Skip to contentSkip to conversation

Allow updating sockmap / sockhash from BPF

PATCH In review Lorenz Bauer 3 participants 11 messages Aug 21, 2020

Subsystems (9)
lore
Share & tools
lore
#0 lore raw

v3 cover Allow updating sockmap / sockhash from BPF

We're currently building a control plane for our BPF socket dispatch work. As part of that, we have a need to create a copy of an existing sockhash, to allow us to change the keys. I previously proposed allowing privileged userspace to look up sockets, which doesn't work due to security concerns (see [1]).

In follow up discussions during BPF office hours we identified bpf_iter as a possible solution: instead of accessing sockets from user space we can iterate the source sockhash, and insert the values into a new map. Enabling this requires two pieces: the ability to iterate sockmap and sockhash, as well as being able to call map_update_elem from BPF.

This patch set implements the latter: it's now possible to update sockmap from BPF context. As a next step, we can implement bpf_iter for sockmap.

===

I've done some more fixups, and audited the safe contexts more thoroughly. As a result I'm removing CGROUP_SKB, SK_MSG and SK_SKB for now.

Changes in v3:
- Use CHECK as much as possible (Yonghong)
- Reject ARG_PTR_TO_MAP_VALUE_OR_NULL for sockmap (Yonghong)
- Remove CGROUP_SKB, SK_MSG, SK_SKB from safe contexts
- Test that the verifier rejects update from unsafe context

Changes in v2:
- Fix warning in patch #2 (Jakub K)
- Renamed override_map_arg_type (John)
- Only allow updating sockmap from known safe contexts (John)
- Use __s64 for sockmap updates from user space (Yonghong)
- Various small test fixes around test macros and such (Yonghong)

Thank your for your reviews!

1: https://lore.kernel.org/bpf/[email protected]/

Lorenz Bauer (6):
net: sk_msg: simplify sk_psock initialization
bpf: sockmap: merge sockmap and sockhash update functions
bpf: sockmap: call sock_map_update_elem directly
bpf: override the meaning of ARG_PTR_TO_MAP_VALUE for sockmap and
sockhash
bpf: sockmap: allow update from BPF
selftests: bpf: test sockmap update from BPF

 include/linux/bpf.h                           |   7 +
 include/linux/skmsg.h                         |  17 ---
 kernel/bpf/syscall.c                          |   5 +-
 kernel/bpf/verifier.c                         |  73 +++++++++-
 net/core/skmsg.c                              |  34 ++++-
 net/core/sock_map.c                           | 137 ++++++++----------
 net/ipv4/tcp_bpf.c                            |  13 +-
 net/ipv4/udp_bpf.c                            |   9 +-
 .../selftests/bpf/prog_tests/sockmap_basic.c  |  78 ++++++++++
 .../bpf/progs/test_sockmap_invalid_update.c   |  23 +++
 .../selftests/bpf/progs/test_sockmap_update.c |  48 ++++++
 11 files changed, 326 insertions(+), 118 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/progs/test_sockmap_invalid_update.c
 create mode 100644 tools/testing/selftests/bpf/progs/test_sockmap_update.c
Signature · 1 line
2.25.1
#1 lore raw

PATCH v3 1/6 net: sk_msg: simplify sk_psock initialization

Initializing psock->sk_proto and other saved callbacks is only done in sk_psock_update_proto, after sk_psock_init has returned. The logic for this is difficult to follow, and needlessly complex.

Instead, initialize psock->sk_proto whenever we allocate a new psock. Additionally, assert the following invariants:

* The SK has no ULP: ULP does it's own finagling of sk->sk_prot
* sk_user_data is unused: we need it to store sk_psock

Protect our access to sk_user_data with sk_callback_lock, which is what other users like reuseport arrays, etc. do.

The result is that an sk_psock is always fully initialized, and that psock->sk_proto is always the "original" struct proto. The latter allows us to use psock->sk_proto when initializing IPv6 TCP / UDP callbacks for sockmap.

---
 include/linux/skmsg.h | 17 -----------------
 net/core/skmsg.c      | 34 ++++++++++++++++++++++++++++------
 net/core/sock_map.c   | 14 ++++----------
 net/ipv4/tcp_bpf.c    | 13 +++++--------
 net/ipv4/udp_bpf.c    |  9 ++++-----
 5 files changed, 41 insertions(+), 46 deletions(-)
include/linux/skmsg.h +0 -17
index 1e9ed840b9fc..3119928fc103 100644
@@ -340,23 +340,6 @@ static inline void sk_psock_update_proto(struct sock *sk,
 					 struct sk_psock *psock,
 					 struct proto *ops)
 {
-	/* 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;
-	}
-
 	/* Pairs with lockless read in sk_clone_lock() */
 	WRITE_ONCE(sk->sk_prot, ops);
 }
net/core/skmsg.c +28 -6
index 6a32a1fd34f8..1c81caf9630f 100644
@@ -494,14 +494,34 @@ static void sk_psock_backlog(struct work_struct *work)
 
 struct sk_psock *sk_psock_init(struct sock *sk, int node)
 {
-	struct sk_psock *psock = kzalloc_node(sizeof(*psock),
-					      GFP_ATOMIC | __GFP_NOWARN,
-					      node);
-	if (!psock)
-		return NULL;
+	struct sk_psock *psock;
+	struct proto *prot;
 
+	write_lock_bh(&sk->sk_callback_lock);
+
+	if (inet_csk_has_ulp(sk)) {
+		psock = ERR_PTR(-EINVAL);
+		goto out;
+	}
+
+	if (sk->sk_user_data) {
+		psock = ERR_PTR(-EBUSY);
+		goto out;
+	}
+
+	psock = kzalloc_node(sizeof(*psock), GFP_ATOMIC | __GFP_NOWARN, node);
+	if (!psock) {
+		psock = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+
+	prot = READ_ONCE(sk->sk_prot);
 	psock->sk = sk;
-	psock->eval =  __SK_NONE;
+	psock->eval = __SK_NONE;
+	psock->sk_proto = prot;
+	psock->saved_unhash = prot->unhash;
+	psock->saved_close = prot->close;
+	psock->saved_write_space = sk->sk_write_space;
 
 	INIT_LIST_HEAD(&psock->link);
 	spin_lock_init(&psock->link_lock);
@@ -516,6 +536,8 @@ struct sk_psock *sk_psock_init(struct sock *sk, int node)
 	rcu_assign_sk_user_data_nocopy(sk, psock);
 	sock_hold(sk);
 
+out:
+	write_unlock_bh(&sk->sk_callback_lock);
 	return psock;
 }
 EXPORT_SYMBOL_GPL(sk_psock_init);
net/core/sock_map.c +4 -10
index 119f52a99dc1..abe4bac40db9 100644
@@ -184,8 +184,6 @@ static int sock_map_init_proto(struct sock *sk, struct sk_psock *psock)
 {
 	struct proto *prot;
 
-	sock_owned_by_me(sk);
-
 	switch (sk->sk_type) {
 	case SOCK_STREAM:
 		prot = tcp_bpf_get_proto(sk, psock);
@@ -272,8 +270,8 @@ static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
 		}
 	} else {
 		psock = sk_psock_init(sk, map->numa_node);
-		if (!psock) {
-			ret = -ENOMEM;
+		if (IS_ERR(psock)) {
+			ret = PTR_ERR(psock);
 			goto out_progs;
 		}
 	}
@@ -322,8 +320,8 @@ static int sock_map_link_no_progs(struct bpf_map *map, struct sock *sk)
 
 	if (!psock) {
 		psock = sk_psock_init(sk, map->numa_node);
-		if (!psock)
-			return -ENOMEM;
+		if (IS_ERR(psock))
+			return PTR_ERR(psock);
 	}
 
 	ret = sock_map_init_proto(sk, psock);
@@ -478,8 +476,6 @@ static int sock_map_update_common(struct bpf_map *map, u32 idx,
 		return -EINVAL;
 	if (unlikely(idx >= map->max_entries))
 		return -E2BIG;
-	if (inet_csk_has_ulp(sk))
-		return -EINVAL;
 
 	link = sk_psock_init_link();
 	if (!link)
@@ -855,8 +851,6 @@ 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 (inet_csk_has_ulp(sk))
-		return -EINVAL;
 
 	link = sk_psock_init_link();
 	if (!link)
net/ipv4/tcp_bpf.c +5 -8
index 7aa68f4aae6c..37f4cb2bba5c 100644
@@ -567,10 +567,9 @@ static void tcp_bpf_rebuild_protos(struct proto prot[TCP_BPF_NUM_CFGS],
 	prot[TCP_BPF_TX].sendpage		= tcp_bpf_sendpage;
 }
 
-static void tcp_bpf_check_v6_needs_rebuild(struct sock *sk, struct proto *ops)
+static void tcp_bpf_check_v6_needs_rebuild(struct proto *ops)
 {
-	if (sk->sk_family == AF_INET6 &&
-	    unlikely(ops != smp_load_acquire(&tcpv6_prot_saved))) {
+	if (unlikely(ops != smp_load_acquire(&tcpv6_prot_saved))) {
 		spin_lock_bh(&tcpv6_prot_lock);
 		if (likely(ops != tcpv6_prot_saved)) {
 			tcp_bpf_rebuild_protos(tcp_bpf_prots[TCP_BPF_IPV6], ops);
@@ -603,13 +602,11 @@ 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;
 
-	if (!psock->sk_proto) {
-		struct proto *ops = READ_ONCE(sk->sk_prot);
-
-		if (tcp_bpf_assert_proto_ops(ops))
+	if (sk->sk_family == AF_INET6) {
+		if (tcp_bpf_assert_proto_ops(psock->sk_proto))
 			return ERR_PTR(-EINVAL);
 
-		tcp_bpf_check_v6_needs_rebuild(sk, ops);
+		tcp_bpf_check_v6_needs_rebuild(psock->sk_proto);
 	}
 
 	return &tcp_bpf_prots[family][config];
net/ipv4/udp_bpf.c +4 -5
index eddd973e6575..7a94791efc1a 100644
@@ -22,10 +22,9 @@ static void udp_bpf_rebuild_protos(struct proto *prot, const struct proto *base)
 	prot->close  = sock_map_close;
 }
 
-static void udp_bpf_check_v6_needs_rebuild(struct sock *sk, struct proto *ops)
+static void udp_bpf_check_v6_needs_rebuild(struct proto *ops)
 {
-	if (sk->sk_family == AF_INET6 &&
-	    unlikely(ops != smp_load_acquire(&udpv6_prot_saved))) {
+	if (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);
@@ -46,8 +45,8 @@ 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));
+	if (sk->sk_family == AF_INET6)
+		udp_bpf_check_v6_needs_rebuild(psock->sk_proto);
 
 	return &udp_bpf_prots[family];
 }
Signature · 1 line
2.25.1
#2 lore raw

PATCH v3 2/6 bpf: sockmap: merge sockmap and sockhash update functions

In reply to Lorenz Bauer

Merge the two very similar functions sock_map_update_elem and sock_hash_update_elem into one.

---
 net/core/sock_map.c | 49 +++++++--------------------------------------
 1 file changed, 7 insertions(+), 42 deletions(-)
net/core/sock_map.c +7 -42
index abe4bac40db9..905e2dd765aa 100644
@@ -559,10 +559,12 @@ static bool sock_map_sk_state_allowed(const struct sock *sk)
 	return false;
 }
 
+static int sock_hash_update_common(struct bpf_map *map, void *key,
+				   struct sock *sk, u64 flags);
+
 static int sock_map_update_elem(struct bpf_map *map, void *key,
 				void *value, u64 flags)
 {
-	u32 idx = *(u32 *)key;
 	struct socket *sock;
 	struct sock *sk;
 	int ret;
@@ -591,8 +593,10 @@ static int sock_map_update_elem(struct bpf_map *map, void *key,
 	sock_map_sk_acquire(sk);
 	if (!sock_map_sk_state_allowed(sk))
 		ret = -EOPNOTSUPP;
+	else if (map->map_type == BPF_MAP_TYPE_SOCKMAP)
+		ret = sock_map_update_common(map, *(u32 *)key, sk, flags);
 	else
-		ret = sock_map_update_common(map, idx, sk, flags);
+		ret = sock_hash_update_common(map, key, sk, flags);
 	sock_map_sk_release(sk);
 out:
 	fput(sock->file);
@@ -909,45 +913,6 @@ static int sock_hash_update_common(struct bpf_map *map, void *key,
 	return ret;
 }
 
-static int sock_hash_update_elem(struct bpf_map *map, void *key,
-				 void *value, u64 flags)
-{
-	struct socket *sock;
-	struct sock *sk;
-	int ret;
-	u64 ufd;
-
-	if (map->value_size == sizeof(u64))
-		ufd = *(u64 *)value;
-	else
-		ufd = *(u32 *)value;
-	if (ufd > S32_MAX)
-		return -EINVAL;
-
-	sock = sockfd_lookup(ufd, &ret);
-	if (!sock)
-		return ret;
-	sk = sock->sk;
-	if (!sk) {
-		ret = -EINVAL;
-		goto out;
-	}
-	if (!sock_map_sk_is_suitable(sk)) {
-		ret = -EOPNOTSUPP;
-		goto out;
-	}
-
-	sock_map_sk_acquire(sk);
-	if (!sock_map_sk_state_allowed(sk))
-		ret = -EOPNOTSUPP;
-	else
-		ret = sock_hash_update_common(map, key, sk, flags);
-	sock_map_sk_release(sk);
-out:
-	fput(sock->file);
-	return ret;
-}
-
 static int sock_hash_get_next_key(struct bpf_map *map, void *key,
 				  void *key_next)
 {
@@ -1216,7 +1181,7 @@ const struct bpf_map_ops sock_hash_ops = {
 	.map_alloc		= sock_hash_alloc,
 	.map_free		= sock_hash_free,
 	.map_get_next_key	= sock_hash_get_next_key,
-	.map_update_elem	= sock_hash_update_elem,
+	.map_update_elem	= sock_map_update_elem,
 	.map_delete_elem	= sock_hash_delete_elem,
 	.map_lookup_elem	= sock_hash_lookup,
 	.map_lookup_elem_sys_only = sock_hash_lookup_sys,
Signature · 1 line
2.25.1
#3 lore raw

PATCH v3 3/6 bpf: sockmap: call sock_map_update_elem directly

In reply to Lorenz Bauer

Don't go via map->ops to call sock_map_update_elem, since we know what function to call in bpf_map_update_value. Since we currently don't allow calling map_update_elem from BPF context, we can remove ops->map_update_elem and rename the function to sock_map_update_elem_sys.

---
 include/linux/bpf.h  | 7 +++++++
 kernel/bpf/syscall.c | 5 +++--
 net/core/sock_map.c  | 6 ++----
 3 files changed, 12 insertions(+), 6 deletions(-)
include/linux/bpf.h +7 -0
index cef4ef0d2b4e..cf3416d1b8c2 100644
@@ -1635,6 +1635,7 @@ int sock_map_prog_update(struct bpf_map *map, struct bpf_prog *prog,
 			 struct bpf_prog *old, u32 which);
 int sock_map_get_from_fd(const union bpf_attr *attr, struct bpf_prog *prog);
 int sock_map_prog_detach(const union bpf_attr *attr, enum bpf_prog_type ptype);
+int sock_map_update_elem_sys(struct bpf_map *map, void *key, void *value, u64 flags);
 void sock_map_unhash(struct sock *sk);
 void sock_map_close(struct sock *sk, long timeout);
 #else
@@ -1656,6 +1657,12 @@ static inline int sock_map_prog_detach(const union bpf_attr *attr,
 {
 	return -EOPNOTSUPP;
 }
+
+static inline int sock_map_update_elem_sys(struct bpf_map *map, void *key, void *value,
+					   u64 flags)
+{
+	return -EOPNOTSUPP;
+}
 #endif /* CONFIG_BPF_STREAM_PARSER */
 
 #if defined(CONFIG_INET) && defined(CONFIG_BPF_SYSCALL)
kernel/bpf/syscall.c +3 -2
index 2f343ce15747..5867cf615a3c 100644
@@ -157,10 +157,11 @@ static int bpf_map_update_value(struct bpf_map *map, struct fd f, void *key,
 	if (bpf_map_is_dev_bound(map)) {
 		return bpf_map_offload_update_elem(map, key, value, flags);
 	} else if (map->map_type == BPF_MAP_TYPE_CPUMAP ||
-		   map->map_type == BPF_MAP_TYPE_SOCKHASH ||
-		   map->map_type == BPF_MAP_TYPE_SOCKMAP ||
 		   map->map_type == BPF_MAP_TYPE_STRUCT_OPS) {
 		return map->ops->map_update_elem(map, key, value, flags);
+	} else if (map->map_type == BPF_MAP_TYPE_SOCKHASH ||
+		   map->map_type == BPF_MAP_TYPE_SOCKMAP) {
+		return sock_map_update_elem_sys(map, key, value, flags);
 	} else if (IS_FD_PROG_ARRAY(map)) {
 		return bpf_fd_array_map_update_elem(map, f.file, key, value,
 						    flags);
net/core/sock_map.c +2 -4
index 905e2dd765aa..48e83f93ee66 100644
@@ -562,8 +562,8 @@ static bool sock_map_sk_state_allowed(const struct sock *sk)
 static int sock_hash_update_common(struct bpf_map *map, void *key,
 				   struct sock *sk, u64 flags);
 
-static int sock_map_update_elem(struct bpf_map *map, void *key,
-				void *value, u64 flags)
+int sock_map_update_elem_sys(struct bpf_map *map, void *key, void *value,
+			     u64 flags)
 {
 	struct socket *sock;
 	struct sock *sk;
@@ -687,7 +687,6 @@ const struct bpf_map_ops sock_map_ops = {
 	.map_free		= sock_map_free,
 	.map_get_next_key	= sock_map_get_next_key,
 	.map_lookup_elem_sys_only = sock_map_lookup_sys,
-	.map_update_elem	= sock_map_update_elem,
 	.map_delete_elem	= sock_map_delete_elem,
 	.map_lookup_elem	= sock_map_lookup,
 	.map_release_uref	= sock_map_release_progs,
@@ -1181,7 +1180,6 @@ const struct bpf_map_ops sock_hash_ops = {
 	.map_alloc		= sock_hash_alloc,
 	.map_free		= sock_hash_free,
 	.map_get_next_key	= sock_hash_get_next_key,
-	.map_update_elem	= sock_map_update_elem,
 	.map_delete_elem	= sock_hash_delete_elem,
 	.map_lookup_elem	= sock_hash_lookup,
 	.map_lookup_elem_sys_only = sock_hash_lookup_sys,
Signature · 1 line
2.25.1
#4 lore raw

PATCH v3 4/6 bpf: override the meaning of ARG_PTR_TO_MAP_VALUE for sockmap and sockhash

In reply to Lorenz Bauer

The verifier assumes that map values are simple blobs of memory, and therefore treats ARG_PTR_TO_MAP_VALUE, etc. as such. However, there are map types where this isn't true. For example, sockmap and sockhash store sockets. In general this isn't a big problem: we can just write helpers that explicitly requests PTR_TO_SOCKET instead of ARG_PTR_TO_MAP_VALUE.

The one exception are the standard map helpers like map_update_elem, map_lookup_elem, etc. Here it would be nice we could overload the function prototype for different kinds of maps. Unfortunately, this isn't entirely straight forward:
We only know the type of the map once we have resolved meta->map_ptr in check_func_arg. This means we can't swap out the prototype in check_helper_call until we're half way through the function.

Instead, modify check_func_arg to treat ARG_PTR_TO_MAP_VALUE to mean "the native type for the map" instead of "pointer to memory" for sockmap and sockhash. This means we don't have to modify the function prototype at all

---
 kernel/bpf/verifier.c | 35 +++++++++++++++++++++++++++++++++++
 1 file changed, 35 insertions(+)
kernel/bpf/verifier.c +35 -0
index b6ccfce3bf4c..7e15866c5184 100644
@@ -3872,6 +3872,33 @@ static int int_ptr_type_to_size(enum bpf_arg_type type)
 	return -EINVAL;
 }
 
+static int resolve_map_arg_type(struct bpf_verifier_env *env,
+				 const struct bpf_call_arg_meta *meta,
+				 enum bpf_arg_type *arg_type)
+{
+	if (!meta->map_ptr) {
+		/* kernel subsystem misconfigured verifier */
+		verbose(env, "invalid map_ptr to access map->type\n");
+		return -EACCES;
+	}
+
+	switch (meta->map_ptr->map_type) {
+	case BPF_MAP_TYPE_SOCKMAP:
+	case BPF_MAP_TYPE_SOCKHASH:
+		if (*arg_type == ARG_PTR_TO_MAP_VALUE) {
+			*arg_type = ARG_PTR_TO_SOCKET;
+		} else {
+			verbose(env, "invalid arg_type for sockmap/sockhash\n");
+			return -EINVAL;
+		}
+		break;
+
+	default:
+		break;
+	}
+	return 0;
+}
+
 static int check_func_arg(struct bpf_verifier_env *env, u32 arg,
 			  struct bpf_call_arg_meta *meta,
 			  const struct bpf_func_proto *fn)
@@ -3904,6 +3931,14 @@ static int check_func_arg(struct bpf_verifier_env *env, u32 arg,
 		return -EACCES;
 	}
 
+	if (arg_type == ARG_PTR_TO_MAP_VALUE ||
+	    arg_type == ARG_PTR_TO_UNINIT_MAP_VALUE ||
+	    arg_type == ARG_PTR_TO_MAP_VALUE_OR_NULL) {
+		err = resolve_map_arg_type(env, meta, &arg_type);
+		if (err)
+			return err;
+	}
+
 	if (arg_type == ARG_PTR_TO_MAP_KEY ||
 	    arg_type == ARG_PTR_TO_MAP_VALUE ||
 	    arg_type == ARG_PTR_TO_UNINIT_MAP_VALUE ||
Signature · 1 line
2.25.1
Acked-by on v3 4/6
#5 lore raw

Re: [PATCH bpf-next v3 4/6] bpf: override the meaning of ARG_PTR_TO_MAP_VALUE for sockmap and sockhash

Quoted 15 lines · Lorenz Bauer

The verifier assumes that map values are simple blobs of memory, and therefore treats ARG_PTR_TO_MAP_VALUE, etc. as such. However, there are map types where this isn't true. For example, sockmap and sockhash store sockets. In general this isn't a big problem: we can just write helpers that explicitly requests PTR_TO_SOCKET instead of ARG_PTR_TO_MAP_VALUE.

The one exception are the standard map helpers like map_update_elem, map_lookup_elem, etc. Here it would be nice we could overload the function prototype for different kinds of maps. Unfortunately, this isn't entirely straight forward:
We only know the type of the map once we have resolved meta->map_ptr in check_func_arg. This means we can't swap out the prototype in check_helper_call until we're half way through the function.

Quoted 6 lines

Instead, modify check_func_arg to treat ARG_PTR_TO_MAP_VALUE to mean "the native type for the map" instead of "pointer to memory" for sockmap and sockhash. This means we don't have to modify the function prototype at all

#6 lore raw

PATCH v3 5/6 bpf: sockmap: allow update from BPF

In reply to Lorenz Bauer

Allow calling bpf_map_update_elem on sockmap and sockhash from a BPF context. The synchronization required for this is a bit fiddly: we need to prevent the socket from changing its state while we add it to the sockmap, since we rely on getting a callback via sk_prot->unhash. However, we can't just lock_sock like in sock_map_sk_acquire because that might sleep. So instead we disable softirq processing and use bh_lock_sock to prevent further modification.

Yet, this is still not enough. BPF can be called in contexts where the current CPU might have locked a socket. If the BPF can get a hold of such a socket, inserting it into a sockmap would lead to a deadlock. One straight forward example are sock_ops programs that have ctx->sk, but the same problem exists for kprobes, etc. We deal with this by allowing sockmap updates only from known safe contexts. Improper usage is rejected by the verifier.

I've audited the enabled contexts to make sure they can't run in a locked context. It's possible that CGROUP_SKB and others are safe as well, but the auditing here is much more difficult. In any case, we can extend the safe contexts when the need arises.

---
 kernel/bpf/verifier.c | 38 ++++++++++++++++++++++++++++++++++++--
 net/core/sock_map.c   | 24 ++++++++++++++++++++++++
 2 files changed, 60 insertions(+), 2 deletions(-)
kernel/bpf/verifier.c +36 -2
index 7e15866c5184..7ba2f7bf81f4 100644
@@ -4178,6 +4178,38 @@ static int check_func_arg(struct bpf_verifier_env *env, u32 arg,
 	return -EACCES;
 }
 
+static bool may_update_sockmap(struct bpf_verifier_env *env, int func_id)
+{
+	enum bpf_attach_type eatype = env->prog->expected_attach_type;
+	enum bpf_prog_type type = env->prog->type;
+
+	if (func_id != BPF_FUNC_map_update_elem)
+		return false;
+
+	/* It's not possible to get access to a locked struct sock in these
+	 * contexts, so updating is safe.
+	 */
+	switch (type) {
+	case BPF_PROG_TYPE_TRACING:
+		if (eatype == BPF_TRACE_ITER)
+			return true;
+		break;
+	case BPF_PROG_TYPE_SOCKET_FILTER:
+	case BPF_PROG_TYPE_SCHED_CLS:
+	case BPF_PROG_TYPE_SCHED_ACT:
+	case BPF_PROG_TYPE_XDP:
+	case BPF_PROG_TYPE_SK_REUSEPORT:
+	case BPF_PROG_TYPE_FLOW_DISSECTOR:
+	case BPF_PROG_TYPE_SK_LOOKUP:
+		return true;
+	default:
+		break;
+	}
+
+	verbose(env, "cannot update sockmap in this context\n");
+	return false;
+}
+
 static int check_map_func_compatibility(struct bpf_verifier_env *env,
 					struct bpf_map *map, int func_id)
 {
@@ -4249,7 +4281,8 @@ static int check_map_func_compatibility(struct bpf_verifier_env *env,
 		    func_id != BPF_FUNC_map_delete_elem &&
 		    func_id != BPF_FUNC_msg_redirect_map &&
 		    func_id != BPF_FUNC_sk_select_reuseport &&
-		    func_id != BPF_FUNC_map_lookup_elem)
+		    func_id != BPF_FUNC_map_lookup_elem &&
+		    !may_update_sockmap(env, func_id))
 			goto error;
 		break;
 	case BPF_MAP_TYPE_SOCKHASH:
@@ -4258,7 +4291,8 @@ static int check_map_func_compatibility(struct bpf_verifier_env *env,
 		    func_id != BPF_FUNC_map_delete_elem &&
 		    func_id != BPF_FUNC_msg_redirect_hash &&
 		    func_id != BPF_FUNC_sk_select_reuseport &&
-		    func_id != BPF_FUNC_map_lookup_elem)
+		    func_id != BPF_FUNC_map_lookup_elem &&
+		    !may_update_sockmap(env, func_id))
 			goto error;
 		break;
 	case BPF_MAP_TYPE_REUSEPORT_SOCKARRAY:
net/core/sock_map.c +24 -0
index 48e83f93ee66..d6c6e1e312fc 100644
@@ -603,6 +603,28 @@ int sock_map_update_elem_sys(struct bpf_map *map, void *key, void *value,
 	return ret;
 }
 
+static int sock_map_update_elem(struct bpf_map *map, void *key,
+				void *value, u64 flags)
+{
+	struct sock *sk = (struct sock *)value;
+	int ret;
+
+	if (!sock_map_sk_is_suitable(sk))
+		return -EOPNOTSUPP;
+
+	local_bh_disable();
+	bh_lock_sock(sk);
+	if (!sock_map_sk_state_allowed(sk))
+		ret = -EOPNOTSUPP;
+	else if (map->map_type == BPF_MAP_TYPE_SOCKMAP)
+		ret = sock_map_update_common(map, *(u32 *)key, sk, flags);
+	else
+		ret = sock_hash_update_common(map, key, sk, flags);
+	bh_unlock_sock(sk);
+	local_bh_enable();
+	return ret;
+}
+
 BPF_CALL_4(bpf_sock_map_update, struct bpf_sock_ops_kern *, sops,
 	   struct bpf_map *, map, void *, key, u64, flags)
 {
@@ -687,6 +709,7 @@ const struct bpf_map_ops sock_map_ops = {
 	.map_free		= sock_map_free,
 	.map_get_next_key	= sock_map_get_next_key,
 	.map_lookup_elem_sys_only = sock_map_lookup_sys,
+	.map_update_elem	= sock_map_update_elem,
 	.map_delete_elem	= sock_map_delete_elem,
 	.map_lookup_elem	= sock_map_lookup,
 	.map_release_uref	= sock_map_release_progs,
@@ -1180,6 +1203,7 @@ const struct bpf_map_ops sock_hash_ops = {
 	.map_alloc		= sock_hash_alloc,
 	.map_free		= sock_hash_free,
 	.map_get_next_key	= sock_hash_get_next_key,
+	.map_update_elem	= sock_map_update_elem,
 	.map_delete_elem	= sock_hash_delete_elem,
 	.map_lookup_elem	= sock_hash_lookup,
 	.map_lookup_elem_sys_only = sock_hash_lookup_sys,
Signature · 1 line
2.25.1
Acked-by on v3 5/6
#7 lore raw

Re: [PATCH bpf-next v3 5/6] bpf: sockmap: allow update from BPF

Quoted 17 lines · Lorenz Bauer

Allow calling bpf_map_update_elem on sockmap and sockhash from a BPF context. The synchronization required for this is a bit fiddly: we need to prevent the socket from changing its state while we add it to the sockmap, since we rely on getting a callback via sk_prot->unhash. However, we can't just lock_sock like in sock_map_sk_acquire because that might sleep. So instead we disable softirq processing and use bh_lock_sock to prevent further modification.

Yet, this is still not enough. BPF can be called in contexts where the current CPU might have locked a socket. If the BPF can get a hold of such a socket, inserting it into a sockmap would lead to a deadlock. One straight forward example are sock_ops programs that have ctx->sk, but the same problem exists for kprobes, etc. We deal with this by allowing sockmap updates only from known safe contexts. Improper usage is rejected by the verifier.

Quoted 6 lines

I've audited the enabled contexts to make sure they can't run in a locked context. It's possible that CGROUP_SKB and others are safe as well, but the auditing here is much more difficult. In any case, we can extend the safe contexts when the need arises.

#8 lore raw

PATCH v3 6/6 selftests: bpf: test sockmap update from BPF

In reply to Lorenz Bauer

Add a test which copies a socket from a sockmap into another sockmap or sockhash. This excercises bpf_map_update_elem support from BPF context. Compare the socket cookies from source and destination to ensure that the copy succeeded.

Also check that the verifier rejects map_update from unsafe contexts.

---
 .../selftests/bpf/prog_tests/sockmap_basic.c  | 78 +++++++++++++++++++
 .../bpf/progs/test_sockmap_invalid_update.c   | 23 ++++++
 .../selftests/bpf/progs/test_sockmap_update.c | 48 ++++++++++++
 3 files changed, 149 insertions(+)
 create mode 100644 tools/testing/selftests/bpf/progs/test_sockmap_invalid_update.c
 create mode 100644 tools/testing/selftests/bpf/progs/test_sockmap_update.c
tools/testing/selftests/bpf/prog_tests/sockmap_basic.c +78 -0
index 96e7b7f84c65..65ce7c289534 100644
@@ -4,6 +4,8 @@
 
 #include "test_progs.h"
 #include "test_skmsg_load_helpers.skel.h"
+#include "test_sockmap_update.skel.h"
+#include "test_sockmap_invalid_update.skel.h"
 
 #define TCP_REPAIR		19	/* TCP sock is under repair right now */
 
@@ -101,6 +103,76 @@ static void test_skmsg_helpers(enum bpf_map_type map_type)
 	test_skmsg_load_helpers__destroy(skel);
 }
 
+static void test_sockmap_update(enum bpf_map_type map_type)
+{
+	struct bpf_prog_test_run_attr tattr;
+	int err, prog, src, dst, duration = 0;
+	struct test_sockmap_update *skel;
+	__u64 src_cookie, dst_cookie;
+	const __u32 zero = 0;
+	char dummy[14] = {0};
+	__s64 sk;
+
+	sk = connected_socket_v4();
+	if (CHECK(sk == -1, "connected_socket_v4", "cannot connect\n"))
+		return;
+
+	skel = test_sockmap_update__open_and_load();
+	if (CHECK(!skel, "open_and_load", "cannot load skeleton\n")) {
+		close(sk);
+		return;
+	}
+
+	prog = bpf_program__fd(skel->progs.copy_sock_map);
+	src = bpf_map__fd(skel->maps.src);
+	if (map_type == BPF_MAP_TYPE_SOCKMAP)
+		dst = bpf_map__fd(skel->maps.dst_sock_map);
+	else
+		dst = bpf_map__fd(skel->maps.dst_sock_hash);
+
+	err = bpf_map_update_elem(src, &zero, &sk, BPF_NOEXIST);
+	if (CHECK(err, "update_elem(src)", "errno=%u\n", errno))
+		goto out;
+
+	err = bpf_map_lookup_elem(src, &zero, &src_cookie);
+	if (CHECK(err, "lookup_elem(src, cookie)", "errno=%u\n", errno))
+		goto out;
+
+	tattr = (struct bpf_prog_test_run_attr){
+		.prog_fd = prog,
+		.repeat = 1,
+		.data_in = dummy,
+		.data_size_in = sizeof(dummy),
+	};
+
+	err = bpf_prog_test_run_xattr(&tattr);
+	if (CHECK_ATTR(err || !tattr.retval, "bpf_prog_test_run",
+		       "errno=%u retval=%u\n", errno, tattr.retval))
+		goto out;
+
+	err = bpf_map_lookup_elem(dst, &zero, &dst_cookie);
+	if (CHECK(err, "lookup_elem(dst, cookie)", "errno=%u\n", errno))
+		goto out;
+
+	CHECK(dst_cookie != src_cookie, "cookie mismatch", "%llu != %llu\n",
+	      dst_cookie, src_cookie);
+
+out:
+	close(sk);
+	test_sockmap_update__destroy(skel);
+}
+
+static void test_sockmap_invalid_update(void)
+{
+	struct test_sockmap_invalid_update *skel;
+	int duration = 0;
+
+	skel = test_sockmap_invalid_update__open_and_load();
+	CHECK(skel, "open_and_load", "verifier accepted map_update\n");
+	if (skel)
+		test_sockmap_invalid_update__destroy(skel);
+}
+
 void test_sockmap_basic(void)
 {
 	if (test__start_subtest("sockmap create_update_free"))
@@ -111,4 +183,10 @@ void test_sockmap_basic(void)
 		test_skmsg_helpers(BPF_MAP_TYPE_SOCKMAP);
 	if (test__start_subtest("sockhash sk_msg load helpers"))
 		test_skmsg_helpers(BPF_MAP_TYPE_SOCKHASH);
+	if (test__start_subtest("sockmap update"))
+		test_sockmap_update(BPF_MAP_TYPE_SOCKMAP);
+	if (test__start_subtest("sockhash update"))
+		test_sockmap_update(BPF_MAP_TYPE_SOCKHASH);
+	if (test__start_subtest("sockmap update in unsafe context"))
+		test_sockmap_invalid_update();
 }
tools/testing/selftests/bpf/progs/test_sockmap_invalid_update.c new file +23 -0
new file mode 100644
index 000000000000..02a59e220cbc
@@ -0,0 +1,23 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (c) 2020 Cloudflare
+#include "vmlinux.h"
+#include <bpf/bpf_helpers.h>
+
+struct {
+	__uint(type, BPF_MAP_TYPE_SOCKMAP);
+	__uint(max_entries, 1);
+	__type(key, __u32);
+	__type(value, __u64);
+} map SEC(".maps");
+
+SEC("sockops")
+int bpf_sockmap(struct bpf_sock_ops *skops)
+{
+	__u32 key = 0;
+
+	if (skops->sk)
+		bpf_map_update_elem(&map, &key, skops->sk, 0);
+	return 0;
+}
+
+char _license[] SEC("license") = "GPL";
tools/testing/selftests/bpf/progs/test_sockmap_update.c new file +48 -0

48 lines not on this page · raw message

Signature · 1 line
2.25.1
Acked-by on v3 6/6
#9 lore raw

Re: [PATCH bpf-next v3 6/6] selftests: bpf: test sockmap update from BPF

Quoted 8 lines · Lorenz Bauer

Add a test which copies a socket from a sockmap into another sockmap or sockhash. This excercises bpf_map_update_elem support from BPF context. Compare the socket cookies from source and destination to ensure that the copy succeeded.

Also check that the verifier rejects map_update from unsafe contexts.

A few nits below.

Quoted 72 lines
---
  .../selftests/bpf/prog_tests/sockmap_basic.c  | 78 +++++++++++++++++++
  .../bpf/progs/test_sockmap_invalid_update.c   | 23 ++++++
  .../selftests/bpf/progs/test_sockmap_update.c | 48 ++++++++++++
  3 files changed, 149 insertions(+)

create mode 100644 tools/testing/selftests/bpf/progs/test_sockmap_invalid_update.c create mode 100644 tools/testing/selftests/bpf/progs/test_sockmap_update.c

tools/testing/selftests/bpf/prog_tests/sockmap_basic.c +49 -0
index 96e7b7f84c65..65ce7c289534 100644
@@ -4,6 +4,8 @@
 
  #include "test_progs.h"
  #include "test_skmsg_load_helpers.skel.h"
+#include "test_sockmap_update.skel.h"
+#include "test_sockmap_invalid_update.skel.h"
 
  #define TCP_REPAIR		19	/* TCP sock is under repair right now */
 
@@ -101,6 +103,76 @@ static void test_skmsg_helpers(enum bpf_map_type map_type)
  	test_skmsg_load_helpers__destroy(skel);
  }
 
+static void test_sockmap_update(enum bpf_map_type map_type)
+{
+	struct bpf_prog_test_run_attr tattr;
+	int err, prog, src, dst, duration = 0;
+	struct test_sockmap_update *skel;
+	__u64 src_cookie, dst_cookie;
+	const __u32 zero = 0;
+	char dummy[14] = {0};
+	__s64 sk;
+
+	sk = connected_socket_v4();
+	if (CHECK(sk == -1, "connected_socket_v4", "cannot connect\n"))
+		return;
+
+	skel = test_sockmap_update__open_and_load();
+	if (CHECK(!skel, "open_and_load", "cannot load skeleton\n")) {
+		close(sk);
+		return;
+	}
+
+	prog = bpf_program__fd(skel->progs.copy_sock_map);
+	src = bpf_map__fd(skel->maps.src);
+	if (map_type == BPF_MAP_TYPE_SOCKMAP)
+		dst = bpf_map__fd(skel->maps.dst_sock_map);
+	else
+		dst = bpf_map__fd(skel->maps.dst_sock_hash);
+
+	err = bpf_map_update_elem(src, &zero, &sk, BPF_NOEXIST);
+	if (CHECK(err, "update_elem(src)", "errno=%u\n", errno))
+		goto out;
+
+	err = bpf_map_lookup_elem(src, &zero, &src_cookie);
+	if (CHECK(err, "lookup_elem(src, cookie)", "errno=%u\n", errno))
+		goto out;
+
+	tattr = (struct bpf_prog_test_run_attr){
+		.prog_fd = prog,
+		.repeat = 1,
+		.data_in = dummy,
+		.data_size_in = sizeof(dummy),
+	};
+
+	err = bpf_prog_test_run_xattr(&tattr);
+	if (CHECK_ATTR(err || !tattr.retval, "bpf_prog_test_run",
+		       "errno=%u retval=%u\n", errno, tattr.retval))
+		goto out;
+

tools/testing/selftests/bpf/prog_tests/sockmap_basic.c static void test_skmsg_helpers(enum bpf_map_type map_type) · line 162

Quoted 10 lines
tools/testing/selftests/bpf/prog_tests/sockmap_basic.c +10 -0
+	err = bpf_map_lookup_elem(dst, &zero, &dst_cookie);
+	if (CHECK(err, "lookup_elem(dst, cookie)", "errno=%u\n", errno))
+		goto out;
+
+	CHECK(dst_cookie != src_cookie, "cookie mismatch", "%llu != %llu\n",
+	      dst_cookie, src_cookie);
+
+out:
+	close(sk);
+	test_sockmap_update__destroy(skel);

nit. In the beginning of the function, 'sk' is available and then skel is opened and loaded. You can have

out:

	test_sockmap_update__destroy(skel);
close_sk:
	close(sk);

this probably more conforms to linux coding style.

Then you can have

	if (CHECK(!skel, "open_and_load", "cannot load skeleton\n"))
		goto close_sk;

tools/testing/selftests/bpf/prog_tests/sockmap_basic.c static void test_skmsg_helpers(enum bpf_map_type map_type) · line 173

Quoted 11 lines
tools/testing/selftests/bpf/prog_tests/sockmap_basic.c +11 -0
+}
+
+static void test_sockmap_invalid_update(void)
+{
+	struct test_sockmap_invalid_update *skel;
+	int duration = 0;
+
+	skel = test_sockmap_invalid_update__open_and_load();
+	CHECK(skel, "open_and_load", "verifier accepted map_update\n");
+	if (skel)
+		test_sockmap_invalid_update__destroy(skel);

nit, you can just have

	if (CHECK(skel, "open_and_load", "verifier accepted map_update\n"))
		test_sockmap_invalid_update__destroy(skel);
Quoted 100 lines
tools/testing/selftests/bpf/prog_tests/sockmap_basic.c +8 -0
+}
+
  void test_sockmap_basic(void)
  {
  	if (test__start_subtest("sockmap create_update_free"))
@@ -111,4 +183,10 @@ void test_sockmap_basic(void)
  		test_skmsg_helpers(BPF_MAP_TYPE_SOCKMAP);
  	if (test__start_subtest("sockhash sk_msg load helpers"))
  		test_skmsg_helpers(BPF_MAP_TYPE_SOCKHASH);
+	if (test__start_subtest("sockmap update"))
+		test_sockmap_update(BPF_MAP_TYPE_SOCKMAP);
+	if (test__start_subtest("sockhash update"))
+		test_sockmap_update(BPF_MAP_TYPE_SOCKHASH);
+	if (test__start_subtest("sockmap update in unsafe context"))
+		test_sockmap_invalid_update();
  }
tools/testing/selftests/bpf/progs/test_sockmap_invalid_update.c new file +23 -0
new file mode 100644
index 000000000000..02a59e220cbc
@@ -0,0 +1,23 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (c) 2020 Cloudflare
+#include "vmlinux.h"
+#include <bpf/bpf_helpers.h>
+
+struct {
+	__uint(type, BPF_MAP_TYPE_SOCKMAP);
+	__uint(max_entries, 1);
+	__type(key, __u32);
+	__type(value, __u64);
+} map SEC(".maps");
+
+SEC("sockops")
+int bpf_sockmap(struct bpf_sock_ops *skops)
+{
+	__u32 key = 0;
+
+	if (skops->sk)
+		bpf_map_update_elem(&map, &key, skops->sk, 0);
+	return 0;
+}
+
+char _license[] SEC("license") = "GPL";
tools/testing/selftests/bpf/progs/test_sockmap_update.c new file +48 -0
new file mode 100644
index 000000000000..9d0c9f28cab2
@@ -0,0 +1,48 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (c) 2020 Cloudflare
+#include "vmlinux.h"
+#include <bpf/bpf_helpers.h>
+
+struct {
+	__uint(type, BPF_MAP_TYPE_SOCKMAP);
+	__uint(max_entries, 1);
+	__type(key, __u32);
+	__type(value, __u64);
+} src SEC(".maps");
+
+struct {
+	__uint(type, BPF_MAP_TYPE_SOCKMAP);
+	__uint(max_entries, 1);
+	__type(key, __u32);
+	__type(value, __u64);
+} dst_sock_map SEC(".maps");
+
+struct {
+	__uint(type, BPF_MAP_TYPE_SOCKHASH);
+	__uint(max_entries, 1);
+	__type(key, __u32);
+	__type(value, __u64);
+} dst_sock_hash SEC(".maps");
+
+SEC("classifier/copy_sock_map")
+int copy_sock_map(void *ctx)
+{
+	struct bpf_sock *sk;
+	bool failed = false;
+	__u32 key = 0;
+
+	sk = bpf_map_lookup_elem(&src, &key);
+	if (!sk)
+		return SK_DROP;
+
+	if (bpf_map_update_elem(&dst_sock_map, &key, sk, 0))
+		failed = true;
+
+	if (bpf_map_update_elem(&dst_sock_hash, &key, sk, 0))
+		failed = true;
+
+	bpf_sk_release(sk);
+	return failed ? SK_DROP : SK_PASS;
+}
+
+char _license[] SEC("license") = "GPL";
#10 lore raw
In reply to Lorenz Bauer
Quoted 24 lines · Lorenz Bauer

We're currently building a control plane for our BPF socket dispatch work. As part of that, we have a need to create a copy of an existing sockhash, to allow us to change the keys. I previously proposed allowing privileged userspace to look up sockets, which doesn't work due to security concerns (see [1]).

In follow up discussions during BPF office hours we identified bpf_iter as a possible solution: instead of accessing sockets from user space we can iterate the source sockhash, and insert the values into a new map. Enabling this requires two pieces: the ability to iterate sockmap and sockhash, as well as being able to call map_update_elem from BPF.

This patch set implements the latter: it's now possible to update sockmap from BPF context. As a next step, we can implement bpf_iter for sockmap.

===

I've done some more fixups, and audited the safe contexts more thoroughly. As a result I'm removing CGROUP_SKB, SK_MSG and SK_SKB for now.

Quoted 5 lines

Changes in v3:
- Use CHECK as much as possible (Yonghong)
- Reject ARG_PTR_TO_MAP_VALUE_OR_NULL for sockmap (Yonghong)
- Remove CGROUP_SKB, SK_MSG, SK_SKB from safe contexts
- Test that the verifier rejects update from unsafe context

All looks good to me.
I've applied the set.
Please follow up with a cleanup to selftests as Yonghong suggested.