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;
Subject: Re: [PATCH 12/13] KVM: x86: Replace late check_nested_events() hack
 with more precise fix
To:     Sean Christopherson <sean.j.christopherson@intel.com>
Cc:     Jim Mattson <jmattson@google.com>,
        Vitaly Kuznetsov <vkuznets@redhat.com>,
        Wanpeng Li <wanpengli@tencent.com>,
        Joerg Roedel <joro@8bytes.org>, kvm list <kvm@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Oliver Upton <oupton@google.com>,
        Peter Shier <pshier@google.com>
References: <20200423022550.15113-1-sean.j.christopherson@intel.com>
 <20200423022550.15113-13-sean.j.christopherson@intel.com>
 <CALMp9eTiGdYPpejAOLNz7zzqP1wPXb_zSL02F27VMHeHGzANJg@mail.gmail.com>
 <20200428222010.GN12735@linux.intel.com>
 <6b35ec9b-9565-ea6c-3de5-0957a9f76257@redhat.com>
 <20200429164547.GF15992@linux.intel.com>
From:   Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <286738de-c268-f0b6-f589-6d9d9ad3dc4a@redhat.com>
Date:   Wed, 29 Apr 2020 18:58:45 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.6.0
MIME-Version: 1.0
In-Reply-To: <20200429164547.GF15992@linux.intel.com>
Content-Type: text/plain; charset=windows-1252
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 29/04/20 18:45, Sean Christopherson wrote:
> 
> Can you just drop 9/13, "Prioritize SMI over nested IRQ/NMI" from kvm/queue?
> It's probably best to deal with this in a new series rather than trying to
> squeeze it in.

With AMD we just have IRQ/NMI/SMI, and it's important to handle SMI in
check_nested_events because you can turn SMIs into vmexit without stuff
such as dual-monitor treatment.  On the other hand there is no MTF and
we're not handling exceptions yet.  So, since SMIs should be pretty rare
anyway, I'd rather just add a comment detailing the correct order and
why we're not following it.  The minimal fix would be to move SMI above
the preemption timer, right?

Paolo