Re: [PATCH v2 2/2] x86: Increase MAXSMP CPU count to 8192

[Josh Boyer <jwboyer@redhat.com>]

On Wed, Nov 06, 2013 at 06:56:34AM +0100, Ingo Molnar wrote:
> 
> * Josh Boyer <jwboyer@redhat.com> wrote:
> 
> > The MAXSMP option is intended to enable silly large numbers of CPUs for
> > testing purposes.  The current value of 4096 isn't very silly any longer
> > as there are actual machines that approach 6096 CPUs when taking HT into
> > account.  Increase the value to 8192 to account for this and short term
> > future increases.
> > 
> > Signed-off-by: Josh Boyer <jwboyer@fedoraproject.org>
> > ---
> >  arch/x86/Kconfig | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> > index f03e428..09f05cd 100644
> > --- a/arch/x86/Kconfig
> > +++ b/arch/x86/Kconfig
> > @@ -828,7 +828,7 @@ config NR_CPUS
> >  	range 2 512 if SMP && !MAXSMP && !CPUMASK_OFFSTACK
> >  	range 2 4096 if SMP && !MAXSMP && CPUMASK_OFFSTACK && X86_64
> >  	default "1" if !SMP
> > -	default "4096" if MAXSMP
> > +	default "8192" if MAXSMP
> 
> So I changed the patch to up the regular range to 8192 as well: it's more 
> consistent if MAXSMP tracks the normal range, as long as the defaults are 
> sane. See the commit below.

OK.  That looks fine to me.

> Note that 'normal users' won't ever see any max CPU count beyond the 
> current 512 (which is already kind of crazy for normal use), because 
> CPUMASK_OFFSTACK is turned off on all but MAXSMP configs (unless the
> user explicitly enables the option).
> 
> So it's not really possible to accidentally wander into that option and 
> set it to an insane value - it takes some effort.
> 
> [ I've Cc:-ed Linus, who had a strong opinion about these max-cpu limits
>   in the past - I hope the changes are fine in this form. ]

Me too.

josh

> 
> Thanks,
> 
> 	Ingo
> 
> =====================>
> From 07bce9fd69e8409740e63db0ab0755b84ee61a0c Mon Sep 17 00:00:00 2001
> From: Josh Boyer <jwboyer@redhat.com>
> Date: Tue, 5 Nov 2013 09:38:16 -0500
> Subject: [PATCH] x86/cpu: Increase max CPU count to 8192
> 
> The MAXSMP option is intended to enable silly large numbers of
> CPUs for testing purposes.  The current value of 4096 isn't very
> silly any longer as there are actual SGI machines that approach
> 6096 CPUs when taking HT into account.
> 
> Increase the value to a nice round 8192 to account for this and
> allow for short term future increases.
> 
> Signed-off-by: Josh Boyer <jwboyer@fedoraproject.org>
> Cc: prarit@redhat.com
> Cc: Russ Anderson <rja@sgi.com>
> Cc: Linus Torvalds <torvalds@linux-foundation.org>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Link: http://lkml.kernel.org/r/20131105143816.GK9944@hansolo.jdub.homelinux.org
> [ Tweaked it so that MAXSMP simply sets the maximum of the normal range. ]
> Signed-off-by: Ingo Molnar <mingo@kernel.org>
> ---
>  arch/x86/Kconfig | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> index f03e428..805469a 100644
> --- a/arch/x86/Kconfig
> +++ b/arch/x86/Kconfig
> @@ -826,9 +826,9 @@ config NR_CPUS
>  	int "Maximum number of CPUs" if SMP && !MAXSMP
>  	range 2 8 if SMP && X86_32 && !X86_BIGSMP
>  	range 2 512 if SMP && !MAXSMP && !CPUMASK_OFFSTACK
> -	range 2 4096 if SMP && !MAXSMP && CPUMASK_OFFSTACK && X86_64
> +	range 2 8192 if SMP && !MAXSMP && CPUMASK_OFFSTACK && X86_64
>  	default "1" if !SMP
> -	default "4096" if MAXSMP
> +	default "8192" if MAXSMP
>  	default "32" if SMP && (X86_NUMAQ || X86_SUMMIT || X86_BIGSMP || X86_ES7000)
>  	default "8" if SMP
>  	---help---