Re: [PATCH] memcg: skip cgroup_file_notify if spinning is not allowed

[Michal Hocko <mhocko@suse.com>]

On Mon 08-09-25 10:11:29, Alexei Starovoitov wrote:
> On Mon, Sep 8, 2025 at 2:08 AM Michal Hocko <mhocko@suse.com> wrote:
> >
> > On Fri 05-09-25 20:48:46, Peilin Ye wrote:
> > > On Fri, Sep 05, 2025 at 01:16:06PM -0700, Shakeel Butt wrote:
> > > > Generally memcg charging is allowed from all the contexts including NMI
> > > > where even spinning on spinlock can cause locking issues. However one
> > > > call chain was missed during the addition of memcg charging from any
> > > > context support. That is try_charge_memcg() -> memcg_memory_event() ->
> > > > cgroup_file_notify().
> > > >
> > > > The possible function call tree under cgroup_file_notify() can acquire
> > > > many different spin locks in spinning mode. Some of them are
> > > > cgroup_file_kn_lock, kernfs_notify_lock, pool_workqeue's lock. So, let's
> > > > just skip cgroup_file_notify() from memcg charging if the context does
> > > > not allow spinning.
> > > >
> > > > Signed-off-by: Shakeel Butt <shakeel.butt@linux.dev>
> > >
> > > Tested-by: Peilin Ye <yepeilin@google.com>
> > >
> > > The repro described in [1] no longer triggers locking issues after
> > > applying this patch and making __bpf_async_init() use __GFP_HIGH
> > > instead of GFP_ATOMIC:
> > >
> > > --- a/kernel/bpf/helpers.c
> > > +++ b/kernel/bpf/helpers.c
> > > @@ -1275,7 +1275,7 @@ static int __bpf_async_init(struct bpf_async_kern *async, struct bpf_map *map, u
> > >         }
> > >
> > >         /* allocate hrtimer via map_kmalloc to use memcg accounting */
> > > -       cb = bpf_map_kmalloc_node(map, size, GFP_ATOMIC, map->numa_node);
> > > +       cb = bpf_map_kmalloc_node(map, size, __GFP_HIGH, map->numa_node);
> >
> > Why do you need to consume memory reserves? Shouldn't kmalloc_nolock be
> > used instead here?
> 
> Yes. That's a plan. We'll convert most of bpf allocations to kmalloc_nolock()
> when it lands.

OK, I thought this was merged already. I suspect __GFP_HIGH is used here
as a result of manual GFP_ATOMIC & ~GFP_RECLAIM. A TODO/FIXME referring
to kmalloc_nolock would clarify the situation.

-- 
Michal Hocko
SUSE Labs