[3278] in SIPB_Linux_Development
Re: more 3c90x boot disk take 2
daemon@ATHENA.MIT.EDU (seph)
Sat Jan 6 15:07:38 2001
Message-Id: <200101062007.PAA19324@ten-thousand-dollar-bill.mit.edu>
To: Derek Atkins <warlord@MIT.EDU>
Cc: seph <seph@MIT.EDU>, linux-dev@MIT.EDU
In-Reply-To: Your message of "06 Jan 2001 10:32:44 EST."
<sjmy9wopug3.fsf@rcn.ihtfp.org>
Date: Sat, 06 Jan 2001 15:07:33 -0500
From: seph <seph@MIT.EDU>
you're makefile basicly does:
make-boots: update-loader
./mkboot -f $(IMAGES)
it does not do a mkinitrd. anyhow, the update-loader taget is fine
and "./mkboot -f network" produces an "incorrect usage" error
message. and mkbot runs mkinitrd, I don't think it really needs to be
run twice.
> 1) I believe that pcitable data needs to me in order -- your
> added line is not in the right place.
hum. I didn't think so, but I fixed this anyhow.
> 2) We should verify what the PCI data is for their network
> card; It appears there is an entry for the 3c905C-TX at
> 0x10b7 0x9200 which refers to the 3c59x driver; I don't
> know if this should be changed to the 3c90x driver. It's
> possible that this "new" card has different PCI info.
I took my pcitable info straight from ..//anaconda/kudzu/pcitable,
which would appear to be redhat's stuff. I would guess that the
installer uses it to tell what driver goes with what, so we could try
forcing more things to use the 3c90x driver. but I'm not sure it's needed.
> As I asked before, is there any reason you left out the 3c50[1357]
> drivers?
this is the first time I've responded to any of your mail. I left them
out cause when I tried with them, because I can't fit the 3x90x driver
without removing something, and I rather doubt anyone uses those
anymore. I should probably pull them out of the modules-info and
pcitable though. (and yes, I did try to build the disk without the
.olds, and with them)
cfox> I tried seph's boot floppy image on snork. If I select "3Com
cfox> 3c590/3c595 Vortex", it punts me back to the "what boot media" screen,
cfox> with "error 2 reading header: Success" at the bottom of it.
the "3Com 3c590/3c595 Vortex" is the 3c59x driver, sp that might have
been the problem the new driver is current called "3Com 3c509[bc]
Vortex" yes, I see the typo, if it works, we can pretty easily change
the description.
alex tested the boot disk, and got the same error on both snork and a
cluster machine, using either of the drivers. so I dunno...
seph
