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;
To:     Lai Jiangshan <jiangshanlai+lkml@gmail.com>
Cc:     Sean Christopherson <seanjc@google.com>,
        LKML <linux-kernel@vger.kernel.org>, kvm@vger.kernel.org,
        Filippo Sironi <sironi@amazon.de>,
        David Woodhouse <dwmw@amazon.co.uk>,
        "v4.7+" <stable@vger.kernel.org>,
        Wanpeng Li <wanpengli@tencent.com>
References: <20201127112114.3219360-1-pbonzini@redhat.com>
 <20201127112114.3219360-3-pbonzini@redhat.com>
 <CAJhGHyCdqgtvK98_KieG-8MUfg1Jghd+H99q+FkgL0ZuqnvuAw@mail.gmail.com>
 <YHS/BxMiO6I1VOEY@google.com>
 <CAJhGHyAcnwkCfTcnxXcgAHnF=wPbH2EDp7H+e74ce+oNOWJ=_Q@mail.gmail.com>
 <80b013dc-0078-76f4-1299-3cff261ef7d8@redhat.com>
 <CAJhGHyChfXdcAMzzD7P3aC8tnhFW5GvOt88vOY=D3pyb7hgNAA@mail.gmail.com>
 <6d9dafb1-b8ff-82ef-93dc-da869fe7ba0f@redhat.com>
 <CAJhGHyA=v_va2QTvo7Ve8JyZO4j5LjiCdB9CLnvRXGwGwa3e+A@mail.gmail.com>
From:   Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [PATCH 2/2] KVM: x86: Fix split-irqchip vs interrupt injection
 window request
Message-ID: <5b422691-ffc5-d73a-1bda-f1ee61116756@redhat.com>
Date:   Thu, 15 Apr 2021 08:07:06 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.7.0
MIME-Version: 1.0
In-Reply-To: <CAJhGHyA=v_va2QTvo7Ve8JyZO4j5LjiCdB9CLnvRXGwGwa3e+A@mail.gmail.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 15/04/21 02:59, Lai Jiangshan wrote:
> The next call to inject_pending_event() will reach here AT FIRST with
> vcpu->arch.exception.injected==false and vcpu->arch.exception.pending==false
> 
>>           ... if (!vcpu->arch.exception.pending) {
>>                   if (vcpu->arch.nmi_injected) {
>>                           static_call(kvm_x86_set_nmi)(vcpu);
>>                           can_inject = false;
>>                   } else if (vcpu->arch.interrupt.injected) {
>>                           static_call(kvm_x86_set_irq)(vcpu);
>>                           can_inject = false;
>
> And comes here and vcpu->arch.interrupt.injected is true for there is
> an interrupt queued by KVM_INTERRUPT for pure user irqchip. It then does
> the injection of the interrupt without checking the EFLAGS.IF.

Ok, understood now.  Yeah, that could be a problem for userspace irqchip 
so we should switch it to use pending_external_vector instead.  Are you 
going to write the patch or should I?

Thanks!

Paolo

> My question is that what stops the next call to inject_pending_event()
> to reach here when KVM_INTERRUPT is called with exepction pending.