On Thu, Jan 26, 2012 at 10:00:50AM +0100, Marek Szyprowski wrote:
From: Michal Nazarewicz mina86@mina86.com @@ -875,10 +895,15 @@ struct page *__rmqueue_smallest(struct zone *zone, unsigned int order,
- This array describes the order lists are fallen back to when
- the free lists for the desirable migrate type are depleted
*/ -static int fallbacks[MIGRATE_TYPES][3] = { +static int fallbacks[MIGRATE_TYPES][4] = { [MIGRATE_UNMOVABLE] = { MIGRATE_RECLAIMABLE, MIGRATE_MOVABLE, MIGRATE_RESERVE }, [MIGRATE_RECLAIMABLE] = { MIGRATE_UNMOVABLE, MIGRATE_MOVABLE, MIGRATE_RESERVE }, +#ifdef CONFIG_CMA
- [MIGRATE_MOVABLE] = { MIGRATE_RECLAIMABLE, MIGRATE_UNMOVABLE, MIGRATE_CMA , MIGRATE_RESERVE },
On Mon, 30 Jan 2012 13:35:42 +0100, Mel Gorman mel@csn.ul.ie wrote:
This is a curious choice. MIGRATE_CMA is allowed to contain movable pages. By using MIGRATE_RECLAIMABLE and MIGRATE_UNMOVABLE for movable pages instead of MIGRATE_CMA, you increase the changes that unmovable pages will need to use MIGRATE_MOVABLE in the future which impacts fragmentation avoidance. I would recommend that you change this to
{ MIGRATE_CMA, MIGRATE_RECLAIMABLE, MIGRATE_UNMOVABLE, MIGRATE_RESERVE }
At the beginning the idea was to try hard not to get pages from MIGRATE_CMA allocated at all, thus it was put at the end of the fallbacks list, but on a busy system this probably won't help anyway, so I'll change it per your suggestion.
@@ -1017,11 +1049,14 @@ __rmqueue_fallback(struct zone *zone, int order, int start_migratetype) rmv_page_order(page);
/* Take ownership for orders >= pageblock_order */
if (current_order >= pageblock_order)
if (current_order >= pageblock_order &&!is_pageblock_cma(page)) change_pageblock_range(page, current_order, start_migratetype);
expand(zone, page, order, current_order, area, migratetype);
expand(zone, page, order, current_order, area,is_migrate_cma(start_migratetype)? start_migratetype : migratetype);What is this check meant to be doing?
start_migratetype is determined by allocflags_to_migratetype() and that never will be MIGRATE_CMA so is_migrate_cma(start_migratetype) should always be false.
Right, thanks! This should be the other way around, ie.:
+ expand(zone, page, order, current_order, area, + is_migrate_cma(migratetype) + ? migratetype : start_migratetype);
I'll fix this and the calls to is_pageblock_cma().