Re: VM-related Oops: 2.4.15pre1

[Rik van Riel]

On Mon, 19 Nov 2001, Linus Torvalds wrote:
> On Mon, 19 Nov 2001, Marcelo Tosatti wrote:
> >
> > We ended up talking about the possibility of a reschedule (IRQ) happening
> > before after the "spin_unlock(pagecache_lock)" but before the
> > "lru_cache_add()".
>
> So?
>
> The worst that happens is that the page is not on the LRU list, which
> just means that it won't be free'd until we add it (which we will do
> when the lru_cache_add() resumes..

I wonder if the following scenario is possible:

CPU 0				CPU 1

add_to_page_cache()		truncate_list_pages()

spin_lock(&pagecache_lock);
__add_to_page_cache()
spin_unlock(&pagecache_lock);

==> network irq
    ...				remove_inode_page()
    ...
==> softirqs			__free_pages_ok()
    ...
    ...
	*** page now on free list ***

lru_cache_add(page);

	*** BOOM ***

regards,

Rik
-- 
DMCA, SSSCA, W3C?  Who cares?  http://thefreeworld.net/

http://www.surriel.com/		http://distro.conectiva.com/