2.6.1 kernel question

[Alan]

On Thu, 2004-01-22 at 10:56, Dave Jones wrote:
> On Thu, 2004-01-22 at 02:22, Alan wrote:
> 
> > > > Probably the version from the development tree is your best bet.
> > > > (currently 2.5.26-2). Or the ones from Arjan's page at
> > > > people.redhat.com/arjanv (which is a few revisions behind right now).
> > > 
> > > I just tried that.  Does not work.  Something is not functioning as it
> > > should.  I am going to run strace and a few other things and see if I
> > > can narrow it down as to cause.
> > 
> > Well, I found out what went wrong. Not certain what caused it.
> > 
> > /proc was no longer functional!  (Not certain why. Maybe this system
> > needs a /proctologist.)
> 
> There's a bugzilla open for this problem. It's ACPI related iirc.
> Similar problem occured with sysfs too. It's being looked into.

That would explain the two ACPI errors in my kernel log.  (I log kernel
messages to a separate non-messages location just to make it easier to
read.)

The triggering event is something in modprobeing of drivers that cannot
be loaded in the first place due to missing hardware.

Here is how you can reproduce it:

- Dual proc machine. (Mine is an dual Athlon 2100+ MP, if that matters.)

- Turn off the parallel port on the motherboard.

- MPS 1.1, not 1.4.  (1.4 has been evil in my experience.) Might not
effect this.

- Get the patched version of vmware for 2.6 and install.  The config
script will hang when it tried to unload the module(s) for the parellel
port.

Hope that helps.

-- 
"Push that big, big granite sphere way up there from way down here!
Gasp and sweat and pant and wheeze! Uh-oh! Feel momentum cease!
Watch it tumble down and then roll the boulder up again!"
    - The story of Sisyphus by Dr. Zeus in Frazz 12/18/2003