[lkml]   [2010]   [Dec]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    Subject[RFC][RT][PATCH 0/4] rtmutex: Simplify PI code
    This is still a little buggy. I've been hitting a hang when running
    dbench 100 in a loop. But I also hit that hang with vanilla -rt.

    But I'm sure this still has some things that need to be straighten out.
    Anyway, you can either apply the following patches or pull the code
    from the listed git repo.

    This still needs to be worked out, but since I'm taking off for the
    rest of the year, I might as well show you where I ended up at.

    The following patches are in:


    branch: rt/tip/rt/devel

    Lai Jiangshan (1):
    rtmutex: Ensure only the top waiter or higher priority task can take the lock

    Steven Rostedt (3):
    rtmutex: Only save lock depth once in spin_slowlock
    rtmutex: Try to take lock early in rt_spin_lock_slowlock()
    rtmutex: Revert Optimize rt lock wakeup

    kernel/futex.c | 22 +--
    kernel/rt.c | 10 +-
    kernel/rtmutex-debug.c | 1 -
    kernel/rtmutex.c | 458 ++++++++++++++++++-----------------------------
    kernel/rtmutex_common.h | 18 +--
    5 files changed, 188 insertions(+), 321 deletions(-)

     \ /
      Last update: 2010-12-23 23:53    [W:0.020 / U:4.808 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site