I also encountered the same issue. Could you tell me how to fix this ?
Best Regards 在 2014年2月16日星期日UTC+8上午3时29分01秒，Siarhei Siamashka写道： > > Based on reading the A20 user manual, it looks like the same > single PLE/PERFMU IRQ is supposed to be used for the CPU > performance counters as on Allwinner A10. This is already suspicious, > because normally each CPU core in a dual-core system should have > its own performance monitoring unit IRQ as explained in: > > http://thread.gmane.org/gmane.linux.ports.arm.kernel/181392/focus=181597 > Still assuming that this is the case and borrowing the code > from db8500 does not work. No interrupts are arriving at all. > > Is there anything special needed to make sure that the performance > monitoring unit interrupts are routed correctly on Allwinner A20? > > Now that Cortex-A7 and GIC are both ARM IP, one might assume that > they perhaps should be able to play nicely together without > involving anything that is Allwinner specific. > > Any ideas? The perf/oprofile tools are very useful for identifying > performance bottlenecks and optimizing software. They work nicely > on Allwinner A10. It would be great if Allwinner A20 could be > supported too. > --- > arch/arm/plat-sunxi/devices.c | 36 ++++++++++++++++++++++++++++++++---- > 1 file changed, 32 insertions(+), 4 deletions(-) > > diff --git a/arch/arm/plat-sunxi/devices.c b/arch/arm/plat-sunxi/devices.c > index fdddc56..f5ecff7 100644 > --- a/arch/arm/plat-sunxi/devices.c > +++ b/arch/arm/plat-sunxi/devices.c > @@ -113,7 +113,35 @@ struct platform_device sw_pdev_nand = > .dev = {} > }; > > -#ifndef CONFIG_ARCH_SUN7I > +#ifdef CONFIG_ARCH_SUN7I > +/* > + * The PMU IRQ lines of two cores are wired together into a single > interrupt. > + * Bounce the interrupt to the other core if it's not ours. > + * > + * This code is borrowed from db8500_pmu_handler as advised by: > + * > http://thread.gmane.org/gmane.linux.ports.arm.kernel/181392/focus=181597 > + */ > +static irqreturn_t sun7i_pmu_handler(int irq, void *dev, irq_handler_t > handler) > +{ > + irqreturn_t ret = handler(irq, dev); > + int other = !smp_processor_id(); > + > + if (ret == IRQ_NONE && cpu_online(other)) > + irq_set_affinity(irq, cpumask_of(other)); > + > + /* > + * We should be able to get away with the amount of IRQ_NONEs we > give, > + * while still having the spurious IRQ detection code kick in if > the > + * interrupt really starts hitting spuriously. > + */ > + return ret; > +} > + > +static struct arm_pmu_platdata sun7i_pmu_platdata = { > + .handle_irq = sun7i_pmu_handler, > +}; > +#endif > + > static struct resource sunxi_pmu_resources[] = { > { > .start = SW_INT_IRQNO_PLE_PFM, > @@ -127,8 +155,10 @@ struct platform_device sunxi_pmu_device = { > .id = ARM_PMU_DEVICE_CPU, > .resource = sunxi_pmu_resources, > .num_resources = ARRAY_SIZE(sunxi_pmu_resources), > -}; > +#ifdef CONFIG_ARCH_SUN7I > + .dev.platform_data = &sun7i_pmu_platdata, > #endif > +}; > > #if defined(CONFIG_MALI_DRM) || defined(CONFIG_MALI_DRM_MODULE) > static struct platform_device sunxi_device_mali_drm = { > @@ -143,9 +173,7 @@ static struct platform_device *sw_pdevs[] __initdata = > { > #endif > &sw_pdev_dmac, > &sw_pdev_nand, > -#ifndef CONFIG_ARCH_SUN7I > &sunxi_pmu_device, > -#endif > #if defined(CONFIG_MALI_DRM) || defined(CONFIG_MALI_DRM_MODULE) > &sunxi_device_mali_drm, > #endif > -- > 1.8.3.2 > > -- You received this message because you are subscribed to the Google Groups "linux-sunxi" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
