Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!columbia!rutgers!ll-xn!mit-eddie!genrad!decvax!decwrl!sun!guy
From: guy@sun.UUCP
Newsgroups: net.unix-wizards
Subject: Re: Writing on shared texts
Message-ID: <8576@sun.uucp>
Date: Mon, 27-Oct-86 13:53:57 EST
Article-I.D.: sun.8576
Posted: Mon Oct 27 13:53:57 1986
Date-Received: Mon, 27-Oct-86 23:20:23 EST
References: <115@tijc02.UUCP> <735@hropus.UUCP> <1040@ho95e.UUCP> <1269@ttrdc.UUCP> <8545@sun.uucp> <14976@onfcanim.UUCP>
Organization: Sun Microsystems, Inc.
Lines: 23

> To do this, just zero the inode pointer in the text table,

After having made *damn* sure you've caught all the code that assumes that
"x_iptr" is never null for a valid text table entry, and changed "xalloc"
not to use a null "x_iptr" as an indication of a free text table entry.
(You definitely have to do this with a non-paging S5R2; as I remember, that
code hadn't changed much since V7.)

> In the case of a system that doesn't load the entire binary at once
> (e.g. demand paging), it is much messier.  There is still no problem
> in providing a clean copy of the new code for new execs, contrary to
> what Guy suggests - just "detach" the text structure as before.

Gee, I wasn't aware that I suggested that you can't do it even with the
changes you mention.  My point was that *with the V7/4BSD/unpaged S5R2
implementation* there is a reason for forbidding writes to active shared
text files.  Another possibility is to do what's done with NFS and kill the
process if it tries to fetch a page from the file and the file has been
written since the last time you looked at it.
-- 
	Guy Harris
	{ihnp4, decvax, seismo, decwrl, ...}!sun!guy
	guy@sun.com (or guy@sun.arpa)