From: Yinghai Lu on
On Thu, Jun 24, 2010 at 12:34 PM, Jack Steiner <steiner(a)sgi.com> wrote:
> Try to assign irqs to cpus on the correct node & fall back to global
> assignment only if node-local fails. This reduces the chances of
> using all of the interrupt vectors of a single cpu.
>
> Signed-off-by: Jack Steiner <steiner(a)sgi.com>
>
>
> ---
> �arch/x86/kernel/apic/io_apic.c | � �4 +++-
> �1 file changed, 3 insertions(+), 1 deletion(-)
>
> Index: linux/arch/x86/kernel/apic/io_apic.c
> ===================================================================
> --- linux.orig/arch/x86/kernel/apic/io_apic.c � 2010-06-24 12:58:11.658933610 -0500
> +++ linux/arch/x86/kernel/apic/io_apic.c � � � �2010-06-24 13:56:26.734380290 -0500
> @@ -3259,7 +3259,9 @@ unsigned int create_irq_nr(unsigned int
> � � � � � � � �desc_new = move_irq_desc(desc_new, node);
> � � � � � � � �cfg_new = desc_new->chip_data;
>
> - � � � � � � � if (__assign_irq_vector(new, cfg_new, apic->target_cpus()) == 0)
> + � � � � � � � if (node >= 0 && __assign_irq_vector(new, cfg_new, node_to_cpumask_map[node]) == 0)
> + � � � � � � � � � � � irq = new;
> + � � � � � � � else if (__assign_irq_vector(new, cfg_new, apic->target_cpus()) == 0)
> � � � � � � � � � � � �irq = new;
> � � � � � � � �break;
> � � � �}

how about update apic::target_cpus() to take node?

YH
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/