STREIF@UWEC.BITNET (VINCE STREIF) (03/30/89)
I just installed the core of a new ethernet and hoped to use the KA9Q
package on this net, however I can't get my machines to talk. The
current setup involves a Zenith 386 machine and a Zenith AT clone,
both running MS-DOS (version 3.21 in one case and 3.3+ in the other)
and they have Western Digital wd8003e's installed using IRQ3. Neither
of the machines has a second serial port installed, so that interrupt
ought to be safe to use (I hope). Both machines pass the entire set
of WD's diagnostics, including the external loopback so I'm convinced
that the cabling is ok.
To verify that I knew something about the software, I first connected
the two machines via their serial ports (COM1) using the KA9Q software
and managed to do FTP, Mail, etc... Then I switched their ATTACH
commands to the following:
attach packet 3 ec0 5 1500
I load the driver with the following command:
wd8003e 3 2 0x280 0xd8000
I had to override the address (defaults to 0xC4000) because of the
configuration of the 386 (vga etc...). Unfortunately, I can find
no sign that either machine will speak to the other across the
ethernet once I start NET. Have I done something truly silly?
There's obviously a lot more information I could send (e.g. copies
of my configuration files) but since it worked with an async connection,
my best guess is that it has something to do directly with the driver.
Thanks in advance for any info anyone can give me.
Vince Streif streif@uwec
streif@uwecvaxa
etc...
University of Wisconsin - Eau Claire
- Vince -clements@bbn.com (Bob Clements) (04/03/89)
In article <890330.08112901.059112@UWEC.CP6> STREIF@UWEC.BITNET (VINCE STREIF) writes: > ... Then I switched their ATTACH >commands to the following: > attach packet 3 ec0 5 1500 > wd8003e 3 2 0x280 0xd8000 > > ... Have I done something truly silly? Just slightly silly. Change the two instances of "3" in those commands to "0x7e" or some other number in the range of 0x60 to 0x7f. That is the software interrupt number used to communicated to/from the packet driver, and must be in the range of reserved software interrupts for MS-DOS. I'm surprised that neither program complained about the "3", but it's not a good number. I hope this fixes it for you; there's no guarantee that only one thing is wrong, of course... >Vince Streif streif@uwec >- Vince - /Rcc clements@bbn.com