RFC: amba: remove NO_IRQ

[Alessandro Rubini <rubini@gnudd.com>]

I'm using PrimeCell devices under PCIe, x86 platform. We discussed
this issue on Mar 1st, and you (Russell) reported that NO_IRQ is
expected to go away for ARM, as 0 is the new official "no irq" marker.

The trivial patch below is what I'm currently using to compile my
stuff, but I'd like to get an update about NO_IRQ in ARM (which you
complained was completely on your shoulders). If needed I can take it
over and propose patches.

I'm well aware that the few users within arch/arm are matched by a
much bigger number of users in drivers/ .  The work is not trivial,
but at least I hope to be able to remove NO_IRQ from amba/bus.c by
auditing all its users.

thanks
/alessandro
---
 drivers/amba/bus.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/amba/bus.c b/drivers/amba/bus.c
index 64ddedd..490a308 100644
--- a/drivers/amba/bus.c
+++ b/drivers/amba/bus.c
@@ -546,10 +546,10 @@ int amba_device_add(struct amba_device *dev, struct resource *parent)
 		goto err_release;
 
 	printk(KERN_INFO "%s %i\n", __func__, __LINE__);
-	if (dev->irq[0] && dev->irq[0] != NO_IRQ)
+	if (dev->irq[0])
 		ret = device_create_file(&dev->dev, &dev_attr_irq0);
 	printk(KERN_INFO "%s %i\n", __func__, __LINE__);
-	if (ret == 0 && dev->irq[1] && dev->irq[1] != NO_IRQ)
+	if (ret == 0 && dev->irq[1])
 		ret = device_create_file(&dev->dev, &dev_attr_irq1);
 	printk(KERN_INFO "%s %i\n", __func__, __LINE__);
 	if (ret == 0)
-- 
1.7.7.2