On Thu, 2010-05-27 at 16:34 -0700, Luck, Tony wrote:
> Hi Linus,
>
> please pull from:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6.git release
>
> This will update the files shown below.
>
> Thanks!
>
> -Tony
>
> arch/ia64/kernel/smpboot.c | 4 ++++
> arch/ia64/mm/numa.c | 17 +++++++++++++++++
> 2 files changed, 21 insertions(+), 0 deletions(-)
>
> Tony Luck (1):
> [IA64] Fix build breakage
>
> Andrew: These changes appear to have gone to Linus without passing
> through linux-next (where I'd have caught the breakage earlier).
> Did they slip through a crack?
>
> ---
>
> commit 4ec37de89d8c758ee8115e0e64b3f994910789ee
> Author: Tony Luck <tony.luck@xxxxxxxxx>
> Date: Thu May 27 15:35:13 2010 -0700
>
> [IA64] Fix build breakage
>
> In commit 0ac0c0d0f837c499afd02a802f9cf52d3027fa3b
> cpusets: randomize node rotor used in cpuset_mem_spread_node()
>
> Jack Steiner fixed a problem with too many small tasks being
> assigned to node 0. Copy his code to ia64 to avoid build error.
>
> arch/ia64/kernel/smpboot.c:641: error: ‘cpu_to_node_map’ undeclared (first use in this function)
>
> In commit 3bccd996276b108c138e8176793a26ecef54d573
> numa: ia64: use generic percpu var numa_node_id() implementation
>
> Lee Schermerhorn added some set_numa_node() calls - but these
> only work on CONFIG_NUMA=y configurations. Surround the calls
> with #ifdef CONFIG_NUMA
Tony:
Can I get your .config? I thought I tested ia64 !NUMA, but I see now
that when I edit the .config and 'make silentoldconfig', IA64_GENERIC
just selects NUMA again. Even 'make allnoconfig' on ia64 generates
a .config with NUMA=y. I tried changing IA64_GENERIC to IA64_HP_ZX1
that does not select NUMA, but the build fails elsewhere because
NODE_DATA is apparently not defined for !NUMA. So, I'd like to get
your .config to duplicate the build error that you're fixing here.
I had intended that asm-generic/topology.h define the stubs for the
missing calls when !NUMA as it does for several other functions and
avoid the ifdefs in the code. Apparently I just dreamed that I did
that...
I'll send an update patch when I get it figured out.
Thanks,
Lee
>
> Signed-off-by: Tony Luck <tony.luck@xxxxxxxxx>
>
> diff --git a/arch/ia64/kernel/smpboot.c b/arch/ia64/kernel/smpboot.c
> index 518e876..6a1380e 100644
> --- a/arch/ia64/kernel/smpboot.c
> +++ b/arch/ia64/kernel/smpboot.c
> @@ -390,11 +390,13 @@ smp_callin (void)
>
> fix_b0_for_bsp();
>
> +#ifdef CONFIG_NUMA
> /*
> * numa_node_id() works after this.
> */
> set_numa_node(cpu_to_node_map[cpuid]);
> set_numa_mem(local_memory_node(cpu_to_node_map[cpuid]));
> +#endif
>
> ipi_call_lock_irq();
> spin_lock(&vector_lock);
> @@ -638,7 +640,9 @@ void __devinit smp_prepare_boot_cpu(void)
> {
> cpu_set(smp_processor_id(), cpu_online_map);
> cpu_set(smp_processor_id(), cpu_callin_map);
> +#ifdef CONFIG_NUMA
> set_numa_node(cpu_to_node_map[smp_processor_id()]);
> +#endif
> per_cpu(cpu_state, smp_processor_id()) = CPU_ONLINE;
> paravirt_post_smp_prepare_boot_cpu();
> }
> diff --git a/arch/ia64/mm/numa.c b/arch/ia64/mm/numa.c
> index 3efea7d..2437718 100644
> --- a/arch/ia64/mm/numa.c
> +++ b/arch/ia64/mm/numa.c
> @@ -17,6 +17,7 @@
> #include <linux/init.h>
> #include <linux/bootmem.h>
> #include <linux/module.h>
> +#include <linux/random.h>
> #include <asm/mmzone.h>
> #include <asm/numa.h>
>
> @@ -50,6 +51,22 @@ paddr_to_nid(unsigned long paddr)
> return (i < num_node_memblks) ? node_memblk[i].nid : (num_node_memblks ? -1 : 0);
> }
>
> +/*
> + * Return the bit number of a random bit set in the nodemask.
> + * (returns -1 if nodemask is empty)
> + */
> +int __node_random(const nodemask_t *maskp)
> +{
> + int w, bit = -1;
> +
> + w = nodes_weight(*maskp);
> + if (w)
> + bit = bitmap_ord_to_pos(maskp->bits,
> + get_random_int() % w, MAX_NUMNODES);
> + return bit;
> +}
> +EXPORT_SYMBOL(__node_random);
> +
> #if defined(CONFIG_SPARSEMEM) && defined(CONFIG_NUMA)
> /*
> * Because of holes evaluate on section limits.
--
To unsubscribe from this list: send the line "unsubscribe linux-ia64" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
|