Most recent kernel where this bug did not occur: n/a Distribution: Ubuntu/edgy Hardware Environment: DP965LT Board, Intel Core2 Duo processor, 512MB RAM, Nvidia graphic. Software Environment: Ubuntu desktop installation for AMD64, running in X86_64 mode. Problem Description: System is unable to power off. I can see messages: Power off. Acpi_power_off. Steps to reproduce: just trying to halt.
Is it possible to build a couple of older kernels to see if this is a regression versus a failure that has always existed? Probably 2.6.16 would be a good place to start.
Looks like being caused by some unneeded config options. The same kernel with different config does power-off correctly. (It seems, that omiting acpi-p-states for cpufreq helps)
I think it's the same bug as reported on the ubuntu launchpad https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/71040
Laurent, In the ubuntu report, you wrote about your acer travelmate 8000 series > ...this bug only appears if you connect to ac/dc power supply. > If you shutdown on battery supply it works correctly. And Adel confirms the same on an ... > Acer Aspire 1681 LCi Note that the DP965LT that this bug is filed against is a desktop board with no concept of AC vs DC. Exactly what config options makes the problem go away with a kernel.org kernel?
>Note that the DP965LT that this bug is filed against is a desktop board >with no concept of AC vs DC. Yes I know, but some other people also report this bug against an asus board (P4P800)
Hi, Lukas, Does this bug still exist in 2.6.22? >Exactly what config options makes the problem go away >with a kernel.org kernel? Can you provide this information? Thanks, Ray
Hello, the bug has been fixed in BIOS. Anyway, if e1000 driver is compiled as a module and the module is removed before boot, it is OK. (Linux driver leaves E1000 card in D3, while windows in D0)
Good news. Thanks. >the bug has been fixed in BIOS. disable the onboard nic? >Anyway, if e1000 driver is compiled as a module and the module is removed >before boot, it is OK. You probably mean reboot? So this's the symptom in 2.6.16 and before you fix it in the BIOS, right? Is this fixed in recent kernel release, with the same bios as 2.6.16?
> disable the onboard nic? no. the new BIOS for DP965LT board fixed some issues regarding freeze while power off. Kernel 2.6.19-rc1-git9 was the first that did not hang. 2.6.18 was broken. I don't know whether 2.6.16 is or is not ok. Board name: DP965LT BIOS version: 1458 2.6.18, 2.6.19-rc1 halts OK. BIOS version: 816 2.6.18 halts OK. BIOS version: 1162, 1176, 1250 2.6.18 cannot halt. (rmmod e1000 causes halt OK) 2.6.19-rc1 halts OK (no additional patches)
Hmm, this is already fixed in BIOS. So reject this bug and mark it as invalid. Please reopen it if you still have some problems. :)
I confirm, it is fixed and may be closed.