2002-06-25 00:01:12

by Shawn Starr

[permalink] [raw]
Subject: Linux 2.5.24 - Strange power off problems (ACPI support on)

Briefly, I have an Athlon MP 2000+ Asus A7M266-D (Dual CPU) system:

Here's the dmesg (this may also provide other info on other errors)


When I run /sbin/poweroff it also dumps to console that it had one
preempt left (I forgot the actual message it outputs).

The problem is this:

When I run the halt/poweroff program, it powers off the hard disks, and
displays the 'Poweroff System' message. If I press the power button, It
turns the system off HARD, Ie, i don't see the NIC card's aux power on,
and when I try to turn the machine on again it won't until I physically
remove all power (turn off powerbar) for a few seconds, then It will let
me power it on (or it will power on itself).

This did not happen in the 2.4.x kernel, I assume a new ACPI bug? ;-)

Shawn.


--
Shawn Starr, sh0n.net, <[email protected]>
Maintainer: -shawn kernel patches: http://xfs.sh0n.net/2.4/
Developer Support Engineer
Datawire Communication Networks Inc.
10 Carlson Court, Suite 300
Toronto, ON, M9W 6L2
T: 416.213.2001 ext 179 F: 416.213.2008


Attachments:
dmesg (11.92 kB)

2002-06-25 00:15:15

by Andrew Grover

[permalink] [raw]
Subject: RE: Linux 2.5.24 - Strange power off problems (ACPI support on)

> From: Shawn Starr [mailto:[email protected]]
> When I run the halt/poweroff program, it powers off the hard
> disks, and
> displays the 'Poweroff System' message. If I press the power
> button, It
> turns the system off HARD, Ie, i don't see the NIC card's aux
> power on,
> and when I try to turn the machine on again it won't until I
> physically
> remove all power (turn off powerbar) for a few seconds, then
> It will let
> me power it on (or it will power on itself).
>
> This did not happen in the 2.4.x kernel, I assume a new ACPI bug? ;-)

I'm heartened to see that ACPI PCI IRQ routing and CPU/IOAPIC enumeration is
working on your system. ;-)

2.4 ACPI is very different from 2.5 ACPI. You can try the latest ACPI patch
against 2.4 at http://sf.net/projects/acpi .

I bet it will do the same thing, though. I'll have to think about why this
might be happening.

Regards -- Andy