fix broken irq affinity
PATCH 1/1 fix broken irq affinity
It's been broken for a while ...
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 <[email protected]>
arch/ia64/kernel/iosapic.c renamed from arch/ia64/kernel/iosapic.c.orig +2 -2
@@ -385,7 +385,7 @@ 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); }
@@ -427,7 +427,7 @@ 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