Merge branch 'irq/core' into x86/irq

Reason: Enabling irq threads and update to latest genirq functionality
	requires the core code

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

Documentation/kernel-parameters.txt

@@ -2444,6 +2444,10 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
 			<deci-seconds>: poll all this frequency
 			0: no polling (default)
 
+	threadirqs	[KNL]
+			Force threading of all interrupt handlers except those
+			marked explicitely IRQF_NO_THREAD.
+
 	topology=	[S390]
 			Format: {off | on}
 			Specify if the kernel should make use of the cpu

include/linux/interrupt.h

@@ -14,6 +14,8 @@
 #include <linux/smp.h>
 #include <linux/percpu.h>
 #include <linux/hrtimer.h>
+#include <linux/kref.h>
+#include <linux/workqueue.h>
 
 #include <asm/atomic.h>
 #include <asm/ptrace.h>
@@ -55,7 +57,8 @@
  *                Used by threaded interrupts which need to keep the
  *                irq line disabled until the threaded handler has been run.
  * IRQF_NO_SUSPEND - Do not disable this IRQ during suspend
- *
+ * IRQF_FORCE_RESUME - Force enable it on resume even if IRQF_NO_SUSPEND is set
+ * IRQF_NO_THREAD - Interrupt cannot be threaded
  */
 #define IRQF_DISABLED		0x00000020
 #define IRQF_SAMPLE_RANDOM	0x00000040
@@ -67,22 +70,10 @@
 #define IRQF_IRQPOLL		0x00001000
 #define IRQF_ONESHOT		0x00002000
 #define IRQF_NO_SUSPEND		0x00004000
+#define IRQF_FORCE_RESUME	0x00008000
+#define IRQF_NO_THREAD		0x00010000
 
-#define IRQF_TIMER		(__IRQF_TIMER | IRQF_NO_SUSPEND)
-
-/*
- * Bits used by threaded handlers:
- * IRQTF_RUNTHREAD - signals that the interrupt handler thread should run
- * IRQTF_DIED      - handler thread died
- * IRQTF_WARNED    - warning "IRQ_WAKE_THREAD w/o thread_fn" has been printed
- * IRQTF_AFFINITY  - irq thread is requested to adjust affinity
- */
-enum {
-	IRQTF_RUNTHREAD,
-	IRQTF_DIED,
-	IRQTF_WARNED,
-	IRQTF_AFFINITY,
-};
+#define IRQF_TIMER		(__IRQF_TIMER | IRQF_NO_SUSPEND | IRQF_NO_THREAD)
 
 /*
  * These values can be returned by request_any_context_irq() and
@@ -110,6 +101,7 @@ typedef irqreturn_t (*irq_handler_t)(int, void *);
  * @thread_fn:	interupt handler function for threaded interrupts
  * @thread:	thread pointer for threaded interrupts
  * @thread_flags:	flags related to @thread
+ * @thread_mask:	bitmask for keeping track of @thread activity
  */
 struct irqaction {
 	irq_handler_t handler;
@@ -120,6 +112,7 @@ struct irqaction {
 	irq_handler_t thread_fn;
 	struct task_struct *thread;
 	unsigned long thread_flags;
+	unsigned long thread_mask;
 	const char *name;
 	struct proc_dir_entry *dir;
 } ____cacheline_internodealigned_in_smp;
@@ -240,6 +233,35 @@ extern int irq_can_set_affinity(unsigned int irq);
 extern int irq_select_affinity(unsigned int irq);
 
 extern int irq_set_affinity_hint(unsigned int irq, const struct cpumask *m);
+
+/**
+ * struct irq_affinity_notify - context for notification of IRQ affinity changes
+ * @irq:		Interrupt to which notification applies
+ * @kref:		Reference count, for internal use
+ * @work:		Work item, for internal use
+ * @notify:		Function to be called on change.  This will be
+ *			called in process context.
+ * @release:		Function to be called on release.  This will be
+ *			called in process context.  Once registered, the
+ *			structure must only be freed when this function is
+ *			called or later.
+ */
+struct irq_affinity_notify {
+	unsigned int irq;
+	struct kref kref;
+	struct work_struct work;
+	void (*notify)(struct irq_affinity_notify *, const cpumask_t *mask);
+	void (*release)(struct kref *ref);
+};
+
+extern int
+irq_set_affinity_notifier(unsigned int irq, struct irq_affinity_notify *notify);
+
+static inline void irq_run_affinity_notifiers(void)
+{
+	flush_scheduled_work();
+}
+
 #else /* CONFIG_SMP */
 
 static inline int irq_set_affinity(unsigned int irq, const struct cpumask *m)
@@ -255,7 +277,7 @@ static inline int irq_can_set_affinity(unsigned int irq)
 static inline int irq_select_affinity(unsigned int irq)  { return 0; }
 
 static inline int irq_set_affinity_hint(unsigned int irq,
-                                        const struct cpumask *m)
+					const struct cpumask *m)
 {
 	return -EINVAL;
 }
@@ -314,16 +336,24 @@ static inline void enable_irq_lockdep_irqrestore(unsigned int irq, unsigned long
 }
 
 /* IRQ wakeup (PM) control: */
-extern int set_irq_wake(unsigned int irq, unsigned int on);
+extern int irq_set_irq_wake(unsigned int irq, unsigned int on);
+
+#ifndef CONFIG_GENERIC_HARDIRQS_NO_COMPAT
+/* Please do not use: Use the replacement functions instead */
+static inline int set_irq_wake(unsigned int irq, unsigned int on)
+{
+	return irq_set_irq_wake(irq, on);
+}
+#endif
 
 static inline int enable_irq_wake(unsigned int irq)
 {
-	return set_irq_wake(irq, 1);
+	return irq_set_irq_wake(irq, 1);
 }
 
 static inline int disable_irq_wake(unsigned int irq)
 {
-	return set_irq_wake(irq, 0);
+	return irq_set_irq_wake(irq, 0);
 }
 
 #else /* !CONFIG_GENERIC_HARDIRQS */
@@ -353,6 +383,13 @@ static inline int disable_irq_wake(unsigned int irq)
 }
 #endif /* CONFIG_GENERIC_HARDIRQS */
 
+
+#ifdef CONFIG_IRQ_FORCED_THREADING
+extern bool force_irqthreads;
+#else
+#define force_irqthreads	(0)
+#endif
+
 #ifndef __ARCH_SET_SOFTIRQ_PENDING
 #define set_softirq_pending(x) (local_softirq_pending() = (x))
 #define or_softirq_pending(x)  (local_softirq_pending() |= (x))
@@ -645,6 +682,7 @@ static inline void init_irq_proc(void)
 
 struct seq_file;
 int show_interrupts(struct seq_file *p, void *v);
+int arch_show_interrupts(struct seq_file *p, int prec);
 
 extern int early_irq_init(void);
 extern int arch_probe_nr_irqs(void);

include/linux/irqdesc.h

@@ -8,6 +8,7 @@
  * For now it's included from <linux/irq.h>
  */
 
+struct irq_affinity_notify;
 struct proc_dir_entry;
 struct timer_rand_state;
 /**
@@ -18,13 +19,16 @@ struct timer_rand_state;
  * @handle_irq:		highlevel irq-events handler [if NULL, __do_IRQ()]
  * @action:		the irq action chain
  * @status:		status information
+ * @core_internal_state__do_not_mess_with_it: core internal status information
  * @depth:		disable-depth, for nested irq_disable() calls
  * @wake_depth:		enable depth, for multiple set_irq_wake() callers
  * @irq_count:		stats field to detect stalled irqs
  * @last_unhandled:	aging timer for unhandled count
  * @irqs_unhandled:	stats field for spurious unhandled interrupts
  * @lock:		locking for SMP
+ * @affinity_notify:	context for notification of affinity changes
  * @pending_mask:	pending rebalanced interrupts
+ * @threads_oneshot:	bitfield to handle shared oneshot threads
  * @threads_active:	number of irqaction threads currently running
  * @wait_for_threads:	wait queue for sync_irq to wait for threaded handlers
  * @dir:		/proc/irq/ procfs entry
@@ -45,6 +49,7 @@ struct irq_desc {
 		struct {
 			unsigned int		irq;
 			unsigned int		node;
+			unsigned int		pad_do_not_even_think_about_it;
 			struct irq_chip		*chip;
 			void			*handler_data;
 			void			*chip_data;
@@ -59,9 +64,16 @@ struct irq_desc {
 	struct timer_rand_state *timer_rand_state;
 	unsigned int __percpu	*kstat_irqs;
 	irq_flow_handler_t	handle_irq;
+#ifdef CONFIG_IRQ_PREFLOW_FASTEOI
+	irq_preflow_handler_t	preflow_handler;
+#endif
 	struct irqaction	*action;	/* IRQ action list */
+#ifdef CONFIG_GENERIC_HARDIRQS_NO_COMPAT
+	unsigned int		status_use_accessors;
+#else
 	unsigned int		status;		/* IRQ status */
-
+#endif
+	unsigned int		core_internal_state__do_not_mess_with_it;
 	unsigned int		depth;		/* nested irq disables */
 	unsigned int		wake_depth;	/* nested wake enables */
 	unsigned int		irq_count;	/* For detecting broken IRQs */
@@ -70,10 +82,12 @@ struct irq_desc {
 	raw_spinlock_t		lock;
 #ifdef CONFIG_SMP
 	const struct cpumask	*affinity_hint;
+	struct irq_affinity_notify *affinity_notify;
 #ifdef CONFIG_GENERIC_PENDING_IRQ
 	cpumask_var_t		pending_mask;
 #endif
 #endif
+	unsigned long		threads_oneshot;
 	atomic_t		threads_active;
 	wait_queue_head_t       wait_for_threads;
 #ifdef CONFIG_PROC_FS
@@ -95,10 +109,51 @@ static inline struct irq_desc *move_irq_desc(struct irq_desc *desc, int node)
 
 #ifdef CONFIG_GENERIC_HARDIRQS
 
-#define get_irq_desc_chip(desc)		((desc)->irq_data.chip)
-#define get_irq_desc_chip_data(desc)	((desc)->irq_data.chip_data)
-#define get_irq_desc_data(desc)		((desc)->irq_data.handler_data)
-#define get_irq_desc_msi(desc)		((desc)->irq_data.msi_desc)
+static inline struct irq_data *irq_desc_get_irq_data(struct irq_desc *desc)
+{
+	return &desc->irq_data;
+}
+
+static inline struct irq_chip *irq_desc_get_chip(struct irq_desc *desc)
+{
+	return desc->irq_data.chip;
+}
+
+static inline void *irq_desc_get_chip_data(struct irq_desc *desc)
+{
+	return desc->irq_data.chip_data;
+}
+
+static inline void *irq_desc_get_handler_data(struct irq_desc *desc)
+{
+	return desc->irq_data.handler_data;
+}
+
+static inline struct msi_desc *irq_desc_get_msi_desc(struct irq_desc *desc)
+{
+	return desc->irq_data.msi_desc;
+}
+
+#ifndef CONFIG_GENERIC_HARDIRQS_NO_COMPAT
+static inline struct irq_chip *get_irq_desc_chip(struct irq_desc *desc)
+{
+	return irq_desc_get_chip(desc);
+}
+static inline void *get_irq_desc_data(struct irq_desc *desc)
+{
+	return irq_desc_get_handler_data(desc);
+}
+
+static inline void *get_irq_desc_chip_data(struct irq_desc *desc)
+{
+	return irq_desc_get_chip_data(desc);
+}
+
+static inline struct msi_desc *get_irq_desc_msi(struct irq_desc *desc)
+{
+	return irq_desc_get_msi_desc(desc);
+}
+#endif
 
 /*
  * Architectures call this to let the generic IRQ layer
@@ -123,6 +178,7 @@ static inline int irq_has_action(unsigned int irq)
 	return desc->action != NULL;
 }
 
+#ifndef CONFIG_GENERIC_HARDIRQS_NO_COMPAT
 static inline int irq_balancing_disabled(unsigned int irq)
 {
 	struct irq_desc *desc;
@@ -130,6 +186,7 @@ static inline int irq_balancing_disabled(unsigned int irq)
 	desc = irq_to_desc(irq);
 	return desc->status & IRQ_NO_BALANCING_MASK;
 }
+#endif
 
 /* caller has locked the irq_desc and both params are valid */
 static inline void __set_irq_handler_unlocked(int irq,
@@ -140,6 +197,17 @@ static inline void __set_irq_handler_unlocked(int irq,
 	desc = irq_to_desc(irq);
 	desc->handle_irq = handler;
 }
+
+#ifdef CONFIG_IRQ_PREFLOW_FASTEOI
+static inline void
+__irq_set_preflow_handler(unsigned int irq, irq_preflow_handler_t handler)
+{
+	struct irq_desc *desc;
+
+	desc = irq_to_desc(irq);
+	desc->preflow_handler = handler;
+}
+#endif
 #endif
 
 #endif

kernel/irq/Kconfig

@@ -1,5 +1,6 @@
+# Select this to activate the generic irq options below
 config HAVE_GENERIC_HARDIRQS
-	def_bool n
+	bool
 
 if HAVE_GENERIC_HARDIRQS
 menu "IRQ subsystem"
@@ -11,26 +12,44 @@ config GENERIC_HARDIRQS
 
 # Select this to disable the deprecated stuff
 config GENERIC_HARDIRQS_NO_DEPRECATED
-       def_bool n
+       bool
+
+config GENERIC_HARDIRQS_NO_COMPAT
+       bool
 
 # Options selectable by the architecture code
+
+# Make sparse irq Kconfig switch below available
 config HAVE_SPARSE_IRQ
-       def_bool n
+       bool
 
+# Enable the generic irq autoprobe mechanism
 config GENERIC_IRQ_PROBE
-	def_bool n
+	bool
 
+# Use the generic /proc/interrupts implementation
+config GENERIC_IRQ_SHOW
+       bool
+
+# Support for delayed migration from interrupt context
 config GENERIC_PENDING_IRQ
-	def_bool n
+	bool
 
+# Alpha specific irq affinity mechanism
 config AUTO_IRQ_AFFINITY
-       def_bool n
-
-config IRQ_PER_CPU
-       def_bool n
+       bool
 
+# Tasklet based software resend for pending interrupts on enable_irq()
 config HARDIRQS_SW_RESEND
-       def_bool n
+       bool
+
+# Preflow handler support for fasteoi (sparc64)
+config IRQ_PREFLOW_FASTEOI
+       bool
+
+# Support forced irq threading
+config IRQ_FORCED_THREADING
+       bool
 
 config SPARSE_IRQ
 	bool "Support sparse irq numbering"

kernel/irq/autoprobe.c

@@ -17,7 +17,7 @@
 /*
  * Autodetection depends on the fact that any interrupt that
  * comes in on to an unassigned handler will get stuck with
- * "IRQ_WAITING" cleared and the interrupt disabled.
+ * "IRQS_WAITING" cleared and the interrupt disabled.
  */
 static DEFINE_MUTEX(probing_active);
 
@@ -32,7 +32,6 @@ unsigned long probe_irq_on(void)
 {
 	struct irq_desc *desc;
 	unsigned long mask = 0;
-	unsigned int status;
 	int i;
 
 	/*
@@ -46,13 +45,7 @@ unsigned long probe_irq_on(void)
 	 */
 	for_each_irq_desc_reverse(i, desc) {
 		raw_spin_lock_irq(&desc->lock);
-		if (!desc->action && !(desc->status & IRQ_NOPROBE)) {
-			/*
-			 * An old-style architecture might still have
-			 * the handle_bad_irq handler there:
-			 */
-			compat_irq_chip_set_default_handler(desc);
-
+		if (!desc->action && irq_settings_can_probe(desc)) {
 			/*
 			 * Some chips need to know about probing in
 			 * progress:
@@ -60,7 +53,7 @@ unsigned long probe_irq_on(void)
 			if (desc->irq_data.chip->irq_set_type)
 				desc->irq_data.chip->irq_set_type(&desc->irq_data,
 							 IRQ_TYPE_PROBE);
-			desc->irq_data.chip->irq_startup(&desc->irq_data);
+			irq_startup(desc);
 		}
 		raw_spin_unlock_irq(&desc->lock);
 	}
@@ -75,10 +68,12 @@ unsigned long probe_irq_on(void)
 	 */
 	for_each_irq_desc_reverse(i, desc) {
 		raw_spin_lock_irq(&desc->lock);
-		if (!desc->action && !(desc->status & IRQ_NOPROBE)) {
-			desc->status |= IRQ_AUTODETECT | IRQ_WAITING;
-			if (desc->irq_data.chip->irq_startup(&desc->irq_data))
-				desc->status |= IRQ_PENDING;
+		if (!desc->action && irq_settings_can_probe(desc)) {
+			desc->istate |= IRQS_AUTODETECT | IRQS_WAITING;
+			if (irq_startup(desc)) {
+				irq_compat_set_pending(desc);
+				desc->istate |= IRQS_PENDING;
+			}
 		}
 		raw_spin_unlock_irq(&desc->lock);
 	}
@@ -93,13 +88,12 @@ unsigned long probe_irq_on(void)
 	 */
 	for_each_irq_desc(i, desc) {
 		raw_spin_lock_irq(&desc->lock);
-		status = desc->status;
 
-		if (status & IRQ_AUTODETECT) {
+		if (desc->istate & IRQS_AUTODETECT) {
 			/* It triggered already - consider it spurious. */
-			if (!(status & IRQ_WAITING)) {
-				desc->status = status & ~IRQ_AUTODETECT;
-				desc->irq_data.chip->irq_shutdown(&desc->irq_data);
+			if (!(desc->istate & IRQS_WAITING)) {
+				desc->istate &= ~IRQS_AUTODETECT;
+				irq_shutdown(desc);
 			} else
 				if (i < 32)
 					mask |= 1 << i;
@@ -125,20 +119,18 @@ EXPORT_SYMBOL(probe_irq_on);
  */
 unsigned int probe_irq_mask(unsigned long val)
 {
-	unsigned int status, mask = 0;
+	unsigned int mask = 0;
 	struct irq_desc *desc;
 	int i;
 
 	for_each_irq_desc(i, desc) {
 		raw_spin_lock_irq(&desc->lock);
-		status = desc->status;
-
-		if (status & IRQ_AUTODETECT) {
-			if (i < 16 && !(status & IRQ_WAITING))
+		if (desc->istate & IRQS_AUTODETECT) {
+			if (i < 16 && !(desc->istate & IRQS_WAITING))
 				mask |= 1 << i;
 
-			desc->status = status & ~IRQ_AUTODETECT;
-			desc->irq_data.chip->irq_shutdown(&desc->irq_data);
+			desc->istate &= ~IRQS_AUTODETECT;
+			irq_shutdown(desc);
 		}
 		raw_spin_unlock_irq(&desc->lock);
 	}
@@ -169,20 +161,18 @@ int probe_irq_off(unsigned long val)
 {
 	int i, irq_found = 0, nr_of_irqs = 0;
 	struct irq_desc *desc;
-	unsigned int status;
 
 	for_each_irq_desc(i, desc) {
 		raw_spin_lock_irq(&desc->lock);
-		status = desc->status;
 
-		if (status & IRQ_AUTODETECT) {
-			if (!(status & IRQ_WAITING)) {
+		if (desc->istate & IRQS_AUTODETECT) {
+			if (!(desc->istate & IRQS_WAITING)) {
 				if (!nr_of_irqs)
 					irq_found = i;
 				nr_of_irqs++;
 			}
-			desc->status = status & ~IRQ_AUTODETECT;
-			desc->irq_data.chip->irq_shutdown(&desc->irq_data);
+			desc->istate &= ~IRQS_AUTODETECT;
+			irq_shutdown(desc);
 		}
 		raw_spin_unlock_irq(&desc->lock);
 	}

kernel/irq/compat.h

@@ -0,0 +1,72 @@
+/*
+ * Compat layer for transition period
+ */
+#ifndef CONFIG_GENERIC_HARDIRQS_NO_COMPAT
+static inline void irq_compat_set_progress(struct irq_desc *desc)
+{
+	desc->status |= IRQ_INPROGRESS;
+}
+
+static inline void irq_compat_clr_progress(struct irq_desc *desc)
+{
+	desc->status &= ~IRQ_INPROGRESS;
+}
+static inline void irq_compat_set_disabled(struct irq_desc *desc)
+{
+	desc->status |= IRQ_DISABLED;
+}
+static inline void irq_compat_clr_disabled(struct irq_desc *desc)
+{
+	desc->status &= ~IRQ_DISABLED;
+}
+static inline void irq_compat_set_pending(struct irq_desc *desc)
+{
+	desc->status |= IRQ_PENDING;
+}
+
+static inline void irq_compat_clr_pending(struct irq_desc *desc)
+{
+	desc->status &= ~IRQ_PENDING;
+}
+static inline void irq_compat_set_masked(struct irq_desc *desc)
+{
+	desc->status |= IRQ_MASKED;
+}
+
+static inline void irq_compat_clr_masked(struct irq_desc *desc)
+{
+	desc->status &= ~IRQ_MASKED;
+}
+static inline void irq_compat_set_move_pending(struct irq_desc *desc)
+{
+	desc->status |= IRQ_MOVE_PENDING;
+}
+
+static inline void irq_compat_clr_move_pending(struct irq_desc *desc)
+{
+	desc->status &= ~IRQ_MOVE_PENDING;
+}
+static inline void irq_compat_set_affinity(struct irq_desc *desc)
+{
+	desc->status |= IRQ_AFFINITY_SET;
+}
+
+static inline void irq_compat_clr_affinity(struct irq_desc *desc)
+{
+	desc->status &= ~IRQ_AFFINITY_SET;
+}
+#else
+static inline void irq_compat_set_progress(struct irq_desc *desc) { }
+static inline void irq_compat_clr_progress(struct irq_desc *desc) { }
+static inline void irq_compat_set_disabled(struct irq_desc *desc) { }
+static inline void irq_compat_clr_disabled(struct irq_desc *desc) { }
+static inline void irq_compat_set_pending(struct irq_desc *desc) { }
+static inline void irq_compat_clr_pending(struct irq_desc *desc) { }
+static inline void irq_compat_set_masked(struct irq_desc *desc) { }
+static inline void irq_compat_clr_masked(struct irq_desc *desc) { }
+static inline void irq_compat_set_move_pending(struct irq_desc *desc) { }
+static inline void irq_compat_clr_move_pending(struct irq_desc *desc) { }
+static inline void irq_compat_set_affinity(struct irq_desc *desc) { }
+static inline void irq_compat_clr_affinity(struct irq_desc *desc) { }
+#endif
+

kernel/irq/debug.h

@@ -0,0 +1,40 @@
+/*
+ * Debugging printout:
+ */
+
+#include <linux/kallsyms.h>
+
+#define P(f) if (desc->status & f) printk("%14s set\n", #f)
+#define PS(f) if (desc->istate & f) printk("%14s set\n", #f)
+
+static inline void print_irq_desc(unsigned int irq, struct irq_desc *desc)
+{
+	printk("irq %d, desc: %p, depth: %d, count: %d, unhandled: %d\n",
+		irq, desc, desc->depth, desc->irq_count, desc->irqs_unhandled);
+	printk("->handle_irq():  %p, ", desc->handle_irq);
+	print_symbol("%s\n", (unsigned long)desc->handle_irq);
+	printk("->irq_data.chip(): %p, ", desc->irq_data.chip);
+	print_symbol("%s\n", (unsigned long)desc->irq_data.chip);
+	printk("->action(): %p\n", desc->action);
+	if (desc->action) {
+		printk("->action->handler(): %p, ", desc->action->handler);
+		print_symbol("%s\n", (unsigned long)desc->action->handler);
+	}
+
+	P(IRQ_LEVEL);
+	P(IRQ_PER_CPU);
+	P(IRQ_NOPROBE);
+	P(IRQ_NOREQUEST);
+	P(IRQ_NOAUTOEN);
+
+	PS(IRQS_AUTODETECT);
+	PS(IRQS_INPROGRESS);
+	PS(IRQS_REPLAY);
+	PS(IRQS_WAITING);
+	PS(IRQS_DISABLED);
+	PS(IRQS_PENDING);
+	PS(IRQS_MASKED);
+}
+
+#undef P
+#undef PS

kernel/irq/handle.c

@@ -51,30 +51,92 @@ static void warn_no_thread(unsigned int irq, struct irqaction *action)
 	       "but no thread function available.", irq, action->name);
 }
 
-/**
- * handle_IRQ_event - irq action chain handler
- * @irq:	the interrupt number
- * @action:	the interrupt action chain for this irq
- *
- * Handles the action chain of an irq event
- */
-irqreturn_t handle_IRQ_event(unsigned int irq, struct irqaction *action)
+static void irq_wake_thread(struct irq_desc *desc, struct irqaction *action)
 {
-	irqreturn_t ret, retval = IRQ_NONE;
-	unsigned int status = 0;
+	/*
+	 * Wake up the handler thread for this action. In case the
+	 * thread crashed and was killed we just pretend that we
+	 * handled the interrupt. The hardirq handler has disabled the
+	 * device interrupt, so no irq storm is lurking. If the
+	 * RUNTHREAD bit is already set, nothing to do.
+	 */
+	if (test_bit(IRQTF_DIED, &action->thread_flags) ||
+	    test_and_set_bit(IRQTF_RUNTHREAD, &action->thread_flags))
+		return;
+
+	/*
+	 * It's safe to OR the mask lockless here. We have only two
+	 * places which write to threads_oneshot: This code and the
+	 * irq thread.
+	 *
+	 * This code is the hard irq context and can never run on two
+	 * cpus in parallel. If it ever does we have more serious
+	 * problems than this bitmask.
+	 *
+	 * The irq threads of this irq which clear their "running" bit
+	 * in threads_oneshot are serialized via desc->lock against
+	 * each other and they are serialized against this code by
+	 * IRQS_INPROGRESS.
+	 *
+	 * Hard irq handler:
+	 *
+	 *	spin_lock(desc->lock);
+	 *	desc->state |= IRQS_INPROGRESS;
+	 *	spin_unlock(desc->lock);
+	 *	set_bit(IRQTF_RUNTHREAD, &action->thread_flags);
+	 *	desc->threads_oneshot |= mask;
+	 *	spin_lock(desc->lock);
+	 *	desc->state &= ~IRQS_INPROGRESS;
+	 *	spin_unlock(desc->lock);
+	 *
+	 * irq thread:
+	 *
+	 * again:
+	 *	spin_lock(desc->lock);
+	 *	if (desc->state & IRQS_INPROGRESS) {
+	 *		spin_unlock(desc->lock);
+	 *		while(desc->state & IRQS_INPROGRESS)
+	 *			cpu_relax();
+	 *		goto again;
+	 *	}
+	 *	if (!test_bit(IRQTF_RUNTHREAD, &action->thread_flags))
+	 *		desc->threads_oneshot &= ~mask;
+	 *	spin_unlock(desc->lock);
+	 *
+	 * So either the thread waits for us to clear IRQS_INPROGRESS
+	 * or we are waiting in the flow handler for desc->lock to be
+	 * released before we reach this point. The thread also checks
+	 * IRQTF_RUNTHREAD under desc->lock. If set it leaves
+	 * threads_oneshot untouched and runs the thread another time.
+	 */
+	desc->threads_oneshot |= action->thread_mask;
+	wake_up_process(action->thread);
+}
+
+irqreturn_t
+handle_irq_event_percpu(struct irq_desc *desc, struct irqaction *action)
+{
+	irqreturn_t retval = IRQ_NONE;
+	unsigned int random = 0, irq = desc->irq_data.irq;
 
 	do {
-		trace_irq_handler_entry(irq, action);
-		ret = action->handler(irq, action->dev_id);
-		trace_irq_handler_exit(irq, action, ret);
+		irqreturn_t res;
 
-		switch (ret) {
+		trace_irq_handler_entry(irq, action);
+		res = action->handler(irq, action->dev_id);
+		trace_irq_handler_exit(irq, action, res);
+
+		if (WARN_ONCE(!irqs_disabled(),"irq %u handler %pF enabled interrupts\n",
+			      irq, action->handler))
+			local_irq_disable();
+
+		switch (res) {
 		case IRQ_WAKE_THREAD:
 			/*
 			 * Set result to handled so the spurious check
 			 * does not trigger.
 			 */
-			ret = IRQ_HANDLED;
+			res = IRQ_HANDLED;
 
 			/*
 			 * Catch drivers which return WAKE_THREAD but
@@ -85,36 +147,56 @@ irqreturn_t handle_IRQ_event(unsigned int irq, struct irqaction *action)
 				break;
 			}
 
-			/*
-			 * Wake up the handler thread for this
-			 * action. In case the thread crashed and was
-			 * killed we just pretend that we handled the
-			 * interrupt. The hardirq handler above has
-			 * disabled the device interrupt, so no irq
-			 * storm is lurking.
-			 */
-			if (likely(!test_bit(IRQTF_DIED,
-					     &action->thread_flags))) {
-				set_bit(IRQTF_RUNTHREAD, &action->thread_flags);
-				wake_up_process(action->thread);
-			}
+			irq_wake_thread(desc, action);
 
 			/* Fall through to add to randomness */
 		case IRQ_HANDLED:
-			status |= action->flags;
+			random |= action->flags;
 			break;
 
 		default:
 			break;
 		}
 
-		retval |= ret;
+		retval |= res;
 		action = action->next;
 	} while (action);
 
-	if (status & IRQF_SAMPLE_RANDOM)
+	if (random & IRQF_SAMPLE_RANDOM)
 		add_interrupt_randomness(irq);
-	local_irq_disable();
 
+	if (!noirqdebug)
+		note_interrupt(irq, desc, retval);
 	return retval;
 }
+
+irqreturn_t handle_irq_event(struct irq_desc *desc)
+{
+	struct irqaction *action = desc->action;
+	irqreturn_t ret;
+
+	irq_compat_clr_pending(desc);
+	desc->istate &= ~IRQS_PENDING;
+	irq_compat_set_progress(desc);
+	desc->istate |= IRQS_INPROGRESS;
+	raw_spin_unlock(&desc->lock);
+
+	ret = handle_irq_event_percpu(desc, action);
+
+	raw_spin_lock(&desc->lock);
+	desc->istate &= ~IRQS_INPROGRESS;
+	irq_compat_clr_progress(desc);
+	return ret;
+}
+
+/**
+ * handle_IRQ_event - irq action chain handler
+ * @irq:	the interrupt number
+ * @action:	the interrupt action chain for this irq
+ *
+ * Handles the action chain of an irq event
+ */
+irqreturn_t handle_IRQ_event(unsigned int irq, struct irqaction *action)
+{
+	return handle_irq_event_percpu(irq_to_desc(irq), action);
+}

kernel/irq/internals.h

@@ -1,27 +1,101 @@
 /*
  * IRQ subsystem internal functions and variables:
+ *
+ * Do not ever include this file from anything else than
+ * kernel/irq/. Do not even think about using any information outside
+ * of this file for your non core code.
  */
 #include <linux/irqdesc.h>
 
+#ifdef CONFIG_SPARSE_IRQ
+# define IRQ_BITMAP_BITS	(NR_IRQS + 8196)
+#else
+# define IRQ_BITMAP_BITS	NR_IRQS
+#endif
+
+#define istate core_internal_state__do_not_mess_with_it
+
+#ifdef CONFIG_GENERIC_HARDIRQS_NO_COMPAT
+# define status status_use_accessors
+#endif
+
 extern int noirqdebug;
 
+/*
+ * Bits used by threaded handlers:
+ * IRQTF_RUNTHREAD - signals that the interrupt handler thread should run
+ * IRQTF_DIED      - handler thread died
+ * IRQTF_WARNED    - warning "IRQ_WAKE_THREAD w/o thread_fn" has been printed
+ * IRQTF_AFFINITY  - irq thread is requested to adjust affinity
+ * IRQTF_FORCED_THREAD  - irq action is force threaded
+ */
+enum {
+	IRQTF_RUNTHREAD,
+	IRQTF_DIED,
+	IRQTF_WARNED,
+	IRQTF_AFFINITY,
+	IRQTF_FORCED_THREAD,
+};
+
+/*
+ * Bit masks for desc->state
+ *
+ * IRQS_AUTODETECT		- autodetection in progress
+ * IRQS_SPURIOUS_DISABLED	- was disabled due to spurious interrupt
+ *				  detection
+ * IRQS_POLL_INPROGRESS		- polling in progress
+ * IRQS_INPROGRESS		- Interrupt in progress
+ * IRQS_ONESHOT			- irq is not unmasked in primary handler
+ * IRQS_REPLAY			- irq is replayed
+ * IRQS_WAITING			- irq is waiting
+ * IRQS_DISABLED		- irq is disabled
+ * IRQS_PENDING			- irq is pending and replayed later
+ * IRQS_MASKED			- irq is masked
+ * IRQS_SUSPENDED		- irq is suspended
+ */
+enum {
+	IRQS_AUTODETECT		= 0x00000001,
+	IRQS_SPURIOUS_DISABLED	= 0x00000002,
+	IRQS_POLL_INPROGRESS	= 0x00000008,
+	IRQS_INPROGRESS		= 0x00000010,
+	IRQS_ONESHOT		= 0x00000020,
+	IRQS_REPLAY		= 0x00000040,
+	IRQS_WAITING		= 0x00000080,
+	IRQS_DISABLED		= 0x00000100,
+	IRQS_PENDING		= 0x00000200,
+	IRQS_MASKED		= 0x00000400,
+	IRQS_SUSPENDED		= 0x00000800,
+};
+
+#include "compat.h"
+#include "debug.h"
+#include "settings.h"
+
 #define irq_data_to_desc(data)	container_of(data, struct irq_desc, irq_data)
 
 /* Set default functions for irq_chip structures: */
 extern void irq_chip_set_defaults(struct irq_chip *chip);
 
-/* Set default handler: */
-extern void compat_irq_chip_set_default_handler(struct irq_desc *desc);
-
 extern int __irq_set_trigger(struct irq_desc *desc, unsigned int irq,
 		unsigned long flags);
 extern void __disable_irq(struct irq_desc *desc, unsigned int irq, bool susp);
 extern void __enable_irq(struct irq_desc *desc, unsigned int irq, bool resume);
 
+extern int irq_startup(struct irq_desc *desc);
+extern void irq_shutdown(struct irq_desc *desc);
+extern void irq_enable(struct irq_desc *desc);
+extern void irq_disable(struct irq_desc *desc);
+extern void mask_irq(struct irq_desc *desc);
+extern void unmask_irq(struct irq_desc *desc);
+
 extern void init_kstat_irqs(struct irq_desc *desc, int node, int nr);
 
+irqreturn_t handle_irq_event_percpu(struct irq_desc *desc, struct irqaction *action);
+irqreturn_t handle_irq_event(struct irq_desc *desc);
+
 /* Resending of interrupts :*/
 void check_irq_resend(struct irq_desc *desc, unsigned int irq);
+bool irq_wait_for_poll(struct irq_desc *desc);
 
 #ifdef CONFIG_PROC_FS
 extern void register_irq_proc(unsigned int irq, struct irq_desc *desc);
@@ -37,20 +111,10 @@ static inline void unregister_handler_proc(unsigned int irq,
 					   struct irqaction *action) { }
 #endif
 
-extern int irq_select_affinity_usr(unsigned int irq);
+extern int irq_select_affinity_usr(unsigned int irq, struct cpumask *mask);
 
 extern void irq_set_thread_affinity(struct irq_desc *desc);
 
-#ifndef CONFIG_GENERIC_HARDIRQS_NO_DEPRECATED
-static inline void irq_end(unsigned int irq, struct irq_desc *desc)
-{
-	if (desc->irq_data.chip && desc->irq_data.chip->end)
-		desc->irq_data.chip->end(irq);
-}
-#else
-static inline void irq_end(unsigned int irq, struct irq_desc *desc) { }
-#endif
-
 /* Inline functions for support of irq chips on slow busses */
 static inline void chip_bus_lock(struct irq_desc *desc)
 {
@@ -64,43 +128,60 @@ static inline void chip_bus_sync_unlock(struct irq_desc *desc)
 		desc->irq_data.chip->irq_bus_sync_unlock(&desc->irq_data);
 }
 
-/*
- * Debugging printout:
- */
+struct irq_desc *
+__irq_get_desc_lock(unsigned int irq, unsigned long *flags, bool bus);
+void __irq_put_desc_unlock(struct irq_desc *desc, unsigned long flags, bool bus);
 
-#include <linux/kallsyms.h>
-
-#define P(f) if (desc->status & f) printk("%14s set\n", #f)
-
-static inline void print_irq_desc(unsigned int irq, struct irq_desc *desc)
+static inline struct irq_desc *
+irq_get_desc_buslock(unsigned int irq, unsigned long *flags)
 {
-	printk("irq %d, desc: %p, depth: %d, count: %d, unhandled: %d\n",
-		irq, desc, desc->depth, desc->irq_count, desc->irqs_unhandled);
-	printk("->handle_irq():  %p, ", desc->handle_irq);
-	print_symbol("%s\n", (unsigned long)desc->handle_irq);
-	printk("->irq_data.chip(): %p, ", desc->irq_data.chip);
-	print_symbol("%s\n", (unsigned long)desc->irq_data.chip);
-	printk("->action(): %p\n", desc->action);
-	if (desc->action) {
-		printk("->action->handler(): %p, ", desc->action->handler);
-		print_symbol("%s\n", (unsigned long)desc->action->handler);
-	}
-
-	P(IRQ_INPROGRESS);
-	P(IRQ_DISABLED);
-	P(IRQ_PENDING);
-	P(IRQ_REPLAY);
-	P(IRQ_AUTODETECT);
-	P(IRQ_WAITING);
-	P(IRQ_LEVEL);
-	P(IRQ_MASKED);
-#ifdef CONFIG_IRQ_PER_CPU
-	P(IRQ_PER_CPU);
-#endif
-	P(IRQ_NOPROBE);
-	P(IRQ_NOREQUEST);
-	P(IRQ_NOAUTOEN);
+	return __irq_get_desc_lock(irq, flags, true);
 }
 
-#undef P
+static inline void
+irq_put_desc_busunlock(struct irq_desc *desc, unsigned long flags)
+{
+	__irq_put_desc_unlock(desc, flags, true);
+}
 
+static inline struct irq_desc *
+irq_get_desc_lock(unsigned int irq, unsigned long *flags)
+{
+	return __irq_get_desc_lock(irq, flags, false);
+}
+
+static inline void
+irq_put_desc_unlock(struct irq_desc *desc, unsigned long flags)
+{
+	__irq_put_desc_unlock(desc, flags, false);
+}
+
+/*
+ * Manipulation functions for irq_data.state
+ */
+static inline void irqd_set_move_pending(struct irq_data *d)
+{
+	d->state_use_accessors |= IRQD_SETAFFINITY_PENDING;
+	irq_compat_set_move_pending(irq_data_to_desc(d));
+}
+
+static inline void irqd_clr_move_pending(struct irq_data *d)
+{
+	d->state_use_accessors &= ~IRQD_SETAFFINITY_PENDING;
+	irq_compat_clr_move_pending(irq_data_to_desc(d));
+}
+
+static inline void irqd_clear(struct irq_data *d, unsigned int mask)
+{
+	d->state_use_accessors &= ~mask;
+}
+
+static inline void irqd_set(struct irq_data *d, unsigned int mask)
+{
+	d->state_use_accessors |= mask;
+}
+
+static inline bool irqd_has_set(struct irq_data *d, unsigned int mask)
+{
+	return d->state_use_accessors & mask;
+}

kernel/irq/irqdesc.c

@@ -79,7 +79,8 @@ static void desc_set_defaults(unsigned int irq, struct irq_desc *desc, int node)
 	desc->irq_data.chip_data = NULL;
 	desc->irq_data.handler_data = NULL;
 	desc->irq_data.msi_desc = NULL;
-	desc->status = IRQ_DEFAULT_INIT_FLAGS;
+	irq_settings_clr_and_set(desc, ~0, _IRQ_DEFAULT_INIT_FLAGS);
+	desc->istate = IRQS_DISABLED;
 	desc->handle_irq = handle_bad_irq;
 	desc->depth = 1;
 	desc->irq_count = 0;
@@ -94,7 +95,7 @@ int nr_irqs = NR_IRQS;
 EXPORT_SYMBOL_GPL(nr_irqs);
 
 static DEFINE_MUTEX(sparse_irq_lock);
-static DECLARE_BITMAP(allocated_irqs, NR_IRQS);
+static DECLARE_BITMAP(allocated_irqs, IRQ_BITMAP_BITS);
 
 #ifdef CONFIG_SPARSE_IRQ
 
@@ -206,6 +207,14 @@ struct irq_desc * __ref irq_to_desc_alloc_node(unsigned int irq, int node)
 	return NULL;
 }
 
+static int irq_expand_nr_irqs(unsigned int nr)
+{
+	if (nr > IRQ_BITMAP_BITS)
+		return -ENOMEM;
+	nr_irqs = nr;
+	return 0;
+}
+
 int __init early_irq_init(void)
 {
 	int i, initcnt, node = first_online_node;
@@ -217,6 +226,15 @@ int __init early_irq_init(void)
 	initcnt = arch_probe_nr_irqs();
 	printk(KERN_INFO "NR_IRQS:%d nr_irqs:%d %d\n", NR_IRQS, nr_irqs, initcnt);
 
+	if (WARN_ON(nr_irqs > IRQ_BITMAP_BITS))
+		nr_irqs = IRQ_BITMAP_BITS;
+
+	if (WARN_ON(initcnt > IRQ_BITMAP_BITS))
+		initcnt = IRQ_BITMAP_BITS;
+
+	if (initcnt > nr_irqs)
+		nr_irqs = initcnt;
+
 	for (i = 0; i < initcnt; i++) {
 		desc = alloc_desc(i, node);
 		set_bit(i, allocated_irqs);
@@ -229,7 +247,7 @@ int __init early_irq_init(void)
 
 struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = {
 	[0 ... NR_IRQS-1] = {
-		.status		= IRQ_DEFAULT_INIT_FLAGS,
+		.istate		= IRQS_DISABLED,
 		.handle_irq	= handle_bad_irq,
 		.depth		= 1,
 		.lock		= __RAW_SPIN_LOCK_UNLOCKED(irq_desc->lock),
@@ -251,8 +269,8 @@ int __init early_irq_init(void)
 	for (i = 0; i < count; i++) {
 		desc[i].irq_data.irq = i;
 		desc[i].irq_data.chip = &no_irq_chip;
-		/* TODO : do this allocation on-demand ... */
 		desc[i].kstat_irqs = alloc_percpu(unsigned int);
+		irq_settings_clr_and_set(desc, ~0, _IRQ_DEFAULT_INIT_FLAGS);
 		alloc_masks(desc + i, GFP_KERNEL, node);
 		desc_smp_init(desc + i, node);
 		lockdep_set_class(&desc[i].lock, &irq_desc_lock_class);
@@ -277,24 +295,14 @@ static void free_desc(unsigned int irq)
 
 static inline int alloc_descs(unsigned int start, unsigned int cnt, int node)
 {
-#if defined(CONFIG_KSTAT_IRQS_ONDEMAND)
-	struct irq_desc *desc;
-	unsigned int i;
-
-	for (i = 0; i < cnt; i++) {
-		desc = irq_to_desc(start + i);
-		if (desc && !desc->kstat_irqs) {
-			unsigned int __percpu *stats = alloc_percpu(unsigned int);
-
-			if (!stats)
-				return -1;
-			if (cmpxchg(&desc->kstat_irqs, NULL, stats) != NULL)
-				free_percpu(stats);
-		}
-	}
-#endif
 	return start;
 }
+
+static int irq_expand_nr_irqs(unsigned int nr)
+{
+	return -ENOMEM;
+}
+
 #endif /* !CONFIG_SPARSE_IRQ */
 
 /* Dynamic interrupt handling */
@@ -338,14 +346,17 @@ irq_alloc_descs(int irq, unsigned int from, unsigned int cnt, int node)
 
 	mutex_lock(&sparse_irq_lock);
 
-	start = bitmap_find_next_zero_area(allocated_irqs, nr_irqs, from, cnt, 0);
+	start = bitmap_find_next_zero_area(allocated_irqs, IRQ_BITMAP_BITS,
+					   from, cnt, 0);
 	ret = -EEXIST;
 	if (irq >=0 && start != irq)
 		goto err;
 
-	ret = -ENOMEM;
-	if (start >= nr_irqs)
-		goto err;
+	if (start + cnt > nr_irqs) {
+		ret = irq_expand_nr_irqs(start + cnt);
+		if (ret)
+			goto err;
+	}
 
 	bitmap_set(allocated_irqs, start, cnt);
 	mutex_unlock(&sparse_irq_lock);
@@ -392,6 +403,26 @@ unsigned int irq_get_next_irq(unsigned int offset)
 	return find_next_bit(allocated_irqs, nr_irqs, offset);
 }
 
+struct irq_desc *
+__irq_get_desc_lock(unsigned int irq, unsigned long *flags, bool bus)
+{
+	struct irq_desc *desc = irq_to_desc(irq);
+
+	if (desc) {
+		if (bus)
+			chip_bus_lock(desc);
+		raw_spin_lock_irqsave(&desc->lock, *flags);
+	}
+	return desc;
+}
+
+void __irq_put_desc_unlock(struct irq_desc *desc, unsigned long flags, bool bus)
+{
+	raw_spin_unlock_irqrestore(&desc->lock, flags);
+	if (bus)
+		chip_bus_sync_unlock(desc);
+}
+
 /**
  * dynamic_irq_cleanup - cleanup a dynamically allocated irq
  * @irq:	irq number to initialize

kernel/irq/migration.c

@@ -4,23 +4,23 @@
 
 #include "internals.h"
 
-void move_masked_irq(int irq)
+void irq_move_masked_irq(struct irq_data *idata)
 {
-	struct irq_desc *desc = irq_to_desc(irq);
-	struct irq_chip *chip = desc->irq_data.chip;
+	struct irq_desc *desc = irq_data_to_desc(idata);
+	struct irq_chip *chip = idata->chip;
 
-	if (likely(!(desc->status & IRQ_MOVE_PENDING)))
+	if (likely(!irqd_is_setaffinity_pending(&desc->irq_data)))
 		return;
 
 	/*
 	 * Paranoia: cpu-local interrupts shouldn't be calling in here anyway.
 	 */
-	if (CHECK_IRQ_PER_CPU(desc->status)) {
+	if (!irqd_can_balance(&desc->irq_data)) {
 		WARN_ON(1);
 		return;
 	}
 
-	desc->status &= ~IRQ_MOVE_PENDING;
+	irqd_clr_move_pending(&desc->irq_data);
 
 	if (unlikely(cpumask_empty(desc->pending_mask)))
 		return;
@@ -53,15 +53,20 @@ void move_masked_irq(int irq)
 	cpumask_clear(desc->pending_mask);
 }
 
-void move_native_irq(int irq)
+void move_masked_irq(int irq)
 {
-	struct irq_desc *desc = irq_to_desc(irq);
+	irq_move_masked_irq(irq_get_irq_data(irq));
+}
+
+void irq_move_irq(struct irq_data *idata)
+{
+	struct irq_desc *desc = irq_data_to_desc(idata);
 	bool masked;
 
-	if (likely(!(desc->status & IRQ_MOVE_PENDING)))
+	if (likely(!irqd_is_setaffinity_pending(idata)))
 		return;
 
-	if (unlikely(desc->status & IRQ_DISABLED))
+	if (unlikely(desc->istate & IRQS_DISABLED))
 		return;
 
 	/*
@@ -69,10 +74,15 @@ void move_native_irq(int irq)
 	 * threaded interrupt with ONESHOT set, we can end up with an
 	 * interrupt storm.
 	 */
-	masked = desc->status & IRQ_MASKED;
+	masked = desc->istate & IRQS_MASKED;
 	if (!masked)
-		desc->irq_data.chip->irq_mask(&desc->irq_data);
-	move_masked_irq(irq);
+		idata->chip->irq_mask(idata);
+	irq_move_masked_irq(idata);
 	if (!masked)
-		desc->irq_data.chip->irq_unmask(&desc->irq_data);
+		idata->chip->irq_unmask(idata);
+}
+
+void move_native_irq(int irq)
+{
+	irq_move_irq(irq_get_irq_data(irq));
 }

kernel/irq/pm.c

@@ -18,7 +18,7 @@
  * During system-wide suspend or hibernation device drivers need to be prevented
  * from receiving interrupts and this function is provided for this purpose.
  * It marks all interrupt lines in use, except for the timer ones, as disabled
- * and sets the IRQ_SUSPENDED flag for each of them.
+ * and sets the IRQS_SUSPENDED flag for each of them.
  */
 void suspend_device_irqs(void)
 {
@@ -34,7 +34,7 @@ void suspend_device_irqs(void)
 	}
 
 	for_each_irq_desc(irq, desc)
-		if (desc->status & IRQ_SUSPENDED)
+		if (desc->istate & IRQS_SUSPENDED)
 			synchronize_irq(irq);
 }
 EXPORT_SYMBOL_GPL(suspend_device_irqs);
@@ -43,7 +43,7 @@ EXPORT_SYMBOL_GPL(suspend_device_irqs);
  * resume_device_irqs - enable interrupt lines disabled by suspend_device_irqs()
  *
  * Enable all interrupt lines previously disabled by suspend_device_irqs() that
- * have the IRQ_SUSPENDED flag set.
+ * have the IRQS_SUSPENDED flag set.
  */
 void resume_device_irqs(void)
 {
@@ -53,9 +53,6 @@ void resume_device_irqs(void)
 	for_each_irq_desc(irq, desc) {
 		unsigned long flags;
 
-		if (!(desc->status & IRQ_SUSPENDED))
-			continue;
-
 		raw_spin_lock_irqsave(&desc->lock, flags);
 		__enable_irq(desc, irq, true);
 		raw_spin_unlock_irqrestore(&desc->lock, flags);
@@ -71,9 +68,24 @@ int check_wakeup_irqs(void)
 	struct irq_desc *desc;
 	int irq;
 
-	for_each_irq_desc(irq, desc)
-		if ((desc->status & IRQ_WAKEUP) && (desc->status & IRQ_PENDING))
-			return -EBUSY;
+	for_each_irq_desc(irq, desc) {
+		if (irqd_is_wakeup_set(&desc->irq_data)) {
+			if (desc->istate & IRQS_PENDING)
+				return -EBUSY;
+			continue;
+		}
+		/*
+		 * Check the non wakeup interrupts whether they need
+		 * to be masked before finally going into suspend
+		 * state. That's for hardware which has no wakeup
+		 * source configuration facility. The chip
+		 * implementation indicates that with
+		 * IRQCHIP_MASK_ON_SUSPEND.
+		 */
+		if (desc->istate & IRQS_SUSPENDED &&
+		    irq_desc_get_chip(desc)->flags & IRQCHIP_MASK_ON_SUSPEND)
+			mask_irq(desc);
+	}
 
 	return 0;
 }

kernel/irq/proc.c

@@ -11,6 +11,7 @@
 #include <linux/proc_fs.h>
 #include <linux/seq_file.h>
 #include <linux/interrupt.h>
+#include <linux/kernel_stat.h>
 
 #include "internals.h"
 
@@ -24,7 +25,7 @@ static int irq_affinity_proc_show(struct seq_file *m, void *v)
 	const struct cpumask *mask = desc->irq_data.affinity;
 
 #ifdef CONFIG_GENERIC_PENDING_IRQ
-	if (desc->status & IRQ_MOVE_PENDING)
+	if (irqd_is_setaffinity_pending(&desc->irq_data))
 		mask = desc->pending_mask;
 #endif
 	seq_cpumask(m, mask);
@@ -65,8 +66,7 @@ static ssize_t irq_affinity_proc_write(struct file *file,
 	cpumask_var_t new_value;
 	int err;
 
-	if (!irq_to_desc(irq)->irq_data.chip->irq_set_affinity || no_irq_affinity ||
-	    irq_balancing_disabled(irq))
+	if (!irq_can_set_affinity(irq) || no_irq_affinity)
 		return -EIO;
 
 	if (!alloc_cpumask_var(&new_value, GFP_KERNEL))
@@ -89,7 +89,7 @@ static ssize_t irq_affinity_proc_write(struct file *file,
 	if (!cpumask_intersects(new_value, cpu_online_mask)) {
 		/* Special case for empty set - allow the architecture
 		   code to set default SMP affinity. */
-		err = irq_select_affinity_usr(irq) ? -EINVAL : count;
+		err = irq_select_affinity_usr(irq, new_value) ? -EINVAL : count;
 	} else {
 		irq_set_affinity(irq, new_value);
 		err = count;
@@ -357,3 +357,65 @@ void init_irq_proc(void)
 	}
 }
 
+#ifdef CONFIG_GENERIC_IRQ_SHOW
+
+int __weak arch_show_interrupts(struct seq_file *p, int prec)
+{
+	return 0;
+}
+
+int show_interrupts(struct seq_file *p, void *v)
+{
+	static int prec;
+
+	unsigned long flags, any_count = 0;
+	int i = *(loff_t *) v, j;
+	struct irqaction *action;
+	struct irq_desc *desc;
+
+	if (i > nr_irqs)
+		return 0;
+
+	if (i == nr_irqs)
+		return arch_show_interrupts(p, prec);
+
+	/* print header and calculate the width of the first column */
+	if (i == 0) {
+		for (prec = 3, j = 1000; prec < 10 && j <= nr_irqs; ++prec)
+			j *= 10;
+
+		seq_printf(p, "%*s", prec + 8, "");
+		for_each_online_cpu(j)
+			seq_printf(p, "CPU%-8d", j);
+		seq_putc(p, '\n');
+	}
+
+	desc = irq_to_desc(i);
+	if (!desc)
+		return 0;
+
+	raw_spin_lock_irqsave(&desc->lock, flags);
+	for_each_online_cpu(j)
+		any_count |= kstat_irqs_cpu(i, j);
+	action = desc->action;
+	if (!action && !any_count)
+		goto out;
+
+	seq_printf(p, "%*d: ", prec, i);
+	for_each_online_cpu(j)
+		seq_printf(p, "%10u ", kstat_irqs_cpu(i, j));
+	seq_printf(p, " %8s", desc->irq_data.chip->name);
+	seq_printf(p, "-%-8s", desc->name);
+
+	if (action) {
+		seq_printf(p, "  %s", action->name);
+		while ((action = action->next) != NULL)
+			seq_printf(p, ", %s", action->name);
+	}
+
+	seq_putc(p, '\n');
+out:
+	raw_spin_unlock_irqrestore(&desc->lock, flags);
+	return 0;
+}
+#endif

kernel/irq/resend.c

@@ -23,7 +23,7 @@
 #ifdef CONFIG_HARDIRQS_SW_RESEND
 
 /* Bitmap to handle software resend of interrupts: */
-static DECLARE_BITMAP(irqs_resend, NR_IRQS);
+static DECLARE_BITMAP(irqs_resend, IRQ_BITMAP_BITS);
 
 /*
  * Run software resends of IRQ's
@@ -55,20 +55,19 @@ static DECLARE_TASKLET(resend_tasklet, resend_irqs, 0);
  */
 void check_irq_resend(struct irq_desc *desc, unsigned int irq)
 {
-	unsigned int status = desc->status;
-
-	/*
-	 * Make sure the interrupt is enabled, before resending it:
-	 */
-	desc->irq_data.chip->irq_enable(&desc->irq_data);
-
 	/*
 	 * We do not resend level type interrupts. Level type
 	 * interrupts are resent by hardware when they are still
 	 * active.
 	 */
-	if ((status & (IRQ_LEVEL | IRQ_PENDING | IRQ_REPLAY)) == IRQ_PENDING) {
-		desc->status = (status & ~IRQ_PENDING) | IRQ_REPLAY;
+	if (irq_settings_is_level(desc))
+		return;
+	if (desc->istate & IRQS_REPLAY)
+		return;
+	if (desc->istate & IRQS_PENDING) {
+		irq_compat_clr_pending(desc);
+		desc->istate &= ~IRQS_PENDING;
+		desc->istate |= IRQS_REPLAY;
 
 		if (!desc->irq_data.chip->irq_retrigger ||
 		    !desc->irq_data.chip->irq_retrigger(&desc->irq_data)) {

kernel/irq/settings.h

@@ -0,0 +1,138 @@
+/*
+ * Internal header to deal with irq_desc->status which will be renamed
+ * to irq_desc->settings.
+ */
+enum {
+	_IRQ_DEFAULT_INIT_FLAGS	= IRQ_DEFAULT_INIT_FLAGS,
+	_IRQ_PER_CPU		= IRQ_PER_CPU,
+	_IRQ_LEVEL		= IRQ_LEVEL,
+	_IRQ_NOPROBE		= IRQ_NOPROBE,
+	_IRQ_NOREQUEST		= IRQ_NOREQUEST,
+	_IRQ_NOAUTOEN		= IRQ_NOAUTOEN,
+	_IRQ_MOVE_PCNTXT	= IRQ_MOVE_PCNTXT,
+	_IRQ_NO_BALANCING	= IRQ_NO_BALANCING,
+	_IRQ_NESTED_THREAD	= IRQ_NESTED_THREAD,
+	_IRQF_MODIFY_MASK	= IRQF_MODIFY_MASK,
+};
+
+#define IRQ_INPROGRESS		GOT_YOU_MORON
+#define IRQ_REPLAY		GOT_YOU_MORON
+#define IRQ_WAITING		GOT_YOU_MORON
+#define IRQ_DISABLED		GOT_YOU_MORON
+#define IRQ_PENDING		GOT_YOU_MORON
+#define IRQ_MASKED		GOT_YOU_MORON
+#define IRQ_WAKEUP		GOT_YOU_MORON
+#define IRQ_MOVE_PENDING	GOT_YOU_MORON
+#define IRQ_PER_CPU		GOT_YOU_MORON
+#define IRQ_NO_BALANCING	GOT_YOU_MORON
+#define IRQ_AFFINITY_SET	GOT_YOU_MORON
+#define IRQ_LEVEL		GOT_YOU_MORON
+#define IRQ_NOPROBE		GOT_YOU_MORON
+#define IRQ_NOREQUEST		GOT_YOU_MORON
+#define IRQ_NOAUTOEN		GOT_YOU_MORON
+#define IRQ_NESTED_THREAD	GOT_YOU_MORON
+#undef IRQF_MODIFY_MASK
+#define IRQF_MODIFY_MASK	GOT_YOU_MORON
+
+static inline void
+irq_settings_clr_and_set(struct irq_desc *desc, u32 clr, u32 set)
+{
+	desc->status &= ~(clr & _IRQF_MODIFY_MASK);
+	desc->status |= (set & _IRQF_MODIFY_MASK);
+}
+
+static inline bool irq_settings_is_per_cpu(struct irq_desc *desc)
+{
+	return desc->status & _IRQ_PER_CPU;
+}
+
+static inline void irq_settings_set_per_cpu(struct irq_desc *desc)
+{
+	desc->status |= _IRQ_PER_CPU;
+}
+
+static inline void irq_settings_set_no_balancing(struct irq_desc *desc)
+{
+	desc->status |= _IRQ_NO_BALANCING;
+}
+
+static inline bool irq_settings_has_no_balance_set(struct irq_desc *desc)
+{
+	return desc->status & _IRQ_NO_BALANCING;
+}
+
+static inline u32 irq_settings_get_trigger_mask(struct irq_desc *desc)
+{
+	return desc->status & IRQ_TYPE_SENSE_MASK;
+}
+
+static inline void
+irq_settings_set_trigger_mask(struct irq_desc *desc, u32 mask)
+{
+	desc->status &= ~IRQ_TYPE_SENSE_MASK;
+	desc->status |= mask & IRQ_TYPE_SENSE_MASK;
+}
+
+static inline bool irq_settings_is_level(struct irq_desc *desc)
+{
+	return desc->status & _IRQ_LEVEL;
+}
+
+static inline void irq_settings_clr_level(struct irq_desc *desc)
+{
+	desc->status &= ~_IRQ_LEVEL;
+}
+
+static inline void irq_settings_set_level(struct irq_desc *desc)
+{
+	desc->status |= _IRQ_LEVEL;
+}
+
+static inline bool irq_settings_can_request(struct irq_desc *desc)
+{
+	return !(desc->status & _IRQ_NOREQUEST);
+}
+
+static inline void irq_settings_clr_norequest(struct irq_desc *desc)
+{
+	desc->status &= ~_IRQ_NOREQUEST;
+}
+
+static inline void irq_settings_set_norequest(struct irq_desc *desc)
+{
+	desc->status |= _IRQ_NOREQUEST;
+}
+
+static inline bool irq_settings_can_probe(struct irq_desc *desc)
+{
+	return !(desc->status & _IRQ_NOPROBE);
+}
+
+static inline void irq_settings_clr_noprobe(struct irq_desc *desc)
+{
+	desc->status &= ~_IRQ_NOPROBE;
+}
+
+static inline void irq_settings_set_noprobe(struct irq_desc *desc)
+{
+	desc->status |= _IRQ_NOPROBE;
+}
+
+static inline bool irq_settings_can_move_pcntxt(struct irq_desc *desc)
+{
+	return desc->status & _IRQ_MOVE_PCNTXT;
+}
+
+static inline bool irq_settings_can_autoenable(struct irq_desc *desc)
+{
+	return !(desc->status & _IRQ_NOAUTOEN);
+}
+
+static inline bool irq_settings_is_nested_thread(struct irq_desc *desc)
+{
+	return desc->status & _IRQ_NESTED_THREAD;
+}
+
+/* Nothing should touch desc->status from now on */
+#undef status
+#define status		USE_THE_PROPER_WRAPPERS_YOU_MORON

kernel/irq/spurious.c

@@ -21,70 +21,94 @@ static int irqfixup __read_mostly;
 #define POLL_SPURIOUS_IRQ_INTERVAL (HZ/10)
 static void poll_spurious_irqs(unsigned long dummy);
 static DEFINE_TIMER(poll_spurious_irq_timer, poll_spurious_irqs, 0, 0);
+static int irq_poll_cpu;
+static atomic_t irq_poll_active;
+
+/*
+ * We wait here for a poller to finish.
+ *
+ * If the poll runs on this CPU, then we yell loudly and return
+ * false. That will leave the interrupt line disabled in the worst
+ * case, but it should never happen.
+ *
+ * We wait until the poller is done and then recheck disabled and
+ * action (about to be disabled). Only if it's still active, we return
+ * true and let the handler run.
+ */
+bool irq_wait_for_poll(struct irq_desc *desc)
+{
+	if (WARN_ONCE(irq_poll_cpu == smp_processor_id(),
+		      "irq poll in progress on cpu %d for irq %d\n",
+		      smp_processor_id(), desc->irq_data.irq))
+		return false;
+
+#ifdef CONFIG_SMP
+	do {
+		raw_spin_unlock(&desc->lock);
+		while (desc->istate & IRQS_INPROGRESS)
+			cpu_relax();
+		raw_spin_lock(&desc->lock);
+	} while (desc->istate & IRQS_INPROGRESS);
+	/* Might have been disabled in meantime */
+	return !(desc->istate & IRQS_DISABLED) && desc->action;
+#else
+	return false;
+#endif
+}
+
 
 /*
  * Recovery handler for misrouted interrupts.
  */
-static int try_one_irq(int irq, struct irq_desc *desc)
+static int try_one_irq(int irq, struct irq_desc *desc, bool force)
 {
+	irqreturn_t ret = IRQ_NONE;
 	struct irqaction *action;
-	int ok = 0, work = 0;
 
 	raw_spin_lock(&desc->lock);
+
+	/* PER_CPU and nested thread interrupts are never polled */
+	if (irq_settings_is_per_cpu(desc) || irq_settings_is_nested_thread(desc))
+		goto out;
+
+	/*
+	 * Do not poll disabled interrupts unless the spurious
+	 * disabled poller asks explicitely.
+	 */
+	if ((desc->istate & IRQS_DISABLED) && !force)
+		goto out;
+
+	/*
+	 * All handlers must agree on IRQF_SHARED, so we test just the
+	 * first. Check for action->next as well.
+	 */
+	action = desc->action;
+	if (!action || !(action->flags & IRQF_SHARED) ||
+	    (action->flags & __IRQF_TIMER) || !action->next)
+		goto out;
+
 	/* Already running on another processor */
-	if (desc->status & IRQ_INPROGRESS) {
+	if (desc->istate & IRQS_INPROGRESS) {
 		/*
 		 * Already running: If it is shared get the other
 		 * CPU to go looking for our mystery interrupt too
 		 */
-		if (desc->action && (desc->action->flags & IRQF_SHARED))
-			desc->status |= IRQ_PENDING;
-		raw_spin_unlock(&desc->lock);
-		return ok;
+		irq_compat_set_pending(desc);
+		desc->istate |= IRQS_PENDING;
+		goto out;
 	}
-	/* Honour the normal IRQ locking */
-	desc->status |= IRQ_INPROGRESS;
-	action = desc->action;
+
+	/* Mark it poll in progress */
+	desc->istate |= IRQS_POLL_INPROGRESS;
+	do {
+		if (handle_irq_event(desc) == IRQ_HANDLED)
+			ret = IRQ_HANDLED;
+		action = desc->action;
+	} while ((desc->istate & IRQS_PENDING) && action);
+	desc->istate &= ~IRQS_POLL_INPROGRESS;
+out:
 	raw_spin_unlock(&desc->lock);
-
-	while (action) {
-		/* Only shared IRQ handlers are safe to call */
-		if (action->flags & IRQF_SHARED) {
-			if (action->handler(irq, action->dev_id) ==
-				IRQ_HANDLED)
-				ok = 1;
-		}
-		action = action->next;
-	}
-	local_irq_disable();
-	/* Now clean up the flags */
-	raw_spin_lock(&desc->lock);
-	action = desc->action;
-
-	/*
-	 * While we were looking for a fixup someone queued a real
-	 * IRQ clashing with our walk:
-	 */
-	while ((desc->status & IRQ_PENDING) && action) {
-		/*
-		 * Perform real IRQ processing for the IRQ we deferred
-		 */
-		work = 1;
-		raw_spin_unlock(&desc->lock);
-		handle_IRQ_event(irq, action);
-		raw_spin_lock(&desc->lock);
-		desc->status &= ~IRQ_PENDING;
-	}
-	desc->status &= ~IRQ_INPROGRESS;
-	/*
-	 * If we did actual work for the real IRQ line we must let the
-	 * IRQ controller clean up too
-	 */
-	if (work)
-		irq_end(irq, desc);
-	raw_spin_unlock(&desc->lock);
-
-	return ok;
+	return ret == IRQ_HANDLED;
 }
 
 static int misrouted_irq(int irq)
@@ -92,6 +116,11 @@ static int misrouted_irq(int irq)
 	struct irq_desc *desc;
 	int i, ok = 0;
 
+	if (atomic_inc_return(&irq_poll_active) == 1)
+		goto out;
+
+	irq_poll_cpu = smp_processor_id();
+
 	for_each_irq_desc(i, desc) {
 		if (!i)
 			 continue;
@@ -99,9 +128,11 @@ static int misrouted_irq(int irq)
 		if (i == irq)	/* Already tried */
 			continue;
 
-		if (try_one_irq(i, desc))
+		if (try_one_irq(i, desc, false))
 			ok = 1;
 	}
+out:
+	atomic_dec(&irq_poll_active);
 	/* So the caller can adjust the irq error counts */
 	return ok;
 }
@@ -111,23 +142,28 @@ static void poll_spurious_irqs(unsigned long dummy)
 	struct irq_desc *desc;
 	int i;
 
+	if (atomic_inc_return(&irq_poll_active) != 1)
+		goto out;
+	irq_poll_cpu = smp_processor_id();
+
 	for_each_irq_desc(i, desc) {
-		unsigned int status;
+		unsigned int state;
 
 		if (!i)
 			 continue;
 
 		/* Racy but it doesn't matter */
-		status = desc->status;
+		state = desc->istate;
 		barrier();
-		if (!(status & IRQ_SPURIOUS_DISABLED))
+		if (!(state & IRQS_SPURIOUS_DISABLED))
 			continue;
 
 		local_irq_disable();
-		try_one_irq(i, desc);
+		try_one_irq(i, desc, true);
 		local_irq_enable();
 	}
-
+out:
+	atomic_dec(&irq_poll_active);
 	mod_timer(&poll_spurious_irq_timer,
 		  jiffies + POLL_SPURIOUS_IRQ_INTERVAL);
 }
@@ -139,15 +175,13 @@ static void poll_spurious_irqs(unsigned long dummy)
  *
  * (The other 100-of-100,000 interrupts may have been a correctly
  *  functioning device sharing an IRQ with the failing one)
- *
- * Called under desc->lock
  */
-
 static void
 __report_bad_irq(unsigned int irq, struct irq_desc *desc,
 		 irqreturn_t action_ret)
 {
 	struct irqaction *action;
+	unsigned long flags;
 
 	if (action_ret != IRQ_HANDLED && action_ret != IRQ_NONE) {
 		printk(KERN_ERR "irq event %d: bogus return value %x\n",
@@ -159,6 +193,13 @@ __report_bad_irq(unsigned int irq, struct irq_desc *desc,
 	dump_stack();
 	printk(KERN_ERR "handlers:\n");
 
+	/*
+	 * We need to take desc->lock here. note_interrupt() is called
+	 * w/o desc->lock held, but IRQ_PROGRESS set. We might race
+	 * with something else removing an action. It's ok to take
+	 * desc->lock here. See synchronize_irq().
+	 */
+	raw_spin_lock_irqsave(&desc->lock, flags);
 	action = desc->action;
 	while (action) {
 		printk(KERN_ERR "[<%p>]", action->handler);
@@ -167,6 +208,7 @@ __report_bad_irq(unsigned int irq, struct irq_desc *desc,
 		printk("\n");
 		action = action->next;
 	}
+	raw_spin_unlock_irqrestore(&desc->lock, flags);
 }
 
 static void
@@ -218,6 +260,9 @@ try_misrouted_irq(unsigned int irq, struct irq_desc *desc,
 void note_interrupt(unsigned int irq, struct irq_desc *desc,
 		    irqreturn_t action_ret)
 {
+	if (desc->istate & IRQS_POLL_INPROGRESS)
+		return;
+
 	if (unlikely(action_ret != IRQ_HANDLED)) {
 		/*
 		 * If we are seeing only the odd spurious IRQ caused by
@@ -254,9 +299,9 @@ void note_interrupt(unsigned int irq, struct irq_desc *desc,
 		 * Now kill the IRQ
 		 */
 		printk(KERN_EMERG "Disabling IRQ #%d\n", irq);
-		desc->status |= IRQ_DISABLED | IRQ_SPURIOUS_DISABLED;
+		desc->istate |= IRQS_SPURIOUS_DISABLED;
 		desc->depth++;
-		desc->irq_data.chip->irq_disable(&desc->irq_data);
+		irq_disable(desc);
 
 		mod_timer(&poll_spurious_irq_timer,
 			  jiffies + POLL_SPURIOUS_IRQ_INTERVAL);

kernel/sched.c

@@ -2224,7 +2224,10 @@ unsigned long wait_task_inactive(struct task_struct *p, long match_state)
 		 * yield - it could be a while.
 		 */
 		if (unlikely(on_rq)) {
-			schedule_timeout_uninterruptible(1);
+			ktime_t to = ktime_set(0, NSEC_PER_SEC/HZ);
+
+			set_current_state(TASK_UNINTERRUPTIBLE);
+			schedule_hrtimeout(&to, HRTIMER_MODE_REL);
 			continue;
 		}