From 567e4b79731c352a17d73c483959f795d3593e03 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 6 Feb 2015 12:59:01 -0800 Subject: net: rfs: add hash collision detection Receive Flow Steering is a nice solution but suffers from hash collisions when a mix of connected and unconnected traffic is received on the host, when flow hash table is populated. Also, clearing flow in inet_release() makes RFS not very good for short lived flows, as many packets can follow close(). (FIN , ACK packets, ...) This patch extends the information stored into global hash table to not only include cpu number, but upper part of the hash value. I use a 32bit value, and dynamically split it in two parts. For host with less than 64 possible cpus, this gives 6 bits for the cpu number, and 26 (32-6) bits for the upper part of the hash. Since hash bucket selection use low order bits of the hash, we have a full hash match, if /proc/sys/net/core/rps_sock_flow_entries is big enough. If the hash found in flow table does not match, we fallback to RPS (if it is enabled for the rxqueue). This means that a packet for an non connected flow can avoid the IPI through a unrelated/victim CPU. This also means we no longer have to clear the table at socket close time, and this helps short lived flows performance. Signed-off-by: Eric Dumazet Acked-by: Tom Herbert Signed-off-by: David S. Miller --- net/core/sysctl_net_core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/core/sysctl_net_core.c') diff --git a/net/core/sysctl_net_core.c b/net/core/sysctl_net_core.c index fde21d19e61b..7a31be5e361f 100644 --- a/net/core/sysctl_net_core.c +++ b/net/core/sysctl_net_core.c @@ -65,7 +65,7 @@ static int rps_sock_flow_sysctl(struct ctl_table *table, int write, mutex_unlock(&sock_flow_mutex); return -ENOMEM; } - + rps_cpu_mask = roundup_pow_of_two(nr_cpu_ids) - 1; sock_table->mask = size - 1; } else sock_table = orig_sock_table; -- cgit v1.2.3