[PATCH] x86_64: Add idle notifiers

This adds a new notifier chain that is called with IDLE_START
when a CPU goes idle and IDLE_END when it goes out of idle.
The context can be idle thread or interrupt context.

Since we cannot rely on MONITOR/MWAIT existing the idle
end check currently has to be done in all interrupt
handlers.

They were originally inspired by the similar s390 implementation.

They have a variety of applications:
- They will be needed for CONFIG_NO_IDLE_HZ
- They can be used for oprofile to fix up the missing time
in idle when performance counters don't tick.
- They can be used for better C state management in ACPI
- They could be used for microstate accounting.

This is just infrastructure so far, no users.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

6 files changed, 60 insertions, 0 deletions

diff --git a/arch/x86_64/kernel/apic.c b/arch/x86_64/kernel/apic.c
index 40cbd60e2d0b..628aebf9f9aa 100644
--- a/arch/x86_64/kernel/apic.c
+++ b/arch/x86_64/kernel/apic.c
@@ -34,6 +34,7 @@
 #include <asm/pgalloc.h>
 #include <asm/mach_apic.h>
 #include <asm/nmi.h>
+#include <asm/idle.h>
 
 int apic_verbosity;
 
@@ -922,6 +923,7 @@ void smp_apic_timer_interrupt(struct pt_regs *regs)
 	 * Besides, if we don't timer interrupts ignore the global
 	 * interrupt lock, which is the WrongThing (tm) to do.
 	 */
+	exit_idle();
 	irq_enter();
 	smp_local_timer_interrupt(regs);
 	irq_exit();
@@ -981,6 +983,7 @@ __init int oem_force_hpet_timer(void)
 asmlinkage void smp_spurious_interrupt(void)
 {
 	unsigned int v;
+	exit_idle();
 	irq_enter();
 	/*
 	 * Check if this really is a spurious interrupt and ACK it
@@ -1016,6 +1019,7 @@ asmlinkage void smp_error_interrupt(void)
 {
 	unsigned int v, v1;
 
+	exit_idle();
 	irq_enter();
 	/* First tickle the hardware, only then report what went on. -- REW */
 	v = apic_read(APIC_ESR);
diff --git a/arch/x86_64/kernel/irq.c b/arch/x86_64/kernel/irq.c
index d6a04a8320a3..8bf57f78d923 100644
--- a/arch/x86_64/kernel/irq.c
+++ b/arch/x86_64/kernel/irq.c
@@ -17,6 +17,7 @@
 #include <linux/delay.h>
 #include <asm/uaccess.h>
 #include <asm/io_apic.h>
+#include <asm/idle.h>
 
 atomic_t irq_err_count;
 #ifdef CONFIG_X86_IO_APIC
@@ -98,6 +99,7 @@ asmlinkage unsigned int do_IRQ(struct pt_regs *regs)
 	/* high bits used in ret_from_ code  */
 	unsigned irq = regs->orig_rax & 0xff;
 
+	exit_idle();
 	irq_enter();
 
 	__do_IRQ(irq, regs);
diff --git a/arch/x86_64/kernel/mce_amd.c b/arch/x86_64/kernel/mce_amd.c
index 1f76175ace02..0b62a44ab34b 100644
--- a/arch/x86_64/kernel/mce_amd.c
+++ b/arch/x86_64/kernel/mce_amd.c
@@ -27,6 +27,7 @@
 #include <asm/mce.h>
 #include <asm/msr.h>
 #include <asm/percpu.h>
+#include <asm/idle.h>
 
 #define PFX "mce_threshold: "
 #define VERSION "version 1.00.9"
@@ -140,6 +141,7 @@ asmlinkage void mce_threshold_interrupt(void)
 	struct mce m;
 
 	ack_APIC_irq();
+	exit_idle();
 	irq_enter();
 
 	memset(&m, 0, sizeof(m));
diff --git a/arch/x86_64/kernel/mce_intel.c b/arch/x86_64/kernel/mce_intel.c
index 0be0a7959814..3ef845e8d3b0 100644
--- a/arch/x86_64/kernel/mce_intel.c
+++ b/arch/x86_64/kernel/mce_intel.c
@@ -10,6 +10,7 @@
 #include <asm/msr.h>
 #include <asm/mce.h>
 #include <asm/hw_irq.h>
+#include <asm/idle.h>
 
 static DEFINE_PER_CPU(unsigned long, next_check);
 
@@ -19,6 +20,7 @@ asmlinkage void smp_thermal_interrupt(void)
 
 	ack_APIC_irq();
 
+	exit_idle();
 	irq_enter();
 	if (time_before(jiffies, __get_cpu_var(next_check)))
 		goto done;
diff --git a/arch/x86_64/kernel/process.c b/arch/x86_64/kernel/process.c
index c9df991150bb..669cf0ed3266 100644
--- a/arch/x86_64/kernel/process.c
+++ b/arch/x86_64/kernel/process.c
@@ -36,6 +36,7 @@
 #include <linux/utsname.h>
 #include <linux/random.h>
 #include <linux/kprobes.h>
+#include <linux/notifier.h>
 
 #include <asm/uaccess.h>
 #include <asm/pgtable.h>
@@ -50,6 +51,7 @@
 #include <asm/desc.h>
 #include <asm/proto.h>
 #include <asm/ia32.h>
+#include <asm/idle.h>
 
 asmlinkage extern void ret_from_fork(void);
 
@@ -64,6 +66,50 @@ EXPORT_SYMBOL(boot_option_idle_override);
 void (*pm_idle)(void);
 static DEFINE_PER_CPU(unsigned int, cpu_idle_state);
 
+static struct notifier_block *idle_notifier;
+static DEFINE_SPINLOCK(idle_notifier_lock);
+
+void idle_notifier_register(struct notifier_block *n)
+{
+	unsigned long flags;
+	spin_lock_irqsave(&idle_notifier_lock, flags);
+	notifier_chain_register(&idle_notifier, n);
+	spin_unlock_irqrestore(&idle_notifier_lock, flags);
+}
+EXPORT_SYMBOL_GPL(idle_notifier_register);
+
+void idle_notifier_unregister(struct notifier_block *n)
+{
+	unsigned long flags;
+	spin_lock_irqsave(&idle_notifier_lock, flags);
+	notifier_chain_unregister(&idle_notifier, n);
+	spin_unlock_irqrestore(&idle_notifier_lock, flags);
+}
+EXPORT_SYMBOL(idle_notifier_unregister);
+
+enum idle_state { CPU_IDLE, CPU_NOT_IDLE };
+static DEFINE_PER_CPU(enum idle_state, idle_state) = CPU_NOT_IDLE;
+
+void enter_idle(void)
+{
+	__get_cpu_var(idle_state) = CPU_IDLE;
+	notifier_call_chain(&idle_notifier, IDLE_START, NULL);
+}
+
+static void __exit_idle(void)
+{
+	__get_cpu_var(idle_state) = CPU_NOT_IDLE;
+	notifier_call_chain(&idle_notifier, IDLE_END, NULL);
+}
+
+/* Called from interrupts to signify idle end */
+void exit_idle(void)
+{
+	if (current->pid | read_pda(irqcount))
+		return;
+	__exit_idle();
+}
+
 /*
  * We use this if we don't have any better
  * idle routine..
@@ -180,7 +226,9 @@ void cpu_idle (void)
 				idle = default_idle;
 			if (cpu_is_offline(smp_processor_id()))
 				play_dead();
+			enter_idle();
 			idle();
+			__exit_idle();
 		}
 
 		preempt_enable_no_resched();
diff --git a/arch/x86_64/kernel/smp.c b/arch/x86_64/kernel/smp.c
index cfc3d9dccbd9..6a666d248c12 100644
--- a/arch/x86_64/kernel/smp.c
+++ b/arch/x86_64/kernel/smp.c
@@ -27,6 +27,7 @@
 #include <asm/mmu_context.h>
 #include <asm/proto.h>
 #include <asm/apicdef.h>
+#include <asm/idle.h>
 
 /*
  *	Smarter SMP flushing macros. 
@@ -512,6 +513,7 @@ asmlinkage void smp_call_function_interrupt(void)
 	/*
 	 * At this point the info structure may be out of scope unless wait==1
 	 */
+	exit_idle();
 	irq_enter();
 	(*func)(info);
 	irq_exit();