Trailing-Edge
-
PDP-10 Archives
-
BB-T573C-DD_1986
-
5,14/tgha.doc
There are 19 other files named tgha.doc in the archive. Click here to see a list.
COPYRIGHT (c) DIGITAL EQUIPMENT CORPORATION 1979,1981,1984,1986. ALL RIGHTS RESERVED.
THIS SOFTWARE IS FURNISHED UNDER A LICENSE AND MAY BE USED AND COPIED
ONLY IN ACCORDANCE WITH THE TERMS OF SUCH LICENSE AND WITH THE
INCLUSION OF THE ABOVE COPYRIGHT NOTICE. THIS SOFTWARE OR ANY OTHER
COPIES THEREOF MAY NOT BE PROVIDED OR OTHERWISE MADE AVAILABLE TO ANY
OTHER PERSON. NO TITLE TO AND OWNERSHIP OF THE SOFTWARE IS HEREBY
TRANSFERRED.
THE INFORMATION IN THIS SOFTWARE IS SUBJECT TO CHANGE WITHOUT NOTICE
AND SHOULD NOT BE CONSTRUED AS A COMMITMENT BY DIGITAL EQUIPMENT
CORPORATION.
DIGITAL ASSUMES NO RESPONSIBILITY FOR THE USE OR RELIABILITY OF ITS
SOFTWARE ON EQUIPMENT WHICH IS NOT SUPPLIED BY DIGITAL.
Page 2
Table of Contents
1.0 ABSTRACT . . . . . . . . . . . . . . . . . . . . . . 3
2.0 INSTALLATION INSTRUCTIONS . . . . . . . . . . . . . 3
2.1 INSTALLATION ON TOPS10 . . . . . . . . . . . . . . 4
2.2 INSTALLATION ON TOPS20 . . . . . . . . . . . . . . 4
3.0 AN OVERVIEW OF TGHA . . . . . . . . . . . . . . . . 4
3.1 RUNNING TGHA . . . . . . . . . . . . . . . . . . . 4
3.2 MAKING LIST FILES . . . . . . . . . . . . . . . . 4
3.3 TGHA FILE REQUIREMENTS . . . . . . . . . . . . . . 5
3.4 INITIAL SYSTEM START UP . . . . . . . . . . . . . 6
3.5 MF20 ERRORS . . . . . . . . . . . . . . . . . . . 6
3.6 KNOWN ERROR DETERMINATION . . . . . . . . . . . . 6
3.7 CORRECTIVE ACTION . . . . . . . . . . . . . . . . 7
3.8 PARITY ERRORS . . . . . . . . . . . . . . . . . . 7
4.0 HISTORY FILE . . . . . . . . . . . . . . . . . . . . 8
4.1 DIRECTORY PAGE . . . . . . . . . . . . . . . . . . 8
4.2 GROUP PAGE . . . . . . . . . . . . . . . . . . . . 8
4.3 STORAGE MODULE PAGE . . . . . . . . . . . . . . . 9
4.4 UNUSED STORAGE MODULES . . . . . . . . . . . . . 10
5.0 TRACE FILE . . . . . . . . . . . . . . . . . . . . 10
5.1 SPARE BIT SWAP . . . . . . . . . . . . . . . . . 10
5.2 PARITY ERROR . . . . . . . . . . . . . . . . . 11
5.3 SERIOUS MF20 HARDWARE FAILURES . . . . . . . . . 11
5.4 USE OF SPARE BIT BY KLI . . . . . . . . . . . . 11
5.5 KLI DETECTED BAD BLOCK . . . . . . . . . . . . . 11
5.6 MISSING DATABASE . . . . . . . . . . . . . . . . 12
5.6.1 RECOVERY PROCEDURE . . . . . . . . . . . . . . 12
5.7 SOFTWARE STATE WORD . . . . . . . . . . . . . . 12
6.0 SPEAR . . . . . . . . . . . . . . . . . . . . . . 13
6.1 NEW KNOWN ERROR: . . . . . . . . . . . . . . . . 13
6.2 SERIOUS MF20 HARDWARE FAILURE . . . . . . . . . 13
6.3 USE OF SPARE BIT BY KLI . . . . . . . . . . . . 13
6.4 KLI DETECTED BAD BLOCK . . . . . . . . . . . . . 14
6.5 MISSING DATABASE . . . . . . . . . . . . . . . . 14
6.6 SOFTWARE STATE WORD . . . . . . . . . . . . . . 14
7.0 MF20 MEMORY CONTROLLER SBUS DIAGNOSTIC FUNCTIONS . 15
8.0 MF20 ARRAY BOARD . . . . . . . . . . . . . . . . . 22
9.0 MG20 ARRAY BOARD (M8570) . . . . . . . . . . . . . 23
10.0 MF20 ERRORS . . . . . . . . . . . . . . . . . . . 24
10.1 CORRECTION CODE (ECC) CALCULATION . . . . . . . 24
10.2 DETECTION AND CORRECTION . . . . . . . . . . . . 26
10.3 DOUBLE BIT ERROR SYNDROMES . . . . . . . . . . . 27
11.0 KL10 I/O INSTRUCTIONS RELATING TO MEMORY . . . . . 29
Page 3
1.0 ABSTRACT
TGHA - THE GREAT HEURISTIC ALGORITHM
TGHA IS THE MOS MEMORY ANALYZER PROGRAM THAT IS RUN BY THE MONITOR,
WHENEVER A CORRECTABLE ERROR OCCURS IN AN MF20. IT HAS THE
RESPONSIBILITY FOR KEEPING THE MF20 MEMORIES IN A VIABLE CONDITION USING
THE MF20 MAINTAINENCE FEATURES. WHEN RUN BY THE MONITOR, TGHA OBTAINS
DATA ABOUT THE MF20 ERROR FROM THE MONITOR. TGHA CAN ALSO BE RUN BY A
USER WITH AT LEAST MAINTAINANCE PRIVILEGES. IN THIS MODE, TGHA IS USED
TO CREATE ASCII FILES (BIT TO TEXT) FROM TGHA'S DATA BASE. HISTORY.LST
CONTAINS INFORMATION ABOUT THE HARDWARE CONFIGURATION, THE STATE OF THE
CONTROLLER CORRECTIVE LOGIC, AND A LIST OF LOGGED ERRORS. TRACE.LST
CONTAINS ENTRIES TRACKING THE CORRECTIVE ACTION TAKEN BY TGHA. TGHA
4(0) CAN BE RUN WITH EITHER OR BOTH MF20 (16K) AND MG20 (64K) ARRAY
MODULES IN THE MEMORY SYSTEM. HOWEVER, BOTH TYPES CANNOT BE INSTALLED
WITHIN THE SAME CONTROLLER
1. TGHA SOURCES WILL NOT BE AVAILABLE TO CUSTOMERS.
2. THIS DOCUMENT APPLIES TO BOTH TOPS10 AND TOPS20 UNLESS
OTHERWISE SPECIFIED.
3. IF THE MEMORY BOOT PORTION OF KLI DOES NOT CONFIGURE A BLOCK OF
MF20 MEMORY DUE TO MULTIPLE BIT ERRORS, TGHA WILL INDICATE WHAT
MEMORY AND HOW MUCH MEMORY WAS NOT CONFIGURED. THE MESSAGE
WILL APPEAR ON THE CTY, IN THE TGHA TRACE FILE, AND IN
ERROR.SYS.
4. IF THE MEMORY BOOT PORTION OF KLI FINDS A PARITY ERROR, KLI
WILL USE THE SPARE BIT AND CONFIGURE THE BLOCK. TGHA, UPON
FINDING THAT A SPARE BIT WAS SET INTO USE BY KLI, WILL NEVER
USE THAT SPARE BIT FOR ANY OTHER PURPOSE. UPON FINDING SUCH A
CONDITION, TGHA WILL OUTPUT AN APPROPRIATE MESSAGE ON THE CTY,
IN THE TRACE FILE, AND IN ERROR.SYS.
5. IF A PARITY ERROR EVER OCCURS DURING SWAPPING IN OF A SPARE
BIT, THE SPARE BIT WILL THEN BE FLAGGED AS IF KLI HAD SET THE
SPARE BIT INTO USE (SEE PREVIOUS PARAGRAPH). SIMILAR CTY,
TRACE FILE AND SPEAR ENTRIES WILL BE MADE.
2.0 INSTALLATION INSTRUCTIONS
TGHA DATA AND TRACE FILE NAMES ARE TGHA.DAT AND TGHA.TRA RESPECTIVELY.
THE OLD FILES TGHAV2 OR TGHAV3 DAT/TRA SHOULD BE RENAMED TO PRESERVE THE
DATABASE.
Page 4
2.1 INSTALLATION ON TOPS10
FOR TGHA TO RUN PROPERLY ON TOPS10, THE FOLLOWING ENTRY MUST BE MADE IN
OPR.ATO AFTER DAEMON IS STARTED::
1. :SLOG
2. :DEF TGHA=
3. R TGHA
SINCE TGHA MAKES ENTRIES IN ERROR.SYS VIA DAEMON, DAEMON MUST BE RUNNING
BEFORE TGHA IS STARTED. TGHA.EXE MUST RESIDE IN THE [1,4] AREA:
2.2 INSTALLATION ON TOPS20
TGHA.EXE MUST RESIDE IN PS:<SYSTEM> AREA:
3.0 AN OVERVIEW OF TGHA
3.1 RUNNING TGHA
1. THE MONITOR RUNS TGHA AT INITIAL SYSTEM START UP
2. THE MONITOR RUNS TGHA IN RESPONSE TO MF20 ERRORS
3. FIELD SERVICE CAN RUN TGHA TO OBTAIN LIST FILES
3.2 MAKING LIST FILES
************************************************************************
NOTE:
IN ORDER TO RUN TGHA MANUALLY, THE USER MUST HAVE THE MINIMUM OF
MAINTAINENCE CAPABILITIES AND BE ENABLED. THIS IS REQUIRED BECAUSE TGHA
LOOKS AT THE MF20 MEMORY TO DETERMINE THE CURRENT CONFIGURATION.
************************************************************************
TGHA CAN ALSO BE RUN BY FIELD SERVICE PERSONNEL. IN THIS MODE, TGHA
KNOWS THAT IT IS NOT THE SYSTEM TGHA. THEREFORE, TGHA DOES NOT
AUTOMATICALLY LOOK FOR MF20 ERRORS.
TGHA WILL RESPOND WITH A PROMPT OF:
TGHA>
IF "HELP"<CR> IS TYPED, THE FOLLOWING TEXT WILL BE PRINTED
Page 5
TGHA HELP - COMMAND PURPOSE
EXIT EXIT FROM TGHA.
HELP TYPE THIS TEXT.
HISTORY DUMP HISTORY FILE.
TRACE DUMP TRACE FILE.
TGHA>
THE HISTORY AND TRACE DUMP FILES WILL BE CREATED IN THE AREA FROM WHICH
TGHA HAS BEEN RUN. THEY WILL BE HISTRY.LST FOR TOPS10 AND HISTORY.LST
FOR TOPS20 AND TRACE.LST RESPECTIVELY. THE HISTORY AND TRACE DUMP FILES
WILL BE CREATED IN THE AREA FROM WHICH TGHA HAS BEEN RUN.
THESE FILES ARE LISTABLE VERSIONS OF THE TGHA DATABASE AND THE TRACE
FILE. ALTHOUGH TGHA.TRA MAY LOOK READABLE, REMEMBER THAT IT IS THE
EQUIVALENT OF A RING BUFFER. IE. THE OLDEST DATA IS OVER WRITTEN WHEN
THE FILE BECOMES LARGE ENOUGH. THIS PREVENTS THE TRACE FILE FROM
BECOMING USELESSLY LARGE. TRACE.LST IS IN CHRONOLOGICAL ORDER, THE
OLDEST ENTRIES FIRST.
WHEN TGHA IS RUN BY FIELD SERVICE, THE ONLY FUNCTIONS AVAILABLE ARE THE
DUMPING OF THE LIST FILES. NO CHANGES IN THE MEMORY CONFIGURATION OR
THE USE OF THE SPARE BITS CAN BE DONE BY TGHA IN USER MODE.
3.3 TGHA FILE REQUIREMENTS
THE TGHA FILE SYSTEM FOR BOTH TOPS10 AND TOPS20 INCLUDES THE FOLLOWING
FILES:
FILE PURPOSE CREATION MECHANISM
---- ------- -------------------
TGHA.EXE TGHA EXECUTION FILE MONITOR TAPE
TGHA.HLP TGHA HELP FILE MONITOR TAPE
TGHA.DOC TGHA OVERVIEW FILE MONITOR TAPE
TGHA.DAT TGHA HISTORY/DATA FILE TGHA.EXE WHEN FIRST RUN
TGHA.TRA TGHA TRACE FILE TGHA.EXE WHEN FIRST RUN
TGHA.BAD BAD COPY OF TGHA.DAT TGHA.EXE WHEN THE DATABASE GETS
CONFUSED
WHEN THE MONITOR FIRST STARTS UP ON A SYSTEM RELOAD, IT LOOKS FOR TGHA
IN THE APPROPRIATE AREA. IF IT IS NOT THERE, TGHA CANNOT BE RUN BY THE
MONITOR. IF TGHA.EXE IS NOT FOUND, THE MONITOR WILL THROW AWAY THE
ERROR DATA THAT MIGHT HAVE BEEN COLLECTED BY TGHA. TGHA.DAT AND
TGHA.TRA WILL BE CREATED BY TGHA WHEN NECESSARY.
TGHA.DAT WILL NOT GROW AS ERRORS ARE INCURRED. TGHA WILL PURGE OLD DATA
AS ITS DATABASES FILL UP. THE ONLY WAY THAT TGHA.DAT WILL GROW IS IF
NEW HARDWARE IS BROUGHT ON LINE (IE. A STORAGE MODULE IS SWAPPED OR A
NEW CONTROLLER IS ADDED TO THE SYSTEM). TGHA REQUIRES ALL OF THE
HISTORY DATA THAT IT HAS COLLECTED TO REMAIN INTACT SO THAT IT CAN MAKE
CORRECT DECISIONS ABOUT CORRECTIVE ACTION.
Page 6
***TGHA.DAT SHOULD NOT BE DELETED. IT CONTAINS IMPORTANT HISTORY
INFORMATION.
3.4 INITIAL SYSTEM START UP
AT INITIAL SYSTEM START UP, THE MONITOR RUNS TGHA IN START UP MODE. IN
THIS MODE, TGHA WILL FIRST ENABLE SINGLE BIT ERROR REPORTING THROUGHOUT
MF20 MEMORY. TGHA WILL THEN EITHER BUILD THE HISTORY FILE IF IT DOES
NOT ALREADY EXIST, OR VERIFY THAT IT KNOWS ABOUT ALL OF THE MF20
HARDWARE THAT IS ON LINE. IF THE HISTORY FILE EXISTS AND NEW MF20
HARDWARE APPEARS, TGHA WILL ADD THIS NEW HARDWARE TO ITS HISTORY FILE.
ONCE THE INITIAL START UP INITIALIZATION IS COMPLETE, TGHA THEN LOOKS
FOR ANY MF20 ERRORS THAT HAVE OCCURED SINCE THE INITIAL SYSTEM START UP.
3.5 MF20 ERRORS
FOR TOPS10, WHEN AN MF20 CORRECTABLE ERROR OCCURS, THE MONITOR WAKES UP
THE SYSTEM TGHA. FOR TOPS20, WHEN AN MF20 CORRECTABLE ERROR OCCURS, THE
MONITOR CALLS TGHA WITH A JOB NUMBER OF 0. TGHA THEN GETS THE DATA FOR
THE MF20 ERROR FROM THE MONITOR. THIS IS CALLED A CHRONOLOGICAL ERROR.
THEY ARE STORED IN ORDER OF OCCURRANCE IN THE HISTORY DATABASE. THERE
IS A SEPARATE CHRONOLOGICAL ERROR LIST FOR EACH MODULE.
ONLY WHEN THE CHRONOLOGICAL ERROR LIST IS FULL DOES TGHA GO OFF AND
ATTEMPT TO RESOLVE A KNOWN ERROR FROM THIS LIST.
3.6 KNOWN ERROR DETERMINATION
KNOWN ERROR DETERMINATION IS DONE STATISTICALLY USING THE CHI SQUARED
GOODNESS OF FIT FORMULA. STARTING WITH THE FIRST CHRONOLOGICAL ERROR IN
THE STORAGE MODULE, THE FORMULA IS USED TO TEST THE DISTRIBUTION OF THE
OTHER ERRORS IN THE CHRONOLOGICAL ERROR LIST FOR EACH TYPE OF POSSIBLE
HARDWARE ERROR.
FOR INSTANCE, THE FIRST HARDWARE FAILURE CONSIDERED IS A FULL MUX
FAILURE. A TABLE IS BUILT BY SCANNING ALL OF THE OTHER ERRORS IN THE
STORAGE MODULE AND TALLYING ONLY THOSE ERRORS THAT ARE COVERED BY THE
SPECIFIC TYPE OF HARDWARE ERROR BEING CONSIDERED, IN THIS CASE, A FULL
MUX FAILURE. THE TABLE FOR A FULL MUX FAILURE IS DISTRIBUTED BY BLOCK
AND SUBBLOCK NUMBER. THIS TRANSLATES INTO A DISTRIBUTION BY CHIP SINCE
THERE ARE 16 CHIPS INVOLVED IN A FULL MUX FAILURE. IF THE THE
DISTRIBUTION IS EVEN ENOUGH, THE CHI 2'D GOODNESS OF FIT TEST WILL
SUCCEED, AND A KNOWN ERROR IS DECLARED.
THIS CONTINUES UNTIL EITHER THE HARDWARE ERROR TYPE IS FOUND OR ALL OF
THE ERROR TYPES HAVE BEEN TRIED.
THE LAST TYPE OF HARDWARE ERROR CONSIDERED IS A CELL ERROR. IF THERE
ARE MORE THAN A MINIMUM NUMBER OF ERRORS OF THE SAME CELL (TYPICALLY 5),
THEN A CELL ERROR IS DECLARED.
Page 7
AFTER AN ERROR TYPE HAS BEEN FOUND, ALL OF THE CHRONOLOGICAL ERRORS THAT
ARE INFLUENCED BY THE KNOWN ERROR TYPE ARE ELIMINATED FROM THE
CHRONOLOGICAL ERROR LIST. THIS PROCEDURE CONTINUES FOR EACH REMAINING
ERROR IN THE CHRONOLOGICAL ERROR LIST.
AFTER THE KNOWN ERROR ROUTINE HAS BEEN RUN, THE CORRECTIVE ACTION
ROUTINE IS CALLED.
3.7 CORRECTIVE ACTION
THE GOAL OF THE CORRECTIVE ACTION ROUTINES IS TO DETERMINE THE OPTIMUM
CORRECTIVE ACTION GIVEN THE CURRENT KNOWN ERRORS. THIS IS DONE ON A PER
GROUP BASIS.
FIRST, THE WORST ERROR IN THE GROUP IS FOUND. THIS IS THE ERROR THAT
AFFECT THE MOST AMOUNT OF MEMORY. ONCE THE WORST ERROR HAS BEEN FOUND,
THE APPLICABLE SPARE BITS ARE USED TO COVER IT. IF THERE IS MORE THAN
ONE KNOWN ERROR IN THE GROUP, A SCAN IS DONE TO SET ALL ICE (INHIBIT
CORRECTABLE ERROR REPORTING) BITS WITHIN THE SCOPE OF THE WORST ERROR.
THIS PROCEDURE IS REPEATED UNTIL ALL KNOWN ERRORS FOR THE GROUP HAVE
BEEN ANALYZED.
NOTE: THERE IS ONE EXCEPTION TO THE PREVIOUS PROCEDURE. IF THE MEMORY
BOOT PORTION OF KLI HAS (DURING THE DBE SCAN - DOUBLE BIT ERROR) USED A
SPARE BIT TO HANDLE A DOUBLE BIT ERROR, TGHA WILL NOT CHANGE THE USE OF
THAT SPECIFIC SPARE BIT. IF TGHA WERE TO ATTEMPT TO USE SUCH A SPARE
BIT FOR ANOTHER PURPOSE, THE POSSIBILITY OF PARITY ERRORS RESULTING
WOULD BE TOO GREAT TO RISK.
ALSO, IF A PARITY ERROR OCCURS DURING THE SETTING OF A SPARE BIT, TGHA
WILL SET THE SAME CONSIDERATIONS ON THAT SPARE BIT AS IF KLI HAD SET IT.
ON THE OCCURRENCE OF THE PREVIOUS CONDITIONS, APPROPRIATE ERROR.SYS
ENTRIES WILL BE MADE.
3.8 PARITY ERRORS
PARITY ERRORS ARE HANDLED LIKE PARITY ERRORS ALWAYS HAVE BEEN. IE. IF
THE MONITOR SUCCESSFULLY CONTINUES AFTER THE PARITY ERROR, IT WILL
ATTEMPT TO TAKE THE PAGE OF MEMORY WITH THE PARITY ERROR OFFLINE. THE
MONITOR THEN RUNS TGHA IF THE PAGE WAS SUCCESSFULLY REMOVED. THE
MONITOR MAY NOT BE ABLE TO TAKE THE PAGE WITH THE PARITY ERROR OFF LINE
IF ,FOR INSTANCE, THE PAGE WAS PART OF THE RESIDENT MONITOR. TGHA WILL
ENTER PARITY ERRORS IN ITS TRACE FILE, NOT THE HISTORY FILE DATABASE.
TGHA WILL MAKE AN ENTRY IN THE TRACE FILE AND ERROR.SYS INDICATING
PHYSICAL ADDRESS, BLOCK OF MEMORY, AND THE ORIENTATION OF THE 4 STORAGE
MODULES CONTAINING THE ERROR.
Page 8
4.0 HISTORY FILE
THE HISTORY FILE IS MADE UP OF 3 DIFFERENT TYPES OF PAGES. THEY ARE THE
DIRECTORY PAGE, GROUP PAGES, AND STORAGE MODULE PAGES. THE SEQUENCE
STARTS OUT WITH ONE DIRECTORY PAGE. THIS PAGE CONTAINS CONFIGURATION
INFORMATION ABOUT ALL OF THE MF20'S ON THE SYSTEM. THE NEXT PAGE IS A
GROUP PAGE FOLLOWED BY FOUR STORAGE MODULE PAGES. THE PATTERN OF A
GROUP AND FOUR SUCCEEDING STORAGE MODULE PAGES IS REPEATED FOR EACH
GROUP CURRENTLY ON LINE IN MF20'S. A STORAGE MODULE PAGE FOR EACH
STORAGE MODULE THAT HAS GONE AWAY IS THEN LISTED.
4.1 DIRECTORY PAGE
THE FIRST PAGE OF THE HISTORY FILE CONTAINS THE CURRENT CONFIGURATION OF
THE MF20 MEMORY, INCLUDING THE LOCATION OF EACH STORAGE MODULE BY SERIAL
NUMBER. NOTE THAT THE SERIAL NUMBER IN THE HISTORY FILE IS IN THE SAME
FORMAT AS THE STICKER ON THE STORAGE MODULE ITSELF.
THE LOCATION OF THE STORAGE MODULES IN MF20 MEMORY IS DOCUMENTED IN THE
HISTORY LIST FILE FOR ARCHIVAL AND FIELD SERVICE REASONS. THE ERRORS
LOGGED IN THE HISTORY FILE ARE DEPENDENT UPON THE ORIENTATION OF THE
STORAGE MODULES. IF THEY ARE MOVED, THE ERROR CORRECTIONS IN THE MF20
MAY HAVE DIFFERENT CHARACTERISTICS. IF FIELD SERVICE WISHES TO REPLACE
A STORAGE MODULE, THE EXACT LOCATION BY SERIAL NUMBER IS THEREFORE
DOCUMENTED.
4.2 GROUP PAGE
THE GROUP PAGE CONTAINS INFORMATION SPECIFIC TO THE CONFIGURATION OF THE
GROUP. THE LAST TIME THE GROUP WAS USED IN THIS CONFIGURATION MAY BE
USEFUL IF GROUPS OF STORAGE MODULES ARE REMOVED AND PUT BACK INTO THE
MEMORY LATER. THE TIME OF THE LAST CHANGE TO THE GROUP IS USEFUL WHEN
DETERMINING WHERE THE MOST RECENT ERRORS HAVE OCCURED IN THE MF20.
LET IT BE NOTED HERE THAT THE CHRONOLOGICAL LIST OF ERRORS IN THE
HISTORY FILE AND THE ENTRIES IN THE TRACE FILE ARE TAGGED WITH THE DATE
AND TIME. THIS IS REQUIRED WHEN CORRELATING CORRECTIVE ACTION WITH
MF20.
THE GROUP PAGE ALSO CONTAINS A MAP OF THE CURRENT STATE OF THE BIT
SUBSTITUTION RAM. THE BIT SUBSTITUTION RAM IS USED TO DIRECT BIT
REPLACEMENT WITH THE SPARE BIT RAMS.
THE VALUE OF THE SPARE BIT RAM CONTAINS 4 FIELDS THAT HAVE THE FOLLOWING
MEANING:
FIELD BITS MEANING
400 EITHER KLI USED THIS SPARE BIT TO COVER A PARITY
ERROR OR A PARITY OCCURED DURING A BIT SWAP.
THE 400 BIT IS IN THE TGHA DATABASE ONLY, AND IS
NOT USED IN THE ACTUAL BITSWAP PROCESS.
374 THE OCTAL VALUE OF THE BIT BEING REPLACED *
Page 9
2 SET - CORRECTABLE ERROR REPORTING IS DISABLED
CLEARED - CORRECTABLE ERROR REPORTING IS ENABLED
1 BIT SUB RAM VALUE PARITY BIT
* TO FIND THE VALUE OF THE BIT BEING REPLACED, THIS FIELD MUST
BE SHIFTED TO THE RIGHT BY 2 PLACES, AND THE DECIMAL
EQUIVALENT VALUE OF THE OCTAL DIGITS IS THE WORD BIT BEING
REPLACED.
BIT SUBSTITUTION RAM EXAMPLES:
-----------------------------
VALUE MEANING
256 - THE SPARE BIT IS NOT IN USE. THE VALUE OF
THE BIT TO BE SWAPPED POINTS TO THE SPARE BIT.
THE SPARE BIT IS IN DECIMAL BIT POSITION 43
- CORRECTABLE ERROR REPORTING IS DISABLED FOR
THIS BIT SUB RAM ADDRESS
255 - SPARE BIT IS NOT IN USE (SAME AS PREVIOUS EXAMPLE)
- CORRECTABLE ERROR REPORTING IS ENABLED
652 - THE 400 BIT ON INDICATES THAT THIS USE OF
THIS SPARE BIT PREVENTS A PARITY ERROR
FROM OCCURING. TGHA WILL NOT CHANGE THIS
VALUE. THE BITSUB RAM DATA REQUIRES ODD
PARITY.
- BIT 42 OF THE MF20 DATA PATH IS BEING SWAPPED
(THE PARITY BIT).
THE ACTUAL SPARE BIT ADDRESS IN SBUS DIAG FUNCTION 7 IS 7 BITS WIDE.
THE ADDRESS IN THE GROUP PAGE ACCOUNTS FOR ONLY THE LOWER 5 BITS. THE
HIGH ORDER 2 BITS IS THE GROUP NUMBER WITHIN THE CONTROLLER. THE GROUP
FIELD IS DETERMINED FROM THE GROUP POSITION IN THE MF20 CONTROLLER, AND
NOT THE GROUP DATABASE. THEREFORE THE GROUP PORTION OF THE FIELD IS
NOT, AND INDEED CANNOT BE KEPT WITHIN THE SPARE BIT RAM ADDRESS TABLE.
THIS PORTION OF THE SBUS DIAG FUNCTION IS FILLED IN AT EXECUTION TIME.
4.3 STORAGE MODULE PAGE
A STORAGE MODULE PAGE CONTAINS INFORMATION RELATING SPECIFICLY TO THE
STORAGE MODULE. THE SERIAL NUMBER APPEARS IN THE SAME FORMAT AS THE
STICKER ON THE MODULE ITSELF. THIS IS CRITICAL FOR FIELD RETURNS, WHICH
REQUIRE THAT THIS PAGE BE ATTACHED TO THE STORAGE MODULE IF IT IS TO BE
RETURNED TO THE FACTORY FOR REPAIRS. THE LAST TIME THE MODULE WAS USED
CAN BE USEFUL IN TRACKING MODULE SWAPS MADE IN THE MF20 AND FIELD
RETURNS.
THE KNOWN ERROR LIST CONTAINS THE ERROR INFORMATION RELATED TO THE KNOWN
ERRORS RESOLVED FROM CHRONOLOGICAL ERRORS BY TGHA. MOST OF THE
INFORMATION IN EACH ENTRY IS USED BY TGHA TO DETERMINE OVERLAP OF
ERRORS. THIS OVERLAP IS IMPORTANT WHEN CONSIDERING THE OPTIMUM USE OF
THE SPARE BITS.
Page 10
THE STORAGE MODULE CHIP LOCATION IS INCLUDED TO FACILITATE MODULE
REPAIR.
THE TIME OF RESOLUTION OF THE ERROR MAY GIVE AN INDICATION OF THE RATE
OF DETERIORATION OF FAILURE OF THE MODULE.
THE CURRENT CHRONOLOGICAL ERROR LIST FOR THE STORAGE MODULE WILL INCLUDE
ANY ERRORS THAT HAVE NOT BEEN RESOLVED INTO KNOWN ERRORS YET. THERE MAY
BE INCLUDED IN THIS LIST, ANY SOFT ERRORS THAT HAVE OCCURED. THESE SOFT
ERRORS WILL BE REMOVED FROM THE LIST SOMETIME AFTER A MINIMUM DEFAULT
(ABOUT 1 WEEK).
4.4 UNUSED STORAGE MODULES
ANY STORAGE MODULES THAT ARE IN THE TGHA HISTORY FILE, YET ARE NOT
PRESENTLY IN USE THEN FOLLOW IN A SIMILAR MANNER. THE INCLUSION OF THE
UNUSED STORAGE MODULES IN THE HISTORY FILE IS REQUIRED FOR THE CASE WHEN
FIELD SERVICE HAS REPLACED A STORAGE MODULE, RELOADED THE SYSTEM, AND
WANTS A COPY OF THE HISTORY FILE FOR THAT STORAGE MODULE SO THAT THE
ERROR DATA CAN BE INCLUDED WITH THE MODULE WHEN THE MODULE IS RETURNED
FOR REPAIRS.
SINCE THESE STORAGE MODULES ARE NO LONGER PART OF ANY CONFIGURED GROUP,
NO GROUP INFORMATION IS GIVEN.
5.0 TRACE FILE
THE TRACE FILE CONTAINS DATED ENTRIES INDICATING WHAT CORRECTIVE ACTION
HAS BEEN TAKEN BY TGHA. IF THE TGHA SOFTWARE RUNS INTO CONFUSION OR A
DATABASE GETS FULL, ENTRIES INTO THE TRACE FILE WILL ALSO REFLECT THE
DIFFICULTY. SERIOUS MF20 HARDWARE ERRORS WILL ALSO BE ENTERED IN THE
TRACE FILE.
5.1 SPARE BIT SWAP
11-Oct-82 15:50:12 - IN CONTROLLER 10, GROUP 0, VECTOR 3,THE SPARE BIT IS GOING TO BE CHANGED TO 124 (OCTAL).
ONCE THE SPARE BIT SWAP HAS COMPLETED, THE RESULTS OF THE BITSWAP ARE INDICATED.
11-Oct-82 15:50:12 - BIT SUBSTITUTION - EVERYTHING COMPLETED NORMALLY
WHEN CORRECTABLE ERROR REPORTING IS DISABLED FOR A SPARE BIT SUBSTITUTION RAM ADDRESS, THE FOLLOWING ENTRY IS MADE:
11-Oct-82 15:50:13 - IN CONTROLLER 10, GROUP 0, VECTOR 3, THE ICE BIT IS GOING TO BE SET.
Page 11
5.2 PARITY ERROR
**********
11-OCT-82 19:20:02
* PARITY ERROR AT ADDRESS 1001475, BLOCK 0
* STORAGE MODULE SERIAL NUMBERS BY FIELD: 0 = 7460209 1 = 7460108 2 = 7460117 3 = 7460200
**********
5.3 SERIOUS MF20 HARDWARE FAILURES
************************************************************************
*11-OCT-82 19:20:02
* TGHA HAS TEMPORARILY CORRECTED A SERIOUS MOS MEMORY FAILURE.
*** CALL FIELD SERVICE TO REPORT THIS CONDITION ***
************************************************************************
THIS MESSAGE ALSO APPEARS ON THE CTY. ALTHOUGH THE ERROR HAS BEEN
CORRECTED USING THE SPARE BITS, THE PROBABILITY OF A PARITY ERROR DUE TO
FURTHER HARDWARE DEGRADATION HAS RISEN TO AN UNCOMFORTABLE STATE. THE
MF20 DIAGNOSTICS SHOULD BE RUN AND THE OFFENDING STORAGE MODULE
REPLACED.
5.4 USE OF SPARE BIT BY KLI
*****
5-AUG-79 13:20:29
THE MEMORY BOOT IN KLI HAS USED THE SPARE BIT TO PREVENT A PARITY ERROR.
THIS CONDITION SHOULD BE CORRECTED AS SOON AS POSSIBLE.
CONTROLLER GROUP BLOCK WORD (BITS 33-35)
10 0 1 5
*****
5.5 KLI DETECTED BAD BLOCK
IF KLI HAS NOT CONFIGURED A BLOCK OF MF20 MEMORY BECAUSE OF AN ERROR
THAT CANNOT BE REPAIRED USING THE SPARE BIT, THE BLOCK WILL NOT BE
CONFIGURED. THIS MESSAGE ALSO APPEARS ON THE CTY. AN ENTRY WILL BE
MADE IN THE TRACE FILE AS FOLLOWS:
************************************************************************
*5-AUG-79 13:40:51
* THE FOLLOWING BLOCKS ARE MARKED AS BAD
* AND ARE NOT ON LINE:
* CONTROLLER GROUP BLOCK
* 10 1 3
* 10 2 1
* THIS CONSISTS OF 128K OF MEMORY THAT IS OFF LINE.
*** CALL FIELD SERVICE TO REPORT THIS CONDITION ***
************************************************************************
Page 12
5.6 MISSING DATABASE
IF TGHA FINDS THAT THERE IS NO DATABASE AND THAT AN MF20 CONTROLLER WAS
LAST INITIALIZED BY TGHA, TGHA WILL OUTPUT THE FOLLOWING MESSAGE:
*****
15-AUG-1979 15:10:20
* MF20 CONTROLLER 10 WAS LAST INITIALIZED BY TGHA AND
* TGHA HAS LOST ITS DATABASE. REFER TO TRACE OR SPEAR
* DOCUMENTATION IN TGHA.DOC FOR FURTHER DETAILS AND RECOVERY INSTRUCTIONS.
*****
IN THIS SITUATION, TGHA CANNOT DETERMINE WHETHER ANY SPARE BITS IN USE
WERE SET BY THE MEMORY BOOT IN KLI OR TGHA. THIS INFORMATION WAS STORED
IN THE TGHA DATABASE, WHICH WENT AWAY FOR SOME REASON. TGHA MUST
THEREFORE ASSUME THAT ALL SPARE BITS IN USE WERE SET BY KLI TO PREVENT
PARITY ERRORS. THIS MEANS THAT THESE SPARE BITS WILL NOT BE AVAILABLE
FOR CORRECTION OF FURTHER HARDWARE FAILURES IN THE MF20 MEMORY. THIS IS
NOT AN IDEAL SITUATION FOR TGHA TO BE IN, AND WILL NOT RECTIFY ITS SELF
AUTOMATICALLY. ALTHOUGH THIS SITUATION IS NOT IMMEDIATELY CRITICAL, IT
DEGRADES THE PERFORMANCE OF TGHA. THE RECOVERY PROCEDURE CAN BE
DEFERRED TO A CONVENIENT TIME, LIKE A FIELD SERVICE PM OR A SCHEDULED
SYSTEM RELOAD.
5.6.1 RECOVERY PROCEDURE
1. DELETE TGHA.DAT FROM PS:<SYSTEM>
2. BRING THE SYSTEM DOWN
3. RECONFIGURE THE MEMORY SYSTEM BY USING THE 'FORCE' OPTION IN
THE MEMORY CONFIGURATION PORTION OF KLI.
4. REBOOT THE SYSTEM
THIS IS THE -ONLY- TIME THAT THE TGHA DATABASE FILE (TGHA.DAT) SHOULD BE
EXPLICITELY DELETED. ALTHOUGH THIS DESTROYS ALL OF THE KNOWN ERROR
HISTORY, THE KNOWN ERRORS WILL BE RESOLVED BY TGHA IN THE SAME MANNER AS
BEFORE.
THIS IS THE ONLY TIME THAT THE TGHA DATABASE FILE (TGHA.DAT) SHOULD BE
EXPLICITELY DELETED. ALTHOUGH THIS DESTROYS ALL OF THE KNOWN ERROR
HISTORY, THE KNOWN ERRORS WILL BE RESOLVED BY TGHA IN THE SAME MANNER AS
BEFORE. THE PROBABILITY OF THIS SITUATION ARISIG IS VERY LOW. THIS
DRASTIC ACTION IS EXPLAINED ONLY TO COVER THIS SITUATION.
5.7 SOFTWARE STATE WORD
THE SOFTWARE STATE OF THE MEMORY INDICATES WHETHER KLI OR TGHA WAS THE
LAST PROGRAM TO MAKE ANY CHANGES IN THE BIT SUB RAM. ALTHOUGH THE
SOFTWARE BITS IN SBUS FUNCTION WERE DESIGNED FOR THIS PURPOSE, CONFLICTS
MADE THEM IMPRACTICAL. THE INFORMATION IS FLAGGED IN A WORD IN THE
Page 13
BITSUB RAM INSTEAD OF USING THE SOFTWARE STATE BITS. IF, WHEN TGHA GOES
TO CHECK THE STATE OF THE SOFTWARE STATE WORD IN THE BITSUB RAM, TGHA
FINDS THAT THE DATA IS NOT OF THE FORMAT THAT EITHER KLI OR TGHA USES,
THE FOLLOWING MESSAGE WILL BE PUT OUT:
*****
17-AUG-1979 13:02:59
IN CONTROLLER 10 THE BITSUB RAM STATE WORD WAS GARBAGE.
ASSUMIMG A DBE SCAN WAS DONE LAST. SEE TGHA.DOC FOR FURTHER DETAILS.
*****
THIS IS A VERY STRANGE STATE SINCE THE SPARE BIT SUB RAM ADDRESS USED
FOR THE SOFTWARE STATE WORD IS NOT IN THE ADDRESS SPACE USED FOR
SWAPPING OF MEMORY BITS. IT IS SUGGESTED THAT THE MF20 DIAGNOSTICS BE
RUN AT THE NEXT CONVENIENT TIME.
6.0 SPEAR
ERROR.SYS ENTRIES MADE BY TGHA HAVE THE FOLLOWING FORMAT:
6.1 NEW KNOWN ERROR:
A NEW MF20 KNOWN ERROR HAS BEEN DECLARED. DATA: STORAGE MODULE SERIAL
NUMBER: 838009 BLOCK: 1, SUBBLOCK: 0, BIT IN FIELD (10): 11, ROW:
71, COLUMN: 24, E NUMBER 192, ERROR TYPE: CHIP
6.2 SERIOUS MF20 HARDWARE FAILURE
************************************************************************
*11-OCT-82 19:20:02
* TGHA HAS TEMPORARILY CORRECTED A SERIOUS MOS MEMORY FAILURE.
*** CALL FIELD SERVICE TO REPORT THIS CONDITION ***
************************************************************************
ALTHOUGH THE ERROR HAS BEEN CORRECTED USING THE SPARE BITS, THE
PROBABILITY OF A PARITY ERROR DUE TO FURTHER HARDWARE DEGRADATION HAS
RISEN TO AN UNCOMFORTABLE STATE. THE MF20 DIAGNOSTICS SHOULD BE RUN AND
THE OFFENDING STORAGE MODULE REPLACED.
6.3 USE OF SPARE BIT BY KLI
*****
5-AUG-83 13:20:29
THE MEMORY BOOT IN KLI HAS USED THE SPARE BIT TO PREVENT A PARITY ERROR.
THIS CONDITION SHOULD BE CORRECTED AS SOON AS POSSIBLE.
CONTROLLER GROUP BLOCK WORD (BITS 33-35)
10 0 1 5
*****
Page 14
6.4 KLI DETECTED BAD BLOCK
IF KLI HAS NOT CONFIGURED A BLOCK OF MF20 MEMORY BECAUSE OF AN ERROR
THAT CANNOT BE REPAIRED USING THE SPARE BIT, THE BLOCK WILL NOT BE
CONFIGURED. AN ENTRY WILL BE MADE IN ERROR.SYS AS FOLLOWS:
************************************************************************
*5-AUG-83 13:40:51
* THE FOLLOWING BLOCKS ARE MARKED AS BAD
* AND ARE NOT ON LINE:
* CONTROLLER GROUP BLOCK
* 10 1 3
* 10 2 1
* THIS CONSISTS OF 128K OF MEMORY THAT IS OFF LINE.
*** CALL FIELD SERVICE TO REPORT THIS CONDITION ***
**********************************************************************
6.5 MISSING DATABASE
IF TGHA FINDS THAT THERE IS NO DATABASE AND THAT AN MF20 CONTROLLER WAS
LAST INITIALIZED BY TGHA, TGHA WILL OUTPUT THE FOLLOWING MESSAGE:
*****
15-AUG-1983 15:10:20
* MF20 CONTROLLER 10 WAS LAST INITIALIZED BY TGHA AND
* TGHA HAS LOST ITS DATABASE. REFER TO TRACE OR SPEAR
* DOCUMENTATION IN TGHA.DOC FOR FURTHER DETAILS AND RECOVERY INSTRUCTIONS.
*****
FOR FURTHER EXPLANATION REFER TO MISSING DATABASE SECTION UNDER TRACE
FILE ENTRIES.
6.6 SOFTWARE STATE WORD
THE SOFTWARE STATE OF THE MEMORY INDICATES WHETHER KLI OR TGHA WAS THE
LAST PROGRAM TO MAKE ANY CHANGES IN THE BIT SUB RAM. ALTHOUGH THE
SOFTWARE BITS IN SBUS FUNCTION WERE DESIGNED FOR THIS PURPOSE, CONFLICTS
MADE THEM IMPRACTICAL. THE INFORMATION IS FLAGGED IN A WORD IN THE
BITSUB RAM INSTEAD OF USING THE SOFTWARE STATE BITS. IF, WHEN TGHA GOES
TO CHECK THE STATE OF THE SOFTWARE STATE WORD IN THE BITSUB RAM, TGHA
FINDS THAT THE DATA IS NOT OF THE FORMAT THAT EITHER KLI OR TGHA USES,
THE FOLLOWING MESSAGE WILL BE PUT OUT:
*****
17-AUG-1983 13:02:59
IN CONTROLLER 10 THE BITSUB RAM STATE WORD WAS GARBAGE.
ASSUMIMG A DBE SCAN WAS DONE LAST. SEE TGHA.DOC FOR FURTHER DETAILS.
*****
THIS IS A VERY STRANGE STATE SINCE THE SPARE BIT SUB RAM ADDRESS USED
FOR THE SOFTWARE STATE WORD IS NOT IN THE ADDRESS SPACE USED FOR
SWAPPING OF MEMORY BITS. IT IS SUGGESTED THAT THE MF20 DIAGNOSTICS BE
RUN AT THE NEXT CONVENIENT TIME.
Page 15
7.0 MF20 MEMORY CONTROLLER SBUS DIAGNOSTIC FUNCTIONS
FUNCTION 00 00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E) 1* < CONTROLLER NUMBER > !CLEAR* *
TO MEM 2* 0 ! 1 ! ? * ? ! ? ! 0-5 * ! ! * ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E) 3* ! < FUNCTION NUMBER > *
TO MEM 4* ! ! * ! ! * ! ! * ! ! * ! 0 ! 0 * 0 ! 0 ! 0 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E+1) 5*CONTR!CORR !INCMP* <PARITY ERRORS> *INTERLEAVE ! <LAST WORD REQUESTS > *<LAST TYPE>! < ERROR ADDRESS *
FROM MEM 6* ERR ! ERR !CYCLE*READ !WRITE! ADR * 1 ! 1 ! RQ0 * RQ1 ! RQ2 ! RQ3 * RD ! WR ! 14 * 15 ! 16 ! 17 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E+1) 7* ERROR ADDRESS (CONTINUED) > *
FROM MEM 8* 18 ! 19 ! 20 * 21 ! 22 ! 23 * 24 ! 25 ! 26 * 27 ! 28 ! 29 * 30 ! 31 ! 32 * 33 ! 34 ! 35 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
FUNCTION 01 00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E) 1* < CONTROLLER NUMBER > ! *<GROUP LOOPBACK >* *
TO MEM 2* 0 ! 1 ! ? * ? ! ? ! * ! ! * ! ! *ENABL! GN2 ! GN1 * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E) 3* ! < STATUS > !LD EN! ! < FUNCTION NUMBER > *
TO MEM 4* ! ! * ! ! * !DSABL! SF2 * SF1 !25-27! * ! 0 ! 0 * 0 ! 0 ! 1 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E+1) 5* ! <MEM CONTROLLER TYPE> * <LPBK GRP SEL> * *
FROM MEM 6* ! ! * ! ! * ! ! 0 * 1 ! 0 ! 1 * 0 ! 1 ! 2 * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E+1) 7* ! < STATUS > ! *
FROM MEM 8* ! ! * ! ! * !DSABL! SF2 * SF1 ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
SF2 & SF1 ARE SOFTWARE FLAGS DO NOT EFFECT THE HARDWARE, NOR DOES THE HARDWARE AFFECT THEM EXCEPT TO CLEAR THEM ON POWER-UP. THEIR
COMBINED MEANINGS ARE: 0/CONTROLLER WAS JUST POWERED UP; 1/ALL RAMS BUT ADRESP ARE LOADED, CRUDE PATCHING IS DONE AND BITSUB RAM
SAYS WHICH BLOCKS MAY BE USED; 2/SAME AS 1 BUT CONTROLLER IS CONFIGURED; 3/SAME AS 2 EXCEPT THAT TGHA HAS FINISHED ITS WORK.
Page 16
FUNCTION 02 00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E) 1* < CONTROLLER NUMBER > ! * < SM PROM SELECT > !<PROM BYTE>* *
TO MEM 2* 0 ! 1 ! ? * ? ! ? ! * ! ! * GN2 ! GN1 ! FN2 * FN1 ! 2 ! 1 * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E) 3* ! < DIAG MIXER INPUT SELECT > ! ! < FUNCTION NUMBER > *
TO MEM 4* ! ! * ! ! 2 * 3 ! 4 ! 5 * 6 ! ! * ! 0 ! 0 * 0 ! 1 ! 0 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E+1) 5* !<ERROR WD#>! < SELECTED DIAG DATA OR PROM DATA > * *
FROM MEM 6* ! ! * ! ! 2 * 1 ! 0 ! 1 * 2 ! 3 ! 4 * 5 ! 6 ! 7 * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E+1) 7* ! < MOS ADDRESS (SINGLE STEP ONLY) > !<CTL RAM PAR ERR>! *
FROM MEM 8* ! ! 0 * 1 ! 2 ! 3 * 4 ! 5 ! 6 * 7 ! TIM ! SUB * ADR ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
NOTE: BITS 23-27 ARE ONLY USED DURING SINGLE STEP DIAGNOSTIC TESTS. THE LEFTMOST 1 BIT SELECTS THE CORRESPONDING MIXER INPUT
FOR READING DIAGNOSTIC SIGNALS. IF 23-27 ARE 0 THEN THE DATA RETURNED IS THE SELECTED SM PROM WORD. NOTE ALSO THAT
BITS 26-27 ARE USED TO SELECT THE ADDRESS WHICH IS SENT BACK IN THE MOS ADDRESS FIELD.
FUNCTION 03 00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E) 1* < CONTROLLER NUMBER > ! !FAST * !<FIXED VAL LOGIC RAMS >!LD EN*LD EN! *
TO MEM 2* 0 ! 1 ! ? * ? ! ? ! * ! ! BUS * !ACKN !D VLD*RDA34!RDA35! 10 *11-13! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E) 3* ! < FIXED VALUE LOGIC RAM LOAD ADDRESS > ! ! < FUNCTION NUMBER > *
TO MEM 4* ! ! 0 * 1 ! 2 ! 3 * 4 ! 5 ! 6 * 7 ! ! * ! 0 ! 0 * 0 ! 1 ! 1 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E+1) 5* !FAST * !<FIXED VAL LOGIC RAMS >! *
FROM MEM 6* ! ! * ! ! * ! ! BUS * !ACKN !D VLD*RDA34!RDA35! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E+1) 7* *
FROM MEM 8* ! ! * ! ! * ! ! * ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
NOTES: THE ACKN RAM (BIT 10) USES ADR BITS 1-7 ONLY.
THE FAST-BUS BIT IS DISCONNECTED INTERNALLY AND IS LEFT IN ONLY FOR DIAGNOSTIC COMPATABILITY.
Page 17
FUNCTION 04 00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E) 1* < CONTROLLER NUMBER > !PORT * * < SINGLE STEP CLOCK CONTROL *
TO MEM 2* 0 ! 1 ! ? * ? ! ? !LPBK * ! ! *A COM!B COM!D VLD* ! !SSCLK*SSMOD! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E) 3* SS CLK CTL > *REFR !LD EN! * < REFRESH INTERVAL (GATED CLOCK/16) > ! < FUNCTION NUMBER > *
TO MEM 4* ! !RNOW *ALLOW!24-30! * 1 ! 2 ! 3 * 4 ! 5 ! 6 * 7 ! 0 ! 0 * 1 ! 0 ! 0 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E+1) 5* * < SS CLOCK CONTROL ECHO > ! *
FROM MEM 6* ! ! * ! ! * ! ! *A COM!B COM! * ! ! *SSMOD! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E+1) 7* *REFR ! ! < REFRESH INTERVAL COUNTER > ! *
FROM MEM 8* ! ! *ALLOW! ! 0 * 1 ! 2 ! 3 * 4 ! 5 ! 6 * 7 ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
NOTE: IF PORT LOOPBACK IS SET (BIT 5) THE CONTROLLER WILL JUST ECHO ANYTHING SENT OUT BY AN SBUS DIAG AS LONG AS THE PROPER
CONTROLLER NUMBER IS GIVEN. ONLY AN SBUS RESET CAN CLEAR THIS CONDITION.
FUNCTION 05 00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E) 1* < CONTROLLER NUMBER > ! * < SINGLE STEP SIMULATED REQUEST BITS > ! *LD EN! *
TO MEM 2* 0 ! 1 ! ? * ? ! ? ! *ST A !ST B ! RQ0 * RQ1 ! RQ2 ! RQ3 * RD ! WR ! * SS ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E) 3* ! < SINGLE STEP REQUEST SIMULATED ADDRESS BITS > ! < FUNCTION NUMBER > *
TO MEM 4* ! ! 14 * 15 ! 16 ! 17 * 18 ! 19 ! 20 * 21 ! ! 34 * 35 ! 0 ! 0 * 1 ! 0 ! 1 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E+1) 5* *
FROM MEM 6* ! ! * ! ! * ! ! * ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E+1) 7* * < SS MOS RAS > ! *
FROM MEM 8* ! ! * ! ! * 0 ! 1 ! 2 * 3 ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
Page 18
FUNCTION 06 00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E) 1* < CONTROLLER NUMBER > ! ! < DIAGNOSTIC DATA TO SY OR WP BOARD > *LD EN! *
TO MEM 2* 0 ! 1 ! ? * ? ! ? ! * ! 32 ! 16 * 8 ! 4 ! 2 * 1 !MISC !MISC *07-14! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E) 3* !<ECC DIAG SUBFCN>! ! < FUNCTION NUMBER > *
TO MEM 4* ! ! * ! ! * ! 4 ! 2 * 1 ! ! * ! 0 ! 0 * 1 ! 1 ! 0 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E+1) 5* ! < DIAGNOSTIC DATA FROM SY OR WP BOARD > * *
FROM MEM 6* ! ! * ! ! * ! 32 ! 16 * 8 ! 4 ! 2 * 1 !MISC !MISC * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E+1) 7* * < SS MOS CAS > ! *
FROM MEM 8* ! ! * ! ! * 0 ! 1 ! 2 * 3 ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
NOTE: SBDIAG FCN 6 DOES DIFFERENT THINGS DEPENDING ON THE SUBFUNCTION CODE IN BITS 25-27:
6.0 READ THE ECC REGISTER ON THE SYN BOARD.
6.1 READ THE SYNDROME BUFFER ON THE SYN BOARD. THIS FUNCTION IS ALSO NORMALLY USED BY MONITOR.
6.2 SEL DIAG BITS 07-13 IN PLACE OF MOS BITS 36-42, FORCE ZEROS ON 00-35, RUN THRU A CORRECTION PASS, AND RETURN 00-35.
*** PLEASE NOTE THAT FCN 6.2 RETURNS THE ENTIRE WORD, NOT JUST BITS 7-14. ***
6.3 UNUSED
6.4 WR ECC COMPLEMENT REG IF BIT 15 SET, THEN READ IT BACK.
6.5 WR ECC COMPLEMENT REG IF BIT 15 SET, THEN ENABLE IT TO BE SENT TO MEM IN PLACE OF D36-42 ON NEXT WR TO MEM.
6.6 READ D36-43 MIXER.
6.7 ENABLE LATCHING OF D36-43 MIXER AFTER NEXT WRITE.
7 8 9 10 11 12 13 14 7 8 9 10 11 12 13 14
+-----+-----*-----+-----+-----*-----+-----+-----* +-----+-----*-----+-----+-----*-----+-----+-----*
SUBFUNCTION ! < MEM BITS 36-43 LATCHED ON LAST RD ERROR > * SUBFUNCTION ! < ECC BITS TO BE COMPLEMENTED ON WR > ! PAR *
6.0 !ECC32!ECC16*ECC8 !ECC4 !ECC2 *ECC1 !E PAR!SPARE* 6.4 ! 32 ! 16 * 8 ! 4 ! 2 * 1 !E PAR! CTL *
+-----+-----*-----+-----+-----*-----+-----+-----* +-----+-----*-----+-----+-----*-----+-----+-----*
+-----+-----*-----+-----+-----*-----+-----+-----* +-----+-----*-----+-----+-----*-----+-----+-----*
SUBFUNCTION ! <SYNDROME LATCHED ON LAST RD ERR> !<ERR TYPE >* SUBFUNCTION ! <BITS 36-43 AS WRITTEN TO MEM AFTER FCN 6.7 > *
6.1 ! 32 ! 16 * 8 ! 4 ! 2 * 1 ! CORR! DBL * 6.6 !ECC32!ECC16*ECC8 !ECC4 !ECC2 *ECC1 !E PAR!SPARE*
+-----+-----*-----+-----+-----*-----+-----+-----* +-----+-----*-----+-----+-----*-----+-----+-----*
Page 19
FUNCTION 07 00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E) 1* < CONTROLLER NUMBER > ! ! < BIT SUBSTITUTION RAM DATA (BIT NUMBERS) > *LD EN! *
TO MEM 2* 0 ! 1 ! ? * ? ! ? ! * ! 32 ! 16 * 8 ! 4 ! 2 * 1 ! ICE ! PAR *07-14! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E) 3* * < BIT SUBSTITUTION RAM LOAD ADDRESS > ! ! < FUNCTION NUMBER > *
TO MEM 4* ! ! * GN2 ! GN1 ! BN2 * BN1 ! 33 ! 34 * 35 ! ! * ! 0 ! 0 * 1 ! 1 ! 1 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E+1) 5* ! < BIT SUBSTITUTION RAM DATA ECHO > * *
FROM MEM 6* ! ! * ! ! * ! 32 ! 16 * 8 ! 4 ! 2 * 1 ! ICE ! PAR * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E+1) 7* * < SS MOS WR EN > ! *
FROM MEM 8* ! ! * ! ! * 0 ! 1 ! 2 * 3 ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
"ICE" MEANS IGNORE CORRECTABLE ERROR (IE DON'T SET CORR ERR FLAG)
FUNCTION 10 00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E) 1* < CONTROLLER NUMBER > ! ! < VOLTAGE MARGINS > ! < V MARGIN ENABLES > *LD EN! *
TO MEM 2* 0 ! 1 ! ? * ? ! ? ! * !12.60!5.25 *-2.10!-5.46! 12 * 5 ! -2 !-5.2 *07-14! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E) 3* !CORR * CLR ! ! < FUNCTION NUMBER > *
TO MEM 4* ! ! * ! ! * ! ! DIS *DCBAD! ! * ! 0 ! 1 * 0 ! 0 ! 0 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E+1) 5* ! < VOLTAGE MARGINS > ! < V MARGINS ENABLED > * *
FROM MEM 6* ! ! * ! ! * !12.60!5.25 *-2.10!-5.46! 12 * 5 ! -2 !-5.2 * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E+1) 7* !CORR * DC ! *
FROM MEM 8* ! ! * ! ! * ! ! DIS * BAD ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
NOTE: THE VOLTAGES GIVEN ABOVE APPLY FOR ONES IN BITS 7-10 WHERE ENABLED BY BITS 11-14. THE VOLTAGES CORRESPONDING TO ZEROS
IN BITS 7-10 ARE 11.40, 4.75, -1.90, AND -4.94. MORE THAN ONE MARGIN MAY BE SET CONCURRENTLY.
Page 20
FUNCTION 11 00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E) 1* < CONTROLLER NUMBER > ! ! < TIMING RAM LOAD ADR... *
TO MEM 2* 0 ! 1 ! ? * ? ! ? ! * ! ! * ! ! * ! 0 ! 1 * 2 ! 3 ! 4 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E) 3* ...ADR > !LD EN* < TIMING RAM DATA > ! ! < FUNCTION NUMBER > *
TO MEM 4* 5 ! 6 !21-30* RAS ! CAS ! PAR *WR EN!2ND A!D RDY*B CLR! ! * ! 0 ! 1 * 0 ! 0 ! 1 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E+1) 5* *
FROM MEM 6* ! ! * ! ! * ! ! * ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E+1) 7* * < TIMING RAM DATA ECHO > ! ! 4K *-DSEL! *
FROM MEM 8* ! ! * RAS ! CAS ! PAR *WR EN!2ND A!D RDY*B CLR! ! EN *CYCEN! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
FUNCTION 12 00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E) 1* < CONTROLLER NUMBER > ! ! < ADDRESS RESPONSE RAM DATA > *LD EN! *
TO MEM 2* 0 ! 1 ! ? * ? ! ? ! * ! ! PAR *TYPE ! GN2 ! GN1 * BN2 ! BN1 !DESEL*08-14! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E) 3* ! < ADDRESS RESPONSE RAM LOAD ADDRESS > ! ! < FUNCTION NUMBER > *
TO MEM 4* ! ! 14 * 15 ! 16 ! 17 * 18 ! 19 ! 20 * 21 ! ! * ! 0 ! 1 * 0 ! 1 ! 0 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
LH(E+1) 5* ! < ADDRESS RESPONSE RAM ECHO > * *
FROM MEM 6* ! ! * ! ! * ! ! PAR *TYPE ! GN2 ! GN1 * BN2 ! BN1 !DESEL* ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
RH(E+1) 7* *
FROM MEM 8* ! ! * ! ! * ! ! * ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
NOTE: "DESEL" (BIT 14) IS "BOX DESELECTED ON 1".
Page 21
*** MF20 STORAGE ORGANIZATION ***
MF20 STORAGE IS COMPOSED OF GROUPS, FIELDS, BLOCKS, & SUBBLOCKS. KL10 BIT NUMBER VS. "FIELD" LOCATION
A GROUP IS A SET OF FOUR STORAGE BOARDS. A FIELD IS ONE BOARD (11 BITS) ----------------------------------------
WIDE. 4 FIELDS USED IN PARALLEL GIVE A 44-BIT WORD: 36 DATA BITS, 6 ECC
(ERROR CHECK & CORRECT) BITS, ONE 43-BIT PARITY BIT, AND ONE SPARE BIT. KL10 BIT BIT IN SYNDROME
A BIT'S GROUP AND FIELD NUMBERS UNIQUELY PICK THE BIT'S STORAGE BOARD. BIT USE FIELD FIELD (NOTE 1)
A GROUP HAS 1 TO 4 BLOCKS (USUALLY 4) NUMBERED FROM 0. A BLOCK'S SIZE ---- ------ ----- ------ --------
IS 4 TIMES RAM SIZE. ITS START ADDRESS IS PROGRAMMABLE (SBDIAG FCN 12). 00 DATA 0 02 014
EACH BLOCK HAS 4 SUBBLOCKS. SUBBLOCKS ARE THE BASIC MF20 STORAGE UNIT, 01 DATA 1 02 024
ARE 1 WORD (44 BITS) WIDE, AND ARE SELECTED BY ADDRESS 34-35. EVERY 4TH 02 DATA 0 03 030
WORD OF A BLOCK'S MEMORY BELONGS TO A GIVEN SUBBLOCK. THERE IS 1 SPARE 03 DATA 1 03 034
BIT RAM PER SUBBLOCK. DURING A MEMORY CYCLE ALL OF THE BLOCK'S FIELDS 04 DATA 0 04 044
AND SUBBLOCKS ARE ALWAYS CYCLED, WITH THE 4 XBUS REQUEST BITS SPECIFYING 05 DATA 1 04 050
WHICH SUBBLOCK'S WORDS ARE ACTUALLY TO BE USED. 06 DATA 0 05 054
07 DATA 1 05 060
TABLE OF SLOT NUMBERS VS. GROUP, FIELD, AND BIT. 08 DATA 0 06 064
-------------------------------------------------------- 09 DATA 1 06 070
10 DATA 0 07 074
SLOT GROUP FIELD BITS 11 DATA 1 07 104
---- ----- ----- -------------------------------- 12 DATA 0 08 110
8 2 0 36,38,00,02,04,06,08,10,12,14,16 13 DATA 1 08 114
9 1 0 SAME 14 DATA 0 09 120
10 0 0 SAME 15 DATA 1 09 124
11 2 1 37,39,01,03,05,07,09,11,13,15,17 16 DATA 0 10 130
12 1 1 SAME 17 DATA 1 10 134
13 0 1 SAME 18 DATA 2 02 140
14 2 2 40,42,18,20,22,24,26,28,30,32,34 19 DATA 3 02 144
15 1 2 SAME 20 DATA 2 03 150
16 0 2 SAME 21 DATA 3 03 154
17 2 3 41,43,19,21,23,25,27,29,31,33,35 22 DATA 2 04 160
18 1 3 SAME 23 DATA 3 04 164
19 0 3 SAME 24 DATA 2 05 170
25 DATA 3 05 174
------------------------------------------------------------------------ 26 DATA 2 06 204
27 DATA 3 06 210
CALCULATING A FAILED RAM "E" NUMBER 28 DATA 2 07 214
----------------------------------- 29 DATA 3 07 220
30 DATA 2 08 224
TO CALCULATE THE "E" NUMBER OF A RAM ON AN M8579 BOARD YOU NEED 31 DATA 3 08 230
THE BLOCK NUMBER (BN), WORD NUMBER (WN), AND BIT-IN-FIELD NUMBER (BIFN). 32 DATA 2 09 234
USE ERROR ADDRESS BITS 14-21 IN AN SBDIAG FUNCTION 12 TO READ THE 33 DATA 3 09 240
ADDRESS RESPONSE RAM. BITS 12-13 OF THE SBDIAG ECHO ARE BN. ERROR 34 DATA 2 10 244
ADDRESS BITS 34-35 ARE WN. BIFN CAN BE DERIVED FROM THE TABLE ON THE 35 DATA 3 10 250
RIGHT USING EITHER THE SYNDROME (FROM FUNCTION 6.1) OR KNOWN ERROR DATA 36 ECC32 0 00 200
PATTERN. USE BIFN TO SELECT A NUMBER FROM THE TABLE BELOW AND PLUG THAT 37 ECC16 1 00 100
NUMBER INTO THE FORMULA TO GET THE ACTUAL "E" NUMBER. 38 ECC8 0 01 040
39 ECC4 1 01 020
BIFN: 0 1 2 3 4 5 6 7 8 9 10 40 ECC2 2 00 010
TABLE: 19. 38. 57. 76. 96. 116. 149. 167. 188. 204. 224. 41 ECC1 3 00 004
42 ECCPAR 2 01 000
FORMULA: E# = TABLE(BIFN) - (4 * WN) - BN 43 SPARE 3 01 (NOTE 2)
NOTE 1: THE SYNDROME IS A SIX BIT NUMBER AND IS SHOWN HERE OCTALLY GROUPED AS IT WOULD APPEAR IN AN SBDIAG FUNCTION 6.1 ECHO
IN BITS 07-12. NOTE THAT THE 43-BIT PARITY BIT IS NOT INCLUDED IN THIS LIST.
NOTE 2: THE SPARE BIT HAS NO ASSOCIATED SYNDROME. IF THE SPARE BIT HAPPENS TO BE REPLACING THE BIT WHICH YOUR SYNDROME
POINTS TO, THEN THE ERROR IS IN THE SPARE BIT AND NOT THE BIT WHICH THE SYNDROME POINTS TO.
Page 22
8.0 MF20 ARRAY BOARD
MF20 ARRAY BOARD PROM DATA. M8579 MF20 STORAGE ARRAY BOARD E NUMBER LAYOUT
EVERY MF20 STORAGE ARRAY BOARD HAS ASSOCIATED WITH IT A SERIAL THE M8579 STORAGE BOARDS DO NOT HAVE ANY E NUMBERS
NUMBER AND OTHER RELATED DATA. THERE IS ONE 32-BIT PROM "WORD" PER ETCHED ON THEM, MAKING IT DIFFICULT TO LOCATE A
BOARD WHICH IS READ USING FUNCTION 2. WHICH WORD YOU GET DEPENDS UPON SPECIFIC DIP. USE THE MAP BELOW TO LOCATE MOS RAM
BITS 9-12. FOUR SBDIAGS ARE NECESSARY TO READ THE ENTIRE WORD BY DIPS OR DC008 MUX DIPS. EXAMPLE: THE E NUMBER FOR
VARYING THE BYTE NUMBER IN BITS 13 & 14. THE LAYOUT OF A PROM WORD IS BLOCK (BLK) 2, SUBBLOCK (SBLK) 1, BIT 18 WOULD BE
AS FOLLOWS: 51. USE BLK & SBLK TO FIND THE ROW. USING THE BIT
NUMBER, LOOK IN THE LOWER TABLE TO FIND THE COLUMN.
PROM BYTE: 0 1 2 3 THE MAP IS ORIENTED AS IF YOU HELD THE BOARD WITH
PROM BIT: 01234567012345670123456701234567 THE COMPONENT SIDE UP AND THE CONNECTOR EDGE TOWARDS
CONTENT: YYYYWWWWWW###########PPABNNSSMMM YOU. NOTE THAT THE ROW OF MUX CHIPS IS STAGGERED.
YYYY IS THE 4 BIT YEAR NUMBER WHICH IS THE BCD EQUIVALENT OF THE BLK SBLK MOS RAM AND MUX CHIP E NUMBERS
LAST DIGIT OF THE CALENDAR YEAR. --------------------------------------------------
WWWWWW IS THE 6 BIT WEEK NUMBER. 0 0 224 204 188 167 149 116 96 76 57 38 19
1 0 223 203 187 166 148 115 95 75 56 37 18
##...## IS AN 11 BIT SERIAL NUMBER WHICH IS RESTARTED FROM 1 EACH WEEK. 2 0 222 202 186 165 147 114 94 74 55 36 17
3 0 221 201 185 164 146 113 93 73 54 35 16
PP IS A MOS RAM POPULATION CODE: 00 - MOS RAM BLOCK 0 EXISTS 0 1 220 200 184 163 145 112 92 72 53 34 15
01 - BLOCKS 0 & 1 EXIST 1 1 219 199 183 162 144 111 91 71 52 33 14
10 - BLOCKS 0, 1, & 2 EXIST 2 1 218 198 182 161 143 110 90 70 51 32 13
11 - BOARD IS FULLY POPULATED 3 1 217 197 181 160 142 109 89 69 50 31 12
0 2 216 196 180 159 141 108 88 68 49 30 11
A PARITY BIT SUCH THAT BYTES 0-2 OF THE WORD HAVE EVEN PARITY. 1 2 215 195 179 158 140 107 87 67 48 29 10
2 2 214 194 178 157 139 106 86 66 47 28 9
B ANOTHER PARITY BIT SUCH THAT BYTE 3 HAS ODD PARITY. 3 2 213 193 177 156 138 105 85 65 46 27 8
0 3 212 192 176 155 137 104 84 64 45 26 7
NN IS THE 2 BIT "TIMING NUMBER". THIS NUMBER DETERMINES WHICH 1 3 211 191 175 154 136 103 83 63 44 25 6
DISTINCT TIMING IS TO BE USED WITH THE RAMS ON THIS BOARD. 2 3 210 190 174 153 135 102 82 62 43 24 5
3 3 209 189 173 152 134 101 81 61 42 23 4
SS IS THE MOS RAM SIZE CODE: 11 - INDICATES 4K RAMS MUX ==> 207 205 171 169 131 98 79 59 40 20 2
10 - 16K RAMS
01 - 32K RAMS FIELD KL10 BIT NUMBERS AS POSITIONED IN FIELD
00 - 64K RAMS ----- -------------------------------------------
0 16 14 12 10 08 06 04 02 00 38 36
MMM IS THE 3 BIT MOS RAM MANUFACTURER CODE: 1 17 15 13 11 09 07 05 03 01 39 37
2 34 32 30 28 26 24 22 20 18 42 40
000 - MOSTEK 100 - MOTOROLA 3 35 33 31 29 27 25 23 21 19 43 41
001 - INTEL 101 - T.I.
010 - FUJITSU 110 - OTHER (DEC, DEC PART NUMBER, ETC.)
011 - HITACHI 111 - UNASSIGNED =================================================
THERE ARE ACTUALLY 8 PROM WORDS PER ARRAY BOARD, OF WHICH ONLY
ONE IS ALLOWED OUT BY THE STATE OF THE 3 JUMPERS. IF NO JUMPER IS CUT CONTROL BOARD SLOT NUMBERS & BASIC FUNCTIONS
THEN WORD 0 IS SELECTED, IF ALL ARE CUT THEN WORD 7 IS SELECTED. THESE ---------------------------------------------------
JUMPERS ARE CUT IN RELATION TO THE MANUFACTURER OF THE MOS RAMS ON THE
BOARD. NOTE ALSO THAT THE SERIAL NUMBER IS ALSO PRINTED ON THE BOARD IN NUMBER NAME SLOT FUNCTION
THE FORM:
YWW#### M8574 WRP 04 XBUS DATA/WR PATH/ECC GEN/SPARE OUT
M8576 CTL 05 XBUS & CYC CTL/WD & GRP SEL/PROM RD
**** PLEASE MAKE SURE WHEN AN ARRAY BOARD IS SENT IN FOR REPAIR THAT THE M8575 SYN 06 SPARE IN/CHK/CORR/ADR 14-21/V MARG
**** ERROR TYPEOUT SHOWING THE SERIAL NUMBER IS ATTACHED TO THE BOARD. M8577 ADT 07 MOS ADR/TIMING/REFRESH/BLK SEL/ERR
Page 23
9.0 MG20 ARRAY BOARD (M8570)
CALCULATING A FAILED RAM "E" NUMBER
-----------------------------------
TO CALCULATE THE "E" NUMBER OF A RAM ON AN M8570 BOARD YOU NEED
THE BLOCK NUMBER (BN), WORD NUMBER (WN), AND BIT-IN-FIELD NUMBER (BIFN).
USE ERROR ADDRESS BITS 14-21 IN AN SBDIAG FUNCTION 12 TO READ THE
ADDRESS RESPONSE RAM. BITS 12-13 OF THE SBDIAG ECHO ARE BN. ERROR
ADDRESS BITS 34-35 ARE WN. BIFN CAN BE DERIVED FROM THE TABLE ON THE
RIGHT USING EITHER THE SYNDROME (FROM FUNCTION 6.1) OR KNOWN ERROR DATA
PATTERN. USE BIFN TO SELECT A NUMBER FROM THE TABLE BELOW AND PLUG THAT
NUMBER INTO THE FORMULA TO GET THE ACTUAL "E" NUMBER.
BIFN: 0 1 2 3 4 5 6 7 8 9 10
TABLE: 19. 38. 57. 75. 92. 112. 146. 165. 183. 202. 221.
FORMULA: E# = TABLE(BIFN) - (4 * WN) - BN
M8570 MG20 STORAGE ARRAY BOARD E NUMBER LAYOUT
M8570 STORAGE BOARDS DO NOT HAVE ANY E NUMBERS ETCHED ON THEM, MAKING IT DIFFICULT TO LOCATE A SPECIFIC DIP. USE THE MAP BELOW
TO LOCATE MOS RAM DIPS OR DC008 MUX DIPS. EXAMPLE: THE E NUMBER FOR BLOCK (BLK) 2, SUBBLOCK (SBLK) 1, BIT 18 WOULD BE 51. USE
BLK & SBLK TO FIND THE ROW. USING THE BIT NUMBER, LOOK IN THE LOWER TABLE TO FIND THE COLUMN. THE MAP IS ORIENTED AS IF YOU
HELD THE BOARD WITH THE COMPONENT SIDE UP AND THE CONNECTOR EDGE TOWARDS YOU. NOTE THAT THE ROW OF MUX CHIPS IS STAGGERED.
BLK SBLK MOS RAM AND MUX CHIP E NUMBERS
-----------------------------------------------------
0 0 221 202 183 165 146 112 92 75 57 38 19
1 0 220 201 182 164 145 111 91 74 56 37 18
2 0 219 200 181 163 144 110 90 73 55 36 17
3 0 218 199 180 162 143 109 89 72 54 35 16
0 1 217 198 179 161 142 108 88 71 53 34 15
1 1 216 197 178 160 141 107 87 70 52 33 14
2 1 215 196 177 159 140 106 86 69 51 32 13
3 1 214 195 176 158 139 105 85 68 50 31 12
0 2 213 194 175 157 138 104 84 67 49 30 11
1 2 212 193 174 156 137 103 83 66 48 29 10
2 2 211 192 173 155 136 102 82 65 47 28 9
3 2 210 191 172 154 135 101 81 64 46 27 8
0 3 209 190 171 153 134 100 80 63 45 26 7
1 3 208 189 170 152 133 99 79 62 44 25 6
2 3 207 188 169 151 132 98 78 61 43 24 5
3 3 206 187 168 150 131 97 77 60 42 23 4
MUX ==> 204 203 185 148 128 114 95 58 40 20 2
FIELD KL10 BIT NUMBERS AS POSITIONED IN FIELD
----- -------------------------------------------
0 16 14 12 10 08 06 04 02 00 38 36
1 17 15 13 11 09 07 05 03 01 39 37
2 34 32 30 28 26 24 22 20 18 42 40
3 35 33 31 29 27 25 23 21 19 43 41
====================================================
Page 24
10.0 MF20 ERRORS
10.1 CORRECTION CODE (ECC) CALCULATION
CALCULATION OF THE ECC IS BASED UPON BIT CHANGES FROM A DATA WORD OF ALL ZEROS. THE ECC FOR ALL ZEROS IS DEFINED TO BE
11 111 11 (OR 376 AS IT WOULD APPEAR IN BITS 07-13 OF AN SBDIAG FUNC 6.0). TO CALCULATE THE ECC FOR ANY GIVEN WORD. TAKE
EACH "1" BIT FROM THE DATA WORD, GET THE SYNDROME FOR THAT BIT FROM THE TABLE BELOW, AND XOR IT INTO THE PREVIOUS ECC. INITIALLY
THE ECC IS THE ECC FOR A DATA WORD OF ZERO.
IF THERE ARE MORE ONES THAN ZEROS IN THE WORD THEN DO THE SAME AS ABOVE ONLY APPLY THE SYNDROMES FOR EACH "0" BIT AND USE
THE INITIAL ECC VALUE OF 10 101 11 (256 OCTAL). NOTE THAT THE FOLLOWING RELATION ALWAYS HOLDS:
ECC(X)=120.XOR.ECC(.NOT.X)
WHICH MEANS THAT THE ECC FOR THE COMPLEMENT OF A WORD DIFFERS FROM THE ECC OF THE WORD BY 120 OCTAL.
*** SYNDROME & 43-BIT PARITY TABLE ***
BIT FCN 6 SYN SYN SYN SYN SYN SYN 43B
# OCTAL 32 16 8 4 2 1 PAR
--- ----- --- --- --- --- --- --- ---
00 016 0 0 0 0 1 1 1 EXAMPLE: CALCULATE THE ECC & 43-BIT PARITY FOR A DATA WORD OF 3
01 026 0 0 0 1 0 1 1
02 032 0 0 0 1 1 0 1 376 IS ECC FOR ZERO
03 034 0 0 0 1 1 1 0 250 IS SYNDROME & 43B PAR FOR BIT 35
04 046 0 0 1 0 0 1 1 244 IS SYNDROME & 43B PAR FOR BIT 34
05 052 0 0 1 0 1 0 1 --- .XOR. ALL OF THE ABOVE
06 054 0 0 1 0 1 1 0 362 IS THE ECC FOR 3
07 062 0 0 1 1 0 0 1
08 064 0 0 1 1 0 1 0
09 070 0 0 1 1 1 0 0
10 076 0 0 1 1 1 1 1
11 106 0 1 0 0 0 1 1 ANOTHER METHOD FOR COMPUTING THE ECC IS AS FOLLOWS: FOR EACH OF THE 7
12 112 0 1 0 0 1 0 1 ECC BITS TAKE THE DATA WORD, .AND. IT WITH THE APPROPRIATE MASK BELOW,
13 114 0 1 0 0 1 1 0 AND COMPUTE THE RESULT'S ODD PARITY. THAT BIT IS ONE OF THE ECC BITS.
14 122 0 1 0 1 0 0 1
15 124 0 1 0 1 0 1 0 000000 001777 ECC 32
16 130 0 1 0 1 1 0 0 000177 776000 ECC 16
17 136 0 1 0 1 1 1 1 037600 776007 ECC 8
18 142 0 1 1 0 0 0 1 343617 036170 ECC 4
19 144 0 1 1 0 0 1 0 554663 146631 ECC 2
20 150 0 1 1 0 1 0 0 665325 253252 ECC 1
21 156 0 1 1 0 1 1 1 732351 455514 43B PAR
22 160 0 1 1 1 0 0 0
23 166 0 1 1 1 0 1 1
24 172 0 1 1 1 1 0 1
25 174 0 1 1 1 1 1 0
26 206 1 0 0 0 0 1 1
27 212 1 0 0 0 1 0 1 BITS 03, 09, 10, 13, 15, 25, AND 32 (040624 002010) FORM A SET WHICH IS
28 214 1 0 0 0 1 1 0 ONE OF MANY POSSIBLE "INDEPENDENT" BIT SETS. BY VARYING THE VALUES OF
29 222 1 0 0 1 0 0 1 THESE 7 BITS ONE CAN GENERATE ALL 128 ECC CODES, THOUGH NOT IN ANY
30 224 1 0 0 1 0 1 0 SPECIAL ORDER. ADDING BIT 35 ALLOWS PLAYING WITH THE SPARE BIT BUT DOES
31 230 1 0 0 1 1 0 0 NOT ALTER THE INDEPENDENCE OF THIS BIT SET.
32 236 1 0 0 1 1 1 1
33 242 1 0 1 0 0 0 1
34 244 1 0 1 0 0 1 0 NOTE THAT THE OCTAL NUMBER 200,,2217 (33555599 DECIMAL) IS A SMALL,
35 250 1 0 1 0 1 0 0 RELATIVELY EASY TO REMEMBER NUMBER WHICH PRODUCES AN ALL ZEROS ECC.
Page 25
-- -- -- -- -- -- --
10 15 18 19 20 20 20 <== SUM OF SYNDROME ONE BITS
Page 26
10.2 DETECTION AND CORRECTION
HOW MF20 ERROR CORRECTION WORKS. AN EXAMPLE OF ERROR CORRECTION.
-------------------------------- -------------------------------
WHEN A WRITE IS DONE THE MF20 CALCULATES AN ECC BASED UPON THE WRITE A 3 TO MEMORY, ASSUME THAT BIT 27 IS STUCK
DATA IT GOT FROM THE CPU. THE DATA AND THE ECC ARE THEN PUT IN MEMORY. HIGH IN MEMORY. AS SHOWN ON THE PREVIOUS PAGE
WHEN A READ IS DONE THE MF20 GETS THE DATA AND THE OLD ECC FROM MEMORY. THE ECC FOR 3 IS 362, THEREFORE:
IT THEN COMPUTES A NEW ECC ON THE DATA AND XOR'S THE OLD AND NEW ECC'S 3<362> GOES TO MEMORY.
WHICH RESULTS IN A SYNDROME. OBVIOUSLY IF THERE WAS NO ERROR IN MEMORY ON THE READ BIT 27 IS STUCK HIGH SO THAT WE SEE:
THE OLD AND NEW ECC'S WILL BE THE SAME, AND THE SYNDROME WILL BE ZERO 403<362> FROM MEMORY.
SINCE XOR'ING EQUAL VALUES PRODUCES 0. WE CALCULATE THE NEW ECC:
374 IS ECC FOR 0 (LESS 43B PAR BIT)
NOW, WHEN THERE IS AN ERROR IN ONE BIT THE NEW SYNDROME WILL 250 IS SYNDROME FOR BIT 35
DIFFER FROM WHAT IT WOULD HAVE BEEN BY A VALUE EQUAL TO THE SYNDROME OF 244 IS SYNDROME FOR BIT 34
THE BAD BIT. THUS XOR'ING THE NEW SYNDROME WITH THE OLD ONE FROM MEMORY 210 IS SYNDROME FOR BIT 27
WILL RESULT IN A SYNDROME EQUAL TO THAT OF THE BAD BIT. THE SYNDROME --- XOR THE WHOLE MESS TOGETHER
MAPPED TO A BIT WHICH, IN TURN, IS USED TO COMPLEMENT (IE CORRECT) THE 170 IS THE NEW ECC.
BAD BIT. SEE EXAMPLE ON RIGHT AT TOP. WE NOW CALCULATE THE SYNDROME:
360 IS THE OLD ECC (LESS 43B PAR BIT)
WHEN ONE BIT IS BAD ON A READ THEN, OBVIOUSLY, THE PARITY OF THE 170 IS THE NEW ECC
43-BIT WORD FROM MEMORY WILL BE BAD. BAD PARITY IS WHAT CAUSES THE MF20 --- XOR
TO RECOGNIZE A CORRECTABLE ERROR. THE MF20 THEN GOES OFF AND COMPUTES 210 IS THE SYNDROME
THE SYNDROME. IF THE SYNDROME IS THAT OF ONE OF THE ECC BITS THEN THE THE HARDWARE MAPS THE SYNDROME TO THE PROPER BIT
MAPPED CORRECTION BIT GOES NOWHERE SINCE WE DON'T SEND THE ECC BACK TO AND CHANGES THE BIT:
THE CPU. 403 IS DATA FROM MEMORY
400 IS THE MAPPED BIT
IF THERE IS NO PARITY ERROR AND THE SYNDROME IS ZERO THEN THERE --- XOR
IS NO CORRECTION NECESSARY. HOWEVER A DOUBLE BIT (UNCORRECTABLE) ERROR 3 IS WHAT GOES BACK TO THE CPU.
GIVES GOOD PARITY (SINCE 2 BITS ARE BAD) AND A NONZERO SYNDROME. THIS HOCUS-POCUS-PRESTO-CHANGO, THE ERROR IS GONE!
IS HOW AN MF20 RECOGNIZES AN DOUBLE BIT ERROR. WHEN A DOUBLE BIT ERROR
IS DETECTED THE MF20 SHIPS THE DATA, AS IS, TO THE CPU WHILE FORCING BAD --------------------------------------------------
XBUS PARITY. SEE THE TABLE ON THE RIGHT AT THE BOTTOM FOR A CONCISE
LIST OF ERROR CONDITIONS. SYNDROME PARITY MEANING & ACTION TAKEN BY MF20
-------- ------ ------------------------------
ZERO OK EVERY THING IS OK. TAKE DATA
AS IT IS.
ZERO BAD 43-BIT PARITY BIT IS BAD.
USE THE DATA AS IT IS.
1,2,4,8, BAD AN ECC BIT IS BAD. USE THE
16,32 DATA AS IT IS.
NOT 0 OR BAD A DATA BIT IS BAD. CORRECT IT
1,2, ETC AND SEND BACK CORRECTED WORD.
NOT 0 OK THIS IS A NON-CORRECTABLE
ERROR. PUT DATA ON XBUS AS
IT IS PLUS BAD PARITY.
Page 27
10.3 DOUBLE BIT ERROR SYNDROMES
THIS IS THE TABLE OF SYNDROMES VS. ALL POSSIBLE BIT PAIRS WHICH COULD CAUSE A DOUBLE
BIT ERROR WITH THAT SYNDROME. THE SYNDROME IS THE OCTAL NUMBER ON THE LEFT. REMEMBER THAT
IT IS ONLY 6 BITS BUT IS SHOWN ALIGNED AS IT IS IN AN SBUS DIAG FUNCTION 6.1. ALL THE BIT
PAIRS WHICH WOULD RESULT IN A GIVEN SYNDROME ARE ON THE LINE TO THE RIGHT OF THE SYNDROME.
THESE PAIRS ARE DECIMAL KL10 BIT NUMBERS WITH THE LOWER NUMBER OF THE PAIR TO THE LEFT OF THE
COMMA. THE PAIRS ARE IN ORDER BY THE LOWER NUMBER. THE TABLE CONTINUES ON THE NEXT PAGE.
SYN BIT PAIRS WHICH COULD CAUSE A DBE WITH THE GIVEN SYNDROME
--- ---------------------------------------------------------
374 15,35 16,34 17,33 18,32 19,31 20,30 21,29 22,28 23,27 24,26 25,36
370 14,35 16,33 17,34 18,31 19,32 20,29 21,30 22,27 23,28 24,36 25,26
364 14,34 15,33 17,35 18,30 19,29 20,32 21,31 22,26 23,36 24,28 25,27
360 14,33 15,34 16,35 18,29 19,30 20,31 21,32 22,36 23,26 24,27 25,28
354 11,35 12,34 13,33 18,28 19,27 20,26 21,36 22,32 23,31 24,30 25,29
350 12,33 13,34 18,27 19,28 20,36 21,26 22,31 23,32 24,29 25,30 35,37
344 11,33 13,35 18,26 19,36 20,28 21,27 22,30 23,29 24,32 25,31 34,37
340 11,34 12,35 18,36 19,26 20,27 21,28 22,29 23,30 24,31 25,32 33,37
334 11,31 12,30 13,29 14,28 15,27 16,26 17,36 23,35 24,34 25,33 32,37
330 11,32 12,29 13,30 14,27 15,28 16,36 17,26 22,35 24,33 25,34 31,37
324 11,29 12,32 13,31 14,26 15,36 16,28 17,27 22,34 23,33 25,35 30,37
320 11,30 12,31 13,32 14,36 15,26 16,27 17,28 22,33 23,34 24,35 29,37
314 11,27 12,26 13,36 14,32 15,31 16,30 17,29 19,35 20,34 21,33 28,37
310 11,28 12,36 13,26 14,31 15,32 16,29 17,30 18,35 20,33 21,34 27,37
304 11,36 12,28 13,27 14,30 15,29 16,32 17,31 18,34 19,33 21,35 26,37
300 11,26 12,27 13,28 14,29 15,30 16,31 17,32 18,33 19,34 20,35 36,37
274 01,35 02,34 03,33 04,31 05,30 06,29 07,28 08,27 09,26 10,36 32,38
270 02,33 03,34 04,32 05,29 06,30 07,27 08,28 09,36 10,26 31,38 35,39
264 01,33 03,35 04,29 05,32 06,31 07,26 08,36 09,28 10,27 30,38 34,39
260 01,34 02,35 04,30 05,31 06,32 07,36 08,26 09,27 10,28 29,38 33,39
254 00,33 04,27 05,26 06,36 07,32 08,31 09,30 10,29 28,38 34,40 35,41
250 00,34 04,28 05,36 06,26 07,31 08,32 09,29 10,30 27,38 33,40 35,42
244 00,35 04,36 05,28 06,27 07,30 08,29 09,32 10,31 26,38 33,41 34,42
240 04,26 05,27 06,28 07,29 08,30 09,31 10,32 33,42 34,41 35,40 36,38
234 00,29 01,27 02,26 03,36 08,35 09,34 10,33 28,39 30,40 31,41 32,42
230 00,30 01,28 02,36 03,26 07,35 09,33 10,34 27,39 29,40 31,42 32,41
224 00,31 01,36 02,28 03,27 07,34 08,33 10,35 26,39 29,41 30,42 32,40
220 00,32 01,26 02,27 03,28 07,33 08,34 09,35 29,42 30,41 31,40 36,39
214 00,36 01,31 02,30 03,29 04,35 05,34 06,33 26,40 27,41 28,42 32,39
210 00,26 01,32 02,29 03,30 05,33 06,34 27,42 28,41 31,39 35,38 36,40
Page 28
204 00,27 01,29 02,32 03,31 04,33 06,35 26,42 28,40 30,39 34,38 36,41
200 00,28 01,30 02,31 03,32 04,34 05,35 26,41 27,40 29,39 33,38 36,42
174 00,22 01,20 02,19 03,18 04,16 05,15 06,14 07,13 08,12 09,11 10,37 17,38 21,39 23,40 24,41 25,42
170 00,23 01,21 02,18 03,19 04,17 05,14 06,15 07,12 08,13 09,37 10,11 16,38 20,39 22,40 24,42 25,41
164 00,24 01,18 02,21 03,20 04,14 05,17 06,16 07,11 08,37 09,13 10,12 15,38 19,39 22,41 23,42 25,40
160 00,25 01,19 02,20 03,21 04,15 05,16 06,17 07,37 08,11 09,12 10,13 14,38 18,39 22,42 23,41 24,40
154 00,18 01,24 02,23 03,22 04,12 05,11 06,37 07,17 08,16 09,15 10,14 13,38 19,40 20,41 21,42 25,39
150 00,19 01,25 02,22 03,23 04,13 05,37 06,11 07,16 08,17 09,14 10,15 12,38 18,40 20,42 21,41 24,39
144 00,20 01,22 02,25 03,24 04,37 05,13 06,12 07,15 08,14 09,17 10,16 11,38 18,41 19,42 21,40 23,39
140 00,21 01,23 02,24 03,25 04,11 05,12 06,13 07,14 08,15 09,16 10,17 18,42 19,41 20,40 22,39 37,38
134 00,14 01,12 02,11 03,37 04,24 05,23 06,22 07,21 08,20 09,19 10,18 13,39 15,40 16,41 17,42 25,38
130 00,15 01,13 02,37 03,11 04,25 05,22 06,23 07,20 08,21 09,18 10,19 12,39 14,40 16,42 17,41 24,38
124 00,16 01,37 02,13 03,12 04,22 05,25 06,24 07,19 08,18 09,21 10,20 11,39 14,41 15,42 17,40 23,38
120 00,17 01,11 02,12 03,13 04,23 05,24 06,25 07,18 08,19 09,20 10,21 14,42 15,41 16,40 22,38 37,39
114 00,37 01,16 02,15 03,14 04,20 05,19 06,18 07,25 08,24 09,23 10,22 11,40 12,41 13,42 17,39 21,38
110 00,11 01,17 02,14 03,15 04,21 05,18 06,19 07,24 08,25 09,22 10,23 12,42 13,41 16,39 20,38 37,40
104 00,12 01,14 02,17 03,16 04,18 05,21 06,20 07,23 08,22 09,25 10,24 11,42 13,40 15,39 19,38 37,41
100 00,13 01,15 02,16 03,17 04,19 05,20 06,21 07,22 08,23 09,24 10,25 11,41 12,40 14,39 18,38 37,42
074 00,07 01,05 02,04 03,38 06,39 08,40 09,41 10,42 11,24 12,23 13,22 14,21 15,20 16,19 17,18 25,37 30,35 31,34 32,33
070 00,08 01,06 02,38 03,04 05,39 07,40 09,42 10,41 11,25 12,22 13,23 14,20 15,21 16,18 17,19 24,37 29,35 31,33 32,34
064 00,09 01,38 02,06 03,05 04,39 07,41 08,42 10,40 11,22 12,25 13,24 14,19 15,18 16,21 17,20 23,37 29,34 30,33 32,35
060 00,10 01,04 02,05 03,06 07,42 08,41 09,40 11,23 12,24 13,25 14,18 15,19 16,20 17,21 22,37 29,33 30,34 31,35 38,39
054 00,38 01,09 02,08 03,07 04,40 05,41 06,42 10,39 11,20 12,19 13,18 14,25 15,24 16,23 17,22 21,37 26,35 27,34 28,33
050 00,04 01,10 02,07 03,08 05,42 06,41 09,39 11,21 12,18 13,19 14,24 15,25 16,22 17,23 20,37 27,33 28,34 35,36 38,40
044 00,05 01,07 02,10 03,09 04,42 06,40 08,39 11,18 12,21 13,20 14,23 15,22 16,25 17,24 19,37 26,33 28,35 34,36 38,41
040 00,06 01,08 02,09 03,10 04,41 05,40 07,39 11,19 12,20 13,21 14,22 15,23 16,24 17,25 18,37 26,34 27,35 33,36 38,42
034 00,39 01,40 02,41 03,42 04,09 05,08 06,07 10,38 11,16 12,15 13,14 17,37 18,25 19,24 20,23 21,22 26,31 27,30 28,29 32,36
030 00,01 02,42 03,41 04,10 05,07 06,08 09,38 11,17 12,14 13,15 16,37 18,24 19,25 20,22 21,23 26,32 27,29 28,30 31,36 39,40
024 00,02 01,42 03,40 04,07 05,10 06,09 08,38 11,14 12,17 13,16 15,37 18,23 19,22 20,25 21,24 26,29 27,32 28,31 30,36 39,41
020 00,03 01,41 02,40 04,08 05,09 06,10 07,38 11,15 12,16 13,17 14,37 18,22 19,23 20,24 21,25 26,30 27,31 28,32 29,36 39,42
014 00,42 01,02 03,39 04,05 06,38 07,10 08,09 11,12 13,37 14,17 15,16 18,21 19,20 22,25 23,24 26,27 28,36 29,32 30,31 34,35 40,41
010 00,41 01,03 02,39 04,06 05,38 07,09 08,10 11,13 12,37 14,16 15,17 18,20 19,21 22,24 23,25 26,28 27,36 29,31 30,32 33,35 40,42
004 00,40 01,39 02,03 04,38 05,06 07,08 09,10 11,37 12,13 14,15 16,17 18,19 20,21 22,23 24,25 26,36 27,28 29,30 31,32 33,34 41,42
000 A ZERO SYNDROME NEVER CAUSES A DOUBLE BIT ERROR.
Page 29
11.0 KL10 I/O INSTRUCTIONS RELATING TO MEMORY
00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
APRID *-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
700000 * < MICROCODE OPTIONS > * < MICROCODE VERSION NUMBER > *
LH(E) *KLPAG!XADDR!UNSTD* ! ! * ! !TRCKS* ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
* < HARDWARE OPTIONS > * < PROCESSOR SERIAL NUMBER > *
RH(E) *50 HZ!CACHE!CHANL* XADR!MSTR ! * ! ! * ! ! * ! ! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
CONO APR, *-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
700200 * ! I/O ! SELECTED FLAGS * < SELECT FLAG > ! * PI LEVEL *
E * !RESET! EN * DIS ! CLR ! SET * SBUS! NXM ! IOPF*MBPAR!C DIR!ADR P*POWER!SWEEP! * 4 ! 2 ! 1 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
CONI APR, *-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
700240 * * < ENABLED FLAGS > ! *
LH(E) * ! ! * ! ! * SBUS! NXM ! IOPF*MBPAR!C DIR!ADR P*POWER!SWEEP! * ! ! *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
* !SWEEP! * SBUS! ! I/O * MB !CACHE! ADDR*POWER!SWEEP! INT * PI LEVEL *
RH(E) * ! BUSY! * ! ! * ERR ! NXM ! PGF * PAR ! DIR ! PAR * FAIL! DONE! REQ * 4 ! 2 ! 1 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
00/18 01/19 02/20 03/21 04/22 05/23 06/24 07/25 08/26 09/27 10/28 11/29 12/30 13/31 14/32 15/33 16/34 17/35
RDERA *-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
700400 * WORD NO !SWEEP* CHAN! DATA SRC *WRITE! ! PHYSICAL ADDRESS *
LH(E) * ! ! REF * REF ! ! * REF ! JUNK! JUNK* ! ! * ! ! 14 * 15 ! 16 ! 17 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
* < PHYSICAL ADDRESS OF FIRST WORD OF TRANSFER > *
RH(E) * 18 ! 19 ! 20 * 21 ! 22 ! 23 * 24 ! 25 ! 26 * 27 ! 28 ! 29 * 30 ! 31 ! 32 * 33 ! 34 ! 35 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
DATA SRC WRITE = 0 WRITE = 1
00 MEMORY (READ, RPW) CHANNEL (STORE STATUS)
01 CHANNEL (DATA STORE)
10 EBOX STORE FROM AR
11 READ FROM CACHE WRITEBACK FROM CACHE
(PG REFILL OR CHAN READ)
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
CONO PI, *-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
700600 * WRITE EVEN PAR * ! DROP!CLEAR* REQ ! TURN CHAN * TURN SYS ! < SELECT CHANNEL > *
E * ADDR! DATA! DIR * ! INT ! SYS * INT ! ON ! OFF * OFF ! ON ! 1 * 2 ! 3 ! 4 * 5 ! 6 ! 7 *
*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*-----+-----+-----*
[END OF TGHA.DOC]