RE: [PATCH 10/10] clocksource/drivers/hyper-v: Move handling of STIMER0 interrupts

[Michael Kelley]

From: Wei Liu <wei.liu@xxxxxxxxxx> Sent: Monday, February 1, 2021 11:53 AM
> 
> On Wed, Jan 27, 2021 at 12:23:45PM -0800, Michael Kelley wrote:
> [...]
> > +static int hv_setup_stimer0_irq(void)
> > +{
> > +	int ret;
> > +
> > +	ret = acpi_register_gsi(NULL, HYPERV_STIMER0_VECTOR,
> > +			ACPI_EDGE_SENSITIVE, ACPI_ACTIVE_HIGH);
> 
> When IO-APIC is enabled on x86, acpi_register_gsi calls
> mp_map_gsi_to_irq. That function then calls mp_find_ioapic. Is
> HYPERV_STIMER0_VECTOR, when used as a GSI, associated with an IO-APIC?
> If not, wouldn't mp_find_ioapic return an error causing
> acpi_register_gsi to fail?
> 
> Ah, it appears that this function is not called on x86. I haven't
> checked how ARM handles GSI, but presumably this works for you.  It
> would be good if a comment can be added to clarify things.
> 
> Wei.

Will do.

Michael