cirrus_laguna.o vs cirrus_laguna_drv.o (was: Re: User problems with the DLLoader)

[Alex Deucher]

On Thu, 27 Jan 2005 17:00:54 +0000, Derek Fawcus <dfawcus at cisco.com> wrote:
> On Thu, Jan 27, 2005 at 01:05:22PM +1100, Daniel Stone wrote:
> > I would vastly prefer to drag the drivers into line with the rough standard, but
> > I suppose we could do both?  Maybe someone could shed some light into why these
> > drivers don't have _drv appended.
> 
> Well I created them when I split the Alpine code from the Laguna code.  They don't
> have _drv.o because they are sub modules,  they are loaded by the main cirrus
> driver (which I believe is named cirrus_drv.o).
> 
> As to their existance - political.  I actually wanted to create a alpine_drv.o and
> a laguna_drv.o but at the time was told one _must_ have the top level "cirrus"
> driver known as cirrus.  So I complied.
> 

That expalins why we have "ati" as a wrapper for r128, radeon, and
atimisc.  I wonder why that wasn't enforced on S3 (wrapping s3, virge,
and savage).  That might also explain the riva module.

Alex

> I'd rather they were split to be two drivers build - "laguna" and "alpine".
> 
> Or possibly "cirrus" and "cirrus_laguna" as the laguna is the one odd member of
> the cirrus family of chips,  having a completely different programming interface.
> 
> DF