On Mon, 4 Jan 1999, Manuel Bouyer wrote:

> On Jan 2, Havard.Eidnes@runit.sintef.no wrote
> > Hi,
> > 
> > although I am not a member of core, I'll speak up anyway.  I can
> > sympathize with those who think there should be some form of buy-in,
> > at least to the general direction of importing CAM to NetBSD.  I,
> > for one, would like to see it happen, for at least a couple of
> > different reasons:
> > 
> >  o Support for newer HBAs (e.g. the newer aic7xxx chips which
> >    started this particular discussion)
> 
> I'm not sure CAM is needed for this. We just need something that
> provides the same services than CAM. Maybe the best way to achieve this 
> is really to adapt CAM to NetBSD. Maybe we just don't want this.
> We still need a discution about this.


To be fair about it, the answer is 'no, you do not need CAm to support new
hardware'. It just makes it easier to port Justin's latest && greatest.

When I get finished porting my Qlogic target mode (which I've had working
under Solaris for a month or so) back into the common driver (common for
FreeBSD, NetBSD, Linux) there'll be a framework ready for it in
FreeBSD/CAM but not as much of a one in NetBSD.

> 
> > 
> >  o Better code sharing with FreeBSD, the hope being that
> >    development resources can be more efficiently used
> 
> This would be nice, for sure ...
> 
> > 
> >  o Better error recovery in SCSI (from what I understand)
> > 
> 
> Can be achieved in other ways.

Sure. There are incremental changes going on all the time in the current
SCSI layer. I even added a few even though I'm more interested in CAM
development at this time.