Re: [PATCH] usb: gadget: uvc: fix returnvar.cocci warnings

Re: [PATCH] usb: gadget: uvc: fix returnvar.cocci warnings

[Julia Lawall]

Coccinelle suggests the following patch.  But the code is curious.  Is the
function expected to always return a failure value?

thanks,
julia

On Thu, 17 Sep 2015, kbuild test robot wrote:

> TO: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
> CC: kbuild-all@01.org
> CC: Felipe Balbi <balbi@ti.com>
> CC: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> CC: "Greg Kroah-Hartman" <gregkh@linuxfoundation.org>
> CC: linux-usb@vger.kernel.org
> CC: linux-kernel@vger.kernel.org
>
> drivers/usb/gadget/function/uvc_configfs.c:866:5-8: Unneeded variable: "ret". Return "- EINVAL" on line 891
>
>
>  Remove unneeded variable used to store return value.
>
> Generated by: scripts/coccinelle/misc/returnvar.cocci
>
> CC: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
> Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
> ---
>
> Please take the patch only if it's a positive warning. Thanks!
>
>  uvc_configfs.c |    3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
>
> --- a/drivers/usb/gadget/function/uvc_configfs.c
> +++ b/drivers/usb/gadget/function/uvc_configfs.c
> @@ -863,7 +863,6 @@ static int uvcg_streaming_header_drop_li
>  	struct uvcg_streaming_header *src_hdr;
>  	struct uvcg_format *target_fmt = NULL;
>  	struct uvcg_format_ptr *format_ptr, *tmp;
> -	int ret = -EINVAL;
>
>  	src_hdr = to_uvcg_streaming_header(src);
>  	mutex_lock(su_mutex); /* for navigating configfs hierarchy */
> @@ -888,7 +887,7 @@ static int uvcg_streaming_header_drop_li
>  out:
>  	mutex_unlock(&opts->lock);
>  	mutex_unlock(su_mutex);
> -	return ret;
> +	return -EINVAL;
>
>  }
>
>

Re: [PATCH] usb: gadget: uvc: fix returnvar.cocci warnings

[Andrzej Pietrasiewicz]

Hi Julia,

W dniu 17.09.2015 o 10:57, Julia Lawall pisze:
> Coccinelle suggests the following patch.  But the code is curious.  Is the
> function expected to always return a failure value?
>

Thank you for catching this. The function is not expected to always
return a failure value. Fortunately it does not matter anyway because
the return value of the drop_link() operation is silently ignored by
its caller in fs/configfs/symlink.c, functions configfs_symlink()
and configfs_unlink(). For my comments see inline.

> thanks,
> julia
>
> On Thu, 17 Sep 2015, kbuild test robot wrote:
>
>> TO: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
>> CC: kbuild-all@01.org
>> CC: Felipe Balbi <balbi@ti.com>
>> CC: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
>> CC: "Greg Kroah-Hartman" <gregkh@linuxfoundation.org>
>> CC: linux-usb@vger.kernel.org
>> CC: linux-kernel@vger.kernel.org
>>
>> drivers/usb/gadget/function/uvc_configfs.c:866:5-8: Unneeded variable: "ret". Return "- EINVAL" on line 891
>>
>>
>>   Remove unneeded variable used to store return value.
>>
>> Generated by: scripts/coccinelle/misc/returnvar.cocci
>>
>> CC: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
>> Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
>> ---
>>
>> Please take the patch only if it's a positive warning. Thanks!
>>
>>   uvc_configfs.c |    3 +--
>>   1 file changed, 1 insertion(+), 2 deletions(-)
>>
>> --- a/drivers/usb/gadget/function/uvc_configfs.c
>> +++ b/drivers/usb/gadget/function/uvc_configfs.c
>> @@ -863,7 +863,6 @@ static int uvcg_streaming_header_drop_li
>>   	struct uvcg_streaming_header *src_hdr;
>>   	struct uvcg_format *target_fmt = NULL;
>>   	struct uvcg_format_ptr *format_ptr, *tmp;
>> -	int ret = -EINVAL;
>>
>>   	src_hdr = to_uvcg_streaming_header(src);
>>   	mutex_lock(su_mutex); /* for navigating configfs hierarchy */
>> @@ -888,7 +887,7 @@ static int uvcg_streaming_header_drop_li
>>   out:
>>   	mutex_unlock(&opts->lock);
>>   	mutex_unlock(su_mutex);
>> -	return ret;
>> +	return -EINVAL;

return 0;


Thanks,

AP

Re: [PATCH] usb: gadget: uvc: fix returnvar.cocci warnings

[SF Markus Elfring]

> Fortunately it does not matter anyway because the return value
> of the drop_link() operation is silently ignored by its caller in
> fs/configfs/symlink.c, functions configfs_symlink() and configfs_unlink().

Should such an implementation detail be also reconsidered once more?

Regards,
Markus

Re: [PATCH] usb: gadget: uvc: fix returnvar.cocci warnings

[Laurent Pinchart]

Hi Andrzej and Julia,

Could one of you please submit a patch to fix this ?

On Thursday 17 Sep 2015 13:18:04 Andrzej Pietrasiewicz wrote:
> Hi Julia,
> 
> W dniu 17.09.2015 o 10:57, Julia Lawall pisze:
> > Coccinelle suggests the following patch.  But the code is curious.  Is the
> > function expected to always return a failure value?
> 
> Thank you for catching this. The function is not expected to always
> return a failure value. Fortunately it does not matter anyway because
> the return value of the drop_link() operation is silently ignored by
> its caller in fs/configfs/symlink.c, functions configfs_symlink()
> and configfs_unlink(). For my comments see inline.
> 
> > thanks,
> > julia
> > 
> > On Thu, 17 Sep 2015, kbuild test robot wrote:
> >> TO: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
> >> CC: kbuild-all@01.org
> >> CC: Felipe Balbi <balbi@ti.com>
> >> CC: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> >> CC: "Greg Kroah-Hartman" <gregkh@linuxfoundation.org>
> >> CC: linux-usb@vger.kernel.org
> >> CC: linux-kernel@vger.kernel.org
> >> 
> >> drivers/usb/gadget/function/uvc_configfs.c:866:5-8: Unneeded variable:
> >> "ret". Return "- EINVAL" on line 891>> 
> >>   Remove unneeded variable used to store return value.
> >> 
> >> Generated by: scripts/coccinelle/misc/returnvar.cocci
> >> 
> >> CC: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
> >> Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
> >> ---
> >> 
> >> Please take the patch only if it's a positive warning. Thanks!
> >> 
> >>   uvc_configfs.c |    3 +--
> >>   1 file changed, 1 insertion(+), 2 deletions(-)
> >> 
> >> --- a/drivers/usb/gadget/function/uvc_configfs.c
> >> +++ b/drivers/usb/gadget/function/uvc_configfs.c
> >> @@ -863,7 +863,6 @@ static int uvcg_streaming_header_drop_li
> >>   	struct uvcg_streaming_header *src_hdr;
> >>   	struct uvcg_format *target_fmt = NULL;
> >>   	struct uvcg_format_ptr *format_ptr, *tmp;
> >> -	int ret = -EINVAL;
> >> 
> >>   	src_hdr = to_uvcg_streaming_header(src);
> >>   	mutex_lock(su_mutex); /* for navigating configfs hierarchy */
> >> @@ -888,7 +887,7 @@ static int uvcg_streaming_header_drop_li
> >>   out:
> >>   	mutex_unlock(&opts->lock);
> >>   	mutex_unlock(su_mutex);
> >> 
> >> -	return ret;
> >> +	return -EINVAL;
> 
> return 0;

-- 
Regards,

Laurent Pinchart

Re: [PATCH] usb: gadget: uvc: fix returnvar.cocci warnings

[Andrzej Pietrasiewicz]

Hi Laurent,

Thanks for a reminder.  Please see inline.

W dniu 22.11.2016 o 18:27, Laurent Pinchart pisze:
> Hi Andrzej and Julia,
>
> Could one of you please submit a patch to fix this ?
>
> On Thursday 17 Sep 2015 13:18:04 Andrzej Pietrasiewicz wrote:
>> Hi Julia,
>>
>> W dniu 17.09.2015 o 10:57, Julia Lawall pisze:
>>> Coccinelle suggests the following patch.  But the code is curious.  Is the
>>> function expected to always return a failure value?

As a matter of fact it seems it should not return anything at all,
because...

>>
>> Thank you for catching this. The function is not expected to always
>> return a failure value. Fortunately it does not matter anyway because

...because
>> the return value of the drop_link() operation is silently ignored by

And the Documentation/filesystems/configfs/configfs.txt says here:

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/Documentation/filesystems/configfs/configfs.txt#n397

"When unlink(2) is called on the symbolic link, the source item is
notified via the ->drop_link() method.  Like the ->drop_item() method,
this is a void function and cannot return failure."

The ->drop_item() is indeed a void function, the ->drop_link() is
actually not. This, together with the fact that the value of ->drop_link()
is silently ignored suggests, that it is the ->drop_link() return
type that should be corrected and changed to void.

@Joel: What is your opinion? Should return type be changed to void?
Is there any reason why it should still be declared int?

I'm sending a copy of this mail to target-devel and linux-nvme,
because other potentially affected users of configfs live there.

AP

Re: [PATCH] usb: gadget: uvc: fix returnvar.cocci warnings

[Christoph Hellwig]

On Wed, Nov 23, 2016 at 09:35:36AM +0100, Andrzej Pietrasiewicz wrote:
> The ->drop_item() is indeed a void function, the ->drop_link() is
> actually not. This, together with the fact that the value of ->drop_link()
> is silently ignored suggests, that it is the ->drop_link() return
> type that should be corrected and changed to void.

Please send a patch to change it.