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

Re: TRAP #1 from TRAP #2?



Claus Brod <claus_brod@hpbbn.bbn.hp.com> writes:

>Scott Bigham wrote:

>>    [...]  My guess is that I have to somehow invoke the GEMDOS trap
>> handler "manually"; my question is, how?

>I don't see problems in calling GEMDOS functions from within an AES
>function.

No?  This surprises me.  Once I hit the AES trap, I'm "inside the
kernel", so to speak.  I'm in supervisor mode instead of user mode.
I'm on the back side of a non-reentrant kernel interface.  If you say
this isn't a problem, I'll believe you; it just comes as something of
a surprise.

>> In fact, now that I think about it, I'm going to have the same problem
>> when I try to call e.g. evnt_multi(), but worse, since not only will I
>> have to invoke the TRAP #2 handler "manually", but I'll also have to
>> provide my own set of intin, intout, et al. arrays -- correct? incorrect?

>That's a more interesting question, indeed. I would think you're safe if
>you have your own intin etc. fields.

And this comes as even more a surprise, because ISTR the Compendium
making much more explicit noises about not hitting the AES trap from
inside its own trap handler.  [Or maybe it's another trap I'm thinking
of; I'm at work ATM and can't check.]

						-sbigham