Signficant bugfix to my emulator

[Mouse <mouse%Rodents-Montreal.ORG@localhost>]

I think at least a few of you are doing things with my MicroVAX-II
emulator.  Those of you who are might want to pull the latest version.

I had it hang when doing a certain compile.  Investigating, I found it
was livelocking, repeatedly taking the same translation-not-valid fault
on the same instruction, calling pmap_simulref, returning to userland,
and taking the same fault.

Adding a TBIA to pmap_simulref fixed this.  But, looking at what it
did, that shouldn't make any difference, because all it does is set
some valid bits.  Digging more, it turned out I was entering invalid
PTEs in the TB!  (The TB is defined to never cache info from invalid
PTEs.)

So I added two things: (1) I check that PTE_V is set before making TB
entries, and (2) there is a firewall test when making TB entries, to
ensure invalid PTEs don't slip in.  (I also took the TBIA back out from
pmap_simulref, but that's not part of the emulator proper.)

There also may be assorted other changes, depending on when you last
pulled.

Curiously, fixing this also fixed a *substantial* performance problem
that's been bothering me for quite some time.  (The perceived symptom
was that simple commands would take multiple seconds to start.)

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse%rodents-montreal.org@localhost
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B