Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <1587632507-18997-1-git-send-email-wanpengli@tencent.com>
 <1587632507-18997-6-git-send-email-wanpengli@tencent.com> <99d81fa5-dc37-b22f-be1e-4aa0449e6c26@redhat.com>
 <CANRm+CyyKwFoSns31gK=_v0j1VQrOwDhgTqWZOLZS9iGZeC3Gw@mail.gmail.com> <ce15ea08-4d5d-6e7b-9413-b5fcf1309697@redhat.com>
In-Reply-To: <ce15ea08-4d5d-6e7b-9413-b5fcf1309697@redhat.com>
From:   Wanpeng Li <kernellwp@gmail.com>
Date:   Fri, 24 Apr 2020 14:38:03 +0800
Message-ID: <CANRm+CxxQ1oL0fKmnxAv739nsjDg_V7Pgkmm==7CfPQUxBHW4w@mail.gmail.com>
Subject: Re: [PATCH v2 5/5] KVM: VMX: Handle preemption timer fastpath
To:     Paolo Bonzini <pbonzini@redhat.com>
Cc:     LKML <linux-kernel@vger.kernel.org>, kvm <kvm@vger.kernel.org>,
        Sean Christopherson <sean.j.christopherson@intel.com>,
        Vitaly Kuznetsov <vkuznets@redhat.com>,
        Wanpeng Li <wanpengli@tencent.com>,
        Jim Mattson <jmattson@google.com>,
        Joerg Roedel <joro@8bytes.org>,
        Haiwei Li <lihaiwei@tencent.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, 23 Apr 2020 at 18:29, Paolo Bonzini <pbonzini@redhat.com> wrote:
>
> On 23/04/20 11:56, Wanpeng Li wrote:
> >> Please re-evaluate if this is needed (or which parts are needed) after
> >> cleaning up patch 4.  Anyway again---this is already better, I don't
> >> like the duplicated code but at least I can understand what's going on.
> > Except the apic_lvtt_tscdeadline(apic) check, others are duplicated,
> > what do you think about apic_lvtt_tscdeadline(apic) check?
>
> We have to take a look again after you clean up patch 4.  My hope is to
> reuse the slowpath code as much as possible, by introducing some
> optimizations here and there.

I found we are not need to move the if (vcpu->arch.apicv_active) from
__apic_accept_irq() to a separate function if I understand you
correctly. Please see patch v3 3/5. In addition, I observe
kvm-unit-tests #UD etc if check need_cancel_enter_guest() after the
generic fastpath handler, I didn't dig too much, just move it before
the generic fastpath handler for safe in patch v3 2/5.

    Wanpeng