Date: Tue, 14 Jan 2014 08:30:01 GMT From: Jens Kassel <Jens.Kassel@aptilo.com> To: freebsd-bugs@FreeBSD.org Subject: RE: kern/185596: in_cksum_skip: out of data by 41310 Message-ID: <201401140830.s0E8U1Z3019104@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR kern/185596; it has been noted by GNATS. From: Jens Kassel <Jens.Kassel@aptilo.com> To: Gleb Smirnoff <glebius@FreeBSD.org> Cc: "freebsd-gnats-submit@FreeBSD.org" <freebsd-gnats-submit@FreeBSD.org> Subject: RE: kern/185596: in_cksum_skip: out of data by 41310 Date: Tue, 14 Jan 2014 08:01:59 +0000 Hi, I have not tested this on FreeBSD 10. But I found that the problem can only be reproduced when using the bce driv= er.=20 Tested with both re & bge drivers and with these the problem cannot be repr= oduces. Checked SVN history for bce but could not find any commit that seemed to be= related to this issue. Regards, Jens -----Original Message----- From: Gleb Smirnoff [mailto:glebius@FreeBSD.org]=20 Sent: den 9 januari 2014 11:16 To: Jens Kassel Cc: freebsd-gnats-submit@FreeBSD.org Subject: Re: kern/185596: in_cksum_skip: out of data by 41310 Jens, On Thu, Jan 09, 2014 at 06:26:14AM +0000, Jens Kassel wrote: J> >Originator: Jens Kassel J> >Release: 8.4 J> Aptilo Networks J> >Environment: J> FreeBSD orbytebuild-freebsd84.sth.aptilo.com 8.4-RELEASE FreeBSD 8.4-REL= EASE8 #0 r251259: Mon Jun 3 01:14:28 UTC 2013 root@bake.isc.freebsd.or= g:/usr/obj/usr/src/sys/GENERIC i386 J>=20 J> We sometimes see messages like this in dmesg and /var/log/messages J>=20 J> in_cksum_skip: out of data by 15810 J> in_cksum_skip: out of data by 15810 J> in_cksum_skip: out of data by 15810 J> in_cksum_skip: out of data by 15810 J> in_cksum_skip: out of data by 41310 J>=20 J> I have checked with tcpdump and these messages seems to be triggered whe= n receiving "ICMP destination unreachable" packets. J>=20 J> Is this a known issue in this release? My guessing is that this messages comes off when a packet in wrong byte ord= er enters in_cksum(). Starting with 10.0, our stack doesn't change byte order of IP packets, so i= f my guessing is right, the bug should vanish. We do not plan any more releases from stable/8 branch. Thus, I'd like to cl= ose the PR if you don't mind. -- Totus tuus, Glebius.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201401140830.s0E8U1Z3019104>