[RFC PATCH 4/6] proc: Reduce cache miss in xfrm_statistics_seq_show
From: Jia He <hidden>
Date: 2016-08-29 16:04:58
Also in:
linux-sctp, lkml
Subsystem:
networking [general], networking [ipsec], the rest · Maintainers:
"David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Steffen Klassert, Herbert Xu, Linus Torvalds
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 <redacted> --- 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