[PATCH v2] ARM: at91: pit add DT support

On Thu, Jan 05, 2012 at 03:50:09PM +0100, Nicolas Ferre wrote:

On 01/05/2012 01:00 PM, Jamie Iles :

quoted

Hi Nicolas, Jean-Christophe,

On Thu, Jan 05, 2012 at 02:48:28PM +0100, Nicolas Ferre wrote:

quoted

From: Jean-Christophe PLAGNIOL-VILLARD <redacted>

Retreive registers address and IRQ from device tree entry. Fall back
to built-in values if an error occurs.

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <redacted>
[nicolas.ferre at atmel.com: change error path and interrupts property handling]
Signed-off-by: Nicolas Ferre <redacted>
---

[...]

quoted

diff --git a/arch/arm/mach-at91/at91sam926x_time.c 

b/arch/arm/mach-at91/at91sam926x_time.c
index d89ead7..5d4c308 100644

--- a/arch/arm/mach-at91/at91sam926x_time.c
+++ b/arch/arm/mach-at91/at91sam926x_time.c

@@ -14,6 +14,8 @@

[...]

quoted

+static int __init of_at91sam926x_pit_init(void)
+{
+	struct device_node *np;
+
+	np = of_find_matching_node(NULL, timer_ids);
+	if (!np)
+		goto err;
+
+	pit_base_addr = of_iomap(np, 0);
+	if (!pit_base_addr)
+		goto node_err;
+
+	/* Get the interrupts property */
+	if (of_property_read_u32(np, "interrupts", &at91sam926x_pit_irq.irq))
+		goto ioremap_err;

I think you want:

	at91sam926x_pit_irq.irq = irq_of_parse_and_map(np, 0);

to make sure the IRQ is converted to a Linux IRQ and not a hwirq.

Yes, that definitively makes sense. BTW, without a DT entry, how should
I map a hwirq, with the irq_domain_to_irq() helper? But I will need
access to the interrupt controller "struct irq_domain"...

If you aren't booting with DT then at the moment I think on AT91 you 
have statically assigned irq_descs and IRQ numbers so you don't need to 
map them, you just know what IRQ number to request at compile time.

Jamie