[lkml]   [2009]   [Sep]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    Subject[v5 RFC PATCH 0/7]: cpuidle/x86/POWER (REDESIGN): Cleanup idle power management code in x86, cleanup drivers/cpuidle/cpuidle.c and introduce cpuidle to POWER.

    ******** This is an RFC, not for inclusion **********

    This patchset introduces cpuidle infrastructure to POWER, prototyping
    for pSeries, and also does a major refactoring of current x86 idle
    power management and a cleanup of cpuidle infrastructure.

    My earlier iterations can be found at:

    v4 -->
    v3 -->
    v2 -->
    v1 -->

    Major Changes in this iteration:

    * Refactoring x86 idle power management code
    Remove all instances of pm_idle and make cpuidle_idle_call as
    _the_ idle routine in x86. So, cpuidle_idle_call will be
    called from the main idle loop, cpu_idle instead of the
    function pointer pm_idle. Also, pm_idle was used by apm module
    and xen. Change those instances such that they register to
    cpuidle instead.

    * Cleanup drivers/cpuidle/cpuidle.c
    Currently, the cpuidle implementation has weakness in the
    framework where an exported pm_idle function pointer is
    manipulated by various subsystem. The proposed framework has
    a registration mechanism to cleanly add and remove new idle
    routines from different subsystems.

    * Implement cpuidle for pSeries
    Implement the processor_idle module for pseries, which
    registers idle loops to cpuidle and also cleanup
    arch/powerpc/platforms/pseries/setup.c and remove the
    redundant pseries_dedicated/shared_idle_sleep which is
    implemented in processor_idle.c
    Also, remove all instances of ppc_md.power_save, for the same
    reason as that given for pm_idle.


    * Currently, the list based approach that I'm using here is not
    working in a clean way on a few x86 platforms which have multiple
    sleep states, leading to kernel panics. So working on resolving

    * ppc_md.power_save has been replaced by cpuidle_idle_call only for
    pseries. So this needs to be done for all POWER platforms so that
    ppc_md.power_save is completely removed.

    Patches included in this series:

    1/7 - cleanup drivers/cpuidle/cpuidle.c
    2/7 - implement a list based approach to register a set of idle
    3/7 - refactor x86 idle power management code and remove all instances
    of pm_idle.
    4/7 - enable cpuidle for POWER.
    5/7 - remove dedicate/shared idle loops, which will be moved to
    6/7 - add a default_idle idle loop for POWER.
    7/7 - implement pSeries processor idle module.

    Any comments on the design is welcome.


     \ /
      Last update: 2009-09-22 07:37    [W:0.037 / U:61.500 seconds]
    ©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site