Keven writes:

> When I trap system traps with user trap procedure set with XSYSTRAP,
> there are 4 integers passed in (int32). The 1st integer is named TRAPCODE.
...
> I get value 1003.

I armed XSYSTRAP and triggered it with 3 kinds of traps:

  1) call to ACTIVATE (and my process didn't have PH capability (hi Chris! :))

      Systrap caught: parameters = $000003eb, $00000068, $00000002, $00000000

  2) call to GETPRIVMODE (and my process didn't have PH capability (hi Chris! :))

      Systrap caught: parameters = $000003eb, $000000c8, $00000002, $00000000

  3) illegal memory access

      **** Data memory protection trap (TRAPS 68).

So...the "thing" that varied (for a successful trap) was paramter two.
As it turns out,   genmsgu (5, $68) produces "ACTIVATE",
and genmsgu (5, $c8) produces GETPRIVMODE.

I.e., it would appear that the second parameter is the intrinsic number
that caused the trap.

From CATALOG.PUB.SYS, the title for message set 5:

   $SET 5    INTRINSIC - FOR ABORT


BTW, XCODETRAP would have caught that third trap above.  For XCODETRAP,
the 4 parameters are:

   R26: unknown address of some kind ... does not appear to be pointing
        to the interrupt stack marker :(
   R25: space ID of instruction causing trap  (i.e, upper 32-bits of address)
   R24: offset of instruction causing trap  (i.e, bottom 32-bits of address)
   R23: instruction causing trap (in debug, do:  DIS R23 to see it)

--
Stan Sieler                                          [log in to unmask]
                                         http://www.allegro.com/sieler/