Re: software suspend in 2.5.70-mm3.

hugang (hugang@soulinfo.com)
Sun, 8 Jun 2003 10:26:26 +0800


On Tue, 3 Jun 2003 20:55:51 +0200
Pavel Machek <pavel@ucw.cz> wrote:

> Locating Ben's patch and forward-porting
> it would be way better...

I was tried Ben's patch, It's cool, Very stable in my laptop.

Here is two patch.
* suspend.c can swap more pages into swap space.
* vmscan.c can make swap faster.

--- linux-2.5.70/kernel/suspend.c.old Sun Jun 8 11:09:41 2003
+++ linux-2.5.70/kernel/suspend.c Sun Jun 8 10:54:55 2003
@@ -621,9 +621,17 @@
*/
static void free_some_memory(void)
{
+ unsigned int count = 10;
+
printk("Freeing memory: ");
- while (shrink_all_memory(10000))
+ while (count) {
+ unsigned int ret = shrink_all_memory(4 * 1024 * 1024 / PAGE_SIZE);
+ if (ret == 0) {
+ count--;
+ continue;
+ }
printk(".");
+ }
printk("|\n");
}
--- linux-2.5.70/mm/vmscan.c.old Sun Jun 8 11:08:27 2003
+++ linux-2.5.70/mm/vmscan.c Sun Jun 8 11:02:27 2003
@@ -882,7 +882,8 @@
* dead and from now on, only perform a short scan. Basically we're polling
* the zone for when the problem goes away.
*/
-static int balance_pgdat(pg_data_t *pgdat, int nr_pages, struct page_state *ps)
+static int balance_pgdat(pg_data_t *pgdat, int nr_pages,
+ struct page_state *ps, unsigned int time)
{
int to_free = nr_pages;
int priority;
@@ -930,7 +931,7 @@
}
if (all_zones_ok)
break;
- blk_congestion_wait(WRITE, HZ/10);
+ blk_congestion_wait(WRITE, HZ/time);
}
return nr_pages - to_free;
}
@@ -984,7 +985,7 @@
schedule();
finish_wait(&pgdat->kswapd_wait, &wait);
get_page_state(&ps);
- balance_pgdat(pgdat, 0, &ps);
+ balance_pgdat(pgdat, 0, &ps, 10);
}
}

@@ -1020,7 +1021,7 @@
struct page_state ps;

get_page_state(&ps);
- freed = balance_pgdat(pgdat, nr_to_free, &ps);
+ freed = balance_pgdat(pgdat, nr_to_free, &ps, 200);
ret += freed;
nr_to_free -= freed;
if (nr_to_free <= 0)

-- 
Hu Gang / Steve
Email        : huagng@soulinfo.com, steve@soulinfo.com
GPG FinePrint: 4099 3F1D AE01 1817 68F7  D499 A6C2 C418 86C8 610E
http://soulinfo.com/~hugang/HuGang.asc
ICQ#         : 205800361
Registered Linux User : 204016
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/