Re: PROBLEM: in_atomic() misuse all over the place

[Andrew Morton <akpm@linux-foundation.org>]

On Wed, 28 Jan 2009 13:18:50 +0100
Andi Kleen <andi@firstfloor.org> wrote:

> > file:  include/net/sock.h
> >
> >    static inline gfp_t gfp_any(void)
> >    {
> >      return in_atomic() ? GFP_ATOMIC : GFP_KERNEL;
> >    }
> 
> That's typically for softirq vs non softirq, which is important
> for the network stack.
> 

There's a bit of a problem here.  If someone accidentally uses
gfp_any() inside a spinlock, it will do a sleeping allocation on
non-preempt kernels and will do an atomic allocation on preemptible
kernels, so we won't get to see the warning which would allow us to fix
the bug.

Would using irq_count() work?  If so, that would fix this up.