Re: Possible memory ordering bug in page reclaim?

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Sat, 2005-10-15 at 07:17 +0100, Hugh Dickins wrote:
> On Sat, 15 Oct 2005, Nick Piggin wrote:
> > 
> > Is there anything that prevents PageDirty from theoretically being
> > speculatively loaded before page_count here? (see patch)
> > 
> > It would result in pagecache corruption in the following situation:
> > 
> > 1                                2
> > find_get_page();
> > write to page                    write_lock(tree_lock);
> > SetPageDirty();                  if (page_count != 2
> > put_page();                          || PageDirty())
> > 
> > Now I'm worried that 2 might see PageDirty *before* SetPageDirty in
>                                   page->flags
> > 1, and page_count *after* put_page in 1.
> 
> I think you're right.  But I'm the last person to ask
> barrier/ordering questions of.  CC'ed Ben and Andrea.

yup, now the question is wether PG_Dirty will be visible to CPU 2 before
the page count is decremented right ? That depends on put_page, I
suppose. If it's doing a simple atomic, there is an issue. But atomics
with return has been so often abused as locks that they may have been
implemented with a barrier... (On ppc64, it will do an eieio, thus I
think it should be ok).

There is also a problem the other way around. Write to page, then set
page dirty... those writes may be visible to CPU 2 (that is the page
content be dirty) before find_get_page even increased the page count,
unless there is a barrier in there too.

Paul, Anton ?

Ben.