[NEWSBU.TXT]

	   REPTILE'S BUG REPORTS


This is the first of a regular(?) buglist. PLEASE
send a machine readable copy of your favorite SPR,
bug or "feature". If you see somthing here or get
a fix, let us know so we can warn all. 

1.	M, M-Plus Versions 4.1/2.1, updates C & D
	-----------------------------------------
Level of urgency: MANDATORY

VMR will screw the TT driver data base if you have
Muxes in your configuration. (ie DH-11,
DZ-11,DZV-11,DHU-11,DHV-11s). 

FIX: get VMR off an update E tape.

2.	RSX-11M-PLUS, Ver 2.1 E, V 3.0 Field test.
	------------------------------------------
LOU: RARE but hard to work around.

LOAd won't load a driver that has multiple
controller NAMES and thus multiple CTBs and DCBs
(for example a terminal driver with YL, YH, YZ,
YV). LOA claims the first S.KRB in the data base
is invalid. 

FIX: In next release??

3.	RSX11-M, RSX-11M-PLUS 4.1/2.1 Up D
	----------------------------------
LOU: Work around simple.

INI crashes system if you INI a RA81 on a UDA-50B
(note well: ALL UDA's should be -B's, also known
as the UDA-52. Field service should have replaced
all originals and may wish to know if you have one
they missed, even if you are not on DEC
maintance). 

WORKAROUND: Use BRU to tranfer a non existent file
/INI and specify all the parm's you want on the
command. 

FIX: Update E. I think the fix made it to the
Dispatch but don't hold me to it. 

4.	EMULEX CS02. Early Revs of microcode.
	-------------------------------------
LOU: see under.

A CS02 in DHV mode will crash your system. Known
problem and new microms fix it. Beat up your
Emulex salesman if you have an old one. PS they
work fine if you run them in DH mode. (Of course
you don't mind another sysgen, do you!) 


5.	RSX-11M V 4.1 Up C.
	-------------------
LOU: One bit fix.

If you have a 44 or 73 that is running like VMS,
read on. One of the undocumented features of a 44
is a hook for multi-processing a la 11/74 style.
Specifically you can by-pass the cache on a
per-APR basis. There is a jumper in the CPU that
hides this little quirk normally BUT, it would
appear that at some stage DEC stopped installing
the jumper (wire costs money you know...) AND poor
little RSX sets the cache bypass bit in PDR0. This
is the high order bit in the PDR and and 1xxxxxx
is loaded insted of 0xxxxxx. Bug is in INITAL
and/or SAVe. Sysgen not needed. Fix, reassemble,
and build RSX and/or SAVe. No need to rebuild Priv
tasks, as nothing shifts. Not all systems may be
hit by this, as I know that at least some early
44's had it disabled. ALL KDJ's will be affected.
(ie 11/73's, Micro-73's and 11/84's). 


5.	VAX/RSX V2
	----------
LOU: MANDATORY. No DEC fix known.

Problem: the AME will blow you away if an ALUN on
lun 0 is done. 

FCS will do an ALUN on lun 0 in some
circumstances. On a REAL COMPUTER it gets an
"Illegal Lun" DSW and happily goes about its
lawful business. With the 'New Improved
We-Know-Better User-Friendly Gee-Whizz
Quota-exceded (note the user friendly lower case)'
thing you get blown out of the water. 

FIX: Extract ASSLUN for syslib and DOB it. (if you
haven't got DOB or some other disassembler, the
DECUS library will be glad of your custom. And
your money) Fire up your favorite editor and find
the ALUN (NB it will be an EMT), just before it,
test R4, and if zero, cram $DSW with the error
code, set Carry,and exit. If DEC ship a fix for
this, please let all know. 

