[lkml]   [2011]   [Feb]   [8]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    SubjectRe: [PATCH 01/12] export kernel call get_task_comm()
    > On Tue, Feb 08, 2011 at 11:44:16AM -0800,
    > wrote:
    >> That is originally what I did, but after talking with Greg KH, Arjan
    >> vdv,
    >> and Alan Cox I made this patch instead.
    > Care to explain why exactly?

    If I can jog my own memory as this was a while ago...

    Well for starters, Arjan van de Ven first suggested it, and I trust his
    judgement because of his Linux kernel experience.

    But looking at the code, the kernel function get_task_comm() provides a
    level of safety with a lock, so the global resource 'current' can be used
    without someone else trying to use it at the same time. And using this
    function facilitates code reuse which is something Greg KH was trying to
    point out to me at the time:

    In fact, I remember finding existing, accepted drivers in the kernel that
    wrote their own locks before accessing current (I quickly just now found
    block/dfdb/drbd_main.c as an example), so Greg is making a valid point
    about code reuse. So not exporting get_task_comm() encourages code
    locking re-write of every driver that wants to access 'current'.

    In that email, Greg clearly reminded me my goal is to get my work into the
    kernel. Thus, I listened to their advice and made this patch.

     \ /
      Last update: 2011-02-08 21:37    [W:0.020 / U:5.904 seconds]
    ©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site