Something has happened internally timing-wise, and the 20 ms magic timeout has become too short.
Even 60 ms can still fail, as demonstrated here (some output snipped, gtest arguments from earlier failure-reducer). 70 or 80 ms works, but to be safe, we should probably go higher. (Fix incoming, just creating an issue for it so that the fix is more searchable.)
$ cd core
$ FIFO_TEST_TIMEOUT=60 ./all_test --gtest_shuffle \
--gtest_random_seed=58276 --gtest_filter=EmTableTest.*:FifoTestFixture.*
Note: Google Test filter = EmTableTest.*:FifoTestFixture.*
Note: Randomizing tests' orders with a seed of 58276 .
[==========] Running 11 tests from 2 test cases.
[----------] Global test environment set-up.
[----------] 7 tests from EmTableTest
[ RUN ] EmTableTest.FindMakeKeysPktBatch
...
child 2: failed getting message from parent
utils/fifo_test.cc:309: Failure
Value of: ok
Actual: false
Expected: true
failed to send <close> to child 2
[ FAILED ] FifoTestFixture.MultipleFancyFifos (615 ms)
[----------] 4 tests from FifoTestFixture (1043 ms total)
[----------] Global test environment tear-down
[==========] 11 tests from 2 test cases ran. (1317 ms total)
[ PASSED ] 10 tests.
[ FAILED ] 1 test, listed below:
[ FAILED ] FifoTestFixture.MultipleFancyFifos
Something has happened internally timing-wise, and the 20 ms magic timeout has become too short.
Even 60 ms can still fail, as demonstrated here (some output snipped, gtest arguments from earlier failure-reducer). 70 or 80 ms works, but to be safe, we should probably go higher. (Fix incoming, just creating an issue for it so that the fix is more searchable.)