net: openvswitch: reorder masks array based on usage

This patch reorders the masks array every 4 seconds based on their
usage count. This greatly reduces the masks per packet hit, and
hence the overall performance. Especially in the OVS/OVN case for
OpenShift.

Here are some results from the OVS/OVN OpenShift test, which use
8 pods, each pod having 512 uperf connections, each connection
sends a 64-byte request and gets a 1024-byte response (TCP).
All uperf clients are on 1 worker node while all uperf servers are
on the other worker node.

Kernel without this patch     :  7.71 Gbps
Kernel with this patch applied: 14.52 Gbps

We also run some tests to verify the rebalance activity does not
lower the flow insertion rate, which does not.

Signed-off-by: Eelco Chaudron <echaudro@redhat.com>
Tested-by: Andrew Theurer <atheurer@redhat.com>
Reviewed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Eelco Chaudron 2020-07-15 14:09:28 +02:00 committed by David S. Miller
parent b18432c5a4
commit eac87c413b
4 changed files with 207 additions and 7 deletions

View file

@ -20,8 +20,9 @@
#include "meter.h"
#include "vport-internal_dev.h"
#define DP_MAX_PORTS USHRT_MAX
#define DP_VPORT_HASH_BUCKETS 1024
#define DP_MAX_PORTS USHRT_MAX
#define DP_VPORT_HASH_BUCKETS 1024
#define DP_MASKS_REBALANCE_INTERVAL 4000
/**
* struct dp_stats_percpu - per-cpu packet processing statistics for a given
@ -83,6 +84,9 @@ struct datapath {
/* Switch meters. */
struct dp_meter_table meter_tbl;
/* re-balance flow masks timer */
struct delayed_work masks_rebalance;
};
/**