DHCPOFFER not crossing virtual bridge on Linux?

I have a physical host machine running Ubuntu 20.4LTS (fully patched) attached to my network, which retrieves it’s IP address via DHCP. This works.

I have a VirtualBox VM on that host attached to a bridged interface (macvlan on the physical NIC) that is trying to obtain an IP address from the same DHCP server.

Using tcpdump -i ethXXX port 67 or port 68 -e -n -vv on the Host, I see multiple repeats of this message going out:

15:04:52.845988 66:66:66:66:66:12 > ff:ff:ff:ff:ff:ff, ethertype IPv4 (0x0800), length 342: (tos 0x10, ttl 128, id 0, offset 0, flags (none), proto UDP (17), length 328)
    0.0.0.0.68 > 255.255.255.255.67: (udp sum ok) BOOTP/DHCP, Request from 66:66:66:66:66:12, length 300, xid 0x84ce756d, secs 6, Flags (none) (0x0000)
          Client-Ethernet-Address 66:66:66:66:66:12
          Vendor-rfc1048 Extensions
            Magic Cookie 0x63825363
            DHCP-Message Option 53, length 1: Discover
            Hostname Option 12, length 6: "nodename"
            Parameter-Request Option 55, length 17: 
              Subnet-Mask, BR, Time-Zone, Default-Gateway
              Domain-Name, Domain-Name-Server, Option 119, Hostname
              Netbios-Name-Server, Netbios-Scope, MTU, Classless-Static-Route
              NTP, Classless-Static-Route-Microsoft, Static-Route, Option 252
              RP

and on the DHCP server, I see it come in, and a reply sent out:

15:05:57.193384 66:66:66:66:66:12 > ff:ff:ff:ff:ff:ff, ethertype IPv4 (0x0800), length 342: (tos 0x10, ttl 128, id 0, offset 0, flags (none), proto UDP (17), length 328)
    0.0.0.0.68 > 255.255.255.255.67: (udp sum ok) BOOTP/DHCP, Request from 66:66:66:66:66:12, length 300, xid 0xcede823f, secs 22, Flags (none) (0x0000)
          Client-Ethernet-Address 66:66:66:66:66:12
          Vendor-rfc1048 Extensions
            Magic Cookie 0x63825363
            DHCP-Message Option 53, length 1: Discover
            Hostname Option 12, length 6: "nodename"
            Parameter-Request Option 55, length 17:
              Subnet-Mask, BR, Time-Zone, Default-Gateway
              Domain-Name, Domain-Name-Server, Option 119, Hostname
              Netbios-Name-Server, Netbios-Scope, MTU, Classless-Static-Route
              NTP, Classless-Static-Route-Microsoft, Static-Route, Option 252
              RP
15:05:57.193585 00:0c:29:cc:f8:f6 > 66:66:66:66:66:12, ethertype IPv4 (0x0800), length 342: (tos 0x10, ttl 128, id 0, offset 0, flags (none), proto UDP (17), length 328)
    10.0.0.20.67 > 10.0.0.180.68: (udp sum ok) BOOTP/DHCP, Reply, length 300, xid 0xcede823f, secs 22, Flags (none) (0x0000)
          Your-IP 10.0.0.180
          Client-Ethernet-Address 66:66:66:66:66:12
          Vendor-rfc1048 Extensions
            Magic Cookie 0x63825363
            DHCP-Message Option 53, length 1: Offer
            Server-ID Option 54, length 4: 10.0.0.20
            Lease-Time Option 51, length 4: 14400
            Subnet-Mask Option 1, length 4: 255.255.255.0
            Default-Gateway Option 3, length 4: 10.0.0.20
            Domain-Name Option 15, length 17: "my.domain.for.dhcp"
            Domain-Name-Server Option 6, length 4: 10.0.0.20

What I do not see is the reply arriving at the Host (or at the requesting VM, obviously).

I had originally thought this was a VirtualBox problem, but now I am thinking there is some othher setting I am missing in Linux to allow layer 2 messages to macvlan MAC addresses to be routed correctly.

I do have ip_forward set to 1 in /proc/sys/net/ipv4/ip_forward.