Received-SPF: pass (google.com: domain of linux-kernel+bounces-179162-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Message-ID: <0300cd4e-46d6-499a-98d5-72360c94ae49@acm.org>
Date: Tue, 14 May 2024 14:54:14 -0600
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH v3] scsi: ufs: Allow RTT negotiation
To: Avri Altman <Avri.Altman@wdc.com>,
 "Martin K . Petersen" <martin.petersen@oracle.com>
Cc: Bean Huo <beanhuo@micron.com>, Peter Wang <peter.wang@mediatek.com>,
 "linux-scsi@vger.kernel.org" <linux-scsi@vger.kernel.org>,
 "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
References: <20240514050823.735-1-avri.altman@wdc.com>
 <34c50f23-82dc-4b53-b8cb-e5c07c6e0106@acm.org>
 <DM6PR04MB6575CE65772D92073360FE64FCE32@DM6PR04MB6575.namprd04.prod.outlook.com>
Content-Language: en-US
From: Bart Van Assche <bvanassche@acm.org>
In-Reply-To: <DM6PR04MB6575CE65772D92073360FE64FCE32@DM6PR04MB6575.namprd04.prod.outlook.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

On 5/14/24 14:34, Avri Altman wrote:
>> On 5/13/24 23:08, Avri Altman wrote:
>>> +/* bMaxNumOfRTT is equal to two after device manufacturing */
>>> +#define DEFAULT_MAX_NUM_RTT 2
>>> [ ... ]
>>> +     /* do not override if it was already written */
>>> +     if (dev_rtt != DEFAULT_MAX_NUM_RTT)
>>> +             return;
>>
>> I haven't found any text in the UFSHCI 4.0 specification that says
>> that the default value for the number of outstanding RTT requests
>> should be 2. Did I perhaps overlook something? If I didn't overlook
>> anything, the driver should not try to check whether dev_rtt is at its
>> default value.
> JEDEC Standard No. 220F Page 150 Line 2837 says: "bMaxNumOfRTT is equal to two after device manufacturing,"

Thanks Avri for having looked this up.

My understanding is that the above check won't work as intended if
ufshcd_rtt_set() does not modify the RTT value. Wouldn't it be better
to add a boolean in struct ufs_hba that indicates whether or not
ufshcd_rtt_set() has been called before?

Thanks,

Bart.