~jj mentioned on the Jammy MP that another feasible option would be to
backport 2.91 due to the complexity of the upstream commit and later
followups to fix DNSSEC retry woes. This is not unprecedented: we did
already bump major versions in Jammy to address CVEs.

I looked at the diff between questing-devel and noble-devel and the
upstream changelog, and here are what I think are the most relevant
changes.

1. New --do-0x20-encoding option: explicit opt-in, does not change existing 
behavior if not set.
2. New --dhcp-option-pxe: also explicit opt-in.
3. --fast-dns-retries is now enabled by default when DNSSEC is active.
4. Changed behavior of --synth-domain when using IPv6 to be IDNA2008 compliant.
5. Truncated upstream answers return partial RRs instead of empty.
6. Some EDNS0 UDP packet size handling changes.
7. Improved TCP behavior for non-responsive/slow upstream servers.

(1) and (2) are probably safe since they are opt-in, but the others
represent non-trivial behavior differences that could trigger some
regressions for LTS users. Even considering that, I think it's probably
safer to do a full backport of 2.91 instead of trying to pick and
backport the specific changes related to DNSSEC+truncation+cache to
2.90.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2138412

Title:
  DNSSEC validation with stale cache enabled does not properly retry
  truncated response

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/dnsmasq/+bug/2138412/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to