On Mon, Dec 02, 2013 at 09:34:03AM +0000, Vinayak Kale wrote:

 static void
+armpmu_disable_percpu_irq(void *data)
+{
+	struct arm_pmu *armpmu = data;
+	struct platform_device *pmu_device = armpmu->plat_device;
+	int irq = platform_get_irq(pmu_device, 0);
+
+	cpumask_test_and_clear_cpu(smp_processor_id(), &armpmu->active_irqs);

Why not just cpumask_clear_cpu?

+	disable_percpu_irq(irq);
+}
+
+static void
 armpmu_release_hardware(struct arm_pmu *armpmu)
 {
-	int i, irq, irqs;
+	int irq;
+	unsigned int i, irqs;
 	struct platform_device *pmu_device = armpmu->plat_device;
 
 	irqs = min(pmu_device->num_resources, num_possible_cpus());
+	if (!irqs)
+		return;
 
-	for (i = 0; i < irqs; ++i) {
-		if (!cpumask_test_and_clear_cpu(i, &armpmu->active_irqs))
-			continue;
-		irq = platform_get_irq(pmu_device, i);
-		if (irq >= 0)
-			free_irq(irq, armpmu);
+	irq = platform_get_irq(pmu_device, 0);
+	if (irq <= 0)
+		return;
+
+	if (irq_is_percpu(irq)) {
+		on_each_cpu(armpmu_disable_percpu_irq, armpmu, 1);
+		free_percpu_irq(irq, &cpu_hw_events);
+	} else {
+		for (i = 0; i < irqs; ++i) {
+			if (!cpumask_test_and_clear_cpu(i, &armpmu->active_irqs))
+				continue;
+			irq = platform_get_irq(pmu_device, i);
+			if (irq > 0)
+				free_irq(irq, armpmu);
+		}
 	}
 }
 
+static void
+armpmu_enable_percpu_irq(void *data)
+{
+	struct arm_pmu *armpmu = data;
+	struct platform_device *pmu_device = armpmu->plat_device;
+	int irq = platform_get_irq(pmu_device, 0);
+
+	enable_percpu_irq(irq, IRQ_TYPE_NONE);
+	cpumask_set_cpu(smp_processor_id(), &armpmu->active_irqs);

Hmm, wouldn't it make more sense to pass the irq in data, then deal with the
mask in the caller? (since the mask will *always* be updated by each CPU).

Similarly for the disable path.

Will