[Bug 94297] [regression, bisected] spec.arb_tessellation_shader.compiler.barrier-switch-always.tesc

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=94297

Kenneth Graunke <kenneth at whitecape.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|medium                      |low

--- Comment #1 from Kenneth Graunke <kenneth at whitecape.org> ---
Although this is technically a regression, I don't really consider it critical.

The problem is that the GLSL compiler checks for barrier being in an allowed
position too late.  We passed this test previously because our switch
statements always generated lousy code (pointless loops that ran for a single
iteration). Now, we don't.

Also, the result of this bug is that we accept programs that we should reject
according to the rules.  We also happen to compile them and do what the author
intended.  So it's a portability problem (apps written against our driver may
work on our driver but not work on other drivers), but not going to break any
valid apps.

-- 
You are receiving this mail because:
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/20160601/b172032b/attachment.html>