On Thu, Apr 13, 2017 at 9:09 AM, David Miller <da...@davemloft.net> wrote:
>
[snip]
> +static u32 netif_receive_generic_xdp(struct sk_buff *skb,
> + struct bpf_prog *xdp_prog)
> +{
> + struct xdp_buff xdp;
> + u32 act = XDP_DROP;
> + void *orig_data;
> + int hlen, off;
> +
> + if (skb_linearize(skb))
> + goto do_drop;
> +
> + /* The XDP program wants to see the packet starting at the MAC
> + * header.
> + */
> + hlen = skb_headlen(skb) + skb->mac_len;
> + xdp.data = skb->data - skb->mac_len;
> + xdp.data_end = xdp.data + hlen;
> + xdp.data_hard_start = xdp.data - skb_headroom(skb);
> + orig_data = xdp.data;
> +
> + act = bpf_prog_run_xdp(xdp_prog, &xdp);
> +
> + off = xdp.data - orig_data;
should we do "orig_data - xdp.data" instead?
The 'off' might be < 0, when pushing new header.
> + if (off)
> + __skb_push(skb, off);
When doing encapsulation by calling xdp_adjust_head(skb, offset), the
offset is < 0 in order to make extra room in the front ; so xdp.data <
orig_data, off < 0.
But if we are decapsulating protocol, then off > 0, and maybe we
should call __skb_pull()?
Thanks,
William
