Path: utzoo!utgpu!watmath!maytag!water!cgeisler
From: cgeisler@water.waterloo.edu (Craig Eisler)
Newsgroups: comp.sys.ibm.pc
Subject: Fork and process creation
Message-ID: <2361@water.waterloo.edu>
Date: 25 May 89 18:43:49 GMT
References: <2968@cps3xx.UUCP> <75800069@p.cs.uiuc.edu> <2332@water.waterloo.edu> <593@rna.UUCP>
Reply-To: cgeisler@water.waterloo.edu (Craig Eisler)
Organization: U. of Waterloo, Ontario
Lines: 44

In article <593@rna.UUCP> kc@rna.UUCP (Kaare Christian) writes:
>"An initial Send", whatever that is, must dupicate I/O connections,
	     ^^^^
>signal handling and other aspects of the original process' state,

All of these things have no place in a process creation primitive.
Create() is the type of primitive the kernel needs to spawn a
new process.  Duplication of I/O connections is a job for a file
server, not the kernel.

>Forks make it easy to have pipes (cheapo, flexible ipc)
						    ^^^
Pipes are a nice feature, but they are anything but cheap.
They are file I/O.  Too damn slow.  Since you mention IPC, I
assume you know what Send,Receive, and Reply are.  If you want 
cheap communication, that is what you use.  And you do not need
fork to implement pipes. 

>Unix is still the only system that allows most programs to work together
>easily, using ad hoc connections. Fork is an essential part of that
>flexibility.

I disagree.  Fork is an unnecessary primitive.  Of course, the UNIX
kernel is filled with unnecessary primitives (270 or something similar).

>And now a few numbers. How fast is "fast enough" for fork? 
>[times omitted]

At a cost of 35ms for a fork, you can forget about using it in any
kind of real time environment.

In an operating system I implemented on a intel 186, process creation
was 2.8 ms.  And this was barely good enough for arbitrary process
creation while trying run a system of trains in real time.

Currently I am involved with the Multi-processor Systems Group here
at the university, and we are designing a firmware (microcoded) kernel
to run 4 68030's.  There will be no fork, only process creation.
All other features of fork have no reason for being in the kernel.

craig
-- 
Craig Eisler, Applied Math, University of Waterloo
"Either way, I'm afraid to try it" - Calvin.