Re: Regression: unable to boot after commit bd9240a18edf ("x86/apic: Add TSC_DEADLINE quirk due to errata") - Surface Pro 4 SKL

[Peter Zijlstra]

On Tue, Dec 19, 2017 at 05:01:55PM +0100, Peter Zijlstra wrote:
> On Tue, Dec 19, 2017 at 11:42:41PM +0800, Zhang Rui wrote:
> > On Tue, 2017-12-19 at 16:23 +0100, Peter Zijlstra wrote:
> 
> > > [    0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 2 dfl dfl)
> > > [    0.000000] ACPI: IRQ0 used by override.
> > > 
> > > So your ACPI table has an override for IRQ2 and routes it to IRQ0.
> 
> ^ this
> 
> > >  The HPET document says:
> > > 
> > >   If the ENABLE_CNF bit and the LEG_RT_CNF bit are both set, then the
> > >   interrupts will be routed as follows:
> > > 
> > >     Timer 0 will be routed to IRQ0 in Non-APIC or IRQ2 in the I/O APIC
> > 
> > But AFAICS, the HPET emulated timer interrupts goes to IRQ0
> 
> Right, so see that ACPI override, that routes I/O APIC IRQ2 to IRQ0, or
> it _should_.
> 
> Clearly something is messed up here.. but I've no idea what. That whole
> IRQ routing stuff is confusing.

Does this help?

diff --git a/arch/x86/kernel/time.c b/arch/x86/kernel/time.c
index 749d189f8cd4..45675072771c 100644
--- a/arch/x86/kernel/time.c
+++ b/arch/x86/kernel/time.c
@@ -69,8 +69,6 @@ static struct irqaction irq0  = {
 
 static void __init setup_default_timer_irq(void)
 {
-	if (!nr_legacy_irqs())
-		return;
 	setup_irq(0, &irq0);
 }