[PATCH] ASoC: amd: Return ENODEV if acp63 is not found.

[PATCH] ASoC: amd: Return ENODEV if acp63 is not found.

[Tom Rix]

The clang build fails with
sound/soc/amd/ps/pci-ps.c:218:2: error: variable 'ret' is used
  uninitialized whenever switch default is taken [-Werror,-Wsometimes-uninitialized]
        default:
        ^~~~~~~

When no device is found -ENODEV should be returned.
A switch with a single case is overkill, change to if-else.

Fixes: 1d325cdaf7a2 ("ASoC: amd: ps: refactor platform device creation logic")
Signed-off-by: Tom Rix <trix@redhat.com>
---
 sound/soc/amd/ps/pci-ps.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/sound/soc/amd/ps/pci-ps.c b/sound/soc/amd/ps/pci-ps.c
index 401cfd0036be..cba8800ab4ea 100644
--- a/sound/soc/amd/ps/pci-ps.c
+++ b/sound/soc/amd/ps/pci-ps.c
@@ -205,8 +205,7 @@ static int create_acp63_platform_devs(struct pci_dev *pci, struct acp63_dev_data
 		memset(&pdevinfo, 0, sizeof(pdevinfo));
 	}
 
-	switch (adata->pdev_mask) {
-	case ACP63_PDM_DEV_MASK:
+	if (adata->pdev_mask == ACP63_PDM_DEV_MASK) {
 		adata->pdm_dev_index  = 0;
 		acp63_fill_platform_dev_info(&pdevinfo[0], parent, NULL, "acp_ps_pdm_dma",
 					     0, adata->res, 1, NULL, 0);
@@ -214,8 +213,8 @@ static int create_acp63_platform_devs(struct pci_dev *pci, struct acp63_dev_data
 					     0, NULL, 0, NULL, 0);
 		acp63_fill_platform_dev_info(&pdevinfo[2], parent, NULL, "acp_ps_mach",
 					     0, NULL, 0, NULL, 0);
-		break;
-	default:
+	} else {
+		ret = -ENODEV;
 		dev_dbg(&pci->dev, "No PDM devices found\n");
 		goto de_init;
 	}
-- 
2.27.0

Re: [PATCH] ASoC: amd: Return ENODEV if acp63 is not found.

[Nathan Chancellor]

Hi Tom,

On Thu, Jan 05, 2023 at 04:19:12PM -0500, Tom Rix wrote:
> The clang build fails with
> sound/soc/amd/ps/pci-ps.c:218:2: error: variable 'ret' is used
>   uninitialized whenever switch default is taken [-Werror,-Wsometimes-uninitialized]
>         default:
>         ^~~~~~~
> 
> When no device is found -ENODEV should be returned.
> A switch with a single case is overkill, change to if-else.
> 
> Fixes: 1d325cdaf7a2 ("ASoC: amd: ps: refactor platform device creation logic")
> Signed-off-by: Tom Rix <trix@redhat.com>

Thanks for the patch! I sent basically the same thing earlier today
(sorry for forgetting to Cc you directly) and was told it was not
correct:

https://lore.kernel.org/6bb126b7-1cb4-0c4c-d357-fadc3ffdd3f9@amd.com/

I am just waiting for some feedback before sending a v2.

> ---
>  sound/soc/amd/ps/pci-ps.c | 7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)
> 
> diff --git a/sound/soc/amd/ps/pci-ps.c b/sound/soc/amd/ps/pci-ps.c
> index 401cfd0036be..cba8800ab4ea 100644
> --- a/sound/soc/amd/ps/pci-ps.c
> +++ b/sound/soc/amd/ps/pci-ps.c
> @@ -205,8 +205,7 @@ static int create_acp63_platform_devs(struct pci_dev *pci, struct acp63_dev_data
>  		memset(&pdevinfo, 0, sizeof(pdevinfo));
>  	}
>  
> -	switch (adata->pdev_mask) {
> -	case ACP63_PDM_DEV_MASK:
> +	if (adata->pdev_mask == ACP63_PDM_DEV_MASK) {
>  		adata->pdm_dev_index  = 0;
>  		acp63_fill_platform_dev_info(&pdevinfo[0], parent, NULL, "acp_ps_pdm_dma",
>  					     0, adata->res, 1, NULL, 0);
> @@ -214,8 +213,8 @@ static int create_acp63_platform_devs(struct pci_dev *pci, struct acp63_dev_data
>  					     0, NULL, 0, NULL, 0);
>  		acp63_fill_platform_dev_info(&pdevinfo[2], parent, NULL, "acp_ps_mach",
>  					     0, NULL, 0, NULL, 0);
> -		break;
> -	default:
> +	} else {
> +		ret = -ENODEV;
>  		dev_dbg(&pci->dev, "No PDM devices found\n");
>  		goto de_init;
>  	}
> -- 
> 2.27.0
> 
> 

Re: [PATCH] ASoC: amd: Return ENODEV if acp63 is not found.

[Tom Rix]

On 1/5/23 1:49 PM, Nathan Chancellor wrote:
> Hi Tom,
>
> On Thu, Jan 05, 2023 at 04:19:12PM -0500, Tom Rix wrote:
>> The clang build fails with
>> sound/soc/amd/ps/pci-ps.c:218:2: error: variable 'ret' is used
>>    uninitialized whenever switch default is taken [-Werror,-Wsometimes-uninitialized]
>>          default:
>>          ^~~~~~~
>>
>> When no device is found -ENODEV should be returned.
>> A switch with a single case is overkill, change to if-else.
>>
>> Fixes: 1d325cdaf7a2 ("ASoC: amd: ps: refactor platform device creation logic")
>> Signed-off-by: Tom Rix <trix@redhat.com>
> Thanks for the patch! I sent basically the same thing earlier today
> (sorry for forgetting to Cc you directly) and was told it was not
> correct:
>
> https://lore.kernel.org/6bb126b7-1cb4-0c4c-d357-fadc3ffdd3f9@amd.com/
>
> I am just waiting for some feedback before sending a v2.

Looking closer, the pdev_mask is only set to ACP63_PDM_DEV_MASK so the 
case statement can be folded into the above if-check of pdev_mask.

And the default: dropped,  it looks like dead code.

Please cc me on the v2,

Tom

>
>> ---
>>   sound/soc/amd/ps/pci-ps.c | 7 +++----
>>   1 file changed, 3 insertions(+), 4 deletions(-)
>>
>> diff --git a/sound/soc/amd/ps/pci-ps.c b/sound/soc/amd/ps/pci-ps.c
>> index 401cfd0036be..cba8800ab4ea 100644
>> --- a/sound/soc/amd/ps/pci-ps.c
>> +++ b/sound/soc/amd/ps/pci-ps.c
>> @@ -205,8 +205,7 @@ static int create_acp63_platform_devs(struct pci_dev *pci, struct acp63_dev_data
>>   		memset(&pdevinfo, 0, sizeof(pdevinfo));
>>   	}
>>   
>> -	switch (adata->pdev_mask) {
>> -	case ACP63_PDM_DEV_MASK:
>> +	if (adata->pdev_mask == ACP63_PDM_DEV_MASK) {
>>   		adata->pdm_dev_index  = 0;
>>   		acp63_fill_platform_dev_info(&pdevinfo[0], parent, NULL, "acp_ps_pdm_dma",
>>   					     0, adata->res, 1, NULL, 0);
>> @@ -214,8 +213,8 @@ static int create_acp63_platform_devs(struct pci_dev *pci, struct acp63_dev_data
>>   					     0, NULL, 0, NULL, 0);
>>   		acp63_fill_platform_dev_info(&pdevinfo[2], parent, NULL, "acp_ps_mach",
>>   					     0, NULL, 0, NULL, 0);
>> -		break;
>> -	default:
>> +	} else {
>> +		ret = -ENODEV;
>>   		dev_dbg(&pci->dev, "No PDM devices found\n");
>>   		goto de_init;
>>   	}
>> -- 
>> 2.27.0
>>
>>