forked from luck/tmp_suning_uos_patched
da882c1f2e
While investigating TCP performance problems on 10Gb+ links, we found a tcp sender was dropping lot of incoming ACKS because of sk_rcvbuf limit in sk_add_backlog(), especially if receiver doesnt use GRO/LRO and sends one ACK every two MSS segments. A sender usually tweaks sk_sndbuf, but sk_rcvbuf stays at its default value (87380), allowing a too small backlog. A TCP ACK, even being small, can consume nearly same truesize space than outgoing packets. Using sk_rcvbuf + sk_sndbuf as a limit makes sense and is fast to compute. Performance results on netperf, single flow, receiver with disabled GRO/LRO : 7500 Mbits instead of 6050 Mbits, no more TCPBacklogDrop increments at sender. Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Neal Cardwell <ncardwell@google.com> Cc: Tom Herbert <therbert@google.com> Cc: Maciej Żenczykowski <maze@google.com> Cc: Yuchung Cheng <ycheng@google.com> Cc: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Cc: Rick Jones <rick.jones2@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net> |
||
---|---|---|
.. | ||
netfilter | ||
addrconf_core.c | ||
addrconf.c | ||
addrlabel.c | ||
af_inet6.c | ||
ah6.c | ||
anycast.c | ||
datagram.c | ||
esp6.c | ||
exthdrs_core.c | ||
exthdrs.c | ||
fib6_rules.c | ||
icmp.c | ||
inet6_connection_sock.c | ||
inet6_hashtables.c | ||
ip6_fib.c | ||
ip6_flowlabel.c | ||
ip6_input.c | ||
ip6_output.c | ||
ip6_tunnel.c | ||
ip6mr.c | ||
ipcomp6.c | ||
ipv6_sockglue.c | ||
Kconfig | ||
Makefile | ||
mcast.c | ||
mip6.c | ||
ndisc.c | ||
netfilter.c | ||
proc.c | ||
protocol.c | ||
raw.c | ||
reassembly.c | ||
route.c | ||
sit.c | ||
syncookies.c | ||
sysctl_net_ipv6.c | ||
tcp_ipv6.c | ||
tunnel6.c | ||
udp_impl.h | ||
udp.c | ||
udplite.c | ||
xfrm6_input.c | ||
xfrm6_mode_beet.c | ||
xfrm6_mode_ro.c | ||
xfrm6_mode_transport.c | ||
xfrm6_mode_tunnel.c | ||
xfrm6_output.c | ||
xfrm6_policy.c | ||
xfrm6_state.c | ||
xfrm6_tunnel.c |