bpf: Post-hooks for sys_bind

"Post-hooks" are hooks that are called right before returning from
sys_bind. At this time IP and port are already allocated and no further
changes to `struct sock` can happen before returning from sys_bind but
BPF program has a chance to inspect the socket and change sys_bind
result.

Specifically it can e.g. inspect what port was allocated and if it
doesn't satisfy some policy, BPF program can force sys_bind to fail and
return EPERM to user.

Another example of usage is recording the IP:port pair to some map to
use it in later calls to sys_connect. E.g. if some TCP server inside
cgroup was bound to some IP:port_n, it can be recorded to a map. And
later when some TCP client inside same cgroup is trying to connect to
127.0.0.1:port_n, BPF hook for sys_connect can override the destination
and connect application to IP:port_n instead of 127.0.0.1:port_n. That
helps forcing all applications inside a cgroup to use desired IP and not
break those applications if they e.g. use localhost to communicate
between each other.

== Implementation details ==

Post-hooks are implemented as two new attach types
`BPF_CGROUP_INET4_POST_BIND` and `BPF_CGROUP_INET6_POST_BIND` for
existing prog type `BPF_PROG_TYPE_CGROUP_SOCK`.

Separate attach types for IPv4 and IPv6 are introduced to avoid access
to IPv6 field in `struct sock` from `inet_bind()` and to IPv4 field from
`inet6_bind()` since those fields might not make sense in such cases.

Signed-off-by: Andrey Ignatov <[email protected]>
Signed-off-by: Alexei Starovoitov <[email protected]>
Signed-off-by: Daniel Borkmann <[email protected]>

Author: rdna

include/linux/bpf-cgroup.h

@@ -98,16 +98,24 @@ int __cgroup_bpf_check_dev_permission(short dev_type, u32 major, u32 minor,
 	__ret;								       \
 })
 
-#define BPF_CGROUP_RUN_PROG_INET_SOCK(sk)				       \
+#define BPF_CGROUP_RUN_SK_PROG(sk, type)				       \
 ({									       \
 	int __ret = 0;							       \
 	if (cgroup_bpf_enabled) {					       \
-		__ret = __cgroup_bpf_run_filter_sk(sk,			       \
-						 BPF_CGROUP_INET_SOCK_CREATE); \
+		__ret = __cgroup_bpf_run_filter_sk(sk, type);		       \
 	}								       \
 	__ret;								       \
 })
 
+#define BPF_CGROUP_RUN_PROG_INET_SOCK(sk)				       \
+	BPF_CGROUP_RUN_SK_PROG(sk, BPF_CGROUP_INET_SOCK_CREATE)
+
+#define BPF_CGROUP_RUN_PROG_INET4_POST_BIND(sk)				       \
+	BPF_CGROUP_RUN_SK_PROG(sk, BPF_CGROUP_INET4_POST_BIND)
+
+#define BPF_CGROUP_RUN_PROG_INET6_POST_BIND(sk)				       \
+	BPF_CGROUP_RUN_SK_PROG(sk, BPF_CGROUP_INET6_POST_BIND)
+
 #define BPF_CGROUP_RUN_SA_PROG(sk, uaddr, type)				       \
 ({									       \
 	int __ret = 0;							       \
@@ -183,6 +191,8 @@ static inline int cgroup_bpf_inherit(struct cgroup *cgrp) { return 0; }
 #define BPF_CGROUP_RUN_PROG_INET_SOCK(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET4_BIND(sk, uaddr) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET6_BIND(sk, uaddr) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET4_POST_BIND(sk) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET6_POST_BIND(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr) ({ 0; })

include/uapi/linux/bpf.h

@@ -152,6 +152,8 @@ enum bpf_attach_type {
 	BPF_CGROUP_INET6_BIND,
 	BPF_CGROUP_INET4_CONNECT,
 	BPF_CGROUP_INET6_CONNECT,
+	BPF_CGROUP_INET4_POST_BIND,
+	BPF_CGROUP_INET6_POST_BIND,
 	__MAX_BPF_ATTACH_TYPE
 };
 
@@ -948,6 +950,15 @@ struct bpf_sock {
 	__u32 protocol;
 	__u32 mark;
 	__u32 priority;
+	__u32 src_ip4;		/* Allows 1,2,4-byte read.
+				 * Stored in network byte order.
+				 */
+	__u32 src_ip6[4];	/* Allows 1,2,4-byte read.
+				 * Stored in network byte order.
+				 */
+	__u32 src_port;		/* Allows 4-byte read.
+				 * Stored in host byte order
+				 */
 };
 
 #define XDP_PACKET_HEADROOM 256

kernel/bpf/syscall.c

@@ -1171,11 +1171,46 @@ struct bpf_prog *bpf_prog_get_type_dev(u32 ufd, enum bpf_prog_type type,
 }
 EXPORT_SYMBOL_GPL(bpf_prog_get_type_dev);
 
+/* Initially all BPF programs could be loaded w/o specifying
+ * expected_attach_type. Later for some of them specifying expected_attach_type
+ * at load time became required so that program could be validated properly.
+ * Programs of types that are allowed to be loaded both w/ and w/o (for
+ * backward compatibility) expected_attach_type, should have the default attach
+ * type assigned to expected_attach_type for the latter case, so that it can be
+ * validated later at attach time.
+ *
+ * bpf_prog_load_fixup_attach_type() sets expected_attach_type in @attr if
+ * prog type requires it but has some attach types that have to be backward
+ * compatible.
+ */
+static void bpf_prog_load_fixup_attach_type(union bpf_attr *attr)
+{
+	switch (attr->prog_type) {
+	case BPF_PROG_TYPE_CGROUP_SOCK:
+		/* Unfortunately BPF_ATTACH_TYPE_UNSPEC enumeration doesn't
+		 * exist so checking for non-zero is the way to go here.
+		 */
+		if (!attr->expected_attach_type)
+			attr->expected_attach_type =
+				BPF_CGROUP_INET_SOCK_CREATE;
+		break;
+	}
+}
+
 static int
 bpf_prog_load_check_attach_type(enum bpf_prog_type prog_type,
 				enum bpf_attach_type expected_attach_type)
 {
 	switch (prog_type) {
+	case BPF_PROG_TYPE_CGROUP_SOCK:
+		switch (expected_attach_type) {
+		case BPF_CGROUP_INET_SOCK_CREATE:
+		case BPF_CGROUP_INET4_POST_BIND:
+		case BPF_CGROUP_INET6_POST_BIND:
+			return 0;
+		default:
+			return -EINVAL;
+		}
 	case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
 		switch (expected_attach_type) {
 		case BPF_CGROUP_INET4_BIND:
@@ -1195,6 +1230,7 @@ static int bpf_prog_attach_check_attach_type(const struct bpf_prog *prog,
 					     enum bpf_attach_type attach_type)
 {
 	switch (prog->type) {
+	case BPF_PROG_TYPE_CGROUP_SOCK:
 	case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
 		return attach_type == prog->expected_attach_type ? 0 : -EINVAL;
 	default:
@@ -1240,6 +1276,7 @@ static int bpf_prog_load(union bpf_attr *attr)
 	    !capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
+	bpf_prog_load_fixup_attach_type(attr);
 	if (bpf_prog_load_check_attach_type(type, attr->expected_attach_type))
 		return -EINVAL;
 
@@ -1489,6 +1526,8 @@ static int bpf_prog_attach(const union bpf_attr *attr)
 		ptype = BPF_PROG_TYPE_CGROUP_SKB;
 		break;
 	case BPF_CGROUP_INET_SOCK_CREATE:
+	case BPF_CGROUP_INET4_POST_BIND:
+	case BPF_CGROUP_INET6_POST_BIND:
 		ptype = BPF_PROG_TYPE_CGROUP_SOCK;
 		break;
 	case BPF_CGROUP_INET4_BIND:
@@ -1557,6 +1596,8 @@ static int bpf_prog_detach(const union bpf_attr *attr)
 		ptype = BPF_PROG_TYPE_CGROUP_SKB;
 		break;
 	case BPF_CGROUP_INET_SOCK_CREATE:
+	case BPF_CGROUP_INET4_POST_BIND:
+	case BPF_CGROUP_INET6_POST_BIND:
 		ptype = BPF_PROG_TYPE_CGROUP_SOCK;
 		break;
 	case BPF_CGROUP_INET4_BIND:
@@ -1616,6 +1657,8 @@ static int bpf_prog_query(const union bpf_attr *attr,
 	case BPF_CGROUP_INET_SOCK_CREATE:
 	case BPF_CGROUP_INET4_BIND:
 	case BPF_CGROUP_INET6_BIND:
+	case BPF_CGROUP_INET4_POST_BIND:
+	case BPF_CGROUP_INET6_POST_BIND:
 	case BPF_CGROUP_INET4_CONNECT:
 	case BPF_CGROUP_INET6_CONNECT:
 	case BPF_CGROUP_SOCK_OPS:

net/core/filter.c

@@ -4097,30 +4097,80 @@ static bool lwt_is_valid_access(int off, int size,
 	return bpf_skb_is_valid_access(off, size, type, prog, info);
 }
 
-static bool sock_filter_is_valid_access(int off, int size,
-					enum bpf_access_type type,
-					const struct bpf_prog *prog,
-					struct bpf_insn_access_aux *info)
+
+/* Attach type specific accesses */
+static bool __sock_filter_check_attach_type(int off,
+					    enum bpf_access_type access_type,
+					    enum bpf_attach_type attach_type)
 {
-	if (type == BPF_WRITE) {
-		switch (off) {
-		case offsetof(struct bpf_sock, bound_dev_if):
-		case offsetof(struct bpf_sock, mark):
-		case offsetof(struct bpf_sock, priority):
-			break;
+	switch (off) {
+	case offsetof(struct bpf_sock, bound_dev_if):
+	case offsetof(struct bpf_sock, mark):
+	case offsetof(struct bpf_sock, priority):
+		switch (attach_type) {
+		case BPF_CGROUP_INET_SOCK_CREATE:
+			goto full_access;
+		default:
+			return false;
+		}
+	case bpf_ctx_range(struct bpf_sock, src_ip4):
+		switch (attach_type) {
+		case BPF_CGROUP_INET4_POST_BIND:
+			goto read_only;
+		default:
+			return false;
+		}
+	case bpf_ctx_range_till(struct bpf_sock, src_ip6[0], src_ip6[3]):
+		switch (attach_type) {
+		case BPF_CGROUP_INET6_POST_BIND:
+			goto read_only;
+		default:
+			return false;
+		}
+	case bpf_ctx_range(struct bpf_sock, src_port):
+		switch (attach_type) {
+		case BPF_CGROUP_INET4_POST_BIND:
+		case BPF_CGROUP_INET6_POST_BIND:
+			goto read_only;
 		default:
 			return false;
 		}
 	}
+read_only:
+	return access_type == BPF_READ;
+full_access:
+	return true;
+}
+
+static bool __sock_filter_check_size(int off, int size,
+				     struct bpf_insn_access_aux *info)
+{
+	const int size_default = sizeof(__u32);
 
-	if (off < 0 || off + size > sizeof(struct bpf_sock))
+	switch (off) {
+	case bpf_ctx_range(struct bpf_sock, src_ip4):
+	case bpf_ctx_range_till(struct bpf_sock, src_ip6[0], src_ip6[3]):
+		bpf_ctx_record_field_size(info, size_default);
+		return bpf_ctx_narrow_access_ok(off, size, size_default);
+	}
+
+	return size == size_default;
+}
+
+static bool sock_filter_is_valid_access(int off, int size,
+					enum bpf_access_type type,
+					const struct bpf_prog *prog,
+					struct bpf_insn_access_aux *info)
+{
+	if (off < 0 || off >= sizeof(struct bpf_sock))
 		return false;
-	/* The verifier guarantees that size > 0. */
 	if (off % size != 0)
 		return false;
-	if (size != sizeof(__u32))
+	if (!__sock_filter_check_attach_type(off, type,
+					     prog->expected_attach_type))
+		return false;
+	if (!__sock_filter_check_size(off, size, info))
 		return false;
-
 	return true;
 }
 
@@ -4728,6 +4778,7 @@ static u32 sock_filter_convert_ctx_access(enum bpf_access_type type,
 					  struct bpf_prog *prog, u32 *target_size)
 {
 	struct bpf_insn *insn = insn_buf;
+	int off;
 
 	switch (si->off) {
 	case offsetof(struct bpf_sock, bound_dev_if):
@@ -4783,6 +4834,43 @@ static u32 sock_filter_convert_ctx_access(enum bpf_access_type type,
 		*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, SK_FL_PROTO_MASK);
 		*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg, SK_FL_PROTO_SHIFT);
 		break;
+
+	case offsetof(struct bpf_sock, src_ip4):
+		*insn++ = BPF_LDX_MEM(
+			BPF_SIZE(si->code), si->dst_reg, si->src_reg,
+			bpf_target_off(struct sock_common, skc_rcv_saddr,
+				       FIELD_SIZEOF(struct sock_common,
+						    skc_rcv_saddr),
+				       target_size));
+		break;
+
+	case bpf_ctx_range_till(struct bpf_sock, src_ip6[0], src_ip6[3]):
+#if IS_ENABLED(CONFIG_IPV6)
+		off = si->off;
+		off -= offsetof(struct bpf_sock, src_ip6[0]);
+		*insn++ = BPF_LDX_MEM(
+			BPF_SIZE(si->code), si->dst_reg, si->src_reg,
+			bpf_target_off(
+				struct sock_common,
+				skc_v6_rcv_saddr.s6_addr32[0],
+				FIELD_SIZEOF(struct sock_common,
+					     skc_v6_rcv_saddr.s6_addr32[0]),
+				target_size) + off);
+#else
+		(void)off;
+		*insn++ = BPF_MOV32_IMM(si->dst_reg, 0);
+#endif
+		break;
+
+	case offsetof(struct bpf_sock, src_port):
+		*insn++ = BPF_LDX_MEM(
+			BPF_FIELD_SIZEOF(struct sock_common, skc_num),
+			si->dst_reg, si->src_reg,
+			bpf_target_off(struct sock_common, skc_num,
+				       FIELD_SIZEOF(struct sock_common,
+						    skc_num),
+				       target_size));
+		break;
 	}
 
 	return insn - insn_buf;

net/ipv4/af_inet.c

@@ -519,12 +519,18 @@ int __inet_bind(struct sock *sk, struct sockaddr *uaddr, int addr_len,
 		inet->inet_saddr = 0;  /* Use device */
 
 	/* Make sure we are allowed to bind here. */
-	if ((snum || !(inet->bind_address_no_port ||
-		       force_bind_address_no_port)) &&
-	    sk->sk_prot->get_port(sk, snum)) {
-		inet->inet_saddr = inet->inet_rcv_saddr = 0;
-		err = -EADDRINUSE;
-		goto out_release_sock;
+	if (snum || !(inet->bind_address_no_port ||
+		      force_bind_address_no_port)) {
+		if (sk->sk_prot->get_port(sk, snum)) {
+			inet->inet_saddr = inet->inet_rcv_saddr = 0;
+			err = -EADDRINUSE;
+			goto out_release_sock;
+		}
+		err = BPF_CGROUP_RUN_PROG_INET4_POST_BIND(sk);
+		if (err) {
+			inet->inet_saddr = inet->inet_rcv_saddr = 0;
+			goto out_release_sock;
+		}
 	}
 
 	if (inet->inet_rcv_saddr)

net/ipv6/af_inet6.c

@@ -412,13 +412,20 @@ int __inet6_bind(struct sock *sk, struct sockaddr *uaddr, int addr_len,
 		sk->sk_ipv6only = 1;
 
 	/* Make sure we are allowed to bind here. */
-	if ((snum || !(inet->bind_address_no_port ||
-		       force_bind_address_no_port)) &&
-	    sk->sk_prot->get_port(sk, snum)) {
-		sk->sk_ipv6only = saved_ipv6only;
-		inet_reset_saddr(sk);
-		err = -EADDRINUSE;
-		goto out;
+	if (snum || !(inet->bind_address_no_port ||
+		      force_bind_address_no_port)) {
+		if (sk->sk_prot->get_port(sk, snum)) {
+			sk->sk_ipv6only = saved_ipv6only;
+			inet_reset_saddr(sk);
+			err = -EADDRINUSE;
+			goto out;
+		}
+		err = BPF_CGROUP_RUN_PROG_INET6_POST_BIND(sk);
+		if (err) {
+			sk->sk_ipv6only = saved_ipv6only;
+			inet_reset_saddr(sk);
+			goto out;
+		}
 	}
 
 	if (addr_type != IPV6_ADDR_ANY)