Description of problem: When connected using CIPE, this protocol will not recognize too small packets passing through the tunnel, it will act as the packets never existed!! I found on the CIPE developer's site that this is really a bug on all versions of cipe previous to the 1.5.4. Bellow, I cut and pasted the information I found on http://sites.inka.de/bigred/archive/cipe-l/2002-01/msg00011.html -------------------------------------------------------------------- To: cipe-l Subject: BUG: crasher [IMPORTANT PATCH] From: Olaf Titz <olaf.de> Date: Mon, 07 Jan 2002 21:17:18 +0100 -------------------------------------------------------------------------------- This must be an old problem, why was it never found? :-) To my knowledge it exists in all published versions of CIPE. It causes a crash when CIPE receives too small packets. Thanks to Larry McVoy for alerting me to this bug. The attached patch is from the CVS but applies cleanly to 1.5.2. Olaf Index: cipe/sock.c =================================================================== RCS file: /home/olaf/common-home/CVSrepos/security/cipe-linux/cipe/sock.c,v retrieving revision 1.36 diff -u -r1.36 sock.c --- cipe/sock.c 2001/12/29 20:23:04 1.36 +++ cipe/sock.c 2002/01/06 18:28:56 @@ -199,6 +199,11 @@ goto framerr; } #endif + if (length<cipehdrlen+(c->sockshost?sizeof(struct sockshdr):0)) { + printk(KERN_INFO "%s: got short packet from %s\n", c->dev- >name, + cipe_ntoa(saddr(skb))); + goto framerr; + } n=alloc_skb(skb->len, GFP_KERNEL); if (!n) { @@ -390,10 +395,8 @@ c->stat.rx_packets++; return NULL; -#if 0 framerr: ++c->stat.rx_frame_errors; /* slightly abuse this */ -#endif error: ++c->stat.rx_errors; if (n) === end of patch === Version-Release number of selected component (if applicable): cipe-1.4.5-11 How reproducible: Always Steps to Reproduce: To reproduce the problem, try a 'ping' with the '-s 0' option, this will produce icmp packets carrying 0 bytes of data, what generates a packet of only 8 bytes (originated of the ICMP header). This way: ping -s 0 10.0.0.1 Where 10.0.0.1 is the p-t-p addres of your peer machine. Actual Results: The ping will end with 100% of loss. But if you simply try a 'ping 10.0.0.1' it will work. Why? Because the ping command sends by default a 56 bytes of data in each packet, which raises a packet with a total of 64 bytes. Expected Results: 'ping -s 0' should work as it works for machines not using CIPE. Additional info: Of course when this problem happens on small ICMP packets this is not serious, but, when happening on small TCP packets(or other protocols), it could cause real problems.
I confirm that this is a serios problem on tcp connections - i waisted a full working day troubleshooting a problem that in the end was caused by this bug. Building the kernel module with the one compiled from cipe 1.5.4 solved the problem. 1.5.4 also works fine with 1.4.5 machines, so i don't see any compatibility issues. This is a real problem.
just saw that the new beta, severn still comes with cipe 1.4.5 ??? Come on, how hard is it to update ??? Hell, just use my modifications http://mail.incredimail.com/howto/lvs/install/ there is a mondified kernel that includes lvs and cipe 1.5.4
a link to fix the problem described above is here - this is one line patch, that is taken from the latest, 1.5.4 cipe version. please consider integrating it to main kernel, if you are not upgrading to cipe 1.5.4 http://www.nrh-up2date.org/howto/lvs/src/linux-2.4.20-cipe-shortpacket.patch
This is also an issue with the current (as of this note) version of Fedora Core 1. The userspace daemon is still 1.4.5 and I regularly suffer from tunnel crashes.
Red Hat apologizes that these issues have not been resolved yet. We do want to make sure that no important bugs slip through the cracks. Red Hat Linux 7.3 and Red Hat Linux 9 are no longer supported by Red Hat, Inc. They are maintained by the Fedora Legacy project (http://www.fedoralegacy.org/) for security updates only. If this is a security issue, please reassign to the 'Fedora Legacy' product in bugzilla. Please note that Legacy security update support for these products will stop on December 31st, 2006. If this is not a security issue, please check if this issue is still present in a current Fedora Core release. If so, please change the product and version to match, and check the box indicating that the requested information has been provided. If you are currently still running Red Hat Linux 7.3 or 9, please note that Fedora Legacy security update support for these products will stop on December 31st, 2006. You are strongly advised to upgrade to a current Fedora Core release or Red Hat Enterprise Linux or comparable. Some information on which option may be right for you is available at http://www.redhat.com/rhel/migrate/redhatlinux/. Any bug still open against Red Hat Linux 7.3 or 9 at the end of 2006 will be closed 'CANTFIX'. Again, if this bug still exists in a current release, or is a security issue, please change the product as necessary. We thank you for your help, and apologize again that we haven't handled these issues to this point.
Red Hat Linux 7.3 and Red Hat Linux 9 are no longer supported by Red Hat, Inc. f you are currently still running Red Hat Linux 7.3 or 9, you are strongly advised to upgrade to a current Fedora Core release or Red Hat Enterprise Linux or comparable. Some information on which option may be right for you is available at http://www.redhat.com/rhel/migrate/redhatlinux/. Closing as CANTFIX.