Messages in this thread Patch in this message | | | From | "Siddha, Suresh B" <> | Subject | Re: [PATCH] consolidate sched domains | Date | Fri, 23 Jul 2004 14:50:46 -0700 |
| |
On Friday 23 July 2004 08:30, Dimitri Sivanich wrote: > On Fri, Jul 23, 2004 at 01:18:30PM +1000, Nick Piggin wrote: > > The attached patch is against 2.6.8-rc1-mm1. Tested on SMP, UP and SMP+HT > > here and it seems to be OK. > > > > I have included the cpu_sibling_map for ppc64, although Anton said he did > > have an implementation floating around which he would probably prefer, > > but I'll let him deal with that. > > Do other architectures need to define their own cpu_sibling_maps, or am I > missing something that would define that for IA64 and others?
Nick means, all the architectures which use CONFIG_SCHED_SMT needs to define cpu_sibling_map.
Nick, aren't you missing the attached fix in your patch?
thanks, suresh --- linux-2.6.8-rc1/kernel/sched.c~ 2004-07-23 13:19:48.000000000 -0700 +++ linux-2.6.8-rc1/kernel/sched.c 2004-07-23 13:34:49.000000000 -0700 @@ -3845,6 +3845,8 @@ sd->groups->cpu_power = power; #ifdef CONFIG_NUMA + if (i != first_cpu(sd->groups->cpumask)) + continue; sd = &per_cpu(node_domains, i); sd->groups->cpu_power += power; #endif | |