For the machines I have access to, Reloading the same pointer from
memory seems to inhibit complier optimization somewhat.
In emc_processing(), using a single packet pointer, instead reloading
it from memory with packets[i], improves performance by 0.3 Mpps (tested
with 10G NIC pushing 64 byte packets, with the base line of 12.2 Mpps).
Besides improving performance, this patch should also improves code
readability.
Signed-off-by: Andy Zhou <az...@ovn.org>
---
lib/dpif-netdev.c | 13 +++++++------
1 file changed, 7 insertions(+), 6 deletions(-)
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c
index 833174a..205176a 100644
--- a/lib/dpif-netdev.c
+++ b/lib/dpif-netdev.c
@@ -3302,9 +3302,10 @@ emc_processing(struct dp_netdev_pmd_thread *pmd, struct
dp_packet **packets,
for (i = 0; i < cnt; i++) {
struct dp_netdev_flow *flow;
+ struct dp_packet *packet = packets[i];
- if (OVS_UNLIKELY(dp_packet_size(packets[i]) < ETH_HEADER_LEN)) {
- dp_packet_delete(packets[i]);
+ if (OVS_UNLIKELY(dp_packet_size(packet) < ETH_HEADER_LEN)) {
+ dp_packet_delete(packet);
n_dropped++;
continue;
}
@@ -3314,18 +3315,18 @@ emc_processing(struct dp_netdev_pmd_thread *pmd, struct
dp_packet **packets,
OVS_PREFETCH(dp_packet_data(packets[i+1]));
}
- miniflow_extract(packets[i], &key.mf);
+ miniflow_extract(packet, &key.mf);
key.len = 0; /* Not computed yet. */
- key.hash = dpif_netdev_packet_get_rss_hash(packets[i], &key.mf);
+ key.hash = dpif_netdev_packet_get_rss_hash(packet, &key.mf);
flow = emc_lookup(flow_cache, &key);
if (OVS_LIKELY(flow)) {
- dp_netdev_queue_batches(packets[i], flow, &key.mf, batches,
+ dp_netdev_queue_batches(packet, flow, &key.mf, batches,
n_batches);
} else {
/* Exact match cache missed. Group missed packets together at
* the beginning of the 'packets' array. */
- packets[n_missed] = packets[i];
+ packets[n_missed] = packet;
keys[n_missed++] = key;
}
}
--
1.9.1
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
