SGI O2 PANIC on mem=xxxM, and strange initrd behavior

SGI O2 PANIC on mem=xxxM, and strange initrd behavior

[Robin H. Johnson]

[-- Attachment #1: Type: text/plain, Size: 1631 bytes --]

Hi,

I've recently picked up an SGI O2, 300mhz R12k, 512Mb RAM.
Playing with getting it working, using Ilya's minimal patches on top of
the LMO CVS HEAD, I noticed that it wasn't pulling up the correct amount
of RAM. I threw mem=512M onto the params, and found a kernel crash.

Boots mostly fine:
http://www.orbis-terrarum.net/~robbat2/sgio2/kernel.boot
Crashes:
http://www.orbis-terrarum.net/~robbat2/sgio2/kernel.crash
Decoded:
http://www.orbis-terrarum.net/~robbat2/sgio2/decoded.panic
Kernel config:
http://www.orbis-terrarum.net/~robbat2/sgio2/kernel.config

I've also got one weird problem where it never gets any further than the
init call to my busybox initrd.

Output using an old debian netboot initrd (from debian-mips-2.4.19.img):
serial console detected.  Disabling virtual terminals.  init started:
BusyBox v0.60.3-pre (2002.01.22-06:31+0000) multi-call binary

Output for using bleeding-edge Gentoo initrd that works perfectly for
an Indy and an I2:
init started:  BusyBox v1.00-pre7 (2004.02.10-08:42+0000) multi-call binary

In both cases, right after the init message, absolutely nothing happens.
I've tried putting commands in the linuxrc that would cause some
externally visible activity, but nothing.

I am wondering if it may have to with the virtual consoles (which I had
to turn off to get any serial output), but I doubt they are the problem.

-- 
Robin Hugh Johnson
E-Mail     : robbat2@orbis-terrarum.net
Home Page  : http://www.orbis-terrarum.net/?l=people.robbat2
ICQ#       : 30269588 or 41961639
GnuPG FP   : 11AC BA4F 4778 E3F6 E4ED  F38E B27B 944E 3488 4E85

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: SGI O2 PANIC on mem=xxxM, and strange initrd behavior

[Ilya A. Volynets-Evenbakh]

As it is now, you are limited to 256M of RAM. pass mem=256M, or die ;-) 
I am working now on fixing that, but hit a brick wall in few places. 
We'll see by the end of next week though.

Also, pass console=ttyS0 to get serial output.

As for death after starting init, you may be running into problem
with R1[02]K, and your init just crashes in loop, or something like 
that. I get it from time to time.

    Ilya.

Robin H. Johnson wrote:

>Hi,
>
>I've recently picked up an SGI O2, 300mhz R12k, 512Mb RAM.
>Playing with getting it working, using Ilya's minimal patches on top of
>the LMO CVS HEAD, I noticed that it wasn't pulling up the correct amount
>of RAM. I threw mem=512M onto the params, and found a kernel crash.
>
>Boots mostly fine:
>http://www.orbis-terrarum.net/~robbat2/sgio2/kernel.boot
>Crashes:
>http://www.orbis-terrarum.net/~robbat2/sgio2/kernel.crash
>Decoded:
>http://www.orbis-terrarum.net/~robbat2/sgio2/decoded.panic
>Kernel config:
>http://www.orbis-terrarum.net/~robbat2/sgio2/kernel.config
>
>I've also got one weird problem where it never gets any further than the
>init call to my busybox initrd.
>
>Output using an old debian netboot initrd (from debian-mips-2.4.19.img):
>serial console detected.  Disabling virtual terminals.  init started:
>BusyBox v0.60.3-pre (2002.01.22-06:31+0000) multi-call binary
>
>Output for using bleeding-edge Gentoo initrd that works perfectly for
>an Indy and an I2:
>init started:  BusyBox v1.00-pre7 (2004.02.10-08:42+0000) multi-call binary
>
>In both cases, right after the init message, absolutely nothing happens.
>I've tried putting commands in the linuxrc that would cause some
>externally visible activity, but nothing.
>
>I am wondering if it may have to with the virtual consoles (which I had
>to turn off to get any serial output), but I doubt they are the problem.
>
>  
>