Linux Virtualization - Aug 2013 - [PATCH V12 3/5] kvm : Fold pv_unhalt flag into GET_MP_STATE ioctl to aid migration

On Tue, Aug 6, 2013 at 11:23 AM, Raghavendra K T
<raghavendra.kt at linux.vnet.ibm.com> wrote:
> kvm : Fold pv_unhalt flag into GET_MP_STATE ioctl to aid migration
>
> From: Raghavendra K T <raghavendra.kt at linux.vnet.ibm.com>
>
> During migration, any vcpu that got kicked but did not become runnable
> (still in halted state) should be runnable after migration.
If this is about migration correctness, could it get folded into the
previous patch 2/5, so that there's not a broken commit which could
hurt bisection?
>
> Signed-off-by: Raghavendra K T <raghavendra.kt at linux.vnet.ibm.com>
> Acked-by: Gleb Natapov <gleb at redhat.com>
> Acked-by: Ingo Molnar <mingo at kernel.org>
> ---
>  arch/x86/kvm/x86.c |    7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index dae4575..1e73dab 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -6284,7 +6284,12 @@ int kvm_arch_vcpu_ioctl_get_mpstate(struct kvm_vcpu
*vcpu,
>                                     struct kvm_mp_state *mp_state)
>  {
>         kvm_apic_accept_events(vcpu);
> -       mp_state->mp_state = vcpu->arch.mp_state;
> +       if (vcpu->arch.mp_state == KVM_MP_STATE_HALTED &&
> +                                       vcpu->arch.pv.pv_unhalted)
> +               mp_state->mp_state = KVM_MP_STATE_RUNNABLE;
> +       else
> +               mp_state->mp_state = vcpu->arch.mp_state;
> +
>         return 0;
>  }
>
>
> _______________________________________________
> Virtualization mailing list
> Virtualization at lists.linux-foundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/virtualization

On 08/07/2013 12:02 AM, Eric Northup wrote:
>
> If this is about migration correctness, could it get folded into the
> previous patch 2/5, so that there's not a broken commit which could
> hurt bisection?
Yes. It could be. Only reason I maintained like that was,
original author in the previous patch is different (Srivatsa) and I did
not want to merge this hunk when the patch series got evolved to mix
the sign-offs.

Gleb, Paolo please let me know.

On Wed, Aug 07, 2013 at 12:40:36AM +0530, Raghavendra K T
wrote:
> On 08/07/2013 12:02 AM, Eric Northup wrote:
> >
> >If this is about migration correctness, could it get folded into the
> >previous patch 2/5, so that there's not a broken commit which could
> >hurt bisection?
> 
> Yes. It could be. Only reason I maintained like that was,
> original author in the previous patch is different (Srivatsa) and I did
> not want to merge this hunk when the patch series got evolved to mix
> the sign-offs.
> 
> Gleb, Paolo please let me know.
> 
Yes please, do so.

--
			Gleb.