Add SPDX license tags and checking tool

[robn]

Motivation and Context

After another fun cycle of OpenZFS licensing opinions being shared, I decided I would like to better understand what our licensing situation actually is, with a view to maybe someday setting a clear policy for new contributions and being able to enforce it.

A reasonable place to start is to add SPDX license identifier tags to every source file, and then write a program to enforce these for new files.

Description

Most of this PR is adding license tags to files that have clear and unambiguous license headers. It's arranged as multiple commits, but that's mostly a side effect of my method, which was a glorified search-and-replace.

The final commit adds a program called spdxcheck, to be run as part of checkstyle. This interrogates git for a canonical file list, searches them all for license tags, then decides if they have the correct license tag based on a number of factors. It reports on:

• files that do not have a tag that should
• files that have a tag that should not
• files that have the "wrong" tag for their path

I've documented a lot of this in spdxcheck itself, so I won't repeat it all here. It's sort of written with an eye towards a license policy that doesn't currently exist, but would likely say something like: we prefer X, we tolerate Y, and in these specific places, these different rules apply.

One important thing to note is that this PR is not trying to change any licenses, but rather reflect what's already there. Along the way, I found that we have quite a few files that don't have clear licenses (either ambiguous or omitted entirely), and many that are "wrong" for what they are, usually either GPL code in the core or FreeBSD platform code, or CDDL code in the Linux platform code. So, spdxcheck is carrying a lot of exclusions and overrides that I intend to sort out, but that will take time, as it's going to involve a lot of study of the history and a lot of actually asking original authors to clarify.

How Has This Been Tested?

No code changes. spdxcheck was developed in tandem with applying tags to files and building out exception lists, until it ran clean.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Performance enhancement (non-breaking change which improves efficiency)
• Code cleanup (non-breaking change which makes code smaller or more readable)
• Breaking change (fix or feature that would cause existing functionality to change)
• Library ABI change (libzfs, libzfs_core, libnvpair, libuutil and libzfsbootenv)
• Documentation (a change to man pages or other documentation)

Checklist:

• My code follows the OpenZFS code style requirements.
• I have updated the documentation accordingly.
• I have read the contributing document.
• I have added tests to cover my changes.
• I have run the ZFS Test Suite with this change applied.
• All commit messages are properly formatted and contain Signed-off-by.

[tonyhutter]

For the curious:

license	count
0BSD	1
LGPL-2.1-or-later	1
OpenSSL-standalone	1
Brian-Gladman-3-Clause	4
LicenseRef-OpenZFS-ThirdParty-PublicDomain	7
Apache-2.0	24
MIT	53
BSD-3-Clause	64
GPL-2.0-or-later	71
BSD-2-Clause	101
CDDL-1.0	2912

[tonyhutter]

I manually tested with two fake files. The changes have to be in your commit as the PR description mentions, but it worked correctly:

$ scripts/spdxcheck.pl
missing license tag: module/zfs/hello.c
invalid license tag: module/zfs/hello2.c
    (got Fake-license; expected: CDDL-1.0 0BSD BSD-2-Clause BSD-3-Clause MIT)

[robn]

FYI, if you were to remove all the exceptions and overrides in spdxcheck.pl (robn@41e4b6e), you get this output: gist robn/073a88c5. These are things that I intend to sort out over the coming months, as mentioned above.

[behlendorf]

Looks good. Thanks for helping to somewhat clarify the situation and adding a tool for this.

[robn]

Last push rebased to master and conflicts resolved.