Skip to content
Snippets Groups Projects

Thinkpad z13 with Ryzen 7 6850u "Balanced" profile will not set, only "Power Saver" and "Performance"

    • Closed Issue created by Britt Yazel @brittyazel

    I'm not sure when this started happening, but recently on my Thinkpad z13 with the Ryzen 7 6850u the "Balanced" power profile stopped working. It used to be that on battery the "Power Saver" profile would be used, but then when plugged in would automatically switch to "Balanced". However, now it stays in "Power Saver" regardless of being plugged in or not. Likewise, when I try to manually set the power profile from within GNOME, I can successfully get "Performance" and "Power Saver" to work reliably. While "Balanced" does show in my list, selecting it immediately reset to "Power Saver".

    Using doesn't give me any sort of error, it just doesn't work.

    Designs

      Child items ...

      2. Activity

      • Bastien Nocera
        Bastien Nocera @hadess ·
        Owner

        Please follow the README to gather debug information.

      • Britt Yazel
        Britt Yazel @brittyazel ·
        Author

        Here is the debug information when starting in "performance" mode and then manually switching to "Balanced"

         britt  ~  sudo G_MESSAGES_DEBUG=all /usr/lib/power-profiles-daemon -r -v                                           1
** (power-profiles-daemon:66359): DEBUG: 08:45:42.099: Starting power-profiles-daemon version 0.12
(power-profiles-daemon:66359): GLib-GIO-DEBUG: 08:45:42.100: Using cross-namespace EXTERNAL authentication (this will deadlock if server is GDBus < 2.73.3)
** (power-profiles-daemon:66359): DEBUG: 08:45:42.102: Handling driver 'fake'
** (power-profiles-daemon:66359): DEBUG: 08:45:42.102: probe() failed for driver fake, skipping
** (power-profiles-daemon:66359): DEBUG: 08:45:42.102: Handling driver 'platform_profile'
(power-profiles-daemon:66359): GLib-GIO-DEBUG: 08:45:42.103: _g_io_module_get_default: Found default implementation local (GLocalVfs) for ‘gio-vfs’
** (power-profiles-daemon:66359): DEBUG: 08:45:42.107: ACPI performance_profile is now p, so profile is detected as performance
** (power-profiles-daemon:66359): DEBUG: 08:45:42.107: Found a dytc_lapmode sysfs attribute to thinkpad_acpi
** (power-profiles-daemon:66359): DEBUG: 08:45:42.107: Handling driver 'intel_pstate'
** (power-profiles-daemon:66359): DEBUG: 08:45:42.107: Driver 'platform_profile' already probed, skipping driver 'intel_pstate'
** (power-profiles-daemon:66359): DEBUG: 08:45:42.107: Handling driver 'placeholder'
** (power-profiles-daemon:66359): DEBUG: 08:45:42.107: Driver 'platform_profile' already probed, skipping driver 'placeholder'
** (power-profiles-daemon:66359): DEBUG: 08:45:42.107: Handling action 'trickle_charge'
** (power-profiles-daemon:66359): DEBUG: 08:45:42.107: Applying profile 'performance' from configuration file
** (power-profiles-daemon:66359): DEBUG: 08:45:42.107: Setting active profile 'performance' for reason 'reset' (current: 'performance')
** (power-profiles-daemon:66359): DEBUG: 08:45:42.107: Can't switch to performance mode, already there
** (power-profiles-daemon:66359): DEBUG: 08:46:17.205: Transitioning active profile from 'performance' to 'balanced' by user request
** (power-profiles-daemon:66359): DEBUG: 08:46:17.205: Setting active profile 'balanced' for reason 'user' (current: 'performance')
** (power-profiles-daemon:66359): DEBUG: 08:46:17.205: Writing 'balanced' to '/sys/firmware/acpi/platform_profile'
** (power-profiles-daemon:66359): DEBUG: 08:46:17.283: Successfully switched to profile balanced
** (power-profiles-daemon:66359): DEBUG: 08:46:17.289: /sys/firmware/acpi/platform_profile changed
** (power-profiles-daemon:66359): DEBUG: 08:46:17.290: ACPI performance_profile is now l, so profile is detected as power-saver
** (power-profiles-daemon:66359): DEBUG: 08:46:17.290: Driver 'platform_profile' switched internally to profile 'power-saver' (current: 'balanced')
** (power-profiles-daemon:66359): DEBUG: 08:46:17.290: Setting active profile 'power-saver' for reason 'internal' (current: 'balanced')
** (power-profiles-daemon:66359): DEBUG: 08:46:17.290: Can't switch to power-saver mode, already there
** (power-profiles-daemon:66359): DEBUG: 08:46:17.296: /sys/firmware/acpi/platform_profile changed
** (power-profiles-daemon:66359): DEBUG: 08:46:17.296: ACPI performance_profile is now l, so profile is detected as power-saver
** (power-profiles-daemon:66359): DEBUG: 08:46:17.296: /sys/firmware/acpi/platform_profile changed
** (power-profiles-daemon:66359): DEBUG: 08:46:17.296: ACPI performance_profile is now l, so profile is detected as power-saver
** (power-profiles-daemon:66359): DEBUG: 08:46:17.296: /sys/firmware/acpi/platform_profile changed
** (power-profiles-daemon:66359): DEBUG: 08:46:17.296: ACPI performance_profile is now l, so profile is detected as power-saver
** (power-profiles-daemon:66359): DEBUG: 08:46:19.287: /sys/firmware/acpi/platform_profile changed
** (power-profiles-daemon:66359): DEBUG: 08:46:19.287: ACPI performance_profile is now l, so profile is detected as power-saver
      • SpaceRaven
        SpaceRaven @drolyk ·

        I can confirm the same problem on Z13 gen1 with AMD Ryzen™ 7 PRO 6860Z

      • Bastien Nocera
        Bastien Nocera @hadess ·
        Owner

        Have you tried updating to the latest firmware on your machines? If so, which versions of the firmwares and kernel did you reproduce those problems with?

        @markpearson, any idea on those ones?

      • Britt Yazel
        Britt Yazel @brittyazel ·
        Author

        Yep, fwupd has been great for keeping my firmware up to date. Lenovo System Firmware: 0.1.24 Lenovo Embedded Controller: 0.1.53

        I have been aware of the issue since at least firmware version 0.1.21 (it could have happened earlier, I just wasn't aware of it until this revision). I believe my laptop first shipped with 0.1.13.

        Kernel is Archlinux's version 6.0.11

        Likewise, I tested with Archlinux's Linux-LTS kernel and had the same issue (5.15.82)

        • Mark Pearson
          Mark Pearson @mpearson ·

          Oh...that is interesting. I can reproduce this on my system and I'm slightly embarrassed as this is my 'evening' daily driver (surfing/netflix/etc) and I hadn't noticed. I'm finding on mine I can go between performance and low-power; but not balanced.

          There was a problem where the profile mode wasn't being retained over suspend and resume; that was fixed - but I think is a newly introduced problem :disappointed:

          I've created internal ticket LO-2191. Pretty sure this is a Lenovo FW problem but I'll confirm with the FW team what is going on.

      • hexchain
        hexchain @hexchain ·

        I roughly recalled that this only started to happen after 6.0, and just verified that by using a 5.19.9 kernel and the power profile could be changed to "balanced" and stayed that way.

        Not sure about 5.15 LTS though. Maybe something got backported and broke it?

      • Mark Pearson
        Mark Pearson @mpearson ·

        I think I've gotten to the root cause of the issue - I've posted a longer post on our forums here: https://forums.lenovo.com/topic/findpost/15028/5169096/5856412

        Summary: FW enabled another thermal mode (somewhat unexpectedly), which Linux tries to use. Unfortunately there are some bugs and it means balanced mode gets broken.

        My plan is

        • patch to fix this in the short term by disabling the new mode, and get that backported to stable kernels.
        • Fix the problems with the new mode so that works - expecting that to give better performance and battery.

        I'm leaning towards closing this issue as it's not a ppd issue - but Linux kernel. Are people OK using the above linked Lenovo forum thread for getting updates?

        Mark

      • Mike Lothian mentioned in commit FireBurn/linux@fde5f74c

        mentioned in commit FireBurn/linux@fde5f74c

      • Mario Limonciello
        Mario Limonciello @superm1 ·
        Maintainer

        Mark's patch is in 6.2-rc4 and 6.1.7. I think this issue should be closed.

        Edited by Mario Limonciello
      • Britt Yazel
        Britt Yazel @brittyazel ·
        Author

        Thank you @mpearson for your help on this one. I can't wait for things to get back to normal, finally :-)

        Closing, since this is effectively fixed.

        Edited by Britt Yazel
      • Britt Yazel closed

        closed

      • Iyan
        Iyan @iyanmv ·

        @superm1 well, maybe the patch fixes things for @brittyazel but it broke the switch for me :see_no_evil:

        https://bugzilla.kernel.org/show_bug.cgi?id=216963

      • Mark Pearson
        Mark Pearson @mpearson ·

        Ah crud...and that's my patch :disappointed:

        I'll check it on my system...as this got back-ported to all the stable kernels this is going to blow up on me. Thanks for letting me know.

        If you're able to do any debug and figure out what's going please let me know. It should be going into the MMC mode path for the Intel platforms and nothing should have changed there. I'm wondering what stupid thing I've done...

        Mark

      • Mario Limonciello
        Mario Limonciello @superm1 ·
        Maintainer

        As a "simple" solution, you can try to "mask" the AMT capability when the CPU is Intel. Then the whole switch/case should keep working otherwise.

      Please register or sign in to reply