Messages in this thread Patch in this message | | | Date | Wed, 30 Jun 2010 10:08:29 -0500 | From | Jack Steiner <> | Subject | Re: [PATCH V2] x86: irq vector assignment |
| |
Try to assign irq vectors 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@sgi.com>
---
V2 - Fix error for builds without CONFIG_NUMA.
On Thu, Jun 24, 2010 at 01:45:40PM -0700, Yinghai Lu wrote: > how about update apic::target_cpus() to take node?
I may be overlooking something but I don't see the benefit of passing node to target_cpus(). It looks like only 1 place that calls target_cpus() has a valid "node". The other places would pass -1. In addition, I'm not sure how some of the other subarches (ex., es7000_32) would use the node number. Many would just ignore it.
--- arch/x86/kernel/apic/io_apic.c | 5 +++++ 1 file changed, 5 insertions(+)
Index: linux/arch/x86/kernel/apic/io_apic.c =================================================================== --- linux.orig/arch/x86/kernel/apic/io_apic.c 2010-06-25 08:11:07.186867196 -0500 +++ linux/arch/x86/kernel/apic/io_apic.c 2010-06-25 08:41:51.790274150 -0500 @@ -3265,6 +3265,11 @@ unsigned int create_irq_nr(unsigned int desc_new = move_irq_desc(desc_new, node); cfg_new = desc_new->chip_data; +#ifdef CONFIG_NUMA + if (node >= 0 && __assign_irq_vector(new, cfg_new, node_to_cpumask_map[node]) == 0) + irq = new; + else +#endif if (__assign_irq_vector(new, cfg_new, apic->target_cpus()) == 0) irq = new; break;
| |