1 files changed, 102 insertions, 50 deletions

diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c
index b2bb5aa3b05..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>
@@ -26,8 +28,6 @@
 
 #include "xen-ops.h"
 
-#define XEN_SHIFT 22
-
 /* Xen may fire a timer up to this many ns early */
 #define TIMER_SLOP	100000
 #define NS_PER_TICK	(1000000000LL / HZ)
@@ -38,9 +38,8 @@ static DEFINE_PER_CPU(struct vcpu_runstate_info, xen_runstate);
 /* snapshots of runstate info */
 static DEFINE_PER_CPU(struct vcpu_runstate_info, xen_runstate_snapshot);
 
-/* unused ns of stolen and blocked time */
+/* unused ns of stolen time */
 static DEFINE_PER_CPU(u64, xen_residual_stolen);
-static DEFINE_PER_CPU(u64, xen_residual_blocked);
 
 /* return an consistent snapshot of 64-bit time/counter value */
 static u64 get64(const u64 *p)
@@ -117,7 +116,7 @@ static void do_stolen_accounting(void)
 {
 	struct vcpu_runstate_info state;
 	struct vcpu_runstate_info *snap;
-	s64 blocked, runnable, offline, stolen;
+	s64 runnable, offline, stolen;
 	cputime_t ticks;
 
 	get_runstate_snapshot(&state);
@@ -127,7 +126,6 @@ static void do_stolen_accounting(void)
 	snap = &__get_cpu_var(xen_runstate_snapshot);
 
 	/* work out how much time the VCPU has not been runn*ing*  */
-	blocked = state.time[RUNSTATE_blocked] - snap->time[RUNSTATE_blocked];
 	runnable = state.time[RUNSTATE_runnable] - snap->time[RUNSTATE_runnable];
 	offline = state.time[RUNSTATE_offline] - snap->time[RUNSTATE_offline];
 
@@ -135,25 +133,14 @@ static void do_stolen_accounting(void)
 
 	/* Add the appropriate number of ticks of stolen time,
 	   including any left-overs from last time. */
-	stolen = runnable + offline + __get_cpu_var(xen_residual_stolen);
+	stolen = runnable + offline + __this_cpu_read(xen_residual_stolen);
 
 	if (stolen < 0)
 		stolen = 0;
 
 	ticks = iter_div_u64_rem(stolen, NS_PER_TICK, &stolen);
-	__get_cpu_var(xen_residual_stolen) = stolen;
+	__this_cpu_write(xen_residual_stolen, stolen);
 	account_steal_ticks(ticks);
-
-	/* Add the appropriate number of ticks of blocked time,
-	   including any left-overs from last time. */
-	blocked += __get_cpu_var(xen_residual_blocked);
-
-	if (blocked < 0)
-		blocked = 0;
-
-	ticks = iter_div_u64_rem(blocked, NS_PER_TICK, &blocked);
-	__get_cpu_var(xen_residual_blocked) = blocked;
-	account_idle_ticks(ticks);
 }
 
 /* Get the TSC speed from Xen */
@@ -170,9 +157,10 @@ cycle_t xen_clocksource_read(void)
         struct pvclock_vcpu_time_info *src;
 	cycle_t ret;
 
-	src = &get_cpu_var(xen_vcpu)->time;
+	preempt_disable_notrace();
+	src = &__get_cpu_var(xen_vcpu)->time;
 	ret = pvclock_clocksource_read(src);
-	put_cpu_var(xen_vcpu);
+	preempt_enable_notrace();
 	return ret;
 }
 
@@ -192,27 +180,61 @@ 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(&ts);
-	return ts.tv_sec;
+	xen_read_wallclock(now);
 }
 
-static int xen_set_wallclock(unsigned long now)
+static int xen_set_wallclock(const struct timespec *now)
 {
-	/* do nothing for domU */
 	return -1;
 }
 
+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;
+	struct timespec now;
+
+	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.tv_sec;
+	op.u.settime.nsecs = now.tv_nsec;
+	op.u.settime.system_time = xen_clocksource_read();
+
+	(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 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,
 	.read = xen_clocksource_get_cycles,
 	.mask = ~0,
-	.mult = 1<<XEN_SHIFT,		/* time directly in nanoseconds */
-	.shift = XEN_SHIFT,
 	.flags = CLOCK_SOURCE_IS_CONTINUOUS,
 };
 
@@ -366,11 +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;
@@ -384,12 +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);
@@ -397,35 +443,32 @@ 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)
 {
 	int cpu;
 
+	pvclock_resume();
+
 	if (xen_clockevent != &xen_vcpuop_clockevent)
 		return;
 
@@ -435,16 +478,16 @@ void xen_timer_resume(void)
 	}
 }
 
-static const struct pv_time_ops xen_time_ops __initdata = {
+static const struct pv_time_ops xen_time_ops __initconst = {
 	.sched_clock = xen_clocksource_read,
 };
 
-static __init void xen_time_init(void)
+static void __init xen_time_init(void)
 {
 	int cpu = smp_processor_id();
 	struct timespec tp;
 
-	clocksource_register(&xen_clocksource);
+	clocksource_register_hz(&xen_clocksource, NSEC_PER_SEC);
 
 	if (HYPERVISOR_vcpu_op(VCPUOP_stop_periodic_timer, cpu, NULL) == 0) {
 		/* Successfully turned off 100Hz tick, so we have the
@@ -462,9 +505,12 @@ static __init void 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);
 }
 
-__init void xen_init_time_ops(void)
+void __init xen_init_time_ops(void)
 {
 	pv_time_ops = xen_time_ops;
 
@@ -474,7 +520,9 @@ __init void 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
@@ -482,11 +530,15 @@ static void xen_hvm_setup_cpu_clockevents(void)
 {
 	int cpu = smp_processor_id();
 	xen_setup_runstate_info(cpu);
-	xen_setup_timer(cpu);
+	/*
+	 * xen_setup_timer(cpu) - snprintf is bad in atomic context. Hence
+	 * doing it xen_hvm_cpu_notify (which gets called by smp_init during
+	 * early bootup and also during CPU hotplug events).
+	 */
 	xen_setup_cpu_clockevents();
 }
 
-__init void xen_hvm_init_time_ops(void)
+void __init xen_hvm_init_time_ops(void)
 {
 	/* vector callback is needed otherwise we cannot receive interrupts
 	 * on cpu > 0 and at this point we don't know how many cpus are