[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <162dd41ee6717ad46e0a37003d922ea1@dev.tdt.de>
Date: Tue, 29 Sep 2020 06:52:12 +0200
From: Martin Schiller <ms@....tdt.de>
To: David Miller <davem@...emloft.net>
Cc: andrew.hendry@...il.com, kuba@...nel.org, edumazet@...gle.com,
xiyuyang19@...an.edu.cn, linux-x25@...r.kernel.org,
netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] net/x25: Fix null-ptr-deref in x25_connect
On 2020-09-29 03:43, David Miller wrote:
> From: Martin Schiller <ms@....tdt.de>
> Date: Mon, 28 Sep 2020 11:23:27 +0200
>
>> diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c
>> index 0bbb283f23c9..0524a5530b91 100644
>> --- a/net/x25/af_x25.c
>> +++ b/net/x25/af_x25.c
>> @@ -820,7 +820,7 @@ static int x25_connect(struct socket *sock, struct
>> sockaddr *uaddr,
>>
>> rc = x25_wait_for_connection_establishment(sk);
>> if (rc)
>> - goto out_put_neigh;
>> + goto out;
>
> If x25_wait_for_connection_establishment() returns because of an
> interrupting
> signal, we are not going to call x25_disconnect().
>
> The case you are fixing only applies _sometimes_ when
> x25_wait_for_connection_establishment() returns. But not always.
>
> That neighbour has to be released at this spot otherwise.
OK, thanks for the hint. So I think the simplest solution would be to
check
that x25->neighbour is != NULL like this:
diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c
index 0bbb283f23c9..046d3fee66a9 100644
--- a/net/x25/af_x25.c
+++ b/net/x25/af_x25.c
@@ -825,7 +825,7 @@ static int x25_connect(struct socket *sock, struct
sockaddr *uaddr,
sock->state = SS_CONNECTED;
rc = 0;
out_put_neigh:
- if (rc) {
+ if (rc && x25->neighbour) {
read_lock_bh(&x25_list_lock);
x25_neigh_put(x25->neighbour);
x25->neighbour = NULL;
--
What do you think?
If that would be OK, I'll send a v2 of the Patch.
- Martin
Powered by blists - more mailing lists