Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!mcsun!ukc!cam-cl!news
From: bdb@cl.cam.ac.uk (Brian Brunswick)
Newsgroups: comp.sys.acorn
Subject: RISCOS Tiny Source: Write Protect low memory
Message-ID: <1991Feb28.232432.9008@cl.cam.ac.uk>
Date: 28 Feb 91 23:24:32 GMT
Reply-To: bdb@cl.cam.ac.uk (Brian Brunswick)
Organization: U of Cambridge Comp Lab, UK
Lines: 36

The other day I was trying to debug something with a pointer bug in
it, and it was crashing the system totally each time it ran. I whined
about the lack of proper memory protection in RiscOs, and my friend
suggested that the memory below 32K was only not protected due to
ancient compatibility - most of the system would work with it write
protected for user mode. So I whipped up a quick proglet to do this,
and he was indeed correct - enough of the system works, certainly
sufficiently to easily find my bug from the backtrace due to the
abort.

Here follows the few lines of basic needed.
--- cut ---
REM>PROT0
REM By Brian.Brunswick@uk.ac.cam.cl
SYS"OS_ReadMemMapInfo" TO pagesize,pages
DIM p (pages+1)*12
FOR I=0 TO pages-1
p!(I*12)=I
NEXT
p!(pages*12)=-1
SYS"OS_ReadMemMapEntries",p
FOR I=0 TO pages-1
IF p!(I*12+4)=0 THEN p!(I*12+8)=1
NEXT
SYS"OS_SetMemMapEntries",p
--- cut ---

I have since found that printing seems to die with this thing in
though.  And I would suspect that suspicious applications might well
do too.  It might be interesting to build up a list of things grungy
enough to die, as indications of nasty hackery at work.

Now if only we could protect the RMA space too, those system crashes
might be a thing of the past....

Brian.Brunswick@uk.ac.cam.cl  Disclaimer.  Short sig rules!