[PATCH] tick-management: broadcast functionality

With Ingo Molnar <mingo@elte.hu>

Add broadcast functionality, so per cpu clock event devices can be registered
as dummy devices or switched from/to broadcast on demand.  The broadcast
function distributes the events via the broadcast function of the clock event
device.  This is primarily designed to replace the switch apic timer to / from
IPI in power states, where the apic stops.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: john stultz <johnstul@us.ibm.com>
Cc: Roman Zippel <zippel@linux-m68k.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

4 files changed, 382 insertions, 11 deletions

diff --git a/kernel/time/Makefile b/kernel/time/Makefile
index 337daef72755..a941743c3ff8 100644
--- a/kernel/time/Makefile
+++ b/kernel/time/Makefile
@@ -1,4 +1,5 @@
 obj-y += ntp.o clocksource.o jiffies.o
 
-obj-$(CONFIG_GENERIC_CLOCKEVENTS)       += clockevents.o
-obj-$(CONFIG_GENERIC_CLOCKEVENTS)       += tick-common.o
+obj-$(CONFIG_GENERIC_CLOCKEVENTS)               += clockevents.o
+obj-$(CONFIG_GENERIC_CLOCKEVENTS)               += tick-common.o
+obj-$(CONFIG_GENERIC_CLOCKEVENTS_BROADCAST)     += tick-broadcast.o
diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c
new file mode 100644
index 000000000000..0ee4968ff791
--- /dev/null
+++ b/kernel/time/tick-broadcast.c
@@ -0,0 +1,270 @@
+/*
+ * linux/kernel/time/tick-broadcast.c
+ *
+ * This file contains functions which emulate a local clock-event
+ * device via a broadcast event source.
+ *
+ * Copyright(C) 2005-2006, Thomas Gleixner <tglx@linutronix.de>
+ * Copyright(C) 2005-2007, Red Hat, Inc., Ingo Molnar
+ * Copyright(C) 2006-2007, Timesys Corp., Thomas Gleixner
+ *
+ * This code is licenced under the GPL version 2. For details see
+ * kernel-base/COPYING.
+ */
+#include <linux/cpu.h>
+#include <linux/err.h>
+#include <linux/hrtimer.h>
+#include <linux/irq.h>
+#include <linux/percpu.h>
+#include <linux/profile.h>
+#include <linux/sched.h>
+#include <linux/tick.h>
+
+#include "tick-internal.h"
+
+/*
+ * Broadcast support for broken x86 hardware, where the local apic
+ * timer stops in C3 state.
+ */
+
+struct tick_device tick_broadcast_device;
+static cpumask_t tick_broadcast_mask;
+DEFINE_SPINLOCK(tick_broadcast_lock);
+
+/*
+ * Start the device in periodic mode
+ */
+static void tick_broadcast_start_periodic(struct clock_event_device *bc)
+{
+        if (bc && bc->mode == CLOCK_EVT_MODE_SHUTDOWN)
+                tick_setup_periodic(bc, 1);
+}
+
+/*
+ * Check, if the device can be utilized as broadcast device:
+ */
+int tick_check_broadcast_device(struct clock_event_device *dev)
+{
+        if (tick_broadcast_device.evtdev ||
+            (dev->features & CLOCK_EVT_FEAT_C3STOP))
+                return 0;
+
+        clockevents_exchange_device(NULL, dev);
+        tick_broadcast_device.evtdev = dev;
+        if (!cpus_empty(tick_broadcast_mask))
+                tick_broadcast_start_periodic(dev);
+        return 1;
+}
+
+/*
+ * Check, if the device is the broadcast device
+ */
+int tick_is_broadcast_device(struct clock_event_device *dev)
+{
+        return (dev && tick_broadcast_device.evtdev == dev);
+}
+
+/*
+ * Check, if the device is disfunctional and a place holder, which
+ * needs to be handled by the broadcast device.
+ */
+int tick_device_uses_broadcast(struct clock_event_device *dev, int cpu)
+{
+        unsigned long flags;
+        int ret = 0;
+
+        spin_lock_irqsave(&tick_broadcast_lock, flags);
+
+        /*
+         * Devices might be registered with both periodic and oneshot
+         * mode disabled. This signals, that the device needs to be
+         * operated from the broadcast device and is a placeholder for
+         * the cpu local device.
+         */
+        if (!tick_device_is_functional(dev)) {
+                dev->event_handler = tick_handle_periodic;
+                cpu_set(cpu, tick_broadcast_mask);
+                tick_broadcast_start_periodic(tick_broadcast_device.evtdev);
+                ret = 1;
+        }
+
+        spin_unlock_irqrestore(&tick_broadcast_lock, flags);
+        return ret;
+}
+
+/*
+ * Broadcast the event to the cpus, which are set in the mask
+ */
+int tick_do_broadcast(cpumask_t mask)
+{
+        int ret = 0, cpu = smp_processor_id();
+        struct tick_device *td;
+
+        /*
+         * Check, if the current cpu is in the mask
+         */
+        if (cpu_isset(cpu, mask)) {
+                cpu_clear(cpu, mask);
+                td = &per_cpu(tick_cpu_device, cpu);
+                td->evtdev->event_handler(td->evtdev);
+                ret = 1;
+        }
+
+        if (!cpus_empty(mask)) {
+                /*
+                 * It might be necessary to actually check whether the devices
+                 * have different broadcast functions. For now, just use the
+                 * one of the first device. This works as long as we have this
+                 * misfeature only on x86 (lapic)
+                 */
+                cpu = first_cpu(mask);
+                td = &per_cpu(tick_cpu_device, cpu);
+                td->evtdev->broadcast(mask);
+                ret = 1;
+        }
+        return ret;
+}
+
+/*
+ * Periodic broadcast:
+ * - invoke the broadcast handlers
+ */
+static void tick_do_periodic_broadcast(void)
+{
+        cpumask_t mask;
+
+        spin_lock(&tick_broadcast_lock);
+
+        cpus_and(mask, cpu_online_map, tick_broadcast_mask);
+        tick_do_broadcast(mask);
+
+        spin_unlock(&tick_broadcast_lock);
+}
+
+/*
+ * Event handler for periodic broadcast ticks
+ */
+static void tick_handle_periodic_broadcast(struct clock_event_device *dev)
+{
+        dev->next_event.tv64 = KTIME_MAX;
+
+        tick_do_periodic_broadcast();
+
+        /*
+         * The device is in periodic mode. No reprogramming necessary:
+         */
+        if (dev->mode == CLOCK_EVT_MODE_PERIODIC)
+                return;
+
+        /*
+         * Setup the next period for devices, which do not have
+         * periodic mode:
+         */
+        for (;;) {
+                ktime_t next = ktime_add(dev->next_event, tick_period);
+
+                if (!clockevents_program_event(dev, next, ktime_get()))
+                        return;
+                tick_do_periodic_broadcast();
+        }
+}
+
+/*
+ * Powerstate information: The system enters/leaves a state, where
+ * affected devices might stop
+ */
+static void tick_do_broadcast_on_off(void *why)
+{
+        struct clock_event_device *bc, *dev;
+        struct tick_device *td;
+        unsigned long flags, *reason = why;
+        int cpu;
+
+        spin_lock_irqsave(&tick_broadcast_lock, flags);
+