Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
Date:   Tue, 26 Apr 2022 17:25:40 +0800
From:   Muchun Song <songmuchun@bytedance.com>
To:     David Rientjes <rientjes@google.com>
Cc:     cl@linux.com, penberg@kernel.org, iamjoonsoo.kim@lge.com,
        akpm@linux-foundation.org, vbabka@suse.cz,
        roman.gushchin@linux.dev, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org, smuchun@gmail.com
Subject: Re: [PATCH] mm: slab: optimize memcg_slab_free_hook()
Message-ID: <Yme6lFrVj8FHegjN@FVFYT0MHHV2J.usts.net>
References: <20220425035406.9337-1-songmuchun@bytedance.com>
 <74f6b949-f9f-deb1-2786-47c92a619fd8@google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <74f6b949-f9f-deb1-2786-47c92a619fd8@google.com>
Precedence: bulk

On Mon, Apr 25, 2022 at 12:42:23PM -0700, David Rientjes wrote:
> On Mon, 25 Apr 2022, Muchun Song wrote:
> 
> > Most callers of memcg_slab_free_hook() already know the slab,  which could
> > be passed to memcg_slab_free_hook() directly to reduce the overhead of an
> > another call of virt_to_slab().  For bulk freeing of objects, the call of
> > slab_objcgs() in the loop in memcg_slab_free_hook() is redundant as well.
> > Rework memcg_slab_free_hook() and build_detached_freelist() to reduce
> > those unnecessary overhead and make memcg_slab_free_hook() can handle bulk
> > freeing in slab_free().
> > 
> 
> I agree that for the SLUB case the code flows better after this, but I'm 
> wondering if you have any numbers to share on how much better this is once 
> we've avoided the virt_to_slab() calls?
> 
> IOW, I'm struggling to figure out if this is code cleanup or a performance 
> optimization.
>

I think it is both. The numbers are listed in another thread [1].

[1] https://lore.kernel.org/linux-mm/Yme5zE8ZU%2FFN63Av@FVFYT0MHHV2J.usts.net/T/#m7e5c4ae3e92128781664b3396ce300c8076a5b63

Thanks.