On Sat, Jan 21, 2006 at 01:04:06PM +1100, James Haggerty wrote:
> 
> Still no improvement with diff1, diff2 reports 'Drive flags 0x51', so yes,
> that bit is set.

OK, so DMA is probably used, but the controller or the drive can't
run fast.
Could you run the test program with a large number, and check the
system and interrupt CPU usage with top or systat vm ?

Anyway I'm going to commit the patch, as at last it prevents the panic.

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--