Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <1648216709-44755-1-git-send-email-wanpengli@tencent.com>
 <1648216709-44755-4-git-send-email-wanpengli@tencent.com> <YkOfJeXm8MiMOEyh@google.com>
In-Reply-To: <YkOfJeXm8MiMOEyh@google.com>
From:   Wanpeng Li <kernellwp@gmail.com>
Date:   Wed, 30 Mar 2022 09:18:38 +0800
Message-ID: <CANRm+CxtyyGvBKMt2XpqvWwukAmZHvE7-SBKc7wLcYLR5D7v9g@mail.gmail.com>
Subject: Re: [PATCH RESEND 3/5] KVM: X86: Boost vCPU which is in critical section
To:     Sean Christopherson <seanjc@google.com>
Cc:     LKML <linux-kernel@vger.kernel.org>, kvm <kvm@vger.kernel.org>,
        Paolo Bonzini <pbonzini@redhat.com>,
        Vitaly Kuznetsov <vkuznets@redhat.com>,
        Wanpeng Li <wanpengli@tencent.com>,
        Jim Mattson <jmattson@google.com>,
        Joerg Roedel <joro@8bytes.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Wed, 30 Mar 2022 at 08:07, Sean Christopherson <seanjc@google.com> wrote:
>
> On Fri, Mar 25, 2022, Wanpeng Li wrote:
> > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> > index 425fd7f38fa9..6b300496bbd0 100644
> > --- a/arch/x86/kvm/x86.c
> > +++ b/arch/x86/kvm/x86.c
> > @@ -10375,6 +10375,28 @@ static int vcpu_run(struct kvm_vcpu *vcpu)
> >       return r;
> >  }
> >
> > +static int kvm_vcpu_non_preemptable(struct kvm_vcpu *vcpu)
>
> s/preemtable/preemptible
>
> And I'd recommend inverting the return, and also return a bool, i.e.
>
> static bool kvm_vcpu_is_preemptible(struct kvm_vcpu *vcpu)

Good suggestion.

>
> > +{
> > +     int count;
> > +
> > +     if (!vcpu->arch.pv_pc.preempt_count_enabled)
> > +             return 0;
> > +
> > +     if (!kvm_read_guest_cached(vcpu->kvm, &vcpu->arch.pv_pc.preempt_count_cache,
> > +         &count, sizeof(int)))
> > +             return (count & ~PREEMPT_NEED_RESCHED);
>
> This cements PREEMPT_NEED_RESCHED into KVM's guest/host ABI.  I doubt the sched
> folks will be happy with that.
>
> > +
> > +     return 0;
> > +}
> > +