[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAMj1kXFeRiFRsNwTE0ggFkiO8LG6maWmFJaY=jq8xk_JvFo=pQ@mail.gmail.com>
Date: Thu, 15 May 2025 11:57:22 +0100
From: Ard Biesheuvel <ardb@...nel.org>
To: Borislav Petkov <bp@...en8.de>
Cc: Ard Biesheuvel <ardb+git@...gle.com>, linux-kernel@...r.kernel.org, x86@...nel.org,
Ingo Molnar <mingo@...nel.org>, Linus Torvalds <torvalds@...ux-foundation.org>,
Brian Gerst <brgerst@...il.com>
Subject: Re: [PATCH v3 1/7] x86/cpu: Use a new feature flag for 5 level paging
On Thu, 15 May 2025 at 11:39, Borislav Petkov <bp@...en8.de> wrote:
>
> On Thu, May 15, 2025 at 11:17:47AM +0100, Ard Biesheuvel wrote:
> > I'd lean towards not lying about whether the CPU is la57 capable in
> > /proc/cpuinfo if we don't have to - this flag directly reflects the
> > CPUID leaf.
>
> We have a whole documentation about that whole "not lying" thing: short
> version - use kcpuid or some other cpuid tool:
>
> Documentation/arch/x86/cpuinfo.rst
>
> > This is arguably the important one: as long as "la57" does not change
> > meaning, we should be fine from compatibility pov.
>
> Yes, those flags in /proc/cpuinfo should show that something is enabled and in
> use. And it makes sense here. And you can't change that string anyway anymore.
>
> Whether the hw is capable is not important for userspace and there are a lot
> better means for it.
>
Fair enough.
Powered by blists - more mailing lists