AF_VSOCK stream receives never fill msghdr.msg_inq, so io_uring cannot set IORING_CQE_F_SOCK_NONEMPTY and retries a multishot receive even after the queue has been drained.
Fill the hint at the common receive exit using the transport callback that SIOCINQ already uses, and report 1 once the connection is finished so the caller performs the receive which observes EOF, as TCP does after a FIN. A vsock loopback ping-pong with io_uring multishot receive drops entries into __vsock_connectible_recvmsg from 1.97 to 1.00 per delivered message, and receiver CPU time by about 3% (25 runs of 50000 messages, p=0.006). Signed-off-by: David Carlier <[email protected]> --- net/vmw_vsock/af_vsock.c | 35 +++++++++++++++++++++++++++++++++++ 1 file changed, 35 insertions(+) diff --git a/net/vmw_vsock/af_vsock.c b/net/vmw_vsock/af_vsock.c index f840498b58af..20d6f9ca6a96 100644 --- a/net/vmw_vsock/af_vsock.c +++ b/net/vmw_vsock/af_vsock.c @@ -2543,6 +2543,35 @@ static int __vsock_seqpacket_recvmsg(struct sock *sk, struct msghdr *msg, return err; } +/* Bytes a following receive can consume, 1 if it would only see EOF, or -1 + * if the transport cannot tell. + * + * Called under the socket lock after a nonnegative stream receive, so a NULL + * transport implies SOCK_DONE. + */ +static int vsock_stream_inq_hint(struct sock *sk) +{ + struct vsock_sock *vsk = vsock_sk(sk); + s64 data; + + if ((sk->sk_shutdown & RCV_SHUTDOWN) || !vsk->transport || + (sock_flag(sk, SOCK_DONE) && sk->sk_state != TCP_ESTABLISHED)) + return 1; + + data = vsock_stream_has_data(vsk); + if (data < 0) + return -1; + if (data > 0) + return min_t(s64, data, INT_MAX); + + /* Empty but finished: keep the caller reading so it sees EOF. */ + if (sock_flag(sk, SOCK_DONE) || + (READ_ONCE(vsk->peer_shutdown) & SEND_SHUTDOWN)) + return 1; + + return 0; +} + int __vsock_connectible_recvmsg(struct socket *sock, struct msghdr *msg, size_t len, int flags) @@ -2606,6 +2635,12 @@ __vsock_connectible_recvmsg(struct socket *sock, struct msghdr *msg, size_t len, err = __vsock_seqpacket_recvmsg(sk, msg, len, flags); out: + /* Seqpacket has_data counts messages, while io_uring treats msg_inq as + * a byte length when sizing retries, so only streams report a hint. + */ + if (msg->msg_get_inq && err >= 0 && sk->sk_type == SOCK_STREAM) + msg->msg_inq = vsock_stream_inq_hint(sk); + release_sock(sk); return err; } -- 2.55.0
