Subject: Re: MI "sbc" vs. MD "ncrscsi" driver for NCR 5380
To: None <port-mac68k@netbsd.org>
From: Chuck Silvers <chuq@chuq.com>
List: port-mac68k
Date: 01/18/2006 08:33:29
On Tue, Jan 17, 2006 at 12:55:26PM -0600, Dave Huang wrote:
> On Tue, Jan 17, 2006 at 08:48:27AM -0800, Chuck Silvers wrote:
> > does anyone know what problems still exist with the sbc driver that
> > would prevent us from switching to that and getting rid of the
> > mac68k-specific ncrscsi driver? the MI driver is 50% faster and uses
> > 1/3 less CPU time than the MD driver.
>
> I never ran NetBSD on a Mac that used the 5380, so I didn't pay too
> much attention to the problems; what I remember was that some drives
> worked properly with sbc, while others only worked properly with
> ncrscsi. Someone (Allen Briggs?) suggested that a SCSI bus analyzer
> might be helpful in figuring out what the exact problem is.
>
> Searching the list archives for messages that mention SCSI bus
> analyzer turns up these threads:
> http://mail-index.netbsd.org/port-mac68k/2000/09/04/0012.html
> http://mail-index.netbsd.org/port-mac68k/2000/10/27/0007.html
the only problem with sbc mentioned in those messages is that it reports
parity errors whereas the ncrscsi driver doesn't. the ncrscsi driver
appears to always ignore parity errors by default. it would be easy enough
to add a hook to sbc to allow the same behaviour (but I'd leave it off
by default there). if someone has such a device to test with, I could
provide a patch.
-Chuck