On 1/28/19 6:12 PM, Peter Oskolkov wrote
> @@ -2583,7 +2594,15 @@ enum bpf_ret_code {
> BPF_DROP = 2,
> /* 3-6 reserved */
> BPF_REDIRECT = 7,
> - /* >127 are reserved for prog type specific return codes */
> + /* >127 are reserved for prog type specific return codes.
> + *
> + * BPF_LWT_REROUTE: used by BPF_PROG_TYPE_LWT_IN and
> + * BPF_PROG_TYPE_LWT_XMIT to indicate that skb's dst
> + * has changed and appropriate dst_input() or dst_output()
> + * action has to be taken (this is an L3 redirect, as
> + * opposed to L2 redirect represented by BPF_REDIRECT above).
> + */
> + BPF_LWT_REROUTE = 128,
> };
What happens if a program pushes a new header onto the skb and does not
return BPF_LWT_REROUTE?
Might be better to move the route lookup and dst swap to run_lwt_bpf and
only do it if the program returns BPF_LWT_REROUTE. That allows calling
bpf_push_ip_encap without requiring a route lookup. That might be fine
as long as their is not a protocol mismatch (ipv4 packet gets an ipv6
header or vice versa). But then, I think you have the mismatch problem
now if the program does not return BPF_LWT_REROUTE.
