[PATCH v3 bpf-next 0/2] bpf tcp save syn set/get sockoptions

[PATCH v3 bpf-next 0/2] bpf tcp save syn set/get sockoptions

[Nikita V. Shirokov]

adding supprot for two new bpf's tcp sockopts:
TCP_SAVE_SYN (set) and TCP_SAVED_SYN (get)
this would allow for tcp-bpf program to build some logic based on fields from
ingress syn packet (e.g. doing tcp's tos/tclass reflection (see sample prog))
and do it transparently from userspace program point of view

v2->v3:
 - make patch series public
v1->v2:
 - adding proper SPDX license

Nikita V. Shirokov (2):
  new options for bpf_(set|get)sockopt
  new sample bpf prog

 net/core/filter.c                  | 25 +++++++--
 samples/bpf/Makefile               |  1 +
 samples/bpf/tcp_tos_reflect_kern.c | 87 ++++++++++++++++++++++++++++++
 3 files changed, 109 insertions(+), 4 deletions(-)
 create mode 100644 samples/bpf/tcp_tos_reflect_kern.c

-- 
2.17.1

[PATCH v3 bpf-next 1/2] new options for bpf_(set|get)sockopt

[Nikita V. Shirokov]

adding support for two new bpf's get/set sockopts: TCP_SAVE_SYN (set)
and TCP_SAVED_SYN (get). this would allow for bpf program to build
logic based on data from ingress SYN packet

Signed-off-by: Nikita V. Shirokov <tehnerd@fb.com>
---
 net/core/filter.c | 25 +++++++++++++++++++++----
 1 file changed, 21 insertions(+), 4 deletions(-)

diff --git a/net/core/filter.c b/net/core/filter.c
index c25eb36f1320..feb578506009 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -4007,6 +4007,12 @@ BPF_CALL_5(bpf_setsockopt, struct bpf_sock_ops_kern *, bpf_sock,
 					tp->snd_ssthresh = val;
 				}
 				break;
+			case TCP_SAVE_SYN:
+				if (val < 0 || val > 1)
+					ret = -EINVAL;
+				else
+					tp->save_syn = val;
+				break;
 			default:
 				ret = -EINVAL;
 			}
@@ -4032,21 +4038,32 @@ static const struct bpf_func_proto bpf_setsockopt_proto = {
 BPF_CALL_5(bpf_getsockopt, struct bpf_sock_ops_kern *, bpf_sock,
 	   int, level, int, optname, char *, optval, int, optlen)
 {
+	struct inet_connection_sock *icsk;
 	struct sock *sk = bpf_sock->sk;
+	struct tcp_sock *tp;
 
 	if (!sk_fullsock(sk))
 		goto err_clear;
-
 #ifdef CONFIG_INET
 	if (level == SOL_TCP && sk->sk_prot->getsockopt == tcp_getsockopt) {
-		if (optname == TCP_CONGESTION) {
-			struct inet_connection_sock *icsk = inet_csk(sk);
+		switch (optname) {
+		case TCP_CONGESTION:
+			icsk = inet_csk(sk);
 
 			if (!icsk->icsk_ca_ops || optlen <= 1)
 				goto err_clear;
 			strncpy(optval, icsk->icsk_ca_ops->name, optlen);
 			optval[optlen - 1] = 0;
-		} else {
+			break;
+		case TCP_SAVED_SYN:
+			tp = tcp_sk(sk);
+
+			if (optlen <= 0 || !tp->saved_syn ||
+			    optlen > tp->saved_syn[0])
+				goto err_clear;
+			memcpy(optval, tp->saved_syn + 1, optlen);
+			break;
+		default:
 			goto err_clear;
 		}
 	} else if (level == SOL_IP) {
-- 
2.17.1

[PATCH v3 bpf-next 2/2] new sample bpf prog

[Nikita V. Shirokov]

sample program which shows TCP_SAVE_SYN/TCP_SAVED_SYN usage example:
bpf's program which is doing TOS/TCLASS reflection (server would reply
with a same TOS/TCLASS as client)

Signed-off-by: Nikita V. Shirokov <tehnerd@fb.com>
---
 samples/bpf/Makefile               |  1 +
 samples/bpf/tcp_tos_reflect_kern.c | 87 ++++++++++++++++++++++++++++++
 2 files changed, 88 insertions(+)
 create mode 100644 samples/bpf/tcp_tos_reflect_kern.c

diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
index 36f9f41d094b..be0a961450bc 100644
--- a/samples/bpf/Makefile
+++ b/samples/bpf/Makefile
@@ -153,6 +153,7 @@ always += tcp_cong_kern.o
 always += tcp_iw_kern.o
 always += tcp_clamp_kern.o
 always += tcp_basertt_kern.o
+always += tcp_tos_reflect_kern.o
 always += xdp_redirect_kern.o
 always += xdp_redirect_map_kern.o
 always += xdp_redirect_cpu_kern.o
diff --git a/samples/bpf/tcp_tos_reflect_kern.c b/samples/bpf/tcp_tos_reflect_kern.c
new file mode 100644
index 000000000000..d51dab19eca6
--- /dev/null
+++ b/samples/bpf/tcp_tos_reflect_kern.c
@@ -0,0 +1,87 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2018 Facebook
+ *
+ * BPF program to automatically reflect TOS option from received syn packet
+ *
+ * Use load_sock_ops to load this BPF program.
+ */
+
+#include <uapi/linux/bpf.h>
+#include <uapi/linux/tcp.h>
+#include <uapi/linux/if_ether.h>
+#include <uapi/linux/if_packet.h>
+#include <uapi/linux/ip.h>
+#include <uapi/linux/ipv6.h>
+#include <uapi/linux/in.h>
+#include <linux/socket.h>
+#include "bpf_helpers.h"
+#include "bpf_endian.h"
+
+#define DEBUG 1
+
+#define bpf_printk(fmt, ...)					\
+({								\
+	       char ____fmt[] = fmt;				\
+	       bpf_trace_printk(____fmt, sizeof(____fmt),	\
+				##__VA_ARGS__);			\
+})
+
+SEC("sockops")
+int bpf_basertt(struct bpf_sock_ops *skops)
+{
+	char header[sizeof(struct ipv6hdr)];
+	struct ipv6hdr *hdr6;
+	struct iphdr *hdr;
+	int hdr_size = 0;
+	int save_syn = 1;
+	int tos = 0;
+	int rv = 0;
+	int op;
+
+	op = (int) skops->op;
+
+#ifdef DEBUG
+	bpf_printk("BPF command: %d\n", op);
+#endif
+	switch (op) {
+	case BPF_SOCK_OPS_TCP_LISTEN_CB:
+		rv = bpf_setsockopt(skops, SOL_TCP, TCP_SAVE_SYN,
+				   &save_syn, sizeof(save_syn));
+		break;
+	case BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB:
+		if (skops->family == AF_INET)
+			hdr_size = sizeof(struct iphdr);
+		else
+			hdr_size = sizeof(struct ipv6hdr);
+		rv = bpf_getsockopt(skops, SOL_TCP, TCP_SAVED_SYN,
+				    header, hdr_size);
+		if (!rv) {
+			if (skops->family == AF_INET) {
+				hdr = (struct iphdr *) header;
+				tos = hdr->tos;
+				if (tos != 0)
+					bpf_setsockopt(skops, SOL_IP, IP_TOS,
+						       &tos, sizeof(tos));
+			} else {
+				hdr6 = (struct ipv6hdr *) header;
+				tos = ((hdr6->priority) << 4 |
+				       (hdr6->flow_lbl[0]) >>  4);
+				if (tos)
+					bpf_setsockopt(skops, SOL_IPV6,
+						       IPV6_TCLASS,
+						       &tos, sizeof(tos));
+			}
+			rv = 0;
+		}
+		break;
+	default:
+		rv = -1;
+	}
+#ifdef DEBUG
+	bpf_printk("Returning %d\n", rv);
+#endif
+	skops->reply = rv;
+	return 1;
+}
+char _license[] SEC("license") = "GPL";
-- 
2.17.1

Re: [PATCH v3 bpf-next 0/2] bpf tcp save syn set/get sockoptions

[Alexei Starovoitov]

On Thu, Aug 30, 2018 at 07:51:52AM -0700, Nikita V. Shirokov wrote:
> 
> adding supprot for two new bpf's tcp sockopts:
> TCP_SAVE_SYN (set) and TCP_SAVED_SYN (get)
> this would allow for tcp-bpf program to build some logic based on fields from
> ingress syn packet (e.g. doing tcp's tos/tclass reflection (see sample prog))
> and do it transparently from userspace program point of view

Applied, Thanks
but please convert the sample code into selftest.