Benjamin Herrenschmidt wrote:

On Mon, 2011-06-27 at 11:00 +0200, Sebastian Andrzej Siewior wrote:

quoted

* Grant Likely | 2011-05-26 00:54:38 [-0600]:

quoted

diff --git a/arch/x86/kernel/devicetree.c b/arch/x86/kernel/devicetree.c
index 39645b6..9891cd4 100644
--- a/arch/x86/kernel/devicetree.c
+++ b/arch/x86/kernel/devicetree.c

@@ -371,36 +321,49 @@ static struct of_ioapic_type of_ioapic_type[] =

	},
};

-static int ioapic_xlate(struct irq_domain *id, const u32 *intspec, u32 intsize,
-			u32 *out_hwirq, u32 *out_type)
+static int ioapic_dt_translate(struct irq_domain *domain,
+				struct device_node *controller,
+				const u32 *intspec, u32 intsize,
+				irq_hw_number_t *out_hwirq, u32 *out_type)
{
-	struct mp_ioapic_gsi *gsi_cfg;
	struct io_apic_irq_attr attr;
	struct of_ioapic_type *it;
	u32 line, idx, type;
+	int rc;

-	if (intsize < 2)
+	if (controller != domain->of_node)
		return -EINVAL;

Is there a reason not havining the (controller != domain->of_node) check
in irq_create_of_mapping()?

Not all domains are associated with a single OF node.

Take xics, where there can be quite a few "source controllers" which act
as device-tree interrupt parents but there's a single global domain.

I see. No, actually I don't. xics is pseries where I don't see the .dts. 
So you are saying that we have one irq_domain but 2+ different
interrupt-parents nodes?

How do you distinguish then between two different controllers lets say
xics and a gpio based controller? This implementation calls ->dt_translate
until one controller returns 0 which looks like brute force.

xics_host_xlate() returns always zero so you would have to go for
the compatible and check it.
Every device has an interrupt-parent node. Shouldn't the code call exact 
this irq controller xlate function instead of trying them all?

 
Cheers,
Ben.

Sebastian