Re: 2.6.22.1-rt4 lockups

[Ingo Molnar <mingo@elte.hu>]

* Daniel Walker <dwalker@mvista.com> wrote:

> It looks like sched_class->enqueue_task() is NULL and that's why the 
> system hangs ..
> 
> The reason why that happens is because check_pgt_cache() is called 
> from the idle thread, and with PREEMPT_RT check_pgt_cache() locks at 
> least one mutex .. Once the idle thread is on a wait_list, as soon as 
> it's woke by the mutex owner the system will crash in enqueue_task. 
> Since the idle thread has a NULL sched_class->enqueue_task ..
> 
> check_pgt_cache() is already getting called from the desched_thread() 
> , so I think it could just be removed from i386 cpu_idle().
> 
> Anyone have comments on the theory above?

yeah, that call definitely looks wrong in cpu_idle(). Most of the other 
check_pgd_cache() calls introduced by commit f1d1a842 look wrong too in 
an -rt context. Fix is below.

	Ingo

Index: linux-rt.q/arch/i386/kernel/process.c
===================================================================
--- linux-rt.q.orig/arch/i386/kernel/process.c
+++ linux-rt.q/arch/i386/kernel/process.c
@@ -189,7 +189,6 @@ void cpu_idle(void)
 
 			tick_nohz_stop_sched_tick();
 
-			check_pgt_cache();
 			rmb();
 			idle = pm_idle;