Use real time pcp locking for page draining during cpu unplug Looks like a merging mistake that happened at some point. This is the only place in the file that disables interrupts directly. This fixes one case of CPU hotunplug failing on RT, but there are still more. Signed-off-by: Andi Kleen <ak@xxxxxxx> Index: linux-2.6.24-rc2-rt1-hack/mm/page_alloc.c =================================================================== --- linux-2.6.24-rc2-rt1-hack.orig/mm/page_alloc.c +++ linux-2.6.24-rc2-rt1-hack/mm/page_alloc.c @@ -4049,10 +4049,11 @@ static int page_alloc_cpu_notify(struct int cpu = (unsigned long)hcpu; if (action == CPU_DEAD || action == CPU_DEAD_FROZEN) { - local_irq_disable(); + unsigned long flags; + __lock_cpu_pcp(&flags, cpu); __drain_pages(cpu); vm_events_fold_cpu(cpu); - local_irq_enable(); + unlock_cpu_pcp(flags, cpu); refresh_cpu_vm_stats(cpu); } return NOTIFY_OK; - To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html