[PATCH] net: arp, ipv6: handle special case of tap device

From: Vladis Dronov
Date: Wed Sep 26 2018 - 05:30:29 EST


dev->type can be set to any value for tun/tap devices by TUNSETLINK ioctl.
Nevertheless, all other dev's fields related to the link level like addr_len
and broadcast remain the same as configured previously for ARPHRD_ETHER type,
making dev inconsistent. This can lead to read from uninitialized memory.
Fix this by checking for the case of tun/tap device and act according to the
real link level type of dev while mapping a multicast IP onto multicast MAC

Reported-by: syzbot+d3402c47f680ff24b29c@xxxxxxxxxxxxxxxxxxxxxxxxx
Signed-off-by: Vladis Dronov <vdronov@xxxxxxxxxx>
---
net/ipv4/arp.c | 9 ++++++++-
net/ipv6/ndisc.c | 9 ++++++++-
2 files changed, 16 insertions(+), 2 deletions(-)

diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c
index e90c89ef8c08..9ce472cf98a3 100644
--- a/net/ipv4/arp.c
+++ b/net/ipv4/arp.c
@@ -187,7 +187,14 @@ EXPORT_SYMBOL(arp_tbl);

int arp_mc_map(__be32 addr, u8 *haddr, struct net_device *dev, int dir)
{
- switch (dev->type) {
+ unsigned short type = dev->type;
+
+#if IS_ENABLED(CONFIG_TAP)
+ if (dev->rtnl_link_ops && !strcmp(dev->rtnl_link_ops->kind, "tun"))
+ type = ARPHRD_ETHER;
+#endif /* CONFIG_TAP */
+
+ switch (type) {
case ARPHRD_ETHER:
case ARPHRD_FDDI:
case ARPHRD_IEEE802:
diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c
index 0ec273997d1d..9371ff4454f7 100644
--- a/net/ipv6/ndisc.c
+++ b/net/ipv6/ndisc.c
@@ -283,7 +283,14 @@ struct ndisc_options *ndisc_parse_options(const struct net_device *dev,

int ndisc_mc_map(const struct in6_addr *addr, char *buf, struct net_device *dev, int dir)
{
- switch (dev->type) {
+ unsigned short type = dev->type;
+
+#if IS_ENABLED(CONFIG_TAP)
+ if (dev->rtnl_link_ops && !strcmp(dev->rtnl_link_ops->kind, "tun"))
+ type = ARPHRD_ETHER;
+#endif /* CONFIG_TAP */
+
+ switch (type) {
case ARPHRD_ETHER:
case ARPHRD_IEEE802: /* Not sure. Check it later. --ANK */
case ARPHRD_FDDI:
--
2.19.0