On Thursday 29 November 2007, Tony Breeds wrote:
> why not?
In the past I've seen run-time problems because some parts of the
code kmalloc NR_CPUS sized arrays, which can easily grow beyond
the kmalloc limit of 128kb. Have you tried running some tests
on a small machine with the big NR_CPUS?
Hi Tony, :-)
On Thu, 29 Nov 2007 15:16:03 +1100 [EMAIL PROTECTED] (Tony Breeds) wrote:
>
> Signed-off-by: Tony Breeds <[EMAIL PROTECTED]>
NAK ... you need to change the static initialisation of the paca
structures to match ...
--
Cheers,
Stephen Rothwell[EMAIL PROTECTED]
htt
On Thu, 2007-11-29 at 15:23 +1100, Tony Breeds wrote:
> On Thu, Nov 29, 2007 at 03:17:16PM +1100, Michael Ellerman wrote:
> > On Thu, 2007-11-29 at 15:16 +1100, Tony Breeds wrote:
> > > Signed-off-by: Tony Breeds <[EMAIL PROTECTED]>
> > >
> > > ---
> > > why not?
> >
> > How big is say a pseries
On Thu, Nov 29, 2007 at 03:17:16PM +1100, Michael Ellerman wrote:
> On Thu, 2007-11-29 at 15:16 +1100, Tony Breeds wrote:
> > Signed-off-by: Tony Breeds <[EMAIL PROTECTED]>
> >
> > ---
> > why not?
>
> How big is say a pseries_defconfig with NR_CPUS = 1024 ?
This is a ppc64_defconfig, with a cou
On Thu, 2007-11-29 at 15:16 +1100, Tony Breeds wrote:
> Signed-off-by: Tony Breeds <[EMAIL PROTECTED]>
>
> ---
> why not?
How big is say a pseries_defconfig with NR_CPUS = 1024 ?
cheers
--
Michael Ellerman
OzLabs, IBM Australia Development Lab
wwweb: http://michael.ellerman.id.au
phone: +61 2
