Re: [PATCH 1/1] platform/x86: surface3-wmi: Balance locking on error path

[Darren Hart]

On Thu, Dec 15, 2016 at 03:24:58AM +0200, Andy Shevchenko wrote:
> There is a possibility that lock will be left acquired.
> Consolidate error path under out_free_unlock label.
> 
> Reported-by: kbuild test robot <lkp@xxxxxxxxx>
> Cc: Benjamin Tissoires <benjamin.tissoires@xxxxxxxxxx>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
> ---
>  drivers/platform/x86/surface3-wmi.c | 11 ++++++-----
>  1 file changed, 6 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/platform/x86/surface3-wmi.c b/drivers/platform/x86/surface3-wmi.c
> index 5553b2b85e0a..ed7a3b77a0cc 100644
> --- a/drivers/platform/x86/surface3-wmi.c
> +++ b/drivers/platform/x86/surface3-wmi.c
> @@ -60,10 +60,10 @@ static DEFINE_MUTEX(s3_wmi_lock);
>  
>  static int s3_wmi_query_block(const char *guid, int instance, int *ret)
>  {
> +	struct acpi_buffer output = { ACPI_ALLOCATE_BUFFER, NULL };
>  	acpi_status status;
>  	union acpi_object *obj;
> -
> -	struct acpi_buffer output = { ACPI_ALLOCATE_BUFFER, NULL };
> +	int error = 0;

I'd prefer ret throughout for consistency with the kernel in general, but as
error is used already, it's better to be self-consistent.

>  
>  	mutex_lock(&s3_wmi_lock);
>  	status = wmi_query_block(guid, instance, &output);
> @@ -77,13 +77,14 @@ static int s3_wmi_query_block(const char *guid, int instance, int *ret)
>  			       obj->type == ACPI_TYPE_BUFFER ?
>  						obj->buffer.length : 0);
>  		}
> -		kfree(obj);
> -		return -EINVAL;
> +		error = -EINVAL;
> +		goto out_free_unlock;
>  	}
>  	*ret = obj->integer.value;
> +out_free_unlock:

Please lead labels with a space:

 out_free_unlock:

This makes diffs a bit nicer as the label isn't used in lieu of the function
name. This is also consistent with the rest of the file.

Thanks,

-- 
Darren Hart
Intel Open Source Technology Center