Fwd: EDID/DP fixes for proper bpc detection of displays.

[Daniel Vetter]

On Sat, Jul 30, 2016 at 05:08:03PM +0200, Mario Kleiner wrote:
> Three weeks have passed, so resending. Could i maybe excite somebody to
> review/merge at least some of the really trivial but important patches to
> fix regressions back-ported to all stable kernels, mostly in Intel kms + DP
> legacy in this series before the 4.8 merge window closes?
> 
> Especially patch 2, "[PATCH 2/5] drm/i915/dp: Revert "drm/i915/dp: fall back
> to 18 bpp when sink capability is unknown" a revert acked by Jani for a
> patch of him, and patch 5 "[PATCH 5/5] drm/edid: Set 8 bpc color depth for
> displays with "DFP 1.x compliant TMDS"?
> 
> Patch 5/5 would make our EDID parsing for DVI more complete and if also
> back-ported to stable would at least prevent the worst fallout for current
> users if patch 2/5 doesn't get applied.

I don't have the spec handy to review the details, but it all seems to
make sense. Ack: me.

Dave, probably simplest if you just pick up the entire pile into drm-fixes
directly? Or still offline because cold, and I should carry it in
drm-fixes?
-Daniel

> 
> thanks,
> -mario
> 
>  Forwarded Message 
> Subject: EDID/DP fixes for proper bpc detection of displays.
> Date: Wed,  6 Jul 2016 12:05:43 +0200
> From: Mario Kleiner 
> To: dri-devel at lists.freedesktop.org
> CC: mario.kleiner.de at gmail.com
> 
> Updated series to fix the regressions introduced into stable
> kernels for active Displayport to dual-link DVI converters and
> for active Displayport to VGA converters. Also to fix this stuff
> better for Linux 4.8.
> 
> Patch 1 + 2 are needed to fix the regressions in the stable
> kernels. These are meant for easy backporting to all affected
> kernels.
> 
> Patch 3: A DP helper for parsing sink bpc from DP aux data to
> assist drivers if they can't get bpc from EDID.
> 
> Patch 4: Make intel kms use patch 3.
> 
> Patch 5: Make the EDID parser recognize the "DFP1.x compatible TMDS"
> bit by which DVI sinks can declare they support at least 8 bpc according
> to DFP spec. This one if applied would let me sleep so much better, as
> the common special display equipment for neuroscience research, which is
> all DVI based atm., sets that bit in its EDID 1.3.
> 
> Thanks,
> -mario
> 

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

Fwd: EDID/DP fixes for proper bpc detection of displays.

[Mario Kleiner]

Three weeks have passed, so resending. Could i maybe excite somebody to 
review/merge at least some of the really trivial but important patches 
to fix regressions back-ported to all stable kernels, mostly in Intel 
kms + DP legacy in this series before the 4.8 merge window closes?

Especially patch 2, "[PATCH 2/5] drm/i915/dp: Revert "drm/i915/dp: fall 
back to 18 bpp when sink capability is unknown" a revert acked by Jani 
for a patch of him, and patch 5 "[PATCH 5/5] drm/edid: Set 8 bpc color 
depth for displays with "DFP 1.x compliant TMDS"?

Patch 5/5 would make our EDID parsing for DVI more complete and if also 
back-ported to stable would at least prevent the worst fallout for 
current users if patch 2/5 doesn't get applied.

thanks,
-mario

 Forwarded Message 
Subject: EDID/DP fixes for proper bpc detection of displays.
Date: Wed,  6 Jul 2016 12:05:43 +0200
From: Mario Kleiner 
To: dri-devel at lists.freedesktop.org
CC: mario.kleiner.de at gmail.com

Updated series to fix the regressions introduced into stable
kernels for active Displayport to dual-link DVI converters and
for active Displayport to VGA converters. Also to fix this stuff
better for Linux 4.8.

Patch 1 + 2 are needed to fix the regressions in the stable
kernels. These are meant for easy backporting to all affected
kernels.

Patch 3: A DP helper for parsing sink bpc from DP aux data to
assist drivers if they can't get bpc from EDID.

Patch 4: Make intel kms use patch 3.

Patch 5: Make the EDID parser recognize the "DFP1.x compatible TMDS"
bit by which DVI sinks can declare they support at least 8 bpc according
to DFP spec. This one if applied would let me sleep so much better, as
the common special display equipment for neuroscience research, which is
all DVI based atm., sets that bit in its EDID 1.3.

Thanks,
-mario