Re: [PATCH 01/15] mm: poison struct page for ptlock

[Andrew Morton <akpm@osdl.org>]

Hugh Dickins <hugh@veritas.com> wrote:
>
> On Tue, 15 Nov 2005, Andrew Morton wrote:
> > 
> > It occurs to me that we can do the above if (__GNUC__ > 2), or whatever.
> > 
> > That way, the only people who have a 4-byte-larger pageframe are those who
> > use CONFIG_PREEMPT, NR_CPUS>=4 and gcc-2.x.y.  An acceptably small
> > community, I suspect.
> 
> I can't really think of this at the moment (though the PageReserved
> fixups going smoother this evening).  Acceptably small community, yes.
> But wouldn't it plunge us into the very mess of wrappers we were trying
> to avoid with anony structunions, to handle the __GNUC__ differences?

Nope, all the changes would be constrained to the definition of struct
page, and struct page is special.

struct page {
	...
#if __GNUC__ > 2
	union {
		spinlock_t ptl;
		struct {
			unsigned long private;
			struct address_space *mapping;
		}
	}
#else
	union {
		unsigned long private;
		spinlock_t ptl;
	} u;
	struct address_space *mapping;
#endif


and

#if __GNUC__ > 2
#define page_private(page)		((page)->private)
#define set_page_private(page, v)	((page)->private = (v))
#else
#define page_private(page)		((page)->u.private)
#define set_page_private(page, v)	((page)->u.private = (v))
#endif

Of course, adding "u." and "u.s." all over the place would be a sane
solution, but we can do that later - I'm sure we'll be changing struct page
again.

View the above as "a space optimisation made possible by gcc-3.x".