On Thu, Jan 30, 2014 at 09:58:21AM -0800, John Stultz wrote:
> On 01/30/2014 06:44 AM, Luis Henriques wrote:
> > Hi John,
> >
> > On Wed, Jan 22, 2014 at 05:34:53PM -0800, John Stultz wrote:
> >> Just wanted to send out a few timekeeping fixes that were merged
> >> in 3.14 which are appropriate for -stable.
> >>
> >> This queue backports the following fixes:
> >> -----------------------------------------
> >> f55c07607a38f84b5c7e6066ee1cfe433fa5643c timekeeping: Fix lost updates to tai adjustment
> >> 5258d3f25c76f6ab86e9333abf97a55a877d3870 timekeeping: Fix potential lost pv notification of time change
> >> 6fdda9a9c5db367130cf32df5d6618d08b89f46a timekeeping: Avoid possible deadlock from clock_was_set_delayed
> >> 04005f6011e3b504cd4d791d9769f7cb9a3b2eae timekeeping: Fix CLOCK_TAI timer/nanosleep delays
> >> 330a1617b0a6268d427aa5922c94d082b1d3e96d timekeeping: Fix missing timekeeping_update in suspend path
> >> d5a1c7e3fc38d9c7d629e1e47f32f863acbdec3d rtc-cmos: Add an alarm disable quirk
> > Two questions on these patches:
> >
> > 1) Should they be applied to 3.13 stable only? At least some of them seem
> > to be applicable to older kernels as well.
>
> Yes, and I have already tested (x86,i386,arm) backports ready for 3.13,
> 3.10 and 3.4 here:
>
> git://git.linaro.org/people/john.stultz/linux.git stable/3.13/timefix
> git://git.linaro.org/people/john.stultz/linux.git stable/3.10/timefix
> git://git.linaro.org/people/john.stultz/linux.git stable/3.4/timefix
>
> (Note: The 3.4 tree also has an additional fix thats been hanging out in
> the AOSP tree for months).
>
> But since these landed in the 3.14 merge window, I'll be waiting till
> after -rc2 to submit them.
>
> For 2.6.34 and 2.6.32, the clock_was_set_delayed() call raises a softirq
> instead of calling sched_work, so it shouldn't have the deadlock issue
> fixed in the series above.
Great, thanks a lot for the clarification (and the backports!). I'll
queue these for the 3.5 and 3.11 kernels once -rc1 is out.
> > 2) Shouldn't 47a1b796 ("tick/timekeeping: Call update_wall_time outside
> > the jiffies lock") be applied as well? There's a compilation warning
> > introduced by 6fdda9a:
> >
> > kernel/time/timekeeping.c:1446:6: warning: the address of 'clock_was_set' will always evaluate as 'true' [-Waddress]
> >
> > that seems to be a bug, fixed by 47a1b796.
>
> Yes. :( If you notice in the patch series I sent out, there is a fix for
> the thinko (which was why I sent out this series, instead of just
> waiting...). Greg has already clarified I shouldn't fold such a fix in,
> so in my current 3.13 branch above, you'll notice there's a new patch:
> http://git.linaro.org/people/john.stultz/linux.git/commit/129f8bf1beaec75c2bdaa02d7d2b63989b6a7ed6
>
> Which pulls just the needed correction from 47a1b796.
Doh! I completely missed that the fix was already in your backport. All
makes sense now, and I'll use the additional patch as well. Again, thanks
for lot for the clarifications.
Cheers,
--
Luis
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html