Index: content/test/gpu/gpu_tests/webgl2_conformance_expectations.py |
diff --git a/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py b/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py |
index 05a6ce8125c82bba11c1350f9b4ea8f54a0d6ab3..6361f30b37d2b04f761085e64c74f55ef96c1714 100644 |
--- a/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py |
+++ b/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py |
@@ -115,6 +115,12 @@ class WebGL2ConformanceExpectations(WebGLConformanceExpectations): |
# and updated driver. The older drivers won't ever get fixes from AMD. |
# Use ['win', ('amd', 0x6613)] for the R7 240 devices. |
+ # Have seen this time out. Think it may be because it's currently |
+ # the first test that runs in the shard, and the browser might not |
+ # be coming up correctly. |
+ self.Flaky('deqp/functional/gles3/multisample.html', |
+ ['win', ('amd', 0x6613)], bug=687374) |
+ |
# It's unfortunate that these suppressions need to be so broad, but |
# basically any test that uses readPixels is potentially flaky, and |
# it's infeasible to suppress individual failures one by one. |