[Intel-gfx] Regression in drm-intel caused by d91a2cb8e510

[Alan Stern]

On Mon, 1 Sep 2014, Damien Lespiau wrote:

> On Mon, Sep 01, 2014 at 04:19:23PM -0400, Alan Stern wrote:
> > Daniel:
> > 
> > I'm encountering a problem with the drm-intel-nightly branch in your
> > drm-intel repository.  The symptom is that when I boot my laptop, at
> > some point during the start-up procedure the screen goes totally blank,
> > as though the backlight were turned off, and it remains that way.  I
> > can't tell if the backlight really is off -- it may be that there's
> > just nothing being displayed -- and playing with the screen-brightness
> > function keys doesn't help.
> > 
> > Git bisect located the source of the problem as commit d91a2cb8e510
> > (drm/i915: Add 180 degree primary plane rotation support).  The problem
> > occurs when I boot a kernel built from that commit, and it doesn't
> > occur when I boot a kernel built from the preceding commit
> > (19ac737dc3c3).
> > 
> > If necessary, I can get debugging info by logging in over a network 
> > connection.  Tell me what you need.
> 
> It seems like Ville and Mika already caught a similar bug today, could
> you give Ville's patch a go (available on the intel-gfx):
> 
>   drm/i915: Don't call intel_plane_restore() when the prop value didn't change

Yes, that patch fixed the problem.  Thank you.

Tested-by: Alan Stern <stern at rowland.harvard.edu>

Alan Stern