[Mesa-stable] [PATCH (10.6)] egl/dri2: implement platform_surfaceless

[Emil Velikov]

Hi all,

On 25 August 2015 at 18:09, Chad Versace <chad.versace at intel.com> wrote:
> From: Haixia Shi <hshi at chromium.org>
>
> The surfaceless platform is for off-screen rendering only. Render node support
> is required.
>
> Only consider the render nodes. Do not use normal nodes as they require
> auth hooks.
>
> v3: change platform_null to platform_surfaceless
> v4: make libdrm required for surfaceless
> v5: remove modified include guards with defined(HAVE_SURFACELESS_PLATFORM)
> v6: use O_CLOEXEC for drm fd
>
> Cc: "10.6" <mesa-stable at lists.freedesktop.org>
> Cc: Mark Janes <mark.a.janes at intel.com>
> Signed-off-by: Haixia Shi <hshi at chromium.org>
> Signed-off-by: Zach Reizner <zachr at google.com>
> Reviewed-by: Emil Velikov <emil.l.velikov at gmail.com>
> Reviewed-by: Chad Versace <chad.versace at intel.com>
> (cherry picked from commit 6b8accb36b541f77774109ea42533c02bb90bc68)
> ---
>
> This is a cherry-pick of EGL's surfaceless backend to the 10.6 branch.
> If upstream 10.6 included this feature, that would make validation
> easier for several groups inside Intel (and probably Google too).
>
> I think this is safe to cherry-pick into a stable release because it
> will have zero impact on stability:
>
>   - The feature isn't enabled (or even compiled into libEGL) unless it's
>     explicitly enabled at configure time.  In other words, this
>     cherry-pick will add no new feature to distro Mesa packages.
>
>   - The patch's changes are largely self-contained and confined to a new
>     file (platform_surfaceless.c). That file is built only if the
>     feature is enabled at configure time.
>
>   - Changes to existing source files are minimal (+/- 14 lines) and
>     safe. They consist of adding the new enum _EGL_PLATFORM_SURFACELESS
>     and some code that switches on it.
>
> I understand there's not much precedent for adding patches like this to
> a stable branch. So my feelings won't get hurt if the patch is rejected.
>
I'm a bit ambivalent about this bth.

On one hand it's nicely isolated from everything else, requires
additional configure switch and even an env variable to get it running
(EGL_*_platform_surfaceless anyone ?). On the other hand it's never a
good idea to add new features (platform support in this case) in
bugfix releases esp this late on. The fact that there is no upstream
user for it (that I know of) doesn't make it much better :'(

I'm be ok with pulling this, if we get a positive response from other
(non intel/google) team/developer.

Having someone spend a few minutes to remove (or forward to the ML)
the 40 or so patches in the chromiumos tree could also help ;-)

Thanks
Emil