Hi Tang,
On 5/24/22 03:44, tangbin wrote:
> Hi Mark & Olivier：
>
> On 2022/5/24 2:57, Mark Brown wrote:
>> On Mon, May 23, 2022 at 03:28:48PM +0200, Olivier MOYSAN wrote:
>>
>>> The current patch requires a change in the driver.
>>> Either changing STM_SAI_x_ID enums, or replacing data by a struct.
>>> For instance:
>>> struct stm32_sai_comp_data {
>>> unsigned int id;
>>> }
>>> struct stm32_sai_comp_data stm32_sai_comp_data_a = {
>>> .id = STM_SAI_A_ID;
>>> }
>>> struct of_device_id stm32_sai_sub_ids[] = {
>>> .data = &stm32_sai_comp_data_a},
>>> }
>> Either approach works for me (or a revert for that matter).
>
> Thanks for your advice, I was thoughtless.
>
> I think change the date of STM_SAI_x_ID maybe simple. But if we
> don't change the id,
>
> what about add a "#define" like the line 47:
>
> #define STM_SAI_IS_SUB(x) ((x)->id == STM_SAI_A_ID || (x)->id ==
> STM_SAI_B_ID)
>
> then in the judgement, wu use:
>
> sai->id = (uintptr_t)of_device_get_match_data(&pdev->dev);
>
> if (!STM_SAI_IS_SUB(sai))
>
> return -EINVAL;
>
>
> if you think that's ok, I will send patch v2 for you .
>
If we allow null value in STM_SAI_IS_SUB(sai) check, we can miss real
NULL pointer error from of_device_get_match_data().
The simplest way is to change STM_SAI_x_ID enums I think.
But honnestly, I feel more comfortable to let the driver unchanged.
BRs
Olivier
> Thanks
>
> Tang Bin
>
>
Hi Olivier：
On 2022/5/24 22:30, Olivier MOYSAN wrote:
> Hi Tang,
>
> On 5/24/22 03:44, tangbin wrote:
>> Hi Mark & Olivier：
>>
>> On 2022/5/24 2:57, Mark Brown wrote:
>>> On Mon, May 23, 2022 at 03:28:48PM +0200, Olivier MOYSAN wrote:
>>>
>>>> The current patch requires a change in the driver.
>>>> Either changing STM_SAI_x_ID enums, or replacing data by a struct.
>>>> For instance:
>>>> struct stm32_sai_comp_data {
>>>> unsigned int id;
>>>> }
>>>> struct stm32_sai_comp_data stm32_sai_comp_data_a = {
>>>> .id = STM_SAI_A_ID;
>>>> }
>>>> struct of_device_id stm32_sai_sub_ids[] = {
>>>> .data = &stm32_sai_comp_data_a},
>>>> }
>>> Either approach works for me (or a revert for that matter).
>>
>> Thanks for your advice, I was thoughtless.
>>
>> I think change the date of STM_SAI_x_ID maybe simple. But if we
>> don't change the id,
>>
>> what about add a "#define" like the line 47:
>>
>> #define STM_SAI_IS_SUB(x) ((x)->id == STM_SAI_A_ID || (x)->id ==
>> STM_SAI_B_ID)
>>
>> then in the judgement, wu use:
>>
>> sai->id = (uintptr_t)of_device_get_match_data(&pdev->dev);
>>
>> if (!STM_SAI_IS_SUB(sai))
>>
>> return -EINVAL;
>>
>>
>> if you think that's ok, I will send patch v2 for you .
>>
>
> If we allow null value in STM_SAI_IS_SUB(sai) check, we can miss real
> NULL pointer error from of_device_get_match_data().
>
> The simplest way is to change STM_SAI_x_ID enums I think.
> But honnestly, I feel more comfortable to let the driver unchanged.
>
Oh，you are right, I am sorry.
Please forget this patch, I'm sorry to have wasted your time.
But I saw some codes is useless in the line 48 & line 49, I think we can
remove it.
If you think so, I will send this patch for you.
Thanks
Tang Bin
> BRs
> Olivier
>
>> Thanks
>>
>> Tang Bin
>>
>>
Hi Tang,
On 5/25/22 09:36, tangbin wrote:
> Hi Olivier：
>
> On 2022/5/24 22:30, Olivier MOYSAN wrote:
>> Hi Tang,
>>
>> On 5/24/22 03:44, tangbin wrote:
>>> Hi Mark & Olivier：
>>>
>>> On 2022/5/24 2:57, Mark Brown wrote:
>>>> On Mon, May 23, 2022 at 03:28:48PM +0200, Olivier MOYSAN wrote:
>>>>
>>>>> The current patch requires a change in the driver.
>>>>> Either changing STM_SAI_x_ID enums, or replacing data by a struct.
>>>>> For instance:
>>>>> struct stm32_sai_comp_data {
>>>>> unsigned int id;
>>>>> }
>>>>> struct stm32_sai_comp_data stm32_sai_comp_data_a = {
>>>>> .id = STM_SAI_A_ID;
>>>>> }
>>>>> struct of_device_id stm32_sai_sub_ids[] = {
>>>>> .data = &stm32_sai_comp_data_a},
>>>>> }
>>>> Either approach works for me (or a revert for that matter).
>>>
>>> Thanks for your advice, I was thoughtless.
>>>
>>> I think change the date of STM_SAI_x_ID maybe simple. But if we
>>> don't change the id,
>>>
>>> what about add a "#define" like the line 47:
>>>
>>> #define STM_SAI_IS_SUB(x) ((x)->id == STM_SAI_A_ID || (x)->id ==
>>> STM_SAI_B_ID)
>>>
>>> then in the judgement, wu use:
>>>
>>> sai->id = (uintptr_t)of_device_get_match_data(&pdev->dev);
>>>
>>> if (!STM_SAI_IS_SUB(sai))
>>>
>>> return -EINVAL;
>>>
>>>
>>> if you think that's ok, I will send patch v2 for you .
>>>
>>
>> If we allow null value in STM_SAI_IS_SUB(sai) check, we can miss real
>> NULL pointer error from of_device_get_match_data().
>>
>> The simplest way is to change STM_SAI_x_ID enums I think.
>> But honnestly, I feel more comfortable to let the driver unchanged.
>>
> Oh，you are right, I am sorry.
>
> Please forget this patch, I'm sorry to have wasted your time.
>
> But I saw some codes is useless in the line 48 & line 49, I think we can
> remove it.
>
Yes, these two defines are no more useful.
Feel free to send a cleanup patch.
BRs
Olivier
> If you think so, I will send this patch for you.
>
>
> Thanks
>
> Tang Bin
>
>
>> BRs
>> Olivier
>>
>>> Thanks
>>>
>>> Tang Bin
>>>
>>>
>
>