[Intel-gfx] Supporting fused display configurations

[Chris Wilson]

On Thu, Dec 05, 2013 at 01:56:15PM +0000, Damien Lespiau wrote:
> A while ago, Ben added support for Quanta devices that had the display hardware
> fused off. We can a bit more general than that by reading the FUSE_STRAP
> register and setting num_pipes to 0 when detecting such condition.
> 
> To achieve that, the intel_device_info structure in dev_priv needed to be made
> writable, that's patch 1.
> 
> Patches 2 to 5 build the support for reading out FUSE_STRAP on IVB+, client
> platforms only to detect both "display fused off" and "pipe C fused off"
> configurations.
> 
> Patch 6 removes the Quanta special-case with the (reasonable) hope that it's
> now covered by the DISPLAY_DISABLE bit.
> 
> Patch 7 is a bonus, somewhat unrelated, little clean-up patch.
> 
> This series is not tested due to lack of such a device, I was hoping Ben still
> had the device around somewhere and could give this series a go.

Patches look fine. Just relying on fuse registers being correct is like
relying on vbt being correct, fraught with anxiety.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre