timekeeping: Fix CLOCK_TAI timer/nanosleep delays
authorJohn Stultz <john.stultz@linaro.org>
Wed, 11 Dec 2013 01:13:35 +0000 (17:13 -0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 13 Feb 2014 21:48:03 +0000 (13:48 -0800)
commit 04005f6011e3b504cd4d791d9769f7cb9a3b2eae upstream.

A think-o in the calculation of the monotonic -> tai time offset
results in CLOCK_TAI timers and nanosleeps to expire late (the
latency is ~2x the tai offset).

Fix this by adding the tai offset from the realtime offset instead
of subtracting.

Cc: Sasha Levin <sasha.levin@oracle.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
kernel/time/timekeeping.c

index 88c2b8b..40f75a0 100644 (file)
@@ -72,7 +72,7 @@ static void tk_set_wall_to_mono(struct timekeeper *tk, struct timespec wtm)
        tk->wall_to_monotonic = wtm;
        set_normalized_timespec(&tmp, -wtm.tv_sec, -wtm.tv_nsec);
        tk->offs_real = timespec_to_ktime(tmp);
-       tk->offs_tai = ktime_sub(tk->offs_real, ktime_set(tk->tai_offset, 0));
+       tk->offs_tai = ktime_add(tk->offs_real, ktime_set(tk->tai_offset, 0));
 }
 
 static void tk_set_sleep_time(struct timekeeper *tk, struct timespec t)
@@ -590,7 +590,7 @@ s32 timekeeping_get_tai_offset(void)
 static void __timekeeping_set_tai_offset(struct timekeeper *tk, s32 tai_offset)
 {
        tk->tai_offset = tai_offset;
-       tk->offs_tai = ktime_sub(tk->offs_real, ktime_set(tai_offset, 0));
+       tk->offs_tai = ktime_add(tk->offs_real, ktime_set(tai_offset, 0));
 }
 
 /**