bootmem: micro optimize freeing pages in bulk
authorUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Tue, 10 Jan 2012 23:08:00 +0000 (15:08 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 11 Jan 2012 00:30:44 +0000 (16:30 -0800)
The first entry of bdata->node_bootmem_map holds the data for
bdata->node_min_pfn up to bdata->node_min_pfn + BITS_PER_LONG - 1.  So the
test for freeing all pages of a single map entry can be slightly relaxed.

Moreover use DIV_ROUND_UP in another place instead of open coding it.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Cc: Johannes Weiner <hannes@saeurebad.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/bootmem.c

index 1a77012..3e6f152 100644 (file)
@@ -56,7 +56,7 @@ early_param("bootmem_debug", bootmem_debug_setup);
 
 static unsigned long __init bootmap_bytes(unsigned long pages)
 {
-       unsigned long bytes = (pages + 7) / 8;
+       unsigned long bytes = DIV_ROUND_UP(pages, 8);
 
        return ALIGN(bytes, sizeof(long));
 }
@@ -197,7 +197,7 @@ static unsigned long __init free_all_bootmem_core(bootmem_data_t *bdata)
                idx = start - bdata->node_min_pfn;
                vec = ~map[idx / BITS_PER_LONG];
 
-               if (aligned && vec == ~0UL && start + BITS_PER_LONG < end) {
+               if (aligned && vec == ~0UL && start + BITS_PER_LONG <= end) {
                        int order = ilog2(BITS_PER_LONG);
 
                        __free_pages_bootmem(pfn_to_page(start), order);