Asymetric network speed question

[David G. Miller]

Tim <ignored_mailbox at yahoo.com.au> wrote: 

>On Wed, 2006-07-05 at 11:21 -0600, David G. Miller wrote:
>  
>
>>> I guess next thing is to try a different NIC in the workstation and
>>> see if its the NIC.  The NICs were surplussed to me from where I used
>>> to work so I can't complain too much if I have trouble with them.
>>> They probably wouldn't have gotten rid of them unless there was a
>>> reason.
>>    
>>
>
>Even "good" ones don't always work well for someone else.  They might
>work well with Windows drivers, but not Linux, or vice versa.
>
>I have an issue with one of mine that way.  On a dual-boot box, it's
>rather painful in one direction between Windows and Linux.  But if I
>reboot that box to Linux, it works fine between itself and another Linux
>box.  Go figure...
>
>(Yes, I'd checked through the duplex options in both OSs.)
>  
>
Tried a different 3c2000-t (box was still shrink wrapped) and behavior 
was the same.  Not sure when the skge package replaced sk98lin in FC4 
but it looks like the version of the kernel on the original install cd 
still had sk98lin.  I just did a --oldpackage install of that kernel so 
I'll see what happens when this box has sk98lin as the driver.  I'll 
post the result.

I've always considered networking somewhat of a "black art" since the 
time in the late '80s when two adjacent HP workstations on my project 
decided to stop talking to each other and we fixed the problem by 
replacing the 3 ft cable between them with a 6 ft cable.  This was back 
in the coax days.

>  
>
>>> BTW, gigabit uses all four wire pairs in the cable.
>>    
>>
>
>Yep, as I said, "two pairs in each direction" (that's four pairs in one
>cable).
>
Oops.  Sorry.  I've had way too much advice about not needing all four 
pairs.

Cheers,
Dave

-- 
Politics, n. Strife of interests masquerading as a contest of principles.
-- Ambrose Bierce