lkml.org 
[lkml]   [2010]   [Nov]   [9]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: [PATCH 1/2] mm: page allocator: Adjust the per-cpu counter threshold when memory is low
On Fri, Oct 29, 2010 at 12:40:02PM -0700, Andrew Morton wrote:
> > ...
> >
> > Follow-on patch?
>
> Sometime, please.
>

How does this look?

==== CUT HERE ====
mm: vmscan: Comment on why kswapd reduces the per-cpu vmstat threshold

While kswapd is awake, the per-cpu vmstat threshold is reduced to
reduce per-cpu drift to acceptable levels. Add a comment explaining
why.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
---
mm/vmscan.c | 13 +++++++++++++
1 files changed, 13 insertions(+), 0 deletions(-)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index 7966110..ba39948 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2378,6 +2378,19 @@ static int kswapd(void *p)
*/
if (!sleeping_prematurely(pgdat, order, remaining)) {
trace_mm_vmscan_kswapd_sleep(pgdat->node_id);
+
+ /*
+ * vmstat counters are not perfectly
+ * accurate and the estimated value
+ * for counters such as NR_FREE_PAGES
+ * can deviate from the true value by
+ * nr_online_cpus * threshold. To
+ * avoid the zone watermarks being
+ * breached while under pressure, we
+ * reduce the per-cpu vmstat threshold
+ * while kswapd is awake and restore
+ * them before going back to sleep.
+ */
set_pgdat_percpu_threshold(pgdat,
calculate_normal_threshold);
schedule();

\
 
 \ /
  Last update: 2010-11-09 12:37    [W:1.216 / U:0.004 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site