[Devel] [PATCH RHEL8 COMMIT] push_backup: Make the target !immutable
Konstantin Khorenko
khorenko at virtuozzo.com
Fri Sep 10 20:16:51 MSK 2021
The commit is pushed to "branch-rh8-4.18.0-305.3.1.vz8.7.x-ovz" and will appear at https://src.openvz.org/scm/ovz/vzkernel.git
after rh8-4.18.0-305.3.1.vz8.7.11
------>
commit 0aa4a43a9753d65f6a81d1bd036765bf3bbd39be
Author: Kirill Tkhai <ktkhai at virtuozzo.com>
Date: Fri Sep 10 20:16:51 2021 +0300
push_backup: Make the target !immutable
Then we will be able to replace it with another target.
!immutable target can't use per_io_data_size, so
we allocate that memory in driver.
Signed-off-by: Kirill Tkhai <ktkhai at virtuozzo.com>
====================
push_backup: Make target !immutable.
https://jira.sw.ru/browse/PSBM-127989
Kirill Tkhai (14):
push_backup: Rename ppb_map
push_backup: Add unsigned long alignment
push_backup: Add pending_map
push_backup: Kill find_node_pbio_range()
push_backup: Use nr_delayed in postpone_if_required_for_backup()
push_backup: Introduce hash table
push_backup: Leave pending pbio in pending queue
push_backup: Do not split bios by cluster size
dm: Allow singleton target with devices attached
dm: Introduce dm_requeue_original_rq()
push_backup: Make it request based
push_backup: Change retval postpone_if_required_for_backup()
push_backup: Change arguments of calc_bio_clusters()
push_backup: Make the target !immutable
---
drivers/md/dm-push-backup.c | 54 ++++++++++++++++++++++++++++++---------------
1 file changed, 36 insertions(+), 18 deletions(-)
diff --git a/drivers/md/dm-push-backup.c b/drivers/md/dm-push-backup.c
index 4f9aa558b9ad..4b46bff42260 100644
--- a/drivers/md/dm-push-backup.c
+++ b/drivers/md/dm-push-backup.c
@@ -21,6 +21,10 @@
#define PB_HASH_TABLE_BITS 5
#define PB_HASH_TABLE_SIZE (1 << PB_HASH_TABLE_BITS)
+#define PUSH_BACKUP_POOL_SIZE 128
+
+static struct kmem_cache *pbio_cache;
+
static inline struct hlist_head *pb_htable_slot(struct hlist_head head[], u32 clu)
{
return &head[hash_32(clu, PB_HASH_TABLE_BITS)];
@@ -38,6 +42,7 @@ struct pb_bio {
struct push_backup {
struct dm_target *ti;
struct dm_dev *origin_dev;
+ mempool_t *pbio_pool;
u64 clu_size;
u64 nr_clus;
@@ -106,20 +111,26 @@ static void calc_bio_clusters(struct push_backup *pb, struct request *rq,
*end_clu = (off + blk_rq_bytes(rq) - 1) / pb->clu_size;
}
-static int setup_if_required_for_backup(struct push_backup *pb, struct request *rq,
- struct pb_bio *pbio)
+static int setup_if_required_for_backup(struct push_backup *pb,
+ struct request *rq,
+ struct pb_bio **ret_pbio)
{
u64 start_clu, end_clu, key;
+ struct pb_bio *pbio;
calc_bio_clusters(pb, rq, &start_clu, &end_clu);
key = last_required_for_backup_clu(pb, start_clu, end_clu);
if (key != U64_MAX) {
+ pbio = mempool_alloc(pb->pbio_pool, GFP_ATOMIC);
+ if (!pbio)
+ return -ENOMEM;
init_pb_bio(pbio);
pbio->rq = rq;
pbio->start_clu = start_clu;
pbio->end_clu = end_clu;
pbio->key_clu = key;
+ *ret_pbio = pbio;
return 1;
}
return 0;
@@ -188,6 +199,7 @@ static void resubmit_pbios(struct push_backup *pb, struct list_head *list)
while ((pbio = list_first_entry_or_null(list, struct pb_bio, list)) != NULL) {
list_del_init(&pbio->list);
dm_requeue_original_rq(pbio->rq);
+ mempool_free(pbio, pb->pbio_pool);
}
}
@@ -239,10 +251,10 @@ static void pb_timer_func(struct timer_list *timer)
}
static int postpone_if_required_for_backup(struct push_backup *pb,
- struct request *rq,
- struct pb_bio *pbio)
+ struct request *rq)
{
bool queue_timer = false;
+ struct pb_bio *pbio;
unsigned long flags;
int ret = 0;
@@ -250,7 +262,7 @@ static int postpone_if_required_for_backup(struct push_backup *pb,
spin_lock_irqsave(&pb->lock, flags);
if (!pb->alive)
goto unlock;
- ret = setup_if_required_for_backup(pb, rq, pbio);
+ ret = setup_if_required_for_backup(pb, rq, &pbio);
if (ret <= 0)
goto unlock;
@@ -276,17 +288,11 @@ static int postpone_if_required_for_backup(struct push_backup *pb,
return ret;
}
-static inline struct pb_bio *map_info_to_embedded_pbio(union map_info *info)
-{
- return (void *)info->ptr;
-}
-
static int pb_clone_and_map(struct dm_target *ti, struct request *rq,
union map_info *map_context,
struct request **__clone)
{
- struct pb_bio *pbio = map_info_to_embedded_pbio(map_context);
struct push_backup *pb = ti->private;
struct block_device *bdev = pb->origin_dev->bdev;
struct request_queue *q;
@@ -294,7 +300,7 @@ static int pb_clone_and_map(struct dm_target *ti, struct request *rq,
int ret;
if (blk_rq_bytes(rq) && op_is_write(req_op(rq))) {
- ret = postpone_if_required_for_backup(pb, rq, pbio);
+ ret = postpone_if_required_for_backup(pb, rq);
if (ret < 0) /* ENOMEM */
return DM_MAPIO_REQUEUE;
if (ret > 0) /* Postponed */
@@ -628,6 +634,7 @@ static void pb_destroy(struct push_backup *pb)
kvfree(pb->map); /* Is's not zero if stop was not called */
kvfree(pb->pending_map);
kvfree(pb->pending_htable);
+ mempool_destroy(pb->pbio_pool);
if (pb->origin_dev)
dm_put_device(pb->ti, pb->origin_dev);
kfree(pb);
@@ -650,10 +657,12 @@ static int pb_ctr(struct dm_target *ti, unsigned int argc, char **argv)
if (!pb)
goto err;
+ pb->pbio_pool = mempool_create_slab_pool(PUSH_BACKUP_POOL_SIZE,
+ pbio_cache);
pb->pending_htable = kcalloc(PB_HASH_TABLE_SIZE,
sizeof(struct hlist_head),
GFP_KERNEL);
- if (!pb->pending_htable)
+ if (!pb->pbio_pool || !pb->pending_htable)
goto err;
pb->suspended = true;
@@ -694,7 +703,6 @@ static int pb_ctr(struct dm_target *ti, unsigned int argc, char **argv)
goto err;
}
- ti->per_io_data_size = sizeof(struct pb_bio);
ti->num_flush_bios = 1;
ti->flush_supported = true;
ti->num_discard_bios = 1;
@@ -765,7 +773,7 @@ static void pb_resume(struct dm_target *ti)
static struct target_type pb_target = {
.name = "push_backup",
.version = {1, 0, 0},
- .features = DM_TARGET_SINGLETON|DM_TARGET_IMMUTABLE,
+ .features = DM_TARGET_SINGLETON,
.module = THIS_MODULE,
.ctr = pb_ctr,
.dtr = pb_dtr,
@@ -780,18 +788,28 @@ static struct target_type pb_target = {
static int __init dm_pb_init(void)
{
- int r = -ENOMEM;
+ int r;
+
+ pbio_cache = kmem_cache_create("pb_bio", sizeof(struct pb_bio), 0, 0, NULL);
+ if (!pbio_cache)
+ return -ENOMEM;
r = dm_register_target(&pb_target);
- if (r)
+ if (r) {
DMERR("pb target registration failed: %d", r);
-
+ goto err;
+ }
+out:
return r;
+err:
+ kmem_cache_destroy(pbio_cache);
+ goto out;
}
static void __exit dm_pb_exit(void)
{
dm_unregister_target(&pb_target);
+ kmem_cache_destroy(pbio_cache);
}
module_init(dm_pb_init);
More information about the Devel
mailing list