On Sonntag, 21. August 2016 05:25:32 CEST Linus Lüssing wrote: [...]
@@ -334,8 +333,9 @@ int batadv_v_elp_iface_enable(struct batadv_hard_iface *hard_iface) goto out;
skb_reserve(hard_iface->bat_v.elp_skb, ETH_HLEN + NET_IP_ALIGN);
- elp_buff = skb_push(hard_iface->bat_v.elp_skb, BATADV_ELP_HLEN);
- elp_packet = (struct batadv_elp_packet *)elp_buff;
skb_put(hard_iface->bat_v.elp_skb, BATADV_ELP_HLEN);
elp_packet = (struct batadv_elp_packet *)
hard_iface->bat_v.elp_skb->data;memset(elp_packet, 0, BATADV_ELP_HLEN);
elp_packet->packet_type = BATADV_ELP;
I don't get right now why you did the of split the skb_put into two different "ugly" lines (skb_put + the elp_packet assignment without elp_buff). I fear that this weird (non)-alignment you've created will bite us when the patch is submitted upstream.
Maybe you can tell us more about why this removal of elp_buff is necessary.
Btw. please use the prefix "batman-adv" in the subject and not "batamn-adv" ;)
And it is at the moment not important whether it goes into next or maint. Both will be submitted by Simon to net.git because we are currently completely off with our timing (compared to the upstream submissions). I personally would go for maint.
Kind regards, Sven