diff options
| author | erikwright@chromium.org <erikwright@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2012-09-14 18:11:21 +0000 |
|---|---|---|
| committer | erikwright@chromium.org <erikwright@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2012-09-14 18:11:21 +0000 |
| commit | 0390722347f3e4b3dc8feb3a721d20bc510ed8d8 (patch) | |
| tree | 5279c6bbb7c117438df0e378e36903c9338a471a /tools | |
| parent | eb666c51391b28cbdb37fb4a9f371caa77120424 (diff) | |
| download | chromium_src-0390722347f3e4b3dc8feb3a721d20bc510ed8d8.zip chromium_src-0390722347f3e4b3dc8feb3a721d20bc510ed8d8.tar.gz chromium_src-0390722347f3e4b3dc8feb3a721d20bc510ed8d8.tar.bz2 | |
A speculative fix for TSAN and HeapCheck failures in InMemoryURLIndexCacheTest.
These failures may be due to poor assumptions that were previously validated as a side-effect of code that I removed.
Basically, the test is reaching into the history service and grabbing its DB, then accessing it off of the service's thread. This is fundamentally wrong but would still work if you wait until the history service is completely idle. Previously BlockUntilBookmarkModelLoaded might have had this side-effect.
I was unable to repro the failure locally with TSan, so removing the suppressions and observing to see if the problem recurs.
R=timurrrr@chromium.org,sky@chromium.org
BUG=146322,146265
TEST=observe TSan/Heapcheck bots.
Review URL: https://chromiumcodereview.appspot.com/10914125
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@156840 0039d316-1c4b-4281-b951-d872f2087c98
Diffstat (limited to 'tools')
| -rw-r--r-- | tools/heapcheck/unit_tests.gtest-heapcheck.txt | 3 | ||||
| -rw-r--r-- | tools/valgrind/tsan/suppressions.txt | 21 |
2 files changed, 0 insertions, 24 deletions
diff --git a/tools/heapcheck/unit_tests.gtest-heapcheck.txt b/tools/heapcheck/unit_tests.gtest-heapcheck.txt index c7ddacf..5dc3fa3 100644 --- a/tools/heapcheck/unit_tests.gtest-heapcheck.txt +++ b/tools/heapcheck/unit_tests.gtest-heapcheck.txt @@ -35,9 +35,6 @@ SyncNewNonFrontendDataTypeControllerTest.StartAfterSyncShutdown # Randomly fails on the linux heapcheck bot HistoryQuickProviderTest.* -# http://crbug.com/146265 -InMemoryURLIndexTest.* - # Don't run FLAKY or FAILS tests under the heapchecker. # They may cause flaky leak reports and timeouts. *FLAKY_* diff --git a/tools/valgrind/tsan/suppressions.txt b/tools/valgrind/tsan/suppressions.txt index 8407925..c1b686c 100644 --- a/tools/valgrind/tsan/suppressions.txt +++ b/tools/valgrind/tsan/suppressions.txt @@ -989,24 +989,3 @@ fun:media::MockMediaSource::AppendAtTime fun:media::PipelineIntegrationTest_MediaSource_ConfigChange_WebM_Test::TestBody } -{ - bug_146322_a - ThreadSanitizer:Race - ... - fun:sql::Connection::StatementRefCreated - fun:sql::Connection::StatementRef::StatementRef - fun:sql::Connection::GetUniqueStatement - ... - fun:history::* -} -{ - bug_146322_b - ThreadSanitizer:Race - ... - fun:sql::Connection::StatementRefDeleted - fun:sql::Connection::StatementRef::~StatementRef - fun:base::RefCounted::Release - fun:scoped_refptr::~scoped_refptr - fun:sql::Statement::~Statement - fun:history::InMemoryURLIndexTest::SetUp -} |