Support setting user properties in a channel program

[jasonbking]

This adds support for setting user properties in a
zfs channel program by adding 'zfs.sync.set_prop'
and 'zfs.check.set_prop' to the ZFS LUA API.

Motivation and Context

Work in SmartOS utilities this features, and it seems generally useful enough that the broader ZFS community might also find it useful. This change is limited to only zfs user properties. Setting user properties is all our work required, and that seems a useful delineation point--handling non-user properties in a channel program is a far more complex proposition (though nothing in this work should preclude extending it to support non-user properties in the future).

Description

This just adds the two new channel programs zfs.sync.set_prop and zfs.check.set_prop. They validate the property is a user property (and return EINVAL if the property is not a user property), and then set the property.

How Has This Been Tested?

This has been in use in a SmartOS development branch for several months (aside from paths, the code is identical between OpenZFS and illumos ZFS) without issue. In addition, a new test for the zfs test suite has been added which sets and verifies a user property in a channel program.

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)
• Documentation (a change to man pages or other documentation)

Checklist:

• My code follows the ZFS on Linux 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.

Contributions-by: Jason King jason.king@joyent.com
Signed-off-by: Sara Hartse sara.hartse@delphix.com
Signed-off-by: Jason King jason.king@joyent.com

[codecov]

Codecov Report

Merging #9950 into master will increase coverage by 0.02%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #9950      +/-   ##
==========================================
+ Coverage   79.18%   79.21%   +0.02%     
==========================================
  Files         385      386       +1     
  Lines      122382   122424      +42     
==========================================
+ Hits        96904    96973      +69     
+ Misses      25478    25451      -27     

Flag	Coverage Δ
#kernel	79.37% <100.00%> (-0.02%)	⬇️
#user	66.54% <0.00%> (-0.16%)	⬇️

Impacted Files	Coverage Δ
module/os/linux/spl/spl-kmem-cache.c	75.45% <0.00%> (-9.46%)	⬇️
module/zfs/bpobj.c	86.86% <0.00%> (-4.29%)	⬇️
module/lua/lmem.c	83.33% <0.00%> (-4.17%)	⬇️
lib/libzpool/kernel.c	64.09% <0.00%> (-2.96%)	⬇️
module/zfs/dsl_synctask.c	92.40% <0.00%> (-2.54%)	⬇️
lib/libzfs/libzfs_changelist.c	84.37% <0.00%> (-1.96%)	⬇️
module/os/linux/zfs/vdev_file.c	80.37% <0.00%> (-1.87%)	⬇️
module/zfs/vdev_queue.c	94.49% <0.00%> (-1.84%)	⬇️
module/zfs/vdev_indirect_births.c	89.53% <0.00%> (-1.17%)	⬇️
cmd/ztest/ztest.c	82.70% <0.00%> (-0.81%)	⬇️
... and 49 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 610eec4...4d24afc. Read the comment docs.

[behlendorf]

Thanks. This looks good just a couple tiny nits.

[ahrens]

Since this is essentially a helper for the _sync() func, I'm not sure it makes sense for this to return a value, since it "can't fail" (except via longjmp)

[jasonbking]

I'll change it to void

[jasonbking]

Should be all squashed and rebased to current now... I think I've addressed all the comments (at least so far)..

[shartse]

Looks good!