[PATCH 00/23] drm: Eliminate plane->fb/crtc usage for atomic drivers

Emil Velikov emil.l.velikov at gmail.com
Thu Mar 22 18:34:19 UTC 2018


On 22 March 2018 at 18:03, Harry Wentland <harry.wentland at amd.com> wrote:
> On 2018-03-22 01:54 PM, Emil Velikov wrote:
>> Hi Ville,
>>
>> On 22 March 2018 at 15:22, Ville Syrjala <ville.syrjala at linux.intel.com> wrote:
>>> From: Ville Syrjälä <ville.syrjala at linux.intel.com>
>>>
>>> I really just wanted to fix i915 to re-enable its planes afer load
>>> detection (a two line patch). This is what I actually ended up with
>>> after I ran into a framebuffer refcount leak with said two line patch.
>>>
>>> I've tested this on a few i915 boxes and so far it's looking
>>> good. Everything else is just compile tested.
>>>
>> Mostly thinking out loud:
>>
>> Wondering if one cannot somehow (re)move plane->fb/crtc altogether.
>> Otherwise drivers will reintroduce similar code, despite the WARNs and
>> beefy documentation :-\
>
> Wouldn't that require an atomic conversion of all remaining drivers?
>
That or maybe move into plane->legacy->{fb,crtc}. Feel free to swap
'legacy' with flashier name.

Hmm back in 2015 we had a GSoC that updated BOCHS and CIRRUS drivers,
but they never got merged.
Don't recall the details - from memory the conversion seemed fine, but
there was either shortage on review/other.

Might be worth reviving that... regardless it's getting a bit off-topic.
-Emil

[1] https://www.google-melange.com/archive/gsoc/2015/orgs/xorg/projects/johnhunter.html


More information about the amd-gfx mailing list