Jurij Smakov wrote: [snip] > >The thing is, that not all telnet-connections fail: > > > >- telnet to Linux-Host (try 'telnet mailgw1 25'): works > >- telnet to Solaris-Host (try 'telnet esslingen'): works > >- telnet to Ascend MAX4000 Access-Server (m-nas1): fails > >- telnet to Cisco 7200 VXR (try 'telnet ol-gw'): fails > > Right, I was able to figure out that these bad checksum messages are > completely bogus (they are also produced by a tcpdump when I run it on my > Ultra5 and telnet to it). So that's not really it :-(. At the moment I am > out of ideas. I'll try to build a pristine kernel tomorrow and make a > comparison between tcpdump output in these two cases.
This suggests the checksum inline assembly in arch/sparc64/lib is somewhat broken, but it seem to be in line with upstream 2.4. It is much more complicated than the 2.6 code, though. Thiemo -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
