[PATCH] avoid timer interrupt replay effect when onlining cpu

[Joel Schopp]

> Signed-off-by: Nathan Lynch <ntl at pobox.com>
> 
> 
> --- powerpc-timer_interrupt-replay.orig/arch/powerpc/kernel/smp.c
> +++ powerpc-timer_interrupt-replay/arch/powerpc/kernel/smp.c
> @@ -540,6 +540,9 @@ int __devinit start_secondary(void *unus
>  	if (smp_ops->take_timebase)
>  		smp_ops->take_timebase();
>  
> +	if (system_state > SYSTEM_BOOTING)
> +		per_cpu(last_jiffy, cpu) = get_tb();
> +
>  	spin_lock(&call_lock);
>  	cpu_set(cpu, cpu_online_map);
>  	spin_unlock(&call_lock);
> _______________________________________________

Yep, this bug has been seen in SUSE & Redhat distro kernels and this 
patch fixes it.

While we are here, is there any reason we still have 
next_jiffy_update_tb in the paca?  It isn't used anywhere anymore.

Acked-by: Joel Schopp <jschopp at austin.ibm.com>