tlug: Sparc Linux, TPE on eth0

[Frank BENNETT <bennett@example.com>]

I have just compiled a 2.2.1 kernel for an SS5 here on the kind
suggestion of Chris Sekiya, and with the kind advice of Steve
Turnbull.  It works, and more's the marvel, its SCSI bus appears
(superficially at least) to have stabilized after once repartitioning
the external disk and laying down a new ext2 filesystem on it.
The machine has successfully rebooted without resetting the
SCSI bus on that channel.  A first in our lab.  Hurray!

There is a problem, though, with ethernet support.  With kernel
2.0.3x (Redhat), I would get "Trying DHCP on eth0 ..." in the
boot sequence, followed after a 10-20 second pause by
something like "failed, trying TPE", after which the network
would pick up and the boot would proceed.

I now just get a failure.  The "Trying TPE" error recovery
message will come up, oddly, if I do a full ifconfig eth0 blah
blah blah on the command line, and routing seems to kick in
automatically after that, yielding full network support.  I could
probably mess around with the boot scripts to engineer similar
behavior at boot time, but it seems to me that what I OUGHT to do is
just force the driver to use TPE at boot time.  At the moment the
driver is compiled into the kernel.  Is there some boot option
I can use to force (what I think is) the sunlance driver to use
TPE rather than AUI?

Or does this suggest deeper and altogether (gulp) more interesting
problems?

Many thanks for any advice anyone can give.

With best regards,
Frank Bennett
Nagoya


-------------------------------------------------------------------
Next Nomikai: March 19 (Fri), 19:30  Tengu TokyoEkiMae 03-3275-3691
Next Technical Meeting: April 10 (Sat), 12:30   place: Temple Univ.
-------------------------------------------------------------------
more info: http://tlug.linux.or.jp                     Sponsor: PHT