Re: [PATCH] staging: vt6656: remove unused variable

[Quentin Deslandes <quentin.deslandes@itdev.co.uk>]

On Thu, May 16, 2019 at 12:19:53PM +0200, Greg Kroah-Hartman wrote:
> On Thu, May 16, 2019 at 09:50:38AM +0000, Quentin Deslandes wrote:
> > On Thu, May 16, 2019 at 11:39:51AM +0200, Greg Kroah-Hartman wrote:
> > > On Thu, May 16, 2019 at 09:31:05AM +0000, Quentin Deslandes wrote:
> > > > Fixed 'set but not used' warning message on a status variable. The
> > > > called function returning the status code 'vnt_start_interrupt_urb()'
> > > > clean up after itself and the caller function
> > > > 'vnt_int_start_interrupt()' does not returns any value.
> > > > 
> > > > Signed-off-by: Quentin Deslandes <quentin.deslandes@itdev.co.uk>
> > > > ---
> > > >  drivers/staging/vt6656/int.c | 3 +--
> > > >  1 file changed, 1 insertion(+), 2 deletions(-)
> > > > 
> > > > diff --git a/drivers/staging/vt6656/int.c b/drivers/staging/vt6656/int.c
> > > > index 504424b19fcf..ac30ce72db5a 100644
> > > > --- a/drivers/staging/vt6656/int.c
> > > > +++ b/drivers/staging/vt6656/int.c
> > > > @@ -42,13 +42,12 @@ static const u8 fallback_rate1[5][5] = {
> > > >  void vnt_int_start_interrupt(struct vnt_private *priv)
> > > >  {
> > > >  	unsigned long flags;
> > > > -	int status;
> > > >  
> > > >  	dev_dbg(&priv->usb->dev, "---->Interrupt Polling Thread\n");
> > > >  
> > > >  	spin_lock_irqsave(&priv->lock, flags);
> > > >  
> > > > -	status = vnt_start_interrupt_urb(priv);
> > > > +	vnt_start_interrupt_urb(priv);
> > > 
> > > Shouldn't you fix this by erroring out if this fails?  Why ignore the
> > > errors?
> > > 
> > > thanks,
> > > 
> > > greg k-h
> > 
> > I could, however 'vnt_start_interrupt_urb()' already call 'dev_dbg()' if
> > it fails. Nothing is done after this debug call except returning an
> > error code.
> 
> Returning an error code is fine for that function.  But then you have to
> do something with that error.
> 
> > 'vnt_int_start_interrupt()' should, IMHO, return a status code, but the
> > original developer may have good reasons not to do so.
> 
> I think that is the real problem that needs to be fixed here, don't
> paper over the issue by ignoring the return value.
> 
> thanks,
> 
> greg k-h

Thus I'll return an error value to handle this in the caller function
and then send a v2.

Thank you for your help,
Quentin