Updated: 24-SEP-2003 (Use your browsers' Reload button to ensure you're viewing the most recent version)
ALPCPU0C09_062 Alpha V6.2 - V6.2-1H3 CPU Diagnostics ECO Summary
Copyright (c) Compaq Computer Corporation 1998, 1999. All rights reserved.
Modification Date: 06-JUL-1999
Modification Type: Documentation Update:
Removed reference to SYS$CPU_ROUTINES_1E05
PRODUCT: OpenVMS Alpha
COMPONENT: CPU Diagnostic Routines
SOURCE: Compaq Computer Corporation
ECO INFORMATION:
ECO Kit Name: ALPCPU0C09_062
ECO Kits Superseded by This ECO Kit: ALPCPU0C08_062
ALPCPUC07_062
ALPCPUC06_062
ALPCPU005_062 aka ALPCPUC05_062
(Only Released Through Factory
Installed Software - FIS)
ALPCPUC04_062
ALPCPUC03_062
ALPCPUC02_062
ALPCPUC01_062
ECO Kit Approximate Size: 17172 Blocks
Kit Applies To: OpenVMS Alpha V6.2, V6.2-1H1, V6.2-1H2, V6.2-1H3
System/Cluster Reboot Necessary: Yes
Rolling Reboot Supported: No
Installation Rating: INSTALL_2
2 - To be installed on all systems running
the listed version of OpenVMS and using
the following feature:
Turbo-Laser systems
KIT DEPENDENCIES:
The following remedial kit(s) must be installed BEFORE
installation of this kit:
None.
In order to receive the all the corrections listed in this
kit, the following remedial kits should also be installed:
ALPLAN05_062 - For RACOR and DE500-BA adapter support
ECO KIT SUMMARY:
An ECO kit exists for the CPU Diagnostic Routines on OpenVMS Alpha V6.2
through V6.2-1H3. This kit addresses the following problems:
Problems Addressed in the ALPCPU0C09_062 Kit:
o A potential, but so far unseen data problem, was fixed.
Specifically, an extra read of the Scatter Gather (S/G) map was
provided to correct this problem.
o On Turbo_Laser systems, the Scatter Gather (S/G) map was not
visible in main memory since it was on the PCIA's RAM instead,
which is just needed for debugging purposes. A copy of the Scatter
Gather into main memory during shutdown was provided to aid in
debugging S/G register crashes.
o During shutdown, a copy of the Scatter Gather (S/G) map would
require some Turbo-Laser users to modify their systems to contain
more non-paged pool. To avoid impact on users, changes were made
to the copying of the S/G map.
o If a Turbo-Laser crashes before the final executive initial
breakpoint during a reboot, the user will see a kernel stack not
valid error.
Problems Addressed in the ALPCPU0C08_062 Kit:
o Add support for the DE500-BA adapter and the Gigabit Ethernet
adapter.
o The RACORE (Token Ring) adapter was never supported in V6.2.
The supported adapter, Thomas Conrad (Token Ring adapter) is
being phased out so this new adapter must be supported.
Problems Addressed in the ALPCPUC07_062 Kit:
o Crash due to TIOP soft corrected adapter error. The crash
footprint is the following:
**** OpenVMS (TM) Alpha Operating System V6.2-1H3 - BUGCHECK
**** ** Code=00000215: MACHINECHK, Machine check while in
kernel mode ** Crash CPU: 00 Primary CPU: 00 Active CPUs:
0000000F ** Current Process = NULL ** Image Name = ****
Starting Memory Dump... **** Memory Dump complete - not all
processes saved
The entry in the errror log would be similar to the following
sample entry:
Logging OS 1. OpenVMS
System Architecture 2. Alpha
OS version V6.2-1H3
Event sequence number 8.
Timestamp of occurrence 09-DEC-1997 09:42:29
System uptime in seconds 184.
Flags x0000
Host name KRYPT2
Alpha HW model AlphaServer 8400 Model 5/350
Entry type 28. Adapter Error
SWI Minor class 8. Adapter Error
SWI Minor sub class 4. TIOP
Software Flags x00008000 Channel Subpacket Present
* TLaser I/O Registers *
Base Phys IO Addr x000000FF8A000000
TLSB Node 8
TLDEV x2000 Turbo-Laser I/O Module
TLBER x00140000 CORRECTABLE READ DATA ERROR
DATA SYNDROME 0
TLFADR0 x00000000 Failing Adr<39:32> x00000000
Failing Command = NOP
LFADR1 x00000000
TLESR0 x0020D9D9 ECC Syndrome 0 x000000D9
ECC Syndrome 1 x000000D9
CORRECTABLE READ ECC ERROR
TLESR1 x00000000
TLESR2 x00000000
TLESR3 x00000000
CPU Interrupt Mask x00000001 Cpu Interrupt Mask = x00000001
ICCMSR x00000000 Arbitration Control Minimum
Latency Mode
Supress Control Suppress after
16 Transations
ICCNSE x80000000 Interrupt Enable on NSES Set
ICCMTR x00000008 Mbox Trans in Prog, Hose 3
IDPNSE-0 x00000000
IDPNSE-1 x00000000
IDPNSE-2 x00000000
IDPNSE-3 x00000006 Hose Power OK
Hose Cable OK
IDPVR x00000810
ICCWTR x00000000
TLMBPR x0000000000000000
IDPDR0 x00000000
IDPDR1 x00000000
IDPDR2 x00000000
IDPDR3 x00000000
Channel Flags x0008 Adapter on Hose 4
-- XMI Subpacket --
Channel Number 3. HOSE 3
Valid Bits x00000FFF
XDEV x0208102A Device Type DWLMA
Dev rev x00000008
XBER x00000202 More Protocol Disabled
Disable XMI Timeout
Failing Commander ID x00000020
XFADR x61880104 Failing Address x21880104
Length x00000001
XFAER x700000FF Mask x000000FF
Address Extension x00000000
FCMD x00000007
LDIAG x81000000 Force Addr Bit <39> to 0
XMI Timeout = 16 ms
ARB Suppress Modes 6 Free Q
locations remaining
Force Parity Errors Normal
Operation, correct parity
asserted
Force Read Responses Normal
Lockout Deasserted
XMI Node Asserting ARB Suppress
DMA Loopback Length Hexword
XMI Node ID x00000008
IMSK x0BFFA070 Unknown Interrupt Value
LEVR x00000800
LERR x00000000 IVINTR Source ID x00000000
Failed Parity Bit(s),
Val x00000000
LGPR x00000000
IPR1 x00000000 Int Pending Node 1 x00000000
Int Pending Node 2 x00000000
Int Pending Node 3 x00000000
Int Pending Node 4 x00000000
Int Pending Node 5 x00000000
Int Pending Node 6 x00000000
Int Pending Node 7 x00000000
Int Pending Node 8 x00000000
IPR2 x00000000 Int Pending Node 9 x00000000
Int Pending Node 10 x00000000
Int Pending Node 11 x00000000
Int Pending Node 12 x00000000
Int Pending Node 13 x00000000
Int Pending Node 14 x00000000
LAMB Err Int Pending x00000000
IIPR x00000000 IPL 14 Int in Prog Node
x00000000
IPL 15 Int in Prog Node
x00000000
IPL 16 Int in Prog Node
x00000000
IPL 17 Int in Prog Node
x00000000
Cmnd Node ID x00000000
o Correctable read error logged as an individual error log entry
as opposed to being logged as an extended correctable read
error.
o Aborted 660 IBOX timeout machine check resulting in a
CPUSPINWAIT crash.
Problems Addressed in the ALPCPUC06_062 Kit:
o CRD errors on the Turbo Laser 4GB memory module (TMEM4) always
contain zero in the failing string register. A hardware
problem in the TMEM4 causes a failure to report the correct
failing string on CRD errors. The inability to determine the
correct failing string prevents DECEVENT from determining the
failing memory SIMM.
Problems Addressed in the ALPCPUC05_062 Kit:
o SMPSTART_MIXEDCPU has been added to the Turbo Laser Link file
so that it will link in the new SMPSTART_MIXEDCPU module which
contains code patches to allow mixing of EV5/EV6 and mixed CPU
passes.
Problems Addressed in the ALPCPUC04_062 Kit:
o The message "***ERROR_ROUTINES_0C05 version 960424-2***" is
displayed during the boot process.
o When a MachineCheck 660/670 occurs while in MachineCheck
620/630 processing, the system will crash and not place the
660/670 information in the errorlog. This means that the
crash dump is the only place that contains the 660/670
information. This fix separates buffers used for 620/630
and 660/670 processing to allow a 660/670 to process and
create an errorlog entry then crash the system as planned.
Problems Addressed in the ALPCPUC03_062 Kit:
o When a MachineCheck 660/670 occurs while in MachineCheck
620/630 processing, the system will crash with an INVEXCEPTN
at SYS$CPU_ROUTINES_0C05_NPRO+07808 and not place the 660/670
information in the errorlog. This means that the crash dump
is the only place that contains the 660/670 information.
Problems Addressed in the ALPCPUC02_062 Kit:
o If a device's EISA configuration utility CFG file requests two or
more IRQs, the device can crash an EISA-based system. This problem
can occur on Digital AlphaServer 2100 5/250, AlphaServer 8400 5/275,
AlphaStation 600 5/266, Digital AlphaServer 1000 4/200 and DEC
2000-300 platforms.
Problems Addressed in the ALPCPUC01_062 Kit:
o Current CPU error routines on AlphaServer 8200 5/300 and AlphaServer
8400 5/300 systems do not allow On-site Service Engineers to
properly diagnose failing components. This ECO kit enhances those
error routines.
According to OpenVMS Engineering, this fix has been included in
OpenVMS Alpha V7.0.
RELATED ARTICLES:
Detailed articles describing the problems listed above may exist in
the OPENVMS database. To view these articles, open the appropriate
product database and perform a query using either of the following
search strings: 'ALPCPU0C09_062', 'ALPCPU0C', or 'ALPCPU'.
ECO KIT ORDERING INSTRUCTIONS:
If after an evaluation you wish to obtain this kit, request it
electronically using the appropriate Advanced Electronic Services
(AES) Service Tool. If you are not familiar with how to request
kits electronically, open the DIA, WIS or DSNLINK database and
review the article entitled:
[AES] How To Electronically Request ECO Kits Using Service Tools
INSTALLATION NOTES:
This kit cannot be booted from a user prompt because it contains a fix
for a problem that prevents OpenVMS from booting. The problem occurs
when devices that use PCI-to-PCI bridged devices are used on Turbo-Laser
systems. To correct this problem, perform the following steps:
o Remove the devices from the system.
* the PCI-PCI devices for a Turbo-Laser system are:
- ATM350
- KZPDA
o Reboot the system
o Use the the VMSINSTAL utility to install the TIMA kit.
- log into the SYSTEM account and type the following at the
DCL prompt:
@SYS$UPDATE:VMSINSTAL ALPCPU0C09_062 [saveset location]
The saveset location may be a tape drive or a disk directory
that contains the kit saveset.
o Shut down the system.
o Re-install the PCI-PCI devices
o Re-boot the system
The images in this kit will not take effect until the system is rebooted.
The fix only applies to Turbo-Laser systems with PCI-to-PCI bridged
devices installed in them. If you have other Turbo-Laser nodes with
these devices in your OpenVMS cluster, they must also be rebooted in
order to make use of the new image(s). If the Turbo-Laser systems do
not use a shared system disk, before each system is rebooted, the above
installation procedure must be performed.
As an alternative to removing the PCI-to-PCI bridged devices, you can
install the TIMA saveset to the destination disk from a working system.
To do this, the destination disk must be a shared disk or be manually
relocated from the destination system to the working system for
installation purposes.
The images in this kit will not take effect until the system is rebooted.
==========================================================================
| Table of Kit Image Information |
+----------------------------+----------+-----------------+--------------+
| | Overall | Image File | Image Link |
| Image Name | Checksum | Identification | Date/Time |
+----------------------------+----------+-----------------+--------------+
| SYS$CPU_ROUTINES_0C05.EXE | D1A92FC4 | X-5 | 2-SEP-1998 |
| | | 17:47:45.08 |
+----------------------------+----------+-----------------+--------------+
|
