Re: powerpc + i2c/rtc : where is the 11 min update ?

Re: powerpc + i2c/rtc : where is the 11 min update ?

[Philippe De Muyter]

CCing lkml

On Tue, May 20, 2008 at 09:10:25PM +1000, Paul Mackerras wrote:
> Philippe De Muyter writes:
> 
> > I have a ARCH=powerpc linux-2.6.25-rc6 linux running with an i2c rtc chip,
> > and synchronized to a ntp server.
> > 
> > I noticed that my rtc chip does not get updated by the kernel, just like
> > it would be on all other architectures (included ppc).
> 
> That is now supposed to be done by the code inside #ifdef
> CONFIG_GENERIC_CMOS_UPDATE in kernel/time/ntp.c.
> 
> Paul.

Thanks.  I agree that is a good place.

But, here CONFIG_GENERIC_CMOS_UPDATE is defined and an i2c clock also,
and my rtc clock is not updated.  I see kernel/time/ntp.c::sync_cmos_clock
calling arch/powerpc/kernel/time.c::update_persistent_clock, where
ppc_md.set_rtc_time is NULL.

Who is supposed to initialize ppc_md.set_rtc_time to use an i2c clock and
when in the boot process may that happen ?

Or alternatively, should update_persistent_clock not be part of the rtc
subsystem ?

Philippe

RTC subsystem and 11 min rtc update

[Philippe De Muyter]

Hello everybody,

I have linux-2.6.25-rc6 running on a board with an i2c rtc chip,
and synchronized to a ntp server.  When my board starts, the current
time is correctly read from the rtc chip, but ...

I noticed that my rtc chip does not get updated by the kernel every 11 min,
on my powerpc board, just like it would be on all other architectures.

Paul Mackerras told me that that is now supposed to be done by the code
inside #ifdef CONFIG_GENERIC_CMOS_UPDATE in kernel/time/ntp.c.

I agree that is a good place.

But, here CONFIG_GENERIC_CMOS_UPDATE is well defined and an i2c clock also,
and my rtc clock is not updated.  I see kernel/time/ntp.c::sync_cmos_clock
calling arch/powerpc/kernel/time.c::update_persistent_clock, where
ppc_md.set_rtc_time is NULL. ppc_md.get_rtc_time is NULL as well, but
that does not prevent the initial reading of the rtc clock during the startup
of the kernel.

Should update_persistent_clock not be directly part of the rtc subsystem,
instead of falling in a detour or a dead-end in the arch subsystem ?

Philippe

Re: RTC subsystem and 11 min rtc update

[Maciej W. Rozycki]

On Wed, 21 May 2008, Philippe De Muyter wrote:

> Should update_persistent_clock not be directly part of the rtc subsystem,
> instead of falling in a detour or a dead-end in the arch subsystem ?

 See the recent discussion around my proposal:

http://www.ussg.iu.edu/hypermail/linux/kernel/0805.1/1772.html

I am about to update some of the architecture bits in arch/mips/sibyte,
but otherwise the patchset is ready to go.  Feel free to join the
discussion and/or update your platform to make use of the generic
rtc_read_persistent_clock() and rtc_update_persistent_clock() functions.  
Testing is always welcome too.

  Maciej