[PATCH] net: core: Remove redundant call to 'nf_reset' in 'dev_forward_skb'

[PATCH] net: core: Remove redundant call to 'nf_reset' in 'dev_forward_skb'

[Shmulik Ladkani]

'nf_reset' is called just prior calling 'netif_rx'.
No need to call it twice.

Reported-by: Igor Michailov <rgohita@gmail.com>
Signed-off-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
---
 net/core/dev.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

diff --git a/net/core/dev.c b/net/core/dev.c
index 2db88df..071f398 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -1624,7 +1624,6 @@ int dev_forward_skb(struct net_device *dev, struct sk_buff *skb)
 	}
 
 	skb_orphan(skb);
-	nf_reset(skb);
 
 	if (unlikely(!is_skb_forwardable(dev, skb))) {
 		atomic_long_inc(&dev->rx_dropped);
-- 
1.7.9

Re: [PATCH] net: core: Remove redundant call to 'nf_reset' in 'dev_forward_skb'

[Eric Dumazet]

On Thu, 2013-03-28 at 11:13 +0200, Shmulik Ladkani wrote:
> 'nf_reset' is called just prior calling 'netif_rx'.
> No need to call it twice.
> 
> Reported-by: Igor Michailov <rgohita@gmail.com>
> Signed-off-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
> ---
>  net/core/dev.c |    1 -
>  1 files changed, 0 insertions(+), 1 deletions(-)
> 
> diff --git a/net/core/dev.c b/net/core/dev.c
> index 2db88df..071f398 100644
> --- a/net/core/dev.c
> +++ b/net/core/dev.c
> @@ -1624,7 +1624,6 @@ int dev_forward_skb(struct net_device *dev, struct sk_buff *skb)
>  	}
>  
>  	skb_orphan(skb);
> -	nf_reset(skb);
>  
>  	if (unlikely(!is_skb_forwardable(dev, skb))) {
>  		atomic_long_inc(&dev->rx_dropped);

Acked-by: Eric Dumazet <edumazet@google.com>

Re: [PATCH] net: core: Remove redundant call to 'nf_reset' in 'dev_forward_skb'

[David Miller]

From: Shmulik Ladkani <shmulik.ladkani@gmail.com>
Date: Thu, 28 Mar 2013 11:13:26 +0200

> 'nf_reset' is called just prior calling 'netif_rx'.
> No need to call it twice.
> 
> Reported-by: Igor Michailov <rgohita@gmail.com>
> Signed-off-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>

I do not see this happening in the:

	macvlan_start_xmit()

		 --> macvlan_queue_xmit()

code path.

I'm not applying this patch.  There seems to be no real agreement
that the caller of dev_forward_skb() takes care of the nf_reset().

And wouldn't it be better to consolidate the nf_reset() calls
into one place instead of several, increasing the audit burdon.

Re: [PATCH] net: core: Remove redundant call to 'nf_reset' in 'dev_forward_skb'

[Eric Dumazet]

On Fri, 2013-03-29 at 15:11 -0400, David Miller wrote:
> From: Shmulik Ladkani <shmulik.ladkani@gmail.com>
> Date: Thu, 28 Mar 2013 11:13:26 +0200
> 
> > 'nf_reset' is called just prior calling 'netif_rx'.
> > No need to call it twice.
> > 
> > Reported-by: Igor Michailov <rgohita@gmail.com>
> > Signed-off-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
> 
> I do not see this happening in the:
> 
> 	macvlan_start_xmit()
> 
> 		 --> macvlan_queue_xmit()
> 
> code path.
> 
> I'm not applying this patch.  There seems to be no real agreement
> that the caller of dev_forward_skb() takes care of the nf_reset().
> 
> And wouldn't it be better to consolidate the nf_reset() calls
> into one place instead of several, increasing the audit burdon.
> 

Hmm, I believe you misread this patch :

nf_reset() is called _twice_ in dev_forward_skb()

int dev_forward_skb(struct net_device *dev, struct sk_buff *skb)
{
        if (skb_shinfo(skb)->tx_flags & SKBTX_DEV_ZEROCOPY) {
                if (skb_copy_ubufs(skb, GFP_ATOMIC)) {
                        atomic_long_inc(&dev->rx_dropped);
                        kfree_skb(skb);
                        return NET_RX_DROP;
                }
        }

        skb_orphan(skb);
<1>     nf_reset(skb);

        if (unlikely(!is_skb_forwardable(dev, skb))) {
                atomic_long_inc(&dev->rx_dropped);
                kfree_skb(skb);
                return NET_RX_DROP;
        }
        skb->skb_iif = 0;
        skb->dev = dev;
        skb_dst_drop(skb);
        skb->tstamp.tv64 = 0;
        skb->pkt_type = PACKET_HOST;
        skb->protocol = eth_type_trans(skb, dev);
        skb->mark = 0;
        secpath_reset(skb);
<2>     nf_reset(skb);
        return netif_rx(skb);
}

Re: [PATCH] net: core: Remove redundant call to 'nf_reset' in 'dev_forward_skb'

[David Miller]

From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Fri, 29 Mar 2013 12:17:36 -0700

> On Fri, 2013-03-29 at 15:11 -0400, David Miller wrote:
>> From: Shmulik Ladkani <shmulik.ladkani@gmail.com>
>> Date: Thu, 28 Mar 2013 11:13:26 +0200
>> 
>> > 'nf_reset' is called just prior calling 'netif_rx'.
>> > No need to call it twice.
>> > 
>> > Reported-by: Igor Michailov <rgohita@gmail.com>
>> > Signed-off-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
>> 
>> I do not see this happening in the:
>> 
>> 	macvlan_start_xmit()
>> 
>> 		 --> macvlan_queue_xmit()
>> 
>> code path.
>> 
>> I'm not applying this patch.  There seems to be no real agreement
>> that the caller of dev_forward_skb() takes care of the nf_reset().
>> 
>> And wouldn't it be better to consolidate the nf_reset() calls
>> into one place instead of several, increasing the audit burdon.
>> 
> 
> Hmm, I believe you misread this patch :
> 
> nf_reset() is called _twice_ in dev_forward_skb()

Aha, that makes more sense, applied.