Re: [PATCH] mm, list_lru: refactor the locking code

[SeongJae Park <sj@kernel.org>]

Hi Kairui,

On Tue, 27 May 2025 02:06:38 +0800 Kairui Song <ryncsn@gmail.com> wrote:

> From: Kairui Song <kasong@tencent.com>
> 
> Cocci is confused by the try lock then release RCU and return logic
> here. So separate the try lock part out into a standalone helper. The
> code is easier to follow too.
> 
> No feature change, fixes:
> 
> cocci warnings: (new ones prefixed by >>)
> >> mm/list_lru.c:82:3-9: preceding lock on line 77
> >> mm/list_lru.c:82:3-9: preceding lock on line 77
>    mm/list_lru.c:82:3-9: preceding lock on line 75
>    mm/list_lru.c:82:3-9: preceding lock on line 75
> 
> Reported-by: kernel test robot <lkp@intel.com>
> Reported-by: Julia Lawall <julia.lawall@inria.fr>
> Closes: https://lore.kernel.org/r/202505252043.pbT1tBHJ-lkp@intel.com/
> Signed-off-by: Kairui Song <kasong@tencent.com>

Reviewed-by: SeongJae Park <sj@kernel.org>

> ---
>  mm/list_lru.c | 34 +++++++++++++++++++---------------
>  1 file changed, 19 insertions(+), 15 deletions(-)
> 
> diff --git a/mm/list_lru.c b/mm/list_lru.c
> index 490473af3122..ec48b5dadf51 100644
> --- a/mm/list_lru.c
> +++ b/mm/list_lru.c
> @@ -60,30 +60,34 @@ list_lru_from_memcg_idx(struct list_lru *lru, int nid, int idx)
>  	return &lru->node[nid].lru;
>  }
>  
> +static inline bool lock_list_lru(struct list_lru_one *l, bool irq)
> +{
> +	if (irq)
> +		spin_lock_irq(&l->lock);
> +	else
> +		spin_lock(&l->lock);
> +	if (unlikely(READ_ONCE(l->nr_items) == LONG_MIN)) {
> +		if (irq)
> +			spin_unlock_irq(&l->lock);
> +		else
> +			spin_unlock(&l->lock);
> +		return false;
> +	}

I'd prefer 'if (likely(...)) return true;' to reduce indentation and stop
wondering what goes to the likely case earlier.  But that's my personal
preferrence that shouldn't block this.

> +	return true;
> +}
> +
>  static inline struct list_lru_one *
>  lock_list_lru_of_memcg(struct list_lru *lru, int nid, struct mem_cgroup *memcg,
>  		       bool irq, bool skip_empty)
>  {
>  	struct list_lru_one *l;
> -	long nr_items;
>  
>  	rcu_read_lock();
>  again:
>  	l = list_lru_from_memcg_idx(lru, nid, memcg_kmem_id(memcg));
> -	if (likely(l)) {
> -		if (irq)
> -			spin_lock_irq(&l->lock);
> -		else
> -			spin_lock(&l->lock);
> -		nr_items = READ_ONCE(l->nr_items);
> -		if (likely(nr_items != LONG_MIN)) {
> -			rcu_read_unlock();
> -			return l;
> -		}
> -		if (irq)
> -			spin_unlock_irq(&l->lock);
> -		else
> -			spin_unlock(&l->lock);
> +	if (likely(l) && lock_list_lru(l, irq)) {
> +		rcu_read_unlock();
> +		return l;
>  	}

Much easier to read, indeed :)

>  	/*
>  	 * Caller may simply bail out if raced with reparenting or
> -- 
> 2.49.0


Thanks,
SJ

[...]