kernel-packages team mailing list archive

[Renardo <1452421@xxxxxxxxxxxxxxxxxx>]

Thanks for the hint. I did that today and after a few minutes of video
streaming (www.tagesschau.de) "top" showed me the now well-known
"kidle_inject" processes but the box held out for a bit over ten
minutes, which is more but not significantly more than before. Then it
switched off. I booted the 3.16 kernel, reloaded the same video stream
(it is available for two hours), and it played the full 15 minutes
without any kidle_injects appearing. And without powering off the
machine, of course.

This time "top" (under 3.19) never showed me any processes (firefox,
plugin-container) going anywhere near 100% of CPU load; everything
remained under 50% most of the time, more or less as with the old
kernel. And still those kidle_inject processes come up. Could it be a
false alarm (despite the very warm air coming from the fan slots)? I
know that kidle_inject generates no CPU load, just blocks other
processes from taking a CPU. Could it still have influence on the
temperature sensor and the fan?

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1452421

Title:
  ubuntu 15.04 overheats

Status in linux package in Ubuntu:
  Incomplete

Bug description:
  On 2015-04-26 I upgraded to Ubuntu 15.04 (from 14.10). Since then, my
  notebook (HP envy dv6) fan is nearly always on, and the notebook
  switched off several times, probably due to overheating. Each time I
  was working with Firefox; watching the "livestream" of
  www.tagesschau.de seems to be a good way of reproducing the problem.

  Going back to the 3.16.0-31 kernel (still offered by grub) seems to
  circumvent the problem. When I boot 3.16 the fan runs only
  occasionally. My first 15.04 kernel was 3.19.0-15, today I got -16 but
  the problem is the same.

  Before the machine switches off (it does not shut down Linux but just
  goes out) "top" shows me eight instances of "kidle_inject", obviously
  desperately but hopelessly trying to keep processor load down. Under
  the 3.19 kernel the "plugin-container" process goes to loads of more
  than 100 % (on a four-core box), under 3.16 it never goes beyond 60 %.
  Note that between the "overheat" and the "normal" scenario I change
  only the kernel in the boot menu, all other binaries are the same
  (15.04). Especially, it's the same firefox.

  I fear that within a couple of weeks more kernel versions will arrive
  at my system and eventually push the working 3.16 kernel out of my
  grub menu, and then I shall have no way left of running my system for
  more than 20 or 30 minutes.

  ProblemType: Bug
  DistroRelease: Ubuntu 15.04
  Package: kernel-common (not installed)
  ProcVersionSignature: Ubuntu 3.16.0-34.47-generic 3.16.7-ckt8
  Uname: Linux 3.16.0-34-generic x86_64
  ApportVersion: 2.17.2-0ubuntu1
  Architecture: amd64
  CurrentDesktop: Unity
  Date: Wed May  6 20:57:39 2015
  InstallationDate: Installed on 2014-06-09 (331 days ago)
  InstallationMedia: Ubuntu 14.04 LTS "Trusty Tahr" - Release amd64 (20140417)
  SourcePackage: kernel-package
  UpgradeStatus: Upgraded to vivid on 2015-04-24 (12 days ago)

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1452421/+subscriptions