Two patches here, first fixes the issue where tunnel core doesn't actually extract DF bit from the outer IP header, even though both OVS and TC flower allow matching on it. More details in the commit message.
The second is a selftest for openvswitch that reproduces the issue, but also just adds some basic coverage for the tunnel metadata extraction and related openvswitch uAPI.
Version 2: * Added missing tun_dst NULL check. * Added Reviewed-by from Aaron for the selftest.
Version 1: https://lore.kernel.org/netdev/20250905133105.3940420-1-i.maximets@ovn.org/
Ilya Maximets (2): net: dst_metadata: fix IP_DF bit not extracted from tunnel headers selftests: openvswitch: add a simple test for tunnel metadata
include/net/dst_metadata.h | 11 ++- .../selftests/net/openvswitch/openvswitch.sh | 88 +++++++++++++++++-- 2 files changed, 90 insertions(+), 9 deletions(-)
Both OVS and TC flower allow extracting and matching on the DF bit of the outer IP header via OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT in the OVS_KEY_ATTR_TUNNEL and TCA_FLOWER_KEY_FLAGS_TUNNEL_DONT_FRAGMENT in the TCA_FLOWER_KEY_ENC_FLAGS respectively. Flow dissector extracts this information as FLOW_DIS_F_TUNNEL_DONT_FRAGMENT from the tunnel info key.
However, the IP_TUNNEL_DONT_FRAGMENT_BIT in the tunnel key is never actually set, because the tunneling code doesn't actually extract it from the IP header. OAM and CRIT_OPT are extracted by the tunnel implementation code, same code also sets the KEY flag, if present. UDP tunnel core takes care of setting the CSUM flag if the checksum is present in the UDP header, but the DONT_FRAGMENT is not handled at any layer.
Fix that by checking the bit and setting the corresponding flag while populating the tunnel info in the IP layer where it belongs.
Not using __assign_bit as we don't really need to clear the bit in a just initialized field. It also doesn't seem like using __assign_bit will make the code look better.
Clearly, users didn't rely on this functionality for anything very important until now. The reason why this doesn't break OVS logic is that it only matches on what kernel previously parsed out and if kernel consistently reports this bit as zero, OVS will only match on it to be zero, which sort of works. But it is still a bug that the uAPI reports and allows matching on the field that is not actually checked in the packet. And this is causing misleading -df reporting in OVS datapath flows, while the tunnel traffic actually has the bit set in most cases.
This may also cause issues if a hardware properly implements support for tunnel flag matching as it will disagree with the implementation in a software path of TC flower.
Fixes: 7d5437c709de ("openvswitch: Add tunneling interface.") Fixes: 1d17568e74de ("net/sched: cls_flower: add support for matching tunnel control flags") Signed-off-by: Ilya Maximets i.maximets@ovn.org --- include/net/dst_metadata.h | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/include/net/dst_metadata.h b/include/net/dst_metadata.h index 4160731dcb6e..1fc2fb03ce3f 100644 --- a/include/net/dst_metadata.h +++ b/include/net/dst_metadata.h @@ -3,6 +3,7 @@ #define __NET_DST_METADATA_H 1
#include <linux/skbuff.h> +#include <net/ip.h> #include <net/ip_tunnels.h> #include <net/macsec.h> #include <net/dst.h> @@ -220,9 +221,15 @@ static inline struct metadata_dst *ip_tun_rx_dst(struct sk_buff *skb, int md_size) { const struct iphdr *iph = ip_hdr(skb); + struct metadata_dst *tun_dst; + + tun_dst = __ip_tun_set_dst(iph->saddr, iph->daddr, iph->tos, iph->ttl, + 0, flags, tunnel_id, md_size);
- return __ip_tun_set_dst(iph->saddr, iph->daddr, iph->tos, iph->ttl, - 0, flags, tunnel_id, md_size); + if (tun_dst && (iph->frag_off & htons(IP_DF))) + __set_bit(IP_TUNNEL_DONT_FRAGMENT_BIT, + tun_dst->u.tun_info.key.tun_flags); + return tun_dst; }
static inline struct metadata_dst *__ipv6_tun_set_dst(const struct in6_addr *saddr,
On Tue, Sep 09, 2025 at 06:54:15PM +0200, Ilya Maximets wrote:
Both OVS and TC flower allow extracting and matching on the DF bit of the outer IP header via OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT in the OVS_KEY_ATTR_TUNNEL and TCA_FLOWER_KEY_FLAGS_TUNNEL_DONT_FRAGMENT in the TCA_FLOWER_KEY_ENC_FLAGS respectively. Flow dissector extracts this information as FLOW_DIS_F_TUNNEL_DONT_FRAGMENT from the tunnel info key.
However, the IP_TUNNEL_DONT_FRAGMENT_BIT in the tunnel key is never actually set, because the tunneling code doesn't actually extract it from the IP header. OAM and CRIT_OPT are extracted by the tunnel implementation code, same code also sets the KEY flag, if present. UDP tunnel core takes care of setting the CSUM flag if the checksum is present in the UDP header, but the DONT_FRAGMENT is not handled at any layer.
Fix that by checking the bit and setting the corresponding flag while populating the tunnel info in the IP layer where it belongs.
Not using __assign_bit as we don't really need to clear the bit in a just initialized field. It also doesn't seem like using __assign_bit will make the code look better.
Clearly, users didn't rely on this functionality for anything very important until now. The reason why this doesn't break OVS logic is that it only matches on what kernel previously parsed out and if kernel consistently reports this bit as zero, OVS will only match on it to be zero, which sort of works. But it is still a bug that the uAPI reports and allows matching on the field that is not actually checked in the packet. And this is causing misleading -df reporting in OVS datapath flows, while the tunnel traffic actually has the bit set in most cases.
This may also cause issues if a hardware properly implements support for tunnel flag matching as it will disagree with the implementation in a software path of TC flower.
Fixes: 7d5437c709de ("openvswitch: Add tunneling interface.") Fixes: 1d17568e74de ("net/sched: cls_flower: add support for matching tunnel control flags") Signed-off-by: Ilya Maximets i.maximets@ovn.org
Reviewed-by: Ido Schimmel idosch@nvidia.com
This test ensures that upon receiving decapsulated packets from a tunnel interface in openvswitch, the tunnel metadata fields are properly populated. This partially covers interoperability of the kernel tunnel ports and openvswitch tunnels (LWT) and parsing and formatting of the tunnel metadata fields of the openvswitch netlink uAPI. Doing so, this test also ensures that fields and flags are properly extracted during decapsulation by the tunnel core code, serving as a regression test for the previously fixed issue with the DF bit not being extracted from the outer IP header.
The ovs-dpctl.py script already supports all that is necessary for the tunnel ports for this test, so we only need to adjust the ovs_add_if() function to pass the '-t' port type argument in order to be able to create tunnel ports in the openvswitch datapath.
Reviewed-by: Aaron Conole aconole@redhat.com Signed-off-by: Ilya Maximets i.maximets@ovn.org --- .../selftests/net/openvswitch/openvswitch.sh | 88 +++++++++++++++++-- 1 file changed, 81 insertions(+), 7 deletions(-)
diff --git a/tools/testing/selftests/net/openvswitch/openvswitch.sh b/tools/testing/selftests/net/openvswitch/openvswitch.sh index 3c8d3455d8e7..b327d3061ed5 100755 --- a/tools/testing/selftests/net/openvswitch/openvswitch.sh +++ b/tools/testing/selftests/net/openvswitch/openvswitch.sh @@ -25,6 +25,7 @@ tests=" nat_related_v4 ip4-nat-related: ICMP related matches work with SNAT netlink_checks ovsnl: validate netlink attrs and settings upcall_interfaces ovs: test the upcall interfaces + tunnel_metadata ovs: test extraction of tunnel metadata drop_reason drop: test drop reasons are emitted psample psample: Sampling packets with psample"
@@ -113,13 +114,13 @@ ovs_add_dp () { }
ovs_add_if () { - info "Adding IF to DP: br:$2 if:$3" - if [ "$4" != "-u" ]; then - ovs_sbx "$1" python3 $ovs_base/ovs-dpctl.py add-if "$2" "$3" \ - || return 1 + info "Adding IF to DP: br:$3 if:$4 ($2)" + if [ "$5" != "-u" ]; then + ovs_sbx "$1" python3 $ovs_base/ovs-dpctl.py add-if \ + -t "$2" "$3" "$4" || return 1 else python3 $ovs_base/ovs-dpctl.py add-if \ - -u "$2" "$3" >$ovs_dir/$3.out 2>$ovs_dir/$3.err & + -u -t "$2" "$3" "$4" >$ovs_dir/$4.out 2>$ovs_dir/$4.err & pid=$! on_exit "ovs_sbx $1 kill -TERM $pid 2>/dev/null" fi @@ -166,9 +167,9 @@ ovs_add_netns_and_veths () { fi
if [ "$7" != "-u" ]; then - ovs_add_if "$1" "$2" "$4" || return 1 + ovs_add_if "$1" "netdev" "$2" "$4" || return 1 else - ovs_add_if "$1" "$2" "$4" -u || return 1 + ovs_add_if "$1" "netdev" "$2" "$4" -u || return 1 fi
if [ $TRACING -eq 1 ]; then @@ -756,6 +757,79 @@ test_upcall_interfaces() { return 0 }
+ovs_add_kernel_tunnel() { + local sbxname=$1; shift + local ns=$1; shift + local tnl_type=$1; shift + local name=$1; shift + local addr=$1; shift + + info "setting up kernel ${tnl_type} tunnel ${name}" + ovs_sbx "${sbxname}" ip -netns ${ns} link add dev ${name} type ${tnl_type} $* || return 1 + on_exit "ovs_sbx ${sbxname} ip -netns ${ns} link del ${name} >/dev/null 2>&1" + ovs_sbx "${sbxname}" ip -netns ${ns} addr add dev ${name} ${addr} || return 1 + ovs_sbx "${sbxname}" ip -netns ${ns} link set dev ${name} mtu 1450 up || return 1 +} + +test_tunnel_metadata() { + which arping >/dev/null 2>&1 || return $ksft_skip + + sbxname="test_tunnel_metadata" + sbx_add "${sbxname}" || return 1 + + info "setting up new DP" + ovs_add_dp "${sbxname}" tdp0 -V 2:1 || return 1 + + ovs_add_netns_and_veths "${sbxname}" tdp0 tns left0 l0 \ + 172.31.110.1/24 || return 1 + + info "removing veth interface from openvswitch and setting IP" + ovs_del_if "${sbxname}" tdp0 left0 || return 1 + ovs_sbx "${sbxname}" ip addr add 172.31.110.2/24 dev left0 || return 1 + ovs_sbx "${sbxname}" ip link set left0 up || return 1 + + info "setting up tunnel port in openvswitch" + ovs_add_if "${sbxname}" "vxlan" tdp0 ovs-vxlan0 -u || return 1 + on_exit "ovs_sbx ${sbxname} ip link del ovs-vxlan0" + ovs_wait ip link show ovs-vxlan0 &>/dev/null || return 1 + ovs_sbx "${sbxname}" ip link set ovs-vxlan0 up || return 1 + + configs=$(echo ' + 1 172.31.221.1/24 1155332 32 set udpcsum flags(df|csum) + 2 172.31.222.1/24 1234567 45 set noudpcsum flags(df) + 3 172.31.223.1/24 1020304 23 unset udpcsum flags(csum) + 4 172.31.224.1/24 1357986 15 unset noudpcsum' | sed '/^$/d') + + while read -r i addr id ttl df csum flags; do + ovs_add_kernel_tunnel "${sbxname}" tns vxlan vxlan${i} ${addr} \ + remote 172.31.110.2 id ${id} dstport 4789 \ + ttl ${ttl} df ${df} ${csum} || return 1 + done <<< "${configs}" + + ovs_wait grep -q 'listening on upcall packet handler' \ + ${ovs_dir}/ovs-vxlan0.out || return 1 + + info "sending arping" + for i in 1 2 3 4; do + ovs_sbx "${sbxname}" ip netns exec tns \ + arping -I vxlan${i} 172.31.22${i}.2 -c 1 \ + >${ovs_dir}/arping.stdout 2>${ovs_dir}/arping.stderr + done + + info "checking that received decapsulated packets carry correct metadata" + while read -r i addr id ttl df csum flags; do + arp_hdr="arp\(sip=172.31.22${i}.1,tip=172.31.22${i}.2,op=1,sha=" + addrs="src=172.31.110.1,dst=172.31.110.2" + ports="tp_src=[0-9]*,tp_dst=4789" + tnl_md="tunnel\(tun_id=${id},${addrs},ttl=${ttl},${ports},${flags}\)" + + ovs_sbx "${sbxname}" grep -qE "MISS upcall.*${tnl_md}.*${arp_hdr}" \ + ${ovs_dir}/ovs-vxlan0.out || return 1 + done <<< "${configs}" + + return 0 +} + run_test() { ( tname="$1"
linux-kselftest-mirror@lists.linaro.org