Hi,
the patch below should make page_launder() more well-behaved
than it is in -ac8 and -ac9 ... note, however, that this thing
is still completely untested and only in theory makes page_launder
behave better ;)
Since there seems to be a lot of VM testing going on at the
moment I thought I might as well send it out now so I can get
some feedback before I get into the airplane towards sweden
tomorrow...
cheers,
Rik
--
Linux MM bugzilla: http://linux-mm.org/bugzilla.shtml
Virtual memory is like a game you can't win;
However, without VM there's truly nothing to lose...
http://www.surriel.com/
http://www.conectiva.com/ http://distro.conectiva.com/
--- linux-2.4.1-ac8/mm/vmscan.c.orig Fri Feb 9 15:04:16 2001
+++ linux-2.4.1-ac8/mm/vmscan.c Sat Feb 10 20:50:40 2001
@@ -413,7 +413,7 @@
* This code is heavily inspired by the FreeBSD source code. Thanks
* go out to Matthew Dillon.
*
- * XXX: restrict number of pageouts in flight...
+ * XXX: restrict number of pageouts in flight by ->writepage...
*/
#define MAX_LAUNDER (1 << page_cluster)
int page_launder(int gfp_mask, int user)
@@ -514,7 +514,10 @@
spin_unlock(&pagemap_lru_lock);
writepage(page);
- flushed_pages++;
+ /* XXX: all ->writepage()s should use nr_async_pages */
+ if (!PageSwapCache(page))
+ flushed_pages++;
+ maxlaunder--;
page_cache_release(page);
/* And re-start the thing.. */
@@ -636,14 +639,16 @@
* with the paging load in the system and doesn't have
* the IO storm problem, so it just flushes all pages
* needed to fix the free shortage.
- *
- * XXX: keep track of nr_async_pages like the old swap
- * code did?
*/
- if (user)
+ maxlaunder = shortage;
+ maxlaunder -= flushed_pages;
+ maxlaunder -= atomic_read(&nr_async_pages);
+
+ if (maxlaunder <= 0)
+ goto out;
+
+ if (user && maxlaunder > MAX_LAUNDER)
maxlaunder = MAX_LAUNDER;
- else
- maxlaunder = shortage;
/*
* If we are called by a user program, we need to free
@@ -667,6 +672,7 @@
/*
* Return the amount of pages we freed or made freeable.
*/
+out:
return freed_pages + flushed_pages;
}
I just tested it here and it seems to behave pretty well.
On Sat, 10 Feb 2001, Rik van Riel wrote:
> Hi,
>
> the patch below should make page_launder() more well-behaved
> than it is in -ac8 and -ac9 ... note, however, that this thing
> is still completely untested and only in theory makes page_launder
> behave better ;)
>
> Since there seems to be a lot of VM testing going on at the
> moment I thought I might as well send it out now so I can get
> some feedback before I get into the airplane towards sweden
> tomorrow...
>
> cheers,
>