Re: [PATCH v1] scsi: ufs: remove clk_scaling_lock when clkscaling isn't supported.

[Adrian Hunter]

On 12/02/2022 06:44, Kiwoong Kim wrote:
>> The error handler really should have exclusive access.  One of the places
>> you change does explain that:
>>
>>  		 * Hold the scaling lock just in case dev cmds
>>  		 * are sent via bsg and/or sysfs.
>>  		 */
>> -		down_write(&hba->clk_scaling_lock);
>> +		if (ufshcd_is_clkscaling_supported(hba))
>> +			down_write(&hba->clk_scaling_lock); 
> 
> 
> Yeah.., I saw the comment but didn't get why.
> 
> Is there anyone who knows why it's necessary for all SoCs?
> At lease, I know there is no reason to forbid concurrent executions of dev cmd and power mode change.
> 
> If there's nothing, how about adding a quick to ignore it?

Is it worth it?

The error handler really should have exclusive access.
Have you considered, for example, races of ufshcd_reset_and restore() and dev commands, tm commands, UIC commands.
I suspect more locking is needed not less.