tbuf errors

[rbbb at RICE.ARPA]

It is still possible to get the tbuf errors with rev 7 hardware.  We have
always had rev 7 hardware, and we have never had the writeable control
store.  We once had lots of errors, and now we do not.  How did we fix
this, back in the days before microcode updates?  I believe that we
replaced the L0001 module with one that used a different brand of chips.
This fix was discovered through a great deal of board-swapping; if anyone
has a similar story, please let me know, because this all happened a
looong time ago (but we still don't have the PCS, and we still run rev 7
hardware, and we don't get tbuf errors, and we swapped a pile of boards; I
just can't remember exactly which one).

And it is also true that Unix seems to get more of these than VMS
(including the user-mode machine checks).  Weird as shit.

drc