Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Tue, 25 Oct 2022 19:21:17 +0800
From:   Ming Lei <ming.lei@redhat.com>
To:     John Garry <john.garry@huawei.com>
Cc:     axboe@kernel.dk, linux-kernel@vger.kernel.org,
        linux-block@vger.kernel.org, hch@lst.de,
        Bart Van Assche <bvanassche@acm.org>
Subject: Re: [PATCH] blk-mq: Properly init bios from
 blk_mq_alloc_request_hctx()
Message-ID: <Y1fGrfHqbha6l+hz@T590>
References: <Y1U9zNZtZjRHQBww@T590>
 <99c6ca81-746d-85f4-04d3-49d7a3de611b@huawei.com>
 <Y1aS3vIbuQTNGWJL@T590>
 <360c78dc-65ce-362f-389d-075f2259ce5b@huawei.com>
 <Y1cvJ4/uwUScAQq4@T590>
 <3513b14c-14e0-b865-628e-a83521090de9@huawei.com>
 <CAFj5m9JnSBBVGrp5CqeH99-+VOGRuroUAi3c-3=6XKa891Sfmw@mail.gmail.com>
 <cf7f8f88-7d3e-8818-8584-e2276e7a1f30@huawei.com>
 <Y1epeuwonmjQhrXW@T590>
 <399a2c2d-0b56-e4e7-c309-a6b9537d8939@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <399a2c2d-0b56-e4e7-c309-a6b9537d8939@huawei.com>
Precedence: bulk

On Tue, Oct 25, 2022 at 10:32:28AM +0100, John Garry wrote:
> On 25/10/2022 10:16, Ming Lei wrote:
> > > > > I mentioned before that if no hctx->cpumask is online then we don't need
> > > > > to allocate a request. That is because if no hctx->cpumask is online,
> > > > > this means that original erroneous IO must be completed due to nature of
> > > > > how blk-mq cpu hotplug handler works, i.e. drained, and then we don't
> > > > > actually need to abort it any longer, so ok to not get a request.
> > > > No, it is really not OK, if all cpus in hctx->cpumask are offline, you
> > > > can't allocate
> > > > request on the specified hw queue, then the erroneous IO can't be handled,
> > > > then cpu hotplug handler may hang for ever.
> > > If the erroneous IO is still in-flight from blk-mq perspective, then how can
> > > hctx->cpumask still be offline? I thought that we guarantee that
> > > hctx->cpumask cannot go offline until drained.
> > Yeah, the draining is done before the cpu is offline. But the drain is
> > simply waiting for the inflight IO to be completed. If the IO is failed
> > during the waiting, you can't allocate such reserved request for error
> > handling, then hang ever in blk_mq_hctx_notify_offline().
> 
> Actually if final cpu in hctx->cpumask is going offline, then hctx won't
> queue any more requests, right? In this case I don't think we can queue on
> that hctx anyway. I need to think about this more.

It can be queued actually, but interrupt may not be delivered if managed
irq is used.

> 
> > 
> > If you just make it one driver private command, there can't be such
> > issue.
> 
> Well we're trying to use reserved requests for EH commands, which that goes
> against.
> 
> > Block layer is supposed for handling common case(normal io and pt io),
> > I'd suggest to not put such special cases into block layer.
> 
> It also supports reserved commands, which I would assume would be suitable
> for EH scenarios.

Then you have to be careful, as I mentioned, EH has to provide forward
progress, if you let blk-mq allocate & submit EH request, the implied
dependency from blk-mq has to be payed attention.


Thanks,
Ming