LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - Hardware (https://www.linuxquestions.org/questions/linux-hardware-18/)
-   -   What device caused my system to wake? (https://www.linuxquestions.org/questions/linux-hardware-18/what-device-caused-my-system-to-wake-4175537448/)

malch 03-21-2015 11:25 AM

What device caused my system to wake?
 
I have been struggling to solve a problem with a system running Ubuntu that unexpectedly wakes up (from an S5 state).

I have searched the log files to no avail. Mr Google reveals others will the same problem, but no meaningful solutions as far as I can tell.

Even when I intentionally wake the system from S5 (via WOL for example) I am unable to find that event reflected anywhere in the logs. Booting with acpi.debug_layer=0xffffffff and acpi.debug_level=0x2 didn't seem to help either.

Any idea how to trace/catch the source of this wake event?

michaelk 03-21-2015 12:46 PM

Welcome to LinuxQuestions.

Once the computer is configured for WOL and shutdown the power on process is completely independent of the OS. The WOL magic packet is decoded by the ethernet adapter so would not expect any log event. Have you checked the BIOS for wake on events? Some have timers, power on after loss of main etc.

malch 03-21-2015 04:14 PM

Quote:

Originally Posted by michaelk (Post 5335637)
Once the computer is configured for WOL and shutdown the power on process is completely independent of the OS. The WOL magic packet is decoded by the ethernet adapter so would not expect any log event. Have you checked the BIOS for wake on events? Some have timers, power on after loss of main etc.

Yes, I have. Unfortunately, the BIOS was of no help. It does give me the option to disable RTC wakes which I have done. However, the system still wakes at random times.

I realize the OS is not running when the wake is initiated (and hence can't do anything). However, once the system starts booting up, the kernel can, through the ACPI protocols and tables, see what device initiated the wake. This is presumably how the Windows "powercfg -lastwake" command actually works.

ACPI is quite complex and deals with a lot more than simple power on/off events. And Linux includes pretty comprehensive ACPI support. Hence I suspect that the kernel is indeed getting that information. The information exists and is accessible -- the question remains concerning how to make the kernel reveal it.


All times are GMT -5. The time now is 01:05 PM.