Re: [PATCH] x86: Add check for number of available vectors beforeCPU down [v2]

From: Chen, Gong
Date: Mon Dec 30 2013 - 03:03:35 EST


On Sat, Dec 28, 2013 at 12:10:38PM -0500, Prarit Bhargava wrote:
> Gong and Rui,
>
> After looking at this in detail I realized I made a mistake in my patch by
> including the check for the smp_affinity. Simply put, it shouldn't be there
> given Rui's explanation above.
>
> So I think the patch simply needs to do:
>
> this_count = 0;
> for (vector = FIRST_EXTERNAL_VECTOR; vector < NR_VECTORS; vector++) {
> irq = __this_cpu_read(vector_irq[vector]);
> if (irq >= 0) {
> desc = irq_to_desc(irq);
> data = irq_desc_get_irq_data(desc);
> affinity = data->affinity;
> if (irq_has_action(irq) && !irqd_is_per_cpu(data))
> this_count++;
> }
> }
>
> Can the two of you confirm the above is correct? It would be greatly appreciated.
>

No, I don't think it is correct. We still need to consider smp_affinity.

fixup_irqs
irq_set_affinity(native_ioapic_set_affinity)
__ioapic_set_affinity
assign_irq_vector
__assign_irq_vector
cpu_mask_to_apicid_and
/* now begin to set ioapic RET */

__assign_irq_vector(int irq, struct irq_cfg *cfg, const struct cpumask *mask)
{
...
apic->vector_allocation_domain(cpu, tmp_mask, mask);
...
for_each_cpu_and(new_cpu, tmp_mask, cpu_online_mask)
per_cpu(vector_irq, new_cpu)[vector] = irq;
cfg->vector = vector;
cpumask_copy(cfg->domain, tmp_mask);
...
}

On same vecotr on all related vector_irq, irq is set. So such kind of
irq should happen in multiple vector_irq. In cpu_mask_to_apicid_and(e.g.
x2apic_cpu_mask_to_apicid_and for cluster mode), apic is updated
depending on new mask. That's why I think this kind of interrupt
should be bypassed.

Attachment: signature.asc
Description: Digital signature