On Wed, 2014-02-12 at 17:02 -0800, John Stultz wrote: > On 02/12/2014 03:52 PM, Ben Hutchings wrote: > > On Mon, 2014-02-10 at 13:16 -0800, John Stultz wrote: > >> Just wanted to send out a few timekeeping fixes that were merged > >> in 3.14 which are appropriate for -stable (now that -rc2 is otut), > >> as well as an extra fix I noticed hadn't been submitted back to 3.4. > >> > >> This queue backports the following fixes: > >> ----------------------------------------- > >> d5a1c7e3fc38d9c7d629e1e47f32f863acbdec3d rtc-cmos: Add an alarm > disable quirk > >> 6fdda9a9c5db367130cf32df5d6618d08b89f46a timekeeping: Avoid > possible deadlock from clock_was_set_delayed > >> > >> This also includes a simple fix from Colin Cross that has > >> been in the AOSP common.git tree for awhile that handles > >> potential overflows that were fixed in 3.6 w/: > >> ec145babe754f9ea1079034a108104b6001e001c time: Fix > timeekeping_get_ns overflow on 32bit systems > >> > >> > >> I've run these through my timetest suite w/ kvm on arm, i386, > >> & x86_64. But more testing would be of course appreciated. > >> https://github.com/johnstultz-work/timetests > > [...] > > > > Are these also applicable to 3.2? > > Maybe? Sorry for missing 3.2 was still an active stable tree. :( It's active, but only releasing about once a month - I can't keep pace with Greg! (Even 2.6.32 is still active, but hopefully it doesn't suffer from these bugs.) > The rtc-cmos should apply without issue. > > The deadlock item we shouldn't need because 3.2 still raises the softirq > in clock_was_set_delayed instead of calling schedule_work(). So that > should be ok. > > The get_monotonic_boottime fix looks like it should be needed though > (and does apply with fuzz). > > I'm running tests now... would you be able to pick up the patches I've > already sent or do I need to resubmit? I saved the patches and will queue them up for 3.2.56 if your tests passed. > Thanks for the ping, and sorry again for missing this the first time. Thanks. Ben. -- Ben Hutchings The world is coming to an end. Please log off.
Attachment:
signature.asc
Description: This is a digitally signed message part