[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <472A2523.2030201@zytor.com>
Date: Thu, 01 Nov 2007 12:12:35 -0700
From: "H. Peter Anvin" <hpa@...or.com>
To: Glauber de Oliveira Costa <gcosta@...hat.com>
CC: Andreas Herrmann3 <andreas.herrmann3@....com>,
linux-kernel@...r.kernel.org, mingo@...e.hu, tglx@...utronix.de,
akpm@...ux-foundation.org, travis@....com
Subject: Re: [PATCH] x86: show cpuinfo only for online CPUs
Glauber de Oliveira Costa wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Andreas Herrmann3 escreveu:
>> On Thu, Nov 01, 2007 at 06:35:43PM +0100, Andreas Herrmann3 wrote:
>>> On Thu, Nov 01, 2007 at 03:12:25PM -0200, Glauber de Oliveira Costa wrote:
>>>> So it could even work, but as accident. Unless I'm wrong about it, I'd
>>>> prefer to see an explicit attribution of cpu_index = 0 somewhere for the
>>>> boot cpu.
>>> Hmm, will look at this as well.
>> BTW, isn't it zero initialized anyway?
>> So, no need to explicitely set cpuinfo->cpu_index=0 for the boot cpu.
>
> Well, it should be, but as far as I know it is not exactly a guarantee
> given by all compilers. So it should be safer to do it explicitly, with
> no prejudice. Unless it's really guaranteed. If it is, yeah, no need.
Any uninitialized field in a static section (.data or .bss) is
guaranteed to be initialized to zero. This is guaranteed by the C
standard. In the former case, it is the responsibility of the compiler
and in the latter by the runtime.
The Linux percpu handling creates a .data.percpu section which is then
replicated into each of the CPU data blocks; thus, percpu data counts as
static data for this purpose, and initialization is guaranteed.
-hpa
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists