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;
From:   Vitaly Kuznetsov <vkuznets@redhat.com>
To:     Sean Christopherson <sean.j.christopherson@intel.com>,
        Paolo Bonzini <pbonzini@redhat.com>
Cc:     kvm@vger.kernel.org, Wanpeng Li <wanpengli@tencent.com>,
        Jim Mattson <jmattson@google.com>, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] KVM: nVMX: Inject #GP when nested_vmx_get_vmptr() fails to read guest memory
In-Reply-To: <20200604145357.GA30223@linux.intel.com>
References: <20200604143158.484651-1-vkuznets@redhat.com> <da7acd6f-204d-70e2-52aa-915a4d9163ef@redhat.com> <20200604145357.GA30223@linux.intel.com>
Date:   Thu, 04 Jun 2020 17:33:25 +0200
Message-ID: <87k10meth6.fsf@vitty.brq.redhat.com>
MIME-Version: 1.0
Content-Type: text/plain
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Sean Christopherson <sean.j.christopherson@intel.com> writes:

> On Thu, Jun 04, 2020 at 04:40:52PM +0200, Paolo Bonzini wrote:
>> On 04/06/20 16:31, Vitaly Kuznetsov wrote:
>
> ...
>
>> > KVM could've handled the request correctly by going to userspace and
>> > performing I/O but there doesn't seem to be a good need for such requests
>> > in the first place. Sane guests should not call VMXON/VMPTRLD/VMCLEAR with
>> > anything but normal memory. Just inject #GP to find insane ones.
>> > 

...

>> 
>> looks good but we need to do the same in handle_vmread, handle_vmwrite,
>> handle_invept and handle_invvpid.  Which probably means adding something
>> like nested_inject_emulation_fault to commonize the inner "if".
>
> Can we just kill the guest already instead of throwing more hacks at this
> and hoping something sticks?  We already have one in
> kvm_write_guest_virt_system...
>
>   commit 541ab2aeb28251bf7135c7961f3a6080eebcc705
>   Author: Fuqian Huang <huangfq.daxian@gmail.com>
>   Date:   Thu Sep 12 12:18:17 2019 +0800
>
>     KVM: x86: work around leak of uninitialized stack contents
>

Oh I see...

[...]

Let's get back to 'vm_bugged' idea then? 

https://lore.kernel.org/kvm/87muadnn1t.fsf@vitty.brq.redhat.com/

-- 
Vitaly