Thomas Huth <th...@redhat.com> writes: > On 13/12/2024 01.26, Nabih Estefan wrote: >> Recent CDMSK Watchdog changes (eff9dc5660fad3a610171c56a5ec3fada245e519) >> updated the CDMSK APB Watchdog to not free run out of reset. That led to >> this test failing since it never triggers the watchdog to start running. >> No watchdog running means that the timer and counter in the test cannot >> start, leading to failures in the assert statements throughout the test. >> Adding a reset and enable of the watchdog to the reset function solves >> this problem by enabling the watchdog and thus letting the timer and >> counter run as expected >> >> Also renaming the reset_counter_and_timer function since it now also >> affects the watchdog. >> >> To reproduce the failure at HEAD: >> ./configure --target-list=arm-softmmu >> make -j check-report-qtest-arm.junit.xml >> >> Signed-off-by: Nabih Estefan <nabiheste...@google.com>
I don't see any failure with the sse-timer-test on master (83aaec1d5a). Is it supposed to be intermittent? Nabih, can you confirm whether this is the same issue Thomas has seen? > I think this fixes the problem that I was seeing on Travis-CI: > > Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2702 I did see this^ issue, however, with Philippe's series [1] (and this patch doesn't fix it). Here's a CI job with both series applied: https://gitlab.com/farosas/qemu/-/jobs/8638126011 1- https://lore.kernel.org/r/20241211233727.98923-2-phi...@linaro.org
