[PATCH 0/2] drm/i915: 830M regression fixes

[PATCH 0/2] drm/i915: 830M regression fixes

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Looks like I managed to break 830M in a few different ways recently. But I
recently found one for myself so hopefully that'll not happen again.

I have a few other things lined up for 830M, but these are the quick fixes
that should get us back to the state we had before.

Ville Syrjälä (2):
  drm/i915: Don't grab crtc mutexes in intel_modeset_gem_init()
  drm/i915: Don't swap planes on 830M

 drivers/gpu/drm/i915/i915_irq.c      | 4 ++--
 drivers/gpu/drm/i915/intel_display.c | 6 +++---
 2 files changed, 5 insertions(+), 5 deletions(-)

-- 
1.8.3.2

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 0/2] drm/i915: 830M regression fixes

[Thomas Richter]

On 01/07/2014 03:15 PM, ville.syrjala@linux.intel.com wrote:
> From: Ville Syrjälä<ville.syrjala@linux.intel.com>
>
> Looks like I managed to break 830M in a few different ways recently. But I
> recently found one for myself so hopefully that'll not happen again.
>
> I have a few other things lined up for 830M, but these are the quick fixes
> that should get us back to the state we had before.

Thanks, Ville!

As a related note, I found that with the latest kernel, "uxa" drawing of 
filled rectangles became very slow (and sna is still too broken to be 
useful, sorry). Would you know which file I should look into to check 
for something like this? The client application only submits plain X11 
calls, the render acceleration or composition should (?) not play a role 
here. It's really a plain simple filled rectangle (or many of them) that 
became notably slow.

BTW, the latest kernel version fixed one issue, namely that on 830 the 
boot console could become blank or vertically misaligned. It is now 
always fine.

So long,
     Thomas

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 0/2] drm/i915: 830M regression fixes

[Chris Wilson]

On Tue, Jan 07, 2014 at 04:12:06PM +0100, Thomas Richter wrote:
> On 01/07/2014 03:15 PM, ville.syrjala@linux.intel.com wrote:
> >From: Ville Syrjälä<ville.syrjala@linux.intel.com>
> >
> >Looks like I managed to break 830M in a few different ways recently. But I
> >recently found one for myself so hopefully that'll not happen again.
> >
> >I have a few other things lined up for 830M, but these are the quick fixes
> >that should get us back to the state we had before.
> 
> Thanks, Ville!
> 
> As a related note, I found that with the latest kernel, "uxa"
> drawing of filled rectangles became very slow (and sna is still too
> broken to be useful, sorry). Would you know which file I should look
> into to check for something like this? The client application only
> submits plain X11 calls, the render acceleration or composition
> should (?) not play a role here. It's really a plain simple filled
> rectangle (or many of them) that became notably slow.

Wtf? What's up with SNA? There's a solitary report that cache coherency
on 830GM requires some workaround, on which I'd hoped that Daniel could
run igt to confirm (and now perhaps Ville might fancy the challenge).
Other than that, judging by my 845g, it should be solid.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre

Re: [PATCH 0/2] drm/i915: 830M regression fixes

[Ville Syrjälä]

On Tue, Jan 07, 2014 at 04:59:10PM +0000, Chris Wilson wrote:
> On Tue, Jan 07, 2014 at 04:12:06PM +0100, Thomas Richter wrote:
> > On 01/07/2014 03:15 PM, ville.syrjala@linux.intel.com wrote:
> > >From: Ville Syrjälä<ville.syrjala@linux.intel.com>
> > >
> > >Looks like I managed to break 830M in a few different ways recently. But I
> > >recently found one for myself so hopefully that'll not happen again.
> > >
> > >I have a few other things lined up for 830M, but these are the quick fixes
> > >that should get us back to the state we had before.
> > 
> > Thanks, Ville!
> > 
> > As a related note, I found that with the latest kernel, "uxa"
> > drawing of filled rectangles became very slow (and sna is still too
> > broken to be useful, sorry). Would you know which file I should look
> > into to check for something like this? The client application only
> > submits plain X11 calls, the render acceleration or composition
> > should (?) not play a role here. It's really a plain simple filled
> > rectangle (or many of them) that became notably slow.
> 
> Wtf? What's up with SNA? There's a solitary report that cache coherency
> on 830GM requires some workaround, on which I'd hoped that Daniel could
> run igt to confirm (and now perhaps Ville might fancy the challenge).
> Other than that, judging by my 845g, it should be solid.

FWIW I haven't seen any issues w/ SNA on either 830 or 855, but I must
admit that I've not done any extensive testing on either.

-- 
Ville Syrjälä
Intel OTC