On Mon, Aug 16, 2010 at 8:49 PM, Rik van Riel <[email protected]> wrote: > On 08/16/2010 03:48 AM, Parmenides wrote: > >> if (!current_thread_info->preempt_count&& !irqs_disabled()) { >> current_thread_info->preempt_count = PREEMPT_ACTIVE; >> schedule(); >> current_thread_info->preempt_count = 0; >> } >> >> The if statement checks whether irqs are enabled. If so, the >> preemption will be carried out. I wonder why this check is necessary. >> Why preemption is prohibited with irqs disabled? > > Because the thread you are scheduling to, won't know that > it needs to reenable interrupts. > > If interrupts stay disabled "forever", the CPU will not > get any more interrupts, like interrupts that tell it to > reschedule, move the clock forward, etc...
Yes. And one more important point here is that the current thread may have intentionally disabled the interrupts to disallow pre-emption because it may be in some critical section and may want uninterrupted access to some shared resource. Hence, no other thread should be given control in this case. > > -- > To unsubscribe from this list: send an email with > "unsubscribe kernelnewbies" to [email protected] > Please read the FAQ at http://kernelnewbies.org/FAQ > > -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to [email protected] Please read the FAQ at http://kernelnewbies.org/FAQ
