net: Get rid of consume_skb when tracing is off

The function consume_skb is only meaningful when tracing is enabled.
This patch makes it conditional on CONFIG_TRACEPOINTS.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

On 8/21/20 3:23 PM, Herbert Xu wrote:
> The function consume_skb is only meaningful when tracing is enabled.
> This patch makes it conditional on CONFIG_TRACEPOINTS.
> 
> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
> 


I am not completely familiar with CONFIG_TRACEPOINTS

Is "perf probe" support requiring it ?

We want the following to be supported.

perf probe consume_skb

On Fri, Aug 21, 2020 at 04:40:49PM -0700, Eric Dumazet wrote:
>
> I am not completely familiar with CONFIG_TRACEPOINTS
> 
> Is "perf probe" support requiring it ?

Yes.

perf probe requires CONFIG_KPROBE_EVENTS which selects CONFIG_TRACING
which selects CONFIG_TRACEPOINTS.

> We want the following to be supported.
> 
> perf probe consume_skb

That should continue to work as this patch does not change anything
when CONFIG_TRACEPOINTS is enabled.

Cheers,

On Sat, Aug 22, 2020 at 08:23:29AM +1000, Herbert Xu wrote:
> The function consume_skb is only meaningful when tracing is enabled.
> This patch makes it conditional on CONFIG_TRACEPOINTS.
> 
> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
> 
> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
> index 46881d902124..e8bca74857a3 100644
> --- a/include/linux/skbuff.h
> +++ b/include/linux/skbuff.h
> @@ -1056,7 +1056,16 @@ void kfree_skb(struct sk_buff *skb);
>  void kfree_skb_list(struct sk_buff *segs);
>  void skb_dump(const char *level, const struct sk_buff *skb, bool full_pkt);
>  void skb_tx_error(struct sk_buff *skb);
> +
> +#ifdef CONFIG_TRACEPOINTS
>  void consume_skb(struct sk_buff *skb);
> +#else
> +static inline void consume_skb(struct sk_buff *skb)
> +{
> +	return kfree_skb(skb);
> +}
> +#endif
> +
Wouldn't it be better to make this:
#define consume_skb(x) kfree_skb(x)
?
Best
Neil

>  void __consume_stateless_skb(struct sk_buff *skb);
>  void  __kfree_skb(struct sk_buff *skb);
>  extern struct kmem_cache *skbuff_head_cache;
> diff --git a/net/core/skbuff.c b/net/core/skbuff.c
> index 7e2e502ef519..593fe73d4993 100644
> --- a/net/core/skbuff.c
> +++ b/net/core/skbuff.c
> @@ -820,6 +820,7 @@ void skb_tx_error(struct sk_buff *skb)
>  }
>  EXPORT_SYMBOL(skb_tx_error);
>  
> +#ifdef CONFIG_TRACEPOINTS
>  /**
>   *	consume_skb - free an skbuff
>   *	@skb: buffer to free
> @@ -837,6 +838,7 @@ void consume_skb(struct sk_buff *skb)
>  	__kfree_skb(skb);
>  }
>  EXPORT_SYMBOL(consume_skb);
> +#endif
>  
>  /**
>   *	consume_stateless_skb - free an skbuff, assuming it is stateless
> -- 
> Email: Herbert Xu <herbert@gondor.apana.org.au>
> Home Page: http://gondor.apana.org.au/~herbert/
> PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
>

From: Neil Horman <nhorman@localhost.localdomain>

Neil, you might want to fix this so people can reply to you :-)

On Sat, Aug 22, 2020 at 01:54:19PM -0400, Neil Horman wrote:
>
> Wouldn't it be better to make this:
> #define consume_skb(x) kfree_skb(x)

Either way is fine but I prefer inline functions over macros.

Cheers,

On Sat, Aug 22, 2020 at 12:49:02PM -0700, David Miller wrote:
> 
> From: Neil Horman <nhorman@localhost.localdomain>
> 
> Neil, you might want to fix this so people can reply to you :-)
Crap, thanks, new computer.

Neil

>

From: Herbert Xu <herbert@gondor.apana.org.au>
Date: Sat, 22 Aug 2020 08:23:29 +1000

> The function consume_skb is only meaningful when tracing is enabled.
> This patch makes it conditional on CONFIG_TRACEPOINTS.
> 
> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Fair enough, applied, thanks.