Micro-optimize zrl_remove().

[amotin]

atomic_dec_32() should be a bit lighter than atomic_dec_32_nv().

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.

[ryao]

Is the micro-optimization here just reduced register pressure from the compiler not having to worry about a register being clobbered with the result, or is there some micro-architectural detail of how these operations work that make one more expensive than the other beyond the register pressure from one version clobbering a register?

For reference, in the FreeBSD source code on amd64, atomic_dec_32_nv() becomes atomic_fetchadd_int() and atomic_dec_32() becomes atomic_subtract_int(). The relevant lines are:

https://github.com/freebsd/freebsd-src/blob/424bf1d5ca5bdba0597b546ec3931eb05bfabcab/sys/amd64/include/atomic.h#L192-L203

https://github.com/freebsd/freebsd-src/blob/424bf1d5ca5bdba0597b546ec3931eb05bfabcab/sys/amd64/include/atomic.h#L370

https://github.com/freebsd/freebsd-src/blob/424bf1d5ca5bdba0597b546ec3931eb05bfabcab/sys/amd64/include/atomic.h#L117-L123

[amotin]

@ryao It was MOV + LOCK XADD, now it is just LOCK SUB. Minus register, extra instruction and extra micro-operation. Not much, but it is a primitive to be used by other code, so better be optimal.

[behlendorf]

@amotin would you mind rebasing and force updating this PR. That should hopefully clear up the checkstyle failure at least.

[amotin]

@behlendorf done.