On Fri, May 03, 2019 at 06:24:06PM -0500, Jeremy Linton wrote:
> ACPI 6.3 adds additional fields to the MADT GICC
> structure to describe SPE PPI's. We pick these out
> of the cached reference to the madt_gicc structure
> similarly to the core PMU code. We then create a platform
> device referring to the IRQ and let the user/module loader
> decide whether to load the SPE driver.
>
> Signed-off-by: Jeremy Linton <jeremy.linton@xxxxxxx>
> ---
> arch/arm64/include/asm/acpi.h | 3 ++
> drivers/perf/Kconfig | 5 +++
> drivers/perf/arm_pmu_acpi.c | 76 +++++++++++++++++++++++++++++++++++
> include/linux/perf/arm_pmu.h | 2 +
> 4 files changed, 86 insertions(+)
>
> diff --git a/arch/arm64/include/asm/acpi.h b/arch/arm64/include/asm/acpi.h
> index 7628efbe6c12..d10399b9f998 100644
> --- a/arch/arm64/include/asm/acpi.h
> +++ b/arch/arm64/include/asm/acpi.h
> @@ -41,6 +41,9 @@
> (!(entry) || (entry)->header.length < ACPI_MADT_GICC_MIN_LENGTH || \
> (unsigned long)(entry) + (entry)->header.length > (end))
>
> +#define ACPI_MADT_GICC_SPE (ACPI_OFFSET(struct acpi_madt_generic_interrupt, \
> + spe_interrupt) + sizeof(u16))
> +
> /* Basic configuration for ACPI */
> #ifdef CONFIG_ACPI
> pgprot_t __acpi_get_mem_attribute(phys_addr_t addr);
> diff --git a/drivers/perf/Kconfig b/drivers/perf/Kconfig
> index af9bc178495d..bc2647c64c9d 100644
> --- a/drivers/perf/Kconfig
> +++ b/drivers/perf/Kconfig
> @@ -52,6 +52,11 @@ config ARM_PMU_ACPI
> depends on ARM_PMU && ACPI
> def_bool y
>
> +config ARM_SPE_ACPI
> + depends on ARM_PMU_ACPI && ARM_SPE_PMU
> + def_bool y
> +
> +
> config ARM_DSU_PMU
> tristate "ARM DynamIQ Shared Unit (DSU) PMU"
> depends on ARM64
> diff --git a/drivers/perf/arm_pmu_acpi.c b/drivers/perf/arm_pmu_acpi.c
> index 0f197516d708..b0244e1e8c91 100644
> --- a/drivers/perf/arm_pmu_acpi.c
> +++ b/drivers/perf/arm_pmu_acpi.c
> @@ -74,6 +74,80 @@ static void arm_pmu_acpi_unregister_irq(int cpu)
> acpi_unregister_gsi(gsi);
> }
>
> +#ifdef CONFIG_ARM_SPE_ACPI
> +static struct resource spe_resources[] = {
> + {
> + /* irq */
> + .flags = IORESOURCE_IRQ,
> + }
> +};
> +
> +static struct platform_device spe_dev = {
> + .name = ARMV8_SPE_PDEV_NAME,
> + .id = -1,
> + .resource = spe_resources,
> + .num_resources = ARRAY_SIZE(spe_resources)
> +};
> +
> +/*
> + * For lack of a better place, hook the normal PMU MADT walk
> + * and create a SPE device if we detect a recent MADT with
> + * a homogeneous PPI mapping.
> + */
> +static int arm_spe_acpi_register_device(void)
> +{
> + int cpu, ret, irq;
> + int hetid;
> + u16 gsi = 0;
> + bool first = true;
> +
> + struct acpi_madt_generic_interrupt *gicc;
> +
> + /*
> + * sanity check all the GICC tables for the same interrupt number
> + * for now we only support homogeneous ACPI/SPE machines.
> + */
> + for_each_possible_cpu(cpu) {
> + gicc = acpi_cpu_get_madt_gicc(cpu);
> +
> + if (gicc->header.length < ACPI_MADT_GICC_SPE)
> + return -ENODEV;
> + if (first) {
> + gsi = gicc->spe_interrupt;
> + if (!gsi)
> + return -ENODEV;
> + hetid = find_acpi_cpu_topology_hetero_id(cpu);
> + first = false;
> + } else if ((gsi != gicc->spe_interrupt) ||
> + (hetid != find_acpi_cpu_topology_hetero_id(cpu))) {
Sorry, I should have noticed this in patch 2 itself. Won't this break for
multi-socket system ? The hetid in that case will be package id, no ?
Otherwise the patch looks good.
--
Regards,
Sudeep
