On Sun, Mar 11, 2007 at 06:34:44PM +0100, Havard Eidnes wrote:
> > I browsed through the diffs to catch possible caddr_t fallout, and I
> > stumbled accross this one:
> >
> > you changed the assignment from assigning the pointer to pointer (previous)
> > to assigning the pointer. Unless I missed something, from reading the last
> > few versions of the driver, it seems to me that sc_dmaaddr is really a 
> > pointer to the address pointer, so your change would be wrong. 
> >
> > Am I missing something?
> 
> This must be characterized as a "thinko" from my side, sorry
> about that.  Sometimes making the compiler shut up the simplest
> way is not the right way...  This same bug was present in all of
> bzivsc.c bzsc.c bztzsc.c cbiisc.c cbsc.c flsc.c.  At least I was
> consistent...

Of course - all use the same chip family, IIRC, and are partial
copies from each other.

> I've now committed a fix which reverts this particular change and
> instead does an explicit cast to (char **), and the result still
> builds...

Thanks - will save me from worrying about that, and help me concentrate
on the graphics drivers.

Regards,
	Ignatios