linux-kernel - Re: [PATCH v2] KVM: SVM: Set/clear SRSO's BP_SPEC

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: <174674531183.1512799.1644836485465401885.b4-ty@google.com>
Date: Thu,  8 May 2025 16:04:13 -0700
From: Sean Christopherson <seanjc@...gle.com>
To: Sean Christopherson <seanjc@...gle.com>, Paolo Bonzini <pbonzini@...hat.com>
Cc: kvm@...r.kernel.org, linux-kernel@...r.kernel.org, 
	Michael Larabel <Michael@...haellarabel.com>, Borislav Petkov <bp@...en8.de>
Subject: Re: [PATCH v2] KVM: SVM: Set/clear SRSO's BP_SPEC_REDUCE on 0 <=> 1
 VM count transitions

On Mon, 05 May 2025 11:03:00 -0700, Sean Christopherson wrote:
> Set the magic BP_SPEC_REDUCE bit to mitigate SRSO when running VMs if and
> only if KVM has at least one active VM.  Leaving the bit set at all times
> unfortunately degrades performance by a wee bit more than expected.
> 
> Use a dedicated spinlock and counter instead of hooking virtualization
> enablement, as changing the behavior of kvm.enable_virt_at_load based on
> SRSO_BP_SPEC_REDUCE is painful, and has its own drawbacks, e.g. could
> result in performance issues for flows that are sensitive to VM creation
> latency.
> 
> [...]

Applied to kvm-x86 fixes.  Assuming -next doesn't explode overnight, I'll get
a pull request sent to Paolo tomorrow.

[1/1] KVM: SVM: Set/clear SRSO's BP_SPEC_REDUCE on 0 <=> 1 VM count transitions
      https://github.com/kvm-x86/linux/commit/e3417ab75ab2

--
https://github.com/kvm-x86/linux/tree/next