[Mesa-dev] [Bug 100613] Regression in Mesa 17 on s390x (zSystems)

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=100613

--- Comment #17 from Ray Strode <halfline at gmail.com> ---
Created attachment 131171
  --> https://bugs.freedesktop.org/attachment.cgi?id=131171&action=edit
another patch that may aid in understanding what's going on

Just to follow up again, we played around a bit more and figured some stuff out
but we're still kind of banging around in the dark.

What we learned:

- The reason vector_justify = TRUE fixed one of the tests, but kept some
failing is because it was reversing X and Y coordinates, so only the first 3
component sshort test run was getting placed properly (because the test case
has x and y start with the same value).  The subsequent ones were growing down
instead of right.

- of course we can then get it working by changing the swizzle order in
u_format.csv to yxz1 for the 3 component short and 3 component half float
tests. clearly not the "right" fix, but was interesting to see all the tests
working.

- if i leave vector_justify to FALSE, but change attachment 130980 to use
fetch_width instead of format_desc->block.bits then all the 3 component sshort
tests and half float tests start working, but then r32g32b32_sscaled start
failing.  So, the above patch fudges it to make everything in draw-vertices and
draw-vertices-half-float work. Obviously not shippable, but might aid in
figuring out what's going on?

-- 
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/mesa-dev/attachments/20170502/e8b44174/attachment.html>