Re: Re: Ethernet Trouble

At 02:31 AM 10/7/00 +0200, you wrote:

In message [off-list ref] you wrote:

quoted

Looking up port of RPC 100003/2 on 100.100.100.11
Looking up port of RPC 100005/1 on 100.100.100.11
VFS: Mounted root (NFS filesystem).
Freeing unused kernel memory: 40k init

It hangs here.

It seems like it is not running init and/or loading inittab.

Whats up with this?????

You're running vwith Root over NFS, so you can use  a  sniffer  (like
ethereal)  to  check  which  files  it's  trying to access. Check all
failing accesses...

Either it can't find init, or it's a shared library problem, or  your
console device is missing, or ...

        I did some tracing with ethereal. The first thing it does after
mounting the NFS drive look up the dev directory then the console. Next it
looks in the sbin directory (realy just a link to the bin directory) and
loads the init program. Finally it goes to the lib directory an loads
ld.so.1 (which is a link to ld-2.1.3.so). After this I can't decipher the
binary messages in the packets. Then it quits.

        Right now I think that it can't talk to the console or init uses a
function in ld.so.1 to spit a message to the console and is failing to load
the library correctly. Does PPCBOOT tell the kernel to use its console or do
I need to do that somehow? I have tried to stick "console=/dev/ttyS00" in
the bootargs but then I don't get any messages after the uncompresssing
kernel message.

        I am using Hardhat version 1.2 for my development and using thier
target directory for my NFS.

You may install a STATICALLY linked shell (like the standalone  shell
"sash" or the simple shell "ash") and pass the boot arguments

... init=/bin/sash
or	... init=/bin/ash

This will make the situation less complex...

        Tried this. It looked up the /dev/console just as before then It
loaded sash but, I still didn't get anything on the console screen.

Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
Any technology distinguishable from magic is insufficiently advanced.

        Sorry to be such a pain in your side lately.

        Conn Clark


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

        Right now I think that it can't talk to the console or init uses a
function in ld.so.1 to spit a message to the console and is failing to load
the library correctly. Does PPCBOOT tell the kernel to use its console or do
I need to do that somehow? I have tried to stick "console=/dev/ttyS00" in
the bootargs but then I don't get any messages after the uncompresssing
kernel message.

Unless its a typo, try "console=/dev/ttyS0" (1 zero, not 2).

Also, if its just a console print problem, you should still be able to
telnet into the box and do things like "cat <file> > /dev/ttyS0" and
"od -x /dev/ttyS0", just to show that the serial device is good and
the problem is with the console device attached to the serial port.

Your console,tty should look something like this:

ls -l console ttyS0

crw-------   1 root     root       5,   1 Aug  4 10:34 console
crw-------   1 root     tty        4,  64 Aug  4 10:34 ttyS0

I also like to use the following inttab file, instead of
default we ship.   It means there are more places for problems
to exist, but I've seen direct bash drop errors that getty
picks up, like "serial input buffer overrun".

cat etc/inittab

id:3:initdefault:

# System initialization.
si::sysinit:/etc/rc.sysinit

# Single user shell
1:2345:respawn:/sbin/getty console DT9600 linux

l0:0:wait:/sbin/halt -dip
l1:1:wait:/etc/rc.d/rc 1
l2:2:wait:/etc/rc.d/rc 2
l3:3:wait:/etc/rc.d/rc 3
l4:4:wait:/etc/rc.d/rc 4
l5:5:wait:/etc/rc.d/rc 5
l6:6:wait:/sbin/reboot -dip

cat etc/default/getty

CLEAR=NO


--
Michael Pruznick, michael_pruznick@mvista.com, 408-328-9215
MontaVista Software, 490 Potrero Ave, Sunnyvale, CA 94085
http://www.hardhatlinux.com, V:408-328-9200, F:408-328-9204

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/