slirp: domainname and classic stateless for DHCP

@Benjamin - please comment what you think and we will go from there together

I agree on point 3.

I incorporated the feedback on 0002-slirp-Add-classless-static-routes-support-to-DHCP-se.patch and posted patch v3: http://lists.nongnu.org/archive/html/qemu-devel/2018-03/msg04437.html

there was a reply yesterday which looked more like a nack still :-/
I pinged Benjamin on IRC.

But as deadlines are closing and this is a post Feature Freeze addition which might end up super hard to maintain I hope you get it accepted so that we can update it to what is upstream.

FYI: I discussed with a few others and the general recommendation seems to be to otherwise (not accepted in time) drop it from qemu in 18.04 so that you can bring it in later on.

Sadly since this can't get upstream in time I'm (for now) forced to remove it.
It is just that sort of non-upstream Delta that might become a long term maintenance pain e.g. as discussed upstream "... Sadly, once the QMP is released, we don't want to be changing it ...".

That combined with the fact that it was post FF and is a feature and also had no info ahead of time what it is about or anything else for now only lets me revert it to avoid issues inthe 18.04 life cycle.

Get me right, I like the change but for now need to follow the rules and think on 18.04 full lifecycle.

I'd love to see it in qemu thou, so if you:
- get it upstreamed in the next ~2 weeks please open a new bug against qmeu (Ubuntu)
- Link/mention the new bug here and refer to this bug from the new one
- make the new bug a Feature Freeze exception
- explain in the new bug all you can to convince the release Team.
- Get an FFe ack of the release Team on here
- put the final patches in qemu (patchname or dep3 header should refer to the new bug this time)
- If you want review provide commits against git.debian.org/git/pkg-qemu/qemu.git branch ubuntu-bionic-2.11
- I can help you then with regression tests, uploads, ... as needed

This bug was fixed in the package qemu - 1:2.11+dfsg-1ubuntu5

---------------
qemu (1:2.11+dfsg-1ubuntu5) bionic; urgency=medium

  * Revert the slirp changes of 1:2.11+dfsg-1ubuntu3 until they are upstream
    accepted to be better long term maintainable (LP: #1753938)

 -- Christian Ehrhardt <email address hidden> Thu, 22 Mar 2018 10:31:23 +0100

Sorry for missing the IRC ping. I agree to drop the classless static routes support (patch 2) since this needs bigger changes than expected. So let's get this done upstream.

The domainname option (patch 1) is uncontroversial and small. It was review and Samuel Thibault and he applied it to his tree: http://patchwork.ozlabs.org/patch/878667/ I don't know the workflow. What is the next step from his tree? Can I consider this patch upstreamed?

It isn't really upstream yet, Samuel usually asks for pulls on the start of a version.
fd06527b was the last for 2.12 so I'd assume it only gets "really upstream" around mid-May.

Of course under the current conditions we can assume this patch will make it.

The question is now your urgency:
a) is having #1 without #2 in Ubuntu 18.04 super important for you(only then worth the hazzle so late), then lets consider taking that one back in.
b) or are you just as fine with the qemu in Ubuntu 18.10. In that case given history we would look at 2.12 maybe 2.12.1 (Expecting rather late 2018 release of 2.13). But at the time we would work on qemu 2.12 for Ubuntu 18.10 your changes would be upstream and we could pick them into "our" 2.12 from qemu/master.

So it is up to you, if you think you can (and want to) convince the Release Team of an FFE for change #1.
If so please as I mentioned before open a new bug, make it an FFE, explain all you need why you think so, and if possible subscribe me right away.

Otherwise if b) is enough lets work together on it later this year.

Having #1 without #2 is important for me (option a). Patch #2 is nice to have, but patch #1 is required for making testing live-boot with QEMU working. Without the domainname support the boot will fail (at least if you have custom boot scripts that operate on the FQDN).

Ok Benjamin, I filed bug 1762315 for a) on change #1

I'm rather sure it is a good change since it is already tested by me and was in Bionic a while without complaints. Never the less I need an FFE given the time. So lets see if we run into any trouble there being so late.

If you get to upstream #2 completely let me know for some later Ubuntu versions to integrate it (if the normal pick via upstream merge isn't fast enough).