Thanks for your advice!
Michael
On Wed, Jul 21, 2010 at 12:56 AM, Michael Blizek <michi1@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
Hi!
This is probably related to the "sleepable spinlock" feature in the realtime
On 17:42 Tue 20 Jul , Michael Zhu wrote:
> Hi,
>
> In my box, there is an error:
> BUG: scheduling with irqs disabled: ledd/0x00000000/224
> caller is rt_spin_lock_slowlock+0x13c/0x20c
>
> If process ledd is killed, BUG message is gone.
>
> Would you like to give me some suggestion to debug ?
kernel. Usually spinlocks do not sleep and thus may be used in codepaths which
cannot sleep, like interrupt handlers. The realtime kernel makes them
sleepable and created "raw" spinlock for situations where sleeping in not
possible. This looks like a situation where a sleepable spinlock is used, but
a raw spinlock in needed.
In order to debug this, it would be nice to have the full kernel log output,
especially the stack trace below rt_spin_lock_slowlock.
-Michi
--
programing a layer 3+4 network protocol for mesh networks
see http://michaelblizek.twilightparadox.com
