Hi, Michal
在 2023/08/11 17:17, Michal Koutný 写道:
On Fri, Aug 11, 2023 at 04:53:44PM +0800, Yu Kuai <yukuai1@xxxxxxxxxxxxxxx> wrote:
Yes, it'm implemented in the upper layer that rq_qos_add() and
blkcg_activate_policy() should be atmoic, and currently there is no
comments for that.
The check (iolat_rq_qos()) and use (activating the policy) should be the
atomic pair.
Perhaps it's better to add some comments like following in rq_qos_add()
instead?
Honestly, I find the current variant (v3) good as it is -- closest to
the pair of the operations.
(But it's merely a comment so ¯\_(ツ)_/¯)
Yes, it's just a comment.
Lingfeng, can you resend this patch with the following fixed?
> lockdep_assert_held(ctx.bdev->bd_queue->rq_qos_mutex);
should be:
lockdep_assert_held(&ctx.bdev->bd_queue->rq_qos_mutex);
And you can keep my review tag.
Thanks,
Kuai
Michal
