Re: [tip:timers/core] ptp_clock: Allow for it to be optional

From: Arnd Bergmann
Date: Fri Nov 18 2016 - 04:19:15 EST

On Thursday, November 17, 2016 7:48:00 PM CET Nicolas Pitre wrote:
> >
> > It was introduced in linux-next today, but it only happens if you
> > don't do a 'make clean' or 'make mrproper'. Apparently patch 1
> > of the series changes kconfig but that change does not trigger
> > a rebuild of the kconfig binary for me. After removing kconfig
> > from the object directory, it works again.
> That's odd. If I do a git checkout from linux-next of a commit before
> this series, then "make oldconfig", and then checkout of master and
> "make oldconfig again, in both cases I always get:
> $ make oldconfig
> HOSTCC scripts/kconfig/conf.o
> SHIPPED scripts/kconfig/
> SHIPPED scripts/kconfig/zconf.hash.c
> HOSTCC scripts/kconfig/
> HOSTLD scripts/kconfig/conf
> scripts/kconfig/conf --oldconfig Kconfig
> [...]
> And obviously no errors.

I usually build with "make -skj30", so it's possible that it only
happens in a parallel build environment on the first run.

> > I also ran into a problem with CONFIG_TIMERFD enabled but
> > CONFIG_POSIX_TIMERS turned off. This could be related to some
> > of my own patches though, haven't tried if that happens
> > with just your patches applied.
> Tried that combination in my checkout of linux-next, and once again no
> build errors here.

Ok, thanks for testing this. It must be something in the y2038
patches I have on top of next. They clashed quite a bit with your
work already, this added dependency is just one more detail for
me to remember with my series, nothing you need to worry about.