Module: Mesa
Branch: main
Commit: dbe5b8b5a468981e793dde2b285f6cbb43773304
URL:    
http://cgit.freedesktop.org/mesa/mesa/commit/?id=dbe5b8b5a468981e793dde2b285f6cbb43773304

Author: Rob Clark <[email protected]>
Date:   Wed Dec 20 11:09:57 2023 -0800

ci: More context for color_clear skips for Wayland

Add some more notes, so context isn't lost to time when someone gets
around to digging deeper.

Signed-off-by: Rob Clark <[email protected]>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26779>

---

 .gitlab-ci/all-skips.txt | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/.gitlab-ci/all-skips.txt b/.gitlab-ci/all-skips.txt
index 762706dc4d2..8096892098b 100644
--- a/.gitlab-ci/all-skips.txt
+++ b/.gitlab-ci/all-skips.txt
@@ -39,4 +39,19 @@ glx@glx-copy-sub-buffer.*
 
 # A majority of the tests introduced in CTS 1.3.7.0 are experiencing failures 
and flakes.
 # Disable these tests until someone with a more deeper understanding of EGL 
examines them.
+#
+# Note: on sc8280xp/a690 I get identical results (same passes and fails)
+# between freedreno, zink, and llvmpipe, so I believe this is either a
+# deqp bug or egl/wayland bug, rather than driver issue.
+#
+# With llvmpipe, the failing tests have the error message:
+#
+#   "Illegal sampler view creation without bind flag"
+#
+# which might be a hint.  (But some passing tests also have the same
+# error message.)
+#
+# more context from David Heidelberg on IRC: the deqp commit where these
+# started failing is: 
https://github.com/KhronosGroup/VK-GL-CTS/commit/79b25659bcbced0cfc2c3fe318951c585f682abe
+# prior to that they were skipping.
 wayland-dEQP-EGL.functional.color_clears.*

Reply via email to