In message <19970428172237.1972.qmail@noxious.muscat.co.uk>,
     Olly Betts wrote:

> [...]
> 
> http://noxious.muscat.co.uk/riscbsd/
> 
> This has a good search facility, but the messages are just text at present
> (I'll be adding threading, HTML-ising, and other features at some point).

Thanks, the index has worked well enough for me. Good job!

The index showed that there's an early SA110 kernel (4740) that can detect
an etherh card. The index also showed that there were some problems with
the etherh driver at that time resulting in "timeout" messages or
"carrier lost" messages. I got these messages with kernel sa4740 as well.

I then decided to build my own kernel. I found an old 1.2A kernel source
archive and followed the configuration instruction. I made sure
that "eh" isn't commented out in the config file. Building the kernel
worked fine but when I booted the new kernel I got the "podule8: not
configured" message again.

My first thought was that podulebus/if_eh.c hasn't been compiled but there
is a new if_eh.o in the compile directory so it has been linked into the
new kernel.

A closer look at the message that signals the detection of the etherh card
showed that the manufacturer (I-cubed) was recognized fine. The product id
however was not recognized. The card obviously reported 0x011e but in the
kernel sources the product id was specified as 0x00ec (?). So I changed
the product id in podulebus/podules, did "awk -f podulelist.awk podules"
to create a new podules/podules.h and rebuild the kernel.

After booting the new kernel the card could be configured without
problems. I even don't get "timeout" or "carrier lost" messages. 

I'm now trying to get a newer kernel source so don't be surprised if my
next posting is about sup'ing ;-)

> 
> Cheers,
> Olly
> -- 
> Don't send a man to do a buoy's job.
> 

Regards,
   Stefan
   
-- 
Stefan Voss
(voss@yoda.in-berlin.de)