lkml.org 
[lkml]   [2009]   [Nov]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: [tip:timers/urgent] hrtimer: Fix /proc/timer_list regression
    On Thu, 19 Nov 2009, Ingo Molnar wrote:
    > i've bisected back to it:
    >
    > | 887a29f59b93cf54e21814869a4ab6e80b6fa623 is the first bad commit
    > | commit 887a29f59b93cf54e21814869a4ab6e80b6fa623
    > | Author: Feng Tang <feng.tang@intel.com>
    > | Date: Thu Sep 3 16:32:53 2009 +0800
    > |
    > | hrtimer: Fix /proc/timer_list regression
    >
    > Config attached.
    >
    > I've removed it from timers/urgent for now.

    Come on, this is a patently wrong conclusion.

    We call timer_stats_update_stats() which returns on top of the
    function due to:

    if (likely(!timer_stats_active))
    return;

    timer_stats_active is 0 during boot and you can only activate it by
    writing to /proc/timer_stats which you certainly did not at this
    point.

    Can you please explain how a call to a function which returns right
    away can cause that problem ?

    That patch unearthed some other bug and your revert is just papering
    over that fact.

    Thanks,

    tglx


    \
     
     \ /
      Last update: 2009-11-19 11:11    [W:0.034 / U:31.176 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site