debcrafters-packages team mailing list archive

[Kai-Chuan Hsieh <2115041@xxxxxxxxxxxxxxxxxx>]

@superm1 @ahasenack

Hello,

I try to update the description by referencing @superm1's comment. Could
you help to review if it is good to proceed. If it is ok, I'll try to
align with certification team to add the test job into kernel SRU test
plan to ensure the test coverage. I think orcaular/plucky/questing can
be covered as well.

Thanks,

** Description changed:

  [ Impact ]
  
-  * The thermal management policy may impacts cooling device operation
- like fan, for example, it can keep the fan speed high when user select
- performance in case a throttling happened. Without the change, it won't
- change the fan mode, so that the fan speed might be limited for not
- being so noisy, however, the real outcome depends on the platform design
- (some platforms may not have fan). In conclusion, the new AMD platform
- is less likely being throttled with the change.
+  * The kernel 6.14+ support new string 'custom' platform profile to
+ accommodate multiple driver to support different positions.
  
- The mapping relationship power mode and platform profiles:
- performance <-> performance
- balanced <-> balanced/balanced_performance/cool
- power saver <-> low-power/quiet
+  * On Dell's platform have a BIOS option that configures the user
+ selectable thermal tables which are read by the dell-pc driver. It's
+ possible to boot up with the two drivers having different values and
+ then the power slider being in an undefined position
+ 
+  * PPD had to be modified to account for this new string. If PPD isn't
+ modified then the system will be in an undefined power state.
  
  [ Test Plan ]
  
-  * On AMD Strix Point/Krackan Point platform which has sysfs node
- /sys/firmware/acpi/platform_profile
+  * Checkbox job com.canonical.certification::power-
+ management/switch_power_mode can verify if all platform supported
+ profiles can be set and read correctly.
  
-  * Try to modify power mode by $ sudo powerprofilectl set <power-mode>
+  *
+ https://github.com/canonical/checkbox/blob/main/providers/base/units/power-
+ management/jobs.pxu#L119
  
-  * Check if /sys/firmware/acpi/platform_profile being updated
- 
-  * After reboot, check if BIOS update the ThermalManagement setting by $
- fwupdmgr get-bios-setting ThermalManagement
+  * Add the job into kernel SRU plan to ensure the test coverage
  
  [ Where problems could occur ]
  
   * The platform will list supported platform_profile in
  /sys/firmware/acpi/platform_profile_choices, the DPP power mode should
  map to one of it, otherwise, it can't select the correct thermal policy.
  If the platform set the unexpected string which the mapping is not
  defined in the PPD. The thermal policy can't be switched successfully.
  
  [ Other Info ]
  
   * Upstream commit: https://gitlab.freedesktop.org/upower/power-
  profiles-daemon/-/commit/6b565b0b05a2bc92e3c546a567b394f415937265
  
   * The patch is in 25.04 and 25.10 already.
  
   * To the kernel without platform_profile node [1], the user experience
  is just as before.
  
  [1] https://docs.kernel.org/userspace-api/sysfs-platform_profile.html

-- 
You received this bug notification because you are a member of
Debcrafters packages, which is subscribed to power-profiles-daemon in
Ubuntu.
https://bugs.launchpad.net/bugs/2115041

Title:
  Can't recognize custom ACPI platform profile

Status in OEM Priority Project:
  New
Status in power-profiles-daemon package in Ubuntu:
  In Progress

Bug description:
  [ Impact ]

   * The kernel 6.14+ support new string 'custom' platform profile to
  accommodate multiple driver to support different positions.

   * On Dell's platform have a BIOS option that configures the user
  selectable thermal tables which are read by the dell-pc driver. It's
  possible to boot up with the two drivers having different values and
  then the power slider being in an undefined position

   * PPD had to be modified to account for this new string. If PPD isn't
  modified then the system will be in an undefined power state.

  [ Test Plan ]

   * Checkbox job com.canonical.certification::power-
  management/switch_power_mode can verify if all platform supported
  profiles can be set and read correctly.

   *
  https://github.com/canonical/checkbox/blob/main/providers/base/units/power-
  management/jobs.pxu#L119

   * Add the job into kernel SRU plan to ensure the test coverage

  [ Where problems could occur ]

   * The platform will list supported platform_profile in
  /sys/firmware/acpi/platform_profile_choices, the PPD power mode should
  map to one of it, otherwise, it can't select the correct thermal
  policy. If the platform set the unexpected string which the mapping is
  not defined in the PPD. The thermal policy can't be switched
  successfully.

  [ Other Info ]

   * Upstream commit: https://gitlab.freedesktop.org/upower/power-
  profiles-daemon/-/commit/6b565b0b05a2bc92e3c546a567b394f415937265

   * The patch is in 25.04 and 25.10 already.

   * To the kernel without platform_profile node [1], the user
  experience is just as before.

  [1] https://docs.kernel.org/userspace-api/sysfs-platform_profile.html

To manage notifications about this bug go to:
https://bugs.launchpad.net/oem-priority/+bug/2115041/+subscriptions