summaryrefslogtreecommitdiff
path: root/drivers/md
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/md')
-rw-r--r--drivers/md/Kconfig1
-rw-r--r--drivers/md/bitmap.c13
-rw-r--r--drivers/md/dm-cache-metadata.c104
-rw-r--r--drivers/md/dm-cache-target.c89
-rw-r--r--drivers/md/dm-thin.c35
-rw-r--r--drivers/md/dm.c11
-rw-r--r--drivers/md/raid5.c5
7 files changed, 203 insertions, 55 deletions
diff --git a/drivers/md/Kconfig b/drivers/md/Kconfig
index 5bdedf6df153..c355a226a024 100644
--- a/drivers/md/Kconfig
+++ b/drivers/md/Kconfig
@@ -5,6 +5,7 @@
menuconfig MD
bool "Multiple devices driver support (RAID and LVM)"
depends on BLOCK
+ select SRCU
help
Support multiple physical spindles through a single logical device.
Required for RAID and logical volume management.
diff --git a/drivers/md/bitmap.c b/drivers/md/bitmap.c
index da3604e73e8a..1695ee5f3ffc 100644
--- a/drivers/md/bitmap.c
+++ b/drivers/md/bitmap.c
@@ -72,6 +72,19 @@ __acquires(bitmap->lock)
/* this page has not been allocated yet */
spin_unlock_irq(&bitmap->lock);
+ /* It is possible that this is being called inside a
+ * prepare_to_wait/finish_wait loop from raid5c:make_request().
+ * In general it is not permitted to sleep in that context as it
+ * can cause the loop to spin freely.
+ * That doesn't apply here as we can only reach this point
+ * once with any loop.
+ * When this function completes, either bp[page].map or
+ * bp[page].hijacked. In either case, this function will
+ * abort before getting to this point again. So there is
+ * no risk of a free-spin, and so it is safe to assert
+ * that sleeping here is allowed.
+ */
+ sched_annotate_sleep();
mappage = kzalloc(PAGE_SIZE, GFP_NOIO);
spin_lock_irq(&bitmap->lock);
diff --git a/drivers/md/dm-cache-metadata.c b/drivers/md/dm-cache-metadata.c
index 9fc616c2755e..c1c010498a21 100644
--- a/drivers/md/dm-cache-metadata.c
+++ b/drivers/md/dm-cache-metadata.c
@@ -94,6 +94,9 @@ struct cache_disk_superblock {
} __packed;
struct dm_cache_metadata {
+ atomic_t ref_count;
+ struct list_head list;
+
struct block_device *bdev;
struct dm_block_manager *bm;
struct dm_space_map *metadata_sm;
@@ -669,10 +672,10 @@ static void unpack_value(__le64 value_le, dm_oblock_t *block, unsigned *flags)
/*----------------------------------------------------------------*/
-struct dm_cache_metadata *dm_cache_metadata_open(struct block_device *bdev,
- sector_t data_block_size,
- bool may_format_device,
- size_t policy_hint_size)
+static struct dm_cache_metadata *metadata_open(struct block_device *bdev,
+ sector_t data_block_size,
+ bool may_format_device,
+ size_t policy_hint_size)
{
int r;
struct dm_cache_metadata *cmd;
@@ -680,9 +683,10 @@ struct dm_cache_metadata *dm_cache_metadata_open(struct block_device *bdev,
cmd = kzalloc(sizeof(*cmd), GFP_KERNEL);
if (!cmd) {
DMERR("could not allocate metadata struct");
- return NULL;
+ return ERR_PTR(-ENOMEM);
}
+ atomic_set(&cmd->ref_count, 1);
init_rwsem(&cmd->root_lock);
cmd->bdev = bdev;
cmd->data_block_size = data_block_size;
@@ -705,10 +709,96 @@ struct dm_cache_metadata *dm_cache_metadata_open(struct block_device *bdev,
return cmd;
}
+/*
+ * We keep a little list of ref counted metadata objects to prevent two
+ * different target instances creating separate bufio instances. This is
+ * an issue if a table is reloaded before the suspend.
+ */
+static DEFINE_MUTEX(table_lock);
+static LIST_HEAD(table);
+
+static struct dm_cache_metadata *lookup(struct block_device *bdev)
+{
+ struct dm_cache_metadata *cmd;
+
+ list_for_each_entry(cmd, &table, list)
+ if (cmd->bdev == bdev) {
+ atomic_inc(&cmd->ref_count);
+ return cmd;
+ }
+
+ return NULL;
+}
+
+static struct dm_cache_metadata *lookup_or_open(struct block_device *bdev,
+ sector_t data_block_size,
+ bool may_format_device,
+ size_t policy_hint_size)
+{
+ struct dm_cache_metadata *cmd, *cmd2;
+
+ mutex_lock(&table_lock);
+ cmd = lookup(bdev);
+ mutex_unlock(&table_lock);
+
+ if (cmd)
+ return cmd;
+
+ cmd = metadata_open(bdev, data_block_size, may_format_device, policy_hint_size);
+ if (!IS_ERR(cmd)) {
+ mutex_lock(&table_lock);
+ cmd2 = lookup(bdev);
+ if (cmd2) {
+ mutex_unlock(&table_lock);
+ __destroy_persistent_data_objects(cmd);
+ kfree(cmd);
+ return cmd2;
+ }
+ list_add(&cmd->list, &table);
+ mutex_unlock(&table_lock);
+ }
+
+ return cmd;
+}
+
+static bool same_params(struct dm_cache_metadata *cmd, sector_t data_block_size)
+{
+ if (cmd->data_block_size != data_block_size) {
+ DMERR("data_block_size (%llu) different from that in metadata (%llu)\n",
+ (unsigned long long) data_block_size,
+ (unsigned long long) cmd->data_block_size);
+ return false;
+ }
+
+ return true;
+}
+
+struct dm_cache_metadata *dm_cache_metadata_open(struct block_device *bdev,
+ sector_t data_block_size,
+ bool may_format_device,
+ size_t policy_hint_size)
+{
+ struct dm_cache_metadata *cmd = lookup_or_open(bdev, data_block_size,
+ may_format_device, policy_hint_size);
+
+ if (!IS_ERR(cmd) && !same_params(cmd, data_block_size)) {
+ dm_cache_metadata_close(cmd);
+ return ERR_PTR(-EINVAL);
+ }
+
+ return cmd;
+}
+
void dm_cache_metadata_close(struct dm_cache_metadata *cmd)
{
- __destroy_persistent_data_objects(cmd);
- kfree(cmd);
+ if (atomic_dec_and_test(&cmd->ref_count)) {
+ mutex_lock(&table_lock);
+ list_del(&cmd->list);
+ mutex_unlock(&table_lock);
+
+ __destroy_persistent_data_objects(cmd);
+ kfree(cmd);
+ }
}
/*
diff --git a/drivers/md/dm-cache-target.c b/drivers/md/dm-cache-target.c
index 1e96d7889f51..e1650539cc2f 100644
--- a/drivers/md/dm-cache-target.c
+++ b/drivers/md/dm-cache-target.c
@@ -221,7 +221,13 @@ struct cache {
struct list_head need_commit_migrations;
sector_t migration_threshold;
wait_queue_head_t migration_wait;
- atomic_t nr_migrations;
+ atomic_t nr_allocated_migrations;
+
+ /*
+ * The number of in flight migrations that are performing
+ * background io. eg, promotion, writeback.
+ */
+ atomic_t nr_io_migrations;
wait_queue_head_t quiescing_wait;
atomic_t quiescing;
@@ -258,7 +264,6 @@ struct cache {
struct dm_deferred_set *all_io_ds;
mempool_t *migration_pool;
- struct dm_cache_migration *next_migration;
struct dm_cache_policy *policy;
unsigned policy_nr_args;
@@ -350,10 +355,31 @@ static void free_prison_cell(struct cache *cache, struct dm_bio_prison_cell *cel
dm_bio_prison_free_cell(cache->prison, cell);
}
+static struct dm_cache_migration *alloc_migration(struct cache *cache)
+{
+ struct dm_cache_migration *mg;
+
+ mg = mempool_alloc(cache->migration_pool, GFP_NOWAIT);
+ if (mg) {
+ mg->cache = cache;
+ atomic_inc(&mg->cache->nr_allocated_migrations);
+ }
+
+ return mg;
+}
+
+static void free_migration(struct dm_cache_migration *mg)
+{
+ if (atomic_dec_and_test(&mg->cache->nr_allocated_migrations))
+ wake_up(&mg->cache->migration_wait);
+
+ mempool_free(mg, mg->cache->migration_pool);
+}
+
static int prealloc_data_structs(struct cache *cache, struct prealloc *p)
{
if (!p->mg) {
- p->mg = mempool_alloc(cache->migration_pool, GFP_NOWAIT);
+ p->mg = alloc_migration(cache);
if (!p->mg)
return -ENOMEM;
}
@@ -382,7 +408,7 @@ static void prealloc_free_structs(struct cache *cache, struct prealloc *p)
free_prison_cell(cache, p->cell1);
if (p->mg)
- mempool_free(p->mg, cache->migration_pool);
+ free_migration(p->mg);
}
static struct dm_cache_migration *prealloc_get_migration(struct prealloc *p)
@@ -854,24 +880,14 @@ static void remap_to_origin_then_cache(struct cache *cache, struct bio *bio,
* Migration covers moving data from the origin device to the cache, or
* vice versa.
*--------------------------------------------------------------*/
-static void free_migration(struct dm_cache_migration *mg)
-{
- mempool_free(mg, mg->cache->migration_pool);
-}
-
-static void inc_nr_migrations(struct cache *cache)
+static void inc_io_migrations(struct cache *cache)
{
- atomic_inc(&cache->nr_migrations);
+ atomic_inc(&cache->nr_io_migrations);
}
-static void dec_nr_migrations(struct cache *cache)
+static void dec_io_migrations(struct cache *cache)
{
- atomic_dec(&cache->nr_migrations);
-
- /*
- * Wake the worker in case we're suspending the target.
- */
- wake_up(&cache->migration_wait);
+ atomic_dec(&cache->nr_io_migrations);
}
static void __cell_defer(struct cache *cache, struct dm_bio_prison_cell *cell,
@@ -894,11 +910,10 @@ static void cell_defer(struct cache *cache, struct dm_bio_prison_cell *cell,
wake_worker(cache);
}
-static void cleanup_migration(struct dm_cache_migration *mg)
+static void free_io_migration(struct dm_cache_migration *mg)
{
- struct cache *cache = mg->cache;
+ dec_io_migrations(mg->cache);
free_migration(mg);
- dec_nr_migrations(cache);
}
static void migration_failure(struct dm_cache_migration *mg)
@@ -923,7 +938,7 @@ static void migration_failure(struct dm_cache_migration *mg)
cell_defer(cache, mg->new_ocell, true);
}
- cleanup_migration(mg);
+ free_io_migration(mg);
}
static void migration_success_pre_commit(struct dm_cache_migration *mg)
@@ -934,7 +949,7 @@ static void migration_success_pre_commit(struct dm_cache_migration *mg)
if (mg->writeback) {
clear_dirty(cache, mg->old_oblock, mg->cblock);
cell_defer(cache, mg->old_ocell, false);
- cleanup_migration(mg);
+ free_io_migration(mg);
return;
} else if (mg->demote) {
@@ -944,14 +959,14 @@ static void migration_success_pre_commit(struct dm_cache_migration *mg)
mg->old_oblock);
if (mg->promote)
cell_defer(cache, mg->new_ocell, true);
- cleanup_migration(mg);
+ free_io_migration(mg);
return;
}
} else {
if (dm_cache_insert_mapping(cache->cmd, mg->cblock, mg->new_oblock)) {
DMWARN_LIMIT("promotion failed; couldn't update on disk metadata");
policy_remove_mapping(cache->policy, mg->new_oblock);
- cleanup_migration(mg);
+ free_io_migration(mg);
return;
}
}
@@ -984,7 +999,7 @@ static void migration_success_post_commit(struct dm_cache_migration *mg)
} else {
if (mg->invalidate)
policy_remove_mapping(cache->policy, mg->old_oblock);
- cleanup_migration(mg);
+ free_io_migration(mg);
}
} else {
@@ -999,7 +1014,7 @@ static void migration_success_post_commit(struct dm_cache_migration *mg)
bio_endio(mg->new_ocell->holder, 0);
cell_defer(cache, mg->new_ocell, false);
}
- cleanup_migration(mg);
+ free_io_migration(mg);
}
}
@@ -1251,7 +1266,7 @@ static void promote(struct cache *cache, struct prealloc *structs,
mg->new_ocell = cell;
mg->start_jiffies = jiffies;
- inc_nr_migrations(cache);
+ inc_io_migrations(cache);
quiesce_migration(mg);
}
@@ -1275,7 +1290,7 @@ static void writeback(struct cache *cache, struct prealloc *structs,
mg->new_ocell = NULL;
mg->start_jiffies = jiffies;
- inc_nr_migrations(cache);
+ inc_io_migrations(cache);
quiesce_migration(mg);
}
@@ -1302,7 +1317,7 @@ static void demote_then_promote(struct cache *cache, struct prealloc *structs,
mg->new_ocell = new_ocell;
mg->start_jiffies = jiffies;
- inc_nr_migrations(cache);
+ inc_io_migrations(cache);
quiesce_migration(mg);
}
@@ -1330,7 +1345,7 @@ static void invalidate(struct cache *cache, struct prealloc *structs,
mg->new_ocell = NULL;
mg->start_jiffies = jiffies;
- inc_nr_migrations(cache);
+ inc_io_migrations(cache);
quiesce_migration(mg);
}
@@ -1412,7 +1427,7 @@ static void process_discard_bio(struct cache *cache, struct prealloc *structs,
static bool spare_migration_bandwidth(struct cache *cache)
{
- sector_t current_volume = (atomic_read(&cache->nr_migrations) + 1) *
+ sector_t current_volume = (atomic_read(&cache->nr_io_migrations) + 1) *
cache->sectors_per_block;
return current_volume < cache->migration_threshold;
}
@@ -1764,7 +1779,7 @@ static void stop_quiescing(struct cache *cache)
static void wait_for_migrations(struct cache *cache)
{
- wait_event(cache->migration_wait, !atomic_read(&cache->nr_migrations));
+ wait_event(cache->migration_wait, !atomic_read(&cache->nr_allocated_migrations));
}
static void stop_worker(struct cache *cache)
@@ -1876,9 +1891,6 @@ static void destroy(struct cache *cache)
{
unsigned i;
- if (cache->next_migration)
- mempool_free(cache->next_migration, cache->migration_pool);
-
if (cache->migration_pool)
mempool_destroy(cache->migration_pool);
@@ -2424,7 +2436,8 @@ static int cache_create(struct cache_args *ca, struct cache **result)
INIT_LIST_HEAD(&cache->quiesced_migrations);
INIT_LIST_HEAD(&cache->completed_migrations);
INIT_LIST_HEAD(&cache->need_commit_migrations);
- atomic_set(&cache->nr_migrations, 0);
+ atomic_set(&cache->nr_allocated_migrations, 0);
+ atomic_set(&cache->nr_io_migrations, 0);
init_waitqueue_head(&cache->migration_wait);
init_waitqueue_head(&cache->quiescing_wait);
@@ -2487,8 +2500,6 @@ static int cache_create(struct cache_args *ca, struct cache **result)
goto bad;
}
- cache->next_migration = NULL;
-
cache->need_tick_bio = true;
cache->sized = false;
cache->invalidate = false;
diff --git a/drivers/md/dm-thin.c b/drivers/md/dm-thin.c
index 8735543eacdb..07705ee181e3 100644
--- a/drivers/md/dm-thin.c
+++ b/drivers/md/dm-thin.c
@@ -1127,6 +1127,24 @@ static void schedule_external_copy(struct thin_c *tc, dm_block_t virt_block,
schedule_zero(tc, virt_block, data_dest, cell, bio);
}
+static void set_pool_mode(struct pool *pool, enum pool_mode new_mode);
+
+static void check_for_space(struct pool *pool)
+{
+ int r;
+ dm_block_t nr_free;
+
+ if (get_pool_mode(pool) != PM_OUT_OF_DATA_SPACE)
+ return;
+
+ r = dm_pool_get_free_block_count(pool->pmd, &nr_free);
+ if (r)
+ return;
+
+ if (nr_free)
+ set_pool_mode(pool, PM_WRITE);
+}
+
/*
* A non-zero return indicates read_only or fail_io mode.
* Many callers don't care about the return value.
@@ -1141,6 +1159,8 @@ static int commit(struct pool *pool)
r = dm_pool_commit_metadata(pool->pmd);
if (r)
metadata_operation_failed(pool, "dm_pool_commit_metadata", r);
+ else
+ check_for_space(pool);
return r;
}
@@ -1159,8 +1179,6 @@ static void check_low_water_mark(struct pool *pool, dm_block_t free_blocks)
}
}
-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;
@@ -2155,7 +2173,7 @@ static void set_pool_mode(struct pool *pool, enum pool_mode new_mode)
pool->process_cell = process_cell_read_only;
pool->process_discard_cell = process_discard_cell;
pool->process_prepared_mapping = process_prepared_mapping;
- pool->process_prepared_discard = process_prepared_discard_passdown;
+ pool->process_prepared_discard = process_prepared_discard;
if (!pool->pf.error_if_no_space && no_space_timeout)
queue_delayed_work(pool->wq, &pool->no_space_timeout, no_space_timeout);
@@ -3367,6 +3385,12 @@ static int pool_message(struct dm_target *ti, unsigned argc, char **argv)
struct pool_c *pt = ti->private;
struct pool *pool = pt->pool;
+ if (get_pool_mode(pool) >= PM_READ_ONLY) {
+ DMERR("%s: unable to service pool target messages in READ_ONLY or FAIL mode",
+ dm_device_name(pool->pool_md));
+ return -EINVAL;
+ }
+
if (!strcasecmp(argv[0], "create_thin"))
r = process_create_thin_mesg(argc, argv, pool);
@@ -3814,6 +3838,8 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv)
r = -EINVAL;
goto bad;
}
+ atomic_set(&tc->refcount, 1);
+ init_completion(&tc->can_destroy);
list_add_tail_rcu(&tc->list, &tc->pool->active_thins);
spin_unlock_irqrestore(&tc->pool->lock, flags);
/*
@@ -3826,9 +3852,6 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv)
dm_put(pool_md);
- atomic_set(&tc->refcount, 1);
- init_completion(&tc->can_destroy);
-
return 0;
bad:
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 4c06585bf165..2caf5b374649 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -206,6 +206,9 @@ struct mapped_device {
/* zero-length flush that will be cloned and submitted to targets */
struct bio flush_bio;
+ /* the number of internal suspends */
+ unsigned internal_suspend_count;
+
struct dm_stats stats;
};
@@ -899,7 +902,7 @@ static void disable_write_same(struct mapped_device *md)
static void clone_endio(struct bio *bio, int error)
{
- int r = 0;
+ int r = error;
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;
@@ -2928,7 +2931,7 @@ static void __dm_internal_suspend(struct mapped_device *md, unsigned suspend_fla
{
struct dm_table *map = NULL;
- if (dm_suspended_internally_md(md))
+ if (md->internal_suspend_count++)
return; /* nested internal suspend */
if (dm_suspended_md(md)) {
@@ -2953,7 +2956,9 @@ static void __dm_internal_suspend(struct mapped_device *md, unsigned suspend_fla
static void __dm_internal_resume(struct mapped_device *md)
{
- if (!dm_suspended_internally_md(md))
+ BUG_ON(!md->internal_suspend_count);
+
+ if (--md->internal_suspend_count)
return; /* resume from nested internal suspend */
if (dm_suspended_md(md))
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index c1b0d52bfcb0..b98765f6f77f 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -3195,6 +3195,11 @@ static void handle_stripe_dirtying(struct r5conf *conf,
(unsigned long long)sh->sector,
rcw, qread, test_bit(STRIPE_DELAYED, &sh->state));
}
+
+ if (rcw > disks && rmw > disks &&
+ !test_bit(STRIPE_PREREAD_ACTIVE, &sh->state))
+ set_bit(STRIPE_DELAYED, &sh->state);
+
/* now if nothing is locked, and if we have enough data,
* we can start a write request
*/
View Lorry Controller Status