* [PATCH 2/2] geneve: handle ipv6 priority like ipv4 tos
2015-10-20 15:11 ` [PATCH v4 1/2] " John W. Linville
@ 2015-10-20 15:11 ` John W. Linville
2015-10-21 5:13 ` Jesse Gross
2015-10-20 22:55 ` [PATCH v4 1/2] geneve: implement support for IPv6-based tunnels kbuild test robot
` (3 subsequent siblings)
4 siblings, 1 reply; 37+ messages in thread
From: John W. Linville @ 2015-10-20 15:11 UTC (permalink / raw)
To: netdev
Cc: Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc, John W. Linville
Other callers of udp_tunnel6_xmit_skb just pass 0 for the prio
argument. Jesse Gross <jesse@nicira.com> suggested that prio is really
the same as IPv4's tos and should be handled the same, so this is my
interpretation of that suggestion.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Reported-by: Jesse Gross <jesse@nicira.com>
---
drivers/net/geneve.c | 19 +++++++++++++++++--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index 217b472ab9e7..cef71bdda13d 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -748,6 +748,7 @@ static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
struct geneve_dev *geneve = netdev_priv(dev);
struct geneve_sock *gs6 = geneve->sock6;
struct dst_entry *dst = NULL;
+ __u8 prio;
memset(fl6, 0, sizeof(*fl6));
fl6->flowi6_mark = skb->mark;
@@ -756,7 +757,16 @@ static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
if (info) {
fl6->daddr = info->key.u.ipv6.dst;
fl6->saddr = info->key.u.ipv6.src;
+ fl6->flowi6_tos = RT_TOS(info->key.tos);
} else {
+ prio = geneve->tos;
+ if (prio == 1) {
+ const struct iphdr *iip = ip_hdr(skb);
+
+ prio = ip_tunnel_get_dsfield(iip, skb);
+ }
+
+ fl6->flowi6_tos = RT_TOS(prio);
fl6->daddr = geneve->remote.sin6.sin6_addr;
}
@@ -876,8 +886,9 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
struct geneve_dev *geneve = netdev_priv(dev);
struct geneve_sock *gs6 = geneve->sock6;
struct dst_entry *dst = NULL;
+ const struct iphdr *iip; /* interior IP header */
struct flowi6 fl6;
- __u8 ttl;
+ __u8 prio, ttl;
__be16 sport;
bool udp_csum;
int err;
@@ -906,6 +917,8 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
skb_reset_mac_header(skb);
+ iip = ip_hdr(skb);
+
if (info) {
const struct ip_tunnel_key *key = &info->key;
u8 *opts = NULL;
@@ -922,6 +935,7 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
if (unlikely(err))
goto err;
+ prio = ip_tunnel_ecn_encap(key->tos, iip, skb);
ttl = key->ttl;
} else {
udp_csum = false;
@@ -930,13 +944,14 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
if (unlikely(err))
goto err;
+ prio = ip_tunnel_ecn_encap(fl6.flowi6_tos, iip, skb);
ttl = geneve->ttl;
if (!ttl && ipv6_addr_is_multicast(&fl6.daddr))
ttl = 1;
ttl = ttl ? : ip6_dst_hoplimit(dst);
}
err = udp_tunnel6_xmit_skb(dst, gs6->sock->sk, skb, dev,
- &fl6.saddr, &fl6.daddr, 0, ttl,
+ &fl6.saddr, &fl6.daddr, prio, ttl,
sport, geneve->dst_port, !udp_csum);
iptunnel_xmit_stats(err, &dev->stats, dev->tstats);
--
2.4.3
^ permalink raw reply related [flat|nested] 37+ messages in thread
* Re: [PATCH 2/2] geneve: handle ipv6 priority like ipv4 tos
2015-10-20 15:11 ` [PATCH 2/2] geneve: handle ipv6 priority like ipv4 tos John W. Linville
@ 2015-10-21 5:13 ` Jesse Gross
0 siblings, 0 replies; 37+ messages in thread
From: Jesse Gross @ 2015-10-21 5:13 UTC (permalink / raw)
To: John W. Linville; +Cc: netdev, Dave Miller, Pravin B Shelar, Jiri Benc
On Tue, Oct 20, 2015 at 11:11 PM, John W. Linville
<linville@tuxdriver.com> wrote:
> Other callers of udp_tunnel6_xmit_skb just pass 0 for the prio
> argument. Jesse Gross <jesse@nicira.com> suggested that prio is really
> the same as IPv4's tos and should be handled the same, so this is my
> interpretation of that suggestion.
>
> Signed-off-by: John W. Linville <linville@tuxdriver.com>
> Reported-by: Jesse Gross <jesse@nicira.com>
Reviewed-by: Jesse Gross <jesse@nicira.com>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH v4 1/2] geneve: implement support for IPv6-based tunnels
2015-10-20 15:11 ` [PATCH v4 1/2] " John W. Linville
2015-10-20 15:11 ` [PATCH 2/2] geneve: handle ipv6 priority like ipv4 tos John W. Linville
@ 2015-10-20 22:55 ` kbuild test robot
2015-10-21 1:52 ` YOSHIFUJI Hideaki/吉藤英明
` (2 subsequent siblings)
4 siblings, 0 replies; 37+ messages in thread
From: kbuild test robot @ 2015-10-20 22:55 UTC (permalink / raw)
To: John W. Linville
Cc: kbuild-all, netdev, Dave Miller, Pravin B Shelar, Jesse Gross,
Jiri Benc, John W. Linville
[-- Attachment #1: Type: text/plain, Size: 2479 bytes --]
Hi John,
[auto build test WARNING on v4.3-rc6 -- if it's inappropriate base, please suggest rules for selecting the more suitable base]
url: https://github.com/0day-ci/linux/commits/John-W-Linville/geneve-implement-support-for-IPv6-based-tunnels/20151020-232128
config: x86_64-randconfig-a0-10210538 (attached as .config)
reproduce:
# save the attached .config to linux build tree
make ARCH=x86_64
All warnings (new ones prefixed by >>):
In file included from include/linux/linkage.h:4:0,
from include/linux/kernel.h:6,
from drivers/net/geneve.c:13:
drivers/net/geneve.c: In function 'geneve_rx':
drivers/net/geneve.c:239:6: error: 'ip6h' undeclared (first use in this function)
if (ip6h)
^
include/linux/compiler.h:147:28: note: in definition of macro '__trace_if'
if (__builtin_constant_p((cond)) ? !!(cond) : \
^
>> drivers/net/geneve.c:239:2: note: in expansion of macro 'if'
if (ip6h)
^
drivers/net/geneve.c:239:6: note: each undeclared identifier is reported only once for each function it appears in
if (ip6h)
^
include/linux/compiler.h:147:28: note: in definition of macro '__trace_if'
if (__builtin_constant_p((cond)) ? !!(cond) : \
^
>> drivers/net/geneve.c:239:2: note: in expansion of macro 'if'
if (ip6h)
^
vim +/if +239 drivers/net/geneve.c
223 skb_reset_mac_header(skb);
224 skb_scrub_packet(skb, !net_eq(geneve->net, dev_net(geneve->dev)));
225 skb->protocol = eth_type_trans(skb, geneve->dev);
226 skb_postpull_rcsum(skb, eth_hdr(skb), ETH_HLEN);
227
228 if (tun_dst)
229 skb_dst_set(skb, &tun_dst->dst);
230
231 /* Ignore packet loops (and multicast echo) */
232 if (ether_addr_equal(eth_hdr(skb)->h_source, geneve->dev->dev_addr))
233 goto drop;
234
235 skb_reset_network_header(skb);
236
237 if (iph)
238 err = IP_ECN_decapsulate(iph, skb);
> 239 if (ip6h)
240 err = IP6_ECN_decapsulate(ip6h, skb);
241
242 if (unlikely(err)) {
243 if (log_ecn_error) {
244 if (iph)
245 net_info_ratelimited("non-ECT from %pI4 "
246 "with TOS=%#x\n",
247 &iph->saddr, iph->tos);
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
[-- Attachment #2: .config.gz --]
[-- Type: application/octet-stream, Size: 24631 bytes --]
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH v4 1/2] geneve: implement support for IPv6-based tunnels
2015-10-20 15:11 ` [PATCH v4 1/2] " John W. Linville
2015-10-20 15:11 ` [PATCH 2/2] geneve: handle ipv6 priority like ipv4 tos John W. Linville
2015-10-20 22:55 ` [PATCH v4 1/2] geneve: implement support for IPv6-based tunnels kbuild test robot
@ 2015-10-21 1:52 ` YOSHIFUJI Hideaki/吉藤英明
2015-10-21 18:58 ` John W. Linville
2015-10-21 5:06 ` Jesse Gross
2015-10-22 19:45 ` [PATCH v5 " John W. Linville
4 siblings, 1 reply; 37+ messages in thread
From: YOSHIFUJI Hideaki/吉藤英明 @ 2015-10-21 1:52 UTC (permalink / raw)
To: John W. Linville, netdev
Cc: hideaki.yoshifuji, Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc
Hi,
John W. Linville wrote:
> Signed-off-by: John W. Linville <linville@tuxdriver.com>
> ---
> v4:
> - treat mode field of ip_tunnel_info as flags
> - add a missing IS_ENABLED(CONFIG_IPV6) to geneve_rx
> - remove unneeded flags field in geneve_dev
> - NULL-check parameter for __geneve_sock_release
> - check remote socket family for AF_UNSPEC in geneve_configure
> - rename geneve_get_{rt,dst} as geneve_get_{v4_rt,v6_dst}
> - refactor some error handling in the xmit paths
>
> v3:
> - declare geneve_remote_unspec as static
>
> v2:
> - do not require remote address for tx on metadata tunnels
> - pass correct sockaddr family to udp_tun_rx_dst in geneve_rx
> - accommodate both ipv4 and ipv6 sockets open on same tunnel
> - move declaration of geneve_get_dst for aesthetic purposes
>
> drivers/net/geneve.c | 459 +++++++++++++++++++++++++++++++++++--------
> include/uapi/linux/if_link.h | 1 +
> 2 files changed, 377 insertions(+), 83 deletions(-)
>
> diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
> index 8f5c02eed47d..217b472ab9e7 100644
> --- a/drivers/net/geneve.c
> +++ b/drivers/net/geneve.c
> @@ -46,16 +46,25 @@ struct geneve_net {
>
> static int geneve_net_id;
>
> +union geneve_addr {
> + struct sockaddr_in sin;
> + struct sockaddr_in6 sin6;
> + struct sockaddr sa;
> +};
> +
> +static union geneve_addr geneve_remote_unspec = { .sa.sa_family = AF_UNSPEC, };
> +
> /* Pseudo network device */
> struct geneve_dev {
> struct hlist_node hlist; /* vni hash table */
> struct net *net; /* netns for packet i/o */
> struct net_device *dev; /* netdev for geneve tunnel */
> - struct geneve_sock *sock; /* socket used for geneve tunnel */
> + struct geneve_sock *sock4; /* IPv4 socket used for geneve tunnel */
> + struct geneve_sock *sock6; /* IPv6 socket used for geneve tunnel */
> u8 vni[3]; /* virtual network ID for tunnel */
> u8 ttl; /* TTL override */
> u8 tos; /* TOS override */
> - struct sockaddr_in remote; /* IPv4 address for link partner */
> + union geneve_addr remote; /* IP address for link partner */
> struct list_head next; /* geneve's per namespace list */
> __be16 dst_port;
> bool collect_md;
> @@ -103,11 +112,32 @@ static struct geneve_dev *geneve_lookup(struct geneve_sock *gs,
> vni_list_head = &gs->vni_list[hash];
> hlist_for_each_entry_rcu(geneve, vni_list_head, hlist) {
> if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
> - addr == geneve->remote.sin_addr.s_addr)
> + addr == geneve->remote.sin.sin_addr.s_addr)
> + return geneve;
> + }
> + return NULL;
> +}
> +
> +#if IS_ENABLED(CONFIG_IPV6)
> +static struct geneve_dev *geneve6_lookup(struct geneve_sock *gs,
> + struct in6_addr addr6, u8 vni[])
> +{
> + struct hlist_head *vni_list_head;
> + struct geneve_dev *geneve;
> + __u32 hash;
> +
> + /* Find the device for this VNI */
> + hash = geneve_net_vni_hash(vni);
> + vni_list_head = &gs->vni_list[hash];
> + hlist_for_each_entry_rcu(geneve, vni_list_head, hlist) {
> + if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
> + !memcmp(&addr6, &geneve->remote.sin6.sin6_addr,
> + sizeof(addr6)))
Please use ipv6_addr_equal().
How do you handle link-local addresses here?
> return geneve;
> }
> return NULL;
> }
> +#endif
>
> static inline struct genevehdr *geneve_hdr(const struct sk_buff *skb)
> {
> @@ -121,24 +151,49 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
> struct metadata_dst *tun_dst = NULL;
> struct geneve_dev *geneve = NULL;
> struct pcpu_sw_netstats *stats;
> - struct iphdr *iph;
> - u8 *vni;
> + struct iphdr *iph = NULL;
> __be32 addr;
> - int err;
> + static u8 zero_vni[3];
> + u8 *vni;
> + int err = 0;
> + sa_family_t sa_family;
> +#if IS_ENABLED(CONFIG_IPV6)
> + struct ipv6hdr *ip6h = NULL;
> + struct in6_addr addr6;
> + static struct in6_addr zero_addr6;
> +#endif
>
> - iph = ip_hdr(skb); /* outer IP header... */
> + sa_family = gs->sock->sk->sk_family;
>
> - if (gs->collect_md) {
> - static u8 zero_vni[3];
> + if (sa_family == AF_INET) {
> + iph = ip_hdr(skb); /* outer IP header... */
>
> - vni = zero_vni;
> - addr = 0;
> - } else {
> - vni = gnvh->vni;
> - addr = iph->saddr;
> - }
> + if (gs->collect_md) {
> + vni = zero_vni;
> + addr = 0;
> + } else {
> + vni = gnvh->vni;
> +
> + addr = iph->saddr;
> + }
> +
> + geneve = geneve_lookup(gs, addr, vni);
> +#if IS_ENABLED(CONFIG_IPV6)
> + } else if (sa_family == AF_INET6) {
> + ip6h = ipv6_hdr(skb); /* outer IPv6 header... */
> +
> + if (gs->collect_md) {
> + vni = zero_vni;
> + addr6 = zero_addr6;
> + } else {
> + vni = gnvh->vni;
>
> - geneve = geneve_lookup(gs, addr, vni);
> + addr6 = ip6h->saddr;
> + }
> +
> + geneve = geneve6_lookup(gs, addr6, vni);
> +#endif
> + }
> if (!geneve)
> goto drop;
>
> @@ -149,7 +204,7 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
> (gnvh->oam ? TUNNEL_OAM : 0) |
> (gnvh->critical ? TUNNEL_CRIT_OPT : 0);
>
> - tun_dst = udp_tun_rx_dst(skb, AF_INET, flags,
> + tun_dst = udp_tun_rx_dst(skb, sa_family, flags,
> vni_to_tunnel_id(gnvh->vni),
> gnvh->opt_len * 4);
> if (!tun_dst)
> @@ -179,12 +234,21 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
>
> skb_reset_network_header(skb);
>
> - err = IP_ECN_decapsulate(iph, skb);
> + if (iph)
> + err = IP_ECN_decapsulate(iph, skb);
> + if (ip6h)
> + err = IP6_ECN_decapsulate(ip6h, skb);
>
> if (unlikely(err)) {
> - if (log_ecn_error)
> - net_info_ratelimited("non-ECT from %pI4 with TOS=%#x\n",
> - &iph->saddr, iph->tos);
> + if (log_ecn_error) {
> + if (iph)
> + net_info_ratelimited("non-ECT from %pI4 "
> + "with TOS=%#x\n",
> + &iph->saddr, iph->tos);
> + if (ip6h)
> + net_info_ratelimited("non-ECT from %pI6\n",
> + &ip6h->saddr);
> + }
> if (err > 1) {
> ++geneve->dev->stats.rx_frame_errors;
> ++geneve->dev->stats.rx_errors;
> @@ -284,6 +348,7 @@ static struct socket *geneve_create_sock(struct net *net, bool ipv6,
>
> if (ipv6) {
> udp_conf.family = AF_INET6;
> + udp_conf.ipv6_v6only = 1;
> } else {
> udp_conf.family = AF_INET;
> udp_conf.local_ip.s_addr = htonl(INADDR_ANY);
> @@ -458,9 +523,9 @@ static void geneve_notify_del_rx_port(struct geneve_sock *gs)
> udp_del_offload(&gs->udp_offloads);
> }
>
> -static void geneve_sock_release(struct geneve_sock *gs)
> +static void __geneve_sock_release(struct geneve_sock *gs)
> {
> - if (--gs->refcnt)
> + if (!gs || --gs->refcnt)
> return;
>
> list_del(&gs->list);
> @@ -469,58 +534,107 @@ static void geneve_sock_release(struct geneve_sock *gs)
> kfree_rcu(gs, rcu);
> }
>
> +static void geneve_sock_release(struct geneve_dev *geneve)
> +{
> + __geneve_sock_release(geneve->sock4);
> +#if IS_ENABLED(CONFIG_IPV6)
> + __geneve_sock_release(geneve->sock6);
> +#endif
> +}
> +
> static struct geneve_sock *geneve_find_sock(struct geneve_net *gn,
> + sa_family_t family,
> __be16 dst_port)
> {
> struct geneve_sock *gs;
>
> list_for_each_entry(gs, &gn->sock_list, list) {
> if (inet_sk(gs->sock->sk)->inet_sport == dst_port &&
> - inet_sk(gs->sock->sk)->sk.sk_family == AF_INET) {
> + inet_sk(gs->sock->sk)->sk.sk_family == family) {
> return gs;
> }
> }
> return NULL;
> }
>
> -static int geneve_open(struct net_device *dev)
> +static int geneve_sock_add(struct geneve_dev *geneve, bool ipv6)
> {
> - struct geneve_dev *geneve = netdev_priv(dev);
> struct net *net = geneve->net;
> struct geneve_net *gn = net_generic(net, geneve_net_id);
> struct geneve_sock *gs;
> __u32 hash;
>
> - gs = geneve_find_sock(gn, geneve->dst_port);
> + gs = geneve_find_sock(gn, ipv6 ? AF_INET6 : AF_INET, geneve->dst_port);
> if (gs) {
> gs->refcnt++;
> goto out;
> }
>
> - gs = geneve_socket_create(net, geneve->dst_port, false);
> + gs = geneve_socket_create(net, geneve->dst_port, ipv6);
> if (IS_ERR(gs))
> return PTR_ERR(gs);
>
> out:
> gs->collect_md = geneve->collect_md;
> - geneve->sock = gs;
> +#if IS_ENABLED(CONFIG_IPV6)
> + if (ipv6)
> + geneve->sock6 = gs;
> + else
> +#endif
> + geneve->sock4 = gs;
>
> hash = geneve_net_vni_hash(geneve->vni);
> hlist_add_head_rcu(&geneve->hlist, &gs->vni_list[hash]);
> return 0;
> }
>
> +static int geneve_open(struct net_device *dev)
> +{
> + struct geneve_dev *geneve = netdev_priv(dev);
> + bool ipv6 = geneve->remote.sa.sa_family == AF_INET6;
> + bool metadata = !!geneve->collect_md;
> + int ret = 0;
> +
> + geneve->sock4 = NULL;
> +#if IS_ENABLED(CONFIG_IPV6)
> + geneve->sock6 = NULL;
> + if (ipv6 || metadata)
> + ret = geneve_sock_add(geneve, true);
> +#endif
> + if (!ret && (!ipv6 || metadata))
> + ret = geneve_sock_add(geneve, false);
> + if (ret < 0)
> + geneve_sock_release(geneve);
> +
> + return ret;
> +}
> +
> static int geneve_stop(struct net_device *dev)
> {
> struct geneve_dev *geneve = netdev_priv(dev);
> - struct geneve_sock *gs = geneve->sock;
>
> if (!hlist_unhashed(&geneve->hlist))
> hlist_del_rcu(&geneve->hlist);
> - geneve_sock_release(gs);
> + geneve_sock_release(geneve);
> return 0;
> }
>
> +static void geneve_build_header(struct genevehdr *geneveh,
> + __be16 tun_flags, u8 vni[3],
> + u8 options_len, u8 *options)
> +{
> + geneveh->ver = GENEVE_VER;
> + geneveh->opt_len = options_len / 4;
> + geneveh->oam = !!(tun_flags & TUNNEL_OAM);
> + geneveh->critical = !!(tun_flags & TUNNEL_CRIT_OPT);
> + geneveh->rsvd1 = 0;
> + memcpy(geneveh->vni, vni, 3);
> + geneveh->proto_type = htons(ETH_P_TEB);
> + geneveh->rsvd2 = 0;
> +
> + memcpy(geneveh->options, options, options_len);
> +}
> +
> static int geneve_build_skb(struct rtable *rt, struct sk_buff *skb,
> __be16 tun_flags, u8 vni[3], u8 opt_len, u8 *opt,
> bool csum)
> @@ -544,15 +658,7 @@ static int geneve_build_skb(struct rtable *rt, struct sk_buff *skb,
> }
>
> gnvh = (struct genevehdr *)__skb_push(skb, sizeof(*gnvh) + opt_len);
> - gnvh->ver = GENEVE_VER;
> - gnvh->opt_len = opt_len / 4;
> - gnvh->oam = !!(tun_flags & TUNNEL_OAM);
> - gnvh->critical = !!(tun_flags & TUNNEL_CRIT_OPT);
> - gnvh->rsvd1 = 0;
> - memcpy(gnvh->vni, vni, 3);
> - gnvh->proto_type = htons(ETH_P_TEB);
> - gnvh->rsvd2 = 0;
> - memcpy(gnvh->options, opt, opt_len);
> + geneve_build_header(gnvh, tun_flags, vni, opt_len, opt);
>
> skb_set_inner_protocol(skb, htons(ETH_P_TEB));
> return 0;
> @@ -562,10 +668,47 @@ free_rt:
> return err;
> }
>
> -static struct rtable *geneve_get_rt(struct sk_buff *skb,
> - struct net_device *dev,
> - struct flowi4 *fl4,
> - struct ip_tunnel_info *info)
> +#if IS_ENABLED(CONFIG_IPV6)
> +static int geneve6_build_skb(struct dst_entry *dst, struct sk_buff *skb,
> + __be16 tun_flags, u8 vni[3], u8 opt_len, u8 *opt,
> + bool csum, bool xnet)
> +{
> + struct genevehdr *gnvh;
> + int min_headroom;
> + int err;
> +
> + skb_scrub_packet(skb, xnet);
> +
> + min_headroom = LL_RESERVED_SPACE(dst->dev) + dst->header_len
> + + GENEVE_BASE_HLEN + opt_len + sizeof(struct ipv6hdr);
> + err = skb_cow_head(skb, min_headroom);
> + if (unlikely(err)) {
> + kfree_skb(skb);
> + goto free_dst;
> + }
> +
> + skb = udp_tunnel_handle_offloads(skb, csum);
> + if (IS_ERR(skb)) {
> + err = PTR_ERR(skb);
> + goto free_dst;
> + }
> +
> + gnvh = (struct genevehdr *)__skb_push(skb, sizeof(*gnvh) + opt_len);
> + geneve_build_header(gnvh, tun_flags, vni, opt_len, opt);
> +
> + skb_set_inner_protocol(skb, htons(ETH_P_TEB));
> + return 0;
> +
> +free_dst:
> + dst_release(dst);
> + return err;
> +}
> +#endif
> +
> +static struct rtable *geneve_get_v4_rt(struct sk_buff *skb,
> + struct net_device *dev,
> + struct flowi4 *fl4,
> + struct ip_tunnel_info *info)
> {
> struct geneve_dev *geneve = netdev_priv(dev);
> struct rtable *rt = NULL;
> @@ -588,24 +731,42 @@ static struct rtable *geneve_get_rt(struct sk_buff *skb,
> }
>
> fl4->flowi4_tos = RT_TOS(tos);
> - fl4->daddr = geneve->remote.sin_addr.s_addr;
> + fl4->daddr = geneve->remote.sin.sin_addr.s_addr;
> }
>
> rt = ip_route_output_key(geneve->net, fl4);
> - if (IS_ERR(rt)) {
> - netdev_dbg(dev, "no route to %pI4\n", &fl4->daddr);
> - dev->stats.tx_carrier_errors++;
> - return rt;
> - }
> - if (rt->dst.dev == dev) { /* is this necessary? */
> - netdev_dbg(dev, "circular route to %pI4\n", &fl4->daddr);
> - dev->stats.collisions++;
> - ip_rt_put(rt);
> - return ERR_PTR(-EINVAL);
> - }
> +
> return rt;
> }
>
> +#if IS_ENABLED(CONFIG_IPV6)
> +static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
> + struct net_device *dev,
> + struct flowi6 *fl6,
> + struct ip_tunnel_info *info)
> +{
> + struct geneve_dev *geneve = netdev_priv(dev);
> + struct geneve_sock *gs6 = geneve->sock6;
> + struct dst_entry *dst = NULL;
> +
> + memset(fl6, 0, sizeof(*fl6));
> + fl6->flowi6_mark = skb->mark;
> + fl6->flowi6_proto = IPPROTO_UDP;
> +
> + if (info) {
> + fl6->daddr = info->key.u.ipv6.dst;
> + fl6->saddr = info->key.u.ipv6.src;
> + } else {
> + fl6->daddr = geneve->remote.sin6.sin6_addr;
> + }
> +
> + if (ipv6_stub->ipv6_dst_lookup(geneve->net, gs6->sock->sk, &dst, fl6))
> + return ERR_PTR(-EHOSTUNREACH);
> +
> + return dst;
> +}
> +#endif
> +
> /* Convert 64 bit tunnel ID to 24 bit VNI. */
> static void tunnel_id_to_vni(__be64 tun_id, __u8 *vni)
> {
> @@ -620,11 +781,11 @@ static void tunnel_id_to_vni(__be64 tun_id, __u8 *vni)
> #endif
> }
>
> -static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
> +static netdev_tx_t geneve_xmit_skb(struct sk_buff *skb, struct net_device *dev,
> + struct ip_tunnel_info *info)
> {
> struct geneve_dev *geneve = netdev_priv(dev);
> - struct geneve_sock *gs = geneve->sock;
> - struct ip_tunnel_info *info = NULL;
> + struct geneve_sock *gs4 = geneve->sock4;
> struct rtable *rt = NULL;
> const struct iphdr *iip; /* interior IP header */
> struct flowi4 fl4;
> @@ -635,7 +796,6 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
> int err;
>
> if (geneve->collect_md) {
> - info = skb_tunnel_info(skb);
> if (unlikely(info && !(info->mode & IP_TUNNEL_INFO_TX))) {
> netdev_dbg(dev, "no tunnel metadata\n");
> goto tx_error;
> @@ -644,12 +804,18 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
> goto tx_error;
> }
>
> - rt = geneve_get_rt(skb, dev, &fl4, info);
> + rt = geneve_get_v4_rt(skb, dev, &fl4, info);
> if (IS_ERR(rt)) {
> netdev_dbg(dev, "no route to %pI4\n", &fl4.daddr);
> dev->stats.tx_carrier_errors++;
> goto tx_error;
> }
> + if (rt->dst.dev == dev) { /* is this necessary? */
> + netdev_dbg(dev, "circular route to %pI4\n", &fl4.daddr);
> + dev->stats.collisions++;
> + ip_rt_put(rt);
> + goto tx_error;
> + }
>
> sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
> skb_reset_mac_header(skb);
> @@ -688,7 +854,7 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
> ttl = ttl ? : ip4_dst_hoplimit(&rt->dst);
> df = 0;
> }
> - err = udp_tunnel_xmit_skb(rt, gs->sock->sk, skb, fl4.saddr, fl4.daddr,
> + err = udp_tunnel_xmit_skb(rt, gs4->sock->sk, skb, fl4.saddr, fl4.daddr,
> tos, ttl, df, sport, geneve->dst_port,
> !net_eq(geneve->net, dev_net(geneve->dev)),
> !udp_csum);
> @@ -703,6 +869,103 @@ err:
> return NETDEV_TX_OK;
> }
>
> +#if IS_ENABLED(CONFIG_IPV6)
> +static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
> + struct ip_tunnel_info *info)
> +{
> + struct geneve_dev *geneve = netdev_priv(dev);
> + struct geneve_sock *gs6 = geneve->sock6;
> + struct dst_entry *dst = NULL;
> + struct flowi6 fl6;
> + __u8 ttl;
> + __be16 sport;
> + bool udp_csum;
> + int err;
> + bool xnet = !net_eq(geneve->net, dev_net(geneve->dev));
> +
> + if (geneve->collect_md) {
> + if (unlikely(info && !(info->mode & IP_TUNNEL_INFO_TX))) {
> + netdev_dbg(dev, "no tunnel metadata\n");
> + goto tx_error;
> + }
> + }
> +
> + dst = geneve_get_v6_dst(skb, dev, &fl6, info);
> + if (IS_ERR(dst)) {
> + netdev_dbg(dev, "no route to %pI6\n", &fl6.daddr);
> + dev->stats.tx_carrier_errors++;
> + goto tx_error;
> + }
> + if (dst->dev == dev) { /* is this necessary? */
> + netdev_dbg(dev, "circular route to %pI6\n", &fl6.daddr);
> + dev->stats.collisions++;
> + dst_release(dst);
> + goto tx_error;
> + }
> +
> + sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
> + skb_reset_mac_header(skb);
> +
> + if (info) {
> + const struct ip_tunnel_key *key = &info->key;
> + u8 *opts = NULL;
> + u8 vni[3];
> +
> + tunnel_id_to_vni(key->tun_id, vni);
> + if (key->tun_flags & TUNNEL_GENEVE_OPT)
> + opts = ip_tunnel_info_opts(info);
> +
> + udp_csum = !!(key->tun_flags & TUNNEL_CSUM);
> + err = geneve6_build_skb(dst, skb, key->tun_flags, vni,
> + info->options_len, opts,
> + udp_csum, xnet);
> + if (unlikely(err))
> + goto err;
> +
> + ttl = key->ttl;
> + } else {
> + udp_csum = false;
> + err = geneve6_build_skb(dst, skb, 0, geneve->vni,
> + 0, NULL, udp_csum, xnet);
> + if (unlikely(err))
> + goto err;
> +
> + ttl = geneve->ttl;
> + if (!ttl && ipv6_addr_is_multicast(&fl6.daddr))
> + ttl = 1;
> + ttl = ttl ? : ip6_dst_hoplimit(dst);
> + }
> + err = udp_tunnel6_xmit_skb(dst, gs6->sock->sk, skb, dev,
> + &fl6.saddr, &fl6.daddr, 0, ttl,
> + sport, geneve->dst_port, !udp_csum);
> +
> + iptunnel_xmit_stats(err, &dev->stats, dev->tstats);
> + return NETDEV_TX_OK;
> +
> +tx_error:
> + dev_kfree_skb(skb);
> +err:
> + dev->stats.tx_errors++;
> + return NETDEV_TX_OK;
> +}
> +#endif
> +
> +static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
> +{
> + struct geneve_dev *geneve = netdev_priv(dev);
> + struct ip_tunnel_info *info = NULL;
> +
> + if (geneve->collect_md)
> + info = skb_tunnel_info(skb);
> +
> +#if IS_ENABLED(CONFIG_IPV6)
> + if ((info && ip_tunnel_info_af(info) == AF_INET6) ||
> + (!info && geneve->remote.sa.sa_family == AF_INET6))
> + return geneve6_xmit_skb(skb, dev, info);
> +#endif
> + return geneve_xmit_skb(skb, dev, info);
> +}
> +
> static const struct net_device_ops geneve_netdev_ops = {
> .ndo_init = geneve_init,
> .ndo_uninit = geneve_uninit,
> @@ -759,6 +1022,7 @@ static void geneve_setup(struct net_device *dev)
> static const struct nla_policy geneve_policy[IFLA_GENEVE_MAX + 1] = {
> [IFLA_GENEVE_ID] = { .type = NLA_U32 },
> [IFLA_GENEVE_REMOTE] = { .len = FIELD_SIZEOF(struct iphdr, daddr) },
> + [IFLA_GENEVE_REMOTE6] = { .len = sizeof(struct in6_addr) },
> [IFLA_GENEVE_TTL] = { .type = NLA_U8 },
> [IFLA_GENEVE_TOS] = { .type = NLA_U8 },
> [IFLA_GENEVE_PORT] = { .type = NLA_U16 },
> @@ -790,7 +1054,7 @@ static int geneve_validate(struct nlattr *tb[], struct nlattr *data[])
>
> static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
> __be16 dst_port,
> - __be32 rem_addr,
> + union geneve_addr *remote,
> u8 vni[],
> bool *tun_on_same_port,
> bool *tun_collect_md)
> @@ -806,7 +1070,7 @@ static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
> *tun_on_same_port = true;
> }
> if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
> - rem_addr == geneve->remote.sin_addr.s_addr &&
> + !memcmp(remote, &geneve->remote, sizeof(geneve->remote)) &&
> dst_port == geneve->dst_port)
> t = geneve;
> }
> @@ -814,18 +1078,20 @@ static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
> }
>
> static int geneve_configure(struct net *net, struct net_device *dev,
> - __be32 rem_addr, __u32 vni, __u8 ttl, __u8 tos,
> - __be16 dst_port, bool metadata)
> + union geneve_addr *remote,
> + __u32 vni, __u8 ttl, __u8 tos, __be16 dst_port,
> + bool metadata)
> {
> struct geneve_net *gn = net_generic(net, geneve_net_id);
> struct geneve_dev *t, *geneve = netdev_priv(dev);
> bool tun_collect_md, tun_on_same_port;
> int err;
>
> - if (metadata) {
> - if (rem_addr || vni || tos || ttl)
> - return -EINVAL;
> - }
> + if (!remote)
> + return -EINVAL;
> + if (metadata &&
> + (remote->sa.sa_family != AF_UNSPEC || vni || tos || ttl))
> + return -EINVAL;
>
> geneve->net = net;
> geneve->dev = dev;
> @@ -834,16 +1100,19 @@ static int geneve_configure(struct net *net, struct net_device *dev,
> geneve->vni[1] = (vni & 0x0000ff00) >> 8;
> geneve->vni[2] = vni & 0x000000ff;
>
> - geneve->remote.sin_addr.s_addr = rem_addr;
> - if (IN_MULTICAST(ntohl(geneve->remote.sin_addr.s_addr)))
> + if ((remote->sa.sa_family == AF_INET &&
> + IN_MULTICAST(ntohl(remote->sin.sin_addr.s_addr))) ||
> + (remote->sa.sa_family == AF_INET6 &&
> + ipv6_addr_is_multicast(&remote->sin6.sin6_addr)))
> return -EINVAL;
> + geneve->remote = *remote;
>
> geneve->ttl = ttl;
> geneve->tos = tos;
> geneve->dst_port = dst_port;
> geneve->collect_md = metadata;
>
> - t = geneve_find_dev(gn, dst_port, rem_addr, geneve->vni,
> + t = geneve_find_dev(gn, dst_port, remote, geneve->vni,
> &tun_on_same_port, &tun_collect_md);
> if (t)
> return -EBUSY;
> @@ -870,14 +1139,29 @@ static int geneve_newlink(struct net *net, struct net_device *dev,
> __be16 dst_port = htons(GENEVE_UDP_PORT);
> __u8 ttl = 0, tos = 0;
> bool metadata = false;
> - __be32 rem_addr;
> + union geneve_addr remote;
> __u32 vni;
>
> - if (!data[IFLA_GENEVE_ID] || !data[IFLA_GENEVE_REMOTE])
> + if (!data[IFLA_GENEVE_ID] ||
> + (data[IFLA_GENEVE_REMOTE] && data[IFLA_GENEVE_REMOTE6]) ||
> + (!data[IFLA_GENEVE_REMOTE] && !data[IFLA_GENEVE_REMOTE6]))
> return -EINVAL;
>
> vni = nla_get_u32(data[IFLA_GENEVE_ID]);
> - rem_addr = nla_get_in_addr(data[IFLA_GENEVE_REMOTE]);
> +
> + memset(&remote, 0, sizeof(remote));
> + if (data[IFLA_GENEVE_REMOTE]) {
> + remote.sa.sa_family = AF_INET;
> + remote.sin.sin_addr.s_addr =
> + nla_get_in_addr(data[IFLA_GENEVE_REMOTE]);
> + } else if (data[IFLA_GENEVE_REMOTE6]) {
> + if (!IS_ENABLED(CONFIG_IPV6))
> + return -EPFNOSUPPORT;
> +
> + remote.sa.sa_family = AF_INET6;
> + remote.sin6.sin6_addr =
> + nla_get_in6_addr(data[IFLA_GENEVE_REMOTE6]);
> + }
>
> if (data[IFLA_GENEVE_TTL])
> ttl = nla_get_u8(data[IFLA_GENEVE_TTL]);
> @@ -891,8 +1175,8 @@ static int geneve_newlink(struct net *net, struct net_device *dev,
> if (data[IFLA_GENEVE_COLLECT_METADATA])
> metadata = true;
>
> - return geneve_configure(net, dev, rem_addr, vni,
> - ttl, tos, dst_port, metadata);
> + return geneve_configure(net, dev, &remote, vni, ttl, tos, dst_port,
> + metadata);
> }
>
> static void geneve_dellink(struct net_device *dev, struct list_head *head)
> @@ -906,7 +1190,7 @@ static void geneve_dellink(struct net_device *dev, struct list_head *head)
> static size_t geneve_get_size(const struct net_device *dev)
> {
> return nla_total_size(sizeof(__u32)) + /* IFLA_GENEVE_ID */
> - nla_total_size(sizeof(struct in_addr)) + /* IFLA_GENEVE_REMOTE */
> + nla_total_size(sizeof(struct in6_addr)) + /* IFLA_GENEVE_REMOTE{6} */
> nla_total_size(sizeof(__u8)) + /* IFLA_GENEVE_TTL */
> nla_total_size(sizeof(__u8)) + /* IFLA_GENEVE_TOS */
> nla_total_size(sizeof(__be16)) + /* IFLA_GENEVE_PORT */
> @@ -923,9 +1207,17 @@ static int geneve_fill_info(struct sk_buff *skb, const struct net_device *dev)
> if (nla_put_u32(skb, IFLA_GENEVE_ID, vni))
> goto nla_put_failure;
>
> - if (nla_put_in_addr(skb, IFLA_GENEVE_REMOTE,
> - geneve->remote.sin_addr.s_addr))
> - goto nla_put_failure;
> + if (geneve->remote.sa.sa_family == AF_INET) {
> + if (nla_put_in_addr(skb, IFLA_GENEVE_REMOTE,
> + geneve->remote.sin.sin_addr.s_addr))
> + goto nla_put_failure;
> +#if IS_ENABLED(CONFIG_IPV6)
> + } else {
> + if (nla_put_in6_addr(skb, IFLA_GENEVE_REMOTE6,
> + &geneve->remote.sin6.sin6_addr))
> + goto nla_put_failure;
> +#endif
> + }
>
> if (nla_put_u8(skb, IFLA_GENEVE_TTL, geneve->ttl) ||
> nla_put_u8(skb, IFLA_GENEVE_TOS, geneve->tos))
> @@ -971,7 +1263,8 @@ struct net_device *geneve_dev_create_fb(struct net *net, const char *name,
> if (IS_ERR(dev))
> return dev;
>
> - err = geneve_configure(net, dev, 0, 0, 0, 0, htons(dst_port), true);
> + err = geneve_configure(net, dev, &geneve_remote_unspec,
> + 0, 0, 0, htons(dst_port), true);
> if (err) {
> free_netdev(dev);
> return ERR_PTR(err);
> diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
> index e3b6217f34f1..45e3a48550f9 100644
> --- a/include/uapi/linux/if_link.h
> +++ b/include/uapi/linux/if_link.h
> @@ -461,6 +461,7 @@ enum {
> IFLA_GENEVE_TOS,
> IFLA_GENEVE_PORT, /* destination port */
> IFLA_GENEVE_COLLECT_METADATA,
> + IFLA_GENEVE_REMOTE6,
> __IFLA_GENEVE_MAX
> };
> #define IFLA_GENEVE_MAX (__IFLA_GENEVE_MAX - 1)
>
--
吉藤英明 <hideaki.yoshifuji@miraclelinux.com>
ミラクル・リナックス株式会社 技術本部 サポート部
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH v4 1/2] geneve: implement support for IPv6-based tunnels
2015-10-21 1:52 ` YOSHIFUJI Hideaki/吉藤英明
@ 2015-10-21 18:58 ` John W. Linville
0 siblings, 0 replies; 37+ messages in thread
From: John W. Linville @ 2015-10-21 18:58 UTC (permalink / raw)
To: YOSHIFUJI Hideaki/吉藤英明
Cc: netdev, Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc
On Wed, Oct 21, 2015 at 10:52:43AM +0900, YOSHIFUJI Hideaki/吉藤英明 wrote:
> Hi,
Yoshifuji-san -- thank you for taking a look at my proposed patch!
> John W. Linville wrote:
> > Signed-off-by: John W. Linville <linville@tuxdriver.com>
> > ---
> > v4:
> > - treat mode field of ip_tunnel_info as flags
> > - add a missing IS_ENABLED(CONFIG_IPV6) to geneve_rx
> > - remove unneeded flags field in geneve_dev
> > - NULL-check parameter for __geneve_sock_release
> > - check remote socket family for AF_UNSPEC in geneve_configure
> > - rename geneve_get_{rt,dst} as geneve_get_{v4_rt,v6_dst}
> > - refactor some error handling in the xmit paths
> >
> > v3:
> > - declare geneve_remote_unspec as static
> >
> > v2:
> > - do not require remote address for tx on metadata tunnels
> > - pass correct sockaddr family to udp_tun_rx_dst in geneve_rx
> > - accommodate both ipv4 and ipv6 sockets open on same tunnel
> > - move declaration of geneve_get_dst for aesthetic purposes
> >
> > drivers/net/geneve.c | 459 +++++++++++++++++++++++++++++++++++--------
> > include/uapi/linux/if_link.h | 1 +
> > 2 files changed, 377 insertions(+), 83 deletions(-)
> >
> > diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
> > index 8f5c02eed47d..217b472ab9e7 100644
> > --- a/drivers/net/geneve.c
> > +++ b/drivers/net/geneve.c
> > @@ -46,16 +46,25 @@ struct geneve_net {
> >
> > static int geneve_net_id;
> >
> > +union geneve_addr {
> > + struct sockaddr_in sin;
> > + struct sockaddr_in6 sin6;
> > + struct sockaddr sa;
> > +};
> > +
> > +static union geneve_addr geneve_remote_unspec = { .sa.sa_family = AF_UNSPEC, };
> > +
> > /* Pseudo network device */
> > struct geneve_dev {
> > struct hlist_node hlist; /* vni hash table */
> > struct net *net; /* netns for packet i/o */
> > struct net_device *dev; /* netdev for geneve tunnel */
> > - struct geneve_sock *sock; /* socket used for geneve tunnel */
> > + struct geneve_sock *sock4; /* IPv4 socket used for geneve tunnel */
> > + struct geneve_sock *sock6; /* IPv6 socket used for geneve tunnel */
> > u8 vni[3]; /* virtual network ID for tunnel */
> > u8 ttl; /* TTL override */
> > u8 tos; /* TOS override */
> > - struct sockaddr_in remote; /* IPv4 address for link partner */
> > + union geneve_addr remote; /* IP address for link partner */
> > struct list_head next; /* geneve's per namespace list */
> > __be16 dst_port;
> > bool collect_md;
> > @@ -103,11 +112,32 @@ static struct geneve_dev *geneve_lookup(struct geneve_sock *gs,
> > vni_list_head = &gs->vni_list[hash];
> > hlist_for_each_entry_rcu(geneve, vni_list_head, hlist) {
> > if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
> > - addr == geneve->remote.sin_addr.s_addr)
> > + addr == geneve->remote.sin.sin_addr.s_addr)
> > + return geneve;
> > + }
> > + return NULL;
> > +}
> > +
> > +#if IS_ENABLED(CONFIG_IPV6)
> > +static struct geneve_dev *geneve6_lookup(struct geneve_sock *gs,
> > + struct in6_addr addr6, u8 vni[])
> > +{
> > + struct hlist_head *vni_list_head;
> > + struct geneve_dev *geneve;
> > + __u32 hash;
> > +
> > + /* Find the device for this VNI */
> > + hash = geneve_net_vni_hash(vni);
> > + vni_list_head = &gs->vni_list[hash];
> > + hlist_for_each_entry_rcu(geneve, vni_list_head, hlist) {
> > + if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
> > + !memcmp(&addr6, &geneve->remote.sin6.sin6_addr,
> > + sizeof(addr6)))
>
> Please use ipv6_addr_equal().
Sure, no problem.
> How do you handle link-local addresses here?
Hmmmm...TBH, I had completely overlooked link-local addresses.
Do you have any suggestions for how to address this shortcoming?
Thanks,
John
--
John W. Linville Someday the world will need a hero, and you
linville@tuxdriver.com might be all we have. Be ready.
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH v4 1/2] geneve: implement support for IPv6-based tunnels
2015-10-20 15:11 ` [PATCH v4 1/2] " John W. Linville
` (2 preceding siblings ...)
2015-10-21 1:52 ` YOSHIFUJI Hideaki/吉藤英明
@ 2015-10-21 5:06 ` Jesse Gross
2015-10-22 19:45 ` [PATCH v5 " John W. Linville
4 siblings, 0 replies; 37+ messages in thread
From: Jesse Gross @ 2015-10-21 5:06 UTC (permalink / raw)
To: John W. Linville; +Cc: netdev, Dave Miller, Pravin B Shelar, Jiri Benc
All of this looks pretty good to me, I just have a few last things
that I noticed:
On Tue, Oct 20, 2015 at 11:11 PM, John W. Linville
<linville@tuxdriver.com> wrote:
> diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
> index 8f5c02eed47d..217b472ab9e7 100644
> --- a/drivers/net/geneve.c
> +++ b/drivers/net/geneve.c
> /* Pseudo network device */
> struct geneve_dev {
> struct hlist_node hlist; /* vni hash table */
> struct net *net; /* netns for packet i/o */
> struct net_device *dev; /* netdev for geneve tunnel */
> - struct geneve_sock *sock; /* socket used for geneve tunnel */
> + struct geneve_sock *sock4; /* IPv4 socket used for geneve tunnel */
> + struct geneve_sock *sock6; /* IPv6 socket used for geneve tunnel */
It might be worth wrapping sock6 in #if IS_ENABLED(CONFIG_IPV6). Not
because I'm trying to save space but because we make initializing it
conditional on this, so it would catch if somebody tries to access it
uninitialized.
> +static int geneve_open(struct net_device *dev)
> +{
> + struct geneve_dev *geneve = netdev_priv(dev);
> + bool ipv6 = geneve->remote.sa.sa_family == AF_INET6;
> + bool metadata = !!geneve->collect_md;
A small thing but geneve->collect_md is also a bool, so I guess we
probably don't need the !!.
> +#if IS_ENABLED(CONFIG_IPV6)
> +static int geneve6_build_skb(struct dst_entry *dst, struct sk_buff *skb,
> + __be16 tun_flags, u8 vni[3], u8 opt_len, u8 *opt,
> + bool csum, bool xnet)
> +{
[...]
> + skb_scrub_packet(skb, xnet);
I realized that I wasn't really all that clear with my previous
comment here. I was just asking if we should also use
skb_scrub_packet() in the IPv4 to keep things consistent. I agree that
the rt/dst differences makes sharing code a bit difficult in the
general sense.
> +static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
> + struct ip_tunnel_info *info)
[...]
> + if (geneve->collect_md) {
> + if (unlikely(info && !(info->mode & IP_TUNNEL_INFO_TX))) {
> + netdev_dbg(dev, "no tunnel metadata\n");
> + goto tx_error;
It seems like this should also be checking for !info here - that's
perhaps the main cause of not having any tunnel metadata. This is the
same with IPv4 though too.
> @@ -870,14 +1139,29 @@ static int geneve_newlink(struct net *net, struct net_device *dev,
> - if (!data[IFLA_GENEVE_ID] || !data[IFLA_GENEVE_REMOTE])
> + if (!data[IFLA_GENEVE_ID] ||
> + (data[IFLA_GENEVE_REMOTE] && data[IFLA_GENEVE_REMOTE6]) ||
> + (!data[IFLA_GENEVE_REMOTE] && !data[IFLA_GENEVE_REMOTE6]))
> return -EINVAL;
I think this will conflict with/revert my change in -net. Obviously,
the conflict will need to be resolved at some point, but it's probably
just best to remove the whole block here so the resolution is obvious.
^ permalink raw reply [flat|nested] 37+ messages in thread
* [PATCH v5 1/2] geneve: implement support for IPv6-based tunnels
2015-10-20 15:11 ` [PATCH v4 1/2] " John W. Linville
` (3 preceding siblings ...)
2015-10-21 5:06 ` Jesse Gross
@ 2015-10-22 19:45 ` John W. Linville
2015-10-22 19:45 ` [PATCH v5 2/2] geneve: handle ipv6 priority like ipv4 tos John W. Linville
` (2 more replies)
4 siblings, 3 replies; 37+ messages in thread
From: John W. Linville @ 2015-10-22 19:45 UTC (permalink / raw)
To: netdev
Cc: Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc,
YOSHIFUJI Hideaki/吉藤英明,
John W. Linville
NOTE: Link-local IPv6 addresses for remote endpoints are not supported,
since the driver currently has no capacity for binding a geneve
interface to a specific link.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
---
v5:
- wrap declaration of sock6 in geneve_dev with IS_ENABLED(CONFIG_IPV6)
- remove superfluous '!!' when assigning geneve->collect_md to bool
- use skb_scrub_packet in IPv4 tx path as well
- check for NULL ip_tunnel_info pointer in geneve[6]_xmit_skb
- use ipv6_addr_equal for comparing IPv6 addresses
- more use of IS_ENABLED(CONFIG_IPV6) for preserving build integrity
- reject link-local ipv6 address for remote tunnel endpoint
v4:
- treat mode field of ip_tunnel_info as flags
- add a missing IS_ENABLED(CONFIG_IPV6) to geneve_rx
- remove unneeded flags field in geneve_dev
- NULL-check parameter for __geneve_sock_release
- check remote socket family for AF_UNSPEC in geneve_configure
- rename geneve_get_{rt,dst} as geneve_get_{v4_rt,v6_dst}
- refactor some error handling in the xmit paths
v3:
- declare geneve_remote_unspec as static
v2:
- do not require remote address for tx on metadata tunnels
- pass correct sockaddr family to udp_tun_rx_dst in geneve_rx
- accommodate both ipv4 and ipv6 sockets open on same tunnel
- move declaration of geneve_get_dst for aesthetic purposes
drivers/net/geneve.c | 482 +++++++++++++++++++++++++++++++++++--------
include/uapi/linux/if_link.h | 1 +
2 files changed, 395 insertions(+), 88 deletions(-)
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index cde29f8a37bf..47f7512f02cf 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -46,16 +46,27 @@ struct geneve_net {
static int geneve_net_id;
+union geneve_addr {
+ struct sockaddr_in sin;
+ struct sockaddr_in6 sin6;
+ struct sockaddr sa;
+};
+
+static union geneve_addr geneve_remote_unspec = { .sa.sa_family = AF_UNSPEC, };
+
/* Pseudo network device */
struct geneve_dev {
struct hlist_node hlist; /* vni hash table */
struct net *net; /* netns for packet i/o */
struct net_device *dev; /* netdev for geneve tunnel */
- struct geneve_sock *sock; /* socket used for geneve tunnel */
+ struct geneve_sock *sock4; /* IPv4 socket used for geneve tunnel */
+#if IS_ENABLED(CONFIG_IPV6)
+ struct geneve_sock *sock6; /* IPv6 socket used for geneve tunnel */
+#endif
u8 vni[3]; /* virtual network ID for tunnel */
u8 ttl; /* TTL override */
u8 tos; /* TOS override */
- struct sockaddr_in remote; /* IPv4 address for link partner */
+ union geneve_addr remote; /* IP address for link partner */
struct list_head next; /* geneve's per namespace list */
__be16 dst_port;
bool collect_md;
@@ -103,12 +114,32 @@ static struct geneve_dev *geneve_lookup(struct geneve_sock *gs,
vni_list_head = &gs->vni_list[hash];
hlist_for_each_entry_rcu(geneve, vni_list_head, hlist) {
if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
- addr == geneve->remote.sin_addr.s_addr)
+ addr == geneve->remote.sin.sin_addr.s_addr)
return geneve;
}
return NULL;
}
+#if IS_ENABLED(CONFIG_IPV6)
+static struct geneve_dev *geneve6_lookup(struct geneve_sock *gs,
+ struct in6_addr addr6, u8 vni[])
+{
+ struct hlist_head *vni_list_head;
+ struct geneve_dev *geneve;
+ __u32 hash;
+
+ /* Find the device for this VNI */
+ hash = geneve_net_vni_hash(vni);
+ vni_list_head = &gs->vni_list[hash];
+ hlist_for_each_entry_rcu(geneve, vni_list_head, hlist) {
+ if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
+ ipv6_addr_equal(&addr6, &geneve->remote.sin6.sin6_addr))
+ return geneve;
+ }
+ return NULL;
+}
+#endif
+
static inline struct genevehdr *geneve_hdr(const struct sk_buff *skb)
{
return (struct genevehdr *)(udp_hdr(skb) + 1);
@@ -121,24 +152,49 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
struct metadata_dst *tun_dst = NULL;
struct geneve_dev *geneve = NULL;
struct pcpu_sw_netstats *stats;
- struct iphdr *iph;
- u8 *vni;
+ struct iphdr *iph = NULL;
__be32 addr;
- int err;
+ static u8 zero_vni[3];
+ u8 *vni;
+ int err = 0;
+ sa_family_t sa_family;
+#if IS_ENABLED(CONFIG_IPV6)
+ struct ipv6hdr *ip6h = NULL;
+ struct in6_addr addr6;
+ static struct in6_addr zero_addr6;
+#endif
+
+ sa_family = gs->sock->sk->sk_family;
- iph = ip_hdr(skb); /* outer IP header... */
+ if (sa_family == AF_INET) {
+ iph = ip_hdr(skb); /* outer IP header... */
- if (gs->collect_md) {
- static u8 zero_vni[3];
+ if (gs->collect_md) {
+ vni = zero_vni;
+ addr = 0;
+ } else {
+ vni = gnvh->vni;
- vni = zero_vni;
- addr = 0;
- } else {
- vni = gnvh->vni;
- addr = iph->saddr;
- }
+ addr = iph->saddr;
+ }
- geneve = geneve_lookup(gs, addr, vni);
+ geneve = geneve_lookup(gs, addr, vni);
+#if IS_ENABLED(CONFIG_IPV6)
+ } else if (sa_family == AF_INET6) {
+ ip6h = ipv6_hdr(skb); /* outer IPv6 header... */
+
+ if (gs->collect_md) {
+ vni = zero_vni;
+ addr6 = zero_addr6;
+ } else {
+ vni = gnvh->vni;
+
+ addr6 = ip6h->saddr;
+ }
+
+ geneve = geneve6_lookup(gs, addr6, vni);
+#endif
+ }
if (!geneve)
goto drop;
@@ -149,7 +205,7 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
(gnvh->oam ? TUNNEL_OAM : 0) |
(gnvh->critical ? TUNNEL_CRIT_OPT : 0);
- tun_dst = udp_tun_rx_dst(skb, AF_INET, flags,
+ tun_dst = udp_tun_rx_dst(skb, sa_family, flags,
vni_to_tunnel_id(gnvh->vni),
gnvh->opt_len * 4);
if (!tun_dst)
@@ -179,12 +235,25 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
skb_reset_network_header(skb);
- err = IP_ECN_decapsulate(iph, skb);
+ if (iph)
+ err = IP_ECN_decapsulate(iph, skb);
+#if IS_ENABLED(CONFIG_IPV6)
+ if (ip6h)
+ err = IP6_ECN_decapsulate(ip6h, skb);
+#endif
if (unlikely(err)) {
- if (log_ecn_error)
- net_info_ratelimited("non-ECT from %pI4 with TOS=%#x\n",
- &iph->saddr, iph->tos);
+ if (log_ecn_error) {
+ if (iph)
+ net_info_ratelimited("non-ECT from %pI4 "
+ "with TOS=%#x\n",
+ &iph->saddr, iph->tos);
+#if IS_ENABLED(CONFIG_IPV6)
+ if (ip6h)
+ net_info_ratelimited("non-ECT from %pI6\n",
+ &ip6h->saddr);
+#endif
+ }
if (err > 1) {
++geneve->dev->stats.rx_frame_errors;
++geneve->dev->stats.rx_errors;
@@ -284,6 +353,7 @@ static struct socket *geneve_create_sock(struct net *net, bool ipv6,
if (ipv6) {
udp_conf.family = AF_INET6;
+ udp_conf.ipv6_v6only = 1;
} else {
udp_conf.family = AF_INET;
udp_conf.local_ip.s_addr = htonl(INADDR_ANY);
@@ -458,9 +528,9 @@ static void geneve_notify_del_rx_port(struct geneve_sock *gs)
udp_del_offload(&gs->udp_offloads);
}
-static void geneve_sock_release(struct geneve_sock *gs)
+static void __geneve_sock_release(struct geneve_sock *gs)
{
- if (--gs->refcnt)
+ if (!gs || --gs->refcnt)
return;
list_del(&gs->list);
@@ -469,66 +539,117 @@ static void geneve_sock_release(struct geneve_sock *gs)
kfree_rcu(gs, rcu);
}
+static void geneve_sock_release(struct geneve_dev *geneve)
+{
+ __geneve_sock_release(geneve->sock4);
+#if IS_ENABLED(CONFIG_IPV6)
+ __geneve_sock_release(geneve->sock6);
+#endif
+}
+
static struct geneve_sock *geneve_find_sock(struct geneve_net *gn,
+ sa_family_t family,
__be16 dst_port)
{
struct geneve_sock *gs;
list_for_each_entry(gs, &gn->sock_list, list) {
if (inet_sk(gs->sock->sk)->inet_sport == dst_port &&
- inet_sk(gs->sock->sk)->sk.sk_family == AF_INET) {
+ inet_sk(gs->sock->sk)->sk.sk_family == family) {
return gs;
}
}
return NULL;
}
-static int geneve_open(struct net_device *dev)
+static int geneve_sock_add(struct geneve_dev *geneve, bool ipv6)
{
- struct geneve_dev *geneve = netdev_priv(dev);
struct net *net = geneve->net;
struct geneve_net *gn = net_generic(net, geneve_net_id);
struct geneve_sock *gs;
__u32 hash;
- gs = geneve_find_sock(gn, geneve->dst_port);
+ gs = geneve_find_sock(gn, ipv6 ? AF_INET6 : AF_INET, geneve->dst_port);
if (gs) {
gs->refcnt++;
goto out;
}
- gs = geneve_socket_create(net, geneve->dst_port, false);
+ gs = geneve_socket_create(net, geneve->dst_port, ipv6);
if (IS_ERR(gs))
return PTR_ERR(gs);
out:
gs->collect_md = geneve->collect_md;
- geneve->sock = gs;
+#if IS_ENABLED(CONFIG_IPV6)
+ if (ipv6)
+ geneve->sock6 = gs;
+ else
+#endif
+ geneve->sock4 = gs;
hash = geneve_net_vni_hash(geneve->vni);
hlist_add_head_rcu(&geneve->hlist, &gs->vni_list[hash]);
return 0;
}
+static int geneve_open(struct net_device *dev)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ bool ipv6 = geneve->remote.sa.sa_family == AF_INET6;
+ bool metadata = geneve->collect_md;
+ int ret = 0;
+
+ geneve->sock4 = NULL;
+#if IS_ENABLED(CONFIG_IPV6)
+ geneve->sock6 = NULL;
+ if (ipv6 || metadata)
+ ret = geneve_sock_add(geneve, true);
+#endif
+ if (!ret && (!ipv6 || metadata))
+ ret = geneve_sock_add(geneve, false);
+ if (ret < 0)
+ geneve_sock_release(geneve);
+
+ return ret;
+}
+
static int geneve_stop(struct net_device *dev)
{
struct geneve_dev *geneve = netdev_priv(dev);
- struct geneve_sock *gs = geneve->sock;
if (!hlist_unhashed(&geneve->hlist))
hlist_del_rcu(&geneve->hlist);
- geneve_sock_release(gs);
+ geneve_sock_release(geneve);
return 0;
}
+static void geneve_build_header(struct genevehdr *geneveh,
+ __be16 tun_flags, u8 vni[3],
+ u8 options_len, u8 *options)
+{
+ geneveh->ver = GENEVE_VER;
+ geneveh->opt_len = options_len / 4;
+ geneveh->oam = !!(tun_flags & TUNNEL_OAM);
+ geneveh->critical = !!(tun_flags & TUNNEL_CRIT_OPT);
+ geneveh->rsvd1 = 0;
+ memcpy(geneveh->vni, vni, 3);
+ geneveh->proto_type = htons(ETH_P_TEB);
+ geneveh->rsvd2 = 0;
+
+ memcpy(geneveh->options, options, options_len);
+}
+
static int geneve_build_skb(struct rtable *rt, struct sk_buff *skb,
__be16 tun_flags, u8 vni[3], u8 opt_len, u8 *opt,
- bool csum)
+ bool csum, bool xnet)
{
struct genevehdr *gnvh;
int min_headroom;
int err;
+ skb_scrub_packet(skb, xnet);
+
min_headroom = LL_RESERVED_SPACE(rt->dst.dev) + rt->dst.header_len
+ GENEVE_BASE_HLEN + opt_len + sizeof(struct iphdr);
err = skb_cow_head(skb, min_headroom);
@@ -544,15 +665,7 @@ static int geneve_build_skb(struct rtable *rt, struct sk_buff *skb,
}
gnvh = (struct genevehdr *)__skb_push(skb, sizeof(*gnvh) + opt_len);
- gnvh->ver = GENEVE_VER;
- gnvh->opt_len = opt_len / 4;
- gnvh->oam = !!(tun_flags & TUNNEL_OAM);
- gnvh->critical = !!(tun_flags & TUNNEL_CRIT_OPT);
- gnvh->rsvd1 = 0;
- memcpy(gnvh->vni, vni, 3);
- gnvh->proto_type = htons(ETH_P_TEB);
- gnvh->rsvd2 = 0;
- memcpy(gnvh->options, opt, opt_len);
+ geneve_build_header(gnvh, tun_flags, vni, opt_len, opt);
skb_set_inner_protocol(skb, htons(ETH_P_TEB));
return 0;
@@ -562,10 +675,47 @@ free_rt:
return err;
}
-static struct rtable *geneve_get_rt(struct sk_buff *skb,
- struct net_device *dev,
- struct flowi4 *fl4,
- struct ip_tunnel_info *info)
+#if IS_ENABLED(CONFIG_IPV6)
+static int geneve6_build_skb(struct dst_entry *dst, struct sk_buff *skb,
+ __be16 tun_flags, u8 vni[3], u8 opt_len, u8 *opt,
+ bool csum, bool xnet)
+{
+ struct genevehdr *gnvh;
+ int min_headroom;
+ int err;
+
+ skb_scrub_packet(skb, xnet);
+
+ min_headroom = LL_RESERVED_SPACE(dst->dev) + dst->header_len
+ + GENEVE_BASE_HLEN + opt_len + sizeof(struct ipv6hdr);
+ err = skb_cow_head(skb, min_headroom);
+ if (unlikely(err)) {
+ kfree_skb(skb);
+ goto free_dst;
+ }
+
+ skb = udp_tunnel_handle_offloads(skb, csum);
+ if (IS_ERR(skb)) {
+ err = PTR_ERR(skb);
+ goto free_dst;
+ }
+
+ gnvh = (struct genevehdr *)__skb_push(skb, sizeof(*gnvh) + opt_len);
+ geneve_build_header(gnvh, tun_flags, vni, opt_len, opt);
+
+ skb_set_inner_protocol(skb, htons(ETH_P_TEB));
+ return 0;
+
+free_dst:
+ dst_release(dst);
+ return err;
+}
+#endif
+
+static struct rtable *geneve_get_v4_rt(struct sk_buff *skb,
+ struct net_device *dev,
+ struct flowi4 *fl4,
+ struct ip_tunnel_info *info)
{
struct geneve_dev *geneve = netdev_priv(dev);
struct rtable *rt = NULL;
@@ -588,24 +738,42 @@ static struct rtable *geneve_get_rt(struct sk_buff *skb,
}
fl4->flowi4_tos = RT_TOS(tos);
- fl4->daddr = geneve->remote.sin_addr.s_addr;
+ fl4->daddr = geneve->remote.sin.sin_addr.s_addr;
}
rt = ip_route_output_key(geneve->net, fl4);
- if (IS_ERR(rt)) {
- netdev_dbg(dev, "no route to %pI4\n", &fl4->daddr);
- dev->stats.tx_carrier_errors++;
- return rt;
- }
- if (rt->dst.dev == dev) { /* is this necessary? */
- netdev_dbg(dev, "circular route to %pI4\n", &fl4->daddr);
- dev->stats.collisions++;
- ip_rt_put(rt);
- return ERR_PTR(-EINVAL);
- }
+
return rt;
}
+#if IS_ENABLED(CONFIG_IPV6)
+static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
+ struct net_device *dev,
+ struct flowi6 *fl6,
+ struct ip_tunnel_info *info)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ struct geneve_sock *gs6 = geneve->sock6;
+ struct dst_entry *dst = NULL;
+
+ memset(fl6, 0, sizeof(*fl6));
+ fl6->flowi6_mark = skb->mark;
+ fl6->flowi6_proto = IPPROTO_UDP;
+
+ if (info) {
+ fl6->daddr = info->key.u.ipv6.dst;
+ fl6->saddr = info->key.u.ipv6.src;
+ } else {
+ fl6->daddr = geneve->remote.sin6.sin6_addr;
+ }
+
+ if (ipv6_stub->ipv6_dst_lookup(geneve->net, gs6->sock->sk, &dst, fl6))
+ return ERR_PTR(-EHOSTUNREACH);
+
+ return dst;
+}
+#endif
+
/* Convert 64 bit tunnel ID to 24 bit VNI. */
static void tunnel_id_to_vni(__be64 tun_id, __u8 *vni)
{
@@ -620,11 +788,11 @@ static void tunnel_id_to_vni(__be64 tun_id, __u8 *vni)
#endif
}
-static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
+static netdev_tx_t geneve_xmit_skb(struct sk_buff *skb, struct net_device *dev,
+ struct ip_tunnel_info *info)
{
struct geneve_dev *geneve = netdev_priv(dev);
- struct geneve_sock *gs = geneve->sock;
- struct ip_tunnel_info *info = NULL;
+ struct geneve_sock *gs4 = geneve->sock4;
struct rtable *rt = NULL;
const struct iphdr *iip; /* interior IP header */
struct flowi4 fl4;
@@ -633,10 +801,10 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
bool udp_csum;
__be16 df;
int err;
+ bool xnet = !net_eq(geneve->net, dev_net(geneve->dev));
if (geneve->collect_md) {
- info = skb_tunnel_info(skb);
- if (unlikely(info && !(info->mode & IP_TUNNEL_INFO_TX))) {
+ if (unlikely(!info || !(info->mode & IP_TUNNEL_INFO_TX))) {
netdev_dbg(dev, "no tunnel metadata\n");
goto tx_error;
}
@@ -644,12 +812,18 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
goto tx_error;
}
- rt = geneve_get_rt(skb, dev, &fl4, info);
+ rt = geneve_get_v4_rt(skb, dev, &fl4, info);
if (IS_ERR(rt)) {
netdev_dbg(dev, "no route to %pI4\n", &fl4.daddr);
dev->stats.tx_carrier_errors++;
goto tx_error;
}
+ if (rt->dst.dev == dev) { /* is this necessary? */
+ netdev_dbg(dev, "circular route to %pI4\n", &fl4.daddr);
+ dev->stats.collisions++;
+ ip_rt_put(rt);
+ goto tx_error;
+ }
sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
skb_reset_mac_header(skb);
@@ -667,7 +841,7 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
udp_csum = !!(key->tun_flags & TUNNEL_CSUM);
err = geneve_build_skb(rt, skb, key->tun_flags, vni,
- info->options_len, opts, udp_csum);
+ info->options_len, opts, udp_csum, xnet);
if (unlikely(err))
goto err;
@@ -677,7 +851,7 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
} else {
udp_csum = false;
err = geneve_build_skb(rt, skb, 0, geneve->vni,
- 0, NULL, udp_csum);
+ 0, NULL, udp_csum, xnet);
if (unlikely(err))
goto err;
@@ -688,7 +862,7 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
ttl = ttl ? : ip4_dst_hoplimit(&rt->dst);
df = 0;
}
- err = udp_tunnel_xmit_skb(rt, gs->sock->sk, skb, fl4.saddr, fl4.daddr,
+ err = udp_tunnel_xmit_skb(rt, gs4->sock->sk, skb, fl4.saddr, fl4.daddr,
tos, ttl, df, sport, geneve->dst_port,
!net_eq(geneve->net, dev_net(geneve->dev)),
!udp_csum);
@@ -703,6 +877,103 @@ err:
return NETDEV_TX_OK;
}
+#if IS_ENABLED(CONFIG_IPV6)
+static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
+ struct ip_tunnel_info *info)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ struct geneve_sock *gs6 = geneve->sock6;
+ struct dst_entry *dst = NULL;
+ struct flowi6 fl6;
+ __u8 ttl;
+ __be16 sport;
+ bool udp_csum;
+ int err;
+ bool xnet = !net_eq(geneve->net, dev_net(geneve->dev));
+
+ if (geneve->collect_md) {
+ if (unlikely(!info || !(info->mode & IP_TUNNEL_INFO_TX))) {
+ netdev_dbg(dev, "no tunnel metadata\n");
+ goto tx_error;
+ }
+ }
+
+ dst = geneve_get_v6_dst(skb, dev, &fl6, info);
+ if (IS_ERR(dst)) {
+ netdev_dbg(dev, "no route to %pI6\n", &fl6.daddr);
+ dev->stats.tx_carrier_errors++;
+ goto tx_error;
+ }
+ if (dst->dev == dev) { /* is this necessary? */
+ netdev_dbg(dev, "circular route to %pI6\n", &fl6.daddr);
+ dev->stats.collisions++;
+ dst_release(dst);
+ goto tx_error;
+ }
+
+ sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
+ skb_reset_mac_header(skb);
+
+ if (info) {
+ const struct ip_tunnel_key *key = &info->key;
+ u8 *opts = NULL;
+ u8 vni[3];
+
+ tunnel_id_to_vni(key->tun_id, vni);
+ if (key->tun_flags & TUNNEL_GENEVE_OPT)
+ opts = ip_tunnel_info_opts(info);
+
+ udp_csum = !!(key->tun_flags & TUNNEL_CSUM);
+ err = geneve6_build_skb(dst, skb, key->tun_flags, vni,
+ info->options_len, opts,
+ udp_csum, xnet);
+ if (unlikely(err))
+ goto err;
+
+ ttl = key->ttl;
+ } else {
+ udp_csum = false;
+ err = geneve6_build_skb(dst, skb, 0, geneve->vni,
+ 0, NULL, udp_csum, xnet);
+ if (unlikely(err))
+ goto err;
+
+ ttl = geneve->ttl;
+ if (!ttl && ipv6_addr_is_multicast(&fl6.daddr))
+ ttl = 1;
+ ttl = ttl ? : ip6_dst_hoplimit(dst);
+ }
+ err = udp_tunnel6_xmit_skb(dst, gs6->sock->sk, skb, dev,
+ &fl6.saddr, &fl6.daddr, 0, ttl,
+ sport, geneve->dst_port, !udp_csum);
+
+ iptunnel_xmit_stats(err, &dev->stats, dev->tstats);
+ return NETDEV_TX_OK;
+
+tx_error:
+ dev_kfree_skb(skb);
+err:
+ dev->stats.tx_errors++;
+ return NETDEV_TX_OK;
+}
+#endif
+
+static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ struct ip_tunnel_info *info = NULL;
+
+ if (geneve->collect_md)
+ info = skb_tunnel_info(skb);
+
+#if IS_ENABLED(CONFIG_IPV6)
+ if ((info && ip_tunnel_info_af(info) == AF_INET6) ||
+ (!info && geneve->remote.sa.sa_family == AF_INET6))
+ return geneve6_xmit_skb(skb, dev, info);
+#endif
+ return geneve_xmit_skb(skb, dev, info);
+}
+
static const struct net_device_ops geneve_netdev_ops = {
.ndo_init = geneve_init,
.ndo_uninit = geneve_uninit,
@@ -759,6 +1030,7 @@ static void geneve_setup(struct net_device *dev)
static const struct nla_policy geneve_policy[IFLA_GENEVE_MAX + 1] = {
[IFLA_GENEVE_ID] = { .type = NLA_U32 },
[IFLA_GENEVE_REMOTE] = { .len = FIELD_SIZEOF(struct iphdr, daddr) },
+ [IFLA_GENEVE_REMOTE6] = { .len = sizeof(struct in6_addr) },
[IFLA_GENEVE_TTL] = { .type = NLA_U8 },
[IFLA_GENEVE_TOS] = { .type = NLA_U8 },
[IFLA_GENEVE_PORT] = { .type = NLA_U16 },
@@ -790,7 +1062,7 @@ static int geneve_validate(struct nlattr *tb[], struct nlattr *data[])
static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
__be16 dst_port,
- __be32 rem_addr,
+ union geneve_addr *remote,
u8 vni[],
bool *tun_on_same_port,
bool *tun_collect_md)
@@ -806,7 +1078,7 @@ static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
*tun_on_same_port = true;
}
if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
- rem_addr == geneve->remote.sin_addr.s_addr &&
+ !memcmp(remote, &geneve->remote, sizeof(geneve->remote)) &&
dst_port == geneve->dst_port)
t = geneve;
}
@@ -814,18 +1086,20 @@ static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
}
static int geneve_configure(struct net *net, struct net_device *dev,
- __be32 rem_addr, __u32 vni, __u8 ttl, __u8 tos,
- __be16 dst_port, bool metadata)
+ union geneve_addr *remote,
+ __u32 vni, __u8 ttl, __u8 tos, __be16 dst_port,
+ bool metadata)
{
struct geneve_net *gn = net_generic(net, geneve_net_id);
struct geneve_dev *t, *geneve = netdev_priv(dev);
bool tun_collect_md, tun_on_same_port;
int err;
- if (metadata) {
- if (rem_addr || vni || tos || ttl)
- return -EINVAL;
- }
+ if (!remote)
+ return -EINVAL;
+ if (metadata &&
+ (remote->sa.sa_family != AF_UNSPEC || vni || tos || ttl))
+ return -EINVAL;
geneve->net = net;
geneve->dev = dev;
@@ -834,16 +1108,19 @@ static int geneve_configure(struct net *net, struct net_device *dev,
geneve->vni[1] = (vni & 0x0000ff00) >> 8;
geneve->vni[2] = vni & 0x000000ff;
- geneve->remote.sin_addr.s_addr = rem_addr;
- if (IN_MULTICAST(ntohl(geneve->remote.sin_addr.s_addr)))
+ if ((remote->sa.sa_family == AF_INET &&
+ IN_MULTICAST(ntohl(remote->sin.sin_addr.s_addr))) ||
+ (remote->sa.sa_family == AF_INET6 &&
+ ipv6_addr_is_multicast(&remote->sin6.sin6_addr)))
return -EINVAL;
+ geneve->remote = *remote;
geneve->ttl = ttl;
geneve->tos = tos;
geneve->dst_port = dst_port;
geneve->collect_md = metadata;
- t = geneve_find_dev(gn, dst_port, rem_addr, geneve->vni,
+ t = geneve_find_dev(gn, dst_port, remote, geneve->vni,
&tun_on_same_port, &tun_collect_md);
if (t)
return -EBUSY;
@@ -870,15 +1147,35 @@ static int geneve_newlink(struct net *net, struct net_device *dev,
__be16 dst_port = htons(GENEVE_UDP_PORT);
__u8 ttl = 0, tos = 0;
bool metadata = false;
- __be32 rem_addr = 0;
+ union geneve_addr remote = geneve_remote_unspec;
__u32 vni = 0;
+ if (data[IFLA_GENEVE_REMOTE] && data[IFLA_GENEVE_REMOTE6])
+ return -EINVAL;
+
+ if (data[IFLA_GENEVE_REMOTE]) {
+ remote.sa.sa_family = AF_INET;
+ remote.sin.sin_addr.s_addr =
+ nla_get_in_addr(data[IFLA_GENEVE_REMOTE]);
+ }
+
+ if (data[IFLA_GENEVE_REMOTE6]) {
+ if (!IS_ENABLED(CONFIG_IPV6))
+ return -EPFNOSUPPORT;
+
+ remote.sa.sa_family = AF_INET6;
+ remote.sin6.sin6_addr =
+ nla_get_in6_addr(data[IFLA_GENEVE_REMOTE6]);
+
+ if (ipv6_addr_type(&remote.sin6.sin6_addr) &
+ IPV6_ADDR_LINKLOCAL)
+ netdev_dbg(dev, "link-local remote is unsupported\n");
+ return -EINVAL;
+ }
+
if (data[IFLA_GENEVE_ID])
vni = nla_get_u32(data[IFLA_GENEVE_ID]);
- if (data[IFLA_GENEVE_REMOTE])
- rem_addr = nla_get_in_addr(data[IFLA_GENEVE_REMOTE]);
-
if (data[IFLA_GENEVE_TTL])
ttl = nla_get_u8(data[IFLA_GENEVE_TTL]);
@@ -891,8 +1188,8 @@ static int geneve_newlink(struct net *net, struct net_device *dev,
if (data[IFLA_GENEVE_COLLECT_METADATA])
metadata = true;
- return geneve_configure(net, dev, rem_addr, vni,
- ttl, tos, dst_port, metadata);
+ return geneve_configure(net, dev, &remote, vni, ttl, tos, dst_port,
+ metadata);
}
static void geneve_dellink(struct net_device *dev, struct list_head *head)
@@ -906,7 +1203,7 @@ static void geneve_dellink(struct net_device *dev, struct list_head *head)
static size_t geneve_get_size(const struct net_device *dev)
{
return nla_total_size(sizeof(__u32)) + /* IFLA_GENEVE_ID */
- nla_total_size(sizeof(struct in_addr)) + /* IFLA_GENEVE_REMOTE */
+ nla_total_size(sizeof(struct in6_addr)) + /* IFLA_GENEVE_REMOTE{6} */
nla_total_size(sizeof(__u8)) + /* IFLA_GENEVE_TTL */
nla_total_size(sizeof(__u8)) + /* IFLA_GENEVE_TOS */
nla_total_size(sizeof(__be16)) + /* IFLA_GENEVE_PORT */
@@ -923,9 +1220,17 @@ static int geneve_fill_info(struct sk_buff *skb, const struct net_device *dev)
if (nla_put_u32(skb, IFLA_GENEVE_ID, vni))
goto nla_put_failure;
- if (nla_put_in_addr(skb, IFLA_GENEVE_REMOTE,
- geneve->remote.sin_addr.s_addr))
- goto nla_put_failure;
+ if (geneve->remote.sa.sa_family == AF_INET) {
+ if (nla_put_in_addr(skb, IFLA_GENEVE_REMOTE,
+ geneve->remote.sin.sin_addr.s_addr))
+ goto nla_put_failure;
+#if IS_ENABLED(CONFIG_IPV6)
+ } else {
+ if (nla_put_in6_addr(skb, IFLA_GENEVE_REMOTE6,
+ &geneve->remote.sin6.sin6_addr))
+ goto nla_put_failure;
+#endif
+ }
if (nla_put_u8(skb, IFLA_GENEVE_TTL, geneve->ttl) ||
nla_put_u8(skb, IFLA_GENEVE_TOS, geneve->tos))
@@ -971,7 +1276,8 @@ struct net_device *geneve_dev_create_fb(struct net *net, const char *name,
if (IS_ERR(dev))
return dev;
- err = geneve_configure(net, dev, 0, 0, 0, 0, htons(dst_port), true);
+ err = geneve_configure(net, dev, &geneve_remote_unspec,
+ 0, 0, 0, htons(dst_port), true);
if (err) {
free_netdev(dev);
return ERR_PTR(err);
diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
index e3b6217f34f1..45e3a48550f9 100644
--- a/include/uapi/linux/if_link.h
+++ b/include/uapi/linux/if_link.h
@@ -461,6 +461,7 @@ enum {
IFLA_GENEVE_TOS,
IFLA_GENEVE_PORT, /* destination port */
IFLA_GENEVE_COLLECT_METADATA,
+ IFLA_GENEVE_REMOTE6,
__IFLA_GENEVE_MAX
};
#define IFLA_GENEVE_MAX (__IFLA_GENEVE_MAX - 1)
--
2.4.3
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH v5 2/2] geneve: handle ipv6 priority like ipv4 tos
2015-10-22 19:45 ` [PATCH v5 " John W. Linville
@ 2015-10-22 19:45 ` John W. Linville
2015-10-23 4:48 ` [PATCH v5 1/2] geneve: implement support for IPv6-based tunnels YOSHIFUJI Hideaki
2015-10-23 14:40 ` [PATCH v6 " John W. Linville
2 siblings, 0 replies; 37+ messages in thread
From: John W. Linville @ 2015-10-22 19:45 UTC (permalink / raw)
To: netdev
Cc: Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc,
YOSHIFUJI Hideaki/吉藤英明,
John W. Linville
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Reported-by: Jesse Gross <jesse@nicira.com>
Reviewed-by: Jesse Gross <jesse@nicira.com>
---
v5 -- same as previous revision
drivers/net/geneve.c | 19 +++++++++++++++++--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index 47f7512f02cf..4d4d8ca9eb7a 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -755,6 +755,7 @@ static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
struct geneve_dev *geneve = netdev_priv(dev);
struct geneve_sock *gs6 = geneve->sock6;
struct dst_entry *dst = NULL;
+ __u8 prio;
memset(fl6, 0, sizeof(*fl6));
fl6->flowi6_mark = skb->mark;
@@ -763,7 +764,16 @@ static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
if (info) {
fl6->daddr = info->key.u.ipv6.dst;
fl6->saddr = info->key.u.ipv6.src;
+ fl6->flowi6_tos = RT_TOS(info->key.tos);
} else {
+ prio = geneve->tos;
+ if (prio == 1) {
+ const struct iphdr *iip = ip_hdr(skb);
+
+ prio = ip_tunnel_get_dsfield(iip, skb);
+ }
+
+ fl6->flowi6_tos = RT_TOS(prio);
fl6->daddr = geneve->remote.sin6.sin6_addr;
}
@@ -884,8 +894,9 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
struct geneve_dev *geneve = netdev_priv(dev);
struct geneve_sock *gs6 = geneve->sock6;
struct dst_entry *dst = NULL;
+ const struct iphdr *iip; /* interior IP header */
struct flowi6 fl6;
- __u8 ttl;
+ __u8 prio, ttl;
__be16 sport;
bool udp_csum;
int err;
@@ -914,6 +925,8 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
skb_reset_mac_header(skb);
+ iip = ip_hdr(skb);
+
if (info) {
const struct ip_tunnel_key *key = &info->key;
u8 *opts = NULL;
@@ -930,6 +943,7 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
if (unlikely(err))
goto err;
+ prio = ip_tunnel_ecn_encap(key->tos, iip, skb);
ttl = key->ttl;
} else {
udp_csum = false;
@@ -938,13 +952,14 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
if (unlikely(err))
goto err;
+ prio = ip_tunnel_ecn_encap(fl6.flowi6_tos, iip, skb);
ttl = geneve->ttl;
if (!ttl && ipv6_addr_is_multicast(&fl6.daddr))
ttl = 1;
ttl = ttl ? : ip6_dst_hoplimit(dst);
}
err = udp_tunnel6_xmit_skb(dst, gs6->sock->sk, skb, dev,
- &fl6.saddr, &fl6.daddr, 0, ttl,
+ &fl6.saddr, &fl6.daddr, prio, ttl,
sport, geneve->dst_port, !udp_csum);
iptunnel_xmit_stats(err, &dev->stats, dev->tstats);
--
2.4.3
^ permalink raw reply related [flat|nested] 37+ messages in thread
* Re: [PATCH v5 1/2] geneve: implement support for IPv6-based tunnels
2015-10-22 19:45 ` [PATCH v5 " John W. Linville
2015-10-22 19:45 ` [PATCH v5 2/2] geneve: handle ipv6 priority like ipv4 tos John W. Linville
@ 2015-10-23 4:48 ` YOSHIFUJI Hideaki
2015-10-23 13:38 ` John W. Linville
2015-10-23 14:40 ` [PATCH v6 " John W. Linville
2 siblings, 1 reply; 37+ messages in thread
From: YOSHIFUJI Hideaki @ 2015-10-23 4:48 UTC (permalink / raw)
To: John W. Linville, netdev
Cc: hideaki.yoshifuji, Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc
Hi,
John W. Linville wrote:
> NOTE: Link-local IPv6 addresses for remote endpoints are not supported,
> since the driver currently has no capacity for binding a geneve
> interface to a specific link.
>
> Signed-off-by: John W. Linville <linville@tuxdriver.com>
> ---
> v5:
> - wrap declaration of sock6 in geneve_dev with IS_ENABLED(CONFIG_IPV6)
> - remove superfluous '!!' when assigning geneve->collect_md to bool
> - use skb_scrub_packet in IPv4 tx path as well
> - check for NULL ip_tunnel_info pointer in geneve[6]_xmit_skb
> - use ipv6_addr_equal for comparing IPv6 addresses
> - more use of IS_ENABLED(CONFIG_IPV6) for preserving build integrity
> - reject link-local ipv6 address for remote tunnel endpoint
:
> @@ -870,15 +1147,35 @@ static int geneve_newlink(struct net *net, struct net_device *dev,
:
> +
> + if (ipv6_addr_type(&remote.sin6.sin6_addr) &
> + IPV6_ADDR_LINKLOCAL)
> + netdev_dbg(dev, "link-local remote is unsupported\n");
> + return -EINVAL;
> + }
> +
This always returns -EINVAL; {} is missing.
--
Hideaki Yoshifuji <hideaki.yoshifuji@miraclelinux.com>
Technical Division, MIRACLE LINUX CORPORATION
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH v5 1/2] geneve: implement support for IPv6-based tunnels
2015-10-23 4:48 ` [PATCH v5 1/2] geneve: implement support for IPv6-based tunnels YOSHIFUJI Hideaki
@ 2015-10-23 13:38 ` John W. Linville
0 siblings, 0 replies; 37+ messages in thread
From: John W. Linville @ 2015-10-23 13:38 UTC (permalink / raw)
To: YOSHIFUJI Hideaki
Cc: netdev, Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc
On Fri, Oct 23, 2015 at 01:48:49PM +0900, YOSHIFUJI Hideaki wrote:
> Hi,
>
> John W. Linville wrote:
> > NOTE: Link-local IPv6 addresses for remote endpoints are not supported,
> > since the driver currently has no capacity for binding a geneve
> > interface to a specific link.
> >
> > Signed-off-by: John W. Linville <linville@tuxdriver.com>
> > ---
> > v5:
> > - wrap declaration of sock6 in geneve_dev with IS_ENABLED(CONFIG_IPV6)
> > - remove superfluous '!!' when assigning geneve->collect_md to bool
> > - use skb_scrub_packet in IPv4 tx path as well
> > - check for NULL ip_tunnel_info pointer in geneve[6]_xmit_skb
> > - use ipv6_addr_equal for comparing IPv6 addresses
> > - more use of IS_ENABLED(CONFIG_IPV6) for preserving build integrity
> > - reject link-local ipv6 address for remote tunnel endpoint
> :
>
> > @@ -870,15 +1147,35 @@ static int geneve_newlink(struct net *net, struct net_device *dev,
> :
> > +
> > + if (ipv6_addr_type(&remote.sin6.sin6_addr) &
> > + IPV6_ADDR_LINKLOCAL)
> > + netdev_dbg(dev, "link-local remote is unsupported\n");
> > + return -EINVAL;
> > + }
> > +
>
> This always returns -EINVAL; {} is missing.
Yikes! I posted the wrong patch...good eyes!
I will repost...
John
--
John W. Linville Someday the world will need a hero, and you
linville@tuxdriver.com might be all we have. Be ready.
^ permalink raw reply [flat|nested] 37+ messages in thread
* [PATCH v6 1/2] geneve: implement support for IPv6-based tunnels
2015-10-22 19:45 ` [PATCH v5 " John W. Linville
2015-10-22 19:45 ` [PATCH v5 2/2] geneve: handle ipv6 priority like ipv4 tos John W. Linville
2015-10-23 4:48 ` [PATCH v5 1/2] geneve: implement support for IPv6-based tunnels YOSHIFUJI Hideaki
@ 2015-10-23 14:40 ` John W. Linville
2015-10-23 14:40 ` [PATCH v6 2/2] geneve: handle ipv6 priority like ipv4 tos John W. Linville
` (2 more replies)
2 siblings, 3 replies; 37+ messages in thread
From: John W. Linville @ 2015-10-23 14:40 UTC (permalink / raw)
To: netdev
Cc: Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc,
YOSHIFUJI Hideaki/吉藤英明,
John W. Linville
NOTE: Link-local IPv6 addresses for remote endpoints are not supported,
since the driver currently has no capacity for binding a geneve
interface to a specific link.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
---
v6:
- fix a typo (missing {}'s)
v5:
- wrap declaration of sock6 in geneve_dev with IS_ENABLED(CONFIG_IPV6)
- remove superfluous '!!' when assigning geneve->collect_md to bool
- use skb_scrub_packet in IPv4 tx path as well
- check for NULL ip_tunnel_info pointer in geneve[6]_xmit_skb
- use ipv6_addr_equal for comparing IPv6 addresses
- more use of IS_ENABLED(CONFIG_IPV6) for preserving build integrity
- reject link-local ipv6 address for remote tunnel endpoint
v4:
- treat mode field of ip_tunnel_info as flags
- add a missing IS_ENABLED(CONFIG_IPV6) to geneve_rx
- remove unneeded flags field in geneve_dev
- NULL-check parameter for __geneve_sock_release
- check remote socket family for AF_UNSPEC in geneve_configure
- rename geneve_get_{rt,dst} as geneve_get_{v4_rt,v6_dst}
- refactor some error handling in the xmit paths
v3:
- declare geneve_remote_unspec as static
v2:
- do not require remote address for tx on metadata tunnels
- pass correct sockaddr family to udp_tun_rx_dst in geneve_rx
- accommodate both ipv4 and ipv6 sockets open on same tunnel
- move declaration of geneve_get_dst for aesthetic purposes
drivers/net/geneve.c | 483 +++++++++++++++++++++++++++++++++++--------
include/uapi/linux/if_link.h | 1 +
2 files changed, 396 insertions(+), 88 deletions(-)
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index cde29f8a37bf..46d4aa714893 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -46,16 +46,27 @@ struct geneve_net {
static int geneve_net_id;
+union geneve_addr {
+ struct sockaddr_in sin;
+ struct sockaddr_in6 sin6;
+ struct sockaddr sa;
+};
+
+static union geneve_addr geneve_remote_unspec = { .sa.sa_family = AF_UNSPEC, };
+
/* Pseudo network device */
struct geneve_dev {
struct hlist_node hlist; /* vni hash table */
struct net *net; /* netns for packet i/o */
struct net_device *dev; /* netdev for geneve tunnel */
- struct geneve_sock *sock; /* socket used for geneve tunnel */
+ struct geneve_sock *sock4; /* IPv4 socket used for geneve tunnel */
+#if IS_ENABLED(CONFIG_IPV6)
+ struct geneve_sock *sock6; /* IPv6 socket used for geneve tunnel */
+#endif
u8 vni[3]; /* virtual network ID for tunnel */
u8 ttl; /* TTL override */
u8 tos; /* TOS override */
- struct sockaddr_in remote; /* IPv4 address for link partner */
+ union geneve_addr remote; /* IP address for link partner */
struct list_head next; /* geneve's per namespace list */
__be16 dst_port;
bool collect_md;
@@ -103,12 +114,32 @@ static struct geneve_dev *geneve_lookup(struct geneve_sock *gs,
vni_list_head = &gs->vni_list[hash];
hlist_for_each_entry_rcu(geneve, vni_list_head, hlist) {
if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
- addr == geneve->remote.sin_addr.s_addr)
+ addr == geneve->remote.sin.sin_addr.s_addr)
return geneve;
}
return NULL;
}
+#if IS_ENABLED(CONFIG_IPV6)
+static struct geneve_dev *geneve6_lookup(struct geneve_sock *gs,
+ struct in6_addr addr6, u8 vni[])
+{
+ struct hlist_head *vni_list_head;
+ struct geneve_dev *geneve;
+ __u32 hash;
+
+ /* Find the device for this VNI */
+ hash = geneve_net_vni_hash(vni);
+ vni_list_head = &gs->vni_list[hash];
+ hlist_for_each_entry_rcu(geneve, vni_list_head, hlist) {
+ if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
+ ipv6_addr_equal(&addr6, &geneve->remote.sin6.sin6_addr))
+ return geneve;
+ }
+ return NULL;
+}
+#endif
+
static inline struct genevehdr *geneve_hdr(const struct sk_buff *skb)
{
return (struct genevehdr *)(udp_hdr(skb) + 1);
@@ -121,24 +152,49 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
struct metadata_dst *tun_dst = NULL;
struct geneve_dev *geneve = NULL;
struct pcpu_sw_netstats *stats;
- struct iphdr *iph;
- u8 *vni;
+ struct iphdr *iph = NULL;
__be32 addr;
- int err;
+ static u8 zero_vni[3];
+ u8 *vni;
+ int err = 0;
+ sa_family_t sa_family;
+#if IS_ENABLED(CONFIG_IPV6)
+ struct ipv6hdr *ip6h = NULL;
+ struct in6_addr addr6;
+ static struct in6_addr zero_addr6;
+#endif
+
+ sa_family = gs->sock->sk->sk_family;
- iph = ip_hdr(skb); /* outer IP header... */
+ if (sa_family == AF_INET) {
+ iph = ip_hdr(skb); /* outer IP header... */
- if (gs->collect_md) {
- static u8 zero_vni[3];
+ if (gs->collect_md) {
+ vni = zero_vni;
+ addr = 0;
+ } else {
+ vni = gnvh->vni;
- vni = zero_vni;
- addr = 0;
- } else {
- vni = gnvh->vni;
- addr = iph->saddr;
- }
+ addr = iph->saddr;
+ }
- geneve = geneve_lookup(gs, addr, vni);
+ geneve = geneve_lookup(gs, addr, vni);
+#if IS_ENABLED(CONFIG_IPV6)
+ } else if (sa_family == AF_INET6) {
+ ip6h = ipv6_hdr(skb); /* outer IPv6 header... */
+
+ if (gs->collect_md) {
+ vni = zero_vni;
+ addr6 = zero_addr6;
+ } else {
+ vni = gnvh->vni;
+
+ addr6 = ip6h->saddr;
+ }
+
+ geneve = geneve6_lookup(gs, addr6, vni);
+#endif
+ }
if (!geneve)
goto drop;
@@ -149,7 +205,7 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
(gnvh->oam ? TUNNEL_OAM : 0) |
(gnvh->critical ? TUNNEL_CRIT_OPT : 0);
- tun_dst = udp_tun_rx_dst(skb, AF_INET, flags,
+ tun_dst = udp_tun_rx_dst(skb, sa_family, flags,
vni_to_tunnel_id(gnvh->vni),
gnvh->opt_len * 4);
if (!tun_dst)
@@ -179,12 +235,25 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
skb_reset_network_header(skb);
- err = IP_ECN_decapsulate(iph, skb);
+ if (iph)
+ err = IP_ECN_decapsulate(iph, skb);
+#if IS_ENABLED(CONFIG_IPV6)
+ if (ip6h)
+ err = IP6_ECN_decapsulate(ip6h, skb);
+#endif
if (unlikely(err)) {
- if (log_ecn_error)
- net_info_ratelimited("non-ECT from %pI4 with TOS=%#x\n",
- &iph->saddr, iph->tos);
+ if (log_ecn_error) {
+ if (iph)
+ net_info_ratelimited("non-ECT from %pI4 "
+ "with TOS=%#x\n",
+ &iph->saddr, iph->tos);
+#if IS_ENABLED(CONFIG_IPV6)
+ if (ip6h)
+ net_info_ratelimited("non-ECT from %pI6\n",
+ &ip6h->saddr);
+#endif
+ }
if (err > 1) {
++geneve->dev->stats.rx_frame_errors;
++geneve->dev->stats.rx_errors;
@@ -284,6 +353,7 @@ static struct socket *geneve_create_sock(struct net *net, bool ipv6,
if (ipv6) {
udp_conf.family = AF_INET6;
+ udp_conf.ipv6_v6only = 1;
} else {
udp_conf.family = AF_INET;
udp_conf.local_ip.s_addr = htonl(INADDR_ANY);
@@ -458,9 +528,9 @@ static void geneve_notify_del_rx_port(struct geneve_sock *gs)
udp_del_offload(&gs->udp_offloads);
}
-static void geneve_sock_release(struct geneve_sock *gs)
+static void __geneve_sock_release(struct geneve_sock *gs)
{
- if (--gs->refcnt)
+ if (!gs || --gs->refcnt)
return;
list_del(&gs->list);
@@ -469,66 +539,117 @@ static void geneve_sock_release(struct geneve_sock *gs)
kfree_rcu(gs, rcu);
}
+static void geneve_sock_release(struct geneve_dev *geneve)
+{
+ __geneve_sock_release(geneve->sock4);
+#if IS_ENABLED(CONFIG_IPV6)
+ __geneve_sock_release(geneve->sock6);
+#endif
+}
+
static struct geneve_sock *geneve_find_sock(struct geneve_net *gn,
+ sa_family_t family,
__be16 dst_port)
{
struct geneve_sock *gs;
list_for_each_entry(gs, &gn->sock_list, list) {
if (inet_sk(gs->sock->sk)->inet_sport == dst_port &&
- inet_sk(gs->sock->sk)->sk.sk_family == AF_INET) {
+ inet_sk(gs->sock->sk)->sk.sk_family == family) {
return gs;
}
}
return NULL;
}
-static int geneve_open(struct net_device *dev)
+static int geneve_sock_add(struct geneve_dev *geneve, bool ipv6)
{
- struct geneve_dev *geneve = netdev_priv(dev);
struct net *net = geneve->net;
struct geneve_net *gn = net_generic(net, geneve_net_id);
struct geneve_sock *gs;
__u32 hash;
- gs = geneve_find_sock(gn, geneve->dst_port);
+ gs = geneve_find_sock(gn, ipv6 ? AF_INET6 : AF_INET, geneve->dst_port);
if (gs) {
gs->refcnt++;
goto out;
}
- gs = geneve_socket_create(net, geneve->dst_port, false);
+ gs = geneve_socket_create(net, geneve->dst_port, ipv6);
if (IS_ERR(gs))
return PTR_ERR(gs);
out:
gs->collect_md = geneve->collect_md;
- geneve->sock = gs;
+#if IS_ENABLED(CONFIG_IPV6)
+ if (ipv6)
+ geneve->sock6 = gs;
+ else
+#endif
+ geneve->sock4 = gs;
hash = geneve_net_vni_hash(geneve->vni);
hlist_add_head_rcu(&geneve->hlist, &gs->vni_list[hash]);
return 0;
}
+static int geneve_open(struct net_device *dev)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ bool ipv6 = geneve->remote.sa.sa_family == AF_INET6;
+ bool metadata = geneve->collect_md;
+ int ret = 0;
+
+ geneve->sock4 = NULL;
+#if IS_ENABLED(CONFIG_IPV6)
+ geneve->sock6 = NULL;
+ if (ipv6 || metadata)
+ ret = geneve_sock_add(geneve, true);
+#endif
+ if (!ret && (!ipv6 || metadata))
+ ret = geneve_sock_add(geneve, false);
+ if (ret < 0)
+ geneve_sock_release(geneve);
+
+ return ret;
+}
+
static int geneve_stop(struct net_device *dev)
{
struct geneve_dev *geneve = netdev_priv(dev);
- struct geneve_sock *gs = geneve->sock;
if (!hlist_unhashed(&geneve->hlist))
hlist_del_rcu(&geneve->hlist);
- geneve_sock_release(gs);
+ geneve_sock_release(geneve);
return 0;
}
+static void geneve_build_header(struct genevehdr *geneveh,
+ __be16 tun_flags, u8 vni[3],
+ u8 options_len, u8 *options)
+{
+ geneveh->ver = GENEVE_VER;
+ geneveh->opt_len = options_len / 4;
+ geneveh->oam = !!(tun_flags & TUNNEL_OAM);
+ geneveh->critical = !!(tun_flags & TUNNEL_CRIT_OPT);
+ geneveh->rsvd1 = 0;
+ memcpy(geneveh->vni, vni, 3);
+ geneveh->proto_type = htons(ETH_P_TEB);
+ geneveh->rsvd2 = 0;
+
+ memcpy(geneveh->options, options, options_len);
+}
+
static int geneve_build_skb(struct rtable *rt, struct sk_buff *skb,
__be16 tun_flags, u8 vni[3], u8 opt_len, u8 *opt,
- bool csum)
+ bool csum, bool xnet)
{
struct genevehdr *gnvh;
int min_headroom;
int err;
+ skb_scrub_packet(skb, xnet);
+
min_headroom = LL_RESERVED_SPACE(rt->dst.dev) + rt->dst.header_len
+ GENEVE_BASE_HLEN + opt_len + sizeof(struct iphdr);
err = skb_cow_head(skb, min_headroom);
@@ -544,15 +665,7 @@ static int geneve_build_skb(struct rtable *rt, struct sk_buff *skb,
}
gnvh = (struct genevehdr *)__skb_push(skb, sizeof(*gnvh) + opt_len);
- gnvh->ver = GENEVE_VER;
- gnvh->opt_len = opt_len / 4;
- gnvh->oam = !!(tun_flags & TUNNEL_OAM);
- gnvh->critical = !!(tun_flags & TUNNEL_CRIT_OPT);
- gnvh->rsvd1 = 0;
- memcpy(gnvh->vni, vni, 3);
- gnvh->proto_type = htons(ETH_P_TEB);
- gnvh->rsvd2 = 0;
- memcpy(gnvh->options, opt, opt_len);
+ geneve_build_header(gnvh, tun_flags, vni, opt_len, opt);
skb_set_inner_protocol(skb, htons(ETH_P_TEB));
return 0;
@@ -562,10 +675,47 @@ free_rt:
return err;
}
-static struct rtable *geneve_get_rt(struct sk_buff *skb,
- struct net_device *dev,
- struct flowi4 *fl4,
- struct ip_tunnel_info *info)
+#if IS_ENABLED(CONFIG_IPV6)
+static int geneve6_build_skb(struct dst_entry *dst, struct sk_buff *skb,
+ __be16 tun_flags, u8 vni[3], u8 opt_len, u8 *opt,
+ bool csum, bool xnet)
+{
+ struct genevehdr *gnvh;
+ int min_headroom;
+ int err;
+
+ skb_scrub_packet(skb, xnet);
+
+ min_headroom = LL_RESERVED_SPACE(dst->dev) + dst->header_len
+ + GENEVE_BASE_HLEN + opt_len + sizeof(struct ipv6hdr);
+ err = skb_cow_head(skb, min_headroom);
+ if (unlikely(err)) {
+ kfree_skb(skb);
+ goto free_dst;
+ }
+
+ skb = udp_tunnel_handle_offloads(skb, csum);
+ if (IS_ERR(skb)) {
+ err = PTR_ERR(skb);
+ goto free_dst;
+ }
+
+ gnvh = (struct genevehdr *)__skb_push(skb, sizeof(*gnvh) + opt_len);
+ geneve_build_header(gnvh, tun_flags, vni, opt_len, opt);
+
+ skb_set_inner_protocol(skb, htons(ETH_P_TEB));
+ return 0;
+
+free_dst:
+ dst_release(dst);
+ return err;
+}
+#endif
+
+static struct rtable *geneve_get_v4_rt(struct sk_buff *skb,
+ struct net_device *dev,
+ struct flowi4 *fl4,
+ struct ip_tunnel_info *info)
{
struct geneve_dev *geneve = netdev_priv(dev);
struct rtable *rt = NULL;
@@ -588,24 +738,42 @@ static struct rtable *geneve_get_rt(struct sk_buff *skb,
}
fl4->flowi4_tos = RT_TOS(tos);
- fl4->daddr = geneve->remote.sin_addr.s_addr;
+ fl4->daddr = geneve->remote.sin.sin_addr.s_addr;
}
rt = ip_route_output_key(geneve->net, fl4);
- if (IS_ERR(rt)) {
- netdev_dbg(dev, "no route to %pI4\n", &fl4->daddr);
- dev->stats.tx_carrier_errors++;
- return rt;
- }
- if (rt->dst.dev == dev) { /* is this necessary? */
- netdev_dbg(dev, "circular route to %pI4\n", &fl4->daddr);
- dev->stats.collisions++;
- ip_rt_put(rt);
- return ERR_PTR(-EINVAL);
- }
+
return rt;
}
+#if IS_ENABLED(CONFIG_IPV6)
+static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
+ struct net_device *dev,
+ struct flowi6 *fl6,
+ struct ip_tunnel_info *info)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ struct geneve_sock *gs6 = geneve->sock6;
+ struct dst_entry *dst = NULL;
+
+ memset(fl6, 0, sizeof(*fl6));
+ fl6->flowi6_mark = skb->mark;
+ fl6->flowi6_proto = IPPROTO_UDP;
+
+ if (info) {
+ fl6->daddr = info->key.u.ipv6.dst;
+ fl6->saddr = info->key.u.ipv6.src;
+ } else {
+ fl6->daddr = geneve->remote.sin6.sin6_addr;
+ }
+
+ if (ipv6_stub->ipv6_dst_lookup(geneve->net, gs6->sock->sk, &dst, fl6))
+ return ERR_PTR(-EHOSTUNREACH);
+
+ return dst;
+}
+#endif
+
/* Convert 64 bit tunnel ID to 24 bit VNI. */
static void tunnel_id_to_vni(__be64 tun_id, __u8 *vni)
{
@@ -620,11 +788,11 @@ static void tunnel_id_to_vni(__be64 tun_id, __u8 *vni)
#endif
}
-static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
+static netdev_tx_t geneve_xmit_skb(struct sk_buff *skb, struct net_device *dev,
+ struct ip_tunnel_info *info)
{
struct geneve_dev *geneve = netdev_priv(dev);
- struct geneve_sock *gs = geneve->sock;
- struct ip_tunnel_info *info = NULL;
+ struct geneve_sock *gs4 = geneve->sock4;
struct rtable *rt = NULL;
const struct iphdr *iip; /* interior IP header */
struct flowi4 fl4;
@@ -633,10 +801,10 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
bool udp_csum;
__be16 df;
int err;
+ bool xnet = !net_eq(geneve->net, dev_net(geneve->dev));
if (geneve->collect_md) {
- info = skb_tunnel_info(skb);
- if (unlikely(info && !(info->mode & IP_TUNNEL_INFO_TX))) {
+ if (unlikely(!info || !(info->mode & IP_TUNNEL_INFO_TX))) {
netdev_dbg(dev, "no tunnel metadata\n");
goto tx_error;
}
@@ -644,12 +812,18 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
goto tx_error;
}
- rt = geneve_get_rt(skb, dev, &fl4, info);
+ rt = geneve_get_v4_rt(skb, dev, &fl4, info);
if (IS_ERR(rt)) {
netdev_dbg(dev, "no route to %pI4\n", &fl4.daddr);
dev->stats.tx_carrier_errors++;
goto tx_error;
}
+ if (rt->dst.dev == dev) { /* is this necessary? */
+ netdev_dbg(dev, "circular route to %pI4\n", &fl4.daddr);
+ dev->stats.collisions++;
+ ip_rt_put(rt);
+ goto tx_error;
+ }
sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
skb_reset_mac_header(skb);
@@ -667,7 +841,7 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
udp_csum = !!(key->tun_flags & TUNNEL_CSUM);
err = geneve_build_skb(rt, skb, key->tun_flags, vni,
- info->options_len, opts, udp_csum);
+ info->options_len, opts, udp_csum, xnet);
if (unlikely(err))
goto err;
@@ -677,7 +851,7 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
} else {
udp_csum = false;
err = geneve_build_skb(rt, skb, 0, geneve->vni,
- 0, NULL, udp_csum);
+ 0, NULL, udp_csum, xnet);
if (unlikely(err))
goto err;
@@ -688,7 +862,7 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
ttl = ttl ? : ip4_dst_hoplimit(&rt->dst);
df = 0;
}
- err = udp_tunnel_xmit_skb(rt, gs->sock->sk, skb, fl4.saddr, fl4.daddr,
+ err = udp_tunnel_xmit_skb(rt, gs4->sock->sk, skb, fl4.saddr, fl4.daddr,
tos, ttl, df, sport, geneve->dst_port,
!net_eq(geneve->net, dev_net(geneve->dev)),
!udp_csum);
@@ -703,6 +877,103 @@ err:
return NETDEV_TX_OK;
}
+#if IS_ENABLED(CONFIG_IPV6)
+static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
+ struct ip_tunnel_info *info)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ struct geneve_sock *gs6 = geneve->sock6;
+ struct dst_entry *dst = NULL;
+ struct flowi6 fl6;
+ __u8 ttl;
+ __be16 sport;
+ bool udp_csum;
+ int err;
+ bool xnet = !net_eq(geneve->net, dev_net(geneve->dev));
+
+ if (geneve->collect_md) {
+ if (unlikely(!info || !(info->mode & IP_TUNNEL_INFO_TX))) {
+ netdev_dbg(dev, "no tunnel metadata\n");
+ goto tx_error;
+ }
+ }
+
+ dst = geneve_get_v6_dst(skb, dev, &fl6, info);
+ if (IS_ERR(dst)) {
+ netdev_dbg(dev, "no route to %pI6\n", &fl6.daddr);
+ dev->stats.tx_carrier_errors++;
+ goto tx_error;
+ }
+ if (dst->dev == dev) { /* is this necessary? */
+ netdev_dbg(dev, "circular route to %pI6\n", &fl6.daddr);
+ dev->stats.collisions++;
+ dst_release(dst);
+ goto tx_error;
+ }
+
+ sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
+ skb_reset_mac_header(skb);
+
+ if (info) {
+ const struct ip_tunnel_key *key = &info->key;
+ u8 *opts = NULL;
+ u8 vni[3];
+
+ tunnel_id_to_vni(key->tun_id, vni);
+ if (key->tun_flags & TUNNEL_GENEVE_OPT)
+ opts = ip_tunnel_info_opts(info);
+
+ udp_csum = !!(key->tun_flags & TUNNEL_CSUM);
+ err = geneve6_build_skb(dst, skb, key->tun_flags, vni,
+ info->options_len, opts,
+ udp_csum, xnet);
+ if (unlikely(err))
+ goto err;
+
+ ttl = key->ttl;
+ } else {
+ udp_csum = false;
+ err = geneve6_build_skb(dst, skb, 0, geneve->vni,
+ 0, NULL, udp_csum, xnet);
+ if (unlikely(err))
+ goto err;
+
+ ttl = geneve->ttl;
+ if (!ttl && ipv6_addr_is_multicast(&fl6.daddr))
+ ttl = 1;
+ ttl = ttl ? : ip6_dst_hoplimit(dst);
+ }
+ err = udp_tunnel6_xmit_skb(dst, gs6->sock->sk, skb, dev,
+ &fl6.saddr, &fl6.daddr, 0, ttl,
+ sport, geneve->dst_port, !udp_csum);
+
+ iptunnel_xmit_stats(err, &dev->stats, dev->tstats);
+ return NETDEV_TX_OK;
+
+tx_error:
+ dev_kfree_skb(skb);
+err:
+ dev->stats.tx_errors++;
+ return NETDEV_TX_OK;
+}
+#endif
+
+static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ struct ip_tunnel_info *info = NULL;
+
+ if (geneve->collect_md)
+ info = skb_tunnel_info(skb);
+
+#if IS_ENABLED(CONFIG_IPV6)
+ if ((info && ip_tunnel_info_af(info) == AF_INET6) ||
+ (!info && geneve->remote.sa.sa_family == AF_INET6))
+ return geneve6_xmit_skb(skb, dev, info);
+#endif
+ return geneve_xmit_skb(skb, dev, info);
+}
+
static const struct net_device_ops geneve_netdev_ops = {
.ndo_init = geneve_init,
.ndo_uninit = geneve_uninit,
@@ -759,6 +1030,7 @@ static void geneve_setup(struct net_device *dev)
static const struct nla_policy geneve_policy[IFLA_GENEVE_MAX + 1] = {
[IFLA_GENEVE_ID] = { .type = NLA_U32 },
[IFLA_GENEVE_REMOTE] = { .len = FIELD_SIZEOF(struct iphdr, daddr) },
+ [IFLA_GENEVE_REMOTE6] = { .len = sizeof(struct in6_addr) },
[IFLA_GENEVE_TTL] = { .type = NLA_U8 },
[IFLA_GENEVE_TOS] = { .type = NLA_U8 },
[IFLA_GENEVE_PORT] = { .type = NLA_U16 },
@@ -790,7 +1062,7 @@ static int geneve_validate(struct nlattr *tb[], struct nlattr *data[])
static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
__be16 dst_port,
- __be32 rem_addr,
+ union geneve_addr *remote,
u8 vni[],
bool *tun_on_same_port,
bool *tun_collect_md)
@@ -806,7 +1078,7 @@ static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
*tun_on_same_port = true;
}
if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
- rem_addr == geneve->remote.sin_addr.s_addr &&
+ !memcmp(remote, &geneve->remote, sizeof(geneve->remote)) &&
dst_port == geneve->dst_port)
t = geneve;
}
@@ -814,18 +1086,20 @@ static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
}
static int geneve_configure(struct net *net, struct net_device *dev,
- __be32 rem_addr, __u32 vni, __u8 ttl, __u8 tos,
- __be16 dst_port, bool metadata)
+ union geneve_addr *remote,
+ __u32 vni, __u8 ttl, __u8 tos, __be16 dst_port,
+ bool metadata)
{
struct geneve_net *gn = net_generic(net, geneve_net_id);
struct geneve_dev *t, *geneve = netdev_priv(dev);
bool tun_collect_md, tun_on_same_port;
int err;
- if (metadata) {
- if (rem_addr || vni || tos || ttl)
- return -EINVAL;
- }
+ if (!remote)
+ return -EINVAL;
+ if (metadata &&
+ (remote->sa.sa_family != AF_UNSPEC || vni || tos || ttl))
+ return -EINVAL;
geneve->net = net;
geneve->dev = dev;
@@ -834,16 +1108,19 @@ static int geneve_configure(struct net *net, struct net_device *dev,
geneve->vni[1] = (vni & 0x0000ff00) >> 8;
geneve->vni[2] = vni & 0x000000ff;
- geneve->remote.sin_addr.s_addr = rem_addr;
- if (IN_MULTICAST(ntohl(geneve->remote.sin_addr.s_addr)))
+ if ((remote->sa.sa_family == AF_INET &&
+ IN_MULTICAST(ntohl(remote->sin.sin_addr.s_addr))) ||
+ (remote->sa.sa_family == AF_INET6 &&
+ ipv6_addr_is_multicast(&remote->sin6.sin6_addr)))
return -EINVAL;
+ geneve->remote = *remote;
geneve->ttl = ttl;
geneve->tos = tos;
geneve->dst_port = dst_port;
geneve->collect_md = metadata;
- t = geneve_find_dev(gn, dst_port, rem_addr, geneve->vni,
+ t = geneve_find_dev(gn, dst_port, remote, geneve->vni,
&tun_on_same_port, &tun_collect_md);
if (t)
return -EBUSY;
@@ -870,15 +1147,36 @@ static int geneve_newlink(struct net *net, struct net_device *dev,
__be16 dst_port = htons(GENEVE_UDP_PORT);
__u8 ttl = 0, tos = 0;
bool metadata = false;
- __be32 rem_addr = 0;
+ union geneve_addr remote = geneve_remote_unspec;
__u32 vni = 0;
+ if (data[IFLA_GENEVE_REMOTE] && data[IFLA_GENEVE_REMOTE6])
+ return -EINVAL;
+
+ if (data[IFLA_GENEVE_REMOTE]) {
+ remote.sa.sa_family = AF_INET;
+ remote.sin.sin_addr.s_addr =
+ nla_get_in_addr(data[IFLA_GENEVE_REMOTE]);
+ }
+
+ if (data[IFLA_GENEVE_REMOTE6]) {
+ if (!IS_ENABLED(CONFIG_IPV6))
+ return -EPFNOSUPPORT;
+
+ remote.sa.sa_family = AF_INET6;
+ remote.sin6.sin6_addr =
+ nla_get_in6_addr(data[IFLA_GENEVE_REMOTE6]);
+
+ if (ipv6_addr_type(&remote.sin6.sin6_addr) &
+ IPV6_ADDR_LINKLOCAL) {
+ netdev_dbg(dev, "link-local remote is unsupported\n");
+ return -EINVAL;
+ }
+ }
+
if (data[IFLA_GENEVE_ID])
vni = nla_get_u32(data[IFLA_GENEVE_ID]);
- if (data[IFLA_GENEVE_REMOTE])
- rem_addr = nla_get_in_addr(data[IFLA_GENEVE_REMOTE]);
-
if (data[IFLA_GENEVE_TTL])
ttl = nla_get_u8(data[IFLA_GENEVE_TTL]);
@@ -891,8 +1189,8 @@ static int geneve_newlink(struct net *net, struct net_device *dev,
if (data[IFLA_GENEVE_COLLECT_METADATA])
metadata = true;
- return geneve_configure(net, dev, rem_addr, vni,
- ttl, tos, dst_port, metadata);
+ return geneve_configure(net, dev, &remote, vni, ttl, tos, dst_port,
+ metadata);
}
static void geneve_dellink(struct net_device *dev, struct list_head *head)
@@ -906,7 +1204,7 @@ static void geneve_dellink(struct net_device *dev, struct list_head *head)
static size_t geneve_get_size(const struct net_device *dev)
{
return nla_total_size(sizeof(__u32)) + /* IFLA_GENEVE_ID */
- nla_total_size(sizeof(struct in_addr)) + /* IFLA_GENEVE_REMOTE */
+ nla_total_size(sizeof(struct in6_addr)) + /* IFLA_GENEVE_REMOTE{6} */
nla_total_size(sizeof(__u8)) + /* IFLA_GENEVE_TTL */
nla_total_size(sizeof(__u8)) + /* IFLA_GENEVE_TOS */
nla_total_size(sizeof(__be16)) + /* IFLA_GENEVE_PORT */
@@ -923,9 +1221,17 @@ static int geneve_fill_info(struct sk_buff *skb, const struct net_device *dev)
if (nla_put_u32(skb, IFLA_GENEVE_ID, vni))
goto nla_put_failure;
- if (nla_put_in_addr(skb, IFLA_GENEVE_REMOTE,
- geneve->remote.sin_addr.s_addr))
- goto nla_put_failure;
+ if (geneve->remote.sa.sa_family == AF_INET) {
+ if (nla_put_in_addr(skb, IFLA_GENEVE_REMOTE,
+ geneve->remote.sin.sin_addr.s_addr))
+ goto nla_put_failure;
+#if IS_ENABLED(CONFIG_IPV6)
+ } else {
+ if (nla_put_in6_addr(skb, IFLA_GENEVE_REMOTE6,
+ &geneve->remote.sin6.sin6_addr))
+ goto nla_put_failure;
+#endif
+ }
if (nla_put_u8(skb, IFLA_GENEVE_TTL, geneve->ttl) ||
nla_put_u8(skb, IFLA_GENEVE_TOS, geneve->tos))
@@ -971,7 +1277,8 @@ struct net_device *geneve_dev_create_fb(struct net *net, const char *name,
if (IS_ERR(dev))
return dev;
- err = geneve_configure(net, dev, 0, 0, 0, 0, htons(dst_port), true);
+ err = geneve_configure(net, dev, &geneve_remote_unspec,
+ 0, 0, 0, htons(dst_port), true);
if (err) {
free_netdev(dev);
return ERR_PTR(err);
diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
index e3b6217f34f1..45e3a48550f9 100644
--- a/include/uapi/linux/if_link.h
+++ b/include/uapi/linux/if_link.h
@@ -461,6 +461,7 @@ enum {
IFLA_GENEVE_TOS,
IFLA_GENEVE_PORT, /* destination port */
IFLA_GENEVE_COLLECT_METADATA,
+ IFLA_GENEVE_REMOTE6,
__IFLA_GENEVE_MAX
};
#define IFLA_GENEVE_MAX (__IFLA_GENEVE_MAX - 1)
--
2.4.3
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH v6 2/2] geneve: handle ipv6 priority like ipv4 tos
2015-10-23 14:40 ` [PATCH v6 " John W. Linville
@ 2015-10-23 14:40 ` John W. Linville
2015-10-26 4:08 ` [PATCH v6 1/2] geneve: implement support for IPv6-based tunnels Jesse Gross
2015-10-26 21:01 ` [PATCH v7 1/3] " John W. Linville
2 siblings, 0 replies; 37+ messages in thread
From: John W. Linville @ 2015-10-23 14:40 UTC (permalink / raw)
To: netdev
Cc: Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc,
YOSHIFUJI Hideaki/吉藤英明,
John W. Linville
Other callers of udp_tunnel6_xmit_skb just pass 0 for the prio
argument. Jesse Gross <jesse@nicira.com> suggested that prio is really
the same as IPv4's tos and should be handled the same, so this is my
interpretation of that suggestion.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Reported-by: Jesse Gross <jesse@nicira.com>
Reviewed-by: Jesse Gross <jesse@nicira.com>
---
v6 -- same as previous revision
drivers/net/geneve.c | 19 +++++++++++++++++--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index 46d4aa714893..1327694c932e 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -755,6 +755,7 @@ static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
struct geneve_dev *geneve = netdev_priv(dev);
struct geneve_sock *gs6 = geneve->sock6;
struct dst_entry *dst = NULL;
+ __u8 prio;
memset(fl6, 0, sizeof(*fl6));
fl6->flowi6_mark = skb->mark;
@@ -763,7 +764,16 @@ static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
if (info) {
fl6->daddr = info->key.u.ipv6.dst;
fl6->saddr = info->key.u.ipv6.src;
+ fl6->flowi6_tos = RT_TOS(info->key.tos);
} else {
+ prio = geneve->tos;
+ if (prio == 1) {
+ const struct iphdr *iip = ip_hdr(skb);
+
+ prio = ip_tunnel_get_dsfield(iip, skb);
+ }
+
+ fl6->flowi6_tos = RT_TOS(prio);
fl6->daddr = geneve->remote.sin6.sin6_addr;
}
@@ -884,8 +894,9 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
struct geneve_dev *geneve = netdev_priv(dev);
struct geneve_sock *gs6 = geneve->sock6;
struct dst_entry *dst = NULL;
+ const struct iphdr *iip; /* interior IP header */
struct flowi6 fl6;
- __u8 ttl;
+ __u8 prio, ttl;
__be16 sport;
bool udp_csum;
int err;
@@ -914,6 +925,8 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
skb_reset_mac_header(skb);
+ iip = ip_hdr(skb);
+
if (info) {
const struct ip_tunnel_key *key = &info->key;
u8 *opts = NULL;
@@ -930,6 +943,7 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
if (unlikely(err))
goto err;
+ prio = ip_tunnel_ecn_encap(key->tos, iip, skb);
ttl = key->ttl;
} else {
udp_csum = false;
@@ -938,13 +952,14 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
if (unlikely(err))
goto err;
+ prio = ip_tunnel_ecn_encap(fl6.flowi6_tos, iip, skb);
ttl = geneve->ttl;
if (!ttl && ipv6_addr_is_multicast(&fl6.daddr))
ttl = 1;
ttl = ttl ? : ip6_dst_hoplimit(dst);
}
err = udp_tunnel6_xmit_skb(dst, gs6->sock->sk, skb, dev,
- &fl6.saddr, &fl6.daddr, 0, ttl,
+ &fl6.saddr, &fl6.daddr, prio, ttl,
sport, geneve->dst_port, !udp_csum);
iptunnel_xmit_stats(err, &dev->stats, dev->tstats);
--
2.4.3
^ permalink raw reply related [flat|nested] 37+ messages in thread
* Re: [PATCH v6 1/2] geneve: implement support for IPv6-based tunnels
2015-10-23 14:40 ` [PATCH v6 " John W. Linville
2015-10-23 14:40 ` [PATCH v6 2/2] geneve: handle ipv6 priority like ipv4 tos John W. Linville
@ 2015-10-26 4:08 ` Jesse Gross
2015-10-26 21:01 ` [PATCH v7 1/3] " John W. Linville
2 siblings, 0 replies; 37+ messages in thread
From: Jesse Gross @ 2015-10-26 4:08 UTC (permalink / raw)
To: John W. Linville
Cc: netdev, Dave Miller, Pravin B Shelar, Jiri Benc,
YOSHIFUJI Hideaki/吉藤英明
On Fri, Oct 23, 2015 at 10:40 PM, John W. Linville
<linville@tuxdriver.com> wrote:
> NOTE: Link-local IPv6 addresses for remote endpoints are not supported,
> since the driver currently has no capacity for binding a geneve
> interface to a specific link.
>
> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Reviewed-by: Jesse Gross <jesse@nicira.com>
Thanks!
By the way, there are now some merge conflicts against current
net-next due to the merge from net.
^ permalink raw reply [flat|nested] 37+ messages in thread
* [PATCH v7 1/3] geneve: implement support for IPv6-based tunnels
2015-10-23 14:40 ` [PATCH v6 " John W. Linville
2015-10-23 14:40 ` [PATCH v6 2/2] geneve: handle ipv6 priority like ipv4 tos John W. Linville
2015-10-26 4:08 ` [PATCH v6 1/2] geneve: implement support for IPv6-based tunnels Jesse Gross
@ 2015-10-26 21:01 ` John W. Linville
2015-10-26 21:01 ` [PATCH v7 2/3] geneve: handle ipv6 priority like ipv4 tos John W. Linville
` (2 more replies)
2 siblings, 3 replies; 37+ messages in thread
From: John W. Linville @ 2015-10-26 21:01 UTC (permalink / raw)
To: netdev
Cc: Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc,
YOSHIFUJI Hideaki/吉藤英明,
John W. Linville
NOTE: Link-local IPv6 addresses for remote endpoints are not supported,
since the driver currently has no capacity for binding a geneve
interface to a specific link.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Reviewed-by: Jesse Gross <jesse@nicira.com>
---
v7:
- rebase on top of commit fc4099f17240 ("openvswitch: Fix egress tunnel info.")
- revise error handling in ipv6 tx path to match ipv4 tx path (as above)
- Added Reviewed-by from Jesse based on v6 review -- changes above are minor
v6:
- fix a typo (missing {}'s)
v5:
- wrap declaration of sock6 in geneve_dev with IS_ENABLED(CONFIG_IPV6)
- remove superfluous '!!' when assigning geneve->collect_md to bool
- use skb_scrub_packet in IPv4 tx path as well
- check for NULL ip_tunnel_info pointer in geneve[6]_xmit_skb
- use ipv6_addr_equal for comparing IPv6 addresses
- more use of IS_ENABLED(CONFIG_IPV6) for preserving build integrity
- reject link-local ipv6 address for remote tunnel endpoint
v4:
- treat mode field of ip_tunnel_info as flags
- add a missing IS_ENABLED(CONFIG_IPV6) to geneve_rx
- remove unneeded flags field in geneve_dev
- NULL-check parameter for __geneve_sock_release
- check remote socket family for AF_UNSPEC in geneve_configure
- rename geneve_get_{rt,dst} as geneve_get_{v4_rt,v6_dst}
- refactor some error handling in the xmit paths
v3:
- declare geneve_remote_unspec as static
v2:
- do not require remote address for tx on metadata tunnels
- pass correct sockaddr family to udp_tun_rx_dst in geneve_rx
- accommodate both ipv4 and ipv6 sockets open on same tunnel
- move declaration of geneve_get_dst for aesthetic purposes
drivers/net/geneve.c | 473 +++++++++++++++++++++++++++++++++++--------
include/uapi/linux/if_link.h | 1 +
2 files changed, 395 insertions(+), 79 deletions(-)
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index 445071c163cb..393b0bddf7cf 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -46,16 +46,27 @@ struct geneve_net {
static int geneve_net_id;
+union geneve_addr {
+ struct sockaddr_in sin;
+ struct sockaddr_in6 sin6;
+ struct sockaddr sa;
+};
+
+static union geneve_addr geneve_remote_unspec = { .sa.sa_family = AF_UNSPEC, };
+
/* Pseudo network device */
struct geneve_dev {
struct hlist_node hlist; /* vni hash table */
struct net *net; /* netns for packet i/o */
struct net_device *dev; /* netdev for geneve tunnel */
- struct geneve_sock *sock; /* socket used for geneve tunnel */
+ struct geneve_sock *sock4; /* IPv4 socket used for geneve tunnel */
+#if IS_ENABLED(CONFIG_IPV6)
+ struct geneve_sock *sock6; /* IPv6 socket used for geneve tunnel */
+#endif
u8 vni[3]; /* virtual network ID for tunnel */
u8 ttl; /* TTL override */
u8 tos; /* TOS override */
- struct sockaddr_in remote; /* IPv4 address for link partner */
+ union geneve_addr remote; /* IP address for link partner */
struct list_head next; /* geneve's per namespace list */
__be16 dst_port;
bool collect_md;
@@ -103,11 +114,31 @@ static struct geneve_dev *geneve_lookup(struct geneve_sock *gs,
vni_list_head = &gs->vni_list[hash];
hlist_for_each_entry_rcu(geneve, vni_list_head, hlist) {
if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
- addr == geneve->remote.sin_addr.s_addr)
+ addr == geneve->remote.sin.sin_addr.s_addr)
+ return geneve;
+ }
+ return NULL;
+}
+
+#if IS_ENABLED(CONFIG_IPV6)
+static struct geneve_dev *geneve6_lookup(struct geneve_sock *gs,
+ struct in6_addr addr6, u8 vni[])
+{
+ struct hlist_head *vni_list_head;
+ struct geneve_dev *geneve;
+ __u32 hash;
+
+ /* Find the device for this VNI */
+ hash = geneve_net_vni_hash(vni);
+ vni_list_head = &gs->vni_list[hash];
+ hlist_for_each_entry_rcu(geneve, vni_list_head, hlist) {
+ if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
+ ipv6_addr_equal(&addr6, &geneve->remote.sin6.sin6_addr))
return geneve;
}
return NULL;
}
+#endif
static inline struct genevehdr *geneve_hdr(const struct sk_buff *skb)
{
@@ -121,24 +152,49 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
struct metadata_dst *tun_dst = NULL;
struct geneve_dev *geneve = NULL;
struct pcpu_sw_netstats *stats;
- struct iphdr *iph;
- u8 *vni;
+ struct iphdr *iph = NULL;
__be32 addr;
- int err;
+ static u8 zero_vni[3];
+ u8 *vni;
+ int err = 0;
+ sa_family_t sa_family;
+#if IS_ENABLED(CONFIG_IPV6)
+ struct ipv6hdr *ip6h = NULL;
+ struct in6_addr addr6;
+ static struct in6_addr zero_addr6;
+#endif
- iph = ip_hdr(skb); /* outer IP header... */
+ sa_family = gs->sock->sk->sk_family;
- if (gs->collect_md) {
- static u8 zero_vni[3];
+ if (sa_family == AF_INET) {
+ iph = ip_hdr(skb); /* outer IP header... */
- vni = zero_vni;
- addr = 0;
- } else {
- vni = gnvh->vni;
- addr = iph->saddr;
- }
+ if (gs->collect_md) {
+ vni = zero_vni;
+ addr = 0;
+ } else {
+ vni = gnvh->vni;
+
+ addr = iph->saddr;
+ }
+
+ geneve = geneve_lookup(gs, addr, vni);
+#if IS_ENABLED(CONFIG_IPV6)
+ } else if (sa_family == AF_INET6) {
+ ip6h = ipv6_hdr(skb); /* outer IPv6 header... */
- geneve = geneve_lookup(gs, addr, vni);
+ if (gs->collect_md) {
+ vni = zero_vni;
+ addr6 = zero_addr6;
+ } else {
+ vni = gnvh->vni;
+
+ addr6 = ip6h->saddr;
+ }
+
+ geneve = geneve6_lookup(gs, addr6, vni);
+#endif
+ }
if (!geneve)
goto drop;
@@ -149,7 +205,7 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
(gnvh->oam ? TUNNEL_OAM : 0) |
(gnvh->critical ? TUNNEL_CRIT_OPT : 0);
- tun_dst = udp_tun_rx_dst(skb, AF_INET, flags,
+ tun_dst = udp_tun_rx_dst(skb, sa_family, flags,
vni_to_tunnel_id(gnvh->vni),
gnvh->opt_len * 4);
if (!tun_dst)
@@ -179,12 +235,25 @@ static void geneve_rx(struct geneve_sock *gs, struct sk_buff *skb)
skb_reset_network_header(skb);
- err = IP_ECN_decapsulate(iph, skb);
+ if (iph)
+ err = IP_ECN_decapsulate(iph, skb);
+#if IS_ENABLED(CONFIG_IPV6)
+ if (ip6h)
+ err = IP6_ECN_decapsulate(ip6h, skb);
+#endif
if (unlikely(err)) {
- if (log_ecn_error)
- net_info_ratelimited("non-ECT from %pI4 with TOS=%#x\n",
- &iph->saddr, iph->tos);
+ if (log_ecn_error) {
+ if (iph)
+ net_info_ratelimited("non-ECT from %pI4 "
+ "with TOS=%#x\n",
+ &iph->saddr, iph->tos);
+#if IS_ENABLED(CONFIG_IPV6)
+ if (ip6h)
+ net_info_ratelimited("non-ECT from %pI6\n",
+ &ip6h->saddr);
+#endif
+ }
if (err > 1) {
++geneve->dev->stats.rx_frame_errors;
++geneve->dev->stats.rx_errors;
@@ -284,6 +353,7 @@ static struct socket *geneve_create_sock(struct net *net, bool ipv6,
if (ipv6) {
udp_conf.family = AF_INET6;
+ udp_conf.ipv6_v6only = 1;
} else {
udp_conf.family = AF_INET;
udp_conf.local_ip.s_addr = htonl(INADDR_ANY);
@@ -458,9 +528,9 @@ static void geneve_notify_del_rx_port(struct geneve_sock *gs)
udp_del_offload(&gs->udp_offloads);
}
-static void geneve_sock_release(struct geneve_sock *gs)
+static void __geneve_sock_release(struct geneve_sock *gs)
{
- if (--gs->refcnt)
+ if (!gs || --gs->refcnt)
return;
list_del(&gs->list);
@@ -469,66 +539,117 @@ static void geneve_sock_release(struct geneve_sock *gs)
kfree_rcu(gs, rcu);
}
+static void geneve_sock_release(struct geneve_dev *geneve)
+{
+ __geneve_sock_release(geneve->sock4);
+#if IS_ENABLED(CONFIG_IPV6)
+ __geneve_sock_release(geneve->sock6);
+#endif
+}
+
static struct geneve_sock *geneve_find_sock(struct geneve_net *gn,
+ sa_family_t family,
__be16 dst_port)
{
struct geneve_sock *gs;
list_for_each_entry(gs, &gn->sock_list, list) {
if (inet_sk(gs->sock->sk)->inet_sport == dst_port &&
- inet_sk(gs->sock->sk)->sk.sk_family == AF_INET) {
+ inet_sk(gs->sock->sk)->sk.sk_family == family) {
return gs;
}
}
return NULL;
}
-static int geneve_open(struct net_device *dev)
+static int geneve_sock_add(struct geneve_dev *geneve, bool ipv6)
{
- struct geneve_dev *geneve = netdev_priv(dev);
struct net *net = geneve->net;
struct geneve_net *gn = net_generic(net, geneve_net_id);
struct geneve_sock *gs;
__u32 hash;
- gs = geneve_find_sock(gn, geneve->dst_port);
+ gs = geneve_find_sock(gn, ipv6 ? AF_INET6 : AF_INET, geneve->dst_port);
if (gs) {
gs->refcnt++;
goto out;
}
- gs = geneve_socket_create(net, geneve->dst_port, false);
+ gs = geneve_socket_create(net, geneve->dst_port, ipv6);
if (IS_ERR(gs))
return PTR_ERR(gs);
out:
gs->collect_md = geneve->collect_md;
- geneve->sock = gs;
+#if IS_ENABLED(CONFIG_IPV6)
+ if (ipv6)
+ geneve->sock6 = gs;
+ else
+#endif
+ geneve->sock4 = gs;
hash = geneve_net_vni_hash(geneve->vni);
hlist_add_head_rcu(&geneve->hlist, &gs->vni_list[hash]);
return 0;
}
+static int geneve_open(struct net_device *dev)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ bool ipv6 = geneve->remote.sa.sa_family == AF_INET6;
+ bool metadata = geneve->collect_md;
+ int ret = 0;
+
+ geneve->sock4 = NULL;
+#if IS_ENABLED(CONFIG_IPV6)
+ geneve->sock6 = NULL;
+ if (ipv6 || metadata)
+ ret = geneve_sock_add(geneve, true);
+#endif
+ if (!ret && (!ipv6 || metadata))
+ ret = geneve_sock_add(geneve, false);
+ if (ret < 0)
+ geneve_sock_release(geneve);
+
+ return ret;
+}
+
static int geneve_stop(struct net_device *dev)
{
struct geneve_dev *geneve = netdev_priv(dev);
- struct geneve_sock *gs = geneve->sock;
if (!hlist_unhashed(&geneve->hlist))
hlist_del_rcu(&geneve->hlist);
- geneve_sock_release(gs);
+ geneve_sock_release(geneve);
return 0;
}
+static void geneve_build_header(struct genevehdr *geneveh,
+ __be16 tun_flags, u8 vni[3],
+ u8 options_len, u8 *options)
+{
+ geneveh->ver = GENEVE_VER;
+ geneveh->opt_len = options_len / 4;
+ geneveh->oam = !!(tun_flags & TUNNEL_OAM);
+ geneveh->critical = !!(tun_flags & TUNNEL_CRIT_OPT);
+ geneveh->rsvd1 = 0;
+ memcpy(geneveh->vni, vni, 3);
+ geneveh->proto_type = htons(ETH_P_TEB);
+ geneveh->rsvd2 = 0;
+
+ memcpy(geneveh->options, options, options_len);
+}
+
static int geneve_build_skb(struct rtable *rt, struct sk_buff *skb,
__be16 tun_flags, u8 vni[3], u8 opt_len, u8 *opt,
- bool csum)
+ bool csum, bool xnet)
{
struct genevehdr *gnvh;
int min_headroom;
int err;
+ skb_scrub_packet(skb, xnet);
+
min_headroom = LL_RESERVED_SPACE(rt->dst.dev) + rt->dst.header_len
+ GENEVE_BASE_HLEN + opt_len + sizeof(struct iphdr);
err = skb_cow_head(skb, min_headroom);
@@ -544,15 +665,7 @@ static int geneve_build_skb(struct rtable *rt, struct sk_buff *skb,
}
gnvh = (struct genevehdr *)__skb_push(skb, sizeof(*gnvh) + opt_len);
- gnvh->ver = GENEVE_VER;
- gnvh->opt_len = opt_len / 4;
- gnvh->oam = !!(tun_flags & TUNNEL_OAM);
- gnvh->critical = !!(tun_flags & TUNNEL_CRIT_OPT);
- gnvh->rsvd1 = 0;
- memcpy(gnvh->vni, vni, 3);
- gnvh->proto_type = htons(ETH_P_TEB);
- gnvh->rsvd2 = 0;
- memcpy(gnvh->options, opt, opt_len);
+ geneve_build_header(gnvh, tun_flags, vni, opt_len, opt);
skb_set_inner_protocol(skb, htons(ETH_P_TEB));
return 0;
@@ -562,10 +675,47 @@ free_rt:
return err;
}
-static struct rtable *geneve_get_rt(struct sk_buff *skb,
- struct net_device *dev,
- struct flowi4 *fl4,
- struct ip_tunnel_info *info)
+#if IS_ENABLED(CONFIG_IPV6)
+static int geneve6_build_skb(struct dst_entry *dst, struct sk_buff *skb,
+ __be16 tun_flags, u8 vni[3], u8 opt_len, u8 *opt,
+ bool csum, bool xnet)
+{
+ struct genevehdr *gnvh;
+ int min_headroom;
+ int err;
+
+ skb_scrub_packet(skb, xnet);
+
+ min_headroom = LL_RESERVED_SPACE(dst->dev) + dst->header_len
+ + GENEVE_BASE_HLEN + opt_len + sizeof(struct ipv6hdr);
+ err = skb_cow_head(skb, min_headroom);
+ if (unlikely(err)) {
+ kfree_skb(skb);
+ goto free_dst;
+ }
+
+ skb = udp_tunnel_handle_offloads(skb, csum);
+ if (IS_ERR(skb)) {
+ err = PTR_ERR(skb);
+ goto free_dst;
+ }
+
+ gnvh = (struct genevehdr *)__skb_push(skb, sizeof(*gnvh) + opt_len);
+ geneve_build_header(gnvh, tun_flags, vni, opt_len, opt);
+
+ skb_set_inner_protocol(skb, htons(ETH_P_TEB));
+ return 0;
+
+free_dst:
+ dst_release(dst);
+ return err;
+}
+#endif
+
+static struct rtable *geneve_get_v4_rt(struct sk_buff *skb,
+ struct net_device *dev,
+ struct flowi4 *fl4,
+ struct ip_tunnel_info *info)
{
struct geneve_dev *geneve = netdev_priv(dev);
struct rtable *rt = NULL;
@@ -588,7 +738,7 @@ static struct rtable *geneve_get_rt(struct sk_buff *skb,
}
fl4->flowi4_tos = RT_TOS(tos);
- fl4->daddr = geneve->remote.sin_addr.s_addr;
+ fl4->daddr = geneve->remote.sin.sin_addr.s_addr;
}
rt = ip_route_output_key(geneve->net, fl4);
@@ -604,6 +754,41 @@ static struct rtable *geneve_get_rt(struct sk_buff *skb,
return rt;
}
+#if IS_ENABLED(CONFIG_IPV6)
+static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
+ struct net_device *dev,
+ struct flowi6 *fl6,
+ struct ip_tunnel_info *info)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ struct geneve_sock *gs6 = geneve->sock6;
+ struct dst_entry *dst = NULL;
+
+ memset(fl6, 0, sizeof(*fl6));
+ fl6->flowi6_mark = skb->mark;
+ fl6->flowi6_proto = IPPROTO_UDP;
+
+ if (info) {
+ fl6->daddr = info->key.u.ipv6.dst;
+ fl6->saddr = info->key.u.ipv6.src;
+ } else {
+ fl6->daddr = geneve->remote.sin6.sin6_addr;
+ }
+
+ if (ipv6_stub->ipv6_dst_lookup(geneve->net, gs6->sock->sk, &dst, fl6)) {
+ netdev_dbg(dev, "no route to %pI6\n", &fl6->daddr);
+ return ERR_PTR(-ENETUNREACH);
+ }
+ if (dst->dev == dev) { /* is this necessary? */
+ netdev_dbg(dev, "circular route to %pI6\n", &fl6->daddr);
+ dst_release(dst);
+ return ERR_PTR(-ELOOP);
+ }
+
+ return dst;
+}
+#endif
+
/* Convert 64 bit tunnel ID to 24 bit VNI. */
static void tunnel_id_to_vni(__be64 tun_id, __u8 *vni)
{
@@ -618,11 +803,11 @@ static void tunnel_id_to_vni(__be64 tun_id, __u8 *vni)
#endif
}
-static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
+static netdev_tx_t geneve_xmit_skb(struct sk_buff *skb, struct net_device *dev,
+ struct ip_tunnel_info *info)
{
struct geneve_dev *geneve = netdev_priv(dev);
- struct geneve_sock *gs = geneve->sock;
- struct ip_tunnel_info *info = NULL;
+ struct geneve_sock *gs4 = geneve->sock4;
struct rtable *rt = NULL;
const struct iphdr *iip; /* interior IP header */
int err = -EINVAL;
@@ -631,10 +816,10 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
__be16 sport;
bool udp_csum;
__be16 df;
+ bool xnet = !net_eq(geneve->net, dev_net(geneve->dev));
if (geneve->collect_md) {
- info = skb_tunnel_info(skb);
- if (unlikely(info && !(info->mode & IP_TUNNEL_INFO_TX))) {
+ if (unlikely(!info || !(info->mode & IP_TUNNEL_INFO_TX))) {
netdev_dbg(dev, "no tunnel metadata\n");
goto tx_error;
}
@@ -642,9 +827,8 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
goto tx_error;
}
- rt = geneve_get_rt(skb, dev, &fl4, info);
+ rt = geneve_get_v4_rt(skb, dev, &fl4, info);
if (IS_ERR(rt)) {
- netdev_dbg(dev, "no route to %pI4\n", &fl4.daddr);
err = PTR_ERR(rt);
goto tx_error;
}
@@ -665,7 +849,7 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
udp_csum = !!(key->tun_flags & TUNNEL_CSUM);
err = geneve_build_skb(rt, skb, key->tun_flags, vni,
- info->options_len, opts, udp_csum);
+ info->options_len, opts, udp_csum, xnet);
if (unlikely(err))
goto err;
@@ -675,7 +859,7 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
} else {
udp_csum = false;
err = geneve_build_skb(rt, skb, 0, geneve->vni,
- 0, NULL, udp_csum);
+ 0, NULL, udp_csum, xnet);
if (unlikely(err))
goto err;
@@ -686,7 +870,7 @@ static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
ttl = ttl ? : ip4_dst_hoplimit(&rt->dst);
df = 0;
}
- err = udp_tunnel_xmit_skb(rt, gs->sock->sk, skb, fl4.saddr, fl4.daddr,
+ err = udp_tunnel_xmit_skb(rt, gs4->sock->sk, skb, fl4.saddr, fl4.daddr,
tos, ttl, df, sport, geneve->dst_port,
!net_eq(geneve->net, dev_net(geneve->dev)),
!udp_csum);
@@ -706,6 +890,101 @@ err:
return NETDEV_TX_OK;
}
+#if IS_ENABLED(CONFIG_IPV6)
+static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
+ struct ip_tunnel_info *info)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ struct geneve_sock *gs6 = geneve->sock6;
+ struct dst_entry *dst = NULL;
+ int err = -EINVAL;
+ struct flowi6 fl6;
+ __u8 ttl;
+ __be16 sport;
+ bool udp_csum;
+ bool xnet = !net_eq(geneve->net, dev_net(geneve->dev));
+
+ if (geneve->collect_md) {
+ if (unlikely(!info || !(info->mode & IP_TUNNEL_INFO_TX))) {
+ netdev_dbg(dev, "no tunnel metadata\n");
+ goto tx_error;
+ }
+ }
+
+ dst = geneve_get_v6_dst(skb, dev, &fl6, info);
+ if (IS_ERR(dst)) {
+ err = PTR_ERR(dst);
+ goto tx_error;
+ }
+
+ sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
+ skb_reset_mac_header(skb);
+
+ if (info) {
+ const struct ip_tunnel_key *key = &info->key;
+ u8 *opts = NULL;
+ u8 vni[3];
+
+ tunnel_id_to_vni(key->tun_id, vni);
+ if (key->tun_flags & TUNNEL_GENEVE_OPT)
+ opts = ip_tunnel_info_opts(info);
+
+ udp_csum = !!(key->tun_flags & TUNNEL_CSUM);
+ err = geneve6_build_skb(dst, skb, key->tun_flags, vni,
+ info->options_len, opts,
+ udp_csum, xnet);
+ if (unlikely(err))
+ goto err;
+
+ ttl = key->ttl;
+ } else {
+ udp_csum = false;
+ err = geneve6_build_skb(dst, skb, 0, geneve->vni,
+ 0, NULL, udp_csum, xnet);
+ if (unlikely(err))
+ goto err;
+
+ ttl = geneve->ttl;
+ if (!ttl && ipv6_addr_is_multicast(&fl6.daddr))
+ ttl = 1;
+ ttl = ttl ? : ip6_dst_hoplimit(dst);
+ }
+ err = udp_tunnel6_xmit_skb(dst, gs6->sock->sk, skb, dev,
+ &fl6.saddr, &fl6.daddr, 0, ttl,
+ sport, geneve->dst_port, !udp_csum);
+
+ iptunnel_xmit_stats(err, &dev->stats, dev->tstats);
+ return NETDEV_TX_OK;
+
+tx_error:
+ dev_kfree_skb(skb);
+err:
+ if (err == -ELOOP)
+ dev->stats.collisions++;
+ else if (err == -ENETUNREACH)
+ dev->stats.tx_carrier_errors++;
+ else
+ dev->stats.tx_errors++;
+ return NETDEV_TX_OK;
+}
+#endif
+
+static netdev_tx_t geneve_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+ struct geneve_dev *geneve = netdev_priv(dev);
+ struct ip_tunnel_info *info = NULL;
+
+ if (geneve->collect_md)
+ info = skb_tunnel_info(skb);
+
+#if IS_ENABLED(CONFIG_IPV6)
+ if ((info && ip_tunnel_info_af(info) == AF_INET6) ||
+ (!info && geneve->remote.sa.sa_family == AF_INET6))
+ return geneve6_xmit_skb(skb, dev, info);
+#endif
+ return geneve_xmit_skb(skb, dev, info);
+}
+
static int geneve_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
{
struct ip_tunnel_info *info = skb_tunnel_info(skb);
@@ -716,7 +995,7 @@ static int geneve_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
if (ip_tunnel_info_af(info) != AF_INET)
return -EINVAL;
- rt = geneve_get_rt(skb, dev, &fl4, info);
+ rt = geneve_get_v4_rt(skb, dev, &fl4, info);
if (IS_ERR(rt))
return PTR_ERR(rt);
@@ -785,6 +1064,7 @@ static void geneve_setup(struct net_device *dev)
static const struct nla_policy geneve_policy[IFLA_GENEVE_MAX + 1] = {
[IFLA_GENEVE_ID] = { .type = NLA_U32 },
[IFLA_GENEVE_REMOTE] = { .len = FIELD_SIZEOF(struct iphdr, daddr) },
+ [IFLA_GENEVE_REMOTE6] = { .len = sizeof(struct in6_addr) },
[IFLA_GENEVE_TTL] = { .type = NLA_U8 },
[IFLA_GENEVE_TOS] = { .type = NLA_U8 },
[IFLA_GENEVE_PORT] = { .type = NLA_U16 },
@@ -816,7 +1096,7 @@ static int geneve_validate(struct nlattr *tb[], struct nlattr *data[])
static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
__be16 dst_port,
- __be32 rem_addr,
+ union geneve_addr *remote,
u8 vni[],
bool *tun_on_same_port,
bool *tun_collect_md)
@@ -832,7 +1112,7 @@ static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
*tun_on_same_port = true;
}
if (!memcmp(vni, geneve->vni, sizeof(geneve->vni)) &&
- rem_addr == geneve->remote.sin_addr.s_addr &&
+ !memcmp(remote, &geneve->remote, sizeof(geneve->remote)) &&
dst_port == geneve->dst_port)
t = geneve;
}
@@ -840,18 +1120,20 @@ static struct geneve_dev *geneve_find_dev(struct geneve_net *gn,
}
static int geneve_configure(struct net *net, struct net_device *dev,
- __be32 rem_addr, __u32 vni, __u8 ttl, __u8 tos,
- __be16 dst_port, bool metadata)
+ union geneve_addr *remote,
+ __u32 vni, __u8 ttl, __u8 tos, __be16 dst_port,
+ bool metadata)
{
struct geneve_net *gn = net_generic(net, geneve_net_id);
struct geneve_dev *t, *geneve = netdev_priv(dev);
bool tun_collect_md, tun_on_same_port;
int err;
- if (metadata) {
- if (rem_addr || vni || tos || ttl)
- return -EINVAL;
- }
+ if (!remote)
+ return -EINVAL;
+ if (metadata &&
+ (remote->sa.sa_family != AF_UNSPEC || vni || tos || ttl))
+ return -EINVAL;
geneve->net = net;
geneve->dev = dev;
@@ -860,16 +1142,19 @@ static int geneve_configure(struct net *net, struct net_device *dev,
geneve->vni[1] = (vni & 0x0000ff00) >> 8;
geneve->vni[2] = vni & 0x000000ff;
- geneve->remote.sin_addr.s_addr = rem_addr;
- if (IN_MULTICAST(ntohl(geneve->remote.sin_addr.s_addr)))
+ if ((remote->sa.sa_family == AF_INET &&
+ IN_MULTICAST(ntohl(remote->sin.sin_addr.s_addr))) ||
+ (remote->sa.sa_family == AF_INET6 &&
+ ipv6_addr_is_multicast(&remote->sin6.sin6_addr)))
return -EINVAL;
+ geneve->remote = *remote;
geneve->ttl = ttl;
geneve->tos = tos;
geneve->dst_port = dst_port;
geneve->collect_md = metadata;
- t = geneve_find_dev(gn, dst_port, rem_addr, geneve->vni,
+ t = geneve_find_dev(gn, dst_port, remote, geneve->vni,
&tun_on_same_port, &tun_collect_md);
if (t)
return -EBUSY;
@@ -896,15 +1181,36 @@ static int geneve_newlink(struct net *net, struct net_device *dev,
__be16 dst_port = htons(GENEVE_UDP_PORT);
__u8 ttl = 0, tos = 0;
bool metadata = false;
- __be32 rem_addr = 0;
+ union geneve_addr remote = geneve_remote_unspec;
__u32 vni = 0;
+ if (data[IFLA_GENEVE_REMOTE] && data[IFLA_GENEVE_REMOTE6])
+ return -EINVAL;
+
+ if (data[IFLA_GENEVE_REMOTE]) {
+ remote.sa.sa_family = AF_INET;
+ remote.sin.sin_addr.s_addr =
+ nla_get_in_addr(data[IFLA_GENEVE_REMOTE]);
+ }
+
+ if (data[IFLA_GENEVE_REMOTE6]) {
+ if (!IS_ENABLED(CONFIG_IPV6))
+ return -EPFNOSUPPORT;
+
+ remote.sa.sa_family = AF_INET6;
+ remote.sin6.sin6_addr =
+ nla_get_in6_addr(data[IFLA_GENEVE_REMOTE6]);
+
+ if (ipv6_addr_type(&remote.sin6.sin6_addr) &
+ IPV6_ADDR_LINKLOCAL) {
+ netdev_dbg(dev, "link-local remote is unsupported\n");
+ return -EINVAL;
+ }
+ }
+
if (data[IFLA_GENEVE_ID])
vni = nla_get_u32(data[IFLA_GENEVE_ID]);
- if (data[IFLA_GENEVE_REMOTE])
- rem_addr = nla_get_in_addr(data[IFLA_GENEVE_REMOTE]);
-
if (data[IFLA_GENEVE_TTL])
ttl = nla_get_u8(data[IFLA_GENEVE_TTL]);
@@ -917,8 +1223,8 @@ static int geneve_newlink(struct net *net, struct net_device *dev,
if (data[IFLA_GENEVE_COLLECT_METADATA])
metadata = true;
- return geneve_configure(net, dev, rem_addr, vni,
- ttl, tos, dst_port, metadata);
+ return geneve_configure(net, dev, &remote, vni, ttl, tos, dst_port,
+ metadata);
}
static void geneve_dellink(struct net_device *dev, struct list_head *head)
@@ -932,7 +1238,7 @@ static void geneve_dellink(struct net_device *dev, struct list_head *head)
static size_t geneve_get_size(const struct net_device *dev)
{
return nla_total_size(sizeof(__u32)) + /* IFLA_GENEVE_ID */
- nla_total_size(sizeof(struct in_addr)) + /* IFLA_GENEVE_REMOTE */
+ nla_total_size(sizeof(struct in6_addr)) + /* IFLA_GENEVE_REMOTE{6} */
nla_total_size(sizeof(__u8)) + /* IFLA_GENEVE_TTL */
nla_total_size(sizeof(__u8)) + /* IFLA_GENEVE_TOS */
nla_total_size(sizeof(__be16)) + /* IFLA_GENEVE_PORT */
@@ -949,9 +1255,17 @@ static int geneve_fill_info(struct sk_buff *skb, const struct net_device *dev)
if (nla_put_u32(skb, IFLA_GENEVE_ID, vni))
goto nla_put_failure;
- if (nla_put_in_addr(skb, IFLA_GENEVE_REMOTE,
- geneve->remote.sin_addr.s_addr))
- goto nla_put_failure;
+ if (geneve->remote.sa.sa_family == AF_INET) {
+ if (nla_put_in_addr(skb, IFLA_GENEVE_REMOTE,
+ geneve->remote.sin.sin_addr.s_addr))
+ goto nla_put_failure;
+#if IS_ENABLED(CONFIG_IPV6)
+ } else {
+ if (nla_put_in6_addr(skb, IFLA_GENEVE_REMOTE6,
+ &geneve->remote.sin6.sin6_addr))
+ goto nla_put_failure;
+#endif
+ }
if (nla_put_u8(skb, IFLA_GENEVE_TTL, geneve->ttl) ||
nla_put_u8(skb, IFLA_GENEVE_TOS, geneve->tos))
@@ -997,7 +1311,8 @@ struct net_device *geneve_dev_create_fb(struct net *net, const char *name,
if (IS_ERR(dev))
return dev;
- err = geneve_configure(net, dev, 0, 0, 0, 0, htons(dst_port), true);
+ err = geneve_configure(net, dev, &geneve_remote_unspec,
+ 0, 0, 0, htons(dst_port), true);
if (err) {
free_netdev(dev);
return ERR_PTR(err);
diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
index a7aea8418abb..5ad57375a99f 100644
--- a/include/uapi/linux/if_link.h
+++ b/include/uapi/linux/if_link.h
@@ -461,6 +461,7 @@ enum {
IFLA_GENEVE_TOS,
IFLA_GENEVE_PORT, /* destination port */
IFLA_GENEVE_COLLECT_METADATA,
+ IFLA_GENEVE_REMOTE6,
__IFLA_GENEVE_MAX
};
#define IFLA_GENEVE_MAX (__IFLA_GENEVE_MAX - 1)
--
2.4.3
^ permalink raw reply related [flat|nested] 37+ messages in thread
* [PATCH v7 2/3] geneve: handle ipv6 priority like ipv4 tos
2015-10-26 21:01 ` [PATCH v7 1/3] " John W. Linville
@ 2015-10-26 21:01 ` John W. Linville
2015-10-30 3:11 ` David Miller
2015-10-26 21:01 ` [PATCH v7 3/3] geneve: add IPv6 bits to geneve_fill_metadata_dst John W. Linville
2015-10-30 3:11 ` [PATCH v7 1/3] geneve: implement support for IPv6-based tunnels David Miller
2 siblings, 1 reply; 37+ messages in thread
From: John W. Linville @ 2015-10-26 21:01 UTC (permalink / raw)
To: netdev
Cc: Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc,
YOSHIFUJI Hideaki/吉藤英明,
John W. Linville
Other callers of udp_tunnel6_xmit_skb just pass 0 for the prio
argument. Jesse Gross <jesse@nicira.com> suggested that prio is really
the same as IPv4's tos and should be handled the same, so this is my
interpretation of that suggestion.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Reported-by: Jesse Gross <jesse@nicira.com>
Reviewed-by: Jesse Gross <jesse@nicira.com>
---
v7 -- same as previous revisions
drivers/net/geneve.c | 19 +++++++++++++++++--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index 393b0bddf7cf..44e724508c55 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -763,6 +763,7 @@ static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
struct geneve_dev *geneve = netdev_priv(dev);
struct geneve_sock *gs6 = geneve->sock6;
struct dst_entry *dst = NULL;
+ __u8 prio;
memset(fl6, 0, sizeof(*fl6));
fl6->flowi6_mark = skb->mark;
@@ -771,7 +772,16 @@ static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb,
if (info) {
fl6->daddr = info->key.u.ipv6.dst;
fl6->saddr = info->key.u.ipv6.src;
+ fl6->flowi6_tos = RT_TOS(info->key.tos);
} else {
+ prio = geneve->tos;
+ if (prio == 1) {
+ const struct iphdr *iip = ip_hdr(skb);
+
+ prio = ip_tunnel_get_dsfield(iip, skb);
+ }
+
+ fl6->flowi6_tos = RT_TOS(prio);
fl6->daddr = geneve->remote.sin6.sin6_addr;
}
@@ -897,9 +907,10 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
struct geneve_dev *geneve = netdev_priv(dev);
struct geneve_sock *gs6 = geneve->sock6;
struct dst_entry *dst = NULL;
+ const struct iphdr *iip; /* interior IP header */
int err = -EINVAL;
struct flowi6 fl6;
- __u8 ttl;
+ __u8 prio, ttl;
__be16 sport;
bool udp_csum;
bool xnet = !net_eq(geneve->net, dev_net(geneve->dev));
@@ -920,6 +931,8 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
sport = udp_flow_src_port(geneve->net, skb, 1, USHRT_MAX, true);
skb_reset_mac_header(skb);
+ iip = ip_hdr(skb);
+
if (info) {
const struct ip_tunnel_key *key = &info->key;
u8 *opts = NULL;
@@ -936,6 +949,7 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
if (unlikely(err))
goto err;
+ prio = ip_tunnel_ecn_encap(key->tos, iip, skb);
ttl = key->ttl;
} else {
udp_csum = false;
@@ -944,13 +958,14 @@ static netdev_tx_t geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
if (unlikely(err))
goto err;
+ prio = ip_tunnel_ecn_encap(fl6.flowi6_tos, iip, skb);
ttl = geneve->ttl;
if (!ttl && ipv6_addr_is_multicast(&fl6.daddr))
ttl = 1;
ttl = ttl ? : ip6_dst_hoplimit(dst);
}
err = udp_tunnel6_xmit_skb(dst, gs6->sock->sk, skb, dev,
- &fl6.saddr, &fl6.daddr, 0, ttl,
+ &fl6.saddr, &fl6.daddr, prio, ttl,
sport, geneve->dst_port, !udp_csum);
iptunnel_xmit_stats(err, &dev->stats, dev->tstats);
--
2.4.3
^ permalink raw reply related [flat|nested] 37+ messages in thread
* Re: [PATCH v7 2/3] geneve: handle ipv6 priority like ipv4 tos
2015-10-26 21:01 ` [PATCH v7 2/3] geneve: handle ipv6 priority like ipv4 tos John W. Linville
@ 2015-10-30 3:11 ` David Miller
0 siblings, 0 replies; 37+ messages in thread
From: David Miller @ 2015-10-30 3:11 UTC (permalink / raw)
To: linville; +Cc: netdev, pshelar, jesse, jbenc, hideaki.yoshifuji
From: "John W. Linville" <linville@tuxdriver.com>
Date: Mon, 26 Oct 2015 17:01:45 -0400
> Other callers of udp_tunnel6_xmit_skb just pass 0 for the prio
> argument. Jesse Gross <jesse@nicira.com> suggested that prio is really
> the same as IPv4's tos and should be handled the same, so this is my
> interpretation of that suggestion.
>
> Signed-off-by: John W. Linville <linville@tuxdriver.com>
> Reported-by: Jesse Gross <jesse@nicira.com>
> Reviewed-by: Jesse Gross <jesse@nicira.com>
Applied.
^ permalink raw reply [flat|nested] 37+ messages in thread
* [PATCH v7 3/3] geneve: add IPv6 bits to geneve_fill_metadata_dst
2015-10-26 21:01 ` [PATCH v7 1/3] " John W. Linville
2015-10-26 21:01 ` [PATCH v7 2/3] geneve: handle ipv6 priority like ipv4 tos John W. Linville
@ 2015-10-26 21:01 ` John W. Linville
2015-10-27 12:48 ` Sergei Shtylyov
2015-10-27 13:49 ` [PATCH v8 " John W. Linville
2015-10-30 3:11 ` [PATCH v7 1/3] geneve: implement support for IPv6-based tunnels David Miller
2 siblings, 2 replies; 37+ messages in thread
From: John W. Linville @ 2015-10-26 21:01 UTC (permalink / raw)
To: netdev
Cc: Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc,
YOSHIFUJI Hideaki/吉藤英明,
John W. Linville
Signed-off-by: John W. Linville <linville@tuxdriver.com>
---
v7 -- initial version (numbered to match earlier patches in series)
drivers/net/geneve.c | 28 +++++++++++++++++++++-------
1 file changed, 21 insertions(+), 7 deletions(-)
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index 44e724508c55..be532d7b879d 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -1006,16 +1006,30 @@ static int geneve_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
struct geneve_dev *geneve = netdev_priv(dev);
struct rtable *rt;
struct flowi4 fl4;
+#if IS_ENABLED(CONFIG_IPV6)
+ struct dst_entry *dst;
+ struct flowi6 fl6;
+#endif
- if (ip_tunnel_info_af(info) != AF_INET)
- return -EINVAL;
+ if (ip_tunnel_info_af(info) == AF_INET) {
+ rt = geneve_get_v4_rt(skb, dev, &fl4, info);
+ if (IS_ERR(rt))
+ return PTR_ERR(rt);
- rt = geneve_get_v4_rt(skb, dev, &fl4, info);
- if (IS_ERR(rt))
- return PTR_ERR(rt);
+ ip_rt_put(rt);
+ info->key.u.ipv4.src = fl4.saddr;
+#if IS_ENABLED(CONFIG_IPV6)
+ } else if (ip_tunnel_info_af(info) == AF_INET6) {
+ dst = geneve_get_v6_dst(skb, dev, &fl6, info);
+ if (IS_ERR(dst))
+ return PTR_ERR(dst);
+
+ dst_release(dst);
+ info->key.u.ipv6.src = fl6.saddr;
+#endif
+ } else
+ return -EINVAL;
- ip_rt_put(rt);
- info->key.u.ipv4.src = fl4.saddr;
info->key.tp_src = udp_flow_src_port(geneve->net, skb,
1, USHRT_MAX, true);
info->key.tp_dst = geneve->dst_port;
--
2.4.3
^ permalink raw reply related [flat|nested] 37+ messages in thread
* Re: [PATCH v7 3/3] geneve: add IPv6 bits to geneve_fill_metadata_dst
2015-10-26 21:01 ` [PATCH v7 3/3] geneve: add IPv6 bits to geneve_fill_metadata_dst John W. Linville
@ 2015-10-27 12:48 ` Sergei Shtylyov
2015-10-27 13:49 ` [PATCH v8 " John W. Linville
1 sibling, 0 replies; 37+ messages in thread
From: Sergei Shtylyov @ 2015-10-27 12:48 UTC (permalink / raw)
To: John W. Linville, netdev
Cc: Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc,
YOSHIFUJI Hideaki/吉藤英明
Hello.
On 10/27/2015 12:01 AM, John W. Linville wrote:
> Signed-off-by: John W. Linville <linville@tuxdriver.com>
> ---
> v7 -- initial version (numbered to match earlier patches in series)
>
> drivers/net/geneve.c | 28 +++++++++++++++++++++-------
> 1 file changed, 21 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
> index 44e724508c55..be532d7b879d 100644
> --- a/drivers/net/geneve.c
> +++ b/drivers/net/geneve.c
> @@ -1006,16 +1006,30 @@ static int geneve_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
> struct geneve_dev *geneve = netdev_priv(dev);
> struct rtable *rt;
> struct flowi4 fl4;
> +#if IS_ENABLED(CONFIG_IPV6)
> + struct dst_entry *dst;
> + struct flowi6 fl6;
> +#endif
>
> - if (ip_tunnel_info_af(info) != AF_INET)
> - return -EINVAL;
> + if (ip_tunnel_info_af(info) == AF_INET) {
> + rt = geneve_get_v4_rt(skb, dev, &fl4, info);
> + if (IS_ERR(rt))
> + return PTR_ERR(rt);
>
> - rt = geneve_get_v4_rt(skb, dev, &fl4, info);
> - if (IS_ERR(rt))
> - return PTR_ERR(rt);
> + ip_rt_put(rt);
> + info->key.u.ipv4.src = fl4.saddr;
> +#if IS_ENABLED(CONFIG_IPV6)
> + } else if (ip_tunnel_info_af(info) == AF_INET6) {
> + dst = geneve_get_v6_dst(skb, dev, &fl6, info);
> + if (IS_ERR(dst))
> + return PTR_ERR(dst);
> +
> + dst_release(dst);
> + info->key.u.ipv6.src = fl6.saddr;
> +#endif
> + } else
> + return -EINVAL;
You need {} on this branch too, according to Documentation/CodingStyle.
MBR, Sergei
^ permalink raw reply [flat|nested] 37+ messages in thread
* [PATCH v8 3/3] geneve: add IPv6 bits to geneve_fill_metadata_dst
2015-10-26 21:01 ` [PATCH v7 3/3] geneve: add IPv6 bits to geneve_fill_metadata_dst John W. Linville
2015-10-27 12:48 ` Sergei Shtylyov
@ 2015-10-27 13:49 ` John W. Linville
2015-10-27 14:24 ` Jesse Gross
2015-10-30 3:12 ` David Miller
1 sibling, 2 replies; 37+ messages in thread
From: John W. Linville @ 2015-10-27 13:49 UTC (permalink / raw)
To: netdev
Cc: Dave Miller, Pravin B Shelar, Jesse Gross, Jiri Benc,
YOSHIFUJI Hideaki/吉藤英明,
Sergei Shtylyov, John W. Linville
Signed-off-by: John W. Linville <linville@tuxdriver.com>
---
v7 -- initial version (numbered to match earlier patches in series)
v8 -- fixup of bracing on an "else { return -EINVAL; }"
drivers/net/geneve.c | 29 ++++++++++++++++++++++-------
1 file changed, 22 insertions(+), 7 deletions(-)
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index 44e724508c55..de5c30c9f059 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -1006,16 +1006,31 @@ static int geneve_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
struct geneve_dev *geneve = netdev_priv(dev);
struct rtable *rt;
struct flowi4 fl4;
+#if IS_ENABLED(CONFIG_IPV6)
+ struct dst_entry *dst;
+ struct flowi6 fl6;
+#endif
- if (ip_tunnel_info_af(info) != AF_INET)
- return -EINVAL;
+ if (ip_tunnel_info_af(info) == AF_INET) {
+ rt = geneve_get_v4_rt(skb, dev, &fl4, info);
+ if (IS_ERR(rt))
+ return PTR_ERR(rt);
- rt = geneve_get_v4_rt(skb, dev, &fl4, info);
- if (IS_ERR(rt))
- return PTR_ERR(rt);
+ ip_rt_put(rt);
+ info->key.u.ipv4.src = fl4.saddr;
+#if IS_ENABLED(CONFIG_IPV6)
+ } else if (ip_tunnel_info_af(info) == AF_INET6) {
+ dst = geneve_get_v6_dst(skb, dev, &fl6, info);
+ if (IS_ERR(dst))
+ return PTR_ERR(dst);
+
+ dst_release(dst);
+ info->key.u.ipv6.src = fl6.saddr;
+#endif
+ } else {
+ return -EINVAL;
+ }
- ip_rt_put(rt);
- info->key.u.ipv4.src = fl4.saddr;
info->key.tp_src = udp_flow_src_port(geneve->net, skb,
1, USHRT_MAX, true);
info->key.tp_dst = geneve->dst_port;
--
2.4.3
^ permalink raw reply related [flat|nested] 37+ messages in thread
* Re: [PATCH v8 3/3] geneve: add IPv6 bits to geneve_fill_metadata_dst
2015-10-27 13:49 ` [PATCH v8 " John W. Linville
@ 2015-10-27 14:24 ` Jesse Gross
2015-10-30 3:12 ` David Miller
1 sibling, 0 replies; 37+ messages in thread
From: Jesse Gross @ 2015-10-27 14:24 UTC (permalink / raw)
To: John W. Linville
Cc: netdev, Dave Miller, Pravin B Shelar, Jiri Benc,
YOSHIFUJI Hideaki/吉藤英明,
Sergei Shtylyov
On Tue, Oct 27, 2015 at 9:49 PM, John W. Linville
<linville@tuxdriver.com> wrote:
> Signed-off-by: John W. Linville <linville@tuxdriver.com>
> ---
> v7 -- initial version (numbered to match earlier patches in series)
> v8 -- fixup of bracing on an "else { return -EINVAL; }"
Reviewed-by: Jesse Gross <jesse@nicira.com>
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH v8 3/3] geneve: add IPv6 bits to geneve_fill_metadata_dst
2015-10-27 13:49 ` [PATCH v8 " John W. Linville
2015-10-27 14:24 ` Jesse Gross
@ 2015-10-30 3:12 ` David Miller
1 sibling, 0 replies; 37+ messages in thread
From: David Miller @ 2015-10-30 3:12 UTC (permalink / raw)
To: linville
Cc: netdev, pshelar, jesse, jbenc, hideaki.yoshifuji, sergei.shtylyov
From: "John W. Linville" <linville@tuxdriver.com>
Date: Tue, 27 Oct 2015 09:49:00 -0400
> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Applied.
^ permalink raw reply [flat|nested] 37+ messages in thread
* Re: [PATCH v7 1/3] geneve: implement support for IPv6-based tunnels
2015-10-26 21:01 ` [PATCH v7 1/3] " John W. Linville
2015-10-26 21:01 ` [PATCH v7 2/3] geneve: handle ipv6 priority like ipv4 tos John W. Linville
2015-10-26 21:01 ` [PATCH v7 3/3] geneve: add IPv6 bits to geneve_fill_metadata_dst John W. Linville
@ 2015-10-30 3:11 ` David Miller
2 siblings, 0 replies; 37+ messages in thread
From: David Miller @ 2015-10-30 3:11 UTC (permalink / raw)
To: linville; +Cc: netdev, pshelar, jesse, jbenc, hideaki.yoshifuji
From: "John W. Linville" <linville@tuxdriver.com>
Date: Mon, 26 Oct 2015 17:01:44 -0400
> NOTE: Link-local IPv6 addresses for remote endpoints are not supported,
> since the driver currently has no capacity for binding a geneve
> interface to a specific link.
>
> Signed-off-by: John W. Linville <linville@tuxdriver.com>
> Reviewed-by: Jesse Gross <jesse@nicira.com>
Applied.
^ permalink raw reply [flat|nested] 37+ messages in thread