| Message ID | 20170527142126.26079-6-ming.lei@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Sat, 2017-05-27 at 22:21 +0800, Ming Lei wrote: > /** > - * blk_mq_quiesce_queue() - wait until all ongoing queue_rq calls have finished > + * blk_mq_quiesce_queue() - wait until all ongoing dispatching have finished > * @q: request queue. > * Hello Ming, The concept of dispatching does not have a meaning to block driver authors who are not familiar with the block layer internals. However, every author of a blk-mq driver knows what the .queue_rq() function is. Additionally, the new comment is grammatically incorrect. So the above change looks like a step in the wrong direction to me. Bart.
On Tue, May 30, 2017 at 05:14:44PM +0000, Bart Van Assche wrote: > On Sat, 2017-05-27 at 22:21 +0800, Ming Lei wrote: > > /** > > - * blk_mq_quiesce_queue() - wait until all ongoing queue_rq calls have finished > > + * blk_mq_quiesce_queue() - wait until all ongoing dispatching have finished > > * @q: request queue. > > * > > Hello Ming, > > The concept of dispatching does not have a meaning to block driver authors who are > not familiar with the block layer internals. However, every author of a blk-mq driver > knows what the .queue_rq() function is. Unfortunately it isn't enough to just block .queue_rq(), did you read the commit log? > Additionally, the new comment is grammatically > incorrect. > So the above change looks like a step in the wrong direction to me. Sorry, I simply don't agree, and we have to make it explicit.
diff --git a/block/blk-mq.c b/block/blk-mq.c index 470ee5514ea9..032045841856 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -155,12 +155,12 @@ void blk_mq_unfreeze_queue(struct request_queue *q) EXPORT_SYMBOL_GPL(blk_mq_unfreeze_queue); /** - * blk_mq_quiesce_queue() - wait until all ongoing queue_rq calls have finished + * blk_mq_quiesce_queue() - wait until all ongoing dispatching have finished * @q: request queue. * * Note: this function does not prevent that the struct request end_io() - * callback function is invoked. Additionally, it is not prevented that - * new queue_rq() calls occur unless the queue has been stopped first. + * callback function is invoked. Once this function is returned, we make + * sure no dispatching can happen. */ void blk_mq_quiesce_queue(struct request_queue *q) {
Actually what we want to get from blk_mq_quiesce_queue() isn't only to wait for completion of all ongoing .queue_rq(). In the typical context of canceling requests, we need to make sure that the following is done in the dispatch path before starting to cancel requests: - failed dispatched request is finished - busy dispatched request is requeued, and the STARTED flag is cleared So update comment to keep code, doc and our expection consistent. Signed-off-by: Ming Lei <ming.lei@redhat.com> --- block/blk-mq.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)