Re: [External] Re: [PATCH] mm: memcontrol: remove rcu_read_lock from get_mem_cgroup_from_page

[Michal Hocko]

On Fri 05-02-21 13:15:40, Johannes Weiner wrote:
> On Fri, Feb 05, 2021 at 11:32:24AM +0100, Michal Hocko wrote:
> > On Fri 05-02-21 17:14:30, Muchun Song wrote:
> > > On Fri, Feb 5, 2021 at 4:36 PM Michal Hocko <mhocko@xxxxxxxx> wrote:
> > > >
> > > > On Fri 05-02-21 14:27:19, Muchun Song wrote:
> > > > > The get_mem_cgroup_from_page() is called under page lock, so the page
> > > > > memcg cannot be changed under us.
> > > >
> > > > Where is the page lock enforced?
> > > 
> > > Because it is called from alloc_page_buffers(). This path is under
> > > page lock.
> > 
> > I do not see any page lock enforecement there. There is not even a
> > comment requiring that. Can we grow more users where this is not the
> > case? There is no actual relation between alloc_page_buffers and
> > get_mem_cgroup_from_page except that the former is the only _current_
> > existing user. I would be careful to dictate locking based solely on
> > that.
> 
> Since alloc_page_buffers() holds the page lock throughout the entire
> time it uses the memcg, there is no actual reason for it to use RCU or
> even acquire an additional reference on the css. We know it's pinned,
> the charge pins it, and the page lock pins the charge. It can neither
> move to a different cgroup nor be uncharged.
> 
> So what do you say we switch alloc_page_buffers() to page_memcg()?
> 
> And because that removes the last user of get_mem_cgroup_from_page(),
> we can kill it off and worry about a good interface once a consumer
> materializes for it.

Yes, this makes much more sense than impose a weird locking rules to a
more general purpose helper. Thanks!
-- 
Michal Hocko
SUSE Labs