Exclude gltestpersistentcache from TSAN bot.

This bot occasionally produces empty GM images. We believe this is due
to TSAN's thread scheduling interacting poorly with threaded GL drivers.

gltestpersistentcache fails if the cold cache and hot cache images don't
match and this flakiness causes it to fail intermittently.

Change-Id: Ib914c547120e05fac921d7e70de8e36a53bfb61b
Reviewed-on: https://skia-review.googlesource.com/142591
Commit-Queue: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: Brian Salomon <bsalomon@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This commit is contained in:
Brian Salomon 2018-07-19 17:01:09 -04:00 committed by Skia Commit-Bot
parent a76b7a9ded
commit 6a3957c1f9

View File

@ -248,7 +248,7 @@ def dm_flags(api, bot):
# Test 1010102 on our Linux/NVIDIA bots and the persistent cache config
# on the GL bots.
if ('QuadroP400' in bot and 'PreAbandonGpuContext' not in bot and
api.vars.is_linux):
'TSAN' not in bot and api.vars.is_linux):
if 'Vulkan' in bot:
configs.append('vk1010102')
# Decoding transparent images to 1010102 just looks bad