[PATCH] Fix i8259 cascade IRQ

[David Woodhouse]

setup_irq() aborts immediately if there's no handler for the IRQ in
question. So i8259_init() should set up its handlers before trying to
set up the cascade on IRQ 2.

With this and the patch I sent a few days ago to fix initrd on ppc32, my
Pegasos now runs the arch/powerpc kernel.

Signed-off-by: David Woodhouse <dwmw2 at infradead.org>

diff --git a/arch/powerpc/sysdev/i8259.c b/arch/powerpc/sysdev/i8259.c
index 90bce6e..b7ac32f 100644
--- a/arch/powerpc/sysdev/i8259.c
+++ b/arch/powerpc/sysdev/i8259.c
@@ -207,6 +207,9 @@ void __init i8259_init(unsigned long int
 
 	spin_unlock_irqrestore(&i8259_lock, flags);
 
+	for (i = 0; i < NUM_ISA_INTERRUPTS; ++i)
+		irq_desc[offset + i].handler = &i8259_pic;
+
 	/* reserve our resources */
 	setup_irq(offset + 2, &i8259_irqaction);
 	request_resource(&ioport_resource, &pic1_iores);
@@ -216,6 +219,4 @@ void __init i8259_init(unsigned long int
 	if (intack_addr != 0)
 		pci_intack = ioremap(intack_addr, 1);
 
-	for (i = 0; i < NUM_ISA_INTERRUPTS; ++i)
-		irq_desc[offset + i].handler = &i8259_pic;
 }

-- 
dwmw2