Re: Setexc question

[Konrad Kokoszkiewicz <draco@mi.com.pl>]

Hi,

> I didn't have time to check... But now that a process is able to set
> it's termination vector, haven't we introduced a backdoor to get into
> supervisor mode? In that case I would propose to run the code in
> user mode...

Yes, possibly. And all such stuff should be slowly fixed step by step. But
anyways, some 3/4 of the system calls is freely available for all users
(almost all the XBIOS, all the VDI and all the AES, if loaded) and it is
prolly impossible to fix all this stuff in one cast. The backdoor you
pointed out is worth chacking and fixing now, because it is a hole in the
new (slowly being born) protection system :-)

BTW. I think that MiNT should catch all the unassigned traps. Any
objections? Now it is easy to paint bombs on the console (even via telnet
:-)) trap #3 draws 34 bombs AFAIR. :-)

> Generelly (Super and Supexec): wouldn't it be a good idea not to
> terminate a process, but let it continue to run in user mode? If it then
> tries to do something it needs super priviliges for, it will generate
> some sort of signal which could then be catched and handled...

Well, it may be for Super() (the call would be just ignored in that 
case and the program would probably die soon due to privilege violation
or bus error). But Supexec() calls user subroutine (supposed to be ended
by RTS), so how? Call this in user mode? (just thinking).

Gtx,

--
Konrad M.Kokoszkiewicz
|mail: draco@mi.com.pl                  | Atari Falcon030/TT030/65XE |
|http://www.orient.uw.edu.pl/~conradus/ | ** FreeMiNT development ** |

** Ea natura multitudinis est,
** aut servit humiliter, aut superbe dominatur (Liv. XXIV,25)
*************************************************************
** U pospolstwa normalne jest, ze albo sluzy ono unizenie,
** albo bezczelnie sie panoszy.