On Tue, 20 Sept 2022 at 17:45, Thomas Huth <th...@redhat.com> wrote: > this seems to break some Avocado based test(s) in our CI: > > make check-venv > ./tests/venv/bin/avocado run > tests/avocado/replay_kernel.py:ReplayKernelNormal.test_aarch64_virt > > ... fails with commit 01765386a88868ae993bcb but still passes > with the preceeding commit. Could you please have a look?
Thanks for the report. I can reproduce it, it seems to happen with all kinds of icount, not specifically with record-replay. We hit the "Bad icount read" exit in icount_get_raw_locked(). I'll investigate further. (Rather unhelpfully, the avocado framework has ignored the fact that the QEMU process exited with an non-zero status, has failed to log the "Bad icount read" message in the log, and has only failed the test for hitting the 120s timeout :-( ) thanks -- PMM
