[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Would setting the CONFIG_RTC option break the powerpc kernel on your machine ?



On Wed, Dec 17, 2003 at 05:24:54PM +1100, Benjamin Herrenschmidt wrote:
> On Tue, 2003-12-16 at 21:56, Sven Luther wrote:
> > Hello,
> > 
> > On my Pegasos/CHRP box, i noticed that not setting the CONFIG_RTC option
> > in the Character device part of the configuration (not to be confused with
> > the CONFIG_GENERIC_RTC and CONFIG_PPC_RTC options in the General setup
> > part) will cause some strange delay when setting the hardware clock on
> > my box, and i believe the clock will be badly set (have to check again
> > to confirm though). Setting CONFIG_RTC will fix that, and furthermore,
> > in the help of the CONFIG_RTC there is some mention of this option being
> > needed for using the RTC clock in a SMP compatible fashion.
> 
> Hrm... that SMP thingy is weird. CONFIG_RTC shouldn't be used on

Well, i guess it is a x86 SMP documentation stuff, not sure.

> most PPCs. If GENERIC_RTC & PPC_RTC doesn't work properly, then we
> need to fix the pegasos patch.

Well, i use GENERIC_RTC also, but i get a delay when setting the hw
clock, maybe even a freeze in older versions not sure. I believe the
clock doesn't get set, but i have still to check that.

If the pegasos patch needs fixing, i am all for it.

BTW, the bigger initrd image is still causing problems on chrp, you said you
would look into it, did you have any chance to it ?

> > So, i believe this option is needed for me, and probably also for the
> > powerpc-smp config.
> > 
> > But, since the powerpc configurations are common to many subarch, there
> > is maybe a reason for not having this options set, and i would like to
> > know if setting it per default in the powerpc and powerpc-smp
> > configurations, but not in the powerpc-small one, is likely to break on
> > some of the supported powerps subarches out there.
> 
> CONFIG_RTC will definitely break a pmac

Any idea on the other non-pmac ?

Friendly,

Sven Luther



Reply to: