From: Ming Lei <ming.lei@redhat.com>
To: Jens Axboe <axboe@kernel.dk>, linux-block@vger.kernel.org
Cc: "Nilay Shroff" <nilay@linux.ibm.com>,
"Shinichiro Kawasaki" <shinichiro.kawasaki@wdc.com>,
"Thomas Hellström" <thomas.hellstrom@linux.intel.com>,
"Christoph Hellwig" <hch@lst.de>,
"Ming Lei" <ming.lei@redhat.com>,
"Hannes Reinecke" <hare@suse.de>
Subject: [PATCH V5 22/25] block: move hctx debugfs/sysfs registering out of freezing queue
Date: Mon, 5 May 2025 22:18:00 +0800 [thread overview]
Message-ID: <20250505141805.2751237-23-ming.lei@redhat.com> (raw)
In-Reply-To: <20250505141805.2751237-1-ming.lei@redhat.com>
Move hctx debugfs/sysfs register out of freezing queue in
__blk_mq_update_nr_hw_queues(), so that the following lockdep dependency
can be killed:
#2 (&q->q_usage_counter(io)#16){++++}-{0:0}:
#1 (fs_reclaim){+.+.}-{0:0}:
#0 (&sb->s_type->i_mutex_key#3){+.+.}-{4:4}: //debugfs
And registering/un-registering hctx debugfs/sysfs does not require queue to
be frozen:
- hctx sysfs attributes show() are drained when removing kobject, and
there isn't store() implementation for hctx sysfs attributes
- debugfs entry read() is drained too when removing debugfs directory,
and there isn't write() implementation for hctx debugfs too
- so it is safe to register/unregister hctx sysfs/debugfs without
freezing queue because the cod paths changes nothing, and we just
need to keep hctx live
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Nilay Shroff <nilay@linux.ibm.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
---
block/blk-mq.c | 15 +++++++--------
1 file changed, 7 insertions(+), 8 deletions(-)
diff --git a/block/blk-mq.c b/block/blk-mq.c
index d1b5b75840eb..32526d120bf4 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -5007,14 +5007,14 @@ static void __blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set,
return;
memflags = memalloc_noio_save();
- list_for_each_entry(q, &set->tag_list, tag_set_list)
- blk_mq_freeze_queue_nomemsave(q);
-
list_for_each_entry(q, &set->tag_list, tag_set_list) {
blk_mq_debugfs_unregister_hctxs(q);
blk_mq_sysfs_unregister_hctxs(q);
}
+ list_for_each_entry(q, &set->tag_list, tag_set_list)
+ blk_mq_freeze_queue_nomemsave(q);
+
if (blk_mq_realloc_tag_set_tags(set, nr_hw_queues) < 0)
goto reregister;
@@ -5037,16 +5037,15 @@ static void __blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set,
blk_mq_map_swqueue(q);
}
+ /* elv_update_nr_hw_queues() unfreeze queue for us */
+ list_for_each_entry(q, &set->tag_list, tag_set_list)
+ elv_update_nr_hw_queues(q);
+
reregister:
list_for_each_entry(q, &set->tag_list, tag_set_list) {
blk_mq_sysfs_register_hctxs(q);
blk_mq_debugfs_register_hctxs(q);
}
-
- /* elv_update_nr_hw_queues() unfreeze queue for us */
- list_for_each_entry(q, &set->tag_list, tag_set_list)
- elv_update_nr_hw_queues(q);
-
memalloc_noio_restore(memflags);
/* Free the excess tags when nr_hw_queues shrink. */
--
2.47.0
next prev parent reply other threads:[~2025-05-05 14:19 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-05-05 14:17 [PATCH V5 00/25] block: unify elevator changing and fix lockdep warning Ming Lei
2025-05-05 14:17 ` [PATCH V5 01/25] block: move blk_mq_add_queue_tag_set() after blk_mq_map_swqueue() Ming Lei
2025-05-05 14:17 ` [PATCH V5 02/25] block: move ELEVATOR_FLAG_DISABLE_WBT a request queue flag Ming Lei
2025-05-05 14:17 ` [PATCH V5 03/25] block: don't call freeze queue in elevator_switch() and elevator_disable() Ming Lei
2025-05-05 14:17 ` [PATCH V5 04/25] block: use q->elevator with ->elevator_lock held in elv_iosched_show() Ming Lei
2025-05-05 14:17 ` [PATCH V5 05/25] block: add two helpers for registering/un-registering sched debugfs Ming Lei
2025-05-05 14:17 ` [PATCH V5 06/25] block: move sched debugfs register into elvevator_register_queue Ming Lei
2025-05-05 14:17 ` [PATCH V5 07/25] block: add helper add_disk_final() Ming Lei
2025-05-06 4:40 ` Christoph Hellwig
2025-05-06 7:43 ` Nilay Shroff
2025-05-06 11:02 ` Hannes Reinecke
2025-05-05 14:17 ` [PATCH V5 08/25] block: prevent adding/deleting disk during updating nr_hw_queues Ming Lei
2025-05-05 14:17 ` [PATCH V5 09/25] block: don't allow to switch elevator if updating nr_hw_queues is in-progress Ming Lei
2025-05-06 4:41 ` Christoph Hellwig
2025-05-06 6:26 ` Nilay Shroff
2025-05-05 14:17 ` [PATCH V5 10/25] block: look up the elevator type in elevator_switch Ming Lei
2025-05-05 14:17 ` [PATCH V5 11/25] block: fold elevator_disable into elevator_switch Ming Lei
2025-05-05 14:17 ` [PATCH V5 12/25] block: move blk_queue_registered() check into elv_iosched_store() Ming Lei
2025-05-06 4:41 ` Christoph Hellwig
2025-05-06 7:47 ` Nilay Shroff
2025-05-05 14:17 ` [PATCH V5 13/25] block: simplify elevator reattachment for updating nr_hw_queues Ming Lei
2025-05-05 14:17 ` [PATCH V5 14/25] block: move queue freezing & elevator_lock into elevator_change() Ming Lei
2025-05-05 14:17 ` [PATCH V5 15/25] block: add `struct elv_change_ctx` for unifying elevator change Ming Lei
2025-05-06 4:42 ` Christoph Hellwig
2025-05-05 14:17 ` [PATCH V5 16/25] block: " Ming Lei
2025-05-05 14:17 ` [PATCH V5 17/25] block: pass elevator_queue to elv_register_queue & unregister_queue Ming Lei
2025-05-05 14:17 ` [PATCH V5 18/25] block: remove elevator queue's type check in elv_attr_show/store() Ming Lei
2025-05-05 14:17 ` [PATCH V5 19/25] block: fail to show/store elevator sysfs attribute if elevator is dying Ming Lei
2025-05-06 11:09 ` Hannes Reinecke
2025-05-05 14:17 ` [PATCH V5 20/25] block: add new helper for disabling elevator switch when deleting disk Ming Lei
2025-05-06 6:32 ` Nilay Shroff
2025-05-05 14:17 ` [PATCH V5 21/25] block: move elv_register[unregister]_queue out of elevator_lock Ming Lei
2025-05-06 4:43 ` Christoph Hellwig
2025-05-06 6:36 ` Nilay Shroff
2025-05-05 14:18 ` Ming Lei [this message]
2025-05-05 14:18 ` [PATCH V5 23/25] block: don't acquire ->elevator_lock in blk_mq_map_swqueue and blk_mq_realloc_hw_ctxs Ming Lei
2025-05-06 4:44 ` Christoph Hellwig
2025-05-05 14:18 ` [PATCH V5 24/25] block: move hctx cpuhp add/del out of queue freezing Ming Lei
2025-05-06 4:44 ` Christoph Hellwig
2025-05-05 14:18 ` [PATCH V5 25/25] block: move wbt_enable_default() out of queue freezing from sched ->exit() Ming Lei
2025-05-06 4:44 ` Christoph Hellwig
2025-05-06 13:48 ` [PATCH V5 00/25] block: unify elevator changing and fix lockdep warning Jens Axboe
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250505141805.2751237-23-ming.lei@redhat.com \
--to=ming.lei@redhat.com \
--cc=axboe@kernel.dk \
--cc=hare@suse.de \
--cc=hch@lst.de \
--cc=linux-block@vger.kernel.org \
--cc=nilay@linux.ibm.com \
--cc=shinichiro.kawasaki@wdc.com \
--cc=thomas.hellstrom@linux.intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox