Re: [PATCH] net/xdp: remove unused macro REG_STATE_NEW

[Jesper Dangaard Brouer]

On Mon, 09 Nov 2020 13:44:48 -0800
John Fastabend <john.fastabend@xxxxxxxxx> wrote:

> Alex Shi wrote:
> > 
> > 
> > 在 2020/11/7 上午12:13, Jesper Dangaard Brouer 写道:  
> > > Hmm... REG_STATE_NEW is zero, so it is implicitly set via memset zero.
> > > But it is true that it is technically not directly used or referenced.
> > > 
> > > It is mentioned in a comment, so please send V2 with this additional change:  
> > 
> > Hi Jesper,
> > 
> > Thanks a lot for comments. here is the v2:
> > 
> > From 2908d25bf2e1c90ad71a83ba056743f45da283e8 Mon Sep 17 00:00:00 2001
> > From: Alex Shi <alex.shi@xxxxxxxxxxxxxxxxx>
> > Date: Fri, 6 Nov 2020 13:41:58 +0800
> > Subject: [PATCH v2] net/xdp: remove unused macro REG_STATE_NEW
> > 
> > To tame gcc warning on it:
> > net/core/xdp.c:20:0: warning: macro "REG_STATE_NEW" is not used
> > [-Wunused-macros]
> > And change related comments as Jesper Dangaard Brouer suggested.
> > 
> > Signed-off-by: Alex Shi <alex.shi@xxxxxxxxxxxxxxxxx>
> > ---  
> 
> >  net/core/xdp.c | 3 +--
> >  1 file changed, 1 insertion(+), 2 deletions(-)
> > 
> > diff --git a/net/core/xdp.c b/net/core/xdp.c
> > index 48aba933a5a8..0df5ee5682d9 100644
> > --- a/net/core/xdp.c
> > +++ b/net/core/xdp.c
> > @@ -19,7 +19,6 @@
> >  #include <trace/events/xdp.h>
> >  #include <net/xdp_sock_drv.h>
> >  
> > -#define REG_STATE_NEW		0x0
> >  #define REG_STATE_REGISTERED	0x1
> >  #define REG_STATE_UNREGISTERED	0x2
> >  #define REG_STATE_UNUSED	0x3  
> 
> I think having the define there makes it more readable and clear what
> the zero state is. But if we run with unused-macros I guess its even
> uglier to try and mark it with unused attribute.

I  agree having the define there makes it more readable and clear what
the zero state is.

We can also add code that replace the comment, that check/use these
defines.  It is slow-path code, so it doesn't hurt to add this extra
code.  Generally I find it strange to "fix" these kind of warnings, but
also don't care that we do fix them if it helps someone else spot code
where it actually matters.

> Acked-by: John Fastabend <john.fastabend@xxxxxxxxx>
> 
> > @@ -175,7 +174,7 @@ int xdp_rxq_info_reg(struct xdp_rxq_info *xdp_rxq,
> >  		return -ENODEV;
> >  	}
> >  
> > -	/* State either UNREGISTERED or NEW */
> > +	/* State either UNREGISTERED or zero */
> >  	xdp_rxq_info_init(xdp_rxq);
> >  	xdp_rxq->dev = dev;
> >  	xdp_rxq->queue_index = queue_index;
> > -- 
> > 1.8.3.1
> >   
> 
> 



-- 
Best regards,
  Jesper Dangaard Brouer
  MSc.CS, Principal Kernel Engineer at Red Hat
  LinkedIn: http://www.linkedin.com/in/brouer