rcu: Don't allocate bootmem from rcu_init()
authorSasha Levin <sasha.levin@oracle.com>
Tue, 26 Mar 2013 15:35:16 +0000 (11:35 -0400)
committerPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Wed, 15 May 2013 17:41:12 +0000 (10:41 -0700)
When rcu_init() is called we already have slab working, allocating
bootmem at that point results in warnings and an allocation from
slab.  This commit therefore changes alloc_bootmem_cpumask_var() to
alloc_cpumask_var() in rcu_bootup_announce_oddness(), which is called
from rcu_init().

Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Tested-by: Robin Holt <holt@sgi.com>
[paulmck: convert to zalloc_cpumask_var(), as suggested by Yinghai Lu.]

kernel/rcutree_plugin.h

index 6d939a6..3db5a37 100644 (file)
@@ -88,7 +88,7 @@ static void __init rcu_bootup_announce_oddness(void)
 #ifdef CONFIG_RCU_NOCB_CPU
 #ifndef CONFIG_RCU_NOCB_CPU_NONE
        if (!have_rcu_nocb_mask) {
-               alloc_bootmem_cpumask_var(&rcu_nocb_mask);
+               zalloc_cpumask_var(&rcu_nocb_mask, GFP_KERNEL);
                have_rcu_nocb_mask = true;
        }
 #ifdef CONFIG_RCU_NOCB_CPU_ZERO