Bugzilla – Bug 227384
(ATX board) power off button does not work correctly
Last modified: 2007-01-12 04:27:55 UTC
ASUS P3B-F (ACPI BIOS rev. 1006) it is the ATX motherboard. KDE 3.5.5. In opensuse 10.0 I was able to turn off the computer via "shutdown", in 10.2 I can almost do the same. The system is shutting down to the exactly last step but then... it just stays on. The last piece is missing -- real turning off the computer. And since opensuse shows by default splashscreen the user has no idea what is going on. It is impossible to see the log at the end because computer is frozen.
Please attach your /var/log/messages. Thank you.
Created attachment 109007 [details] /var/log/messages
Created attachment 109106 [details] /var/log/messages
I've got the same problem. After "system will be halted immediately" message computer is just stoped, but it's not turned off. No matter how I do shutdown (from GUI or console: shutdown -h now). I've got an old motherboard: ACORP 7KTA3. On 10.1 i was able to turn off computer. BTW, rebooting works OK.
I found workaround. For my system it works when I pass acpi=force parameter to kernel. Now my default /boot/grub/menu.lst entry looks like this: title openSUSE 10.2 root (hd0,3) kernel /boot/vmlinuz-2.6.18.2-34-default root=/dev/hda4 vga=0x317 resume=/dev/hda3 splash=silent showopts acpi=force initrd /boot/initrd-2.6.18.2-34-default
Works like you described. But previously (without acpi=force) the hardware off button kills the computer (cold-down, immediate turn-off) which was bad of course. With acpi=force the hardware off is completely ignored. Which is bad of course too (but I rather like the latter since it is safer for the system).
Now, you can immediately turn-off computer by holding power button down for a few seconds. You can also assign some actions for normal power button use (short one). See Yast->System->/etc/sysconfig->System->Powermanagement->Powersave->Events->EVENT_BUTTON_POWER
> Now, you can immediately turn-off computer by holding power button down for > a few seconds. a) it is a brute-force thing -- you can pull the power plug as well b) I don't want to do this (get this effect -- see (a)) c) I want to be able to properly shutdown the system > You can also assign some actions for normal power button use > (short one). See Yast->System->/etc/sysconfig->System->Powermanagement->Powersave->Events->EVENT_BUTTON_POWER a) it should be shutdown by default b) it is c) it does not work (same as in 10.0, worked in 9.x)
> a) it is a brute-force thing -- you can pull the power plug as well > b) I don't want to do this (get this effect -- see (a)) > c) I want to be able to properly shutdown the system True. > a) it should be shutdown by default > b) it is > c) it does not work (same as in 10.0, worked in 9.x) > True. Now (with acpi=force) on _my_ system it works
(In reply to comment #5) > I found workaround. For my system it works when I pass acpi=force parameter to > kernel. I have an old system too, but unfortunately the ACPI implementation is broken, so this is not an option for me. So far, using APM there has never been a problem with this and in previous versions it always did power off cleanly.
Known problem; just add "power-off" to kernel command line. (Changing kernel to work around this one would be ugly/incompatible; but could we make yast add power-off by default?)
Please test the Pavel's suggestion and report if you are successful.
power-off at grub level, I mean just like acpi=force? Does not work, I mean I still can't shutdown system via hardware button.
It is just getting better... I upgraded another computer from 10.0 to 10.2 -- Dell Latitude D610. In 10.0 power off button worked. Lid close -- worked (suspend to ram). In 10.2 none of them work. I retitled this report since it is no longer "some old ATX board" case.
Retitled again (back; sorry). This laptop story is quite other case -- see: https://bugzilla.novell.com/show_bug.cgi?id=230841 Superficially they looked the same but they are not the same. So, here is the case as described in the report and #4 comment.
*** This bug has been marked as a duplicate of bug 221667 ***