[PATCH v2 3/3] mtd: rawnand: marvell: set reg_clk to NULL if it can't be obtained

[daniel@xxxxxxxxxx (Daniel Mack)]

On Saturday, July 07, 2018 08:36 AM, Boris Brezillon wrote:
> On Sat,  7 Jul 2018 08:28:52 +0200
> Daniel Mack <daniel at zonque.org> wrote:
>> diff --git a/drivers/mtd/nand/raw/marvell_nand.c b/drivers/mtd/nand/raw/marvell_nand.c
>> index e72070fe8af9..9246c3ea6ebc 100644
>> --- a/drivers/mtd/nand/raw/marvell_nand.c
>> +++ b/drivers/mtd/nand/raw/marvell_nand.c
>> @@ -2752,15 +2752,19 @@ static int marvell_nfc_probe(struct platform_device *pdev)
>>   		return ret;
>>   
>>   	nfc->reg_clk = devm_clk_get(&pdev->dev, "reg");
>> -	if (PTR_ERR(nfc->reg_clk) != -ENOENT) {
>> -		if (!IS_ERR(nfc->reg_clk)) {
>> -			ret = clk_prepare_enable(nfc->reg_clk);
>> -			if (ret)
>> -				goto unprepare_core_clk;
>> -		} else {
>> +	if (IS_ERR(nfc->reg_clk)) {
>> +		if (PTR_ERR(nfc->reg_clk) != -ENOENT) {
>>   			ret = PTR_ERR(nfc->reg_clk);
>>   			goto unprepare_core_clk;
>>   		}
>> +
>> +		nfc->reg_clk = NULL;
>> +	}
>> +
>> +	if (nfc->reg_clk) {
>> +		ret = clk_prepare_enable(nfc->reg_clk);
>> +		if (ret)
>> +			goto unprepare_core_clk;
> 
> No need to check nfc->reg_clk value here:
> 
> 	ret = clk_prepare_enable(nfc->reg_clk);
> 	if (ret)
> 		goto unprepare_core_clk;


Ah, true.