Re: kernel 2.6.39 (user mode linux) crashes (2.6.38 works fine)

[Darren Hart <dvhart@linux.intel.com>]

On 05/20/2011 12:56 AM, richard -rw- weinberger wrote:
> 2011/5/20 Toralf Förster <toralf.foerster@gmx.de>:
>> ...
>> Kernel panic - not syncing: Kernel mode fault at addr 0x0, ip 0x80a9f6b
> 
> Looks like a NULL-pointer bug.
> What code is at address 80a9f6b?
> Use "objdump -d -S | less" to find it.
> Please note, kernel binary and log message have to match!
> 
>> The file /var/log/messages of the UML says :
>>
>> 2011-05-20T09:33:03.455+02:00 n22_uml kernel: ------------[ cut here ]------------
>> 2011-05-20T09:33:03.455+02:00 n22_uml kernel: WARNING: at kernel/futex.c:789 wake_futex+0x28/0x60()
> 
> Is this really 2.6.39?
> Line 789 contains no WARN*().
> http://lxr.linux.no/#linux+v2.6.39/kernel/futex.c#L789
> 

I suspect Toralf is hitting the WARN_ON in __unqueue_futex:

	if (WARN_ON(!q->lock_ptr || !spin_is_locked(q->lock_ptr)
			|| plist_node_empty(&q->list)))

Toralf, can you instrument that let us know which of conditions is
triggering the WARN_ON? Something like the following should be adequate
to get you the line number. I suspect it is plist_node_empty give the
git bisect results you reported.


diff --git a/kernel/futex.c b/kernel/futex.c
index abd5324..7f31bca 100644
--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -782,8 +782,11 @@ static void __unqueue_futex(struct futex_q *q)
 {
 	struct futex_hash_bucket *hb;

-	if (WARN_ON(!q->lock_ptr || !spin_is_locked(q->lock_ptr)
-			|| plist_node_empty(&q->list)))
+	if (WARN_ON(!q->lock_ptr))
+		return;
+	if (!spin_is_locked(q->lock_ptr))
+		return;
+	if (plist_node_empty(&q->list))
 		return;

 	hb = container_of(q->lock_ptr, struct futex_hash_bucket, lock);




-- 
Darren Hart
Intel Open Source Technology Center
Yocto Project - Linux Kernel