Refactor spa, vdev, and metaslab commands

[sdimitro]

These commands were hacked together during the initial prototype of sdb based on crash-python and then they were shoved into drgn-based prototype from which point they were never revisited. As a result their code is far from optimal and since they are used so often we'll want to refactor them soon. Below is a list of problems with that code:

[1] man/help messages are missing - also the descriptions of each of their command line options is lacking

[2] The arg_string convention used/passed between these commands is very ad-hoc and MDB-like (meaning Python should allow us to do better) - here is what I'm talking about

spa.py:

    def __init__(self, args: str = "", name: str = "_") -> None:
        super().__init__(args, name)
        self.arg_string = ""
        if self.args.metaslab:
            self.arg_string += "-m "
        if self.args.histogram:
            self.arg_string += "-H "
        if self.args.weight:
            self.arg_string += "-w "

...

    def pretty_print(self, spas: Iterable[drgn.Object]) -> None:
            ...
            if self.args.vdevs:
                vdevs = sdb.execute_pipeline([spa], [Vdev()])
                Vdev(self.arg_string).pretty_print(vdevs, 5)   // <---- arg_string passed here

then in vdev.py we do the same for metaslab.py.

[3] It is annoying that in order to print just the metaslabs for now through spa (e.g. ::spa -m in mdb) we are forced to provide the -v option on top of -m. E.g.

sdb> spa -m                           // <----- The -m option is basically a no-op here
ADDR               NAME
------------------------------------------------------------
0xffff94234ca54000 rpool


sdb> spa -vm
ADDR               NAME
------------------------------------------------------------
0xffff94234ca54000 rpool
      ADDR               STATE   AUX  DESCRIPTION
      ------------------------------------------------------------
      0xffff94234cc8c000 HEALTHY NONE  root
      0xffff94234cde4000 HEALTHY NONE    /dev/sda1
           ADDR                 ID           OFFSET     FREE  FRAG     UCMU
           -----------------------------------------------------------------
           0xffff94234cf94000    0              0x0     59MB   54%       0B
           0xffff94234cf93000    1       0x20000000    301MB    6%       0B
           0xffff94234cf92000    2       0x40000000    307MB    0%       0B

In general we need to rethink what we want the behavior of the options to be and think of a clean way to pass them between those commands (see [2]).

[4] As it is visible in the regression output introduced in the histogram PR #207 , we have a pool (rpool) where the histograms feature is not enabled. It would be nice to check if that version is enabled before attempting to print anything histogram related.