Does the KA630 have a TB?

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

Does anyone happen to know whether the KA630 (the MicroVAX-II) has a
translation buffer, and if so how big it is and what its replacement
policy is?

I added a translation buffer to my emulator and promptly got crashes
from the NetBSD version I'm running on it.  My translation buffer is
big enough to hold a full 8M of mappings (I could have got away with
6M, but it didn't seem worth the additional code complexity).  On
digging, it seems that the NetBSD version I'm trying to use is missing
some mtpr(0,PR_TBIA) calls.

This led me to wonder how it ever worked on that hardware (which it
did; it's what I ran back when I was running my real KA630).  The only
conclusions I can come to are (1) either the KA630 doesn't have a TB
(which sounds very nonperformant - each userland VM access would
involve two additional trips to main memory) or (2) it does, but its
size combined witih its replacement policy are such that the stale
mappings got flushed before they mattered.

I could just fix the kernel to TBIA/TBIS more correctly (and as an
interim step that's exactly what I'm doing).  But I'd like to make my
emulation more accurate, if enough information to do so is publicly
available.

/~\ 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