From: Christoph Paasch cpaasch@openai.com
Accelerated Receive Flow Steering (aRFS) relies on sockets recording their RX flow hash into the rps_sock_flow_table so that incoming packets are steered to the CPU where the application runs.
With MPTCP, the application interacts with the parent MPTCP socket while data is carried over per-subflow TCP sockets. Without recording these subflows, aRFS cannot steer interrupts and RX processing for the flows to the desired CPU.
Record all subflows in the RPS table by calling sock_rps_record_flow() for each subflow at the start of mptcp_sendmsg(), mptcp_recvmsg() and mptcp_stream_accept(), by using the new helper mptcp_rps_record_subflows().
It does not by itself improve throughput, but ensures that IRQ and RX processing are directed to the right CPU, which is a prerequisite for effective aRFS.
Signed-off-by: Christoph Paasch cpaasch@openai.com Reviewed-by: Matthieu Baerts (NGI0) matttbe@kernel.org Signed-off-by: Matthieu Baerts (NGI0) matttbe@kernel.org --- net/mptcp/protocol.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+)
diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index ad41c48126e44fda646f1ec1c81957db1407a6cc..a8d57b88578dfea807d3d55e430849aa8005c637 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -12,6 +12,7 @@ #include <linux/sched/signal.h> #include <linux/atomic.h> #include <net/aligned_data.h> +#include <net/rps.h> #include <net/sock.h> #include <net/inet_common.h> #include <net/inet_hashtables.h> @@ -1740,6 +1741,20 @@ static u32 mptcp_send_limit(const struct sock *sk) return limit - not_sent; }
+static void mptcp_rps_record_subflows(const struct mptcp_sock *msk) +{ + struct mptcp_subflow_context *subflow; + + if (!rfs_is_needed()) + return; + + mptcp_for_each_subflow(msk, subflow) { + struct sock *ssk = mptcp_subflow_tcp_sock(subflow); + + sock_rps_record_flow(ssk); + } +} + static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len) { struct mptcp_sock *msk = mptcp_sk(sk); @@ -1753,6 +1768,8 @@ static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
lock_sock(sk);
+ mptcp_rps_record_subflows(msk); + if (unlikely(inet_test_bit(DEFER_CONNECT, sk) || msg->msg_flags & MSG_FASTOPEN)) { int copied_syn = 0; @@ -2131,6 +2148,8 @@ static int mptcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, goto out_err; }
+ mptcp_rps_record_subflows(msk); + timeo = sock_rcvtimeo(sk, flags & MSG_DONTWAIT);
len = min_t(size_t, len, INT_MAX); @@ -3922,6 +3941,8 @@ static int mptcp_stream_accept(struct socket *sock, struct socket *newsock, mptcp_sock_graft(ssk, newsock); }
+ mptcp_rps_record_subflows(msk); + /* Do late cleanup for the first subflow as necessary. Also * deal with bad peers not doing a complete shutdown. */