KAMEZAWA Hiroyuki wrote: > How about this one ? > tested on x86-64 + mmotm-Nov10, works well. > (test on other arch is welcome.) > > -Kame > == > As Jan Blunck <jblunck@xxxxxxx> pointed out, allocating > per-cpu stat for memcg to the size of NR_CPUS is not good. > > This patch changes mem_cgroup's cpustat allocation not based > on NR_CPUS but based on nr_cpu_ids. > > From: Jan Blunck <jblunck@xxxxxxx> > Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx> > > --- > mm/memcontrol.c | 34 ++++++++++++++++++---------------- > 1 file changed, 18 insertions(+), 16 deletions(-) > > Index: mmotm-2.6.28-Nov10/mm/memcontrol.c > =================================================================== > --- mmotm-2.6.28-Nov10.orig/mm/memcontrol.c > +++ mmotm-2.6.28-Nov10/mm/memcontrol.c > @@ -60,7 +60,7 @@ struct mem_cgroup_stat_cpu { > } ____cacheline_aligned_in_smp; > > struct mem_cgroup_stat { > - struct mem_cgroup_stat_cpu cpustat[NR_CPUS]; > + struct mem_cgroup_stat_cpu cpustat[0]; > }; > > /* > @@ -129,11 +129,10 @@ struct mem_cgroup { > > int prev_priority; /* for recording reclaim priority */ > /* > - * statistics. > + * statistics. This must be placed at the end of memcg. > */ > struct mem_cgroup_stat stat; > }; > -static struct mem_cgroup init_mem_cgroup; > > enum charge_type { > MEM_CGROUP_CHARGE_TYPE_CACHE = 0, > @@ -1292,42 +1291,45 @@ static void free_mem_cgroup_per_zone_inf > kfree(mem->info.nodeinfo[node]); > } > > +static int mem_cgroup_size(void) inline this function? Other than that, I think the cont->parent check for freeing has already been spotted and pointed out -- Balbir _______________________________________________ Containers mailing list Containers@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linux-foundation.org/mailman/listinfo/containers