[ECOS] NIC problems running Redboot

["Mike at VirginNet" <mister dot chips at virgin dot net>]

Hi,

When I run the 'ip_address' command having booted in the absence of any NIC,
the console prompt corrupts (the text I'm seeing is
"$T0b08:ffffffff;04:1c9c0100;#bd"). Also there doesn't seem to be any way of
assigning the local IP address to a channel, nor is there a way of
specifying IP addresses in any other command. I had assumed that RedBoot
could operate TFTP over serial using PPP, and that this would be the best
way to connect to a host in the absence of a NIC. Am I wrong?

I'm doing this serial-line stuff because :-

(a) the 21143-based PC build I have produced, seems to have problems driving
the NIC in question. It reads the ethernet address off the card as all
zeroes, and here's a selection of other diagnostics I see on startup :-

    Note that board is active. Probed sizes and CPU addresses invalid!
    eth0 configured
    I/O address = 0x00e0000
    *** device enabled for I/O only and Bus Master
    eth_set_mac_address: 482 loops: bufdesc status tdes0 7fffffff, tdes1
0a0000c0
    eth_set_mac_address: tx done after 0 iters (status f0600004)
    eth_set_mac_address: tx stopped after 0 iters (status f0000006)
    i21143_start 0 flg ff07 Link = Down, 10Mbps, half duplex
    Can't get BOOTP info, using default IP address

...and the results of ping are 'network unreachable' (because Link = Down?
Is there an equivalent of "ifconfig eth0 up"?).  Any ideas what may be
causing these problems?

(b) the 83816-based PC build I've just come up with, doesn't recognise the
existance of a known working 83815 NIC (though the NS web site provides the
same Linux driver source for both chipsets, so I'd have thought the eCos one
should work on both also). The target (moab) that uses the 83816 driver in
the CVS version of ecos.db uses three other DEVS_ETH components, two of
which are PPC-specific, and I assume I won't need to produce PC equivalents
of them (or will I?).  The fourth is CYGPKG_DEVS_ETH_PHY - should I be using
that, rather than some hack of say the 82559 PC-specific driver part, as
now?

Maybe I'll need to do some fairly substantial code tailoring to get the
21143 and/or 83816 driver to operate in the PC environment. Is that to be
expected?

By the way, many thanks for the ongoing help I'm getting here!

Cheers,

Mike



---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.530 / Virus Database: 325 - Release Date: 23/10/03


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss