mm/page_alloc: use get_pfnblock_migratetype to avoid extra page_to_pfn

We have get_pageblock_migratetype and get_pfnblock_migratetype to get
migratetype of page.  get_pfnblock_migratetype accepts both page and pfn
from caller while get_pageblock_migratetype only accept page and get pfn
with page_to_pfn from page.

In case we already record pfn of page, we can simply call
get_pfnblock_migratetype to avoid a page_to_pfn.

Link: https://lkml.kernel.org/r/20230811115945.3423894-3-shikemeng@huaweicloud.com
Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
This commit is contained in:
Kemeng Shi 2023-08-11 19:59:45 +08:00 committed by Andrew Morton
parent a04d12c248
commit b5ffd29733

View File

@ -813,7 +813,7 @@ static inline void __free_one_page(struct page *page,
* pageblock isolation could cause incorrect freepage or CMA
* accounting or HIGHATOMIC accounting.
*/
int buddy_mt = get_pageblock_migratetype(buddy);
int buddy_mt = get_pfnblock_migratetype(buddy, buddy_pfn);
if (migratetype != buddy_mt
&& (!migratetype_is_mergeable(migratetype) ||
@ -889,7 +889,7 @@ int split_free_page(struct page *free_page,
goto out;
}
mt = get_pageblock_migratetype(free_page);
mt = get_pfnblock_migratetype(free_page, free_page_pfn);
if (likely(!is_migrate_isolate(mt)))
__mod_zone_freepage_state(zone, -(1UL << order), mt);