Re: [PATCH v2 5/5] x86/pti: Do not enable PTI on fixed Intel processors

[David Woodhouse <dwmw2@infradead.org>]

[-- Attachment #1: Type: text/plain, Size: 1775 bytes --]

On Wed, 2018-01-24 at 18:40 +0000, Alan Cox wrote:
> Nobody has published official statements on Cyrix or AMD 32bit processors
> so we don't know if they are vulnerable to meltdown. One problem I
> suspect is that as with things like Alpha 21264 - the people who knew are
> probably long retired. We do know the Intel ones I listed are OK and the
> Centaur.
> 
> If someone can figure out the Cyrix and AMD cases that would be great.

Well Tom already submitted a patch to turn it off for *all* AMD, 32-bit 
and 64-bit.

> > 
> > By the time the dust settles we might end up with a bunch of different
> > match tables, *one* of which is "does not speculate at all". And the
> > conditions for the different bugs will each use different sets of match
> > tables. For example
> > 
> >  if (!x86_match_cpu(cpu_no_speculation_at_all) &&
> >      !x86_match_cpu(speculation_but_no_meltdown) &&
> >      !cpu_sets_rdcl_no())
> > 	setup_force_cpu_bug(X86_BUG_CPU_MELTDOWN);
> > 
> >  if (!x86_match_cpu(cpu_no_speculation_at_all) &&
> >      !x86_match_cpu(no_branch_target_buffer))
> > 	setup_force_cpu_bug(X86_BUG_SPECTRE_V2);
> There are afaik no x86 processors that speculate and don't have a BTB.
> It's a bit like building a racing car with no gearbox.

Right, "has a BTB and doesn't tag/flush it according to privilege level
and context". Which is the thing that, as lamented, Intel hasn't even
proposed a way to *tell* us that's the case, if/when they finally
manage to fix it up in the next generation *after* the IBRS_ALL hack.

So yeah, X86_BUG_SPECTRE_V2 probably *does* end up being "speculates".

I'll change !Meltdown from ANY,5 to INTEL,5 + CENTAUR,5 as you suggest,
and we can work on Spectre separately.

Thanks.

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5213 bytes --]