Skip to contentSkip to conversation

implement bpf iterator for tcp and udp sockets

PATCH Applied Yonghong Song 2 participants 17 messages Jun 23, 2020 – Jun 25, 2020

Subsystems (12)
lore
Share & tools
lore
#0 lore raw

v5 cover implement bpf iterator for tcp and udp sockets

bpf iterator implments traversal of kernel data structures and these data structures are passed to a bpf program for processing. This gives great flexibility for users to examine kernel data structure without using e.g. /proc/net which has limited and fixed format.

Commit 138d0be35b14 ("net: bpf: Add netlink and ipv6_route bpf_iter targets") implemented bpf iterators for netlink and ipv6_route.
This patch set intends to implement bpf iterators for tcp and udp.

Currently, /proc/net/tcp is used to print tcp4 stats and /proc/net/tcp6 is used to print tcp6 stats. /proc/net/udp[6] have similar usage model. In contrast, only one tcp iterator is implemented and it is bpf program resposibility to filter based on socket family. The same is for udp. This will avoid another unnecessary traversal pass if users want to check both tcp4 and tcp6.

Several helpers are also implemented in this patch
bpf_skc_to_{tcp, tcp6, tcp_timewait, tcp_request, udp6}_sock
The argument for these helpers is not a fixed btf_id. For example,
bpf_skc_to_tcp(struct sock_common *), or bpf_skc_to_tcp(struct sock *), or
bpf_skc_to_tcp(struct inet_sock *), ...
are all valid. At runtime, the helper will check whether pointer cast is legal or not. Please see Patch #5 for details.

Since btf_id's for both arguments and return value are known at build time, the btf_id's are pre-computed once vmlinux btf becomes valid. Jiri's "adding d_path helper" patch set
https://lore.kernel.org/bpf/[email protected]/T/
provides a way to pre-compute btf id during vmlinux build time. This can be applied here as well. A followup patch can convert to build time btf id computation after Jiri's patch landed.

Changelogs:
v4 -> v5:
- fix bpf_skc_to_udp6_sock helper as besides sk_protocol, sk_family, sk_type == SOCK_DGRAM is also needed to differentiate from SOCK_RAW (Eric)
v3 -> v4:
- fix bpf_skc_to_{tcp_timewait, tcp_request}_sock helper implementation as just checking sk->sk_state is not enough (Martin)
- fix a few kernel test robot reported failures
- move bpf_tracing_net.h from libbpf to selftests (Andrii)
- remove __weak attribute from selftests CONFIG_HZ variables (Andrii)
v2 -> v3:
- change sock_cast*/SOCK_CAST* names to btf_sock* names for generality (Martin)
- change gpl_license to false (Martin)
- fix helper to cast to tcp timewait/request socket. (Martin)
v1 -> v2:
- guard init_sock_cast_types() defination properly with CONFIG_NET (Martin)
- reuse the btf_ids, computed for new helper argument, for return values (Martin)
- using BTF_TYPE_EMIT to express intent of btf type generation (Andrii)
- abstract out common net macros into bpf_tracing_net.h (Andrii)

Yonghong Song (15):
net: bpf: add bpf_seq_afinfo in tcp_iter_state
net: bpf: implement bpf iterator for tcp
bpf: support 'X' in bpf_seq_printf() helper
bpf: allow tracing programs to use bpf_jiffies64() helper
bpf: add bpf_skc_to_tcp6_sock() helper
bpf: add bpf_skc_to_{tcp,tcp_timewait,tcp_request}_sock() helpers
net: bpf: add bpf_seq_afinfo in udp_iter_state
net: bpf: implement bpf iterator for udp
bpf: add bpf_skc_to_udp6_sock() helper
selftests/bpf: move newer bpf_iter_* type redefining to a new header
file
selftests/bpf: refactor some net macros to bpf_tracing_net.h selftests/bpf: add more common macros to bpf_tracing_net.h selftests/bpf: implement sample tcp/tcp6 bpf_iter programs selftests/bpf: implement sample udp/udp6 bpf_iter programs selftests/bpf: add tcp/udp iterator programs to selftests

 include/linux/bpf.h                           |  16 ++
 include/net/tcp.h                             |   1 +
 include/net/udp.h                             |   1 +
 include/uapi/linux/bpf.h                      |  37 ++-
 kernel/bpf/btf.c                              |   1 +
 kernel/bpf/verifier.c                         |  43 ++-
 kernel/trace/bpf_trace.c                      |  15 +-
 net/core/filter.c                             | 166 ++++++++++++
 net/ipv4/tcp_ipv4.c                           | 153 ++++++++++-
 net/ipv4/udp.c                                | 144 +++++++++-
 scripts/bpf_helpers_doc.py                    |  10 +
 tools/include/uapi/linux/bpf.h                |  37 ++-
 .../selftests/bpf/prog_tests/bpf_iter.c       |  68 +++++
 tools/testing/selftests/bpf/progs/bpf_iter.h  |  80 ++++++
 .../selftests/bpf/progs/bpf_iter_bpf_map.c    |  18 +-
 .../selftests/bpf/progs/bpf_iter_ipv6_route.c |  25 +-
 .../selftests/bpf/progs/bpf_iter_netlink.c    |  22 +-
 .../selftests/bpf/progs/bpf_iter_task.c       |  18 +-
 .../selftests/bpf/progs/bpf_iter_task_file.c  |  20 +-
 .../selftests/bpf/progs/bpf_iter_tcp4.c       | 234 ++++++++++++++++
 .../selftests/bpf/progs/bpf_iter_tcp6.c       | 250 ++++++++++++++++++
 .../selftests/bpf/progs/bpf_iter_test_kern3.c |  17 +-
 .../selftests/bpf/progs/bpf_iter_test_kern4.c |  17 +-
 .../bpf/progs/bpf_iter_test_kern_common.h     |  18 +-
 .../selftests/bpf/progs/bpf_iter_udp4.c       |  71 +++++
 .../selftests/bpf/progs/bpf_iter_udp6.c       |  79 ++++++
 .../selftests/bpf/progs/bpf_tracing_net.h     |  51 ++++
 27 files changed, 1443 insertions(+), 169 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_iter.h
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_iter_tcp4.c
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_iter_tcp6.c
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_iter_udp4.c
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_iter_udp6.c
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_tracing_net.h
Signature · 1 line
2.24.1
#1 lore raw

PATCH v5 1/15 net: bpf: add bpf_seq_afinfo in tcp_iter_state

A new field bpf_seq_afinfo is added to tcp_iter_state to provide bpf tcp iterator afinfo. There are two reasons on why we did this.

First, the current way to get afinfo from PDE_DATA does not work for bpf iterator as its seq_file inode does not conform to /proc/net/{tcp,tcp6} inode structures. More specifically, anonymous bpf iterator will use an anonymous inode which is shared in the system and we cannot change inode private data structure at all.

Second, bpf iterator for tcp/tcp6 wants to traverse all tcp and tcp6 sockets in one pass and bpf program can control whether they want to skip one sk_family or not. Having a different afinfo with family AF_UNSPEC make it easier to understand in the code.

This patch does not change /proc/net/{tcp,tcp6} behavior as the bpf_seq_afinfo will be NULL for these two proc files.

---
 include/net/tcp.h   |  1 +
 net/ipv4/tcp_ipv4.c | 30 ++++++++++++++++++++++++------
 2 files changed, 25 insertions(+), 6 deletions(-)
include/net/tcp.h +1 -0
index 4de9485f73d9..eab1c7d0facb 100644
@@ -1935,6 +1935,7 @@ struct tcp_iter_state {
 	struct seq_net_private	p;
 	enum tcp_seq_states	state;
 	struct sock		*syn_wait_sk;
+	struct tcp_seq_afinfo	*bpf_seq_afinfo;
 	int			bucket, offset, sbucket, num;
 	loff_t			last_pos;
 };
net/ipv4/tcp_ipv4.c +24 -6
index ad6435ba6d72..9cb65ee4ec63 100644
@@ -2211,13 +2211,18 @@ EXPORT_SYMBOL(tcp_v4_destroy_sock);
  */
 static void *listening_get_next(struct seq_file *seq, void *cur)
 {
-	struct tcp_seq_afinfo *afinfo = PDE_DATA(file_inode(seq->file));
+	struct tcp_seq_afinfo *afinfo;
 	struct tcp_iter_state *st = seq->private;
 	struct net *net = seq_file_net(seq);
 	struct inet_listen_hashbucket *ilb;
 	struct hlist_nulls_node *node;
 	struct sock *sk = cur;
 
+	if (st->bpf_seq_afinfo)
+		afinfo = st->bpf_seq_afinfo;
+	else
+		afinfo = PDE_DATA(file_inode(seq->file));
+
 	if (!sk) {
 get_head:
 		ilb = &tcp_hashinfo.listening_hash[st->bucket];
@@ -2235,7 +2240,8 @@ static void *listening_get_next(struct seq_file *seq, void *cur)
 	sk_nulls_for_each_from(sk, node) {
 		if (!net_eq(sock_net(sk), net))
 			continue;
-		if (sk->sk_family == afinfo->family)
+		if (afinfo->family == AF_UNSPEC ||
+		    sk->sk_family == afinfo->family)
 			return sk;
 	}
 	spin_unlock(&ilb->lock);
@@ -2272,11 +2278,16 @@ static inline bool empty_bucket(const struct tcp_iter_state *st)
  */
 static void *established_get_first(struct seq_file *seq)
 {
-	struct tcp_seq_afinfo *afinfo = PDE_DATA(file_inode(seq->file));
+	struct tcp_seq_afinfo *afinfo;
 	struct tcp_iter_state *st = seq->private;
 	struct net *net = seq_file_net(seq);
 	void *rc = NULL;
 
+	if (st->bpf_seq_afinfo)
+		afinfo = st->bpf_seq_afinfo;
+	else
+		afinfo = PDE_DATA(file_inode(seq->file));
+
 	st->offset = 0;
 	for (; st->bucket <= tcp_hashinfo.ehash_mask; ++st->bucket) {
 		struct sock *sk;
@@ -2289,7 +2300,8 @@ static void *established_get_first(struct seq_file *seq)
 
 		spin_lock_bh(lock);
 		sk_nulls_for_each(sk, node, &tcp_hashinfo.ehash[st->bucket].chain) {
-			if (sk->sk_family != afinfo->family ||
+			if ((afinfo->family != AF_UNSPEC &&
+			     sk->sk_family != afinfo->family) ||
 			    !net_eq(sock_net(sk), net)) {
 				continue;
 			}
@@ -2304,19 +2316,25 @@ static void *established_get_first(struct seq_file *seq)
 
 static void *established_get_next(struct seq_file *seq, void *cur)
 {
-	struct tcp_seq_afinfo *afinfo = PDE_DATA(file_inode(seq->file));
+	struct tcp_seq_afinfo *afinfo;
 	struct sock *sk = cur;
 	struct hlist_nulls_node *node;
 	struct tcp_iter_state *st = seq->private;
 	struct net *net = seq_file_net(seq);
 
+	if (st->bpf_seq_afinfo)
+		afinfo = st->bpf_seq_afinfo;
+	else
+		afinfo = PDE_DATA(file_inode(seq->file));
+
 	++st->num;
 	++st->offset;
 
 	sk = sk_nulls_next(sk);
 
 	sk_nulls_for_each_from(sk, node) {
-		if (sk->sk_family == afinfo->family &&
+		if ((afinfo->family == AF_UNSPEC ||
+		     sk->sk_family == afinfo->family) &&
 		    net_eq(sock_net(sk), net))
 			return sk;
 	}
Signature · 1 line
2.24.1
#2 lore raw

PATCH v5 2/15 net: bpf: implement bpf iterator for tcp

In reply to Yonghong Song

The bpf iterator for tcp is implemented. Both tcp4 and tcp6 sockets will be traversed. It is up to bpf program to filter for tcp4 or tcp6 only, or both families of sockets.

---
 net/ipv4/tcp_ipv4.c | 123 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 123 insertions(+)
net/ipv4/tcp_ipv4.c +123 -0
index 9cb65ee4ec63..ea0df9fd7618 100644
@@ -2613,6 +2613,74 @@ static int tcp4_seq_show(struct seq_file *seq, void *v)
 	return 0;
 }
 
+#ifdef CONFIG_BPF_SYSCALL
+struct bpf_iter__tcp {
+	__bpf_md_ptr(struct bpf_iter_meta *, meta);
+	__bpf_md_ptr(struct sock_common *, sk_common);
+	uid_t uid __aligned(8);
+};
+
+static int tcp_prog_seq_show(struct bpf_prog *prog, struct bpf_iter_meta *meta,
+			     struct sock_common *sk_common, uid_t uid)
+{
+	struct bpf_iter__tcp ctx;
+
+	meta->seq_num--;  /* skip SEQ_START_TOKEN */
+	ctx.meta = meta;
+	ctx.sk_common = sk_common;
+	ctx.uid = uid;
+	return bpf_iter_run_prog(prog, &ctx);
+}
+
+static int bpf_iter_tcp_seq_show(struct seq_file *seq, void *v)
+{
+	struct bpf_iter_meta meta;
+	struct bpf_prog *prog;
+	struct sock *sk = v;
+	uid_t uid;
+
+	if (v == SEQ_START_TOKEN)
+		return 0;
+
+	if (sk->sk_state == TCP_TIME_WAIT) {
+		uid = 0;
+	} else if (sk->sk_state == TCP_NEW_SYN_RECV) {
+		const struct request_sock *req = v;
+
+		uid = from_kuid_munged(seq_user_ns(seq),
+				       sock_i_uid(req->rsk_listener));
+	} else {
+		uid = from_kuid_munged(seq_user_ns(seq), sock_i_uid(sk));
+	}
+
+	meta.seq = seq;
+	prog = bpf_iter_get_info(&meta, false);
+	return tcp_prog_seq_show(prog, &meta, v, uid);
+}
+
+static void bpf_iter_tcp_seq_stop(struct seq_file *seq, void *v)
+{
+	struct bpf_iter_meta meta;
+	struct bpf_prog *prog;
+
+	if (!v) {
+		meta.seq = seq;
+		prog = bpf_iter_get_info(&meta, true);
+		if (prog)
+			(void)tcp_prog_seq_show(prog, &meta, v, 0);
+	}
+
+	tcp_seq_stop(seq, v);
+}
+
+static const struct seq_operations bpf_iter_tcp_seq_ops = {
+	.show		= bpf_iter_tcp_seq_show,
+	.start		= tcp_seq_start,
+	.next		= tcp_seq_next,
+	.stop		= bpf_iter_tcp_seq_stop,
+};
+#endif
+
 static const struct seq_operations tcp4_seq_ops = {
 	.show		= tcp4_seq_show,
 	.start		= tcp_seq_start,
@@ -2844,8 +2912,63 @@ static struct pernet_operations __net_initdata tcp_sk_ops = {
        .exit_batch = tcp_sk_exit_batch,
 };
 
+#if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_PROC_FS)
+DEFINE_BPF_ITER_FUNC(tcp, struct bpf_iter_meta *meta,
+		     struct sock_common *sk_common, uid_t uid)
+
+static int bpf_iter_init_tcp(void *priv_data)
+{
+	struct tcp_iter_state *st = priv_data;
+	struct tcp_seq_afinfo *afinfo;
+	int ret;
+
+	afinfo = kmalloc(sizeof(*afinfo), GFP_USER | __GFP_NOWARN);
+	if (!afinfo)
+		return -ENOMEM;
+
+	afinfo->family = AF_UNSPEC;
+	st->bpf_seq_afinfo = afinfo;
+	ret = bpf_iter_init_seq_net(priv_data);
+	if (ret)
+		kfree(afinfo);
+	return ret;
+}
+
+static void bpf_iter_fini_tcp(void *priv_data)
+{
+	struct tcp_iter_state *st = priv_data;
+
+	kfree(st->bpf_seq_afinfo);
+	bpf_iter_fini_seq_net(priv_data);
+}
+
+static const struct bpf_iter_reg tcp_reg_info = {
+	.target			= "tcp",
+	.seq_ops		= &bpf_iter_tcp_seq_ops,
+	.init_seq_private	= bpf_iter_init_tcp,
+	.fini_seq_private	= bpf_iter_fini_tcp,
+	.seq_priv_size		= sizeof(struct tcp_iter_state),
+	.ctx_arg_info_size	= 1,
+	.ctx_arg_info		= {
+		{ offsetof(struct bpf_iter__tcp, sk_common),
+		  PTR_TO_BTF_ID_OR_NULL },
+	},
+};
+
+static void __init bpf_iter_register(void)
+{
+	if (bpf_iter_reg_target(&tcp_reg_info))
+		pr_warn("Warning: could not register bpf iterator tcp\n");
+}
+
+#endif
+
 void __init tcp_v4_init(void)
 {
 	if (register_pernet_subsys(&tcp_sk_ops))
 		panic("Failed to create the TCP control socket.\n");
+
+#if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_PROC_FS)
+	bpf_iter_register();
+#endif
 }
Signature · 1 line
2.24.1
#3 lore raw

PATCH v5 3/15 bpf: support 'X' in bpf_seq_printf() helper

In reply to Yonghong Song

'X' tells kernel to print hex with upper case letters. /proc/net/tcp{4,6} seq_file show() used this, and supports it in bpf_seq_printf() helper too.

---
 kernel/trace/bpf_trace.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/trace/bpf_trace.c +2 -1
index e729c9e587a0..dbee30e2ad91 100644
@@ -681,7 +681,8 @@ BPF_CALL_5(bpf_seq_printf, struct seq_file *, m, char *, fmt, u32, fmt_size,
 		}
 
 		if (fmt[i] != 'i' && fmt[i] != 'd' &&
-		    fmt[i] != 'u' && fmt[i] != 'x') {
+		    fmt[i] != 'u' && fmt[i] != 'x' &&
+		    fmt[i] != 'X') {
 			err = -EINVAL;
 			goto out;
 		}
Signature · 1 line
2.24.1
#4 lore raw

PATCH v5 4/15 bpf: allow tracing programs to use bpf_jiffies64() helper

In reply to Yonghong Song

/proc/net/tcp{4,6} uses jiffies for various computations. Let us add bpf_jiffies64() helper to tracing program so bpf_iter and other programs can use it.

---
 kernel/trace/bpf_trace.c | 2 ++
 1 file changed, 2 insertions(+)
kernel/trace/bpf_trace.c +2 -0
index dbee30e2ad91..afaec7e082d9 100644
@@ -1135,6 +1135,8 @@ bpf_tracing_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_ringbuf_discard_proto;
 	case BPF_FUNC_ringbuf_query:
 		return &bpf_ringbuf_query_proto;
+	case BPF_FUNC_jiffies64:
+		return &bpf_jiffies64_proto;
 	default:
 		return NULL;
 	}
Signature · 1 line
2.24.1
#5 lore raw

PATCH v5 5/15 bpf: add bpf_skc_to_tcp6_sock() helper

In reply to Yonghong Song

The helper is used in tracing programs to cast a socket pointer to a tcp6_sock pointer.
The return value could be NULL if the casting is illegal.

A new helper return type RET_PTR_TO_BTF_ID_OR_NULL is added so the verifier is able to deduce proper return types for the helper.

Different from the previous BTF_ID based helpers,
the bpf_skc_to_tcp6_sock() argument can be several possible btf_ids. More specifically, all possible socket data structures with sock_common appearing in the first in the memory layout. This patch only added socket types related to tcp and udp.

All possible argument btf_id and return value btf_id for helper bpf_skc_to_tcp6_sock() are pre-calculcated and cached. In the future, it is even possible to precompute these btf_id's at kernel build time.

---
 include/linux/bpf.h            | 12 +++++
 include/uapi/linux/bpf.h       |  9 +++-
 kernel/bpf/btf.c               |  1 +
 kernel/bpf/verifier.c          | 43 +++++++++++++-----
 kernel/trace/bpf_trace.c       |  2 +
 net/core/filter.c              | 82 ++++++++++++++++++++++++++++++++++
 scripts/bpf_helpers_doc.py     |  2 +
 tools/include/uapi/linux/bpf.h |  9 +++-
 8 files changed, 148 insertions(+), 12 deletions(-)
include/linux/bpf.h +12 -0
index 1e1501ee53ce..c0e38ad07848 100644
@@ -265,6 +265,7 @@ enum bpf_return_type {
 	RET_PTR_TO_TCP_SOCK_OR_NULL,	/* returns a pointer to a tcp_sock or NULL */
 	RET_PTR_TO_SOCK_COMMON_OR_NULL,	/* returns a pointer to a sock_common or NULL */
 	RET_PTR_TO_ALLOC_MEM_OR_NULL,	/* returns a pointer to dynamically allocated memory or NULL */
+	RET_PTR_TO_BTF_ID_OR_NULL,	/* returns a pointer to a btf_id or NULL */
 };
 
 /* eBPF function prototype used by verifier to allow BPF_CALLs from eBPF programs
@@ -287,6 +288,12 @@ struct bpf_func_proto {
 		enum bpf_arg_type arg_type[5];
 	};
 	int *btf_id; /* BTF ids of arguments */
+	bool (*check_btf_id)(u32 btf_id, u32 arg); /* if the argument btf_id is
+						    * valid. Often used if more
+						    * than one btf id is permitted
+						    * for this argument.
+						    */
+	int *ret_btf_id; /* return value btf_id */
 };
 
 /* bpf_context is intentionally undefined structure. Pointer to bpf_context is
@@ -1524,6 +1531,7 @@ static inline bool bpf_map_is_dev_bound(struct bpf_map *map)
 
 struct bpf_map *bpf_map_offload_map_alloc(union bpf_attr *attr);
 void bpf_map_offload_map_free(struct bpf_map *map);
+void init_btf_sock_ids(struct btf *btf);
 #else
 static inline int bpf_prog_offload_init(struct bpf_prog *prog,
 					union bpf_attr *attr)
@@ -1549,6 +1557,9 @@ static inline struct bpf_map *bpf_map_offload_map_alloc(union bpf_attr *attr)
 static inline void bpf_map_offload_map_free(struct bpf_map *map)
 {
 }
+static inline void init_btf_sock_ids(struct btf *btf)
+{
+}
 #endif /* CONFIG_NET && CONFIG_BPF_SYSCALL */
 
 #if defined(CONFIG_BPF_STREAM_PARSER)
@@ -1638,6 +1649,7 @@ extern const struct bpf_func_proto bpf_ringbuf_reserve_proto;
 extern const struct bpf_func_proto bpf_ringbuf_submit_proto;
 extern const struct bpf_func_proto bpf_ringbuf_discard_proto;
 extern const struct bpf_func_proto bpf_ringbuf_query_proto;
+extern const struct bpf_func_proto bpf_skc_to_tcp6_sock_proto;
 
 const struct bpf_func_proto *bpf_tracing_func_proto(
 	enum bpf_func_id func_id, const struct bpf_prog *prog);
include/uapi/linux/bpf.h +8 -1
index 19684813faae..394fcba27b6a 100644
@@ -3252,6 +3252,12 @@ union bpf_attr {
  * 		case of **BPF_CSUM_LEVEL_QUERY**, the current skb->csum_level
  * 		is returned or the error code -EACCES in case the skb is not
  * 		subject to CHECKSUM_UNNECESSARY.
+ *
+ * struct tcp6_sock *bpf_skc_to_tcp6_sock(void *sk)
+ *	Description
+ *		Dynamically cast a *sk* pointer to a *tcp6_sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or NULL otherwise.
  */
 #define __BPF_FUNC_MAPPER(FN)		\
 	FN(unspec),			\
@@ -3389,7 +3395,8 @@ union bpf_attr {
 	FN(ringbuf_submit),		\
 	FN(ringbuf_discard),		\
 	FN(ringbuf_query),		\
-	FN(csum_level),
+	FN(csum_level),			\
+	FN(skc_to_tcp6_sock),
 
 /* integer value in 'imm' field of BPF_CALL instruction selects which helper
  * function eBPF program intends to call
kernel/bpf/btf.c +1 -0
index e377d1981730..4c3007f428b1 100644
@@ -3674,6 +3674,7 @@ struct btf *btf_parse_vmlinux(void)
 		goto errout;
 
 	bpf_struct_ops_init(btf, log);
+	init_btf_sock_ids(btf);
 
 	btf_verifier_env_free(env);
 	refcount_set(&btf->refcnt, 1);
kernel/bpf/verifier.c +33 -10

70 lines not on this page · raw message

kernel/trace/bpf_trace.c +2 -0
index afaec7e082d9..478c10d1ec33 100644
@@ -1515,6 +1515,8 @@ tracing_prog_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_skb_output_proto;
 	case BPF_FUNC_xdp_output:
 		return &bpf_xdp_output_proto;
+	case BPF_FUNC_skc_to_tcp6_sock:
+		return &bpf_skc_to_tcp6_sock_proto;
 #endif
 	case BPF_FUNC_seq_printf:
 		return prog->expected_attach_type == BPF_TRACE_ITER ?
net/core/filter.c +82 -0

91 lines not on this page · raw message

scripts/bpf_helpers_doc.py +2 -0
index 91fa668fa860..6c2f64118651 100755
@@ -421,6 +421,7 @@ class PrinterHelpers(Printer):
             'struct sockaddr',
             'struct tcphdr',
             'struct seq_file',
+            'struct tcp6_sock',
 
             'struct __sk_buff',
             'struct sk_msg_md',
@@ -458,6 +459,7 @@ class PrinterHelpers(Printer):
             'struct sockaddr',
             'struct tcphdr',
             'struct seq_file',
+            'struct tcp6_sock',
     }
     mapped_types = {
             'u8': '__u8',
tools/include/uapi/linux/bpf.h +8 -1
index 19684813faae..394fcba27b6a 100644
@@ -3252,6 +3252,12 @@ union bpf_attr {
  * 		case of **BPF_CSUM_LEVEL_QUERY**, the current skb->csum_level
  * 		is returned or the error code -EACCES in case the skb is not
  * 		subject to CHECKSUM_UNNECESSARY.
+ *
+ * struct tcp6_sock *bpf_skc_to_tcp6_sock(void *sk)
+ *	Description
+ *		Dynamically cast a *sk* pointer to a *tcp6_sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or NULL otherwise.
  */
 #define __BPF_FUNC_MAPPER(FN)		\
 	FN(unspec),			\
@@ -3389,7 +3395,8 @@ union bpf_attr {
 	FN(ringbuf_submit),		\
 	FN(ringbuf_discard),		\
 	FN(ringbuf_query),		\
-	FN(csum_level),
+	FN(csum_level),			\
+	FN(skc_to_tcp6_sock),
 
 /* integer value in 'imm' field of BPF_CALL instruction selects which helper
  * function eBPF program intends to call
Signature · 1 line
2.24.1
#6 lore raw

PATCH v5 6/15 bpf: add bpf_skc_to_{tcp,tcp_timewait,tcp_request}_sock() helpers

In reply to Yonghong Song

Three more helpers are added to cast a sock_common pointer to an tcp_sock, tcp_timewait_sock or a tcp_request_sock for tracing programs.

---
 include/linux/bpf.h            |  3 ++
 include/uapi/linux/bpf.h       | 23 ++++++++++++-
 kernel/trace/bpf_trace.c       |  6 ++++
 net/core/filter.c              | 62 ++++++++++++++++++++++++++++++++++
 scripts/bpf_helpers_doc.py     |  6 ++++
 tools/include/uapi/linux/bpf.h | 23 ++++++++++++-
 6 files changed, 121 insertions(+), 2 deletions(-)
include/linux/bpf.h +3 -0
index c0e38ad07848..c23998cf6699 100644
@@ -1650,6 +1650,9 @@ extern const struct bpf_func_proto bpf_ringbuf_submit_proto;
 extern const struct bpf_func_proto bpf_ringbuf_discard_proto;
 extern const struct bpf_func_proto bpf_ringbuf_query_proto;
 extern const struct bpf_func_proto bpf_skc_to_tcp6_sock_proto;
+extern const struct bpf_func_proto bpf_skc_to_tcp_sock_proto;
+extern const struct bpf_func_proto bpf_skc_to_tcp_timewait_sock_proto;
+extern const struct bpf_func_proto bpf_skc_to_tcp_request_sock_proto;
 
 const struct bpf_func_proto *bpf_tracing_func_proto(
 	enum bpf_func_id func_id, const struct bpf_prog *prog);
include/uapi/linux/bpf.h +22 -1
index 394fcba27b6a..e256417d94c2 100644
@@ -3258,6 +3258,24 @@ union bpf_attr {
  *		Dynamically cast a *sk* pointer to a *tcp6_sock* pointer.
  *	Return
  *		*sk* if casting is valid, or NULL otherwise.
+ *
+ * struct tcp_sock *bpf_skc_to_tcp_sock(void *sk)
+ *	Description
+ *		Dynamically cast a *sk* pointer to a *tcp_sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or NULL otherwise.
+ *
+ * struct tcp_timewait_sock *bpf_skc_to_tcp_timewait_sock(void *sk)
+ * 	Description
+ *		Dynamically cast a *sk* pointer to a *tcp_timewait_sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or NULL otherwise.
+ *
+ * struct tcp_request_sock *bpf_skc_to_tcp_request_sock(void *sk)
+ * 	Description
+ *		Dynamically cast a *sk* pointer to a *tcp_request_sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or NULL otherwise.
  */
 #define __BPF_FUNC_MAPPER(FN)		\
 	FN(unspec),			\
@@ -3396,7 +3414,10 @@ union bpf_attr {
 	FN(ringbuf_discard),		\
 	FN(ringbuf_query),		\
 	FN(csum_level),			\
-	FN(skc_to_tcp6_sock),
+	FN(skc_to_tcp6_sock),		\
+	FN(skc_to_tcp_sock),		\
+	FN(skc_to_tcp_timewait_sock),	\
+	FN(skc_to_tcp_request_sock),
 
 /* integer value in 'imm' field of BPF_CALL instruction selects which helper
  * function eBPF program intends to call
kernel/trace/bpf_trace.c +6 -0
index 478c10d1ec33..de5fbe66e1ca 100644
@@ -1517,6 +1517,12 @@ tracing_prog_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_xdp_output_proto;
 	case BPF_FUNC_skc_to_tcp6_sock:
 		return &bpf_skc_to_tcp6_sock_proto;
+	case BPF_FUNC_skc_to_tcp_sock:
+		return &bpf_skc_to_tcp_sock_proto;
+	case BPF_FUNC_skc_to_tcp_timewait_sock:
+		return &bpf_skc_to_tcp_timewait_sock_proto;
+	case BPF_FUNC_skc_to_tcp_request_sock:
+		return &bpf_skc_to_tcp_request_sock_proto;
 #endif
 	case BPF_FUNC_seq_printf:
 		return prog->expected_attach_type == BPF_TRACE_ITER ?
net/core/filter.c +62 -0

71 lines not on this page · raw message

scripts/bpf_helpers_doc.py +6 -0
index 6c2f64118651..d886657c6aaa 100755
@@ -422,6 +422,9 @@ class PrinterHelpers(Printer):
             'struct tcphdr',
             'struct seq_file',
             'struct tcp6_sock',
+            'struct tcp_sock',
+            'struct tcp_timewait_sock',
+            'struct tcp_request_sock',
 
             'struct __sk_buff',
             'struct sk_msg_md',
@@ -460,6 +463,9 @@ class PrinterHelpers(Printer):
             'struct tcphdr',
             'struct seq_file',
             'struct tcp6_sock',
+            'struct tcp_sock',
+            'struct tcp_timewait_sock',
+            'struct tcp_request_sock',
     }
     mapped_types = {
             'u8': '__u8',
tools/include/uapi/linux/bpf.h +22 -1
index 394fcba27b6a..e256417d94c2 100644
@@ -3258,6 +3258,24 @@ union bpf_attr {
  *		Dynamically cast a *sk* pointer to a *tcp6_sock* pointer.
  *	Return
  *		*sk* if casting is valid, or NULL otherwise.
+ *
+ * struct tcp_sock *bpf_skc_to_tcp_sock(void *sk)
+ *	Description
+ *		Dynamically cast a *sk* pointer to a *tcp_sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or NULL otherwise.
+ *
+ * struct tcp_timewait_sock *bpf_skc_to_tcp_timewait_sock(void *sk)
+ * 	Description
+ *		Dynamically cast a *sk* pointer to a *tcp_timewait_sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or NULL otherwise.
+ *
+ * struct tcp_request_sock *bpf_skc_to_tcp_request_sock(void *sk)
+ * 	Description
+ *		Dynamically cast a *sk* pointer to a *tcp_request_sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or NULL otherwise.
  */
 #define __BPF_FUNC_MAPPER(FN)		\
 	FN(unspec),			\
@@ -3396,7 +3414,10 @@ union bpf_attr {
 	FN(ringbuf_discard),		\
 	FN(ringbuf_query),		\
 	FN(csum_level),			\
-	FN(skc_to_tcp6_sock),
+	FN(skc_to_tcp6_sock),		\
+	FN(skc_to_tcp_sock),		\
+	FN(skc_to_tcp_timewait_sock),	\
+	FN(skc_to_tcp_request_sock),
 
 /* integer value in 'imm' field of BPF_CALL instruction selects which helper
  * function eBPF program intends to call
Signature · 1 line
2.24.1
#7 lore raw

PATCH v5 7/15 net: bpf: add bpf_seq_afinfo in udp_iter_state

In reply to Yonghong Song

Similar to tcp_iter_state, a new field bpf_seq_afinfo is added to udp_iter_state to provide bpf udp iterator afinfo.

This does not change /proc/net/{udp, udp6} behavior. But it enables bpf iterator to avoid get afinfo from PDE_DATA and iterate through all udp and udp6 sockets in one pass.

---
 include/net/udp.h |  1 +
 net/ipv4/udp.c    | 28 +++++++++++++++++++++++-----
 2 files changed, 24 insertions(+), 5 deletions(-)
include/net/udp.h +1 -0
index a8fa6c0c6ded..67c8b7368845 100644
@@ -440,6 +440,7 @@ struct udp_seq_afinfo {
 struct udp_iter_state {
 	struct seq_net_private  p;
 	int			bucket;
+	struct udp_seq_afinfo	*bpf_seq_afinfo;
 };
 
 void *udp_seq_start(struct seq_file *seq, loff_t *pos);
net/ipv4/udp.c +23 -5
index 1b7ebbcae497..90355301b266 100644
@@ -2826,10 +2826,15 @@ EXPORT_SYMBOL(udp_prot);
 static struct sock *udp_get_first(struct seq_file *seq, int start)
 {
 	struct sock *sk;
-	struct udp_seq_afinfo *afinfo = PDE_DATA(file_inode(seq->file));
+	struct udp_seq_afinfo *afinfo;
 	struct udp_iter_state *state = seq->private;
 	struct net *net = seq_file_net(seq);
 
+	if (state->bpf_seq_afinfo)
+		afinfo = state->bpf_seq_afinfo;
+	else
+		afinfo = PDE_DATA(file_inode(seq->file));
+
 	for (state->bucket = start; state->bucket <= afinfo->udp_table->mask;
 	     ++state->bucket) {
 		struct udp_hslot *hslot = &afinfo->udp_table->hash[state->bucket];
@@ -2841,7 +2846,8 @@ static struct sock *udp_get_first(struct seq_file *seq, int start)
 		sk_for_each(sk, &hslot->head) {
 			if (!net_eq(sock_net(sk), net))
 				continue;
-			if (sk->sk_family == afinfo->family)
+			if (afinfo->family == AF_UNSPEC ||
+			    sk->sk_family == afinfo->family)
 				goto found;
 		}
 		spin_unlock_bh(&hslot->lock);
@@ -2853,13 +2859,20 @@ static struct sock *udp_get_first(struct seq_file *seq, int start)
 
 static struct sock *udp_get_next(struct seq_file *seq, struct sock *sk)
 {
-	struct udp_seq_afinfo *afinfo = PDE_DATA(file_inode(seq->file));
+	struct udp_seq_afinfo *afinfo;
 	struct udp_iter_state *state = seq->private;
 	struct net *net = seq_file_net(seq);
 
+	if (state->bpf_seq_afinfo)
+		afinfo = state->bpf_seq_afinfo;
+	else
+		afinfo = PDE_DATA(file_inode(seq->file));
+
 	do {
 		sk = sk_next(sk);
-	} while (sk && (!net_eq(sock_net(sk), net) || sk->sk_family != afinfo->family));
+	} while (sk && (!net_eq(sock_net(sk), net) ||
+			(afinfo->family != AF_UNSPEC &&
+			 sk->sk_family != afinfo->family)));
 
 	if (!sk) {
 		if (state->bucket <= afinfo->udp_table->mask)
@@ -2904,9 +2917,14 @@ EXPORT_SYMBOL(udp_seq_next);
 
 void udp_seq_stop(struct seq_file *seq, void *v)
 {
-	struct udp_seq_afinfo *afinfo = PDE_DATA(file_inode(seq->file));
+	struct udp_seq_afinfo *afinfo;
 	struct udp_iter_state *state = seq->private;
 
+	if (state->bpf_seq_afinfo)
+		afinfo = state->bpf_seq_afinfo;
+	else
+		afinfo = PDE_DATA(file_inode(seq->file));
+
 	if (state->bucket <= afinfo->udp_table->mask)
 		spin_unlock_bh(&afinfo->udp_table->hash[state->bucket].lock);
 }
Signature · 1 line
2.24.1
#8 lore raw

PATCH v5 8/15 net: bpf: implement bpf iterator for udp

In reply to Yonghong Song

The bpf iterator for udp is implemented. Both udp4 and udp6 sockets will be traversed. It is up to bpf program to filter for udp4 or udp6 only, or both families of sockets.

---
 net/ipv4/udp.c | 116 +++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 116 insertions(+)
net/ipv4/udp.c +116 -0
index 90355301b266..31530129f137 100644
@@ -2968,6 +2968,67 @@ int udp4_seq_show(struct seq_file *seq, void *v)
 	return 0;
 }
 
+#ifdef CONFIG_BPF_SYSCALL
+struct bpf_iter__udp {
+	__bpf_md_ptr(struct bpf_iter_meta *, meta);
+	__bpf_md_ptr(struct udp_sock *, udp_sk);
+	uid_t uid __aligned(8);
+	int bucket __aligned(8);
+};
+
+static int udp_prog_seq_show(struct bpf_prog *prog, struct bpf_iter_meta *meta,
+			     struct udp_sock *udp_sk, uid_t uid, int bucket)
+{
+	struct bpf_iter__udp ctx;
+
+	meta->seq_num--;  /* skip SEQ_START_TOKEN */
+	ctx.meta = meta;
+	ctx.udp_sk = udp_sk;
+	ctx.uid = uid;
+	ctx.bucket = bucket;
+	return bpf_iter_run_prog(prog, &ctx);
+}
+
+static int bpf_iter_udp_seq_show(struct seq_file *seq, void *v)
+{
+	struct udp_iter_state *state = seq->private;
+	struct bpf_iter_meta meta;
+	struct bpf_prog *prog;
+	struct sock *sk = v;
+	uid_t uid;
+
+	if (v == SEQ_START_TOKEN)
+		return 0;
+
+	uid = from_kuid_munged(seq_user_ns(seq), sock_i_uid(sk));
+	meta.seq = seq;
+	prog = bpf_iter_get_info(&meta, false);
+	return udp_prog_seq_show(prog, &meta, v, uid, state->bucket);
+}
+
+static void bpf_iter_udp_seq_stop(struct seq_file *seq, void *v)
+{
+	struct bpf_iter_meta meta;
+	struct bpf_prog *prog;
+
+	if (!v) {
+		meta.seq = seq;
+		prog = bpf_iter_get_info(&meta, true);
+		if (prog)
+			(void)udp_prog_seq_show(prog, &meta, v, 0, 0);
+	}
+
+	udp_seq_stop(seq, v);
+}
+
+static const struct seq_operations bpf_iter_udp_seq_ops = {
+	.start		= udp_seq_start,
+	.next		= udp_seq_next,
+	.stop		= bpf_iter_udp_seq_stop,
+	.show		= bpf_iter_udp_seq_show,
+};
+#endif
+
 const struct seq_operations udp_seq_ops = {
 	.start		= udp_seq_start,
 	.next		= udp_seq_next,
@@ -3085,6 +3146,57 @@ static struct pernet_operations __net_initdata udp_sysctl_ops = {
 	.init	= udp_sysctl_init,
 };
 
+#if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_PROC_FS)
+DEFINE_BPF_ITER_FUNC(udp, struct bpf_iter_meta *meta,
+		     struct udp_sock *udp_sk, uid_t uid, int bucket)
+
+static int bpf_iter_init_udp(void *priv_data)
+{
+	struct udp_iter_state *st = priv_data;
+	struct udp_seq_afinfo *afinfo;
+	int ret;
+
+	afinfo = kmalloc(sizeof(*afinfo), GFP_USER | __GFP_NOWARN);
+	if (!afinfo)
+		return -ENOMEM;
+
+	afinfo->family = AF_UNSPEC;
+	afinfo->udp_table = &udp_table;
+	st->bpf_seq_afinfo = afinfo;
+	ret = bpf_iter_init_seq_net(priv_data);
+	if (ret)
+		kfree(afinfo);
+	return ret;
+}
+
+static void bpf_iter_fini_udp(void *priv_data)
+{
+	struct udp_iter_state *st = priv_data;
+
+	kfree(st->bpf_seq_afinfo);
+	bpf_iter_fini_seq_net(priv_data);
+}
+
+static const struct bpf_iter_reg udp_reg_info = {
+	.target			= "udp",
+	.seq_ops		= &bpf_iter_udp_seq_ops,
+	.init_seq_private	= bpf_iter_init_udp,
+	.fini_seq_private	= bpf_iter_fini_udp,
+	.seq_priv_size		= sizeof(struct udp_iter_state),
+	.ctx_arg_info_size	= 1,
+	.ctx_arg_info		= {
+		{ offsetof(struct bpf_iter__udp, udp_sk),
+		  PTR_TO_BTF_ID_OR_NULL },
+	},
+};
+
+static void __init bpf_iter_register(void)
+{
+	if (bpf_iter_reg_target(&udp_reg_info))
+		pr_warn("Warning: could not register bpf iterator udp\n");
+}
+#endif
+
 void __init udp_init(void)
 {
 	unsigned long limit;
@@ -3110,4 +3222,8 @@ void __init udp_init(void)
 
 	if (register_pernet_subsys(&udp_sysctl_ops))
 		panic("UDP: failed to init sysctl parameters.\n");
+
+#if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_PROC_FS)
+	bpf_iter_register();
+#endif
 }
Signature · 1 line
2.24.1
#9 lore raw

PATCH v5 9/15 bpf: add bpf_skc_to_udp6_sock() helper

In reply to Yonghong Song

The helper is used in tracing programs to cast a socket pointer to a udp6_sock pointer.
The return value could be NULL if the casting is illegal.

---
 include/linux/bpf.h            |  1 +
 include/uapi/linux/bpf.h       |  9 ++++++++-
 kernel/trace/bpf_trace.c       |  2 ++
 net/core/filter.c              | 22 ++++++++++++++++++++++
 scripts/bpf_helpers_doc.py     |  2 ++
 tools/include/uapi/linux/bpf.h |  9 ++++++++-
 6 files changed, 43 insertions(+), 2 deletions(-)
include/linux/bpf.h +1 -0
index c23998cf6699..3d2ade703a35 100644
@@ -1653,6 +1653,7 @@ extern const struct bpf_func_proto bpf_skc_to_tcp6_sock_proto;
 extern const struct bpf_func_proto bpf_skc_to_tcp_sock_proto;
 extern const struct bpf_func_proto bpf_skc_to_tcp_timewait_sock_proto;
 extern const struct bpf_func_proto bpf_skc_to_tcp_request_sock_proto;
+extern const struct bpf_func_proto bpf_skc_to_udp6_sock_proto;
 
 const struct bpf_func_proto *bpf_tracing_func_proto(
 	enum bpf_func_id func_id, const struct bpf_prog *prog);
include/uapi/linux/bpf.h +8 -1
index e256417d94c2..3f4b12c5c563 100644
@@ -3276,6 +3276,12 @@ union bpf_attr {
  *		Dynamically cast a *sk* pointer to a *tcp_request_sock* pointer.
  *	Return
  *		*sk* if casting is valid, or NULL otherwise.
+ *
+ * struct udp6_sock *bpf_skc_to_udp6_sock(void *sk)
+ * 	Description
+ *		Dynamically cast a *sk* pointer to a *udp6_sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or NULL otherwise.
  */
 #define __BPF_FUNC_MAPPER(FN)		\
 	FN(unspec),			\
@@ -3417,7 +3423,8 @@ union bpf_attr {
 	FN(skc_to_tcp6_sock),		\
 	FN(skc_to_tcp_sock),		\
 	FN(skc_to_tcp_timewait_sock),	\
-	FN(skc_to_tcp_request_sock),
+	FN(skc_to_tcp_request_sock),	\
+	FN(skc_to_udp6_sock),
 
 /* integer value in 'imm' field of BPF_CALL instruction selects which helper
  * function eBPF program intends to call
kernel/trace/bpf_trace.c +2 -0
index de5fbe66e1ca..d10ab16c4a2f 100644
@@ -1523,6 +1523,8 @@ tracing_prog_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_skc_to_tcp_timewait_sock_proto;
 	case BPF_FUNC_skc_to_tcp_request_sock:
 		return &bpf_skc_to_tcp_request_sock_proto;
+	case BPF_FUNC_skc_to_udp6_sock:
+		return &bpf_skc_to_udp6_sock_proto;
 #endif
 	case BPF_FUNC_seq_printf:
 		return prog->expected_attach_type == BPF_TRACE_ITER ?
net/core/filter.c +22 -0
index a285a2d00edc..416c6b232b01 100644
@@ -9335,3 +9335,25 @@ const struct bpf_func_proto bpf_skc_to_tcp_request_sock_proto = {
 	.check_btf_id		= check_arg_btf_id,
 	.ret_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_TCP_REQ],
 };
+
+BPF_CALL_1(bpf_skc_to_udp6_sock, struct sock *, sk)
+{
+	/* udp6_sock type is not generated in dwarf and hence btf,
+	 * trigger an explicit type generation here.
+	 */
+	BTF_TYPE_EMIT(struct udp6_sock);
+	if (sk_fullsock(sk) && sk->sk_protocol == IPPROTO_UDP &&
+	    sk->sk_type == SOCK_DGRAM && sk->sk_family == AF_INET6)
+		return (unsigned long)sk;
+
+	return (unsigned long)NULL;
+}
+
+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,
+	.check_btf_id		= check_arg_btf_id,
+	.ret_btf_id		= &btf_sock_ids[BTF_SOCK_TYPE_UDP6],
+};
scripts/bpf_helpers_doc.py +2 -0
index d886657c6aaa..6bab40ff442e 100755
@@ -425,6 +425,7 @@ class PrinterHelpers(Printer):
             'struct tcp_sock',
             'struct tcp_timewait_sock',
             'struct tcp_request_sock',
+            'struct udp6_sock',
 
             'struct __sk_buff',
             'struct sk_msg_md',
@@ -466,6 +467,7 @@ class PrinterHelpers(Printer):
             'struct tcp_sock',
             'struct tcp_timewait_sock',
             'struct tcp_request_sock',
+            'struct udp6_sock',
     }
     mapped_types = {
             'u8': '__u8',
tools/include/uapi/linux/bpf.h +8 -1
index e256417d94c2..3f4b12c5c563 100644
@@ -3276,6 +3276,12 @@ union bpf_attr {
  *		Dynamically cast a *sk* pointer to a *tcp_request_sock* pointer.
  *	Return
  *		*sk* if casting is valid, or NULL otherwise.
+ *
+ * struct udp6_sock *bpf_skc_to_udp6_sock(void *sk)
+ * 	Description
+ *		Dynamically cast a *sk* pointer to a *udp6_sock* pointer.
+ *	Return
+ *		*sk* if casting is valid, or NULL otherwise.
  */
 #define __BPF_FUNC_MAPPER(FN)		\
 	FN(unspec),			\
@@ -3417,7 +3423,8 @@ union bpf_attr {
 	FN(skc_to_tcp6_sock),		\
 	FN(skc_to_tcp_sock),		\
 	FN(skc_to_tcp_timewait_sock),	\
-	FN(skc_to_tcp_request_sock),
+	FN(skc_to_tcp_request_sock),	\
+	FN(skc_to_udp6_sock),
 
 /* integer value in 'imm' field of BPF_CALL instruction selects which helper
  * function eBPF program intends to call
Signature · 1 line
2.24.1
#10 lore raw

PATCH v5 10/15 selftests/bpf: move newer bpf_iter_* type redefining to a new header file

In reply to Yonghong Song

Commit b9f4c01f3e0b ("selftest/bpf: Make bpf_iter selftest compilable against old vmlinux.h") and Commit dda18a5c0b75 ("selftests/bpf: Convert bpf_iter_test_kern{3, 4}.c to define own bpf_iter_meta") redefined newly introduced types in bpf programs so the bpf program can still compile properly with old kernels although loading may fail.

Since this patch set introduced new types and the same workaround is needed, so let us move the workaround to a separate header file so they do not clutter bpf programs.

---
 tools/testing/selftests/bpf/progs/bpf_iter.h  | 49 +++++++++++++++++++
 .../selftests/bpf/progs/bpf_iter_bpf_map.c    | 18 +------
 .../selftests/bpf/progs/bpf_iter_ipv6_route.c | 18 +------
 .../selftests/bpf/progs/bpf_iter_netlink.c    | 18 +------
 .../selftests/bpf/progs/bpf_iter_task.c       | 18 +------
 .../selftests/bpf/progs/bpf_iter_task_file.c  | 20 +-------
 .../selftests/bpf/progs/bpf_iter_test_kern3.c | 17 +------
 .../selftests/bpf/progs/bpf_iter_test_kern4.c | 17 +------
 .../bpf/progs/bpf_iter_test_kern_common.h     | 18 +------
 9 files changed, 57 insertions(+), 136 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_iter.h
tools/testing/selftests/bpf/progs/bpf_iter.h new file +49 -0
new file mode 100644
index 000000000000..3757e88c6406
@@ -0,0 +1,49 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2020 Facebook */
+/* "undefine" structs in vmlinux.h, because we "override" them below */
+#define bpf_iter_meta bpf_iter_meta___not_used
+#define bpf_iter__bpf_map bpf_iter__bpf_map___not_used
+#define bpf_iter__ipv6_route bpf_iter__ipv6_route___not_used
+#define bpf_iter__netlink bpf_iter__netlink___not_used
+#define bpf_iter__task bpf_iter__task___not_used
+#define bpf_iter__task_file bpf_iter__task_file___not_used
+#include "vmlinux.h"
+#undef bpf_iter_meta
+#undef bpf_iter__bpf_map
+#undef bpf_iter__ipv6_route
+#undef bpf_iter__netlink
+#undef bpf_iter__task
+#undef bpf_iter__task_file
+
+struct bpf_iter_meta {
+	struct seq_file *seq;
+	__u64 session_id;
+	__u64 seq_num;
+} __attribute__((preserve_access_index));
+
+struct bpf_iter__ipv6_route {
+	struct bpf_iter_meta *meta;
+	struct fib6_info *rt;
+} __attribute__((preserve_access_index));
+
+struct bpf_iter__netlink {
+	struct bpf_iter_meta *meta;
+	struct netlink_sock *sk;
+} __attribute__((preserve_access_index));
+
+struct bpf_iter__task {
+	struct bpf_iter_meta *meta;
+	struct task_struct *task;
+} __attribute__((preserve_access_index));
+
+struct bpf_iter__task_file {
+	struct bpf_iter_meta *meta;
+	struct task_struct *task;
+	__u32 fd;
+	struct file *file;
+} __attribute__((preserve_access_index));
+
+struct bpf_iter__bpf_map {
+	struct bpf_iter_meta *meta;
+	struct bpf_map *map;
+} __attribute__((preserve_access_index));
tools/testing/selftests/bpf/progs/bpf_iter_bpf_map.c +1 -17
index b57bd6fef208..08651b23edba 100644
@@ -1,27 +1,11 @@
 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2020 Facebook */
-/* "undefine" structs in vmlinux.h, because we "override" them below */
-#define bpf_iter_meta bpf_iter_meta___not_used
-#define bpf_iter__bpf_map bpf_iter__bpf_map___not_used
-#include "vmlinux.h"
-#undef bpf_iter_meta
-#undef bpf_iter__bpf_map
+#include "bpf_iter.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 
 char _license[] SEC("license") = "GPL";
 
-struct bpf_iter_meta {
-	struct seq_file *seq;
-	__u64 session_id;
-	__u64 seq_num;
-} __attribute__((preserve_access_index));
-
-struct bpf_iter__bpf_map {
-	struct bpf_iter_meta *meta;
-	struct bpf_map *map;
-} __attribute__((preserve_access_index));
-
 SEC("iter/bpf_map")
 int dump_bpf_map(struct bpf_iter__bpf_map *ctx)
 {
tools/testing/selftests/bpf/progs/bpf_iter_ipv6_route.c +1 -17
index c8e9ca74c87b..93a452d1d136 100644
@@ -1,25 +1,9 @@
 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2020 Facebook */
-/* "undefine" structs in vmlinux.h, because we "override" them below */
-#define bpf_iter_meta bpf_iter_meta___not_used
-#define bpf_iter__ipv6_route bpf_iter__ipv6_route___not_used
-#include "vmlinux.h"
-#undef bpf_iter_meta
-#undef bpf_iter__ipv6_route
+#include "bpf_iter.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 
-struct bpf_iter_meta {
-	struct seq_file *seq;
-	__u64 session_id;
-	__u64 seq_num;
-} __attribute__((preserve_access_index));
-
-struct bpf_iter__ipv6_route {
-	struct bpf_iter_meta *meta;
-	struct fib6_info *rt;
-} __attribute__((preserve_access_index));
-
 char _license[] SEC("license") = "GPL";
 
 extern bool CONFIG_IPV6_SUBTREES __kconfig __weak;
tools/testing/selftests/bpf/progs/bpf_iter_netlink.c +1 -17
index e7b8753eac0b..fda5036fdf75 100644
@@ -1,11 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2020 Facebook */
-/* "undefine" structs in vmlinux.h, because we "override" them below */
-#define bpf_iter_meta bpf_iter_meta___not_used
-#define bpf_iter__netlink bpf_iter__netlink___not_used
-#include "vmlinux.h"
-#undef bpf_iter_meta
-#undef bpf_iter__netlink
+#include "bpf_iter.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 
@@ -14,17 +9,6 @@ char _license[] SEC("license") = "GPL";
 #define sk_rmem_alloc	sk_backlog.rmem_alloc
 #define sk_refcnt	__sk_common.skc_refcnt
 
-struct bpf_iter_meta {
-	struct seq_file *seq;
-	__u64 session_id;
-	__u64 seq_num;
-} __attribute__((preserve_access_index));
-
-struct bpf_iter__netlink {
-	struct bpf_iter_meta *meta;
-	struct netlink_sock *sk;
-} __attribute__((preserve_access_index));
-
 static inline struct inode *SOCK_INODE(struct socket *socket)
 {
 	return &container_of(socket, struct socket_alloc, socket)->vfs_inode;
tools/testing/selftests/bpf/progs/bpf_iter_task.c +1 -17
index ee754021f98e..4983087852a0 100644
@@ -1,27 +1,11 @@
 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2020 Facebook */
-/* "undefine" structs in vmlinux.h, because we "override" them below */
-#define bpf_iter_meta bpf_iter_meta___not_used
-#define bpf_iter__task bpf_iter__task___not_used
-#include "vmlinux.h"
-#undef bpf_iter_meta
-#undef bpf_iter__task
+#include "bpf_iter.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 
 char _license[] SEC("license") = "GPL";
 
-struct bpf_iter_meta {
-	struct seq_file *seq;
-	__u64 session_id;
-	__u64 seq_num;
-} __attribute__((preserve_access_index));
-
-struct bpf_iter__task {
-	struct bpf_iter_meta *meta;
-	struct task_struct *task;
-} __attribute__((preserve_access_index));
-
 SEC("iter/task")
 int dump_task(struct bpf_iter__task *ctx)
 {
tools/testing/selftests/bpf/progs/bpf_iter_task_file.c +1 -19
index 0f0ec3db20ba..8b787baa2654 100644
@@ -1,29 +1,11 @@
 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2020 Facebook */
-/* "undefine" structs in vmlinux.h, because we "override" them below */
-#define bpf_iter_meta bpf_iter_meta___not_used
-#define bpf_iter__task_file bpf_iter__task_file___not_used
-#include "vmlinux.h"
-#undef bpf_iter_meta
-#undef bpf_iter__task_file
+#include "bpf_iter.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 
 char _license[] SEC("license") = "GPL";
 
-struct bpf_iter_meta {
-	struct seq_file *seq;
-	__u64 session_id;
-	__u64 seq_num;
-} __attribute__((preserve_access_index));
-
-struct bpf_iter__task_file {
-	struct bpf_iter_meta *meta;
-	struct task_struct *task;
-	__u32 fd;
-	struct file *file;
-} __attribute__((preserve_access_index));
-
 SEC("iter/task_file")
 int dump_task_file(struct bpf_iter__task_file *ctx)
 {
tools/testing/selftests/bpf/progs/bpf_iter_test_kern3.c +1 -16
index 13c2c90c835f..2a4647f20c46 100644
@@ -1,25 +1,10 @@
 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2020 Facebook */
-#define bpf_iter_meta bpf_iter_meta___not_used
-#define bpf_iter__task bpf_iter__task___not_used
-#include "vmlinux.h"
-#undef bpf_iter_meta
-#undef bpf_iter__task
+#include "bpf_iter.h"
 #include <bpf/bpf_helpers.h>
 
 char _license[] SEC("license") = "GPL";
 
-struct bpf_iter_meta {
-	struct seq_file *seq;
-	__u64 session_id;
-	__u64 seq_num;
-} __attribute__((preserve_access_index));
-
-struct bpf_iter__task {
-	struct bpf_iter_meta *meta;
-	struct task_struct *task;
-} __attribute__((preserve_access_index));
-
 SEC("iter/task")
 int dump_task(struct bpf_iter__task *ctx)
 {
tools/testing/selftests/bpf/progs/bpf_iter_test_kern4.c +1 -16
index 0aa71b333cf3..ee49493dc125 100644
@@ -1,25 +1,10 @@
 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2020 Facebook */
-#define bpf_iter_meta bpf_iter_meta___not_used
-#define bpf_iter__bpf_map bpf_iter__bpf_map___not_used
-#include "vmlinux.h"
-#undef bpf_iter_meta
-#undef bpf_iter__bpf_map
+#include "bpf_iter.h"
 #include <bpf/bpf_helpers.h>
 
 char _license[] SEC("license") = "GPL";
 
-struct bpf_iter_meta {
-	struct seq_file *seq;
-	__u64 session_id;
-	__u64 seq_num;
-} __attribute__((preserve_access_index));
-
-struct bpf_iter__bpf_map {
-	struct bpf_iter_meta *meta;
-	struct bpf_map *map;
-} __attribute__((preserve_access_index));
-
 __u32 map1_id = 0, map2_id = 0;
 __u32 map1_accessed = 0, map2_accessed = 0;
 __u64 map1_seqnum = 0, map2_seqnum1 = 0, map2_seqnum2 = 0;
tools/testing/selftests/bpf/progs/bpf_iter_test_kern_common.h +1 -17
index dee1339e6905..d5e3df66ad9a 100644
@@ -1,27 +1,11 @@
 /* SPDX-License-Identifier: GPL-2.0 */
 /* Copyright (c) 2020 Facebook */
-/* "undefine" structs in vmlinux.h, because we "override" them below */
-#define bpf_iter_meta bpf_iter_meta___not_used
-#define bpf_iter__task bpf_iter__task___not_used
-#include "vmlinux.h"
-#undef bpf_iter_meta
-#undef bpf_iter__task
+#include "bpf_iter.h"
 #include <bpf/bpf_helpers.h>
 
 char _license[] SEC("license") = "GPL";
 int count = 0;
 
-struct bpf_iter_meta {
-	struct seq_file *seq;
-	__u64 session_id;
-	__u64 seq_num;
-} __attribute__((preserve_access_index));
-
-struct bpf_iter__task {
-	struct bpf_iter_meta *meta;
-	struct task_struct *task;
-} __attribute__((preserve_access_index));
-
 SEC("iter/task")
 int dump_task(struct bpf_iter__task *ctx)
 {
Signature · 1 line
2.24.1
#11 lore raw

PATCH v5 11/15 selftests/bpf: refactor some net macros to bpf_tracing_net.h

In reply to Yonghong Song

Refactor bpf_iter_ipv6_route.c and bpf_iter_netlink.c so net macros, originally from various include/linux header files, are moved to a new header file bpf_tracing_net.h. The goal is to improve reuse so networking tracing programs do not need to
copy these macros every time they use them.

---
 .../selftests/bpf/progs/bpf_iter_ipv6_route.c    |  7 +------
 .../selftests/bpf/progs/bpf_iter_netlink.c       |  4 +---
 .../selftests/bpf/progs/bpf_tracing_net.h        | 16 ++++++++++++++++
 3 files changed, 18 insertions(+), 9 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_tracing_net.h
tools/testing/selftests/bpf/progs/bpf_iter_ipv6_route.c +1 -6
index 93a452d1d136..d58d9f1642b5 100644
@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2020 Facebook */
 #include "bpf_iter.h"
+#include "bpf_tracing_net.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 
@@ -8,12 +9,6 @@ char _license[] SEC("license") = "GPL";
 
 extern bool CONFIG_IPV6_SUBTREES __kconfig __weak;
 
-#define RTF_GATEWAY		0x0002
-#define IFNAMSIZ		16
-#define fib_nh_gw_family	nh_common.nhc_gw_family
-#define fib_nh_gw6		nh_common.nhc_gw.ipv6
-#define fib_nh_dev		nh_common.nhc_dev
-
 SEC("iter/ipv6_route")
 int dump_ipv6_route(struct bpf_iter__ipv6_route *ctx)
 {
tools/testing/selftests/bpf/progs/bpf_iter_netlink.c +1 -3
index fda5036fdf75..cec82a419800 100644
@@ -1,14 +1,12 @@
 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2020 Facebook */
 #include "bpf_iter.h"
+#include "bpf_tracing_net.h"
 #include <bpf/bpf_helpers.h>
 #include <bpf/bpf_tracing.h>
 
 char _license[] SEC("license") = "GPL";
 
-#define sk_rmem_alloc	sk_backlog.rmem_alloc
-#define sk_refcnt	__sk_common.skc_refcnt
-
 static inline struct inode *SOCK_INODE(struct socket *socket)
 {
 	return &container_of(socket, struct socket_alloc, socket)->vfs_inode;
tools/testing/selftests/bpf/progs/bpf_tracing_net.h new file +16 -0
new file mode 100644
index 000000000000..1f38a1098727
@@ -0,0 +1,16 @@
+/* SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) */
+#ifndef __BPF_TRACING_NET_H__
+#define __BPF_TRACING_NET_H__
+
+#define IFNAMSIZ		16
+
+#define RTF_GATEWAY		0x0002
+
+#define fib_nh_dev		nh_common.nhc_dev
+#define fib_nh_gw_family	nh_common.nhc_gw_family
+#define fib_nh_gw6		nh_common.nhc_gw.ipv6
+
+#define sk_rmem_alloc		sk_backlog.rmem_alloc
+#define sk_refcnt		__sk_common.skc_refcnt
+
+#endif
Signature · 1 line
2.24.1
#12 lore raw

PATCH v5 12/15 selftests/bpf: add more common macros to bpf_tracing_net.h

In reply to Yonghong Song

These newly added macros will be used in subsequent bpf iterator tcp{4,6} and udp{4,6} programs.

---
 .../selftests/bpf/progs/bpf_tracing_net.h     | 35 +++++++++++++++++++
 1 file changed, 35 insertions(+)
tools/testing/selftests/bpf/progs/bpf_tracing_net.h +35 -0
index 1f38a1098727..01378911252b 100644
@@ -2,15 +2,50 @@
 #ifndef __BPF_TRACING_NET_H__
 #define __BPF_TRACING_NET_H__
 
+#define AF_INET			2
+#define AF_INET6		10
+
+#define ICSK_TIME_RETRANS	1
+#define ICSK_TIME_PROBE0	3
+#define ICSK_TIME_LOSS_PROBE	5
+#define ICSK_TIME_REO_TIMEOUT	6
+
 #define IFNAMSIZ		16
 
 #define RTF_GATEWAY		0x0002
 
+#define TCP_INFINITE_SSTHRESH	0x7fffffff
+#define TCP_PINGPONG_THRESH	3
+
 #define fib_nh_dev		nh_common.nhc_dev
 #define fib_nh_gw_family	nh_common.nhc_gw_family
 #define fib_nh_gw6		nh_common.nhc_gw.ipv6
 
+#define inet_daddr		sk.__sk_common.skc_daddr
+#define inet_rcv_saddr		sk.__sk_common.skc_rcv_saddr
+#define inet_dport		sk.__sk_common.skc_dport
+
+#define ir_loc_addr		req.__req_common.skc_rcv_saddr
+#define ir_num			req.__req_common.skc_num
+#define ir_rmt_addr		req.__req_common.skc_daddr
+#define ir_rmt_port		req.__req_common.skc_dport
+#define ir_v6_rmt_addr		req.__req_common.skc_v6_daddr
+#define ir_v6_loc_addr		req.__req_common.skc_v6_rcv_saddr
+
+#define sk_family		__sk_common.skc_family
 #define sk_rmem_alloc		sk_backlog.rmem_alloc
 #define sk_refcnt		__sk_common.skc_refcnt
+#define sk_state		__sk_common.skc_state
+#define sk_v6_daddr		__sk_common.skc_v6_daddr
+#define sk_v6_rcv_saddr		__sk_common.skc_v6_rcv_saddr
+
+#define s6_addr32		in6_u.u6_addr32
+
+#define tw_daddr		__tw_common.skc_daddr
+#define tw_rcv_saddr		__tw_common.skc_rcv_saddr
+#define tw_dport		__tw_common.skc_dport
+#define tw_refcnt		__tw_common.skc_refcnt
+#define tw_v6_daddr		__tw_common.skc_v6_daddr
+#define tw_v6_rcv_saddr		__tw_common.skc_v6_rcv_saddr
 
 #endif
Signature · 1 line
2.24.1
#13 lore raw

PATCH v5 13/15 selftests/bpf: implement sample tcp/tcp6 bpf_iter programs

In reply to Yonghong Song

In my VM, I got identical result compared to /proc/net/{tcp,tcp6}. For tcp6:

  $ cat /proc/net/tcp6
    sl  local_address                         remote_address                        st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode
     0: 00000000000000000000000000000000:0016 00000000000000000000000000000000:0000 0A 00000000:00000000 00:00000001 00000000     0        0 17955 1 000000003eb3102e 100 0 0 10 0
  $ cat /sys/fs/bpf/p1
    sl  local_address                         remote_address                        st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode
     0: 00000000000000000000000000000000:0016 00000000000000000000000000000000:0000 0A 00000000:00000000 00:00000000 00000000     0        0 17955 1 000000003eb3102e 100 0 0 10 0

For tcp:

  $ cat /proc/net/tcp
  sl  local_address rem_address   st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode
   0: 00000000:0016 00000000:0000 0A 00000000:00000000 00:00000000 00000000     0        0 2666 1 000000007152e43f 100 0 0 10 0
  $ cat /sys/fs/bpf/p2
  sl  local_address                         remote_address                        st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode
   1: 00000000:0016 00000000:0000 0A 00000000:00000000 00:00000000 00000000     0        0 2666 1 000000007152e43f 100 0 0 10 0
---
 tools/testing/selftests/bpf/progs/bpf_iter.h  |  15 ++
 .../selftests/bpf/progs/bpf_iter_tcp4.c       | 234 ++++++++++++++++
 .../selftests/bpf/progs/bpf_iter_tcp6.c       | 250 ++++++++++++++++++
 3 files changed, 499 insertions(+)
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_iter_tcp4.c
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_iter_tcp6.c
tools/testing/selftests/bpf/progs/bpf_iter.h +15 -0
index 3757e88c6406..bde23e16e777 100644
@@ -7,6 +7,8 @@
 #define bpf_iter__netlink bpf_iter__netlink___not_used
 #define bpf_iter__task bpf_iter__task___not_used
 #define bpf_iter__task_file bpf_iter__task_file___not_used
+#define bpf_iter__tcp bpf_iter__tcp___not_used
+#define tcp6_sock tcp6_sock___not_used
 #include "vmlinux.h"
 #undef bpf_iter_meta
 #undef bpf_iter__bpf_map
@@ -14,6 +16,8 @@
 #undef bpf_iter__netlink
 #undef bpf_iter__task
 #undef bpf_iter__task_file
+#undef bpf_iter__tcp
+#undef tcp6_sock
 
 struct bpf_iter_meta {
 	struct seq_file *seq;
@@ -47,3 +51,14 @@ struct bpf_iter__bpf_map {
 	struct bpf_iter_meta *meta;
 	struct bpf_map *map;
 } __attribute__((preserve_access_index));
+
+struct bpf_iter__tcp {
+	struct bpf_iter_meta *meta;
+	struct sock_common *sk_common;
+	uid_t uid;
+} __attribute__((preserve_access_index));
+
+struct tcp6_sock {
+	struct tcp_sock	tcp;
+	struct ipv6_pinfo inet6;
+} __attribute__((preserve_access_index));
tools/testing/selftests/bpf/progs/bpf_iter_tcp4.c new file +234 -0

234 lines not on this page · raw message

tools/testing/selftests/bpf/progs/bpf_iter_tcp6.c new file +250 -0

250 lines not on this page · raw message

Signature · 1 line
2.24.1
#14 lore raw

PATCH v5 14/15 selftests/bpf: implement sample udp/udp6 bpf_iter programs

In reply to Yonghong Song

On my VM, I got identical results between /proc/net/udp[6] and the udp{4,6} bpf iterator.

For udp6:

  $ cat /sys/fs/bpf/p1
    sl  local_address                         remote_address                        st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode ref pointer drops
   1405: 000080FE00000000FF7CC4D0D9EFE4FE:0222 00000000000000000000000000000000:0000 07 00000000:00000000 00:00000000 00000000   193        0 19183 2 0000000029eab111 0
  $ cat /proc/net/udp6
    sl  local_address                         remote_address                        st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode ref pointer drops
   1405: 000080FE00000000FF7CC4D0D9EFE4FE:0222 00000000000000000000000000000000:0000 07 00000000:00000000 00:00000000 00000000   193        0 19183 2 0000000029eab111 0

For udp4:

  $ cat /sys/fs/bpf/p4
    sl  local_address rem_address   st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode ref pointer drops
   2007: 00000000:1F90 00000000:0000 07 00000000:00000000 00:00000000 00000000     0        0 72540 2 000000004ede477a 0
  $ cat /proc/net/udp
    sl  local_address rem_address   st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode ref pointer drops
   2007: 00000000:1F90 00000000:0000 07 00000000:00000000 00:00000000 00000000     0        0 72540 2 000000004ede477a 0
---
 tools/testing/selftests/bpf/progs/bpf_iter.h  | 16 ++++
 .../selftests/bpf/progs/bpf_iter_udp4.c       | 71 +++++++++++++++++
 .../selftests/bpf/progs/bpf_iter_udp6.c       | 79 +++++++++++++++++++
 3 files changed, 166 insertions(+)
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_iter_udp4.c
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_iter_udp6.c
tools/testing/selftests/bpf/progs/bpf_iter.h +16 -0
index bde23e16e777..17db3bac518b 100644
@@ -9,6 +9,8 @@
 #define bpf_iter__task_file bpf_iter__task_file___not_used
 #define bpf_iter__tcp bpf_iter__tcp___not_used
 #define tcp6_sock tcp6_sock___not_used
+#define bpf_iter__udp bpf_iter__udp___not_used
+#define udp6_sock udp6_sock___not_used
 #include "vmlinux.h"
 #undef bpf_iter_meta
 #undef bpf_iter__bpf_map
@@ -18,6 +20,8 @@
 #undef bpf_iter__task_file
 #undef bpf_iter__tcp
 #undef tcp6_sock
+#undef bpf_iter__udp
+#undef udp6_sock
 
 struct bpf_iter_meta {
 	struct seq_file *seq;
@@ -62,3 +66,15 @@ struct tcp6_sock {
 	struct tcp_sock	tcp;
 	struct ipv6_pinfo inet6;
 } __attribute__((preserve_access_index));
+
+struct bpf_iter__udp {
+	struct bpf_iter_meta *meta;
+	struct udp_sock *udp_sk;
+	uid_t uid __attribute__((aligned(8)));
+	int bucket __attribute__((aligned(8)));
+} __attribute__((preserve_access_index));
+
+struct udp6_sock {
+	struct udp_sock	udp;
+	struct ipv6_pinfo inet6;
+} __attribute__((preserve_access_index));
tools/testing/selftests/bpf/progs/bpf_iter_udp4.c new file +71 -0

71 lines not on this page · raw message

tools/testing/selftests/bpf/progs/bpf_iter_udp6.c new file +79 -0

79 lines not on this page · raw message

Signature · 1 line
2.24.1
#15 lore raw

PATCH v5 15/15 selftests/bpf: add tcp/udp iterator programs to selftests

In reply to Yonghong Song

Added tcp{4,6} and udp{4,6} bpf programs into test_progs selftest so that they at least can load successfully.

  $ ./test_progs -n 3
  ...
  #3/7 tcp4:OK
  #3/8 tcp6:OK
  #3/9 udp4:OK
  #3/10 udp6:OK
  ...
  #3 bpf_iter:OK
  Summary: 1/16 PASSED, 0 SKIPPED, 0 FAILED
---
 .../selftests/bpf/prog_tests/bpf_iter.c       | 68 +++++++++++++++++++
 1 file changed, 68 insertions(+)
tools/testing/selftests/bpf/prog_tests/bpf_iter.c +68 -0
index 87c29dde1cf9..1e2e0fced6e8 100644
@@ -6,6 +6,10 @@
 #include "bpf_iter_bpf_map.skel.h"
 #include "bpf_iter_task.skel.h"
 #include "bpf_iter_task_file.skel.h"
+#include "bpf_iter_tcp4.skel.h"
+#include "bpf_iter_tcp6.skel.h"
+#include "bpf_iter_udp4.skel.h"
+#include "bpf_iter_udp6.skel.h"
 #include "bpf_iter_test_kern1.skel.h"
 #include "bpf_iter_test_kern2.skel.h"
 #include "bpf_iter_test_kern3.skel.h"
@@ -120,6 +124,62 @@ static void test_task_file(void)
 	bpf_iter_task_file__destroy(skel);
 }
 
+static void test_tcp4(void)
+{
+	struct bpf_iter_tcp4 *skel;
+
+	skel = bpf_iter_tcp4__open_and_load();
+	if (CHECK(!skel, "bpf_iter_tcp4__open_and_load",
+		  "skeleton open_and_load failed\n"))
+		return;
+
+	do_dummy_read(skel->progs.dump_tcp4);
+
+	bpf_iter_tcp4__destroy(skel);
+}
+
+static void test_tcp6(void)
+{
+	struct bpf_iter_tcp6 *skel;
+
+	skel = bpf_iter_tcp6__open_and_load();
+	if (CHECK(!skel, "bpf_iter_tcp6__open_and_load",
+		  "skeleton open_and_load failed\n"))
+		return;
+
+	do_dummy_read(skel->progs.dump_tcp6);
+
+	bpf_iter_tcp6__destroy(skel);
+}
+
+static void test_udp4(void)
+{
+	struct bpf_iter_udp4 *skel;
+
+	skel = bpf_iter_udp4__open_and_load();
+	if (CHECK(!skel, "bpf_iter_udp4__open_and_load",
+		  "skeleton open_and_load failed\n"))
+		return;
+
+	do_dummy_read(skel->progs.dump_udp4);
+
+	bpf_iter_udp4__destroy(skel);
+}
+
+static void test_udp6(void)
+{
+	struct bpf_iter_udp6 *skel;
+
+	skel = bpf_iter_udp6__open_and_load();
+	if (CHECK(!skel, "bpf_iter_udp6__open_and_load",
+		  "skeleton open_and_load failed\n"))
+		return;
+
+	do_dummy_read(skel->progs.dump_udp6);
+
+	bpf_iter_udp6__destroy(skel);
+}
+
 /* The expected string is less than 16 bytes */
 static int do_read_with_fd(int iter_fd, const char *expected,
 			   bool read_one_char)
@@ -394,6 +454,14 @@ void test_bpf_iter(void)
 		test_task();
 	if (test__start_subtest("task_file"))
 		test_task_file();
+	if (test__start_subtest("tcp4"))
+		test_tcp4();
+	if (test__start_subtest("tcp6"))
+		test_tcp6();
+	if (test__start_subtest("udp4"))
+		test_udp4();
+	if (test__start_subtest("udp6"))
+		test_udp6();
 	if (test__start_subtest("anon"))
 		test_anon_iter(false);
 	if (test__start_subtest("anon-read-one-char"))
Signature · 1 line
2.24.1
#16 lore raw
In reply to Yonghong Song
Quoted 35 lines · Yonghong Song

bpf iterator implments traversal of kernel data structures and these data structures are passed to a bpf program for processing. This gives great flexibility for users to examine kernel data structure without using e.g. /proc/net which has limited and fixed format.

Commit 138d0be35b14 ("net: bpf: Add netlink and ipv6_route bpf_iter targets") implemented bpf iterators for netlink and ipv6_route.
This patch set intends to implement bpf iterators for tcp and udp.

Currently, /proc/net/tcp is used to print tcp4 stats and /proc/net/tcp6 is used to print tcp6 stats. /proc/net/udp[6] have similar usage model. In contrast, only one tcp iterator is implemented and it is bpf program resposibility to filter based on socket family. The same is for udp. This will avoid another unnecessary traversal pass if users want to check both tcp4 and tcp6.

Several helpers are also implemented in this patch
bpf_skc_to_{tcp, tcp6, tcp_timewait, tcp_request, udp6}_sock
The argument for these helpers is not a fixed btf_id. For example,
bpf_skc_to_tcp(struct sock_common *), or bpf_skc_to_tcp(struct sock *), or
bpf_skc_to_tcp(struct inet_sock *), ...
are all valid. At runtime, the helper will check whether pointer cast is legal or not. Please see Patch #5 for details.

Since btf_id's for both arguments and return value are known at build time, the btf_id's are pre-computed once vmlinux btf becomes valid. Jiri's "adding d_path helper" patch set
https://lore.kernel.org/bpf/[email protected]/T/
provides a way to pre-compute btf id during vmlinux build time. This can be applied here as well. A followup patch can convert to build time btf id computation after Jiri's patch landed.

Quoted 5 lines

Changelogs:
v4 -> v5:
- fix bpf_skc_to_udp6_sock helper as besides sk_protocol, sk_family, sk_type == SOCK_DGRAM is also needed to differentiate from SOCK_RAW (Eric)

Applied. Thanks