linux-kernel - Re: [PATCH] KVM: x86: Inject #UD on "unsupported" hypercall if patching fails

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <87wnk8v43z.fsf@redhat.com>
Date:   Mon, 13 Dec 2021 17:53:04 +0100
From:   Vitaly Kuznetsov <vkuznets@...hat.com>
To:     Sean Christopherson <seanjc@...gle.com>,
        Paolo Bonzini <pbonzini@...hat.com>
Cc:     Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Joerg Roedel <joro@...tes.org>, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Hou Wenlong <houwenlong93@...ux.alibaba.com>
Subject: Re: [PATCH] KVM: x86: Inject #UD on "unsupported" hypercall if
 patching fails

Sean Christopherson <seanjc@...gle.com> writes:

> Ideally, KVM wouldn't patch at all; it's the guest's responsibility to
> identify and use the correct hypercall instruction (VMCALL vs. VMMCALL).
> Sadly, older Linux kernels prior to commit c1118b3602c2 ("x86: kvm: use
> alternatives for VMCALL vs. VMMCALL if kernel text is read-only") do the
> wrong thing and blindly use VMCALL, i.e. removing the patching would
> break running VMs with older kernels.
>

FWIW, we also use hypercall patching for Hyper-V emulation (when
HV_X64_MSR_HYPERCALL is written) and this complies with TLFS, we can't
get rid of this. It's a different 'patching' though...

-- 
Vitaly