Re: [MiNT] Proposal for SLB extension

["Konrad M. Kokoszkiewicz" <draco@obta.uw.edu.pl>]

Hi Joerg

> To me the way Thomas and Julian are discussing (using act_pd) seems to be a
> very good aproach, as it has no/only little overhead even if there are
> libraries needing the basepage pointer

Yes, except this creates some security/stability risk, as it was already
explained in Thomas' mails. We should rather tend to reduce the number of
memory areas which are marked 'global', instead of increasing it, or doing
things so that regions currently marked global may be harder to be
converted into protected regions in the future. MiNT's mp is already way
too perforated.

By the way, I noticed some problem with MiNT's SLB implementation, which
can be potentially incompatible with TOS/MagiC libs. Namely, when
implementing tfork(), the thread startup code CANNOT be located within the
SLB, because memory violation occurs as soon as Pexec(104) executes the
code. It must be located in the program's TEXT segment to avoid that,
though it would be nicer to have it within the SLB. Proposals? Maybe
changing Pexec(106) into Pexec(104) within Slbopen() is enough? Any risk
if so?

--
Konrad M.Kokoszkiewicz
mail: draco@atari.org
http://draco.atari.org

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