timekeeping: add arch_offset hook to ktime_get functions
Hector Palacios [Mon, 14 Nov 2011 10:15:25 +0000 (11:15 +0100)]
commit d004e024058a0eaca097513ce62cbcf978913e0a upstream.

ktime_get and ktime_get_ts were calling timekeeping_get_ns()
but later they were not calling arch_gettimeoffset() so architectures
using this mechanism returned 0 ns when calling these functions.

This happened for example when running Busybox's ping which calls
syscall(__NR_clock_gettime, CLOCK_MONOTONIC, ts) which eventually
calls ktime_get. As a result the returned ping travel time was zero.

Signed-off-by: Hector Palacios <hector.palacios@digi.com>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

kernel/time/timekeeping.c

index 2b021b0..da4a778 100644 (file)
@@ -249,6 +249,8 @@ ktime_t ktime_get(void)
                secs = xtime.tv_sec + wall_to_monotonic.tv_sec;
                nsecs = xtime.tv_nsec + wall_to_monotonic.tv_nsec;
                nsecs += timekeeping_get_ns();
+               /* If arch requires, add in gettimeoffset() */
+               nsecs += arch_gettimeoffset();
 
        } while (read_seqretry(&xtime_lock, seq));
        /*
@@ -280,6 +282,8 @@ void ktime_get_ts(struct timespec *ts)
                *ts = xtime;
                tomono = wall_to_monotonic;
                nsecs = timekeeping_get_ns();
+               /* If arch requires, add in gettimeoffset() */
+               nsecs += arch_gettimeoffset();
 
        } while (read_seqretry(&xtime_lock, seq));