Re: getting the console output for s390 cdrom-test?

[Peter Maydell <peter.maydell@linaro.org>]

On Fri, 12 Feb 2021 at 11:45, Thomas Huth <thuth@redhat.com> wrote:
> I guess Cornelia simply meant a:
>
>  asm volatile("nop":::"memory");

That will force the compiler not to reorder, but it doesn't
seem to me that it would really force the memory accesses to
appear to the other host CPU thread that's running the device
emulation. So either it's insufficient in the s390 memory model,
or our s390 emulation isn't being sufficiently rigorous about
implementing the memory model...

> Anyway, I've now succeeded in getting my hands on an aarch64 host and tried
> to reproduce the issue, but even after running it in a loop like this:
>
>   for ((x=0;x<200;x++)) ; do \
>    QTEST_QEMU_BINARY=./qemu-system-s390x tests/qtest/cdrom-test \
>    || break ; \
>   done
>
> ... I was not able to reproduce the issue. What kind of host distro are you
> using there? Could the exact host CPU type matter here, too? (I was running
> my code on an HPE Apollo system, with Fedora 33, gcc 10.2)

I expect the host CPU type matters a lot, yes. dmesg reports it
as "cavium,thunder-88xx", with 96 cores.

thanks
-- PMM