netdev - Re: [PATCH net] can: j1939: j1939_xtp_rx_dat_one(): cancel session if receive TP.DT with error length

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20210930074206.GB7502@x1.vandijck-laurijssen.be>
Date:   Thu, 30 Sep 2021 09:42:06 +0200
From:   Kurt Van Dijck <dev.kurt@...dijck-laurijssen.be>
To:     Zhang Changzhong <zhangchangzhong@...wei.com>
Cc:     Robin van der Gracht <robin@...tonic.nl>,
        Oleksij Rempel <linux@...pel-privat.de>, kernel@...gutronix.de,
        Oliver Hartkopp <socketcan@...tkopp.net>,
        Marc Kleine-Budde <mkl@...gutronix.de>,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Maxime Jayat <maxime.jayat@...ile-devices.fr>,
        linux-can@...r.kernel.org, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH net] can: j1939: j1939_xtp_rx_dat_one(): cancel session
 if receive TP.DT with error length

On Thu, 30 Sep 2021 11:33:20 +0800, Zhang Changzhong wrote:
> According to SAE-J1939-21, the data length of TP.DT must be 8 bytes, so
> cancel session when receive unexpected TP.DT message.

SAE-j1939-21 indeed says that all TP.DT must be 8 bytes.
However, the last TP.DT may contain up to 6 stuff bytes, which have no meaning.
If I remember well, they are even not 'reserved'.

> 
> Fixes: 9d71dd0c7009 ("can: add support of SAE J1939 protocol")
> Signed-off-by: Zhang Changzhong <zhangchangzhong@...wei.com>
> ---
>  net/can/j1939/transport.c | 7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/net/can/j1939/transport.c b/net/can/j1939/transport.c
> index bb5c4b8..eedaeaf 100644
> --- a/net/can/j1939/transport.c
> +++ b/net/can/j1939/transport.c
> @@ -1789,6 +1789,7 @@ static void j1939_xtp_rx_dpo(struct j1939_priv *priv, struct sk_buff *skb,
>  static void j1939_xtp_rx_dat_one(struct j1939_session *session,
>  				 struct sk_buff *skb)
>  {
> +	enum j1939_xtp_abort abort = J1939_XTP_ABORT_FAULT;
>  	struct j1939_priv *priv = session->priv;
>  	struct j1939_sk_buff_cb *skcb, *se_skcb;
>  	struct sk_buff *se_skb = NULL;
> @@ -1803,9 +1804,11 @@ static void j1939_xtp_rx_dat_one(struct j1939_session *session,
>  
>  	skcb = j1939_skb_to_cb(skb);
>  	dat = skb->data;
> -	if (skb->len <= 1)
> +	if (skb->len != 8) {
>  		/* makes no sense */
> +		abort = J1939_XTP_ABORT_UNEXPECTED_DATA;
>  		goto out_session_cancel;

I think this is a situation of
"be strict on what you send, be tolerant on what you receive".

Did you find a technical reason to abort a session because the last frame didn't
bring overhead that you don't use?

Kind regards,
Kurt
> +	}
>  
>  	switch (session->last_cmd) {
>  	case 0xff:
> @@ -1904,7 +1907,7 @@ static void j1939_xtp_rx_dat_one(struct j1939_session *session,
>   out_session_cancel:
>  	kfree_skb(se_skb);
>  	j1939_session_timers_cancel(session);
> -	j1939_session_cancel(session, J1939_XTP_ABORT_FAULT);
> +	j1939_session_cancel(session, abort);
>  	j1939_session_put(session);
>  }
>  
> -- 
> 2.9.5
>