sched: Fix potential NULL derference of doms_cur

If CONFIG_CPUMASK_OFFSTACK is enabled but doms_cur alloc failed in
arch_init_sched_domains(), doms_cur will move back to
fallback_doms. But this time, fallback_doms has not been
initialized yet.

Signed-off-by: Yong Zhang <yong.zhang0@gmail.com>
Cc: a.p.zijlstra@chello.nl
LKML-Reference: <1252930816-7672-1-git-send-email-yong.zhang0@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
This commit is contained in:
Yong Zhang 2009-09-14 20:20:16 +08:00 committed by Ingo Molnar
parent 89f19f04dc
commit cb5fd13f11

View file

@ -9165,6 +9165,7 @@ void __init sched_init_smp(void)
cpumask_var_t non_isolated_cpus;
alloc_cpumask_var(&non_isolated_cpus, GFP_KERNEL);
alloc_cpumask_var(&fallback_doms, GFP_KERNEL);
#if defined(CONFIG_NUMA)
sched_group_nodes_bycpu = kzalloc(nr_cpu_ids * sizeof(void **),
@ -9196,7 +9197,6 @@ void __init sched_init_smp(void)
sched_init_granularity();
free_cpumask_var(non_isolated_cpus);
alloc_cpumask_var(&fallback_doms, GFP_KERNEL);
init_sched_rt_class();
}
#else