On Wed, 23 Aug 2000, Michael Rothwell wrote:
> Ingo Molnar wrote:
> > we should unexport it. Wakeups should be done via waitqueues
>
> Do those work on 2.2? [...]
are you asking wether there are waitqueues in the 2.2? Yes, of course,
waitqueues were around and were the primary process-wakeup mechanism since
around 1992 or so.
> > only the 'in-kernel' kernel threads like bdflush
> > (which never exit) should use it.
>
> Exactly the use we're putting it to.
my point is that you lose future optimizations (already in my kernel) done
at the wake_up() level.
> We do not have any rmmod crash problems with it. We are doing our own
> locking to prevent that.
well, if you do your own locking then yes. With waitqueues you need no
extra locking as far as thread exits go, and they are more generic.
wake_up_process() is used only a few very specific and kernel-internal
cases.
Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Aug 23 2000 - 21:00:09 EST