From: Martin Bligh <mbligh@mbligh.org>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Christoph Lameter <clameter@sgi.com>,
linux-mm@kvack.org, Nick Piggin <nickpiggin@yahoo.com.au>,
Peter Zijlstra <a.p.zijlstra@chello.nl>,
KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
Rik van Riel <riel@redhat.com>
Subject: Re: [RFC] Remove unswappable anonymous pages off the LRU
Date: Thu, 15 Feb 2007 17:40:09 -0800 [thread overview]
Message-ID: <45D50B79.5080002@mbligh.org> (raw)
In-Reply-To: <20070215171355.67c7e8b4.akpm@linux-foundation.org>
Andrew Morton wrote:
> On Thu, 15 Feb 2007 13:05:47 -0800 (PST)
> Christoph Lameter <clameter@sgi.com> wrote:
>
>> If we do not have any swap or we have run out of swap then anonymous pages
>> can no longer be removed from memory. In that case we simply treat them
>> like mlocked pages. For a kernel compiled CONFIG_SWAP off this means
>> that all anonymous pages are marked mlocked when they are allocated.
>
> It's nice and simple, but I think I'd prefer to wait for the existing mlock
> changes to crash a bit less before we do this.
>
> Is it true that PageMlocked() pages are never on the LRU? If so, perhaps
> we could overload the lru.next/prev on these pages to flag an mlocked page.
>
> #define PageMlocked(page) (page->lru.next == some_address_which_isnt_used_for_anwything_else)
Mine just created a locked list. If you stick them there, there's no
need for a page flag ... and we don't abuse the lru pointers AGAIN! ;-)
Suspect most of the rest of my patch is crap, but that might be useful?
M.
--- linux-2.6.17/include/linux/mm_inline.h 2006-06-17
18:49:35.000000000 -0
700
+++ linux-2.6.17-mlock_lru/include/linux/mm_inline.h 2006-07-28
15:53:15.0000
00000 -0700
@@ -28,6 +27,20 @@ del_page_from_inactive_list(struct zone
}
static inline void
+add_page_to_mlocked_list(struct zone *zone, struct page *page)
+{
+ list_add(&page->lru, &zone->mlocked_list);
+ zone->nr_mlocked--;
+}
+
+static inline void
+del_page_from_mlocked_list(struct zone *zone, struct page *page)
+{
+ list_del(&page->lru);
+ zone->nr_mlocked--;
+}
+
+static inline void
del_page_from_lru(struct zone *zone, struct page *page)
{
list_del(&page->lru);
diff -aurpN -X /home/mbligh/.diff.exclude
linux-2.6.17/include/linux/mmzone.h li
nux-2.6.17-mlock_lru/include/linux/mmzone.h
--- linux-2.6.17/include/linux/mmzone.h 2006-06-17 18:49:35.000000000 -0700
+++ linux-2.6.17-mlock_lru/include/linux/mmzone.h 2006-07-28
15:49:05.0000
00000 -0700
@@ -156,10 +156,12 @@ struct zone {
spinlock_t lru_lock;
struct list_head active_list;
struct list_head inactive_list;
+ struct list_head mlocked_list;
unsigned long nr_scan_active;
unsigned long nr_scan_inactive;
unsigned long nr_active;
unsigned long nr_inactive;
+ unsigned long nr_mlocked;
unsigned long pages_scanned; /* since last reclaim */
int all_unreclaimable; /* All pages pinned */
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2007-02-16 1:40 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-02-15 21:05 [RFC] Remove unswappable anonymous pages off the LRU Christoph Lameter
2007-02-15 22:31 ` Rik van Riel
2007-02-15 22:41 ` Christoph Lameter
2007-02-15 22:50 ` Rik van Riel
2007-02-15 22:53 ` Christoph Lameter
2007-02-15 23:19 ` Andrew Morton
2007-02-15 23:20 ` Lee Schermerhorn
2007-02-16 0:15 ` Andrew Morton
2007-02-16 1:13 ` Andrew Morton
2007-02-16 1:24 ` KAMEZAWA Hiroyuki
2007-02-16 1:40 ` Martin Bligh [this message]
2007-02-16 1:49 ` Andrew Morton
2007-02-16 2:21 ` Martin Bligh
2007-02-16 2:34 ` Christoph Lameter
2007-02-16 2:48 ` Andrew Morton
2007-02-16 2:50 ` Christoph Lameter
2007-02-16 3:18 ` Andrew Morton
2007-02-16 3:36 ` Christoph Lameter
2007-02-16 3:42 ` Andrew Morton
2007-02-16 3:50 ` Christoph Lameter
2007-02-16 4:02 ` Andrew Morton
2007-02-16 4:07 ` Christoph Lameter
2007-02-16 4:03 ` Andrew Morton
2007-02-16 4:14 ` Rik van Riel
2007-02-16 4:15 ` Christoph Lameter
2007-02-16 4:57 ` KAMEZAWA Hiroyuki
2007-02-16 5:16 ` Andrew Morton
2007-02-16 5:25 ` Christoph Lameter
2007-02-16 5:41 ` Andrew Morton
2007-02-16 5:19 ` Christoph Lameter
2007-02-16 4:24 ` Andrew Morton
2007-02-16 8:15 ` Peter Zijlstra
2007-02-16 9:11 ` Rafael J. Wysocki
2007-02-16 9:19 ` Peter Zijlstra
2007-02-16 10:10 ` Christoph Lameter
2007-02-16 10:17 ` Peter Zijlstra
2007-02-16 11:04 ` Rafael J. Wysocki
2007-02-16 2:16 ` Christoph Lameter
2007-02-16 3:17 ` Martin Bligh
2007-02-16 3:29 ` Christoph Lameter
2007-02-16 8:10 ` Peter Zijlstra
2007-02-16 2:15 ` Christoph Lameter
2007-02-16 2:55 ` Christoph Lameter
2007-02-16 5:02 ` Christoph Lameter
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=45D50B79.5080002@mbligh.org \
--to=mbligh@mbligh.org \
--cc=a.p.zijlstra@chello.nl \
--cc=akpm@linux-foundation.org \
--cc=clameter@sgi.com \
--cc=kamezawa.hiroyu@jp.fujitsu.com \
--cc=linux-mm@kvack.org \
--cc=nickpiggin@yahoo.com.au \
--cc=riel@redhat.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.