Bug 770470

Summary: radeon default power_profile "default" makes laptop overheat (Mobility Radeon HD 3650)
Product: Red Hat Enterprise Linux 6 Reporter: Pasi Karkkainen <pasik>
Component: xorg-x11-drv-atiAssignee: Jérôme Glisse <jglisse>
Status: CLOSED CURRENTRELEASE QA Contact: Desktop QE <desktop-qa-list>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 6.8CC: collura, lyude, pasik, tpelka
Target Milestone: rcKeywords: OtherQA
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2015-10-13 21:06:21 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
dmesg with drm.debug=0x2
none
HP Elitebook 8530p laptop ATI Mobility Radeon HD3650 VBIOS ROM none

Description Pasi Karkkainen 2011-12-26 21:40:16 UTC
Description of problem:

I'm using HP EliteBook 8530P laptop with Mobility Radeon HD 3650, and it's overheating with RHEL 6.x. I've tested with RHEL 6.0 and RHEL 6.1, and now I tested also with CentOS 6.2 liveCD (2.6.32-220.el6.x86_64).

This problem also happens with Fedora 14, Fedora 15 and Fedora 16.

As a default radeon power_profile "default" is being used, and it causes laptop
overheating and emergency thermal shutdowns.. on a completely idle system! It
seems the power management doesn't work with the opensource radeon driver on
this laptop. See below for analysis and more info.

When I manually switch the radeon power_profile to "low" (see below) the
temperature of the laptop decreases over 10-20 degrees celsius, making the laptop much cooler and also the fans will immediately slow down because the
temperature goes down. Any idea why the "default" power_profile doesn't
automatically run the card at lower engine/memory clock speeds?

More information from CentOS 6.2 liveCD (x64) below.. notice how the
temperature goes down from 84C-89C to 72C when manually switching to power_profile "low".


Version-Release number of selected component (if applicable):
RHEL 6.2, 2.6.32-220.el6.x86_64.

How reproducible:
Always.

Steps to Reproduce:
1. Install RHEL 6.2 on HP EliteBook 8530P with the default Gnome desktop.
2. Notice how the laptop runs very hot and fans are running at high speed all the time, even when the system is idle, only running the empty desktop.
3. Manually switch to power_profile "low" and the temperature immediately goes down and fans slow down.
  
Actual results:
With the default settings laptop overheats and does emergency thermal shutdowns.

Expected results:
Laptop runs cool and works normally.

Additional info:

Upstream bug:
"radeon default power_profile "default" makes laptop overheat (Mobility Radeon HD 3650)":
https://bugs.freedesktop.org/show_bug.cgi?id=41762


$ uname -a
Linux livecd.centos 2.6.32-220.el6.x86_64 #1 SMP Tue Dec 6 19:48:22 GMT 2011 x86_64 x86_64 x86_64 GNU/Linux

$ cat /sys/class/drm/card0/device/power_profile 
default

$ cat /sys/devices/virtual/thermal/thermal_zone3/device/path 
\_TZ_.CPUZ

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
83000

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
84000

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
89000


# echo low > /sys/class/drm/card0/device/power_profile 


$ cat /sys/class/drm/card0/device/power_profile 
low

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
81000

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
80000

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
79000

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
77000

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
76000

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
75000

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
74000

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
73000

$ cat /sys/devices/virtual/thermal/thermal_zone3/temp
72000

Comment 1 Pasi Karkkainen 2011-12-26 21:41:04 UTC
Created attachment 549625 [details]
dmesg with drm.debug=0x2

Comment 3 Pasi Karkkainen 2012-01-29 19:42:24 UTC
Btw this issue is not HP/Elitebook specific. The same overheating problem happens also with other laptop brands/models. There are multiple problem reports also with Apple Macbooks and Lenovo Thinkpads.

Comment 4 Jérôme Glisse 2012-02-06 16:04:58 UTC
We would need hw. In the mean time can you provide your video bios.

(change 01:05.0 by whatever is correct for you gpu lspci will tell you)
cd /sys/bus/pci/devices/0000:01:05.0
sudo sh -c "echo 1 > rom"
sudo sh -c "cat rom > ~/bios.rom"

Comment 6 Pasi Karkkainen 2012-02-13 09:14:25 UTC
Sorry for the delay. Here's the vbios rom.

Comment 7 Pasi Karkkainen 2012-02-13 09:15:16 UTC
Created attachment 561472 [details]
HP Elitebook 8530p laptop ATI Mobility Radeon HD3650 VBIOS ROM

Comment 8 Suzanne Logcher 2012-02-14 23:25:40 UTC
This request was evaluated by Red Hat Product Management for
inclusion in the current release of Red Hat Enterprise Linux.
Because the affected component is not scheduled to be updated
in the current release, Red Hat is unfortunately unable to
address this request at this time. Red Hat invites you to
ask your support representative to propose this request, if
appropriate and relevant, in the next release of Red Hat
Enterprise Linux. If you would like it considered as an
exception in the current release, please ask your support
representative.

Comment 9 Jérôme Glisse 2014-03-19 21:59:33 UTC
Moving to 6.6 so it has a chance of being retested there and close consequently

Comment 10 Tomas Pelka 2014-03-20 06:46:17 UTC
Hi Pasi,

would be able to get a commitment from you to provide test feedback once test package available?

Thanks
Tom

Comment 11 Pasi Karkkainen 2014-03-20 09:18:17 UTC
Hmm, I haven't used that laptop for around two years now, but I think I'll be able to still find it :) so yeah, I can test..

Comment 12 Jérôme Glisse 2014-10-30 19:41:08 UTC
So rhel 6.6 should fix this issue as we backported upstream power management. This need to be tested so we can close the bug.

Comment 15 Lyude 2015-10-13 21:06:21 UTC
When this bug was filed we didn't have power management enabled by default. We do now, so this shouldn't be an issue.
(Sources: Jérôme Glisse)