Re: [PATCH v2 000/144] KVM: selftests: Overhaul APIs, purge VCPU_ID

[Sean Christopherson]

+s390 folks...

On Fri, Jun 10, 2022, Sean Christopherson wrote:
> On Thu, Jun 09, 2022, Sean Christopherson wrote:
> > On Thu, Jun 09, 2022, Anup Patel wrote:
> > > On Wed, Jun 8, 2022 at 9:26 PM Anup Patel <anup@xxxxxxxxxxxxxx> wrote:
> > > >
> > > > On Tue, Jun 7, 2022 at 8:57 PM Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote:
> > > > >
> > > > > Marc, Christian, Anup, can you please give this a go?
> > > >
> > > > Sure, I will try this series.
> > > 
> > > I tried to apply this series on top of kvm/next and kvm/queue but
> > > I always get conflicts. It seems this series is dependent on other
> > > in-flight patches.
> > 
> > Hrm, that's odd, it's based directly on kvm/queue, commit 55371f1d0c01 ("KVM: ...).
> 
> Duh, Paolo updated kvm/queue.  Where's Captain Obvious when you need him...
> 
> > > Is there a branch somewhere in a public repo ?
> > 
> > https://github.com/sean-jc/linux/tree/x86/selftests_overhaul
> 
> I pushed a new version that's based on the current kvm/queue, commit 5e9402ac128b.
> arm and x86 look good (though I've yet to test on AMD).
> 
> Thomas,
> If you get a chance, could you rerun the s390 tests?  The recent refactorings to
> use TAP generated some fun conflicts.
> 
> Speaking of TAP, I added a patch to convert __TEST_REQUIRE to use ksft_exit_skip()
> instead of KVM's custom print_skip().  The s390 tests are being converted to use
> TAP output, I couldn't see any advantage of KVM's arbitrary "skipping test" over
> TAP-friendly output, and converting everything is far easier than special casing s390.