RE: [MiNT] Security again

[Jo-Even.Skarstein@gjensidige.no]

> -----Original Message-----
> From: Joerg_Westheide@su.maus.de [mailto:Joerg_Westheide@su.maus.de]
> Sent: Wednesday, December 08, 1999 2:00 AM
> To: mint@fishpool.com
> Subject: [MiNT] Security again
> 
JE>your MiNT-application (including TSRs), then there's nothing that
> JE>prevents you from using TraPatch or something similar.
> The only argument against TraPatch that I ever heard:
> I do not want to rely on the presence of another program and 
> I also don't
> want to urge someone to install another program in the auto folder

Well, is it any better to make it dependant on MiNT? Atleast TraPatch will
work with TOS and MagiC as well.

> JE>What concerns me is stability, and I can't see how that can be
> JE>improved by promoting vector-bending.
> I do not consider it promoting vector bending (as I don't see 
> Setexc() doing

Setexc() is a relic from a time where vector-bending was the only way to
extend the OS.

> so). If someone thinks he has to bend a vector, he will do 
> it, if not, he
> won't.

That's true. But he's probably not programming for MiNT in the first place.

> JE>Instead of trying to find new ways to bend vectors, let's 
> try to find
> JE>ways to eliminate it. Here's a few suggestions:
> Sorry, but your suggestions are only solutions for some special TSRs.
> There's missing a concept how you can get rid of vector bending

I'm not quite sure what you mean. Is there something that you can't
implement as a XFS or XDD today? Remember, you can bend vectors as much as
you like inside the kernel, so basically anything can be implemented as a
XDD. TraPatch-like functionality would be nice in this case though (for
efficiency), but it should only be available through the XDD/XFS-interface.

> JE>3. Let the various AES-developers agree on a common way to hook a
> JE>fileselector into the AES.
> Not only file selectors hook into the GEM trap. Have a look 
> at my homepage
> and you will find 3 of them (and I have some more here). None 
> of them is a
> file selector

That's why I suggested that a common dispatcher for the GEM-trap would be a
good thing. But this shouldn't be of any concern to MiNT, the correct place
for this should IMO be in the VDI.

Jo Even Skarstein