block: reduce kblockd_mod_delayed_work_on() CPU consumption
authorJens Axboe <axboe@kernel.dk>
Tue, 14 Dec 2021 14:03:24 +0000 (07:03 -0700)
committerJens Axboe <axboe@kernel.dk>
Wed, 15 Dec 2021 03:08:05 +0000 (20:08 -0700)
commitcb2ac2912a9ca7d3d26291c511939a41361d2d83
tree5e7397cf98c33d5ea4287116aeded901fc7d19db
parentedaa26334c117a584add6053f48d63a988d25a6e
block: reduce kblockd_mod_delayed_work_on() CPU consumption

Dexuan reports that he's seeing spikes of very heavy CPU utilization when
running 24 disks and using the 'none' scheduler. This happens off the
sched restart path, because SCSI requires the queue to be restarted async,
and hence we're hammering on mod_delayed_work_on() to ensure that the work
item gets run appropriately.

Avoid hammering on the timer and just use queue_work_on() if no delay
has been specified.

Reported-and-tested-by: Dexuan Cui <decui@microsoft.com>
Link: https://lore.kernel.org/linux-block/BYAPR21MB1270C598ED214C0490F47400BF719@BYAPR21MB1270.namprd21.prod.outlook.com/
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
block/blk-core.c