Re: [PATCH] interconnect: Fix a GCC compiler warning

[Georgi Djakov]

Hi Zhong and Greg,

On 09/17/2018 02:24 AM, Greg KH wrote:
> On Mon, Sep 17, 2018 at 05:19:18PM +0800, zhong jiang wrote:
>> On 2018/9/17 16:59, Greg KH wrote:
>>> On Mon, Sep 17, 2018 at 04:53:31PM +0800, zhong jiang wrote:
>>>> +to Greg and Andrew.
>>>>
>>>> On 2018/9/16 22:31, zhong jiang wrote:
>>>>> Fix the following compile warning:
>>>>>
>>>>> drivers/interconnect/core.c:26:23: warning: 'icc_debugfs_dir' defined but not used [-Wunused-variable]
>>>>>  static struct dentry *icc_debugfs_dir;
>>>>>
>>>>> Signed-off-by: zhong jiang <zhongjiang@xxxxxxxxxx>
>>>>> ---
>>>>>  drivers/interconnect/core.c | 3 +++
>>> This file is not in a released tree, where are you seeing it?
>> I see it in linux-next.

Yes, it has been recently pulled into linux-next for wider testing
coverage. Thank you Zhong for reporting this.

> 
> Then please contact the owner of the branch where it came from.  They
> are the only ones that can handle such a patch, not anyone else.
> 
>>>>>  1 file changed, 3 insertions(+)
>>>>>
>>>>> diff --git a/drivers/interconnect/core.c b/drivers/interconnect/core.c
>>>>> index 2740f1a..9ea9959 100644
>>>>> --- a/drivers/interconnect/core.c
>>>>> +++ b/drivers/interconnect/core.c
>>>>> @@ -23,7 +23,10 @@
>>>>>  static DEFINE_IDR(icc_idr);
>>>>>  static LIST_HEAD(icc_provider_list);
>>>>>  static DEFINE_MUTEX(icc_lock);
>>>>> +
>>>>> +#ifdef CONFIG_DEBUG_FS
>>>>>  static struct dentry *icc_debugfs_dir;
>>>>> +#endif
>>> Why not move the variable to where it is used?  It should be able to go
>>> under an existing #ifdef section, right?
>> Because the variable need to be used in __init and __exit.  So it should be identified as
>> a gloabl variable.  I fix it under the #ifdef condition.
> 
> If the variable is being used this way, then the code is written
> incorrectly, such an #ifdef should not be needed for any debugfs
> calls/variables.

Thanks for the comments. Will improve it.

BR,
Georgi