[RFC PATCH v1 4/4] arm/arm64: fix a migrating irq bug when hotplug cpu

[jiang.liu@linux.intel.com (Jiang Liu)]

On 2015/9/6 12:23, Yang Yingliang wrote:
> When cpu is disabled, all irqs will be migratged to another cpu.
> In some cases, a new affinity is different, it needed to be coppied
> to irq's affinity. But if the type of irq is LPI, it's affinity will
> not be coppied because of irq_set_affinity's return value. Fix it by
> using irq_do_set_affinity.
> 
> And migrating interrupts is a core code matter, so move the code to
> kernel/irq/migration.c and select CONFIG_GENERIC_IRQ_MIGRATION when
> CONFIG_HOTPLUG_CPU and CONFIG_SMP is enabled.
> 
> Cc: Jiang Liu <jiang.liu@linux.intel.com>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Marc Zyngier <marc.zyngier@arm.com>
> Cc: Mark Rutland <mark.rutland@arm.com>
> Cc: Will Deacon <will.deacon@arm.com>
> Cc: Russell King - ARM Linux <linux@arm.linux.org.uk>
> Cc: Hanjun Guo <hanjun.guo@linaro.org>
> Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
> ---
>  arch/arm/Kconfig             |  1 +
>  arch/arm/include/asm/irq.h   |  1 -
>  arch/arm/kernel/irq.c        | 62 --------------------------------------------
>  arch/arm64/Kconfig           |  1 +
>  arch/arm64/include/asm/irq.h |  1 -
>  arch/arm64/kernel/irq.c      | 62 --------------------------------------------
>  kernel/irq/migration.c       | 62 ++++++++++++++++++++++++++++++++++++++++++++
>  7 files changed, 64 insertions(+), 126 deletions(-)
> 
> diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
> index 41cbb4a..ebc8a33 100644
> --- a/arch/arm/Kconfig
> +++ b/arch/arm/Kconfig
> @@ -22,6 +22,7 @@ config ARM
>  	select GENERIC_CLOCKEVENTS_BROADCAST if SMP
>  	select GENERIC_IDLE_POLL_SETUP
>  	select GENERIC_IRQ_PROBE
> +	select GENERIC_IRQ_MIGRATION if SMP && HOTPLUG_CPU
>  	select GENERIC_IRQ_SHOW
>  	select GENERIC_IRQ_SHOW_LEVEL
>  	select GENERIC_PCI_IOMAP
> diff --git a/arch/arm/include/asm/irq.h b/arch/arm/include/asm/irq.h
> index 53c15de..d17fc900 100644
> --- a/arch/arm/include/asm/irq.h
> +++ b/arch/arm/include/asm/irq.h
> @@ -24,7 +24,6 @@
>  #ifndef __ASSEMBLY__
>  struct irqaction;
>  struct pt_regs;
> -extern void migrate_irqs(void);
>  
>  extern void asm_do_IRQ(unsigned int, struct pt_regs *);
>  void handle_IRQ(unsigned int, struct pt_regs *);
> diff --git a/arch/arm/kernel/irq.c b/arch/arm/kernel/irq.c
> index baf8ede..2efdb40 100644
> --- a/arch/arm/kernel/irq.c
> +++ b/arch/arm/kernel/irq.c
> @@ -31,7 +31,6 @@
>  #include <linux/smp.h>
>  #include <linux/init.h>
>  #include <linux/seq_file.h>
> -#include <linux/ratelimit.h>
>  #include <linux/errno.h>
>  #include <linux/list.h>
>  #include <linux/kallsyms.h>
> @@ -135,64 +134,3 @@ int __init arch_probe_nr_irqs(void)
>  	return nr_irqs;
>  }
>  #endif
> -
> -#ifdef CONFIG_HOTPLUG_CPU
> -static bool migrate_one_irq(struct irq_desc *desc)
> -{
> -	struct irq_data *d = irq_desc_get_irq_data(desc);
> -	const struct cpumask *affinity = irq_data_get_affinity_mask(d);
> -	struct irq_chip *c;
> -	bool ret = false;
> -
> -	/*
> -	 * If this is a per-CPU interrupt, or the affinity does not
> -	 * include this CPU, then we have nothing to do.
> -	 */
> -	if (irqd_is_per_cpu(d) || !cpumask_test_cpu(smp_processor_id(), affinity))
> -		return false;
> -
> -	if (cpumask_any_and(affinity, cpu_online_mask) >= nr_cpu_ids) {
> -		affinity = cpu_online_mask;
> -		ret = true;
> -	}
> -
> -	c = irq_data_get_irq_chip(d);
> -	if (!c->irq_set_affinity)
> -		pr_debug("IRQ%u: unable to set affinity\n", d->irq);
> -	else if (c->irq_set_affinity(d, affinity, false) == IRQ_SET_MASK_OK && ret)
> -		cpumask_copy(irq_data_get_affinity_mask(d), affinity);
> -
> -	return ret;
> -}
> -
> -/*
> - * The current CPU has been marked offline.  Migrate IRQs off this CPU.
> - * If the affinity settings do not allow other CPUs, force them onto any
> - * available CPU.
> - *
> - * Note: we must iterate over all IRQs, whether they have an attached
> - * action structure or not, as we need to get chained interrupts too.
> - */
> -void migrate_irqs(void)
> -{
> -	unsigned int i;
> -	struct irq_desc *desc;
> -	unsigned long flags;
> -
> -	local_irq_save(flags);
> -
> -	for_each_irq_desc(i, desc) {
> -		bool affinity_broken;
> -
> -		raw_spin_lock(&desc->lock);
> -		affinity_broken = migrate_one_irq(desc);
> -		raw_spin_unlock(&desc->lock);
> -
> -		if (affinity_broken)
> -			pr_warn_ratelimited("IRQ%u no longer affine to CPU%u\n",
> -				i, smp_processor_id());
> -	}
> -
> -	local_irq_restore(flags);
> -}
> -#endif /* CONFIG_HOTPLUG_CPU */
> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> index b7b9cea..6ffe411 100644
> --- a/arch/arm64/Kconfig
> +++ b/arch/arm64/Kconfig
> @@ -32,6 +32,7 @@ config ARM64
>  	select GENERIC_CPU_AUTOPROBE
>  	select GENERIC_EARLY_IOREMAP
>  	select GENERIC_IRQ_PROBE
> +	select GENERIC_IRQ_MIGRATION if SMP && HOTPLUG_CPU
>  	select GENERIC_IRQ_SHOW
>  	select GENERIC_IRQ_SHOW_LEVEL
>  	select GENERIC_PCI_IOMAP
> diff --git a/arch/arm64/include/asm/irq.h b/arch/arm64/include/asm/irq.h
> index bbb251b..0916929 100644
> --- a/arch/arm64/include/asm/irq.h
> +++ b/arch/arm64/include/asm/irq.h
> @@ -7,7 +7,6 @@
>  
>  struct pt_regs;
>  
> -extern void migrate_irqs(void);
>  extern void set_handle_irq(void (*handle_irq)(struct pt_regs *));
>  
>  static inline void acpi_irq_init(void)
> diff --git a/arch/arm64/kernel/irq.c b/arch/arm64/kernel/irq.c
> index 463fa2e..04ac1f6 100644
> --- a/arch/arm64/kernel/irq.c
> +++ b/arch/arm64/kernel/irq.c
> @@ -27,7 +27,6 @@
>  #include <linux/init.h>
>  #include <linux/irqchip.h>
>  #include <linux/seq_file.h>
> -#include <linux/ratelimit.h>
>  
>  unsigned long irq_err_count;
>  
> @@ -56,64 +55,3 @@ void __init init_IRQ(void)
>  	if (!handle_arch_irq)
>  		panic("No interrupt controller found.");
>  }
> -
> -#ifdef CONFIG_HOTPLUG_CPU
> -static bool migrate_one_irq(struct irq_desc *desc)
> -{
> -	struct irq_data *d = irq_desc_get_irq_data(desc);
> -	const struct cpumask *affinity = irq_data_get_affinity_mask(d);
> -	struct irq_chip *c;
> -	bool ret = false;
> -
> -	/*
> -	 * If this is a per-CPU interrupt, or the affinity does not
> -	 * include this CPU, then we have nothing to do.
> -	 */
> -	if (irqd_is_per_cpu(d) || !cpumask_test_cpu(smp_processor_id(), affinity))
> -		return false;
> -
> -	if (cpumask_any_and(affinity, cpu_online_mask) >= nr_cpu_ids) {
> -		affinity = cpu_online_mask;
> -		ret = true;
> -	}
> -
> -	c = irq_data_get_irq_chip(d);
> -	if (!c->irq_set_affinity)
> -		pr_debug("IRQ%u: unable to set affinity\n", d->irq);
> -	else if (c->irq_set_affinity(d, affinity, false) == IRQ_SET_MASK_OK && ret)
> -		cpumask_copy(irq_data_get_affinity_mask(d), affinity);
> -
> -	return ret;
> -}
> -
> -/*
> - * The current CPU has been marked offline.  Migrate IRQs off this CPU.
> - * If the affinity settings do not allow other CPUs, force them onto any
> - * available CPU.
> - *
> - * Note: we must iterate over all IRQs, whether they have an attached
> - * action structure or not, as we need to get chained interrupts too.
> - */
> -void migrate_irqs(void)
> -{
> -	unsigned int i;
> -	struct irq_desc *desc;
> -	unsigned long flags;
> -
> -	local_irq_save(flags);
> -
> -	for_each_irq_desc(i, desc) {
> -		bool affinity_broken;
> -
> -		raw_spin_lock(&desc->lock);
> -		affinity_broken = migrate_one_irq(desc);
> -		raw_spin_unlock(&desc->lock);
> -
> -		if (affinity_broken)
> -			pr_warn_ratelimited("IRQ%u no longer affine to CPU%u\n",
> -					    i, smp_processor_id());
> -	}
> -
> -	local_irq_restore(flags);
> -}
> -#endif /* CONFIG_HOTPLUG_CPU */
> diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c
> index 37ddb7b..5801c79 100644
> --- a/kernel/irq/migration.c
> +++ b/kernel/irq/migration.c
> @@ -1,6 +1,7 @@
>  
>  #include <linux/irq.h>
>  #include <linux/interrupt.h>
> +#include <linux/ratelimit.h>
>  
>  #include "internals.h"
>  
> @@ -77,3 +78,64 @@ void irq_move_irq(struct irq_data *idata)
>  	if (!masked)
>  		idata->chip->irq_unmask(idata);
>  }
> +
> +#ifdef CONFIG_HOTPLUG_CPU
> +static bool migrate_one_irq(struct irq_desc *desc)
> +{
> +	struct irq_data *d = irq_desc_get_irq_data(desc);
> +	const struct cpumask *affinity = d->affinity;
> +	struct irq_chip *c;
> +	bool ret = false;
> +
> +	/*
> +	 * If this is a per-CPU interrupt, or the affinity does not
> +	 * include this CPU, then we have nothing to do.
> +	 */
> +	if (irqd_is_per_cpu(d) || !cpumask_test_cpu(smp_processor_id(), affinity))
> +		return false;
> +
> +	if (cpumask_any_and(affinity, cpu_online_mask) >= nr_cpu_ids) {
> +		affinity = cpu_online_mask;
> +		ret = true;
> +	}
> +
> +	c = irq_data_get_irq_chip(d);
> +	if (!c->irq_set_affinity)
> +		pr_debug("IRQ%u: unable to set affinity\n", d->irq);
How about pr_warn here? It may cause serious drawback if this happens.

> +	else
> +		irq_do_set_affinity(d, affinity, false);
Should we check return value here?

> +
> +	return ret;
> +}
> +
> +/*
> + * The current CPU has been marked offline.  Migrate IRQs off this CPU.
> + * If the affinity settings do not allow other CPUs, force them onto any
> + * available CPU.
> + *
> + * Note: we must iterate over all IRQs, whether they have an attached
> + * action structure or not, as we need to get chained interrupts too.
> + */
> +void migrate_irqs(void)
> +{
> +	unsigned int i;
> +	struct irq_desc *desc;
> +	unsigned long flags;
> +
> +	local_irq_save(flags);
> +
> +	for_each_irq_desc(i, desc) {
Should we use for_each_active_irq() here to iterate over active
irqs only?

> +		bool affinity_broken;
> +
> +		raw_spin_lock(&desc->lock);
> +		affinity_broken = migrate_one_irq(desc);
> +		raw_spin_unlock(&desc->lock);
> +
> +		if (affinity_broken)
> +			pr_warn_ratelimited("IRQ%u no longer affine to CPU%u\n",
> +					    i, smp_processor_id());
> +	}
> +
> +	local_irq_restore(flags);
> +}
> +#endif /* CONFIG_HOTPLUG_CPU */
>