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: <46b4a936-c59a-ba04-4395-ed53fb72a67d@acm.org>
Date:   Thu, 29 Dec 2022 06:12:27 -0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.4.0
Subject: Re: [PATCH 1/1] ufs: update the timeout timer after resume
Content-Language: en-US
To:     Ed Tsai <ed.tsai@mediatek.com>, alim.akhtar@samsung.com,
        avri.altman@wdc.com, jejb@linux.ibm.com,
        martin.petersen@oracle.com, matthias.bgg@gmail.com,
        linux-scsi@vger.kernel.org
Cc:     peter.wang@mediatek.com, alice.chao@mediatek.com,
        powen.kao@mediatek.com, naomi.chu@mediatek.com,
        stanley.chu@mediatek.com, wsd_upstream@mediatek.com,
        linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
        linux-mediatek@lists.infradead.org
References: <20221229030645.11558-1-ed.tsai@mediatek.com>
From:   Bart Van Assche <bvanassche@acm.org>
In-Reply-To: <20221229030645.11558-1-ed.tsai@mediatek.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 12/28/22 19:06, Ed Tsai wrote:
> The tags allocation is limited by the number of active queues and a
> queue is marked as inactive by the queue timeout worker after up to 30Hz
> by default.
> 
> Therefore, tags for the general I/O may be limited to half of the max
> depth up to 30HZ after resume. To make sure the ufs request queue for pm
> usage can be inactive immediately, trigger the timeout worker to release
> the tag set.
> 
> Signed-off-by: Ed Tsai <ed.tsai@mediatek.com>
> ---
>   drivers/ufs/core/ufshcd.c | 1 +
>   1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
> index e18c9f4463ec..c77570caa3a8 100644
> --- a/drivers/ufs/core/ufshcd.c
> +++ b/drivers/ufs/core/ufshcd.c
> @@ -8842,6 +8842,7 @@ static int ufshcd_set_dev_pwr_mode(struct ufs_hba *hba,
>   		hba->curr_dev_pwr_mode = pwr_mode;
>   	}
>   
> +	mod_timer(&sdp->request_queue->timeout, 0);
>   	scsi_device_put(sdp);
>   	hba->host->eh_noresume = 0;
>   	return ret;

Where is the code that restores the timeout to the original value?

Additionally, I think that only the block layer core should manipulate 
the request queue timer directly. Please add a helper function in the 
block layer and call that helper function from the UFS driver.

Thanks,

Bart.