net: ipv4/ipv6: new option to accept garp/untracked na only if in-network
Subsystems (5)
cover net: ipv4/ipv6: new option to accept garp/untracked na only if in-network
The first patch adds an option to learn a neighbor from garp only if the src ip is in the same subnet of addresses configured on the interface. The option has been added to arp_accept in ipv4.
The same feature has been added to ndisc (patch 2). For ipv6, the subnet filtering knob is an extension of the accept_untracked_na option introduced in these patches:
https://lore.kernel.org/all/[email protected]/t/
https://lore.kernel.org/netdev/[email protected]/T/
The third patch contains selftests for testing the different options for accepting arp and neighbor advertisements.
Jaehee Park (3):
net: ipv4: new arp_accept option to accept garp only if in-network
net: ipv6: new accept_untracked_na option to accept na only if
in-network
selftests: net: arp_ndisc_untracked_subnets: test for arp_accept and
accept_untracked_na
Documentation/networking/ip-sysctl.rst | 48 +-- include/linux/inetdevice.h | 2 +- net/ipv4/arp.c | 24 +- net/ipv6/addrconf.c | 2 +- net/ipv6/ndisc.c | 29 +- tools/testing/selftests/net/Makefile | 1 + .../net/arp_ndisc_untracked_subnets.sh | 281 ++++++++++++++++++ 7 files changed, 360 insertions(+), 27 deletions(-) create mode 100755 tools/testing/selftests/net/arp_ndisc_untracked_subnets.sh
Signature · 1 line
2.30.2
PATCH 1/3 net: ipv4: new arp_accept option to accept garp only if in-network
In many deployments, we want the option to not learn a neighbor from garp if the src ip is not in the subnet of addresses configured on the interface. net.ipv4.arp_accept sysctl is currently used to control creation of a neigh from a received garp packet. This patch adds a new option '2' to net.ipv4.arp_accept which extends option '1' by including the subnet check.
--- Documentation/networking/ip-sysctl.rst | 4 +++- include/linux/inetdevice.h | 2 +- net/ipv4/arp.c | 24 ++++++++++++++++++++++-- 3 files changed, 26 insertions(+), 4 deletions(-)
Documentation/networking/ip-sysctl.rst +3 -1
@@ -1633,12 +1633,14 @@ arp_notify - BOOLEAN or hardware address changes. == ========================================================== -arp_accept - BOOLEAN +arp_accept - INTEGER Define behavior for gratuitous ARP frames who's IP is not already present in the ARP table: - 0 - don't create new entries in the ARP table - 1 - create new entries in the ARP table + - 2 - create new entries only if src ip is in the same subnet as + the configured address on the received interface Both replies and requests type gratuitous arp will trigger the ARP table to be updated, if this setting is on.
include/linux/inetdevice.h +1 -1
@@ -131,7 +131,7 @@ static inline void ipv4_devconf_setall(struct in_device *in_dev) IN_DEV_ORCONF((in_dev), IGNORE_ROUTES_WITH_LINKDOWN)#define IN_DEV_ARPFILTER (in_dev) IN_DEV_ORCONF((in_dev), ARPFILTER) -#define IN_DEV_ARP_ACCEPT (in_dev) IN_DEV_ORCONF((in_dev), ARP_ACCEPT) +#define IN_DEV_ARP_ACCEPT (in_dev) IN_DEV_MAXCONF((in_dev), ARP_ACCEPT)#define IN_DEV_ARP_ANNOUNCE (in_dev) IN_DEV_MAXCONF((in_dev), ARP_ANNOUNCE)#define IN_DEV_ARP_IGNORE (in_dev) IN_DEV_MAXCONF((in_dev), ARP_IGNORE)#define IN_DEV_ARP_NOTIFY (in_dev) IN_DEV_MAXCONF((in_dev), ARP_NOTIFY)
net/ipv4/arp.c +22 -2
@@ -429,6 +429,26 @@ static int arp_ignore(struct in_device *in_dev, __be32 sip, __be32 tip)return !inet_confirm_addr (net, in_dev, sip, tip, scope); } +static int arp_accept (struct in_device *in_dev,__be32 sip) +{ +struct net *net =dev_net (in_dev->dev); +int scope =RT_SCOPE_LINK ; + +switch (IN_DEV_ARP_ACCEPT (in_dev)) { +case 0 :/* don't create new entries from garp */ +return 0 ; +case 1 :/* create new entries from garp */ +return 1 ; +case 2 :/* +* create garp only if sip is in the same subnet +* as an address configured on the incoming interface +*/ +return inet_confirm_addr (net, in_dev, sip,0 , scope) ?1 :0 ; +default : +return 0 ; + } +} +static int arp_filter (__be32 sip,__be32 tip,struct net_device *dev) {struct rtable *rt;
@@ -868,12 +888,12 @@ static int arp_process(struct net *net, struct sock *sk, struct sk_buff *skb) n =__neigh_lookup (&arp_tbl, &sip, dev,0 ); addr_type =-1 ; -if (n ||IN_DEV_ARP_ACCEPT (in_dev)) { +if (n ||arp_accept (in_dev, sip)) { is_garp =arp_is_garp (net, dev, &addr_type, arp->ar_op, sip, tip, sha, tha); } -if (IN_DEV_ARP_ACCEPT (in_dev)) { +if (arp_accept (in_dev, sip)) {/* Unsolicited ARP is not accepted by default. It is possible, that this option should be enabled for some devices (strip is candidate)
Signature · 1 line
2.30.2
PATCH 2/3 net: ipv6: new accept_untracked_na option to accept na only if in-network
In reply to Jaehee ParkThis patch adds a third knob, '2', which extends the accept_untracked_na option to learn a neighbor only if the src ip is in the same subnet of addresses configured on the interfaces. This is similar to the arp_accept configuration for ipv4.
--- Documentation/networking/ip-sysctl.rst | 50 +++++++++++++++----------- net/ipv6/addrconf.c | 2 +- net/ipv6/ndisc.c | 29 ++++++++++++--- 3 files changed, 55 insertions(+), 26 deletions(-)
Documentation/networking/ip-sysctl.rst +29 -21
@@ -2482,27 +2482,36 @@ drop_unsolicited_na - BOOLEAN By default this is turned off. -accept_untracked_na - BOOLEAN - Add a new neighbour cache entry in STALE state for routers on receiving a - neighbour advertisement (either solicited or unsolicited) with target - link-layer address option specified if no neighbour entry is already - present for the advertised IPv6 address. Without this knob, NAs received - for untracked addresses (absent in neighbour cache) are silently ignored. - - This is as per router-side behaviour documented in RFC9131. - - This has lower precedence than drop_unsolicited_na. - - This will optimize the return path for the initial off-link communication - that is initiated by a directly connected host, by ensuring that - the first-hop router which turns on this setting doesn't have to - buffer the initial return packets to do neighbour-solicitation. - The prerequisite is that the host is configured to send - unsolicited neighbour advertisements on interface bringup. - This setting should be used in conjunction with the ndisc_notify setting - on the host to satisfy this prerequisite. +accept_untracked_na - INTEGER + Define behavior for accepting neighbor advertisements from devices that + are absent in the neighbor cache: - By default this is turned off. + - 0 - (default) Do not accept unsolicited and untracked neighbor + advertisements. + + - 1 - Add a new neighbor cache entry in STALE state for routers on + receiving a neighbor advertisement (either solicited or unsolicited) + with target link-layer address option specified if no neighbor entry + is already present for the advertised IPv6 address. Without this knob, + NAs received for untracked addresses (absent in neighbor cache) are + silently ignored. + + This is as per router-side behavior documented in RFC9131. + + This has lower precedence than drop_unsolicited_na. + + This will optimize the return path for the initial off-link + communication that is initiated by a directly connected host, by + ensuring that the first-hop router which turns on this setting doesn't + have to buffer the initial return packets to do neighbor-solicitation. + The prerequisite is that the host is configured to send unsolicited + neighbor advertisements on interface bringup. This setting should be + used in conjunction with the ndisc_notify setting on the host to + satisfy this prerequisite. + + - 2 - Extend option (1) to add a new neighbor cache entry only if the + source IP is in the same subnet as the configured address on the + received interface. enhanced_dad - BOOLEAN Include a nonce option in the IPv6 neighbor solicitation messages used for
@@ -2984,4 +2993,3 @@ max_dgram_qlen - INTEGER The maximum length of dgram socket receive queue Default: 10 -
net/ipv6/addrconf.c +1 -1
@@ -7042,7 +7042,7 @@ static const struct ctl_table addrconf_sysctl[] = { .data = &ipv6_devconf.accept_untracked_na, .maxlen = sizeof(int), .mode = 0644, - .proc_handler = proc_dointvec_minmax, + .proc_handler = proc_dointvec, .extra1 = (void *)SYSCTL_ZERO, .extra2 = (void *)SYSCTL_ONE, },
net/ipv6/ndisc.c +25 -4
@@ -967,6 +967,26 @@ static void ndisc_recv_ns(struct sk_buff *skb)in6_dev_put (idev); } +static int accept_untracked_na (struct net_device *dev,struct in6_addr *saddr) +{ +struct inet6_dev *idev =__in6_dev_get (dev); + +switch (idev->cnf.accept_untracked_na) { +case 0 :/* don't accept untracked (absent in neighbor cache) */ +return 0 ; +case 1 :/* create new entries if entry currently untracked */ +return 1 ; +case 2 :/* +* create new entries from untracked only if saddr is in the +* same subnet as an address configured on the incoming +* interface +*/ +return ipv6_chk_prefix (saddr, dev) ?1 :0 ; +default : +return 0 ; + } +} +static void ndisc_recv_na (struct sk_buff *skb) {struct nd_msg *msg = (struct nd_msg *)skb_transport_header (skb);
@@ -1062,10 +1082,11 @@ static void ndisc_recv_na(struct sk_buff *skb)*/ new_state = msg->icmph.icmp6_solicited ?NUD_REACHABLE :NUD_STALE ;if (!neigh && lladdr && - idev && idev->cnf.forwarding && - idev->cnf.accept_untracked_na) { - neigh =neigh_create (&nd_tbl, &msg->target, dev); - new_state =NUD_STALE ; + idev && idev->cnf.forwarding) { +if (accept_untracked_na (dev, saddr)) { + neigh =neigh_create (&nd_tbl, &msg->target, dev); + new_state =NUD_STALE ; + } }if (neigh && !IS_ERR (neigh)) {
Signature · 1 line
2.30.2
PATCH 3/3 selftests: net: arp_ndisc_untracked_subnets: test for arp_accept and accept_untracked_na
In reply to Jaehee Parkipv4 arp_accept has a new option '2' to create new neighbor entries only if the src ip is in the subnet of configured address on the interface. This selftest tests all options in arp_accept.
ipv6 has a sysctl endpoint, accept_untracked_na, that defines the behavior for accepting untracked neighbor advertisements. A new option similar to that of arp_accept for learning only from the same subnet is added to accept_untracked_na. This selftest tests this new feature.
--- tools/testing/selftests/net/Makefile | 1 + .../net/arp_ndisc_untracked_subnets.sh | 281 ++++++++++++++++++ 2 files changed, 282 insertions(+) create mode 100755 tools/testing/selftests/net/arp_ndisc_untracked_subnets.sh
tools/testing/selftests/net/Makefile +1 -0
@@ -38,6 +38,7 @@ TEST_PROGS += srv6_end_dt6_l3vpn_test.sh TEST_PROGS += vrf_strict_mode_test.sh TEST_PROGS += arp_ndisc_evict_nocarrier.sh TEST_PROGS += ndisc_unsolicited_na_test.sh +TEST_PROGS += arp_ndisc_untracked_subnets.sh TEST_PROGS += stress_reuseport_listen.sh TEST_PROGS_EXTENDED := in_netns.sh setup_loopback.sh setup_veth.sh TEST_PROGS_EXTENDED += toeplitz_client.sh toeplitz.sh
tools/testing/selftests/net/arp_ndisc_untracked_subnets.sh new file +281 -0
281 lines not on this page · raw message
Signature · 1 line
2.30.2
Quoted 31 lines · Jaehee Park
The first patch adds an option to learn a neighbor from garp only if the src ip is in the same subnet of addresses configured on the interface. The option has been added to
arp_acceptin ipv4.The same feature has been added to ndisc (patch 2). For ipv6, the subnet filtering knob is an extension of the
accept_untracked_naoption introduced in these patches:
https://lore.kernel.org/all/[email protected]/t/
https://lore.kernel.org/netdev/[email protected]/T/The third patch contains selftests for testing the different options for accepting arp and neighbor advertisements.
Jaehee Park (3):
net: ipv4: newarp_acceptoption to accept garp only if in-network
net: ipv6: newaccept_untracked_naoption to accept na only if
in-network
selftests: net:arp_ndisc_untracked_subnets: test forarp_acceptand
accept_untracked_naDocumentation/networking/ip-sysctl.rst | 48 +-- include/linux/inetdevice.h | 2 +- net/ipv4/arp.c | 24 +- net/ipv6/addrconf.c | 2 +- net/ipv6/ndisc.c | 29 +- tools/testing/selftests/net/Makefile | 1 + .../net/arp_ndisc_untracked_subnets.sh | 281 ++++++++++++++++++ 7 files changed, 360 insertions(+), 27 deletions(-) create mode 100755 tools/testing/selftests/net/arp_ndisc_untracked_subnets.sh
Quoted 3 lines
--
2.30.2
I forgot a few cleanups. I will post a v2 soon.
Sorry about that!
Thanks,
Jaehee