Re: [PATCH v2 1/2] KVM: X86: Implement PV IPI in linux guest

[Wanpeng Li]

On Mon, 2 Jul 2018 at 19:32, Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote:
>
> On 02/07/2018 12:14, Wanpeng Li wrote:
> > +     unsigned long flags, ipi_bitmap_low = 0, ipi_bitmap_high = 0, icr = 0;
> > +     int cpu, apic_id, ret = 1;
> > +
> > +     if (cpumask_empty(mask))
> > +             return true;
> > +
> > +     local_irq_save(flags);
> > +
> > +     for_each_cpu(cpu, mask) {
> > +             apic_id = per_cpu(x86_cpu_to_apicid, cpu);
> > +             if (apic_id < BITS_PER_LONG)
> > +                     __set_bit(apic_id, &ipi_bitmap_low);
> > +             else if (apic_id < 2 * BITS_PER_LONG)
> > +                     __set_bit(apic_id - BITS_PER_LONG, &ipi_bitmap_high);
> > +             else
> > +                     goto ipi_mask_done;
> > +     }
>
> CPU masks are themselves bitmaps made of longs, so you should be able to
> avoid the loop here.

As we discuss offline, loop is needed since I need to convert
processor id which is allocated by OS to apic id.

>
> > +static void kvm_send_ipi_mask(const struct cpumask *mask, int vector)
> > +{
> > +     if (!__send_ipi_mask(mask, vector))
> > +             orig_apic.send_IPI_mask(mask, vector);
> > +}
> > +
> > +static void kvm_send_ipi_mask_allbutself(const struct cpumask *mask, int vector)
> > +{
> > +     unsigned int this_cpu = smp_processor_id();
> > +     struct cpumask new_mask;
> > +     const struct cpumask *local_mask;
> > +
> > +     cpumask_copy(&new_mask, mask);
> > +     cpumask_clear_cpu(this_cpu, &new_mask);
> > +     local_mask = &new_mask;
> > +     if (!__send_ipi_mask(local_mask, vector))
> > +             orig_apic.send_IPI_mask_allbutself(mask, vector);
> > +}
>
> Likewise, here it should be possible to check the highest bit in the
> mask before copying it.

__send_ipi_mask() has already handled sparse APIC ID and > 128 APID ID
scenarios.

Regards,
Wanpeng Li