VMware Fusion Player vms cannot use IPV6?

Host is a Mac mini X86 running Sonoma 14.2.1, VMware Fusion Player Version 13.5.0 (which are current as of this writing.). Guest vms include Windows 11, macOS Sonoma 14.2.1, and a Linux, where I was best able to isolate and diagnose the problem. Consider this transcript (with some extra blank lines for readability), which starts in Terminal on the host, and continues via ssh on a fresh (File/New/Create custom ...) vm running an installation CD:

scott@Mac-mini-x86 ~ $ scutil --dns |  tail -9
DNS configuration (for scoped queries)

resolver #1
  search domain : attlocal.net
  nameserver : 2600:1700:6b0:e60::1
  nameserver : 192.168.1.254
  if_index : 11 (en1)
  flags    : Scoped, Request A records, Request AAAA records
  reach    : 0x00020002 (Reachable,Directly Reachable Address)


scott@Mac-mini-x86 ~ $ ping -c 1 www.ibm.com
PING user-att-162-198-200-0.e7817.dscx.akamaiedge.net (23.0.23.183): 56 data bytes
64 bytes from 23.0.23.183: icmp_seq=0 ttl=53 time=14.879 ms

--- user-att-162-198-200-0.e7817.dscx.akamaiedge.net ping statistics ---
1 packets transmitted, 1 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 14.879/14.879/14.879/0.000 ms


scott@Mac-mini-x86 ~ $ ping6 -c 1 www.ibm.com
PING6(56=40+8+8 bytes) 2600:1700:6b0:e60:b9a0:4f76:4b85:6a1d --> 2600:1407:7400:d87::1e89
16 bytes from 2600:1407:7400:d87::1e89, icmp_seq=0 hlim=50 time=32.135 ms

--- user-att-162-198-200-0.e7817.dscx.akamaiedge.net ping6 statistics ---
1 packets transmitted, 1 packets received, 0.0% packet loss
round-trip min/avg/max/std-dev = 32.135/32.135/32.135/0.000 ms


scott@Mac-mini-x86 ~ $ ssh $vm



livecd ~ # ping -c 1 23.0.23.183
PING 23.0.23.183 (23.0.23.183) 56(84) bytes of data.
64 bytes from 23.0.23.183: icmp_seq=1 ttl=53 time=21.6 ms

--- 23.0.23.183 ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
rtt min/avg/max/mdev = 21.590/21.590/21.590/0.000 ms


livecd ~ # ping6 -c 1 2600:1407:7400:d87::1e89
PING 2600:1407:7400:d87::1e89(2600:1407:7400:d87::1e89) 56 data bytes

--- 2600:1407:7400:d87::1e89 ping statistics ---
1 packets transmitted, 0 received, 100% packet loss, time 0ms


livecd ~ # ping -c 1 www.ibm.com
PING www.ibm.com(g2600-1407-7400-0d94-0000-0000-0000-1e89.deploy.static.akamaitechnologies.com (2600:1407:7400:d94::1e89)) 56 data bytes

--- www.ibm.com ping statistics ---
1 packets transmitted, 0 received, 100% packet loss, time 0ms


livecd ~ # cat /etc/resolv.conf 
# Generated by dhcpcd from enp2s1.dhcp, enp2s1.dhcp6, enp2s1.ra
# /etc/resolv.conf.head can replace this line
domain attlocal.net
nameserver 192.168.1.254
nameserver 2600:1700:6b0:e60::1
# /etc/resolv.conf.tail can replace this line

First thing to note is that although this example was run on an installation image in a fresh vm, essentially the same thing happens in a Windows VM or a macOS VM -- ipv6 doesn't work! Back to the story: I ping www.ibm.com, both with ping (ipv4) and ping6 (ipv6) while at the host -- works great, and quickly. In the guest vm, the ping still works, and quickly, but the ping6 just times out. Why did I use explicit addresses on the guest? The last experiment shows that on the guest, www.ibm.com will resolve to an ipv6 address (DNS lookup working, right?) but then can't be reached 🧐. Note also that scutil (on the Mac) or cat /etc/resolv.conf (on the guest) both show that same addresses, ipv4 and ipv6, for dsldevice.attlocal.net, my router. Perhaps there is some configuration magic in the .vmx file or something similar that could make ipv6 just work, or if not, make ipv6 fail in a way that things don't get confused, and just use ipv4 like the Before Times.