diff options
author | Yinghai Lu <yinghai@kernel.org> | 2009-11-21 03:23:37 -0500 |
---|---|---|
committer | Ingo Molnar <mingo@elte.hu> | 2009-11-23 04:06:24 -0500 |
commit | d9c2d5ac6af87b4491bff107113aaf16f6c2b2d9 (patch) | |
tree | 5cb605d6396a0572dabf2b6152bb26fea425e376 /arch/x86/mm | |
parent | 021428ad1418cf3c386a1a0157140c3ea29b17ef (diff) |
x86, numa: Use near(er) online node instead of roundrobin for NUMA
CPU to node mapping is set via the following sequence:
1. numa_init_array(): Set up roundrobin from cpu to online node
2. init_cpu_to_node(): Set that according to apicid_to_node[]
according to srat only handle the node that
is online, and leave other cpu on node
without ram (aka not online) to still
roundrobin.
3. later call srat_detect_node for Intel/AMD, will use first_online
node or nearby node.
Problem is that setup_per_cpu_areas() is not called between 2 and 3,
the per_cpu for cpu on node with ram is on different node, and could
put that on node with two hops away.
So try to optimize this and add find_near_online_node() and call
init_cpu_to_node().
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: David Rientjes <rientjes@google.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
LKML-Reference: <4B07A739.3030104@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Diffstat (limited to 'arch/x86/mm')
-rw-r--r-- | arch/x86/mm/numa_64.c | 21 |
1 files changed, 20 insertions, 1 deletions
diff --git a/arch/x86/mm/numa_64.c b/arch/x86/mm/numa_64.c index 3acd870d316a..83bbc70d11bb 100644 --- a/arch/x86/mm/numa_64.c +++ b/arch/x86/mm/numa_64.c | |||
@@ -764,6 +764,25 @@ static __init int numa_setup(char *opt) | |||
764 | early_param("numa", numa_setup); | 764 | early_param("numa", numa_setup); |
765 | 765 | ||
766 | #ifdef CONFIG_NUMA | 766 | #ifdef CONFIG_NUMA |
767 | |||
768 | static __init int find_near_online_node(int node) | ||
769 | { | ||
770 | int n, val; | ||
771 | int min_val = INT_MAX; | ||
772 | int best_node = -1; | ||
773 | |||
774 | for_each_online_node(n) { | ||
775 | val = node_distance(node, n); | ||
776 | |||
777 | if (val < min_val) { | ||
778 | min_val = val; | ||
779 | best_node = n; | ||
780 | } | ||
781 | } | ||
782 | |||
783 | return best_node; | ||
784 | } | ||
785 | |||
767 | /* | 786 | /* |
768 | * Setup early cpu_to_node. | 787 | * Setup early cpu_to_node. |
769 | * | 788 | * |
@@ -795,7 +814,7 @@ void __init init_cpu_to_node(void) | |||
795 | if (node == NUMA_NO_NODE) | 814 | if (node == NUMA_NO_NODE) |
796 | continue; | 815 | continue; |
797 | if (!node_online(node)) | 816 | if (!node_online(node)) |
798 | continue; | 817 | node = find_near_online_node(node); |
799 | numa_set_node(cpu, node); | 818 | numa_set_node(cpu, node); |
800 | } | 819 | } |
801 | } | 820 | } |