> I doubt the SCSI pointer reload interrupt can be used until the
> spl*() and interrupt handling is significantly revised.  I suspect
> that the SCSI pointer reload interrupt would need to be kept
> enabled all the time, even at splhigh() to ensure that the interrupt
> can be processed quickly.

By manupilating enable/disable bits in IOASIC CSR?

(And I'm wondering how "the new way" can be fit in IOASIC Alphas)

Tohru Nishimura
Nara Institute of Science and Technology