Re: [GIT PULL] parisc fixes for 2.6.39-rc5

Re: [GIT PULL] parisc fixes for 2.6.39-rc5

[Geert Uytterhoeven]

On Wed, Apr 27, 2011 at 22:18, James Bottomley
<James.Bottomley@hansenpartnership.com> wrote:
> These two fixes contain the agreed fix for our slub panic (agreed with
> the mm folks that we'll define SLUB broken in the !NUMA && DISCONTIGMEM
> case, which should fix m86k as well) and another buglet that turned up
> examining the parisc discontigmem code.

Do we still need to mark SLUB broken?
On m68k, the issue seems to have been fixed by setting the N_NORMAL_MEMORY flag
(pull request scheduled after one more linux-next build cycle).

Or do you plan to unmark it broken once every affected arch sets the
N_NORMAL_MEMORY flag?

> The patches are here:
>
> master.kernel.org:/pub/scm/linux/kernel/git/jejb/parisc-2.6.git#fixes
>
> The short changelog is:
>
> David Rientjes (1):
>      [PARISC] set memory ranges in N_NORMAL_MEMORY when onlined
>
> James Bottomley (1):
>      [PARISC] slub: fix panic with DISCONTIGMEM
>
> The diffstat is:
>
>  arch/parisc/mm/init.c |    4 +++-
>  init/Kconfig          |    1 +
>  2 files changed, 4 insertions(+), 1 deletions(-)
>
> And the full diff are below.
>
> James
>
> ---
>
> diff --git a/arch/parisc/mm/init.c b/arch/parisc/mm/init.c
> index b7ed8d7..b1d1262 100644
> --- a/arch/parisc/mm/init.c
> +++ b/arch/parisc/mm/init.c
> @@ -266,8 +266,10 @@ static void __init setup_bootmem(void)
>        }
>        memset(pfnnid_map, 0xff, sizeof(pfnnid_map));
>
> -       for (i = 0; i < npmem_ranges; i++)
> +       for (i = 0; i < npmem_ranges; i++) {
> +               node_set_state(i, N_NORMAL_MEMORY);
>                node_set_online(i);
> +       }
>  #endif
>
>        /*
> diff --git a/init/Kconfig b/init/Kconfig
> index 56240e7..a7ad8fb 100644
> --- a/init/Kconfig
> +++ b/init/Kconfig
> @@ -1226,6 +1226,7 @@ config SLAB
>          per cpu and per node queues.
>
>  config SLUB
> +       depends on BROKEN || NUMA || !DISCONTIGMEM
>        bool "SLUB (Unqueued Allocator)"
>        help
>           SLUB is a slab allocator that minimizes cache line usage

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [GIT PULL] parisc fixes for 2.6.39-rc5

[James Bottomley]

On Wed, 2011-04-27 at 22:36 +0200, Geert Uytterhoeven wrote:
> On Wed, Apr 27, 2011 at 22:18, James Bottomley
> <James.Bottomley@hansenpartnership.com> wrote:
> > These two fixes contain the agreed fix for our slub panic (agreed with
> > the mm folks that we'll define SLUB broken in the !NUMA && DISCONTIGMEM
> > case, which should fix m86k as well) and another buglet that turned up
> > examining the parisc discontigmem code.
> 
> Do we still need to mark SLUB broken?
> On m68k, the issue seems to have been fixed by setting the N_NORMAL_MEMORY flag
> (pull request scheduled after one more linux-next build cycle).
> 
> Or do you plan to unmark it broken once every affected arch sets the
> N_NORMAL_MEMORY flag?

Yes ... simply doing N_NORMAL_MEMORY doesn't fix parisc to not oops.

James