== Description of problem ==

When adding an MPLS flow, ovs-ofctl ignores the mpls_ttl field.
For example:                                            
# ovs-ofctl add-flow br0 "in_port=veth10,mpls,mpls_label=20,mpls_bos=1,mpls_ttl=64 actions=pop_mpls:0x0800,output:veth12"
# ovs-ofctl dump-flows br0                                                         
 cookie=0x0, duration=9.743s, table=0, n_packets=0, n_bytes=0, mpls,in_port=veth10,mpls_label=20,mpls_bos=1 actions=pop_mpls:0x0800,output:veth12
            
The mpls_ttl field is absent from the dump-flows command output.

== Version-Release number of selected component (if applicable) ==

Problem originally found on RHEL 8 and reproduced with the latest upstream git tree, currently at 89b522aee379 ("ovsdb-idl: Add function to reset min_index.").
# ovs-ofctl --version
ovs-ofctl (Open vSwitch) 2.13.90
OpenFlow versions 0x1:0x6

== How reproducible ==

Always.

== Steps to Reproduce ==

Here's a more complete scenario:

Create two network namespaces, ns0 and ns2.
# ip netns add ns0
# ip netns add ns2
# ip -n ns0 link set dev lo up
# ip -n ns2 link set dev lo up

Create a veth pair in each namespace, connected with the current namespace.
# ip link add name veth01 netns ns0 type veth peer name veth10
# ip link add name veth12 type veth peer name veth21 netns ns2
# ip -n ns0 link set dev veth01 up
# ip -n ns2 link set dev veth21 up
# ip link set dev veth10 up
# ip link set dev veth12 up

The veth of ns0 has IP 192.0.2.10, ns2 has IP 192.0.2.12
# ip -n ns0 address add 192.0.2.10/24 dev veth01
# ip -n ns2 address add 192.0.2.12/24 dev veth21

Use tc to add an MPLS header to packets leaving ns0 or ns2. The TTL is set to 20 in ns0 and 30 in ns2.
# tc -n ns0 qdisc replace dev veth01 root prio
# tc -n ns2 qdisc replace dev veth21 root prio
# tc -n ns0 filter add dev veth01 proto ipv4 flower dst_ip 192.0.2.12/32 action mpls push label 20 ttl 20
# tc -n ns2 filter add dev veth21 proto ipv4 flower dst_ip 192.0.2.10/32 action mpls push label 30 ttl 30

Use OVS to bridge the veths of ns0 and ns2 in the current namespace.
# ovs-vsctl del-br br0
# ovs-vsctl add-br br0
# ovs-vsctl add-port br0 veth10
# ovs-vsctl add-port br0 veth12

Add flow for ARP.
# ovs-ofctl del-flows br0
# ovs-ofctl add-flow br0 "table=0,priority=10 arp action=normal"

Drop the MPLS header previously added by tc, but match of the wrong TTL (64 instead of 20 and 30).
# ovs-ofctl add-flow br0 "in_port=veth10,mpls,mpls_label=20,mpls_bos=1,mpls_ttl=64 actions=pop_mpls:0x0800,output:veth12"
# ovs-ofctl add-flow br0 "in_port=veth12,mpls,mpls_label=30,mpls_bos=1,mpls_ttl=64 actions=pop_mpls:0x0800,output:veth10"

== Actual results ==
The MPLS flows are matched, even though the TTL is different.
Therefore ns0 can ping ns2:
# ip netns exec ns0 ping -c 1 192.0.2.12
PING 192.0.2.12 (192.0.2.12) 56(84) bytes of data.
64 bytes from 192.0.2.12: icmp_seq=1 ttl=64 time=2.13 ms

--- 192.0.2.12 ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
rtt min/avg/max/mdev = 2.125/2.125/2.125/0.000 ms

== Expected results ==
The MPLS flows don't match, because the TTL is different, and ping fails.