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;
Subject: Re: [PATCH 2/2] ufs: don't use the fair tag sharings
To:     Bart Van Assche <bvanassche@acm.org>,
        Yu Kuai <yukuai1@huaweicloud.com>,
        Christoph Hellwig <hch@infradead.org>
Cc:     Ed Tsai <ed.tsai@mediatek.com>, axboe@kernel.dk,
        linux-block@vger.kernel.org, linux-scsi@vger.kernel.org,
        linux-kernel@vger.kernel.org, martin.petersen@oracle.com,
        stanley.chu@mediatek.com, peter.wang@mediatek.com,
        chun-hung.wu@mediatek.com, alice.chao@mediatek.com,
        powen.kao@mediatek.com, naomi.chu@mediatek.com,
        wsd_upstream@mediatek.com, "yukuai (C)" <yukuai3@huawei.com>
References: <20230509065230.32552-1-ed.tsai@mediatek.com>
 <20230509065230.32552-3-ed.tsai@mediatek.com>
 <ZF0K7A6G2cYBjSgn@infradead.org>
 <aa9af9ae-62a4-6469-244c-b5d9106bb044@acm.org>
 <ZF5G5ztMng8Xbd1W@infradead.org>
 <2740ee82-e35f-1cbf-f5d0-373f94eb14a5@acm.org>
 <de3f41a0-b13d-d4f6-765a-19b857bce53e@huaweicloud.com>
 <86065501-ab2e-09b4-71cd-c0b18ede00ed@acm.org>
 <a26e28a6-91e0-e803-749e-2ce957711c64@huaweicloud.com>
 <097caed2-10b3-7cd1-7c06-90f983e5c720@acm.org>
From:   Yu Kuai <yukuai1@huaweicloud.com>
Message-ID: <f9ccab59-91a1-69d5-6d20-2c6ea0e24b5a@huaweicloud.com>
Date:   Thu, 18 May 2023 09:49:06 +0800
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:60.0) Gecko/20100101
 Thunderbird/60.8.0
MIME-Version: 1.0
In-Reply-To: <097caed2-10b3-7cd1-7c06-90f983e5c720@acm.org>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk

Hi,

在 2023/05/18 2:23, Bart Van Assche 写道:
> On 5/17/23 00:49, Yu Kuai wrote:
>> 在 2023/05/16 23:12, Bart Van Assche 写道:
>>> I propose that we switch to one of these two approaches:
>>
>> How about a smoothing method that the device with more io will share
>> more tag, and each device will get at least one tag?
> 
> Hi Yu,
> 
> hctx_may_queue() is called from the hot path (blk_mq_get_tag()). I'm 
> pretty sure that adding any nontrivial code in that path will cause a 
> performance (IOPS) regression. So I don't think that adding a smoothing 
> method in hctx_may_queue() is a realistic option.
> 

Currently, fair share from hctx_may_queue() requires two
atomic_read(active_queues and active_requests), I think this smoothing
method can be placed into get_tag fail path, for example, the more times
a disk failed to get tag in a period of time, the more tag this disk can
get, and all the information can be updated here(perhaps directly
record how many tags a disk can get, then hctx_may_queue() still only
require 2 atomic_read()).

Thanks,
Bart