Re: rtc-ppc

Re: rtc-ppc

[David Woodhouse]

On Wed, 2008-11-05 at 02:19 +0100, Alessandro Zummo wrote:
>   I was checking a few drivers for compliance and noticed that
>  the rtc-ppc driver registers a platform device instead of
>  a platform driver.
> 
>   Wouldn't be possible to have the ppc arch doing the device
>  thing and cleanup the driver?

It would be possible, yes -- but it would be better to have the various
PPC platforms just register RTC-class devices _directly_, and ditch the
RTC bits from ppc_md altogether.

I think we're waiting until the RTC class works with NTP before we can
contemplate that, though.

-- 
David Woodhouse                            Open Source Technology Centre
David.Woodhouse@intel.com                              Intel Corporation

Re: rtc-ppc

[Alessandro Zummo]

On Wed, 05 Nov 2008 23:18:46 +0100
David Woodhouse <dwmw2@infradead.org> wrote:

> It would be possible, yes -- but it would be better to have the various
> PPC platforms just register RTC-class devices _directly_, and ditch the
> RTC bits from ppc_md altogether.

 That would be real fine :)

> I think we're waiting until the RTC class works with NTP before we can
> contemplate that, though.

 That will be a long wait, I don't think anyone volunteered for
 the ntp code. I've been told some platforms are already using the class.


-- 

 Best regards,

 Alessandro Zummo,
  Tower Technologies - Torino, Italy

  http://www.towertech.it

Re: rtc-ppc

[David Gibson]

On Wed, Nov 05, 2008 at 11:18:46PM +0100, David Woodhouse wrote:
> On Wed, 2008-11-05 at 02:19 +0100, Alessandro Zummo wrote:
> >   I was checking a few drivers for compliance and noticed that
> >  the rtc-ppc driver registers a platform device instead of
> >  a platform driver.
> > 
> >   Wouldn't be possible to have the ppc arch doing the device
> >  thing and cleanup the driver?
> 
> It would be possible, yes -- but it would be better to have the various
> PPC platforms just register RTC-class devices _directly_, and ditch the
> RTC bits from ppc_md altogether.
> 
> I think we're waiting until the RTC class works with NTP before we can
> contemplate that, though.

Yeah, I tried writing some ppc_md hooks which backended onto the rtc
class drivers as an interim measure, but the hooks are called from
places the rtc class stuff can't safely be called (too early, IIRC,
but it's a while since I checked).

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

Re: [rtc-linux] Re: rtc-ppc

[Alessandro Zummo]

On Thu, 6 Nov 2008 10:08:25 +1100
David Gibson <david@gibson.dropbear.id.au> wrote:

> > I think we're waiting until the RTC class works with NTP before we can
> > contemplate that, though.  
> 
> Yeah, I tried writing some ppc_md hooks which backended onto the rtc
> class drivers as an interim measure, but the hooks are called from
> places the rtc class stuff can't safely be called (too early, IIRC,
> but it's a while since I checked).

 I think the rtc drivers are too heterogeneous to have a valid
 ntp strategy. 

-- 

 Best regards,

 Alessandro Zummo,
  Tower Technologies - Torino, Italy

  http://www.towertech.it