* [3.0-rc3] i915: potentially stale ring data during resume
@ 2011-06-18 6:48 Daniel J Blueman
2011-06-18 16:12 ` Ben Widawsky
0 siblings, 1 reply; 2+ messages in thread
From: Daniel J Blueman @ 2011-06-18 6:48 UTC (permalink / raw)
To: Eric Anholt, Chris Wilson; +Cc: intel-gfx
Hi Eric, Chris,
There are quite a lot of callers of I915_READ/WRITE which aren't
holding the dev->struct_mutex lock across device suspend and resume
operations, to prevent the GT going to sleep early if two callers are
in the gen6_gt_force_wake_get->gen6_gt_force_wake_put critical region.
I started cooking up a patch, but ran out of time.
Thanks,
Daniel
--
Daniel J Blueman
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [3.0-rc3] i915: potentially stale ring data during resume
2011-06-18 6:48 [3.0-rc3] i915: potentially stale ring data during resume Daniel J Blueman
@ 2011-06-18 16:12 ` Ben Widawsky
0 siblings, 0 replies; 2+ messages in thread
From: Ben Widawsky @ 2011-06-18 16:12 UTC (permalink / raw)
To: Daniel J Blueman; +Cc: intel-gfx
On Sat, Jun 18, 2011 at 02:48:29PM +0800, Daniel J Blueman wrote:
> Hi Eric, Chris,
>
> There are quite a lot of callers of I915_READ/WRITE which aren't
> holding the dev->struct_mutex lock across device suspend and resume
> operations, to prevent the GT going to sleep early if two callers are
> in the gen6_gt_force_wake_get->gen6_gt_force_wake_put critical region.
>
> I started cooking up a patch, but ran out of time.
Please note the only register range that matters is 0-0x40000
(don't ask how we know that).
Ben
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2011-06-18 16:12 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-06-18 6:48 [3.0-rc3] i915: potentially stale ring data during resume Daniel J Blueman
2011-06-18 16:12 ` Ben Widawsky
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox