Christian Borntraeger wrote:
quoted text
> Am Montag, 15. Oktober 2007 schrieb Avi Kivity:
>> We can clear it a bit later, after local_irq_enable() in __vcpu_run().=
 =20
quoted text
>> However we need a nop instruction first because "sti" keeps interrupts=
=20
quoted text
>> disabled for one more instruction.
>=20
> Ah, I see. The host interrupt  behaves different and instead of running=
 the
quoted text
> interrupt handler, it exits the vmrun on x86? The interrupt handler wil=
l be
quoted text
> called some cycles after the sti?
> That is different to s390. We can run the guest code for a long time an=
d the
quoted text
> host instruction pointer stays on the sie instruction. That means, we c=
an=20
quoted text
> interrupt sie and continue by simply setting the instrution pointer (PS=
W)=20
quoted text
> back to the sie instruction.
>=20
> Any idea how to make this proper on all architectures? I will have a lo=
ok.

I think the solution is to have an arch dependent kvm_guest_exit(): empty=
 for
x86, clearing the bit for s390.

Regards,
Laurent
--=20
---------------- Laurent.Vivier@bull.net  -----------------
"Given enough eyeballs, all bugs are shallow" E. S. Raymond