laptop switches off

sycamorex · 08-29-2009, 04:25 PM

This morning I installed Slackware 13 i386 and started configuring it. I've got a problem as my laptop has turned off unexpectedly twice since I started using Slack 13. I don't know if it's related to Slackware 13 but it has never happened before on this laptop with Slackware 12.2 or any other system.
The first guess would be overheating. I checked the laptop's BIOS but could not find anything related to cooling down. The laptop seems quite hot but it's nothing unusual for it. I also noticed that acpi is not installed by default. Would it make much difference?
The laptop is pretty new (6 months) and the cooling module seems to be clean. The only thing that differs from my previous installations is that for the first time I installed ATI's fglrx and enabled desktop effects. Could it be a buggy video driver?

tredegar · 08-30-2009, 10:37 AM

See if there are any interesting messages in /var/log/messages
If it's a software controlled shutdown, it should say why.
(Eg overheating -> Shutting down NOW)

Improve the cooling - does this stop the shutdowns?

sycamorex · 08-30-2009, 12:59 PM

It seems to be fine now. I left the laptop on overnight and so far so good. It didn't look like a software controlled shutdown as it just switched off as if you had removed all sources of power at once.
I didn't find anything suspicious in /var/log/messages. Unfortunately the logging system in BIOS is disabled.
I'll just wait and see if it happens again. I was thinking of logging in the temperature reading every 10 seconds. I've got some problems with it though. I don't seem to be able to read temperature. The /proc/acpi/thermal_zone is empty (I installed acpi). After some googling I'm still none the wiser:
lm_sensors are installed:

Code:

ls /var/log/packages/ | grep lm_sensors
lm_sensors-3.1.1-i486-1

Apparently, no sensors have been detected on my laptop:

Code:

# sensors
No sensors found!
Make sure you loaded all the kernel drivers you need.
Try sensors-detect to find out which these are.

After going through all the stages, sensor-detect also concluded that I've got no sensors whatsoever.

Code:

CONFIG_ACPI_THERMAL=m

Code:

ls /lib/modules/$(uname -r)/kernel/drivers/acpi
ac.ko  battery.ko  button.ko  container.ko  fan.ko  pci_slot.ko  processor.ko  thermal.ko  video.ko

Code:

lsmod | grep thermal
thermal                15096  0
processor              38600  2 powernow_k8,thermal
thermal_sys            10756  3 thermal,processor,video
hwmon                   2104  1 thermal_sys

Code:

less /proc/acpi/processor/CPU0
processor id:            0
acpi id:                 0
bus mastering control:   yes
power management:        no
throttling control:      yes
limit interface:         yes

acpitool

Code:

...

  Thermal info   : <not available>

   Device       S-state   Status   Sysfs node
  ---------------------------------------
  1. PB5          S5     disabled  pci:0000:00:05.0
  2. PB6          S4     disabled  pci:0000:00:06.0
  3. PB7          S5     disabled
  4. PB9          S5     disabled
  5. PB10         S5     disabled  pci:0000:00:0a.0
  6. PS2K         S3     disabled  pnp:00:07
  7. PS2M         S3     disabled  pnp:00:08

Code:

dmesg | grep -C 10 thermal

ACPI: Lid Switch [LID]
ACPI: WMI: Mapper loaded
i2c_core: exports duplicate symbol i2c_smbus_xfer (owned by kernel)
ACPI: Video Device [DVGA] (multi-head: yes  rom: no  post: no)
processor ACPI_CPU:00: registered as cooling_device1
ACPI: Processor [CPU0] (supports 8 throttling states)
processor ACPI_CPU:01: registered as cooling_device2
Linux video capture interface: v2.00
ACPI: I/O resource piix4_smbus [0xb00-0xb07] conflicts with ACPI region SMBI [0xb00-0xb0f]
ACPI: Device needs an ACPI driver
ACPI Exception (thermal-0377): AE_OK, No or invalid critical threshold [20081204]
piix4_smbus 0000:00:14.0: SMBus Host Controller at 0xb00, revision 0
rtc_cmos 00:05: RTC can wake from S4
rtc_cmos 00:05: rtc core: registered rtc_cmos as rtc0
rtc0: alarms up to one day, 114 bytes nvram, hpet irqs
jmb38x_ms 0000:0a:00.3: PCI INT A -> GSI 18 (level, low) -> IRQ 18
jmb38x_ms 0000:0a:00.3: setting latency timer to 64
uvcvideo: Found UVC 1.00 device HP Webcam (064e:c107)
input: HP Webcam as /devices/pci0000:00/0000:00:12.2/usb1/1-3/1-3:1.0/input/input8
sdhci: Secure Digital Host Controller Interface driver
sdhci: Copyright(c) Pierre Ossman

Any help on this?
thanks

I

tredegar · 08-30-2009, 04:05 PM

I never had any fun setting up lm_sensors

The first I knew of it was when I installed a new distro about 3-4y ago, and it kept printing "Over temperature PANIC" all over my terminal, when, in fact everything was OK. Something wasn't calibrated properly and I couldn't be a**ed to fix it.

I don't overclock, my MoBos are dusted sometimes, the PCs continue to work, so I have always disabled it. I hate being nagged. One of the reasons I quit win.

Bottom line: If there's nothing about forcing a shutdown in messages then it's probably a hardware fault. Try running memtest on a hot day, or put a blanket over it.

Last thought: It's a laptop: It has batteries / needs power. Was everything plugged in and switched on?
A couple of yrs ago I was halfway through a presentation when my linux laptop powered off.
Damn! I had plugged in the PSU (good), but not switched it on at the socket (bad).
To my great surprise, when power was restored, the presentation was where I left it (or it had left me).

sycamorex · 08-30-2009, 04:23 PM

Thanks for your reply.

Yeah, the laptop was connected properly and running on AC power. Right after it switched off I booted it and checked the battery indicator - it was 100% charged.

It hasn't happened again though so I'm getting less and less concerned. I'll do the memtest though.

cgtueno · 09-01-2009, 08:48 PM

Hi

Just reviewing the posts.
What make and model laptop are you using ?

It's curious that you are experiencing system shutdowns (which I take to be power off events), and have not experienced any systems hangs (which would be indicative of a hardware, or over heating problem).

On really old laptops I have seen problems with incorrect hardware responses to ACPI activities; powering off instead of entering standby, and vice versa type events), but not of late.

Is it possible that the external power supply failed ? (External PSU has an intermittent fault, external PSU flyleads are faulty, AC wall outlet is faulty, etc ?) The reason I ask is it sounds suspiciously like the laptop wound up running on batteries and ran out of power.

May I suggest that you thoroughly examine the external power supply and associated cables, and try using a different AC wall outlet. Also test the wall outlet with an AC test plug (in Australia these can be purchased from a local electrical products supplier - they indicate whether the wiring is correct).

Just a thought. I have encountered many laptops with faulty external PSU connectors, stretched leads, etc over the years. Often the simplest explanation is the cause.

Regards

Chris

sycamorex · 09-02-2009, 06:56 AM

Thanks for your suggstions.
I might be talking complete rubbish now but I'm inclined to think that it's something to do with acpi.
After it shut down twice on me on Saturday I installed acpi (not installed by default). The laptop was constantly on till last night (4 days). Nothing happened. I also successfully completed memtest86 (with a towel covering the laptop

) with no errors. Everything was brilliant.
Last night I booted into another partition with a fresh install of Slackware (without acpi installed by default) to play with fluxbox. Guess what? It shut down after 1 hour. As I said I don't know if it's possible that acpi (or some other power management program) can be responsible for it, but that's what my inexperienced mind would think.

business_kid · 09-02-2009, 01:23 PM

What's your cpu, and laptop?

The kernel will only find sensors it can load the modules for.
Here's what acpitool -t looks like on my box
bash-3.1$ acpitool -t
Thermal zone 1 : ok, 50 C
Trip points :
-------------
critical (S5): 105 C
passive: 95 C: tc1=1 tc2=2 tsp=100 devices=C000 C001
active[0]: 75 C: devices=C39B
active[1]: 65 C: devices=C39C
active[2]: 55 C: devices=C39D
active[3]: 40 C: devices=C39E

This is actually 4 speeds of the same fan. Here's the modules I load for the sensors (which I added to /etc/rc.d/rc.lm_sensors.conf)

MODULE_0=i2c-piix4
MODULE_1=lm90
MODULE_2=k8temp
MODULE_4=adm1021

If they are not in the kernel, the chip can be misdetected, or missed. You need the ones for your laptop.

sycamorex · 09-02-2009, 01:49 PM

It's HP Pavilion dv5-1110em and the CPU: AMD Turion X2 Dual-Core Mobile Processor RM-72.
lshw:

Code:

*-cpu                                                                                                                                                   
          description:CPU                                                                                                                                   
          product: AMD Turion(tm) X2 Dual-Core Mobile RM-72                                                                                                  
          vendor: Advanced Micro Devices [AMD]                                                                                                               
          physical id: a                                                                                                                                     
          bus info: cpu@0                                                                                                                                    
          version: 15.3.1                                                                                                                                    
          serial: NotSupport                                                                                                                              
          slot: Socket M2/S1G1                                                                                                                               
          size: 500MHz                                                                                                                                       
          capacity: 2100MHz                                                                                                                                  
          width: 64 bits                                                                                                                                     
          clock: 1800MHz                                                                                                                                     
          capabilities: fpu fpu_exception wp vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp x86-64 3dnowext 3dnow constant_tsc nonstop_tsc pni cx16 lahf_lm cmp_legacy svm extapic cr8_legacy 3dnowprefetch osvw skinit cpufreq

see my above posts for acpitool

business_kid · 09-03-2009, 03:03 AM

From post #1, You have slackware i386, a 32 bit os loaded. Come back to what I posted earlier
MODULE_0=i2c-piix4
MODULE_1=lm90
MODULE_2=k8temp
MODULE_4=adm1021

I very much doubt if you have either the k8temp module, which would be fairly critical for you. Neither would you have any of the 64 bit processor extended features. My box is similar. ACPI doesn't work on my box on that kernel. I suggest a 64 bit OS.

cgtueno · 09-03-2009, 10:21 AM

Hi

Have you considered removing Slackware from the equation by booting the device with DOS and leaving the machine to see if it resets while it is sitting idle ? (ie. indicating that it is some kind of hardware error).

Just get the machine to boot and display the date and time and let it stand idle to see if it reboots unexpectedly.

Or run a live CD distro and see if the same fault occurs.

Just a thought.

Chris

sycamorex · 09-03-2009, 01:37 PM

I created /etc/rc.d/rc.lm_sensors.conf and added the modules. The result of it was that sensors-detect actually gave me a 'yes'

Code:

Lastly, we can probe the I2C/SMBus adapters for connected hardware
monitoring devices. This is the most risky part, and while it works
reasonably well on most systems, it has been reported to cause trouble
on some systems.
Do you want to probe the I2C/SMBus adapters now? (YES/no): yes
Using driver `i2c-piix4' for device 0000:00:14.0: ATI Technologies Inc SB600 SMBus

Next adapter: SMBus PIIX4 adapter at 0b00 (i2c-0)
Do you want to scan it? (YES/no/selectively): yes
Client found at address 0x50
Probing for `Analog Devices ADM1033'...                     No
Probing for `Analog Devices ADM1034'...                     No
Probing for `SPD EEPROM'...                                 Yes
    (confidence 8, not a hardware monitoring chip)
Probing for `EDID EEPROM'...                                No
Client found at address 0x51
Probing for `Analog Devices ADM1033'...                     No
Probing for `Analog Devices ADM1034'...                     No
Probing for `SPD EEPROM'...                                 Yes
    (confidence 8, not a hardware monitoring chip)

Sorry, no sensors were detected.
This is relatively common on laptops, where thermal management is
handled by ACPI rather than the OS.

Still, /proc/acpi/thermal_zone is empty

Quote:

Have you considered removing Slackware from the equation by booting the device with DOS and leaving the machine to see if it resets while it is sitting idle ? (ie. indicating that it is some kind of hardware error).

Just get the machine to boot and display the date and time and let it stand idle to see if it reboots unexpectedly.

Or run a live CD distro and see if the same fault occurs.

It's been running ok for a few days now.
I'm going to try 64 bit as well.

business_kid · 09-04-2009, 06:39 AM

Okay, you have an sb600 southbridge, same as me.
Get a 64 bit OS. I have Slamd64 and Fedora installed. Either works. I had a bad time with ubuntu, which basically didn't work.
It had so many issues out of the box I just ran away.

As for /proc/acpi/thermal_zone, there's a kernel option for creating decpacated (=obsolete) /proc files for old software. That stuff now should be in /sys
cd /sys && find -name '*thermal*'

sycamorex · 09-04-2009, 06:52 AM

Thanks business_kid.
I was going to try slackware64 anyway

I installed it last night, did some configuration for a couple of hours and left it on. Before I went to work I checked it and it was fine. I'll see if it's ok after I get back from work.