Future directions - !SMB1 by 2020 ?

Garming Sam garming at catalyst.net.nz
Thu Dec 13 21:04:15 UTC 2018

Does removing NetBIOS require implementing the new SOAP based discovery?


I now have some experience dealing with these SOAP protocols because of
investigating AD Powershell support (at least from a Python perspective)
-- we have a PoC working which I should actually write up about later.
Honestly, implementing this kind of stuff in C is just a bit of a mess
(if you read on the bug: using xml in C scares metze, and he's not
alone), and I would be curious about actually writing new daemons in
some other language. If we are given a chance to ditch some of the
legacy code, we do have a bit more freedom in how we implement newer
features. That would be nice to take advantage of sometimes.



On 13/12/18 10:59 AM, Jeremy Allison via samba-technical wrote:
> So I've been thinking some long-term thoughts
> about the Samba project.
> I'd like to propose a radical idea.
> I'd like to remove the SMB1 and NetBIOS
> (nmbd etc.) code from Samba by the end of
> year 2020.
> I'm guessing that will be a Samba version
> we can call Samba 5.x (if we haven't already
> moved before then).
> Of course I'll have to get off my ass and
> finish the SMB3+ UNIX extensions (ongoing)
> in order to match current Linux -> Samba
> functionality.
> But use of SMB1 is small and diminishing,
> and keeping this code functional will
> eventually require wasting resources that
> would be better spent on SMB2+ improvements.
> It'll be easier to restructure the server
> around clustered SMB3 when we're not continuously
> worrying about breaking SMB1 rap call tests :-).
> For people who require SMB1 support for the
> forseeable future, they can remain on Samba
> 4.x which which will get maintained by vendors
> and consulting companies (in the same way
> we sometimes do patches for Samba 3.0.x
> for security).
> The need for dual access from SMB1 and
> SMB3 simultaneously I would argue will
> be so small that it's safe to say if
> people want SMB3+ they use the new
> supported code, and if they have to
> maintain SMB1 then they use the legacy
> Samba 4.x code.
> Thoughts everyone ?
> Jeremy.

More information about the samba-technical mailing list