[Mesa-dev] [PATCH 7/7] st/va: add headless support, i.e. VA_DISPLAY_DRM

[Emil Velikov]

On 29 October 2015 at 23:04, Arnaud Vrac <rawoul at gmail.com> wrote:
> On Thu, Oct 29, 2015 at 11:28 PM, Emil Velikov <emil.l.velikov at gmail.com>
> wrote:
>>
>> On 20 October 2015 at 17:40, Arnaud Vrac <rawoul at gmail.com> wrote:
>> > On Tue, Oct 20, 2015 at 6:35 PM, Emil Velikov <emil.l.velikov at gmail.com>
>> > wrote:
>> >>
>> >> On 20 October 2015 at 17:06, Julien Isorce <julien.isorce at gmail.com>
>> >> wrote:
>> >> >
>> >> >
>> >> > On 19 October 2015 at 17:16, Emil Velikov <emil.l.velikov at gmail.com>
>> >> > wrote:
>> >> >>
>> >> >> On 17 October 2015 at 00:14, Julien Isorce <julien.isorce at gmail.com>
>> >> >> wrote:
>> >> >> > This patch allows to use gallium vaapi without requiring
>> >> >> > a X server running for your second graphic card.
>> >> >> >
>> >> >> I've sent a lengthy series which should mitigate the need of some
>> >> >> hunks.
>> >> >
>> >> >
>> >> > Ok I'll wait for your patches to land before going further on this
>> >> > patch.
>> >> > Should I expect vl_winsy_drm.c in your patches ? Not sure do
>> >> > understood
>> >> > that
>> >> > part. Actually I though about having "vl_screen_create_drm" and
>> >> > renames
>> >> > vl_screen_create to vl_screen_create_x11 (because it takes XDisplay
>> >> > in
>> >> > params) but then I got confused because vl_winsys.h includes Xlib.h.
>> >> > Should
>> >> > future vl_screen_create_drm be in another header, vl_drm.h ?
>> >> >
>> >> My series flattens the if GALLIUM_STATIC_TARGETS spaghetti. Although
>> >> it's more of a FYI rather than "wait until they land".
>> >>
>> >> On the winsys_dri vs winsys_drm side - I'm not planning to do any work
>> >> there, neither I did notice the Xlib.h dependency in vl_winsys.h.
>> >>
>> >> What I'm pondering is about having a 'proper' drm backend, although
>> >> admittedly I haven't looked exactly what libva{-intel-driver,}'s
>> >> definition of that is. I'd assume that moving the non-winsys specifics
>> >> (from vl_winsys_dri.c) to vl_winsys.h and adding a
>> >> vl_screen_texture_from_drawable() equivalent for drm (amongst others).
>> >> As you can tell much of this is guesswork, so if you don't have the
>> >> time and others are happy with the approach as is, feel free to
>> >> ignore.
>> >
>> >
>> > A wayland backend would be nice too.
>> I'm afraid not many of us have the time and/or interest to work on
>> that. Patches implementing it will be kindly accepted :-)
>>
>> > Right now vainfo under wayland just
>> > crashes.
>> >
>> I guess we can separate the VA_DISPLAY_foo switch statement into a
>> separate patch which can also go into stable. Personally I don't mind
>> either way.
>
>
> Julien added the VA_DISPLAY_WAYLAND in the switch in the latest patches, so
> at least an error is now returned instead of crashing.
>
True. But as is the patch won't end up in stable. Which means that
mesa 11.0.x and earlier(?) will continue crashing.
Although my systems is fine - no wayland no problems :-P

-Emil