Re: [kvm-unit-tests PATCH v3 0/3] GIC ITS tests

[Andrew Jones]

On Tue, Nov 30, 2021 at 02:11:34PM +0000, Alex Bennée wrote:
> 
> Andrew Jones <drjones@redhat.com> writes:
> 
> > On Fri, Nov 19, 2021 at 04:30:47PM +0000, Alex Bennée wrote:
> >> 
> >> Andrew Jones <drjones@redhat.com> writes:
> >> 
> >> > On Fri, Nov 12, 2021 at 02:08:01PM +0000, Alex Bennée wrote:
> >> >> 
> >> >> Andrew Jones <drjones@redhat.com> writes:
> >> >> 
> >> >> > On Fri, Nov 12, 2021 at 11:47:31AM +0000, Alex Bennée wrote:
> >> >> >> Hi,
> >> >> >> 
> >> >> >> Sorry this has been sitting in my tree so long. The changes are 
> >> >> >> fairly
> >> >> >> minor from v2. I no longer split the tests up into TCG and KVM
> >> >> >> versions and instead just ensure that ERRATA_FORCE is always set when
> >> >> >> run under TCG.
> >> >> >> 
> >> >> >> Alex Bennée (3):
> >> >> >>   arm64: remove invalid check from its-trigger test
> >> >> >>   arm64: enable its-migration tests for TCG
> >> >> >>   arch-run: do not process ERRATA when running under TCG
> >> >> >> 
> >> >> >>  scripts/arch-run.bash |  4 +++-
> >> >> >>  arm/gic.c             | 16 ++++++----------
> >> >> >>  arm/unittests.cfg     |  3 ---
> >> >> >>  3 files changed, 9 insertions(+), 14 deletions(-)
> >> >> >> 
> >> >> >> -- 
> >> >> >> 2.30.2
> >> >> >> 
> >> >> >> _______________________________________________
> >> >> >> kvmarm mailing list
> >> >> >> kvmarm@lists.cs.columbia.edu
> >> >> >> https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
> >> >> >
> >> >> > Hi Alex,
> >> >> >
> >> >> > Thanks for this. I've applied to arm/queue, but I see that
> >> >> >
> >> >> > FAIL: gicv3: its-trigger: inv/invall: dev2/eventid=20 pending LPI is 
> >> >> > received
> >> >> >
> >> >> > consistently fails for me. Is that expected? Does it work for you?
> >> >> 
> >> >> doh - looks like I cocked up the merge conflict...
> >> >> 
> >> >> Did it fail for TCG or for KVM (or both)?
> >> >
> >> > Just TCG, which was why I was wondering if it was expected. I've never 
> >> > run
> >> > these tests with TCG before.
> >> 
> >> Hmm I think expecting the IRQ at all is broken so I think I should
> >> delete the whole pending test.
> >
> > Feel free to repost. I'll update the patches in arm/queue before my next
> > MR.
> 
> Actually I think the problem was with a regression in the TCG ITS
> support (now fixed in master). So I believe as of v3 everything is
> correct (and v4 should be ignored).
> 
> Are you happy to apply this series or do you want me to repost it as v5?

No need to repost. I'll retest v3 with latest QEMU.

Thanks,
drew