1 files changed, 75 insertions, 29 deletions

diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c
index d50a821ec89..7b78f88c170 100644
--- a/arch/x86/xen/time.c
+++ b/arch/x86/xen/time.c
@@ -14,6 +14,8 @@
 #include <linux/kernel_stat.h>
 #include <linux/math64.h>
 #include <linux/gfp.h>
+#include <linux/slab.h>
+#include <linux/pvclock_gtod.h>
 
 #include <asm/pvclock.h>
 #include <asm/xen/hypervisor.h>
@@ -178,34 +180,56 @@ static void xen_read_wallclock(struct timespec *ts)
 	put_cpu_var(xen_vcpu);
 }
 
-static unsigned long xen_get_wallclock(void)
+static void xen_get_wallclock(struct timespec *now)
 {
-	struct timespec ts;
+	xen_read_wallclock(now);
+}
 
-	xen_read_wallclock(&ts);
-	return ts.tv_sec;
+static int xen_set_wallclock(const struct timespec *now)
+{
+	return -1;
 }
 
-static int xen_set_wallclock(unsigned long now)
+static int xen_pvclock_gtod_notify(struct notifier_block *nb,
+				   unsigned long was_set, void *priv)
 {
+	/* Protected by the calling core code serialization */
+	static struct timespec next_sync;
+
 	struct xen_platform_op op;
-	int rc;
+	struct timespec now;
 
-	/* do nothing for domU */
-	if (!xen_initial_domain())
-		return -1;
+	now = __current_kernel_time();
+
+	/*
+	 * We only take the expensive HV call when the clock was set
+	 * or when the 11 minutes RTC synchronization time elapsed.
+	 */
+	if (!was_set && timespec_compare(&now, &next_sync) < 0)
+		return NOTIFY_OK;
 
 	op.cmd = XENPF_settime;
-	op.u.settime.secs = now;
-	op.u.settime.nsecs = 0;
+	op.u.settime.secs = now.tv_sec;
+	op.u.settime.nsecs = now.tv_nsec;
 	op.u.settime.system_time = xen_clocksource_read();
 
-	rc = HYPERVISOR_dom0_op(&op);
-	WARN(rc != 0, "XENPF_settime failed: now=%ld\n", now);
+	(void)HYPERVISOR_dom0_op(&op);
+
+	/*
+	 * Move the next drift compensation time 11 minutes
+	 * ahead. That's emulating the sync_cmos_clock() update for
+	 * the hardware RTC.
+	 */
+	next_sync = now;
+	next_sync.tv_sec += 11 * 60;
 
-	return rc;
+	return NOTIFY_OK;
 }
 
+static struct notifier_block xen_pvclock_gtod_notifier = {
+	.notifier_call = xen_pvclock_gtod_notify,
+};
+
 static struct clocksource xen_clocksource __read_mostly = {
 	.name = "xen",
 	.rating = 400,
@@ -364,11 +388,16 @@ static const struct clock_event_device xen_vcpuop_clockevent = {
 
 static const struct clock_event_device *xen_clockevent =
 	&xen_timerop_clockevent;
-static DEFINE_PER_CPU(struct clock_event_device, xen_clock_events);
+
+struct xen_clock_event_device {
+	struct clock_event_device evt;
+	char *name;
+};
+static DEFINE_PER_CPU(struct xen_clock_event_device, xen_clock_events) = { .evt.irq = -1 };
 
 static irqreturn_t xen_timer_interrupt(int irq, void *dev_id)
 {
-	struct clock_event_device *evt = &__get_cpu_var(xen_clock_events);
+	struct clock_event_device *evt = &__get_cpu_var(xen_clock_events).evt;
 	irqreturn_t ret;
 
 	ret = IRQ_NONE;
@@ -382,12 +411,31 @@ static irqreturn_t xen_timer_interrupt(int irq, void *dev_id)
 	return ret;
 }
 
+void xen_teardown_timer(int cpu)
+{
+	struct clock_event_device *evt;
+	BUG_ON(cpu == 0);
+	evt = &per_cpu(xen_clock_events, cpu).evt;
+
+	if (evt->irq >= 0) {
+		unbind_from_irqhandler(evt->irq, NULL);
+		evt->irq = -1;
+		kfree(per_cpu(xen_clock_events, cpu).name);
+		per_cpu(xen_clock_events, cpu).name = NULL;
+	}
+}
+
 void xen_setup_timer(int cpu)
 {
-	const char *name;
+	char *name;
 	struct clock_event_device *evt;
 	int irq;
 
+	evt = &per_cpu(xen_clock_events, cpu).evt;
+	WARN(evt->irq >= 0, "IRQ%d for CPU%d is already allocated\n", evt->irq, cpu);
+	if (evt->irq >= 0)
+		xen_teardown_timer(cpu);
+
 	printk(KERN_INFO "installing Xen timer for CPU %d\n", cpu);
 
 	name = kasprintf(GFP_KERNEL, "timer%d", cpu);
@@ -395,31 +443,24 @@ void xen_setup_timer(int cpu)
 		name = "<timer kasprintf failed>";
 
 	irq = bind_virq_to_irqhandler(VIRQ_TIMER, cpu, xen_timer_interrupt,
-				      IRQF_DISABLED|IRQF_PERCPU|
-				      IRQF_NOBALANCING|IRQF_TIMER|
+				      IRQF_PERCPU|IRQF_NOBALANCING|IRQF_TIMER|
 				      IRQF_FORCE_RESUME,
 				      name, NULL);
+	(void)xen_set_irq_priority(irq, XEN_IRQ_PRIORITY_MAX);
 
-	evt = &per_cpu(xen_clock_events, cpu);
 	memcpy(evt, xen_clockevent, sizeof(*evt));
 
 	evt->cpumask = cpumask_of(cpu);
 	evt->irq = irq;
+	per_cpu(xen_clock_events, cpu).name = name;
 }
 
-void xen_teardown_timer(int cpu)
-{
-	struct clock_event_device *evt;
-	BUG_ON(cpu == 0);
-	evt = &per_cpu(xen_clock_events, cpu);
-	unbind_from_irqhandler(evt->irq, NULL);
-}
 
 void xen_setup_cpu_clockevents(void)
 {
 	BUG_ON(preemptible());
 
-	clockevents_register_device(&__get_cpu_var(xen_clock_events));
+	clockevents_register_device(&__get_cpu_var(xen_clock_events).evt);
 }
 
 void xen_timer_resume(void)
@@ -464,6 +505,9 @@ static void __init xen_time_init(void)
 	xen_setup_runstate_info(cpu);
 	xen_setup_timer(cpu);
 	xen_setup_cpu_clockevents();
+
+	if (xen_initial_domain())
+		pvclock_gtod_register_notifier(&xen_pvclock_gtod_notifier);
 }
 
 void __init xen_init_time_ops(void)
@@ -476,7 +520,9 @@ void __init xen_init_time_ops(void)
 
 	x86_platform.calibrate_tsc = xen_tsc_khz;
 	x86_platform.get_wallclock = xen_get_wallclock;
-	x86_platform.set_wallclock = xen_set_wallclock;
+	/* Dom0 uses the native method to set the hardware RTC. */
+	if (!xen_initial_domain())
+		x86_platform.set_wallclock = xen_set_wallclock;
 }
 
 #ifdef CONFIG_XEN_PVHVM