From: ijackson@nyx.cs.du.edu (Ian Jackson) Subject: Draconian memory protection (Re: **** Protection ****) Date: Thu, 17 Jun 1993 00:01:43 GMT
In article <1vmqio$aqu@irz401.inf.tu-dresden.de> ss10@irz.inf.tu-dresden.de (Sebastian Schoenberg) writes:
<snip - `near' pointers ! :-)>
>For instance,
> char *c,*ptr = (char *)malloc (2048);
>
>if I write now this
>
> c = p+2049;
> *c = 123;
>
>will a protection-violation follows or not ?
As someone has already noted, talk of `near pointers' is an anathema
to proper operating systems (and indeed proper CPU architectures). [*]
The answer to your question is: probably not, unless you're lucky.
You might however be interested in a package called `electric fence',
posted recently to csm or csu [*2]. If anyone out there has
compiled/ported it already please speak up - otherwise I may find time
to have a look at it myself.
[*] No, I don't like Intel's processors. Yes, I bought one.
Architecture rants by email please. That's the last thing col needs!
[*2] csm = comp.sources.misc; csu = comp.sources.unix
Both of these newsgroups are invaluable - I suggest anyone interested
in running the latest free software who is competent to configure and
build them read these groups. Oh - you'll need a fast modem :-).
Csm and csu - like all the other comp.sources groups - are archived.
The archives are available at all major FTP sites.
PS: Please try not to *SHOUT* in subject lines - many people, myself
included, have a strong tendency to ignore postings with subjects in
capitals or containing lots of asterisks, because they're often stupid
questions from people who don't know what they're doing.
-- Ian Jackson <ijackson@nyx.cs.du.edu> (non urgent email only - see below) home: 35 Molewood Close, Cambridge, CB4 3SR, England; phone: +44 223 327029 work & urgent email: <iwj@cam-orl.co.uk> Olivetti Research Ltd; +44 223 343398 PGP2 public key on request; fingerprint = 5906F687 BD03ACAD 0D8E602E FCF37657