2003-09-05 03:21:33

by carbonated beverage

[permalink] [raw]
Subject: cpu not being found by 2.6.0-test4, input event bug too

Hi all,

Can anyone think of a change that went in recently to the 2.6.0-test*
tree that could have caused one of my CPUs to not be detected by the kernel?
It's a dual P2/333 system based on an Intel 440DX chipset, don't remember
the exact model of the motherboard off-hand. The latest BIOS update (rather
old) has been installed on the system already.

Debian/woody-based system.

The dmesg is also being FLOODED with the following:
barbeque/kittens:doc: dmesg|head -2 ; dmesg | tail -2
ue: 0
evbug.c: Event. Dev: isa0060/serio1/input0, Type: 2, Code: 0, Value: -3
evbug.c: Event. Dev: isa0060/serio0/input0, Type: 1, Code: 28, Value: 1
evbug.c: Event. Dev: isa0060/serio0/input0, Type: 0, Code: 0, Value: 0

As you can see from the truncation, the kernel buffer's being wrapped.
Once I get a login prompt and hop in, that's all there is. The dmesg
saved by the init scripts in /var/log/dmesg is also filled with just that.

tail -f'ing the syslog file when I was still enabling the logging of that
from evbug.c:evbug_event() pretty much scrolled by looking like a pretty
text flood.

As there's a TON of information, I put the config and other stuff
on the 'net at:

http://www.net-ronin.org/~ramune/kconf

-- DN
Daniel


2003-09-05 10:14:39

by carbonated beverage

[permalink] [raw]
Subject: Re: cpu not being found by 2.6.0-test4, input event bug too

Just tried reverting the one change I made, which was to the BIOS to
disable power management. Poof. 2nd CPU shows up when I reboot.

The dmesg flood is still going on, though.

Also, the motherboard is an Intel DK440LX (not a 440DX-based chipset).

Any other infor people need, or suggestions on what to look for to track
down the reason SMP won't work unless power management is enabled?

I'm assuming that means APM, considering when the motherboard was made...

-- DN
Daniel

2003-09-07 20:38:13

by carbonated beverage

[permalink] [raw]
Subject: Re: cpu not being found by 2.6.0-test4, input event bug too

Tried an older kernel (2.5.6x-series) and the event input kernel log flood
is happening there, too.

Can't really test for SMP, etc. as -test3 was the first kernel I tried since
2.5.4x that booted properly on this system, and didn't panic quickly on
various stuff. I can try a binary search across kernel versions, if I knew
what to test for.

Anyone have an idea?

-- DN
Daniel