On Tue, 5 Mar 2019 05:30:02 -0500
Christian Borntraeger <borntraeger@xxxxxxxxxx> wrote:
> We do track the current steal time of the host CPUs. Let us use
> this value to disable halt polling if the steal time goes beyond
> a predefined value. (12.5% for now).
Might make sense to make it tweakable. On the other hand, a reasonable
value might be good enough :)
>
> Signed-off-by: Christian Borntraeger <borntraeger@xxxxxxxxxx>
> ---
> arch/s390/include/asm/kvm_host.h | 6 ++++++
> arch/s390/kvm/Kconfig | 1 +
> 2 files changed, 7 insertions(+)
>
> diff --git a/arch/s390/include/asm/kvm_host.h b/arch/s390/include/asm/kvm_host.h
> index a4962761e1dd..f699983a8d09 100644
> --- a/arch/s390/include/asm/kvm_host.h
> +++ b/arch/s390/include/asm/kvm_host.h
> @@ -884,6 +884,12 @@ struct kvm_arch_async_pf {
> unsigned long pfault_token;
> };
>
> +static inline bool kvm_arch_no_poll(void)
> +{
> + /* do not poll with more than 12.5% steal */
> + return S390_lowcore.avg_steal_timer > (TICK_USEC << 12) / 8;
That's an average anyway, correct?
> +}
> +
> bool kvm_arch_can_inject_async_page_present(struct kvm_vcpu *vcpu);
>
> void kvm_arch_async_page_ready(struct kvm_vcpu *vcpu,
> diff --git a/arch/s390/kvm/Kconfig b/arch/s390/kvm/Kconfig
> index 767453faacfc..e987e73738b5 100644
> --- a/arch/s390/kvm/Kconfig
> +++ b/arch/s390/kvm/Kconfig
> @@ -31,6 +31,7 @@ config KVM
> select HAVE_KVM_IRQFD
> select HAVE_KVM_IRQ_ROUTING
> select HAVE_KVM_INVALID_WAKEUPS
> + select HAVE_KVM_NO_POLL
> select SRCU
> select KVM_VFIO
> ---help---
