laptop overheats and suddenly shuts down/off

I can confirm this issue in a Sony Vaio VGN-NS150J.

Also, battery time has been considerably reduced from Ubuntu 8.10. Powertop output:

Top causes for wakeups:
  96.1% (43222.2) <interrupt> : extra timer interrupt
   2.0% (887.4) pidgin : hrtimer_start_range_ns (hrtimer_wakeup)
   0.6% (286.6) firefox : hrtimer_start_range_ns (hrtimer_wakeup)
   0.5% (229.4) <kernel core> : hrtimer_start_range_ns (tick_sched_timer)
   0.3% (115.8) amarok : hrtimer_start_range_ns (hrtimer_wakeup)
   0.2% ( 98.6) <interrupt> : HDA Intel
   0.0% ( 21.6) gnome-panel : hrtimer_start_range_ns (hrtimer_wakeup)
   0.0% ( 21.0) skype : hrtimer_start_range_ns (hrtimer_wakeup)
   0.0% ( 20.0) <interrupt> : ahci
   0.0% ( 15.2) <interrupt> : iwlagn
   0.0% ( 10.0) async/0 : timer_action (ehci_watchdog)
   0.0% ( 5.6) python : schedule_timeout_uninterruptible (process_timeout)
   0.0% ( 5.4) <interrupt> : i915@pci:0000:00:02.0
   0.0% ( 4.2) VBoxSVC : hrtimer_start_range_ns (hrtimer_wakeup)
   0.0% ( 3.6) <interrupt> : acpi
   0.0% ( 2.4) <kernel core> : hrtimer_start (tick_sched_timer)

I have seen this behaviour for the first time today. I've had Jaunty running for a couple of weeks on my Compaq Armada M700 (PIII-800). Today I installed the latest batch of updates (I usually update whenever it tells me, so it should only be the last day or so's worth of updates) and then noticed that after gentle web browsing (not even video clips/flash) the laptop was insanely hot and the fan wasn't running.

I also had constant notification popups saying something about not being able to read 'sensor CPU C19A' from /proc/cpu/therm...(? from memory).

When it's cooled down I'll try it again and add more detail.

Is there any way to find out what updates have just been installed in the latest batch?

If I hadn't spotted it this could have been nasty for the laptop, or for any heat-damageable furniture it was sat on.

OK, after a cool down and reboot:

It was /proc/acpi/thermal_zone/C19A/<something> it was popping up and complaining about. After the reboot it hasn't been warning me. It also seems to be running cooler, and I've felt the fan running.as well.

I guess the update left it in an odd state.

I've now tested it with cpuburn (actually burnMMX) and it went up to 78C with the fan going full belt, then cooled immediately when I stopped burnMMX. It felt *a lot* hotter when it overheated than it did after running at 78C for a few minutes...

I've attached syslog for the time from when I did the update (19:10) to when I manually shut it down as it overheated (19:30) then the next boot (19:58) to now. Some interesting messages in there.

If you need any other files let me know.

I had this issue last week when I installed Ubuntu 9.04, on my Gateway laptop. My specifications are:

Gateway T-1625
AMD Turio 64 X2 TL-60
Ubuntu 9.04 Jaunty Jackalope 32 bits

But I followed this thread http://ubuntuforums.org/showthread.php?t=597998 and at least my laptop doesn't suddenly shutdown anymore, it's hot sometimes anyway, but now, I'm constantly monitoring the activity of the CPU to prevent that from happening. I hope in the future we don't have to adjust this and look for help with everyone. I love Ubuntu but I will try others Linux distros that support better my laptop.

Hope this can help anyone who go through the same issue.

I can confirm this issue on my hp tx2110us, using Jaunty 64.

When using some intensive app for a minute or two, processor temp goes over 90 Celsius. When using the same app on Intrepid 64, everything works fine.

My acer aspire 5315 runs at 80 celsius idle. After watching videos or using flash it shuts down giving an overheated promt. Having a dualboot with intrepid, hardy and vista, but I only encounter this problem in jaunty.

Acer Aspire One D150--Fresh install
Same issue

I have the same problem with Acer Aspire 5315. I installed Ubuntu 9.04.

Same problem with MSI Megabook L715
AMD Turion 64 mobile ML-40
Ubuntu 9.04 Jaunty Jackalope 64 bits

Just a comment: i've downgraded to intrepid 64 and it does not overheat anymore.

AMD Turion 64 X2 TL-62

same with lenovo x200s with jaunty

My laptop has not reached a critical temp yet, but I can feel alot of heat coming from the keyboard when idle. This was not the case with hardy-32bit (Clean install of jaunty-32bit currently)
AMD Turion 64 ML-32
ATI 200m (currently using xorg-ati drivers, while in hardy I was using fgrlx, which is no longer available for my video card now)

My HP dv1430us has done it about 4-5 times a day (acpi critical_trip_point) despite me altering the fan points, frequency stepping, etc. It did it this morning while I was reading text on a non-cpu intensive website. It also seems to have corrupted grub and made my entire installation useless, so I can't even post system logs.

I would like to note that this is *not* a hardware issue, I ran 8.10 for months without a hiccup.

got this with a dell inspiron 640m, intel core solo processor with integrated graphics

Having the same problem:

Could it be nVidia graphics card?

Specs: HP dv5242ea | Core Duo 1.60GHz | 2GB RAM | 320GB SATA | DVD-RAM Matshita UJ 840S | nVidia GeForce Go 7400 256mb | Intel PRO 3945ABG

i don't think its nVidia because my x200s has intel integrated

Heat issues partially solved on 2 laptops both have Intel 965 chips

gateway 6860 fx w/ 9.04 64bit 2.6.28.12-generic

acer extensa 5620-6830 w/ 9.04 64bit 2.6.28.12-generic

both have custom DSDT files w/ 0,0,0 problems reported in iasl

both use acpi_osi="Linux" permanently added to /boot/grub/menu.list

both have been taken down to motherboard for cleaning and re-assembled

6860 video is nvidia 185.18.10

5620 video is integrated 965

Same slow fan issue on both systems, the fans NEVER STEP HIGHER THAN 66%

I can reboot the 6860 from JJ64 with sensors reading 78C on 8800m video with fan barely blowing boot into vista 64 ultmt where the fan immediately comes on and stays on for 5-6 minutes then slows down to JJ64 like speeds, I reboot into JJ64 and within 1/2 hour the temps are back up.

Keep in mind that ntune doesnt support fans on the 8800M so I cant change it in windows either but the fan does spin up to full speed and in JJ64 it never does.

The 5620 fans never seem to get above idle-33% ever for any reason (I just shut it down @ 65C. In my mind no hardware that isnt abused due to neglect or environment should see 75C ).

I'm happy to submit any info for either system that will be helpful.

Like CylnZ said, my fan seems to not spin as fast as it used to.
I don't have a way to get actual numbers, but I can tell by the sound the fan makes that it doesn't reach its maximum speed now when at 100% cpu for a couple of mins.

I can confirm on two different laptops:
After upgrading to jaunty 9.04 they both run overheat even when closed
Acer aspire 5630 with 2.6.28.11-generic | intel core duo 1.66GHz
Acer aspire 5730Z with 2.6.28.11-generic | intel core duo 2.16GHz
both are integrated video

This is definitely an os issue. Both were fine running on gutsy and intrpid.
Both do not overheat when I boot into windows.

same bug?
https://bugs.launchpad.net/ubuntu/+bug/361123

I found a very clumsy workaround. Either I unplug the battery or the AC and the temperature stays around 55-65 celsius. So could this be a charging issue mainly?

Jaunty
Linux 2.6.28-11-generic
Intel(R) Celeron(R) CPU 540 @ 1.86GHz
ACER aspire 5315

This is the same bug mentioned above (e13) - https://bugs.launchpad.net/ubuntu/+bug/361123

The issue is the fan speed. NOT that the processor is working at capacity (set to performance, etc). My Lenovo T61 frequently serves as a desktop replacement. Even with the cpu throttled it runs not. Overnight backups, or software like BOINC or even a flashy screen saver can all send the laptop into thermal shutdown. If I want to run my cpu at it's full speed, the laptop should be capable of that without overheating. Dumbing it down is not the answer.

Here is the workaround for Lenovo, which involves disengaging acpi control of the fan so that it runs at the correct speed. If you look at your fan speed, you will see that it is *slow* and that it gets *delayed* turning on when the cpu overheats. This is why the cpu overheats - the fan is too slow and does not turn on soon enough - at least in my case.

A person should be able to run cpu intensive applications overnight without causing their laptop to go into thermal shutdown!

 I found a work around, which is to let the fan run full speed all the time. (This can also be adjusted easily)

The laptop is now running at 44C as I surf and write this post, and goes to 94-95C under a cpu stress test but no higher.

1. you can monitor your fan speed from the thinkpad-acpi module: cat /proc/acpi/ibm/fan (check path under /proc for your laptop)

2. you can adjust the fan speed to stay on at full speed by executing the following command as root: echo level disengaged > /proc/acpi/ibm/fan

(you will need to uninstall any other fan control applications or scripts you may have tried, including tpfand)

One I did that, the fan kicked up to 4000+ rpm and stayed there. The CPU temp stays at 94C under full load and goes no higher. There are additional options, and scripts you can run if you wish to control the fan based on temperature and load.

With the fan running between 4000-5000 rpm cooling is not a problem.

When fan control is engaged in the thinkpad-acpi module, the fan spins *only* between 0 and about 3000 rpm and takes quite a while to kick in - usually too late

Note that I tested the same software on my wife's T61 running 8.04 and did *NOT* have any thermal issue. I have replicated my heat problem on the T61 using both the 64-bit and 32-bit kernels.

For more info, see the following: http://www.thinkwiki.org/wiki/How_to_control_fan_speed

@CharleyS:

Your fix sounds encouraging, but are there any side-effects of leaving it on all the time like that?

Is there no way of using a different setting in Jaunty that controls the fan a bit better. I mean I had no such problem in Intrepid?

@e13:
Yeah, seems the same bug, but not only on Lenovo, so I marked #361123 as a duplicate, even though it was registered earlier.

@David:
This might be weird and subjective, but I've noticed that if the battery is unplugged temperature rises fast and high. With the battery laptop is more stable. Need to check and confirm that though.

I think you may have something with the battery. After reporting this a while ago, my Compaq Armada M700 has not drastically over heated again - although it has been running 10 degrees or so hotter than it did with Gutsy-Intrepid (58 not 48 when idling).

Until last night. I usually run it plugged in, with the battery in place. I noticed the orange 'battery charge' light was on which is rare, then a few minutes later there was a 'pop' (I think just the speakers) and it froze. When I held it to turn it off it didn't respond to the keyboard, mouse or soft power button, and it was scorching hot.

After a hard reset, then a cool down, it's been fine again. No charging light.

Just to confuse things even more...

Since I created the original bug report I've not witnessed the behavior EXCEPT when I have the laptop plugged in w/ battery and connected to a 2nd monitor with Xinerma enabled. I have a feeling that, in that situation, the temperature increase is probably more due to the video card working extra hard for dual-monitor rather than any CPU freq scaling issues.

Again - This is NOT a "cpu scaling" problem. Unless you are OVER clocking your cpu, you are running it at the speed it was DESIGNED to operate at.

The issue is cooling.

The FAN is not running fast enough, or often enough, or at the right time, to sufficiently cool the processor.

This is NOT a CPU "scaling" issue.

It is a heat issue caused by the fan control (ACPI).

You also need to REPLICATE the issue with a CONTROL.

If you just "wait" for the bug to "happen" you will ge nowhere.

I can replicate the issue EVERY time by running stress.

sudo apt-get install stress

Then run it: stress --cpu 16 --vm 2 --vm-bytes 128M

Watch your laptop shutdown.

I guess that it can be a cpu scaling problem, 'cause Jaunty always boot in PERFORMANCE governor, instead of the default ondemand.

The bug is labled as an overheat and shutdown issue. Normal CPU's do not overheat when run at their designed speed. "Performance" does NOT mean "over clock". It simply means the cpu is running at it's highest (and least battery friendly) mode.

This should have nothing to do with an overheating problem.

The cpu is designed to run comfortably at it's rated speed so long as the cooling is adequate as designed.

No, it is NOT a cpu scaling problem. It is a head problem.

Booting in one mode or another is an entirely different issue. The laptop should NOT be overheating in ANY mode. Period.

I'm also affected; a Thinkpad R61i model 7650-range, preinstalled Ubuntu 8.04, currently running kernel 2.6.24-23, as had boot issues with -24.
(probably got the most recent ACPI patches for Hardy)
Fan running slowly - was cooking my laptop more.

I have had to buy a laptop cooler so I can safely use it.

PLEASE fix it for us on Hardy as well!

@CharleyS
I can successfully replicate the issue with three or four opened Flash pages in browser, or Skype with video. So no 'stress' need for me. Who could do the same?

@Alex Cockell
Hmm, we should somehow ascertain that in Hardy you have the same issue, because we experience this only since Jaunty. When your Hardy started to cook your laptop, Alex, — since the beginning or not so long ago?

@ Idle sign - since not so long ago. So it's definitely recent changes that came down.

Being a "regular user", I generally trust the release people to get it right - I am currently running the machine atop a cooler - which at least makes it usable.

Temperatures are still around what I quoted... but before I bought the machine - it would freeze up... or not boot due to heat during the last mini-heatwave...

Although I did notice it getting rather warm around the CPU fan outlet.. but I would have expected that to ramp the fan if it WAS getting warm...

When was the last batch of acpi patches released?

What was VERY scary was that the following lines started showing in my /var/log/messages...

May 30 22:32:21 ubuntu -- MARK --
May 30 22:52:21 ubuntu -- MARK --
May 30 23:10:05 ubuntu kernel: [531782.411921] thinkpad_acpi: unhandled HKEY event 0x6022
May 30 23:32:21 ubuntu -- MARK --
May 30 23:52:21 ubuntu -- MARK --

Jun 2 23:52:22 ubuntu -- MARK --
Jun 3 00:12:22 ubuntu -- MARK --
Jun 3 00:32:22 ubuntu -- MARK --
Jun 3 00:50:15 ubuntu kernel: [644574.555901] thinkpad_acpi: unhandled HKEY event 0x6022
Jun 3 01:06:48 ubuntu kernel: [645562.273688] thinkpad_acpi: unhandled HKEY event 0x6022
Jun 3 01:07:44 ubuntu exiting on signal 15

... and *didn't* shut down.

This happened a lot when the lappie overheated... I've got Linux Emporium investigating it as well (preinstalled machine from them).

Looks as though it's known about... http://marc.info/?l=git-commits-head&m=123213959414589&w=2

In both cases - the fan ought to kick in and run at close to max speed - it doesn't. Only speed up from, say 3003rpm to 3050 for a few seconds. I'd have expected to feel the fan really ramp up and run at abotu 4000-5000 or even max out (I think they're rated to approx 5500rpm for a long time ot get the temp right back down before slowing back down again. Hence the cooler...

I use this thing as a desktop replacement, so often run it without the battery on mains, and often all day.

From the above - possible leads - thinkpad_acpi.. needs to be released to ALL VERSIONS.

PLEASE!!!!!

I have collated all the various posts into one ubuntuforums post...please help advice there...

http://ubuntuforums.org/showthread.php?p=7399158#post7399158

This is NOT just a thinkpad issue. Both my laptops are overheating and they are not thinkpads. It happens every time, with battery or without. I don't even have to run any programs, just walk away from it and it gets so hot you can't touch it.

I have to work in windows until this is fixed #@$%
There are many who are affected by this bug. It should be marked as high importance.

Please, pretty please! with cherries on top

   I also see this issue on my lenovo t61p. Hard for me to tell if it happens all the time or just occassionally. I will notice the keyboard burning up and realize that my temps are at 90C even though I am not doing anything processor intensive. No 3D (I don't use compiz or other compositing). Fan is sitting at a fairly low speed. When I switch the fans to manual unlocked mode the temps settle down to 58/69 (temp1 and temp2 respectively) which still seems pretty hot especially for max cooling.

   I think this might be a contributor to the periodic instabilities I see under Jaunty. I'll start paying more attention to the temps.

James

The fan fix suggested sounds interesting...but there is nothing under the /proc/acpi/fan folder.

@abhiropb -

On my Lenovo it is /proc/acpi/ibm/fan - using the ibm acpi module.

On your laptop it may be different. Look under /proc/acpi for something else. I thing there is a module called acer-acpi.

I have the following under /proc/acpi:

Folders:
battery
embedded_controller
fan
processor
video
button
ac_adapter
powinfo
sleep
wakeuper_resource
thermal_zone

Files:
event
dsdt
fadt
info
sleep
wakeup

The folders and files are mostly empty (the fan directory is empty).

Also I'm using a HP laptop

I seem to be having this problem on a Toshiba Satellite A200 1 GB. The CPU heats up to 78C (in Windows it never gets past 70C) at max usage, then the laptop shuts down. Sometimes it even refuses to boot for several minutes after I shut it down (presumably until it cools down). I can confirm this began happening when I upgraded from intrepid to jaunty.

It can't be a dust problem because the laptop was cleaned less than a month ago. Also, it runs fine on Windows Vista at full load.

I should note that even setting the CPU scaling governor to powersave, the problem persists. This renders the machine pretty much unusable until the problem is fixed.

I'll be more than happy to provide you with more information if you request it.

I can confirm Jaunty's (incl. fresh install) overheating. My laptop is HP 6510b with Intel x3100 GPU, and Intel Core 2 Duo.

I suggest to mark this bug as CRITICAL because I can see my battery capacity lowered :( So this heat affects the hardware.

    I believe one aspect of this "bug" may be related to the fact that my laptop has a single fan for both processor and GPU. My GPU seems to stay on its "high performance" mode most of the time, possibly because I have VirtualBox running, but I'm not exactly sure why this is.(in any case this is the nvidia driver's responsibility) My GPU runs VERY hot (92C as I write) and the fan does not seem to react to GPU temperature/load. None of the 10+ temperatures reported to acpi seem to correspond to the GPU temp reported by the nvidia tools.

    There is a lot of reporting in this thread but not much investigation. I think we first have to determine what mechanism is actually controlling the fans on these machines. If the kernel is controlling the fans, then it is a kernel bug. If in fact the machine BIOS's are controlling them... then this is in fact a hardware bug.(windows drivers may gloss over the hardware bug by overriding the BIOS fan control) I believe that on my Lenovo T61p the fan speeds are controlled by the hardware and not the kernel. Again it is possible that under windows this control is overridden.

James

James - http://www.thinkwiki.org/wiki/How_to_control_fan_speed

I'm using a T61

I will repeat what I posted earlier, the Lenovo T61 fan is controlled by the thinkpad-acpi module. There is no question on this. It IS a kernel module. It is called thinkpad-acpi - in the following you will find a successful method to override control of the fan so you can choose the speed or disengage control completely (so it runs full speed). Also note that fan speed control and cpu controls are for power saving only. There is no issue with running your fan at full speed (some bios allow you to override this on ac power - HP does) - and there should be no issue running your cpu at its full speed either. You aren't "over clocking" anything. The only reason for controlling these things in this context is to save battery life. On AC I'm not worried about battery, so letting the fan run full speed is not an issue for me.

For more info, see the following: http://www.thinkwiki.org/wiki/How_to_control_fan_speed - this is specific to ThinkPads.

You can monitor your fan speed from the thinkpad-acpi module: cat /proc/acpi/ibm/fan (check path under /proc for your laptop)

You can adjust the fan speed to stay on at full speed by executing the following command as root: echo level disengaged > /proc/acpi/ibm/fan

(you will need to uninstall any other fan control applications or scripts you may have tried, including tpfand)

With the fan running between 4000-5000 rpm cooling is not a problem.

When fan control is engaged in the thinkpad-acpi module, the fan spins *only* between 0 and about 3000 rpm and takes quite a while to kick in - usually too late. With the control disengaged the fan runs at full speed. (but is not noisy)

For more info, see the following: http://www.thinkwiki.org/wiki/How_to_control_fan_speed

The issue is that the fan does not spin up soon enough or fast enough when the processor is running at full speed.

THAT is the bug.

For the link challenged - how to work around your fan issue on a Lenovo running a late version kernel

Linux Kernel 2.6.22 and Above (ibm-acpi has been replaced by thinkpad-acpi)

To enable fan control, the module parameter fan_control=1 must be given to thinkpad-acpi. (find the thinkpad-acpi configuration file under /etc/modprobe.d)

Having done so, reboot and you can use the following commands to control fan speed:
NOTE!
Instead of rebooting reloading module is enough (rmmod thinkpad_acpi && modprobe thinkpad_acpi as root).

# echo level 0 > /proc/acpi/ibm/fan (fan off)

# echo level 2 > /proc/acpi/ibm/fan (low speed)

# echo level 4 > /proc/acpi/ibm/fan (medium speed)

# echo level 7 > /proc/acpi/ibm/fan (maximum speed)

# echo level auto > /proc/acpi/ibm/fan (automatic - default)

# echo level disengaged > /proc/acpi/ibm/fan (disengaged)

The thinkpad-acpi module can <b>override</b> the fan speed, but when the mode is set to "auto" (the default) it is controlled by the hardware. This is documented in the thinkpad-acpi documentation. From the thinkpad-acpi documentation:
"Level "auto" means the EC changes the fan level according to some
internal algorithm, usually based on readings from the thermal sensors."

I have not checked the latest source but I assume this documentation is up to date.

I have been setting my fan to disengaged when I notice it is hot... but I don't consider this a fix.

James

James -

Auto is *not* hardware controlled. When set to Auto, thinkpad-acpi is controlling the fan speed directly. When you set it to disengage, this means thinkpad-acpi is *disengaged* - which is the fan will run at full speed. Thinkpad-acpi's "algorithm" does not spin the fan up soon enough or fast enough. "Auto" means acpi is controlling the fan. At least on my T61 running the latest kernel and Jaunty. Perhaps you are reading documentation related to an older version or different kernel.

I have not been able to duplicate the overheating issue in any other version of Ubuntu, nor in Windows XP SP3.

Auto is the default setting and the one that causes the issue.

With the acpi module disengaged, my fan runs between 4000 and 5000 rpm. The laptop now idles at 45C. I can drive the cpu temp to the mid 90's by running cpu stress, but the temp does *not exceed* 95C. The high temp is only achieved when artificially stressing the cpu using the stress application.

Charley -

I suggest you check the thinkpad-acpi source. I just did. There is NO automatic fan control in there. When you set "auto" it is simply setting the hardware to regulate the fans automatically. This is the default setting for thinkpad-acpi and the default setting when the machine turns on before any OS drivers come into play. The file is in git under drivers/platform/x86/thinkpad-acpi.c. The fan part isn't much code at all.

There is actually quite a bit of interesting stuff in the source. For one it says that the hardware controller will only set a maximum of fan level 6 and that 7 is only used for "thermal emergencies". Level 6 on my machine corresponds to ~3000rpm... which is basically where it sits all the time. This means the hardware controller is basically sitting at max cooling all the time. Is it possible that there is a windows driver that preemptively "disengages" the automatic control to kick up the fan during heavy GPU/CPU usage? Since the machine runs extremely hot and the automatic fan controller is at its max speed, I don't see any other way the machine would run cooler under similar loading.

Regardless of this though since nothing about the fan control has changed, if this is a truly new issue it must be related to changes in CPU or GPU power management.

James

James,

I apologize as it appears you are *correct* about the embedded controller and firmware - could it be that the hardware is expecting some type of control from the OS? Could it be that the firmware was not designed to regulate the temperature independent of the OS (assuming the OS is also governing CPU speed, GPU, etc?)

However - disengaging the fan control (thus allowing the fan to run full speed) does "solve" heat the issue for me. I have experienced no thermal shutdowns and the laptop runs *much* cooler.

The top fan speed on the T61 is about 5000 rpm. If you monitor the fan speeds while running stress, you'll see that regardless, the fan never goes over much more than about 3000 rpm and frequently does not run at all!

Is it unreasonable to think that one ought to be able to let the laptop run in performance mode, on AC power, with an application running the cpu at or near 100% and not experience temps in excess of 100C? This is why I don't accept the CPU governance issue as this should only be a function of keeping the laptop quiet and extending battery life (or perhaps pleasing the more 'green minded' among us). Other than modifying the hardware or over-clocking the cpu or gpu beyond spec, it shouldn't be possible to overheat... Perhaps that assumption is not valid?

For the mean time, I'm experiencing normal temps with fan control disengaged. I'm happy with this solution for the time being, but I agree it would be great to know the real issue and have a fix.

You are correct on the auto mode leaving it up to the firmware. Hats off.

Cheers,

Charley

  I did some testing under that other closed source operating system and I have decided that in my case at least I believe the thermal behavior is pretty much the same. I think my issue under ubuntu is that the nvidia driver does not seem to be throttling the GPU (despite it being set to, and me not being in a 3d app). Under windows the GPU does get throttled correctly but when I am running 3D applications for an extended period (games etc) the system runs just as hot as it does under Ubuntu. Windows does not kick up the fan speed... it sits at ~3000rpm and gets extremely hot (GPU at 99C, CPU at ~90). Frankly... it seems the T61p just does not have enough cooling to keep things under 100C.

   I'm going to fool with the nvidia driver settings to see if I can get the power management to work properly. The current behavior is that it boots at a lower speed (cooler), but once it kicks it up to the high performance mode it never goes back. I suspect probably an nvidia problem... not kernel or ubuntu.

  Note this is on the Lenovo T61p specifically... some of the other thermal issues discussed above on other machines could be (and probably are) for different reasons.

James

@James,
Presumably the issue has nothing to do with GPU for the reason that we all have the same problem even though having different hardware. OS - that's what is the same for us all. Since the issue was introduced with Jaunty and later on Hardy, according to Alex Cockell's report, it sticks me to the idea of ACPI built in kernel that causes the mess.

Well, it looks like ibm/lenovo folks have some hope :)

Anyone know of a util similar to "thinkpad-acpi" for compaq/gateway/acer ??

I'm really confused by this bug...

On the one hand where before I could leave my laptop lying on its own rubber feet and not have it turn off due to heat, I need to raise it a few inches to ensure that there is sufficient airflow and that it does not shut down due to heat.

On the other hand NOTHING has changed on my computer except the OS. I use the same scripts and programmes as I did before.

On another note, the fan SEEMS to be running as fast as it always did. That is I can hear it running hard when the CPU gets hot.

So, I can't see where the problem lies!

Could this be a lead? http://forums.lenovo.com/lnv/board/message?board.id=R_Series_Thinkpads&thread.id=3409
Possible heat issue with the graphics processor even with Vista...

I found the above on the Lenovo forums yesterday; maybe it's the GPU that's heating stuff up.. but the main thermal sensors listened to are on the CPUs? Is the GPU kinda central on the motherboard? Just that I do remember my laptop's underside getting warm especially when running on mains with the battery out ever since I bought the laptop (I used to cook batteries in an old Toshiba - so I often do that if I'm running mostly on mains at home).

Laptop fan kicks up to around 3000 periodically - but mostly being handled by the cooler at the mo.

Could it be that the GPU needs the fan to run longer and faster to cool it down?

Just that I did notice once when it WAS hot that the mouse would respond to movement but not to clicks. Could the GPU be thrashing? There were also these lines from /var/log/messages...

May 30 23:10:05 ubuntu kernel: [531782.411921] thinkpad_acpi: unhandled HKEY event 0x6022

But with no shutdown action as these events weren't being handed off in Hardy...

But - since I bought the cooler - no heat issues at all. However, I DID notice CPU0 go to 100% when the GPU wakes up (I hit shift to wake the display up after leaving the machine for a bit)... in 2.6.24-24 (kernel that gave me grief), I did also notice a bit of a memory leak.

In -23 (my uname - Linux ubuntu 2.6.24-23-generic #1 SMP Wed Apr 1 21:47:28 UTC 2009 i686 GNU/Linux), it isn't immediately hitting swap, and running at about 100Mb less memory used with Evolution, Firefox (6 tabs), Sysmon, gedit. Unlike under -24, where there was a definitely memory leak (committed memory climbing towards 50% rather than 30 - I run 2Gb onboard).

Could the bugs be in -24-restricted?

Well its happening on my HP, but I have a nVidia geforce 7400

  Those of you who have nvidia GPUs and are running the "nvidia" driver (the closed source one), you can check the powermizer status of your card by running the nvidia-settings tool. My machine starts up at level 1, but if I run a 3D application it switches to 2 and then never downclocks. If the GPU is at level 1, the machine stays reasonably cool (CPU temps ~50C after hours of use, GPU at 68). For those of you running compiz, your video card probably runs at the highest speed at all times.

James

   I have discovered that the nvidia driver locks the card in the highest performance mode when you have an external monitor enabled (which I usually do), which causes it to run very hot. When I am running just with the built in screen the card correctly declocks itself when doing normal "desktop" 2d work. Due to the fact that the GPU and CPU share a heatsink on this machine (T61p)... higher GPU temp most definitely affects CPU temps as well.

   I'm not sure if the external monitor/high performance mode is a bug or a feature... it seems unnecessary to me though.

James

I apologize if this isn't the forum to ask this question, but has this heat issue been reported for new Dell Inspiron's with Ubuntu preinstalled? I was planning on supporting Linux with my money by not paying the M$ tax, but if there is a heat issue, I'll have to wait to get my laptop. I'm not even sure which distro the inspirons come with.

So what is to do?

Have you tried playing Glest? or Wesnoth?

Don't speak about flightgear.

I own a HP dv6000, the unly way I can run those programs is using "cpulimit"

sudo apt-get install cpulimit

But framerate goes to hell, This has never happened before, not since Edgy...

This is happening to a huge percentage of laptops rendering them unusable for intensive tasks... counting of course flash as "intensive"

this is a MAJOR issue

Please backport fixes to Hardy as well when resolved.

So is anything happening with this bug at all? I would gladly submit any information required, if only I knew that someone handles this bug.
But so far seems like the way to fix that is just to switch to a different Linux distribution which is not Jaunty-based. That's sad. :(

Yeah, i have been using my own patch for my Dell Laptop. Basically i
have just been speeding up the fan to max. Its looped to fight to
keep the speed at max, it kind of breaths again. Its a dirty hack
for sure, but not as dirty as Canonical for releasing such as shitty
bug.

#!/usr/bin/python
import os
import time
#apt-get install i8kutils
#sudo modprobe i8k force=1
while(True):
 os.system("i8kfan speed 10000 > /dev/null")
 time.sleep(.5)

On Mon, Jun 15, 2009 at 3:56 AM, kblc<email address hidden> wrote:
> So is anything happening with this bug at all? I would gladly submit any information required, if only I knew that someone handles this bug.
> But so far seems like the way to fix that is just to switch to a different Linux distribution which is not Jaunty-based. That's sad. :(
>
> --
> Ubuntu 9.04 laptop overheat and shutdown
> https://bugs.launchpad.net/bugs/370173
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>
> Status in “linux” source package in Ubuntu: Confirmed
>
> Bug description:
> Ubuntu 9.04, fresh install on Acer Aspire 5005.
> Not sure about package, but it seems the problem is with CPU frequency adjustment or fan control.
> Laptop shuts down right in the middle of CPU-greedy operation overheated. I haven't seen it before (since 7.10).
> The issue is discussed at several forums (eg. http://ubuntuforums.org/showthread.php?t=1130450)
>
> ----
>     *-cpu
>          product: AMD Turion(tm) 64 Mobile Technology ML-37
>          vendor: Advanced Micro Devices [AMD]
>          physical id: 1
>          bus info: cpu@0
>          version: 15.4.2
>          size: 2GHz
>          capacity: 2GHz
>          width: 64 bits
>          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 syscall nx mmxext fxsr_opt x86-64 3dnowext 3dnow up pni lahf_lm cpufreq
>

Well, worse looking at this bug suggests it's a collection of different
problems, and instead of trying to triage, reproduce it, and working on
the bugs one-on-one, they are collecting it into one big bug, with no
hope of ever fixing it.

In my case, the buggy overheating leads to a situation where my laptop
is not capable of running Ubuntu, without major nursing by a Linux
old-hand. That basically destroys the value-proposition of Ubuntu. If I
need to tweak, hack, and so on, I do not need Ubuntu, there are
certainly distributions better suited to this.

Basically, it's an irritating situation for me, that I can survive, in
the hope that it will happen to get better, by random change, because
one does not get the impression that Ubuntu even cares. Would the laptop
be owned by some 3rd party that is not as well versed in Linux, this
would be a major disaster.

I wonder what would happen if I buy an installation support package from
Canonical, and then end up with this situation? Would that make anyone
care for this problem?

As far as I've sampled this bug, Jaunty seems to have often problems
with AMD based laptop, which makes one wonder how common this problem
is?

Andreas

Andreas you are wrong. The problem is very simple and my python patch
works GREAT on my system. If you speed up the fan, then the machine
works like a CHAMP. By default the fan is too fucking slow and the
machine overheats and dies. I'm going to stick with Ubuntu until I
get burned again. (No Pun intended...) What concerns me most is that
the problem is SIMPLE but there isn't an official patch which is just
pissing everyone off.

Peace

On Mon, Jun 15, 2009 at 5:21 AM, Andreas Kostyrka<email address hidden> wrote:
> Well, worse looking at this bug suggests it's a collection of different
> problems, and instead of trying to triage, reproduce it, and working on
> the bugs one-on-one, they are collecting it into one big bug, with no
> hope of ever fixing it.
>
> In my case, the buggy overheating leads to a situation where my laptop
> is not capable of running Ubuntu, without major nursing by a Linux
> old-hand. That basically destroys the value-proposition of Ubuntu. If I
> need to tweak, hack, and so on, I do not need Ubuntu, there are
> certainly distributions better suited to this.
>
> Basically, it's an irritating situation for me, that I can survive, in
> the hope that it will happen to get better, by random change, because
> one does not get the impression that Ubuntu even cares. Would the laptop
> be owned by some 3rd party that is not as well versed in Linux, this
> would be a major disaster.
>
> I wonder what would happen if I buy an installation support package from
> Canonical, and then end up with this situation? Would that make anyone
> care for this problem?
>
> As far as I've sampled this bug, Jaunty seems to have often problems
> with AMD based laptop, which makes one wonder how common this problem
> is?
>
> Andreas
>
> --
> Ubuntu 9.04 laptop overheat and shutdown
> https://bugs.launchpad.net/bugs/370173
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>
> Status in “linux” source package in Ubuntu: Confirmed
>
> Bug description:
> Ubuntu 9.04, fresh install on Acer Aspire 5005.
> Not sure about package, but it seems the problem is with CPU frequency adjustment or fan control.
> Laptop shuts down right in the middle of CPU-greedy operation overheated. I haven't seen it before (since 7.10).
> The issue is discussed at several forums (eg. http://ubuntuforums.org/showthread.php?t=1130450)
>
> ----
>     *-cpu
>          product: AMD Turion(tm) 64 Mobile Technology ML-37
>          vendor: Advanced Micro Devices [AMD]
>          physical id: 1
>          bus info: cpu@0
>          version: 15.4.2
>          size: 2GHz
>          capacity: 2GHz
>          width: 64 bits
>          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 syscall nx mmxext fxsr_opt x86-64 3dnowext 3dnow up pni lahf_lm cpufreq
>

Michael Brooks: Firstly, there's whole community involved even though it's Jaunty-only problem. Secondly, without Canonical you wouldn't have anything to complain about.

However, I suggest to mark it as CRITICAL again. Can anybody do it? I can't.

CRITICAL because
1) it damages hardware
2) it disables people to use computer properly (ana and her dv 6000)
3) because it shutdowns the computer without any warning so one's unsaved work can be done again
4) people switch to another distributions (I think about it as well)
5) people switch to another OS (back to trees?)

I'm sure this bug would be also much more visible in case of higher priority.

Yep, it's already 1,5 month this bug is here, and I assume that most of those who have encountered this bug didn't bother filling a bug report (or maybe they don't even know about it).
Still, most of the reports seem to be somehow connected to either NVidia GPU or chipset (correct me if I'm wrong), so maybe Canonical can't really do anything about this bug. But still that silent "ignore" mode for this bug is annoying.

Maybe this bug is assigned to the wrong team? I've checked stuff assigned to them - there are some bugs from 2006 which are still unresolved.

I've personally installed freshest Fedora on my laptop, and no similar heat issue so far. Even though it is somewhat flawed compared to Ubuntu, it doesn't destroy machine with constant overheat.

Hint: on "your" machine.

And it basically validates my points. You machine is not capable of
running Ubuntu without hacks, meaning your machine + Linux newbie =
Ubuntu does not work.

In my case the workaround is to use an older kernel. Not cool, but
workable. Although something in the system (and I have no idea what)
seems to limit which frequencies I'm capable to enable via cpufreq-set.
(might be as well the hardware that disable higher frequencies because
it is feeling "hot")

Andreas

Am Montag, den 15.06.2009, 12:40 +0000 schrieb Michael Brooks:
> Andreas you are wrong. The problem is very simple and my python patch
> works GREAT on my system. If you speed up the fan, then the machine
> works like a CHAMP. By default the fan is too fucking slow and the
> machine overheats and dies. I'm going to stick with Ubuntu until I
> get burned again. (No Pun intended...) What concerns me most is that
> the problem is SIMPLE but there isn't an official patch which is just
> pissing everyone off.
>
> Peace
>
> On Mon, Jun 15, 2009 at 5:21 AM, Andreas Kostyrka<email address hidden> wrote:
> > Well, worse looking at this bug suggests it's a collection of different
> > problems, and instead of trying to triage, reproduce it, and working on
> > the bugs one-on-one, they are collecting it into one big bug, with no
> > hope of ever fixing it.
> >
> > In my case, the buggy overheating leads to a situation where my laptop
> > is not capable of running Ubuntu, without major nursing by a Linux
> > old-hand. That basically destroys the value-proposition of Ubuntu. If I
> > need to tweak, hack, and so on, I do not need Ubuntu, there are
> > certainly distributions better suited to this.
> >
> > Basically, it's an irritating situation for me, that I can survive, in
> > the hope that it will happen to get better, by random change, because
> > one does not get the impression that Ubuntu even cares. Would the laptop
> > be owned by some 3rd party that is not as well versed in Linux, this
> > would be a major disaster.
> >
> > I wonder what would happen if I buy an installation support package from
> > Canonical, and then end up with this situation? Would that make anyone
> > care for this problem?
> >
> > As far as I've sampled this bug, Jaunty seems to have often problems
> > with AMD based laptop, which makes one wonder how common this problem
> > is?
> >
> > Andreas
> >
> > --
> > Ubuntu 9.04 laptop overheat and shutdown
> > https://bugs.launchpad.net/bugs/370173
> > You received this bug notification because you are a direct subscriber
> > of a duplicate bug.
> >
> > Status in “linux” source package in Ubuntu: Confirmed
> >
> > Bug description:
> > Ubuntu 9.04, fresh install on Acer Aspire 5005.
> > Not sure about package, but it seems the problem is with CPU frequency adjustment or fan control.
> > Laptop shuts down right in the middle of CPU-greedy operation overheated. I haven't seen it before (since 7.10).
> > The issue is discussed at several forums (eg. http://ubuntuforums.org/showthread.php?t=1130450)
> >
> > ----
> > *-cpu
> > product: AMD Turion(tm) 64 Mobile Technology ML-37
> > vendor: Advanced Micro Devices [AMD]
> >...

Hi people.

In fact, it's a CRITICAL bug and it will be nice if it can be solved quickly.

BUT I guess that forcing fan to it's maximum speed is not the right hack. You could damage your fan in a long way.

I'm just waiting for this fix moving back to Intrepid Ibex. But, you can also keep your cpu running in a lowest frequency to keep using jaunty.

kblc wrote:
> Yep, it's already 1,5 month this bug is here, and I assume that most of those who have encountered this bug didn't bother filling a bug report (or maybe they don't even know about it).
> Still, most of the reports seem to be somehow connected to either NVidia GPU or chipset (correct me if I'm wrong), so maybe Canonical can't really do anything about this bug. But still that silent "ignore" mode for this bug is annoying.
>

Seriously? I hadn't bothered to read the bugs I got merged into yet,
but if 370173 is in any way related to nVidia that merge shouldn't have
happened. The original bug that got me here, 213818, has been
reproduced on a bunch of Thinkpad t60 laptops - they've got not a bit of
nVidia inside I'm pretty sure. My graphics card is ATI anyway.

I can only frown.

--
Dominic van Berkel
"Hey, stoic one. We're alive. Dance with me."

My GPU is Intel X3100 in HP 6510b, so I think it's not GPU related.

I think Andreas hit the nail here... this "bug" thread isn't so much a single bug it is a single -symptom- of multiple bugs which may or may not be related.

Without some more detailed data on all the different machine configurations I don't see this bug moving forward. Fan control hack scripts on your own machine aren't going to help fix it.

Folks should report:
  Machine, chipset, graphics chip, graphics driver being used
  Temperature data under both jaunty and another OS (windows or previous ubuntu if the issue is changed). Idle temps, idle fan speeds, cpu loaded temps, cpu loaded fan speeds, heavy graphics temperature data and fan speeds.

With more data we might start to see some trends and be able to boil this down to where the bugs really are.

James

Well, that's what I said. This is a collective bug, where anything with
the keyword "overheat" ends in.

Collecting different bugs, without any directed effort to even verify
what the bug is, won't solve the problems.

In my original bugreport, I did include most of this, although I did not
mention that I've got NVidia graphics.

In my case the problem started with Jaunty, is bad with
2.6.27-14-generic, and got intolerable in 2.6.28 (as in burnK7 manages
to make the machine reboot in less than 60 seconds).

I've documented that with script that logged all kind of potentially
relevant stuff, but nobody even got interested in that.

My current work around is to stick with 2.6.27, although it's impossible
to get more than 1.6GHz, and very hard to persuade the box to speed up
from 800MHz. It's highly unfunny, the Atom netbook of my wife feels
snappier.

I'm also wondering if there is a way to make the fan to spin up
manually, but my /proc/acpi/fan directory is empty, loading fan.ko does
not change that, nor does it produce any dmesg output.

Andreas

Am Montag, den 15.06.2009, 13:17 +0000 schrieb arsenix:
> I think Andreas hit the nail here... this "bug" thread isn't so much a
> single bug it is a single -symptom- of multiple bugs which may or may
> not be related.
>
> Without some more detailed data on all the different machine
> configurations I don't see this bug moving forward. Fan control hack
> scripts on your own machine aren't going to help fix it.
>
> Folks should report:
> Machine, chipset, graphics chip, graphics driver being used
> Temperature data under both jaunty and another OS (windows or previous ubuntu if the issue is changed). Idle temps, idle fan speeds, cpu loaded temps, cpu loaded fan speeds, heavy graphics temperature data and fan speeds.
>
> With more data we might start to see some trends and be able to boil
> this down to where the bugs really are.
>
>
> James
>

HP 6510b, Intel Core 2 Duo, Intel 965GM, Intel GMA X3100 using xserver-xorg-video-intel (I hope so). uname and sensors (just browser running follows)

uname -a
Linux barny 2.6.28-13-generic #44-Ubuntu SMP Tue Jun 2 07:55:09 UTC 2009 x86_64 GNU/Linux

sensors
acpitz-virtual-0
Adapter: Virtual device
temp1: +50.0°C (crit = +256.0°C)
temp2: +48.0°C (crit = +108.0°C)
temp3: +47.0°C (crit = +105.0°C)
temp4: +34.5°C (crit = +108.0°C)
temp5: +50.0°C (crit = +110.0°C)

coretemp-isa-0000
Adapter: ISA adapter
Core 0: +49.0°C (high = +100.0°C, crit = +100.0°C)

coretemp-isa-0001
Adapter: ISA adapter
Core 1: +49.0°C (high = +100.0°C, crit = +100.0°C)

This is copypasting from my original post back when this first happened
to me, in #213818. I haven't used Linux on this machine for some time,
but I recall testing Jaunty and finding that it didn't fix anything.

[paste]
Similar issue here, not quite sure whether it's the same.
Running 8.04 on a Thinkpad t60p (Intel Core2Duo proc), my fan speed for
some reason will not go above 3500RPM while set to "auto". I haven't had
a shutdown yet, but I've been unable to boot once. Temperature mostly
hangs on 62C, which is simply too darn high for an idling laptop on AC.
Funny bit is that I'm for some reason unable to manually set fan speed,
even though I turned the necessary module option (thinkpad_acpi
fan_control=1) - I get a Permission Denied, whether I try as root or as
a plain user.
[paste]

The t60p uses an ATI FireGL V5250, which I've used with both the free
and the non-free drivers.

I really do hope this gets cleared up. And, probably, split up.

Dominic

Michal I think 50C is pretty much normal operating temp for a Core2 Duo laptop. I think if you run under windows you will find it is similar. Thermal shutdown temp is well over 100C.

I think we should focus on users who are getting thermal shutdowns since that is pretty much guaranteed to be a critical bug. Temperatures which are just "higher than I think they should be" are very difficult to judge quantitatively.

Andreas can you post some fan speed information? Before thermal shutdown of the machine what are the fans doing?

James

Am Montag, den 15.06.2009, 13:57 +0000 schrieb arsenix:
> Michal I think 50C is pretty much normal operating temp for a Core2 Duo
> laptop. I think if you run under windows you will find it is similar.
> Thermal shutdown temp is well over 100C.
>
> I think we should focus on users who are getting thermal shutdowns since
> that is pretty much guaranteed to be a critical bug. Temperatures which
> are just "higher than I think they should be" are very difficult to
> judge quantitatively.
>
> Andreas can you post some fan speed information? Before thermal
> shutdown of the machine what are the fans doing?

How can I query the fan speed? /proc/acpi/fan is empty.

Subjectivly, it's running fast, but no idea how fast in exact numbers.

I've got powertop, temperature logs, and cpufreq logs.

Interestingly, looking at the data it seems that 2.6.28 even crashed
with the CPUs running at 800MHz.

Andreas

After I have put my laptop on some stands and given more space under it, I am not having any thermal shutdowns. However, the laptop does seem to be hotter than in Intrepid.

I agree this is subjective, but in Intrepid I would not need to keep my laptop raised, and it would never have a thermal shutdown. I think this is key, because I have not changed my habits, and it is ONLY because of the upgrade that I am having a thermal shutdown.

Specs: HP dv5242ea | Core Duo 1.60GHz | 2GB RAM | 320GB SATA | DVD-RAM Matshita UJ 840S | nVidia GeForce Go 7400 256mb | Intel PRO 3945ABG

I guess fan speed monitoring is not supported by the kernel yet on your
machine. I assume your machine has an AMD chipset? (I think all AMD
notebooks do)

Have you done any testing on windows or with earlier ubuntu distro's to
verify that the behavior is different under Jaunty?

James

Andreas Kostyrka wrote:
> Am Montag, den 15.06.2009, 13:57 +0000 schrieb arsenix:
>
>> Michal I think 50C is pretty much normal operating temp for a Core2 Duo
>> laptop. I think if you run under windows you will find it is similar.
>> Thermal shutdown temp is well over 100C.
>>
>> I think we should focus on users who are getting thermal shutdowns since
>> that is pretty much guaranteed to be a critical bug. Temperatures which
>> are just "higher than I think they should be" are very difficult to
>> judge quantitatively.
>>
>> Andreas can you post some fan speed information? Before thermal
>> shutdown of the machine what are the fans doing?
>>
>
> How can I query the fan speed? /proc/acpi/fan is empty.
>
> Subjectivly, it's running fast, but no idea how fast in exact numbers.
>
> I've got powertop, temperature logs, and cpufreq logs.
>
> Interestingly, looking at the data it seems that 2.6.28 even crashed
> with the CPUs running at 800MHz.
>
> Andreas
>
>

Am Montag, den 15.06.2009, 15:06 +0000 schrieb arsenix:
> I guess fan speed monitoring is not supported by the kernel yet on your
> machine. I assume your machine has an AMD chipset? (I think all AMD
> notebooks do)

Well, I guess it's an nVidia chipset:

00:00.0 RAM memory: nVidia Corporation C51 Host Bridge (rev a2)
00:00.1 RAM memory: nVidia Corporation C51 Memory Controller 0 (rev a2)
00:00.2 RAM memory: nVidia Corporation C51 Memory Controller 1 (rev a2)
00:00.3 RAM memory: nVidia Corporation C51 Memory Controller 5 (rev a2)
00:00.4 RAM memory: nVidia Corporation C51 Memory Controller 4 (rev a2)
00:00.5 RAM memory: nVidia Corporation C51 Host Bridge (rev a2)
00:00.6 RAM memory: nVidia Corporation C51 Memory Controller 3 (rev a2)
00:00.7 RAM memory: nVidia Corporation C51 Memory Controller 2 (rev a2)
00:02.0 PCI bridge: nVidia Corporation C51 PCI Express Bridge (rev a1)
00:03.0 PCI bridge: nVidia Corporation C51 PCI Express Bridge (rev a1)
00:05.0 VGA compatible controller: nVidia Corporation C51 [Geforce 6150 Go] (rev a2)
00:09.0 RAM memory: nVidia Corporation MCP51 Host Bridge (rev a2)
00:0a.0 ISA bridge: nVidia Corporation MCP51 LPC Bridge (rev a3)
00:0a.1 SMBus: nVidia Corporation MCP51 SMBus (rev a3)
00:0a.3 Co-processor: nVidia Corporation MCP51 PMU (rev a3)
00:0b.0 USB Controller: nVidia Corporation MCP51 USB Controller (rev a3)
00:0b.1 USB Controller: nVidia Corporation MCP51 USB Controller (rev a3)
00:0d.0 IDE interface: nVidia Corporation MCP51 IDE (rev f1)
00:0e.0 IDE interface: nVidia Corporation MCP51 Serial ATA Controller (rev f1)
00:10.0 PCI bridge: nVidia Corporation MCP51 PCI Bridge (rev a2)
00:10.1 Audio device: nVidia Corporation MCP51 High Definition Audio (rev a2)
00:14.0 Bridge: nVidia Corporation MCP51 Ethernet Controller (rev a3)
00:18.0 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] HyperTransport Technology Configuration
00:18.1 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] Address Map
00:18.2 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] DRAM Controller
00:18.3 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] Miscellaneous Control
03:00.0 Network controller: Broadcom Corporation BCM4328 802.11a/b/g/n (rev 03)

>
> Have you done any testing on windows or with earlier ubuntu distro's to
> verify that the behavior is different under Jaunty?

Windows no, this laptop has run exactly 6 hours Vista while creating the
recovery DVDs, which I have no idea even where they are.

Hardy, Ibex have had no problems with this hardware, at least not
temperature/kernel related.

On Jaunty it's strongly varies on the kernel release. 2.6.27-1{1,4} work
somewhat, 2.6.28 does not even survive 60 seconds of burnK7.

See https://bugs.launchpad.net/ubuntu/+source/linux/+bug/378065 for my
logs and experiences.

Andreas

>
>
> James
>
>
> Andreas Kostyrka wrote:
> > Am Montag, den 15.06.2009, 13:57 +0000 schrieb arsenix:
> >
> >> Michal I think 50C is pretty much normal operating temp for a Core2 Duo
> >> laptop. I think if you run under windows you will find it is similar.
> >> Thermal shutdown temp is well over 100C.
> >>
> >> I think we should focus on use...

Lenovo Thinkpad R61i - I was running 2.6.24-24.53 when suffering thermal emergencies - but with no shutdowns although the laptop was asking for them.
Intel X3100 embedded graphics
Intel 3945ABG wireless, Broadcom ethernet controller
Core 2 Duo, 2x 1.8GHz.

I daren't run this not on the Axasa laptop cooler.

RUnning 8.04.2 (or most of it), but running 2.6.24-23generic..

Canonical - you have a Severity 2 incident here - just fix it already!

Andres I have a desktop with the MPC51 chipset. Running the "sensors-detect" script that is part of the lmsensors package correctly detected the proper modules to measure the onboard sensors. This should allow fan speed, temperatures etc to be monitored through the lm-sensors package. Try it and report back.

James

This is a Jaunty bug... if you aren't running Jaunty please post on another bug.

Unless this isn't already clear to folks... if you don't have (or aren't willing to obtain) comparative thermal testing data from another OS where the problem does not occur... your contribution is not that useful. Many laptops have thermal design problems, or were just designed to run hot to keep noise and size to a minimum. If you aren't sure that the issue you are seeing doesn't happen under all OS's... then it is very hard to say if the problem you are seeing is truly isolated to ubuntu jaunty. If you want this bug fixed... we need more data.

More people need to post information about what their fan's do under Ubuntu vs under Os's where the machines don't run hot. This will determine whether the issue is related to fan control or some type of CPU/GPU heat output increase.

James

Am Montag, den 15.06.2009, 19:26 +0000 schrieb arsenix:
> This is a Jaunty bug... if you aren't running Jaunty please post on
> another bug.
>
> Unless this isn't already clear to folks... if you don't have (or aren't
> willing to obtain) comparative thermal testing data from another OS
> where the problem does not occur... your contribution is not that
> useful. Many laptops have thermal design problems, or were just

Well, as this is my main working machine, just installing a different OS
is not a really an option.

OTOH, this machine has run on since last August without problems till I
upgrade to Jaunty, and the real trouble started when I upgraded to
2.6.28.

I think comparative temperature logs for different Linux kernels should
qualify.

> designed to run hot to keep noise and size to a minimum. If you aren't
> sure that the issue you are seeing doesn't happen under all OS's... then
> it is very hard to say if the problem you are seeing is truly isolated
> to ubuntu jaunty. If you want this bug fixed... we need more data.
>
> More people need to post information about what their fan's do under
> Ubuntu vs under Os's where the machines don't run hot. This will
> determine whether the issue is related to fan control or some type of
> CPU/GPU heat output increase.
>
>
> James
>

> I think comparative temperature logs for different Linux kernels should
> qualify.

Indeed your experience definitely qualifies. If we can figure out how to get some fan speed data as well we might be able to start digging down into your issue.

James

Am Montag, den 15.06.2009, 19:52 +0000 schrieb arsenix:
> > I think comparative temperature logs for different Linux kernels should
> > qualify.
>
> Indeed your experience definitely qualifies. If we can figure out how
> to get some fan speed data as well we might be able to start digging
> down into your issue.
>
> James
>

Ok, I've run sensors-detect, and while it detected some additional
temperature sensors, it seems to have found no fan data:

root@andi-lap:~# sensors
acpitz-virtual-0
Adapter: Virtual device
temp1: +85.0°C (crit = +95.0°C)

k8temp-pci-00c3
Adapter: PCI adapter
Core0 Temp: +82.0°C
Core0 Temp: +81.0°C
Core1 Temp: +81.0°C
Core1 Temp: +82.0°C

root@andi-lap:~# acpi -t
     Battery 0: Full, 100%, rate information unavailable
     Thermal 0: ok, 84.0 degrees C

Andreas

This is not a hardware specific issue. Some people here have ATI
graphics or AMD chips. I have a nVidia chip with a dual core 2.4ghz
Intel. The heat control on Jaunty was changed, the fans stopped
the oscillated "breathing" which to be honest was annoying. All of
the evidence i have seen is that if you have a powerful system then it
will overheat more often. This is a very easy to problem to
reproduce, just run any heavy process of your choice the system will
die. ~82.0°C I hit 104 and shutdown just about every day until i
wrote my python hack. I haven't had my system over heat sense then,
so this is PROOF that the fan speed is too slow. This is a very
easy problem to fix, just speed up the fan.

So maybe like 4 people on the planet including me have a fix for this,
 too bad everyone else is getting FUCKED while you guys chase your
tails. This problem is so obvious it HURTS.

Peace out.

On Mon, Jun 15, 2009 at 1:11 PM, Andreas Kostyrka<email address hidden> wrote:
> Am Montag, den 15.06.2009, 19:52 +0000 schrieb arsenix:
>> > I think comparative temperature logs for different Linux kernels should
>> > qualify.
>>
>> Indeed your experience definitely qualifies.  If we can figure out how
>> to get some fan speed data as well we might be able to start digging
>> down into your issue.
>>
>> James
>>
>
> Ok, I've run sensors-detect, and while it detected some additional
> temperature sensors, it seems to have found no fan data:
>
> root@andi-lap:~# sensors
> acpitz-virtual-0
> Adapter: Virtual device
> temp1:       +85.0°C  (crit = +95.0°C)
>
> k8temp-pci-00c3
> Adapter: PCI adapter
> Core0 Temp:  +82.0°C
> Core0 Temp:  +81.0°C
> Core1 Temp:  +81.0°C
> Core1 Temp:  +82.0°C
>
> root@andi-lap:~# acpi -t
>     Battery 0: Full, 100%, rate information unavailable
>     Thermal 0: ok, 84.0 degrees C
>
> Andreas
>
> --
> Ubuntu 9.04 laptop overheat and shutdown
> https://bugs.launchpad.net/bugs/370173
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>
> Status in “linux” source package in Ubuntu: Confirmed
>
> Bug description:
> Ubuntu 9.04, fresh install on Acer Aspire 5005.
> Not sure about package, but it seems the problem is with CPU frequency adjustment or fan control.
> Laptop shuts down right in the middle of CPU-greedy operation overheated. I haven't seen it before (since 7.10).
> The issue is discussed at several forums (eg. http://ubuntuforums.org/showthread.php?t=1130450)
>
> ----
>     *-cpu
>          product: AMD Turion(tm) 64 Mobile Technology ML-37
>          vendor: Advanced Micro Devices [AMD]
>          physical id: 1
>          bus info: cpu@0
>          version: 15.4.2
>          size: 2GHz
>          capacity: 2GHz
>          width: 64 bits
>          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 syscall nx mmxext fxsr_opt x86-64 3dnowext 3dnow up pni lahf_lm cpufreq
>

Am Montag, den 15.06.2009, 22:32 +0000 schrieb Michael Brooks:
> This is a very
> easy problem to fix, just speed up the fan.
>

1.) Certainly this fix is more workable if you can "address" the fan in
the system. My laptop does not report a fan, nor can I control it, at
least from userspace.

2.) Older kernels do not require any hacks, fixes, whatever. What
changed?

Andreas

I'll contain my wonder and summarize by saying let's move on.

Andreas can you try a different distro with a similar vintage kernel? If you try your burn in with knoppix 6.0.1 it also uses 2.6.28 kernel. I'm assuming it is a kernel issue, but this will also see if it is an ubuntu kernel specific issue. If it does still overheat we can move back in kernel version until the problem disappears.

BTW... I took my laptop apart and reapplied thermal grease to the CPU and GPU. My temps are now way down from where they were before. Evidently when lenovo replaced my motherboard they failed to reapply the grease well. Just a reminder that hardware issues are not out of the question for everyone chasing thermal issues...

James

For the record, I am your putative Regular User who bought his laptop preinstalled with Ubuntu 8.04. I have no other OS, and this is my primary machine (Thinkpad R61i, 7650-series model).

All I know is that my machine overheated when running 2.6.24-24.53, and my vendor told me to go back to booting -23. And in order to keep the CPU temps around the mid-40s (which I believe is OK for Core 2 Duo THinkpads according to these long threads), I have to run it on a cooler.

I gave it a blast with compressed air - but anything else would mean me taking the laptop to an IBM Service Centre.

Currently, it's running OK atop the cooler- but I am sure it's the cooler keeping the temperature down.

Something happened with the system updates around my first post on this incident report.

Please, could Canonical fix the issue, and get a fix into my Update Manager once it's fully tested.

Alex:

The Ubuntu 8.04 you are running is Hardy, a different and older release than this bug is discussing (9.04, Jaunty).

My thinkpad idles at nearly 50C... in windows... in ubuntu etc. Thermal shutdown on Core2 Duo's is over 100C. It does not sound like you are experiencing the issue discussed here and frankly it doesn't sound to me like you have a thermal issue at all.

James

All I know is that back last week, my machine was desperately hot to the touch, and froze up. I hard-powered off, called my vendor, left it most of a day to cool down... but am now running it on top of a cooler.

SOMETHING happened after system updates (accepting 2.6.24-24.53 onto my machine) that I took on about two weeks back, which caused the laptop to be extremely hot to the touch, whacked the thermal warnings into my logs, AND DID NOT SHUT DOWN. In order to run my machine, I HAVE to run it on the cooler, and the fan does not go about about 3100rpm.

It was unbearably hot, and it is only with the cooler running, blowing cold air at the base on the unit, that the CPU stays within functional temps.

Since then - I run it on top of the cooler.. but if I ever did boot into -24.53 again, it would probably overheat.

As stuff is often rolled in.. PLEASE JUST FIX IT.

I am *not* a developer.

Nor do I want to see my laptop break if I had to run it away from the cooler.

Apologies if I come across as strident. I am an end-user - I bought the laptop preinstalled. Therefore, to me, the OS is part of the product.

For the record, with the cooler running, my temps are down at
alex@ubuntu:~$ acpi -t
     Thermal 1: ok, 42.0 degrees C
     Thermal 2: ok, 40.0 degrees C

whereas I recall them being at the 52 deg C point... and the GPU running very hard.

On mains, the BIOS is set to run at Performance level, GNOME has Ondemand/ondemand.

Just that I'm worried that eventually when I accept an upgrade to the next LTS, it would cook my machine.

And this is 650 quid of laptop...

uname is Linux ubuntu 2.6.24-23-generic #1 SMP Wed Apr 1 21:47:28 UTC 2009 i686 GNU/Linux. Last kernel accepted via Update Manager was -24.53 which had the issue.

Hm, might this issue then somehow be connected to that particular kernel update?
Fedora 11 has even newer kernel (2.6.30.something, I think), and no heat issue. Might the kernel update resolve this issue then?

Somebody stated that my ~50°C in powersave state is normal. It isn't :) I can subjectively feel that is is much more hotter than in Intrepid. I'll do some tests with Interpid or Fedora which I'm going to switch this week. I'll post it here.

Michael Brooks thinks that fans aren't working enough. It seems racial, however it is not as obvious because some people also stated that their battery life (I don't mean permanent capacity now) has decreased. And if the fans work more, it decreases it even more. But maybe the heat is decreasing the battery life as well.

Regarding the permanent capacity, I had 52 Wh, now I have 42 Wh :( The heat really decreases the permanent (factory) capacity. It damages our hardware unfortunately :(

kblc: If you're brave, you can try to install 2.6.30 kernel using this manual:
http://ubuntuforums.org/showthread.php?t=1130582

Actually I don't think it'll help. But don't forget to run "sensors" before this upgrade in order to have some material to compare ;)

Michal, I would definitely do so, but I've already installed Fedora on that laptop, and since it's really my girlfriends laptop, I don't think she'll be happy that I'm playing around it for the whole weekend again. :)

Michal here is a review for your notebook:
  http://www.notebookreview.com/default.asp?newsID=3763

"I installed Notebook Hardware Control and according to its measures the CPU temperature remained at about 50C when idling."

If you disagree, lets see some data.

James

arsenix: thanks for the link...then I think the heat isn't measured properly :( I can really feel the hot on my palms and I really do sweat. However, don't know any other way how to measure it :(

I know for a fact the fans are too slow, I have proved it. If you
speed up the fan, its stops over heating, plain and simple.

There might be other issues at play eating up processor time and thus
battery life. I wish a had a mode just for battery life....

Peace

On Tue, Jun 16, 2009 at 6:12 AM, Michal Pěnka<email address hidden> wrote:
> arsenix: thanks for the link...then I think the heat isn't measured
> properly :( I can really feel the hot on my palms and I really do sweat.
> However, don't know any other way how to measure it :(
>
> --
> Ubuntu 9.04 laptop overheat and shutdown
> https://bugs.launchpad.net/bugs/370173
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>
> Status in "linux" source package in Ubuntu: Confirmed
>
> Bug description:
> Ubuntu 9.04, fresh install on Acer Aspire 5005.
> Not sure about package, but it seems the problem is with CPU frequency adjustment or fan control.
> Laptop shuts down right in the middle of CPU-greedy operation overheated. I haven't seen it before (since 7.10).
> The issue is discussed at several forums (eg. http://ubuntuforums.org/showthread.php?t=1130450)
>
> ----
> *-cpu
> product: AMD Turion(tm) 64 Mobile Technology ML-37
> vendor: Advanced Micro Devices [AMD]
> physical id: 1
> bus info: cpu@0
> version: 15.4.2
> size: 2GHz
> capacity: 2GHz
> width: 64 bits
> 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 syscall nx mmxext fxsr_opt x86-64 3dnowext 3dnow up pni lahf_lm cpufreq
>

Michael your dell (unlike every other machine being discussed here) uses software fan control, using the kernel i8k driver. There are several usermode tools to control the fans automatically and keep the temperatures lower on these machines both under linux and windows. Lots of folks have issues on these machines and run the usermode tools to keep them cooler. One of them is called dellfand. Search and you will find issues with the i8k driver have been discussed at length for years (well before ubuntu came onto the scene). In nearly all cases these issues stem from hardware problems (some of them inherent design problems) even if you can work around them with a bandaid fix with software tools.

James

Now I'm really angry. Yesterday my computer went shutdown because of overheating. I was running some virtual machine and after cooling down and starting it again, it didn't want to start againg, because filesystem in it is corrupted. I've lost many hours of work which can be fortunately done again.

I'm wondering why isn't this bug marked CRITICAL yet and why assigned team is Ubuntu Laptop Team. It seems they aren't active there so I suggest to change the team and marking it CRITICAL already!

Here you are my log when computer runs at 2 GHz/both cores:
------------------------------------------------------------------------
acpitz-virtual-0
Adapter: Virtual device
temp1: +65.0°C (crit = +256.0°C)
temp2: +64.0°C (crit = +108.0°C)
temp3: +51.0°C (crit = +105.0°C)
temp4: +36.1°C (crit = +108.0°C)
temp5: +80.0°C (crit = +110.0°C)

coretemp-isa-0000
Adapter: ISA adapter
Core 0: +66.0°C (high = +100.0°C, crit = +100.0°C)

coretemp-isa-0001
Adapter: ISA adapter
Core 1: +65.0°C (high = +100.0°C, crit = +100.0°C)
------------------------------------------------------------------------

Please compare it to my powersave states (above), pasted again:
------------------------------------------------------------------------
acpitz-virtual-0
Adapter: Virtual device
temp1: +50.0°C (crit = +256.0°C)
temp2: +48.0°C (crit = +108.0°C)
temp3: +47.0°C (crit = +105.0°C)
temp4: +34.5°C (crit = +108.0°C)
temp5: +50.0°C (crit = +110.0°C)

coretemp-isa-0000
Adapter: ISA adapter
Core 0: +49.0°C (high = +100.0°C, crit = +100.0°C)

coretemp-isa-0001
Adapter: ISA adapter
Core 1: +49.0°C (high = +100.0°C, crit = +100.0°C)
------------------------------------------------------------------------

I think this team or kernel team should be better

Please run the following so we have the appropriate debugging info for you system.

apport-collect -p linux-image-`uname -r` 370173

Thanks

--
Pete Graner
<email address hidden>
Manager
Ubuntu Kernel Team
Canonical Ltd.

Architecture: amd64
DistroRelease: Ubuntu 9.04
HibernationDevice: RESUME=UUID=d4e29aec-2b62-42ab-a6ad-ed56e222d32b
MachineType: Hewlett-Packard HP Compaq 6510b (RJ559AV)
Package: linux-image-2.6.28-13-generic 2.6.28-13.44
PackageArchitecture: amd64
ProcCmdLine: root=UUID=4f344b01-bc9b-48c8-b599-e557175bd864 ro quiet splash
ProcEnviron:
 SHELL=/bin/bash
 LANG=cs_CZ.UTF-8
ProcVersionSignature: Ubuntu 2.6.28-13.44-generic
Uname: Linux 2.6.28-13-generic x86_64
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare

Michal are the temperatures you posted above for the full speed running as hot as the machine got? Those temperatures are not that hot. If your machine crashed from seeing a peak temperature of 80C... you have a hardware problem. Thermal shutdown temperatures are more like 110C.

James

No that is a mistype. Of course the machine shuts down at 104c! So it
sounds like you all are still chasing your tails or denying the bug. I
don't give a fuck about this tread anymore, I have my patch that's all I
need. If you are still waiting for canonical to even acknowledge this
issue then your got bigger problems on your hands.

peace

On Thu, Jun 18, 2009 at 4:01 AM, arsenix <email address hidden> wrote:

> Michal are the temperatures you posted above for the full speed running
> as hot as the machine got? Those temperatures are not that hot. If
> your machine crashed from seeing a peak temperature of 80C... you have a
> hardware problem. Thermal shutdown temperatures are more like 110C.
>
> James
>
> --
> Ubuntu 9.04 laptop overheat and shutdown
> https://bugs.launchpad.net/bugs/370173
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>
> Status in “linux” source package in Ubuntu: Confirmed
>
> Bug description:
> Ubuntu 9.04, fresh install on Acer Aspire 5005.
> Not sure about package, but it seems the problem is with CPU frequency
> adjustment or fan control.
> Laptop shuts down right in the middle of CPU-greedy operation overheated. I
> haven't seen it before (since 7.10).
> The issue is discussed at several forums (eg.
> http://ubuntuforums.org/showthread.php?t=1130450)
>
> ----
> *-cpu
> product: AMD Turion(tm) 64 Mobile Technology ML-37
> vendor: Advanced Micro Devices [AMD]
> physical id: 1
> bus info: cpu@0
> version: 15.4.2
> size: 2GHz
> capacity: 2GHz
> width: 64 bits
> 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 syscall nx
> mmxext fxsr_opt x86-64 3dnowext 3dnow up pni lahf_lm cpufreq
>

I don't know with which temperature my laptop went down, but I don't think 80°C is normal. It is really hot. And I can prove it damages the hardware - my permanent (factory) battery capacity went down! I think not only capacity suffers :(

Now as I'm writing you, the computer runs for about an hour and is in 800 MHz powersave state. The CPU temp is (shows) about 61°C...

Michael: your patch doesn't apply my computer and I'm not as skilled programmer, so I'm switching to Fedora this weekend. I've got install medium already.

Architecture: amd64
DistroRelease: Ubuntu 9.04
NonfreeKernelModules: nvidia
ProcEnviron:
 SHELL=/bin/bash
 LANG=en_US.UTF-8
Uname: Linux 2.6.28-11-generic x86_64
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare

Michal make sure you report back how well fedora works out. Thanks!

James

I've set /proc/acpi/ibm/fan to "disengaged" - which causes the fan to run at full speed without any hardware or software control. My idle temps are way down, but I am still periodically experiencing the thermal shutdowns - usually while processes are running with the laptop unattended. The fan speed helps, but has not solved the issue for me.

I've submitted apport-collect data above.

Note that on my Lenovo T61, 2000-3000 rpm is actually normal by design - which contradicts my earlier assertion that the fan was running too slowly. Disengaged, the fan is currently running 4000-5000 rpm (which as I said helps a great deal but hasn't solved the issue entirely for me)

I can replicate the issue using the stress application (see earlier posts)

I tried to replicate the issue on another laptop running 8.04 Hardy (32-bit in that case) and was not able to do so.

Charley you have a hardware problem. Talk to lenovo and send it for repairs. Ubuntu is not involved in the fan control, your fans are running fine... and regardless of what ubuntu does with your machine's power management it should be able to run without shutting down.

James

Yes James, thank you. I understand Ubuntu is not involved in the fan control. Got it.

The thing is that I am unable to replicate the issue in any other version of Ubuntu, nor am I able to replicate the issue in Windows XP SP3 - running a variety of stress tests encompassing video, i/o, memory, and cpu.

So - back to your earlier questions about GPU perhaps... or earlier comments about CPU governor... I don't know.

What I can assert is that the thermal issues are unique to this version of Ubuntu and kernel -

What I have ruled out is that the issue is related to the fan, which seemed like the obvious issue at first, and was also the focus (most likely the incorrect focus) of attention in the other threads around this issue...

Note that I'm not bashing Canonical, or Ubuntu or anyone else here, unlike some of our peers. I do have the option of booting Windows XP if I need to, or running some earlier distro. I'm quite happy to try and contribute whatever I can here. I've been a network/support manager/engineer for twelve years, primarily in Windows shops. Most recently interested in FOSS solutions and Ubuntu as a user-FRIENDLY distribution I can eagerly and confidently promote to friends, clients, and family.

Yes, I understand it isn't the fan control.

But I also understand there is a thermal shutdown issue (perhaps secondary to some as yet unidentified issue) that is UNIQUE to this version.

Cheers.

Charley -

I'm sticking by my rec that you need to get the hardware fixed. Maybe windows throttles the machine more agressively when it gets hot instead of shutting down... maybe windows has better GPU thermal control (this is almost certainly true)... in any case your machine should not be shutting down with the fans at max under any loading situation. Lots of other users have T60's and no thermal shutdown issues in Jaunty. My machine (T61p) has the same fan control and power management system as yours... and has no issues. The thermal issues that originally led me to this thread were hardware issue related.

I'm not saying there isn't some power management bug that makes your machine runs hotter under only Jaunty... I'm just saying you have a hardware problem if your machine gets itself hot enough to shutdown.

James

Well, while the fans are certainly the core mechanism of keeping a
laptop cool nowadays, the OS/firmware/hardware have other mechanisms
too.

So while I concur that a laptop that shutdowns itself with the fans
running suggests a hardware problem, it does not have to be so.

The central question is, why does it run on Windows? Hard to answer
because Windows is SO different, but there are more than enough reports
where other Linux distributions or even different kernels (as in my
case) have kept the hardware going.

In my experience btw, there are hardware setups that have rather badly
configured trippoints, and another thing is that although my laptop
claims to have a critical trippoint at 100°C I have never seen the
temperature approach it, it always reboots below 90°C => that behavior
I've noticed with different systems, so I guess it's meant to be so.

Other things to consider, because modern systems seem to be rather
choosy sometimes:

-) are you running it on a hard desk? I know laptops that overheat even
from a simple tablecloth beneath them.
-) are the fans clear? Depending upon the environment fans can get
really really fast clogged.

Andreas

Am Donnerstag, den 18.06.2009, 14:03 +0000 schrieb arsenix:
> Charley -
>
> I'm sticking by my rec that you need to get the hardware fixed. Maybe
> windows throttles the machine more agressively when it gets hot instead
> of shutting down... maybe windows has better GPU thermal control (this
> is almost certainly true)... in any case your machine should not be
> shutting down with the fans at max under any loading situation. Lots of
> other users have T60's and no thermal shutdown issues in Jaunty. My
> machine (T61p) has the same fan control and power management system as
> yours... and has no issues. The thermal issues that originally led me
> to this thread were hardware issue related.
>
> I'm not saying there isn't some power management bug that makes your
> machine runs hotter under only Jaunty... I'm just saying you have a
> hardware problem if your machine gets itself hot enough to shutdown.
>
>
> James
>

-) are you running it on a hard desk? I know laptops that overheat even
from a simple tablecloth beneath them.

Answer: Yes, an oak table with no tablecloth.

-) are the fans clear? Depending upon the environment fans can get
really really fast clogged.

Answer: Yes, completely free of dust and debris. I can feel the airflow, especially with the fan running 4000-5000 rpm.

The laptop idles at 40-45C now, which is quite cool - even in comparison to other users with T61's. Operating temps ranging in the 60's. NORMAL.

Note: this is a 15" T61 with the 2.5 GHz, not 14" with a slower CPU.

Note: the issue does not occur in Hardy either.

Note: the thermal shutdowns occur while I'm not actively using the laptop, typically in the middle of the night. I wake up in the morning to a cold laptop, and log messages about thermal trip points reached in the middle of the night.

Note: I can reproduce the issue by running stress.

If Hardy runs cool, can it still be a hardware issue? If the thermal paste was bad, wouldn't it overheat all the time?

What is unique to Jaunty that that would be the only place I see the issue? Why doesn't it overheat *while I'm using the laptop*? (when it is indeed on my *lap* and not on a table)

Why are *so* many people *other* reporting this issue?

LP#370173

It seems there is a mix of issues here. For this bug we shall consider the issue that the latest Jaunty kernel is perceived to fail to manage the CPU temperature correctly and that it used to work correctly on Intrepid. If you have a different issue please file a separate bug.

To confirm this issue that we need to compare at least these two releases in a controlled manner. This will better allow us to confirm that this is a regression. To better identify the location of this regression and any potential fix it would be helpful to test the kernels below. For those of you comfortable installing kernels and also able to get their temperature information (either via sensors or another method), it would be helpful to get comparisons of a number of kernels (please pick one of the i386 or amd64 kernels to match your system). If you have another version where you know the bug does not exhibit, it would be helpful to include the same test results for that kernel too:

Intrepid Latest -- 2.6.27-14.34
    http://launchpadlibrarian.net/27522947/linux-image-2.6.27-14-generic_2.6.27-14.34_i386.deb
    http://launchpadlibrarian.net/27522807/linux-image-2.6.27-14-generic_2.6.27-14.34_amd64.deb
Jaunty Latest -- 2.6.28-13.44
    http://launchpadlibrarian.net/27405830/linux-image-2.6.28-13-generic_2.6.28-13.44_i386.deb
    http://launchpadlibrarian.net/27405998/linux-image-2.6.28-13-generic_2.6.28-13.44_amd64.deb
Mainline 2.6.28.x
    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.28.10/linux-image-2.6.28-02062810-generic_2.6.28-02062810_i386.deb
    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.28.10/linux-image-2.6.28-02062810-generic_2.6.28-02062810_amd64.deb
Mainline 2.6.30
    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.30/linux-image-2.6.30-020630-generic_2.6.30-020630_i386.deb
    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.30/linux-image-2.6.30-020630-generic_2.6.30-020630_amd64.deb

for each test using stress (sudo apt-get install stress):

    stress --cpu 16 --vm 2 --vm-bytes 128M

Run this for a measured period of time say 15 minutes and then gather the temperatures, I use sensors for this:

    $ sensors
    acpitz-virtual-0
    Adapter: Virtual device
    temp1: +46.0°C (crit = +100.0°C)
    temp2: +43.0°C (crit = +100.0°C)
    temp3: +0.0°C (crit = +100.0°C)

If you are able to identify your your fan speeds it would also be helpful to include that information in your report. In all cases please report your results in this bug. Thanks.

Hm... that is interesting that it doesn't happen while you are actively using the laptop...

Are you running xscreensaver with 3d graphics displays? If the GPU is heating things up that could make it run hot while you aren't around.

Do you close it when you aren't using it? A lot of heat is radiated through the keyboard.

James

I disabled the screen saver to see if that was the cause, even through the power manager is set to turn the display off after 20 mins.

The lid is open.

I will attempt to try the kernels in Andy's post this weekend if I possibly can and forward my results.

Maybe there is some hardware issue... I was wrong about the fan as you so well illustrated!

But it seems like its Jaunty-centric to me - perhaps i'll be able to get more data over the weekend.

It's nice to see more people are working on it.

Thank you for involving Andy, however, can you please describe how to install the kernels?

Shall I install all of them (amd64 only of course for my notebook) just by downloading and using dpkg? Wouldn't apt protest because of some dependencies? Can I install Intrepid's kernels in my updated Jaunty?

Good to see something happens finally. Thanks!

I'll add my 2 cents. My laptop, Pentium M 2.0 GHz, is experiencing CPU temps 10-20 degrees C hotter in Jaunty than in Intrepid. My fan is running properly, and actually more in Jaunty because of the higher temperatures. In Intrepid the fan would remain on low or off most of the time, with temps 35-40 idle, but with Jaunty the fan runs on medium or high constantly, with temps 45+ when idle.

The laptop is not overheating and shutting down, but it is definitely running hotter. So there are two separate issues at play: some people are not getting proper fan control. Others have proper fan control, but for a currently-unknown reason are experiencing higher CPU temps in Jaunty, in many cases leading to shutdowns even with the fan running on max.

In Intrepid my CPU temp never went over 69 C, whereas in Jaunty it regularly shoots up to 79 C. I luck out because my fan prevents it from going so high to initiate a shutdown. There are enough people working on debugging and I don't have time to install/test multiple kernels. My point is that this bug should be tackled as two issues: (1) fan problems, and (2) something unknown in Jaunty that is causing higher CPU temps separate from fan issues.

m4cph1sto:
My fan spins in circles derp derp.

I sped up the fan and the problem went away. Can you tell me why would that
happen?

On Thu, Jun 18, 2009 at 10:56 AM, m4cph1sto <email address hidden> wrote:

> I'll add my 2 cents. My laptop, Pentium M 2.0 GHz, is experiencing CPU
> temps 10-20 degrees C hotter in Jaunty than in Intrepid. My fan is
> running properly, and actually more in Jaunty because of the higher
> temperatures. In Intrepid the fan would remain on low or off most of
> the time, with temps 35-40 idle, but with Jaunty the fan runs on medium
> or high constantly, with temps 45+ when idle.
>
> The laptop is not overheating and shutting down, but it is definitely
> running hotter. So there are two separate issues at play: some people
> are not getting proper fan control. Others have proper fan control, but
> for a currently-unknown reason are experiencing higher CPU temps in
> Jaunty, in many cases leading to shutdowns even with the fan running on
> max.
>
> In Intrepid my CPU temp never went over 69 C, whereas in Jaunty it
> regularly shoots up to 79 C. I luck out because my fan prevents it from
> going so high to initiate a shutdown. There are enough people working
> on debugging and I don't have time to install/test multiple kernels. My
> point is that this bug should be tackled as two issues: (1) fan
> problems, and (2) something unknown in Jaunty that is causing higher CPU
> temps separate from fan issues.
>
> --
> Ubuntu 9.04 laptop overheat and shutdown
> https://bugs.launchpad.net/bugs/370173
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>
> Status in “linux” source package in Ubuntu: In Progress
>
> Bug description:
> Ubuntu 9.04, fresh install on Acer Aspire 5005.
> Not sure about package, but it seems the problem is with CPU frequency
> adjustment or fan control.
> Laptop shuts down right in the middle of CPU-greedy operation overheated. I
> haven't seen it before (since 7.10).
> The issue is discussed at several forums (eg.
> http://ubuntuforums.org/showthread.php?t=1130450)
>
> ----
> *-cpu
> product: AMD Turion(tm) 64 Mobile Technology ML-37
> vendor: Advanced Micro Devices [AMD]
> physical id: 1
> bus info: cpu@0
> version: 15.4.2
> size: 2GHz
> capacity: 2GHz
> width: 64 bits
> 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 syscall nx
> mmxext fxsr_opt x86-64 3dnowext 3dnow up pni lahf_lm cpufreq
>

Brooks: you clearly have "issue 1" that I referred to in my comment. I have "issue 2".

after week on windows, ubuntu has updates including -13 kernel... after which machine fails to boot! physically on-off, then boot. no programs running except boinc client with extra low processor/memory usage settings (intrepid ran same shit with default settings) and darn thing keeps stepping both cores up to 2G. idle intrepid ran at 800M. always.

now start firefox and... i am sorry to say, week on the windows kinda got me used to actually being able to use this thing as a laptop! after jaunty upgrade, same shock first made me search for a solution but after seeing longlasting inactivity/undeceided-ness @https://bugs.launchpad.net/ubuntu/+bug/361123 i dont know anymore what to think.

nice -applications should not affect cpu@ondemand. and they didnot @intrepid. looks like they do @jaunty. or at least someting weird goes on at this up/down stepping department, to say the least.

what i can see for sure, is that god-damn 2 core's speed/govenor can/should be set independently since jaunty. is that not the direction to look at?

anyway, that bs really got me now. back in a day i already lost 2 hard-disks thanks to those laptop-mode (or whatever) issues. now it seems like its really time for os-change, before this shit burns my pants off or catches fire by itself.

hopefully will those rushed-out releases stop eventually, if there's any users left to release them for.

Perhaps there is some magical "issue 2" . But how do you know your
fan is spinning properly if you haven't tried to speed it up? Being
unscientific is only creating more problems for this bug. m4cph1sto
you one of the people being very unscientific about this bug. I know
for a fact that we are all running the same fan triggers and that they
where not properly tested before shipping them to us. Speeding up
the fan is a quick fix, But like i said Cannonical doesn't give a
fuck about us, they would rather our hardware burn then pay for real
regression testing. This issue is still completely ignored even when
there is a known fix.

peace

On Sun, Jun 21, 2009 at 9:49 PM, e13<email address hidden> wrote:
> after week on windows, ubuntu has updates including -13 kernel... after
> which machine fails to boot! physically on-off, then boot. no programs
> running except boinc client with extra low processor/memory usage
> settings (intrepid ran same shit with default settings) and darn thing
> keeps stepping both cores up to 2G. idle intrepid ran at 800M. always.
>
> now start firefox and... i am sorry to say, week on the windows kinda
> got me used to actually being able to use this thing as a laptop! after
> jaunty upgrade, same shock first made me search for a solution but after
> seeing longlasting inactivity/undeceided-ness
> @https://bugs.launchpad.net/ubuntu/+bug/361123 i dont know anymore what
> to think.
>
> nice -applications should not affect cpu@ondemand. and they didnot
> @intrepid. looks like they do @jaunty. or at least someting weird goes
> on at this up/down stepping department, to say the least.
>
> what i can see for sure, is that god-damn 2 core's speed/govenor
> can/should be set independently since jaunty. is that not the direction
> to look at?
>
> anyway, that bs really got me now. back in a day i already lost 2 hard-
> disks thanks to those laptop-mode (or whatever) issues. now it seems
> like its really time for os-change, before this shit burns my pants off
> or catches fire by itself.
>
> hopefully will those rushed-out releases stop eventually, if there's any
> users left to release them for.
>
> --
> Ubuntu 9.04 laptop overheat and shutdown
> https://bugs.launchpad.net/bugs/370173
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>
> Status in “linux” source package in Ubuntu: In Progress
>
> Bug description:
> Ubuntu 9.04, fresh install on Acer Aspire 5005.
> Not sure about package, but it seems the problem is with CPU frequency adjustment or fan control.
> Laptop shuts down right in the middle of CPU-greedy operation overheated. I haven't seen it before (since 7.10).
> The issue is discussed at several forums (eg. http://ubuntuforums.org/showthread.php?t=1130450)
>
> ----
>     *-cpu
>          product: AMD Turion(tm) 64 Mobile Technology ML-37
>          vendor: Advanced Micro Devices [AMD]
>          physical id: 1
>          bus info: cpu@0
>          version: 15.4.2
>          size: 2GHz
>          capacity: 2GHz
>          width: 64 bits
>          capabilities: fpu fpu_exception wp vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 cl...

Michael you should try running your machine in a colder environment, like Antarctica. I think you will find your machine runs cooler and by your logic this PROVES that the problem is that you don't live in Antarctica.

Study this link: http://www.google.com/search?q=define%3A+causality

There are a lot of factors at play here and a lot of different hardware configurations, all of which control fans through completely different mechanisms. Problems will be fixed when they are isolated and identified... nothing in this thread has given enough information to do that yet. If you want to try to speed up your dell fans in the kernel, check out the "i8k" module from git and go to town.

James

Gush...

I've been experiencing the same problem over a year ago, and it'd been a topic reporting such a bug somewhere else.
Those days I had to change my OS to Windows, now I decided to give Ubuntu a try, and (sorry for being upset) it's all the same! The problem didn't vanish magically, it's still here, isn't it annoying?
Please, could anyone do something with it, and not make workarounds, that just does not work...
And btw. isn't this bug worth fixing enought to be marked as critical?

michael, having this as my primary computer for about 7 months now and using it every single day with all the same programs running the same way in quiet environment, one gets used to how this thing behaves, how fast fan rotates and ho hot it gets. as does one notice how often and for how long ondemand cpu gets stepped to higher speed.

my laptop does not shut down. it just jumps its speed up way too often compared to intrepid. fan noise is very much the same and when getting hotter, fan also speeds up. never ran stresstest to cause thermal shutdowns. fan speeds up anyway, when i let this thing jump to whatever speeds it feels like.

but because i cannot be bothered to start using that as a desktop-computer, this heat thing is just not acceptable. as would not be increased fan speeds. this just is not justified for things i run on that computer while cpu set to ondemand. it needs to be cool and not too noisy. and it used to be like that with intrepid. still is with windows. not with jaunty.

clear change after jaunty were feature that 2 cores can now be set independently to run whatever speed. and right after upgrade, all the same programs running and fan making all the same noise, all that was different, was cpu speedups occuring more often and lasting far longer... talk all about you want about fan speed being not correct. i know how much processor my programs used to need and how hot that made my computer and how fast fan used to run to cope with that. fan still runs the same, but same temp can only be achieved by setting cores to 800M or 1.2G. they used to be set to ondemand and mostly idle machine with nice-d boinc running used to stay at 800 all the time.

Architecture: amd64
DistroRelease: Ubuntu 9.04
HibernationDevice: RESUME=UUID=322532a4-3479-46ec-a1b4-1643352ff1b9
MachineType: Gateway P-6860FX
NonfreeKernelModules: nvidia
Package: linux-image-2.6.28-13-generic 2.6.28-13.44
PackageArchitecture: amd64
ProcCmdLine: root=UUID=d140a130-2891-42b5-8055-e30dfafe24d2 ro quiet splash acpi_osi="Linux"
ProcEnviron:
 SHELL=/bin/bash
 LANG=en_US.UTF-8
ProcVersionSignature: Ubuntu 2.6.28-13.44-generic
Uname: Linux 2.6.28-13-generic x86_64
UserGroups: adm admin audio avahi avahi-autoipd cdrom crontab dialout dip fax fuse gdm haldaemon klog libuuid lpadmin messagebus mlocate netdev plugdev polkituser pulse pulse-access pulse-rt root sambashare saned ssh ssl-cert syslog tape users video

Hi, gateway 6860fx and an acer 5620-6830, both core duo both 965m.
I submitted a "apport-collect -p linux-image-`uname -r` 370173"
ran stress, hit 68C in 1 min 30 sec, fans never moved off idle speed (hand check only. I'm not disassembling it again to put an amp meter on it to check voltages). I stopped the test @ 75C after 4 minutes 15 sec because I'm not willing to let my hw get any hotter than that, sorry.

I had a similar problem: 373921 but i solved it by cleaning the fan exhaust conducts.
I'm another one of the timeless persons who can't install other kernels, but as a reference,
and in case this is a hardware issue, here are my reads using Andy's parameters only on one kernel.
Maybe they can serve as a reference.
Toshiba satellite a215-s5825
Athlon 64x2 TK-55
Ati x1200
2.6.28-13-generic #44-Ubuntu SMP Tue Jun 2 07:55:09 UTC 2009
temp1: +54.0°C (crit = +102.0°C) (idle)
temp1: +68.0°C (crit = +102.0°C) (when fan speeds up, about a minute after starting stress)
temp1: +72.0°C (crit = +102.0°C) (5 minutes of stress)
temp1: +77.0°C (crit = +102.0°C) (10 minutes of stress)
temp1: +80.0°C (crit = +102.0°C) (15 minutes of stress)
        After killing stress
temp1: +64.0°C (crit = +102.0°C) (1 minute after)
temp1: +60.0°C (crit = +102.0°C) (5 minutes after)
temp1: +58.0°C (crit = +102.0°C) (10 minutes after)

@Michal Pěnka -- yes you can just download the appropriate kernel for you i386/amd64 and just install those with dpkg. They should install in parallel and you can simply select them in the main grub menu.

@all -- the key desire here is to isolate exactly when this issue appeared. There seems to be enough people indicating they have an issue to make it likely that later kernels are triggering this behaviour. As generally the kernel does not get involved in fan control directly it must be something else triggering the change. As there are some 20k changes in every mainline release and normally at least 2 mainline releases in every Ubuntu release, there is a huge amount of change in each kernel. This huge change makes it hard to pinpoint such a behaviour. The normal way to do this is to identify kernel versions which exhibit the behaviour and those which do not and bisect the gap to narrow down the the exact mainline release produced the change, and then within that release. This is the source of my requests for testing of specific kernel versions.

For me, the 2.6.28 kernel does trigger it. Running 2.6.27-14 works fine.

(Actually, subjectivly 2.6.27-14 is not perfect either, but I haven't
managed yet to kill it via burnK7, while 2.6.28 takes less than 60s of
2xburnK7 (dualcore) to shutdown)

To summarize:

Hardy/Intrepid ran perfect on the hardware (temperature-wise)

2.6.27-14 feels slightly to loud on fan noise, but that's strictly
subjective. The scripts that I wrote to burn and log have been not able
to force a thermal shutdown for over 15 minutes.

2.6.28 can be easily crashed by
So it's really only the kernel, or something in userland that depends
upon the currently running kernel. Demonstrated via logs/scripts that I
posted to my bug report. Thermal shutdown happens after less than 60
seconds.

So just booting a different kernel makes my laptop go into thermal
shutdown, so it is a kernel or kernel-specific userland problem.

Andreas

Am Donnerstag, den 25.06.2009, 08:31 +0000 schrieb Andy Whitcroft:
> @Michal Pěnka -- yes you can just download the appropriate kernel for
> you i386/amd64 and just install those with dpkg. They should install in
> parallel and you can simply select them in the main grub menu.
>
> @all -- the key desire here is to isolate exactly when this issue
> appeared. There seems to be enough people indicating they have an issue
> to make it likely that later kernels are triggering this behaviour. As
> generally the kernel does not get involved in fan control directly it
> must be something else triggering the change. As there are some 20k
> changes in every mainline release and normally at least 2 mainline
> releases in every Ubuntu release, there is a huge amount of change in
> each kernel. This huge change makes it hard to pinpoint such a
> behaviour. The normal way to do this is to identify kernel versions
> which exhibit the behaviour and those which do not and bisect the gap to
> narrow down the the exact mainline release produced the change, and then
> within that release. This is the source of my requests for testing of
> specific kernel versions.
>

Architecture: i386
DistroRelease: Ubuntu 9.04
HibernationDevice: RESUME=UUID=84817c49-1557-4bf3-af9b-5679d04468f7
Lsusb:
 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
 Bus 003 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
 Bus 002 Device 002: ID 0458:003a KYE Systems Corp. (Mouse Systems)
 Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
MachineType: Acer, inc. Aspire 5000
Package: linux-image-2.6.28-13-generic 2.6.28-13.44
PackageArchitecture: i386
ProcCmdLine: root=UUID=11a14c8c-ec74-40d4-8cff-924a320898ca ro quiet splash
ProcEnviron:
 SHELL=/bin/bash
 PATH=(custom, user)
 LANG=ru_RU.UTF-8
ProcVersionSignature: Ubuntu 2.6.28-13.44-generic
Uname: Linux 2.6.28-13-generic i686
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare

Architecture: i386
DistroRelease: Ubuntu 9.04
HibernationDevice: RESUME=UUID=91117883-e8de-4553-b567-27afcf45749c
Lsusb:
 Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
 Bus 001 Device 002: ID 0c45:62c0 Microdia Pavilion Webcam
 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
MachineType: Hewlett-Packard HP Pavilion dv2000 (EZ644UA#ABA)
NonfreeKernelModules: nvidia
Package: linux-image-2.6.28-11-generic 2.6.28-11.42
PackageArchitecture: i386
ProcCmdLine: root=UUID=53bb2779-d451-4c02-88d9-5fe4eb5fcad2 ro quiet splash
ProcEnviron:
 SHELL=/bin/bash
 PATH=(custom, no user)
 LANG=en_US.UTF-8
ProcVersionSignature: Ubuntu 2.6.28-11.42-generic
Uname: Linux 2.6.28-11-generic i686
UserGroups: adm admin audio cdrom dialout dip floppy fuse lpadmin plugdev uml-net video

The report to https://bugs.launchpad.net/ubuntu/+bug/370173/comments/129, take it from me

I've started with my current kernel:
Linux version 2.6.28-13-generic (buildd@palmer) (gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4) ) #44-Ubuntu SMP Tue Jun 2 07:57:31 UTC 2009

Before stress $ sensors
acpitz-virtual-0
Adapter: Virtual device
temp1: +53.0°C (crit = +97.0°C)
k8temp-pci-00c3
Adapter: PCI adapter
Core0 Temp: +57.0°C

Laptop shutdown after 8 min 30 sec under stress -cpu 16 --vm 2 --vm-bytes 128M
8 min 30 sec - that's the time I will use for all further tests
---

Next I switched back to Jaunty kernel:
Linux version 2.6.27-14-generic (buildd@vernadsky) (gcc version 4.3.2 (Ubuntu 4.3.2-1ubuntu12) ) #1 SMP Fri Jun 5 10:14:59 UTC 2009

Before stress $ sensors
acpitz-virtual-0
Adapter: Virtual device
temp1: +53.0°C (crit = +97.0°C)
k8temp-pci-00c3
Adapter: PCI adapter
Core0 Temp: +63.0°C

After 8 min 30 sec stress $ sensors
acpitz-virtual-0
Adapter: Virtual device
temp1: +74.0°C (crit = +97.0°C)
k8temp-pci-00c3
Adapter: PCI adapter
Core0 Temp: +85.0°C
---

While installing 2.6.30 upcoming kernel got the following errors:
* vboxdrv (2.1.4) Cannot install this module.
* vboxnetflt (2.1.4) Cannot install this module.

Nevertheless
Linux version 2.6.30-020630-generic (root@zinc) (gcc version 4.2.3 (Ubuntu 4.2.3-2ubuntu7)) #020630 SMP Wed Jun 10 09:45:40 UTC 2009

Before stress $ sensors
acpitz-virtual-0
Adapter: Virtual device
temp1: +41.0°C (crit = +97.0°C)
k8temp-pci-00c3
Adapter: PCI adapter
Core0 Temp: +45.0°C

After 8 min 30 sec stress $ sensors
acpitz-virtual-0
Adapter: Virtual device
temp1: +74.0°C (crit = +97.0°C)
k8temp-pci-00c3
Adapter: PCI adapter
Core0 Temp: +75.0°C
---

Just like that: 2.6.30 seems promissing.

Trying to supply more data but for me there seems to be no visible issue:

Kernel: 2.6.28-13-generic (32bit)
Host: Dell Inspiron 1525 (AMD Turion(tm) 64 X2 Mobile Technology TL-64)

With "stress --cpu 16 --vm 2 --vm-bytes 128M", I only get a temperature rise of 44°C to 66°C while the fan speeds up accordingly.

I am getting the same issue here too
Fresh install of 9.04 today. The installer failed and shut down the PC at about 25-27%. I tried to install 3 times.

I made a bootable USB to install Ubuntu and this time it worked.

With nothing running, first boot I noticed the PC was very slow. I checked the performance meter, 1/2 CPU was constantly at 100%.
It was switching CPU over every 60 sec or so.

Even on LiveCD, same issue. Very sluggish, 100% CPU used with no application running. CPU FAN running high speed all the time.
No process is using 100%. Only process to use about£% is the Gnome manager. Everything else at 0%.

My laptop is a Compaq Presario CQ60, AMD Turion 64 X2.

HTH

Yeah ok I just discovered, this is a known issue since the last couple of years.

You can fix it by adding "noapic nolapic" to your boot options.

HTH

This bug is affecting me as well. I've attached my lshw just in case it could be of any help.

All my testing on my kit here shows the basic same behaviour with 2.6.27/28/30/31 based kernels, no particular kernel seems to show any advantage. This is on a T30, a Dell 1537, and also on a quad core server. Only the T30 shows any strong shift up in temperature when loaded, but does consistently on all those tested kernels.

@arsenix - I took the Lenovo T61 apart and applied a liberal coat of thermal paste on the cpu and gpu. I then tightened down the heatsink/heatpipe fan unit firmly.

It seems to be running *much* cooler - I ran the stress test again and didn't hit 70C...

This is with the fan running disengaged just shy of 5000 rpm.

This may be specific to Lenovo T61's - but if you have one, its worth checking out. There is full documentation of the procedure to remove the fan/heatsink unit on Lenovo's site. Just search for your model in the support section.

I don't think this is a fix necessarily, but the temps are much improved.

Perhaps there is an issue with the T30 also. It seems Lenovo must have been running out of thermal grease when they put this one together...

Using a Dell latitude D620 running Ubuntu 8.10 using kernel - 2.6.27-14-generic, I can overheat the laptop in ~ 25 minutes running the program stress with the following syntax: stress --cpu 8 --io 4 --vm 2 --vm-bytes 128M

While running the program stress I have loaded the i8k module and gkrellm to force the fan on full time. The fan is running at full speed prior to initiating the test with no load. Even with the fan still running, the laptop will shutdown after ~ 25 minutes. At start of the run with nothing running -
acpitz-virtual-0
Adapter: Virtual device
temp1: +55.5°C (crit = +126.0°C)

coretemp-isa-0000
Adapter: ISA adapter
Core 0: +56.0°C (high = +100.0°C, crit = +100.0°C)

Shutdown at -
acpitz-virtual-0
Adapter: Virtual device
temp1: +90.5°C (crit = +126.0°C)

coretemp-isa-0000
Adapter: ISA adapter
Core 0: +89.0°C (high = +100.0°C, crit = +100.0°C)

I have a HP DV1240 running Ubuntu Jaunty running kernel - 2.6.28-11-generic which is running really hot also. I haven't attempted to stress test it to determine its thermal threshold, but it gets uncomfortably hot also and the fan doesn't appear to kick nearly enough.

I'll attach a lshw of the dell 620.

dlf

We may have found one potential trigger for this behaviour in testing. We have seen some correlation between hibernate and CPU overheating. Therefore could those of you definatly suffering this issue, ie. those who have had thermal shutdowns, please compare the behaviour of the machine following a clean reboot, and then after a hibernate. As before could you record the temperature reported under sensors when running the stress applications as detailed above. It might also be instructive to try the same after just a suspend. Please report back here. Thanks!

never used hibernate on my t61. excessive heating starts right after clean restart.

i've never used hibernate on my tx2110.
seems that fan does not spin up early enough.
i tryed to install the powernowd package, as it's default on intrepid, but this makes no difference on jaunty 64.

@Andy Whitcroft:
I don't use hibernate on my Acer Aspire 5005. Still need that report?

I have a Lenovo X61, and have experienced shutdown due to excessive heat. I don't use hibernate.

Same here, I dont use hibernate and from clean boot (cycle power) The system overheats. I've lifted both laptops with yahtzi dice and run deskfans under each to keep cool, and I just boot into windows if I have to go portable. with the deskfans on and proc set to perf i never hit 55C even playing games just like it should be.

Again, both systems have pm965 chips which have no way to control fans.

6860fx temps with deskfan blowing across system, cpu 35c, hdd0 41c, hdd1 43c, 8800gts 43c.
6860fx temps with deskfan off, cpu 58c, hdd0 53c, hdd1 55c, 8800gts 49c.

obviously, i just need a way to turn the fans on and leave them on, I could give a @#$@ less about battery life or fan noise. I didnt buy a desktop replacement to simulate a palmtop or ibook. I bought it for power and performance. this comes at the cost of heat. Active cooling is the answer to heat, if I can turn it on. I know, you're working on it. It's just a bit frustrating to read the money and excitement being poured into ubuntu cloud suites when the systems overheat before you can connect to said cloud.

Do my eyes deceive me? Did canonical finely recognize this after 96
emails to my inbox? Andy you are testing this on a desktop? This is
a problem with laptops!!!!! Mr Prabhakar, downgrade to 8 or use
Fedora if you want an immediate fix.

Peace out

On Thu, Jul 9, 2009 at 12:26 AM, Prabhakar Nagral<email address hidden> wrote:
>
> Hi Andy,
>
> I have Lenovo W500 and after loading 9.04, it gets heated tremendously.
> Can you please fix this bug ASAP? Or else, provide me with workaround.
>
> Regards,
> -Prabhakar Nagral
>
> --- On Mon, 6/7/09, Andy Whitcroft <email address hidden> wrote:
>
>> From: Andy Whitcroft <email address hidden>
>> Subject: [Bug 370173] Re: Ubuntu 9.04 laptop overheat and shutdown
>> To: <email address hidden>
>> Date: Monday, 6 July, 2009, 8:47 PM
>> All my testing on my kit here shows
>> the basic same behaviour with
>> 2..6.27/28/30/31 based kernels, no particular kernel seems
>> to show any
>> advantage.  This is on a T30, a Dell 1537, and also on
>> a quad core
>> server.  Only the T30 shows any strong shift up in
>> temperature when
>> loaded, but does consistently on all those tested kernels.
>>
>> --
>> Ubuntu 9.04 laptop overheat and shutdown
>> https://bugs.launchpad.net/bugs/370173
>> You received this bug notification because you are a direct
>> subscriber
>> of the bug.
>>
>> Status in “linux” package in Ubuntu: In Progress
>>
>> Bug description:
>> Ubuntu 9.04, fresh install on Acer Aspire 5005.
>> Not sure about package, but it seems the problem is with
>> CPU frequency adjustment or fan control.
>> Laptop shuts down right in the middle of CPU-greedy
>> operation overheated. I haven't seen it before (since
>> 7.10).
>> The issue is discussed at several forums (eg. http://ubuntuforums.org/showthread.php?t=1130450)
>>
>> ----
>>      *-cpu
>>           product: AMD Turion(tm)
>> 64 Mobile Technology ML-37
>>           vendor: Advanced Micro
>> Devices [AMD]
>>           physical id: 1
>>           bus info: cpu@0
>>           version: 15.4.2
>>           size: 2GHz
>>           capacity: 2GHz
>>           width: 64 bits
>>           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
>> syscall nx mmxext fxsr_opt x86-64 3dnowext 3dnow up pni
>> lahf_lm cpufreq
>>
>
>
>      Love Cricket? Check out live scores, photos, video highlights and more. Click here http://cricket.yahoo.com
>
> --
> Ubuntu 9.04 laptop overheat and shutdown
> https://bugs.launchpad.net/bugs/370173
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>
> Status in “linux” package in Ubuntu: In Progress
>
> Bug description:
> Ubuntu 9.04, fresh install on Acer Aspire 5005.
> Not sure about package, but it seems the problem is with CPU frequency adjustment or fan control.
> Laptop shuts down right in the middle of CPU-greedy operation overheated. I haven't seen it before (since 7.10).
> The issue is discussed at several forums (eg. http://ubuntuforums.org/showthread.php?t=1130450)
>
> ----
>     *-cpu
>          product: AMD Turion(tm) 64 Mobile Technology ML-37
>          vendor: Advanced Micro Devices [AMD]
>        ...

@Michael -- there are many bugs reported against the kernel. This one came to our attention as there were a number of duplicates which raises its profile and its priority for attention. I actually said that I had tested it on all the kit I had here and had not been able to reproduce it. That included on two laptops neither of which show the issue. This is not an issue affecting every laptop therefore not trivial to isolate. We have now since found one person who has a reproducible trigger for what appears to be this behaviour and they are attempting to bisect the kernel to find the cause. The problem is looking like it is in mainline also, and likely was introduced in the 2.6.28 release, and so you may well hit the issue in any distro.

For those of you who have experienced thermal shutdowns could we get your acpidump and dmidecode output attached to this bug, run the two commands below and attach the result to this bug:
    sudo acpidump -o acpidump.txt
    sudo dmidecode >dmidecode.txt

For those of you comfortable testing kernels we currently suspect the problem was introduced in v2.6.28 mainline, you might test the v2.6.27 and v2.6.28 kernels from the Mainline Kernel repository to confirm it is not present in v2.6.27 and is in v2.6.28. Those kernel can be found here:

    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.27/
    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.28/

and instructions found here:

    https://wiki.ubuntu.com/KernelTeam/MainlineBuilds

If the problem is found to be in v2.6.28 and not in v2.6.27 then it would also be worth testing the following stable updates kernels also:

    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.27.23/
    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.27.12/

In all cases please report your testing results back here. Thanks.

Hey Andy sounds like you have a heap of fail on your hands. I have
had a patch for a few weeks now. Ubuntu's fan speed is too slow, and
this is the cause of 99% of the 100 posts on this thread. You probalby
can't reproduce the issue because your laptops are not powerful
enough. I can't even run VMWare on my intel dual 2.4ghz dell for more
than 20 min without speeding up the fan. For me its not a matter of
reproducing this, it is a huge barrier that i faced every day until I
fixed it. People are really pissed off, and i was one of them.

peace

On Thu, Jul 9, 2009 at 2:28 AM, Andy Whitcroft<email address hidden> wrote:
> @Michael -- there are many bugs reported against the kernel.  This one
> came to our attention as there were a number of duplicates which raises
> its profile and its priority for attention.  I actually said that I had
> tested it on all the kit I had here and had not been able to reproduce
> it.  That included on two laptops neither of which show the issue.  This
> is not an issue affecting every laptop therefore not trivial to isolate.
> We have now since found one person who has a reproducible trigger for
> what appears to be this behaviour and they are attempting to bisect the
> kernel to find the cause.  The problem is looking like it is in mainline
> also, and likely was introduced in the 2.6.28 release, and so you may
> well hit the issue in any distro.
>
> For those of you who have experienced thermal shutdowns could we get your acpidump and dmidecode output attached to this bug, run the two commands below and attach the result to this bug:
>    sudo acpidump -o acpidump.txt
>    sudo dmidecode >dmidecode.txt
>
> For those of you comfortable testing kernels we currently suspect the
> problem was introduced in v2.6.28 mainline, you might test the v2.6.27
> and v2.6.28 kernels from the Mainline Kernel repository to confirm it is
> not present in v2.6.27 and is in v2.6.28. Those kernel can be found
> here:
>
>    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.27/
>    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.28/
>
> and instructions found here:
>
>    https://wiki.ubuntu.com/KernelTeam/MainlineBuilds
>
> If the problem is found to be in v2.6.28 and not in v2.6.27 then it
> would also be worth testing the following stable updates kernels also:
>
>    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.27.23/
>    http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.27.12/
>
> In all cases please report your testing results back here.  Thanks.
>
> --
> Ubuntu 9.04 laptop overheat and shutdown
> https://bugs.launchpad.net/bugs/370173
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>
> Status in “linux” package in Ubuntu: In Progress
>
> Bug description:
> Ubuntu 9.04, fresh install on Acer Aspire 5005.
> Not sure about package, but it seems the problem is with CPU frequency adjustment or fan control.
> Laptop shuts down right in the middle of CPU-greedy operation overheated. I haven't seen it before (since 7.10).
> The issue is discussed at several forums (eg. http://ubuntuforums.org/showthread.php?t=1130450)
>
> ----
>     *-cpu
>          product: AMD Turion(...

ACPI dump

DMIDECODE dump

Ok, my case is clearly not hibernate related, as it happens as demonstrated immediatly after boot with 2.6.28 (burnK7 kills the laptop in less than a minute).

Now, as far as I've been able to tell, 2.6.30 mainline does run stable on my laptop.

I've installed it from http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.30/

The only issues, and not sure how to fix it, are:

-) no restricted kernel modules, so no WLAN for me.
-) no nvidia kernel module so no X11 for me. (this might work with the nouveau driver, but I haven't tested it yet).

Any idea how to fix these two things?

Andreas

@Michael -- indeed the T30 is not very powerful, but the other has a top of the range dual core cpu in it. The fact is there is an underlying trigger which only affects a small subset of users. That is how this bug got released in the first place, you can be sure if it was routinely hitting any of the development teams during the release cycle we would have known about it much sooner. As for your fix I am glad you have a way to keep working, but that is a work around which can only work on a small subset of machines, it is not a solution to the underlying issue either for you or anyone else.

In the majority of the reported cases we are not meant to have any control or intervention in the fan control. We rely on the BIOS setup for these components and there is no software component in the monitoring of temperature nor in the control of the fan speed in response. It is therefore very hard to identify what has changed in the kernel to trigger this change in behaviour. Clearly there is a kernel component or interaction triggering the issue as only changing the kernel changes the bahaviour. The tests I have requested would help us eliminate some of the kernel delta which is suspected as triggering the issue (the difference between v2.6.27 and v2.6.28).

One thing that I forgot, my WLAN would be 03:00.0 Network controller: Broadcom Corporation BCM4328 802.11a/b/g/n (rev 03)

@Prabhakar Nagral:

Simply using the power manager from a different release is anything but simple.
It's not as if "the power manager" is one thing, power management is a cooperation of:

-) hardware
-) firmware (BIOS/ACPI, whatever)
-) kernel
-) userspace

Worse, some of these parts are almost certainly "defective" (in a tight sense), e.g. I wouldn't count on 99% of the BIOSes implementing full specs, beyond what a certain legacy OS from the north west needs. So even changes how the kernel interacts with the BIOS that have no direct relevance (at first glance) to power management, can influence the situation. Furthermore, it all can depend upon:

-) configuration (hardware/software)
-) production variance/tolerance. (E.g. some devices X of manufacturer Y work, some don't.)

The problems that this issue has, are IMHO:

-) it's still not really reproducible.
-) it's still probably a mix of different causes (e.g. there are enough cases that are not hibernation related)
-) it seems to be not reproducible even on the same hardware, there seems to be people running officially the same hardware that have/have not problems.
-) it's dependent upon the environment. E.g. while some hard cases have the problem always, others only experience them in warm weather.
-) some cases might be hardware related. Enough cases are not so.

So generally speaking, as a software developer, this issue is classification wise pure hell to fix. At least Ubuntu has assigned somebody that takes a look into this. Don't expect a solution that fixes all problems quickly :(

@Andy

Wondering, do we have a git repo of the kernels used by ubuntu, wondering if I might try out git bisect myself on my laptop?

Andreas