Re: [RFC PATCH 27/34] x86/bugs: Add attack vector controls for spectre_v1

[Borislav Petkov]

On Thu, Sep 12, 2024 at 07:57:50PM +0000, Kaplan, David wrote:
> > There's also added complexity from having 'enum vulnerabilities' which
> > basically duplicate the X86_BUG_* space.  If the infrastructure was, for
> > instance, built around X86_BUG bits, it might have enabled this patch to be
> > something like:
> >
> > -       if (!boot_cpu_has_bug(X86_BUG_SPECTRE_V1) ||
> > -           cpu_mitigations_off())
> > +       if (!should_mitigate_vuln(X86_BUG_SPECTRE_V1))
> >                 spectre_v1_mitigation = SPECTRE_V1_MITIGATION_NONE;
> 
> That's a reasonable idea.  One issue I see is that there is no separation in
> the X86_BUG* space for spectre_v2 vs spectre_v2_user, but they do have
> separate mitigations.  But I think that is the only missing one, so maybe it
> just makes sense to add a X86_BUG bit for that?

I think we should do that. That's less complexity in the mitigations area and
those are always welcome.

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette