irq: better lost/spurious irq handling

index 038f24d..6be86e9 100644
@@ -82,9 +82,6 @@ static void handle_prio_irq(unsigned int irq, struct irq_desc *desc)
 
 	action_ret = handle_IRQ_event(irq, action);
 
-	/* XXX: There is no direct way to access noirqdebug, so check
-	 * unconditionally for spurious irqs...
-	 * Maybe this function should go to kernel/irq/chip.c? */
 	note_interrupt(irq, desc, action_ret);
 
 	raw_spin_lock(&desc->lock);

index 10eb1a4..e638fb3 100644
@@ -233,8 +233,6 @@ static int iic_host_match(struct irq_host *h, struct device_node *node)
 				    "IBM,CBEA-Internal-Interrupt-Controller");
 }
 
-extern int noirqdebug;
-
 static void handle_iic_irq(unsigned int irq, struct irq_desc *desc)
 {
 	raw_spin_lock(&desc->lock);

@@ -267,8 +265,7 @@ static void handle_iic_irq(unsigned int irq, struct irq_desc *desc)
 		desc->status &= ~IRQ_PENDING;
 		raw_spin_unlock(&desc->lock);
 		action_ret = handle_IRQ_event(irq, action);
-		if (!noirqdebug)
-			note_interrupt(irq, desc, action_ret);
+		note_interrupt(irq, desc, action_ret);
 		raw_spin_lock(&desc->lock);
 
 	} while ((desc->status & (IRQ_PENDING | IRQ_DISABLED)) == IRQ_PENDING);

index c03243a..ec93be4 100644
@@ -324,8 +324,17 @@ static inline void generic_handle_irq(unsigned int irq)
 }
 
 /* Handling of unhandled and spurious interrupts: */
-extern void note_interrupt(unsigned int irq, struct irq_desc *desc,
-			   irqreturn_t action_ret);
+extern void __note_interrupt(unsigned int irq, struct irq_desc *desc,
+			     irqreturn_t action_ret);
+
+static inline void note_interrupt(unsigned int irq, struct irq_desc *desc,
+				  irqreturn_t action_ret)
+{
+	extern int noirqdebug;
+
+	if (!noirqdebug)
+		__note_interrupt(irq, desc, action_ret);
+}
 
 /* Resending of interrupts :*/
 void check_irq_resend(struct irq_desc *desc, unsigned int irq);

index b7091d5..db26ff0 100644
@@ -406,8 +406,7 @@ void handle_nested_irq(unsigned int irq)
 	raw_spin_unlock_irq(&desc->lock);
 
 	action_ret = action->thread_fn(action->irq, action->dev_id);
-	if (!noirqdebug)
-		note_interrupt(irq, desc, action_ret);
+	note_interrupt(irq, desc, action_ret);
 
 	raw_spin_lock_irq(&desc->lock);
 	desc->status &= ~IRQ_INPROGRESS;

@@ -450,8 +449,7 @@ handle_simple_irq(unsigned int irq, struct irq_desc *desc)
 	raw_spin_unlock(&desc->lock);
 
 	action_ret = handle_IRQ_event(irq, action);
-	if (!noirqdebug)
-		note_interrupt(irq, desc, action_ret);
+	note_interrupt(irq, desc, action_ret);
 
 	raw_spin_lock(&desc->lock);
 	desc->status &= ~IRQ_INPROGRESS;

@@ -495,8 +493,7 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc)
 	raw_spin_unlock(&desc->lock);
 
 	action_ret = handle_IRQ_event(irq, action);
-	if (!noirqdebug)
-		note_interrupt(irq, desc, action_ret);
+	note_interrupt(irq, desc, action_ret);
 
 	raw_spin_lock(&desc->lock);
 	desc->status &= ~IRQ_INPROGRESS;

@@ -548,8 +545,7 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc)
 	raw_spin_unlock(&desc->lock);
 
 	action_ret = handle_IRQ_event(irq, action);
-	if (!noirqdebug)
-		note_interrupt(irq, desc, action_ret);
+	note_interrupt(irq, desc, action_ret);
 
 	raw_spin_lock(&desc->lock);
 	desc->status &= ~IRQ_INPROGRESS;

@@ -625,8 +621,7 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc)
 		desc->status &= ~IRQ_PENDING;
 		raw_spin_unlock(&desc->lock);
 		action_ret = handle_IRQ_event(irq, action);
-		if (!noirqdebug)
-			note_interrupt(irq, desc, action_ret);
+		note_interrupt(irq, desc, action_ret);
 		raw_spin_lock(&desc->lock);
 
 	} while ((desc->status & (IRQ_PENDING | IRQ_DISABLED)) == IRQ_PENDING);

@@ -654,8 +649,7 @@ handle_percpu_irq(unsigned int irq, struct irq_desc *desc)
 		desc->chip->ack(irq);
 
 	action_ret = handle_IRQ_event(irq, desc->action);
-	if (!noirqdebug)
-		note_interrupt(irq, desc, action_ret);
+	note_interrupt(irq, desc, action_ret);
 
 	if (desc->chip->eoi)
 		desc->chip->eoi(irq);

index 27e5c69..3ae50bf 100644
@@ -461,8 +461,7 @@ unsigned int __do_IRQ(unsigned int irq)
 			desc->chip->ack(irq);
 		if (likely(!(desc->status & IRQ_DISABLED))) {
 			action_ret = handle_IRQ_event(irq, desc->action);
-			if (!noirqdebug)
-				note_interrupt(irq, desc, action_ret);
+			note_interrupt(irq, desc, action_ret);
 		}
 		desc->chip->end(irq);
 		return 1;

@@ -515,8 +514,7 @@ unsigned int __do_IRQ(unsigned int irq)
 		raw_spin_unlock(&desc->lock);
 
 		action_ret = handle_IRQ_event(irq, action);
-		if (!noirqdebug)
-			note_interrupt(irq, desc, action_ret);
+		note_interrupt(irq, desc, action_ret);
 
 		raw_spin_lock(&desc->lock);
 		if (likely(!(desc->status & IRQ_PENDING)))

index c63f3bc..341f952 100644
@@ -2,8 +2,6 @@
  * IRQ subsystem internal functions and variables:
  */
 
-extern int noirqdebug;
-
 /* Set default functions for irq_chip structures: */
 extern void irq_chip_set_defaults(struct irq_chip *chip);
 

index 89fb90a..5da60a2 100644
@@ -14,7 +14,15 @@
 #include <linux/moduleparam.h>
 #include <linux/timer.h>
 
-static int irqfixup __read_mostly;
+enum {
+	/* irqfixup levels */
+	IRQFIXUP_SPURIOUS		= 0,		/* spurious storm detection */
+	IRQFIXUP_MISROUTED		= 1,		/* misrouted IRQ fixup */
+	IRQFIXUP_POLL			= 2,		/* enable polling by default */
+};
+
+int noirqdebug __read_mostly;
+static int irqfixup __read_mostly = IRQFIXUP_SPURIOUS;
 
 #define POLL_SPURIOUS_IRQ_INTERVAL (HZ/10)
 static void poll_spurious_irqs(unsigned long dummy);

@@ -184,7 +192,7 @@ try_misrouted_irq(unsigned int irq, struct irq_desc *desc,
 {
 	struct irqaction *action;
 
-	if (!irqfixup)
+	if (irqfixup < IRQFIXUP_MISROUTED)
 		return 0;
 
 	/* We didn't actually handle the IRQ - see if it was misrouted? */

@@ -192,11 +200,11 @@ try_misrouted_irq(unsigned int irq, struct irq_desc *desc,
 		return 1;
 
 	/*
-	 * But for 'irqfixup == 2' we also do it for handled interrupts if
-	 * they are marked as IRQF_IRQPOLL (or for irq zero, which is the
-	 * traditional PC timer interrupt.. Legacy)
+	 * But for IRQFIXUP_POLL we also do it for handled interrupts
+	 * if they are marked as IRQF_IRQPOLL (or for irq zero, which
+	 * is the traditional PC timer interrupt.. Legacy)
 	 */
-	if (irqfixup < 2)
+	if (irqfixup < IRQFIXUP_POLL)
 		return 0;
 
 	if (!irq)

@@ -213,8 +221,8 @@ try_misrouted_irq(unsigned int irq, struct irq_desc *desc,
 	return action && (action->flags & IRQF_IRQPOLL);
 }
 
-void note_interrupt(unsigned int irq, struct irq_desc *desc,
-		    irqreturn_t action_ret)
+void __note_interrupt(unsigned int irq, struct irq_desc *desc,
+		      irqreturn_t action_ret)
 {
 	if (unlikely(action_ret != IRQ_HANDLED)) {
 		/*

@@ -262,8 +270,6 @@ void note_interrupt(unsigned int irq, struct irq_desc *desc,
 	desc->irqs_unhandled = 0;
 }
 
-int noirqdebug __read_mostly;
-
 int noirqdebug_setup(char *str)
 {
 	noirqdebug = 1;

@@ -278,7 +284,7 @@ MODULE_PARM_DESC(noirqdebug, "Disable irq lockup detection when true");
 
 static int __init irqfixup_setup(char *str)
 {
-	irqfixup = 1;
+	irqfixup = max(irqfixup, IRQFIXUP_MISROUTED);
 	printk(KERN_WARNING "Misrouted IRQ fixup support enabled.\n");
 	printk(KERN_WARNING "This may impact system performance.\n");
 

@@ -290,7 +296,7 @@ module_param(irqfixup, int, 0644);
 
 static int __init irqpoll_setup(char *str)
 {
-	irqfixup = 2;
+	irqfixup = IRQFIXUP_POLL;
 	printk(KERN_WARNING "Misrouted IRQ fixup and polling support "
 				"enabled\n");
 	printk(KERN_WARNING "This may significantly impact system "

index 038f24d..6be86e9 100644
@@ -82,9 +82,6 @@ static void handle_prio_irq(unsigned int irq, struct irq_desc *desc)
 
 	action_ret = handle_IRQ_event(irq, action);
 
-	/* XXX: There is no direct way to access noirqdebug, so check
-	 * unconditionally for spurious irqs...
-	 * Maybe this function should go to kernel/irq/chip.c? */
 	note_interrupt(irq, desc, action_ret);
 
 	raw_spin_lock(&desc->lock);

index 10eb1a4..e638fb3 100644
@@ -233,8 +233,6 @@ static int iic_host_match(struct irq_host *h, struct device_node *node)
 				    "IBM,CBEA-Internal-Interrupt-Controller");
 }
 
-extern int noirqdebug;
-
 static void handle_iic_irq(unsigned int irq, struct irq_desc *desc)
 {
 	raw_spin_lock(&desc->lock);

@@ -267,8 +265,7 @@ static void handle_iic_irq(unsigned int irq, struct irq_desc *desc)
 		desc->status &= ~IRQ_PENDING;
 		raw_spin_unlock(&desc->lock);
 		action_ret = handle_IRQ_event(irq, action);
-		if (!noirqdebug)
-			note_interrupt(irq, desc, action_ret);
+		note_interrupt(irq, desc, action_ret);
 		raw_spin_lock(&desc->lock);
 
 	} while ((desc->status & (IRQ_PENDING | IRQ_DISABLED)) == IRQ_PENDING);

index c03243a..ec93be4 100644
@@ -324,8 +324,17 @@ static inline void generic_handle_irq(unsigned int irq)
 }
 
 /* Handling of unhandled and spurious interrupts: */
-extern void note_interrupt(unsigned int irq, struct irq_desc *desc,
-			   irqreturn_t action_ret);
+extern void __note_interrupt(unsigned int irq, struct irq_desc *desc,
+			     irqreturn_t action_ret);
+
+static inline void note_interrupt(unsigned int irq, struct irq_desc *desc,
+				  irqreturn_t action_ret)
+{
+	extern int noirqdebug;
+
+	if (!noirqdebug)
+		__note_interrupt(irq, desc, action_ret);
+}
 
 /* Resending of interrupts :*/
 void check_irq_resend(struct irq_desc *desc, unsigned int irq);

index b7091d5..db26ff0 100644
@@ -406,8 +406,7 @@ void handle_nested_irq(unsigned int irq)
 	raw_spin_unlock_irq(&desc->lock);
 
 	action_ret = action->thread_fn(action->irq, action->dev_id);
-	if (!noirqdebug)
-		note_interrupt(irq, desc, action_ret);
+	note_interrupt(irq, desc, action_ret);
 
 	raw_spin_lock_irq(&desc->lock);
 	desc->status &= ~IRQ_INPROGRESS;

@@ -450,8 +449,7 @@ handle_simple_irq(unsigned int irq, struct irq_desc *desc)
 	raw_spin_unlock(&desc->lock);
 
 	action_ret = handle_IRQ_event(irq, action);
-	if (!noirqdebug)
-		note_interrupt(irq, desc, action_ret);
+	note_interrupt(irq, desc, action_ret);
 
 	raw_spin_lock(&desc->lock);
 	desc->status &= ~IRQ_INPROGRESS;

@@ -495,8 +493,7 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc)
 	raw_spin_unlock(&desc->lock);
 
 	action_ret = handle_IRQ_event(irq, action);
-	if (!noirqdebug)
-		note_interrupt(irq, desc, action_ret);
+	note_interrupt(irq, desc, action_ret);
 
 	raw_spin_lock(&desc->lock);
 	desc->status &= ~IRQ_INPROGRESS;

@@ -548,8 +545,7 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc)
 	raw_spin_unlock(&desc->lock);
 
 	action_ret = handle_IRQ_event(irq, action);
-	if (!noirqdebug)
-		note_interrupt(irq, desc, action_ret);
+	note_interrupt(irq, desc, action_ret);
 
 	raw_spin_lock(&desc->lock);
 	desc->status &= ~IRQ_INPROGRESS;

@@ -625,8 +621,7 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc)
 		desc->status &= ~IRQ_PENDING;
 		raw_spin_unlock(&desc->lock);
 		action_ret = handle_IRQ_event(irq, action);
-		if (!noirqdebug)
-			note_interrupt(irq, desc, action_ret);
+		note_interrupt(irq, desc, action_ret);
 		raw_spin_lock(&desc->lock);
 
 	} while ((desc->status & (IRQ_PENDING | IRQ_DISABLED)) == IRQ_PENDING);

@@ -654,8 +649,7 @@ handle_percpu_irq(unsigned int irq, struct irq_desc *desc)
 		desc->chip->ack(irq);
 
 	action_ret = handle_IRQ_event(irq, desc->action);
-	if (!noirqdebug)
-		note_interrupt(irq, desc, action_ret);
+	note_interrupt(irq, desc, action_ret);
 
 	if (desc->chip->eoi)
 		desc->chip->eoi(irq);

index 27e5c69..3ae50bf 100644
@@ -461,8 +461,7 @@ unsigned int __do_IRQ(unsigned int irq)
 			desc->chip->ack(irq);
 		if (likely(!(desc->status & IRQ_DISABLED))) {
 			action_ret = handle_IRQ_event(irq, desc->action);
-			if (!noirqdebug)
-				note_interrupt(irq, desc, action_ret);
+			note_interrupt(irq, desc, action_ret);
 		}
 		desc->chip->end(irq);
 		return 1;

@@ -515,8 +514,7 @@ unsigned int __do_IRQ(unsigned int irq)
 		raw_spin_unlock(&desc->lock);
 
 		action_ret = handle_IRQ_event(irq, action);
-		if (!noirqdebug)
-			note_interrupt(irq, desc, action_ret);
+		note_interrupt(irq, desc, action_ret);
 
 		raw_spin_lock(&desc->lock);
 		if (likely(!(desc->status & IRQ_PENDING)))

index c63f3bc..341f952 100644
@@ -2,8 +2,6 @@
  * IRQ subsystem internal functions and variables:
  */
 
-extern int noirqdebug;
-
 /* Set default functions for irq_chip structures: */
 extern void irq_chip_set_defaults(struct irq_chip *chip);
 

index 89fb90a..5da60a2 100644
@@ -14,7 +14,15 @@
 #include <linux/moduleparam.h>
 #include <linux/timer.h>
 
-static int irqfixup __read_mostly;
+enum {
+	/* irqfixup levels */
+	IRQFIXUP_SPURIOUS		= 0,		/* spurious storm detection */
+	IRQFIXUP_MISROUTED		= 1,		/* misrouted IRQ fixup */
+	IRQFIXUP_POLL			= 2,		/* enable polling by default */
+};
+
+int noirqdebug __read_mostly;
+static int irqfixup __read_mostly = IRQFIXUP_SPURIOUS;
 
 #define POLL_SPURIOUS_IRQ_INTERVAL (HZ/10)
 static void poll_spurious_irqs(unsigned long dummy);

@@ -184,7 +192,7 @@ try_misrouted_irq(unsigned int irq, struct irq_desc *desc,
 {
 	struct irqaction *action;
 
-	if (!irqfixup)
+	if (irqfixup < IRQFIXUP_MISROUTED)
 		return 0;
 
 	/* We didn't actually handle the IRQ - see if it was misrouted? */

@@ -192,11 +200,11 @@ try_misrouted_irq(unsigned int irq, struct irq_desc *desc,
 		return 1;
 
 	/*
-	 * But for 'irqfixup == 2' we also do it for handled interrupts if
-	 * they are marked as IRQF_IRQPOLL (or for irq zero, which is the
-	 * traditional PC timer interrupt.. Legacy)
+	 * But for IRQFIXUP_POLL we also do it for handled interrupts
+	 * if they are marked as IRQF_IRQPOLL (or for irq zero, which
+	 * is the traditional PC timer interrupt.. Legacy)
 	 */
-	if (irqfixup < 2)
+	if (irqfixup < IRQFIXUP_POLL)
 		return 0;
 
 	if (!irq)

@@ -213,8 +221,8 @@ try_misrouted_irq(unsigned int irq, struct irq_desc *desc,
 	return action && (action->flags & IRQF_IRQPOLL);
 }
 
-void note_interrupt(unsigned int irq, struct irq_desc *desc,
-		    irqreturn_t action_ret)
+void __note_interrupt(unsigned int irq, struct irq_desc *desc,
+		      irqreturn_t action_ret)
 {
 	if (unlikely(action_ret != IRQ_HANDLED)) {
 		/*

@@ -262,8 +270,6 @@ void note_interrupt(unsigned int irq, struct irq_desc *desc,
 	desc->irqs_unhandled = 0;
 }
 
-int noirqdebug __read_mostly;
-
 int noirqdebug_setup(char *str)
 {
 	noirqdebug = 1;

@@ -278,7 +284,7 @@ MODULE_PARM_DESC(noirqdebug, "Disable irq lockup detection when true");
 
 static int __init irqfixup_setup(char *str)
 {
-	irqfixup = 1;
+	irqfixup = max(irqfixup, IRQFIXUP_MISROUTED);
 	printk(KERN_WARNING "Misrouted IRQ fixup support enabled.\n");
 	printk(KERN_WARNING "This may impact system performance.\n");
 

@@ -290,7 +296,7 @@ module_param(irqfixup, int, 0644);
 
 static int __init irqpoll_setup(char *str)
 {
-	irqfixup = 2;
+	irqfixup = IRQFIXUP_POLL;
 	printk(KERN_WARNING "Misrouted IRQ fixup and polling support "
 				"enabled\n");
 	printk(KERN_WARNING "This may significantly impact system "

index ec93be4..50a77f9 100644
@@ -21,6 +21,7 @@
 #include <linux/irqreturn.h>
 #include <linux/irqnr.h>
 #include <linux/errno.h>
+#include <linux/timer.h>
 #include <linux/topology.h>
 #include <linux/wait.h>
 

@@ -169,6 +170,7 @@ struct irq_2_iommu;
  * @pending_mask:	pending rebalanced interrupts
  * @threads_active:	number of irqaction threads currently running
  * @wait_for_threads:	wait queue for sync_irq to wait for threaded handlers
+ * @poll_timer:		timer for IRQ polling
  * @dir:		/proc/irq/ procfs entry
  * @name:		flow handler name for /proc/interrupts output
  */

@@ -203,6 +205,9 @@ struct irq_desc {
 #endif
 	atomic_t		threads_active;
 	wait_queue_head_t       wait_for_threads;
+
+	struct timer_list	poll_timer;
+
 #ifdef CONFIG_PROC_FS
 	struct proc_dir_entry	*dir;
 #endif

index 341f952..088e5d6 100644
@@ -38,6 +38,12 @@ extern int irq_select_affinity_usr(unsigned int irq);
 
 extern void irq_set_thread_affinity(struct irq_desc *desc);
 
+extern void poll_irq(unsigned long arg);
+extern void irq_poll_action_added(struct irq_desc *desc,
+				  struct irqaction *action);
+extern void irq_poll_action_removed(struct irq_desc *desc,
+				    struct irqaction *action);
+
 /* Inline functions for support of irq chips on slow busses */
 static inline void chip_bus_lock(unsigned int irq, struct irq_desc *desc)
 {

index 3164ba7..cf9ab65 100644
@@ -752,6 +752,7 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new)
 		irq_chip_set_defaults(desc->chip);
 
 		init_waitqueue_head(&desc->wait_for_threads);
+		setup_timer(&desc->poll_timer, poll_irq, (unsigned long)desc);
 
 		/* Setup the type (level, edge polarity) if configured: */
 		if (new->flags & IRQF_TRIGGER_MASK) {

@@ -804,17 +805,10 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new)
 	desc->irq_count = 0;
 	desc->irqs_unhandled = 0;
 
-	/*
-	 * Check whether we disabled the irq via the spurious handler
-	 * before. Reenable it and give it another chance.
-	 */
-	if (shared && (desc->status & IRQ_SPURIOUS_DISABLED)) {
-		desc->status &= ~IRQ_SPURIOUS_DISABLED;
-		__enable_irq(desc, irq, false);
-	}
-
 	raw_spin_unlock_irqrestore(&desc->lock, flags);
 
+	irq_poll_action_added(desc, new);
+
 	/*
 	 * Strictly no need to wake it up, but hung_task complains
 	 * when no hard interrupt wakes the thread up.

@@ -930,6 +924,8 @@ static struct irqaction *__free_irq(unsigned int irq, void *dev_id)
 
 	raw_spin_unlock_irqrestore(&desc->lock, flags);
 
+	irq_poll_action_removed(desc, action);
+
 	unregister_handler_proc(irq, action);
 
 	/* Make sure it's not being used on another CPU: */

151 lines not on this page · raw message

index 545f730..fc18a13 100644
@@ -159,41 +159,6 @@ report_bad_irq(unsigned int irq, struct irq_desc *desc, irqreturn_t action_ret)
 	}
 }
 
-static inline int
-try_misrouted_irq(unsigned int irq, struct irq_desc *desc,
-		  irqreturn_t action_ret)
-{
-	struct irqaction *action;
-
-	if (irqfixup < IRQFIXUP_MISROUTED)
-		return 0;
-
-	/* We didn't actually handle the IRQ - see if it was misrouted? */
-	if (action_ret == IRQ_NONE)
-		return 1;
-
-	/*
-	 * But for IRQFIXUP_POLL we also do it for handled interrupts
-	 * if they are marked as IRQF_IRQPOLL (or for irq zero, which
-	 * is the traditional PC timer interrupt.. Legacy)
-	 */
-	if (irqfixup < IRQFIXUP_POLL)
-		return 0;
-
-	if (!irq)
-		return 1;
-
-	/*
-	 * Since we don't get the descriptor lock, "action" can
-	 * change under us.  We don't really care, but we don't
-	 * want to follow a NULL pointer. So tell the compiler to
-	 * just load it once by using a barrier.
-	 */
-	action = desc->action;
-	barrier();
-	return action && (action->flags & IRQF_IRQPOLL);
-}
-
 void __note_interrupt(unsigned int irq, struct irq_desc *desc,
 		      irqreturn_t action_ret)
 {

@@ -213,7 +178,8 @@ void __note_interrupt(unsigned int irq, struct irq_desc *desc,
 			report_bad_irq(irq, desc, action_ret);
 	}
 
-	if (unlikely(try_misrouted_irq(irq, desc, action_ret))) {
+	if (unlikely(irqfixup >= IRQFIXUP_MISROUTED &&
+		     action_ret == IRQ_NONE)) {
 		int ok = misrouted_irq(irq);
 		if (action_ret == IRQ_NONE)
 			desc->irqs_unhandled -= ok;

@@ -269,6 +235,9 @@ void irq_poll_action_added(struct irq_desc *desc, struct irqaction *action)
 	}
 
 	raw_spin_unlock_irqrestore(&desc->lock, flags);
+
+	if ((action->flags & IRQF_SHARED) && irqfixup >= IRQFIXUP_POLL)
+		mod_timer(&desc->poll_timer, jiffies + IRQ_POLL_INTV);
 }
 
 void irq_poll_action_removed(struct irq_desc *desc, struct irqaction *action)

index 545f730..fc18a13 100644
@@ -159,41 +159,6 @@ report_bad_irq(unsigned int irq, struct irq_desc *desc, irqreturn_t action_ret)
 	}
 }
 
-static inline int
-try_misrouted_irq(unsigned int irq, struct irq_desc *desc,
-		  irqreturn_t action_ret)
-{
-	struct irqaction *action;
-
-	if (irqfixup < IRQFIXUP_MISROUTED)
-		return 0;
-
-	/* We didn't actually handle the IRQ - see if it was misrouted? */
-	if (action_ret == IRQ_NONE)
-		return 1;
-
-	/*
-	 * But for IRQFIXUP_POLL we also do it for handled interrupts
-	 * if they are marked as IRQF_IRQPOLL (or for irq zero, which
-	 * is the traditional PC timer interrupt.. Legacy)
-	 */
-	if (irqfixup < IRQFIXUP_POLL)
-		return 0;
-
-	if (!irq)
-		return 1;
-
-	/*
-	 * Since we don't get the descriptor lock, "action" can
-	 * change under us.  We don't really care, but we don't
-	 * want to follow a NULL pointer. So tell the compiler to
-	 * just load it once by using a barrier.
-	 */
-	action = desc->action;
-	barrier();
-	return action && (action->flags & IRQF_IRQPOLL);
-}
-
 void __note_interrupt(unsigned int irq, struct irq_desc *desc,
 		      irqreturn_t action_ret)
 {

@@ -213,7 +178,8 @@ void __note_interrupt(unsigned int irq, struct irq_desc *desc,
 			report_bad_irq(irq, desc, action_ret);
 	}
 
-	if (unlikely(try_misrouted_irq(irq, desc, action_ret))) {
+	if (unlikely(irqfixup >= IRQFIXUP_MISROUTED &&
+		     action_ret == IRQ_NONE)) {
 		int ok = misrouted_irq(irq);
 		if (action_ret == IRQ_NONE)
 			desc->irqs_unhandled -= ok;

@@ -269,6 +235,9 @@ void irq_poll_action_added(struct irq_desc *desc, struct irqaction *action)
 	}
 
 	raw_spin_unlock_irqrestore(&desc->lock, flags);
+
+	if ((action->flags & IRQF_SHARED) && irqfixup >= IRQFIXUP_POLL)
+		mod_timer(&desc->poll_timer, jiffies + IRQ_POLL_INTV);
 }
 
 void irq_poll_action_removed(struct irq_desc *desc, struct irqaction *action)

index 3ef6833..ac87535 100644
@@ -139,7 +139,7 @@ aaec2000_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction aaec2000_timer_irq = {
 	.name		= "AAEC-2000 Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= aaec2000_timer_interrupt,
 };
 

index 2500f41..a4a0049 100644
@@ -87,7 +87,7 @@ static irqreturn_t at91rm9200_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction at91rm9200_timer_irq = {
 	.name		= "at91_tick",
-	.flags		= IRQF_SHARED | IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_SHARED | IRQF_DISABLED | IRQF_TIMER,
 	.handler	= at91rm9200_timer_interrupt
 };
 

index 608a632..22a290b 100644
@@ -123,7 +123,7 @@ static irqreturn_t at91sam926x_pit_interrupt(int irq, void *dev_id)
 
 static struct irqaction at91sam926x_pit_irq = {
 	.name		= "at91_tick",
-	.flags		= IRQF_SHARED | IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_SHARED | IRQF_DISABLED | IRQF_TIMER,
 	.handler	= at91sam926x_pit_interrupt
 };
 

index 72e405d..99e10ff 100644
@@ -266,7 +266,7 @@ static irqreturn_t bcmring_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction bcmring_timer_irq = {
 	.name = "bcmring Timer Tick",
-	.flags = IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags = IRQF_DISABLED | IRQF_TIMER,
 	.handler = bcmring_timer_interrupt,
 };
 

index d581ef0..9dffaa6 100644
@@ -56,7 +56,7 @@ p720t_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction clps711x_timer_irq = {
 	.name		= "CLPS711x Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= p720t_timer_interrupt,
 };
 

index 9ca4d58..9dcea9f 100644
@@ -178,7 +178,7 @@ static irqreturn_t cns3xxx_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction cns3xxx_timer_irq = {
 	.name		= "timer",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= cns3xxx_timer_interrupt,
 };
 

index c7bc7fb..efb7734 100644
@@ -195,7 +195,7 @@ ebsa110_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction ebsa110_timer_irq = {
 	.name		= "EBSA110 Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= ebsa110_timer_interrupt,
 };
 

index 9092677..2910f6c 100644
@@ -132,7 +132,7 @@ static irqreturn_t ep93xx_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction ep93xx_timer_irq = {
 	.name		= "ep93xx timer",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= ep93xx_timer_interrupt,
 };
 

index bc5e83f..42b0bd7 100644
@@ -41,7 +41,7 @@ timer1_interrupt(int irq, void *dev_id)
 static struct irqaction footbridge_timer_irq = {
 	.name		= "Timer1 timer tick",
 	.handler	= timer1_interrupt,
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 };
 
 /*

index f488fa2..ca1932e 100644
@@ -71,7 +71,7 @@ isa_timer_interrupt(int irq, void *dev_id)
 static struct irqaction isa_timer_irq = {
 	.name		= "ISA timer tick",
 	.handler	= isa_timer_interrupt,
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 };
 
 static void __init isa_timer_init(void)

index 24df2a3..be1db54 100644
@@ -37,7 +37,7 @@ h7201_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction h7201_timer_irq = {
 	.name		= "h7201 Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= h7201_timer_interrupt,
 };
 

index fd33a19..e40deea 100644
@@ -166,7 +166,7 @@ static struct irq_chip h7202_timerx_chip = {
 
 static struct irqaction h7202_timer_irq = {
 	.name		= "h7202 Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= h7202_timer_interrupt,
 };
 

index 227cf4d..60f741e 100644
@@ -447,7 +447,7 @@ static struct clock_event_device integrator_clockevent = {
 
 static struct irqaction integrator_timer_irq = {
 	.name		= "timer",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= integrator_timer_interrupt,
 	.dev_id		= &integrator_clockevent,
 };

index babb225..693275d 100644
@@ -213,7 +213,7 @@ static int ixp2000_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction ixp2000_timer_irq = {
 	.name		= "IXP2000 Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= ixp2000_timer_interrupt,
 };
 

index aa4c442..8967392 100644
@@ -359,7 +359,7 @@ ixp23xx_timer_interrupt(int irq, void *dev_id)
 static struct irqaction ixp23xx_timer_irq = {
 	.name		= "IXP23xx Timer Tick",
 	.handler	= ixp23xx_timer_interrupt,
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 };
 
 void __init ixp23xx_init_timer(void)

index 0bce097..9a574fb 100644
@@ -278,7 +278,7 @@ static irqreturn_t ixp4xx_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction ixp4xx_timer_irq = {
 	.name		= "timer1",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= ixp4xx_timer_interrupt,
 	.dev_id		= &clockevent_ixp4xx,
 };

index 4601e42..841fe8c 100644
@@ -49,7 +49,7 @@ lh7a40x_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction lh7a40x_timer_irq = {
 	.name		= "LHA740x Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= lh7a40x_timer_interrupt,
 };
 

index cf75694..3d6957a 100644
@@ -177,7 +177,7 @@ static void __init timer_config(void)
 
 static struct irqaction timer_irq = {
 	.name		= "timer",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= timer_interrupt,
 	.dev_id		= &ckevt,
 };

index 82801db..c0cc836 100644
@@ -100,7 +100,7 @@ netx_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction netx_timer_irq = {
 	.name		= "NetX Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= netx_timer_interrupt,
 };
 

index 7728126..e60627f 100644
@@ -121,7 +121,7 @@ static irqreturn_t ns9360_clockevent_handler(int irq, void *dev_id)
 
 static struct irqaction ns9360_clockevent_action = {
 	.name		= "ns9360-timer" __stringify(TIMER_CLOCKEVENT),
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= ns9360_clockevent_handler,
 };
 

index 2f90f9d..8e0dbea 100644
@@ -56,7 +56,7 @@ static irqreturn_t nuc93x_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction nuc93x_timer_irq = {
 	.name		= "nuc93x Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= nuc93x_timer_interrupt,
 };
 

index 1be6a21..c62fa79 100644
@@ -157,7 +157,7 @@ static irqreturn_t omap_mpu_timer1_interrupt(int irq, void *dev_id)
 
 static struct irqaction omap_mpu_timer1_irq = {
 	.name		= "mpu_timer1",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= omap_mpu_timer1_interrupt,
 };
 

index 20cfbcc..8ad901b 100644
@@ -156,7 +156,7 @@ static irqreturn_t omap_32k_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction omap_32k_timer_irq = {
 	.name		= "32KHz timer",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= omap_32k_timer_interrupt,
 };
 

index 74fbed8..ddf9fae 100644
@@ -62,7 +62,7 @@ static irqreturn_t omap2_gp_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction omap2_gp_timer_irq = {
 	.name		= "gp timer",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= omap2_gp_timer_interrupt,
 };
 

index 0c8aad4..1d5b2da 100644
@@ -80,7 +80,7 @@ static irqreturn_t pnx4008_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction pnx4008_timer_irq = {
 	.name = "PNX4008 Tick Timer",
-	.flags = IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags = IRQF_DISABLED | IRQF_TIMER,
 	.handler = pnx4008_timer_interrupt
 };
 

index 293e40a..9fa6e1a 100644
@@ -133,7 +133,7 @@ static struct clocksource cksrc_pxa_oscr0 = {
 
 static struct irqaction pxa_ost0_irq = {
 	.name		= "ost0",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= pxa_ost0_interrupt,
 	.dev_id		= &ckevt_pxa_osmr0,
 };

index 74b6e0e..7ec781d 100644
@@ -87,7 +87,7 @@ static struct clocksource cksrc_sa1100_oscr = {
 
 static struct irqaction sa1100_timer_irq = {
 	.name		= "ost0",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= sa1100_ost0_interrupt,
 	.dev_id		= &ckevt_sa1100_osmr0,
 };

index 358d875..d1d6ea5 100644
@@ -130,7 +130,7 @@ shark_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction shark_timer_irq = {
 	.name		= "Shark Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= shark_timer_interrupt,
 };
 

index 26d26f5..63db057 100644
@@ -326,7 +326,7 @@ static irqreturn_t u300_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction u300_timer_irq = {
 	.name           = "U300 Timer Tick",
-	.flags          = IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags          = IRQF_DISABLED | IRQF_TIMER,
 	.handler        = u300_timer_interrupt,
 };
 

index b80f769..9cda835 100644
@@ -111,7 +111,7 @@ static irqreturn_t nuc900_timer0_interrupt(int irq, void *dev_id)
 
 static struct irqaction nuc900_timer0_irq = {
 	.name		= "nuc900-timer0",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= nuc900_timer0_interrupt,
 };
 

index 6c8a02a..e320148 100644
@@ -164,7 +164,7 @@ iop_timer_interrupt(int irq, void *dev_id)
 static struct irqaction iop_timer_irq = {
 	.name		= "IOP Timer Tick",
 	.handler	= iop_timer_interrupt,
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.dev_id		= &iop_clockevent,
 };
 

index f9a1b05..e5ca2bd 100644
@@ -258,7 +258,7 @@ static irqreturn_t mxc_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction mxc_timer_irq = {
 	.name		= "i.MX Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= mxc_timer_interrupt,
 };
 

index 2231d80..133069a 100644
@@ -138,7 +138,7 @@ s3c2410_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction s3c2410_timer_irq = {
 	.name		= "S3C2410 Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= s3c2410_timer_interrupt,
 };
 

index fb0d1c2..62066b4 100644
@@ -135,7 +135,7 @@ static struct clock_event_device sp804_clockevent = {
 
 static struct irqaction sp804_timer_irq = {
 	.name		= "timer",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_DISABLED | IRQF_TIMER,
 	.handler	= sp804_timer_interrupt,
 	.dev_id		= &sp804_clockevent,
 };

index 8c9a43d..6fefad4 100644
@@ -213,8 +213,7 @@ irqreturn_t bfin_gptmr0_interrupt(int irq, void *dev_id)
 
 static struct irqaction gptmr0_irq = {
 	.name		= "Blackfin GPTimer0",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | \
-			  IRQF_IRQPOLL | IRQF_PERCPU,
+	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_PERCPU,
 	.handler	= bfin_gptmr0_interrupt,
 };
 

@@ -322,8 +321,7 @@ irqreturn_t bfin_coretmr_interrupt(int irq, void *dev_id)
 
 static struct irqaction coretmr_irq = {
 	.name		= "Blackfin CoreTimer",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | \
-			  IRQF_IRQPOLL | IRQF_PERCPU,
+	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_PERCPU,
 	.handler	= bfin_coretmr_interrupt,
 };
 

index 653b3c4..8b366c4 100644
@@ -411,7 +411,7 @@ static cycle_t itc_get_cycles(struct clocksource *cs)
 
 static struct irqaction timer_irqaction = {
 	.handler =	timer_interrupt,
-	.flags =	IRQF_DISABLED | IRQF_IRQPOLL,
+	.flags =	IRQF_DISABLED,
 	.name =		"timer"
 };
 

index efbcee5..24681d5 100644
@@ -383,7 +383,7 @@ void do_cpu_irq_mask(struct pt_regs *regs)
 static struct irqaction timer_action = {
 	.handler = timer_interrupt,
 	.name = "timer",
-	.flags = IRQF_DISABLED | IRQF_TIMER | IRQF_PERCPU | IRQF_IRQPOLL,
+	.flags = IRQF_DISABLED | IRQF_TIMER | IRQF_PERCPU,
 };
 
 #ifdef CONFIG_SMP

index fb5cc5e..476e2fe 100644
@@ -88,7 +88,7 @@ static irqreturn_t timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction irq0  = {
 	.handler = timer_interrupt,
-	.flags = IRQF_DISABLED | IRQF_NOBALANCING | IRQF_IRQPOLL | IRQF_TIMER,
+	.flags = IRQF_DISABLED | IRQF_NOBALANCING | IRQF_TIMER,
 	.name = "timer"
 };
 

index f3d3898..60b681c 100644
@@ -610,8 +610,7 @@ static int sh_cmt_setup(struct sh_cmt_priv *p, struct platform_device *pdev)
 	p->irqaction.name = dev_name(&p->pdev->dev);
 	p->irqaction.handler = sh_cmt_interrupt;
 	p->irqaction.dev_id = p;
-	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | \
-			     IRQF_IRQPOLL  | IRQF_NOBALANCING;
+	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | IRQF_NOBALANCING;
 
 	/* get hold of clock */
 	p->clk = clk_get(&p->pdev->dev, "cmt_fck");

index ef7a5be..9fe3507 100644
@@ -281,8 +281,7 @@ static int sh_mtu2_setup(struct sh_mtu2_priv *p, struct platform_device *pdev)
 	p->irqaction.handler = sh_mtu2_interrupt;
 	p->irqaction.dev_id = p;
 	p->irqaction.irq = irq;
-	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | \
-			     IRQF_IRQPOLL  | IRQF_NOBALANCING;
+	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | IRQF_NOBALANCING;
 
 	/* get hold of clock */
 	p->clk = clk_get(&p->pdev->dev, "mtu2_fck");

index de71590..75967f8 100644
@@ -387,8 +387,7 @@ static int sh_tmu_setup(struct sh_tmu_priv *p, struct platform_device *pdev)
 	p->irqaction.handler = sh_tmu_interrupt;
 	p->irqaction.dev_id = p;
 	p->irqaction.irq = irq;
-	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | \
-			     IRQF_IRQPOLL  | IRQF_NOBALANCING;
+	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | IRQF_NOBALANCING;
 
 	/* get hold of clock */
 	p->clk = clk_get(&p->pdev->dev, "tmu_fck");

index c233113..61857f1 100644
@@ -47,9 +47,6 @@
  * IRQF_TIMER - Flag to mark this interrupt as timer interrupt
  * IRQF_PERCPU - Interrupt is per cpu
  * IRQF_NOBALANCING - Flag to exclude this interrupt from irq balancing
- * IRQF_IRQPOLL - Interrupt is used for polling (only the interrupt that is
- *                registered first in an shared interrupt is considered for
- *                performance reasons)
  * IRQF_ONESHOT - Interrupt is not reenabled after the hardirq handler finished.
  *                Used by threaded interrupts which need to keep the
  *                irq line disabled until the threaded handler has been run.

@@ -61,7 +58,6 @@
 #define IRQF_TIMER		0x00000200
 #define IRQF_PERCPU		0x00000400
 #define IRQF_NOBALANCING	0x00000800
-#define IRQF_IRQPOLL		0x00001000
 #define IRQF_ONESHOT		0x00002000
 
 /*

index 61857f1..b20bd65 100644
@@ -102,16 +102,16 @@ typedef irqreturn_t (*irq_handler_t)(int, void *);
  * @thread_flags:	flags related to @thread
  */
 struct irqaction {
-	irq_handler_t handler;
-	unsigned long flags;
-	const char *name;
-	void *dev_id;
-	struct irqaction *next;
-	int irq;
-	struct proc_dir_entry *dir;
-	irq_handler_t thread_fn;
-	struct task_struct *thread;
-	unsigned long thread_flags;
+	irq_handler_t		handler;
+	unsigned long		flags;
+	const char		*name;
+	void			*dev_id;
+	struct irqaction	*next;
+	int			irq;
+	struct proc_dir_entry	*dir;
+	irq_handler_t		thread_fn;
+	struct task_struct	*thread;
+	unsigned long		thread_flags;
 };
 
 extern irqreturn_t no_action(int cpl, void *dev_id);

index fc18a13..0bce0e3 100644
@@ -28,6 +28,21 @@ enum {
 int noirqdebug __read_mostly;
 static int irqfixup __read_mostly = IRQFIXUP_SPURIOUS;
 
+static void print_irq_handlers(struct irq_desc *desc)
+{
+	struct irqaction *action;
+
+	printk(KERN_ERR "handlers:\n");
+
+	action = desc->action;
+	while (action) {
+		printk(KERN_ERR "[<%p>]", action->handler);
+		print_symbol(" (%s)", (unsigned long)action->handler);
+		printk("\n");
+		action = action->next;
+	}
+}
+
 /*
  * Recovery handler for misrouted interrupts.
  */

@@ -126,8 +141,6 @@ static void
 __report_bad_irq(unsigned int irq, struct irq_desc *desc,
 		 irqreturn_t action_ret)
 {
-	struct irqaction *action;
-
 	if (action_ret != IRQ_HANDLED && action_ret != IRQ_NONE) {
 		printk(KERN_ERR "irq event %d: bogus return value %x\n",
 				irq, action_ret);

@@ -136,16 +149,7 @@ __report_bad_irq(unsigned int irq, struct irq_desc *desc,
 				"the \"irqpoll\" option)\n", irq);
 	}
 	dump_stack();
-	printk(KERN_ERR "handlers:\n");
-
-	action = desc->action;
-	while (action) {
-		printk(KERN_ERR "[<%p>]", action->handler);
-		print_symbol(" (%s)",
-			(unsigned long)action->handler);
-		printk("\n");
-		action = action->next;
-	}
+	print_irq_handlers(desc);
 }
 
 static void

index 61857f1..b20bd65 100644
@@ -102,16 +102,16 @@ typedef irqreturn_t (*irq_handler_t)(int, void *);
  * @thread_flags:	flags related to @thread
  */
 struct irqaction {
-	irq_handler_t handler;
-	unsigned long flags;
-	const char *name;
-	void *dev_id;
-	struct irqaction *next;
-	int irq;
-	struct proc_dir_entry *dir;
-	irq_handler_t thread_fn;
-	struct task_struct *thread;
-	unsigned long thread_flags;
+	irq_handler_t		handler;
+	unsigned long		flags;
+	const char		*name;
+	void			*dev_id;
+	struct irqaction	*next;
+	int			irq;
+	struct proc_dir_entry	*dir;
+	irq_handler_t		thread_fn;
+	struct task_struct	*thread;
+	unsigned long		thread_flags;
 };
 
 extern irqreturn_t no_action(int cpl, void *dev_id);

index fc18a13..0bce0e3 100644
@@ -28,6 +28,21 @@ enum {
 int noirqdebug __read_mostly;
 static int irqfixup __read_mostly = IRQFIXUP_SPURIOUS;
 
+static void print_irq_handlers(struct irq_desc *desc)
+{
+	struct irqaction *action;
+
+	printk(KERN_ERR "handlers:\n");
+
+	action = desc->action;
+	while (action) {
+		printk(KERN_ERR "[<%p>]", action->handler);
+		print_symbol(" (%s)", (unsigned long)action->handler);
+		printk("\n");
+		action = action->next;
+	}
+}
+
 /*
  * Recovery handler for misrouted interrupts.
  */

@@ -126,8 +141,6 @@ static void
 __report_bad_irq(unsigned int irq, struct irq_desc *desc,
 		 irqreturn_t action_ret)
 {
-	struct irqaction *action;
-
 	if (action_ret != IRQ_HANDLED && action_ret != IRQ_NONE) {
 		printk(KERN_ERR "irq event %d: bogus return value %x\n",
 				irq, action_ret);

@@ -136,16 +149,7 @@ __report_bad_irq(unsigned int irq, struct irq_desc *desc,
 				"the \"irqpoll\" option)\n", irq);
 	}
 	dump_stack();
-	printk(KERN_ERR "handlers:\n");
-
-	action = desc->action;
-	while (action) {
-		printk(KERN_ERR "[<%p>]", action->handler);
-		print_symbol(" (%s)",
-			(unsigned long)action->handler);
-		printk("\n");
-		action = action->next;
-	}
+	print_irq_handlers(desc);
 }
 
 static void

index 0bce0e3..0b8fd0a 100644
@@ -23,6 +23,8 @@ enum {
 
 	/* IRQ polling common parameters */
 	IRQ_POLL_INTV			= HZ / 100,	/* from the good ol' 100HZ tick */
+
+	IRQ_POLL_SLACK			= HZ / 1000,	/* 10% slack */
 };
 
 int noirqdebug __read_mostly;

@@ -43,6 +45,38 @@ static void print_irq_handlers(struct irq_desc *desc)
 	}
 }
 
+static unsigned long irq_poll_slack(unsigned long intv)
+{
+	return IRQ_POLL_SLACK;
+}
+
+/**
+ * irq_schedule_poll - schedule IRQ poll
+ * @desc: IRQ desc to schedule poll for
+ * @intv: poll interval
+ *
+ * Schedules @desc->poll_timer.  If the timer is already scheduled,
+ * it's modified iff jiffies + @intv + slack is before the timer's
+ * expires.  poll_timers aren't taken offline behind this function's
+ * back and the users of this function are guaranteed that poll_irq()
+ * will be called at or before jiffies + @intv + slack.
+ *
+ * CONTEXT:
+ * desc->lock
+ */
+static void irq_schedule_poll(struct irq_desc *desc, unsigned long intv)
+{
+	unsigned long expires = jiffies + intv;
+	int slack = irq_poll_slack(intv);
+
+	if (timer_pending(&desc->poll_timer) &&
+	    time_before_eq(desc->poll_timer.expires, expires + slack))
+		return;
+
+	set_timer_slack(&desc->poll_timer, slack);
+	mod_timer(&desc->poll_timer, expires);
+}
+
 /*
  * Recovery handler for misrouted interrupts.
  */

@@ -207,7 +241,9 @@ void __note_interrupt(unsigned int irq, struct irq_desc *desc,
 		desc->depth++;
 		desc->chip->disable(irq);
 
-		mod_timer(&desc->poll_timer, jiffies + IRQ_POLL_INTV);
+		raw_spin_lock(&desc->lock);
+		irq_schedule_poll(desc, IRQ_POLL_INTV);
+		raw_spin_unlock(&desc->lock);
 	}
 	desc->irqs_unhandled = 0;
 }

@@ -221,9 +257,8 @@ void poll_irq(unsigned long arg)
 
 	raw_spin_lock_irq(&desc->lock);
 	try_one_irq(desc->irq, desc);
+	irq_schedule_poll(desc, IRQ_POLL_INTV);
 	raw_spin_unlock_irq(&desc->lock);
-
-	mod_timer(&desc->poll_timer, jiffies + IRQ_POLL_INTV);
 }
 
 void irq_poll_action_added(struct irq_desc *desc, struct irqaction *action)

@@ -238,10 +273,10 @@ void irq_poll_action_added(struct irq_desc *desc, struct irqaction *action)
 		__enable_irq(desc, desc->irq, false);
 	}
 
-	raw_spin_unlock_irqrestore(&desc->lock, flags);
-
 	if ((action->flags & IRQF_SHARED) && irqfixup >= IRQFIXUP_POLL)
-		mod_timer(&desc->poll_timer, jiffies + IRQ_POLL_INTV);
+		irq_schedule_poll(desc, IRQ_POLL_INTV);
+
+	raw_spin_unlock_irqrestore(&desc->lock, flags);
 }
 
 void irq_poll_action_removed(struct irq_desc *desc, struct irqaction *action)

Index: work/kernel/irq/spurious.c
===================================================================
@@ -23,6 +23,8 @@ enum {
 
 	/* IRQ polling common parameters */
 	IRQ_POLL_INTV			= HZ / 100,	/* from the good ol' 100HZ tick */
+
+	IRQ_POLL_SLACK			= HZ / 250,	/* 1 tick slack w/ the popular 250HZ config */
 };
 
 int noirqdebug __read_mostly;

@@ -43,6 +45,38 @@ static void print_irq_handlers(struct ir
 	}
 }
 
+static unsigned long irq_poll_slack(unsigned long intv)
+{
+	return IRQ_POLL_SLACK;
+}
+
+/**
+ * irq_schedule_poll - schedule IRQ poll
+ * @desc: IRQ desc to schedule poll for
+ * @intv: poll interval
+ *
+ * Schedules @desc->poll_timer.  If the timer is already scheduled,
+ * it's modified iff jiffies + @intv + slack is before the timer's
+ * expires.  poll_timers aren't taken offline behind this function's
+ * back and the users of this function are guaranteed that poll_irq()
+ * will be called at or before jiffies + @intv + slack.
+ *
+ * CONTEXT:
+ * desc->lock
+ */
+static void irq_schedule_poll(struct irq_desc *desc, unsigned long intv)
+{
+	unsigned long expires = jiffies + intv;
+	int slack = irq_poll_slack(intv);
+
+	if (timer_pending(&desc->poll_timer) &&
+	    time_before_eq(desc->poll_timer.expires, expires + slack))
+		return;
+
+	set_timer_slack(&desc->poll_timer, slack);
+	mod_timer(&desc->poll_timer, expires);
+}
+
 /*
  * Recovery handler for misrouted interrupts.
  */

@@ -207,7 +241,9 @@ void __note_interrupt(unsigned int irq,
 		desc->depth++;
 		desc->chip->disable(irq);
 
-		mod_timer(&desc->poll_timer, jiffies + IRQ_POLL_INTV);
+		raw_spin_lock(&desc->lock);
+		irq_schedule_poll(desc, IRQ_POLL_INTV);
+		raw_spin_unlock(&desc->lock);
 	}
 	desc->irqs_unhandled = 0;
 }

@@ -221,9 +257,8 @@ void poll_irq(unsigned long arg)
 
 	raw_spin_lock_irq(&desc->lock);
 	try_one_irq(desc->irq, desc);
+	irq_schedule_poll(desc, IRQ_POLL_INTV);
 	raw_spin_unlock_irq(&desc->lock);
-
-	mod_timer(&desc->poll_timer, jiffies + IRQ_POLL_INTV);
 }
 
 void irq_poll_action_added(struct irq_desc *desc, struct irqaction *action)

@@ -238,10 +273,10 @@ void irq_poll_action_added(struct irq_de
 		__enable_irq(desc, desc->irq, false);
 	}
 
-	raw_spin_unlock_irqrestore(&desc->lock, flags);
-
 	if ((action->flags & IRQF_SHARED) && irqfixup >= IRQFIXUP_POLL)
-		mod_timer(&desc->poll_timer, jiffies + IRQ_POLL_INTV);
+		irq_schedule_poll(desc, IRQ_POLL_INTV);
+
+	raw_spin_unlock_irqrestore(&desc->lock, flags);
 }
 
 void irq_poll_action_removed(struct irq_desc *desc, struct irqaction *action)

index 50a77f9..b2f73ba 100644
@@ -145,6 +145,17 @@ struct irq_chip {
 
 struct timer_rand_state;
 struct irq_2_iommu;
+
+/* spurious IRQ tracking and handling */
+struct irq_spr {
+	unsigned long		last_bad;	/* when was the last bad? */
+	unsigned long		period_start;	/* period start jiffies */
+	unsigned int		nr_samples;	/* nr of irqs in this period */
+	unsigned int		nr_bad;		/* nr of bad deliveries */
+	unsigned int		poll_cnt;	/* nr to poll once activated */
+	unsigned int		poll_rem;	/* how many polls are left? */
+};
+
 /**
  * struct irq_desc - interrupt descriptor
  * @irq:		interrupt number for this descriptor

@@ -161,15 +172,13 @@ struct irq_2_iommu;
  * @status:		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:		IRQ affinity on SMP
  * @node:		node index useful for balancing
  * @pending_mask:	pending rebalanced interrupts
  * @threads_active:	number of irqaction threads currently running
  * @wait_for_threads:	wait queue for sync_irq to wait for threaded handlers
+ * @spr:		data for spurious IRQ handling
  * @poll_timer:		timer for IRQ polling
  * @dir:		/proc/irq/ procfs entry
  * @name:		flow handler name for /proc/interrupts output

@@ -191,9 +200,6 @@ struct irq_desc {
 
 	unsigned int		depth;		/* nested irq disables */
 	unsigned int		wake_depth;	/* nested wake enables */
-	unsigned int		irq_count;	/* For detecting broken IRQs */
-	unsigned long		last_unhandled;	/* Aging timer for unhandled count */
-	unsigned int		irqs_unhandled;
 	raw_spinlock_t		lock;
 #ifdef CONFIG_SMP
 	cpumask_var_t		affinity;

@@ -206,6 +212,7 @@ struct irq_desc {
 	atomic_t		threads_active;
 	wait_queue_head_t       wait_for_threads;
 
+	struct irq_spr		spr;
 	struct timer_list	poll_timer;
 
 #ifdef CONFIG_PROC_FS

index db26ff0..45a87f5 100644
@@ -40,8 +40,6 @@ static void dynamic_irq_init_x(unsigned int irq, bool keep_chip_data)
 	if (!keep_chip_data)
 		desc->chip_data = NULL;
 	desc->action = NULL;
-	desc->irq_count = 0;
-	desc->irqs_unhandled = 0;
 #ifdef CONFIG_SMP
 	cpumask_setall(desc->affinity);
 #ifdef CONFIG_GENERIC_PENDING_IRQ

index 088e5d6..1b24309 100644
@@ -68,7 +68,7 @@ static inline void chip_bus_sync_unlock(unsigned int irq, struct irq_desc *desc)
 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);
+		irq, desc, desc->depth, desc->spr.nr_samples, desc->spr.nr_bad);
 	printk("->handle_irq():  %p, ", desc->handle_irq);
 	print_symbol("%s\n", (unsigned long)desc->handle_irq);
 	printk("->chip(): %p, ", desc->chip);

index cf9ab65..5862bfc 100644
@@ -801,10 +801,6 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new)
 	new->irq = irq;
 	*old_ptr = new;
 
-	/* Reset broken irq detection when installing new handler */
-	desc->irq_count = 0;
-	desc->irqs_unhandled = 0;
-
 	raw_spin_unlock_irqrestore(&desc->lock, flags);
 
 	irq_poll_action_added(desc, new);

index 09a2ee5..b072460 100644
@@ -205,10 +205,11 @@ static const struct file_operations irq_node_proc_fops = {
 static int irq_spurious_proc_show(struct seq_file *m, void *v)
 {
 	struct irq_desc *desc = irq_to_desc((long) m->private);
+	struct irq_spr *spr = &desc->spr;
 
 	seq_printf(m, "count %u\n" "unhandled %u\n" "last_unhandled %u ms\n",
-		   desc->irq_count, desc->irqs_unhandled,
-		   jiffies_to_msecs(desc->last_unhandled));
+		   spr->nr_samples, spr->nr_bad,
+		   jiffies_to_msecs(spr->last_bad));
 	return 0;
 }
 

368 lines not on this page · raw message

index 50a77f9..b2f73ba 100644
@@ -145,6 +145,17 @@ struct irq_chip {
 
 struct timer_rand_state;
 struct irq_2_iommu;
+
+/* spurious IRQ tracking and handling */
+struct irq_spr {
+	unsigned long		last_bad;	/* when was the last bad? */
+	unsigned long		period_start;	/* period start jiffies */
+	unsigned int		nr_samples;	/* nr of irqs in this period */
+	unsigned int		nr_bad;		/* nr of bad deliveries */
+	unsigned int		poll_cnt;	/* nr to poll once activated */
+	unsigned int		poll_rem;	/* how many polls are left? */
+};
+
 /**
  * struct irq_desc - interrupt descriptor
  * @irq:		interrupt number for this descriptor

@@ -161,15 +172,13 @@ struct irq_2_iommu;
  * @status:		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:		IRQ affinity on SMP
  * @node:		node index useful for balancing
  * @pending_mask:	pending rebalanced interrupts
  * @threads_active:	number of irqaction threads currently running
  * @wait_for_threads:	wait queue for sync_irq to wait for threaded handlers
+ * @spr:		data for spurious IRQ handling
  * @poll_timer:		timer for IRQ polling
  * @dir:		/proc/irq/ procfs entry
  * @name:		flow handler name for /proc/interrupts output

@@ -191,9 +200,6 @@ struct irq_desc {
 
 	unsigned int		depth;		/* nested irq disables */
 	unsigned int		wake_depth;	/* nested wake enables */
-	unsigned int		irq_count;	/* For detecting broken IRQs */
-	unsigned long		last_unhandled;	/* Aging timer for unhandled count */
-	unsigned int		irqs_unhandled;
 	raw_spinlock_t		lock;
 #ifdef CONFIG_SMP
 	cpumask_var_t		affinity;

@@ -206,6 +212,7 @@ struct irq_desc {
 	atomic_t		threads_active;
 	wait_queue_head_t       wait_for_threads;
 
+	struct irq_spr		spr;
 	struct timer_list	poll_timer;
 
 #ifdef CONFIG_PROC_FS

index db26ff0..45a87f5 100644
@@ -40,8 +40,6 @@ static void dynamic_irq_init_x(unsigned int irq, bool keep_chip_data)
 	if (!keep_chip_data)
 		desc->chip_data = NULL;
 	desc->action = NULL;
-	desc->irq_count = 0;
-	desc->irqs_unhandled = 0;
 #ifdef CONFIG_SMP
 	cpumask_setall(desc->affinity);
 #ifdef CONFIG_GENERIC_PENDING_IRQ

index 088e5d6..1b24309 100644
@@ -68,7 +68,7 @@ static inline void chip_bus_sync_unlock(unsigned int irq, struct irq_desc *desc)
 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);
+		irq, desc, desc->depth, desc->spr.nr_samples, desc->spr.nr_bad);
 	printk("->handle_irq():  %p, ", desc->handle_irq);
 	print_symbol("%s\n", (unsigned long)desc->handle_irq);
 	printk("->chip(): %p, ", desc->chip);

index cf9ab65..5862bfc 100644
@@ -801,10 +801,6 @@ __setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new)
 	new->irq = irq;
 	*old_ptr = new;
 
-	/* Reset broken irq detection when installing new handler */
-	desc->irq_count = 0;
-	desc->irqs_unhandled = 0;
-
 	raw_spin_unlock_irqrestore(&desc->lock, flags);
 
 	irq_poll_action_added(desc, new);

index 09a2ee5..b072460 100644
@@ -205,10 +205,11 @@ static const struct file_operations irq_node_proc_fops = {
 static int irq_spurious_proc_show(struct seq_file *m, void *v)
 {
 	struct irq_desc *desc = irq_to_desc((long) m->private);
+	struct irq_spr *spr = &desc->spr;
 
 	seq_printf(m, "count %u\n" "unhandled %u\n" "last_unhandled %u ms\n",
-		   desc->irq_count, desc->irqs_unhandled,
-		   jiffies_to_msecs(desc->last_unhandled));
+		   spr->nr_samples, spr->nr_bad,
+		   jiffies_to_msecs(spr->last_bad));
 	return 0;
 }
 

368 lines not on this page · raw message

index b20bd65..bc0cdbc 100644
@@ -88,6 +88,14 @@ enum {
 
 typedef irqreturn_t (*irq_handler_t)(int, void *);
 
+struct irq_watch {
+	irqreturn_t		last_ret;
+	unsigned int		flags;
+	unsigned long		started;
+	unsigned int		nr_samples;
+	unsigned int		nr_polled;
+};
+
 /**
  * struct irqaction - per interrupt action descriptor
  * @handler:	interrupt handler function

@@ -100,6 +108,7 @@ 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
+ * @watch:	data for irq watching
  */
 struct irqaction {
 	irq_handler_t		handler;

@@ -112,6 +121,7 @@ struct irqaction {
 	irq_handler_t		thread_fn;
 	struct task_struct	*thread;
 	unsigned long		thread_flags;
+	struct irq_watch	watch;
 };
 
 extern irqreturn_t no_action(int cpl, void *dev_id);

@@ -184,6 +194,8 @@ devm_request_irq(struct device *dev, unsigned int irq, irq_handler_t handler,
 
 extern void devm_free_irq(struct device *dev, unsigned int irq, void *dev_id);
 
+extern void watch_irq(unsigned int irq, void *dev_id);
+
 /*
  * On lockdep we dont want to enable hardirqs in hardirq
  * context. Use local_irq_enable_in_hardirq() to annotate

index b2f73ba..e31954f 100644
@@ -72,6 +72,7 @@ typedef	void (*irq_flow_handler_t)(unsigned int irq,
 #define IRQ_SUSPENDED		0x04000000	/* IRQ has gone through suspend sequence */
 #define IRQ_ONESHOT		0x08000000	/* IRQ is not unmasked after hardirq */
 #define IRQ_NESTED_THREAD	0x10000000	/* IRQ is nested into another, no own handler thread */
+#define IRQ_CHECK_WATCHES	0x40000000	/* IRQ watch enabled */
 
 #ifdef CONFIG_IRQ_PER_CPU
 # define CHECK_IRQ_PER_CPU(var) ((var) & IRQ_PER_CPU)

@@ -214,6 +215,7 @@ struct irq_desc {
 
 	struct irq_spr		spr;
 	struct timer_list	poll_timer;
+	bool			poll_warned;
 
 #ifdef CONFIG_PROC_FS
 	struct proc_dir_entry	*dir;

index 3ae50bf..685c3b3 100644
@@ -416,6 +416,7 @@ irqreturn_t handle_IRQ_event(unsigned int irq, struct irqaction *action)
 		}
 
 		retval |= ret;
+		action->watch.last_ret = ret;
 		action = action->next;
 	} while (action);
 

264 lines not on this page · raw message

index b20bd65..bc0cdbc 100644
@@ -88,6 +88,14 @@ enum {
 
 typedef irqreturn_t (*irq_handler_t)(int, void *);
 
+struct irq_watch {
+	irqreturn_t		last_ret;
+	unsigned int		flags;
+	unsigned long		started;
+	unsigned int		nr_samples;
+	unsigned int		nr_polled;
+};
+
 /**
  * struct irqaction - per interrupt action descriptor
  * @handler:	interrupt handler function

@@ -100,6 +108,7 @@ 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
+ * @watch:	data for irq watching
  */
 struct irqaction {
 	irq_handler_t		handler;

@@ -112,6 +121,7 @@ struct irqaction {
 	irq_handler_t		thread_fn;
 	struct task_struct	*thread;
 	unsigned long		thread_flags;
+	struct irq_watch	watch;
 };
 
 extern irqreturn_t no_action(int cpl, void *dev_id);

@@ -184,6 +194,8 @@ devm_request_irq(struct device *dev, unsigned int irq, irq_handler_t handler,
 
 extern void devm_free_irq(struct device *dev, unsigned int irq, void *dev_id);
 
+extern void watch_irq(unsigned int irq, void *dev_id);
+
 /*
  * On lockdep we dont want to enable hardirqs in hardirq
  * context. Use local_irq_enable_in_hardirq() to annotate

index b2f73ba..e31954f 100644
@@ -72,6 +72,7 @@ typedef	void (*irq_flow_handler_t)(unsigned int irq,
 #define IRQ_SUSPENDED		0x04000000	/* IRQ has gone through suspend sequence */
 #define IRQ_ONESHOT		0x08000000	/* IRQ is not unmasked after hardirq */
 #define IRQ_NESTED_THREAD	0x10000000	/* IRQ is nested into another, no own handler thread */
+#define IRQ_CHECK_WATCHES	0x40000000	/* IRQ watch enabled */
 
 #ifdef CONFIG_IRQ_PER_CPU
 # define CHECK_IRQ_PER_CPU(var) ((var) & IRQ_PER_CPU)

@@ -214,6 +215,7 @@ struct irq_desc {
 
 	struct irq_spr		spr;
 	struct timer_list	poll_timer;
+	bool			poll_warned;
 
 #ifdef CONFIG_PROC_FS
 	struct proc_dir_entry	*dir;

index 3ae50bf..685c3b3 100644
@@ -416,6 +416,7 @@ irqreturn_t handle_IRQ_event(unsigned int irq, struct irqaction *action)
 		}
 
 		retval |= ret;
+		action->watch.last_ret = ret;
 		action = action->next;
 	} while (action);
 

264 lines not on this page · raw message

index bc0cdbc..8bbd9dc 100644
@@ -88,6 +88,8 @@ enum {
 
 typedef irqreturn_t (*irq_handler_t)(int, void *);
 
+struct irq_expect;
+
 struct irq_watch {
 	irqreturn_t		last_ret;
 	unsigned int		flags;

@@ -109,6 +111,7 @@ struct irq_watch {
  * @thread:	thread pointer for threaded interrupts
  * @thread_flags:	flags related to @thread
  * @watch:	data for irq watching
+ * @expects:	data for irq expecting
  */
 struct irqaction {
 	irq_handler_t		handler;

@@ -122,6 +125,7 @@ struct irqaction {
 	struct task_struct	*thread;
 	unsigned long		thread_flags;
 	struct irq_watch	watch;
+	struct irq_expect	*expects;
 };
 
 extern irqreturn_t no_action(int cpl, void *dev_id);

@@ -194,6 +198,9 @@ devm_request_irq(struct device *dev, unsigned int irq, irq_handler_t handler,
 
 extern void devm_free_irq(struct device *dev, unsigned int irq, void *dev_id);
 
+extern struct irq_expect *init_irq_expect(unsigned int irq, void *dev_id);
+extern void expect_irq(struct irq_expect *exp);
+extern void unexpect_irq(struct irq_expect *exp, bool timedout);
 extern void watch_irq(unsigned int irq, void *dev_id);
 
 /*

index e31954f..98530ef 100644
@@ -72,6 +72,7 @@ typedef	void (*irq_flow_handler_t)(unsigned int irq,
 #define IRQ_SUSPENDED		0x04000000	/* IRQ has gone through suspend sequence */
 #define IRQ_ONESHOT		0x08000000	/* IRQ is not unmasked after hardirq */
 #define IRQ_NESTED_THREAD	0x10000000	/* IRQ is nested into another, no own handler thread */
+#define IRQ_IN_POLLING		0x20000000	/* IRQ polling in progress */
 #define IRQ_CHECK_WATCHES	0x40000000	/* IRQ watch enabled */
 
 #ifdef CONFIG_IRQ_PER_CPU

340 lines not on this page · raw message

index 2d92113..329555f 100644
@@ -2,8 +2,66 @@
  * linux/kernel/irq/spurious.c
  *
  * Copyright (C) 1992, 1998-2004 Linus Torvalds, Ingo Molnar
+ * Copyright (C) 2010            SUSE Linux Products GmbH
+ * Copyright (C) 2010            Tejun Heo <[email protected]>
  *
- * This file contains spurious interrupt handling.
+ * There are two ways interrupt handling can go wrong - too few or too
+ * many.  Due to misrouting or other issues, sometimes IRQs don't
+ * reach the driver while at other times an interrupt line gets stuck
+ * and a continuous spurious interrupts are generated.
+ *
+ * This file implements workaround for both cases.  Lost interrupts
+ * are handled by IRQ expecting and watching, and spurious interrupts
+ * by spurious polling.  All mechanisms need IRQF_SHARED to be set on
+ * the irqaction in question.
+ *
+ * Both lost interrupt workarounds require cooperation from drivers
+ * and can be chosen depending on how much information the driver can
+ * provide.
+ *
+ * - IRQ expecting
+ *
+ *   IRQ expecting is useful when the driver can tell when IRQs can be
+ *   expected; in other words, when IRQs are used to signal completion
+ *   of host initiated operations.  This is the surest way to work
+ *   around lost interrupts.
+ *
+ *   When the controller is expected to raise an IRQ, the driver
+ *   should call expect_irq() and, when the expected event happens or
+ *   times out, unexpect_irq().  IRQ subsystem polls the interrupt
+ *   inbetween.
+ *
+ *   As interrupts tend to keep working if it works at the beginning,
+ *   IRQ expecting implements "verified state".  After certain number
+ *   of successful IRQ deliveries, the irqaction becomes verified and
+ *   much longer polling interval is used.
+ *
+ * - IRQ watching
+ *
+ *   This can be used when the driver doesn't know when to exactly
+ *   expect and unexpect IRQs.  Once watch_irq() is called, the
+ *   irqaction is slowly polled for certain amount of time (1min).  If
+ *   IRQs are missed during that time, the irqaction is marked and
+ *   actively polled; otherwise, the watching is stopped.
+ *
+ *   In the most basic case, drivers can call this right after
+ *   registering an irqaction to verify IRQ delivery.  In many cases,
+ *   if IRQ works at the beginning, it keeps working, so just calling
+ *   watch_irq() once can provide decent protection against misrouted
+ *   IRQs.  It would also be a good idea to call watch_irq() when
+ *   timeouts are detected.
+ *
+ * - Spurious IRQ handling
+ *
+ *   All IRQs are continuously monitored and spurious IRQ handling
+ *   kicks in if there are too many spurious IRQs.  The IRQ is
+ *   disabled and the registered irqactions are polled.  The IRQ is
+ *   given another shot after certain number IRQs are handled or an
+ *   irqaction is added or removed.
+ *
+ * All of the above three mechanisms can be used together.  Spurious
+ * IRQ handling is enabled by default and drivers are free to expect
+ * and watch IRQs as they see fit.
  */
 
 #include <linux/jiffies.h>

@@ -17,6 +75,20 @@
 
 #include "internals.h"
 
+/*
+ * I spent quite some time thinking about each parameter but they
+ * still are just numbers pulled out of my ass.  If you think your ass
+ * is prettier than mine, please go ahead and suggest better ones.
+ *
+ * Most parameters are intentionally fixed constants and not
+ * adjustable through API.  The nature of IRQ delivery failures isn't
+ * usually dependent on specific drivers and the timing parameters are
+ * more about human perceivable latencies rather than any specific
+ * controller timing details, so figuring out constant values which
+ * can work for most cases shouldn't be too hard.  This allows tighter
+ * control over polling behaviors, eases future changes and makes the
+ * interface easy for drivers.
+ */
 enum {
 	/* irqfixup levels */
 	IRQFIXUP_SPURIOUS		= 0,		/* spurious storm detection */

index ddf8e48..9a0aaa0 100644
@@ -4972,6 +4972,8 @@ void ata_qc_complete(struct ata_queued_cmd *qc)
 {
 	struct ata_port *ap = qc->ap;
 
+	unexpect_irq(ap->irq_expect, false);
+
 	/* XXX: New EH and old EH use different mechanisms to
 	 * synchronize EH with regular execution path.
 	 *

@@ -5087,6 +5089,9 @@ int ata_qc_complete_multiple(struct ata_port *ap, u32 qc_active)
 		done_mask &= ~(1 << tag);
 	}
 
+	if (ap->qc_active)
+		expect_irq(ap->irq_expect);
+
 	return nr_done;
 }
 

@@ -5153,6 +5158,7 @@ void ata_qc_issue(struct ata_queued_cmd *qc)
 	qc->err_mask |= ap->ops->qc_issue(qc);
 	if (unlikely(qc->err_mask))
 		goto err;
+	expect_irq(ap->irq_expect);
 	return;
 
 sg_err:

@@ -6185,8 +6191,13 @@ int ata_host_activate(struct ata_host *host, int irq,
 	if (rc)
 		return rc;
 
-	for (i = 0; i < host->n_ports; i++)
-		ata_port_desc(host->ports[i], "irq %d", irq);
+	for (i = 0; i < host->n_ports; i++) {
+		struct ata_port *ap = host->ports[i];
+
+		if (!ata_port_is_dummy(ap))
+			ap->irq_expect = init_irq_expect(irq, host);
+		ata_port_desc(ap, "irq %d%s", irq, ap->irq_expect ? "+" : "");
+	}
 
 	rc = ata_host_register(host, sht);
 	/* if failed, just free the IRQ and leave ports alone */

index f77a673..f1ae3ec 100644
@@ -619,8 +619,10 @@ void ata_scsi_error(struct Scsi_Host *host)
 		 * handler doesn't diddle with those qcs.  This must
 		 * be done atomically w.r.t. setting QCFLAG_FAILED.
 		 */
-		if (nr_timedout)
+		if (nr_timedout) {
+			unexpect_irq(ap->irq_expect, true);
 			__ata_port_freeze(ap);
+		}
 
 		spin_unlock_irqrestore(ap->lock, flags);
 

index efa4a18..cc96f36 100644
@@ -2388,7 +2388,8 @@ int ata_pci_sff_activate_host(struct ata_host *host,
 	struct device *dev = host->dev;
 	struct pci_dev *pdev = to_pci_dev(dev);
 	const char *drv_name = dev_driver_string(host->dev);
-	int legacy_mode = 0, rc;
+	struct ata_port *ap[2] = { host->ports[0], host->ports[1] };
+	int legacy_mode = 0, i, rc;
 
 	rc = ata_host_start(host);
 	if (rc)

@@ -2422,29 +2423,29 @@ int ata_pci_sff_activate_host(struct ata_host *host,
 		if (rc)
 			goto out;
 
-		ata_port_desc(host->ports[0], "irq %d", pdev->irq);
-		ata_port_desc(host->ports[1], "irq %d", pdev->irq);
+		for (i = 0; i < 2; i++) {
+			if (!ata_port_is_dummy(ap[i]))
+				ap[i]->irq_expect =
+					init_irq_expect(pdev->irq, host);
+			ata_port_desc(ap[i], "irq %d%s",
+				      pdev->irq, ap[i]->irq_expect ? "+" : "");
+		}
 	} else if (legacy_mode) {
-		if (!ata_port_is_dummy(host->ports[0])) {
-			rc = devm_request_irq(dev, ATA_PRIMARY_IRQ(pdev),
-					      irq_handler, IRQF_SHARED,
-					      drv_name, host);
-			if (rc)
-				goto out;
+		unsigned int irqs[2] = { ATA_PRIMARY_IRQ(pdev),
+					 ATA_SECONDARY_IRQ(pdev) };
 
-			ata_port_desc(host->ports[0], "irq %d",
-				      ATA_PRIMARY_IRQ(pdev));
-		}
+		for (i = 0; i < 2; i++) {
+			if (ata_port_is_dummy(ap[i]))
+				continue;
 
-		if (!ata_port_is_dummy(host->ports[1])) {
-			rc = devm_request_irq(dev, ATA_SECONDARY_IRQ(pdev),
-					      irq_handler, IRQF_SHARED,
-					      drv_name, host);
+			rc = devm_request_irq(dev, irqs[i], irq_handler,
+					      IRQF_SHARED, drv_name, host);
 			if (rc)
 				goto out;
 
-			ata_port_desc(host->ports[1], "irq %d",
-				      ATA_SECONDARY_IRQ(pdev));
+			ap[i]->irq_expect = init_irq_expect(irqs[i], host);
+			ata_port_desc(ap[i], "irq %d%s",
+				      irqs[i], ap[i]->irq_expect ? "+" : "");
 		}
 	}
 

index b85f3ff..3f5f159 100644
@@ -751,6 +751,8 @@ struct ata_port {
 	struct ata_host		*host;
 	struct device 		*dev;
 
+	struct irq_expect	*irq_expect;	/* for irq expecting */
+
 	struct delayed_work	hotplug_task;
 	struct work_struct	scsi_rescan_task;
 

index ddf8e48..9a0aaa0 100644
@@ -4972,6 +4972,8 @@ void ata_qc_complete(struct ata_queued_cmd *qc)
 {
 	struct ata_port *ap = qc->ap;
 
+	unexpect_irq(ap->irq_expect, false);
+
 	/* XXX: New EH and old EH use different mechanisms to
 	 * synchronize EH with regular execution path.
 	 *

@@ -5087,6 +5089,9 @@ int ata_qc_complete_multiple(struct ata_port *ap, u32 qc_active)
 		done_mask &= ~(1 << tag);
 	}
 
+	if (ap->qc_active)
+		expect_irq(ap->irq_expect);
+
 	return nr_done;
 }
 

@@ -5153,6 +5158,7 @@ void ata_qc_issue(struct ata_queued_cmd *qc)
 	qc->err_mask |= ap->ops->qc_issue(qc);
 	if (unlikely(qc->err_mask))
 		goto err;
+	expect_irq(ap->irq_expect);
 	return;
 
 sg_err:

@@ -6185,8 +6191,13 @@ int ata_host_activate(struct ata_host *host, int irq,
 	if (rc)
 		return rc;
 
-	for (i = 0; i < host->n_ports; i++)
-		ata_port_desc(host->ports[i], "irq %d", irq);
+	for (i = 0; i < host->n_ports; i++) {
+		struct ata_port *ap = host->ports[i];
+
+		if (!ata_port_is_dummy(ap))
+			ap->irq_expect = init_irq_expect(irq, host);
+		ata_port_desc(ap, "irq %d%s", irq, ap->irq_expect ? "+" : "");
+	}
 
 	rc = ata_host_register(host, sht);
 	/* if failed, just free the IRQ and leave ports alone */

index f77a673..f1ae3ec 100644
@@ -619,8 +619,10 @@ void ata_scsi_error(struct Scsi_Host *host)
 		 * handler doesn't diddle with those qcs.  This must
 		 * be done atomically w.r.t. setting QCFLAG_FAILED.
 		 */
-		if (nr_timedout)
+		if (nr_timedout) {
+			unexpect_irq(ap->irq_expect, true);
 			__ata_port_freeze(ap);
+		}
 
 		spin_unlock_irqrestore(ap->lock, flags);
 

index efa4a18..cc96f36 100644
@@ -2388,7 +2388,8 @@ int ata_pci_sff_activate_host(struct ata_host *host,
 	struct device *dev = host->dev;
 	struct pci_dev *pdev = to_pci_dev(dev);
 	const char *drv_name = dev_driver_string(host->dev);
-	int legacy_mode = 0, rc;
+	struct ata_port *ap[2] = { host->ports[0], host->ports[1] };
+	int legacy_mode = 0, i, rc;
 
 	rc = ata_host_start(host);
 	if (rc)

@@ -2422,29 +2423,29 @@ int ata_pci_sff_activate_host(struct ata_host *host,
 		if (rc)
 			goto out;
 
-		ata_port_desc(host->ports[0], "irq %d", pdev->irq);
-		ata_port_desc(host->ports[1], "irq %d", pdev->irq);
+		for (i = 0; i < 2; i++) {
+			if (!ata_port_is_dummy(ap[i]))
+				ap[i]->irq_expect =
+					init_irq_expect(pdev->irq, host);
+			ata_port_desc(ap[i], "irq %d%s",
+				      pdev->irq, ap[i]->irq_expect ? "+" : "");
+		}
 	} else if (legacy_mode) {
-		if (!ata_port_is_dummy(host->ports[0])) {
-			rc = devm_request_irq(dev, ATA_PRIMARY_IRQ(pdev),
-					      irq_handler, IRQF_SHARED,
-					      drv_name, host);
-			if (rc)
-				goto out;
+		unsigned int irqs[2] = { ATA_PRIMARY_IRQ(pdev),
+					 ATA_SECONDARY_IRQ(pdev) };
 
-			ata_port_desc(host->ports[0], "irq %d",
-				      ATA_PRIMARY_IRQ(pdev));
-		}
+		for (i = 0; i < 2; i++) {
+			if (ata_port_is_dummy(ap[i]))
+				continue;
 
-		if (!ata_port_is_dummy(host->ports[1])) {
-			rc = devm_request_irq(dev, ATA_SECONDARY_IRQ(pdev),
-					      irq_handler, IRQF_SHARED,
-					      drv_name, host);
+			rc = devm_request_irq(dev, irqs[i], irq_handler,
+					      IRQF_SHARED, drv_name, host);
 			if (rc)
 				goto out;
 
-			ata_port_desc(host->ports[1], "irq %d",
-				      ATA_SECONDARY_IRQ(pdev));
+			ap[i]->irq_expect = init_irq_expect(irqs[i], host);
+			ata_port_desc(ap[i], "irq %d%s",
+				      irqs[i], ap[i]->irq_expect ? "+" : "");
 		}
 	}
 

index b85f3ff..3f5f159 100644
@@ -751,6 +751,8 @@ struct ata_port {
 	struct ata_host		*host;
 	struct device 		*dev;
 
+	struct irq_expect	*irq_expect;	/* for irq expecting */
+
 	struct delayed_work	hotplug_task;
 	struct work_struct	scsi_rescan_task;
 

Index: ata/drivers/ata/sata_fsl.c
===================================================================
@@ -1096,7 +1096,7 @@ static void sata_fsl_host_intr(struct at
 {
 	struct sata_fsl_host_priv *host_priv = ap->host->private_data;
 	void __iomem *hcr_base = host_priv->hcr_base;
-	u32 hstatus, qc_active = 0;
+	u32 hstatus, done_mask = 0;
 	struct ata_queued_cmd *qc;
 	u32 SError;
 

@@ -1116,28 +1116,28 @@ static void sata_fsl_host_intr(struct at
 	}
 
 	/* Read command completed register */
-	qc_active = ioread32(hcr_base + CC);
+	done_mask = ioread32(hcr_base + CC);
 
 	VPRINTK("Status of all queues :\n");
-	VPRINTK("qc_active/CC = 0x%x, CA = 0x%x, CE=0x%x,CQ=0x%x,apqa=0x%x\n",
-		qc_active,
+	VPRINTK("done_mask/CC = 0x%x, CA = 0x%x, CE=0x%x,CQ=0x%x,apqa=0x%x\n",
+		done_mask,
 		ioread32(hcr_base + CA),
 		ioread32(hcr_base + CE),
 		ioread32(hcr_base + CQ),
 		ap->qc_active);
 
-	if (qc_active & ap->qc_active) {
+	if (done_mask & ap->qc_active) {
 		int i;
 		/* clear CC bit, this will also complete the interrupt */
-		iowrite32(qc_active, hcr_base + CC);
+		iowrite32(done_mask, hcr_base + CC);
 
 		DPRINTK("Status of all queues :\n");
-		DPRINTK("qc_active/CC = 0x%x, CA = 0x%x, CE=0x%x\n",
-			qc_active, ioread32(hcr_base + CA),
+		DPRINTK("done_mask/CC = 0x%x, CA = 0x%x, CE=0x%x\n",
+			done_mask, ioread32(hcr_base + CA),
 			ioread32(hcr_base + CE));
 
 		for (i = 0; i < SATA_FSL_QUEUE_DEPTH; i++) {
-			if (qc_active & (1 << i)) {
+			if (done_mask & (1 << i)) {
 				qc = ata_qc_from_tag(ap, i);
 				if (qc) {
 					ata_qc_complete(qc);

@@ -1164,7 +1164,7 @@ static void sata_fsl_host_intr(struct at
 		/* Spurious Interrupt!! */
 		DPRINTK("spurious interrupt!!, CC = 0x%x\n",
 			ioread32(hcr_base + CC));
-		iowrite32(qc_active, hcr_base + CC);
+		iowrite32(done_mask, hcr_base + CC);
 		return;
 	}
 }

58 lines not on this page · raw message

Index: ata/drivers/ata/sata_nv.c
===================================================================
@@ -1018,7 +1018,7 @@ static irqreturn_t nv_adma_interrupt(int
 			      NV_ADMA_STAT_CPBERR |
 			      NV_ADMA_STAT_CMD_COMPLETE)) {
 			u32 check_commands = notifier_clears[i];
-			int pos, error = 0;
+			int pos, rc;
 
 			if (status & NV_ADMA_STAT_CPBERR) {
 				/* check all active commands */

@@ -1030,10 +1030,12 @@ static irqreturn_t nv_adma_interrupt(int
 			}
 
 			/* check CPBs for completed commands */
-			while ((pos = ffs(check_commands)) && !error) {
+			while ((pos = ffs(check_commands))) {
 				pos--;
-				error = nv_adma_check_cpb(ap, pos,
+				rc = nv_adma_check_cpb(ap, pos,
 						notifier_error & (1 << pos));
+				if (unlikely(rc))
+					check_commands = 0;
 				check_commands &= ~(1 << pos);
 			}
 		}

@@ -2129,7 +2131,6 @@ static int nv_swncq_sdbfis(struct ata_po
 	struct nv_swncq_port_priv *pp = ap->private_data;
 	struct ata_eh_info *ehi = &ap->link.eh_info;
 	u32 sactive;
-	int nr_done = 0;
 	u32 done_mask;
 	int i;
 	u8 host_stat;

@@ -2170,22 +2171,21 @@ static int nv_swncq_sdbfis(struct ata_po
 			pp->dhfis_bits &= ~(1 << i);
 			pp->dmafis_bits &= ~(1 << i);
 			pp->sdbfis_bits |= (1 << i);
-			nr_done++;
 		}
 	}
 
 	if (!ap->qc_active) {
 		DPRINTK("over\n");
 		nv_swncq_pp_reinit(ap);
-		return nr_done;
+		return 0;
 	}
 
 	if (pp->qc_active & pp->dhfis_bits)
-		return nr_done;
+		return 0;
 
 	if ((pp->ncq_flags & ncq_saw_backout) ||
 	    (pp->qc_active ^ pp->dhfis_bits))
-		/* if the controller cann't get a device to host register FIS,
+		/* if the controller can't get a device to host register FIS,
 		 * The driver needs to reissue the new command.
 		 */
 		lack_dhfis = 1;

@@ -2202,7 +2202,7 @@ static int nv_swncq_sdbfis(struct ata_po
 	if (lack_dhfis) {
 		qc = ata_qc_from_tag(ap, pp->last_issue_tag);
 		nv_swncq_issue_atacmd(ap, qc);
-		return nr_done;
+		return 0;
 	}
 
 	if (pp->defer_queue.defer_bits) {

@@ -2212,7 +2212,7 @@ static int nv_swncq_sdbfis(struct ata_po
 		nv_swncq_issue_atacmd(ap, qc);
 	}
 
-	return nr_done;
+	return 0;
 }
 
 static inline u32 nv_swncq_tag(struct ata_port *ap)

@@ -2224,7 +2224,7 @@ static inline u32 nv_swncq_tag(struct at
 	return (tag & 0x1f);
 }
 
-static int nv_swncq_dmafis(struct ata_port *ap)
+static void nv_swncq_dmafis(struct ata_port *ap)
 {
 	struct ata_queued_cmd *qc;
 	unsigned int rw;

@@ -2239,7 +2239,7 @@ static int nv_swncq_dmafis(struct ata_po
 	qc = ata_qc_from_tag(ap, tag);
 
 	if (unlikely(!qc))
-		return 0;
+		return;
 
 	rw = qc->tf.flags & ATA_TFLAG_WRITE;
 

@@ -2254,8 +2254,6 @@ static int nv_swncq_dmafis(struct ata_po
 		dmactl |= ATA_DMA_WR;
 
 	iowrite8(dmactl | ATA_DMA_START, ap->ioaddr.bmdma_addr + ATA_DMA_CMD);
-
-	return 1;
 }
 
 static void nv_swncq_host_interrupt(struct ata_port *ap, u16 fis)

@@ -2265,7 +2263,6 @@ static void nv_swncq_host_interrupt(stru
 	struct ata_eh_info *ehi = &ap->link.eh_info;
 	u32 serror;
 	u8 ata_stat;
-	int rc = 0;
 
 	ata_stat = ap->ops->sff_check_status(ap);
 	nv_swncq_irq_clear(ap, fis);

@@ -2310,8 +2307,7 @@ static void nv_swncq_host_interrupt(stru
 			"dhfis 0x%X dmafis 0x%X sactive 0x%X\n",
 			ap->print_id, pp->qc_active, pp->dhfis_bits,
 			pp->dmafis_bits, readl(pp->sactive_block));
-		rc = nv_swncq_sdbfis(ap);
-		if (rc < 0)
+		if (nv_swncq_sdbfis(ap) < 0)
 			goto irq_error;
 	}
 

@@ -2348,7 +2344,7 @@ static void nv_swncq_host_interrupt(stru
 		 */
 		pp->dmafis_bits |= (0x1 << nv_swncq_tag(ap));
 		pp->ncq_flags |= ncq_saw_dmas;
-		rc = nv_swncq_dmafis(ap);
+		nv_swncq_dmafis(ap);
 	}
 
 irq_exit:

Index: ata/drivers/ata/libata-core.c
===================================================================
@@ -4962,8 +4962,13 @@ static void ata_verify_xfer(struct ata_q
  *	ata_qc_complete - Complete an active ATA command
  *	@qc: Command to complete
  *
- *	Indicate to the mid and upper layers that an ATA
- *	command has completed, with either an ok or not-ok status.
+ *	Indicate to the mid and upper layers that an ATA command has
+ *	completed, with either an ok or not-ok status.
+ *
+ *	Refrain from calling this function multiple times when
+ *	successfully completing multiple NCQ commands.
+ *	ata_qc_complete_multiple() should be used instead, which will
+ *	properly update IRQ expect state.
  *
  *	LOCKING:
  *	spin_lock_irqsave(host lock)

@@ -5056,6 +5061,10 @@ void ata_qc_complete(struct ata_queued_c
  *	requests normally.  ap->qc_active and @qc_active is compared
  *	and commands are completed accordingly.
  *
+ *	Always use this function when completing multiple NCQ commands
+ *	from IRQ handlers instead of calling ata_qc_complete()
+ *	multiple times to keep IRQ expect status properly in sync.
+ *
  *	LOCKING:
  *	spin_lock_irqsave(host lock)
  *

Index: ata/drivers/ata/sata_fsl.c
===================================================================
@@ -1137,17 +1137,13 @@ static void sata_fsl_host_intr(struct at
 			ioread32(hcr_base + CE));
 
 		for (i = 0; i < SATA_FSL_QUEUE_DEPTH; i++) {
-			if (done_mask & (1 << i)) {
-				qc = ata_qc_from_tag(ap, i);
-				if (qc) {
-					ata_qc_complete(qc);
-				}
+			if (done_mask & (1 << i))
 				DPRINTK
 				    ("completing ncq cmd,tag=%d,CC=0x%x,CA=0x%x\n",
 				     i, ioread32(hcr_base + CC),
 				     ioread32(hcr_base + CA));
-			}
 		}
+		ata_qc_complete_multiple(ap, ap->qc_active ^ done_mask);
 		return;
 
 	} else if ((ap->qc_active & (1 << ATA_TAG_INTERNAL))) {

Index: ata/drivers/ata/sata_mv.c
===================================================================
@@ -2713,18 +2713,11 @@ static void mv_err_intr(struct ata_port
 	}
 }
 
-static void mv_process_crpb_response(struct ata_port *ap,
+static bool mv_process_crpb_response(struct ata_port *ap,
 		struct mv_crpb *response, unsigned int tag, int ncq_enabled)
 {
 	u8 ata_status;
 	u16 edma_status = le16_to_cpu(response->flags);
-	struct ata_queued_cmd *qc = ata_qc_from_tag(ap, tag);
-
-	if (unlikely(!qc)) {
-		ata_port_printk(ap, KERN_ERR, "%s: no qc for tag=%d\n",
-				__func__, tag);
-		return;
-	}
 
 	/*
 	 * edma_status from a response queue entry: