Re: next68k softintr(9) patch

On Tue, Mar 06, 2007 at 12:56:02AM +0900, Izumi Tsutsui wrote:
> Could anyone try or review attached patch?
> 
> There are several possible issues though:
> - According to next68k/include/cpu.h, TIMER is at ipl 6
>   but current IPL_CLOCK is (PSL_S|PSL_IPL3).
>   It was chaned after DMA fixes were integrated, but
>   is it really required?
> - NEXT_I_SOFTINT[01] are defined in cpu.h and it seems
>   that they can generate real hardware interrupts,
>   but current siron() doesn't touch any hardware and
>   it uses hp300 derived simulated sir to handle softintrs.
>   Are these bits are just taken from NEXTSTEP but unused?
> - next68k seems to have its own interrupt control register,
>   so maybe it should have own spl control routines like x86
>   rather than traditional m68k splN() based implementation.

good work ! I can't anserw your questions, but My Next (68040) boots multiuser
with this patch. Both disk and network looks fine.

thanks !

-- 
Manuel Bouyer, LIP6, Universite Paris VI.           [email protected]
     NetBSD: 26 ans d'experience feront toujours la difference
--