Re: [RFC] netlink broadcast return value

[Patrick McHardy <kaber@trash.net>]

Pablo Neira Ayuso wrote:
> Patrick McHardy wrote:
>> So you're returning an error when at least one of the "reliable"
>> sockets doesn't get its delivery.
> 
> Patrick, I like it, I'm fine with this approach as soon as it let me add
> the "reliable" ctnetlink state-change reporting. I can add the following
> on top of the patch that David already applied:
> 
> --- a/net/netlink/af_netlink.c
> +++ b/net/netlink/af_netlink.c
> [...]
> @@ -999,6 +1000,7 @@ static inline int do_one_broadcast(struct sock *sk,
>  		p->skb2 = NULL;
>  	} else if ((val = netlink_broadcast_deliver(sk, p->skb2)) < 0) {
>  		netlink_overrun(sk);
> +		p->delivery_failure = 1;
>                 ^^^^^^^^^^^^^^^^^^^^^^^^
> Replace this by:
> +		if (nlk->flags & NETLINK_HIGHLY_RELIABLE)
> +			p->delivery_failure = 1;
> 
> And include the flag definition and setsockopt() operations in the new
> patch, of course.

Sounds good. Maybe a nicer name for the flag :)

> Please, find the previous patch that was applied to net-next tree
> enclosed to save you some time in case that you don't know what patch I
> was refering to. I think that the changes (several drivers and such) are
> still useful, as they should ignore the return value of
> netlink_broadcast() since it's not of any use for them (as we already
> discussed, they printk the error, that's useless).

Agreed. The remaining question would be what to do about
xfrm_state. I think it can stay as it is if you add this
flag, *swan could use it if desired.