hrtimer: Make offset update smarter

On every tick/hrtimer interrupt we update the offset variables of the
clock bases. That's silly because these offsets change very seldom.

Add a sequence counter to the time keeping code which keeps track of
the offset updates (clock_was_set()). Have a sequence cache in the
hrtimer cpu bases to evaluate whether the offsets must be updated or
not. This allows us later to avoid pointless cacheline pollution.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: John Stultz <john.stultz@linaro.org>
Link: http://lkml.kernel.org/r/20150414203501.132820245@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: John Stultz <john.stultz@linaro.org>

include/linux/hrtimer.h

@@ -163,7 +163,7 @@ enum  hrtimer_base_type {
  *			and timers
  * @cpu:		cpu number
  * @active_bases:	Bitfield to mark bases with active timers
- * @clock_was_set:	Indicates that clock was set from irq context.
+ * @clock_was_set_seq:	Sequence counter of clock was set events
  * @expires_next:	absolute time of the next event which was scheduled
  *			via clock_set_next_event()
  * @in_hrtirq:		hrtimer_interrupt() is currently executing
@@ -179,7 +179,7 @@ struct hrtimer_cpu_base {
 	raw_spinlock_t			lock;
 	unsigned int			cpu;
 	unsigned int			active_bases;
-	unsigned int			clock_was_set;
+	unsigned int			clock_was_set_seq;
 #ifdef CONFIG_HIGH_RES_TIMERS
 	ktime_t				expires_next;
 	int				in_hrtirq;

include/linux/timekeeper_internal.h

@@ -49,6 +49,7 @@ struct tk_read_base {
  * @offs_boot:		Offset clock monotonic -> clock boottime
  * @offs_tai:		Offset clock monotonic -> clock tai
  * @tai_offset:		The current UTC to TAI offset in seconds
+ * @clock_was_set_seq:	The sequence number of clock was set events
  * @raw_time:		Monotonic raw base time in timespec64 format
  * @cycle_interval:	Number of clock cycles in one NTP interval
  * @xtime_interval:	Number of clock shifted nano seconds in one NTP
@@ -85,6 +86,7 @@ struct timekeeper {
 	ktime_t			offs_boot;
 	ktime_t			offs_tai;
 	s32			tai_offset;
+	unsigned int		clock_was_set_seq;
 	struct timespec64	raw_time;
 
 	/* The following members are for timekeeping internal use */

kernel/time/hrtimer.c

@@ -451,7 +451,8 @@ static inline ktime_t hrtimer_update_base(struct hrtimer_cpu_base *base)
 	ktime_t *offs_boot = &base->clock_base[HRTIMER_BASE_BOOTTIME].offset;
 	ktime_t *offs_tai = &base->clock_base[HRTIMER_BASE_TAI].offset;
 
-	return ktime_get_update_offsets_now(offs_real, offs_boot, offs_tai);
+	return ktime_get_update_offsets_now(&base->clock_was_set_seq,
+					    offs_real, offs_boot, offs_tai);
 }
 
 /* High resolution timer related functions */

kernel/time/timekeeping.c

@@ -602,6 +602,9 @@ static void timekeeping_update(struct timekeeper *tk, unsigned int action)
 
 	update_fast_timekeeper(&tk->tkr_mono, &tk_fast_mono);
 	update_fast_timekeeper(&tk->tkr_raw,  &tk_fast_raw);
+
+	if (action & TK_CLOCK_WAS_SET)
+		tk->clock_was_set_seq++;
 }
 
 /**
@@ -1927,15 +1930,19 @@ void do_timer(unsigned long ticks)
 
 /**
  * ktime_get_update_offsets_now - hrtimer helper
+ * @cwsseq:	pointer to check and store the clock was set sequence number
  * @offs_real:	pointer to storage for monotonic -> realtime offset
  * @offs_boot:	pointer to storage for monotonic -> boottime offset
  * @offs_tai:	pointer to storage for monotonic -> clock tai offset
  *
- * Returns current monotonic time and updates the offsets
+ * Returns current monotonic time and updates the offsets if the
+ * sequence number in @cwsseq and timekeeper.clock_was_set_seq are
+ * different.
+ *
  * Called from hrtimer_interrupt() or retrigger_next_event()
  */
-ktime_t ktime_get_update_offsets_now(ktime_t *offs_real, ktime_t *offs_boot,
-							ktime_t *offs_tai)
+ktime_t ktime_get_update_offsets_now(unsigned int *cwsseq, ktime_t *offs_real,
+				     ktime_t *offs_boot, ktime_t *offs_tai)
 {
 	struct timekeeper *tk = &tk_core.timekeeper;
 	unsigned int seq;
@@ -1947,10 +1954,12 @@ ktime_t ktime_get_update_offsets_now(ktime_t *offs_real, ktime_t *offs_boot,
 
 		base = tk->tkr_mono.base;
 		nsecs = timekeeping_get_ns(&tk->tkr_mono);
-
-		*offs_real = tk->offs_real;
-		*offs_boot = tk->offs_boot;
-		*offs_tai = tk->offs_tai;
+		if (*cwsseq != tk->clock_was_set_seq) {
+			*cwsseq = tk->clock_was_set_seq;
+			*offs_real = tk->offs_real;
+			*offs_boot = tk->offs_boot;
+			*offs_tai = tk->offs_tai;
+		}
 	} while (read_seqcount_retry(&tk_core.seq, seq));
 
 	return ktime_add_ns(base, nsecs);

kernel/time/timekeeping.h

@@ -3,9 +3,10 @@
 /*
  * Internal interfaces for kernel/time/
  */
-extern ktime_t ktime_get_update_offsets_now(ktime_t *offs_real,
-						ktime_t *offs_boot,
-						ktime_t *offs_tai);
+extern ktime_t ktime_get_update_offsets_now(unsigned int *cwsseq,
+					    ktime_t *offs_real,
+					    ktime_t *offs_boot,
+					    ktime_t *offs_tai);
 
 extern int timekeeping_valid_for_hres(void);
 extern u64 timekeeping_max_deferment(void);