All of lore.kernel.org
 help / color / mirror / Atom feed
From: Daniel Vetter <daniel@ffwll.ch>
To: Chris Wilson <chris@chris-wilson.co.uk>,
	Daniel Vetter <daniel@ffwll.ch>, Tomas Elf <tomas.elf@intel.com>,
	Intel-GFX@Lists.FreeDesktop.Org
Subject: Re: [PATCH 7/8] drm/i915: Grab execlist spinlock to avoid post-reset concurrency issues.
Date: Tue, 13 Oct 2015 13:46:38 +0200	[thread overview]
Message-ID: <20151013114638.GQ26718@phenom.ffwll.local> (raw)
In-Reply-To: <20151009084516.GG27939@nuc-i3427.alporthouse.com>

On Fri, Oct 09, 2015 at 09:45:16AM +0100, Chris Wilson wrote:
> On Fri, Oct 09, 2015 at 10:38:18AM +0200, Daniel Vetter wrote:
> > On Thu, Oct 08, 2015 at 07:31:39PM +0100, Tomas Elf wrote:
> > > Grab execlist lock when cleaning up execlist queues after GPU reset to avoid
> > > concurrency problems between the context event interrupt handler and the reset
> > > path immediately following a GPU reset.
> > > 
> > > Signed-off-by: Tomas Elf <tomas.elf@intel.com>
> > 
> > Should we instead just stop any irqs from the GT completely while we do
> > the reset (plus a synchronize_irq)? It's a bit heavy-weight, but probably
> > safer. Well not the entire GT, but all the rings under reset (as prep for
> > per-ring reset).
> 
> Bah, stopping IRQs is not enough for error state capture though since
> requests complete asynchronously just by polling a memory address. (If
> that is the goal here, this patch just makes execlist_queue access
> consistent and should only be run once the GPU has been reset and so is
> categorically idle.)

This is the execlist ELSP tracking, which is execlusively driven by the
CTX_SWITCH interrupt signal from each engine.

At least that's been my assumption, and under that assumption I think
stalling interrupts should be good enough.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

  reply	other threads:[~2015-10-13 11:43 UTC|newest]

Thread overview: 68+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-10-08 18:31 [PATCH 0/8] Stability improvements to error state capture Tomas Elf
2015-10-08 18:31 ` [PATCH 1/8] drm/i915: Early exit from semaphore_waits_for for execlist mode Tomas Elf
2015-10-08 18:31 ` [PATCH 2/8] drm/i915: Migrate to safe iterators in error state capture Tomas Elf
2015-10-09  7:49   ` Chris Wilson
2015-10-09 11:38     ` Tomas Elf
2015-10-09  8:27   ` Daniel Vetter
2015-10-09 11:40     ` Tomas Elf
2015-10-13 11:37       ` Daniel Vetter
2015-10-13 11:47         ` Chris Wilson
2015-10-08 18:31 ` [PATCH 3/8] drm/i915: Cope with request list state change during " Tomas Elf
2015-10-09  7:48   ` Chris Wilson
2015-10-09 11:25     ` Tomas Elf
2015-10-13 11:39       ` Daniel Vetter
2015-10-14 11:46         ` Tomas Elf
2015-10-14 12:45           ` Daniel Vetter
2015-10-09  8:28   ` Daniel Vetter
2015-10-09 11:45     ` Tomas Elf
2015-10-13 11:40       ` Daniel Vetter
2015-10-08 18:31 ` [PATCH 4/8] drm/i915: NULL checking when capturing buffer objects " Tomas Elf
2015-10-09  7:49   ` Chris Wilson
2015-10-09 11:34     ` Tomas Elf
2015-10-09  8:32   ` Daniel Vetter
2015-10-09  8:47     ` Chris Wilson
2015-10-09 11:52       ` Tomas Elf
2015-10-09 11:45     ` Tomas Elf
2015-10-08 18:31 ` [PATCH 5/8] drm/i915: vma NULL pointer check Tomas Elf
2015-10-09  7:48   ` Chris Wilson
2015-10-09 11:30     ` Tomas Elf
2015-10-09 11:59       ` Chris Wilson
2015-10-13 11:43         ` Daniel Vetter
2015-10-09  8:33   ` Daniel Vetter
2015-10-09 11:46     ` Tomas Elf
2015-10-08 18:31 ` [PATCH 6/8] drm/i915: Use safe list iterators Tomas Elf
2015-10-09  7:41   ` Chris Wilson
2015-10-09 10:27     ` Tomas Elf
2015-10-09 10:38       ` Chris Wilson
2015-10-09 12:00         ` Tomas Elf
2015-10-08 18:31 ` [PATCH 7/8] drm/i915: Grab execlist spinlock to avoid post-reset concurrency issues Tomas Elf
2015-10-09  7:45   ` Chris Wilson
2015-10-09 10:28     ` Tomas Elf
2015-10-09  8:38   ` Daniel Vetter
2015-10-09  8:45     ` Chris Wilson
2015-10-13 11:46       ` Daniel Vetter [this message]
2015-10-13 11:45         ` Chris Wilson
2015-10-13 13:46           ` Daniel Vetter
2015-10-13 14:00             ` Chris Wilson
2015-10-19 15:32   ` [PATCH v2 " Tomas Elf
2015-10-22 16:49     ` Dave Gordon
2015-10-22 17:35       ` Daniel Vetter
2015-10-23  8:42     ` Tvrtko Ursulin
2015-10-23  8:59       ` Daniel Vetter
2015-10-23 11:02         ` Tomas Elf
2015-10-23 12:49           ` Dave Gordon
2015-10-23 13:08     ` [PATCH v3 " Tomas Elf
2015-10-23 14:53       ` Daniel, Thomas
2015-10-23 17:02     ` [PATCH] drm/i915: Update to post-reset execlist queue clean-up Tomas Elf
2015-12-01 11:46       ` Tvrtko Ursulin
2015-12-11 14:14         ` Dave Gordon
2015-12-11 16:40           ` Daniel Vetter
2015-12-14 10:21           ` Mika Kuoppala
2015-10-08 18:31 ` [PATCH 8/8] drm/i915: NULL check of unpin_work Tomas Elf
2015-10-09  7:46   ` Chris Wilson
2015-10-09  8:39     ` Daniel Vetter
2015-10-09 11:50       ` Tomas Elf
2015-10-09 10:30     ` Tomas Elf
2015-10-09 10:44       ` Chris Wilson
2015-10-09 12:06         ` Tomas Elf
2015-10-13 11:51           ` Daniel Vetter

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20151013114638.GQ26718@phenom.ffwll.local \
    --to=daniel@ffwll.ch \
    --cc=Intel-GFX@Lists.FreeDesktop.Org \
    --cc=chris@chris-wilson.co.uk \
    --cc=tomas.elf@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.