On Sat, Jan 22, 2022 at 07:10:52PM +0800, Ming Lei wrote:
> 3 files changed, 41 insertions(+), 17 deletions(-)
>
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index d51b0aa2e4e4..72ae9955cf27 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -3101,6 +3101,9 @@ void blk_mq_free_rqs(struct blk_mq_tag_set *set, struct blk_mq_tags *tags,
> struct blk_mq_tags *drv_tags;
> struct page *page;
>
> + if (list_empty(&tags->page_list))
> + return;
Please split this out and document why it is needed.
> +/* Unconfigure the I/O scheduler and dissociate from the cgroup controller. */
> +static void blk_exit_queue(struct request_queue *q)
> +{
> + /*
> + * Since the I/O scheduler exit code may access cgroup information,
> + * perform I/O scheduler exit before disassociating from the block
> + * cgroup controller.
> + */
> + if (q->elevator) {
> + ioc_clear_queue(q);
> +
> + mutex_lock(&q->sysfs_lock);
> + blk_mq_sched_free_rqs(q);
> + elevator_exit(q);
> + mutex_unlock(&q->sysfs_lock);
> + }
> +}
Given that this all deals with the I/O scheduler the naming seems
wrong. It almost seems like we'd want to move ioc_clear_queue and
blk_mq_sched_free_rqs into elevator_exit to have somewhat sensible
helpers. That would add extra locking of q->sysfs_lock for
ioc_clear_queue, but given that elevator_switch_mq already does
that, this seems harmless.
