linux-kernel - Re: [PATCH] memcg: decouple memcg_hotplug_cpu_dead from stock

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <0e9e2d5d-ec64-4ad4-a184-0c53832ff565@suse.cz>
Date: Fri, 11 Apr 2025 10:55:31 +0200
From: Vlastimil Babka <vbabka@...e.cz>
To: Shakeel Butt <shakeel.butt@...ux.dev>,
 Andrew Morton <akpm@...ux-foundation.org>,
 Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Cc: Johannes Weiner <hannes@...xchg.org>, Michal Hocko <mhocko@...nel.org>,
 Roman Gushchin <roman.gushchin@...ux.dev>,
 Muchun Song <muchun.song@...ux.dev>, Yosry Ahmed <yosry.ahmed@...ux.dev>,
 Waiman Long <llong@...hat.com>, linux-mm@...ck.org, cgroups@...r.kernel.org,
 linux-kernel@...r.kernel.org, Meta kernel team <kernel-team@...a.com>
Subject: Re: [PATCH] memcg: decouple memcg_hotplug_cpu_dead from stock_lock

On 4/10/25 23:06, Shakeel Butt wrote:
> The function memcg_hotplug_cpu_dead works on the stock of a remote dead
> CPU and drain_obj_stock works on the given stock instead of local stock,
> so there is no need to take local stock_lock anymore.
> 
> Signed-off-by: Shakeel Butt <shakeel.butt@...ux.dev>
> ---
>  mm/memcontrol.c | 20 ++++++++++++--------
>  1 file changed, 12 insertions(+), 8 deletions(-)
> 
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index f23a4d0ad239..2178a051bd09 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -1789,7 +1789,7 @@ static DEFINE_PER_CPU(struct memcg_stock_pcp, memcg_stock) = {
>  };
>  static DEFINE_MUTEX(percpu_charge_mutex);
>  
> -static void drain_obj_stock(struct memcg_stock_pcp *stock);
> +static void __drain_obj_stock(struct memcg_stock_pcp *stock);
>  static bool obj_stock_flush_required(struct memcg_stock_pcp *stock,
>  				     struct mem_cgroup *root_memcg);
>  
> @@ -1873,7 +1873,7 @@ static void drain_local_stock(struct work_struct *dummy)
>  	local_lock_irqsave(&memcg_stock.stock_lock, flags);
>  
>  	stock = this_cpu_ptr(&memcg_stock);
> -	drain_obj_stock(stock);
> +	__drain_obj_stock(stock);
>  	drain_stock(stock);
>  	clear_bit(FLUSHING_CACHED_CHARGE, &stock->flags);
>  
> @@ -1964,10 +1964,10 @@ static int memcg_hotplug_cpu_dead(unsigned int cpu)
>  
>  	stock = &per_cpu(memcg_stock, cpu);
>  
> -	/* drain_obj_stock requires stock_lock */
> -	local_lock_irqsave(&memcg_stock.stock_lock, flags);
> -	drain_obj_stock(stock);
> -	local_unlock_irqrestore(&memcg_stock.stock_lock, flags);
> +	local_irq_save(flag);

I think for RT this is not great? At least in theory, probably it's not
actually used together with cpu hotplug? As it relies on memcg_stats_lock()
I think no irq save/enable is necessary there. local_lock_irqsave wasn't
actually a irq disable on RT. I don't know if there's a handy wrapper for this.

> +	/* stock of a remote dead cpu, no need for stock_lock. */
> +	__drain_obj_stock(stock);
> +	local_irq_restore(flag);
>  
>  	drain_stock(stock);
>  
> @@ -2837,7 +2837,11 @@ static bool consume_obj_stock(struct obj_cgroup *objcg, unsigned int nr_bytes,
>  	return ret;
>  }
>  
> -static void drain_obj_stock(struct memcg_stock_pcp *stock)
> +/*
> + * Works on the given stock. The callers are responsible for the proper locking
> + * for the local or remote stocks.
> + */
> +static void __drain_obj_stock(struct memcg_stock_pcp *stock)
>  {
>  	struct obj_cgroup *old = READ_ONCE(stock->cached_objcg);
>  
> @@ -2925,7 +2929,7 @@ static void refill_obj_stock(struct obj_cgroup *objcg, unsigned int nr_bytes,
>  
>  	stock = this_cpu_ptr(&memcg_stock);
>  	if (READ_ONCE(stock->cached_objcg) != objcg) { /* reset if necessary */
> -		drain_obj_stock(stock);
> +		__drain_obj_stock(stock);
>  		obj_cgroup_get(objcg);
>  		stock->nr_bytes = atomic_read(&objcg->nr_charged_bytes)
>  				? atomic_xchg(&objcg->nr_charged_bytes, 0) : 0;