Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH v2] x86/kvm: Disable KVM_ASYNC_PF_SEND_ALWAYS
To:     Andy Lutomirski <luto@amacapital.net>
Cc:     Andrew Cooper <andrew.cooper3@citrix.com>,
        Andy Lutomirski <luto@kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Sean Christopherson <sean.j.christopherson@intel.com>,
        Vivek Goyal <vgoyal@redhat.com>,
        Peter Zijlstra <peterz@infradead.org>,
        LKML <linux-kernel@vger.kernel.org>, X86 ML <x86@kernel.org>,
        kvm list <kvm@vger.kernel.org>, stable <stable@vger.kernel.org>
References: <c09dd91f-c280-85a6-c2a2-d44a0d378bbc@redhat.com>
 <4EB5D96F-F322-45BB-9169-6BF932D413D4@amacapital.net>
From:   Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <931f6e6d-ac17-05f9-0605-ac8f89f40b2b@redhat.com>
Date:   Thu, 9 Apr 2020 17:17:32 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.5.0
MIME-Version: 1.0
In-Reply-To: <4EB5D96F-F322-45BB-9169-6BF932D413D4@amacapital.net>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 09/04/20 17:03, Andy Lutomirski wrote:
>> No, I think we wouldn't use a paravirt #VE at this point, we would
>> use the real thing if available.
>> 
>> It would still be possible to switch from the IST to the main
>> kernel stack before writing 0 to the reentrancy word.
> 
> Almost but not quite. We do this for NMI-from-usermode, and it’s
> ugly. But we can’t do this for NMI-from-kernel or #VE-from-kernel
> because there might not be a kernel stack.  Trying to hack around
> this won’t be pretty.
> 
> Frankly, I think that we shouldn’t even try to report memory failure
> to the guest if it happens with interrupts off. Just kill the guest
> cleanly and keep it simple. Or inject an intentionally unrecoverable
> IST exception.

But it would be nice to use #VE for all host-side page faults, not just
for memory failure.

So the solution would be the same as for NMIs, duplicating the stack
frame and patching the outer handler's stack from the recursive #VE
(https://lwn.net/Articles/484932/).  It's ugly but it's a known ugliness.

Paolo