Re: [PATCH 1/4] sched: move IO scheduling accounting from io_schedule_timeout() to __schedule()

From	Tejun Heo <tj@kernel.org>
Newsgroups	linux.kernel
Subject	Re: [PATCH 1/4] sched: move IO scheduling accounting from io_schedule_timeout() to __schedule()
Date	2016-10-31 17:50 +0100
Message-ID	<synWx-4Wn-1@gated-at.bofh.it> (permalink)
References	(1 earlier) <sxiPg-2Iq-25@gated-at.bofh.it> <sxk4L-3ru-25@gated-at.bofh.it> <sxkHn-3Tx-3@gated-at.bofh.it> <sxsvg-tH-7@gated-at.bofh.it> <sxsvg-tH-5@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

Hello,

On Sat, Oct 29, 2016 at 05:21:26AM +0200, Peter Zijlstra wrote:
> On Fri, Oct 28, 2016 at 03:12:32PM -0400, Tejun Heo wrote:
> > Hello, Peter.
> > 
> > On Fri, Oct 28, 2016 at 09:07:02PM +0200, Peter Zijlstra wrote:
> > > One alternative is to inherit the iowait state of the task we block on.
> > > That'll not get rid of the branches much, but it will remove the new
> > > mutex APIs.
> > 
> > Yeah, thought about that briefly but we don't necessarily track mutex
> 
> This one I actually fixed and should be in -next. And it would be
> sufficient to cover the use case here.

Tracking the owners of mutexes and rwsems does help quite a bit.  I
don't think it's as simple as inheriting io sleep state from the
current owner tho.  The owner might be running or in a non-IO sleep
when others try to grab the mutex.  It is an option to ignore those
cases but this would have a real possibility to lead to surprising
results in some corner cases.  If we choose to propagate dynamically,
it becomes an a lot more complex problem and I don't think it'd be
justfiable.

Unless there can be a simple enough and reliable solution, I think
it'd be better to stick with explicit marking.

Thanks.

-- 
tejun

Thread

[PATCHSET RFC] sched, jbd2: mark sleeps on journal->j_checkpoint_mutex as iowait Tejun Heo <tj@kernel.org> - 2016-10-28 19:00 +0200
  [PATCH 2/4] sched: separate out io_schedule_prepare() and io_schedule_finish() Tejun Heo <tj@kernel.org> - 2016-10-28 19:00 +0200
  [PATCH 1/4] sched: move IO scheduling accounting from io_schedule_timeout() to __schedule() Tejun Heo <tj@kernel.org> - 2016-10-28 19:10 +0200
    Re: [PATCH 1/4] sched: move IO scheduling accounting from  io_schedule_timeout() to __schedule() Peter Zijlstra <peterz@infradead.org> - 2016-10-28 20:30 +0200
      Re: [PATCH 1/4] sched: move IO scheduling accounting from  io_schedule_timeout() to __schedule() Peter Zijlstra <peterz@infradead.org> - 2016-10-28 21:10 +0200
        Re: [PATCH 1/4] sched: move IO scheduling accounting from  io_schedule_timeout() to __schedule() Peter Zijlstra <peterz@infradead.org> - 2016-10-29 05:30 +0200
          Re: [PATCH 1/4] sched: move IO scheduling accounting from  io_schedule_timeout() to __schedule() Tejun Heo <tj@kernel.org> - 2016-10-31 17:50 +0100
    Re: [PATCH 1/4] sched: move IO scheduling accounting from  io_schedule_timeout() to __schedule() Pavan Kondeti <pkondeti@codeaurora.org> - 2016-11-03 16:40 +0100
      Re: [PATCH 1/4] sched: move IO scheduling accounting from  io_schedule_timeout() to __schedule() Tejun Heo <tj@kernel.org> - 2016-11-09 00:00 +0100

csiph-web