Ok, clear now.
Even though I don't understand what to set to avoid triggering the
WARN_ON(!skb_dst(skb));
inside dev_loopback_xmit.
I just would like to send the skb in loopback, i.e. moving the packet
from the sending to the receiving queue of a certain struct net_device.
On 24/04/2018 00:36, Eric Dumazet wrote:
On 04/23/2018 02:40 PM, Emanuele wrote:
Hello,
I don't know if this is the right place where to ask, but I was wondering why
the dev_loopback_xmit function defined in /net/core/dev.c takes struct net *
and struct sock * as parameters. They are never used, so I believe passing
only the struct sk_buff * should be enough.
Look at net/ipv6/ip6_output.c where NF_HOOK() uses dev_loopback_xmit().
In addition, it would like to know where I can read what is and how to set a
skb dst_entry, since I don't really understand it.
Thanks a lot,
Emanuele