From: Hugh Dickins <hugh@veritas.com>
To: Andrew Morton <akpm@osdl.org>
Cc: linux-kernel@vger.kernel.org
Subject: [PATCH 04/13] swap extent list is ordered
Date: Sat, 9 Jul 2005 01:03:47 +0100 (BST) [thread overview]
Message-ID: <Pine.LNX.4.61.0507090102540.13391@goblin.wat.veritas.com> (raw)
In-Reply-To: <Pine.LNX.4.61.0507090057340.13391@goblin.wat.veritas.com>
There are several comments that swap's extent_list.prev points to the
lowest extent: that's not so, it's extent_list.next which points to it,
as you'd expect. And a couple of loops in add_swap_extent which go all
the way through the list, when they should just add to the other end.
Fix those up, and let map_swap_page search the list forwards: profiles
shows it to be twice as quick that way - because prefetch works better
on how the structs are typically kmalloc'ed? or because usually more
is written to than read from swap, and swap is allocated ascendingly?
Signed-off-by: Hugh Dickins <hugh@veritas.com>
---
include/linux/swap.h | 2 --
mm/swapfile.c | 27 +++++++++------------------
2 files changed, 9 insertions(+), 20 deletions(-)
--- swap3/include/linux/swap.h 2005-07-07 12:33:21.000000000 +0100
+++ swap4/include/linux/swap.h 2005-07-08 19:14:00.000000000 +0100
@@ -115,8 +115,6 @@ enum {
/*
* The in-memory structure used to track swap areas.
- * extent_list.prev points at the lowest-index extent. That list is
- * sorted.
*/
struct swap_info_struct {
unsigned int flags;
--- swap3/mm/swapfile.c 2005-07-08 19:13:46.000000000 +0100
+++ swap4/mm/swapfile.c 2005-07-08 19:14:00.000000000 +0100
@@ -830,9 +830,9 @@ sector_t map_swap_page(struct swap_info_
offset < (se->start_page + se->nr_pages)) {
return se->start_block + (offset - se->start_page);
}
- lh = se->list.prev;
+ lh = se->list.next;
if (lh == &sis->extent_list)
- lh = lh->prev;
+ lh = lh->next;
se = list_entry(lh, struct swap_extent, list);
sis->curr_swap_extent = se;
BUG_ON(se == start_se); /* It *must* be present */
@@ -857,10 +857,9 @@ static void destroy_swap_extents(struct
/*
* Add a block range (and the corresponding page range) into this swapdev's
- * extent list. The extent list is kept sorted in block order.
+ * extent list. The extent list is kept sorted in page order.
*
- * This function rather assumes that it is called in ascending sector_t order.
- * It doesn't look for extent coalescing opportunities.
+ * This function rather assumes that it is called in ascending page order.
*/
static int
add_swap_extent(struct swap_info_struct *sis, unsigned long start_page,
@@ -870,16 +869,15 @@ add_swap_extent(struct swap_info_struct
struct swap_extent *new_se;
struct list_head *lh;
- lh = sis->extent_list.next; /* The highest-addressed block */
- while (lh != &sis->extent_list) {
+ lh = sis->extent_list.prev; /* The highest page extent */
+ if (lh != &sis->extent_list) {
se = list_entry(lh, struct swap_extent, list);
- if (se->start_block + se->nr_pages == start_block &&
- se->start_page + se->nr_pages == start_page) {
+ BUG_ON(se->start_page + se->nr_pages != start_page);
+ if (se->start_block + se->nr_pages == start_block) {
/* Merge it */
se->nr_pages += nr_pages;
return 0;
}
- lh = lh->next;
}
/*
@@ -892,14 +890,7 @@ add_swap_extent(struct swap_info_struct
new_se->nr_pages = nr_pages;
new_se->start_block = start_block;
- lh = sis->extent_list.prev; /* The lowest block */
- while (lh != &sis->extent_list) {
- se = list_entry(lh, struct swap_extent, list);
- if (se->start_block > start_block)
- break;
- lh = lh->prev;
- }
- list_add_tail(&new_se->list, lh);
+ list_add_tail(&new_se->list, &sis->extent_list);
sis->nr_extents++;
return 0;
}
next prev parent reply other threads:[~2005-07-09 0:05 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-07-09 0:00 [PATCH 00/13] some swapfile patches Hugh Dickins
2005-07-09 0:01 ` [PATCH 01/13] update swapfile i_sem comment Hugh Dickins
2005-07-09 0:01 ` [PATCH 02/13] correct swapfile nr_good_pages Hugh Dickins
2005-07-09 0:02 ` [PATCH 03/13] move destroy_swap_extents calls Hugh Dickins
2005-07-09 0:03 ` Hugh Dickins [this message]
2005-07-09 0:04 ` [PATCH 05/13] show span of swap extents Hugh Dickins
2005-07-09 0:05 ` [PATCH 06/13] swap unsigned int consistency Hugh Dickins
2005-07-09 0:06 ` [PATCH 07/13] freeing update swap_list.next Hugh Dickins
2005-07-09 0:07 ` [PATCH 08/13] get_swap_page drop swap_list_lock Hugh Dickins
2005-07-09 0:08 ` [PATCH 09/13] scan_swap_map restyled Hugh Dickins
2005-07-09 0:09 ` [PATCH 10/13] scan_swap_map drop swap_device_lock Hugh Dickins
2005-07-09 0:10 ` [PATCH 11/13] scan_swap_map latency breaks Hugh Dickins
2005-07-09 0:11 ` [PATCH 12/13] swap_lock replace list+device Hugh Dickins
2005-07-09 0:15 ` [PATCH 13/13] update swsusp use of swap_info Hugh Dickins
2005-07-09 11:13 ` [PATCH 14/13] swsusp mod needed parentheses Hugh Dickins
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Pine.LNX.4.61.0507090102540.13391@goblin.wat.veritas.com \
--to=hugh@veritas.com \
--cc=akpm@osdl.org \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®