[PATCH] neigh: print nud_state in neigh timer handler.

[PATCH] neigh: print nud_state in neigh timer handler.

[Daniel Baluta]

From: Daniel Baluta <daniel.baluta@gmail.com>

For debugging purposes it is useful to know the exact state of a
non NUD_IN_TIMER neighbour entry whose timer handler just expired.

Signed-off-by: Daniel Baluta <dbaluta@ixiacom.com>
---
 net/core/neighbour.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/net/core/neighbour.c b/net/core/neighbour.c
index 909ecb3..6a8a311 100644
--- a/net/core/neighbour.c
+++ b/net/core/neighbour.c
@@ -874,7 +874,7 @@ static void neigh_timer_handler(unsigned long arg)
 
 	if (!(state & NUD_IN_TIMER)) {
 #ifndef CONFIG_SMP
-		printk(KERN_WARNING "neigh: timer & !nud_in_timer\n");
+		printk(KERN_WARNING "neigh: timer & !nud_in_timer, state:0x%x\n", state);
 #endif
 		goto out;
 	}
-- 
1.7.1

Re: [PATCH] neigh: print nud_state in neigh timer handler.

[David Miller]

From: Daniel Baluta <dbaluta@ixiacom.com>
Date: Mon, 31 Oct 2011 23:10:48 +0200

> From: Daniel Baluta <daniel.baluta@gmail.com>
> 
> For debugging purposes it is useful to know the exact state of a
> non NUD_IN_TIMER neighbour entry whose timer handler just expired.
> 
> Signed-off-by: Daniel Baluta <dbaluta@ixiacom.com>

This debugging log message is dubious.

Also, it is protected by !SMP and I can almost guarentee you that whatever
condition makes it happen leigitmately on SMP also can be triggered with
preemption enabled on !SMP.

So I'm going to simply remove this debug message entirely instead.

Re: [PATCH] neigh: print nud_state in neigh timer handler.

[Daniel Baluta]

On Tue, Nov 1, 2011 at 11:44 PM, David Miller <davem@davemloft.net> wrote:
> From: Daniel Baluta <dbaluta@ixiacom.com>
> Date: Mon, 31 Oct 2011 23:10:48 +0200
>
>> From: Daniel Baluta <daniel.baluta@gmail.com>
>>
>> For debugging purposes it is useful to know the exact state of a
>> non NUD_IN_TIMER neighbour entry whose timer handler just expired.
>>
>> Signed-off-by: Daniel Baluta <dbaluta@ixiacom.com>
>
> This debugging log message is dubious.
>
> Also, it is protected by !SMP and I can almost guarentee you that whatever
> condition makes it happen leigitmately on SMP also can be triggered with
> preemption enabled on !SMP.
>
> So I'm going to simply remove this debug message entirely instead.

Agree. If you can figure out when this condition is true, it is nice
to have a comment.

Daniel.