[conspire] ZFS, Ubuntu, and Eben Moglen & Mishi Choudhary

[Rick Moen]

A fully thing happened on the way to the filesystem. 

Over this past decade, some Linux folks have had ZFS envy.
Recently, Canonical, Ltd. announced intent to cut corners on copyright
legality by shipping native ZFS support in Ubuntu 16.04 LTS, next month.
And, last, Eben Moglen & Mishi Choudhary of Software Freedom Law Center
wrote a _very_ peculiar analysis piece talking around the copyright
violation issue and suggesting that if the Linux kernel coders object to
this violation of their copyright, that they are somehow being
unreasonable.  (The most peculiar bit of that is where lawyers Moglen
and Choudhary advocate that permitting this violation is within the 
'"equity of the license" or its "spirit"' -- despite, as lawyers,
knowing that the wording of written formal licenses always trumps 
allegations about its 'spirit' and that equitable remedies are no
defence against copyright infringement.

Some background:

In 2005, Sun Microsystems developed ZFS (Zettabyte File System) as a
self-healing, self-checking, sophisticated filesystem and volume manager
for very large storage arrays, shipped in all releases of Sun
OpenSolaris since OpenSolaris 10 in 2005.  It's been a big hit for a
number of reasons -- among them eliminating the huge problem of massive
fsck times for large volumes.  (For context, if you don't store at least
terabytes of data, you're not yet in ZFS's league, which would be a godo
thing, as you won't need its huge RAM cost or complexity, either.)

For licensing, Sun applied to all of OpenSolaris (except a few
proprietary bits over which it didn't have ownership) CDDL, Sun's
improved version of the Mozilla Public Licence.  MPL was that rare
thing, an innovative open source licence creating a whole new category
of open source licences, sometimes called 'weak copyleft', where
licensors can choose to apply copyleft obligations or not to parts of
the codebase on a file-by-file basis.  CDDL (Common Development and
Distribution License) was Sun's slight improvement on MPL.

Meantime, Oracle Corporation gobbled up Sun, and terminated OpenSolaris,
which persists in independent forks but in general ZFS is little used
outside Oracle's storage appliances -- and in filesystem code ports to
other OSes, particularly porting of OpenZFS
(https://en.wikipedia.org/wiki/OpenZFS).


So, CDDL, a pretty good weak copyleft licence, became modestly
successful through persistence of OpenSolaris and some Java and other
codebases, but forms an island of code, because it clashes with the
requirements of GNU GPLv2 and later.  Some ex-Sun people claim this is
not true, and Canonical flacks such as James E.J. Bottomley claim it
doesn't matter.  They're wrong.

The BSDs and Apple OS X are able to use CDDL-licensed ZFS drivers
because the permissive licensing of their kernels is maximally
compatible.  The Linux kernel, though:  nope.  Linux's kernel devs 
have been very clear and consistent in their legal stance:  Anything you
create that is a derivative work of the Linux kernel, if you wish to
release it, must be licensed compatibly with the kernel's GPLv2 terms of
use.  CDDL ain't in that.

'Derivative work' is a legal technical term (term of art) in copyright
law.  Whether one work is derivative of another depends on factors
including what major copyright-protected 'expressive' elements of the
original have been copied, and other factors.  E.g., West Side Story was
derivative of Romeo and Juliet, and it's lucky for Sondheim that
Shakespeare is public domain.

Is a filesystem developed for a kernel derivative of it.  Very probably.
Is one ported to it, as IBM OpenAFS and more lately Sun ZFS have been?
Maybe, maybe not.  As the lawyers say, that is a triable claim of fact, 
e.g., we'd know for certain at the end of someone suing.


Before the Canonical, Ltd. announcement, there were two approaches to 
tacking ZFS onto Linux.  One was running a port of the code in userspace
as a 'FUSE' (Filesystem in Userspace) filesystem.   FUSE being a defined
interface specifically for running weird, differently licensed code,
this has been considered to work around the GPL/CDDL problem, not
creating a derivative work.  _But_ the performance hit is pretty bad.

Separately, there is 'ZFS on Linux', a native port in source form:
http://zfsonlinux.org/ .  The idea here is that, yes, applying these
patches to your Linux kernel quite possibly creates a derivative work of
that kernel, so _you_ get to do that and not distribute the result,
hence the devs commit no copyright violation.  _But_ this is a
troublesome workaround and (obviously) precludes distributing prebuilt
kernels.

A third alternative is to use Linux's workalike btrfs filesystem, which
is becoming respectable _but_ still has reliability issues and is
nowhere near as mature as ZFS is.


Everyone treated the GPL/CDDL clash as serious, and stuck to one of
those three options (or used other filesystem types for less insane
volume sizes) -- until Canonical came along recently and said 'Gosh, we
see no problem here.  We're going to just ship OpenZFS as part of our
Linux distro.'

http://arstechnica.com/gadgets/2016/02/zfs-filesystem-will-be-built-into-ubuntu-16-04-lts-by-default/
http://blog.dustinkirkland.com/2016/02/zfs-is-fs-for-containers-in-ubuntu-1604.html

  Dustin Kirkland
 
  Our lawyers have examined the licenses in detail and we see no problem.
  ZFS remains CDDL and the Kernel GPL. zfs.ko ships as a kernel module.


  Matthew Garrett
  +Dustin Kirkland Your lawyers disagree with (a) other lawyers, (b) the
  FSF and (c) several kernel copyright holders. It should be made clear
  that while you feel that you're legally justified in doing this, you're
  definitely violating the expectations of many other parts of the
  community in doing so.


  Simon Phipps
  +Matthew Garrett Setting aside the question of license compatibility
  itself for a moment, and given it's 100% free software that's involved
  here, why would a license incompatibility in this case provoke a fury
  usually reserved for proprietary abuse?


  Matthew Garrett
  +Simon Phipps The legal arguments we have against proprietary drivers
  are fundamentally about license compatibility, and normalising the idea
  that we can ignore license compatibility as long as the components in
  question are free weakens the legitimacy of our arguments against
  proprietary drivers.


So, we basically have yet another example of Canonical, Ltd. behaving in
a sociopathic way, in this case deciding to ignore the expressed views,
preferences, and explicit licensing terms of Linux kernel coders and
dare them to sue.  Hilarity may ensue.


Into this parade of badness now steps attroneys Eben Moglen & Mishi
Choudhary of Software Freedom Law Center, with a major display of
gratuitous WTFery:

http://softwarefreedom.org/resources/2016/linux-kernel-cddl.html

  The Linux Kernel, CDDL and Related Issues

  Eben Moglen & Mishi Choudhary
  February 26, 2016

This is a quite long, closely reasoned piece but pretty lucide even for
novices on the subject -- by two expert authors and open source / free
software licensing pundits.  Prof. Moglen has had very long and close
association with FSF and lead the legal side of the drafting of GPLv3,
for example.  

What is Software Freedom Law Center?  It's Moglen, Choudhary, and
Jonathan D. Bean operating as a semi-public interest law firm.
http://softwarefreedom.org/services/

In the current piece, Moglen & Choudhary recap the GPL/CDDL issue in
detail, and it's possible to get lulled into inattention by the sonorous
and lucid prose.  The WTFery starts near the end:

  Richard Stallman and I both felt, and said publicly, that the issue
  was a close one under a literal interpretation of GPLv2, but that the
  combination involved plainly fell within the equity or spirit of GPLv2. 
  [...]  In US copyright law, the meaning of any license is determined
  by the intentions of the licensor: because a license is a unilateral
  permission by a holder of otherwise exclusive rights, the law holds that
  the licensor gives neither more nor less permission than the evidence
  shows it actually intended.
  [...]
  No existing record conclusively or convincingly demonstrates whether
  the only relevant licensing community, the holders of kernel copyright,
  intends a literal or equitable interpretation of its license terms under
  present circumstances. As so often in the long history of our law, both
  literal and equitable postures of interpretation are completely tenable,
  and reasonable people in the relevant roles may justifiably disagree. 

WTF?  In any case of copyright infringement, judges look to the wording
of any explicit licence terms.  There is no weighing of 'equity or
sprit'.  This is nonsense, and Moglen knows it.

  [Discusses the Canonical announcement, then:]

  If there exists a consensus among the licensing copyright holders to
  prefer the literal meaning to the equity of the license, the copyright
  holders can, at their discretion, object to the distribution of such
  combinations. They would be asserting not that the binary so compiled
  infringes their copyright, which it does not, but that their exclusive
  right to the copying and redistribution of their source code, on which
  their copyright is maximally strong, is infringed by the publication of
  a source tree which includes their code under GPLv2 and ZFS filesystem
  files under CDDL, when that source tree is offered to downstream users
  as the complete and corresponding source code for the GPL'd binary.

  In response to such an objection, all distributors would no doubt cease
  distributing such combinations, which it would remain perfectly legal
  and appropriate for users to make for themselves. An objectively
  reasonable good-faith belief that the conduct falls within the equity of
  the license, until such time as the licensors state that they are
  interpreting the license literally, would be a full defense against
  claims of intentional infringement. A party swiftly adjusting its
  behavior in response to complaint is not an appropriate object of
  request for injunctive relief, [...]

So, basically, Moglen & Choudhary are pronouncing in advance that the 
Linux kernel coders insisting that GPLv2 _means GPLv2_ would be
unreasonable and mean, given that Canonical intends to comply with the 
'equity or sprit' of the terms they specified, and, gosh, although they 
could object and get Canonical to avoid infringing their copyright, 
it'd be wrongful for them to do that, and they should not be awarded an
injuction.

What the Actual Fsck, Moglen and Choudhary?  When did you become PR
shills for copyright infringement?  


Here is a third-party analysis of this bizarre situation, by one Serge
Wroclawski:

http://blog.emacsen.net/blog/2016/02/28/why-is-sflc-siding-with-oracle-over-linux-developers/

And one by Tom Marble:
http://info9.net/wiki//tmarble/posts/is-slfc-shooting-open-source-in-the-foot/


Both are worth reading.  For anyone reading Tom Marble's comment, he 
overstates the truth here:  

   A lawyer I respect called this out: "Equity" has no place in US law.
   [link]  The point is that for lawyers software licenses work because 
   they have clear, written rules to guarantee the spirit is upheld; 
   but spirit doesn't work in front of a judge -- clear rules do.

The link is to a Twitter exchange with Pamela S. Chestek, 'IP' attorney in
Raliegh, talking to ex-Sun computer scientist Simon Phipps:

   Simon Phipps:
   Again, doesn't SFLC's theory deal with that?

   Pamela S. Chestek:
   It does not. It is no place in US contract law for "equity" when
   there a written agreement. You follow the terms, period.

Exactly so.