On Tue, Feb 20 2024 at 22:51, Anup Patel wrote:
> On Tue, Feb 20, 2024 at 7:05 PM Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:
>>
>> On Tue, Feb 20 2024 at 11:37, Anup Patel wrote:
>> > static bool imsic_init_dev_msi_info(struct device *dev,
>> > struct irq_domain *domain,
>> > struct irq_domain *real_parent,
>> > @@ -218,6 +241,7 @@ static bool imsic_init_dev_msi_info(struct device *dev,
>> >
>> > /* MSI parent domain specific settings */
>> > switch (real_parent->bus_token) {
>> > + case DOMAIN_BUS_PCI_MSI:
>>
>> case DOMAIN_BUS_PCI_DEVICE_MSIX:
>>
>> ?
>
> Actually, the DOMAIN_BUS_PCI_MSI is not needed because
> the real parent domain is always the IMSIC base irq_domain
> so DOMAIN_BUS_NEXUS is sufficient.
Indeed. Obviously I can't read.
> Better to just drop DOMAIN_BUS_PCI_MSI from this switch case ?
Yes. I actually would be a bug if that ends up as the real parent
domain.
Thanks,
tglx
