net/rose: Fix Use-After-Free in rose_ioctl
authorHyunwoo Kim <v4bel@theori.io>
Sat, 9 Dec 2023 10:05:38 +0000 (05:05 -0500)
committerPaolo Abeni <pabeni@redhat.com>
Tue, 12 Dec 2023 12:24:58 +0000 (13:24 +0100)
Because rose_ioctl() accesses sk->sk_receive_queue
without holding a sk->sk_receive_queue.lock, it can
cause a race with rose_accept().
A use-after-free for skb occurs with the following flow.
```
rose_ioctl() -> skb_peek()
rose_accept() -> skb_dequeue() -> kfree_skb()
```
Add sk->sk_receive_queue.lock to rose_ioctl() to fix this issue.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Hyunwoo Kim <v4bel@theori.io>
Link: https://lore.kernel.org/r/20231209100538.GA407321@v4bel-B760M-AORUS-ELITE-AX
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
net/rose/af_rose.c

index 0cc5a4e19900e10b31172433f36f5835101908ed..ecb91ad4ce639e7f8f3d2a9698e005cf696b123f 100644 (file)
@@ -1315,9 +1315,11 @@ static int rose_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
        case TIOCINQ: {
                struct sk_buff *skb;
                long amount = 0L;
-               /* These two are safe on a single CPU system as only user tasks fiddle here */
+
+               spin_lock_irq(&sk->sk_receive_queue.lock);
                if ((skb = skb_peek(&sk->sk_receive_queue)) != NULL)
                        amount = skb->len;
+               spin_unlock_irq(&sk->sk_receive_queue.lock);
                return put_user(amount, (unsigned int __user *) argp);
        }