[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: bind9 LTS



Hello,

On Sun 31 Mar 2024 at 09:51pm +08, Sean Whitton wrote:

> I've started looking at the first vulnerability, CVE-2023-4408, and have
> some confusions/questions.
>
> The ISC website that 9.11 is EOL as of March 2022.  But there is a lot
> of activity on the 9.11 branch, including a fix for this CVE.  Are we
> generally able to assume that changes are intended not to break anything
> for users?
>
> For example, commit 2fc28056b3 is a backport of API changes, and I can
> do the work to *confirm* that they don't appear to break anything for
> users, but I wouldn't like to rely on my own *discovery* as to whether
> they might break anything.
>
> At any point did you consider just backporting snapshots of upstream's
> 9.11 branch into LTS?  Do you know if any other vendors do that?  I'm
> wondering if, on balance, that might be safest -- if, that is, upstream
> are indeed not intending to break anything.
>
> Finally, do you you have any notes on testing?

Some follow-up.

- looks like backporting the old branches is what's done in bullseye and
  bookworm; do you know of some reason we're not doing this for buster too?

- CVE-2023-50387 and CVE-2023-50868 are both DoS vulnerabilities for
  DNSSEC.  The fixes for CVE-2023-50387 is large, and I am not sure
  there is one for CVE-2023-50868 for bind-9.11.

  I think that these fixes are too intrusive to fix by backporting,
  unless we decide to start backporting whole upstream 9.11.y releases.
  Would you agree?

-- 
Sean Whitton

Attachment: signature.asc
Description: PGP signature


Reply to: