Bug 21762 - Reboot on power cable plug in/remove - Samsung P560-AS04DE
Summary: Reboot on power cable plug in/remove - Samsung P560-AS04DE
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: ACPI
Classification: Unclassified
Component: Other (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: acpi_other
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-11-02 11:38 UTC by Rudolf Polzer
Modified: 2012-01-18 02:22 UTC (History)
3 users (show)

See Also:
Kernel Version: at least 2.6.31.14 to 2.6.36
Subsystem:
Regression: No
Bisected commit-id:


Attachments
ver_linux output, kernel config, modules list, and other possibly interesting stuff (33.13 KB, application/octet-stream)
2010-11-02 11:38 UTC, Rudolf Polzer
Details
acpidump output (159.79 KB, text/plain)
2010-11-03 10:38 UTC, Rudolf Polzer
Details
/proc/interrupts (1.55 KB, text/plain)
2010-11-09 09:54 UTC, Rudolf Polzer
Details
/sys/firmware/acpi/interrupts/* (3.91 KB, text/plain)
2010-11-09 09:55 UTC, Rudolf Polzer
Details
custom DSDT : remove SMM call in _Q51/_Q52 (288.56 KB, application/octet-stream)
2010-11-10 00:56 UTC, Zhang Rui
Details
dmesg output with DSDT overridden (39.66 KB, text/plain)
2010-11-10 10:34 UTC, Rudolf Polzer
Details
debug DSDT (288.49 KB, application/octet-stream)
2010-12-27 06:30 UTC, Zhang Rui
Details

Description Rudolf Polzer 2010-11-02 11:38:15 UTC
Created attachment 35892 [details]
ver_linux output, kernel config, modules list, and other possibly interesting stuff

My notebook is a Samsung P560-AS04DE. When I unplug the power cord (or plug it in), the system does a hard reboot. I tested the following kernels:

- v2.6.36 from linux-2.6.git with config from arch linux 2.6.35.8 package
- arch linux's build of 2.6.35.8
- arch linux's build of 2.6.31
- arch linux's build of 2.6.32.25
- v2.6.31.14 from linux-2.6.git with config from arch linux 2.6.31 kernel package does NOT BOOT AT ALL and ends with a panic which is longer than the screen, also is unrelated to my issue as v2.6.36 from git works (apart from the ACPI issue)

and the issue seems identical on all of them. Interestingly, Ubuntu 9.10's 2.6.31 kernel does not exhibit this problem, so this might rather be an issue that can be worked around by the kernel config (or is worked around by one of Ubuntu's patches). So I attached the kernel config I tested last (of v2.6.31.14).

Other actions that trigger an instant reboot (likely related, as these are ACPI modules):

modprobe processor
modprobe video

Well-intentioned workarounds that will not help me:
- acpi=off (then for some critical devices like the SATA controller, no interrupts are allocated and thus the system cannot boot)
- acpi=ht (crash still happens)
- BIOS upgrade (the updater requires Windows, and I do not have a Windows partition and not even free space where I could install Windows - also, ReactOS won't boot on this notebook and the Samsung BIOS upgrader is known to not work within BartPE)
Comment 1 Zhang Rui 2010-11-03 06:07:05 UTC
please attach the acpidump output of this netbook.
Comment 2 Rudolf Polzer 2010-11-03 10:38:09 UTC
Created attachment 35982 [details]
acpidump output
Comment 3 Len Brown 2010-11-09 02:36:10 UTC
upon the power cable plug or unplug, is it an immediate hardware RESET
that causes a reboot, or does the kernel actually go through
an orderly shutdown/reboot process?

# modprobe processor
or
# modprobe video

also provoke an immediate reboot?

what if you have a kernel with
CONFIG_ACPI_PROCESSOR=y
CONFIG_ACPI_VIDEO=y

does it boot and probe those drivers properly,
or does it provoke a rebot during boot?

please show /proc/interrupts and
grep . /sys/firmware/acpi/iterrupts/*
Comment 4 Rudolf Polzer 2010-11-09 06:38:40 UTC
It is an immediate hardware reset. Will try with the kernel configuration options, but I expect them to also cause an immediate reset on kernel startup.

Also, as an additional detail, the system does not power off on shutdown, but reboots (most likely the same issue).

In some hours (when I am on that notebook) I will attach /proc/interrupts and /sys/firmware/acpi/interrupts/*.
Comment 5 Rudolf Polzer 2010-11-09 09:54:45 UTC
Created attachment 36902 [details]
/proc/interrupts
Comment 6 Rudolf Polzer 2010-11-09 09:55:14 UTC
Created attachment 36912 [details]
/sys/firmware/acpi/interrupts/*
Comment 7 Rudolf Polzer 2010-11-09 11:31:35 UTC
I now tested compiling "processor" and "video" into the kernel. No crash on startup, surprisingly, and the respective directories do appear under /proc/acpi with sensible content.

However, even in that kernel, the system still does a hard reboot when I unplug AC. I will now try to build a kernel without the ACPI features that might have to do with unplugging the power cable (I'll first disable both battery and ac).
Comment 8 Zhang Rui 2010-11-10 00:56:14 UTC
Created attachment 36982 [details]
custom DSDT : remove SMM call in _Q51/_Q52

please apply this customized DSDT and see if the problem still exists.
Comment 9 Rudolf Polzer 2010-11-10 10:34:39 UTC
Created attachment 37052 [details]
dmesg output with DSDT overridden

The bug persists, even with overriding the DSDT. Both on power cable unplug, and on "modprobe processor", the system instant hard reboots.
Comment 10 Zhang Rui 2010-12-27 06:30:34 UTC
Created attachment 41662 [details]
debug DSDT

please try this DSDT instead.
Comment 11 Zhang Rui 2012-01-18 02:22:16 UTC
Bug closed as there is no response from the bug reporter.
Please feel free to reopen it if the problem still exists in the latest upstream kernel.

Note You need to log in before you can comment on or make changes to this bug.