Re: [PATCH RFC 4/6] KVM: x86: acknowledgment mechanism for async pf page ready notifications

From: Vitaly Kuznetsov
Date: Thu Apr 30 2020 - 04:40:26 EST


Paolo Bonzini <pbonzini@xxxxxxxxxx> writes:

> On 29/04/20 11:36, Vitaly Kuznetsov wrote:
>> + case MSR_KVM_ASYNC_PF_ACK:
>> + if (data & 0x1)
>> + kvm_check_async_pf_completion(vcpu);
>> + break;
>
> Does this work if interrupts are turned off?
> I think in that case
> kvm_check_async_pf_completion will refuse to make progress.
> You need to make this bit stateful (e.g. 1 = async PF in progress, 0 =
> not in progress), and check that for page ready notifications instead of
> EFLAGS.IF.
> This probably means that;
>
> - it might be simpler to move it to the vector MSR

I didn't want to merge 'ACK' with the vector MSR as it forces the guest
to remember the setting. It doesn't matter at all for Linux as we
hardcode the interrupt number but I can imaging an OS assigning IRQ
numbers dynamically, it'll need to keep record to avoid doing rdmsr.

> - it's definitely much simpler to remove the #PF-based mechanism for
> injecting page ready notifications.

Yea, the logic in kvm_can_do_async_pf()/kvm_can_deliver_async_pf()
becomes cumbersome. If we are to drop #PF-based mechanism I'd split it
completely from the remaining synchronious #PF for page-not-present:
basically, we only need to check that the slot (which we agreed becomes
completely separate) is empty, interrupt/pending expception/... state
becomes irrelevant.

--
Vitaly