Re: [PATCH] KVM: x86: Ensure deadline timer has truly expired before posting its IRQ

[Paolo Bonzini]

On 05/03/21 03:18, Sean Christopherson wrote:
> When posting a deadline timer interrupt, open code the checks guarding
> __kvm_wait_lapic_expire() in order to skip the lapic_timer_int_injected()
> check in kvm_wait_lapic_expire().  The injection check will always fail
> since the interrupt has not yet be injected.  Moving the call after
> injection would also be wrong as that wouldn't actually delay delivery
> of the IRQ if it is indeed sent via posted interrupt.
>
> Fixes: 010fd37fddf6 ("KVM: LAPIC: Reduce world switch latency caused by timer_advance_ns")
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Sean Christopherson <seanjc@xxxxxxxxxx>
> ---
>   arch/x86/kvm/lapic.c | 11 ++++++++++-
>   1 file changed, 10 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
> index 45d40bfacb7c..cb8ebfaccfb6 100644
> --- a/arch/x86/kvm/lapic.c
> +++ b/arch/x86/kvm/lapic.c
> @@ -1642,7 +1642,16 @@ static void apic_timer_expired(struct kvm_lapic *apic, bool from_timer_fn)
>   	}
>   
>   	if (kvm_use_posted_timer_interrupt(apic->vcpu)) {
> -		kvm_wait_lapic_expire(vcpu);
> +		/*
> +		 * Ensure the guest's timer has truly expired before posting an
> +		 * interrupt.  Open code the relevant checks to avoid querying
> +		 * lapic_timer_int_injected(), which will be false since the
> +		 * interrupt isn't yet injected.  Waiting until after injecting
> +		 * is not an option since that won't help a posted interrupt.
> +		 */
> +		if (vcpu->arch.apic->lapic_timer.expired_tscdeadline &&
> +		    vcpu->arch.apic->lapic_timer.timer_advance_ns)
> +			__kvm_wait_lapic_expire(vcpu);
>   		kvm_apic_inject_pending_timer_irqs(apic);
>   		return;
>   	}

Queued, thanks.

Paolo