Shut Down no more

tkoco · #1 21st May 2007, 11:57 PM

I recently did a total update of FC6. Prior to the update, I could do a shutdown from the Gnome desktop and the box would power down. Now, the desktop will go through the power down sequence and the screen shows that the box is halted, BUT, the box will not power down. I have to either hold down the power button until the BIOS shuts off the box, or hit the power switch in the rear of the box. What in the latest FC6 kernel changed to cause this sudden change in behavior?

multescugeorge@ · #2 1st June 2007, 07:47 AM

run in a terminal, as root if needed
poweroff
see if it works

edzxp · #3 4th June 2007, 10:15 PM

Quote:

Originally Posted by tkoco

I recently did a total update of FC6. Prior to the update, I could do a shutdown from the Gnome desktop and the box would power down. Now, the desktop will go through the power down sequence and the screen shows that the box is halted, BUT, the box will not power down. I have to either hold down the power button until the BIOS shuts off the box, or hit the power switch in the rear of the box. What in the latest FC6 kernel changed to cause this sudden change in behavior?

I have the same problem since fedora 6, I have tried all the suggested boot paramters including acpi=force on my dual opteron Tyan k8w tiger. Does anybody have further ideas or a fix?

cr4ck3r · #4 7th June 2007, 02:05 AM

I have the same problem with my F7 installation. Does anyone have a solution for this?

Thetargos · #5 7th June 2007, 08:07 AM

Now this is odd... This is usually due to ACPI and/or AMP conflicting. The only way I've been able to get around these issues has been with a custom kernel and APM completely disabled, but I believe APM is disabled in the main Fedora kernel since 2.6.5 or so (FC2), alternatively you may want to check you Motherboard's settings for power management. I stopped using the menu to shut down the PC and use the power button which will cause Fedora to go through the shutdown process (by signal interception), very convenient.

cr4ck3r · #6 7th June 2007, 03:53 PM

Or should we file a bug against this issue?

Thetargos · #7 7th June 2007, 04:47 PM

I'd say, file it. Or check if someone else has. It maybe useful to have a couple hardware profiles on which this problem has been seen too.

stepher · #8 7th June 2007, 06:23 PM

My system is a PC Chips M810LR (AMD K7) so it's not an x64 CPU. However, I've noticed similar problems with both FC6 and FC7 (FC7, BTW, was a breeze to install and update over FC6. Thanks folks!!!) in that the system gets all the way to the "System Shutdown" message and then never turns off the machine. The hardware did turn off with FC5. I have changed nothing on the M/B since FC5.

Any help is appreciated.

Thanks in advance...Steph

Thetargos · #9 7th June 2007, 06:27 PM

Deffinitely a time for bug reports, please people, someone file a bug report on bugzilla and post back the bug number so others can add info and hardware profiles to that report, so the devs can solve it ASAP.

stepher · #10 8th June 2007, 08:12 PM

I wrote earlier (see below) that my PC Chips M810L shut down when loaded with FC5, but not with FC6 or FC7. I spoke to a friend yesterday who loaded FC7 onto an ASUS A7V133 M/B and he reported his system does just fine (shut down, that is). The BIOS rev on his machine is Rev. 1009 which he said he thinks is the latest (in case that helps anyone). I asked him what the last message was on the screen right before the computer turned off and, while I can't remember what he said, it certainly wasn't "System Shutdown" (which may only mean that the power disappears moments after the message is displayed but not long enough to see it).

So, while it looks like the problem is more wide spread than most of us would like, it appears that FC7 is just fine in some situations.

Cheers to all...Steph

Quote:

Originally Posted by stepher

My system is a PC Chips M810LR (AMD K7) so it's not an x64 CPU. However, I've noticed similar problems with both FC6 and FC7 (FC7, BTW, was a breeze to install and update over FC6. Thanks folks!!!) in that the system gets all the way to the "System Shutdown" message and then never turns off the machine. The hardware did turn off with FC5. I have changed nothing on the M/B since FC5.

Any help is appreciated.

Thanks in advance...Steph

Thetargos · #11 9th June 2007, 03:57 AM

The problem is with how does the BIOS interprets the ACPI commands sent by the kernel. It is not a problem as such, as with AT systems you had to manually shut down the machine, and pretty much the sequence was the same. It may required to use either a special kernel command-line argument for the boot process, so that the BIOS will react better to system calls to the ACPI (power management) system, and as such successfully powering off the system. The problem is that in some cases this is local to the hardware/software combination, try any of the ACPI kernel command-line arguments to see if they affect in any way the outcome when sutting down... particualrly try 'pic=noacpi' (without the quotes) and see how it goes.

stepher · #12 9th June 2007, 03:38 PM

Hi Thetargos...

Thanks for the suggestion(s) and the pointers. I'll read up on kernel commands re:ACPI, perform some experiments on my system and report back with my results.

BTW - While I agree that it may be h/w-s/w specific or some combination, do we have any idea how the kernel might have changed from FC5 to FC6/7 (since things shut down just fine in FC5).

Thanks again and great weekend to all....Steph

Quote:

Originally Posted by Thetargos

The problem is with how does the BIOS interprets the ACPI commands sent by the kernel. It is not a problem as such, as with AT systems you had to manually shut down the machine, and pretty much the sequence was the same. It may required to use either a special kernel command-line argument for the boot process, so that the BIOS will react better to system calls to the ACPI (power management) system, and as such successfully powering off the system. The problem is that in some cases this is local to the hardware/software combination, try any of the ACPI kernel command-line arguments to see if they affect in any way the outcome when sutting down... particualrly try 'pic=noacpi' (without the quotes) and see how it goes.

stepher · #13 9th June 2007, 05:57 PM

I had a quick moment to try the 'pic=noacpi' suggestion and that didn't seem to help. When I get a bit more time I'll go do some more experimenting.

In the mean time, I came across this thread

Fedorahttp://readlist.com/lists/vger.kernel.org/linux-kernel/15/77088.html

which might mean something to someone with a better background than me in Fedora. It seems like this is a problem that may have crept in unknowingly by upgrading or "backporting" certain codes and routines. Can anyone make sense of this?

Cheers...Steph

Quote:

Originally Posted by stepher

Hi Thetargos...

Thanks for the suggestion(s) and the pointers. I'll read up on kernel commands re:ACPI, perform some experiments on my system and report back with my results.

BTW - While I agree that it may be h/w-s/w specific or some combination, do we have any idea how the kernel might have changed from FC5 to FC6/7 (since things shut down just fine in FC5).

Thanks again and great weekend to all....Steph

Thetargos · #14 10th June 2007, 02:08 AM

Oops, my fault, the argument should be pci=noacpi, sorry.

cr4ck3r · #15 10th June 2007, 03:58 AM

I filed it https://bugzilla.redhat.com/bugzilla....cgi?id=243579
Hope that somebody will fix it soon.