[PATCH bpf-next 1/2] bpf: allow to retrieve sol_socket opts from sock_addr progs

Subsystems: bpf [general] (safe dynamic programs and tools), bpf [networking] (tcx & tc bpf, sock_addr), networking [general], the rest

STALE2038d

5 messages, 2 authors, 2021-01-11 · open the first message on its own page

[PATCH bpf-next 1/2] bpf: allow to retrieve sol_socket opts from sock_addr progs

From: Daniel Borkmann <daniel@iogearbox.net>
Date: 2021-01-11 16:18:52

The _bpf_setsockopt() is able to set some of the SOL_SOCKET level options,
however, _bpf_getsockopt() has little support to actually retrieve them.
This small patch adds few misc options such as SO_MARK, SO_PRIORITY and
SO_BINDTOIFINDEX. For the latter getter and setter are added. The mark and
priority in particular allow to retrieve the options from BPF cgroup hooks
to then implement custom behavior / settings on the syscall hooks compared
to other sockets that stick to the defaults, for example.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
---
 net/core/filter.c | 25 +++++++++++++++++++++++--
 1 file changed, 23 insertions(+), 2 deletions(-)

diff --git a/net/core/filter.c b/net/core/filter.c
index 255aeee72402..9ab94e90d660 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c

@@ -4770,6 +4770,10 @@ static int _bpf_setsockopt(struct sock *sk, int level, int optname,
 				ifindex = dev->ifindex;
 				dev_put(dev);
 			}
+			fallthrough;
+		case SO_BINDTOIFINDEX:
+			if (optname == SO_BINDTOIFINDEX)
+				ifindex = val;
 			ret = sock_bindtoindex(sk, ifindex, false);
 			break;
 		case SO_KEEPALIVE:

@@ -4932,8 +4936,25 @@ static int _bpf_getsockopt(struct sock *sk, int level, int optname,
 
 	sock_owned_by_me(sk);
 
+	if (level == SOL_SOCKET) {
+		if (optlen != sizeof(int))
+			goto err_clear;
+
+		switch (optname) {
+		case SO_MARK:
+			*((int *)optval) = sk->sk_mark;
+			break;
+		case SO_PRIORITY:
+			*((int *)optval) = sk->sk_priority;
+			break;
+		case SO_BINDTOIFINDEX:
+			*((int *)optval) = sk->sk_bound_dev_if;
+			break;
+		default:
+			goto err_clear;
+		}
 #ifdef CONFIG_INET
-	if (level == SOL_TCP && sk->sk_prot->getsockopt == tcp_getsockopt) {
+	} else if (level == SOL_TCP && sk->sk_prot->getsockopt == tcp_getsockopt) {
 		struct inet_connection_sock *icsk;
 		struct tcp_sock *tp;

@@ -4986,12 +5007,12 @@ static int _bpf_getsockopt(struct sock *sk, int level, int optname,
 		default:
 			goto err_clear;
 		}
+#endif
 #endif
 	} else {
 		goto err_clear;
 	}
 	return 0;
-#endif
 err_clear:
 	memset(optval, 0, optlen);
 	return -EINVAL;

-- 
2.21.0

[PATCH bpf-next 2/2] bpf: extend bind v4/v6 selftests for mark/prio/bindtoifindex

From: Daniel Borkmann <daniel@iogearbox.net>
Date: 2021-01-11 16:18:52

Extend existing cgroup bind4/bind6 tests to add coverage for setting and
retrieving SO_MARK, SO_PRIORITY and SO_BINDTOIFINDEX at the bind hook.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
---
 .../testing/selftests/bpf/progs/bind4_prog.c  | 41 +++++++++++++++++--
 .../testing/selftests/bpf/progs/bind6_prog.c  | 41 +++++++++++++++++--
 2 files changed, 74 insertions(+), 8 deletions(-)

diff --git a/tools/testing/selftests/bpf/progs/bind4_prog.c b/tools/testing/selftests/bpf/progs/bind4_prog.c
index c6520f21f5f5..4479ac27b1d3 100644
--- a/tools/testing/selftests/bpf/progs/bind4_prog.c
+++ b/tools/testing/selftests/bpf/progs/bind4_prog.c

@@ -29,18 +29,47 @@ static __inline int bind_to_device(struct bpf_sock_addr *ctx)
 	char veth2[IFNAMSIZ] = "test_sock_addr2";
 	char missing[IFNAMSIZ] = "nonexistent_dev";
 	char del_bind[IFNAMSIZ] = "";
+	int veth1_idx, veth2_idx;
 
 	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&veth1, sizeof(veth1)))
+			   &veth1, sizeof(veth1)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth1_idx, sizeof(veth1_idx)) || !veth1_idx)
 		return 1;
 	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&veth2, sizeof(veth2)))
+			   &veth2, sizeof(veth2)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth2_idx, sizeof(veth2_idx)) || !veth2_idx ||
+	    veth1_idx == veth2_idx)
 		return 1;
 	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&missing, sizeof(missing)) != -ENODEV)
+			   &missing, sizeof(missing)) != -ENODEV)
+		return 1;
+	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth1_idx, sizeof(veth1_idx)))
 		return 1;
 	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&del_bind, sizeof(del_bind)))
+			   &del_bind, sizeof(del_bind)))
+		return 1;
+
+	return 0;
+}
+
+static __inline int misc_opts(struct bpf_sock_addr *ctx, int opt)
+{
+	int old, tmp, new = 0xeb9f;
+
+	if (bpf_getsockopt(ctx, SOL_SOCKET, opt, &old, sizeof(old)) ||
+	    old == new)
+		return 1;
+	if (bpf_setsockopt(ctx, SOL_SOCKET, opt, &new, sizeof(new)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, opt, &tmp, sizeof(tmp)) ||
+	    tmp != new)
+		return 1;
+	if (bpf_setsockopt(ctx, SOL_SOCKET, opt, &old, sizeof(old)))
 		return 1;
 
 	return 0;

@@ -93,6 +122,10 @@ int bind_v4_prog(struct bpf_sock_addr *ctx)
 	if (bind_to_device(ctx))
 		return 0;
 
+	/* Test for misc socket options. */
+	if (misc_opts(ctx, SO_MARK) || misc_opts(ctx, SO_PRIORITY))
+		return 0;
+
 	ctx->user_ip4 = bpf_htonl(SERV4_REWRITE_IP);
 	ctx->user_port = bpf_htons(SERV4_REWRITE_PORT);

diff --git a/tools/testing/selftests/bpf/progs/bind6_prog.c b/tools/testing/selftests/bpf/progs/bind6_prog.c
index 4358e44dcf47..1b4142fcdd4b 100644
--- a/tools/testing/selftests/bpf/progs/bind6_prog.c
+++ b/tools/testing/selftests/bpf/progs/bind6_prog.c

@@ -35,18 +35,47 @@ static __inline int bind_to_device(struct bpf_sock_addr *ctx)
 	char veth2[IFNAMSIZ] = "test_sock_addr2";
 	char missing[IFNAMSIZ] = "nonexistent_dev";
 	char del_bind[IFNAMSIZ] = "";
+	int veth1_idx, veth2_idx;
 
 	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&veth1, sizeof(veth1)))
+			   &veth1, sizeof(veth1)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth1_idx, sizeof(veth1_idx)) || !veth1_idx)
 		return 1;
 	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&veth2, sizeof(veth2)))
+			   &veth2, sizeof(veth2)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth2_idx, sizeof(veth2_idx)) || !veth2_idx ||
+	    veth1_idx == veth2_idx)
 		return 1;
 	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&missing, sizeof(missing)) != -ENODEV)
+			   &missing, sizeof(missing)) != -ENODEV)
+		return 1;
+	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth1_idx, sizeof(veth1_idx)))
 		return 1;
 	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&del_bind, sizeof(del_bind)))
+			   &del_bind, sizeof(del_bind)))
+		return 1;
+
+	return 0;
+}
+
+static __inline int misc_opts(struct bpf_sock_addr *ctx, int opt)
+{
+	int old, tmp, new = 0xeb9f;
+
+	if (bpf_getsockopt(ctx, SOL_SOCKET, opt, &old, sizeof(old)) ||
+	    old == new)
+		return 1;
+	if (bpf_setsockopt(ctx, SOL_SOCKET, opt, &new, sizeof(new)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, opt, &tmp, sizeof(tmp)) ||
+	    tmp != new)
+		return 1;
+	if (bpf_setsockopt(ctx, SOL_SOCKET, opt, &old, sizeof(old)))
 		return 1;
 
 	return 0;

@@ -107,6 +136,10 @@ int bind_v6_prog(struct bpf_sock_addr *ctx)
 	if (bind_to_device(ctx))
 		return 0;
 
+	/* Test for misc socket options. */
+	if (misc_opts(ctx, SO_MARK) || misc_opts(ctx, SO_PRIORITY))
+		return 0;
+
 	ctx->user_ip6[0] = bpf_htonl(SERV6_REWRITE_IP_0);
 	ctx->user_ip6[1] = bpf_htonl(SERV6_REWRITE_IP_1);
 	ctx->user_ip6[2] = bpf_htonl(SERV6_REWRITE_IP_2);

-- 
2.21.0

Re: [PATCH bpf-next 1/2] bpf: allow to retrieve sol_socket opts from sock_addr progs

From: Yonghong Song <hidden>
Date: 2021-01-11 20:04:39


On 1/11/21 8:17 AM, Daniel Borkmann wrote:

The _bpf_setsockopt() is able to set some of the SOL_SOCKET level options,
however, _bpf_getsockopt() has little support to actually retrieve them.
This small patch adds few misc options such as SO_MARK, SO_PRIORITY and
SO_BINDTOIFINDEX. For the latter getter and setter are added. The mark and
priority in particular allow to retrieve the options from BPF cgroup hooks
to then implement custom behavior / settings on the syscall hooks compared
to other sockets that stick to the defaults, for example.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Acked-by: Yonghong Song <redacted>

Re: [PATCH bpf-next 2/2] bpf: extend bind v4/v6 selftests for mark/prio/bindtoifindex

From: Yonghong Song <hidden>
Date: 2021-01-11 20:16:15


On 1/11/21 8:17 AM, Daniel Borkmann wrote:

Extend existing cgroup bind4/bind6 tests to add coverage for setting and
retrieving SO_MARK, SO_PRIORITY and SO_BINDTOIFINDEX at the bind hook.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Ack with a minor comments below.

Acked-by: Yonghong Song <redacted>

quoted hunk

---
  .../testing/selftests/bpf/progs/bind4_prog.c  | 41 +++++++++++++++++--
  .../testing/selftests/bpf/progs/bind6_prog.c  | 41 +++++++++++++++++--
  2 files changed, 74 insertions(+), 8 deletions(-)

diff --git a/tools/testing/selftests/bpf/progs/bind4_prog.c b/tools/testing/selftests/bpf/progs/bind4_prog.c
index c6520f21f5f5..4479ac27b1d3 100644
--- a/tools/testing/selftests/bpf/progs/bind4_prog.c
+++ b/tools/testing/selftests/bpf/progs/bind4_prog.c

@@ -29,18 +29,47 @@ static __inline int bind_to_device(struct bpf_sock_addr *ctx)
  	char veth2[IFNAMSIZ] = "test_sock_addr2";
  	char missing[IFNAMSIZ] = "nonexistent_dev";
  	char del_bind[IFNAMSIZ] = "";
+	int veth1_idx, veth2_idx;
  
  	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&veth1, sizeof(veth1)))
+			   &veth1, sizeof(veth1)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth1_idx, sizeof(veth1_idx)) || !veth1_idx)
  		return 1;
  	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&veth2, sizeof(veth2)))
+			   &veth2, sizeof(veth2)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth2_idx, sizeof(veth2_idx)) || !veth2_idx ||
+	    veth1_idx == veth2_idx)
  		return 1;
  	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&missing, sizeof(missing)) != -ENODEV)
+			   &missing, sizeof(missing)) != -ENODEV)
+		return 1;
+	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth1_idx, sizeof(veth1_idx)))
  		return 1;
  	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&del_bind, sizeof(del_bind)))
+			   &del_bind, sizeof(del_bind)))
+		return 1;
+
+	return 0;
+}
+
+static __inline int misc_opts(struct bpf_sock_addr *ctx, int opt)
+{
+	int old, tmp, new = 0xeb9f;
+
+	if (bpf_getsockopt(ctx, SOL_SOCKET, opt, &old, sizeof(old)) ||
+	    old == new)
+		return 1;

Here, we assume old never equals to new. it would be good to add
a comment to explicitly state this is true. Maybe in the future
somebody will try to add more misc_opts which might have conflict
here.

Alternatively, you could pass in "new" values
from user space with global variables for each option,
but that may be an overkill.

quoted hunk

+	if (bpf_setsockopt(ctx, SOL_SOCKET, opt, &new, sizeof(new)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, opt, &tmp, sizeof(tmp)) ||
+	    tmp != new)
+		return 1;
+	if (bpf_setsockopt(ctx, SOL_SOCKET, opt, &old, sizeof(old)))
  		return 1;
  
  	return 0;

@@ -93,6 +122,10 @@ int bind_v4_prog(struct bpf_sock_addr *ctx)
  	if (bind_to_device(ctx))
  		return 0;
  
+	/* Test for misc socket options. */
+	if (misc_opts(ctx, SO_MARK) || misc_opts(ctx, SO_PRIORITY))
+		return 0;
+
  	ctx->user_ip4 = bpf_htonl(SERV4_REWRITE_IP);
  	ctx->user_port = bpf_htons(SERV4_REWRITE_PORT);

diff --git a/tools/testing/selftests/bpf/progs/bind6_prog.c b/tools/testing/selftests/bpf/progs/bind6_prog.c
index 4358e44dcf47..1b4142fcdd4b 100644
--- a/tools/testing/selftests/bpf/progs/bind6_prog.c
+++ b/tools/testing/selftests/bpf/progs/bind6_prog.c

@@ -35,18 +35,47 @@ static __inline int bind_to_device(struct bpf_sock_addr *ctx)
  	char veth2[IFNAMSIZ] = "test_sock_addr2";
  	char missing[IFNAMSIZ] = "nonexistent_dev";
  	char del_bind[IFNAMSIZ] = "";
+	int veth1_idx, veth2_idx;
  
  	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&veth1, sizeof(veth1)))
+			   &veth1, sizeof(veth1)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth1_idx, sizeof(veth1_idx)) || !veth1_idx)
  		return 1;
  	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&veth2, sizeof(veth2)))
+			   &veth2, sizeof(veth2)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth2_idx, sizeof(veth2_idx)) || !veth2_idx ||
+	    veth1_idx == veth2_idx)
  		return 1;
  	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&missing, sizeof(missing)) != -ENODEV)
+			   &missing, sizeof(missing)) != -ENODEV)
+		return 1;
+	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+			   &veth1_idx, sizeof(veth1_idx)))
  		return 1;
  	if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-				&del_bind, sizeof(del_bind)))
+			   &del_bind, sizeof(del_bind)))
+		return 1;
+
+	return 0;
+}
+
+static __inline int misc_opts(struct bpf_sock_addr *ctx, int opt)
+{
+	int old, tmp, new = 0xeb9f;
+
+	if (bpf_getsockopt(ctx, SOL_SOCKET, opt, &old, sizeof(old)) ||
+	    old == new)
+		return 1;
+	if (bpf_setsockopt(ctx, SOL_SOCKET, opt, &new, sizeof(new)))
+		return 1;
+	if (bpf_getsockopt(ctx, SOL_SOCKET, opt, &tmp, sizeof(tmp)) ||
+	    tmp != new)
+		return 1;
+	if (bpf_setsockopt(ctx, SOL_SOCKET, opt, &old, sizeof(old)))
  		return 1;
  
  	return 0;

@@ -107,6 +136,10 @@ int bind_v6_prog(struct bpf_sock_addr *ctx)
  	if (bind_to_device(ctx))
  		return 0;
  
+	/* Test for misc socket options. */
+	if (misc_opts(ctx, SO_MARK) || misc_opts(ctx, SO_PRIORITY))
+		return 0;
+
  	ctx->user_ip6[0] = bpf_htonl(SERV6_REWRITE_IP_0);
  	ctx->user_ip6[1] = bpf_htonl(SERV6_REWRITE_IP_1);
  	ctx->user_ip6[2] = bpf_htonl(SERV6_REWRITE_IP_2);

Re: [PATCH bpf-next 2/2] bpf: extend bind v4/v6 selftests for mark/prio/bindtoifindex

From: Daniel Borkmann <daniel@iogearbox.net>
Date: 2021-01-11 21:02:27

On 1/11/21 9:15 PM, Yonghong Song wrote:

On 1/11/21 8:17 AM, Daniel Borkmann wrote:

quoted

Extend existing cgroup bind4/bind6 tests to add coverage for setting and
retrieving SO_MARK, SO_PRIORITY and SO_BINDTOIFINDEX at the bind hook.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Ack with a minor comments below.

Acked-by: Yonghong Song <redacted>

quoted

---
  .../testing/selftests/bpf/progs/bind4_prog.c  | 41 +++++++++++++++++--
  .../testing/selftests/bpf/progs/bind6_prog.c  | 41 +++++++++++++++++--
  2 files changed, 74 insertions(+), 8 deletions(-)

diff --git a/tools/testing/selftests/bpf/progs/bind4_prog.c b/tools/testing/selftests/bpf/progs/bind4_prog.c
index c6520f21f5f5..4479ac27b1d3 100644
--- a/tools/testing/selftests/bpf/progs/bind4_prog.c
+++ b/tools/testing/selftests/bpf/progs/bind4_prog.c

@@ -29,18 +29,47 @@ static __inline int bind_to_device(struct bpf_sock_addr *ctx)

      char veth2[IFNAMSIZ] = "test_sock_addr2";
      char missing[IFNAMSIZ] = "nonexistent_dev";
      char del_bind[IFNAMSIZ] = "";
+    int veth1_idx, veth2_idx;
      if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-                &veth1, sizeof(veth1)))
+               &veth1, sizeof(veth1)))
+        return 1;
+    if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+               &veth1_idx, sizeof(veth1_idx)) || !veth1_idx)
          return 1;
      if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-                &veth2, sizeof(veth2)))
+               &veth2, sizeof(veth2)))
+        return 1;
+    if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+               &veth2_idx, sizeof(veth2_idx)) || !veth2_idx ||
+        veth1_idx == veth2_idx)
          return 1;
      if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-                &missing, sizeof(missing)) != -ENODEV)
+               &missing, sizeof(missing)) != -ENODEV)
+        return 1;
+    if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX,
+               &veth1_idx, sizeof(veth1_idx)))
          return 1;
      if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE,
-                &del_bind, sizeof(del_bind)))
+               &del_bind, sizeof(del_bind)))
+        return 1;
+
+    return 0;
+}
+
+static __inline int misc_opts(struct bpf_sock_addr *ctx, int opt)
+{
+    int old, tmp, new = 0xeb9f;
+
+    if (bpf_getsockopt(ctx, SOL_SOCKET, opt, &old, sizeof(old)) ||
+        old == new)
+        return 1;

Here, we assume old never equals to new. it would be good to add
a comment to explicitly state this is true. Maybe in the future
somebody will try to add more misc_opts which might have conflict
here.

I thought it's obvious, but yes I can add a comment.

Alternatively, you could pass in "new" values
from user space with global variables for each option,
but that may be an overkill.

Agree, that's overkill.

Thanks,
Daniel

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help