Re: Solved! (was: why is this needed?)

[Joerg Sonnenberger <joerg%britannica.bec.de@localhost>]

On Mon, Aug 17, 2009 at 02:26:41PM -0500, David Young wrote:
> On Mon, Aug 17, 2009 at 09:06:14PM +0200, Joerg Sonnenberger wrote:
> > On Mon, Aug 17, 2009 at 01:57:09PM -0500, David Young wrote:
> > > On Mon, Aug 17, 2009 at 04:04:08PM +0200, Marc Balmer wrote:
> > > > While here, I removed the unneeded and empty gscpcib_childdetach  
> > > > function.
> > > 
> > > The gscpcib_childdetach function, empty though it is, is needed so that
> > > children can be detached.  Please put that back. :-)
> > 
> > Yes and no. config_detach works, but detachdevbyname rejects that.
> > I am not sure if the latter is correct to do so.
> 
> I am sure that the latter is correct to do so. config_detach() should
> probably require the parent to provide a child-detached routine, too,
> but it's not necessary to discuss that before fixing the regression.

There should be a flag for that affect. I do not mind if this is a
*generic* childdetach routine or a device flag, but I don't think that
having an empty function is the right approach.

Joerg