Wu, Feng wrote on 2014-12-19:
>
>
> Zhang, Yang Z wrote on 2014-12-18:
>> jiang.liu@xxxxxxxxxxxxxxx
>> Cc: eric.auger@xxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
>> iommu@xxxxxxxxxxxxxxxxxxxxxxxxxx; kvm@xxxxxxxxxxxxxxx; Wu, Feng
>> Subject: RE: [v3 06/26] iommu, x86: No need to migrating irq for
>> VT-d Posted-Interrupts
>>
>> Feng Wu wrote on 2014-12-12:
>>> We don't need to migrate the irqs for VT-d Posted-Interrupts here.
>>> When 'pst' is set in IRTE, the associated irq will be posted to
>>> guests instead of interrupt remapping. The destination of the
>>> interrupt is set in Posted-Interrupts Descriptor, and the
>>> migration happens during vCPU scheduling.
>>>
>>> However, we still update the cached irte here, which can be used
>>> when changing back to remapping mode.
>>>
>>> Signed-off-by: Feng Wu <feng.wu@xxxxxxxxx>
>>> Reviewed-by: Jiang Liu <jiang.liu@xxxxxxxxxxxxxxx>
>>> ---
>>> drivers/iommu/intel_irq_remapping.c | 6 +++++-
>>> 1 file changed, 5 insertions(+), 1 deletion(-) diff --git
>>> a/drivers/iommu/intel_irq_remapping.c
>>> b/drivers/iommu/intel_irq_remapping.c index 48c2051..ab9057a
>>> 100644
>>> --- a/drivers/iommu/intel_irq_remapping.c +++
>>> b/drivers/iommu/intel_irq_remapping.c @@ -977,6 +977,7 @@
>>> intel_ir_set_affinity(struct irq_data *data, const struct cpumask
>>> *mask, {
>>> struct intel_ir_data *ir_data = data->chip_data; struct irte *irte =
>>> &ir_data->irte_entry; + struct irte_pi *irte_pi = (struct irte_pi
>>> *)irte; struct irq_cfg *cfg = irqd_cfg(data); struct irq_data *parent
>>> = data->parent_data; int ret;
>>> @@ -991,7 +992,10 @@ intel_ir_set_affinity(struct irq_data *data,
>>> const struct cpumask *mask,
>>> */
>>> irte->vector = cfg->vector;
>>> irte->dest_id = IRTE_DEST(cfg->dest_apicid);
>>> - modify_irte(&ir_data->irq_2_iommu, irte);
>>> +
>>> + /* We don't need to modify irte if the interrupt is for posting. */
>>> + if (irte_pi->pst != 1)
>>> + modify_irte(&ir_data->irq_2_iommu, irte);
>>
>> What happens if user changes the IRQ affinity manually?
>
> If the IRQ is posted, its affinity is controlled by guest (irq <--->
> vCPU <----> pCPU), it has no effect when host changes its affinity.
That's the problem: User is able to changes it in host but it never takes effect since it is actually controlled by guest. I guess it will break the IRQ balance too.
>
> Thanks,
> Feng
>
>>
>>>
>>> /*
>>> * After this point, all the interrupts will start arriving
>>
>>
>> Best regards,
>> Yang
>>
Best regards,
Yang
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
