[Bug 105301] The big SKQP bug

[bugzilla-daemon at freedesktop.org]

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

--- Comment #18 from Kenneth Graunke <kenneth at whitecape.org> ---
About mixedtextblobs...

https://gitlab.freedesktop.org/kwg/mesa/commits/skia has a patch, "i965: Always
use dual source RT writes if src1 blend factors are enabled", which makes the
mixedtextblobs tests pass.

Note that the SKQP test is broken.  It's enabled dual source blend factors
(SRC1_COLOR etc.) but provided a shader that only writes one color output, when
it ought to write two colors.  For reference, the one that is hanging is the
fragment shader with the 'discard', which only outputs a single color.

In fact, none of the fragment shaders appear to have a layout(index = 1)
qualifier indicating which is the second color.  Nor does SQKP call
glBindFragDataLocationIndexed to set one (which is the other mechanism for
doing this).  So, the results are undefined.  Search
https://www.khronos.org/registry/OpenGL/extensions/ARB/ARB_blend_func_extended.txt
for "not defined" to find the spec citation.

That said, we should not hang in this case - just render some undefined
garbage.  Doing so appears to suit SKQP just fine (as does rendering nothing at
all, amusingly).

If you guys could pass that on to the SKQP authors, it would be great to get
the bug in the test fixed.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20181030/80eb9d7f/attachment.html>