Re: slab: start_cpu_timer/cache_reap CONFIG_HOTPLUG_CPU problems

[Oleg Nesterov <oleg@tv-sign.ru>]

On 01/29, Christoph Lameter wrote:
>
> On Tue, 30 Jan 2007, Oleg Nesterov wrote:
> 
> > Now we have 2 additional events, CPU_LOCK_ACQUIRE/CPU_LOCK_RELEASE,
> > so cpuup_callback() can use them to lock/unlock cache_chain_mutex,
> > but this is not related.
> 
> Then we wont need to do the mutex_lock/unlock in CPU_DOWN_XX 
> anymore, right? Which brings us to this form of the patch:

Yes, if CPU_LOCK_ACQUIRE takes cache_chain_mutex, we don't need to do so
on CPU_DOWN_PREPARE. I didn't know cpuup_callback() was already converted,
sorry for the confusion!

Note: with the patch below we are doing cancel_rearming_delayed_work()
under cache_chain_mutex. This is ok since cache_reap() does mutex_trylock(),
so deadlock is not possible. However, this means that mutex_trylock() becomes
mandatory for cache_reap(), probably a little comment will be good.

> Shutdown cache_reaper when cpu goes down
> 
> Shutdown the cache_reaper in slab.c if the cpu is brought down
> and set the cache_reap.func to NULL. Otherwise hotplug shuts
> down the reaper for good.
> 
> Signed-off-by: Christoph Lameter <clameter@sgi.com>
> 
> Index: linux-2.6.20-rc6-mm2/mm/slab.c
> ===================================================================
> --- linux-2.6.20-rc6-mm2.orig/mm/slab.c	2007-01-29 14:27:34.199229828 -0600
> +++ linux-2.6.20-rc6-mm2/mm/slab.c	2007-01-29 15:47:18.293962726 -0600
> @@ -1271,6 +1271,14 @@ static int __cpuinit cpuup_callback(stru
>  		start_cpu_timer(cpu);
>  		break;
>  #ifdef CONFIG_HOTPLUG_CPU
> +  	case CPU_DOWN_PREPARE:
> +		/* Shutdown cache reaper */
> +		cancel_rearming_delayed_work(&per_cpu(reap_work, cpu));
> +		per_cpu(reap_work, cpu).work.func = NULL;
> +  		break;
> +  	case CPU_DOWN_FAILED:
> +		start_cpu_timer(cpu);
> +  		break;
>  	case CPU_DEAD:
>  		/*
>  		 * Even if all the cpus of a node are down, we don't free the

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/