Re: 2.6.24-rc1 fails with lockup and BUG:

[Peter Zijlstra]

On Wed, 2007-10-24 at 17:55 +0200, Ingo Molnar wrote:
> * Romano Giannetti <romanol@xxxxxxxxxxxxx> wrote:
> 
> > Done. The results are at:
> > 
> > http://www.dea.icai.upcomillas.es/romano/linux/info/2624rc1_2/
> > 
> > in the syslog-after-failed-suspend.txt file. After the failed suspend 
> > (at line 15766) there where the bunch of things in D-state. I have 
> > left the file intact.
> > 
> > At line 17646 there is:
> > 
> > WARNING: at kernel/lockdep.c:2033 trace_hardirqs_on()
> 
> hm, this lockdep warning caused lockdep to turn itself off - hence we 
> wont get to the really interesting warnings. We'll try to come up with a 
> solution for this.

Does this help?

---
Subject: lockdep: invalid irq usage

this function can be called from hardirq context.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
---

Index: linux-2.6-2/kernel/sched_debug.c
===================================================================
--- linux-2.6-2.orig/kernel/sched_debug.c
+++ linux-2.6-2/kernel/sched_debug.c
@@ -80,6 +80,7 @@ print_task(struct seq_file *m, struct rq
 static void print_rq(struct seq_file *m, struct rq *rq, int rq_cpu)
 {
 	struct task_struct *g, *p;
+	unsigned long flags;
 
 	SEQ_printf(m,
 	"\nrunnable tasks:\n"
@@ -88,7 +89,7 @@ static void print_rq(struct seq_file *m,
 	"------------------------------------------------------"
 	"----------------------------------------------------\n");
 
-	read_lock_irq(&tasklist_lock);
+	read_lock_irqsave(&tasklist_lock, flags);
 
 	do_each_thread(g, p) {
 		if (!p->se.on_rq || task_cpu(p) != rq_cpu)
@@ -97,7 +98,7 @@ static void print_rq(struct seq_file *m,
 		print_task(m, rq, p);
 	} while_each_thread(g, p);
 
-	read_unlock_irq(&tasklist_lock);
+	read_unlock_irqrestore(&tasklist_lock, flags);
 }
 
 void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)

Attachment: signature.asc
Description: This is a digitally signed message part