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: <bad9f1e3-bbf5-7fa6-946c-55bccb70ec65@quicinc.com>
Date:   Thu, 1 Jun 2023 18:48:57 +0530
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101
 Thunderbird/91.9.1
Subject: Re: [PATCH] mm: madvise: fix uneven accounting of psi
Content-Language: en-US
To:     Andrew Morton <akpm@linux-foundation.org>
CC:     <hannes@cmpxchg.org>, <minchan@kernel.org>,
        <quic_pkondeti@quicinc.com>, <linux-mm@kvack.org>,
        <linux-kernel@vger.kernel.org>
References: <1685531374-6091-1-git-send-email-quic_charante@quicinc.com>
 <20230531150449.56311ca2ea6184b413517672@linux-foundation.org>
From:   Charan Teja Kalla <quic_charante@quicinc.com>
In-Reply-To: <20230531150449.56311ca2ea6184b413517672@linux-foundation.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
Precedence: bulk

Thanks Andrew!!

On 6/1/2023 3:34 AM, Andrew Morton wrote:
>> This patch is tested on Android, Snapdragon SoC with 8Gb RAM, 4GB swap
>> mounted on zram which has 2GB of backingdev. The test case involved
>> launching some memory hungry apps in an order and do the proactive
>> reclaim for the app that went to background using madvise(MADV_PAGEOUT).
>> We are seeing ~40% less total values of psi mem some and full when this
>> patch is combined with [1].
> Does this accounting inaccuracy have any perceptible runtime effects?
> 
With the testcase mentioned, we are able to achieve better app
concurrency(It is a measure on Android of how many apps that can survive
by not getting killed by lmkd(which relies on PSI events to take the
decissions) by the end of the testcase).

The total memsome and memfull psi events(measured in usec) are as below:
mem PSI	  	 w/o patch	with patch
some		61255313	21867736
full		31353138	15391454

Thanks,
Charan
>