Re: [PATCH 5/5] blk-mq: Rework blk-mq timeout handling again

From: Bart Van Assche <Bart.VanAssche@wdc.com>
To: "keith.busch@intel.com" <keith.busch@intel.com>
Cc: "hch@lst.de" <hch@lst.de>,
	"ming.lei@redhat.com" <ming.lei@redhat.com>,
	"linux-block@vger.kernel.org" <linux-block@vger.kernel.org>,
	"martin.petersen@oracle.com" <martin.petersen@oracle.com>,
	"axboe@kernel.dk" <axboe@kernel.dk>,
	"jianchao.w.wang@oracle.com" <jianchao.w.wang@oracle.com>
Subject: Re: [PATCH 5/5] blk-mq: Rework blk-mq timeout handling again
Date: Fri, 27 Jul 2018 16:51:05 +0000	[thread overview]
Message-ID: <08bd5d1c691e2c8f9950f7ebcf04716507a02675.camel@wdc.com> (raw)
In-Reply-To: <20180727164627.GB20829@localhost.localdomain>

On Fri, 2018-07-27 at 10:46 -0600, Keith Busch wrote:
> On Fri, Jul 27, 2018 at 09:20:42AM -0700, Bart Van Assche wrote:
> > +	ret = req->q->mq_ops->timeout(req, reser=
ved);
> > +	/*
> > +	 * BLK_EH_DONT_RESET_TIMER means that th=
e block driver either
> > +	 * completed the request or still owns the request and w=
ill
> > +	 * continue processing the timeout asynchronously. In th=
e
> > +	 * latter case, if blk_mq_complete_request()=
 was called while
> > +	 * the timeout handler was in progress, ignore that call=
.
> > +	 */
> > +	if (ret == BLK_EH_DONT_RESET_TIMER)
> > +		return;
>=20
> This is how completions get lost.

The new approach for handling completions that occur while the .timeout()
callback in progress is as follows:
* blk_mq_complete_request() executes the following code:
       if (blk_mq_change_rq_state(rq, MQ_RQ_TIMED=
F8-OUT, MQ_RQ_COMPLETE))
               return;
* blk_mq_rq_timed_out() executes the following code:
       if (blk_mq_rq_state(req) == MQ_RQ_COMPLETE=
) {
               __blk_mq_complete_request(req);
               return;
       }

As one can see __blk_mq_complete_request() gets called if=
 this race occurs.

Bart.