[Devel] Re: [PATCH 1/7] introduce atomic_dec_and_lock_irqsave()

[Roman Zippel]

Hi,

On Wed, 30 Aug 2006, Oleg Nesterov wrote:

> > AFAICT it's called via rcu, does that mean anything released via rcu has 
> > to be protected against interrupts?
> 
> RCU means softirq, probably we can use spin_lock_bh() to protect against deadlock.
> But free_uid() may be called with irqs disabled, we can't use local_bh_enable()
> in such a case.

Eek, I wasn't really aware of it and this would really suck. We should 
move things out of the interrupt context and not into it. :(
I would call it a bug in the rcu system.

bye, Roman