[RFC PATCH 4/6] proc: Reduce cache miss in xfrm_statistics_seq_show

From: Jia He
Date: Mon Aug 29 2016 - 12:05:06 EST


This patch exchanges the two loop for collecting the percpu
statistics data. This can reduce cache misses by going through
all the items of each cpu sequentially.

Signed-off-by: Jia He <hejianet@xxxxxxxxx>
---
net/xfrm/xfrm_proc.c | 17 +++++++++++++----
1 file changed, 13 insertions(+), 4 deletions(-)

diff --git a/net/xfrm/xfrm_proc.c b/net/xfrm/xfrm_proc.c
index 9c4fbd8..c9df546 100644
--- a/net/xfrm/xfrm_proc.c
+++ b/net/xfrm/xfrm_proc.c
@@ -51,11 +51,20 @@ static const struct snmp_mib xfrm_mib_list[] = {
static int xfrm_statistics_seq_show(struct seq_file *seq, void *v)
{
struct net *net = seq->private;
- int i;
- for (i = 0; xfrm_mib_list[i].name; i++)
+ int i, c;
+ unsigned long buff[LINUX_MIB_XFRMMAX];
+
+ memset(buff, 0, sizeof(unsigned long) * LINUX_MIB_XFRMMAX);
+
+ for_each_possible_cpu(c)
+ for (i = 0; xfrm_mib_list[i].name != NULL; i++)
+ buff[i] += snmp_get_cpu_field(
+ net->mib.xfrm_statistics,
+ c, xfrm_mib_list[i].entry);
+ for (i = 0; xfrm_mib_list[i].name != NULL; i++)
seq_printf(seq, "%-24s\t%lu\n", xfrm_mib_list[i].name,
- snmp_fold_field(net->mib.xfrm_statistics,
- xfrm_mib_list[i].entry));
+ buff[i]);
+
return 0;
}

--
1.8.3.1