On Thursday 26 August 2004 02:36 am, [EMAIL PROTECTED] wrote:
> Now "tsleep(&ident,PRIBIO,NULL,1)" solved the problem. Maybe yield() can't
> do the job in kernel. And mi_switch() is not enough to.Thank you.On
yield() can just switch back to you immediately since you (the current thread)
are still runnable. Use a sleep, either via sleep/wakeup, or doing a tsleep
with a timeout forces the current thread to not be runnable for a while
giving the other thread time to run. Note that your tsleep(..., 1) is still
racey, you really need to use sleep/wakeup to truly close the race.
> Wednesday 25 August 2004 03:06 am, yangshazhou at hotmail.com wrote:
> > Hi all,
> > This problem delay me for quite a long time.
> > I've built two klds, I'd like to see the result one's running to the
> > other's. In one kld's MOD_LOAD subroutine, I make a long 'for' loop, in
> > which it DELAY(1000) and then call yield(curthread,NULL). When the module
> > was loaded by kldload, the system failed to response until the loop quit.
> > The loop seems like that:
> > for(i=0;i<8000;i++){
> > ......
> > DELAY(1000);
> > yield(curthread,NULL);
> > }
> >
> > My system is 5.2.1-release. Thanks.
>
> You could try having the first one wait on a condition variable that the
> second module's MOD_LOAD does a wakeup on.
--
John Baldwin <[EMAIL PROTECTED]> <>< http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve" = http://www.FreeBSD.org
_______________________________________________
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
To unsubscribe, send any mail to "[EMAIL PROTECTED]"
