Re: [PATCH v1 6/9] s390x: kvm: topology: interception of PTF instruction

[Pierre Morel]

On 7/16/21 11:22 AM, Cornelia Huck wrote:
> On Wed, Jul 14 2021, Pierre Morel <pmorel@linux.ibm.com> wrote:
>
> > Interception of the PTF instruction depending on the new
> > KVM_CAP_S390_CPU_TOPOLOGY KVM extension.
>
> Wasn't that the capability that you dropped?

yes,

> Is PTF supposed to be always intercepting? If that isn't configurable,
> wouldn't older QEMUs generate exceptions for it? I'm a bit confused.

Yes, PTF generated an OPERATION exception on old QEMU, but was not used 
by the guest if it has not the topology facility 11.

So just as for STSI, I think we need the KVM_CAP_S390_CPU_TOPOLOGY I 
dropped because otherwise, now that the kernel will advertise facility 
11, the guest will use it and it will get the exception that it should 
not get.

Regards,
Pierre

> > A global value is used to remember if a Topology change occured since
> > the last interception of a PTF instruction with function code 0.
> >
> > Signed-off-by: Pierre Morel <pmorel@linux.ibm.com>
> > ---
> >   hw/s390x/cpu-topology.c            | 19 +++++++++++
> >   include/hw/s390x/cpu-topology.h    |  8 +++++
> >   include/hw/s390x/s390-virtio-ccw.h |  1 +
> >   target/s390x/cpu.c                 |  4 +++
> >   target/s390x/cpu.h                 |  1 +
> >   target/s390x/kvm/kvm.c             | 52 ++++++++++++++++++++++++++++++
> >   6 files changed, 85 insertions(+)

--
Pierre Morel
IBM Lab Boeblingen