[Mesa-dev] [PATCH] i965: Add RGBX8888 and RGBA8888 to EGL configure and reorder the list

[Xu, Randy]

Thanks, Emil

Yes, I have discussed it with Tapani, my proposed solution is not good for desktop Linux.
We will keep the change for Android only. And I will take care of it.


Thanks,
Randy


> -----Original Message-----
> From: Emil Velikov [mailto:emil.l.velikov at gmail.com]
> Sent: Thursday, May 11, 2017 8:31 PM
> To: Xu, Randy <randy.xu at intel.com>
> Cc: ML mesa-dev <mesa-dev at lists.freedesktop.org>
> Subject: Re: [Mesa-dev] [PATCH] i965: Add RGBX8888 and RGBA8888 to EGL
> configure and reorder the list
> 
> Hi Randy,
> 
> Found this stuck in my Drafts.
> 
> On 11 January 2017 at 08:04, Randy Xu <randy.xu at intel.com> wrote:
> > On Android platform, the EGL configuration query (eglGetConfigs) will
> > check the DRI backend. Add R8G8B8A8 and R8G8B8X8 to meet the native
> > app test requirement (like gl2_basic etc.).
> >
> > Reorder the list and move the B5G6R5 to the tail, it can solve two
> > dEQP failures on Android platform, they are dEQP-
> GLES3.functional.fragment_ops.
> >
> blend.default_framebuffer.rgb_func_alpha_func.src.dst_color_one_minus_
> > dst_alpha & .dst_color_constant_alpha. These two failures are all
> > caused by B5G6R5 low precision and the dEQP case always choose the
> > first EGL configure id in list returned by eglGetConfigs. The failure
> > cannot be reproduced on X11 platform as Intel does not support 16bit
> > visuals with X driver
> >
> When developing changes please make sure that you've tested those against
> the normal Linux builds.
> Either handle this yourself, or check with the devs on #intel-gfx about going
> through the Intel CI.
> 
> An identical change broke KDE/Qt in the past [1] and there is some discussion
> how to fix it properly [2].
> 
> Thanks
> Emil
> 
> [1] https://bugs.freedesktop.org/show_bug.cgi?id=95071
> [2] https://lists.freedesktop.org/archives/mesa-dev/2016-April/113654.html
> (check through the thread)