** KERNEL CRASH ON SHUTDOWN! ** Ran simple 'halt' and kernel died horrbily! Got following error dump (some cut out like register info) Code <1> Unable to handle kernel paging request at virtual address 00008875 current->tss.cr3 = 089ad00, %cr3= 089ad000 *pde = 00000000 Ooops: 0000 CPU: 0 <snip register contents> Process halt (pid 1119, process nr=46, STACKpage=c89dd000) <snip remaining register dumps and Code: dump> The halt (or even shutdown -h now) command, the death takes place exactly as it says it's stopping al md devices. It pauses for a second and then blows up. APM is off in the BIOS, APM was turned off via chkconfig and service stopped prior to shut down.
This is a bug in your BIOS, not in apmd (which isn't even running at that point) or the kernel. What happens is the following: - You tell the computer to shutdown. - SysVInit kills all running processes (including APMD) - the kernel shuts down, and tells the APM BIOS to turn off power - your APM BIOS is buggy, can't execute the request and doesn't return an error code either, so it crashes. This is normally related to the APM BIOS not handling 32-bit mode calls (that's why the problem doesn't occur with Windows). Turning APM off in the CMOS setup has nothing to do with this: it only turns off stuff like disk spindown. It is nothing to really worry about though: Since everything has been stopped, the crash is harmless. Consider it a fancy version of saying "It is safe to turn off your computer now". ;)