diff options
| author | Kevin Krakauer <krakauer@google.com> | 2021-07-23 12:04:49 -0700 |
|---|---|---|
| committer | gVisor bot <gvisor-bot@google.com> | 2021-07-23 12:07:57 -0700 |
| commit | ea0d9a8f48f13bd15f9450848bd515a337750192 (patch) | |
| tree | 38760954413a69c17e8dfa7a00be3ed07bf367e6 | |
| parent | c3c5c55d134dcc22480984e3882072b936bae899 (diff) | |
buildkite: deflake itimer test
The self-admittedly arbitrary threshold of 20% of alarms being sent to the main
thread was being breached. I saw a run of of ~20-23% [1]. We should allow for a
little more breathing room.
1 - https://buildkite.com/gvisor/pipeline/builds/7400
PiperOrigin-RevId: 386503482
| -rw-r--r-- | test/syscalls/linux/itimer.cc | 7 |
1 files changed, 4 insertions, 3 deletions
diff --git a/test/syscalls/linux/itimer.cc b/test/syscalls/linux/itimer.cc index ac113e6da..9fb04eae6 100644 --- a/test/syscalls/linux/itimer.cc +++ b/test/syscalls/linux/itimer.cc @@ -197,9 +197,9 @@ int TestSIGALRMToMainThread() { // (but don't guarantee it), so we expect to see most samples on the main // thread. // - // The number of SIGALRMs delivered to a worker should not exceed 20% + // The number of SIGALRMs delivered to a worker should not exceed 40% // of the number of total signals expected (this is somewhat arbitrary). - const int worker_threshold = result.expected_total / 5; + const int worker_threshold = result.expected_total / 5 * 2; // // Linux only guarantees timers will never expire before the requested time. @@ -230,7 +230,8 @@ TEST(ItimerTest, DeliversSIGALRMToMainThread) { // Not required anymore. kill.Release(); - EXPECT_TRUE(WIFEXITED(status) && WEXITSTATUS(status) == 0) << status; + EXPECT_EQ(WIFEXITED(status) && WEXITSTATUS(status), 0) + << WIFEXITED(status) << " " << WEXITSTATUS(status); } // Signals are delivered to threads fairly. |