On Fri, May 27, 2016 at 09:42:24AM +0800, Chen Feng wrote:

Hi Joonsoo,

quoted

-/* Free whole pageblock and set its migration type to MIGRATE_CMA. */
+/* Free whole pageblock and set its migration type to MIGRATE_MOVABLE. */
 void __init init_cma_reserved_pageblock(struct page *page)
 {
 	unsigned i = pageblock_nr_pages;

@@ -1605,7 +1602,7 @@ void __init init_cma_reserved_pageblock(struct page *page)
 
 	adjust_present_page_count(page, pageblock_nr_pages);
 
-	set_pageblock_migratetype(page, MIGRATE_CMA);
+	set_pageblock_migratetype(page, MIGRATE_MOVABLE);

I have a question here, if the ZONE_CMA pages are all movable.

Then the unmovable alloc will also use CMA memory. Is this right?

No, previous patch changes that the CMA memory is on separate zone,
ZONE_CMA. We allow that zone when gfp is GFP_HIGHUSER_MOVABLE so
unmovable allocation cannot happen on CMA memory.

Thanks.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>