timekeeping fix patch got mis-applied
authorThomas Gleixner <tglx@linutronix.de>
Mon, 14 May 2007 09:10:02 +0000 (11:10 +0200)
committerLinus Torvalds <torvalds@woody.linux-foundation.org>
Mon, 14 May 2007 19:13:11 +0000 (12:13 -0700)
The time keeping code move to kernel/time/timekeeping.c broke the
clocksource resume logic patch, which got applied to the old file by a
fuzzy application.  Fix it up and move the clocksource_resume() call to
the appropriate place.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
[ tssk, tssk, everybody should use --fuzz=0 ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kernel/time/timekeeping.c
kernel/timer.c

index f9217bf..3d1042f 100644 (file)
@@ -273,6 +273,8 @@ static int timekeeping_resume(struct sys_device *dev)
        unsigned long flags;
        unsigned long now = read_persistent_clock();
 
+       clocksource_resume();
+
        write_seqlock_irqsave(&xtime_lock, flags);
 
        if (now && (now > timekeeping_suspend_time)) {
index a6c580a..5ec5490 100644 (file)
@@ -1499,8 +1499,6 @@ unregister_time_interpolator(struct time_interpolator *ti)
                prev = &curr->next;
        }
 
-       clocksource_resume();
-
        write_seqlock_irqsave(&xtime_lock, flags);
        if (ti == time_interpolator) {
                /* we lost the best time-interpolator: */