powertop incorrectly reports multiple display brightness devices to use power

Bug #949221 reported by Christian Mertes
40
This bug affects 9 people
Affects Status Importance Assigned to Milestone
powertop (Ubuntu)
Confirmed
Undecided
Unassigned

Bug Description

I am running PowerTOP 1.97 beta 1 on Lubuntu Precise. There are three reported devices called "Display backlight" on my Samsung N220 with a Pixel Qi display. I made a couple of measurements to figure out what these devices might be:

Power est. Usage Device name
pt_bright
  1.13 W 100.0% Display backlight
  1.13 W 100.0% Display backlight
  183 mW 28.6% Display backlight
pt_dark
  1.13 W 100.0% Display backlight
  220 mW 28.6% Display backlight
  220 mW 0.0% Display backlight
pt_off
  1.15 W 100.0% Display backlight
  1.15 W 100.0% Display backlight
  212 mW 28.6% Display backlight
pt_off2
  1.14 W 100.0% Display backlight
  211 mW 28.6% Display backlight
  211 mW 0.0% Display backlight
pt_displayoff
    0 mW 100.0% Display backlight
    0 mW 28.6% Display backlight
    0 mW 0.0% Display backlight

bright is taken with 100% display brightness, dark with the minimum the brightness keys in Lubuntu can do (0d according to setpci on its 00-ff scale). off is measured with the backlight completely off, set with setpci -s 00:02.0 F4.B=00; off2 is a second measurement under the same conditions. displayoff was done with xset dpms force off.

So it appears to me that the first "Display backlight" is the display itself, the last one is the actual display backlight, although is still consumes 0,2W even when switched off. The second device seems to behave totally erratic, showing the measurement of one of the other two devices at random. This doesn't completely hold true when running with AC power but then there seem to be so many other issues that it almost doesn't seem worth mentioning.

I couldn't figure out any such pattern for the usage percentage though.

Any help on how to dig deeper into this is appreciated. Also, any hints how to really switch off the backlight would be great. Other than that, powertop should show only existing devices and name them correctly. Should I also report this to power [at] bughost org?

Revision history for this message
Christian Mertes (cmertes) wrote :

One small addendum: just saw that the second device actually sometimes shows a different measurement. Something like 1.17W say when the first device is 1.25. In these cases it always seems to be a bit below the first device. When it mimics the third device, then it always seems to be spot on. When adding all reported devices to the reported baseline, the second 1W device is also needed to come close to the reported discharge rate.

Revision history for this message
Launchpad Janitor (janitor) wrote :

Status changed to 'Confirmed' because the bug affects multiple users.

Changed in powertop (Ubuntu):
status: New → Confirmed
Revision history for this message
madbiologist (me-again) wrote :

Does Ubuntu 12.10 "Quantal Quetzal" alpha 1 simplify things?

It's 3.4-based kernel contains the follwing commit (taken from the upstream kernel changelog):

author Corentin Chary <email address hidden>
Thu, 22 Mar 2012 13:08:19 +0000 (14:08 +0100)
committer Matthew Garrett <email address hidden>
Mon, 26 Mar 2012 19:45:25 +0000 (15:45 -0400)
commit a979e2e2af7d5b4bb3b20f6a716c627bb23a6753
tree 02e944b7a63f86947205a9a389a9737cdb2b3099
parent 5719b81988f3c24ff694dc3a37e35b35630a3966

samsung-laptop: unregister ACPI video module for some well known laptops

On these laptops, the ACPI video is not functional, and very unlikely
to be fixed by the vendor. Note that intel_backlight works for some
of these laptops, and the backlight from samsung-laptop always work.

The good news is that newer laptops have functional ACPI video device
and won't end up growing this list.

Signed-off-by: Corentin Chary
Signed-off-by: Matthew Garrett

tags: added: precise
Revision history for this message
Karol Szkudlarek (karol-mikronika) wrote :

Any chances to port it to Precise 12.04?

Revision history for this message
madbiologist (me-again) wrote :

That's up to the Ubuntu kernel team (I'm not a member) but the first step is to know if it helps. Do you see any improvement in Qantal beta 2?

Revision history for this message
Haw Loeung (hloeung) wrote :

I'm also seeing this in Raring alpha 2:

System baseline power is estimated at 4.05 W

Power est. Usage Device name
  1.44 W 6.6% Display backlight
  1.34 W 0.0% Display backlight
  294 mW 2.7% CPU use

powertop:
  Installed: 2.1-0ubuntu1
  Candidate: 2.1-0ubuntu1
  Version table:
 *** 2.1-0ubuntu1 0
        500 http://archive.ubuntu.com/ubuntu/ raring/main amd64 Packages
        100 /var/lib/dpkg/status

Revision history for this message
Nicholas Camp (nicksnicksnicks) wrote :

Me too:

213 mW 73,3% Display backlight
195 mW 51,7% Display backlight

I'm using Manjaro (Arch Linux) x64, Lenovo G480.

Revision history for this message
madbiologist (me-again) wrote :

@Nicholas Camp - You have non-Samsung hardware so you should file a separate bug report.

tags: added: trusty
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.