home *** CD-ROM | disk | FTP | other *** search
- Path: aargh.incubus.sub.org!marc
- From: marc@aargh.incubus.sub.org (Marc 'Nepomuk' Heuler)
- Newsgroups: comp.sys.amiga.misc
- Subject: SHAREWARE PROGRAMMERS: Protection Patch (was Re: OS features)
- Message-ID: <bAAYx*mla@aargh.incubus.sub.org>
- Date: Thu, 18 Jan 1996 17:48:59 CET
- Reply-To: marc@aargh.incubus.sub.org
- References: <92747544038@PAPA.NORTH.DE> <4b3h9s$1st@alterdial.UU.NET> <hmAVx*Y3f@yaps.rhein.de>
- <cg.75pf@ami-cg.GraySage.Edmonton.AB.CA> <4d3t9e$mkv@hermes.louisville.edu>
- <m4d20p0mn2v.fsf@freud.hut.fi>
- Distribution: inet
- Organization: Data Design
- X-Newsreader: Arn V1.03a
-
- In article <m4d20p0mn2v.fsf@freud.hut.fi>, Janne Jalkanen writes:
-
- > This means the OS also should be more picky about the data it gets.
-
- Anyone know SCRATCH (by Bill Hawes)? It SetFunctions() all known library
- functions to scratch CPU registers D1/A0/A1.
-
- One could easily make a Protection Patch program, that enables _todays'_
- AmigaOS to carefully watch a functions input.
-
- It probably won't be reliable enough to withstand a CRASH program like on
- Unix, because as long as it is not part of the OS itself, it won't protect
- future OS functions - but anyway, I think it'll catch most wrong function
- input and should be able to help you sort out another bunch of dirty
- programs that ENFORCER didn't expose already.
-
- One could create a consistency check subfunction for every type of input
- that is used in AmigaOS. I didn't count, but I think less than 100 small
- subfunctions will suffice.
-
- Parsing the FD files enables us to automatically SetFunction() all
- functions of the OS, adding the appropriate consistency check subfunction
- for each OS function.
-
- A false function input should either reveal debug output (like ENFORCER),
- or a Requester offering CONTINUE|SUSPEND TASK|REBOOT (like Amiga GURU), or
- both - depending on user preferences.
-
- Masking out time-critical functions for speed reasons should be an option,
- too. Interrupt-proven functions must be excluded from the Requester
- option, and for future compatibility the list of those must be extenable as
- configuration option.
-
- Future OS' are partially supported: new functions that accept already
- known input types will be included to the patch & watch list when applying
- the future FD files to the parser.
-
- Now, who's willing to program this?
-