Messages in this thread | | | From | John Stultz <> | Subject | [PATCH 0/3] Cleanup ADJ_SETOFFSET patch | Date | Mon, 27 Dec 2010 15:40:41 -0800 |
| |
I was looking to queue Richard's ADJ_SETOFFSET patch to see if it could be merged into -tip for 2.6.38, but on second glance I noticed the ugly local_irq_disable bits as well as the fact that adding the offset uses a gettime/add/settime pattern which will also add a small error as the action isn't atomic.
So I implemented a timekeeping function to add a fixed offset: timekeeping_inject_offset(), and reworked Richard's code to make use of it.
Richard: Any objection here? Mind testing this with the rest of your patch queue?
Thomas: Any comments? Does this seem reasonable for 2.6.38? Should I fold my cleanups into Richard's patch?
thanks -john
CC: Richard Cochran <richard.cochran@omicron.at> CC: Thomas Gleixner <tglx@linutronix.de>
John Stultz (2): time: Introduce timekeeping_inject_offset ntp: Change ADJ_SETOFFSET implementation to use timekeeping_inject_offset
Richard Cochran (1): ntp: add ADJ_SETOFFSET mode bit
include/linux/time.h | 1 + include/linux/timex.h | 3 ++- kernel/time/ntp.c | 13 +++++++++++++ kernel/time/timekeeping.c | 36 ++++++++++++++++++++++++++++++++++++ 4 files changed, 52 insertions(+), 1 deletions(-)
-- 1.7.3.2.146.gca209
| |