On Sat, 15 Oct 2011 01:29:33 +0200, Andrew Morton [off-list ref] wrote:

On Thu, 06 Oct 2011 15:54:42 +0200
Marek Szyprowski [off-list ref] wrote:

quoted

From: KAMEZAWA Hiroyuki <redacted>

This commit introduces alloc_contig_freed_pages() function

The "freed" seems redundant to me.  Wouldn't "alloc_contig_pages" be a
better name?

The ?freed? is there because the function operates on pages that are in
buddy system, ie. it is given a range of PFNs that are to be removed
 from buddy system.

There's also a alloc_contig_range() function (added by next patch)
which frees pages in given range and then calls
alloc_contig_free_pages() to allocate them.

IMO, if there was an alloc_contig_pages() function, it would have to
be one level up (ie. it would figure out where to allocate memory and
then call alloc_contig_range()).  (That's really what CMA is doing).

Still, as I think of it now, maybe alloc_contig_free_range() would be
better?