[PATCH bpf v2 0/2] sockmap fix for KASAN_VMALLOC and af_unix

[PATCH bpf v2 0/2] sockmap fix for KASAN_VMALLOC and af_unix

[John Fastabend]

The af_unix tests in sockmap_listen causes a splat from KASAN_VMALLOC.
Fix it here and include an extra test to catch case where both pairs
of the af_unix socket are included in a BPF sockmap.

Also it seems the test infra is not passing type through correctly when
testing unix_inet_redir_to_connected. Unfortunately, the simple fix
also caused some CI tests to fail so investigating that now.

v2: drop changes to dgram side its fine per Jakub's point it graps a
    reference on the peer socket from each sendmsg.

John Fastabend (2):
  bpf: sockmap, af_unix stream sockets need to hold ref for pair sock
  bpf: sockmap, add af_unix test with both sockets in map

 include/linux/skmsg.h                         |  1 +
 include/net/af_unix.h                         |  1 +
 net/core/skmsg.c                              |  2 +
 net/unix/af_unix.c                            |  2 -
 net/unix/unix_bpf.c                           |  5 +++
 .../selftests/bpf/prog_tests/sockmap_listen.c | 39 ++++++++++++++++---
 .../selftests/bpf/progs/test_sockmap_listen.c |  7 ++++
 7 files changed, 49 insertions(+), 8 deletions(-)

-- 
2.33.0

[PATCH bpf v2 1/2] bpf: sockmap, af_unix stream sockets need to hold ref for pair sock

[John Fastabend]

AF_UNIX stream sockets are a paired socket. So sending on one of the pairs
will lookup the paired socket as part of the send operation. It is possible
however to put just one of the pairs in a BPF map. This currently
increments the refcnt on the sock in the sockmap to ensure it is not
free'd by the stack before sockmap cleans up its state and stops any
skbs being sent/recv'd to that socket.

But we missed a case. If the peer socket is closed it will be
free'd by the stack. However, the paired socket can still be
referenced from BPF sockmap side because we hold a reference
there. Then if we are sending traffic through BPF sockmap to
that socket it will try to dereference the free'd pair in its
send logic creating a use after free.  And following splat,

   [59.900375] BUG: KASAN: slab-use-after-free in sk_wake_async+0x31/0x1b0
   [59.901211] Read of size 8 at addr ffff88811acbf060 by task kworker/1:2/954
   [...]
   [59.905468] Call Trace:
   [59.905787]  <TASK>
   [59.906066]  dump_stack_lvl+0x130/0x1d0
   [59.908877]  print_report+0x16f/0x740
   [59.910629]  kasan_report+0x118/0x160
   [59.912576]  sk_wake_async+0x31/0x1b0
   [59.913554]  sock_def_readable+0x156/0x2a0
   [59.914060]  unix_stream_sendmsg+0x3f9/0x12a0
   [59.916398]  sock_sendmsg+0x20e/0x250
   [59.916854]  skb_send_sock+0x236/0xac0
   [59.920527]  sk_psock_backlog+0x287/0xaa0

To fix let BPF sockmap hold a refcnt on both the socket in the
sockmap and its paired socket.  It wasn't obvious how to contain
the fix to bpf_unix logic. The primarily problem with keeping this
logic in bpf_unix was: In the sock close() we could handle the
deref by having a close handler. But, when we are destroying the
psock through a map delete operation we wouldn't have gotten any
signal thorugh the proto struct other than it being replaced.
If we do the deref from the proto replace its too early because
we need to deref the skpair after the backlog worker has been
stopped.

Given all this it seems best to just cache it at the end of the
psock and eat 8B for the af_unix and vsock users.

Notice dgram sockets are OK because they handle locking already.

Fixes: 94531cfcbe79 ("af_unix: Add unix_stream_proto for sockmap")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
---
 include/linux/skmsg.h | 1 +
 include/net/af_unix.h | 1 +
 net/core/skmsg.c      | 2 ++
 net/unix/af_unix.c    | 2 --
 net/unix/unix_bpf.c   | 5 +++++
 5 files changed, 9 insertions(+), 2 deletions(-)

diff --git a/include/linux/skmsg.h b/include/linux/skmsg.h
index c1637515a8a4..fbe628961cf8 100644
--- a/include/linux/skmsg.h
+++ b/include/linux/skmsg.h
@@ -106,6 +106,7 @@ struct sk_psock {
 	struct mutex			work_mutex;
 	struct sk_psock_work_state	work_state;
 	struct delayed_work		work;
+	struct sock			*skpair;
 	struct rcu_work			rwork;
 };
 
diff --git a/include/net/af_unix.h b/include/net/af_unix.h
index 824c258143a3..49c4640027d8 100644
--- a/include/net/af_unix.h
+++ b/include/net/af_unix.h
@@ -75,6 +75,7 @@ struct unix_sock {
 };
 
 #define unix_sk(ptr) container_of_const(ptr, struct unix_sock, sk)
+#define unix_peer(sk) (unix_sk(sk)->peer)
 
 #define peer_wait peer_wq.wait
 
diff --git a/net/core/skmsg.c b/net/core/skmsg.c
index 6c31eefbd777..6236164b9bce 100644
--- a/net/core/skmsg.c
+++ b/net/core/skmsg.c
@@ -826,6 +826,8 @@ static void sk_psock_destroy(struct work_struct *work)
 
 	if (psock->sk_redir)
 		sock_put(psock->sk_redir);
+	if (psock->skpair)
+		sock_put(psock->skpair);
 	sock_put(psock->sk);
 	kfree(psock);
 }
diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
index 3e8a04a13668..87dd723aacf9 100644
--- a/net/unix/af_unix.c
+++ b/net/unix/af_unix.c
@@ -212,8 +212,6 @@ static inline bool unix_secdata_eq(struct scm_cookie *scm, struct sk_buff *skb)
 }
 #endif /* CONFIG_SECURITY_NETWORK */
 
-#define unix_peer(sk) (unix_sk(sk)->peer)
-
 static inline int unix_our_peer(struct sock *sk, struct sock *osk)
 {
 	return unix_peer(osk) == sk;
diff --git a/net/unix/unix_bpf.c b/net/unix/unix_bpf.c
index 2f9d8271c6ec..7e7791029198 100644
--- a/net/unix/unix_bpf.c
+++ b/net/unix/unix_bpf.c
@@ -159,12 +159,17 @@ int unix_dgram_bpf_update_proto(struct sock *sk, struct sk_psock *psock, bool re
 
 int unix_stream_bpf_update_proto(struct sock *sk, struct sk_psock *psock, bool restore)
 {
+	struct sock *skpair = unix_peer(sk);
+
 	if (restore) {
 		sk->sk_write_space = psock->saved_write_space;
 		sock_replace_proto(sk, psock->sk_proto);
 		return 0;
 	}
 
+	skpair = unix_peer(sk);
+	sock_hold(skpair);
+	psock->skpair = skpair;
 	unix_stream_bpf_check_needs_rebuild(psock->sk_proto);
 	sock_replace_proto(sk, &unix_stream_bpf_prot);
 	return 0;
-- 
2.33.0

Re: [PATCH bpf v2 1/2] bpf: sockmap, af_unix stream sockets need to hold ref for pair sock

[Yonghong Song]

On 11/22/23 2:24 PM, John Fastabend wrote:
> AF_UNIX stream sockets are a paired socket. So sending on one of the pairs
> will lookup the paired socket as part of the send operation. It is possible
> however to put just one of the pairs in a BPF map. This currently
> increments the refcnt on the sock in the sockmap to ensure it is not
> free'd by the stack before sockmap cleans up its state and stops any
> skbs being sent/recv'd to that socket.
>
> But we missed a case. If the peer socket is closed it will be
> free'd by the stack. However, the paired socket can still be
> referenced from BPF sockmap side because we hold a reference
> there. Then if we are sending traffic through BPF sockmap to
> that socket it will try to dereference the free'd pair in its
> send logic creating a use after free.  And following splat,
>
>     [59.900375] BUG: KASAN: slab-use-after-free in sk_wake_async+0x31/0x1b0
>     [59.901211] Read of size 8 at addr ffff88811acbf060 by task kworker/1:2/954
>     [...]
>     [59.905468] Call Trace:
>     [59.905787]  <TASK>
>     [59.906066]  dump_stack_lvl+0x130/0x1d0
>     [59.908877]  print_report+0x16f/0x740
>     [59.910629]  kasan_report+0x118/0x160
>     [59.912576]  sk_wake_async+0x31/0x1b0
>     [59.913554]  sock_def_readable+0x156/0x2a0
>     [59.914060]  unix_stream_sendmsg+0x3f9/0x12a0
>     [59.916398]  sock_sendmsg+0x20e/0x250
>     [59.916854]  skb_send_sock+0x236/0xac0
>     [59.920527]  sk_psock_backlog+0x287/0xaa0
>
> To fix let BPF sockmap hold a refcnt on both the socket in the
> sockmap and its paired socket.  It wasn't obvious how to contain
> the fix to bpf_unix logic. The primarily problem with keeping this
> logic in bpf_unix was: In the sock close() we could handle the
> deref by having a close handler. But, when we are destroying the
> psock through a map delete operation we wouldn't have gotten any
> signal thorugh the proto struct other than it being replaced.
> If we do the deref from the proto replace its too early because
> we need to deref the skpair after the backlog worker has been
> stopped.
>
> Given all this it seems best to just cache it at the end of the
> psock and eat 8B for the af_unix and vsock users.
>
> Notice dgram sockets are OK because they handle locking already.
>
> Fixes: 94531cfcbe79 ("af_unix: Add unix_stream_proto for sockmap")
> Signed-off-by: John Fastabend <john.fastabend@gmail.com>
> ---
>   include/linux/skmsg.h | 1 +
>   include/net/af_unix.h | 1 +
>   net/core/skmsg.c      | 2 ++
>   net/unix/af_unix.c    | 2 --
>   net/unix/unix_bpf.c   | 5 +++++
>   5 files changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/include/linux/skmsg.h b/include/linux/skmsg.h
> index c1637515a8a4..fbe628961cf8 100644
> --- a/include/linux/skmsg.h
> +++ b/include/linux/skmsg.h
> @@ -106,6 +106,7 @@ struct sk_psock {
>   	struct mutex			work_mutex;
>   	struct sk_psock_work_state	work_state;
>   	struct delayed_work		work;
> +	struct sock			*skpair;
>   	struct rcu_work			rwork;
>   };
>   
> diff --git a/include/net/af_unix.h b/include/net/af_unix.h
> index 824c258143a3..49c4640027d8 100644
> --- a/include/net/af_unix.h
> +++ b/include/net/af_unix.h
> @@ -75,6 +75,7 @@ struct unix_sock {
>   };
>   
>   #define unix_sk(ptr) container_of_const(ptr, struct unix_sock, sk)
> +#define unix_peer(sk) (unix_sk(sk)->peer)
>   
>   #define peer_wait peer_wq.wait
>   
> diff --git a/net/core/skmsg.c b/net/core/skmsg.c
> index 6c31eefbd777..6236164b9bce 100644
> --- a/net/core/skmsg.c
> +++ b/net/core/skmsg.c
> @@ -826,6 +826,8 @@ static void sk_psock_destroy(struct work_struct *work)
>   
>   	if (psock->sk_redir)
>   		sock_put(psock->sk_redir);
> +	if (psock->skpair)
> +		sock_put(psock->skpair);
>   	sock_put(psock->sk);
>   	kfree(psock);
>   }
> diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
> index 3e8a04a13668..87dd723aacf9 100644
> --- a/net/unix/af_unix.c
> +++ b/net/unix/af_unix.c
> @@ -212,8 +212,6 @@ static inline bool unix_secdata_eq(struct scm_cookie *scm, struct sk_buff *skb)
>   }
>   #endif /* CONFIG_SECURITY_NETWORK */
>   
> -#define unix_peer(sk) (unix_sk(sk)->peer)
> -
>   static inline int unix_our_peer(struct sock *sk, struct sock *osk)
>   {
>   	return unix_peer(osk) == sk;
> diff --git a/net/unix/unix_bpf.c b/net/unix/unix_bpf.c
> index 2f9d8271c6ec..7e7791029198 100644
> --- a/net/unix/unix_bpf.c
> +++ b/net/unix/unix_bpf.c
> @@ -159,12 +159,17 @@ int unix_dgram_bpf_update_proto(struct sock *sk, struct sk_psock *psock, bool re
>   
>   int unix_stream_bpf_update_proto(struct sock *sk, struct sk_psock *psock, bool restore)
>   {
> +	struct sock *skpair = unix_peer(sk);

The above assignment is redundant.

> +
>   	if (restore) {
>   		sk->sk_write_space = psock->saved_write_space;
>   		sock_replace_proto(sk, psock->sk_proto);
>   		return 0;
>   	}
>   
> +	skpair = unix_peer(sk);
> +	sock_hold(skpair);
> +	psock->skpair = skpair;
>   	unix_stream_bpf_check_needs_rebuild(psock->sk_proto);
>   	sock_replace_proto(sk, &unix_stream_bpf_prot);
>   	return 0;

Re: [PATCH bpf v2 1/2] bpf: sockmap, af_unix stream sockets need to hold ref for pair sock

[Cong Wang]

On Wed, Nov 22, 2023 at 11:24:51AM -0800, John Fastabend wrote:
> AF_UNIX stream sockets are a paired socket. So sending on one of the pairs
> will lookup the paired socket as part of the send operation. It is possible
> however to put just one of the pairs in a BPF map. This currently
> increments the refcnt on the sock in the sockmap to ensure it is not
> free'd by the stack before sockmap cleans up its state and stops any
> skbs being sent/recv'd to that socket.
> 
> But we missed a case. If the peer socket is closed it will be
> free'd by the stack. However, the paired socket can still be
> referenced from BPF sockmap side because we hold a reference
> there. Then if we are sending traffic through BPF sockmap to
> that socket it will try to dereference the free'd pair in its
> send logic creating a use after free.  And following splat,

Hmm, how could it pass the SOCK_DEAD test in unix_stream_sendmsg()?

2285                 unix_state_lock(other);
2286
2287                 if (sock_flag(other, SOCK_DEAD) ||
2288                     (other->sk_shutdown & RCV_SHUTDOWN))
2289                         goto pipe_err_free;


Thanks.

Re: [PATCH bpf v2 1/2] bpf: sockmap, af_unix stream sockets need to hold ref for pair sock

[Jakub Sitnicki]

On Thu, Nov 23, 2023 at 12:24 PM -08, Cong Wang wrote:
> On Wed, Nov 22, 2023 at 11:24:51AM -0800, John Fastabend wrote:
>> AF_UNIX stream sockets are a paired socket. So sending on one of the pairs
>> will lookup the paired socket as part of the send operation. It is possible
>> however to put just one of the pairs in a BPF map. This currently
>> increments the refcnt on the sock in the sockmap to ensure it is not
>> free'd by the stack before sockmap cleans up its state and stops any
>> skbs being sent/recv'd to that socket.
>> 
>> But we missed a case. If the peer socket is closed it will be
>> free'd by the stack. However, the paired socket can still be
>> referenced from BPF sockmap side because we hold a reference
>> there. Then if we are sending traffic through BPF sockmap to
>> that socket it will try to dereference the free'd pair in its
>> send logic creating a use after free.  And following splat,
>
> Hmm, how could it pass the SOCK_DEAD test in unix_stream_sendmsg()?
>
> 2285                 unix_state_lock(other);
> 2286
> 2287                 if (sock_flag(other, SOCK_DEAD) ||
> 2288                     (other->sk_shutdown & RCV_SHUTDOWN))
> 2289                         goto pipe_err_free;

The quoted UAF happens after unix_state_unlock(other):

  2285                  unix_state_lock(other);
  2286
  2287                  if (sock_flag(other, SOCK_DEAD) ||
  2288                      (other->sk_shutdown & RCV_SHUTDOWN))
  2289                          goto pipe_err_free;
  2290
  2291                  maybe_add_creds(skb, sock, other);
  2292                  scm_stat_add(other, skb);
  2293                  skb_queue_tail(&other->sk_receive_queue, skb);
  2294                  unix_state_unlock(other);
  2295                  other->sk_data_ready(other); <-- UAF

Although, I think I saw it happen at unix_state_lock(other) as well.

We don't hold a ref on other, so we're racing with __sock_release /
unix_release_sock.

[PATCH bpf v2 2/2] bpf: sockmap, add af_unix test with both sockets in map

[John Fastabend]

This adds a test where both pairs of a af_unix paired socket are put into
a BPF map. This ensures that when we tear down the af_unix pair we don't
have any issues on sockmap side with ordering and reference counting.

Signed-off-by: John Fastabend <john.fastabend@gmail.com>
---
 .../selftests/bpf/prog_tests/sockmap_listen.c | 39 ++++++++++++++++---
 .../selftests/bpf/progs/test_sockmap_listen.c |  7 ++++
 2 files changed, 40 insertions(+), 6 deletions(-)

diff --git a/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c b/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
index 8df8cbb447f1..90e97907c1c1 100644
--- a/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
+++ b/tools/testing/selftests/bpf/prog_tests/sockmap_listen.c
@@ -1824,8 +1824,10 @@ static void inet_unix_skb_redir_to_connected(struct test_sockmap_listen *skel,
 	xbpf_prog_detach2(verdict, sock_map, BPF_SK_SKB_VERDICT);
 }
 
-static void unix_inet_redir_to_connected(int family, int type, int sock_mapfd,
-					int verd_mapfd, enum redir_mode mode)
+static void unix_inet_redir_to_connected(int family, int type,
+					int sock_mapfd, int nop_mapfd,
+					int verd_mapfd,
+					enum redir_mode mode)
 {
 	const char *log_prefix = redir_mode_str(mode);
 	int c0, c1, p0, p1;
@@ -1849,6 +1851,12 @@ static void unix_inet_redir_to_connected(int family, int type, int sock_mapfd,
 	if (err)
 		goto close;
 
+	if (nop_mapfd >= 0) {
+		err = add_to_sockmap(nop_mapfd, c0, c1);
+		if (err)
+			goto close;
+	}
+
 	n = write(c1, "a", 1);
 	if (n < 0)
 		FAIL_ERRNO("%s: write", log_prefix);
@@ -1883,6 +1891,7 @@ static void unix_inet_skb_redir_to_connected(struct test_sockmap_listen *skel,
 					    struct bpf_map *inner_map, int family)
 {
 	int verdict = bpf_program__fd(skel->progs.prog_skb_verdict);
+	int nop_map = bpf_map__fd(skel->maps.nop_map);
 	int verdict_map = bpf_map__fd(skel->maps.verdict_map);
 	int sock_map = bpf_map__fd(inner_map);
 	int err;
@@ -1892,14 +1901,32 @@ static void unix_inet_skb_redir_to_connected(struct test_sockmap_listen *skel,
 		return;
 
 	skel->bss->test_ingress = false;
-	unix_inet_redir_to_connected(family, SOCK_DGRAM, sock_map, verdict_map,
+	unix_inet_redir_to_connected(family, SOCK_DGRAM,
+				     sock_map, -1, verdict_map,
+				     REDIR_EGRESS);
+	unix_inet_redir_to_connected(family, SOCK_DGRAM,
+				     sock_map, -1, verdict_map,
 				     REDIR_EGRESS);
-	unix_inet_redir_to_connected(family, SOCK_STREAM, sock_map, verdict_map,
+
+	unix_inet_redir_to_connected(family, SOCK_DGRAM,
+				     sock_map, nop_map, verdict_map,
+				     REDIR_EGRESS);
+	unix_inet_redir_to_connected(family, SOCK_STREAM,
+				     sock_map, nop_map, verdict_map,
 				     REDIR_EGRESS);
 	skel->bss->test_ingress = true;
-	unix_inet_redir_to_connected(family, SOCK_DGRAM, sock_map, verdict_map,
+	unix_inet_redir_to_connected(family, SOCK_DGRAM,
+				     sock_map, -1, verdict_map,
+				     REDIR_INGRESS);
+	unix_inet_redir_to_connected(family, SOCK_STREAM,
+				     sock_map, -1, verdict_map,
+				     REDIR_INGRESS);
+
+	unix_inet_redir_to_connected(family, SOCK_DGRAM,
+				     sock_map, nop_map, verdict_map,
 				     REDIR_INGRESS);
-	unix_inet_redir_to_connected(family, SOCK_STREAM, sock_map, verdict_map,
+	unix_inet_redir_to_connected(family, SOCK_STREAM,
+				     sock_map, nop_map, verdict_map,
 				     REDIR_INGRESS);
 
 	xbpf_prog_detach2(verdict, sock_map, BPF_SK_SKB_VERDICT);
diff --git a/tools/testing/selftests/bpf/progs/test_sockmap_listen.c b/tools/testing/selftests/bpf/progs/test_sockmap_listen.c
index 464d35bd57c7..b7250eb9c30c 100644
--- a/tools/testing/selftests/bpf/progs/test_sockmap_listen.c
+++ b/tools/testing/selftests/bpf/progs/test_sockmap_listen.c
@@ -14,6 +14,13 @@ struct {
 	__type(value, __u64);
 } sock_map SEC(".maps");
 
+struct {
+	__uint(type, BPF_MAP_TYPE_SOCKMAP);
+	__uint(max_entries, 2);
+	__type(key, __u32);
+	__type(value, __u64);
+} nop_map SEC(".maps");
+
 struct {
 	__uint(type, BPF_MAP_TYPE_SOCKHASH);
 	__uint(max_entries, 2);
-- 
2.33.0

Re: [PATCH bpf v2 0/2] sockmap fix for KASAN_VMALLOC and af_unix

[Jakub Sitnicki]

On Wed, Nov 22, 2023 at 11:24 AM -08, John Fastabend wrote:
> The af_unix tests in sockmap_listen causes a splat from KASAN_VMALLOC.
> Fix it here and include an extra test to catch case where both pairs
> of the af_unix socket are included in a BPF sockmap.
>
> Also it seems the test infra is not passing type through correctly when
> testing unix_inet_redir_to_connected. Unfortunately, the simple fix
> also caused some CI tests to fail so investigating that now.
>
> v2: drop changes to dgram side its fine per Jakub's point it graps a
>     reference on the peer socket from each sendmsg.
>
> John Fastabend (2):
>   bpf: sockmap, af_unix stream sockets need to hold ref for pair sock
>   bpf: sockmap, add af_unix test with both sockets in map
>
>  include/linux/skmsg.h                         |  1 +
>  include/net/af_unix.h                         |  1 +
>  net/core/skmsg.c                              |  2 +
>  net/unix/af_unix.c                            |  2 -
>  net/unix/unix_bpf.c                           |  5 +++
>  .../selftests/bpf/prog_tests/sockmap_listen.c | 39 ++++++++++++++++---
>  .../selftests/bpf/progs/test_sockmap_listen.c |  7 ++++
>  7 files changed, 49 insertions(+), 8 deletions(-)

Short of the nit pointed out by Yonghong Song:

Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>