Power6 High-End System Firmware

12/01/10

System firmware changes that affect all systems

• A problem was fixed that caused the HMC to show a status of "Incomplete" for the managed system, and numerous service processor dumps to be generated.
• The firmware was enhanced to log SRC B181D30B as informational instead of predictive.
• The firmware was enhanced to list the attached devices when viewing the adapter information for a partition profile on the HMC GUI.
• A problem was fixed that caused the HMC2 port on the advanced system management interface (ASMI) to erroneously default to static IP addressing instead of dynamic.
  

• A problem was fixed that prevented the timed-power-on function from turning the system back on if the service processor's clock was adjusted to an earlier time.  This problem could occur during the fall when clocks are set back when daylight savings time ends, for example.
• A problem was fixed that caused a partition to fail to reboot, or fail to boot if it had been shut down once since the platform was booted, with SRC B2001230 and word 3 = 000000BF.  This failure can be seen on a partition that owns a PCI, PCI-E, or PCI-X slot.
  

05/24/10

System firmware changes that affect all systems

• The firmware was enhanced to improve the callouts when NVRAM corruption is detected in the bulk power controller's (BPC's) service processor.
  

• A problem was fixed that caused the system to crash with SRC B7000103 when a concurrent maintenance operation was performed on an I/O slot directly from a partition (using AIX SMIT or IBM i HST).
• A problem was fixed that caused a system or partition running Linux to crash when the "serv_config -l" command was run.
  
  

04/07/10

System firmware changes that affect all systems

• DEFERRED:   This fix corrects the handling of a specific processor instruction sequence that has the potential to result in undetected data errors.  This specific instruction sequence has only been observed in a small number of highly tuned floating point-intensive applications.  However, it is strongly recommended that this fix be applied to all POWER6 systems.  This fix has the potential to decrease system performance on applications that make extensive use of floating point divide, square root, or estimate instructions.
• A problem was fixed that prevented an SRC from being recorded in the service processor dump produced by a host-initiated reset.
• A problem was fixed that caused the system to become unresponsive and appear to hang  when page migration occurred on a PCIe slot.

• A problem was fixed that caused a virtual SCSI or virtual fibre channel adapter to be seen by the operating system as not bootable when it was added to a partition using a dynamic LPAR (DLPAR) operation.
• In partitions running AIX or Linux, a problem was fixed that caused the addition of an I/O slot to a partition using a dynamic LPAR (DLPAR) add operation to fail.
• On systems running redundant VIOS partitions, a problem was fixed that prevented Ethernet traffic from being properly bridged between the two partitions.  This problem also prevented shared Ethernet adapter failover from working correctly.
• On systems using InfiniBand switches for processor clustering, a problem was fixed that caused InfiniBand ports to intermittently drop out.
  

12/16/09

System firmware changes that affect all systems

• HIPER:  A problem was fixed that might cause the system to crash if the server is running AIX  and has a F/C 5802 or 5877 drawer (in a 19" rack), or F/C  5803 or 5873 drawer (in a 24"rack), attached.
• A problem was fixed that prevented an automatic reboot after some types of memory-related crashes.
• The firmware was enhanced such that SRCs B181F126, B181F127, and B181F129 are correctly handled, and no longer cause unnecessary calls home to be made.
• A problem was fixed that caused the system to hang with SRCs B182953C, B182954C, and B17BE434 being logged.
• A problem was fixed that caused SRC 10009135, followed by 10009139, to be erroneously logged.  These SRCs indicate a system power control network (SPCN) loop is being broken, then re-established. 

• On a single system running Oracle in multiple partitions, with multiple IBM LHCAs connected in the same subnet, a problem was fixed that caused the remaining partitions to lose their reliable datagram socket (RDS) heartbeat connections after the reboot of a single partition.  There is a greater probability of encountering this problem if the partition being rebooted has a large partition memory assigned to it.

09/23/09

System firmware changes that affect all systems

• DEFERRED:  The firmware was enhanced to eliminate correctable errors (CEs) being erroneously logged against the memory bus with SRC B124E504.  This change affects only 9117-MMA systems equipped with 4.2GHz quad core processor cards (FC 7540) and all 8234-EMA systems.  This change is not critical.
• The firmware was enhanced such that SRC B181F126 is correctly managed, and no longer calls home  unnecessarily for this problem.

08/20/09

System firmware changes that affect all systems

• DEFERRED:  This fix corrects the handling of a specific processor instruction sequence that was generated on a particular heavily-tuned High Performance Computing (HPC) application. This specific instruction sequence has the potential to produce an incorrect result. This instruction sequence has only been observed in a single HPC application.  However, it is strongly recommended that you apply this fix. 

• A problem was fixed that under certain rare circumstances caused a partition to crash when a 24" InfiniBand I/O drawer (feature code 5797 or 5798) drawer was concurrently added.  When this problem occurred, rebooting the system was required to recover.
• On systems running system firmware ES340_075 and Active Memory Sharing, a problem was fixed that might have caused a partition to lose I/O entitlement after the partition was moved from one system to another using PowerVM Mobility.
• On systems running system firmware ES340_075 and Active Memory Sharing, a problem was fixed that might have caused a partition to fail to boot with SRC B700F103 if the partition had more than 24 virtual processors assigned to it.
• On systems with F/C 5802 or 5877 drawers attached, a problem was fixed that prevented an I/O slot's power LED from accurately reflecting the state of the I/O slot in a 5802 or 5877 drawer, under certain circumstances.
• On systems with external I/O towers attached, the firmware was enhanced so that the system will not crash when SRC B7006981 is logged for certain types of I/O hardware failures. 

06/22/09

New features and functions:

- DEFERRED: Support for F/C 5803 (24" I/O drawer) and F/C 5873 (diskless 24" I/O drawer).

Attention: After this level of firmware is installed, the platform must be powered off, then powered on, before the 5803 or 5873 I/O drawer is added to the system.

- DEFERRED: Support for POWER VM Active Memory Sharing.

Attention: After this level of firmware is installed, the platform must be powered off, then powered on to activate the POWER VM Active Memory Sharing function.

Attention: If ES340_075 has been installed, and the new POWER VM Active Memory Sharing function has been activated, and you want to back-level the system firmware, the active memory sharing pool must be deactivated and deleted prior to back-leveling the system firmware. IBM does not recommend back-leveling the system firmware.

System firmware changes that affect all systems:

• A problem was fixed that caused hardware to be deconfigured when the system encountered network errors, even though the SRCs were being logged as informational.
• A problem was fixed that caused the detailed data at the end of an "early power off warning type 5" AIX error log entry to be filled with invalid data instead of zeros.
• On systems running firmware release ES340, a problem was fixed that caused data in the platform dump to be invalid.
• The firmware was enhanced so that SRCs B181720D, B1818A13, and B1818A0F, and occasionally a service processor dump, will not be generated when the service processor's two Ethernet interfaces are on the same subnet. (This is an invalid configuration.)
• The firmware was enhanced such that error logs with relevant information will be created when a system crashes under certain circumstances, rather than a generic SRC (B1813410), with very little debug information, being logged.
• A problem was fixed that caused the system to hang when terminating if the system had been in power save mode.
• The firmware was enhanced so that a call home will be made if the hypervisor issues a "terminate immediate" interrupt.
• A problem was fixed that caused incorrect field replaceable unit (FRU) part numbers to be returned for the BPC scroll assembly, UEPO panel and the CEC MDA scroll assembly.
• The firmware was enhanced so that the service processor only logs SRC B1A38B24 when a valid network setup error is found. The callouts for this SRC were also improved.
• A problem was fixed that caused a system with I/O drawers attached to crash, and a SYSDUMP to be taken, with SRCs B7000103 and SRC B181D138 being logged.  Another symptom of this failure is informational SRC B7006970 entries constantly posting in the iqyylog.log.

• In systems using InfiniBand switches for processor clustering, a problem was fixed that caused packets to be dropped under certain circumstances.
• A problem was fixed that caused the migration of a partition with more that 900 virtual slots defined, from a system running firmware ES320 to a system running firmware ES340, to fail.
• On systems using on/off (temporary) memory capacity on demand (COD), the firmware was enhanced to improve memory COD's interaction with other tools (such as Inventory Scout in AIX), and to make the billing process easier. 

04/20/09

System firmware changes that affect all systems:

• HIPER: The firmware was enhanced to improve the service processor's capability to recover from bad bits in the flash memory. A predictive error, or an unrecoverable error, will be logged against the card that contains the system firmware if the number of correctable or uncorrectable errors exceeds the threshold.
• HIPER: A problem was fixed that caused nodes to guard out processor cores, or checkstop, during the transition to nominal voltage from "power save" mode.
• A problem was fixed that caused the service processor diagnostics to report a "TOD (time-of-day) overflow" error, instead of an uncorrectable memory error, when failures occurred on memory DIMMs.
• A problem was fixed that prevented the service processor from automatically booting from the permanent (or P) side if the temporary (or T) side of the firmware flash was corrupted. When the problem occurred, the service processor stopped instead of booting from the P side.
• A problem was fixed that might have caused the system to crash when a processor was dynamically removed when the system was running.
• The firmware was enhanced such that VPD data corruption in the Anchor (VPD) will be corrected by the firmware, rather than having to have the Anchor card replaced.
• A problem was fixed that caused the system to crash, under certain circumstances, with SRC B112E504 being logged, followed by SRC B181C350, when a system dump was initiated.
• A problem was fixed that caused a partition being migrated to crash on the target system.
• On systems running the ES340 release of system firmware, a problem was fixed that caused an abort code to be logged in the virtual input/output system (VIOS) error log on the source system after a successful partition migration.
• A problem was fixed that caused a partition being migrated to become unresponsive on the target system when firmware-assisted dump was enabled.
• The firmware was enhanced so that SRC BA210012 will not generate a call home when logged.
• A problem was fixed that caused hardware to be deconfigured when the system encountered network errors, even though the SRCs were being logged as informational.

• On systems with external I/O drawers, a problem was fixed that could cause the system to hang on checkpoint C700406E during a "warm" reboot (a reboot in which the processor drawer is power-cycled but the I/O drawers are not).
• On systems with virtual fiber channel disks, a problem was fixed that prevented the system management services (SMS) from displaying the virtual fiber channel disks if the virtual fiber channel server reported that any of them were reserved. 

01/16/09

System firmware changes that affect all systems:

• HIPER: A problem was fixed that caused nodes to guard out processor cores, or checkstop, during the transition to nominal voltage from "power save" mode.
• The firmware was enhanced so that SRC B1xx3409, which indicates an invalid state change (such as pushing the power on button twice quickly) will be logged as informational instead of predictive, and will not call home.
• A problem was fixed that caused a service processor dump to be taken and SRC B181EF88 to be logged, even though the operation of the system was not affected.
• A problem was fixed that caused the location codes for multi-port PCI adapters, such as the 4-port Ethernet adapters, to be incorrect.
• On systems with multiple host channel adapter (HCA) cards, a problem was fixed that logical ports on the HCA cards to be intermittently inactive.
• A problem was fixed that caused the system to crash and a SYSDUMP to be taken, with SRCs B170E540, 

B181D138, or B700F105, with a bad PCI-E adapter installed and in use, or while running a heavy network load.

• In networks using a time server, a problem was fixed that caused the date on a client system to be reset to 1969 if the client system lost power.