Hi,

> I had that problem too, and I used a hack.  If curproc is 0 I use a timeout()
> and do the rest of the processing in the timeout function, otherwise I
> use tsleep().  The timeout() version would work all the time, of course,
> but in this particular instance I had reasons not to do it that way.

So are you saying that your code might call tsleep in an interrupt handler?
As others have already said, this would be a gross error!  It's not enough
to test curproc against 0 to test for an interrupt context.  That pointer
is only 0 when the machine is in the idle loop, but you can of course get
interrupts at other times, too.

I have to admit however, that I haven't looked at any of your code yet.

Ciao,
Wolfgang
-- 
ws@TooLs.DE     (Wolfgang Solfrank, TooLs GmbH) 	+49-228-985800