In article <1993Mar9.194509.2344@super.org> becker@super.org (Donald J. Becker) writes:
quoted text
>In article <1993Mar8.160930.6123@searchtech.com> mra@searchtech.com (Michael Almond) writes:
>>In article <8659.1189.uupcb@satalink.com> john.will@satalink.com (John Will)  writes:
>>>....
>>>The ethernet probe seems to be confused, it finds the WD8003, but it
>>>thinks it's on IRQ3, which is obviously false.  In point of fact, it's
>>>on IRQ2, and works fine there with a MS-DOS network.  This is what I
>>>see during the boot:
>>>
>>>WD80x3 ethercard probe at 0-280 00 00 C0 76 A5 12 using address 0xd0000,
>>>WD8003 found, using IRQ 3.
>>
>>I did have a problem with my eth setup, however, the probe was finding my
>>card properlly.  My problem was I didn't change the IP address on SLS from
>>the one that comes with the system (I'm on an isolated PC network).  In
>>the rc.net it checks to see if the IP address is equal to softland's and
>>doesn't complete the configuration of the eth if the IP is softland's.
>>
>>You may want to try IRQ 10 or 5.
>
>In this case the problem is likely due to using an old version of the
>WD8003.  The old versions didn't have the registers that report the
>IRQ line and shared memory location, yet the driver looks anyway:-<.
>The immediate fix is to compile a new kernel with the shared memory
>and IRQ line explicitly specified in /usr/src/linux/net/tcp/GNUmakefile.
>
>The next version of the driver should correctly detect an old WD8003,
>but I haven't yet found a safe way to autoIRQ or detect the shared
>memory address on those boards.  It might be time to add an ioctl() to
>set ethercard values and enhance /etc/config... I know Fred would like
>this.
>
>Donald Becker                                 becker@super.org
>Supercomputing Research Center
>17100 Science Drive, Bowie MD 20715               301-805-7482

Hmm... with the current SLS, I HAD to use an old-style 8003 (w/jumpers)
since the new-style (mostly software config) simply hung.  Maybe, though
the board was bad.  Anyway, will the updated driver support the *OLD*
style board?

BTW- I tried a 3c503 first, but either the kernel was compiled for thin-
     net or the "eth_if" problem could not be overcome.  Getting the *OLD*
     8003 solved my problem.  I am trying to use "off-the-shelf" SLS, so
     kernel-compiling was not an option.
-- 
==============================================================================
  phil perucci                       | "Any opinions expressed are my views, 
  dsc3pzp@nmrdc1.nmrdc.nnmc.navy.mil |  not the position of any organization"
==============================================================================