[PATCH v5 25/45] md/raid5: dynamically allocate the md-raid5 shrinker

Subsystems: software raid (multiple disks) support, the rest

STALE1132d REVIEWED: 1 (0M)

Revision v5 of 6 in this series; 1 review trailer.

2 messages, 2 authors, 2023-08-24 · open the first message on its own page

[PATCH v5 25/45] md/raid5: dynamically allocate the md-raid5 shrinker

From: Qi Zheng <hidden>
Date: 2023-08-24 03:50:15

In preparation for implementing lockless slab shrink, use new APIs to
dynamically allocate the md-raid5 shrinker, so that it can be freed
asynchronously via RCU. Then it doesn't need to wait for RCU read-side
critical section when releasing the struct r5conf.

Signed-off-by: Qi Zheng <redacted>
Reviewed-by: Muchun Song <redacted>
CC: Song Liu <song@kernel.org>
CC: linux-raid@vger.kernel.org
---
 drivers/md/raid5.c | 26 +++++++++++++++-----------
 drivers/md/raid5.h |  2 +-
 2 files changed, 16 insertions(+), 12 deletions(-)
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 4cb9c608ee19..c8d2c6e50aa1 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -7401,7 +7401,7 @@ static void free_conf(struct r5conf *conf)
 
 	log_exit(conf);
 
-	unregister_shrinker(&conf->shrinker);
+	shrinker_free(conf->shrinker);
 	free_thread_groups(conf);
 	shrink_stripes(conf);
 	raid5_free_percpu(conf);
@@ -7449,7 +7449,7 @@ static int raid5_alloc_percpu(struct r5conf *conf)
 static unsigned long raid5_cache_scan(struct shrinker *shrink,
 				      struct shrink_control *sc)
 {
-	struct r5conf *conf = container_of(shrink, struct r5conf, shrinker);
+	struct r5conf *conf = shrink->private_data;
 	unsigned long ret = SHRINK_STOP;
 
 	if (mutex_trylock(&conf->cache_size_mutex)) {
@@ -7470,7 +7470,7 @@ static unsigned long raid5_cache_scan(struct shrinker *shrink,
 static unsigned long raid5_cache_count(struct shrinker *shrink,
 				       struct shrink_control *sc)
 {
-	struct r5conf *conf = container_of(shrink, struct r5conf, shrinker);
+	struct r5conf *conf = shrink->private_data;
 
 	if (conf->max_nr_stripes < conf->min_nr_stripes)
 		/* unlikely, but not impossible */
@@ -7705,18 +7705,22 @@ static struct r5conf *setup_conf(struct mddev *mddev)
 	 * it reduces the queue depth and so can hurt throughput.
 	 * So set it rather large, scaled by number of devices.
 	 */
-	conf->shrinker.seeks = DEFAULT_SEEKS * conf->raid_disks * 4;
-	conf->shrinker.scan_objects = raid5_cache_scan;
-	conf->shrinker.count_objects = raid5_cache_count;
-	conf->shrinker.batch = 128;
-	conf->shrinker.flags = 0;
-	ret = register_shrinker(&conf->shrinker, "md-raid5:%s", mdname(mddev));
-	if (ret) {
-		pr_warn("md/raid:%s: couldn't register shrinker.\n",
+	conf->shrinker = shrinker_alloc(0, "md-raid5:%s", mdname(mddev));
+	if (!conf->shrinker) {
+		ret = -ENOMEM;
+		pr_warn("md/raid:%s: couldn't allocate shrinker.\n",
 			mdname(mddev));
 		goto abort;
 	}
 
+	conf->shrinker->seeks = DEFAULT_SEEKS * conf->raid_disks * 4;
+	conf->shrinker->scan_objects = raid5_cache_scan;
+	conf->shrinker->count_objects = raid5_cache_count;
+	conf->shrinker->batch = 128;
+	conf->shrinker->private_data = conf;
+
+	shrinker_register(conf->shrinker);
+
 	sprintf(pers_name, "raid%d", mddev->new_level);
 	rcu_assign_pointer(conf->thread,
 			   md_register_thread(raid5d, mddev, pers_name));
diff --git a/drivers/md/raid5.h b/drivers/md/raid5.h
index 97a795979a35..22bea20eccbd 100644
--- a/drivers/md/raid5.h
+++ b/drivers/md/raid5.h
@@ -670,7 +670,7 @@ struct r5conf {
 	wait_queue_head_t	wait_for_stripe;
 	wait_queue_head_t	wait_for_overlap;
 	unsigned long		cache_state;
-	struct shrinker		shrinker;
+	struct shrinker		*shrinker;
 	int			pool_size; /* number of disks in stripeheads in pool */
 	spinlock_t		device_lock;
 	struct disk_info	*disks;
-- 
2.30.2

Re: [PATCH v5 25/45] md/raid5: dynamically allocate the md-raid5 shrinker

From: Song Liu <song@kernel.org>
Date: 2023-08-24 03:58:34

On Wed, Aug 23, 2023 at 8:47 PM Qi Zheng [off-list ref] wrote:
In preparation for implementing lockless slab shrink, use new APIs to
dynamically allocate the md-raid5 shrinker, so that it can be freed
asynchronously via RCU. Then it doesn't need to wait for RCU read-side
critical section when releasing the struct r5conf.

Signed-off-by: Qi Zheng <redacted>
Reviewed-by: Muchun Song <redacted>
CC: Song Liu <song@kernel.org>
CC: linux-raid@vger.kernel.org
LGTM!

Reviewed-by: Song Liu <song@kernel.org>
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help