[PATCH 01/11] scsi: ufs: core: Introduce a new ufshcd vops negotiate_pwr_mode()

Can Guo can.guo at oss.qualcomm.com
Sun Mar 1 06:26:20 PST 2026


Hi Bart,

On 2/28/2026 3:31 AM, Bart Van Assche wrote:
> On 2/27/26 8:07 AM, Can Guo wrote:
>> Before power mode change to a target power mode, TX Equalzation Training
>
> "Equalzation" -> "Equalization"
Done.
>
>> (EQTR) needs be done for that target power mode. In addition, before TX
>> EQTR we need to change the power mode to HS-G1. These cannot happen
>> before the vops pwr_change_notify(PRE_CHANGE) because we don't know the
>> negotiated target power mode yet. It is neither approprite if all these
>> happen post the vops pwr_change_notify(PRE_CHANGE) as we are going to
>> change the power mode to HS-G1 for TX EQTR.
>
> approprite -> appropriate
Done.
>
> Additionally, if "neither" occurs in a sentence, "nor" should occur in
> the same sentence. I don't see "nor" in the above sentence?
Will improve the commit message in next version.
>
>> Introduce a new ufshcd vops negotiate_pwr_mode(), so that TX EQTR can be
>> done after vops negotiate_pwr_mode() and before vops 
>> pwr_change_notify().
>
> This patch does much more than only introducing a new vendor operation.
> Please make sure the patch description is complete.
Done.
>
>> -    return -ENOTSUPP;
>> +    return -EOPNOTSUPP;
>
> Why has ENOTSUPP been changed into EOPNOTSUPP?
I got a warning from checkpatch.pl when I add the new vops, so I changed 
the same for
ufshcd_vops_pwr_change_notify() too.

WARNING: ENOTSUPP is not a SUSV4 error code, prefer EOPNOTSUPP
#59: FILE: drivers/ufs/core/ufshcd-priv.h:178:
+       return -ENOTSUPP;
>
>> -static int ufshcd_change_power_mode(struct ufs_hba *hba,
>> -                 struct ufs_pa_layer_attr *pwr_mode)
>> +static int __ufshcd_change_power_mode(struct ufs_hba *hba,
>> +                      struct ufs_pa_layer_attr *pwr_mode)
>>   {
>>       int ret;
>
> The double underscore prefix is typically used in the Linux kernel to
> indicate that the caller holds a lock. That is not the case here. Please
> choose another name for this function, e.g.
> ufshcd_dme_change_power_mode().
Done.
>
> Thanks,
>
> Bart.




More information about the Linux-mediatek mailing list