[IA64] fix broken irq affinity
authorChen, Kenneth W <kenneth.w.chen@intel.com>
Tue, 16 May 2006 23:29:00 +0000 (16:29 -0700)
committerTony Luck <tony.luck@intel.com>
Wed, 17 May 2006 13:20:23 +0000 (06:20 -0700)
When CONFIG_PCI_MSI is set, move_irq() is an empty function, causing
grief when sys admin tries to bind interrupt to CPU.

Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
arch/ia64/kernel/iosapic.c

index 7956eb9058fcb873a65b5889d66f6eb6208e4c76..d58c1c5c903a9cfdae0981411ae53604fce020f1 100644 (file)
@@ -416,7 +416,7 @@ iosapic_end_level_irq (unsigned int irq)
        ia64_vector vec = irq_to_vector(irq);
        struct iosapic_rte_info *rte;
 
-       move_irq(irq);
+       move_native_irq(irq);
        list_for_each_entry(rte, &iosapic_intr_info[vec].rtes, rte_list)
                iosapic_eoi(rte->addr, vec);
 }
@@ -458,7 +458,7 @@ iosapic_ack_edge_irq (unsigned int irq)
 {
        irq_desc_t *idesc = irq_descp(irq);
 
-       move_irq(irq);
+       move_native_irq(irq);
        /*
         * Once we have recorded IRQ_PENDING already, we can mask the
         * interrupt for real. This prevents IRQ storms from unhandled