*/
#include "dm-thin-metadata.h"
+#include "dm-bio-prison.h"
#include "dm.h"
#include <linux/device-mapper.h>
* Tunable constants
*/
#define ENDIO_HOOK_POOL_SIZE 1024
-#define DEFERRED_SET_SIZE 64
#define MAPPING_POOL_SIZE 1024
#define PRISON_CELLS 1024
#define COMMIT_PERIOD HZ
* i) plug io further to this physical block. (see bio_prison code).
*
* ii) quiesce any read io to that shared data block. Obviously
- * including all devices that share this block. (see deferred_set code)
+ * including all devices that share this block. (see dm_deferred_set code)
*
* iii) copy the data block to a newly allocate block. This step can be
* missed out if the io covers the block. (schedule_copy).
/*----------------------------------------------------------------*/
/*
- * Sometimes we can't deal with a bio straight away. We put them in prison
- * where they can't cause any mischief. Bios are put in a cell identified
- * by a key, multiple bios can be in the same cell. When the cell is
- * subsequently unlocked the bios become available.
- */
-struct bio_prison;
-
-struct cell_key {
- int virtual;
- dm_thin_id dev;
- dm_block_t block;
-};
-
-struct dm_bio_prison_cell {
- struct hlist_node list;
- struct bio_prison *prison;
- struct cell_key key;
- struct bio *holder;
- struct bio_list bios;
-};
-
-struct bio_prison {
- spinlock_t lock;
- mempool_t *cell_pool;
-
- unsigned nr_buckets;
- unsigned hash_mask;
- struct hlist_head *cells;
-};
-
-static uint32_t calc_nr_buckets(unsigned nr_cells)
-{
- uint32_t n = 128;
-
- nr_cells /= 4;
- nr_cells = min(nr_cells, 8192u);
-
- while (n < nr_cells)
- n <<= 1;
-
- return n;
-}
-
-static struct kmem_cache *_cell_cache;
-
-/*
- * @nr_cells should be the number of cells you want in use _concurrently_.
- * Don't confuse it with the number of distinct keys.
- */
-static struct bio_prison *prison_create(unsigned nr_cells)
-{
- unsigned i;
- uint32_t nr_buckets = calc_nr_buckets(nr_cells);
- size_t len = sizeof(struct bio_prison) +
- (sizeof(struct hlist_head) * nr_buckets);
- struct bio_prison *prison = kmalloc(len, GFP_KERNEL);
-
- if (!prison)
- return NULL;
-
- spin_lock_init(&prison->lock);
- prison->cell_pool = mempool_create_slab_pool(nr_cells, _cell_cache);
- if (!prison->cell_pool) {
- kfree(prison);
- return NULL;
- }
-
- prison->nr_buckets = nr_buckets;
- prison->hash_mask = nr_buckets - 1;
- prison->cells = (struct hlist_head *) (prison + 1);
- for (i = 0; i < nr_buckets; i++)
- INIT_HLIST_HEAD(prison->cells + i);
-
- return prison;
-}
-
-static void prison_destroy(struct bio_prison *prison)
-{
- mempool_destroy(prison->cell_pool);
- kfree(prison);
-}
-
-static uint32_t hash_key(struct bio_prison *prison, struct cell_key *key)
-{
- const unsigned long BIG_PRIME = 4294967291UL;
- uint64_t hash = key->block * BIG_PRIME;
-
- return (uint32_t) (hash & prison->hash_mask);
-}
-
-static int keys_equal(struct cell_key *lhs, struct cell_key *rhs)
-{
- return (lhs->virtual == rhs->virtual) &&
- (lhs->dev == rhs->dev) &&
- (lhs->block == rhs->block);
-}
-
-static struct dm_bio_prison_cell *__search_bucket(struct hlist_head *bucket,
- struct cell_key *key)
-{
- struct dm_bio_prison_cell *cell;
- struct hlist_node *tmp;
-
- hlist_for_each_entry(cell, tmp, bucket, list)
- if (keys_equal(&cell->key, key))
- return cell;
-
- return NULL;
-}
-
-/*
- * This may block if a new cell needs allocating. You must ensure that
- * cells will be unlocked even if the calling thread is blocked.
- *
- * Returns 1 if the cell was already held, 0 if @inmate is the new holder.
- */
-static int bio_detain(struct bio_prison *prison, struct cell_key *key,
- struct bio *inmate, struct dm_bio_prison_cell **ref)
-{
- int r = 1;
- unsigned long flags;
- uint32_t hash = hash_key(prison, key);
- struct dm_bio_prison_cell *cell, *cell2;
-
- BUG_ON(hash > prison->nr_buckets);
-
- spin_lock_irqsave(&prison->lock, flags);
-
- cell = __search_bucket(prison->cells + hash, key);
- if (cell) {
- bio_list_add(&cell->bios, inmate);
- goto out;
- }
-
- /*
- * Allocate a new cell
- */
- spin_unlock_irqrestore(&prison->lock, flags);
- cell2 = mempool_alloc(prison->cell_pool, GFP_NOIO);
- spin_lock_irqsave(&prison->lock, flags);
-
- /*
- * We've been unlocked, so we have to double check that
- * nobody else has inserted this cell in the meantime.
- */
- cell = __search_bucket(prison->cells + hash, key);
- if (cell) {
- mempool_free(cell2, prison->cell_pool);
- bio_list_add(&cell->bios, inmate);
- goto out;
- }
-
- /*
- * Use new cell.
- */
- cell = cell2;
-
- cell->prison = prison;
- memcpy(&cell->key, key, sizeof(cell->key));
- cell->holder = inmate;
- bio_list_init(&cell->bios);
- hlist_add_head(&cell->list, prison->cells + hash);
-
- r = 0;
-
-out:
- spin_unlock_irqrestore(&prison->lock, flags);
-
- *ref = cell;
-
- return r;
-}
-
-/*
- * @inmates must have been initialised prior to this call
- */
-static void __cell_release(struct dm_bio_prison_cell *cell, struct bio_list *inmates)
-{
- struct bio_prison *prison = cell->prison;
-
- hlist_del(&cell->list);
-
- if (inmates) {
- bio_list_add(inmates, cell->holder);
- bio_list_merge(inmates, &cell->bios);
- }
-
- mempool_free(cell, prison->cell_pool);
-}
-
-static void cell_release(struct dm_bio_prison_cell *cell, struct bio_list *bios)
-{
- unsigned long flags;
- struct bio_prison *prison = cell->prison;
-
- spin_lock_irqsave(&prison->lock, flags);
- __cell_release(cell, bios);
- spin_unlock_irqrestore(&prison->lock, flags);
-}
-
-/*
- * There are a couple of places where we put a bio into a cell briefly
- * before taking it out again. In these situations we know that no other
- * bio may be in the cell. This function releases the cell, and also does
- * a sanity check.
- */
-static void __cell_release_singleton(struct dm_bio_prison_cell *cell, struct bio *bio)
-{
- BUG_ON(cell->holder != bio);
- BUG_ON(!bio_list_empty(&cell->bios));
-
- __cell_release(cell, NULL);
-}
-
-static void cell_release_singleton(struct dm_bio_prison_cell *cell, struct bio *bio)
-{
- unsigned long flags;
- struct bio_prison *prison = cell->prison;
-
- spin_lock_irqsave(&prison->lock, flags);
- __cell_release_singleton(cell, bio);
- spin_unlock_irqrestore(&prison->lock, flags);
-}
-
-/*
- * Sometimes we don't want the holder, just the additional bios.
- */
-static void __cell_release_no_holder(struct dm_bio_prison_cell *cell,
- struct bio_list *inmates)
-{
- struct bio_prison *prison = cell->prison;
-
- hlist_del(&cell->list);
- bio_list_merge(inmates, &cell->bios);
-
- mempool_free(cell, prison->cell_pool);
-}
-
-static void cell_release_no_holder(struct dm_bio_prison_cell *cell,
- struct bio_list *inmates)
-{
- unsigned long flags;
- struct bio_prison *prison = cell->prison;
-
- spin_lock_irqsave(&prison->lock, flags);
- __cell_release_no_holder(cell, inmates);
- spin_unlock_irqrestore(&prison->lock, flags);
-}
-
-static void cell_error(struct dm_bio_prison_cell *cell)
-{
- struct bio_prison *prison = cell->prison;
- struct bio_list bios;
- struct bio *bio;
- unsigned long flags;
-
- bio_list_init(&bios);
-
- spin_lock_irqsave(&prison->lock, flags);
- __cell_release(cell, &bios);
- spin_unlock_irqrestore(&prison->lock, flags);
-
- while ((bio = bio_list_pop(&bios)))
- bio_io_error(bio);
-}
-
-/*----------------------------------------------------------------*/
-
-/*
- * We use the deferred set to keep track of pending reads to shared blocks.
- * We do this to ensure the new mapping caused by a write isn't performed
- * until these prior reads have completed. Otherwise the insertion of the
- * new mapping could free the old block that the read bios are mapped to.
- */
-
-struct deferred_set;
-struct deferred_entry {
- struct deferred_set *ds;
- unsigned count;
- struct list_head work_items;
-};
-
-struct deferred_set {
- spinlock_t lock;
- unsigned current_entry;
- unsigned sweeper;
- struct deferred_entry entries[DEFERRED_SET_SIZE];
-};
-
-static void ds_init(struct deferred_set *ds)
-{
- int i;
-
- spin_lock_init(&ds->lock);
- ds->current_entry = 0;
- ds->sweeper = 0;
- for (i = 0; i < DEFERRED_SET_SIZE; i++) {
- ds->entries[i].ds = ds;
- ds->entries[i].count = 0;
- INIT_LIST_HEAD(&ds->entries[i].work_items);
- }
-}
-
-static struct deferred_entry *ds_inc(struct deferred_set *ds)
-{
- unsigned long flags;
- struct deferred_entry *entry;
-
- spin_lock_irqsave(&ds->lock, flags);
- entry = ds->entries + ds->current_entry;
- entry->count++;
- spin_unlock_irqrestore(&ds->lock, flags);
-
- return entry;
-}
-
-static unsigned ds_next(unsigned index)
-{
- return (index + 1) % DEFERRED_SET_SIZE;
-}
-
-static void __sweep(struct deferred_set *ds, struct list_head *head)
-{
- while ((ds->sweeper != ds->current_entry) &&
- !ds->entries[ds->sweeper].count) {
- list_splice_init(&ds->entries[ds->sweeper].work_items, head);
- ds->sweeper = ds_next(ds->sweeper);
- }
-
- if ((ds->sweeper == ds->current_entry) && !ds->entries[ds->sweeper].count)
- list_splice_init(&ds->entries[ds->sweeper].work_items, head);
-}
-
-static void ds_dec(struct deferred_entry *entry, struct list_head *head)
-{
- unsigned long flags;
-
- spin_lock_irqsave(&entry->ds->lock, flags);
- BUG_ON(!entry->count);
- --entry->count;
- __sweep(entry->ds, head);
- spin_unlock_irqrestore(&entry->ds->lock, flags);
-}
-
-/*
- * Returns 1 if deferred or 0 if no pending items to delay job.
- */
-static int ds_add_work(struct deferred_set *ds, struct list_head *work)
-{
- int r = 1;
- unsigned long flags;
- unsigned next_entry;
-
- spin_lock_irqsave(&ds->lock, flags);
- if ((ds->sweeper == ds->current_entry) &&
- !ds->entries[ds->current_entry].count)
- r = 0;
- else {
- list_add(work, &ds->entries[ds->current_entry].work_items);
- next_entry = ds_next(ds->current_entry);
- if (!ds->entries[next_entry].count)
- ds->current_entry = next_entry;
- }
- spin_unlock_irqrestore(&ds->lock, flags);
-
- return r;
-}
-
-/*----------------------------------------------------------------*/
-
-/*
* Key building.
*/
static void build_data_key(struct dm_thin_device *td,
- dm_block_t b, struct cell_key *key)
+ dm_block_t b, struct dm_cell_key *key)
{
key->virtual = 0;
key->dev = dm_thin_dev_id(td);
}
static void build_virtual_key(struct dm_thin_device *td, dm_block_t b,
- struct cell_key *key)
+ struct dm_cell_key *key)
{
key->virtual = 1;
key->dev = dm_thin_dev_id(td);
unsigned low_water_triggered:1; /* A dm event has been sent */
unsigned no_free_space:1; /* A -ENOSPC warning has been issued */
- struct bio_prison *prison;
+ struct dm_bio_prison *prison;
struct dm_kcopyd_client *copier;
struct workqueue_struct *wq;
struct bio_list retry_on_resume_list;
- struct deferred_set shared_read_ds;
- struct deferred_set all_io_ds;
+ struct dm_deferred_set *shared_read_ds;
+ struct dm_deferred_set *all_io_ds;
struct dm_thin_new_mapping *next_mapping;
mempool_t *mapping_pool;
- mempool_t *endio_hook_pool;
process_bio_fn process_bio;
process_bio_fn process_discard;
struct dm_thin_endio_hook {
struct thin_c *tc;
- struct deferred_entry *shared_read_entry;
- struct deferred_entry *all_io_entry;
+ struct dm_deferred_entry *shared_read_entry;
+ struct dm_deferred_entry *all_io_entry;
struct dm_thin_new_mapping *overwrite_mapping;
};
bio_list_init(master);
while ((bio = bio_list_pop(&bios))) {
- struct dm_thin_endio_hook *h = dm_get_mapinfo(bio)->ptr;
+ struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
if (h->tc == tc)
bio_endio(bio, DM_ENDIO_REQUEUE);
dm_thin_changed_this_transaction(tc->td);
}
+static void inc_all_io_entry(struct pool *pool, struct bio *bio)
+{
+ struct dm_thin_endio_hook *h;
+
+ if (bio->bi_rw & REQ_DISCARD)
+ return;
+
+ h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
+ h->all_io_entry = dm_deferred_entry_inc(pool->all_io_ds);
+}
+
static void issue(struct thin_c *tc, struct bio *bio)
{
struct pool *pool = tc->pool;
static void overwrite_endio(struct bio *bio, int err)
{
unsigned long flags;
- struct dm_thin_endio_hook *h = dm_get_mapinfo(bio)->ptr;
+ struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
struct dm_thin_new_mapping *m = h->overwrite_mapping;
struct pool *pool = m->tc->pool;
/*
* This sends the bios in the cell back to the deferred_bios list.
*/
-static void cell_defer(struct thin_c *tc, struct dm_bio_prison_cell *cell,
- dm_block_t data_block)
+static void cell_defer(struct thin_c *tc, struct dm_bio_prison_cell *cell)
{
struct pool *pool = tc->pool;
unsigned long flags;
spin_lock_irqsave(&pool->lock, flags);
- cell_release(cell, &pool->deferred_bios);
+ dm_cell_release(cell, &pool->deferred_bios);
spin_unlock_irqrestore(&tc->pool->lock, flags);
wake_worker(pool);
}
/*
- * Same as cell_defer above, except it omits one particular detainee,
- * a write bio that covers the block and has already been processed.
+ * Same as cell_defer except it omits the original holder of the cell.
*/
-static void cell_defer_except(struct thin_c *tc, struct dm_bio_prison_cell *cell)
+static void cell_defer_no_holder(struct thin_c *tc, struct dm_bio_prison_cell *cell)
{
- struct bio_list bios;
struct pool *pool = tc->pool;
unsigned long flags;
- bio_list_init(&bios);
-
spin_lock_irqsave(&pool->lock, flags);
- cell_release_no_holder(cell, &pool->deferred_bios);
+ dm_cell_release_no_holder(cell, &pool->deferred_bios);
spin_unlock_irqrestore(&pool->lock, flags);
wake_worker(pool);
{
if (m->bio)
m->bio->bi_end_io = m->saved_bi_end_io;
- cell_error(m->cell);
+ dm_cell_error(m->cell);
list_del(&m->list);
mempool_free(m, m->tc->pool->mapping_pool);
}
bio->bi_end_io = m->saved_bi_end_io;
if (m->err) {
- cell_error(m->cell);
+ dm_cell_error(m->cell);
goto out;
}
*/
r = dm_thin_insert_block(tc->td, m->virt_block, m->data_block);
if (r) {
- DMERR("dm_thin_insert_block() failed");
- cell_error(m->cell);
+ DMERR_LIMIT("dm_thin_insert_block() failed");
+ dm_cell_error(m->cell);
goto out;
}
* the bios in the cell.
*/
if (bio) {
- cell_defer_except(tc, m->cell);
+ cell_defer_no_holder(tc, m->cell);
bio_endio(bio, 0);
} else
- cell_defer(tc, m->cell, m->data_block);
+ cell_defer(tc, m->cell);
out:
list_del(&m->list);
struct thin_c *tc = m->tc;
bio_io_error(m->bio);
- cell_defer_except(tc, m->cell);
- cell_defer_except(tc, m->cell2);
+ cell_defer_no_holder(tc, m->cell);
+ cell_defer_no_holder(tc, m->cell2);
mempool_free(m, tc->pool->mapping_pool);
}
{
struct thin_c *tc = m->tc;
+ inc_all_io_entry(tc->pool, m->bio);
+ cell_defer_no_holder(tc, m->cell);
+ cell_defer_no_holder(tc, m->cell2);
+
if (m->pass_discard)
remap_and_issue(tc, m->bio, m->data_block);
else
bio_endio(m->bio, 0);
- cell_defer_except(tc, m->cell);
- cell_defer_except(tc, m->cell2);
mempool_free(m, tc->pool->mapping_pool);
}
r = dm_thin_remove_block(tc->td, m->virt_block);
if (r)
- DMERR("dm_thin_remove_block() failed");
+ DMERR_LIMIT("dm_thin_remove_block() failed");
process_prepared_discard_passdown(m);
}
m->err = 0;
m->bio = NULL;
- if (!ds_add_work(&pool->shared_read_ds, &m->list))
+ if (!dm_deferred_set_add_work(pool->shared_read_ds, &m->list))
m->quiesced = 1;
/*
* bio immediately. Otherwise we use kcopyd to clone the data first.
*/
if (io_overwrites_block(pool, bio)) {
- struct dm_thin_endio_hook *h = dm_get_mapinfo(bio)->ptr;
+ struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
h->overwrite_mapping = m;
m->bio = bio;
save_and_set_endio(bio, &m->saved_bi_end_io, overwrite_endio);
+ inc_all_io_entry(pool, bio);
remap_and_issue(tc, bio, data_dest);
} else {
struct dm_io_region from, to;
0, copy_complete, m);
if (r < 0) {
mempool_free(m, pool->mapping_pool);
- DMERR("dm_kcopyd_copy() failed");
- cell_error(cell);
+ DMERR_LIMIT("dm_kcopyd_copy() failed");
+ dm_cell_error(cell);
}
}
}
process_prepared_mapping(m);
else if (io_overwrites_block(pool, bio)) {
- struct dm_thin_endio_hook *h = dm_get_mapinfo(bio)->ptr;
+ struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
h->overwrite_mapping = m;
m->bio = bio;
save_and_set_endio(bio, &m->saved_bi_end_io, overwrite_endio);
+ inc_all_io_entry(pool, bio);
remap_and_issue(tc, bio, data_block);
} else {
int r;
r = dm_kcopyd_zero(pool->copier, 1, &to, 0, copy_complete, m);
if (r < 0) {
mempool_free(m, pool->mapping_pool);
- DMERR("dm_kcopyd_zero() failed");
- cell_error(cell);
+ DMERR_LIMIT("dm_kcopyd_zero() failed");
+ dm_cell_error(cell);
}
}
}
r = dm_pool_commit_metadata(pool->pmd);
if (r)
- DMERR("commit failed, error = %d", r);
+ DMERR_LIMIT("commit failed: error = %d", r);
return r;
}
*/
static void retry_on_resume(struct bio *bio)
{
- struct dm_thin_endio_hook *h = dm_get_mapinfo(bio)->ptr;
+ struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
struct thin_c *tc = h->tc;
struct pool *pool = tc->pool;
unsigned long flags;
struct bio_list bios;
bio_list_init(&bios);
- cell_release(cell, &bios);
+ dm_cell_release(cell, &bios);
while ((bio = bio_list_pop(&bios)))
retry_on_resume(bio);
unsigned long flags;
struct pool *pool = tc->pool;
struct dm_bio_prison_cell *cell, *cell2;
- struct cell_key key, key2;
+ struct dm_cell_key key, key2;
dm_block_t block = get_bio_block(tc, bio);
struct dm_thin_lookup_result lookup_result;
struct dm_thin_new_mapping *m;
build_virtual_key(tc->td, block, &key);
- if (bio_detain(tc->pool->prison, &key, bio, &cell))
+ if (dm_bio_detain(tc->pool->prison, &key, bio, &cell))
return;
r = dm_thin_find_block(tc->td, block, 1, &lookup_result);
* on this block.
*/
build_data_key(tc->td, lookup_result.block, &key2);
- if (bio_detain(tc->pool->prison, &key2, bio, &cell2)) {
- cell_release_singleton(cell, bio);
+ if (dm_bio_detain(tc->pool->prison, &key2, bio, &cell2)) {
+ cell_defer_no_holder(tc, cell);
break;
}
m->err = 0;
m->bio = bio;
- if (!ds_add_work(&pool->all_io_ds, &m->list)) {
+ if (!dm_deferred_set_add_work(pool->all_io_ds, &m->list)) {
spin_lock_irqsave(&pool->lock, flags);
list_add(&m->list, &pool->prepared_discards);
spin_unlock_irqrestore(&pool->lock, flags);
wake_worker(pool);
}
} else {
+ inc_all_io_entry(pool, bio);
+ cell_defer_no_holder(tc, cell);
+ cell_defer_no_holder(tc, cell2);
+
/*
* The DM core makes sure that the discard doesn't span
* a block boundary. So we submit the discard of a
* partial block appropriately.
*/
- cell_release_singleton(cell, bio);
- cell_release_singleton(cell2, bio);
if ((!lookup_result.shared) && pool->pf.discard_passdown)
remap_and_issue(tc, bio, lookup_result.block);
else
/*
* It isn't provisioned, just forget it.
*/
- cell_release_singleton(cell, bio);
+ cell_defer_no_holder(tc, cell);
bio_endio(bio, 0);
break;
default:
- DMERR("discard: find block unexpectedly returned %d", r);
- cell_release_singleton(cell, bio);
+ DMERR_LIMIT("%s: dm_thin_find_block() failed: error = %d",
+ __func__, r);
+ cell_defer_no_holder(tc, cell);
bio_io_error(bio);
break;
}
}
static void break_sharing(struct thin_c *tc, struct bio *bio, dm_block_t block,
- struct cell_key *key,
+ struct dm_cell_key *key,
struct dm_thin_lookup_result *lookup_result,
struct dm_bio_prison_cell *cell)
{
break;
default:
- DMERR("%s: alloc_data_block() failed, error = %d", __func__, r);
- cell_error(cell);
+ DMERR_LIMIT("%s: alloc_data_block() failed: error = %d",
+ __func__, r);
+ dm_cell_error(cell);
break;
}
}
{
struct dm_bio_prison_cell *cell;
struct pool *pool = tc->pool;
- struct cell_key key;
+ struct dm_cell_key key;
/*
* If cell is already occupied, then sharing is already in the process
* of being broken so we have nothing further to do here.
*/
build_data_key(tc->td, lookup_result->block, &key);
- if (bio_detain(pool->prison, &key, bio, &cell))
+ if (dm_bio_detain(pool->prison, &key, bio, &cell))
return;
if (bio_data_dir(bio) == WRITE && bio->bi_size)
break_sharing(tc, bio, block, &key, lookup_result, cell);
else {
- struct dm_thin_endio_hook *h = dm_get_mapinfo(bio)->ptr;
+ struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
- h->shared_read_entry = ds_inc(&pool->shared_read_ds);
+ h->shared_read_entry = dm_deferred_entry_inc(pool->shared_read_ds);
+ inc_all_io_entry(pool, bio);
+ cell_defer_no_holder(tc, cell);
- cell_release_singleton(cell, bio);
remap_and_issue(tc, bio, lookup_result->block);
}
}
* Remap empty bios (flushes) immediately, without provisioning.
*/
if (!bio->bi_size) {
- cell_release_singleton(cell, bio);
+ inc_all_io_entry(tc->pool, bio);
+ cell_defer_no_holder(tc, cell);
+
remap_and_issue(tc, bio, 0);
return;
}
*/
if (bio_data_dir(bio) == READ) {
zero_fill_bio(bio);
- cell_release_singleton(cell, bio);
+ cell_defer_no_holder(tc, cell);
bio_endio(bio, 0);
return;
}
break;
default:
- DMERR("%s: alloc_data_block() failed, error = %d", __func__, r);
+ DMERR_LIMIT("%s: alloc_data_block() failed: error = %d",
+ __func__, r);
set_pool_mode(tc->pool, PM_READ_ONLY);
- cell_error(cell);
+ dm_cell_error(cell);
break;
}
}
int r;
dm_block_t block = get_bio_block(tc, bio);
struct dm_bio_prison_cell *cell;
- struct cell_key key;
+ struct dm_cell_key key;
struct dm_thin_lookup_result lookup_result;
/*
* being provisioned so we have nothing further to do here.
*/
build_virtual_key(tc->td, block, &key);
- if (bio_detain(tc->pool->prison, &key, bio, &cell))
+ if (dm_bio_detain(tc->pool->prison, &key, bio, &cell))
return;
r = dm_thin_find_block(tc->td, block, 1, &lookup_result);
switch (r) {
case 0:
- /*
- * We can release this cell now. This thread is the only
- * one that puts bios into a cell, and we know there were
- * no preceding bios.
- */
- /*
- * TODO: this will probably have to change when discard goes
- * back in.
- */
- cell_release_singleton(cell, bio);
-
- if (lookup_result.shared)
+ if (lookup_result.shared) {
process_shared_bio(tc, bio, block, &lookup_result);
- else
+ cell_defer_no_holder(tc, cell);
+ } else {
+ inc_all_io_entry(tc->pool, bio);
+ cell_defer_no_holder(tc, cell);
+
remap_and_issue(tc, bio, lookup_result.block);
+ }
break;
case -ENODATA:
if (bio_data_dir(bio) == READ && tc->origin_dev) {
- cell_release_singleton(cell, bio);
+ inc_all_io_entry(tc->pool, bio);
+ cell_defer_no_holder(tc, cell);
+
remap_to_origin_and_issue(tc, bio);
} else
provision_block(tc, bio, block, cell);
break;
default:
- DMERR("dm_thin_find_block() failed, error = %d", r);
- cell_release_singleton(cell, bio);
+ DMERR_LIMIT("%s: dm_thin_find_block() failed: error = %d",
+ __func__, r);
+ cell_defer_no_holder(tc, cell);
bio_io_error(bio);
break;
}
case 0:
if (lookup_result.shared && (rw == WRITE) && bio->bi_size)
bio_io_error(bio);
- else
+ else {
+ inc_all_io_entry(tc->pool, bio);
remap_and_issue(tc, bio, lookup_result.block);
+ }
break;
case -ENODATA:
}
if (tc->origin_dev) {
+ inc_all_io_entry(tc->pool, bio);
remap_to_origin_and_issue(tc, bio);
break;
}
break;
default:
- DMERR("dm_thin_find_block() failed, error = %d", r);
+ DMERR_LIMIT("%s: dm_thin_find_block() failed: error = %d",
+ __func__, r);
bio_io_error(bio);
break;
}
spin_unlock_irqrestore(&pool->lock, flags);
while ((bio = bio_list_pop(&bios))) {
- struct dm_thin_endio_hook *h = dm_get_mapinfo(bio)->ptr;
+ struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
struct thin_c *tc = h->tc;
/*
wake_worker(pool);
}
-static struct dm_thin_endio_hook *thin_hook_bio(struct thin_c *tc, struct bio *bio)
+static void thin_hook_bio(struct thin_c *tc, struct bio *bio)
{
- struct pool *pool = tc->pool;
- struct dm_thin_endio_hook *h = mempool_alloc(pool->endio_hook_pool, GFP_NOIO);
+ struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
h->tc = tc;
h->shared_read_entry = NULL;
- h->all_io_entry = bio->bi_rw & REQ_DISCARD ? NULL : ds_inc(&pool->all_io_ds);
+ h->all_io_entry = NULL;
h->overwrite_mapping = NULL;
-
- return h;
}
/*
dm_block_t block = get_bio_block(tc, bio);
struct dm_thin_device *td = tc->td;
struct dm_thin_lookup_result result;
+ struct dm_bio_prison_cell *cell1, *cell2;
+ struct dm_cell_key key;
- map_context->ptr = thin_hook_bio(tc, bio);
+ thin_hook_bio(tc, bio);
if (get_pool_mode(tc->pool) == PM_FAIL) {
bio_io_error(bio);
* shared flag will be set in their case.
*/
thin_defer_bio(tc, bio);
- r = DM_MAPIO_SUBMITTED;
- } else {
- remap(tc, bio, result.block);
- r = DM_MAPIO_REMAPPED;
+ return DM_MAPIO_SUBMITTED;
+ }
+
+ build_virtual_key(tc->td, block, &key);
+ if (dm_bio_detain(tc->pool->prison, &key, bio, &cell1))
+ return DM_MAPIO_SUBMITTED;
+
+ build_data_key(tc->td, result.block, &key);
+ if (dm_bio_detain(tc->pool->prison, &key, bio, &cell2)) {
+ cell_defer_no_holder(tc, cell1);
+ return DM_MAPIO_SUBMITTED;
}
- break;
+
+ inc_all_io_entry(tc->pool, bio);
+ cell_defer_no_holder(tc, cell2);
+ cell_defer_no_holder(tc, cell1);
+
+ remap(tc, bio, result.block);
+ return DM_MAPIO_REMAPPED;
case -ENODATA:
if (get_pool_mode(tc->pool) == PM_READ_ONLY) {
* of doing so. Just error it.
*/
bio_io_error(bio);
- r = DM_MAPIO_SUBMITTED;
- break;
+ return DM_MAPIO_SUBMITTED;
}
/* fall through */
* provide the hint to load the metadata into cache.
*/
thin_defer_bio(tc, bio);
- r = DM_MAPIO_SUBMITTED;
- break;
+ return DM_MAPIO_SUBMITTED;
default:
/*
* pool is switched to fail-io mode.
*/
bio_io_error(bio);
- r = DM_MAPIO_SUBMITTED;
- break;
+ return DM_MAPIO_SUBMITTED;
}
-
- return r;
}
static int pool_is_congested(struct dm_target_callbacks *cb, int bdi_bits)
if (dm_pool_metadata_close(pool->pmd) < 0)
DMWARN("%s: dm_pool_metadata_close() failed.", __func__);
- prison_destroy(pool->prison);
+ dm_bio_prison_destroy(pool->prison);
dm_kcopyd_client_destroy(pool->copier);
if (pool->wq)
if (pool->next_mapping)
mempool_free(pool->next_mapping, pool->mapping_pool);
mempool_destroy(pool->mapping_pool);
- mempool_destroy(pool->endio_hook_pool);
+ dm_deferred_set_destroy(pool->shared_read_ds);
+ dm_deferred_set_destroy(pool->all_io_ds);
kfree(pool);
}
static struct kmem_cache *_new_mapping_cache;
-static struct kmem_cache *_endio_hook_cache;
static struct pool *pool_create(struct mapped_device *pool_md,
struct block_device *metadata_dev,
pool->sectors_per_block_shift = __ffs(block_size);
pool->low_water_blocks = 0;
pool_features_init(&pool->pf);
- pool->prison = prison_create(PRISON_CELLS);
+ pool->prison = dm_bio_prison_create(PRISON_CELLS);
if (!pool->prison) {
*error = "Error creating pool's bio prison";
err_p = ERR_PTR(-ENOMEM);
pool->low_water_triggered = 0;
pool->no_free_space = 0;
bio_list_init(&pool->retry_on_resume_list);
- ds_init(&pool->shared_read_ds);
- ds_init(&pool->all_io_ds);
+
+ pool->shared_read_ds = dm_deferred_set_create();
+ if (!pool->shared_read_ds) {
+ *error = "Error creating pool's shared read deferred set";
+ err_p = ERR_PTR(-ENOMEM);
+ goto bad_shared_read_ds;
+ }
+
+ pool->all_io_ds = dm_deferred_set_create();
+ if (!pool->all_io_ds) {
+ *error = "Error creating pool's all io deferred set";
+ err_p = ERR_PTR(-ENOMEM);
+ goto bad_all_io_ds;
+ }
pool->next_mapping = NULL;
pool->mapping_pool = mempool_create_slab_pool(MAPPING_POOL_SIZE,
goto bad_mapping_pool;
}
- pool->endio_hook_pool = mempool_create_slab_pool(ENDIO_HOOK_POOL_SIZE,
- _endio_hook_cache);
- if (!pool->endio_hook_pool) {
- *error = "Error creating pool's endio_hook mempool";
- err_p = ERR_PTR(-ENOMEM);
- goto bad_endio_hook_pool;
- }
pool->ref_count = 1;
pool->last_commit_jiffies = jiffies;
pool->pool_md = pool_md;
return pool;
-bad_endio_hook_pool:
- mempool_destroy(pool->mapping_pool);
bad_mapping_pool:
+ dm_deferred_set_destroy(pool->all_io_ds);
+bad_all_io_ds:
+ dm_deferred_set_destroy(pool->shared_read_ds);
+bad_shared_read_ds:
destroy_workqueue(pool->wq);
bad_wq:
dm_kcopyd_client_destroy(pool->copier);
bad_kcopyd_client:
- prison_destroy(pool->prison);
+ dm_bio_prison_destroy(pool->prison);
bad_prison:
kfree(pool);
bad_pool:
goto out_flags_changed;
}
- /*
- * The block layer requires discard_granularity to be a power of 2.
- */
- if (pf.discard_enabled && !is_power_of_2(block_size)) {
- ti->error = "Discard support must be disabled when the block size is not a power of 2";
- r = -EINVAL;
- goto out_flags_changed;
- }
-
pt->pool = pool;
pt->ti = ti;
pt->metadata_dev = metadata_dev;
else
DMEMIT("rw ");
- if (pool->pf.discard_enabled && pool->pf.discard_passdown)
+ if (!pool->pf.discard_enabled)
+ DMEMIT("ignore_discard");
+ else if (pool->pf.discard_passdown)
DMEMIT("discard_passdown");
else
DMEMIT("no_discard_passdown");
return min(max_size, q->merge_bvec_fn(q, bvm, biovec));
}
+static bool block_size_is_power_of_two(struct pool *pool)
+{
+ return pool->sectors_per_block_shift >= 0;
+}
+
static void set_discard_limits(struct pool_c *pt, struct queue_limits *limits)
{
struct pool *pool = pt->pool;
if (pt->adjusted_pf.discard_passdown) {
data_limits = &bdev_get_queue(pt->data_dev->bdev)->limits;
limits->discard_granularity = data_limits->discard_granularity;
- } else
+ } else if (block_size_is_power_of_two(pool))
limits->discard_granularity = pool->sectors_per_block << SECTOR_SHIFT;
+ else
+ /*
+ * Use largest power of 2 that is a factor of sectors_per_block
+ * but at least DATA_DEV_BLOCK_SIZE_MIN_SECTORS.
+ */
+ limits->discard_granularity = max(1 << (ffs(pool->sectors_per_block) - 1),
+ DATA_DEV_BLOCK_SIZE_MIN_SECTORS) << SECTOR_SHIFT;
}
static void pool_io_hints(struct dm_target *ti, struct queue_limits *limits)
.name = "thin-pool",
.features = DM_TARGET_SINGLETON | DM_TARGET_ALWAYS_WRITEABLE |
DM_TARGET_IMMUTABLE,
- .version = {1, 4, 0},
+ .version = {1, 6, 0},
.module = THIS_MODULE,
.ctr = pool_ctr,
.dtr = pool_dtr,
ti->num_flush_requests = 1;
ti->flush_supported = true;
+ ti->per_bio_data_size = sizeof(struct dm_thin_endio_hook);
/* In case the pool supports discards, pass them on. */
if (tc->pool->pf.discard_enabled) {
union map_info *map_context)
{
unsigned long flags;
- struct dm_thin_endio_hook *h = map_context->ptr;
+ struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
struct list_head work;
struct dm_thin_new_mapping *m, *tmp;
struct pool *pool = h->tc->pool;
if (h->shared_read_entry) {
INIT_LIST_HEAD(&work);
- ds_dec(h->shared_read_entry, &work);
+ dm_deferred_entry_dec(h->shared_read_entry, &work);
spin_lock_irqsave(&pool->lock, flags);
list_for_each_entry_safe(m, tmp, &work, list) {
if (h->all_io_entry) {
INIT_LIST_HEAD(&work);
- ds_dec(h->all_io_entry, &work);
- spin_lock_irqsave(&pool->lock, flags);
- list_for_each_entry_safe(m, tmp, &work, list)
- list_add(&m->list, &pool->prepared_discards);
- spin_unlock_irqrestore(&pool->lock, flags);
+ dm_deferred_entry_dec(h->all_io_entry, &work);
+ if (!list_empty(&work)) {
+ spin_lock_irqsave(&pool->lock, flags);
+ list_for_each_entry_safe(m, tmp, &work, list)
+ list_add(&m->list, &pool->prepared_discards);
+ spin_unlock_irqrestore(&pool->lock, flags);
+ wake_worker(pool);
+ }
}
- mempool_free(h, pool->endio_hook_pool);
-
return 0;
}
static struct target_type thin_target = {
.name = "thin",
- .version = {1, 4, 0},
+ .version = {1, 6, 0},
.module = THIS_MODULE,
.ctr = thin_ctr,
.dtr = thin_dtr,
r = -ENOMEM;
- _cell_cache = KMEM_CACHE(dm_bio_prison_cell, 0);
- if (!_cell_cache)
- goto bad_cell_cache;
-
_new_mapping_cache = KMEM_CACHE(dm_thin_new_mapping, 0);
if (!_new_mapping_cache)
goto bad_new_mapping_cache;
- _endio_hook_cache = KMEM_CACHE(dm_thin_endio_hook, 0);
- if (!_endio_hook_cache)
- goto bad_endio_hook_cache;
-
return 0;
-bad_endio_hook_cache:
- kmem_cache_destroy(_new_mapping_cache);
bad_new_mapping_cache:
- kmem_cache_destroy(_cell_cache);
-bad_cell_cache:
dm_unregister_target(&pool_target);
bad_pool_target:
dm_unregister_target(&thin_target);
dm_unregister_target(&thin_target);
dm_unregister_target(&pool_target);
- kmem_cache_destroy(_cell_cache);
kmem_cache_destroy(_new_mapping_cache);
- kmem_cache_destroy(_endio_hook_cache);
}
module_init(dm_thin_init);