[PATCH V4 net-next 1/2] tcp: send in-queue bytes in cmsg upon read

[PATCH V4 net-next 1/2] tcp: send in-queue bytes in cmsg upon read

[Soheil Hassas Yeganeh]

From: Soheil Hassas Yeganeh <soheil@google.com>

Applications with many concurrent connections, high variance
in receive queue length and tight memory bounds cannot
allocate worst-case buffer size to drain sockets. Knowing
the size of receive queue length, applications can optimize
how they allocate buffers to read from the socket.

The number of bytes pending on the socket is directly
available through ioctl(FIONREAD/SIOCINQ) and can be
approximated using getsockopt(MEMINFO) (rmem_alloc includes
skb overheads in addition to application data). But, both of
these options add an extra syscall per recvmsg. Moreover,
ioctl(FIONREAD/SIOCINQ) takes the socket lock.

Add the TCP_INQ socket option to TCP. When this socket
option is set, recvmsg() relays the number of bytes available
on the socket for reading to the application via the
TCP_CM_INQ control message.

Calculate the number of bytes after releasing the socket lock
to include the processed backlog, if any. To avoid an extra
branch in the hot path of recvmsg() for this new control
message, move all cmsg processing inside an existing branch for
processing receive timestamps. Since the socket lock is not held
when calculating the size of receive queue, TCP_INQ is a hint.
For example, it can overestimate the queue size by one byte,
if FIN is received.

With this method, applications can start reading from the socket
using a small buffer, and then use larger buffers based on the
remaining data when needed.

V3 change-log:
	As suggested by David Miller, added loads with barrier
	to check whether we have multiple threads calling recvmsg
	in parallel. When that happens we lock the socket to
	calculate inq.
V4 change-log:
	Removed inline from a static function.

Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Suggested-by: David Miller <davem@davemloft.net>
---
 include/linux/tcp.h      |  2 +-
 include/uapi/linux/tcp.h |  3 +++
 net/ipv4/tcp.c           | 43 ++++++++++++++++++++++++++++++++++++----
 3 files changed, 43 insertions(+), 5 deletions(-)

diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 20585d5c4e1c3..807776928cb86 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -228,7 +228,7 @@ struct tcp_sock {
 		unused:2;
 	u8	nonagle     : 4,/* Disable Nagle algorithm?             */
 		thin_lto    : 1,/* Use linear timeouts for thin streams */
-		unused1	    : 1,
+		recvmsg_inq : 1,/* Indicate # of bytes in queue upon recvmsg */
 		repair      : 1,
 		frto        : 1;/* F-RTO (RFC5682) activated in CA_Loss */
 	u8	repair_queue;
diff --git a/include/uapi/linux/tcp.h b/include/uapi/linux/tcp.h
index e9e8373b34b9d..29eb659aa77a1 100644
--- a/include/uapi/linux/tcp.h
+++ b/include/uapi/linux/tcp.h
@@ -123,6 +123,9 @@ enum {
 #define TCP_FASTOPEN_KEY	33	/* Set the key for Fast Open (cookie) */
 #define TCP_FASTOPEN_NO_COOKIE	34	/* Enable TFO without a TFO cookie */
 #define TCP_ZEROCOPY_RECEIVE	35
+#define TCP_INQ			36	/* Notify bytes available to read as a cmsg on read */
+
+#define TCP_CM_INQ		TCP_INQ
 
 struct tcp_repair_opt {
 	__u32	opt_code;
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 4028ddd14dd5a..868ed74a76a80 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -1889,6 +1889,22 @@ static void tcp_recv_timestamp(struct msghdr *msg, const struct sock *sk,
 	}
 }
 
+static int tcp_inq_hint(struct sock *sk)
+{
+	const struct tcp_sock *tp = tcp_sk(sk);
+	u32 copied_seq = READ_ONCE(tp->copied_seq);
+	u32 rcv_nxt = READ_ONCE(tp->rcv_nxt);
+	int inq;
+
+	inq = rcv_nxt - copied_seq;
+	if (unlikely(inq < 0 || copied_seq != READ_ONCE(tp->copied_seq))) {
+		lock_sock(sk);
+		inq = tp->rcv_nxt - tp->copied_seq;
+		release_sock(sk);
+	}
+	return inq;
+}
+
 /*
  *	This routine copies from a sock struct into the user buffer.
  *
@@ -1905,13 +1921,14 @@ int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
 	u32 peek_seq;
 	u32 *seq;
 	unsigned long used;
-	int err;
+	int err, inq;
 	int target;		/* Read at least this many bytes */
 	long timeo;
 	struct sk_buff *skb, *last;
 	u32 urg_hole = 0;
 	struct scm_timestamping tss;
 	bool has_tss = false;
+	bool has_cmsg;
 
 	if (unlikely(flags & MSG_ERRQUEUE))
 		return inet_recv_error(sk, msg, len, addr_len);
@@ -1926,6 +1943,7 @@ int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
 	if (sk->sk_state == TCP_LISTEN)
 		goto out;
 
+	has_cmsg = tp->recvmsg_inq;
 	timeo = sock_rcvtimeo(sk, nonblock);
 
 	/* Urgent data needs to be handled specially. */
@@ -2112,6 +2130,7 @@ int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
 		if (TCP_SKB_CB(skb)->has_rxtstamp) {
 			tcp_update_recv_tstamps(skb, &tss);
 			has_tss = true;
+			has_cmsg = true;
 		}
 		if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)
 			goto found_fin_ok;
@@ -2131,13 +2150,20 @@ int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
 	 * on connected socket. I was just happy when found this 8) --ANK
 	 */
 
-	if (has_tss)
-		tcp_recv_timestamp(msg, sk, &tss);
-
 	/* Clean up data we have read: This will do ACK frames. */
 	tcp_cleanup_rbuf(sk, copied);
 
 	release_sock(sk);
+
+	if (has_cmsg) {
+		if (has_tss)
+			tcp_recv_timestamp(msg, sk, &tss);
+		if (tp->recvmsg_inq) {
+			inq = tcp_inq_hint(sk);
+			put_cmsg(msg, SOL_TCP, TCP_CM_INQ, sizeof(inq), &inq);
+		}
+	}
+
 	return copied;
 
 out:
@@ -3006,6 +3032,12 @@ static int do_tcp_setsockopt(struct sock *sk, int level,
 		tp->notsent_lowat = val;
 		sk->sk_write_space(sk);
 		break;
+	case TCP_INQ:
+		if (val > 1 || val < 0)
+			err = -EINVAL;
+		else
+			tp->recvmsg_inq = val;
+		break;
 	default:
 		err = -ENOPROTOOPT;
 		break;
@@ -3431,6 +3463,9 @@ static int do_tcp_getsockopt(struct sock *sk, int level,
 	case TCP_NOTSENT_LOWAT:
 		val = tp->notsent_lowat;
 		break;
+	case TCP_INQ:
+		val = tp->recvmsg_inq;
+		break;
 	case TCP_SAVE_SYN:
 		val = tp->save_syn;
 		break;
-- 
2.17.0.441.gb46fe60e1d-goog

[PATCH V4 net-next 2/2] selftest: add test for TCP_INQ

[Soheil Hassas Yeganeh]

From: Soheil Hassas Yeganeh <soheil@google.com>

Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
---
 tools/testing/selftests/net/Makefile  |   3 +-
 tools/testing/selftests/net/tcp_inq.c | 189 ++++++++++++++++++++++++++
 2 files changed, 191 insertions(+), 1 deletion(-)
 create mode 100644 tools/testing/selftests/net/tcp_inq.c

diff --git a/tools/testing/selftests/net/Makefile b/tools/testing/selftests/net/Makefile
index df9102ec7b7af..0a1821f8dfb18 100644
--- a/tools/testing/selftests/net/Makefile
+++ b/tools/testing/selftests/net/Makefile
@@ -9,7 +9,7 @@ TEST_PROGS += fib_tests.sh fib-onlink-tests.sh in_netns.sh pmtu.sh udpgso.sh
 TEST_PROGS += udpgso_bench.sh
 TEST_GEN_FILES =  socket
 TEST_GEN_FILES += psock_fanout psock_tpacket msg_zerocopy
-TEST_GEN_FILES += tcp_mmap
+TEST_GEN_FILES += tcp_mmap tcp_inq
 TEST_GEN_PROGS = reuseport_bpf reuseport_bpf_cpu reuseport_bpf_numa
 TEST_GEN_PROGS += reuseport_dualstack reuseaddr_conflict
 TEST_GEN_PROGS += udpgso udpgso_bench_tx udpgso_bench_rx
@@ -18,3 +18,4 @@ include ../lib.mk
 
 $(OUTPUT)/reuseport_bpf_numa: LDFLAGS += -lnuma
 $(OUTPUT)/tcp_mmap: LDFLAGS += -lpthread
+$(OUTPUT)/tcp_inq: LDFLAGS += -lpthread
diff --git a/tools/testing/selftests/net/tcp_inq.c b/tools/testing/selftests/net/tcp_inq.c
new file mode 100644
index 0000000000000..d044b29ddabcc
--- /dev/null
+++ b/tools/testing/selftests/net/tcp_inq.c
@@ -0,0 +1,189 @@
+/*
+ * Copyright 2018 Google Inc.
+ * Author: Soheil Hassas Yeganeh (soheil@google.com)
+ *
+ * Simple example on how to use TCP_INQ and TCP_CM_INQ.
+ *
+ * License (GPLv2):
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. * See the GNU General Public License for
+ * more details.
+ */
+#define _GNU_SOURCE
+
+#include <error.h>
+#include <netinet/in.h>
+#include <netinet/tcp.h>
+#include <pthread.h>
+#include <stdio.h>
+#include <errno.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/socket.h>
+#include <unistd.h>
+
+#ifndef TCP_INQ
+#define TCP_INQ 36
+#endif
+
+#ifndef TCP_CM_INQ
+#define TCP_CM_INQ TCP_INQ
+#endif
+
+#define BUF_SIZE 8192
+#define CMSG_SIZE 32
+
+static int family = AF_INET6;
+static socklen_t addr_len = sizeof(struct sockaddr_in6);
+static int port = 4974;
+
+static void setup_loopback_addr(int family, struct sockaddr_storage *sockaddr)
+{
+	struct sockaddr_in6 *addr6 = (void *) sockaddr;
+	struct sockaddr_in *addr4 = (void *) sockaddr;
+
+	switch (family) {
+	case PF_INET:
+		memset(addr4, 0, sizeof(*addr4));
+		addr4->sin_family = AF_INET;
+		addr4->sin_addr.s_addr = htonl(INADDR_LOOPBACK);
+		addr4->sin_port = htons(port);
+		break;
+	case PF_INET6:
+		memset(addr6, 0, sizeof(*addr6));
+		addr6->sin6_family = AF_INET6;
+		addr6->sin6_addr = in6addr_loopback;
+		addr6->sin6_port = htons(port);
+		break;
+	default:
+		error(1, 0, "illegal family");
+	}
+}
+
+void *start_server(void *arg)
+{
+	int server_fd = (int)(unsigned long)arg;
+	struct sockaddr_in addr;
+	socklen_t addrlen = sizeof(addr);
+	char *buf;
+	int fd;
+	int r;
+
+	buf = malloc(BUF_SIZE);
+
+	for (;;) {
+		fd = accept(server_fd, (struct sockaddr *)&addr, &addrlen);
+		if (fd == -1) {
+			perror("accept");
+			break;
+		}
+		do {
+			r = send(fd, buf, BUF_SIZE, 0);
+		} while (r < 0 && errno == EINTR);
+		if (r < 0)
+			perror("send");
+		if (r != BUF_SIZE)
+			fprintf(stderr, "can only send %d bytes\n", r);
+		/* TCP_INQ can overestimate in-queue by one byte if we send
+		 * the FIN packet. Sleep for 1 second, so that the client
+		 * likely invoked recvmsg().
+		 */
+		sleep(1);
+		close(fd);
+	}
+
+	free(buf);
+	close(server_fd);
+	pthread_exit(0);
+}
+
+int main(int argc, char *argv[])
+{
+	struct sockaddr_storage listen_addr, addr;
+	int c, one = 1, inq = -1;
+	pthread_t server_thread;
+	char cmsgbuf[CMSG_SIZE];
+	struct iovec iov[1];
+	struct cmsghdr *cm;
+	struct msghdr msg;
+	int server_fd, fd;
+	char *buf;
+
+	while ((c = getopt(argc, argv, "46p:")) != -1) {
+		switch (c) {
+		case '4':
+			family = PF_INET;
+			addr_len = sizeof(struct sockaddr_in);
+			break;
+		case '6':
+			family = PF_INET6;
+			addr_len = sizeof(struct sockaddr_in6);
+			break;
+		case 'p':
+			port = atoi(optarg);
+			break;
+		}
+	}
+
+	server_fd = socket(family, SOCK_STREAM, 0);
+	if (server_fd < 0)
+		error(1, errno, "server socket");
+	setup_loopback_addr(family, &listen_addr);
+	if (setsockopt(server_fd, SOL_SOCKET, SO_REUSEADDR,
+		       &one, sizeof(one)) != 0)
+		error(1, errno, "setsockopt(SO_REUSEADDR)");
+	if (bind(server_fd, (const struct sockaddr *)&listen_addr,
+		 addr_len) == -1)
+		error(1, errno, "bind");
+	if (listen(server_fd, 128) == -1)
+		error(1, errno, "listen");
+	if (pthread_create(&server_thread, NULL, start_server,
+			   (void *)(unsigned long)server_fd) != 0)
+		error(1, errno, "pthread_create");
+
+	fd = socket(family, SOCK_STREAM, 0);
+	if (fd < 0)
+		error(1, errno, "client socket");
+	setup_loopback_addr(family, &addr);
+	if (connect(fd, (const struct sockaddr *)&addr, addr_len) == -1)
+		error(1, errno, "connect");
+	if (setsockopt(fd, SOL_TCP, TCP_INQ, &one, sizeof(one)) != 0)
+		error(1, errno, "setsockopt(TCP_INQ)");
+
+	msg.msg_name = NULL;
+	msg.msg_namelen = 0;
+	msg.msg_iov = iov;
+	msg.msg_iovlen = 1;
+	msg.msg_control = cmsgbuf;
+	msg.msg_controllen = sizeof(cmsgbuf);
+	msg.msg_flags = 0;
+
+	buf = malloc(BUF_SIZE);
+	iov[0].iov_base = buf;
+	iov[0].iov_len = BUF_SIZE / 2;
+
+	if (recvmsg(fd, &msg, 0) != iov[0].iov_len)
+		error(1, errno, "recvmsg");
+	if (msg.msg_flags & MSG_CTRUNC)
+		error(1, 0, "control message is truncated");
+
+	for (cm = CMSG_FIRSTHDR(&msg); cm; cm = CMSG_NXTHDR(&msg, cm))
+		if (cm->cmsg_level == SOL_TCP && cm->cmsg_type == TCP_CM_INQ)
+			inq = *((int *) CMSG_DATA(cm));
+
+	if (inq != BUF_SIZE - iov[0].iov_len) {
+		fprintf(stderr, "unexpected inq: %d\n", inq);
+		exit(1);
+	}
+
+	printf("PASSED\n");
+	free(buf);
+	close(fd);
+	return 0;
+}
-- 
2.17.0.441.gb46fe60e1d-goog

Re: [PATCH V4 net-next 1/2] tcp: send in-queue bytes in cmsg upon read

[David Miller]

From: Soheil Hassas Yeganeh <soheil.kdev@gmail.com>
Date: Tue,  1 May 2018 15:39:15 -0400

> Add the TCP_INQ socket option to TCP. When this socket
> option is set, recvmsg() relays the number of bytes available
> on the socket for reading to the application via the
> TCP_CM_INQ control message.

Applied.

Re: [PATCH V4 net-next 2/2] selftest: add test for TCP_INQ

[David Miller]

From: Soheil Hassas Yeganeh <soheil.kdev@gmail.com>
Date: Tue,  1 May 2018 15:39:16 -0400

> From: Soheil Hassas Yeganeh <soheil@google.com>
> 
> Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
> Signed-off-by: Yuchung Cheng <ycheng@google.com>
> Signed-off-by: Willem de Bruijn <willemb@google.com>
> Reviewed-by: Eric Dumazet <edumazet@google.com>
> Reviewed-by: Neal Cardwell <ncardwell@google.com>

Applied.