Re: [PATCH v2] ASoC: soc-card: soc-card-test: Fix some error handling in probe()

[Richard Fitzgerald <rf@opensource.cirrus.com>]

On 10/4/24 16:31, Dan Carpenter wrote:
> On Wed, Apr 10, 2024 at 03:43:45PM +0100, Richard Fitzgerald wrote:
>> On 10/4/24 15:22, Dan Carpenter wrote:
>>> Fix this reversed if statement and call put_device() before returning
>>> the error code.
>>>
>>> Fixes: ef7784e41db7 ("ASoC: soc-card: Add KUnit test case for snd_soc_card_get_kcontrol")
>>> Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
>>> ---
>>> v2: call put_device()
>>>
>>>    sound/soc/soc-card-test.c | 4 +++-
>>>    1 file changed, 3 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/sound/soc/soc-card-test.c b/sound/soc/soc-card-test.c
>>> index 075c52fe82e5..faf9a3d46884 100644
>>> --- a/sound/soc/soc-card-test.c
>>> +++ b/sound/soc/soc-card-test.c
>>> @@ -148,8 +148,10 @@ static int soc_card_test_case_init(struct kunit *test)
>>>    	priv->card->owner = THIS_MODULE;
>>>    	ret = snd_soc_register_card(priv->card);
>>> -	if (!ret)
>>> +	if (ret) {
>>> +		put_device(priv->card_dev);
>>>    		return ret;
>>> +	}
>>>    	return 0;
>>>    }
>>
>> Thanks.
>> Reviewed-by: Richard Fitzgerald <rf@opensource.cirrus.com>
>>
>> I can see that put_device() is also missing earlier in the
>> function:
>>
>> 	if (!priv->card)
>> 		return -ENOMEM;
>>
>> I can send a fix for that.
> 
> No.  Let me resend.  I'm sorry, this patch has not been up to proper
> standards.  

The same could be said for my original code here.

I suggest moving this block of code _before_ the kunit_device_register()
so there's no need to put_device() if the alloc fails:

	priv->card = kunit_kzalloc(test, sizeof(*priv->card), GFP_KERNEL);
	if (!priv->card)
		return -ENOMEM;

Also I should fix Smatch to warn about missing put_device()
> calls to prevent this sort of thing going forward.
> 
> regards,
> dan carpenter