Yes, the bug does exist in 3.4.29 The patch is quite easy to apply if you don't want to upgrade to the latest version...
You'll probably only notice the bug if (as I did) you try to transfer large (10s of MB) amounts of data through ipnat using something like FTP or rcp. David -- Dr David Hedley Director, Inty Ltd www.inty.com, 01454 640500 > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED]]On Behalf Of Chris Johnson > Sent: 07 February 2003 15:08 > To: [EMAIL PROTECTED] > Subject: RE: Bug in fr_tcpsum on *BSD on 3.4.31 > > > Does this same bug exist in 3.4.29? Or is the code in error a recent > introduction? > > 3.4.29 is the standard version on 4.7-RELEASE, so without good reason to > upgrade, that's what I'm running on my hosts. I'd guess that would be the > case with other folks, too. > > Thanks, > ..chris > > > -----Original Message----- > > From: [EMAIL PROTECTED] > > [mailto:[EMAIL PROTECTED]]On Behalf Of David Hedley > > > Subject: Bug in fr_tcpsum on *BSD on 3.4.31 > > > Version: 3.4.31 > > Platform: FreeBSD 4.7-RELEASE > > > > There is a bug in the calculation of TCP checksums on *BSD (and > > possibly SunOS) platforms. This will cause random timeouts on any > > connections using any of the proxy modules (e.g. ftp, rcmd etc). > > > > The bug is caused by the incorrect combination of the two partial > > sums computed for each packet. When the two partial sums are the > > same, the result return is 0xffff instead of the correct 0x0000. > > > > intY has scanned this email for all known viruses (www.inty.com) > > > intY has scanned this email for all known viruses (www.inty.com)
