netdev - Re: [PATCH v2] l2tp: Avoid possible recursive deadlock in l2tp_tunnel

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: <Y+pbMgEq0epVbB4P@debian>
Date:   Mon, 13 Feb 2023 16:45:54 +0100
From:   Guillaume Nault <gnault@...hat.com>
To:     Alexander Lobakin <alexandr.lobakin@...el.com>
Cc:     Shigeru Yoshida <syoshida@...hat.com>, jchapman@...alix.com,
        davem@...emloft.net, edumazet@...gle.com, kuba@...nel.org,
        pabeni@...hat.com, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] l2tp: Avoid possible recursive deadlock in
 l2tp_tunnel_register()

On Mon, Feb 13, 2023 at 04:05:59PM +0100, Alexander Lobakin wrote:
> From: Shigeru Yoshida <syoshida@...hat.com>
> Date: Mon, 13 Feb 2023 01:26:23 +0900
> 
> > When a file descriptor of pppol2tp socket is passed as file descriptor
> > of UDP socket, a recursive deadlock occurs in l2tp_tunnel_register().
> > This situation is reproduced by the following program:
> 
> [...]
> 
> > +static struct l2tp_tunnel *pppol2tp_tunnel_get(struct net *net,
> > +					       struct l2tp_connect_info *info,
> > +					       bool *new_tunnel)
> > +{
> > +	struct l2tp_tunnel *tunnel;
> > +	int error;
> > +
> > +	*new_tunnel = false;
> > +
> > +	tunnel = l2tp_tunnel_get(net, info->tunnel_id);
> > +
> > +	/* Special case: create tunnel context if session_id and
> > +	 * peer_session_id is 0. Otherwise look up tunnel using supplied
> > +	 * tunnel id.
> > +	 */
> > +	if (!info->session_id && !info->peer_session_id) {
> > +		if (!tunnel) {
> 
> This `if` is the sole thing the outer `if` contains, could we combine them?

The logic of this code is a bit convoluted, sure, but if we want to
rework it, let's simplify it for real:

	tunnel = l2tp_tunnel_get(...)
	if (tunnel)
		return tunnel; /* the original !tunnel->sock test is useless I believe */

	/* Tunnel not found. Try to create one if both session_id and
	 * peer_session_id are 0. Fail otherwise.
	 */
	if (info->session_id || info->peer_session_id)
		return ERR_PTR(-ENOENT);

	[...] /* Tunnel creation code */


However, I'd prefer to keep such refactoring for later. Keeping the
same structure in pppol2tp_tunnel_get() as in the original code helps
reviewing the patch.

> > +			struct l2tp_tunnel_cfg tcfg = {
> > +				.encap = L2TP_ENCAPTYPE_UDP,
> > +			};
> > +
> > +			/* Prevent l2tp_tunnel_register() from trying to set up
> > +			 * a kernel socket.
> > +			 */
> > +			if (info->fd < 0)
> > +				return ERR_PTR(-EBADF);
> > +
> > +			error = l2tp_tunnel_create(info->fd,
> > +						   info->version,
> 
> This fits into the prev line.
> 
> > +						   info->tunnel_id,
> > +						   info->peer_tunnel_id, &tcfg,
> > +						   &tunnel);
> > +			if (error < 0)
> > +				return ERR_PTR(error);
> > +
> > +			l2tp_tunnel_inc_refcount(tunnel);
> > +			error = l2tp_tunnel_register(tunnel, net, &tcfg);
> > +			if (error < 0) {
> > +				kfree(tunnel);
> > +				return ERR_PTR(error);
> > +			}
> > +
> > +			*new_tunnel = true;
> > +		}
> > +	} else {
> > +		/* Error if we can't find the tunnel */
> > +		if (!tunnel)
> > +			return ERR_PTR(-ENOENT);
> 
> [...]
> 
> Thanks,
> Olek
>