On Tue, Sep 18, 2018 at 05:50:21AM -0700, Tejun Heo wrote: > On Tue, Sep 18, 2018 at 06:19:44PM +0800, Ming Lei wrote: > > Now percpu_ref_reinit() can only be done on one percpu refcounter when it > > drops zero. And the limit shouldn't be so strict, and it is quite > > straightforward that we can do it when the refcount doesn't drop zero > > because it is at atomic mode. > > > > This patch introduces percpu_ref_resurge() in which the above limit is > > relaxed, so we may avoid extra change[1] for NVMe timeout's requirement. > > For now, > > Nacked-by: Tejun Heo <tj@xxxxxxxxxx> > > Please see the original discussion thread. Your comment in that thread supposes that synchronize_rcu() is used for avoiding race with .release(). But this patchset doesn't use that approach at all for avoiding this race. Thanks, Ming