[Devel] [PATCH rh7 v2 07/10] mm/mem_cgroup_iter: Invalidate iterators only if needed

[Kirill Tkhai]

On 24.02.2021 13:48, Konstantin Khorenko wrote:
> Previously we invalidated all iterators up to the memcg root,
> but this is overkill: we can check if currently dying memcg is
> stored in iterator's 'last_visited' and invalidate only those
> unlucky iterators.
> 
> https://jira.sw.ru/browse/PSBM-123655
> 
> Signed-off-by: Konstantin Khorenko <khorenko at virtuozzo.com>
> ---
>  mm/memcontrol.c | 48 +++++++++++++++++++++++++++++++++++++++++-------
>  1 file changed, 41 insertions(+), 7 deletions(-)
> 
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index eec944f712b0..b6059f46049b 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -192,8 +192,8 @@ struct mem_cgroup_stat2_cpu {
>  
>  struct mem_cgroup_reclaim_iter {
>  	/*
> -	 * last scanned hierarchy member. Valid only if last_dead_count
> -	 * matches memcg->dead_count of the hierarchy root group.
> +	 * Last scanned hierarchy member.
> +	 * If stored memcg is destroyed, the field is wiped.
>  	 */
>  	struct mem_cgroup *last_visited;
>  	unsigned long last_dead_count;
> @@ -1570,19 +1570,54 @@ static void mem_cgroup_iter_invalidate(struct mem_cgroup *root,
>  {
>  	struct mem_cgroup_reclaim_iter *iter;
>  	struct mem_cgroup_per_zone *mz;
> +	struct mem_cgroup *pos;
>  	int zone, node, i;
>  
> +	/*
> +	 * When a group in the hierarchy below root is destroyed,
> +	 * the hierarchy iterator can no longer be trusted iif
> +	 * iter->last_visited contains the cgroup being destroyed.
> +	 * Check if we get this unlikely case and invalidate the iterator
> +	 * if so.
> +	 *
> +	 * Note, only "break-ed" iterators can store iter->last_visited
> +	 * == dead_memcg because normally 'last_visited' is assigned
> +	 * in mem_cgroup_iter_update() and 'new_position' is just after
> +	 * css_tryget() there (ref inc-ed in __mem_cgroup_iter_next())
> +	 * and thus cgroup is not offlined yet.
> +	 *
> +	 * mem_cgroup_iter_break() in its turn puts memcg's css but does
> +	 * not wipe it from iter->last_visited.
> +	 *
> +	 * Q: Why dying memcg (dead_memcg) cannot get into
> +	 *    iter->last_visited a bit later after we wipe it here?
> +	 * A: Because up to the moment of the current function execution
> +	 *    css_tryget() is guaranteed to fail on 'dead_memcg'.
> +	 */
> +
> +	rcu_read_lock();

Looking this again, I think we do not need rcu_read_lock() here, so it does
not protect anything.

>  	for_each_node(node) {
>  		for (zone = 0; zone < MAX_NR_ZONES; zone++) {
>  			mz = mem_cgroup_zoneinfo(root, node, zone);
>  
>  			for (i = 0; i < ARRAY_SIZE(mz->reclaim_iter); i++) {
>  				iter = &mz->reclaim_iter[i];
> -				rcu_assign_pointer(iter->last_visited, NULL);
> +
> +				pos = rcu_access_pointer(iter->last_visited);
> +				/*
> +				 * It's OK to race with mem_cgroup_iter_update()
> +				 * here because it cannot write new
> +				 * position == dead_memcg as
> +				 * css_tryget() for it should fail already.
> +				 */
> +				if (pos == dead_memcg) {
> +					rcu_assign_pointer(iter->last_visited,
> +							   NULL);
>  				}
>  			}
>  		}
>  	}
> +	rcu_read_unlock();
>  
>  	/*
>  	 * When a group in the hierarchy below root is destroyed, the
> @@ -1631,10 +1666,9 @@ static void mem_cgroup_iter_update(struct mem_cgroup_reclaim_iter *iter,
>  				   int sequence)
>  {
>  	/*
> -	 * We store the sequence count from the time @last_visited was
> -	 * loaded successfully instead of rereading it here so that we
> -	 * don't lose destruction events in between.  We could have
> -	 * raced with the destruction of @new_position after all.
> +	 * The position saved in 'last_visited' is always valid.
> +	 * If the stored corresponding cgroup is destroyed,
> +	 * 'last_visited' is NULLed.
>  	 */
>  	rcu_assign_pointer(iter->last_visited, new_position);
>  	iter->last_dead_count = sequence;
>