[Bug 105301] The big SKQP bug

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=105301

--- Comment #6 from Tapani Pälli <lemody at gmail.com> ---
(In reply to Aditya Swarup from comment #2)
> Another point I would like to add is - In the specs it is not mentioned that
> EGLImages should only be created with EGL_EXT_image_dma_buf_import. I think
> we(Intel) are imposing the restriction that EGL image will only be supported
> through Linux dma buf by using extension EGL_LINUX_DMA_BUF_EXT.

There is no such restriction. We do support creating EGLImages from regular 2D
textures, native platform pixmaps and also renderbuffers and pass the tests
written for this exact purpose (Piglit, deqp).

However we don't support "mixmatching" of resources in the way this test does.

> Also, this test passes on ARM(Pixel 2 phones) and seems to be only affecting
> Intel platforms.
> 
> I understand that 2D textures are different from external textures.
> 
> From the specs it seems to be wrt usage of following operations TexImage2D,
> TexSubImage2D,CompressedTexImage2D, CompressedTexSubImage2D, CopyTexImage2D,
> or
> CopyTexSubImage2D not permitted for external textures(which should be an
> INVALID operation for external textures in any case and since the texture
> object is bound to TEXTURE_EXTERNAL_OES, it shouldn't matter if EGLImage
> resource was created with a regular texture. It would be considered an
> external texture till it is deleted).

But in practice from implementation point of view it can make a big difference
how do we treat the resource. Note that at same time, we would need to make
sure the resource is good for external use-cases but also all of those APIs
would need to work when using the regular handle.

It seems removing the restriction causes problems on Android so I believe there
are some corner cases that do not work correctly without this restriction in
place.

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20181023/8d0de5f4/attachment-0001.html>