Re: [PATCH 5/5] watchdog: introduce platform_data and remove cpu conditional code

[Felipe Balbi <me@felipebalbi.com>]

On Fri, Sep 19, 2008 at 08:46:34AM +0100, Russell King - ARM Linux wrote:
> On Fri, Sep 19, 2008 at 02:23:25AM +0300, Felipe Balbi wrote:
> > On Fri, Sep 19, 2008 at 12:20:06AM +0100, Russell King - ARM Linux wrote:
> > > On Fri, Sep 19, 2008 at 01:23:42AM +0300, Felipe Balbi wrote:
> > > > +struct omap_wdt_platform_data {
> > > > +	const char	*ick;
> > > > +	const char	*fck;
> > > 
> > > Grumble.  The clk API is supposed to make this stuff completely
> > > unnecessary - but OMAP doesn't implement the clk API correctly -
> > > by using the 'name' as the sole key.  The 'name' argument is
> > > supposed to be used to disambiguate between multiple clock
> > > inputs on a single device. ;(
> > 
> > Yes, and we have two clocks for the same device_id on the bus. Right ?
> 
> That's not a problem.  I have an solution to solve this, but for the
> time being, let's just keep the logic inside the omap_wdt driver.

Ok, I'll wait on this patch then. The sync with mainline and ioremap()
fix will be going to Wim soon, I just need to test for a while.

-- 
balbi