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;
Message-ID: <eb13cb3e5e1625afe1bb783810f4d6b52a66a2f6.camel@linux.intel.com>
Subject: Re: [PATCH v4] ipc/msg: mitigate the lock contention with percpu
 counter
From:   Tim Chen <tim.c.chen@linux.intel.com>
To:     Jiebin Sun <jiebin.sun@intel.com>, akpm@linux-foundation.org,
        vasily.averin@linux.dev, shakeelb@google.com, dennis@kernel.org,
        tj@kernel.org, cl@linux.com, ebiederm@xmission.com,
        legion@kernel.org, manfred@colorfullife.com,
        alexander.mikhalitsyn@virtuozzo.com, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
Cc:     tim.c.chen@intel.com, feng.tang@intel.com, ying.huang@intel.com,
        tianyou.li@intel.com, wangyang.guo@intel.com
Date:   Wed, 07 Sep 2022 09:01:53 -0700
In-Reply-To: <20220907172516.1210842-1-jiebin.sun@intel.com>
References: <CALvZod44uUFnwfF4StC24t+d1s_XE10hkmSCgb04FjtTATo6xQ@mail.gmail.com>
         <20220907172516.1210842-1-jiebin.sun@intel.com>
Content-Type: text/plain; charset="UTF-8"
User-Agent: Evolution 3.34.4 (3.34.4-1.fc31) 
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Thu, 2022-09-08 at 01:25 +0800, Jiebin Sun wrote:
> The msg_bytes and msg_hdrs atomic counters are frequently
> updated when IPC msg queue is in heavy use, causing heavy
> cache bounce and overhead. Change them to percpu_counter
> greatly improve the performance. Since there is one percpu
> struct per namespace, additional memory cost is minimal.
> Reading of the count done in msgctl call, which is infrequent.
> So the need to sum up the counts in each CPU is infrequent.
> 
> 
> Apply the patch and test the pts/stress-ng-1.4.0
> -- system v message passing (160 threads).
> 
> Score gain: 3.17x
> 
> 
...
>  
> +/* large batch size could reduce the times to sum up percpu counter */
> +#define MSG_PERCPU_COUNTER_BATCH 1024
> +

Jiebin, 

1024 is a small size (1/4 page). 
The local per cpu counter could overflow to the gloabal count quickly
if it is limited to this size, since our count tracks msg size.
  
I'll suggest something larger, say 8*1024*1024, about
8MB to accommodate about 2 large page worth of data.  Maybe that
will further improve throughput on stress-ng by reducing contention
on adding to the global count.

Tim