I'm unable to get 2.2.12 (it's my first 2.2.x attempt) to run on a SS20,
currently running RedHat Linux 5.1 and kernel 2.0.35. The main difference
between myself and most of the rest of the world is that I MUST be able to
manage this machine via a serial console. Thus, the following inittab entry:
s2:12345:respawn:/sbin/getty ttyS0 9600
I took RedHat's kernel .config file for a single-processor setup, and built
and installed 2.2.12 painlessly. I added the image to silo.conf, and booted
from the new kernel. I got as far as:
Turning on user and group quotas for local filesystems Setting clock: Sun Sep 5 18:09:11 PDT 1999 Enabling swap space. Initializing random number generator... zs_open ttyS0, tty overwrite. zs_open ttyS0, tty overwrite. zs_open ttyS0, tty overwrite.
[snip about 20-30 of these]
zs_open ttyS0, tty overwrite. zs_open ttyS0, tty overwrite. eth0: Carrier Lost, trying TPE
At this point the machine is ping-able, but does not seem to fire up any
services, and I don't see anything more which looks like a boot sequence on
the serial console. So I hit return on the serial link just to see if it would
still scroll.
The kernel OOPSed at this point. All kinds of nasty stuff written to the
console, which I could have hardly transcribed to paper and which is now lost
forever :-(. I tried sending a few breaks via minicom, but whatever they did,
if anything, I never got back to the PROM's "ok" prompt. Eventually the screen
just settled into a tight, seemingly uninterruptible loop:
Unable to handle kernel paging request at virtual address 40001000 Unable to handle kernel paging request at virtual address 40001000 Unable to handle kernel paging request at virtual address 40001000 Unable to handle kernel paging request at virtual address 40001000 Unable to handle kernel paging request at virtual address 40001000 Unable to handle kernel paging request at virtual address 40001000
which scrolls endlessly, very rapidly, and I was unable to get a BREAK to get
me back to the PROM.
Someone else has seen this problem and posted on sparc-list@redhat.com:
Message-ID: <Pine.LNX.4.10.9908300852270.21084-100000@brechin.fpcc.net>
Kirk Rafferty <kirk@rafferty.org> said:
> I've got a Sparc 20 with 2 60Mhz Sparc processors running Red Hat
> Linux 5.2. I've put the packages Red Hat recommends for running a 2.2
> kernel on 5.2. I'm running the machine headless, and when I boot the
> kernel (anything from 2.2.5 to 2.2.12) I get this:
> zs_open cua0, tty overwrite. zs_open cua0, tty overwrite. zs_open
> cua0, tty overwrite. zs_open cua0, tty overwrite. zs_open cua0, tty
> overwrite. [ ... About a zillion times and then ... ]
> Unable to handle kernel NULL pointer dereference<1>tsk->mm->context =
> 00000001 tsk->mm->pgd = f2e35000
> \|/ ____ \|/
> "@'/ ,. \`@"
> /_| \__/ |_\
> \__U_/ ifup(125): Oops PSR: 00900fc1
> PC: f0141e24 NPC: f0141e28 Y: 10e69800 g0: f0041d64 g1: 00400fe3 g2:
> 00000021 g3: 00000021 g4: e00e1040 g5: 00000000 g6 : ff00e000 g7:
> 00000000 o0: 00000001 o1: 00000003 o2: f2f18130 o3: e00e1050 o4:
> 00000000 o5: 00000002 sp : ff00fc70 o7: f0141dd8 l0: 0000000d l1:
> f2f18000 l2: f018fde0 l3: f018fc00 l4: 00000001 l5: f017f800 l6
> I seem to recall seeing this problem, and the fix being to turn the
> getty off the serial port, but this is not an option for me. Anyone
> know a way around this?
Any advice on how to troubleshoot this welcomed, with thanks. Is this likely
to be a genuine kernel bug? If so, I guess my frightening task will be to try
to get some more firm diagnostics on the exact oops . . . frightening because
I risk losing the ability to manage/reset the machine.
-Darren
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/