.write_prio - ? WATERMARK_MOVINGGC - : WATERMARK_NONE; - - spin_unlock(&c->data_bucket_lock); - - if (bch_bucket_alloc_set(c, watermark, &alloc.key, 1, w)) - return false; - - spin_lock(&c->data_bucket_lock); - } + size_t oldsize = bch_keylist_nkeys(l); + size_t newsize = oldsize + u64s; /* - * If we had to allocate, we might race and not need to allocate the - * second time we call find_data_bucket(). If we allocated a bucket but - * didn't use it, drop the refcount bch_bucket_alloc_set() took: + * 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, + * bio_insert() and bio_invalidate() will insert the keys created so far + * and finish the rest when the keylist is empty. */ - 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: */ + if (newsize * sizeof(uint64_t) > block_bytes(c) - sizeof(struct jset)) + return -ENOMEM; - for (i = 0; i < KEY_PTRS(&b->key); i++) - k->ptr[i] = b->key.ptr[i]; + return __bch_keylist_realloc(l, u64s); +} - sectors = min(sectors, b->sectors_free); +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; - SET_KEY_OFFSET(k, KEY_OFFSET(k) + sectors); - SET_KEY_SIZE(k, sectors); - SET_KEY_PTRS(k, KEY_PTRS(&b->key)); + pr_debug("invalidating %i sectors from %llu", + bio_sectors(bio), (uint64_t) bio->bi_iter.bi_sector); - /* - * 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 = s->task; + while (bio_sectors(bio)) { + unsigned sectors = min(bio_sectors(bio), + 1U << (KEY_SIZE_BITS - 1)); - b->sectors_free -= sectors; + if (bch_keylist_realloc(&op->insert_keys, 2, op->c)) + goto out; - for (i = 0; i < KEY_PTRS(&b->key); i++) { - SET_PTR_OFFSET(&b->key, i, PTR_OFFSET(&b->key, i) + sectors); + bio->bi_iter.bi_sector += sectors; + bio->bi_iter.bi_size -= sectors << 9; - atomic_long_add(sectors, - &PTR_CACHE(c, &b->key, i)->sectors_written); + bch_keylist_add(&op->insert_keys, + &KEY(op->inode, bio->bi_iter.bi_sector, sectors)); } - if (b->sectors_free < c->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; + op->insert_data_done = true; + bio_put(bio); +out: + continue_at(cl, bch_data_insert_keys, op->wq); } -static void bch_insert_data_error(struct closure *cl) +static void bch_data_insert_error(struct closure *cl) { - struct btree_op *op = container_of(cl, struct btree_op, 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 @@ -439,35 +153,34 @@ static void bch_insert_data_error(struct closure *cl) * from the keys we'll accomplish just that. */ - struct bkey *src = op->keys.bottom, *dst = op->keys.bottom; + struct bkey *src = op->insert_keys.keys, *dst = op->insert_keys.keys; - while (src != op->keys.top) { + while (src != op->insert_keys.top) { struct bkey *n = bkey_next(src); SET_KEY_PTRS(src, 0); - bkey_copy(dst, src); + memmove(dst, src, bkey_bytes(src)); dst = bkey_next(dst); src = n; } - op->keys.top = dst; + op->insert_keys.top = dst; - bch_journal(cl); + bch_data_insert_keys(cl); } -static void bch_insert_data_endio(struct bio *bio, int error) +static void bch_data_insert_endio(struct bio *bio, int error) { struct closure *cl = bio->bi_private; - struct btree_op *op = container_of(cl, struct btree_op, cl); - struct search *s = container_of(op, struct search, op); + struct data_insert_op *op = container_of(cl, struct data_insert_op, cl); if (error) { /* TODO: We could try to recover from this. */ - if (s->writeback) - s->error = error; - else if (s->write) - set_closure_fn(cl, bch_insert_data_error, bcache_wq); + if (op->writeback) + op->error = error; + else if (!op->replace) + set_closure_fn(cl, bch_data_insert_error, op->wq); else set_closure_fn(cl, NULL, NULL); } @@ -475,50 +188,52 @@ static void bch_insert_data_endio(struct bio *bio, int error) bch_bbio_endio(op->c, bio, error, "writing data to cache"); } -static void bch_insert_data_loop(struct closure *cl) +static void bch_data_insert_start(struct closure *cl) { - struct btree_op *op = container_of(cl, struct btree_op, cl); - struct search *s = container_of(op, struct search, op); - struct bio *bio = op->cache_bio, *n; - - if (op->skip) - return bio_invalidate(cl); + struct data_insert_op *op = container_of(cl, struct data_insert_op, cl); + struct bio *bio = op->bio, *n; if (atomic_sub_return(bio_sectors(bio), &op->c->sectors_to_gc) < 0) { set_gc_sectors(op->c); - bch_queue_gc(op->c); + wake_up_gc(op->c); } + if (op->bypass) + return bch_data_invalidate(cl); + + /* + * Journal writes are marked REQ_FLUSH; if the original write was a + * flush, it'll wait on the journal write. + */ + bio->bi_rw &= ~(REQ_FLUSH|REQ_FUA); + do { unsigned i; struct bkey *k; - struct bio_set *split = s->d - ? s->d->bio_split : op->c->bio_split; + struct bio_set *split = op->c->bio_split; /* 1 for the device pointer and 1 for the chksum */ - if (bch_keylist_realloc(&op->keys, - 1 + (op->csum ? 1 : 0), + if (bch_keylist_realloc(&op->insert_keys, + 3 + (op->csum ? 1 : 0), op->c)) - continue_at(cl, bch_journal, bcache_wq); + continue_at(cl, bch_data_insert_keys, op->wq); - k = op->keys.top; + k = op->insert_keys.top; bkey_init(k); SET_KEY_INODE(k, op->inode); - SET_KEY_OFFSET(k, bio->bi_sector); + SET_KEY_OFFSET(k, bio->bi_iter.bi_sector); - if (!bch_alloc_sectors(k, bio_sectors(bio), s)) + if (!bch_alloc_sectors(op->c, k, bio_sectors(bio), + op->write_point, op->write_prio, + op->writeback)) goto err; - n = bch_bio_split(bio, KEY_SIZE(k), GFP_NOIO, split); - if (!n) { - __bkey_put(op->c, k); - continue_at(cl, bch_insert_data_loop, bcache_wq); - } + n = bio_next_split(bio, KEY_SIZE(k), GFP_NOIO, split); - n->bi_end_io = bch_insert_data_endio; + n->bi_end_io = bch_data_insert_endio; n->bi_private = cl; - if (s->writeback) { + if (op->writeback) { SET_KEY_DIRTY(k, true); for (i = 0; i < KEY_PTRS(k); i++) @@ -530,19 +245,18 @@ static void bch_insert_data_loop(struct closure *cl) if (KEY_CSUM(k)) bio_csum(n, k); - pr_debug("%s", pkey(k)); - bch_keylist_push(&op->keys); + trace_bcache_cache_insert(k); + bch_keylist_push(&op->insert_keys); - trace_bcache_cache_insert(n, n->bi_sector, n->bi_bdev); n->bi_rw |= REQ_WRITE; bch_submit_bbio(n, op->c, k, 0); } while (n != bio); op->insert_data_done = true; - continue_at(cl, bch_journal, bcache_wq); + continue_at(cl, bch_data_insert_keys, op->wq); err: /* bch_alloc_sectors() blocks if s->writeback = true */ - BUG_ON(s->writeback); + BUG_ON(op->writeback); /* * But if it's not a writeback write we'd rather just bail out if @@ -550,15 +264,15 @@ err: * we might be starving btree writes for gc or something. */ - if (s->write) { + 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->skip = true; - return bio_invalidate(cl); + op->bypass = true; + return bch_data_invalidate(cl); } else { /* * From a cache miss, we can just insert the keys for the data @@ -567,15 +281,15 @@ err: op->insert_data_done = true; bio_put(bio); - if (!bch_keylist_empty(&op->keys)) - continue_at(cl, bch_journal, bcache_wq); + if (!bch_keylist_empty(&op->insert_keys)) + continue_at(cl, bch_data_insert_keys, op->wq); else closure_return(cl); } } /** - * bch_insert_data - stick some data in the cache + * bch_data_insert - stick some data in the cache * * 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 @@ -587,56 +301,176 @@ err: * 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->cache_bio; bi_sector is used for the key offset, + * It inserts the data in s->cache_bio; bi_sector is used for the key offset, * and op->inode is used for the key inode. * - * If op->skip is true, instead of inserting the data it invalidates the region - * of the cache represented by op->cache_bio and op->inode. + * If s->bypass is true, instead of inserting the data it invalidates the + * region of the cache represented by s->cache_bio and op->inode. */ -void bch_insert_data(struct closure *cl) +void bch_data_insert(struct closure *cl) { - struct btree_op *op = container_of(cl, struct btree_op, cl); + struct data_insert_op *op = container_of(cl, struct data_insert_op, cl); + + trace_bcache_write(op->bio, op->writeback, op->bypass); - bch_keylist_init(&op->keys); - bio_get(op->cache_bio); - bch_insert_data_loop(cl); + bch_keylist_init(&op->insert_keys); + bio_get(op->bio); + bch_data_insert_start(cl); } -void bch_btree_insert_async(struct closure *cl) +/* Congested? */ + +unsigned bch_get_congested(struct cache_set *c) { - struct btree_op *op = container_of(cl, struct btree_op, cl); - struct search *s = container_of(op, struct search, op); + int i; + long rand; - if (bch_btree_insert(op, op->c)) { - s->error = -ENOMEM; - op->insert_data_done = true; - } + 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; - if (op->insert_data_done) { - bch_keylist_free(&op->keys); - closure_return(cl); - } else - continue_at(cl, bch_insert_data_loop, bcache_wq); + i += CONGESTED_MAX; + + if (i > 0) + i = fract_exp_two(i, 6); + + rand = get_random_int(); + i -= bitmap_weight(&rand, BITS_PER_LONG); + + return i > 0 ? i : 1; } -/* Common code for the make_request functions */ +static void add_sequential(struct task_struct *t) +{ + ewma_add(t->sequential_io_avg, + t->sequential_io, 8, 0); -static void request_endio(struct bio *bio, int error) + t->sequential_io = 0; +} + +static struct hlist_head *iohash(struct cached_dev *dc, uint64_t k) { - struct closure *cl = bio->bi_private; + return &dc->io_hash[hash_64(k, RECENT_IO_BITS)]; +} - if (error) { - struct search *s = container_of(cl, struct search, cl); - s->error = error; - /* Only cache read errors are recoverable */ - s->recoverable = false; +static bool check_should_bypass(struct cached_dev *dc, struct bio *bio) +{ + struct cache_set *c = dc->disk.c; + unsigned mode = cache_mode(dc, bio); + unsigned sectors, congested = bch_get_congested(c); + 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->bi_rw & REQ_DISCARD)) + goto skip; + + if (mode == CACHE_MODE_NONE || + (mode == CACHE_MODE_WRITEAROUND && + (bio->bi_rw & REQ_WRITE))) + goto skip; + + if (bio->bi_iter.bi_sector & (c->sb.block_size - 1) || + bio_sectors(bio) & (c->sb.block_size - 1)) { + pr_debug("skipping unaligned io"); + goto skip; } - bio_put(bio); - closure_put(cl); + if (bypass_torture_test(dc)) { + if ((get_random_int() & 3) == 3) + goto skip; + else + goto rescale; + } + + if (!congested && !dc->sequential_cutoff) + goto rescale; + + if (!congested && + mode == CACHE_MODE_WRITEBACK && + (bio->bi_rw & REQ_WRITE) && + (bio->bi_rw & REQ_SYNC)) + 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; } -void bch_cache_read_endio(struct bio *bio, int error) +/* 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 insert_bio_sectors; + unsigned recoverable:1; + unsigned write:1; + unsigned read_dirty_data:1; + + unsigned long start_time; + + struct btree_op op; + struct data_insert_op iop; +}; + +static void bch_cache_read_endio(struct bio *bio, int error) { struct bbio *b = container_of(bio, struct bbio, bio); struct closure *cl = bio->bi_private; @@ -650,13 +484,117 @@ void bch_cache_read_endio(struct bio *bio, int error) */ if (error) - s->error = error; - else if (ptr_stale(s->op.c, &b->key, 0)) { - atomic_long_inc(&s->op.c->cache_read_races); - s->error = -EINTR; + s->iop.error = error; + 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.error = -EINTR; + } + + bch_bbio_endio(s->iop.c, bio, error, "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 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 bio_sectors = bio_sectors(bio); + unsigned 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; + 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); + + closure_return(cl); +} + +/* Common code for the make_request functions */ + +static void request_endio(struct bio *bio, int error) +{ + struct closure *cl = bio->bi_private; + + if (error) { + struct search *s = container_of(cl, struct search, cl); + s->iop.error = error; + /* Only cache read errors are recoverable */ + s->recoverable = false; } - bch_bbio_endio(s->op.c, bio, error, "reading from cache"); + bio_put(bio); + closure_put(cl); } static void bio_complete(struct search *s) @@ -670,19 +608,21 @@ static void bio_complete(struct search *s) part_stat_add(cpu, &s->d->disk->part0, ticks[rw], duration); part_stat_unlock(); - trace_bcache_request_end(s, s->orig_bio); - bio_endio(s->orig_bio, s->error); + trace_bcache_request_end(s->d, s->orig_bio); + bio_endio(s->orig_bio, s->iop.error); s->orig_bio = NULL; } } -static void do_bio_hook(struct search *s) +static void do_bio_hook(struct search *s, struct bio *orig_bio) { struct bio *bio = &s->bio.bio; - memcpy(bio, s->orig_bio, sizeof(struct bio)); + bio_init(bio); + __bio_clone_fast(bio, orig_bio); bio->bi_end_io = request_endio; bio->bi_private = &s->cl; + atomic_set(&bio->bi_cnt, 3); } @@ -691,61 +631,44 @@ static void search_free(struct closure *cl) struct search *s = container_of(cl, struct search, cl); bio_complete(s); - if (s->op.cache_bio) - bio_put(s->op.cache_bio); - - if (s->unaligned_bvec) - mempool_free(s->bio.bio.bi_io_vec, s->d->unaligned_bvec); + if (s->iop.bio) + bio_put(s->iop.bio); closure_debug_destroy(cl); mempool_free(s, s->d->c->search); } -static struct search *search_alloc(struct bio *bio, struct bcache_device *d) +static inline struct search *search_alloc(struct bio *bio, + struct bcache_device *d) { - struct bio_vec *bv; - struct search *s = mempool_alloc(d->c->search, GFP_NOIO); - memset(s, 0, offsetof(struct search, op.keys)); + struct search *s; - __closure_init(&s->cl, NULL); + s = mempool_alloc(d->c->search, GFP_NOIO); + + closure_init(&s->cl, NULL); + do_bio_hook(s, bio); - s->op.inode = d->id; - s->op.c = d->c; - s->d = d; - s->op.lock = -1; - s->task = current; s->orig_bio = bio; - s->write = (bio->bi_rw & REQ_WRITE) != 0; - s->op.flush_journal = (bio->bi_rw & REQ_FLUSH) != 0; - s->op.skip = (bio->bi_rw & REQ_DISCARD) != 0; + s->cache_miss = NULL; + s->d = d; s->recoverable = 1; + s->write = (bio->bi_rw & REQ_WRITE) != 0; + s->read_dirty_data = 0; s->start_time = jiffies; - do_bio_hook(s); - - if (bio->bi_size != bio_segments(bio) * PAGE_SIZE) { - bv = mempool_alloc(d->unaligned_bvec, GFP_NOIO); - memcpy(bv, bio_iovec(bio), - sizeof(struct bio_vec) * bio_segments(bio)); - s->bio.bio.bi_io_vec = bv; - s->unaligned_bvec = 1; - } + 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.error = 0; + s->iop.flags = 0; + s->iop.flush_journal = (bio->bi_rw & (REQ_FLUSH|REQ_FUA)) != 0; + s->iop.wq = bcache_wq; return s; } -static void btree_read_async(struct closure *cl) -{ - struct btree_op *op = container_of(cl, struct btree_op, cl); - - int ret = btree_root(search_recurse, op->c, op); - - if (ret == -EAGAIN) - continue_at(cl, btree_read_async, bcache_wq); - - closure_return(cl); -} - /* Cached devices */ static void cached_dev_bio_complete(struct closure *cl) @@ -759,236 +682,179 @@ static void cached_dev_bio_complete(struct closure *cl) /* Process reads */ -static void cached_dev_read_complete(struct closure *cl) +static void cached_dev_cache_miss_done(struct closure *cl) { struct search *s = container_of(cl, struct search, cl); - if (s->op.insert_collision) - bch_mark_cache_miss_collision(s); + if (s->iop.replace_collision) + bch_mark_cache_miss_collision(s->iop.c, s->d); - if (s->op.cache_bio) { + if (s->iop.bio) { int i; struct bio_vec *bv; - __bio_for_each_segment(bv, s->op.cache_bio, i, 0) + bio_for_each_segment_all(bv, s->iop.bio, i) __free_page(bv->bv_page); } cached_dev_bio_complete(cl); } -static void request_read_error(struct closure *cl) +static void cached_dev_read_error(struct closure *cl) { struct search *s = container_of(cl, struct search, cl); - struct bio_vec *bv; - int i; + struct bio *bio = &s->bio.bio; if (s->recoverable) { - /* The cache read failed, but we can retry from the backing - * device. - */ - pr_debug("recovering at sector %llu", - (uint64_t) s->orig_bio->bi_sector); - - s->error = 0; - bv = s->bio.bio.bi_io_vec; - do_bio_hook(s); - s->bio.bio.bi_io_vec = bv; + /* Retry from the backing device: */ + trace_bcache_read_retry(s->orig_bio); - if (!s->unaligned_bvec) - bio_for_each_segment(bv, s->orig_bio, i) - bv->bv_offset = 0, bv->bv_len = PAGE_SIZE; - else - memcpy(s->bio.bio.bi_io_vec, - bio_iovec(s->orig_bio), - sizeof(struct bio_vec) * - bio_segments(s->orig_bio)); + s->iop.error = 0; + do_bio_hook(s, s->orig_bio); /* XXX: invalidate cache */ - trace_bcache_read_retry(&s->bio.bio); - closure_bio_submit(&s->bio.bio, &s->cl, s->d); + closure_bio_submit(bio, cl, s->d); } - continue_at(cl, cached_dev_read_complete, NULL); + continue_at(cl, cached_dev_cache_miss_done, NULL); } -static void request_read_done(struct closure *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); /* - * s->cache_bio != NULL implies that we had a cache miss; cache_bio now - * contains data ready to be inserted into the cache. + * 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->op.cache_bio) { - struct bio_vec *src, *dst; - unsigned src_offset, dst_offset, bytes; - void *dst_ptr; - - bio_reset(s->op.cache_bio); - s->op.cache_bio->bi_sector = s->cache_miss->bi_sector; - s->op.cache_bio->bi_bdev = s->cache_miss->bi_bdev; - s->op.cache_bio->bi_size = s->cache_bio_sectors << 9; - bch_bio_map(s->op.cache_bio, NULL); - - src = bio_iovec(s->op.cache_bio); - dst = bio_iovec(s->cache_miss); - src_offset = src->bv_offset; - dst_offset = dst->bv_offset; - dst_ptr = kmap(dst->bv_page); - - while (1) { - if (dst_offset == dst->bv_offset + dst->bv_len) { - kunmap(dst->bv_page); - dst++; - if (dst == bio_iovec_idx(s->cache_miss, - s->cache_miss->bi_vcnt)) - break; - - dst_offset = dst->bv_offset; - dst_ptr = kmap(dst->bv_page); - } - - if (src_offset == src->bv_offset + src->bv_len) { - src++; - if (src == bio_iovec_idx(s->op.cache_bio, - s->op.cache_bio->bi_vcnt)) - BUG(); - - src_offset = src->bv_offset; - } - - bytes = min(dst->bv_offset + dst->bv_len - dst_offset, - src->bv_offset + src->bv_len - src_offset); - - memcpy(dst_ptr + dst_offset, - page_address(src->bv_page) + src_offset, - bytes); - - src_offset += bytes; - dst_offset += bytes; - } + if (s->iop.bio) { + bio_reset(s->iop.bio); + s->iop.bio->bi_iter.bi_sector = s->cache_miss->bi_iter.bi_sector; + s->iop.bio->bi_bdev = s->cache_miss->bi_bdev; + 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->bio.bio) && s->recoverable) - bch_data_verify(s); + if (verify(dc, &s->bio.bio) && s->recoverable && !s->read_dirty_data) + bch_data_verify(dc, s->orig_bio); bio_complete(s); - if (s->op.cache_bio && - !test_bit(CACHE_SET_STOPPING, &s->op.c->flags)) { - s->op.type = BTREE_REPLACE; - closure_call(&s->op.cl, bch_insert_data, NULL, cl); + 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_read_complete, NULL); + continue_at(cl, cached_dev_cache_miss_done, NULL); } -static void request_read_done_bh(struct closure *cl) +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, !s->cache_miss, s->op.skip); + bch_mark_cache_accounting(s->iop.c, s->d, + !s->cache_miss, s->iop.bypass); + trace_bcache_read(s->orig_bio, !s->cache_miss, s->iop.bypass); - if (s->error) - continue_at_nobarrier(cl, request_read_error, bcache_wq); - else if (s->op.cache_bio || verify(dc, &s->bio.bio)) - continue_at_nobarrier(cl, request_read_done, bcache_wq); + if (s->iop.error) + continue_at_nobarrier(cl, cached_dev_read_error, bcache_wq); + else if (s->iop.bio || verify(dc, &s->bio.bio)) + continue_at_nobarrier(cl, cached_dev_read_done, bcache_wq); else - continue_at_nobarrier(cl, cached_dev_read_complete, NULL); + 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 sectors) { - int ret = 0; - unsigned reada; + int ret = MAP_CONTINUE; + unsigned reada = 0; struct cached_dev *dc = container_of(s->d, struct cached_dev, disk); - struct bio *miss; + struct bio *miss, *cache_bio; - miss = bch_bio_split(bio, sectors, GFP_NOIO, s->d->bio_split); - if (!miss) - return -EAGAIN; + 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 (miss == bio) - s->op.lookup_done = true; + if (!(bio->bi_rw & REQ_RAHEAD) && + !(bio->bi_rw & REQ_META) && + s->iop.c->gc_stats.in_use < CUTOFF_CACHE_READA) + reada = min_t(sector_t, dc->readahead >> 9, + bdev_sectors(bio->bi_bdev) - bio_end_sector(bio)); - miss->bi_end_io = request_endio; - miss->bi_private = &s->cl; + s->insert_bio_sectors = min(sectors, bio_sectors(bio) + reada); - if (s->cache_miss || s->op.skip) - goto out_submit; + s->iop.replace_key = KEY(s->iop.inode, + bio->bi_iter.bi_sector + s->insert_bio_sectors, + s->insert_bio_sectors); - if (miss != bio || - (bio->bi_rw & REQ_RAHEAD) || - (bio->bi_rw & REQ_META) || - s->op.c->gc_stats.in_use >= CUTOFF_CACHE_READA) - reada = 0; - else { - reada = min(dc->readahead >> 9, - sectors - bio_sectors(miss)); - - if (bio_end(miss) + reada > bdev_sectors(miss->bi_bdev)) - reada = bdev_sectors(miss->bi_bdev) - bio_end(miss); - } + ret = bch_btree_insert_check_key(b, &s->op, &s->iop.replace_key); + if (ret) + return ret; - s->cache_bio_sectors = bio_sectors(miss) + reada; - s->op.cache_bio = bio_alloc_bioset(GFP_NOWAIT, - DIV_ROUND_UP(s->cache_bio_sectors, PAGE_SECTORS), - dc->disk.bio_split); + s->iop.replace = true; - if (!s->op.cache_bio) - goto out_submit; + 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; - s->op.cache_bio->bi_sector = miss->bi_sector; - s->op.cache_bio->bi_bdev = miss->bi_bdev; - s->op.cache_bio->bi_size = s->cache_bio_sectors << 9; + 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; - s->op.cache_bio->bi_end_io = request_endio; - s->op.cache_bio->bi_private = &s->cl; + cache_bio->bi_iter.bi_sector = miss->bi_iter.bi_sector; + cache_bio->bi_bdev = miss->bi_bdev; + cache_bio->bi_iter.bi_size = s->insert_bio_sectors << 9; - /* btree_search_recurse()'s btree iterator is no good anymore */ - ret = -EINTR; - if (!bch_btree_insert_check_key(b, &s->op, s->op.cache_bio)) - goto out_put; + cache_bio->bi_end_io = request_endio; + cache_bio->bi_private = &s->cl; - bch_bio_map(s->op.cache_bio, NULL); - if (bch_bio_alloc_pages(s->op.cache_bio, __GFP_NOWARN|GFP_NOIO)) + bch_bio_map(cache_bio, NULL); + if (bio_alloc_pages(cache_bio, __GFP_NOWARN|GFP_NOIO)) goto out_put; - s->cache_miss = miss; - bio_get(s->op.cache_bio); + if (reada) + bch_mark_cache_readahead(s->iop.c, s->d); - trace_bcache_cache_miss(s->orig_bio); - closure_bio_submit(s->op.cache_bio, &s->cl, s->d); + s->cache_miss = miss; + s->iop.bio = cache_bio; + bio_get(cache_bio); + closure_bio_submit(cache_bio, &s->cl, s->d); return ret; out_put: - bio_put(s->op.cache_bio); - s->op.cache_bio = NULL; + bio_put(cache_bio); out_submit: + miss->bi_end_io = request_endio; + miss->bi_private = &s->cl; closure_bio_submit(miss, &s->cl, s->d); return ret; } -static void request_read(struct cached_dev *dc, struct search *s) +static void cached_dev_read(struct cached_dev *dc, struct search *s) { struct closure *cl = &s->cl; - check_should_skip(dc, s); - closure_call(&s->op.cl, btree_read_async, NULL, cl); - - continue_at(cl, request_read_done_bh, NULL); + closure_call(&s->iop.cl, cache_lookup, NULL, cl); + continue_at(cl, cached_dev_read_done_bh, NULL); } /* Process writes */ @@ -1002,85 +868,84 @@ static void cached_dev_write_complete(struct closure *cl) cached_dev_bio_complete(cl); } -static bool should_writeback(struct cached_dev *dc, struct bio *bio) -{ - unsigned threshold = (bio->bi_rw & REQ_SYNC) - ? CUTOFF_WRITEBACK_SYNC - : CUTOFF_WRITEBACK; - - return !atomic_read(&dc->disk.detaching) && - cache_mode(dc, bio) == CACHE_MODE_WRITEBACK && - dc->disk.c->gc_stats.in_use < threshold; -} - -static void request_write(struct cached_dev *dc, struct search *s) +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, end; - start = KEY(dc->disk.id, bio->bi_sector, 0); - end = KEY(dc->disk.id, bio_end(bio), 0); + 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->op.c->moving_gc_keys, &start, &end); + bch_keybuf_check_overlapping(&s->iop.c->moving_gc_keys, &start, &end); - check_should_skip(dc, s); down_read_non_owner(&dc->writeback_lock); - if (bch_keybuf_check_overlapping(&dc->writeback_keys, &start, &end)) { - s->op.skip = false; - s->writeback = true; + /* + * 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->bi_rw & REQ_DISCARD) - goto skip; + s->iop.bypass = true; - if (s->op.skip) - goto skip; + if (should_writeback(dc, s->orig_bio, + cache_mode(dc, bio), + s->iop.bypass)) { + s->iop.bypass = false; + s->iop.writeback = true; + } - if (should_writeback(dc, s->orig_bio)) - s->writeback = true; + if (s->iop.bypass) { + s->iop.bio = s->orig_bio; + bio_get(s->iop.bio); - if (!s->writeback) { - s->op.cache_bio = bio_clone_bioset(bio, GFP_NOIO, - dc->disk.bio_split); + if (!(bio->bi_rw & REQ_DISCARD) || + blk_queue_discard(bdev_get_queue(dc->bdev))) + closure_bio_submit(bio, cl, s->d); + } else if (s->iop.writeback) { + bch_writeback_add(dc); + s->iop.bio = bio; - trace_bcache_writethrough(s->orig_bio); - closure_bio_submit(bio, cl, s->d); + if (bio->bi_rw & REQ_FLUSH) { + /* Also need to send a flush to the backing device */ + struct bio *flush = bio_alloc_bioset(GFP_NOIO, 0, + dc->disk.bio_split); + + flush->bi_rw = WRITE_FLUSH; + flush->bi_bdev = bio->bi_bdev; + flush->bi_end_io = request_endio; + flush->bi_private = cl; + + closure_bio_submit(flush, cl, s->d); + } } else { - s->op.cache_bio = bio; - trace_bcache_writeback(s->orig_bio); - bch_writeback_add(dc, bio_sectors(bio)); - } -out: - closure_call(&s->op.cl, bch_insert_data, NULL, cl); - continue_at(cl, cached_dev_write_complete, NULL); -skip: - s->op.skip = true; - s->op.cache_bio = s->orig_bio; - bio_get(s->op.cache_bio); - trace_bcache_write_skip(s->orig_bio); + s->iop.bio = bio_clone_fast(bio, GFP_NOIO, dc->disk.bio_split); - if ((bio->bi_rw & REQ_DISCARD) && - !blk_queue_discard(bdev_get_queue(dc->bdev))) - goto out; + closure_bio_submit(bio, cl, s->d); + } - closure_bio_submit(bio, cl, s->d); - goto out; + closure_call(&s->iop.cl, bch_data_insert, NULL, cl); + continue_at(cl, cached_dev_write_complete, NULL); } -static void request_nodata(struct cached_dev *dc, struct search *s) +static void cached_dev_nodata(struct closure *cl) { - struct closure *cl = &s->cl; + struct search *s = container_of(cl, struct search, cl); struct bio *bio = &s->bio.bio; - if (bio->bi_rw & REQ_DISCARD) { - request_write(dc, s); - return; - } - - if (s->op.flush_journal) - bch_journal_meta(s->op.c, cl); + 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 */ closure_bio_submit(bio, cl, s->d); continue_at(cl, cached_dev_bio_complete, NULL); @@ -1088,125 +953,6 @@ static void request_nodata(struct cached_dev *dc, struct search *s) /* Cached devices - read & write stuff */ -int bch_get_congested(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; - - return i <= 0 ? 1 : fract_exp_two(i, 6); -} - -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 void check_should_skip(struct cached_dev *dc, struct search *s) -{ - struct cache_set *c = s->op.c; - struct bio *bio = &s->bio.bio; - - long rand; - int cutoff = bch_get_congested(c); - unsigned mode = cache_mode(dc, bio); - - if (atomic_read(&dc->disk.detaching) || - c->gc_stats.in_use > CUTOFF_CACHE_ADD || - (bio->bi_rw & REQ_DISCARD)) - goto skip; - - if (mode == CACHE_MODE_NONE || - (mode == CACHE_MODE_WRITEAROUND && - (bio->bi_rw & REQ_WRITE))) - goto skip; - - if (bio->bi_sector & (c->sb.block_size - 1) || - bio_sectors(bio) & (c->sb.block_size - 1)) { - pr_debug("skipping unaligned io"); - goto skip; - } - - if (!cutoff) { - cutoff = dc->sequential_cutoff >> 9; - - if (!cutoff) - goto rescale; - - if (mode == CACHE_MODE_WRITEBACK && - (bio->bi_rw & REQ_WRITE) && - (bio->bi_rw & REQ_SYNC)) - goto rescale; - } - - if (dc->sequential_merge) { - struct io *i; - - spin_lock(&dc->io_lock); - - hlist_for_each_entry(i, iohash(dc, bio->bi_sector), hash) - if (i->last == bio->bi_sector && - time_before(jiffies, i->jiffies)) - goto found; - - i = list_first_entry(&dc->io_lru, struct io, lru); - - add_sequential(s->task); - i->sequential = 0; -found: - if (i->sequential + bio->bi_size > i->sequential) - i->sequential += bio->bi_size; - - i->last = bio_end(bio); - i->jiffies = jiffies + msecs_to_jiffies(5000); - s->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); - } else { - s->task->sequential_io = bio->bi_size; - - add_sequential(s->task); - } - - rand = get_random_int(); - cutoff -= bitmap_weight(&rand, BITS_PER_LONG); - - if (cutoff <= (int) (max(s->task->sequential_io, - s->task->sequential_io_avg) >> 9)) - goto skip; - -rescale: - bch_rescale_priorities(c, bio_sectors(bio)); - return; -skip: - bch_mark_sectors_bypassed(s, bio_sectors(bio)); - s->op.skip = true; -} - static void cached_dev_make_request(struct request_queue *q, struct bio *bio) { struct search *s; @@ -1220,18 +966,28 @@ static void cached_dev_make_request(struct request_queue *q, struct bio *bio) part_stat_unlock(); bio->bi_bdev = dc->bdev; - bio->bi_sector += dc->sb.data_offset; + bio->bi_iter.bi_sector += dc->sb.data_offset; if (cached_dev_get(dc)) { s = search_alloc(bio, d); - trace_bcache_request_start(s, bio); - - if (!bio_has_data(bio)) - request_nodata(dc, s); - else if (rw) - request_write(dc, s); - else - request_read(dc, s); + trace_bcache_request_start(s->d, bio); + + if (!bio->bi_iter.bi_size) { + /* + * can't call bch_journal_meta from under + * generic_make_request + */ + 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 { if ((bio->bi_rw & REQ_DISCARD) && !blk_queue_discard(bdev_get_queue(dc->bdev))) @@ -1288,32 +1044,28 @@ void bch_cached_dev_request_init(struct cached_dev *dc) static int flash_dev_cache_miss(struct btree *b, struct search *s, struct bio *bio, unsigned sectors) { - /* Zero fill bio */ - - while (bio->bi_idx != bio->bi_vcnt) { - struct bio_vec *bv = bio_iovec(bio); - unsigned j = min(bv->bv_len >> 9, sectors); + unsigned bytes = min(sectors, bio_sectors(bio)) << 9; - void *p = kmap(bv->bv_page); - memset(p + bv->bv_offset, 0, j << 9); - kunmap(bv->bv_page); + swap(bio->bi_iter.bi_size, bytes); + zero_fill_bio(bio); + swap(bio->bi_iter.bi_size, bytes); - bv->bv_len -= j << 9; - bv->bv_offset += j << 9; + bio_advance(bio, bytes); - if (bv->bv_len) - return 0; + if (!bio->bi_iter.bi_size) + return MAP_DONE; - bio->bi_sector += j; - bio->bi_size -= j << 9; + return MAP_CONTINUE; +} - bio->bi_idx++; - sectors -= j; - } +static void flash_dev_nodata(struct closure *cl) +{ + struct search *s = container_of(cl, struct search, cl); - s->op.lookup_done = true; + if (s->iop.flush_journal) + bch_journal_meta(s->iop.c, cl); - return 0; + continue_at(cl, search_free, NULL); } static void flash_dev_make_request(struct request_queue *q, struct bio *bio) @@ -1332,23 +1084,28 @@ static void flash_dev_make_request(struct request_queue *q, struct bio *bio) cl = &s->cl; bio = &s->bio.bio; - trace_bcache_request_start(s, bio); - - if (bio_has_data(bio) && !rw) { - closure_call(&s->op.cl, btree_read_async, NULL, cl); - } else if (bio_has_data(bio) || s->op.skip) { - bch_keybuf_check_overlapping(&s->op.c->moving_gc_keys, - &KEY(d->id, bio->bi_sector, 0), - &KEY(d->id, bio_end(bio), 0)); + trace_bcache_request_start(s->d, bio); - s->writeback = true; - s->op.cache_bio = bio; - - closure_call(&s->op.cl, bch_insert_data, NULL, cl); + if (!bio->bi_iter.bi_size) { + /* + * can't call bch_journal_meta from under + * generic_make_request + */ + continue_at_nobarrier(&s->cl, + flash_dev_nodata, + bcache_wq); + } else if (rw) { + 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->bi_rw & REQ_DISCARD) != 0; + s->iop.writeback = true; + s->iop.bio = bio; + + closure_call(&s->iop.cl, bch_data_insert, NULL, cl); } else { - /* No data - probably a cache flush */ - if (s->op.flush_journal) - bch_journal_meta(s->op.c, cl); + closure_call(&s->iop.cl, cache_lookup, NULL, cl); } continue_at(cl, search_free, NULL); @@ -1388,9 +1145,6 @@ void bch_flash_dev_request_init(struct bcache_device *d) void bch_request_exit(void) { -#ifdef CONFIG_CGROUP_BCACHE - cgroup_unload_subsys(&bcache_subsys); -#endif if (bch_search_cache) kmem_cache_destroy(bch_search_cache); } @@ -1401,11 +1155,5 @@ int __init bch_request_init(void) if (!bch_search_cache) return -ENOMEM; -#ifdef CONFIG_CGROUP_BCACHE - cgroup_load_subsys(&bcache_subsys); - init_bch_cgroup(&bcache_default_cgroup); - - cgroup_add_cftypes(&bcache_subsys, bch_files); -#endif return 0; } diff --git a/drivers/md/bcache/request.h b/drivers/md/bcache/request.h index 254d9ab5707..1ff36875c2b 100644 --- a/drivers/md/bcache/request.h +++ b/drivers/md/bcache/request.h @@ -1,62 +1,43 @@ #ifndef _BCACHE_REQUEST_H_ #define _BCACHE_REQUEST_H_ -#include <linux/cgroup.h> - -struct search { - /* Stack frame for bio_complete */ +struct data_insert_op { struct closure cl; + struct cache_set *c; + struct bio *bio; + struct workqueue_struct *wq; - struct bcache_device *d; - struct task_struct *task; + unsigned inode; + uint16_t write_point; + uint16_t write_prio; + short error; - struct bbio bio; - struct bio *orig_bio; - struct bio *cache_miss; - unsigned cache_bio_sectors; + union { + uint16_t flags; - unsigned recoverable:1; - unsigned unaligned_bvec:1; + struct { + unsigned bypass:1; + unsigned writeback:1; + unsigned flush_journal:1; + unsigned csum:1; - unsigned write:1; - unsigned writeback:1; + unsigned replace:1; + unsigned replace_collision:1; - /* IO error returned to s->bio */ - short error; - unsigned long start_time; + unsigned insert_data_done:1; + }; + }; - /* Anything past op->keys won't get zeroed in do_bio_hook */ - struct btree_op op; + struct keylist insert_keys; + BKEY_PADDED(replace_key); }; -void bch_cache_read_endio(struct bio *, int); -int bch_get_congested(struct cache_set *); -void bch_insert_data(struct closure *cl); -void bch_btree_insert_async(struct closure *); -void bch_cache_read_endio(struct bio *, int); - -void bch_open_buckets_free(struct cache_set *); -int bch_open_buckets_alloc(struct cache_set *); +unsigned bch_get_congested(struct cache_set *); +void bch_data_insert(struct closure *cl); void bch_cached_dev_request_init(struct cached_dev *dc); void bch_flash_dev_request_init(struct bcache_device *d); extern struct kmem_cache *bch_search_cache, *bch_passthrough_cache; -struct bch_cgroup { -#ifdef CONFIG_CGROUP_BCACHE - struct cgroup_subsys_state css; -#endif - /* - * We subtract one from the index into bch_cache_modes[], so that - * default == -1; this makes it so the rest match up with d->cache_mode, - * and we use d->cache_mode if cgrp->cache_mode < 0 - */ - short cache_mode; - bool verify; - struct cache_stat_collector stats; -}; - -struct bch_cgroup *bch_bio_to_cgroup(struct bio *bio); - #endif /* _BCACHE_REQUEST_H_ */ diff --git a/drivers/md/bcache/stats.c b/drivers/md/bcache/stats.c index b8730e714d6..0ca072c20d0 100644 --- a/drivers/md/bcache/stats.c +++ b/drivers/md/bcache/stats.c @@ -7,7 +7,6 @@ #include "bcache.h" #include "stats.h" #include "btree.h" -#include "request.h" #include "sysfs.h" /* @@ -196,35 +195,33 @@ static void mark_cache_stats(struct cache_stat_collector *stats, atomic_inc(&stats->cache_bypass_misses); } -void bch_mark_cache_accounting(struct search *s, bool hit, bool bypass) +void bch_mark_cache_accounting(struct cache_set *c, struct bcache_device *d, + bool hit, bool bypass) { - struct cached_dev *dc = container_of(s->d, struct cached_dev, disk); + struct cached_dev *dc = container_of(d, struct cached_dev, disk); mark_cache_stats(&dc->accounting.collector, hit, bypass); - mark_cache_stats(&s->op.c->accounting.collector, hit, bypass); -#ifdef CONFIG_CGROUP_BCACHE - mark_cache_stats(&(bch_bio_to_cgroup(s->orig_bio)->stats), hit, bypass); -#endif + mark_cache_stats(&c->accounting.collector, hit, bypass); } -void bch_mark_cache_readahead(struct search *s) +void bch_mark_cache_readahead(struct cache_set *c, struct bcache_device *d) { - struct cached_dev *dc = container_of(s->d, struct cached_dev, disk); + struct cached_dev *dc = container_of(d, struct cached_dev, disk); atomic_inc(&dc->accounting.collector.cache_readaheads); - atomic_inc(&s->op.c->accounting.collector.cache_readaheads); + atomic_inc(&c->accounting.collector.cache_readaheads); } -void bch_mark_cache_miss_collision(struct search *s) +void bch_mark_cache_miss_collision(struct cache_set *c, struct bcache_device *d) { - struct cached_dev *dc = container_of(s->d, struct cached_dev, disk); + struct cached_dev *dc = container_of(d, struct cached_dev, disk); atomic_inc(&dc->accounting.collector.cache_miss_collisions); - atomic_inc(&s->op.c->accounting.collector.cache_miss_collisions); + atomic_inc(&c->accounting.collector.cache_miss_collisions); } -void bch_mark_sectors_bypassed(struct search *s, int sectors) +void bch_mark_sectors_bypassed(struct cache_set *c, struct cached_dev *dc, + int sectors) { - struct cached_dev *dc = container_of(s->d, struct cached_dev, disk); atomic_add(sectors, &dc->accounting.collector.sectors_bypassed); - atomic_add(sectors, &s->op.c->accounting.collector.sectors_bypassed); + atomic_add(sectors, &c->accounting.collector.sectors_bypassed); } void bch_cache_accounting_init(struct cache_accounting *acc, diff --git a/drivers/md/bcache/stats.h b/drivers/md/bcache/stats.h index c7c7a8fd29f..adbff141c88 100644 --- a/drivers/md/bcache/stats.h +++ b/drivers/md/bcache/stats.h @@ -38,7 +38,9 @@ struct cache_accounting { struct cache_stats day; }; -struct search; +struct cache_set; +struct cached_dev; +struct bcache_device; void bch_cache_accounting_init(struct cache_accounting *acc, struct closure *parent); @@ -50,9 +52,10 @@ void bch_cache_accounting_clear(struct cache_accounting *acc); void bch_cache_accounting_destroy(struct cache_accounting *acc); -void bch_mark_cache_accounting(struct search *s, bool hit, bool bypass); -void bch_mark_cache_readahead(struct search *s); -void bch_mark_cache_miss_collision(struct search *s); -void bch_mark_sectors_bypassed(struct search *s, int sectors); +void bch_mark_cache_accounting(struct cache_set *, struct bcache_device *, + bool, bool); +void bch_mark_cache_readahead(struct cache_set *, struct bcache_device *); +void bch_mark_cache_miss_collision(struct cache_set *, struct bcache_device *); +void bch_mark_sectors_bypassed(struct cache_set *, struct cached_dev *, int); #endif /* _BCACHE_STATS_H_ */ diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c index f88e2b653a3..926ded8ccbf 100644 --- a/drivers/md/bcache/super.c +++ b/drivers/md/bcache/super.c @@ -9,11 +9,16 @@ #include "bcache.h" #include "btree.h" #include "debug.h" +#include "extents.h" #include "request.h" +#include "writeback.h" +#include <linux/blkdev.h> #include <linux/buffer_head.h> #include <linux/debugfs.h> #include <linux/genhd.h> +#include <linux/idr.h> +#include <linux/kthread.h> #include <linux/module.h> #include <linux/random.h> #include <linux/reboot.h> @@ -42,21 +47,13 @@ const char * const bch_cache_modes[] = { NULL }; -struct uuid_entry_v0 { - uint8_t uuid[16]; - uint8_t label[32]; - uint32_t first_reg; - uint32_t last_reg; - uint32_t invalidated; - uint32_t pad; -}; - static struct kobject *bcache_kobj; struct mutex bch_register_lock; LIST_HEAD(bch_cache_sets); static LIST_HEAD(uncached_devices); -static int bcache_major, bcache_minor; +static int bcache_major; +static DEFINE_IDA(bcache_minor); static wait_queue_head_t unregister_wait; struct workqueue_struct *bcache_wq; @@ -229,7 +226,7 @@ static void write_bdev_super_endio(struct bio *bio, int error) struct cached_dev *dc = bio->bi_private; /* XXX: error checking */ - closure_put(&dc->sb_write.cl); + closure_put(&dc->sb_write); } static void __write_super(struct cache_sb *sb, struct bio *bio) @@ -237,9 +234,9 @@ static void __write_super(struct cache_sb *sb, struct bio *bio) struct cache_sb *out = page_address(bio->bi_io_vec[0].bv_page); unsigned i; - bio->bi_sector = SB_SECTOR; - bio->bi_rw = REQ_SYNC|REQ_META; - bio->bi_size = SB_SIZE; + bio->bi_iter.bi_sector = SB_SECTOR; + bio->bi_rw = REQ_SYNC|REQ_META; + bio->bi_iter.bi_size = SB_SIZE; bch_bio_map(bio, NULL); out->offset = cpu_to_le64(sb->offset); @@ -267,12 +264,20 @@ static void __write_super(struct cache_sb *sb, struct bio *bio) submit_bio(REQ_WRITE, 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.cl; + struct closure *cl = &dc->sb_write; struct bio *bio = &dc->sb_bio; - closure_lock(&dc->sb_write, parent); + down(&dc->sb_write_mutex); + closure_init(cl, parent); bio_reset(bio); bio->bi_bdev = dc->bdev; @@ -282,7 +287,7 @@ void bch_write_bdev_super(struct cached_dev *dc, struct closure *parent) closure_get(cl); __write_super(&dc->sb, bio); - closure_return(cl); + closure_return_with_destructor(cl, bch_write_bdev_super_unlock); } static void write_super_endio(struct bio *bio, int error) @@ -290,16 +295,24 @@ static void write_super_endio(struct bio *bio, int error) struct cache *ca = bio->bi_private; bch_count_io_errors(ca, error, "writing superblock"); - closure_put(&ca->set->sb_write.cl); + 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.cl; + struct closure *cl = &c->sb_write; struct cache *ca; unsigned i; - closure_lock(&c->sb_write, &c->cl); + down(&c->sb_write_mutex); + closure_init(cl, &c->cl); c->sb.seq++; @@ -321,7 +334,7 @@ void bcache_write_super(struct cache_set *c) __write_super(&ca->sb, bio); } - closure_return(cl); + closure_return_with_destructor(cl, bcache_write_super_unlock); } /* UUID io */ @@ -329,28 +342,37 @@ void bcache_write_super(struct cache_set *c) static void uuid_endio(struct bio *bio, int error) { struct closure *cl = bio->bi_private; - struct cache_set *c = container_of(cl, struct cache_set, uuid_write.cl); + struct cache_set *c = container_of(cl, struct cache_set, uuid_write); cache_set_err_on(error, 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, unsigned long rw, struct bkey *k, struct closure *parent) { - struct closure *cl = &c->uuid_write.cl; + struct closure *cl = &c->uuid_write; struct uuid_entry *u; unsigned i; + char buf[80]; BUG_ON(!parent); - closure_lock(&c->uuid_write, 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_rw = REQ_SYNC|REQ_META|rw; - bio->bi_size = KEY_SIZE(k) << 9; + bio->bi_iter.bi_size = KEY_SIZE(k) << 9; bio->bi_end_io = uuid_endio; bio->bi_private = cl; @@ -362,8 +384,8 @@ static void uuid_io(struct cache_set *c, unsigned long rw, break; } - pr_debug("%s UUIDs at %s", rw & REQ_WRITE ? "wrote" : "read", - pkey(&c->uuid_bucket)); + bch_extent_to_text(buf, sizeof(buf), k); + pr_debug("%s UUIDs at %s", rw & REQ_WRITE ? "wrote" : "read", buf); for (u = c->uuids; u < c->uuids + c->nr_uuids; u++) if (!bch_is_zero(u->uuid, 16)) @@ -371,14 +393,14 @@ static void uuid_io(struct cache_set *c, unsigned long rw, u - c->uuids, u->uuid, u->label, u->first_reg, u->last_reg, u->invalidated); - closure_return(cl); + 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_ptr_invalid(c, 1, k)) + if (__bch_btree_ptr_invalid(c, k)) return "bad uuid pointer"; bkey_copy(&c->uuid_bucket, k); @@ -423,7 +445,7 @@ static int __uuid_write(struct cache_set *c) lockdep_assert_held(&bch_register_lock); - if (bch_bucket_alloc_set(c, WATERMARK_METADATA, &k.key, 1, &cl)) + if (bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, 1, true)) return 1; SET_KEY_SIZE(&k.key, c->sb.bucket_size); @@ -431,7 +453,7 @@ static int __uuid_write(struct cache_set *c) closure_sync(&cl); bkey_copy(&c->uuid_bucket, &k.key); - __bkey_put(c, &k.key); + bkey_put(c, &k.key); return 0; } @@ -506,10 +528,10 @@ static void prio_io(struct cache *ca, uint64_t bucket, unsigned long rw) closure_init_stack(cl); - bio->bi_sector = bucket * ca->sb.bucket_size; - bio->bi_bdev = ca->bdev; - bio->bi_rw = REQ_SYNC|REQ_META|rw; - bio->bi_size = bucket_bytes(ca); + bio->bi_iter.bi_sector = bucket * ca->sb.bucket_size; + bio->bi_bdev = ca->bdev; + bio->bi_rw = REQ_SYNC|REQ_META|rw; + bio->bi_iter.bi_size = bucket_bytes(ca); bio->bi_end_io = prio_endio; bio->bi_private = ca; @@ -519,9 +541,6 @@ static void prio_io(struct cache *ca, uint64_t bucket, unsigned long rw) closure_sync(cl); } -#define buckets_free(c) "free %zu, free_inc %zu, unused %zu", \ - fifo_used(&c->free), fifo_used(&c->free_inc), fifo_used(&c->unused) - void bch_prio_write(struct cache *ca) { int i; @@ -532,18 +551,13 @@ void bch_prio_write(struct cache *ca) lockdep_assert_held(&ca->set->bucket_lock); - for (b = ca->buckets; - b < ca->buckets + ca->sb.nbuckets; b++) - b->disk_gen = b->gen; - ca->disk_buckets->seq++; atomic_long_add(ca->sb.bucket_size * prio_buckets(ca), &ca->meta_sectors_written); - pr_debug("free %zu, free_inc %zu, unused %zu", fifo_used(&ca->free), - fifo_used(&ca->free_inc), fifo_used(&ca->unused)); - blktrace_msg(ca, "Starting priorities: " buckets_free(ca)); + //pr_debug("free %zu, free_inc %zu, unused %zu", fifo_used(&ca->free), + // fifo_used(&ca->free_inc), fifo_used(&ca->unused)); for (i = prio_buckets(ca) - 1; i >= 0; --i) { long bucket; @@ -559,10 +573,10 @@ void bch_prio_write(struct cache *ca) } p->next_bucket = ca->prio_buckets[i + 1]; - p->magic = pset_magic(ca); + p->magic = pset_magic(&ca->sb); p->csum = bch_crc64(&p->magic, bucket_bytes(ca) - 8); - bucket = bch_bucket_alloc(ca, WATERMARK_PRIO, &cl); + bucket = bch_bucket_alloc(ca, RESERVE_PRIO, true); BUG_ON(bucket == -1); mutex_unlock(&ca->set->bucket_lock); @@ -580,14 +594,17 @@ void bch_prio_write(struct cache *ca) mutex_lock(&ca->set->bucket_lock); - ca->need_save_prio = 0; - /* * 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++) + 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]; + } } static void prio_read(struct cache *ca, uint64_t bucket) @@ -610,7 +627,7 @@ static void prio_read(struct cache *ca, uint64_t bucket) if (p->csum != bch_crc64(&p->magic, bucket_bytes(ca) - 8)) pr_warn("bad csum reading priorities"); - if (p->magic != pset_magic(ca)) + if (p->magic != pset_magic(&ca->sb)) pr_warn("bad magic reading priorities"); bucket = p->next_bucket; @@ -618,7 +635,7 @@ static void prio_read(struct cache *ca, uint64_t bucket) } b->prio = le16_to_cpu(d->prio); - b->gen = b->disk_gen = b->last_gc = b->gc_gen = d->gen; + b->gen = b->last_gc = d->gen; } } @@ -627,7 +644,7 @@ static void prio_read(struct cache *ca, uint64_t bucket) static int open_dev(struct block_device *b, fmode_t mode) { struct bcache_device *d = b->bd_disk->private_data; - if (atomic_read(&d->closing)) + if (test_bit(BCACHE_DEV_CLOSING, &d->flags)) return -ENXIO; closure_get(&d->cl); @@ -656,20 +673,24 @@ static const struct block_device_operations bcache_ops = { void bcache_device_stop(struct bcache_device *d) { - if (!atomic_xchg(&d->closing, 1)) + if (!test_and_set_bit(BCACHE_DEV_CLOSING, &d->flags)) closure_queue(&d->cl); } static void bcache_device_unlink(struct bcache_device *d) { - unsigned i; - struct cache *ca; + lockdep_assert_held(&bch_register_lock); - sysfs_remove_link(&d->c->kobj, d->name); - sysfs_remove_link(&d->kobj, "cache"); + if (d->c && !test_and_set_bit(BCACHE_DEV_UNLINK_DONE, &d->flags)) { + unsigned i; + struct cache *ca; - for_each_cache(ca, d->c, i) - bd_unlink_disk_holder(ca->bdev, d->disk); + sysfs_remove_link(&d->c->kobj, d->name); + sysfs_remove_link(&d->kobj, "cache"); + + for_each_cache(ca, d->c, i) + bd_unlink_disk_holder(ca->bdev, d->disk); + } } static void bcache_device_link(struct bcache_device *d, struct cache_set *c, @@ -693,15 +714,13 @@ static void bcache_device_detach(struct bcache_device *d) { lockdep_assert_held(&bch_register_lock); - if (atomic_read(&d->detaching)) { + 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(get_seconds()); bch_uuid_write(d->c); - - atomic_set(&d->detaching, 0); } bcache_device_unlink(d); @@ -735,37 +754,82 @@ static void bcache_device_free(struct bcache_device *d) del_gendisk(d->disk); if (d->disk && d->disk->queue) blk_cleanup_queue(d->disk->queue); - if (d->disk) + if (d->disk) { + ida_simple_remove(&bcache_minor, d->disk->first_minor); put_disk(d->disk); + } bio_split_pool_free(&d->bio_split_hook); - if (d->unaligned_bvec) - mempool_destroy(d->unaligned_bvec); if (d->bio_split) bioset_free(d->bio_split); + if (is_vmalloc_addr(d->full_dirty_stripes)) + vfree(d->full_dirty_stripes); + else + kfree(d->full_dirty_stripes); + if (is_vmalloc_addr(d->stripe_sectors_dirty)) + vfree(d->stripe_sectors_dirty); + else + kfree(d->stripe_sectors_dirty); closure_debug_destroy(&d->cl); } -static int bcache_device_init(struct bcache_device *d, unsigned block_size) +static int bcache_device_init(struct bcache_device *d, unsigned block_size, + sector_t sectors) { struct request_queue *q; + size_t n; + int minor; + + if (!d->stripe_size) + d->stripe_size = 1 << 31; + + d->nr_stripes = DIV_ROUND_UP_ULL(sectors, d->stripe_size); + + if (!d->nr_stripes || + d->nr_stripes > INT_MAX || + d->nr_stripes > SIZE_MAX / sizeof(atomic_t)) { + pr_err("nr_stripes too large"); + return -ENOMEM; + } + + n = d->nr_stripes * sizeof(atomic_t); + d->stripe_sectors_dirty = n < PAGE_SIZE << 6 + ? kzalloc(n, GFP_KERNEL) + : vzalloc(n); + if (!d->stripe_sectors_dirty) + return -ENOMEM; + + n = BITS_TO_LONGS(d->nr_stripes) * sizeof(unsigned long); + d->full_dirty_stripes = n < PAGE_SIZE << 6 + ? kzalloc(n, GFP_KERNEL) + : vzalloc(n); + if (!d->full_dirty_stripes) + return -ENOMEM; + + minor = ida_simple_get(&bcache_minor, 0, MINORMASK + 1, GFP_KERNEL); + if (minor < 0) + return minor; if (!(d->bio_split = bioset_create(4, offsetof(struct bbio, bio))) || - !(d->unaligned_bvec = mempool_create_kmalloc_pool(1, - sizeof(struct bio_vec) * BIO_MAX_PAGES)) || bio_split_pool_init(&d->bio_split_hook) || - !(d->disk = alloc_disk(1)) || - !(q = blk_alloc_queue(GFP_KERNEL))) + !(d->disk = alloc_disk(1))) { + ida_simple_remove(&bcache_minor, minor); return -ENOMEM; + } - snprintf(d->disk->disk_name, DISK_NAME_LEN, "bcache%i", bcache_minor); + set_capacity(d->disk, sectors); + snprintf(d->disk->disk_name, DISK_NAME_LEN, "bcache%i", minor); d->disk->major = bcache_major; - d->disk->first_minor = bcache_minor++; + d->disk->first_minor = minor; d->disk->fops = &bcache_ops; d->disk->private_data = d; + q = blk_alloc_queue(GFP_KERNEL); + if (!q) + return -ENOMEM; + blk_queue_make_request(q, NULL); d->disk->queue = q; q->queuedata = d; @@ -775,12 +839,15 @@ static int bcache_device_init(struct bcache_device *d, unsigned block_size) q->limits.max_segment_size = UINT_MAX; q->limits.max_segments = BIO_MAX_PAGES; q->limits.max_discard_sectors = 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; set_bit(QUEUE_FLAG_NONROT, &d->disk->queue->queue_flags); set_bit(QUEUE_FLAG_DISCARD, &d->disk->queue->queue_flags); + blk_queue_flush(q, REQ_FLUSH|REQ_FUA); + return 0; } @@ -800,6 +867,17 @@ static void calc_cached_dev_sectors(struct cache_set *c) void bch_cached_dev_run(struct cached_dev *dc) { struct bcache_device *d = &dc->disk; + char buf[SB_LABEL_SIZE + 1]; + char *env[] = { + "DRIVER=bcache", + kasprintf(GFP_KERNEL, "CACHED_UUID=%pU", dc->sb.uuid), + NULL, + NULL, + }; + + memcpy(buf, dc->sb.label, SB_LABEL_SIZE); + buf[SB_LABEL_SIZE] = '\0'; + env[2] = kasprintf(GFP_KERNEL, "CACHED_LABEL=%s", buf); if (atomic_xchg(&dc->running, 1)) return; @@ -816,10 +894,12 @@ void bch_cached_dev_run(struct cached_dev *dc) add_disk(d->disk); bd_link_disk_holder(dc->bdev, dc->disk.disk); -#if 0 - char *env[] = { "SYMLINK=label" , NULL }; + /* 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); -#endif + kfree(env[1]); + kfree(env[2]); + 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_debug("error creating sysfs link"); @@ -832,7 +912,7 @@ static void cached_dev_detach_finish(struct work_struct *w) struct closure cl; closure_init_stack(&cl); - BUG_ON(!atomic_read(&dc->disk.detaching)); + BUG_ON(!test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags)); BUG_ON(atomic_read(&dc->count)); mutex_lock(&bch_register_lock); @@ -846,6 +926,8 @@ static void cached_dev_detach_finish(struct work_struct *w) bcache_device_detach(&dc->disk); list_move(&dc->list, &uncached_devices); + clear_bit(BCACHE_DEV_DETACHING, &dc->disk.flags); + mutex_unlock(&bch_register_lock); pr_info("Caching disabled for %s", bdevname(dc->bdev, buf)); @@ -858,10 +940,10 @@ void bch_cached_dev_detach(struct cached_dev *dc) { lockdep_assert_held(&bch_register_lock); - if (atomic_read(&dc->disk.closing)) + if (test_bit(BCACHE_DEV_CLOSING, &dc->disk.flags)) return; - if (atomic_xchg(&dc->disk.detaching, 1)) + if (test_and_set_bit(BCACHE_DEV_DETACHING, &dc->disk.flags)) return; /* @@ -960,6 +1042,7 @@ int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c) atomic_set(&dc->count, 1); if (BDEV_STATE(&dc->sb) == BDEV_STATE_DIRTY) { + bch_sectors_dirty_init(dc); atomic_set(&dc->has_dirty, 1); atomic_inc(&dc->count); bch_writeback_queue(dc); @@ -987,6 +1070,7 @@ static void cached_dev_free(struct closure *cl) struct cached_dev *dc = container_of(cl, struct cached_dev, disk.cl); cancel_delayed_work_sync(&dc->writeback_rate_update); + kthread_stop(dc->writeback_thread); mutex_lock(&bch_register_lock); @@ -1014,6 +1098,10 @@ 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); @@ -1032,12 +1120,11 @@ static int cached_dev_init(struct cached_dev *dc, unsigned block_size) 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); - closure_init_unlocked(&dc->sb_write); + 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_merge = true; dc->sequential_cutoff = 4 << 20; for (io = dc->io; io < dc->io + RECENT_IO; io++) { @@ -1045,7 +1132,14 @@ static int cached_dev_init(struct cached_dev *dc, unsigned block_size) hlist_add_head(&io->hash, dc->io_hash + RECENT_IO); } - ret = bcache_device_init(&dc->disk, block_size); + 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); if (ret) return ret; @@ -1144,11 +1238,10 @@ static int flash_dev_run(struct cache_set *c, struct uuid_entry *u) kobject_init(&d->kobj, &bch_flash_dev_ktype); - if (bcache_device_init(d, block_bytes(c))) + if (bcache_device_init(d, block_bytes(c), u->sectors)) goto err; bcache_device_attach(d, c, u - c->uuids); - set_capacity(d->disk, u->sectors); bch_flash_dev_request_init(d); add_disk(d->disk); @@ -1209,7 +1302,8 @@ bool bch_cache_set_error(struct cache_set *c, const char *fmt, ...) { va_list args; - if (test_bit(CACHE_SET_STOPPING, &c->flags)) + if (c->on_error != ON_ERROR_PANIC && + test_bit(CACHE_SET_STOPPING, &c->flags)) return false; /* XXX: we can be called from atomic context @@ -1224,6 +1318,9 @@ bool bch_cache_set_error(struct cache_set *c, const char *fmt, ...) printk(", disabling caching\n"); + if (c->on_error == ON_ERROR_PANIC) + panic("panic forced after error\n"); + bch_cache_set_unregister(c); return true; } @@ -1252,12 +1349,15 @@ static void cache_set_free(struct closure *cl) if (ca) kobject_put(&ca->kobj); + bch_bset_sort_state_free(&c->sort); free_pages((unsigned long) c->uuids, ilog2(bucket_pages(c))); - free_pages((unsigned long) c->sort, ilog2(bucket_pages(c))); - kfree(c->fill_iter); + if (c->moving_gc_wq) + destroy_workqueue(c->moving_gc_wq); if (c->bio_split) bioset_free(c->bio_split); + if (c->fill_iter) + mempool_destroy(c->fill_iter); if (c->bio_meta) mempool_destroy(c->bio_meta); if (c->search) @@ -1278,24 +1378,36 @@ static void cache_set_free(struct closure *cl) static void cache_set_flush(struct closure *cl) { struct cache_set *c = container_of(cl, struct cache_set, caching); + struct cache *ca; struct btree *b; - - /* Shut down allocator threads */ - set_bit(CACHE_SET_STOPPING_2, &c->flags); - wake_up(&c->alloc_wait); + unsigned i; bch_cache_accounting_destroy(&c->accounting); kobject_put(&c->internal); kobject_del(&c->kobj); + if (c->gc_thread) + kthread_stop(c->gc_thread); + if (!IS_ERR_OR_NULL(c->root)) list_add(&c->root->list, &c->btree_cache); /* Should skip this if we're unregistering because of an error */ - list_for_each_entry(b, &c->btree_cache, list) + list_for_each_entry(b, &c->btree_cache, list) { + mutex_lock(&b->write_lock); if (btree_node_dirty(b)) - bch_btree_write(b, true, NULL); + __bch_btree_node_write(b, NULL); + mutex_unlock(&b->write_lock); + } + + for_each_cache(ca, c, i) + if (ca->alloc_thread) + kthread_stop(ca->alloc_thread); + + 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); } @@ -1303,18 +1415,22 @@ static void cache_set_flush(struct closure *cl) static void __cache_set_unregister(struct closure *cl) { struct cache_set *c = container_of(cl, struct cache_set, caching); - struct cached_dev *dc, *t; + struct cached_dev *dc; size_t i; mutex_lock(&bch_register_lock); - if (test_bit(CACHE_SET_UNREGISTERING, &c->flags)) - list_for_each_entry_safe(dc, t, &c->cached_devs, list) - bch_cached_dev_detach(dc); - for (i = 0; i < c->nr_uuids; i++) - if (c->devices[i] && UUID_FLASH_ONLY(&c->uuids[i])) - bcache_device_stop(c->devices[i]); + if (c->devices[i]) { + if (!UUID_FLASH_ONLY(&c->uuids[i]) && + test_bit(CACHE_SET_UNREGISTERING, &c->flags)) { + dc = container_of(c->devices[i], + struct cached_dev, disk); + bch_cached_dev_detach(dc); + } else { + bcache_device_stop(c->devices[i]); + } + } mutex_unlock(&bch_register_lock); @@ -1368,19 +1484,21 @@ struct cache_set *bch_cache_set_alloc(struct cache_sb *sb) c->block_bits = ilog2(sb->block_size); c->nr_uuids = bucket_bytes(c) / sizeof(struct uuid_entry); - c->btree_pages = c->sb.bucket_size / PAGE_SECTORS; + c->btree_pages = bucket_pages(c); if (c->btree_pages > BTREE_MAX_PAGES) c->btree_pages = max_t(int, c->btree_pages / 4, BTREE_MAX_PAGES); - init_waitqueue_head(&c->alloc_wait); + sema_init(&c->sb_write_mutex, 1); mutex_init(&c->bucket_lock); - mutex_init(&c->fill_lock); - mutex_init(&c->sort_lock); - spin_lock_init(&c->sort_time_lock); - closure_init_unlocked(&c->sb_write); - closure_init_unlocked(&c->uuid_write); - spin_lock_init(&c->btree_read_time_lock); + init_waitqueue_head(&c->btree_cache_wait); + init_waitqueue_head(&c->bucket_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); @@ -1401,17 +1519,16 @@ struct cache_set *bch_cache_set_alloc(struct cache_sb *sb) !(c->bio_meta = mempool_create_kmalloc_pool(2, sizeof(struct bbio) + sizeof(struct bio_vec) * bucket_pages(c))) || + !(c->fill_iter = mempool_create_kmalloc_pool(1, iter_size)) || !(c->bio_split = bioset_create(4, offsetof(struct bbio, bio))) || - !(c->fill_iter = kmalloc(iter_size, GFP_KERNEL)) || - !(c->sort = alloc_bucket_pages(GFP_KERNEL, c)) || !(c->uuids = alloc_bucket_pages(GFP_KERNEL, c)) || + !(c->moving_gc_wq = create_workqueue("bcache_gc")) || bch_journal_alloc(c) || bch_btree_cache_alloc(c) || - bch_open_buckets_alloc(c)) + bch_open_buckets_alloc(c) || + bch_bset_sort_state_init(&c->sort, ilog2(c->btree_pages))) goto err; - c->fill_iter->size = sb->bucket_size / sb->block_size; - c->congested_read_threshold_us = 2000; c->congested_write_threshold_us = 20000; c->error_limit = 8 << IO_ERROR_SHIFT; @@ -1427,11 +1544,10 @@ static void run_cache_set(struct cache_set *c) const char *err = "cannot allocate memory"; struct cached_dev *dc, *t; struct cache *ca; + struct closure cl; unsigned i; - struct btree_op op; - bch_btree_op_init_stack(&op); - op.lock = SHRT_MAX; + closure_init_stack(&cl); for_each_cache(ca, c, i) c->nbuckets += ca->sb.nbuckets; @@ -1442,7 +1558,7 @@ static void run_cache_set(struct cache_set *c) struct jset *j; err = "cannot allocate memory for journal"; - if (bch_journal_read(c, &journal, &op)) + if (bch_journal_read(c, &journal)) goto err; pr_debug("btree_journal_read() done"); @@ -1466,27 +1582,27 @@ static void run_cache_set(struct cache_set *c) k = &j->btree_root; err = "bad btree root"; - if (__bch_ptr_invalid(c, j->btree_level + 1, k)) + if (__bch_btree_ptr_invalid(c, k)) goto err; err = "error reading btree root"; - c->root = bch_btree_node_get(c, k, j->btree_level, &op); + c->root = bch_btree_node_get(c, NULL, k, j->btree_level, true); if (IS_ERR_OR_NULL(c->root)) goto err; list_del_init(&c->root->list); rw_unlock(true, c->root); - err = uuid_read(c, j, &op.cl); + err = uuid_read(c, j, &cl); if (err) goto err; err = "error in recovery"; - if (bch_btree_check(c, &op)) + if (bch_btree_check(c)) goto err; bch_journal_mark(c, &journal); - bch_btree_gc_finish(c); + bch_initial_gc_finish(c); pr_debug("btree_check() done"); /* @@ -1496,9 +1612,10 @@ static void run_cache_set(struct cache_set *c) */ bch_journal_next(&c->journal); + err = "error starting allocator thread"; for_each_cache(ca, c, i) - closure_call(&ca->alloc, bch_allocator_thread, - system_wq, &c->cl); + if (bch_cache_allocator_start(ca)) + goto err; /* * First place it's safe to allocate: btree_check() and @@ -1513,11 +1630,9 @@ static void run_cache_set(struct cache_set *c) if (j->version < BCACHE_JSET_VERSION_UUID) __uuid_write(c); - bch_journal_replay(c, &journal, &op); + bch_journal_replay(c, &journal); } else { pr_notice("invalidating existing data"); - /* Don't want invalidate_buckets() to queue a gc yet */ - closure_lock(&c->gc, NULL); for_each_cache(ca, c, i) { unsigned j; @@ -1529,30 +1644,31 @@ static void run_cache_set(struct cache_set *c) ca->sb.d[j] = ca->sb.first_bucket + j; } - bch_btree_gc_finish(c); + bch_initial_gc_finish(c); + err = "error starting allocator thread"; for_each_cache(ca, c, i) - closure_call(&ca->alloc, bch_allocator_thread, - ca->alloc_workqueue, &c->cl); + if (bch_cache_allocator_start(ca)) + goto err; mutex_lock(&c->bucket_lock); for_each_cache(ca, c, i) bch_prio_write(ca); mutex_unlock(&c->bucket_lock); - wake_up(&c->alloc_wait); - err = "cannot allocate new UUID bucket"; if (__uuid_write(c)) - goto err_unlock_gc; + goto err; err = "cannot allocate new btree root"; - c->root = bch_btree_node_alloc(c, 0, &op.cl); + c->root = bch_btree_node_alloc(c, NULL, 0); if (IS_ERR_OR_NULL(c->root)) - goto err_unlock_gc; + goto err; + mutex_lock(&c->root->write_lock); bkey_copy_key(&c->root->key, &MAX_KEY); - bch_btree_write(c->root, true, &op); + bch_btree_node_write(c->root, &cl); + mutex_unlock(&c->root->write_lock); bch_btree_set_root(c->root); rw_unlock(true, c->root); @@ -1565,14 +1681,14 @@ static void run_cache_set(struct cache_set *c) SET_CACHE_SYNC(&c->sb, true); bch_journal_next(&c->journal); - bch_journal_meta(c, &op.cl); - - /* Unlock */ - closure_set_stopped(&c->gc.cl); - closure_put(&c->gc.cl); + bch_journal_meta(c, &cl); } - closure_sync(&op.cl); + err = "error starting gc thread"; + if (bch_gc_thread_start(c)) + goto err; + + closure_sync(&cl); c->sb.last_mount = get_seconds(); bcache_write_super(c); @@ -1582,19 +1698,16 @@ static void run_cache_set(struct cache_set *c) flash_devs_run(c); return; -err_unlock_gc: - closure_set_stopped(&c->gc.cl); - closure_put(&c->gc.cl); err: - closure_sync(&op.cl); + closure_sync(&cl); /* XXX: test this, it's broken */ - bch_cache_set_error(c, err); + bch_cache_set_error(c, "%s", err); } static bool can_attach_cache(struct cache *ca, struct cache_set *c) { return ca->sb.block_size == c->sb.block_size && - ca->sb.bucket_size == c->sb.block_size && + ca->sb.bucket_size == c->sb.bucket_size && ca->sb.nr_in_set == c->sb.nr_in_set; } @@ -1665,25 +1778,22 @@ err: void bch_cache_release(struct kobject *kobj) { struct cache *ca = container_of(kobj, struct cache, kobj); + unsigned i; if (ca->set) ca->set->cache[ca->sb.nr_this_dev] = NULL; - bch_cache_allocator_exit(ca); - bio_split_pool_free(&ca->bio_split_hook); - if (ca->alloc_workqueue) - destroy_workqueue(ca->alloc_workqueue); - free_pages((unsigned long) ca->disk_buckets, ilog2(bucket_pages(ca))); kfree(ca->prio_buckets); vfree(ca->buckets); free_heap(&ca->heap); - free_fifo(&ca->unused); free_fifo(&ca->free_inc); - free_fifo(&ca->free); + + for (i = 0; i < RESERVE_NR; i++) + free_fifo(&ca->free[i]); if (ca->sb_bio.bi_inline_vecs[0].bv_page) put_page(ca->sb_bio.bi_io_vec[0].bv_page); @@ -1705,25 +1815,23 @@ static int cache_alloc(struct cache_sb *sb, struct cache *ca) __module_get(THIS_MODULE); kobject_init(&ca->kobj, &bch_cache_ktype); - INIT_LIST_HEAD(&ca->discards); - bio_init(&ca->journal.bio); ca->journal.bio.bi_max_vecs = 8; ca->journal.bio.bi_io_vec = ca->journal.bio.bi_inline_vecs; - free = roundup_pow_of_two(ca->sb.nbuckets) >> 9; - free = max_t(size_t, free, (prio_buckets(ca) + 8) * 2); + free = roundup_pow_of_two(ca->sb.nbuckets) >> 10; - if (!init_fifo(&ca->free, free, GFP_KERNEL) || + if (!init_fifo(&ca->free[RESERVE_BTREE], 8, GFP_KERNEL) || + !init_fifo(&ca->free[RESERVE_PRIO], prio_buckets(ca), GFP_KERNEL) || + !init_fifo(&ca->free[RESERVE_MOVINGGC], free, GFP_KERNEL) || + !init_fifo(&ca->free[RESERVE_NONE], free, GFP_KERNEL) || !init_fifo(&ca->free_inc, free << 2, GFP_KERNEL) || - !init_fifo(&ca->unused, free << 2, GFP_KERNEL) || !init_heap(&ca->heap, free << 3, GFP_KERNEL) || !(ca->buckets = vzalloc(sizeof(struct bucket) * ca->sb.nbuckets)) || !(ca->prio_buckets = kzalloc(sizeof(uint64_t) * prio_buckets(ca) * 2, GFP_KERNEL)) || !(ca->disk_buckets = alloc_bucket_pages(GFP_KERNEL, ca)) || - !(ca->alloc_workqueue = alloc_workqueue("bch_allocator", 0, 1)) || bio_split_pool_init(&ca->bio_split_hook)) return -ENOMEM; @@ -1732,13 +1840,7 @@ static int cache_alloc(struct cache_sb *sb, struct cache *ca) for_each_bucket(b, ca) atomic_set(&b->pin, 0); - if (bch_cache_allocator_init(ca)) - goto err; - return 0; -err: - kobject_put(&ca->kobj); - return -ENOMEM; } static void register_cache(struct cache_sb *sb, struct page *sb_page, @@ -1767,7 +1869,10 @@ static void register_cache(struct cache_sb *sb, struct page *sb_page, if (kobject_add(&ca->kobj, &part_to_dev(bdev->bd_part)->kobj, "bcache")) goto err; + mutex_lock(&bch_register_lock); err = register_cache_set(ca); + mutex_unlock(&bch_register_lock); + if (err) goto err; @@ -1786,6 +1891,36 @@ static ssize_t register_bcache(struct kobject *, struct kobj_attribute *, kobj_attribute_write(register, register_bcache); kobj_attribute_write(register_quiet, register_bcache); +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; + struct cache *ca; + unsigned i; + + list_for_each_entry_safe(c, tc, &bch_cache_sets, list) + for_each_cache(ca, c, i) + 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); +} + static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr, const char *buffer, size_t size) { @@ -1799,8 +1934,6 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr, if (!try_module_get(THIS_MODULE)) return -EBUSY; - mutex_lock(&bch_register_lock); - if (!(path = kstrndup(buffer, size, GFP_KERNEL)) || !(sb = kmalloc(sizeof(struct cache_sb), GFP_KERNEL))) goto err; @@ -1810,8 +1943,13 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr, FMODE_READ|FMODE_WRITE|FMODE_EXCL, sb); if (IS_ERR(bdev)) { - if (bdev == ERR_PTR(-EBUSY)) - err = "device busy"; + if (bdev == ERR_PTR(-EBUSY)) { + bdev = lookup_bdev(strim(path)); + if (!IS_ERR(bdev) && bch_is_open(bdev)) + err = "device already registered"; + else + err = "device busy"; + } goto err; } @@ -1828,7 +1966,9 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr, if (!dc) goto err_close; + mutex_lock(&bch_register_lock); register_bdev(sb, sb_page, bdev, dc); + mutex_unlock(&bch_register_lock); } else { struct cache *ca = kzalloc(sizeof(*ca), GFP_KERNEL); if (!ca) @@ -1841,7 +1981,6 @@ out: put_page(sb_page); kfree(sb); kfree(path); - mutex_unlock(&bch_register_lock); module_put(THIS_MODULE); return ret; @@ -1919,14 +2058,13 @@ static struct notifier_block reboot = { static void bcache_exit(void) { bch_debug_exit(); - bch_writeback_exit(); bch_request_exit(); - bch_btree_exit(); if (bcache_kobj) kobject_put(bcache_kobj); if (bcache_wq) destroy_workqueue(bcache_wq); - unregister_blkdev(bcache_major, "bcache"); + if (bcache_major) + unregister_blkdev(bcache_major, "bcache"); unregister_reboot_notifier(&reboot); } @@ -1950,9 +2088,7 @@ static int __init bcache_init(void) if (!(bcache_wq = create_workqueue("bcache")) || !(bcache_kobj = kobject_create_and_add("bcache", fs_kobj)) || sysfs_create_files(bcache_kobj, files) || - bch_btree_init() || bch_request_init() || - bch_writeback_init() || bch_debug_init(bcache_kobj)) goto err; diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c index 4d9cca47e4c..b3ff57d61dd 100644 --- a/drivers/md/bcache/sysfs.c +++ b/drivers/md/bcache/sysfs.c @@ -9,7 +9,9 @@ #include "sysfs.h" #include "btree.h" #include "request.h" +#include "writeback.h" +#include <linux/blkdev.h> #include <linux/sort.h> static const char * const cache_replacement_policies[] = { @@ -19,6 +21,12 @@ static const char * const cache_replacement_policies[] = { NULL }; +static const char * const error_actions[] = { + "unregister", + "panic", + NULL +}; + write_attribute(attach); write_attribute(detach); write_attribute(unregister); @@ -46,7 +54,6 @@ 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); -sysfs_time_stats_attribute(try_harder, ms, us); read_attribute(btree_nodes); read_attribute(btree_used_percent); @@ -64,7 +71,6 @@ rw_attribute(congested_read_threshold_us); rw_attribute(congested_write_threshold_us); rw_attribute(sequential_cutoff); -rw_attribute(sequential_merge); rw_attribute(data_csum); rw_attribute(cache_mode); rw_attribute(writeback_metadata); @@ -76,21 +82,25 @@ rw_attribute(writeback_rate); rw_attribute(writeback_rate_update_seconds); rw_attribute(writeback_rate_d_term); rw_attribute(writeback_rate_p_term_inverse); -rw_attribute(writeback_rate_d_smooth); read_attribute(writeback_rate_debug); +read_attribute(stripe_size); +read_attribute(partial_stripes_expensive); + rw_attribute(synchronous); rw_attribute(journal_delay_ms); 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(freelist_percent); +rw_attribute(expensive_debug_checks); rw_attribute(cache_replacement_policy); rw_attribute(btree_shrinker_disabled); rw_attribute(copy_gc_enabled); @@ -111,41 +121,54 @@ SHOW(__bch_cached_dev) 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_print(writeback_rate, dc->writeback_rate.rate); + sysfs_hprint(writeback_rate, dc->writeback_rate.rate << 9); var_print(writeback_rate_update_seconds); var_print(writeback_rate_d_term); var_print(writeback_rate_p_term_inverse); - var_print(writeback_rate_d_smooth); if (attr == &sysfs_writeback_rate_debug) { + char rate[20]; char dirty[20]; - char derivative[20]; char target[20]; - bch_hprint(dirty, - atomic_long_read(&dc->disk.sectors_dirty) << 9); - bch_hprint(derivative, dc->writeback_rate_derivative << 9); + char proportional[20]; + char derivative[20]; + char change[20]; + s64 next_io; + + bch_hprint(rate, dc->writeback_rate.rate << 9); + bch_hprint(dirty, bcache_dev_sectors_dirty(&dc->disk) << 9); bch_hprint(target, dc->writeback_rate_target << 9); + bch_hprint(proportional,dc->writeback_rate_proportional << 9); + bch_hprint(derivative, dc->writeback_rate_derivative << 9); + bch_hprint(change, dc->writeback_rate_change << 9); + + next_io = div64_s64(dc->writeback_rate.next - local_clock(), + NSEC_PER_MSEC); return sprintf(buf, - "rate:\t\t%u\n" - "change:\t\t%i\n" + "rate:\t\t%s/sec\n" "dirty:\t\t%s\n" + "target:\t\t%s\n" + "proportional:\t%s\n" "derivative:\t%s\n" - "target:\t\t%s\n", - dc->writeback_rate.rate, - dc->writeback_rate_change, - dirty, derivative, target); + "change:\t\t%s/sec\n" + "next io:\t%llims\n", + rate, dirty, target, proportional, + derivative, change, next_io); } sysfs_hprint(dirty_data, - atomic_long_read(&dc->disk.sectors_dirty) << 9); + bcache_dev_sectors_dirty(&dc->disk) << 9); + + sysfs_hprint(stripe_size, dc->disk.stripe_size << 9); + var_printf(partial_stripes_expensive, "%u"); - var_printf(sequential_merge, "%i"); var_hprint(sequential_cutoff); var_hprint(readahead); @@ -170,27 +193,28 @@ STORE(__cached_dev) disk.kobj); unsigned v = size; struct cache_set *c; + struct kobj_uevent_env *env; #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); + d_strtoul(bypass_torture_test); d_strtoul(writeback_metadata); d_strtoul(writeback_running); d_strtoul(writeback_delay); - sysfs_strtoul_clamp(writeback_rate, - dc->writeback_rate.rate, 1, 1000000); + sysfs_strtoul_clamp(writeback_percent, dc->writeback_percent, 0, 40); - d_strtoul(writeback_rate_update_seconds); + sysfs_strtoul_clamp(writeback_rate, + dc->writeback_rate.rate, 1, INT_MAX); + + d_strtoul_nonzero(writeback_rate_update_seconds); d_strtoul(writeback_rate_d_term); - d_strtoul(writeback_rate_p_term_inverse); - sysfs_strtoul_clamp(writeback_rate_p_term_inverse, - dc->writeback_rate_p_term_inverse, 1, INT_MAX); - d_strtoul(writeback_rate_d_smooth); + d_strtoul_nonzero(writeback_rate_p_term_inverse); - d_strtoul(sequential_merge); d_strtoi_h(sequential_cutoff); d_strtoi_h(readahead); @@ -214,13 +238,28 @@ STORE(__cached_dev) } if (attr == &sysfs_label) { - memcpy(dc->sb.label, buf, SB_LABEL_SIZE); + 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) { @@ -281,11 +320,11 @@ static struct attribute *bch_cached_dev_files[] = { &sysfs_writeback_rate_update_seconds, &sysfs_writeback_rate_d_term, &sysfs_writeback_rate_p_term_inverse, - &sysfs_writeback_rate_d_smooth, &sysfs_writeback_rate_debug, &sysfs_dirty_data, + &sysfs_stripe_size, + &sysfs_partial_stripes_expensive, &sysfs_sequential_cutoff, - &sysfs_sequential_merge, &sysfs_clear_stats, &sysfs_running, &sysfs_state, @@ -293,6 +332,7 @@ static struct attribute *bch_cached_dev_files[] = { &sysfs_readahead, #ifdef CONFIG_BCACHE_DEBUG &sysfs_verify, + &sysfs_bypass_torture_test, #endif NULL }; @@ -340,7 +380,7 @@ STORE(__bch_flash_dev) } if (attr == &sysfs_unregister) { - atomic_set(&d->detaching, 1); + set_bit(BCACHE_DEV_DETACHING, &d->flags); bcache_device_stop(d); } @@ -359,81 +399,123 @@ static struct attribute *bch_flash_dev_files[] = { }; KTYPE(bch_flash_dev); -SHOW(__bch_cache_set) +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) { - unsigned root_usage(struct cache_set *c) - { - unsigned bytes = 0; - struct bkey *k; - struct btree *b; - struct btree_iter iter; + struct bset_stats_op *op = container_of(b_op, struct bset_stats_op, op); - goto lock_root; + op->nodes++; + bch_btree_keys_stats(&b->keys, &op->stats); - do { - rw_unlock(false, b); -lock_root: - b = c->root; - rw_lock(false, b, b->level); - } while (b != c->root); + 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); - for_each_key_filter(b, k, &iter, bch_ptr_bad) - bytes += bkey_bytes(k); + 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 bch_root_usage(struct cache_set *c) +{ + unsigned 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); - return (bytes * 100) / btree_bytes(c); - } + for_each_key_filter(&b->keys, k, &iter, bch_ptr_bad) + bytes += bkey_bytes(k); - size_t cache_size(struct cache_set *c) - { - size_t ret = 0; - struct btree *b; + rw_unlock(false, b); - mutex_lock(&c->bucket_lock); - list_for_each_entry(b, &c->btree_cache, list) - ret += 1 << (b->page_order + PAGE_SHIFT); + return (bytes * 100) / btree_bytes(c); +} - mutex_unlock(&c->bucket_lock); - return ret; - } +static size_t bch_cache_size(struct cache_set *c) +{ + size_t ret = 0; + struct btree *b; - unsigned cache_max_chain(struct cache_set *c) - { - unsigned ret = 0; - struct hlist_head *h; + 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; +} - mutex_lock(&c->bucket_lock); +static unsigned bch_cache_max_chain(struct cache_set *c) +{ + unsigned ret = 0; + struct hlist_head *h; - for (h = c->bucket_hash; - h < c->bucket_hash + (1 << BUCKET_HASH_BITS); - h++) { - unsigned i = 0; - struct hlist_node *p; + mutex_lock(&c->bucket_lock); - hlist_for_each(p, h) - i++; + for (h = c->bucket_hash; + h < c->bucket_hash + (1 << BUCKET_HASH_BITS); + h++) { + unsigned i = 0; + struct hlist_node *p; - ret = max(ret, i); - } + hlist_for_each(p, h) + i++; - mutex_unlock(&c->bucket_lock); - return ret; + ret = max(ret, i); } - unsigned btree_used(struct cache_set *c) - { - return div64_u64(c->gc_stats.key_bytes * 100, - (c->gc_stats.nodes ?: 1) * btree_bytes(c)); - } + mutex_unlock(&c->bucket_lock); + return ret; +} - unsigned average_key_size(struct cache_set *c) - { - return c->gc_stats.nkeys - ? div64_u64(c->gc_stats.data, c->gc_stats.nkeys) - : 0; - } +static unsigned 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 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->sb)); @@ -441,22 +523,20 @@ lock_root: sysfs_hprint(bucket_size, bucket_bytes(c)); sysfs_hprint(block_size, block_bytes(c)); sysfs_print(tree_depth, c->root->level); - sysfs_print(root_usage_percent, root_usage(c)); + sysfs_print(root_usage_percent, bch_root_usage(c)); - sysfs_hprint(btree_cache_size, cache_size(c)); - sysfs_print(btree_cache_max_chain, cache_max_chain(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->sort.time, btree_sort, ms, us); sysfs_print_time_stats(&c->btree_read_time, btree_read, ms, us); - sysfs_print_time_stats(&c->try_harder_time, try_harder, ms, us); - sysfs_print(btree_used_percent, btree_used(c)); + sysfs_print(btree_used_percent, bch_btree_used(c)); sysfs_print(btree_nodes, c->gc_stats.nodes); - sysfs_hprint(dirty_data, c->gc_stats.dirty); - sysfs_hprint(average_key_size, average_key_size(c)); + sysfs_hprint(average_key_size, bch_average_key_size(c)); sysfs_print(cache_read_races, atomic_long_read(&c->cache_read_races)); @@ -466,6 +546,10 @@ lock_root: 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 >> IO_ERROR_SHIFT); @@ -480,6 +564,8 @@ lock_root: 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); @@ -529,13 +615,13 @@ STORE(__bch_cache_set) } if (attr == &sysfs_trigger_gc) - bch_queue_gc(c); + 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.shrink(&c->shrink, &sc); + c->shrink.scan_objects(&c->shrink, &sc); } sysfs_strtoul(congested_read_threshold_us, @@ -543,6 +629,15 @@ STORE(__bch_cache_set) sysfs_strtoul(congested_write_threshold_us, c->congested_write_threshold_us); + if (attr == &sysfs_errors) { + ssize_t v = bch_read_string_list(buf, error_actions); + + if (v < 0) + return v; + + c->on_error = v; + } + if (attr == &sysfs_io_error_limit) c->error_limit = strtoul_or_return(buf) << IO_ERROR_SHIFT; @@ -553,6 +648,7 @@ STORE(__bch_cache_set) sysfs_strtoul(journal_delay_ms, c->journal_delay_ms); sysfs_strtoul(verify, c->verify); sysfs_strtoul(key_merging_disabled, c->key_merging_disabled); + sysfs_strtoul(expensive_debug_checks, c->expensive_debug_checks); sysfs_strtoul(gc_always_rewrite, c->gc_always_rewrite); sysfs_strtoul(btree_shrinker_disabled, c->shrinker_disabled); sysfs_strtoul(copy_gc_enabled, c->copy_gc_enabled); @@ -592,8 +688,8 @@ static struct attribute *bch_cache_set_files[] = { &sysfs_cache_available_percent, &sysfs_average_key_size, - &sysfs_dirty_data, + &sysfs_errors, &sysfs_io_error_limit, &sysfs_io_error_halflife, &sysfs_congested, @@ -611,7 +707,6 @@ static struct attribute *bch_cache_set_internal_files[] = { 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_time_stats_attribute_list(try_harder, ms, us) &sysfs_btree_nodes, &sysfs_btree_used_percent, @@ -627,6 +722,7 @@ static struct attribute *bch_cache_set_internal_files[] = { #ifdef CONFIG_BCACHE_DEBUG &sysfs_verify, &sysfs_key_merging_disabled, + &sysfs_expensive_debug_checks, #endif &sysfs_gc_always_rewrite, &sysfs_btree_shrinker_disabled, @@ -653,9 +749,6 @@ SHOW(__bch_cache) sysfs_print(io_errors, atomic_read(&ca->io_errors) >> IO_ERROR_SHIFT); - sysfs_print(freelist_percent, ca->free.size * 100 / - ((size_t) ca->sb.nbuckets)); - if (attr == &sysfs_cache_replacement_policy) return bch_snprint_string_list(buf, PAGE_SIZE, cache_replacement_policies, @@ -665,12 +758,12 @@ SHOW(__bch_cache) int cmp(const void *l, const void *r) { return *((uint16_t *) r) - *((uint16_t *) l); } - /* Number of quantiles we compute */ - const unsigned nq = 31; - - size_t n = ca->sb.nbuckets, i, unused, btree; + struct bucket *b; + size_t n = ca->sb.nbuckets, i; + size_t unused = 0, available = 0, dirty = 0, meta = 0; uint64_t sum = 0; - uint16_t q[nq], *p, *cached; + /* Compute 31 quantiles */ + uint16_t q[31], *p, *cached; ssize_t ret; cached = p = vmalloc(ca->sb.nbuckets * sizeof(uint16_t)); @@ -678,6 +771,17 @@ SHOW(__bch_cache) 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); @@ -692,10 +796,7 @@ SHOW(__bch_cache) while (cached < p + n && *cached == BTREE_PRIO) - cached++; - - btree = cached - p; - n -= btree; + cached++, n--; for (i = 0; i < n; i++) sum += INITIAL_PRIO - cached[i]; @@ -703,26 +804,33 @@ SHOW(__bch_cache) if (n) do_div(sum, n); - for (i = 0; i < nq; i++) - q[i] = INITIAL_PRIO - cached[n * (i + 1) / (nq + 1)]; + for (i = 0; i < ARRAY_SIZE(q); i++) + q[i] = INITIAL_PRIO - cached[n * (i + 1) / + (ARRAY_SIZE(q) + 1)]; vfree(p); - ret = snprintf(buf, PAGE_SIZE, - "Unused: %zu%%\n" - "Metadata: %zu%%\n" - "Average: %llu\n" - "Sectors per Q: %zu\n" - "Quantiles: [", - unused * 100 / (size_t) ca->sb.nbuckets, - btree * 100 / (size_t) ca->sb.nbuckets, sum, - n * ca->sb.bucket_size / (nq + 1)); - - for (i = 0; i < nq && ret < (ssize_t) PAGE_SIZE; i++) - ret += snprintf(buf + ret, PAGE_SIZE - ret, - i < nq - 1 ? "%u " : "%u]\n", q[i]); - - buf[PAGE_SIZE - 1] = '\0'; + 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; } @@ -761,32 +869,6 @@ STORE(__bch_cache) } } - if (attr == &sysfs_freelist_percent) { - DECLARE_FIFO(long, free); - long i; - size_t p = strtoul_or_return(buf); - - p = clamp_t(size_t, - ((size_t) ca->sb.nbuckets * p) / 100, - roundup_pow_of_two(ca->sb.nbuckets) >> 9, - ca->sb.nbuckets / 2); - - if (!init_fifo_exact(&free, p, GFP_KERNEL)) - return -ENOMEM; - - mutex_lock(&ca->set->bucket_lock); - - fifo_move(&free, &ca->free); - fifo_swap(&free, &ca->free); - - mutex_unlock(&ca->set->bucket_lock); - - while (fifo_pop(&free, i)) - atomic_dec(&ca->buckets[i].pin); - - free_fifo(&free); - } - if (attr == &sysfs_clear_stats) { atomic_long_set(&ca->sectors_written, 0); atomic_long_set(&ca->btree_sectors_written, 0); @@ -810,7 +892,6 @@ static struct attribute *bch_cache_files[] = { &sysfs_metadata_written, &sysfs_io_errors, &sysfs_clear_stats, - &sysfs_freelist_percent, &sysfs_cache_replacement_policy, NULL }; diff --git a/drivers/md/bcache/trace.c b/drivers/md/bcache/trace.c index 983f9bb411b..b7820b0d262 100644 --- a/drivers/md/bcache/trace.c +++ b/drivers/md/bcache/trace.c @@ -1,7 +1,7 @@ #include "bcache.h" #include "btree.h" -#include "request.h" +#include <linux/blktrace_api.h> #include <linux/module.h> #define CREATE_TRACE_POINTS @@ -9,18 +9,44 @@ EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_request_start); EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_request_end); -EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_passthrough); -EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_cache_hit); -EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_cache_miss); + +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_writethrough); -EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_writeback); -EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_write_skip); + +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_write_dirty); -EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_read_dirty); -EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_journal_write); -EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_cache_insert); + +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 index da3a99e85b1..db3ae4c2b22 100644 --- a/drivers/md/bcache/util.c +++ b/drivers/md/bcache/util.c @@ -168,10 +168,14 @@ int bch_parse_uuid(const char *s, char *uuid) void bch_time_stats_update(struct time_stats *stats, uint64_t start_time) { - uint64_t now = local_clock(); - uint64_t duration = time_after64(now, start_time) + uint64_t now, duration, last; + + spin_lock(&stats->lock); + + now = local_clock(); + duration = time_after64(now, start_time) ? now - start_time : 0; - uint64_t last = time_after64(now, stats->last) + last = time_after64(now, stats->last) ? now - stats->last : 0; stats->max_duration = max(stats->max_duration, duration); @@ -188,13 +192,30 @@ void bch_time_stats_update(struct time_stats *stats, uint64_t start_time) } stats->last = now ?: 1; + + spin_unlock(&stats->lock); } -unsigned bch_next_delay(struct ratelimit *d, uint64_t done) +/** + * bch_next_delay() - increment @d by the amount of work done, and return how + * long to delay until the next time to do some work. + * + * @d - the struct bch_ratelimit to update + * @done - the amount of work done, in arbitrary units + * + * Returns the amount of time to delay by, in jiffies + */ +uint64_t bch_next_delay(struct bch_ratelimit *d, uint64_t done) { uint64_t now = local_clock(); - d->next += div_u64(done, d->rate); + d->next += div_u64(done * NSEC_PER_SEC, d->rate); + + if (time_before64(now + NSEC_PER_SEC, d->next)) + d->next = now + NSEC_PER_SEC; + + 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) @@ -203,10 +224,10 @@ unsigned bch_next_delay(struct ratelimit *d, uint64_t done) void bch_bio_map(struct bio *bio, void *base) { - size_t size = bio->bi_size; + size_t size = bio->bi_iter.bi_size; struct bio_vec *bv = bio->bi_io_vec; - BUG_ON(!bio->bi_size); + BUG_ON(!bio->bi_iter.bi_size); BUG_ON(bio->bi_vcnt); bv->bv_offset = base ? ((unsigned long) base) % PAGE_SIZE : 0; @@ -228,23 +249,6 @@ start: bv->bv_len = min_t(size_t, PAGE_SIZE - bv->bv_offset, } } -int bch_bio_alloc_pages(struct bio *bio, gfp_t gfp) -{ - int i; - struct bio_vec *bv; - - bio_for_each_segment(bv, bio, i) { - bv->bv_page = alloc_page(gfp); - if (!bv->bv_page) { - while (bv-- != bio->bi_io_vec + bio->bi_idx) - __free_page(bv->bv_page); - return -ENOMEM; - } - } - - return 0; -} - /* * Portions Copyright (c) 1996-2001, PostgreSQL Global Development Group (Any * use permitted, subject to terms of PostgreSQL license; see.) diff --git a/drivers/md/bcache/util.h b/drivers/md/bcache/util.h index 577393e38c3..ac7d0d1f70d 100644 --- a/drivers/md/bcache/util.h +++ b/drivers/md/bcache/util.h @@ -2,6 +2,7 @@ #ifndef _BCACHE_UTIL_H #define _BCACHE_UTIL_H +#include <linux/blkdev.h> #include <linux/errno.h> #include <linux/kernel.h> #include <linux/llist.h> @@ -15,30 +16,20 @@ struct closure; -#include <trace/events/bcache.h> - -#ifdef CONFIG_BCACHE_EDEBUG +#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 /* EDEBUG */ +#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 BITMASK(name, type, field, offset, size) \ -static inline uint64_t name(const type *k) \ -{ return (k->field >> offset) & ~(((uint64_t) ~0) << size); } \ - \ -static inline void SET_##name(type *k, uint64_t v) \ -{ \ - k->field &= ~(~((uint64_t) ~0 << size) << offset); \ - k->field |= v << offset; \ -} - #define DECLARE_HEAP(type, name) \ struct { \ size_t size, used; \ @@ -122,7 +113,7 @@ do { \ _r; \ }) -#define heap_peek(h) ((h)->size ? (h)->data[0] : NULL) +#define heap_peek(h) ((h)->used ? (h)->data[0] : NULL) #define heap_full(h) ((h)->used == (h)->size) @@ -390,6 +381,7 @@ ssize_t bch_snprint_string_list(char *buf, size_t size, const char * const list[ ssize_t bch_read_string_list(const char *buf, const char * const list[]); struct time_stats { + spinlock_t lock; /* * all fields are in nanoseconds, averages are ewmas stored left shifted * by 8 @@ -402,6 +394,11 @@ struct time_stats { void bch_time_stats_update(struct time_stats *stats, uint64_t time); +static inline unsigned 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 @@ -452,17 +449,23 @@ read_attribute(name ## _last_ ## frequency_units) (ewma) >> factor; \ }) -struct ratelimit { +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 nanosecond + * The units here correspond to the units passed to bch_next_delay() + */ unsigned rate; }; -static inline void ratelimit_reset(struct ratelimit *d) +static inline void bch_ratelimit_reset(struct bch_ratelimit *d) { d->next = local_clock(); } -unsigned bch_next_delay(struct ratelimit *d, uint64_t done); +uint64_t bch_next_delay(struct bch_ratelimit *d, uint64_t done); #define __DIV_SAFE(n, d, zero) \ ({ \ @@ -566,12 +569,8 @@ static inline unsigned fract_exp_two(unsigned x, unsigned fract_bits) return x; } -#define bio_end(bio) ((bio)->bi_sector + bio_sectors(bio)) - void bch_bio_map(struct bio *bio, void *base); -int bch_bio_alloc_pages(struct bio *bio, gfp_t gfp); - static inline sector_t bdev_sectors(struct block_device *bdev) { return bdev->bd_inode->i_size >> 9; diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c index 2714ed3991d..f4300e4c011 100644 --- a/drivers/md/bcache/writeback.c +++ b/drivers/md/bcache/writeback.c @@ -9,16 +9,12 @@ #include "bcache.h" #include "btree.h" #include "debug.h" +#include "writeback.h" -static struct workqueue_struct *dirty_wq; - -static void read_dirty(struct closure *); - -struct dirty_io { - struct closure cl; - struct cached_dev *dc; - struct bio bio; -}; +#include <linux/delay.h> +#include <linux/freezer.h> +#include <linux/kthread.h> +#include <trace/events/bcache.h> /* Rate limiting */ @@ -34,44 +30,43 @@ static void __update_writeback_rate(struct cached_dev *dc) /* PD controller */ - int change = 0; - int64_t error; - int64_t dirty = atomic_long_read(&dc->disk.sectors_dirty); + int64_t dirty = bcache_dev_sectors_dirty(&dc->disk); int64_t derivative = dirty - dc->disk.sectors_dirty_last; + int64_t proportional = dirty - target; + int64_t change; dc->disk.sectors_dirty_last = dirty; - derivative *= dc->writeback_rate_d_term; - derivative = clamp(derivative, -dirty, dirty); + /* Scale to sectors per second */ - derivative = ewma_add(dc->disk.sectors_dirty_derivative, derivative, - dc->writeback_rate_d_smooth, 0); + proportional *= dc->writeback_rate_update_seconds; + proportional = div_s64(proportional, dc->writeback_rate_p_term_inverse); - /* Avoid divide by zero */ - if (!target) - goto out; + derivative = div_s64(derivative, dc->writeback_rate_update_seconds); + + derivative = ewma_add(dc->disk.sectors_dirty_derivative, derivative, + (dc->writeback_rate_d_term / + dc->writeback_rate_update_seconds) ?: 1, 0); - error = div64_s64((dirty + derivative - target) << 8, target); + derivative *= dc->writeback_rate_d_term; + derivative = div_s64(derivative, dc->writeback_rate_p_term_inverse); - change = div_s64((dc->writeback_rate.rate * error) >> 8, - dc->writeback_rate_p_term_inverse); + change = proportional + derivative; /* Don't increase writeback rate if the device isn't keeping up */ if (change > 0 && time_after64(local_clock(), - dc->writeback_rate.next + 10 * NSEC_PER_MSEC)) + dc->writeback_rate.next + NSEC_PER_MSEC)) change = 0; dc->writeback_rate.rate = - clamp_t(int64_t, dc->writeback_rate.rate + change, + clamp_t(int64_t, (int64_t) dc->writeback_rate.rate + change, 1, NSEC_PER_MSEC); -out: + + dc->writeback_rate_proportional = proportional; dc->writeback_rate_derivative = derivative; dc->writeback_rate_change = change; dc->writeback_rate_target = target; - - schedule_delayed_work(&dc->writeback_rate_update, - dc->writeback_rate_update_seconds * HZ); } static void update_writeback_rate(struct work_struct *work) @@ -87,23 +82,25 @@ static void update_writeback_rate(struct work_struct *work) __update_writeback_rate(dc); up_read(&dc->writeback_lock); + + schedule_delayed_work(&dc->writeback_rate_update, + dc->writeback_rate_update_seconds * HZ); } static unsigned writeback_delay(struct cached_dev *dc, unsigned sectors) { - if (atomic_read(&dc->disk.detaching) || + if (test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags) || !dc->writeback_percent) return 0; - return bch_next_delay(&dc->writeback_rate, sectors * 10000000ULL); + return bch_next_delay(&dc->writeback_rate, sectors); } -/* Background writeback */ - -static bool dirty_pred(struct keybuf *buf, struct bkey *k) -{ - return KEY_DIRTY(k); -} +struct dirty_io { + struct closure cl; + struct cached_dev *dc; + struct bio bio; +}; static void dirty_init(struct keybuf_key *w) { @@ -114,97 +111,13 @@ static void dirty_init(struct keybuf_key *w) if (!io->dc->writeback_percent) bio_set_prio(bio, IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0)); - bio->bi_size = KEY_SIZE(&w->key) << 9; + bio->bi_iter.bi_size = KEY_SIZE(&w->key) << 9; bio->bi_max_vecs = DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS); bio->bi_private = w; bio->bi_io_vec = bio->bi_inline_vecs; bch_bio_map(bio, NULL); } -static void refill_dirty(struct closure *cl) -{ - struct cached_dev *dc = container_of(cl, struct cached_dev, - writeback.cl); - struct keybuf *buf = &dc->writeback_keys; - bool searched_from_start = false; - struct bkey end = MAX_KEY; - SET_KEY_INODE(&end, dc->disk.id); - - if (!atomic_read(&dc->disk.detaching) && - !dc->writeback_running) - closure_return(cl); - - down_write(&dc->writeback_lock); - - if (!atomic_read(&dc->has_dirty)) { - SET_BDEV_STATE(&dc->sb, BDEV_STATE_CLEAN); - bch_write_bdev_super(dc, NULL); - - up_write(&dc->writeback_lock); - closure_return(cl); - } - - if (bkey_cmp(&buf->last_scanned, &end) >= 0) { - buf->last_scanned = KEY(dc->disk.id, 0, 0); - searched_from_start = true; - } - - bch_refill_keybuf(dc->disk.c, buf, &end); - - if (bkey_cmp(&buf->last_scanned, &end) >= 0 && searched_from_start) { - /* Searched the entire btree - delay awhile */ - - if (RB_EMPTY_ROOT(&buf->keys)) { - atomic_set(&dc->has_dirty, 0); - cached_dev_put(dc); - } - - if (!atomic_read(&dc->disk.detaching)) - closure_delay(&dc->writeback, dc->writeback_delay * HZ); - } - - up_write(&dc->writeback_lock); - - ratelimit_reset(&dc->writeback_rate); - - /* Punt to workqueue only so we don't recurse and blow the stack */ - continue_at(cl, read_dirty, dirty_wq); -} - -void bch_writeback_queue(struct cached_dev *dc) -{ - if (closure_trylock(&dc->writeback.cl, &dc->disk.cl)) { - if (!atomic_read(&dc->disk.detaching)) - closure_delay(&dc->writeback, dc->writeback_delay * HZ); - - continue_at(&dc->writeback.cl, refill_dirty, dirty_wq); - } -} - -void bch_writeback_add(struct cached_dev *dc, unsigned sectors) -{ - atomic_long_add(sectors, &dc->disk.sectors_dirty); - - if (!atomic_read(&dc->has_dirty) && - !atomic_xchg(&dc->has_dirty, 1)) { - atomic_inc(&dc->count); - - 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); - - if (dc->writeback_percent) - schedule_delayed_work(&dc->writeback_rate_update, - dc->writeback_rate_update_seconds * HZ); - } -} - -/* Background writeback - IO loop */ - static void dirty_io_destructor(struct closure *cl) { struct dirty_io *io = container_of(cl, struct dirty_io, cl); @@ -216,39 +129,39 @@ 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; - struct bio_vec *bv = bio_iovec_idx(&io->bio, io->bio.bi_vcnt); + struct bio_vec *bv; + int i; - while (bv-- != io->bio.bi_io_vec) + bio_for_each_segment_all(bv, &io->bio, i) __free_page(bv->bv_page); /* This is kind of a dumb way of signalling errors. */ if (KEY_DIRTY(&w->key)) { + int ret; unsigned i; - struct btree_op op; - bch_btree_op_init_stack(&op); + struct keylist keys; - op.type = BTREE_REPLACE; - bkey_copy(&op.replace, &w->key); + bch_keylist_init(&keys); - SET_KEY_DIRTY(&w->key, false); - bch_keylist_add(&op.keys, &w->key); + 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); - pr_debug("clearing %s", pkey(&w->key)); - bch_btree_insert(&op, dc->disk.c); - closure_sync(&op.cl); + ret = bch_btree_insert(dc->disk.c, &keys, NULL, &w->key); + + if (ret) + trace_bcache_writeback_collision(&w->key); - atomic_long_inc(op.insert_collision + atomic_long_inc(ret ? &dc->disk.c->writeback_keys_failed : &dc->disk.c->writeback_keys_done); } bch_keybuf_del(&dc->writeback_keys, w); - atomic_dec_bug(&dc->in_flight); - - closure_wake_up(&dc->writeback_wait); + up(&dc->in_flight); closure_return_with_destructor(cl, dirty_io_destructor); } @@ -271,14 +184,13 @@ static void write_dirty(struct closure *cl) dirty_init(w); io->bio.bi_rw = WRITE; - io->bio.bi_sector = KEY_START(&w->key); + io->bio.bi_iter.bi_sector = KEY_START(&w->key); io->bio.bi_bdev = io->dc->bdev; io->bio.bi_end_io = dirty_endio; - trace_bcache_write_dirty(&io->bio); closure_bio_submit(&io->bio, cl, &io->dc->disk); - continue_at(cl, write_dirty_finish, dirty_wq); + continue_at(cl, write_dirty_finish, system_wq); } static void read_dirty_endio(struct bio *bio, int error) @@ -296,40 +208,38 @@ static void read_dirty_submit(struct closure *cl) { struct dirty_io *io = container_of(cl, struct dirty_io, cl); - trace_bcache_read_dirty(&io->bio); closure_bio_submit(&io->bio, cl, &io->dc->disk); - continue_at(cl, write_dirty, dirty_wq); + continue_at(cl, write_dirty, system_wq); } -static void read_dirty(struct closure *cl) +static void read_dirty(struct cached_dev *dc) { - struct cached_dev *dc = container_of(cl, struct cached_dev, - writeback.cl); - unsigned delay = writeback_delay(dc, 0); + unsigned delay = 0; struct keybuf_key *w; struct dirty_io *io; + struct closure cl; + + closure_init_stack(&cl); /* * XXX: if we error, background writeback just spins. Should use some * mempools. */ - while (1) { + while (!kthread_should_stop()) { + try_to_freeze(); + w = bch_keybuf_next(&dc->writeback_keys); if (!w) break; BUG_ON(ptr_stale(dc->disk.c, &w->key, 0)); - if (delay > 0 && - (KEY_START(&w->key) != dc->last_read || - jiffies_to_msecs(delay) > 50)) { - w->private = NULL; - - closure_delay(&dc->writeback, delay); - continue_at(cl, read_dirty, dirty_wq); - } + if (KEY_START(&w->key) != dc->last_read || + jiffies_to_msecs(delay) > 50) + while (!kthread_should_stop() && delay) + delay = schedule_timeout_uninterruptible(delay); dc->last_read = KEY_OFFSET(&w->key); @@ -343,26 +253,21 @@ static void read_dirty(struct closure *cl) io->dc = dc; dirty_init(w); - io->bio.bi_sector = PTR_OFFSET(&w->key, 0); + io->bio.bi_iter.bi_sector = PTR_OFFSET(&w->key, 0); io->bio.bi_bdev = PTR_CACHE(dc->disk.c, &w->key, 0)->bdev; io->bio.bi_rw = READ; io->bio.bi_end_io = read_dirty_endio; - if (bch_bio_alloc_pages(&io->bio, GFP_KERNEL)) + if (bio_alloc_pages(&io->bio, GFP_KERNEL)) goto err_free; - pr_debug("%s", pkey(&w->key)); + trace_bcache_writeback(&w->key); - closure_call(&io->cl, read_dirty_submit, NULL, &dc->disk.cl); + down(&dc->in_flight); + closure_call(&io->cl, read_dirty_submit, NULL, &cl); delay = writeback_delay(dc, KEY_SIZE(&w->key)); - - atomic_inc(&dc->in_flight); - - if (!closure_wait_event(&dc->writeback_wait, cl, - atomic_read(&dc->in_flight) < 64)) - continue_at(cl, read_dirty, dirty_wq); } if (0) { @@ -372,15 +277,212 @@ err: bch_keybuf_del(&dc->writeback_keys, w); } - refill_dirty(cl); + /* + * Wait for outstanding writeback IOs to finish (and keybuf slots to be + * freed) before refilling again + */ + closure_sync(&cl); } -void bch_cached_dev_writeback_init(struct cached_dev *dc) +/* Scan for dirty data */ + +void bcache_dev_sectors_dirty_add(struct cache_set *c, unsigned inode, + uint64_t offset, int nr_sectors) { - closure_init_unlocked(&dc->writeback); - init_rwsem(&dc->writeback_lock); + struct bcache_device *d = c->devices[inode]; + unsigned stripe_offset, stripe, sectors_dirty; + + if (!d) + return; - bch_keybuf_init(&dc->writeback_keys, dirty_pred); + stripe = offset_to_stripe(d, offset); + stripe_offset = offset & (d->stripe_size - 1); + + while (nr_sectors) { + int s = min_t(unsigned, 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) +{ + return KEY_DIRTY(k); +} + +static void refill_full_stripes(struct cached_dev *dc) +{ + struct keybuf *buf = &dc->writeback_keys; + unsigned start_stripe, stripe, next_stripe; + bool wrapped = false; + + stripe = offset_to_stripe(&dc->disk, KEY_OFFSET(&buf->last_scanned)); + + if (stripe >= dc->disk.nr_stripes) + 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; + } + } +} + +static bool refill_dirty(struct cached_dev *dc) +{ + struct keybuf *buf = &dc->writeback_keys; + struct bkey end = KEY(dc->disk.id, MAX_KEY_OFFSET, 0); + bool searched_from_start = false; + + if (dc->partial_stripes_expensive) { + refill_full_stripes(dc); + if (array_freelist_empty(&buf->freelist)) + return false; + } + + if (bkey_cmp(&buf->last_scanned, &end) >= 0) { + buf->last_scanned = KEY(dc->disk.id, 0, 0); + searched_from_start = true; + } + + bch_refill_keybuf(dc->disk.c, buf, &end, dirty_pred); + + return bkey_cmp(&buf->last_scanned, &end) >= 0 && searched_from_start; +} + +static int bch_writeback_thread(void *arg) +{ + struct cached_dev *dc = arg; + bool searched_full_index; + + while (!kthread_should_stop()) { + down_write(&dc->writeback_lock); + if (!atomic_read(&dc->has_dirty) || + (!test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags) && + !dc->writeback_running)) { + up_write(&dc->writeback_lock); + set_current_state(TASK_INTERRUPTIBLE); + + if (kthread_should_stop()) + return 0; + + try_to_freeze(); + schedule(); + continue; + } + + searched_full_index = refill_dirty(dc); + + if (searched_full_index && + RB_EMPTY_ROOT(&dc->writeback_keys.keys)) { + atomic_set(&dc->has_dirty, 0); + cached_dev_put(dc); + SET_BDEV_STATE(&dc->sb, BDEV_STATE_CLEAN); + bch_write_bdev_super(dc, NULL); + } + + up_write(&dc->writeback_lock); + + bch_ratelimit_reset(&dc->writeback_rate); + read_dirty(dc); + + if (searched_full_index) { + unsigned delay = dc->writeback_delay * HZ; + + while (delay && + !kthread_should_stop() && + !test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags)) + delay = schedule_timeout_uninterruptible(delay); + } + } + + return 0; +} + +/* Init */ + +struct sectors_dirty_init { + struct btree_op op; + unsigned inode; +}; + +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)); + + return MAP_CONTINUE; +} + +void bch_sectors_dirty_init(struct cached_dev *dc) +{ + struct sectors_dirty_init op; + + bch_btree_op_init(&op.op, -1); + op.inode = dc->disk.id; + + bch_btree_map_keys(&op.op, dc->disk.c, &KEY(op.inode, 0, 0), + sectors_dirty_init_fn, 0); + + dc->disk.sectors_dirty_last = bcache_dev_sectors_dirty(&dc->disk); +} + +int 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 = true; @@ -388,27 +490,18 @@ void bch_cached_dev_writeback_init(struct cached_dev *dc) dc->writeback_delay = 30; dc->writeback_rate.rate = 1024; - dc->writeback_rate_update_seconds = 30; - dc->writeback_rate_d_term = 16; - dc->writeback_rate_p_term_inverse = 64; - dc->writeback_rate_d_smooth = 8; + dc->writeback_rate_update_seconds = 5; + dc->writeback_rate_d_term = 30; + dc->writeback_rate_p_term_inverse = 6000; + + dc->writeback_thread = kthread_create(bch_writeback_thread, dc, + "bcache_writeback"); + if (IS_ERR(dc->writeback_thread)) + return PTR_ERR(dc->writeback_thread); INIT_DELAYED_WORK(&dc->writeback_rate_update, update_writeback_rate); schedule_delayed_work(&dc->writeback_rate_update, dc->writeback_rate_update_seconds * HZ); -} - -void bch_writeback_exit(void) -{ - if (dirty_wq) - destroy_workqueue(dirty_wq); -} - -int __init bch_writeback_init(void) -{ - dirty_wq = create_singlethread_workqueue("bcache_writeback"); - if (!dirty_wq) - return -ENOMEM; return 0; } diff --git a/drivers/md/bcache/writeback.h b/drivers/md/bcache/writeback.h new file mode 100644 index 00000000000..e2f8598937a --- /dev/null +++ b/drivers/md/bcache/writeback.h @@ -0,0 +1,90 @@ +#ifndef _BCACHE_WRITEBACK_H +#define _BCACHE_WRITEBACK_H + +#define CUTOFF_WRITEBACK 40 +#define CUTOFF_WRITEBACK_SYNC 70 + +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 unsigned offset_to_stripe(struct bcache_device *d, + uint64_t offset) +{ + do_div(offset, d->stripe_size); + return offset; +} + +static inline bool bcache_dev_stripe_dirty(struct cached_dev *dc, + uint64_t offset, + unsigned nr_sectors) +{ + unsigned stripe = offset_to_stripe(&dc->disk, offset); + + 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++; + } +} + +static inline bool should_writeback(struct cached_dev *dc, struct bio *bio, + unsigned cache_mode, bool would_skip) +{ + unsigned 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 > CUTOFF_WRITEBACK_SYNC) + 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 bio->bi_rw & REQ_SYNC || + in_use <= CUTOFF_WRITEBACK; +} + +static inline void bch_writeback_queue(struct cached_dev *dc) +{ + 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)) { + atomic_inc(&dc->count); + + 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 *, unsigned, uint64_t, int); + +void bch_sectors_dirty_init(struct cached_dev *dc); +int bch_cached_dev_writeback_init(struct cached_dev *); + +#endif diff --git a/drivers/md/bitmap.c b/drivers/md/bitmap.c index 5a2c7549982..67f8b31e205 100644 --- a/drivers/md/bitmap.c +++ b/drivers/md/bitmap.c @@ -669,17 +669,13 @@ static inline unsigned long file_page_offset(struct bitmap_storage *store, /* * return a pointer to the page in the filemap that contains the given bit * - * this lookup is complicated by the fact that the bitmap sb might be exactly - * 1 page (e.g., x86) or less than 1 page -- so the bitmap might start on page - * 0 or page 1 */ static inline struct page *filemap_get_page(struct bitmap_storage *store, unsigned long chunk) { if (file_page_index(store, chunk) >= store->file_pages) return NULL; - return store->filemap[file_page_index(store, chunk) - - file_page_index(store, 0)]; + return store->filemap[file_page_index(store, chunk)]; } static int bitmap_storage_alloc(struct bitmap_storage *store, @@ -1635,7 +1631,7 @@ int bitmap_create(struct mddev *mddev) sector_t blocks = mddev->resync_max_sectors; struct file *file = mddev->bitmap_info.file; int err; - struct sysfs_dirent *bm = NULL; + struct kernfs_node *bm = NULL; BUILD_BUG_ON(sizeof(bitmap_super_t) != 256); @@ -1654,9 +1650,9 @@ int bitmap_create(struct mddev *mddev) bitmap->mddev = mddev; if (mddev->kobj.sd) - bm = sysfs_get_dirent(mddev->kobj.sd, NULL, "bitmap"); + bm = sysfs_get_dirent(mddev->kobj.sd, "bitmap"); if (bm) { - bitmap->sysfs_can_clear = sysfs_get_dirent(bm, NULL, "can_clear"); + bitmap->sysfs_can_clear = sysfs_get_dirent(bm, "can_clear"); sysfs_put(bm); } else bitmap->sysfs_can_clear = NULL; @@ -1988,7 +1984,6 @@ location_store(struct mddev *mddev, const char *buf, size_t len) if (mddev->bitmap_info.file) { struct file *f = mddev->bitmap_info.file; mddev->bitmap_info.file = NULL; - restore_bitmap_write_access(f); fput(f); } } else { @@ -2002,9 +1997,9 @@ location_store(struct mddev *mddev, const char *buf, size_t len) } else { int rv; if (buf[0] == '+') - rv = strict_strtoll(buf+1, 10, &offset); + rv = kstrtoll(buf+1, 10, &offset); else - rv = strict_strtoll(buf, 10, &offset); + rv = kstrtoll(buf, 10, &offset); if (rv) return rv; if (offset == 0) @@ -2139,7 +2134,7 @@ static ssize_t backlog_store(struct mddev *mddev, const char *buf, size_t len) { unsigned long backlog; - int rv = strict_strtoul(buf, 10, &backlog); + int rv = kstrtoul(buf, 10, &backlog); if (rv) return rv; if (backlog > COUNTER_MAX) @@ -2165,7 +2160,7 @@ chunksize_store(struct mddev *mddev, const char *buf, size_t len) unsigned long csize; if (mddev->bitmap) return -EBUSY; - rv = strict_strtoul(buf, 10, &csize); + rv = kstrtoul(buf, 10, &csize); if (rv) return rv; if (csize < 512 || diff --git a/drivers/md/bitmap.h b/drivers/md/bitmap.h index df4aeb6ac6f..30210b9c4ef 100644 --- a/drivers/md/bitmap.h +++ b/drivers/md/bitmap.h @@ -225,7 +225,7 @@ struct bitmap { wait_queue_head_t overflow_wait; wait_queue_head_t behind_wait; - struct sysfs_dirent *sysfs_can_clear; + struct kernfs_node *sysfs_can_clear; }; /* the bitmap API */ diff --git a/drivers/md/dm-bio-prison.c b/drivers/md/dm-bio-prison.c index 85f0b707425..f752d12081f 100644 --- a/drivers/md/dm-bio-prison.c +++ b/drivers/md/dm-bio-prison.c @@ -14,13 +14,17 @@ /*----------------------------------------------------------------*/ -struct dm_bio_prison { +struct bucket { spinlock_t lock; + struct hlist_head cells; +}; + +struct dm_bio_prison { mempool_t *cell_pool; unsigned nr_buckets; unsigned hash_mask; - struct hlist_head *cells; + struct bucket *buckets; }; /*----------------------------------------------------------------*/ @@ -40,6 +44,12 @@ static uint32_t calc_nr_buckets(unsigned nr_cells) static struct kmem_cache *_cell_cache; +static void init_bucket(struct bucket *b) +{ + spin_lock_init(&b->lock); + INIT_HLIST_HEAD(&b->cells); +} + /* * @nr_cells should be the number of cells you want in use _concurrently_. * Don't confuse it with the number of distinct keys. @@ -49,13 +59,12 @@ struct dm_bio_prison *dm_bio_prison_create(unsigned nr_cells) unsigned i; uint32_t nr_buckets = calc_nr_buckets(nr_cells); size_t len = sizeof(struct dm_bio_prison) + - (sizeof(struct hlist_head) * nr_buckets); + (sizeof(struct bucket) * nr_buckets); struct dm_bio_prison *prison = kmalloc(len, GFP_KERNEL); if (!prison) return NULL; - spin_lock_init(&prison->lock); prison->cell_pool = mempool_create_slab_pool(nr_cells, _cell_cache); if (!prison->cell_pool) { kfree(prison); @@ -64,9 +73,9 @@ struct dm_bio_prison *dm_bio_prison_create(unsigned nr_cells) prison->nr_buckets = nr_buckets; prison->hash_mask = nr_buckets - 1; - prison->cells = (struct hlist_head *) (prison + 1); + prison->buckets = (struct bucket *) (prison + 1); for (i = 0; i < nr_buckets; i++) - INIT_HLIST_HEAD(prison->cells + i); + init_bucket(prison->buckets + i); return prison; } @@ -107,40 +116,44 @@ static int keys_equal(struct dm_cell_key *lhs, struct dm_cell_key *rhs) (lhs->block == rhs->block); } -static struct dm_bio_prison_cell *__search_bucket(struct hlist_head *bucket, +static struct bucket *get_bucket(struct dm_bio_prison *prison, + struct dm_cell_key *key) +{ + return prison->buckets + hash_key(prison, key); +} + +static struct dm_bio_prison_cell *__search_bucket(struct bucket *b, struct dm_cell_key *key) { struct dm_bio_prison_cell *cell; - hlist_for_each_entry(cell, bucket, list) + hlist_for_each_entry(cell, &b->cells, list) if (keys_equal(&cell->key, key)) return cell; return NULL; } -static void __setup_new_cell(struct dm_bio_prison *prison, +static void __setup_new_cell(struct bucket *b, struct dm_cell_key *key, struct bio *holder, - uint32_t hash, struct dm_bio_prison_cell *cell) { memcpy(&cell->key, key, sizeof(cell->key)); cell->holder = holder; bio_list_init(&cell->bios); - hlist_add_head(&cell->list, prison->cells + hash); + hlist_add_head(&cell->list, &b->cells); } -static int __bio_detain(struct dm_bio_prison *prison, +static int __bio_detain(struct bucket *b, struct dm_cell_key *key, struct bio *inmate, struct dm_bio_prison_cell *cell_prealloc, struct dm_bio_prison_cell **cell_result) { - uint32_t hash = hash_key(prison, key); struct dm_bio_prison_cell *cell; - cell = __search_bucket(prison->cells + hash, key); + cell = __search_bucket(b, key); if (cell) { if (inmate) bio_list_add(&cell->bios, inmate); @@ -148,7 +161,7 @@ static int __bio_detain(struct dm_bio_prison *prison, return 1; } - __setup_new_cell(prison, key, inmate, hash, cell_prealloc); + __setup_new_cell(b, key, inmate, cell_prealloc); *cell_result = cell_prealloc; return 0; } @@ -161,10 +174,11 @@ static int bio_detain(struct dm_bio_prison *prison, { int r; unsigned long flags; + struct bucket *b = get_bucket(prison, key); - spin_lock_irqsave(&prison->lock, flags); - r = __bio_detain(prison, key, inmate, cell_prealloc, cell_result); - spin_unlock_irqrestore(&prison->lock, flags); + spin_lock_irqsave(&b->lock, flags); + r = __bio_detain(b, key, inmate, cell_prealloc, cell_result); + spin_unlock_irqrestore(&b->lock, flags); return r; } @@ -208,10 +222,11 @@ void dm_cell_release(struct dm_bio_prison *prison, struct bio_list *bios) { unsigned long flags; + struct bucket *b = get_bucket(prison, &cell->key); - spin_lock_irqsave(&prison->lock, flags); + spin_lock_irqsave(&b->lock, flags); __cell_release(cell, bios); - spin_unlock_irqrestore(&prison->lock, flags); + spin_unlock_irqrestore(&b->lock, flags); } EXPORT_SYMBOL_GPL(dm_cell_release); @@ -230,28 +245,25 @@ void dm_cell_release_no_holder(struct dm_bio_prison *prison, struct bio_list *inmates) { unsigned long flags; + struct bucket *b = get_bucket(prison, &cell->key); - spin_lock_irqsave(&prison->lock, flags); + spin_lock_irqsave(&b->lock, flags); __cell_release_no_holder(cell, inmates); - spin_unlock_irqrestore(&prison->lock, flags); + spin_unlock_irqrestore(&b->lock, flags); } EXPORT_SYMBOL_GPL(dm_cell_release_no_holder); void dm_cell_error(struct dm_bio_prison *prison, - struct dm_bio_prison_cell *cell) + struct dm_bio_prison_cell *cell, int error) { struct bio_list bios; struct bio *bio; - unsigned long flags; bio_list_init(&bios); - - spin_lock_irqsave(&prison->lock, flags); - __cell_release(cell, &bios); - spin_unlock_irqrestore(&prison->lock, flags); + dm_cell_release(prison, cell, &bios); while ((bio = bio_list_pop(&bios))) - bio_io_error(bio); + bio_endio(bio, error); } EXPORT_SYMBOL_GPL(dm_cell_error); diff --git a/drivers/md/dm-bio-prison.h b/drivers/md/dm-bio-prison.h index 3f833190ead..6805a142b75 100644 --- a/drivers/md/dm-bio-prison.h +++ b/drivers/md/dm-bio-prison.h @@ -85,7 +85,7 @@ void dm_cell_release_no_holder(struct dm_bio_prison *prison, struct dm_bio_prison_cell *cell, struct bio_list *inmates); void dm_cell_error(struct dm_bio_prison *prison, - struct dm_bio_prison_cell *cell); + struct dm_bio_prison_cell *cell, int error); /*----------------------------------------------------------------*/ diff --git a/drivers/md/dm-bio-record.h b/drivers/md/dm-bio-record.h index 3a8cfa2645c..dd364611156 100644 --- a/drivers/md/dm-bio-record.h +++ b/drivers/md/dm-bio-record.h @@ -17,55 +17,24 @@ * original bio state. */ -struct dm_bio_vec_details { -#if PAGE_SIZE < 65536 - __u16 bv_len; - __u16 bv_offset; -#else - unsigned bv_len; - unsigned bv_offset; -#endif -}; - struct dm_bio_details { - sector_t bi_sector; struct block_device *bi_bdev; - unsigned int bi_size; - unsigned short bi_idx; unsigned long bi_flags; - struct dm_bio_vec_details bi_io_vec[BIO_MAX_PAGES]; + struct bvec_iter bi_iter; }; static inline void dm_bio_record(struct dm_bio_details *bd, struct bio *bio) { - unsigned i; - - bd->bi_sector = bio->bi_sector; bd->bi_bdev = bio->bi_bdev; - bd->bi_size = bio->bi_size; - bd->bi_idx = bio->bi_idx; bd->bi_flags = bio->bi_flags; - - for (i = 0; i < bio->bi_vcnt; i++) { - bd->bi_io_vec[i].bv_len = bio->bi_io_vec[i].bv_len; - bd->bi_io_vec[i].bv_offset = bio->bi_io_vec[i].bv_offset; - } + bd->bi_iter = bio->bi_iter; } static inline void dm_bio_restore(struct dm_bio_details *bd, struct bio *bio) { - unsigned i; - - bio->bi_sector = bd->bi_sector; bio->bi_bdev = bd->bi_bdev; - bio->bi_size = bd->bi_size; - bio->bi_idx = bd->bi_idx; bio->bi_flags = bd->bi_flags; - - for (i = 0; i < bio->bi_vcnt; i++) { - bio->bi_io_vec[i].bv_len = bd->bi_io_vec[i].bv_len; - bio->bi_io_vec[i].bv_offset = bd->bi_io_vec[i].bv_offset; - } + bio->bi_iter = bd->bi_iter; } #endif diff --git a/drivers/md/dm-bufio.c b/drivers/md/dm-bufio.c index 0387e05cdb9..d724459860d 100644 --- a/drivers/md/dm-bufio.c +++ b/drivers/md/dm-bufio.c @@ -104,6 +104,8 @@ struct dm_bufio_client { struct list_head reserved_buffers; unsigned need_reserved_buffers; + unsigned minimum_buffers; + struct hlist_head *cache_hash; wait_queue_head_t free_buffer_wait; @@ -145,6 +147,7 @@ struct dm_buffer { unsigned long state; unsigned long last_accessed; struct dm_bufio_client *c; + struct list_head write_list; struct bio bio; struct bio_vec bio_vec[DM_BUFIO_INLINE_VECS]; }; @@ -349,7 +352,7 @@ static void *alloc_buffer_data(struct dm_bufio_client *c, gfp_t gfp_mask, if (gfp_mask & __GFP_NORETRY) noio_flag = memalloc_noio_save(); - ptr = __vmalloc(c->block_size, gfp_mask, PAGE_KERNEL); + ptr = __vmalloc(c->block_size, gfp_mask | __GFP_HIGHMEM, PAGE_KERNEL); if (gfp_mask & __GFP_NORETRY) memalloc_noio_restore(noio_flag); @@ -537,7 +540,7 @@ static void use_inline_bio(struct dm_buffer *b, int rw, sector_t block, bio_init(&b->bio); b->bio.bi_io_vec = b->bio_vec; b->bio.bi_max_vecs = DM_BUFIO_INLINE_VECS; - b->bio.bi_sector = block << b->c->sectors_per_block_bits; + b->bio.bi_iter.bi_sector = block << b->c->sectors_per_block_bits; b->bio.bi_bdev = b->c->bdev; b->bio.bi_end_io = end_io; @@ -604,9 +607,9 @@ static void write_endio(struct bio *bio, int error) BUG_ON(!test_bit(B_WRITING, &b->state)); - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(B_WRITING, &b->state); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); wake_up_bit(&b->state, B_WRITING); } @@ -630,7 +633,8 @@ static int do_io_schedule(void *word) * - Submit our write and don't wait on it. We set B_WRITING indicating * that there is a write in progress. */ -static void __write_dirty_buffer(struct dm_buffer *b) +static void __write_dirty_buffer(struct dm_buffer *b, + struct list_head *write_list) { if (!test_bit(B_DIRTY, &b->state)) return; @@ -639,7 +643,24 @@ static void __write_dirty_buffer(struct dm_buffer *b) wait_on_bit_lock(&b->state, B_WRITING, do_io_schedule, TASK_UNINTERRUPTIBLE); - submit_io(b, WRITE, b->block, write_endio); + if (!write_list) + submit_io(b, WRITE, b->block, write_endio); + else + list_add_tail(&b->write_list, write_list); +} + +static void __flush_write_list(struct list_head *write_list) +{ + struct blk_plug plug; + blk_start_plug(&plug); + while (!list_empty(write_list)) { + struct dm_buffer *b = + list_entry(write_list->next, struct dm_buffer, write_list); + list_del(&b->write_list); + submit_io(b, WRITE, b->block, write_endio); + dm_bufio_cond_resched(); + } + blk_finish_plug(&plug); } /* @@ -655,7 +676,7 @@ static void __make_buffer_clean(struct dm_buffer *b) return; wait_on_bit(&b->state, B_READING, do_io_schedule, TASK_UNINTERRUPTIBLE); - __write_dirty_buffer(b); + __write_dirty_buffer(b, NULL); wait_on_bit(&b->state, B_WRITING, do_io_schedule, TASK_UNINTERRUPTIBLE); } @@ -802,7 +823,8 @@ static void __free_buffer_wake(struct dm_buffer *b) wake_up(&c->free_buffer_wait); } -static void __write_dirty_buffers_async(struct dm_bufio_client *c, int no_wait) +static void __write_dirty_buffers_async(struct dm_bufio_client *c, int no_wait, + struct list_head *write_list) { struct dm_buffer *b, *tmp; @@ -818,7 +840,7 @@ static void __write_dirty_buffers_async(struct dm_bufio_client *c, int no_wait) if (no_wait && test_bit(B_WRITING, &b->state)) return; - __write_dirty_buffer(b); + __write_dirty_buffer(b, write_list); dm_bufio_cond_resched(); } } @@ -841,8 +863,8 @@ static void __get_memory_limit(struct dm_bufio_client *c, buffers = dm_bufio_cache_size_per_client >> (c->sectors_per_block_bits + SECTOR_SHIFT); - if (buffers < DM_BUFIO_MIN_BUFFERS) - buffers = DM_BUFIO_MIN_BUFFERS; + if (buffers < c->minimum_buffers) + buffers = c->minimum_buffers; *limit_buffers = buffers; *threshold_buffers = buffers * DM_BUFIO_WRITEBACK_PERCENT / 100; @@ -853,7 +875,8 @@ static void __get_memory_limit(struct dm_bufio_client *c, * If we are over threshold_buffers, start freeing buffers. * If we're over "limit_buffers", block until we get under the limit. */ -static void __check_watermark(struct dm_bufio_client *c) +static void __check_watermark(struct dm_bufio_client *c, + struct list_head *write_list) { unsigned long threshold_buffers, limit_buffers; @@ -872,7 +895,7 @@ static void __check_watermark(struct dm_bufio_client *c) } if (c->n_buffers[LIST_DIRTY] > threshold_buffers) - __write_dirty_buffers_async(c, 1); + __write_dirty_buffers_async(c, 1, write_list); } /* @@ -897,7 +920,8 @@ static struct dm_buffer *__find(struct dm_bufio_client *c, sector_t block) *--------------------------------------------------------------*/ static struct dm_buffer *__bufio_new(struct dm_bufio_client *c, sector_t block, - enum new_flag nf, int *need_submit) + enum new_flag nf, int *need_submit, + struct list_head *write_list) { struct dm_buffer *b, *new_b = NULL; @@ -924,7 +948,7 @@ static struct dm_buffer *__bufio_new(struct dm_bufio_client *c, sector_t block, goto found_buffer; } - __check_watermark(c); + __check_watermark(c, write_list); b = new_b; b->hold_count = 1; @@ -973,9 +997,9 @@ static void read_endio(struct bio *bio, int error) BUG_ON(!test_bit(B_READING, &b->state)); - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(B_READING, &b->state); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); wake_up_bit(&b->state, B_READING); } @@ -992,10 +1016,14 @@ static void *new_read(struct dm_bufio_client *c, sector_t block, int need_submit; struct dm_buffer *b; + LIST_HEAD(write_list); + dm_bufio_lock(c); - b = __bufio_new(c, block, nf, &need_submit); + b = __bufio_new(c, block, nf, &need_submit, &write_list); dm_bufio_unlock(c); + __flush_write_list(&write_list); + if (!b) return b; @@ -1047,6 +1075,8 @@ void dm_bufio_prefetch(struct dm_bufio_client *c, { struct blk_plug plug; + LIST_HEAD(write_list); + BUG_ON(dm_bufio_in_request()); blk_start_plug(&plug); @@ -1055,7 +1085,15 @@ void dm_bufio_prefetch(struct dm_bufio_client *c, for (; n_blocks--; block++) { int need_submit; struct dm_buffer *b; - b = __bufio_new(c, block, NF_PREFETCH, &need_submit); + b = __bufio_new(c, block, NF_PREFETCH, &need_submit, + &write_list); + if (unlikely(!list_empty(&write_list))) { + dm_bufio_unlock(c); + blk_finish_plug(&plug); + __flush_write_list(&write_list); + blk_start_plug(&plug); + dm_bufio_lock(c); + } if (unlikely(b != NULL)) { dm_bufio_unlock(c); @@ -1069,7 +1107,6 @@ void dm_bufio_prefetch(struct dm_bufio_client *c, goto flush_plug; dm_bufio_lock(c); } - } dm_bufio_unlock(c); @@ -1126,11 +1163,14 @@ EXPORT_SYMBOL_GPL(dm_bufio_mark_buffer_dirty); void dm_bufio_write_dirty_buffers_async(struct dm_bufio_client *c) { + LIST_HEAD(write_list); + BUG_ON(dm_bufio_in_request()); dm_bufio_lock(c); - __write_dirty_buffers_async(c, 0); + __write_dirty_buffers_async(c, 0, &write_list); dm_bufio_unlock(c); + __flush_write_list(&write_list); } EXPORT_SYMBOL_GPL(dm_bufio_write_dirty_buffers_async); @@ -1147,8 +1187,13 @@ int dm_bufio_write_dirty_buffers(struct dm_bufio_client *c) unsigned long buffers_processed = 0; struct dm_buffer *b, *tmp; + LIST_HEAD(write_list); + + dm_bufio_lock(c); + __write_dirty_buffers_async(c, 0, &write_list); + dm_bufio_unlock(c); + __flush_write_list(&write_list); dm_bufio_lock(c); - __write_dirty_buffers_async(c, 0); again: list_for_each_entry_safe_reverse(b, tmp, &c->lru[LIST_DIRTY], lru_list) { @@ -1274,7 +1319,7 @@ retry: BUG_ON(!b->hold_count); BUG_ON(test_bit(B_READING, &b->state)); - __write_dirty_buffer(b); + __write_dirty_buffer(b, NULL); if (b->hold_count == 1) { wait_on_bit(&b->state, B_WRITING, do_io_schedule, TASK_UNINTERRUPTIBLE); @@ -1307,6 +1352,34 @@ retry: } EXPORT_SYMBOL_GPL(dm_bufio_release_move); +/* + * Free the given buffer. + * + * This is just a hint, if the buffer is in use or dirty, this function + * does nothing. + */ +void dm_bufio_forget(struct dm_bufio_client *c, sector_t block) +{ + struct dm_buffer *b; + + dm_bufio_lock(c); + + b = __find(c, block); + if (b && likely(!b->hold_count) && likely(!b->state)) { + __unlink_buffer(b); + __free_buffer_wake(b); + } + + dm_bufio_unlock(c); +} +EXPORT_SYMBOL(dm_bufio_forget); + +void dm_bufio_set_minimum_buffers(struct dm_bufio_client *c, unsigned n) +{ + c->minimum_buffers = n; +} +EXPORT_SYMBOL(dm_bufio_set_minimum_buffers); + unsigned dm_bufio_get_block_size(struct dm_bufio_client *c) { return c->block_size; @@ -1382,62 +1455,75 @@ static int __cleanup_old_buffer(struct dm_buffer *b, gfp_t gfp, unsigned long max_jiffies) { if (jiffies - b->last_accessed < max_jiffies) - return 1; + return 0; if (!(gfp & __GFP_IO)) { if (test_bit(B_READING, &b->state) || test_bit(B_WRITING, &b->state) || test_bit(B_DIRTY, &b->state)) - return 1; + return 0; } if (b->hold_count) - return 1; + return 0; __make_buffer_clean(b); __unlink_buffer(b); __free_buffer_wake(b); - return 0; + return 1; } -static void __scan(struct dm_bufio_client *c, unsigned long nr_to_scan, - struct shrink_control *sc) +static long __scan(struct dm_bufio_client *c, unsigned long nr_to_scan, + gfp_t gfp_mask) { int l; struct dm_buffer *b, *tmp; + long freed = 0; for (l = 0; l < LIST_SIZE; l++) { - list_for_each_entry_safe_reverse(b, tmp, &c->lru[l], lru_list) - if (!__cleanup_old_buffer(b, sc->gfp_mask, 0) && - !--nr_to_scan) - return; + list_for_each_entry_safe_reverse(b, tmp, &c->lru[l], lru_list) { + freed += __cleanup_old_buffer(b, gfp_mask, 0); + if (!--nr_to_scan) + break; + } dm_bufio_cond_resched(); } + return freed; } -static int shrink(struct shrinker *shrinker, struct shrink_control *sc) +static unsigned long +dm_bufio_shrink_scan(struct shrinker *shrink, struct shrink_control *sc) { - struct dm_bufio_client *c = - container_of(shrinker, struct dm_bufio_client, shrinker); - unsigned long r; - unsigned long nr_to_scan = sc->nr_to_scan; + struct dm_bufio_client *c; + unsigned long freed; + c = container_of(shrink, struct dm_bufio_client, shrinker); if (sc->gfp_mask & __GFP_IO) dm_bufio_lock(c); else if (!dm_bufio_trylock(c)) - return !nr_to_scan ? 0 : -1; + return SHRINK_STOP; - if (nr_to_scan) - __scan(c, nr_to_scan, sc); + freed = __scan(c, sc->nr_to_scan, sc->gfp_mask); + dm_bufio_unlock(c); + return freed; +} - r = c->n_buffers[LIST_CLEAN] + c->n_buffers[LIST_DIRTY]; - if (r > INT_MAX) - r = INT_MAX; +static unsigned long +dm_bufio_shrink_count(struct shrinker *shrink, struct shrink_control *sc) +{ + struct dm_bufio_client *c; + unsigned long count; - dm_bufio_unlock(c); + c = container_of(shrink, struct dm_bufio_client, shrinker); + if (sc->gfp_mask & __GFP_IO) + dm_bufio_lock(c); + else if (!dm_bufio_trylock(c)) + return 0; - return r; + count = c->n_buffers[LIST_CLEAN] + c->n_buffers[LIST_DIRTY]; + dm_bufio_unlock(c); + return count; } /* @@ -1455,7 +1541,7 @@ struct dm_bufio_client *dm_bufio_client_create(struct block_device *bdev, unsign BUG_ON(block_size < 1 << SECTOR_SHIFT || (block_size & (block_size - 1))); - c = kmalloc(sizeof(*c), GFP_KERNEL); + c = kzalloc(sizeof(*c), GFP_KERNEL); if (!c) { r = -ENOMEM; goto bad_client; @@ -1490,6 +1576,8 @@ struct dm_bufio_client *dm_bufio_client_create(struct block_device *bdev, unsign INIT_LIST_HEAD(&c->reserved_buffers); c->need_reserved_buffers = reserved_buffers; + c->minimum_buffers = DM_BUFIO_MIN_BUFFERS; + init_waitqueue_head(&c->free_buffer_wait); c->async_write_error = 0; @@ -1539,7 +1627,8 @@ struct dm_bufio_client *dm_bufio_client_create(struct block_device *bdev, unsign __cache_size_refresh(); mutex_unlock(&dm_bufio_clients_lock); - c->shrinker.shrink = shrink; + c->shrinker.count_objects = dm_bufio_shrink_count; + c->shrinker.scan_objects = dm_bufio_shrink_scan; c->shrinker.seeks = 1; c->shrinker.batch = 0; register_shrinker(&c->shrinker); @@ -1626,7 +1715,7 @@ static void cleanup_old_buffers(void) struct dm_buffer *b; b = list_entry(c->lru[LIST_CLEAN].prev, struct dm_buffer, lru_list); - if (__cleanup_old_buffer(b, 0, max_age * HZ)) + if (!__cleanup_old_buffer(b, 0, max_age * HZ)) break; dm_bufio_cond_resched(); } @@ -1660,6 +1749,11 @@ static int __init dm_bufio_init(void) { __u64 mem; + dm_bufio_allocated_kmem_cache = 0; + dm_bufio_allocated_get_free_pages = 0; + dm_bufio_allocated_vmalloc = 0; + dm_bufio_current_allocated = 0; + memset(&dm_bufio_caches, 0, sizeof dm_bufio_caches); memset(&dm_bufio_cache_names, 0, sizeof dm_bufio_cache_names); diff --git a/drivers/md/dm-bufio.h b/drivers/md/dm-bufio.h index b142946a9e3..c096779a729 100644 --- a/drivers/md/dm-bufio.h +++ b/drivers/md/dm-bufio.h @@ -108,6 +108,18 @@ int dm_bufio_issue_flush(struct dm_bufio_client *c); */ void dm_bufio_release_move(struct dm_buffer *b, sector_t new_block); +/* + * Free the given buffer. + * This is just a hint, if the buffer is in use or dirty, this function + * does nothing. + */ +void dm_bufio_forget(struct dm_bufio_client *c, sector_t block); + +/* + * Set the minimum number of buffers before cleanup happens. + */ +void dm_bufio_set_minimum_buffers(struct dm_bufio_client *c, unsigned n); + unsigned dm_bufio_get_block_size(struct dm_bufio_client *c); sector_t dm_bufio_get_device_size(struct dm_bufio_client *c); sector_t dm_bufio_get_block_number(struct dm_buffer *b); diff --git a/drivers/md/dm-builtin.c b/drivers/md/dm-builtin.c new file mode 100644 index 00000000000..6c9049c51b2 --- /dev/null +++ b/drivers/md/dm-builtin.c @@ -0,0 +1,48 @@ +#include "dm.h" + +/* + * The kobject release method must not be placed in the module itself, + * otherwise we are subject to module unload races. + * + * The release method is called when the last reference to the kobject is + * dropped. It may be called by any other kernel code that drops the last + * reference. + * + * The release method suffers from module unload race. We may prevent the + * module from being unloaded at the start of the release method (using + * increased module reference count or synchronizing against the release + * method), however there is no way to prevent the module from being + * unloaded at the end of the release method. + * + * If this code were placed in the dm module, the following race may + * happen: + * 1. Some other process takes a reference to dm kobject + * 2. The user issues ioctl function to unload the dm device + * 3. dm_sysfs_exit calls kobject_put, however the object is not released + * because of the other reference taken at step 1 + * 4. dm_sysfs_exit waits on the completion + * 5. The other process that took the reference in step 1 drops it, + * dm_kobject_release is called from this process + * 6. dm_kobject_release calls complete() + * 7. a reschedule happens before dm_kobject_release returns + * 8. dm_sysfs_exit continues, the dm device is unloaded, module reference + * count is decremented + * 9. The user unloads the dm module + * 10. The other process that was rescheduled in step 7 continues to run, + * it is now executing code in unloaded module, so it crashes + * + * Note that if the process that takes the foreign reference to dm kobject + * has a low priority and the system is sufficiently loaded with + * higher-priority processes that prevent the low-priority process from + * being scheduled long enough, this bug may really happen. + * + * In order to fix this module unload race, we place the release method + * into a helper code that is compiled directly into the kernel. + */ + +void dm_kobject_release(struct kobject *kobj) +{ + complete(dm_get_completion_from_kobject(kobj)); +} + +EXPORT_SYMBOL(dm_kobject_release); diff --git a/drivers/md/dm-cache-block-types.h b/drivers/md/dm-cache-block-types.h index bed4ad4e1b7..aac0e2df06b 100644 --- a/drivers/md/dm-cache-block-types.h +++ b/drivers/md/dm-cache-block-types.h @@ -19,7 +19,6 @@ typedef dm_block_t __bitwise__ dm_oblock_t; typedef uint32_t __bitwise__ dm_cblock_t; -typedef dm_block_t __bitwise__ dm_dblock_t; static inline dm_oblock_t to_oblock(dm_block_t b) { @@ -41,14 +40,4 @@ static inline uint32_t from_cblock(dm_cblock_t b) return (__force uint32_t) b; } -static inline dm_dblock_t to_dblock(dm_block_t b) -{ - return (__force dm_dblock_t) b; -} - -static inline dm_block_t from_dblock(dm_dblock_t b) -{ - return (__force dm_block_t) b; -} - #endif /* DM_CACHE_BLOCK_TYPES_H */ diff --git a/drivers/md/dm-cache-metadata.c b/drivers/md/dm-cache-metadata.c index 1af7255bbff..d2899e7eb3a 100644 --- a/drivers/md/dm-cache-metadata.c +++ b/drivers/md/dm-cache-metadata.c @@ -20,7 +20,13 @@ #define CACHE_SUPERBLOCK_MAGIC 06142003 #define CACHE_SUPERBLOCK_LOCATION 0 -#define CACHE_VERSION 1 + +/* + * defines a range of metadata versions that this module can handle. + */ +#define MIN_CACHE_VERSION 1 +#define MAX_CACHE_VERSION 1 + #define CACHE_METADATA_CACHE_SIZE 64 /* @@ -103,7 +109,7 @@ struct dm_cache_metadata { dm_block_t discard_root; sector_t discard_block_size; - dm_dblock_t discard_nr_blocks; + dm_oblock_t discard_nr_blocks; sector_t data_block_size; dm_cblock_t cache_blocks; @@ -114,6 +120,12 @@ struct dm_cache_metadata { unsigned policy_version[CACHE_POLICY_VERSION_SIZE]; size_t policy_hint_size; struct dm_cache_statistics stats; + + /* + * Reading the space map root can fail, so we read it into this + * buffer before the superblock is locked and updated. + */ + __u8 metadata_space_map_root[SPACE_MAP_ROOT_SIZE]; }; /*------------------------------------------------------------------- @@ -134,6 +146,18 @@ static void sb_prepare_for_write(struct dm_block_validator *v, SUPERBLOCK_CSUM_XOR)); } +static int check_metadata_version(struct cache_disk_superblock *disk_super) +{ + uint32_t metadata_version = le32_to_cpu(disk_super->version); + if (metadata_version < MIN_CACHE_VERSION || metadata_version > MAX_CACHE_VERSION) { + DMERR("Cache metadata version %u found, but only versions between %u and %u supported.", + metadata_version, MIN_CACHE_VERSION, MAX_CACHE_VERSION); + return -EINVAL; + } + + return 0; +} + static int sb_check(struct dm_block_validator *v, struct dm_block *b, size_t sb_block_size) @@ -164,7 +188,7 @@ static int sb_check(struct dm_block_validator *v, return -EILSEQ; } - return 0; + return check_metadata_version(disk_super); } static struct dm_block_validator sb_validator = { @@ -198,7 +222,7 @@ static int superblock_lock(struct dm_cache_metadata *cmd, /*----------------------------------------------------------------*/ -static int __superblock_all_zeroes(struct dm_block_manager *bm, int *result) +static int __superblock_all_zeroes(struct dm_block_manager *bm, bool *result) { int r; unsigned i; @@ -214,10 +238,10 @@ static int __superblock_all_zeroes(struct dm_block_manager *bm, int *result) return r; data_le = dm_block_data(b); - *result = 1; + *result = true; for (i = 0; i < sb_block_size; i++) { if (data_le[i] != zero) { - *result = 0; + *result = false; break; } } @@ -242,11 +266,31 @@ static void __setup_mapping_info(struct dm_cache_metadata *cmd) } } +static int __save_sm_root(struct dm_cache_metadata *cmd) +{ + int r; + size_t metadata_len; + + r = dm_sm_root_size(cmd->metadata_sm, &metadata_len); + if (r < 0) + return r; + + return dm_sm_copy_root(cmd->metadata_sm, &cmd->metadata_space_map_root, + metadata_len); +} + +static void __copy_sm_root(struct dm_cache_metadata *cmd, + struct cache_disk_superblock *disk_super) +{ + memcpy(&disk_super->metadata_space_map_root, + &cmd->metadata_space_map_root, + sizeof(cmd->metadata_space_map_root)); +} + static int __write_initial_superblock(struct dm_cache_metadata *cmd) { int r; struct dm_block *sblock; - size_t metadata_len; struct cache_disk_superblock *disk_super; sector_t bdev_size = i_size_read(cmd->bdev->bd_inode) >> SECTOR_SHIFT; @@ -254,12 +298,16 @@ static int __write_initial_superblock(struct dm_cache_metadata *cmd) if (bdev_size > DM_CACHE_METADATA_MAX_SECTORS) bdev_size = DM_CACHE_METADATA_MAX_SECTORS; - r = dm_sm_root_size(cmd->metadata_sm, &metadata_len); + r = dm_tm_pre_commit(cmd->tm); if (r < 0) return r; - r = dm_tm_pre_commit(cmd->tm); - if (r < 0) + /* + * dm_sm_copy_root() can fail. So we need to do it before we start + * updating the superblock. + */ + r = __save_sm_root(cmd); + if (r) return r; r = superblock_lock_zero(cmd, &sblock); @@ -270,21 +318,18 @@ static int __write_initial_superblock(struct dm_cache_metadata *cmd) disk_super->flags = 0; memset(disk_super->uuid, 0, sizeof(disk_super->uuid)); disk_super->magic = cpu_to_le64(CACHE_SUPERBLOCK_MAGIC); - disk_super->version = cpu_to_le32(CACHE_VERSION); + disk_super->version = cpu_to_le32(MAX_CACHE_VERSION); memset(disk_super->policy_name, 0, sizeof(disk_super->policy_name)); memset(disk_super->policy_version, 0, sizeof(disk_super->policy_version)); disk_super->policy_hint_size = 0; - r = dm_sm_copy_root(cmd->metadata_sm, &disk_super->metadata_space_map_root, - metadata_len); - if (r < 0) - goto bad_locked; + __copy_sm_root(cmd, disk_super); disk_super->mapping_root = cpu_to_le64(cmd->root); disk_super->hint_root = cpu_to_le64(cmd->hint_root); disk_super->discard_root = cpu_to_le64(cmd->discard_root); disk_super->discard_block_size = cpu_to_le64(cmd->discard_block_size); - disk_super->discard_nr_blocks = cpu_to_le64(from_dblock(cmd->discard_nr_blocks)); + disk_super->discard_nr_blocks = cpu_to_le64(from_oblock(cmd->discard_nr_blocks)); disk_super->metadata_block_size = cpu_to_le32(DM_CACHE_METADATA_BLOCK_SIZE >> SECTOR_SHIFT); disk_super->data_block_size = cpu_to_le32(cmd->data_block_size); disk_super->cache_blocks = cpu_to_le32(0); @@ -295,10 +340,6 @@ static int __write_initial_superblock(struct dm_cache_metadata *cmd) disk_super->write_misses = cpu_to_le32(0); return dm_tm_commit(cmd->tm, sblock); - -bad_locked: - dm_bm_unlock(sblock); - return r; } static int __format_metadata(struct dm_cache_metadata *cmd) @@ -384,6 +425,15 @@ static int __open_metadata(struct dm_cache_metadata *cmd) disk_super = dm_block_data(sblock); + /* Verify the data block size hasn't changed */ + if (le32_to_cpu(disk_super->data_block_size) != cmd->data_block_size) { + DMERR("changing the data block size (from %u to %llu) is not supported", + le32_to_cpu(disk_super->data_block_size), + (unsigned long long)cmd->data_block_size); + r = -EINVAL; + goto bad; + } + r = __check_incompat_features(disk_super, cmd); if (r < 0) goto bad; @@ -411,7 +461,8 @@ bad: static int __open_or_format_metadata(struct dm_cache_metadata *cmd, bool format_device) { - int r, unformatted; + int r; + bool unformatted = false; r = __superblock_all_zeroes(cmd->bm, &unformatted); if (r) @@ -477,7 +528,7 @@ static void read_superblock_fields(struct dm_cache_metadata *cmd, cmd->hint_root = le64_to_cpu(disk_super->hint_root); cmd->discard_root = le64_to_cpu(disk_super->discard_root); cmd->discard_block_size = le64_to_cpu(disk_super->discard_block_size); - cmd->discard_nr_blocks = to_dblock(le64_to_cpu(disk_super->discard_nr_blocks)); + cmd->discard_nr_blocks = to_oblock(le64_to_cpu(disk_super->discard_nr_blocks)); cmd->data_block_size = le32_to_cpu(disk_super->data_block_size); cmd->cache_blocks = to_cblock(le32_to_cpu(disk_super->cache_blocks)); strncpy(cmd->policy_name, disk_super->policy_name, sizeof(cmd->policy_name)); @@ -511,8 +562,9 @@ static int __begin_transaction_flags(struct dm_cache_metadata *cmd, disk_super = dm_block_data(sblock); update_flags(disk_super, mutator); read_superblock_fields(cmd, disk_super); + dm_bm_unlock(sblock); - return dm_bm_flush_and_unlock(cmd->bm, sblock); + return dm_bm_flush(cmd->bm); } static int __begin_transaction(struct dm_cache_metadata *cmd) @@ -540,7 +592,6 @@ static int __commit_transaction(struct dm_cache_metadata *cmd, flags_mutator mutator) { int r; - size_t metadata_len; struct cache_disk_superblock *disk_super; struct dm_block *sblock; @@ -558,8 +609,8 @@ static int __commit_transaction(struct dm_cache_metadata *cmd, if (r < 0) return r; - r = dm_sm_root_size(cmd->metadata_sm, &metadata_len); - if (r < 0) + r = __save_sm_root(cmd); + if (r) return r; r = superblock_lock(cmd, &sblock); @@ -575,7 +626,7 @@ static int __commit_transaction(struct dm_cache_metadata *cmd, disk_super->hint_root = cpu_to_le64(cmd->hint_root); disk_super->discard_root = cpu_to_le64(cmd->discard_root); disk_super->discard_block_size = cpu_to_le64(cmd->discard_block_size); - disk_super->discard_nr_blocks = cpu_to_le64(from_dblock(cmd->discard_nr_blocks)); + disk_super->discard_nr_blocks = cpu_to_le64(from_oblock(cmd->discard_nr_blocks)); disk_super->cache_blocks = cpu_to_le32(from_cblock(cmd->cache_blocks)); strncpy(disk_super->policy_name, cmd->policy_name, sizeof(disk_super->policy_name)); disk_super->policy_version[0] = cpu_to_le32(cmd->policy_version[0]); @@ -586,13 +637,7 @@ static int __commit_transaction(struct dm_cache_metadata *cmd, disk_super->read_misses = cpu_to_le32(cmd->stats.read_misses); disk_super->write_hits = cpu_to_le32(cmd->stats.write_hits); disk_super->write_misses = cpu_to_le32(cmd->stats.write_misses); - - r = dm_sm_copy_root(cmd->metadata_sm, &disk_super->metadata_space_map_root, - metadata_len); - if (r < 0) { - dm_bm_unlock(sblock); - return r; - } + __copy_sm_root(cmd, disk_super); return dm_tm_commit(cmd->tm, sblock); } @@ -666,19 +711,85 @@ void dm_cache_metadata_close(struct dm_cache_metadata *cmd) kfree(cmd); } +/* + * Checks that the given cache block is either unmapped or clean. + */ +static int block_unmapped_or_clean(struct dm_cache_metadata *cmd, dm_cblock_t b, + bool *result) +{ + int r; + __le64 value; + dm_oblock_t ob; + unsigned flags; + + r = dm_array_get_value(&cmd->info, cmd->root, from_cblock(b), &value); + if (r) { + DMERR("block_unmapped_or_clean failed"); + return r; + } + + unpack_value(value, &ob, &flags); + *result = !((flags & M_VALID) && (flags & M_DIRTY)); + + return 0; +} + +static int blocks_are_unmapped_or_clean(struct dm_cache_metadata *cmd, + dm_cblock_t begin, dm_cblock_t end, + bool *result) +{ + int r; + *result = true; + + while (begin != end) { + r = block_unmapped_or_clean(cmd, begin, result); + if (r) + return r; + + if (!*result) { + DMERR("cache block %llu is dirty", + (unsigned long long) from_cblock(begin)); + return 0; + } + + begin = to_cblock(from_cblock(begin) + 1); + } + + return 0; +} + int dm_cache_resize(struct dm_cache_metadata *cmd, dm_cblock_t new_cache_size) { int r; + bool clean; __le64 null_mapping = pack_value(0, 0); down_write(&cmd->root_lock); __dm_bless_for_disk(&null_mapping); + + if (from_cblock(new_cache_size) < from_cblock(cmd->cache_blocks)) { + r = blocks_are_unmapped_or_clean(cmd, new_cache_size, cmd->cache_blocks, &clean); + if (r) { + __dm_unbless_for_disk(&null_mapping); + goto out; + } + + if (!clean) { + DMERR("unable to shrink cache due to dirty blocks"); + r = -EINVAL; + __dm_unbless_for_disk(&null_mapping); + goto out; + } + } + r = dm_array_resize(&cmd->info, cmd->root, from_cblock(cmd->cache_blocks), from_cblock(new_cache_size), &null_mapping, &cmd->root); if (!r) cmd->cache_blocks = new_cache_size; cmd->changed = true; + +out: up_write(&cmd->root_lock); return r; @@ -686,15 +797,15 @@ int dm_cache_resize(struct dm_cache_metadata *cmd, dm_cblock_t new_cache_size) int dm_cache_discard_bitset_resize(struct dm_cache_metadata *cmd, sector_t discard_block_size, - dm_dblock_t new_nr_entries) + dm_oblock_t new_nr_entries) { int r; down_write(&cmd->root_lock); r = dm_bitset_resize(&cmd->discard_info, cmd->discard_root, - from_dblock(cmd->discard_nr_blocks), - from_dblock(new_nr_entries), + from_oblock(cmd->discard_nr_blocks), + from_oblock(new_nr_entries), false, &cmd->discard_root); if (!r) { cmd->discard_block_size = discard_block_size; @@ -707,28 +818,28 @@ int dm_cache_discard_bitset_resize(struct dm_cache_metadata *cmd, return r; } -static int __set_discard(struct dm_cache_metadata *cmd, dm_dblock_t b) +static int __set_discard(struct dm_cache_metadata *cmd, dm_oblock_t b) { return dm_bitset_set_bit(&cmd->discard_info, cmd->discard_root, - from_dblock(b), &cmd->discard_root); + from_oblock(b), &cmd->discard_root); } -static int __clear_discard(struct dm_cache_metadata *cmd, dm_dblock_t b) +static int __clear_discard(struct dm_cache_metadata *cmd, dm_oblock_t b) { return dm_bitset_clear_bit(&cmd->discard_info, cmd->discard_root, - from_dblock(b), &cmd->discard_root); + from_oblock(b), &cmd->discard_root); } -static int __is_discarded(struct dm_cache_metadata *cmd, dm_dblock_t b, +static int __is_discarded(struct dm_cache_metadata *cmd, dm_oblock_t b, bool *is_discarded) { return dm_bitset_test_bit(&cmd->discard_info, cmd->discard_root, - from_dblock(b), &cmd->discard_root, + from_oblock(b), &cmd->discard_root, is_discarded); } static int __discard(struct dm_cache_metadata *cmd, - dm_dblock_t dblock, bool discard) + dm_oblock_t dblock, bool discard) { int r; @@ -741,7 +852,7 @@ static int __discard(struct dm_cache_metadata *cmd, } int dm_cache_set_discard(struct dm_cache_metadata *cmd, - dm_dblock_t dblock, bool discard) + dm_oblock_t dblock, bool discard) { int r; @@ -759,8 +870,8 @@ static int __load_discards(struct dm_cache_metadata *cmd, dm_block_t b; bool discard; - for (b = 0; b < from_dblock(cmd->discard_nr_blocks); b++) { - dm_dblock_t dblock = to_dblock(b); + for (b = 0; b < from_oblock(cmd->discard_nr_blocks); b++) { + dm_oblock_t dblock = to_oblock(b); if (cmd->clean_when_opened) { r = __is_discarded(cmd, dblock, &discard); @@ -1143,22 +1254,12 @@ static int begin_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *po return 0; } -int dm_cache_begin_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *policy) +static int save_hint(void *context, dm_cblock_t cblock, dm_oblock_t oblock, uint32_t hint) { + struct dm_cache_metadata *cmd = context; + __le32 value = cpu_to_le32(hint); int r; - down_write(&cmd->root_lock); - r = begin_hints(cmd, policy); - up_write(&cmd->root_lock); - - return r; -} - -static int save_hint(struct dm_cache_metadata *cmd, dm_cblock_t cblock, - uint32_t hint) -{ - int r; - __le32 value = cpu_to_le32(hint); __dm_bless_for_disk(&value); r = dm_array_set_value(&cmd->hint_info, cmd->hint_root, @@ -1168,17 +1269,31 @@ static int save_hint(struct dm_cache_metadata *cmd, dm_cblock_t cblock, return r; } -int dm_cache_save_hint(struct dm_cache_metadata *cmd, dm_cblock_t cblock, - uint32_t hint) +static int write_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *policy) { int r; - if (!hints_array_initialized(cmd)) - return 0; + r = begin_hints(cmd, policy); + if (r) { + DMERR("begin_hints failed"); + return r; + } + + return policy_walk_mappings(policy, save_hint, cmd); +} + +int dm_cache_write_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *policy) +{ + int r; down_write(&cmd->root_lock); - r = save_hint(cmd, cblock, hint); + r = write_hints(cmd, policy); up_write(&cmd->root_lock); return r; } + +int dm_cache_metadata_all_clean(struct dm_cache_metadata *cmd, bool *result) +{ + return blocks_are_unmapped_or_clean(cmd, 0, cmd->cache_blocks, result); +} diff --git a/drivers/md/dm-cache-metadata.h b/drivers/md/dm-cache-metadata.h index f45cef21f3d..cd70a78623a 100644 --- a/drivers/md/dm-cache-metadata.h +++ b/drivers/md/dm-cache-metadata.h @@ -72,14 +72,14 @@ dm_cblock_t dm_cache_size(struct dm_cache_metadata *cmd); int dm_cache_discard_bitset_resize(struct dm_cache_metadata *cmd, sector_t discard_block_size, - dm_dblock_t new_nr_entries); + dm_oblock_t new_nr_entries); typedef int (*load_discard_fn)(void *context, sector_t discard_block_size, - dm_dblock_t dblock, bool discarded); + dm_oblock_t dblock, bool discarded); int dm_cache_load_discards(struct dm_cache_metadata *cmd, load_discard_fn fn, void *context); -int dm_cache_set_discard(struct dm_cache_metadata *cmd, dm_dblock_t dblock, bool discard); +int dm_cache_set_discard(struct dm_cache_metadata *cmd, dm_oblock_t dblock, bool discard); int dm_cache_remove_mapping(struct dm_cache_metadata *cmd, dm_cblock_t cblock); int dm_cache_insert_mapping(struct dm_cache_metadata *cmd, dm_cblock_t cblock, dm_oblock_t oblock); @@ -128,14 +128,12 @@ void dm_cache_dump(struct dm_cache_metadata *cmd); * rather than querying the policy for each cblock, we let it walk its data * structures and fill in the hints in whatever order it wishes. */ - -int dm_cache_begin_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *p); +int dm_cache_write_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *p); /* - * requests hints for every cblock and stores in the metadata device. + * Query method. Are all the blocks in the cache clean? */ -int dm_cache_save_hint(struct dm_cache_metadata *cmd, - dm_cblock_t cblock, uint32_t hint); +int dm_cache_metadata_all_clean(struct dm_cache_metadata *cmd, bool *result); /*----------------------------------------------------------------*/ diff --git a/drivers/md/dm-cache-policy-internal.h b/drivers/md/dm-cache-policy-internal.h index 0928abdc49f..2256a1f24f7 100644 --- a/drivers/md/dm-cache-policy-internal.h +++ b/drivers/md/dm-cache-policy-internal.h @@ -61,7 +61,12 @@ static inline int policy_writeback_work(struct dm_cache_policy *p, static inline void policy_remove_mapping(struct dm_cache_policy *p, dm_oblock_t oblock) { - return p->remove_mapping(p, oblock); + p->remove_mapping(p, oblock); +} + +static inline int policy_remove_cblock(struct dm_cache_policy *p, dm_cblock_t cblock) +{ + return p->remove_cblock(p, cblock); } static inline void policy_force_mapping(struct dm_cache_policy *p, diff --git a/drivers/md/dm-cache-policy-mq.c b/drivers/md/dm-cache-policy-mq.c index dc112a7137f..0e385e40909 100644 --- a/drivers/md/dm-cache-policy-mq.c +++ b/drivers/md/dm-cache-policy-mq.c @@ -26,19 +26,6 @@ static unsigned next_power(unsigned n, unsigned min) /*----------------------------------------------------------------*/ -static unsigned long *alloc_bitset(unsigned nr_entries) -{ - size_t s = sizeof(unsigned long) * dm_div_up(nr_entries, BITS_PER_LONG); - return vzalloc(s); -} - -static void free_bitset(unsigned long *bits) -{ - vfree(bits); -} - -/*----------------------------------------------------------------*/ - /* * Large, sequential ios are probably better left on the origin device since * spindles tend to have good bandwidth. @@ -85,7 +72,7 @@ static enum io_pattern iot_pattern(struct io_tracker *t) static void iot_update_stats(struct io_tracker *t, struct bio *bio) { - if (bio->bi_sector == from_oblock(t->last_end_oblock) + 1) + if (bio->bi_iter.bi_sector == from_oblock(t->last_end_oblock) + 1) t->nr_seq_samples++; else { /* @@ -100,7 +87,7 @@ static void iot_update_stats(struct io_tracker *t, struct bio *bio) t->nr_rand_samples++; } - t->last_end_oblock = to_oblock(bio->bi_sector + bio_sectors(bio) - 1); + t->last_end_oblock = to_oblock(bio_end_sector(bio) - 1); } static void iot_check_for_pattern_switch(struct io_tracker *t) @@ -151,6 +138,21 @@ static void queue_init(struct queue *q) } /* + * Checks to see if the queue is empty. + * FIXME: reduce cpu usage. + */ +static bool queue_empty(struct queue *q) +{ + unsigned i; + + for (i = 0; i < NR_QUEUE_LEVELS; i++) + if (!list_empty(q->qs + i)) + return false; + + return true; +} + +/* * Insert an entry to the back of the given level. */ static void queue_push(struct queue *q, unsigned level, struct list_head *elt) @@ -218,17 +220,115 @@ struct entry { struct hlist_node hlist; struct list_head list; dm_oblock_t oblock; - dm_cblock_t cblock; /* valid iff in_cache */ /* * FIXME: pack these better */ - bool in_cache:1; + bool dirty:1; unsigned hit_count; unsigned generation; unsigned tick; }; +/* + * Rather than storing the cblock in an entry, we allocate all entries in + * an array, and infer the cblock from the entry position. + * + * Free entries are linked together into a list. + */ +struct entry_pool { + struct entry *entries, *entries_end; + struct list_head free; + unsigned nr_allocated; +}; + +static int epool_init(struct entry_pool *ep, unsigned nr_entries) +{ + unsigned i; + + ep->entries = vzalloc(sizeof(struct entry) * nr_entries); + if (!ep->entries) + return -ENOMEM; + + ep->entries_end = ep->entries + nr_entries; + + INIT_LIST_HEAD(&ep->free); + for (i = 0; i < nr_entries; i++) + list_add(&ep->entries[i].list, &ep->free); + + ep->nr_allocated = 0; + + return 0; +} + +static void epool_exit(struct entry_pool *ep) +{ + vfree(ep->entries); +} + +static struct entry *alloc_entry(struct entry_pool *ep) +{ + struct entry *e; + + if (list_empty(&ep->free)) + return NULL; + + e = list_entry(list_pop(&ep->free), struct entry, list); + INIT_LIST_HEAD(&e->list); + INIT_HLIST_NODE(&e->hlist); + ep->nr_allocated++; + + return e; +} + +/* + * This assumes the cblock hasn't already been allocated. + */ +static struct entry *alloc_particular_entry(struct entry_pool *ep, dm_cblock_t cblock) +{ + struct entry *e = ep->entries + from_cblock(cblock); + + list_del_init(&e->list); + INIT_HLIST_NODE(&e->hlist); + ep->nr_allocated++; + + return e; +} + +static void free_entry(struct entry_pool *ep, struct entry *e) +{ + BUG_ON(!ep->nr_allocated); + ep->nr_allocated--; + INIT_HLIST_NODE(&e->hlist); + list_add(&e->list, &ep->free); +} + +/* + * Returns NULL if the entry is free. + */ +static struct entry *epool_find(struct entry_pool *ep, dm_cblock_t cblock) +{ + struct entry *e = ep->entries + from_cblock(cblock); + return !hlist_unhashed(&e->hlist) ? e : NULL; +} + +static bool epool_empty(struct entry_pool *ep) +{ + return list_empty(&ep->free); +} + +static bool in_pool(struct entry_pool *ep, struct entry *e) +{ + return e >= ep->entries && e < ep->entries_end; +} + +static dm_cblock_t infer_cblock(struct entry_pool *ep, struct entry *e) +{ + return to_cblock(e - ep->entries); +} + +/*----------------------------------------------------------------*/ + struct mq_policy { struct dm_cache_policy policy; @@ -238,13 +338,22 @@ struct mq_policy { struct io_tracker tracker; /* - * We maintain two queues of entries. The cache proper contains - * the currently active mappings. Whereas the pre_cache tracks - * blocks that are being hit frequently and potential candidates - * for promotion to the cache. + * Entries come from two pools, one of pre-cache entries, and one + * for the cache proper. + */ + struct entry_pool pre_cache_pool; + struct entry_pool cache_pool; + + /* + * We maintain three queues of entries. The cache proper, + * consisting of a clean and dirty queue, contains the currently + * active mappings. Whereas the pre_cache tracks blocks that + * are being hit frequently and potential candidates for promotion + * to the cache. */ struct queue pre_cache; - struct queue cache; + struct queue cache_clean; + struct queue cache_dirty; /* * Keeps track of time, incremented by the core. We use this to @@ -281,24 +390,9 @@ struct mq_policy { */ unsigned promote_threshold; - /* - * We need cache_size entries for the cache, and choose to have - * cache_size entries for the pre_cache too. One motivation for - * using the same size is to make the hit counts directly - * comparable between pre_cache and cache. - */ - unsigned nr_entries; - unsigned nr_entries_allocated; - struct list_head free; - - /* - * Cache blocks may be unallocated. We store this info in a - * bitset. - */ - unsigned long *allocation_bitset; - unsigned nr_cblocks_allocated; - unsigned find_free_nr_words; - unsigned find_free_last_word; + unsigned discard_promote_adjustment; + unsigned read_promote_adjustment; + unsigned write_promote_adjustment; /* * The hash table allows us to quickly find an entry by origin @@ -309,48 +403,9 @@ struct mq_policy { struct hlist_head *table; }; -/*----------------------------------------------------------------*/ -/* Free/alloc mq cache entry structures. */ -static void takeout_queue(struct list_head *lh, struct queue *q) -{ - unsigned level; - - for (level = 0; level < NR_QUEUE_LEVELS; level++) - list_splice(q->qs + level, lh); -} - -static void free_entries(struct mq_policy *mq) -{ - struct entry *e, *tmp; - - takeout_queue(&mq->free, &mq->pre_cache); - takeout_queue(&mq->free, &mq->cache); - - list_for_each_entry_safe(e, tmp, &mq->free, list) - kmem_cache_free(mq_entry_cache, e); -} - -static int alloc_entries(struct mq_policy *mq, unsigned elts) -{ - unsigned u = mq->nr_entries; - - INIT_LIST_HEAD(&mq->free); - mq->nr_entries_allocated = 0; - - while (u--) { - struct entry *e = kmem_cache_zalloc(mq_entry_cache, GFP_KERNEL); - - if (!e) { - free_entries(mq); - return -ENOMEM; - } - - - list_add(&e->list, &mq->free); - } - - return 0; -} +#define DEFAULT_DISCARD_PROMOTE_ADJUSTMENT 1 +#define DEFAULT_READ_PROMOTE_ADJUSTMENT 4 +#define DEFAULT_WRITE_PROMOTE_ADJUSTMENT 8 /*----------------------------------------------------------------*/ @@ -388,96 +443,14 @@ static void hash_remove(struct entry *e) /*----------------------------------------------------------------*/ -/* - * Allocates a new entry structure. The memory is allocated in one lump, - * so we just handing it out here. Returns NULL if all entries have - * already been allocated. Cannot fail otherwise. - */ -static struct entry *alloc_entry(struct mq_policy *mq) -{ - struct entry *e; - - if (mq->nr_entries_allocated >= mq->nr_entries) { - BUG_ON(!list_empty(&mq->free)); - return NULL; - } - - e = list_entry(list_pop(&mq->free), struct entry, list); - INIT_LIST_HEAD(&e->list); - INIT_HLIST_NODE(&e->hlist); - - mq->nr_entries_allocated++; - return e; -} - -/*----------------------------------------------------------------*/ - -/* - * Mark cache blocks allocated or not in the bitset. - */ -static void alloc_cblock(struct mq_policy *mq, dm_cblock_t cblock) -{ - BUG_ON(from_cblock(cblock) > from_cblock(mq->cache_size)); - BUG_ON(test_bit(from_cblock(cblock), mq->allocation_bitset)); - - set_bit(from_cblock(cblock), mq->allocation_bitset); - mq->nr_cblocks_allocated++; -} - -static void free_cblock(struct mq_policy *mq, dm_cblock_t cblock) -{ - BUG_ON(from_cblock(cblock) > from_cblock(mq->cache_size)); - BUG_ON(!test_bit(from_cblock(cblock), mq->allocation_bitset)); - - clear_bit(from_cblock(cblock), mq->allocation_bitset); - mq->nr_cblocks_allocated--; -} - static bool any_free_cblocks(struct mq_policy *mq) { - return mq->nr_cblocks_allocated < from_cblock(mq->cache_size); -} - -/* - * Fills result out with a cache block that isn't in use, or return - * -ENOSPC. This does _not_ mark the cblock as allocated, the caller is - * reponsible for that. - */ -static int __find_free_cblock(struct mq_policy *mq, unsigned begin, unsigned end, - dm_cblock_t *result, unsigned *last_word) -{ - int r = -ENOSPC; - unsigned w; - - for (w = begin; w < end; w++) { - /* - * ffz is undefined if no zero exists - */ - if (mq->allocation_bitset[w] != ~0UL) { - *last_word = w; - *result = to_cblock((w * BITS_PER_LONG) + ffz(mq->allocation_bitset[w])); - if (from_cblock(*result) < from_cblock(mq->cache_size)) - r = 0; - - break; - } - } - - return r; + return !epool_empty(&mq->cache_pool); } -static int find_free_cblock(struct mq_policy *mq, dm_cblock_t *result) +static bool any_clean_cblocks(struct mq_policy *mq) { - int r; - - if (!any_free_cblocks(mq)) - return -ENOSPC; - - r = __find_free_cblock(mq, mq->find_free_last_word, mq->find_free_nr_words, result, &mq->find_free_last_word); - if (r == -ENOSPC && mq->find_free_last_word) - r = __find_free_cblock(mq, 0, mq->find_free_last_word, result, &mq->find_free_last_word); - - return r; + return !queue_empty(&mq->cache_clean); } /*----------------------------------------------------------------*/ @@ -496,33 +469,35 @@ static unsigned queue_level(struct entry *e) return min((unsigned) ilog2(e->hit_count), NR_QUEUE_LEVELS - 1u); } +static bool in_cache(struct mq_policy *mq, struct entry *e) +{ + return in_pool(&mq->cache_pool, e); +} + /* * Inserts the entry into the pre_cache or the cache. Ensures the cache - * block is marked as allocated if necc. Inserts into the hash table. Sets the - * tick which records when the entry was last moved about. + * block is marked as allocated if necc. Inserts into the hash table. + * Sets the tick which records when the entry was last moved about. */ static void push(struct mq_policy *mq, struct entry *e) { e->tick = mq->tick; hash_insert(mq, e); - if (e->in_cache) { - alloc_cblock(mq, e->cblock); - queue_push(&mq->cache, queue_level(e), &e->list); - } else + if (in_cache(mq, e)) + queue_push(e->dirty ? &mq->cache_dirty : &mq->cache_clean, + queue_level(e), &e->list); + else queue_push(&mq->pre_cache, queue_level(e), &e->list); } /* * Removes an entry from pre_cache or cache. Removes from the hash table. - * Frees off the cache block if necc. */ static void del(struct mq_policy *mq, struct entry *e) { queue_remove(&e->list); hash_remove(e); - if (e->in_cache) - free_cblock(mq, e->cblock); } /* @@ -531,14 +506,14 @@ static void del(struct mq_policy *mq, struct entry *e) */ static struct entry *pop(struct mq_policy *mq, struct queue *q) { - struct entry *e = container_of(queue_pop(q), struct entry, list); + struct entry *e; + struct list_head *h = queue_pop(q); - if (e) { - hash_remove(e); + if (!h) + return NULL; - if (e->in_cache) - free_cblock(mq, e->cblock); - } + e = container_of(h, struct entry, list); + hash_remove(e); return e; } @@ -556,7 +531,8 @@ static bool updated_this_tick(struct mq_policy *mq, struct entry *e) * of the entries. * * At the moment the threshold is taken by averaging the hit counts of some - * of the entries in the cache (the first 20 entries of the first level). + * of the entries in the cache (the first 20 entries across all levels in + * ascending order, giving preference to the clean entries at each level). * * We can be much cleverer than this though. For example, each promotion * could bump up the threshold helping to prevent churn. Much more to do @@ -571,14 +547,21 @@ static void check_generation(struct mq_policy *mq) struct list_head *head; struct entry *e; - if ((mq->hit_count >= mq->generation_period) && - (mq->nr_cblocks_allocated == from_cblock(mq->cache_size))) { - + if ((mq->hit_count >= mq->generation_period) && (epool_empty(&mq->cache_pool))) { mq->hit_count = 0; mq->generation++; for (level = 0; level < NR_QUEUE_LEVELS && count < MAX_TO_AVERAGE; level++) { - head = mq->cache.qs + level; + head = mq->cache_clean.qs + level; + list_for_each_entry(e, head, list) { + nr++; + total += e->hit_count; + + if (++count >= MAX_TO_AVERAGE) + break; + } + + head = mq->cache_dirty.qs + level; list_for_each_entry(e, head, list) { nr++; total += e->hit_count; @@ -631,19 +614,30 @@ static void requeue_and_update_tick(struct mq_policy *mq, struct entry *e) * - set the hit count to a hard coded value other than 1, eg, is it better * if it goes in at level 2? */ -static dm_cblock_t demote_cblock(struct mq_policy *mq, dm_oblock_t *oblock) +static int demote_cblock(struct mq_policy *mq, dm_oblock_t *oblock) { - dm_cblock_t result; - struct entry *demoted = pop(mq, &mq->cache); + struct entry *demoted = pop(mq, &mq->cache_clean); + + if (!demoted) + /* + * We could get a block from mq->cache_dirty, but that + * would add extra latency to the triggering bio as it + * waits for the writeback. Better to not promote this + * time and hope there's a clean block next time this block + * is hit. + */ + return -ENOSPC; - BUG_ON(!demoted); - result = demoted->cblock; *oblock = demoted->oblock; - demoted->in_cache = false; - demoted->hit_count = 1; - push(mq, demoted); + free_entry(&mq->cache_pool, demoted); + + /* + * We used to put the demoted block into the pre-cache, but I think + * it's simpler to just let it work it's way up from zero again. + * Stops blocks flickering in and out of the cache. + */ - return result; + return 0; } /* @@ -655,24 +649,21 @@ static dm_cblock_t demote_cblock(struct mq_policy *mq, dm_oblock_t *oblock) * We bias towards reads, since they can be demoted at no cost if they * haven't been dirtied. */ -#define DISCARDED_PROMOTE_THRESHOLD 1 -#define READ_PROMOTE_THRESHOLD 4 -#define WRITE_PROMOTE_THRESHOLD 8 - static unsigned adjusted_promote_threshold(struct mq_policy *mq, bool discarded_oblock, int data_dir) { - if (discarded_oblock && any_free_cblocks(mq) && data_dir == WRITE) + if (data_dir == READ) + return mq->promote_threshold + mq->read_promote_adjustment; + + if (discarded_oblock && (any_free_cblocks(mq) || any_clean_cblocks(mq))) { /* * We don't need to do any copying at all, so give this a - * very low threshold. In practice this only triggers - * during initial population after a format. + * very low threshold. */ - return DISCARDED_PROMOTE_THRESHOLD; + return mq->discard_promote_adjustment; + } - return data_dir == READ ? - (mq->promote_threshold + READ_PROMOTE_THRESHOLD) : - (mq->promote_threshold + WRITE_PROMOTE_THRESHOLD); + return mq->promote_threshold + mq->write_promote_adjustment; } static bool should_promote(struct mq_policy *mq, struct entry *e, @@ -688,34 +679,49 @@ static int cache_entry_found(struct mq_policy *mq, { requeue_and_update_tick(mq, e); - if (e->in_cache) { + if (in_cache(mq, e)) { result->op = POLICY_HIT; - result->cblock = e->cblock; + result->cblock = infer_cblock(&mq->cache_pool, e); } return 0; } /* - * Moves and entry from the pre_cache to the cache. The main work is + * Moves an entry from the pre_cache to the cache. The main work is * finding which cache block to use. */ static int pre_cache_to_cache(struct mq_policy *mq, struct entry *e, struct policy_result *result) { - dm_cblock_t cblock; + int r; + struct entry *new_e; - if (find_free_cblock(mq, &cblock) == -ENOSPC) { + /* Ensure there's a free cblock in the cache */ + if (epool_empty(&mq->cache_pool)) { result->op = POLICY_REPLACE; - cblock = demote_cblock(mq, &result->old_oblock); + r = demote_cblock(mq, &result->old_oblock); + if (r) { + result->op = POLICY_MISS; + return 0; + } } else result->op = POLICY_NEW; - result->cblock = e->cblock = cblock; + new_e = alloc_entry(&mq->cache_pool); + BUG_ON(!new_e); + + new_e->oblock = e->oblock; + new_e->dirty = false; + new_e->hit_count = e->hit_count; + new_e->generation = e->generation; + new_e->tick = e->tick; del(mq, e); - e->in_cache = true; - push(mq, e); + free_entry(&mq->pre_cache_pool, e); + push(mq, new_e); + + result->cblock = infer_cblock(&mq->cache_pool, new_e); return 0; } @@ -727,15 +733,18 @@ static int pre_cache_entry_found(struct mq_policy *mq, struct entry *e, int r = 0; bool updated = updated_this_tick(mq, e); - requeue_and_update_tick(mq, e); - if ((!discarded_oblock && updated) || - !should_promote(mq, e, discarded_oblock, data_dir)) + !should_promote(mq, e, discarded_oblock, data_dir)) { + requeue_and_update_tick(mq, e); result->op = POLICY_MISS; - else if (!can_migrate) + + } else if (!can_migrate) r = -EWOULDBLOCK; - else + + else { + requeue_and_update_tick(mq, e); r = pre_cache_to_cache(mq, e, result); + } return r; } @@ -743,7 +752,7 @@ static int pre_cache_entry_found(struct mq_policy *mq, struct entry *e, static void insert_in_pre_cache(struct mq_policy *mq, dm_oblock_t oblock) { - struct entry *e = alloc_entry(mq); + struct entry *e = alloc_entry(&mq->pre_cache_pool); if (!e) /* @@ -757,7 +766,7 @@ static void insert_in_pre_cache(struct mq_policy *mq, return; } - e->in_cache = false; + e->dirty = false; e->oblock = oblock; e->hit_count = 1; e->generation = mq->generation; @@ -767,37 +776,43 @@ static void insert_in_pre_cache(struct mq_policy *mq, static void insert_in_cache(struct mq_policy *mq, dm_oblock_t oblock, struct policy_result *result) { + int r; struct entry *e; - dm_cblock_t cblock; - if (find_free_cblock(mq, &cblock) == -ENOSPC) { - result->op = POLICY_MISS; - insert_in_pre_cache(mq, oblock); - return; - } + if (epool_empty(&mq->cache_pool)) { + result->op = POLICY_REPLACE; + r = demote_cblock(mq, &result->old_oblock); + if (unlikely(r)) { + result->op = POLICY_MISS; + insert_in_pre_cache(mq, oblock); + return; + } - e = alloc_entry(mq); - if (unlikely(!e)) { - result->op = POLICY_MISS; - return; + /* + * This will always succeed, since we've just demoted. + */ + e = alloc_entry(&mq->cache_pool); + BUG_ON(!e); + + } else { + e = alloc_entry(&mq->cache_pool); + result->op = POLICY_NEW; } e->oblock = oblock; - e->cblock = cblock; - e->in_cache = true; + e->dirty = false; e->hit_count = 1; e->generation = mq->generation; push(mq, e); - result->op = POLICY_NEW; - result->cblock = e->cblock; + result->cblock = infer_cblock(&mq->cache_pool, e); } static int no_entry_found(struct mq_policy *mq, dm_oblock_t oblock, bool can_migrate, bool discarded_oblock, int data_dir, struct policy_result *result) { - if (adjusted_promote_threshold(mq, discarded_oblock, data_dir) == 1) { + if (adjusted_promote_threshold(mq, discarded_oblock, data_dir) <= 1) { if (can_migrate) insert_in_cache(mq, oblock, result); else @@ -821,13 +836,16 @@ static int map(struct mq_policy *mq, dm_oblock_t oblock, int r = 0; struct entry *e = hash_lookup(mq, oblock); - if (e && e->in_cache) + if (e && in_cache(mq, e)) r = cache_entry_found(mq, e, result); + else if (iot_pattern(&mq->tracker) == PATTERN_SEQUENTIAL) result->op = POLICY_MISS; + else if (e) r = pre_cache_entry_found(mq, e, can_migrate, discarded_oblock, data_dir, result); + else r = no_entry_found(mq, oblock, can_migrate, discarded_oblock, data_dir, result); @@ -854,9 +872,9 @@ static void mq_destroy(struct dm_cache_policy *p) { struct mq_policy *mq = to_mq_policy(p); - free_bitset(mq->allocation_bitset); - kfree(mq->table); - free_entries(mq); + vfree(mq->table); + epool_exit(&mq->cache_pool); + epool_exit(&mq->pre_cache_pool); kfree(mq); } @@ -904,8 +922,8 @@ static int mq_lookup(struct dm_cache_policy *p, dm_oblock_t oblock, dm_cblock_t return -EWOULDBLOCK; e = hash_lookup(mq, oblock); - if (e && e->in_cache) { - *cblock = e->cblock; + if (e && in_cache(mq, e)) { + *cblock = infer_cblock(&mq->cache_pool, e); r = 0; } else r = -ENOENT; @@ -915,6 +933,36 @@ static int mq_lookup(struct dm_cache_policy *p, dm_oblock_t oblock, dm_cblock_t return r; } +static void __mq_set_clear_dirty(struct mq_policy *mq, dm_oblock_t oblock, bool set) +{ + struct entry *e; + + e = hash_lookup(mq, oblock); + BUG_ON(!e || !in_cache(mq, e)); + + del(mq, e); + e->dirty = set; + push(mq, e); +} + +static void mq_set_dirty(struct dm_cache_policy *p, dm_oblock_t oblock) +{ + struct mq_policy *mq = to_mq_policy(p); + + mutex_lock(&mq->lock); + __mq_set_clear_dirty(mq, oblock, true); + mutex_unlock(&mq->lock); +} + +static void mq_clear_dirty(struct dm_cache_policy *p, dm_oblock_t oblock) +{ + struct mq_policy *mq = to_mq_policy(p); + + mutex_lock(&mq->lock); + __mq_set_clear_dirty(mq, oblock, false); + mutex_unlock(&mq->lock); +} + static int mq_load_mapping(struct dm_cache_policy *p, dm_oblock_t oblock, dm_cblock_t cblock, uint32_t hint, bool hint_valid) @@ -922,13 +970,9 @@ static int mq_load_mapping(struct dm_cache_policy *p, struct mq_policy *mq = to_mq_policy(p); struct entry *e; - e = alloc_entry(mq); - if (!e) - return -ENOMEM; - - e->cblock = cblock; + e = alloc_particular_entry(&mq->cache_pool, cblock); e->oblock = oblock; - e->in_cache = true; + e->dirty = false; /* this gets corrected in a minute */ e->hit_count = hint_valid ? hint : 1; e->generation = mq->generation; push(mq, e); @@ -936,38 +980,50 @@ static int mq_load_mapping(struct dm_cache_policy *p, return 0; } +static int mq_save_hints(struct mq_policy *mq, struct queue *q, + policy_walk_fn fn, void *context) +{ + int r; + unsigned level; + struct entry *e; + + for (level = 0; level < NR_QUEUE_LEVELS; level++) + list_for_each_entry(e, q->qs + level, list) { + r = fn(context, infer_cblock(&mq->cache_pool, e), + e->oblock, e->hit_count); + if (r) + return r; + } + + return 0; +} + static int mq_walk_mappings(struct dm_cache_policy *p, policy_walk_fn fn, void *context) { struct mq_policy *mq = to_mq_policy(p); int r = 0; - struct entry *e; - unsigned level; mutex_lock(&mq->lock); - for (level = 0; level < NR_QUEUE_LEVELS; level++) - list_for_each_entry(e, &mq->cache.qs[level], list) { - r = fn(context, e->cblock, e->oblock, e->hit_count); - if (r) - goto out; - } + r = mq_save_hints(mq, &mq->cache_clean, fn, context); + if (!r) + r = mq_save_hints(mq, &mq->cache_dirty, fn, context); -out: mutex_unlock(&mq->lock); return r; } -static void remove_mapping(struct mq_policy *mq, dm_oblock_t oblock) +static void __remove_mapping(struct mq_policy *mq, dm_oblock_t oblock) { - struct entry *e = hash_lookup(mq, oblock); + struct entry *e; - BUG_ON(!e || !e->in_cache); + e = hash_lookup(mq, oblock); + BUG_ON(!e || !in_cache(mq, e)); del(mq, e); - e->in_cache = false; - push(mq, e); + free_entry(&mq->cache_pool, e); } static void mq_remove_mapping(struct dm_cache_policy *p, dm_oblock_t oblock) @@ -975,20 +1031,75 @@ static void mq_remove_mapping(struct dm_cache_policy *p, dm_oblock_t oblock) struct mq_policy *mq = to_mq_policy(p); mutex_lock(&mq->lock); - remove_mapping(mq, oblock); + __remove_mapping(mq, oblock); mutex_unlock(&mq->lock); } -static void force_mapping(struct mq_policy *mq, - dm_oblock_t current_oblock, dm_oblock_t new_oblock) +static int __remove_cblock(struct mq_policy *mq, dm_cblock_t cblock) { - struct entry *e = hash_lookup(mq, current_oblock); + struct entry *e = epool_find(&mq->cache_pool, cblock); - BUG_ON(!e || !e->in_cache); + if (!e) + return -ENODATA; del(mq, e); - e->oblock = new_oblock; + free_entry(&mq->cache_pool, e); + + return 0; +} + +static int mq_remove_cblock(struct dm_cache_policy *p, dm_cblock_t cblock) +{ + int r; + struct mq_policy *mq = to_mq_policy(p); + + mutex_lock(&mq->lock); + r = __remove_cblock(mq, cblock); + mutex_unlock(&mq->lock); + + return r; +} + +static int __mq_writeback_work(struct mq_policy *mq, dm_oblock_t *oblock, + dm_cblock_t *cblock) +{ + struct entry *e = pop(mq, &mq->cache_dirty); + + if (!e) + return -ENODATA; + + *oblock = e->oblock; + *cblock = infer_cblock(&mq->cache_pool, e); + e->dirty = false; push(mq, e); + + return 0; +} + +static int mq_writeback_work(struct dm_cache_policy *p, dm_oblock_t *oblock, + dm_cblock_t *cblock) +{ + int r; + struct mq_policy *mq = to_mq_policy(p); + + mutex_lock(&mq->lock); + r = __mq_writeback_work(mq, oblock, cblock); + mutex_unlock(&mq->lock); + + return r; +} + +static void __force_mapping(struct mq_policy *mq, + dm_oblock_t current_oblock, dm_oblock_t new_oblock) +{ + struct entry *e = hash_lookup(mq, current_oblock); + + if (e && in_cache(mq, e)) { + del(mq, e); + e->oblock = new_oblock; + e->dirty = true; + push(mq, e); + } } static void mq_force_mapping(struct dm_cache_policy *p, @@ -997,16 +1108,20 @@ static void mq_force_mapping(struct dm_cache_policy *p, struct mq_policy *mq = to_mq_policy(p); mutex_lock(&mq->lock); - force_mapping(mq, current_oblock, new_oblock); + __force_mapping(mq, current_oblock, new_oblock); mutex_unlock(&mq->lock); } static dm_cblock_t mq_residency(struct dm_cache_policy *p) { + dm_cblock_t r; struct mq_policy *mq = to_mq_policy(p); - /* FIXME: lock mutex, not sure we can block here */ - return to_cblock(mq->nr_cblocks_allocated); + mutex_lock(&mq->lock); + r = to_cblock(mq->cache_pool.nr_allocated); + mutex_unlock(&mq->lock); + + return r; } static void mq_tick(struct dm_cache_policy *p) @@ -1023,20 +1138,28 @@ static int mq_set_config_value(struct dm_cache_policy *p, const char *key, const char *value) { struct mq_policy *mq = to_mq_policy(p); - enum io_pattern pattern; unsigned long tmp; - if (!strcasecmp(key, "random_threshold")) - pattern = PATTERN_RANDOM; - else if (!strcasecmp(key, "sequential_threshold")) - pattern = PATTERN_SEQUENTIAL; - else - return -EINVAL; - if (kstrtoul(value, 10, &tmp)) return -EINVAL; - mq->tracker.thresholds[pattern] = tmp; + if (!strcasecmp(key, "random_threshold")) { + mq->tracker.thresholds[PATTERN_RANDOM] = tmp; + + } else if (!strcasecmp(key, "sequential_threshold")) { + mq->tracker.thresholds[PATTERN_SEQUENTIAL] = tmp; + + } else if (!strcasecmp(key, "discard_promote_adjustment")) + mq->discard_promote_adjustment = tmp; + + else if (!strcasecmp(key, "read_promote_adjustment")) + mq->read_promote_adjustment = tmp; + + else if (!strcasecmp(key, "write_promote_adjustment")) + mq->write_promote_adjustment = tmp; + + else + return -EINVAL; return 0; } @@ -1046,9 +1169,16 @@ static int mq_emit_config_values(struct dm_cache_policy *p, char *result, unsign ssize_t sz = 0; struct mq_policy *mq = to_mq_policy(p); - DMEMIT("4 random_threshold %u sequential_threshold %u", + DMEMIT("10 random_threshold %u " + "sequential_threshold %u " + "discard_promote_adjustment %u " + "read_promote_adjustment %u " + "write_promote_adjustment %u", mq->tracker.thresholds[PATTERN_RANDOM], - mq->tracker.thresholds[PATTERN_SEQUENTIAL]); + mq->tracker.thresholds[PATTERN_SEQUENTIAL], + mq->discard_promote_adjustment, + mq->read_promote_adjustment, + mq->write_promote_adjustment); return 0; } @@ -1059,10 +1189,13 @@ static void init_policy_functions(struct mq_policy *mq) mq->policy.destroy = mq_destroy; mq->policy.map = mq_map; mq->policy.lookup = mq_lookup; + mq->policy.set_dirty = mq_set_dirty; + mq->policy.clear_dirty = mq_clear_dirty; mq->policy.load_mapping = mq_load_mapping; mq->policy.walk_mappings = mq_walk_mappings; mq->policy.remove_mapping = mq_remove_mapping; - mq->policy.writeback_work = NULL; + mq->policy.remove_cblock = mq_remove_cblock; + mq->policy.writeback_work = mq_writeback_work; mq->policy.force_mapping = mq_force_mapping; mq->policy.residency = mq_residency; mq->policy.tick = mq_tick; @@ -1074,7 +1207,6 @@ static struct dm_cache_policy *mq_create(dm_cblock_t cache_size, sector_t origin_size, sector_t cache_block_size) { - int r; struct mq_policy *mq = kzalloc(sizeof(*mq), GFP_KERNEL); if (!mq) @@ -1082,47 +1214,48 @@ static struct dm_cache_policy *mq_create(dm_cblock_t cache_size, init_policy_functions(mq); iot_init(&mq->tracker, SEQUENTIAL_THRESHOLD_DEFAULT, RANDOM_THRESHOLD_DEFAULT); - mq->cache_size = cache_size; + + if (epool_init(&mq->pre_cache_pool, from_cblock(cache_size))) { + DMERR("couldn't initialize pool of pre-cache entries"); + goto bad_pre_cache_init; + } + + if (epool_init(&mq->cache_pool, from_cblock(cache_size))) { + DMERR("couldn't initialize pool of cache entries"); + goto bad_cache_init; + } + mq->tick_protected = 0; mq->tick = 0; mq->hit_count = 0; mq->generation = 0; mq->promote_threshold = 0; + mq->discard_promote_adjustment = DEFAULT_DISCARD_PROMOTE_ADJUSTMENT; + mq->read_promote_adjustment = DEFAULT_READ_PROMOTE_ADJUSTMENT; + mq->write_promote_adjustment = DEFAULT_WRITE_PROMOTE_ADJUSTMENT; mutex_init(&mq->lock); spin_lock_init(&mq->tick_lock); - mq->find_free_nr_words = dm_div_up(from_cblock(mq->cache_size), BITS_PER_LONG); - mq->find_free_last_word = 0; queue_init(&mq->pre_cache); - queue_init(&mq->cache); - mq->generation_period = max((unsigned) from_cblock(cache_size), 1024U); - - mq->nr_entries = 2 * from_cblock(cache_size); - r = alloc_entries(mq, mq->nr_entries); - if (r) - goto bad_cache_alloc; + queue_init(&mq->cache_clean); + queue_init(&mq->cache_dirty); - mq->nr_entries_allocated = 0; - mq->nr_cblocks_allocated = 0; + mq->generation_period = max((unsigned) from_cblock(cache_size), 1024U); mq->nr_buckets = next_power(from_cblock(cache_size) / 2, 16); mq->hash_bits = ffs(mq->nr_buckets) - 1; - mq->table = kzalloc(sizeof(*mq->table) * mq->nr_buckets, GFP_KERNEL); + mq->table = vzalloc(sizeof(*mq->table) * mq->nr_buckets); if (!mq->table) goto bad_alloc_table; - mq->allocation_bitset = alloc_bitset(from_cblock(cache_size)); - if (!mq->allocation_bitset) - goto bad_alloc_bitset; - return &mq->policy; -bad_alloc_bitset: - kfree(mq->table); bad_alloc_table: - free_entries(mq); -bad_cache_alloc: + epool_exit(&mq->cache_pool); +bad_cache_init: + epool_exit(&mq->pre_cache_pool); +bad_pre_cache_init: kfree(mq); return NULL; @@ -1132,7 +1265,7 @@ bad_cache_alloc: static struct dm_cache_policy_type mq_policy_type = { .name = "mq", - .version = {1, 0, 0}, + .version = {1, 2, 0}, .hint_size = 4, .owner = THIS_MODULE, .create = mq_create @@ -1140,10 +1273,11 @@ static struct dm_cache_policy_type mq_policy_type = { static struct dm_cache_policy_type default_policy_type = { .name = "default", - .version = {1, 0, 0}, + .version = {1, 2, 0}, .hint_size = 4, .owner = THIS_MODULE, - .create = mq_create + .create = mq_create, + .real = &mq_policy_type }; static int __init mq_init(void) diff --git a/drivers/md/dm-cache-policy.c b/drivers/md/dm-cache-policy.c index 21c03c570c0..c1a3cee99b4 100644 --- a/drivers/md/dm-cache-policy.c +++ b/drivers/md/dm-cache-policy.c @@ -119,13 +119,13 @@ struct dm_cache_policy *dm_cache_policy_create(const char *name, type = get_policy(name); if (!type) { DMWARN("unknown policy type"); - return NULL; + return ERR_PTR(-EINVAL); } p = type->create(cache_size, origin_size, cache_block_size); if (!p) { put_policy(type); - return NULL; + return ERR_PTR(-ENOMEM); } p->private = type; @@ -146,6 +146,10 @@ const char *dm_cache_policy_get_name(struct dm_cache_policy *p) { struct dm_cache_policy_type *t = p->private; + /* if t->real is set then an alias was used (e.g. "default") */ + if (t->real) + return t->real->name; + return t->name; } EXPORT_SYMBOL_GPL(dm_cache_policy_get_name); diff --git a/drivers/md/dm-cache-policy.h b/drivers/md/dm-cache-policy.h index 33369ca9614..f50fe360c54 100644 --- a/drivers/md/dm-cache-policy.h +++ b/drivers/md/dm-cache-policy.h @@ -135,9 +135,6 @@ struct dm_cache_policy { */ int (*lookup)(struct dm_cache_policy *p, dm_oblock_t oblock, dm_cblock_t *cblock); - /* - * oblock must be a mapped block. Must not block. - */ void (*set_dirty)(struct dm_cache_policy *p, dm_oblock_t oblock); void (*clear_dirty)(struct dm_cache_policy *p, dm_oblock_t oblock); @@ -159,8 +156,24 @@ struct dm_cache_policy { void (*force_mapping)(struct dm_cache_policy *p, dm_oblock_t current_oblock, dm_oblock_t new_oblock); - int (*writeback_work)(struct dm_cache_policy *p, dm_oblock_t *oblock, dm_cblock_t *cblock); + /* + * This is called via the invalidate_cblocks message. It is + * possible the particular cblock has already been removed due to a + * write io in passthrough mode. In which case this should return + * -ENODATA. + */ + int (*remove_cblock)(struct dm_cache_policy *p, dm_cblock_t cblock); + /* + * Provide a dirty block to be written back by the core target. + * + * Returns: + * + * 0 and @cblock,@oblock: block to write back provided + * + * -ENODATA: no dirty blocks available + */ + int (*writeback_work)(struct dm_cache_policy *p, dm_oblock_t *oblock, dm_cblock_t *cblock); /* * How full is the cache? @@ -210,6 +223,12 @@ struct dm_cache_policy_type { unsigned version[CACHE_POLICY_VERSION_SIZE]; /* + * For use by an alias dm_cache_policy_type to point to the + * real dm_cache_policy_type. + */ + struct dm_cache_policy_type *real; + + /* * Policies may store a hint for each each cache block. * Currently the size of this hint must be 0 or 4 bytes but we * expect to relax this in future. diff --git a/drivers/md/dm-cache-target.c b/drivers/md/dm-cache-target.c index df44b60e66f..2c63326638b 100644 --- a/drivers/md/dm-cache-target.c +++ b/drivers/md/dm-cache-target.c @@ -61,27 +61,86 @@ static void free_bitset(unsigned long *bits) /*----------------------------------------------------------------*/ +/* + * There are a couple of places where we let a bio run, but want to do some + * work before calling its endio function. We do this by temporarily + * changing the endio fn. + */ +struct dm_hook_info { + bio_end_io_t *bi_end_io; + void *bi_private; +}; + +static void dm_hook_bio(struct dm_hook_info *h, struct bio *bio, + bio_end_io_t *bi_end_io, void *bi_private) +{ + h->bi_end_io = bio->bi_end_io; + h->bi_private = bio->bi_private; + + bio->bi_end_io = bi_end_io; + bio->bi_private = bi_private; +} + +static void dm_unhook_bio(struct dm_hook_info *h, struct bio *bio) +{ + bio->bi_end_io = h->bi_end_io; + bio->bi_private = h->bi_private; + + /* + * Must bump bi_remaining to allow bio to complete with + * restored bi_end_io. + */ + atomic_inc(&bio->bi_remaining); +} + +/*----------------------------------------------------------------*/ + #define PRISON_CELLS 1024 #define MIGRATION_POOL_SIZE 128 #define COMMIT_PERIOD HZ #define MIGRATION_COUNT_WINDOW 10 /* - * The block size of the device holding cache data must be >= 32KB + * The block size of the device holding cache data must be + * between 32KB and 1GB. */ #define DATA_DEV_BLOCK_SIZE_MIN_SECTORS (32 * 1024 >> SECTOR_SHIFT) +#define DATA_DEV_BLOCK_SIZE_MAX_SECTORS (1024 * 1024 * 1024 >> SECTOR_SHIFT) /* * FIXME: the cache is read/write for the time being. */ -enum cache_mode { +enum cache_metadata_mode { CM_WRITE, /* metadata may be changed */ CM_READ_ONLY, /* metadata may not be changed */ }; +enum cache_io_mode { + /* + * Data is written to cached blocks only. These blocks are marked + * dirty. If you lose the cache device you will lose data. + * Potential performance increase for both reads and writes. + */ + CM_IO_WRITEBACK, + + /* + * Data is written to both cache and origin. Blocks are never + * dirty. Potential performance benfit for reads only. + */ + CM_IO_WRITETHROUGH, + + /* + * A degraded mode useful for various cache coherency situations + * (eg, rolling back snapshots). Reads and writes always go to the + * origin. If a write goes to a cached oblock, then the cache + * block is invalidated. + */ + CM_IO_PASSTHROUGH +}; + struct cache_features { - enum cache_mode mode; - bool write_through:1; + enum cache_metadata_mode mode; + enum cache_io_mode io_mode; }; struct cache_stats { @@ -97,10 +156,31 @@ struct cache_stats { atomic_t discard_count; }; +/* + * Defines a range of cblocks, begin to (end - 1) are in the range. end is + * the one-past-the-end value. + */ +struct cblock_range { + dm_cblock_t begin; + dm_cblock_t end; +}; + +struct invalidation_request { + struct list_head list; + struct cblock_range *cblocks; + + atomic_t complete; + int err; + + wait_queue_head_t result_wait; +}; + struct cache { struct dm_target *ti; struct dm_target_callbacks callbacks; + struct dm_cache_metadata *cmd; + /* * Metadata is written to this device. */ @@ -117,11 +197,6 @@ struct cache { struct dm_dev *cache_dev; /* - * Cache features such as write-through. - */ - struct cache_features features; - - /* * Size of the origin device in _complete_ blocks and native sectors. */ dm_oblock_t origin_blocks; @@ -138,8 +213,6 @@ struct cache { uint32_t sectors_per_block; int sectors_per_block_shift; - struct dm_cache_metadata *cmd; - spinlock_t lock; struct bio_list deferred_bios; struct bio_list deferred_flush_bios; @@ -148,22 +221,32 @@ struct cache { struct list_head completed_migrations; struct list_head need_commit_migrations; sector_t migration_threshold; - atomic_t nr_migrations; wait_queue_head_t migration_wait; + atomic_t nr_migrations; + + wait_queue_head_t quiescing_wait; + atomic_t quiescing; + atomic_t quiescing_ack; /* * cache_size entries, dirty if set */ - dm_cblock_t nr_dirty; + atomic_t nr_dirty; unsigned long *dirty_bitset; /* * origin_blocks entries, discarded if set. */ - uint32_t discard_block_size; /* a power of 2 times sectors per block */ - dm_dblock_t discard_nr_blocks; + dm_oblock_t discard_nr_blocks; unsigned long *discard_bitset; + /* + * Rather than reconstructing the table line for the status we just + * save it and regurgitate. + */ + unsigned nr_ctr_args; + const char **ctr_args; + struct dm_kcopyd_client *copier; struct workqueue_struct *wq; struct work_struct worker; @@ -182,25 +265,30 @@ struct cache { bool need_tick_bio:1; bool sized:1; - bool quiescing:1; + bool invalidate:1; bool commit_requested:1; bool loaded_mappings:1; bool loaded_discards:1; + /* + * Cache features such as write-through. + */ + struct cache_features features; + struct cache_stats stats; /* - * Rather than reconstructing the table line for the status we just - * save it and regurgitate. + * Invalidation fields. */ - unsigned nr_ctr_args; - const char **ctr_args; + spinlock_t invalidation_lock; + struct list_head invalidation_requests; }; struct per_bio_data { bool tick:1; unsigned req_nr:2; struct dm_deferred_entry *all_io_entry; + struct dm_hook_info hook_info; /* * writethrough fields. These MUST remain at the end of this @@ -209,7 +297,6 @@ struct per_bio_data { */ struct cache *cache; dm_cblock_t cblock; - bio_end_io_t *saved_bi_end_io; struct dm_bio_details bio_details; }; @@ -226,6 +313,8 @@ struct dm_cache_migration { bool writeback:1; bool demote:1; bool promote:1; + bool requeue_holder:1; + bool invalidate:1; struct dm_bio_prison_cell *old_ocell; struct dm_bio_prison_cell *new_ocell; @@ -403,7 +492,7 @@ static bool is_dirty(struct cache *cache, dm_cblock_t b) static void set_dirty(struct cache *cache, dm_oblock_t oblock, dm_cblock_t cblock) { if (!test_and_set_bit(from_cblock(cblock), cache->dirty_bitset)) { - cache->nr_dirty = to_cblock(from_cblock(cache->nr_dirty) + 1); + atomic_inc(&cache->nr_dirty); policy_set_dirty(cache->policy, oblock); } } @@ -412,8 +501,7 @@ static void clear_dirty(struct cache *cache, dm_oblock_t oblock, dm_cblock_t cbl { if (test_and_clear_bit(from_cblock(cblock), cache->dirty_bitset)) { policy_clear_dirty(cache->policy, oblock); - cache->nr_dirty = to_cblock(from_cblock(cache->nr_dirty) - 1); - if (!from_cblock(cache->nr_dirty)) + if (atomic_dec_return(&cache->nr_dirty) == 0) dm_table_event(cache->ti->table); } } @@ -425,6 +513,10 @@ static bool block_size_is_power_of_two(struct cache *cache) return cache->sectors_per_block_shift >= 0; } +/* gcc on ARM generates spurious references to __udivdi3 and __umoddi3 */ +#if defined(CONFIG_ARM) && __GNUC__ == 4 && __GNUC_MINOR__ <= 6 +__always_inline +#endif static dm_block_t block_div(dm_block_t b, uint32_t n) { do_div(b, n); @@ -432,48 +524,33 @@ static dm_block_t block_div(dm_block_t b, uint32_t n) return b; } -static dm_dblock_t oblock_to_dblock(struct cache *cache, dm_oblock_t oblock) -{ - uint32_t discard_blocks = cache->discard_block_size; - dm_block_t b = from_oblock(oblock); - - if (!block_size_is_power_of_two(cache)) - discard_blocks = discard_blocks / cache->sectors_per_block; - else - discard_blocks >>= cache->sectors_per_block_shift; - - b = block_div(b, discard_blocks); - - return to_dblock(b); -} - -static void set_discard(struct cache *cache, dm_dblock_t b) +static void set_discard(struct cache *cache, dm_oblock_t b) { unsigned long flags; atomic_inc(&cache->stats.discard_count); spin_lock_irqsave(&cache->lock, flags); - set_bit(from_dblock(b), cache->discard_bitset); + set_bit(from_oblock(b), cache->discard_bitset); spin_unlock_irqrestore(&cache->lock, flags); } -static void clear_discard(struct cache *cache, dm_dblock_t b) +static void clear_discard(struct cache *cache, dm_oblock_t b) { unsigned long flags; spin_lock_irqsave(&cache->lock, flags); - clear_bit(from_dblock(b), cache->discard_bitset); + clear_bit(from_oblock(b), cache->discard_bitset); spin_unlock_irqrestore(&cache->lock, flags); } -static bool is_discarded(struct cache *cache, dm_dblock_t b) +static bool is_discarded(struct cache *cache, dm_oblock_t b) { int r; unsigned long flags; spin_lock_irqsave(&cache->lock, flags); - r = test_bit(from_dblock(b), cache->discard_bitset); + r = test_bit(from_oblock(b), cache->discard_bitset); spin_unlock_irqrestore(&cache->lock, flags); return r; @@ -485,8 +562,7 @@ static bool is_discarded_oblock(struct cache *cache, dm_oblock_t b) unsigned long flags; spin_lock_irqsave(&cache->lock, flags); - r = test_bit(from_dblock(oblock_to_dblock(cache, b)), - cache->discard_bitset); + r = test_bit(from_oblock(b), cache->discard_bitset); spin_unlock_irqrestore(&cache->lock, flags); return r; @@ -527,9 +603,24 @@ static void save_stats(struct cache *cache) #define PB_DATA_SIZE_WB (offsetof(struct per_bio_data, cache)) #define PB_DATA_SIZE_WT (sizeof(struct per_bio_data)) +static bool writethrough_mode(struct cache_features *f) +{ + return f->io_mode == CM_IO_WRITETHROUGH; +} + +static bool writeback_mode(struct cache_features *f) +{ + return f->io_mode == CM_IO_WRITEBACK; +} + +static bool passthrough_mode(struct cache_features *f) +{ + return f->io_mode == CM_IO_PASSTHROUGH; +} + static size_t get_per_bio_data_size(struct cache *cache) { - return cache->features.write_through ? PB_DATA_SIZE_WT : PB_DATA_SIZE_WB; + return writethrough_mode(&cache->features) ? PB_DATA_SIZE_WT : PB_DATA_SIZE_WB; } static struct per_bio_data *get_per_bio_data(struct bio *bio, size_t data_size) @@ -561,15 +652,18 @@ static void remap_to_origin(struct cache *cache, struct bio *bio) static void remap_to_cache(struct cache *cache, struct bio *bio, dm_cblock_t cblock) { - sector_t bi_sector = bio->bi_sector; + sector_t bi_sector = bio->bi_iter.bi_sector; + sector_t block = from_cblock(cblock); bio->bi_bdev = cache->cache_dev->bdev; if (!block_size_is_power_of_two(cache)) - bio->bi_sector = (from_cblock(cblock) * cache->sectors_per_block) + - sector_div(bi_sector, cache->sectors_per_block); + bio->bi_iter.bi_sector = + (block * cache->sectors_per_block) + + sector_div(bi_sector, cache->sectors_per_block); else - bio->bi_sector = (from_cblock(cblock) << cache->sectors_per_block_shift) | - (bi_sector & (cache->sectors_per_block - 1)); + bio->bi_iter.bi_sector = + (block << cache->sectors_per_block_shift) | + (bi_sector & (cache->sectors_per_block - 1)); } static void check_if_tick_bio_needed(struct cache *cache, struct bio *bio) @@ -593,22 +687,23 @@ static void remap_to_origin_clear_discard(struct cache *cache, struct bio *bio, check_if_tick_bio_needed(cache, bio); remap_to_origin(cache, bio); if (bio_data_dir(bio) == WRITE) - clear_discard(cache, oblock_to_dblock(cache, oblock)); + clear_discard(cache, oblock); } static void remap_to_cache_dirty(struct cache *cache, struct bio *bio, dm_oblock_t oblock, dm_cblock_t cblock) { + check_if_tick_bio_needed(cache, bio); remap_to_cache(cache, bio, cblock); if (bio_data_dir(bio) == WRITE) { set_dirty(cache, oblock, cblock); - clear_discard(cache, oblock_to_dblock(cache, oblock)); + clear_discard(cache, oblock); } } static dm_oblock_t get_bio_block(struct cache *cache, struct bio *bio) { - sector_t block_nr = bio->bi_sector; + sector_t block_nr = bio->bi_iter.bi_sector; if (!block_size_is_power_of_two(cache)) (void) sector_div(block_nr, cache->sectors_per_block); @@ -656,7 +751,8 @@ static void defer_writethrough_bio(struct cache *cache, struct bio *bio) static void writethrough_endio(struct bio *bio, int err) { struct per_bio_data *pb = get_per_bio_data(bio, PB_DATA_SIZE_WT); - bio->bi_end_io = pb->saved_bi_end_io; + + dm_unhook_bio(&pb->hook_info, bio); if (err) { bio_endio(bio, err); @@ -687,9 +783,8 @@ static void remap_to_origin_then_cache(struct cache *cache, struct bio *bio, pb->cache = cache; pb->cblock = cblock; - pb->saved_bi_end_io = bio->bi_end_io; + dm_hook_bio(&pb->hook_info, bio, writethrough_endio, NULL); dm_bio_record(&pb->bio_details, bio); - bio->bi_end_io = writethrough_endio; remap_to_origin_clear_discard(pb->cache, bio, oblock); } @@ -742,8 +837,9 @@ static void cell_defer(struct cache *cache, struct dm_bio_prison_cell *cell, static void cleanup_migration(struct dm_cache_migration *mg) { - dec_nr_migrations(mg->cache); + struct cache *cache = mg->cache; free_migration(mg); + dec_nr_migrations(cache); } static void migration_failure(struct dm_cache_migration *mg) @@ -759,13 +855,13 @@ static void migration_failure(struct dm_cache_migration *mg) DMWARN_LIMIT("demotion failed; couldn't copy block"); policy_force_mapping(cache->policy, mg->new_oblock, mg->old_oblock); - cell_defer(cache, mg->old_ocell, mg->promote ? 0 : 1); + cell_defer(cache, mg->old_ocell, mg->promote ? false : true); if (mg->promote) - cell_defer(cache, mg->new_ocell, 1); + cell_defer(cache, mg->new_ocell, true); } else { DMWARN_LIMIT("promotion failed; couldn't copy block"); policy_remove_mapping(cache->policy, mg->new_oblock); - cell_defer(cache, mg->new_ocell, 1); + cell_defer(cache, mg->new_ocell, true); } cleanup_migration(mg); @@ -817,7 +913,7 @@ static void migration_success_post_commit(struct dm_cache_migration *mg) return; } else if (mg->demote) { - cell_defer(cache, mg->old_ocell, mg->promote ? 0 : 1); + cell_defer(cache, mg->old_ocell, mg->promote ? false : true); if (mg->promote) { mg->demote = false; @@ -826,11 +922,19 @@ static void migration_success_post_commit(struct dm_cache_migration *mg) list_add_tail(&mg->list, &cache->quiesced_migrations); spin_unlock_irqrestore(&cache->lock, flags); - } else + } else { + if (mg->invalidate) + policy_remove_mapping(cache->policy, mg->old_oblock); cleanup_migration(mg); + } } else { - cell_defer(cache, mg->new_ocell, true); + if (mg->requeue_holder) + cell_defer(cache, mg->new_ocell, true); + else { + bio_endio(mg->new_ocell->holder, 0); + cell_defer(cache, mg->new_ocell, false); + } clear_dirty(cache, mg->new_oblock, mg->cblock); cleanup_migration(mg); } @@ -857,12 +961,13 @@ static void issue_copy_real(struct dm_cache_migration *mg) int r; struct dm_io_region o_region, c_region; struct cache *cache = mg->cache; + sector_t cblock = from_cblock(mg->cblock); o_region.bdev = cache->origin_dev->bdev; o_region.count = cache->sectors_per_block; c_region.bdev = cache->cache_dev->bdev; - c_region.sector = from_cblock(mg->cblock) * cache->sectors_per_block; + c_region.sector = cblock * cache->sectors_per_block; c_region.count = cache->sectors_per_block; if (mg->writeback || mg->demote) { @@ -875,8 +980,48 @@ static void issue_copy_real(struct dm_cache_migration *mg) r = dm_kcopyd_copy(cache->copier, &o_region, 1, &c_region, 0, copy_complete, mg); } - if (r < 0) + if (r < 0) { + DMERR_LIMIT("issuing migration failed"); migration_failure(mg); + } +} + +static void overwrite_endio(struct bio *bio, int err) +{ + struct dm_cache_migration *mg = bio->bi_private; + struct cache *cache = mg->cache; + size_t pb_data_size = get_per_bio_data_size(cache); + struct per_bio_data *pb = get_per_bio_data(bio, pb_data_size); + unsigned long flags; + + dm_unhook_bio(&pb->hook_info, bio); + + if (err) + mg->err = true; + + mg->requeue_holder = false; + + spin_lock_irqsave(&cache->lock, flags); + list_add_tail(&mg->list, &cache->completed_migrations); + spin_unlock_irqrestore(&cache->lock, flags); + + wake_worker(cache); +} + +static void issue_overwrite(struct dm_cache_migration *mg, struct bio *bio) +{ + size_t pb_data_size = get_per_bio_data_size(mg->cache); + struct per_bio_data *pb = get_per_bio_data(bio, pb_data_size); + + dm_hook_bio(&pb->hook_info, bio, overwrite_endio, mg); + remap_to_cache_dirty(mg->cache, bio, mg->new_oblock, mg->cblock); + generic_make_request(bio); +} + +static bool bio_writes_complete_block(struct cache *cache, struct bio *bio) +{ + return (bio_data_dir(bio) == WRITE) && + (bio->bi_iter.bi_size == (cache->sectors_per_block << SECTOR_SHIFT)); } static void avoid_copy(struct dm_cache_migration *mg) @@ -893,9 +1038,17 @@ static void issue_copy(struct dm_cache_migration *mg) if (mg->writeback || mg->demote) avoid = !is_dirty(cache, mg->cblock) || is_discarded_oblock(cache, mg->old_oblock); - else + else { + struct bio *bio = mg->new_ocell->holder; + avoid = is_discarded_oblock(cache, mg->new_oblock); + if (!avoid && bio_writes_complete_block(cache, bio)) { + issue_overwrite(mg, bio); + return; + } + } + avoid ? avoid_copy(mg) : issue_copy_real(mg); } @@ -985,6 +1138,8 @@ static void promote(struct cache *cache, struct prealloc *structs, mg->writeback = false; mg->demote = false; mg->promote = true; + mg->requeue_holder = true; + mg->invalidate = false; mg->cache = cache; mg->new_oblock = oblock; mg->cblock = cblock; @@ -1006,6 +1161,8 @@ static void writeback(struct cache *cache, struct prealloc *structs, mg->writeback = true; mg->demote = false; mg->promote = false; + mg->requeue_holder = true; + mg->invalidate = false; mg->cache = cache; mg->old_oblock = oblock; mg->cblock = cblock; @@ -1029,6 +1186,8 @@ static void demote_then_promote(struct cache *cache, struct prealloc *structs, mg->writeback = false; mg->demote = true; mg->promote = true; + mg->requeue_holder = true; + mg->invalidate = false; mg->cache = cache; mg->old_oblock = old_oblock; mg->new_oblock = new_oblock; @@ -1041,6 +1200,33 @@ static void demote_then_promote(struct cache *cache, struct prealloc *structs, quiesce_migration(mg); } +/* + * Invalidate a cache entry. No writeback occurs; any changes in the cache + * block are thrown away. + */ +static void invalidate(struct cache *cache, struct prealloc *structs, + dm_oblock_t oblock, dm_cblock_t cblock, + struct dm_bio_prison_cell *cell) +{ + struct dm_cache_migration *mg = prealloc_get_migration(structs); + + mg->err = false; + mg->writeback = false; + mg->demote = true; + mg->promote = false; + mg->requeue_holder = true; + mg->invalidate = true; + mg->cache = cache; + mg->old_oblock = oblock; + mg->cblock = cblock; + mg->old_ocell = cell; + mg->new_ocell = NULL; + mg->start_jiffies = jiffies; + + inc_nr_migrations(cache); + quiesce_migration(mg); +} + /*---------------------------------------------------------------- * bio processing *--------------------------------------------------------------*/ @@ -1060,7 +1246,7 @@ static void process_flush_bio(struct cache *cache, struct bio *bio) size_t pb_data_size = get_per_bio_data_size(cache); struct per_bio_data *pb = get_per_bio_data(bio, pb_data_size); - BUG_ON(bio->bi_size); + BUG_ON(bio->bi_iter.bi_size); if (!pb->req_nr) remap_to_origin(cache, bio); else @@ -1083,15 +1269,15 @@ static void process_flush_bio(struct cache *cache, struct bio *bio) */ static void process_discard_bio(struct cache *cache, struct bio *bio) { - dm_block_t start_block = dm_sector_div_up(bio->bi_sector, - cache->discard_block_size); - dm_block_t end_block = bio->bi_sector + bio_sectors(bio); + dm_block_t start_block = dm_sector_div_up(bio->bi_iter.bi_sector, + cache->sectors_per_block); + dm_block_t end_block = bio_end_sector(bio); dm_block_t b; - end_block = block_div(end_block, cache->discard_block_size); + end_block = block_div(end_block, cache->sectors_per_block); for (b = start_block; b < end_block; b++) - set_discard(cache, to_dblock(b)); + set_discard(cache, to_oblock(b)); bio_endio(bio, 0); } @@ -1103,13 +1289,6 @@ static bool spare_migration_bandwidth(struct cache *cache) return current_volume < cache->migration_threshold; } -static bool is_writethrough_io(struct cache *cache, struct bio *bio, - dm_cblock_t cblock) -{ - return bio_data_dir(bio) == WRITE && - cache->features.write_through && !is_dirty(cache, cblock); -} - static void inc_hit_counter(struct cache *cache, struct bio *bio) { atomic_inc(bio_data_dir(bio) == READ ? @@ -1122,6 +1301,15 @@ static void inc_miss_counter(struct cache *cache, struct bio *bio) &cache->stats.read_miss : &cache->stats.write_miss); } +static void issue_cache_bio(struct cache *cache, struct bio *bio, + struct per_bio_data *pb, + dm_oblock_t oblock, dm_cblock_t cblock) +{ + pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds); + remap_to_cache_dirty(cache, bio, oblock, cblock); + issue(cache, bio); +} + static void process_bio(struct cache *cache, struct prealloc *structs, struct bio *bio) { @@ -1133,7 +1321,8 @@ static void process_bio(struct cache *cache, struct prealloc *structs, size_t pb_data_size = get_per_bio_data_size(cache); struct per_bio_data *pb = get_per_bio_data(bio, pb_data_size); bool discarded_block = is_discarded_oblock(cache, block); - bool can_migrate = discarded_block || spare_migration_bandwidth(cache); + bool passthrough = passthrough_mode(&cache->features); + bool can_migrate = !passthrough && (discarded_block || spare_migration_bandwidth(cache)); /* * Check to see if that block is currently migrating. @@ -1154,15 +1343,39 @@ static void process_bio(struct cache *cache, struct prealloc *structs, switch (lookup_result.op) { case POLICY_HIT: - inc_hit_counter(cache, bio); - pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds); + if (passthrough) { + inc_miss_counter(cache, bio); - if (is_writethrough_io(cache, bio, lookup_result.cblock)) - remap_to_origin_then_cache(cache, bio, block, lookup_result.cblock); - else - remap_to_cache_dirty(cache, bio, block, lookup_result.cblock); + /* + * Passthrough always maps to the origin, + * invalidating any cache blocks that are written + * to. + */ + + if (bio_data_dir(bio) == WRITE) { + atomic_inc(&cache->stats.demotion); + invalidate(cache, structs, block, lookup_result.cblock, new_ocell); + release_cell = false; + + } else { + /* FIXME: factor out issue_origin() */ + pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds); + remap_to_origin_clear_discard(cache, bio, block); + issue(cache, bio); + } + } else { + inc_hit_counter(cache, bio); + + if (bio_data_dir(bio) == WRITE && + writethrough_mode(&cache->features) && + !is_dirty(cache, lookup_result.cblock)) { + pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds); + remap_to_origin_then_cache(cache, bio, block, lookup_result.cblock); + issue(cache, bio); + } else + issue_cache_bio(cache, bio, pb, block, lookup_result.cblock); + } - issue(cache, bio); break; case POLICY_MISS: @@ -1221,15 +1434,17 @@ static int need_commit_due_to_time(struct cache *cache) static int commit_if_needed(struct cache *cache) { - if (dm_cache_changed_this_transaction(cache->cmd) && - (cache->commit_requested || need_commit_due_to_time(cache))) { + int r = 0; + + if ((cache->commit_requested || need_commit_due_to_time(cache)) && + dm_cache_changed_this_transaction(cache->cmd)) { atomic_inc(&cache->stats.commit_count); - cache->last_commit_jiffies = jiffies; cache->commit_requested = false; - return dm_cache_commit(cache->cmd, false); + r = dm_cache_commit(cache->cmd, false); + cache->last_commit_jiffies = jiffies; } - return 0; + return r; } static void process_deferred_bios(struct cache *cache) @@ -1338,36 +1553,88 @@ static void writeback_some_dirty_blocks(struct cache *cache) } /*---------------------------------------------------------------- - * Main worker loop + * Invalidations. + * Dropping something from the cache *without* writing back. *--------------------------------------------------------------*/ -static void start_quiescing(struct cache *cache) + +static void process_invalidation_request(struct cache *cache, struct invalidation_request *req) { - unsigned long flags; + int r = 0; + uint64_t begin = from_cblock(req->cblocks->begin); + uint64_t end = from_cblock(req->cblocks->end); - spin_lock_irqsave(&cache->lock, flags); - cache->quiescing = 1; - spin_unlock_irqrestore(&cache->lock, flags); + while (begin != end) { + r = policy_remove_cblock(cache->policy, to_cblock(begin)); + if (!r) { + r = dm_cache_remove_mapping(cache->cmd, to_cblock(begin)); + if (r) + break; + + } else if (r == -ENODATA) { + /* harmless, already unmapped */ + r = 0; + + } else { + DMERR("policy_remove_cblock failed"); + break; + } + + begin++; + } + + cache->commit_requested = true; + + req->err = r; + atomic_set(&req->complete, 1); + + wake_up(&req->result_wait); } -static void stop_quiescing(struct cache *cache) +static void process_invalidation_requests(struct cache *cache) { - unsigned long flags; + struct list_head list; + struct invalidation_request *req, *tmp; - spin_lock_irqsave(&cache->lock, flags); - cache->quiescing = 0; - spin_unlock_irqrestore(&cache->lock, flags); + INIT_LIST_HEAD(&list); + spin_lock(&cache->invalidation_lock); + list_splice_init(&cache->invalidation_requests, &list); + spin_unlock(&cache->invalidation_lock); + + list_for_each_entry_safe (req, tmp, &list, list) + process_invalidation_request(cache, req); } +/*---------------------------------------------------------------- + * Main worker loop + *--------------------------------------------------------------*/ static bool is_quiescing(struct cache *cache) { - int r; - unsigned long flags; + return atomic_read(&cache->quiescing); +} - spin_lock_irqsave(&cache->lock, flags); - r = cache->quiescing; - spin_unlock_irqrestore(&cache->lock, flags); +static void ack_quiescing(struct cache *cache) +{ + if (is_quiescing(cache)) { + atomic_inc(&cache->quiescing_ack); + wake_up(&cache->quiescing_wait); + } +} - return r; +static void wait_for_quiescing_ack(struct cache *cache) +{ + wait_event(cache->quiescing_wait, atomic_read(&cache->quiescing_ack)); +} + +static void start_quiescing(struct cache *cache) +{ + atomic_inc(&cache->quiescing); + wait_for_quiescing_ack(cache); +} + +static void stop_quiescing(struct cache *cache) +{ + atomic_set(&cache->quiescing, 0); + atomic_set(&cache->quiescing_ack, 0); } static void wait_for_migrations(struct cache *cache) @@ -1406,7 +1673,8 @@ static int more_work(struct cache *cache) !bio_list_empty(&cache->deferred_writethrough_bios) || !list_empty(&cache->quiesced_migrations) || !list_empty(&cache->completed_migrations) || - !list_empty(&cache->need_commit_migrations); + !list_empty(&cache->need_commit_migrations) || + cache->invalidate; } static void do_worker(struct work_struct *ws) @@ -1414,16 +1682,16 @@ static void do_worker(struct work_struct *ws) struct cache *cache = container_of(ws, struct cache, worker); do { - if (!is_quiescing(cache)) + if (!is_quiescing(cache)) { + writeback_some_dirty_blocks(cache); + process_deferred_writethrough_bios(cache); process_deferred_bios(cache); + process_invalidation_requests(cache); + } process_migrations(cache, &cache->quiesced_migrations, issue_copy); process_migrations(cache, &cache->completed_migrations, complete_migration); - writeback_some_dirty_blocks(cache); - - process_deferred_writethrough_bios(cache); - if (commit_if_needed(cache)) { process_deferred_flush_bios(cache, false); @@ -1436,6 +1704,9 @@ static void do_worker(struct work_struct *ws) process_migrations(cache, &cache->need_commit_migrations, migration_success_post_commit); } + + ack_quiescing(cache); + } while (more_work(cache)); } @@ -1683,24 +1954,25 @@ static int parse_origin_dev(struct cache_args *ca, struct dm_arg_set *as, static int parse_block_size(struct cache_args *ca, struct dm_arg_set *as, char **error) { - unsigned long tmp; + unsigned long block_size; if (!at_least_one_arg(as, error)) return -EINVAL; - if (kstrtoul(dm_shift_arg(as), 10, &tmp) || !tmp || - tmp < DATA_DEV_BLOCK_SIZE_MIN_SECTORS || - tmp & (DATA_DEV_BLOCK_SIZE_MIN_SECTORS - 1)) { + if (kstrtoul(dm_shift_arg(as), 10, &block_size) || !block_size || + block_size < DATA_DEV_BLOCK_SIZE_MIN_SECTORS || + block_size > DATA_DEV_BLOCK_SIZE_MAX_SECTORS || + block_size & (DATA_DEV_BLOCK_SIZE_MIN_SECTORS - 1)) { *error = "Invalid data block size"; return -EINVAL; } - if (tmp > ca->cache_sectors) { + if (block_size > ca->cache_sectors) { *error = "Data block size is larger than the cache device"; return -EINVAL; } - ca->block_size = tmp; + ca->block_size = block_size; return 0; } @@ -1708,7 +1980,7 @@ static int parse_block_size(struct cache_args *ca, struct dm_arg_set *as, static void init_features(struct cache_features *cf) { cf->mode = CM_WRITE; - cf->write_through = false; + cf->io_mode = CM_IO_WRITEBACK; } static int parse_features(struct cache_args *ca, struct dm_arg_set *as, @@ -1733,10 +2005,13 @@ static int parse_features(struct cache_args *ca, struct dm_arg_set *as, arg = dm_shift_arg(as); if (!strcasecmp(arg, "writeback")) - cf->write_through = false; + cf->io_mode = CM_IO_WRITEBACK; else if (!strcasecmp(arg, "writethrough")) - cf->write_through = true; + cf->io_mode = CM_IO_WRITETHROUGH; + + else if (!strcasecmp(arg, "passthrough")) + cf->io_mode = CM_IO_PASSTHROUGH; else { *error = "Unrecognised cache feature requested"; @@ -1865,47 +2140,19 @@ static int set_config_values(struct cache *cache, int argc, const char **argv) static int create_cache_policy(struct cache *cache, struct cache_args *ca, char **error) { - cache->policy = dm_cache_policy_create(ca->policy_name, - cache->cache_size, - cache->origin_sectors, - cache->sectors_per_block); - if (!cache->policy) { + struct dm_cache_policy *p = dm_cache_policy_create(ca->policy_name, + cache->cache_size, + cache->origin_sectors, + cache->sectors_per_block); + if (IS_ERR(p)) { *error = "Error creating cache's policy"; - return -ENOMEM; + return PTR_ERR(p); } + cache->policy = p; return 0; } -/* - * We want the discard block size to be a power of two, at least the size - * of the cache block size, and have no more than 2^14 discard blocks - * across the origin. - */ -#define MAX_DISCARD_BLOCKS (1 << 14) - -static bool too_many_discard_blocks(sector_t discard_block_size, - sector_t origin_size) -{ - (void) sector_div(origin_size, discard_block_size); - - return origin_size > MAX_DISCARD_BLOCKS; -} - -static sector_t calculate_discard_block_size(sector_t cache_block_size, - sector_t origin_size) -{ - sector_t discard_block_size; - - discard_block_size = roundup_pow_of_two(cache_block_size); - - if (origin_size) - while (too_many_discard_blocks(discard_block_size, origin_size)) - discard_block_size *= 2; - - return discard_block_size; -} - #define DEFAULT_MIGRATION_THRESHOLD 2048 static int cache_create(struct cache_args *ca, struct cache **result) @@ -1930,6 +2177,8 @@ static int cache_create(struct cache_args *ca, struct cache **result) ti->num_discard_bios = 1; ti->discards_supported = true; ti->discard_zeroes_data_unsupported = true; + /* Discard bios must be split on a block boundary */ + ti->split_discard_bios = true; cache->features = ca->features; ti->per_bio_data_size = get_per_bio_data_size(cache); @@ -1988,6 +2237,22 @@ static int cache_create(struct cache_args *ca, struct cache **result) } cache->cmd = cmd; + if (passthrough_mode(&cache->features)) { + bool all_clean; + + r = dm_cache_metadata_all_clean(cache->cmd, &all_clean); + if (r) { + *error = "dm_cache_metadata_all_clean() failed"; + goto bad; + } + + if (!all_clean) { + *error = "Cannot enter passthrough mode unless all blocks are clean"; + r = -EINVAL; + goto bad; + } + } + spin_lock_init(&cache->lock); bio_list_init(&cache->deferred_bios); bio_list_init(&cache->deferred_flush_bios); @@ -1998,8 +2263,12 @@ static int cache_create(struct cache_args *ca, struct cache **result) atomic_set(&cache->nr_migrations, 0); init_waitqueue_head(&cache->migration_wait); + init_waitqueue_head(&cache->quiescing_wait); + atomic_set(&cache->quiescing, 0); + atomic_set(&cache->quiescing_ack, 0); + r = -ENOMEM; - cache->nr_dirty = 0; + atomic_set(&cache->nr_dirty, 0); cache->dirty_bitset = alloc_bitset(from_cblock(cache->cache_size)); if (!cache->dirty_bitset) { *error = "could not allocate dirty bitset"; @@ -2007,16 +2276,13 @@ static int cache_create(struct cache_args *ca, struct cache **result) } clear_bitset(cache->dirty_bitset, from_cblock(cache->cache_size)); - cache->discard_block_size = - calculate_discard_block_size(cache->sectors_per_block, - cache->origin_sectors); - cache->discard_nr_blocks = oblock_to_dblock(cache, cache->origin_blocks); - cache->discard_bitset = alloc_bitset(from_dblock(cache->discard_nr_blocks)); + cache->discard_nr_blocks = cache->origin_blocks; + cache->discard_bitset = alloc_bitset(from_oblock(cache->discard_nr_blocks)); if (!cache->discard_bitset) { *error = "could not allocate discard bitset"; goto bad; } - clear_bitset(cache->discard_bitset, from_dblock(cache->discard_nr_blocks)); + clear_bitset(cache->discard_bitset, from_oblock(cache->discard_nr_blocks)); cache->copier = dm_kcopyd_client_create(&dm_kcopyd_throttle); if (IS_ERR(cache->copier)) { @@ -2057,7 +2323,7 @@ static int cache_create(struct cache_args *ca, struct cache **result) cache->need_tick_bio = true; cache->sized = false; - cache->quiescing = false; + cache->invalidate = false; cache->commit_requested = false; cache->loaded_mappings = false; cache->loaded_discards = false; @@ -2071,6 +2337,9 @@ static int cache_create(struct cache_args *ca, struct cache **result) atomic_set(&cache->stats.commit_count, 0); atomic_set(&cache->stats.discard_count, 0); + spin_lock_init(&cache->invalidation_lock); + INIT_LIST_HEAD(&cache->invalidation_requests); + *result = cache; return 0; @@ -2148,20 +2417,18 @@ static int cache_map(struct dm_target *ti, struct bio *bio) bool discarded_block; struct dm_bio_prison_cell *cell; struct policy_result lookup_result; - struct per_bio_data *pb; + struct per_bio_data *pb = init_per_bio_data(bio, pb_data_size); - if (from_oblock(block) > from_oblock(cache->origin_blocks)) { + if (unlikely(from_oblock(block) >= from_oblock(cache->origin_blocks))) { /* * This can only occur if the io goes to a partial block at * the end of the origin device. We don't cache these. * Just remap to the origin and carry on. */ - remap_to_origin_clear_discard(cache, bio, block); + remap_to_origin(cache, bio); return DM_MAPIO_REMAPPED; } - pb = init_per_bio_data(bio, pb_data_size); - if (bio->bi_rw & (REQ_FLUSH | REQ_FUA | REQ_DISCARD)) { defer_bio(cache, bio); return DM_MAPIO_SUBMITTED; @@ -2200,17 +2467,38 @@ static int cache_map(struct dm_target *ti, struct bio *bio) return DM_MAPIO_SUBMITTED; } + r = DM_MAPIO_REMAPPED; switch (lookup_result.op) { case POLICY_HIT: - inc_hit_counter(cache, bio); - pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds); + if (passthrough_mode(&cache->features)) { + if (bio_data_dir(bio) == WRITE) { + /* + * We need to invalidate this block, so + * defer for the worker thread. + */ + cell_defer(cache, cell, true); + r = DM_MAPIO_SUBMITTED; + + } else { + pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds); + inc_miss_counter(cache, bio); + remap_to_origin_clear_discard(cache, bio, block); + + cell_defer(cache, cell, false); + } - if (is_writethrough_io(cache, bio, lookup_result.cblock)) - remap_to_origin_then_cache(cache, bio, block, lookup_result.cblock); - else - remap_to_cache_dirty(cache, bio, block, lookup_result.cblock); + } else { + inc_hit_counter(cache, bio); + pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds); - cell_defer(cache, cell, false); + if (bio_data_dir(bio) == WRITE && writethrough_mode(&cache->features) && + !is_dirty(cache, lookup_result.cblock)) + remap_to_origin_then_cache(cache, bio, block, lookup_result.cblock); + else + remap_to_cache_dirty(cache, bio, block, lookup_result.cblock); + + cell_defer(cache, cell, false); + } break; case POLICY_MISS: @@ -2235,10 +2523,10 @@ static int cache_map(struct dm_target *ti, struct bio *bio) DMERR_LIMIT("%s: erroring bio: unknown policy op: %u", __func__, (unsigned) lookup_result.op); bio_io_error(bio); - return DM_MAPIO_SUBMITTED; + r = DM_MAPIO_SUBMITTED; } - return DM_MAPIO_REMAPPED; + return r; } static int cache_end_io(struct dm_target *ti, struct bio *bio, int error) @@ -2279,16 +2567,16 @@ static int write_discard_bitset(struct cache *cache) { unsigned i, r; - r = dm_cache_discard_bitset_resize(cache->cmd, cache->discard_block_size, - cache->discard_nr_blocks); + r = dm_cache_discard_bitset_resize(cache->cmd, cache->sectors_per_block, + cache->origin_blocks); if (r) { DMERR("could not resize on-disk discard bitset"); return r; } - for (i = 0; i < from_dblock(cache->discard_nr_blocks); i++) { - r = dm_cache_set_discard(cache->cmd, to_dblock(i), - is_discarded(cache, to_dblock(i))); + for (i = 0; i < from_oblock(cache->discard_nr_blocks); i++) { + r = dm_cache_set_discard(cache->cmd, to_oblock(i), + is_discarded(cache, to_oblock(i))); if (r) return r; } @@ -2296,30 +2584,6 @@ static int write_discard_bitset(struct cache *cache) return 0; } -static int save_hint(void *context, dm_cblock_t cblock, dm_oblock_t oblock, - uint32_t hint) -{ - struct cache *cache = context; - return dm_cache_save_hint(cache->cmd, cblock, hint); -} - -static int write_hints(struct cache *cache) -{ - int r; - - r = dm_cache_begin_hints(cache->cmd, cache->policy); - if (r) { - DMERR("dm_cache_begin_hints failed"); - return r; - } - - r = policy_walk_mappings(cache->policy, save_hint, cache); - if (r) - DMERR("policy_walk_mappings failed"); - - return r; -} - /* * returns true on success */ @@ -2337,7 +2601,7 @@ static bool sync_metadata(struct cache *cache) save_stats(cache); - r3 = write_hints(cache); + r3 = dm_cache_write_hints(cache->cmd, cache->policy); if (r3) DMERR("could not write hints"); @@ -2385,16 +2649,56 @@ static int load_mapping(void *context, dm_oblock_t oblock, dm_cblock_t cblock, } static int load_discard(void *context, sector_t discard_block_size, - dm_dblock_t dblock, bool discard) + dm_oblock_t oblock, bool discard) { struct cache *cache = context; - /* FIXME: handle mis-matched block size */ - if (discard) - set_discard(cache, dblock); + set_discard(cache, oblock); else - clear_discard(cache, dblock); + clear_discard(cache, oblock); + + return 0; +} + +static dm_cblock_t get_cache_dev_size(struct cache *cache) +{ + sector_t size = get_dev_size(cache->cache_dev); + (void) sector_div(size, cache->sectors_per_block); + return to_cblock(size); +} + +static bool can_resize(struct cache *cache, dm_cblock_t new_size) +{ + if (from_cblock(new_size) > from_cblock(cache->cache_size)) + return true; + + /* + * We can't drop a dirty block when shrinking the cache. + */ + while (from_cblock(new_size) < from_cblock(cache->cache_size)) { + new_size = to_cblock(from_cblock(new_size) + 1); + if (is_dirty(cache, new_size)) { + DMERR("unable to shrink cache; cache block %llu is dirty", + (unsigned long long) from_cblock(new_size)); + return false; + } + } + + return true; +} + +static int resize_cache_dev(struct cache *cache, dm_cblock_t new_size) +{ + int r; + + r = dm_cache_resize(cache->cmd, new_size); + if (r) { + DMERR("could not resize cache metadata"); + return r; + } + + cache->cache_size = new_size; return 0; } @@ -2403,22 +2707,25 @@ static int cache_preresume(struct dm_target *ti) { int r = 0; struct cache *cache = ti->private; - sector_t actual_cache_size = get_dev_size(cache->cache_dev); - (void) sector_div(actual_cache_size, cache->sectors_per_block); + dm_cblock_t csize = get_cache_dev_size(cache); /* * Check to see if the cache has resized. */ - if (from_cblock(cache->cache_size) != actual_cache_size || !cache->sized) { - cache->cache_size = to_cblock(actual_cache_size); - - r = dm_cache_resize(cache->cmd, cache->cache_size); - if (r) { - DMERR("could not resize cache metadata"); + if (!cache->sized) { + r = resize_cache_dev(cache, csize); + if (r) return r; - } cache->sized = true; + + } else if (csize != cache->cache_size) { + if (!can_resize(cache, csize)) + return -EINVAL; + + r = resize_cache_dev(cache, csize); + if (r) + return r; } if (!cache->loaded_mappings) { @@ -2456,12 +2763,13 @@ static void cache_resume(struct dm_target *ti) /* * Status format: * - * <#used metadata blocks>/<#total metadata blocks> + * <metadata block size> <#used metadata blocks>/<#total metadata blocks> + * <cache block size> <#used cache blocks>/<#total cache blocks> * <#read hits> <#read misses> <#write hits> <#write misses> - * <#demotions> <#promotions> <#blocks in cache> <#dirty> + * <#demotions> <#promotions> <#dirty> * <#features> <features>* * <#core args> <core args> - * <#policy args> <policy args>* + * <policy name> <#policy args> <policy args>* */ static void cache_status(struct dm_target *ti, status_type_t type, unsigned status_flags, char *result, unsigned maxlen) @@ -2499,24 +2807,38 @@ static void cache_status(struct dm_target *ti, status_type_t type, residency = policy_residency(cache->policy); - DMEMIT("%llu/%llu %u %u %u %u %u %u %llu %u ", + DMEMIT("%u %llu/%llu %u %llu/%llu %u %u %u %u %u %u %lu ", + (unsigned)(DM_CACHE_METADATA_BLOCK_SIZE >> SECTOR_SHIFT), (unsigned long long)(nr_blocks_metadata - nr_free_blocks_metadata), (unsigned long long)nr_blocks_metadata, + cache->sectors_per_block, + (unsigned long long) from_cblock(residency), + (unsigned long long) from_cblock(cache->cache_size), (unsigned) atomic_read(&cache->stats.read_hit), (unsigned) atomic_read(&cache->stats.read_miss), (unsigned) atomic_read(&cache->stats.write_hit), (unsigned) atomic_read(&cache->stats.write_miss), (unsigned) atomic_read(&cache->stats.demotion), (unsigned) atomic_read(&cache->stats.promotion), - (unsigned long long) from_cblock(residency), - cache->nr_dirty); + (unsigned long) atomic_read(&cache->nr_dirty)); - if (cache->features.write_through) + if (writethrough_mode(&cache->features)) DMEMIT("1 writethrough "); - else - DMEMIT("0 "); + + else if (passthrough_mode(&cache->features)) + DMEMIT("1 passthrough "); + + else if (writeback_mode(&cache->features)) + DMEMIT("1 writeback "); + + else { + DMERR("internal error: unknown io mode: %d", (int) cache->features.io_mode); + goto err; + } DMEMIT("2 migration_threshold %llu ", (unsigned long long) cache->migration_threshold); + + DMEMIT("%s ", dm_cache_policy_get_name(cache->policy)); if (sz < maxlen) { r = policy_emit_config_values(cache->policy, result + sz, maxlen - sz); if (r) @@ -2546,7 +2868,128 @@ err: } /* - * Supports <key> <value>. + * A cache block range can take two forms: + * + * i) A single cblock, eg. '3456' + * ii) A begin and end cblock with dots between, eg. 123-234 + */ +static int parse_cblock_range(struct cache *cache, const char *str, + struct cblock_range *result) +{ + char dummy; + uint64_t b, e; + int r; + + /* + * Try and parse form (ii) first. + */ + r = sscanf(str, "%llu-%llu%c", &b, &e, &dummy); + if (r < 0) + return r; + + if (r == 2) { + result->begin = to_cblock(b); + result->end = to_cblock(e); + return 0; + } + + /* + * That didn't work, try form (i). + */ + r = sscanf(str, "%llu%c", &b, &dummy); + if (r < 0) + return r; + + if (r == 1) { + result->begin = to_cblock(b); + result->end = to_cblock(from_cblock(result->begin) + 1u); + return 0; + } + + DMERR("invalid cblock range '%s'", str); + return -EINVAL; +} + +static int validate_cblock_range(struct cache *cache, struct cblock_range *range) +{ + uint64_t b = from_cblock(range->begin); + uint64_t e = from_cblock(range->end); + uint64_t n = from_cblock(cache->cache_size); + + if (b >= n) { + DMERR("begin cblock out of range: %llu >= %llu", b, n); + return -EINVAL; + } + + if (e > n) { + DMERR("end cblock out of range: %llu > %llu", e, n); + return -EINVAL; + } + + if (b >= e) { + DMERR("invalid cblock range: %llu >= %llu", b, e); + return -EINVAL; + } + + return 0; +} + +static int request_invalidation(struct cache *cache, struct cblock_range *range) +{ + struct invalidation_request req; + + INIT_LIST_HEAD(&req.list); + req.cblocks = range; + atomic_set(&req.complete, 0); + req.err = 0; + init_waitqueue_head(&req.result_wait); + + spin_lock(&cache->invalidation_lock); + list_add(&req.list, &cache->invalidation_requests); + spin_unlock(&cache->invalidation_lock); + wake_worker(cache); + + wait_event(req.result_wait, atomic_read(&req.complete)); + return req.err; +} + +static int process_invalidate_cblocks_message(struct cache *cache, unsigned count, + const char **cblock_ranges) +{ + int r = 0; + unsigned i; + struct cblock_range range; + + if (!passthrough_mode(&cache->features)) { + DMERR("cache has to be in passthrough mode for invalidation"); + return -EPERM; + } + + for (i = 0; i < count; i++) { + r = parse_cblock_range(cache, cblock_ranges[i], &range); + if (r) + break; + + r = validate_cblock_range(cache, &range); + if (r) + break; + + /* + * Pass begin and end origin blocks to the worker and wake it. + */ + r = request_invalidation(cache, &range); + if (r) + break; + } + + return r; +} + +/* + * Supports + * "<key> <value>" + * and + * "invalidate_cblocks [(<begin>)|(<begin>-<end>)]* * * The key migration_threshold is supported by the cache target core. */ @@ -2554,6 +2997,12 @@ static int cache_message(struct dm_target *ti, unsigned argc, char **argv) { struct cache *cache = ti->private; + if (!argc) + return -EINVAL; + + if (!strcasecmp(argv[0], "invalidate_cblocks")) + return process_invalidate_cblocks_message(cache, argc - 1, (const char **) argv + 1); + if (argc != 2) return -EINVAL; @@ -2598,16 +3047,24 @@ static void set_discard_limits(struct cache *cache, struct queue_limits *limits) /* * FIXME: these limits may be incompatible with the cache device */ - limits->max_discard_sectors = cache->discard_block_size * 1024; - limits->discard_granularity = cache->discard_block_size << SECTOR_SHIFT; + limits->max_discard_sectors = cache->sectors_per_block; + limits->discard_granularity = cache->sectors_per_block << SECTOR_SHIFT; } static void cache_io_hints(struct dm_target *ti, struct queue_limits *limits) { struct cache *cache = ti->private; + uint64_t io_opt_sectors = limits->io_opt >> SECTOR_SHIFT; - blk_limits_io_min(limits, 0); - blk_limits_io_opt(limits, cache->sectors_per_block << SECTOR_SHIFT); + /* + * If the system-determined stacked limits are compatible with the + * cache's blocksize (io_opt is a factor) do not override them. + */ + if (io_opt_sectors < cache->sectors_per_block || + do_div(io_opt_sectors, cache->sectors_per_block)) { + blk_limits_io_min(limits, 0); + blk_limits_io_opt(limits, cache->sectors_per_block << SECTOR_SHIFT); + } set_discard_limits(cache, limits); } @@ -2615,7 +3072,7 @@ static void cache_io_hints(struct dm_target *ti, struct queue_limits *limits) static struct target_type cache_target = { .name = "cache", - .version = {1, 1, 1}, + .version = {1, 4, 0}, .module = THIS_MODULE, .ctr = cache_ctr, .dtr = cache_dtr, diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c index 6d2d41ae9e3..4cba2d808af 100644 --- a/drivers/md/dm-crypt.c +++ b/drivers/md/dm-crypt.c @@ -1,7 +1,8 @@ /* - * Copyright (C) 2003 Christophe Saout <christophe@saout.de> + * Copyright (C) 2003 Jana Saout <jana@saout.de> * Copyright (C) 2004 Clemens Fruhwirth <clemens@endorphin.org> * Copyright (C) 2006-2009 Red Hat, Inc. All rights reserved. + * Copyright (C) 2013 Milan Broz <gmazyland@gmail.com> * * This file is released under the GPL. */ @@ -18,7 +19,6 @@ #include <linux/crypto.h> #include <linux/workqueue.h> #include <linux/backing-dev.h> -#include <linux/percpu.h> #include <linux/atomic.h> #include <linux/scatterlist.h> #include <asm/page.h> @@ -38,12 +38,11 @@ struct convert_context { struct completion restart; struct bio *bio_in; struct bio *bio_out; - unsigned int offset_in; - unsigned int offset_out; - unsigned int idx_in; - unsigned int idx_out; + struct bvec_iter iter_in; + struct bvec_iter iter_out; sector_t cc_sector; atomic_t cc_pending; + struct ablkcipher_request *req; }; /* @@ -98,6 +97,13 @@ struct iv_lmk_private { u8 *seed; }; +#define TCW_WHITENING_SIZE 16 +struct iv_tcw_private { + struct crypto_shash *crc32_tfm; + u8 *iv_seed; + u8 *whitening; +}; + /* * Crypt: maps a linear range of a block device * and encrypts / decrypts at the same time. @@ -105,15 +111,7 @@ struct iv_lmk_private { enum flags { DM_CRYPT_SUSPENDED, DM_CRYPT_KEY_VALID }; /* - * Duplicated per-CPU state for cipher. - */ -struct crypt_cpu { - struct ablkcipher_request *req; -}; - -/* - * The fields in here must be read only after initialization, - * changing state should be in crypt_cpu. + * The fields in here must be read only after initialization. */ struct crypt_config { struct dm_dev *dev; @@ -139,16 +137,11 @@ struct crypt_config { struct iv_essiv_private essiv; struct iv_benbi_private benbi; struct iv_lmk_private lmk; + struct iv_tcw_private tcw; } iv_gen_private; sector_t iv_offset; unsigned int iv_size; - /* - * Duplicated per cpu state. Access through - * per_cpu_ptr() only. - */ - struct crypt_cpu __percpu *cpu; - /* ESSIV: struct crypto_cipher *essiv_tfm */ void *iv_private; struct crypto_ablkcipher **tfms; @@ -171,7 +164,8 @@ struct crypt_config { unsigned long flags; unsigned int key_size; - unsigned int key_parts; + unsigned int key_parts; /* independent parts in key buffer */ + unsigned int key_extra_size; /* additional keys length */ u8 key[0]; }; @@ -184,11 +178,6 @@ static void clone_init(struct dm_crypt_io *, struct bio *); static void kcryptd_queue_crypt(struct dm_crypt_io *io); static u8 *iv_of_dmreq(struct crypt_config *cc, struct dm_crypt_request *dmreq); -static struct crypt_cpu *this_crypt_config(struct crypt_config *cc) -{ - return this_cpu_ptr(cc->cpu); -} - /* * Use this to access cipher attributes that are the same for each CPU. */ @@ -230,6 +219,16 @@ static struct crypto_ablkcipher *any_tfm(struct crypt_config *cc) * version 3: the same as version 2 with additional IV seed * (it uses 65 keys, last key is used as IV seed) * + * tcw: Compatible implementation of the block chaining mode used + * by the TrueCrypt device encryption system (prior to version 4.1). + * For more info see: http://www.truecrypt.org + * It operates on full 512 byte sectors and uses CBC + * with an IV derived from initial key and the sector number. + * In addition, whitening value is applied on every sector, whitening + * is calculated from initial key, sector number and mixed using CRC32. + * Note that this encryption scheme is vulnerable to watermarking attacks + * and should be used for old compatible containers access only. + * * plumb: unimplemented, see: * http://article.gmane.org/gmane.linux.kernel.device-mapper.dm-crypt/454 */ @@ -530,7 +529,7 @@ static int crypt_iv_lmk_one(struct crypt_config *cc, u8 *iv, char ctx[crypto_shash_descsize(lmk->hash_tfm)]; } sdesc; struct md5_state md5state; - u32 buf[4]; + __le32 buf[4]; int i, r; sdesc.desc.tfm = lmk->hash_tfm; @@ -608,6 +607,153 @@ static int crypt_iv_lmk_post(struct crypt_config *cc, u8 *iv, return r; } +static void crypt_iv_tcw_dtr(struct crypt_config *cc) +{ + struct iv_tcw_private *tcw = &cc->iv_gen_private.tcw; + + kzfree(tcw->iv_seed); + tcw->iv_seed = NULL; + kzfree(tcw->whitening); + tcw->whitening = NULL; + + if (tcw->crc32_tfm && !IS_ERR(tcw->crc32_tfm)) + crypto_free_shash(tcw->crc32_tfm); + tcw->crc32_tfm = NULL; +} + +static int crypt_iv_tcw_ctr(struct crypt_config *cc, struct dm_target *ti, + const char *opts) +{ + struct iv_tcw_private *tcw = &cc->iv_gen_private.tcw; + + if (cc->key_size <= (cc->iv_size + TCW_WHITENING_SIZE)) { + ti->error = "Wrong key size for TCW"; + return -EINVAL; + } + + tcw->crc32_tfm = crypto_alloc_shash("crc32", 0, 0); + if (IS_ERR(tcw->crc32_tfm)) { + ti->error = "Error initializing CRC32 in TCW"; + return PTR_ERR(tcw->crc32_tfm); + } + + tcw->iv_seed = kzalloc(cc->iv_size, GFP_KERNEL); + tcw->whitening = kzalloc(TCW_WHITENING_SIZE, GFP_KERNEL); + if (!tcw->iv_seed || !tcw->whitening) { + crypt_iv_tcw_dtr(cc); + ti->error = "Error allocating seed storage in TCW"; + return -ENOMEM; + } + + return 0; +} + +static int crypt_iv_tcw_init(struct crypt_config *cc) +{ + struct iv_tcw_private *tcw = &cc->iv_gen_private.tcw; + int key_offset = cc->key_size - cc->iv_size - TCW_WHITENING_SIZE; + + memcpy(tcw->iv_seed, &cc->key[key_offset], cc->iv_size); + memcpy(tcw->whitening, &cc->key[key_offset + cc->iv_size], + TCW_WHITENING_SIZE); + + return 0; +} + +static int crypt_iv_tcw_wipe(struct crypt_config *cc) +{ + struct iv_tcw_private *tcw = &cc->iv_gen_private.tcw; + + memset(tcw->iv_seed, 0, cc->iv_size); + memset(tcw->whitening, 0, TCW_WHITENING_SIZE); + + return 0; +} + +static int crypt_iv_tcw_whitening(struct crypt_config *cc, + struct dm_crypt_request *dmreq, + u8 *data) +{ + struct iv_tcw_private *tcw = &cc->iv_gen_private.tcw; + u64 sector = cpu_to_le64((u64)dmreq->iv_sector); + u8 buf[TCW_WHITENING_SIZE]; + struct { + struct shash_desc desc; + char ctx[crypto_shash_descsize(tcw->crc32_tfm)]; + } sdesc; + int i, r; + + /* xor whitening with sector number */ + memcpy(buf, tcw->whitening, TCW_WHITENING_SIZE); + crypto_xor(buf, (u8 *)§or, 8); + crypto_xor(&buf[8], (u8 *)§or, 8); + + /* calculate crc32 for every 32bit part and xor it */ + sdesc.desc.tfm = tcw->crc32_tfm; + sdesc.desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP; + for (i = 0; i < 4; i++) { + r = crypto_shash_init(&sdesc.desc); + if (r) + goto out; + r = crypto_shash_update(&sdesc.desc, &buf[i * 4], 4); + if (r) + goto out; + r = crypto_shash_final(&sdesc.desc, &buf[i * 4]); + if (r) + goto out; + } + crypto_xor(&buf[0], &buf[12], 4); + crypto_xor(&buf[4], &buf[8], 4); + + /* apply whitening (8 bytes) to whole sector */ + for (i = 0; i < ((1 << SECTOR_SHIFT) / 8); i++) + crypto_xor(data + i * 8, buf, 8); +out: + memset(buf, 0, sizeof(buf)); + return r; +} + +static int crypt_iv_tcw_gen(struct crypt_config *cc, u8 *iv, + struct dm_crypt_request *dmreq) +{ + struct iv_tcw_private *tcw = &cc->iv_gen_private.tcw; + u64 sector = cpu_to_le64((u64)dmreq->iv_sector); + u8 *src; + int r = 0; + + /* Remove whitening from ciphertext */ + if (bio_data_dir(dmreq->ctx->bio_in) != WRITE) { + src = kmap_atomic(sg_page(&dmreq->sg_in)); + r = crypt_iv_tcw_whitening(cc, dmreq, src + dmreq->sg_in.offset); + kunmap_atomic(src); + } + + /* Calculate IV */ + memcpy(iv, tcw->iv_seed, cc->iv_size); + crypto_xor(iv, (u8 *)§or, 8); + if (cc->iv_size > 8) + crypto_xor(&iv[8], (u8 *)§or, cc->iv_size - 8); + + return r; +} + +static int crypt_iv_tcw_post(struct crypt_config *cc, u8 *iv, + struct dm_crypt_request *dmreq) +{ + u8 *dst; + int r; + + if (bio_data_dir(dmreq->ctx->bio_in) != WRITE) + return 0; + + /* Apply whitening on ciphertext */ + dst = kmap_atomic(sg_page(&dmreq->sg_out)); + r = crypt_iv_tcw_whitening(cc, dmreq, dst + dmreq->sg_out.offset); + kunmap_atomic(dst); + + return r; +} + static struct crypt_iv_operations crypt_iv_plain_ops = { .generator = crypt_iv_plain_gen }; @@ -643,6 +789,15 @@ static struct crypt_iv_operations crypt_iv_lmk_ops = { .post = crypt_iv_lmk_post }; +static struct crypt_iv_operations crypt_iv_tcw_ops = { + .ctr = crypt_iv_tcw_ctr, + .dtr = crypt_iv_tcw_dtr, + .init = crypt_iv_tcw_init, + .wipe = crypt_iv_tcw_wipe, + .generator = crypt_iv_tcw_gen, + .post = crypt_iv_tcw_post +}; + static void crypt_convert_init(struct crypt_config *cc, struct convert_context *ctx, struct bio *bio_out, struct bio *bio_in, @@ -650,10 +805,10 @@ static void crypt_convert_init(struct crypt_config *cc, { ctx->bio_in = bio_in; ctx->bio_out = bio_out; - ctx->offset_in = 0; - ctx->offset_out = 0; - ctx->idx_in = bio_in ? bio_in->bi_idx : 0; - ctx->idx_out = bio_out ? bio_out->bi_idx : 0; + if (bio_in) + ctx->iter_in = bio_in->bi_iter; + if (bio_out) + ctx->iter_out = bio_out->bi_iter; ctx->cc_sector = sector + cc->iv_offset; init_completion(&ctx->restart); } @@ -681,8 +836,8 @@ static int crypt_convert_block(struct crypt_config *cc, struct convert_context *ctx, struct ablkcipher_request *req) { - struct bio_vec *bv_in = bio_iovec_idx(ctx->bio_in, ctx->idx_in); - struct bio_vec *bv_out = bio_iovec_idx(ctx->bio_out, ctx->idx_out); + struct bio_vec bv_in = bio_iter_iovec(ctx->bio_in, ctx->iter_in); + struct bio_vec bv_out = bio_iter_iovec(ctx->bio_out, ctx->iter_out); struct dm_crypt_request *dmreq; u8 *iv; int r; @@ -693,24 +848,15 @@ static int crypt_convert_block(struct crypt_config *cc, dmreq->iv_sector = ctx->cc_sector; dmreq->ctx = ctx; sg_init_table(&dmreq->sg_in, 1); - sg_set_page(&dmreq->sg_in, bv_in->bv_page, 1 << SECTOR_SHIFT, - bv_in->bv_offset + ctx->offset_in); + sg_set_page(&dmreq->sg_in, bv_in.bv_page, 1 << SECTOR_SHIFT, + bv_in.bv_offset); sg_init_table(&dmreq->sg_out, 1); - sg_set_page(&dmreq->sg_out, bv_out->bv_page, 1 << SECTOR_SHIFT, - bv_out->bv_offset + ctx->offset_out); - - ctx->offset_in += 1 << SECTOR_SHIFT; - if (ctx->offset_in >= bv_in->bv_len) { - ctx->offset_in = 0; - ctx->idx_in++; - } + sg_set_page(&dmreq->sg_out, bv_out.bv_page, 1 << SECTOR_SHIFT, + bv_out.bv_offset); - ctx->offset_out += 1 << SECTOR_SHIFT; - if (ctx->offset_out >= bv_out->bv_len) { - ctx->offset_out = 0; - ctx->idx_out++; - } + bio_advance_iter(ctx->bio_in, &ctx->iter_in, 1 << SECTOR_SHIFT); + bio_advance_iter(ctx->bio_out, &ctx->iter_out, 1 << SECTOR_SHIFT); if (cc->iv_gen_ops) { r = cc->iv_gen_ops->generator(cc, iv, dmreq); @@ -738,16 +884,15 @@ static void kcryptd_async_done(struct crypto_async_request *async_req, static void crypt_alloc_req(struct crypt_config *cc, struct convert_context *ctx) { - struct crypt_cpu *this_cc = this_crypt_config(cc); unsigned key_index = ctx->cc_sector & (cc->tfms_count - 1); - if (!this_cc->req) - this_cc->req = mempool_alloc(cc->req_pool, GFP_NOIO); + if (!ctx->req) + ctx->req = mempool_alloc(cc->req_pool, GFP_NOIO); - ablkcipher_request_set_tfm(this_cc->req, cc->tfms[key_index]); - ablkcipher_request_set_callback(this_cc->req, + ablkcipher_request_set_tfm(ctx->req, cc->tfms[key_index]); + ablkcipher_request_set_callback(ctx->req, CRYPTO_TFM_REQ_MAY_BACKLOG | CRYPTO_TFM_REQ_MAY_SLEEP, - kcryptd_async_done, dmreq_of_req(cc, this_cc->req)); + kcryptd_async_done, dmreq_of_req(cc, ctx->req)); } /* @@ -756,28 +901,26 @@ static void crypt_alloc_req(struct crypt_config *cc, static int crypt_convert(struct crypt_config *cc, struct convert_context *ctx) { - struct crypt_cpu *this_cc = this_crypt_config(cc); int r; atomic_set(&ctx->cc_pending, 1); - while(ctx->idx_in < ctx->bio_in->bi_vcnt && - ctx->idx_out < ctx->bio_out->bi_vcnt) { + while (ctx->iter_in.bi_size && ctx->iter_out.bi_size) { crypt_alloc_req(cc, ctx); atomic_inc(&ctx->cc_pending); - r = crypt_convert_block(cc, ctx, this_cc->req); + r = crypt_convert_block(cc, ctx, ctx->req); switch (r) { /* async */ case -EBUSY: wait_for_completion(&ctx->restart); - INIT_COMPLETION(ctx->restart); + reinit_completion(&ctx->restart); /* fall through*/ case -EINPROGRESS: - this_cc->req = NULL; + ctx->req = NULL; ctx->cc_sector++; continue; @@ -845,7 +988,7 @@ static struct bio *crypt_alloc_buffer(struct dm_crypt_io *io, unsigned size, size -= len; } - if (!clone->bi_size) { + if (!clone->bi_iter.bi_size) { bio_put(clone); return NULL; } @@ -876,6 +1019,7 @@ static struct dm_crypt_io *crypt_io_alloc(struct crypt_config *cc, io->sector = sector; io->error = 0; io->base_io = NULL; + io->ctx.req = NULL; atomic_set(&io->io_pending, 0); return io; @@ -901,6 +1045,8 @@ static void crypt_dec_pending(struct dm_crypt_io *io) if (!atomic_dec_and_test(&io->io_pending)) return; + if (io->ctx.req) + mempool_free(io->ctx.req, cc->req_pool); mempool_free(io, cc->io_pool); if (likely(!base_io)) @@ -985,7 +1131,7 @@ static int kcryptd_io_read(struct dm_crypt_io *io, gfp_t gfp) crypt_inc_pending(io); clone_init(io, clone); - clone->bi_sector = cc->start + io->sector; + clone->bi_iter.bi_sector = cc->start + io->sector; generic_make_request(clone); return 0; @@ -1031,9 +1177,9 @@ static void kcryptd_crypt_write_io_submit(struct dm_crypt_io *io, int async) } /* crypt_convert should have filled the clone bio */ - BUG_ON(io->ctx.idx_out < clone->bi_vcnt); + BUG_ON(io->ctx.iter_out.bi_size); - clone->bi_sector = cc->start + io->sector; + clone->bi_iter.bi_sector = cc->start + io->sector; if (async) kcryptd_queue_io(io); @@ -1048,7 +1194,7 @@ static void kcryptd_crypt_write_convert(struct dm_crypt_io *io) struct dm_crypt_io *new_io; int crypt_finished; unsigned out_of_pages = 0; - unsigned remaining = io->base_bio->bi_size; + unsigned remaining = io->base_bio->bi_iter.bi_size; sector_t sector = io->sector; int r; @@ -1070,9 +1216,9 @@ static void kcryptd_crypt_write_convert(struct dm_crypt_io *io) } io->ctx.bio_out = clone; - io->ctx.idx_out = 0; + io->ctx.iter_out = clone->bi_iter; - remaining -= clone->bi_size; + remaining -= clone->bi_iter.bi_size; sector += bio_sectors(clone); crypt_inc_pending(io); @@ -1114,8 +1260,7 @@ static void kcryptd_crypt_write_convert(struct dm_crypt_io *io) crypt_inc_pending(new_io); crypt_convert_init(cc, &new_io->ctx, NULL, io->base_bio, sector); - new_io->ctx.idx_in = io->ctx.idx_in; - new_io->ctx.offset_in = io->ctx.offset_in; + new_io->ctx.iter_in = io->ctx.iter_in; /* * Fragments after the first use the base_io @@ -1274,9 +1419,12 @@ static int crypt_alloc_tfms(struct crypt_config *cc, char *ciphermode) static int crypt_setkey_allcpus(struct crypt_config *cc) { - unsigned subkey_size = cc->key_size >> ilog2(cc->tfms_count); + unsigned subkey_size; int err = 0, i, r; + /* Ignore extra keys (which are used for IV etc) */ + subkey_size = (cc->key_size - cc->key_extra_size) >> ilog2(cc->tfms_count); + for (i = 0; i < cc->tfms_count; i++) { r = crypto_ablkcipher_setkey(cc->tfms[i], cc->key + (i * subkey_size), @@ -1326,8 +1474,6 @@ static int crypt_wipe_key(struct crypt_config *cc) static void crypt_dtr(struct dm_target *ti) { struct crypt_config *cc = ti->private; - struct crypt_cpu *cpu_cc; - int cpu; ti->private = NULL; @@ -1339,13 +1485,6 @@ static void crypt_dtr(struct dm_target *ti) if (cc->crypt_queue) destroy_workqueue(cc->crypt_queue); - if (cc->cpu) - for_each_possible_cpu(cpu) { - cpu_cc = per_cpu_ptr(cc->cpu, cpu); - if (cpu_cc->req) - mempool_free(cpu_cc->req, cc->req_pool); - } - crypt_free_tfms(cc); if (cc->bs) @@ -1364,9 +1503,6 @@ static void crypt_dtr(struct dm_target *ti) if (cc->dev) dm_put_device(ti, cc->dev); - if (cc->cpu) - free_percpu(cc->cpu); - kzfree(cc->cipher); kzfree(cc->cipher_string); @@ -1409,6 +1545,7 @@ static int crypt_ctr_cipher(struct dm_target *ti, return -EINVAL; } cc->key_parts = cc->tfms_count; + cc->key_extra_size = 0; cc->cipher = kstrdup(cipher, GFP_KERNEL); if (!cc->cipher) @@ -1421,13 +1558,6 @@ static int crypt_ctr_cipher(struct dm_target *ti, if (tmp) DMWARN("Ignoring unexpected additional cipher options"); - cc->cpu = __alloc_percpu(sizeof(*(cc->cpu)), - __alignof__(struct crypt_cpu)); - if (!cc->cpu) { - ti->error = "Cannot allocate per cpu state"; - goto bad_mem; - } - /* * For compatibility with the original dm-crypt mapping format, if * only the cipher name is supplied, use cbc-plain. @@ -1460,13 +1590,6 @@ static int crypt_ctr_cipher(struct dm_target *ti, goto bad; } - /* Initialize and set key */ - ret = crypt_set_key(cc, key); - if (ret < 0) { - ti->error = "Error decoding and setting key"; - goto bad; - } - /* Initialize IV */ cc->iv_size = crypto_ablkcipher_ivsize(any_tfm(cc)); if (cc->iv_size) @@ -1493,18 +1616,33 @@ static int crypt_ctr_cipher(struct dm_target *ti, cc->iv_gen_ops = &crypt_iv_null_ops; else if (strcmp(ivmode, "lmk") == 0) { cc->iv_gen_ops = &crypt_iv_lmk_ops; - /* Version 2 and 3 is recognised according + /* + * Version 2 and 3 is recognised according * to length of provided multi-key string. * If present (version 3), last key is used as IV seed. + * All keys (including IV seed) are always the same size. */ - if (cc->key_size % cc->key_parts) + if (cc->key_size % cc->key_parts) { cc->key_parts++; + cc->key_extra_size = cc->key_size / cc->key_parts; + } + } else if (strcmp(ivmode, "tcw") == 0) { + cc->iv_gen_ops = &crypt_iv_tcw_ops; + cc->key_parts += 2; /* IV + whitening */ + cc->key_extra_size = cc->iv_size + TCW_WHITENING_SIZE; } else { ret = -EINVAL; ti->error = "Invalid IV mode"; goto bad; } + /* Initialize and set key */ + ret = crypt_set_key(cc, key); + if (ret < 0) { + ti->error = "Error decoding and setting key"; + goto bad; + } + /* Allocate IV */ if (cc->iv_gen_ops && cc->iv_gen_ops->ctr) { ret = cc->iv_gen_ops->ctr(cc, ti, ivopts); @@ -1645,20 +1783,14 @@ static int crypt_ctr(struct dm_target *ti, unsigned int argc, char **argv) } ret = -ENOMEM; - cc->io_queue = alloc_workqueue("kcryptd_io", - WQ_NON_REENTRANT| - WQ_MEM_RECLAIM, - 1); + cc->io_queue = alloc_workqueue("kcryptd_io", WQ_MEM_RECLAIM, 1); if (!cc->io_queue) { ti->error = "Couldn't create kcryptd io queue"; goto bad; } cc->crypt_queue = alloc_workqueue("kcryptd", - WQ_NON_REENTRANT| - WQ_CPU_INTENSIVE| - WQ_MEM_RECLAIM, - 1); + WQ_CPU_INTENSIVE | WQ_MEM_RECLAIM, 1); if (!cc->crypt_queue) { ti->error = "Couldn't create kcryptd queue"; goto bad; @@ -1687,11 +1819,12 @@ static int crypt_map(struct dm_target *ti, struct bio *bio) if (unlikely(bio->bi_rw & (REQ_FLUSH | REQ_DISCARD))) { bio->bi_bdev = cc->dev->bdev; if (bio_sectors(bio)) - bio->bi_sector = cc->start + dm_target_offset(ti, bio->bi_sector); + bio->bi_iter.bi_sector = cc->start + + dm_target_offset(ti, bio->bi_iter.bi_sector); return DM_MAPIO_REMAPPED; } - io = crypt_io_alloc(cc, bio, dm_target_offset(ti, bio->bi_sector)); + io = crypt_io_alloc(cc, bio, dm_target_offset(ti, bio->bi_iter.bi_sector)); if (bio_data_dir(io->base_bio) == READ) { if (kcryptd_io_read(io, GFP_NOWAIT)) @@ -1823,7 +1956,7 @@ static int crypt_iterate_devices(struct dm_target *ti, static struct target_type crypt_target = { .name = "crypt", - .version = {1, 12, 1}, + .version = {1, 13, 0}, .module = THIS_MODULE, .ctr = crypt_ctr, .dtr = crypt_dtr, @@ -1863,6 +1996,6 @@ static void __exit dm_crypt_exit(void) module_init(dm_crypt_init); module_exit(dm_crypt_exit); -MODULE_AUTHOR("Christophe Saout <christophe@saout.de>"); +MODULE_AUTHOR("Jana Saout <jana@saout.de>"); MODULE_DESCRIPTION(DM_NAME " target for transparent encryption / decryption"); MODULE_LICENSE("GPL"); diff --git a/drivers/md/dm-delay.c b/drivers/md/dm-delay.c index 496d5f3646a..42c3a27a14c 100644 --- a/drivers/md/dm-delay.c +++ b/drivers/md/dm-delay.c @@ -20,10 +20,10 @@ struct delay_c { struct timer_list delay_timer; struct mutex timer_lock; + struct workqueue_struct *kdelayd_wq; struct work_struct flush_expired_bios; struct list_head delayed_bios; atomic_t may_delay; - mempool_t *delayed_pool; struct dm_dev *dev_read; sector_t start_read; @@ -39,20 +39,16 @@ struct delay_c { struct dm_delay_info { struct delay_c *context; struct list_head list; - struct bio *bio; unsigned long expires; }; static DEFINE_MUTEX(delayed_bios_lock); -static struct workqueue_struct *kdelayd_wq; -static struct kmem_cache *delayed_cache; - static void handle_delayed_timer(unsigned long data) { struct delay_c *dc = (struct delay_c *)data; - queue_work(kdelayd_wq, &dc->flush_expired_bios); + queue_work(dc->kdelayd_wq, &dc->flush_expired_bios); } static void queue_timeout(struct delay_c *dc, unsigned long expires) @@ -87,13 +83,14 @@ static struct bio *flush_delayed_bios(struct delay_c *dc, int flush_all) mutex_lock(&delayed_bios_lock); list_for_each_entry_safe(delayed, next, &dc->delayed_bios, list) { if (flush_all || time_after_eq(jiffies, delayed->expires)) { + struct bio *bio = dm_bio_from_per_bio_data(delayed, + sizeof(struct dm_delay_info)); list_del(&delayed->list); - bio_list_add(&flush_bios, delayed->bio); - if ((bio_data_dir(delayed->bio) == WRITE)) + bio_list_add(&flush_bios, bio); + if ((bio_data_dir(bio) == WRITE)) delayed->context->writes--; else delayed->context->reads--; - mempool_free(delayed, dc->delayed_pool); continue; } @@ -185,10 +182,10 @@ static int delay_ctr(struct dm_target *ti, unsigned int argc, char **argv) } out: - dc->delayed_pool = mempool_create_slab_pool(128, delayed_cache); - if (!dc->delayed_pool) { - DMERR("Couldn't create delayed bio pool."); - goto bad_dev_write; + dc->kdelayd_wq = alloc_workqueue("kdelayd", WQ_MEM_RECLAIM, 0); + if (!dc->kdelayd_wq) { + DMERR("Couldn't start kdelayd"); + goto bad_queue; } setup_timer(&dc->delay_timer, handle_delayed_timer, (unsigned long)dc); @@ -200,10 +197,11 @@ out: ti->num_flush_bios = 1; ti->num_discard_bios = 1; + ti->per_bio_data_size = sizeof(struct dm_delay_info); ti->private = dc; return 0; -bad_dev_write: +bad_queue: if (dc->dev_write) dm_put_device(ti, dc->dev_write); bad_dev_read: @@ -217,14 +215,13 @@ static void delay_dtr(struct dm_target *ti) { struct delay_c *dc = ti->private; - flush_workqueue(kdelayd_wq); + destroy_workqueue(dc->kdelayd_wq); dm_put_device(ti, dc->dev_read); if (dc->dev_write) dm_put_device(ti, dc->dev_write); - mempool_destroy(dc->delayed_pool); kfree(dc); } @@ -236,10 +233,9 @@ static int delay_bio(struct delay_c *dc, int delay, struct bio *bio) if (!delay || !atomic_read(&dc->may_delay)) return 1; - delayed = mempool_alloc(dc->delayed_pool, GFP_NOIO); + delayed = dm_per_bio_data(bio, sizeof(struct dm_delay_info)); delayed->context = dc; - delayed->bio = bio; delayed->expires = expires = jiffies + (delay * HZ / 1000); mutex_lock(&delayed_bios_lock); @@ -281,14 +277,15 @@ static int delay_map(struct dm_target *ti, struct bio *bio) if ((bio_data_dir(bio) == WRITE) && (dc->dev_write)) { bio->bi_bdev = dc->dev_write->bdev; if (bio_sectors(bio)) - bio->bi_sector = dc->start_write + - dm_target_offset(ti, bio->bi_sector); + bio->bi_iter.bi_sector = dc->start_write + + dm_target_offset(ti, bio->bi_iter.bi_sector); return delay_bio(dc, dc->write_delay, bio); } bio->bi_bdev = dc->dev_read->bdev; - bio->bi_sector = dc->start_read + dm_target_offset(ti, bio->bi_sector); + bio->bi_iter.bi_sector = dc->start_read + + dm_target_offset(ti, bio->bi_iter.bi_sector); return delay_bio(dc, dc->read_delay, bio); } @@ -348,19 +345,7 @@ static struct target_type delay_target = { static int __init dm_delay_init(void) { - int r = -ENOMEM; - - kdelayd_wq = alloc_workqueue("kdelayd", WQ_MEM_RECLAIM, 0); - if (!kdelayd_wq) { - DMERR("Couldn't start kdelayd"); - goto bad_queue; - } - - delayed_cache = KMEM_CACHE(dm_delay_info, 0); - if (!delayed_cache) { - DMERR("Couldn't create delayed bio cache."); - goto bad_memcache; - } + int r; r = dm_register_target(&delay_target); if (r < 0) { @@ -371,18 +356,12 @@ static int __init dm_delay_init(void) return 0; bad_register: - kmem_cache_destroy(delayed_cache); -bad_memcache: - destroy_workqueue(kdelayd_wq); -bad_queue: return r; } static void __exit dm_delay_exit(void) { dm_unregister_target(&delay_target); - kmem_cache_destroy(delayed_cache); - destroy_workqueue(kdelayd_wq); } /* Module hooks */ diff --git a/drivers/md/dm-era-target.c b/drivers/md/dm-era-target.c new file mode 100644 index 00000000000..ad913cd4ade --- /dev/null +++ b/drivers/md/dm-era-target.c @@ -0,0 +1,1747 @@ +#include "dm.h" +#include "persistent-data/dm-transaction-manager.h" +#include "persistent-data/dm-bitset.h" +#include "persistent-data/dm-space-map.h" + +#include <linux/dm-io.h> +#include <linux/dm-kcopyd.h> +#include <linux/init.h> +#include <linux/mempool.h> +#include <linux/module.h> +#include <linux/slab.h> +#include <linux/vmalloc.h> + +#define DM_MSG_PREFIX "era" + +#define SUPERBLOCK_LOCATION 0 +#define SUPERBLOCK_MAGIC 2126579579 +#define SUPERBLOCK_CSUM_XOR 146538381 +#define MIN_ERA_VERSION 1 +#define MAX_ERA_VERSION 1 +#define INVALID_WRITESET_ROOT SUPERBLOCK_LOCATION +#define MIN_BLOCK_SIZE 8 + +/*---------------------------------------------------------------- + * Writeset + *--------------------------------------------------------------*/ +struct writeset_metadata { + uint32_t nr_bits; + dm_block_t root; +}; + +struct writeset { + struct writeset_metadata md; + + /* + * An in core copy of the bits to save constantly doing look ups on + * disk. + */ + unsigned long *bits; +}; + +/* + * This does not free off the on disk bitset as this will normally be done + * after digesting into the era array. + */ +static void writeset_free(struct writeset *ws) +{ + vfree(ws->bits); +} + +static int setup_on_disk_bitset(struct dm_disk_bitset *info, + unsigned nr_bits, dm_block_t *root) +{ + int r; + + r = dm_bitset_empty(info, root); + if (r) + return r; + + return dm_bitset_resize(info, *root, 0, nr_bits, false, root); +} + +static size_t bitset_size(unsigned nr_bits) +{ + return sizeof(unsigned long) * dm_div_up(nr_bits, BITS_PER_LONG); +} + +/* + * Allocates memory for the in core bitset. + */ +static int writeset_alloc(struct writeset *ws, dm_block_t nr_blocks) +{ + ws->md.nr_bits = nr_blocks; + ws->md.root = INVALID_WRITESET_ROOT; + ws->bits = vzalloc(bitset_size(nr_blocks)); + if (!ws->bits) { + DMERR("%s: couldn't allocate in memory bitset", __func__); + return -ENOMEM; + } + + return 0; +} + +/* + * Wipes the in-core bitset, and creates a new on disk bitset. + */ +static int writeset_init(struct dm_disk_bitset *info, struct writeset *ws) +{ + int r; + + memset(ws->bits, 0, bitset_size(ws->md.nr_bits)); + + r = setup_on_disk_bitset(info, ws->md.nr_bits, &ws->md.root); + if (r) { + DMERR("%s: setup_on_disk_bitset failed", __func__); + return r; + } + + return 0; +} + +static bool writeset_marked(struct writeset *ws, dm_block_t block) +{ + return test_bit(block, ws->bits); +} + +static int writeset_marked_on_disk(struct dm_disk_bitset *info, + struct writeset_metadata *m, dm_block_t block, + bool *result) +{ + dm_block_t old = m->root; + + /* + * The bitset was flushed when it was archived, so we know there'll + * be no change to the root. + */ + int r = dm_bitset_test_bit(info, m->root, block, &m->root, result); + if (r) { + DMERR("%s: dm_bitset_test_bit failed", __func__); + return r; + } + + BUG_ON(m->root != old); + + return r; +} + +/* + * Returns < 0 on error, 0 if the bit wasn't previously set, 1 if it was. + */ +static int writeset_test_and_set(struct dm_disk_bitset *info, + struct writeset *ws, uint32_t block) +{ + int r; + + if (!test_and_set_bit(block, ws->bits)) { + r = dm_bitset_set_bit(info, ws->md.root, block, &ws->md.root); + if (r) { + /* FIXME: fail mode */ + return r; + } + + return 0; + } + + return 1; +} + +/*---------------------------------------------------------------- + * On disk metadata layout + *--------------------------------------------------------------*/ +#define SPACE_MAP_ROOT_SIZE 128 +#define UUID_LEN 16 + +struct writeset_disk { + __le32 nr_bits; + __le64 root; +} __packed; + +struct superblock_disk { + __le32 csum; + __le32 flags; + __le64 blocknr; + + __u8 uuid[UUID_LEN]; + __le64 magic; + __le32 version; + + __u8 metadata_space_map_root[SPACE_MAP_ROOT_SIZE]; + + __le32 data_block_size; + __le32 metadata_block_size; + __le32 nr_blocks; + + __le32 current_era; + struct writeset_disk current_writeset; + + /* + * Only these two fields are valid within the metadata snapshot. + */ + __le64 writeset_tree_root; + __le64 era_array_root; + + __le64 metadata_snap; +} __packed; + +/*---------------------------------------------------------------- + * Superblock validation + *--------------------------------------------------------------*/ +static void sb_prepare_for_write(struct dm_block_validator *v, + struct dm_block *b, + size_t sb_block_size) +{ + struct superblock_disk *disk = dm_block_data(b); + + disk->blocknr = cpu_to_le64(dm_block_location(b)); + disk->csum = cpu_to_le32(dm_bm_checksum(&disk->flags, + sb_block_size - sizeof(__le32), + SUPERBLOCK_CSUM_XOR)); +} + +static int check_metadata_version(struct superblock_disk *disk) +{ + uint32_t metadata_version = le32_to_cpu(disk->version); + if (metadata_version < MIN_ERA_VERSION || metadata_version > MAX_ERA_VERSION) { + DMERR("Era metadata version %u found, but only versions between %u and %u supported.", + metadata_version, MIN_ERA_VERSION, MAX_ERA_VERSION); + return -EINVAL; + } + + return 0; +} + +static int sb_check(struct dm_block_validator *v, + struct dm_block *b, + size_t sb_block_size) +{ + struct superblock_disk *disk = dm_block_data(b); + __le32 csum_le; + + if (dm_block_location(b) != le64_to_cpu(disk->blocknr)) { + DMERR("sb_check failed: blocknr %llu: wanted %llu", + le64_to_cpu(disk->blocknr), + (unsigned long long)dm_block_location(b)); + return -ENOTBLK; + } + + if (le64_to_cpu(disk->magic) != SUPERBLOCK_MAGIC) { + DMERR("sb_check failed: magic %llu: wanted %llu", + le64_to_cpu(disk->magic), + (unsigned long long) SUPERBLOCK_MAGIC); + return -EILSEQ; + } + + csum_le = cpu_to_le32(dm_bm_checksum(&disk->flags, + sb_block_size - sizeof(__le32), + SUPERBLOCK_CSUM_XOR)); + if (csum_le != disk->csum) { + DMERR("sb_check failed: csum %u: wanted %u", + le32_to_cpu(csum_le), le32_to_cpu(disk->csum)); + return -EILSEQ; + } + + return check_metadata_version(disk); +} + +static struct dm_block_validator sb_validator = { + .name = "superblock", + .prepare_for_write = sb_prepare_for_write, + .check = sb_check +}; + +/*---------------------------------------------------------------- + * Low level metadata handling + *--------------------------------------------------------------*/ +#define DM_ERA_METADATA_BLOCK_SIZE 4096 +#define DM_ERA_METADATA_CACHE_SIZE 64 +#define ERA_MAX_CONCURRENT_LOCKS 5 + +struct era_metadata { + struct block_device *bdev; + struct dm_block_manager *bm; + struct dm_space_map *sm; + struct dm_transaction_manager *tm; + + dm_block_t block_size; + uint32_t nr_blocks; + + uint32_t current_era; + + /* + * We preallocate 2 writesets. When an era rolls over we + * switch between them. This means the allocation is done at + * preresume time, rather than on the io path. + */ + struct writeset writesets[2]; + struct writeset *current_writeset; + + dm_block_t writeset_tree_root; + dm_block_t era_array_root; + + struct dm_disk_bitset bitset_info; + struct dm_btree_info writeset_tree_info; + struct dm_array_info era_array_info; + + dm_block_t metadata_snap; + + /* + * A flag that is set whenever a writeset has been archived. + */ + bool archived_writesets; + + /* + * Reading the space map root can fail, so we read it into this + * buffer before the superblock is locked and updated. + */ + __u8 metadata_space_map_root[SPACE_MAP_ROOT_SIZE]; +}; + +static int superblock_read_lock(struct era_metadata *md, + struct dm_block **sblock) +{ + return dm_bm_read_lock(md->bm, SUPERBLOCK_LOCATION, + &sb_validator, sblock); +} + +static int superblock_lock_zero(struct era_metadata *md, + struct dm_block **sblock) +{ + return dm_bm_write_lock_zero(md->bm, SUPERBLOCK_LOCATION, + &sb_validator, sblock); +} + +static int superblock_lock(struct era_metadata *md, + struct dm_block **sblock) +{ + return dm_bm_write_lock(md->bm, SUPERBLOCK_LOCATION, + &sb_validator, sblock); +} + +/* FIXME: duplication with cache and thin */ +static int superblock_all_zeroes(struct dm_block_manager *bm, bool *result) +{ + int r; + unsigned i; + struct dm_block *b; + __le64 *data_le, zero = cpu_to_le64(0); + unsigned sb_block_size = dm_bm_block_size(bm) / sizeof(__le64); + + /* + * We can't use a validator here - it may be all zeroes. + */ + r = dm_bm_read_lock(bm, SUPERBLOCK_LOCATION, NULL, &b); + if (r) + return r; + + data_le = dm_block_data(b); + *result = true; + for (i = 0; i < sb_block_size; i++) { + if (data_le[i] != zero) { + *result = false; + break; + } + } + + return dm_bm_unlock(b); +} + +/*----------------------------------------------------------------*/ + +static void ws_pack(const struct writeset_metadata *core, struct writeset_disk *disk) +{ + disk->nr_bits = cpu_to_le32(core->nr_bits); + disk->root = cpu_to_le64(core->root); +} + +static void ws_unpack(const struct writeset_disk *disk, struct writeset_metadata *core) +{ + core->nr_bits = le32_to_cpu(disk->nr_bits); + core->root = le64_to_cpu(disk->root); +} + +static void ws_inc(void *context, const void *value) +{ + struct era_metadata *md = context; + struct writeset_disk ws_d; + dm_block_t b; + + memcpy(&ws_d, value, sizeof(ws_d)); + b = le64_to_cpu(ws_d.root); + + dm_tm_inc(md->tm, b); +} + +static void ws_dec(void *context, const void *value) +{ + struct era_metadata *md = context; + struct writeset_disk ws_d; + dm_block_t b; + + memcpy(&ws_d, value, sizeof(ws_d)); + b = le64_to_cpu(ws_d.root); + + dm_bitset_del(&md->bitset_info, b); +} + +static int ws_eq(void *context, const void *value1, const void *value2) +{ + return !memcmp(value1, value2, sizeof(struct writeset_metadata)); +} + +/*----------------------------------------------------------------*/ + +static void setup_writeset_tree_info(struct era_metadata *md) +{ + struct dm_btree_value_type *vt = &md->writeset_tree_info.value_type; + md->writeset_tree_info.tm = md->tm; + md->writeset_tree_info.levels = 1; + vt->context = md; + vt->size = sizeof(struct writeset_disk); + vt->inc = ws_inc; + vt->dec = ws_dec; + vt->equal = ws_eq; +} + +static void setup_era_array_info(struct era_metadata *md) + +{ + struct dm_btree_value_type vt; + vt.context = NULL; + vt.size = sizeof(__le32); + vt.inc = NULL; + vt.dec = NULL; + vt.equal = NULL; + + dm_array_info_init(&md->era_array_info, md->tm, &vt); +} + +static void setup_infos(struct era_metadata *md) +{ + dm_disk_bitset_init(md->tm, &md->bitset_info); + setup_writeset_tree_info(md); + setup_era_array_info(md); +} + +/*----------------------------------------------------------------*/ + +static int create_fresh_metadata(struct era_metadata *md) +{ + int r; + + r = dm_tm_create_with_sm(md->bm, SUPERBLOCK_LOCATION, + &md->tm, &md->sm); + if (r < 0) { + DMERR("dm_tm_create_with_sm failed"); + return r; + } + + setup_infos(md); + + r = dm_btree_empty(&md->writeset_tree_info, &md->writeset_tree_root); + if (r) { + DMERR("couldn't create new writeset tree"); + goto bad; + } + + r = dm_array_empty(&md->era_array_info, &md->era_array_root); + if (r) { + DMERR("couldn't create era array"); + goto bad; + } + + return 0; + +bad: + dm_sm_destroy(md->sm); + dm_tm_destroy(md->tm); + + return r; +} + +static int save_sm_root(struct era_metadata *md) +{ + int r; + size_t metadata_len; + + r = dm_sm_root_size(md->sm, &metadata_len); + if (r < 0) + return r; + + return dm_sm_copy_root(md->sm, &md->metadata_space_map_root, + metadata_len); +} + +static void copy_sm_root(struct era_metadata *md, struct superblock_disk *disk) +{ + memcpy(&disk->metadata_space_map_root, + &md->metadata_space_map_root, + sizeof(md->metadata_space_map_root)); +} + +/* + * Writes a superblock, including the static fields that don't get updated + * with every commit (possible optimisation here). 'md' should be fully + * constructed when this is called. + */ +static void prepare_superblock(struct era_metadata *md, struct superblock_disk *disk) +{ + disk->magic = cpu_to_le64(SUPERBLOCK_MAGIC); + disk->flags = cpu_to_le32(0ul); + + /* FIXME: can't keep blanking the uuid (uuid is currently unused though) */ + memset(disk->uuid, 0, sizeof(disk->uuid)); + disk->version = cpu_to_le32(MAX_ERA_VERSION); + + copy_sm_root(md, disk); + + disk->data_block_size = cpu_to_le32(md->block_size); + disk->metadata_block_size = cpu_to_le32(DM_ERA_METADATA_BLOCK_SIZE >> SECTOR_SHIFT); + disk->nr_blocks = cpu_to_le32(md->nr_blocks); + disk->current_era = cpu_to_le32(md->current_era); + + ws_pack(&md->current_writeset->md, &disk->current_writeset); + disk->writeset_tree_root = cpu_to_le64(md->writeset_tree_root); + disk->era_array_root = cpu_to_le64(md->era_array_root); + disk->metadata_snap = cpu_to_le64(md->metadata_snap); +} + +static int write_superblock(struct era_metadata *md) +{ + int r; + struct dm_block *sblock; + struct superblock_disk *disk; + + r = save_sm_root(md); + if (r) { + DMERR("%s: save_sm_root failed", __func__); + return r; + } + + r = superblock_lock_zero(md, &sblock); + if (r) + return r; + + disk = dm_block_data(sblock); + prepare_superblock(md, disk); + + return dm_tm_commit(md->tm, sblock); +} + +/* + * Assumes block_size and the infos are set. + */ +static int format_metadata(struct era_metadata *md) +{ + int r; + + r = create_fresh_metadata(md); + if (r) + return r; + + r = write_superblock(md); + if (r) { + dm_sm_destroy(md->sm); + dm_tm_destroy(md->tm); + return r; + } + + return 0; +} + +static int open_metadata(struct era_metadata *md) +{ + int r; + struct dm_block *sblock; + struct superblock_disk *disk; + + r = superblock_read_lock(md, &sblock); + if (r) { + DMERR("couldn't read_lock superblock"); + return r; + } + + disk = dm_block_data(sblock); + r = dm_tm_open_with_sm(md->bm, SUPERBLOCK_LOCATION, + disk->metadata_space_map_root, + sizeof(disk->metadata_space_map_root), + &md->tm, &md->sm); + if (r) { + DMERR("dm_tm_open_with_sm failed"); + goto bad; + } + + setup_infos(md); + + md->block_size = le32_to_cpu(disk->data_block_size); + md->nr_blocks = le32_to_cpu(disk->nr_blocks); + md->current_era = le32_to_cpu(disk->current_era); + + md->writeset_tree_root = le64_to_cpu(disk->writeset_tree_root); + md->era_array_root = le64_to_cpu(disk->era_array_root); + md->metadata_snap = le64_to_cpu(disk->metadata_snap); + md->archived_writesets = true; + + return dm_bm_unlock(sblock); + +bad: + dm_bm_unlock(sblock); + return r; +} + +static int open_or_format_metadata(struct era_metadata *md, + bool may_format) +{ + int r; + bool unformatted = false; + + r = superblock_all_zeroes(md->bm, &unformatted); + if (r) + return r; + + if (unformatted) + return may_format ? format_metadata(md) : -EPERM; + + return open_metadata(md); +} + +static int create_persistent_data_objects(struct era_metadata *md, + bool may_format) +{ + int r; + + md->bm = dm_block_manager_create(md->bdev, DM_ERA_METADATA_BLOCK_SIZE, + DM_ERA_METADATA_CACHE_SIZE, + ERA_MAX_CONCURRENT_LOCKS); + if (IS_ERR(md->bm)) { + DMERR("could not create block manager"); + return PTR_ERR(md->bm); + } + + r = open_or_format_metadata(md, may_format); + if (r) + dm_block_manager_destroy(md->bm); + + return r; +} + +static void destroy_persistent_data_objects(struct era_metadata *md) +{ + dm_sm_destroy(md->sm); + dm_tm_destroy(md->tm); + dm_block_manager_destroy(md->bm); +} + +/* + * This waits until all era_map threads have picked up the new filter. + */ +static void swap_writeset(struct era_metadata *md, struct writeset *new_writeset) +{ + rcu_assign_pointer(md->current_writeset, new_writeset); + synchronize_rcu(); +} + +/*---------------------------------------------------------------- + * Writesets get 'digested' into the main era array. + * + * We're using a coroutine here so the worker thread can do the digestion, + * thus avoiding synchronisation of the metadata. Digesting a whole + * writeset in one go would cause too much latency. + *--------------------------------------------------------------*/ +struct digest { + uint32_t era; + unsigned nr_bits, current_bit; + struct writeset_metadata writeset; + __le32 value; + struct dm_disk_bitset info; + + int (*step)(struct era_metadata *, struct digest *); +}; + +static int metadata_digest_lookup_writeset(struct era_metadata *md, + struct digest *d); + +static int metadata_digest_remove_writeset(struct era_metadata *md, + struct digest *d) +{ + int r; + uint64_t key = d->era; + + r = dm_btree_remove(&md->writeset_tree_info, md->writeset_tree_root, + &key, &md->writeset_tree_root); + if (r) { + DMERR("%s: dm_btree_remove failed", __func__); + return r; + } + + d->step = metadata_digest_lookup_writeset; + return 0; +} + +#define INSERTS_PER_STEP 100 + +static int metadata_digest_transcribe_writeset(struct era_metadata *md, + struct digest *d) +{ + int r; + bool marked; + unsigned b, e = min(d->current_bit + INSERTS_PER_STEP, d->nr_bits); + + for (b = d->current_bit; b < e; b++) { + r = writeset_marked_on_disk(&d->info, &d->writeset, b, &marked); + if (r) { + DMERR("%s: writeset_marked_on_disk failed", __func__); + return r; + } + + if (!marked) + continue; + + __dm_bless_for_disk(&d->value); + r = dm_array_set_value(&md->era_array_info, md->era_array_root, + b, &d->value, &md->era_array_root); + if (r) { + DMERR("%s: dm_array_set_value failed", __func__); + return r; + } + } + + if (b == d->nr_bits) + d->step = metadata_digest_remove_writeset; + else + d->current_bit = b; + + return 0; +} + +static int metadata_digest_lookup_writeset(struct era_metadata *md, + struct digest *d) +{ + int r; + uint64_t key; + struct writeset_disk disk; + + r = dm_btree_find_lowest_key(&md->writeset_tree_info, + md->writeset_tree_root, &key); + if (r < 0) + return r; + + d->era = key; + + r = dm_btree_lookup(&md->writeset_tree_info, + md->writeset_tree_root, &key, &disk); + if (r) { + if (r == -ENODATA) { + d->step = NULL; + return 0; + } + + DMERR("%s: dm_btree_lookup failed", __func__); + return r; + } + + ws_unpack(&disk, &d->writeset); + d->value = cpu_to_le32(key); + + d->nr_bits = min(d->writeset.nr_bits, md->nr_blocks); + d->current_bit = 0; + d->step = metadata_digest_transcribe_writeset; + + return 0; +} + +static int metadata_digest_start(struct era_metadata *md, struct digest *d) +{ + if (d->step) + return 0; + + memset(d, 0, sizeof(*d)); + + /* + * We initialise another bitset info to avoid any caching side + * effects with the previous one. + */ + dm_disk_bitset_init(md->tm, &d->info); + d->step = metadata_digest_lookup_writeset; + + return 0; +} + +/*---------------------------------------------------------------- + * High level metadata interface. Target methods should use these, and not + * the lower level ones. + *--------------------------------------------------------------*/ +static struct era_metadata *metadata_open(struct block_device *bdev, + sector_t block_size, + bool may_format) +{ + int r; + struct era_metadata *md = kzalloc(sizeof(*md), GFP_KERNEL); + + if (!md) + return NULL; + + md->bdev = bdev; + md->block_size = block_size; + + md->writesets[0].md.root = INVALID_WRITESET_ROOT; + md->writesets[1].md.root = INVALID_WRITESET_ROOT; + md->current_writeset = &md->writesets[0]; + + r = create_persistent_data_objects(md, may_format); + if (r) { + kfree(md); + return ERR_PTR(r); + } + + return md; +} + +static void metadata_close(struct era_metadata *md) +{ + destroy_persistent_data_objects(md); + kfree(md); +} + +static bool valid_nr_blocks(dm_block_t n) +{ + /* + * dm_bitset restricts us to 2^32. test_bit & co. restrict us + * further to 2^31 - 1 + */ + return n < (1ull << 31); +} + +static int metadata_resize(struct era_metadata *md, void *arg) +{ + int r; + dm_block_t *new_size = arg; + __le32 value; + + if (!valid_nr_blocks(*new_size)) { + DMERR("Invalid number of origin blocks %llu", + (unsigned long long) *new_size); + return -EINVAL; + } + + writeset_free(&md->writesets[0]); + writeset_free(&md->writesets[1]); + + r = writeset_alloc(&md->writesets[0], *new_size); + if (r) { + DMERR("%s: writeset_alloc failed for writeset 0", __func__); + return r; + } + + r = writeset_alloc(&md->writesets[1], *new_size); + if (r) { + DMERR("%s: writeset_alloc failed for writeset 1", __func__); + return r; + } + + value = cpu_to_le32(0u); + __dm_bless_for_disk(&value); + r = dm_array_resize(&md->era_array_info, md->era_array_root, + md->nr_blocks, *new_size, + &value, &md->era_array_root); + if (r) { + DMERR("%s: dm_array_resize failed", __func__); + return r; + } + + md->nr_blocks = *new_size; + return 0; +} + +static int metadata_era_archive(struct era_metadata *md) +{ + int r; + uint64_t keys[1]; + struct writeset_disk value; + + r = dm_bitset_flush(&md->bitset_info, md->current_writeset->md.root, + &md->current_writeset->md.root); + if (r) { + DMERR("%s: dm_bitset_flush failed", __func__); + return r; + } + + ws_pack(&md->current_writeset->md, &value); + md->current_writeset->md.root = INVALID_WRITESET_ROOT; + + keys[0] = md->current_era; + __dm_bless_for_disk(&value); + r = dm_btree_insert(&md->writeset_tree_info, md->writeset_tree_root, + keys, &value, &md->writeset_tree_root); + if (r) { + DMERR("%s: couldn't insert writeset into btree", __func__); + /* FIXME: fail mode */ + return r; + } + + md->archived_writesets = true; + + return 0; +} + +static struct writeset *next_writeset(struct era_metadata *md) +{ + return (md->current_writeset == &md->writesets[0]) ? + &md->writesets[1] : &md->writesets[0]; +} + +static int metadata_new_era(struct era_metadata *md) +{ + int r; + struct writeset *new_writeset = next_writeset(md); + + r = writeset_init(&md->bitset_info, new_writeset); + if (r) { + DMERR("%s: writeset_init failed", __func__); + return r; + } + + swap_writeset(md, new_writeset); + md->current_era++; + + return 0; +} + +static int metadata_era_rollover(struct era_metadata *md) +{ + int r; + + if (md->current_writeset->md.root != INVALID_WRITESET_ROOT) { + r = metadata_era_archive(md); + if (r) { + DMERR("%s: metadata_archive_era failed", __func__); + /* FIXME: fail mode? */ + return r; + } + } + + r = metadata_new_era(md); + if (r) { + DMERR("%s: new era failed", __func__); + /* FIXME: fail mode */ + return r; + } + + return 0; +} + +static bool metadata_current_marked(struct era_metadata *md, dm_block_t block) +{ + bool r; + struct writeset *ws; + + rcu_read_lock(); + ws = rcu_dereference(md->current_writeset); + r = writeset_marked(ws, block); + rcu_read_unlock(); + + return r; +} + +static int metadata_commit(struct era_metadata *md) +{ + int r; + struct dm_block *sblock; + + if (md->current_writeset->md.root != SUPERBLOCK_LOCATION) { + r = dm_bitset_flush(&md->bitset_info, md->current_writeset->md.root, + &md->current_writeset->md.root); + if (r) { + DMERR("%s: bitset flush failed", __func__); + return r; + } + } + + r = save_sm_root(md); + if (r) { + DMERR("%s: save_sm_root failed", __func__); + return r; + } + + r = dm_tm_pre_commit(md->tm); + if (r) { + DMERR("%s: pre commit failed", __func__); + return r; + } + + r = superblock_lock(md, &sblock); + if (r) { + DMERR("%s: superblock lock failed", __func__); + return r; + } + + prepare_superblock(md, dm_block_data(sblock)); + + return dm_tm_commit(md->tm, sblock); +} + +static int metadata_checkpoint(struct era_metadata *md) +{ + /* + * For now we just rollover, but later I want to put a check in to + * avoid this if the filter is still pretty fresh. + */ + return metadata_era_rollover(md); +} + +/* + * Metadata snapshots allow userland to access era data. + */ +static int metadata_take_snap(struct era_metadata *md) +{ + int r, inc; + struct dm_block *clone; + + if (md->metadata_snap != SUPERBLOCK_LOCATION) { + DMERR("%s: metadata snapshot already exists", __func__); + return -EINVAL; + } + + r = metadata_era_rollover(md); + if (r) { + DMERR("%s: era rollover failed", __func__); + return r; + } + + r = metadata_commit(md); + if (r) { + DMERR("%s: pre commit failed", __func__); + return r; + } + + r = dm_sm_inc_block(md->sm, SUPERBLOCK_LOCATION); + if (r) { + DMERR("%s: couldn't increment superblock", __func__); + return r; + } + + r = dm_tm_shadow_block(md->tm, SUPERBLOCK_LOCATION, + &sb_validator, &clone, &inc); + if (r) { + DMERR("%s: couldn't shadow superblock", __func__); + dm_sm_dec_block(md->sm, SUPERBLOCK_LOCATION); + return r; + } + BUG_ON(!inc); + + r = dm_sm_inc_block(md->sm, md->writeset_tree_root); + if (r) { + DMERR("%s: couldn't inc writeset tree root", __func__); + dm_tm_unlock(md->tm, clone); + return r; + } + + r = dm_sm_inc_block(md->sm, md->era_array_root); + if (r) { + DMERR("%s: couldn't inc era tree root", __func__); + dm_sm_dec_block(md->sm, md->writeset_tree_root); + dm_tm_unlock(md->tm, clone); + return r; + } + + md->metadata_snap = dm_block_location(clone); + + r = dm_tm_unlock(md->tm, clone); + if (r) { + DMERR("%s: couldn't unlock clone", __func__); + md->metadata_snap = SUPERBLOCK_LOCATION; + return r; + } + + return 0; +} + +static int metadata_drop_snap(struct era_metadata *md) +{ + int r; + dm_block_t location; + struct dm_block *clone; + struct superblock_disk *disk; + + if (md->metadata_snap == SUPERBLOCK_LOCATION) { + DMERR("%s: no snap to drop", __func__); + return -EINVAL; + } + + r = dm_tm_read_lock(md->tm, md->metadata_snap, &sb_validator, &clone); + if (r) { + DMERR("%s: couldn't read lock superblock clone", __func__); + return r; + } + + /* + * Whatever happens now we'll commit with no record of the metadata + * snap. + */ + md->metadata_snap = SUPERBLOCK_LOCATION; + + disk = dm_block_data(clone); + r = dm_btree_del(&md->writeset_tree_info, + le64_to_cpu(disk->writeset_tree_root)); + if (r) { + DMERR("%s: error deleting writeset tree clone", __func__); + dm_tm_unlock(md->tm, clone); + return r; + } + + r = dm_array_del(&md->era_array_info, le64_to_cpu(disk->era_array_root)); + if (r) { + DMERR("%s: error deleting era array clone", __func__); + dm_tm_unlock(md->tm, clone); + return r; + } + + location = dm_block_location(clone); + dm_tm_unlock(md->tm, clone); + + return dm_sm_dec_block(md->sm, location); +} + +struct metadata_stats { + dm_block_t used; + dm_block_t total; + dm_block_t snap; + uint32_t era; +}; + +static int metadata_get_stats(struct era_metadata *md, void *ptr) +{ + int r; + struct metadata_stats *s = ptr; + dm_block_t nr_free, nr_total; + + r = dm_sm_get_nr_free(md->sm, &nr_free); + if (r) { + DMERR("dm_sm_get_nr_free returned %d", r); + return r; + } + + r = dm_sm_get_nr_blocks(md->sm, &nr_total); + if (r) { + DMERR("dm_pool_get_metadata_dev_size returned %d", r); + return r; + } + + s->used = nr_total - nr_free; + s->total = nr_total; + s->snap = md->metadata_snap; + s->era = md->current_era; + + return 0; +} + +/*----------------------------------------------------------------*/ + +struct era { + struct dm_target *ti; + struct dm_target_callbacks callbacks; + + struct dm_dev *metadata_dev; + struct dm_dev *origin_dev; + + dm_block_t nr_blocks; + uint32_t sectors_per_block; + int sectors_per_block_shift; + struct era_metadata *md; + + struct workqueue_struct *wq; + struct work_struct worker; + + spinlock_t deferred_lock; + struct bio_list deferred_bios; + + spinlock_t rpc_lock; + struct list_head rpc_calls; + + struct digest digest; + atomic_t suspended; +}; + +struct rpc { + struct list_head list; + + int (*fn0)(struct era_metadata *); + int (*fn1)(struct era_metadata *, void *); + void *arg; + int result; + + struct completion complete; +}; + +/*---------------------------------------------------------------- + * Remapping. + *---------------------------------------------------------------*/ +static bool block_size_is_power_of_two(struct era *era) +{ + return era->sectors_per_block_shift >= 0; +} + +static dm_block_t get_block(struct era *era, struct bio *bio) +{ + sector_t block_nr = bio->bi_iter.bi_sector; + + if (!block_size_is_power_of_two(era)) + (void) sector_div(block_nr, era->sectors_per_block); + else + block_nr >>= era->sectors_per_block_shift; + + return block_nr; +} + +static void remap_to_origin(struct era *era, struct bio *bio) +{ + bio->bi_bdev = era->origin_dev->bdev; +} + +/*---------------------------------------------------------------- + * Worker thread + *--------------------------------------------------------------*/ +static void wake_worker(struct era *era) +{ + if (!atomic_read(&era->suspended)) + queue_work(era->wq, &era->worker); +} + +static void process_old_eras(struct era *era) +{ + int r; + + if (!era->digest.step) + return; + + r = era->digest.step(era->md, &era->digest); + if (r < 0) { + DMERR("%s: digest step failed, stopping digestion", __func__); + era->digest.step = NULL; + + } else if (era->digest.step) + wake_worker(era); +} + +static void process_deferred_bios(struct era *era) +{ + int r; + struct bio_list deferred_bios, marked_bios; + struct bio *bio; + bool commit_needed = false; + bool failed = false; + + bio_list_init(&deferred_bios); + bio_list_init(&marked_bios); + + spin_lock(&era->deferred_lock); + bio_list_merge(&deferred_bios, &era->deferred_bios); + bio_list_init(&era->deferred_bios); + spin_unlock(&era->deferred_lock); + + while ((bio = bio_list_pop(&deferred_bios))) { + r = writeset_test_and_set(&era->md->bitset_info, + era->md->current_writeset, + get_block(era, bio)); + if (r < 0) { + /* + * This is bad news, we need to rollback. + * FIXME: finish. + */ + failed = true; + + } else if (r == 0) + commit_needed = true; + + bio_list_add(&marked_bios, bio); + } + + if (commit_needed) { + r = metadata_commit(era->md); + if (r) + failed = true; + } + + if (failed) + while ((bio = bio_list_pop(&marked_bios))) + bio_io_error(bio); + else + while ((bio = bio_list_pop(&marked_bios))) + generic_make_request(bio); +} + +static void process_rpc_calls(struct era *era) +{ + int r; + bool need_commit = false; + struct list_head calls; + struct rpc *rpc, *tmp; + + INIT_LIST_HEAD(&calls); + spin_lock(&era->rpc_lock); + list_splice_init(&era->rpc_calls, &calls); + spin_unlock(&era->rpc_lock); + + list_for_each_entry_safe(rpc, tmp, &calls, list) { + rpc->result = rpc->fn0 ? rpc->fn0(era->md) : rpc->fn1(era->md, rpc->arg); + need_commit = true; + } + + if (need_commit) { + r = metadata_commit(era->md); + if (r) + list_for_each_entry_safe(rpc, tmp, &calls, list) + rpc->result = r; + } + + list_for_each_entry_safe(rpc, tmp, &calls, list) + complete(&rpc->complete); +} + +static void kick_off_digest(struct era *era) +{ + if (era->md->archived_writesets) { + era->md->archived_writesets = false; + metadata_digest_start(era->md, &era->digest); + } +} + +static void do_work(struct work_struct *ws) +{ + struct era *era = container_of(ws, struct era, worker); + + kick_off_digest(era); + process_old_eras(era); + process_deferred_bios(era); + process_rpc_calls(era); +} + +static void defer_bio(struct era *era, struct bio *bio) +{ + spin_lock(&era->deferred_lock); + bio_list_add(&era->deferred_bios, bio); + spin_unlock(&era->deferred_lock); + + wake_worker(era); +} + +/* + * Make an rpc call to the worker to change the metadata. + */ +static int perform_rpc(struct era *era, struct rpc *rpc) +{ + rpc->result = 0; + init_completion(&rpc->complete); + + spin_lock(&era->rpc_lock); + list_add(&rpc->list, &era->rpc_calls); + spin_unlock(&era->rpc_lock); + + wake_worker(era); + wait_for_completion(&rpc->complete); + + return rpc->result; +} + +static int in_worker0(struct era *era, int (*fn)(struct era_metadata *)) +{ + struct rpc rpc; + rpc.fn0 = fn; + rpc.fn1 = NULL; + + return perform_rpc(era, &rpc); +} + +static int in_worker1(struct era *era, + int (*fn)(struct era_metadata *, void *), void *arg) +{ + struct rpc rpc; + rpc.fn0 = NULL; + rpc.fn1 = fn; + rpc.arg = arg; + + return perform_rpc(era, &rpc); +} + +static void start_worker(struct era *era) +{ + atomic_set(&era->suspended, 0); +} + +static void stop_worker(struct era *era) +{ + atomic_set(&era->suspended, 1); + flush_workqueue(era->wq); +} + +/*---------------------------------------------------------------- + * Target methods + *--------------------------------------------------------------*/ +static int dev_is_congested(struct dm_dev *dev, int bdi_bits) +{ + struct request_queue *q = bdev_get_queue(dev->bdev); + return bdi_congested(&q->backing_dev_info, bdi_bits); +} + +static int era_is_congested(struct dm_target_callbacks *cb, int bdi_bits) +{ + struct era *era = container_of(cb, struct era, callbacks); + return dev_is_congested(era->origin_dev, bdi_bits); +} + +static void era_destroy(struct era *era) +{ + if (era->md) + metadata_close(era->md); + + if (era->wq) + destroy_workqueue(era->wq); + + if (era->origin_dev) + dm_put_device(era->ti, era->origin_dev); + + if (era->metadata_dev) + dm_put_device(era->ti, era->metadata_dev); + + kfree(era); +} + +static dm_block_t calc_nr_blocks(struct era *era) +{ + return dm_sector_div_up(era->ti->len, era->sectors_per_block); +} + +static bool valid_block_size(dm_block_t block_size) +{ + bool greater_than_zero = block_size > 0; + bool multiple_of_min_block_size = (block_size & (MIN_BLOCK_SIZE - 1)) == 0; + + return greater_than_zero && multiple_of_min_block_size; +} + +/* + * <metadata dev> <data dev> <data block size (sectors)> + */ +static int era_ctr(struct dm_target *ti, unsigned argc, char **argv) +{ + int r; + char dummy; + struct era *era; + struct era_metadata *md; + + if (argc != 3) { + ti->error = "Invalid argument count"; + return -EINVAL; + } + + era = kzalloc(sizeof(*era), GFP_KERNEL); + if (!era) { + ti->error = "Error allocating era structure"; + return -ENOMEM; + } + + era->ti = ti; + + r = dm_get_device(ti, argv[0], FMODE_READ | FMODE_WRITE, &era->metadata_dev); + if (r) { + ti->error = "Error opening metadata device"; + era_destroy(era); + return -EINVAL; + } + + r = dm_get_device(ti, argv[1], FMODE_READ | FMODE_WRITE, &era->origin_dev); + if (r) { + ti->error = "Error opening data device"; + era_destroy(era); + return -EINVAL; + } + + r = sscanf(argv[2], "%u%c", &era->sectors_per_block, &dummy); + if (r != 1) { + ti->error = "Error parsing block size"; + era_destroy(era); + return -EINVAL; + } + + r = dm_set_target_max_io_len(ti, era->sectors_per_block); + if (r) { + ti->error = "could not set max io len"; + era_destroy(era); + return -EINVAL; + } + + if (!valid_block_size(era->sectors_per_block)) { + ti->error = "Invalid block size"; + era_destroy(era); + return -EINVAL; + } + if (era->sectors_per_block & (era->sectors_per_block - 1)) + era->sectors_per_block_shift = -1; + else + era->sectors_per_block_shift = __ffs(era->sectors_per_block); + + md = metadata_open(era->metadata_dev->bdev, era->sectors_per_block, true); + if (IS_ERR(md)) { + ti->error = "Error reading metadata"; + era_destroy(era); + return PTR_ERR(md); + } + era->md = md; + + era->nr_blocks = calc_nr_blocks(era); + + r = metadata_resize(era->md, &era->nr_blocks); + if (r) { + ti->error = "couldn't resize metadata"; + era_destroy(era); + return -ENOMEM; + } + + era->wq = alloc_ordered_workqueue("dm-" DM_MSG_PREFIX, WQ_MEM_RECLAIM); + if (!era->wq) { + ti->error = "could not create workqueue for metadata object"; + era_destroy(era); + return -ENOMEM; + } + INIT_WORK(&era->worker, do_work); + + spin_lock_init(&era->deferred_lock); + bio_list_init(&era->deferred_bios); + + spin_lock_init(&era->rpc_lock); + INIT_LIST_HEAD(&era->rpc_calls); + + ti->private = era; + ti->num_flush_bios = 1; + ti->flush_supported = true; + + ti->num_discard_bios = 1; + ti->discards_supported = true; + era->callbacks.congested_fn = era_is_congested; + dm_table_add_target_callbacks(ti->table, &era->callbacks); + + return 0; +} + +static void era_dtr(struct dm_target *ti) +{ + era_destroy(ti->private); +} + +static int era_map(struct dm_target *ti, struct bio *bio) +{ + struct era *era = ti->private; + dm_block_t block = get_block(era, bio); + + /* + * All bios get remapped to the origin device. We do this now, but + * it may not get issued until later. Depending on whether the + * block is marked in this era. + */ + remap_to_origin(era, bio); + + /* + * REQ_FLUSH bios carry no data, so we're not interested in them. + */ + if (!(bio->bi_rw & REQ_FLUSH) && + (bio_data_dir(bio) == WRITE) && + !metadata_current_marked(era->md, block)) { + defer_bio(era, bio); + return DM_MAPIO_SUBMITTED; + } + + return DM_MAPIO_REMAPPED; +} + +static void era_postsuspend(struct dm_target *ti) +{ + int r; + struct era *era = ti->private; + + r = in_worker0(era, metadata_era_archive); + if (r) { + DMERR("%s: couldn't archive current era", __func__); + /* FIXME: fail mode */ + } + + stop_worker(era); +} + +static int era_preresume(struct dm_target *ti) +{ + int r; + struct era *era = ti->private; + dm_block_t new_size = calc_nr_blocks(era); + + if (era->nr_blocks != new_size) { + r = in_worker1(era, metadata_resize, &new_size); + if (r) + return r; + + era->nr_blocks = new_size; + } + + start_worker(era); + + r = in_worker0(era, metadata_new_era); + if (r) { + DMERR("%s: metadata_era_rollover failed", __func__); + return r; + } + + return 0; +} + +/* + * Status format: + * + * <metadata block size> <#used metadata blocks>/<#total metadata blocks> + * <current era> <held metadata root | '-'> + */ +static void era_status(struct dm_target *ti, status_type_t type, + unsigned status_flags, char *result, unsigned maxlen) +{ + int r; + struct era *era = ti->private; + ssize_t sz = 0; + struct metadata_stats stats; + char buf[BDEVNAME_SIZE]; + + switch (type) { + case STATUSTYPE_INFO: + r = in_worker1(era, metadata_get_stats, &stats); + if (r) + goto err; + + DMEMIT("%u %llu/%llu %u", + (unsigned) (DM_ERA_METADATA_BLOCK_SIZE >> SECTOR_SHIFT), + (unsigned long long) stats.used, + (unsigned long long) stats.total, + (unsigned) stats.era); + + if (stats.snap != SUPERBLOCK_LOCATION) + DMEMIT(" %llu", stats.snap); + else + DMEMIT(" -"); + break; + + case STATUSTYPE_TABLE: + format_dev_t(buf, era->metadata_dev->bdev->bd_dev); + DMEMIT("%s ", buf); + format_dev_t(buf, era->origin_dev->bdev->bd_dev); + DMEMIT("%s %u", buf, era->sectors_per_block); + break; + } + + return; + +err: + DMEMIT("Error"); +} + +static int era_message(struct dm_target *ti, unsigned argc, char **argv) +{ + struct era *era = ti->private; + + if (argc != 1) { + DMERR("incorrect number of message arguments"); + return -EINVAL; + } + + if (!strcasecmp(argv[0], "checkpoint")) + return in_worker0(era, metadata_checkpoint); + + if (!strcasecmp(argv[0], "take_metadata_snap")) + return in_worker0(era, metadata_take_snap); + + if (!strcasecmp(argv[0], "drop_metadata_snap")) + return in_worker0(era, metadata_drop_snap); + + DMERR("unsupported message '%s'", argv[0]); + return -EINVAL; +} + +static sector_t get_dev_size(struct dm_dev *dev) +{ + return i_size_read(dev->bdev->bd_inode) >> SECTOR_SHIFT; +} + +static int era_iterate_devices(struct dm_target *ti, + iterate_devices_callout_fn fn, void *data) +{ + struct era *era = ti->private; + return fn(ti, era->origin_dev, 0, get_dev_size(era->origin_dev), data); +} + +static int era_merge(struct dm_target *ti, struct bvec_merge_data *bvm, + struct bio_vec *biovec, int max_size) +{ + struct era *era = ti->private; + struct request_queue *q = bdev_get_queue(era->origin_dev->bdev); + + if (!q->merge_bvec_fn) + return max_size; + + bvm->bi_bdev = era->origin_dev->bdev; + + return min(max_size, q->merge_bvec_fn(q, bvm, biovec)); +} + +static void era_io_hints(struct dm_target *ti, struct queue_limits *limits) +{ + struct era *era = ti->private; + uint64_t io_opt_sectors = limits->io_opt >> SECTOR_SHIFT; + + /* + * If the system-determined stacked limits are compatible with the + * era device's blocksize (io_opt is a factor) do not override them. + */ + if (io_opt_sectors < era->sectors_per_block || + do_div(io_opt_sectors, era->sectors_per_block)) { + blk_limits_io_min(limits, 0); + blk_limits_io_opt(limits, era->sectors_per_block << SECTOR_SHIFT); + } +} + +/*----------------------------------------------------------------*/ + +static struct target_type era_target = { + .name = "era", + .version = {1, 0, 0}, + .module = THIS_MODULE, + .ctr = era_ctr, + .dtr = era_dtr, + .map = era_map, + .postsuspend = era_postsuspend, + .preresume = era_preresume, + .status = era_status, + .message = era_message, + .iterate_devices = era_iterate_devices, + .merge = era_merge, + .io_hints = era_io_hints +}; + +static int __init dm_era_init(void) +{ + int r; + + r = dm_register_target(&era_target); + if (r) { + DMERR("era target registration failed: %d", r); + return r; + } + + return 0; +} + +static void __exit dm_era_exit(void) +{ + dm_unregister_target(&era_target); +} + +module_init(dm_era_init); +module_exit(dm_era_exit); + +MODULE_DESCRIPTION(DM_NAME " era target"); +MODULE_AUTHOR("Joe Thornber <ejt@redhat.com>"); +MODULE_LICENSE("GPL"); diff --git a/drivers/md/dm-flakey.c b/drivers/md/dm-flakey.c index 7fcf21cb4ff..b257e46876d 100644 --- a/drivers/md/dm-flakey.c +++ b/drivers/md/dm-flakey.c @@ -176,7 +176,7 @@ static int flakey_ctr(struct dm_target *ti, unsigned int argc, char **argv) fc = kzalloc(sizeof(*fc), GFP_KERNEL); if (!fc) { - ti->error = "Cannot allocate linear context"; + ti->error = "Cannot allocate context"; return -ENOMEM; } fc->start_time = jiffies; @@ -248,7 +248,8 @@ static void flakey_map_bio(struct dm_target *ti, struct bio *bio) bio->bi_bdev = fc->dev->bdev; if (bio_sectors(bio)) - bio->bi_sector = flakey_map_sector(ti, bio->bi_sector); + bio->bi_iter.bi_sector = + flakey_map_sector(ti, bio->bi_iter.bi_sector); } static void corrupt_bio_data(struct bio *bio, struct flakey_c *fc) @@ -265,8 +266,8 @@ static void corrupt_bio_data(struct bio *bio, struct flakey_c *fc) DMDEBUG("Corrupting data bio=%p by writing %u to byte %u " "(rw=%c bi_rw=%lu bi_sector=%llu cur_bytes=%u)\n", bio, fc->corrupt_bio_value, fc->corrupt_bio_byte, - (bio_data_dir(bio) == WRITE) ? 'w' : 'r', - bio->bi_rw, (unsigned long long)bio->bi_sector, bio_bytes); + (bio_data_dir(bio) == WRITE) ? 'w' : 'r', bio->bi_rw, + (unsigned long long)bio->bi_iter.bi_sector, bio_bytes); } } diff --git a/drivers/md/dm-io.c b/drivers/md/dm-io.c index ea49834377c..db404a0f7e2 100644 --- a/drivers/md/dm-io.c +++ b/drivers/md/dm-io.c @@ -10,6 +10,7 @@ #include <linux/device-mapper.h> #include <linux/bio.h> +#include <linux/completion.h> #include <linux/mempool.h> #include <linux/module.h> #include <linux/sched.h> @@ -19,8 +20,6 @@ #define DM_MSG_PREFIX "io" #define DM_IO_MAX_REGIONS BITS_PER_LONG -#define MIN_IOS 16 -#define MIN_BIOS 16 struct dm_io_client { mempool_t *pool; @@ -34,7 +33,7 @@ struct dm_io_client { struct io { unsigned long error_bits; atomic_t count; - struct task_struct *sleeper; + struct completion *wait; struct dm_io_client *client; io_notify_fn callback; void *context; @@ -50,16 +49,17 @@ static struct kmem_cache *_dm_io_cache; struct dm_io_client *dm_io_client_create(void) { struct dm_io_client *client; + unsigned min_ios = dm_get_reserved_bio_based_ios(); client = kmalloc(sizeof(*client), GFP_KERNEL); if (!client) return ERR_PTR(-ENOMEM); - client->pool = mempool_create_slab_pool(MIN_IOS, _dm_io_cache); + client->pool = mempool_create_slab_pool(min_ios, _dm_io_cache); if (!client->pool) goto bad; - client->bios = bioset_create(MIN_BIOS, 0); + client->bios = bioset_create(min_ios, 0); if (!client->bios) goto bad; @@ -122,8 +122,8 @@ static void dec_count(struct io *io, unsigned int region, int error) invalidate_kernel_vmap_range(io->vma_invalidate_address, io->vma_invalidate_size); - if (io->sleeper) - wake_up_process(io->sleeper); + if (io->wait) + complete(io->wait); else { unsigned long r = io->error_bits; @@ -202,26 +202,28 @@ static void list_dp_init(struct dpages *dp, struct page_list *pl, unsigned offse /* * Functions for getting the pages from a bvec. */ -static void bvec_get_page(struct dpages *dp, - struct page **p, unsigned long *len, unsigned *offset) +static void bio_get_page(struct dpages *dp, struct page **p, + unsigned long *len, unsigned *offset) { - struct bio_vec *bvec = (struct bio_vec *) dp->context_ptr; + struct bio_vec *bvec = dp->context_ptr; *p = bvec->bv_page; - *len = bvec->bv_len; - *offset = bvec->bv_offset; + *len = bvec->bv_len - dp->context_u; + *offset = bvec->bv_offset + dp->context_u; } -static void bvec_next_page(struct dpages *dp) +static void bio_next_page(struct dpages *dp) { - struct bio_vec *bvec = (struct bio_vec *) dp->context_ptr; + struct bio_vec *bvec = dp->context_ptr; dp->context_ptr = bvec + 1; + dp->context_u = 0; } -static void bvec_dp_init(struct dpages *dp, struct bio_vec *bvec) +static void bio_dp_init(struct dpages *dp, struct bio *bio) { - dp->get_page = bvec_get_page; - dp->next_page = bvec_next_page; - dp->context_ptr = bvec; + dp->get_page = bio_get_page; + dp->next_page = bio_next_page; + dp->context_ptr = __bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter); + dp->context_u = bio->bi_iter.bi_bvec_done; } /* @@ -305,14 +307,14 @@ static void do_region(int rw, unsigned region, struct dm_io_region *where, dm_sector_div_up(remaining, (PAGE_SIZE >> SECTOR_SHIFT))); bio = bio_alloc_bioset(GFP_NOIO, num_bvecs, io->client->bios); - bio->bi_sector = where->sector + (where->count - remaining); + bio->bi_iter.bi_sector = where->sector + (where->count - remaining); bio->bi_bdev = where->bdev; bio->bi_end_io = endio; store_io_and_region_in_bio(bio, io, region); if (rw & REQ_DISCARD) { num_sectors = min_t(sector_t, q->limits.max_discard_sectors, remaining); - bio->bi_size = num_sectors << SECTOR_SHIFT; + bio->bi_iter.bi_size = num_sectors << SECTOR_SHIFT; remaining -= num_sectors; } else if (rw & REQ_WRITE_SAME) { /* @@ -321,7 +323,7 @@ static void do_region(int rw, unsigned region, struct dm_io_region *where, dp->get_page(dp, &page, &len, &offset); bio_add_page(bio, page, logical_block_size, offset); num_sectors = min_t(sector_t, q->limits.max_write_same_sectors, remaining); - bio->bi_size = num_sectors << SECTOR_SHIFT; + bio->bi_iter.bi_size = num_sectors << SECTOR_SHIFT; offset = 0; remaining -= num_sectors; @@ -386,6 +388,7 @@ static int sync_io(struct dm_io_client *client, unsigned int num_regions, */ volatile char io_[sizeof(struct io) + __alignof__(struct io) - 1]; struct io *io = (struct io *)PTR_ALIGN(&io_, __alignof__(struct io)); + DECLARE_COMPLETION_ONSTACK(wait); if (num_regions > 1 && (rw & RW_MASK) != WRITE) { WARN_ON(1); @@ -394,7 +397,7 @@ static int sync_io(struct dm_io_client *client, unsigned int num_regions, io->error_bits = 0; atomic_set(&io->count, 1); /* see dispatch_io() */ - io->sleeper = current; + io->wait = &wait; io->client = client; io->vma_invalidate_address = dp->vma_invalidate_address; @@ -402,15 +405,7 @@ static int sync_io(struct dm_io_client *client, unsigned int num_regions, dispatch_io(rw, num_regions, where, dp, io, 1); - while (1) { - set_current_state(TASK_UNINTERRUPTIBLE); - - if (!atomic_read(&io->count)) - break; - - io_schedule(); - } - set_current_state(TASK_RUNNING); + wait_for_completion_io(&wait); if (error_bits) *error_bits = io->error_bits; @@ -433,7 +428,7 @@ static int async_io(struct dm_io_client *client, unsigned int num_regions, io = mempool_alloc(client->pool, GFP_NOIO); io->error_bits = 0; atomic_set(&io->count, 1); /* see dispatch_io() */ - io->sleeper = NULL; + io->wait = NULL; io->client = client; io->callback = fn; io->context = context; @@ -458,8 +453,8 @@ static int dp_init(struct dm_io_request *io_req, struct dpages *dp, list_dp_init(dp, io_req->mem.ptr.pl, io_req->mem.offset); break; - case DM_IO_BVEC: - bvec_dp_init(dp, io_req->mem.ptr.bvec); + case DM_IO_BIO: + bio_dp_init(dp, io_req->mem.ptr.bio); break; case DM_IO_VMA: diff --git a/drivers/md/dm-ioctl.c b/drivers/md/dm-ioctl.c index aa04f022464..51521429fb5 100644 --- a/drivers/md/dm-ioctl.c +++ b/drivers/md/dm-ioctl.c @@ -36,6 +36,14 @@ struct hash_cell { struct dm_table *new_map; }; +/* + * A dummy definition to make RCU happy. + * struct dm_table should never be dereferenced in this file. + */ +struct dm_table { + int undefined__; +}; + struct vers_iter { size_t param_size; struct dm_target_versions *vers, *old_vers; @@ -49,7 +57,7 @@ struct vers_iter { static struct list_head _name_buckets[NUM_BUCKETS]; static struct list_head _uuid_buckets[NUM_BUCKETS]; -static void dm_hash_remove_all(int keep_open_devices); +static void dm_hash_remove_all(bool keep_open_devices, bool mark_deferred, bool only_deferred); /* * Guards access to both hash tables. @@ -78,7 +86,7 @@ static int dm_hash_init(void) static void dm_hash_exit(void) { - dm_hash_remove_all(0); + dm_hash_remove_all(false, false, false); } /*----------------------------------------------------------------- @@ -242,9 +250,10 @@ static int dm_hash_insert(const char *name, const char *uuid, struct mapped_devi return -EBUSY; } -static void __hash_remove(struct hash_cell *hc) +static struct dm_table *__hash_remove(struct hash_cell *hc) { struct dm_table *table; + int srcu_idx; /* remove from the dev hash */ list_del(&hc->uuid_list); @@ -253,23 +262,26 @@ static void __hash_remove(struct hash_cell *hc) dm_set_mdptr(hc->md, NULL); mutex_unlock(&dm_hash_cells_mutex); - table = dm_get_live_table(hc->md); - if (table) { + table = dm_get_live_table(hc->md, &srcu_idx); + if (table) dm_table_event(table); - dm_table_put(table); - } + dm_put_live_table(hc->md, srcu_idx); + table = NULL; if (hc->new_map) - dm_table_destroy(hc->new_map); + table = hc->new_map; dm_put(hc->md); free_cell(hc); + + return table; } -static void dm_hash_remove_all(int keep_open_devices) +static void dm_hash_remove_all(bool keep_open_devices, bool mark_deferred, bool only_deferred) { int i, dev_skipped; struct hash_cell *hc; struct mapped_device *md; + struct dm_table *t; retry: dev_skipped = 0; @@ -281,16 +293,21 @@ retry: md = hc->md; dm_get(md); - if (keep_open_devices && dm_lock_for_deletion(md)) { + if (keep_open_devices && + dm_lock_for_deletion(md, mark_deferred, only_deferred)) { dm_put(md); dev_skipped++; continue; } - __hash_remove(hc); + t = __hash_remove(hc); up_write(&_hash_lock); + if (t) { + dm_sync_table(md); + dm_table_destroy(t); + } dm_put(md); if (likely(keep_open_devices)) dm_destroy(md); @@ -356,6 +373,7 @@ static struct mapped_device *dm_hash_rename(struct dm_ioctl *param, struct dm_table *table; struct mapped_device *md; unsigned change_uuid = (param->flags & DM_UUID_FLAG) ? 1 : 0; + int srcu_idx; /* * duplicate new. @@ -418,11 +436,10 @@ static struct mapped_device *dm_hash_rename(struct dm_ioctl *param, /* * Wake up any dm event waiters. */ - table = dm_get_live_table(hc->md); - if (table) { + table = dm_get_live_table(hc->md, &srcu_idx); + if (table) dm_table_event(table); - dm_table_put(table); - } + dm_put_live_table(hc->md, srcu_idx); if (!dm_kobject_uevent(hc->md, KOBJ_CHANGE, param->event_nr)) param->flags |= DM_UEVENT_GENERATED_FLAG; @@ -434,6 +451,11 @@ static struct mapped_device *dm_hash_rename(struct dm_ioctl *param, return md; } +void dm_deferred_remove(void) +{ + dm_hash_remove_all(true, false, true); +} + /*----------------------------------------------------------------- * Implementation of the ioctl commands *---------------------------------------------------------------*/ @@ -445,7 +467,7 @@ typedef int (*ioctl_fn)(struct dm_ioctl *param, size_t param_size); static int remove_all(struct dm_ioctl *param, size_t param_size) { - dm_hash_remove_all(1); + dm_hash_remove_all(true, !!(param->flags & DM_DEFERRED_REMOVE), false); param->data_size = 0; return 0; } @@ -620,11 +642,14 @@ static int check_name(const char *name) * _hash_lock without first calling dm_table_put, because dm_table_destroy * waits for this dm_table_put and could be called under this lock. */ -static struct dm_table *dm_get_inactive_table(struct mapped_device *md) +static struct dm_table *dm_get_inactive_table(struct mapped_device *md, int *srcu_idx) { struct hash_cell *hc; struct dm_table *table = NULL; + /* increment rcu count, we don't care about the table pointer */ + dm_get_live_table(md, srcu_idx); + down_read(&_hash_lock); hc = dm_get_mdptr(md); if (!hc || hc->md != md) { @@ -633,8 +658,6 @@ static struct dm_table *dm_get_inactive_table(struct mapped_device *md) } table = hc->new_map; - if (table) - dm_table_get(table); out: up_read(&_hash_lock); @@ -643,10 +666,11 @@ out: } static struct dm_table *dm_get_live_or_inactive_table(struct mapped_device *md, - struct dm_ioctl *param) + struct dm_ioctl *param, + int *srcu_idx) { return (param->flags & DM_QUERY_INACTIVE_TABLE_FLAG) ? - dm_get_inactive_table(md) : dm_get_live_table(md); + dm_get_inactive_table(md, srcu_idx) : dm_get_live_table(md, srcu_idx); } /* @@ -657,6 +681,7 @@ static void __dev_status(struct mapped_device *md, struct dm_ioctl *param) { struct gendisk *disk = dm_disk(md); struct dm_table *table; + int srcu_idx; param->flags &= ~(DM_SUSPEND_FLAG | DM_READONLY_FLAG | DM_ACTIVE_PRESENT_FLAG); @@ -664,6 +689,9 @@ static void __dev_status(struct mapped_device *md, struct dm_ioctl *param) if (dm_suspended_md(md)) param->flags |= DM_SUSPEND_FLAG; + if (dm_test_deferred_remove_flag(md)) + param->flags |= DM_DEFERRED_REMOVE; + param->dev = huge_encode_dev(disk_devt(disk)); /* @@ -676,26 +704,27 @@ static void __dev_status(struct mapped_device *md, struct dm_ioctl *param) param->event_nr = dm_get_event_nr(md); param->target_count = 0; - table = dm_get_live_table(md); + table = dm_get_live_table(md, &srcu_idx); if (table) { if (!(param->flags & DM_QUERY_INACTIVE_TABLE_FLAG)) { if (get_disk_ro(disk)) param->flags |= DM_READONLY_FLAG; param->target_count = dm_table_get_num_targets(table); } - dm_table_put(table); param->flags |= DM_ACTIVE_PRESENT_FLAG; } + dm_put_live_table(md, srcu_idx); if (param->flags & DM_QUERY_INACTIVE_TABLE_FLAG) { - table = dm_get_inactive_table(md); + int srcu_idx; + table = dm_get_inactive_table(md, &srcu_idx); if (table) { if (!(dm_table_get_mode(table) & FMODE_WRITE)) param->flags |= DM_READONLY_FLAG; param->target_count = dm_table_get_num_targets(table); - dm_table_put(table); } + dm_put_live_table(md, srcu_idx); } } @@ -796,6 +825,7 @@ static int dev_remove(struct dm_ioctl *param, size_t param_size) struct hash_cell *hc; struct mapped_device *md; int r; + struct dm_table *t; down_write(&_hash_lock); hc = __find_device_hash_cell(param); @@ -811,17 +841,29 @@ static int dev_remove(struct dm_ioctl *param, size_t param_size) /* * Ensure the device is not open and nothing further can open it. */ - r = dm_lock_for_deletion(md); + r = dm_lock_for_deletion(md, !!(param->flags & DM_DEFERRED_REMOVE), false); if (r) { + if (r == -EBUSY && param->flags & DM_DEFERRED_REMOVE) { + up_write(&_hash_lock); + dm_put(md); + return 0; + } DMDEBUG_LIMIT("unable to remove open device %s", hc->name); up_write(&_hash_lock); dm_put(md); return r; } - __hash_remove(hc); + t = __hash_remove(hc); up_write(&_hash_lock); + if (t) { + dm_sync_table(md); + dm_table_destroy(t); + } + + param->flags &= ~DM_DEFERRED_REMOVE; + if (!dm_kobject_uevent(md, KOBJ_REMOVE, param->event_nr)) param->flags |= DM_UEVENT_GENERATED_FLAG; @@ -851,7 +893,7 @@ static int dev_rename(struct dm_ioctl *param, size_t param_size) unsigned change_uuid = (param->flags & DM_UUID_FLAG) ? 1 : 0; if (new_data < param->data || - invalid_str(new_data, (void *) param + param_size) || + invalid_str(new_data, (void *) param + param_size) || !*new_data || strlen(new_data) > (change_uuid ? DM_UUID_LEN - 1 : DM_NAME_LEN - 1)) { DMWARN("Invalid new mapped device name or uuid string supplied."); return -EINVAL; @@ -986,6 +1028,7 @@ static int do_resume(struct dm_ioctl *param) old_map = dm_swap_table(md, new_map); if (IS_ERR(old_map)) { + dm_sync_table(md); dm_table_destroy(new_map); dm_put(md); return PTR_ERR(old_map); @@ -1003,6 +1046,10 @@ static int do_resume(struct dm_ioctl *param) param->flags |= DM_UEVENT_GENERATED_FLAG; } + /* + * Since dm_swap_table synchronizes RCU, nobody should be in + * read-side critical section already. + */ if (old_map) dm_table_destroy(old_map); @@ -1125,6 +1172,7 @@ static int dev_wait(struct dm_ioctl *param, size_t param_size) int r = 0; struct mapped_device *md; struct dm_table *table; + int srcu_idx; md = find_device(param); if (!md) @@ -1145,11 +1193,10 @@ static int dev_wait(struct dm_ioctl *param, size_t param_size) */ __dev_status(md, param); - table = dm_get_live_or_inactive_table(md, param); - if (table) { + table = dm_get_live_or_inactive_table(md, param, &srcu_idx); + if (table) retrieve_status(table, param, param_size); - dm_table_put(table); - } + dm_put_live_table(md, srcu_idx); out: dm_put(md); @@ -1221,7 +1268,7 @@ static int table_load(struct dm_ioctl *param, size_t param_size) { int r; struct hash_cell *hc; - struct dm_table *t; + struct dm_table *t, *old_map = NULL; struct mapped_device *md; struct target_type *immutable_target_type; @@ -1231,44 +1278,37 @@ static int table_load(struct dm_ioctl *param, size_t param_size) r = dm_table_create(&t, get_mode(param), param->target_count, md); if (r) - goto out; + goto err; + /* Protect md->type and md->queue against concurrent table loads. */ + dm_lock_md_type(md); r = populate_table(t, param, param_size); - if (r) { - dm_table_destroy(t); - goto out; - } + if (r) + goto err_unlock_md_type; immutable_target_type = dm_get_immutable_target_type(md); if (immutable_target_type && (immutable_target_type != dm_table_get_immutable_target_type(t))) { DMWARN("can't replace immutable target type %s", immutable_target_type->name); - dm_table_destroy(t); r = -EINVAL; - goto out; + goto err_unlock_md_type; } - /* Protect md->type and md->queue against concurrent table loads. */ - dm_lock_md_type(md); if (dm_get_md_type(md) == DM_TYPE_NONE) /* Initial table load: acquire type of table. */ dm_set_md_type(md, dm_table_get_type(t)); else if (dm_get_md_type(md) != dm_table_get_type(t)) { DMWARN("can't change device type after initial table load."); - dm_table_destroy(t); - dm_unlock_md_type(md); r = -EINVAL; - goto out; + goto err_unlock_md_type; } /* setup md->queue to reflect md's type (may block) */ r = dm_setup_md_queue(md); if (r) { DMWARN("unable to set up device queue for new table."); - dm_table_destroy(t); - dm_unlock_md_type(md); - goto out; + goto err_unlock_md_type; } dm_unlock_md_type(md); @@ -1277,21 +1317,33 @@ static int table_load(struct dm_ioctl *param, size_t param_size) hc = dm_get_mdptr(md); if (!hc || hc->md != md) { DMWARN("device has been removed from the dev hash table."); - dm_table_destroy(t); up_write(&_hash_lock); r = -ENXIO; - goto out; + goto err_destroy_table; } if (hc->new_map) - dm_table_destroy(hc->new_map); + old_map = hc->new_map; hc->new_map = t; up_write(&_hash_lock); param->flags |= DM_INACTIVE_PRESENT_FLAG; __dev_status(md, param); -out: + if (old_map) { + dm_sync_table(md); + dm_table_destroy(old_map); + } + + dm_put(md); + + return 0; + +err_unlock_md_type: + dm_unlock_md_type(md); +err_destroy_table: + dm_table_destroy(t); +err: dm_put(md); return r; @@ -1301,6 +1353,7 @@ static int table_clear(struct dm_ioctl *param, size_t param_size) { struct hash_cell *hc; struct mapped_device *md; + struct dm_table *old_map = NULL; down_write(&_hash_lock); @@ -1312,7 +1365,7 @@ static int table_clear(struct dm_ioctl *param, size_t param_size) } if (hc->new_map) { - dm_table_destroy(hc->new_map); + old_map = hc->new_map; hc->new_map = NULL; } @@ -1321,6 +1374,10 @@ static int table_clear(struct dm_ioctl *param, size_t param_size) __dev_status(hc->md, param); md = hc->md; up_write(&_hash_lock); + if (old_map) { + dm_sync_table(md); + dm_table_destroy(old_map); + } dm_put(md); return 0; @@ -1370,6 +1427,7 @@ static int table_deps(struct dm_ioctl *param, size_t param_size) { struct mapped_device *md; struct dm_table *table; + int srcu_idx; md = find_device(param); if (!md) @@ -1377,11 +1435,10 @@ static int table_deps(struct dm_ioctl *param, size_t param_size) __dev_status(md, param); - table = dm_get_live_or_inactive_table(md, param); - if (table) { + table = dm_get_live_or_inactive_table(md, param, &srcu_idx); + if (table) retrieve_deps(table, param, param_size); - dm_table_put(table); - } + dm_put_live_table(md, srcu_idx); dm_put(md); @@ -1396,6 +1453,7 @@ static int table_status(struct dm_ioctl *param, size_t param_size) { struct mapped_device *md; struct dm_table *table; + int srcu_idx; md = find_device(param); if (!md) @@ -1403,31 +1461,44 @@ static int table_status(struct dm_ioctl *param, size_t param_size) __dev_status(md, param); - table = dm_get_live_or_inactive_table(md, param); - if (table) { + table = dm_get_live_or_inactive_table(md, param, &srcu_idx); + if (table) retrieve_status(table, param, param_size); - dm_table_put(table); - } + dm_put_live_table(md, srcu_idx); dm_put(md); return 0; } -static bool buffer_test_overflow(char *result, unsigned maxlen) -{ - return !maxlen || strlen(result) + 1 >= maxlen; -} - /* - * Process device-mapper dependent messages. + * Process device-mapper dependent messages. Messages prefixed with '@' + * are processed by the DM core. All others are delivered to the target. * Returns a number <= 1 if message was processed by device mapper. * Returns 2 if message should be delivered to the target. */ static int message_for_md(struct mapped_device *md, unsigned argc, char **argv, char *result, unsigned maxlen) { - return 2; + int r; + + if (**argv != '@') + return 2; /* no '@' prefix, deliver to target */ + + if (!strcasecmp(argv[0], "@cancel_deferred_remove")) { + if (argc != 1) { + DMERR("Invalid arguments for @cancel_deferred_remove"); + return -EINVAL; + } + return dm_cancel_deferred_remove(md); + } + + r = dm_stats_message(md, argc, argv, result, maxlen); + if (r < 2) + return r; + + DMERR("Unsupported message sent to DM core: %s", argv[0]); + return -EINVAL; } /* @@ -1443,6 +1514,7 @@ static int target_message(struct dm_ioctl *param, size_t param_size) struct dm_target_msg *tmsg = (void *) param + param->data_start; size_t maxlen; char *result = get_result_buffer(param, param_size, &maxlen); + int srcu_idx; md = find_device(param); if (!md) @@ -1470,9 +1542,9 @@ static int target_message(struct dm_ioctl *param, size_t param_size) if (r <= 1) goto out_argv; - table = dm_get_live_table(md); + table = dm_get_live_table(md, &srcu_idx); if (!table) - goto out_argv; + goto out_table; if (dm_deleting_md(md)) { r = -ENXIO; @@ -1491,7 +1563,7 @@ static int target_message(struct dm_ioctl *param, size_t param_size) } out_table: - dm_table_put(table); + dm_put_live_table(md, srcu_idx); out_argv: kfree(argv); out: @@ -1500,7 +1572,7 @@ static int target_message(struct dm_ioctl *param, size_t param_size) if (r == 1) { param->flags |= DM_DATA_OUT_FLAG; - if (buffer_test_overflow(result, maxlen)) + if (dm_message_test_buffer_overflow(result, maxlen)) param->flags |= DM_BUFFER_FULL_FLAG; else param->data_size = param->data_start + strlen(result) + 1; @@ -1644,7 +1716,10 @@ static int copy_params(struct dm_ioctl __user *user, struct dm_ioctl *param_kern } if (!dmi) { - dmi = __vmalloc(param_kernel->data_size, GFP_NOIO | __GFP_REPEAT | __GFP_HIGH, PAGE_KERNEL); + unsigned noio_flag; + noio_flag = memalloc_noio_save(); + dmi = __vmalloc(param_kernel->data_size, GFP_NOIO | __GFP_REPEAT | __GFP_HIGH | __GFP_HIGHMEM, PAGE_KERNEL); + memalloc_noio_restore(noio_flag); if (dmi) *param_flags |= DM_PARAMS_VMALLOC; } diff --git a/drivers/md/dm-kcopyd.c b/drivers/md/dm-kcopyd.c index d581fe5d2fa..3a7cade5e27 100644 --- a/drivers/md/dm-kcopyd.c +++ b/drivers/md/dm-kcopyd.c @@ -833,8 +833,7 @@ struct dm_kcopyd_client *dm_kcopyd_client_create(struct dm_kcopyd_throttle *thro goto bad_slab; INIT_WORK(&kc->kcopyd_work, do_work); - kc->kcopyd_wq = alloc_workqueue("kcopyd", - WQ_NON_REENTRANT | WQ_MEM_RECLAIM, 0); + kc->kcopyd_wq = alloc_workqueue("kcopyd", WQ_MEM_RECLAIM, 0); if (!kc->kcopyd_wq) goto bad_workqueue; diff --git a/drivers/md/dm-linear.c b/drivers/md/dm-linear.c index 4f99d267340..53e848c1093 100644 --- a/drivers/md/dm-linear.c +++ b/drivers/md/dm-linear.c @@ -85,7 +85,8 @@ static void linear_map_bio(struct dm_target *ti, struct bio *bio) bio->bi_bdev = lc->dev->bdev; if (bio_sectors(bio)) - bio->bi_sector = linear_map_sector(ti, bio->bi_sector); + bio->bi_iter.bi_sector = + linear_map_sector(ti, bio->bi_iter.bi_sector); } static int linear_map(struct dm_target *ti, struct bio *bio) diff --git a/drivers/md/dm-log-userspace-base.c b/drivers/md/dm-log-userspace-base.c index 9429159d9ee..b953db6cc22 100644 --- a/drivers/md/dm-log-userspace-base.c +++ b/drivers/md/dm-log-userspace-base.c @@ -10,10 +10,11 @@ #include <linux/device-mapper.h> #include <linux/dm-log-userspace.h> #include <linux/module.h> +#include <linux/workqueue.h> #include "dm-log-userspace-transfer.h" -#define DM_LOG_USERSPACE_VSN "1.1.0" +#define DM_LOG_USERSPACE_VSN "1.3.0" struct flush_entry { int type; @@ -58,6 +59,18 @@ struct log_c { spinlock_t flush_lock; struct list_head mark_list; struct list_head clear_list; + + /* + * Workqueue for flush of clear region requests. + */ + struct workqueue_struct *dmlog_wq; + struct delayed_work flush_log_work; + atomic_t sched_flush; + + /* + * Combine userspace flush and mark requests for efficiency. + */ + uint32_t integrated_flush; }; static mempool_t *flush_entry_pool; @@ -122,6 +135,9 @@ static int build_constructor_string(struct dm_target *ti, *ctr_str = NULL; + /* + * Determine overall size of the string. + */ for (i = 0, str_size = 0; i < argc; i++) str_size += strlen(argv[i]) + 1; /* +1 for space between args */ @@ -141,18 +157,39 @@ static int build_constructor_string(struct dm_target *ti, return str_size; } +static void do_flush(struct work_struct *work) +{ + int r; + struct log_c *lc = container_of(work, struct log_c, flush_log_work.work); + + atomic_set(&lc->sched_flush, 0); + + r = userspace_do_request(lc, lc->uuid, DM_ULOG_FLUSH, NULL, 0, NULL, NULL); + + if (r) + dm_table_event(lc->ti->table); +} + /* * userspace_ctr * * argv contains: - * <UUID> <other args> - * Where 'other args' is the userspace implementation specific log - * arguments. An example might be: - * <UUID> clustered-disk <arg count> <log dev> <region_size> [[no]sync] + * <UUID> [integrated_flush] <other args> + * Where 'other args' are the userspace implementation-specific log + * arguments. + * + * Example: + * <UUID> [integrated_flush] clustered-disk <arg count> <log dev> + * <region_size> [[no]sync] + * + * This module strips off the <UUID> and uses it for identification + * purposes when communicating with userspace about a log. * - * So, this module will strip off the <UUID> for identification purposes - * when communicating with userspace about a log; but will pass on everything - * else. + * If integrated_flush is defined, the kernel combines flush + * and mark requests. + * + * The rest of the line, beginning with 'clustered-disk', is passed + * to the userspace ctr function. */ static int userspace_ctr(struct dm_dirty_log *log, struct dm_target *ti, unsigned argc, char **argv) @@ -188,12 +225,22 @@ static int userspace_ctr(struct dm_dirty_log *log, struct dm_target *ti, return -EINVAL; } + lc->usr_argc = argc; + strncpy(lc->uuid, argv[0], DM_UUID_LEN); + argc--; + argv++; spin_lock_init(&lc->flush_lock); INIT_LIST_HEAD(&lc->mark_list); INIT_LIST_HEAD(&lc->clear_list); - str_size = build_constructor_string(ti, argc - 1, argv + 1, &ctr_str); + if (!strcasecmp(argv[0], "integrated_flush")) { + lc->integrated_flush = 1; + argc--; + argv++; + } + + str_size = build_constructor_string(ti, argc, argv, &ctr_str); if (str_size < 0) { kfree(lc); return str_size; @@ -246,6 +293,19 @@ static int userspace_ctr(struct dm_dirty_log *log, struct dm_target *ti, DMERR("Failed to register %s with device-mapper", devices_rdata); } + + if (lc->integrated_flush) { + lc->dmlog_wq = alloc_workqueue("dmlogd", WQ_MEM_RECLAIM, 0); + if (!lc->dmlog_wq) { + DMERR("couldn't start dmlogd"); + r = -ENOMEM; + goto out; + } + + INIT_DELAYED_WORK(&lc->flush_log_work, do_flush); + atomic_set(&lc->sched_flush, 0); + } + out: kfree(devices_rdata); if (r) { @@ -253,7 +313,6 @@ out: kfree(ctr_str); } else { lc->usr_argv_str = ctr_str; - lc->usr_argc = argc; log->context = lc; } @@ -264,9 +323,16 @@ static void userspace_dtr(struct dm_dirty_log *log) { struct log_c *lc = log->context; + if (lc->integrated_flush) { + /* flush workqueue */ + if (atomic_read(&lc->sched_flush)) + flush_delayed_work(&lc->flush_log_work); + + destroy_workqueue(lc->dmlog_wq); + } + (void) dm_consult_userspace(lc->uuid, lc->luid, DM_ULOG_DTR, - NULL, 0, - NULL, NULL); + NULL, 0, NULL, NULL); if (lc->log_dev) dm_put_device(lc->ti, lc->log_dev); @@ -283,8 +349,7 @@ static int userspace_presuspend(struct dm_dirty_log *log) struct log_c *lc = log->context; r = dm_consult_userspace(lc->uuid, lc->luid, DM_ULOG_PRESUSPEND, - NULL, 0, - NULL, NULL); + NULL, 0, NULL, NULL); return r; } @@ -294,9 +359,14 @@ static int userspace_postsuspend(struct dm_dirty_log *log) int r; struct log_c *lc = log->context; + /* + * Run planned flush earlier. + */ + if (lc->integrated_flush && atomic_read(&lc->sched_flush)) + flush_delayed_work(&lc->flush_log_work); + r = dm_consult_userspace(lc->uuid, lc->luid, DM_ULOG_POSTSUSPEND, - NULL, 0, - NULL, NULL); + NULL, 0, NULL, NULL); return r; } @@ -308,8 +378,7 @@ static int userspace_resume(struct dm_dirty_log *log) lc->in_sync_hint = 0; r = dm_consult_userspace(lc->uuid, lc->luid, DM_ULOG_RESUME, - NULL, 0, - NULL, NULL); + NULL, 0, NULL, NULL); return r; } @@ -405,7 +474,8 @@ static int flush_one_by_one(struct log_c *lc, struct list_head *flush_list) return r; } -static int flush_by_group(struct log_c *lc, struct list_head *flush_list) +static int flush_by_group(struct log_c *lc, struct list_head *flush_list, + int flush_with_payload) { int r = 0; int count; @@ -431,15 +501,29 @@ static int flush_by_group(struct log_c *lc, struct list_head *flush_list) break; } - r = userspace_do_request(lc, lc->uuid, type, - (char *)(group), - count * sizeof(uint64_t), - NULL, NULL); - if (r) { - /* Group send failed. Attempt one-by-one. */ - list_splice_init(&tmp_list, flush_list); - r = flush_one_by_one(lc, flush_list); - break; + if (flush_with_payload) { + r = userspace_do_request(lc, lc->uuid, DM_ULOG_FLUSH, + (char *)(group), + count * sizeof(uint64_t), + NULL, NULL); + /* + * Integrated flush failed. + */ + if (r) + break; + } else { + r = userspace_do_request(lc, lc->uuid, type, + (char *)(group), + count * sizeof(uint64_t), + NULL, NULL); + if (r) { + /* + * Group send failed. Attempt one-by-one. + */ + list_splice_init(&tmp_list, flush_list); + r = flush_one_by_one(lc, flush_list); + break; + } } } @@ -476,6 +560,8 @@ static int userspace_flush(struct dm_dirty_log *log) struct log_c *lc = log->context; LIST_HEAD(mark_list); LIST_HEAD(clear_list); + int mark_list_is_empty; + int clear_list_is_empty; struct flush_entry *fe, *tmp_fe; spin_lock_irqsave(&lc->flush_lock, flags); @@ -483,23 +569,51 @@ static int userspace_flush(struct dm_dirty_log *log) list_splice_init(&lc->clear_list, &clear_list); spin_unlock_irqrestore(&lc->flush_lock, flags); - if (list_empty(&mark_list) && list_empty(&clear_list)) + mark_list_is_empty = list_empty(&mark_list); + clear_list_is_empty = list_empty(&clear_list); + + if (mark_list_is_empty && clear_list_is_empty) return 0; - r = flush_by_group(lc, &mark_list); + r = flush_by_group(lc, &clear_list, 0); if (r) - goto fail; + goto out; - r = flush_by_group(lc, &clear_list); + if (!lc->integrated_flush) { + r = flush_by_group(lc, &mark_list, 0); + if (r) + goto out; + r = userspace_do_request(lc, lc->uuid, DM_ULOG_FLUSH, + NULL, 0, NULL, NULL); + goto out; + } + + /* + * Send integrated flush request with mark_list as payload. + */ + r = flush_by_group(lc, &mark_list, 1); if (r) - goto fail; + goto out; - r = userspace_do_request(lc, lc->uuid, DM_ULOG_FLUSH, - NULL, 0, NULL, NULL); + if (mark_list_is_empty && !atomic_read(&lc->sched_flush)) { + /* + * When there are only clear region requests, + * we schedule a flush in the future. + */ + queue_delayed_work(lc->dmlog_wq, &lc->flush_log_work, 3 * HZ); + atomic_set(&lc->sched_flush, 1); + } else { + /* + * Cancel pending flush because we + * have already flushed in mark_region. + */ + cancel_delayed_work(&lc->flush_log_work); + atomic_set(&lc->sched_flush, 0); + } -fail: +out: /* - * We can safely remove these entries, even if failure. + * We can safely remove these entries, even after failure. * Calling code will receive an error and will know that * the log facility has failed. */ @@ -603,8 +717,7 @@ static int userspace_get_resync_work(struct dm_dirty_log *log, region_t *region) rdata_size = sizeof(pkg); r = userspace_do_request(lc, lc->uuid, DM_ULOG_GET_RESYNC_WORK, - NULL, 0, - (char *)&pkg, &rdata_size); + NULL, 0, (char *)&pkg, &rdata_size); *region = pkg.r; return (r) ? r : (int)pkg.i; @@ -630,8 +743,7 @@ static void userspace_set_region_sync(struct dm_dirty_log *log, pkg.i = (int64_t)in_sync; r = userspace_do_request(lc, lc->uuid, DM_ULOG_SET_REGION_SYNC, - (char *)&pkg, sizeof(pkg), - NULL, NULL); + (char *)&pkg, sizeof(pkg), NULL, NULL); /* * It would be nice to be able to report failures. @@ -657,8 +769,7 @@ static region_t userspace_get_sync_count(struct dm_dirty_log *log) rdata_size = sizeof(sync_count); r = userspace_do_request(lc, lc->uuid, DM_ULOG_GET_SYNC_COUNT, - NULL, 0, - (char *)&sync_count, &rdata_size); + NULL, 0, (char *)&sync_count, &rdata_size); if (r) return 0; @@ -685,8 +796,7 @@ static int userspace_status(struct dm_dirty_log *log, status_type_t status_type, switch (status_type) { case STATUSTYPE_INFO: r = userspace_do_request(lc, lc->uuid, DM_ULOG_STATUS_INFO, - NULL, 0, - result, &sz); + NULL, 0, result, &sz); if (r) { sz = 0; @@ -699,8 +809,10 @@ static int userspace_status(struct dm_dirty_log *log, status_type_t status_type, BUG_ON(!table_args); /* There will always be a ' ' */ table_args++; - DMEMIT("%s %u %s %s ", log->type->name, lc->usr_argc, - lc->uuid, table_args); + DMEMIT("%s %u %s ", log->type->name, lc->usr_argc, lc->uuid); + if (lc->integrated_flush) + DMEMIT("integrated_flush "); + DMEMIT("%s ", table_args); break; } return (r) ? 0 : (int)sz; diff --git a/drivers/md/dm-log-userspace-transfer.c b/drivers/md/dm-log-userspace-transfer.c index 08d9a207259..b428c0ae63d 100644 --- a/drivers/md/dm-log-userspace-transfer.c +++ b/drivers/md/dm-log-userspace-transfer.c @@ -66,7 +66,7 @@ static int dm_ulog_sendto_server(struct dm_ulog_request *tfr) msg->seq = tfr->seq; msg->len = sizeof(struct dm_ulog_request) + tfr->data_size; - r = cn_netlink_send(msg, 0, gfp_any()); + r = cn_netlink_send(msg, 0, 0, gfp_any()); return r; } diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c index bdf26f5bd32..f4167b013d9 100644 --- a/drivers/md/dm-mpath.c +++ b/drivers/md/dm-mpath.c @@ -7,6 +7,7 @@ #include <linux/device-mapper.h> +#include "dm.h" #include "dm-path-selector.h" #include "dm-uevent.h" @@ -86,15 +87,12 @@ struct multipath { unsigned queue_if_no_path:1; /* Queue I/O if last path fails? */ unsigned saved_queue_if_no_path:1; /* Saved state during suspension */ unsigned retain_attached_hw_handler:1; /* If there's already a hw_handler present, don't change it. */ + unsigned pg_init_disabled:1; /* pg_init is not currently allowed */ unsigned pg_init_retries; /* Number of times to retry pg_init */ unsigned pg_init_count; /* Number of times pg_init called */ unsigned pg_init_delay_msecs; /* Number of msecs before pg_init retry */ - unsigned queue_size; - struct work_struct process_queued_ios; - struct list_head queued_ios; - struct work_struct trigger_event; /* @@ -116,14 +114,12 @@ struct dm_mpath_io { typedef int (*action_fn) (struct pgpath *pgpath); -#define MIN_IOS 256 /* Mempool size */ - static struct kmem_cache *_mpio_cache; static struct workqueue_struct *kmultipathd, *kmpath_handlerd; -static void process_queued_ios(struct work_struct *work); static void trigger_event(struct work_struct *work); static void activate_path(struct work_struct *work); +static int __pgpath_busy(struct pgpath *pgpath); /*----------------------------------------------- @@ -190,19 +186,18 @@ static void free_priority_group(struct priority_group *pg, static struct multipath *alloc_multipath(struct dm_target *ti) { struct multipath *m; + unsigned min_ios = dm_get_reserved_rq_based_ios(); m = kzalloc(sizeof(*m), GFP_KERNEL); if (m) { INIT_LIST_HEAD(&m->priority_groups); - INIT_LIST_HEAD(&m->queued_ios); spin_lock_init(&m->lock); m->queue_io = 1; m->pg_init_delay_msecs = DM_PG_INIT_DELAY_DEFAULT; - INIT_WORK(&m->process_queued_ios, process_queued_ios); INIT_WORK(&m->trigger_event, trigger_event); init_waitqueue_head(&m->pg_init_wait); mutex_init(&m->work_mutex); - m->mpio_pool = mempool_create_slab_pool(MIN_IOS, _mpio_cache); + m->mpio_pool = mempool_create_slab_pool(min_ios, _mpio_cache); if (!m->mpio_pool) { kfree(m); return NULL; @@ -255,13 +250,21 @@ static void clear_mapinfo(struct multipath *m, union map_info *info) * Path selection *-----------------------------------------------*/ -static void __pg_init_all_paths(struct multipath *m) +static int __pg_init_all_paths(struct multipath *m) { struct pgpath *pgpath; unsigned long pg_init_delay = 0; + if (m->pg_init_in_progress || m->pg_init_disabled) + return 0; + m->pg_init_count++; m->pg_init_required = 0; + + /* Check here to reset pg_init_required */ + if (!m->current_pg) + return 0; + if (m->pg_init_delay_retry) pg_init_delay = msecs_to_jiffies(m->pg_init_delay_msecs != DM_PG_INIT_DELAY_DEFAULT ? m->pg_init_delay_msecs : DM_PG_INIT_DELAY_MSECS); @@ -273,6 +276,7 @@ static void __pg_init_all_paths(struct multipath *m) pg_init_delay)) m->pg_init_in_progress++; } + return m->pg_init_in_progress; } static void __switch_pg(struct multipath *m, struct pgpath *pgpath) @@ -364,19 +368,26 @@ failed: */ static int __must_push_back(struct multipath *m) { - return (m->queue_if_no_path != m->saved_queue_if_no_path && - dm_noflush_suspending(m->ti)); + return (m->queue_if_no_path || + (m->queue_if_no_path != m->saved_queue_if_no_path && + dm_noflush_suspending(m->ti))); } -static int map_io(struct multipath *m, struct request *clone, - union map_info *map_context, unsigned was_queued) +#define pg_ready(m) (!(m)->queue_io && !(m)->pg_init_required) + +/* + * Map cloned requests + */ +static int multipath_map(struct dm_target *ti, struct request *clone, + union map_info *map_context) { - int r = DM_MAPIO_REMAPPED; + struct multipath *m = (struct multipath *) ti->private; + int r = DM_MAPIO_REQUEUE; size_t nr_bytes = blk_rq_bytes(clone); unsigned long flags; struct pgpath *pgpath; struct block_device *bdev; - struct dm_mpath_io *mpio = map_context->ptr; + struct dm_mpath_io *mpio; spin_lock_irqsave(&m->lock, flags); @@ -387,35 +398,33 @@ static int map_io(struct multipath *m, struct request *clone, pgpath = m->current_pgpath; - if (was_queued) - m->queue_size--; - - if ((pgpath && m->queue_io) || - (!pgpath && m->queue_if_no_path)) { - /* Queue for the daemon to resubmit */ - list_add_tail(&clone->queuelist, &m->queued_ios); - m->queue_size++; - if ((m->pg_init_required && !m->pg_init_in_progress) || - !m->queue_io) - queue_work(kmultipathd, &m->process_queued_ios); - pgpath = NULL; - r = DM_MAPIO_SUBMITTED; - } else if (pgpath) { - bdev = pgpath->path.dev->bdev; - clone->q = bdev_get_queue(bdev); - clone->rq_disk = bdev->bd_disk; - } else if (__must_push_back(m)) - r = DM_MAPIO_REQUEUE; - else - r = -EIO; /* Failed */ + if (!pgpath) { + if (!__must_push_back(m)) + r = -EIO; /* Failed */ + goto out_unlock; + } + if (!pg_ready(m)) { + __pg_init_all_paths(m); + goto out_unlock; + } + if (set_mapinfo(m, map_context) < 0) + /* ENOMEM, requeue */ + goto out_unlock; + bdev = pgpath->path.dev->bdev; + clone->q = bdev_get_queue(bdev); + clone->rq_disk = bdev->bd_disk; + clone->cmd_flags |= REQ_FAILFAST_TRANSPORT; + mpio = map_context->ptr; mpio->pgpath = pgpath; mpio->nr_bytes = nr_bytes; - - if (r == DM_MAPIO_REMAPPED && pgpath->pg->ps.type->start_io) - pgpath->pg->ps.type->start_io(&pgpath->pg->ps, &pgpath->path, + if (pgpath->pg->ps.type->start_io) + pgpath->pg->ps.type->start_io(&pgpath->pg->ps, + &pgpath->path, nr_bytes); + r = DM_MAPIO_REMAPPED; +out_unlock: spin_unlock_irqrestore(&m->lock, flags); return r; @@ -436,73 +445,12 @@ static int queue_if_no_path(struct multipath *m, unsigned queue_if_no_path, else m->saved_queue_if_no_path = queue_if_no_path; m->queue_if_no_path = queue_if_no_path; - if (!m->queue_if_no_path && m->queue_size) - queue_work(kmultipathd, &m->process_queued_ios); - spin_unlock_irqrestore(&m->lock, flags); - return 0; -} - -/*----------------------------------------------------------------- - * The multipath daemon is responsible for resubmitting queued ios. - *---------------------------------------------------------------*/ - -static void dispatch_queued_ios(struct multipath *m) -{ - int r; - unsigned long flags; - union map_info *info; - struct request *clone, *n; - LIST_HEAD(cl); - - spin_lock_irqsave(&m->lock, flags); - list_splice_init(&m->queued_ios, &cl); - spin_unlock_irqrestore(&m->lock, flags); - - list_for_each_entry_safe(clone, n, &cl, queuelist) { - list_del_init(&clone->queuelist); - - info = dm_get_rq_mapinfo(clone); - - r = map_io(m, clone, info, 1); - if (r < 0) { - clear_mapinfo(m, info); - dm_kill_unmapped_request(clone, r); - } else if (r == DM_MAPIO_REMAPPED) - dm_dispatch_request(clone); - else if (r == DM_MAPIO_REQUEUE) { - clear_mapinfo(m, info); - dm_requeue_unmapped_request(clone); - } - } -} - -static void process_queued_ios(struct work_struct *work) -{ - struct multipath *m = - container_of(work, struct multipath, process_queued_ios); - struct pgpath *pgpath = NULL; - unsigned must_queue = 1; - unsigned long flags; - - spin_lock_irqsave(&m->lock, flags); - - if (!m->current_pgpath) - __choose_pgpath(m, 0); - - pgpath = m->current_pgpath; - - if ((pgpath && !m->queue_io) || - (!pgpath && !m->queue_if_no_path)) - must_queue = 0; - - if (m->pg_init_required && !m->pg_init_in_progress && pgpath) - __pg_init_all_paths(m); + if (!queue_if_no_path) + dm_table_run_md_queue_async(m->ti->table); - spin_unlock_irqrestore(&m->lock, flags); - if (!must_queue) - dispatch_queued_ios(m); + return 0; } /* @@ -942,10 +890,20 @@ static void multipath_wait_for_pg_init_completion(struct multipath *m) static void flush_multipath_work(struct multipath *m) { + unsigned long flags; + + spin_lock_irqsave(&m->lock, flags); + m->pg_init_disabled = 1; + spin_unlock_irqrestore(&m->lock, flags); + flush_workqueue(kmpath_handlerd); multipath_wait_for_pg_init_completion(m); flush_workqueue(kmultipathd); flush_work(&m->trigger_event); + + spin_lock_irqsave(&m->lock, flags); + m->pg_init_disabled = 0; + spin_unlock_irqrestore(&m->lock, flags); } static void multipath_dtr(struct dm_target *ti) @@ -957,27 +915,6 @@ static void multipath_dtr(struct dm_target *ti) } /* - * Map cloned requests - */ -static int multipath_map(struct dm_target *ti, struct request *clone, - union map_info *map_context) -{ - int r; - struct multipath *m = (struct multipath *) ti->private; - - if (set_mapinfo(m, map_context) < 0) - /* ENOMEM, requeue */ - return DM_MAPIO_REQUEUE; - - clone->cmd_flags |= REQ_FAILFAST_TRANSPORT; - r = map_io(m, clone, map_context, 0); - if (r < 0 || r == DM_MAPIO_REQUEUE) - clear_mapinfo(m, map_context); - - return r; -} - -/* * Take a path out of use. */ static int fail_path(struct pgpath *pgpath) @@ -1017,7 +954,7 @@ out: */ static int reinstate_path(struct pgpath *pgpath) { - int r = 0; + int r = 0, run_queue = 0; unsigned long flags; struct multipath *m = pgpath->pg->m; @@ -1039,9 +976,9 @@ static int reinstate_path(struct pgpath *pgpath) pgpath->is_active = 1; - if (!m->nr_valid_paths++ && m->queue_size) { + if (!m->nr_valid_paths++) { m->current_pgpath = NULL; - queue_work(kmultipathd, &m->process_queued_ios); + run_queue = 1; } else if (m->hw_handler_name && (m->current_pg == pgpath->pg)) { if (queue_work(kmpath_handlerd, &pgpath->activate_path.work)) m->pg_init_in_progress++; @@ -1054,6 +991,8 @@ static int reinstate_path(struct pgpath *pgpath) out: spin_unlock_irqrestore(&m->lock, flags); + if (run_queue) + dm_table_run_md_queue_async(m->ti->table); return r; } @@ -1164,7 +1103,7 @@ static int pg_init_limit_reached(struct multipath *m, struct pgpath *pgpath) spin_lock_irqsave(&m->lock, flags); - if (m->pg_init_count <= m->pg_init_retries) + if (m->pg_init_count <= m->pg_init_retries && !m->pg_init_disabled) m->pg_init_required = 1; else limit_reached = 1; @@ -1237,11 +1176,12 @@ static void pg_init_done(void *data, int errors) /* Activations of other paths are still on going */ goto out; - if (!m->pg_init_required) - m->queue_io = 0; - - m->pg_init_delay_retry = delay_retry; - queue_work(kmultipathd, &m->process_queued_ios); + if (m->pg_init_required) { + m->pg_init_delay_retry = delay_retry; + if (__pg_init_all_paths(m)) + goto out; + } + m->queue_io = 0; /* * Wake up any thread waiting to suspend. @@ -1257,8 +1197,26 @@ static void activate_path(struct work_struct *work) struct pgpath *pgpath = container_of(work, struct pgpath, activate_path.work); - scsi_dh_activate(bdev_get_queue(pgpath->path.dev->bdev), - pg_init_done, pgpath); + if (pgpath->is_active) + scsi_dh_activate(bdev_get_queue(pgpath->path.dev->bdev), + pg_init_done, pgpath); + else + pg_init_done(pgpath, SCSI_DH_DEV_OFFLINED); +} + +static int noretry_error(int error) +{ + switch (error) { + case -EOPNOTSUPP: + case -EREMOTEIO: + case -EILSEQ: + case -ENODATA: + case -ENOSPC: + return 1; + } + + /* Anything else could be a path failure, so should be retried */ + return 0; } /* @@ -1284,7 +1242,7 @@ static int do_end_io(struct multipath *m, struct request *clone, if (!error && !clone->errors) return 0; /* I/O complete */ - if (error == -EOPNOTSUPP || error == -EREMOTEIO || error == -EILSEQ) + if (noretry_error(error)) return error; if (mpio->pgpath) @@ -1394,7 +1352,7 @@ static void multipath_status(struct dm_target *ti, status_type_t type, /* Features */ if (type == STATUSTYPE_INFO) - DMEMIT("2 %u %u ", m->queue_size, m->pg_init_count); + DMEMIT("2 %u %u ", m->queue_io, m->pg_init_count); else { DMEMIT("%u ", m->queue_if_no_path + (m->pg_init_retries > 0) * 2 + @@ -1513,7 +1471,7 @@ static int multipath_message(struct dm_target *ti, unsigned argc, char **argv) } if (argc != 2) { - DMWARN("Unrecognised multipath message received."); + DMWARN("Invalid multipath message arguments. Expected 2 arguments, got %d.", argc); goto out; } @@ -1531,7 +1489,7 @@ static int multipath_message(struct dm_target *ti, unsigned argc, char **argv) else if (!strcasecmp(argv[0], "fail_path")) action = fail_path; else { - DMWARN("Unrecognised multipath message received."); + DMWARN("Unrecognised multipath message received: %s", argv[0]); goto out; } @@ -1561,7 +1519,6 @@ static int multipath_ioctl(struct dm_target *ti, unsigned int cmd, unsigned long flags; int r; -again: bdev = NULL; mode = 0; r = 0; @@ -1579,7 +1536,7 @@ again: } if ((pgpath && m->queue_io) || (!pgpath && m->queue_if_no_path)) - r = -EAGAIN; + r = -ENOTCONN; else if (!bdev) r = -EIO; @@ -1588,13 +1545,22 @@ again: /* * Only pass ioctls through if the device sizes match exactly. */ - if (!r && ti->len != i_size_read(bdev->bd_inode) >> SECTOR_SHIFT) - r = scsi_verify_blk_ioctl(NULL, cmd); + if (!bdev || ti->len != i_size_read(bdev->bd_inode) >> SECTOR_SHIFT) { + int err = scsi_verify_blk_ioctl(NULL, cmd); + if (err) + r = err; + } - if (r == -EAGAIN && !fatal_signal_pending(current)) { - queue_work(kmultipathd, &m->process_queued_ios); - msleep(10); - goto again; + if (r == -ENOTCONN && !fatal_signal_pending(current)) { + spin_lock_irqsave(&m->lock, flags); + if (!m->current_pg) { + /* Path status changed, redo selection */ + __choose_pgpath(m, 0); + } + if (m->pg_init_required) + __pg_init_all_paths(m); + spin_unlock_irqrestore(&m->lock, flags); + dm_table_run_md_queue_async(m->ti->table); } return r ? : __blkdev_driver_ioctl(bdev, mode, cmd, arg); @@ -1645,6 +1611,12 @@ static int multipath_busy(struct dm_target *ti) spin_lock_irqsave(&m->lock, flags); + /* pg_init in progress or no paths available */ + if (m->pg_init_in_progress || + (!m->nr_valid_paths && m->queue_if_no_path)) { + busy = 1; + goto out; + } /* Guess which priority_group will be used at next mapping time */ if (unlikely(!m->current_pgpath && m->next_pg)) pg = m->next_pg; @@ -1694,7 +1666,7 @@ out: *---------------------------------------------------------------*/ static struct target_type multipath_target = { .name = "multipath", - .version = {1, 5, 1}, + .version = {1, 7, 0}, .module = THIS_MODULE, .ctr = multipath_ctr, .dtr = multipath_dtr, diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c index 1d3fe1a40a9..4880b69e2e9 100644 --- a/drivers/md/dm-raid.c +++ b/drivers/md/dm-raid.c @@ -380,7 +380,7 @@ static int validate_region_size(struct raid_set *rs, unsigned long region_size) static int validate_raid_redundancy(struct raid_set *rs) { unsigned i, rebuild_cnt = 0; - unsigned rebuilds_per_group, copies, d; + unsigned rebuilds_per_group = 0, copies, d; unsigned group_size, last_group_start; for (i = 0; i < rs->md.raid_disks; i++) @@ -504,7 +504,7 @@ static int parse_raid_params(struct raid_set *rs, char **argv, * First, parse the in-order required arguments * "chunk_size" is the only argument of this type. */ - if ((strict_strtoul(argv[0], 10, &value) < 0)) { + if ((kstrtoul(argv[0], 10, &value) < 0)) { rs->ti->error = "Bad chunk size"; return -EINVAL; } else if (rs->raid_type->level == 1) { @@ -585,7 +585,7 @@ static int parse_raid_params(struct raid_set *rs, char **argv, continue; } - if (strict_strtoul(argv[i], 10, &value) < 0) { + if (kstrtoul(argv[i], 10, &value) < 0) { rs->ti->error = "Bad numerical argument given in raid params"; return -EINVAL; } @@ -1181,7 +1181,7 @@ static int raid_ctr(struct dm_target *ti, unsigned argc, char **argv) argv++; /* number of RAID parameters */ - if (strict_strtoul(argv[0], 10, &num_raid_params) < 0) { + if (kstrtoul(argv[0], 10, &num_raid_params) < 0) { ti->error = "Cannot understand number of RAID parameters"; return -EINVAL; } @@ -1194,7 +1194,7 @@ static int raid_ctr(struct dm_target *ti, unsigned argc, char **argv) return -EINVAL; } - if ((strict_strtoul(argv[num_raid_params], 10, &num_raid_devs) < 0) || + if ((kstrtoul(argv[num_raid_params], 10, &num_raid_devs) < 0) || (num_raid_devs >= INT_MAX)) { ti->error = "Cannot understand number of raid devices"; return -EINVAL; @@ -1388,6 +1388,7 @@ static void raid_status(struct dm_target *ti, status_type_t type, * performing a "check" of the array. */ DMEMIT(" %llu", + (strcmp(rs->md.last_sync_action, "check")) ? 0 : (unsigned long long) atomic64_read(&rs->md.resync_mismatches)); break; @@ -1572,6 +1573,62 @@ static void raid_postsuspend(struct dm_target *ti) mddev_suspend(&rs->md); } +static void attempt_restore_of_faulty_devices(struct raid_set *rs) +{ + int i; + uint64_t failed_devices, cleared_failed_devices = 0; + unsigned long flags; + struct dm_raid_superblock *sb; + struct md_rdev *r; + + for (i = 0; i < rs->md.raid_disks; i++) { + r = &rs->dev[i].rdev; + if (test_bit(Faulty, &r->flags) && r->sb_page && + sync_page_io(r, 0, r->sb_size, r->sb_page, READ, 1)) { + DMINFO("Faulty %s device #%d has readable super block." + " Attempting to revive it.", + rs->raid_type->name, i); + + /* + * Faulty bit may be set, but sometimes the array can + * be suspended before the personalities can respond + * by removing the device from the array (i.e. calling + * 'hot_remove_disk'). If they haven't yet removed + * the failed device, its 'raid_disk' number will be + * '>= 0' - meaning we must call this function + * ourselves. + */ + if ((r->raid_disk >= 0) && + (r->mddev->pers->hot_remove_disk(r->mddev, r) != 0)) + /* Failed to revive this device, try next */ + continue; + + r->raid_disk = i; + r->saved_raid_disk = i; + flags = r->flags; + clear_bit(Faulty, &r->flags); + clear_bit(WriteErrorSeen, &r->flags); + clear_bit(In_sync, &r->flags); + if (r->mddev->pers->hot_add_disk(r->mddev, r)) { + r->raid_disk = -1; + r->saved_raid_disk = -1; + r->flags = flags; + } else { + r->recovery_offset = 0; + cleared_failed_devices |= 1 << i; + } + } + } + if (cleared_failed_devices) { + rdev_for_each(r, &rs->md) { + sb = page_address(r->sb_page); + failed_devices = le64_to_cpu(sb->failed_devices); + failed_devices &= ~cleared_failed_devices; + sb->failed_devices = cpu_to_le64(failed_devices); + } + } +} + static void raid_resume(struct dm_target *ti) { struct raid_set *rs = ti->private; @@ -1580,6 +1637,13 @@ static void raid_resume(struct dm_target *ti) if (!rs->bitmap_loaded) { bitmap_load(&rs->md); rs->bitmap_loaded = 1; + } else { + /* + * A secondary resume while the device is active. + * Take this opportunity to check whether any failed + * devices are reachable again. + */ + attempt_restore_of_faulty_devices(rs); } clear_bit(MD_RECOVERY_FROZEN, &rs->md.recovery); @@ -1588,7 +1652,7 @@ static void raid_resume(struct dm_target *ti) static struct target_type raid_target = { .name = "raid", - .version = {1, 5, 0}, + .version = {1, 5, 2}, .module = THIS_MODULE, .ctr = raid_ctr, .dtr = raid_dtr, diff --git a/drivers/md/dm-raid1.c b/drivers/md/dm-raid1.c index 699b5be68d3..7dfdb5c746d 100644 --- a/drivers/md/dm-raid1.c +++ b/drivers/md/dm-raid1.c @@ -432,7 +432,7 @@ static int mirror_available(struct mirror_set *ms, struct bio *bio) region_t region = dm_rh_bio_to_region(ms->rh, bio); if (log->type->in_sync(log, region, 0)) - return choose_mirror(ms, bio->bi_sector) ? 1 : 0; + return choose_mirror(ms, bio->bi_iter.bi_sector) ? 1 : 0; return 0; } @@ -442,15 +442,15 @@ static int mirror_available(struct mirror_set *ms, struct bio *bio) */ static sector_t map_sector(struct mirror *m, struct bio *bio) { - if (unlikely(!bio->bi_size)) + if (unlikely(!bio->bi_iter.bi_size)) return 0; - return m->offset + dm_target_offset(m->ms->ti, bio->bi_sector); + return m->offset + dm_target_offset(m->ms->ti, bio->bi_iter.bi_sector); } static void map_bio(struct mirror *m, struct bio *bio) { bio->bi_bdev = m->dev->bdev; - bio->bi_sector = map_sector(m, bio); + bio->bi_iter.bi_sector = map_sector(m, bio); } static void map_region(struct dm_io_region *io, struct mirror *m, @@ -526,8 +526,8 @@ static void read_async_bio(struct mirror *m, struct bio *bio) struct dm_io_region io; struct dm_io_request io_req = { .bi_rw = READ, - .mem.type = DM_IO_BVEC, - .mem.ptr.bvec = bio->bi_io_vec + bio->bi_idx, + .mem.type = DM_IO_BIO, + .mem.ptr.bio = bio, .notify.fn = read_callback, .notify.context = bio, .client = m->ms->io_client, @@ -559,7 +559,7 @@ static void do_reads(struct mirror_set *ms, struct bio_list *reads) * We can only read balance if the region is in sync. */ if (likely(region_in_sync(ms, region, 1))) - m = choose_mirror(ms, bio->bi_sector); + m = choose_mirror(ms, bio->bi_iter.bi_sector); else if (m && atomic_read(&m->error_count)) m = NULL; @@ -629,8 +629,8 @@ static void do_write(struct mirror_set *ms, struct bio *bio) struct mirror *m; struct dm_io_request io_req = { .bi_rw = WRITE | (bio->bi_rw & WRITE_FLUSH_FUA), - .mem.type = DM_IO_BVEC, - .mem.ptr.bvec = bio->bi_io_vec + bio->bi_idx, + .mem.type = DM_IO_BIO, + .mem.ptr.bio = bio, .notify.fn = write_callback, .notify.context = bio, .client = ms->io_client, @@ -1080,8 +1080,7 @@ static int mirror_ctr(struct dm_target *ti, unsigned int argc, char **argv) ti->per_bio_data_size = sizeof(struct dm_raid1_bio_record); ti->discard_zeroes_data_unsupported = true; - ms->kmirrord_wq = alloc_workqueue("kmirrord", - WQ_NON_REENTRANT | WQ_MEM_RECLAIM, 0); + ms->kmirrord_wq = alloc_workqueue("kmirrord", WQ_MEM_RECLAIM, 0); if (!ms->kmirrord_wq) { DMERR("couldn't start kmirrord"); r = -ENOMEM; @@ -1182,7 +1181,7 @@ static int mirror_map(struct dm_target *ti, struct bio *bio) * The region is in-sync and we can perform reads directly. * Store enough information so we can retry if it fails. */ - m = choose_mirror(ms, bio->bi_sector); + m = choose_mirror(ms, bio->bi_iter.bi_sector); if (unlikely(!m)) return -EIO; @@ -1245,6 +1244,9 @@ static int mirror_end_io(struct dm_target *ti, struct bio *bio, int error) dm_bio_restore(bd, bio); bio_record->details.bi_bdev = NULL; + + atomic_inc(&bio->bi_remaining); + queue_bio(ms, bio, rw); return DM_ENDIO_INCOMPLETE; } diff --git a/drivers/md/dm-region-hash.c b/drivers/md/dm-region-hash.c index 69732e03eb3..b929fd5f498 100644 --- a/drivers/md/dm-region-hash.c +++ b/drivers/md/dm-region-hash.c @@ -126,7 +126,8 @@ EXPORT_SYMBOL_GPL(dm_rh_region_to_sector); region_t dm_rh_bio_to_region(struct dm_region_hash *rh, struct bio *bio) { - return dm_rh_sector_to_region(rh, bio->bi_sector - rh->target_begin); + return dm_rh_sector_to_region(rh, bio->bi_iter.bi_sector - + rh->target_begin); } EXPORT_SYMBOL_GPL(dm_rh_bio_to_region); diff --git a/drivers/md/dm-snap-persistent.c b/drivers/md/dm-snap-persistent.c index 3ac415675b6..d6e88178d22 100644 --- a/drivers/md/dm-snap-persistent.c +++ b/drivers/md/dm-snap-persistent.c @@ -13,10 +13,13 @@ #include <linux/export.h> #include <linux/slab.h> #include <linux/dm-io.h> +#include "dm-bufio.h" #define DM_MSG_PREFIX "persistent snapshot" #define DM_CHUNK_SIZE_DEFAULT_SECTORS 32 /* 16KB */ +#define DM_PREFETCH_CHUNKS 12 + /*----------------------------------------------------------------- * Persistent snapshots, by persistent we mean that the snapshot * will survive a reboot. @@ -256,7 +259,8 @@ static int chunk_io(struct pstore *ps, void *area, chunk_t chunk, int rw, */ INIT_WORK_ONSTACK(&req.work, do_metadata); queue_work(ps->metadata_wq, &req.work); - flush_work(&req.work); + flush_workqueue(ps->metadata_wq); + destroy_work_on_stack(&req.work); return req.result; } @@ -269,6 +273,14 @@ static chunk_t area_location(struct pstore *ps, chunk_t area) return NUM_SNAPSHOT_HDR_CHUNKS + ((ps->exceptions_per_area + 1) * area); } +static void skip_metadata(struct pstore *ps) +{ + uint32_t stride = ps->exceptions_per_area + 1; + chunk_t next_free = ps->next_free; + if (sector_div(next_free, stride) == NUM_SNAPSHOT_HDR_CHUNKS) + ps->next_free++; +} + /* * Read or write a metadata area. Remembering to skip the first * chunk which holds the header. @@ -393,17 +405,18 @@ static int write_header(struct pstore *ps) /* * Access functions for the disk exceptions, these do the endian conversions. */ -static struct disk_exception *get_exception(struct pstore *ps, uint32_t index) +static struct disk_exception *get_exception(struct pstore *ps, void *ps_area, + uint32_t index) { BUG_ON(index >= ps->exceptions_per_area); - return ((struct disk_exception *) ps->area) + index; + return ((struct disk_exception *) ps_area) + index; } -static void read_exception(struct pstore *ps, +static void read_exception(struct pstore *ps, void *ps_area, uint32_t index, struct core_exception *result) { - struct disk_exception *de = get_exception(ps, index); + struct disk_exception *de = get_exception(ps, ps_area, index); /* copy it */ result->old_chunk = le64_to_cpu(de->old_chunk); @@ -413,7 +426,7 @@ static void read_exception(struct pstore *ps, static void write_exception(struct pstore *ps, uint32_t index, struct core_exception *e) { - struct disk_exception *de = get_exception(ps, index); + struct disk_exception *de = get_exception(ps, ps->area, index); /* copy it */ de->old_chunk = cpu_to_le64(e->old_chunk); @@ -422,7 +435,7 @@ static void write_exception(struct pstore *ps, static void clear_exception(struct pstore *ps, uint32_t index) { - struct disk_exception *de = get_exception(ps, index); + struct disk_exception *de = get_exception(ps, ps->area, index); /* clear it */ de->old_chunk = 0; @@ -434,7 +447,7 @@ static void clear_exception(struct pstore *ps, uint32_t index) * 'full' is filled in to indicate if the area has been * filled. */ -static int insert_exceptions(struct pstore *ps, +static int insert_exceptions(struct pstore *ps, void *ps_area, int (*callback)(void *callback_context, chunk_t old, chunk_t new), void *callback_context, @@ -448,7 +461,7 @@ static int insert_exceptions(struct pstore *ps, *full = 1; for (i = 0; i < ps->exceptions_per_area; i++) { - read_exception(ps, i, &e); + read_exception(ps, ps_area, i, &e); /* * If the new_chunk is pointing at the start of @@ -485,24 +498,75 @@ static int read_exceptions(struct pstore *ps, void *callback_context) { int r, full = 1; + struct dm_bufio_client *client; + chunk_t prefetch_area = 0; + + client = dm_bufio_client_create(dm_snap_cow(ps->store->snap)->bdev, + ps->store->chunk_size << SECTOR_SHIFT, + 1, 0, NULL, NULL); + + if (IS_ERR(client)) + return PTR_ERR(client); + + /* + * Setup for one current buffer + desired readahead buffers. + */ + dm_bufio_set_minimum_buffers(client, 1 + DM_PREFETCH_CHUNKS); /* * Keeping reading chunks and inserting exceptions until * we find a partially full area. */ for (ps->current_area = 0; full; ps->current_area++) { - r = area_io(ps, READ); - if (r) - return r; + struct dm_buffer *bp; + void *area; + chunk_t chunk; + + if (unlikely(prefetch_area < ps->current_area)) + prefetch_area = ps->current_area; + + if (DM_PREFETCH_CHUNKS) do { + chunk_t pf_chunk = area_location(ps, prefetch_area); + if (unlikely(pf_chunk >= dm_bufio_get_device_size(client))) + break; + dm_bufio_prefetch(client, pf_chunk, 1); + prefetch_area++; + if (unlikely(!prefetch_area)) + break; + } while (prefetch_area <= ps->current_area + DM_PREFETCH_CHUNKS); + + chunk = area_location(ps, ps->current_area); + + area = dm_bufio_read(client, chunk, &bp); + if (unlikely(IS_ERR(area))) { + r = PTR_ERR(area); + goto ret_destroy_bufio; + } - r = insert_exceptions(ps, callback, callback_context, &full); - if (r) - return r; + r = insert_exceptions(ps, area, callback, callback_context, + &full); + + if (!full) + memcpy(ps->area, area, ps->store->chunk_size << SECTOR_SHIFT); + + dm_bufio_release(bp); + + dm_bufio_forget(client, chunk); + + if (unlikely(r)) + goto ret_destroy_bufio; } ps->current_area--; - return 0; + skip_metadata(ps); + + r = 0; + +ret_destroy_bufio: + dm_bufio_client_destroy(client); + + return r; } static struct pstore *get_info(struct dm_exception_store *store) @@ -616,8 +680,6 @@ static int persistent_prepare_exception(struct dm_exception_store *store, struct dm_exception *e) { struct pstore *ps = get_info(store); - uint32_t stride; - chunk_t next_free; sector_t size = get_dev_size(dm_snap_cow(store->snap)->bdev); /* Is there enough room ? */ @@ -630,10 +692,8 @@ static int persistent_prepare_exception(struct dm_exception_store *store, * Move onto the next free pending, making sure to take * into account the location of the metadata chunks. */ - stride = (ps->exceptions_per_area + 1); - next_free = ++ps->next_free; - if (sector_div(next_free, stride) == 1) - ps->next_free++; + ps->next_free++; + skip_metadata(ps); atomic_inc(&ps->pending_count); return 0; @@ -727,7 +787,7 @@ static int persistent_prepare_merge(struct dm_exception_store *store, ps->current_committed = ps->exceptions_per_area; } - read_exception(ps, ps->current_committed - 1, &ce); + read_exception(ps, ps->area, ps->current_committed - 1, &ce); *last_old_chunk = ce.old_chunk; *last_new_chunk = ce.new_chunk; @@ -737,8 +797,8 @@ static int persistent_prepare_merge(struct dm_exception_store *store, */ for (nr_consecutive = 1; nr_consecutive < ps->current_committed; nr_consecutive++) { - read_exception(ps, ps->current_committed - 1 - nr_consecutive, - &ce); + read_exception(ps, ps->area, + ps->current_committed - 1 - nr_consecutive, &ce); if (ce.old_chunk != *last_old_chunk - nr_consecutive || ce.new_chunk != *last_new_chunk - nr_consecutive) break; diff --git a/drivers/md/dm-snap.c b/drivers/md/dm-snap.c index c434e5aab2d..5bd2290cfb1 100644 --- a/drivers/md/dm-snap.c +++ b/drivers/md/dm-snap.c @@ -66,6 +66,18 @@ struct dm_snapshot { atomic_t pending_exceptions_count; + /* Protected by "lock" */ + sector_t exception_start_sequence; + + /* Protected by kcopyd single-threaded callback */ + sector_t exception_complete_sequence; + + /* + * A list of pending exceptions that completed out of order. + * Protected by kcopyd single-threaded callback. + */ + struct list_head out_of_order_list; + mempool_t *pending_pool; struct dm_exception_table pending; @@ -173,6 +185,14 @@ struct dm_snap_pending_exception { */ int started; + /* There was copying error. */ + int copy_error; + + /* A sequence number, it is used for in-order completion. */ + sector_t exception_sequence; + + struct list_head out_of_order_entry; + /* * For writing a complete chunk, bypassing the copy. */ @@ -590,12 +610,12 @@ static struct dm_exception *dm_lookup_exception(struct dm_exception_table *et, return NULL; } -static struct dm_exception *alloc_completed_exception(void) +static struct dm_exception *alloc_completed_exception(gfp_t gfp) { struct dm_exception *e; - e = kmem_cache_alloc(exception_cache, GFP_NOIO); - if (!e) + e = kmem_cache_alloc(exception_cache, gfp); + if (!e && gfp == GFP_NOIO) e = kmem_cache_alloc(exception_cache, GFP_ATOMIC); return e; @@ -622,7 +642,7 @@ static void free_pending_exception(struct dm_snap_pending_exception *pe) struct dm_snapshot *s = pe->snap; mempool_free(pe, s->pending_pool); - smp_mb__before_atomic_dec(); + smp_mb__before_atomic(); atomic_dec(&s->pending_exceptions_count); } @@ -677,7 +697,7 @@ static int dm_add_exception(void *context, chunk_t old, chunk_t new) struct dm_snapshot *s = context; struct dm_exception *e; - e = alloc_completed_exception(); + e = alloc_completed_exception(GFP_KERNEL); if (!e) return -ENOMEM; @@ -725,17 +745,16 @@ static int calc_max_buckets(void) */ static int init_hash_tables(struct dm_snapshot *s) { - sector_t hash_size, cow_dev_size, origin_dev_size, max_buckets; + sector_t hash_size, cow_dev_size, max_buckets; /* * Calculate based on the size of the original volume or * the COW volume... */ cow_dev_size = get_dev_size(s->cow->bdev); - origin_dev_size = get_dev_size(s->origin->bdev); max_buckets = calc_max_buckets(); - hash_size = min(origin_dev_size, cow_dev_size) >> s->store->chunk_shift; + hash_size = cow_dev_size >> s->store->chunk_shift; hash_size = min(hash_size, max_buckets); if (hash_size < 64) @@ -764,7 +783,7 @@ static int init_hash_tables(struct dm_snapshot *s) static void merge_shutdown(struct dm_snapshot *s) { clear_bit_unlock(RUNNING_MERGE, &s->state_bits); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); wake_up_bit(&s->state_bits, RUNNING_MERGE); } @@ -1095,6 +1114,9 @@ static int snapshot_ctr(struct dm_target *ti, unsigned int argc, char **argv) s->valid = 1; s->active = 0; atomic_set(&s->pending_exceptions_count, 0); + s->exception_start_sequence = 0; + s->exception_complete_sequence = 0; + INIT_LIST_HEAD(&s->out_of_order_list); init_rwsem(&s->lock); INIT_LIST_HEAD(&s->list); spin_lock_init(&s->pe_lock); @@ -1383,7 +1405,7 @@ static void pending_complete(struct dm_snap_pending_exception *pe, int success) goto out; } - e = alloc_completed_exception(); + e = alloc_completed_exception(GFP_NOIO); if (!e) { down_write(&s->lock); __invalidate_snapshot(s, -ENOMEM); @@ -1416,6 +1438,7 @@ out: if (full_bio) { full_bio->bi_end_io = pe->full_bio_end_io; full_bio->bi_private = pe->full_bio_private; + atomic_inc(&full_bio->bi_remaining); } free_pending_exception(pe); @@ -1444,6 +1467,19 @@ static void commit_callback(void *context, int success) pending_complete(pe, success); } +static void complete_exception(struct dm_snap_pending_exception *pe) +{ + struct dm_snapshot *s = pe->snap; + + if (unlikely(pe->copy_error)) + pending_complete(pe, 0); + + else + /* Update the metadata if we are persistent */ + s->store->type->commit_exception(s->store, &pe->e, + commit_callback, pe); +} + /* * Called when the copy I/O has finished. kcopyd actually runs * this code so don't block. @@ -1453,13 +1489,32 @@ static void copy_callback(int read_err, unsigned long write_err, void *context) struct dm_snap_pending_exception *pe = context; struct dm_snapshot *s = pe->snap; - if (read_err || write_err) - pending_complete(pe, 0); + pe->copy_error = read_err || write_err; - else - /* Update the metadata if we are persistent */ - s->store->type->commit_exception(s->store, &pe->e, - commit_callback, pe); + if (pe->exception_sequence == s->exception_complete_sequence) { + s->exception_complete_sequence++; + complete_exception(pe); + + while (!list_empty(&s->out_of_order_list)) { + pe = list_entry(s->out_of_order_list.next, + struct dm_snap_pending_exception, out_of_order_entry); + if (pe->exception_sequence != s->exception_complete_sequence) + break; + s->exception_complete_sequence++; + list_del(&pe->out_of_order_entry); + complete_exception(pe); + } + } else { + struct list_head *lh; + struct dm_snap_pending_exception *pe2; + + list_for_each_prev(lh, &s->out_of_order_list) { + pe2 = list_entry(lh, struct dm_snap_pending_exception, out_of_order_entry); + if (pe2->exception_sequence < pe->exception_sequence) + break; + } + list_add(&pe->out_of_order_entry, lh); + } } /* @@ -1554,6 +1609,8 @@ __find_pending_exception(struct dm_snapshot *s, return NULL; } + pe->exception_sequence = s->exception_start_sequence++; + dm_insert_exception(&s->pending, &pe->e); return pe; @@ -1563,11 +1620,10 @@ static void remap_exception(struct dm_snapshot *s, struct dm_exception *e, struct bio *bio, chunk_t chunk) { bio->bi_bdev = s->cow->bdev; - bio->bi_sector = chunk_to_sector(s->store, - dm_chunk_number(e->new_chunk) + - (chunk - e->old_chunk)) + - (bio->bi_sector & - s->store->chunk_mask); + bio->bi_iter.bi_sector = + chunk_to_sector(s->store, dm_chunk_number(e->new_chunk) + + (chunk - e->old_chunk)) + + (bio->bi_iter.bi_sector & s->store->chunk_mask); } static int snapshot_map(struct dm_target *ti, struct bio *bio) @@ -1585,7 +1641,7 @@ static int snapshot_map(struct dm_target *ti, struct bio *bio) return DM_MAPIO_REMAPPED; } - chunk = sector_to_chunk(s->store, bio->bi_sector); + chunk = sector_to_chunk(s->store, bio->bi_iter.bi_sector); /* Full snapshots are not usable */ /* To get here the table must be live so s->active is always set. */ @@ -1646,7 +1702,8 @@ static int snapshot_map(struct dm_target *ti, struct bio *bio) r = DM_MAPIO_SUBMITTED; if (!pe->started && - bio->bi_size == (s->store->chunk_size << SECTOR_SHIFT)) { + bio->bi_iter.bi_size == + (s->store->chunk_size << SECTOR_SHIFT)) { pe->started = 1; up_write(&s->lock); start_full_bio(pe, bio); @@ -1702,7 +1759,7 @@ static int snapshot_merge_map(struct dm_target *ti, struct bio *bio) return DM_MAPIO_REMAPPED; } - chunk = sector_to_chunk(s->store, bio->bi_sector); + chunk = sector_to_chunk(s->store, bio->bi_iter.bi_sector); down_write(&s->lock); @@ -2039,7 +2096,7 @@ static int do_origin(struct dm_dev *origin, struct bio *bio) down_read(&_origins_lock); o = __lookup_origin(origin->bdev); if (o) - r = __origin_write(&o->snapshots, bio->bi_sector, bio); + r = __origin_write(&o->snapshots, bio->bi_iter.bi_sector, bio); up_read(&_origins_lock); return r; @@ -2084,6 +2141,11 @@ static int origin_write_extent(struct dm_snapshot *merging_snap, * Origin: maps a linear range of a device, with hooks for snapshotting. */ +struct dm_origin { + struct dm_dev *dev; + unsigned split_boundary; +}; + /* * Construct an origin mapping: <dev_path> * The context for an origin is merely a 'struct dm_dev *' @@ -2092,41 +2154,65 @@ static int origin_write_extent(struct dm_snapshot *merging_snap, static int origin_ctr(struct dm_target *ti, unsigned int argc, char **argv) { int r; - struct dm_dev *dev; + struct dm_origin *o; if (argc != 1) { ti->error = "origin: incorrect number of arguments"; return -EINVAL; } - r = dm_get_device(ti, argv[0], dm_table_get_mode(ti->table), &dev); + o = kmalloc(sizeof(struct dm_origin), GFP_KERNEL); + if (!o) { + ti->error = "Cannot allocate private origin structure"; + r = -ENOMEM; + goto bad_alloc; + } + + r = dm_get_device(ti, argv[0], dm_table_get_mode(ti->table), &o->dev); if (r) { ti->error = "Cannot get target device"; - return r; + goto bad_open; } - ti->private = dev; + ti->private = o; ti->num_flush_bios = 1; return 0; + +bad_open: + kfree(o); +bad_alloc: + return r; } static void origin_dtr(struct dm_target *ti) { - struct dm_dev *dev = ti->private; - dm_put_device(ti, dev); + struct dm_origin *o = ti->private; + dm_put_device(ti, o->dev); + kfree(o); } static int origin_map(struct dm_target *ti, struct bio *bio) { - struct dm_dev *dev = ti->private; - bio->bi_bdev = dev->bdev; + struct dm_origin *o = ti->private; + unsigned available_sectors; - if (bio->bi_rw & REQ_FLUSH) + bio->bi_bdev = o->dev->bdev; + + if (unlikely(bio->bi_rw & REQ_FLUSH)) + return DM_MAPIO_REMAPPED; + + if (bio_rw(bio) != WRITE) return DM_MAPIO_REMAPPED; + available_sectors = o->split_boundary - + ((unsigned)bio->bi_iter.bi_sector & (o->split_boundary - 1)); + + if (bio_sectors(bio) > available_sectors) + dm_accept_partial_bio(bio, available_sectors); + /* Only tell snapshots if this is a write */ - return (bio_rw(bio) == WRITE) ? do_origin(dev, bio) : DM_MAPIO_REMAPPED; + return do_origin(o->dev, bio); } /* @@ -2135,15 +2221,15 @@ static int origin_map(struct dm_target *ti, struct bio *bio) */ static void origin_resume(struct dm_target *ti) { - struct dm_dev *dev = ti->private; + struct dm_origin *o = ti->private; - ti->max_io_len = get_origin_minimum_chunksize(dev->bdev); + o->split_boundary = get_origin_minimum_chunksize(o->dev->bdev); } static void origin_status(struct dm_target *ti, status_type_t type, unsigned status_flags, char *result, unsigned maxlen) { - struct dm_dev *dev = ti->private; + struct dm_origin *o = ti->private; switch (type) { case STATUSTYPE_INFO: @@ -2151,7 +2237,7 @@ static void origin_status(struct dm_target *ti, status_type_t type, break; case STATUSTYPE_TABLE: - snprintf(result, maxlen, "%s", dev->name); + snprintf(result, maxlen, "%s", o->dev->name); break; } } @@ -2159,13 +2245,13 @@ static void origin_status(struct dm_target *ti, status_type_t type, static int origin_merge(struct dm_target *ti, struct bvec_merge_data *bvm, struct bio_vec *biovec, int max_size) { - struct dm_dev *dev = ti->private; - struct request_queue *q = bdev_get_queue(dev->bdev); + struct dm_origin *o = ti->private; + struct request_queue *q = bdev_get_queue(o->dev->bdev); if (!q->merge_bvec_fn) return max_size; - bvm->bi_bdev = dev->bdev; + bvm->bi_bdev = o->dev->bdev; return min(max_size, q->merge_bvec_fn(q, bvm, biovec)); } @@ -2173,9 +2259,9 @@ static int origin_merge(struct dm_target *ti, struct bvec_merge_data *bvm, static int origin_iterate_devices(struct dm_target *ti, iterate_devices_callout_fn fn, void *data) { - struct dm_dev *dev = ti->private; + struct dm_origin *o = ti->private; - return fn(ti, dev, 0, ti->len, data); + return fn(ti, o->dev, 0, ti->len, data); } static struct target_type origin_target = { @@ -2193,7 +2279,7 @@ static struct target_type origin_target = { static struct target_type snapshot_target = { .name = "snapshot", - .version = {1, 11, 1}, + .version = {1, 12, 0}, .module = THIS_MODULE, .ctr = snapshot_ctr, .dtr = snapshot_dtr, diff --git a/drivers/md/dm-stats.c b/drivers/md/dm-stats.c new file mode 100644 index 00000000000..28a90122a5a --- /dev/null +++ b/drivers/md/dm-stats.c @@ -0,0 +1,981 @@ +#include <linux/errno.h> +#include <linux/numa.h> +#include <linux/slab.h> +#include <linux/rculist.h> +#include <linux/threads.h> +#include <linux/preempt.h> +#include <linux/irqflags.h> +#include <linux/vmalloc.h> +#include <linux/mm.h> +#include <linux/module.h> +#include <linux/device-mapper.h> + +#include "dm.h" +#include "dm-stats.h" + +#define DM_MSG_PREFIX "stats" + +static int dm_stat_need_rcu_barrier; + +/* + * Using 64-bit values to avoid overflow (which is a + * problem that block/genhd.c's IO accounting has). + */ +struct dm_stat_percpu { + unsigned long long sectors[2]; + unsigned long long ios[2]; + unsigned long long merges[2]; + unsigned long long ticks[2]; + unsigned long long io_ticks[2]; + unsigned long long io_ticks_total; + unsigned long long time_in_queue; +}; + +struct dm_stat_shared { + atomic_t in_flight[2]; + unsigned long stamp; + struct dm_stat_percpu tmp; +}; + +struct dm_stat { + struct list_head list_entry; + int id; + size_t n_entries; + sector_t start; + sector_t end; + sector_t step; + const char *program_id; + const char *aux_data; + struct rcu_head rcu_head; + size_t shared_alloc_size; + size_t percpu_alloc_size; + struct dm_stat_percpu *stat_percpu[NR_CPUS]; + struct dm_stat_shared stat_shared[0]; +}; + +struct dm_stats_last_position { + sector_t last_sector; + unsigned last_rw; +}; + +/* + * A typo on the command line could possibly make the kernel run out of memory + * and crash. To prevent the crash we account all used memory. We fail if we + * exhaust 1/4 of all memory or 1/2 of vmalloc space. + */ +#define DM_STATS_MEMORY_FACTOR 4 +#define DM_STATS_VMALLOC_FACTOR 2 + +static DEFINE_SPINLOCK(shared_memory_lock); + +static unsigned long shared_memory_amount; + +static bool __check_shared_memory(size_t alloc_size) +{ + size_t a; + + a = shared_memory_amount + alloc_size; + if (a < shared_memory_amount) + return false; + if (a >> PAGE_SHIFT > totalram_pages / DM_STATS_MEMORY_FACTOR) + return false; +#ifdef CONFIG_MMU + if (a > (VMALLOC_END - VMALLOC_START) / DM_STATS_VMALLOC_FACTOR) + return false; +#endif + return true; +} + +static bool check_shared_memory(size_t alloc_size) +{ + bool ret; + + spin_lock_irq(&shared_memory_lock); + + ret = __check_shared_memory(alloc_size); + + spin_unlock_irq(&shared_memory_lock); + + return ret; +} + +static bool claim_shared_memory(size_t alloc_size) +{ + spin_lock_irq(&shared_memory_lock); + + if (!__check_shared_memory(alloc_size)) { + spin_unlock_irq(&shared_memory_lock); + return false; + } + + shared_memory_amount += alloc_size; + + spin_unlock_irq(&shared_memory_lock); + + return true; +} + +static void free_shared_memory(size_t alloc_size) +{ + unsigned long flags; + + spin_lock_irqsave(&shared_memory_lock, flags); + + if (WARN_ON_ONCE(shared_memory_amount < alloc_size)) { + spin_unlock_irqrestore(&shared_memory_lock, flags); + DMCRIT("Memory usage accounting bug."); + return; + } + + shared_memory_amount -= alloc_size; + + spin_unlock_irqrestore(&shared_memory_lock, flags); +} + +static void *dm_kvzalloc(size_t alloc_size, int node) +{ + void *p; + + if (!claim_shared_memory(alloc_size)) + return NULL; + + if (alloc_size <= KMALLOC_MAX_SIZE) { + p = kzalloc_node(alloc_size, GFP_KERNEL | __GFP_NORETRY | __GFP_NOMEMALLOC | __GFP_NOWARN, node); + if (p) + return p; + } + p = vzalloc_node(alloc_size, node); + if (p) + return p; + + free_shared_memory(alloc_size); + + return NULL; +} + +static void dm_kvfree(void *ptr, size_t alloc_size) +{ + if (!ptr) + return; + + free_shared_memory(alloc_size); + + if (is_vmalloc_addr(ptr)) + vfree(ptr); + else + kfree(ptr); +} + +static void dm_stat_free(struct rcu_head *head) +{ + int cpu; + struct dm_stat *s = container_of(head, struct dm_stat, rcu_head); + + kfree(s->program_id); + kfree(s->aux_data); + for_each_possible_cpu(cpu) + dm_kvfree(s->stat_percpu[cpu], s->percpu_alloc_size); + dm_kvfree(s, s->shared_alloc_size); +} + +static int dm_stat_in_flight(struct dm_stat_shared *shared) +{ + return atomic_read(&shared->in_flight[READ]) + + atomic_read(&shared->in_flight[WRITE]); +} + +void dm_stats_init(struct dm_stats *stats) +{ + int cpu; + struct dm_stats_last_position *last; + + mutex_init(&stats->mutex); + INIT_LIST_HEAD(&stats->list); + stats->last = alloc_percpu(struct dm_stats_last_position); + for_each_possible_cpu(cpu) { + last = per_cpu_ptr(stats->last, cpu); + last->last_sector = (sector_t)ULLONG_MAX; + last->last_rw = UINT_MAX; + } +} + +void dm_stats_cleanup(struct dm_stats *stats) +{ + size_t ni; + struct dm_stat *s; + struct dm_stat_shared *shared; + + while (!list_empty(&stats->list)) { + s = container_of(stats->list.next, struct dm_stat, list_entry); + list_del(&s->list_entry); + for (ni = 0; ni < s->n_entries; ni++) { + shared = &s->stat_shared[ni]; + if (WARN_ON(dm_stat_in_flight(shared))) { + DMCRIT("leaked in-flight counter at index %lu " + "(start %llu, end %llu, step %llu): reads %d, writes %d", + (unsigned long)ni, + (unsigned long long)s->start, + (unsigned long long)s->end, + (unsigned long long)s->step, + atomic_read(&shared->in_flight[READ]), + atomic_read(&shared->in_flight[WRITE])); + } + } + dm_stat_free(&s->rcu_head); + } + free_percpu(stats->last); +} + +static int dm_stats_create(struct dm_stats *stats, sector_t start, sector_t end, + sector_t step, const char *program_id, const char *aux_data, + void (*suspend_callback)(struct mapped_device *), + void (*resume_callback)(struct mapped_device *), + struct mapped_device *md) +{ + struct list_head *l; + struct dm_stat *s, *tmp_s; + sector_t n_entries; + size_t ni; + size_t shared_alloc_size; + size_t percpu_alloc_size; + struct dm_stat_percpu *p; + int cpu; + int ret_id; + int r; + + if (end < start || !step) + return -EINVAL; + + n_entries = end - start; + if (dm_sector_div64(n_entries, step)) + n_entries++; + + if (n_entries != (size_t)n_entries || !(size_t)(n_entries + 1)) + return -EOVERFLOW; + + shared_alloc_size = sizeof(struct dm_stat) + (size_t)n_entries * sizeof(struct dm_stat_shared); + if ((shared_alloc_size - sizeof(struct dm_stat)) / sizeof(struct dm_stat_shared) != n_entries) + return -EOVERFLOW; + + percpu_alloc_size = (size_t)n_entries * sizeof(struct dm_stat_percpu); + if (percpu_alloc_size / sizeof(struct dm_stat_percpu) != n_entries) + return -EOVERFLOW; + + if (!check_shared_memory(shared_alloc_size + num_possible_cpus() * percpu_alloc_size)) + return -ENOMEM; + + s = dm_kvzalloc(shared_alloc_size, NUMA_NO_NODE); + if (!s) + return -ENOMEM; + + s->n_entries = n_entries; + s->start = start; + s->end = end; + s->step = step; + s->shared_alloc_size = shared_alloc_size; + s->percpu_alloc_size = percpu_alloc_size; + + s->program_id = kstrdup(program_id, GFP_KERNEL); + if (!s->program_id) { + r = -ENOMEM; + goto out; + } + s->aux_data = kstrdup(aux_data, GFP_KERNEL); + if (!s->aux_data) { + r = -ENOMEM; + goto out; + } + + for (ni = 0; ni < n_entries; ni++) { + atomic_set(&s->stat_shared[ni].in_flight[READ], 0); + atomic_set(&s->stat_shared[ni].in_flight[WRITE], 0); + } + + for_each_possible_cpu(cpu) { + p = dm_kvzalloc(percpu_alloc_size, cpu_to_node(cpu)); + if (!p) { + r = -ENOMEM; + goto out; + } + s->stat_percpu[cpu] = p; + } + + /* + * Suspend/resume to make sure there is no i/o in flight, + * so that newly created statistics will be exact. + * + * (note: we couldn't suspend earlier because we must not + * allocate memory while suspended) + */ + suspend_callback(md); + + mutex_lock(&stats->mutex); + s->id = 0; + list_for_each(l, &stats->list) { + tmp_s = container_of(l, struct dm_stat, list_entry); + if (WARN_ON(tmp_s->id < s->id)) { + r = -EINVAL; + goto out_unlock_resume; + } + if (tmp_s->id > s->id) + break; + if (unlikely(s->id == INT_MAX)) { + r = -ENFILE; + goto out_unlock_resume; + } + s->id++; + } + ret_id = s->id; + list_add_tail_rcu(&s->list_entry, l); + mutex_unlock(&stats->mutex); + + resume_callback(md); + + return ret_id; + +out_unlock_resume: + mutex_unlock(&stats->mutex); + resume_callback(md); +out: + dm_stat_free(&s->rcu_head); + return r; +} + +static struct dm_stat *__dm_stats_find(struct dm_stats *stats, int id) +{ + struct dm_stat *s; + + list_for_each_entry(s, &stats->list, list_entry) { + if (s->id > id) + break; + if (s->id == id) + return s; + } + + return NULL; +} + +static int dm_stats_delete(struct dm_stats *stats, int id) +{ + struct dm_stat *s; + int cpu; + + mutex_lock(&stats->mutex); + + s = __dm_stats_find(stats, id); + if (!s) { + mutex_unlock(&stats->mutex); + return -ENOENT; + } + + list_del_rcu(&s->list_entry); + mutex_unlock(&stats->mutex); + + /* + * vfree can't be called from RCU callback + */ + for_each_possible_cpu(cpu) + if (is_vmalloc_addr(s->stat_percpu)) + goto do_sync_free; + if (is_vmalloc_addr(s)) { +do_sync_free: + synchronize_rcu_expedited(); + dm_stat_free(&s->rcu_head); + } else { + ACCESS_ONCE(dm_stat_need_rcu_barrier) = 1; + call_rcu(&s->rcu_head, dm_stat_free); + } + return 0; +} + +static int dm_stats_list(struct dm_stats *stats, const char *program, + char *result, unsigned maxlen) +{ + struct dm_stat *s; + sector_t len; + unsigned sz = 0; + + /* + * Output format: + * <region_id>: <start_sector>+<length> <step> <program_id> <aux_data> + */ + + mutex_lock(&stats->mutex); + list_for_each_entry(s, &stats->list, list_entry) { + if (!program || !strcmp(program, s->program_id)) { + len = s->end - s->start; + DMEMIT("%d: %llu+%llu %llu %s %s\n", s->id, + (unsigned long long)s->start, + (unsigned long long)len, + (unsigned long long)s->step, + s->program_id, + s->aux_data); + } + } + mutex_unlock(&stats->mutex); + + return 1; +} + +static void dm_stat_round(struct dm_stat_shared *shared, struct dm_stat_percpu *p) +{ + /* + * This is racy, but so is part_round_stats_single. + */ + unsigned long now = jiffies; + unsigned in_flight_read; + unsigned in_flight_write; + unsigned long difference = now - shared->stamp; + + if (!difference) + return; + in_flight_read = (unsigned)atomic_read(&shared->in_flight[READ]); + in_flight_write = (unsigned)atomic_read(&shared->in_flight[WRITE]); + if (in_flight_read) + p->io_ticks[READ] += difference; + if (in_flight_write) + p->io_ticks[WRITE] += difference; + if (in_flight_read + in_flight_write) { + p->io_ticks_total += difference; + p->time_in_queue += (in_flight_read + in_flight_write) * difference; + } + shared->stamp = now; +} + +static void dm_stat_for_entry(struct dm_stat *s, size_t entry, + unsigned long bi_rw, sector_t len, bool merged, + bool end, unsigned long duration) +{ + unsigned long idx = bi_rw & REQ_WRITE; + struct dm_stat_shared *shared = &s->stat_shared[entry]; + struct dm_stat_percpu *p; + + /* + * For strict correctness we should use local_irq_save/restore + * instead of preempt_disable/enable. + * + * preempt_disable/enable is racy if the driver finishes bios + * from non-interrupt context as well as from interrupt context + * or from more different interrupts. + * + * On 64-bit architectures the race only results in not counting some + * events, so it is acceptable. On 32-bit architectures the race could + * cause the counter going off by 2^32, so we need to do proper locking + * there. + * + * part_stat_lock()/part_stat_unlock() have this race too. + */ +#if BITS_PER_LONG == 32 + unsigned long flags; + local_irq_save(flags); +#else + preempt_disable(); +#endif + p = &s->stat_percpu[smp_processor_id()][entry]; + + if (!end) { + dm_stat_round(shared, p); + atomic_inc(&shared->in_flight[idx]); + } else { + dm_stat_round(shared, p); + atomic_dec(&shared->in_flight[idx]); + p->sectors[idx] += len; + p->ios[idx] += 1; + p->merges[idx] += merged; + p->ticks[idx] += duration; + } + +#if BITS_PER_LONG == 32 + local_irq_restore(flags); +#else + preempt_enable(); +#endif +} + +static void __dm_stat_bio(struct dm_stat *s, unsigned long bi_rw, + sector_t bi_sector, sector_t end_sector, + bool end, unsigned long duration, + struct dm_stats_aux *stats_aux) +{ + sector_t rel_sector, offset, todo, fragment_len; + size_t entry; + + if (end_sector <= s->start || bi_sector >= s->end) + return; + if (unlikely(bi_sector < s->start)) { + rel_sector = 0; + todo = end_sector - s->start; + } else { + rel_sector = bi_sector - s->start; + todo = end_sector - bi_sector; + } + if (unlikely(end_sector > s->end)) + todo -= (end_sector - s->end); + + offset = dm_sector_div64(rel_sector, s->step); + entry = rel_sector; + do { + if (WARN_ON_ONCE(entry >= s->n_entries)) { + DMCRIT("Invalid area access in region id %d", s->id); + return; + } + fragment_len = todo; + if (fragment_len > s->step - offset) + fragment_len = s->step - offset; + dm_stat_for_entry(s, entry, bi_rw, fragment_len, + stats_aux->merged, end, duration); + todo -= fragment_len; + entry++; + offset = 0; + } while (unlikely(todo != 0)); +} + +void dm_stats_account_io(struct dm_stats *stats, unsigned long bi_rw, + sector_t bi_sector, unsigned bi_sectors, bool end, + unsigned long duration, struct dm_stats_aux *stats_aux) +{ + struct dm_stat *s; + sector_t end_sector; + struct dm_stats_last_position *last; + + if (unlikely(!bi_sectors)) + return; + + end_sector = bi_sector + bi_sectors; + + if (!end) { + /* + * A race condition can at worst result in the merged flag being + * misrepresented, so we don't have to disable preemption here. + */ + last = __this_cpu_ptr(stats->last); + stats_aux->merged = + (bi_sector == (ACCESS_ONCE(last->last_sector) && + ((bi_rw & (REQ_WRITE | REQ_DISCARD)) == + (ACCESS_ONCE(last->last_rw) & (REQ_WRITE | REQ_DISCARD))) + )); + ACCESS_ONCE(last->last_sector) = end_sector; + ACCESS_ONCE(last->last_rw) = bi_rw; + } + + rcu_read_lock(); + + list_for_each_entry_rcu(s, &stats->list, list_entry) + __dm_stat_bio(s, bi_rw, bi_sector, end_sector, end, duration, stats_aux); + + rcu_read_unlock(); +} + +static void __dm_stat_init_temporary_percpu_totals(struct dm_stat_shared *shared, + struct dm_stat *s, size_t x) +{ + int cpu; + struct dm_stat_percpu *p; + + local_irq_disable(); + p = &s->stat_percpu[smp_processor_id()][x]; + dm_stat_round(shared, p); + local_irq_enable(); + + memset(&shared->tmp, 0, sizeof(shared->tmp)); + for_each_possible_cpu(cpu) { + p = &s->stat_percpu[cpu][x]; + shared->tmp.sectors[READ] += ACCESS_ONCE(p->sectors[READ]); + shared->tmp.sectors[WRITE] += ACCESS_ONCE(p->sectors[WRITE]); + shared->tmp.ios[READ] += ACCESS_ONCE(p->ios[READ]); + shared->tmp.ios[WRITE] += ACCESS_ONCE(p->ios[WRITE]); + shared->tmp.merges[READ] += ACCESS_ONCE(p->merges[READ]); + shared->tmp.merges[WRITE] += ACCESS_ONCE(p->merges[WRITE]); + shared->tmp.ticks[READ] += ACCESS_ONCE(p->ticks[READ]); + shared->tmp.ticks[WRITE] += ACCESS_ONCE(p->ticks[WRITE]); + shared->tmp.io_ticks[READ] += ACCESS_ONCE(p->io_ticks[READ]); + shared->tmp.io_ticks[WRITE] += ACCESS_ONCE(p->io_ticks[WRITE]); + shared->tmp.io_ticks_total += ACCESS_ONCE(p->io_ticks_total); + shared->tmp.time_in_queue += ACCESS_ONCE(p->time_in_queue); + } +} + +static void __dm_stat_clear(struct dm_stat *s, size_t idx_start, size_t idx_end, + bool init_tmp_percpu_totals) +{ + size_t x; + struct dm_stat_shared *shared; + struct dm_stat_percpu *p; + + for (x = idx_start; x < idx_end; x++) { + shared = &s->stat_shared[x]; + if (init_tmp_percpu_totals) + __dm_stat_init_temporary_percpu_totals(shared, s, x); + local_irq_disable(); + p = &s->stat_percpu[smp_processor_id()][x]; + p->sectors[READ] -= shared->tmp.sectors[READ]; + p->sectors[WRITE] -= shared->tmp.sectors[WRITE]; + p->ios[READ] -= shared->tmp.ios[READ]; + p->ios[WRITE] -= shared->tmp.ios[WRITE]; + p->merges[READ] -= shared->tmp.merges[READ]; + p->merges[WRITE] -= shared->tmp.merges[WRITE]; + p->ticks[READ] -= shared->tmp.ticks[READ]; + p->ticks[WRITE] -= shared->tmp.ticks[WRITE]; + p->io_ticks[READ] -= shared->tmp.io_ticks[READ]; + p->io_ticks[WRITE] -= shared->tmp.io_ticks[WRITE]; + p->io_ticks_total -= shared->tmp.io_ticks_total; + p->time_in_queue -= shared->tmp.time_in_queue; + local_irq_enable(); + } +} + +static int dm_stats_clear(struct dm_stats *stats, int id) +{ + struct dm_stat *s; + + mutex_lock(&stats->mutex); + + s = __dm_stats_find(stats, id); + if (!s) { + mutex_unlock(&stats->mutex); + return -ENOENT; + } + + __dm_stat_clear(s, 0, s->n_entries, true); + + mutex_unlock(&stats->mutex); + + return 1; +} + +/* + * This is like jiffies_to_msec, but works for 64-bit values. + */ +static unsigned long long dm_jiffies_to_msec64(unsigned long long j) +{ + unsigned long long result = 0; + unsigned mult; + + if (j) + result = jiffies_to_msecs(j & 0x3fffff); + if (j >= 1 << 22) { + mult = jiffies_to_msecs(1 << 22); + result += (unsigned long long)mult * (unsigned long long)jiffies_to_msecs((j >> 22) & 0x3fffff); + } + if (j >= 1ULL << 44) + result += (unsigned long long)mult * (unsigned long long)mult * (unsigned long long)jiffies_to_msecs(j >> 44); + + return result; +} + +static int dm_stats_print(struct dm_stats *stats, int id, + size_t idx_start, size_t idx_len, + bool clear, char *result, unsigned maxlen) +{ + unsigned sz = 0; + struct dm_stat *s; + size_t x; + sector_t start, end, step; + size_t idx_end; + struct dm_stat_shared *shared; + + /* + * Output format: + * <start_sector>+<length> counters + */ + + mutex_lock(&stats->mutex); + + s = __dm_stats_find(stats, id); + if (!s) { + mutex_unlock(&stats->mutex); + return -ENOENT; + } + + idx_end = idx_start + idx_len; + if (idx_end < idx_start || + idx_end > s->n_entries) + idx_end = s->n_entries; + + if (idx_start > idx_end) + idx_start = idx_end; + + step = s->step; + start = s->start + (step * idx_start); + + for (x = idx_start; x < idx_end; x++, start = end) { + shared = &s->stat_shared[x]; + end = start + step; + if (unlikely(end > s->end)) + end = s->end; + + __dm_stat_init_temporary_percpu_totals(shared, s, x); + + DMEMIT("%llu+%llu %llu %llu %llu %llu %llu %llu %llu %llu %d %llu %llu %llu %llu\n", + (unsigned long long)start, + (unsigned long long)step, + shared->tmp.ios[READ], + shared->tmp.merges[READ], + shared->tmp.sectors[READ], + dm_jiffies_to_msec64(shared->tmp.ticks[READ]), + shared->tmp.ios[WRITE], + shared->tmp.merges[WRITE], + shared->tmp.sectors[WRITE], + dm_jiffies_to_msec64(shared->tmp.ticks[WRITE]), + dm_stat_in_flight(shared), + dm_jiffies_to_msec64(shared->tmp.io_ticks_total), + dm_jiffies_to_msec64(shared->tmp.time_in_queue), + dm_jiffies_to_msec64(shared->tmp.io_ticks[READ]), + dm_jiffies_to_msec64(shared->tmp.io_ticks[WRITE])); + + if (unlikely(sz + 1 >= maxlen)) + goto buffer_overflow; + } + + if (clear) + __dm_stat_clear(s, idx_start, idx_end, false); + +buffer_overflow: + mutex_unlock(&stats->mutex); + + return 1; +} + +static int dm_stats_set_aux(struct dm_stats *stats, int id, const char *aux_data) +{ + struct dm_stat *s; + const char *new_aux_data; + + mutex_lock(&stats->mutex); + + s = __dm_stats_find(stats, id); + if (!s) { + mutex_unlock(&stats->mutex); + return -ENOENT; + } + + new_aux_data = kstrdup(aux_data, GFP_KERNEL); + if (!new_aux_data) { + mutex_unlock(&stats->mutex); + return -ENOMEM; + } + + kfree(s->aux_data); + s->aux_data = new_aux_data; + + mutex_unlock(&stats->mutex); + + return 0; +} + +static int message_stats_create(struct mapped_device *md, + unsigned argc, char **argv, + char *result, unsigned maxlen) +{ + int id; + char dummy; + unsigned long long start, end, len, step; + unsigned divisor; + const char *program_id, *aux_data; + + /* + * Input format: + * <range> <step> [<program_id> [<aux_data>]] + */ + + if (argc < 3 || argc > 5) + return -EINVAL; + + if (!strcmp(argv[1], "-")) { + start = 0; + len = dm_get_size(md); + if (!len) + len = 1; + } else if (sscanf(argv[1], "%llu+%llu%c", &start, &len, &dummy) != 2 || + start != (sector_t)start || len != (sector_t)len) + return -EINVAL; + + end = start + len; + if (start >= end) + return -EINVAL; + + if (sscanf(argv[2], "/%u%c", &divisor, &dummy) == 1) { + step = end - start; + if (do_div(step, divisor)) + step++; + if (!step) + step = 1; + } else if (sscanf(argv[2], "%llu%c", &step, &dummy) != 1 || + step != (sector_t)step || !step) + return -EINVAL; + + program_id = "-"; + aux_data = "-"; + + if (argc > 3) + program_id = argv[3]; + + if (argc > 4) + aux_data = argv[4]; + + /* + * If a buffer overflow happens after we created the region, + * it's too late (the userspace would retry with a larger + * buffer, but the region id that caused the overflow is already + * leaked). So we must detect buffer overflow in advance. + */ + snprintf(result, maxlen, "%d", INT_MAX); + if (dm_message_test_buffer_overflow(result, maxlen)) + return 1; + + id = dm_stats_create(dm_get_stats(md), start, end, step, program_id, aux_data, + dm_internal_suspend, dm_internal_resume, md); + if (id < 0) + return id; + + snprintf(result, maxlen, "%d", id); + + return 1; +} + +static int message_stats_delete(struct mapped_device *md, + unsigned argc, char **argv) +{ + int id; + char dummy; + + if (argc != 2) + return -EINVAL; + + if (sscanf(argv[1], "%d%c", &id, &dummy) != 1 || id < 0) + return -EINVAL; + + return dm_stats_delete(dm_get_stats(md), id); +} + +static int message_stats_clear(struct mapped_device *md, + unsigned argc, char **argv) +{ + int id; + char dummy; + + if (argc != 2) + return -EINVAL; + + if (sscanf(argv[1], "%d%c", &id, &dummy) != 1 || id < 0) + return -EINVAL; + + return dm_stats_clear(dm_get_stats(md), id); +} + +static int message_stats_list(struct mapped_device *md, + unsigned argc, char **argv, + char *result, unsigned maxlen) +{ + int r; + const char *program = NULL; + + if (argc < 1 || argc > 2) + return -EINVAL; + + if (argc > 1) { + program = kstrdup(argv[1], GFP_KERNEL); + if (!program) + return -ENOMEM; + } + + r = dm_stats_list(dm_get_stats(md), program, result, maxlen); + + kfree(program); + + return r; +} + +static int message_stats_print(struct mapped_device *md, + unsigned argc, char **argv, bool clear, + char *result, unsigned maxlen) +{ + int id; + char dummy; + unsigned long idx_start = 0, idx_len = ULONG_MAX; + + if (argc != 2 && argc != 4) + return -EINVAL; + + if (sscanf(argv[1], "%d%c", &id, &dummy) != 1 || id < 0) + return -EINVAL; + + if (argc > 3) { + if (strcmp(argv[2], "-") && + sscanf(argv[2], "%lu%c", &idx_start, &dummy) != 1) + return -EINVAL; + if (strcmp(argv[3], "-") && + sscanf(argv[3], "%lu%c", &idx_len, &dummy) != 1) + return -EINVAL; + } + + return dm_stats_print(dm_get_stats(md), id, idx_start, idx_len, clear, + result, maxlen); +} + +static int message_stats_set_aux(struct mapped_device *md, + unsigned argc, char **argv) +{ + int id; + char dummy; + + if (argc != 3) + return -EINVAL; + + if (sscanf(argv[1], "%d%c", &id, &dummy) != 1 || id < 0) + return -EINVAL; + + return dm_stats_set_aux(dm_get_stats(md), id, argv[2]); +} + +int dm_stats_message(struct mapped_device *md, unsigned argc, char **argv, + char *result, unsigned maxlen) +{ + int r; + + if (dm_request_based(md)) { + DMWARN("Statistics are only supported for bio-based devices"); + return -EOPNOTSUPP; + } + + /* All messages here must start with '@' */ + if (!strcasecmp(argv[0], "@stats_create")) + r = message_stats_create(md, argc, argv, result, maxlen); + else if (!strcasecmp(argv[0], "@stats_delete")) + r = message_stats_delete(md, argc, argv); + else if (!strcasecmp(argv[0], "@stats_clear")) + r = message_stats_clear(md, argc, argv); + else if (!strcasecmp(argv[0], "@stats_list")) + r = message_stats_list(md, argc, argv, result, maxlen); + else if (!strcasecmp(argv[0], "@stats_print")) + r = message_stats_print(md, argc, argv, false, result, maxlen); + else if (!strcasecmp(argv[0], "@stats_print_clear")) + r = message_stats_print(md, argc, argv, true, result, maxlen); + else if (!strcasecmp(argv[0], "@stats_set_aux")) + r = message_stats_set_aux(md, argc, argv); + else + return 2; /* this wasn't a stats message */ + + if (r == -EINVAL) + DMWARN("Invalid parameters for message %s", argv[0]); + + return r; +} + +int __init dm_statistics_init(void) +{ + shared_memory_amount = 0; + dm_stat_need_rcu_barrier = 0; + return 0; +} + +void dm_statistics_exit(void) +{ + if (dm_stat_need_rcu_barrier) + rcu_barrier(); + if (WARN_ON(shared_memory_amount)) + DMCRIT("shared_memory_amount leaked: %lu", shared_memory_amount); +} + +module_param_named(stats_current_allocated_bytes, shared_memory_amount, ulong, S_IRUGO); +MODULE_PARM_DESC(stats_current_allocated_bytes, "Memory currently used by statistics"); diff --git a/drivers/md/dm-stats.h b/drivers/md/dm-stats.h new file mode 100644 index 00000000000..e7c4984bf23 --- /dev/null +++ b/drivers/md/dm-stats.h @@ -0,0 +1,40 @@ +#ifndef DM_STATS_H +#define DM_STATS_H + +#include <linux/types.h> +#include <linux/mutex.h> +#include <linux/list.h> + +int dm_statistics_init(void); +void dm_statistics_exit(void); + +struct dm_stats { + struct mutex mutex; + struct list_head list; /* list of struct dm_stat */ + struct dm_stats_last_position __percpu *last; + sector_t last_sector; + unsigned last_rw; +}; + +struct dm_stats_aux { + bool merged; +}; + +void dm_stats_init(struct dm_stats *st); +void dm_stats_cleanup(struct dm_stats *st); + +struct mapped_device; + +int dm_stats_message(struct mapped_device *md, unsigned argc, char **argv, + char *result, unsigned maxlen); + +void dm_stats_account_io(struct dm_stats *stats, unsigned long bi_rw, + sector_t bi_sector, unsigned bi_sectors, bool end, + unsigned long duration, struct dm_stats_aux *aux); + +static inline bool dm_stats_used(struct dm_stats *st) +{ + return !list_empty(&st->list); +} + +#endif diff --git a/drivers/md/dm-stripe.c b/drivers/md/dm-stripe.c index d907ca6227c..d1600d2aa2e 100644 --- a/drivers/md/dm-stripe.c +++ b/drivers/md/dm-stripe.c @@ -4,6 +4,7 @@ * This file is released under the GPL. */ +#include "dm.h" #include <linux/device-mapper.h> #include <linux/module.h> @@ -258,13 +259,15 @@ static int stripe_map_range(struct stripe_c *sc, struct bio *bio, { sector_t begin, end; - stripe_map_range_sector(sc, bio->bi_sector, target_stripe, &begin); + stripe_map_range_sector(sc, bio->bi_iter.bi_sector, + target_stripe, &begin); stripe_map_range_sector(sc, bio_end_sector(bio), target_stripe, &end); if (begin < end) { bio->bi_bdev = sc->stripe[target_stripe].dev->bdev; - bio->bi_sector = begin + sc->stripe[target_stripe].physical_start; - bio->bi_size = to_bytes(end - begin); + bio->bi_iter.bi_sector = begin + + sc->stripe[target_stripe].physical_start; + bio->bi_iter.bi_size = to_bytes(end - begin); return DM_MAPIO_REMAPPED; } else { /* The range doesn't map to the target stripe */ @@ -292,9 +295,10 @@ static int stripe_map(struct dm_target *ti, struct bio *bio) return stripe_map_range(sc, bio, target_bio_nr); } - stripe_map_sector(sc, bio->bi_sector, &stripe, &bio->bi_sector); + stripe_map_sector(sc, bio->bi_iter.bi_sector, + &stripe, &bio->bi_iter.bi_sector); - bio->bi_sector += sc->stripe[stripe].physical_start; + bio->bi_iter.bi_sector += sc->stripe[stripe].physical_start; bio->bi_bdev = sc->stripe[stripe].dev->bdev; return DM_MAPIO_REMAPPED; diff --git a/drivers/md/dm-switch.c b/drivers/md/dm-switch.c new file mode 100644 index 00000000000..09a688b3d48 --- /dev/null +++ b/drivers/md/dm-switch.c @@ -0,0 +1,538 @@ +/* + * Copyright (C) 2010-2012 by Dell Inc. All rights reserved. + * Copyright (C) 2011-2013 Red Hat, Inc. + * + * This file is released under the GPL. + * + * dm-switch is a device-mapper target that maps IO to underlying block + * devices efficiently when there are a large number of fixed-sized + * address regions but there is no simple pattern to allow for a compact + * mapping representation such as dm-stripe. + */ + +#include <linux/device-mapper.h> + +#include <linux/module.h> +#include <linux/init.h> +#include <linux/vmalloc.h> + +#define DM_MSG_PREFIX "switch" + +/* + * One region_table_slot_t holds <region_entries_per_slot> region table + * entries each of which is <region_table_entry_bits> in size. + */ +typedef unsigned long region_table_slot_t; + +/* + * A device with the offset to its start sector. + */ +struct switch_path { + struct dm_dev *dmdev; + sector_t start; +}; + +/* + * Context block for a dm switch device. + */ +struct switch_ctx { + struct dm_target *ti; + + unsigned nr_paths; /* Number of paths in path_list. */ + + unsigned region_size; /* Region size in 512-byte sectors */ + unsigned long nr_regions; /* Number of regions making up the device */ + signed char region_size_bits; /* log2 of region_size or -1 */ + + unsigned char region_table_entry_bits; /* Number of bits in one region table entry */ + unsigned char region_entries_per_slot; /* Number of entries in one region table slot */ + signed char region_entries_per_slot_bits; /* log2 of region_entries_per_slot or -1 */ + + region_table_slot_t *region_table; /* Region table */ + + /* + * Array of dm devices to switch between. + */ + struct switch_path path_list[0]; +}; + +static struct switch_ctx *alloc_switch_ctx(struct dm_target *ti, unsigned nr_paths, + unsigned region_size) +{ + struct switch_ctx *sctx; + + sctx = kzalloc(sizeof(struct switch_ctx) + nr_paths * sizeof(struct switch_path), + GFP_KERNEL); + if (!sctx) + return NULL; + + sctx->ti = ti; + sctx->region_size = region_size; + + ti->private = sctx; + + return sctx; +} + +static int alloc_region_table(struct dm_target *ti, unsigned nr_paths) +{ + struct switch_ctx *sctx = ti->private; + sector_t nr_regions = ti->len; + sector_t nr_slots; + + if (!(sctx->region_size & (sctx->region_size - 1))) + sctx->region_size_bits = __ffs(sctx->region_size); + else + sctx->region_size_bits = -1; + + sctx->region_table_entry_bits = 1; + while (sctx->region_table_entry_bits < sizeof(region_table_slot_t) * 8 && + (region_table_slot_t)1 << sctx->region_table_entry_bits < nr_paths) + sctx->region_table_entry_bits++; + + sctx->region_entries_per_slot = (sizeof(region_table_slot_t) * 8) / sctx->region_table_entry_bits; + if (!(sctx->region_entries_per_slot & (sctx->region_entries_per_slot - 1))) + sctx->region_entries_per_slot_bits = __ffs(sctx->region_entries_per_slot); + else + sctx->region_entries_per_slot_bits = -1; + + if (sector_div(nr_regions, sctx->region_size)) + nr_regions++; + + sctx->nr_regions = nr_regions; + if (sctx->nr_regions != nr_regions || sctx->nr_regions >= ULONG_MAX) { + ti->error = "Region table too large"; + return -EINVAL; + } + + nr_slots = nr_regions; + if (sector_div(nr_slots, sctx->region_entries_per_slot)) + nr_slots++; + + if (nr_slots > ULONG_MAX / sizeof(region_table_slot_t)) { + ti->error = "Region table too large"; + return -EINVAL; + } + + sctx->region_table = vmalloc(nr_slots * sizeof(region_table_slot_t)); + if (!sctx->region_table) { + ti->error = "Cannot allocate region table"; + return -ENOMEM; + } + + return 0; +} + +static void switch_get_position(struct switch_ctx *sctx, unsigned long region_nr, + unsigned long *region_index, unsigned *bit) +{ + if (sctx->region_entries_per_slot_bits >= 0) { + *region_index = region_nr >> sctx->region_entries_per_slot_bits; + *bit = region_nr & (sctx->region_entries_per_slot - 1); + } else { + *region_index = region_nr / sctx->region_entries_per_slot; + *bit = region_nr % sctx->region_entries_per_slot; + } + + *bit *= sctx->region_table_entry_bits; +} + +/* + * Find which path to use at given offset. + */ +static unsigned switch_get_path_nr(struct switch_ctx *sctx, sector_t offset) +{ + unsigned long region_index; + unsigned bit, path_nr; + sector_t p; + + p = offset; + if (sctx->region_size_bits >= 0) + p >>= sctx->region_size_bits; + else + sector_div(p, sctx->region_size); + + switch_get_position(sctx, p, ®ion_index, &bit); + path_nr = (ACCESS_ONCE(sctx->region_table[region_index]) >> bit) & + ((1 << sctx->region_table_entry_bits) - 1); + + /* This can only happen if the processor uses non-atomic stores. */ + if (unlikely(path_nr >= sctx->nr_paths)) + path_nr = 0; + + return path_nr; +} + +static void switch_region_table_write(struct switch_ctx *sctx, unsigned long region_nr, + unsigned value) +{ + unsigned long region_index; + unsigned bit; + region_table_slot_t pte; + + switch_get_position(sctx, region_nr, ®ion_index, &bit); + + pte = sctx->region_table[region_index]; + pte &= ~((((region_table_slot_t)1 << sctx->region_table_entry_bits) - 1) << bit); + pte |= (region_table_slot_t)value << bit; + sctx->region_table[region_index] = pte; +} + +/* + * Fill the region table with an initial round robin pattern. + */ +static void initialise_region_table(struct switch_ctx *sctx) +{ + unsigned path_nr = 0; + unsigned long region_nr; + + for (region_nr = 0; region_nr < sctx->nr_regions; region_nr++) { + switch_region_table_write(sctx, region_nr, path_nr); + if (++path_nr >= sctx->nr_paths) + path_nr = 0; + } +} + +static int parse_path(struct dm_arg_set *as, struct dm_target *ti) +{ + struct switch_ctx *sctx = ti->private; + unsigned long long start; + int r; + + r = dm_get_device(ti, dm_shift_arg(as), dm_table_get_mode(ti->table), + &sctx->path_list[sctx->nr_paths].dmdev); + if (r) { + ti->error = "Device lookup failed"; + return r; + } + + if (kstrtoull(dm_shift_arg(as), 10, &start) || start != (sector_t)start) { + ti->error = "Invalid device starting offset"; + dm_put_device(ti, sctx->path_list[sctx->nr_paths].dmdev); + return -EINVAL; + } + + sctx->path_list[sctx->nr_paths].start = start; + + sctx->nr_paths++; + + return 0; +} + +/* + * Destructor: Don't free the dm_target, just the ti->private data (if any). + */ +static void switch_dtr(struct dm_target *ti) +{ + struct switch_ctx *sctx = ti->private; + + while (sctx->nr_paths--) + dm_put_device(ti, sctx->path_list[sctx->nr_paths].dmdev); + + vfree(sctx->region_table); + kfree(sctx); +} + +/* + * Constructor arguments: + * <num_paths> <region_size> <num_optional_args> [<optional_args>...] + * [<dev_path> <offset>]+ + * + * Optional args are to allow for future extension: currently this + * parameter must be 0. + */ +static int switch_ctr(struct dm_target *ti, unsigned argc, char **argv) +{ + static struct dm_arg _args[] = { + {1, (KMALLOC_MAX_SIZE - sizeof(struct switch_ctx)) / sizeof(struct switch_path), "Invalid number of paths"}, + {1, UINT_MAX, "Invalid region size"}, + {0, 0, "Invalid number of optional args"}, + }; + + struct switch_ctx *sctx; + struct dm_arg_set as; + unsigned nr_paths, region_size, nr_optional_args; + int r; + + as.argc = argc; + as.argv = argv; + + r = dm_read_arg(_args, &as, &nr_paths, &ti->error); + if (r) + return -EINVAL; + + r = dm_read_arg(_args + 1, &as, ®ion_size, &ti->error); + if (r) + return r; + + r = dm_read_arg_group(_args + 2, &as, &nr_optional_args, &ti->error); + if (r) + return r; + /* parse optional arguments here, if we add any */ + + if (as.argc != nr_paths * 2) { + ti->error = "Incorrect number of path arguments"; + return -EINVAL; + } + + sctx = alloc_switch_ctx(ti, nr_paths, region_size); + if (!sctx) { + ti->error = "Cannot allocate redirection context"; + return -ENOMEM; + } + + r = dm_set_target_max_io_len(ti, region_size); + if (r) + goto error; + + while (as.argc) { + r = parse_path(&as, ti); + if (r) + goto error; + } + + r = alloc_region_table(ti, nr_paths); + if (r) + goto error; + + initialise_region_table(sctx); + + /* For UNMAP, sending the request down any path is sufficient */ + ti->num_discard_bios = 1; + + return 0; + +error: + switch_dtr(ti); + + return r; +} + +static int switch_map(struct dm_target *ti, struct bio *bio) +{ + struct switch_ctx *sctx = ti->private; + sector_t offset = dm_target_offset(ti, bio->bi_iter.bi_sector); + unsigned path_nr = switch_get_path_nr(sctx, offset); + + bio->bi_bdev = sctx->path_list[path_nr].dmdev->bdev; + bio->bi_iter.bi_sector = sctx->path_list[path_nr].start + offset; + + return DM_MAPIO_REMAPPED; +} + +/* + * We need to parse hex numbers in the message as quickly as possible. + * + * This table-based hex parser improves performance. + * It improves a time to load 1000000 entries compared to the condition-based + * parser. + * table-based parser condition-based parser + * PA-RISC 0.29s 0.31s + * Opteron 0.0495s 0.0498s + */ +static const unsigned char hex_table[256] = { +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 255, 255, 255, 255, 255, 255, +255, 10, 11, 12, 13, 14, 15, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 10, 11, 12, 13, 14, 15, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, +255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 +}; + +static __always_inline unsigned long parse_hex(const char **string) +{ + unsigned char d; + unsigned long r = 0; + + while ((d = hex_table[(unsigned char)**string]) < 16) { + r = (r << 4) | d; + (*string)++; + } + + return r; +} + +static int process_set_region_mappings(struct switch_ctx *sctx, + unsigned argc, char **argv) +{ + unsigned i; + unsigned long region_index = 0; + + for (i = 1; i < argc; i++) { + unsigned long path_nr; + const char *string = argv[i]; + + if (*string == ':') + region_index++; + else { + region_index = parse_hex(&string); + if (unlikely(*string != ':')) { + DMWARN("invalid set_region_mappings argument: '%s'", argv[i]); + return -EINVAL; + } + } + + string++; + if (unlikely(!*string)) { + DMWARN("invalid set_region_mappings argument: '%s'", argv[i]); + return -EINVAL; + } + + path_nr = parse_hex(&string); + if (unlikely(*string)) { + DMWARN("invalid set_region_mappings argument: '%s'", argv[i]); + return -EINVAL; + } + if (unlikely(region_index >= sctx->nr_regions)) { + DMWARN("invalid set_region_mappings region number: %lu >= %lu", region_index, sctx->nr_regions); + return -EINVAL; + } + if (unlikely(path_nr >= sctx->nr_paths)) { + DMWARN("invalid set_region_mappings device: %lu >= %u", path_nr, sctx->nr_paths); + return -EINVAL; + } + + switch_region_table_write(sctx, region_index, path_nr); + } + + return 0; +} + +/* + * Messages are processed one-at-a-time. + * + * Only set_region_mappings is supported. + */ +static int switch_message(struct dm_target *ti, unsigned argc, char **argv) +{ + static DEFINE_MUTEX(message_mutex); + + struct switch_ctx *sctx = ti->private; + int r = -EINVAL; + + mutex_lock(&message_mutex); + + if (!strcasecmp(argv[0], "set_region_mappings")) + r = process_set_region_mappings(sctx, argc, argv); + else + DMWARN("Unrecognised message received."); + + mutex_unlock(&message_mutex); + + return r; +} + +static void switch_status(struct dm_target *ti, status_type_t type, + unsigned status_flags, char *result, unsigned maxlen) +{ + struct switch_ctx *sctx = ti->private; + unsigned sz = 0; + int path_nr; + + switch (type) { + case STATUSTYPE_INFO: + result[0] = '\0'; + break; + + case STATUSTYPE_TABLE: + DMEMIT("%u %u 0", sctx->nr_paths, sctx->region_size); + for (path_nr = 0; path_nr < sctx->nr_paths; path_nr++) + DMEMIT(" %s %llu", sctx->path_list[path_nr].dmdev->name, + (unsigned long long)sctx->path_list[path_nr].start); + break; + } +} + +/* + * Switch ioctl: + * + * Passthrough all ioctls to the path for sector 0 + */ +static int switch_ioctl(struct dm_target *ti, unsigned cmd, + unsigned long arg) +{ + struct switch_ctx *sctx = ti->private; + struct block_device *bdev; + fmode_t mode; + unsigned path_nr; + int r = 0; + + path_nr = switch_get_path_nr(sctx, 0); + + bdev = sctx->path_list[path_nr].dmdev->bdev; + mode = sctx->path_list[path_nr].dmdev->mode; + + /* + * Only pass ioctls through if the device sizes match exactly. + */ + if (ti->len + sctx->path_list[path_nr].start != i_size_read(bdev->bd_inode) >> SECTOR_SHIFT) + r = scsi_verify_blk_ioctl(NULL, cmd); + + return r ? : __blkdev_driver_ioctl(bdev, mode, cmd, arg); +} + +static int switch_iterate_devices(struct dm_target *ti, + iterate_devices_callout_fn fn, void *data) +{ + struct switch_ctx *sctx = ti->private; + int path_nr; + int r; + + for (path_nr = 0; path_nr < sctx->nr_paths; path_nr++) { + r = fn(ti, sctx->path_list[path_nr].dmdev, + sctx->path_list[path_nr].start, ti->len, data); + if (r) + return r; + } + + return 0; +} + +static struct target_type switch_target = { + .name = "switch", + .version = {1, 0, 0}, + .module = THIS_MODULE, + .ctr = switch_ctr, + .dtr = switch_dtr, + .map = switch_map, + .message = switch_message, + .status = switch_status, + .ioctl = switch_ioctl, + .iterate_devices = switch_iterate_devices, +}; + +static int __init dm_switch_init(void) +{ + int r; + + r = dm_register_target(&switch_target); + if (r < 0) + DMERR("dm_register_target() failed %d", r); + + return r; +} + +static void __exit dm_switch_exit(void) +{ + dm_unregister_target(&switch_target); +} + +module_init(dm_switch_init); +module_exit(dm_switch_exit); + +MODULE_DESCRIPTION(DM_NAME " dynamic path switching target"); +MODULE_AUTHOR("Kevin D. O'Kelley <Kevin_OKelley@dell.com>"); +MODULE_AUTHOR("Narendran Ganapathy <Narendran_Ganapathy@dell.com>"); +MODULE_AUTHOR("Jim Ramsay <Jim_Ramsay@dell.com>"); +MODULE_AUTHOR("Mikulas Patocka <mpatocka@redhat.com>"); +MODULE_LICENSE("GPL"); diff --git a/drivers/md/dm-sysfs.c b/drivers/md/dm-sysfs.c index 84d2b91e4ef..c62c5ab6aed 100644 --- a/drivers/md/dm-sysfs.c +++ b/drivers/md/dm-sysfs.c @@ -86,6 +86,7 @@ static const struct sysfs_ops dm_sysfs_ops = { static struct kobj_type dm_ktype = { .sysfs_ops = &dm_sysfs_ops, .default_attrs = dm_attrs, + .release = dm_kobject_release, }; /* @@ -104,5 +105,7 @@ int dm_sysfs_init(struct mapped_device *md) */ void dm_sysfs_exit(struct mapped_device *md) { - kobject_put(dm_kobject(md)); + struct kobject *kobj = dm_kobject(md); + kobject_put(kobj); + wait_for_completion(dm_get_completion_from_kobject(kobj)); } diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c index 1ff252ab7d4..5f59f1e3e5b 100644 --- a/drivers/md/dm-table.c +++ b/drivers/md/dm-table.c @@ -26,22 +26,8 @@ #define KEYS_PER_NODE (NODE_SIZE / sizeof(sector_t)) #define CHILDREN_PER_NODE (KEYS_PER_NODE + 1) -/* - * The table has always exactly one reference from either mapped_device->map - * or hash_cell->new_map. This reference is not counted in table->holders. - * A pair of dm_create_table/dm_destroy_table functions is used for table - * creation/destruction. - * - * Temporary references from the other code increase table->holders. A pair - * of dm_table_get/dm_table_put functions is used to manipulate it. - * - * When the table is about to be destroyed, we wait for table->holders to - * drop to zero. - */ - struct dm_table { struct mapped_device *md; - atomic_t holders; unsigned type; /* btree table */ @@ -169,7 +155,6 @@ static int alloc_targets(struct dm_table *t, unsigned int num) { sector_t *n_highs; struct dm_target *n_targets; - int n = t->num_targets; /* * Allocate both the target array and offset array at once. @@ -183,12 +168,7 @@ static int alloc_targets(struct dm_table *t, unsigned int num) n_targets = (struct dm_target *) (n_highs + num); - if (n) { - memcpy(n_highs, t->highs, sizeof(*n_highs) * n); - memcpy(n_targets, t->targets, sizeof(*n_targets) * n); - } - - memset(n_highs + n, -1, sizeof(*n_highs) * (num - n)); + memset(n_highs, -1, sizeof(*n_highs) * num); vfree(t->highs); t->num_allocated = num; @@ -208,13 +188,17 @@ int dm_table_create(struct dm_table **result, fmode_t mode, INIT_LIST_HEAD(&t->devices); INIT_LIST_HEAD(&t->target_callbacks); - atomic_set(&t->holders, 0); if (!num_targets) num_targets = KEYS_PER_NODE; num_targets = dm_round_up(num_targets, KEYS_PER_NODE); + if (!num_targets) { + kfree(t); + return -ENOMEM; + } + if (alloc_targets(t, num_targets)) { kfree(t); return -ENOMEM; @@ -246,10 +230,6 @@ void dm_table_destroy(struct dm_table *t) if (!t) return; - while (atomic_read(&t->holders)) - msleep(1); - smp_mb(); - /* free the indexes */ if (t->depth >= 2) vfree(t->index[t->depth - 2]); @@ -274,33 +254,6 @@ void dm_table_destroy(struct dm_table *t) kfree(t); } -void dm_table_get(struct dm_table *t) -{ - atomic_inc(&t->holders); -} -EXPORT_SYMBOL(dm_table_get); - -void dm_table_put(struct dm_table *t) -{ - if (!t) - return; - - smp_mb__before_atomic_dec(); - atomic_dec(&t->holders); -} -EXPORT_SYMBOL(dm_table_put); - -/* - * Checks to see if we need to extend highs or targets. - */ -static inline int check_space(struct dm_table *t) -{ - if (t->num_targets >= t->num_allocated) - return alloc_targets(t, t->num_allocated * 2); - - return 0; -} - /* * See if we've already got a device in the list. */ @@ -512,8 +465,8 @@ int dm_get_device(struct dm_target *ti, const char *path, fmode_t mode, } EXPORT_SYMBOL(dm_get_device); -int dm_set_device_limits(struct dm_target *ti, struct dm_dev *dev, - sector_t start, sector_t len, void *data) +static int dm_set_device_limits(struct dm_target *ti, struct dm_dev *dev, + sector_t start, sector_t len, void *data) { struct queue_limits *limits = data; struct block_device *bdev = dev->bdev; @@ -546,7 +499,6 @@ int dm_set_device_limits(struct dm_target *ti, struct dm_dev *dev, (unsigned int) (PAGE_SIZE >> 9)); return 0; } -EXPORT_SYMBOL_GPL(dm_set_device_limits); /* * Decrement a device's use count and remove it if necessary. @@ -580,14 +532,28 @@ static int adjoin(struct dm_table *table, struct dm_target *ti) /* * Used to dynamically allocate the arg array. + * + * We do first allocation with GFP_NOIO because dm-mpath and dm-thin must + * process messages even if some device is suspended. These messages have a + * small fixed number of arguments. + * + * On the other hand, dm-switch needs to process bulk data using messages and + * excessive use of GFP_NOIO could cause trouble. */ static char **realloc_argv(unsigned *array_size, char **old_argv) { char **argv; unsigned new_size; + gfp_t gfp; - new_size = *array_size ? *array_size * 2 : 64; - argv = kmalloc(new_size * sizeof(*argv), GFP_KERNEL); + if (*array_size) { + new_size = *array_size * 2; + gfp = GFP_KERNEL; + } else { + new_size = 8; + gfp = GFP_NOIO; + } + argv = kmalloc(new_size * sizeof(*argv), gfp); if (argv) { memcpy(argv, old_argv, *array_size * sizeof(*argv)); *array_size = new_size; @@ -747,8 +713,7 @@ int dm_table_add_target(struct dm_table *t, const char *type, return -EINVAL; } - if ((r = check_space(t))) - return r; + BUG_ON(t->num_targets >= t->num_allocated); tgt = t->targets + t->num_targets; memset(tgt, 0, sizeof(*tgt)); @@ -895,14 +860,17 @@ EXPORT_SYMBOL(dm_consume_args); static int dm_table_set_type(struct dm_table *t) { unsigned i; - unsigned bio_based = 0, request_based = 0; + unsigned bio_based = 0, request_based = 0, hybrid = 0; struct dm_target *tgt; struct dm_dev_internal *dd; struct list_head *devices; + unsigned live_md_type; for (i = 0; i < t->num_targets; i++) { tgt = t->targets + i; - if (dm_target_request_based(tgt)) + if (dm_target_hybrid(tgt)) + hybrid = 1; + else if (dm_target_request_based(tgt)) request_based = 1; else bio_based = 1; @@ -914,6 +882,19 @@ static int dm_table_set_type(struct dm_table *t) } } + if (hybrid && !bio_based && !request_based) { + /* + * The targets can work either way. + * Determine the type from the live device. + * Default to bio-based if device is new. + */ + live_md_type = dm_get_md_type(t->md); + if (live_md_type == DM_TYPE_REQUEST_BASED) + request_based = 1; + else + bio_based = 1; + } + if (bio_based) { /* We must use this table as bio-based */ t->type = DM_TYPE_BIO_BASED; @@ -963,7 +944,7 @@ bool dm_table_request_based(struct dm_table *t) return dm_table_get_type(t) == DM_TYPE_REQUEST_BASED; } -int dm_table_alloc_md_mempools(struct dm_table *t) +static int dm_table_alloc_md_mempools(struct dm_table *t) { unsigned type = dm_table_get_type(t); unsigned per_bio_data_size = 0; @@ -1567,8 +1548,11 @@ int dm_table_resume_targets(struct dm_table *t) continue; r = ti->type->preresume(ti); - if (r) + if (r) { + DMERR("%s: %s: preresume failed, error = %d", + dm_device_name(t->md), ti->type->name, r); return r; + } } for (i = 0; i < t->num_targets; i++) { @@ -1633,6 +1617,25 @@ struct mapped_device *dm_table_get_md(struct dm_table *t) } EXPORT_SYMBOL(dm_table_get_md); +void dm_table_run_md_queue_async(struct dm_table *t) +{ + struct mapped_device *md; + struct request_queue *queue; + unsigned long flags; + + if (!dm_table_request_based(t)) + return; + + md = dm_table_get_md(t); + queue = dm_get_md_queue(md); + if (queue) { + spin_lock_irqsave(queue->queue_lock, flags); + blk_run_queue_async(queue); + spin_unlock_irqrestore(queue->queue_lock, flags); + } +} +EXPORT_SYMBOL(dm_table_run_md_queue_async); + static int device_discard_capable(struct dm_target *ti, struct dm_dev *dev, sector_t start, sector_t len, void *data) { diff --git a/drivers/md/dm-target.c b/drivers/md/dm-target.c index 37ba5db71cd..242e3cec397 100644 --- a/drivers/md/dm-target.c +++ b/drivers/md/dm-target.c @@ -131,12 +131,19 @@ static int io_err_map(struct dm_target *tt, struct bio *bio) return -EIO; } +static int io_err_map_rq(struct dm_target *ti, struct request *clone, + union map_info *map_context) +{ + return -EIO; +} + static struct target_type error_target = { .name = "error", - .version = {1, 1, 0}, + .version = {1, 2, 0}, .ctr = io_err_ctr, .dtr = io_err_dtr, .map = io_err_map, + .map_rq = io_err_map_rq, }; int __init dm_target_init(void) diff --git a/drivers/md/dm-thin-metadata.c b/drivers/md/dm-thin-metadata.c index 60bce435f4f..e9d33ad59df 100644 --- a/drivers/md/dm-thin-metadata.c +++ b/drivers/md/dm-thin-metadata.c @@ -76,7 +76,7 @@ #define THIN_SUPERBLOCK_MAGIC 27022010 #define THIN_SUPERBLOCK_LOCATION 0 -#define THIN_VERSION 1 +#define THIN_VERSION 2 #define THIN_METADATA_CACHE_SIZE 64 #define SECTOR_TO_BLOCK_SHIFT 3 @@ -192,6 +192,13 @@ struct dm_pool_metadata { * operation possible in this state is the closing of the device. */ bool fail_io:1; + + /* + * Reading the space map roots can fail, so we read it into these + * buffers before the superblock is locked and updated. + */ + __u8 data_space_map_root[SPACE_MAP_ROOT_SIZE]; + __u8 metadata_space_map_root[SPACE_MAP_ROOT_SIZE]; }; struct dm_thin_device { @@ -431,26 +438,53 @@ static void __setup_btree_details(struct dm_pool_metadata *pmd) pmd->details_info.value_type.equal = NULL; } +static int save_sm_roots(struct dm_pool_metadata *pmd) +{ + int r; + size_t len; + + r = dm_sm_root_size(pmd->metadata_sm, &len); + if (r < 0) + return r; + + r = dm_sm_copy_root(pmd->metadata_sm, &pmd->metadata_space_map_root, len); + if (r < 0) + return r; + + r = dm_sm_root_size(pmd->data_sm, &len); + if (r < 0) + return r; + + return dm_sm_copy_root(pmd->data_sm, &pmd->data_space_map_root, len); +} + +static void copy_sm_roots(struct dm_pool_metadata *pmd, + struct thin_disk_superblock *disk) +{ + memcpy(&disk->metadata_space_map_root, + &pmd->metadata_space_map_root, + sizeof(pmd->metadata_space_map_root)); + + memcpy(&disk->data_space_map_root, + &pmd->data_space_map_root, + sizeof(pmd->data_space_map_root)); +} + static int __write_initial_superblock(struct dm_pool_metadata *pmd) { int r; struct dm_block *sblock; - size_t metadata_len, data_len; struct thin_disk_superblock *disk_super; sector_t bdev_size = i_size_read(pmd->bdev->bd_inode) >> SECTOR_SHIFT; if (bdev_size > THIN_METADATA_MAX_SECTORS) bdev_size = THIN_METADATA_MAX_SECTORS; - r = dm_sm_root_size(pmd->metadata_sm, &metadata_len); - if (r < 0) - return r; - - r = dm_sm_root_size(pmd->data_sm, &data_len); + r = dm_sm_commit(pmd->data_sm); if (r < 0) return r; - r = dm_sm_commit(pmd->data_sm); + r = save_sm_roots(pmd); if (r < 0) return r; @@ -471,27 +505,15 @@ static int __write_initial_superblock(struct dm_pool_metadata *pmd) disk_super->trans_id = 0; disk_super->held_root = 0; - r = dm_sm_copy_root(pmd->metadata_sm, &disk_super->metadata_space_map_root, - metadata_len); - if (r < 0) - goto bad_locked; - - r = dm_sm_copy_root(pmd->data_sm, &disk_super->data_space_map_root, - data_len); - if (r < 0) - goto bad_locked; + copy_sm_roots(pmd, disk_super); disk_super->data_mapping_root = cpu_to_le64(pmd->root); disk_super->device_details_root = cpu_to_le64(pmd->details_root); - disk_super->metadata_block_size = cpu_to_le32(THIN_METADATA_BLOCK_SIZE >> SECTOR_SHIFT); + disk_super->metadata_block_size = cpu_to_le32(THIN_METADATA_BLOCK_SIZE); disk_super->metadata_nr_blocks = cpu_to_le64(bdev_size >> SECTOR_TO_BLOCK_SHIFT); disk_super->data_block_size = cpu_to_le32(pmd->data_block_size); return dm_tm_commit(pmd->tm, sblock); - -bad_locked: - dm_bm_unlock(sblock); - return r; } static int __format_metadata(struct dm_pool_metadata *pmd) @@ -591,6 +613,15 @@ static int __open_metadata(struct dm_pool_metadata *pmd) disk_super = dm_block_data(sblock); + /* Verify the data block size hasn't changed */ + if (le32_to_cpu(disk_super->data_block_size) != pmd->data_block_size) { + DMERR("changing the data block size (from %u to %llu) is not supported", + le32_to_cpu(disk_super->data_block_size), + (unsigned long long)pmd->data_block_size); + r = -EINVAL; + goto bad_unlock_sblock; + } + r = __check_incompat_features(disk_super, pmd); if (r < 0) goto bad_unlock_sblock; @@ -651,7 +682,7 @@ static int __create_persistent_data_objects(struct dm_pool_metadata *pmd, bool f { int r; - pmd->bm = dm_block_manager_create(pmd->bdev, THIN_METADATA_BLOCK_SIZE, + pmd->bm = dm_block_manager_create(pmd->bdev, THIN_METADATA_BLOCK_SIZE << SECTOR_SHIFT, THIN_METADATA_CACHE_SIZE, THIN_MAX_CONCURRENT_LOCKS); if (IS_ERR(pmd->bm)) { @@ -769,6 +800,10 @@ static int __commit_transaction(struct dm_pool_metadata *pmd) if (r < 0) return r; + r = save_sm_roots(pmd); + if (r < 0) + return r; + r = superblock_lock(pmd, &sblock); if (r) return r; @@ -780,21 +815,9 @@ static int __commit_transaction(struct dm_pool_metadata *pmd) disk_super->trans_id = cpu_to_le64(pmd->trans_id); disk_super->flags = cpu_to_le32(pmd->flags); - r = dm_sm_copy_root(pmd->metadata_sm, &disk_super->metadata_space_map_root, - metadata_len); - if (r < 0) - goto out_locked; - - r = dm_sm_copy_root(pmd->data_sm, &disk_super->data_space_map_root, - data_len); - if (r < 0) - goto out_locked; + copy_sm_roots(pmd, disk_super); return dm_tm_commit(pmd->tm, sblock); - -out_locked: - dm_bm_unlock(sblock); - return r; } struct dm_pool_metadata *dm_pool_metadata_open(struct block_device *bdev, @@ -1349,6 +1372,12 @@ dm_thin_id dm_thin_dev_id(struct dm_thin_device *td) return td->id; } +/* + * Check whether @time (of block creation) is older than @td's last snapshot. + * If so then the associated block is shared with the last snapshot device. + * Any block on a device created *after* the device last got snapshotted is + * necessarily not shared. + */ static bool __snapshotted_since(struct dm_thin_device *td, uint32_t time) { return td->snapshotted_time > time; @@ -1458,6 +1487,20 @@ int dm_thin_remove_block(struct dm_thin_device *td, dm_block_t block) return r; } +int dm_pool_block_is_used(struct dm_pool_metadata *pmd, dm_block_t b, bool *result) +{ + int r; + uint32_t ref_count; + + down_read(&pmd->root_lock); + r = dm_sm_get_count(pmd->data_sm, b, &ref_count); + if (!r) + *result = (ref_count != 0); + up_read(&pmd->root_lock); + + return r; +} + bool dm_thin_changed_this_transaction(struct dm_thin_device *td) { int r; @@ -1469,6 +1512,23 @@ bool dm_thin_changed_this_transaction(struct dm_thin_device *td) return r; } +bool dm_pool_changed_this_transaction(struct dm_pool_metadata *pmd) +{ + bool r = false; + struct dm_thin_device *td, *tmp; + + down_read(&pmd->root_lock); + list_for_each_entry_safe(td, tmp, &pmd->thin_devices, list) { + if (td->changed) { + r = td->changed; + break; + } + } + up_read(&pmd->root_lock); + + return r; +} + bool dm_thin_aborted_changes(struct dm_thin_device *td) { bool r; @@ -1697,6 +1757,14 @@ void dm_pool_metadata_read_only(struct dm_pool_metadata *pmd) up_write(&pmd->root_lock); } +void dm_pool_metadata_read_write(struct dm_pool_metadata *pmd) +{ + down_write(&pmd->root_lock); + pmd->read_only = false; + dm_bm_set_read_write(pmd->bm); + up_write(&pmd->root_lock); +} + int dm_pool_register_metadata_threshold(struct dm_pool_metadata *pmd, dm_block_t threshold, dm_sm_threshold_fn fn, @@ -1710,3 +1778,38 @@ int dm_pool_register_metadata_threshold(struct dm_pool_metadata *pmd, return r; } + +int dm_pool_metadata_set_needs_check(struct dm_pool_metadata *pmd) +{ + int r; + struct dm_block *sblock; + struct thin_disk_superblock *disk_super; + + down_write(&pmd->root_lock); + pmd->flags |= THIN_METADATA_NEEDS_CHECK_FLAG; + + r = superblock_lock(pmd, &sblock); + if (r) { + DMERR("couldn't read superblock"); + goto out; + } + + disk_super = dm_block_data(sblock); + disk_super->flags = cpu_to_le32(pmd->flags); + + dm_bm_unlock(sblock); +out: + up_write(&pmd->root_lock); + return r; +} + +bool dm_pool_metadata_needs_check(struct dm_pool_metadata *pmd) +{ + bool needs_check; + + down_read(&pmd->root_lock); + needs_check = pmd->flags & THIN_METADATA_NEEDS_CHECK_FLAG; + up_read(&pmd->root_lock); + + return needs_check; +} diff --git a/drivers/md/dm-thin-metadata.h b/drivers/md/dm-thin-metadata.h index 845ebbe589a..e3c857db195 100644 --- a/drivers/md/dm-thin-metadata.h +++ b/drivers/md/dm-thin-metadata.h @@ -9,16 +9,14 @@ #include "persistent-data/dm-block-manager.h" #include "persistent-data/dm-space-map.h" +#include "persistent-data/dm-space-map-metadata.h" -#define THIN_METADATA_BLOCK_SIZE 4096 +#define THIN_METADATA_BLOCK_SIZE DM_SM_METADATA_BLOCK_SIZE /* * The metadata device is currently limited in size. - * - * We have one block of index, which can hold 255 index entries. Each - * index entry contains allocation info about 16k metadata blocks. */ -#define THIN_METADATA_MAX_SECTORS (255 * (1 << 14) * (THIN_METADATA_BLOCK_SIZE / (1 << SECTOR_SHIFT))) +#define THIN_METADATA_MAX_SECTORS DM_SM_METADATA_MAX_SECTORS /* * A metadata device larger than 16GB triggers a warning. @@ -27,6 +25,11 @@ /*----------------------------------------------------------------*/ +/* + * Thin metadata superblock flags. + */ +#define THIN_METADATA_NEEDS_CHECK_FLAG (1 << 0) + struct dm_pool_metadata; struct dm_thin_device; @@ -131,7 +134,7 @@ dm_thin_id dm_thin_dev_id(struct dm_thin_device *td); struct dm_thin_lookup_result { dm_block_t block; - unsigned shared:1; + bool shared:1; }; /* @@ -161,6 +164,8 @@ int dm_thin_remove_block(struct dm_thin_device *td, dm_block_t block); */ bool dm_thin_changed_this_transaction(struct dm_thin_device *td); +bool dm_pool_changed_this_transaction(struct dm_pool_metadata *pmd); + bool dm_thin_aborted_changes(struct dm_thin_device *td); int dm_thin_get_highest_mapped_block(struct dm_thin_device *td, @@ -181,6 +186,8 @@ int dm_pool_get_data_block_size(struct dm_pool_metadata *pmd, sector_t *result); int dm_pool_get_data_dev_size(struct dm_pool_metadata *pmd, dm_block_t *result); +int dm_pool_block_is_used(struct dm_pool_metadata *pmd, dm_block_t b, bool *result); + /* * Returns -ENOSPC if the new size is too small and already allocated * blocks would be lost. @@ -193,12 +200,19 @@ int dm_pool_resize_metadata_dev(struct dm_pool_metadata *pmd, dm_block_t new_siz * that nothing is changing. */ void dm_pool_metadata_read_only(struct dm_pool_metadata *pmd); +void dm_pool_metadata_read_write(struct dm_pool_metadata *pmd); int dm_pool_register_metadata_threshold(struct dm_pool_metadata *pmd, dm_block_t threshold, dm_sm_threshold_fn fn, void *context); +/* + * Updates the superblock immediately. + */ +int dm_pool_metadata_set_needs_check(struct dm_pool_metadata *pmd); +bool dm_pool_metadata_needs_check(struct dm_pool_metadata *pmd); + /*----------------------------------------------------------------*/ #endif diff --git a/drivers/md/dm-thin.c b/drivers/md/dm-thin.c index 88f2f802d52..fc9c848a60c 100644 --- a/drivers/md/dm-thin.c +++ b/drivers/md/dm-thin.c @@ -12,9 +12,11 @@ #include <linux/dm-io.h> #include <linux/dm-kcopyd.h> #include <linux/list.h> +#include <linux/rculist.h> #include <linux/init.h> #include <linux/module.h> #include <linux/slab.h> +#include <linux/rbtree.h> #define DM_MSG_PREFIX "thin" @@ -25,6 +27,9 @@ #define MAPPING_POOL_SIZE 1024 #define PRISON_CELLS 1024 #define COMMIT_PERIOD HZ +#define NO_SPACE_TIMEOUT_SECS 60 + +static unsigned no_space_timeout_secs = NO_SPACE_TIMEOUT_SECS; DECLARE_DM_KCOPYD_THROTTLE_WITH_MODULE_PARM(snapshot_copy_throttle, "A percentage of time allocated for copy on write"); @@ -130,10 +135,11 @@ static void build_virtual_key(struct dm_thin_device *td, dm_block_t b, struct dm_thin_new_mapping; /* - * The pool runs in 3 modes. Ordered in degraded order for comparisons. + * The pool runs in 4 modes. Ordered in degraded order for comparisons. */ enum pool_mode { PM_WRITE, /* metadata may be changed */ + PM_OUT_OF_DATA_SPACE, /* metadata may be changed, though data may not be allocated */ PM_READ_ONLY, /* metadata may not be changed */ PM_FAIL, /* all I/O fails */ }; @@ -144,6 +150,7 @@ struct pool_features { bool zero_new_blocks:1; bool discard_enabled:1; bool discard_passdown:1; + bool error_if_no_space:1; }; struct thin_c; @@ -163,8 +170,7 @@ struct pool { int sectors_per_block_shift; struct pool_features pf; - unsigned low_water_triggered:1; /* A dm event has been sent */ - unsigned no_free_space:1; /* A -ENOSPC warning has been issued */ + bool low_water_triggered:1; /* A dm event has been sent */ struct dm_bio_prison *prison; struct dm_kcopyd_client *copier; @@ -172,17 +178,16 @@ struct pool { struct workqueue_struct *wq; struct work_struct worker; struct delayed_work waker; + struct delayed_work no_space_timeout; unsigned long last_commit_jiffies; unsigned ref_count; spinlock_t lock; - struct bio_list deferred_bios; struct bio_list deferred_flush_bios; struct list_head prepared_mappings; struct list_head prepared_discards; - - struct bio_list retry_on_resume_list; + struct list_head active_thins; struct dm_deferred_set *shared_read_ds; struct dm_deferred_set *all_io_ds; @@ -198,7 +203,7 @@ struct pool { }; static enum pool_mode get_pool_mode(struct pool *pool); -static void set_pool_mode(struct pool *pool, enum pool_mode mode); +static void metadata_operation_failed(struct pool *pool, const char *op, int r); /* * Target context for a pool. @@ -219,12 +224,25 @@ struct pool_c { * Target context for a thin. */ struct thin_c { + struct list_head list; struct dm_dev *pool_dev; struct dm_dev *origin_dev; dm_thin_id dev_id; struct pool *pool; struct dm_thin_device *td; + bool requeue_mode:1; + spinlock_t lock; + struct bio_list deferred_bio_list; + struct bio_list retry_on_resume_list; + struct rb_root sort_bio_list; /* sorted list of deferred bios */ + + /* + * Ensures the thin is not destroyed until the worker has finished + * iterating the active_thins list. + */ + atomic_t refcount; + struct completion can_destroy; }; /*----------------------------------------------------------------*/ @@ -285,20 +303,25 @@ static void cell_defer_no_holder_no_free(struct thin_c *tc, struct pool *pool = tc->pool; unsigned long flags; - spin_lock_irqsave(&pool->lock, flags); - dm_cell_release_no_holder(pool->prison, cell, &pool->deferred_bios); - spin_unlock_irqrestore(&pool->lock, flags); + spin_lock_irqsave(&tc->lock, flags); + dm_cell_release_no_holder(pool->prison, cell, &tc->deferred_bio_list); + spin_unlock_irqrestore(&tc->lock, flags); wake_worker(pool); } -static void cell_error(struct pool *pool, - struct dm_bio_prison_cell *cell) +static void cell_error_with_code(struct pool *pool, + struct dm_bio_prison_cell *cell, int error_code) { - dm_cell_error(pool->prison, cell); + dm_cell_error(pool->prison, cell, error_code); dm_bio_prison_free_cell(pool->prison, cell); } +static void cell_error(struct pool *pool, struct dm_bio_prison_cell *cell) +{ + cell_error_with_code(pool, cell, -EIO); +} + /*----------------------------------------------------------------*/ /* @@ -366,36 +389,57 @@ struct dm_thin_endio_hook { struct dm_deferred_entry *shared_read_entry; struct dm_deferred_entry *all_io_entry; struct dm_thin_new_mapping *overwrite_mapping; + struct rb_node rb_node; }; -static void __requeue_bio_list(struct thin_c *tc, struct bio_list *master) +static void requeue_bio_list(struct thin_c *tc, struct bio_list *master) { struct bio *bio; struct bio_list bios; + unsigned long flags; bio_list_init(&bios); + + spin_lock_irqsave(&tc->lock, flags); bio_list_merge(&bios, master); bio_list_init(master); + spin_unlock_irqrestore(&tc->lock, flags); - while ((bio = bio_list_pop(&bios))) { - struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook)); - - if (h->tc == tc) - bio_endio(bio, DM_ENDIO_REQUEUE); - else - bio_list_add(master, bio); - } + while ((bio = bio_list_pop(&bios))) + bio_endio(bio, DM_ENDIO_REQUEUE); } static void requeue_io(struct thin_c *tc) { - struct pool *pool = tc->pool; + requeue_bio_list(tc, &tc->deferred_bio_list); + requeue_bio_list(tc, &tc->retry_on_resume_list); +} + +static void error_thin_retry_list(struct thin_c *tc) +{ + struct bio *bio; unsigned long flags; + struct bio_list bios; - spin_lock_irqsave(&pool->lock, flags); - __requeue_bio_list(tc, &pool->deferred_bios); - __requeue_bio_list(tc, &pool->retry_on_resume_list); - spin_unlock_irqrestore(&pool->lock, flags); + bio_list_init(&bios); + + spin_lock_irqsave(&tc->lock, flags); + bio_list_merge(&bios, &tc->retry_on_resume_list); + bio_list_init(&tc->retry_on_resume_list); + spin_unlock_irqrestore(&tc->lock, flags); + + while ((bio = bio_list_pop(&bios))) + bio_io_error(bio); +} + +static void error_retry_list(struct pool *pool) +{ + struct thin_c *tc; + + rcu_read_lock(); + list_for_each_entry_rcu(tc, &pool->active_thins, list) + error_thin_retry_list(tc); + rcu_read_unlock(); } /* @@ -413,7 +457,7 @@ static bool block_size_is_power_of_two(struct pool *pool) static dm_block_t get_bio_block(struct thin_c *tc, struct bio *bio) { struct pool *pool = tc->pool; - sector_t block_nr = bio->bi_sector; + sector_t block_nr = bio->bi_iter.bi_sector; if (block_size_is_power_of_two(pool)) block_nr >>= pool->sectors_per_block_shift; @@ -426,14 +470,15 @@ static dm_block_t get_bio_block(struct thin_c *tc, struct bio *bio) static void remap(struct thin_c *tc, struct bio *bio, dm_block_t block) { struct pool *pool = tc->pool; - sector_t bi_sector = bio->bi_sector; + sector_t bi_sector = bio->bi_iter.bi_sector; bio->bi_bdev = tc->pool_dev->bdev; if (block_size_is_power_of_two(pool)) - bio->bi_sector = (block << pool->sectors_per_block_shift) | - (bi_sector & (pool->sectors_per_block - 1)); + bio->bi_iter.bi_sector = + (block << pool->sectors_per_block_shift) | + (bi_sector & (pool->sectors_per_block - 1)); else - bio->bi_sector = (block * pool->sectors_per_block) + + bio->bi_iter.bi_sector = (block * pool->sectors_per_block) + sector_div(bi_sector, pool->sectors_per_block); } @@ -509,15 +554,16 @@ static void remap_and_issue(struct thin_c *tc, struct bio *bio, struct dm_thin_new_mapping { struct list_head list; - unsigned quiesced:1; - unsigned prepared:1; - unsigned pass_discard:1; + bool quiesced:1; + bool prepared:1; + bool pass_discard:1; + bool definitely_not_shared:1; + int err; struct thin_c *tc; dm_block_t virt_block; dm_block_t data_block; struct dm_bio_prison_cell *cell, *cell2; - int err; /* * If the bio covers the whole area of a block then we can avoid @@ -534,7 +580,7 @@ static void __maybe_add_mapping(struct dm_thin_new_mapping *m) struct pool *pool = m->tc->pool; if (m->quiesced && m->prepared) { - list_add(&m->list, &pool->prepared_mappings); + list_add_tail(&m->list, &pool->prepared_mappings); wake_worker(pool); } } @@ -548,7 +594,7 @@ static void copy_complete(int read_err, unsigned long write_err, void *context) m->err = read_err || write_err ? -EIO : 0; spin_lock_irqsave(&pool->lock, flags); - m->prepared = 1; + m->prepared = true; __maybe_add_mapping(m); spin_unlock_irqrestore(&pool->lock, flags); } @@ -563,7 +609,7 @@ static void overwrite_endio(struct bio *bio, int err) m->err = err; spin_lock_irqsave(&pool->lock, flags); - m->prepared = 1; + m->prepared = true; __maybe_add_mapping(m); spin_unlock_irqrestore(&pool->lock, flags); } @@ -586,9 +632,9 @@ static void cell_defer(struct thin_c *tc, struct dm_bio_prison_cell *cell) struct pool *pool = tc->pool; unsigned long flags; - spin_lock_irqsave(&pool->lock, flags); - cell_release(pool, cell, &pool->deferred_bios); - spin_unlock_irqrestore(&tc->pool->lock, flags); + spin_lock_irqsave(&tc->lock, flags); + cell_release(pool, cell, &tc->deferred_bio_list); + spin_unlock_irqrestore(&tc->lock, flags); wake_worker(pool); } @@ -601,17 +647,19 @@ static void cell_defer_no_holder(struct thin_c *tc, struct dm_bio_prison_cell *c struct pool *pool = tc->pool; unsigned long flags; - spin_lock_irqsave(&pool->lock, flags); - cell_release_no_holder(pool, cell, &pool->deferred_bios); - spin_unlock_irqrestore(&pool->lock, flags); + spin_lock_irqsave(&tc->lock, flags); + cell_release_no_holder(pool, cell, &tc->deferred_bio_list); + spin_unlock_irqrestore(&tc->lock, flags); wake_worker(pool); } static void process_prepared_mapping_fail(struct dm_thin_new_mapping *m) { - if (m->bio) + if (m->bio) { m->bio->bi_end_io = m->saved_bi_end_io; + atomic_inc(&m->bio->bi_remaining); + } cell_error(m->tc->pool, m->cell); list_del(&m->list); mempool_free(m, m->tc->pool->mapping_pool); @@ -625,8 +673,10 @@ static void process_prepared_mapping(struct dm_thin_new_mapping *m) int r; bio = m->bio; - if (bio) + if (bio) { bio->bi_end_io = m->saved_bi_end_io; + atomic_inc(&bio->bi_remaining); + } if (m->err) { cell_error(pool, m->cell); @@ -640,7 +690,7 @@ static void process_prepared_mapping(struct dm_thin_new_mapping *m) */ r = dm_thin_insert_block(tc->td, m->virt_block, m->data_block); if (r) { - DMERR_LIMIT("dm_thin_insert_block() failed"); + metadata_operation_failed(pool, "dm_thin_insert_block", r); cell_error(pool, m->cell); goto out; } @@ -681,7 +731,15 @@ static void process_prepared_discard_passdown(struct dm_thin_new_mapping *m) cell_defer_no_holder(tc, m->cell2); if (m->pass_discard) - remap_and_issue(tc, m->bio, m->data_block); + if (m->definitely_not_shared) + remap_and_issue(tc, m->bio, m->data_block); + else { + bool used = false; + if (dm_pool_block_is_used(tc->pool->pmd, m->data_block, &used) || used) + bio_endio(m->bio, 0); + else + remap_and_issue(tc, m->bio, m->data_block); + } else bio_endio(m->bio, 0); @@ -721,7 +779,8 @@ static void process_prepared(struct pool *pool, struct list_head *head, */ static int io_overlaps_block(struct pool *pool, struct bio *bio) { - return bio->bi_size == (pool->sectors_per_block << SECTOR_SHIFT); + return bio->bi_iter.bi_size == + (pool->sectors_per_block << SECTOR_SHIFT); } static int io_overwrites_block(struct pool *pool, struct bio *bio) @@ -749,13 +808,17 @@ static int ensure_next_mapping(struct pool *pool) static struct dm_thin_new_mapping *get_next_mapping(struct pool *pool) { - struct dm_thin_new_mapping *r = pool->next_mapping; + struct dm_thin_new_mapping *m = pool->next_mapping; BUG_ON(!pool->next_mapping); + memset(m, 0, sizeof(struct dm_thin_new_mapping)); + INIT_LIST_HEAD(&m->list); + m->bio = NULL; + pool->next_mapping = NULL; - return r; + return m; } static void schedule_copy(struct thin_c *tc, dm_block_t virt_block, @@ -767,18 +830,13 @@ static void schedule_copy(struct thin_c *tc, dm_block_t virt_block, struct pool *pool = tc->pool; struct dm_thin_new_mapping *m = get_next_mapping(pool); - INIT_LIST_HEAD(&m->list); - m->quiesced = 0; - m->prepared = 0; m->tc = tc; m->virt_block = virt_block; m->data_block = data_dest; m->cell = cell; - m->err = 0; - m->bio = NULL; if (!dm_deferred_set_add_work(pool->shared_read_ds, &m->list)) - m->quiesced = 1; + m->quiesced = true; /* * IO to pool_dev remaps to the pool target's data_dev. @@ -838,15 +896,12 @@ static void schedule_zero(struct thin_c *tc, dm_block_t virt_block, struct pool *pool = tc->pool; struct dm_thin_new_mapping *m = get_next_mapping(pool); - INIT_LIST_HEAD(&m->list); - m->quiesced = 1; - m->prepared = 0; + m->quiesced = true; + m->prepared = false; m->tc = tc; m->virt_block = virt_block; m->data_block = data_block; m->cell = cell; - m->err = 0; - m->bio = NULL; /* * If the whole block of data is being overwritten or we are not @@ -881,87 +936,83 @@ static void schedule_zero(struct thin_c *tc, dm_block_t virt_block, } } -static int commit(struct pool *pool) -{ - int r; - - r = dm_pool_commit_metadata(pool->pmd); - if (r) - DMERR_LIMIT("commit failed: error = %d", r); - - return r; -} - /* * A non-zero return indicates read_only or fail_io mode. * Many callers don't care about the return value. */ -static int commit_or_fallback(struct pool *pool) +static int commit(struct pool *pool) { int r; - if (get_pool_mode(pool) != PM_WRITE) + if (get_pool_mode(pool) >= PM_READ_ONLY) return -EINVAL; - r = commit(pool); + r = dm_pool_commit_metadata(pool->pmd); if (r) - set_pool_mode(pool, PM_READ_ONLY); + metadata_operation_failed(pool, "dm_pool_commit_metadata", r); return r; } -static int alloc_data_block(struct thin_c *tc, dm_block_t *result) +static void check_low_water_mark(struct pool *pool, dm_block_t free_blocks) { - int r; - dm_block_t free_blocks; unsigned long flags; - struct pool *pool = tc->pool; - - r = dm_pool_get_free_block_count(pool->pmd, &free_blocks); - if (r) - return r; if (free_blocks <= pool->low_water_blocks && !pool->low_water_triggered) { DMWARN("%s: reached low water mark for data device: sending event.", dm_device_name(pool->pool_md)); spin_lock_irqsave(&pool->lock, flags); - pool->low_water_triggered = 1; + pool->low_water_triggered = true; spin_unlock_irqrestore(&pool->lock, flags); dm_table_event(pool->ti->table); } +} + +static void set_pool_mode(struct pool *pool, enum pool_mode new_mode); + +static int alloc_data_block(struct thin_c *tc, dm_block_t *result) +{ + int r; + dm_block_t free_blocks; + struct pool *pool = tc->pool; + + if (WARN_ON(get_pool_mode(pool) != PM_WRITE)) + return -EINVAL; + + r = dm_pool_get_free_block_count(pool->pmd, &free_blocks); + if (r) { + metadata_operation_failed(pool, "dm_pool_get_free_block_count", r); + return r; + } + + check_low_water_mark(pool, free_blocks); if (!free_blocks) { - if (pool->no_free_space) - return -ENOSPC; - else { - /* - * Try to commit to see if that will free up some - * more space. - */ - (void) commit_or_fallback(pool); + /* + * Try to commit to see if that will free up some + * more space. + */ + r = commit(pool); + if (r) + return r; - r = dm_pool_get_free_block_count(pool->pmd, &free_blocks); - if (r) - return r; + r = dm_pool_get_free_block_count(pool->pmd, &free_blocks); + if (r) { + metadata_operation_failed(pool, "dm_pool_get_free_block_count", r); + return r; + } - /* - * If we still have no space we set a flag to avoid - * doing all this checking and return -ENOSPC. - */ - if (!free_blocks) { - DMWARN("%s: no free space available.", - dm_device_name(pool->pool_md)); - spin_lock_irqsave(&pool->lock, flags); - pool->no_free_space = 1; - spin_unlock_irqrestore(&pool->lock, flags); - return -ENOSPC; - } + if (!free_blocks) { + set_pool_mode(pool, PM_OUT_OF_DATA_SPACE); + return -ENOSPC; } } r = dm_pool_alloc_data_block(pool->pmd, result); - if (r) + if (r) { + metadata_operation_failed(pool, "dm_pool_alloc_data_block", r); return r; + } return 0; } @@ -974,24 +1025,68 @@ static void retry_on_resume(struct bio *bio) { struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook)); struct thin_c *tc = h->tc; - struct pool *pool = tc->pool; unsigned long flags; - spin_lock_irqsave(&pool->lock, flags); - bio_list_add(&pool->retry_on_resume_list, bio); - spin_unlock_irqrestore(&pool->lock, flags); + spin_lock_irqsave(&tc->lock, flags); + bio_list_add(&tc->retry_on_resume_list, bio); + spin_unlock_irqrestore(&tc->lock, flags); } -static void no_space(struct pool *pool, struct dm_bio_prison_cell *cell) +static int should_error_unserviceable_bio(struct pool *pool) +{ + enum pool_mode m = get_pool_mode(pool); + + switch (m) { + case PM_WRITE: + /* Shouldn't get here */ + DMERR_LIMIT("bio unserviceable, yet pool is in PM_WRITE mode"); + return -EIO; + + case PM_OUT_OF_DATA_SPACE: + return pool->pf.error_if_no_space ? -ENOSPC : 0; + + case PM_READ_ONLY: + case PM_FAIL: + return -EIO; + default: + /* Shouldn't get here */ + DMERR_LIMIT("bio unserviceable, yet pool has an unknown mode"); + return -EIO; + } +} + +static void handle_unserviceable_bio(struct pool *pool, struct bio *bio) +{ + int error = should_error_unserviceable_bio(pool); + + if (error) + bio_endio(bio, error); + else + retry_on_resume(bio); +} + +static void retry_bios_on_resume(struct pool *pool, struct dm_bio_prison_cell *cell) { struct bio *bio; struct bio_list bios; + int error; + + error = should_error_unserviceable_bio(pool); + if (error) { + cell_error_with_code(pool, cell, error); + return; + } bio_list_init(&bios); cell_release(pool, cell, &bios); - while ((bio = bio_list_pop(&bios))) - retry_on_resume(bio); + error = should_error_unserviceable_bio(pool); + if (error) + while ((bio = bio_list_pop(&bios))) + bio_endio(bio, error); + else + while ((bio = bio_list_pop(&bios))) + retry_on_resume(bio); } static void process_discard(struct thin_c *tc, struct bio *bio) @@ -1030,17 +1125,17 @@ static void process_discard(struct thin_c *tc, struct bio *bio) */ m = get_next_mapping(pool); m->tc = tc; - m->pass_discard = (!lookup_result.shared) && pool->pf.discard_passdown; + m->pass_discard = pool->pf.discard_passdown; + m->definitely_not_shared = !lookup_result.shared; m->virt_block = block; m->data_block = lookup_result.block; m->cell = cell; m->cell2 = cell2; - m->err = 0; m->bio = bio; if (!dm_deferred_set_add_work(pool->all_io_ds, &m->list)) { spin_lock_irqsave(&pool->lock, flags); - list_add(&m->list, &pool->prepared_discards); + list_add_tail(&m->list, &pool->prepared_discards); spin_unlock_irqrestore(&pool->lock, flags); wake_worker(pool); } @@ -1085,6 +1180,7 @@ static void break_sharing(struct thin_c *tc, struct bio *bio, dm_block_t block, { int r; dm_block_t data_block; + struct pool *pool = tc->pool; r = alloc_data_block(tc, &data_block); switch (r) { @@ -1094,13 +1190,13 @@ static void break_sharing(struct thin_c *tc, struct bio *bio, dm_block_t block, break; case -ENOSPC: - no_space(tc->pool, cell); + retry_bios_on_resume(pool, cell); break; default: DMERR_LIMIT("%s: alloc_data_block() failed: error = %d", __func__, r); - cell_error(tc->pool, cell); + cell_error(pool, cell); break; } } @@ -1121,7 +1217,7 @@ static void process_shared_bio(struct thin_c *tc, struct bio *bio, if (bio_detain(pool, &key, bio, &cell)) return; - if (bio_data_dir(bio) == WRITE && bio->bi_size) + if (bio_data_dir(bio) == WRITE && bio->bi_iter.bi_size) break_sharing(tc, bio, block, &key, lookup_result, cell); else { struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook)); @@ -1144,7 +1240,7 @@ static void provision_block(struct thin_c *tc, struct bio *bio, dm_block_t block /* * Remap empty bios (flushes) immediately, without provisioning. */ - if (!bio->bi_size) { + if (!bio->bi_iter.bi_size) { inc_all_io_entry(pool, bio); cell_defer_no_holder(tc, cell); @@ -1172,13 +1268,12 @@ static void provision_block(struct thin_c *tc, struct bio *bio, dm_block_t block break; case -ENOSPC: - no_space(pool, cell); + retry_bios_on_resume(pool, cell); break; default: DMERR_LIMIT("%s: alloc_data_block() failed: error = %d", __func__, r); - set_pool_mode(pool, PM_READ_ONLY); cell_error(pool, cell); break; } @@ -1244,8 +1339,8 @@ static void process_bio_read_only(struct thin_c *tc, struct bio *bio) r = dm_thin_find_block(tc->td, block, 1, &lookup_result); switch (r) { case 0: - if (lookup_result.shared && (rw == WRITE) && bio->bi_size) - bio_io_error(bio); + if (lookup_result.shared && (rw == WRITE) && bio->bi_iter.bi_size) + handle_unserviceable_bio(tc->pool, bio); else { inc_all_io_entry(tc->pool, bio); remap_and_issue(tc, bio, lookup_result.block); @@ -1254,7 +1349,7 @@ static void process_bio_read_only(struct thin_c *tc, struct bio *bio) case -ENODATA: if (rw != READ) { - bio_io_error(bio); + handle_unserviceable_bio(tc->pool, bio); break; } @@ -1276,6 +1371,11 @@ static void process_bio_read_only(struct thin_c *tc, struct bio *bio) } } +static void process_bio_success(struct thin_c *tc, struct bio *bio) +{ + bio_endio(bio, 0); +} + static void process_bio_fail(struct thin_c *tc, struct bio *bio) { bio_io_error(bio); @@ -1291,33 +1391,111 @@ static int need_commit_due_to_time(struct pool *pool) jiffies > pool->last_commit_jiffies + COMMIT_PERIOD; } -static void process_deferred_bios(struct pool *pool) +#define thin_pbd(node) rb_entry((node), struct dm_thin_endio_hook, rb_node) +#define thin_bio(pbd) dm_bio_from_per_bio_data((pbd), sizeof(struct dm_thin_endio_hook)) + +static void __thin_bio_rb_add(struct thin_c *tc, struct bio *bio) +{ + struct rb_node **rbp, *parent; + struct dm_thin_endio_hook *pbd; + sector_t bi_sector = bio->bi_iter.bi_sector; + + rbp = &tc->sort_bio_list.rb_node; + parent = NULL; + while (*rbp) { + parent = *rbp; + pbd = thin_pbd(parent); + + if (bi_sector < thin_bio(pbd)->bi_iter.bi_sector) + rbp = &(*rbp)->rb_left; + else + rbp = &(*rbp)->rb_right; + } + + pbd = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook)); + rb_link_node(&pbd->rb_node, parent, rbp); + rb_insert_color(&pbd->rb_node, &tc->sort_bio_list); +} + +static void __extract_sorted_bios(struct thin_c *tc) { + struct rb_node *node; + struct dm_thin_endio_hook *pbd; + struct bio *bio; + + for (node = rb_first(&tc->sort_bio_list); node; node = rb_next(node)) { + pbd = thin_pbd(node); + bio = thin_bio(pbd); + + bio_list_add(&tc->deferred_bio_list, bio); + rb_erase(&pbd->rb_node, &tc->sort_bio_list); + } + + WARN_ON(!RB_EMPTY_ROOT(&tc->sort_bio_list)); +} + +static void __sort_thin_deferred_bios(struct thin_c *tc) +{ + struct bio *bio; + struct bio_list bios; + + bio_list_init(&bios); + bio_list_merge(&bios, &tc->deferred_bio_list); + bio_list_init(&tc->deferred_bio_list); + + /* Sort deferred_bio_list using rb-tree */ + while ((bio = bio_list_pop(&bios))) + __thin_bio_rb_add(tc, bio); + + /* + * Transfer the sorted bios in sort_bio_list back to + * deferred_bio_list to allow lockless submission of + * all bios. + */ + __extract_sorted_bios(tc); +} + +static void process_thin_deferred_bios(struct thin_c *tc) +{ + struct pool *pool = tc->pool; unsigned long flags; struct bio *bio; struct bio_list bios; + struct blk_plug plug; + + if (tc->requeue_mode) { + requeue_bio_list(tc, &tc->deferred_bio_list); + return; + } bio_list_init(&bios); - spin_lock_irqsave(&pool->lock, flags); - bio_list_merge(&bios, &pool->deferred_bios); - bio_list_init(&pool->deferred_bios); - spin_unlock_irqrestore(&pool->lock, flags); + spin_lock_irqsave(&tc->lock, flags); - while ((bio = bio_list_pop(&bios))) { - struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook)); - struct thin_c *tc = h->tc; + if (bio_list_empty(&tc->deferred_bio_list)) { + spin_unlock_irqrestore(&tc->lock, flags); + return; + } + + __sort_thin_deferred_bios(tc); + bio_list_merge(&bios, &tc->deferred_bio_list); + bio_list_init(&tc->deferred_bio_list); + + spin_unlock_irqrestore(&tc->lock, flags); + + blk_start_plug(&plug); + while ((bio = bio_list_pop(&bios))) { /* * If we've got no free new_mapping structs, and processing * this bio might require one, we pause until there are some * prepared mappings to process. */ if (ensure_next_mapping(pool)) { - spin_lock_irqsave(&pool->lock, flags); - bio_list_merge(&pool->deferred_bios, &bios); - spin_unlock_irqrestore(&pool->lock, flags); - + spin_lock_irqsave(&tc->lock, flags); + bio_list_add(&tc->deferred_bio_list, bio); + bio_list_merge(&tc->deferred_bio_list, &bios); + spin_unlock_irqrestore(&tc->lock, flags); break; } @@ -1326,6 +1504,60 @@ static void process_deferred_bios(struct pool *pool) else pool->process_bio(tc, bio); } + blk_finish_plug(&plug); +} + +static void thin_get(struct thin_c *tc); +static void thin_put(struct thin_c *tc); + +/* + * We can't hold rcu_read_lock() around code that can block. So we + * find a thin with the rcu lock held; bump a refcount; then drop + * the lock. + */ +static struct thin_c *get_first_thin(struct pool *pool) +{ + struct thin_c *tc = NULL; + + rcu_read_lock(); + if (!list_empty(&pool->active_thins)) { + tc = list_entry_rcu(pool->active_thins.next, struct thin_c, list); + thin_get(tc); + } + rcu_read_unlock(); + + return tc; +} + +static struct thin_c *get_next_thin(struct pool *pool, struct thin_c *tc) +{ + struct thin_c *old_tc = tc; + + rcu_read_lock(); + list_for_each_entry_continue_rcu(tc, &pool->active_thins, list) { + thin_get(tc); + thin_put(old_tc); + rcu_read_unlock(); + return tc; + } + thin_put(old_tc); + rcu_read_unlock(); + + return NULL; +} + +static void process_deferred_bios(struct pool *pool) +{ + unsigned long flags; + struct bio *bio; + struct bio_list bios; + struct thin_c *tc; + + tc = get_first_thin(pool); + while (tc) { + process_thin_deferred_bios(tc); + tc = get_next_thin(pool, tc); + } /* * If there are any deferred flush bios, we must commit @@ -1337,10 +1569,11 @@ static void process_deferred_bios(struct pool *pool) bio_list_init(&pool->deferred_flush_bios); spin_unlock_irqrestore(&pool->lock, flags); - if (bio_list_empty(&bios) && !need_commit_due_to_time(pool)) + if (bio_list_empty(&bios) && + !(dm_pool_changed_this_transaction(pool->pmd) && need_commit_due_to_time(pool))) return; - if (commit_or_fallback(pool)) { + if (commit(pool)) { while ((bio = bio_list_pop(&bios))) bio_io_error(bio); return; @@ -1371,6 +1604,81 @@ static void do_waker(struct work_struct *ws) queue_delayed_work(pool->wq, &pool->waker, COMMIT_PERIOD); } +/* + * We're holding onto IO to allow userland time to react. After the + * timeout either the pool will have been resized (and thus back in + * PM_WRITE mode), or we degrade to PM_READ_ONLY and start erroring IO. + */ +static void do_no_space_timeout(struct work_struct *ws) +{ + struct pool *pool = container_of(to_delayed_work(ws), struct pool, + no_space_timeout); + + if (get_pool_mode(pool) == PM_OUT_OF_DATA_SPACE && !pool->pf.error_if_no_space) + set_pool_mode(pool, PM_READ_ONLY); +} + +/*----------------------------------------------------------------*/ + +struct pool_work { + struct work_struct worker; + struct completion complete; +}; + +static struct pool_work *to_pool_work(struct work_struct *ws) +{ + return container_of(ws, struct pool_work, worker); +} + +static void pool_work_complete(struct pool_work *pw) +{ + complete(&pw->complete); +} + +static void pool_work_wait(struct pool_work *pw, struct pool *pool, + void (*fn)(struct work_struct *)) +{ + INIT_WORK_ONSTACK(&pw->worker, fn); + init_completion(&pw->complete); + queue_work(pool->wq, &pw->worker); + wait_for_completion(&pw->complete); +} + +/*----------------------------------------------------------------*/ + +struct noflush_work { + struct pool_work pw; + struct thin_c *tc; +}; + +static struct noflush_work *to_noflush(struct work_struct *ws) +{ + return container_of(to_pool_work(ws), struct noflush_work, pw); +} + +static void do_noflush_start(struct work_struct *ws) +{ + struct noflush_work *w = to_noflush(ws); + w->tc->requeue_mode = true; + requeue_io(w->tc); + pool_work_complete(&w->pw); +} + +static void do_noflush_stop(struct work_struct *ws) +{ + struct noflush_work *w = to_noflush(ws); + w->tc->requeue_mode = false; + pool_work_complete(&w->pw); +} + +static void noflush_work(struct thin_c *tc, void (*fn)(struct work_struct *)) +{ + struct noflush_work w; + + w.tc = tc; + pool_work_wait(&w.pw, tc->pool, fn); +} + /*----------------------------------------------------------------*/ static enum pool_mode get_pool_mode(struct pool *pool) @@ -1378,43 +1686,127 @@ static enum pool_mode get_pool_mode(struct pool *pool) return pool->pf.mode; } -static void set_pool_mode(struct pool *pool, enum pool_mode mode) +static void notify_of_pool_mode_change(struct pool *pool, const char *new_mode) { - int r; + dm_table_event(pool->ti->table); + DMINFO("%s: switching pool to %s mode", + dm_device_name(pool->pool_md), new_mode); +} - pool->pf.mode = mode; +static void set_pool_mode(struct pool *pool, enum pool_mode new_mode) +{ + struct pool_c *pt = pool->ti->private; + bool needs_check = dm_pool_metadata_needs_check(pool->pmd); + enum pool_mode old_mode = get_pool_mode(pool); + unsigned long no_space_timeout = ACCESS_ONCE(no_space_timeout_secs) * HZ; - switch (mode) { + /* + * Never allow the pool to transition to PM_WRITE mode if user + * intervention is required to verify metadata and data consistency. + */ + if (new_mode == PM_WRITE && needs_check) { + DMERR("%s: unable to switch pool to write mode until repaired.", + dm_device_name(pool->pool_md)); + if (old_mode != new_mode) + new_mode = old_mode; + else + new_mode = PM_READ_ONLY; + } + /* + * If we were in PM_FAIL mode, rollback of metadata failed. We're + * not going to recover without a thin_repair. So we never let the + * pool move out of the old mode. + */ + if (old_mode == PM_FAIL) + new_mode = old_mode; + + switch (new_mode) { case PM_FAIL: - DMERR("switching pool to failure mode"); + if (old_mode != new_mode) + notify_of_pool_mode_change(pool, "failure"); + dm_pool_metadata_read_only(pool->pmd); pool->process_bio = process_bio_fail; pool->process_discard = process_bio_fail; pool->process_prepared_mapping = process_prepared_mapping_fail; pool->process_prepared_discard = process_prepared_discard_fail; + + error_retry_list(pool); break; case PM_READ_ONLY: - DMERR("switching pool to read-only mode"); - r = dm_pool_abort_metadata(pool->pmd); - if (r) { - DMERR("aborting transaction failed"); - set_pool_mode(pool, PM_FAIL); - } else { - dm_pool_metadata_read_only(pool->pmd); - pool->process_bio = process_bio_read_only; - pool->process_discard = process_discard; - pool->process_prepared_mapping = process_prepared_mapping_fail; - pool->process_prepared_discard = process_prepared_discard_passdown; - } + if (old_mode != new_mode) + notify_of_pool_mode_change(pool, "read-only"); + dm_pool_metadata_read_only(pool->pmd); + pool->process_bio = process_bio_read_only; + pool->process_discard = process_bio_success; + pool->process_prepared_mapping = process_prepared_mapping_fail; + pool->process_prepared_discard = process_prepared_discard_passdown; + + error_retry_list(pool); + break; + + case PM_OUT_OF_DATA_SPACE: + /* + * Ideally we'd never hit this state; the low water mark + * would trigger userland to extend the pool before we + * completely run out of data space. However, many small + * IOs to unprovisioned space can consume data space at an + * alarming rate. Adjust your low water mark if you're + * frequently seeing this mode. + */ + if (old_mode != new_mode) + notify_of_pool_mode_change(pool, "out-of-data-space"); + pool->process_bio = process_bio_read_only; + pool->process_discard = process_discard; + pool->process_prepared_mapping = process_prepared_mapping; + pool->process_prepared_discard = process_prepared_discard_passdown; + + if (!pool->pf.error_if_no_space && no_space_timeout) + queue_delayed_work(pool->wq, &pool->no_space_timeout, no_space_timeout); break; case PM_WRITE: + if (old_mode != new_mode) + notify_of_pool_mode_change(pool, "write"); + dm_pool_metadata_read_write(pool->pmd); pool->process_bio = process_bio; pool->process_discard = process_discard; pool->process_prepared_mapping = process_prepared_mapping; pool->process_prepared_discard = process_prepared_discard; break; } + + pool->pf.mode = new_mode; + /* + * The pool mode may have changed, sync it so bind_control_target() + * doesn't cause an unexpected mode transition on resume. + */ + pt->adjusted_pf.mode = new_mode; +} + +static void abort_transaction(struct pool *pool) +{ + const char *dev_name = dm_device_name(pool->pool_md); + + DMERR_LIMIT("%s: aborting current metadata transaction", dev_name); + if (dm_pool_abort_metadata(pool->pmd)) { + DMERR("%s: failed to abort metadata transaction", dev_name); + set_pool_mode(pool, PM_FAIL); + } + + if (dm_pool_metadata_set_needs_check(pool->pmd)) { + DMERR("%s: failed to set 'needs_check' flag in metadata", dev_name); + set_pool_mode(pool, PM_FAIL); + } +} + +static void metadata_operation_failed(struct pool *pool, const char *op, int r) +{ + DMERR_LIMIT("%s: metadata operation '%s' failed: error = %d", + dm_device_name(pool->pool_md), op, r); + + abort_transaction(pool); + set_pool_mode(pool, PM_READ_ONLY); } /*----------------------------------------------------------------*/ @@ -1431,9 +1823,9 @@ static void thin_defer_bio(struct thin_c *tc, struct bio *bio) unsigned long flags; struct pool *pool = tc->pool; - spin_lock_irqsave(&pool->lock, flags); - bio_list_add(&pool->deferred_bios, bio); - spin_unlock_irqrestore(&pool->lock, flags); + spin_lock_irqsave(&tc->lock, flags); + bio_list_add(&tc->deferred_bio_list, bio); + spin_unlock_irqrestore(&tc->lock, flags); wake_worker(pool); } @@ -1464,6 +1856,11 @@ static int thin_bio_map(struct dm_target *ti, struct bio *bio) thin_hook_bio(tc, bio); + if (tc->requeue_mode) { + bio_endio(bio, DM_ENDIO_REQUEUE); + return DM_MAPIO_SUBMITTED; + } + if (get_pool_mode(tc->pool) == PM_FAIL) { bio_io_error(bio); return DM_MAPIO_SUBMITTED; @@ -1521,9 +1918,9 @@ static int thin_bio_map(struct dm_target *ti, struct bio *bio) if (get_pool_mode(tc->pool) == PM_READ_ONLY) { /* * This block isn't provisioned, and we have no way - * of doing so. Just error it. + * of doing so. */ - bio_io_error(bio); + handle_unserviceable_bio(tc->pool, bio); return DM_MAPIO_SUBMITTED; } /* fall through */ @@ -1549,26 +1946,29 @@ static int thin_bio_map(struct dm_target *ti, struct bio *bio) static int pool_is_congested(struct dm_target_callbacks *cb, int bdi_bits) { - int r; - unsigned long flags; struct pool_c *pt = container_of(cb, struct pool_c, callbacks); + struct request_queue *q; - spin_lock_irqsave(&pt->pool->lock, flags); - r = !bio_list_empty(&pt->pool->retry_on_resume_list); - spin_unlock_irqrestore(&pt->pool->lock, flags); - - if (!r) { - struct request_queue *q = bdev_get_queue(pt->data_dev->bdev); - r = bdi_congested(&q->backing_dev_info, bdi_bits); - } + if (get_pool_mode(pt->pool) == PM_OUT_OF_DATA_SPACE) + return 1; - return r; + q = bdev_get_queue(pt->data_dev->bdev); + return bdi_congested(&q->backing_dev_info, bdi_bits); } -static void __requeue_bios(struct pool *pool) +static void requeue_bios(struct pool *pool) { - bio_list_merge(&pool->deferred_bios, &pool->retry_on_resume_list); - bio_list_init(&pool->retry_on_resume_list); + unsigned long flags; + struct thin_c *tc; + + rcu_read_lock(); + list_for_each_entry_rcu(tc, &pool->active_thins, list) { + spin_lock_irqsave(&tc->lock, flags); + bio_list_merge(&tc->deferred_bio_list, &tc->retry_on_resume_list); + bio_list_init(&tc->retry_on_resume_list); + spin_unlock_irqrestore(&tc->lock, flags); + } + rcu_read_unlock(); } /*---------------------------------------------------------------- @@ -1625,17 +2025,21 @@ static int bind_control_target(struct pool *pool, struct dm_target *ti) struct pool_c *pt = ti->private; /* - * We want to make sure that degraded pools are never upgraded. + * We want to make sure that a pool in PM_FAIL mode is never upgraded. */ - enum pool_mode old_mode = pool->pf.mode; + enum pool_mode old_mode = get_pool_mode(pool); enum pool_mode new_mode = pt->adjusted_pf.mode; - if (old_mode > new_mode) - new_mode = old_mode; + /* + * Don't change the pool's mode until set_pool_mode() below. + * Otherwise the pool's process_* function pointers may + * not match the desired pool mode. + */ + pt->adjusted_pf.mode = old_mode; pool->ti = ti; - pool->low_water_blocks = pt->low_water_blocks; pool->pf = pt->adjusted_pf; + pool->low_water_blocks = pt->low_water_blocks; set_pool_mode(pool, new_mode); @@ -1658,6 +2062,7 @@ static void pool_features_init(struct pool_features *pf) pf->zero_new_blocks = true; pf->discard_enabled = true; pf->discard_passdown = true; + pf->error_if_no_space = false; } static void __pool_destroy(struct pool *pool) @@ -1743,14 +2148,13 @@ static struct pool *pool_create(struct mapped_device *pool_md, INIT_WORK(&pool->worker, do_worker); INIT_DELAYED_WORK(&pool->waker, do_waker); + INIT_DELAYED_WORK(&pool->no_space_timeout, do_no_space_timeout); spin_lock_init(&pool->lock); - bio_list_init(&pool->deferred_bios); bio_list_init(&pool->deferred_flush_bios); INIT_LIST_HEAD(&pool->prepared_mappings); INIT_LIST_HEAD(&pool->prepared_discards); - pool->low_water_triggered = 0; - pool->no_free_space = 0; - bio_list_init(&pool->retry_on_resume_list); + INIT_LIST_HEAD(&pool->active_thins); + pool->low_water_triggered = false; pool->shared_read_ds = dm_deferred_set_create(); if (!pool->shared_read_ds) { @@ -1874,7 +2278,7 @@ static int parse_pool_features(struct dm_arg_set *as, struct pool_features *pf, const char *arg_name; static struct dm_arg _args[] = { - {0, 3, "Invalid number of pool feature arguments"}, + {0, 4, "Invalid number of pool feature arguments"}, }; /* @@ -1903,6 +2307,9 @@ static int parse_pool_features(struct dm_arg_set *as, struct pool_features *pf, else if (!strcasecmp(arg_name, "read_only")) pf->mode = PM_READ_ONLY; + else if (!strcasecmp(arg_name, "error_if_no_space")) + pf->error_if_no_space = true; + else { ti->error = "Unrecognised pool feature requested"; r = -EINVAL; @@ -1923,16 +2330,27 @@ static void metadata_low_callback(void *context) dm_table_event(pool->ti->table); } -static sector_t get_metadata_dev_size(struct block_device *bdev) +static sector_t get_dev_size(struct block_device *bdev) +{ + return i_size_read(bdev->bd_inode) >> SECTOR_SHIFT; +} + +static void warn_if_metadata_device_too_big(struct block_device *bdev) { - sector_t metadata_dev_size = i_size_read(bdev->bd_inode) >> SECTOR_SHIFT; + sector_t metadata_dev_size = get_dev_size(bdev); char buffer[BDEVNAME_SIZE]; - if (metadata_dev_size > THIN_METADATA_MAX_SECTORS_WARNING) { + if (metadata_dev_size > THIN_METADATA_MAX_SECTORS_WARNING) DMWARN("Metadata device %s is larger than %u sectors: excess space will not be used.", bdevname(bdev, buffer), THIN_METADATA_MAX_SECTORS); - metadata_dev_size = THIN_METADATA_MAX_SECTORS_WARNING; - } +} + +static sector_t get_metadata_dev_size(struct block_device *bdev) +{ + sector_t metadata_dev_size = get_dev_size(bdev); + + if (metadata_dev_size > THIN_METADATA_MAX_SECTORS) + metadata_dev_size = THIN_METADATA_MAX_SECTORS; return metadata_dev_size; } @@ -1941,7 +2359,7 @@ static dm_block_t get_metadata_dev_size_in_blocks(struct block_device *bdev) { sector_t metadata_dev_size = get_metadata_dev_size(bdev); - sector_div(metadata_dev_size, THIN_METADATA_BLOCK_SIZE >> SECTOR_SHIFT); + sector_div(metadata_dev_size, THIN_METADATA_BLOCK_SIZE); return metadata_dev_size; } @@ -1973,6 +2391,8 @@ static dm_block_t calc_metadata_threshold(struct pool_c *pt) * skip_block_zeroing: skips the zeroing of newly-provisioned blocks. * ignore_discard: disable discard * no_discard_passdown: don't pass discards down to the data device + * read_only: Don't allow any changes to be made to the pool metadata. + * error_if_no_space: error IOs, instead of queueing, if no space. */ static int pool_ctr(struct dm_target *ti, unsigned argc, char **argv) { @@ -2017,12 +2437,7 @@ static int pool_ctr(struct dm_target *ti, unsigned argc, char **argv) ti->error = "Error opening metadata block device"; goto out_unlock; } - - /* - * Run for the side-effect of possibly issuing a warning if the - * device is too big. - */ - (void) get_metadata_dev_size(metadata_dev->bdev); + warn_if_metadata_device_too_big(metadata_dev->bdev); r = dm_get_device(ti, argv[1], FMODE_READ | FMODE_WRITE, &data_dev); if (r) { @@ -2083,6 +2498,7 @@ static int pool_ctr(struct dm_target *ti, unsigned argc, char **argv) * them down to the data device. The thin device's discard * processing will cause mappings to be removed from the btree. */ + ti->discard_zeroes_data_unsupported = true; if (pf.discard_enabled && pf.discard_passdown) { ti->num_discard_bios = 1; @@ -2092,7 +2508,6 @@ static int pool_ctr(struct dm_target *ti, unsigned argc, char **argv) * thin devices' discard limits consistent). */ ti->discards_supported = true; - ti->discard_zeroes_data_unsupported = true; } ti->private = pt; @@ -2156,20 +2571,31 @@ static int maybe_resize_data_dev(struct dm_target *ti, bool *need_commit) r = dm_pool_get_data_dev_size(pool->pmd, &sb_data_size); if (r) { - DMERR("failed to retrieve data device size"); + DMERR("%s: failed to retrieve data device size", + dm_device_name(pool->pool_md)); return r; } if (data_size < sb_data_size) { - DMERR("pool target (%llu blocks) too small: expected %llu", + DMERR("%s: pool target (%llu blocks) too small: expected %llu", + dm_device_name(pool->pool_md), (unsigned long long)data_size, sb_data_size); return -EINVAL; } else if (data_size > sb_data_size) { + if (dm_pool_metadata_needs_check(pool->pmd)) { + DMERR("%s: unable to grow the data device until repaired.", + dm_device_name(pool->pool_md)); + return 0; + } + + if (sb_data_size) + DMINFO("%s: growing the data device from %llu to %llu blocks", + dm_device_name(pool->pool_md), + sb_data_size, (unsigned long long)data_size); r = dm_pool_resize_data_dev(pool->pmd, data_size); if (r) { - DMERR("failed to resize data device"); - set_pool_mode(pool, PM_READ_ONLY); + metadata_operation_failed(pool, "dm_pool_resize_data_dev", r); return r; } @@ -2192,19 +2618,31 @@ static int maybe_resize_metadata_dev(struct dm_target *ti, bool *need_commit) r = dm_pool_get_metadata_dev_size(pool->pmd, &sb_metadata_dev_size); if (r) { - DMERR("failed to retrieve data device size"); + DMERR("%s: failed to retrieve metadata device size", + dm_device_name(pool->pool_md)); return r; } if (metadata_dev_size < sb_metadata_dev_size) { - DMERR("metadata device (%llu blocks) too small: expected %llu", + DMERR("%s: metadata device (%llu blocks) too small: expected %llu", + dm_device_name(pool->pool_md), metadata_dev_size, sb_metadata_dev_size); return -EINVAL; } else if (metadata_dev_size > sb_metadata_dev_size) { + if (dm_pool_metadata_needs_check(pool->pmd)) { + DMERR("%s: unable to grow the metadata device until repaired.", + dm_device_name(pool->pool_md)); + return 0; + } + + warn_if_metadata_device_too_big(pool->md_dev); + DMINFO("%s: growing the metadata device from %llu to %llu blocks", + dm_device_name(pool->pool_md), + sb_metadata_dev_size, metadata_dev_size); r = dm_pool_resize_metadata_dev(pool->pmd, metadata_dev_size); if (r) { - DMERR("failed to resize metadata device"); + metadata_operation_failed(pool, "dm_pool_resize_metadata_dev", r); return r; } @@ -2248,7 +2686,7 @@ static int pool_preresume(struct dm_target *ti) return r; if (need_commit1 || need_commit2) - (void) commit_or_fallback(pool); + (void) commit(pool); return 0; } @@ -2260,10 +2698,9 @@ static void pool_resume(struct dm_target *ti) unsigned long flags; spin_lock_irqsave(&pool->lock, flags); - pool->low_water_triggered = 0; - pool->no_free_space = 0; - __requeue_bios(pool); + pool->low_water_triggered = false; spin_unlock_irqrestore(&pool->lock, flags); + requeue_bios(pool); do_waker(&pool->waker.work); } @@ -2274,8 +2711,9 @@ static void pool_postsuspend(struct dm_target *ti) struct pool *pool = pt->pool; cancel_delayed_work(&pool->waker); + cancel_delayed_work(&pool->no_space_timeout); flush_workqueue(pool->wq); - (void) commit_or_fallback(pool); + (void) commit(pool); } static int check_arg_count(unsigned argc, unsigned args_required) @@ -2409,7 +2847,7 @@ static int process_reserve_metadata_snap_mesg(unsigned argc, char **argv, struct if (r) return r; - (void) commit_or_fallback(pool); + (void) commit(pool); r = dm_pool_reserve_metadata_snap(pool->pmd); if (r) @@ -2471,7 +2909,7 @@ static int pool_message(struct dm_target *ti, unsigned argc, char **argv) DMWARN("Unrecognised thin pool target message received: %s", argv[0]); if (!r) - (void) commit_or_fallback(pool); + (void) commit(pool); return r; } @@ -2480,7 +2918,8 @@ static void emit_flags(struct pool_features *pf, char *result, unsigned sz, unsigned maxlen) { unsigned count = !pf->zero_new_blocks + !pf->discard_enabled + - !pf->discard_passdown + (pf->mode == PM_READ_ONLY); + !pf->discard_passdown + (pf->mode == PM_READ_ONLY) + + pf->error_if_no_space; DMEMIT("%u ", count); if (!pf->zero_new_blocks) @@ -2494,6 +2933,9 @@ static void emit_flags(struct pool_features *pf, char *result, if (pf->mode == PM_READ_ONLY) DMEMIT("read_only "); + + if (pf->error_if_no_space) + DMEMIT("error_if_no_space "); } /* @@ -2526,41 +2968,47 @@ static void pool_status(struct dm_target *ti, status_type_t type, /* Commit to ensure statistics aren't out-of-date */ if (!(status_flags & DM_STATUS_NOFLUSH_FLAG) && !dm_suspended(ti)) - (void) commit_or_fallback(pool); + (void) commit(pool); r = dm_pool_get_metadata_transaction_id(pool->pmd, &transaction_id); if (r) { - DMERR("dm_pool_get_metadata_transaction_id returned %d", r); + DMERR("%s: dm_pool_get_metadata_transaction_id returned %d", + dm_device_name(pool->pool_md), r); goto err; } r = dm_pool_get_free_metadata_block_count(pool->pmd, &nr_free_blocks_metadata); if (r) { - DMERR("dm_pool_get_free_metadata_block_count returned %d", r); + DMERR("%s: dm_pool_get_free_metadata_block_count returned %d", + dm_device_name(pool->pool_md), r); goto err; } r = dm_pool_get_metadata_dev_size(pool->pmd, &nr_blocks_metadata); if (r) { - DMERR("dm_pool_get_metadata_dev_size returned %d", r); + DMERR("%s: dm_pool_get_metadata_dev_size returned %d", + dm_device_name(pool->pool_md), r); goto err; } r = dm_pool_get_free_block_count(pool->pmd, &nr_free_blocks_data); if (r) { - DMERR("dm_pool_get_free_block_count returned %d", r); + DMERR("%s: dm_pool_get_free_block_count returned %d", + dm_device_name(pool->pool_md), r); goto err; } r = dm_pool_get_data_dev_size(pool->pmd, &nr_blocks_data); if (r) { - DMERR("dm_pool_get_data_dev_size returned %d", r); + DMERR("%s: dm_pool_get_data_dev_size returned %d", + dm_device_name(pool->pool_md), r); goto err; } r = dm_pool_get_metadata_snap(pool->pmd, &held_root); if (r) { - DMERR("dm_pool_get_metadata_snap returned %d", r); + DMERR("%s: dm_pool_get_metadata_snap returned %d", + dm_device_name(pool->pool_md), r); goto err; } @@ -2576,17 +3024,24 @@ static void pool_status(struct dm_target *ti, status_type_t type, else DMEMIT("- "); - if (pool->pf.mode == PM_READ_ONLY) + if (pool->pf.mode == PM_OUT_OF_DATA_SPACE) + DMEMIT("out_of_data_space "); + else if (pool->pf.mode == PM_READ_ONLY) DMEMIT("ro "); else DMEMIT("rw "); if (!pool->pf.discard_enabled) - DMEMIT("ignore_discard"); + DMEMIT("ignore_discard "); else if (pool->pf.discard_passdown) - DMEMIT("discard_passdown"); + DMEMIT("discard_passdown "); + else + DMEMIT("no_discard_passdown "); + + if (pool->pf.error_if_no_space) + DMEMIT("error_if_no_space "); else - DMEMIT("no_discard_passdown"); + DMEMIT("queue_if_no_space "); break; @@ -2639,7 +3094,8 @@ static void set_discard_limits(struct pool_c *pt, struct queue_limits *limits) */ if (pt->adjusted_pf.discard_passdown) { data_limits = &bdev_get_queue(pt->data_dev->bdev)->limits; - limits->discard_granularity = data_limits->discard_granularity; + limits->discard_granularity = max(data_limits->discard_granularity, + pool->sectors_per_block << SECTOR_SHIFT); } else limits->discard_granularity = pool->sectors_per_block << SECTOR_SHIFT; } @@ -2648,17 +3104,33 @@ static void pool_io_hints(struct dm_target *ti, struct queue_limits *limits) { struct pool_c *pt = ti->private; struct pool *pool = pt->pool; + uint64_t io_opt_sectors = limits->io_opt >> SECTOR_SHIFT; - blk_limits_io_min(limits, 0); - blk_limits_io_opt(limits, pool->sectors_per_block << SECTOR_SHIFT); + /* + * If the system-determined stacked limits are compatible with the + * pool's blocksize (io_opt is a factor) do not override them. + */ + if (io_opt_sectors < pool->sectors_per_block || + do_div(io_opt_sectors, pool->sectors_per_block)) { + blk_limits_io_min(limits, 0); + blk_limits_io_opt(limits, pool->sectors_per_block << SECTOR_SHIFT); + } /* * pt->adjusted_pf is a staging area for the actual features to use. * They get transferred to the live pool in bind_control_target() * called from pool_preresume(). */ - if (!pt->adjusted_pf.discard_enabled) + if (!pt->adjusted_pf.discard_enabled) { + /* + * Must explicitly disallow stacking discard limits otherwise the + * block layer will stack them if pool's data device has support. + * QUEUE_FLAG_DISCARD wouldn't be set but there is no way for the + * user to see that, so make sure to set all discard limits to 0. + */ + limits->discard_granularity = 0; return; + } disable_passdown_if_not_supported(pt); @@ -2669,7 +3141,7 @@ static struct target_type pool_target = { .name = "thin-pool", .features = DM_TARGET_SINGLETON | DM_TARGET_ALWAYS_WRITEABLE | DM_TARGET_IMMUTABLE, - .version = {1, 8, 0}, + .version = {1, 12, 0}, .module = THIS_MODULE, .ctr = pool_ctr, .dtr = pool_dtr, @@ -2687,9 +3159,29 @@ static struct target_type pool_target = { /*---------------------------------------------------------------- * Thin target methods *--------------------------------------------------------------*/ +static void thin_get(struct thin_c *tc) +{ + atomic_inc(&tc->refcount); +} + +static void thin_put(struct thin_c *tc) +{ + if (atomic_dec_and_test(&tc->refcount)) + complete(&tc->can_destroy); +} + static void thin_dtr(struct dm_target *ti) { struct thin_c *tc = ti->private; + unsigned long flags; + + thin_put(tc); + wait_for_completion(&tc->can_destroy); + + spin_lock_irqsave(&tc->pool->lock, flags); + list_del_rcu(&tc->list); + spin_unlock_irqrestore(&tc->pool->lock, flags); + synchronize_rcu(); mutex_lock(&dm_thin_pool_table.mutex); @@ -2721,6 +3213,7 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv) struct thin_c *tc; struct dm_dev *pool_dev, *origin_dev; struct mapped_device *pool_md; + unsigned long flags; mutex_lock(&dm_thin_pool_table.mutex); @@ -2736,6 +3229,10 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv) r = -ENOMEM; goto out_unlock; } + spin_lock_init(&tc->lock); + bio_list_init(&tc->deferred_bio_list); + bio_list_init(&tc->retry_on_resume_list); + tc->sort_bio_list = RB_ROOT; if (argc == 3) { r = dm_get_device(ti, argv[2], FMODE_READ, &origin_dev); @@ -2776,6 +3273,7 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv) if (get_pool_mode(tc->pool) == PM_FAIL) { ti->error = "Couldn't open thin device, Pool is in fail mode"; + r = -EINVAL; goto bad_thin_open; } @@ -2787,17 +3285,17 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv) r = dm_set_target_max_io_len(ti, tc->pool->sectors_per_block); if (r) - goto bad_thin_open; + goto bad_target_max_io_len; ti->num_flush_bios = 1; ti->flush_supported = true; ti->per_bio_data_size = sizeof(struct dm_thin_endio_hook); /* In case the pool supports discards, pass them on. */ + ti->discard_zeroes_data_unsupported = true; if (tc->pool->pf.discard_enabled) { ti->discards_supported = true; ti->num_discard_bios = 1; - ti->discard_zeroes_data_unsupported = true; /* Discard bios must be split on a block boundary */ ti->split_discard_bios = true; } @@ -2806,8 +3304,24 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv) mutex_unlock(&dm_thin_pool_table.mutex); + atomic_set(&tc->refcount, 1); + init_completion(&tc->can_destroy); + + spin_lock_irqsave(&tc->pool->lock, flags); + list_add_tail_rcu(&tc->list, &tc->pool->active_thins); + spin_unlock_irqrestore(&tc->pool->lock, flags); + /* + * This synchronize_rcu() call is needed here otherwise we risk a + * wake_worker() call finding no bios to process (because the newly + * added tc isn't yet visible). So this reduces latency since we + * aren't then dependent on the periodic commit to wake_worker(). + */ + synchronize_rcu(); + return 0; +bad_target_max_io_len: + dm_pool_close_thin_device(tc->td); bad_thin_open: __pool_dec(tc->pool); bad_pool_lookup: @@ -2827,7 +3341,7 @@ out_unlock: static int thin_map(struct dm_target *ti, struct bio *bio) { - bio->bi_sector = dm_target_offset(ti, bio->bi_sector); + bio->bi_iter.bi_sector = dm_target_offset(ti, bio->bi_iter.bi_sector); return thin_bio_map(ti, bio); } @@ -2847,7 +3361,7 @@ static int thin_endio(struct dm_target *ti, struct bio *bio, int err) spin_lock_irqsave(&pool->lock, flags); list_for_each_entry_safe(m, tmp, &work, list) { list_del(&m->list); - m->quiesced = 1; + m->quiesced = true; __maybe_add_mapping(m); } spin_unlock_irqrestore(&pool->lock, flags); @@ -2859,7 +3373,7 @@ static int thin_endio(struct dm_target *ti, struct bio *bio, int err) if (!list_empty(&work)) { spin_lock_irqsave(&pool->lock, flags); list_for_each_entry_safe(m, tmp, &work, list) - list_add(&m->list, &pool->prepared_discards); + list_add_tail(&m->list, &pool->prepared_discards); spin_unlock_irqrestore(&pool->lock, flags); wake_worker(pool); } @@ -2868,10 +3382,23 @@ static int thin_endio(struct dm_target *ti, struct bio *bio, int err) return 0; } -static void thin_postsuspend(struct dm_target *ti) +static void thin_presuspend(struct dm_target *ti) { + struct thin_c *tc = ti->private; + if (dm_noflush_suspending(ti)) - requeue_io((struct thin_c *)ti->private); + noflush_work(tc, do_noflush_start); +} + +static void thin_postsuspend(struct dm_target *ti) +{ + struct thin_c *tc = ti->private; + + /* + * The dm_noflush_suspending flag has been cleared by now, so + * unfortunately we must always run this. + */ + noflush_work(tc, do_noflush_stop); } /* @@ -2956,12 +3483,13 @@ static int thin_iterate_devices(struct dm_target *ti, static struct target_type thin_target = { .name = "thin", - .version = {1, 8, 0}, + .version = {1, 12, 0}, .module = THIS_MODULE, .ctr = thin_ctr, .dtr = thin_dtr, .map = thin_map, .end_io = thin_endio, + .presuspend = thin_presuspend, .postsuspend = thin_postsuspend, .status = thin_status, .iterate_devices = thin_iterate_devices, @@ -3010,6 +3538,9 @@ static void dm_thin_exit(void) module_init(dm_thin_init); module_exit(dm_thin_exit); +module_param_named(no_space_timeout, no_space_timeout_secs, uint, S_IRUGO | S_IWUSR); +MODULE_PARM_DESC(no_space_timeout, "Out of data space queue IO timeout in seconds"); + MODULE_DESCRIPTION(DM_NAME " thin provisioning target"); MODULE_AUTHOR("Joe Thornber <dm-devel@redhat.com>"); MODULE_LICENSE("GPL"); diff --git a/drivers/md/dm-verity.c b/drivers/md/dm-verity.c index b948fd864d4..7a7bab8947a 100644 --- a/drivers/md/dm-verity.c +++ b/drivers/md/dm-verity.c @@ -73,15 +73,10 @@ struct dm_verity_io { sector_t block; unsigned n_blocks; - /* saved bio vector */ - struct bio_vec *io_vec; - unsigned io_vec_size; + struct bvec_iter iter; struct work_struct work; - /* A space for short vectors; longer vectors are allocated separately. */ - struct bio_vec io_vec_inline[DM_VERITY_IO_VEC_INLINE]; - /* * Three variably-size fields follow this struct: * @@ -284,9 +279,10 @@ release_ret_r: static int verity_verify_io(struct dm_verity_io *io) { struct dm_verity *v = io->v; + struct bio *bio = dm_bio_from_per_bio_data(io, + v->ti->per_bio_data_size); unsigned b; int i; - unsigned vector = 0, offset = 0; for (b = 0; b < io->n_blocks; b++) { struct shash_desc *desc; @@ -334,31 +330,25 @@ test_block_hash: return r; } } - todo = 1 << v->data_dev_block_bits; do { - struct bio_vec *bv; u8 *page; unsigned len; + struct bio_vec bv = bio_iter_iovec(bio, io->iter); - BUG_ON(vector >= io->io_vec_size); - bv = &io->io_vec[vector]; - page = kmap_atomic(bv->bv_page); - len = bv->bv_len - offset; + page = kmap_atomic(bv.bv_page); + len = bv.bv_len; if (likely(len >= todo)) len = todo; - r = crypto_shash_update(desc, - page + bv->bv_offset + offset, len); + r = crypto_shash_update(desc, page + bv.bv_offset, len); kunmap_atomic(page); + if (r < 0) { DMERR("crypto_shash_update failed: %d", r); return r; } - offset += len; - if (likely(offset == bv->bv_len)) { - offset = 0; - vector++; - } + + bio_advance_iter(bio, &io->iter, len); todo -= len; } while (todo); @@ -383,8 +373,6 @@ test_block_hash: return -EIO; } } - BUG_ON(vector != io->io_vec_size); - BUG_ON(offset); return 0; } @@ -400,10 +388,7 @@ static void verity_finish_io(struct dm_verity_io *io, int error) bio->bi_end_io = io->orig_bi_end_io; bio->bi_private = io->orig_bi_private; - if (io->io_vec != io->io_vec_inline) - mempool_free(io->io_vec, v->vec_mempool); - - bio_endio(bio, error); + bio_endio_nodec(bio, error); } static void verity_work(struct work_struct *w) @@ -451,7 +436,7 @@ static void verity_prefetch_io(struct work_struct *work) goto no_prefetch_cluster; if (unlikely(cluster & (cluster - 1))) - cluster = 1 << (fls(cluster) - 1); + cluster = 1 << __fls(cluster); hash_block_start &= ~(sector_t)(cluster - 1); hash_block_end |= cluster - 1; @@ -493,9 +478,9 @@ static int verity_map(struct dm_target *ti, struct bio *bio) struct dm_verity_io *io; bio->bi_bdev = v->data_dev->bdev; - bio->bi_sector = verity_map_sector(v, bio->bi_sector); + bio->bi_iter.bi_sector = verity_map_sector(v, bio->bi_iter.bi_sector); - if (((unsigned)bio->bi_sector | bio_sectors(bio)) & + if (((unsigned)bio->bi_iter.bi_sector | bio_sectors(bio)) & ((1 << (v->data_dev_block_bits - SECTOR_SHIFT)) - 1)) { DMERR_LIMIT("unaligned io"); return -EIO; @@ -514,18 +499,12 @@ static int verity_map(struct dm_target *ti, struct bio *bio) io->v = v; io->orig_bi_end_io = bio->bi_end_io; io->orig_bi_private = bio->bi_private; - io->block = bio->bi_sector >> (v->data_dev_block_bits - SECTOR_SHIFT); - io->n_blocks = bio->bi_size >> v->data_dev_block_bits; + io->block = bio->bi_iter.bi_sector >> (v->data_dev_block_bits - SECTOR_SHIFT); + io->n_blocks = bio->bi_iter.bi_size >> v->data_dev_block_bits; bio->bi_end_io = verity_end_io; bio->bi_private = io; - io->io_vec_size = bio_segments(bio); - if (io->io_vec_size < DM_VERITY_IO_VEC_INLINE) - io->io_vec = io->io_vec_inline; - else - io->io_vec = mempool_alloc(v->vec_mempool, GFP_NOIO); - memcpy(io->io_vec, bio_iovec(bio), - io->io_vec_size * sizeof(struct bio_vec)); + io->iter = bio->bi_iter; verity_submit_prefetch(v, io); @@ -695,8 +674,8 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv) goto bad; } - if (sscanf(argv[0], "%d%c", &num, &dummy) != 1 || - num < 0 || num > 1) { + if (sscanf(argv[0], "%u%c", &num, &dummy) != 1 || + num > 1) { ti->error = "Invalid version"; r = -EINVAL; goto bad; @@ -723,7 +702,7 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv) r = -EINVAL; goto bad; } - v->data_dev_block_bits = ffs(num) - 1; + v->data_dev_block_bits = __ffs(num); if (sscanf(argv[4], "%u%c", &num, &dummy) != 1 || !num || (num & (num - 1)) || @@ -733,7 +712,7 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv) r = -EINVAL; goto bad; } - v->hash_dev_block_bits = ffs(num) - 1; + v->hash_dev_block_bits = __ffs(num); if (sscanf(argv[5], "%llu%c", &num_ll, &dummy) != 1 || (sector_t)(num_ll << (v->data_dev_block_bits - SECTOR_SHIFT)) @@ -812,7 +791,7 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv) } v->hash_per_block_bits = - fls((1 << v->hash_dev_block_bits) / v->digest_size) - 1; + __fls((1 << v->hash_dev_block_bits) / v->digest_size); v->levels = 0; if (v->data_blocks) @@ -831,9 +810,8 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv) for (i = v->levels - 1; i >= 0; i--) { sector_t s; v->hash_level_block[i] = hash_position; - s = verity_position_at_level(v, v->data_blocks, i); - s = (s >> v->hash_per_block_bits) + - !!(s & ((1 << v->hash_per_block_bits) - 1)); + s = (v->data_blocks + ((sector_t)1 << ((i + 1) * v->hash_per_block_bits)) - 1) + >> ((i + 1) * v->hash_per_block_bits); if (hash_position + s < hash_position) { ti->error = "Hash device offset overflow"; r = -E2BIG; diff --git a/drivers/md/dm-zero.c b/drivers/md/dm-zero.c index c99003e0d47..b9a64bbce30 100644 --- a/drivers/md/dm-zero.c +++ b/drivers/md/dm-zero.c @@ -1,5 +1,5 @@ /* - * Copyright (C) 2003 Christophe Saout <christophe@saout.de> + * Copyright (C) 2003 Jana Saout <jana@saout.de> * * This file is released under the GPL. */ @@ -79,6 +79,6 @@ static void __exit dm_zero_exit(void) module_init(dm_zero_init) module_exit(dm_zero_exit) -MODULE_AUTHOR("Christophe Saout <christophe@saout.de>"); +MODULE_AUTHOR("Jana Saout <jana@saout.de>"); MODULE_DESCRIPTION(DM_NAME " dummy target returning zeros"); MODULE_LICENSE("GPL"); diff --git a/drivers/md/dm.c b/drivers/md/dm.c index d5370a94b2c..32b958dbc49 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -49,6 +49,13 @@ static unsigned int _major = 0; static DEFINE_IDR(_minor_idr); static DEFINE_SPINLOCK(_minor_lock); + +static void do_deferred_remove(struct work_struct *w); + +static DECLARE_WORK(deferred_remove_work, do_deferred_remove); + +static struct workqueue_struct *deferred_remove_workqueue; + /* * For bio-based dm. * One of these is allocated per bio. @@ -60,6 +67,7 @@ struct dm_io { struct bio *bio; unsigned long start_time; spinlock_t endio_lock; + struct dm_stats_aux stats_aux; }; /* @@ -88,13 +96,6 @@ struct dm_rq_clone_bio_info { struct bio clone; }; -union map_info *dm_get_mapinfo(struct bio *bio) -{ - if (bio && bio->bi_private) - return &((struct dm_target_io *)bio->bi_private)->info; - return NULL; -} - union map_info *dm_get_rq_mapinfo(struct request *rq) { if (rq && rq->end_io_data) @@ -115,17 +116,32 @@ EXPORT_SYMBOL_GPL(dm_get_rq_mapinfo); #define DMF_DELETING 4 #define DMF_NOFLUSH_SUSPENDING 5 #define DMF_MERGE_IS_OPTIONAL 6 +#define DMF_DEFERRED_REMOVE 7 + +/* + * A dummy definition to make RCU happy. + * struct dm_table should never be dereferenced in this file. + */ +struct dm_table { + int undefined__; +}; /* * Work processed by per-device workqueue. */ struct mapped_device { - struct rw_semaphore io_lock; + struct srcu_struct io_barrier; struct mutex suspend_lock; - rwlock_t map_lock; atomic_t holders; atomic_t open_count; + /* + * The current mapping. + * Use dm_get_live_table{_fast} or take suspend_lock for + * dereference. + */ + struct dm_table *map; + unsigned long flags; struct request_queue *queue; @@ -155,11 +171,6 @@ struct mapped_device { struct workqueue_struct *wq; /* - * The current mapping. - */ - struct dm_table *map; - - /* * io objects are allocated from here. */ mempool_t *io_pool; @@ -184,11 +195,13 @@ struct mapped_device { /* forced geometry settings */ struct hd_geometry geometry; - /* sysfs handle */ - struct kobject kobj; + /* kobject and completion */ + struct dm_kobject_holder kobj_holder; /* zero-length flush that will be cloned and submitted to targets */ struct bio flush_bio; + + struct dm_stats stats; }; /* @@ -199,10 +212,55 @@ struct dm_md_mempools { struct bio_set *bs; }; -#define MIN_IOS 256 +#define RESERVED_BIO_BASED_IOS 16 +#define RESERVED_REQUEST_BASED_IOS 256 +#define RESERVED_MAX_IOS 1024 static struct kmem_cache *_io_cache; static struct kmem_cache *_rq_tio_cache; +/* + * Bio-based DM's mempools' reserved IOs set by the user. + */ +static unsigned reserved_bio_based_ios = RESERVED_BIO_BASED_IOS; + +/* + * Request-based DM's mempools' reserved IOs set by the user. + */ +static unsigned reserved_rq_based_ios = RESERVED_REQUEST_BASED_IOS; + +static unsigned __dm_get_reserved_ios(unsigned *reserved_ios, + unsigned def, unsigned max) +{ + unsigned ios = ACCESS_ONCE(*reserved_ios); + unsigned modified_ios = 0; + + if (!ios) + modified_ios = def; + else if (ios > max) + modified_ios = max; + + if (modified_ios) { + (void)cmpxchg(reserved_ios, ios, modified_ios); + ios = modified_ios; + } + + return ios; +} + +unsigned dm_get_reserved_bio_based_ios(void) +{ + return __dm_get_reserved_ios(&reserved_bio_based_ios, + RESERVED_BIO_BASED_IOS, RESERVED_MAX_IOS); +} +EXPORT_SYMBOL_GPL(dm_get_reserved_bio_based_ios); + +unsigned dm_get_reserved_rq_based_ios(void) +{ + return __dm_get_reserved_ios(&reserved_rq_based_ios, + RESERVED_REQUEST_BASED_IOS, RESERVED_MAX_IOS); +} +EXPORT_SYMBOL_GPL(dm_get_reserved_rq_based_ios); + static int __init local_init(void) { int r = -ENOMEM; @@ -220,16 +278,24 @@ static int __init local_init(void) if (r) goto out_free_rq_tio_cache; + deferred_remove_workqueue = alloc_workqueue("kdmremove", WQ_UNBOUND, 1); + if (!deferred_remove_workqueue) { + r = -ENOMEM; + goto out_uevent_exit; + } + _major = major; r = register_blkdev(_major, _name); if (r < 0) - goto out_uevent_exit; + goto out_free_workqueue; if (!_major) _major = r; return 0; +out_free_workqueue: + destroy_workqueue(deferred_remove_workqueue); out_uevent_exit: dm_uevent_exit(); out_free_rq_tio_cache: @@ -242,6 +308,9 @@ out_free_io_cache: static void local_exit(void) { + flush_scheduled_work(); + destroy_workqueue(deferred_remove_workqueue); + kmem_cache_destroy(_rq_tio_cache); kmem_cache_destroy(_io_cache); unregister_blkdev(_major, _name); @@ -260,6 +329,7 @@ static int (*_inits[])(void) __initdata = { dm_io_init, dm_kcopyd_init, dm_interface_init, + dm_statistics_init, }; static void (*_exits[])(void) = { @@ -270,6 +340,7 @@ static void (*_exits[])(void) = { dm_io_exit, dm_kcopyd_exit, dm_interface_exit, + dm_statistics_exit, }; static int __init dm_init(void) @@ -345,7 +416,10 @@ static void dm_blk_close(struct gendisk *disk, fmode_t mode) spin_lock(&_minor_lock); - atomic_dec(&md->open_count); + if (atomic_dec_and_test(&md->open_count) && + (test_bit(DMF_DEFERRED_REMOVE, &md->flags))) + queue_work(deferred_remove_workqueue, &deferred_remove_work); + dm_put(md); spin_unlock(&_minor_lock); @@ -359,14 +433,18 @@ int dm_open_count(struct mapped_device *md) /* * Guarantees nothing is using the device before it's deleted. */ -int dm_lock_for_deletion(struct mapped_device *md) +int dm_lock_for_deletion(struct mapped_device *md, bool mark_deferred, bool only_deferred) { int r = 0; spin_lock(&_minor_lock); - if (dm_open_count(md)) + if (dm_open_count(md)) { r = -EBUSY; + if (mark_deferred) + set_bit(DMF_DEFERRED_REMOVE, &md->flags); + } else if (only_deferred && !test_bit(DMF_DEFERRED_REMOVE, &md->flags)) + r = -EEXIST; else set_bit(DMF_DELETING, &md->flags); @@ -375,6 +453,42 @@ int dm_lock_for_deletion(struct mapped_device *md) return r; } +int dm_cancel_deferred_remove(struct mapped_device *md) +{ + int r = 0; + + spin_lock(&_minor_lock); + + if (test_bit(DMF_DELETING, &md->flags)) + r = -EBUSY; + else + clear_bit(DMF_DEFERRED_REMOVE, &md->flags); + + spin_unlock(&_minor_lock); + + return r; +} + +static void do_deferred_remove(struct work_struct *w) +{ + dm_deferred_remove(); +} + +sector_t dm_get_size(struct mapped_device *md) +{ + return get_capacity(md->disk); +} + +struct request_queue *dm_get_md_queue(struct mapped_device *md) +{ + return md->queue; +} + +struct dm_stats *dm_get_stats(struct mapped_device *md) +{ + return &md->stats; +} + static int dm_blk_getgeo(struct block_device *bdev, struct hd_geometry *geo) { struct mapped_device *md = bdev->bd_disk->private_data; @@ -386,10 +500,14 @@ static int dm_blk_ioctl(struct block_device *bdev, fmode_t mode, unsigned int cmd, unsigned long arg) { struct mapped_device *md = bdev->bd_disk->private_data; - struct dm_table *map = dm_get_live_table(md); + int srcu_idx; + struct dm_table *map; struct dm_target *tgt; int r = -ENOTTY; +retry: + map = dm_get_live_table(md, &srcu_idx); + if (!map || !dm_table_get_size(map)) goto out; @@ -408,7 +526,12 @@ static int dm_blk_ioctl(struct block_device *bdev, fmode_t mode, r = tgt->type->ioctl(tgt, cmd, arg); out: - dm_table_put(map); + dm_put_live_table(md, srcu_idx); + + if (r == -ENOTCONN) { + msleep(10); + goto retry; + } return r; } @@ -448,8 +571,9 @@ static int md_in_flight(struct mapped_device *md) static void start_io_acct(struct dm_io *io) { struct mapped_device *md = io->md; + struct bio *bio = io->bio; int cpu; - int rw = bio_data_dir(io->bio); + int rw = bio_data_dir(bio); io->start_time = jiffies; @@ -458,6 +582,10 @@ static void start_io_acct(struct dm_io *io) part_stat_unlock(); atomic_set(&dm_disk(md)->part0.in_flight[rw], atomic_inc_return(&md->pending[rw])); + + if (unlikely(dm_stats_used(&md->stats))) + dm_stats_account_io(&md->stats, bio->bi_rw, bio->bi_iter.bi_sector, + bio_sectors(bio), false, 0, &io->stats_aux); } static void end_io_acct(struct dm_io *io) @@ -473,6 +601,10 @@ static void end_io_acct(struct dm_io *io) part_stat_add(cpu, &dm_disk(md)->part0, ticks[rw], duration); part_stat_unlock(); + if (unlikely(dm_stats_used(&md->stats))) + dm_stats_account_io(&md->stats, bio->bi_rw, bio->bi_iter.bi_sector, + bio_sectors(bio), true, duration, &io->stats_aux); + /* * After this is decremented the bio must not be touched if it is * a flush. @@ -502,20 +634,39 @@ static void queue_io(struct mapped_device *md, struct bio *bio) /* * Everyone (including functions in this file), should use this * function to access the md->map field, and make sure they call - * dm_table_put() when finished. + * dm_put_live_table() when finished. */ -struct dm_table *dm_get_live_table(struct mapped_device *md) +struct dm_table *dm_get_live_table(struct mapped_device *md, int *srcu_idx) __acquires(md->io_barrier) { - struct dm_table *t; - unsigned long flags; + *srcu_idx = srcu_read_lock(&md->io_barrier); - read_lock_irqsave(&md->map_lock, flags); - t = md->map; - if (t) - dm_table_get(t); - read_unlock_irqrestore(&md->map_lock, flags); + return srcu_dereference(md->map, &md->io_barrier); +} - return t; +void dm_put_live_table(struct mapped_device *md, int srcu_idx) __releases(md->io_barrier) +{ + srcu_read_unlock(&md->io_barrier, srcu_idx); +} + +void dm_sync_table(struct mapped_device *md) +{ + synchronize_srcu(&md->io_barrier); + synchronize_rcu_expedited(); +} + +/* + * A fast alternative to dm_get_live_table/dm_put_live_table. + * The caller must not block between these two functions. + */ +static struct dm_table *dm_get_live_table_fast(struct mapped_device *md) __acquires(RCU) +{ + rcu_read_lock(); + return rcu_dereference(md->map); +} + +static void dm_put_live_table_fast(struct mapped_device *md) __releases(RCU) +{ + rcu_read_unlock(); } /* @@ -600,7 +751,7 @@ static void dec_pending(struct dm_io *io, int error) if (io_error == DM_ENDIO_REQUEUE) return; - if ((bio->bi_rw & REQ_FLUSH) && bio->bi_size) { + if ((bio->bi_rw & REQ_FLUSH) && bio->bi_iter.bi_size) { /* * Preflush done for flush with data, reissue * without REQ_FLUSH. @@ -615,10 +766,18 @@ static void dec_pending(struct dm_io *io, int error) } } +static void disable_write_same(struct mapped_device *md) +{ + struct queue_limits *limits = dm_get_queue_limits(md); + + /* device doesn't really support WRITE SAME, disable it */ + limits->max_write_same_sectors = 0; +} + static void clone_endio(struct bio *bio, int error) { int r = 0; - struct dm_target_io *tio = bio->bi_private; + struct dm_target_io *tio = container_of(bio, struct dm_target_io, clone); struct dm_io *io = tio->io; struct mapped_device *md = tio->io->md; dm_endio_fn endio = tio->ti->type->end_io; @@ -643,6 +802,10 @@ static void clone_endio(struct bio *bio, int error) } } + if (unlikely(r == -EREMOTEIO && (bio->bi_rw & REQ_WRITE_SAME) && + !bdev_get_queue(bio->bi_bdev)->limits.max_write_same_sectors)) + disable_write_same(md); + free_tio(md, tio); dec_pending(io, error); } @@ -652,10 +815,11 @@ static void clone_endio(struct bio *bio, int error) */ static void end_clone_bio(struct bio *clone, int error) { - struct dm_rq_clone_bio_info *info = clone->bi_private; + struct dm_rq_clone_bio_info *info = + container_of(clone, struct dm_rq_clone_bio_info, clone); struct dm_rq_target_io *tio = info->tio; struct bio *bio = info->orig; - unsigned int nr_bytes = info->orig->bi_size; + unsigned int nr_bytes = info->orig->bi_iter.bi_size; bio_put(clone); @@ -836,6 +1000,10 @@ static void dm_done(struct request *clone, int error, bool mapped) r = rq_end_io(tio->ti, clone, error, &tio->info); } + if (unlikely(r == -EREMOTEIO && (clone->cmd_flags & REQ_WRITE_SAME) && + !clone->q->limits.max_write_same_sectors)) + disable_write_same(tio->md); + if (r <= 0) /* The target wants to complete the I/O */ dm_end_request(clone, r); @@ -969,6 +1137,46 @@ int dm_set_target_max_io_len(struct dm_target *ti, sector_t len) } EXPORT_SYMBOL_GPL(dm_set_target_max_io_len); +/* + * A target may call dm_accept_partial_bio only from the map routine. It is + * allowed for all bio types except REQ_FLUSH. + * + * dm_accept_partial_bio informs the dm that the target only wants to process + * additional n_sectors sectors of the bio and the rest of the data should be + * sent in a next bio. + * + * A diagram that explains the arithmetics: + * +--------------------+---------------+-------+ + * | 1 | 2 | 3 | + * +--------------------+---------------+-------+ + * + * <-------------- *tio->len_ptr ---------------> + * <------- bi_size -------> + * <-- n_sectors --> + * + * Region 1 was already iterated over with bio_advance or similar function. + * (it may be empty if the target doesn't use bio_advance) + * Region 2 is the remaining bio size that the target wants to process. + * (it may be empty if region 1 is non-empty, although there is no reason + * to make it empty) + * The target requires that region 3 is to be sent in the next bio. + * + * If the target wants to receive multiple copies of the bio (via num_*bios, etc), + * the partially processed part (the sum of regions 1+2) must be the same for all + * copies of the bio. + */ +void dm_accept_partial_bio(struct bio *bio, unsigned n_sectors) +{ + struct dm_target_io *tio = container_of(bio, struct dm_target_io, clone); + unsigned bi_size = bio->bi_iter.bi_size >> SECTOR_SHIFT; + BUG_ON(bio->bi_rw & REQ_FLUSH); + BUG_ON(bi_size > *tio->len_ptr); + BUG_ON(n_sectors > bi_size); + *tio->len_ptr -= bi_size - n_sectors; + bio->bi_iter.bi_size = n_sectors << SECTOR_SHIFT; +} +EXPORT_SYMBOL_GPL(dm_accept_partial_bio); + static void __map_bio(struct dm_target_io *tio) { int r; @@ -978,7 +1186,6 @@ static void __map_bio(struct dm_target_io *tio) struct dm_target *ti = tio->ti; clone->bi_end_io = clone_endio; - clone->bi_private = tio; /* * Map the clone. If r == 0 we don't need to do @@ -986,7 +1193,7 @@ static void __map_bio(struct dm_target_io *tio) * this io. */ atomic_inc(&tio->io->io_count); - sector = clone->bi_sector; + sector = clone->bi_iter.bi_sector; r = ti->type->map(ti, clone); if (r == DM_MAPIO_REMAPPED) { /* the bio has been remapped so dispatch it */ @@ -1012,77 +1219,33 @@ struct clone_info { struct bio *bio; struct dm_io *io; sector_t sector; - sector_t sector_count; - unsigned short idx; + unsigned sector_count; }; -static void bio_setup_sector(struct bio *bio, sector_t sector, sector_t len) -{ - bio->bi_sector = sector; - bio->bi_size = to_bytes(len); -} - -static void bio_setup_bv(struct bio *bio, unsigned short idx, unsigned short bv_count) +static void bio_setup_sector(struct bio *bio, sector_t sector, unsigned len) { - bio->bi_idx = idx; - bio->bi_vcnt = idx + bv_count; - bio->bi_flags &= ~(1 << BIO_SEG_VALID); -} - -static void clone_bio_integrity(struct bio *bio, struct bio *clone, - unsigned short idx, unsigned len, unsigned offset, - unsigned trim) -{ - if (!bio_integrity(bio)) - return; - - bio_integrity_clone(clone, bio, GFP_NOIO); - - if (trim) - bio_integrity_trim(clone, bio_sector_offset(bio, idx, offset), len); -} - -/* - * Creates a little bio that just does part of a bvec. - */ -static void clone_split_bio(struct dm_target_io *tio, struct bio *bio, - sector_t sector, unsigned short idx, - unsigned offset, unsigned len) -{ - struct bio *clone = &tio->clone; - struct bio_vec *bv = bio->bi_io_vec + idx; - - *clone->bi_io_vec = *bv; - - bio_setup_sector(clone, sector, len); - - clone->bi_bdev = bio->bi_bdev; - clone->bi_rw = bio->bi_rw; - clone->bi_vcnt = 1; - clone->bi_io_vec->bv_offset = offset; - clone->bi_io_vec->bv_len = clone->bi_size; - clone->bi_flags |= 1 << BIO_CLONED; - - clone_bio_integrity(bio, clone, idx, len, offset, 1); + bio->bi_iter.bi_sector = sector; + bio->bi_iter.bi_size = to_bytes(len); } /* * Creates a bio that consists of range of complete bvecs. */ static void clone_bio(struct dm_target_io *tio, struct bio *bio, - sector_t sector, unsigned short idx, - unsigned short bv_count, unsigned len) + sector_t sector, unsigned len) { struct bio *clone = &tio->clone; - unsigned trim = 0; - __bio_clone(clone, bio); - bio_setup_sector(clone, sector, len); - bio_setup_bv(clone, idx, bv_count); + __bio_clone_fast(clone, bio); + + if (bio_integrity(bio)) + bio_integrity_clone(clone, bio, GFP_NOIO); - if (idx != bio->bi_idx || clone->bi_size < bio->bi_size) - trim = 1; - clone_bio_integrity(bio, clone, idx, len, 0, trim); + bio_advance(clone, to_bytes(sector - clone->bi_iter.bi_sector)); + clone->bi_iter.bi_size = to_bytes(len); + + if (bio_integrity(bio)) + bio_integrity_trim(clone, 0, len); } static struct dm_target_io *alloc_tio(struct clone_info *ci, @@ -1097,7 +1260,6 @@ static struct dm_target_io *alloc_tio(struct clone_info *ci, tio->io = ci->io; tio->ti = ti; - memset(&tio->info, 0, sizeof(tio->info)); tio->target_bio_nr = target_bio_nr; return tio; @@ -1105,25 +1267,27 @@ static struct dm_target_io *alloc_tio(struct clone_info *ci, static void __clone_and_map_simple_bio(struct clone_info *ci, struct dm_target *ti, - unsigned target_bio_nr, sector_t len) + unsigned target_bio_nr, unsigned *len) { struct dm_target_io *tio = alloc_tio(ci, ti, ci->bio->bi_max_vecs, target_bio_nr); struct bio *clone = &tio->clone; + tio->len_ptr = len; + /* * Discard requests require the bio's inline iovecs be initialized. * ci->bio->bi_max_vecs is BIO_INLINE_VECS anyway, for both flush * and discard, so no need for concern about wasted bvec allocations. */ - __bio_clone(clone, ci->bio); + __bio_clone_fast(clone, ci->bio); if (len) - bio_setup_sector(clone, ci->sector, len); + bio_setup_sector(clone, ci->sector, *len); __map_bio(tio); } static void __send_duplicate_bios(struct clone_info *ci, struct dm_target *ti, - unsigned num_bios, sector_t len) + unsigned num_bios, unsigned *len) { unsigned target_bio_nr; @@ -1138,16 +1302,13 @@ static int __send_empty_flush(struct clone_info *ci) BUG_ON(bio_has_data(ci->bio)); while ((ti = dm_table_get_target(ci->map, target_nr++))) - __send_duplicate_bios(ci, ti, ti->num_flush_bios, 0); + __send_duplicate_bios(ci, ti, ti->num_flush_bios, NULL); return 0; } static void __clone_and_map_data_bio(struct clone_info *ci, struct dm_target *ti, - sector_t sector, int nr_iovecs, - unsigned short idx, unsigned short bv_count, - unsigned offset, unsigned len, - unsigned split_bvec) + sector_t sector, unsigned *len) { struct bio *bio = ci->bio; struct dm_target_io *tio; @@ -1161,11 +1322,9 @@ static void __clone_and_map_data_bio(struct clone_info *ci, struct dm_target *ti num_target_bios = ti->num_write_bios(ti, bio); for (target_bio_nr = 0; target_bio_nr < num_target_bios; target_bio_nr++) { - tio = alloc_tio(ci, ti, nr_iovecs, target_bio_nr); - if (split_bvec) - clone_split_bio(tio, bio, sector, idx, offset, len); - else - clone_bio(tio, bio, sector, idx, bv_count, len); + tio = alloc_tio(ci, ti, 0, target_bio_nr); + tio->len_ptr = len; + clone_bio(tio, bio, sector, *len); __map_bio(tio); } } @@ -1194,7 +1353,7 @@ static int __send_changing_extent_only(struct clone_info *ci, is_split_required_fn is_split_required) { struct dm_target *ti; - sector_t len; + unsigned len; unsigned num_bios; do { @@ -1213,11 +1372,11 @@ static int __send_changing_extent_only(struct clone_info *ci, return -EOPNOTSUPP; if (is_split_required && !is_split_required(ti)) - len = min(ci->sector_count, max_io_len_target_boundary(ci->sector, ti)); + len = min((sector_t)ci->sector_count, max_io_len_target_boundary(ci->sector, ti)); else - len = min(ci->sector_count, max_io_len(ci->sector, ti)); + len = min((sector_t)ci->sector_count, max_io_len(ci->sector, ti)); - __send_duplicate_bios(ci, ti, num_bios, len); + __send_duplicate_bios(ci, ti, num_bios, &len); ci->sector += len; } while (ci->sector_count -= len); @@ -1237,68 +1396,13 @@ static int __send_write_same(struct clone_info *ci) } /* - * Find maximum number of sectors / bvecs we can process with a single bio. - */ -static sector_t __len_within_target(struct clone_info *ci, sector_t max, int *idx) -{ - struct bio *bio = ci->bio; - sector_t bv_len, total_len = 0; - - for (*idx = ci->idx; max && (*idx < bio->bi_vcnt); (*idx)++) { - bv_len = to_sector(bio->bi_io_vec[*idx].bv_len); - - if (bv_len > max) - break; - - max -= bv_len; - total_len += bv_len; - } - - return total_len; -} - -static int __split_bvec_across_targets(struct clone_info *ci, - struct dm_target *ti, sector_t max) -{ - struct bio *bio = ci->bio; - struct bio_vec *bv = bio->bi_io_vec + ci->idx; - sector_t remaining = to_sector(bv->bv_len); - unsigned offset = 0; - sector_t len; - - do { - if (offset) { - ti = dm_table_find_target(ci->map, ci->sector); - if (!dm_target_is_valid(ti)) - return -EIO; - - max = max_io_len(ci->sector, ti); - } - - len = min(remaining, max); - - __clone_and_map_data_bio(ci, ti, ci->sector, 1, ci->idx, 0, - bv->bv_offset + offset, len, 1); - - ci->sector += len; - ci->sector_count -= len; - offset += to_bytes(len); - } while (remaining -= len); - - ci->idx++; - - return 0; -} - -/* * Select the correct strategy for processing a non-flush bio. */ static int __split_and_process_non_flush(struct clone_info *ci) { struct bio *bio = ci->bio; struct dm_target *ti; - sector_t len, max; - int idx; + unsigned len; if (unlikely(bio->bi_rw & REQ_DISCARD)) return __send_discard(ci); @@ -1309,57 +1413,31 @@ static int __split_and_process_non_flush(struct clone_info *ci) if (!dm_target_is_valid(ti)) return -EIO; - max = max_io_len(ci->sector, ti); - - /* - * Optimise for the simple case where we can do all of - * the remaining io with a single clone. - */ - if (ci->sector_count <= max) { - __clone_and_map_data_bio(ci, ti, ci->sector, bio->bi_max_vecs, - ci->idx, bio->bi_vcnt - ci->idx, 0, - ci->sector_count, 0); - ci->sector_count = 0; - return 0; - } - - /* - * There are some bvecs that don't span targets. - * Do as many of these as possible. - */ - if (to_sector(bio->bi_io_vec[ci->idx].bv_len) <= max) { - len = __len_within_target(ci, max, &idx); + len = min_t(sector_t, max_io_len(ci->sector, ti), ci->sector_count); - __clone_and_map_data_bio(ci, ti, ci->sector, bio->bi_max_vecs, - ci->idx, idx - ci->idx, 0, len, 0); + __clone_and_map_data_bio(ci, ti, ci->sector, &len); - ci->sector += len; - ci->sector_count -= len; - ci->idx = idx; + ci->sector += len; + ci->sector_count -= len; - return 0; - } - - /* - * Handle a bvec that must be split between two or more targets. - */ - return __split_bvec_across_targets(ci, ti, max); + return 0; } /* * Entry point to split a bio into clones and submit them to the targets. */ -static void __split_and_process_bio(struct mapped_device *md, struct bio *bio) +static void __split_and_process_bio(struct mapped_device *md, + struct dm_table *map, struct bio *bio) { struct clone_info ci; int error = 0; - ci.map = dm_get_live_table(md); - if (unlikely(!ci.map)) { + if (unlikely(!map)) { bio_io_error(bio); return; } + ci.map = map; ci.md = md; ci.io = alloc_io(md); ci.io->error = 0; @@ -1367,8 +1445,7 @@ static void __split_and_process_bio(struct mapped_device *md, struct bio *bio) ci.io->bio = bio; ci.io->md = md; spin_lock_init(&ci.io->endio_lock); - ci.sector = bio->bi_sector; - ci.idx = bio->bi_idx; + ci.sector = bio->bi_iter.bi_sector; start_io_acct(ci.io); @@ -1386,7 +1463,6 @@ static void __split_and_process_bio(struct mapped_device *md, struct bio *bio) /* drop the extra reference count */ dec_pending(ci.io, error); - dm_table_put(ci.map); } /*----------------------------------------------------------------- * CRUD END @@ -1397,7 +1473,7 @@ static int dm_merge_bvec(struct request_queue *q, struct bio_vec *biovec) { struct mapped_device *md = q->queuedata; - struct dm_table *map = dm_get_live_table(md); + struct dm_table *map = dm_get_live_table_fast(md); struct dm_target *ti; sector_t max_sectors; int max_size = 0; @@ -1407,7 +1483,7 @@ static int dm_merge_bvec(struct request_queue *q, ti = dm_table_find_target(map, bvm->bi_sector); if (!dm_target_is_valid(ti)) - goto out_table; + goto out; /* * Find maximum amount of I/O that won't need splitting @@ -1433,13 +1509,10 @@ static int dm_merge_bvec(struct request_queue *q, * just one page. */ else if (queue_max_hw_sectors(q) <= PAGE_SIZE >> 9) - max_size = 0; -out_table: - dm_table_put(map); - out: + dm_put_live_table_fast(md); /* * Always allow an entire first page */ @@ -1458,8 +1531,10 @@ static void _dm_request(struct request_queue *q, struct bio *bio) int rw = bio_data_dir(bio); struct mapped_device *md = q->queuedata; int cpu; + int srcu_idx; + struct dm_table *map; - down_read(&md->io_lock); + map = dm_get_live_table(md, &srcu_idx); cpu = part_stat_lock(); part_stat_inc(cpu, &dm_disk(md)->part0, ios[rw]); @@ -1468,7 +1543,7 @@ static void _dm_request(struct request_queue *q, struct bio *bio) /* if we're suspended, we have to queue this io for later */ if (unlikely(test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags))) { - up_read(&md->io_lock); + dm_put_live_table(md, srcu_idx); if (bio_rw(bio) != READA) queue_io(md, bio); @@ -1477,12 +1552,12 @@ static void _dm_request(struct request_queue *q, struct bio *bio) return; } - __split_and_process_bio(md, bio); - up_read(&md->io_lock); + __split_and_process_bio(md, map, bio); + dm_put_live_table(md, srcu_idx); return; } -static int dm_request_based(struct mapped_device *md) +int dm_request_based(struct mapped_device *md) { return blk_queue_stackable(md->queue); } @@ -1521,7 +1596,6 @@ static int dm_rq_bio_constructor(struct bio *bio, struct bio *bio_orig, info->orig = bio_orig; info->tio = tio; bio->bi_end_io = end_clone_bio; - bio->bi_private = info; return 0; } @@ -1539,7 +1613,6 @@ static int setup_clone(struct request *clone, struct request *rq, clone->cmd = rq->cmd; clone->cmd_len = rq->cmd_len; clone->sense = rq->sense; - clone->buffer = rq->buffer; clone->end_io = end_clone_request; clone->end_io_data = tio; @@ -1664,7 +1737,8 @@ static struct request *dm_start_request(struct mapped_device *md, struct request static void dm_request_fn(struct request_queue *q) { struct mapped_device *md = q->queuedata; - struct dm_table *map = dm_get_live_table(md); + int srcu_idx; + struct dm_table *map = dm_get_live_table(md, &srcu_idx); struct dm_target *ti; struct request *rq, *clone; sector_t pos; @@ -1719,7 +1793,7 @@ requeued: delay_and_out: blk_delay_queue(q, HZ / 10); out: - dm_table_put(map); + dm_put_live_table(md, srcu_idx); } int dm_underlying_device_busy(struct request_queue *q) @@ -1732,14 +1806,14 @@ static int dm_lld_busy(struct request_queue *q) { int r; struct mapped_device *md = q->queuedata; - struct dm_table *map = dm_get_live_table(md); + struct dm_table *map = dm_get_live_table_fast(md); if (!map || test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags)) r = 1; else r = dm_table_any_busy_target(map); - dm_table_put(map); + dm_put_live_table_fast(md); return r; } @@ -1751,7 +1825,7 @@ static int dm_any_congested(void *congested_data, int bdi_bits) struct dm_table *map; if (!test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags)) { - map = dm_get_live_table(md); + map = dm_get_live_table_fast(md); if (map) { /* * Request-based dm cares about only own queue for @@ -1762,9 +1836,8 @@ static int dm_any_congested(void *congested_data, int bdi_bits) bdi_bits; else r = dm_table_any_congested(map, bdi_bits); - - dm_table_put(map); } + dm_put_live_table_fast(md); } return r; @@ -1869,12 +1942,14 @@ static struct mapped_device *alloc_dev(int minor) if (r < 0) goto bad_minor; + r = init_srcu_struct(&md->io_barrier); + if (r < 0) + goto bad_io_barrier; + md->type = DM_TYPE_NONE; - init_rwsem(&md->io_lock); mutex_init(&md->suspend_lock); mutex_init(&md->type_lock); spin_lock_init(&md->deferred_lock); - rwlock_init(&md->map_lock); atomic_set(&md->holders, 1); atomic_set(&md->open_count, 0); atomic_set(&md->event_nr, 0); @@ -1897,6 +1972,7 @@ static struct mapped_device *alloc_dev(int minor) init_waitqueue_head(&md->wait); INIT_WORK(&md->work, dm_wq_work); init_waitqueue_head(&md->eventq); + init_completion(&md->kobj_holder.completion); md->disk->major = _major; md->disk->first_minor = minor; @@ -1907,8 +1983,7 @@ static struct mapped_device *alloc_dev(int minor) add_disk(md->disk); format_dev_t(md->name, MKDEV(_major, minor)); - md->wq = alloc_workqueue("kdmflush", - WQ_NON_REENTRANT | WQ_MEM_RECLAIM, 0); + md->wq = alloc_workqueue("kdmflush", WQ_MEM_RECLAIM, 0); if (!md->wq) goto bad_thread; @@ -1920,6 +1995,8 @@ static struct mapped_device *alloc_dev(int minor) md->flush_bio.bi_bdev = md->bdev; md->flush_bio.bi_rw = WRITE_FLUSH; + dm_stats_init(&md->stats); + /* Populate the mapping, nobody knows we exist yet */ spin_lock(&_minor_lock); old_md = idr_replace(&_minor_idr, md, minor); @@ -1937,6 +2014,8 @@ bad_thread: bad_disk: blk_cleanup_queue(md->queue); bad_queue: + cleanup_srcu_struct(&md->io_barrier); +bad_io_barrier: free_minor(minor); bad_minor: module_put(THIS_MODULE); @@ -1960,6 +2039,7 @@ static void free_dev(struct mapped_device *md) bioset_free(md->bs); blk_integrity_unregister(md->disk); del_gendisk(md->disk); + cleanup_srcu_struct(&md->io_barrier); free_minor(minor); spin_lock(&_minor_lock); @@ -1968,6 +2048,7 @@ static void free_dev(struct mapped_device *md) put_disk(md->disk); blk_cleanup_queue(md->queue); + dm_stats_cleanup(&md->stats); module_put(THIS_MODULE); kfree(md); } @@ -2102,7 +2183,6 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t, struct dm_table *old_map; struct request_queue *q = md->queue; sector_t size; - unsigned long flags; int merge_is_optional; size = dm_table_get_size(t); @@ -2110,7 +2190,7 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t, /* * Wipe any geometry if the size of the table changed. */ - if (size != get_capacity(md->disk)) + if (size != dm_get_size(md)) memset(&md->geometry, 0, sizeof(md->geometry)); __set_size(md, size); @@ -2131,9 +2211,8 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t, merge_is_optional = dm_table_merge_is_optional(t); - write_lock_irqsave(&md->map_lock, flags); old_map = md->map; - md->map = t; + rcu_assign_pointer(md->map, t); md->immutable_target_type = dm_table_get_immutable_target_type(t); dm_table_set_restrictions(t, q, limits); @@ -2141,7 +2220,7 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t, set_bit(DMF_MERGE_IS_OPTIONAL, &md->flags); else clear_bit(DMF_MERGE_IS_OPTIONAL, &md->flags); - write_unlock_irqrestore(&md->map_lock, flags); + dm_sync_table(md); return old_map; } @@ -2152,15 +2231,13 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t, static struct dm_table *__unbind(struct mapped_device *md) { struct dm_table *map = md->map; - unsigned long flags; if (!map) return NULL; dm_table_event_callback(map, NULL, NULL); - write_lock_irqsave(&md->map_lock, flags); - md->map = NULL; - write_unlock_irqrestore(&md->map_lock, flags); + RCU_INIT_POINTER(md->map, NULL); + dm_sync_table(md); return map; } @@ -2198,11 +2275,13 @@ void dm_unlock_md_type(struct mapped_device *md) void dm_set_md_type(struct mapped_device *md, unsigned type) { + BUG_ON(!mutex_is_locked(&md->type_lock)); md->type = type; } unsigned dm_get_md_type(struct mapped_device *md) { + BUG_ON(!mutex_is_locked(&md->type_lock)); return md->type; } @@ -2212,6 +2291,17 @@ struct target_type *dm_get_immutable_target_type(struct mapped_device *md) } /* + * The queue_limits are only valid as long as you have a reference + * count on 'md'. + */ +struct queue_limits *dm_get_queue_limits(struct mapped_device *md) +{ + BUG_ON(!atomic_read(&md->holders)); + return &md->queue->limits; +} +EXPORT_SYMBOL_GPL(dm_get_queue_limits); + +/* * Fully initialize a request-based queue (->elevator, ->request_fn, etc). */ static int dm_init_request_based_queue(struct mapped_device *md) @@ -2312,11 +2402,12 @@ EXPORT_SYMBOL_GPL(dm_device_name); static void __dm_destroy(struct mapped_device *md, bool wait) { struct dm_table *map; + int srcu_idx; might_sleep(); spin_lock(&_minor_lock); - map = dm_get_live_table(md); + map = dm_get_live_table(md, &srcu_idx); idr_replace(&_minor_idr, MINOR_ALLOCED, MINOR(disk_devt(dm_disk(md)))); set_bit(DMF_FREEING, &md->flags); spin_unlock(&_minor_lock); @@ -2326,6 +2417,9 @@ static void __dm_destroy(struct mapped_device *md, bool wait) dm_table_postsuspend_targets(map); } + /* dm_put_live_table must be before msleep, otherwise deadlock is possible */ + dm_put_live_table(md, srcu_idx); + /* * Rare, but there may be I/O requests still going to complete, * for example. Wait for all references to disappear. @@ -2340,7 +2434,6 @@ static void __dm_destroy(struct mapped_device *md, bool wait) dm_device_name(md), atomic_read(&md->holders)); dm_sysfs_exit(md); - dm_table_put(map); dm_table_destroy(__unbind(md)); free_dev(md); } @@ -2397,8 +2490,10 @@ static void dm_wq_work(struct work_struct *work) struct mapped_device *md = container_of(work, struct mapped_device, work); struct bio *c; + int srcu_idx; + struct dm_table *map; - down_read(&md->io_lock); + map = dm_get_live_table(md, &srcu_idx); while (!test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags)) { spin_lock_irq(&md->deferred_lock); @@ -2408,23 +2503,19 @@ static void dm_wq_work(struct work_struct *work) if (!c) break; - up_read(&md->io_lock); - if (dm_request_based(md)) generic_make_request(c); else - __split_and_process_bio(md, c); - - down_read(&md->io_lock); + __split_and_process_bio(md, map, c); } - up_read(&md->io_lock); + dm_put_live_table(md, srcu_idx); } static void dm_queue_flush(struct mapped_device *md) { clear_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); queue_work(md->wq, &md->work); } @@ -2450,10 +2541,10 @@ struct dm_table *dm_swap_table(struct mapped_device *md, struct dm_table *table) * reappear. */ if (dm_table_has_no_data_devices(table)) { - live_map = dm_get_live_table(md); + live_map = dm_get_live_table_fast(md); if (live_map) limits = md->queue->limits; - dm_table_put(live_map); + dm_put_live_table_fast(md); } if (!live_map) { @@ -2533,7 +2624,7 @@ int dm_suspend(struct mapped_device *md, unsigned suspend_flags) goto out_unlock; } - map = dm_get_live_table(md); + map = md->map; /* * DMF_NOFLUSH_SUSPENDING must be set before presuspend. @@ -2554,7 +2645,7 @@ int dm_suspend(struct mapped_device *md, unsigned suspend_flags) if (!noflush && do_lockfs) { r = lock_fs(md); if (r) - goto out; + goto out_unlock; } /* @@ -2569,9 +2660,8 @@ int dm_suspend(struct mapped_device *md, unsigned suspend_flags) * (dm_wq_work), we set BMF_BLOCK_IO_FOR_SUSPEND and call * flush_workqueue(md->wq). */ - down_write(&md->io_lock); set_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags); - up_write(&md->io_lock); + synchronize_srcu(&md->io_barrier); /* * Stop md->queue before flushing md->wq in case request-based @@ -2589,10 +2679,9 @@ int dm_suspend(struct mapped_device *md, unsigned suspend_flags) */ r = dm_wait_for_completion(md, TASK_INTERRUPTIBLE); - down_write(&md->io_lock); if (noflush) clear_bit(DMF_NOFLUSH_SUSPENDING, &md->flags); - up_write(&md->io_lock); + synchronize_srcu(&md->io_barrier); /* were we interrupted ? */ if (r < 0) { @@ -2602,7 +2691,7 @@ int dm_suspend(struct mapped_device *md, unsigned suspend_flags) start_queue(md->queue); unlock_fs(md); - goto out; /* pushback list is already flushed, so skip flush */ + goto out_unlock; /* pushback list is already flushed, so skip flush */ } /* @@ -2615,9 +2704,6 @@ int dm_suspend(struct mapped_device *md, unsigned suspend_flags) dm_table_postsuspend_targets(map); -out: - dm_table_put(map); - out_unlock: mutex_unlock(&md->suspend_lock); return r; @@ -2632,7 +2718,7 @@ int dm_resume(struct mapped_device *md) if (!dm_suspended_md(md)) goto out; - map = dm_get_live_table(md); + map = md->map; if (!map || !dm_table_get_size(map)) goto out; @@ -2656,12 +2742,43 @@ int dm_resume(struct mapped_device *md) r = 0; out: - dm_table_put(map); mutex_unlock(&md->suspend_lock); return r; } +/* + * Internal suspend/resume works like userspace-driven suspend. It waits + * until all bios finish and prevents issuing new bios to the target drivers. + * It may be used only from the kernel. + * + * Internal suspend holds md->suspend_lock, which prevents interaction with + * userspace-driven suspend. + */ + +void dm_internal_suspend(struct mapped_device *md) +{ + mutex_lock(&md->suspend_lock); + if (dm_suspended_md(md)) + return; + + set_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags); + synchronize_srcu(&md->io_barrier); + flush_workqueue(md->wq); + dm_wait_for_completion(md, TASK_UNINTERRUPTIBLE); +} + +void dm_internal_resume(struct mapped_device *md) +{ + if (dm_suspended_md(md)) + goto done; + + dm_queue_flush(md); + +done: + mutex_unlock(&md->suspend_lock); +} + /*----------------------------------------------------------------- * Event notification. *---------------------------------------------------------------*/ @@ -2717,20 +2834,14 @@ struct gendisk *dm_disk(struct mapped_device *md) struct kobject *dm_kobject(struct mapped_device *md) { - return &md->kobj; + return &md->kobj_holder.kobj; } -/* - * struct mapped_device should not be exported outside of dm.c - * so use this check to verify that kobj is part of md structure - */ struct mapped_device *dm_get_from_kobject(struct kobject *kobj) { struct mapped_device *md; - md = container_of(kobj, struct mapped_device, kobj); - if (&md->kobj != kobj) - return NULL; + md = container_of(kobj, struct mapped_device, kobj_holder.kobj); if (test_bit(DMF_FREEING, &md->flags) || dm_deleting_md(md)) @@ -2745,6 +2856,11 @@ int dm_suspended_md(struct mapped_device *md) return test_bit(DMF_SUSPENDED, &md->flags); } +int dm_test_deferred_remove_flag(struct mapped_device *md) +{ + return test_bit(DMF_DEFERRED_REMOVE, &md->flags); +} + int dm_suspended(struct dm_target *ti) { return dm_suspended_md(dm_table_get_md(ti->table)); @@ -2769,18 +2885,18 @@ struct dm_md_mempools *dm_alloc_md_mempools(unsigned type, unsigned integrity, u if (type == DM_TYPE_BIO_BASED) { cachep = _io_cache; - pool_size = 16; + pool_size = dm_get_reserved_bio_based_ios(); front_pad = roundup(per_bio_data_size, __alignof__(struct dm_target_io)) + offsetof(struct dm_target_io, clone); } else if (type == DM_TYPE_REQUEST_BASED) { cachep = _rq_tio_cache; - pool_size = MIN_IOS; + pool_size = dm_get_reserved_rq_based_ios(); front_pad = offsetof(struct dm_rq_clone_bio_info, clone); /* per_bio_data_size is not used. See __bind_mempools(). */ WARN_ON(per_bio_data_size != 0); } else goto out; - pools->io_pool = mempool_create_slab_pool(MIN_IOS, cachep); + pools->io_pool = mempool_create_slab_pool(pool_size, cachep); if (!pools->io_pool) goto out; @@ -2821,8 +2937,6 @@ static const struct block_device_operations dm_blk_dops = { .owner = THIS_MODULE }; -EXPORT_SYMBOL(dm_get_mapinfo); - /* * module hooks */ @@ -2831,6 +2945,13 @@ module_exit(dm_exit); module_param(major, uint, 0); MODULE_PARM_DESC(major, "The major number of the device mapper"); + +module_param(reserved_bio_based_ios, uint, S_IRUGO | S_IWUSR); +MODULE_PARM_DESC(reserved_bio_based_ios, "Reserved IOs in bio-based mempools"); + +module_param(reserved_rq_based_ios, uint, S_IRUGO | S_IWUSR); +MODULE_PARM_DESC(reserved_rq_based_ios, "Reserved IOs in request-based mempools"); + MODULE_DESCRIPTION(DM_NAME " driver"); MODULE_AUTHOR("Joe Thornber <dm-devel@redhat.com>"); MODULE_LICENSE("GPL"); diff --git a/drivers/md/dm.h b/drivers/md/dm.h index 45b97da1bd0..ed76126aac5 100644 --- a/drivers/md/dm.h +++ b/drivers/md/dm.h @@ -15,6 +15,10 @@ #include <linux/list.h> #include <linux/blkdev.h> #include <linux/hdreg.h> +#include <linux/completion.h> +#include <linux/kobject.h> + +#include "dm-stats.h" /* * Suspend feature flags @@ -69,7 +73,6 @@ unsigned dm_table_get_type(struct dm_table *t); struct target_type *dm_table_get_immutable_target_type(struct dm_table *t); bool dm_table_request_based(struct dm_table *t); bool dm_table_supports_discards(struct dm_table *t); -int dm_table_alloc_md_mempools(struct dm_table *t); void dm_table_free_md_mempools(struct dm_table *t); struct dm_md_mempools *dm_table_get_md_mempools(struct dm_table *t); @@ -89,10 +92,21 @@ int dm_setup_md_queue(struct mapped_device *md); #define dm_target_is_valid(t) ((t)->table) /* + * To check whether the target type is bio-based or not (request-based). + */ +#define dm_target_bio_based(t) ((t)->type->map != NULL) + +/* * To check whether the target type is request-based or not (bio-based). */ #define dm_target_request_based(t) ((t)->type->map_rq != NULL) +/* + * To check whether the target type is a hybrid (capable of being + * either request-based or bio-based). + */ +#define dm_target_hybrid(t) (dm_target_bio_based(t) && dm_target_request_based(t)) + /*----------------------------------------------------------------- * A registry of target types. *---------------------------------------------------------------*/ @@ -116,6 +130,16 @@ int dm_deleting_md(struct mapped_device *md); int dm_suspended_md(struct mapped_device *md); /* + * Test if the device is scheduled for deferred remove. + */ +int dm_test_deferred_remove_flag(struct mapped_device *md); + +/* + * Try to remove devices marked for deferred removal. + */ +void dm_deferred_remove(void); + +/* * The device-mapper can be driven through one of two interfaces; * ioctl or filesystem, depending which patch you have applied. */ @@ -125,12 +149,27 @@ void dm_interface_exit(void); /* * sysfs interface */ +struct dm_kobject_holder { + struct kobject kobj; + struct completion completion; +}; + +static inline struct completion *dm_get_completion_from_kobject(struct kobject *kobj) +{ + return &container_of(kobj, struct dm_kobject_holder, kobj)->completion; +} + int dm_sysfs_init(struct mapped_device *md); void dm_sysfs_exit(struct mapped_device *md); struct kobject *dm_kobject(struct mapped_device *md); struct mapped_device *dm_get_from_kobject(struct kobject *kobj); /* + * The kobject helper + */ +void dm_kobject_release(struct kobject *kobj); + +/* * Targets for linear and striped mappings */ int dm_linear_init(void); @@ -145,11 +184,19 @@ void dm_stripe_exit(void); void dm_destroy(struct mapped_device *md); void dm_destroy_immediate(struct mapped_device *md); int dm_open_count(struct mapped_device *md); -int dm_lock_for_deletion(struct mapped_device *md); +int dm_lock_for_deletion(struct mapped_device *md, bool mark_deferred, bool only_deferred); +int dm_cancel_deferred_remove(struct mapped_device *md); +int dm_request_based(struct mapped_device *md); +sector_t dm_get_size(struct mapped_device *md); +struct request_queue *dm_get_md_queue(struct mapped_device *md); +struct dm_stats *dm_get_stats(struct mapped_device *md); int dm_kobject_uevent(struct mapped_device *md, enum kobject_action action, unsigned cookie); +void dm_internal_suspend(struct mapped_device *md); +void dm_internal_resume(struct mapped_device *md); + int dm_io_init(void); void dm_io_exit(void); @@ -162,4 +209,15 @@ void dm_kcopyd_exit(void); struct dm_md_mempools *dm_alloc_md_mempools(unsigned type, unsigned integrity, unsigned per_bio_data_size); void dm_free_md_mempools(struct dm_md_mempools *pools); +/* + * Helpers that are used by DM core + */ +unsigned dm_get_reserved_bio_based_ios(void); +unsigned dm_get_reserved_rq_based_ios(void); + +static inline bool dm_message_test_buffer_overflow(char *result, unsigned maxlen) +{ + return !maxlen || strlen(result) + 1 >= maxlen; +} + #endif diff --git a/drivers/md/faulty.c b/drivers/md/faulty.c index 3193aefe982..e8b4574956c 100644 --- a/drivers/md/faulty.c +++ b/drivers/md/faulty.c @@ -74,8 +74,8 @@ static void faulty_fail(struct bio *bio, int error) { struct bio *b = bio->bi_private; - b->bi_size = bio->bi_size; - b->bi_sector = bio->bi_sector; + b->bi_iter.bi_size = bio->bi_iter.bi_size; + b->bi_iter.bi_sector = bio->bi_iter.bi_sector; bio_put(bio); @@ -185,26 +185,31 @@ static void make_request(struct mddev *mddev, struct bio *bio) return; } - if (check_sector(conf, bio->bi_sector, bio_end_sector(bio), WRITE)) + if (check_sector(conf, bio->bi_iter.bi_sector, + bio_end_sector(bio), WRITE)) failit = 1; if (check_mode(conf, WritePersistent)) { - add_sector(conf, bio->bi_sector, WritePersistent); + add_sector(conf, bio->bi_iter.bi_sector, + WritePersistent); failit = 1; } if (check_mode(conf, WriteTransient)) failit = 1; } else { /* read request */ - if (check_sector(conf, bio->bi_sector, bio_end_sector(bio), READ)) + if (check_sector(conf, bio->bi_iter.bi_sector, + bio_end_sector(bio), READ)) failit = 1; if (check_mode(conf, ReadTransient)) failit = 1; if (check_mode(conf, ReadPersistent)) { - add_sector(conf, bio->bi_sector, ReadPersistent); + add_sector(conf, bio->bi_iter.bi_sector, + ReadPersistent); failit = 1; } if (check_mode(conf, ReadFixable)) { - add_sector(conf, bio->bi_sector, ReadFixable); + add_sector(conf, bio->bi_iter.bi_sector, + ReadFixable); failit = 1; } } diff --git a/drivers/md/linear.c b/drivers/md/linear.c index f03fabd2b37..56f534b4a2d 100644 --- a/drivers/md/linear.c +++ b/drivers/md/linear.c @@ -288,65 +288,65 @@ static int linear_stop (struct mddev *mddev) static void linear_make_request(struct mddev *mddev, struct bio *bio) { + char b[BDEVNAME_SIZE]; struct dev_info *tmp_dev; - sector_t start_sector; + struct bio *split; + sector_t start_sector, end_sector, data_offset; if (unlikely(bio->bi_rw & REQ_FLUSH)) { md_flush_request(mddev, bio); return; } - rcu_read_lock(); - tmp_dev = which_dev(mddev, bio->bi_sector); - start_sector = tmp_dev->end_sector - tmp_dev->rdev->sectors; - - - if (unlikely(bio->bi_sector >= (tmp_dev->end_sector) - || (bio->bi_sector < start_sector))) { - char b[BDEVNAME_SIZE]; - - printk(KERN_ERR - "md/linear:%s: make_request: Sector %llu out of bounds on " - "dev %s: %llu sectors, offset %llu\n", - mdname(mddev), - (unsigned long long)bio->bi_sector, - bdevname(tmp_dev->rdev->bdev, b), - (unsigned long long)tmp_dev->rdev->sectors, - (unsigned long long)start_sector); - rcu_read_unlock(); - bio_io_error(bio); - return; - } - if (unlikely(bio_end_sector(bio) > tmp_dev->end_sector)) { - /* This bio crosses a device boundary, so we have to - * split it. - */ - struct bio_pair *bp; - sector_t end_sector = tmp_dev->end_sector; + do { + rcu_read_lock(); - rcu_read_unlock(); - - bp = bio_split(bio, end_sector - bio->bi_sector); + tmp_dev = which_dev(mddev, bio->bi_iter.bi_sector); + start_sector = tmp_dev->end_sector - tmp_dev->rdev->sectors; + end_sector = tmp_dev->end_sector; + data_offset = tmp_dev->rdev->data_offset; + bio->bi_bdev = tmp_dev->rdev->bdev; - linear_make_request(mddev, &bp->bio1); - linear_make_request(mddev, &bp->bio2); - bio_pair_release(bp); - return; - } - - bio->bi_bdev = tmp_dev->rdev->bdev; - bio->bi_sector = bio->bi_sector - start_sector - + tmp_dev->rdev->data_offset; - rcu_read_unlock(); + rcu_read_unlock(); - if (unlikely((bio->bi_rw & REQ_DISCARD) && - !blk_queue_discard(bdev_get_queue(bio->bi_bdev)))) { - /* Just ignore it */ - bio_endio(bio, 0); - return; - } + if (unlikely(bio->bi_iter.bi_sector >= end_sector || + bio->bi_iter.bi_sector < start_sector)) + goto out_of_bounds; + + if (unlikely(bio_end_sector(bio) > end_sector)) { + /* This bio crosses a device boundary, so we have to + * split it. + */ + split = bio_split(bio, end_sector - + bio->bi_iter.bi_sector, + GFP_NOIO, fs_bio_set); + bio_chain(split, bio); + } else { + split = bio; + } - generic_make_request(bio); + split->bi_iter.bi_sector = split->bi_iter.bi_sector - + start_sector + data_offset; + + if (unlikely((split->bi_rw & REQ_DISCARD) && + !blk_queue_discard(bdev_get_queue(split->bi_bdev)))) { + /* Just ignore it */ + bio_endio(split, 0); + } else + generic_make_request(split); + } while (split != bio); + return; + +out_of_bounds: + printk(KERN_ERR + "md/linear:%s: make_request: Sector %llu out of bounds on " + "dev %s: %llu sectors, offset %llu\n", + mdname(mddev), + (unsigned long long)bio->bi_iter.bi_sector, + bdevname(tmp_dev->rdev->bdev, b), + (unsigned long long)tmp_dev->rdev->sectors, + (unsigned long long)start_sector); + bio_io_error(bio); } static void linear_status (struct seq_file *seq, struct mddev *mddev) diff --git a/drivers/md/md.c b/drivers/md/md.c index 9b82377a833..32fc19c540d 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -112,7 +112,7 @@ static inline int speed_max(struct mddev *mddev) static struct ctl_table_header *raid_table_header; -static ctl_table raid_table[] = { +static struct ctl_table raid_table[] = { { .procname = "speed_limit_min", .data = &sysctl_speed_limit_min, @@ -130,7 +130,7 @@ static ctl_table raid_table[] = { { } }; -static ctl_table raid_dir_table[] = { +static struct ctl_table raid_dir_table[] = { { .procname = "raid", .maxlen = 0, @@ -140,7 +140,7 @@ static ctl_table raid_dir_table[] = { { } }; -static ctl_table raid_root_table[] = { +static struct ctl_table raid_root_table[] = { { .procname = "dev", .maxlen = 0, @@ -183,46 +183,6 @@ struct bio *bio_clone_mddev(struct bio *bio, gfp_t gfp_mask, } EXPORT_SYMBOL_GPL(bio_clone_mddev); -void md_trim_bio(struct bio *bio, int offset, int size) -{ - /* 'bio' is a cloned bio which we need to trim to match - * the given offset and size. - * This requires adjusting bi_sector, bi_size, and bi_io_vec - */ - int i; - struct bio_vec *bvec; - int sofar = 0; - - size <<= 9; - if (offset == 0 && size == bio->bi_size) - return; - - clear_bit(BIO_SEG_VALID, &bio->bi_flags); - - bio_advance(bio, offset << 9); - - bio->bi_size = size; - - /* avoid any complications with bi_idx being non-zero*/ - if (bio->bi_idx) { - memmove(bio->bi_io_vec, bio->bi_io_vec+bio->bi_idx, - (bio->bi_vcnt - bio->bi_idx) * sizeof(struct bio_vec)); - bio->bi_vcnt -= bio->bi_idx; - bio->bi_idx = 0; - } - /* Make sure vcnt and last bv are not too big */ - bio_for_each_segment(bvec, bio, i) { - if (sofar + bvec->bv_len > size) - bvec->bv_len = size - sofar; - if (bvec->bv_len == 0) { - bio->bi_vcnt = i; - break; - } - sofar += bvec->bv_len; - } -} -EXPORT_SYMBOL_GPL(md_trim_bio); - /* * We have a system wide 'event count' that is incremented * on any 'interesting' event, and readers of /proc/mdstat @@ -433,7 +393,7 @@ static void md_submit_flush_data(struct work_struct *ws) struct mddev *mddev = container_of(ws, struct mddev, flush_work); struct bio *bio = mddev->flush_bio; - if (bio->bi_size == 0) + if (bio->bi_iter.bi_size == 0) /* an empty barrier - all done */ bio_endio(bio, 0); else { @@ -521,6 +481,7 @@ void mddev_init(struct mddev *mddev) init_waitqueue_head(&mddev->recovery_wait); mddev->reshape_position = MaxSector; mddev->reshape_backwards = 0; + mddev->last_sync_action = "none"; mddev->resync_min = 0; mddev->resync_max = MaxSector; mddev->level = LEVEL_NONE; @@ -601,11 +562,19 @@ static struct mddev * mddev_find(dev_t unit) goto retry; } -static inline int mddev_lock(struct mddev * mddev) +static inline int __must_check mddev_lock(struct mddev * mddev) { return mutex_lock_interruptible(&mddev->reconfig_mutex); } +/* Sometimes we need to take the lock in a situation where + * failure due to interrupts is not acceptable. + */ +static inline void mddev_lock_nointr(struct mddev * mddev) +{ + mutex_lock(&mddev->reconfig_mutex); +} + static inline int mddev_is_locked(struct mddev *mddev) { return mutex_is_locked(&mddev->reconfig_mutex); @@ -785,7 +754,7 @@ void md_super_write(struct mddev *mddev, struct md_rdev *rdev, struct bio *bio = bio_alloc_mddev(GFP_NOIO, 1, mddev); bio->bi_bdev = rdev->meta_bdev ? rdev->meta_bdev : rdev->bdev; - bio->bi_sector = sector; + bio->bi_iter.bi_sector = sector; bio_add_page(bio, page, size, 0); bio->bi_private = rdev; bio->bi_end_io = super_written; @@ -807,36 +776,24 @@ void md_super_wait(struct mddev *mddev) finish_wait(&mddev->sb_wait, &wq); } -static void bi_complete(struct bio *bio, int error) -{ - complete((struct completion*)bio->bi_private); -} - int sync_page_io(struct md_rdev *rdev, sector_t sector, int size, struct page *page, int rw, bool metadata_op) { struct bio *bio = bio_alloc_mddev(GFP_NOIO, 1, rdev->mddev); - struct completion event; int ret; - rw |= REQ_SYNC; - bio->bi_bdev = (metadata_op && rdev->meta_bdev) ? rdev->meta_bdev : rdev->bdev; if (metadata_op) - bio->bi_sector = sector + rdev->sb_start; + bio->bi_iter.bi_sector = sector + rdev->sb_start; else if (rdev->mddev->reshape_position != MaxSector && (rdev->mddev->reshape_backwards == (sector >= rdev->mddev->reshape_position))) - bio->bi_sector = sector + rdev->new_data_offset; + bio->bi_iter.bi_sector = sector + rdev->new_data_offset; else - bio->bi_sector = sector + rdev->data_offset; + bio->bi_iter.bi_sector = sector + rdev->data_offset; bio_add_page(bio, page, size, 0); - init_completion(&event); - bio->bi_private = &event; - bio->bi_end_io = bi_complete; - submit_bio(rw, bio); - wait_for_completion(&event); + submit_bio_wait(rw, bio); ret = test_bit(BIO_UPTODATE, &bio->bi_flags); bio_put(bio); @@ -1118,6 +1075,7 @@ static int super_90_validate(struct mddev *mddev, struct md_rdev *rdev) rdev->raid_disk = -1; clear_bit(Faulty, &rdev->flags); clear_bit(In_sync, &rdev->flags); + clear_bit(Bitmap_sync, &rdev->flags); clear_bit(WriteMostly, &rdev->flags); if (mddev->raid_disks == 0) { @@ -1179,7 +1137,7 @@ static int super_90_validate(struct mddev *mddev, struct md_rdev *rdev) mddev->bitmap_info.offset = mddev->bitmap_info.default_offset; mddev->bitmap_info.space = - mddev->bitmap_info.space; + mddev->bitmap_info.default_space; } } else if (mddev->pers == NULL) { @@ -1196,6 +1154,8 @@ static int super_90_validate(struct mddev *mddev, struct md_rdev *rdev) */ if (ev1 < mddev->bitmap->events_cleared) return 0; + if (ev1 < mddev->events) + set_bit(Bitmap_sync, &rdev->flags); } else { if (ev1 < mddev->events) /* just a hot-add of a new device, leave raid_disk at -1 */ @@ -1211,6 +1171,7 @@ static int super_90_validate(struct mddev *mddev, struct md_rdev *rdev) desc->raid_disk < mddev->raid_disks */) { set_bit(In_sync, &rdev->flags); rdev->raid_disk = desc->raid_disk; + rdev->saved_raid_disk = desc->raid_disk; } else if (desc->state & (1<<MD_DISK_ACTIVE)) { /* active but not in sync implies recovery up to * reshape position. We don't know exactly where @@ -1604,6 +1565,7 @@ static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev) rdev->raid_disk = -1; clear_bit(Faulty, &rdev->flags); clear_bit(In_sync, &rdev->flags); + clear_bit(Bitmap_sync, &rdev->flags); clear_bit(WriteMostly, &rdev->flags); if (mddev->raid_disks == 0) { @@ -1686,6 +1648,8 @@ static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev) */ if (ev1 < mddev->bitmap->events_cleared) return 0; + if (ev1 < mddev->events) + set_bit(Bitmap_sync, &rdev->flags); } else { if (ev1 < mddev->events) /* just a hot-add of a new device, leave raid_disk at -1 */ @@ -1706,10 +1670,14 @@ static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev) set_bit(Faulty, &rdev->flags); break; default: + rdev->saved_raid_disk = role; if ((le32_to_cpu(sb->feature_map) & - MD_FEATURE_RECOVERY_OFFSET)) + MD_FEATURE_RECOVERY_OFFSET)) { rdev->recovery_offset = le64_to_cpu(sb->recovery_offset); - else + if (!(le32_to_cpu(sb->feature_map) & + MD_FEATURE_RECOVERY_BITMAP)) + rdev->saved_raid_disk = -1; + } else set_bit(In_sync, &rdev->flags); rdev->raid_disk = role; break; @@ -1771,6 +1739,9 @@ static void super_1_sync(struct mddev *mddev, struct md_rdev *rdev) cpu_to_le32(MD_FEATURE_RECOVERY_OFFSET); sb->recovery_offset = cpu_to_le64(rdev->recovery_offset); + if (rdev->saved_raid_disk >= 0 && mddev->bitmap) + sb->feature_map |= + cpu_to_le32(MD_FEATURE_RECOVERY_BITMAP); } if (test_bit(Replacement, &rdev->flags)) sb->feature_map |= @@ -2512,8 +2483,7 @@ repeat: if (rdev->sb_loaded != 1) continue; /* no noise on spare devices */ - if (!test_bit(Faulty, &rdev->flags) && - rdev->saved_raid_disk == -1) { + if (!test_bit(Faulty, &rdev->flags)) { md_super_write(mddev,rdev, rdev->sb_start, rdev->sb_size, rdev->sb_page); @@ -2529,11 +2499,9 @@ repeat: rdev->badblocks.size = 0; } - } else if (test_bit(Faulty, &rdev->flags)) + } else pr_debug("md: %s (skipping faulty)\n", bdevname(rdev->bdev, b)); - else - pr_debug("(skipping incremental s/r "); if (mddev->level == LEVEL_MULTIPATH) /* only need to write one superblock... */ @@ -2649,6 +2617,8 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len) * blocked - sets the Blocked flags * -blocked - clears the Blocked and possibly simulates an error * insync - sets Insync providing device isn't active + * -insync - clear Insync for a device with a slot assigned, + * so that it gets rebuilt based on bitmap * write_error - sets WriteErrorSeen * -write_error - clears WriteErrorSeen */ @@ -2697,6 +2667,11 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len) } else if (cmd_match(buf, "insync") && rdev->raid_disk == -1) { set_bit(In_sync, &rdev->flags); err = 0; + } else if (cmd_match(buf, "-insync") && rdev->raid_disk >= 0) { + clear_bit(In_sync, &rdev->flags); + rdev->saved_raid_disk = rdev->raid_disk; + rdev->raid_disk = -1; + err = 0; } else if (cmd_match(buf, "write_error")) { set_bit(WriteErrorSeen, &rdev->flags); err = 0; @@ -2829,6 +2804,7 @@ slot_store(struct md_rdev *rdev, const char *buf, size_t len) else rdev->saved_raid_disk = -1; clear_bit(In_sync, &rdev->flags); + clear_bit(Bitmap_sync, &rdev->flags); err = rdev->mddev->pers-> hot_add_disk(rdev->mddev, rdev); if (err) { @@ -2867,7 +2843,7 @@ static ssize_t offset_store(struct md_rdev *rdev, const char *buf, size_t len) { unsigned long long offset; - if (strict_strtoull(buf, 10, &offset) < 0) + if (kstrtoull(buf, 10, &offset) < 0) return -EINVAL; if (rdev->mddev->pers && rdev->raid_disk >= 0) return -EBUSY; @@ -2895,7 +2871,7 @@ static ssize_t new_offset_store(struct md_rdev *rdev, unsigned long long new_offset; struct mddev *mddev = rdev->mddev; - if (strict_strtoull(buf, 10, &new_offset) < 0) + if (kstrtoull(buf, 10, &new_offset) < 0) return -EINVAL; if (mddev->sync_thread) @@ -2961,7 +2937,7 @@ static int strict_blocks_to_sectors(const char *buf, sector_t *sectors) unsigned long long blocks; sector_t new; - if (strict_strtoull(buf, 10, &blocks) < 0) + if (kstrtoull(buf, 10, &blocks) < 0) return -EINVAL; if (blocks & 1ULL << (8 * sizeof(blocks) - 1)) @@ -3017,7 +2993,7 @@ rdev_size_store(struct md_rdev *rdev, const char *buf, size_t len) for_each_mddev(mddev, tmp) { struct md_rdev *rdev2; - mddev_lock(mddev); + mddev_lock_nointr(mddev); rdev_for_each(rdev2, mddev) if (rdev->bdev == rdev2->bdev && rdev != rdev2 && @@ -3033,7 +3009,7 @@ rdev_size_store(struct md_rdev *rdev, const char *buf, size_t len) break; } } - mddev_lock(my_mddev); + mddev_lock_nointr(my_mddev); if (overlap) { /* Someone else could have slipped in a size * change here, but doing so is just silly. @@ -3069,7 +3045,7 @@ static ssize_t recovery_start_store(struct md_rdev *rdev, const char *buf, size_ if (cmd_match(buf, "none")) recovery_start = MaxSector; - else if (strict_strtoull(buf, 10, &recovery_start)) + else if (kstrtoull(buf, 10, &recovery_start)) return -EINVAL; if (rdev->mddev->pers && @@ -3428,7 +3404,7 @@ safe_delay_store(struct mddev *mddev, const char *cbuf, size_t len) mddev->safemode_delay = (msec*HZ)/1000; if (mddev->safemode_delay == 0) mddev->safemode_delay = 1; - if (mddev->safemode_delay < old_delay) + if (mddev->safemode_delay < old_delay || old_delay == 0) md_safemode_timeout((unsigned long)mddev); } return len; @@ -3472,6 +3448,8 @@ level_store(struct mddev *mddev, const char *buf, size_t len) mddev->level = LEVEL_NONE; return rv; } + if (mddev->ro) + return -EROFS; /* request to change the personality. Need to ensure: * - array is not engaged in resync/recovery/reshape @@ -3497,7 +3475,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len) if (clevel[len-1] == '\n') len--; clevel[len] = 0; - if (strict_strtol(clevel, 10, &level)) + if (kstrtol(clevel, 10, &level)) level = LEVEL_NONE; if (request_module("md-%s", clevel) != 0) @@ -3554,7 +3532,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len) printk(KERN_WARNING "md: cannot register extra attributes for %s\n", mdname(mddev)); - mddev->sysfs_action = sysfs_get_dirent(mddev->kobj.sd, NULL, "sync_action"); + mddev->sysfs_action = sysfs_get_dirent(mddev->kobj.sd, "sync_action"); } if (mddev->pers->sync_request != NULL && pers->sync_request == NULL) { @@ -3619,9 +3597,12 @@ level_store(struct mddev *mddev, const char *buf, size_t len) mddev->in_sync = 1; del_timer_sync(&mddev->safemode_timer); } + blk_set_stacking_limits(&mddev->queue->limits); pers->run(mddev); set_bit(MD_CHANGE_DEVS, &mddev->flags); mddev_resume(mddev); + if (!mddev->thread) + md_update_sb(mddev, 1); sysfs_notify(&mddev->kobj, NULL, "level"); md_new_event(mddev); return rv; @@ -3655,6 +3636,8 @@ layout_store(struct mddev *mddev, const char *buf, size_t len) int err; if (mddev->pers->check_reshape == NULL) return -EBUSY; + if (mddev->ro) + return -EROFS; mddev->new_layout = n; err = mddev->pers->check_reshape(mddev); if (err) { @@ -3744,6 +3727,8 @@ chunk_size_store(struct mddev *mddev, const char *buf, size_t len) int err; if (mddev->pers->check_reshape == NULL) return -EBUSY; + if (mddev->ro) + return -EROFS; mddev->new_chunk_sectors = n >> 9; err = mddev->pers->check_reshape(mddev); if (err) { @@ -4272,6 +4257,17 @@ action_store(struct mddev *mddev, const char *page, size_t len) return len; } +static struct md_sysfs_entry md_scan_mode = +__ATTR(sync_action, S_IRUGO|S_IWUSR, action_show, action_store); + +static ssize_t +last_sync_action_show(struct mddev *mddev, char *page) +{ + return sprintf(page, "%s\n", mddev->last_sync_action); +} + +static struct md_sysfs_entry md_last_scan_mode = __ATTR_RO(last_sync_action); + static ssize_t mismatch_cnt_show(struct mddev *mddev, char *page) { @@ -4280,10 +4276,6 @@ mismatch_cnt_show(struct mddev *mddev, char *page) atomic64_read(&mddev->resync_mismatches)); } -static struct md_sysfs_entry md_scan_mode = -__ATTR(sync_action, S_IRUGO|S_IWUSR, action_show, action_store); - - static struct md_sysfs_entry md_mismatches = __ATTR_RO(mismatch_cnt); static ssize_t @@ -4356,7 +4348,7 @@ sync_force_parallel_store(struct mddev *mddev, const char *buf, size_t len) { long n; - if (strict_strtol(buf, 10, &n)) + if (kstrtol(buf, 10, &n)) return -EINVAL; if (n != 0 && n != 1) @@ -4424,7 +4416,7 @@ static ssize_t min_sync_store(struct mddev *mddev, const char *buf, size_t len) { unsigned long long min; - if (strict_strtoull(buf, 10, &min)) + if (kstrtoull(buf, 10, &min)) return -EINVAL; if (min > mddev->resync_max) return -EINVAL; @@ -4461,7 +4453,7 @@ max_sync_store(struct mddev *mddev, const char *buf, size_t len) mddev->resync_max = MaxSector; else { unsigned long long max; - if (strict_strtoull(buf, 10, &max)) + if (kstrtoull(buf, 10, &max)) return -EINVAL; if (max < mddev->resync_min) return -EINVAL; @@ -4686,6 +4678,7 @@ static struct attribute *md_default_attrs[] = { static struct attribute *md_redundancy_attrs[] = { &md_scan_mode.attr, + &md_last_scan_mode.attr, &md_mismatches.attr, &md_sync_min.attr, &md_sync_max.attr, @@ -5135,7 +5128,7 @@ int md_run(struct mddev *mddev) set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); - if (mddev->flags) + if (mddev->flags & MD_UPDATE_SB_FLAGS) md_update_sb(mddev, 0); md_new_event(mddev); @@ -5194,32 +5187,6 @@ static int restart_array(struct mddev *mddev) return 0; } -/* similar to deny_write_access, but accounts for our holding a reference - * to the file ourselves */ -static int deny_bitmap_write_access(struct file * file) -{ - struct inode *inode = file->f_mapping->host; - - spin_lock(&inode->i_lock); - if (atomic_read(&inode->i_writecount) > 1) { - spin_unlock(&inode->i_lock); - return -ETXTBSY; - } - atomic_set(&inode->i_writecount, -1); - spin_unlock(&inode->i_lock); - - return 0; -} - -void restore_bitmap_write_access(struct file *file) -{ - struct inode *inode = file->f_mapping->host; - - spin_lock(&inode->i_lock); - atomic_set(&inode->i_writecount, 1); - spin_unlock(&inode->i_lock); -} - static void md_clean(struct mddev *mddev) { mddev->array_sectors = 0; @@ -5280,7 +5247,7 @@ static void __md_stop_writes(struct mddev *mddev) md_super_wait(mddev); if (mddev->ro == 0 && - (!mddev->in_sync || mddev->flags)) { + (!mddev->in_sync || (mddev->flags & MD_UPDATE_SB_FLAGS))) { /* mark array as shutdown cleanly */ mddev->in_sync = 1; md_update_sb(mddev, 1); @@ -5289,7 +5256,7 @@ static void __md_stop_writes(struct mddev *mddev) void md_stop_writes(struct mddev *mddev) { - mddev_lock(mddev); + mddev_lock_nointr(mddev); __md_stop_writes(mddev); mddev_unlock(mddev); } @@ -5322,14 +5289,35 @@ EXPORT_SYMBOL_GPL(md_stop); static int md_set_readonly(struct mddev *mddev, struct block_device *bdev) { int err = 0; + int did_freeze = 0; + + if (!test_bit(MD_RECOVERY_FROZEN, &mddev->recovery)) { + did_freeze = 1; + set_bit(MD_RECOVERY_FROZEN, &mddev->recovery); + md_wakeup_thread(mddev->thread); + } + if (mddev->sync_thread) { + set_bit(MD_RECOVERY_INTR, &mddev->recovery); + /* Thread might be blocked waiting for metadata update + * which will now never happen */ + wake_up_process(mddev->sync_thread->tsk); + } + mddev_unlock(mddev); + wait_event(resync_wait, mddev->sync_thread == NULL); + mddev_lock_nointr(mddev); + mutex_lock(&mddev->open_mutex); - if (atomic_read(&mddev->openers) > !!bdev) { + if (atomic_read(&mddev->openers) > !!bdev || + mddev->sync_thread || + (bdev && !test_bit(MD_STILL_CLOSED, &mddev->flags))) { printk("md: %s still in use.\n",mdname(mddev)); + if (did_freeze) { + clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery); + md_wakeup_thread(mddev->thread); + } err = -EBUSY; goto out; } - if (bdev) - sync_blockdev(bdev); if (mddev->pers) { __md_stop_writes(mddev); @@ -5340,7 +5328,7 @@ static int md_set_readonly(struct mddev *mddev, struct block_device *bdev) set_disk_ro(mddev->gendisk, 1); clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery); sysfs_notify_dirent_safe(mddev->sysfs_state); - err = 0; + err = 0; } out: mutex_unlock(&mddev->open_mutex); @@ -5356,22 +5344,36 @@ static int do_md_stop(struct mddev * mddev, int mode, { struct gendisk *disk = mddev->gendisk; struct md_rdev *rdev; + int did_freeze = 0; + + if (!test_bit(MD_RECOVERY_FROZEN, &mddev->recovery)) { + did_freeze = 1; + set_bit(MD_RECOVERY_FROZEN, &mddev->recovery); + md_wakeup_thread(mddev->thread); + } + if (mddev->sync_thread) { + set_bit(MD_RECOVERY_INTR, &mddev->recovery); + /* Thread might be blocked waiting for metadata update + * which will now never happen */ + wake_up_process(mddev->sync_thread->tsk); + } + mddev_unlock(mddev); + wait_event(resync_wait, mddev->sync_thread == NULL); + mddev_lock_nointr(mddev); mutex_lock(&mddev->open_mutex); if (atomic_read(&mddev->openers) > !!bdev || - mddev->sysfs_active) { + mddev->sysfs_active || + mddev->sync_thread || + (bdev && !test_bit(MD_STILL_CLOSED, &mddev->flags))) { printk("md: %s still in use.\n",mdname(mddev)); mutex_unlock(&mddev->open_mutex); + if (did_freeze) { + clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery); + md_wakeup_thread(mddev->thread); + } return -EBUSY; } - if (bdev) - /* It is possible IO was issued on some other - * open file which was closed before we took ->open_mutex. - * As that was not the last close __blkdev_put will not - * have called sync_blockdev, so we must. - */ - sync_blockdev(bdev); - if (mddev->pers) { if (mddev->ro) set_disk_ro(disk, 0); @@ -5405,7 +5407,6 @@ static int do_md_stop(struct mddev * mddev, int mode, bitmap_destroy(mddev); if (mddev->bitmap_info.file) { - restore_bitmap_write_access(mddev->bitmap_info.file); fput(mddev->bitmap_info.file); mddev->bitmap_info.file = NULL; } @@ -5598,7 +5599,7 @@ static int get_array_info(struct mddev * mddev, void __user * arg) if (mddev->in_sync) info.state = (1<<MD_SB_CLEAN); if (mddev->bitmap && mddev->bitmap_info.offset) - info.state = (1<<MD_SB_BITMAP_PRESENT); + info.state |= (1<<MD_SB_BITMAP_PRESENT); info.active_disks = insync; info.working_disks = working; info.failed_disks = failed; @@ -5619,10 +5620,7 @@ static int get_bitmap_file(struct mddev * mddev, void __user * arg) char *ptr, *buf = NULL; int err = -ENOMEM; - if (md_allow_write(mddev)) - file = kmalloc(sizeof(*file), GFP_NOIO); - else - file = kmalloc(sizeof(*file), GFP_KERNEL); + file = kmalloc(sizeof(*file), GFP_NOIO); if (!file) goto out; @@ -5760,8 +5758,10 @@ static int add_new_disk(struct mddev * mddev, mdu_disk_info_t *info) info->raid_disk < mddev->raid_disks) { rdev->raid_disk = info->raid_disk; set_bit(In_sync, &rdev->flags); + clear_bit(Bitmap_sync, &rdev->flags); } else rdev->raid_disk = -1; + rdev->saved_raid_disk = rdev->raid_disk; } else super_types[mddev->major_version]. validate_super(mddev, rdev); @@ -5774,11 +5774,6 @@ static int add_new_disk(struct mddev * mddev, mdu_disk_info_t *info) return -EINVAL; } - if (test_bit(In_sync, &rdev->flags)) - rdev->saved_raid_disk = rdev->raid_disk; - else - rdev->saved_raid_disk = -1; - clear_bit(In_sync, &rdev->flags); /* just to be sure */ if (info->state & (1<<MD_DISK_WRITEMOSTLY)) set_bit(WriteMostly, &rdev->flags); @@ -5963,7 +5958,7 @@ abort_export: static int set_bitmap_file(struct mddev *mddev, int fd) { - int err; + int err = 0; if (mddev->pers) { if (!mddev->pers->quiesce) @@ -5975,6 +5970,7 @@ static int set_bitmap_file(struct mddev *mddev, int fd) if (fd >= 0) { + struct inode *inode; if (mddev->bitmap) return -EEXIST; /* cannot add when bitmap is present */ mddev->bitmap_info.file = fget(fd); @@ -5985,10 +5981,21 @@ static int set_bitmap_file(struct mddev *mddev, int fd) return -EBADF; } - err = deny_bitmap_write_access(mddev->bitmap_info.file); - if (err) { + inode = mddev->bitmap_info.file->f_mapping->host; + if (!S_ISREG(inode->i_mode)) { + printk(KERN_ERR "%s: error: bitmap file must be a regular file\n", + mdname(mddev)); + err = -EBADF; + } else if (!(mddev->bitmap_info.file->f_mode & FMODE_WRITE)) { + printk(KERN_ERR "%s: error: bitmap file must open for write\n", + mdname(mddev)); + err = -EBADF; + } else if (atomic_read(&inode->i_writecount) != 1) { printk(KERN_ERR "%s: error: bitmap file is already in use\n", mdname(mddev)); + err = -EBUSY; + } + if (err) { fput(mddev->bitmap_info.file); mddev->bitmap_info.file = NULL; return err; @@ -6011,10 +6018,8 @@ static int set_bitmap_file(struct mddev *mddev, int fd) mddev->pers->quiesce(mddev, 0); } if (fd < 0) { - if (mddev->bitmap_info.file) { - restore_bitmap_write_access(mddev->bitmap_info.file); + if (mddev->bitmap_info.file) fput(mddev->bitmap_info.file); - } mddev->bitmap_info.file = NULL; } @@ -6136,6 +6141,8 @@ static int update_size(struct mddev *mddev, sector_t num_sectors) */ if (mddev->sync_thread) return -EBUSY; + if (mddev->ro) + return -EROFS; rdev_for_each(rdev, mddev) { sector_t avail = rdev->sectors; @@ -6158,6 +6165,8 @@ static int update_raid_disks(struct mddev *mddev, int raid_disks) /* change the number of raid disks */ if (mddev->pers->check_reshape == NULL) return -EINVAL; + if (mddev->ro) + return -EROFS; if (raid_disks <= 0 || (mddev->max_disks && raid_disks >= mddev->max_disks)) return -EINVAL; @@ -6328,6 +6337,32 @@ static int md_getgeo(struct block_device *bdev, struct hd_geometry *geo) return 0; } +static inline bool md_ioctl_valid(unsigned int cmd) +{ + switch (cmd) { + case ADD_NEW_DISK: + case BLKROSET: + case GET_ARRAY_INFO: + case GET_BITMAP_FILE: + case GET_DISK_INFO: + case HOT_ADD_DISK: + case HOT_REMOVE_DISK: + case PRINT_RAID_DEBUG: + case RAID_AUTORUN: + case RAID_VERSION: + case RESTART_ARRAY_RW: + case RUN_ARRAY: + case SET_ARRAY_INFO: + case SET_BITMAP_FILE: + case SET_DISK_FAULTY: + case STOP_ARRAY: + case STOP_ARRAY_RO: + return true; + default: + return false; + } +} + static int md_ioctl(struct block_device *bdev, fmode_t mode, unsigned int cmd, unsigned long arg) { @@ -6336,6 +6371,9 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode, struct mddev *mddev = NULL; int ro; + if (!md_ioctl_valid(cmd)) + return -ENOTTY; + switch (cmd) { case RAID_VERSION: case GET_ARRAY_INFO: @@ -6405,6 +6443,26 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode, /* need to ensure md_delayed_delete() has completed */ flush_workqueue(md_misc_wq); + if (cmd == HOT_REMOVE_DISK) + /* need to ensure recovery thread has run */ + wait_event_interruptible_timeout(mddev->sb_wait, + !test_bit(MD_RECOVERY_NEEDED, + &mddev->flags), + msecs_to_jiffies(5000)); + if (cmd == STOP_ARRAY || cmd == STOP_ARRAY_RO) { + /* Need to flush page cache, and ensure no-one else opens + * and writes + */ + mutex_lock(&mddev->open_mutex); + if (atomic_read(&mddev->openers) > 1) { + mutex_unlock(&mddev->open_mutex); + err = -EBUSY; + goto abort; + } + set_bit(MD_STILL_CLOSED, &mddev->flags); + mutex_unlock(&mddev->open_mutex); + sync_blockdev(bdev); + } err = mddev_lock(mddev); if (err) { printk(KERN_INFO @@ -6559,7 +6617,7 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode, wait_event(mddev->sb_wait, !test_bit(MD_CHANGE_DEVS, &mddev->flags) && !test_bit(MD_CHANGE_PENDING, &mddev->flags)); - mddev_lock(mddev); + mddev_lock_nointr(mddev); } } else { err = -EROFS; @@ -6658,6 +6716,7 @@ static int md_open(struct block_device *bdev, fmode_t mode) err = 0; atomic_inc(&mddev->openers); + clear_bit(MD_STILL_CLOSED, &mddev->flags); mutex_unlock(&mddev->open_mutex); check_disk_change(bdev); @@ -7116,11 +7175,14 @@ static int md_seq_open(struct inode *inode, struct file *file) return error; } +static int md_unloading; static unsigned int mdstat_poll(struct file *filp, poll_table *wait) { struct seq_file *seq = filp->private_data; int mask; + if (md_unloading) + return POLLIN|POLLRDNORM|POLLERR|POLLPRI;; poll_wait(filp, &md_event_waiters, wait); /* always allow read */ @@ -7323,27 +7385,33 @@ void md_do_sync(struct md_thread *thread) sector_t last_check; int skipped = 0; struct md_rdev *rdev; - char *desc; + char *desc, *action = NULL; struct blk_plug plug; /* just incase thread restarts... */ if (test_bit(MD_RECOVERY_DONE, &mddev->recovery)) return; - if (mddev->ro) /* never try to sync a read-only array */ + if (mddev->ro) {/* never try to sync a read-only array */ + set_bit(MD_RECOVERY_INTR, &mddev->recovery); return; + } if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery)) { - if (test_bit(MD_RECOVERY_CHECK, &mddev->recovery)) + if (test_bit(MD_RECOVERY_CHECK, &mddev->recovery)) { desc = "data-check"; - else if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery)) + action = "check"; + } else if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery)) { desc = "requested-resync"; - else + action = "repair"; + } else desc = "resync"; } else if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery)) desc = "reshape"; else desc = "recovery"; + mddev->last_sync_action = action ?: desc; + /* we overload curr_resync somewhat here. * 0 == not engaged in resync at all * 2 == checking that there is no conflict with another sync @@ -7364,9 +7432,6 @@ void md_do_sync(struct md_thread *thread) mddev->curr_resync = 2; try_again: - if (kthread_should_stop()) - set_bit(MD_RECOVERY_INTR, &mddev->recovery); - if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) goto skip; for_each_mddev(mddev2, tmp) { @@ -7391,7 +7456,7 @@ void md_do_sync(struct md_thread *thread) * be caught by 'softlockup' */ prepare_to_wait(&resync_wait, &wq, TASK_INTERRUPTIBLE); - if (!kthread_should_stop() && + if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery) && mddev2->curr_resync >= mddev->curr_resync) { printk(KERN_INFO "md: delaying %s of %s" " until %s has finished (they" @@ -7436,6 +7501,19 @@ void md_do_sync(struct md_thread *thread) rdev->recovery_offset < j) j = rdev->recovery_offset; rcu_read_unlock(); + + /* If there is a bitmap, we need to make sure all + * writes that started before we added a spare + * complete before we start doing a recovery. + * Otherwise the write might complete and (via + * bitmap_endwrite) set a bit in the bitmap after the + * recovery has checked that bit and skipped that + * region. + */ + if (mddev->bitmap) { + mddev->pers->quiesce(mddev, 1); + mddev->pers->quiesce(mddev, 0); + } } printk(KERN_INFO "md: %s of RAID array %s\n", desc, mdname(mddev)); @@ -7467,7 +7545,7 @@ void md_do_sync(struct md_thread *thread) last_check = 0; if (j>2) { - printk(KERN_INFO + printk(KERN_INFO "md: resuming %s of %s from checkpoint.\n", desc, mdname(mddev)); mddev->curr_resync = j; @@ -7504,7 +7582,8 @@ void md_do_sync(struct md_thread *thread) sysfs_notify(&mddev->kobj, NULL, "sync_completed"); } - while (j >= mddev->resync_max && !kthread_should_stop()) { + while (j >= mddev->resync_max && + !test_bit(MD_RECOVERY_INTR, &mddev->recovery)) { /* As this condition is controlled by user-space, * we can block indefinitely, so use '_interruptible' * to avoid triggering warnings. @@ -7512,17 +7591,18 @@ void md_do_sync(struct md_thread *thread) flush_signals(current); /* just in case */ wait_event_interruptible(mddev->recovery_wait, mddev->resync_max > j - || kthread_should_stop()); + || test_bit(MD_RECOVERY_INTR, + &mddev->recovery)); } - if (kthread_should_stop()) - goto interrupted; + if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) + break; sectors = mddev->pers->sync_request(mddev, j, &skipped, currspeed < speed_min(mddev)); if (sectors == 0) { set_bit(MD_RECOVERY_INTR, &mddev->recovery); - goto out; + break; } if (!skipped) { /* actual IO requested */ @@ -7559,10 +7639,8 @@ void md_do_sync(struct md_thread *thread) last_mark = next; } - - if (kthread_should_stop()) - goto interrupted; - + if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) + break; /* * this loop exits only if either when we are slower than @@ -7585,11 +7663,12 @@ void md_do_sync(struct md_thread *thread) } } } - printk(KERN_INFO "md: %s: %s done.\n",mdname(mddev), desc); + printk(KERN_INFO "md: %s: %s %s.\n",mdname(mddev), desc, + test_bit(MD_RECOVERY_INTR, &mddev->recovery) + ? "interrupted" : "done"); /* * this also signals 'finished resyncing' to md_stop */ - out: blk_finish_plug(&plug); wait_event(mddev->recovery_wait, !atomic_read(&mddev->recovery_active)); @@ -7643,16 +7722,6 @@ void md_do_sync(struct md_thread *thread) set_bit(MD_RECOVERY_DONE, &mddev->recovery); md_wakeup_thread(mddev->thread); return; - - interrupted: - /* - * got a signal, exit. - */ - printk(KERN_INFO - "md: md_do_sync() got signal ... exiting\n"); - set_bit(MD_RECOVERY_INTR, &mddev->recovery); - goto out; - } EXPORT_SYMBOL_GPL(md_do_sync); @@ -7693,24 +7762,12 @@ static int remove_and_add_spares(struct mddev *mddev, if (test_bit(Faulty, &rdev->flags)) continue; if (mddev->ro && - rdev->saved_raid_disk < 0) + ! (rdev->saved_raid_disk >= 0 && + !test_bit(Bitmap_sync, &rdev->flags))) continue; - rdev->recovery_offset = 0; - if (rdev->saved_raid_disk >= 0 && mddev->in_sync) { - spin_lock_irq(&mddev->write_lock); - if (mddev->in_sync) - /* OK, this device, which is in_sync, - * will definitely be noticed before - * the next write, so recovery isn't - * needed. - */ - rdev->recovery_offset = mddev->recovery_cp; - spin_unlock_irq(&mddev->write_lock); - } - if (mddev->ro && rdev->recovery_offset != MaxSector) - /* not safe to add this disk now */ - continue; + if (rdev->saved_raid_disk < 0) + rdev->recovery_offset = 0; if (mddev->pers-> hot_add_disk(mddev, rdev) == 0) { if (sysfs_link_rdev(mddev, rdev)) @@ -7768,7 +7825,7 @@ void md_check_recovery(struct mddev *mddev) if (mddev->ro && !test_bit(MD_RECOVERY_NEEDED, &mddev->recovery)) return; if ( ! ( - (mddev->flags & ~ (1<<MD_CHANGE_PENDING)) || + (mddev->flags & MD_UPDATE_SB_FLAGS & ~ (1<<MD_CHANGE_PENDING)) || test_bit(MD_RECOVERY_NEEDED, &mddev->recovery) || test_bit(MD_RECOVERY_DONE, &mddev->recovery) || (mddev->external == 0 && mddev->safemode == 1) || @@ -7788,9 +7845,13 @@ void md_check_recovery(struct mddev *mddev) * As we only add devices that are already in-sync, * we can activate the spares immediately. */ - clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery); remove_and_add_spares(mddev, NULL); - mddev->pers->spare_active(mddev); + /* There is no thread, but we need to call + * ->spare_active and clear saved_raid_disk + */ + set_bit(MD_RECOVERY_INTR, &mddev->recovery); + md_reap_sync_thread(mddev); + clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery); goto unlock; } @@ -7812,7 +7873,7 @@ void md_check_recovery(struct mddev *mddev) sysfs_notify_dirent_safe(mddev->sysfs_state); } - if (mddev->flags) + if (mddev->flags & MD_UPDATE_SB_FLAGS) md_update_sb(mddev, 0); if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery) && @@ -7892,6 +7953,8 @@ void md_check_recovery(struct mddev *mddev) md_new_event(mddev); } unlock: + wake_up(&mddev->sb_wait); + if (!mddev->sync_thread) { clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery); if (test_and_clear_bit(MD_RECOVERY_RECOVER, @@ -7909,6 +7972,7 @@ void md_reap_sync_thread(struct mddev *mddev) /* resync has finished, collect result */ md_unregister_thread(&mddev->sync_thread); + wake_up(&resync_wait); if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery) && !test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery)) { /* success...*/ @@ -7924,14 +7988,10 @@ void md_reap_sync_thread(struct mddev *mddev) mddev->pers->finish_reshape(mddev); /* If array is no-longer degraded, then any saved_raid_disk - * information must be scrapped. Also if any device is now - * In_sync we must scrape the saved_raid_disk for that device - * do the superblock for an incrementally recovered device - * written out. + * information must be scrapped. */ - rdev_for_each(rdev, mddev) - if (!mddev->degraded || - test_bit(In_sync, &rdev->flags)) + if (!mddev->degraded) + rdev_for_each(rdev, mddev) rdev->saved_raid_disk = -1; md_update_sb(mddev, 1); @@ -8086,6 +8146,7 @@ static int md_set_badblocks(struct badblocks *bb, sector_t s, int sectors, u64 *p; int lo, hi; int rv = 1; + unsigned long flags; if (bb->shift < 0) /* badblocks are disabled */ @@ -8100,7 +8161,7 @@ static int md_set_badblocks(struct badblocks *bb, sector_t s, int sectors, sectors = next - s; } - write_seqlock_irq(&bb->lock); + write_seqlock_irqsave(&bb->lock, flags); p = bb->page; lo = 0; @@ -8216,7 +8277,7 @@ static int md_set_badblocks(struct badblocks *bb, sector_t s, int sectors, bb->changed = 1; if (!acknowledged) bb->unacked_exist = 1; - write_sequnlock_irq(&bb->lock); + write_sequnlock_irqrestore(&bb->lock, flags); return rv; } @@ -8295,7 +8356,7 @@ static int md_clear_badblocks(struct badblocks *bb, sector_t s, int sectors) if (a < s) { /* we need to split this range */ if (bb->count >= MD_MAX_BADBLOCKS) { - rv = 0; + rv = -ENOSPC; goto out; } memmove(p+lo+1, p+lo, (bb->count - lo) * 8); @@ -8481,7 +8542,8 @@ static int md_notify_reboot(struct notifier_block *this, if (mddev_trylock(mddev)) { if (mddev->pers) __md_stop_writes(mddev); - mddev->safemode = 2; + if (mddev->persistent) + mddev->safemode = 2; mddev_unlock(mddev); } need_delay = 1; @@ -8623,6 +8685,7 @@ static __exit void md_exit(void) { struct mddev *mddev; struct list_head *tmp; + int delay = 1; blk_unregister_region(MKDEV(MD_MAJOR,0), 1U << MINORBITS); blk_unregister_region(MKDEV(mdp_major,0), 1U << MINORBITS); @@ -8631,7 +8694,19 @@ static __exit void md_exit(void) unregister_blkdev(mdp_major, "mdp"); unregister_reboot_notifier(&md_notifier); unregister_sysctl_table(raid_table_header); + + /* We cannot unload the modules while some process is + * waiting for us in select() or poll() - wake them up + */ + md_unloading = 1; + while (waitqueue_active(&md_event_waiters)) { + /* not safe to leave yet */ + wake_up(&md_event_waiters); + msleep(delay); + delay += delay; + } remove_proc_entry("mdstat", NULL); + for_each_mddev(mddev, tmp) { export_array(mddev); mddev->hold_active = 0; diff --git a/drivers/md/md.h b/drivers/md/md.h index 653f992b687..a49d991f3fe 100644 --- a/drivers/md/md.h +++ b/drivers/md/md.h @@ -106,7 +106,7 @@ struct md_rdev { */ struct work_struct del_work; /* used for delayed sysfs removal */ - struct sysfs_dirent *sysfs_state; /* handle for 'state' + struct kernfs_node *sysfs_state; /* handle for 'state' * sysfs entry */ struct badblocks { @@ -129,6 +129,9 @@ struct md_rdev { enum flag_bits { Faulty, /* device is known to have a fault */ In_sync, /* device is in_sync with rest of array */ + Bitmap_sync, /* ..actually, not quite In_sync. Need a + * bitmap-based recovery to get fully in sync + */ Unmerged, /* device is being added to array and should * be considerred for bvec_merge_fn but not * yet for actual IO @@ -204,12 +207,16 @@ struct mddev { struct md_personality *pers; dev_t unit; int md_minor; - struct list_head disks; + struct list_head disks; unsigned long flags; #define MD_CHANGE_DEVS 0 /* Some device status has changed */ #define MD_CHANGE_CLEAN 1 /* transition to or from 'clean' */ #define MD_CHANGE_PENDING 2 /* switch from 'clean' to 'active' in progress */ +#define MD_UPDATE_SB_FLAGS (1 | 2 | 4) /* If these are set, md_update_sb needed */ #define MD_ARRAY_FIRST_USE 3 /* First use of array, needs initialization */ +#define MD_STILL_CLOSED 4 /* If set, then array has not been opened since + * md_ioctl checked on it. + */ int suspended; atomic_t active_io; @@ -218,7 +225,7 @@ struct mddev { * are happening, so run/ * takeover/stop are not safe */ - int ready; /* See when safe to pass + int ready; /* See when safe to pass * IO requests down */ struct gendisk *gendisk; @@ -268,6 +275,14 @@ struct mddev { struct md_thread *thread; /* management thread */ struct md_thread *sync_thread; /* doing resync or reconstruct */ + + /* 'last_sync_action' is initialized to "none". It is set when a + * sync operation (i.e "data-check", "requested-resync", "resync", + * "recovery", or "reshape") is started. It holds this value even + * when the sync thread is "frozen" (interrupted) or "idle" (stopped + * or finished). It is overwritten when a new sync operation is begun. + */ + char *last_sync_action; sector_t curr_resync; /* last block scheduled */ /* As resync requests can complete out of order, we cannot easily track * how much resync has been completed. So we occasionally pause until @@ -364,10 +379,10 @@ struct mddev { sector_t resync_max; /* resync should pause * when it gets here */ - struct sysfs_dirent *sysfs_state; /* handle for 'array_state' + struct kernfs_node *sysfs_state; /* handle for 'array_state' * file in sysfs. */ - struct sysfs_dirent *sysfs_action; /* handle for 'sync_action' */ + struct kernfs_node *sysfs_action; /* handle for 'sync_action' */ struct work_struct del_work; /* used for delayed sysfs removal */ @@ -486,13 +501,13 @@ struct md_sysfs_entry { }; extern struct attribute_group md_bitmap_group; -static inline struct sysfs_dirent *sysfs_get_dirent_safe(struct sysfs_dirent *sd, char *name) +static inline struct kernfs_node *sysfs_get_dirent_safe(struct kernfs_node *sd, char *name) { if (sd) - return sysfs_get_dirent(sd, NULL, name); + return sysfs_get_dirent(sd, name); return sd; } -static inline void sysfs_notify_dirent_safe(struct sysfs_dirent *sd) +static inline void sysfs_notify_dirent_safe(struct kernfs_node *sd) { if (sd) sysfs_notify_dirent(sd); @@ -590,7 +605,6 @@ extern int md_check_no_bitmap(struct mddev *mddev); extern int md_integrity_register(struct mddev *mddev); extern void md_integrity_add_rdev(struct md_rdev *rdev, struct mddev *mddev); extern int strict_strtoul_scaled(const char *cp, unsigned long *res, int scale); -extern void restore_bitmap_write_access(struct file *file); extern void mddev_init(struct mddev *mddev); extern int md_run(struct mddev *mddev); @@ -605,7 +619,6 @@ extern struct bio *bio_clone_mddev(struct bio *bio, gfp_t gfp_mask, struct mddev *mddev); extern struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs, struct mddev *mddev); -extern void md_trim_bio(struct bio *bio, int offset, int size); extern void md_unplug(struct blk_plug_cb *cb, bool from_schedule); static inline int mddev_check_plugged(struct mddev *mddev) diff --git a/drivers/md/multipath.c b/drivers/md/multipath.c index 1642eae75a3..849ad39f547 100644 --- a/drivers/md/multipath.c +++ b/drivers/md/multipath.c @@ -100,7 +100,7 @@ static void multipath_end_request(struct bio *bio, int error) md_error (mp_bh->mddev, rdev); printk(KERN_ERR "multipath: %s: rescheduling sector %llu\n", bdevname(rdev->bdev,b), - (unsigned long long)bio->bi_sector); + (unsigned long long)bio->bi_iter.bi_sector); multipath_reschedule_retry(mp_bh); } else multipath_end_bh_io(mp_bh, error); @@ -132,7 +132,7 @@ static void multipath_make_request(struct mddev *mddev, struct bio * bio) multipath = conf->multipaths + mp_bh->path; mp_bh->bio = *bio; - mp_bh->bio.bi_sector += multipath->rdev->data_offset; + mp_bh->bio.bi_iter.bi_sector += multipath->rdev->data_offset; mp_bh->bio.bi_bdev = multipath->rdev->bdev; mp_bh->bio.bi_rw |= REQ_FAILFAST_TRANSPORT; mp_bh->bio.bi_end_io = multipath_end_request; @@ -355,21 +355,22 @@ static void multipathd(struct md_thread *thread) spin_unlock_irqrestore(&conf->device_lock, flags); bio = &mp_bh->bio; - bio->bi_sector = mp_bh->master_bio->bi_sector; + bio->bi_iter.bi_sector = mp_bh->master_bio->bi_iter.bi_sector; if ((mp_bh->path = multipath_map (conf))<0) { printk(KERN_ALERT "multipath: %s: unrecoverable IO read" " error for block %llu\n", bdevname(bio->bi_bdev,b), - (unsigned long long)bio->bi_sector); + (unsigned long long)bio->bi_iter.bi_sector); multipath_end_bh_io(mp_bh, -EIO); } else { printk(KERN_ERR "multipath: %s: redirecting sector %llu" " to another IO path\n", bdevname(bio->bi_bdev,b), - (unsigned long long)bio->bi_sector); + (unsigned long long)bio->bi_iter.bi_sector); *bio = *(mp_bh->master_bio); - bio->bi_sector += conf->multipaths[mp_bh->path].rdev->data_offset; + bio->bi_iter.bi_sector += + conf->multipaths[mp_bh->path].rdev->data_offset; bio->bi_bdev = conf->multipaths[mp_bh->path].rdev->bdev; bio->bi_rw |= REQ_FAILFAST_TRANSPORT; bio->bi_end_io = multipath_end_request; diff --git a/drivers/md/persistent-data/Kconfig b/drivers/md/persistent-data/Kconfig index 19b26879541..0c2dec7aec2 100644 --- a/drivers/md/persistent-data/Kconfig +++ b/drivers/md/persistent-data/Kconfig @@ -6,3 +6,13 @@ config DM_PERSISTENT_DATA ---help--- Library providing immutable on-disk data structure support for device-mapper targets such as the thin provisioning target. + +config DM_DEBUG_BLOCK_STACK_TRACING + boolean "Keep stack trace of persistent data block lock holders" + depends on STACKTRACE_SUPPORT && DM_PERSISTENT_DATA + select STACKTRACE + ---help--- + Enable this for messages that may help debug problems with the + block manager locking used by thin provisioning and caching. + + If unsure, say N. diff --git a/drivers/md/persistent-data/dm-array.c b/drivers/md/persistent-data/dm-array.c index 172147eb1d4..1d75b1dc1e2 100644 --- a/drivers/md/persistent-data/dm-array.c +++ b/drivers/md/persistent-data/dm-array.c @@ -317,8 +317,16 @@ static int shadow_ablock(struct dm_array_info *info, dm_block_t *root, * The shadow op will often be a noop. Only insert if it really * copied data. */ - if (dm_block_location(*block) != b) + if (dm_block_location(*block) != b) { + /* + * dm_tm_shadow_block will have already decremented the old + * block, but it is still referenced by the btree. We + * increment to stop the insert decrementing it below zero + * when overwriting the old value. + */ + dm_tm_inc(info->btree_info.tm, b); r = insert_ablock(info, index, *block, root); + } return r; } @@ -509,15 +517,18 @@ static int grow_add_tail_block(struct resize *resize) static int grow_needs_more_blocks(struct resize *resize) { int r; + unsigned old_nr_blocks = resize->old_nr_full_blocks; if (resize->old_nr_entries_in_last_block > 0) { + old_nr_blocks++; + r = grow_extend_tail_block(resize, resize->max_entries); if (r) return r; } r = insert_full_ablocks(resize->info, resize->size_of_block, - resize->old_nr_full_blocks, + old_nr_blocks, resize->new_nr_full_blocks, resize->max_entries, resize->value, &resize->root); diff --git a/drivers/md/persistent-data/dm-bitset.c b/drivers/md/persistent-data/dm-bitset.c index cd9a86d4cdf..36f7cc2c710 100644 --- a/drivers/md/persistent-data/dm-bitset.c +++ b/drivers/md/persistent-data/dm-bitset.c @@ -65,7 +65,7 @@ int dm_bitset_flush(struct dm_disk_bitset *info, dm_block_t root, int r; __le64 value; - if (!info->current_index_set) + if (!info->current_index_set || !info->dirty) return 0; value = cpu_to_le64(info->current_bits); @@ -77,6 +77,8 @@ int dm_bitset_flush(struct dm_disk_bitset *info, dm_block_t root, return r; info->current_index_set = false; + info->dirty = false; + return 0; } EXPORT_SYMBOL_GPL(dm_bitset_flush); @@ -94,6 +96,8 @@ static int read_bits(struct dm_disk_bitset *info, dm_block_t root, info->current_bits = le64_to_cpu(value); info->current_index_set = true; info->current_index = array_index; + info->dirty = false; + return 0; } @@ -126,6 +130,8 @@ int dm_bitset_set_bit(struct dm_disk_bitset *info, dm_block_t root, return r; set_bit(b, (unsigned long *) &info->current_bits); + info->dirty = true; + return 0; } EXPORT_SYMBOL_GPL(dm_bitset_set_bit); @@ -141,6 +147,8 @@ int dm_bitset_clear_bit(struct dm_disk_bitset *info, dm_block_t root, return r; clear_bit(b, (unsigned long *) &info->current_bits); + info->dirty = true; + return 0; } EXPORT_SYMBOL_GPL(dm_bitset_clear_bit); diff --git a/drivers/md/persistent-data/dm-bitset.h b/drivers/md/persistent-data/dm-bitset.h index e1b9bea14aa..c2287d672ef 100644 --- a/drivers/md/persistent-data/dm-bitset.h +++ b/drivers/md/persistent-data/dm-bitset.h @@ -71,6 +71,7 @@ struct dm_disk_bitset { uint64_t current_bits; bool current_index_set:1; + bool dirty:1; }; /* diff --git a/drivers/md/persistent-data/dm-block-manager.c b/drivers/md/persistent-data/dm-block-manager.c index 81b513890e2..087411c95ff 100644 --- a/drivers/md/persistent-data/dm-block-manager.c +++ b/drivers/md/persistent-data/dm-block-manager.c @@ -104,7 +104,7 @@ static int __check_holder(struct block_lock *lock) for (i = 0; i < MAX_HOLDERS; i++) { if (lock->holders[i] == current) { - DMERR("recursive lock detected in pool metadata"); + DMERR("recursive lock detected in metadata"); #ifdef CONFIG_DM_DEBUG_BLOCK_STACK_TRACING DMERR("previously held here:"); print_stack_trace(lock->traces + i, 4); @@ -595,25 +595,19 @@ int dm_bm_unlock(struct dm_block *b) } EXPORT_SYMBOL_GPL(dm_bm_unlock); -int dm_bm_flush_and_unlock(struct dm_block_manager *bm, - struct dm_block *superblock) +int dm_bm_flush(struct dm_block_manager *bm) { - int r; - if (bm->read_only) return -EPERM; - r = dm_bufio_write_dirty_buffers(bm->bufio); - if (unlikely(r)) { - dm_bm_unlock(superblock); - return r; - } - - dm_bm_unlock(superblock); - return dm_bufio_write_dirty_buffers(bm->bufio); } -EXPORT_SYMBOL_GPL(dm_bm_flush_and_unlock); +EXPORT_SYMBOL_GPL(dm_bm_flush); + +void dm_bm_prefetch(struct dm_block_manager *bm, dm_block_t b) +{ + dm_bufio_prefetch(bm->bufio, b, 1); +} void dm_bm_set_read_only(struct dm_block_manager *bm) { @@ -621,6 +615,12 @@ void dm_bm_set_read_only(struct dm_block_manager *bm) } EXPORT_SYMBOL_GPL(dm_bm_set_read_only); +void dm_bm_set_read_write(struct dm_block_manager *bm) +{ + bm->read_only = false; +} +EXPORT_SYMBOL_GPL(dm_bm_set_read_write); + u32 dm_bm_checksum(const void *data, size_t len, u32 init_xor) { return crc32c(~(u32) 0, data, len) ^ init_xor; diff --git a/drivers/md/persistent-data/dm-block-manager.h b/drivers/md/persistent-data/dm-block-manager.h index be5bff61be2..1b95dfc1778 100644 --- a/drivers/md/persistent-data/dm-block-manager.h +++ b/drivers/md/persistent-data/dm-block-manager.h @@ -105,8 +105,12 @@ int dm_bm_unlock(struct dm_block *b); * * This method always blocks. */ -int dm_bm_flush_and_unlock(struct dm_block_manager *bm, - struct dm_block *superblock); +int dm_bm_flush(struct dm_block_manager *bm); + +/* + * Request data is prefetched into the cache. + */ +void dm_bm_prefetch(struct dm_block_manager *bm, dm_block_t b); /* * Switches the bm to a read only mode. Once read-only mode @@ -120,6 +124,7 @@ int dm_bm_flush_and_unlock(struct dm_block_manager *bm, * be returned if you do. */ void dm_bm_set_read_only(struct dm_block_manager *bm); +void dm_bm_set_read_write(struct dm_block_manager *bm); u32 dm_bm_checksum(const void *data, size_t len, u32 init_xor); diff --git a/drivers/md/persistent-data/dm-btree.c b/drivers/md/persistent-data/dm-btree.c index 35865425e4b..416060c2570 100644 --- a/drivers/md/persistent-data/dm-btree.c +++ b/drivers/md/persistent-data/dm-btree.c @@ -161,6 +161,7 @@ struct frame { }; struct del_stack { + struct dm_btree_info *info; struct dm_transaction_manager *tm; int top; struct frame spine[MAX_SPINE_DEPTH]; @@ -183,6 +184,20 @@ static int unprocessed_frames(struct del_stack *s) return s->top >= 0; } +static void prefetch_children(struct del_stack *s, struct frame *f) +{ + unsigned i; + struct dm_block_manager *bm = dm_tm_get_bm(s->tm); + + for (i = 0; i < f->nr_children; i++) + dm_bm_prefetch(bm, value64(f->n, i)); +} + +static bool is_internal_level(struct dm_btree_info *info, struct frame *f) +{ + return f->level < (info->levels - 1); +} + static int push_frame(struct del_stack *s, dm_block_t b, unsigned level) { int r; @@ -205,6 +220,7 @@ static int push_frame(struct del_stack *s, dm_block_t b, unsigned level) dm_tm_dec(s->tm, b); else { + uint32_t flags; struct frame *f = s->spine + ++s->top; r = dm_tm_read_lock(s->tm, b, &btree_node_validator, &f->b); @@ -217,6 +233,10 @@ static int push_frame(struct del_stack *s, dm_block_t b, unsigned level) f->level = level; f->nr_children = le32_to_cpu(f->n->header.nr_entries); f->current_child = 0; + + flags = le32_to_cpu(f->n->header.flags); + if (flags & INTERNAL_NODE || is_internal_level(s->info, f)) + prefetch_children(s, f); } return 0; @@ -230,11 +250,6 @@ static void pop_frame(struct del_stack *s) dm_tm_unlock(s->tm, f->b); } -static bool is_internal_level(struct dm_btree_info *info, struct frame *f) -{ - return f->level < (info->levels - 1); -} - int dm_btree_del(struct dm_btree_info *info, dm_block_t root) { int r; @@ -243,6 +258,7 @@ int dm_btree_del(struct dm_btree_info *info, dm_block_t root) s = kmalloc(sizeof(*s), GFP_KERNEL); if (!s) return -ENOMEM; + s->info = info; s->tm = info->tm; s->top = -1; @@ -287,7 +303,7 @@ int dm_btree_del(struct dm_btree_info *info, dm_block_t root) info->value_type.dec(info->value_type.context, value_ptr(f->n, i)); } - f->current_child = f->nr_children; + pop_frame(s); } } @@ -754,8 +770,8 @@ EXPORT_SYMBOL_GPL(dm_btree_insert_notify); /*----------------------------------------------------------------*/ -static int find_highest_key(struct ro_spine *s, dm_block_t block, - uint64_t *result_key, dm_block_t *next_block) +static int find_key(struct ro_spine *s, dm_block_t block, bool find_highest, + uint64_t *result_key, dm_block_t *next_block) { int i, r; uint32_t flags; @@ -772,7 +788,11 @@ static int find_highest_key(struct ro_spine *s, dm_block_t block, else i--; - *result_key = le64_to_cpu(ro_node(s)->keys[i]); + if (find_highest) + *result_key = le64_to_cpu(ro_node(s)->keys[i]); + else + *result_key = le64_to_cpu(ro_node(s)->keys[0]); + if (next_block || flags & INTERNAL_NODE) block = value64(ro_node(s), i); @@ -783,16 +803,16 @@ static int find_highest_key(struct ro_spine *s, dm_block_t block, return 0; } -int dm_btree_find_highest_key(struct dm_btree_info *info, dm_block_t root, - uint64_t *result_keys) +static int dm_btree_find_key(struct dm_btree_info *info, dm_block_t root, + bool find_highest, uint64_t *result_keys) { int r = 0, count = 0, level; struct ro_spine spine; init_ro_spine(&spine, info); for (level = 0; level < info->levels; level++) { - r = find_highest_key(&spine, root, result_keys + level, - level == info->levels - 1 ? NULL : &root); + r = find_key(&spine, root, find_highest, result_keys + level, + level == info->levels - 1 ? NULL : &root); if (r == -ENODATA) { r = 0; break; @@ -806,8 +826,23 @@ int dm_btree_find_highest_key(struct dm_btree_info *info, dm_block_t root, return r ? r : count; } + +int dm_btree_find_highest_key(struct dm_btree_info *info, dm_block_t root, + uint64_t *result_keys) +{ + return dm_btree_find_key(info, root, true, result_keys); +} EXPORT_SYMBOL_GPL(dm_btree_find_highest_key); +int dm_btree_find_lowest_key(struct dm_btree_info *info, dm_block_t root, + uint64_t *result_keys) +{ + return dm_btree_find_key(info, root, false, result_keys); +} +EXPORT_SYMBOL_GPL(dm_btree_find_lowest_key); + +/*----------------------------------------------------------------*/ + /* * FIXME: We shouldn't use a recursive algorithm when we have limited stack * space. Also this only works for single level trees. diff --git a/drivers/md/persistent-data/dm-btree.h b/drivers/md/persistent-data/dm-btree.h index 8672d159e0b..dacfc34180b 100644 --- a/drivers/md/persistent-data/dm-btree.h +++ b/drivers/md/persistent-data/dm-btree.h @@ -137,6 +137,14 @@ int dm_btree_remove(struct dm_btree_info *info, dm_block_t root, /* * Returns < 0 on failure. Otherwise the number of key entries that have * been filled out. Remember trees can have zero entries, and as such have + * no lowest key. + */ +int dm_btree_find_lowest_key(struct dm_btree_info *info, dm_block_t root, + uint64_t *result_keys); + +/* + * Returns < 0 on failure. Otherwise the number of key entries that have + * been filled out. Remember trees can have zero entries, and as such have * no highest key. */ int dm_btree_find_highest_key(struct dm_btree_info *info, dm_block_t root, diff --git a/drivers/md/persistent-data/dm-space-map-common.c b/drivers/md/persistent-data/dm-space-map-common.c index 3e7a88d99eb..aacbe70c2c2 100644 --- a/drivers/md/persistent-data/dm-space-map-common.c +++ b/drivers/md/persistent-data/dm-space-map-common.c @@ -245,6 +245,10 @@ int sm_ll_extend(struct ll_disk *ll, dm_block_t extra_blocks) return -EINVAL; } + /* + * We need to set this before the dm_tm_new_block() call below. + */ + ll->nr_blocks = nr_blocks; for (i = old_blocks; i < blocks; i++) { struct dm_block *b; struct disk_index_entry idx; @@ -252,6 +256,7 @@ int sm_ll_extend(struct ll_disk *ll, dm_block_t extra_blocks) r = dm_tm_new_block(ll->tm, &dm_sm_bitmap_validator, &b); if (r < 0) return r; + idx.blocknr = cpu_to_le64(dm_block_location(b)); r = dm_tm_unlock(ll->tm, b); @@ -266,7 +271,6 @@ int sm_ll_extend(struct ll_disk *ll, dm_block_t extra_blocks) return r; } - ll->nr_blocks = nr_blocks; return 0; } @@ -292,16 +296,11 @@ int sm_ll_lookup_bitmap(struct ll_disk *ll, dm_block_t b, uint32_t *result) return dm_tm_unlock(ll->tm, blk); } -int sm_ll_lookup(struct ll_disk *ll, dm_block_t b, uint32_t *result) +static int sm_ll_lookup_big_ref_count(struct ll_disk *ll, dm_block_t b, + uint32_t *result) { __le32 le_rc; - int r = sm_ll_lookup_bitmap(ll, b, result); - - if (r) - return r; - - if (*result != 3) - return r; + int r; r = dm_btree_lookup(&ll->ref_count_info, ll->ref_count_root, &b, &le_rc); if (r < 0) @@ -312,6 +311,19 @@ int sm_ll_lookup(struct ll_disk *ll, dm_block_t b, uint32_t *result) return r; } +int sm_ll_lookup(struct ll_disk *ll, dm_block_t b, uint32_t *result) +{ + int r = sm_ll_lookup_bitmap(ll, b, result); + + if (r) + return r; + + if (*result != 3) + return r; + + return sm_ll_lookup_big_ref_count(ll, b, result); +} + int sm_ll_find_free_block(struct ll_disk *ll, dm_block_t begin, dm_block_t end, dm_block_t *result) { @@ -372,11 +384,12 @@ int sm_ll_find_free_block(struct ll_disk *ll, dm_block_t begin, return -ENOSPC; } -int sm_ll_insert(struct ll_disk *ll, dm_block_t b, - uint32_t ref_count, enum allocation_event *ev) +static int sm_ll_mutate(struct ll_disk *ll, dm_block_t b, + int (*mutator)(void *context, uint32_t old, uint32_t *new), + void *context, enum allocation_event *ev) { int r; - uint32_t bit, old; + uint32_t bit, old, ref_count; struct dm_block *nb; dm_block_t index = b; struct disk_index_entry ie_disk; @@ -399,6 +412,20 @@ int sm_ll_insert(struct ll_disk *ll, dm_block_t b, bm_le = dm_bitmap_data(nb); old = sm_lookup_bitmap(bm_le, bit); + if (old > 2) { + r = sm_ll_lookup_big_ref_count(ll, b, &old); + if (r < 0) { + dm_tm_unlock(ll->tm, nb); + return r; + } + } + + r = mutator(context, old, &ref_count); + if (r) { + dm_tm_unlock(ll->tm, nb); + return r; + } + if (ref_count <= 2) { sm_set_bitmap(bm_le, bit, ref_count); @@ -448,31 +475,43 @@ int sm_ll_insert(struct ll_disk *ll, dm_block_t b, return ll->save_ie(ll, index, &ie_disk); } -int sm_ll_inc(struct ll_disk *ll, dm_block_t b, enum allocation_event *ev) +static int set_ref_count(void *context, uint32_t old, uint32_t *new) { - int r; - uint32_t rc; - - r = sm_ll_lookup(ll, b, &rc); - if (r) - return r; + *new = *((uint32_t *) context); + return 0; +} - return sm_ll_insert(ll, b, rc + 1, ev); +int sm_ll_insert(struct ll_disk *ll, dm_block_t b, + uint32_t ref_count, enum allocation_event *ev) +{ + return sm_ll_mutate(ll, b, set_ref_count, &ref_count, ev); } -int sm_ll_dec(struct ll_disk *ll, dm_block_t b, enum allocation_event *ev) +static int inc_ref_count(void *context, uint32_t old, uint32_t *new) { - int r; - uint32_t rc; + *new = old + 1; + return 0; +} - r = sm_ll_lookup(ll, b, &rc); - if (r) - return r; +int sm_ll_inc(struct ll_disk *ll, dm_block_t b, enum allocation_event *ev) +{ + return sm_ll_mutate(ll, b, inc_ref_count, NULL, ev); +} - if (!rc) +static int dec_ref_count(void *context, uint32_t old, uint32_t *new) +{ + if (!old) { + DMERR_LIMIT("unable to decrement a reference count below 0"); return -EINVAL; + } - return sm_ll_insert(ll, b, rc - 1, ev); + *new = old - 1; + return 0; +} + +int sm_ll_dec(struct ll_disk *ll, dm_block_t b, enum allocation_event *ev) +{ + return sm_ll_mutate(ll, b, dec_ref_count, NULL, ev); } int sm_ll_commit(struct ll_disk *ll) diff --git a/drivers/md/persistent-data/dm-space-map-disk.c b/drivers/md/persistent-data/dm-space-map-disk.c index e735a6d5a79..cfbf9617e46 100644 --- a/drivers/md/persistent-data/dm-space-map-disk.c +++ b/drivers/md/persistent-data/dm-space-map-disk.c @@ -140,26 +140,10 @@ static int sm_disk_inc_block(struct dm_space_map *sm, dm_block_t b) static int sm_disk_dec_block(struct dm_space_map *sm, dm_block_t b) { - int r; - uint32_t old_count; enum allocation_event ev; struct sm_disk *smd = container_of(sm, struct sm_disk, sm); - r = sm_ll_dec(&smd->ll, b, &ev); - if (!r && (ev == SM_FREE)) { - /* - * It's only free if it's also free in the last - * transaction. - */ - r = sm_ll_lookup(&smd->old_ll, b, &old_count); - if (r) - return r; - - if (!old_count) - smd->nr_allocated_this_transaction--; - } - - return r; + return sm_ll_dec(&smd->ll, b, &ev); } static int sm_disk_new_block(struct dm_space_map *sm, dm_block_t *b) diff --git a/drivers/md/persistent-data/dm-space-map-metadata.c b/drivers/md/persistent-data/dm-space-map-metadata.c index 1c959684cae..786b689bdfc 100644 --- a/drivers/md/persistent-data/dm-space-map-metadata.c +++ b/drivers/md/persistent-data/dm-space-map-metadata.c @@ -91,6 +91,69 @@ struct block_op { dm_block_t block; }; +struct bop_ring_buffer { + unsigned begin; + unsigned end; + struct block_op bops[MAX_RECURSIVE_ALLOCATIONS + 1]; +}; + +static void brb_init(struct bop_ring_buffer *brb) +{ + brb->begin = 0; + brb->end = 0; +} + +static bool brb_empty(struct bop_ring_buffer *brb) +{ + return brb->begin == brb->end; +} + +static unsigned brb_next(struct bop_ring_buffer *brb, unsigned old) +{ + unsigned r = old + 1; + return (r >= (sizeof(brb->bops) / sizeof(*brb->bops))) ? 0 : r; +} + +static int brb_push(struct bop_ring_buffer *brb, + enum block_op_type type, dm_block_t b) +{ + struct block_op *bop; + unsigned next = brb_next(brb, brb->end); + + /* + * We don't allow the last bop to be filled, this way we can + * differentiate between full and empty. + */ + if (next == brb->begin) + return -ENOMEM; + + bop = brb->bops + brb->end; + bop->type = type; + bop->block = b; + + brb->end = next; + + return 0; +} + +static int brb_pop(struct bop_ring_buffer *brb, struct block_op *result) +{ + struct block_op *bop; + + if (brb_empty(brb)) + return -ENODATA; + + bop = brb->bops + brb->begin; + result->type = bop->type; + result->block = bop->block; + + brb->begin = brb_next(brb, brb->begin); + + return 0; +} + +/*----------------------------------------------------------------*/ + struct sm_metadata { struct dm_space_map sm; @@ -101,25 +164,20 @@ struct sm_metadata { unsigned recursion_count; unsigned allocated_this_transaction; - unsigned nr_uncommitted; - struct block_op uncommitted[MAX_RECURSIVE_ALLOCATIONS]; + struct bop_ring_buffer uncommitted; struct threshold threshold; }; static int add_bop(struct sm_metadata *smm, enum block_op_type type, dm_block_t b) { - struct block_op *op; + int r = brb_push(&smm->uncommitted, type, b); - if (smm->nr_uncommitted == MAX_RECURSIVE_ALLOCATIONS) { + if (r) { DMERR("too many recursive allocations"); return -ENOMEM; } - op = smm->uncommitted + smm->nr_uncommitted++; - op->type = type; - op->block = b; - return 0; } @@ -158,11 +216,17 @@ static int out(struct sm_metadata *smm) return -ENOMEM; } - if (smm->recursion_count == 1 && smm->nr_uncommitted) { - while (smm->nr_uncommitted && !r) { - smm->nr_uncommitted--; - r = commit_bop(smm, smm->uncommitted + - smm->nr_uncommitted); + if (smm->recursion_count == 1) { + while (!brb_empty(&smm->uncommitted)) { + struct block_op bop; + + r = brb_pop(&smm->uncommitted, &bop); + if (r) { + DMERR("bug in bop ring buffer"); + break; + } + + r = commit_bop(smm, &bop); if (r) break; } @@ -217,7 +281,8 @@ static int sm_metadata_get_nr_free(struct dm_space_map *sm, dm_block_t *count) static int sm_metadata_get_count(struct dm_space_map *sm, dm_block_t b, uint32_t *result) { - int r, i; + int r; + unsigned i; struct sm_metadata *smm = container_of(sm, struct sm_metadata, sm); unsigned adjustment = 0; @@ -225,8 +290,10 @@ static int sm_metadata_get_count(struct dm_space_map *sm, dm_block_t b, * We may have some uncommitted adjustments to add. This list * should always be really short. */ - for (i = 0; i < smm->nr_uncommitted; i++) { - struct block_op *op = smm->uncommitted + i; + for (i = smm->uncommitted.begin; + i != smm->uncommitted.end; + i = brb_next(&smm->uncommitted, i)) { + struct block_op *op = smm->uncommitted.bops + i; if (op->block != b) continue; @@ -254,7 +321,8 @@ static int sm_metadata_get_count(struct dm_space_map *sm, dm_block_t b, static int sm_metadata_count_is_more_than_one(struct dm_space_map *sm, dm_block_t b, int *result) { - int r, i, adjustment = 0; + int r, adjustment = 0; + unsigned i; struct sm_metadata *smm = container_of(sm, struct sm_metadata, sm); uint32_t rc; @@ -262,8 +330,11 @@ static int sm_metadata_count_is_more_than_one(struct dm_space_map *sm, * We may have some uncommitted adjustments to add. This list * should always be really short. */ - for (i = 0; i < smm->nr_uncommitted; i++) { - struct block_op *op = smm->uncommitted + i; + for (i = smm->uncommitted.begin; + i != smm->uncommitted.end; + i = brb_next(&smm->uncommitted, i)) { + + struct block_op *op = smm->uncommitted.bops + i; if (op->block != b) continue; @@ -384,12 +455,16 @@ static int sm_metadata_new_block(struct dm_space_map *sm, dm_block_t *b) struct sm_metadata *smm = container_of(sm, struct sm_metadata, sm); int r = sm_metadata_new_block_(sm, b); - if (r) - DMERR("unable to allocate new metadata block"); + if (r) { + DMERR_LIMIT("unable to allocate new metadata block"); + return r; + } r = sm_metadata_get_nr_free(sm, &count); - if (r) - DMERR("couldn't get free block count"); + if (r) { + DMERR_LIMIT("couldn't get free block count"); + return r; + } check_threshold(&smm->threshold, count); @@ -604,20 +679,38 @@ static int sm_metadata_extend(struct dm_space_map *sm, dm_block_t extra_blocks) * Flick into a mode where all blocks get allocated in the new area. */ smm->begin = old_len; - memcpy(&smm->sm, &bootstrap_ops, sizeof(smm->sm)); + memcpy(sm, &bootstrap_ops, sizeof(*sm)); /* * Extend. */ r = sm_ll_extend(&smm->ll, extra_blocks); + if (r) + goto out; /* - * Switch back to normal behaviour. + * We repeatedly increment then commit until the commit doesn't + * allocate any new blocks. */ - memcpy(&smm->sm, &ops, sizeof(smm->sm)); - for (i = old_len; !r && i < smm->begin; i++) - r = sm_ll_inc(&smm->ll, i, &ev); + do { + for (i = old_len; !r && i < smm->begin; i++) { + r = sm_ll_inc(&smm->ll, i, &ev); + if (r) + goto out; + } + old_len = smm->begin; + + r = sm_ll_commit(&smm->ll); + if (r) + goto out; + + } while (old_len != smm->begin); +out: + /* + * Switch back to normal behaviour. + */ + memcpy(sm, &ops, sizeof(*sm)); return r; } @@ -649,7 +742,7 @@ int dm_sm_metadata_create(struct dm_space_map *sm, smm->begin = superblock + 1; smm->recursion_count = 0; smm->allocated_this_transaction = 0; - smm->nr_uncommitted = 0; + brb_init(&smm->uncommitted); threshold_init(&smm->threshold); memcpy(&smm->sm, &bootstrap_ops, sizeof(smm->sm)); @@ -658,6 +751,8 @@ int dm_sm_metadata_create(struct dm_space_map *sm, if (r) return r; + if (nr_blocks > DM_SM_METADATA_MAX_BLOCKS) + nr_blocks = DM_SM_METADATA_MAX_BLOCKS; r = sm_ll_extend(&smm->ll, nr_blocks); if (r) return r; @@ -691,7 +786,7 @@ int dm_sm_metadata_open(struct dm_space_map *sm, smm->begin = 0; smm->recursion_count = 0; smm->allocated_this_transaction = 0; - smm->nr_uncommitted = 0; + brb_init(&smm->uncommitted); threshold_init(&smm->threshold); memcpy(&smm->old_ll, &smm->ll, sizeof(smm->old_ll)); diff --git a/drivers/md/persistent-data/dm-space-map-metadata.h b/drivers/md/persistent-data/dm-space-map-metadata.h index 39bba0801cf..64df923974d 100644 --- a/drivers/md/persistent-data/dm-space-map-metadata.h +++ b/drivers/md/persistent-data/dm-space-map-metadata.h @@ -9,6 +9,17 @@ #include "dm-transaction-manager.h" +#define DM_SM_METADATA_BLOCK_SIZE (4096 >> SECTOR_SHIFT) + +/* + * The metadata device is currently limited in size. + * + * We have one block of index, which can hold 255 index entries. Each + * index entry contains allocation info about ~16k metadata blocks. + */ +#define DM_SM_METADATA_MAX_BLOCKS (255 * ((1 << 14) - 64)) +#define DM_SM_METADATA_MAX_SECTORS (DM_SM_METADATA_MAX_BLOCKS * DM_SM_METADATA_BLOCK_SIZE) + /* * Unfortunately we have to use two-phase construction due to the cycle * between the tm and sm. diff --git a/drivers/md/persistent-data/dm-transaction-manager.c b/drivers/md/persistent-data/dm-transaction-manager.c index 81da1a26042..3bc30a0ae3d 100644 --- a/drivers/md/persistent-data/dm-transaction-manager.c +++ b/drivers/md/persistent-data/dm-transaction-manager.c @@ -154,7 +154,7 @@ int dm_tm_pre_commit(struct dm_transaction_manager *tm) if (r < 0) return r; - return 0; + return dm_bm_flush(tm->bm); } EXPORT_SYMBOL_GPL(dm_tm_pre_commit); @@ -164,8 +164,9 @@ int dm_tm_commit(struct dm_transaction_manager *tm, struct dm_block *root) return -EWOULDBLOCK; wipe_shadow_table(tm); + dm_bm_unlock(root); - return dm_bm_flush_and_unlock(tm->bm, root); + return dm_bm_flush(tm->bm); } EXPORT_SYMBOL_GPL(dm_tm_commit); diff --git a/drivers/md/persistent-data/dm-transaction-manager.h b/drivers/md/persistent-data/dm-transaction-manager.h index b5b139076ca..2772ed2a781 100644 --- a/drivers/md/persistent-data/dm-transaction-manager.h +++ b/drivers/md/persistent-data/dm-transaction-manager.h @@ -38,18 +38,17 @@ struct dm_transaction_manager *dm_tm_create_non_blocking_clone(struct dm_transac /* * We use a 2-phase commit here. * - * i) In the first phase the block manager is told to start flushing, and - * the changes to the space map are written to disk. You should interrogate - * your particular space map to get detail of its root node etc. to be - * included in your superblock. + * i) Make all changes for the transaction *except* for the superblock. + * Then call dm_tm_pre_commit() to flush them to disk. * - * ii) @root will be committed last. You shouldn't use more than the - * first 512 bytes of @root if you wish the transaction to survive a power - * failure. You *must* have a write lock held on @root for both stage (i) - * and (ii). The commit will drop the write lock. + * ii) Lock your superblock. Update. Then call dm_tm_commit() which will + * unlock the superblock and flush it. No other blocks should be updated + * during this period. Care should be taken to never unlock a partially + * updated superblock; perform any operations that could fail *before* you + * take the superblock lock. */ int dm_tm_pre_commit(struct dm_transaction_manager *tm); -int dm_tm_commit(struct dm_transaction_manager *tm, struct dm_block *root); +int dm_tm_commit(struct dm_transaction_manager *tm, struct dm_block *superblock); /* * These methods are the only way to get hold of a writeable block. diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c index fcf65e512cf..407a99e46f6 100644 --- a/drivers/md/raid0.c +++ b/drivers/md/raid0.c @@ -501,10 +501,11 @@ static inline int is_io_in_chunk_boundary(struct mddev *mddev, unsigned int chunk_sects, struct bio *bio) { if (likely(is_power_of_2(chunk_sects))) { - return chunk_sects >= ((bio->bi_sector & (chunk_sects-1)) + return chunk_sects >= + ((bio->bi_iter.bi_sector & (chunk_sects-1)) + bio_sectors(bio)); } else{ - sector_t sector = bio->bi_sector; + sector_t sector = bio->bi_iter.bi_sector; return chunk_sects >= (sector_div(sector, chunk_sects) + bio_sectors(bio)); } @@ -512,64 +513,44 @@ static inline int is_io_in_chunk_boundary(struct mddev *mddev, static void raid0_make_request(struct mddev *mddev, struct bio *bio) { - unsigned int chunk_sects; - sector_t sector_offset; struct strip_zone *zone; struct md_rdev *tmp_dev; + struct bio *split; if (unlikely(bio->bi_rw & REQ_FLUSH)) { md_flush_request(mddev, bio); return; } - chunk_sects = mddev->chunk_sectors; - if (unlikely(!is_io_in_chunk_boundary(mddev, chunk_sects, bio))) { - sector_t sector = bio->bi_sector; - struct bio_pair *bp; - /* Sanity check -- queue functions should prevent this happening */ - if (bio_segments(bio) > 1) - goto bad_map; - /* This is a one page bio that upper layers - * refuse to split for us, so we need to split it. - */ - if (likely(is_power_of_2(chunk_sects))) - bp = bio_split(bio, chunk_sects - (sector & - (chunk_sects-1))); - else - bp = bio_split(bio, chunk_sects - - sector_div(sector, chunk_sects)); - raid0_make_request(mddev, &bp->bio1); - raid0_make_request(mddev, &bp->bio2); - bio_pair_release(bp); - return; - } + do { + sector_t sector = bio->bi_iter.bi_sector; + unsigned chunk_sects = mddev->chunk_sectors; - sector_offset = bio->bi_sector; - zone = find_zone(mddev->private, §or_offset); - tmp_dev = map_sector(mddev, zone, bio->bi_sector, - §or_offset); - bio->bi_bdev = tmp_dev->bdev; - bio->bi_sector = sector_offset + zone->dev_start + - tmp_dev->data_offset; - - if (unlikely((bio->bi_rw & REQ_DISCARD) && - !blk_queue_discard(bdev_get_queue(bio->bi_bdev)))) { - /* Just ignore it */ - bio_endio(bio, 0); - return; - } + unsigned sectors = chunk_sects - + (likely(is_power_of_2(chunk_sects)) + ? (sector & (chunk_sects-1)) + : sector_div(sector, chunk_sects)); - generic_make_request(bio); - return; - -bad_map: - printk("md/raid0:%s: make_request bug: can't convert block across chunks" - " or bigger than %dk %llu %d\n", - mdname(mddev), chunk_sects / 2, - (unsigned long long)bio->bi_sector, bio_sectors(bio) / 2); + if (sectors < bio_sectors(bio)) { + split = bio_split(bio, sectors, GFP_NOIO, fs_bio_set); + bio_chain(split, bio); + } else { + split = bio; + } - bio_io_error(bio); - return; + zone = find_zone(mddev->private, §or); + tmp_dev = map_sector(mddev, zone, sector, §or); + split->bi_bdev = tmp_dev->bdev; + split->bi_iter.bi_sector = sector + zone->dev_start + + tmp_dev->data_offset; + + if (unlikely((split->bi_rw & REQ_DISCARD) && + !blk_queue_discard(bdev_get_queue(split->bi_bdev)))) { + /* Just ignore it */ + bio_endio(split, 0); + } else + generic_make_request(split); + } while (split != bio); } static void raid0_status(struct seq_file *seq, struct mddev *mddev) @@ -597,6 +578,7 @@ static void *raid0_takeover_raid45(struct mddev *mddev) mdname(mddev)); return ERR_PTR(-EINVAL); } + rdev->sectors = mddev->dev_sectors; } /* Set new parameters */ diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index 6e17f8181c4..56e24c072b6 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -66,7 +66,8 @@ */ static int max_queued_requests = 1024; -static void allow_barrier(struct r1conf *conf); +static void allow_barrier(struct r1conf *conf, sector_t start_next_window, + sector_t bi_sector); static void lower_barrier(struct r1conf *conf); static void * r1bio_pool_alloc(gfp_t gfp_flags, void *data) @@ -84,16 +85,19 @@ static void r1bio_pool_free(void *r1_bio, void *data) } #define RESYNC_BLOCK_SIZE (64*1024) -//#define RESYNC_BLOCK_SIZE PAGE_SIZE +#define RESYNC_DEPTH 32 #define RESYNC_SECTORS (RESYNC_BLOCK_SIZE >> 9) #define RESYNC_PAGES ((RESYNC_BLOCK_SIZE + PAGE_SIZE-1) / PAGE_SIZE) -#define RESYNC_WINDOW (2048*1024) +#define RESYNC_WINDOW (RESYNC_BLOCK_SIZE * RESYNC_DEPTH) +#define RESYNC_WINDOW_SECTORS (RESYNC_WINDOW >> 9) +#define NEXT_NORMALIO_DISTANCE (3 * RESYNC_WINDOW_SECTORS) static void * r1buf_pool_alloc(gfp_t gfp_flags, void *data) { struct pool_info *pi = data; struct r1bio *r1_bio; struct bio *bio; + int need_pages; int i, j; r1_bio = r1bio_pool_alloc(gfp_flags, pi); @@ -116,15 +120,15 @@ static void * r1buf_pool_alloc(gfp_t gfp_flags, void *data) * RESYNC_PAGES for each bio. */ if (test_bit(MD_RECOVERY_REQUESTED, &pi->mddev->recovery)) - j = pi->raid_disks; + need_pages = pi->raid_disks; else - j = 1; - while(j--) { + need_pages = 1; + for (j = 0; j < need_pages; j++) { bio = r1_bio->bios[j]; bio->bi_vcnt = RESYNC_PAGES; if (bio_alloc_pages(bio, gfp_flags)) - goto out_free_bio; + goto out_free_pages; } /* If not user-requests, copy the page pointers to all bios */ if (!test_bit(MD_RECOVERY_REQUESTED, &pi->mddev->recovery)) { @@ -138,6 +142,14 @@ static void * r1buf_pool_alloc(gfp_t gfp_flags, void *data) return r1_bio; +out_free_pages: + while (--j >= 0) { + struct bio_vec *bv; + + bio_for_each_segment_all(bv, r1_bio->bios[j], i) + __free_page(bv->bv_page); + } + out_free_bio: while (++j < pi->raid_disks) bio_put(r1_bio->bios[j]); @@ -225,6 +237,8 @@ static void call_bio_endio(struct r1bio *r1_bio) struct bio *bio = r1_bio->master_bio; int done; struct r1conf *conf = r1_bio->mddev->private; + sector_t start_next_window = r1_bio->start_next_window; + sector_t bi_sector = bio->bi_iter.bi_sector; if (bio->bi_phys_segments) { unsigned long flags; @@ -232,6 +246,11 @@ static void call_bio_endio(struct r1bio *r1_bio) bio->bi_phys_segments--; done = (bio->bi_phys_segments == 0); spin_unlock_irqrestore(&conf->device_lock, flags); + /* + * make_request() might be waiting for + * bi_phys_segments to decrease + */ + wake_up(&conf->wait_barrier); } else done = 1; @@ -243,7 +262,7 @@ static void call_bio_endio(struct r1bio *r1_bio) * Wake up any possible resync thread that waits for the device * to go idle. */ - allow_barrier(conf); + allow_barrier(conf, start_next_window, bi_sector); } } @@ -255,9 +274,8 @@ static void raid_end_bio_io(struct r1bio *r1_bio) if (!test_and_set_bit(R1BIO_Returned, &r1_bio->state)) { pr_debug("raid1: sync end %s on sectors %llu-%llu\n", (bio_data_dir(bio) == WRITE) ? "write" : "read", - (unsigned long long) bio->bi_sector, - (unsigned long long) bio->bi_sector + - bio_sectors(bio) - 1); + (unsigned long long) bio->bi_iter.bi_sector, + (unsigned long long) bio_end_sector(bio) - 1); call_bio_endio(r1_bio); } @@ -456,9 +474,8 @@ static void raid1_end_write_request(struct bio *bio, int error) struct bio *mbio = r1_bio->master_bio; pr_debug("raid1: behind end write sectors" " %llu-%llu\n", - (unsigned long long) mbio->bi_sector, - (unsigned long long) mbio->bi_sector + - bio_sectors(mbio) - 1); + (unsigned long long) mbio->bi_iter.bi_sector, + (unsigned long long) bio_end_sector(mbio) - 1); call_bio_endio(r1_bio); } } @@ -814,8 +831,6 @@ static void flush_pending_writes(struct r1conf *conf) * there is no normal IO happeing. It must arrange to call * lower_barrier when the particular background IO completes. */ -#define RESYNC_DEPTH 32 - static void raise_barrier(struct r1conf *conf) { spin_lock_irq(&conf->resync_lock); @@ -827,9 +842,19 @@ static void raise_barrier(struct r1conf *conf) /* block any new IO from starting */ conf->barrier++; - /* Now wait for all pending IO to complete */ + /* For these conditions we must wait: + * A: while the array is in frozen state + * B: while barrier >= RESYNC_DEPTH, meaning resync reach + * the max count which allowed. + * C: next_resync + RESYNC_SECTORS > start_next_window, meaning + * next resync will reach to the window which normal bios are + * handling. + */ wait_event_lock_irq(conf->wait_barrier, - !conf->nr_pending && conf->barrier < RESYNC_DEPTH, + !conf->array_frozen && + conf->barrier < RESYNC_DEPTH && + (conf->start_next_window >= + conf->next_resync + RESYNC_SECTORS), conf->resync_lock); spin_unlock_irq(&conf->resync_lock); @@ -845,10 +870,33 @@ static void lower_barrier(struct r1conf *conf) wake_up(&conf->wait_barrier); } -static void wait_barrier(struct r1conf *conf) +static bool need_to_wait_for_sync(struct r1conf *conf, struct bio *bio) { + bool wait = false; + + if (conf->array_frozen || !bio) + wait = true; + else if (conf->barrier && bio_data_dir(bio) == WRITE) { + if (conf->next_resync < RESYNC_WINDOW_SECTORS) + wait = true; + else if ((conf->next_resync - RESYNC_WINDOW_SECTORS + >= bio_end_sector(bio)) || + (conf->next_resync + NEXT_NORMALIO_DISTANCE + <= bio->bi_iter.bi_sector)) + wait = false; + else + wait = true; + } + + return wait; +} + +static sector_t wait_barrier(struct r1conf *conf, struct bio *bio) +{ + sector_t sector = 0; + spin_lock_irq(&conf->resync_lock); - if (conf->barrier) { + if (need_to_wait_for_sync(conf, bio)) { conf->nr_waiting++; /* Wait for the barrier to drop. * However if there are already pending @@ -860,22 +908,66 @@ static void wait_barrier(struct r1conf *conf) * count down. */ wait_event_lock_irq(conf->wait_barrier, - !conf->barrier || - (conf->nr_pending && + !conf->array_frozen && + (!conf->barrier || + ((conf->start_next_window < + conf->next_resync + RESYNC_SECTORS) && current->bio_list && - !bio_list_empty(current->bio_list)), + !bio_list_empty(current->bio_list))), conf->resync_lock); conf->nr_waiting--; } + + if (bio && bio_data_dir(bio) == WRITE) { + if (conf->next_resync + NEXT_NORMALIO_DISTANCE + <= bio->bi_iter.bi_sector) { + if (conf->start_next_window == MaxSector) + conf->start_next_window = + conf->next_resync + + NEXT_NORMALIO_DISTANCE; + + if ((conf->start_next_window + NEXT_NORMALIO_DISTANCE) + <= bio->bi_iter.bi_sector) + conf->next_window_requests++; + else + conf->current_window_requests++; + sector = conf->start_next_window; + } + } + conf->nr_pending++; spin_unlock_irq(&conf->resync_lock); + return sector; } -static void allow_barrier(struct r1conf *conf) +static void allow_barrier(struct r1conf *conf, sector_t start_next_window, + sector_t bi_sector) { unsigned long flags; + spin_lock_irqsave(&conf->resync_lock, flags); conf->nr_pending--; + if (start_next_window) { + if (start_next_window == conf->start_next_window) { + if (conf->start_next_window + NEXT_NORMALIO_DISTANCE + <= bi_sector) + conf->next_window_requests--; + else + conf->current_window_requests--; + } else + conf->current_window_requests--; + + if (!conf->current_window_requests) { + if (conf->next_window_requests) { + conf->current_window_requests = + conf->next_window_requests; + conf->next_window_requests = 0; + conf->start_next_window += + NEXT_NORMALIO_DISTANCE; + } else + conf->start_next_window = MaxSector; + } + } spin_unlock_irqrestore(&conf->resync_lock, flags); wake_up(&conf->wait_barrier); } @@ -884,8 +976,7 @@ static void freeze_array(struct r1conf *conf, int extra) { /* stop syncio and normal IO and wait for everything to * go quite. - * We increment barrier and nr_waiting, and then - * wait until nr_pending match nr_queued+extra + * We wait until nr_pending match nr_queued+extra * This is called in the context of one normal IO request * that has failed. Thus any sync request that might be pending * will be blocked by nr_pending, and we need to wait for @@ -895,8 +986,7 @@ static void freeze_array(struct r1conf *conf, int extra) * we continue. */ spin_lock_irq(&conf->resync_lock); - conf->barrier++; - conf->nr_waiting++; + conf->array_frozen = 1; wait_event_lock_irq_cmd(conf->wait_barrier, conf->nr_pending == conf->nr_queued+extra, conf->resync_lock, @@ -907,8 +997,7 @@ static void unfreeze_array(struct r1conf *conf) { /* reverse the effect of the freeze */ spin_lock_irq(&conf->resync_lock); - conf->barrier--; - conf->nr_waiting--; + conf->array_frozen = 0; wake_up(&conf->wait_barrier); spin_unlock_irq(&conf->resync_lock); } @@ -945,7 +1034,8 @@ do_sync_io: if (bvecs[i].bv_page) put_page(bvecs[i].bv_page); kfree(bvecs); - pr_debug("%dB behind alloc failed, doing sync I/O\n", bio->bi_size); + pr_debug("%dB behind alloc failed, doing sync I/O\n", + bio->bi_iter.bi_size); } struct raid1_plug_cb { @@ -1013,6 +1103,7 @@ static void make_request(struct mddev *mddev, struct bio * bio) int first_clone; int sectors_handled; int max_sectors; + sector_t start_next_window; /* * Register the new request and wait if the reconstruction @@ -1024,7 +1115,7 @@ static void make_request(struct mddev *mddev, struct bio * bio) if (bio_data_dir(bio) == WRITE && bio_end_sector(bio) > mddev->suspend_lo && - bio->bi_sector < mddev->suspend_hi) { + bio->bi_iter.bi_sector < mddev->suspend_hi) { /* As the suspend_* range is controlled by * userspace, we want an interruptible * wait. @@ -1035,14 +1126,14 @@ static void make_request(struct mddev *mddev, struct bio * bio) prepare_to_wait(&conf->wait_barrier, &w, TASK_INTERRUPTIBLE); if (bio_end_sector(bio) <= mddev->suspend_lo || - bio->bi_sector >= mddev->suspend_hi) + bio->bi_iter.bi_sector >= mddev->suspend_hi) break; schedule(); } finish_wait(&conf->wait_barrier, &w); } - wait_barrier(conf); + start_next_window = wait_barrier(conf, bio); bitmap = mddev->bitmap; @@ -1057,7 +1148,7 @@ static void make_request(struct mddev *mddev, struct bio * bio) r1_bio->sectors = bio_sectors(bio); r1_bio->state = 0; r1_bio->mddev = mddev; - r1_bio->sector = bio->bi_sector; + r1_bio->sector = bio->bi_iter.bi_sector; /* We might need to issue multiple reads to different * devices if there are bad blocks around, so we keep @@ -1097,12 +1188,13 @@ read_again: r1_bio->read_disk = rdisk; read_bio = bio_clone_mddev(bio, GFP_NOIO, mddev); - md_trim_bio(read_bio, r1_bio->sector - bio->bi_sector, - max_sectors); + bio_trim(read_bio, r1_bio->sector - bio->bi_iter.bi_sector, + max_sectors); r1_bio->bios[rdisk] = read_bio; - read_bio->bi_sector = r1_bio->sector + mirror->rdev->data_offset; + read_bio->bi_iter.bi_sector = r1_bio->sector + + mirror->rdev->data_offset; read_bio->bi_bdev = mirror->rdev->bdev; read_bio->bi_end_io = raid1_end_read_request; read_bio->bi_rw = READ | do_sync; @@ -1114,7 +1206,7 @@ read_again: */ sectors_handled = (r1_bio->sector + max_sectors - - bio->bi_sector); + - bio->bi_iter.bi_sector); r1_bio->sectors = max_sectors; spin_lock_irq(&conf->device_lock); if (bio->bi_phys_segments == 0) @@ -1135,7 +1227,8 @@ read_again: r1_bio->sectors = bio_sectors(bio) - sectors_handled; r1_bio->state = 0; r1_bio->mddev = mddev; - r1_bio->sector = bio->bi_sector + sectors_handled; + r1_bio->sector = bio->bi_iter.bi_sector + + sectors_handled; goto read_again; } else generic_make_request(read_bio); @@ -1163,6 +1256,7 @@ read_again: disks = conf->raid_disks * 2; retry_write: + r1_bio->start_next_window = start_next_window; blocked_rdev = NULL; rcu_read_lock(); max_sectors = r1_bio->sectors; @@ -1231,14 +1325,24 @@ read_again: if (unlikely(blocked_rdev)) { /* Wait for this device to become unblocked */ int j; + sector_t old = start_next_window; for (j = 0; j < i; j++) if (r1_bio->bios[j]) rdev_dec_pending(conf->mirrors[j].rdev, mddev); r1_bio->state = 0; - allow_barrier(conf); + allow_barrier(conf, start_next_window, bio->bi_iter.bi_sector); md_wait_for_blocked_rdev(blocked_rdev, mddev); - wait_barrier(conf); + start_next_window = wait_barrier(conf, bio); + /* + * We must make sure the multi r1bios of bio have + * the same value of bi_phys_segments + */ + if (bio->bi_phys_segments && old && + old != start_next_window) + /* Wait for the former r1bio(s) to complete */ + wait_event(conf->wait_barrier, + bio->bi_phys_segments == 1); goto retry_write; } @@ -1254,7 +1358,7 @@ read_again: bio->bi_phys_segments++; spin_unlock_irq(&conf->device_lock); } - sectors_handled = r1_bio->sector + max_sectors - bio->bi_sector; + sectors_handled = r1_bio->sector + max_sectors - bio->bi_iter.bi_sector; atomic_set(&r1_bio->remaining, 1); atomic_set(&r1_bio->behind_remaining, 0); @@ -1266,7 +1370,7 @@ read_again: continue; mbio = bio_clone_mddev(bio, GFP_NOIO, mddev); - md_trim_bio(mbio, r1_bio->sector - bio->bi_sector, max_sectors); + bio_trim(mbio, r1_bio->sector - bio->bi_iter.bi_sector, max_sectors); if (first_clone) { /* do behind I/O ? @@ -1300,7 +1404,7 @@ read_again: r1_bio->bios[i] = mbio; - mbio->bi_sector = (r1_bio->sector + + mbio->bi_iter.bi_sector = (r1_bio->sector + conf->mirrors[i].rdev->data_offset); mbio->bi_bdev = conf->mirrors[i].rdev->bdev; mbio->bi_end_io = raid1_end_write_request; @@ -1340,7 +1444,7 @@ read_again: r1_bio->sectors = bio_sectors(bio) - sectors_handled; r1_bio->state = 0; r1_bio->mddev = mddev; - r1_bio->sector = bio->bi_sector + sectors_handled; + r1_bio->sector = bio->bi_iter.bi_sector + sectors_handled; goto retry_write; } @@ -1438,11 +1542,14 @@ static void print_conf(struct r1conf *conf) static void close_sync(struct r1conf *conf) { - wait_barrier(conf); - allow_barrier(conf); + wait_barrier(conf, NULL); + allow_barrier(conf, 0, 0); mempool_destroy(conf->r1buf_pool); conf->r1buf_pool = NULL; + + conf->next_resync = 0; + conf->start_next_window = MaxSector; } static int raid1_spare_active(struct mddev *mddev) @@ -1479,6 +1586,7 @@ static int raid1_spare_active(struct mddev *mddev) } } if (rdev + && rdev->recovery_offset == MaxSector && !test_bit(Faulty, &rdev->flags) && !test_and_set_bit(In_sync, &rdev->flags)) { count++; @@ -1519,8 +1627,9 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev) p = conf->mirrors+mirror; if (!p->rdev) { - disk_stack_limits(mddev->gendisk, rdev->bdev, - rdev->data_offset << 9); + if (mddev->gendisk) + disk_stack_limits(mddev->gendisk, rdev->bdev, + rdev->data_offset << 9); p->head_position = 0; rdev->raid_disk = mirror; @@ -1559,7 +1668,7 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev) clear_bit(Unmerged, &rdev->flags); } md_integrity_add_rdev(rdev, mddev); - if (blk_queue_discard(bdev_get_queue(rdev->bdev))) + if (mddev->queue && blk_queue_discard(bdev_get_queue(rdev->bdev))) queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, mddev->queue); print_conf(conf); return err; @@ -1848,6 +1957,40 @@ static int process_checks(struct r1bio *r1_bio) int i; int vcnt; + /* Fix variable parts of all bios */ + vcnt = (r1_bio->sectors + PAGE_SIZE / 512 - 1) >> (PAGE_SHIFT - 9); + for (i = 0; i < conf->raid_disks * 2; i++) { + int j; + int size; + int uptodate; + struct bio *b = r1_bio->bios[i]; + if (b->bi_end_io != end_sync_read) + continue; + /* fixup the bio for reuse, but preserve BIO_UPTODATE */ + uptodate = test_bit(BIO_UPTODATE, &b->bi_flags); + bio_reset(b); + if (!uptodate) + clear_bit(BIO_UPTODATE, &b->bi_flags); + b->bi_vcnt = vcnt; + b->bi_iter.bi_size = r1_bio->sectors << 9; + b->bi_iter.bi_sector = r1_bio->sector + + conf->mirrors[i].rdev->data_offset; + b->bi_bdev = conf->mirrors[i].rdev->bdev; + b->bi_end_io = end_sync_read; + b->bi_private = r1_bio; + + size = b->bi_iter.bi_size; + for (j = 0; j < vcnt ; j++) { + struct bio_vec *bi; + bi = &b->bi_io_vec[j]; + bi->bv_offset = 0; + if (size > PAGE_SIZE) + bi->bv_len = PAGE_SIZE; + else + bi->bv_len = size; + size -= PAGE_SIZE; + } + } for (primary = 0; primary < conf->raid_disks * 2; primary++) if (r1_bio->bios[primary]->bi_end_io == end_sync_read && test_bit(BIO_UPTODATE, &r1_bio->bios[primary]->bi_flags)) { @@ -1856,17 +1999,18 @@ static int process_checks(struct r1bio *r1_bio) break; } r1_bio->read_disk = primary; - vcnt = (r1_bio->sectors + PAGE_SIZE / 512 - 1) >> (PAGE_SHIFT - 9); for (i = 0; i < conf->raid_disks * 2; i++) { int j; struct bio *pbio = r1_bio->bios[primary]; struct bio *sbio = r1_bio->bios[i]; - int size; + int uptodate = test_bit(BIO_UPTODATE, &sbio->bi_flags); if (sbio->bi_end_io != end_sync_read) continue; + /* Now we can 'fixup' the BIO_UPTODATE flag */ + set_bit(BIO_UPTODATE, &sbio->bi_flags); - if (test_bit(BIO_UPTODATE, &sbio->bi_flags)) { + if (uptodate) { for (j = vcnt; j-- ; ) { struct page *p, *s; p = pbio->bi_io_vec[j].bv_page; @@ -1881,33 +2025,12 @@ static int process_checks(struct r1bio *r1_bio) if (j >= 0) atomic64_add(r1_bio->sectors, &mddev->resync_mismatches); if (j < 0 || (test_bit(MD_RECOVERY_CHECK, &mddev->recovery) - && test_bit(BIO_UPTODATE, &sbio->bi_flags))) { + && uptodate)) { /* No need to write to this device. */ sbio->bi_end_io = NULL; rdev_dec_pending(conf->mirrors[i].rdev, mddev); continue; } - /* fixup the bio for reuse */ - bio_reset(sbio); - sbio->bi_vcnt = vcnt; - sbio->bi_size = r1_bio->sectors << 9; - sbio->bi_sector = r1_bio->sector + - conf->mirrors[i].rdev->data_offset; - sbio->bi_bdev = conf->mirrors[i].rdev->bdev; - sbio->bi_end_io = end_sync_read; - sbio->bi_private = r1_bio; - - size = sbio->bi_size; - for (j = 0; j < vcnt ; j++) { - struct bio_vec *bi; - bi = &sbio->bi_io_vec[j]; - bi->bv_offset = 0; - if (size > PAGE_SIZE) - bi->bv_len = PAGE_SIZE; - else - bi->bv_len = size; - size -= PAGE_SIZE; - } bio_copy_data(sbio, pbio); } @@ -2114,11 +2237,11 @@ static int narrow_write_error(struct r1bio *r1_bio, int i) } wbio->bi_rw = WRITE; - wbio->bi_sector = r1_bio->sector; - wbio->bi_size = r1_bio->sectors << 9; + wbio->bi_iter.bi_sector = r1_bio->sector; + wbio->bi_iter.bi_size = r1_bio->sectors << 9; - md_trim_bio(wbio, sector - r1_bio->sector, sectors); - wbio->bi_sector += rdev->data_offset; + bio_trim(wbio, sector - r1_bio->sector, sectors); + wbio->bi_iter.bi_sector += rdev->data_offset; wbio->bi_bdev = rdev->bdev; if (submit_bio_wait(WRITE, wbio) == 0) /* failure! */ @@ -2232,7 +2355,8 @@ read_more: } r1_bio->read_disk = disk; bio = bio_clone_mddev(r1_bio->master_bio, GFP_NOIO, mddev); - md_trim_bio(bio, r1_bio->sector - bio->bi_sector, max_sectors); + bio_trim(bio, r1_bio->sector - bio->bi_iter.bi_sector, + max_sectors); r1_bio->bios[r1_bio->read_disk] = bio; rdev = conf->mirrors[disk].rdev; printk_ratelimited(KERN_ERR @@ -2241,7 +2365,7 @@ read_more: mdname(mddev), (unsigned long long)r1_bio->sector, bdevname(rdev->bdev, b)); - bio->bi_sector = r1_bio->sector + rdev->data_offset; + bio->bi_iter.bi_sector = r1_bio->sector + rdev->data_offset; bio->bi_bdev = rdev->bdev; bio->bi_end_io = raid1_end_read_request; bio->bi_rw = READ | do_sync; @@ -2250,7 +2374,7 @@ read_more: /* Drat - have to split this up more */ struct bio *mbio = r1_bio->master_bio; int sectors_handled = (r1_bio->sector + max_sectors - - mbio->bi_sector); + - mbio->bi_iter.bi_sector); r1_bio->sectors = max_sectors; spin_lock_irq(&conf->device_lock); if (mbio->bi_phys_segments == 0) @@ -2268,7 +2392,8 @@ read_more: r1_bio->state = 0; set_bit(R1BIO_ReadError, &r1_bio->state); r1_bio->mddev = mddev; - r1_bio->sector = mbio->bi_sector + sectors_handled; + r1_bio->sector = mbio->bi_iter.bi_sector + + sectors_handled; goto read_more; } else @@ -2492,7 +2617,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int *skipp } if (bio->bi_end_io) { atomic_inc(&rdev->nr_pending); - bio->bi_sector = sector_nr + rdev->data_offset; + bio->bi_iter.bi_sector = sector_nr + rdev->data_offset; bio->bi_bdev = rdev->bdev; bio->bi_private = r1_bio; } @@ -2592,7 +2717,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int *skipp continue; /* remove last page from this bio */ bio->bi_vcnt--; - bio->bi_size -= len; + bio->bi_iter.bi_size -= len; bio->bi_flags &= ~(1<< BIO_SEG_VALID); } goto bio_full; @@ -2705,6 +2830,9 @@ static struct r1conf *setup_conf(struct mddev *mddev) conf->pending_count = 0; conf->recovery_disabled = mddev->recovery_disabled - 1; + conf->start_next_window = MaxSector; + conf->current_window_requests = conf->next_window_requests = 0; + err = -EIO; for (i = 0; i < conf->raid_disks * 2; i++) { @@ -2862,8 +2990,8 @@ static int stop(struct mddev *mddev) atomic_read(&bitmap->behind_writes) == 0); } - raise_barrier(conf); - lower_barrier(conf); + freeze_array(conf, 0); + unfreeze_array(conf); md_unregister_thread(&mddev->thread); if (conf->r1bio_pool) @@ -3022,10 +3150,10 @@ static void raid1_quiesce(struct mddev *mddev, int state) wake_up(&conf->wait_barrier); break; case 1: - raise_barrier(conf); + freeze_array(conf, 0); break; case 0: - lower_barrier(conf); + unfreeze_array(conf); break; } } @@ -3042,7 +3170,8 @@ static void *raid1_takeover(struct mddev *mddev) mddev->new_chunk_sectors = 0; conf = setup_conf(mddev); if (!IS_ERR(conf)) - conf->barrier = 1; + /* Array must appear to be quiesced */ + conf->array_frozen = 1; return conf; } return ERR_PTR(-EINVAL); diff --git a/drivers/md/raid1.h b/drivers/md/raid1.h index 0ff3715fb7e..9bebca7bff2 100644 --- a/drivers/md/raid1.h +++ b/drivers/md/raid1.h @@ -41,6 +41,19 @@ struct r1conf { */ sector_t next_resync; + /* When raid1 starts resync, we divide array into four partitions + * |---------|--------------|---------------------|-------------| + * next_resync start_next_window end_window + * start_next_window = next_resync + NEXT_NORMALIO_DISTANCE + * end_window = start_next_window + NEXT_NORMALIO_DISTANCE + * current_window_requests means the count of normalIO between + * start_next_window and end_window. + * next_window_requests means the count of normalIO after end_window. + * */ + sector_t start_next_window; + int current_window_requests; + int next_window_requests; + spinlock_t device_lock; /* list of 'struct r1bio' that need to be processed by raid1d, @@ -65,6 +78,7 @@ struct r1conf { int nr_waiting; int nr_queued; int barrier; + int array_frozen; /* Set to 1 if a full sync is needed, (fresh device added). * Cleared when a sync completes. @@ -111,6 +125,7 @@ struct r1bio { * in this BehindIO request */ sector_t sector; + sector_t start_next_window; int sectors; unsigned long state; struct mddev *mddev; diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index 6ddae2501b9..cb882aae9e2 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -97,7 +97,7 @@ static int max_queued_requests = 1024; static void allow_barrier(struct r10conf *conf); static void lower_barrier(struct r10conf *conf); -static int enough(struct r10conf *conf, int ignore); +static int _enough(struct r10conf *conf, int previous, int ignore); static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr, int *skipped); static void reshape_request_write(struct mddev *mddev, struct r10bio *r10_bio); @@ -392,11 +392,9 @@ static void raid10_end_read_request(struct bio *bio, int error) * than fail the last device. Here we redefine * "uptodate" to mean "Don't want to retry" */ - unsigned long flags; - spin_lock_irqsave(&conf->device_lock, flags); - if (!enough(conf, rdev->raid_disk)) + if (!_enough(conf, test_bit(R10BIO_Previous, &r10_bio->state), + rdev->raid_disk)) uptodate = 1; - spin_unlock_irqrestore(&conf->device_lock, flags); } if (uptodate) { raid_end_bio_io(r10_bio); @@ -1154,14 +1152,12 @@ static void raid10_unplug(struct blk_plug_cb *cb, bool from_schedule) kfree(plug); } -static void make_request(struct mddev *mddev, struct bio * bio) +static void __make_request(struct mddev *mddev, struct bio *bio) { struct r10conf *conf = mddev->private; struct r10bio *r10_bio; struct bio *read_bio; int i; - sector_t chunk_mask = (conf->geo.chunk_mask & conf->prev.chunk_mask); - int chunk_sects = chunk_mask + 1; const int rw = bio_data_dir(bio); const unsigned long do_sync = (bio->bi_rw & REQ_SYNC); const unsigned long do_fua = (bio->bi_rw & REQ_FUA); @@ -1176,61 +1172,6 @@ static void make_request(struct mddev *mddev, struct bio * bio) int max_sectors; int sectors; - if (unlikely(bio->bi_rw & REQ_FLUSH)) { - md_flush_request(mddev, bio); - return; - } - - /* If this request crosses a chunk boundary, we need to - * split it. This will only happen for 1 PAGE (or less) requests. - */ - if (unlikely((bio->bi_sector & chunk_mask) + bio_sectors(bio) - > chunk_sects - && (conf->geo.near_copies < conf->geo.raid_disks - || conf->prev.near_copies < conf->prev.raid_disks))) { - struct bio_pair *bp; - /* Sanity check -- queue functions should prevent this happening */ - if (bio_segments(bio) > 1) - goto bad_map; - /* This is a one page bio that upper layers - * refuse to split for us, so we need to split it. - */ - bp = bio_split(bio, - chunk_sects - (bio->bi_sector & (chunk_sects - 1)) ); - - /* Each of these 'make_request' calls will call 'wait_barrier'. - * If the first succeeds but the second blocks due to the resync - * thread raising the barrier, we will deadlock because the - * IO to the underlying device will be queued in generic_make_request - * and will never complete, so will never reduce nr_pending. - * So increment nr_waiting here so no new raise_barriers will - * succeed, and so the second wait_barrier cannot block. - */ - spin_lock_irq(&conf->resync_lock); - conf->nr_waiting++; - spin_unlock_irq(&conf->resync_lock); - - make_request(mddev, &bp->bio1); - make_request(mddev, &bp->bio2); - - spin_lock_irq(&conf->resync_lock); - conf->nr_waiting--; - wake_up(&conf->wait_barrier); - spin_unlock_irq(&conf->resync_lock); - - bio_pair_release(bp); - return; - bad_map: - printk("md/raid10:%s: make_request bug: can't convert block across chunks" - " or bigger than %dk %llu %d\n", mdname(mddev), chunk_sects/2, - (unsigned long long)bio->bi_sector, bio_sectors(bio) / 2); - - bio_io_error(bio); - return; - } - - md_write_start(mddev, bio); - /* * Register the new request and wait if the reconstruction * thread has put up a bar for new requests. @@ -1240,24 +1181,25 @@ static void make_request(struct mddev *mddev, struct bio * bio) sectors = bio_sectors(bio); while (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) && - bio->bi_sector < conf->reshape_progress && - bio->bi_sector + sectors > conf->reshape_progress) { + bio->bi_iter.bi_sector < conf->reshape_progress && + bio->bi_iter.bi_sector + sectors > conf->reshape_progress) { /* IO spans the reshape position. Need to wait for * reshape to pass */ allow_barrier(conf); wait_event(conf->wait_barrier, - conf->reshape_progress <= bio->bi_sector || - conf->reshape_progress >= bio->bi_sector + sectors); + conf->reshape_progress <= bio->bi_iter.bi_sector || + conf->reshape_progress >= bio->bi_iter.bi_sector + + sectors); wait_barrier(conf); } if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) && bio_data_dir(bio) == WRITE && (mddev->reshape_backwards - ? (bio->bi_sector < conf->reshape_safe && - bio->bi_sector + sectors > conf->reshape_progress) - : (bio->bi_sector + sectors > conf->reshape_safe && - bio->bi_sector < conf->reshape_progress))) { + ? (bio->bi_iter.bi_sector < conf->reshape_safe && + bio->bi_iter.bi_sector + sectors > conf->reshape_progress) + : (bio->bi_iter.bi_sector + sectors > conf->reshape_safe && + bio->bi_iter.bi_sector < conf->reshape_progress))) { /* Need to update reshape_position in metadata */ mddev->reshape_position = conf->reshape_progress; set_bit(MD_CHANGE_DEVS, &mddev->flags); @@ -1275,7 +1217,7 @@ static void make_request(struct mddev *mddev, struct bio * bio) r10_bio->sectors = sectors; r10_bio->mddev = mddev; - r10_bio->sector = bio->bi_sector; + r10_bio->sector = bio->bi_iter.bi_sector; r10_bio->state = 0; /* We might need to issue multiple reads to different @@ -1304,13 +1246,13 @@ read_again: slot = r10_bio->read_slot; read_bio = bio_clone_mddev(bio, GFP_NOIO, mddev); - md_trim_bio(read_bio, r10_bio->sector - bio->bi_sector, - max_sectors); + bio_trim(read_bio, r10_bio->sector - bio->bi_iter.bi_sector, + max_sectors); r10_bio->devs[slot].bio = read_bio; r10_bio->devs[slot].rdev = rdev; - read_bio->bi_sector = r10_bio->devs[slot].addr + + read_bio->bi_iter.bi_sector = r10_bio->devs[slot].addr + choose_data_offset(r10_bio, rdev); read_bio->bi_bdev = rdev->bdev; read_bio->bi_end_io = raid10_end_read_request; @@ -1321,15 +1263,15 @@ read_again: /* Could not read all from this device, so we will * need another r10_bio. */ - sectors_handled = (r10_bio->sectors + max_sectors - - bio->bi_sector); + sectors_handled = (r10_bio->sector + max_sectors + - bio->bi_iter.bi_sector); r10_bio->sectors = max_sectors; spin_lock_irq(&conf->device_lock); if (bio->bi_phys_segments == 0) bio->bi_phys_segments = 2; else bio->bi_phys_segments++; - spin_unlock(&conf->device_lock); + spin_unlock_irq(&conf->device_lock); /* Cannot call generic_make_request directly * as that will be queued in __generic_make_request * and subsequent mempool_alloc might block @@ -1343,7 +1285,8 @@ read_again: r10_bio->sectors = bio_sectors(bio) - sectors_handled; r10_bio->state = 0; r10_bio->mddev = mddev; - r10_bio->sector = bio->bi_sector + sectors_handled; + r10_bio->sector = bio->bi_iter.bi_sector + + sectors_handled; goto read_again; } else generic_make_request(read_bio); @@ -1501,7 +1444,8 @@ retry_write: bio->bi_phys_segments++; spin_unlock_irq(&conf->device_lock); } - sectors_handled = r10_bio->sector + max_sectors - bio->bi_sector; + sectors_handled = r10_bio->sector + max_sectors - + bio->bi_iter.bi_sector; atomic_set(&r10_bio->remaining, 1); bitmap_startwrite(mddev->bitmap, r10_bio->sector, r10_bio->sectors, 0); @@ -1512,11 +1456,11 @@ retry_write: if (r10_bio->devs[i].bio) { struct md_rdev *rdev = conf->mirrors[d].rdev; mbio = bio_clone_mddev(bio, GFP_NOIO, mddev); - md_trim_bio(mbio, r10_bio->sector - bio->bi_sector, - max_sectors); + bio_trim(mbio, r10_bio->sector - bio->bi_iter.bi_sector, + max_sectors); r10_bio->devs[i].bio = mbio; - mbio->bi_sector = (r10_bio->devs[i].addr+ + mbio->bi_iter.bi_sector = (r10_bio->devs[i].addr+ choose_data_offset(r10_bio, rdev)); mbio->bi_bdev = rdev->bdev; @@ -1555,11 +1499,11 @@ retry_write: rdev = conf->mirrors[d].rdev; } mbio = bio_clone_mddev(bio, GFP_NOIO, mddev); - md_trim_bio(mbio, r10_bio->sector - bio->bi_sector, - max_sectors); + bio_trim(mbio, r10_bio->sector - bio->bi_iter.bi_sector, + max_sectors); r10_bio->devs[i].repl_bio = mbio; - mbio->bi_sector = (r10_bio->devs[i].addr + + mbio->bi_iter.bi_sector = (r10_bio->devs[i].addr + choose_data_offset( r10_bio, rdev)); mbio->bi_bdev = rdev->bdev; @@ -1593,11 +1537,51 @@ retry_write: r10_bio->sectors = bio_sectors(bio) - sectors_handled; r10_bio->mddev = mddev; - r10_bio->sector = bio->bi_sector + sectors_handled; + r10_bio->sector = bio->bi_iter.bi_sector + sectors_handled; r10_bio->state = 0; goto retry_write; } one_write_done(r10_bio); +} + +static void make_request(struct mddev *mddev, struct bio *bio) +{ + struct r10conf *conf = mddev->private; + sector_t chunk_mask = (conf->geo.chunk_mask & conf->prev.chunk_mask); + int chunk_sects = chunk_mask + 1; + + struct bio *split; + + if (unlikely(bio->bi_rw & REQ_FLUSH)) { + md_flush_request(mddev, bio); + return; + } + + md_write_start(mddev, bio); + + + do { + + /* + * If this request crosses a chunk boundary, we need to split + * it. + */ + if (unlikely((bio->bi_iter.bi_sector & chunk_mask) + + bio_sectors(bio) > chunk_sects + && (conf->geo.near_copies < conf->geo.raid_disks + || conf->prev.near_copies < + conf->prev.raid_disks))) { + split = bio_split(bio, chunk_sects - + (bio->bi_iter.bi_sector & + (chunk_sects - 1)), + GFP_NOIO, fs_bio_set); + bio_chain(split, bio); + } else { + split = bio; + } + + __make_request(mddev, split); + } while (split != bio); /* In case raid10d snuck in to freeze_array */ wake_up(&conf->wait_barrier); @@ -1632,37 +1616,58 @@ static void status(struct seq_file *seq, struct mddev *mddev) * Don't consider the device numbered 'ignore' * as we might be about to remove it. */ -static int _enough(struct r10conf *conf, struct geom *geo, int ignore) +static int _enough(struct r10conf *conf, int previous, int ignore) { int first = 0; + int has_enough = 0; + int disks, ncopies; + if (previous) { + disks = conf->prev.raid_disks; + ncopies = conf->prev.near_copies; + } else { + disks = conf->geo.raid_disks; + ncopies = conf->geo.near_copies; + } + rcu_read_lock(); do { int n = conf->copies; int cnt = 0; int this = first; while (n--) { - if (conf->mirrors[this].rdev && - this != ignore) + struct md_rdev *rdev; + if (this != ignore && + (rdev = rcu_dereference(conf->mirrors[this].rdev)) && + test_bit(In_sync, &rdev->flags)) cnt++; - this = (this+1) % geo->raid_disks; + this = (this+1) % disks; } if (cnt == 0) - return 0; - first = (first + geo->near_copies) % geo->raid_disks; + goto out; + first = (first + ncopies) % disks; } while (first != 0); - return 1; + has_enough = 1; +out: + rcu_read_unlock(); + return has_enough; } static int enough(struct r10conf *conf, int ignore) { - return _enough(conf, &conf->geo, ignore) && - _enough(conf, &conf->prev, ignore); + /* when calling 'enough', both 'prev' and 'geo' must + * be stable. + * This is ensured if ->reconfig_mutex or ->device_lock + * is held. + */ + return _enough(conf, 0, ignore) && + _enough(conf, 1, ignore); } static void error(struct mddev *mddev, struct md_rdev *rdev) { char b[BDEVNAME_SIZE]; struct r10conf *conf = mddev->private; + unsigned long flags; /* * If it is not operational, then we have already marked it as dead @@ -1670,18 +1675,18 @@ static void error(struct mddev *mddev, struct md_rdev *rdev) * next level up know. * else mark the drive as failed */ + spin_lock_irqsave(&conf->device_lock, flags); if (test_bit(In_sync, &rdev->flags) - && !enough(conf, rdev->raid_disk)) + && !enough(conf, rdev->raid_disk)) { /* * Don't fail the drive, just return an IO error. */ + spin_unlock_irqrestore(&conf->device_lock, flags); return; + } if (test_and_clear_bit(In_sync, &rdev->flags)) { - unsigned long flags; - spin_lock_irqsave(&conf->device_lock, flags); mddev->degraded++; - spin_unlock_irqrestore(&conf->device_lock, flags); - /* + /* * if recovery is running, make sure it aborts. */ set_bit(MD_RECOVERY_INTR, &mddev->recovery); @@ -1689,6 +1694,7 @@ static void error(struct mddev *mddev, struct md_rdev *rdev) set_bit(Blocked, &rdev->flags); set_bit(Faulty, &rdev->flags); set_bit(MD_CHANGE_DEVS, &mddev->flags); + spin_unlock_irqrestore(&conf->device_lock, flags); printk(KERN_ALERT "md/raid10:%s: Disk failure on %s, disabling device.\n" "md/raid10:%s: Operation continuing on %d devices.\n", @@ -1762,6 +1768,7 @@ static int raid10_spare_active(struct mddev *mddev) } sysfs_notify_dirent_safe(tmp->replacement->sysfs_state); } else if (tmp->rdev + && tmp->rdev->recovery_offset == MaxSector && !test_bit(Faulty, &tmp->rdev->flags) && !test_and_set_bit(In_sync, &tmp->rdev->flags)) { count++; @@ -1791,7 +1798,7 @@ static int raid10_add_disk(struct mddev *mddev, struct md_rdev *rdev) * very different from resync */ return -EBUSY; - if (rdev->saved_raid_disk < 0 && !_enough(conf, &conf->prev, -1)) + if (rdev->saved_raid_disk < 0 && !_enough(conf, 1, -1)) return -EINVAL; if (rdev->raid_disk >= 0) @@ -1819,15 +1826,17 @@ static int raid10_add_disk(struct mddev *mddev, struct md_rdev *rdev) set_bit(Replacement, &rdev->flags); rdev->raid_disk = mirror; err = 0; - disk_stack_limits(mddev->gendisk, rdev->bdev, - rdev->data_offset << 9); + if (mddev->gendisk) + disk_stack_limits(mddev->gendisk, rdev->bdev, + rdev->data_offset << 9); conf->fullsync = 1; rcu_assign_pointer(p->replacement, rdev); break; } - disk_stack_limits(mddev->gendisk, rdev->bdev, - rdev->data_offset << 9); + if (mddev->gendisk) + disk_stack_limits(mddev->gendisk, rdev->bdev, + rdev->data_offset << 9); p->head_position = 0; p->recovery_disabled = mddev->recovery_disabled - 1; @@ -2075,11 +2084,17 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) * both 'first' and 'i', so we just compare them. * All vec entries are PAGE_SIZE; */ - for (j = 0; j < vcnt; j++) + int sectors = r10_bio->sectors; + for (j = 0; j < vcnt; j++) { + int len = PAGE_SIZE; + if (sectors < (len / 512)) + len = sectors * 512; if (memcmp(page_address(fbio->bi_io_vec[j].bv_page), page_address(tbio->bi_io_vec[j].bv_page), - fbio->bi_io_vec[j].bv_len)) + len)) break; + sectors -= len/512; + } if (j == vcnt) continue; atomic64_add(r10_bio->sectors, &mddev->resync_mismatches); @@ -2095,10 +2110,10 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) bio_reset(tbio); tbio->bi_vcnt = vcnt; - tbio->bi_size = r10_bio->sectors << 9; + tbio->bi_iter.bi_size = r10_bio->sectors << 9; tbio->bi_rw = WRITE; tbio->bi_private = r10_bio; - tbio->bi_sector = r10_bio->devs[i].addr; + tbio->bi_iter.bi_sector = r10_bio->devs[i].addr; for (j=0; j < vcnt ; j++) { tbio->bi_io_vec[j].bv_offset = 0; @@ -2115,7 +2130,7 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio) atomic_inc(&r10_bio->remaining); md_sync_acct(conf->mirrors[d].rdev->bdev, bio_sectors(tbio)); - tbio->bi_sector += conf->mirrors[d].rdev->data_offset; + tbio->bi_iter.bi_sector += conf->mirrors[d].rdev->data_offset; tbio->bi_bdev = conf->mirrors[d].rdev->bdev; generic_make_request(tbio); } @@ -2262,12 +2277,18 @@ static void recovery_request_write(struct mddev *mddev, struct r10bio *r10_bio) d = r10_bio->devs[1].devnum; wbio = r10_bio->devs[1].bio; wbio2 = r10_bio->devs[1].repl_bio; + /* Need to test wbio2->bi_end_io before we call + * generic_make_request as if the former is NULL, + * the latter is free to free wbio2. + */ + if (wbio2 && !wbio2->bi_end_io) + wbio2 = NULL; if (wbio->bi_end_io) { atomic_inc(&conf->mirrors[d].rdev->nr_pending); md_sync_acct(conf->mirrors[d].rdev->bdev, bio_sectors(wbio)); generic_make_request(wbio); } - if (wbio2 && wbio2->bi_end_io) { + if (wbio2) { atomic_inc(&conf->mirrors[d].replacement->nr_pending); md_sync_acct(conf->mirrors[d].replacement->bdev, bio_sectors(wbio2)); @@ -2579,8 +2600,8 @@ static int narrow_write_error(struct r10bio *r10_bio, int i) sectors = sect_to_write; /* Write at 'sector' for 'sectors' */ wbio = bio_clone_mddev(bio, GFP_NOIO, mddev); - md_trim_bio(wbio, sector - bio->bi_sector, sectors); - wbio->bi_sector = (r10_bio->devs[i].addr+ + bio_trim(wbio, sector - bio->bi_iter.bi_sector, sectors); + wbio->bi_iter.bi_sector = (r10_bio->devs[i].addr+ choose_data_offset(r10_bio, rdev) + (sector - r10_bio->sector)); wbio->bi_bdev = rdev->bdev; @@ -2652,12 +2673,10 @@ read_more: (unsigned long long)r10_bio->sector); bio = bio_clone_mddev(r10_bio->master_bio, GFP_NOIO, mddev); - md_trim_bio(bio, - r10_bio->sector - bio->bi_sector, - max_sectors); + bio_trim(bio, r10_bio->sector - bio->bi_iter.bi_sector, max_sectors); r10_bio->devs[slot].bio = bio; r10_bio->devs[slot].rdev = rdev; - bio->bi_sector = r10_bio->devs[slot].addr + bio->bi_iter.bi_sector = r10_bio->devs[slot].addr + choose_data_offset(r10_bio, rdev); bio->bi_bdev = rdev->bdev; bio->bi_rw = READ | do_sync; @@ -2668,7 +2687,7 @@ read_more: struct bio *mbio = r10_bio->master_bio; int sectors_handled = r10_bio->sector + max_sectors - - mbio->bi_sector; + - mbio->bi_iter.bi_sector; r10_bio->sectors = max_sectors; spin_lock_irq(&conf->device_lock); if (mbio->bi_phys_segments == 0) @@ -2686,7 +2705,7 @@ read_more: set_bit(R10BIO_ReadError, &r10_bio->state); r10_bio->mddev = mddev; - r10_bio->sector = mbio->bi_sector + r10_bio->sector = mbio->bi_iter.bi_sector + sectors_handled; goto read_more; @@ -2909,14 +2928,13 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, */ if (mddev->bitmap == NULL && mddev->recovery_cp == MaxSector && + mddev->reshape_position == MaxSector && + !test_bit(MD_RECOVERY_SYNC, &mddev->recovery) && !test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery) && + !test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) && conf->fullsync == 0) { *skipped = 1; - max_sector = mddev->dev_sectors; - if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery) || - test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery)) - max_sector = mddev->resync_max_sectors; - return max_sector - sector_nr; + return mddev->dev_sectors - sector_nr; } skipped: @@ -3125,7 +3143,8 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, bio->bi_end_io = end_sync_read; bio->bi_rw = READ; from_addr = r10_bio->devs[j].addr; - bio->bi_sector = from_addr + rdev->data_offset; + bio->bi_iter.bi_sector = from_addr + + rdev->data_offset; bio->bi_bdev = rdev->bdev; atomic_inc(&rdev->nr_pending); /* and we write to 'i' (if not in_sync) */ @@ -3149,7 +3168,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, bio->bi_private = r10_bio; bio->bi_end_io = end_sync_write; bio->bi_rw = WRITE; - bio->bi_sector = to_addr + bio->bi_iter.bi_sector = to_addr + rdev->data_offset; bio->bi_bdev = rdev->bdev; atomic_inc(&r10_bio->remaining); @@ -3178,7 +3197,8 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, bio->bi_private = r10_bio; bio->bi_end_io = end_sync_write; bio->bi_rw = WRITE; - bio->bi_sector = to_addr + rdev->data_offset; + bio->bi_iter.bi_sector = to_addr + + rdev->data_offset; bio->bi_bdev = rdev->bdev; atomic_inc(&r10_bio->remaining); break; @@ -3186,10 +3206,6 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, if (j == conf->copies) { /* Cannot recover, so abort the recovery or * record a bad block */ - put_buf(r10_bio); - if (rb2) - atomic_dec(&rb2->remaining); - r10_bio = rb2; if (any_working) { /* problem is that there are bad blocks * on other device(s) @@ -3221,6 +3237,10 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, mirror->recovery_disabled = mddev->recovery_disabled; } + put_buf(r10_bio); + if (rb2) + atomic_dec(&rb2->remaining); + r10_bio = rb2; break; } } @@ -3296,7 +3316,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, bio->bi_private = r10_bio; bio->bi_end_io = end_sync_read; bio->bi_rw = READ; - bio->bi_sector = sector + + bio->bi_iter.bi_sector = sector + conf->mirrors[d].rdev->data_offset; bio->bi_bdev = conf->mirrors[d].rdev->bdev; count++; @@ -3318,7 +3338,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, bio->bi_private = r10_bio; bio->bi_end_io = end_sync_write; bio->bi_rw = WRITE; - bio->bi_sector = sector + + bio->bi_iter.bi_sector = sector + conf->mirrors[d].replacement->data_offset; bio->bi_bdev = conf->mirrors[d].replacement->bdev; count++; @@ -3365,7 +3385,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, bio2 = bio2->bi_next) { /* remove last page from this bio */ bio2->bi_vcnt--; - bio2->bi_size -= len; + bio2->bi_iter.bi_size -= len; bio2->bi_flags &= ~(1<< BIO_SEG_VALID); } goto bio_full; @@ -3386,6 +3406,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, if (bio->bi_end_io == end_sync_read) { md_sync_acct(bio->bi_bdev, nr_sectors); + set_bit(BIO_UPTODATE, &bio->bi_flags); generic_make_request(bio); } } @@ -3532,7 +3553,7 @@ static struct r10conf *setup_conf(struct mddev *mddev) /* FIXME calc properly */ conf->mirrors = kzalloc(sizeof(struct raid10_info)*(mddev->raid_disks + - max(0,mddev->delta_disks)), + max(0,-mddev->delta_disks)), GFP_KERNEL); if (!conf->mirrors) goto out; @@ -3691,7 +3712,7 @@ static int run(struct mddev *mddev) conf->geo.far_offset == 0) goto out_free_conf; if (conf->prev.far_copies != 1 && - conf->geo.far_offset == 0) + conf->prev.far_offset == 0) goto out_free_conf; } @@ -3714,7 +3735,8 @@ static int run(struct mddev *mddev) !test_bit(In_sync, &disk->rdev->flags)) { disk->head_position = 0; mddev->degraded++; - if (disk->rdev) + if (disk->rdev && + disk->rdev->saved_raid_disk < 0) conf->fullsync = 1; } disk->recovery_disabled = mddev->recovery_disabled - 1; @@ -4351,7 +4373,11 @@ static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr, set_bit(MD_CHANGE_DEVS, &mddev->flags); md_wakeup_thread(mddev->thread); wait_event(mddev->sb_wait, mddev->flags == 0 || - kthread_should_stop()); + test_bit(MD_RECOVERY_INTR, &mddev->recovery)); + if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) { + allow_barrier(conf); + return sectors_done; + } conf->reshape_safe = mddev->reshape_position; allow_barrier(conf); } @@ -4380,7 +4406,7 @@ read_more: read_bio = bio_alloc_mddev(GFP_KERNEL, RESYNC_PAGES, mddev); read_bio->bi_bdev = rdev->bdev; - read_bio->bi_sector = (r10_bio->devs[r10_bio->read_slot].addr + read_bio->bi_iter.bi_sector = (r10_bio->devs[r10_bio->read_slot].addr + rdev->data_offset); read_bio->bi_private = r10_bio; read_bio->bi_end_io = end_sync_read; @@ -4388,7 +4414,7 @@ read_more: read_bio->bi_flags &= ~(BIO_POOL_MASK - 1); read_bio->bi_flags |= 1 << BIO_UPTODATE; read_bio->bi_vcnt = 0; - read_bio->bi_size = 0; + read_bio->bi_iter.bi_size = 0; r10_bio->master_bio = read_bio; r10_bio->read_slot = r10_bio->devs[r10_bio->read_slot].devnum; @@ -4414,7 +4440,8 @@ read_more: bio_reset(b); b->bi_bdev = rdev2->bdev; - b->bi_sector = r10_bio->devs[s/2].addr + rdev2->new_data_offset; + b->bi_iter.bi_sector = r10_bio->devs[s/2].addr + + rdev2->new_data_offset; b->bi_private = r10_bio; b->bi_end_io = end_reshape_write; b->bi_rw = WRITE; @@ -4441,7 +4468,7 @@ read_more: bio2 = bio2->bi_next) { /* Remove last page from this bio */ bio2->bi_vcnt--; - bio2->bi_size -= len; + bio2->bi_iter.bi_size -= len; bio2->bi_flags &= ~(1<<BIO_SEG_VALID); } goto bio_full; diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 05e4a105b9c..6234b2e8458 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -53,6 +53,7 @@ #include <linux/cpu.h> #include <linux/slab.h> #include <linux/ratelimit.h> +#include <linux/nodemask.h> #include <trace/events/block.h> #include "md.h" @@ -60,6 +61,10 @@ #include "raid0.h" #include "bitmap.h" +#define cpu_to_group(cpu) cpu_to_node(cpu) +#define ANY_GROUP NUMA_NO_NODE + +static struct workqueue_struct *raid5_wq; /* * Stripe cache */ @@ -72,6 +77,7 @@ #define BYPASS_THRESHOLD 1 #define NR_HASH (PAGE_SIZE / sizeof(struct hlist_head)) #define HASH_MASK (NR_HASH - 1) +#define MAX_STRIPE_BATCH 8 static inline struct hlist_head *stripe_hash(struct r5conf *conf, sector_t sect) { @@ -79,6 +85,42 @@ static inline struct hlist_head *stripe_hash(struct r5conf *conf, sector_t sect) return &conf->stripe_hashtbl[hash]; } +static inline int stripe_hash_locks_hash(sector_t sect) +{ + return (sect >> STRIPE_SHIFT) & STRIPE_HASH_LOCKS_MASK; +} + +static inline void lock_device_hash_lock(struct r5conf *conf, int hash) +{ + spin_lock_irq(conf->hash_locks + hash); + spin_lock(&conf->device_lock); +} + +static inline void unlock_device_hash_lock(struct r5conf *conf, int hash) +{ + spin_unlock(&conf->device_lock); + spin_unlock_irq(conf->hash_locks + hash); +} + +static inline void lock_all_device_hash_locks_irq(struct r5conf *conf) +{ + int i; + local_irq_disable(); + spin_lock(conf->hash_locks); + for (i = 1; i < NR_STRIPE_HASH_LOCKS; i++) + spin_lock_nest_lock(conf->hash_locks + i, conf->hash_locks); + spin_lock(&conf->device_lock); +} + +static inline void unlock_all_device_hash_locks_irq(struct r5conf *conf) +{ + int i; + spin_unlock(&conf->device_lock); + for (i = NR_STRIPE_HASH_LOCKS; i; i--) + spin_unlock(conf->hash_locks + i - 1); + local_irq_enable(); +} + /* bio's attached to a stripe+device for I/O are linked together in bi_sector * order without overlap. There may be several bio's per stripe+device, and * a bio could span several devices. @@ -91,7 +133,7 @@ static inline struct hlist_head *stripe_hash(struct r5conf *conf, sector_t sect) static inline struct bio *r5_next_bio(struct bio *bio, sector_t sector) { int sectors = bio_sectors(bio); - if (bio->bi_sector + sectors < sector + STRIPE_SECTORS) + if (bio->bi_iter.bi_sector + sectors < sector + STRIPE_SECTORS) return bio->bi_next; else return NULL; @@ -183,7 +225,7 @@ static void return_io(struct bio *return_bi) return_bi = bi->bi_next; bi->bi_next = NULL; - bi->bi_size = 0; + bi->bi_iter.bi_size = 0; trace_block_bio_complete(bdev_get_queue(bi->bi_bdev), bi, 0); bio_endio(bi, 0); @@ -200,21 +242,73 @@ static int stripe_operations_active(struct stripe_head *sh) test_bit(STRIPE_COMPUTE_RUN, &sh->state); } -static void do_release_stripe(struct r5conf *conf, struct stripe_head *sh) +static void raid5_wakeup_stripe_thread(struct stripe_head *sh) +{ + struct r5conf *conf = sh->raid_conf; + struct r5worker_group *group; + int thread_cnt; + int i, cpu = sh->cpu; + + if (!cpu_online(cpu)) { + cpu = cpumask_any(cpu_online_mask); + sh->cpu = cpu; + } + + if (list_empty(&sh->lru)) { + struct r5worker_group *group; + group = conf->worker_groups + cpu_to_group(cpu); + list_add_tail(&sh->lru, &group->handle_list); + group->stripes_cnt++; + sh->group = group; + } + + if (conf->worker_cnt_per_group == 0) { + md_wakeup_thread(conf->mddev->thread); + return; + } + + group = conf->worker_groups + cpu_to_group(sh->cpu); + + group->workers[0].working = true; + /* at least one worker should run to avoid race */ + queue_work_on(sh->cpu, raid5_wq, &group->workers[0].work); + + thread_cnt = group->stripes_cnt / MAX_STRIPE_BATCH - 1; + /* wakeup more workers */ + for (i = 1; i < conf->worker_cnt_per_group && thread_cnt > 0; i++) { + if (group->workers[i].working == false) { + group->workers[i].working = true; + queue_work_on(sh->cpu, raid5_wq, + &group->workers[i].work); + thread_cnt--; + } + } +} + +static void do_release_stripe(struct r5conf *conf, struct stripe_head *sh, + struct list_head *temp_inactive_list) { BUG_ON(!list_empty(&sh->lru)); BUG_ON(atomic_read(&conf->active_stripes)==0); if (test_bit(STRIPE_HANDLE, &sh->state)) { if (test_bit(STRIPE_DELAYED, &sh->state) && - !test_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) + !test_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) { list_add_tail(&sh->lru, &conf->delayed_list); - else if (test_bit(STRIPE_BIT_DELAY, &sh->state) && + if (atomic_read(&conf->preread_active_stripes) + < IO_THRESHOLD) + md_wakeup_thread(conf->mddev->thread); + } else if (test_bit(STRIPE_BIT_DELAY, &sh->state) && sh->bm_seq - conf->seq_write > 0) list_add_tail(&sh->lru, &conf->bitmap_list); else { clear_bit(STRIPE_DELAYED, &sh->state); clear_bit(STRIPE_BIT_DELAY, &sh->state); - list_add_tail(&sh->lru, &conf->handle_list); + if (conf->worker_cnt_per_group == 0) { + list_add_tail(&sh->lru, &conf->handle_list); + } else { + raid5_wakeup_stripe_thread(sh); + return; + } } md_wakeup_thread(conf->mddev->thread); } else { @@ -224,30 +318,125 @@ static void do_release_stripe(struct r5conf *conf, struct stripe_head *sh) < IO_THRESHOLD) md_wakeup_thread(conf->mddev->thread); atomic_dec(&conf->active_stripes); - if (!test_bit(STRIPE_EXPANDING, &sh->state)) { - list_add_tail(&sh->lru, &conf->inactive_list); - wake_up(&conf->wait_for_stripe); - if (conf->retry_read_aligned) - md_wakeup_thread(conf->mddev->thread); - } + if (!test_bit(STRIPE_EXPANDING, &sh->state)) + list_add_tail(&sh->lru, temp_inactive_list); } } -static void __release_stripe(struct r5conf *conf, struct stripe_head *sh) +static void __release_stripe(struct r5conf *conf, struct stripe_head *sh, + struct list_head *temp_inactive_list) { if (atomic_dec_and_test(&sh->count)) - do_release_stripe(conf, sh); + do_release_stripe(conf, sh, temp_inactive_list); +} + +/* + * @hash could be NR_STRIPE_HASH_LOCKS, then we have a list of inactive_list + * + * Be careful: Only one task can add/delete stripes from temp_inactive_list at + * given time. Adding stripes only takes device lock, while deleting stripes + * only takes hash lock. + */ +static void release_inactive_stripe_list(struct r5conf *conf, + struct list_head *temp_inactive_list, + int hash) +{ + int size; + bool do_wakeup = false; + unsigned long flags; + + if (hash == NR_STRIPE_HASH_LOCKS) { + size = NR_STRIPE_HASH_LOCKS; + hash = NR_STRIPE_HASH_LOCKS - 1; + } else + size = 1; + while (size) { + struct list_head *list = &temp_inactive_list[size - 1]; + + /* + * We don't hold any lock here yet, get_active_stripe() might + * remove stripes from the list + */ + if (!list_empty_careful(list)) { + spin_lock_irqsave(conf->hash_locks + hash, flags); + if (list_empty(conf->inactive_list + hash) && + !list_empty(list)) + atomic_dec(&conf->empty_inactive_list_nr); + list_splice_tail_init(list, conf->inactive_list + hash); + do_wakeup = true; + spin_unlock_irqrestore(conf->hash_locks + hash, flags); + } + size--; + hash--; + } + + if (do_wakeup) { + wake_up(&conf->wait_for_stripe); + if (conf->retry_read_aligned) + md_wakeup_thread(conf->mddev->thread); + } +} + +/* should hold conf->device_lock already */ +static int release_stripe_list(struct r5conf *conf, + struct list_head *temp_inactive_list) +{ + struct stripe_head *sh; + int count = 0; + struct llist_node *head; + + head = llist_del_all(&conf->released_stripes); + head = llist_reverse_order(head); + while (head) { + int hash; + + sh = llist_entry(head, struct stripe_head, release_list); + head = llist_next(head); + /* sh could be readded after STRIPE_ON_RELEASE_LIST is cleard */ + smp_mb(); + clear_bit(STRIPE_ON_RELEASE_LIST, &sh->state); + /* + * Don't worry the bit is set here, because if the bit is set + * again, the count is always > 1. This is true for + * STRIPE_ON_UNPLUG_LIST bit too. + */ + hash = sh->hash_lock_index; + __release_stripe(conf, sh, &temp_inactive_list[hash]); + count++; + } + + return count; } static void release_stripe(struct stripe_head *sh) { struct r5conf *conf = sh->raid_conf; unsigned long flags; + struct list_head list; + int hash; + bool wakeup; + /* Avoid release_list until the last reference. + */ + if (atomic_add_unless(&sh->count, -1, 1)) + return; + + if (unlikely(!conf->mddev->thread) || + test_and_set_bit(STRIPE_ON_RELEASE_LIST, &sh->state)) + goto slow_path; + wakeup = llist_add(&sh->release_list, &conf->released_stripes); + if (wakeup) + md_wakeup_thread(conf->mddev->thread); + return; +slow_path: local_irq_save(flags); + /* we are ok here if STRIPE_ON_RELEASE_LIST is set or not */ if (atomic_dec_and_lock(&sh->count, &conf->device_lock)) { - do_release_stripe(conf, sh); + INIT_LIST_HEAD(&list); + hash = sh->hash_lock_index; + do_release_stripe(conf, sh, &list); spin_unlock(&conf->device_lock); + release_inactive_stripe_list(conf, &list, hash); } local_irq_restore(flags); } @@ -272,18 +461,21 @@ static inline void insert_hash(struct r5conf *conf, struct stripe_head *sh) /* find an idle stripe, make sure it is unhashed, and return it. */ -static struct stripe_head *get_free_stripe(struct r5conf *conf) +static struct stripe_head *get_free_stripe(struct r5conf *conf, int hash) { struct stripe_head *sh = NULL; struct list_head *first; - if (list_empty(&conf->inactive_list)) + if (list_empty(conf->inactive_list + hash)) goto out; - first = conf->inactive_list.next; + first = (conf->inactive_list + hash)->next; sh = list_entry(first, struct stripe_head, lru); list_del_init(first); remove_hash(sh); atomic_inc(&conf->active_stripes); + BUG_ON(hash != sh->hash_lock_index); + if (list_empty(conf->inactive_list + hash)) + atomic_inc(&conf->empty_inactive_list_nr); out: return sh; } @@ -295,6 +487,7 @@ static void shrink_buffers(struct stripe_head *sh) int num = sh->raid_conf->pool_size; for (i = 0; i < num ; i++) { + WARN_ON(sh->dev[i].page != sh->dev[i].orig_page); p = sh->dev[i].page; if (!p) continue; @@ -315,6 +508,7 @@ static int grow_buffers(struct stripe_head *sh) return 1; } sh->dev[i].page = page; + sh->dev[i].orig_page = page; } return 0; } @@ -326,7 +520,7 @@ static void stripe_set_idx(sector_t stripe, struct r5conf *conf, int previous, static void init_stripe(struct stripe_head *sh, sector_t sector, int previous) { struct r5conf *conf = sh->raid_conf; - int i; + int i, seq; BUG_ON(atomic_read(&sh->count) != 0); BUG_ON(test_bit(STRIPE_HANDLE, &sh->state)); @@ -336,7 +530,8 @@ static void init_stripe(struct stripe_head *sh, sector_t sector, int previous) (unsigned long long)sh->sector); remove_hash(sh); - +retry: + seq = read_seqcount_begin(&conf->gen_lock); sh->generation = conf->generation - previous; sh->disks = previous ? conf->previous_raid_disks : conf->raid_disks; sh->sector = sector; @@ -358,7 +553,10 @@ static void init_stripe(struct stripe_head *sh, sector_t sector, int previous) dev->flags = 0; raid5_build_block(sh, i, previous); } + if (read_seqcount_retry(&conf->gen_lock, seq)) + goto retry; insert_hash(conf, sh); + sh->cpu = smp_processor_id(); } static struct stripe_head *__find_stripe(struct r5conf *conf, sector_t sector, @@ -461,52 +659,55 @@ get_active_stripe(struct r5conf *conf, sector_t sector, int previous, int noblock, int noquiesce) { struct stripe_head *sh; + int hash = stripe_hash_locks_hash(sector); pr_debug("get_stripe, sector %llu\n", (unsigned long long)sector); - spin_lock_irq(&conf->device_lock); + spin_lock_irq(conf->hash_locks + hash); do { wait_event_lock_irq(conf->wait_for_stripe, conf->quiesce == 0 || noquiesce, - conf->device_lock); + *(conf->hash_locks + hash)); sh = __find_stripe(conf, sector, conf->generation - previous); if (!sh) { if (!conf->inactive_blocked) - sh = get_free_stripe(conf); + sh = get_free_stripe(conf, hash); if (noblock && sh == NULL) break; if (!sh) { conf->inactive_blocked = 1; - wait_event_lock_irq(conf->wait_for_stripe, - !list_empty(&conf->inactive_list) && - (atomic_read(&conf->active_stripes) - < (conf->max_nr_stripes *3/4) - || !conf->inactive_blocked), - conf->device_lock); + wait_event_lock_irq( + conf->wait_for_stripe, + !list_empty(conf->inactive_list + hash) && + (atomic_read(&conf->active_stripes) + < (conf->max_nr_stripes * 3 / 4) + || !conf->inactive_blocked), + *(conf->hash_locks + hash)); conf->inactive_blocked = 0; - } else - init_stripe(sh, sector, previous); - } else { - if (atomic_read(&sh->count)) { - BUG_ON(!list_empty(&sh->lru) - && !test_bit(STRIPE_EXPANDING, &sh->state) - && !test_bit(STRIPE_ON_UNPLUG_LIST, &sh->state)); } else { + init_stripe(sh, sector, previous); + atomic_inc(&sh->count); + } + } else if (!atomic_inc_not_zero(&sh->count)) { + spin_lock(&conf->device_lock); + if (!atomic_read(&sh->count)) { if (!test_bit(STRIPE_HANDLE, &sh->state)) atomic_inc(&conf->active_stripes); - if (list_empty(&sh->lru) && - !test_bit(STRIPE_EXPANDING, &sh->state)) - BUG(); + BUG_ON(list_empty(&sh->lru) && + !test_bit(STRIPE_EXPANDING, &sh->state)); list_del_init(&sh->lru); + if (sh->group) { + sh->group->stripes_cnt--; + sh->group = NULL; + } } + atomic_inc(&sh->count); + spin_unlock(&conf->device_lock); } } while (sh == NULL); - if (sh) - atomic_inc(&sh->count); - - spin_unlock_irq(&conf->device_lock); + spin_unlock_irq(conf->hash_locks + hash); return sh; } @@ -656,18 +857,27 @@ static void ops_run_io(struct stripe_head *sh, struct stripe_head_state *s) bi->bi_rw, i); atomic_inc(&sh->count); if (use_new_offset(conf, sh)) - bi->bi_sector = (sh->sector + bi->bi_iter.bi_sector = (sh->sector + rdev->new_data_offset); else - bi->bi_sector = (sh->sector + bi->bi_iter.bi_sector = (sh->sector + rdev->data_offset); if (test_bit(R5_ReadNoMerge, &sh->dev[i].flags)) - bi->bi_rw |= REQ_FLUSH; + bi->bi_rw |= REQ_NOMERGE; + if (test_bit(R5_SkipCopy, &sh->dev[i].flags)) + WARN_ON(test_bit(R5_UPTODATE, &sh->dev[i].flags)); + sh->dev[i].vec.bv_page = sh->dev[i].page; bi->bi_vcnt = 1; bi->bi_io_vec[0].bv_len = STRIPE_SIZE; bi->bi_io_vec[0].bv_offset = 0; - bi->bi_size = STRIPE_SIZE; + bi->bi_iter.bi_size = STRIPE_SIZE; + /* + * If this is discard request, set bi_vcnt 0. We don't + * want to confuse SCSI because SCSI will replace payload + */ + if (rw & REQ_DISCARD) + bi->bi_vcnt = 0; if (rrdev) set_bit(R5_DOUBLE_LOCKED, &sh->dev[i].flags); @@ -697,15 +907,24 @@ static void ops_run_io(struct stripe_head *sh, struct stripe_head_state *s) rbi->bi_rw, i); atomic_inc(&sh->count); if (use_new_offset(conf, sh)) - rbi->bi_sector = (sh->sector + rbi->bi_iter.bi_sector = (sh->sector + rrdev->new_data_offset); else - rbi->bi_sector = (sh->sector + rbi->bi_iter.bi_sector = (sh->sector + rrdev->data_offset); + if (test_bit(R5_SkipCopy, &sh->dev[i].flags)) + WARN_ON(test_bit(R5_UPTODATE, &sh->dev[i].flags)); + sh->dev[i].rvec.bv_page = sh->dev[i].page; rbi->bi_vcnt = 1; rbi->bi_io_vec[0].bv_len = STRIPE_SIZE; rbi->bi_io_vec[0].bv_offset = 0; - rbi->bi_size = STRIPE_SIZE; + rbi->bi_iter.bi_size = STRIPE_SIZE; + /* + * If this is discard request, set bi_vcnt 0. We don't + * want to confuse SCSI because SCSI will replace payload + */ + if (rw & REQ_DISCARD) + rbi->bi_vcnt = 0; if (conf->mddev->gendisk) trace_block_bio_remap(bdev_get_queue(rbi->bi_bdev), rbi, disk_devt(conf->mddev->gendisk), @@ -724,27 +943,28 @@ static void ops_run_io(struct stripe_head *sh, struct stripe_head_state *s) } static struct dma_async_tx_descriptor * -async_copy_data(int frombio, struct bio *bio, struct page *page, - sector_t sector, struct dma_async_tx_descriptor *tx) +async_copy_data(int frombio, struct bio *bio, struct page **page, + sector_t sector, struct dma_async_tx_descriptor *tx, + struct stripe_head *sh) { - struct bio_vec *bvl; + struct bio_vec bvl; + struct bvec_iter iter; struct page *bio_page; - int i; int page_offset; struct async_submit_ctl submit; enum async_tx_flags flags = 0; - if (bio->bi_sector >= sector) - page_offset = (signed)(bio->bi_sector - sector) * 512; + if (bio->bi_iter.bi_sector >= sector) + page_offset = (signed)(bio->bi_iter.bi_sector - sector) * 512; else - page_offset = (signed)(sector - bio->bi_sector) * -512; + page_offset = (signed)(sector - bio->bi_iter.bi_sector) * -512; if (frombio) flags |= ASYNC_TX_FENCE; init_async_submit(&submit, flags, tx, NULL, NULL, NULL); - bio_for_each_segment(bvl, bio, i) { - int len = bvl->bv_len; + bio_for_each_segment(bvl, bio, iter) { + int len = bvl.bv_len; int clen; int b_offset = 0; @@ -760,13 +980,18 @@ async_copy_data(int frombio, struct bio *bio, struct page *page, clen = len; if (clen > 0) { - b_offset += bvl->bv_offset; - bio_page = bvl->bv_page; - if (frombio) - tx = async_memcpy(page, bio_page, page_offset, + b_offset += bvl.bv_offset; + bio_page = bvl.bv_page; + if (frombio) { + if (sh->raid_conf->skip_copy && + b_offset == 0 && page_offset == 0 && + clen == STRIPE_SIZE) + *page = bio_page; + else + tx = async_memcpy(*page, bio_page, page_offset, b_offset, clen, &submit); - else - tx = async_memcpy(bio_page, page, b_offset, + } else + tx = async_memcpy(bio_page, *page, b_offset, page_offset, clen, &submit); } /* chain the operations */ @@ -804,7 +1029,7 @@ static void ops_complete_biofill(void *stripe_head_ref) BUG_ON(!dev->read); rbi = dev->read; dev->read = NULL; - while (rbi && rbi->bi_sector < + while (rbi && rbi->bi_iter.bi_sector < dev->sector + STRIPE_SECTORS) { rbi2 = r5_next_bio(rbi, dev->sector); if (!raid5_dec_bi_active_stripes(rbi)) { @@ -840,10 +1065,10 @@ static void ops_run_biofill(struct stripe_head *sh) dev->read = rbi = dev->toread; dev->toread = NULL; spin_unlock_irq(&sh->stripe_lock); - while (rbi && rbi->bi_sector < + while (rbi && rbi->bi_iter.bi_sector < dev->sector + STRIPE_SECTORS) { - tx = async_copy_data(0, rbi, dev->page, - dev->sector, tx); + tx = async_copy_data(0, rbi, &dev->page, + dev->sector, tx, sh); rbi = r5_next_bio(rbi, dev->sector); } } @@ -1181,8 +1406,9 @@ ops_run_biodrain(struct stripe_head *sh, struct dma_async_tx_descriptor *tx) BUG_ON(dev->written); wbi = dev->written = chosen; spin_unlock_irq(&sh->stripe_lock); + WARN_ON(dev->page != dev->orig_page); - while (wbi && wbi->bi_sector < + while (wbi && wbi->bi_iter.bi_sector < dev->sector + STRIPE_SECTORS) { if (wbi->bi_rw & REQ_FUA) set_bit(R5_WantFUA, &dev->flags); @@ -1190,9 +1416,15 @@ ops_run_biodrain(struct stripe_head *sh, struct dma_async_tx_descriptor *tx) set_bit(R5_SyncIO, &dev->flags); if (wbi->bi_rw & REQ_DISCARD) set_bit(R5_Discard, &dev->flags); - else - tx = async_copy_data(1, wbi, dev->page, - dev->sector, tx); + else { + tx = async_copy_data(1, wbi, &dev->page, + dev->sector, tx, sh); + if (dev->page != dev->orig_page) { + set_bit(R5_SkipCopy, &dev->flags); + clear_bit(R5_UPTODATE, &dev->flags); + clear_bit(R5_OVERWRITE, &dev->flags); + } + } wbi = r5_next_bio(wbi, dev->sector); } } @@ -1223,7 +1455,7 @@ static void ops_complete_reconstruct(void *stripe_head_ref) struct r5dev *dev = &sh->dev[i]; if (dev->written || i == pd_idx || i == qd_idx) { - if (!discard) + if (!discard && !test_bit(R5_SkipCopy, &dev->flags)) set_bit(R5_UPTODATE, &dev->flags); if (fua) set_bit(R5_WantFUA, &dev->flags); @@ -1474,7 +1706,7 @@ static void raid_run_ops(struct stripe_head *sh, unsigned long ops_request) put_cpu(); } -static int grow_one_stripe(struct r5conf *conf) +static int grow_one_stripe(struct r5conf *conf, int hash) { struct stripe_head *sh; sh = kmem_cache_zalloc(conf->slab_cache, GFP_KERNEL); @@ -1490,6 +1722,7 @@ static int grow_one_stripe(struct r5conf *conf) kmem_cache_free(conf->slab_cache, sh); return 0; } + sh->hash_lock_index = hash; /* we just created an active stripe so... */ atomic_set(&sh->count, 1); atomic_inc(&conf->active_stripes); @@ -1502,6 +1735,7 @@ static int grow_stripes(struct r5conf *conf, int num) { struct kmem_cache *sc; int devs = max(conf->raid_disks, conf->previous_raid_disks); + int hash; if (conf->mddev->gendisk) sprintf(conf->cache_name[0], @@ -1519,9 +1753,13 @@ static int grow_stripes(struct r5conf *conf, int num) return 1; conf->slab_cache = sc; conf->pool_size = devs; - while (num--) - if (!grow_one_stripe(conf)) + hash = conf->max_nr_stripes % NR_STRIPE_HASH_LOCKS; + while (num--) { + if (!grow_one_stripe(conf, hash)) return 1; + conf->max_nr_stripes++; + hash = (hash + 1) % NR_STRIPE_HASH_LOCKS; + } return 0; } @@ -1579,6 +1817,7 @@ static int resize_stripes(struct r5conf *conf, int newsize) int err; struct kmem_cache *sc; int i; + int hash, cnt; if (newsize <= conf->pool_size) return 0; /* never bother to shrink */ @@ -1618,19 +1857,31 @@ static int resize_stripes(struct r5conf *conf, int newsize) * OK, we have enough stripes, start collecting inactive * stripes and copying them over */ + hash = 0; + cnt = 0; list_for_each_entry(nsh, &newstripes, lru) { - spin_lock_irq(&conf->device_lock); - wait_event_lock_irq(conf->wait_for_stripe, - !list_empty(&conf->inactive_list), - conf->device_lock); - osh = get_free_stripe(conf); - spin_unlock_irq(&conf->device_lock); + lock_device_hash_lock(conf, hash); + wait_event_cmd(conf->wait_for_stripe, + !list_empty(conf->inactive_list + hash), + unlock_device_hash_lock(conf, hash), + lock_device_hash_lock(conf, hash)); + osh = get_free_stripe(conf, hash); + unlock_device_hash_lock(conf, hash); atomic_set(&nsh->count, 1); - for(i=0; i<conf->pool_size; i++) + for(i=0; i<conf->pool_size; i++) { nsh->dev[i].page = osh->dev[i].page; + nsh->dev[i].orig_page = osh->dev[i].page; + } for( ; i<newsize; i++) nsh->dev[i].page = NULL; + nsh->hash_lock_index = hash; kmem_cache_free(conf->slab_cache, osh); + cnt++; + if (cnt >= conf->max_nr_stripes / NR_STRIPE_HASH_LOCKS + + !!((conf->max_nr_stripes % NR_STRIPE_HASH_LOCKS) > hash)) { + hash++; + cnt = 0; + } } kmem_cache_destroy(conf->slab_cache); @@ -1676,6 +1927,7 @@ static int resize_stripes(struct r5conf *conf, int newsize) if (nsh->dev[i].page == NULL) { struct page *p = alloc_page(GFP_NOIO); nsh->dev[i].page = p; + nsh->dev[i].orig_page = p; if (!p) err = -ENOMEM; } @@ -1689,13 +1941,13 @@ static int resize_stripes(struct r5conf *conf, int newsize) return err; } -static int drop_one_stripe(struct r5conf *conf) +static int drop_one_stripe(struct r5conf *conf, int hash) { struct stripe_head *sh; - spin_lock_irq(&conf->device_lock); - sh = get_free_stripe(conf); - spin_unlock_irq(&conf->device_lock); + spin_lock_irq(conf->hash_locks + hash); + sh = get_free_stripe(conf, hash); + spin_unlock_irq(conf->hash_locks + hash); if (!sh) return 0; BUG_ON(atomic_read(&sh->count)); @@ -1707,8 +1959,10 @@ static int drop_one_stripe(struct r5conf *conf) static void shrink_stripes(struct r5conf *conf) { - while (drop_one_stripe(conf)) - ; + int hash; + for (hash = 0; hash < NR_STRIPE_HASH_LOCKS; hash++) + while (drop_one_stripe(conf, hash)) + ; if (conf->slab_cache) kmem_cache_destroy(conf->slab_cache); @@ -1813,6 +2067,9 @@ static void raid5_end_read_request(struct bio * bi, int error) mdname(conf->mddev), bdn); else retry = 1; + if (set_bad && test_bit(In_sync, &rdev->flags) + && !test_bit(R5_ReadNoMerge, &sh->dev[i].flags)) + retry = 1; if (retry) if (test_bit(R5_ReadNoMerge, &sh->dev[i].flags)) { set_bit(R5_ReadError, &sh->dev[i].flags); @@ -1881,6 +2138,7 @@ static void raid5_end_write_request(struct bio *bi, int error) set_bit(R5_MadeGoodRepl, &sh->dev[i].flags); } else { if (!uptodate) { + set_bit(STRIPE_DEGRADED, &sh->state); set_bit(WriteErrorSeen, &rdev->flags); set_bit(R5_WriteError, &sh->dev[i].flags); if (!test_and_set_bit(WantReplacement, &rdev->flags)) @@ -1907,24 +2165,20 @@ static void raid5_end_write_request(struct bio *bi, int error) } static sector_t compute_blocknr(struct stripe_head *sh, int i, int previous); - + static void raid5_build_block(struct stripe_head *sh, int i, int previous) { struct r5dev *dev = &sh->dev[i]; bio_init(&dev->req); dev->req.bi_io_vec = &dev->vec; - dev->req.bi_vcnt++; - dev->req.bi_max_vecs++; + dev->req.bi_max_vecs = 1; dev->req.bi_private = sh; - dev->vec.bv_page = dev->page; bio_init(&dev->rreq); dev->rreq.bi_io_vec = &dev->rvec; - dev->rreq.bi_vcnt++; - dev->rreq.bi_max_vecs++; + dev->rreq.bi_max_vecs = 1; dev->rreq.bi_private = sh; - dev->rvec.bv_page = dev->page; dev->flags = 0; dev->sector = compute_blocknr(sh, i, previous); @@ -2384,7 +2638,7 @@ static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, in int firstwrite=0; pr_debug("adding bi b#%llu to stripe s#%llu\n", - (unsigned long long)bi->bi_sector, + (unsigned long long)bi->bi_iter.bi_sector, (unsigned long long)sh->sector); /* @@ -2402,12 +2656,12 @@ static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, in firstwrite = 1; } else bip = &sh->dev[dd_idx].toread; - while (*bip && (*bip)->bi_sector < bi->bi_sector) { - if (bio_end_sector(*bip) > bi->bi_sector) + while (*bip && (*bip)->bi_iter.bi_sector < bi->bi_iter.bi_sector) { + if (bio_end_sector(*bip) > bi->bi_iter.bi_sector) goto overlap; bip = & (*bip)->bi_next; } - if (*bip && (*bip)->bi_sector < bio_end_sector(bi)) + if (*bip && (*bip)->bi_iter.bi_sector < bio_end_sector(bi)) goto overlap; BUG_ON(*bip && bi->bi_next && (*bip) != bi->bi_next); @@ -2421,7 +2675,7 @@ static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, in sector_t sector = sh->dev[dd_idx].sector; for (bi=sh->dev[dd_idx].towrite; sector < sh->dev[dd_idx].sector + STRIPE_SECTORS && - bi && bi->bi_sector <= sector; + bi && bi->bi_iter.bi_sector <= sector; bi = r5_next_bio(bi, sh->dev[dd_idx].sector)) { if (bio_end_sector(bi) >= sector) sector = bio_end_sector(bi); @@ -2431,7 +2685,7 @@ static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, in } pr_debug("added bi b#%llu to stripe s#%llu, disk %d.\n", - (unsigned long long)(*bip)->bi_sector, + (unsigned long long)(*bip)->bi_iter.bi_sector, (unsigned long long)sh->sector, dd_idx); spin_unlock_irq(&sh->stripe_lock); @@ -2506,7 +2760,7 @@ handle_failed_stripe(struct r5conf *conf, struct stripe_head *sh, if (test_and_clear_bit(R5_Overlap, &sh->dev[i].flags)) wake_up(&conf->wait_for_overlap); - while (bi && bi->bi_sector < + while (bi && bi->bi_iter.bi_sector < sh->dev[i].sector + STRIPE_SECTORS) { struct bio *nextbi = r5_next_bio(bi, sh->dev[i].sector); clear_bit(BIO_UPTODATE, &bi->bi_flags); @@ -2524,8 +2778,13 @@ handle_failed_stripe(struct r5conf *conf, struct stripe_head *sh, /* and fail all 'written' */ bi = sh->dev[i].written; sh->dev[i].written = NULL; + if (test_and_clear_bit(R5_SkipCopy, &sh->dev[i].flags)) { + WARN_ON(test_bit(R5_UPTODATE, &sh->dev[i].flags)); + sh->dev[i].page = sh->dev[i].orig_page; + } + if (bi) bitmap_end = 1; - while (bi && bi->bi_sector < + while (bi && bi->bi_iter.bi_sector < sh->dev[i].sector + STRIPE_SECTORS) { struct bio *bi2 = r5_next_bio(bi, sh->dev[i].sector); clear_bit(BIO_UPTODATE, &bi->bi_flags); @@ -2549,7 +2808,7 @@ handle_failed_stripe(struct r5conf *conf, struct stripe_head *sh, spin_unlock_irq(&sh->stripe_lock); if (test_and_clear_bit(R5_Overlap, &sh->dev[i].flags)) wake_up(&conf->wait_for_overlap); - while (bi && bi->bi_sector < + while (bi && bi->bi_iter.bi_sector < sh->dev[i].sector + STRIPE_SECTORS) { struct bio *nextbi = r5_next_bio(bi, sh->dev[i].sector); @@ -2660,8 +2919,11 @@ static int fetch_block(struct stripe_head *sh, struct stripe_head_state *s, (s->failed >= 1 && fdev[0]->toread) || (s->failed >= 2 && fdev[1]->toread) || (sh->raid_conf->level <= 5 && s->failed && fdev[0]->towrite && + (!test_bit(R5_Insync, &dev->flags) || test_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) && !test_bit(R5_OVERWRITE, &fdev[0]->flags)) || - (sh->raid_conf->level == 6 && s->failed && s->to_write))) { + (sh->raid_conf->level == 6 && s->failed && s->to_write && + s->to_write < sh->raid_conf->raid_disks - 2 && + (!test_bit(R5_Insync, &dev->flags) || test_bit(STRIPE_PREREAD_ACTIVE, &sh->state))))) { /* we would like to get this block, possibly by computing it, * otherwise read it if the backing disk is insync */ @@ -2765,15 +3027,20 @@ static void handle_stripe_clean_event(struct r5conf *conf, dev = &sh->dev[i]; if (!test_bit(R5_LOCKED, &dev->flags) && (test_bit(R5_UPTODATE, &dev->flags) || - test_bit(R5_Discard, &dev->flags))) { + test_bit(R5_Discard, &dev->flags) || + test_bit(R5_SkipCopy, &dev->flags))) { /* We can return any write requests */ struct bio *wbi, *wbi2; pr_debug("Return write for disc %d\n", i); if (test_and_clear_bit(R5_Discard, &dev->flags)) clear_bit(R5_UPTODATE, &dev->flags); + if (test_and_clear_bit(R5_SkipCopy, &dev->flags)) { + WARN_ON(test_bit(R5_UPTODATE, &dev->flags)); + dev->page = dev->orig_page; + } wbi = dev->written; dev->written = NULL; - while (wbi && wbi->bi_sector < + while (wbi && wbi->bi_iter.bi_sector < dev->sector + STRIPE_SECTORS) { wbi2 = r5_next_bio(wbi, dev->sector); if (!raid5_dec_bi_active_stripes(wbi)) { @@ -2789,6 +3056,8 @@ static void handle_stripe_clean_event(struct r5conf *conf, 0); } else if (test_bit(R5_Discard, &dev->flags)) discard_pending = 1; + WARN_ON(test_bit(R5_SkipCopy, &dev->flags)); + WARN_ON(dev->page != dev->orig_page); } if (!discard_pending && test_bit(R5_Discard, &sh->dev[sh->pd_idx].flags)) { @@ -2800,6 +3069,14 @@ static void handle_stripe_clean_event(struct r5conf *conf, } /* now that discard is done we can proceed with any sync */ clear_bit(STRIPE_DISCARD, &sh->state); + /* + * SCSI discard will change some bio fields and the stripe has + * no updated data, so remove it from hash list and the stripe + * will be reinitialized + */ + spin_lock_irq(&conf->device_lock); + remove_hash(sh); + spin_unlock_irq(&conf->device_lock); if (test_bit(STRIPE_SYNC_REQUESTED, &sh->state)) set_bit(STRIPE_HANDLE, &sh->state); @@ -2852,7 +3129,8 @@ static void handle_stripe_dirtying(struct r5conf *conf, !test_bit(R5_LOCKED, &dev->flags) && !(test_bit(R5_UPTODATE, &dev->flags) || test_bit(R5_Wantcompute, &dev->flags))) { - if (test_bit(R5_Insync, &dev->flags)) rcw++; + if (test_bit(R5_Insync, &dev->flags)) + rcw++; else rcw += 2*disks; } @@ -2873,10 +3151,10 @@ static void handle_stripe_dirtying(struct r5conf *conf, !(test_bit(R5_UPTODATE, &dev->flags) || test_bit(R5_Wantcompute, &dev->flags)) && test_bit(R5_Insync, &dev->flags)) { - if ( - test_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) { - pr_debug("Read_old block " - "%d for r-m-w\n", i); + if (test_bit(STRIPE_PREREAD_ACTIVE, + &sh->state)) { + pr_debug("Read_old block %d for r-m-w\n", + i); set_bit(R5_LOCKED, &dev->flags); set_bit(R5_Wantread, &dev->flags); s->locked++; @@ -2899,10 +3177,9 @@ static void handle_stripe_dirtying(struct r5conf *conf, !(test_bit(R5_UPTODATE, &dev->flags) || test_bit(R5_Wantcompute, &dev->flags))) { rcw++; - if (!test_bit(R5_Insync, &dev->flags)) - continue; /* it's a failed drive */ - if ( - test_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) { + if (test_bit(R5_Insync, &dev->flags) && + test_bit(STRIPE_PREREAD_ACTIVE, + &sh->state)) { pr_debug("Read_old block " "%d for Reconstruct\n", i); set_bit(R5_LOCKED, &dev->flags); @@ -3371,7 +3648,7 @@ static void analyse_stripe(struct stripe_head *sh, struct stripe_head_state *s) */ set_bit(R5_Insync, &dev->flags); - if (rdev && test_bit(R5_WriteError, &dev->flags)) { + if (test_bit(R5_WriteError, &dev->flags)) { /* This flag does not apply to '.replacement' * only to .rdev, so make sure to check that*/ struct md_rdev *rdev2 = rcu_dereference( @@ -3384,7 +3661,7 @@ static void analyse_stripe(struct stripe_head *sh, struct stripe_head_state *s) } else clear_bit(R5_WriteError, &dev->flags); } - if (rdev && test_bit(R5_MadeGood, &dev->flags)) { + if (test_bit(R5_MadeGood, &dev->flags)) { /* This flag does not apply to '.replacement' * only to .rdev, so make sure to check that*/ struct md_rdev *rdev2 = rcu_dereference( @@ -3462,6 +3739,7 @@ static void handle_stripe(struct stripe_head *sh) test_and_clear_bit(STRIPE_SYNC_REQUESTED, &sh->state)) { set_bit(STRIPE_SYNCING, &sh->state); clear_bit(STRIPE_INSYNC, &sh->state); + clear_bit(STRIPE_REPLACED, &sh->state); } spin_unlock(&sh->stripe_lock); } @@ -3607,19 +3885,23 @@ static void handle_stripe(struct stripe_head *sh) handle_parity_checks5(conf, sh, &s, disks); } - if (s.replacing && s.locked == 0 - && !test_bit(STRIPE_INSYNC, &sh->state)) { + if ((s.replacing || s.syncing) && s.locked == 0 + && !test_bit(STRIPE_COMPUTE_RUN, &sh->state) + && !test_bit(STRIPE_REPLACED, &sh->state)) { /* Write out to replacement devices where possible */ for (i = 0; i < conf->raid_disks; i++) - if (test_bit(R5_UPTODATE, &sh->dev[i].flags) && - test_bit(R5_NeedReplace, &sh->dev[i].flags)) { + if (test_bit(R5_NeedReplace, &sh->dev[i].flags)) { + WARN_ON(!test_bit(R5_UPTODATE, &sh->dev[i].flags)); set_bit(R5_WantReplace, &sh->dev[i].flags); set_bit(R5_LOCKED, &sh->dev[i].flags); s.locked++; } - set_bit(STRIPE_INSYNC, &sh->state); + if (s.replacing) + set_bit(STRIPE_INSYNC, &sh->state); + set_bit(STRIPE_REPLACED, &sh->state); } if ((s.syncing || s.replacing) && s.locked == 0 && + !test_bit(STRIPE_COMPUTE_RUN, &sh->state) && test_bit(STRIPE_INSYNC, &sh->state)) { md_done_sync(conf->mddev, STRIPE_SECTORS, 1); clear_bit(STRIPE_SYNCING, &sh->state); @@ -3774,11 +4056,13 @@ static void raid5_activate_delayed(struct r5conf *conf) if (!test_and_set_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) atomic_inc(&conf->preread_active_stripes); list_add_tail(&sh->lru, &conf->hold_list); + raid5_wakeup_stripe_thread(sh); } } } -static void activate_bit_delay(struct r5conf *conf) +static void activate_bit_delay(struct r5conf *conf, + struct list_head *temp_inactive_list) { /* device_lock is held */ struct list_head head; @@ -3786,9 +4070,11 @@ static void activate_bit_delay(struct r5conf *conf) list_del_init(&conf->bitmap_list); while (!list_empty(&head)) { struct stripe_head *sh = list_entry(head.next, struct stripe_head, lru); + int hash; list_del_init(&sh->lru); atomic_inc(&sh->count); - __release_stripe(conf, sh); + hash = sh->hash_lock_index; + __release_stripe(conf, sh, &temp_inactive_list[hash]); } } @@ -3804,7 +4090,7 @@ int md_raid5_congested(struct mddev *mddev, int bits) return 1; if (conf->quiesce) return 1; - if (list_empty_careful(&conf->inactive_list)) + if (atomic_read(&conf->empty_inactive_list_nr)) return 1; return 0; @@ -3848,7 +4134,7 @@ static int raid5_mergeable_bvec(struct request_queue *q, static int in_chunk_boundary(struct mddev *mddev, struct bio *bio) { - sector_t sector = bio->bi_sector + get_start_sect(bio->bi_bdev); + sector_t sector = bio->bi_iter.bi_sector + get_start_sect(bio->bi_bdev); unsigned int chunk_sectors = mddev->chunk_sectors; unsigned int bio_sectors = bio_sectors(bio); @@ -3985,9 +4271,9 @@ static int chunk_aligned_read(struct mddev *mddev, struct bio * raid_bio) /* * compute position */ - align_bi->bi_sector = raid5_compute_sector(conf, raid_bio->bi_sector, - 0, - &dd_idx, NULL); + align_bi->bi_iter.bi_sector = + raid5_compute_sector(conf, raid_bio->bi_iter.bi_sector, + 0, &dd_idx, NULL); end_sector = bio_end_sector(align_bi); rcu_read_lock(); @@ -4012,7 +4298,8 @@ static int chunk_aligned_read(struct mddev *mddev, struct bio * raid_bio) align_bi->bi_flags &= ~(1 << BIO_SEG_VALID); if (!bio_fits_rdev(align_bi) || - is_badblock(rdev, align_bi->bi_sector, bio_sectors(align_bi), + is_badblock(rdev, align_bi->bi_iter.bi_sector, + bio_sectors(align_bi), &first_bad, &bad_sectors)) { /* too big in some way, or has a known bad block */ bio_put(align_bi); @@ -4021,7 +4308,7 @@ static int chunk_aligned_read(struct mddev *mddev, struct bio * raid_bio) } /* No reshape active, so we can trust rdev->data_offset */ - align_bi->bi_sector += rdev->data_offset; + align_bi->bi_iter.bi_sector += rdev->data_offset; spin_lock_irq(&conf->device_lock); wait_event_lock_irq(conf->wait_for_stripe, @@ -4033,7 +4320,7 @@ static int chunk_aligned_read(struct mddev *mddev, struct bio * raid_bio) if (mddev->gendisk) trace_block_bio_remap(bdev_get_queue(align_bi->bi_bdev), align_bi, disk_devt(mddev->gendisk), - raid_bio->bi_sector); + raid_bio->bi_iter.bi_sector); generic_make_request(align_bi); return 1; } else { @@ -4053,18 +4340,35 @@ static int chunk_aligned_read(struct mddev *mddev, struct bio * raid_bio) * head of the hold_list has changed, i.e. the head was promoted to the * handle_list. */ -static struct stripe_head *__get_priority_stripe(struct r5conf *conf) +static struct stripe_head *__get_priority_stripe(struct r5conf *conf, int group) { - struct stripe_head *sh; + struct stripe_head *sh = NULL, *tmp; + struct list_head *handle_list = NULL; + struct r5worker_group *wg = NULL; + + if (conf->worker_cnt_per_group == 0) { + handle_list = &conf->handle_list; + } else if (group != ANY_GROUP) { + handle_list = &conf->worker_groups[group].handle_list; + wg = &conf->worker_groups[group]; + } else { + int i; + for (i = 0; i < conf->group_cnt; i++) { + handle_list = &conf->worker_groups[i].handle_list; + wg = &conf->worker_groups[i]; + if (!list_empty(handle_list)) + break; + } + } pr_debug("%s: handle: %s hold: %s full_writes: %d bypass_count: %d\n", __func__, - list_empty(&conf->handle_list) ? "empty" : "busy", + list_empty(handle_list) ? "empty" : "busy", list_empty(&conf->hold_list) ? "empty" : "busy", atomic_read(&conf->pending_full_writes), conf->bypass_count); - if (!list_empty(&conf->handle_list)) { - sh = list_entry(conf->handle_list.next, typeof(*sh), lru); + if (!list_empty(handle_list)) { + sh = list_entry(handle_list->next, typeof(*sh), lru); if (list_empty(&conf->hold_list)) conf->bypass_count = 0; @@ -4082,23 +4386,41 @@ static struct stripe_head *__get_priority_stripe(struct r5conf *conf) ((conf->bypass_threshold && conf->bypass_count > conf->bypass_threshold) || atomic_read(&conf->pending_full_writes) == 0)) { - sh = list_entry(conf->hold_list.next, - typeof(*sh), lru); - conf->bypass_count -= conf->bypass_threshold; - if (conf->bypass_count < 0) - conf->bypass_count = 0; - } else + + list_for_each_entry(tmp, &conf->hold_list, lru) { + if (conf->worker_cnt_per_group == 0 || + group == ANY_GROUP || + !cpu_online(tmp->cpu) || + cpu_to_group(tmp->cpu) == group) { + sh = tmp; + break; + } + } + + if (sh) { + conf->bypass_count -= conf->bypass_threshold; + if (conf->bypass_count < 0) + conf->bypass_count = 0; + } + wg = NULL; + } + + if (!sh) return NULL; + if (wg) { + wg->stripes_cnt--; + sh->group = NULL; + } list_del_init(&sh->lru); - atomic_inc(&sh->count); - BUG_ON(atomic_read(&sh->count) != 1); + BUG_ON(atomic_inc_return(&sh->count) != 1); return sh; } struct raid5_plug_cb { struct blk_plug_cb cb; struct list_head list; + struct list_head temp_inactive_list[NR_STRIPE_HASH_LOCKS]; }; static void raid5_unplug(struct blk_plug_cb *blk_cb, bool from_schedule) @@ -4109,6 +4431,7 @@ static void raid5_unplug(struct blk_plug_cb *blk_cb, bool from_schedule) struct mddev *mddev = cb->cb.data; struct r5conf *conf = mddev->private; int cnt = 0; + int hash; if (cb->list.next && !list_empty(&cb->list)) { spin_lock_irq(&conf->device_lock); @@ -4120,13 +4443,20 @@ static void raid5_unplug(struct blk_plug_cb *blk_cb, bool from_schedule) * STRIPE_ON_UNPLUG_LIST clear but the stripe * is still in our list */ - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(STRIPE_ON_UNPLUG_LIST, &sh->state); - __release_stripe(conf, sh); + /* + * STRIPE_ON_RELEASE_LIST could be set here. In that + * case, the count is always > 1 here + */ + hash = sh->hash_lock_index; + __release_stripe(conf, sh, &cb->temp_inactive_list[hash]); cnt++; } spin_unlock_irq(&conf->device_lock); } + release_inactive_stripe_list(conf, cb->temp_inactive_list, + NR_STRIPE_HASH_LOCKS); if (mddev->queue) trace_block_unplug(mddev->queue, cnt, !from_schedule); kfree(cb); @@ -4147,8 +4477,12 @@ static void release_stripe_plug(struct mddev *mddev, cb = container_of(blk_cb, struct raid5_plug_cb, cb); - if (cb->list.next == NULL) + if (cb->list.next == NULL) { + int i; INIT_LIST_HEAD(&cb->list); + for (i = 0; i < NR_STRIPE_HASH_LOCKS; i++) + INIT_LIST_HEAD(cb->temp_inactive_list + i); + } if (!test_and_set_bit(STRIPE_ON_UNPLUG_LIST, &sh->state)) list_add_tail(&sh->lru, &cb->list); @@ -4168,8 +4502,8 @@ static void make_discard_request(struct mddev *mddev, struct bio *bi) /* Skip discard while reshape is happening */ return; - logical_sector = bi->bi_sector & ~((sector_t)STRIPE_SECTORS-1); - last_sector = bi->bi_sector + (bi->bi_size>>9); + logical_sector = bi->bi_iter.bi_sector & ~((sector_t)STRIPE_SECTORS-1); + last_sector = bi->bi_iter.bi_sector + (bi->bi_iter.bi_size>>9); bi->bi_next = NULL; bi->bi_phys_segments = 1; /* over-loaded to count active stripes */ @@ -4255,6 +4589,8 @@ static void make_request(struct mddev *mddev, struct bio * bi) struct stripe_head *sh; const int rw = bio_data_dir(bi); int remaining; + DEFINE_WAIT(w); + bool do_prepare; if (unlikely(bi->bi_rw & REQ_FLUSH)) { md_flush_request(mddev, bi); @@ -4273,18 +4609,23 @@ static void make_request(struct mddev *mddev, struct bio * bi) return; } - logical_sector = bi->bi_sector & ~((sector_t)STRIPE_SECTORS-1); + logical_sector = bi->bi_iter.bi_sector & ~((sector_t)STRIPE_SECTORS-1); last_sector = bio_end_sector(bi); bi->bi_next = NULL; bi->bi_phys_segments = 1; /* over-loaded to count active stripes */ + prepare_to_wait(&conf->wait_for_overlap, &w, TASK_UNINTERRUPTIBLE); for (;logical_sector < last_sector; logical_sector += STRIPE_SECTORS) { - DEFINE_WAIT(w); int previous; + int seq; + do_prepare = false; retry: + seq = read_seqcount_begin(&conf->gen_lock); previous = 0; - prepare_to_wait(&conf->wait_for_overlap, &w, TASK_UNINTERRUPTIBLE); + if (do_prepare) + prepare_to_wait(&conf->wait_for_overlap, &w, + TASK_UNINTERRUPTIBLE); if (unlikely(conf->reshape_progress != MaxSector)) { /* spinlock is needed as reshape_progress may be * 64bit on a 32bit platform, and so it might be @@ -4305,6 +4646,7 @@ static void make_request(struct mddev *mddev, struct bio * bi) : logical_sector >= conf->reshape_safe) { spin_unlock_irq(&conf->device_lock); schedule(); + do_prepare = true; goto retry; } } @@ -4315,7 +4657,7 @@ static void make_request(struct mddev *mddev, struct bio * bi) previous, &dd_idx, NULL); pr_debug("raid456: make_request, sector %llu logical %llu\n", - (unsigned long long)new_sector, + (unsigned long long)new_sector, (unsigned long long)logical_sector); sh = get_active_stripe(conf, new_sector, previous, @@ -4341,9 +4683,17 @@ static void make_request(struct mddev *mddev, struct bio * bi) if (must_retry) { release_stripe(sh); schedule(); + do_prepare = true; goto retry; } } + if (read_seqcount_retry(&conf->gen_lock, seq)) { + /* Might have got the wrong stripe_head + * by accident + */ + release_stripe(sh); + goto retry; + } if (rw == WRITE && logical_sector >= mddev->suspend_lo && @@ -4357,8 +4707,10 @@ static void make_request(struct mddev *mddev, struct bio * bi) prepare_to_wait(&conf->wait_for_overlap, &w, TASK_INTERRUPTIBLE); if (logical_sector >= mddev->suspend_lo && - logical_sector < mddev->suspend_hi) + logical_sector < mddev->suspend_hi) { schedule(); + do_prepare = true; + } goto retry; } @@ -4371,9 +4723,9 @@ static void make_request(struct mddev *mddev, struct bio * bi) md_wakeup_thread(mddev->thread); release_stripe(sh); schedule(); + do_prepare = true; goto retry; } - finish_wait(&conf->wait_for_overlap, &w); set_bit(STRIPE_HANDLE, &sh->state); clear_bit(STRIPE_DELAYED, &sh->state); if ((bi->bi_rw & REQ_SYNC) && @@ -4383,10 +4735,10 @@ static void make_request(struct mddev *mddev, struct bio * bi) } else { /* cannot get stripe for read-ahead, just give-up */ clear_bit(BIO_UPTODATE, &bi->bi_flags); - finish_wait(&conf->wait_for_overlap, &w); break; } } + finish_wait(&conf->wait_for_overlap, &w); remaining = raid5_dec_bi_active_stripes(bi); if (remaining == 0) { @@ -4522,14 +4874,19 @@ static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr, int *sk time_after(jiffies, conf->reshape_checkpoint + 10*HZ)) { /* Cannot proceed until we've updated the superblock... */ wait_event(conf->wait_for_overlap, - atomic_read(&conf->reshape_stripes)==0); + atomic_read(&conf->reshape_stripes)==0 + || test_bit(MD_RECOVERY_INTR, &mddev->recovery)); + if (atomic_read(&conf->reshape_stripes) != 0) + return 0; mddev->reshape_position = conf->reshape_progress; mddev->curr_resync_completed = sector_nr; conf->reshape_checkpoint = jiffies; set_bit(MD_CHANGE_DEVS, &mddev->flags); md_wakeup_thread(mddev->thread); wait_event(mddev->sb_wait, mddev->flags == 0 || - kthread_should_stop()); + test_bit(MD_RECOVERY_INTR, &mddev->recovery)); + if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) + return 0; spin_lock_irq(&conf->device_lock); conf->reshape_safe = mddev->reshape_position; spin_unlock_irq(&conf->device_lock); @@ -4612,7 +4969,10 @@ static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr, int *sk >= mddev->resync_max - mddev->curr_resync_completed) { /* Cannot proceed until we've updated the superblock... */ wait_event(conf->wait_for_overlap, - atomic_read(&conf->reshape_stripes) == 0); + atomic_read(&conf->reshape_stripes) == 0 + || test_bit(MD_RECOVERY_INTR, &mddev->recovery)); + if (atomic_read(&conf->reshape_stripes) != 0) + goto ret; mddev->reshape_position = conf->reshape_progress; mddev->curr_resync_completed = sector_nr; conf->reshape_checkpoint = jiffies; @@ -4620,13 +4980,16 @@ static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr, int *sk md_wakeup_thread(mddev->thread); wait_event(mddev->sb_wait, !test_bit(MD_CHANGE_DEVS, &mddev->flags) - || kthread_should_stop()); + || test_bit(MD_RECOVERY_INTR, &mddev->recovery)); + if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) + goto ret; spin_lock_irq(&conf->device_lock); conf->reshape_safe = mddev->reshape_position; spin_unlock_irq(&conf->device_lock); wake_up(&conf->wait_for_overlap); sysfs_notify(&mddev->kobj, NULL, "sync_completed"); } +ret: return reshape_sectors; } @@ -4711,8 +5074,8 @@ static inline sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int bitmap_start_sync(mddev->bitmap, sector_nr, &sync_blocks, still_degraded); set_bit(STRIPE_SYNC_REQUESTED, &sh->state); + set_bit(STRIPE_HANDLE, &sh->state); - handle_stripe(sh); release_stripe(sh); return STRIPE_SECTORS; @@ -4737,7 +5100,8 @@ static int retry_aligned_read(struct r5conf *conf, struct bio *raid_bio) int remaining; int handled = 0; - logical_sector = raid_bio->bi_sector & ~((sector_t)STRIPE_SECTORS-1); + logical_sector = raid_bio->bi_iter.bi_sector & + ~((sector_t)STRIPE_SECTORS-1); sector = raid5_compute_sector(conf, logical_sector, 0, &dd_idx, NULL); last_sector = bio_end_sector(raid_bio); @@ -4751,7 +5115,7 @@ static int retry_aligned_read(struct r5conf *conf, struct bio *raid_bio) /* already done this stripe */ continue; - sh = get_active_stripe(conf, sector, 0, 1, 0); + sh = get_active_stripe(conf, sector, 0, 1, 1); if (!sh) { /* failed to get a stripe - must wait */ @@ -4783,31 +5147,83 @@ static int retry_aligned_read(struct r5conf *conf, struct bio *raid_bio) return handled; } -#define MAX_STRIPE_BATCH 8 -static int handle_active_stripes(struct r5conf *conf) +static int handle_active_stripes(struct r5conf *conf, int group, + struct r5worker *worker, + struct list_head *temp_inactive_list) { struct stripe_head *batch[MAX_STRIPE_BATCH], *sh; - int i, batch_size = 0; + int i, batch_size = 0, hash; + bool release_inactive = false; while (batch_size < MAX_STRIPE_BATCH && - (sh = __get_priority_stripe(conf)) != NULL) + (sh = __get_priority_stripe(conf, group)) != NULL) batch[batch_size++] = sh; - if (batch_size == 0) - return batch_size; + if (batch_size == 0) { + for (i = 0; i < NR_STRIPE_HASH_LOCKS; i++) + if (!list_empty(temp_inactive_list + i)) + break; + if (i == NR_STRIPE_HASH_LOCKS) + return batch_size; + release_inactive = true; + } spin_unlock_irq(&conf->device_lock); + release_inactive_stripe_list(conf, temp_inactive_list, + NR_STRIPE_HASH_LOCKS); + + if (release_inactive) { + spin_lock_irq(&conf->device_lock); + return 0; + } + for (i = 0; i < batch_size; i++) handle_stripe(batch[i]); cond_resched(); spin_lock_irq(&conf->device_lock); - for (i = 0; i < batch_size; i++) - __release_stripe(conf, batch[i]); + for (i = 0; i < batch_size; i++) { + hash = batch[i]->hash_lock_index; + __release_stripe(conf, batch[i], &temp_inactive_list[hash]); + } return batch_size; } +static void raid5_do_work(struct work_struct *work) +{ + struct r5worker *worker = container_of(work, struct r5worker, work); + struct r5worker_group *group = worker->group; + struct r5conf *conf = group->conf; + int group_id = group - conf->worker_groups; + int handled; + struct blk_plug plug; + + pr_debug("+++ raid5worker active\n"); + + blk_start_plug(&plug); + handled = 0; + spin_lock_irq(&conf->device_lock); + while (1) { + int batch_size, released; + + released = release_stripe_list(conf, worker->temp_inactive_list); + + batch_size = handle_active_stripes(conf, group_id, worker, + worker->temp_inactive_list); + worker->working = false; + if (!batch_size && !released) + break; + handled += batch_size; + } + pr_debug("%d stripes handled\n", handled); + + spin_unlock_irq(&conf->device_lock); + blk_finish_plug(&plug); + + pr_debug("--- raid5worker inactive\n"); +} + /* * This is our raid5 kernel thread. * @@ -4831,7 +5247,9 @@ static void raid5d(struct md_thread *thread) spin_lock_irq(&conf->device_lock); while (1) { struct bio *bio; - int batch_size; + int batch_size, released; + + released = release_stripe_list(conf, conf->temp_inactive_list); if ( !list_empty(&conf->bitmap_list)) { @@ -4841,7 +5259,7 @@ static void raid5d(struct md_thread *thread) bitmap_unplug(mddev->bitmap); spin_lock_irq(&conf->device_lock); conf->seq_write = conf->seq_flush; - activate_bit_delay(conf); + activate_bit_delay(conf, conf->temp_inactive_list); } raid5_activate_delayed(conf); @@ -4855,8 +5273,9 @@ static void raid5d(struct md_thread *thread) handled++; } - batch_size = handle_active_stripes(conf); - if (!batch_size) + batch_size = handle_active_stripes(conf, ANY_GROUP, NULL, + conf->temp_inactive_list); + if (!batch_size && !released) break; handled += batch_size; @@ -4891,22 +5310,29 @@ raid5_set_cache_size(struct mddev *mddev, int size) { struct r5conf *conf = mddev->private; int err; + int hash; if (size <= 16 || size > 32768) return -EINVAL; + hash = (conf->max_nr_stripes - 1) % NR_STRIPE_HASH_LOCKS; while (size < conf->max_nr_stripes) { - if (drop_one_stripe(conf)) + if (drop_one_stripe(conf, hash)) conf->max_nr_stripes--; else break; + hash--; + if (hash < 0) + hash = NR_STRIPE_HASH_LOCKS - 1; } err = md_allow_write(mddev); if (err) return err; + hash = conf->max_nr_stripes % NR_STRIPE_HASH_LOCKS; while (size > conf->max_nr_stripes) { - if (grow_one_stripe(conf)) + if (grow_one_stripe(conf, hash)) conf->max_nr_stripes++; else break; + hash = (hash + 1) % NR_STRIPE_HASH_LOCKS; } return 0; } @@ -4924,7 +5350,7 @@ raid5_store_stripe_cache_size(struct mddev *mddev, const char *page, size_t len) if (!conf) return -ENODEV; - if (strict_strtoul(page, 10, &new)) + if (kstrtoul(page, 10, &new)) return -EINVAL; err = raid5_set_cache_size(mddev, new); if (err) @@ -4957,7 +5383,7 @@ raid5_store_preread_threshold(struct mddev *mddev, const char *page, size_t len) if (!conf) return -ENODEV; - if (strict_strtoul(page, 10, &new)) + if (kstrtoul(page, 10, &new)) return -EINVAL; if (new > conf->max_nr_stripes) return -EINVAL; @@ -4972,6 +5398,50 @@ raid5_preread_bypass_threshold = __ATTR(preread_bypass_threshold, raid5_store_preread_threshold); static ssize_t +raid5_show_skip_copy(struct mddev *mddev, char *page) +{ + struct r5conf *conf = mddev->private; + if (conf) + return sprintf(page, "%d\n", conf->skip_copy); + else + return 0; +} + +static ssize_t +raid5_store_skip_copy(struct mddev *mddev, const char *page, size_t len) +{ + struct r5conf *conf = mddev->private; + unsigned long new; + if (len >= PAGE_SIZE) + return -EINVAL; + if (!conf) + return -ENODEV; + + if (kstrtoul(page, 10, &new)) + return -EINVAL; + new = !!new; + if (new == conf->skip_copy) + return len; + + mddev_suspend(mddev); + conf->skip_copy = new; + if (new) + mddev->queue->backing_dev_info.capabilities |= + BDI_CAP_STABLE_WRITES; + else + mddev->queue->backing_dev_info.capabilities &= + ~BDI_CAP_STABLE_WRITES; + mddev_resume(mddev); + return len; +} + +static struct md_sysfs_entry +raid5_skip_copy = __ATTR(skip_copy, S_IRUGO | S_IWUSR, + raid5_show_skip_copy, + raid5_store_skip_copy); + + +static ssize_t stripe_cache_active_show(struct mddev *mddev, char *page) { struct r5conf *conf = mddev->private; @@ -4984,10 +5454,79 @@ stripe_cache_active_show(struct mddev *mddev, char *page) static struct md_sysfs_entry raid5_stripecache_active = __ATTR_RO(stripe_cache_active); +static ssize_t +raid5_show_group_thread_cnt(struct mddev *mddev, char *page) +{ + struct r5conf *conf = mddev->private; + if (conf) + return sprintf(page, "%d\n", conf->worker_cnt_per_group); + else + return 0; +} + +static int alloc_thread_groups(struct r5conf *conf, int cnt, + int *group_cnt, + int *worker_cnt_per_group, + struct r5worker_group **worker_groups); +static ssize_t +raid5_store_group_thread_cnt(struct mddev *mddev, const char *page, size_t len) +{ + struct r5conf *conf = mddev->private; + unsigned long new; + int err; + struct r5worker_group *new_groups, *old_groups; + int group_cnt, worker_cnt_per_group; + + if (len >= PAGE_SIZE) + return -EINVAL; + if (!conf) + return -ENODEV; + + if (kstrtoul(page, 10, &new)) + return -EINVAL; + + if (new == conf->worker_cnt_per_group) + return len; + + mddev_suspend(mddev); + + old_groups = conf->worker_groups; + if (old_groups) + flush_workqueue(raid5_wq); + + err = alloc_thread_groups(conf, new, + &group_cnt, &worker_cnt_per_group, + &new_groups); + if (!err) { + spin_lock_irq(&conf->device_lock); + conf->group_cnt = group_cnt; + conf->worker_cnt_per_group = worker_cnt_per_group; + conf->worker_groups = new_groups; + spin_unlock_irq(&conf->device_lock); + + if (old_groups) + kfree(old_groups[0].workers); + kfree(old_groups); + } + + mddev_resume(mddev); + + if (err) + return err; + return len; +} + +static struct md_sysfs_entry +raid5_group_thread_cnt = __ATTR(group_thread_cnt, S_IRUGO | S_IWUSR, + raid5_show_group_thread_cnt, + raid5_store_group_thread_cnt); + static struct attribute *raid5_attrs[] = { &raid5_stripecache_size.attr, &raid5_stripecache_active.attr, &raid5_preread_bypass_threshold.attr, + &raid5_group_thread_cnt.attr, + &raid5_skip_copy.attr, NULL, }; static struct attribute_group raid5_attrs_group = { @@ -4995,6 +5534,61 @@ static struct attribute_group raid5_attrs_group = { .attrs = raid5_attrs, }; +static int alloc_thread_groups(struct r5conf *conf, int cnt, + int *group_cnt, + int *worker_cnt_per_group, + struct r5worker_group **worker_groups) +{ + int i, j, k; + ssize_t size; + struct r5worker *workers; + + *worker_cnt_per_group = cnt; + if (cnt == 0) { + *group_cnt = 0; + *worker_groups = NULL; + return 0; + } + *group_cnt = num_possible_nodes(); + size = sizeof(struct r5worker) * cnt; + workers = kzalloc(size * *group_cnt, GFP_NOIO); + *worker_groups = kzalloc(sizeof(struct r5worker_group) * + *group_cnt, GFP_NOIO); + if (!*worker_groups || !workers) { + kfree(workers); + kfree(*worker_groups); + return -ENOMEM; + } + + for (i = 0; i < *group_cnt; i++) { + struct r5worker_group *group; + + group = &(*worker_groups)[i]; + INIT_LIST_HEAD(&group->handle_list); + group->conf = conf; + group->workers = workers + i * cnt; + + for (j = 0; j < cnt; j++) { + struct r5worker *worker = group->workers + j; + worker->group = group; + INIT_WORK(&worker->work, raid5_do_work); + + for (k = 0; k < NR_STRIPE_HASH_LOCKS; k++) + INIT_LIST_HEAD(worker->temp_inactive_list + k); + } + } + + return 0; +} + +static void free_thread_groups(struct r5conf *conf) +{ + if (conf->worker_groups) + kfree(conf->worker_groups[0].workers); + kfree(conf->worker_groups); + conf->worker_groups = NULL; +} + static sector_t raid5_size(struct mddev *mddev, sector_t sectors, int raid_disks) { @@ -5011,23 +5605,43 @@ raid5_size(struct mddev *mddev, sector_t sectors, int raid_disks) return sectors * (raid_disks - conf->max_degraded); } +static void free_scratch_buffer(struct r5conf *conf, struct raid5_percpu *percpu) +{ + safe_put_page(percpu->spare_page); + kfree(percpu->scribble); + percpu->spare_page = NULL; + percpu->scribble = NULL; +} + +static int alloc_scratch_buffer(struct r5conf *conf, struct raid5_percpu *percpu) +{ + if (conf->level == 6 && !percpu->spare_page) + percpu->spare_page = alloc_page(GFP_KERNEL); + if (!percpu->scribble) + percpu->scribble = kmalloc(conf->scribble_len, GFP_KERNEL); + + if (!percpu->scribble || (conf->level == 6 && !percpu->spare_page)) { + free_scratch_buffer(conf, percpu); + return -ENOMEM; + } + + return 0; +} + static void raid5_free_percpu(struct r5conf *conf) { - struct raid5_percpu *percpu; unsigned long cpu; if (!conf->percpu) return; - get_online_cpus(); - for_each_possible_cpu(cpu) { - percpu = per_cpu_ptr(conf->percpu, cpu); - safe_put_page(percpu->spare_page); - kfree(percpu->scribble); - } #ifdef CONFIG_HOTPLUG_CPU unregister_cpu_notifier(&conf->cpu_notify); #endif + + get_online_cpus(); + for_each_possible_cpu(cpu) + free_scratch_buffer(conf, per_cpu_ptr(conf->percpu, cpu)); put_online_cpus(); free_percpu(conf->percpu); @@ -5035,6 +5649,7 @@ static void raid5_free_percpu(struct r5conf *conf) static void free_conf(struct r5conf *conf) { + free_thread_groups(conf); shrink_stripes(conf); raid5_free_percpu(conf); kfree(conf->disks); @@ -5053,15 +5668,7 @@ static int raid456_cpu_notify(struct notifier_block *nfb, unsigned long action, switch (action) { case CPU_UP_PREPARE: case CPU_UP_PREPARE_FROZEN: - if (conf->level == 6 && !percpu->spare_page) - percpu->spare_page = alloc_page(GFP_KERNEL); - if (!percpu->scribble) - percpu->scribble = kmalloc(conf->scribble_len, GFP_KERNEL); - - if (!percpu->scribble || - (conf->level == 6 && !percpu->spare_page)) { - safe_put_page(percpu->spare_page); - kfree(percpu->scribble); + if (alloc_scratch_buffer(conf, percpu)) { pr_err("%s: failed memory allocation for cpu%ld\n", __func__, cpu); return notifier_from_errno(-ENOMEM); @@ -5069,10 +5676,7 @@ static int raid456_cpu_notify(struct notifier_block *nfb, unsigned long action, break; case CPU_DEAD: case CPU_DEAD_FROZEN: - safe_put_page(percpu->spare_page); - kfree(percpu->scribble); - percpu->spare_page = NULL; - percpu->scribble = NULL; + free_scratch_buffer(conf, per_cpu_ptr(conf->percpu, cpu)); break; default: break; @@ -5084,40 +5688,29 @@ static int raid456_cpu_notify(struct notifier_block *nfb, unsigned long action, static int raid5_alloc_percpu(struct r5conf *conf) { unsigned long cpu; - struct page *spare_page; - struct raid5_percpu __percpu *allcpus; - void *scribble; - int err; + int err = 0; - allcpus = alloc_percpu(struct raid5_percpu); - if (!allcpus) + conf->percpu = alloc_percpu(struct raid5_percpu); + if (!conf->percpu) return -ENOMEM; - conf->percpu = allcpus; + +#ifdef CONFIG_HOTPLUG_CPU + conf->cpu_notify.notifier_call = raid456_cpu_notify; + conf->cpu_notify.priority = 0; + err = register_cpu_notifier(&conf->cpu_notify); + if (err) + return err; +#endif get_online_cpus(); - err = 0; for_each_present_cpu(cpu) { - if (conf->level == 6) { - spare_page = alloc_page(GFP_KERNEL); - if (!spare_page) { - err = -ENOMEM; - break; - } - per_cpu_ptr(conf->percpu, cpu)->spare_page = spare_page; - } - scribble = kmalloc(conf->scribble_len, GFP_KERNEL); - if (!scribble) { - err = -ENOMEM; + err = alloc_scratch_buffer(conf, per_cpu_ptr(conf->percpu, cpu)); + if (err) { + pr_err("%s: failed memory allocation for cpu%ld\n", + __func__, cpu); break; } - per_cpu_ptr(conf->percpu, cpu)->scribble = scribble; } -#ifdef CONFIG_HOTPLUG_CPU - conf->cpu_notify.notifier_call = raid456_cpu_notify; - conf->cpu_notify.priority = 0; - if (err == 0) - err = register_cpu_notifier(&conf->cpu_notify); -#endif put_online_cpus(); return err; @@ -5130,6 +5723,9 @@ static struct r5conf *setup_conf(struct mddev *mddev) struct md_rdev *rdev; struct disk_info *disk; char pers_name[6]; + int i; + int group_cnt, worker_cnt_per_group; + struct r5worker_group *new_group; if (mddev->new_level != 5 && mddev->new_level != 4 @@ -5163,14 +5759,23 @@ static struct r5conf *setup_conf(struct mddev *mddev) conf = kzalloc(sizeof(struct r5conf), GFP_KERNEL); if (conf == NULL) goto abort; + /* Don't enable multi-threading by default*/ + if (!alloc_thread_groups(conf, 0, &group_cnt, &worker_cnt_per_group, + &new_group)) { + conf->group_cnt = group_cnt; + conf->worker_cnt_per_group = worker_cnt_per_group; + conf->worker_groups = new_group; + } else + goto abort; spin_lock_init(&conf->device_lock); + seqcount_init(&conf->gen_lock); init_waitqueue_head(&conf->wait_for_stripe); init_waitqueue_head(&conf->wait_for_overlap); INIT_LIST_HEAD(&conf->handle_list); INIT_LIST_HEAD(&conf->hold_list); INIT_LIST_HEAD(&conf->delayed_list); INIT_LIST_HEAD(&conf->bitmap_list); - INIT_LIST_HEAD(&conf->inactive_list); + init_llist_head(&conf->released_stripes); atomic_set(&conf->active_stripes, 0); atomic_set(&conf->preread_active_stripes, 0); atomic_set(&conf->active_aligned_reads, 0); @@ -5195,6 +5800,21 @@ static struct r5conf *setup_conf(struct mddev *mddev) if ((conf->stripe_hashtbl = kzalloc(PAGE_SIZE, GFP_KERNEL)) == NULL) goto abort; + /* We init hash_locks[0] separately to that it can be used + * as the reference lock in the spin_lock_nest_lock() call + * in lock_all_device_hash_locks_irq in order to convince + * lockdep that we know what we are doing. + */ + spin_lock_init(conf->hash_locks); + for (i = 1; i < NR_STRIPE_HASH_LOCKS; i++) + spin_lock_init(conf->hash_locks + i); + + for (i = 0; i < NR_STRIPE_HASH_LOCKS; i++) + INIT_LIST_HEAD(conf->inactive_list + i); + + for (i = 0; i < NR_STRIPE_HASH_LOCKS; i++) + INIT_LIST_HEAD(conf->temp_inactive_list + i); + conf->level = mddev->new_level; if (raid5_alloc_percpu(conf) != 0) goto abort; @@ -5235,7 +5855,6 @@ static struct r5conf *setup_conf(struct mddev *mddev) else conf->max_degraded = 1; conf->algorithm = mddev->new_layout; - conf->max_nr_stripes = NR_STRIPES; conf->reshape_progress = mddev->reshape_position; if (conf->reshape_progress != MaxSector) { conf->prev_chunk_sectors = mddev->chunk_sectors; @@ -5244,7 +5863,8 @@ static struct r5conf *setup_conf(struct mddev *mddev) memory = conf->max_nr_stripes * (sizeof(struct stripe_head) + max_disks * ((sizeof(struct bio) + PAGE_SIZE))) / 1024; - if (grow_stripes(conf, conf->max_nr_stripes)) { + atomic_set(&conf->empty_inactive_list_nr, NR_STRIPE_HASH_LOCKS); + if (grow_stripes(conf, NR_STRIPES)) { printk(KERN_ERR "md/raid:%s: couldn't allocate %dkB for buffers\n", mdname(mddev), memory); @@ -5572,6 +6192,7 @@ static int run(struct mddev *mddev) blk_queue_io_min(mddev->queue, chunk_size); blk_queue_io_opt(mddev->queue, chunk_size * (conf->raid_disks - conf->max_degraded)); + mddev->queue->limits.raid_partial_stripes_expensive = 1; /* * We can only discard a whole stripe. It doesn't make sense to * discard data disk but write parity disk @@ -5914,7 +6535,7 @@ static int check_reshape(struct mddev *mddev) return 0; /* nothing to do */ if (has_failed(conf)) return -EINVAL; - if (mddev->delta_disks < 0) { + if (mddev->delta_disks < 0 && mddev->reshape_position == MaxSector) { /* We might be able to shrink, but the devices must * be made bigger first. * For raid6, 4 is the minimum size. @@ -5975,6 +6596,7 @@ static int raid5_start_reshape(struct mddev *mddev) atomic_set(&conf->reshape_stripes, 0); spin_lock_irq(&conf->device_lock); + write_seqcount_begin(&conf->gen_lock); conf->previous_raid_disks = conf->raid_disks; conf->raid_disks += mddev->delta_disks; conf->prev_chunk_sectors = conf->chunk_sectors; @@ -5991,8 +6613,16 @@ static int raid5_start_reshape(struct mddev *mddev) else conf->reshape_progress = 0; conf->reshape_safe = conf->reshape_progress; + write_seqcount_end(&conf->gen_lock); spin_unlock_irq(&conf->device_lock); + /* Now make sure any requests that proceeded on the assumption + * the reshape wasn't running - like Discard or Read - have + * completed. + */ + mddev_suspend(mddev); + mddev_resume(mddev); + /* Add some new drives, as many as will fit. * We know there are enough to make the newly sized array work. * Don't add devices if we are reducing the number of @@ -6041,12 +6671,18 @@ static int raid5_start_reshape(struct mddev *mddev) if (!mddev->sync_thread) { mddev->recovery = 0; spin_lock_irq(&conf->device_lock); + write_seqcount_begin(&conf->gen_lock); mddev->raid_disks = conf->raid_disks = conf->previous_raid_disks; + mddev->new_chunk_sectors = + conf->chunk_sectors = conf->prev_chunk_sectors; + mddev->new_layout = conf->algorithm = conf->prev_algo; rdev_for_each(rdev, mddev) rdev->new_data_offset = rdev->data_offset; smp_wmb(); + conf->generation --; conf->reshape_progress = MaxSector; mddev->reshape_position = MaxSector; + write_seqcount_end(&conf->gen_lock); spin_unlock_irq(&conf->device_lock); return -EAGAIN; } @@ -6134,27 +6770,28 @@ static void raid5_quiesce(struct mddev *mddev, int state) break; case 1: /* stop all writes */ - spin_lock_irq(&conf->device_lock); + lock_all_device_hash_locks_irq(conf); /* '2' tells resync/reshape to pause so that all * active stripes can drain */ conf->quiesce = 2; - wait_event_lock_irq(conf->wait_for_stripe, + wait_event_cmd(conf->wait_for_stripe, atomic_read(&conf->active_stripes) == 0 && atomic_read(&conf->active_aligned_reads) == 0, - conf->device_lock); + unlock_all_device_hash_locks_irq(conf), + lock_all_device_hash_locks_irq(conf)); conf->quiesce = 1; - spin_unlock_irq(&conf->device_lock); + unlock_all_device_hash_locks_irq(conf); /* allow reshape to continue */ wake_up(&conf->wait_for_overlap); break; case 0: /* re-enable writes */ - spin_lock_irq(&conf->device_lock); + lock_all_device_hash_locks_irq(conf); conf->quiesce = 0; wake_up(&conf->wait_for_stripe); wake_up(&conf->wait_for_overlap); - spin_unlock_irq(&conf->device_lock); + unlock_all_device_hash_locks_irq(conf); break; } } @@ -6467,6 +7104,10 @@ static struct md_personality raid4_personality = static int __init raid5_init(void) { + raid5_wq = alloc_workqueue("raid5wq", + WQ_UNBOUND|WQ_MEM_RECLAIM|WQ_CPU_INTENSIVE|WQ_SYSFS, 0); + if (!raid5_wq) + return -ENOMEM; register_md_personality(&raid6_personality); register_md_personality(&raid5_personality); register_md_personality(&raid4_personality); @@ -6478,6 +7119,7 @@ static void raid5_exit(void) unregister_md_personality(&raid6_personality); unregister_md_personality(&raid5_personality); unregister_md_personality(&raid4_personality); + destroy_workqueue(raid5_wq); } module_init(raid5_init); diff --git a/drivers/md/raid5.h b/drivers/md/raid5.h index b0b663b119a..bc72cd4be5f 100644 --- a/drivers/md/raid5.h +++ b/drivers/md/raid5.h @@ -49,7 +49,7 @@ * can't distinguish between a clean block that has been generated * from parity calculations, and a clean block that has been * successfully written to the spare ( or to parity when resyncing). - * To distingush these states we have a stripe bit STRIPE_INSYNC that + * To distinguish these states we have a stripe bit STRIPE_INSYNC that * is set whenever a write is scheduled to the spare, or to the parity * disc if there is no spare. A sync request clears this bit, and * when we find it set with no buffers locked, we know the sync is @@ -197,6 +197,7 @@ enum reconstruct_states { struct stripe_head { struct hlist_node hash; struct list_head lru; /* inactive_list or handle_list */ + struct llist_node release_list; struct r5conf *raid_conf; short generation; /* increments with every * reshape */ @@ -204,6 +205,7 @@ struct stripe_head { short pd_idx; /* parity disk index */ short qd_idx; /* 'Q' disk index for raid6 */ short ddf_layout;/* use DDF ordering to calculate Q */ + short hash_lock_index; unsigned long state; /* state flags */ atomic_t count; /* nr of active thread/requests */ int bm_seq; /* sequence number for bitmap flushes */ @@ -211,6 +213,8 @@ struct stripe_head { enum check_states check_state; enum reconstruct_states reconstruct_state; spinlock_t stripe_lock; + int cpu; + struct r5worker_group *group; /** * struct stripe_operations * @target - STRIPE_OP_COMPUTE_BLK target @@ -228,7 +232,7 @@ struct stripe_head { */ struct bio req, rreq; struct bio_vec vec, rvec; - struct page *page; + struct page *page, *orig_page; struct bio *toread, *read, *towrite, *written; sector_t sector; /* sector of this page */ unsigned long flags; @@ -295,6 +299,7 @@ enum r5dev_flags { * data in, and now is a good time to write it out. */ R5_Discard, /* Discard the stripe */ + R5_SkipCopy, /* Don't copy data from bio to stripe cache */ }; /* @@ -306,6 +311,7 @@ enum { STRIPE_SYNC_REQUESTED, STRIPE_SYNCING, STRIPE_INSYNC, + STRIPE_REPLACED, STRIPE_PREREAD_ACTIVE, STRIPE_DELAYED, STRIPE_DEGRADED, @@ -320,6 +326,7 @@ enum { STRIPE_OPS_REQ_PENDING, STRIPE_ON_UNPLUG_LIST, STRIPE_DISCARD, + STRIPE_ON_RELEASE_LIST, }; /* @@ -362,8 +369,32 @@ struct disk_info { struct md_rdev *rdev, *replacement; }; +/* NOTE NR_STRIPE_HASH_LOCKS must remain below 64. + * This is because we sometimes take all the spinlocks + * and creating that much locking depth can cause + * problems. + */ +#define NR_STRIPE_HASH_LOCKS 8 +#define STRIPE_HASH_LOCKS_MASK (NR_STRIPE_HASH_LOCKS - 1) + +struct r5worker { + struct work_struct work; + struct r5worker_group *group; + struct list_head temp_inactive_list[NR_STRIPE_HASH_LOCKS]; + bool working; +}; + +struct r5worker_group { + struct list_head handle_list; + struct r5conf *conf; + struct r5worker *workers; + int stripes_cnt; +}; + struct r5conf { struct hlist_head *stripe_hashtbl; + /* only protect corresponding hash list and inactive_list */ + spinlock_t hash_locks[NR_STRIPE_HASH_LOCKS]; struct mddev *mddev; int chunk_sectors; int level, algorithm; @@ -385,6 +416,7 @@ struct r5conf { int prev_chunk_sectors; int prev_algo; short generation; /* increments with every reshape */ + seqcount_t gen_lock; /* lock against generation changes */ unsigned long reshape_checkpoint; /* Time we last updated * metadata */ long long min_offset_diff; /* minimum difference between @@ -405,6 +437,7 @@ struct r5conf { atomic_t pending_full_writes; /* full write backlog */ int bypass_count; /* bypassed prereads */ int bypass_threshold; /* preread nice */ + int skip_copy; /* Don't copy data from bio to stripe cache */ struct list_head *last_hold; /* detect hold_list promotions */ atomic_t reshape_stripes; /* stripes with pending writes for reshape */ @@ -443,7 +476,9 @@ struct r5conf { * Free stripes pool */ atomic_t active_stripes; - struct list_head inactive_list; + struct list_head inactive_list[NR_STRIPE_HASH_LOCKS]; + atomic_t empty_inactive_list_nr; + struct llist_head released_stripes; wait_queue_head_t wait_for_stripe; wait_queue_head_t wait_for_overlap; int inactive_blocked; /* release of inactive stripes blocked, @@ -457,6 +492,10 @@ struct r5conf { * the new thread here until we fully activate the array. */ struct md_thread *thread; + struct list_head temp_inactive_list[NR_STRIPE_HASH_LOCKS]; + struct r5worker_group *worker_groups; + int group_cnt; + int worker_cnt_per_group; }; /* |