Re: [PATCH 1/2] scsi: ufs: Protect some contexts from unexpected clock scaling

From: Can Guo <cang@codeaurora.org>
To: Bean Huo <huobean@gmail.com>
Cc: asutoshd@codeaurora.org, nguyenb@codeaurora.org,
	hongwus@codeaurora.org, rnayak@codeaurora.org,
	linux-scsi@vger.kernel.org, kernel-team@android.com,
	saravanak@google.com, salyzyn@google.com,
	Alim Akhtar <alim.akhtar@samsung.com>,
	Avri Altman <avri.altman@wdc.com>,
	"James E.J. Bottomley" <jejb@linux.ibm.com>,
	"Martin K. Petersen" <martin.petersen@oracle.com>,
	Stanley Chu <stanley.chu@mediatek.com>,
	Bean Huo <beanhuo@micron.com>,
	Bart Van Assche <bvanassche@acm.org>,
	Satya Tangirala <satyat@google.com>,
	open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 1/2] scsi: ufs: Protect some contexts from unexpected clock scaling
Date: Fri, 11 Dec 2020 09:09:18 +0800	[thread overview]
Message-ID: <75b3b8496bb0a41cad1fe09f93a4ca10@codeaurora.org> (raw)
In-Reply-To: <a2338ef6da3d4ed4093547ba87e13e94d8dd2a45.camel@gmail.com>

On 2020-12-11 01:34, Bean Huo wrote:
> Hi Can
> 
> On Wed, 2020-12-09 at 05:35 -0800, Can Guo wrote:
>> 
>> 
>> @@ -1160,6 +1166,7 @@ static void
>> ufshcd_clock_scaling_unprepare(struct ufs_hba *hba)
>>  {
>>  	up_write(&hba->clk_scaling_lock);
>>  	ufshcd_scsi_unblock_requests(hba);
>> +	ufshcd_release(hba);
>>  }
>> 
>>  /**
>> @@ -1175,12 +1182,9 @@ static int ufshcd_devfreq_scale(struct ufs_hba
>> *hba, bool scale_up)
>>  {
>>  	int ret = 0;
>> 
>> -	/* let's not get into low power until clock scaling is
>> completed */
>> -	ufshcd_hold(hba, false);
>> -
>>  	ret = ufshcd_clock_scaling_prepare(hba);
>>  	if (ret)
>> -		goto out;
>> +		return ret;
>> 
>>  	/* scale down the gear before scaling down clocks */
>>  	if (!scale_up) {
>> @@ -1212,8 +1216,6 @@ static int ufshcd_devfreq_scale(struct ufs_hba
>> *hba, bool scale_up)
>> 
>>  out_unprepare:
>>  	ufshcd_clock_scaling_unprepare(hba);
>> -out:
>> -	ufshcd_release(hba);
>>  	return ret;
>>  }
> 
> I didn't understand why moving ufshcd_hold/ufshcd_release into
> ufshcd_clock_scaling_prepare()/ufshcd_clock_scaling_unprepare().
> 

Say you change devfreq's governor to performance after UFS host
has entered runtime suspend.

governor_store
  devfreq_performance_handler
   update_devfreq
    devfreq_set_target
     ufshcd_devfreq_target
      ufshcd_devfreq_scale

When ufshcd_devfreq_scale() calls ufshcd_hold() when host is already
runtime suspended, guess what, NoC issues. So clk_scaling.is_allowed
should be checked first.

Regards,

Can Guo.

> 
>> 
>> @@ -1294,15 +1296,8 @@ static int ufshcd_devfreq_target(struct device
>> *dev,
>>  	}
>>  	spin_unlock_irqrestore(hba->host->host_lock, irq_flags);
>> 
>> -	pm_runtime_get_noresume(hba->dev);
>> -	if (!pm_runtime_active(hba->dev)) {
>> -		pm_runtime_put_noidle(hba->dev);
>> -		ret = -EAGAIN;
>> -		goto out;
>> -	}
>>  	start = ktime_get();
>>  	ret = ufshcd_devfreq_scale(hba, scale_up);
>> -	pm_runtime_put(hba->dev);
>> 
> 
> which branch are you working on?  I didn't see this part codes in the
> branch 5.11/scsi-queue and 5.11/scsi-staging.
> 
> Bean