On Fri, 25 May 2018 13:37:07 +0200 David Hildenbrand <da...@redhat.com> wrote:
> If the CPU data is migrated after the TOD clock, the CKC timer of a CPU > is not rearmed. Let's rearm it when loading the CPU state. > > Signed-off-by: David Hildenbrand <da...@redhat.com> > --- > target/s390x/machine.c | 7 +++++++ > 1 file changed, 7 insertions(+) > > diff --git a/target/s390x/machine.c b/target/s390x/machine.c > index 84b4928755..53d226ff48 100644 > --- a/target/s390x/machine.c > +++ b/target/s390x/machine.c > @@ -34,6 +34,13 @@ static int cpu_post_load(void *opaque, int version_id) > return kvm_s390_vcpu_interrupt_post_load(cpu); > } > > +#ifdef CONFIG_TCG > + if (tcg_enabled()) { > + /* Rearm the CKC timer if necessary */ > + tcg_s390_tod_updated(CPU(cpu), RUN_ON_CPU_NULL); Stub this out so the explicit CONFIG_TCG is not required? But I don't care too much. > + } > +#endif > + > return 0; > } >