mm: fix off-by-two in __zone_watermark_ok()

Commit 88f5acf ("mm: page allocator: adjust the per-cpu counter
threshold when memory is low") changed the form how free_pages is
calculated but it forgot that we used to do free_pages - ((1 << order) -
1) so we ended up with off-by-two when calculating free_pages.

Reported-by: Wang Sheng-Hui <shhuiw@gmail.com>
Signed-off-by: Michal Hocko <mhocko@suse.cz>
Acked-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/page_alloc.c

@@ -1474,7 +1474,7 @@ static bool __zone_watermark_ok(struct zone *z, int order, unsigned long mark,
 	long min = mark;
 	int o;
 
-	free_pages -= (1 << order) + 1;
+	free_pages -= (1 << order) - 1;
 	if (alloc_flags & ALLOC_HIGH)
 		min -= min / 2;
 	if (alloc_flags & ALLOC_HARDER)