On 15/04/11 09:51 -0500, ext Christoph Lameter wrote:
> On Fri, 15 Apr 2011, Phil Carmody wrote:
>
> > +++ b/include/linux/mm.h
> > @@ -353,9 +353,16 @@ static inline struct page *compound_head(struct page *page)
> > return page;
> > }
> >
> > -static inline int page_count(struct page *page)
> > +static inline const struct page *compound_head_ro(const struct page *page)
> > {
> > - return atomic_read(&compound_head(page)->_count);
> > + if (unlikely(PageTail(page)))
> > + return page->first_page;
> > + return page;
> > +}
>
> Can you make compound_head take a const pointer too to avoid this?
Not in C, alas. As it returns what it's given I wouldn't want it to lie
about the type of what it returns, and some of its clients want it to
return something writeable.
The simplest macro would have multiple-evaluation issues:
#define compound_head(page) (PageTail(page) ? (page)->first_page : (page))
Not that there are any clients who would misuse that currently, but setting
traps isn't a good way to make things cleaner.
Phil
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>