Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.kernel > #65742 > unrolled thread

Bug#945618: regression: 5.3 and above kernels in Debian are not allowing Intel processor to enter package c states

Back to article view | Back to linux.debian.kernel

Contents

  Bug#945618: regression: 5.3 and above kernels in Debian are not allowing Intel processor to enter package c states thomasw@fastmail.cn - 2019-11-28 00:50 +0100
    Bug#945618:  thomasw@fastmail.cn - 2019-11-29 03:00 +0100
      Bug#945618:  Ben Hutchings <ben@decadent.org.uk> - 2019-11-29 22:20 +0100
    Processed: Re: Bug#945618: "Debian Bug Tracking System" <owner@bugs.debian.org> - 2019-11-29 22:20 +0100

#65742 — Bug#945618: regression: 5.3 and above kernels in Debian are not allowing Intel processor to enter package c states

package: src:linux
1. run powertop --auto-tune to ensure settings are optimal for power saving. This may require a laptop.
2. Run powertop, press tab, and notice that the values in pc2 through pc10 are not incrementing.
Reproduced on Broadwell and Whiskey Lake. The first version of 5.3 posted to experimental was fine but this regressed after that and it is still a problem in 5.4 in experimental. Also, things work fine in Arch kernel so its probably something with the Debian config. On a side note, if you want to test in Skylake or newer, you need the non-free firmware package for the Intel DMC firmware which allows the graphics controller to enter a low power state. I used to see pc8 on one of my machines and pc7 on the other   as the lowest power states incrementing in Debian. These still increment in Arch with the latest Kernels. This bug results in a 5 degree c higher idle temperature on the Whiskey Lake machine.

[toc] | [next] | [standalone]

#65749 — Bug#945618:

I have to appologize for an error I made when reporting this. It can actually be reproduced in both Arch and Debian. The difference was that I didn't have pulseaudio installed on my Arch system. I am blind and use Orca. If I have Orca playing through pulse, the pc states are not entered in both Arch and Debian with the newer kernels. If pulse is on but Orca is disabled, I can enter the pc states. I am not sure if all sounds playing through pulse cause the pc states to not be entered or if this is specifically an Orca/speech-dispatcher/espeak-ng problem. I had sighted assistance reading powertop with Orca disabled and could only test this one case with no sounds playing (sighted help is limited in my current living situation). The information in my first message about the kernels where this occurs and those in which this does not occur is still accurate. I don't think this matters but I will include that I use the Mate desktop.

[toc] | [prev] | [next] | [standalone]

#65767 — Bug#945618:

[Multipart message — attachments visible in raw view] — view raw

Control: reassign -1 orca

I'm reassigning this to orca, in case there is something that can be
done there to allow power saving when it's not actively speaking.

Ben.

On Thu, 2019-11-28 at 20:54 -0500, thomasw@fastmail.cn wrote:
> I have to appologize for an error I made when reporting this. It can
> actually be reproduced in both Arch and Debian. The difference was
> that I didn't have pulseaudio installed on my Arch system. I am blind
> and use Orca. If I have Orca playing through pulse, the pc states are
> not entered in both Arch and Debian with the newer kernels. If pulse
> is on but Orca is disabled, I can enter the pc states. I am not sure
> if all sounds playing through pulse cause the pc states to not be
> entered or if this is specifically an Orca/speech-dispatcher/espeak-
> ng problem. I had sighted assistance reading powertop with Orca
> disabled and could only test this one case with no sounds playing
> (sighted help is limited in my current living situation). The
> information in my first message about the kernels where this occurs
> and those in which this does not occur is still accurate. I don't
> think this matters but I will include that I use the Mate desktop.
> 
-- 
Ben Hutchings
Every program is either trivial or else contains at least one bug

[toc] | [prev] | [next] | [standalone]

#65766 — Processed: Re: Bug#945618:

Processing control commands:

> reassign -1 orca
Bug #945618 [src:linux] regression: 5.3 and above kernels in Debian are not allowing Intel processor to enter package c states
Bug reassigned from package 'src:linux' to 'orca'.
Ignoring request to alter found versions of bug #945618 to the same values previously set
Ignoring request to alter fixed versions of bug #945618 to the same values previously set

-- 
945618: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=945618
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.kernel

csiph-web