[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [MiNT] Problem
>> new = (BASEPAGE *)Pexec(7L, (void *)flags, "", 0L);
> I haven't found any doc on Pexec(7).
Pexec(7) is just the same as Pexec(5), except that the second argument are
prgflags. It was first implemented in TOS 3.0, and MiNT has it too.
> Maybe programmer/man/man2/Pexec.2 should be updated.
Yes, it should :-)
>> static short child_global;
> The use of GEMlib seems good.
So we have a problem, why it doesn't work.
> BTW, just a stupid question: is the static array child_global part
> of the child memory ? Or is it owned by the parent only ?
This is a part of the parent BSS. However, Pexec(104) guarantees that the
child and parent share one common TEXT/DATA/BSS (the child is a "thread" of
the parent), and only the basepage/stack are separate.
I did some experiment, and changed the child code into something simpler:
static short childarg;
Cconws("Child: started up\r\n");
ca = childarg;
Cconws("Child: argument OK\r\n");
Now, THIS DOES'NT WORK EITHER. So the problem does not seem to be
AES-related at all. This is what appears on the screen:
Child: started up
Child: argument OK
pid 58 (child): MEMORY VIOLATION: type=private RW=w AA=867FBE PC=85428A
or (yes, results are different):
Child: started up
pid 213 (child): MEMORY VIOLATION: type=private RW=w AA=867FD2 PC=854270
Notice that the access address (AA) that causes the bus error, is
$17FE/$1FD2 (6142/8146) bytes bigger, than the child's basepage address
(BP). Since the child's basepage is shrunk earlier to $2000 (8192) bytes by
the parent, then it seems that the fault occurs, when the child accesses
it's own stack.
If I don't overlook anything, then it seems that something is seriously
Konrad M.Kokoszkiewicz, http://draco.atari.org
** Ea natura multitudinis est, aut seruit humiliter, aut superbe dominatur.
** Taka to już natura pospólstwa, albo służalczo się płaszczy,
** albo bezczelnie się panoszy. (T. Liuius XXIV, 25).