linux-kernel - Re: [PATCH] Revert "KVM: nVMX: Do not expose MPX VMX controls when guest MPX disabled"

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: <YtrdJWmk/2RkcQi7@google.com>
Date:   Fri, 22 Jul 2022 17:23:49 +0000
From:   Sean Christopherson <seanjc@...gle.com>
To:     Paolo Bonzini <pbonzini@...hat.com>
Cc:     linux-kernel@...r.kernel.org, kvm@...r.kernel.org,
        oliver.upton@...ux.dev
Subject: Re: [PATCH] Revert "KVM: nVMX: Do not expose MPX VMX controls when
 guest MPX disabled"

On Fri, Jul 22, 2022, Paolo Bonzini wrote:
> On 7/22/22 17:27, Sean Christopherson wrote:
> > > So revert it, at the potential cost
> > > of breaking L1s with a 6 year old kernel.
> > I would further qualify this with "breaking L1s with an_unpatched_  6 year old
> > kernel".  That fix was tagged for stable and made it way to at least the 4.9 and
> > 4.4 LTS releases.
> > 
> 
> Well, there _are_ people that use very old kernels and keep them up-to-date
> with fixes for only critical CVEs (for example by, ehm, paying my employer
> to do so).

Heh, I'm sure that's a winning strategy.

> But still it's way way unlikely for them to be used as L1 in a nested setup,
> whether on their own hardware or in the cloud.
>
> I pushed everything to kvm/queue, but depending on what you post it may be
> deferred to 5.21.

Can you drop the PERF_GLOBAL_CTRL revert?  I figured out how to achieve what you
intended, but in a more robust (and IMO more logical) manner.

If you don't drop it before I concoct the series, I'll just include a throwaway
patch to revert it.