(t) avoid recalculating pool free

[phillxnet]

(t) = testing branch. While working on #2384 and revising mocking of such things as pool free space, we see a redundant re-calculation of the same in our PUT (remove) pool command when the exact same calculation has already been made via an earlier Pool object @Property instantiation still held.

Pool "free" property definition:

rockstor-core/src/rockstor/storageadmin/models/pool.py

Lines 97 to 104 in afa3e41

	@property
	def free(self, *args, **kwargs):
	# Why do we compute pool usage on the fly like this and not like
	# share usage as part of state refresh? This is a lot simpler and
	# less code. For share usage, this type of logic could slow things
	# down quite a bit because there can be 100's of Shares, but number
	# of Pools even on a large instance is usually no more than a few.
	return self.size - pool_usage(self.mnt_pt_var)

storageadmin.views.pool.PoolDetailView -put()

rockstor-core/src/rockstor/storageadmin/views/pool.py

Lines 700 to 705 in afa3e41

	usage = pool_usage("/{}/{}".format(settings.MNT_PT, pool.name))
	size_cut = 0
	for d in disks: # to be removed
	size_cut += d.allocated
	available_free = pool.size - usage
	if size_cut >= available_free:

This is obviously redundant. And given the second in-transaction re-use/repeat-call is the only other user of fs.btrfs.pool_usage() we can improve/ease our traceability and mocking by centralising this methods use.

[FroggyFlox]

Closing as fixed by #2387 .