BISECTED: Suspend regression on drm-intel-next

BISECTED: Suspend regression on drm-intel-next

[Peter Clifton]

Hi Chris,

I bisected the suspend regression I mentioned, and came to this commit:

commit 0cdab21f9a1fca50dd27e488839f5a6578e333b2
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Sun Dec 5 17:27:06 2010 +0000

    drm/i915: Uncouple render/power ctx before suspending
    
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>


Disabling power context tear down on suspend with this diff:

diff --git a/drivers/gpu/drm/i915/i915_suspend.c
b/drivers/gpu/drm/i915/i915_suspend.c
index a311809..6e98764 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -820,7 +820,7 @@ int i915_save_state(struct drm_device *dev)
        if (HAS_PCH_SPLIT(dev))
                ironlake_disable_drps(dev);
 
-       intel_disable_clock_gating(dev);
+//     intel_disable_clock_gating(dev);
 
        /* Cache mode state */
        dev_priv->saveCACHE_MODE_0 = I915_READ(CACHE_MODE_0);



NB: I've not tested later than c57802706aad9f179f2219415717896b3c177845,
which does not suspend. (I'm actually pretty certain that I tested until
around a8e93126a6f10d0a14ba8407ec112b1b3a5e2e97, but it was still bad,
and I didn't need to include it in my bisect, I did not make any notes
about it.

Best regards,

-- 
Peter Clifton

Electrical Engineering Division,
Engineering Department,
University of Cambridge,
9, JJ Thomson Avenue,
Cambridge
CB3 0FA

Tel: +44 (0)7729 980173 - (No signal in the lab!)
Tel: +44 (0)1223 748328 - (Shared lab phone, ask for me)

Re: BISECTED: Suspend regression on drm-intel-next

[Chris Wilson]

On Sat, 18 Dec 2010 20:02:21 +0000, Peter Clifton <pcjc2@cam.ac.uk> wrote:
> Hi Chris,
> 
> I bisected the suspend regression I mentioned, and came to this commit:
> 
> commit 0cdab21f9a1fca50dd27e488839f5a6578e333b2
> Author: Chris Wilson <chris@chris-wilson.co.uk>
> Date:   Sun Dec 5 17:27:06 2010 +0000
> 
>     drm/i915: Uncouple render/power ctx before suspending
>     
>     Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> 
> 
> Disabling power context tear down on suspend with this diff:

And you verified the revert against drm-intel-next HEAD? I'll revert just
as soon as I see your confirmation. :)

[The patch was only trying to see if it made Ironlake stable on suspending
with renderctx, to no avail so there should be no repercussions in
reverting.]

Thanks,
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre

Re: BISECTED: Suspend regression on drm-intel-next

[Peter Clifton]

On Sat, 2010-12-18 at 21:42 +0000, Chris Wilson wrote:
> On Sat, 18 Dec 2010 20:02:21 +0000, Peter Clifton <pcjc2@cam.ac.uk> wrote:
> > Hi Chris,
> > 
> > I bisected the suspend regression I mentioned, and came to this commit:
> > 
> > commit 0cdab21f9a1fca50dd27e488839f5a6578e333b2
> > Author: Chris Wilson <chris@chris-wilson.co.uk>
> > Date:   Sun Dec 5 17:27:06 2010 +0000
> > 
> >     drm/i915: Uncouple render/power ctx before suspending
> >     
> >     Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> > 
> > 
> > Disabling power context tear down on suspend with this diff:
> 
> And you verified the revert against drm-intel-next HEAD? I'll revert just
> as soon as I see your confirmation. :)

Confirmed.. drm-intel-next is still broken.

I'm vaguely curious why it makes it lock up, but I'll settle for it not
doing so ;)  Only since suspend has been broken did I realise just how
much I rely upon it!

> [The patch was only trying to see if it made Ironlake stable on suspending
> with renderctx, to no avail so there should be no repercussions in
> reverting.]

Good to hear.

Best wishes,

-- 
Peter Clifton

Electrical Engineering Division,
Engineering Department,
University of Cambridge,
9, JJ Thomson Avenue,
Cambridge
CB3 0FA

Tel: +44 (0)7729 980173 - (No signal in the lab!)
Tel: +44 (0)1223 748328 - (Shared lab phone, ask for me)