lkml.org 
[lkml]   [2008]   [Feb]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: [PATCH 2/8] Add interface for queuing work on a specific CPU
    Sorry for delay,

    On 02/07, Andrew Morton wrote:
    >
    > On Thu, 7 Feb 2008 10:18:59 +0100 Jens Axboe <jens.axboe@oracle.com> wrote:
    >
    > > --- a/kernel/workqueue.c
    > > +++ b/kernel/workqueue.c
    > > @@ -175,6 +175,21 @@ int fastcall queue_work(struct workqueue_struct *wq, struct work_struct *work)
    > > }
    > > EXPORT_SYMBOL_GPL(queue_work);
    > >
    > > +int fastcall queue_work_on_cpu(struct workqueue_struct *wq,
    > > + struct work_struct *work, int cpu)
    > > +{
    > > + int ret = 0;
    > > +
    > > + if (!test_and_set_bit(WORK_STRUCT_PENDING, work_data_bits(work))) {
    > > + BUG_ON(!list_empty(&work->entry));
    > > + __queue_work(per_cpu_ptr(wq->cpu_wq, cpu), work);
    > > + ret = 1;
    > > + }
    > > +
    > > + return ret;
    > > +}
    > > +EXPORT_SYMBOL_GPL(queue_work_on_cpu);
    >
    > Might as well change queue_work() to call this?

    This is possible, but in that case queue_work_on_cpu() should use wq_per_cpu(),
    not per_cpu_ptr(). (otherwise queue_work(single_threaded_wq) won't work).


    A bit off-topic, the comment near queue_work() says

    * We queue the work to the CPU it was submitted, but there is no
    * guarantee that it will be processed by that CPU.

    This is wrong. Unless cpu_down() happens, we do guarantee it will be processed
    by that CPU. Perhaps it makes sense to fix the comment as well?

    Oleg.



    \
     
     \ /
      Last update: 2008-02-11 11:55    [W:0.024 / U:33.800 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site