[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4c8da935-35c0-d2b3-af12-9557209ac5dd@gmail.com>
Date: Tue, 6 Nov 2018 10:59:45 -0700
From: David Ahern <dsahern@...il.com>
To: Mike Manning <mmanning@...tta.att-mail.com>, netdev@...r.kernel.org
Subject: Re: [PATCH net-next v4 3/9] net: ensure unbound datagram socket to be
chosen when not in a VRF
On 11/2/18 1:10 PM, Mike Manning wrote:
> diff --git a/include/net/udp.h b/include/net/udp.h
> index 9e82cb391dea..057972d0eea5 100644
> --- a/include/net/udp.h
> +++ b/include/net/udp.h
> @@ -252,6 +252,17 @@ static inline int udp_rqueue_get(struct sock *sk)
> return sk_rmem_alloc_get(sk) - READ_ONCE(udp_sk(sk)->forward_deficit);
> }
>
> +static inline bool udp_sk_bound_dev_eq(struct net *net, int bound_dev_if,
> + int dif, int sdif)
> +{
> +#if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
> + return inet_bound_dev_eq(net->ipv4.sysctl_udp_l3mdev_accept,
!!net->ipv4.sysctl_udp_l3mdev_accept since first arg is a bool
> + bound_dev_if, dif, sdif);
> +#else
> + return inet_bound_dev_eq(1, bound_dev_if, dif, sdif);
> +#endif
> +}
> +
> /* net/ipv4/udp.c */
> void udp_destruct_sock(struct sock *sk);
> void skb_consume_udp(struct sock *sk, struct sk_buff *skb, int len);
...
> diff --git a/net/ipv6/datagram.c b/net/ipv6/datagram.c
> index 1ede7a16a0be..4813293d4fad 100644
> --- a/net/ipv6/datagram.c
> +++ b/net/ipv6/datagram.c
> @@ -782,7 +782,10 @@ int ip6_datagram_send_ctl(struct net *net, struct sock *sk,
>
> if (src_info->ipi6_ifindex) {
> if (fl6->flowi6_oif &&
> - src_info->ipi6_ifindex != fl6->flowi6_oif)
> + src_info->ipi6_ifindex != fl6->flowi6_oif &&
> + (sk->sk_bound_dev_if != fl6->flowi6_oif ||
> + !sk_dev_equal_l3scope(
> + sk, src_info->ipi6_ifindex)))
That's getting ugly to read. A tmp variable
int idx = rc_info->ipi6_ifindex;
should shorten that to fit on one line.
> return -EINVAL;
> fl6->flowi6_oif = src_info->ipi6_ifindex;
> }
Powered by blists - more mailing lists