clocksource / timekeeping rework V4 (resend V3 + bug fix)

Index: linux-2.6/include/linux/time.h
===================================================================
@@ -147,6 +147,7 @@ extern struct timespec timespec_trunc(st
 extern int timekeeping_valid_for_hres(void);
 extern void update_wall_time(void);
 extern void update_xtime_cache(u64 nsec);
+extern void timekeeping_leap_insert(int leapsecond);
 
 struct tms;
 extern void do_sys_times(struct tms *);

Index: linux-2.6/kernel/time/ntp.c
===================================================================
@@ -194,8 +194,7 @@ static enum hrtimer_restart ntp_leap_sec
 	case TIME_OK:
 		break;
 	case TIME_INS:
-		xtime.tv_sec--;
-		wall_to_monotonic.tv_sec++;
+		timekeeping_leap_insert(-1);
 		time_state = TIME_OOP;
 		printk(KERN_NOTICE
 			"Clock: inserting leap second 23:59:60 UTC\n");

@@ -203,9 +202,8 @@ static enum hrtimer_restart ntp_leap_sec
 		res = HRTIMER_RESTART;
 		break;
 	case TIME_DEL:
-		xtime.tv_sec++;
+		timekeeping_leap_insert(1);
 		time_tai--;
-		wall_to_monotonic.tv_sec--;
 		time_state = TIME_WAIT;
 		printk(KERN_NOTICE
 			"Clock: deleting leap second 23:59:59 UTC\n");

@@ -219,7 +217,6 @@ static enum hrtimer_restart ntp_leap_sec
 			time_state = TIME_OK;
 		break;
 	}
-	update_vsyscall(&xtime, clock);
 
 	write_sequnlock(&xtime_lock);
 

Index: linux-2.6/kernel/time/timekeeping.c
===================================================================
@@ -58,6 +58,13 @@ void update_xtime_cache(u64 nsec)
 
 struct clocksource *clock;
 
+/* must hold xtime_lock */
+void timekeeping_leap_insert(int leapsecond)
+{
+	xtime.tv_sec += leapsecond;
+	wall_to_monotonic.tv_sec -= leapsecond;
+	update_vsyscall(&xtime, clock);
+}
 
 #ifdef CONFIG_GENERIC_TIME
 /**

index ea16c1a..e7c8445 100644
@@ -147,6 +147,7 @@ extern struct timespec timespec_trunc(struct timespec t, unsigned gran);
 extern int timekeeping_valid_for_hres(void);
 extern void update_wall_time(void);
 extern void update_xtime_cache(u64 nsec);
+extern void timekeeping_leap_insert(int leapsecond);
 
 struct tms;
 extern void do_sys_times(struct tms *);

index 7fc6437..4800f93 100644
@@ -194,8 +194,7 @@ static enum hrtimer_restart ntp_leap_second(struct hrtimer *timer)
 	case TIME_OK:
 		break;
 	case TIME_INS:
-		xtime.tv_sec--;
-		wall_to_monotonic.tv_sec++;
+		timekeeping_leap_insert(-1);
 		time_state = TIME_OOP;
 		printk(KERN_NOTICE
 			"Clock: inserting leap second 23:59:60 UTC\n");

@@ -203,9 +202,8 @@ static enum hrtimer_restart ntp_leap_second(struct hrtimer *timer)
 		res = HRTIMER_RESTART;
 		break;
 	case TIME_DEL:
-		xtime.tv_sec++;
+		timekeeping_leap_insert(1);
 		time_tai--;
-		wall_to_monotonic.tv_sec--;
 		time_state = TIME_WAIT;
 		printk(KERN_NOTICE
 			"Clock: deleting leap second 23:59:59 UTC\n");

@@ -219,7 +217,6 @@ static enum hrtimer_restart ntp_leap_second(struct hrtimer *timer)
 			time_state = TIME_OK;
 		break;
 	}
-	update_vsyscall(&xtime, clock);
 
 	write_sequnlock(&xtime_lock);
 

index 02c0b2c..b8b70fb 100644
@@ -58,6 +58,13 @@ void update_xtime_cache(u64 nsec)
 
 struct clocksource *clock;
 
+/* must hold xtime_lock */
+void timekeeping_leap_insert(int leapsecond)
+{
+	xtime.tv_sec += leapsecond;
+	wall_to_monotonic.tv_sec -= leapsecond;
+	update_vsyscall(&xtime, clock);
+}
 
 #ifdef CONFIG_GENERIC_TIME
 /**

Index: linux-2.6/kernel/time/timekeeping.c
===================================================================
@@ -79,7 +79,7 @@ static void clocksource_forward_now(void
 	cycle_t cycle_now, cycle_delta;
 	s64 nsec;
 
-	cycle_now = clocksource_read(clock);
+	cycle_now = clock->read(clock);
 	cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
 	clock->cycle_last = cycle_now;
 

@@ -114,7 +114,7 @@ void getnstimeofday(struct timespec *ts)
 		*ts = xtime;
 
 		/* read clocksource: */
-		cycle_now = clocksource_read(clock);
+		cycle_now = clock->read(clock);
 
 		/* calculate the delta since the last update_wall_time: */
 		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;

@@ -146,7 +146,7 @@ ktime_t ktime_get(void)
 		nsecs = xtime.tv_nsec + wall_to_monotonic.tv_nsec;
 
 		/* read clocksource: */
-		cycle_now = clocksource_read(clock);
+		cycle_now = clock->read(clock);
 
 		/* calculate the delta since the last update_wall_time: */
 		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;

@@ -186,7 +186,7 @@ void ktime_get_ts(struct timespec *ts)
 		tomono = wall_to_monotonic;
 
 		/* read clocksource: */
-		cycle_now = clocksource_read(clock);
+		cycle_now = clock->read(clock);
 
 		/* calculate the delta since the last update_wall_time: */
 		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;

@@ -274,16 +274,29 @@ static void change_clocksource(void)
 
 	clocksource_forward_now();
 
-	if (clocksource_enable(new))
+	if (new->enable && ! new->enable(new))
 		return;
+	/*
+	 * The frequency may have changed while the clocksource
+	 * was disabled. If so the code in ->enable() must update
+	 * the mult value to reflect the new frequency. Make sure
+	 * mult_orig follows this change.
+	 */
+	new->mult_orig = new->mult;
 
 	new->raw_time = clock->raw_time;
 	old = clock;
 	clock = new;
-	clocksource_disable(old);
+	/*
+	 * Save mult_orig in mult so that the value can be restored
+	 * regardless if ->enable() updates the value of mult or not.
+	 */
+	old->mult = old->mult_orig;
+	if (old->disable)
+		old->disable(old);
 
 	clock->cycle_last = 0;
-	clock->cycle_last = clocksource_read(clock);
+	clock->cycle_last = clock->read(clock);
 	clock->error = 0;
 	clock->xtime_nsec = 0;
 	clocksource_calculate_interval(clock, NTP_INTERVAL_LENGTH);

@@ -373,7 +386,7 @@ void getrawmonotonic(struct timespec *ts
 		seq = read_seqbegin(&xtime_lock);
 
 		/* read clocksource: */
-		cycle_now = clocksource_read(clock);
+		cycle_now = clock->read(clock);
 
 		/* calculate the delta since the last update_wall_time: */
 		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;

@@ -435,9 +448,12 @@ void __init timekeeping_init(void)
 	ntp_init();
 
 	clock = clocksource_get_next();
-	clocksource_enable(clock);
+	if (clock->enable)
+		clock->enable(clock);
+	/* set mult_orig on enable */
+	clock->mult_orig = clock->mult;
 	clocksource_calculate_interval(clock, NTP_INTERVAL_LENGTH);
-	clock->cycle_last = clocksource_read(clock);
+	clock->cycle_last = clock->read(clock);
 
 	xtime.tv_sec = sec;
 	xtime.tv_nsec = 0;

@@ -477,8 +493,7 @@ static int timekeeping_resume(struct sys
 	}
 	update_xtime_cache(0);
 	/* re-base the last cycle value */
-	clock->cycle_last = 0;
-	clock->cycle_last = clocksource_read(clock);
+	clock->cycle_last = clock->read(clock);
 	clock->error = 0;
 	timekeeping_suspended = 0;
 	write_sequnlock_irqrestore(&xtime_lock, flags);

@@ -630,7 +645,7 @@ void update_wall_time(void)
 		return;
 
 #ifdef CONFIG_GENERIC_TIME
-	offset = (clocksource_read(clock) - clock->cycle_last) & clock->mask;
+	offset = (clock->read(clock) - clock->cycle_last) & clock->mask;
 #else
 	offset = clock->cycle_interval;
 #endif

64 lines not on this page · raw message

index 1219be4..a1ef46f 100644
@@ -268,64 +268,6 @@ static inline u32 clocksource_hz2mult(u32 hz, u32 shift_constant)
 }
 
 /**
- * clocksource_read: - Access the clocksource's current cycle value
- * @cs:		pointer to clocksource being read
- *
- * Uses the clocksource to return the current cycle_t value
- */
-static inline cycle_t clocksource_read(struct clocksource *cs)
-{
-	return cs->read(cs);
-}
-
-/**
- * clocksource_enable: - enable clocksource
- * @cs:		pointer to clocksource
- *
- * Enables the specified clocksource. The clocksource callback
- * function should start up the hardware and setup mult and field
- * members of struct clocksource to reflect hardware capabilities.
- */
-static inline int clocksource_enable(struct clocksource *cs)
-{
-	int ret = 0;
-
-	if (cs->enable)
-		ret = cs->enable(cs);
-
-	/*
-	 * The frequency may have changed while the clocksource
-	 * was disabled. If so the code in ->enable() must update
-	 * the mult value to reflect the new frequency. Make sure
-	 * mult_orig follows this change.
-	 */
-	cs->mult_orig = cs->mult;
-
-	return ret;
-}
-
-/**
- * clocksource_disable: - disable clocksource
- * @cs:		pointer to clocksource
- *
- * Disables the specified clocksource. The clocksource callback
- * function should power down the now unused hardware block to
- * save power.
- */
-static inline void clocksource_disable(struct clocksource *cs)
-{
-	/*
-	 * Save mult_orig in mult so clocksource_enable() can
-	 * restore the value regardless if ->enable() updates
-	 * the value of mult or not.
-	 */
-	cs->mult = cs->mult_orig;
-
-	if (cs->disable)
-		cs->disable(cs);
-}
-
-/**
  * cyc2ns - converts clocksource cycles to nanoseconds
  * @cs:		Pointer to clocksource
  * @cycles:	Cycles

103 lines not on this page · raw message

Index: linux-2.6/arch/x86/kernel/tsc.c
===================================================================
@@ -744,10 +744,16 @@ static cycle_t __vsyscall_fn vread_tsc(v
 }
 #endif
 
+static void resume_tsc(void)
+{
+	clocksource_tsc.cycle_last = 0;
+}
+
 static struct clocksource clocksource_tsc = {
 	.name                   = "tsc",
 	.rating                 = 300,
 	.read                   = read_tsc,
+	.resume			= resume_tsc,
 	.mask                   = CLOCKSOURCE_MASK(64),
 	.shift                  = 22,
 	.flags                  = CLOCK_SOURCE_IS_CONTINUOUS |

Index: linux-2.6/kernel/time/timekeeping.c
===================================================================
@@ -295,7 +295,6 @@ static void change_clocksource(void)
 	if (old->disable)
 		old->disable(old);
 
-	clock->cycle_last = 0;
 	clock->cycle_last = clock->read(clock);
 	clock->error = 0;
 	clock->xtime_nsec = 0;

index 71f4368..9684254 100644
@@ -744,10 +744,16 @@ static cycle_t __vsyscall_fn vread_tsc(void)
 }
 #endif
 
+static void resume_tsc(void)
+{
+	clocksource_tsc.cycle_last = 0;
+}
+
 static struct clocksource clocksource_tsc = {
 	.name                   = "tsc",
 	.rating                 = 300,
 	.read                   = read_tsc,
+	.resume			= resume_tsc,
 	.mask                   = CLOCKSOURCE_MASK(64),
 	.shift                  = 22,
 	.flags                  = CLOCK_SOURCE_IS_CONTINUOUS |

index 016a259..b567301 100644
@@ -295,7 +295,6 @@ static void change_clocksource(void)
 	if (old->disable)
 		old->disable(old);
 
-	clock->cycle_last = 0;
 	clock->cycle_last = clock->read(clock);
 	clock->error = 0;
 	clock->xtime_nsec = 0;

Index: linux-2.6/kernel/time/clocksource.c
===================================================================
@@ -21,7 +21,6 @@
  *
  * TODO WishList:
  *   o Allow clocksource drivers to be unregistered
- *   o get rid of clocksource_jiffies extern
  */
 
 #include <linux/clocksource.h>

@@ -107,12 +106,9 @@ u64 timecounter_cyc2time(struct timecoun
 }
 EXPORT_SYMBOL(timecounter_cyc2time);
 
-/* XXX - Would like a better way for initializing curr_clocksource */
-extern struct clocksource clocksource_jiffies;
-
 /*[Clocksource internal variables]---------
  * curr_clocksource:
- *	currently selected clocksource. Initialized to clocksource_jiffies.
+ *	currently selected clocksource.
  * next_clocksource:
  *	pending next selected clocksource.
  * clocksource_list:

@@ -123,9 +119,8 @@ extern struct clocksource clocksource_ji
  * override_name:
  *	Name of the user-specified clocksource.
  */
-static struct clocksource *curr_clocksource = &clocksource_jiffies;
+static struct clocksource *curr_clocksource;
 static struct clocksource *next_clocksource;
-static struct clocksource *clocksource_override;
 static LIST_HEAD(clocksource_list);
 static DEFINE_SPINLOCK(clocksource_lock);
 static char override_name[32];

@@ -320,6 +315,7 @@ void clocksource_touch_watchdog(void)
 	clocksource_resume_watchdog();
 }
 
+#ifdef CONFIG_GENERIC_TIME
 /**
  * clocksource_get_next - Returns the selected clocksource
  *

@@ -339,56 +335,65 @@ struct clocksource *clocksource_get_next
 }
 
 /**
- * select_clocksource - Selects the best registered clocksource.
+ * clocksource_select - Select the best clocksource available
  *
  * Private function. Must hold clocksource_lock when called.
  *
  * Select the clocksource with the best rating, or the clocksource,
  * which is selected by userspace override.
  */
-static struct clocksource *select_clocksource(void)
+static void clocksource_select(void)
 {
-	struct clocksource *next;
+	struct clocksource *best, *cs;
 
 	if (list_empty(&clocksource_list))
-		return NULL;
+		return;
+	/* First clocksource on the list has the best rating. */
+	best = list_first_entry(&clocksource_list, struct clocksource, list);
+	/* Check for the override clocksource. */
+	list_for_each_entry(cs, &clocksource_list, list) {
+		if (strcmp(cs->name, override_name) != 0)
+			continue;
+		/*
+		 * Check to make sure we don't switch to a non-highres
+		 * capable clocksource if the tick code is in oneshot
+		 * mode (highres or nohz)
+		 */
+		if (!(cs->flags & CLOCK_SOURCE_VALID_FOR_HRES) &&
+		    tick_oneshot_mode_active()) {
+			/* Override clocksource cannot be used. */
+			printk(KERN_WARNING "Override clocksource %s is not "
+			       "HRT compatible. Cannot switch while in "
+			       "HRT/NOHZ mode\n", cs->name);
+			override_name[0] = 0;
+		} else
+			/* Override clocksource can be used. */
+			best = cs;
+		break;
+	}
+	if (curr_clocksource != best)
+		next_clocksource = best;
+}
 
-	if (clocksource_override)
-		next = clocksource_override;
-	else
-		next = list_entry(clocksource_list.next, struct clocksource,
-				  list);
+#else /* CONFIG_GENERIC_TIME */
 
-	if (next == curr_clocksource)
-		return NULL;
+static void clocksource_select(void) { }
 
-	return next;
-}
+#endif
 
 /*
  * Enqueue the clocksource sorted by rating
  */
-static int clocksource_enqueue(struct clocksource *c)
+static void clocksource_enqueue(struct clocksource *cs)
 {
-	struct list_head *tmp, *entry = &clocksource_list;
-
-	list_for_each(tmp, &clocksource_list) {
-		struct clocksource *cs;
+	struct list_head *entry = &clocksource_list;
+	struct clocksource *tmp;
 
-		cs = list_entry(tmp, struct clocksource, list);
-		if (cs == c)
-			return -EBUSY;
+	list_for_each_entry(tmp, &clocksource_list, list)
 		/* Keep track of the place, where to insert */
-		if (cs->rating >= c->rating)
-			entry = tmp;
-	}
-	list_add(&c->list, entry);
-
-	if (strlen(c->name) == strlen(override_name) &&
-	    !strcmp(c->name, override_name))
-		clocksource_override = c;
-
-	return 0;
+		if (tmp->rating >= cs->rating)
+			entry = &tmp->list;
+	list_add(&cs->list, entry);
 }
 
 /**

@@ -397,19 +402,16 @@ static int clocksource_enqueue(struct cl
  *
  * Returns -EBUSY if registration fails, zero otherwise.
  */
-int clocksource_register(struct clocksource *c)
+int clocksource_register(struct clocksource *cs)
 {
 	unsigned long flags;
-	int ret;
 
 	spin_lock_irqsave(&clocksource_lock, flags);
-	ret = clocksource_enqueue(c);
-	if (!ret)
-		next_clocksource = select_clocksource();
+	clocksource_enqueue(cs);
+	clocksource_select();
 	spin_unlock_irqrestore(&clocksource_lock, flags);
-	if (!ret)
-		clocksource_check_watchdog(c);
-	return ret;
+	clocksource_check_watchdog(cs);
+	return 0;
 }
 EXPORT_SYMBOL(clocksource_register);
 

@@ -425,7 +427,7 @@ void clocksource_change_rating(struct cl
 	list_del(&cs->list);
 	cs->rating = rating;
 	clocksource_enqueue(cs);
-	next_clocksource = select_clocksource();
+	clocksource_select();
 	spin_unlock_irqrestore(&clocksource_lock, flags);
 }
 

@@ -438,9 +440,7 @@ void clocksource_unregister(struct clock
 
 	spin_lock_irqsave(&clocksource_lock, flags);
 	list_del(&cs->list);
-	if (clocksource_override == cs)
-		clocksource_override = NULL;
-	next_clocksource = select_clocksource();
+	clocksource_select();
 	spin_unlock_irqrestore(&clocksource_lock, flags);
 }
 

@@ -478,10 +478,6 @@ static ssize_t sysfs_override_clocksourc
 					  struct sysdev_attribute *attr,
 					  const char *buf, size_t count)
 {
-	struct clocksource *ovr = NULL;
-	size_t ret = count;
-	int len;
-
 	/* strings from sysfs write are not 0 terminated! */
 	if (count >= sizeof(override_name))
 		return -EINVAL;

@@ -495,41 +491,11 @@ static ssize_t sysfs_override_clocksourc
 	if (count > 0)
 		memcpy(override_name, buf, count);
 	override_name[count] = 0;
-
-	len = strlen(override_name);
-	if (len) {
-		struct clocksource *cs;
-
-		ovr = clocksource_override;
-		/* try to select it: */
-		list_for_each_entry(cs, &clocksource_list, list) {
-			if (strlen(cs->name) == len &&
-			    !strcmp(cs->name, override_name))
-				ovr = cs;
-		}
-	}
-
-	/*
-	 * Check to make sure we don't switch to a non-highres capable
-	 * clocksource if the tick code is in oneshot mode (highres or nohz)
-	 */
-	if (tick_oneshot_mode_active() && ovr &&
-	    !(ovr->flags & CLOCK_SOURCE_VALID_FOR_HRES)) {
-		printk(KERN_WARNING "%s clocksource is not HRT compatible. "
-			"Cannot switch while in HRT/NOHZ mode\n", ovr->name);
-		ovr = NULL;
-		override_name[0] = 0;
-	}
-
-	/* Reselect, when the override name has changed */
-	if (ovr != clocksource_override) {
-		clocksource_override = ovr;
-		next_clocksource = select_clocksource();
-	}
+	clocksource_select();
 
 	spin_unlock_irq(&clocksource_lock);
 
-	return ret;
+	return count;
 }
 
 /**

Index: linux-2.6/include/linux/clocksource.h
===================================================================
@@ -14,6 +14,7 @@
 #include <linux/list.h>
 #include <linux/cache.h>
 #include <linux/timer.h>
+#include <linux/init.h>
 #include <asm/div64.h>
 #include <asm/io.h>
 

@@ -322,6 +323,7 @@ extern void clocksource_touch_watchdog(v
 extern struct clocksource* clocksource_get_next(void);
 extern void clocksource_change_rating(struct clocksource *cs, int rating);
 extern void clocksource_resume(void);
+extern struct clocksource * __init __weak clocksource_default_clock(void);
 
 #ifdef CONFIG_GENERIC_TIME_VSYSCALL
 extern void update_vsyscall(struct timespec *ts, struct clocksource *c);

Index: linux-2.6/kernel/time/timekeeping.c
===================================================================
@@ -269,7 +269,7 @@ static void change_clocksource(void)
 
 	new = clocksource_get_next();
 
-	if (clock == new)
+	if (!new || clock == new)
 		return;
 
 	clocksource_forward_now();

@@ -446,7 +446,7 @@ void __init timekeeping_init(void)
 
 	ntp_init();
 
-	clock = clocksource_get_next();
+	clock = clocksource_default_clock();
 	if (clock->enable)
 		clock->enable(clock);
 	/* set mult_orig on enable */

Index: linux-2.6/kernel/time/jiffies.c
===================================================================
@@ -61,7 +61,6 @@ struct clocksource clocksource_jiffies =
 	.read		= jiffies_read,
 	.mask		= 0xffffffff, /*32bits*/
 	.mult		= NSEC_PER_JIFFY << JIFFIES_SHIFT, /* details above */
-	.mult_orig	= NSEC_PER_JIFFY << JIFFIES_SHIFT,
 	.shift		= JIFFIES_SHIFT,
 };
 

@@ -71,3 +70,8 @@ static int __init init_jiffies_clocksour
 }
 
 core_initcall(init_jiffies_clocksource);
+
+struct clocksource * __init __weak clocksource_default_clock(void)
+{
+	return &clocksource_jiffies;
+}

Index: linux-2.6/arch/s390/kernel/time.c
===================================================================
@@ -205,6 +205,10 @@ static struct clocksource clocksource_to
 	.flags		= CLOCK_SOURCE_IS_CONTINUOUS,
 };
 
+struct clocksource * __init clocksource_default_clock(void)
+{
+	return &clocksource_tod;
+}
 
 void update_vsyscall(struct timespec *wall_time, struct clocksource *clock)
 {

index d4c8e9c..afefe51 100644
@@ -205,6 +205,10 @@ static struct clocksource clocksource_tod = {
 	.flags		= CLOCK_SOURCE_IS_CONTINUOUS,
 };
 
+struct clocksource * __init clocksource_default_clock(void)
+{
+	return &clocksource_tod;
+}
 
 void update_vsyscall(struct timespec *wall_time, struct clocksource *clock)
 {

index a1ef46f..f263b3a 100644
@@ -14,6 +14,7 @@
 #include <linux/list.h>
 #include <linux/cache.h>
 #include <linux/timer.h>
+#include <linux/init.h>
 #include <asm/div64.h>
 #include <asm/io.h>
 

@@ -322,6 +323,7 @@ extern void clocksource_touch_watchdog(void);
 extern struct clocksource* clocksource_get_next(void);
 extern void clocksource_change_rating(struct clocksource *cs, int rating);
 extern void clocksource_resume(void);
+extern struct clocksource * __init __weak clocksource_default_clock(void);
 
 #ifdef CONFIG_GENERIC_TIME_VSYSCALL
 extern void update_vsyscall(struct timespec *ts, struct clocksource *c);

222 lines not on this page · raw message

index c3f6c30..5404a84 100644
@@ -61,7 +61,6 @@ struct clocksource clocksource_jiffies = {
 	.read		= jiffies_read,
 	.mask		= 0xffffffff, /*32bits*/
 	.mult		= NSEC_PER_JIFFY << JIFFIES_SHIFT, /* details above */
-	.mult_orig	= NSEC_PER_JIFFY << JIFFIES_SHIFT,
 	.shift		= JIFFIES_SHIFT,
 };
 

@@ -71,3 +70,8 @@ static int __init init_jiffies_clocksource(void)
 }
 
 core_initcall(init_jiffies_clocksource);
+
+struct clocksource * __init __weak clocksource_default_clock(void)
+{
+	return &clocksource_jiffies;
+}

index b567301..325a9b6 100644
@@ -269,7 +269,7 @@ static void change_clocksource(void)
 
 	new = clocksource_get_next();
 
-	if (clock == new)
+	if (!new || clock == new)
 		return;
 
 	clocksource_forward_now();

@@ -446,7 +446,7 @@ void __init timekeeping_init(void)
 
 	ntp_init();
 
-	clock = clocksource_get_next();
+	clock = clocksource_default_clock();
 	if (clock->enable)
 		clock->enable(clock);
 	/* set mult_orig on enable */

Index: linux-2.6/kernel/time/clocksource.c
===================================================================
@@ -153,11 +153,8 @@ static unsigned long watchdog_resumed;
 #define WATCHDOG_INTERVAL (HZ >> 1)
 #define WATCHDOG_THRESHOLD (NSEC_PER_SEC >> 4)
 
-static void clocksource_ratewd(struct clocksource *cs, int64_t delta)
+static void clocksource_unstable(struct clocksource *cs, int64_t delta)
 {
-	if (delta > -WATCHDOG_THRESHOLD && delta < WATCHDOG_THRESHOLD)
-		return;
-
 	printk(KERN_WARNING "Clocksource %s unstable (delta = %Ld ns)\n",
 	       cs->name, delta);
 	cs->flags &= ~(CLOCK_SOURCE_VALID_FOR_HRES | CLOCK_SOURCE_WATCHDOG);

@@ -183,31 +180,31 @@ static void clocksource_watchdog(unsigne
 	list_for_each_entry_safe(cs, tmp, &watchdog_list, wd_list) {
 		csnow = cs->read(cs);
 
-		if (unlikely(resumed)) {
+		/* Clocksource initialized ? */
+		if (!(cs->flags & CLOCK_SOURCE_WATCHDOG)) {
+			cs->flags |= CLOCK_SOURCE_WATCHDOG;
 			cs->wd_last = csnow;
 			continue;
 		}
 
-		/* Initialized ? */
-		if (!(cs->flags & CLOCK_SOURCE_WATCHDOG)) {
-			if ((cs->flags & CLOCK_SOURCE_IS_CONTINUOUS) &&
-			    (watchdog->flags & CLOCK_SOURCE_IS_CONTINUOUS)) {
-				cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES;
-				/*
-				 * We just marked the clocksource as
-				 * highres-capable, notify the rest of the
-				 * system as well so that we transition
-				 * into high-res mode:
-				 */
-				tick_clock_notify();
-			}
-			cs->flags |= CLOCK_SOURCE_WATCHDOG;
-			cs->wd_last = csnow;
-		} else {
-			cs_nsec = cyc2ns(cs, (csnow - cs->wd_last) & cs->mask);
-			cs->wd_last = csnow;
-			/* Check the delta. Might remove from the list ! */
-			clocksource_ratewd(cs, cs_nsec - wd_nsec);
+		/* Check the deviation from the watchdog clocksource. */
+		cs_nsec = cyc2ns(cs, (csnow - cs->wd_last) & cs->mask);
+		cs->wd_last = csnow;
+		if (abs(cs_nsec - wd_nsec) > WATCHDOG_THRESHOLD) {
+			clocksource_unstable(cs, cs_nsec - wd_nsec);
+			continue;
+		}
+
+		if (!(cs->flags & CLOCK_SOURCE_VALID_FOR_HRES) &&
+		    (cs->flags & CLOCK_SOURCE_IS_CONTINUOUS) &&
+		    (watchdog->flags & CLOCK_SOURCE_IS_CONTINUOUS)) {
+			cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES;
+			/*
+			 * We just marked the clocksource as highres-capable,
+			 * notify the rest of the system as well so that we
+			 * transition into high-res mode:
+			 */
+			tick_clock_notify();
 		}
 	}
 

index e91662e..76256c5 100644
@@ -153,11 +153,8 @@ static unsigned long watchdog_resumed;
 #define WATCHDOG_INTERVAL (HZ >> 1)
 #define WATCHDOG_THRESHOLD (NSEC_PER_SEC >> 4)
 
-static void clocksource_ratewd(struct clocksource *cs, int64_t delta)
+static void clocksource_unstable(struct clocksource *cs, int64_t delta)
 {
-	if (delta > -WATCHDOG_THRESHOLD && delta < WATCHDOG_THRESHOLD)
-		return;
-
 	printk(KERN_WARNING "Clocksource %s unstable (delta = %Ld ns)\n",
 	       cs->name, delta);
 	cs->flags &= ~(CLOCK_SOURCE_VALID_FOR_HRES | CLOCK_SOURCE_WATCHDOG);

@@ -183,31 +180,31 @@ static void clocksource_watchdog(unsigned long data)
 	list_for_each_entry_safe(cs, tmp, &watchdog_list, wd_list) {
 		csnow = cs->read(cs);
 
-		if (unlikely(resumed)) {
+		/* Clocksource initialized ? */
+		if (!(cs->flags & CLOCK_SOURCE_WATCHDOG)) {
+			cs->flags |= CLOCK_SOURCE_WATCHDOG;
 			cs->wd_last = csnow;
 			continue;
 		}
 
-		/* Initialized ? */
-		if (!(cs->flags & CLOCK_SOURCE_WATCHDOG)) {
-			if ((cs->flags & CLOCK_SOURCE_IS_CONTINUOUS) &&
-			    (watchdog->flags & CLOCK_SOURCE_IS_CONTINUOUS)) {
-				cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES;
-				/*
-				 * We just marked the clocksource as
-				 * highres-capable, notify the rest of the
-				 * system as well so that we transition
-				 * into high-res mode:
-				 */
-				tick_clock_notify();
-			}
-			cs->flags |= CLOCK_SOURCE_WATCHDOG;
-			cs->wd_last = csnow;
-		} else {
-			cs_nsec = cyc2ns(cs, (csnow - cs->wd_last) & cs->mask);
-			cs->wd_last = csnow;
-			/* Check the delta. Might remove from the list ! */
-			clocksource_ratewd(cs, cs_nsec - wd_nsec);
+		/* Check the deviation from the watchdog clocksource. */
+		cs_nsec = cyc2ns(cs, (csnow - cs->wd_last) & cs->mask);
+		cs->wd_last = csnow;
+		if (abs(cs_nsec - wd_nsec) > WATCHDOG_THRESHOLD) {
+			clocksource_unstable(cs, cs_nsec - wd_nsec);
+			continue;
+		}
+
+		if (!(cs->flags & CLOCK_SOURCE_VALID_FOR_HRES) &&
+		    (cs->flags & CLOCK_SOURCE_IS_CONTINUOUS) &&
+		    (watchdog->flags & CLOCK_SOURCE_IS_CONTINUOUS)) {
+			cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES;
+			/*
+			 * We just marked the clocksource as highres-capable,
+			 * notify the rest of the system as well so that we
+			 * transition into high-res mode:
+			 */
+			tick_clock_notify();
 		}
 	}
 

Index: linux-2.6/kernel/time/clocksource.c
===================================================================
@@ -145,7 +145,6 @@ static struct clocksource *watchdog;
 static struct timer_list watchdog_timer;
 static DEFINE_SPINLOCK(watchdog_lock);
 static cycle_t watchdog_last;
-static unsigned long watchdog_resumed;
 
 /*
  * Interval: 0.5sec Threshold: 0.0625s

@@ -167,12 +166,9 @@ static void clocksource_watchdog(unsigne
 	struct clocksource *cs, *tmp;
 	cycle_t csnow, wdnow;
 	int64_t wd_nsec, cs_nsec;
-	int resumed;
 
 	spin_lock(&watchdog_lock);
 
-	resumed = test_and_clear_bit(0, &watchdog_resumed);
-
 	wdnow = watchdog->read(watchdog);
 	wd_nsec = cyc2ns(watchdog, (wdnow - watchdog_last) & watchdog->mask);
 	watchdog_last = wdnow;

@@ -223,14 +219,26 @@ static void clocksource_watchdog(unsigne
 	}
 	spin_unlock(&watchdog_lock);
 }
+
+static inline void clocksource_reset_watchdog(void)
+{
+	struct clocksource *cs;
+
+	list_for_each_entry(cs, &watchdog_list, wd_list)
+		cs->flags &= ~CLOCK_SOURCE_WATCHDOG;
+}
+
 static void clocksource_resume_watchdog(void)
 {
-	set_bit(0, &watchdog_resumed);
+	unsigned long flags;
+
+	spin_lock_irqsave(&watchdog_lock, flags);
+	clocksource_reset_watchdog();
+	spin_unlock_irqrestore(&watchdog_lock, flags);
 }
 
 static void clocksource_check_watchdog(struct clocksource *cs)
 {
-	struct clocksource *cse;
 	unsigned long flags;
 
 	spin_lock_irqsave(&watchdog_lock, flags);

@@ -256,8 +264,7 @@ static void clocksource_check_watchdog(s
 			watchdog_timer.function = clocksource_watchdog;
 
 			/* Reset watchdog cycles */
-			list_for_each_entry(cse, &watchdog_list, wd_list)
-				cse->flags &= ~CLOCK_SOURCE_WATCHDOG;
+			clocksource_reset_watchdog();
 			/* Start if list is not empty */
 			if (!list_empty(&watchdog_list)) {
 				watchdog_last = watchdog->read(watchdog);

index 76256c5..89a7b91 100644
@@ -145,7 +145,6 @@ static struct clocksource *watchdog;
 static struct timer_list watchdog_timer;
 static DEFINE_SPINLOCK(watchdog_lock);
 static cycle_t watchdog_last;
-static unsigned long watchdog_resumed;
 
 /*
  * Interval: 0.5sec Threshold: 0.0625s

@@ -167,12 +166,9 @@ static void clocksource_watchdog(unsigned long data)
 	struct clocksource *cs, *tmp;
 	cycle_t csnow, wdnow;
 	int64_t wd_nsec, cs_nsec;
-	int resumed;
 
 	spin_lock(&watchdog_lock);
 
-	resumed = test_and_clear_bit(0, &watchdog_resumed);
-
 	wdnow = watchdog->read(watchdog);
 	wd_nsec = cyc2ns(watchdog, (wdnow - watchdog_last) & watchdog->mask);
 	watchdog_last = wdnow;

@@ -223,14 +219,26 @@ static void clocksource_watchdog(unsigned long data)
 	}
 	spin_unlock(&watchdog_lock);
 }
+
+static inline void clocksource_reset_watchdog(void)
+{
+	struct clocksource *cs;
+
+	list_for_each_entry(cs, &watchdog_list, wd_list)
+		cs->flags &= ~CLOCK_SOURCE_WATCHDOG;
+}
+
 static void clocksource_resume_watchdog(void)
 {
-	set_bit(0, &watchdog_resumed);
+	unsigned long flags;
+
+	spin_lock_irqsave(&watchdog_lock, flags);
+	clocksource_reset_watchdog();
+	spin_unlock_irqrestore(&watchdog_lock, flags);
 }
 
 static void clocksource_check_watchdog(struct clocksource *cs)
 {
-	struct clocksource *cse;
 	unsigned long flags;
 
 	spin_lock_irqsave(&watchdog_lock, flags);

@@ -256,8 +264,7 @@ static void clocksource_check_watchdog(struct clocksource *cs)
 			watchdog_timer.function = clocksource_watchdog;
 
 			/* Reset watchdog cycles */
-			list_for_each_entry(cse, &watchdog_list, wd_list)
-				cse->flags &= ~CLOCK_SOURCE_WATCHDOG;
+			clocksource_reset_watchdog();
 			/* Start if list is not empty */
 			if (!list_empty(&watchdog_list)) {
 				watchdog_last = watchdog->read(watchdog);

Index: linux-2.6/kernel/time/clocksource.c
===================================================================
@@ -145,6 +145,7 @@ static struct clocksource *watchdog;
 static struct timer_list watchdog_timer;
 static DEFINE_SPINLOCK(watchdog_lock);
 static cycle_t watchdog_last;
+static int watchdog_running;
 
 /*
  * Interval: 0.5sec Threshold: 0.0625s

@@ -168,6 +169,8 @@ static void clocksource_watchdog(unsigne
 	int64_t wd_nsec, cs_nsec;
 
 	spin_lock(&watchdog_lock);
+	if (!watchdog_running)
+		goto out;
 
 	wdnow = watchdog->read(watchdog);
 	wd_nsec = cyc2ns(watchdog, (wdnow - watchdog_last) & watchdog->mask);

@@ -217,9 +220,30 @@ static void clocksource_watchdog(unsigne
 		watchdog_timer.expires += WATCHDOG_INTERVAL;
 		add_timer_on(&watchdog_timer, next_cpu);
 	}
+out:
 	spin_unlock(&watchdog_lock);
 }
 
+static inline void clocksource_start_watchdog(void)
+{
+	if (watchdog_running || !watchdog || list_empty(&watchdog_list))
+		return;
+	init_timer(&watchdog_timer);
+	watchdog_timer.function = clocksource_watchdog;
+	watchdog_last = watchdog->read(watchdog);
+	watchdog_timer.expires = jiffies + WATCHDOG_INTERVAL;
+	add_timer_on(&watchdog_timer, cpumask_first(cpu_online_mask));
+	watchdog_running = 1;
+}
+
+static inline void clocksource_stop_watchdog(void)
+{
+	if (!watchdog_running || (watchdog && !list_empty(&watchdog_list)))
+		return;
+	del_timer(&watchdog_timer);
+	watchdog_running = 0;
+}
+
 static inline void clocksource_reset_watchdog(void)
 {
 	struct clocksource *cs;

@@ -237,55 +261,70 @@ static void clocksource_resume_watchdog(
 	spin_unlock_irqrestore(&watchdog_lock, flags);
 }
 
-static void clocksource_check_watchdog(struct clocksource *cs)
+static void clocksource_enqueue_watchdog(struct clocksource *cs)
 {
 	unsigned long flags;
 
 	spin_lock_irqsave(&watchdog_lock, flags);
 	if (cs->flags & CLOCK_SOURCE_MUST_VERIFY) {
-		int started = !list_empty(&watchdog_list);
-
+		/* cs is a clocksource to be watched. */
 		list_add(&cs->wd_list, &watchdog_list);
-		if (!started && watchdog) {
-			watchdog_last = watchdog->read(watchdog);
-			watchdog_timer.expires = jiffies + WATCHDOG_INTERVAL;
-			add_timer_on(&watchdog_timer,
-				     cpumask_first(cpu_online_mask));
-		}
+		cs->flags &= ~CLOCK_SOURCE_WATCHDOG;
 	} else {
+		/* cs is a watchdog. */
 		if (cs->flags & CLOCK_SOURCE_IS_CONTINUOUS)
 			cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES;
-
+		/* Pick the best watchdog. */
 		if (!watchdog || cs->rating > watchdog->rating) {
-			if (watchdog)
-				del_timer(&watchdog_timer);
 			watchdog = cs;
-			init_timer(&watchdog_timer);
-			watchdog_timer.function = clocksource_watchdog;
-
 			/* Reset watchdog cycles */
 			clocksource_reset_watchdog();
-			/* Start if list is not empty */
-			if (!list_empty(&watchdog_list)) {
-				watchdog_last = watchdog->read(watchdog);
-				watchdog_timer.expires =
-					jiffies + WATCHDOG_INTERVAL;
-				add_timer_on(&watchdog_timer,
-					     cpumask_first(cpu_online_mask));
-			}
 		}
 	}
+	/* Check if the watchdog timer needs to be started. */
+	clocksource_start_watchdog();
 	spin_unlock_irqrestore(&watchdog_lock, flags);
 }
-#else
-static void clocksource_check_watchdog(struct clocksource *cs)
+
+static void clocksource_dequeue_watchdog(struct clocksource *cs)
+{
+	struct clocksource *tmp;
+	unsigned long flags;
+
+	spin_lock_irqsave(&watchdog_lock, flags);
+	if (cs->flags & CLOCK_SOURCE_MUST_VERIFY) {
+		/* cs is a watched clocksource. */
+		list_del_init(&cs->wd_list);
+	} else if (cs == watchdog) {
+		/* Reset watchdog cycles */
+		clocksource_reset_watchdog();
+		/* Current watchdog is removed. Find an alternative. */
+		watchdog = NULL;
+		list_for_each_entry(tmp, &clocksource_list, list) {
+			if (tmp == cs || tmp->flags & CLOCK_SOURCE_MUST_VERIFY)
+				continue;
+			if (!watchdog || tmp->rating > watchdog->rating)
+				watchdog = tmp;
+		}
+	}
+	cs->flags &= ~CLOCK_SOURCE_WATCHDOG;
+	/* Check if the watchdog timer needs to be stopped. */
+	clocksource_stop_watchdog();
+	spin_unlock_irqrestore(&watchdog_lock, flags);
+}
+
+#else /* CONFIG_CLOCKSOURCE_WATCHDOG */
+
+static void clocksource_enqueue_watchdog(struct clocksource *cs)
 {
 	if (cs->flags & CLOCK_SOURCE_IS_CONTINUOUS)
 		cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES;
 }
 
+static inline void clocksource_dequeue_watchdog(struct clocksource *cs) { }
 static inline void clocksource_resume_watchdog(void) { }
-#endif
+
+#endif /* CONFIG_CLOCKSOURCE_WATCHDOG */
 
 /**
  * clocksource_resume - resume the clocksource(s)

@@ -414,14 +453,13 @@ int clocksource_register(struct clocksou
 	clocksource_enqueue(cs);
 	clocksource_select();
 	spin_unlock_irqrestore(&clocksource_lock, flags);
-	clocksource_check_watchdog(cs);
+	clocksource_enqueue_watchdog(cs);
 	return 0;
 }
 EXPORT_SYMBOL(clocksource_register);
 
 /**
  * clocksource_change_rating - Change the rating of a registered clocksource
- *
  */
 void clocksource_change_rating(struct clocksource *cs, int rating)
 {

@@ -434,6 +472,7 @@ void clocksource_change_rating(struct cl
 	clocksource_select();
 	spin_unlock_irqrestore(&clocksource_lock, flags);
 }
+EXPORT_SYMBOL(clocksource_change_rating);
 
 /**
  * clocksource_unregister - remove a registered clocksource

@@ -442,11 +481,13 @@ void clocksource_unregister(struct clock
 {
 	unsigned long flags;
 
+	clocksource_dequeue_watchdog(cs);
 	spin_lock_irqsave(&clocksource_lock, flags);
 	list_del(&cs->list);
 	clocksource_select();
 	spin_unlock_irqrestore(&clocksource_lock, flags);
 }
+EXPORT_SYMBOL(clocksource_unregister);
 
 #ifdef CONFIG_SYSFS
 /**

index 89a7b91..56aaa74 100644
@@ -145,6 +145,7 @@ static struct clocksource *watchdog;
 static struct timer_list watchdog_timer;
 static DEFINE_SPINLOCK(watchdog_lock);
 static cycle_t watchdog_last;
+static int watchdog_running;
 
 /*
  * Interval: 0.5sec Threshold: 0.0625s

@@ -168,6 +169,8 @@ static void clocksource_watchdog(unsigned long data)
 	int64_t wd_nsec, cs_nsec;
 
 	spin_lock(&watchdog_lock);
+	if (!watchdog_running)
+		goto out;
 
 	wdnow = watchdog->read(watchdog);
 	wd_nsec = cyc2ns(watchdog, (wdnow - watchdog_last) & watchdog->mask);

@@ -217,9 +220,30 @@ static void clocksource_watchdog(unsigned long data)
 		watchdog_timer.expires += WATCHDOG_INTERVAL;
 		add_timer_on(&watchdog_timer, next_cpu);
 	}
+out:
 	spin_unlock(&watchdog_lock);
 }
 
+static inline void clocksource_start_watchdog(void)
+{
+	if (watchdog_running || !watchdog || list_empty(&watchdog_list))
+		return;
+	init_timer(&watchdog_timer);
+	watchdog_timer.function = clocksource_watchdog;
+	watchdog_last = watchdog->read(watchdog);
+	watchdog_timer.expires = jiffies + WATCHDOG_INTERVAL;
+	add_timer_on(&watchdog_timer, cpumask_first(cpu_online_mask));
+	watchdog_running = 1;
+}
+
+static inline void clocksource_stop_watchdog(void)
+{
+	if (!watchdog_running || (watchdog && !list_empty(&watchdog_list)))
+		return;
+	del_timer(&watchdog_timer);
+	watchdog_running = 0;
+}
+
 static inline void clocksource_reset_watchdog(void)
 {
 	struct clocksource *cs;

@@ -237,55 +261,70 @@ static void clocksource_resume_watchdog(void)
 	spin_unlock_irqrestore(&watchdog_lock, flags);
 }
 
-static void clocksource_check_watchdog(struct clocksource *cs)
+static void clocksource_enqueue_watchdog(struct clocksource *cs)
 {
 	unsigned long flags;
 
 	spin_lock_irqsave(&watchdog_lock, flags);
 	if (cs->flags & CLOCK_SOURCE_MUST_VERIFY) {
-		int started = !list_empty(&watchdog_list);
-
+		/* cs is a clocksource to be watched. */
 		list_add(&cs->wd_list, &watchdog_list);
-		if (!started && watchdog) {
-			watchdog_last = watchdog->read(watchdog);
-			watchdog_timer.expires = jiffies + WATCHDOG_INTERVAL;
-			add_timer_on(&watchdog_timer,
-				     cpumask_first(cpu_online_mask));
-		}
+		cs->flags &= ~CLOCK_SOURCE_WATCHDOG;
 	} else {
+		/* cs is a watchdog. */
 		if (cs->flags & CLOCK_SOURCE_IS_CONTINUOUS)
 			cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES;
-
+		/* Pick the best watchdog. */
 		if (!watchdog || cs->rating > watchdog->rating) {
-			if (watchdog)
-				del_timer(&watchdog_timer);
 			watchdog = cs;
-			init_timer(&watchdog_timer);
-			watchdog_timer.function = clocksource_watchdog;
-
 			/* Reset watchdog cycles */
 			clocksource_reset_watchdog();
-			/* Start if list is not empty */
-			if (!list_empty(&watchdog_list)) {
-				watchdog_last = watchdog->read(watchdog);
-				watchdog_timer.expires =
-					jiffies + WATCHDOG_INTERVAL;
-				add_timer_on(&watchdog_timer,
-					     cpumask_first(cpu_online_mask));
-			}
 		}
 	}
+	/* Check if the watchdog timer needs to be started. */
+	clocksource_start_watchdog();
 	spin_unlock_irqrestore(&watchdog_lock, flags);
 }
-#else
-static void clocksource_check_watchdog(struct clocksource *cs)
+
+static void clocksource_dequeue_watchdog(struct clocksource *cs)
+{
+	struct clocksource *tmp;
+	unsigned long flags;
+
+	spin_lock_irqsave(&watchdog_lock, flags);
+	if (cs->flags & CLOCK_SOURCE_MUST_VERIFY) {
+		/* cs is a watched clocksource. */
+		list_del_init(&cs->wd_list);
+	} else if (cs == watchdog) {
+		/* Reset watchdog cycles */
+		clocksource_reset_watchdog();
+		/* Current watchdog is removed. Find an alternative. */
+		watchdog = NULL;
+		list_for_each_entry(tmp, &clocksource_list, list) {
+			if (tmp == cs || tmp->flags & CLOCK_SOURCE_MUST_VERIFY)
+				continue;
+			if (!watchdog || tmp->rating > watchdog->rating)
+				watchdog = tmp;
+		}
+	}
+	cs->flags &= ~CLOCK_SOURCE_WATCHDOG;
+	/* Check if the watchdog timer needs to be stopped. */
+	clocksource_stop_watchdog();
+	spin_unlock_irqrestore(&watchdog_lock, flags);
+}
+
+#else /* CONFIG_CLOCKSOURCE_WATCHDOG */
+
+static void clocksource_enqueue_watchdog(struct clocksource *cs)
 {
 	if (cs->flags & CLOCK_SOURCE_IS_CONTINUOUS)
 		cs->flags |= CLOCK_SOURCE_VALID_FOR_HRES;
 }
 
+static inline void clocksource_dequeue_watchdog(struct clocksource *cs) { }
 static inline void clocksource_resume_watchdog(void) { }
-#endif
+
+#endif /* CONFIG_CLOCKSOURCE_WATCHDOG */
 
 /**
  * clocksource_resume - resume the clocksource(s)

@@ -414,14 +453,13 @@ int clocksource_register(struct clocksource *cs)
 	clocksource_enqueue(cs);
 	clocksource_select();
 	spin_unlock_irqrestore(&clocksource_lock, flags);
-	clocksource_check_watchdog(cs);
+	clocksource_enqueue_watchdog(cs);
 	return 0;
 }
 EXPORT_SYMBOL(clocksource_register);
 
 /**
  * clocksource_change_rating - Change the rating of a registered clocksource
- *
  */
 void clocksource_change_rating(struct clocksource *cs, int rating)
 {

@@ -434,6 +472,7 @@ void clocksource_change_rating(struct clocksource *cs, int rating)
 	clocksource_select();
 	spin_unlock_irqrestore(&clocksource_lock, flags);
 }
+EXPORT_SYMBOL(clocksource_change_rating);
 
 /**
  * clocksource_unregister - remove a registered clocksource

@@ -442,11 +481,13 @@ void clocksource_unregister(struct clocksource *cs)
 {
 	unsigned long flags;
 
+	clocksource_dequeue_watchdog(cs);
 	spin_lock_irqsave(&clocksource_lock, flags);
 	list_del(&cs->list);
 	clocksource_select();
 	spin_unlock_irqrestore(&clocksource_lock, flags);
 }
+EXPORT_SYMBOL(clocksource_unregister);
 
 #ifdef CONFIG_SYSFS
 /**

Index: linux-2.6/include/linux/clocksource.h
===================================================================
@@ -213,6 +213,7 @@ extern struct clocksource *clock;	/* cur
 
 #define CLOCK_SOURCE_WATCHDOG			0x10
 #define CLOCK_SOURCE_VALID_FOR_HRES		0x20
+#define CLOCK_SOURCE_UNSTABLE			0x40
 
 /* simplify initialization of mask field */
 #define CLOCKSOURCE_MASK(bits) (cycle_t)((bits) < 64 ? ((1ULL<<(bits))-1) : -1)

102 lines not on this page · raw message

index f263b3a..19ad43a 100644
@@ -213,6 +213,7 @@ extern struct clocksource *clock;	/* current clocksource */
 
 #define CLOCK_SOURCE_WATCHDOG			0x10
 #define CLOCK_SOURCE_VALID_FOR_HRES		0x20
+#define CLOCK_SOURCE_UNSTABLE			0x40
 
 /* simplify initialization of mask field */
 #define CLOCKSOURCE_MASK(bits) (cycle_t)((bits) < 64 ? ((1ULL<<(bits))-1) : -1)

102 lines not on this page · raw message

Index: linux-2.6/arch/s390/kernel/time.c
===================================================================
@@ -280,7 +280,6 @@ void __init time_init(void)
 	now = get_clock();
 	tod_to_timeval(now - TOD_UNIX_EPOCH, &xtime);
 	clocksource_tod.cycle_last = now;
-	clocksource_tod.raw_time = xtime;
 	tod_to_timeval(sched_clock_base_cc - TOD_UNIX_EPOCH, &ts);
 	set_normalized_timespec(&wall_to_monotonic, -ts.tv_sec, -ts.tv_nsec);
 	write_sequnlock_irqrestore(&xtime_lock, flags);

89 lines not on this page · raw message

Index: linux-2.6/kernel/time/clocksource.c
===================================================================
@@ -177,7 +177,8 @@ static void clocksource_watchdog(unsigne
 		goto out;
 
 	wdnow = watchdog->read(watchdog);
-	wd_nsec = cyc2ns(watchdog, (wdnow - watchdog_last) & watchdog->mask);
+	wd_nsec = clocksource_cyc2ns((wdnow - watchdog_last) & watchdog->mask,
+				     watchdog->mult, watchdog->shift);
 	watchdog_last = wdnow;
 
 	list_for_each_entry(cs, &watchdog_list, wd_list) {

@@ -196,7 +197,8 @@ static void clocksource_watchdog(unsigne
 		}
 
 		/* Check the deviation from the watchdog clocksource. */
-		cs_nsec = cyc2ns(cs, (csnow - cs->wd_last) & cs->mask);
+		cs_nsec = clocksource_cyc2ns((csnow - cs->wd_last) &
+					     cs->mask, cs->mult, cs->shift);
 		cs->wd_last = csnow;
 		if (abs(cs_nsec - wd_nsec) > WATCHDOG_THRESHOLD) {
 			clocksource_unstable(cs, cs_nsec - wd_nsec);

514 lines not on this page · raw message

index afefe51..e76c2e7 100644
@@ -280,7 +280,6 @@ void __init time_init(void)
 	now = get_clock();
 	tod_to_timeval(now - TOD_UNIX_EPOCH, &xtime);
 	clocksource_tod.cycle_last = now;
-	clocksource_tod.raw_time = xtime;
 	tod_to_timeval(sched_clock_base_cc - TOD_UNIX_EPOCH, &ts);
 	set_normalized_timespec(&wall_to_monotonic, -ts.tv_sec, -ts.tv_nsec);
 	write_sequnlock_irqrestore(&xtime_lock, flags);

89 lines not on this page · raw message

index f150801..f18c9a6 100644
@@ -177,7 +177,8 @@ static void clocksource_watchdog(unsigned long data)
 		goto out;
 
 	wdnow = watchdog->read(watchdog);
-	wd_nsec = cyc2ns(watchdog, (wdnow - watchdog_last) & watchdog->mask);
+	wd_nsec = clocksource_cyc2ns((wdnow - watchdog_last) & watchdog->mask,
+				     watchdog->mult, watchdog->shift);
 	watchdog_last = wdnow;
 
 	list_for_each_entry(cs, &watchdog_list, wd_list) {

@@ -196,7 +197,8 @@ static void clocksource_watchdog(unsigned long data)
 		}
 
 		/* Check the deviation from the watchdog clocksource. */
-		cs_nsec = cyc2ns(cs, (csnow - cs->wd_last) & cs->mask);
+		cs_nsec = clocksource_cyc2ns((csnow - cs->wd_last) &
+					     cs->mask, cs->mult, cs->shift);
 		cs->wd_last = csnow;
 		if (abs(cs_nsec - wd_nsec) > WATCHDOG_THRESHOLD) {
 			clocksource_unstable(cs, cs_nsec - wd_nsec);

516 lines not on this page · raw message

Index: linux-2.6/kernel/time/timekeeping.c
===================================================================
@@ -23,6 +23,8 @@
 struct timekeeper {
 	/* Current clocksource used for timekeeping. */
 	struct clocksource *clock;
+	/* The shift value of the current clocksource. */
+	int	shift;
 
 	/* Number of clock cycles in one NTP interval. */
 	cycle_t cycle_interval;

@@ -36,6 +38,9 @@ struct timekeeper {
 	/* Difference between accumulated time and NTP time in ntp
 	 * shifted nano seconds. */
 	s64	ntp_error;
+	/* Shift conversion between clock shifted nano seconds and
+	 * ntp shifted nano seconds. */
+	int	ntp_error_shift;
 };
 
 struct timekeeper timekeeper;

@@ -75,8 +80,10 @@ static void timekeeper_setup_internals(s
 		((u64) interval * clock->mult_orig) >> clock->shift;
 
 	timekeeper.xtime_nsec = 0;
+	timekeeper.shift = clock->shift;
 
 	timekeeper.ntp_error = 0;
+	timekeeper.ntp_error_shift = NTP_SCALE_SHIFT - clock->shift;
 }
 
 /*

@@ -641,8 +648,7 @@ static __always_inline int timekeeping_b
 	 * Now calculate the error in (1 << look_ahead) ticks, but first
 	 * remove the single look ahead already included in the error.
 	 */
-	tick_error = tick_length >>
-			(NTP_SCALE_SHIFT - timekeeper.clock->shift + 1);
+	tick_error = tick_length >> (timekeeper.ntp_error_shift + 1);
 	tick_error -= timekeeper.xtime_interval >> 1;
 	error = ((error - tick_error) >> look_ahead) + tick_error;
 

@@ -673,8 +679,7 @@ static void timekeeping_adjust(s64 offse
 	s64 error, interval = timekeeper.cycle_interval;
 	int adj;
 
-	error = timekeeper.ntp_error >>
-		(NTP_SCALE_SHIFT - timekeeper.clock->shift - 1);
+	error = timekeeper.ntp_error >> (timekeeper.ntp_error_shift - 1);
 	if (error > interval) {
 		error >>= 2;
 		if (likely(error <= interval))

@@ -696,7 +701,7 @@ static void timekeeping_adjust(s64 offse
 	timekeeper.xtime_interval += interval;
 	timekeeper.xtime_nsec -= offset;
 	timekeeper.ntp_error -= (interval - offset) <<
-			(NTP_SCALE_SHIFT - timekeeper.clock->shift);
+ 				timekeeper.ntp_error_shift;
 }
 
 /**

@@ -708,7 +713,7 @@ void update_wall_time(void)
 {
 	struct clocksource *clock;
 	cycle_t offset;
-	s64 nsecs;
+	u64 nsecs;
 
 	/* Make sure we're fully resumed: */
 	if (unlikely(timekeeping_suspended))

@@ -720,7 +725,7 @@ void update_wall_time(void)
 #else
 	offset = timekeeper.cycle_interval;
 #endif
-	timekeeper.xtime_nsec = (s64)xtime.tv_nsec << clock->shift;
+	timekeeper.xtime_nsec = (s64)xtime.tv_nsec << timekeeper.shift;
 
 	/* normally this loop will run just once, however in the
 	 * case of lost or late ticks, it will accumulate correctly.

@@ -731,8 +736,9 @@ void update_wall_time(void)
 		clock->cycle_last += timekeeper.cycle_interval;
 
 		timekeeper.xtime_nsec += timekeeper.xtime_interval;
-		if (timekeeper.xtime_nsec >= (u64)NSEC_PER_SEC << clock->shift) {
-			timekeeper.xtime_nsec -= (u64)NSEC_PER_SEC << clock->shift;
+		nsecs = (u64) NSEC_PER_SEC << timekeeper.shift;
+		if (timekeeper.xtime_nsec >= nsecs) {
+			timekeeper.xtime_nsec -= nsecs;
 			xtime.tv_sec++;
 			second_overflow();
 		}

@@ -746,7 +752,7 @@ void update_wall_time(void)
 		/* accumulate error between NTP and clock interval */
 		timekeeper.ntp_error += tick_length;
 		timekeeper.ntp_error -= timekeeper.xtime_interval <<
-					(NTP_SCALE_SHIFT - clock->shift);
+					timekeeper.ntp_error_shift;
 	}
 
 	/* correct the clock when NTP error is too big */

@@ -771,16 +777,16 @@ void update_wall_time(void)
 	if (unlikely((s64)timekeeper.xtime_nsec < 0)) {
 		s64 neg = -(s64)timekeeper.xtime_nsec;
 		timekeeper.xtime_nsec = 0;
-		timekeeper.ntp_error += neg << (NTP_SCALE_SHIFT - clock->shift);
+		timekeeper.ntp_error += neg << timekeeper.ntp_error_shift;
 	}
 
 	/* store full nanoseconds into xtime after rounding it up and
 	 * add the remainder to the error difference.
 	 */
-	xtime.tv_nsec = ((s64)timekeeper.xtime_nsec >> clock->shift) + 1;
-	timekeeper.xtime_nsec -= (s64)xtime.tv_nsec << clock->shift;
-	timekeeper.ntp_error += timekeeper.xtime_nsec <<
-				(NTP_SCALE_SHIFT - clock->shift);
+	xtime.tv_nsec =	((s64) timekeeper.xtime_nsec >> timekeeper.shift) + 1;
+	timekeeper.xtime_nsec -= (s64) xtime.tv_nsec << timekeeper.shift;
+	timekeeper.ntp_error +=	timekeeper.xtime_nsec <<
+				timekeeper.ntp_error_shift;
 
 	nsecs = clocksource_cyc2ns(offset, clock->mult, clock->shift);
 	update_xtime_cache(nsecs);

index 7af45cb..dfdab1c 100644
@@ -23,6 +23,8 @@
 struct timekeeper {
 	/* Current clocksource used for timekeeping. */
 	struct clocksource *clock;
+	/* The shift value of the current clocksource. */
+	int	shift;
 
 	/* Number of clock cycles in one NTP interval. */
 	cycle_t cycle_interval;

@@ -36,6 +38,9 @@ struct timekeeper {
 	/* Difference between accumulated time and NTP time in ntp
 	 * shifted nano seconds. */
 	s64	ntp_error;
+	/* Shift conversion between clock shifted nano seconds and
+	 * ntp shifted nano seconds. */
+	int	ntp_error_shift;
 };
 
 struct timekeeper timekeeper;

@@ -75,8 +80,10 @@ static void timekeeper_setup_internals(struct clocksource *clock)
 		((u64) interval * clock->mult_orig) >> clock->shift;
 
 	timekeeper.xtime_nsec = 0;
+	timekeeper.shift = clock->shift;
 
 	timekeeper.ntp_error = 0;
+	timekeeper.ntp_error_shift = NTP_SCALE_SHIFT - clock->shift;
 }
 
 /*

@@ -641,8 +648,7 @@ static __always_inline int timekeeping_bigadjust(s64 error, s64 *interval,
 	 * Now calculate the error in (1 << look_ahead) ticks, but first
 	 * remove the single look ahead already included in the error.
 	 */
-	tick_error = tick_length >>
-			(NTP_SCALE_SHIFT - timekeeper.clock->shift + 1);
+	tick_error = tick_length >> (timekeeper.ntp_error_shift + 1);
 	tick_error -= timekeeper.xtime_interval >> 1;
 	error = ((error - tick_error) >> look_ahead) + tick_error;
 

@@ -673,8 +679,7 @@ static void timekeeping_adjust(s64 offset)
 	s64 error, interval = timekeeper.cycle_interval;
 	int adj;
 
-	error = timekeeper.ntp_error >>
-		(NTP_SCALE_SHIFT - timekeeper.clock->shift - 1);
+	error = timekeeper.ntp_error >> (timekeeper.ntp_error_shift - 1);
 	if (error > interval) {
 		error >>= 2;
 		if (likely(error <= interval))

@@ -696,7 +701,7 @@ static void timekeeping_adjust(s64 offset)
 	timekeeper.xtime_interval += interval;
 	timekeeper.xtime_nsec -= offset;
 	timekeeper.ntp_error -= (interval - offset) <<
-			(NTP_SCALE_SHIFT - timekeeper.clock->shift);
+				timekeeper.ntp_error_shift;
 }
 
 /**

@@ -708,7 +713,7 @@ void update_wall_time(void)
 {
 	struct clocksource *clock;
 	cycle_t offset;
-	s64 nsecs;
+	u64 nsecs;
 
 	/* Make sure we're fully resumed: */
 	if (unlikely(timekeeping_suspended))

@@ -720,13 +725,13 @@ void update_wall_time(void)
 #else
 	offset = timekeeper.cycle_interval;
 #endif
-	timekeeper.xtime_nsec = (s64)xtime.tv_nsec << clock->shift;
+	timekeeper.xtime_nsec = (s64)xtime.tv_nsec << timekeeper.shift;
 
 	/* normally this loop will run just once, however in the
 	 * case of lost or late ticks, it will accumulate correctly.
 	 */
 	while (offset >= timekeeper.cycle_interval) {
-		u64 nsecps = (u64)NSEC_PER_SEC << clock->shift;
+		u64 nsecps = (u64)NSEC_PER_SEC << timekeeper.shift;
 
 		/* accumulate one interval */
 		offset -= timekeeper.cycle_interval;

@@ -748,7 +753,7 @@ void update_wall_time(void)
 		/* accumulate error between NTP and clock interval */
 		timekeeper.ntp_error += tick_length;
 		timekeeper.ntp_error -= timekeeper.xtime_interval <<
-					(NTP_SCALE_SHIFT - clock->shift);
+					timekeeper.ntp_error_shift;
 	}
 
 	/* correct the clock when NTP error is too big */

@@ -773,16 +778,16 @@ void update_wall_time(void)
 	if (unlikely((s64)timekeeper.xtime_nsec < 0)) {
 		s64 neg = -(s64)timekeeper.xtime_nsec;
 		timekeeper.xtime_nsec = 0;
-		timekeeper.ntp_error += neg << (NTP_SCALE_SHIFT - clock->shift);
+		timekeeper.ntp_error += neg << timekeeper.ntp_error_shift;
 	}
 
 	/* store full nanoseconds into xtime after rounding it up and
 	 * add the remainder to the error difference.
 	 */
-	xtime.tv_nsec = ((s64)timekeeper.xtime_nsec >> clock->shift) + 1;
-	timekeeper.xtime_nsec -= (s64)xtime.tv_nsec << clock->shift;
-	timekeeper.ntp_error += timekeeper.xtime_nsec <<
-				(NTP_SCALE_SHIFT - clock->shift);
+	xtime.tv_nsec =	((s64) timekeeper.xtime_nsec >> timekeeper.shift) + 1;
+	timekeeper.xtime_nsec -= (s64) xtime.tv_nsec << timekeeper.shift;
+	timekeeper.ntp_error +=	timekeeper.xtime_nsec <<
+				timekeeper.ntp_error_shift;
 
 	nsecs = clocksource_cyc2ns(offset, clock->mult, clock->shift);
 	update_xtime_cache(nsecs);

Index: linux-2.6/include/linux/clocksource.h
===================================================================
@@ -149,8 +149,7 @@ extern u64 timecounter_cyc2time(struct t
  * @disable:		optional function to disable the clocksource
  * @mask:		bitmask for two's complement
  *			subtraction of non 64 bit counters
- * @mult:		cycle to nanosecond multiplier (adjusted by NTP)
- * @mult_orig:		cycle to nanosecond multiplier (unadjusted by NTP)
+ * @mult:		cycle to nanosecond multiplier
  * @shift:		cycle to nanosecond divisor (power of two)
  * @flags:		flags describing special properties
  * @vread:		vsyscall based read

@@ -168,7 +167,6 @@ struct clocksource {
 	void (*disable)(struct clocksource *cs);
 	cycle_t mask;
 	u32 mult;
-	u32 mult_orig;
 	u32 shift;
 	unsigned long flags;
 	cycle_t (*vread)(void);

141 lines not on this page · raw message

Index: linux-2.6/arch/arm/plat-omap/common.c
===================================================================
@@ -253,11 +253,8 @@ static struct clocksource clocksource_32
  */
 unsigned long long sched_clock(void)
 {
-	unsigned long long ret;
-
-	ret = (unsigned long long)clocksource_32k.read(&clocksource_32k);
-	ret = (ret * clocksource_32k.mult_orig) >> clocksource_32k.shift;
-	return ret;
+	return clocksource_cyc2ns(clocksource_32k.read(&clocksource_32k),
+				  clocksource_32k.mult, clocksource_32k.shift);
 }
 
 static int __init omap_init_clocksource_32k(void)

index ebcf006..95587b6 100644
@@ -253,11 +253,8 @@ static struct clocksource clocksource_32k = {
  */
 unsigned long long sched_clock(void)
 {
-	unsigned long long ret;
-
-	ret = (unsigned long long)clocksource_32k.read(&clocksource_32k);
-	ret = (ret * clocksource_32k.mult_orig) >> clocksource_32k.shift;
-	return ret;
+	return clocksource_cyc2ns(clocksource_32k.read(&clocksource_32k),
+				  clocksource_32k.mult, clocksource_32k.shift);
 }
 
 static int __init omap_init_clocksource_32k(void)

index e12e309..e34015e 100644
@@ -149,8 +149,7 @@ extern u64 timecounter_cyc2time(struct timecounter *tc,
  * @disable:		optional function to disable the clocksource
  * @mask:		bitmask for two's complement
  *			subtraction of non 64 bit counters
- * @mult:		cycle to nanosecond multiplier (adjusted by NTP)
- * @mult_orig:		cycle to nanosecond multiplier (unadjusted by NTP)
+ * @mult:		cycle to nanosecond multiplier
  * @shift:		cycle to nanosecond divisor (power of two)
  * @flags:		flags describing special properties
  * @vread:		vsyscall based read

@@ -168,7 +167,6 @@ struct clocksource {
 	void (*disable)(struct clocksource *cs);
 	cycle_t mask;
 	u32 mult;
-	u32 mult_orig;
 	u32 shift;
 	unsigned long flags;
 	cycle_t (*vread)(void);

141 lines not on this page · raw message

Index: linux-2.6/kernel/time/timekeeping.c
===================================================================
@@ -95,6 +95,40 @@ static void timekeeper_setup_internals(s
 	timekeeper.mult = clock->mult;
 }
 
+/* Timekeeper helper functions. */
+static inline s64 timekeeping_get_ns(void)
+{
+	cycle_t cycle_now, cycle_delta;
+	struct clocksource *clock;
+
+	/* read clocksource: */
+	clock = timekeeper.clock;
+	cycle_now = clock->read(clock);
+
+	/* calculate the delta since the last update_wall_time: */
+	cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
+
+	/* return delta convert to nanoseconds using ntp adjusted mult. */
+	return clocksource_cyc2ns(cycle_delta, timekeeper.mult,
+				  timekeeper.shift);
+}
+
+static inline s64 timekeeping_get_ns_raw(void)
+{
+	cycle_t cycle_now, cycle_delta;
+	struct clocksource *clock;
+
+	/* read clocksource: */
+	clock = timekeeper.clock;
+	cycle_now = clock->read(clock);
+
+	/* calculate the delta since the last update_wall_time: */
+	cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
+
+	/* return delta convert to nanoseconds using ntp adjusted mult. */
+	return clocksource_cyc2ns(cycle_delta, clock->mult, clock->shift);
+}
+
 /*
  * This read-write spinlock protects us from races in SMP while
  * playing with xtime.

@@ -183,8 +217,6 @@ static void timekeeping_forward_now(void
  */
 void getnstimeofday(struct timespec *ts)
 {
-	cycle_t cycle_now, cycle_delta;
-	struct clocksource *clock;
 	unsigned long seq;
 	s64 nsecs;
 

@@ -194,17 +226,7 @@ void getnstimeofday(struct timespec *ts)
 		seq = read_seqbegin(&xtime_lock);
 
 		*ts = xtime;
-
-		/* read clocksource: */
-		clock = timekeeper.clock;
-		cycle_now = clock->read(clock);
-
-		/* calculate the delta since the last update_wall_time: */
-		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
-
-		/* convert to nanoseconds: */
-		nsecs = clocksource_cyc2ns(cycle_delta, timekeeper.mult,
-					   timekeeper.shift);
+		nsecs = timekeeping_get_ns();
 
 		/* If arch requires, add in gettimeoffset() */
 		nsecs += arch_gettimeoffset();

@@ -218,8 +240,6 @@ EXPORT_SYMBOL(getnstimeofday);
 
 ktime_t ktime_get(void)
 {
-	cycle_t cycle_now, cycle_delta;
-	struct clocksource *clock;
 	unsigned int seq;
 	s64 secs, nsecs;
 

@@ -229,17 +249,7 @@ ktime_t ktime_get(void)
 		seq = read_seqbegin(&xtime_lock);
 		secs = xtime.tv_sec + wall_to_monotonic.tv_sec;
 		nsecs = xtime.tv_nsec + wall_to_monotonic.tv_nsec;
-
-		/* read clocksource: */
-		clock = timekeeper.clock;
-		cycle_now = clock->read(clock);
-
-		/* calculate the delta since the last update_wall_time: */
-		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
-
-		/* convert to nanoseconds: */
-		nsecs += clocksource_cyc2ns(cycle_delta, timekeeper.mult,
-					    timekeeper.shift);
+		nsecs += timekeeping_get_ns();
 
 	} while (read_seqretry(&xtime_lock, seq));
 	/*

@@ -260,8 +270,6 @@ EXPORT_SYMBOL_GPL(ktime_get);
  */
 void ktime_get_ts(struct timespec *ts)
 {
-	cycle_t cycle_now, cycle_delta;
-	struct clocksource *clock;
 	struct timespec tomono;
 	unsigned int seq;
 	s64 nsecs;

@@ -272,17 +280,7 @@ void ktime_get_ts(struct timespec *ts)
 		seq = read_seqbegin(&xtime_lock);
 		*ts = xtime;
 		tomono = wall_to_monotonic;
-
-		/* read clocksource: */
-		clock = timekeeper.clock;
-		cycle_now = clock->read(clock);
-
-		/* calculate the delta since the last update_wall_time: */
-		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
-
-		/* convert to nanoseconds: */
-		nsecs = clocksource_cyc2ns(cycle_delta, timekeeper.mult,
-					   timekeeper.shift);
+		nsecs = timekeeping_get_ns();
 
 	} while (read_seqretry(&xtime_lock, seq));
 

@@ -445,23 +443,10 @@ void getrawmonotonic(struct timespec *ts
 {
 	unsigned long seq;
 	s64 nsecs;
-	cycle_t cycle_now, cycle_delta;
-	struct clocksource *clock;
 
 	do {
 		seq = read_seqbegin(&xtime_lock);
-
-		/* read clocksource: */
-		clock = timekeeper.clock;
-		cycle_now = clock->read(clock);
-
-		/* calculate the delta since the last update_wall_time: */
-		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
-
-		/* convert to nanoseconds: */
-		nsecs = clocksource_cyc2ns(cycle_delta, clock->mult,
-					   clock->shift);
-
+		nsecs = timekeeping_get_ns_raw();
 		*ts = raw_time;
 
 	} while (read_seqretry(&xtime_lock, seq));

index f4056f6..27ae01b 100644
@@ -95,6 +95,40 @@ static void timekeeper_setup_internals(struct clocksource *clock)
 	timekeeper.mult = clock->mult;
 }
 
+/* Timekeeper helper functions. */
+static inline s64 timekeeping_get_ns(void)
+{
+	cycle_t cycle_now, cycle_delta;
+	struct clocksource *clock;
+
+	/* read clocksource: */
+	clock = timekeeper.clock;
+	cycle_now = clock->read(clock);
+
+	/* calculate the delta since the last update_wall_time: */
+	cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
+
+	/* return delta convert to nanoseconds using ntp adjusted mult. */
+	return clocksource_cyc2ns(cycle_delta, timekeeper.mult,
+				  timekeeper.shift);
+}
+
+static inline s64 timekeeping_get_ns_raw(void)
+{
+	cycle_t cycle_now, cycle_delta;
+	struct clocksource *clock;
+
+	/* read clocksource: */
+	clock = timekeeper.clock;
+	cycle_now = clock->read(clock);
+
+	/* calculate the delta since the last update_wall_time: */
+	cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
+
+	/* return delta convert to nanoseconds using ntp adjusted mult. */
+	return clocksource_cyc2ns(cycle_delta, clock->mult, clock->shift);
+}
+
 /*
  * This read-write spinlock protects us from races in SMP while
  * playing with xtime.

@@ -183,8 +217,6 @@ static void timekeeping_forward_now(void)
  */
 void getnstimeofday(struct timespec *ts)
 {
-	cycle_t cycle_now, cycle_delta;
-	struct clocksource *clock;
 	unsigned long seq;
 	s64 nsecs;
 

@@ -194,17 +226,7 @@ void getnstimeofday(struct timespec *ts)
 		seq = read_seqbegin(&xtime_lock);
 
 		*ts = xtime;
-
-		/* read clocksource: */
-		clock = timekeeper.clock;
-		cycle_now = clock->read(clock);
-
-		/* calculate the delta since the last update_wall_time: */
-		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
-
-		/* convert to nanoseconds: */
-		nsecs = clocksource_cyc2ns(cycle_delta, timekeeper.mult,
-					   timekeeper.shift);
+		nsecs = timekeeping_get_ns();
 
 		/* If arch requires, add in gettimeoffset() */
 		nsecs += arch_gettimeoffset();

@@ -218,8 +240,6 @@ EXPORT_SYMBOL(getnstimeofday);
 
 ktime_t ktime_get(void)
 {
-	cycle_t cycle_now, cycle_delta;
-	struct clocksource *clock;
 	unsigned int seq;
 	s64 secs, nsecs;
 

@@ -229,17 +249,7 @@ ktime_t ktime_get(void)
 		seq = read_seqbegin(&xtime_lock);
 		secs = xtime.tv_sec + wall_to_monotonic.tv_sec;
 		nsecs = xtime.tv_nsec + wall_to_monotonic.tv_nsec;
-
-		/* read clocksource: */
-		clock = timekeeper.clock;
-		cycle_now = clock->read(clock);
-
-		/* calculate the delta since the last update_wall_time: */
-		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
-
-		/* convert to nanoseconds: */
-		nsecs += clocksource_cyc2ns(cycle_delta, timekeeper.mult,
-					    timekeeper.shift);
+		nsecs += timekeeping_get_ns();
 
 	} while (read_seqretry(&xtime_lock, seq));
 	/*

@@ -260,8 +270,6 @@ EXPORT_SYMBOL_GPL(ktime_get);
  */
 void ktime_get_ts(struct timespec *ts)
 {
-	cycle_t cycle_now, cycle_delta;
-	struct clocksource *clock;
 	struct timespec tomono;
 	unsigned int seq;
 	s64 nsecs;

@@ -272,17 +280,7 @@ void ktime_get_ts(struct timespec *ts)
 		seq = read_seqbegin(&xtime_lock);
 		*ts = xtime;
 		tomono = wall_to_monotonic;
-
-		/* read clocksource: */
-		clock = timekeeper.clock;
-		cycle_now = clock->read(clock);
-
-		/* calculate the delta since the last update_wall_time: */
-		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
-
-		/* convert to nanoseconds: */
-		nsecs = clocksource_cyc2ns(cycle_delta, timekeeper.mult,
-					   timekeeper.shift);
+		nsecs = timekeeping_get_ns();
 
 	} while (read_seqretry(&xtime_lock, seq));
 

@@ -445,23 +443,10 @@ void getrawmonotonic(struct timespec *ts)
 {
 	unsigned long seq;
 	s64 nsecs;
-	cycle_t cycle_now, cycle_delta;
-	struct clocksource *clock;
 
 	do {
 		seq = read_seqbegin(&xtime_lock);
-
-		/* read clocksource: */
-		clock = timekeeper.clock;
-		cycle_now = clock->read(clock);
-
-		/* calculate the delta since the last update_wall_time: */
-		cycle_delta = (cycle_now - clock->cycle_last) & clock->mask;
-
-		/* convert to nanoseconds: */
-		nsecs = clocksource_cyc2ns(cycle_delta, clock->mult,
-					   clock->shift);
-
+		nsecs = timekeeping_get_ns_raw();
 		*ts = raw_time;
 
 	} while (read_seqretry(&xtime_lock, seq));

Index: linux-2.6/include/linux/clocksource.h
===================================================================
@@ -291,4 +291,6 @@ static inline void update_vsyscall_tz(vo
 }
 #endif
 
+extern void timekeeping_notify(struct clocksource *clock);
+
 #endif /* _LINUX_CLOCKSOURCE_H */

233 lines not on this page · raw message

82 lines not on this page · raw message

index e34015e..9ea40ff 100644
@@ -291,4 +291,6 @@ static inline void update_vsyscall_tz(void)
 }
 #endif
 
+extern void timekeeping_notify(struct clocksource *clock);
+
 #endif /* _LINUX_CLOCKSOURCE_H */

233 lines not on this page · raw message

82 lines not on this page · raw message

Index: linux-2.6/arch/m68knommu/kernel/time.c
===================================================================
@@ -72,9 +72,10 @@ static unsigned long read_rtc_mmss(void)
 	return  mktime(year, mon, day, hour, min, sec);;
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	return read_rtc_mmss();
+	ts->tv_sec = read_rtc_mmss();
+	ts->tv_nsec = 0;
 }
 
 int update_persistent_clock(struct timespec now)

Index: linux-2.6/arch/mips/dec/time.c
===================================================================
@@ -18,7 +18,7 @@
 #include <asm/dec/ioasic.h>
 #include <asm/dec/machtype.h>
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
 	unsigned int year, mon, day, hour, min, sec, real_year;
 	unsigned long flags;

@@ -53,7 +53,8 @@ unsigned long read_persistent_clock(void
 
 	year += real_year - 72 + 2000;
 
-	return mktime(year, mon, day, hour, min, sec);
+	ts->tv_sec = mktime(year, mon, day, hour, min, sec);
+	ts->tv_nsec = 0;
 }
 
 /*

Index: linux-2.6/arch/mips/lasat/ds1603.c
===================================================================
@@ -135,7 +135,7 @@ static void rtc_end_op(void)
 	lasat_ndelay(1000);
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
 	unsigned long word;
 	unsigned long flags;

@@ -147,7 +147,8 @@ unsigned long read_persistent_clock(void
 	rtc_end_op();
 	spin_unlock_irqrestore(&rtc_lock, flags);
 
-	return word;
+	ts->tv_sec = word;
+	ts->tv_nsec = 0;
 }
 
 int rtc_mips_set_mmss(unsigned long time)

Index: linux-2.6/arch/mips/lasat/sysctl.c
===================================================================
@@ -92,10 +92,12 @@ static int rtctmp;
 int proc_dolasatrtc(ctl_table *table, int write, struct file *filp,
 		       void *buffer, size_t *lenp, loff_t *ppos)
 {
+	struct timespec ts;
 	int r;
 
 	if (!write) {
-		rtctmp = read_persistent_clock();
+		read_persistent_clock(&ts);
+		rtctmp = ts.tv_sec;
 		/* check for time < 0 and set to 0 */
 		if (rtctmp < 0)
 			rtctmp = 0;

@@ -134,9 +136,11 @@ int sysctl_lasat_rtc(ctl_table *table,
 		    void *oldval, size_t *oldlenp,
 		    void *newval, size_t newlen)
 {
+	struct timespec ts;
 	int r;
 
-	rtctmp = read_persistent_clock();
+	read_persistent_clock(&ts);
+	rtctmp = ts.tv_sec;
 	if (rtctmp < 0)
 		rtctmp = 0;
 	r = sysctl_intvec(table, oldval, oldlenp, newval, newlen);

Index: linux-2.6/arch/mips/lemote/lm2e/setup.c
===================================================================
@@ -54,9 +54,10 @@ void __init plat_time_init(void)
 	mips_hpt_frequency = cpu_clock_freq / 2;
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	return mc146818_get_cmos_time();
+	ts->tv_sec = mc146818_get_cmos_time();
+	ts->tv_nsec = 0;
 }
 
 void (*__wbflush)(void);

Index: linux-2.6/arch/mips/mti-malta/malta-time.c
===================================================================
@@ -100,9 +100,10 @@ static unsigned int __init estimate_cpu_
 	return count;
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	return mc146818_get_cmos_time();
+	ts->tv_sec = mc146818_get_cmos_time();
+	ts->tv_nsec = 0;
 }
 
 static void __init plat_perf_setup(void)

Index: linux-2.6/arch/mips/pmc-sierra/yosemite/setup.c
===================================================================
@@ -70,7 +70,7 @@ void __init bus_error_init(void)
 }
 
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
 	unsigned int year, month, day, hour, min, sec;
 	unsigned long flags;

@@ -92,7 +92,8 @@ unsigned long read_persistent_clock(void
 	m48t37_base->control = 0x00;
 	spin_unlock_irqrestore(&rtc_lock, flags);
 
-	return mktime(year, month, day, hour, min, sec);
+	ts->tv_sec = mktime(year, month, day, hour, min, sec);
+	ts->tv_nsec = 0;
 }
 
 int rtc_mips_set_time(unsigned long tim)

Index: linux-2.6/arch/mips/sibyte/swarm/setup.c
===================================================================
@@ -87,19 +87,26 @@ enum swarm_rtc_type {
 
 enum swarm_rtc_type swarm_rtc_type;
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
+	unsigned long sec;
+
 	switch (swarm_rtc_type) {
 	case RTC_XICOR:
-		return xicor_get_time();
+		sec = xicor_get_time();
+		break;
 
 	case RTC_M4LT81:
-		return m41t81_get_time();
+		sec = m41t81_get_time();
+		break;
 
 	case RTC_NONE:
 	default:
-		return mktime(2000, 1, 1, 0, 0, 0);
+		sec = mktime(2000, 1, 1, 0, 0, 0);
+		break;
 	}
+	ts->tv_sec = sec;
+	tv->tv_nsec = 0;
 }
 
 int rtc_mips_set_time(unsigned long sec)

Index: linux-2.6/arch/mips/sni/time.c
===================================================================
@@ -182,7 +182,8 @@ void __init plat_time_init(void)
 	setup_pit_timer();
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	return -1;
+	ts->tv_sec = -1;
+	ts->tv_nsec = 0;
 }

Index: linux-2.6/arch/powerpc/kernel/time.c
===================================================================
@@ -772,7 +772,7 @@ int update_persistent_clock(struct times
 	return ppc_md.set_rtc_time(&tm);
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
 	struct rtc_time tm;
 	static int first = 1;

@@ -790,8 +790,9 @@ unsigned long read_persistent_clock(void
 	if (!ppc_md.get_rtc_time)
 		return 0;
 	ppc_md.get_rtc_time(&tm);
-	return mktime(tm.tm_year+1900, tm.tm_mon+1, tm.tm_mday,
-		      tm.tm_hour, tm.tm_min, tm.tm_sec);
+	ts->tv_sec = mktime(tm.tm_year+1900, tm.tm_mon+1, tm.tm_mday,
+			    tm.tm_hour, tm.tm_min, tm.tm_sec);
+	ts->tv_nsec = 0;
 }
 
 /* clocksource code */

Index: linux-2.6/arch/s390/kernel/time.c
===================================================================
@@ -182,12 +182,9 @@ static void timing_alert_interrupt(__u16
 static void etr_reset(void);
 static void stp_reset(void);
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	struct timespec ts;
-
-	tod_to_timeval(get_clock() - TOD_UNIX_EPOCH, &ts);
-	return ts.tv_sec;
+	tod_to_timeval(get_clock() - TOD_UNIX_EPOCH, ts);
 }
 
 static cycle_t read_tod_clock(struct clocksource *cs)

@@ -248,7 +245,6 @@ void __init time_init(void)
 {
 	struct timespec ts;
 	unsigned long flags;
-	cycle_t now;
 
 	/* Reset time synchronization interfaces. */
 	etr_reset();

@@ -266,20 +262,10 @@ void __init time_init(void)
 		panic("Could not register TOD clock source");
 
 	/*
-	 * The TOD clock is an accurate clock. The xtime should be
-	 * initialized in a way that the difference between TOD and
-	 * xtime is reasonably small. Too bad that timekeeping_init
-	 * sets xtime.tv_nsec to zero. In addition the clock source
-	 * change from the jiffies clock source to the TOD clock
-	 * source add another error of up to 1/HZ second. The same
-	 * function sets wall_to_monotonic to a value that is too
-	 * small for /proc/uptime to be accurate.
-	 * Reset xtime and wall_to_monotonic to sane values.
+	 * Reset wall_to_monotonic to the initial timestamp created
+	 * in head.S to get a precise value in /proc/uptime.
 	 */
 	write_seqlock_irqsave(&xtime_lock, flags);
-	now = get_clock();
-	tod_to_timeval(now - TOD_UNIX_EPOCH, &xtime);
-	clocksource_tod.cycle_last = now;
 	tod_to_timeval(sched_clock_base_cc - TOD_UNIX_EPOCH, &ts);
 	set_normalized_timespec(&wall_to_monotonic, -ts.tv_sec, -ts.tv_nsec);
 	write_sequnlock_irqrestore(&xtime_lock, flags);

Index: linux-2.6/arch/sh/kernel/time.c
===================================================================
@@ -39,11 +39,9 @@ void (*rtc_sh_get_time)(struct timespec 
 int (*rtc_sh_set_time)(const time_t) = null_rtc_set_time;
 
 #ifdef CONFIG_GENERIC_CMOS_UPDATE
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	struct timespec tv;
-	rtc_sh_get_time(&tv);
-	return tv.tv_sec;
+	rtc_sh_get_time(&ts);
 }
 
 int update_persistent_clock(struct timespec now)

Index: linux-2.6/arch/x86/kernel/rtc.c
===================================================================
@@ -178,7 +178,7 @@ static int set_rtc_mmss(unsigned long no
 }
 
 /* not static: needed by APM */
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
 	unsigned long retval, flags;
 

@@ -186,7 +186,8 @@ unsigned long read_persistent_clock(void
 	retval = get_wallclock();
 	spin_unlock_irqrestore(&rtc_lock, flags);
 
-	return retval;
+	ts->tv_sec = retval;
+	ts->tv_nsec = 0;
 }
 
 int update_persistent_clock(struct timespec now)

Index: linux-2.6/arch/xtensa/kernel/time.c
===================================================================
@@ -59,9 +59,8 @@ static struct irqaction timer_irqaction 
 
 void __init time_init(void)
 {
-	xtime.tv_nsec = 0;
-	xtime.tv_sec = read_persistent_clock();
-
+	/* FIXME: xtime&wall_to_monotonic are set in timekeeping_init. */
+	read_persistent_clock(&xtime);
 	set_normalized_timespec(&wall_to_monotonic,
 		-xtime.tv_sec, -xtime.tv_nsec);
 

Index: linux-2.6/include/linux/time.h
===================================================================
@@ -101,7 +101,7 @@ extern struct timespec xtime;
 extern struct timespec wall_to_monotonic;
 extern seqlock_t xtime_lock;
 
-extern unsigned long read_persistent_clock(void);
+extern void read_persistent_clock(struct timespec *ts);
 extern int update_persistent_clock(struct timespec now);
 extern int no_sync_cmos_clock __read_mostly;
 void timekeeping_init(void);

120 lines not on this page · raw message

index d182b2f..6843224 100644
@@ -72,9 +72,10 @@ static unsigned long read_rtc_mmss(void)
 	return  mktime(year, mon, day, hour, min, sec);;
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	return read_rtc_mmss();
+	ts->tv_sec = read_rtc_mmss();
+	ts->tv_nsec = 0;
 }
 
 int update_persistent_clock(struct timespec now)

index 463136e..02f505f 100644
@@ -18,7 +18,7 @@
 #include <asm/dec/ioasic.h>
 #include <asm/dec/machtype.h>
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
 	unsigned int year, mon, day, hour, min, sec, real_year;
 	unsigned long flags;

@@ -53,7 +53,8 @@ unsigned long read_persistent_clock(void)
 
 	year += real_year - 72 + 2000;
 
-	return mktime(year, mon, day, hour, min, sec);
+	ts->tv_sec = mktime(year, mon, day, hour, min, sec);
+	ts->tv_nsec = 0;
 }
 
 /*

index 52cb143..c6fd96f 100644
@@ -135,7 +135,7 @@ static void rtc_end_op(void)
 	lasat_ndelay(1000);
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
 	unsigned long word;
 	unsigned long flags;

@@ -147,7 +147,8 @@ unsigned long read_persistent_clock(void)
 	rtc_end_op();
 	spin_unlock_irqrestore(&rtc_lock, flags);
 
-	return word;
+	ts->tv_sec = word;
+	ts->tv_nsec = 0;
 }
 
 int rtc_mips_set_mmss(unsigned long time)

index 8f88886..3f04d4c 100644
@@ -92,10 +92,12 @@ static int rtctmp;
 int proc_dolasatrtc(ctl_table *table, int write, struct file *filp,
 		       void *buffer, size_t *lenp, loff_t *ppos)
 {
+	struct timespec ts;
 	int r;
 
 	if (!write) {
-		rtctmp = read_persistent_clock();
+		read_persistent_clock(&ts);
+		rtctmp = ts.tv_sec;
 		/* check for time < 0 and set to 0 */
 		if (rtctmp < 0)
 			rtctmp = 0;

@@ -134,9 +136,11 @@ int sysctl_lasat_rtc(ctl_table *table,
 		    void *oldval, size_t *oldlenp,
 		    void *newval, size_t newlen)
 {
+	struct timespec ts;
 	int r;
 
-	rtctmp = read_persistent_clock();
+	read_persistent_clock(&ts);
+	rtctmp = ts.tv_sec;
 	if (rtctmp < 0)
 		rtctmp = 0;
 	r = sysctl_intvec(table, oldval, oldlenp, newval, newlen);

index ebd6cea..24b355d 100644
@@ -54,9 +54,10 @@ void __init plat_time_init(void)
 	mips_hpt_frequency = cpu_clock_freq / 2;
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	return mc146818_get_cmos_time();
+	ts->tv_sec = mc146818_get_cmos_time();
+	ts->tv_nsec = 0;
 }
 
 void (*__wbflush)(void);

index 0b97d47..3c6f190 100644
@@ -100,9 +100,10 @@ static unsigned int __init estimate_cpu_frequency(void)
 	return count;
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	return mc146818_get_cmos_time();
+	ts->tv_sec = mc146818_get_cmos_time();
+	ts->tv_nsec = 0;
 }
 
 static void __init plat_perf_setup(void)

index 2d3c0dc..3498ac9 100644
@@ -70,7 +70,7 @@ void __init bus_error_init(void)
 }
 
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
 	unsigned int year, month, day, hour, min, sec;
 	unsigned long flags;

@@ -92,7 +92,8 @@ unsigned long read_persistent_clock(void)
 	m48t37_base->control = 0x00;
 	spin_unlock_irqrestore(&rtc_lock, flags);
 
-	return mktime(year, month, day, hour, min, sec);
+	ts->tv_sec = mktime(year, month, day, hour, min, sec);
+	ts->tv_nsec = 0;
 }
 
 int rtc_mips_set_time(unsigned long tim)

index 672e45d..623ffc9 100644
@@ -87,19 +87,26 @@ enum swarm_rtc_type {
 
 enum swarm_rtc_type swarm_rtc_type;
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
+	unsigned long sec;
+
 	switch (swarm_rtc_type) {
 	case RTC_XICOR:
-		return xicor_get_time();
+		sec = xicor_get_time();
+		break;
 
 	case RTC_M4LT81:
-		return m41t81_get_time();
+		sec = m41t81_get_time();
+		break;
 
 	case RTC_NONE:
 	default:
-		return mktime(2000, 1, 1, 0, 0, 0);
+		sec = mktime(2000, 1, 1, 0, 0, 0);
+		break;
 	}
+	ts->tv_sec = sec;
+	tv->tv_nsec = 0;
 }
 
 int rtc_mips_set_time(unsigned long sec)

index 0d9ec1a..62df6a5 100644
@@ -182,7 +182,8 @@ void __init plat_time_init(void)
 	setup_pit_timer();
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	return -1;
+	ts->tv_sec = -1;
+	ts->tv_nsec = 0;
 }

index eae4511..ad63f30 100644
@@ -769,7 +769,7 @@ int update_persistent_clock(struct timespec now)
 	return ppc_md.set_rtc_time(&tm);
 }
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
 	struct rtc_time tm;
 	static int first = 1;

@@ -787,8 +787,9 @@ unsigned long read_persistent_clock(void)
 	if (!ppc_md.get_rtc_time)
 		return 0;
 	ppc_md.get_rtc_time(&tm);
-	return mktime(tm.tm_year+1900, tm.tm_mon+1, tm.tm_mday,
-		      tm.tm_hour, tm.tm_min, tm.tm_sec);
+	ts->tv_sec = mktime(tm.tm_year+1900, tm.tm_mon+1, tm.tm_mday,
+			    tm.tm_hour, tm.tm_min, tm.tm_sec);
+	ts->tv_nsec = 0;
 }
 
 /* clocksource code */

index e76c2e7..a94ec48 100644
@@ -182,12 +182,9 @@ static void timing_alert_interrupt(__u16 code)
 static void etr_reset(void);
 static void stp_reset(void);
 
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	struct timespec ts;
-
-	tod_to_timeval(get_clock() - TOD_UNIX_EPOCH, &ts);
-	return ts.tv_sec;
+	tod_to_timeval(get_clock() - TOD_UNIX_EPOCH, ts);
 }
 
 static cycle_t read_tod_clock(struct clocksource *cs)

@@ -248,7 +245,6 @@ void __init time_init(void)
 {
 	struct timespec ts;
 	unsigned long flags;
-	cycle_t now;
 
 	/* Reset time synchronization interfaces. */
 	etr_reset();

@@ -266,20 +262,10 @@ void __init time_init(void)
 		panic("Could not register TOD clock source");
 
 	/*
-	 * The TOD clock is an accurate clock. The xtime should be
-	 * initialized in a way that the difference between TOD and
-	 * xtime is reasonably small. Too bad that timekeeping_init
-	 * sets xtime.tv_nsec to zero. In addition the clock source
-	 * change from the jiffies clock source to the TOD clock
-	 * source add another error of up to 1/HZ second. The same
-	 * function sets wall_to_monotonic to a value that is too
-	 * small for /proc/uptime to be accurate.
-	 * Reset xtime and wall_to_monotonic to sane values.
+	 * Reset wall_to_monotonic to the initial timestamp created
+	 * in head.S to get a precise value in /proc/uptime.
 	 */
 	write_seqlock_irqsave(&xtime_lock, flags);
-	now = get_clock();
-	tod_to_timeval(now - TOD_UNIX_EPOCH, &xtime);
-	clocksource_tod.cycle_last = now;
 	tod_to_timeval(sched_clock_base_cc - TOD_UNIX_EPOCH, &ts);
 	set_normalized_timespec(&wall_to_monotonic, -ts.tv_sec, -ts.tv_nsec);
 	write_sequnlock_irqrestore(&xtime_lock, flags);

index 9b352a1..3f4706a 100644
@@ -39,11 +39,9 @@ void (*rtc_sh_get_time)(struct timespec *) = null_rtc_get_time;
 int (*rtc_sh_set_time)(const time_t) = null_rtc_set_time;
 
 #ifdef CONFIG_GENERIC_CMOS_UPDATE
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
-	struct timespec tv;
-	rtc_sh_get_time(&tv);
-	return tv.tv_sec;
+	rtc_sh_get_time(&ts);
 }
 
 int update_persistent_clock(struct timespec now)

index 5d465b2..bf67dcb 100644
@@ -178,7 +178,7 @@ static int set_rtc_mmss(unsigned long nowtime)
 }
 
 /* not static: needed by APM */
-unsigned long read_persistent_clock(void)
+void read_persistent_clock(struct timespec *ts)
 {
 	unsigned long retval, flags;
 

@@ -186,7 +186,8 @@ unsigned long read_persistent_clock(void)
 	retval = get_wallclock();
 	spin_unlock_irqrestore(&rtc_lock, flags);
 
-	return retval;
+	ts->tv_sec = retval;
+	ts->tv_nsec = 0;
 }
 
 int update_persistent_clock(struct timespec now)

index 8848120..19085ff 100644
@@ -59,9 +59,8 @@ static struct irqaction timer_irqaction = {
 
 void __init time_init(void)
 {
-	xtime.tv_nsec = 0;
-	xtime.tv_sec = read_persistent_clock();
-
+	/* FIXME: xtime&wall_to_monotonic are set in timekeeping_init. */
+	read_persistent_clock(&xtime);
 	set_normalized_timespec(&wall_to_monotonic,
 		-xtime.tv_sec, -xtime.tv_nsec);
 

index e7c8445..53a3216 100644
@@ -101,7 +101,7 @@ extern struct timespec xtime;
 extern struct timespec wall_to_monotonic;
 extern seqlock_t xtime_lock;
 
-extern unsigned long read_persistent_clock(void);
+extern void read_persistent_clock(struct timespec *ts);
 extern int update_persistent_clock(struct timespec now);
 extern int no_sync_cmos_clock __read_mostly;
 void timekeeping_init(void);

120 lines not on this page · raw message