[PATCH v3 0/2] blk-mq: fix update nr_requests regressions

[PATCH v3 0/2] blk-mq: fix update nr_requests regressions

[Yu Kuai]

From: Yu Kuai <yukuai3@huawei.com>

Changes in v3:
 - call depth_updated() directly in init_sched() method in patch 1;
 - fix typos in patch 2;
 - add review for patch 2;
Changes in v2:
 - instead of refactor and cleanups and fix updating nr_requests
 thoroughly, fix the regression in patch 2 the easy way, and dealy
 refactor and cleanups to next merge window.

patch 1 fix regression that elevator async_depth is not updated correctly
if nr_requests changes, first from error path and then for mq-deadline,
and recently for bfq and kyber.

patch 2 fix regression that if nr_requests grow, kernel will panic due
to tags double free.

Yu Kuai (2):
  blk-mq: fix elevator depth_updated method
  blk-mq: fix blk_mq_tags double free while nr_requests grown

 block/bfq-iosched.c   | 22 +++++-----------------
 block/blk-mq-sched.h  | 11 +++++++++++
 block/blk-mq-tag.c    |  1 +
 block/blk-mq.c        | 23 ++++++++++++-----------
 block/elevator.h      |  2 +-
 block/kyber-iosched.c | 19 +++++++++----------
 block/mq-deadline.c   | 16 +++-------------
 7 files changed, 42 insertions(+), 52 deletions(-)

-- 
2.39.2

[PATCH v3 1/2] blk-mq: fix elevator depth_updated method

[Yu Kuai]

From: Yu Kuai <yukuai3@huawei.com>

Current depth_updated has some problems:

1) depth_updated() will be called for each hctx, while all elevators
will update async_depth for the disk level, this is not related to hctx;
2) In blk_mq_update_nr_requests(), if previous hctx update succeed and
this hctx update failed, q->nr_requests will not be updated, while
async_depth is already updated with new nr_reqeuests in previous
depth_updated();
3) All elevators are using q->nr_requests to calculate async_depth now,
however, q->nr_requests is still the old value when depth_updated() is
called from blk_mq_update_nr_requests();

Those problems are first from error path, then mq-deadline, and recently
for bfq and kyber, fix those problems by:

- pass in request_queue instead of hctx;
- move depth_updated() after q->nr_requests is updated in
  blk_mq_update_nr_requests();
- add depth_updated() call inside init_sched() method to initialize
  async_depth;
- remove init_hctx() method for mq-deadline and bfq that is useless now;

Fixes: 77f1e0a52d26 ("bfq: update internal depth state when queue depth changes")
Fixes: 39823b47bbd4 ("block/mq-deadline: Fix the tag reservation code")
Fixes: 42e6c6ce03fd ("lib/sbitmap: convert shallow_depth from one word to the whole sbitmap")
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
---
 block/bfq-iosched.c   | 22 +++++-----------------
 block/blk-mq-sched.h  | 11 +++++++++++
 block/blk-mq.c        | 23 ++++++++++++-----------
 block/elevator.h      |  2 +-
 block/kyber-iosched.c | 19 +++++++++----------
 block/mq-deadline.c   | 16 +++-------------
 6 files changed, 41 insertions(+), 52 deletions(-)

diff --git a/block/bfq-iosched.c b/block/bfq-iosched.c
index 50e51047e1fe..4a8d3d96bfe4 100644
--- a/block/bfq-iosched.c
+++ b/block/bfq-iosched.c
@@ -7109,9 +7109,10 @@ void bfq_put_async_queues(struct bfq_data *bfqd, struct bfq_group *bfqg)
  * See the comments on bfq_limit_depth for the purpose of
  * the depths set in the function. Return minimum shallow depth we'll use.
  */
-static void bfq_update_depths(struct bfq_data *bfqd, struct sbitmap_queue *bt)
+static void bfq_depth_updated(struct request_queue *q)
 {
-	unsigned int nr_requests = bfqd->queue->nr_requests;
+	struct bfq_data *bfqd = q->elevator->elevator_data;
+	unsigned int nr_requests = q->nr_requests;
 
 	/*
 	 * In-word depths if no bfq_queue is being weight-raised:
@@ -7143,21 +7144,8 @@ static void bfq_update_depths(struct bfq_data *bfqd, struct sbitmap_queue *bt)
 	bfqd->async_depths[1][0] = max((nr_requests * 3) >> 4, 1U);
 	/* no more than ~37% of tags for sync writes (~20% extra tags) */
 	bfqd->async_depths[1][1] = max((nr_requests * 6) >> 4, 1U);
-}
-
-static void bfq_depth_updated(struct blk_mq_hw_ctx *hctx)
-{
-	struct bfq_data *bfqd = hctx->queue->elevator->elevator_data;
-	struct blk_mq_tags *tags = hctx->sched_tags;
 
-	bfq_update_depths(bfqd, &tags->bitmap_tags);
-	sbitmap_queue_min_shallow_depth(&tags->bitmap_tags, 1);
-}
-
-static int bfq_init_hctx(struct blk_mq_hw_ctx *hctx, unsigned int index)
-{
-	bfq_depth_updated(hctx);
-	return 0;
+	blk_mq_set_min_shallow_depth(q, 1);
 }
 
 static void bfq_exit_queue(struct elevator_queue *e)
@@ -7369,6 +7357,7 @@ static int bfq_init_queue(struct request_queue *q, struct elevator_queue *eq)
 		goto out_free;
 	bfq_init_root_group(bfqd->root_group, bfqd);
 	bfq_init_entity(&bfqd->oom_bfqq.entity, bfqd->root_group);
+	bfq_depth_updated(q);
 
 	/* We dispatch from request queue wide instead of hw queue */
 	blk_queue_flag_set(QUEUE_FLAG_SQ_SCHED, q);
@@ -7628,7 +7617,6 @@ static struct elevator_type iosched_bfq_mq = {
 		.request_merged		= bfq_request_merged,
 		.has_work		= bfq_has_work,
 		.depth_updated		= bfq_depth_updated,
-		.init_hctx		= bfq_init_hctx,
 		.init_sched		= bfq_init_queue,
 		.exit_sched		= bfq_exit_queue,
 	},
diff --git a/block/blk-mq-sched.h b/block/blk-mq-sched.h
index b554e1d55950..fe83187f41db 100644
--- a/block/blk-mq-sched.h
+++ b/block/blk-mq-sched.h
@@ -92,4 +92,15 @@ static inline bool blk_mq_sched_needs_restart(struct blk_mq_hw_ctx *hctx)
 	return test_bit(BLK_MQ_S_SCHED_RESTART, &hctx->state);
 }
 
+static inline void blk_mq_set_min_shallow_depth(struct request_queue *q,
+						unsigned int depth)
+{
+	struct blk_mq_hw_ctx *hctx;
+	unsigned long i;
+
+	queue_for_each_hw_ctx(q, hctx, i)
+		sbitmap_queue_min_shallow_depth(&hctx->sched_tags->bitmap_tags,
+						depth);
+}
+
 #endif
diff --git a/block/blk-mq.c b/block/blk-mq.c
index b67d6c02eceb..9c68749124c6 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -4951,20 +4951,21 @@ int blk_mq_update_nr_requests(struct request_queue *q, unsigned int nr)
 						      false);
 		}
 		if (ret)
-			break;
-		if (q->elevator && q->elevator->type->ops.depth_updated)
-			q->elevator->type->ops.depth_updated(hctx);
+			goto out;
 	}
-	if (!ret) {
-		q->nr_requests = nr;
-		if (blk_mq_is_shared_tags(set->flags)) {
-			if (q->elevator)
-				blk_mq_tag_update_sched_shared_tags(q);
-			else
-				blk_mq_tag_resize_shared_tags(set, nr);
-		}
+
+	q->nr_requests = nr;
+	if (q->elevator && q->elevator->type->ops.depth_updated)
+		q->elevator->type->ops.depth_updated(q);
+
+	if (blk_mq_is_shared_tags(set->flags)) {
+		if (q->elevator)
+			blk_mq_tag_update_sched_shared_tags(q);
+		else
+			blk_mq_tag_resize_shared_tags(set, nr);
 	}
 
+out:
 	blk_mq_unquiesce_queue(q);
 
 	return ret;
diff --git a/block/elevator.h b/block/elevator.h
index adc5c157e17e..c4d20155065e 100644
--- a/block/elevator.h
+++ b/block/elevator.h
@@ -37,7 +37,7 @@ struct elevator_mq_ops {
 	void (*exit_sched)(struct elevator_queue *);
 	int (*init_hctx)(struct blk_mq_hw_ctx *, unsigned int);
 	void (*exit_hctx)(struct blk_mq_hw_ctx *, unsigned int);
-	void (*depth_updated)(struct blk_mq_hw_ctx *);
+	void (*depth_updated)(struct request_queue *);
 
 	bool (*allow_merge)(struct request_queue *, struct request *, struct bio *);
 	bool (*bio_merge)(struct request_queue *, struct bio *, unsigned int);
diff --git a/block/kyber-iosched.c b/block/kyber-iosched.c
index 70cbc7b2deb4..18efd6ef2a2b 100644
--- a/block/kyber-iosched.c
+++ b/block/kyber-iosched.c
@@ -399,6 +399,14 @@ static struct kyber_queue_data *kyber_queue_data_alloc(struct request_queue *q)
 	return ERR_PTR(ret);
 }
 
+static void kyber_depth_updated(struct request_queue *q)
+{
+	struct kyber_queue_data *kqd = q->elevator->elevator_data;
+
+	kqd->async_depth = q->nr_requests * KYBER_ASYNC_PERCENT / 100U;
+	blk_mq_set_min_shallow_depth(q, kqd->async_depth);
+}
+
 static int kyber_init_sched(struct request_queue *q, struct elevator_queue *eq)
 {
 	struct kyber_queue_data *kqd;
@@ -413,6 +421,7 @@ static int kyber_init_sched(struct request_queue *q, struct elevator_queue *eq)
 
 	eq->elevator_data = kqd;
 	q->elevator = eq;
+	kyber_depth_updated(q);
 
 	return 0;
 }
@@ -440,15 +449,6 @@ static void kyber_ctx_queue_init(struct kyber_ctx_queue *kcq)
 		INIT_LIST_HEAD(&kcq->rq_list[i]);
 }
 
-static void kyber_depth_updated(struct blk_mq_hw_ctx *hctx)
-{
-	struct kyber_queue_data *kqd = hctx->queue->elevator->elevator_data;
-	struct blk_mq_tags *tags = hctx->sched_tags;
-
-	kqd->async_depth = hctx->queue->nr_requests * KYBER_ASYNC_PERCENT / 100U;
-	sbitmap_queue_min_shallow_depth(&tags->bitmap_tags, kqd->async_depth);
-}
-
 static int kyber_init_hctx(struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx)
 {
 	struct kyber_hctx_data *khd;
@@ -493,7 +493,6 @@ static int kyber_init_hctx(struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx)
 	khd->batching = 0;
 
 	hctx->sched_data = khd;
-	kyber_depth_updated(hctx);
 
 	return 0;
 
diff --git a/block/mq-deadline.c b/block/mq-deadline.c
index b9b7cdf1d3c9..2e689b2c4021 100644
--- a/block/mq-deadline.c
+++ b/block/mq-deadline.c
@@ -507,22 +507,12 @@ static void dd_limit_depth(blk_opf_t opf, struct blk_mq_alloc_data *data)
 }
 
 /* Called by blk_mq_update_nr_requests(). */
-static void dd_depth_updated(struct blk_mq_hw_ctx *hctx)
+static void dd_depth_updated(struct request_queue *q)
 {
-	struct request_queue *q = hctx->queue;
 	struct deadline_data *dd = q->elevator->elevator_data;
-	struct blk_mq_tags *tags = hctx->sched_tags;
 
 	dd->async_depth = q->nr_requests;
-
-	sbitmap_queue_min_shallow_depth(&tags->bitmap_tags, 1);
-}
-
-/* Called by blk_mq_init_hctx() and blk_mq_init_sched(). */
-static int dd_init_hctx(struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx)
-{
-	dd_depth_updated(hctx);
-	return 0;
+	blk_mq_set_min_shallow_depth(q, 1);
 }
 
 static void dd_exit_sched(struct elevator_queue *e)
@@ -587,6 +577,7 @@ static int dd_init_sched(struct request_queue *q, struct elevator_queue *eq)
 	blk_queue_flag_set(QUEUE_FLAG_SQ_SCHED, q);
 
 	q->elevator = eq;
+	dd_depth_updated(q);
 	return 0;
 }
 
@@ -1048,7 +1039,6 @@ static struct elevator_type mq_deadline = {
 		.has_work		= dd_has_work,
 		.init_sched		= dd_init_sched,
 		.exit_sched		= dd_exit_sched,
-		.init_hctx		= dd_init_hctx,
 	},
 
 #ifdef CONFIG_BLK_DEBUG_FS
-- 
2.39.2

Re: [PATCH v3 1/2] blk-mq: fix elevator depth_updated method

[Nilay Shroff]

On 8/21/25 11:36 AM, Yu Kuai wrote:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> Current depth_updated has some problems:
> 
> 1) depth_updated() will be called for each hctx, while all elevators
> will update async_depth for the disk level, this is not related to hctx;
> 2) In blk_mq_update_nr_requests(), if previous hctx update succeed and
> this hctx update failed, q->nr_requests will not be updated, while
> async_depth is already updated with new nr_reqeuests in previous
> depth_updated();
> 3) All elevators are using q->nr_requests to calculate async_depth now,
> however, q->nr_requests is still the old value when depth_updated() is
> called from blk_mq_update_nr_requests();
> 
> Those problems are first from error path, then mq-deadline, and recently
> for bfq and kyber, fix those problems by:
> 
> - pass in request_queue instead of hctx;
> - move depth_updated() after q->nr_requests is updated in
>   blk_mq_update_nr_requests();
> - add depth_updated() call inside init_sched() method to initialize
>   async_depth;
> - remove init_hctx() method for mq-deadline and bfq that is useless now;
> 
> Fixes: 77f1e0a52d26 ("bfq: update internal depth state when queue depth changes")
> Fixes: 39823b47bbd4 ("block/mq-deadline: Fix the tag reservation code")
> Fixes: 42e6c6ce03fd ("lib/sbitmap: convert shallow_depth from one word to the whole sbitmap")
> Signed-off-by: Yu Kuai <yukuai3@huawei.com>

Looks good to me:
Reviewed-by: Nilay Shroff <nilay@linux.ibm.com>

Re: [PATCH v3 1/2] blk-mq: fix elevator depth_updated method

[Li Nan]

在 2025/8/21 14:06, Yu Kuai 写道:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> Current depth_updated has some problems:
> 
> 1) depth_updated() will be called for each hctx, while all elevators
> will update async_depth for the disk level, this is not related to hctx;
> 2) In blk_mq_update_nr_requests(), if previous hctx update succeed and
> this hctx update failed, q->nr_requests will not be updated, while
> async_depth is already updated with new nr_reqeuests in previous
> depth_updated();
> 3) All elevators are using q->nr_requests to calculate async_depth now,
> however, q->nr_requests is still the old value when depth_updated() is
> called from blk_mq_update_nr_requests();
> 
> Those problems are first from error path, then mq-deadline, and recently
> for bfq and kyber, fix those problems by:
> 
> - pass in request_queue instead of hctx;
> - move depth_updated() after q->nr_requests is updated in
>    blk_mq_update_nr_requests();
> - add depth_updated() call inside init_sched() method to initialize
>    async_depth;
> - remove init_hctx() method for mq-deadline and bfq that is useless now;
> 
> Fixes: 77f1e0a52d26 ("bfq: update internal depth state when queue depth changes")
> Fixes: 39823b47bbd4 ("block/mq-deadline: Fix the tag reservation code")
> Fixes: 42e6c6ce03fd ("lib/sbitmap: convert shallow_depth from one word to the whole sbitmap")
> Signed-off-by: Yu Kuai <yukuai3@huawei.com>
> ---
>   block/bfq-iosched.c   | 22 +++++-----------------
>   block/blk-mq-sched.h  | 11 +++++++++++
>   block/blk-mq.c        | 23 ++++++++++++-----------
>   block/elevator.h      |  2 +-
>   block/kyber-iosched.c | 19 +++++++++----------
>   block/mq-deadline.c   | 16 +++-------------
>   6 files changed, 41 insertions(+), 52 deletions(-)
> 

LGTM

Reviewed-by: Li Nan <linan122@huawei.com>

-- 
Thanks,
Nan

Re: [PATCH v3 1/2] blk-mq: fix elevator depth_updated method

[Hannes Reinecke]

On 8/21/25 08:06, Yu Kuai wrote:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> Current depth_updated has some problems:
> 
> 1) depth_updated() will be called for each hctx, while all elevators
> will update async_depth for the disk level, this is not related to hctx;
> 2) In blk_mq_update_nr_requests(), if previous hctx update succeed and
> this hctx update failed, q->nr_requests will not be updated, while
> async_depth is already updated with new nr_reqeuests in previous
> depth_updated();
> 3) All elevators are using q->nr_requests to calculate async_depth now,
> however, q->nr_requests is still the old value when depth_updated() is
> called from blk_mq_update_nr_requests();
> 
> Those problems are first from error path, then mq-deadline, and recently
> for bfq and kyber, fix those problems by:
> 
> - pass in request_queue instead of hctx;
> - move depth_updated() after q->nr_requests is updated in
>    blk_mq_update_nr_requests();
> - add depth_updated() call inside init_sched() method to initialize
>    async_depth;
> - remove init_hctx() method for mq-deadline and bfq that is useless now;
> 
> Fixes: 77f1e0a52d26 ("bfq: update internal depth state when queue depth changes")
> Fixes: 39823b47bbd4 ("block/mq-deadline: Fix the tag reservation code")
> Fixes: 42e6c6ce03fd ("lib/sbitmap: convert shallow_depth from one word to the whole sbitmap")
> Signed-off-by: Yu Kuai <yukuai3@huawei.com>
> ---
>   block/bfq-iosched.c   | 22 +++++-----------------
>   block/blk-mq-sched.h  | 11 +++++++++++
>   block/blk-mq.c        | 23 ++++++++++++-----------
>   block/elevator.h      |  2 +-
>   block/kyber-iosched.c | 19 +++++++++----------
>   block/mq-deadline.c   | 16 +++-------------
>   6 files changed, 41 insertions(+), 52 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

[PATCH v3 2/2] blk-mq: fix blk_mq_tags double free while nr_requests grown

[Yu Kuai]

From: Yu Kuai <yukuai3@huawei.com>

In the case user trigger tags grow by queue sysfs attribute nr_requests,
hctx->sched_tags will be freed directly and replaced with a new
allocated tags, see blk_mq_tag_update_depth().

The problem is that hctx->sched_tags is from elevator->et->tags, while
et->tags is still the freed tags, hence later elevator exit will try to
free the tags again, causing kernel panic.

Fix this problem by replacing et->tags with new allocated tags as well.

Noted there are still some long term problems that will require some
refactor to be fixed thoroughly[1].

[1] https://lore.kernel.org/all/20250815080216.410665-1-yukuai1@huaweicloud.com/
Fixes: f5a6604f7a44 ("block: fix lockdep warning caused by lock dependency in elv_iosched_store")

Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Nilay Shroff<nilay@linux.ibm.com>
---
 block/blk-mq-tag.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c
index d880c50629d6..5cffa5668d0c 100644
--- a/block/blk-mq-tag.c
+++ b/block/blk-mq-tag.c
@@ -622,6 +622,7 @@ int blk_mq_tag_update_depth(struct blk_mq_hw_ctx *hctx,
 			return -ENOMEM;
 
 		blk_mq_free_map_and_rqs(set, *tagsptr, hctx->queue_num);
+		hctx->queue->elevator->et->tags[hctx->queue_num] = new;
 		*tagsptr = new;
 	} else {
 		/*
-- 
2.39.2

Re: [PATCH v3 2/2] blk-mq: fix blk_mq_tags double free while nr_requests grown

[Li Nan]

在 2025/8/21 14:06, Yu Kuai 写道:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> In the case user trigger tags grow by queue sysfs attribute nr_requests,
> hctx->sched_tags will be freed directly and replaced with a new
> allocated tags, see blk_mq_tag_update_depth().
> 
> The problem is that hctx->sched_tags is from elevator->et->tags, while
> et->tags is still the freed tags, hence later elevator exit will try to
> free the tags again, causing kernel panic.
> 
> Fix this problem by replacing et->tags with new allocated tags as well.
> 
> Noted there are still some long term problems that will require some
> refactor to be fixed thoroughly[1].
> 
> [1] https://lore.kernel.org/all/20250815080216.410665-1-yukuai1@huaweicloud.com/
> Fixes: f5a6604f7a44 ("block: fix lockdep warning caused by lock dependency in elv_iosched_store")
> 
> Signed-off-by: Yu Kuai <yukuai3@huawei.com>
> Reviewed-by: Ming Lei <ming.lei@redhat.com>
> Reviewed-by: Nilay Shroff<nilay@linux.ibm.com>
> ---

LGTM

Reviewed-by: Li Nan <linan122@huawei.com>

-- 
Thanks,
Nan

Re: [PATCH v3 2/2] blk-mq: fix blk_mq_tags double free while nr_requests grown

[Hannes Reinecke]

On 8/21/25 08:06, Yu Kuai wrote:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> In the case user trigger tags grow by queue sysfs attribute nr_requests,
> hctx->sched_tags will be freed directly and replaced with a new
> allocated tags, see blk_mq_tag_update_depth().
> 
> The problem is that hctx->sched_tags is from elevator->et->tags, while
> et->tags is still the freed tags, hence later elevator exit will try to
> free the tags again, causing kernel panic.
> 
> Fix this problem by replacing et->tags with new allocated tags as well.
> 
> Noted there are still some long term problems that will require some
> refactor to be fixed thoroughly[1].
> 
> [1] https://lore.kernel.org/all/20250815080216.410665-1-yukuai1@huaweicloud.com/
> Fixes: f5a6604f7a44 ("block: fix lockdep warning caused by lock dependency in elv_iosched_store")
> 
> Signed-off-by: Yu Kuai <yukuai3@huawei.com>
> Reviewed-by: Ming Lei <ming.lei@redhat.com>
> Reviewed-by: Nilay Shroff<nilay@linux.ibm.com>
> ---
>   block/blk-mq-tag.c | 1 +
>   1 file changed, 1 insertion(+)
> 
> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c
> index d880c50629d6..5cffa5668d0c 100644
> --- a/block/blk-mq-tag.c
> +++ b/block/blk-mq-tag.c
> @@ -622,6 +622,7 @@ int blk_mq_tag_update_depth(struct blk_mq_hw_ctx *hctx,
>   			return -ENOMEM;
>   
>   		blk_mq_free_map_and_rqs(set, *tagsptr, hctx->queue_num);
> +		hctx->queue->elevator->et->tags[hctx->queue_num] = new;
>   		*tagsptr = new;
>   	} else {
>   		/*

Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH v3 0/2] blk-mq: fix update nr_requests regressions

[Yu Kuai]

Hi, Jens

在 2025/08/21 14:06, Yu Kuai 写道:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> Changes in v3:
>   - call depth_updated() directly in init_sched() method in patch 1;
>   - fix typos in patch 2;
>   - add review for patch 2;
> Changes in v2:
>   - instead of refactor and cleanups and fix updating nr_requests
>   thoroughly, fix the regression in patch 2 the easy way, and dealy
>   refactor and cleanups to next merge window.
> 
> patch 1 fix regression that elevator async_depth is not updated correctly
> if nr_requests changes, first from error path and then for mq-deadline,
> and recently for bfq and kyber.
> 
> patch 2 fix regression that if nr_requests grow, kernel will panic due
> to tags double free.
> 
> Yu Kuai (2):
>    blk-mq: fix elevator depth_updated method
>    blk-mq: fix blk_mq_tags double free while nr_requests grown
> 
>   block/bfq-iosched.c   | 22 +++++-----------------
>   block/blk-mq-sched.h  | 11 +++++++++++
>   block/blk-mq-tag.c    |  1 +
>   block/blk-mq.c        | 23 ++++++++++++-----------
>   block/elevator.h      |  2 +-
>   block/kyber-iosched.c | 19 +++++++++----------
>   block/mq-deadline.c   | 16 +++-------------
>   7 files changed, 42 insertions(+), 52 deletions(-)
> 

Friendly ping, please consider this set in this merge window.

BTW, I see that for-6.18/block branch was created, however, I have
a pending set[1] for the next merge window that will have conflicts with
this set, not sure if you want to rebase for-6.18/block with block-6.17
or handle conflicts later for 6.18-rc1.

[1] 
https://lore.kernel.org/all/20250815080216.410665-1-yukuai1@huaweicloud.com/

Thanks,
Kuai

Re: [PATCH v3 0/2] blk-mq: fix update nr_requests regressions

[Yu Kuai]

Hi, Jens

在 2025/08/26 14:27, Yu Kuai 写道:
> Hi, Jens
> 
> 在 2025/08/21 14:06, Yu Kuai 写道:
>> From: Yu Kuai <yukuai3@huawei.com>
>>
>> Changes in v3:
>>   - call depth_updated() directly in init_sched() method in patch 1;
>>   - fix typos in patch 2;
>>   - add review for patch 2;
>> Changes in v2:
>>   - instead of refactor and cleanups and fix updating nr_requests
>>   thoroughly, fix the regression in patch 2 the easy way, and dealy
>>   refactor and cleanups to next merge window.
>>
>> patch 1 fix regression that elevator async_depth is not updated correctly
>> if nr_requests changes, first from error path and then for mq-deadline,
>> and recently for bfq and kyber.
>>
>> patch 2 fix regression that if nr_requests grow, kernel will panic due
>> to tags double free.
>>
>> Yu Kuai (2):
>>    blk-mq: fix elevator depth_updated method
>>    blk-mq: fix blk_mq_tags double free while nr_requests grown
>>
>>   block/bfq-iosched.c   | 22 +++++-----------------
>>   block/blk-mq-sched.h  | 11 +++++++++++
>>   block/blk-mq-tag.c    |  1 +
>>   block/blk-mq.c        | 23 ++++++++++++-----------
>>   block/elevator.h      |  2 +-
>>   block/kyber-iosched.c | 19 +++++++++----------
>>   block/mq-deadline.c   | 16 +++-------------
>>   7 files changed, 42 insertions(+), 52 deletions(-)
>>
> 
> Friendly ping, please consider this set in this merge window.
> 
> BTW, I see that for-6.18/block branch was created, however, I have
> a pending set[1] for the next merge window that will have conflicts with
> this set, not sure if you want to rebase for-6.18/block with block-6.17
> or handle conflicts later for 6.18-rc1.
> 
> [1] 
> https://lore.kernel.org/all/20250815080216.410665-1-yukuai1@huaweicloud.com/ 
> 
Friendly ping ...

Thanks,
Kuai

> 
> Thanks,
> Kuai
> 
> .
> 

Re: [PATCH v3 0/2] blk-mq: fix update nr_requests regressions

[Jens Axboe]

On 9/5/25 1:20 AM, Yu Kuai wrote:
> Hi, Jens
> 
> ? 2025/08/26 14:27, Yu Kuai ??:
>> Hi, Jens
>>
>> ? 2025/08/21 14:06, Yu Kuai ??:
>>> From: Yu Kuai <yukuai3@huawei.com>
>>>
>>> Changes in v3:
>>>   - call depth_updated() directly in init_sched() method in patch 1;
>>>   - fix typos in patch 2;
>>>   - add review for patch 2;
>>> Changes in v2:
>>>   - instead of refactor and cleanups and fix updating nr_requests
>>>   thoroughly, fix the regression in patch 2 the easy way, and dealy
>>>   refactor and cleanups to next merge window.
>>>
>>> patch 1 fix regression that elevator async_depth is not updated correctly
>>> if nr_requests changes, first from error path and then for mq-deadline,
>>> and recently for bfq and kyber.
>>>
>>> patch 2 fix regression that if nr_requests grow, kernel will panic due
>>> to tags double free.
>>>
>>> Yu Kuai (2):
>>>    blk-mq: fix elevator depth_updated method
>>>    blk-mq: fix blk_mq_tags double free while nr_requests grown
>>>
>>>   block/bfq-iosched.c   | 22 +++++-----------------
>>>   block/blk-mq-sched.h  | 11 +++++++++++
>>>   block/blk-mq-tag.c    |  1 +
>>>   block/blk-mq.c        | 23 ++++++++++++-----------
>>>   block/elevator.h      |  2 +-
>>>   block/kyber-iosched.c | 19 +++++++++----------
>>>   block/mq-deadline.c   | 16 +++-------------
>>>   7 files changed, 42 insertions(+), 52 deletions(-)
>>>
>>
>> Friendly ping, please consider this set in this merge window.
>>
>> BTW, I see that for-6.18/block branch was created, however, I have
>> a pending set[1] for the next merge window that will have conflicts with
>> this set, not sure if you want to rebase for-6.18/block with block-6.17
>> or handle conflicts later for 6.18-rc1.

I think we're just a bit late on this one, given that they'd go into
-rc6 at this point. Going to queue this up for 6.18 and then we just get
it into stable instead, that gives us a lot more time to shake out any
potential issues.

-- 
Jens Axboe

Re: [PATCH v3 0/2] blk-mq: fix update nr_requests regressions

[Jens Axboe]

On Thu, 21 Aug 2025 14:06:10 +0800, Yu Kuai wrote:
> Changes in v3:
>  - call depth_updated() directly in init_sched() method in patch 1;
>  - fix typos in patch 2;
>  - add review for patch 2;
> Changes in v2:
>  - instead of refactor and cleanups and fix updating nr_requests
>  thoroughly, fix the regression in patch 2 the easy way, and dealy
>  refactor and cleanups to next merge window.
> 
> [...]

Applied, thanks!

[1/2] blk-mq: fix elevator depth_updated method
      commit: 7d337eef4affc5e26e0570513168c69ddbc40f92
[2/2] blk-mq: fix blk_mq_tags double free while nr_requests grown
      commit: ba28afbd9eff2a6370f23ef4e6a036ab0cfda409

Best regards,
-- 
Jens Axboe