[PATCH bpf-next v4 00/12] bpf: sockmap, sockhash: support storing UDP sockets

[PATCH bpf-next v4 00/12] bpf: sockmap, sockhash: support storing UDP sockets

[Lorenz Bauer]

I've adressed John's nit in patch 3, and added the reviews and acks.

Changes since v3:
- Clarify !psock check in sock_map_link_no_progs

Changes since v2:
- Remove sk_psock_hooks based on Jakub's idea
- Fix reference to tcp_bpf_clone in commit message
- Add inet_csk_has_ulp helper

Changes since v1:
- Check newsk->sk_prot in tcp_bpf_clone
- Fix compilation with BPF_STREAM_PARSER disabled
- Use spin_lock_init instead of static initializer
- Elaborate on TCPF_SYN_RECV
- Cosmetic changes to TEST macros, and more tests
- Add Jakub and me as maintainers

Lorenz Bauer (12):
  bpf: sockmap: only check ULP for TCP sockets
  skmsg: update saved hooks only once
  bpf: tcp: move assertions into tcp_bpf_get_proto
  bpf: tcp: guard declarations with CONFIG_NET_SOCK_MSG
  bpf: sockmap: move generic sockmap hooks from BPF TCP
  bpf: sockmap: simplify sock_map_init_proto
  bpf: add sockmap hooks for UDP sockets
  bpf: sockmap: add UDP support
  selftests: bpf: don't listen() on UDP sockets
  selftests: bpf: add tests for UDP sockets in sockmap
  selftests: bpf: enable UDP sockmap reuseport tests
  bpf, doc: update maintainers for L7 BPF

 MAINTAINERS                                   |   3 +
 include/linux/bpf.h                           |   4 +-
 include/linux/skmsg.h                         |  56 ++---
 include/net/inet_connection_sock.h            |   6 +
 include/net/tcp.h                             |  20 +-
 include/net/udp.h                             |   5 +
 net/core/sock_map.c                           | 157 +++++++++++---
 net/ipv4/Makefile                             |   1 +
 net/ipv4/tcp_bpf.c                            | 114 ++--------
 net/ipv4/tcp_ulp.c                            |   7 -
 net/ipv4/udp_bpf.c                            |  53 +++++
 .../bpf/prog_tests/select_reuseport.c         |   6 -
 .../selftests/bpf/prog_tests/sockmap_listen.c | 204 +++++++++++++-----
 13 files changed, 402 insertions(+), 234 deletions(-)
 create mode 100644 net/ipv4/udp_bpf.c

-- 
2.20.1

[PATCH bpf-next v4 01/12] bpf: sockmap: only check ULP for TCP sockets

[Lorenz Bauer]

The sock map code checks that a socket does not have an active upper
layer protocol before inserting it into the map. This requires casting
via inet_csk, which isn't valid for UDP sockets.

Guard checks for ULP by checking inet_sk(sk)->is_icsk first.

Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 include/linux/skmsg.h              | 8 +++++++-
 include/net/inet_connection_sock.h | 6 ++++++
 net/core/sock_map.c                | 6 ++----
 net/ipv4/tcp_ulp.c                 | 7 -------
 4 files changed, 15 insertions(+), 12 deletions(-)

diff --git a/include/linux/skmsg.h b/include/linux/skmsg.h
index 112765bd146d..4d3d75d63066 100644
--- a/include/linux/skmsg.h
+++ b/include/linux/skmsg.h
@@ -360,7 +360,13 @@ static inline void sk_psock_restore_proto(struct sock *sk,
 					  struct sk_psock *psock)
 {
 	sk->sk_prot->unhash = psock->saved_unhash;
-	tcp_update_ulp(sk, psock->sk_proto, psock->saved_write_space);
+	if (inet_csk_has_ulp(sk)) {
+		tcp_update_ulp(sk, psock->sk_proto, psock->saved_write_space);
+	} else {
+		sk->sk_write_space = psock->saved_write_space;
+		/* Pairs with lockless read in sk_clone_lock() */
+		WRITE_ONCE(sk->sk_prot, psock->sk_proto);
+	}
 }
 
 static inline void sk_psock_set_state(struct sk_psock *psock,
diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h
index 895546058a20..a3f076befa4f 100644
--- a/include/net/inet_connection_sock.h
+++ b/include/net/inet_connection_sock.h
@@ -335,4 +335,10 @@ static inline void inet_csk_inc_pingpong_cnt(struct sock *sk)
 	if (icsk->icsk_ack.pingpong < U8_MAX)
 		icsk->icsk_ack.pingpong++;
 }
+
+static inline bool inet_csk_has_ulp(struct sock *sk)
+{
+	return inet_sk(sk)->is_icsk && !!inet_csk(sk)->icsk_ulp_ops;
+}
+
 #endif /* _INET_CONNECTION_SOCK_H */
diff --git a/net/core/sock_map.c b/net/core/sock_map.c
index 2e0f465295c3..cb8f740f7949 100644
--- a/net/core/sock_map.c
+++ b/net/core/sock_map.c
@@ -384,7 +384,6 @@ static int sock_map_update_common(struct bpf_map *map, u32 idx,
 				  struct sock *sk, u64 flags)
 {
 	struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
-	struct inet_connection_sock *icsk = inet_csk(sk);
 	struct sk_psock_link *link;
 	struct sk_psock *psock;
 	struct sock *osk;
@@ -395,7 +394,7 @@ static int sock_map_update_common(struct bpf_map *map, u32 idx,
 		return -EINVAL;
 	if (unlikely(idx >= map->max_entries))
 		return -E2BIG;
-	if (unlikely(rcu_access_pointer(icsk->icsk_ulp_data)))
+	if (inet_csk_has_ulp(sk))
 		return -EINVAL;
 
 	link = sk_psock_init_link();
@@ -738,7 +737,6 @@ static int sock_hash_update_common(struct bpf_map *map, void *key,
 				   struct sock *sk, u64 flags)
 {
 	struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
-	struct inet_connection_sock *icsk = inet_csk(sk);
 	u32 key_size = map->key_size, hash;
 	struct bpf_htab_elem *elem, *elem_new;
 	struct bpf_htab_bucket *bucket;
@@ -749,7 +747,7 @@ static int sock_hash_update_common(struct bpf_map *map, void *key,
 	WARN_ON_ONCE(!rcu_read_lock_held());
 	if (unlikely(flags > BPF_EXIST))
 		return -EINVAL;
-	if (unlikely(icsk->icsk_ulp_data))
+	if (inet_csk_has_ulp(sk))
 		return -EINVAL;
 
 	link = sk_psock_init_link();
diff --git a/net/ipv4/tcp_ulp.c b/net/ipv4/tcp_ulp.c
index 2703f24c5d1a..7c27aa629af1 100644
--- a/net/ipv4/tcp_ulp.c
+++ b/net/ipv4/tcp_ulp.c
@@ -105,13 +105,6 @@ void tcp_update_ulp(struct sock *sk, struct proto *proto,
 {
 	struct inet_connection_sock *icsk = inet_csk(sk);
 
-	if (!icsk->icsk_ulp_ops) {
-		sk->sk_write_space = write_space;
-		/* Pairs with lockless read in sk_clone_lock() */
-		WRITE_ONCE(sk->sk_prot, proto);
-		return;
-	}
-
 	if (icsk->icsk_ulp_ops->update)
 		icsk->icsk_ulp_ops->update(sk, proto, write_space);
 }
-- 
2.20.1

[PATCH bpf-next v4 02/12] skmsg: update saved hooks only once

[Lorenz Bauer]

Only update psock->saved_* if psock->sk_proto has not been initialized
yet. This allows us to get rid of tcp_bpf_reinit_sk_prot.

Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 include/linux/skmsg.h | 20 ++++++++++++++++----
 net/ipv4/tcp_bpf.c    | 16 +---------------
 2 files changed, 17 insertions(+), 19 deletions(-)

diff --git a/include/linux/skmsg.h b/include/linux/skmsg.h
index 4d3d75d63066..2be51b7a5800 100644
--- a/include/linux/skmsg.h
+++ b/include/linux/skmsg.h
@@ -347,11 +347,23 @@ static inline void sk_psock_update_proto(struct sock *sk,
 					 struct sk_psock *psock,
 					 struct proto *ops)
 {
-	psock->saved_unhash = sk->sk_prot->unhash;
-	psock->saved_close = sk->sk_prot->close;
-	psock->saved_write_space = sk->sk_write_space;
+	/* Initialize saved callbacks and original proto only once, since this
+	 * function may be called multiple times for a psock, e.g. when
+	 * psock->progs.msg_parser is updated.
+	 *
+	 * Since we've not installed the new proto, psock is not yet in use and
+	 * we can initialize it without synchronization.
+	 */
+	if (!psock->sk_proto) {
+		struct proto *orig = READ_ONCE(sk->sk_prot);
+
+		psock->saved_unhash = orig->unhash;
+		psock->saved_close = orig->close;
+		psock->saved_write_space = sk->sk_write_space;
+
+		psock->sk_proto = orig;
+	}
 
-	psock->sk_proto = sk->sk_prot;
 	/* Pairs with lockless read in sk_clone_lock() */
 	WRITE_ONCE(sk->sk_prot, ops);
 }
diff --git a/net/ipv4/tcp_bpf.c b/net/ipv4/tcp_bpf.c
index 7d6e1b75d4d4..3327afa05c3d 100644
--- a/net/ipv4/tcp_bpf.c
+++ b/net/ipv4/tcp_bpf.c
@@ -637,20 +637,6 @@ static void tcp_bpf_update_sk_prot(struct sock *sk, struct sk_psock *psock)
 	sk_psock_update_proto(sk, psock, &tcp_bpf_prots[family][config]);
 }
 
-static void tcp_bpf_reinit_sk_prot(struct sock *sk, struct sk_psock *psock)
-{
-	int family = sk->sk_family == AF_INET6 ? TCP_BPF_IPV6 : TCP_BPF_IPV4;
-	int config = psock->progs.msg_parser   ? TCP_BPF_TX   : TCP_BPF_BASE;
-
-	/* Reinit occurs when program types change e.g. TCP_BPF_TX is removed
-	 * or added requiring sk_prot hook updates. We keep original saved
-	 * hooks in this case.
-	 *
-	 * Pairs with lockless read in sk_clone_lock().
-	 */
-	WRITE_ONCE(sk->sk_prot, &tcp_bpf_prots[family][config]);
-}
-
 static int tcp_bpf_assert_proto_ops(struct proto *ops)
 {
 	/* In order to avoid retpoline, we make assumptions when we call
@@ -670,7 +656,7 @@ void tcp_bpf_reinit(struct sock *sk)
 
 	rcu_read_lock();
 	psock = sk_psock(sk);
-	tcp_bpf_reinit_sk_prot(sk, psock);
+	tcp_bpf_update_sk_prot(sk, psock);
 	rcu_read_unlock();
 }
 
-- 
2.20.1

[PATCH bpf-next v4 03/12] bpf: tcp: move assertions into tcp_bpf_get_proto

[Lorenz Bauer]

We need to ensure that sk->sk_prot uses certain callbacks, so that
code that directly calls e.g. tcp_sendmsg in certain corner cases
works. To avoid spurious asserts, we must to do this only if
sk_psock_update_proto has not yet been called. The same invariants
apply for tcp_bpf_check_v6_needs_rebuild, so move the call as well.

Doing so allows us to merge tcp_bpf_init and tcp_bpf_reinit.

Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 include/net/tcp.h   |  1 -
 net/core/sock_map.c | 25 +++++++++----------------
 net/ipv4/tcp_bpf.c  | 42 ++++++++++++++++++++++--------------------
 3 files changed, 31 insertions(+), 37 deletions(-)

diff --git a/include/net/tcp.h b/include/net/tcp.h
index 07f947cc80e6..ccf39d80b695 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -2196,7 +2196,6 @@ struct sk_msg;
 struct sk_psock;
 
 int tcp_bpf_init(struct sock *sk);
-void tcp_bpf_reinit(struct sock *sk);
 int tcp_bpf_sendmsg_redir(struct sock *sk, struct sk_msg *msg, u32 bytes,
 			  int flags);
 int tcp_bpf_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
diff --git a/net/core/sock_map.c b/net/core/sock_map.c
index cb8f740f7949..fafcbd22ecba 100644
--- a/net/core/sock_map.c
+++ b/net/core/sock_map.c
@@ -145,8 +145,8 @@ static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
 			 struct sock *sk)
 {
 	struct bpf_prog *msg_parser, *skb_parser, *skb_verdict;
-	bool skb_progs, sk_psock_is_new = false;
 	struct sk_psock *psock;
+	bool skb_progs;
 	int ret;
 
 	skb_verdict = READ_ONCE(progs->skb_verdict);
@@ -191,18 +191,14 @@ static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
 			ret = -ENOMEM;
 			goto out_progs;
 		}
-		sk_psock_is_new = true;
 	}
 
 	if (msg_parser)
 		psock_set_prog(&psock->progs.msg_parser, msg_parser);
-	if (sk_psock_is_new) {
-		ret = tcp_bpf_init(sk);
-		if (ret < 0)
-			goto out_drop;
-	} else {
-		tcp_bpf_reinit(sk);
-	}
+
+	ret = tcp_bpf_init(sk);
+	if (ret < 0)
+		goto out_drop;
 
 	write_lock_bh(&sk->sk_callback_lock);
 	if (skb_progs && !psock->parser.enabled) {
@@ -239,15 +235,12 @@ static int sock_map_link_no_progs(struct bpf_map *map, struct sock *sk)
 	if (IS_ERR(psock))
 		return PTR_ERR(psock);
 
-	if (psock) {
-		tcp_bpf_reinit(sk);
-		return 0;
+	if (!psock) {
+		psock = sk_psock_init(sk, map->numa_node);
+		if (!psock)
+			return -ENOMEM;
 	}
 
-	psock = sk_psock_init(sk, map->numa_node);
-	if (!psock)
-		return -ENOMEM;
-
 	ret = tcp_bpf_init(sk);
 	if (ret < 0)
 		sk_psock_put(sk, psock);
diff --git a/net/ipv4/tcp_bpf.c b/net/ipv4/tcp_bpf.c
index 3327afa05c3d..ed8a8f3c9afe 100644
--- a/net/ipv4/tcp_bpf.c
+++ b/net/ipv4/tcp_bpf.c
@@ -629,14 +629,6 @@ static int __init tcp_bpf_v4_build_proto(void)
 }
 core_initcall(tcp_bpf_v4_build_proto);
 
-static void tcp_bpf_update_sk_prot(struct sock *sk, struct sk_psock *psock)
-{
-	int family = sk->sk_family == AF_INET6 ? TCP_BPF_IPV6 : TCP_BPF_IPV4;
-	int config = psock->progs.msg_parser   ? TCP_BPF_TX   : TCP_BPF_BASE;
-
-	sk_psock_update_proto(sk, psock, &tcp_bpf_prots[family][config]);
-}
-
 static int tcp_bpf_assert_proto_ops(struct proto *ops)
 {
 	/* In order to avoid retpoline, we make assumptions when we call
@@ -648,34 +640,44 @@ static int tcp_bpf_assert_proto_ops(struct proto *ops)
 	       ops->sendpage == tcp_sendpage ? 0 : -ENOTSUPP;
 }
 
-void tcp_bpf_reinit(struct sock *sk)
+static struct proto *tcp_bpf_get_proto(struct sock *sk, struct sk_psock *psock)
 {
-	struct sk_psock *psock;
+	int family = sk->sk_family == AF_INET6 ? TCP_BPF_IPV6 : TCP_BPF_IPV4;
+	int config = psock->progs.msg_parser   ? TCP_BPF_TX   : TCP_BPF_BASE;
 
-	sock_owned_by_me(sk);
+	if (!psock->sk_proto) {
+		struct proto *ops = READ_ONCE(sk->sk_prot);
 
-	rcu_read_lock();
-	psock = sk_psock(sk);
-	tcp_bpf_update_sk_prot(sk, psock);
-	rcu_read_unlock();
+		if (tcp_bpf_assert_proto_ops(ops))
+			return ERR_PTR(-EINVAL);
+
+		tcp_bpf_check_v6_needs_rebuild(sk, ops);
+	}
+
+	return &tcp_bpf_prots[family][config];
 }
 
 int tcp_bpf_init(struct sock *sk)
 {
-	struct proto *ops = READ_ONCE(sk->sk_prot);
 	struct sk_psock *psock;
+	struct proto *prot;
 
 	sock_owned_by_me(sk);
 
 	rcu_read_lock();
 	psock = sk_psock(sk);
-	if (unlikely(!psock || psock->sk_proto ||
-		     tcp_bpf_assert_proto_ops(ops))) {
+	if (unlikely(!psock)) {
 		rcu_read_unlock();
 		return -EINVAL;
 	}
-	tcp_bpf_check_v6_needs_rebuild(sk, ops);
-	tcp_bpf_update_sk_prot(sk, psock);
+
+	prot = tcp_bpf_get_proto(sk, psock);
+	if (IS_ERR(prot)) {
+		rcu_read_unlock();
+		return PTR_ERR(prot);
+	}
+
+	sk_psock_update_proto(sk, psock, prot);
 	rcu_read_unlock();
 	return 0;
 }
-- 
2.20.1

[PATCH bpf-next v4 04/12] bpf: tcp: guard declarations with CONFIG_NET_SOCK_MSG

[Lorenz Bauer]

tcp_bpf.c is only included in the build if CONFIG_NET_SOCK_MSG is
selected. The declaration should therefore be guarded as such.

Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 include/net/tcp.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/include/net/tcp.h b/include/net/tcp.h
index ccf39d80b695..ad3abeaa703e 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -2195,6 +2195,7 @@ void tcp_update_ulp(struct sock *sk, struct proto *p,
 struct sk_msg;
 struct sk_psock;
 
+#ifdef CONFIG_NET_SOCK_MSG
 int tcp_bpf_init(struct sock *sk);
 int tcp_bpf_sendmsg_redir(struct sock *sk, struct sk_msg *msg, u32 bytes,
 			  int flags);
@@ -2202,13 +2203,12 @@ int tcp_bpf_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 		    int nonblock, int flags, int *addr_len);
 int __tcp_bpf_recvmsg(struct sock *sk, struct sk_psock *psock,
 		      struct msghdr *msg, int len, int flags);
-#ifdef CONFIG_NET_SOCK_MSG
 void tcp_bpf_clone(const struct sock *sk, struct sock *newsk);
 #else
 static inline void tcp_bpf_clone(const struct sock *sk, struct sock *newsk)
 {
 }
-#endif
+#endif /* CONFIG_NET_SOCK_MSG */
 
 /* Call BPF_SOCK_OPS program that returns an int. If the return value
  * is < 0, then the BPF op failed (for example if the loaded BPF
-- 
2.20.1

[PATCH bpf-next v4 05/12] bpf: sockmap: move generic sockmap hooks from BPF TCP

[Lorenz Bauer]

The init, close and unhash handlers from TCP sockmap are generic,
and can be reused by UDP sockmap. Move the helpers into the sockmap code
base and expose them. This requires tcp_bpf_get_proto and tcp_bpf_clone to
be conditional on BPF_STREAM_PARSER.

The moved functions are unmodified, except that sk_psock_unlink is
renamed to sock_map_unlink to better match its behaviour.

Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 include/linux/bpf.h   |   4 +-
 include/linux/skmsg.h |  28 -----------
 include/net/tcp.h     |  15 +++---
 net/core/sock_map.c   | 106 ++++++++++++++++++++++++++++++++++++++++--
 net/ipv4/tcp_bpf.c    |  84 ++-------------------------------
 5 files changed, 118 insertions(+), 119 deletions(-)

diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 40c53924571d..94a329b9da81 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -1419,6 +1419,8 @@ static inline void bpf_map_offload_map_free(struct bpf_map *map)
 #if defined(CONFIG_BPF_STREAM_PARSER)
 int sock_map_prog_update(struct bpf_map *map, struct bpf_prog *prog, u32 which);
 int sock_map_get_from_fd(const union bpf_attr *attr, struct bpf_prog *prog);
+void sock_map_unhash(struct sock *sk);
+void sock_map_close(struct sock *sk, long timeout);
 #else
 static inline int sock_map_prog_update(struct bpf_map *map,
 				       struct bpf_prog *prog, u32 which)
@@ -1431,7 +1433,7 @@ static inline int sock_map_get_from_fd(const union bpf_attr *attr,
 {
 	return -EINVAL;
 }
-#endif
+#endif /* CONFIG_BPF_STREAM_PARSER */
 
 #if defined(CONFIG_INET) && defined(CONFIG_BPF_SYSCALL)
 void bpf_sk_reuseport_detach(struct sock *sk);
diff --git a/include/linux/skmsg.h b/include/linux/skmsg.h
index 2be51b7a5800..8a709f63c5e5 100644
--- a/include/linux/skmsg.h
+++ b/include/linux/skmsg.h
@@ -323,14 +323,6 @@ static inline void sk_psock_free_link(struct sk_psock_link *link)
 }
 
 struct sk_psock_link *sk_psock_link_pop(struct sk_psock *psock);
-#if defined(CONFIG_BPF_STREAM_PARSER)
-void sk_psock_unlink(struct sock *sk, struct sk_psock_link *link);
-#else
-static inline void sk_psock_unlink(struct sock *sk,
-				   struct sk_psock_link *link)
-{
-}
-#endif
 
 void __sk_psock_purge_ingress_msg(struct sk_psock *psock);
 
@@ -399,26 +391,6 @@ static inline bool sk_psock_test_state(const struct sk_psock *psock,
 	return test_bit(bit, &psock->state);
 }
 
-static inline struct sk_psock *sk_psock_get_checked(struct sock *sk)
-{
-	struct sk_psock *psock;
-
-	rcu_read_lock();
-	psock = sk_psock(sk);
-	if (psock) {
-		if (sk->sk_prot->recvmsg != tcp_bpf_recvmsg) {
-			psock = ERR_PTR(-EBUSY);
-			goto out;
-		}
-
-		if (!refcount_inc_not_zero(&psock->refcnt))
-			psock = ERR_PTR(-EBUSY);
-	}
-out:
-	rcu_read_unlock();
-	return psock;
-}
-
 static inline struct sk_psock *sk_psock_get(struct sock *sk)
 {
 	struct sk_psock *psock;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index ad3abeaa703e..43fa07a36fa6 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -2195,19 +2195,22 @@ void tcp_update_ulp(struct sock *sk, struct proto *p,
 struct sk_msg;
 struct sk_psock;
 
+#ifdef CONFIG_BPF_STREAM_PARSER
+struct proto *tcp_bpf_get_proto(struct sock *sk, struct sk_psock *psock);
+void tcp_bpf_clone(const struct sock *sk, struct sock *newsk);
+#else
+static inline void tcp_bpf_clone(const struct sock *sk, struct sock *newsk)
+{
+}
+#endif /* CONFIG_BPF_STREAM_PARSER */
+
 #ifdef CONFIG_NET_SOCK_MSG
-int tcp_bpf_init(struct sock *sk);
 int tcp_bpf_sendmsg_redir(struct sock *sk, struct sk_msg *msg, u32 bytes,
 			  int flags);
 int tcp_bpf_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 		    int nonblock, int flags, int *addr_len);
 int __tcp_bpf_recvmsg(struct sock *sk, struct sk_psock *psock,
 		      struct msghdr *msg, int len, int flags);
-void tcp_bpf_clone(const struct sock *sk, struct sock *newsk);
-#else
-static inline void tcp_bpf_clone(const struct sock *sk, struct sock *newsk)
-{
-}
 #endif /* CONFIG_NET_SOCK_MSG */
 
 /* Call BPF_SOCK_OPS program that returns an int. If the return value
diff --git a/net/core/sock_map.c b/net/core/sock_map.c
index fafcbd22ecba..cb240d87e068 100644
--- a/net/core/sock_map.c
+++ b/net/core/sock_map.c
@@ -141,6 +141,51 @@ static void sock_map_unref(struct sock *sk, void *link_raw)
 	}
 }
 
+static int sock_map_init_proto(struct sock *sk)
+{
+	struct sk_psock *psock;
+	struct proto *prot;
+
+	sock_owned_by_me(sk);
+
+	rcu_read_lock();
+	psock = sk_psock(sk);
+	if (unlikely(!psock)) {
+		rcu_read_unlock();
+		return -EINVAL;
+	}
+
+	prot = tcp_bpf_get_proto(sk, psock);
+	if (IS_ERR(prot)) {
+		rcu_read_unlock();
+		return PTR_ERR(prot);
+	}
+
+	sk_psock_update_proto(sk, psock, prot);
+	rcu_read_unlock();
+	return 0;
+}
+
+static struct sk_psock *sock_map_psock_get_checked(struct sock *sk)
+{
+	struct sk_psock *psock;
+
+	rcu_read_lock();
+	psock = sk_psock(sk);
+	if (psock) {
+		if (sk->sk_prot->recvmsg != tcp_bpf_recvmsg) {
+			psock = ERR_PTR(-EBUSY);
+			goto out;
+		}
+
+		if (!refcount_inc_not_zero(&psock->refcnt))
+			psock = ERR_PTR(-EBUSY);
+	}
+out:
+	rcu_read_unlock();
+	return psock;
+}
+
 static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
 			 struct sock *sk)
 {
@@ -172,7 +217,7 @@ static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
 		}
 	}
 
-	psock = sk_psock_get_checked(sk);
+	psock = sock_map_psock_get_checked(sk);
 	if (IS_ERR(psock)) {
 		ret = PTR_ERR(psock);
 		goto out_progs;
@@ -196,7 +241,7 @@ static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
 	if (msg_parser)
 		psock_set_prog(&psock->progs.msg_parser, msg_parser);
 
-	ret = tcp_bpf_init(sk);
+	ret = sock_map_init_proto(sk);
 	if (ret < 0)
 		goto out_drop;
 
@@ -231,7 +276,7 @@ static int sock_map_link_no_progs(struct bpf_map *map, struct sock *sk)
 	struct sk_psock *psock;
 	int ret;
 
-	psock = sk_psock_get_checked(sk);
+	psock = sock_map_psock_get_checked(sk);
 	if (IS_ERR(psock))
 		return PTR_ERR(psock);
 
@@ -241,7 +286,7 @@ static int sock_map_link_no_progs(struct bpf_map *map, struct sock *sk)
 			return -ENOMEM;
 	}
 
-	ret = tcp_bpf_init(sk);
+	ret = sock_map_init_proto(sk);
 	if (ret < 0)
 		sk_psock_put(sk, psock);
 	return ret;
@@ -1120,7 +1165,7 @@ int sock_map_prog_update(struct bpf_map *map, struct bpf_prog *prog,
 	return 0;
 }
 
-void sk_psock_unlink(struct sock *sk, struct sk_psock_link *link)
+static void sock_map_unlink(struct sock *sk, struct sk_psock_link *link)
 {
 	switch (link->map->map_type) {
 	case BPF_MAP_TYPE_SOCKMAP:
@@ -1133,3 +1178,54 @@ void sk_psock_unlink(struct sock *sk, struct sk_psock_link *link)
 		break;
 	}
 }
+
+static void sock_map_remove_links(struct sock *sk, struct sk_psock *psock)
+{
+	struct sk_psock_link *link;
+
+	while ((link = sk_psock_link_pop(psock))) {
+		sock_map_unlink(sk, link);
+		sk_psock_free_link(link);
+	}
+}
+
+void sock_map_unhash(struct sock *sk)
+{
+	void (*saved_unhash)(struct sock *sk);
+	struct sk_psock *psock;
+
+	rcu_read_lock();
+	psock = sk_psock(sk);
+	if (unlikely(!psock)) {
+		rcu_read_unlock();
+		if (sk->sk_prot->unhash)
+			sk->sk_prot->unhash(sk);
+		return;
+	}
+
+	saved_unhash = psock->saved_unhash;
+	sock_map_remove_links(sk, psock);
+	rcu_read_unlock();
+	saved_unhash(sk);
+}
+
+void sock_map_close(struct sock *sk, long timeout)
+{
+	void (*saved_close)(struct sock *sk, long timeout);
+	struct sk_psock *psock;
+
+	lock_sock(sk);
+	rcu_read_lock();
+	psock = sk_psock(sk);
+	if (unlikely(!psock)) {
+		rcu_read_unlock();
+		release_sock(sk);
+		return sk->sk_prot->close(sk, timeout);
+	}
+
+	saved_close = psock->saved_close;
+	sock_map_remove_links(sk, psock);
+	rcu_read_unlock();
+	release_sock(sk);
+	saved_close(sk, timeout);
+}
diff --git a/net/ipv4/tcp_bpf.c b/net/ipv4/tcp_bpf.c
index ed8a8f3c9afe..fe7b4fbc31c1 100644
--- a/net/ipv4/tcp_bpf.c
+++ b/net/ipv4/tcp_bpf.c
@@ -528,57 +528,7 @@ static int tcp_bpf_sendpage(struct sock *sk, struct page *page, int offset,
 	return copied ? copied : err;
 }
 
-static void tcp_bpf_remove(struct sock *sk, struct sk_psock *psock)
-{
-	struct sk_psock_link *link;
-
-	while ((link = sk_psock_link_pop(psock))) {
-		sk_psock_unlink(sk, link);
-		sk_psock_free_link(link);
-	}
-}
-
-static void tcp_bpf_unhash(struct sock *sk)
-{
-	void (*saved_unhash)(struct sock *sk);
-	struct sk_psock *psock;
-
-	rcu_read_lock();
-	psock = sk_psock(sk);
-	if (unlikely(!psock)) {
-		rcu_read_unlock();
-		if (sk->sk_prot->unhash)
-			sk->sk_prot->unhash(sk);
-		return;
-	}
-
-	saved_unhash = psock->saved_unhash;
-	tcp_bpf_remove(sk, psock);
-	rcu_read_unlock();
-	saved_unhash(sk);
-}
-
-static void tcp_bpf_close(struct sock *sk, long timeout)
-{
-	void (*saved_close)(struct sock *sk, long timeout);
-	struct sk_psock *psock;
-
-	lock_sock(sk);
-	rcu_read_lock();
-	psock = sk_psock(sk);
-	if (unlikely(!psock)) {
-		rcu_read_unlock();
-		release_sock(sk);
-		return sk->sk_prot->close(sk, timeout);
-	}
-
-	saved_close = psock->saved_close;
-	tcp_bpf_remove(sk, psock);
-	rcu_read_unlock();
-	release_sock(sk);
-	saved_close(sk, timeout);
-}
-
+#ifdef CONFIG_BPF_STREAM_PARSER
 enum {
 	TCP_BPF_IPV4,
 	TCP_BPF_IPV6,
@@ -599,8 +549,8 @@ static void tcp_bpf_rebuild_protos(struct proto prot[TCP_BPF_NUM_CFGS],
 				   struct proto *base)
 {
 	prot[TCP_BPF_BASE]			= *base;
-	prot[TCP_BPF_BASE].unhash		= tcp_bpf_unhash;
-	prot[TCP_BPF_BASE].close		= tcp_bpf_close;
+	prot[TCP_BPF_BASE].unhash		= sock_map_unhash;
+	prot[TCP_BPF_BASE].close		= sock_map_close;
 	prot[TCP_BPF_BASE].recvmsg		= tcp_bpf_recvmsg;
 	prot[TCP_BPF_BASE].stream_memory_read	= tcp_bpf_stream_read;
 
@@ -640,7 +590,7 @@ static int tcp_bpf_assert_proto_ops(struct proto *ops)
 	       ops->sendpage == tcp_sendpage ? 0 : -ENOTSUPP;
 }
 
-static struct proto *tcp_bpf_get_proto(struct sock *sk, struct sk_psock *psock)
+struct proto *tcp_bpf_get_proto(struct sock *sk, struct sk_psock *psock)
 {
 	int family = sk->sk_family == AF_INET6 ? TCP_BPF_IPV6 : TCP_BPF_IPV4;
 	int config = psock->progs.msg_parser   ? TCP_BPF_TX   : TCP_BPF_BASE;
@@ -657,31 +607,6 @@ static struct proto *tcp_bpf_get_proto(struct sock *sk, struct sk_psock *psock)
 	return &tcp_bpf_prots[family][config];
 }
 
-int tcp_bpf_init(struct sock *sk)
-{
-	struct sk_psock *psock;
-	struct proto *prot;
-
-	sock_owned_by_me(sk);
-
-	rcu_read_lock();
-	psock = sk_psock(sk);
-	if (unlikely(!psock)) {
-		rcu_read_unlock();
-		return -EINVAL;
-	}
-
-	prot = tcp_bpf_get_proto(sk, psock);
-	if (IS_ERR(prot)) {
-		rcu_read_unlock();
-		return PTR_ERR(prot);
-	}
-
-	sk_psock_update_proto(sk, psock, prot);
-	rcu_read_unlock();
-	return 0;
-}
-
 /* If a child got cloned from a listening socket that had tcp_bpf
  * protocol callbacks installed, we need to restore the callbacks to
  * the default ones because the child does not inherit the psock state
@@ -695,3 +620,4 @@ void tcp_bpf_clone(const struct sock *sk, struct sock *newsk)
 	if (prot == &tcp_bpf_prots[family][TCP_BPF_BASE])
 		newsk->sk_prot = sk->sk_prot_creator;
 }
+#endif /* CONFIG_BPF_STREAM_PARSER */
-- 
2.20.1

[PATCH bpf-next v4 06/12] bpf: sockmap: simplify sock_map_init_proto

[Lorenz Bauer]

We can take advantage of the fact that both callers of
sock_map_init_proto are holding a RCU read lock, and
have verified that psock is valid.

Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 net/core/sock_map.c | 19 ++++---------------
 1 file changed, 4 insertions(+), 15 deletions(-)

diff --git a/net/core/sock_map.c b/net/core/sock_map.c
index cb240d87e068..edfdce17b951 100644
--- a/net/core/sock_map.c
+++ b/net/core/sock_map.c
@@ -141,28 +141,17 @@ static void sock_map_unref(struct sock *sk, void *link_raw)
 	}
 }
 
-static int sock_map_init_proto(struct sock *sk)
+static int sock_map_init_proto(struct sock *sk, struct sk_psock *psock)
 {
-	struct sk_psock *psock;
 	struct proto *prot;
 
 	sock_owned_by_me(sk);
 
-	rcu_read_lock();
-	psock = sk_psock(sk);
-	if (unlikely(!psock)) {
-		rcu_read_unlock();
-		return -EINVAL;
-	}
-
 	prot = tcp_bpf_get_proto(sk, psock);
-	if (IS_ERR(prot)) {
-		rcu_read_unlock();
+	if (IS_ERR(prot))
 		return PTR_ERR(prot);
-	}
 
 	sk_psock_update_proto(sk, psock, prot);
-	rcu_read_unlock();
 	return 0;
 }
 
@@ -241,7 +230,7 @@ static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
 	if (msg_parser)
 		psock_set_prog(&psock->progs.msg_parser, msg_parser);
 
-	ret = sock_map_init_proto(sk);
+	ret = sock_map_init_proto(sk, psock);
 	if (ret < 0)
 		goto out_drop;
 
@@ -286,7 +275,7 @@ static int sock_map_link_no_progs(struct bpf_map *map, struct sock *sk)
 			return -ENOMEM;
 	}
 
-	ret = sock_map_init_proto(sk);
+	ret = sock_map_init_proto(sk, psock);
 	if (ret < 0)
 		sk_psock_put(sk, psock);
 	return ret;
-- 
2.20.1

[PATCH bpf-next v4 07/12] bpf: add sockmap hooks for UDP sockets

[Lorenz Bauer]

Add basic psock hooks for UDP sockets. This allows adding and
removing sockets, as well as automatic removal on unhash and close.

Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 MAINTAINERS        |  1 +
 include/net/udp.h  |  5 +++++
 net/ipv4/Makefile  |  1 +
 net/ipv4/udp_bpf.c | 53 ++++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 60 insertions(+)
 create mode 100644 net/ipv4/udp_bpf.c

diff --git a/MAINTAINERS b/MAINTAINERS
index c23884e084be..14554bde1c06 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -9370,6 +9370,7 @@ F:	include/linux/skmsg.h
 F:	net/core/skmsg.c
 F:	net/core/sock_map.c
 F:	net/ipv4/tcp_bpf.c
+F:	net/ipv4/udp_bpf.c
 
 LANTIQ / INTEL Ethernet drivers
 M:	Hauke Mehrtens <hauke@hauke-m.de>
diff --git a/include/net/udp.h b/include/net/udp.h
index e55d5f765807..a8fa6c0c6ded 100644
--- a/include/net/udp.h
+++ b/include/net/udp.h
@@ -503,4 +503,9 @@ static inline struct sk_buff *udp_rcv_segment(struct sock *sk,
 	return segs;
 }
 
+#ifdef CONFIG_BPF_STREAM_PARSER
+struct sk_psock;
+struct proto *udp_bpf_get_proto(struct sock *sk, struct sk_psock *psock);
+#endif /* BPF_STREAM_PARSER */
+
 #endif	/* _UDP_H */
diff --git a/net/ipv4/Makefile b/net/ipv4/Makefile
index 9d97bace13c8..9e1a186a3671 100644
--- a/net/ipv4/Makefile
+++ b/net/ipv4/Makefile
@@ -61,6 +61,7 @@ obj-$(CONFIG_TCP_CONG_LP) += tcp_lp.o
 obj-$(CONFIG_TCP_CONG_YEAH) += tcp_yeah.o
 obj-$(CONFIG_TCP_CONG_ILLINOIS) += tcp_illinois.o
 obj-$(CONFIG_NET_SOCK_MSG) += tcp_bpf.o
+obj-$(CONFIG_BPF_STREAM_PARSER) += udp_bpf.o
 obj-$(CONFIG_NETLABEL) += cipso_ipv4.o
 
 obj-$(CONFIG_XFRM) += xfrm4_policy.o xfrm4_state.o xfrm4_input.o \
diff --git a/net/ipv4/udp_bpf.c b/net/ipv4/udp_bpf.c
new file mode 100644
index 000000000000..eddd973e6575
--- /dev/null
+++ b/net/ipv4/udp_bpf.c
@@ -0,0 +1,53 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2020 Cloudflare Ltd https://cloudflare.com */
+
+#include <linux/skmsg.h>
+#include <net/sock.h>
+#include <net/udp.h>
+
+enum {
+	UDP_BPF_IPV4,
+	UDP_BPF_IPV6,
+	UDP_BPF_NUM_PROTS,
+};
+
+static struct proto *udpv6_prot_saved __read_mostly;
+static DEFINE_SPINLOCK(udpv6_prot_lock);
+static struct proto udp_bpf_prots[UDP_BPF_NUM_PROTS];
+
+static void udp_bpf_rebuild_protos(struct proto *prot, const struct proto *base)
+{
+	*prot        = *base;
+	prot->unhash = sock_map_unhash;
+	prot->close  = sock_map_close;
+}
+
+static void udp_bpf_check_v6_needs_rebuild(struct sock *sk, struct proto *ops)
+{
+	if (sk->sk_family == AF_INET6 &&
+	    unlikely(ops != smp_load_acquire(&udpv6_prot_saved))) {
+		spin_lock_bh(&udpv6_prot_lock);
+		if (likely(ops != udpv6_prot_saved)) {
+			udp_bpf_rebuild_protos(&udp_bpf_prots[UDP_BPF_IPV6], ops);
+			smp_store_release(&udpv6_prot_saved, ops);
+		}
+		spin_unlock_bh(&udpv6_prot_lock);
+	}
+}
+
+static int __init udp_bpf_v4_build_proto(void)
+{
+	udp_bpf_rebuild_protos(&udp_bpf_prots[UDP_BPF_IPV4], &udp_prot);
+	return 0;
+}
+core_initcall(udp_bpf_v4_build_proto);
+
+struct proto *udp_bpf_get_proto(struct sock *sk, struct sk_psock *psock)
+{
+	int family = sk->sk_family == AF_INET ? UDP_BPF_IPV4 : UDP_BPF_IPV6;
+
+	if (!psock->sk_proto)
+		udp_bpf_check_v6_needs_rebuild(sk, READ_ONCE(sk->sk_prot));
+
+	return &udp_bpf_prots[family];
+}
-- 
2.20.1

[PATCH bpf-next v4 08/12] bpf: sockmap: add UDP support

[Lorenz Bauer]

Allow adding hashed UDP sockets to sockmaps.

Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 net/core/sock_map.c | 37 +++++++++++++++++++++++++++++++++----
 1 file changed, 33 insertions(+), 4 deletions(-)

diff --git a/net/core/sock_map.c b/net/core/sock_map.c
index edfdce17b951..a7075b3b4489 100644
--- a/net/core/sock_map.c
+++ b/net/core/sock_map.c
@@ -11,6 +11,7 @@
 #include <linux/list.h>
 #include <linux/jhash.h>
 #include <linux/sock_diag.h>
+#include <net/udp.h>
 
 struct bpf_stab {
 	struct bpf_map map;
@@ -147,7 +148,19 @@ static int sock_map_init_proto(struct sock *sk, struct sk_psock *psock)
 
 	sock_owned_by_me(sk);
 
-	prot = tcp_bpf_get_proto(sk, psock);
+	switch (sk->sk_type) {
+	case SOCK_STREAM:
+		prot = tcp_bpf_get_proto(sk, psock);
+		break;
+
+	case SOCK_DGRAM:
+		prot = udp_bpf_get_proto(sk, psock);
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
 	if (IS_ERR(prot))
 		return PTR_ERR(prot);
 
@@ -162,7 +175,7 @@ static struct sk_psock *sock_map_psock_get_checked(struct sock *sk)
 	rcu_read_lock();
 	psock = sk_psock(sk);
 	if (psock) {
-		if (sk->sk_prot->recvmsg != tcp_bpf_recvmsg) {
+		if (sk->sk_prot->close != sock_map_close) {
 			psock = ERR_PTR(-EBUSY);
 			goto out;
 		}
@@ -474,15 +487,31 @@ static bool sock_map_op_okay(const struct bpf_sock_ops_kern *ops)
 	       ops->op == BPF_SOCK_OPS_TCP_LISTEN_CB;
 }
 
-static bool sock_map_sk_is_suitable(const struct sock *sk)
+static bool sk_is_tcp(const struct sock *sk)
 {
 	return sk->sk_type == SOCK_STREAM &&
 	       sk->sk_protocol == IPPROTO_TCP;
 }
 
+static bool sk_is_udp(const struct sock *sk)
+{
+	return sk->sk_type == SOCK_DGRAM &&
+	       sk->sk_protocol == IPPROTO_UDP;
+}
+
+static bool sock_map_sk_is_suitable(const struct sock *sk)
+{
+	return sk_is_tcp(sk) || sk_is_udp(sk);
+}
+
 static bool sock_map_sk_state_allowed(const struct sock *sk)
 {
-	return (1 << sk->sk_state) & (TCPF_ESTABLISHED | TCPF_LISTEN);
+	if (sk_is_tcp(sk))
+		return (1 << sk->sk_state) & (TCPF_ESTABLISHED | TCPF_LISTEN);
+	else if (sk_is_udp(sk))
+		return sk_hashed(sk);
+
+	return false;
 }
 
 static int sock_map_update_elem(struct bpf_map *map, void *key,
-- 
2.20.1

[PATCH bpf-next v4 09/12] selftests: bpf: don't listen() on UDP sockets

[Lorenz Bauer]

Most tests for TCP sockmap can be adapted to UDP sockmap if the
listen call is skipped. Rename listen_loopback, etc. to socket_loopback
and skip listen() for SOCK_DGRAM.

Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 .../selftests/bpf/prog_tests/sockmap_listen.c | 47 ++++++++++---------
 1 file changed, 25 insertions(+), 22 deletions(-)

diff --git a/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c b/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
index b1b2acea0638..4ba41dd26d6b 100644
--- a/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
+++ b/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
@@ -230,7 +230,7 @@ static int enable_reuseport(int s, int progfd)
 	return 0;
 }
 
-static int listen_loopback_reuseport(int family, int sotype, int progfd)
+static int socket_loopback_reuseport(int family, int sotype, int progfd)
 {
 	struct sockaddr_storage addr;
 	socklen_t len;
@@ -249,6 +249,9 @@ static int listen_loopback_reuseport(int family, int sotype, int progfd)
 	if (err)
 		goto close;
 
+	if (sotype == SOCK_DGRAM)
+		return s;
+
 	err = xlisten(s, SOMAXCONN);
 	if (err)
 		goto close;
@@ -259,9 +262,9 @@ static int listen_loopback_reuseport(int family, int sotype, int progfd)
 	return -1;
 }
 
-static int listen_loopback(int family, int sotype)
+static int socket_loopback(int family, int sotype)
 {
-	return listen_loopback_reuseport(family, sotype, -1);
+	return socket_loopback_reuseport(family, sotype, -1);
 }
 
 static void test_insert_invalid(int family, int sotype, int mapfd)
@@ -333,7 +336,7 @@ static void test_insert_listening(int family, int sotype, int mapfd)
 	u32 key;
 	int s;
 
-	s = listen_loopback(family, sotype);
+	s = socket_loopback(family, sotype);
 	if (s < 0)
 		return;
 
@@ -349,7 +352,7 @@ static void test_delete_after_insert(int family, int sotype, int mapfd)
 	u32 key;
 	int s;
 
-	s = listen_loopback(family, sotype);
+	s = socket_loopback(family, sotype);
 	if (s < 0)
 		return;
 
@@ -366,7 +369,7 @@ static void test_delete_after_close(int family, int sotype, int mapfd)
 	u64 value;
 	u32 key;
 
-	s = listen_loopback(family, sotype);
+	s = socket_loopback(family, sotype);
 	if (s < 0)
 		return;
 
@@ -390,7 +393,7 @@ static void test_lookup_after_insert(int family, int sotype, int mapfd)
 	u32 key;
 	int s;
 
-	s = listen_loopback(family, sotype);
+	s = socket_loopback(family, sotype);
 	if (s < 0)
 		return;
 
@@ -417,7 +420,7 @@ static void test_lookup_after_delete(int family, int sotype, int mapfd)
 	u64 value;
 	u32 key;
 
-	s = listen_loopback(family, sotype);
+	s = socket_loopback(family, sotype);
 	if (s < 0)
 		return;
 
@@ -439,7 +442,7 @@ static void test_lookup_32_bit_value(int family, int sotype, int mapfd)
 	u32 key, value32;
 	int err, s;
 
-	s = listen_loopback(family, sotype);
+	s = socket_loopback(family, sotype);
 	if (s < 0)
 		return;
 
@@ -470,11 +473,11 @@ static void test_update_listening(int family, int sotype, int mapfd)
 	u64 value;
 	u32 key;
 
-	s1 = listen_loopback(family, sotype);
+	s1 = socket_loopback(family, sotype);
 	if (s1 < 0)
 		return;
 
-	s2 = listen_loopback(family, sotype);
+	s2 = socket_loopback(family, sotype);
 	if (s2 < 0)
 		goto close_s1;
 
@@ -500,7 +503,7 @@ static void test_destroy_orphan_child(int family, int sotype, int mapfd)
 	u64 value;
 	u32 key;
 
-	s = listen_loopback(family, sotype);
+	s = socket_loopback(family, sotype);
 	if (s < 0)
 		return;
 
@@ -534,7 +537,7 @@ static void test_clone_after_delete(int family, int sotype, int mapfd)
 	u64 value;
 	u32 key;
 
-	s = listen_loopback(family, sotype);
+	s = socket_loopback(family, sotype);
 	if (s < 0)
 		return;
 
@@ -570,7 +573,7 @@ static void test_accept_after_delete(int family, int sotype, int mapfd)
 	socklen_t len;
 	u64 value;
 
-	s = listen_loopback(family, sotype);
+	s = socket_loopback(family, sotype);
 	if (s == -1)
 		return;
 
@@ -624,7 +627,7 @@ static void test_accept_before_delete(int family, int sotype, int mapfd)
 	socklen_t len;
 	u64 value;
 
-	s = listen_loopback(family, sotype);
+	s = socket_loopback(family, sotype);
 	if (s == -1)
 		return;
 
@@ -735,7 +738,7 @@ static void test_syn_recv_insert_delete(int family, int sotype, int mapfd)
 	int err, s;
 	u64 value;
 
-	s = listen_loopback(family, sotype | SOCK_NONBLOCK);
+	s = socket_loopback(family, sotype | SOCK_NONBLOCK);
 	if (s < 0)
 		return;
 
@@ -877,7 +880,7 @@ static void redir_to_connected(int family, int sotype, int sock_mapfd,
 
 	zero_verdict_count(verd_mapfd);
 
-	s = listen_loopback(family, sotype | SOCK_NONBLOCK);
+	s = socket_loopback(family, sotype | SOCK_NONBLOCK);
 	if (s < 0)
 		return;
 
@@ -1009,7 +1012,7 @@ static void redir_to_listening(int family, int sotype, int sock_mapfd,
 
 	zero_verdict_count(verd_mapfd);
 
-	s = listen_loopback(family, sotype | SOCK_NONBLOCK);
+	s = socket_loopback(family, sotype | SOCK_NONBLOCK);
 	if (s < 0)
 		return;
 
@@ -1120,7 +1123,7 @@ static void test_reuseport_select_listening(int family, int sotype,
 
 	zero_verdict_count(verd_map);
 
-	s = listen_loopback_reuseport(family, sotype, reuseport_prog);
+	s = socket_loopback_reuseport(family, sotype, reuseport_prog);
 	if (s < 0)
 		return;
 
@@ -1174,7 +1177,7 @@ static void test_reuseport_select_connected(int family, int sotype,
 
 	zero_verdict_count(verd_map);
 
-	s = listen_loopback_reuseport(family, sotype, reuseport_prog);
+	s = socket_loopback_reuseport(family, sotype, reuseport_prog);
 	if (s < 0)
 		return;
 
@@ -1249,11 +1252,11 @@ static void test_reuseport_mixed_groups(int family, int sotype, int sock_map,
 	zero_verdict_count(verd_map);
 
 	/* Create two listeners, each in its own reuseport group */
-	s1 = listen_loopback_reuseport(family, sotype, reuseport_prog);
+	s1 = socket_loopback_reuseport(family, sotype, reuseport_prog);
 	if (s1 < 0)
 		return;
 
-	s2 = listen_loopback_reuseport(family, sotype, reuseport_prog);
+	s2 = socket_loopback_reuseport(family, sotype, reuseport_prog);
 	if (s2 < 0)
 		goto close_srv1;
 
-- 
2.20.1

[PATCH bpf-next v4 10/12] selftests: bpf: add tests for UDP sockets in sockmap

[Lorenz Bauer]

Expand the TCP sockmap test suite to also check UDP sockets.

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 .../selftests/bpf/prog_tests/sockmap_listen.c | 157 ++++++++++++++----
 1 file changed, 127 insertions(+), 30 deletions(-)

diff --git a/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c b/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
index 4ba41dd26d6b..52aa468bdccd 100644
--- a/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
+++ b/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
@@ -108,6 +108,22 @@
 		__ret;                                                         \
 	})
 
+#define xsend(fd, buf, len, flags)                                             \
+	({                                                                     \
+		ssize_t __ret = send((fd), (buf), (len), (flags));             \
+		if (__ret == -1)                                               \
+			FAIL_ERRNO("send");                                    \
+		__ret;                                                         \
+	})
+
+#define xrecv(fd, buf, len, flags)                                             \
+	({                                                                     \
+		ssize_t __ret = recv((fd), (buf), (len), (flags));             \
+		if (__ret == -1)                                               \
+			FAIL_ERRNO("recv");                                    \
+		__ret;                                                         \
+	})
+
 #define xsocket(family, sotype, flags)                                         \
 	({                                                                     \
 		int __ret = socket(family, sotype, flags);                     \
@@ -330,7 +346,7 @@ static void test_insert_bound(int family, int sotype, int mapfd)
 	xclose(s);
 }
 
-static void test_insert_listening(int family, int sotype, int mapfd)
+static void test_insert(int family, int sotype, int mapfd)
 {
 	u64 value;
 	u32 key;
@@ -467,7 +483,7 @@ static void test_lookup_32_bit_value(int family, int sotype, int mapfd)
 	xclose(s);
 }
 
-static void test_update_listening(int family, int sotype, int mapfd)
+static void test_update_existing(int family, int sotype, int mapfd)
 {
 	int s1, s2;
 	u64 value;
@@ -1116,7 +1132,7 @@ static void test_reuseport_select_listening(int family, int sotype,
 {
 	struct sockaddr_storage addr;
 	unsigned int pass;
-	int s, c, p, err;
+	int s, c, err;
 	socklen_t len;
 	u64 value;
 	u32 key;
@@ -1145,19 +1161,33 @@ static void test_reuseport_select_listening(int family, int sotype,
 	if (err)
 		goto close_cli;
 
-	p = xaccept(s, NULL, NULL);
-	if (p < 0)
-		goto close_cli;
+	if (sotype == SOCK_STREAM) {
+		int p;
+
+		p = xaccept(s, NULL, NULL);
+		if (p < 0)
+			goto close_cli;
+		xclose(p);
+	} else {
+		char b = 'a';
+		ssize_t n;
+
+		n = xsend(c, &b, sizeof(b), 0);
+		if (n == -1)
+			goto close_cli;
+
+		n = xrecv(s, &b, sizeof(b), 0);
+		if (n == -1)
+			goto close_cli;
+	}
 
 	key = SK_PASS;
 	err = xbpf_map_lookup_elem(verd_map, &key, &pass);
 	if (err)
-		goto close_peer;
+		goto close_cli;
 	if (pass != 1)
 		FAIL("want pass count 1, have %d", pass);
 
-close_peer:
-	xclose(p);
 close_cli:
 	xclose(c);
 close_srv:
@@ -1201,9 +1231,24 @@ static void test_reuseport_select_connected(int family, int sotype,
 	if (err)
 		goto close_cli0;
 
-	p0 = xaccept(s, NULL, NULL);
-	if (err)
-		goto close_cli0;
+	if (sotype == SOCK_STREAM) {
+		p0 = xaccept(s, NULL, NULL);
+		if (p0 < 0)
+			goto close_cli0;
+	} else {
+		p0 = xsocket(family, sotype, 0);
+		if (p0 < 0)
+			goto close_cli0;
+
+		len = sizeof(addr);
+		err = xgetsockname(c0, sockaddr(&addr), &len);
+		if (err)
+			goto close_cli0;
+
+		err = xconnect(p0, sockaddr(&addr), len);
+		if (err)
+			goto close_cli0;
+	}
 
 	/* Update sock_map[0] to redirect to a connected socket */
 	key = 0;
@@ -1216,8 +1261,24 @@ static void test_reuseport_select_connected(int family, int sotype,
 	if (c1 < 0)
 		goto close_peer0;
 
+	len = sizeof(addr);
+	err = xgetsockname(s, sockaddr(&addr), &len);
+	if (err)
+		goto close_srv;
+
 	errno = 0;
 	err = connect(c1, sockaddr(&addr), len);
+	if (sotype == SOCK_DGRAM) {
+		char b = 'a';
+		ssize_t n;
+
+		n = xsend(c1, &b, sizeof(b), 0);
+		if (n == -1)
+			goto close_cli1;
+
+		n = recv(c1, &b, sizeof(b), 0);
+		err = n == -1;
+	}
 	if (!err || errno != ECONNREFUSED)
 		FAIL_ERRNO("connect: expected ECONNREFUSED");
 
@@ -1281,7 +1342,18 @@ static void test_reuseport_mixed_groups(int family, int sotype, int sock_map,
 		goto close_srv2;
 
 	err = connect(c, sockaddr(&addr), len);
-	if (err && errno != ECONNREFUSED) {
+	if (sotype == SOCK_DGRAM) {
+		char b = 'a';
+		ssize_t n;
+
+		n = xsend(c, &b, sizeof(b), 0);
+		if (n == -1)
+			goto close_cli;
+
+		n = recv(c, &b, sizeof(b), 0);
+		err = n == -1;
+	}
+	if (!err || errno != ECONNREFUSED) {
 		FAIL_ERRNO("connect: expected ECONNREFUSED");
 		goto close_cli;
 	}
@@ -1302,9 +1374,9 @@ static void test_reuseport_mixed_groups(int family, int sotype, int sock_map,
 	xclose(s1);
 }
 
-#define TEST(fn)                                                               \
+#define TEST(fn, ...)                                                          \
 	{                                                                      \
-		fn, #fn                                                        \
+		fn, #fn, __VA_ARGS__                                           \
 	}
 
 static void test_ops_cleanup(const struct bpf_map *map)
@@ -1353,18 +1425,31 @@ static const char *map_type_str(const struct bpf_map *map)
 	}
 }
 
+static const char *sotype_str(int sotype)
+{
+	switch (sotype) {
+	case SOCK_DGRAM:
+		return "UDP";
+	case SOCK_STREAM:
+		return "TCP";
+	default:
+		return "unknown";
+	}
+}
+
 static void test_ops(struct test_sockmap_listen *skel, struct bpf_map *map,
 		     int family, int sotype)
 {
 	const struct op_test {
 		void (*fn)(int family, int sotype, int mapfd);
 		const char *name;
+		int sotype;
 	} tests[] = {
 		/* insert */
 		TEST(test_insert_invalid),
 		TEST(test_insert_opened),
-		TEST(test_insert_bound),
-		TEST(test_insert_listening),
+		TEST(test_insert_bound, SOCK_STREAM),
+		TEST(test_insert),
 		/* delete */
 		TEST(test_delete_after_insert),
 		TEST(test_delete_after_close),
@@ -1373,28 +1458,32 @@ static void test_ops(struct test_sockmap_listen *skel, struct bpf_map *map,
 		TEST(test_lookup_after_delete),
 		TEST(test_lookup_32_bit_value),
 		/* update */
-		TEST(test_update_listening),
+		TEST(test_update_existing),
 		/* races with insert/delete */
-		TEST(test_destroy_orphan_child),
-		TEST(test_syn_recv_insert_delete),
-		TEST(test_race_insert_listen),
+		TEST(test_destroy_orphan_child, SOCK_STREAM),
+		TEST(test_syn_recv_insert_delete, SOCK_STREAM),
+		TEST(test_race_insert_listen, SOCK_STREAM),
 		/* child clone */
-		TEST(test_clone_after_delete),
-		TEST(test_accept_after_delete),
-		TEST(test_accept_before_delete),
+		TEST(test_clone_after_delete, SOCK_STREAM),
+		TEST(test_accept_after_delete, SOCK_STREAM),
+		TEST(test_accept_before_delete, SOCK_STREAM),
 	};
-	const char *family_name, *map_name;
+	const char *family_name, *map_name, *sotype_name;
 	const struct op_test *t;
 	char s[MAX_TEST_NAME];
 	int map_fd;
 
 	family_name = family_str(family);
 	map_name = map_type_str(map);
+	sotype_name = sotype_str(sotype);
 	map_fd = bpf_map__fd(map);
 
 	for (t = tests; t < tests + ARRAY_SIZE(tests); t++) {
-		snprintf(s, sizeof(s), "%s %s %s", map_name, family_name,
-			 t->name);
+		snprintf(s, sizeof(s), "%s %s %s %s", map_name, family_name,
+			 sotype_name, t->name);
+
+		if (t->sotype != 0 && t->sotype != sotype)
+			continue;
 
 		if (!test__start_subtest(s))
 			continue;
@@ -1427,6 +1516,7 @@ static void test_redir(struct test_sockmap_listen *skel, struct bpf_map *map,
 	for (t = tests; t < tests + ARRAY_SIZE(tests); t++) {
 		snprintf(s, sizeof(s), "%s %s %s", map_name, family_name,
 			 t->name);
+
 		if (!test__start_subtest(s))
 			continue;
 
@@ -1441,26 +1531,31 @@ static void test_reuseport(struct test_sockmap_listen *skel,
 		void (*fn)(int family, int sotype, int socket_map,
 			   int verdict_map, int reuseport_prog);
 		const char *name;
+		int sotype;
 	} tests[] = {
 		TEST(test_reuseport_select_listening),
 		TEST(test_reuseport_select_connected),
 		TEST(test_reuseport_mixed_groups),
 	};
 	int socket_map, verdict_map, reuseport_prog;
-	const char *family_name, *map_name;
+	const char *family_name, *map_name, *sotype_name;
 	const struct reuseport_test *t;
 	char s[MAX_TEST_NAME];
 
 	family_name = family_str(family);
 	map_name = map_type_str(map);
+	sotype_name = sotype_str(sotype);
 
 	socket_map = bpf_map__fd(map);
 	verdict_map = bpf_map__fd(skel->maps.verdict_map);
 	reuseport_prog = bpf_program__fd(skel->progs.prog_reuseport);
 
 	for (t = tests; t < tests + ARRAY_SIZE(tests); t++) {
-		snprintf(s, sizeof(s), "%s %s %s", map_name, family_name,
-			 t->name);
+		snprintf(s, sizeof(s), "%s %s %s %s", map_name, family_name,
+			 sotype_name, t->name);
+
+		if (t->sotype != 0 && t->sotype != sotype)
+			continue;
 
 		if (!test__start_subtest(s))
 			continue;
@@ -1473,8 +1568,10 @@ static void run_tests(struct test_sockmap_listen *skel, struct bpf_map *map,
 		      int family)
 {
 	test_ops(skel, map, family, SOCK_STREAM);
+	test_ops(skel, map, family, SOCK_DGRAM);
 	test_redir(skel, map, family, SOCK_STREAM);
 	test_reuseport(skel, map, family, SOCK_STREAM);
+	test_reuseport(skel, map, family, SOCK_DGRAM);
 }
 
 void test_sockmap_listen(void)
-- 
2.20.1

[PATCH bpf-next v4 11/12] selftests: bpf: enable UDP sockmap reuseport tests

[Lorenz Bauer]

Remove the guard that disables UDP tests now that sockmap
has support for them.

Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 tools/testing/selftests/bpf/prog_tests/select_reuseport.c | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/tools/testing/selftests/bpf/prog_tests/select_reuseport.c b/tools/testing/selftests/bpf/prog_tests/select_reuseport.c
index a1dd13b34d4b..821b4146b7b6 100644
--- a/tools/testing/selftests/bpf/prog_tests/select_reuseport.c
+++ b/tools/testing/selftests/bpf/prog_tests/select_reuseport.c
@@ -805,12 +805,6 @@ static void test_config(int sotype, sa_family_t family, bool inany)
 	char s[MAX_TEST_NAME];
 	const struct test *t;
 
-	/* SOCKMAP/SOCKHASH don't support UDP yet */
-	if (sotype == SOCK_DGRAM &&
-	    (inner_map_type == BPF_MAP_TYPE_SOCKMAP ||
-	     inner_map_type == BPF_MAP_TYPE_SOCKHASH))
-		return;
-
 	for (t = tests; t < tests + ARRAY_SIZE(tests); t++) {
 		if (t->need_sotype && t->need_sotype != sotype)
 			continue; /* test not compatible with socket type */
-- 
2.20.1

[PATCH bpf-next v4 12/12] bpf, doc: update maintainers for L7 BPF

[Lorenz Bauer]

Add Jakub and myself as maintainers for sockmap related code.

Signed-off-by: Lorenz Bauer <lmb@cloudflare.com>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
---
 MAINTAINERS | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index 14554bde1c06..adc7fa8e5880 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -9363,6 +9363,8 @@ F:	include/net/l3mdev.h
 L7 BPF FRAMEWORK
 M:	John Fastabend <john.fastabend@gmail.com>
 M:	Daniel Borkmann <daniel@iogearbox.net>
+M:	Jakub Sitnicki <jakub@cloudflare.com>
+M:	Lorenz Bauer <lmb@cloudflare.com>
 L:	netdev@vger.kernel.org
 L:	bpf@vger.kernel.org
 S:	Maintained
-- 
2.20.1

Re: [PATCH bpf-next v4 00/12] bpf: sockmap, sockhash: support storing UDP sockets

[Daniel Borkmann]

On 3/9/20 12:12 PM, Lorenz Bauer wrote:
> I've adressed John's nit in patch 3, and added the reviews and acks.
> 
> Changes since v3:
> - Clarify !psock check in sock_map_link_no_progs
> 
> Changes since v2:
> - Remove sk_psock_hooks based on Jakub's idea
> - Fix reference to tcp_bpf_clone in commit message
> - Add inet_csk_has_ulp helper
> 
> Changes since v1:
> - Check newsk->sk_prot in tcp_bpf_clone
> - Fix compilation with BPF_STREAM_PARSER disabled
> - Use spin_lock_init instead of static initializer
> - Elaborate on TCPF_SYN_RECV
> - Cosmetic changes to TEST macros, and more tests
> - Add Jakub and me as maintainers
> 
> Lorenz Bauer (12):
>    bpf: sockmap: only check ULP for TCP sockets
>    skmsg: update saved hooks only once
>    bpf: tcp: move assertions into tcp_bpf_get_proto
>    bpf: tcp: guard declarations with CONFIG_NET_SOCK_MSG
>    bpf: sockmap: move generic sockmap hooks from BPF TCP
>    bpf: sockmap: simplify sock_map_init_proto
>    bpf: add sockmap hooks for UDP sockets
>    bpf: sockmap: add UDP support
>    selftests: bpf: don't listen() on UDP sockets
>    selftests: bpf: add tests for UDP sockets in sockmap
>    selftests: bpf: enable UDP sockmap reuseport tests
>    bpf, doc: update maintainers for L7 BPF
> 
>   MAINTAINERS                                   |   3 +
>   include/linux/bpf.h                           |   4 +-
>   include/linux/skmsg.h                         |  56 ++---
>   include/net/inet_connection_sock.h            |   6 +
>   include/net/tcp.h                             |  20 +-
>   include/net/udp.h                             |   5 +
>   net/core/sock_map.c                           | 157 +++++++++++---
>   net/ipv4/Makefile                             |   1 +
>   net/ipv4/tcp_bpf.c                            | 114 ++--------
>   net/ipv4/tcp_ulp.c                            |   7 -
>   net/ipv4/udp_bpf.c                            |  53 +++++
>   .../bpf/prog_tests/select_reuseport.c         |   6 -
>   .../selftests/bpf/prog_tests/sockmap_listen.c | 204 +++++++++++++-----
>   13 files changed, 402 insertions(+), 234 deletions(-)
>   create mode 100644 net/ipv4/udp_bpf.c
> 

Applied, thanks!