Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!samsung!dali.cs.montana.edu!fuquad!simon
From: simon@fuquad.westford.ccur.com (Simon Rosenthal)
Newsgroups: comp.unix.internals
Subject: Re: Quest: Sockets in SYSVR3
Message-ID: <62156@fuquad.westford.ccur.com>
Date: 20 Jun 91 20:07:30 GMT
References: <1991Jun17.190413.29256@sctc.com> <1991Jun18.130539.3579@sctc.com> <1991Jun19.184132.29431@noao.edu>
Reply-To: simon@westford.ccur.com (Simon Rosenthal)
Distribution: comp.unix.internals
Organization: Concurrent Computer Corp
Lines: 27

In article <1991Jun19.184132.29431@noao.edu> rstevens@noao.edu (Rich Stevens) writes:
>> However, SYSVR4 does have sockets based on streams
>> and I am curious how it maintains sockets for dup
>> and fork operations.
> --
>In summary, the paper says that it takes both a user library
>(socklib) and a streams module (sockmod) to provide sockets
>over streams.
>
>	Rich Stevens  (rstevens@noao.edu)

We are just finishing up a sockets-over-SVR4 Streams implementation.
For binary compatibility reasons, we chose to maintain all existing
socket system calls, and move the socket-to-sockmod "glue" that would
be in a user library such as the one referenced above into the kernel.
No problems with forks, dups, or fcntl operations.

The ATT SVR4 'Programmer's Guide : Networking Interfaces' (published by
Prentice-Hall) has a chapter on moving sockets applications to SVR4 and
the differences between Berkeley and SVR4 sockets implementations.

- Simon
_______________________________________________________________________________
Simon Rosenthal:                         ___________
Concurrent Computer Corporation         / _________/_ 
Westford, MA 01886                     /_/________/ /
Internet: simon@westford.ccur.com       /__________/