| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 93.1 | CIBCA-A/B Latest Ucode Rev Files | KERNEL::BLAND | toward 2000 ... | Thu Feb 15 1990 19:53 | 9 |
| I have copied the latest versions of CIBCA.BIN & CIBCB.BIN to directory
RDC$COMMON:[.CI]. As these versions have been available for a long
time, I believe we should be encouraging engineers to install as
soon as convenient. There is some info on what the latest revs fix
in the BI notes file. In case we need to downline load these (65
blocks each), and the fact that they require the EEPROM utility EVGDA
to be available, I will attempt to copy this into the same directory.
Regards, N_B
|
| 93.2 | EEPROM Update Utility EVGDA Revision 2.1 | KERNEL::BLAND | toward 2000 ... | Thu Feb 15 1990 20:29 | 4 |
| OK, we now have EVGDA.EXE revision 2.1 in directory RSDS$COMMON:[.CI].
It is 16 blocks in size. We should not be using revsions before 2.0.
Regards, N_B
|
| 93.3 | | KERNEL::MOUNTFORD | | Wed Feb 28 1990 13:38 | 42 |
| Moved by moderator with permission:
----------------------------------
<<< KERNEL::DISK$APD1:[NOTES$LIBRARY]CSGUK_SYSTEMS.NOTE;1 >>>
-< CSGUK_SYSTEMS >-
================================================================================
Note 97.0 CI Information No replies
KERNEL::ROBB 32 lines 28-FEB-1990 13:08
--------------------------------------------------------------------------------
CIBCA -A or -B ?
If you need to know if you have a CIBCI -A or -B then:-
If the system is down.
Bits <23:22> of the DTYPE register are 00 for CIBCA-A and 11 for CIBCA-B,
Page 6-7 of the VAXBI Options Handbook has it all and more.
>>>e/p/l 2xxyy000
where xx = 2*BI number and yy = 2*NODE number
Device type = 0108
*****
If the system is up
$Show cluster/cont
Command > add circ ! This will tell you what type of CI you have
Command > add rp_rev
40054002
^ ^
|___|___________ These bits are set for a -B and clear for a -A
This example shows a CIBCA-B with a revision of 50002
This may change but it seems to be true so far.
Hope this is of some help.
|
| 93.4 | CIBCA-? and UCODE rev with SYSTEM down | KERNEL::BLAND | toward 2000 ... | Tue Mar 06 1990 13:44 | 66 |
| Here is an example of finding the same info with regards to the flavour of
CIBCA (-A or -B ?), plus the Ucode revision, if the system is down.
System was a 6340 with CIBCA at BI node # 5, XMI node # E.
>>> SHOW CONFIG
Type Rev
......................
......................
......................
XBI D
......................
......................
......................
XBI E
1+ DWMBA/B (2107) 0007
5+ CIBCA (0108) 41C1 !DTYPE REG ("C" is bits 23 & 22), hence -B
6+ TBK70 (410B) 0307 ========
Base address of CIBCA is 3C00A00 (DTYPE REG)
>>> E/P/L 3C00A000
P 3C00A000 41C10108 !Same info as above (in reverse), hence -B
========
We can examine the DTYPE register for the Scorpio & Nautilus systems as well
to get the same info. Assume CI at BI node # 5 & BI # 0 (Nautilus), e.g.
>>> E/P/L 2000A000
P 2000A000 00010108 !DTYPE REG (bits 23:22 are both 0), hence -A
OK, now for the Ucode revision.
Assuming the CIBCA-? has passed self-test, then there are two locations in
the CI Local Store that we can examine to get CI Ucode revision info.
We need to examine the CI base address + 108C & 1090.
FOR 6000 SYSTEMS
================
>>> E/P/L 3C00B08C
P 3C00B08C 00004005 !Note, functional UCODE rev 5 & CIBCA-B
^ ^ ^ ^
| |..................| |
|...............................................|
>>> E/P/L 3C00B090
P 3C00B090 00004002 !Note, self-test UCODE rev 2 & CIBCA-B
^ ^ ^ ^
| |..................| |
|...............................................|
FOR SCORPIO/NAUTILUS SYSTEMS
============================
>>> E/P/L 2000B08C
P 2000B08C 00000007 !Note, functional UCODE rev 7 & CIBCA-A
^ ^ ^ ^
| |..................| |
|...............................................|
>>> E/P/L 2000B090
P 2000B090 00000005 !Note, self-test UCODE rev 5 & CIBCA-A
^ ^ ^ ^
| |..................| |
|...............................................|
|
| 93.5 | L0100/L0118 Jumpers & Switches | KERNEL::BLAND | I wanna be a slug | Thu Jun 27 1991 03:00 | 65 |
| I recently extracted this from an FCO document for someone in our
group; maybe others will find it useful. NB.
============================================================================
SET L0100-E1 COMPATIBILITY JUMPER AS FOLLOWS:
---------------------------------------------
o 7/10-TICK COMPAT. MODE SET BY 1 JUMPER @ E177 stake-pins.
- "7-TICK COMPAT.-MODE": jumper E177-11 to E177-12
- "10-TICK INCOMPAT.-MODE": jumper E177-9 to E177-10
o NO CLUSTER-SIZE OR FUNCTION-MODE JUMPERS/SWITCHES.
o L0100-E1 QUICK-CHECK (PART-REVISION VERIFY):
- ECO wires at E177-13 and E177-9.
- STAKE-PINS (for jumper) @ E177-9,10,11,12.
- ETCH REV-C: P/N "50-14430C-0" marking.
"ECO-FIX OFF" "ECO-FIX ON"
"7-TICK" "10-TICK"
"COMPATIBLE" "INCOMPATIBLE"
20 +-+ 20 12 11
...................|.|.. ........................
| . . . . . . . . | | | | . . . . . . . . . . |
| E177 | | E177 |
| . . . . . . . . . . | | . . . . . . . . . . |
|......................| |..................|.|.|
1 9 10 1 | |
+-+
============================================================================
SET L0100-E2 SWITCH SETTINGS AS FOLLOWS:
----------------------------------------
o 7/10-TICK COMPAT. MODE SET BY DIP-SWITCH SW3-1 LOCATED AT
TOP CENTER OF L0100-E2 TO RIGHT OF E118:
- "7-TICK COMPAT.,ECO OFF": SW3-1 SWITCH "OFF"
- "10-TICK INCOMPAT.,ECO ON ": SW3-1 SWITCH "ON" .
NOTE: SW3-2 IS NOT USED, BUT SET SAME
AS SW3-1 TO AVOID CONFUSION.
o NO CLUSTER-SIZE OR FUNCTION-MODE JUMPERS/SWITCHES.
o L0100-E2 QUICK-CHECK (PART-REVISION VERIFY):
- SW3 DUAL-SWITCH-DIP-PACK AT TOP-CENTER, RIGHT OF
E118.
- ETCH REV-D: P/N "50-14430-0-0 D1-P4" marking.
============================================================================
SET L0118 SWITCH/JUMPER SETTINGS AS FOLLOWS:
--------------------------------------------
o 7/10-TICK COMPAT. MODE SET BY DIP-SWITCH SW3-4 LOCATED AT
TOP CENTER OF L0118 TO LEFT (BELOW IF VERTICAL) OF 2 CI-
NODE-ADDRESS 8-SWITCH PACKS.
- "7- TICK COMPAT.,ECO OFF": SW3-4 SWITCH "OFF".
- "10-TICK INCOMPAT.,ECO ON" : SW3-4 SWITCH "ON" .
- SWITCHES SW3-2 AND SW3-3 MUST BE "OFF" (AFFECTS
DELTA-TIME/QUIET-SLOT) !!
o CLUSTER-SIZE FOR 16/32-NODE ADDRESSING AND ARBITRATION IS
CONTROLLED BY SWITCH SW3-1. DEFAULT = 16-NODE-MODE =
"OFF".
- 16-NODE MODE (DEFAULT WITH NO CISCE): SW3-1 "OFF".
- 32-NODE MODE (CISCE 24-NODE CLUSTER): SW3-1 "ON" .
|
| 93.6 | CI & HSC Switches/Jumpers | KERNEL::BLAND | I wanna be a slug | Wed Aug 07 1991 02:03 | 136 |
| Author : RODNEY BOYLE
User type : USER
Location : CSSE
Vaxmail address : CSSE::BOYLE
+---------------------------+TM
| | | | | | | |
| d | i | g | i | t | a | l | TIME DEPENDENT BLITZ
| | | | | | | |
+---------------------------+
BLITZ TITLE: CI Switch Settings
DATE: 05-AUG-1991
AUTHOR: LARRY CRESS TD #: 000788
DTN: 522-6975
ENET: GENRAL::CRESS CROSS REFERENCE #'s:
DEPARTMENT: Storage Continuation (PRISM/TIME/CLD#'s)
Engineering CSSE
INTENDED AUDIENCE: ALL PRIORITY LEVEL: 1
(U.S./EUROPE/GIA) (1=TIME CRITICAL,
2=NON-TIME CRITICAL)
=====================================================================
PROBLEM:
Incorrect switch and/or jumper settings on the CI interface
modules in both HSC's and CPU's can cause various cluster disturbances,
including:
o disks in Mount Verification
o disks experiencing Mount Verification Timeout
o many Virtual Circuit (VC) Closures, resulting in
shadow copy initiation on shadowed disks
o various VMS bugchecks, most commonly CLUEXIT
The cause of these disturbances is CI collisions caused by mismatched
CI communications mechanisms. There are two mechanisms involved:
o "slot count" or "quiet time"
o "cluster_size"
Settings for these mechanisms are selected by switch positions on the
L0100/L0118 LINK modules in HSC's and older VAX CPU's, and by backplane
jumpers for CIBCA and CIXCD CI port adapters on newer VAX CPU's.
RESOLUTION/WORKAROUND:
Slot count or quiet time has two possible settings, 7 or 10. At the
time of HSCxx installation, it is necessary that all LINK and CI adapter
modules in the cluster be set to a consistent value to avoid the cluster
disturbances mentioned above. It is highly recommended that the slot
count for all CPU and HSC nodes within a cluster be set to a value
of 10 (not 7). Recently manufactured LINK modules are shipped with a slot
count of 10. LINK or CI port adapter modules in CPU's and some older
HSC's may still be set at a slot count of 7. (The CIXCD adapter does not
support a slot count of 7; therefore, if your cluster contains a CIXCD
adapter, all LINK and CI adapter modules should already be set to 10.
If your cluster does not contain a CIXCD adapter and is operating without
problems at a slot count of 7, you may set the HSC LINK module to 7 for
installation, and defer resetting the entire cluster to 10 until a more
convenient time. For your assistance in setting slot count in your cluster,
Table X-x gives a list of LINK and adapter module switch and jumper
settings to select slot count.
In addition, the cluster_size parameter must be consistent throughout
the cluster at a value of either 16 or 32. If all nodes in your cluster
are numbered in the range of 0 to 15, you may select a cluster_size of
*either* 16 or 32; if your cluster has any node numbered in the range 16
to 31, you *must* select a cluster_size of 32. Again, see Table X-x for
a list of LINK and adapter module switch and jumper settings to
select the cluster_size.
TABLE X-x CI ADAPTER MODULE SWITCH AND JUMPER SETTING MATRIX
+------------------------------------------------------------------------------+
| CI ADAPTER MODULE SWITCH and JUMPER SETTING MATRIX |
+------------------------------------------------------------------------------+
| ADAPTER | HOST | LINK MODULE | FUNCTION | SWITCH/ |
| TYPE(s) | SYSTEM(s) | | | JUMPER |
+------------------------------------------------------------------------------+
| CI750 | 11/750 | L0100 Rev E | Slot = 7 | S3-1 "OFF" |
| CI780 | 11/78x,86x0 | | Slot = 10 | S3-1 "ON" |
| CIBCI | 82x0, 83x0 | +---------------------------------+
| | 87x0, 8800 | | Clu_size 16 | FIXED |
| | HSC40,50,70 +--------------+-------------+-------------------+
| | | L0118 Rev B | Slot = 7 | S3-4 "OFF" |
| | | | Slot = 10 | S3-4 "ON" |
| | | +---------------------------------+
| | | | Clu_size 16 | S3-1 OFF |
| | | | Clu_size 32 | S3-1 on |
+------------------------------------------------------------------------------+
| | HSC60,90 | L0118-YA | Slot = 7 | S3-4 "OFF" |
| | | | Slot = 10 | S3-4 "ON" |
| | | +---------------------------------+
| | | | Clu_size 16 | S3-1 OFF |
| | | | Clu_size 32 | S3-1 on |
+------------------------------------------------------------------------------+
| CIBCA-Ax | 82x0, 83x0 | T1025 | Slot = 7 | Jumper E11-41 out |
| CIBCA-Bx | 85x0, 8700 | T1046 | Slot = 10 | Jumper E11-41 in |
| | 88x0, 6000 | +-------------+-------------------+
| | | | Clu_size 16 | Jumper D30-60 out |
| | | | Clu_size 32 | Jumper D30-60 in |
+------------------------------------------------------------------------------+
| CIXCD-xx | 6000, 9000 | T2080 | Slot = 7 | NOT SUPPORTED |
| | | | Slot = 10 | Jumper E15-45 in |
| | | +-------------+-------------------+
| | | | Clu_size 16 | Jumper D15-45 out |
| | | | Clu_size 32 | Jumper D15-45 in |
+------------------------------------------------------------------------------+
NOTES: L0100 Rev "D" modules are not supported.
On L0100 Rev "E" modules, S3-2 serves no function.
On L0118 modules;
(1) all ON-MODULE jumpers MUST be left in their default positions.
(2) S3-2 & 3 MUST remain in the "OFF" position.
For CIBCA options, all reference jumpers are on the backplane slot
for the PORT (T1015/T1045) module.
For CIBCA and CIXCD options, install ONLY those backplane jumpers
listed above and those required to specify the node address.
*** Digital Confidential ***
|
| 93.7 | New CI Ucode with VAXPAX Release 45 | KERNEL::BLAND | Norman Bland 833 3797 CSC, Basingstoke | Sun Dec 08 1991 21:27 | 26 |
| VAXPAX Release 45 has introduced new CI Ucode as follows:-
CIBCA,80005 TITLE: CIBCA FUNCT AND EEPROM UCODE
REASON_FOR_SUBMISSION
Add preferred no_rsp retry algorithm. Add code which
checks destination and complement destination fields
of receives packets and source field. Previous versions
of the code would crash the port this version reduces
the user visible effects of the problem without
sacrificing data integrity.
CIBCB,4007.4002 TITLE: CIBCB FUNCT AND EEPROM UCODE
REASON_FOR_SUBMISSION
Add path-switching eco and random path selection as part
of the 4k compliance requirements.
Fix invalid buffer name bug (#3704)
fixed self-directed packet path status bug (#4001).
NOTE. The CIXCD Ucode is now at 2.04 (44 when you ADD RP_REV in SHOW
CLUSTER/CONTINUOUS). I do no know what the new Ucode fixes.
Norman
|
| 93.8 | CIXCD 2.04 fixes | CSC32::J_MACGOWN | CSC/CS Remote Support | Thu Dec 12 1991 06:05 | 19 |
| Hi,
Heres whats fixed in CIXCD ucode 2.04. It's on VaxPax rel 45:
*****************************************************************
CIXCD,2.4 TITLE: CIXCD OPERATIONAL MICROCODE
REASON_FOR_SUBMISSION
FIXED PROBLEM WHERE VMS COULD NOT BOOT WITH CI CABLE
ON PATH A DISCONNECTED.
FIXED PORT STATUS REGISTER.
THE ABOVE PSR FIX WILL EFFECT HOW THE UCODE INTERRUPTS
THE HOST IN SOME CASES.
THE ERROR CODE FOR A DATA STRUCTURE ERROR IN BITS
15:0 OF THE PORT ERROR STATUS REGISTER
MISCELLANEOUS MINOR BUGS AND CLEANUP
*****************************************************************
Johnny Mac
|
| 93.9 | ci780.bin 200020 | KERNEL::ANTHONY | | Tue Nov 10 1992 14:46 | 10 |
|
Hi I've just seen a new ci780.bin
It's seen as 200020 from show clu/cont , add rp_rev
this new ucode is shipped with the star coupler expander
to increase CI nodes to more than 16.
Brian
|
| 93.10 | Not supplied on VAXPAX | KERNEL::BLAND | Norman Bland 833 3797 CSC, Basingstoke | Wed Nov 11 1992 06:59 | 6 |
| Brian, That is correct. What I would like to know is where we can get
hold of a copy just in case someone needs it one day. I think I will
mail one of my Diagnostic contacts in the states to see if they can
help.
Norman
|
| 93.11 | Found it CI780.BIN REV 20 | KERNEL::BLAND | Norman Bland 833 3797 CSC, Basingstoke | Wed Nov 11 1992 07:25 | 6 |
| Brian, That didn,t take very long. I have managed to find it without
mailing anyone. I have renamed it to CI780_REV20.bin; now I need
a home for it. I guess RDC$COMMON (CI subdirectory) is a good
place.
Norman.
|