Xref: utzoo comp.unix.wizards:20891 comp.unix.questions:20413 comp.realtime:495
Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!lavaca.uh.edu!uhnix1!nuchat!shell!moray!siswat!buck
From: buck@siswat.UUCP (A. Lester Buck)
Newsgroups: comp.unix.wizards,comp.unix.questions,comp.realtime
Subject: true asynchronous I/O in Unix
Keywords: asynchronous I/O implementations RTU Ultrix realtime
Message-ID: <507@siswat.UUCP>
Date: 5 Mar 90 06:32:38 GMT
Organization: Photon Graphics,  Houston
Lines: 24

I have been asked to study how various (mainly realtime) vendors have added
asynchronous I/O to their versions of Unix.  I am interested in existing
or proposed implementations, or any literature on the subject.  I would also
appreciate receiving any personal suggestions for how you think async I/O
should look.  I know lightweight process support a la Mach will someday make
this issue moot, but I am looking for simple and modular solutions that
would be useful today.

To begin, I know some details of the following:

RTU     Uses Asynchronous System Traps as described in "Writing a
        UNIX Device Driver".  Kernel implementation of medium difficulty.

Ultrix  Uses an nbuf scheme, with ioctl interface and select.  Kernel
        implementation relatively easy.


I will summarize if there is any interest.

Thanks alot!

-- 
A. Lester Buck     buck@siswat.lonestar.org  ...!texbell!moray!siswat!buck