[v2,0/3] Documentation: Explain EAS and EM

Message

Quentin Perret Jan. 21, 2019, 11:17 a.m. UTC

The recently introduced Energy Aware Scheduling (EAS) feature relies on
a large set of concepts, assumptions, and design choices that are
probably not obvious for an outsider. Moreover, enabling EAS on a
particular platform isn't straightforward because of all its
dependencies. This series tries to address this by introducing proper
documentation files for the scheduler's part of EAS and for the newly
introduced Energy Model (EM) framework. These are meant to explain not
only the design choices of EAS but also to list its dependencies in a
human-readable location.

Changes in v2:
 - Fixed typos and style in sched-energy.txt (Juri)
 - Moved EM doc under Documentation/driver-api/pm/ (Rafael)
 - Translated EM doc into .rst (Rafael)
 - Fixed EM kerneldoc comments to avoid htmldoc build errors

Quentin Perret (3):
  PM / EM: Fix broken kerneldoc
  PM / EM: Document the Energy Model framework
  sched: Document Energy Aware Scheduling

 Documentation/driver-api/pm/energy-model.rst | 150 +++++++
 Documentation/driver-api/pm/index.rst        |   1 +
 Documentation/scheduler/sched-energy.txt     | 431 +++++++++++++++++++
 include/linux/energy_model.h                 |   4 +-
 kernel/power/energy_model.c                  |   2 +-
 5 files changed, 585 insertions(+), 3 deletions(-)
 create mode 100644 Documentation/driver-api/pm/energy-model.rst
 create mode 100644 Documentation/scheduler/sched-energy.txt

Comments

Jonathan Corbet Feb. 7, 2019, 12:32 a.m. UTC | #1

On Mon, 21 Jan 2019 11:17:21 +0000
Quentin Perret <[e-mail]> wrote:

> The recently introduced Energy Aware Scheduling (EAS) feature relies on
> a large set of concepts, assumptions, and design choices that are
> probably not obvious for an outsider. Moreover, enabling EAS on a
> particular platform isn't straightforward because of all its
> dependencies. This series tries to address this by introducing proper
> documentation files for the scheduler's part of EAS and for the newly
> introduced Energy Model (EM) framework. These are meant to explain not
> only the design choices of EAS but also to list its dependencies in a
> human-readable location.

So these have been sitting in my folder waiting for acks or some other
sort of discussion, but it's been awfully quiet.  Should I take them, or
do they need further work or ... ?

Thanks,

jon

Quentin Perret Feb. 7, 2019, 8:11 a.m. UTC | #2

Hi Jon,

> So these have been sitting in my folder waiting for acks or some other
> sort of discussion, but it's been awfully quiet.  Should I take them, or
> do they need further work or ... ?

These (well, the v1 actually) have been picked recently up by Ingo
in tip/sched/core.

Thanks,
Quentun