| Message ID | 20170809182806.32741-1-bart.vanassche@wdc.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 08/09/2017 12:28 PM, Bart Van Assche wrote: > The blk_mq_delay_kick_requeue_list() function is used by the device > mapper and only by the device mapper to rerun the queue and requeue > list after a delay. This function is called once per request that > gets requeued. Modify this function such that the queue is run once > per path change event instead of once per request that is requeued. Looks good to me, will add for this series. Thanks Bart.
On Wed, Aug 09, 2017 at 11:28:06AM -0700, Bart Van Assche wrote: > The blk_mq_delay_kick_requeue_list() function is used by the device > mapper and only by the device mapper to rerun the queue and requeue > list after a delay. This function is called once per request that > gets requeued. Modify this function such that the queue is run once > per path change event instead of once per request that is requeued. Is there a chance we could look into killing this function and moving dm to run it every time we requeue a request like all the other drivers?
On Thu, 2017-08-10 at 09:14 +0200, Christoph Hellwig wrote: > On Wed, Aug 09, 2017 at 11:28:06AM -0700, Bart Van Assche wrote: > > The blk_mq_delay_kick_requeue_list() function is used by the device > > mapper and only by the device mapper to rerun the queue and requeue > > list after a delay. This function is called once per request that > > gets requeued. Modify this function such that the queue is run once > > per path change event instead of once per request that is requeued. > > Is there a chance we could look into killing this function and moving > dm to run it every time we requeue a request like all the other drivers? Hello Christoph, That's an interesting question. I think it's not possible to remove this function without adding more infrastructure to the block layer. As an example, if the .lld_busy_fn() request queue callback function reports that a request queue is busy the only way for the dm-mpath driver to detect that an underlying request queue is no longer busy is by calling that function again after a certain time (see also multipath_busy() in dm-mpath.c). In other words, if we want to get rid of blk_mq_delay_kick_requeue_list() then notification mechanisms will have to be added for every type of state information that can cause the state of a multipath from "not ready" into "ready". Bart.
diff --git a/block/blk-mq.c b/block/blk-mq.c index 041f7b7fa0d6..8bfea36e92f9 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -679,8 +679,8 @@ EXPORT_SYMBOL(blk_mq_kick_requeue_list); void blk_mq_delay_kick_requeue_list(struct request_queue *q, unsigned long msecs) { - kblockd_schedule_delayed_work(&q->requeue_work, - msecs_to_jiffies(msecs)); + kblockd_mod_delayed_work_on(WORK_CPU_UNBOUND, &q->requeue_work, + msecs_to_jiffies(msecs)); } EXPORT_SYMBOL(blk_mq_delay_kick_requeue_list);
The blk_mq_delay_kick_requeue_list() function is used by the device mapper and only by the device mapper to rerun the queue and requeue list after a delay. This function is called once per request that gets requeued. Modify this function such that the queue is run once per path change event instead of once per request that is requeued. Fixes: commit 2849450ad39d ("blk-mq: introduce blk_mq_delay_kick_requeue_list()") Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com> Cc: Mike Snitzer <snitzer@redhat.com> Cc: Laurence Oberman <loberman@redhat.com> Cc: <stable@vger.kernel.org> --- block/blk-mq.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)