Xref: utzoo comp.lang.c:32222 comp.unix.programmer:101
Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!virtech!cpcahil
From: cpcahil@virtech.uucp (Conor P. Cahill)
Newsgroups: comp.lang.c,comp.unix.programmer
Subject: Re: shared memory
Message-ID: <1990Sep26.230857.29812@virtech.uucp>
Date: 26 Sep 90 23:08:57 GMT
References: <13991@hydra.gatech.EDU> <14140@hydra.gatech.EDU>
Reply-To: cpcahil@virtech.UUCP (Conor P. Cahill)
Organization: Virtual Technologies Inc., Sterling VA
Lines: 23

In article <14140@hydra.gatech.EDU> gt0178a@prism.gatech.EDU (BURNS,JIM) writes:
>HP-UX documents it; it tests out on A/UX 1.1 and Ultrix 4.0 : the first
>call to shmat() (by any process) automatically zeroes out the shm seg, so
>explicit initialization is NOT necessary. Is this standard? Which one?

This is a basic rule of unix: 

	All memory given to a program by the kernel must be cleared to 
	zero.  This applies to new sbrk()'d areas (not always to areas that
	you previously had and then released and then got back) uninitialized
	static data areas, and shared memory segments. It even applies to stack
	the first time it is used.

This is a security measure (makes sure you can't see someone elses data).

Note that it also applies to disk "memory" (if you seek past the end of a
file, write a byte, seek back to the skipped portion and read, you will get
all nulls).

-- 
Conor P. Cahill            (703)430-9247        Virtual Technologies, Inc.,
uunet!virtech!cpcahil                           46030 Manekin Plaza, Suite 160
                                                Sterling, VA 22170