netdev - Re: [PATCH] pppoe: Use workqueue to die properly when a PADT is received

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]

Date:	Fri, 20 Feb 2015 17:10:14 +0100
From:	Christoph Schulz <develop@...stov.de>
To:	Simon Farnsworth <simon@...nz.org.uk>
Cc:	netdev@...r.kernel.org, Dan Williams <dcbw@...hat.com>,
	mostrows@...il.com, linux-ppp@...r.kernel.org
Subject: Re: [PATCH] pppoe: Use workqueue to die properly when a PADT is
 received

(Cc: linux-ppp@...r.kernel.org, mostrows@...il.com)

Hello!

Simon Farnsworth schrieb am Thu, 19 Feb 2015 21:24:28 +0000:

> When a PADT frame is received, the socket may not be in a good state to
> close down the PPP interface. The current implementation handles this by
> simply blocking all further PPP traffic, and hoping that the lack of traffic
> will trigger the user to investigate.
>
> Use schedule_work to get to a process context from which we clear down the
> PPP interface, in a fashion analogous to hangup on a TTY-based PPP
> interface. This causes pppd to disconnect immediately, and allows tools to
> take immediate corrective action.

Tested-by: Christoph Schulz <develop@...stov.de>

I tested your patch von top of 3.14.33 and 3.18.7 together with a  
slightly modified pppd based on Debian's pppd-2.4.6-3 (which is almost  
identical to pppd 2.4.7). The client now indeed receives PADT feedback  
immediately. Closing a PPPoE connection on the server side

Feb 20 16:45:44 swing local2.notice pppoe-server-wrapper[20154]:  
circ4: pppoe-server died with exit code 0, cleaning up

leads to immediate hangup on the client side:

Feb 20 16:45:44 sandbox local2.notice pppd[539]: Modem hangup

Formerly, the LCP echo timeout had to be reached. On the server side:

Feb 20 14:56:53 swing local2.notice pppoe-server-wrapper[12744]:  
circ4: pppoe-server died with exit code 0, cleaning up

On the client side:

Feb 20 14:57:01 sandbox local2.info pppd[15575]: No response to 2  
echo-requests
Feb 20 14:57:01 sandbox local2.notice pppd[15575]: Serial link appears  
to be disconnected.

Note that the short detection of the terminated link (~ 8 s) stems  
from very aggressive LCP echo settings (lcp-echo-interval "3",  
lcp-echo-failure "2"). Typically, these timeouts are much larger to  
accomodate for bad links. Often the client recognizes a dead link only  
after two to three minutes (lcp-echo-interval "30", lcp-echo-failure  
"5").

However, note also that your patch causes pppd (or rather the rp-pppoe  
plugin) to wonder about the socket closed by the kernel:

Feb 20 16:45:44 sandbox local2.err pppd[539]: Failed to disconnect  
PPPoE socket: 114 Operation already in progress

I don't fully understand the code there; it seems that the plugin  
*connects* the PPPoE session socket in order to *disconnect* it:

static void
PPPOEDisconnectDevice(void)
{
     struct sockaddr_pppox sp;

     sp.sa_family = AF_PPPOX;
     sp.sa_protocol = PX_PROTO_OE;
     sp.sa_addr.pppoe.sid = 0;
     memcpy(sp.sa_addr.pppoe.dev, conn->ifName, IFNAMSIZ);
     memcpy(sp.sa_addr.pppoe.remote, conn->peerEth, ETH_ALEN);
     if (connect(conn->sessionSocket, (struct sockaddr *) &sp,
                 sizeof(struct sockaddr_pppox)) < 0)
         error("Failed to disconnect PPPoE socket: %d %m", errno);
     close(conn->sessionSocket);
     /* don't send PADT?? */
     if (conn->discoverySocket >= 0)
         close(conn->discoverySocket);
}


Regards,
-- 
Christoph Schulz

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html