* [PATCH] KVM: x86: move steal time initialization to vcpu entry time
@ 2015-10-14 22:33 Marcelo Tosatti
2015-10-15 17:10 ` David Matlack
0 siblings, 1 reply; 3+ messages in thread
From: Marcelo Tosatti @ 2015-10-14 22:33 UTC (permalink / raw)
To: kvm; +Cc: Rik van Riel, Dr. David Alan Gilbert, Paolo Bonzini, Amit Shah
As reported at https://bugs.launchpad.net/qemu/+bug/1494350,
it is possible to have vcpu->arch.st.last_steal initialized
from a thread other than vcpu thread, say the iothread, via
KVM_SET_MSRS.
Which can cause an overflow later (when subtracting from vcpu threads
sched_info.run_delay).
To avoid that, move steal time accumulation to vcpu entry time,
before copying steal time data to guest.
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 8f0f6ec..0e0332e 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -2030,6 +2030,8 @@ static void accumulate_steal_time(struct kvm_vcpu *vcpu)
static void record_steal_time(struct kvm_vcpu *vcpu)
{
+ accumulate_steal_time(vcpu);
+
if (!(vcpu->arch.st.msr_val & KVM_MSR_ENABLED))
return;
@@ -2182,12 +2184,6 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
if (!(data & KVM_MSR_ENABLED))
break;
- vcpu->arch.st.last_steal = current->sched_info.run_delay;
-
- preempt_disable();
- accumulate_steal_time(vcpu);
- preempt_enable();
-
kvm_make_request(KVM_REQ_STEAL_UPDATE, vcpu);
break;
@@ -2830,7 +2826,6 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
vcpu->cpu = cpu;
}
- accumulate_steal_time(vcpu);
kvm_make_request(KVM_REQ_STEAL_UPDATE, vcpu);
}
^ permalink raw reply related [flat|nested] 3+ messages in thread* Re: [PATCH] KVM: x86: move steal time initialization to vcpu entry time
2015-10-14 22:33 [PATCH] KVM: x86: move steal time initialization to vcpu entry time Marcelo Tosatti
@ 2015-10-15 17:10 ` David Matlack
2015-10-16 8:34 ` Paolo Bonzini
0 siblings, 1 reply; 3+ messages in thread
From: David Matlack @ 2015-10-15 17:10 UTC (permalink / raw)
To: Marcelo Tosatti
Cc: kvm list, Rik van Riel, Dr. David Alan Gilbert, Paolo Bonzini,
Amit Shah
On Wed, Oct 14, 2015 at 3:33 PM, Marcelo Tosatti <mtosatti@redhat.com> wrote:
>
> As reported at https://bugs.launchpad.net/qemu/+bug/1494350,
> it is possible to have vcpu->arch.st.last_steal initialized
> from a thread other than vcpu thread, say the iothread, via
> KVM_SET_MSRS.
>
> Which can cause an overflow later (when subtracting from vcpu threads
> sched_info.run_delay).
>
> To avoid that, move steal time accumulation to vcpu entry time,
> before copying steal time data to guest.
>
> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Reviewed-by: David Matlack <dmatlack@google.com>
>
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 8f0f6ec..0e0332e 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -2030,6 +2030,8 @@ static void accumulate_steal_time(struct kvm_vcpu *vcpu)
>
> static void record_steal_time(struct kvm_vcpu *vcpu)
> {
> + accumulate_steal_time(vcpu);
> +
> if (!(vcpu->arch.st.msr_val & KVM_MSR_ENABLED))
> return;
>
> @@ -2182,12 +2184,6 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
> if (!(data & KVM_MSR_ENABLED))
> break;
>
> - vcpu->arch.st.last_steal = current->sched_info.run_delay;
> -
> - preempt_disable();
> - accumulate_steal_time(vcpu);
> - preempt_enable();
> -
> kvm_make_request(KVM_REQ_STEAL_UPDATE, vcpu);
>
> break;
> @@ -2830,7 +2826,6 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
> vcpu->cpu = cpu;
> }
>
> - accumulate_steal_time(vcpu);
> kvm_make_request(KVM_REQ_STEAL_UPDATE, vcpu);
> }
>
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH] KVM: x86: move steal time initialization to vcpu entry time
2015-10-15 17:10 ` David Matlack
@ 2015-10-16 8:34 ` Paolo Bonzini
0 siblings, 0 replies; 3+ messages in thread
From: Paolo Bonzini @ 2015-10-16 8:34 UTC (permalink / raw)
To: David Matlack, Marcelo Tosatti
Cc: kvm list, Rik van Riel, Dr. David Alan Gilbert, Amit Shah
On 15/10/2015 19:10, David Matlack wrote:
> > As reported at https://bugs.launchpad.net/qemu/+bug/1494350,
> > it is possible to have vcpu->arch.st.last_steal initialized
> > from a thread other than vcpu thread, say the iothread, via
> > KVM_SET_MSRS.
> >
> > Which can cause an overflow later (when subtracting from vcpu threads
> > sched_info.run_delay).
> >
> > To avoid that, move steal time accumulation to vcpu entry time,
> > before copying steal time data to guest.
> >
> > Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
>
> Reviewed-by: David Matlack <dmatlack@google.com>
Applied to kvm/queue. Thanks Marcelo, and thanks David for the review.
Paolo
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2015-10-16 8:34 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-10-14 22:33 [PATCH] KVM: x86: move steal time initialization to vcpu entry time Marcelo Tosatti
2015-10-15 17:10 ` David Matlack
2015-10-16 8:34 ` Paolo Bonzini
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).