On 13 March 2018 at 18:23, Daniel P. Berrangé <berra...@redhat.com> wrote: > On Tue, Mar 13, 2018 at 04:27:08PM +0000, Daniel P. Berrangé wrote: >> This feels unrelated to the series, so possibly a non-deterministic >> failure > > Looking at the test source code, it has a 60 second wait for the > emulator to print expected data on the serial port. My guess > would be that when running with a highly parallel make check > we're sometimes exceeding that 60 seconds. I don't see a > reason why my series would have affected this in a way that > only failed with sparc64 qtest, and none others.
Sounds plausible. Feel free to resubmit the pullreq with the actual bug fixed, for another spin of the intermittent-failure roulette wheel :-) -- PMM