Re: defects for uses of abs(u64) (was: Re: Regression: can't apply frequency offsets above 1000ppm)

2015-09-23 Thread Neil Brown
Joe Perches writes: > On Fri, 2015-09-04 at 18:00 -0700, John Stultz wrote: >> On Fri, Sep 4, 2015 at 5:57 PM, John Stultz wrote: >> > On Thu, Sep 3, 2015 at 4:26 AM, Miroslav Lichvar >> > wrote: >> >> On Wed, Sep 02, 2015 at 04:16:00PM -0700, John Stultz wrote: >> >>> On Tue, Sep 1, 2015 at 6

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-08 Thread Nuno Gonçalves
Yes please. Tested-by: Nuno Goncalves Thanks, Nuno On Wed, Sep 9, 2015 at 1:52 AM, John Stultz wrote: > On Sat, Sep 5, 2015 at 6:41 AM, Nuno Gonçalves wrote: >> Considering your last message I just tried to use abs64 instead. Fixes >> the problem for me. > > Yea, I've since simplified my patc

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-08 Thread John Stultz
On Sat, Sep 5, 2015 at 6:41 AM, Nuno Gonçalves wrote: > Considering your last message I just tried to use abs64 instead. Fixes > the problem for me. Yea, I've since simplified my patch in the same way. Still looking good? Can I get a Tested-by: from you? thanks -john -- To unsubscribe from this

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-05 Thread Nuno Gonçalves
Considering your last message I just tried to use abs64 instead. Fixes the problem for me. diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 946acb7..414d9df 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -1615,7 +1615,7 @@ static __always_inline v

defects for uses of abs(u64) (was: Re: Regression: can't apply frequency offsets above 1000ppm)

2015-09-04 Thread Joe Perches
On Fri, 2015-09-04 at 18:00 -0700, John Stultz wrote: > On Fri, Sep 4, 2015 at 5:57 PM, John Stultz wrote: > > On Thu, Sep 3, 2015 at 4:26 AM, Miroslav Lichvar > > wrote: > >> On Wed, Sep 02, 2015 at 04:16:00PM -0700, John Stultz wrote: > >>> On Tue, Sep 1, 2015 at 6:14 PM, Nuno Gonçalves wrote

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-04 Thread John Stultz
On Fri, Sep 4, 2015 at 5:57 PM, John Stultz wrote: > On Thu, Sep 3, 2015 at 4:26 AM, Miroslav Lichvar wrote: >> On Wed, Sep 02, 2015 at 04:16:00PM -0700, John Stultz wrote: >>> On Tue, Sep 1, 2015 at 6:14 PM, Nuno Gonçalves wrote: >>> > And just installing chrony from the feeds. With any kernel

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-04 Thread John Stultz
On Thu, Sep 3, 2015 at 4:26 AM, Miroslav Lichvar wrote: > On Wed, Sep 02, 2015 at 04:16:00PM -0700, John Stultz wrote: >> On Tue, Sep 1, 2015 at 6:14 PM, Nuno Gonçalves wrote: >> > And just installing chrony from the feeds. With any kernel from 3.17 >> > you'll have wrong estimates at chronyc sou

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-03 Thread Miroslav Lichvar
On Wed, Sep 02, 2015 at 04:16:00PM -0700, John Stultz wrote: > On Tue, Sep 1, 2015 at 6:14 PM, Nuno Gonçalves wrote: > > And just installing chrony from the feeds. With any kernel from 3.17 > > you'll have wrong estimates at chronyc sourcestats. > > Wrong estimates? Could you be more specific abo

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-03 Thread Nuno Gonçalves
Sorry, The default chrony config from the debian package can't bring the clock in sync because it doesn't do steps, and it starts 15 years in the past. You had some other external help. Assuming "all factory config" just make sure to disable timesyncd: sudo systemctl disable systemd-timesyncd A

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-02 Thread John Stultz
On Tue, Sep 1, 2015 at 6:14 PM, Nuno Gonçalves wrote: > On Wed, Sep 2, 2015 at 2:03 AM, John Stultz wrote: >> On Tue, Sep 1, 2015 at 5:36 PM, Nuno Gonçalves wrote: >>> On Tue, Sep 1, 2015 at 9:25 PM, Thomas Gleixner wrote: On Tue, 1 Sep 2015, Nuno Gonçalves wrote: > There is a reg

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-02 Thread Miroslav Lichvar
On Wed, Sep 02, 2015 at 02:14:21AM +0100, Nuno Gonçalves wrote: > On Wed, Sep 2, 2015 at 2:03 AM, John Stultz wrote: > > On Tue, Sep 1, 2015 at 5:36 PM, Nuno Gonçalves wrote: > >> http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=dc491596f6394382fbc74ad331156207d619fa0a >

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-01 Thread Nuno Gonçalves
On Wed, Sep 2, 2015 at 2:03 AM, John Stultz wrote: > On Tue, Sep 1, 2015 at 5:36 PM, Nuno Gonçalves wrote: >> On Tue, Sep 1, 2015 at 9:25 PM, Thomas Gleixner wrote: >>> On Tue, 1 Sep 2015, Nuno Gonçalves wrote: >>> There is a regression on the clock system since v3.16-rc5-111-g4396e05

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-01 Thread John Stultz
On Tue, Sep 1, 2015 at 5:36 PM, Nuno Gonçalves wrote: > On Tue, Sep 1, 2015 at 9:25 PM, Thomas Gleixner wrote: >> On Tue, 1 Sep 2015, Nuno Gonçalves wrote: >> >>> There is a regression on the clock system since v3.16-rc5-111-g4396e05 >>> [1], >> >>> [1] >>> http://git.kernel.org/cgit/linux/kerne

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-01 Thread Nuno Gonçalves
On Tue, Sep 1, 2015 at 9:25 PM, Thomas Gleixner wrote: > On Tue, 1 Sep 2015, Nuno Gonçalves wrote: > >> There is a regression on the clock system since v3.16-rc5-111-g4396e05 >> [1], > >> [1] >> http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=4396e058c52e167729729cf64ea3

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-01 Thread Thomas Gleixner
On Tue, 1 Sep 2015, John Stultz wrote: > On Tue, Sep 1, 2015 at 1:25 PM, Thomas Gleixner wrote: > > On Tue, 1 Sep 2015, Nuno Gonçalves wrote: > > > >> There is a regression on the clock system since v3.16-rc5-111-g4396e05 > >> [1], > > > >> [1] > >> http://git.kernel.org/cgit/linux/kernel/git/tor

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-01 Thread John Stultz
On Tue, Sep 1, 2015 at 1:25 PM, Thomas Gleixner wrote: > On Tue, 1 Sep 2015, Nuno Gonçalves wrote: > >> There is a regression on the clock system since v3.16-rc5-111-g4396e05 >> [1], > >> [1] >> http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=4396e058c52e167729729cf64ea3

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-01 Thread John Stultz
On Tue, Sep 1, 2015 at 1:02 PM, Nuno Gonçalves wrote: > There is a regression on the clock system since v3.16-rc5-111-g4396e05 > [1], where the clock doesn't apply frequency offsets above about > 1000ppm [2]. ??? Hrm. I'm not sure I'm understanding here. Adjtimex frequency limits have been cappe

Re: Regression: can't apply frequency offsets above 1000ppm.

2015-09-01 Thread Thomas Gleixner
On Tue, 1 Sep 2015, Nuno Gonçalves wrote: > There is a regression on the clock system since v3.16-rc5-111-g4396e05 > [1], > [1] > http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=4396e058c52e167729729cf64ea3dfa229637086 That commit has absolutely nothing to do with NTP.

Regression: can't apply frequency offsets above 1000ppm.

2015-09-01 Thread Nuno Gonçalves
There is a regression on the clock system since v3.16-rc5-111-g4396e05 [1], where the clock doesn't apply frequency offsets above about 1000ppm [2]. This was found at a Beaglebone (armv7l, TI AM335x). Thanks, Nuno Goncalves [1] http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/comm