Re: [RFC] net: release dst entry in dev_queue_xmit()

[Eric Dumazet <dada1@cosmosbay.com>]

David Miller a écrit :
> From: Eric Dumazet <dada1@cosmosbay.com>
> Date: Fri, 20 Mar 2009 12:40:22 +0100
> 
>> I believe we can release dst in dev_queue_xmit(), while cpu cache is
>> hot, since caller of dev_queue_xmit() had to hold a reference on dst
>> right before. This reduce work to be done by softirq handler, and
>> decrease cache misses.
>>
> 
> This will break various packet schedulers and classifiers.
> 
> Heck sch_sfq.c even uses skb->dst as part of it's flow hash
> function :-)

Well, as one of the hash perturbator, for other protocols than
IPV4 & IPV6...

        default:
                h = (unsigned long)skb->dst ^ skb->protocol;
                h2 = (unsigned long)skb->sk;
        }

        return sfq_fold_hash(q, h, h2);

But teql indeed mandates dst in __teql_resolve() 

Darn...

This dst freeing should be done very late then, in the NIC driver itself, just
before giving skb to hardware, or right before in dev_hard_start_xmit()

If done in dev_hard_start_xmit(), skb could be requeued (because of NETDEV_TX_BUSY).
Then if requeued, maybe at this time, dst being NULL is not a problem ?

Thanks a lot David