WARNING: Intrepid might burn down MacBook Pro

[oss_test_launchpad]

Binary package hint: linux-image-2.6.27-1-generic

Testing Ubuntu 8.10 Alpha 4 64-Bit on a MacBook Pro rev. 3. Installing today's updates, the MacBook gets extremely hot and automatically goes into standby. This looks like an emergency standby caused by the EFI or something because it got overheated. Unless you put a metal plate or something under it, the fan runs on highest level at full power, which is not the usual.

Is there a tool you can see the CPU temperature with?

[Leann Ogasawara (leannogasawara)]

Hi oss_test_launchpad,

Can you confirm this did not exist with the 2.6.26 kernel previously available in Intrepid? Thanks.

[oss_test_launchpad (oss-test-launchpad)]

I can confirm that.

[oss_test_launchpad (oss-test-launchpad)]

Seems it has become a little better with 2.6.27-2-generic: The notebook still feels as if it was close to boiling after a few hours' runtime, but there was no emergeny standby since then. However, this does not have to be due to kernel improvements: Air temperatures have gone down significantly during the past few days, so maybe the factor that produces the improvement is extrinsic.

After one hour's runtime, 2/3 of it (except the display) feel like a good hot cip of coffee. What worries me is that since the update to 2.6.27-2-generic the fan does not run any more at all, or at least it runs at such a low level that you hardly hear it.

[oss_test_launchpad (oss-test-launchpad)]

Problem persists. After three hours, you'd think about throwing away your microwave and cooking on your notebook instead.

This seems to be a problem that especially occurs during high cpu usage.

[oss_test_launchpad (oss-test-launchpad)]

Would anyone please mark this as critical.

[josh04 (josh04)]

For checking the temperature, try

$ cat /proc/acpi/thermal_zone/*/temperature

[oss_test_launchpad (oss-test-launchpad)]

bash: $: command not found

[parren (peter-arrenbrecht)]

Don't type the "$". It just prefixes commands your supposed to enter at the command line. So please run:

  cat /proc/acpi/thermal_zone/*/temperature

(Interested in this as I've just ordererd a new MacBook Pro.)

[oss_test_launchpad (oss-test-launchpad)]

computer@a-computer:~$ cat /proc/acpi/thermal_zone/*/temperature
cat: /proc/acpi/thermal_zone/*/temperature: No such file or directory

[beefcurry (jonzwong)]

try

watch acpi -V

that should tell you the temperature.

[oss_test_launchpad (oss-test-launchpad)]

Actually, it doesn't. It only tells you time and date plus

----
Every 2,0s: acpi -V Mon Sep 15 15:51:02 2008

Battery 0: Full, 96%
AC Adapter 0: on-line
Cooling 0: Processor 0 of 10
Cooling 1: Processor 0 of 10
----

[Michael Chang (thenewme91)]

What does

lsmod

give you? Perhaps you don't have the right modules loaded, for whatever reason...

Also, the following command tells us which kernel packages you have installed.

dpkg -l 'linux-*'

Also try the instructions at https://wiki.ubuntu.com/DebuggingACPI and see if they are of any help in finding the nature of the problem.

[oss_test_launchpad (oss-test-launchpad)]

user@user-computer:~$ lsmod
Module Size Used by
ipv6 314312 10
binfmt_misc 18700 1
af_packet 29568 4
btusb 21656 2
rfcomm 51104 2
l2cap 33280 9 rfcomm
bluetooth 70692 7 btusb,rfcomm,l2cap
ppdev 16904 0
acpi_cpufreq 16400 1
cpufreq_stats 14468 0
cpufreq_powersave 10368 0
cpufreq_conservative 16392 0
cpufreq_userspace 12420 0
cpufreq_ondemand 16400 1
freq_table 13568 3 acpi_cpufreq,cpufreq_stats,cpufreq_ondemand
sbs 22288 0
wmi 15808 0
container 12288 0
pci_slot 13704 0
sbshc 14592 1 sbs
iptable_filter 11520 0
ip_tables 28176 1 iptable_filter
x_tables 31752 1 ip_tables
parport_pc 44200 0
lp 19588 0
parport 50096 3 ppdev,parport_pc,lp
snd_hda_intel 489264 3
snd_pcm_oss 52608 0
snd_mixer_oss 25088 1 snd_pcm_oss
snd_pcm 99208 2 snd_hda_intel,snd_pcm_oss
snd_seq_dummy 11524 0
snd_seq_oss 42368 0
snd_seq_midi 15872 0
uvcvideo 68488 0
joydev 20736 0
snd_rawmidi 34176 1 snd_seq_midi
compat_ioctl32 18176 1 uvcvideo
arc4 10368 2
snd_seq_midi_event 16768 2 snd_seq_oss,snd_seq_midi
ecb 11520 2
crypto_blkcipher 27780 1 ecb
appletouch 18564 0
videodev 46720 2 uvcvideo,compat_ioctl32
iTCO_wdt 21072 0
iTCO_vendor_support 12420 1 iTCO_wdt
v4l1_compat 24580 2 uvcvideo,videodev
ath9k 296120 0
appleir 13440 0
mac80211 253568 1 ath9k
snd_seq 67168 6 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_seq_midi_event
snd_timer 34320 2 snd_pcm,snd_seq
cfg80211 37136 1 mac80211
snd_seq_device 16404 5 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_rawmidi,snd_seq
snd 79304 15 snd_hda_intel,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_seq_oss,snd_rawmidi,snd_seq,snd_timer,snd_seq_device
soundcore 16800 1 snd
snd_page_alloc 17680 2 snd_hda_intel,snd_pcm
video 28948 0
output 11776 1 video
battery 21128 0
ac 13448 0
shpchp 42140 0
pci_hotplug 39472 1 shpchp
intel_agp 39280 0
button 15904 0
evdev 20512 13
ext3 150544 1
jbd 66472 1 ext3
mbcache 17924 1 ext3
sg 45408 0
sr_mod 24772 0
sd_mod 45864 3
crc_t10dif 10240 1 sd_mod
cdrom 47784 1 sr_mod
ata_generic 14212 0
pata_acpi 13568 0
usbhid 39776 0
hid 59072 1 usbhid
ata_piix 29444 2
libat...

[oss_test_launchpad (oss-test-launchpad)]

user@user-computer:~$ dpkg -l 'linux-*'
Gewünscht=Unbekannt/Installieren/R=Entfernen/P=Vollständig Löschen/Halten
| Status=Nicht/Installiert/Config/U=Entpackt/Fehlgeschl. Konfiguration/
         Halb installiert/Trigger erWartet/Trigger anhängig
|/ Fehler?=(kein)/Halten/R=Neuinst notw/X=beide (Status, Fehler: GROSS=schlecht)
||/ Name Version Beschreibung
+++-==============-==============-============================================
un linux-boot-loa <keine> (keine Beschreibung vorhanden)
un linux-doc-2.6. <keine> (keine Beschreibung vorhanden)
un linux-doc-2.6. <keine> (keine Beschreibung vorhanden)
ii linux-generic 2.6.27.3.3 Complete Generic Linux kernel
un linux-headers <keine> (keine Beschreibung vorhanden)
un linux-headers- <keine> (keine Beschreibung vorhanden)
ii linux-headers- 2.6.26-5.17 Header files related to Linux kernel version
ii linux-headers- 2.6.26-5.17 Linux kernel headers for version 2.6.26 on x
ii linux-headers- 2.6.27-1.2 Header files related to Linux kernel version
ii linux-headers- 2.6.27-1.2 Linux kernel headers for version 2.6.27 on x
ii linux-headers- 2.6.27-2.3 Header files related to Linux kernel version
ii linux-headers- 2.6.27-2.3 Linux kernel headers for version 2.6.27 on x
ii linux-headers- 2.6.27-3.4 Header files related to Linux kernel version
ii linux-headers- 2.6.27-3.4 Linux kernel headers for version 2.6.27 on x
ii linux-headers- 2.6.27.3.3 Generic Linux kernel headers
un linux-image <keine> (keine Beschreibung vorhanden)
un linux-image-2. <keine> (keine Beschreibung vorhanden)
ii linux-image-2. 2.6.26-5.17 Linux kernel image for version 2.6.26 on x86
ii linux-image-2. 2.6.27-1.2 Linux kernel image for version 2.6.27 on x86
ii linux-image-2. 2.6.27-2.3 Linux kernel image for version 2.6.27 on x86
ii linux-image-2. 2.6.27-3.4 Linux kernel image for version 2.6.27 on x86
ii linux-image-ge 2.6.27.3.3 Generic Linux kernel image
un linux-initramf <keine> (keine Beschreibung vorhanden)
un linux-kernel-h <keine> (keine Beschreibung vorhanden)
un linux-kernel-l <keine> (keine Beschreibung vorhanden)
ii linux-libc-dev 2.6.27-3.4 Linux Kernel Headers for development
ii linux-restrict 2.6.26-5.12 Non-free Linux kernel modules for version 2.
ii linux-restrict 2.6.27-1.1 Non-free Linux kernel modules for version 2.
ii linux-restrict 2.6.27-2.3 Non-free Linux kernel modules for version 2.
ii linux-restrict 2.6.27-3.4 Non-free Linux kernel modules for version 2.
ii linux-restrict 2.6.27-3.4 Non-free Linux 2.6.27 modules helper script
ii linux-restrict 2.6.27.3.3 Restricted Linux modules for generic kernels
ii linux-sound-ba 1.0.17.dfsg-2u base package for ALSA and OSS sound systems
un linux-source-2 <keine> (keine Beschreibung vorhanden)
un linux-source-2 <keine> (keine Beschreibung vorhanden)

[oss_test_launchpad (oss-test-launchpad)]

Sorry I don't have time to work myself into this debug thing right now. In case that no one else contributes the information needed, please come back to me in a couple of days.

[oss_test_launchpad (oss-test-launchpad)]

Still no change so far. Once every hour you have to shut down and let it cool for ten minutes.

[Peter Cordes (peter-cordes)]

Is something using a lot of CPU time? run top, and look at the load averages at the top right. (number of tasks waiting running/waiting to run averaged over the last 5/10/15 minutes).

 If it's not close to zero, something's keeping your CPU busy.

 Is your CPU idling at low speed, like it should be?
run cpufreq-info, or if it's not installed, look at /proc/cpuinfo (the clock speed there is updated by the kernel's cpufreq framework). And cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor (ondemand is good, userspace is ok too if there is a userspace daemon checking the load and deciding what speed to set the CPU to.)

 Core 2 CPUs have a built-in temp sensor that reports a relative temp, relative to the max temp.
e.g. on my desktop
$ sensors
coretemp-isa-0000
Adapter: ISA adapter
Core 0: +59.0°C (high = +86.0°C, crit = +100.0°C)

coretemp-isa-0001
Adapter: ISA adapter
Core 1: +58.0°C (high = +86.0°C, crit = +100.0°C)

(sensors is in lm-sensors, and you have to configure it. You can just modprobe coretemp, and do
grep . -n /sys/devices/platform/coretemp.0/temp1_*
(I use grep here as an easy way to combine file names with file contents). I get:
/sys/devices/platform/coretemp.0/temp1_crit:1:100000
/sys/devices/platform/coretemp.0/temp1_crit_alarm:1:0
/sys/devices/platform/coretemp.0/temp1_input:1:59000
/sys/devices/platform/coretemp.0/temp1_label:1:Core 0
/sys/devices/platform/coretemp.0/temp1_max:1:86000
)

see bug 264290 for more info about core 2 temp sensors, and a problem that seemed like it was caused by high temps, but has gone away for me with 2.6.27-3.

[oss_test_launchpad (oss-test-launchpad)]

> Is something using a lot of CPU time?

No, in idle state both cores bounce between 5 and 20 %.

> run top, and look at the load averages at the top right. (number of tasks waiting running/waiting to run averaged over the last 5/10/15 minutes).

Gives you something like 0.22, 0.32, 0.32.

> Is your CPU idling at low speed, like it should be?

How do I know?

> run cpufreq-info, or if it's not installed,

It's not available via Synaptic - why's that?

> look at /proc/cpuinfo

How do I do that, looking at that file?

> cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor (ondemand is good, userspace is ok too if there is a userspace daemon checking the load and deciding what speed to set the CPU to.)

Is set to "ondemand".

> Core 2 CPUs have a built-in temp sensor that reports a relative temp, relative to the max temp.

> do grep . -n /sys/devices/platform/coretemp.0/temp1_*

Gives you

/sys/devices/platform/coretemp.0/temp1_crit:1:100000
/sys/devices/platform/coretemp.0/temp1_crit_alarm:1:0
/sys/devices/platform/coretemp.0/temp1_input:1:66000
/sys/devices/platform/coretemp.0/temp1_label:1:Core 0
/sys/devices/platform/coretemp.0/temp1_max:1:100000

Not that I would know what tht means.

Why doesn't someone should build a simple GUI tool for all temperature related issues which makes checking temperatures easy?

At least I managed to install the GNOME Sensors Applet, and it tells me that my HDD is on 45 degrees Celsius, which seems far too much to me. Maybe it gets overheated because the CPU heats everything up so much.

Unfortunately I didn't get the GNOME Sensors Applet to tell me more about CPU temperatures. Sorry I have to shut down the computer and let it cool for half an hour now anyway.

[oss_test_launchpad (oss-test-launchpad)]

This whole thing seems to have been a problem for quite some time now, see

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/52852
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/73019.

IMHO there really should be some progress before Ubuntu burns down some people's computers.

[oss_test_launchpad (oss-test-launchpad)]

PS A graphical tool that could be used to control the fan would be nice, too. At least as a workaround. Even though the notebook is currently close to boiling, the fan does NOTHING AT ALL (it does work, I have tested it with OSX).

[Steffen Röcker (sroecker)]

Try to load the applesmc module, it should detect the temperature sensors.
Install lm-sensors and see what sensors reports. lm-sensors should also install the fan-control script.
Do you have cpufrequtils installed?

If this works we should try to fix bug #128787 (applesmc not loaded at startup).
(I added applesmc to /etc/modules manually)

[oss_test_launchpad (oss-test-launchpad)]

> Try to load the applesmc module, it should detect the temperature sensors.

I tried around quite a bit with lm-sensors and now it works as long as I have the notebook on AC power BEFORE startup (don't know why). sudo modprobe applesmc leads to nothing (no comment in the shell).

> Try to load the applesmc module, it should detect the temperature sensors.

Now works. Fan now working and temperature now ok. This does, however, seems to have something to do with the AC power (see above).

> Do you have cpufrequtils installed?

I have tested the graphical GNOME applet which to my knowledge does the same. It doesn't seem to change anything.

What does this applesmc thing do?

[Steffen Röcker (sroecker)]

From kernel source:
/ * drivers/hwmon/applesmc.c - driver for Apple's SMC (accelerometer, temperature
 * sensors, fan control, keyboard backlight control) used in Intel-based Apple
 * computers.
 */
You need it to get the temperatures and control the fan.
You should put 'applesmc' in /etc/modules and test again, don't think it has anything to dow with AC power etc.

[Steffen Röcker (sroecker)]

Try dmesg after sudo modprobe applesmc, it should give you something like this:
[ 27.627646] applesmc: Apple MacBook detected:
[ 27.627652] applesmc: - Model with accelerometer
[ 27.627656] applesmc: - Model without light sensors and backlight
[ 27.627661] applesmc: - Model with 10 temperature sensors
[ 27.684481] applesmc: device successfully initialized (0xe0, 0x00).
[ 27.684488] applesmc: device successfully initialized.
[ 27.684700] applesmc: 1 fans found.
[ 27.685603] input: applesmc as /class/input/input11
[ 27.704156] applesmc: driver successfully loaded.

(PS: You can look at files with cat or less)

[oss_test_launchpad (oss-test-launchpad)]

> You should put 'applesmc' in /etc/modules and test again

I already asked that before: What exactly am I supposed to do? Is "modules" a file? Am I supposed to open it with an editor and copy this line into it?

[oss_test_launchpad (oss-test-launchpad)]

sudo modprobe applesmc,
dmesg

gives the following output (among others not reffering to applesmc):

----
[ 379.451963] applesmc: wait status failed: 5 != 0
[ 379.507664] applesmc: wait status failed: 5 != 0
[ 379.545090] applesmc: wait status failed: 5 != 0
[ 379.561116] applesmc: wait status failed: 5 != 0
[ 379.576191] applesmc: wait status failed: 5 != 0
[ 379.591395] applesmc: wait status failed: 5 != 0
[ 379.608716] applesmc: wait status failed: 5 != 0
----

[Steffen Röcker (sroecker)]

Yes, /etc/modules is a file. Just put a line with applesmc in it with a text editor and the module gets loaded when you boot ubuntu.
Why don't you write me an email (see my launchpad profile) so we can discuss this on IRC etc.

[oss_test_launchpad (oss-test-launchpad)]

> Why don't you write me an email (see my launchpad profile) so we can discuss this on IRC etc.

Thanks for this information, but let's keep this here so that more people profit from your information.

[oss_test_launchpad (oss-test-launchpad)]

> You should put 'applesmc' in /etc/modules and test again

Done. I will report.

[Daniel Graziotin (dgraziotin)]

I own a MacBook 3,1 [ Core(TM)2 Duo CPU T7500 @ 2.20GHz]
A fresh installlation of Alpha 6, and a dist-upgrade made yesterday.
After the load of applesmc there are no thermal problems to be reported, it works very well.

[Ali Sabil (asabil)]

I think I am experiencing the same issue with a Macbook 1st generation. There is an overheating problem that either leads to a hardlock (most cases), or a suspend from which I am unable to resume. having applesmc loaded doesn't change anything.

[oss_test_launchpad (oss-test-launchpad)]

Put 'applesmc' in /etc/modules, no change.

After some kernel updates, it worked ok; after others, it was worse again. With the most current kernel (2.6.27-7-generic) it is as bad as it was when this bug was reported.

[Ali Sabil (asabil)]

I can also confirm this with a Macbook 1st generation.

[Ali Sabil (asabil)]

Here is the output of "sensors", with both applesmc and coretemp loaded, please note the +127°C, not sure if those are accurate, but my laptop is burning.

$ sensors
coretemp-isa-0000
Adapter: ISA adapter
Core 0: +57.0°C (crit = +100.0°C)

coretemp-isa-0001
Adapter: ISA adapter
Core 1: +57.0°C (crit = +100.0°C)

applesmc-isa-0300
Adapter: ISA adapter
Exhaust : 3189 RPM (min = 1500 RPM)
temp1: +36.2°C
temp2: +61.5°C
temp3: +59.0°C
temp4: +53.8°C
temp5: +55.2°C
temp6: +55.5°C
temp7: +127.8°C
temp8: +58.5°C
temp9: +57.0°C
temp10: +43.5°C

[Oibaf (oibaf)]

See also:
http://ubuntuforums.org/showthread.php?t=969183

[shclim (shclim)]

Can anyone confirm that this bug has been fixed for the Macbook Pro v3.1 (Santa Rosa) in the 8.10 final release (i386 and amd64 versions).

[Henrik Rydberg (rydberg)]

If the machine gets hot but the fan stays at minimum, one possible cause is that the SMC has been switched to manual. If this was not done by hand, it could have to do with switching between OSX and Ubuntu via reboot. I have put this line in my /etc/rc.local to prevent any such possibilities:

## make sure to use automatic fan control by default
echo 0 > /sys/devices/platform/applesmc.768/fan1_manual

Here is a thread about it: http://ubuntuforums.org/showthread.php?t=969183

[Jamie Bainbridge (superjamie)]

Check dmesg, you may be having the same 100%-cpu-usage bug as here:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/295722

[oss_test_launchpad (oss-test-launchpad)]

Okay, just typed in "dmesg". The part relating to applesmc (what's that anyway?) is:

----
[ 16.203901] applesmc: Apple MacBook Pro 3 detected:
[ 16.203904] applesmc: - Model with accelerometer
[ 16.203905] applesmc: - Model with light sensors and backlight
[ 16.203907] applesmc: - Model with 13 temperature sensors
[ 16.260698] applesmc: device successfully initialized (0xe0, 0x00).
[ 16.260701] applesmc: device successfully initialized.
[ 16.261505] applesmc: 2 fans found.
[ 16.263814] input: applesmc as /devices/platform/applesmc.768/input/input14
----

Guys, this thread is now nearly three months old, and you call Ubuntu 8.10 "stable", and still you don't get along caring that users' notebooks won't burn down. Sorry, but this is not accpetable.

As a first step wouldn't it be a good idea if you programmed a little GUI tool that automatically turns on the fan on full speed so that users won't have to fear that Ubuntu ruins their machines? Just an idea.

As long as things are as they are, IMHO during the installation process Ubuntu should print a huge warning: "DO NOT USE ON APPLE MACBOOK PROS".

Btw. runtime on battery is about 1/3 compared to what it is on Mac OSX.

[Henrik Rydberg (rydberg)]

The rant aside, what is the current problem with temperature on your MBP3? Did you read the linked thread? It seems to me most people do not have this problem anymore,

[polycarbonate1 (polycarbonate1-deactivatedaccount)]

Hi all,

I have suffered from this since Hardy. Something I haven't seen anyone else mention yet it this:
My MB runs way hotter in Leopard than in Ubuntu. Battery life in either OS dropped from 5 hours to 2 after installing Ubuntu. There was an Apple battery firmware update released which fixes the problem in Leopard, only to be broken again (for both OSes) after rebooting to Ubuntu.

I have applesmc loaded, with full thermal readouts in my panel via sensors-applet 2.2.1.
Macbook 2,1
Linux 2.6.27-11-generic #1 SMP Thu Jan 29 19:24:39 UTC 2009 i686 GNU/Linux

[Ricky Campbell (cyberdork33)]

poly, have you tried resetting your SMC controller? It sounds more like a hardware / firmware issue.

[polycarbonate1 (polycarbonate1-deactivatedaccount)]

Ricky, have reset my SMC controller several times recently (related to a Windows VM taking out my video card then rendering my laptop un-bootable...). This is definitely a firmware issue. The problem first appeared, as I said, after installing Hardy last year. Applying this http://www.apple.com/downloads/macosx/apple/firmware_hardware/batteryupdate12.html patch fixes it under Leopard until next boot to Ubuntu, when it's broken for both OSes again. I thought the problem might be fixed when I upgraded to Intrepid so I waited it out, but to no avail.

[polycarbonate1 (polycarbonate1-deactivatedaccount)]

I should probably also add that 65ºC is fairly normal for my CPU; it's the power supply that burns my legs, frequently reaching temperatures of over 80ºC. This happens whether I'm on batteries or AC.

[Ricky Campbell (cyberdork33)]

Sorry that my suggestion was no help. The reason I did so was that there are several paramters that can be set on a Mac and will survive a reboot and even between OS X and Ubuntu (the iSight firmware loading is one). I think that maybe Ubuntu is setting something incorrectly, and then it survives into OS X. resetting the SMC should repair that though until whatever activates the problem does so again.

[polycarbonate1 (polycarbonate1-deactivatedaccount)]

Good news. I'm in OS X now after just having done this http://www.apple.com/downloads/macosx/apple/firmware_hardware/batteryupdate14.html
As well as the power supply getting extremely hot, and short battery life, I was suffering from my batteries not holding their charge properly. I always keep my spare battery fully charged, and found it dead flat the other day after having been out for a week. Also suspended my machine recently with a full battery on Friday only to find it hibernated on Monday morning because the battery was dead. I believe both issues to be related to battery firmware. Haven't actually done any tests yet without a battery; if anyone would care to do so, please report your findings. The good news is that rebooting to Ubuntu cuts battery time to 2:30, but rebooting back to Leopard, it's still 4:30. System's been up for a good 5 minutes now, and the CPU is still on at 40ºC; it would normally be getting fried by now. The CPU itself generates very little heat unless I'm rendering video, the power supply just heats everything else up.
SMC version: 1.13f3
Battery information:
Device name: ASMB013
Pack Lot Code: 0001
PCB Lot Code: 0000
Firmware Version: 0102
Hardware Revision: 0300
Cell Revision: 0200

[shclim (shclim)]

poly, I just went to the link you posted for the Apple battery update 1.4. It says it only works for Macbooks (not Macbook Pros). I just wanted to double check, did you install this on a Macbook Pro?

[polycarbonate1 (polycarbonate1-deactivatedaccount)]

Oh, sorry... I actually have a MB not a MBP. This is the wrong thread then right? Although this bug seems to apply to both.

pc1

[oss_test_launchpad (oss-test-launchpad)]

I confirm this problem for a MacBook Pro rev. 3 with Ubuntu 9.04 Alpha 5+ with all of today's updates. Question is whether this is a new bug or the old one again.

This notebook does have a fan, I don't quite get it why Ubuntu hardly ever switches it on and prefers to have the notebook burnt down.

[Ricky Campbell (cyberdork33)]

Make sure that the applesmc module is loading now by default in Jaunty. This was part of the problem in Intrepid.

[polycarbonate1 (polycarbonate1-deactivatedaccount)]

Scott, what I am experiencing is actually Bug #52852, which was marked as a duplicate of this one although I believe it is not as 53852 was reported on a MacBook not a MBP.

[Ricky Campbell (cyberdork33)]

this should apply to all Macs equally.

[Oibaf (oibaf)]

This is no longer needed on jaunty and nobody appears to care to fix the 8.10 release notes.

[oss_test_launchpad (oss-test-launchpad)]

This bug is back. Just tested Ubuntu 9.10 Alpha 2 64-Bit Live Desktop on the above mentioned MacBook Pro rev. 3. After approx. ten minutes the machine's case's temperature was so high I had to shut down the machine and let it cool down.

[Oibaf (oibaf)]

> This bug is back. Just tested Ubuntu 9.10 Alpha 2 64-Bit Live Desktop on

Just to be sure the module is loaded, can you attach the output of:

$ lsmod | grep applesmc

[Alex Murray (alexmurray)]

I have been running Karmic (alpha 4, but updated to latest) on my MBP 5,1 without any heat issues, so I can at least un-confirm this problem for Karmic - I really think it is a knee-jerk reaction to put it in the release notes - there are many happy users of Karmic on MBP (see the Apple section of ubuntuforums.org).

[Gionn (giovanni.toraldo)]

Yeah, my MBP get hot, but I usually manage to set manually my fans when needed with a script. Btw I never get into a forced standby due to the temperature. Maybe a firmware problem?

Here is the script: http://files.getdropbox.com/u/630394/macbookpro51/setfan.sh

[Νίκος Αλεξανδρής (nikos.alexandris)]

I've experienced sone strange graphics slow-down and freezes on MBP5,1. After installing powernowd the problem is gone. Does anybody experience strange graphics/ freezing system?

[Νίκος Αλεξανδρής (nikos.alexandris)]

Sorry, forgot to mention the system: under Jaunty 64-bit.

[Jim Rorie (jfrorie)]

I've see the screen turn black and as the cursor moves, the screen gets
repainted at an offset. Usually doesn't recover and requires a gnome
restart.

2009/10/5 Νίκος Αλεξανδρής <email address hidden>

> I've experienced sone strange graphics slow-down and freezes on MBP5,1.
> After installing powernowd the problem is gone. Does anybody experience
> strange graphics/ freezing system?
>
> --
> WARNING: Intrepid might burn down MacBook Pro
> https://bugs.launchpad.net/bugs/262550
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>

[Νίκος Αλεξανδρής (nikos.alexandris)]

@Jim Rorie: Yep, exactly that is what I see.

[Alex Murray (alexmurray)]

Am running Karmic beta now, but Jaunty 64-bit was rock solid on my MBP5,1.
Perhaps its a dodgy graphics card? Nvidia did have some problems with some of the 8600M GT's in previous MBP models (http://support.apple.com/kb/TS2377) perhaps it's something similar in your case?

[Νίκος Αλεξανδρής (nikos.alexandris)]

@Alex Murray:
No Alex, (a) since I installed "powernowd" I don't see this problem anymore. (b) I thought of the machine being hot and the graphics chip being damaged or so. So I just launched Mac-OS to see if there is something strange. Nothing, all works fine.

It makes me think it has something to do with frequency scaling management (cpufreqd, cpudyn).

[Oibaf (oibaf)]

I am closing this bug since the original issue was fixed. If someone still has heating problems should open a new bug, otherwise it gets difficult to manage here a different problem (this last one appears to be workarounded installing powernowd) with the same symptoms.

[Steve Langasek (vorlon)]

Please do not change bug statuses without explanation.

[Rolando Blanco (rolando)]

I am running the same MacBook but with Karmic Koala (Kernel 2.6.31-X) And
all runs fine, only have the problem with Batery Live, Tray installing

collectd-sensors -
lm-sensors
sensors-applet

And then try to see it in the GNOME Panel. (Add the Applet)

209/11/27 netsurfer <email address hidden>

> ** Changed in: linux (Ubuntu)
> Status: Fix Released => Fix Committed
>
> --
> WARNING: Intrepid might burn down MacBook Pro
> https://bugs.launchpad.net/bugs/262550
> You received this bug notification because you are a member of Mactel
> Support, which is the registrant for Mactel Support.
>
> Status in Mactel Support: Fix Released
> Status in Ubuntu Release Notes: Fix Released
> Status in “linux” package in Ubuntu: Fix Committed
> Status in “linux” source package in Intrepid: New
>
> Bug description:
> Binary package hint: linux-image-2.6.27-1-generic
>
> Testing Ubuntu 8.10 Alpha 4 64-Bit on a MacBook Pro rev. 3. Installing
> today's updates, the MacBook gets extremely hot and automatically goes into
> standby. This looks like an emergency standby caused by the EFI or
> -something because it got overheated. Unless you put a metal plate or
> something under it, the fan runs on highest level at full power, which is
> not the usual.
>
> Is there a tool you can see the CPU temperature with?
>
> _______________________________________________
> Mailing list: https://launchpad.net/~mactel-support<https://launchpad.net/%7Emactel-support>
> Post to : <email address hidden>
> Unsubscribe : https://launchpad.net/~mactel-support<https://launchpad.net/%7Emactel-support>
> More help : https://help.launchpad.net/ListHelp
>

--
Saludos Fraternos Ubunteros!
________________________________________________________________
"No pases la vida tratando de hacer trabajar a tu computador,...
  ...haz que tu computador trabaje para ti, usa GNU/LINUX"
________________________________________________________________
Rolando F. Blanco C:. (Ubuntu Member)
User: 399645 (GNU) Ubuntu: 3715
http://www.blanco.net.ve/
Ubuntu 9.04 (Jaunty)
OpenPGP keys: E1E3FC41
_______________________________________________________________
ubuntu | ウブントゥ | اُبونتو | उबुन्तु |

[Gambit- (g-launchpad)]

FYI: I added a sensor->meaning mapping to the 5,2 ubuntu Karmic page. My confidence in it's accuracy is low. Comment is being added to this bug to aid future searchers.

[cornbread (corn13read)]

macbook pro 17" works fine with ubuntu karmic.
macbook pro 13" has extreme temperature issues. Can't use it for 10 mins before it's almost too hot to have on my lap.

Installed powernowd and no fix...

[cornbread (corn13read)]

can anything be done or am I going to have to suffer with OSX?

[Oibaf (oibaf)]

> can anything be done or am I going to have to suffer with OSX?

This bug was fixed since 9.04. Please open a new bug if you still have this symptom.

[Henrik (neu242)]

Rolando: Regarding your battery life problems, this forum post might help: http://ubuntuforums.org/showthread.php?t=1215928

[Henrik Rydberg (rydberg)]

I have a theory about the cooling part of the heat problem: the SMC seems to have changed at least twice since 2005 with regard to how fan control works. The attached script will reveal which version applies to a certain machine. If there still is a bug, it should be related at least to machines MB51, MBA21, MBP51, MBP54. If you do file a new bug _please_ insert a link to that page from here.

[Alex Murray (alexmurray)]

@Rydberg: Running the script on my MBP5,1 it says:

Scanning SMC registers...
Mt register found

The machine has a Mt but no Sf register.
It is likely built 2009 or 2010.

The SMC does not have a (known) overheat protection, so the
machine might get really hot under CPU load.

DO NOT rely on the SMC to automatically regulate the temperature,
but use a fan control scheme known to work with your machine.

---

But from experience I can say the SMC seems to regulate temperature quite well and I've never had any problems with potential dangerous overheating even under load - testing by running two instances of

while true; do cat /dev/zero > /dev/null; done

to max out the CPU - once the processor hits about 95C (as gauged by temp5_input from applesmc) the fans ramp up from 2000 RPM to over 3400RPM (although be it quite slowly) and which then seems to hold the CPU temp at ~85C (although it is still dropping very slowing as fan RPM increases) - output of all temp and fan sensors atm under load:

alex@tokyo:~$ for t in /sys/devices/platform/applesmc.768/temp*_input; do echo $t: `cat $t`; done
/sys/devices/platform/applesmc.768/temp10_input: 57500
/sys/devices/platform/applesmc.768/temp11_input: 68250
/sys/devices/platform/applesmc.768/temp12_input: 73750
/sys/devices/platform/applesmc.768/temp13_input: 73500
/sys/devices/platform/applesmc.768/temp14_input: 76250
/sys/devices/platform/applesmc.768/temp15_input: 72500
/sys/devices/platform/applesmc.768/temp16_input: 57500
/sys/devices/platform/applesmc.768/temp17_input: 65500
/sys/devices/platform/applesmc.768/temp18_input: 65750
/sys/devices/platform/applesmc.768/temp19_input: 31500
/sys/devices/platform/applesmc.768/temp1_input: 34250
/sys/devices/platform/applesmc.768/temp20_input: 50250
/sys/devices/platform/applesmc.768/temp2_input: 34250
/sys/devices/platform/applesmc.768/temp3_input: 33500
/sys/devices/platform/applesmc.768/temp4_input: 34250
/sys/devices/platform/applesmc.768/temp5_input: 85000
/sys/devices/platform/applesmc.768/temp6_input: 70000
/sys/devices/platform/applesmc.768/temp7_input: 71250
/sys/devices/platform/applesmc.768/temp8_input: 83000
/sys/devices/platform/applesmc.768/temp9_input: 66000
alex@tokyo:~$ for t in /sys/devices/platform/applesmc.768/fan*_input; do echo $t: `cat $t`; done
/sys/devices/platform/applesmc.768/fan1_input: 3414
/sys/devices/platform/applesmc.768/fan2_input: 3407
alex@tokyo:~$ for t in /sys/devices/platform/applesmc.768/fan*_min; do echo $t: `cat $t`; done
/sys/devices/platform/applesmc.768/fan1_min: 2000
/sys/devices/platform/applesmc.768/fan2_min: 2000
alex@tokyo:~$ for t in /sys/devices/platform/applesmc.768/fan*_manual; do echo $t: `cat $t`; done
/sys/devices/platform/applesmc.768/fan1_manual: 0
/sys/devices/platform/applesmc.768/fan2_manual: 0

[Henrik Rydberg (rydberg)]

Thanks Alex for the input! Apparently the theory is incorrect, in as much as there is still a heat protection function in the newer SMCs. It would however be great to be able to confirm this on one of the other machines, and then preferrably by anyone who has reported fans not moving. Perhaps those experiments were never allowed to proceed beyond the 95 degC required to set them off. Or maybe the temperature limit is what changed? After all, 95 degC is a fairly high temperature :-)

[Alex Murray (alexmurray)]

I agree 95C seems quite high - i had thought it kicked in earlier than this (in real world use) before so perhaps it isnt just based on CPU temp and instead the SMC does some averaging of a number of the sensors...

[Henrik Rydberg (rydberg)]

Repeating your test on my MBA11 yields a threshold of 80 degC (largest value reading sysfs: 80000). The MBA11 belongs to the 2007-2008 range of computers with both a Mt and Sf register. Significantly different, I would say.

[polycarbonate1 (polycarbonate1-deactivatedaccount)]

Since early 2008 my MBA21 has had a CPU idle temperature of 62-67 degC, with any kind of load getting it dangerously hot (>80 degC). On a handful of occasions thermal cut-out (not sure if this is in the CPU itself or the SMC) has kicked in and halted the CPU. My SMC is at 1.14. The heat issues are apparent under Ubuntu, Debian, and Cesium, 32 and 64 bit with various kernels. Booting to Mac OS gets the fans running significantly more, although not bringing temperatures down much. I did not have any temperature issues under any OS before updating from Tiger to Leopard (which included an SMC update). Rather than the CPU, the majority of heat appears to come from the power supply area, leading me to believe that the CPU is possibly being overheated by an external heat source. Having the fans running all the time is of course detrimental to battery life, but I'm not sure that's the only reason my system constantly draws 8-13 Watts.

[Alex Valavanis (valavanisalex)]

Intrepid Ibex reached end-of-life on 30 April 2010 so I am closing the report. The bug has been fixed in newer releases of Ubuntu.

[Thiago Martins (martinx)]

Still a problem on Ubuntu 16.04 on a brand new Macbook Pro 2015.

Under OS X, supercool, under Ubuntu, super hot! Even without running anything, just boot it, login and wait, it will become hot.

This problem have what, 20 years? For God's sake! Time to fix this once for all, Ubuntu should keep a Macbook cool, just like OS X does, by default.

[Thiago Martins (martinx)]

Hey guys! Please, ignore my latest messages! After researching a little bit about this, I realized that by just installing:

macfanctld

and a few more, like:

pommed / gpomme

My Macbook is finally cooler on Ubuntu!

Thanks!
Thiago