Right now source address selection is all screwed up for a number of
use cases. It does not properly take into account VRF centric addresses
or even valid routes for a VRF. Fix by implementating a get_saddr method
similar to what was done for IPv4. The get_saddr6 method does a full
lookup which means pulling a route from the VRF FIB table. Lookup
failures (eg., unreachable) then cause the source address selection
to fail which gets propagated back to the caller.
Since ipv6_dev_get_saddr is already exported move ip6_route_get_saddr to
the header as an inline since it only checks for a preferred source
address prior to calling ipv6_dev_get_saddr.
Signed-off-by: David Ahern <d...@cumulusnetworks.com>
---
drivers/net/vrf.c | 86 +++++++++++++++++++++++++++++++++++++++----------
include/net/ip6_route.h | 21 ++++++++++--
include/net/l3mdev.h | 11 +++++++
net/ipv6/ip6_output.c | 12 +++++--
net/ipv6/route.c | 17 ----------
net/l3mdev/l3mdev.c | 25 ++++++++++++++
6 files changed, 133 insertions(+), 39 deletions(-)
diff --git a/drivers/net/vrf.c b/drivers/net/vrf.c
index fb2d0b2052ea..d83d903dc674 100644
--- a/drivers/net/vrf.c
+++ b/drivers/net/vrf.c
@@ -774,20 +774,11 @@ static bool ipv6_ndisc_frame(const struct sk_buff *skb)
return rc;
}
-static void vrf_ip6_input_dst(struct sk_buff *skb, struct net_device *vrf_dev,
- int ifindex)
+static struct rt6_info *vrf_ip6_route_lookup(struct net_device *dev,
+ struct flowi6 *fl6, int ifindex)
{
- const struct ipv6hdr *iph = ipv6_hdr(skb);
- struct flowi6 fl6 = {
- .daddr = iph->daddr,
- .saddr = iph->saddr,
- .flowlabel = ip6_flowinfo(iph),
- .flowi6_mark = skb->mark,
- .flowi6_proto = iph->nexthdr,
- .flowi6_iif = ifindex,
- };
- struct net_vrf *vrf = netdev_priv(vrf_dev);
- struct net *net = dev_net(vrf_dev);
+ struct net_vrf *vrf = netdev_priv(dev);
+ struct net *net = dev_net(dev);
struct fib6_table *table = NULL;
struct rt6_info *rt6;
@@ -801,14 +792,36 @@ static void vrf_ip6_input_dst(struct sk_buff *skb, struct
net_device *vrf_dev,
rcu_read_unlock();
if (!table)
- return;
+ return NULL;
- rt6 = ip6_pol_route(net, table, ifindex, &fl6,
- RT6_LOOKUP_F_HAS_SADDR | RT6_LOOKUP_F_IFACE);
+ return ip6_pol_route(net, table, ifindex, fl6,
+ RT6_LOOKUP_F_HAS_SADDR | RT6_LOOKUP_F_IFACE);
+}
- if (unlikely(&rt6->dst == &net->ipv6.ip6_null_entry->dst))
+static void vrf_ip6_input_dst(struct sk_buff *skb, struct net_device *dev,
+ int ifindex)
+{
+ const struct ipv6hdr *iph = ipv6_hdr(skb);
+ struct flowi6 fl6 = {
+ .daddr = iph->daddr,
+ .saddr = iph->saddr,
+ .flowlabel = ip6_flowinfo(iph),
+ .flowi6_mark = skb->mark,
+ .flowi6_proto = iph->nexthdr,
+ .flowi6_iif = ifindex,
+ };
+ struct net *net = dev_net(dev);
+ struct rt6_info *rt6;
+
+ rt6 = vrf_ip6_route_lookup(dev, &fl6, ifindex);
+ if (unlikely(!rt6))
return;
+ if (unlikely(&rt6->dst == &net->ipv6.ip6_null_entry->dst)) {
+ dst_release(&rt6->dst);
+ return;
+ }
+
skb_dst_set(skb, &rt6->dst);
}
@@ -836,6 +849,44 @@ static struct sk_buff *vrf_ip6_rcv(struct net_device
*vrf_dev,
return skb;
}
+/* called under rcu_read_lock */
+static int vrf_get_saddr6(struct net_device *dev, const struct sock *sk,
+ struct flowi6 *fl6)
+{
+ struct net *net = dev_net(dev);
+ struct dst_entry *dst;
+ struct rt6_info *rt;
+ int err;
+
+ if (rt6_need_strict(&fl6->daddr)) {
+ rt = vrf_ip6_route_lookup(dev, fl6, fl6->flowi6_oif);
+ if (unlikely(!rt))
+ return 0;
+
+ dst = &rt->dst;
+ } else {
+ __u8 flags = fl6->flowi6_flags;
+
+ fl6->flowi6_flags |= FLOWI_FLAG_L3MDEV_SRC;
+ fl6->flowi6_flags |= FLOWI_FLAG_SKIP_NH_OIF;
+
+ dst = ip6_route_output(net, sk, fl6);
+ rt = (struct rt6_info *)dst;
+
+ fl6->flowi6_flags = flags;
+ }
+
+ err = dst->error;
+ if (!err) {
+ err = ip6_route_get_saddr(net, rt, &fl6->daddr,
+ sk ? inet6_sk(sk)->srcprefs : 0,
+ &fl6->saddr);
+ }
+
+ dst_release(dst);
+
+ return err;
+}
#else
static struct sk_buff *vrf_ip6_rcv(struct net_device *vrf_dev,
struct sk_buff *skb)
@@ -947,6 +998,7 @@ static const struct l3mdev_ops vrf_l3mdev_ops = {
.l3mdev_l3_rcv = vrf_l3_rcv,
#if IS_ENABLED(CONFIG_IPV6)
.l3mdev_get_rt6_dst = vrf_get_rt6_dst,
+ .l3mdev_get_saddr6 = vrf_get_saddr6,
#endif
};
diff --git a/include/net/ip6_route.h b/include/net/ip6_route.h
index f73a65e97597..6886deb45679 100644
--- a/include/net/ip6_route.h
+++ b/include/net/ip6_route.h
@@ -18,6 +18,7 @@ struct route_info {
__u8 prefix[0]; /* 0,8 or 16 */
};
+#include <net/addrconf.h>
#include <net/flow.h>
#include <net/ip6_fib.h>
#include <net/sock.h>
@@ -89,9 +90,23 @@ int ip6_route_add(struct fib6_config *cfg);
int ip6_ins_rt(struct rt6_info *);
int ip6_del_rt(struct rt6_info *);
-int ip6_route_get_saddr(struct net *net, struct rt6_info *rt,
- const struct in6_addr *daddr, unsigned int prefs,
- struct in6_addr *saddr);
+static inline int ip6_route_get_saddr(struct net *net, struct rt6_info *rt,
+ const struct in6_addr *daddr,
+ unsigned int prefs,
+ struct in6_addr *saddr)
+{
+ struct inet6_dev *idev =
+ rt ? ip6_dst_idev((struct dst_entry *)rt) : NULL;
+ int err = 0;
+
+ if (rt && rt->rt6i_prefsrc.plen)
+ *saddr = rt->rt6i_prefsrc.addr;
+ else
+ err = ipv6_dev_get_saddr(net, idev ? idev->dev : NULL,
+ daddr, prefs, saddr);
+
+ return err;
+}
struct rt6_info *rt6_lookup(struct net *net, const struct in6_addr *daddr,
const struct in6_addr *saddr, int oif, int flags);
diff --git a/include/net/l3mdev.h b/include/net/l3mdev.h
index d575185600a5..6ba0a206db45 100644
--- a/include/net/l3mdev.h
+++ b/include/net/l3mdev.h
@@ -38,6 +38,9 @@ struct l3mdev_ops {
struct dst_entry * (*l3mdev_get_rt6_dst)(const struct net_device *dev,
struct flowi6 *fl6,
int flags);
+ int (*l3mdev_get_saddr6)(struct net_device *dev,
+ const struct sock *sk,
+ struct flowi6 *fl6);
};
#ifdef CONFIG_NET_L3_MASTER_DEV
@@ -137,6 +140,8 @@ int l3mdev_get_saddr(struct net *net, int ifindex, struct
flowi4 *fl4);
struct dst_entry *l3mdev_get_rt6_dst(struct net *net, struct flowi6 *fl6,
int flags);
+int l3mdev_get_saddr6(struct net *net, const struct sock *sk,
+ struct flowi6 *fl6);
static inline
struct sk_buff *l3mdev_l3_rcv(struct sk_buff *skb, u16 proto)
@@ -229,6 +234,12 @@ struct dst_entry *l3mdev_get_rt6_dst(struct net *net,
struct flowi6 *fl6,
return NULL;
}
+static inline int l3mdev_get_saddr6(struct net *net, const struct sock *sk,
+ struct flowi6 *fl6)
+{
+ return 0;
+}
+
static inline
struct sk_buff *l3mdev_ip_rcv(struct sk_buff *skb)
{
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index cbf127ae7c67..cfd01782a621 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -910,6 +910,13 @@ static int ip6_dst_lookup_tail(struct net *net, const
struct sock *sk,
int err;
int flags = 0;
+ if (ipv6_addr_any(&fl6->saddr) && fl6->flowi6_oif &&
+ (!*dst || !(*dst)->error)) {
+ err = l3mdev_get_saddr6(net, sk, fl6);
+ if (err)
+ goto out_err;
+ }
+
/* The correct way to handle this would be to do
* ip6_route_get_saddr, and then ip6_route_output; however,
* the route-specific preferred source forces the
@@ -999,10 +1006,11 @@ static int ip6_dst_lookup_tail(struct net *net, const
struct sock *sk,
return 0;
out_err_release:
- if (err == -ENETUNREACH)
- IP6_INC_STATS(net, NULL, IPSTATS_MIB_OUTNOROUTES);
dst_release(*dst);
*dst = NULL;
+out_err:
+ if (err == -ENETUNREACH)
+ IP6_INC_STATS(net, NULL, IPSTATS_MIB_OUTNOROUTES);
return err;
}
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index a87e66d2284f..67ec5594be9c 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -2584,23 +2584,6 @@ struct rt6_info *addrconf_dst_alloc(struct inet6_dev
*idev,
return rt;
}
-int ip6_route_get_saddr(struct net *net,
- struct rt6_info *rt,
- const struct in6_addr *daddr,
- unsigned int prefs,
- struct in6_addr *saddr)
-{
- struct inet6_dev *idev =
- rt ? ip6_dst_idev((struct dst_entry *)rt) : NULL;
- int err = 0;
- if (rt && rt->rt6i_prefsrc.plen)
- *saddr = rt->rt6i_prefsrc.addr;
- else
- err = ipv6_dev_get_saddr(net, idev ? idev->dev : NULL,
- daddr, prefs, saddr);
- return err;
-}
-
/* remove deleted ip from prefsrc entries */
struct arg_dev_net_ip {
struct net_device *dev;
diff --git a/net/l3mdev/l3mdev.c b/net/l3mdev/l3mdev.c
index dceac272b8c4..3e08d3e27a8a 100644
--- a/net/l3mdev/l3mdev.c
+++ b/net/l3mdev/l3mdev.c
@@ -164,3 +164,28 @@ int l3mdev_get_saddr(struct net *net, int ifindex, struct
flowi4 *fl4)
return rc;
}
EXPORT_SYMBOL_GPL(l3mdev_get_saddr);
+
+int l3mdev_get_saddr6(struct net *net, const struct sock *sk,
+ struct flowi6 *fl6)
+{
+ struct net_device *dev;
+ int rc = 0;
+
+ if (fl6->flowi6_oif) {
+ rcu_read_lock();
+
+ dev = dev_get_by_index_rcu(net, fl6->flowi6_oif);
+ if (dev && netif_is_l3_slave(dev))
+ dev = netdev_master_upper_dev_get_rcu(dev);
+
+ if (dev && netif_is_l3_master(dev) &&
+ dev->l3mdev_ops->l3mdev_get_saddr6) {
+ rc = dev->l3mdev_ops->l3mdev_get_saddr6(dev, sk, fl6);
+ }
+
+ rcu_read_unlock();
+ }
+
+ return rc;
+}
+EXPORT_SYMBOL_GPL(l3mdev_get_saddr6);
--
2.1.4
