shutdown "anomoly" with kernel 2.4.9

shutdown "anomoly" with kernel 2.4.9

[Michael G. Mobley]

Folks,

I thought I'd mention something weird I'm seeing with kernel 2.4.9...

BTW, I just recently subscribed to this list, but have scoured the
archives and didn't find anything pertinent to this particular problem.

When running kernel 2.4.9, my system cannot reliably reboot/
halt/shutdown.  It hangs when killall5 sends out the TERM signal, as if
init itself is terminating.  This is VERY repeatable (happens pretty
much every time, whether the shutdown is through 'shutdown -r now',
'reboot', 'halt', whatever...)

I wouldn't have thought this is a kernel problem, but, an identical build
of 2.4.8 does not seem to exhibit this behavior.  And I've played around,
making one change at a time, and that really is the ONLY difference.

BTW, I'm building both kernels to exactly the same config.  (diff of the
.configs show only three commented out sound card options that have been
added to 2.4.9 as differences.)

Brief system setup is:

1xPIII/800, Asus P3V4X MB, 640MB RAM
AHA29160 SCSI controller w/ 3 HDDs, 2 CDROMs
No IDE
USB, AGP, etc...
etc. etc...  (Can provide more details if needed)
GCC version is:  2.96
Binutils version is: 2.10.91.0.2
(Basically it's a stock RH7.1 install right now)

Regards, 
  Michael
-- 
                                              |  Michael G. Mobley
                                              |  mgm@havoc.gtf.org

Re: shutdown "anomoly" with kernel 2.4.9

[Ignacio Vazquez-Abrams]

On Thu, 20 Sep 2001, Michael G. Mobley wrote:

> When running kernel 2.4.9, my system cannot reliably reboot/
> halt/shutdown.  It hangs when killall5 sends out the TERM signal, as if
> init itself is terminating.  This is VERY repeatable (happens pretty
> much every time, whether the shutdown is through 'shutdown -r now',
> 'reboot', 'halt', whatever...)
>
> I wouldn't have thought this is a kernel problem, but, an identical build
> of 2.4.8 does not seem to exhibit this behavior.  And I've played around,
> making one change at a time, and that really is the ONLY difference.
>
> BTW, I'm building both kernels to exactly the same config.  (diff of the
> .configs show only three commented out sound card options that have been
> added to 2.4.9 as differences.)
>
> Brief system setup is:
>
> 1xPIII/800, Asus P3V4X MB, 640MB RAM
> AHA29160 SCSI controller w/ 3 HDDs, 2 CDROMs
> No IDE
> USB, AGP, etc...
> etc. etc...  (Can provide more details if needed)
> GCC version is:  2.96
> Binutils version is: 2.10.91.0.2
> (Basically it's a stock RH7.1 install right now)

Aha! I'm not alone and I'm not nuts. And it doesn't seem to be an Athlon/VIA
thing either.

This is exactly the same thing I'm seeing. And if I put calls to 'ps aux' in
/etc/init.d/halt around the killall5 and sleep calls then it shuts down
properly.

Here's my setup for reference:

1xAthlon 1050/100, Asus A7V, 512 MB PC133 RAM
2xATA/66, hda:CD-ROM, hdc:CDRW
2xATA/100 (Promise 20265 on-board), hde:HD (ext3/vfat), hdg:HD (vfat)
Stock RH7.1+Updates+2.4.9-ac12-preempt1

-- 
Ignacio Vazquez-Abrams  <ignacio@openservices.net>