cvs commit: src/sbin/kldunload kldunload.8 kldunload.c

[M. Warner Losh]

In message: <25436.1089752144 at critter.freebsd.dk>
            "Poul-Henning Kamp" <phk at phk.freebsd.dk> writes:
: In message <20040713.144758.88381443.imp at bsdimp.com>, "M. Warner Losh" writes:
: >In message: <25096.1089751422 at critter.freebsd.dk>
: >            "Poul-Henning Kamp" <phk at phk.freebsd.dk> writes:
: >: In message <40F4477F.9020906 at root.org>, Nate Lawson writes:
: >: >Poul-Henning Kamp wrote:
: >: >> ... somebody who has
: >: >> wasted a couple of months trying to write a newbus-ified GEOM.
: >: >
: >: >I'll happily help review patches/designs for this.
: >: 
: >: No need to.  It doesn't work because newbus doesn't support
: >: multiple interitance.
: >
: >Yes it does.  Or are you talking about multi-pathing...
: 
: It does ?  When did that happen ?

When Doug added the subclassing stuff some months ago.  cardbus now is
a subclass of pci, for example.

: I'm pretty sure i mean "multiple inheritance":
: 
: 	objects A, B, X, Y;
: 	A is a child of X and Y;
: 	B is a child of Y and Z;
: 
: that was called multiple inheritance last I read about OO.

If you are talking types, that's multiple inheritance.  If you are
talking instances in the object tree, that's multi pathing.

newbus definitely supports multiple inheritance like you describe.
Just use DEFINE_CLASS_2 for objects A and B and list X and Y for the
first one and Y and Z for the second.

: (Multi-pathing is the ability to join independent paths, that is
: the correct way to solve the FC redundancy issue, and I belive we
: need both CAM and newbus to learn new tricls before that is realistic)

Agreed.

Warner