Memory protection for lower 16K

[Tony van der Hoff]

On 13 Feb 2002, in message <60c02e84b%Marco at dionin.esw-heim.tu-clausthal.de>,
Marco Baye <Marco.Baye at tu-clausthal.de> wrote:

> Hi!
> 
> Yesterday I learned something that may be of interest for all who
> program in C under RISC OS.
> 
> As you will know, the lower 16K of memory are a system area, and no
> user program should mess with it (and the other 16K before the
> application area starts are "scratch space". Whatever).
> 
> Anyway, the OS does not use memory protection in these lower 16K, so
> if a C program uses an uninitialised or NULL pointer for write
> access, it is likely to corrupt something valuable.
> 
Personally, I don't think that's correct. the amount of protection depends a
bit on the OS version, but I'm pretty sure RO 4 does protect the system
workspace.

[snip]
> 
>     char* tmpnam(char*);
> 
> accesses memory location &FE8 in user mode.

Yes, RO 4 knows about this, and works round it. The 32-bit scl actually fixes
the problem.

I may be wrong on both counts, though, and I think this subject is best aired
on csap; your audience here is really too small.

A Google at csa* for "tmpnam" reveals quite a number of hits; some of them
relevant.

[snip]

-- 
Tony van der Hoff         | MailTo:tony at mk-net.demon.co.uk
                          | MailTo:avanderhoff at iee.org
Buckinghamshire, England  | http:www.mk-net.demon.co.uk