[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MiNT] memory protection on Afterburner04



Hi!

On Sat, May 01, 1999 at 02:40:28PM +0000, Jo Even Skarstein wrote:
> I don't know how desirable this is. Remember, all TSR's are loaded into
> this area, and in many cases they install cookies with pointers to
> structures in their own memory-space. This means that any program that
> access these structures in user-mode (most do, just think about your own
> Clocky...) will be killed.

Exactly. For example, NVDI is /requiered/ to be launched before MiNT to
work properly. For MiNT to be able to protect all memory, it would have
guess which parts if it has to have which protection mode - that's
impossible. Unfortunately, this currently also means that the complete
kernel code is unprotected, too!

A solution for the latter would be to write a tool that loads MiNT to an
8K-aligned address, relocates and runs it (with some special parameter
to tell it it should protect its code). Any volunteers?

For the other problem, I currently see no solution, but maybe somebody
else has a brilliant idea?

> I also noticed this but I wonder if this can have anything to do with the
> fact that I usually enable the copyback-cache after MiNT. Will the toolkit-
> -driver work as usual with memory-protection enabled? If it doesn't, then I'm
> running without copyback which means 20% slowdown...

After MiNT with active memory protection is run, each process gets its
own MMU table, so I don't think enabling the cb-cache will work, then,
as it for sure won't alter all MMU tables (remember, the type of caching
is either set via the MMU or the transparent translation registers,
where the latter are checked by MiNT to not cover the memory it wants to
protect).


Ciao

Thomas


-- 
Thomas Binder (Gryf @ IRCNet)  gryf@hrzpub.tu-darmstadt.de
PGP-key available on request!  binder@rbg.informatik.tu-darmstadt.de

Attachment: pgpyIaWMn1oua.pgp
Description: PGP signature