Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Tue, 17 May 2022 20:53:03 +0900
From:   Hyeonggon Yoo <42.hyeyoo@gmail.com>
To:     Vasily Averin <vvs@openvz.org>
Cc:     Roman Gushchin <roman.gushchin@linux.dev>,
        Vlastimil Babka <vbabka@suse.cz>,
        Andrew Morton <akpm@linux-foundation.org>,
        Joonsoo Kim <iamjoonsoo.kim@lge.com>,
        David Rientjes <rientjes@google.com>,
        Pekka Enberg <penberg@kernel.org>,
        Christoph Lameter <cl@linux.com>, kernel@openvz.org,
        linux-mm@kvack.org, linux-kernel@vger.kernel.org,
        Shakeel Butt <shakeelb@google.com>,
        Michal Hocko <mhocko@suse.com>
Subject: Re: [PATCH] tracing: add ACCOUNT flag for allocations from marked
 slab caches
Message-ID: <YoOMn+Cwwgz/hPnh@hyeyoo>
References: <1fd21e25-b095-e055-fc2e-abda640a0575@openvz.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1fd21e25-b095-e055-fc2e-abda640a0575@openvz.org>
Precedence: bulk

On Mon, May 16, 2022 at 09:53:32PM +0300, Vasily Averin wrote:
> Slab caches marked with SLAB_ACCOUNT force accounting for every
> allocation from this cache even if __GFP_ACCOUNT flag is not passed.
> Unfortunately, at the moment this flag is not visible in ftrace output,
> and this makes it difficult to analyze the accounted allocations.
> 
> This patch adds the __GFP_ACCOUNT flag for allocations from slab caches
> marked with SLAB_ACCOUNT to the ftrace output.
> 
> Signed-off-by: Vasily Averin <vvs@openvz.org>
> ---
>  mm/slab.c | 3 +++
>  mm/slub.c | 3 +++
>  2 files changed, 6 insertions(+)
> 
> diff --git a/mm/slab.c b/mm/slab.c
> index 0edb474edef1..4c3da8dfcbdb 100644
> --- a/mm/slab.c
> +++ b/mm/slab.c
> @@ -3492,6 +3492,9 @@ void *__kmem_cache_alloc_lru(struct kmem_cache *cachep, struct list_lru *lru,
>  {
>  	void *ret = slab_alloc(cachep, lru, flags, cachep->object_size, _RET_IP_);
>  
> +	if (cachep->flags & SLAB_ACCOUNT)
> +		flags |= __GFP_ACCOUNT;
> +
>  	trace_kmem_cache_alloc(_RET_IP_, ret,
>  			       cachep->object_size, cachep->size, flags);
>  
> diff --git a/mm/slub.c b/mm/slub.c
> index ed5c2c03a47a..670bbfef9e49 100644
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -3231,6 +3231,9 @@ void *__kmem_cache_alloc_lru(struct kmem_cache *s, struct list_lru *lru,
>  {
>  	void *ret = slab_alloc(s, lru, gfpflags, _RET_IP_, s->object_size);
>  
> +	if (s->flags & SLAB_ACCOUNT)
> +		gfpflags |= __GFP_ACCOUNT;
> +
>  	trace_kmem_cache_alloc(_RET_IP_, ret, s->object_size,
>  				s->size, gfpflags);
>  
> -- 
> 2.25.1
> 
> 

To me it sounds like it would confuse memory cgroup because:
	1) For now objects are charged only in slab memcg hooks
	2) This patch makes buddy allocator charge the page too

-- 
Thanks,
Hyeonggon