linux-kernel - Re: [PATCH v2 1/5] x86/topology: Export helper to get CPU number from APIC ID

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: <20240712142821.GA10846@yaz-khff2.amd.com>
Date: Fri, 12 Jul 2024 10:28:21 -0400
From: Yazen Ghannam <yazen.ghannam@....com>
To: Borislav Petkov <bp@...en8.de>
Cc: Thomas Gleixner <tglx@...utronix.de>, linux-edac@...r.kernel.org,
	linux-kernel@...r.kernel.org, tony.luck@...el.com, x86@...nel.org,
	avadhut.naik@....com, john.allen@....com
Subject: Re: [PATCH v2 1/5] x86/topology: Export helper to get CPU number
 from APIC ID

On Mon, Jul 01, 2024 at 09:07:04PM +0200, Borislav Petkov wrote:
> On Mon, Jul 01, 2024 at 01:51:42PM -0400, Yazen Ghannam wrote:
> > X86_LOCAL_APIC depends on the logical OR of a bunch of options. So it
> > depends on "any one" of the options to be enabled. But it doesn't need
> > all of them.
> 
> Yes, I can see that.
> 
> How does any of that info answer your initial question?
> 
> IOW, if you don't have LAPIC support in the kernel, what CPU number should we
> return for any APIC ID serving as input, and why?
>

I still think it should return an error code, because theoretically
LAPIC can be disabled and SMP can be enabled.

But I spent some time trying to see if this would work in practice, and
it looks like you can't disable X86_LOCAL_APIC without hitting a bunch
of build errors on x86_64. It seems like a lot of common APIC and SMP
code implicitly depends on X86_LOCAL_APIC. This was true even for a tiny
config. However, this worked for an i386 build (with SMP=n).

I think the most practical option is to keep the local search routine in
mce/apei. I don't think all the additional complexity is worth it for a
simple for-loop.

Regarding the X86_LOCAL_APIC=n build issues, should these be
investigated?

Thanks,
Yazen