Created attachment 181431 [details] content of /var/log/dmesg My newly purchased Medion E4214 notebook hangs late in the reboot / powerdown sequence. The screen becomes black and nothing happens, i.e. for powerdown, the power stays on, and for reboot, it just stays there forever. I need to press the power switch for 5 seconds to switch it off hard. The linux installed is a standard xubuntu LTS 14.04 r2, where the kernel was changed to the 4.1 mainline kernel from the ppa repository (version 4.1.0-040100-generic to be precise) The original 3.16.0-43-generic kernel exhibits the problem as well. I switched off the grub splash screen, stopped most daemons and unloaded all needed modules to narrow down the problem. This is the output of lsmod: Module Size Used by i915 1146880 1 drm_kms_helper 126976 1 i915 drm 352256 3 i915,drm_kms_helper i2c_algo_bit 16384 1 i915 video 28672 1 i915 ahci 36864 2 libahci 32768 1 ahci BIOS is a AMI BIOS version N.1.01 or 2.17.1249 (both appear in the initial BIOS setup screen). I switched the boot mode from "UEFI OS" to "Legacy OS" to be able to install and run Linux. I tried the various reboot=... kernel commandline options, but neither works or exhibits any change. This might be related to bug #100221, but I do not see any errors/warnings in the boot messages.
Created attachment 181441 [details] Output of lspci -v
No immediate idea, please attach acpidump: # acpidump > acpidump.txt Thanks.
Created attachment 182151 [details] Output of acpidump
I attached the acpidump output. Is there anything else I can do, to debug the problem? Compile the kernel with special debugging options, running with other startup parameters, etc.?
The problem looks like to me is that the system does not power off or reboot after we write the value to the IO port according to ACPI table, i.e. to power off, we are supposed to write 0x07(according to _S5 ACPI object) to IO port 0x404:bits10-12, if it doesn't power off, it is either (a): we are not doing this right or (b): the value given by the ACPI table is wrong. I highly doubt it is (b), can you please check if Windows can power off or reboot if you happen to have Windows installed too?
Yes, Windows was preinstalled (I only shrank its partition to install linux) and it can reboot or power off just fine. And Windows is 64bit, just as linux is.
No idea, since you have already tried all the reboot= kernel cmdline options. Do you have vt-d turned on in your BIOS? I saw other places people mention this, maybe worth a check.
According to intel, the processor only supports vt-x, not vt-d, but the BIOS is so stripped down, that it does not allow any setting at all, except selecting the boot device. Now for the reboot options, what I tried where the following reboot options: b,f a,f k,f t,f e,f p,f and none of these work or show any difference. Did I miss something, or can I try something different?
(In reply to Nils Rennebarth from comment #8) > According to intel, the processor only supports vt-x, not vt-d, but the BIOS > is so stripped down, that it does not allow any setting at all, except > selecting the boot device. > > Now for the reboot options, what I tried where the following reboot options: > b,f > a,f > k,f > t,f > e,f > p,f > and none of these work or show any difference. Did I miss something, or can > I try something different? Not something I can think of, perhaps running Linux under EFI mode is worth a try.
I think you can check this simply with a LiveUSB instead of really installing it.
Bingo, that worked. Reboot as well as poweroff. I very much suspect, that the BIOS is a custom BIOS for this model and simply buggy for everything except EFI, because it was only tested with windows.
Almost all laptops are only tested with the pre-installed OS, i.e. Windows :-) I can imagine EFI may solve the reboot problem since it provides a way to do reboot, but I have no idea why poweroff also works now. Anyway, good to know this and I'll close it. I don't find a proper reason so I simply chose OBSOLETE, don't bother that.