[linux-cifs-client] [PATCH 00/15] cifs: implement multisession mounts (RFC)

[Jeff Layton]

On Wed, 31 Mar 2010 13:18:54 +0100
Jamie Lokier <jamie at shareable.org> wrote:

> Jeff Layton wrote:
> > To be honest, I have doubts anyway about whether we should take this
> > patch into what seems to be considered a legacy codebase these days.
> > It's rather difficult to bolt this functionality onto the existing CIFS
> > codebase. It's probably more suited to putting into the newer SMB2
> > codebase. That said, it'll probably be just as difficult to do this
> > there since most of that code was copied from CIFS.
> 
> Legacy?
> 
> Maybe only in environments where everyone is running only Windows
> servers, all of them less than 2 years old :-)  (I've yet to see such
> an environment, btw).
> 
> The stable version of Samba (3.x) supports only SMB1.
> 
> Samba 4, which does support SMB2, is "not yet in a state where it can
> replace existing production deployments"; the Ubuntu package
> description says "experimental, should not be used in production".
> 
> In fact when I wanted to deploy an SMB2 service from Linux recently
> (to get better file link semantics on a Windows client), I gave up on
> it, it was too disruptive to replace Samba 3 with Samba 4.
> 
> This per-user patchset you've produced sounds quite useful, thank you.
> Please don't think you are targetting only a few horribly outdated
> environments with it :-)
> 

Ok...legacy is probably too strong a word... :)

The main problem still stands though. This is a very disruptive patch
set and CIFS really wasn't designed with this in mind.

Part of that paragraph comes from my frustration with the direction
taken by the direction of SMB2 development. Rather than building on the
existing CIFS codebase and abstracting it out where needed, it was
deemed less disruptive to copy large swaths of CIFS wholesale
and build an entirely new SMB2 fs from that.

That of course means that any improvements or bugfixes to CIFS have to
essentially be "ported" to the SMB2 code, which is a tremendous
disincentive to do anything large scale on either code base.

In any case, I plan to keep working on this, but I'm starting to wonder
whether my time would be better spent starting an entirely new SMB
filesystem that's multiuser from the get-go and that could handle
either SMB version.

Thanks,
-- 
Jeff Layton <jlayton at samba.org>