Re: [PATCH v4 08/12] gpio: add ti-ssp gpio driver

[Grant Likely <grant.likely@secretlab.ca>]

On Tue, Nov 02, 2010 at 07:15:16PM -0700, David Brownell wrote:
> 
> 
> --- On Tue, 10/26/10, Cyril Chemparathy <cyril@ti.com> wrote:
> 
> > From: Cyril Chemparathy <cyril@ti.com>
> > Subject: [PATCH v4 08/12] gpio: add ti-ssp gpio driver
> 
> On the assumptions you've tested this *AND* will
> resubmit with the previously-requested change of
> removing all references to non-existent
> "virtual"ness in the driver ... I'll likely
> add my ack to that resubmitted version
> 
> Also, chip2gpio seems an excessively generic name;
> maybe chip2_ti_ssp_gpio or somesuch?
> 
> I'd still be happier seeing this "stack" packaged
> as a hardware library called by various drivers
> (like GPIO, SPI, etc) rather than a core driver
> that's called by other drivers.  That seems like
> a better structure for various vendors' "SSP"
> hardware (multifunction serial interface logic)
> since most function drivers just need to poke
> the registers a bit differently, and don't have
> much to share with a "core driver" beyond a few
> setup routines (if that).

I thought the point of this device was that a single device hosted a
pair of multi-function serial interfaces, with each implementing a
separate function.  If so, then it makes sense for the base driver to
register child devices of the appropriate kinds.

g.