Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit abdf655a authored by Mark Rutland's avatar Mark Rutland Committed by Will Deacon
Browse files

arm: perf: dynamically allocate cpu hardware data 



To support multiple PMUs, each PMU will need its own accounting data.
As we don't know how (in general) many PMUs we'll have to support at
compile-time, we must allocate the data at runtime dynamically

Signed-off-by: default avatarMark Rutland <mark.rutland@arm.com>
Reviewed-by: default avatarStephen Boyd <sboyd@codeaurora.org>
Signed-off-by: default avatarWill Deacon <will.deacon@arm.com>
parent 5ebd9200

arch/arm/kernel/perf_event_cpu.c

+25 −8
Original line number Original line Diff line number Diff line
@@ -35,8 +35,6 @@ 
/* Set at runtime when we know what CPU type we are. */

/* Set at runtime when we know what CPU type we are. */

static struct arm_pmu *cpu_pmu;

static struct arm_pmu *cpu_pmu;





static DEFINE_PER_CPU(struct pmu_hw_events, cpu_hw_events);



/*

/*

 * Despite the names, these two functions are CPU-specific and are used

 * Despite the names, these two functions are CPU-specific and are used

 * by the OProfile/perf code.

 * by the OProfile/perf code.
@@ -162,16 +160,22 @@ static int cpu_pmu_request_irq(struct arm_pmu *cpu_pmu, irq_handler_t handler) 
	return 0;

	return 0;

}

}





static void cpu_pmu_init(struct arm_pmu *cpu_pmu)

static int cpu_pmu_init(struct arm_pmu *cpu_pmu)

{

{

	int cpu;

	int cpu;

	struct pmu_hw_events __percpu *cpu_hw_events;



	cpu_hw_events = alloc_percpu(struct pmu_hw_events);

	if (!cpu_hw_events)

		return -ENOMEM;



	for_each_possible_cpu(cpu) {

	for_each_possible_cpu(cpu) {

		struct pmu_hw_events *events = &per_cpu(cpu_hw_events, cpu);

		struct pmu_hw_events *events = per_cpu_ptr(cpu_hw_events, cpu);

		raw_spin_lock_init(&events->pmu_lock);

		raw_spin_lock_init(&events->pmu_lock);

		events->percpu_pmu = cpu_pmu;

		events->percpu_pmu = cpu_pmu;

	}

	}





	cpu_pmu->hw_events	= &cpu_hw_events;

	cpu_pmu->hw_events	= cpu_hw_events;

	cpu_pmu->request_irq	= cpu_pmu_request_irq;

	cpu_pmu->request_irq	= cpu_pmu_request_irq;

	cpu_pmu->free_irq	= cpu_pmu_free_irq;

	cpu_pmu->free_irq	= cpu_pmu_free_irq;
@@ -182,6 +186,13 @@ static void cpu_pmu_init(struct arm_pmu *cpu_pmu) 
	/* If no interrupts available, set the corresponding capability flag */

	/* If no interrupts available, set the corresponding capability flag */

	if (!platform_get_irq(cpu_pmu->plat_device, 0))

	if (!platform_get_irq(cpu_pmu->plat_device, 0))

		cpu_pmu->pmu.capabilities |= PERF_PMU_CAP_NO_INTERRUPT;

		cpu_pmu->pmu.capabilities |= PERF_PMU_CAP_NO_INTERRUPT;



	return 0;

}



static void cpu_pmu_destroy(struct arm_pmu *cpu_pmu)

{

	free_percpu(cpu_pmu->hw_events);

}

}





/*

/*
@@ -303,12 +314,18 @@ static int cpu_pmu_device_probe(struct platform_device *pdev) 
		goto out_free;

		goto out_free;

	}

	}





	cpu_pmu_init(cpu_pmu);

	ret = cpu_pmu_init(cpu_pmu);

	if (ret)

		goto out_free;



	ret = armpmu_register(cpu_pmu, -1);

	ret = armpmu_register(cpu_pmu, -1);

	if (ret)

		goto out_destroy;





	if (!ret)

	return 0;

	return 0;





out_destroy:

	cpu_pmu_destroy(cpu_pmu);

out_free:

out_free:

	pr_info("failed to register PMU devices!\n");

	pr_info("failed to register PMU devices!\n");

	kfree(pmu);

	kfree(pmu);