Skip to contentSkip to conversation

Add cgroup sockaddr hooks for unix sockets

PATCH Superseded Daan De Meyer 5 participants 18 messages Aug 31, 2023 – Sep 5, 2023

Subsystems (13)

v4 of 11 Superseded Latest is v11 →

lore
Share & tools
lore
#0 lore raw

v4 cover Add cgroup sockaddr hooks for unix sockets

Changes since v3:

* Renamed bpf_sock_addr_set_addr() to bpf_sock_addr_set_unix_addr() and made it only operate on AF_UNIX sockaddrs. This is because for the other families, users usually want to configure more than just the address so a generic interface will not fit the bill here. e.g. for AF_INET and AF_INET6, users would generally also want to be able to configure the port which the current interface doesn't support. So we expose an AF_UNIX specific function instead.
* Made the tests in the new sock addr tests more generic (similar to test_sock_addr.c), this should make it easier to migrate the other sock addr tests in the future.
* Removed the new kfunc hook and attached to BTF_KFUNC_HOOK_COMMON instead
* Set uaddrlen to 0 when the family is AF_UNSPEC
* Pass in the addrlen to the hook from IPv6 code
* Fixed mount directory mkdir() to ignore EEXIST

Changes since v2:

* Configuring the sock addr is now done via a new kfunc bpf_sock_addr_set()
* The addrlen is exposed as u32 in bpf_sock_addr_kern
* Selftests are updated to use the new kfunc
* Selftests are now added as a new sock_addr test in prog_tests/
* Added BTF_KFUNC_HOOK_SOCK_ADDR for BPF_PROG_TYPE_CGROUP_SOCK_ADDR
* __cgroup_bpf_run_filter_sock_addr() now returns the modified addrlen

Changes since v1:

* Split into multiple patches instead of one single patch
* Added unix support for all socket address hooks instead of only connect()
* Switched approach to expose the socket address length to the bpf hook instead of recalculating the socket address length in kernelspace to properly support abstract unix socket addresses
* Modified socket address hook tests to calculate the socket address length once and pass it around everywhere instead of recalculating the actual unix socket address length on demand.
* Added some missing section name tests for getpeername()/getsockname()

This patch series extends the cgroup sockaddr hooks to include support for unix sockets. To add support for unix sockets, struct bpf_sock_addr is extended to expose the unix socket path (sun_path) and the socket address length to the bpf program. For unix sockets, the address length is writable, for the other socket address hook types, the address length is only readable.

I intend to use these new hooks in systemd to reimplement the LogNamespace= feature, which allows running multiple instances of systemd-journald to process the logs of different services. systemd-journald also processes syslog messages, so currently, using log namespaces means all services running in the same log namespace have to live in the same private mount namespace so that systemd can mount the journal namespace's associated syslog socket over /dev/log to properly direct syslog messages from all services running in that log namespace to the correct systemd-journald instance. We want to relax this requirement so that processes running in disjoint mount namespaces can still run in the same log namespace. To achieve this, we can use these new hooks to rewrite the socket address of any connect(), sendto(), ... syscalls to /dev/log to the socket address of the journal namespace's syslog socket instead, which will transparently do the redirection without requiring use of a mount namespace and mounting over /dev/log.

Aside from the above usecase, these hooks can more generally be used to transparently redirect unix sockets to different addresses as required by services.

Daan De Meyer (9):
selftests/bpf: Add missing section name tests for
getpeername/getsockname
bpf: Propagate modified uaddrlen from cgroup sockaddr programs
bpf: Add bpf_sock_addr_set_unix_addr() to allow writing unix sockaddr
from bpf
bpf: Implement cgroup sockaddr hooks for unix sockets
libbpf: Add support for cgroup unix socket address hooks
bpftool: Add support for cgroup unix socket address hooks documentation/bpf: Document cgroup unix socket address hooks selftests/bpf: Make sure mount directory exists selftests/bpf: Add tests for cgroup unix socket address hooks

 Documentation/bpf/libbpf/program_types.rst    |  12 +
 include/linux/bpf-cgroup-defs.h               |   6 +
 include/linux/bpf-cgroup.h                    | 102 ++--
 include/linux/filter.h                        |   1 +
 include/uapi/linux/bpf.h                      |  14 +-
 kernel/bpf/btf.c                              |   1 +
 kernel/bpf/cgroup.c                           |  33 +-
 kernel/bpf/syscall.c                          |  18 +
 kernel/bpf/verifier.c                         |   7 +-
 net/core/filter.c                             |  50 +-
 net/ipv4/af_inet.c                            |   6 +-
 net/ipv4/ping.c                               |   2 +-
 net/ipv4/tcp_ipv4.c                           |   2 +-
 net/ipv4/udp.c                                |   8 +-
 net/ipv6/af_inet6.c                           |   9 +-
 net/ipv6/ping.c                               |   2 +-
 net/ipv6/tcp_ipv6.c                           |   2 +-
 net/ipv6/udp.c                                |   6 +-
 net/unix/af_unix.c                            |  90 +++-
 .../bpftool/Documentation/bpftool-cgroup.rst  |  23 +-
 .../bpftool/Documentation/bpftool-prog.rst    |  10 +-
 tools/bpf/bpftool/bash-completion/bpftool     |  14 +-
 tools/bpf/bpftool/cgroup.c                    |  17 +-
 tools/bpf/bpftool/prog.c                      |   9 +-
 tools/include/uapi/linux/bpf.h                |  14 +-
 tools/lib/bpf/libbpf.c                        |  12 +
 tools/testing/selftests/bpf/bpf_kfuncs.h      |  14 +
 tools/testing/selftests/bpf/cgroup_helpers.c  |   5 +
 tools/testing/selftests/bpf/network_helpers.c |  34 ++
 tools/testing/selftests/bpf/network_helpers.h |   1 +
 .../selftests/bpf/prog_tests/section_names.c  |  50 ++
 .../selftests/bpf/prog_tests/sock_addr.c      | 461 ++++++++++++++++++
 .../testing/selftests/bpf/progs/bindun_prog.c |  39 ++
 .../selftests/bpf/progs/connectun_prog.c      |  40 ++
 .../selftests/bpf/progs/recvmsgun_prog.c      |  39 ++
 .../selftests/bpf/progs/sendmsgun_prog.c      |  40 ++
 36 files changed, 1087 insertions(+), 106 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/sock_addr.c
 create mode 100644 tools/testing/selftests/bpf/progs/bindun_prog.c
 create mode 100644 tools/testing/selftests/bpf/progs/connectun_prog.c
 create mode 100644 tools/testing/selftests/bpf/progs/recvmsgun_prog.c
 create mode 100644 tools/testing/selftests/bpf/progs/sendmsgun_prog.c

--
2.41.0

#1 lore raw

PATCH v3 1/9 selftests/bpf: Add missing section name tests for getpeername/getsockname

These were missed when these hooks were first added so add them now instead to make sure every sockaddr hook has a matching section name test.

---
 .../selftests/bpf/prog_tests/section_names.c  | 20 +++++++++++++++++++
 1 file changed, 20 insertions(+)
tools/testing/selftests/bpf/prog_tests/section_names.c +20 -0
index 8b571890c57e..fc5248e94a01 100644
@@ -158,6 +158,26 @@ static struct sec_name_test tests[] = {
 		{0, BPF_PROG_TYPE_CGROUP_SOCKOPT, BPF_CGROUP_SETSOCKOPT},
 		{0, BPF_CGROUP_SETSOCKOPT},
 	},
+	{
+		"cgroup/getpeername4",
+		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_INET4_GETPEERNAME},
+		{0, BPF_CGROUP_INET4_GETPEERNAME},
+	},
+	{
+		"cgroup/getpeername6",
+		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_INET6_GETPEERNAME},
+		{0, BPF_CGROUP_INET6_GETPEERNAME},
+	},
+	{
+		"cgroup/getsockname4",
+		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_INET4_GETSOCKNAME},
+		{0, BPF_CGROUP_INET4_GETSOCKNAME},
+	},
+	{
+		"cgroup/getsockname6",
+		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_INET6_GETSOCKNAME},
+		{0, BPF_CGROUP_INET6_GETSOCKNAME},
+	},
 };
 
 static void test_prog_type_by_name(const struct sec_name_test *test)
Signature · 1 line
2.41.0
#2 lore raw

PATCH v3 2/9 bpf: Propagate modified uaddrlen from cgroup sockaddr programs

In reply to Daan De Meyer

As prep for adding unix socket support to the cgroup sockaddr hooks, let's propagate the sockaddr length back to the caller after running a bpf cgroup sockaddr hook program. While not important for AF_INET or AF_INET6, the sockaddr length is important when working with AF_UNIX sockaddrs as the size of the sockaddr cannot be determined just from the address family or the sockaddr's contents.

__cgroup_bpf_run_filter_sock_addr() is modified to take the uaddrlen as an input/output argument. After running the program, the modified sockaddr length is stored in the uaddrlen pointer. If no uaddrlen pointer is provided, we determine the uaddrlen based on the socket family. For the existing AF_INET and AF_INET6 use cases, we don't pass in the address length explicitly and just determine it based on the passed in socket family.

---
 include/linux/bpf-cgroup.h | 73 +++++++++++++++++++-------------------
 include/linux/filter.h     |  1 +
 kernel/bpf/cgroup.c        | 20 +++++++++--
 net/ipv4/af_inet.c         |  6 ++--
 net/ipv4/ping.c            |  2 +-
 net/ipv4/tcp_ipv4.c        |  2 +-
 net/ipv4/udp.c             |  8 +++--
 net/ipv6/af_inet6.c        |  9 ++---
 net/ipv6/ping.c            |  2 +-
 net/ipv6/tcp_ipv6.c        |  2 +-
 net/ipv6/udp.c             |  6 ++--
 11 files changed, 76 insertions(+), 55 deletions(-)
include/linux/bpf-cgroup.h +37 -36
index 8506690dbb9c..31561e789715 100644
@@ -120,6 +120,7 @@ int __cgroup_bpf_run_filter_sk(struct sock *sk,
 
 int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
 				      struct sockaddr *uaddr,
+				      int *uaddrlen,
 				      enum cgroup_bpf_attach_type atype,
 				      void *t_ctx,
 				      u32 *flags);
@@ -230,22 +231,22 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
 #define BPF_CGROUP_RUN_PROG_INET6_POST_BIND(sk)				       \
 	BPF_CGROUP_RUN_SK_PROG(sk, CGROUP_INET6_POST_BIND)
 
-#define BPF_CGROUP_RUN_SA_PROG(sk, uaddr, atype)				       \
+#define BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, atype)		       \
 ({									       \
 	int __ret = 0;							       \
 	if (cgroup_bpf_enabled(atype))					       \
-		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, atype,     \
-							  NULL, NULL);	       \
+		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, uaddrlen, \
+							  atype, NULL, NULL);  \
 	__ret;								       \
 })
 
-#define BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, atype, t_ctx)		       \
+#define BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, atype, t_ctx)	       \
 ({									       \
 	int __ret = 0;							       \
 	if (cgroup_bpf_enabled(atype))	{				       \
 		lock_sock(sk);						       \
-		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, atype,     \
-							  t_ctx, NULL);	       \
+		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, uaddrlen, \
+							  atype, t_ctx, NULL); \
 		release_sock(sk);					       \
 	}								       \
 	__ret;								       \
@@ -256,14 +257,14 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
  * (at bit position 0) is to indicate CAP_NET_BIND_SERVICE capability check
  * should be bypassed (BPF_RET_BIND_NO_CAP_NET_BIND_SERVICE).
  */
-#define BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, atype, bind_flags)	       \
+#define BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, uaddrlen, atype, bind_flags) \
 ({									       \
 	u32 __flags = 0;						       \
 	int __ret = 0;							       \
 	if (cgroup_bpf_enabled(atype))	{				       \
 		lock_sock(sk);						       \
-		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, atype,     \
-							  NULL, &__flags);     \
+		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, uaddrlen, \
+							  atype, NULL, &__flags); \
 		release_sock(sk);					       \
 		if (__flags & BPF_RET_BIND_NO_CAP_NET_BIND_SERVICE)	       \
 			*bind_flags |= BIND_NO_CAP_NET_BIND_SERVICE;	       \
@@ -276,29 +277,29 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
 	  cgroup_bpf_enabled(CGROUP_INET6_CONNECT)) &&		       \
 	 (sk)->sk_prot->pre_connect)
 
-#define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr)			       \
-	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, CGROUP_INET4_CONNECT)
+#define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr, uaddrlen)			\
+	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, CGROUP_INET4_CONNECT)
 
-#define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr)			       \
-	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, CGROUP_INET6_CONNECT)
+#define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr, uaddrlen)			\
+	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, CGROUP_INET6_CONNECT)
 
-#define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr)		       \
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_INET4_CONNECT, NULL)
+#define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_INET4_CONNECT, NULL)
 
-#define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr)		       \
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_INET6_CONNECT, NULL)
+#define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_INET6_CONNECT, NULL)
 
-#define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, t_ctx)		       \
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_UDP4_SENDMSG, t_ctx)
+#define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx)	\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP4_SENDMSG, t_ctx)
 
-#define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, t_ctx)		       \
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_UDP6_SENDMSG, t_ctx)
+#define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx)	\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP6_SENDMSG, t_ctx)
 
-#define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr)			\
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_UDP4_RECVMSG, NULL)
+#define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP4_RECVMSG, NULL)
 
-#define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr)			\
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_UDP6_RECVMSG, NULL)
+#define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP6_RECVMSG, NULL)
 
 /* The SOCK_OPS"_SK" macro should be used when sock_ops->sk is not a
  * fullsock and its parent fullsock cannot be traced by
@@ -477,24 +478,24 @@ static inline int bpf_percpu_cgroup_storage_update(struct bpf_map *map,
 }
 
 #define cgroup_bpf_enabled(atype) (0)
-#define BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, atype, t_ctx) ({ 0; })
-#define BPF_CGROUP_RUN_SA_PROG(sk, uaddr, atype) ({ 0; })
+#define BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, atype, t_ctx) ({ 0; })
+#define BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, atype) ({ 0; })
 #define BPF_CGROUP_PRE_CONNECT_ENABLED(sk) (0)
 #define BPF_CGROUP_RUN_PROG_INET_INGRESS(sk,skb) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET_EGRESS(sk,skb) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET_SOCK(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET_SOCK_RELEASE(sk) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, atype, flags) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, uaddrlen, atype, flags) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET4_POST_BIND(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET6_POST_BIND(sk) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, t_ctx) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, t_ctx) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_SOCK_OPS(sock_ops) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_DEVICE_CGROUP(atype, major, minor, access) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_SYSCTL(head,table,write,buf,count,pos) ({ 0; })
include/linux/filter.h +1 -0
index 761af6b3cf2b..77db4263d68d 100644
@@ -1285,6 +1285,7 @@ struct bpf_sock_addr_kern {
 	 */
 	u64 tmp_reg;
 	void *t_ctx;	/* Attach type specific context. */
+	u32 uaddrlen;
 };
 
 struct bpf_sock_ops_kern {
kernel/bpf/cgroup.c +17 -3
index 5b2741aa0d9b..534b6c7f5659 100644
@@ -1449,6 +1449,7 @@ EXPORT_SYMBOL(__cgroup_bpf_run_filter_sk);
  *                                       provided by user sockaddr
  * @sk: sock struct that will use sockaddr
  * @uaddr: sockaddr struct provided by user
+ * @uaddrlen: Pointer to the size of the sockaddr struct provided by user
  * @type: The type of program to be executed
  * @t_ctx: Pointer to attach type specific context
  * @flags: Pointer to u32 which contains higher bits of BPF program
@@ -1461,6 +1462,7 @@ EXPORT_SYMBOL(__cgroup_bpf_run_filter_sk);
  */
 int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
 				      struct sockaddr *uaddr,
+				      int *uaddrlen,
 				      enum cgroup_bpf_attach_type atype,
 				      void *t_ctx,
 				      u32 *flags)
@@ -1472,6 +1474,7 @@ int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
 	};
 	struct sockaddr_storage unspec;
 	struct cgroup *cgrp;
+	int ret;
 
 	/* Check socket family since not all sockets represent network
 	 * endpoint (e.g. AF_UNIX).
@@ -1482,11 +1485,22 @@ int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
 	if (!ctx.uaddr) {
 		memset(&unspec, 0, sizeof(unspec));
 		ctx.uaddr = (struct sockaddr *)&unspec;
-	}
+		ctx.uaddrlen = 0;
+	} else if (uaddrlen)
+		ctx.uaddrlen = *uaddrlen;
+	else if (sk->sk_family == AF_INET)
+		ctx.uaddrlen = sizeof(struct sockaddr_in);
+	else if (sk->sk_family == AF_INET6)
+		ctx.uaddrlen = sizeof(struct sockaddr_in6);
 
 	cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
-	return bpf_prog_run_array_cg(&cgrp->bpf, atype, &ctx, bpf_prog_run,
-				     0, flags);
+	ret = bpf_prog_run_array_cg(&cgrp->bpf, atype, &ctx, bpf_prog_run,
+				    0, flags);
+
+	if (!ret && uaddrlen)
+		*uaddrlen = ctx.uaddrlen;
+
+	return ret;
 }
 EXPORT_SYMBOL(__cgroup_bpf_run_filter_sock_addr);
 
net/ipv4/af_inet.c +3 -3
index e07ee60625d9..ac11d485e3ab 100644
@@ -452,7 +452,7 @@ int inet_bind_sk(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 	/* BPF prog is run before any checks are done so that if the prog
 	 * changes context in a wrong way it will be caught.
 	 */
-	err = BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr,
+	err = BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, NULL,
 						 CGROUP_INET4_BIND, &flags);
 	if (err)
 		return err;
@@ -800,7 +800,7 @@ int inet_getname(struct socket *sock, struct sockaddr *uaddr,
 		}
 		sin->sin_port = inet->inet_dport;
 		sin->sin_addr.s_addr = inet->inet_daddr;
-		BPF_CGROUP_RUN_SA_PROG(sk, (struct sockaddr *)sin,
+		BPF_CGROUP_RUN_SA_PROG(sk, (struct sockaddr *)sin, NULL,
 				       CGROUP_INET4_GETPEERNAME);
 	} else {
 		__be32 addr = inet->inet_rcv_saddr;
@@ -808,7 +808,7 @@ int inet_getname(struct socket *sock, struct sockaddr *uaddr,
 			addr = inet->inet_saddr;
 		sin->sin_port = inet->inet_sport;
 		sin->sin_addr.s_addr = addr;
-		BPF_CGROUP_RUN_SA_PROG(sk, (struct sockaddr *)sin,
+		BPF_CGROUP_RUN_SA_PROG(sk, (struct sockaddr *)sin, NULL,
 				       CGROUP_INET4_GETSOCKNAME);
 	}
 	release_sock(sk);
net/ipv4/ping.c +1 -1
index 75e0aee35eb7..0951837d5138 100644
@@ -301,7 +301,7 @@ static int ping_pre_connect(struct sock *sk, struct sockaddr *uaddr,
 	if (addr_len < sizeof(struct sockaddr_in))
 		return -EINVAL;
 
-	return BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr);
+	return BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr, NULL);
 }
 
 /* Checks the bind address and possibly modifies sk->sk_bound_dev_if. */
net/ipv4/tcp_ipv4.c +1 -1
index 2a662d5f3072..50f34226252a 100644
@@ -194,7 +194,7 @@ static int tcp_v4_pre_connect(struct sock *sk, struct sockaddr *uaddr,
 
 	sock_owned_by_me(sk);
 
-	return BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr);
+	return BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr, NULL);
 }
 
 /* This will initiate an outgoing connection. */
net/ipv4/udp.c +5 -3
index 0794a2c46a56..9ecde90d59b7 100644
@@ -1143,7 +1143,8 @@ int udp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 
 	if (cgroup_bpf_enabled(CGROUP_UDP4_SENDMSG) && !connected) {
 		err = BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk,
-					    (struct sockaddr *)usin, &ipc.addr);
+					    (struct sockaddr *)usin, NULL,
+					    &ipc.addr);
 		if (err)
 			goto out_free;
 		if (usin) {
@@ -1865,7 +1866,8 @@ int udp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int flags,
 		*addr_len = sizeof(*sin);
 
 		BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk,
-						      (struct sockaddr *)sin);
+						      (struct sockaddr *)sin,
+						      NULL);
 	}
 
 	if (udp_sk(sk)->gro_enabled)
@@ -1904,7 +1906,7 @@ int udp_pre_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 	if (addr_len < sizeof(struct sockaddr_in))
 		return -EINVAL;
 
-	return BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr);
+	return BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr, NULL);
 }
 EXPORT_SYMBOL(udp_pre_connect);
 
net/ipv6/af_inet6.c +5 -4
index 368824fe9719..773e4abe55b7 100644
@@ -453,7 +453,7 @@ int inet6_bind_sk(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 	/* BPF prog is run before any checks are done so that if the prog
 	 * changes context in a wrong way it will be caught.
 	 */
-	err = BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr,
+	err = BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, &addr_len,
 						 CGROUP_INET6_BIND, &flags);
 	if (err)
 		return err;
@@ -519,6 +519,7 @@ int inet6_getname(struct socket *sock, struct sockaddr *uaddr,
 		  int peer)
 {
 	struct sockaddr_in6 *sin = (struct sockaddr_in6 *)uaddr;
+	int sin_addr_len = sizeof(*sin);
 	struct sock *sk = sock->sk;
 	struct inet_sock *inet = inet_sk(sk);
 	struct ipv6_pinfo *np = inet6_sk(sk);
@@ -538,7 +539,7 @@ int inet6_getname(struct socket *sock, struct sockaddr *uaddr,
 		sin->sin6_addr = sk->sk_v6_daddr;
 		if (np->sndflow)
 			sin->sin6_flowinfo = np->flow_label;
-		BPF_CGROUP_RUN_SA_PROG(sk, (struct sockaddr *)sin,
+		BPF_CGROUP_RUN_SA_PROG(sk, (struct sockaddr *)sin, &sin_addr_len,
 				       CGROUP_INET6_GETPEERNAME);
 	} else {
 		if (ipv6_addr_any(&sk->sk_v6_rcv_saddr))
@@ -546,13 +547,13 @@ int inet6_getname(struct socket *sock, struct sockaddr *uaddr,
 		else
 			sin->sin6_addr = sk->sk_v6_rcv_saddr;
 		sin->sin6_port = inet->inet_sport;
-		BPF_CGROUP_RUN_SA_PROG(sk, (struct sockaddr *)sin,
+		BPF_CGROUP_RUN_SA_PROG(sk, (struct sockaddr *)sin, &sin_addr_len,
 				       CGROUP_INET6_GETSOCKNAME);
 	}
 	sin->sin6_scope_id = ipv6_iface_scope_id(&sin->sin6_addr,
 						 sk->sk_bound_dev_if);
 	release_sock(sk);
-	return sizeof(*sin);
+	return sin_addr_len;
 }
 EXPORT_SYMBOL(inet6_getname);
 
net/ipv6/ping.c +1 -1
index 1b2772834972..6c594c501cd0 100644
@@ -56,7 +56,7 @@ static int ping_v6_pre_connect(struct sock *sk, struct sockaddr *uaddr,
 	if (addr_len < SIN6_LEN_RFC2133)
 		return -EINVAL;
 
-	return BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr);
+	return BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr, &addr_len);
 }
 
 static int ping_v6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
net/ipv6/tcp_ipv6.c +1 -1
index 3a88545a265d..925f92d4c2d6 100644
@@ -135,7 +135,7 @@ static int tcp_v6_pre_connect(struct sock *sk, struct sockaddr *uaddr,
 
 	sock_owned_by_me(sk);
 
-	return BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr);
+	return BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr, &addr_len);
 }
 
 static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr,
net/ipv6/udp.c +4 -2
index ebc6ae47cfea..26387b71805c 100644
@@ -410,7 +410,8 @@ int udpv6_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 		*addr_len = sizeof(*sin6);
 
 		BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk,
-						      (struct sockaddr *)sin6);
+						      (struct sockaddr *)sin6,
+						      addr_len);
 	}
 
 	if (udp_sk(sk)->gro_enabled)
@@ -1155,7 +1156,7 @@ static int udpv6_pre_connect(struct sock *sk, struct sockaddr *uaddr,
 	if (addr_len < SIN6_LEN_RFC2133)
 		return -EINVAL;
 
-	return BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr);
+	return BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr, &addr_len);
 }
 
 /**
@@ -1508,6 +1509,7 @@ int udpv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 	if (cgroup_bpf_enabled(CGROUP_UDP6_SENDMSG) && !connected) {
 		err = BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk,
 					   (struct sockaddr *)sin6,
+					   &addr_len,
 					   &fl6->saddr);
 		if (err)
 			goto out_no_dst;
Signature · 1 line
2.41.0
#3 lore raw

Re: [PATCH bpf-next v3 2/9] bpf: Propagate modified uaddrlen from cgroup sockaddr programs

Quoted 13 lines · Daan De Meyer

As prep for adding unix socket support to the cgroup sockaddr hooks, let's propagate the sockaddr length back to the caller after running a bpf cgroup sockaddr hook program. While not important for AF_INET or AF_INET6, the sockaddr length is important when working with AF_UNIX sockaddrs as the size of the sockaddr cannot be determined just from the address family or the sockaddr's contents.

__cgroup_bpf_run_filter_sock_addr() is modified to take the uaddrlen as an input/output argument. After running the program, the modified sockaddr length is stored in the uaddrlen pointer. If no uaddrlen pointer is provided, we determine the uaddrlen based on the socket family. For the

Quoted 3 lines

existing AF_INET and AF_INET6 use cases, we don't pass in the address length explicitly and just determine it based on the passed in socket family.

The description on the inet address length needs an update, at least the AF_INET6 one.

Quoted 183 lines
include/linux/bpf-cgroup.h +37 -36
index 8506690dbb9c..31561e789715 100644
@@ -120,6 +120,7 @@ int __cgroup_bpf_run_filter_sk(struct sock *sk,
 
  int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
  				      struct sockaddr *uaddr,
+				      int *uaddrlen,
  				      enum cgroup_bpf_attach_type atype,
  				      void *t_ctx,
  				      u32 *flags);
@@ -230,22 +231,22 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
  #define BPF_CGROUP_RUN_PROG_INET6_POST_BIND(sk)				       \
  	BPF_CGROUP_RUN_SK_PROG(sk, CGROUP_INET6_POST_BIND)
 
-#define BPF_CGROUP_RUN_SA_PROG(sk, uaddr, atype)				       \
+#define BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, atype)		       \
  ({									       \
  	int __ret = 0;							       \
  	if (cgroup_bpf_enabled(atype))					       \
-		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, atype,     \
-							  NULL, NULL);	       \
+		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, uaddrlen, \
+							  atype, NULL, NULL);  \
  	__ret;								       \
  })
 
-#define BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, atype, t_ctx)		       \
+#define BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, atype, t_ctx)	       \
  ({									       \
  	int __ret = 0;							       \
  	if (cgroup_bpf_enabled(atype))	{				       \
  		lock_sock(sk);						       \
-		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, atype,     \
-							  t_ctx, NULL);	       \
+		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, uaddrlen, \
+							  atype, t_ctx, NULL); \
  		release_sock(sk);					       \
  	}								       \
  	__ret;								       \
@@ -256,14 +257,14 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
   * (at bit position 0) is to indicate CAP_NET_BIND_SERVICE capability check
   * should be bypassed (BPF_RET_BIND_NO_CAP_NET_BIND_SERVICE).
   */
-#define BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, atype, bind_flags)	       \
+#define BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, uaddrlen, atype, bind_flags) \
  ({									       \
  	u32 __flags = 0;						       \
  	int __ret = 0;							       \
  	if (cgroup_bpf_enabled(atype))	{				       \
  		lock_sock(sk);						       \
-		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, atype,     \
-							  NULL, &__flags);     \
+		__ret = __cgroup_bpf_run_filter_sock_addr(sk, uaddr, uaddrlen, \
+							  atype, NULL, &__flags); \
  		release_sock(sk);					       \
  		if (__flags & BPF_RET_BIND_NO_CAP_NET_BIND_SERVICE)	       \
  			*bind_flags |= BIND_NO_CAP_NET_BIND_SERVICE;	       \
@@ -276,29 +277,29 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
  	  cgroup_bpf_enabled(CGROUP_INET6_CONNECT)) &&		       \
  	 (sk)->sk_prot->pre_connect)
 
-#define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr)			       \
-	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, CGROUP_INET4_CONNECT)
+#define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr, uaddrlen)			\
+	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, CGROUP_INET4_CONNECT)
 
-#define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr)			       \
-	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, CGROUP_INET6_CONNECT)
+#define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr, uaddrlen)			\
+	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, CGROUP_INET6_CONNECT)
 
-#define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr)		       \
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_INET4_CONNECT, NULL)
+#define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_INET4_CONNECT, NULL)
 
-#define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr)		       \
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_INET6_CONNECT, NULL)
+#define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_INET6_CONNECT, NULL)
 
-#define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, t_ctx)		       \
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_UDP4_SENDMSG, t_ctx)
+#define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx)	\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP4_SENDMSG, t_ctx)
 
-#define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, t_ctx)		       \
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_UDP6_SENDMSG, t_ctx)
+#define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx)	\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP6_SENDMSG, t_ctx)
 
-#define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr)			\
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_UDP4_RECVMSG, NULL)
+#define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP4_RECVMSG, NULL)
 
-#define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr)			\
-	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, CGROUP_UDP6_RECVMSG, NULL)
+#define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP6_RECVMSG, NULL)
 
  /* The SOCK_OPS"_SK" macro should be used when sock_ops->sk is not a
   * fullsock and its parent fullsock cannot be traced by
@@ -477,24 +478,24 @@ static inline int bpf_percpu_cgroup_storage_update(struct bpf_map *map,
  }
 
  #define cgroup_bpf_enabled(atype) (0)
-#define BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, atype, t_ctx) ({ 0; })
-#define BPF_CGROUP_RUN_SA_PROG(sk, uaddr, atype) ({ 0; })
+#define BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, atype, t_ctx) ({ 0; })
+#define BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, atype) ({ 0; })
  #define BPF_CGROUP_PRE_CONNECT_ENABLED(sk) (0)
  #define BPF_CGROUP_RUN_PROG_INET_INGRESS(sk,skb) ({ 0; })
  #define BPF_CGROUP_RUN_PROG_INET_EGRESS(sk,skb) ({ 0; })
  #define BPF_CGROUP_RUN_PROG_INET_SOCK(sk) ({ 0; })
  #define BPF_CGROUP_RUN_PROG_INET_SOCK_RELEASE(sk) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, atype, flags) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, uaddrlen, atype, flags) ({ 0; })
  #define BPF_CGROUP_RUN_PROG_INET4_POST_BIND(sk) ({ 0; })
  #define BPF_CGROUP_RUN_PROG_INET6_POST_BIND(sk) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, t_ctx) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, t_ctx) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr) ({ 0; })
-#define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr, uaddrlen) ({ 0; })
  #define BPF_CGROUP_RUN_PROG_SOCK_OPS(sock_ops) ({ 0; })
  #define BPF_CGROUP_RUN_PROG_DEVICE_CGROUP(atype, major, minor, access) ({ 0; })
  #define BPF_CGROUP_RUN_PROG_SYSCTL(head,table,write,buf,count,pos) ({ 0; })
include/linux/filter.h +1 -0
index 761af6b3cf2b..77db4263d68d 100644
@@ -1285,6 +1285,7 @@ struct bpf_sock_addr_kern {
  	 */
  	u64 tmp_reg;
  	void *t_ctx;	/* Attach type specific context. */
+	u32 uaddrlen;
  };
 
  struct bpf_sock_ops_kern {
kernel/bpf/cgroup.c +3 -0
index 5b2741aa0d9b..534b6c7f5659 100644
@@ -1449,6 +1449,7 @@ EXPORT_SYMBOL(__cgroup_bpf_run_filter_sk);
   *                                       provided by user sockaddr
   * @sk: sock struct that will use sockaddr
   * @uaddr: sockaddr struct provided by user
+ * @uaddrlen: Pointer to the size of the sockaddr struct provided by user
   * @type: The type of program to be executed
   * @t_ctx: Pointer to attach type specific context
   * @flags: Pointer to u32 which contains higher bits of BPF program
@@ -1461,6 +1462,7 @@ EXPORT_SYMBOL(__cgroup_bpf_run_filter_sk);
   */
  int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
  				      struct sockaddr *uaddr,
+				      int *uaddrlen,
  				      enum cgroup_bpf_attach_type atype,
  				      void *t_ctx,
  				      u32 *flags)
@@ -1472,6 +1474,7 @@ int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
  	};
  	struct sockaddr_storage unspec;
  	struct cgroup *cgrp;
+	int ret;
 
  	/* Check socket family since not all sockets represent network
  	 * endpoint (e.g. AF_UNIX).
@@ -1482,11 +1485,22 @@ int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
  	if (!ctx.uaddr) {

kernel/bpf/cgroup.c int __cgroup_bpf_run_filter_sock_addr(struct sock *sk, · line 1494

Quoted 10 lines
kernel/bpf/cgroup.c +7 -1
  		memset(&unspec, 0, sizeof(unspec));
  		ctx.uaddr = (struct sockaddr *)&unspec;
-	}
+		ctx.uaddrlen = 0;
+	} else if (uaddrlen)
+		ctx.uaddrlen = *uaddrlen;
+	else if (sk->sk_family == AF_INET)
+		ctx.uaddrlen = sizeof(struct sockaddr_in);
+	else if (sk->sk_family == AF_INET6)
+		ctx.uaddrlen = sizeof(struct sockaddr_in6);

I was thinking to pass addrlen whenever possible for AF_INET and AF_INET6.

If I read correctly, all AF_INET6 cases have addrlen available in this patch. May be just pass addrlen for AF_INET also, then the new sk->sk_family test here can be avoided?

#4 lore raw

PATCH v3 3/9 bpf: Add bpf_sock_addr_set_unix_addr() to allow writing unix sockaddr from bpf

In reply to Daan De Meyer

As prep for adding unix socket support to the cgroup sockaddr hooks, let's add a kfunc bpf_sock_addr_set_unix_addr() that allows modifying a sockaddr from bpf. While this is already possible for AF_INET and AF_INET6, we'll need this kfunc when we add unix socket support since modifying the address for those requires modifying both the address and the sockaddr length.

---
 kernel/bpf/btf.c  |  1 +
 net/core/filter.c | 32 +++++++++++++++++++++++++++++++-
 2 files changed, 32 insertions(+), 1 deletion(-)
kernel/bpf/btf.c +1 -0
index 249657c466dd..15c972f27574 100644
@@ -7819,6 +7819,7 @@ static int bpf_prog_type_to_kfunc_hook(enum bpf_prog_type prog_type)
 {
 	switch (prog_type) {
 	case BPF_PROG_TYPE_UNSPEC:
+	case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
 		return BTF_KFUNC_HOOK_COMMON;
 	case BPF_PROG_TYPE_XDP:
 		return BTF_KFUNC_HOOK_XDP;
net/core/filter.c +31 -1
index a094694899c9..3ed6cd33b268 100644
@@ -11752,6 +11752,25 @@ __bpf_kfunc int bpf_dynptr_from_xdp(struct xdp_buff *xdp, u64 flags,
 
 	return 0;
 }
+
+__bpf_kfunc int bpf_sock_addr_set_unix_addr(struct bpf_sock_addr_kern *sa_kern,
+					    const u8 *addr, u32 addrlen__sz)
+{
+	struct sockaddr *sa = sa_kern->uaddr;
+	struct sockaddr_un *un;
+
+	if (sa_kern->sk->sk_family != AF_UNIX)
+		return -EINVAL;
+
+	if (addrlen__sz > UNIX_PATH_MAX)
+		return -EINVAL;
+
+	un = (struct sockaddr_un *)sa;
+	memcpy(un->sun_path, addr, addrlen__sz);
+	sa_kern->uaddrlen = offsetof(struct sockaddr_un, sun_path) + addrlen__sz;
+
+	return 0;
+}
 __diag_pop();
 
 int bpf_dynptr_from_skb_rdonly(struct sk_buff *skb, u64 flags,
@@ -11776,6 +11795,10 @@ BTF_SET8_START(bpf_kfunc_check_set_xdp)
 BTF_ID_FLAGS(func, bpf_dynptr_from_xdp)
 BTF_SET8_END(bpf_kfunc_check_set_xdp)
 
+BTF_SET8_START(bpf_kfunc_check_set_sock_addr)
+BTF_ID_FLAGS(func, bpf_sock_addr_set_unix_addr)
+BTF_SET8_END(bpf_kfunc_check_set_sock_addr)
+
 static const struct btf_kfunc_id_set bpf_kfunc_set_skb = {
 	.owner = THIS_MODULE,
 	.set = &bpf_kfunc_check_set_skb,
@@ -11786,6 +11809,11 @@ static const struct btf_kfunc_id_set bpf_kfunc_set_xdp = {
 	.set = &bpf_kfunc_check_set_xdp,
 };
 
+static const struct btf_kfunc_id_set bpf_kfunc_set_sock_addr = {
+	.owner = THIS_MODULE,
+	.set = &bpf_kfunc_check_set_sock_addr,
+};
+
 static int __init bpf_kfunc_init(void)
 {
 	int ret;
@@ -11800,7 +11828,9 @@ static int __init bpf_kfunc_init(void)
 	ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_LWT_XMIT, &bpf_kfunc_set_skb);
 	ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_LWT_SEG6LOCAL, &bpf_kfunc_set_skb);
 	ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_NETFILTER, &bpf_kfunc_set_skb);
-	return ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_XDP, &bpf_kfunc_set_xdp);
+	ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_XDP, &bpf_kfunc_set_xdp);
+	return ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SOCK_ADDR,
+						&bpf_kfunc_set_sock_addr);
 }
 late_initcall(bpf_kfunc_init);
 
Signature · 1 line
2.41.0
#5 lore raw

Re: [PATCH bpf-next v3 3/9] bpf: Add bpf_sock_addr_set_unix_addr() to allow writing unix sockaddr from bpf

Quoted 19 lines · Daan De Meyer

As prep for adding unix socket support to the cgroup sockaddr hooks, let's add a kfunc bpf_sock_addr_set_unix_addr() that allows modifying a sockaddr from bpf. While this is already possible for AF_INET and AF_INET6, we'll need this kfunc when we add unix socket support since modifying the address for those requires modifying both the address and the sockaddr length.

---
 kernel/bpf/btf.c  |  1 +
 net/core/filter.c | 32 +++++++++++++++++++++++++++++++-
 2 files changed, 32 insertions(+), 1 deletion(-)
kernel/bpf/btf.c +0 -0
index 249657c466dd..15c972f27574 100644

kernel/bpf/btf.c static int bpf_prog_type_to_kfunc_hook(enum bpf_prog_type prog_type) · line 7823

Quoted 5 lines
kernel/bpf/btf.c +1 -0
@@ -7819,6 +7819,7 @@ static int bpf_prog_type_to_kfunc_hook(enum bpf_prog_type prog_type)
 {
        switch (prog_type) {
        case BPF_PROG_TYPE_UNSPEC:
+       case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
                return BTF_KFUNC_HOOK_COMMON;

hook_common means that they are shared by all prog types. See btf_kfunc_id_set_contains().

#6 lore raw

Re: [PATCH bpf-next v3 3/9] bpf: Add bpf_sock_addr_set_unix_addr() to allow writing unix sockaddr from bpf

In reply to Daan De Meyer
Quoted 34 lines · Daan De Meyer

As prep for adding unix socket support to the cgroup sockaddr hooks, let's add a kfunc bpf_sock_addr_set_unix_addr() that allows modifying a sockaddr from bpf. While this is already possible for AF_INET and AF_INET6, we'll need this kfunc when we add unix socket support since modifying the address for those requires modifying both the address and the sockaddr length.

---
  kernel/bpf/btf.c  |  1 +
  net/core/filter.c | 32 +++++++++++++++++++++++++++++++-
  2 files changed, 32 insertions(+), 1 deletion(-)
kernel/bpf/btf.c +1 -0
index 249657c466dd..15c972f27574 100644
@@ -7819,6 +7819,7 @@ static int bpf_prog_type_to_kfunc_hook(enum bpf_prog_type prog_type)
  {
  	switch (prog_type) {
  	case BPF_PROG_TYPE_UNSPEC:
+	case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
  		return BTF_KFUNC_HOOK_COMMON;
  	case BPF_PROG_TYPE_XDP:
  		return BTF_KFUNC_HOOK_XDP;
net/core/filter.c +1 -0
index a094694899c9..3ed6cd33b268 100644
@@ -11752,6 +11752,25 @@ __bpf_kfunc int bpf_dynptr_from_xdp(struct xdp_buff *xdp, u64 flags,
 
  	return 0;
  }
+

net/core/filter.c __bpf_kfunc int bpf_dynptr_from_xdp(struct xdp_buff *xdp, u64 flags, · line 11765

Quoted 10 lines
net/core/filter.c +10 -0
+__bpf_kfunc int bpf_sock_addr_set_unix_addr(struct bpf_sock_addr_kern *sa_kern,
+					    const u8 *addr, u32 addrlen__sz)
+{
+	struct sockaddr *sa = sa_kern->uaddr;
+	struct sockaddr_un *un;
+
+	if (sa_kern->sk->sk_family != AF_UNIX)
+		return -EINVAL;
+
+	if (addrlen__sz > UNIX_PATH_MAX)

Is it valid to have addrlen__sz == 0 for AF_UNIX?

Quoted 8 lines
net/core/filter.c +8 -0
+		return -EINVAL;
+
+	un = (struct sockaddr_un *)sa;
+	memcpy(un->sun_path, addr, addrlen__sz);
+	sa_kern->uaddrlen = offsetof(struct sockaddr_un, sun_path) + addrlen__sz;
+
+	return 0;
+}
#7 lore raw

PATCH v3 4/9 bpf: Implement cgroup sockaddr hooks for unix sockets

In reply to Daan De Meyer

These hooks allows intercepting bind(), connect(), getsockname(), getpeername(), sendmsg() and recvmsg() for unix sockets. The unix socket hooks get write access to the address length because the address length is not fixed when dealing with unix sockets and needs to be modified when a unix socket address is modified by the hook. Because abstract socket unix addresses start with a NUL byte, we cannot recalculate the socket address in kernelspace after running the hook by calculating the length of the unix socket path using strlen().

Write support is added for uaddrlen to allow the unix socket hooks to modify the sockaddr length from the bpf program.

This hook can be used when users want to multiplex syscall to a single unix socket to multiple different processes behind the scenes by redirecting the connect() and other syscalls to process specific sockets.

---
 include/linux/bpf-cgroup-defs.h |  6 +++
 include/linux/bpf-cgroup.h      | 29 ++++++++++-
 include/uapi/linux/bpf.h        | 14 +++--
 kernel/bpf/cgroup.c             | 13 ++++-
 kernel/bpf/syscall.c            | 18 +++++++
 kernel/bpf/verifier.c           |  7 ++-
 net/core/filter.c               | 18 ++++++-
 net/unix/af_unix.c              | 90 ++++++++++++++++++++++++++++++---
 tools/include/uapi/linux/bpf.h  | 14 +++--
 9 files changed, 187 insertions(+), 22 deletions(-)
include/linux/bpf-cgroup-defs.h +6 -0
index 7b121bd780eb..8196ccb81915 100644
@@ -26,21 +26,27 @@ enum cgroup_bpf_attach_type {
 	CGROUP_DEVICE,
 	CGROUP_INET4_BIND,
 	CGROUP_INET6_BIND,
+	CGROUP_UNIX_BIND,
 	CGROUP_INET4_CONNECT,
 	CGROUP_INET6_CONNECT,
+	CGROUP_UNIX_CONNECT,
 	CGROUP_INET4_POST_BIND,
 	CGROUP_INET6_POST_BIND,
 	CGROUP_UDP4_SENDMSG,
 	CGROUP_UDP6_SENDMSG,
+	CGROUP_UNIX_SENDMSG,
 	CGROUP_SYSCTL,
 	CGROUP_UDP4_RECVMSG,
 	CGROUP_UDP6_RECVMSG,
+	CGROUP_UNIX_RECVMSG,
 	CGROUP_GETSOCKOPT,
 	CGROUP_SETSOCKOPT,
 	CGROUP_INET4_GETPEERNAME,
 	CGROUP_INET6_GETPEERNAME,
+	CGROUP_UNIX_GETPEERNAME,
 	CGROUP_INET4_GETSOCKNAME,
 	CGROUP_INET6_GETSOCKNAME,
+	CGROUP_UNIX_GETSOCKNAME,
 	CGROUP_INET_SOCK_RELEASE,
 	CGROUP_LSM_START,
 	CGROUP_LSM_END = CGROUP_LSM_START + CGROUP_LSM_NUM - 1,
include/linux/bpf-cgroup.h +28 -1
index 31561e789715..0d572228fa62 100644
@@ -46,21 +46,27 @@ to_cgroup_bpf_attach_type(enum bpf_attach_type attach_type)
 	CGROUP_ATYPE(CGROUP_DEVICE);
 	CGROUP_ATYPE(CGROUP_INET4_BIND);
 	CGROUP_ATYPE(CGROUP_INET6_BIND);
+	CGROUP_ATYPE(CGROUP_UNIX_BIND);
 	CGROUP_ATYPE(CGROUP_INET4_CONNECT);
 	CGROUP_ATYPE(CGROUP_INET6_CONNECT);
+	CGROUP_ATYPE(CGROUP_UNIX_CONNECT);
 	CGROUP_ATYPE(CGROUP_INET4_POST_BIND);
 	CGROUP_ATYPE(CGROUP_INET6_POST_BIND);
 	CGROUP_ATYPE(CGROUP_UDP4_SENDMSG);
 	CGROUP_ATYPE(CGROUP_UDP6_SENDMSG);
+	CGROUP_ATYPE(CGROUP_UNIX_SENDMSG);
 	CGROUP_ATYPE(CGROUP_SYSCTL);
 	CGROUP_ATYPE(CGROUP_UDP4_RECVMSG);
 	CGROUP_ATYPE(CGROUP_UDP6_RECVMSG);
+	CGROUP_ATYPE(CGROUP_UNIX_RECVMSG);
 	CGROUP_ATYPE(CGROUP_GETSOCKOPT);
 	CGROUP_ATYPE(CGROUP_SETSOCKOPT);
 	CGROUP_ATYPE(CGROUP_INET4_GETPEERNAME);
 	CGROUP_ATYPE(CGROUP_INET6_GETPEERNAME);
+	CGROUP_ATYPE(CGROUP_UNIX_GETPEERNAME);
 	CGROUP_ATYPE(CGROUP_INET4_GETSOCKNAME);
 	CGROUP_ATYPE(CGROUP_INET6_GETSOCKNAME);
+	CGROUP_ATYPE(CGROUP_UNIX_GETSOCKNAME);
 	CGROUP_ATYPE(CGROUP_INET_SOCK_RELEASE);
 	default:
 		return CGROUP_BPF_ATTACH_TYPE_INVALID;
@@ -272,9 +278,13 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
 	__ret;								       \
 })
 
+#define BPF_CGROUP_RUN_PROG_UNIX_BIND_LOCK(sk, uaddr, uaddrlen)			\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UNIX_BIND, NULL)
+
 #define BPF_CGROUP_PRE_CONNECT_ENABLED(sk)				       \
 	((cgroup_bpf_enabled(CGROUP_INET4_CONNECT) ||		       \
-	  cgroup_bpf_enabled(CGROUP_INET6_CONNECT)) &&		       \
+	  cgroup_bpf_enabled(CGROUP_INET6_CONNECT) ||		       \
+	  cgroup_bpf_enabled(CGROUP_UNIX_CONNECT)) &&		       \
 	 (sk)->sk_prot->pre_connect)
 
 #define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr, uaddrlen)			\
@@ -283,24 +293,36 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
 #define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr, uaddrlen)			\
 	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, CGROUP_INET6_CONNECT)
 
+#define BPF_CGROUP_RUN_PROG_UNIX_CONNECT(sk, uaddr, uaddrlen)			\
+	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, CGROUP_UNIX_CONNECT)
+
 #define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr, uaddrlen)		\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_INET4_CONNECT, NULL)
 
 #define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr, uaddrlen)		\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_INET6_CONNECT, NULL)
 
+#define BPF_CGROUP_RUN_PROG_UNIX_CONNECT_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UNIX_CONNECT, NULL)
+
 #define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx)	\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP4_SENDMSG, t_ctx)
 
 #define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx)	\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP6_SENDMSG, t_ctx)
 
+#define BPF_CGROUP_RUN_PROG_UNIX_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx)	\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UNIX_SENDMSG, t_ctx)
+
 #define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr, uaddrlen)		\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP4_RECVMSG, NULL)
 
 #define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr, uaddrlen)		\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP6_RECVMSG, NULL)
 
+#define BPF_CGROUP_RUN_PROG_UNIX_RECVMSG_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UNIX_RECVMSG, NULL)
+
 /* The SOCK_OPS"_SK" macro should be used when sock_ops->sk is not a
  * fullsock and its parent fullsock cannot be traced by
  * sk_to_full_sk().
@@ -486,16 +508,21 @@ static inline int bpf_percpu_cgroup_storage_update(struct bpf_map *map,
 #define BPF_CGROUP_RUN_PROG_INET_SOCK(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET_SOCK_RELEASE(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, uaddrlen, atype, flags) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UNIX_BIND_LOCK(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET4_POST_BIND(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET6_POST_BIND(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UNIX_CONNECT(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UNIX_CONNECT_LOCK(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UNIX_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UNIX_RECVMSG_LOCK(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_SOCK_OPS(sock_ops) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_DEVICE_CGROUP(atype, major, minor, access) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_SYSCTL(head,table,write,buf,count,pos) ({ 0; })
include/uapi/linux/bpf.h +10 -4
index 8790b3962e4b..c51889d7e5d8 100644
@@ -1040,6 +1040,12 @@ enum bpf_attach_type {
 	BPF_TCX_INGRESS,
 	BPF_TCX_EGRESS,
 	BPF_TRACE_UPROBE_MULTI,
+	BPF_CGROUP_UNIX_BIND,
+	BPF_CGROUP_UNIX_CONNECT,
+	BPF_CGROUP_UNIX_SENDMSG,
+	BPF_CGROUP_UNIX_RECVMSG,
+	BPF_CGROUP_UNIX_GETPEERNAME,
+	BPF_CGROUP_UNIX_GETSOCKNAME,
 	__MAX_BPF_ATTACH_TYPE
 };
 
@@ -2695,8 +2701,8 @@ union bpf_attr {
  * 		*bpf_socket* should be one of the following:
  *
  * 		* **struct bpf_sock_ops** for **BPF_PROG_TYPE_SOCK_OPS**.
- * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**
- * 		  and **BPF_CGROUP_INET6_CONNECT**.
+ * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**,
+ * 		  **BPF_CGROUP_INET6_CONNECT** and **BPF_CGROUP_UNIX_CONNECT**.
  *
  * 		This helper actually implements a subset of **setsockopt()**.
  * 		It supports the following *level*\ s:
@@ -2934,8 +2940,8 @@ union bpf_attr {
  * 		*bpf_socket* should be one of the following:
  *
  * 		* **struct bpf_sock_ops** for **BPF_PROG_TYPE_SOCK_OPS**.
- * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**
- * 		  and **BPF_CGROUP_INET6_CONNECT**.
+ * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**,
+ * 		  **BPF_CGROUP_INET6_CONNECT** and **BPF_CGROUP_UNIX_CONNECT**.
  *
  * 		This helper actually implements a subset of **getsockopt()**.
  * 		It supports the same set of *optname*\ s that is supported by
kernel/bpf/cgroup.c +11 -2
index 534b6c7f5659..f2304f69016f 100644
@@ -1455,7 +1455,7 @@ EXPORT_SYMBOL(__cgroup_bpf_run_filter_sk);
  * @flags: Pointer to u32 which contains higher bits of BPF program
  *         return value (OR'ed together).
  *
- * socket is expected to be of type INET or INET6.
+ * socket is expected to be of type INET, INET6 or UNIX.
  *
  * This function will return %-EPERM if an attached program is found and
  * returned value != 1 during execution. In all other cases, 0 is returned.
@@ -1479,7 +1479,8 @@ int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
 	/* Check socket family since not all sockets represent network
 	 * endpoint (e.g. AF_UNIX).
 	 */
-	if (sk->sk_family != AF_INET && sk->sk_family != AF_INET6)
+	if (sk->sk_family != AF_INET && sk->sk_family != AF_INET6 &&
+		sk->sk_family != AF_UNIX)
 		return 0;
 
 	if (!ctx.uaddr) {
@@ -1492,6 +1493,8 @@ int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
 		ctx.uaddrlen = sizeof(struct sockaddr_in);
 	else if (sk->sk_family == AF_INET6)
 		ctx.uaddrlen = sizeof(struct sockaddr_in6);
+	else if (sk->sk_family == AF_UNIX)
+		return -EINVAL; /* uaddrlen must always be provided for AF_UNIX. */
 
 	cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
 	ret = bpf_prog_run_array_cg(&cgrp->bpf, atype, &ctx, bpf_prog_run,
@@ -2533,10 +2536,13 @@ cgroup_common_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		case BPF_CGROUP_SOCK_OPS:
 		case BPF_CGROUP_UDP4_RECVMSG:
 		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
 		case BPF_CGROUP_INET4_GETPEERNAME:
 		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
 		case BPF_CGROUP_INET4_GETSOCKNAME:
 		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
 			return NULL;
 		default:
 			return &bpf_get_retval_proto;
@@ -2548,10 +2554,13 @@ cgroup_common_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		case BPF_CGROUP_SOCK_OPS:
 		case BPF_CGROUP_UDP4_RECVMSG:
 		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
 		case BPF_CGROUP_INET4_GETPEERNAME:
 		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
 		case BPF_CGROUP_INET4_GETSOCKNAME:
 		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
 			return NULL;
 		default:
 			return &bpf_set_retval_proto;
kernel/bpf/syscall.c +18 -0
index ebeb0695305a..eb47e93bcce6 100644
@@ -2440,16 +2440,22 @@ bpf_prog_load_check_attach(enum bpf_prog_type prog_type,
 		switch (expected_attach_type) {
 		case BPF_CGROUP_INET4_BIND:
 		case BPF_CGROUP_INET6_BIND:
+		case BPF_CGROUP_UNIX_BIND:
 		case BPF_CGROUP_INET4_CONNECT:
 		case BPF_CGROUP_INET6_CONNECT:
+		case BPF_CGROUP_UNIX_CONNECT:
 		case BPF_CGROUP_INET4_GETPEERNAME:
 		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
 		case BPF_CGROUP_INET4_GETSOCKNAME:
 		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
 		case BPF_CGROUP_UDP4_SENDMSG:
 		case BPF_CGROUP_UDP6_SENDMSG:
+		case BPF_CGROUP_UNIX_SENDMSG:
 		case BPF_CGROUP_UDP4_RECVMSG:
 		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
 			return 0;
 		default:
 			return -EINVAL;
@@ -3670,16 +3676,22 @@ attach_type_to_prog_type(enum bpf_attach_type attach_type)
 		return BPF_PROG_TYPE_CGROUP_SOCK;
 	case BPF_CGROUP_INET4_BIND:
 	case BPF_CGROUP_INET6_BIND:
+	case BPF_CGROUP_UNIX_BIND:
 	case BPF_CGROUP_INET4_CONNECT:
 	case BPF_CGROUP_INET6_CONNECT:
+	case BPF_CGROUP_UNIX_CONNECT:
 	case BPF_CGROUP_INET4_GETPEERNAME:
 	case BPF_CGROUP_INET6_GETPEERNAME:
+	case BPF_CGROUP_UNIX_GETPEERNAME:
 	case BPF_CGROUP_INET4_GETSOCKNAME:
 	case BPF_CGROUP_INET6_GETSOCKNAME:
+	case BPF_CGROUP_UNIX_GETSOCKNAME:
 	case BPF_CGROUP_UDP4_SENDMSG:
 	case BPF_CGROUP_UDP6_SENDMSG:
+	case BPF_CGROUP_UNIX_SENDMSG:
 	case BPF_CGROUP_UDP4_RECVMSG:
 	case BPF_CGROUP_UDP6_RECVMSG:
+	case BPF_CGROUP_UNIX_RECVMSG:
 		return BPF_PROG_TYPE_CGROUP_SOCK_ADDR;
 	case BPF_CGROUP_SOCK_OPS:
 		return BPF_PROG_TYPE_SOCK_OPS;
@@ -3932,18 +3944,24 @@ static int bpf_prog_query(const union bpf_attr *attr,
 	case BPF_CGROUP_INET_SOCK_RELEASE:
 	case BPF_CGROUP_INET4_BIND:
 	case BPF_CGROUP_INET6_BIND:
+	case BPF_CGROUP_UNIX_BIND:
 	case BPF_CGROUP_INET4_POST_BIND:
 	case BPF_CGROUP_INET6_POST_BIND:
 	case BPF_CGROUP_INET4_CONNECT:
 	case BPF_CGROUP_INET6_CONNECT:
+	case BPF_CGROUP_UNIX_CONNECT:
 	case BPF_CGROUP_INET4_GETPEERNAME:
 	case BPF_CGROUP_INET6_GETPEERNAME:
+	case BPF_CGROUP_UNIX_GETPEERNAME:
 	case BPF_CGROUP_INET4_GETSOCKNAME:
 	case BPF_CGROUP_INET6_GETSOCKNAME:
+	case BPF_CGROUP_UNIX_GETSOCKNAME:
 	case BPF_CGROUP_UDP4_SENDMSG:
 	case BPF_CGROUP_UDP6_SENDMSG:
+	case BPF_CGROUP_UNIX_SENDMSG:
 	case BPF_CGROUP_UDP4_RECVMSG:
 	case BPF_CGROUP_UDP6_RECVMSG:
+	case BPF_CGROUP_UNIX_RECVMSG:
 	case BPF_CGROUP_SOCK_OPS:
 	case BPF_CGROUP_DEVICE:
 	case BPF_CGROUP_SYSCTL:
kernel/bpf/verifier.c +6 -1
index 0680569f9bd0..d8f508c56055 100644
@@ -14540,14 +14540,19 @@ static int check_return_code(struct bpf_verifier_env *env)
 	case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
 		if (env->prog->expected_attach_type == BPF_CGROUP_UDP4_RECVMSG ||
 		    env->prog->expected_attach_type == BPF_CGROUP_UDP6_RECVMSG ||
+		    env->prog->expected_attach_type == BPF_CGROUP_UNIX_RECVMSG ||
 		    env->prog->expected_attach_type == BPF_CGROUP_INET4_GETPEERNAME ||
 		    env->prog->expected_attach_type == BPF_CGROUP_INET6_GETPEERNAME ||
+		    env->prog->expected_attach_type == BPF_CGROUP_UNIX_GETPEERNAME ||
 		    env->prog->expected_attach_type == BPF_CGROUP_INET4_GETSOCKNAME ||
-		    env->prog->expected_attach_type == BPF_CGROUP_INET6_GETSOCKNAME)
+		    env->prog->expected_attach_type == BPF_CGROUP_INET6_GETSOCKNAME ||
+		    env->prog->expected_attach_type == BPF_CGROUP_UNIX_GETSOCKNAME)
 			range = tnum_range(1, 1);
 		if (env->prog->expected_attach_type == BPF_CGROUP_INET4_BIND ||
 		    env->prog->expected_attach_type == BPF_CGROUP_INET6_BIND)
 			range = tnum_range(0, 3);
+		if (env->prog->expected_attach_type == BPF_CGROUP_UNIX_BIND)
+			range = tnum_range(0, 1);
 		break;
 	case BPF_PROG_TYPE_CGROUP_SKB:
 		if (env->prog->expected_attach_type == BPF_CGROUP_INET_EGRESS) {
net/core/filter.c +16 -2
index 3ed6cd33b268..be4e0e923aa6 100644
@@ -81,6 +81,7 @@
 #include <net/xdp.h>
 #include <net/mptcp.h>
 #include <net/netfilter/nf_conntrack_bpf.h>
+#include <linux/un.h>
 
 static const struct bpf_func_proto *
 bpf_sk_base_func_proto(enum bpf_func_id func_id);
@@ -7828,6 +7829,7 @@ sock_addr_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		switch (prog->expected_attach_type) {
 		case BPF_CGROUP_INET4_CONNECT:
 		case BPF_CGROUP_INET6_CONNECT:
+		case BPF_CGROUP_UNIX_CONNECT:
 			return &bpf_bind_proto;
 		default:
 			return NULL;
@@ -7856,16 +7858,22 @@ sock_addr_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		switch (prog->expected_attach_type) {
 		case BPF_CGROUP_INET4_BIND:
 		case BPF_CGROUP_INET6_BIND:
+		case BPF_CGROUP_UNIX_BIND:
 		case BPF_CGROUP_INET4_CONNECT:
 		case BPF_CGROUP_INET6_CONNECT:
+		case BPF_CGROUP_UNIX_CONNECT:
 		case BPF_CGROUP_UDP4_RECVMSG:
 		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
 		case BPF_CGROUP_UDP4_SENDMSG:
 		case BPF_CGROUP_UDP6_SENDMSG:
+		case BPF_CGROUP_UNIX_SENDMSG:
 		case BPF_CGROUP_INET4_GETPEERNAME:
 		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
 		case BPF_CGROUP_INET4_GETSOCKNAME:
 		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
 			return &bpf_sock_addr_setsockopt_proto;
 		default:
 			return NULL;
@@ -7874,16 +7882,22 @@ sock_addr_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		switch (prog->expected_attach_type) {
 		case BPF_CGROUP_INET4_BIND:
 		case BPF_CGROUP_INET6_BIND:
+		case BPF_CGROUP_UNIX_BIND:
 		case BPF_CGROUP_INET4_CONNECT:
 		case BPF_CGROUP_INET6_CONNECT:
+		case BPF_CGROUP_UNIX_CONNECT:
 		case BPF_CGROUP_UDP4_RECVMSG:
 		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
 		case BPF_CGROUP_UDP4_SENDMSG:
 		case BPF_CGROUP_UDP6_SENDMSG:
+		case BPF_CGROUP_UNIX_SENDMSG:
 		case BPF_CGROUP_INET4_GETPEERNAME:
 		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
 		case BPF_CGROUP_INET4_GETSOCKNAME:
 		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
 			return &bpf_sock_addr_getsockopt_proto;
 		default:
 			return NULL;
@@ -8931,8 +8945,8 @@ static bool sock_addr_is_valid_access(int off, int size,
 	if (off % size != 0)
 		return false;
 
-	/* Disallow access to IPv6 fields from IPv4 contex and vise
-	 * versa.
+	/* Disallow access to fields not belonging to the attach type's address
+	 * family.
 	 */
 	switch (off) {
 	case bpf_ctx_range(struct bpf_sock_addr, user_ip4):
net/unix/af_unix.c +82 -8

162 lines not on this page · raw message

tools/include/uapi/linux/bpf.h +10 -4
index 8790b3962e4b..c51889d7e5d8 100644
@@ -1040,6 +1040,12 @@ enum bpf_attach_type {
 	BPF_TCX_INGRESS,
 	BPF_TCX_EGRESS,
 	BPF_TRACE_UPROBE_MULTI,
+	BPF_CGROUP_UNIX_BIND,
+	BPF_CGROUP_UNIX_CONNECT,
+	BPF_CGROUP_UNIX_SENDMSG,
+	BPF_CGROUP_UNIX_RECVMSG,
+	BPF_CGROUP_UNIX_GETPEERNAME,
+	BPF_CGROUP_UNIX_GETSOCKNAME,
 	__MAX_BPF_ATTACH_TYPE
 };
 
@@ -2695,8 +2701,8 @@ union bpf_attr {
  * 		*bpf_socket* should be one of the following:
  *
  * 		* **struct bpf_sock_ops** for **BPF_PROG_TYPE_SOCK_OPS**.
- * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**
- * 		  and **BPF_CGROUP_INET6_CONNECT**.
+ * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**,
+ * 		  **BPF_CGROUP_INET6_CONNECT** and **BPF_CGROUP_UNIX_CONNECT**.
  *
  * 		This helper actually implements a subset of **setsockopt()**.
  * 		It supports the following *level*\ s:
@@ -2934,8 +2940,8 @@ union bpf_attr {
  * 		*bpf_socket* should be one of the following:
  *
  * 		* **struct bpf_sock_ops** for **BPF_PROG_TYPE_SOCK_OPS**.
- * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**
- * 		  and **BPF_CGROUP_INET6_CONNECT**.
+ * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**,
+ * 		  **BPF_CGROUP_INET6_CONNECT** and **BPF_CGROUP_UNIX_CONNECT**.
  *
  * 		This helper actually implements a subset of **getsockopt()**.
  * 		It supports the same set of *optname*\ s that is supported by
Signature · 1 line
2.41.0
#8 lore raw

Re: [PATCH bpf-next v3 4/9] bpf: Implement cgroup sockaddr hooks for unix sockets

From: Daan De Meyer <[email protected]>
Date: Thu, 31 Aug 2023 17:34:48 +0200

Quoted 98 lines

These hooks allows intercepting bind(), connect(), getsockname(), getpeername(), sendmsg() and recvmsg() for unix sockets. The unix socket hooks get write access to the address length because the address length is not fixed when dealing with unix sockets and needs to be modified when a unix socket address is modified by the hook. Because abstract socket unix addresses start with a NUL byte, we cannot recalculate the socket address in kernelspace after running the hook by calculating the length of the unix socket path using strlen().

Write support is added for uaddrlen to allow the unix socket hooks to modify the sockaddr length from the bpf program.

This hook can be used when users want to multiplex syscall to a single unix socket to multiple different processes behind the scenes by redirecting the connect() and other syscalls to process specific sockets.

---
 include/linux/bpf-cgroup-defs.h |  6 +++
 include/linux/bpf-cgroup.h      | 29 ++++++++++-
 include/uapi/linux/bpf.h        | 14 +++--
 kernel/bpf/cgroup.c             | 13 ++++-
 kernel/bpf/syscall.c            | 18 +++++++
 kernel/bpf/verifier.c           |  7 ++-
 net/core/filter.c               | 18 ++++++-
 net/unix/af_unix.c              | 90 ++++++++++++++++++++++++++++++---
 tools/include/uapi/linux/bpf.h  | 14 +++--
 9 files changed, 187 insertions(+), 22 deletions(-)
include/linux/bpf-cgroup-defs.h +6 -0
index 7b121bd780eb..8196ccb81915 100644
@@ -26,21 +26,27 @@ enum cgroup_bpf_attach_type {
 	CGROUP_DEVICE,
 	CGROUP_INET4_BIND,
 	CGROUP_INET6_BIND,
+	CGROUP_UNIX_BIND,
 	CGROUP_INET4_CONNECT,
 	CGROUP_INET6_CONNECT,
+	CGROUP_UNIX_CONNECT,
 	CGROUP_INET4_POST_BIND,
 	CGROUP_INET6_POST_BIND,
 	CGROUP_UDP4_SENDMSG,
 	CGROUP_UDP6_SENDMSG,
+	CGROUP_UNIX_SENDMSG,
 	CGROUP_SYSCTL,
 	CGROUP_UDP4_RECVMSG,
 	CGROUP_UDP6_RECVMSG,
+	CGROUP_UNIX_RECVMSG,
 	CGROUP_GETSOCKOPT,
 	CGROUP_SETSOCKOPT,
 	CGROUP_INET4_GETPEERNAME,
 	CGROUP_INET6_GETPEERNAME,
+	CGROUP_UNIX_GETPEERNAME,
 	CGROUP_INET4_GETSOCKNAME,
 	CGROUP_INET6_GETSOCKNAME,
+	CGROUP_UNIX_GETSOCKNAME,
 	CGROUP_INET_SOCK_RELEASE,
 	CGROUP_LSM_START,
 	CGROUP_LSM_END = CGROUP_LSM_START + CGROUP_LSM_NUM - 1,
include/linux/bpf-cgroup.h +6 -0
index 31561e789715..0d572228fa62 100644
@@ -46,21 +46,27 @@ to_cgroup_bpf_attach_type(enum bpf_attach_type attach_type)
 	CGROUP_ATYPE(CGROUP_DEVICE);
 	CGROUP_ATYPE(CGROUP_INET4_BIND);
 	CGROUP_ATYPE(CGROUP_INET6_BIND);
+	CGROUP_ATYPE(CGROUP_UNIX_BIND);
 	CGROUP_ATYPE(CGROUP_INET4_CONNECT);
 	CGROUP_ATYPE(CGROUP_INET6_CONNECT);
+	CGROUP_ATYPE(CGROUP_UNIX_CONNECT);
 	CGROUP_ATYPE(CGROUP_INET4_POST_BIND);
 	CGROUP_ATYPE(CGROUP_INET6_POST_BIND);
 	CGROUP_ATYPE(CGROUP_UDP4_SENDMSG);
 	CGROUP_ATYPE(CGROUP_UDP6_SENDMSG);
+	CGROUP_ATYPE(CGROUP_UNIX_SENDMSG);
 	CGROUP_ATYPE(CGROUP_SYSCTL);
 	CGROUP_ATYPE(CGROUP_UDP4_RECVMSG);
 	CGROUP_ATYPE(CGROUP_UDP6_RECVMSG);
+	CGROUP_ATYPE(CGROUP_UNIX_RECVMSG);
 	CGROUP_ATYPE(CGROUP_GETSOCKOPT);
 	CGROUP_ATYPE(CGROUP_SETSOCKOPT);
 	CGROUP_ATYPE(CGROUP_INET4_GETPEERNAME);
 	CGROUP_ATYPE(CGROUP_INET6_GETPEERNAME);
+	CGROUP_ATYPE(CGROUP_UNIX_GETPEERNAME);
 	CGROUP_ATYPE(CGROUP_INET4_GETSOCKNAME);
 	CGROUP_ATYPE(CGROUP_INET6_GETSOCKNAME);
+	CGROUP_ATYPE(CGROUP_UNIX_GETSOCKNAME);
 	CGROUP_ATYPE(CGROUP_INET_SOCK_RELEASE);
 	default:
 		return CGROUP_BPF_ATTACH_TYPE_INVALID;
@@ -272,9 +278,13 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
 	__ret;								       \
 })

include/linux/bpf-cgroup.h static inline bool cgroup_bpf_sock_enabled(struct sock *sk, · line 288

Quoted 10 lines
include/linux/bpf-cgroup.h +5 -1
 
+#define BPF_CGROUP_RUN_PROG_UNIX_BIND_LOCK(sk, uaddr, uaddrlen)			\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UNIX_BIND, NULL)
+
 #define BPF_CGROUP_PRE_CONNECT_ENABLED(sk)				       \
 	((cgroup_bpf_enabled(CGROUP_INET4_CONNECT) ||		       \
-	  cgroup_bpf_enabled(CGROUP_INET6_CONNECT)) &&		       \
+	  cgroup_bpf_enabled(CGROUP_INET6_CONNECT) ||		       \
+	  cgroup_bpf_enabled(CGROUP_UNIX_CONNECT)) &&		       \
 	 (sk)->sk_prot->pre_connect)

Do we need this change ?
AF_UNIX does not have pre_connect().

Quoted 554 lines
include/linux/bpf-cgroup.h +17 -0
 #define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr, uaddrlen)			\
@@ -283,24 +293,36 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
 #define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr, uaddrlen)			\
 	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, CGROUP_INET6_CONNECT)
 
+#define BPF_CGROUP_RUN_PROG_UNIX_CONNECT(sk, uaddr, uaddrlen)			\
+	BPF_CGROUP_RUN_SA_PROG(sk, uaddr, uaddrlen, CGROUP_UNIX_CONNECT)
+
 #define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr, uaddrlen)		\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_INET4_CONNECT, NULL)
 
 #define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr, uaddrlen)		\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_INET6_CONNECT, NULL)
 
+#define BPF_CGROUP_RUN_PROG_UNIX_CONNECT_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UNIX_CONNECT, NULL)
+
 #define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx)	\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP4_SENDMSG, t_ctx)
 
 #define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx)	\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP6_SENDMSG, t_ctx)
 
+#define BPF_CGROUP_RUN_PROG_UNIX_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx)	\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UNIX_SENDMSG, t_ctx)
+
 #define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr, uaddrlen)		\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP4_RECVMSG, NULL)
 
 #define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr, uaddrlen)		\
 	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UDP6_RECVMSG, NULL)
 
+#define BPF_CGROUP_RUN_PROG_UNIX_RECVMSG_LOCK(sk, uaddr, uaddrlen)		\
+	BPF_CGROUP_RUN_SA_PROG_LOCK(sk, uaddr, uaddrlen, CGROUP_UNIX_RECVMSG, NULL)
+
 /* The SOCK_OPS"_SK" macro should be used when sock_ops->sk is not a
  * fullsock and its parent fullsock cannot be traced by
  * sk_to_full_sk().
@@ -486,16 +508,21 @@ static inline int bpf_percpu_cgroup_storage_update(struct bpf_map *map,
 #define BPF_CGROUP_RUN_PROG_INET_SOCK(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET_SOCK_RELEASE(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET_BIND_LOCK(sk, uaddr, uaddrlen, atype, flags) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UNIX_BIND_LOCK(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET4_POST_BIND(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET6_POST_BIND(sk) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET4_CONNECT(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET4_CONNECT_LOCK(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET6_CONNECT(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_INET6_CONNECT_LOCK(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UNIX_CONNECT(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UNIX_CONNECT_LOCK(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_UDP4_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_UDP6_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UNIX_SENDMSG_LOCK(sk, uaddr, uaddrlen, t_ctx) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_UDP4_RECVMSG_LOCK(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_UDP6_RECVMSG_LOCK(sk, uaddr, uaddrlen) ({ 0; })
+#define BPF_CGROUP_RUN_PROG_UNIX_RECVMSG_LOCK(sk, uaddr, uaddrlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_SOCK_OPS(sock_ops) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_DEVICE_CGROUP(atype, major, minor, access) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_SYSCTL(head,table,write,buf,count,pos) ({ 0; })
include/uapi/linux/bpf.h +10 -4
index 8790b3962e4b..c51889d7e5d8 100644
@@ -1040,6 +1040,12 @@ enum bpf_attach_type {
 	BPF_TCX_INGRESS,
 	BPF_TCX_EGRESS,
 	BPF_TRACE_UPROBE_MULTI,
+	BPF_CGROUP_UNIX_BIND,
+	BPF_CGROUP_UNIX_CONNECT,
+	BPF_CGROUP_UNIX_SENDMSG,
+	BPF_CGROUP_UNIX_RECVMSG,
+	BPF_CGROUP_UNIX_GETPEERNAME,
+	BPF_CGROUP_UNIX_GETSOCKNAME,
 	__MAX_BPF_ATTACH_TYPE
 };
 
@@ -2695,8 +2701,8 @@ union bpf_attr {
  * 		*bpf_socket* should be one of the following:
  *
  * 		* **struct bpf_sock_ops** for **BPF_PROG_TYPE_SOCK_OPS**.
- * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**
- * 		  and **BPF_CGROUP_INET6_CONNECT**.
+ * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**,
+ * 		  **BPF_CGROUP_INET6_CONNECT** and **BPF_CGROUP_UNIX_CONNECT**.
  *
  * 		This helper actually implements a subset of **setsockopt()**.
  * 		It supports the following *level*\ s:
@@ -2934,8 +2940,8 @@ union bpf_attr {
  * 		*bpf_socket* should be one of the following:
  *
  * 		* **struct bpf_sock_ops** for **BPF_PROG_TYPE_SOCK_OPS**.
- * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**
- * 		  and **BPF_CGROUP_INET6_CONNECT**.
+ * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**,
+ * 		  **BPF_CGROUP_INET6_CONNECT** and **BPF_CGROUP_UNIX_CONNECT**.
  *
  * 		This helper actually implements a subset of **getsockopt()**.
  * 		It supports the same set of *optname*\ s that is supported by
kernel/bpf/cgroup.c +11 -2
index 534b6c7f5659..f2304f69016f 100644
@@ -1455,7 +1455,7 @@ EXPORT_SYMBOL(__cgroup_bpf_run_filter_sk);
  * @flags: Pointer to u32 which contains higher bits of BPF program
  *         return value (OR'ed together).
  *
- * socket is expected to be of type INET or INET6.
+ * socket is expected to be of type INET, INET6 or UNIX.
  *
  * This function will return %-EPERM if an attached program is found and
  * returned value != 1 during execution. In all other cases, 0 is returned.
@@ -1479,7 +1479,8 @@ int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
 	/* Check socket family since not all sockets represent network
 	 * endpoint (e.g. AF_UNIX).
 	 */
-	if (sk->sk_family != AF_INET && sk->sk_family != AF_INET6)
+	if (sk->sk_family != AF_INET && sk->sk_family != AF_INET6 &&
+		sk->sk_family != AF_UNIX)
 		return 0;
 
 	if (!ctx.uaddr) {
@@ -1492,6 +1493,8 @@ int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
 		ctx.uaddrlen = sizeof(struct sockaddr_in);
 	else if (sk->sk_family == AF_INET6)
 		ctx.uaddrlen = sizeof(struct sockaddr_in6);
+	else if (sk->sk_family == AF_UNIX)
+		return -EINVAL; /* uaddrlen must always be provided for AF_UNIX. */
 
 	cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
 	ret = bpf_prog_run_array_cg(&cgrp->bpf, atype, &ctx, bpf_prog_run,
@@ -2533,10 +2536,13 @@ cgroup_common_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		case BPF_CGROUP_SOCK_OPS:
 		case BPF_CGROUP_UDP4_RECVMSG:
 		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
 		case BPF_CGROUP_INET4_GETPEERNAME:
 		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
 		case BPF_CGROUP_INET4_GETSOCKNAME:
 		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
 			return NULL;
 		default:
 			return &bpf_get_retval_proto;
@@ -2548,10 +2554,13 @@ cgroup_common_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		case BPF_CGROUP_SOCK_OPS:
 		case BPF_CGROUP_UDP4_RECVMSG:
 		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
 		case BPF_CGROUP_INET4_GETPEERNAME:
 		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
 		case BPF_CGROUP_INET4_GETSOCKNAME:
 		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
 			return NULL;
 		default:
 			return &bpf_set_retval_proto;
kernel/bpf/syscall.c +18 -0
index ebeb0695305a..eb47e93bcce6 100644
@@ -2440,16 +2440,22 @@ bpf_prog_load_check_attach(enum bpf_prog_type prog_type,
 		switch (expected_attach_type) {
 		case BPF_CGROUP_INET4_BIND:
 		case BPF_CGROUP_INET6_BIND:
+		case BPF_CGROUP_UNIX_BIND:
 		case BPF_CGROUP_INET4_CONNECT:
 		case BPF_CGROUP_INET6_CONNECT:
+		case BPF_CGROUP_UNIX_CONNECT:
 		case BPF_CGROUP_INET4_GETPEERNAME:
 		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
 		case BPF_CGROUP_INET4_GETSOCKNAME:
 		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
 		case BPF_CGROUP_UDP4_SENDMSG:
 		case BPF_CGROUP_UDP6_SENDMSG:
+		case BPF_CGROUP_UNIX_SENDMSG:
 		case BPF_CGROUP_UDP4_RECVMSG:
 		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
 			return 0;
 		default:
 			return -EINVAL;
@@ -3670,16 +3676,22 @@ attach_type_to_prog_type(enum bpf_attach_type attach_type)
 		return BPF_PROG_TYPE_CGROUP_SOCK;
 	case BPF_CGROUP_INET4_BIND:
 	case BPF_CGROUP_INET6_BIND:
+	case BPF_CGROUP_UNIX_BIND:
 	case BPF_CGROUP_INET4_CONNECT:
 	case BPF_CGROUP_INET6_CONNECT:
+	case BPF_CGROUP_UNIX_CONNECT:
 	case BPF_CGROUP_INET4_GETPEERNAME:
 	case BPF_CGROUP_INET6_GETPEERNAME:
+	case BPF_CGROUP_UNIX_GETPEERNAME:
 	case BPF_CGROUP_INET4_GETSOCKNAME:
 	case BPF_CGROUP_INET6_GETSOCKNAME:
+	case BPF_CGROUP_UNIX_GETSOCKNAME:
 	case BPF_CGROUP_UDP4_SENDMSG:
 	case BPF_CGROUP_UDP6_SENDMSG:
+	case BPF_CGROUP_UNIX_SENDMSG:
 	case BPF_CGROUP_UDP4_RECVMSG:
 	case BPF_CGROUP_UDP6_RECVMSG:
+	case BPF_CGROUP_UNIX_RECVMSG:
 		return BPF_PROG_TYPE_CGROUP_SOCK_ADDR;
 	case BPF_CGROUP_SOCK_OPS:
 		return BPF_PROG_TYPE_SOCK_OPS;
@@ -3932,18 +3944,24 @@ static int bpf_prog_query(const union bpf_attr *attr,
 	case BPF_CGROUP_INET_SOCK_RELEASE:
 	case BPF_CGROUP_INET4_BIND:
 	case BPF_CGROUP_INET6_BIND:
+	case BPF_CGROUP_UNIX_BIND:
 	case BPF_CGROUP_INET4_POST_BIND:
 	case BPF_CGROUP_INET6_POST_BIND:
 	case BPF_CGROUP_INET4_CONNECT:
 	case BPF_CGROUP_INET6_CONNECT:
+	case BPF_CGROUP_UNIX_CONNECT:
 	case BPF_CGROUP_INET4_GETPEERNAME:
 	case BPF_CGROUP_INET6_GETPEERNAME:
+	case BPF_CGROUP_UNIX_GETPEERNAME:
 	case BPF_CGROUP_INET4_GETSOCKNAME:
 	case BPF_CGROUP_INET6_GETSOCKNAME:
+	case BPF_CGROUP_UNIX_GETSOCKNAME:
 	case BPF_CGROUP_UDP4_SENDMSG:
 	case BPF_CGROUP_UDP6_SENDMSG:
+	case BPF_CGROUP_UNIX_SENDMSG:
 	case BPF_CGROUP_UDP4_RECVMSG:
 	case BPF_CGROUP_UDP6_RECVMSG:
+	case BPF_CGROUP_UNIX_RECVMSG:
 	case BPF_CGROUP_SOCK_OPS:
 	case BPF_CGROUP_DEVICE:
 	case BPF_CGROUP_SYSCTL:
kernel/bpf/verifier.c +6 -1
index 0680569f9bd0..d8f508c56055 100644
@@ -14540,14 +14540,19 @@ static int check_return_code(struct bpf_verifier_env *env)
 	case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
 		if (env->prog->expected_attach_type == BPF_CGROUP_UDP4_RECVMSG ||
 		    env->prog->expected_attach_type == BPF_CGROUP_UDP6_RECVMSG ||
+		    env->prog->expected_attach_type == BPF_CGROUP_UNIX_RECVMSG ||
 		    env->prog->expected_attach_type == BPF_CGROUP_INET4_GETPEERNAME ||
 		    env->prog->expected_attach_type == BPF_CGROUP_INET6_GETPEERNAME ||
+		    env->prog->expected_attach_type == BPF_CGROUP_UNIX_GETPEERNAME ||
 		    env->prog->expected_attach_type == BPF_CGROUP_INET4_GETSOCKNAME ||
-		    env->prog->expected_attach_type == BPF_CGROUP_INET6_GETSOCKNAME)
+		    env->prog->expected_attach_type == BPF_CGROUP_INET6_GETSOCKNAME ||
+		    env->prog->expected_attach_type == BPF_CGROUP_UNIX_GETSOCKNAME)
 			range = tnum_range(1, 1);
 		if (env->prog->expected_attach_type == BPF_CGROUP_INET4_BIND ||
 		    env->prog->expected_attach_type == BPF_CGROUP_INET6_BIND)
 			range = tnum_range(0, 3);
+		if (env->prog->expected_attach_type == BPF_CGROUP_UNIX_BIND)
+			range = tnum_range(0, 1);
 		break;
 	case BPF_PROG_TYPE_CGROUP_SKB:
 		if (env->prog->expected_attach_type == BPF_CGROUP_INET_EGRESS) {
net/core/filter.c +16 -2
index 3ed6cd33b268..be4e0e923aa6 100644
@@ -81,6 +81,7 @@
 #include <net/xdp.h>
 #include <net/mptcp.h>
 #include <net/netfilter/nf_conntrack_bpf.h>
+#include <linux/un.h>
 
 static const struct bpf_func_proto *
 bpf_sk_base_func_proto(enum bpf_func_id func_id);
@@ -7828,6 +7829,7 @@ sock_addr_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		switch (prog->expected_attach_type) {
 		case BPF_CGROUP_INET4_CONNECT:
 		case BPF_CGROUP_INET6_CONNECT:
+		case BPF_CGROUP_UNIX_CONNECT:
 			return &bpf_bind_proto;
 		default:
 			return NULL;
@@ -7856,16 +7858,22 @@ sock_addr_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		switch (prog->expected_attach_type) {
 		case BPF_CGROUP_INET4_BIND:
 		case BPF_CGROUP_INET6_BIND:
+		case BPF_CGROUP_UNIX_BIND:
 		case BPF_CGROUP_INET4_CONNECT:
 		case BPF_CGROUP_INET6_CONNECT:
+		case BPF_CGROUP_UNIX_CONNECT:
 		case BPF_CGROUP_UDP4_RECVMSG:
 		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
 		case BPF_CGROUP_UDP4_SENDMSG:
 		case BPF_CGROUP_UDP6_SENDMSG:
+		case BPF_CGROUP_UNIX_SENDMSG:
 		case BPF_CGROUP_INET4_GETPEERNAME:
 		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
 		case BPF_CGROUP_INET4_GETSOCKNAME:
 		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
 			return &bpf_sock_addr_setsockopt_proto;
 		default:
 			return NULL;
@@ -7874,16 +7882,22 @@ sock_addr_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		switch (prog->expected_attach_type) {
 		case BPF_CGROUP_INET4_BIND:
 		case BPF_CGROUP_INET6_BIND:
+		case BPF_CGROUP_UNIX_BIND:
 		case BPF_CGROUP_INET4_CONNECT:
 		case BPF_CGROUP_INET6_CONNECT:
+		case BPF_CGROUP_UNIX_CONNECT:
 		case BPF_CGROUP_UDP4_RECVMSG:
 		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
 		case BPF_CGROUP_UDP4_SENDMSG:
 		case BPF_CGROUP_UDP6_SENDMSG:
+		case BPF_CGROUP_UNIX_SENDMSG:
 		case BPF_CGROUP_INET4_GETPEERNAME:
 		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
 		case BPF_CGROUP_INET4_GETSOCKNAME:
 		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
 			return &bpf_sock_addr_getsockopt_proto;
 		default:
 			return NULL;
@@ -8931,8 +8945,8 @@ static bool sock_addr_is_valid_access(int off, int size,
 	if (off % size != 0)
 		return false;
 
-	/* Disallow access to IPv6 fields from IPv4 contex and vise
-	 * versa.
+	/* Disallow access to fields not belonging to the attach type's address
+	 * family.
 	 */
 	switch (off) {
 	case bpf_ctx_range(struct bpf_sock_addr, user_ip4):
net/unix/af_unix.c +82 -8
index 86930a8ed012..94fd6f2441d8 100644
@@ -116,6 +116,7 @@
 #include <linux/freezer.h>
 #include <linux/file.h>
 #include <linux/btf_ids.h>
+#include <linux/bpf-cgroup.h>
 
 #include "scm.h"
 
@@ -1323,6 +1324,12 @@ static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
 	struct sock *sk = sock->sk;
 	int err;
 
+	if (cgroup_bpf_enabled(CGROUP_UNIX_BIND)) {
+		err = BPF_CGROUP_RUN_PROG_UNIX_BIND_LOCK(sk, uaddr, &addr_len);
+		if (err)
+			return err;
+	}
+
 	if (addr_len == offsetof(struct sockaddr_un, sun_path) &&
 	    sunaddr->sun_family == AF_UNIX)
 		return unix_autobind(sk);
@@ -1377,6 +1384,13 @@ static int unix_dgram_connect(struct socket *sock, struct sockaddr *addr,
 		goto out;
 
 	if (addr->sa_family != AF_UNSPEC) {
+		if (cgroup_bpf_enabled(CGROUP_UNIX_CONNECT)) {
+			err = BPF_CGROUP_RUN_PROG_UNIX_CONNECT_LOCK(sk, addr,
+								    &alen);
+			if (err)
+				goto out;
+		}
+
 		err = unix_validate_addr(sunaddr, alen);
 		if (err)
 			goto out;
@@ -1486,6 +1500,13 @@ static int unix_stream_connect(struct socket *sock, struct sockaddr *uaddr,
 	int err;
 	int st;
 
+	if (cgroup_bpf_enabled(CGROUP_UNIX_CONNECT)) {
+		err = BPF_CGROUP_RUN_PROG_UNIX_CONNECT_LOCK(sk, uaddr,
+							    &addr_len);
+		if (err)
+			goto out;
+	}
+
 	err = unix_validate_addr(sunaddr, addr_len);
 	if (err)
 		goto out;
@@ -1749,7 +1770,7 @@ static int unix_getname(struct socket *sock, struct sockaddr *uaddr, int peer)
 	struct sock *sk = sock->sk;
 	struct unix_address *addr;
 	DECLARE_SOCKADDR(struct sockaddr_un *, sunaddr, uaddr);
-	int err = 0;
+	int addr_len = 0, err = 0;
 
 	if (peer) {
 		sk = unix_peer_get(sk);
@@ -1766,14 +1787,37 @@ static int unix_getname(struct socket *sock, struct sockaddr *uaddr, int peer)
 	if (!addr) {
 		sunaddr->sun_family = AF_UNIX;
 		sunaddr->sun_path[0] = 0;
-		err = offsetof(struct sockaddr_un, sun_path);
+		addr_len = offsetof(struct sockaddr_un, sun_path);
 	} else {
-		err = addr->len;
+		addr_len = addr->len;
 		memcpy(sunaddr, addr->name, addr->len);
 	}
+
+	if (peer && cgroup_bpf_enabled(CGROUP_UNIX_GETPEERNAME)) {
+		err = BPF_CGROUP_RUN_SA_PROG(sk, uaddr, &addr_len,
+					     CGROUP_UNIX_GETPEERNAME);
+		if (err)
+			goto out;
+
+		err = unix_validate_addr(sunaddr, addr_len);
+		if (err)
+			goto out;
+	}
+
+	if (!peer && cgroup_bpf_enabled(CGROUP_UNIX_GETSOCKNAME)) {
+		err = BPF_CGROUP_RUN_SA_PROG(sk, uaddr, &addr_len,
+					     CGROUP_UNIX_GETSOCKNAME);
+		if (err)
+			goto out;
+
+		err = unix_validate_addr(sunaddr, addr_len);
+		if (err)
+			goto out;
+	}
+
 	sock_put(sk);
 out:
-	return err;
+	return err ?: addr_len;
 }
 
 static void unix_peek_fds(struct scm_cookie *scm, struct sk_buff *skb)
@@ -1919,6 +1963,15 @@ static int unix_dgram_sendmsg(struct socket *sock, struct msghdr *msg,
 		goto out;
 
 	if (msg->msg_namelen) {
+		if (cgroup_bpf_enabled(CGROUP_UNIX_SENDMSG)) {
+			err = BPF_CGROUP_RUN_PROG_UNIX_SENDMSG_LOCK(sk,
+								    msg->msg_name,
+								    &msg->msg_namelen,
+								    NULL);
+			if (err)
+				goto out;
+		}
+
 		err = unix_validate_addr(sunaddr, msg->msg_namelen);
 		if (err)
 			goto out;
@@ -2328,14 +2381,30 @@ static int unix_seqpacket_recvmsg(struct socket *sock, struct msghdr *msg,
 	return unix_dgram_recvmsg(sock, msg, size, flags);
 }
 
-static void unix_copy_addr(struct msghdr *msg, struct sock *sk)
+static int unix_recvmsg_copy_addr(struct msghdr *msg, struct sock *sk)
 {
 	struct unix_address *addr = smp_load_acquire(&unix_sk(sk)->addr);
+	int err;
 
 	if (addr) {
 		msg->msg_namelen = addr->len;
 		memcpy(msg->msg_name, addr->name, addr->len);
+
+		if (cgroup_bpf_enabled(CGROUP_UNIX_RECVMSG)) {
+			err = BPF_CGROUP_RUN_PROG_UNIX_RECVMSG_LOCK(sk,
+								    msg->msg_name,
+								    &msg->msg_namelen);
+			if (err)
+				return err;
+
+			err = unix_validate_addr(msg->msg_name,
+						 msg->msg_namelen);
+			if (err)
+				return err;
+		}
 	}
+
+	return 0;
 }
 
 int __unix_dgram_recvmsg(struct sock *sk, struct msghdr *msg, size_t size,
@@ -2390,8 +2459,11 @@ int __unix_dgram_recvmsg(struct sock *sk, struct msghdr *msg, size_t size,
 						EPOLLOUT | EPOLLWRNORM |
 						EPOLLWRBAND);
 
-	if (msg->msg_name)
-		unix_copy_addr(msg, skb->sk);
+	if (msg->msg_name) {
+		err = unix_recvmsg_copy_addr(msg, skb->sk);
+		if (err)
+			goto out_free;
+	}
 
 	if (size > skb->len - skip)
 		size = skb->len - skip;
@@ -2743,7 +2815,9 @@ static int unix_stream_read_generic(struct unix_stream_read_state *state,
 		if (state->msg && state->msg->msg_name) {
 			DECLARE_SOCKADDR(struct sockaddr_un *, sunaddr,
 					 state->msg->msg_name);
-			unix_copy_addr(state->msg, skb->sk);
+			err = unix_recvmsg_copy_addr(state->msg, skb->sk);
+			if (err)
+				break;
 			sunaddr = NULL;
 		}
 
tools/include/uapi/linux/bpf.h +10 -4
index 8790b3962e4b..c51889d7e5d8 100644
@@ -1040,6 +1040,12 @@ enum bpf_attach_type {
 	BPF_TCX_INGRESS,
 	BPF_TCX_EGRESS,
 	BPF_TRACE_UPROBE_MULTI,
+	BPF_CGROUP_UNIX_BIND,
+	BPF_CGROUP_UNIX_CONNECT,
+	BPF_CGROUP_UNIX_SENDMSG,
+	BPF_CGROUP_UNIX_RECVMSG,
+	BPF_CGROUP_UNIX_GETPEERNAME,
+	BPF_CGROUP_UNIX_GETSOCKNAME,
 	__MAX_BPF_ATTACH_TYPE
 };
 
@@ -2695,8 +2701,8 @@ union bpf_attr {
  * 		*bpf_socket* should be one of the following:
  *
  * 		* **struct bpf_sock_ops** for **BPF_PROG_TYPE_SOCK_OPS**.
- * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**
- * 		  and **BPF_CGROUP_INET6_CONNECT**.
+ * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**,
+ * 		  **BPF_CGROUP_INET6_CONNECT** and **BPF_CGROUP_UNIX_CONNECT**.
  *
  * 		This helper actually implements a subset of **setsockopt()**.
  * 		It supports the following *level*\ s:
@@ -2934,8 +2940,8 @@ union bpf_attr {
  * 		*bpf_socket* should be one of the following:
  *
  * 		* **struct bpf_sock_ops** for **BPF_PROG_TYPE_SOCK_OPS**.
- * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**
- * 		  and **BPF_CGROUP_INET6_CONNECT**.
+ * 		* **struct bpf_sock_addr** for **BPF_CGROUP_INET4_CONNECT**,
+ * 		  **BPF_CGROUP_INET6_CONNECT** and **BPF_CGROUP_UNIX_CONNECT**.
  *
  * 		This helper actually implements a subset of **getsockopt()**.
  * 		It supports the same set of *optname*\ s that is supported by
Signature · 1 line
2.41.0
#9 lore raw

Re: [PATCH bpf-next v3 4/9] bpf: Implement cgroup sockaddr hooks for unix sockets

In reply to Daan De Meyer
Quoted 7 lines · Daan De Meyer
kernel/bpf/verifier.c +0 -0
index 0680569f9bd0..d8f508c56055 100644
@@ -14540,14 +14540,19 @@ static int check_return_code(struct bpf_verifier_env *env)
  	case BPF_PROG_TYPE_CGROUP_SOCK_ADDR:
  		if (env->prog->expected_attach_type == BPF_CGROUP_UDP4_RECVMSG ||

kernel/bpf/verifier.c static int check_return_code(struct bpf_verifier_env *env) · line 14550

Quoted 10 lines
kernel/bpf/verifier.c +4 -1
  		    env->prog->expected_attach_type == BPF_CGROUP_UDP6_RECVMSG ||
+		    env->prog->expected_attach_type == BPF_CGROUP_UNIX_RECVMSG ||
  		    env->prog->expected_attach_type == BPF_CGROUP_INET4_GETPEERNAME ||
  		    env->prog->expected_attach_type == BPF_CGROUP_INET6_GETPEERNAME ||
+		    env->prog->expected_attach_type == BPF_CGROUP_UNIX_GETPEERNAME ||
  		    env->prog->expected_attach_type == BPF_CGROUP_INET4_GETSOCKNAME ||
-		    env->prog->expected_attach_type == BPF_CGROUP_INET6_GETSOCKNAME)
+		    env->prog->expected_attach_type == BPF_CGROUP_INET6_GETSOCKNAME ||
+		    env->prog->expected_attach_type == BPF_CGROUP_UNIX_GETSOCKNAME)
  			range = tnum_range(1, 1);

A note that getpeername, getsockname, and recvmsg cannot return err (err is value 0 for cgroup-bpf). More on this later.

kernel/bpf/verifier.c static int check_return_code(struct bpf_verifier_env *env) · line 14555

Quoted 5 lines
kernel/bpf/verifier.c +2 -0
  		if (env->prog->expected_attach_type == BPF_CGROUP_INET4_BIND ||
  		    env->prog->expected_attach_type == BPF_CGROUP_INET6_BIND)
  			range = tnum_range(0, 3);
+		if (env->prog->expected_attach_type == BPF_CGROUP_UNIX_BIND)
+			range = tnum_range(0, 1);

A few words in the commit message is needed for the difference on the return code between UNIX_BIND and INET[46]_BIND. (ie. the BPF_RET_BIND_NO_CAP_NET_BIND_SERVICE).

Also, the default range should be (0, 1) already (the 'struct tnum range = tnum_range(0, 1)' at the beginning of this function). The same goes for UNIX_SENDMSG (and the existing INET[46]_SENDMSG) which should already have the default (0, 1) range. Thus, no need to have a special test case here.

net/core/filter.c line 84

Quoted 12 lines
kernel/bpf/verifier.c +0 -0
  		break;
  	case BPF_PROG_TYPE_CGROUP_SKB:
  		if (env->prog->expected_attach_type == BPF_CGROUP_INET_EGRESS) {
net/core/filter.c +1 -0
index 3ed6cd33b268..be4e0e923aa6 100644
@@ -81,6 +81,7 @@
  #include <net/xdp.h>
  #include <net/mptcp.h>
  #include <net/netfilter/nf_conntrack_bpf.h>
+#include <linux/un.h>

Is this needed?

Quoted 81 lines
net/core/filter.c +15 -2
  static const struct bpf_func_proto *
  bpf_sk_base_func_proto(enum bpf_func_id func_id);
@@ -7828,6 +7829,7 @@ sock_addr_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
  		switch (prog->expected_attach_type) {
  		case BPF_CGROUP_INET4_CONNECT:
  		case BPF_CGROUP_INET6_CONNECT:
+		case BPF_CGROUP_UNIX_CONNECT:
  			return &bpf_bind_proto;
  		default:
  			return NULL;
@@ -7856,16 +7858,22 @@ sock_addr_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
  		switch (prog->expected_attach_type) {
  		case BPF_CGROUP_INET4_BIND:
  		case BPF_CGROUP_INET6_BIND:
+		case BPF_CGROUP_UNIX_BIND:
  		case BPF_CGROUP_INET4_CONNECT:
  		case BPF_CGROUP_INET6_CONNECT:
+		case BPF_CGROUP_UNIX_CONNECT:
  		case BPF_CGROUP_UDP4_RECVMSG:
  		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
  		case BPF_CGROUP_UDP4_SENDMSG:
  		case BPF_CGROUP_UDP6_SENDMSG:
+		case BPF_CGROUP_UNIX_SENDMSG:
  		case BPF_CGROUP_INET4_GETPEERNAME:
  		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
  		case BPF_CGROUP_INET4_GETSOCKNAME:
  		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
  			return &bpf_sock_addr_setsockopt_proto;
  		default:
  			return NULL;
@@ -7874,16 +7882,22 @@ sock_addr_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
  		switch (prog->expected_attach_type) {
  		case BPF_CGROUP_INET4_BIND:
  		case BPF_CGROUP_INET6_BIND:
+		case BPF_CGROUP_UNIX_BIND:
  		case BPF_CGROUP_INET4_CONNECT:
  		case BPF_CGROUP_INET6_CONNECT:
+		case BPF_CGROUP_UNIX_CONNECT:
  		case BPF_CGROUP_UDP4_RECVMSG:
  		case BPF_CGROUP_UDP6_RECVMSG:
+		case BPF_CGROUP_UNIX_RECVMSG:
  		case BPF_CGROUP_UDP4_SENDMSG:
  		case BPF_CGROUP_UDP6_SENDMSG:
+		case BPF_CGROUP_UNIX_SENDMSG:
  		case BPF_CGROUP_INET4_GETPEERNAME:
  		case BPF_CGROUP_INET6_GETPEERNAME:
+		case BPF_CGROUP_UNIX_GETPEERNAME:
  		case BPF_CGROUP_INET4_GETSOCKNAME:
  		case BPF_CGROUP_INET6_GETSOCKNAME:
+		case BPF_CGROUP_UNIX_GETSOCKNAME:
  			return &bpf_sock_addr_getsockopt_proto;
  		default:
  			return NULL;
@@ -8931,8 +8945,8 @@ static bool sock_addr_is_valid_access(int off, int size,
  	if (off % size != 0)
  		return false;
 
-	/* Disallow access to IPv6 fields from IPv4 contex and vise
-	 * versa.
+	/* Disallow access to fields not belonging to the attach type's address
+	 * family.
  	 */
  	switch (off) {
  	case bpf_ctx_range(struct bpf_sock_addr, user_ip4):
net/unix/af_unix.c +1 -0
index 86930a8ed012..94fd6f2441d8 100644
@@ -116,6 +116,7 @@
  #include <linux/freezer.h>
  #include <linux/file.h>
  #include <linux/btf_ids.h>
+#include <linux/bpf-cgroup.h>
 
  #include "scm.h"
 

net/unix/af_unix.c static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len) · line 1327

Quoted 4 lines
net/unix/af_unix.c +1 -0
@@ -1323,6 +1324,12 @@ static int unix_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
  	struct sock *sk = sock->sk;
  	int err;
 
+	if (cgroup_bpf_enabled(CGROUP_UNIX_BIND)) {

It is a dup test. The same static_key test will be done in BPF_CGROUP_RUN_SA_PROG*() also?

The same comment for other places before calling BPF_CGROUP_RUN_PROG_UNIX_* and BPF_CGROUP_RUN_SA_PROG().

Quoted 51 lines
net/unix/af_unix.c +21 -2
+		err = BPF_CGROUP_RUN_PROG_UNIX_BIND_LOCK(sk, uaddr, &addr_len);
+		if (err)
+			return err;
+	}
+
  	if (addr_len == offsetof(struct sockaddr_un, sun_path) &&
  	    sunaddr->sun_family == AF_UNIX)
  		return unix_autobind(sk);
@@ -1377,6 +1384,13 @@ static int unix_dgram_connect(struct socket *sock, struct sockaddr *addr,
  		goto out;
 
  	if (addr->sa_family != AF_UNSPEC) {
+		if (cgroup_bpf_enabled(CGROUP_UNIX_CONNECT)) {
+			err = BPF_CGROUP_RUN_PROG_UNIX_CONNECT_LOCK(sk, addr,
+								    &alen);
+			if (err)
+				goto out;
+		}
+
  		err = unix_validate_addr(sunaddr, alen);
  		if (err)
  			goto out;
@@ -1486,6 +1500,13 @@ static int unix_stream_connect(struct socket *sock, struct sockaddr *uaddr,
  	int err;
  	int st;
 
+	if (cgroup_bpf_enabled(CGROUP_UNIX_CONNECT)) {
+		err = BPF_CGROUP_RUN_PROG_UNIX_CONNECT_LOCK(sk, uaddr,
+							    &addr_len);
+		if (err)
+			goto out;
+	}
+
  	err = unix_validate_addr(sunaddr, addr_len);
  	if (err)
  		goto out;
@@ -1749,7 +1770,7 @@ static int unix_getname(struct socket *sock, struct sockaddr *uaddr, int peer)
  	struct sock *sk = sock->sk;
  	struct unix_address *addr;
  	DECLARE_SOCKADDR(struct sockaddr_un *, sunaddr, uaddr);
-	int err = 0;
+	int addr_len = 0, err = 0;
 
  	if (peer) {
  		sk = unix_peer_get(sk);
@@ -1766,14 +1787,37 @@ static int unix_getname(struct socket *sock, struct sockaddr *uaddr, int peer)
  	if (!addr) {
  		sunaddr->sun_family = AF_UNIX;
  		sunaddr->sun_path[0] = 0;
-		err = offsetof(struct sockaddr_un, sun_path);
+		addr_len = offsetof(struct sockaddr_un, sun_path);

net/unix/af_unix.c static int unix_getname(struct socket *sock, struct sockaddr *uaddr, int peer) · line 1799

Quoted 10 lines
net/unix/af_unix.c +6 -1
  	} else {
-		err = addr->len;
+		addr_len = addr->len;
  		memcpy(sunaddr, addr->name, addr->len);
  	}
+
+	if (peer && cgroup_bpf_enabled(CGROUP_UNIX_GETPEERNAME)) {
+		err = BPF_CGROUP_RUN_SA_PROG(sk, uaddr, &addr_len,
+					     CGROUP_UNIX_GETPEERNAME);
+		if (err)

UNIX_GETPEERNAME can only have return value 1 (OK), so no need to do err check here.

net/unix/af_unix.c static int unix_getname(struct socket *sock, struct sockaddr *uaddr, int peer) · line 1802

Quoted 3 lines
net/unix/af_unix.c +3 -0
+			goto out;
+
+		err = unix_validate_addr(sunaddr, addr_len);

Since the kfunc is specific to the unix address, how about doing the unix_validate_addr check in the kfunc itself?

net/unix/af_unix.c static int unix_getname(struct socket *sock, struct sockaddr *uaddr, int peer) · line 1810

Quoted 8 lines
net/unix/af_unix.c +8 -0
+		if (err)
+			goto out;
+	}
+
+	if (!peer && cgroup_bpf_enabled(CGROUP_UNIX_GETSOCKNAME)) {
+		err = BPF_CGROUP_RUN_SA_PROG(sk, uaddr, &addr_len,
+					     CGROUP_UNIX_GETSOCKNAME);
+		if (err)

Same here on the unnecessary err check.

Quoted 39 lines
net/unix/af_unix.c +19 -2
+			goto out;
+
+		err = unix_validate_addr(sunaddr, addr_len);
+		if (err)
+			goto out;
+	}
+
  	sock_put(sk);
  out:
-	return err;
+	return err ?: addr_len;
  }
 
  static void unix_peek_fds(struct scm_cookie *scm, struct sk_buff *skb)
@@ -1919,6 +1963,15 @@ static int unix_dgram_sendmsg(struct socket *sock, struct msghdr *msg,
  		goto out;
 
  	if (msg->msg_namelen) {
+		if (cgroup_bpf_enabled(CGROUP_UNIX_SENDMSG)) {
+			err = BPF_CGROUP_RUN_PROG_UNIX_SENDMSG_LOCK(sk,
+								    msg->msg_name,
+								    &msg->msg_namelen,
+								    NULL);
+			if (err)
+				goto out;
+		}
+
  		err = unix_validate_addr(sunaddr, msg->msg_namelen);
  		if (err)
  			goto out;
@@ -2328,14 +2381,30 @@ static int unix_seqpacket_recvmsg(struct socket *sock, struct msghdr *msg,
  	return unix_dgram_recvmsg(sock, msg, size, flags);
  }
 
-static void unix_copy_addr(struct msghdr *msg, struct sock *sk)
+static int unix_recvmsg_copy_addr(struct msghdr *msg, struct sock *sk)
  {
  	struct unix_address *addr = smp_load_acquire(&unix_sk(sk)->addr);
+	int err;

net/unix/af_unix.c static int unix_seqpacket_recvmsg(struct socket *sock, struct msghdr *msg, · line 2397

Quoted 10 lines
net/unix/af_unix.c +6 -0
 
  	if (addr) {
  		msg->msg_namelen = addr->len;
  		memcpy(msg->msg_name, addr->name, addr->len);
+
+		if (cgroup_bpf_enabled(CGROUP_UNIX_RECVMSG)) {
+			err = BPF_CGROUP_RUN_PROG_UNIX_RECVMSG_LOCK(sk,
+								    msg->msg_name,
+								    &msg->msg_namelen);
+			if (err)

Same here on the unnecessary err check.

net/unix/af_unix.c static int unix_seqpacket_recvmsg(struct socket *sock, struct msghdr *msg, · line 2401

Quoted 4 lines
net/unix/af_unix.c +4 -0
+				return err;
+
+			err = unix_validate_addr(msg->msg_name,
+						 msg->msg_namelen);

If unix_validate_addr is done in the kfunc, the unix_recvmsg_copy_addr does not need to return error and the changes in the unix_recvmsg_copy_addr's caller is not needed also.

Quoted 33 lines
net/unix/af_unix.c +13 -3
+			if (err)
+				return err;
+		}
  	}
+
+	return 0;
  }
 
  int __unix_dgram_recvmsg(struct sock *sk, struct msghdr *msg, size_t size,
@@ -2390,8 +2459,11 @@ int __unix_dgram_recvmsg(struct sock *sk, struct msghdr *msg, size_t size,
  						EPOLLOUT | EPOLLWRNORM |
  						EPOLLWRBAND);
 
-	if (msg->msg_name)
-		unix_copy_addr(msg, skb->sk);
+	if (msg->msg_name) {
+		err = unix_recvmsg_copy_addr(msg, skb->sk);
+		if (err)
+			goto out_free;
+	}
 
  	if (size > skb->len - skip)
  		size = skb->len - skip;
@@ -2743,7 +2815,9 @@ static int unix_stream_read_generic(struct unix_stream_read_state *state,
  		if (state->msg && state->msg->msg_name) {
  			DECLARE_SOCKADDR(struct sockaddr_un *, sunaddr,
  					 state->msg->msg_name);
-			unix_copy_addr(state->msg, skb->sk);
+			err = unix_recvmsg_copy_addr(state->msg, skb->sk);
+			if (err)
+				break;
  			sunaddr = NULL;
  		}
#10 lore raw

Re: [PATCH bpf-next v3 4/9] bpf: Implement cgroup sockaddr hooks for unix sockets

Quoted 20 lines · Martin KaFai Lau
Quoted 7 lines
net/unix/af_unix.c +1 -1
@@ -1766,14 +1787,37 @@ static int unix_getname(struct socket *sock, struct 
 sockaddr *uaddr, int peer)
       if (!addr) {
           sunaddr->sun_family = AF_UNIX;
           sunaddr->sun_path[0] = 0;
-        err = offsetof(struct sockaddr_un, sun_path);
+        addr_len = offsetof(struct sockaddr_un, sun_path);

net/unix/af_unix.c static int unix_getname(struct socket *sock, struct · line 1800

Quoted 10 lines
net/unix/af_unix.c +6 -1
       } else {
-        err = addr->len;
+        addr_len = addr->len;
           memcpy(sunaddr, addr->name, addr->len);
       }
+
+    if (peer && cgroup_bpf_enabled(CGROUP_UNIX_GETPEERNAME)) {
+        err = BPF_CGROUP_RUN_SA_PROG(sk, uaddr, &addr_len,
+                         CGROUP_UNIX_GETPEERNAME);
+        if (err)

UNIX_GETPEERNAME can only have return value 1 (OK), so no need to do err check here.

Quoted 5 lines

net/unix/af_unix.c static int unix_getname(struct socket *sock, struct sockaddr *uaddr, int peer) · line 1802

Quoted 3 lines
net/unix/af_unix.c +3 -0
+            goto out;
+
+        err = unix_validate_addr(sunaddr, addr_len);

Since the kfunc is specific to the unix address, how about doing the unix_validate_addr check in the kfunc itself?

When reading patch 3 again, the kfunc has already checked the addrlen with the UNIX_PATH_MAX. It should be as good as unix_validate_addr() check considering the kfunc can only change the sunaddr->sun_path?

#11 lore raw

PATCH v3 5/9 libbpf: Add support for cgroup unix socket address hooks

In reply to Daan De Meyer

Add the necessary plumbing to hook up the new cgroup unix sockaddr hooks into libbpf.

---
 tools/lib/bpf/libbpf.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
tools/lib/bpf/libbpf.c +12 -0
index 96ff1aa4bf6a..99075f7f4e11 100644
@@ -80,19 +80,25 @@ static const char * const attach_type_name[] = {
 	[BPF_CGROUP_DEVICE]		= "cgroup_device",
 	[BPF_CGROUP_INET4_BIND]		= "cgroup_inet4_bind",
 	[BPF_CGROUP_INET6_BIND]		= "cgroup_inet6_bind",
+	[BPF_CGROUP_UNIX_BIND]		= "cgroup_unix_bind",
 	[BPF_CGROUP_INET4_CONNECT]	= "cgroup_inet4_connect",
 	[BPF_CGROUP_INET6_CONNECT]	= "cgroup_inet6_connect",
+	[BPF_CGROUP_UNIX_CONNECT]       = "cgroup_unix_connect",
 	[BPF_CGROUP_INET4_POST_BIND]	= "cgroup_inet4_post_bind",
 	[BPF_CGROUP_INET6_POST_BIND]	= "cgroup_inet6_post_bind",
 	[BPF_CGROUP_INET4_GETPEERNAME]	= "cgroup_inet4_getpeername",
 	[BPF_CGROUP_INET6_GETPEERNAME]	= "cgroup_inet6_getpeername",
+	[BPF_CGROUP_UNIX_GETPEERNAME]	= "cgroup_unix_getpeername",
 	[BPF_CGROUP_INET4_GETSOCKNAME]	= "cgroup_inet4_getsockname",
 	[BPF_CGROUP_INET6_GETSOCKNAME]	= "cgroup_inet6_getsockname",
+	[BPF_CGROUP_UNIX_GETSOCKNAME]	= "cgroup_unix_getsockname",
 	[BPF_CGROUP_UDP4_SENDMSG]	= "cgroup_udp4_sendmsg",
 	[BPF_CGROUP_UDP6_SENDMSG]	= "cgroup_udp6_sendmsg",
+	[BPF_CGROUP_UNIX_SENDMSG]	= "cgroup_unix_sendmsg",
 	[BPF_CGROUP_SYSCTL]		= "cgroup_sysctl",
 	[BPF_CGROUP_UDP4_RECVMSG]	= "cgroup_udp4_recvmsg",
 	[BPF_CGROUP_UDP6_RECVMSG]	= "cgroup_udp6_recvmsg",
+	[BPF_CGROUP_UNIX_RECVMSG]	= "cgroup_unix_recvmsg",
 	[BPF_CGROUP_GETSOCKOPT]		= "cgroup_getsockopt",
 	[BPF_CGROUP_SETSOCKOPT]		= "cgroup_setsockopt",
 	[BPF_SK_SKB_STREAM_PARSER]	= "sk_skb_stream_parser",
@@ -8840,16 +8846,22 @@ static const struct bpf_sec_def section_defs[] = {
 	SEC_DEF("cgroup/post_bind6",	CGROUP_SOCK, BPF_CGROUP_INET6_POST_BIND, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/bind4",		CGROUP_SOCK_ADDR, BPF_CGROUP_INET4_BIND, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/bind6",		CGROUP_SOCK_ADDR, BPF_CGROUP_INET6_BIND, SEC_ATTACHABLE),
+	SEC_DEF("cgroup/bindun",	CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_BIND, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/connect4",	CGROUP_SOCK_ADDR, BPF_CGROUP_INET4_CONNECT, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/connect6",	CGROUP_SOCK_ADDR, BPF_CGROUP_INET6_CONNECT, SEC_ATTACHABLE),
+	SEC_DEF("cgroup/connectun",	CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_CONNECT, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/sendmsg4",	CGROUP_SOCK_ADDR, BPF_CGROUP_UDP4_SENDMSG, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/sendmsg6",	CGROUP_SOCK_ADDR, BPF_CGROUP_UDP6_SENDMSG, SEC_ATTACHABLE),
+	SEC_DEF("cgroup/sendmsgun",	CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_SENDMSG, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/recvmsg4",	CGROUP_SOCK_ADDR, BPF_CGROUP_UDP4_RECVMSG, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/recvmsg6",	CGROUP_SOCK_ADDR, BPF_CGROUP_UDP6_RECVMSG, SEC_ATTACHABLE),
+	SEC_DEF("cgroup/recvmsgun",	CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_RECVMSG, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/getpeername4",	CGROUP_SOCK_ADDR, BPF_CGROUP_INET4_GETPEERNAME, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/getpeername6",	CGROUP_SOCK_ADDR, BPF_CGROUP_INET6_GETPEERNAME, SEC_ATTACHABLE),
+	SEC_DEF("cgroup/getpeernameun", CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_GETPEERNAME, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/getsockname4",	CGROUP_SOCK_ADDR, BPF_CGROUP_INET4_GETSOCKNAME, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/getsockname6",	CGROUP_SOCK_ADDR, BPF_CGROUP_INET6_GETSOCKNAME, SEC_ATTACHABLE),
+	SEC_DEF("cgroup/getsocknameun", CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_GETSOCKNAME, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/sysctl",	CGROUP_SYSCTL, BPF_CGROUP_SYSCTL, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/getsockopt",	CGROUP_SOCKOPT, BPF_CGROUP_GETSOCKOPT, SEC_ATTACHABLE),
 	SEC_DEF("cgroup/setsockopt",	CGROUP_SOCKOPT, BPF_CGROUP_SETSOCKOPT, SEC_ATTACHABLE),
Signature · 1 line
2.41.0
#12 lore raw

PATCH v3 6/9 bpftool: Add support for cgroup unix socket address hooks

In reply to Daan De Meyer

Add the necessary plumbing to hook up the new cgroup unix sockaddr hooks into bpftool.

---
 .../bpftool/Documentation/bpftool-cgroup.rst  | 23 ++++++++++++++-----
 .../bpftool/Documentation/bpftool-prog.rst    | 10 ++++----
 tools/bpf/bpftool/bash-completion/bpftool     | 14 +++++------
 tools/bpf/bpftool/cgroup.c                    | 17 ++++++++------
 tools/bpf/bpftool/prog.c                      |  9 ++++----
 5 files changed, 45 insertions(+), 28 deletions(-)
tools/bpf/bpftool/Documentation/bpftool-cgroup.rst +17 -6
index bd015ec9847b..19dba2b55246 100644
@@ -34,14 +34,17 @@ CGROUP COMMANDS
 |	*ATTACH_TYPE* := { **cgroup_inet_ingress** | **cgroup_inet_egress** |
 |		**cgroup_inet_sock_create** | **cgroup_sock_ops** |
 |		**cgroup_device** | **cgroup_inet4_bind** | **cgroup_inet6_bind** |
-|		**cgroup_inet4_post_bind** | **cgroup_inet6_post_bind** |
-|		**cgroup_inet4_connect** | **cgroup_inet6_connect** |
+|		**cgroup_unix_bind** | **cgroup_inet4_post_bind** |
+|		**cgroup_inet6_post_bind** | **cgroup_inet4_connect** |
+|		**cgroup_inet6_connect** | **cgroup_unix_connect** |
 |		**cgroup_inet4_getpeername** | **cgroup_inet6_getpeername** |
-|		**cgroup_inet4_getsockname** | **cgroup_inet6_getsockname** |
+|		**cgroup_unix_getpeername** | **cgroup_inet4_getsockname** |
+|		**cgroup_inet6_getsockname** | **cgroup_unix_getsockname** |
 |		**cgroup_udp4_sendmsg** | **cgroup_udp6_sendmsg** |
-|		**cgroup_udp4_recvmsg** | **cgroup_udp6_recvmsg** |
-|		**cgroup_sysctl** | **cgroup_getsockopt** | **cgroup_setsockopt** |
-|		**cgroup_inet_sock_release** }
+|		**cgroup_unix_sendmsg** | **cgroup_udp4_recvmsg** |
+|		**cgroup_udp6_recvmsg** | **cgroup_unix_recvmsg** |
+|		**cgroup_sysctl** | **cgroup_getsockopt** |
+|		**cgroup_setsockopt** | **cgroup_inet_sock_release** }
 |	*ATTACH_FLAGS* := { **multi** | **override** }
 
 DESCRIPTION
@@ -98,25 +101,33 @@ DESCRIPTION
 		  **device** device access (since 4.15);
 		  **bind4** call to bind(2) for an inet4 socket (since 4.17);
 		  **bind6** call to bind(2) for an inet6 socket (since 4.17);
+		  **bindun** call to bind(2) for a unix socket (since 6.3);
 		  **post_bind4** return from bind(2) for an inet4 socket (since 4.17);
 		  **post_bind6** return from bind(2) for an inet6 socket (since 4.17);
 		  **connect4** call to connect(2) for an inet4 socket (since 4.17);
 		  **connect6** call to connect(2) for an inet6 socket (since 4.17);
+		  **connectun** call to connect(2) for a unix socket (since 6.3);
 		  **sendmsg4** call to sendto(2), sendmsg(2), sendmmsg(2) for an
 		  unconnected udp4 socket (since 4.18);
 		  **sendmsg6** call to sendto(2), sendmsg(2), sendmmsg(2) for an
 		  unconnected udp6 socket (since 4.18);
+		  **sendmsgun** call to sendto(2), sendmsg(2), sendmmsg(2) for
+		  an unconnected unix socket (since 6.3);
 		  **recvmsg4** call to recvfrom(2), recvmsg(2), recvmmsg(2) for
 		  an unconnected udp4 socket (since 5.2);
 		  **recvmsg6** call to recvfrom(2), recvmsg(2), recvmmsg(2) for
 		  an unconnected udp6 socket (since 5.2);
+		  **recvmsgun** call to recvfrom(2), recvmsg(2), recvmmsg(2) for
+		  an unconnected unix socket (since 6.3);
 		  **sysctl** sysctl access (since 5.2);
 		  **getsockopt** call to getsockopt (since 5.3);
 		  **setsockopt** call to setsockopt (since 5.3);
 		  **getpeername4** call to getpeername(2) for an inet4 socket (since 5.8);
 		  **getpeername6** call to getpeername(2) for an inet6 socket (since 5.8);
+		  **getpeernameun** call to getpeername(2) for a unix socket (since 6.3);
 		  **getsockname4** call to getsockname(2) for an inet4 socket (since 5.8);
 		  **getsockname6** call to getsockname(2) for an inet6 socket (since 5.8).
+		  **getsocknameun** call to getsockname(2) for a unix socket (since 6.3);
 		  **sock_release** closing an userspace inet socket (since 5.9).
 
 	**bpftool cgroup detach** *CGROUP* *ATTACH_TYPE* *PROG*
tools/bpf/bpftool/Documentation/bpftool-prog.rst +6 -4
index dcae81bd27ed..886529dd60f0 100644
@@ -46,10 +46,12 @@ PROG COMMANDS
 |		**tracepoint** | **raw_tracepoint** | **xdp** | **perf_event** | **cgroup/skb** |
 |		**cgroup/sock** | **cgroup/dev** | **lwt_in** | **lwt_out** | **lwt_xmit** |
 |		**lwt_seg6local** | **sockops** | **sk_skb** | **sk_msg** | **lirc_mode2** |
-|		**cgroup/bind4** | **cgroup/bind6** | **cgroup/post_bind4** | **cgroup/post_bind6** |
-|		**cgroup/connect4** | **cgroup/connect6** | **cgroup/getpeername4** | **cgroup/getpeername6** |
-|               **cgroup/getsockname4** | **cgroup/getsockname6** | **cgroup/sendmsg4** | **cgroup/sendmsg6** |
-|		**cgroup/recvmsg4** | **cgroup/recvmsg6** | **cgroup/sysctl** |
+|		**cgroup/bind4** | **cgroup/bind6** | **cgroup/bindun** | **cgroup/post_bind4** |
+|		**cgroup/post_bind6** | **cgroup/connect4** | **cgroup/connect6** | **cgroup/connectun** |
+|		**cgroup/getpeername4** | **cgroup/getpeername6** | **cgroup/getpeernameun** |
+|		**cgroup/getsockname4** | **cgroup/getsockname6** | **cgroup/getsocknameun** |
+|		**cgroup/sendmsg4** | **cgroup/sendmsg6** | **cgroup/sendmsgun** |
+|		**cgroup/recvmsg4** | **cgroup/recvmsg6** | **cgroup/recvmsgun** | **cgroup/sysctl** |
 |		**cgroup/getsockopt** | **cgroup/setsockopt** | **cgroup/sock_release** |
 |		**struct_ops** | **fentry** | **fexit** | **freplace** | **sk_lookup**
 |	}
tools/bpf/bpftool/bash-completion/bpftool +7 -7
index 085bf18f3659..2946749e65bb 100644
@@ -480,13 +480,13 @@ _bpftool()
                                 action tracepoint raw_tracepoint \
                                 xdp perf_event cgroup/skb cgroup/sock \
                                 cgroup/dev lwt_in lwt_out lwt_xmit \
-                                lwt_seg6local sockops sk_skb sk_msg \
-                                lirc_mode2 cgroup/bind4 cgroup/bind6 \
-                                cgroup/connect4 cgroup/connect6 \
-                                cgroup/getpeername4 cgroup/getpeername6 \
-                                cgroup/getsockname4 cgroup/getsockname6 \
-                                cgroup/sendmsg4 cgroup/sendmsg6 \
-                                cgroup/recvmsg4 cgroup/recvmsg6 \
+                                lwt_seg6local sockops sk_skb sk_msg lirc_mode2 \
+                                cgroup/bind4 cgroup/bind6 cgroup/bindun \
+                                cgroup/connect4 cgroup/connect6 cgroup/connectun \
+                                cgroup/getpeername4 cgroup/getpeername6 cgroup/getpeernameun \
+                                cgroup/getsockname4 cgroup/getsockname6 cgroup/getsocknameun \
+                                cgroup/sendmsg4 cgroup/sendmsg6 cgroup/sendmsgun \
+                                cgroup/recvmsg4 cgroup/recvmsg6 cgroup/recvmsgun \
                                 cgroup/post_bind4 cgroup/post_bind6 \
                                 cgroup/sysctl cgroup/getsockopt \
                                 cgroup/setsockopt cgroup/sock_release struct_ops \
tools/bpf/bpftool/cgroup.c +10 -7
index ac846b0805b4..a9700e00064c 100644
@@ -26,13 +26,16 @@
 	"       ATTACH_TYPE := { cgroup_inet_ingress | cgroup_inet_egress |\n" \
 	"                        cgroup_inet_sock_create | cgroup_sock_ops |\n" \
 	"                        cgroup_device | cgroup_inet4_bind |\n" \
-	"                        cgroup_inet6_bind | cgroup_inet4_post_bind |\n" \
-	"                        cgroup_inet6_post_bind | cgroup_inet4_connect |\n" \
-	"                        cgroup_inet6_connect | cgroup_inet4_getpeername |\n" \
-	"                        cgroup_inet6_getpeername | cgroup_inet4_getsockname |\n" \
-	"                        cgroup_inet6_getsockname | cgroup_udp4_sendmsg |\n" \
-	"                        cgroup_udp6_sendmsg | cgroup_udp4_recvmsg |\n" \
-	"                        cgroup_udp6_recvmsg | cgroup_sysctl |\n" \
+	"                        cgroup_inet6_bind | cgroup_unix_bind |\n" \
+	"                        cgroup_inet4_post_bind | cgroup_inet6_post_bind |\n" \
+	"                        cgroup_inet4_connect | cgroup_inet6_connect |\n" \
+	"                        cgroup_unix_connect | cgroup_inet4_getpeername |\n" \
+	"                        cgroup_inet6_getpeername | cgroup_unix_getpeername |\n" \
+	"                        cgroup_inet4_getsockname | cgroup_inet6_getsockname |\n" \
+	"                        cgroup_unix_getsockname | cgroup_udp4_sendmsg |\n" \
+	"                        cgroup_udp6_sendmsg | cgroup_unix_sendmsg |\n" \
+	"                        cgroup_udp4_recvmsg | cgroup_udp6_recvmsg |\n" \
+	"                        cgroup_unix_recvmsg | cgroup_sysctl |\n" \
 	"                        cgroup_getsockopt | cgroup_setsockopt |\n" \
 	"                        cgroup_inet_sock_release }"
 
tools/bpf/bpftool/prog.c +5 -4
index 8443a149dd17..8bfd839cee7a 100644
@@ -2473,11 +2473,12 @@ static int do_help(int argc, char **argv)
 		"                 cgroup/sock | cgroup/dev | lwt_in | lwt_out | lwt_xmit |\n"
 		"                 lwt_seg6local | sockops | sk_skb | sk_msg | lirc_mode2 |\n"
 		"                 sk_reuseport | flow_dissector | cgroup/sysctl |\n"
-		"                 cgroup/bind4 | cgroup/bind6 | cgroup/post_bind4 |\n"
+		"                 cgroup/bind4 | cgroup/bind6 | cgroup/bindun | cgroup/post_bind4 |\n"
 		"                 cgroup/post_bind6 | cgroup/connect4 | cgroup/connect6 |\n"
-		"                 cgroup/getpeername4 | cgroup/getpeername6 |\n"
-		"                 cgroup/getsockname4 | cgroup/getsockname6 | cgroup/sendmsg4 |\n"
-		"                 cgroup/sendmsg6 | cgroup/recvmsg4 | cgroup/recvmsg6 |\n"
+		"                 cgroup/connectun | cgroup/getpeername4 | cgroup/getpeername6 |\n"
+		"                 cgroup/getpeernameun | cgroup/getsockname4 | cgroup/getsockname6 |\n"
+		"                 cgroup/getsocknameun | cgroup/sendmsg4 | cgroup/sendmsg6 |\n"
+		"                 cgroup/sendmsgun | cgroup/recvmsg4 | cgroup/recvmsg6 | cgroup/recvmsgun |\n"
 		"                 cgroup/getsockopt | cgroup/setsockopt | cgroup/sock_release |\n"
 		"                 struct_ops | fentry | fexit | freplace | sk_lookup }\n"
 		"       ATTACH_TYPE := { sk_msg_verdict | sk_skb_verdict | sk_skb_stream_verdict |\n"
Signature · 1 line
2.41.0
Acked-by on v3 6/9
#13 lore raw

Re: [PATCH bpf-next v3 6/9] bpftool: Add support for cgroup unix socket address hooks

Quoted 18 lines · Daan De Meyer

Add the necessary plumbing to hook up the new cgroup unix sockaddr hooks into bpftool.

---
 .../bpftool/Documentation/bpftool-cgroup.rst  | 23 ++++++++++++++-----
 .../bpftool/Documentation/bpftool-prog.rst    | 10 ++++----
 tools/bpf/bpftool/bash-completion/bpftool     | 14 +++++------
 tools/bpf/bpftool/cgroup.c                    | 17 ++++++++------
 tools/bpf/bpftool/prog.c                      |  9 ++++----
 5 files changed, 45 insertions(+), 28 deletions(-)
tools/bpf/bpftool/Documentation/bpftool-cgroup.rst +0 -0
index bd015ec9847b..19dba2b55246 100644

tools/bpf/bpftool/Documentation/bpftool-cgroup.rst DESCRIPTION · line 104

Quoted 4 lines
tools/bpf/bpftool/Documentation/bpftool-cgroup.rst +1 -0
@@ -98,25 +101,33 @@ DESCRIPTION
 		  **device** device access (since 4.15);
 		  **bind4** call to bind(2) for an inet4 socket (since 4.17);
 		  **bind6** call to bind(2) for an inet6 socket (since 4.17);
+		  **bindun** call to bind(2) for a unix socket (since 6.3);

I missed it earlier - kernel version (6.3) won't be correct, please update it for the next iteration.

Bpftool changes look all good otherwise, thank you!

#14 lore raw

PATCH v3 7/9 documentation/bpf: Document cgroup unix socket address hooks

In reply to Daan De Meyer

Update the documentation to mention the new cgroup unix sockaddr hooks.

---
 Documentation/bpf/libbpf/program_types.rst | 12 ++++++++++++
 1 file changed, 12 insertions(+)
Documentation/bpf/libbpf/program_types.rst +12 -0
index ad4d4d5eecb0..06168ad73d5e 100644
@@ -56,6 +56,18 @@ described in more detail in the footnotes.
 |                                           | ``BPF_CGROUP_UDP6_RECVMSG``            | ``cgroup/recvmsg6``              |           |
 +                                           +----------------------------------------+----------------------------------+-----------+
 |                                           | ``BPF_CGROUP_UDP6_SENDMSG``            | ``cgroup/sendmsg6``              |           |
+|                                           +----------------------------------------+----------------------------------+-----------+
+|                                           | ``BPF_CGROUP_UNIX_BIND``               | ``cgroup/bindun``                |           |
+|                                           +----------------------------------------+----------------------------------+-----------+
+|                                           | ``BPF_CGROUP_UNIX_CONNECT``            | ``cgroup/connectun``             |           |
+|                                           +----------------------------------------+----------------------------------+-----------+
+|                                           | ``BPF_CGROUP_UNIX_SENDMSG``            | ``cgroup/sendmsgun``             |           |
+|                                           +----------------------------------------+----------------------------------+-----------+
+|                                           | ``BPF_CGROUP_UNIX_RECVMSG``            | ``cgroup/recvmsgun``             |           |
+|                                           +----------------------------------------+----------------------------------+-----------+
+|                                           | ``BPF_CGROUP_UNIX_GETPEERNAME``        | ``cgroup/getpeernameun``         |           |
+|                                           +----------------------------------------+----------------------------------+-----------+
+|                                           | ``BPF_CGROUP_UNIX_GETSOCKNAME``        | ``cgroup/getsocknameun``         |           |
 +-------------------------------------------+----------------------------------------+----------------------------------+-----------+
 | ``BPF_PROG_TYPE_CGROUP_SOCK``             | ``BPF_CGROUP_INET4_POST_BIND``         | ``cgroup/post_bind4``            |           |
 +                                           +----------------------------------------+----------------------------------+-----------+
Signature · 1 line
2.41.0
#15 lore raw

PATCH v3 8/9 selftests/bpf: Make sure mount directory exists

In reply to Daan De Meyer

The mount directory for the selftests cgroup tree might not exist so let's make sure it does exist by creating it ourselves if it doesn't exist.

---
 tools/testing/selftests/bpf/cgroup_helpers.c | 5 +++++
 1 file changed, 5 insertions(+)
tools/testing/selftests/bpf/cgroup_helpers.c +5 -0
index 2caee8423ee0..6dcf0cd375c4 100644
@@ -195,6 +195,11 @@ int setup_cgroup_environment(void)
 
 	format_cgroup_path(cgroup_workdir, "");
 
+	if (mkdir(CGROUP_MOUNT_PATH, 0777) && errno != EEXIST) {
+		log_err("mkdir mount");
+		return 1;
+	}
+
 	if (unshare(CLONE_NEWNS)) {
 		log_err("unshare");
 		return 1;
Signature · 1 line
2.41.0
#16 lore raw

PATCH v3 9/9 selftests/bpf: Add tests for cgroup unix socket address hooks

In reply to Daan De Meyer

These selftests are written in prog_tests style instead of adding them to the existing test_sock_addr tests. Migrating the existing sock addr tests to prog_tests style is left for future work.

---
 tools/testing/selftests/bpf/bpf_kfuncs.h      |  14 +
 tools/testing/selftests/bpf/network_helpers.c |  34 ++
 tools/testing/selftests/bpf/network_helpers.h |   1 +
 .../selftests/bpf/prog_tests/section_names.c  |  30 ++
 .../selftests/bpf/prog_tests/sock_addr.c      | 461 ++++++++++++++++++
 .../testing/selftests/bpf/progs/bindun_prog.c |  39 ++
 .../selftests/bpf/progs/connectun_prog.c      |  40 ++
 .../selftests/bpf/progs/recvmsgun_prog.c      |  39 ++
 .../selftests/bpf/progs/sendmsgun_prog.c      |  40 ++
 9 files changed, 698 insertions(+)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/sock_addr.c
 create mode 100644 tools/testing/selftests/bpf/progs/bindun_prog.c
 create mode 100644 tools/testing/selftests/bpf/progs/connectun_prog.c
 create mode 100644 tools/testing/selftests/bpf/progs/recvmsgun_prog.c
 create mode 100644 tools/testing/selftests/bpf/progs/sendmsgun_prog.c
tools/testing/selftests/bpf/bpf_kfuncs.h +14 -0
index 642dda0e758a..aa376792c4a4 100644
@@ -1,6 +1,8 @@
 #ifndef __BPF_KFUNCS__
 #define __BPF_KFUNCS__
 
+struct bpf_sock_addr_kern;
+
 /* Description
  *  Initializes an skb-type dynptr
  * Returns
@@ -41,4 +43,16 @@ extern bool bpf_dynptr_is_rdonly(const struct bpf_dynptr *ptr) __ksym;
 extern __u32 bpf_dynptr_size(const struct bpf_dynptr *ptr) __ksym;
 extern int bpf_dynptr_clone(const struct bpf_dynptr *ptr, struct bpf_dynptr *clone__init) __ksym;
 
+/* Description
+ *  Modify the address of a AF_UNIX sockaddr.
+ * Returns__bpf_kfunc
+ *  -EINVAL if the address size is too big or, 0 if the sockaddr was successfully modified.
+ */
+extern int bpf_sock_addr_set_unix_addr(struct bpf_sock_addr_kern *sa_kern,
+				       const __u8 *addr, __u32 addrlen__sz) __ksym;
+
+void *bpf_cast_to_kern_ctx(void *) __ksym;
+
+void *bpf_rdonly_cast(void *obj, __u32 btf_id) __ksym;
+
 #endif
tools/testing/selftests/bpf/network_helpers.c +34 -0
index da72a3a66230..6db27a9088e9 100644
@@ -11,6 +11,7 @@
 #include <arpa/inet.h>
 #include <sys/mount.h>
 #include <sys/stat.h>
+#include <sys/un.h>
 
 #include <linux/err.h>
 #include <linux/in.h>
@@ -257,6 +258,26 @@ static int connect_fd_to_addr(int fd,
 	return 0;
 }
 
+int connect_to_addr(const struct sockaddr_storage *addr, socklen_t addrlen, int type)
+{
+	int fd;
+
+	fd = socket(addr->ss_family, type, 0);
+	if (fd < 0) {
+		log_err("Failed to create client socket");
+		return -1;
+	}
+
+	if (connect_fd_to_addr(fd, addr, addrlen, false))
+		goto error_close;
+
+	return fd;
+
+error_close:
+	save_errno_close(fd);
+	return -1;
+}
+
 static const struct network_helper_opts default_opts;
 
 int connect_to_fd_opts(int server_fd, const struct network_helper_opts *opts)
@@ -380,6 +401,19 @@ int make_sockaddr(int family, const char *addr_str, __u16 port,
 		if (len)
 			*len = sizeof(*sin6);
 		return 0;
+	} else if (family == AF_UNIX) {
+		/* Note that we always use abstract unix sockets to avoid having
+		 * to clean up leftover files.
+		 */
+		struct sockaddr_un *sun = (void *)addr;
+
+		memset(addr, 0, sizeof(*sun));
+		sun->sun_family = family;
+		sun->sun_path[0] = 0;
+		strcpy(sun->sun_path + 1, addr_str);
+		if (len)
+			*len = offsetof(struct sockaddr_un, sun_path) + 1 + strlen(addr_str);
+		return 0;
 	}
 	return -1;
 }
tools/testing/selftests/bpf/network_helpers.h +1 -0
index 5eccc67d1a99..34f1200a781b 100644
@@ -51,6 +51,7 @@ int *start_reuseport_server(int family, int type, const char *addr_str,
 			    __u16 port, int timeout_ms,
 			    unsigned int nr_listens);
 void free_fds(int *fds, unsigned int nr_close_fds);
+int connect_to_addr(const struct sockaddr_storage *addr, socklen_t len, int type);
 int connect_to_fd(int server_fd, int timeout_ms);
 int connect_to_fd_opts(int server_fd, const struct network_helper_opts *opts);
 int connect_fd_to_fd(int client_fd, int server_fd, int timeout_ms);
tools/testing/selftests/bpf/prog_tests/section_names.c +30 -0
index fc5248e94a01..51ebc8e6065d 100644
@@ -113,6 +113,11 @@ static struct sec_name_test tests[] = {
 		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_INET6_BIND},
 		{0, BPF_CGROUP_INET6_BIND},
 	},
+	{
+		"cgroup/bindun",
+		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_BIND},
+		{0, BPF_CGROUP_UNIX_BIND},
+	},
 	{
 		"cgroup/connect4",
 		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_INET4_CONNECT},
@@ -123,6 +128,11 @@ static struct sec_name_test tests[] = {
 		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_INET6_CONNECT},
 		{0, BPF_CGROUP_INET6_CONNECT},
 	},
+	{
+		"cgroup/connectun",
+		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_CONNECT},
+		{0, BPF_CGROUP_UNIX_CONNECT},
+	},
 	{
 		"cgroup/sendmsg4",
 		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_UDP4_SENDMSG},
@@ -133,6 +143,11 @@ static struct sec_name_test tests[] = {
 		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_UDP6_SENDMSG},
 		{0, BPF_CGROUP_UDP6_SENDMSG},
 	},
+	{
+		"cgroup/sendmsgun",
+		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_SENDMSG},
+		{0, BPF_CGROUP_UNIX_SENDMSG},
+	},
 	{
 		"cgroup/recvmsg4",
 		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_UDP4_RECVMSG},
@@ -143,6 +158,11 @@ static struct sec_name_test tests[] = {
 		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_UDP6_RECVMSG},
 		{0, BPF_CGROUP_UDP6_RECVMSG},
 	},
+	{
+		"cgroup/recvmsgun",
+		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_RECVMSG},
+		{0, BPF_CGROUP_UNIX_RECVMSG},
+	},
 	{
 		"cgroup/sysctl",
 		{0, BPF_PROG_TYPE_CGROUP_SYSCTL, BPF_CGROUP_SYSCTL},
@@ -168,6 +188,11 @@ static struct sec_name_test tests[] = {
 		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_INET6_GETPEERNAME},
 		{0, BPF_CGROUP_INET6_GETPEERNAME},
 	},
+	{
+		"cgroup/getpeernameun",
+		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_GETPEERNAME},
+		{0, BPF_CGROUP_UNIX_GETPEERNAME},
+	},
 	{
 		"cgroup/getsockname4",
 		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_INET4_GETSOCKNAME},
@@ -178,6 +203,11 @@ static struct sec_name_test tests[] = {
 		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_INET6_GETSOCKNAME},
 		{0, BPF_CGROUP_INET6_GETSOCKNAME},
 	},
+	{
+		"cgroup/getsocknameun",
+		{0, BPF_PROG_TYPE_CGROUP_SOCK_ADDR, BPF_CGROUP_UNIX_GETSOCKNAME},
+		{0, BPF_CGROUP_UNIX_GETSOCKNAME},
+	},
 };
 
 static void test_prog_type_by_name(const struct sec_name_test *test)
tools/testing/selftests/bpf/prog_tests/sock_addr.c new file +461 -0

461 lines not on this page · raw message

tools/testing/selftests/bpf/progs/bindun_prog.c new file +39 -0
new file mode 100644
index 000000000000..c7bc93d5a904
@@ -0,0 +1,39 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2022 Meta Platforms, Inc. and affiliates. */
+
+#include "vmlinux.h"
+
+#include <string.h>
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_core_read.h>
+#include "bpf_kfuncs.h"
+
+__u8 SERVUN_REWRITE_ADDRESS[] = "\0bpf_cgroup_unix_test_rewrite";
+
+SEC("cgroup/bindun")
+int bindun_prog(struct bpf_sock_addr *ctx)
+{
+	struct bpf_sock_addr_kern *sa_kern = bpf_cast_to_kern_ctx(ctx);
+	struct sockaddr_un *sa_kern_unaddr;
+	__u32 unaddrlen = offsetof(struct sockaddr_un, sun_path) +
+			  sizeof(SERVUN_REWRITE_ADDRESS) - 1;
+	int ret;
+
+	ret = bpf_sock_addr_set_unix_addr(sa_kern, SERVUN_REWRITE_ADDRESS,
+					  sizeof(SERVUN_REWRITE_ADDRESS) - 1);
+	if (ret)
+		return 0;
+
+	if (sa_kern->uaddrlen != unaddrlen)
+		return 0;
+
+	sa_kern_unaddr = bpf_rdonly_cast(sa_kern->uaddr,
+						bpf_core_type_id_kernel(struct sockaddr_un));
+	if (memcmp(sa_kern_unaddr->sun_path, SERVUN_REWRITE_ADDRESS,
+			sizeof(SERVUN_REWRITE_ADDRESS) - 1) != 0)
+		return 0;
+
+	return 1;
+}
+
+char _license[] SEC("license") = "GPL";
tools/testing/selftests/bpf/progs/connectun_prog.c new file +40 -0
new file mode 100644
index 000000000000..8ab5b5d1fc3b
@@ -0,0 +1,40 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2022 Meta Platforms, Inc. and affiliates. */
+
+#include "vmlinux.h"
+
+#include <string.h>
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_core_read.h>
+#include "bpf_kfuncs.h"
+
+__u8 SERVUN_REWRITE_ADDRESS[] = "\0bpf_cgroup_unix_test_rewrite";
+
+SEC("cgroup/connectun")
+int connectun_prog(struct bpf_sock_addr *ctx)
+{
+	struct bpf_sock_addr_kern *sa_kern = bpf_cast_to_kern_ctx(ctx);
+	struct sockaddr_un *sa_kern_unaddr;
+	__u32 unaddrlen = offsetof(struct sockaddr_un, sun_path) +
+			  sizeof(SERVUN_REWRITE_ADDRESS) - 1;
+	int ret;
+
+	/* Rewrite destination. */
+	ret = bpf_sock_addr_set_unix_addr(sa_kern, SERVUN_REWRITE_ADDRESS,
+					  sizeof(SERVUN_REWRITE_ADDRESS) - 1);
+	if (ret)
+		return 0;
+
+	if (sa_kern->uaddrlen != unaddrlen)
+		return 0;
+
+	sa_kern_unaddr = bpf_rdonly_cast(sa_kern->uaddr,
+						bpf_core_type_id_kernel(struct sockaddr_un));
+	if (memcmp(sa_kern_unaddr->sun_path, SERVUN_REWRITE_ADDRESS,
+			sizeof(SERVUN_REWRITE_ADDRESS) - 1) != 0)
+		return 0;
+
+	return 1;
+}
+
+char _license[] SEC("license") = "GPL";
tools/testing/selftests/bpf/progs/recvmsgun_prog.c new file +39 -0
new file mode 100644
index 000000000000..2baeaebac57e
@@ -0,0 +1,39 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2022 Meta Platforms, Inc. and affiliates. */
+
+#include "vmlinux.h"
+
+#include <string.h>
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_core_read.h>
+#include "bpf_kfuncs.h"
+
+__u8 SERVUN_ADDRESS[] = "\0bpf_cgroup_unix_test";
+
+SEC("cgroup/recvmsgun")
+int recvmsgun_prog(struct bpf_sock_addr *ctx)
+{
+	struct bpf_sock_addr_kern *sa_kern = bpf_cast_to_kern_ctx(ctx);
+	struct sockaddr_un *sa_kern_unaddr;
+	__u32 unaddrlen = offsetof(struct sockaddr_un, sun_path) +
+			  sizeof(SERVUN_ADDRESS) - 1;
+	int ret;
+
+	ret = bpf_sock_addr_set_unix_addr(sa_kern, SERVUN_ADDRESS,
+					  sizeof(SERVUN_ADDRESS) - 1);
+	if (ret)
+		return 1;
+
+	if (sa_kern->uaddrlen != unaddrlen)
+		return 1;
+
+	sa_kern_unaddr = bpf_rdonly_cast(sa_kern->uaddr,
+						bpf_core_type_id_kernel(struct sockaddr_un));
+	if (memcmp(sa_kern_unaddr->sun_path, SERVUN_ADDRESS,
+			sizeof(SERVUN_ADDRESS) - 1) != 0)
+		return 1;
+
+	return 1;
+}
+
+char _license[] SEC("license") = "GPL";
tools/testing/selftests/bpf/progs/sendmsgun_prog.c new file +40 -0
new file mode 100644
index 000000000000..23f2135c6eb0
@@ -0,0 +1,40 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2022 Meta Platforms, Inc. and affiliates. */
+
+#include "vmlinux.h"
+
+#include <string.h>
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_core_read.h>
+#include "bpf_kfuncs.h"
+
+__u8 SERVUN_REWRITE_ADDRESS[] = "\0bpf_cgroup_unix_test_rewrite";
+
+SEC("cgroup/sendmsgun")
+int sendmsgun_prog(struct bpf_sock_addr *ctx)
+{
+	struct bpf_sock_addr_kern *sa_kern = bpf_cast_to_kern_ctx(ctx);
+	struct sockaddr_un *sa_kern_unaddr;
+	__u32 unaddrlen = offsetof(struct sockaddr_un, sun_path) +
+			  sizeof(SERVUN_REWRITE_ADDRESS) - 1;
+	int ret;
+
+	/* Rewrite destination. */
+	ret = bpf_sock_addr_set_unix_addr(sa_kern, SERVUN_REWRITE_ADDRESS,
+					  sizeof(SERVUN_REWRITE_ADDRESS) - 1);
+	if (ret)
+		return 0;
+
+	if (sa_kern->uaddrlen != unaddrlen)
+		return 0;
+
+	sa_kern_unaddr = bpf_rdonly_cast(sa_kern->uaddr,
+						bpf_core_type_id_kernel(struct sockaddr_un));
+	if (memcmp(sa_kern_unaddr->sun_path, SERVUN_REWRITE_ADDRESS,
+			sizeof(SERVUN_REWRITE_ADDRESS) - 1) != 0)
+		return 0;
+
+	return 1;
+}
+
+char _license[] SEC("license") = "GPL";
Signature · 1 line
2.41.0
#17 lore raw

Re: [PATCH bpf-next v3 9/9] selftests/bpf: Add tests for cgroup unix socket address hooks

Quoted 16 lines · Daan De Meyer
tools/testing/selftests/bpf/prog_tests/sock_addr.c new file +10 -0
new file mode 100644
index 000000000000..07941dd48efb
@@ -0,0 +1,461 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <sys/un.h>
+
+#include "test_progs.h"
+
+#include "bindun_prog.skel.h"
+#include "connectun_prog.skel.h"
+#include "sendmsgun_prog.skel.h"
+#include "recvmsgun_prog.skel.h"
+#include "network_helpers.h"

tools/testing/selftests/bpf/prog_tests/sock_addr.c line 20

Quoted 10 lines
tools/testing/selftests/bpf/prog_tests/sock_addr.c new file +10 -0
+
+#define SERVUN_ADDRESS         "bpf_cgroup_unix_test"
+#define SERVUN_REWRITE_ADDRESS "bpf_cgroup_unix_test_rewrite"
+#define SRCUN_ADDRESS	       "bpf_cgroup_unix_test_src"
+
+enum sock_addr_test_type {
+	SOCK_ADDR_TEST_BIND,
+	SOCK_ADDR_TEST_CONNECT,
+	SOCK_ADDR_TEST_SENDMSG,
+	SOCK_ADDR_TEST_RECVMSG,

Tests are also needed for cgroup/get{sock,peer}nameun.