[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH 2/2] xen: implement IRQ_WORK_VECTOR handler



On Tue, Apr 17, 2012 at 4:32 AM, Konrad Rzeszutek Wilk
<konrad.wilk@xxxxxxxxxx> wrote:
> On Sun, Apr 15, 2012 at 02:09:32PM +0800, Lin Ming wrote:
>> Signed-off-by: Lin Ming <mlin@xxxxxxxxxxxxx>
>> ---
>>  arch/x86/include/asm/xen/events.h |    1 +
>>  arch/x86/xen/smp.c                |   30 ++++++++++++++++++++++++++++++
>>  2 files changed, 31 insertions(+), 0 deletions(-)
>>
>> diff --git a/arch/x86/include/asm/xen/events.h 
>> b/arch/x86/include/asm/xen/events.h
>> index 1df3541..cc146d5 100644
>> --- a/arch/x86/include/asm/xen/events.h
>> +++ b/arch/x86/include/asm/xen/events.h
>> @@ -6,6 +6,7 @@ enum ipi_vector {
>>       XEN_CALL_FUNCTION_VECTOR,
>>       XEN_CALL_FUNCTION_SINGLE_VECTOR,
>>       XEN_SPIN_UNLOCK_VECTOR,
>> +     XEN_IRQ_WORK_VECTOR,
>>
>>       XEN_NR_IPIS,
>>  };
>> diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c
>> index 2dc6628..92ad12d 100644
>> --- a/arch/x86/xen/smp.c
>> +++ b/arch/x86/xen/smp.c
>> @@ -16,6 +16,7 @@
>>  #include <linux/err.h>
>>  #include <linux/slab.h>
>>  #include <linux/smp.h>
>> +#include <linux/irq_work.h>
>>
>>  #include <asm/paravirt.h>
>>  #include <asm/desc.h>
>> @@ -41,10 +42,12 @@ cpumask_var_t xen_cpu_initialized_map;
>>  static DEFINE_PER_CPU(int, xen_resched_irq);
>>  static DEFINE_PER_CPU(int, xen_callfunc_irq);
>>  static DEFINE_PER_CPU(int, xen_callfuncsingle_irq);
>> +static DEFINE_PER_CPU(int, xen_irq_work);
>>  static DEFINE_PER_CPU(int, xen_debug_irq) = -1;
>>
>>  static irqreturn_t xen_call_function_interrupt(int irq, void *dev_id);
>>  static irqreturn_t xen_call_function_single_interrupt(int irq, void 
>> *dev_id);
>> +static irqreturn_t xen_irq_work_interrupt(int irq, void *dev_id);
>>
>>  /*
>>   * Reschedule call back.
>> @@ -143,6 +146,17 @@ static int xen_smp_intr_init(unsigned int cpu)
>>               goto fail;
>>       per_cpu(xen_callfuncsingle_irq, cpu) = rc;
>>
>> +     callfunc_name = kasprintf(GFP_KERNEL, "irqwork%d", cpu);
>> +     rc = bind_ipi_to_irqhandler(XEN_IRQ_WORK_VECTOR,
>> +                                 cpu,
>> +                                 xen_irq_work_interrupt,
>> +                                 IRQF_DISABLED|IRQF_PERCPU|IRQF_NOBALANCING,
>> +                                 callfunc_name,
>> +                                 NULL);
>> +     if (rc < 0)
>> +             goto fail;
>> +     per_cpu(xen_irq_work, cpu) = rc;
>> +
>>       return 0;
>>
>>   fail:
>> @@ -155,6 +169,8 @@ static int xen_smp_intr_init(unsigned int cpu)
>>       if (per_cpu(xen_callfuncsingle_irq, cpu) >= 0)
>>               unbind_from_irqhandler(per_cpu(xen_callfuncsingle_irq, cpu),
>>                                      NULL);
>> +     if (per_cpu(xen_irq_work, cpu) >= 0)
>> +             unbind_from_irqhandler(per_cpu(xen_irq_work, cpu), NULL);
>>
>>       return rc;
>>  }
>> @@ -509,6 +525,9 @@ static inline int xen_map_vector(int vector)
>>       case CALL_FUNCTION_SINGLE_VECTOR:
>>               xen_vector = XEN_CALL_FUNCTION_SINGLE_VECTOR;
>>               break;
>> +     case IRQ_WORK_VECTOR:
>> +             xen_vector = XEN_IRQ_WORK_VECTOR;
>> +             break;
>>       default:
>>               xen_vector = -1;
>>               printk(KERN_ERR "xen: vector 0x%x is not implemented\n",
>> @@ -588,6 +607,16 @@ static irqreturn_t 
>> xen_call_function_single_interrupt(int irq, void *dev_id)
>>       return IRQ_HANDLED;
>>  }
>>
>> +static irqreturn_t xen_irq_work_interrupt(int irq, void *dev_id)
>> +{
>> +     irq_enter();
>> +     inc_irq_stat(apic_irq_work_irqs);
>> +     irq_work_run();
>
> I think this usually done the other way around:
>
> irq_work_run()
> inc_irq_stat(apic_irq_work_irqs)
>
> Or is there an excellent reason for doing it this way?

Copy & paste from smp_irq_work_interrupt().
But I think there is no much difference.

Anyway, I can change it if needed.

Thanks,
Lin Ming

>
>> +     irq_exit();
>> +
>> +     return IRQ_HANDLED;
>> +}
>> +
>>  static const struct smp_ops xen_smp_ops __initconst = {
>>       .smp_prepare_boot_cpu = xen_smp_prepare_boot_cpu,
>>       .smp_prepare_cpus = xen_smp_prepare_cpus,
>> @@ -634,6 +663,7 @@ static void xen_hvm_cpu_die(unsigned int cpu)
>>       unbind_from_irqhandler(per_cpu(xen_callfunc_irq, cpu), NULL);
>>       unbind_from_irqhandler(per_cpu(xen_debug_irq, cpu), NULL);
>>       unbind_from_irqhandler(per_cpu(xen_callfuncsingle_irq, cpu), NULL);
>> +     unbind_from_irqhandler(per_cpu(xen_irq_work, cpu), NULL);
>>       native_cpu_die(cpu);
>>  }
>>
>> --
>> 1.7.2.5

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.