On 01/24/2018 05:15 PM, Cornelia Huck wrote:
> On Wed, 24 Jan 2018 16:50:54 +0100
> Christian Borntraeger <borntraeger@xxxxxxxxxx> wrote:
>
>> The overall instruction counter is larger than the sum of the
>> single counters. We should try to catch all instruction handlers
>> to make this match the summary counter.
>> Let us add sck,tb,sske,iske,rrbe,tb,tpi,tsch,lpsw,pswe....
>> and remove other unused ones.
>
> This is rebased on top of the diagnose patch, right?
Yes, I did not wanted to change both patches.
>
>>
>> Signed-off-by: Christian Borntraeger <borntraeger@xxxxxxxxxx>
>> Acked-by: Janosch Frank <frankja@xxxxxxxxxxxxxxxxxx>
>> Reviewed-by: David Hildenbrand <david@xxxxxxxxxx>
>> ---
>> arch/s390/include/asm/kvm_host.h | 21 +++++++++++++++++----
>> arch/s390/kvm/diag.c | 1 +
>> arch/s390/kvm/kvm-s390.c | 21 +++++++++++++++++----
>> arch/s390/kvm/priv.c | 32 ++++++++++++++++++++++++++++++--
>> 4 files changed, 65 insertions(+), 10 deletions(-)
>>
>
>> @@ -527,6 +544,7 @@ static int handle_io_inst(struct kvm_vcpu *vcpu)
>> if (vcpu->arch.sie_block->ipa == 0xb235)
>> return handle_tsch(vcpu);
>> /* Handle in userspace. */
>> + vcpu->stat.instruction_io_other++;
>> return -EOPNOTSUPP;
>> } else {
>> /*
>
> As an aside, we don't count I/O instructions if userspace did not
> enable css support. But nobody does that with modern kernels anyway :)
>
> Reviewed-by: Cornelia Huck <cohuck@xxxxxxxxxx>
Thanks
