lkml.org 
[lkml]   [2009]   [Feb]   [12]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
Subject[patch] sched: cpu hotplug fix

The fix below addresses a lockdep warning i triggered with your config.

(You could hit this if you try to use s2ram or hibernation or explicit
CPU hotplug ops.)

Ingo

-------------->
Subject: sched: cpu hotplug fix
From: Ingo Molnar <mingo@elte.hu>
Date: Thu Feb 12 11:35:40 CET 2009
rq_attach_root() does a kfree() with the runqueue lock held.

That's not a very wise move, fix it.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
kernel/sched.c | 15 ++++++++++++---
1 file changed, 12 insertions(+), 3 deletions(-)
Index: tip/kernel/sched.c
===================================================================
--- tip.orig/kernel/sched.c
+++ tip/kernel/sched.c
@@ -7619,20 +7619,26 @@ static void free_rootdomain(struct root_

static void rq_attach_root(struct rq *rq, struct root_domain *rd)
{
+ struct root_domain *old_rd = NULL;
unsigned long flags;

spin_lock_irqsave(&rq->lock, flags);

if (rq->rd) {
- struct root_domain *old_rd = rq->rd;
+ old_rd = rq->rd;

if (cpumask_test_cpu(rq->cpu, old_rd->online))
set_rq_offline(rq);

cpumask_clear_cpu(rq->cpu, old_rd->span);

- if (atomic_dec_and_test(&old_rd->refcount))
- free_rootdomain(old_rd);
+ /*
+ * If we dont want to free the old_rt yet then
+ * set old_rd to NULL to skip the freeing later
+ * in this function:
+ */
+ if (!atomic_dec_and_test(&old_rd->refcount))
+ old_rd = NULL;
}

atomic_inc(&rd->refcount);
@@ -7643,6 +7649,9 @@ static void rq_attach_root(struct rq *rq
set_rq_online(rq);

spin_unlock_irqrestore(&rq->lock, flags);
+
+ if (old_rd)
+ free_rootdomain(old_rd);
}

static int __init_refok init_rootdomain(struct root_domain *rd, bool bootmem)


\
 
 \ /
  Last update: 2009-02-12 11:53    [from the cache]
©2003-2011 Jasper Spaans