Re: percpu-2.5.63-bkcurr

[William Lee Irwin III <wli@holomorphy.com>]

On Fri, Feb 28, 2003 at 11:36:55PM -0800, William Lee Irwin III wrote:
> Tentative followup #1 (thanks Zwane!)
>> Use per-cpu rq's in the sched.c to avoid remote cache misses there.
>> It actually means something now.

On Fri, Feb 28, 2003 at 11:40:35PM -0800, William Lee Irwin III wrote:
> Tentative followup #2 -- totally untested, at some point I have to
> figure out how to avoid breaking the compile for non-NUMA-Q with this.

woops


diff -u sched-2.5/arch/i386/mm/discontig.c sched-2.5/arch/i386/mm/discontig.c
--- sched-2.5/arch/i386/mm/discontig.c	Fri Feb 28 23:12:45 2003
+++ sched-2.5/arch/i386/mm/discontig.c	Fri Feb 28 23:42:15 2003
@@ -230,6 +230,7 @@
 	for (nid = 1; nid < numnodes; nid++) {
 		/* calculate the size of the mem_map needed in pages */
 		size = MEM_MAP_SIZE(nid) + PFN_UP(sizeof(pg_data_t))
+			+ PER_NODE_PAGES
 			+ PER_CPU_PAGES*hweight32(early_node_to_cpumask(nid));
 		/* round up to nearest pmd boundary */
 		size = (size + PTRS_PER_PTE - 1) & ~(PTRS_PER_PTE - 1);