Re: [patch 2/3] mm: speculative get_page

[Nick Piggin]

On Tue, Apr 04, 2006 at 02:47:15AM -0700, Andrew Morton wrote:
> Nick Piggin <npiggin@xxxxxxx> wrote:
> >
> > +static inline struct page *page_cache_get_speculative(struct page **pagep)
> 
> Seems rather large to inline.
> 

Possibly... with all the debugging turned off, it is only atomic_inc
on UP, and atomic_inc_not_zero + several branches and barriers on SMP.

With only two callsites, I figure it is probably OK to be inline. It
probably looks bigger than it is...

> >  +{
> >  +	struct page *page;
> >  +
> >  +	VM_BUG_ON(in_interrupt());
> >  +
> >  +#ifndef CONFIG_SMP
> >  +	page = *pagep;
> >  +	if (unlikely(!page))
> >  +		return NULL;
> >  +
> >  +	VM_BUG_ON(!in_atomic());
> 
> This will go blam if !CONFIG_PREEMPT.

Hmm yes. Is there a safe way to do that? I guess it is pretty trivally
safely under rcu_read_lock , so that can probably just be removed.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/