diff options
Diffstat (limited to 'drivers/md')
37 files changed, 697 insertions, 292 deletions
diff --git a/drivers/md/bcache/alloc.c b/drivers/md/bcache/alloc.c index a1df0d95151c..5310e1f4a282 100644 --- a/drivers/md/bcache/alloc.c +++ b/drivers/md/bcache/alloc.c @@ -49,7 +49,7 @@ * * bch_bucket_alloc() allocates a single bucket from a specific cache. * - * bch_bucket_alloc_set() allocates one or more buckets from different caches + * bch_bucket_alloc_set() allocates one bucket from different caches * out of a cache set. * * free_some_buckets() drives all the processes described above. It's called @@ -488,34 +488,29 @@ void bch_bucket_free(struct cache_set *c, struct bkey *k) } int __bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve, - struct bkey *k, int n, bool wait) + struct bkey *k, bool wait) { - int i; + struct cache *ca; + long b; /* No allocation if CACHE_SET_IO_DISABLE bit is set */ if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &c->flags))) return -1; lockdep_assert_held(&c->bucket_lock); - BUG_ON(!n || n > c->caches_loaded || n > MAX_CACHES_PER_SET); bkey_init(k); - /* sort by free space/prio of oldest data in caches */ - - for (i = 0; i < n; i++) { - struct cache *ca = c->cache_by_alloc[i]; - long b = bch_bucket_alloc(ca, reserve, wait); + ca = c->cache_by_alloc[0]; + b = bch_bucket_alloc(ca, reserve, wait); + if (b == -1) + goto err; - if (b == -1) - goto err; + k->ptr[0] = MAKE_PTR(ca->buckets[b].gen, + bucket_to_sector(c, b), + ca->sb.nr_this_dev); - k->ptr[i] = MAKE_PTR(ca->buckets[b].gen, - bucket_to_sector(c, b), - ca->sb.nr_this_dev); - - SET_KEY_PTRS(k, i + 1); - } + SET_KEY_PTRS(k, 1); return 0; err: @@ -525,12 +520,12 @@ err: } int bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve, - struct bkey *k, int n, bool wait) + struct bkey *k, bool wait) { int ret; mutex_lock(&c->bucket_lock); - ret = __bch_bucket_alloc_set(c, reserve, k, n, wait); + ret = __bch_bucket_alloc_set(c, reserve, k, wait); mutex_unlock(&c->bucket_lock); return ret; } @@ -638,7 +633,7 @@ bool bch_alloc_sectors(struct cache_set *c, spin_unlock(&c->data_bucket_lock); - if (bch_bucket_alloc_set(c, watermark, &alloc.key, 1, wait)) + if (bch_bucket_alloc_set(c, watermark, &alloc.key, wait)) return false; spin_lock(&c->data_bucket_lock); diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h index 36de6f7ddf22..7bce58278845 100644 --- a/drivers/md/bcache/bcache.h +++ b/drivers/md/bcache/bcache.h @@ -265,6 +265,7 @@ struct bcache_device { #define BCACHE_DEV_WB_RUNNING 3 #define BCACHE_DEV_RATE_DW_RUNNING 4 int nr_stripes; +#define BCH_MIN_STRIPE_SZ ((4 << 20) >> SECTOR_SHIFT) unsigned int stripe_size; atomic_t *stripe_sectors_dirty; unsigned long *full_dirty_stripes; @@ -970,9 +971,9 @@ void bch_bucket_free(struct cache_set *c, struct bkey *k); long bch_bucket_alloc(struct cache *ca, unsigned int reserve, bool wait); int __bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve, - struct bkey *k, int n, bool wait); + struct bkey *k, bool wait); int bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve, - struct bkey *k, int n, bool wait); + struct bkey *k, bool wait); bool bch_alloc_sectors(struct cache_set *c, struct bkey *k, unsigned int sectors, unsigned int write_point, unsigned int write_prio, bool wait); diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c index 5a33910aea78..5db893d6a824 100644 --- a/drivers/md/bcache/btree.c +++ b/drivers/md/bcache/btree.c @@ -1020,6 +1020,9 @@ err: * * The btree node will have either a read or a write lock held, depending on * level and op->lock. + * + * Note: Only error code or btree pointer will be returned, it is unncessary + * for callers to check NULL pointer. */ struct btree *bch_btree_node_get(struct cache_set *c, struct btree_op *op, struct bkey *k, int level, bool write, @@ -1132,16 +1135,22 @@ retry: mutex_unlock(&b->c->bucket_lock); } +/* + * Only error code or btree pointer will be returned, it is unncessary for + * callers to check NULL pointer. + */ struct btree *__bch_btree_node_alloc(struct cache_set *c, struct btree_op *op, int level, bool wait, struct btree *parent) { BKEY_PADDED(key) k; - struct btree *b = ERR_PTR(-EAGAIN); + struct btree *b; mutex_lock(&c->bucket_lock); retry: - if (__bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, 1, wait)) + /* return ERR_PTR(-EAGAIN) when it fails */ + b = ERR_PTR(-EAGAIN); + if (__bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, wait)) goto err; bkey_put(c, &k.key); @@ -1186,7 +1195,7 @@ static struct btree *btree_node_alloc_replacement(struct btree *b, { struct btree *n = bch_btree_node_alloc(b->c, op, b->level, b->parent); - if (!IS_ERR_OR_NULL(n)) { + if (!IS_ERR(n)) { mutex_lock(&n->write_lock); bch_btree_sort_into(&b->keys, &n->keys, &b->c->sort); bkey_copy_key(&n->key, &b->key); @@ -1401,7 +1410,7 @@ static int btree_gc_coalesce(struct btree *b, struct btree_op *op, 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])) + if (IS_ERR(new_nodes[i])) goto out_nocoalesce; } @@ -1553,6 +1562,8 @@ static int btree_gc_rewrite_node(struct btree *b, struct btree_op *op, return 0; n = btree_node_alloc_replacement(replace, NULL); + if (IS_ERR(n)) + return 0; /* recheck reserve after allocating replacement node */ if (btree_check_reserve(b, NULL)) { @@ -1718,7 +1729,7 @@ static int bch_btree_gc_root(struct btree *b, struct btree_op *op, if (should_rewrite) { n = btree_node_alloc_replacement(b, NULL); - if (!IS_ERR_OR_NULL(n)) { + if (!IS_ERR(n)) { bch_btree_node_write_sync(n); bch_btree_set_root(n); diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c index efdf6ce0443e..5d1a4eb81669 100644 --- a/drivers/md/bcache/super.c +++ b/drivers/md/bcache/super.c @@ -428,7 +428,7 @@ static int __uuid_write(struct cache_set *c) closure_init_stack(&cl); lockdep_assert_held(&bch_register_lock); - if (bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, 1, true)) + if (bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, true)) return 1; SET_KEY_SIZE(&k.key, c->sb.bucket_size); @@ -822,6 +822,8 @@ static int bcache_device_init(struct bcache_device *d, unsigned int block_size, if (!d->stripe_size) d->stripe_size = 1 << 31; + else if (d->stripe_size < BCH_MIN_STRIPE_SZ) + d->stripe_size = roundup(BCH_MIN_STRIPE_SZ, d->stripe_size); n = DIV_ROUND_UP_ULL(sectors, d->stripe_size); if (!n || n > max_stripes) { @@ -1633,7 +1635,7 @@ static void cache_set_flush(struct closure *cl) if (!IS_ERR_OR_NULL(c->gc_thread)) kthread_stop(c->gc_thread); - if (!IS_ERR_OR_NULL(c->root)) + if (!IS_ERR(c->root)) list_add(&c->root->list, &c->btree_cache); /* @@ -1906,7 +1908,7 @@ static int run_cache_set(struct cache_set *c) c->root = bch_btree_node_get(c, NULL, k, j->btree_level, true, NULL); - if (IS_ERR_OR_NULL(c->root)) + if (IS_ERR(c->root)) goto err; list_del_init(&c->root->list); @@ -2000,7 +2002,7 @@ static int run_cache_set(struct cache_set *c) err = "cannot allocate new btree root"; c->root = __bch_btree_node_alloc(c, NULL, 0, true, NULL); - if (IS_ERR_OR_NULL(c->root)) + if (IS_ERR(c->root)) goto err; mutex_lock(&c->root->write_lock); diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c index 7f0fb4b5755a..30073e4074c1 100644 --- a/drivers/md/bcache/sysfs.c +++ b/drivers/md/bcache/sysfs.c @@ -1057,7 +1057,7 @@ SHOW(__bch_cache) sum += INITIAL_PRIO - cached[i]; if (n) - do_div(sum, n); + sum = div64_u64(sum, n); for (i = 0; i < ARRAY_SIZE(q); i++) q[i] = INITIAL_PRIO - cached[n * (i + 1) / diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c index 0b02210ab435..5767ff6c13e3 100644 --- a/drivers/md/bcache/writeback.c +++ b/drivers/md/bcache/writeback.c @@ -119,27 +119,61 @@ static void __update_writeback_rate(struct cached_dev *dc) dc->writeback_rate_target = target; } +static bool idle_counter_exceeded(struct cache_set *c) +{ + int counter, dev_nr; + + /* + * If c->idle_counter is overflow (idel for really long time), + * reset as 0 and not set maximum rate this time for code + * simplicity. + */ + counter = atomic_inc_return(&c->idle_counter); + if (counter <= 0) { + atomic_set(&c->idle_counter, 0); + return false; + } + + dev_nr = atomic_read(&c->attached_dev_nr); + if (dev_nr == 0) + return false; + + /* + * c->idle_counter is increased by writeback thread of all + * attached backing devices, in order to represent a rough + * time period, counter should be divided by dev_nr. + * Otherwise the idle time cannot be larger with more backing + * device attached. + * The following calculation equals to checking + * (counter / dev_nr) < (dev_nr * 6) + */ + if (counter < (dev_nr * dev_nr * 6)) + return false; + + return true; +} + +/* + * Idle_counter is increased every time when update_writeback_rate() is + * called. If all backing devices attached to the same cache set have + * identical dc->writeback_rate_update_seconds values, it is about 6 + * rounds of update_writeback_rate() on each backing device before + * c->at_max_writeback_rate is set to 1, and then max wrteback rate set + * to each dc->writeback_rate.rate. + * In order to avoid extra locking cost for counting exact dirty cached + * devices number, c->attached_dev_nr is used to calculate the idle + * throushold. It might be bigger if not all cached device are in write- + * back mode, but it still works well with limited extra rounds of + * update_writeback_rate(). + */ static bool set_at_max_writeback_rate(struct cache_set *c, struct cached_dev *dc) { /* Don't set max writeback rate if gc is running */ if (!c->gc_mark_valid) return false; - /* - * Idle_counter is increased everytime when update_writeback_rate() is - * called. If all backing devices attached to the same cache set have - * identical dc->writeback_rate_update_seconds values, it is about 6 - * rounds of update_writeback_rate() on each backing device before - * c->at_max_writeback_rate is set to 1, and then max wrteback rate set - * to each dc->writeback_rate.rate. - * In order to avoid extra locking cost for counting exact dirty cached - * devices number, c->attached_dev_nr is used to calculate the idle - * throushold. It might be bigger if not all cached device are in write- - * back mode, but it still works well with limited extra rounds of - * update_writeback_rate(). - */ - if (atomic_inc_return(&c->idle_counter) < - atomic_read(&c->attached_dev_nr) * 6) + + if (!idle_counter_exceeded(c)) return false; if (atomic_read(&c->at_max_writeback_rate) != 1) @@ -153,13 +187,10 @@ static bool set_at_max_writeback_rate(struct cache_set *c, dc->writeback_rate_change = 0; /* - * Check c->idle_counter and c->at_max_writeback_rate agagain in case - * new I/O arrives during before set_at_max_writeback_rate() returns. - * Then the writeback rate is set to 1, and its new value should be - * decided via __update_writeback_rate(). + * In case new I/O arrives during before + * set_at_max_writeback_rate() returns. */ - if ((atomic_read(&c->idle_counter) < - atomic_read(&c->attached_dev_nr) * 6) || + if (!idle_counter_exceeded(c) || !atomic_read(&c->at_max_writeback_rate)) return false; diff --git a/drivers/md/dm-cache-metadata.c b/drivers/md/dm-cache-metadata.c index af6d4f898e4c..2ecd0db0f294 100644 --- a/drivers/md/dm-cache-metadata.c +++ b/drivers/md/dm-cache-metadata.c @@ -551,11 +551,13 @@ static int __create_persistent_data_objects(struct dm_cache_metadata *cmd, return r; } -static void __destroy_persistent_data_objects(struct dm_cache_metadata *cmd) +static void __destroy_persistent_data_objects(struct dm_cache_metadata *cmd, + bool destroy_bm) { dm_sm_destroy(cmd->metadata_sm); dm_tm_destroy(cmd->tm); - dm_block_manager_destroy(cmd->bm); + if (destroy_bm) + dm_block_manager_destroy(cmd->bm); } typedef unsigned long (*flags_mutator)(unsigned long); @@ -826,7 +828,7 @@ static struct dm_cache_metadata *lookup_or_open(struct block_device *bdev, cmd2 = lookup(bdev); if (cmd2) { mutex_unlock(&table_lock); - __destroy_persistent_data_objects(cmd); + __destroy_persistent_data_objects(cmd, true); kfree(cmd); return cmd2; } @@ -874,7 +876,7 @@ void dm_cache_metadata_close(struct dm_cache_metadata *cmd) mutex_unlock(&table_lock); if (!cmd->fail_io) - __destroy_persistent_data_objects(cmd); + __destroy_persistent_data_objects(cmd, true); kfree(cmd); } } @@ -1808,14 +1810,52 @@ int dm_cache_metadata_needs_check(struct dm_cache_metadata *cmd, bool *result) int dm_cache_metadata_abort(struct dm_cache_metadata *cmd) { - int r; + int r = -EINVAL; + struct dm_block_manager *old_bm = NULL, *new_bm = NULL; + + /* fail_io is double-checked with cmd->root_lock held below */ + if (unlikely(cmd->fail_io)) + return r; + + /* + * Replacement block manager (new_bm) is created and old_bm destroyed outside of + * cmd root_lock to avoid ABBA deadlock that would result (due to life-cycle of + * shrinker associated with the block manager's bufio client vs cmd root_lock). + * - must take shrinker_rwsem without holding cmd->root_lock + */ + new_bm = dm_block_manager_create(cmd->bdev, DM_CACHE_METADATA_BLOCK_SIZE << SECTOR_SHIFT, + CACHE_MAX_CONCURRENT_LOCKS); WRITE_LOCK(cmd); - __destroy_persistent_data_objects(cmd); - r = __create_persistent_data_objects(cmd, false); + if (cmd->fail_io) { + WRITE_UNLOCK(cmd); + goto out; + } + + __destroy_persistent_data_objects(cmd, false); + old_bm = cmd->bm; + if (IS_ERR(new_bm)) { + DMERR("could not create block manager during abort"); + cmd->bm = NULL; + r = PTR_ERR(new_bm); + goto out_unlock; + } + + cmd->bm = new_bm; + r = __open_or_format_metadata(cmd, false); + if (r) { + cmd->bm = NULL; + goto out_unlock; + } + new_bm = NULL; +out_unlock: if (r) cmd->fail_io = true; WRITE_UNLOCK(cmd); + dm_block_manager_destroy(old_bm); +out: + if (new_bm && !IS_ERR(new_bm)) + dm_block_manager_destroy(new_bm); return r; } diff --git a/drivers/md/dm-cache-policy-smq.c b/drivers/md/dm-cache-policy-smq.c index b61aac00ff40..859073193f5b 100644 --- a/drivers/md/dm-cache-policy-smq.c +++ b/drivers/md/dm-cache-policy-smq.c @@ -854,7 +854,13 @@ struct smq_policy { struct background_tracker *bg_work; - bool migrations_allowed; + bool migrations_allowed:1; + + /* + * If this is set the policy will try and clean the whole cache + * even if the device is not idle. + */ + bool cleaner:1; }; /*----------------------------------------------------------------*/ @@ -1133,7 +1139,7 @@ static bool clean_target_met(struct smq_policy *mq, bool idle) * Cache entries may not be populated. So we cannot rely on the * size of the clean queue. */ - if (idle) { + if (idle || mq->cleaner) { /* * We'd like to clean everything. */ @@ -1716,11 +1722,9 @@ static void calc_hotspot_params(sector_t origin_size, *hotspot_block_size /= 2u; } -static struct dm_cache_policy *__smq_create(dm_cblock_t cache_size, - sector_t origin_size, - sector_t cache_block_size, - bool mimic_mq, - bool migrations_allowed) +static struct dm_cache_policy * +__smq_create(dm_cblock_t cache_size, sector_t origin_size, sector_t cache_block_size, + bool mimic_mq, bool migrations_allowed, bool cleaner) { unsigned i; unsigned nr_sentinels_per_queue = 2u * NR_CACHE_LEVELS; @@ -1807,6 +1811,7 @@ static struct dm_cache_policy *__smq_create(dm_cblock_t cache_size, goto bad_btracker; mq->migrations_allowed = migrations_allowed; + mq->cleaner = cleaner; return &mq->policy; @@ -1830,21 +1835,24 @@ static struct dm_cache_policy *smq_create(dm_cblock_t cache_size, sector_t origin_size, sector_t cache_block_size) { - return __smq_create(cache_size, origin_size, cache_block_size, false, true); + return __smq_create(cache_size, origin_size, cache_block_size, + false, true, false); } static struct dm_cache_policy *mq_create(dm_cblock_t cache_size, sector_t origin_size, sector_t cache_block_size) { - return __smq_create(cache_size, origin_size, cache_block_size, true, true); + return __smq_create(cache_size, origin_size, cache_block_size, + true, true, false); } static struct dm_cache_policy *cleaner_create(dm_cblock_t cache_size, sector_t origin_size, sector_t cache_block_size) { - return __smq_create(cache_size, origin_size, cache_block_size, false, false); + return __smq_create(cache_size, origin_size, cache_block_size, + false, false, true); } /*----------------------------------------------------------------*/ diff --git a/drivers/md/dm-cache-target.c b/drivers/md/dm-cache-target.c index f595e9867cbe..08b5e44df324 100644 --- a/drivers/md/dm-cache-target.c +++ b/drivers/md/dm-cache-target.c @@ -1011,16 +1011,16 @@ static void abort_transaction(struct cache *cache) if (get_cache_mode(cache) >= CM_READ_ONLY) return; - if (dm_cache_metadata_set_needs_check(cache->cmd)) { - DMERR("%s: failed to set 'needs_check' flag in metadata", dev_name); - set_cache_mode(cache, CM_FAIL); - } - DMERR_LIMIT("%s: aborting current metadata transaction", dev_name); if (dm_cache_metadata_abort(cache->cmd)) { DMERR("%s: failed to abort metadata transaction", dev_name); set_cache_mode(cache, CM_FAIL); } + + if (dm_cache_metadata_set_needs_check(cache->cmd)) { + DMERR("%s: failed to set 'needs_check' flag in metadata", dev_name); + set_cache_mode(cache, CM_FAIL); + } } static void metadata_operation_failed(struct cache *cache, const char *op, int r) @@ -1910,6 +1910,7 @@ static void process_deferred_bios(struct work_struct *ws) else commit_needed = process_bio(cache, bio) || commit_needed; + cond_resched(); } if (commit_needed) @@ -1932,6 +1933,7 @@ static void requeue_deferred_bios(struct cache *cache) while ((bio = bio_list_pop(&bios))) { bio->bi_status = BLK_STS_DM_REQUEUE; bio_endio(bio); + cond_resched(); } } @@ -1972,6 +1974,8 @@ static void check_migrations(struct work_struct *ws) r = mg_start(cache, op, NULL); if (r) break; + + cond_resched(); } } @@ -1992,6 +1996,7 @@ static void destroy(struct cache *cache) if (cache->prison) dm_bio_prison_destroy_v2(cache->prison); + cancel_delayed_work_sync(&cache->waker); if (cache->wq) destroy_workqueue(cache->wq); diff --git a/drivers/md/dm-clone-target.c b/drivers/md/dm-clone-target.c index eb7a5d3ba81a..355afca2969d 100644 --- a/drivers/md/dm-clone-target.c +++ b/drivers/md/dm-clone-target.c @@ -1977,6 +1977,7 @@ static void clone_dtr(struct dm_target *ti) mempool_exit(&clone->hydration_pool); dm_kcopyd_client_destroy(clone->kcopyd_client); + cancel_delayed_work_sync(&clone->waker); destroy_workqueue(clone->wq); hash_table_exit(clone); dm_clone_metadata_close(clone->cmd); @@ -2231,6 +2232,7 @@ static int __init dm_clone_init(void) r = dm_register_target(&clone_target); if (r < 0) { DMERR("Failed to register clone target"); + kmem_cache_destroy(_hydration_cache); return r; } diff --git a/drivers/md/dm-core.h b/drivers/md/dm-core.h index 3fea121fcbcf..e3c3bbe92677 100644 --- a/drivers/md/dm-core.h +++ b/drivers/md/dm-core.h @@ -18,6 +18,8 @@ #include "dm.h" #define DM_RESERVED_MAX_IOS 1024 +#define DM_MAX_TARGETS 1048576 +#define DM_MAX_TARGET_PARAMS 1024 struct dm_kobject_holder { struct kobject kobj; diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c index fa674e9b6f23..6866aa2aade3 100644 --- a/drivers/md/dm-crypt.c +++ b/drivers/md/dm-crypt.c @@ -46,11 +46,11 @@ struct convert_context { struct completion restart; struct bio *bio_in; - struct bio *bio_out; struct bvec_iter iter_in; + struct bio *bio_out; struct bvec_iter iter_out; - u64 cc_sector; atomic_t cc_pending; + u64 cc_sector; union { struct skcipher_request *req; struct aead_request *req_aead; @@ -1554,6 +1554,7 @@ pop_from_list: io = crypt_io_from_node(rb_first(&write_tree)); rb_erase(&io->rb_node, &write_tree); kcryptd_io_write(io); + cond_resched(); } while (!RB_EMPTY_ROOT(&write_tree)); blk_finish_plug(&plug); } @@ -1626,6 +1627,12 @@ static void kcryptd_crypt_write_convert(struct dm_crypt_io *io) io->ctx.bio_out = clone; io->ctx.iter_out = clone->bi_iter; + if (crypt_integrity_aead(cc)) { + bio_copy_data(clone, io->base_bio); + io->ctx.bio_in = clone; + io->ctx.iter_in = clone->bi_iter; + } + sector += bio_sectors(clone); crypt_inc_pending(io); diff --git a/drivers/md/dm-delay.c b/drivers/md/dm-delay.c index f496213f8b67..7c0e7c662e07 100644 --- a/drivers/md/dm-delay.c +++ b/drivers/md/dm-delay.c @@ -30,7 +30,7 @@ struct delay_c { struct workqueue_struct *kdelayd_wq; struct work_struct flush_expired_bios; struct list_head delayed_bios; - atomic_t may_delay; + bool may_delay; struct delay_class read; struct delay_class write; @@ -191,7 +191,7 @@ static int delay_ctr(struct dm_target *ti, unsigned int argc, char **argv) INIT_WORK(&dc->flush_expired_bios, flush_expired_bios); INIT_LIST_HEAD(&dc->delayed_bios); mutex_init(&dc->timer_lock); - atomic_set(&dc->may_delay, 1); + dc->may_delay = true; dc->argc = argc; ret = delay_class_ctr(ti, &dc->read, argv); @@ -245,7 +245,7 @@ static int delay_bio(struct delay_c *dc, struct delay_class *c, struct bio *bio) struct dm_delay_info *delayed; unsigned long expires = 0; - if (!c->delay || !atomic_read(&dc->may_delay)) + if (!c->delay) return DM_MAPIO_REMAPPED; delayed = dm_per_bio_data(bio, sizeof(struct dm_delay_info)); @@ -254,6 +254,10 @@ static int delay_bio(struct delay_c *dc, struct delay_class *c, struct bio *bio) delayed->expires = expires = jiffies + msecs_to_jiffies(c->delay); mutex_lock(&delayed_bios_lock); + if (unlikely(!dc->may_delay)) { + mutex_unlock(&delayed_bios_lock); + return DM_MAPIO_REMAPPED; + } c->ops++; list_add_tail(&delayed->list, &dc->delayed_bios); mutex_unlock(&delayed_bios_lock); @@ -267,7 +271,10 @@ static void delay_presuspend(struct dm_target *ti) { struct delay_c *dc = ti->private; - atomic_set(&dc->may_delay, 0); + mutex_lock(&delayed_bios_lock); + dc->may_delay = false; + mutex_unlock(&delayed_bios_lock); + del_timer_sync(&dc->delay_timer); flush_bios(flush_delayed_bios(dc, 1)); } @@ -276,7 +283,7 @@ static void delay_resume(struct dm_target *ti) { struct delay_c *dc = ti->private; - atomic_set(&dc->may_delay, 1); + dc->may_delay = true; } static int delay_map(struct dm_target *ti, struct bio *bio) diff --git a/drivers/md/dm-flakey.c b/drivers/md/dm-flakey.c index 2900fbde89b3..967d4ad6d32c 100644 --- a/drivers/md/dm-flakey.c +++ b/drivers/md/dm-flakey.c @@ -124,9 +124,9 @@ static int parse_features(struct dm_arg_set *as, struct flakey_c *fc, * Direction r or w? */ arg_name = dm_shift_arg(as); - if (!strcasecmp(arg_name, "w")) + if (arg_name && !strcasecmp(arg_name, "w")) fc->corrupt_bio_rw = WRITE; - else if (!strcasecmp(arg_name, "r")) + else if (arg_name && !strcasecmp(arg_name, "r")) fc->corrupt_bio_rw = READ; else { ti->error = "Invalid corrupt bio direction (r or w)"; @@ -301,8 +301,11 @@ static void corrupt_bio_data(struct bio *bio, struct flakey_c *fc) */ bio_for_each_segment(bvec, bio, iter) { if (bio_iter_len(bio, iter) > corrupt_bio_byte) { - char *segment = (page_address(bio_iter_page(bio, iter)) - + bio_iter_offset(bio, iter)); + char *segment; + struct page *page = bio_iter_page(bio, iter); + if (unlikely(page == ZERO_PAGE(0))) + break; + segment = (page_address(page) + bio_iter_offset(bio, iter)); segment[corrupt_bio_byte] = fc->corrupt_bio_value; DMDEBUG("Corrupting data bio=%p by writing %u to byte %u " "(rw=%c bi_opf=%u bi_sector=%llu size=%u)\n", @@ -360,9 +363,11 @@ static int flakey_map(struct dm_target *ti, struct bio *bio) /* * Corrupt matching writes. */ - if (fc->corrupt_bio_byte && (fc->corrupt_bio_rw == WRITE)) { - if (all_corrupt_bio_flags_match(bio, fc)) - corrupt_bio_data(bio, fc); + if (fc->corrupt_bio_byte) { + if (fc->corrupt_bio_rw == WRITE) { + if (all_corrupt_bio_flags_match(bio, fc)) + corrupt_bio_data(bio, fc); + } goto map_bio; } @@ -388,13 +393,14 @@ static int flakey_end_io(struct dm_target *ti, struct bio *bio, return DM_ENDIO_DONE; if (!*error && pb->bio_submitted && (bio_data_dir(bio) == READ)) { - if (fc->corrupt_bio_byte && (fc->corrupt_bio_rw == READ) && - all_corrupt_bio_flags_match(bio, fc)) { - /* - * Corrupt successful matching READs while in down state. - */ - corrupt_bio_data(bio, fc); - + if (fc->corrupt_bio_byte) { + if ((fc->corrupt_bio_rw == READ) && + all_corrupt_bio_flags_match(bio, fc)) { + /* + * Corrupt successful matching READs while in down state. + */ + corrupt_bio_data(bio, fc); + } } else if (!test_bit(DROP_WRITES, &fc->flags) && !test_bit(ERROR_WRITES, &fc->flags)) { /* diff --git a/drivers/md/dm-integrity.c b/drivers/md/dm-integrity.c index acbda91e7643..4b634633b4a5 100644 --- a/drivers/md/dm-integrity.c +++ b/drivers/md/dm-integrity.c @@ -31,11 +31,11 @@ #define DEFAULT_BUFFER_SECTORS 128 #define DEFAULT_JOURNAL_WATERMARK 50 #define DEFAULT_SYNC_MSEC 10000 -#define DEFAULT_MAX_JOURNAL_SECTORS 131072 +#define DEFAULT_MAX_JOURNAL_SECTORS (IS_ENABLED(CONFIG_64BIT) ? 131072 : 8192) #define MIN_LOG2_INTERLEAVE_SECTORS 3 #define MAX_LOG2_INTERLEAVE_SECTORS 31 #define METADATA_WORKQUEUE_MAX_ACTIVE 16 -#define RECALC_SECTORS 8192 +#define RECALC_SECTORS (IS_ENABLED(CONFIG_64BIT) ? 32768 : 2048) #define RECALC_WRITE_SUPER 16 #define BITMAP_BLOCK_SIZE 4096 /* don't change it */ #define BITMAP_FLUSH_INTERVAL (10 * HZ) @@ -1582,11 +1582,12 @@ static void integrity_metadata(struct work_struct *w) } __bio_for_each_segment(bv, bio, iter, dio->bio_details.bi_iter) { + struct bio_vec bv_copy = bv; unsigned pos; char *mem, *checksums_ptr; again: - mem = (char *)kmap_atomic(bv.bv_page) + bv.bv_offset; + mem = (char *)kmap_atomic(bv_copy.bv_page) + bv_copy.bv_offset; pos = 0; checksums_ptr = checksums; do { @@ -1595,7 +1596,7 @@ again: sectors_to_process -= ic->sectors_per_block; pos += ic->sectors_per_block << SECTOR_SHIFT; sector += ic->sectors_per_block; - } while (pos < bv.bv_len && sectors_to_process && checksums != checksums_onstack); + } while (pos < bv_copy.bv_len && sectors_to_process && checksums != checksums_onstack); kunmap_atomic(mem); r = dm_integrity_rw_tag(ic, checksums, &dio->metadata_block, &dio->metadata_offset, @@ -1615,9 +1616,9 @@ again: if (!sectors_to_process) break; - if (unlikely(pos < bv.bv_len)) { - bv.bv_offset += pos; - bv.bv_len -= pos; + if (unlikely(pos < bv_copy.bv_len)) { + bv_copy.bv_offset += pos; + bv_copy.bv_len -= pos; goto again; } } @@ -2346,10 +2347,6 @@ static void integrity_writer(struct work_struct *w) unsigned prev_free_sectors; - /* the following test is not needed, but it tests the replay code */ - if (unlikely(dm_post_suspending(ic->ti)) && !ic->meta_dev) - return; - spin_lock_irq(&ic->endio_wait.lock); write_start = ic->committed_section; write_sections = ic->n_committed_sections; @@ -2858,8 +2855,7 @@ static void dm_integrity_postsuspend(struct dm_target *ti) drain_workqueue(ic->commit_wq); if (ic->mode == 'J') { - if (ic->meta_dev) - queue_work(ic->writer_wq, &ic->writer_work); + queue_work(ic->writer_wq, &ic->writer_work); drain_workqueue(ic->writer_wq); dm_integrity_flush_buffers(ic, true); } @@ -3759,7 +3755,7 @@ static int dm_integrity_ctr(struct dm_target *ti, unsigned argc, char **argv) } else if (sscanf(opt_string, "sectors_per_bit:%llu%c", &llval, &dummy) == 1) { log2_sectors_per_bitmap_bit = !llval ? 0 : __ilog2_u64(llval); } else if (sscanf(opt_string, "bitmap_flush_interval:%u%c", &val, &dummy) == 1) { - if (val >= (uint64_t)UINT_MAX * 1000 / HZ) { + if ((uint64_t)val >= (uint64_t)UINT_MAX * 1000 / HZ) { r = -EINVAL; ti->error = "Invalid bitmap_flush_interval argument"; goto bad; @@ -4200,6 +4196,8 @@ static void dm_integrity_dtr(struct dm_target *ti) BUG_ON(!RB_EMPTY_ROOT(&ic->in_progress)); BUG_ON(!list_empty(&ic->wait_list)); + if (ic->mode == 'B') + cancel_delayed_work_sync(&ic->bitmap_flush_work); if (ic->metadata_wq) destroy_workqueue(ic->metadata_wq); if (ic->wait_wq) @@ -4291,11 +4289,13 @@ static int __init dm_integrity_init(void) } r = dm_register_target(&integrity_target); - - if (r < 0) + if (r < 0) { DMERR("register failed %d", r); + kmem_cache_destroy(journal_io_cache); + return r; + } - return r; + return 0; } static void __exit dm_integrity_exit(void) diff --git a/drivers/md/dm-io.c b/drivers/md/dm-io.c index 81ffc59d05c9..4312007d2d34 100644 --- a/drivers/md/dm-io.c +++ b/drivers/md/dm-io.c @@ -306,7 +306,7 @@ static void do_region(int op, int op_flags, unsigned region, struct request_queue *q = bdev_get_queue(where->bdev); unsigned short logical_block_size = queue_logical_block_size(q); sector_t num_sectors; - unsigned int uninitialized_var(special_cmd_max_sectors); + unsigned int special_cmd_max_sectors; /* * Reject unsupported discard and write same requests. diff --git a/drivers/md/dm-ioctl.c b/drivers/md/dm-ioctl.c index 7a73f2fa0ad7..e89e710dd292 100644 --- a/drivers/md/dm-ioctl.c +++ b/drivers/md/dm-ioctl.c @@ -573,7 +573,7 @@ static void list_version_get_needed(struct target_type *tt, void *needed_param) size_t *needed = needed_param; *needed += sizeof(struct dm_target_versions); - *needed += strlen(tt->name); + *needed += strlen(tt->name) + 1; *needed += ALIGN_MASK; } @@ -638,7 +638,7 @@ static int __list_versions(struct dm_ioctl *param, size_t param_size, const char iter_info.old_vers = NULL; iter_info.vers = vers; iter_info.flags = 0; - iter_info.end = (char *)vers+len; + iter_info.end = (char *)vers + needed; /* * Now loop through filling out the names & versions. @@ -1435,11 +1435,12 @@ static int table_clear(struct file *filp, struct dm_ioctl *param, size_t param_s hc->new_map = NULL; } - param->flags &= ~DM_INACTIVE_PRESENT_FLAG; - - __dev_status(hc->md, param); md = hc->md; up_write(&_hash_lock); + + param->flags &= ~DM_INACTIVE_PRESENT_FLAG; + __dev_status(md, param); + if (old_map) { dm_sync_table(md); dm_table_destroy(old_map); @@ -1759,7 +1760,8 @@ static int copy_params(struct dm_ioctl __user *user, struct dm_ioctl *param_kern if (copy_from_user(param_kernel, user, minimum_data_size)) return -EFAULT; - if (param_kernel->data_size < minimum_data_size) + if (unlikely(param_kernel->data_size < minimum_data_size) || + unlikely(param_kernel->data_size > DM_MAX_TARGETS * DM_MAX_TARGET_PARAMS)) return -EINVAL; secure_data = param_kernel->flags & DM_SECURE_DATA_FLAG; @@ -1847,7 +1849,7 @@ static int ctl_ioctl(struct file *file, uint command, struct dm_ioctl __user *us int ioctl_flags; int param_flags; unsigned int cmd; - struct dm_ioctl *uninitialized_var(param); + struct dm_ioctl *param; ioctl_fn fn = NULL; size_t input_param_size; struct dm_ioctl param_kernel; diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c index 42151c9fc6e5..25eecb92f5f3 100644 --- a/drivers/md/dm-raid.c +++ b/drivers/md/dm-raid.c @@ -3284,15 +3284,19 @@ static int raid_ctr(struct dm_target *ti, unsigned int argc, char **argv) /* Try to adjust the raid4/5/6 stripe cache size to the stripe size */ if (rs_is_raid456(rs)) { r = rs_set_raid456_stripe_cache(rs); - if (r) + if (r) { + mddev_unlock(&rs->md); goto bad_stripe_cache; + } } /* Now do an early reshape check */ if (test_bit(RT_FLAG_RESHAPE_RS, &rs->runtime_flags)) { r = rs_check_reshape(rs); - if (r) + if (r) { + mddev_unlock(&rs->md); goto bad_check_reshape; + } /* Restore new, ctr requested layout to perform check */ rs_config_restore(rs, &rs_layout); @@ -3301,6 +3305,7 @@ static int raid_ctr(struct dm_target *ti, unsigned int argc, char **argv) r = rs->md.pers->check_reshape(&rs->md); if (r) { ti->error = "Reshape check failed"; + mddev_unlock(&rs->md); goto bad_check_reshape; } } @@ -3338,14 +3343,14 @@ static int raid_map(struct dm_target *ti, struct bio *bio) struct mddev *mddev = &rs->md; /* - * If we're reshaping to add disk(s)), ti->len and + * If we're reshaping to add disk(s), ti->len and * mddev->array_sectors will differ during the process * (ti->len > mddev->array_sectors), so we have to requeue * bios with addresses > mddev->array_sectors here or * there will occur accesses past EOD of the component * data images thus erroring the raid set. */ - if (unlikely(bio_end_sector(bio) > mddev->array_sectors)) + if (unlikely(bio_has_data(bio) && bio_end_sector(bio) > mddev->array_sectors)) return DM_MAPIO_REQUEUE; md_handle_request(mddev, bio); @@ -3528,7 +3533,7 @@ static void raid_status(struct dm_target *ti, status_type_t type, { struct raid_set *rs = ti->private; struct mddev *mddev = &rs->md; - struct r5conf *conf = mddev->private; + struct r5conf *conf = rs_is_raid456(rs) ? mddev->private : NULL; int i, max_nr_stripes = conf ? conf->max_nr_stripes : 0; unsigned long recovery; unsigned int raid_param_cnt = 1; /* at least 1 for chunksize */ @@ -3808,7 +3813,7 @@ static void attempt_restore_of_faulty_devices(struct raid_set *rs) memset(cleared_failed_devices, 0, sizeof(cleared_failed_devices)); - for (i = 0; i < mddev->raid_disks; i++) { + for (i = 0; i < rs->raid_disks; i++) { r = &rs->dev[i].rdev; /* HM FIXME: enhance journal device recovery processing */ if (test_bit(Journal, &r->flags)) @@ -4022,7 +4027,9 @@ static void raid_resume(struct dm_target *ti) * Take this opportunity to check whether any failed * devices are reachable again. */ + mddev_lock_nointr(mddev); attempt_restore_of_faulty_devices(rs); + mddev_unlock(mddev); } if (test_and_clear_bit(RT_FLAG_RS_SUSPENDED, &rs->runtime_flags)) { diff --git a/drivers/md/dm-snap-persistent.c b/drivers/md/dm-snap-persistent.c index 963d3774c93e..247089c2be25 100644 --- a/drivers/md/dm-snap-persistent.c +++ b/drivers/md/dm-snap-persistent.c @@ -613,7 +613,7 @@ static int persistent_read_metadata(struct dm_exception_store *store, chunk_t old, chunk_t new), void *callback_context) { - int r, uninitialized_var(new_snapshot); + int r, new_snapshot; struct pstore *ps = get_info(store); /* diff --git a/drivers/md/dm-snap.c b/drivers/md/dm-snap.c index e902aae685af..d8902d2b6aa6 100644 --- a/drivers/md/dm-snap.c +++ b/drivers/md/dm-snap.c @@ -685,8 +685,10 @@ static void dm_exception_table_exit(struct dm_exception_table *et, for (i = 0; i < size; i++) { slot = et->table + i; - hlist_bl_for_each_entry_safe(ex, pos, n, slot, hash_list) + hlist_bl_for_each_entry_safe(ex, pos, n, slot, hash_list) { kmem_cache_free(mem, ex); + cond_resched(); + } } vfree(et->table); diff --git a/drivers/md/dm-stats.c b/drivers/md/dm-stats.c index ce6d3bce1b7b..0f42f4a7f382 100644 --- a/drivers/md/dm-stats.c +++ b/drivers/md/dm-stats.c @@ -188,7 +188,7 @@ static int dm_stat_in_flight(struct dm_stat_shared *shared) atomic_read(&shared->in_flight[WRITE]); } -void dm_stats_init(struct dm_stats *stats) +int dm_stats_init(struct dm_stats *stats) { int cpu; struct dm_stats_last_position *last; @@ -196,11 +196,16 @@ void dm_stats_init(struct dm_stats *stats) mutex_init(&stats->mutex); INIT_LIST_HEAD(&stats->list); stats->last = alloc_percpu(struct dm_stats_last_position); + if (!stats->last) + return -ENOMEM; + for_each_possible_cpu(cpu) { last = per_cpu_ptr(stats->last, cpu); last->last_sector = (sector_t)ULLONG_MAX; last->last_rw = UINT_MAX; } + + return 0; } void dm_stats_cleanup(struct dm_stats *stats) diff --git a/drivers/md/dm-stats.h b/drivers/md/dm-stats.h index 2ddfae678f32..dcac11fce03b 100644 --- a/drivers/md/dm-stats.h +++ b/drivers/md/dm-stats.h @@ -22,7 +22,7 @@ struct dm_stats_aux { unsigned long long duration_ns; }; -void dm_stats_init(struct dm_stats *st); +int dm_stats_init(struct dm_stats *st); void dm_stats_cleanup(struct dm_stats *st); struct mapped_device; diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c index 81bc36a43b32..fcb9e2775f78 100644 --- a/drivers/md/dm-table.c +++ b/drivers/md/dm-table.c @@ -184,7 +184,12 @@ static int alloc_targets(struct dm_table *t, unsigned int num) int dm_table_create(struct dm_table **result, fmode_t mode, unsigned num_targets, struct mapped_device *md) { - struct dm_table *t = kzalloc(sizeof(*t), GFP_KERNEL); + struct dm_table *t; + + if (num_targets > DM_MAX_TARGETS) + return -EOVERFLOW; + + t = kzalloc(sizeof(*t), GFP_KERNEL); if (!t) return -ENOMEM; @@ -199,7 +204,7 @@ int dm_table_create(struct dm_table **result, fmode_t mode, if (!num_targets) { kfree(t); - return -ENOMEM; + return -EOVERFLOW; } if (alloc_targets(t, num_targets)) { @@ -668,7 +673,7 @@ static int validate_hardware_logical_block_alignment(struct dm_table *table, */ unsigned short remaining = 0; - struct dm_target *uninitialized_var(ti); + struct dm_target *ti; struct queue_limits ti_limits; unsigned i; diff --git a/drivers/md/dm-thin-metadata.c b/drivers/md/dm-thin-metadata.c index a5ed59eafdc5..92e646d597ea 100644 --- a/drivers/md/dm-thin-metadata.c +++ b/drivers/md/dm-thin-metadata.c @@ -701,6 +701,15 @@ static int __open_metadata(struct dm_pool_metadata *pmd) goto bad_cleanup_data_sm; } + /* + * For pool metadata opening process, root setting is redundant + * because it will be set again in __begin_transaction(). But dm + * pool aborting process really needs to get last transaction's + * root to avoid accessing broken btree. + */ + pmd->root = le64_to_cpu(disk_super->data_mapping_root); + pmd->details_root = le64_to_cpu(disk_super->device_details_root); + __setup_btree_details(pmd); dm_bm_unlock(sblock); @@ -753,13 +762,15 @@ static int __create_persistent_data_objects(struct dm_pool_metadata *pmd, bool f return r; } -static void __destroy_persistent_data_objects(struct dm_pool_metadata *pmd) +static void __destroy_persistent_data_objects(struct dm_pool_metadata *pmd, + bool destroy_bm) { dm_sm_destroy(pmd->data_sm); dm_sm_destroy(pmd->metadata_sm); dm_tm_destroy(pmd->nb_tm); dm_tm_destroy(pmd->tm); - dm_block_manager_destroy(pmd->bm); + if (destroy_bm) + dm_block_manager_destroy(pmd->bm); } static int __begin_transaction(struct dm_pool_metadata *pmd) @@ -966,7 +977,7 @@ int dm_pool_metadata_close(struct dm_pool_metadata *pmd) } pmd_write_unlock(pmd); if (!pmd->fail_io) - __destroy_persistent_data_objects(pmd); + __destroy_persistent_data_objects(pmd, true); kfree(pmd); return 0; @@ -1875,19 +1886,52 @@ static void __set_abort_with_changes_flags(struct dm_pool_metadata *pmd) int dm_pool_abort_metadata(struct dm_pool_metadata *pmd) { int r = -EINVAL; + struct dm_block_manager *old_bm = NULL, *new_bm = NULL; + + /* fail_io is double-checked with pmd->root_lock held below */ + if (unlikely(pmd->fail_io)) + return r; + + /* + * Replacement block manager (new_bm) is created and old_bm destroyed outside of + * pmd root_lock to avoid ABBA deadlock that would result (due to life-cycle of + * shrinker associated with the block manager's bufio client vs pmd root_lock). + * - must take shrinker_rwsem without holding pmd->root_lock + */ + new_bm = dm_block_manager_create(pmd->bdev, THIN_METADATA_BLOCK_SIZE << SECTOR_SHIFT, + THIN_MAX_CONCURRENT_LOCKS); pmd_write_lock(pmd); - if (pmd->fail_io) + if (pmd->fail_io) { + pmd_write_unlock(pmd); goto out; + } __set_abort_with_changes_flags(pmd); - __destroy_persistent_data_objects(pmd); - r = __create_persistent_data_objects(pmd, false); + __destroy_persistent_data_objects(pmd, false); + old_bm = pmd->bm; + if (IS_ERR(new_bm)) { + DMERR("could not create block manager during abort"); + pmd->bm = NULL; + r = PTR_ERR(new_bm); + goto out_unlock; + } + + pmd->bm = new_bm; + r = __open_or_format_metadata(pmd, false); + if (r) { + pmd->bm = NULL; + goto out_unlock; + } + new_bm = NULL; +out_unlock: if (r) pmd->fail_io = true; - -out: pmd_write_unlock(pmd); + dm_block_manager_destroy(old_bm); +out: + if (new_bm && !IS_ERR(new_bm)) + dm_block_manager_destroy(new_bm); return r; } @@ -2060,10 +2104,13 @@ int dm_pool_register_metadata_threshold(struct dm_pool_metadata *pmd, dm_sm_threshold_fn fn, void *context) { - int r; + int r = -EINVAL; pmd_write_lock_in_core(pmd); - r = dm_sm_register_threshold_callback(pmd->metadata_sm, threshold, fn, context); + if (!pmd->fail_io) { + r = dm_sm_register_threshold_callback(pmd->metadata_sm, + threshold, fn, context); + } pmd_write_unlock(pmd); return r; diff --git a/drivers/md/dm-thin.c b/drivers/md/dm-thin.c index 1b2c98b43519..a1abdb3467a9 100644 --- a/drivers/md/dm-thin.c +++ b/drivers/md/dm-thin.c @@ -2224,6 +2224,7 @@ static void process_thin_deferred_bios(struct thin_c *tc) throttle_work_update(&pool->throttle); dm_pool_issue_prefetches(pool->pmd); } + cond_resched(); } blk_finish_plug(&plug); } @@ -2307,6 +2308,7 @@ static void process_thin_deferred_cells(struct thin_c *tc) else pool->process_cell(tc, cell); } + cond_resched(); } while (!list_empty(&cells)); } @@ -2931,6 +2933,8 @@ static void __pool_destroy(struct pool *pool) dm_bio_prison_destroy(pool->prison); dm_kcopyd_client_destroy(pool->copier); + cancel_delayed_work_sync(&pool->waker); + cancel_delayed_work_sync(&pool->no_space_timeout); if (pool->wq) destroy_workqueue(pool->wq); @@ -3403,6 +3407,7 @@ static int pool_ctr(struct dm_target *ti, unsigned argc, char **argv) pt->adjusted_pf = pt->requested_pf = pf; bio_init(&pt->flush_bio, NULL, 0); ti->num_flush_bios = 1; + ti->limit_swap_bios = true; /* * Only need to enable discards if the pool should pass @@ -3425,8 +3430,10 @@ static int pool_ctr(struct dm_target *ti, unsigned argc, char **argv) calc_metadata_threshold(pt), metadata_low_callback, pool); - if (r) + if (r) { + ti->error = "Error registering metadata threshold"; goto out_flags_changed; + } pt->callbacks.congested_fn = pool_is_congested; dm_table_add_target_callbacks(ti->table, &pt->callbacks); @@ -3589,23 +3596,31 @@ static int pool_preresume(struct dm_target *ti) */ r = bind_control_target(pool, ti); if (r) - return r; + goto out; dm_pool_register_pre_commit_callback(pool->pmd, metadata_pre_commit_callback, pt); r = maybe_resize_data_dev(ti, &need_commit1); if (r) - return r; + goto out; r = maybe_resize_metadata_dev(ti, &need_commit2); if (r) - return r; + goto out; if (need_commit1 || need_commit2) (void) commit(pool); +out: + /* + * When a thin-pool is PM_FAIL, it cannot be rebuilt if + * bio is in deferred list. Therefore need to return 0 + * to allow pool_resume() to flush IO. + */ + if (r && get_pool_mode(pool) == PM_FAIL) + r = 0; - return 0; + return r; } static void pool_suspend_active_thins(struct pool *pool) @@ -4278,6 +4293,7 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv) goto bad; ti->num_flush_bios = 1; + ti->limit_swap_bios = true; ti->flush_supported = true; ti->per_io_data_size = sizeof(struct dm_thin_endio_hook); diff --git a/drivers/md/dm-verity-fec.c b/drivers/md/dm-verity-fec.c index cea2b3789736..442437e4e03b 100644 --- a/drivers/md/dm-verity-fec.c +++ b/drivers/md/dm-verity-fec.c @@ -24,7 +24,8 @@ bool verity_fec_is_enabled(struct dm_verity *v) */ static inline struct dm_verity_fec_io *fec_io(struct dm_verity_io *io) { - return (struct dm_verity_fec_io *) verity_io_digest_end(io->v, io); + return (struct dm_verity_fec_io *) + ((char *)io + io->v->ti->per_io_data_size - sizeof(struct dm_verity_fec_io)); } /* diff --git a/drivers/md/dm-verity-target.c b/drivers/md/dm-verity-target.c index 9dcdf34b7e32..cebc7828ea8b 100644 --- a/drivers/md/dm-verity-target.c +++ b/drivers/md/dm-verity-target.c @@ -471,13 +471,14 @@ static int verity_verify_io(struct dm_verity_io *io) struct bvec_iter start; unsigned b; struct crypto_wait wait; + struct bio *bio = dm_bio_from_per_bio_data(io, v->ti->per_io_data_size); for (b = 0; b < io->n_blocks; b++) { int r; sector_t cur_block = io->block + b; struct ahash_request *req = verity_io_hash_req(v, io); - if (v->validated_blocks && + if (v->validated_blocks && bio->bi_status == BLK_STS_OK && likely(test_bit(cur_block, v->validated_blocks))) { verity_bv_skip_block(v, io, &io->iter); continue; @@ -525,9 +526,17 @@ static int verity_verify_io(struct dm_verity_io *io) else if (verity_fec_decode(v, io, DM_VERITY_BLOCK_TYPE_DATA, cur_block, NULL, &start) == 0) continue; - else if (verity_handle_err(v, DM_VERITY_BLOCK_TYPE_DATA, - cur_block)) - return -EIO; + else { + if (bio->bi_status) { + /* + * Error correction failed; Just return error + */ + return -EIO; + } + if (verity_handle_err(v, DM_VERITY_BLOCK_TYPE_DATA, + cur_block)) + return -EIO; + } } return 0; @@ -570,7 +579,9 @@ static void verity_end_io(struct bio *bio) struct dm_verity_io *io = bio->bi_private; if (bio->bi_status && - (!verity_fec_is_enabled(io->v) || verity_is_system_shutting_down())) { + (!verity_fec_is_enabled(io->v) || + verity_is_system_shutting_down() || + (bio->bi_opf & REQ_RAHEAD))) { verity_finish_io(io, bio->bi_status); return; } diff --git a/drivers/md/dm-verity.h b/drivers/md/dm-verity.h index 641b9e3a399b..fee7c7a81ce4 100644 --- a/drivers/md/dm-verity.h +++ b/drivers/md/dm-verity.h @@ -73,11 +73,11 @@ struct dm_verity_io { /* original value of bio->bi_end_io */ bio_end_io_t *orig_bi_end_io; + struct bvec_iter iter; + sector_t block; unsigned n_blocks; - struct bvec_iter iter; - struct work_struct work; /* @@ -110,12 +110,6 @@ static inline u8 *verity_io_want_digest(struct dm_verity *v, return (u8 *)(io + 1) + v->ahash_reqsize + v->digest_size; } -static inline u8 *verity_io_digest_end(struct dm_verity *v, - struct dm_verity_io *io) -{ - return verity_io_want_digest(v, io) + v->digest_size; -} - extern int verity_for_bv_block(struct dm_verity *v, struct dm_verity_io *io, struct bvec_iter *iter, int (*process)(struct dm_verity *v, diff --git a/drivers/md/dm-writecache.c b/drivers/md/dm-writecache.c index ec10fda3f24f..961498d8f406 100644 --- a/drivers/md/dm-writecache.c +++ b/drivers/md/dm-writecache.c @@ -20,7 +20,7 @@ #define HIGH_WATERMARK 50 #define LOW_WATERMARK 45 -#define MAX_WRITEBACK_JOBS 0 +#define MAX_WRITEBACK_JOBS min(0x10000000 / PAGE_SIZE, totalram_pages() / 16) #define ENDIO_LATENCY 16 #define WRITEBACK_LATENCY 64 #define AUTOCOMMIT_BLOCKS_SSD 65536 diff --git a/drivers/md/dm.c b/drivers/md/dm.c index 77e28f77c59f..a7724ba45b43 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -263,7 +263,6 @@ out_uevent_exit: static void local_exit(void) { - flush_scheduled_work(); destroy_workqueue(deferred_remove_workqueue); unregister_blkdev(_major, _name); @@ -1995,7 +1994,9 @@ static struct mapped_device *alloc_dev(int minor) if (!md->bdev) goto bad; - dm_stats_init(&md->stats); + r = dm_stats_init(&md->stats); + if (r < 0) + goto bad; /* Populate the mapping, nobody knows we exist yet */ spin_lock(&_minor_lock); @@ -2808,6 +2809,9 @@ static void __dm_internal_suspend(struct mapped_device *md, unsigned suspend_fla static void __dm_internal_resume(struct mapped_device *md) { + int r; + struct dm_table *map; + BUG_ON(!md->internal_suspend_count); if (--md->internal_suspend_count) @@ -2816,12 +2820,23 @@ static void __dm_internal_resume(struct mapped_device *md) if (dm_suspended_md(md)) goto done; /* resume from nested suspend */ - /* - * NOTE: existing callers don't need to call dm_table_resume_targets - * (which may fail -- so best to avoid it for now by passing NULL map) - */ - (void) __dm_resume(md, NULL); - + map = rcu_dereference_protected(md->map, lockdep_is_held(&md->suspend_lock)); + r = __dm_resume(md, map); + if (r) { + /* + * If a preresume method of some target failed, we are in a + * tricky situation. We can't return an error to the caller. We + * can't fake success because then the "resume" and + * "postsuspend" methods would not be paired correctly, and it + * would break various targets, for example it would cause list + * corruption in the "origin" target. + * + * So, we fake normal suspend here, to make sure that the + * "resume" and "postsuspend" methods will be paired correctly. + */ + DMERR("Preresume method failed: %d", r); + set_bit(DMF_SUSPENDED, &md->flags); + } done: clear_bit(DMF_SUSPENDED_INTERNALLY, &md->flags); smp_mb__after_atomic(); @@ -3080,6 +3095,11 @@ static int dm_call_pr(struct block_device *bdev, iterate_devices_callout_fn fn, goto out; ti = dm_table_get_target(table, 0); + if (dm_suspended_md(md)) { + ret = -EAGAIN; + goto out; + } + ret = -EINVAL; if (!ti->type->iterate_devices) goto out; diff --git a/drivers/md/md-bitmap.c b/drivers/md/md-bitmap.c index a95e20c3d0d4..843139447a96 100644 --- a/drivers/md/md-bitmap.c +++ b/drivers/md/md-bitmap.c @@ -54,14 +54,7 @@ __acquires(bitmap->lock) { unsigned char *mappage; - if (page >= bitmap->pages) { - /* This can happen if bitmap_start_sync goes beyond - * End-of-device while looking for a whole page. - * It is harmless. - */ - return -EINVAL; - } - + WARN_ON_ONCE(page >= bitmap->pages); if (bitmap->bp[page].hijacked) /* it's hijacked, don't try to alloc */ return 0; @@ -489,7 +482,7 @@ void md_bitmap_print_sb(struct bitmap *bitmap) sb = kmap_atomic(bitmap->storage.sb_page); pr_debug("%s: bitmap file superblock:\n", bmname(bitmap)); pr_debug(" magic: %08x\n", le32_to_cpu(sb->magic)); - pr_debug(" version: %d\n", le32_to_cpu(sb->version)); + pr_debug(" version: %u\n", le32_to_cpu(sb->version)); pr_debug(" uuid: %08x.%08x.%08x.%08x\n", le32_to_cpu(*(__le32 *)(sb->uuid+0)), le32_to_cpu(*(__le32 *)(sb->uuid+4)), @@ -500,11 +493,11 @@ void md_bitmap_print_sb(struct bitmap *bitmap) pr_debug("events cleared: %llu\n", (unsigned long long) le64_to_cpu(sb->events_cleared)); pr_debug(" state: %08x\n", le32_to_cpu(sb->state)); - pr_debug(" chunksize: %d B\n", le32_to_cpu(sb->chunksize)); - pr_debug(" daemon sleep: %ds\n", le32_to_cpu(sb->daemon_sleep)); + pr_debug(" chunksize: %u B\n", le32_to_cpu(sb->chunksize)); + pr_debug(" daemon sleep: %us\n", le32_to_cpu(sb->daemon_sleep)); pr_debug(" sync size: %llu KB\n", (unsigned long long)le64_to_cpu(sb->sync_size)/2); - pr_debug("max write behind: %d\n", le32_to_cpu(sb->write_behind)); + pr_debug("max write behind: %u\n", le32_to_cpu(sb->write_behind)); kunmap_atomic(sb); } @@ -1369,6 +1362,14 @@ __acquires(bitmap->lock) sector_t csize; int err; + if (page >= bitmap->pages) { + /* + * This can happen if bitmap_start_sync goes beyond + * End-of-device while looking for a whole page or + * user set a huge number to sysfs bitmap_set_bits. + */ + return NULL; + } err = md_bitmap_checkpage(bitmap, page, create, 0); if (bitmap->bp[page].hijacked || @@ -2110,7 +2111,8 @@ int md_bitmap_resize(struct bitmap *bitmap, sector_t blocks, bytes = DIV_ROUND_UP(chunks, 8); if (!bitmap->mddev->bitmap_info.external) bytes += sizeof(bitmap_super_t); - } while (bytes > (space << 9)); + } while (bytes > (space << 9) && (chunkshift + BITMAP_BLOCK_SHIFT) < + (BITS_PER_BYTE * sizeof(((bitmap_super_t *)0)->chunksize) - 1)); } else chunkshift = ffz(~chunksize) - BITMAP_BLOCK_SHIFT; @@ -2155,7 +2157,7 @@ int md_bitmap_resize(struct bitmap *bitmap, sector_t blocks, bitmap->counts.missing_pages = pages; bitmap->counts.chunkshift = chunkshift; bitmap->counts.chunks = chunks; - bitmap->mddev->bitmap_info.chunksize = 1 << (chunkshift + + bitmap->mddev->bitmap_info.chunksize = 1UL << (chunkshift + BITMAP_BLOCK_SHIFT); blocks = min(old_counts.chunks << old_counts.chunkshift, @@ -2181,8 +2183,8 @@ int md_bitmap_resize(struct bitmap *bitmap, sector_t blocks, bitmap->counts.missing_pages = old_counts.pages; bitmap->counts.chunkshift = old_counts.chunkshift; bitmap->counts.chunks = old_counts.chunks; - bitmap->mddev->bitmap_info.chunksize = 1 << (old_counts.chunkshift + - BITMAP_BLOCK_SHIFT); + bitmap->mddev->bitmap_info.chunksize = + 1UL << (old_counts.chunkshift + BITMAP_BLOCK_SHIFT); blocks = old_counts.chunks << old_counts.chunkshift; pr_warn("Could not pre-allocate in-memory bitmap for cluster raid\n"); break; @@ -2200,20 +2202,23 @@ int md_bitmap_resize(struct bitmap *bitmap, sector_t blocks, if (set) { bmc_new = md_bitmap_get_counter(&bitmap->counts, block, &new_blocks, 1); - if (*bmc_new == 0) { - /* need to set on-disk bits too. */ - sector_t end = block + new_blocks; - sector_t start = block >> chunkshift; - start <<= chunkshift; - while (start < end) { - md_bitmap_file_set_bit(bitmap, block); - start += 1 << chunkshift; + if (bmc_new) { + if (*bmc_new == 0) { + /* need to set on-disk bits too. */ + sector_t end = block + new_blocks; + sector_t start = block >> chunkshift; + + start <<= chunkshift; + while (start < end) { + md_bitmap_file_set_bit(bitmap, block); + start += 1 << chunkshift; + } + *bmc_new = 2; + md_bitmap_count_page(&bitmap->counts, block, 1); + md_bitmap_set_pending(&bitmap->counts, block); } - *bmc_new = 2; - md_bitmap_count_page(&bitmap->counts, block, 1); - md_bitmap_set_pending(&bitmap->counts, block); + *bmc_new |= NEEDED_MASK; } - *bmc_new |= NEEDED_MASK; if (new_blocks < old_blocks) old_blocks = new_blocks; } @@ -2475,11 +2480,35 @@ backlog_store(struct mddev *mddev, const char *buf, size_t len) { unsigned long backlog; unsigned long old_mwb = mddev->bitmap_info.max_write_behind; + struct md_rdev *rdev; + bool has_write_mostly = false; int rv = kstrtoul(buf, 10, &backlog); if (rv) return rv; if (backlog > COUNTER_MAX) return -EINVAL; + + rv = mddev_lock(mddev); + if (rv) + return rv; + + /* + * Without write mostly device, it doesn't make sense to set + * backlog for max_write_behind. + */ + rdev_for_each(rdev, mddev) { + if (test_bit(WriteMostly, &rdev->flags)) { + has_write_mostly = true; + break; + } + } + if (!has_write_mostly) { + pr_warn_ratelimited("%s: can't set backlog, no write mostly device available\n", + mdname(mddev)); + mddev_unlock(mddev); + return -EINVAL; + } + mddev->bitmap_info.max_write_behind = backlog; if (!backlog && mddev->wb_info_pool) { /* wb_info_pool is not needed if backlog is zero */ @@ -2494,6 +2523,8 @@ backlog_store(struct mddev *mddev, const char *buf, size_t len) } if (old_mwb != backlog) md_bitmap_update_sb(mddev->bitmap); + + mddev_unlock(mddev); return len; } @@ -2520,6 +2551,9 @@ chunksize_store(struct mddev *mddev, const char *buf, size_t len) if (csize < 512 || !is_power_of_2(csize)) return -EINVAL; + if (BITS_PER_LONG > 32 && csize >= (1ULL << (BITS_PER_BYTE * + sizeof(((bitmap_super_t *)0)->chunksize)))) + return -EOVERFLOW; mddev->bitmap_info.chunksize = csize; return len; } diff --git a/drivers/md/md.c b/drivers/md/md.c index 11fd3b32b562..61c3e8df1b55 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -481,13 +481,14 @@ static void md_end_flush(struct bio *bio) struct md_rdev *rdev = bio->bi_private; struct mddev *mddev = rdev->mddev; + bio_put(bio); + rdev_dec_pending(rdev, mddev); if (atomic_dec_and_test(&mddev->flush_pending)) { /* The pre-request flush has finished */ queue_work(md_wq, &mddev->flush_work); } - bio_put(bio); } static void md_submit_flush_data(struct work_struct *ws); @@ -885,10 +886,12 @@ static void super_written(struct bio *bio) } else clear_bit(LastDev, &rdev->flags); + bio_put(bio); + + rdev_dec_pending(rdev, mddev); + if (atomic_dec_and_test(&mddev->pending_writes)) wake_up(&mddev->sb_wait); - rdev_dec_pending(rdev, mddev); - bio_put(bio); } void md_super_write(struct mddev *mddev, struct md_rdev *rdev, @@ -1095,6 +1098,7 @@ struct super_type { struct md_rdev *refdev, int minor_version); int (*validate_super)(struct mddev *mddev, + struct md_rdev *freshest, struct md_rdev *rdev); void (*sync_super)(struct mddev *mddev, struct md_rdev *rdev); @@ -1233,8 +1237,9 @@ static int super_90_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor /* * validate_super for 0.90.0 + * note: we are not using "freshest" for 0.9 superblock */ -static int super_90_validate(struct mddev *mddev, struct md_rdev *rdev) +static int super_90_validate(struct mddev *mddev, struct md_rdev *freshest, struct md_rdev *rdev) { mdp_disk_t *desc; mdp_super_t *sb = page_address(rdev->sb_page); @@ -1749,7 +1754,7 @@ static int super_1_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor_ return ret; } -static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev) +static int super_1_validate(struct mddev *mddev, struct md_rdev *freshest, struct md_rdev *rdev) { struct mdp_superblock_1 *sb = page_address(rdev->sb_page); __u64 ev1 = le64_to_cpu(sb->events); @@ -1845,13 +1850,15 @@ static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev) } } else if (mddev->pers == NULL) { /* Insist of good event counter while assembling, except for - * spares (which don't need an event count) */ - ++ev1; + * spares (which don't need an event count). + * Similar to mdadm, we allow event counter difference of 1 + * from the freshest device. + */ if (rdev->desc_nr >= 0 && rdev->desc_nr < le32_to_cpu(sb->max_dev) && (le16_to_cpu(sb->dev_roles[rdev->desc_nr]) < MD_DISK_ROLE_MAX || le16_to_cpu(sb->dev_roles[rdev->desc_nr]) == MD_DISK_ROLE_JOURNAL)) - if (ev1 < mddev->events) + if (ev1 + 1 < mddev->events) return -EINVAL; } else if (mddev->bitmap) { /* If adding to array with a bitmap, then we can accept an @@ -1872,8 +1879,38 @@ static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev) rdev->desc_nr >= le32_to_cpu(sb->max_dev)) { role = MD_DISK_ROLE_SPARE; rdev->desc_nr = -1; - } else + } else if (mddev->pers == NULL && freshest && ev1 < mddev->events) { + /* + * If we are assembling, and our event counter is smaller than the + * highest event counter, we cannot trust our superblock about the role. + * It could happen that our rdev was marked as Faulty, and all other + * superblocks were updated with +1 event counter. + * Then, before the next superblock update, which typically happens when + * remove_and_add_spares() removes the device from the array, there was + * a crash or reboot. + * If we allow current rdev without consulting the freshest superblock, + * we could cause data corruption. + * Note that in this case our event counter is smaller by 1 than the + * highest, otherwise, this rdev would not be allowed into array; + * both kernel and mdadm allow event counter difference of 1. + */ + struct mdp_superblock_1 *freshest_sb = page_address(freshest->sb_page); + u32 freshest_max_dev = le32_to_cpu(freshest_sb->max_dev); + + if (rdev->desc_nr >= freshest_max_dev) { + /* this is unexpected, better not proceed */ + pr_warn("md: %s: rdev[%pg]: desc_nr(%d) >= freshest(%pg)->sb->max_dev(%u)\n", + mdname(mddev), rdev->bdev, rdev->desc_nr, + freshest->bdev, freshest_max_dev); + return -EUCLEAN; + } + + role = le16_to_cpu(freshest_sb->dev_roles[rdev->desc_nr]); + pr_debug("md: %s: rdev[%pg]: role=%d(0x%x) according to freshest %pg\n", + mdname(mddev), rdev->bdev, role, role, freshest->bdev); + } else { role = le16_to_cpu(sb->dev_roles[rdev->desc_nr]); + } switch(role) { case MD_DISK_ROLE_SPARE: /* spare */ break; @@ -2777,7 +2814,7 @@ static int add_bound_rdev(struct md_rdev *rdev) * and should be added immediately. */ super_types[mddev->major_version]. - validate_super(mddev, rdev); + validate_super(mddev, NULL/*freshest*/, rdev); if (add_journal) mddev_suspend(mddev); err = mddev->pers->hot_add_disk(mddev, rdev); @@ -3079,6 +3116,9 @@ slot_store(struct md_rdev *rdev, const char *buf, size_t len) err = kstrtouint(buf, 10, (unsigned int *)&slot); if (err < 0) return err; + if (slot < 0) + /* overflow */ + return -ENOSPC; } if (rdev->mddev->pers && slot == -1) { /* Setting 'slot' on an active array requires also @@ -3684,7 +3724,7 @@ static int analyze_sbs(struct mddev *mddev) } super_types[mddev->major_version]. - validate_super(mddev, freshest); + validate_super(mddev, NULL/*freshest*/, freshest); i = 0; rdev_for_each_safe(rdev, tmp, mddev) { @@ -3699,7 +3739,7 @@ static int analyze_sbs(struct mddev *mddev) } if (rdev != freshest) { if (super_types[mddev->major_version]. - validate_super(mddev, rdev)) { + validate_super(mddev, freshest, rdev)) { pr_warn("md: kicking non-fresh %s from array!\n", bdevname(rdev->bdev,b)); md_kick_rdev_from_array(rdev); @@ -3760,8 +3800,9 @@ int strict_strtoul_scaled(const char *cp, unsigned long *res, int scale) static ssize_t safe_delay_show(struct mddev *mddev, char *page) { - int msec = (mddev->safemode_delay*1000)/HZ; - return sprintf(page, "%d.%03d\n", msec/1000, msec%1000); + unsigned int msec = ((unsigned long)mddev->safemode_delay*1000)/HZ; + + return sprintf(page, "%u.%03u\n", msec/1000, msec%1000); } static ssize_t safe_delay_store(struct mddev *mddev, const char *cbuf, size_t len) @@ -3773,7 +3814,7 @@ safe_delay_store(struct mddev *mddev, const char *cbuf, size_t len) return -EINVAL; } - if (strict_strtoul_scaled(cbuf, &msec, 3) < 0) + if (strict_strtoul_scaled(cbuf, &msec, 3) < 0 || msec > UINT_MAX / HZ) return -EINVAL; if (msec == 0) mddev->safemode_delay = 0; @@ -4434,6 +4475,8 @@ max_corrected_read_errors_store(struct mddev *mddev, const char *buf, size_t len rv = kstrtouint(buf, 10, &n); if (rv < 0) return rv; + if (n > INT_MAX) + return -EINVAL; atomic_set(&mddev->max_corr_read_errors, n); return len; } @@ -4734,11 +4777,21 @@ action_store(struct mddev *mddev, const char *page, size_t len) return -EINVAL; err = mddev_lock(mddev); if (!err) { - if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery)) + if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery)) { err = -EBUSY; - else { + } else if (mddev->reshape_position == MaxSector || + mddev->pers->check_reshape == NULL || + mddev->pers->check_reshape(mddev)) { clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery); err = mddev->pers->start_reshape(mddev); + } else { + /* + * If reshape is still in progress, and + * md_check_recovery() can continue to reshape, + * don't restart reshape because data can be + * corrupted for raid456. + */ + clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery); } mddev_unlock(mddev); } @@ -6094,6 +6147,7 @@ void md_stop(struct mddev *mddev) /* stop the array and free an attached data structures. * This is called from dm-raid */ + __md_stop_writes(mddev); __md_stop(mddev); bioset_exit(&mddev->bio_set); bioset_exit(&mddev->sync_set); @@ -6590,7 +6644,7 @@ static int add_new_disk(struct mddev *mddev, mdu_disk_info_t *info) rdev->saved_raid_disk = rdev->raid_disk; } else super_types[mddev->major_version]. - validate_super(mddev, rdev); + validate_super(mddev, NULL/*freshest*/, rdev); if ((info->state & (1<<MD_DISK_SYNC)) && rdev->raid_disk != info->raid_disk) { /* This was a hot-add request, but events doesn't diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c index 0ead5a7887f1..7f80e86459b1 100644 --- a/drivers/md/raid0.c +++ b/drivers/md/raid0.c @@ -63,8 +63,8 @@ static void dump_zones(struct mddev *mddev) int len = 0; for (k = 0; k < conf->strip_zone[j].nb_dev; k++) - len += snprintf(line+len, 200-len, "%s%s", k?"/":"", - bdevname(conf->devlist[j*raid_disks + len += scnprintf(line+len, 200-len, "%s%s", k?"/":"", + bdevname(conf->devlist[j*raid_disks + k]->bdev, b)); pr_debug("md: zone%d=[%s]\n", j, line); @@ -289,6 +289,18 @@ static int create_strip_zones(struct mddev *mddev, struct r0conf **private_conf) goto abort; } + if (conf->layout == RAID0_ORIG_LAYOUT) { + for (i = 1; i < conf->nr_strip_zones; i++) { + sector_t first_sector = conf->strip_zone[i-1].zone_end; + + sector_div(first_sector, mddev->chunk_sectors); + zone = conf->strip_zone + i; + /* disk_shift is first disk index used in the zone */ + zone->disk_shift = sector_div(first_sector, + zone->nb_dev); + } + } + pr_debug("md/raid0:%s: done.\n", mdname(mddev)); *private_conf = conf; @@ -475,6 +487,20 @@ static inline int is_io_in_chunk_boundary(struct mddev *mddev, } } +/* + * Convert disk_index to the disk order in which it is read/written. + * For example, if we have 4 disks, they are numbered 0,1,2,3. If we + * write the disks starting at disk 3, then the read/write order would + * be disk 3, then 0, then 1, and then disk 2 and we want map_disk_shift() + * to map the disks as follows 0,1,2,3 => 1,2,3,0. So disk 0 would map + * to 1, 1 to 2, 2 to 3, and 3 to 0. That way we can compare disks in + * that 'output' space to understand the read/write disk ordering. + */ +static int map_disk_shift(int disk_index, int num_disks, int disk_shift) +{ + return ((disk_index + num_disks - disk_shift) % num_disks); +} + static void raid0_handle_discard(struct mddev *mddev, struct bio *bio) { struct r0conf *conf = mddev->private; @@ -488,7 +514,9 @@ static void raid0_handle_discard(struct mddev *mddev, struct bio *bio) sector_t end_disk_offset; unsigned int end_disk_index; unsigned int disk; + sector_t orig_start, orig_end; + orig_start = start; zone = find_zone(conf, &start); if (bio_end_sector(bio) > zone->zone_end) { @@ -502,6 +530,7 @@ static void raid0_handle_discard(struct mddev *mddev, struct bio *bio) } else end = bio_end_sector(bio); + orig_end = end; if (zone != conf->strip_zone) end = end - zone[-1].zone_end; @@ -513,13 +542,26 @@ static void raid0_handle_discard(struct mddev *mddev, struct bio *bio) last_stripe_index = end; sector_div(last_stripe_index, stripe_size); - start_disk_index = (int)(start - first_stripe_index * stripe_size) / - mddev->chunk_sectors; + /* In the first zone the original and alternate layouts are the same */ + if ((conf->layout == RAID0_ORIG_LAYOUT) && (zone != conf->strip_zone)) { + sector_div(orig_start, mddev->chunk_sectors); + start_disk_index = sector_div(orig_start, zone->nb_dev); + start_disk_index = map_disk_shift(start_disk_index, + zone->nb_dev, + zone->disk_shift); + sector_div(orig_end, mddev->chunk_sectors); + end_disk_index = sector_div(orig_end, zone->nb_dev); + end_disk_index = map_disk_shift(end_disk_index, + zone->nb_dev, zone->disk_shift); + } else { + start_disk_index = (int)(start - first_stripe_index * stripe_size) / + mddev->chunk_sectors; + end_disk_index = (int)(end - last_stripe_index * stripe_size) / + mddev->chunk_sectors; + } start_disk_offset = ((int)(start - first_stripe_index * stripe_size) % mddev->chunk_sectors) + first_stripe_index * mddev->chunk_sectors; - end_disk_index = (int)(end - last_stripe_index * stripe_size) / - mddev->chunk_sectors; end_disk_offset = ((int)(end - last_stripe_index * stripe_size) % mddev->chunk_sectors) + last_stripe_index * mddev->chunk_sectors; @@ -528,18 +570,22 @@ static void raid0_handle_discard(struct mddev *mddev, struct bio *bio) sector_t dev_start, dev_end; struct bio *discard_bio = NULL; struct md_rdev *rdev; + int compare_disk; + + compare_disk = map_disk_shift(disk, zone->nb_dev, + zone->disk_shift); - if (disk < start_disk_index) + if (compare_disk < start_disk_index) dev_start = (first_stripe_index + 1) * mddev->chunk_sectors; - else if (disk > start_disk_index) + else if (compare_disk > start_disk_index) dev_start = first_stripe_index * mddev->chunk_sectors; else dev_start = start_disk_offset; - if (disk < end_disk_index) + if (compare_disk < end_disk_index) dev_end = (last_stripe_index + 1) * mddev->chunk_sectors; - else if (disk > end_disk_index) + else if (compare_disk > end_disk_index) dev_end = last_stripe_index * mddev->chunk_sectors; else dev_end = end_disk_offset; diff --git a/drivers/md/raid0.h b/drivers/md/raid0.h index 3816e5477db1..8cc761ca7423 100644 --- a/drivers/md/raid0.h +++ b/drivers/md/raid0.h @@ -6,6 +6,7 @@ struct strip_zone { sector_t zone_end; /* Start of the next zone (in sectors) */ sector_t dev_start; /* Zone offset in real dev (in sectors) */ int nb_dev; /* # of devices attached to the zone */ + int disk_shift; /* start disk for the original layout */ }; /* Linux 3.14 (20d0189b101) made an unintended change to diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index e87184645c54..c40237cfdcb0 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -1810,6 +1810,9 @@ static int raid1_remove_disk(struct mddev *mddev, struct md_rdev *rdev) int number = rdev->raid_disk; struct raid1_info *p = conf->mirrors + number; + if (unlikely(number >= conf->raid_disks)) + goto abort; + if (rdev != p->rdev) p = conf->mirrors + conf->raid_disks + number; @@ -3132,6 +3135,7 @@ static int raid1_run(struct mddev *mddev) * RAID1 needs at least one disk in active */ if (conf->raid_disks - mddev->degraded < 1) { + md_unregister_thread(&conf->thread); ret = -EINVAL; goto abort; } diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index deddabfb07d7..3983d5c8b5cd 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -751,8 +751,16 @@ static struct md_rdev *read_balance(struct r10conf *conf, disk = r10_bio->devs[slot].devnum; rdev = rcu_dereference(conf->mirrors[disk].replacement); if (rdev == NULL || test_bit(Faulty, &rdev->flags) || - r10_bio->devs[slot].addr + sectors > rdev->recovery_offset) + r10_bio->devs[slot].addr + sectors > + rdev->recovery_offset) { + /* + * Read replacement first to prevent reading both rdev + * and replacement as NULL during replacement replace + * rdev. + */ + smp_mb(); rdev = rcu_dereference(conf->mirrors[disk].rdev); + } if (rdev == NULL || test_bit(Faulty, &rdev->flags)) continue; @@ -919,6 +927,7 @@ static void flush_pending_writes(struct r10conf *conf) else generic_make_request(bio); bio = next; + cond_resched(); } blk_finish_plug(&plug); } else @@ -1104,6 +1113,7 @@ static void raid10_unplug(struct blk_plug_cb *cb, bool from_schedule) else generic_make_request(bio); bio = next; + cond_resched(); } kfree(plug); } @@ -1363,9 +1373,15 @@ retry_write: for (i = 0; i < conf->copies; i++) { int d = r10_bio->devs[i].devnum; - struct md_rdev *rdev = rcu_dereference(conf->mirrors[d].rdev); - struct md_rdev *rrdev = rcu_dereference( - conf->mirrors[d].replacement); + struct md_rdev *rdev, *rrdev; + + rrdev = rcu_dereference(conf->mirrors[d].replacement); + /* + * Read replacement first to prevent reading both rdev and + * replacement as NULL during replacement replace rdev. + */ + smp_mb(); + rdev = rcu_dereference(conf->mirrors[d].rdev); if (rdev == rrdev) rrdev = NULL; if (rdev && unlikely(test_bit(Blocked, &rdev->flags))) { @@ -1826,9 +1842,12 @@ static int raid10_remove_disk(struct mddev *mddev, struct md_rdev *rdev) int err = 0; int number = rdev->raid_disk; struct md_rdev **rdevp; - struct raid10_info *p = conf->mirrors + number; + struct raid10_info *p; print_conf(conf); + if (unlikely(number >= mddev->raid_disks)) + return 0; + p = conf->mirrors + number; if (rdev == p->rdev) rdevp = &p->rdev; else if (rdev == p->replacement) @@ -2226,11 +2245,22 @@ static void recovery_request_write(struct mddev *mddev, struct r10bio *r10_bio) { struct r10conf *conf = mddev->private; int d; - struct bio *wbio, *wbio2; + struct bio *wbio = r10_bio->devs[1].bio; + struct 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 (!test_bit(R10BIO_Uptodate, &r10_bio->state)) { fix_recovery_read_error(r10_bio); - end_sync_request(r10_bio); + if (wbio->bi_end_io) + end_sync_request(r10_bio); + if (wbio2) + end_sync_request(r10_bio); return; } @@ -2239,14 +2269,6 @@ static void recovery_request_write(struct mddev *mddev, struct r10bio *r10_bio) * and submit the write request */ 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)); @@ -2914,10 +2936,6 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr, sector_t chunk_mask = conf->geo.chunk_mask; int page_idx = 0; - if (!mempool_initialized(&conf->r10buf_pool)) - if (init_resync(conf)) - return 0; - /* * Allow skipping a full rebuild for incremental assembly * of a clean array, like RAID1 does. @@ -2933,6 +2951,10 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr, return mddev->dev_sectors - sector_nr; } + if (!mempool_initialized(&conf->r10buf_pool)) + if (init_resync(conf)) + return 0; + skipped: max_sector = mddev->dev_sectors; if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery) || @@ -3048,7 +3070,6 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr, int must_sync; int any_working; int need_recover = 0; - int need_replace = 0; struct raid10_info *mirror = &conf->mirrors[i]; struct md_rdev *mrdev, *mreplace; @@ -3060,11 +3081,10 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr, !test_bit(Faulty, &mrdev->flags) && !test_bit(In_sync, &mrdev->flags)) need_recover = 1; - if (mreplace != NULL && - !test_bit(Faulty, &mreplace->flags)) - need_replace = 1; + if (mreplace && test_bit(Faulty, &mreplace->flags)) + mreplace = NULL; - if (!need_recover && !need_replace) { + if (!need_recover && !mreplace) { rcu_read_unlock(); continue; } @@ -3080,8 +3100,6 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr, rcu_read_unlock(); continue; } - if (mreplace && test_bit(Faulty, &mreplace->flags)) - mreplace = NULL; /* Unless we are doing a full sync, or a replacement * we only need to recover the block if it is set in * the bitmap @@ -3204,11 +3222,11 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr, bio = r10_bio->devs[1].repl_bio; if (bio) bio->bi_end_io = NULL; - /* Note: if need_replace, then bio + /* Note: if replace is not NULL, then bio * cannot be NULL as r10buf_pool_alloc will * have allocated it. */ - if (!need_replace) + if (!mreplace) break; bio->bi_next = biolist; biolist = bio; @@ -3629,6 +3647,20 @@ static int setup_geo(struct geom *geo, struct mddev *mddev, enum geo_type new) return nc*fc; } +static void raid10_free_conf(struct r10conf *conf) +{ + if (!conf) + return; + + mempool_exit(&conf->r10bio_pool); + kfree(conf->mirrors); + kfree(conf->mirrors_old); + kfree(conf->mirrors_new); + safe_put_page(conf->tmppage); + bioset_exit(&conf->bio_split); + kfree(conf); +} + static struct r10conf *setup_conf(struct mddev *mddev) { struct r10conf *conf = NULL; @@ -3711,20 +3743,24 @@ static struct r10conf *setup_conf(struct mddev *mddev) return conf; out: - if (conf) { - mempool_exit(&conf->r10bio_pool); - kfree(conf->mirrors); - safe_put_page(conf->tmppage); - bioset_exit(&conf->bio_split); - kfree(conf); - } + raid10_free_conf(conf); return ERR_PTR(err); } +static void raid10_set_io_opt(struct r10conf *conf) +{ + int raid_disks = conf->geo.raid_disks; + + if (!(conf->geo.raid_disks % conf->geo.near_copies)) + raid_disks /= conf->geo.near_copies; + blk_queue_io_opt(conf->mddev->queue, (conf->mddev->chunk_sectors << 9) * + raid_disks); +} + static int raid10_run(struct mddev *mddev) { struct r10conf *conf; - int i, disk_idx, chunk_size; + int i, disk_idx; struct raid10_info *disk; struct md_rdev *rdev; sector_t size; @@ -3745,6 +3781,9 @@ static int raid10_run(struct mddev *mddev) if (!conf) goto out; + mddev->thread = conf->thread; + conf->thread = NULL; + if (mddev_is_clustered(conf->mddev)) { int fc, fo; @@ -3757,21 +3796,13 @@ static int raid10_run(struct mddev *mddev) } } - mddev->thread = conf->thread; - conf->thread = NULL; - - chunk_size = mddev->chunk_sectors << 9; if (mddev->queue) { blk_queue_max_discard_sectors(mddev->queue, mddev->chunk_sectors); blk_queue_max_write_same_sectors(mddev->queue, 0); blk_queue_max_write_zeroes_sectors(mddev->queue, 0); - blk_queue_io_min(mddev->queue, chunk_size); - if (conf->geo.raid_disks % conf->geo.near_copies) - blk_queue_io_opt(mddev->queue, chunk_size * conf->geo.raid_disks); - else - blk_queue_io_opt(mddev->queue, chunk_size * - (conf->geo.raid_disks / conf->geo.near_copies)); + blk_queue_io_min(mddev->queue, mddev->chunk_sectors << 9); + raid10_set_io_opt(conf); } rdev_for_each(rdev, mddev) { @@ -3931,10 +3962,7 @@ static int raid10_run(struct mddev *mddev) out_free_conf: md_unregister_thread(&mddev->thread); - mempool_exit(&conf->r10bio_pool); - safe_put_page(conf->tmppage); - kfree(conf->mirrors); - kfree(conf); + raid10_free_conf(conf); mddev->private = NULL; out: return -EIO; @@ -3942,15 +3970,7 @@ out: static void raid10_free(struct mddev *mddev, void *priv) { - struct r10conf *conf = priv; - - mempool_exit(&conf->r10bio_pool); - safe_put_page(conf->tmppage); - kfree(conf->mirrors); - kfree(conf->mirrors_old); - kfree(conf->mirrors_new); - bioset_exit(&conf->bio_split); - kfree(conf); + raid10_free_conf(priv); } static void raid10_quiesce(struct mddev *mddev, int quiesce) @@ -4745,6 +4765,7 @@ static void end_reshape(struct r10conf *conf) stripe /= conf->geo.near_copies; if (conf->mddev->queue->backing_dev_info->ra_pages < 2 * stripe) conf->mddev->queue->backing_dev_info->ra_pages = 2 * stripe; + raid10_set_io_opt(conf); } conf->fullsync = 0; } diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 474cf6abefea..f3d60c4b34b8 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -36,6 +36,7 @@ */ #include <linux/blkdev.h> +#include <linux/delay.h> #include <linux/kthread.h> #include <linux/raid/pq.h> #include <linux/async_tx.h> @@ -2598,7 +2599,7 @@ static void raid5_end_write_request(struct bio *bi) struct stripe_head *sh = bi->bi_private; struct r5conf *conf = sh->raid_conf; int disks = sh->disks, i; - struct md_rdev *uninitialized_var(rdev); + struct md_rdev *rdev; sector_t first_bad; int bad_sectors; int replacement = 0; @@ -2666,10 +2667,10 @@ static void raid5_end_write_request(struct bio *bi) if (!test_and_clear_bit(R5_DOUBLE_LOCKED, &sh->dev[i].flags)) clear_bit(R5_LOCKED, &sh->dev[i].flags); set_bit(STRIPE_HANDLE, &sh->state); - raid5_release_stripe(sh); if (sh->batch_head && sh != sh->batch_head) raid5_release_stripe(sh->batch_head); + raid5_release_stripe(sh); } static void raid5_error(struct mddev *mddev, struct md_rdev *rdev) @@ -3728,7 +3729,7 @@ static void handle_stripe_fill(struct stripe_head *sh, * back cache (prexor with orig_page, and then xor with * page) in the read path */ - if (s->injournal && s->failed) { + if (s->to_read && s->injournal && s->failed) { if (test_bit(STRIPE_R5C_CACHING, &sh->state)) r5c_make_stripe_write_out(sh); goto out; @@ -6334,7 +6335,18 @@ static void raid5d(struct md_thread *thread) spin_unlock_irq(&conf->device_lock); md_check_recovery(mddev); spin_lock_irq(&conf->device_lock); + + /* + * Waiting on MD_SB_CHANGE_PENDING below may deadlock + * seeing md_check_recovery() is needed to clear + * the flag when using mdmon. + */ + continue; } + + wait_event_lock_irq(mddev->sb_wait, + !test_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags), + conf->device_lock); } pr_debug("%d stripes handled\n", handled); @@ -7147,6 +7159,12 @@ static int only_parity(int raid_disk, int algo, int raid_disks, int max_degraded return 0; } +static void raid5_set_io_opt(struct r5conf *conf) +{ + blk_queue_io_opt(conf->mddev->queue, (conf->chunk_sectors << 9) * + (conf->raid_disks - conf->max_degraded)); +} + static int raid5_run(struct mddev *mddev) { struct r5conf *conf; @@ -7436,8 +7454,7 @@ static int raid5_run(struct mddev *mddev) chunk_size = mddev->chunk_sectors << 9; blk_queue_io_min(mddev->queue, chunk_size); - blk_queue_io_opt(mddev->queue, chunk_size * - (conf->raid_disks - conf->max_degraded)); + raid5_set_io_opt(conf); mddev->queue->limits.raid_partial_stripes_expensive = 1; /* * We can only discard a whole stripe. It doesn't make sense to @@ -8031,6 +8048,7 @@ static void end_reshape(struct r5conf *conf) / PAGE_SIZE); if (conf->mddev->queue->backing_dev_info->ra_pages < 2 * stripe) conf->mddev->queue->backing_dev_info->ra_pages = 2 * stripe; + raid5_set_io_opt(conf); } } } |