> T0: Task1 wakeup_kswapd(order=3)
> T1: kswapd enters balance_pgdat
> T2: Task2 wakeup_kswapd(order=2), because pages reclaimed by kswapd are used
> quickly
> T3: kswapd exits balance_pgdat. kswapd will do check. Now new order=2,
> pgdat->kswapd_max_order will become 0, but order=3, if sleeping_prematurely,
> then order will become pgdat->kswapd_max_order(0), while at this time the
> order should 2
> This isn't a big deal, but we do have a small window the order is wrong.
>
> Signed-off-by: Shaohua Li <shaohua.li@xxxxxxxxx>
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index d31d7ce..15cd0d2 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -2450,7 +2450,7 @@ static int kswapd(void *p)
> }
> }
>
> - order = pgdat->kswapd_max_order;
> + order = max_t(unsigned long, new_order, pgdat->kswapd_max_order);
> }
> finish_wait(&pgdat->kswapd_wait, &wait);
Good catch!
But unfortunatelly, the code is not correct. At least, don't fit corrent
design.
1) if "order < new_order" condition is false, we already decided to don't
use new_order. So, we shouldn't use new_order after kswapd_try_to_sleep()
2) if sleeping_prematurely() return false, it probably mean
zone_watermark_ok_safe(zone, order, high_wmark) return false.
therefore, we have to retry reclaim by using old 'order' parameter.
new patch is here.
>From 8f436224219a1da01985fd9644e1307e7c4cb8c3 Mon Sep 17 00:00:00 2001
From: KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx>
Date: Sun, 26 Dec 2010 21:10:55 +0900
Subject: [PATCH] vmscan: make kswapd use a correct order
If sleeping_prematurely() return false, It's a sign of retrying reclaim.
So, we don't have to drop old order value.
Reported-by: Shaohua Li <shaohua.li@xxxxxxxxx>
Cc: Minchan Kim <minchan.kim@xxxxxxxxx>
Cc: Mel Gorman <mel@xxxxxxxxx>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx>
---
mm/vmscan.c | 11 +++++++----
1 files changed, 7 insertions(+), 4 deletions(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 1fcadaf..f052a1a 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2364,13 +2364,13 @@ out:
return sc.nr_reclaimed;
}
-static void kswapd_try_to_sleep(pg_data_t *pgdat, int order)
+static int kswapd_try_to_sleep(pg_data_t *pgdat, int order)
{
long remaining = 0;
DEFINE_WAIT(wait);
if (freezing(current) || kthread_should_stop())
- return;
+ return 0;
prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
@@ -2399,13 +2399,17 @@ static void kswapd_try_to_sleep(pg_data_t *pgdat, int order)
set_pgdat_percpu_threshold(pgdat, calculate_normal_threshold);
schedule();
set_pgdat_percpu_threshold(pgdat, calculate_pressure_threshold);
+ order = pgdat->kswapd_max_order;
} else {
if (remaining)
count_vm_event(KSWAPD_LOW_WMARK_HIT_QUICKLY);
else
count_vm_event(KSWAPD_HIGH_WMARK_HIT_QUICKLY);
+ order = max(order, pgdat->kswapd_max_order);
}
finish_wait(&pgdat->kswapd_wait, &wait);
+
+ return order;
}
/*
@@ -2467,8 +2471,7 @@ static int kswapd(void *p)
*/
order = new_order;
} else {
- kswapd_try_to_sleep(pgdat, order);
- order = pgdat->kswapd_max_order;
+ order = kswapd_try_to_sleep(pgdat, order);
}
ret = try_to_freeze();
--
1.6.5.2
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom policy in Canada: sign http://dissolvethecrtc.ca/
Don't email: <a href