[PATCH net] af_unix: fix struct pid leaks in OOB support

From: Eric Dumazet <edumazet@google.com>
To: "David S . Miller" <davem@davemloft.net>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>
Cc: netdev@vger.kernel.org, eric.dumazet@gmail.com,
	Eric Dumazet <edumazet@google.com>,
	syzbot+7699d9e5635c10253a27@syzkaller.appspotmail.com,
	Rao Shoaib <rao.shoaib@oracle.com>,
	Kuniyuki Iwashima <kuniyu@amazon.com>
Subject: [PATCH net] af_unix: fix struct pid leaks in OOB support
Date: Tue,  7 Mar 2023 16:45:30 +0000	[thread overview]
Message-ID: <20230307164530.771896-1-edumazet@google.com> (raw)

syzbot reported struct pid leak [1].

Issue is that queue_oob() calls maybe_add_creds() which potentially
holds a reference on a pid.

But skb->destructor is not set (either directly or by calling
unix_scm_to_skb())

This means that subsequent kfree_skb() or consume_skb() would leak
this reference.

In this fix, I chose to fully support scm even for the OOB message.

[1]
BUG: memory leak
unreferenced object 0xffff8881053e7f80 (size 128):
comm "syz-executor242", pid 5066, jiffies 4294946079 (age 13.220s)
hex dump (first 32 bytes):
01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:
[<ffffffff812ae26a>] alloc_pid+0x6a/0x560 kernel/pid.c:180
[<ffffffff812718df>] copy_process+0x169f/0x26c0 kernel/fork.c:2285
[<ffffffff81272b37>] kernel_clone+0xf7/0x610 kernel/fork.c:2684
[<ffffffff812730cc>] __do_sys_clone+0x7c/0xb0 kernel/fork.c:2825
[<ffffffff849ad699>] do_syscall_x64 arch/x86/entry/common.c:50 [inline]
[<ffffffff849ad699>] do_syscall_64+0x39/0xb0 arch/x86/entry/common.c:80
[<ffffffff84a0008b>] entry_SYSCALL_64_after_hwframe+0x63/0xcd

Fixes: 314001f0bf92 ("af_unix: Add OOB support")
Reported-by: syzbot+7699d9e5635c10253a27@syzkaller.appspotmail.com
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Rao Shoaib <rao.shoaib@oracle.com>
Cc: Kuniyuki Iwashima <kuniyu@amazon.com>
---
 net/unix/af_unix.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
index 347122c3575eaae597405369e2e9d8324d6ad240..0b0f18ecce4470d6fd21c084a3ea49e04dcbb9bd 100644
--- a/net/unix/af_unix.c
+++ b/net/unix/af_unix.c
@@ -2105,7 +2105,8 @@ static int unix_dgram_sendmsg(struct socket *sock, struct msghdr *msg,
 #define UNIX_SKB_FRAGS_SZ (PAGE_SIZE << get_order(32768))
 
 #if IS_ENABLED(CONFIG_AF_UNIX_OOB)
-static int queue_oob(struct socket *sock, struct msghdr *msg, struct sock *other)
+static int queue_oob(struct socket *sock, struct msghdr *msg, struct sock *other,
+		     struct scm_cookie *scm, bool fds_sent)
 {
 	struct unix_sock *ousk = unix_sk(other);
 	struct sk_buff *skb;
@@ -2116,6 +2117,11 @@ static int queue_oob(struct socket *sock, struct msghdr *msg, struct sock *other
 	if (!skb)
 		return err;
 
+	err = unix_scm_to_skb(scm, skb, !fds_sent);
+	if (err < 0) {
+		kfree_skb(skb);
+		return err;
+	}
 	skb_put(skb, 1);
 	err = skb_copy_datagram_from_iter(skb, 0, &msg->msg_iter, 1);
 
@@ -2243,7 +2249,7 @@ static int unix_stream_sendmsg(struct socket *sock, struct msghdr *msg,
 
 #if IS_ENABLED(CONFIG_AF_UNIX_OOB)
 	if (msg->msg_flags & MSG_OOB) {
-		err = queue_oob(sock, msg, other);
+		err = queue_oob(sock, msg, other, &scm, fds_sent);
 		if (err)
 			goto out_err;
 		sent++;
-- 
2.40.0.rc0.216.gc4246ad0f0-goog

