Re: [PATCH 1/3] stop_machine: kill __stop_machine()

[Suresh Siddha <suresh.b.siddha@intel.com>]

On Thu, 2011-06-16 at 11:36 -0700, Peter Zijlstra wrote:
> On Thu, 2011-06-16 at 20:28 +0200, Tejun Heo wrote:
> > Peter, I don't think it's that simple.  get_online_cpus() itself can't
> > create circular dependency by itself.  It allows recursing.  The chain
> > involves cpu_hotplug_begin() which returns with hotplug mutex held.
> 
> 
> Right, its like:
> 
>  mutex_lock(&a);
>  get_online_cpus();
> 
> vs
> 
>  cpu_hotplug_begin()
>  mutex_lock(&a);
> 
> 
> that will really deadlock

Its actually like:

get_online_cpus();
mutex_lock(&a);
get_online_cpus(); <--- recursive get_online_cpus()

vs

cpu_hotplug_begin()
mutex_lock(&a);

So it can't cause deadlock in practice.

thanks,
suresh