[ On Saturday, December 7, 2002 at 01:38:57 (+0100), Jaromir Dolecek wrote: ]
> Subject: Re: CVS commit: syssrc/sys/kern
>
> FWIW, I fail to see where anyone explained anything like what you are
> saying. So far, I've seen bad example with thttpd. Generalizing on
> basis of thttpd behaviour isn't correct, IMHO.

You keep calling thttpd's behaviour "bad" -- au contraire!

Thttpd's behaviour is exaclty right -- it was designed properly to work
_with_ the system and its resource controls, not against it.  Once you
understand why this is then perhaps you'll understand why the forced
sleep hack is extremely very bad.

As Perry says, and I agree wholeheartedly, processes must not ever block
unless they ask to, period.  (with the mandatory exception to the rule of
course being when they've exhausted their current CPU time slot and are
required to give it up to the next deserving process -- welcome to
timesharing)

-- 
								Greg A. Woods

+1 416 218-0098;            <g.a.woods@ieee.org>;           <woods@robohack.ca>
Planix, Inc. <woods@planix.com>; VE3TCP; Secrets of the Weird <woods@weird.com>