Ubuntu
linux-source-2.6.15 package

  1. Bug #43961
  2. Comment #103

Comment 103 for bug 43961

Revision history for this message
Michael Kofler (michael-kofler) wrote :

I had the same problem, i.e. after 'will now halt' the computer won't actually halt.

The solution for me was acpi=off!

Hardware is a PC with ADM Duron 1300, the BIOS from 2001 says it supports ACPI.

dmesg says:

> dmesg | grep -i acpi
[17179569.184000] BIOS-e820: 000000004fff0000 - 000000004fff8000 (ACPI data)
[17179569.184000] BIOS-e820: 000000004fff8000 - 0000000050000000 (ACPI NVS)
[17179569.184000] ACPI: RSDP (v000 AMI ) @ 0x000fa340
[17179569.184000] ACPI: Unable to map RSDT
[ 39.585043] ACPI: Subsystem revision 20051216
[ 39.585049] ACPI: Interpreter disabled.
[ 39.585070] pnp: PnP ACPI: disabled

I tried acpi=force, then switched to a i386 kernel. Both didn't help.

Then I found out that for some reason, the apm kernel module gets always loaded, no matter whether I use ACPI or not. The apm module is even loaded when I start the kernel with the option apm=no and with an additional line 'blacklist apm' in /etc/modprobe.d/blacklist. (Why is this?)

De-installing apmd is no option, because ubuntu-desktop depends on it ... (Why?)

OK, as I couldn't get rid of APM, I tried to get rid of ACPI.

Then I also tried acpi=off, because I thought, apm and acpi might somehow disturb each other. Bingo!

For reference:

> uname -a
Linux mars 2.6.15-26-386 #1 PREEMPT Thu Aug 3 02:52:00 UTC 2006 i686 GNU/Linux

It seems to me, that somehow apm and acpi disturb each other.

PS: I also tried the apm options power_off=1 and realmode_power_off=1, but these options didn't change anything on my machine. With ACPI (partly) on, they don't help. With ACPI off, they are not needed.