linux-kernel - Re: [PATCH 1/2] x86/bhi: Add BHB clearing for CPUs with larger branch history

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

Message-ID: <20250925204937.d4pabsyg3jxnw332@desk>
Date: Thu, 25 Sep 2025 13:49:37 -0700
From: Pawan Gupta <pawan.kumar.gupta@...ux.intel.com>
To: Jim Mattson <jmattson@...gle.com>
Cc: x86@...nel.org, "H. Peter Anvin" <hpa@...or.com>,
	Josh Poimboeuf <jpoimboe@...nel.org>,
	David Kaplan <david.kaplan@....com>,
	Sean Christopherson <seanjc@...gle.com>,
	Paolo Bonzini <pbonzini@...hat.com>, linux-kernel@...r.kernel.org,
	kvm@...r.kernel.org, Asit Mallick <asit.k.mallick@...el.com>,
	Tao Zhang <tao1.zhang@...el.com>
Subject: Re: [PATCH 1/2] x86/bhi: Add BHB clearing for CPUs with larger
 branch history

On Thu, Sep 25, 2025 at 10:54:48AM -0700, Jim Mattson wrote:
> On Wed, Sep 24, 2025 at 8:09 PM Pawan Gupta
> <pawan.kumar.gupta@...ux.intel.com> wrote:
> >
> > Add a version of clear_bhb_loop() that works on CPUs with larger branch
> > history table such as Alder Lake and newer. This could serve as a cheaper
> > alternative to IBPB mitigation for VMSCAPE.
> 
> Yay!
> 
> Can we also use this longer loop as a BHI mitigation on (virtual)
> processors with larger branch history tables that don't support
> BHI_DIS_S? Today, we just use the short BHB clearing loop and call it
> good.

I believe you are referring to guests that don't enumerate BHI_DIS_S, but
are running on a host that supports BHI_DIS_S. In that case the longer loop
would work to mitigate BHI.

You probably know it already, the longer loop is not an optimal mitigation
compared to BHI_DIS_S. And on CPUs that don't support BHI_DIS_S, short loop
is sufficient.

I think you are talking about some special use cases like a guest migrating
from a CPU that didn't support BHI_DIS_S to a CPU that does. Using the long
loop in that case would be an option.