From: Linus Torvalds Date: Mon, 12 Nov 2007 22:14:44 +0000 (-0800) Subject: Revert "Bias the placement of kernel pages at lower PFNs" X-Git-Tag: v2.6.24-rc3~152 X-Git-Url: https://openfabrics.org/gitweb/?a=commitdiff_plain;h=44048d700bcbfaf4bcca6e2e0a73d89d01ec0878;p=~shefty%2Frdma-dev.git Revert "Bias the placement of kernel pages at lower PFNs" This reverts commit 5adc5be7cd1bcef6bb64f5255d2a33f20a3cf5be. Alexey Dobriyan reports that it causes huge slowdowns under some loads, in his case a "mkfs.ext2" on a 30G partition. With the placement bias, the mkfs took over four minutes, with it reverted it's back to about ten seconds for Alexey. Reported-and-tested-by: Alexey Dobriyan Cc: Mel Gorman Cc: Andrew Morton Signed-off-by: Linus Torvalds --- diff --git a/mm/page_alloc.c b/mm/page_alloc.c index da69d833e06..12376ae3f73 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -749,23 +749,6 @@ int move_freepages_block(struct zone *zone, struct page *page, int migratetype) return move_freepages(zone, start_page, end_page, migratetype); } -/* Return the page with the lowest PFN in the list */ -static struct page *min_page(struct list_head *list) -{ - unsigned long min_pfn = -1UL; - struct page *min_page = NULL, *page;; - - list_for_each_entry(page, list, lru) { - unsigned long pfn = page_to_pfn(page); - if (pfn < min_pfn) { - min_pfn = pfn; - min_page = page; - } - } - - return min_page; -} - /* Remove an element from the buddy allocator from the fallback list */ static struct page *__rmqueue_fallback(struct zone *zone, int order, int start_migratetype) @@ -789,11 +772,8 @@ static struct page *__rmqueue_fallback(struct zone *zone, int order, if (list_empty(&area->free_list[migratetype])) continue; - /* Bias kernel allocations towards low pfns */ page = list_entry(area->free_list[migratetype].next, struct page, lru); - if (unlikely(start_migratetype != MIGRATE_MOVABLE)) - page = min_page(&area->free_list[migratetype]); area->nr_free--; /*