[Intel-gfx] [PATCH] drm/i915: Don't deref pipe->cpu_transcoder in the hangcheck code

[Daniel Vetter]

On Thu, Aug 08, 2013 at 02:44:23PM +0100, Chris Wilson wrote:
> On Thu, Aug 08, 2013 at 03:12:06PM +0200, Daniel Vetter wrote:
> > From: Chris Wilson <chris at chris-wilson.co.uk>
> > 
> > If we get an error event really early in the driver setup sequence,
> > which gen3 is especially prone to with various display GTT faults we
> > Oops. So try to avoid this.
> > 
> > Additionally with Haswell the transcoders are a separate bank of
> > registers from the pipes (4 transcoders, 3 pipes). In event of an
> > error, we want to be sure we have a complete and accurate picture of
> > the machine state, so record all the transcoders in addition to all
> > the active pipes.
> > 
> > This regression has been introduced in
> > 
> > commit 702e7a56af3780d8b3a717f698209bef44187bb0
> > Author: Paulo Zanoni <paulo.r.zanoni at intel.com>
> > Date:   Tue Oct 23 18:29:59 2012 -0200
> > 
> >     drm/i915: convert PIPECONF to use transcoder instead of pipe
> > 
> > Based on the patch "drm/i915: Dump all transcoder registers on error"
> > from Chris Wilson:
> > 
> > v2: Rebase so that we don't try to be clever and try to figure out the
> > cpu transcoder from hw state. That exercise should be done when we
> > analyze the error state offline.
> > 
> > The actual bugfix is to not call intel_pipe_to_cpu_transcoder in the
> > error state capture code in case the pipes aren't fully set up yet.
> > 
> > v3: Simplifiy the err->num_transcoders computation a bit. While at it
> > make the error capture stuff save on systems without a display block.
> > 
> > v4: Fix fail, spotted by Jani.
> > 
> > v5: Completely new commit message, cc: stable.
> > 
> > Cc: Paulo Zanoni <paulo.r.zanoni at intel.com>
> > Cc: Damien Lespiau <damien.lespiau at intel.com>
> > Cc: Jani Nikula <jani.nikula at intel.com>
> > Cc: Chris Wilson <chris at chris-wilson.co.uk>
> > Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=60021
> > Cc: stable at vger.kernel.org
> > Signed-off-by: Daniel Vetter <daniel.vetter at ffwll.ch>
> 
> Lgtm. We may have to modify transcoders[] to be more dynamic in future,
> but for now this works.
> Reviewed-by: Chris Wilson <chris at chris-wilson.co.uk>

Picked up for -fixes, thanks for the review.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch