Merge branch 'enable-bpf_skc-cast-for-networking-progs'

Martin KaFai Lau says:

====================
This set allows networking prog type to directly read fields from
the in-kernel socket type, e.g. "struct tcp_sock".

Patch 2 has the details on the use case.

v3:
- Pass arg_btf_id instead of fn into check_reg_type() in Patch 1 (Lorenz)
- Move arg_btf_id from func_proto to struct bpf_reg_types in Patch 2 (Lorenz)
- Remove test_sock_fields from .gitignore in Patch 8 (Andrii)
- Add tests to have better coverage on the modified helpers (Alexei)
  Patch 13 is added.
- Use "void *sk" as the helper argument in UAPI bpf.h

v3:
- ARG_PTR_TO_SOCK_COMMON_OR_NULL was attempted in v2.  The _OR_NULL was
  needed because the PTR_TO_BTF_ID could be NULL but note that a could be NULL
  PTR_TO_BTF_ID is not a scalar NULL to the verifier.  "_OR_NULL" implicitly
  gives an expectation that the helper can take a scalar NULL which does
  not make sense in most (except one) helpers.  Passing scalar NULL
  should be rejected at the verification time.

  Thus, this patch uses ARG_PTR_TO_BTF_ID_SOCK_COMMON to specify that the
  helper can take both the btf-id ptr or the legacy PTR_TO_SOCK_COMMON but
  not scalar NULL.  It requires the func_proto to explicitly specify the
  arg_btf_id such that there is a very clear expectation that the helper
  can handle a NULL PTR_TO_BTF_ID.

v2:
- Add ARG_PTR_TO_SOCK_COMMON_OR_NULL (Lorenz)
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

include/linux/bpf.h

@@ -292,6 +292,7 @@ enum bpf_arg_type {
 	ARG_PTR_TO_ALLOC_MEM,	/* pointer to dynamically allocated memory */
 	ARG_PTR_TO_ALLOC_MEM_OR_NULL,	/* pointer to dynamically allocated memory or NULL */
 	ARG_CONST_ALLOC_SIZE_OR_ZERO,	/* number of allocated bytes requested */
+	ARG_PTR_TO_BTF_ID_SOCK_COMMON,	/* pointer to in-kernel sock_common or bpf-mirrored bpf_sock */
 	__BPF_ARG_TYPE_MAX,
 };
 

include/net/bpf_sk_storage.h

@@ -20,8 +20,6 @@ void bpf_sk_storage_free(struct sock *sk);
 
 extern const struct bpf_func_proto bpf_sk_storage_get_proto;
 extern const struct bpf_func_proto bpf_sk_storage_delete_proto;
-extern const struct bpf_func_proto sk_storage_get_btf_proto;
-extern const struct bpf_func_proto sk_storage_delete_btf_proto;
 
 struct bpf_local_storage_elem;
 struct bpf_sk_storage_diag;

include/uapi/linux/bpf.h

@@ -2512,7 +2512,7 @@ union bpf_attr {
  *		result is from *reuse*\ **->socks**\ [] using the hash of the
  *		tuple.
  *
- * long bpf_sk_release(struct bpf_sock *sock)
+ * long bpf_sk_release(void *sock)
  *	Description
  *		Release the reference held by *sock*. *sock* must be a
  *		non-**NULL** pointer that was returned from
@@ -2692,7 +2692,7 @@ union bpf_attr {
  *		result is from *reuse*\ **->socks**\ [] using the hash of the
  *		tuple.
  *
- * long bpf_tcp_check_syncookie(struct bpf_sock *sk, void *iph, u32 iph_len, struct tcphdr *th, u32 th_len)
+ * long bpf_tcp_check_syncookie(void *sk, void *iph, u32 iph_len, struct tcphdr *th, u32 th_len)
  * 	Description
  * 		Check whether *iph* and *th* contain a valid SYN cookie ACK for
  * 		the listening socket in *sk*.
@@ -2861,6 +2861,7 @@ union bpf_attr {
  *		0 on success.
  *
  *		**-ENOENT** if the bpf-local-storage cannot be found.
+ *		**-EINVAL** if sk is not a fullsock (e.g. a request_sock).
  *
  * long bpf_send_signal(u32 sig)
  *	Description
@@ -2877,7 +2878,7 @@ union bpf_attr {
  *
  *		**-EAGAIN** if bpf program can try again.
  *
- * s64 bpf_tcp_gen_syncookie(struct bpf_sock *sk, void *iph, u32 iph_len, struct tcphdr *th, u32 th_len)
+ * s64 bpf_tcp_gen_syncookie(void *sk, void *iph, u32 iph_len, struct tcphdr *th, u32 th_len)
  *	Description
  *		Try to issue a SYN cookie for the packet with corresponding
  *		IP/TCP headers, *iph* and *th*, on the listening socket in *sk*.
@@ -3106,7 +3107,7 @@ union bpf_attr {
  * 	Return
  * 		The id is returned or 0 in case the id could not be retrieved.
  *
- * long bpf_sk_assign(struct sk_buff *skb, struct bpf_sock *sk, u64 flags)
+ * long bpf_sk_assign(struct sk_buff *skb, void *sk, u64 flags)
  *	Description
  *		Helper is overloaded depending on BPF program type. This
  *		description applies to **BPF_PROG_TYPE_SCHED_CLS** and
@@ -3234,11 +3235,11 @@ union bpf_attr {
  *
  *		**-EOVERFLOW** if an overflow happened: The same object will be tried again.
  *
- * u64 bpf_sk_cgroup_id(struct bpf_sock *sk)
+ * u64 bpf_sk_cgroup_id(void *sk)
  *	Description
  *		Return the cgroup v2 id of the socket *sk*.
  *
- *		*sk* must be a non-**NULL** pointer to a full socket, e.g. one
+ *		*sk* must be a non-**NULL** pointer to a socket, e.g. one
  *		returned from **bpf_sk_lookup_xxx**\ (),
  *		**bpf_sk_fullsock**\ (), etc. The format of returned id is
  *		same as in **bpf_skb_cgroup_id**\ ().
@@ -3248,7 +3249,7 @@ union bpf_attr {
  *	Return
  *		The id is returned or 0 in case the id could not be retrieved.
  *
- * u64 bpf_sk_ancestor_cgroup_id(struct bpf_sock *sk, int ancestor_level)
+ * u64 bpf_sk_ancestor_cgroup_id(void *sk, int ancestor_level)
  *	Description
  *		Return id of cgroup v2 that is ancestor of cgroup associated
  *		with the *sk* at the *ancestor_level*.  The root cgroup is at

kernel/bpf/bpf_lsm.c

@@ -56,9 +56,9 @@ bpf_lsm_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 	case BPF_FUNC_inode_storage_delete:
 		return &bpf_inode_storage_delete_proto;
 	case BPF_FUNC_sk_storage_get:
-		return &sk_storage_get_btf_proto;
+		return &bpf_sk_storage_get_proto;
 	case BPF_FUNC_sk_storage_delete:
-		return &sk_storage_delete_btf_proto;
+		return &bpf_sk_storage_delete_proto;
 	default:
 		return tracing_prog_func_proto(func_id, prog);
 	}

kernel/bpf/verifier.c

@@ -486,7 +486,12 @@ static bool is_acquire_function(enum bpf_func_id func_id,
 static bool is_ptr_cast_function(enum bpf_func_id func_id)
 {
 	return func_id == BPF_FUNC_tcp_sock ||
-		func_id == BPF_FUNC_sk_fullsock;
+		func_id == BPF_FUNC_sk_fullsock ||
+		func_id == BPF_FUNC_skc_to_tcp_sock ||
+		func_id == BPF_FUNC_skc_to_tcp6_sock ||
+		func_id == BPF_FUNC_skc_to_udp6_sock ||
+		func_id == BPF_FUNC_skc_to_tcp_timewait_sock ||
+		func_id == BPF_FUNC_skc_to_tcp_request_sock;
 }
 
 /* string representation of 'enum bpf_reg_type' */
@@ -3953,6 +3958,7 @@ static int resolve_map_arg_type(struct bpf_verifier_env *env,
 
 struct bpf_reg_types {
 	const enum bpf_reg_type types[10];
+	u32 *btf_id;
 };
 
 static const struct bpf_reg_types map_key_value_types = {
@@ -3973,6 +3979,17 @@ static const struct bpf_reg_types sock_types = {
 	},
 };
 
+static const struct bpf_reg_types btf_id_sock_common_types = {
+	.types = {
+		PTR_TO_SOCK_COMMON,
+		PTR_TO_SOCKET,
+		PTR_TO_TCP_SOCK,
+		PTR_TO_XDP_SOCK,
+		PTR_TO_BTF_ID,
+	},
+	.btf_id = &btf_sock_ids[BTF_SOCK_TYPE_SOCK_COMMON],
+};
+
 static const struct bpf_reg_types mem_types = {
 	.types = {
 		PTR_TO_STACK,
@@ -4014,6 +4031,7 @@ static const struct bpf_reg_types *compatible_reg_types[__BPF_ARG_TYPE_MAX] = {
 	[ARG_PTR_TO_CTX]		= &context_types,
 	[ARG_PTR_TO_CTX_OR_NULL]	= &context_types,
 	[ARG_PTR_TO_SOCK_COMMON]	= &sock_types,
+	[ARG_PTR_TO_BTF_ID_SOCK_COMMON]	= &btf_id_sock_common_types,
 	[ARG_PTR_TO_SOCKET]		= &fullsock_types,
 	[ARG_PTR_TO_SOCKET_OR_NULL]	= &fullsock_types,
 	[ARG_PTR_TO_BTF_ID]		= &btf_ptr_types,
@@ -4028,19 +4046,27 @@ static const struct bpf_reg_types *compatible_reg_types[__BPF_ARG_TYPE_MAX] = {
 };
 
 static int check_reg_type(struct bpf_verifier_env *env, u32 regno,
-			  const struct bpf_reg_types *compatible)
+			  enum bpf_arg_type arg_type,
+			  const u32 *arg_btf_id)
 {
 	struct bpf_reg_state *regs = cur_regs(env), *reg = &regs[regno];
 	enum bpf_reg_type expected, type = reg->type;
+	const struct bpf_reg_types *compatible;
 	int i, j;
 
+	compatible = compatible_reg_types[arg_type];
+	if (!compatible) {
+		verbose(env, "verifier internal error: unsupported arg type %d\n", arg_type);
+		return -EFAULT;
+	}
+
 	for (i = 0; i < ARRAY_SIZE(compatible->types); i++) {
 		expected = compatible->types[i];
 		if (expected == NOT_INIT)
 			break;
 
 		if (type == expected)
-			return 0;
+			goto found;
 	}
 
 	verbose(env, "R%d type=%s expected=", regno, reg_type_str[type]);
@@ -4048,6 +4074,33 @@ static int check_reg_type(struct bpf_verifier_env *env, u32 regno,
 		verbose(env, "%s, ", reg_type_str[compatible->types[j]]);
 	verbose(env, "%s\n", reg_type_str[compatible->types[j]]);
 	return -EACCES;
+
+found:
+	if (type == PTR_TO_BTF_ID) {
+		if (!arg_btf_id) {
+			if (!compatible->btf_id) {
+				verbose(env, "verifier internal error: missing arg compatible BTF ID\n");
+				return -EFAULT;
+			}
+			arg_btf_id = compatible->btf_id;
+		}
+
+		if (!btf_struct_ids_match(&env->log, reg->off, reg->btf_id,
+					  *arg_btf_id)) {
+			verbose(env, "R%d is of type %s but %s is expected\n",
+				regno, kernel_type_name(reg->btf_id),
+				kernel_type_name(*arg_btf_id));
+			return -EACCES;
+		}
+
+		if (!tnum_is_const(reg->var_off) || reg->var_off.value) {
+			verbose(env, "R%d is a pointer to in-kernel struct with non-zero offset\n",
+				regno);
+			return -EACCES;
+		}
+	}
+
+	return 0;
 }
 
 static int check_func_arg(struct bpf_verifier_env *env, u32 arg,
@@ -4057,7 +4110,6 @@ static int check_func_arg(struct bpf_verifier_env *env, u32 arg,
 	u32 regno = BPF_REG_1 + arg;
 	struct bpf_reg_state *regs = cur_regs(env), *reg = &regs[regno];
 	enum bpf_arg_type arg_type = fn->arg_type[arg];
-	const struct bpf_reg_types *compatible;
 	enum bpf_reg_type type = reg->type;
 	int err = 0;
 
@@ -4097,35 +4149,11 @@ static int check_func_arg(struct bpf_verifier_env *env, u32 arg,
 		 */
 		goto skip_type_check;
 
-	compatible = compatible_reg_types[arg_type];
-	if (!compatible) {
-		verbose(env, "verifier internal error: unsupported arg type %d\n", arg_type);
-		return -EFAULT;
-	}
-
-	err = check_reg_type(env, regno, compatible);
+	err = check_reg_type(env, regno, arg_type, fn->arg_btf_id[arg]);
 	if (err)
 		return err;
 
-	if (type == PTR_TO_BTF_ID) {
-		const u32 *btf_id = fn->arg_btf_id[arg];
-
-		if (!btf_id) {
-			verbose(env, "verifier internal error: missing BTF ID\n");
-			return -EFAULT;
-		}
-
-		if (!btf_struct_ids_match(&env->log, reg->off, reg->btf_id, *btf_id)) {
-			verbose(env, "R%d is of type %s but %s is expected\n",
-				regno, kernel_type_name(reg->btf_id), kernel_type_name(*btf_id));
-			return -EACCES;
-		}
-		if (!tnum_is_const(reg->var_off) || reg->var_off.value) {
-			verbose(env, "R%d is a pointer to in-kernel struct with non-zero offset\n",
-				regno);
-			return -EACCES;
-		}
-	} else if (type == PTR_TO_CTX) {
+	if (type == PTR_TO_CTX) {
 		err = check_ctx_reg(env, reg, regno);
 		if (err < 0)
 			return err;
@@ -4573,10 +4601,14 @@ static bool check_btf_id_ok(const struct bpf_func_proto *fn)
 {
 	int i;
 
-	for (i = 0; i < ARRAY_SIZE(fn->arg_type); i++)
+	for (i = 0; i < ARRAY_SIZE(fn->arg_type); i++) {
 		if (fn->arg_type[i] == ARG_PTR_TO_BTF_ID && !fn->arg_btf_id[i])
 			return false;
 
+		if (fn->arg_type[i] != ARG_PTR_TO_BTF_ID && fn->arg_btf_id[i])
+			return false;
+	}
+
 	return true;
 }
 

net/core/bpf_sk_storage.c

@@ -269,7 +269,7 @@ BPF_CALL_4(bpf_sk_storage_get, struct bpf_map *, map, struct sock *, sk,
 {
 	struct bpf_local_storage_data *sdata;
 
-	if (flags > BPF_SK_STORAGE_GET_F_CREATE)
+	if (!sk || !sk_fullsock(sk) || flags > BPF_SK_STORAGE_GET_F_CREATE)
 		return (unsigned long)NULL;
 
 	sdata = sk_storage_lookup(sk, map, true);
@@ -299,6 +299,9 @@ BPF_CALL_4(bpf_sk_storage_get, struct bpf_map *, map, struct sock *, sk,
 
 BPF_CALL_2(bpf_sk_storage_delete, struct bpf_map *, map, struct sock *, sk)
 {
+	if (!sk || !sk_fullsock(sk))
+		return -EINVAL;
+
 	if (refcount_inc_not_zero(&sk->sk_refcnt)) {
 		int err;
 
@@ -355,7 +358,7 @@ const struct bpf_func_proto bpf_sk_storage_get_proto = {
 	.gpl_only	= false,
 	.ret_type	= RET_PTR_TO_MAP_VALUE_OR_NULL,
 	.arg1_type	= ARG_CONST_MAP_PTR,
-	.arg2_type	= ARG_PTR_TO_SOCKET,
+	.arg2_type	= ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 	.arg3_type	= ARG_PTR_TO_MAP_VALUE_OR_NULL,
 	.arg4_type	= ARG_ANYTHING,
 };
@@ -375,27 +378,7 @@ const struct bpf_func_proto bpf_sk_storage_delete_proto = {
 	.gpl_only	= false,
 	.ret_type	= RET_INTEGER,
 	.arg1_type	= ARG_CONST_MAP_PTR,
-	.arg2_type	= ARG_PTR_TO_SOCKET,
-};
-
-const struct bpf_func_proto sk_storage_get_btf_proto = {
-	.func		= bpf_sk_storage_get,
-	.gpl_only	= false,
-	.ret_type	= RET_PTR_TO_MAP_VALUE_OR_NULL,
-	.arg1_type	= ARG_CONST_MAP_PTR,
-	.arg2_type	= ARG_PTR_TO_BTF_ID,
-	.arg2_btf_id	= &btf_sock_ids[BTF_SOCK_TYPE_SOCK],
-	.arg3_type	= ARG_PTR_TO_MAP_VALUE_OR_NULL,
-	.arg4_type	= ARG_ANYTHING,
-};
-
-const struct bpf_func_proto sk_storage_delete_btf_proto = {
-	.func		= bpf_sk_storage_delete,
-	.gpl_only	= false,
-	.ret_type	= RET_INTEGER,
-	.arg1_type	= ARG_CONST_MAP_PTR,
-	.arg2_type	= ARG_PTR_TO_BTF_ID,
-	.arg2_btf_id	= &btf_sock_ids[BTF_SOCK_TYPE_SOCK],
+	.arg2_type	= ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 };
 
 struct bpf_sk_storage_diag {

net/core/filter.c

@@ -77,6 +77,9 @@
 #include <net/transp_v6.h>
 #include <linux/btf_ids.h>
 
+static const struct bpf_func_proto *
+bpf_sk_base_func_proto(enum bpf_func_id func_id);
+
 int copy_bpf_fprog_from_user(struct sock_fprog *dst, sockptr_t src, int len)
 {
 	if (in_compat_syscall()) {
@@ -4085,18 +4088,17 @@ static inline u64 __bpf_sk_cgroup_id(struct sock *sk)
 {
 	struct cgroup *cgrp;
 
+	sk = sk_to_full_sk(sk);
+	if (!sk || !sk_fullsock(sk))
+		return 0;
+
 	cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
 	return cgroup_id(cgrp);
 }
 
 BPF_CALL_1(bpf_skb_cgroup_id, const struct sk_buff *, skb)
 {
-	struct sock *sk = skb_to_full_sk(skb);
-
-	if (!sk || !sk_fullsock(sk))
-		return 0;
-
-	return __bpf_sk_cgroup_id(sk);
+	return __bpf_sk_cgroup_id(skb->sk);
 }
 
 static const struct bpf_func_proto bpf_skb_cgroup_id_proto = {
@@ -4112,6 +4114,10 @@ static inline u64 __bpf_sk_ancestor_cgroup_id(struct sock *sk,
 	struct cgroup *ancestor;
 	struct cgroup *cgrp;
 
+	sk = sk_to_full_sk(sk);
+	if (!sk || !sk_fullsock(sk))
+		return 0;
+
 	cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
 	ancestor = cgroup_ancestor(cgrp, ancestor_level);
 	if (!ancestor)
@@ -4123,12 +4129,7 @@ static inline u64 __bpf_sk_ancestor_cgroup_id(struct sock *sk,
 BPF_CALL_2(bpf_skb_ancestor_cgroup_id, const struct sk_buff *, skb, int,
 	   ancestor_level)
 {
-	struct sock *sk = skb_to_full_sk(skb);
-
-	if (!sk || !sk_fullsock(sk))
-		return 0;
-
-	return __bpf_sk_ancestor_cgroup_id(sk, ancestor_level);
+	return __bpf_sk_ancestor_cgroup_id(skb->sk, ancestor_level);
 }
 
 static const struct bpf_func_proto bpf_skb_ancestor_cgroup_id_proto = {
@@ -4148,7 +4149,7 @@ static const struct bpf_func_proto bpf_sk_cgroup_id_proto = {
 	.func           = bpf_sk_cgroup_id,
 	.gpl_only       = false,
 	.ret_type       = RET_INTEGER,
-	.arg1_type      = ARG_PTR_TO_SOCKET,
+	.arg1_type      = ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 };
 
 BPF_CALL_2(bpf_sk_ancestor_cgroup_id, struct sock *, sk, int, ancestor_level)
@@ -4160,7 +4161,7 @@ static const struct bpf_func_proto bpf_sk_ancestor_cgroup_id_proto = {
 	.func           = bpf_sk_ancestor_cgroup_id,
 	.gpl_only       = false,
 	.ret_type       = RET_INTEGER,
-	.arg1_type      = ARG_PTR_TO_SOCKET,
+	.arg1_type      = ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 	.arg2_type      = ARG_ANYTHING,
 };
 #endif
@@ -5694,7 +5695,7 @@ static const struct bpf_func_proto bpf_sk_lookup_udp_proto = {
 
 BPF_CALL_1(bpf_sk_release, struct sock *, sk)
 {
-	if (sk_is_refcounted(sk))
+	if (sk && sk_is_refcounted(sk))
 		sock_gen_put(sk);
 	return 0;
 }
@@ -5703,7 +5704,7 @@ static const struct bpf_func_proto bpf_sk_release_proto = {
 	.func		= bpf_sk_release,
 	.gpl_only	= false,
 	.ret_type	= RET_INTEGER,
-	.arg1_type	= ARG_PTR_TO_SOCK_COMMON,
+	.arg1_type	= ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 };
 
 BPF_CALL_5(bpf_xdp_sk_lookup_udp, struct xdp_buff *, ctx,
@@ -6085,7 +6086,7 @@ BPF_CALL_5(bpf_tcp_check_syncookie, struct sock *, sk, void *, iph, u32, iph_len
 	u32 cookie;
 	int ret;
 
-	if (unlikely(th_len < sizeof(*th)))
+	if (unlikely(!sk || th_len < sizeof(*th)))
 		return -EINVAL;
 
 	/* sk_listener() allows TCP_NEW_SYN_RECV, which makes no sense here. */
@@ -6138,7 +6139,7 @@ static const struct bpf_func_proto bpf_tcp_check_syncookie_proto = {
 	.gpl_only	= true,
 	.pkt_access	= true,
 	.ret_type	= RET_INTEGER,
-	.arg1_type	= ARG_PTR_TO_SOCK_COMMON,
+	.arg1_type	= ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 	.arg2_type	= ARG_PTR_TO_MEM,
 	.arg3_type	= ARG_CONST_SIZE,
 	.arg4_type	= ARG_PTR_TO_MEM,
@@ -6152,7 +6153,7 @@ BPF_CALL_5(bpf_tcp_gen_syncookie, struct sock *, sk, void *, iph, u32, iph_len,
 	u32 cookie;
 	u16 mss;
 
-	if (unlikely(th_len < sizeof(*th) || th_len != th->doff * 4))
+	if (unlikely(!sk || th_len < sizeof(*th) || th_len != th->doff * 4))
 		return -EINVAL;
 
 	if (sk->sk_protocol != IPPROTO_TCP || sk->sk_state != TCP_LISTEN)
@@ -6207,7 +6208,7 @@ static const struct bpf_func_proto bpf_tcp_gen_syncookie_proto = {
 	.gpl_only	= true, /* __cookie_v*_init_sequence() is GPL */
 	.pkt_access	= true,
 	.ret_type	= RET_INTEGER,
-	.arg1_type	= ARG_PTR_TO_SOCK_COMMON,
+	.arg1_type	= ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 	.arg2_type	= ARG_PTR_TO_MEM,
 	.arg3_type	= ARG_CONST_SIZE,
 	.arg4_type	= ARG_PTR_TO_MEM,
@@ -6216,7 +6217,7 @@ static const struct bpf_func_proto bpf_tcp_gen_syncookie_proto = {
 
 BPF_CALL_3(bpf_sk_assign, struct sk_buff *, skb, struct sock *, sk, u64, flags)
 {
-	if (flags != 0)
+	if (!sk || flags != 0)
 		return -EINVAL;
 	if (!skb_at_tc_ingress(skb))
 		return -EOPNOTSUPP;
@@ -6240,7 +6241,7 @@ static const struct bpf_func_proto bpf_sk_assign_proto = {
 	.gpl_only	= false,
 	.ret_type	= RET_INTEGER,
 	.arg1_type      = ARG_PTR_TO_CTX,
-	.arg2_type      = ARG_PTR_TO_SOCK_COMMON,
+	.arg2_type      = ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 	.arg3_type	= ARG_ANYTHING,
 };
 
@@ -6620,7 +6621,7 @@ sock_addr_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 			return NULL;
 		}
 	default:
-		return bpf_base_func_proto(func_id);
+		return bpf_sk_base_func_proto(func_id);
 	}
 }
 
@@ -6639,7 +6640,7 @@ sk_filter_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 	case BPF_FUNC_perf_event_output:
 		return &bpf_skb_event_output_proto;
 	default:
-		return bpf_base_func_proto(func_id);
+		return bpf_sk_base_func_proto(func_id);
 	}
 }
 
@@ -6800,7 +6801,7 @@ tc_cls_act_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_sk_assign_proto;
 #endif
 	default:
-		return bpf_base_func_proto(func_id);
+		return bpf_sk_base_func_proto(func_id);
 	}
 }
 
@@ -6841,7 +6842,7 @@ xdp_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_tcp_gen_syncookie_proto;
 #endif
 	default:
-		return bpf_base_func_proto(func_id);
+		return bpf_sk_base_func_proto(func_id);
 	}
 }
 
@@ -6883,7 +6884,7 @@ sock_ops_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_tcp_sock_proto;
 #endif /* CONFIG_INET */
 	default:
-		return bpf_base_func_proto(func_id);
+		return bpf_sk_base_func_proto(func_id);
 	}
 }
 
@@ -6929,7 +6930,7 @@ sk_msg_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_get_cgroup_classid_curr_proto;
 #endif
 	default:
-		return bpf_base_func_proto(func_id);
+		return bpf_sk_base_func_proto(func_id);
 	}
 }
 
@@ -6971,7 +6972,7 @@ sk_skb_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_skc_lookup_tcp_proto;
 #endif
 	default:
-		return bpf_base_func_proto(func_id);
+		return bpf_sk_base_func_proto(func_id);
 	}
 }
 
@@ -6982,7 +6983,7 @@ flow_dissector_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 	case BPF_FUNC_skb_load_bytes:
 		return &bpf_flow_dissector_load_bytes_proto;
 	default:
-		return bpf_base_func_proto(func_id);
+		return bpf_sk_base_func_proto(func_id);
 	}
 }
 
@@ -7009,7 +7010,7 @@ lwt_out_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 	case BPF_FUNC_skb_under_cgroup:
 		return &bpf_skb_under_cgroup_proto;
 	default:
-		return bpf_base_func_proto(func_id);
+		return bpf_sk_base_func_proto(func_id);
 	}
 }
 
@@ -9746,7 +9747,7 @@ sk_lookup_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 	case BPF_FUNC_sk_release:
 		return &bpf_sk_release_proto;
 	default:
-		return bpf_base_func_proto(func_id);
+		return bpf_sk_base_func_proto(func_id);
 	}
 }
 
@@ -9913,8 +9914,7 @@ const struct bpf_func_proto bpf_skc_to_tcp6_sock_proto = {
 	.func			= bpf_skc_to_tcp6_sock,
 	.gpl_only		= false,
 	.ret_type		= RET_PTR_TO_BTF_ID_OR_NULL,
-	.arg1_type		= ARG_PTR_TO_BTF_ID,
-	.arg1_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_SOCK_COMMON],
+	.arg1_type		= ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 	.ret_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_TCP6],
 };
 
@@ -9930,8 +9930,7 @@ const struct bpf_func_proto bpf_skc_to_tcp_sock_proto = {
 	.func			= bpf_skc_to_tcp_sock,
 	.gpl_only		= false,
 	.ret_type		= RET_PTR_TO_BTF_ID_OR_NULL,
-	.arg1_type		= ARG_PTR_TO_BTF_ID,
-	.arg1_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_SOCK_COMMON],
+	.arg1_type		= ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 	.ret_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_TCP],
 };
 
@@ -9954,8 +9953,7 @@ const struct bpf_func_proto bpf_skc_to_tcp_timewait_sock_proto = {
 	.func			= bpf_skc_to_tcp_timewait_sock,
 	.gpl_only		= false,
 	.ret_type		= RET_PTR_TO_BTF_ID_OR_NULL,
-	.arg1_type		= ARG_PTR_TO_BTF_ID,
-	.arg1_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_SOCK_COMMON],
+	.arg1_type		= ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 	.ret_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_TCP_TW],
 };
 
@@ -9978,8 +9976,7 @@ const struct bpf_func_proto bpf_skc_to_tcp_request_sock_proto = {
 	.func			= bpf_skc_to_tcp_request_sock,
 	.gpl_only		= false,
 	.ret_type		= RET_PTR_TO_BTF_ID_OR_NULL,
-	.arg1_type		= ARG_PTR_TO_BTF_ID,
-	.arg1_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_SOCK_COMMON],
+	.arg1_type		= ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 	.ret_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_TCP_REQ],
 };
 
@@ -10000,7 +9997,37 @@ const struct bpf_func_proto bpf_skc_to_udp6_sock_proto = {
 	.func			= bpf_skc_to_udp6_sock,
 	.gpl_only		= false,
 	.ret_type		= RET_PTR_TO_BTF_ID_OR_NULL,
-	.arg1_type		= ARG_PTR_TO_BTF_ID,
-	.arg1_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_SOCK_COMMON],
+	.arg1_type		= ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 	.ret_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_UDP6],
 };
+
+static const struct bpf_func_proto *
+bpf_sk_base_func_proto(enum bpf_func_id func_id)
+{
+	const struct bpf_func_proto *func;
+
+	switch (func_id) {
+	case BPF_FUNC_skc_to_tcp6_sock:
+		func = &bpf_skc_to_tcp6_sock_proto;
+		break;
+	case BPF_FUNC_skc_to_tcp_sock:
+		func = &bpf_skc_to_tcp_sock_proto;
+		break;
+	case BPF_FUNC_skc_to_tcp_timewait_sock:
+		func = &bpf_skc_to_tcp_timewait_sock_proto;
+		break;
+	case BPF_FUNC_skc_to_tcp_request_sock:
+		func = &bpf_skc_to_tcp_request_sock_proto;
+		break;
+	case BPF_FUNC_skc_to_udp6_sock:
+		func = &bpf_skc_to_udp6_sock_proto;
+		break;
+	default:
+		return bpf_base_func_proto(func_id);
+	}
+
+	if (!perfmon_capable())
+		return NULL;
+
+	return func;
+}

net/ipv4/bpf_tcp_ca.c

@@ -28,22 +28,6 @@ static u32 unsupported_ops[] = {
 static const struct btf_type *tcp_sock_type;
 static u32 tcp_sock_id, sock_id;
 
-static struct bpf_func_proto btf_sk_storage_get_proto __read_mostly;
-static struct bpf_func_proto btf_sk_storage_delete_proto __read_mostly;
-
-static void convert_sk_func_proto(struct bpf_func_proto *to, const struct bpf_func_proto *from)
-{
-	int i;
-
-	*to = *from;
-	for (i = 0; i < ARRAY_SIZE(to->arg_type); i++) {
-		if (to->arg_type[i] == ARG_PTR_TO_SOCKET) {
-			to->arg_type[i] = ARG_PTR_TO_BTF_ID;
-			to->arg_btf_id[i] = &tcp_sock_id;
-		}
-	}
-}
-
 static int bpf_tcp_ca_init(struct btf *btf)
 {
 	s32 type_id;
@@ -59,9 +43,6 @@ static int bpf_tcp_ca_init(struct btf *btf)
 	tcp_sock_id = type_id;
 	tcp_sock_type = btf_type_by_id(btf, tcp_sock_id);
 
-	convert_sk_func_proto(&btf_sk_storage_get_proto, &bpf_sk_storage_get_proto);
-	convert_sk_func_proto(&btf_sk_storage_delete_proto, &bpf_sk_storage_delete_proto);
-
 	return 0;
 }
 
@@ -188,9 +169,9 @@ bpf_tcp_ca_get_func_proto(enum bpf_func_id func_id,
 	case BPF_FUNC_tcp_send_ack:
 		return &bpf_tcp_send_ack_proto;
 	case BPF_FUNC_sk_storage_get:
-		return &btf_sk_storage_get_proto;
+		return &bpf_sk_storage_get_proto;
 	case BPF_FUNC_sk_storage_delete:
-		return &btf_sk_storage_delete_proto;
+		return &bpf_sk_storage_delete_proto;
 	default:
 		return bpf_base_func_proto(func_id);
 	}

tools/include/uapi/linux/bpf.h

@@ -2512,7 +2512,7 @@ union bpf_attr {
  *		result is from *reuse*\ **->socks**\ [] using the hash of the
  *		tuple.
  *
- * long bpf_sk_release(struct bpf_sock *sock)
+ * long bpf_sk_release(void *sock)
  *	Description
  *		Release the reference held by *sock*. *sock* must be a
  *		non-**NULL** pointer that was returned from
@@ -2692,7 +2692,7 @@ union bpf_attr {
  *		result is from *reuse*\ **->socks**\ [] using the hash of the
  *		tuple.
  *
- * long bpf_tcp_check_syncookie(struct bpf_sock *sk, void *iph, u32 iph_len, struct tcphdr *th, u32 th_len)
+ * long bpf_tcp_check_syncookie(void *sk, void *iph, u32 iph_len, struct tcphdr *th, u32 th_len)
  * 	Description
  * 		Check whether *iph* and *th* contain a valid SYN cookie ACK for
  * 		the listening socket in *sk*.
@@ -2861,6 +2861,7 @@ union bpf_attr {
  *		0 on success.
  *
  *		**-ENOENT** if the bpf-local-storage cannot be found.
+ *		**-EINVAL** if sk is not a fullsock (e.g. a request_sock).
  *
  * long bpf_send_signal(u32 sig)
  *	Description
@@ -2877,7 +2878,7 @@ union bpf_attr {
  *
  *		**-EAGAIN** if bpf program can try again.
  *
- * s64 bpf_tcp_gen_syncookie(struct bpf_sock *sk, void *iph, u32 iph_len, struct tcphdr *th, u32 th_len)
+ * s64 bpf_tcp_gen_syncookie(void *sk, void *iph, u32 iph_len, struct tcphdr *th, u32 th_len)
  *	Description
  *		Try to issue a SYN cookie for the packet with corresponding
  *		IP/TCP headers, *iph* and *th*, on the listening socket in *sk*.
@@ -3106,7 +3107,7 @@ union bpf_attr {
  * 	Return
  * 		The id is returned or 0 in case the id could not be retrieved.
  *
- * long bpf_sk_assign(struct sk_buff *skb, struct bpf_sock *sk, u64 flags)
+ * long bpf_sk_assign(struct sk_buff *skb, void *sk, u64 flags)
  *	Description
  *		Helper is overloaded depending on BPF program type. This
  *		description applies to **BPF_PROG_TYPE_SCHED_CLS** and
@@ -3234,11 +3235,11 @@ union bpf_attr {
  *
  *		**-EOVERFLOW** if an overflow happened: The same object will be tried again.
  *
- * u64 bpf_sk_cgroup_id(struct bpf_sock *sk)
+ * u64 bpf_sk_cgroup_id(void *sk)
  *	Description
  *		Return the cgroup v2 id of the socket *sk*.
  *
- *		*sk* must be a non-**NULL** pointer to a full socket, e.g. one
+ *		*sk* must be a non-**NULL** pointer to a socket, e.g. one
  *		returned from **bpf_sk_lookup_xxx**\ (),
  *		**bpf_sk_fullsock**\ (), etc. The format of returned id is
  *		same as in **bpf_skb_cgroup_id**\ ().
@@ -3248,7 +3249,7 @@ union bpf_attr {
  *	Return
  *		The id is returned or 0 in case the id could not be retrieved.
  *
- * u64 bpf_sk_ancestor_cgroup_id(struct bpf_sock *sk, int ancestor_level)
+ * u64 bpf_sk_ancestor_cgroup_id(void *sk, int ancestor_level)
  *	Description
  *		Return id of cgroup v2 that is ancestor of cgroup associated
  *		with the *sk* at the *ancestor_level*.  The root cgroup is at

tools/testing/selftests/bpf/.gitignore

@@ -13,7 +13,6 @@ test_verifier_log
 feature
 test_sock
 test_sock_addr
-test_sock_fields
 urandom_read
 test_sockmap
 test_lirc_mode2_user

tools/testing/selftests/bpf/Makefile

@@ -35,7 +35,7 @@ TEST_GEN_PROGS = test_verifier test_tag test_maps test_lru_map test_lpm_map test
 	test_verifier_log test_dev_cgroup test_tcpbpf_user \
 	test_sock test_sockmap get_cgroup_id_user test_socket_cookie \
 	test_cgroup_storage \
-	test_netcnt test_tcpnotify_user test_sock_fields test_sysctl \
+	test_netcnt test_tcpnotify_user test_sysctl \
 	test_progs-no_alu32 \
 	test_current_pid_tgid_new_ns
 

tools/testing/selftests/bpf/bpf_tcp_helpers.h

@@ -16,6 +16,7 @@ BPF_PROG(name, args)
 
 struct sock_common {
 	unsigned char	skc_state;
+	__u16		skc_num;
 } __attribute__((preserve_access_index));
 
 enum sk_pacing {
@@ -45,6 +46,10 @@ struct inet_connection_sock {
 	__u64			  icsk_ca_priv[104 / sizeof(__u64)];
 } __attribute__((preserve_access_index));
 
+struct request_sock {
+	struct sock_common		__req_common;
+} __attribute__((preserve_access_index));
+
 struct tcp_sock {
 	struct inet_connection_sock	inet_conn;
 
@@ -115,14 +120,6 @@ enum tcp_ca_event {
 	CA_EVENT_ECN_IS_CE = 5,
 };
 
-enum tcp_ca_state {
-	TCP_CA_Open = 0,
-	TCP_CA_Disorder = 1,
-	TCP_CA_CWR = 2,
-	TCP_CA_Recovery = 3,
-	TCP_CA_Loss = 4
-};
-
 struct ack_sample {
 	__u32 pkts_acked;
 	__s32 rtt_us;

tools/testing/selftests/bpf/prog_tests/btf_skc_cls_ingress.c

+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2020 Facebook */
+
+#define _GNU_SOURCE
+#include <netinet/in.h>
+#include <arpa/inet.h>
+#include <unistd.h>
+#include <stdlib.h>
+#include <string.h>
+#include <errno.h>
+#include <sched.h>
+#include <linux/compiler.h>
+#include <bpf/libbpf.h>
+
+#include "network_helpers.h"
+#include "test_progs.h"
+#include "test_btf_skc_cls_ingress.skel.h"
+
+struct test_btf_skc_cls_ingress *skel;
+struct sockaddr_in6 srv_sa6;
+static __u32 duration;
+
+#define PROG_PIN_FILE "/sys/fs/bpf/btf_skc_cls_ingress"
+
+static int write_sysctl(const char *sysctl, const char *value)
+{
+	int fd, err, len;
+
+	fd = open(sysctl, O_WRONLY);
+	if (CHECK(fd == -1, "open sysctl", "open(%s): %s (%d)\n",
+		  sysctl, strerror(errno), errno))
+		return -1;
+
+	len = strlen(value);
+	err = write(fd, value, len);
+	close(fd);
+	if (CHECK(err != len, "write sysctl",
+		  "write(%s, %s, %d): err:%d %s (%d)\n",
+		  sysctl, value, len, err, strerror(errno), errno))
+		return -1;
+
+	return 0;
+}
+
+static int prepare_netns(void)
+{
+	if (CHECK(unshare(CLONE_NEWNET), "create netns",
+		  "unshare(CLONE_NEWNET): %s (%d)",
+		  strerror(errno), errno))
+		return -1;
+
+	if (CHECK(system("ip link set dev lo up"),
+		  "ip link set dev lo up", "failed\n"))
+		return -1;
+
+	if (CHECK(system("tc qdisc add dev lo clsact"),
+		  "tc qdisc add dev lo clsact", "failed\n"))
+		return -1;
+
+	if (CHECK(system("tc filter add dev lo ingress bpf direct-action object-pinned " PROG_PIN_FILE),
+		  "install tc cls-prog at ingress", "failed\n"))
+		return -1;
+
+	/* Ensure 20 bytes options (i.e. in total 40 bytes tcp header) for the
+	 * bpf_tcp_gen_syncookie() helper.
+	 */
+	if (write_sysctl("/proc/sys/net/ipv4/tcp_window_scaling", "1") ||
+	    write_sysctl("/proc/sys/net/ipv4/tcp_timestamps", "1") ||
+	    write_sysctl("/proc/sys/net/ipv4/tcp_sack", "1"))
+		return -1;
+
+	return 0;
+}
+
+static void reset_test(void)
+{
+	memset(&skel->bss->srv_sa6, 0, sizeof(skel->bss->srv_sa6));
+	skel->bss->listen_tp_sport = 0;
+	skel->bss->req_sk_sport = 0;
+	skel->bss->recv_cookie = 0;
+	skel->bss->gen_cookie = 0;
+	skel->bss->linum = 0;
+}
+
+static void print_err_line(void)
+{
+	if (skel->bss->linum)
+		printf("bpf prog error at line %u\n", skel->bss->linum);
+}
+
+static void test_conn(void)
+{
+	int listen_fd = -1, cli_fd = -1, err;
+	socklen_t addrlen = sizeof(srv_sa6);
+	int srv_port;
+
+	if (write_sysctl("/proc/sys/net/ipv4/tcp_syncookies", "1"))
+		return;
+
+	listen_fd = start_server(AF_INET6, SOCK_STREAM, "::1", 0, 0);
+	if (CHECK_FAIL(listen_fd == -1))
+		return;
+
+	err = getsockname(listen_fd, (struct sockaddr *)&srv_sa6, &addrlen);
+	if (CHECK(err, "getsockname(listen_fd)", "err:%d errno:%d\n", err,
+		  errno))
+		goto done;
+	memcpy(&skel->bss->srv_sa6, &srv_sa6, sizeof(srv_sa6));
+	srv_port = ntohs(srv_sa6.sin6_port);
+
+	cli_fd = connect_to_fd(listen_fd, 0);
+	if (CHECK_FAIL(cli_fd == -1))
+		goto done;
+
+	if (CHECK(skel->bss->listen_tp_sport != srv_port ||
+		  skel->bss->req_sk_sport != srv_port,
+		  "Unexpected sk src port",
+		  "listen_tp_sport:%u req_sk_sport:%u expected:%u\n",
+		  skel->bss->listen_tp_sport, skel->bss->req_sk_sport,
+		  srv_port))
+		goto done;
+
+	if (CHECK(skel->bss->gen_cookie || skel->bss->recv_cookie,
+		  "Unexpected syncookie states",
+		  "gen_cookie:%u recv_cookie:%u\n",
+		  skel->bss->gen_cookie, skel->bss->recv_cookie))
+		goto done;
+
+	CHECK(skel->bss->linum, "bpf prog detected error", "at line %u\n",
+	      skel->bss->linum);
+
+done:
+	if (listen_fd != -1)
+		close(listen_fd);
+	if (cli_fd != -1)
+		close(cli_fd);
+}
+
+static void test_syncookie(void)
+{
+	int listen_fd = -1, cli_fd = -1, err;
+	socklen_t addrlen = sizeof(srv_sa6);
+	int srv_port;
+
+	/* Enforce syncookie mode */
+	if (write_sysctl("/proc/sys/net/ipv4/tcp_syncookies", "2"))
+		return;
+
+	listen_fd = start_server(AF_INET6, SOCK_STREAM, "::1", 0, 0);
+	if (CHECK_FAIL(listen_fd == -1))
+		return;
+
+	err = getsockname(listen_fd, (struct sockaddr *)&srv_sa6, &addrlen);
+	if (CHECK(err, "getsockname(listen_fd)", "err:%d errno:%d\n", err,
+		  errno))
+		goto done;
+	memcpy(&skel->bss->srv_sa6, &srv_sa6, sizeof(srv_sa6));
+	srv_port = ntohs(srv_sa6.sin6_port);
+
+	cli_fd = connect_to_fd(listen_fd, 0);
+	if (CHECK_FAIL(cli_fd == -1))
+		goto done;
+
+	if (CHECK(skel->bss->listen_tp_sport != srv_port,
+		  "Unexpected tp src port",
+		  "listen_tp_sport:%u expected:%u\n",
+		  skel->bss->listen_tp_sport, srv_port))
+		goto done;
+
+	if (CHECK(skel->bss->req_sk_sport,
+		  "Unexpected req_sk src port",
+		  "req_sk_sport:%u expected:0\n",
+		   skel->bss->req_sk_sport))
+		goto done;
+
+	if (CHECK(!skel->bss->gen_cookie ||
+		  skel->bss->gen_cookie != skel->bss->recv_cookie,
+		  "Unexpected syncookie states",
+		  "gen_cookie:%u recv_cookie:%u\n",
+		  skel->bss->gen_cookie, skel->bss->recv_cookie))
+		goto done;
+
+	CHECK(skel->bss->linum, "bpf prog detected error", "at line %u\n",
+	      skel->bss->linum);
+
+done:
+	if (listen_fd != -1)
+		close(listen_fd);
+	if (cli_fd != -1)
+		close(cli_fd);
+}
+
+struct test {
+	const char *desc;
+	void (*run)(void);
+};
+
+#define DEF_TEST(name) { #name, test_##name }
+static struct test tests[] = {
+	DEF_TEST(conn),
+	DEF_TEST(syncookie),
+};
+
+void test_btf_skc_cls_ingress(void)
+{
+	int i, err;
+
+	skel = test_btf_skc_cls_ingress__open_and_load();
+	if (CHECK(!skel, "test_btf_skc_cls_ingress__open_and_load", "failed\n"))
+		return;
+
+	err = bpf_program__pin(skel->progs.cls_ingress, PROG_PIN_FILE);
+	if (CHECK(err, "bpf_program__pin",
+		  "cannot pin bpf prog to %s. err:%d\n", PROG_PIN_FILE, err)) {
+		test_btf_skc_cls_ingress__destroy(skel);
+		return;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(tests); i++) {
+		if (!test__start_subtest(tests[i].desc))
+			continue;
+
+		if (prepare_netns())
+			break;
+
+		tests[i].run();
+
+		print_err_line();
+		reset_test();
+	}
+
+	bpf_program__unpin(skel->progs.cls_ingress, PROG_PIN_FILE);
+	test_btf_skc_cls_ingress__destroy(skel);
+}

tools/testing/selftests/bpf/progs/bpf_cubic.c

@@ -15,6 +15,8 @@
  */
 
 #include <linux/bpf.h>
+#include <linux/stddef.h>
+#include <linux/tcp.h>
 #include "bpf_tcp_helpers.h"
 
 char _license[] SEC("license") = "GPL";

tools/testing/selftests/bpf/progs/bpf_dctcp.c

@@ -9,6 +9,8 @@
 #include <stddef.h>
 #include <linux/bpf.h>
 #include <linux/types.h>
+#include <linux/stddef.h>
+#include <linux/tcp.h>
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 #include "bpf_tcp_helpers.h"

tools/testing/selftests/bpf/progs/test_btf_skc_cls_ingress.c

+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2020 Facebook */
+
+#include <string.h>
+#include <errno.h>
+#include <netinet/in.h>
+#include <linux/stddef.h>
+#include <linux/bpf.h>
+#include <linux/ipv6.h>
+#include <linux/tcp.h>
+#include <linux/if_ether.h>
+#include <linux/pkt_cls.h>
+
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_endian.h>
+#include "bpf_tcp_helpers.h"
+
+struct sockaddr_in6 srv_sa6 = {};
+__u16 listen_tp_sport = 0;
+__u16 req_sk_sport = 0;
+__u32 recv_cookie = 0;
+__u32 gen_cookie = 0;
+__u32 linum = 0;
+
+#define LOG() ({ if (!linum) linum = __LINE__; })
+
+static void test_syncookie_helper(struct ipv6hdr *ip6h, struct tcphdr *th,
+				  struct tcp_sock *tp,
+				  struct __sk_buff *skb)
+{
+	if (th->syn) {
+		__s64 mss_cookie;
+		void *data_end;
+
+		data_end = (void *)(long)(skb->data_end);
+
+		if (th->doff * 4 != 40) {
+			LOG();
+			return;
+		}
+
+		if ((void *)th + 40 > data_end) {
+			LOG();
+			return;
+		}
+
+		mss_cookie = bpf_tcp_gen_syncookie(tp, ip6h, sizeof(*ip6h),
+						   th, 40);
+		if (mss_cookie < 0) {
+			if (mss_cookie != -ENOENT)
+				LOG();
+		} else {
+			gen_cookie = (__u32)mss_cookie;
+		}
+	} else if (gen_cookie) {
+		/* It was in cookie mode */
+		int ret = bpf_tcp_check_syncookie(tp, ip6h, sizeof(*ip6h),
+						  th, sizeof(*th));
+
+		if (ret < 0) {
+			if (ret != -ENOENT)
+				LOG();
+		} else {
+			recv_cookie = bpf_ntohl(th->ack_seq) - 1;
+		}
+	}
+}
+
+static int handle_ip6_tcp(struct ipv6hdr *ip6h, struct __sk_buff *skb)
+{
+	struct bpf_sock_tuple *tuple;
+	struct bpf_sock *bpf_skc;
+	unsigned int tuple_len;
+	struct tcphdr *th;
+	void *data_end;
+
+	data_end = (void *)(long)(skb->data_end);
+
+	th = (struct tcphdr *)(ip6h + 1);
+	if (th + 1 > data_end)
+		return TC_ACT_OK;
+
+	/* Is it the testing traffic? */
+	if (th->dest != srv_sa6.sin6_port)
+		return TC_ACT_OK;
+
+	tuple_len = sizeof(tuple->ipv6);
+	tuple = (struct bpf_sock_tuple *)&ip6h->saddr;
+	if ((void *)tuple + tuple_len > data_end) {
+		LOG();
+		return TC_ACT_OK;
+	}
+
+	bpf_skc = bpf_skc_lookup_tcp(skb, tuple, tuple_len,
+				     BPF_F_CURRENT_NETNS, 0);
+	if (!bpf_skc) {
+		LOG();
+		return TC_ACT_OK;
+	}
+
+	if (bpf_skc->state == BPF_TCP_NEW_SYN_RECV) {
+		struct request_sock *req_sk;
+
+		req_sk = (struct request_sock *)bpf_skc_to_tcp_request_sock(bpf_skc);
+		if (!req_sk) {
+			LOG();
+			goto release;
+		}
+
+		if (bpf_sk_assign(skb, req_sk, 0)) {
+			LOG();
+			goto release;
+		}
+
+		req_sk_sport = req_sk->__req_common.skc_num;
+
+		bpf_sk_release(req_sk);
+		return TC_ACT_OK;
+	} else if (bpf_skc->state == BPF_TCP_LISTEN) {
+		struct tcp_sock *tp;
+
+		tp = bpf_skc_to_tcp_sock(bpf_skc);
+		if (!tp) {
+			LOG();
+			goto release;
+		}
+
+		if (bpf_sk_assign(skb, tp, 0)) {
+			LOG();
+			goto release;
+		}
+
+		listen_tp_sport = tp->inet_conn.icsk_inet.sk.__sk_common.skc_num;
+
+		test_syncookie_helper(ip6h, th, tp, skb);
+		bpf_sk_release(tp);
+		return TC_ACT_OK;
+	}
+
+	if (bpf_sk_assign(skb, bpf_skc, 0))
+		LOG();
+
+release:
+	bpf_sk_release(bpf_skc);
+	return TC_ACT_OK;
+}
+
+SEC("classifier/ingress")
+int cls_ingress(struct __sk_buff *skb)
+{
+	struct ipv6hdr *ip6h;
+	struct ethhdr *eth;
+	void *data_end;
+
+	data_end = (void *)(long)(skb->data_end);
+
+	eth = (struct ethhdr *)(long)(skb->data);
+	if (eth + 1 > data_end)
+		return TC_ACT_OK;
+
+	if (eth->h_proto != bpf_htons(ETH_P_IPV6))
+		return TC_ACT_OK;
+
+	ip6h = (struct ipv6hdr *)(eth + 1);
+	if (ip6h + 1 > data_end)
+		return TC_ACT_OK;
+
+	if (ip6h->nexthdr == IPPROTO_TCP)
+		return handle_ip6_tcp(ip6h, skb);
+
+	return TC_ACT_OK;
+}
+
+char _license[] SEC("license") = "GPL";

tools/testing/selftests/bpf/progs/test_sock_fields_kern.c → tools/testing/selftests/bpf/progs/test_sock_fields.c

@@ -7,19 +7,7 @@
 
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_endian.h>
-
-enum bpf_addr_array_idx {
-	ADDR_SRV_IDX,
-	ADDR_CLI_IDX,
-	__NR_BPF_ADDR_ARRAY_IDX,
-};
-
-enum bpf_result_array_idx {
-	EGRESS_SRV_IDX,
-	EGRESS_CLI_IDX,
-	INGRESS_LISTEN_IDX,
-	__NR_BPF_RESULT_ARRAY_IDX,
-};
+#include "bpf_tcp_helpers.h"
 
 enum bpf_linum_array_idx {
 	EGRESS_LINUM_IDX,
@@ -27,27 +15,6 @@ enum bpf_linum_array_idx {
 	__NR_BPF_LINUM_ARRAY_IDX,
 };
 
-struct {
-	__uint(type, BPF_MAP_TYPE_ARRAY);
-	__uint(max_entries, __NR_BPF_ADDR_ARRAY_IDX);
-	__type(key, __u32);
-	__type(value, struct sockaddr_in6);
-} addr_map SEC(".maps");
-
-struct {
-	__uint(type, BPF_MAP_TYPE_ARRAY);
-	__uint(max_entries, __NR_BPF_RESULT_ARRAY_IDX);
-	__type(key, __u32);
-	__type(value, struct bpf_sock);
-} sock_result_map SEC(".maps");
-
-struct {
-	__uint(type, BPF_MAP_TYPE_ARRAY);
-	__uint(max_entries, __NR_BPF_RESULT_ARRAY_IDX);
-	__type(key, __u32);
-	__type(value, struct bpf_tcp_sock);
-} tcp_sock_result_map SEC(".maps");
-
 struct {
 	__uint(type, BPF_MAP_TYPE_ARRAY);
 	__uint(max_entries, __NR_BPF_LINUM_ARRAY_IDX);
@@ -74,6 +41,17 @@ struct {
 	__type(value, struct bpf_spinlock_cnt);
 } sk_pkt_out_cnt10 SEC(".maps");
 
+struct bpf_tcp_sock listen_tp = {};
+struct sockaddr_in6 srv_sa6 = {};
+struct bpf_tcp_sock cli_tp = {};
+struct bpf_tcp_sock srv_tp = {};
+struct bpf_sock listen_sk = {};
+struct bpf_sock srv_sk = {};
+struct bpf_sock cli_sk = {};
+__u64 parent_cg_id = 0;
+__u64 child_cg_id = 0;
+__u64 lsndtime = 0;
+
 static bool is_loopback6(__u32 *a6)
 {
 	return !a6[0] && !a6[1] && !a6[2] && a6[3] == bpf_htonl(1);
@@ -130,62 +108,86 @@ static void tpcpy(struct bpf_tcp_sock *dst,
 	dst->bytes_acked = src->bytes_acked;
 }
 
-#define RETURN {						\
+/* Always return CG_OK so that no pkt will be filtered out */
+#define CG_OK 1
+
+#define RET_LOG() ({						\
 	linum = __LINE__;					\
-	bpf_map_update_elem(&linum_map, &linum_idx, &linum, 0);	\
-	return 1;						\
-}
+	bpf_map_update_elem(&linum_map, &linum_idx, &linum, BPF_NOEXIST);	\
+	return CG_OK;						\
+})
 
 SEC("cgroup_skb/egress")
 int egress_read_sock_fields(struct __sk_buff *skb)
 {
 	struct bpf_spinlock_cnt cli_cnt_init = { .lock = 0, .cnt = 0xeB9F };
-	__u32 srv_idx = ADDR_SRV_IDX, cli_idx = ADDR_CLI_IDX, result_idx;
 	struct bpf_spinlock_cnt *pkt_out_cnt, *pkt_out_cnt10;
-	struct sockaddr_in6 *srv_sa6, *cli_sa6;
 	struct bpf_tcp_sock *tp, *tp_ret;
 	struct bpf_sock *sk, *sk_ret;
 	__u32 linum, linum_idx;
+	struct tcp_sock *ktp;
 
 	linum_idx = EGRESS_LINUM_IDX;
 
 	sk = skb->sk;
-	if (!sk || sk->state == 10)
-		RETURN;
+	if (!sk)
+		RET_LOG();
 
+	/* Not the testing egress traffic or
+	 * TCP_LISTEN (10) socket will be copied at the ingress side.
+	 */
+	if (sk->family != AF_INET6 || !is_loopback6(sk->src_ip6) ||
+	    sk->state == 10)
+		return CG_OK;
+
+	if (sk->src_port == bpf_ntohs(srv_sa6.sin6_port)) {
+		/* Server socket */
+		sk_ret = &srv_sk;
+		tp_ret = &srv_tp;
+	} else if (sk->dst_port == srv_sa6.sin6_port) {
+		/* Client socket */
+		sk_ret = &cli_sk;
+		tp_ret = &cli_tp;
+	} else {
+		/* Not the testing egress traffic */
+		return CG_OK;
+	}
+
+	/* It must be a fullsock for cgroup_skb/egress prog */
 	sk = bpf_sk_fullsock(sk);
-	if (!sk || sk->family != AF_INET6 || sk->protocol != IPPROTO_TCP ||
-	    !is_loopback6(sk->src_ip6))
-		RETURN;
+	if (!sk)
+		RET_LOG();
+
+	/* Not the testing egress traffic */
+	if (sk->protocol != IPPROTO_TCP)
+		return CG_OK;
 
 	tp = bpf_tcp_sock(sk);
 	if (!tp)
-		RETURN;
+		RET_LOG();
 
-	srv_sa6 = bpf_map_lookup_elem(&addr_map, &srv_idx);
-	cli_sa6 = bpf_map_lookup_elem(&addr_map, &cli_idx);
-	if (!srv_sa6 || !cli_sa6)
-		RETURN;
+	skcpy(sk_ret, sk);
+	tpcpy(tp_ret, tp);
 
-	if (sk->src_port == bpf_ntohs(srv_sa6->sin6_port))
-		result_idx = EGRESS_SRV_IDX;
-	else if (sk->src_port == bpf_ntohs(cli_sa6->sin6_port))
-		result_idx = EGRESS_CLI_IDX;
-	else
-		RETURN;
+	if (sk_ret == &srv_sk) {
+		ktp = bpf_skc_to_tcp_sock(sk);
 
-	sk_ret = bpf_map_lookup_elem(&sock_result_map, &result_idx);
-	tp_ret = bpf_map_lookup_elem(&tcp_sock_result_map, &result_idx);
-	if (!sk_ret || !tp_ret)
-		RETURN;
+		if (!ktp)
+			RET_LOG();
 
-	skcpy(sk_ret, sk);
-	tpcpy(tp_ret, tp);
+		lsndtime = ktp->lsndtime;
+
+		child_cg_id = bpf_sk_cgroup_id(ktp);
+		if (!child_cg_id)
+			RET_LOG();
+
+		parent_cg_id = bpf_sk_ancestor_cgroup_id(ktp, 2);
+		if (!parent_cg_id)
+			RET_LOG();
 
-	if (result_idx == EGRESS_SRV_IDX) {
 		/* The userspace has created it for srv sk */
-		pkt_out_cnt = bpf_sk_storage_get(&sk_pkt_out_cnt, sk, 0, 0);
-		pkt_out_cnt10 = bpf_sk_storage_get(&sk_pkt_out_cnt10, sk,
+		pkt_out_cnt = bpf_sk_storage_get(&sk_pkt_out_cnt, ktp, 0, 0);
+		pkt_out_cnt10 = bpf_sk_storage_get(&sk_pkt_out_cnt10, ktp,
 						   0, 0);
 	} else {
 		pkt_out_cnt = bpf_sk_storage_get(&sk_pkt_out_cnt, sk,
@@ -197,7 +199,7 @@ int egress_read_sock_fields(struct __sk_buff *skb)
 	}
 
 	if (!pkt_out_cnt || !pkt_out_cnt10)
-		RETURN;
+		RET_LOG();
 
 	/* Even both cnt and cnt10 have lock defined in their BTF,
 	 * intentionally one cnt takes lock while one does not
@@ -208,48 +210,44 @@ int egress_read_sock_fields(struct __sk_buff *skb)
 	pkt_out_cnt10->cnt += 10;
 	bpf_spin_unlock(&pkt_out_cnt10->lock);
 
-	RETURN;
+	return CG_OK;
 }
 
 SEC("cgroup_skb/ingress")
 int ingress_read_sock_fields(struct __sk_buff *skb)
 {
-	__u32 srv_idx = ADDR_SRV_IDX, result_idx = INGRESS_LISTEN_IDX;
-	struct bpf_tcp_sock *tp, *tp_ret;
-	struct bpf_sock *sk, *sk_ret;
-	struct sockaddr_in6 *srv_sa6;
+	struct bpf_tcp_sock *tp;
 	__u32 linum, linum_idx;
+	struct bpf_sock *sk;
 
 	linum_idx = INGRESS_LINUM_IDX;
 
 	sk = skb->sk;
-	if (!sk || sk->family != AF_INET6 || !is_loopback6(sk->src_ip6))
-		RETURN;
+	if (!sk)
+		RET_LOG();
 
-	srv_sa6 = bpf_map_lookup_elem(&addr_map, &srv_idx);
-	if (!srv_sa6 || sk->src_port != bpf_ntohs(srv_sa6->sin6_port))
-		RETURN;
+	/* Not the testing ingress traffic to the server */
+	if (sk->family != AF_INET6 || !is_loopback6(sk->src_ip6) ||
+	    sk->src_port != bpf_ntohs(srv_sa6.sin6_port))
+		return CG_OK;
 
-	if (sk->state != 10 && sk->state != 12)
-		RETURN;
+	/* Only interested in TCP_LISTEN */
+	if (sk->state != 10)
+		return CG_OK;
 
-	sk = bpf_get_listener_sock(sk);
+	/* It must be a fullsock for cgroup_skb/ingress prog */
+	sk = bpf_sk_fullsock(sk);
 	if (!sk)
-		RETURN;
+		RET_LOG();
 
 	tp = bpf_tcp_sock(sk);
 	if (!tp)
-		RETURN;
-
-	sk_ret = bpf_map_lookup_elem(&sock_result_map, &result_idx);
-	tp_ret = bpf_map_lookup_elem(&tcp_sock_result_map, &result_idx);
-	if (!sk_ret || !tp_ret)
-		RETURN;
+		RET_LOG();
 
-	skcpy(sk_ret, sk);
-	tpcpy(tp_ret, tp);
+	skcpy(&listen_sk, sk);
+	tpcpy(&listen_tp, tp);
 
-	RETURN;
+	return CG_OK;
 }
 
 char _license[] SEC("license") = "GPL";

tools/testing/selftests/bpf/verifier/ref_tracking.c

@@ -854,3 +854,50 @@
 	.errstr = "Unreleased reference",
 	.result = REJECT,
 },
+{
+	"reference tracking: bpf_sk_release(btf_tcp_sock)",
+	.insns = {
+	BPF_SK_LOOKUP(sk_lookup_tcp),
+	BPF_JMP_IMM(BPF_JNE, BPF_REG_0, 0, 1),
+	BPF_EXIT_INSN(),
+	BPF_MOV64_REG(BPF_REG_6, BPF_REG_0),
+	BPF_MOV64_REG(BPF_REG_1, BPF_REG_0),
+	BPF_EMIT_CALL(BPF_FUNC_skc_to_tcp_sock),
+	BPF_JMP_IMM(BPF_JNE, BPF_REG_0, 0, 3),
+	BPF_MOV64_REG(BPF_REG_1, BPF_REG_6),
+	BPF_EMIT_CALL(BPF_FUNC_sk_release),
+	BPF_EXIT_INSN(),
+	BPF_MOV64_REG(BPF_REG_1, BPF_REG_0),
+	BPF_EMIT_CALL(BPF_FUNC_sk_release),
+	BPF_EXIT_INSN(),
+	},
+	.prog_type = BPF_PROG_TYPE_SCHED_CLS,
+	.result = ACCEPT,
+	.result_unpriv = REJECT,
+	.errstr_unpriv = "unknown func",
+},
+{
+	"reference tracking: use ptr from bpf_skc_to_tcp_sock() after release",
+	.insns = {
+	BPF_SK_LOOKUP(sk_lookup_tcp),
+	BPF_JMP_IMM(BPF_JNE, BPF_REG_0, 0, 1),
+	BPF_EXIT_INSN(),
+	BPF_MOV64_REG(BPF_REG_6, BPF_REG_0),
+	BPF_MOV64_REG(BPF_REG_1, BPF_REG_0),
+	BPF_EMIT_CALL(BPF_FUNC_skc_to_tcp_sock),
+	BPF_JMP_IMM(BPF_JNE, BPF_REG_0, 0, 3),
+	BPF_MOV64_REG(BPF_REG_1, BPF_REG_6),
+	BPF_EMIT_CALL(BPF_FUNC_sk_release),
+	BPF_EXIT_INSN(),
+	BPF_MOV64_REG(BPF_REG_7, BPF_REG_0),
+	BPF_MOV64_REG(BPF_REG_1, BPF_REG_6),
+	BPF_EMIT_CALL(BPF_FUNC_sk_release),
+	BPF_LDX_MEM(BPF_B, BPF_REG_0, BPF_REG_7, 0),
+	BPF_EXIT_INSN(),
+	},
+	.prog_type = BPF_PROG_TYPE_SCHED_CLS,
+	.result = REJECT,
+	.errstr = "invalid mem access",
+	.result_unpriv = REJECT,
+	.errstr_unpriv = "unknown func",
+},