linux-kernel - Re: [RFC PATCH v1 5/5] KVM: Remove kvm

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: <aMs986QHpDqk-MSR@google.com>
Date: Wed, 17 Sep 2025 16:02:11 -0700
From: Sean Christopherson <seanjc@...gle.com>
To: Xin Li <xin@...or.com>
Cc: linux-kernel@...r.kernel.org, kvm@...r.kernel.org, 
	linux-pm@...r.kernel.org, pbonzini@...hat.com, tglx@...utronix.de, 
	mingo@...hat.com, bp@...en8.de, dave.hansen@...ux.intel.com, x86@...nel.org, 
	hpa@...or.com, rafael@...nel.org, pavel@...nel.org, brgerst@...il.com, 
	david.kaplan@....com, peterz@...radead.org, andrew.cooper3@...rix.com, 
	kprateek.nayak@....com, arjan@...ux.intel.com, chao.gao@...el.com, 
	rick.p.edgecombe@...el.com, dan.j.williams@...el.com
Subject: Re: [RFC PATCH v1 5/5] KVM: Remove kvm_rebooting and its references

On Wed, Sep 17, 2025, Xin Li wrote:
> On 9/16/2025 10:56 AM, Sean Christopherson wrote:
> > On Tue, Sep 09, 2025, Xin Li (Intel) wrote:
> > > Drop kvm_rebooting and all related uses.  Virtualization is now disabled
> > > immediately before a CPU shuts down, eliminating any chance of executing
> > > virtualization instructions during reboot.
> > 
> > Wrong.  KVM clears EFER.SVME in reponse to kvm_shutdown(), and thus can trip
> > #UDs on e.g. VMRUN.
> > 
> 
> This patch assumes that AMD SVM enable/disable has been moved to the CPU

Ah, that wasn't exactly obvious.

> startup and shutdown routines.  Accordingly, kvm_shutdown() no longer clears
> EFER.SVME, and the patch demonstrates the resulting simplification from
> removing kvm_rebooting.  However, as noted in the cover letter, no actual
> modifications were made to AMD SVM.

Heh, yeah, that's what's wrong.