[Intel-gfx] [PATCH] drm/i915/lvds: Always use the presence pin for LVDS on PCH

[Daniel Vetter]

On Thu, Feb 09, 2012 at 10:23:37AM -0800, Keith Packard wrote:
> <#part sign=pgpmime>
> On Thu, 09 Feb 2012 17:33:23 +0000, Chris Wilson <chris at chris-wilson.co.uk> wrote:
> > 
> > That code already existed in intel_lvds_init():
> >   if (HAS_PCH_SPLIT(dev)) {
> >      if (I915_READ(PCH_LVDS) & LVDS_DETECTED) == 0)
> >         return false
> >      ...
> >   }
> 
> Tricky. Do we have any idea how reliable this pin is? We've had such bad
> luck in the past...

It's imo guaranteed to blow up somewhere, but we already have the no_lvds
quirk infrastructure in place for that. Bug reporter confirmed that this
patch works for him, so I've decided to include this into the current
-next cycle to get feedback (and no_lvds quirk patches) as early as
possible.

Thanks for the patch and review.
-Daniel
-- 
Daniel Vetter
Mail: daniel at ffwll.ch
Mobile: +41 (0)79 365 57 48