Might simply be a bad card or even a bad setup, but I *believe* there might be a problem with the re driver and was looking for some external collaboration. I didn't think there was a need for a cross-over cable. I tossed it in there trying to be thorough.
Thanks for the response! If this is actually a natural response for GigE then I'm wayyy off tangent in terms of a diagnosis and need to explore something else. I'll shut-up now. Cheers! 2008/11/20 Josh Archambault <[EMAIL PROTECTED]>: > Whyzzi wrote: >> >> Just wondering if anyone else on the list that has a realtek 8169s and >> running 4.4 release and able to run a GigE connection from their >> OpenBSD box to another GigE device, if they can maintain a complete >> connection when they force the re driver via ifconfig to 1000baseT. >> >> Incase you are wondering I do have both a GigE windows machine and >> generic unmanaged GigE switch. I did change the physical network cable >> between the switch and the re rj45 jack with no effect. And plugging >> re directly into my GigE windows box had the driver attempt to make a >> connection, fail, and retry continuously (though I did not use a >> cross-over cable nor try a different network cable -- time was >> somewhat limited). >> >> I skimmed cvs-web on the re driver and didn't see any posts related to >> behaviour. >> >> Flames welcome! > > No flames...just a question: why are you doing this? > > At 10Mb/s and 100Mb/s there are good reasons to fix media speed and duplex, > but I don't know of any reason to do so at 1000Mb/s. While most previous > Ethernet specs did not require or standardize auto-negotiation, the > 1000Base-T spec does both. My experiences have been that working against > the spec and attempting to fix speed or duplex in 1000baseT environments > causes exactly the sorts of weird problems you are describing here. > > (Also, there is no longer a need for "cross-over" cables in the 1000Mb/s > world either...polarity detection is also a part of the 1000baseT spec.) > > -J
