Re: [RFC PATCH 1/4] genirq: Provide basic NMI management for interrupt lines

[Julien Thierry <julien.thierry@arm.com>]

On 17/07/18 12:13, Julien Thierry wrote:
> Hi Thomas,
> 
> Thanks for the review.
> 
> On 17/07/18 10:22, Thomas Gleixner wrote:

[...]

>> On Mon, 9 Jul 2018, Julien Thierry wrote:
>>> +/* This function must be called with desc->lock held */
>>> +static const void *__cleanup_nmi(unsigned int irq, struct irq_desc 
>>> *desc)
>>> +{
>>> +    const char *devname = NULL;
>>> +
>>> +    desc->istate &= ~IRQS_NMI;
>>> +
>>> +#ifdef CONFIG_SMP
>>> +    if (WARN_ON(desc->affinity_notify))
>>> +        desc->affinity_notify = NULL;
>>> +#endif
>>
>> We should not support that affinity notify horror for NMIs at all.
>>
> 
> Makes sense, I'll add it to the constraints for NMIs and remove that part.
> 
>>> +    if (!WARN_ON(desc->action == NULL)) {
>>> +        irq_pm_remove_action(desc, desc->action);
>>> +        devname = desc->action->name;
>>> +        unregister_handler_proc(irq, desc->action);
>>> +
>>> +        kfree(desc->action);
>>> +        desc->action = NULL;
>>> +    }
>>> +
>>> +    irq_settings_clr_disable_unlazy(desc);
>>> +    irq_shutdown(desc);
>>> +
>>> +#ifdef CONFIG_SMP
>>> +    /* make sure affinity_hint is cleaned up */
>>> +    if (WARN_ON_ONCE(desc->affinity_hint))
>>> +        desc->affinity_hint = NULL;
>>> +#endif
>>
>> Especially not twice :)
>>
> 
> Oops... I'll remove that as well.

Re-reading this I realized that second part was about 
desc->affinity_hint and not desc->affinity_notify.

Supporting desc->affinity_hint for NMIs shouldn't be an issue, right?

Thanks,

-- 
Julien Thierry