arm_pmu: note IRQs and PMUs per-cpu

To support ACPI systems, we need to request IRQs before we know the
associated PMU, and thus we need some percpu variable that the IRQ
handler can find the PMU from.

As we're going to request IRQs without the PMU, we can't rely on the
arm_pmu::active_irqs mask, and similarly need to track requested IRQs
with a percpu variable.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
[will: made armpmu_count_irq_users static]
Signed-off-by: Will Deacon <will.deacon@arm.com>

1 files changed, 52 insertions, 17 deletions

diff --git a/drivers/perf/arm_pmu.c b/drivers/perf/arm_pmu.c
index 72118e6f9122..2b2af35db1b6 100644
--- a/drivers/perf/arm_pmu.c
+++ b/drivers/perf/arm_pmu.c
@@ -25,6 +25,9 @@
 
 #include <asm/irq_regs.h>
 
+static DEFINE_PER_CPU(struct arm_pmu *, cpu_armpmu);
+static DEFINE_PER_CPU(int, cpu_irq);
+
 static int
 armpmu_map_cache_event(const unsigned (*cache_map)
 				      [PERF_COUNT_HW_CACHE_MAX]
@@ -332,6 +335,8 @@ static irqreturn_t armpmu_dispatch_irq(int irq, void *dev)
 	 * dereference.
 	 */
 	armpmu = *(void **)dev;
+	if (WARN_ON_ONCE(!armpmu))
+		return IRQ_NONE;
 
 	start_clock = sched_clock();
 	ret = armpmu->handle_irq(irq, armpmu);
@@ -517,29 +522,45 @@ int perf_num_counters(void)
 }
 EXPORT_SYMBOL_GPL(perf_num_counters);
 
-void armpmu_free_irq(struct arm_pmu *armpmu, int cpu)
+static int armpmu_count_irq_users(const int irq)
 {
-	struct pmu_hw_events __percpu *hw_events = armpmu->hw_events;
-	int irq = per_cpu(hw_events->irq, cpu);
+	int cpu, count = 0;
 
-	if (!cpumask_test_and_clear_cpu(cpu, &armpmu->active_irqs))
-		return;
+	for_each_possible_cpu(cpu) {
+		if (per_cpu(cpu_irq, cpu) == irq)
+			count++;
+	}
+
+	return count;
+}
 
-	if (irq_is_percpu_devid(irq)) {
-		free_percpu_irq(irq, &hw_events->percpu_pmu);
-		cpumask_clear(&armpmu->active_irqs);
+void armpmu_free_cpu_irq(int irq, int cpu)
+{
+	if (per_cpu(cpu_irq, cpu) == 0)
 		return;
-	}
+	if (WARN_ON(irq != per_cpu(cpu_irq, cpu)))
+		return;
+
+	if (!irq_is_percpu_devid(irq))
+		free_irq(irq, per_cpu_ptr(&cpu_armpmu, cpu));
+	else if (armpmu_count_irq_users(irq) == 1)
+		free_percpu_irq(irq, &cpu_armpmu);
 
-	free_irq(irq, per_cpu_ptr(&hw_events->percpu_pmu, cpu));
+	per_cpu(cpu_irq, cpu) = 0;
 }
 
-int armpmu_request_irq(struct arm_pmu *armpmu, int cpu)
+void armpmu_free_irq(struct arm_pmu *armpmu, int cpu)
 {
-	int err = 0;
 	struct pmu_hw_events __percpu *hw_events = armpmu->hw_events;
-	const irq_handler_t handler = armpmu_dispatch_irq;
 	int irq = per_cpu(hw_events->irq, cpu);
+
+	armpmu_free_cpu_irq(irq, cpu);
+}
+
+int armpmu_request_cpu_irq(int irq, int cpu)
+{
+	int err = 0;
+	const irq_handler_t handler = armpmu_dispatch_irq;
 	if (!irq)
 		return 0;
 
@@ -560,16 +581,16 @@ int armpmu_request_irq(struct arm_pmu *armpmu, int cpu)
 
 		irq_set_status_flags(irq, IRQ_NOAUTOEN);
 		err = request_irq(irq, handler, irq_flags, "arm-pmu",
-				  per_cpu_ptr(&hw_events->percpu_pmu, cpu));
-	} else if (cpumask_empty(&armpmu->active_irqs)) {
+				  per_cpu_ptr(&cpu_armpmu, cpu));
+	} else if (armpmu_count_irq_users(irq) == 0) {
 		err = request_percpu_irq(irq, handler, "arm-pmu",
-					 &hw_events->percpu_pmu);
+					 &cpu_armpmu);
 	}
 
 	if (err)
 		goto err_out;
 
-	cpumask_set_cpu(cpu, &armpmu->active_irqs);
+	per_cpu(cpu_irq, cpu) = irq;
 	return 0;
 
 err_out:
@@ -577,6 +598,16 @@ err_out:
 	return err;
 }
 
+int armpmu_request_irq(struct arm_pmu *armpmu, int cpu)
+{
+	struct pmu_hw_events __percpu *hw_events = armpmu->hw_events;
+	int irq = per_cpu(hw_events->irq, cpu);
+	if (!irq)
+		return 0;
+
+	return armpmu_request_cpu_irq(irq, cpu);
+}
+
 static int armpmu_get_cpu_irq(struct arm_pmu *pmu, int cpu)
 {
 	struct pmu_hw_events __percpu *hw_events = pmu->hw_events;
@@ -599,6 +630,8 @@ static int arm_perf_starting_cpu(unsigned int cpu, struct hlist_node *node)
 	if (pmu->reset)
 		pmu->reset(pmu);
 
+	per_cpu(cpu_armpmu, cpu) = pmu;
+
 	irq = armpmu_get_cpu_irq(pmu, cpu);
 	if (irq) {
 		if (irq_is_percpu_devid(irq))
@@ -626,6 +659,8 @@ static int arm_perf_teardown_cpu(unsigned int cpu, struct hlist_node *node)
 			disable_irq(irq);
 	}
 
+	per_cpu(cpu_armpmu, cpu) = NULL;
+
 	return 0;
 }