ip: do not set RFS core on error queue reads
authorSoheil Hassas Yeganeh <soheil@google.com>
Thu, 4 Jan 2018 02:47:10 +0000 (21:47 -0500)
committerDavid S. Miller <davem@davemloft.net>
Fri, 5 Jan 2018 16:14:56 +0000 (11:14 -0500)
commite3f2c4a3db1413bebfd502f7ac94fb55e3ba8c84
treef2aede6896cc108e16f9bc785c13d3261d0c1800
parent2e40b823f072e2b57a8d7c43a4bd95988ced5568
ip: do not set RFS core on error queue reads

We should only record RPS on normal reads and writes.
In single threaded processes, all calls record the same state. In
multi-threaded processes where a separate thread processes
errors, the RFS table mispredicts.

Note that, when CONFIG_RPS is disabled, sock_rps_record_flow
is a noop and no branch is added as a result of this patch.

Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv4/af_inet.c