On Tue, Sep 2, 2008 at 9:39 PM, Rik van Riel <[EMAIL PROTECTED]> wrote:
> On Tue, 2 Sep 2008 11:40:59 +0530
> "Venky Shankar" <[EMAIL PROTECTED]> wrote:
>
> > up(&semaphore);
> > ---
> > ---
> > ---
> > prepare_to_wait(queue, &wait, TASK_INTERRUPTIBLE);
> > if(condition())
> > schedule();
> > finish_wait(queue, &wait);
> > ---
> > ---
> > ---
> > down_interruptible(&semaphore);
>
> > 1. The above code can get preempted just after the call to
> > prepare_to_wait(), In that case it would not be run again by the
> scheduler
> > since it
> > is in the TASK_INTERRUPTIBLE state. After that when someone wake's it
> up,
> > then the check is made (if condition) and then the normal flow of code.
>
> This is correct.
>
> > 2. If the code does not get preempted after the call to
> prepare_to_wait(),
> > it will still execute the if condition while in TASK_INTERRUPTIBLE state
> -->
> > but a TASK_INTERRUPTIBLE state is not runnable - i feel this is valid
> after
> > the call to schedule() is made. (i.e. it would not run again after it has
> > given up the CPU)
>
> The task continues to run after it has set its state to TASK_INTERRUPTIBLE.
> Once it enters schedule() it will be taken off the runqueue and stop
> being runnable.
>
> This is ok, because by now it is on a wait queue and eventually it
> should get woken up.
>> Thank you very much.
>
>
> So yes, you understand this code correctly :)
>> I am happy :)
>
>
> --
> All rights reversed.
>
