tasklet_kill: bug or feature?

From: Juergen Quade (quade@hsnr.de)
Date: Sat Dec 14 2002 - 15:16:37 EST


I had a closer look to the tasklet implementation, especially
to tasklet_kill. I do not understand, why at the end of
the function the TASKLET_STATE_SCHED bit is cleared. Is it
a bug or a feature? Maybe someone can help me.

        The function tasklet_schedule does two things:
        1. it sets the schedule_bit (TASKLET_STATE_SCHED)
        2. if the bit wasn't set (cleared) it puts the tasklet on the
           tasklet_vec list.
        By the trick, to set the schedule_bit without putting
        the tasklet on the list, a tasklet can be "killed"
        (it is not on the list and tasklet_schedule is not going
        to put it on the list). But in the current implementation
        of tasklet_kill the schedule_bit is cleared (last line of
        the function)? In this case the next tasklet_schedule
        "works" and the tasklet-function will be called again.
        So, is it a bug? What do we achieve by clearing the
        schedule_bit? Can we remove the "clear_bit" line
        (it is not necessary to call "set_bit", because
        "test_and_set_bit" has set it before)?
        Or what is _exactly_ the purpose of tasklet_kill?


void tasklet_kill(struct tasklet_struct *t)
        if (in_interrupt())
                printk("Attempt to kill tasklet from interrupt\n");

        while (test_and_set_bit(TASKLET_STATE_SCHED, &t->state)) {
                while (test_bit(TASKLET_STATE_SCHED, &t->state));
        clear_bit(TASKLET_STATE_SCHED, &t->state);
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

This archive was generated by hypermail 2b29 : Sun Dec 15 2002 - 22:00:31 EST