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

Re: [MiNT] Fwd: Re: marathon,was:Egale



Jo Even Skarstein wrote:

> > So the AES should test for every trap if it comes from an slb, and if so
> > use the callers apid? I guess it's
> > easier and much cleaner to implement a proper interface in the slb (like
> > other libs do).
>
> Again, I don't understand what you mean. Andreas Kromke say this about SLB
> functions:
>
> It is executed in the context of the caller, and actually with the stack
> of the caller (depending on status usp or ssp).
>
> So the AES shouldn't have to worry about this at all - the call would be
> coming from the calling process and NOT the SLB process. If it doesn't,
> then MiNT's implementation of SLB is not correct.

Possibly the slb starts to run while MiNT is still busy initializing the
slb-stuff (the SLB-flag is not set in appl_init). Maybe a fix in MiNT is
needed.

> If the Draconis mailer calls appl_init() twice (once in it's own code and
> once by calling a SLB function) it should have the same effect as calling
> appl_init() twice from it's own code.

I'm not sure if the slb-appl_init comes from the caller (why should it do this twice?), I assume the
slb does this on its own, and wants to become an AES-client. Someone should check this in the code.

-Helmut