[PATCH] mm: free_pages_and_swap_cache opt
authorHugh Dickins <hugh@veritas.com>
Fri, 6 Jan 2006 08:10:55 +0000 (00:10 -0800)
committerLinus Torvalds <torvalds@g5.osdl.org>
Fri, 6 Jan 2006 16:33:24 +0000 (08:33 -0800)
Minor optimization (though it doesn't help in the PREEMPT case, severely
constrained by small ZAP_BLOCK_SIZE).  free_pages_and_swap_cache works in
chunks of 16, calling release_pages which works in chunks of PAGEVEC_SIZE.
But PAGEVEC_SIZE was dropped from 16 to 14 in 2.6.10, so we're now doing more
spin_lock_irq'ing than necessary: use PAGEVEC_SIZE throughout.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
mm/swap_state.c

index 0df9a57b1de84bca534a9f6cc0c32d1ddefb939a..fc2aecb70a95c9e3559a2d4d1d4a2ae0700ebc50 100644 (file)
@@ -14,6 +14,7 @@
 #include <linux/pagemap.h>
 #include <linux/buffer_head.h>
 #include <linux/backing-dev.h>
+#include <linux/pagevec.h>
 
 #include <asm/pgtable.h>
 
@@ -272,12 +273,11 @@ void free_page_and_swap_cache(struct page *page)
  */
 void free_pages_and_swap_cache(struct page **pages, int nr)
 {
-       int chunk = 16;
        struct page **pagep = pages;
 
        lru_add_drain();
        while (nr) {
-               int todo = min(chunk, nr);
+               int todo = min(nr, PAGEVEC_SIZE);
                int i;
 
                for (i = 0; i < todo; i++)