Can't talk to most of that, but baby-jumbo is supported by the Plusnet setup, so you can configure with 1500 MTU if you set that up. I don't know that it did anything significant to throughput for me, but felt better having the local LAN MTU matching the uplink.
In practical experience terms, when I was running a couple of HE tunnels - one on plusnet and one over Three 5G, I found clamping was more reliable than expecting icmp6 to be correctly returned. ( And I'd configured both to have the same MTU )
Not sure how the pfSense setup would enable baby jumbo, but for Debian, it is a pre-up command on the underlying ethernet port, and then declaring 1500 for mtu and mru in the dsl-providers file. ( I'm still old-school /etc/network/interfaces based but I imagine netplan has similar capability )
##
# The PlusNET PPPOE connection itself which uses enp7s0 physical interface
#
auto plusnet-provider
iface plusnet-provider inet ppp
provider plusnet-provider
pre-up /usr/sbin/ip link set enp7s0 mtu 1508 up || true
pre-up /usr/sbin/sysctl -w net.ipv6.conf.ppp0.accept_ra=2 || true
pre-up /usr/sbin/sysctl -w net.ipv6.conf.ppp0.forwarding=1 || true
post-up /root/bin/on_reboot_policy_routing_setup || true
post-up /usr/sbin/sysctl -w net.ipv6.conf.ppp0.accept_ra=2 || true
post-up /usr/sbin/sysctl -w net.ipv6.conf.ppp0.forwarding=1 || true
# post-up /usr/sbin/tc qdisc del dev ppp0 root || true
# post-up /usr/sbin/tc qdisc add dev ppp0 root sfq perturb 0 depth 127 limit 1023 headdrop || true
post-down /usr/sbin/ip link set enp7s0 down || true
Cheers,
Mark
