unbound segfaults on SIGHUP

[mmac]

pfSense+ 21.02-RELEASE (amd64) on Netgate SG-2220

DNS resolver (unbound) crashed with sefault (signal 11) on one of my firewalls right after dhcpd sent it SIGHUP. Here is relevant section of log file:

Apr  6 01:43:41 bubak dhcpleases[25820]: Sending HUP signal to dns daemon(25159)
Apr  6 01:43:41 bubak kernel: pid 25159 (unbound), jid 0, uid 59: exited on signal 11

It did not restart as evidenced by subsequent failures to deliver SIGHUP to it (and family complaining about "Internet being down" due to non-functional DNS)

Apr  6 01:53:01 bubak dhcpleases[25820]: Sending HUP signal to dns daemon(25159)
Apr  6 01:53:01 bubak dhcpleases[25820]: Could not deliver signal HUP to process 25159: No such process.
...
Apr  6 02:38:54 bubak dhcpleases[25820]: Sending HUP signal to dns daemon(25159)
Apr  6 02:38:54 bubak dhcpleases[25820]: Could not deliver signal HUP to process 25159: No such process.
Apr  6 02:38:54 sklad syslogd: last message repeated 1 times
Apr  6 02:38:54 bubak dhcpleases[25820]: Sending HUP signal to dns daemon(25159)
Apr  6 02:38:54 bubak dhcpleases[25820]: Could not deliver signal HUP to process 25159: No such process.
Apr  6 02:38:54 sklad syslogd: last message repeated 1 times

My home network has only 1 DNS server, so the failure causes all kinds of problems. I understand that the root cause is almost certainly a problem in unbound, however pfSense simply MUST supervise unbound and restart it if it dies. The problem is relatively rare, but the failure mode is unacceptable and AFAICT, fix should quite simple.