Date: Mon, 1 Nov 2004 18:25:06 -0500 (EST) From: Andre Guibert de Bruet <andy@siliconlandmark.com> To: Charles Swiger <cswiger@mac.com> Cc: current@freebsd.org Subject: Re: 6-current 'corrupted mac on input' on cvs co Message-ID: <20041101182204.T82803@alpha.siliconlandmark.com> In-Reply-To: <452EA137-2C43-11D9-A18C-003065ABFD92@mac.com> References: <20041101192959.GD14715@daemon.li> <452EA137-2C43-11D9-A18C-003065ABFD92@mac.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 1 Nov 2004, Charles Swiger wrote: > On Nov 1, 2004, at 2:29 PM, Josef El-Rayes wrote: >> Received disconnect from 192.168.0.5: 2: Corrupted MAC on input. >> cvs[checkout aborted]: received broken pipe signal >> >> any ideas? do you need more input? > > The error message comes from /usr/src/crypto/openssh/packet.c, line 973. > > I haven't seen it happen often, but there seems to be some correlation > between using NAT or having packets get fragmented due to path MTU issues. > Any data corruption in the packets going by ought to be detected and fixed at > the TCP level, but if it doesn't, SSH will notice and drop the connection. As Josef pointed out in his other post to this thread, it is due to a corruption bug in if_sk. I saw this type of thing with if_nge w/mpsafenet set to 1 back in september (Check the archives for the reference), shortly after mpsafenet was made standard. Regards, | Andre Guibert de Bruet | Enterprise Software Consultant > | Silicon Landmark, LLC. | http://siliconlandmark.com/ >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20041101182204.T82803>