Re: [patch 2/2] m68k: Discontinuous memory support

[Andrew Morton <akpm@linux-foundation.org>]

On Mon, 28 May 2007 21:16:32 +0200
Geert Uytterhoeven <geert@linux-m68k.org> wrote:

> +	for_each_online_pgdat(pgdat) {
> +		for (i = 0; i < pgdat->node_spanned_pages; i++) {
> +			struct page *page = pgdat->node_mem_map + i;
> +			total++;
> +			if (PageReserved(page))
> +				reserved++;
> +			else if (PageSwapCache(page))
> +				cached++;
> +			else if (!page_count(page))
> +				free++;

This isn't really true.  Callers of the page allocator don't _have_ to use
page_count(): they can internally perform their own refcounting.  One such
caller is slab, so this "free" count can end up being grossly wrong.

> +			else
> +				shared += page_count(page) - 1;
> +		}