Power7 Mid-Range Systems Firmware

System firmware changes that affect all systems

• HIPER:  IBM testing has uncovered a potential undetected data corruption issue.  The problem can occur in rare instances due to an issue in the firmware and is most likely to impact hypervisor data.  This issue was discovered during internal IBM testing, and has not been reported on any customer system.  However, IBM recommends that  customers running on POWER7 systems with Ax720_090 and earlier firmware move to Ax720_101.  POWER7 systems running with Ax710 firmware do not have an exposure to this issue, so no action is recommended.
•  HIPER:  A problem was fixed that caused the hypervisor to delay dispatching a partition even though it was ready to run, which added latency (delays) that adversely affected performance.  This problem can affect POWER7 systems running any level of Ax720 firmware prior to Ax720_101.
• A problem was fixed that caused certain service processor error log entries with a severity of "predictive", and a failing subsystem of "service processor firmware", to be erroneously converted to "informational".
• A problem was fixed that caused three B181951C SRCs to be erroneously logged, and the system IPL time to increase by as much as an hour.  This problem is more likely to occur on systems with firmware level AL720_082 or AL720_090, AM720_084 or AM720_090, or AH720_082 or AH720_090 installed.
• A problem was fixed that caused the EnergyScale firmware to erroneously go into safe mode when processor 0 was guarded out.
• A problem was fixed that caused SRC B1812A61 to be erroneously logged.
• A problem was fixed that prevented the setting of the boot diagnostic level in the power on/off menu (in the Advanced System Management
  Interface (ASMI)) from being shown correctly after it was changed.
  
• A problem was fixed that prevented a system dump from being off-loaded from the service processor.  When this occurred, additional dumps were not allowed.
  
• The firmware was enhanced so that a message is displayed if setting the brand keyword in the ASMI menu (System Configuration -> Program Vital Product Data -> System brand) fails because the service processor is not in the correct state.
• The firmware was enhanced such that a call home is not made when an error logged by the system controller, node controller, or service processor is informational, or recovered, and the reset/reload bit is set.
• A problem was fixed that caused multiple DR_DMA_MIGRATE_FAIL entries in the AIX error log.
  
• A problem was fixed that caused SRC B7000803 to be erroneously logged multiple times.
• A problem was fixed that prevented processor resources from being moved to another partition by a DLPAR (dynamic LPAR) operation.
• A problem was fixed that prevented partitions from booting.
• A problem was fixed that caused the HMC component interval activity report to always show 100% uncapped CPU available.
• A problem was fixed that caused incorrect data to be displayed in the "Deconfiguration Records" menu option on the ASMI (System Service Aids > Deconfiguration Records) when a service processor was guarded out.
• A problem was fixed that prevented the battery on the secondary service processor from being called out when it needed to be replaced.
• The firmware was enhanced to log SRC 11007610, 11007620, 11007630, 11007640, or 11007650 only when a system fan's speed drops below 2800 RPM.
• A problem was fixed that prevented the system fans from running at the correct speed if a service processor reset to runtime was done, then a fan failure occurred.
• A problem was fixed that caused the green power enclosure LED to be off, instead of blinking at a slow rate, when the system is at standby.
• On systems with AM710 system firmware, a problem was fixed that caused utility capacity on demand (COD) processors to erroneously become enabled when a firmware upgrade was done to AM720.
• A problem was fixed that caused an administrative service processor fail-over (AFO), followed by another AFO without a reset in between, to fail.  If this occurred during a concurrent hot add or repair maintenance operation on a 9117-MMB or 9179-MHB, the operation failed.
• A problem was fixed that caused VIOS partitions to fail to boot.
• A problem was fixed that caused a partition suspend operation to hang.  When this problem occurred, all subsequent suspend operations were locked out as well.
• A problem was fixed that could cause the target partition to crash after a successful P6 to P7 partition migration.  Possible AIX error log entries include:  label: DSI_PROC, resource:  SYSVMM, with description: "DATA STORAGE INTERRUPT, PROCESSOR".  Other partition-related crash descriptors may also be logged.
• A problem was fixed that could cause AIX error log entries following a successful partition migration.  Possible AIX error log entries include: label: RTAS_ERROR, resource: sysplanar0, with description: "INTERNAL ERROR CODE".  Other errors may also be logged.
• A problem was fixed that caused the installation of some versions of Linux to fail.
  

• On systems with two HMCs attached, a problem was fixed that caused one of the HMCs to frequently go to an incomplete state.
• On systems running IBM i partitions, a problem was fixed that caused a RAID array of SCSI disks to be exposed if an MES upgrade was done, or a system plan was created.
• On systems running IBM i partitions, a problem was fixed that caused SRC BA040030 to be erroneously logged, and a call home to be made, even though the partition booted successfully.
• On systems using the host Ethernet adapter (HEA) function, a problem was fixed that caused the HMC to erroneously report that deleting a logical port had failed.
• On partitions running Advanced Memory Sharing (AMS), a problem was fixed that prevented shutdown of a partition when all paging VIOS's servicing the partition were hung and unable to complete outstanding I/O operations.
• On systems running Advanced Memory Sharing (AMS), a problem was fixed that caused an AMS partition to crash with SRC B700F103.  This problem may occur when reducing the size of the AMS pool (or doing a hot node repair on a model MMB or MHB) at the same time as dynamically creating an AMS partition, or changing an AMS partition's maximum memory.
  
• A problem was fixed that caused AIX licensing issues when migrating a partition from a POWER6 to a POWER7 system.
• The "USB Service Functions" option was removed from the ASMI menus on 9117-MMB, 9179-MHB and 9119-FHB systems, which do not support this function.
  
• On systems with a F/C 5802 or 5877 I/O expansion drawer, a problem was fixed that caused SRC 10003144 or 10003154 to be erroneously logged when a repair was done on the I/O drawer.
• On systems with a F/C 5802 or 5877 I/O expansion drawer, a problem was fixed that caused the lamp test on the HMC to turn off all of the LEDs when the test was complete instead of returning them to their original states.
  
• On stand-alone systems running AIX or Linux, and on systems managed by IVM (Integrated Virtualization Manager), a problem was fixed that prevented platform dumps from being off-loaded, or resulted in corrupted or incomplete platform dumps.
• A problem was fixed that caused the exhaust heat index value displayed by IBM Director to be invalid when the system is located near, at, or below sea level.
• On systems on which a NIM installation is being set up using the system management services (SMS) menus, the firmware was changed to limit the packet size options to 512 and 1024 bytes.
• On systems with Selective Memory Mirroring and the Enhanced Cache Option enabled, a problem was fixed that caused unpredictable system behavior when a processor hardware failure occurred.
  

• A problem was fixed that caused the hot repair of a GX adapter to fail, and the system to crash, if the GX adapter had previously logged a non-checkstop type of error.
  
• A problem was fixed that caused the system to crash with SRCs B170E540, B181F02D and B700F103 during a hot node upgrade (memory), or hot node repair, of node A.
  

• HIPER: IBM testing has uncovered a potential undetected data corruption issue when a mobility operation is performed on an AMS (Active Memory Sharing) partition.  The data corruption can occur in rare instances due to a problem in IBM firmware.  This issue was discovered during internal IBM testing, and has not been reported on any customer system.
  
• On systems with a F/C 5802 or 5877 I/O drawer attached, a problem was fixed that caused a partition to crash during a page migration operation.
• On systems with a F/C 5723 communications adapter, a problem was fixed that prevented the adapter from being seen by partition firmware (PFW) if the adapter was not in the first 144 slots that are probed by PFW.
  

New Features and Functions

• Support for partition suspend/resume.  AIX 61 TL6 SP3 or later, or AIX 71 TL0 SP2 or later, is required for partition suspend/resume.
• CEC Hot Node Add & Repair Maintenance (CHARM) support on Power 770 and Power 780 systems.

System firmware changes that affect all systems

• HIPER:  On systems using the HEA (host Ethernet adapter) function, and on which a CEC concurrent maintenance operation that requires a node evacuation is being done, this fix corrects an issue that has the potential to corrupt information stored in the system memory, which may cause undetected data errors.  This issue was discovered during internal IBM testing, and while it has not been reported on any customer systems, IBM strongly recommends that this fix be applied to all model MMB systems that are running AIX partitions.
• HIPER:  A problem was fixed that caused repeated reset/reloads of the service processor, and fail-overs, to occur after a hypervisor-initiated reset/reload of the service processor was completed.  That led to loss of communication between the service processor and the hypervisor (indicated by SRC B182951C).
• A problem was fixed that caused disks that were not bootable to be displayed in the system management services boot menus.  This problem also prevented the operating system level from being displayed for bootable hard disks in the system management services boot menus.
  
• A problem was fixed that caused an error log indicating a dynamic LPAR (DLPAR) error when no DLPAR operations were done, and unrecoverable SRCs BA180010 and BA250010 to be erroneously logged, when a recoverable enhanced error handling (EEH) error was logged on an I/O adapter.
• The firmware was enhanced to use the fan speed signal, as well as the fan present signal, to determine if fans are present in a drawer.  This change keeps the firmware from shutting down a drawer, and logging 110076x1 SRCs, if the fans are functional but the fan presence signals are corrupted.
• A problem was fixed that caused a service processor reset/reload, a service processor dump to be taken, and B181EF88 to be logged.
• A problem was fixed that caused the managed system to go to the incomplete state on the HMC.
• A problem was fixed that caused the system to hang at C700406E during boot.
• A problem was fixed that caused the platform to become unresponsive; this was indicated by an "incomplete" state on the HMC.  When this problem occurred, the partitions on the managed system became unresponsive.
• A problem was fixed that caused SRC B1561111 to be erroneously logged, and the control (operator) panel to erroneously deactivated, if there is no activity on the control panel for several weeks.
  

• On systems with a solid state disk drive (SSD), the fan speeds were increased to provide additional cooling to the SSD drives.
• 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.
• A problem was fixed that caused the system ID to change, which caused software licensing problems, when a live partition mobility operation was done where the target system was an 8203-E4A or an 8204-E8A.
  
• PARTITION-DEFERRED:  A problem was fixed that caused SRC BA210000 to be erroneously logged on the target system when a partition was moved (using Live Partition Mobility) from a Power7 system to a Power6 system.
• A problem was fixed that caused SRC BA280000 to be erroneously logged on the target system when a partition was moved (using Live Partition Mobility) from a Power7 system to a Power6 system.
• A problem was fixed that caused a partition to hang following a partition migration operation (using Live Partition Mobility) from a system running Ax720 system firmware to a system running Ex340, or older, system firmware.
• A problem was fixed that caused a system or partition running Linux to crash when the "serv_config -l".
  
• On systems using the HEA broadcast/multicast application to send and receive millions of packets, such as video streaming, the packet storm mitigation algorithm was enhanced so that a packet will only be dropped when a packet storm is detected.
  
• A problem was fixed that caused a partition to fail to reboot 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.
• On systems with a F/C 5802 or 5877 I/O drawer attached, and a PCI-E adapter in the CEC, a problem was fixed that caused the system to crash during a page migration operation with SRC B700F103.
  

• A problem was fixed that caused the system to hang with SRC B170E540 during a node repair operation.

New Features and Functions

• Support for autonomic IPL, which allows the service processor to decide which diagnostic tests to run at boot time.
• Support for the network installation of the IBM i operating system from the hardware management console (HMC) command line interface (CLI).
• Support for the 7216-1U2 media drawer.
  
• Support for VIOS storage integration.
• Support for CEC Hot Node Add & Repair Maintenance (CHARM) operations on Power 770 and Power 780 systems.  However, IBM recommends installing firmware service pack 7.2.1 (AM720_084) before starting a CHARM operation on a Power 770 or Power 780 system.
  

System firmware changes that affect all systems

• HIPER: A problem was fixed that caused the HMC to show the server's status as incomplete, and SRC B7000602 to be logged against SFLPHMCCMDTASK in serviceable events.  This problem can also cause the system to crash when it occurs.
• HIPER: A problem was fixed that caused an AIX or Linux partition to fail to boot with SRC B2008151, which prevented further access to that partition, and potentially preventing prior LPAR configuration changes from being completed.  A reboot of the server is required to recover from this problem.
• HIPER: A problem was fixed that could have prevented a successful emergency fail-over to the backup service processor if multiple reset/reload commands were issued to both service processors at roughly the same time.  If the emergency fail-over was not successful, the system might have been terminated by either the service processor or the hypervisor.  This problem did not affect an administrative fail-over of the service processor.
• HIPER: A problem was fixed that caused informational SRC B70069DA from a host Ethernet adapter (HEA) to be logged erroneously.  These messages are sent from the hypervisor to the service processor and cause unnecessary loading of the hypervisor-service processor communication link.
• HIPER: A problem was fixed that caused a partition, or the hypervisor, to appear to hang, then recover; the time of the apparent hang varied.  SRC B182953C might also be logged.
  
• A problem was fixed that prevented an SRC from being recorded in a service processor dump produced by a host-initiated reset.
• The wording on the memory deconfiguration menu on the advanced system management interface (ASMI) was enhanced to better differentiate between guarded resources and deconfigured resources.
• The firmware was enhanced to improve the hardware called out with SRC B121B8AB.

• On systems running host Ethernet adapter (HEA), a problem was fixed that caused unrecoverable SRCs BA154050 and BA154070 to be erroneously logged.
• On systems with a F/C 5802 or 5877 I/O drawer attached, the firmware was enhanced to allow the drawer to power on with only one working offline converter assembly (OCA).
• On 9179-MHB systems running the enhanced cache option (ECO), a problem was fixed that periodically caused the system fans to run at full speed.