• Tailing space issue of SNMP MIB of cpqSePCIeDiskModel data has been resolved

• Fixed race condition between BIOS checking link status and LNI completing which resulted in PXE boot failures.
• Changed the default behavior of the hfi1_eprom tool to convert the eprom format to version 2, eliminating the need to supply the -N command line argument.
• Addressed an issue where updating the HfiPcieGen3Loader*.rom file could result in the erasure of HfiPcieGen3Loader*.efi and HFI_TYPE1*.dat files.

• The following issue has been fixed in Unified Extensible Firmware Interafce (UEFI) ROM: On some platforms, the hfi1 device was not showing up in BIOS/UEFI boot menus and was not available as a PXE boot device. This was caused by the platform not loading the UEFI driver for the hfi1 adapter. 

• Tailing space issue of SNMP MIB of cpqSePCIeDiskModel data has been resolved

• Tailing space issue of SNMP MIB of cpqSePCIeDiskModel data has been resolved

• The redundant freeing of a list item could lead to memory corruption, potentially causing the application to crash or incorrect traffic handling.
• The host driver probe did not check whether there are existing SFs which are present in the device, causing the host driver to not recreate those SFs.
• The kernel went into panic mode if there were multiple encapsulations and not all neighbors were valid.

• Creating multiple steering rules that modify a packet and match on the same packet headers could cause an error to be displayed in dmesg when deleting the steering rules.
• Some IB spec QP state behaviour on post_send()/recv() was not being fully enforced. The fix makes the QP complaint to IB spec about when it is allowed to post_send()/recv() and when it should return an error.
• When changing trust state from PCP to DSCP, the TC number changed by default to 8, in some cases, disrupting traffic prioritization if trust state was changed back to PCP.
• The minimum Tx rate limit was not supported with link speed of 1Gb/s.
• Destroying mlxdevm group while SF was attached to it was not supported.
• Using OVS offload with NIC mode (non switchdev mode) caused traffic to drop.
• In some cases VF metering configuration failure caused a deadlock.
• A race condition occasionally caused some connection aging to set to 24 hours instead of 30 seconds.

• The redundant freeing of a list item could lead to memory corruption, potentially causing the application to crash or incorrect traffic handling.
• The host driver probe did not check whether there are existing SFs which are present in the device, causing the host driver to not recreate those SFs.
• The kernel went into panic mode if there were multiple encapsulations and not all neighbors were valid.

• iLO based deployment failure due to empty <products> tag in component XML file.

The following issues have been fixed in version 5.7:

• Creating multiple steering rules that modify a packet and match on the same packet headers could cause an error to be displayed in dmesg when deleting the steering rules.
• Some IB spec QP state behaviour on post_send()/recv() was not being fully enforced. The fix makes the QP complaint to IB spec about when it is allowed to post_send()/recv() and when it should return an error.
• When changing trust state from PCP to DSCP, the TC number changed by default to 8, in some cases, disrupting traffic prioritization if trust state was changed back to PCP.
• The minimum Tx rate limit was not supported with link speed of 1Gb/s.
• Destroying mlxdevm group while SF was attached to it was not supported.
• Using OVS offload with NIC mode (non switchdev mode) caused traffic to drop.
• In some cases VF metering configuration failure caused a deadlock.
• A race condition occasionally caused some connection aging to set to 24 hours instead of 30 seconds.

• Aligned RDE behavior to DSP0266 v1.15.0  table 23.

The following issues have been fixed in version 20.35.1012:

• RDE (Redfish) PATCH operation to LLDPTransmit properties "ManagementAddressIPv4", "ManagementAddressIPv6" and "ManagementAddressMAC" were applied only in the first attempt, but failed in the next.
• The RXT E2E inflight went into an unresposive state occasionally.
• RDMA partition was reported even if NIC+RDMA mode was disabled.
• PLDM AEN event receiver media was changed unexpectedly and destination BDF was overridden with garbage when some PLDM packet were received from the SMBus layer.
• An incorrect address issue for XRQ Event Backend Controller Px Syndrome event in multi-host system during FLR.
• PCIe SKP OS generation interval issue for Gen1 and Gen2.
• InfiniBand L2 QP could not receive RDMA traffic occasionally.
• Running with a debug firmware reduced security as if token was applied.
• PLDM AEN event receiver was required to be cleared on PCIe reset in case the media type was MCTP over PCIe VDM.
• Asynchronous messages were sent over MTCP before endpoint discovery was done.

• Aligned RDE behavior to DSP0266 v1.15.0  table 23.

• Fixed an issue that caused a ParaVirtual VM to lose traffic after resuming from the suspend mode when SR-IOV was enabled in the system.
• Cleaned steering rule from FDB when VF is being quiesced.
• Fixed an issue that caused the driver to configure the incorrect SL value for RoCE with RDMACM.

• Fixed a compatibility issue with VMware Update Manager as it wouldn't accept a bundle with metadata xml with old versioning scheme. The metadata xml now contains the new versioning scheme.

• Fixed an issue that caused a ParaVirtual VM to lose traffic after resuming from the suspend mode when SR-IOV was enabled in the system.
• Cleaned steering rule from FDB when VF is being quiesced.
• Fixed an issue that caused the driver to configure the incorrect SL value for RoCE with RDMACM.

• Fixed an issue that caused a ParaVirtual VM to lose traffic after resuming from the suspend mode when SR-IOV was enabled in the system.
• Cleaned steering rule from FDB when VF is being quiesced.
• Fixed an issue that caused the driver to configure the incorrect SL value for RoCE with RDMACM.

• Aligned RDE behavior to DSP0266 v1.15.0  table 23.

The following issues have been fixed in version 20.35.1012:

• RDE (Redfish) PATCH operation to LLDPTransmit properties "ManagementAddressIPv4", "ManagementAddressIPv6" and "ManagementAddressMAC" were applied only in the first attempt, but failed in the next.
• The RXT E2E inflight went into an unresposive state occasionally.
• RDMA partition was reported even if NIC+RDMA mode was disabled.
• PLDM AEN event receiver media was changed unexpectedly and destination BDF was overridden with garbage when some PLDM packet were received from the SMBus layer.
• An incorrect address issue for XRQ Event Backend Controller Px Syndrome event in multi-host system during FLR.
• PCIe SKP OS generation interval issue for Gen1 and Gen2.
• InfiniBand L2 QP could not receive RDMA traffic occasionally.
• Running with a debug firmware reduced security as if token was applied.
• PLDM AEN event receiver was required to be cleared on PCIe reset in case the media type was MCTP over PCIe VDM.
• Asynchronous messages were sent over MTCP before endpoint discovery was done.

• Aligned RDE behavior to DSP0266 v1.15.0  table 23.

Firmware Dependencies:

Problems Fixed:

Addressed an issue where system may hang at POST during TPM measurement after restoring manufacturing defaults.

Address an issue where the RegistryVersion format in the URI redfish/v1/registrystore/registries did not follow the Redfish Specification. If scripts were written to conform with the incorrect format in previous System ROMs, those scripts may need to be modified to follow the correct, Redfish compliant format which is implemented in this System ROM revision and later.

Addressed an issue where the System Configuration (RBSU) option "Access Control Service" did not disable PCIe Access Control Service when it’s set to disabled.

Known Issues:

None

Firmware Dependencies:

None

Problems Fixed:

This revision of the System ROM includes the latest revision of the AMD reference code for AMD 4th Generation EPYC processors which addresses an issue of potential over voltage condition early in the boot sequence. This issue is not unique to HPE servers.

Known Issues:

None

Firmware Dependencies:

Problems Fixed:

Address an issue that can result in a Red Screen of Death (RSOD) and hang during system boot or a hang during system boot without an RSOD. This issue can occur on systems utilizing the v1.32 or v1.40 System ROM revision and iLO Firmware revision v1.50. The v1.32 System ROM added a boot-time reduction capability (which also required iLO Firmware revision v1.50) which resulted in the boot-time RSOD and/or hang. This capability has been removed in this revision of the System ROM.

Addressed an issue where one-time boot did not work properly after setting one-time boot device from iLO web or via Redfish.

Known Issues:

Firmware Dependencies:

Problems Fixed:

This revision of the System ROM includes the revision of the OpenSSL library 1.1.1u update which provides mitigation for and security vulnerabilities documented as CVE-2023-2650.

Addressed an issue where the unsupported DIMM population IML will report an incorrect processor number on dual-processor systems.

Addressed an issue where the SMBIOS type 197 might contain an incorrect format value of xGMI speed.

Addressed an issue where the CXL root bridge was missing from SMBIOS type 9.

Addressed an issue where dmidecode indicates 4 power supplies are installed when only 2 power supplies are actually installed.

Addressed an issue where number of "Internal L3 cache" displayed under RBSU information page is not correct.

Addressed an issue where the System Health LED is illuminated Red for all memory configuration error cases. Some memory configuration error cases are not critical and should have the System Health LED illuminated Amber.

Known Issues:

Firmware Dependencies:

Problems Fixed:

Updated Server Platform Services (SPS) Firmware to 06.00.04.033.0 to align with the Intel Best Known Configuration (BKC) WW09.

Known Issues:

Temperature Misreading: the adapter was reading incorrect reading of the temperature on the HPE SN1610Q 32Gb Fibre Channel Host Bus adapter's Small Form-factor Pluggable. The symptom would result in a temperature spike and potentially reset the system

• Unsuccessful read of the SFP Temperature which will trigger a Server Shutdown. 
• Message related to a Mailbox 0x48 Firmware dump saved to buffer.

• The Downstream Port Containment (DPC) was not exposed on the downstream ports of the top level PCIe switch in products supporting PCIe switch.
• PRBS lock loss (PRBS_CHK_ERR_CNT_NO_CLR field is raising) occurred when the PRBS mode was first configured on the ConnectX-7 adapter card and then on the Wedge400 switch.
• wqe_based_steering CQEs were not generated upon an error.
• Aligned RDE behavior to DSP0266 v1.15.0  table 23.
• Device link went down, and the device was not able to get traffic, when moving between two states DETECT and POLLING CONFIG in RTL.
• When connecting a ConnectX-7 adapter card to ConnectX-7 adapter card and one side was configured to RM Loopback, and the port was toggled, link flap was experienced occasionally. 
• Live migration of MPV affiliated function pair was not supported when port numbers were changed. Each function should stay on the same port number as before migration.
• When decapsulation on a packet occurred, the FCS indication was not calculated correctly.
• SPDM Get Measurements returned an invalid signature while executed without the included measurements (request param2 = 0).
• The SPDM challenge command returned the hash of all the measurements without their headers.
• Pkey table was not updated, and wrong value was sent, when the MADs handled in a long process were sent using GLOBAL_GVMI instead of vport0_gvmi.
• A wrong lane mapping to serdes when selecting the OSFP port and using only 4 lanes.

• Initial release

• Firmware crashed when setting large payload length values (more than ~1500) in NC-SI command's header.
• Interface type was shown as "unsupported" in CMIS modules.
• Fixed the configuration of the TS1s sent by the DownStream port (DSP) when moving to EQLZ.ph2.
• PLDM firmware update process failed in case 1304 bytes chunk size is chosen.
• An issue that occurred during secure firmware update when decrypting and authenticating each chunk of data using its authentication tag. The issue appeared when the main code chunk was split between the user chunks and any GCM operation (e.g., flash read with decryption). This GCM operation broke the GCM context for main chunk authentication and therefore failed. 
• CNP received, handled, and ignored counters in the hardware counters could not work after moving to Programmable Congestion Control mode.
• Hardware failed to handle an error flow that occurred when accessing the DPA cluster L2 cache from the firmware processor. In this case the firmware processor hardware requested a VA=>PA translation from the internal mmio, and the address translation was broken by the mmio on the 4K page boundary.
• Hardware failed to consume Port-VL and credits, which consequently blocked traffic from being transmitted due to a race condition between the firmware and the hardware when accessing the chip memory (CR space).

• Firmware crashed when setting large payload length values (more than ~1500) in NC-SI command's header.
• Interface type was shown as "unsupported" in CMIS modules.
• Fixed the configuration of the TS1s sent by the DownStream port (DSP) when moving to EQLZ.ph2.
• PLDM firmware update process failed in case 1304 bytes chunk size is chosen.
• An issue that occurred during secure firmware update when decrypting and authenticating each chunk of data using its authentication tag. The issue appeared when the main code chunk was split between the user chunks and any GCM operation (e.g., flash read with decryption). This GCM operation broke the GCM context for main chunk authentication and therefore failed. 
• CNP received, handled, and ignored counters in the hardware counters could not work after moving to Programmable Congestion Control mode.
• Hardware failed to handle an error flow that occurred when accessing the DPA cluster L2 cache from the firmware processor. In this case the firmware processor hardware requested a VA=>PA translation from the internal mmio, and the address translation was broken by the mmio on the 4K page boundary.
• Hardware failed to consume Port-VL and credits, which consequently blocked traffic from being transmitted due to a race condition between the firmware and the hardware when accessing the chip memory (CR space).

Firmware Dependencies:

Problems Fixed:

Updated Server Platform Services (SPS) Firmware to 06.00.04.033.0 to align with the Intel Best Known Configuration (BKC) WW09.

Known Issues:

Temperature Misreading: the adapter was reading incorrect reading of the temperature on the HPE SN1610Q 32Gb Fibre Channel Host Bus adapter's Small Form-factor Pluggable. The symptom would result in a temperature spike and potentially reset the system

• Unsuccessful read of the SFP Temperature which will trigger a Server Shutdown. 
• Message related to a Mailbox 0x48 Firmware dump saved to buffer.

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with certain 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel's description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different versions of internal processor firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with some 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel’s description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different version of internal firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None

This version of the Server Platform Services (SPS) Firmware contains updates aligned with the Intel Best Known Configuration (BKC) WW09.

Firmware Dependencies:

None

Problems Fixed:

Updated Server Platform Services (SPS) Firmware to 06.00.04.033.0 to align with the Intel Best Known Configuration (BKC) WW09.

Known Issues:

None

• RDE (Redfish) PATCH operation to LLDPTransmit properties "ManagementAddressIPv4", "ManagementAddressIPv6" and "ManagementAddressMAC" were applied only in the first attempt, but failed in the next.
• PLDM AEN event receiver media was changed unexpectedly and destination BDF was overridden with garbage when some PLDM packet were received from the SMBus layer.
• Bad configuration of number of VFs and SFs led to the consumption of too many functions and triggered a FW assert 0x888E. The reduction flows behavior was fixed to ensure the configuration does not exceed the total number of supported functions. 
• InfiniBand L2 QP could not receive RDMA traffic occasionally.
• Running with a debug firmware reduced security as if token was applied.

• "set_flow_table_entry" failed when aso_flow_meter action was used.
• A race condition occured between the duplicate read and QP commands (2RST, 2ERR and Destroy) in the signature that caused the command to hang.
• When all traffic applications sharing the same combination of <function, priority, side> were rate limited (for example by congestion control), this limit was enforced on other applications with different combinations of <function, priority, side> under the same VL. For example, requestor flows (RDMA-write) were limited to rate X, however, this rate was also enforced on a QP sending RDMAread responses. This firmware version prevents rate limit enforcement on traffic applications which should not be limited.
• Firmware update failed or timed out when multiple adapters of the same type were present on a system. This was fixed by adding "Command Unsupported" response code in cases when running the MCTP control command "Get Vendor Defined Messages Supported", and there were no supported VDMs.

• Packet loss that occurred when restarting the transmit. 
• An issue that prevented RoCE malformed packets (UDP packet with dest_port equal to RoCE well known udp_dport (0x4791)) from being counted on the vport_counter when the function disables RoCE (through MODIFY_NIC_VPORT_CONTEXT command).
• A memory leakage that occurred when closing connected QPs (Type RC/UC/XRC/DC).
• Added a missing VLAN strip.

• RDE (Redfish) PATCH operation to LLDPTransmit properties "ManagementAddressIPv4", "ManagementAddressIPv6" and "ManagementAddressMAC" were applied only in the first attempt, but failed in the next.
• PLDM AEN event receiver media was changed unexpectedly and destination BDF was overridden with garbage when some PLDM packet were received from the SMBus layer.
• Bad configuration of number of VFs and SFs led to the consumption of too many functions and triggered a FW assert 0x888E. The reduction flows behavior was fixed to ensure the configuration does not exceed the total number of supported functions. 
• InfiniBand L2 QP could not receive RDMA traffic occasionally.
• Running with a debug firmware reduced security as if token was applied.

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with certain 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel's description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different versions of internal processor firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with some 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel’s description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different version of internal firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None

• The patch to the RDE LLDPEnable property in Port schema would not be updated after the host reboot.
• RDE (Redfish) PATCH operation to LLDPTransmit properties "ManagementAddressIPv4", "ManagementAddressIPv6" and "ManagementAddressMAC" were applied only in the first attempt, but failed in the next.
• PLDM AEN event receiver media was changed unexpectedly and destination BDF was overridden with garbage when some PLDM packet were received from the SMBus layer.
• PCIe SKP OS generation interval for Gen1 and Gen2.

• An issue where set_flow_table_entry failed when aso_flow_meter action was used.
• A race condition occured between the duplicate read and QP commands (2RST, 2ERR and Destroy) in the signature that caused the command to hang.
• An issue where vPort counters had wrong values.
• An issue where firmware update failed or timed out when multiple adapters of the same type were present on a system. This was fixed by adding "Command Unsupported" response code in cases when running the MCTP control command "Get Vendor Defined Messages Supported", and there were no supported VDMs.

• An issue that prevented the adapter card from handling WoL packets when the host powered up directly into the standby mode.
• CNP received, handled, and ignored counters in the hardware counters could not work after moving to Programmable Congestion Control mode.

• The patch to the RDE LLDPEnable property in Port schema would not be updated after the host reboot.
• RDE (Redfish) PATCH operation to LLDPTransmit properties "ManagementAddressIPv4", "ManagementAddressIPv6" and "ManagementAddressMAC" were applied only in the first attempt, but failed in the next.
• PLDM AEN event receiver media was changed unexpectedly and destination BDF was overridden with garbage when some PLDM packet were received from the SMBus layer.
• PCIe SKP OS generation interval for Gen1 and Gen2.

Firmware Dependencies:

Problems Fixed:

Addressed an issue where system may hang at POST during TPM measurement after restoring manufacturing defaults.

Address an issue where the RegistryVersion format in the URI redfish/v1/registrystore/registries did not follow the Redfish Specification. If scripts were written to conform with the incorrect format in previous System ROMs, those scripts may need to be modified to follow the correct, Redfish compliant format which is implemented in this System ROM revision and later.

Addressed an issue where the System Configuration (RBSU) option "Access Control Service" did not disable PCIe Access Control Service when it’s set to disabled.

Known Issues:

None

Firmware Dependencies:

None

Problems Fixed:

This revision of the System ROM includes the latest revision of the AMD reference code for AMD 4th Generation EPYC processors which addresses an issue of potential over voltage condition early in the boot sequence. This issue is not unique to HPE servers.

Known Issues:

None

Firmware Dependencies:

Problems Fixed:

Address an issue that can result in a Red Screen of Death (RSOD) and hang during system boot or a hang during system boot without an RSOD. This issue can occur on systems utilizing the v1.32 or v1.40 System ROM revision and iLO Firmware revision v1.50. The v1.32 System ROM added a boot-time reduction capability (which also required iLO Firmware revision v1.50) which resulted in the boot-time RSOD and/or hang. This capability has been removed in this revision of the System ROM.

Addressed an issue where one-time boot did not work properly after setting one-time boot device from iLO web or via Redfish.

Known Issues:

Firmware Dependencies:

Problems Fixed:

This revision of the System ROM includes the revision of the OpenSSL library 1.1.1u update which provides mitigation for and security vulnerabilities documented as CVE-2023-2650.

Addressed an issue where the unsupported DIMM population IML will report an incorrect processor number on dual-processor systems.

Addressed an issue where the SMBIOS type 197 might contain an incorrect format value of xGMI speed.

Addressed an issue where the CXL root bridge was missing from SMBIOS type 9.

Addressed an issue where dmidecode indicates 4 power supplies are installed when only 2 power supplies are actually installed.

Addressed an issue where number of "Internal L3 cache" displayed under RBSU information page is not correct.

Addressed an issue where the System Health LED is illuminated Red for all memory configuration error cases. Some memory configuration error cases are not critical and should have the System Health LED illuminated Amber.

Known Issues:

• Installing "mlnx_tune" on Python3 did not work properly. "mlnx_tune" now supports Python3 in addition to Python2.
• Changing traffic class via the sysfs while modifying QPs in parallel caused a deadlock.
• Due to a bug in the emulation layer, performance degradation was experienced when running GPUDirect over Virtual Functions.
• "debugfs" directories could not be created for representors and sub-functions, thus the log showed error for both the scenarios.
• "knem" did not fully support RHEL8.7 and newer releases.

Firmware Dependencies:

Problems Fixed:

Updated Server Platform Services (SPS) Firmware to 06.00.04.033.0 to align with the Intel Best Known Configuration (BKC) WW09.

Known Issues:

• Aligned RDE behavior to DSP0266 v1.15.0  table 23.

The following issues have been fixed in version 20.35.1012:

• RDE (Redfish) PATCH operation to LLDPTransmit properties "ManagementAddressIPv4", "ManagementAddressIPv6" and "ManagementAddressMAC" were applied only in the first attempt, but failed in the next.
• The RXT E2E inflight went into an unresposive state occasionally.
• RDMA partition was reported even if NIC+RDMA mode was disabled.
• PLDM AEN event receiver media was changed unexpectedly and destination BDF was overridden with garbage when some PLDM packet were received from the SMBus layer.
• An incorrect address issue for XRQ Event Backend Controller Px Syndrome event in multi-host system during FLR.
• PCIe SKP OS generation interval issue for Gen1 and Gen2.
• InfiniBand L2 QP could not receive RDMA traffic occasionally.
• Running with a debug firmware reduced security as if token was applied.
• PLDM AEN event receiver was required to be cleared on PCIe reset in case the media type was MCTP over PCIe VDM.
• Asynchronous messages were sent over MTCP before endpoint discovery was done.

• Aligned RDE behavior to DSP0266 v1.15.0  table 23.

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with certain 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel's description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different versions of internal processor firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with some 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel’s description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different version of internal firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with certain 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel's description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different versions of internal processor firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with some 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel’s description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different version of internal firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None

Firmware Dependencies:

Problems Fixed:

Addressed an issue where system may hang at POST during TPM measurement after restoring manufacturing defaults.

Address an issue where the RegistryVersion format in the URI redfish/v1/registrystore/registries did not follow the Redfish Specification. If scripts were written to conform with the incorrect format in previous System ROMs, those scripts may need to be modified to follow the correct, Redfish compliant format which is implemented in this System ROM revision and later.

Addressed an issue where the System Configuration (RBSU) option "Access Control Service" did not disable PCIe Access Control Service when it’s set to disabled.

Known Issues:

None

Firmware Dependencies:

None

Problems Fixed:

This revision of the System ROM includes the latest revision of the AMD reference code for AMD 4th Generation EPYC processors which addresses an issue of potential over voltage condition early in the boot sequence. This issue is not unique to HPE servers.

Known Issues:

None

Firmware Dependencies:

Problems Fixed:

Address an issue that can result in a Red Screen of Death (RSOD) and hang during system boot or a hang during system boot without an RSOD. This issue can occur on systems utilizing the v1.32 or v1.40 System ROM revision and iLO Firmware revision v1.50. The v1.32 System ROM added a boot-time reduction capability (which also required iLO Firmware revision v1.50) which resulted in the boot-time RSOD and/or hang. This capability has been removed in this revision of the System ROM.

Addressed an issue where one-time boot did not work properly after setting one-time boot device from iLO web or via Redfish.

Known Issues:

Firmware Dependencies:

Problems Fixed:

This revision of the System ROM includes the revision of the OpenSSL library 1.1.1u update which provides mitigation for and security vulnerabilities documented as CVE-2023-2650.

Addressed an issue where the unsupported DIMM population IML will report an incorrect processor number on dual-processor systems.

Addressed an issue where the SMBIOS type 197 might contain an incorrect format value of xGMI speed.

Addressed an issue where the CXL root bridge was missing from SMBIOS type 9.

Addressed an issue where dmidecode indicates 4 power supplies are installed when only 2 power supplies are actually installed.

Addressed an issue where number of "Internal L3 cache" displayed under RBSU information page is not correct.

Addressed an issue where the System Health LED is illuminated Red for all memory configuration error cases. Some memory configuration error cases are not critical and should have the System Health LED illuminated Amber.

Known Issues:

• RDE (Redfish) PATCH operation to LLDPTransmit properties "ManagementAddressIPv4", "ManagementAddressIPv6" and "ManagementAddressMAC" were applied only in the first attempt, but failed in the next.
• Commands sent by the MLNX_OFED driver to the NIC would fail when loading the VirtIO driver. 
• A firmware limitation caused enabling tx_port_ts to result in syndrome 0x5d2974.
• PLDM AEN event receiver media was changed unexpectedly and destination BDF was overridden with garbage when some PLDM packet were received from the SMBus layer.
• PCIe SKP OS generation interval for Gen1 and Gen2.
• CPU handling synchronization required separation (run ptp4l with taskset -c [cpu #] prefix) while running heavy traffic. 
   

• An issue that caused the destroy_match_definer object command to fail after dumping it using resource_dump.
• An issue where set_flow_table_entry failed when aso_flow_meter action was used. 
• An issue where firmware update failed or timed out when multiple adapters of the same type were present on a system. This was fixed by adding "Command Unsupported" response code in cases when running the MCTP control command "Get Vendor Defined Messages Supported", and there were no supported VDMs.

• No ping in NODNIC VF when VLAN stripping was enabled.
• vDPA application initialization failed due to virtual queue creation failure. The failure came with the "0x8f7a: ring address translate failed" assert which indicates incorrect permission supplied by QEMU for virtual ring memory.
• CNP received, handled, and ignored counters in the hardware counters could not work after moving to Programmable Congestion Control mode.

• RDE (Redfish) PATCH operation to LLDPTransmit properties "ManagementAddressIPv4", "ManagementAddressIPv6" and "ManagementAddressMAC" were applied only in the first attempt, but failed in the next.
• Commands sent by the MLNX_OFED driver to the NIC would fail when loading the VirtIO driver. 
• A firmware limitation caused enabling tx_port_ts to result in syndrome 0x5d2974.
• PLDM AEN event receiver media was changed unexpectedly and destination BDF was overridden with garbage when some PLDM packet were received from the SMBus layer.
• PCIe SKP OS generation interval for Gen1 and Gen2.
• CPU handling synchronization required separation (run ptp4l with taskset -c [cpu #] prefix) while running heavy traffic. 
   

Firmware Dependencies:

Problems Fixed:

Updated Server Platform Services (SPS) Firmware to 06.00.04.033.0 to align with the Intel Best Known Configuration (BKC) WW09.

Known Issues:

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with certain 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel's description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different versions of internal processor firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with some 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel’s description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different version of internal firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with certain 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel's description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different versions of internal processor firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None

None

Firmware Dependencies:

None

Problems Fixed:

Address an issue with some 4^th Generation Intel Xeon Processors that can result in an unexpected reset with a machine check error logged to the Integrated Management Log (IML).  This issue is not unique to HPE systems.  The following is Intel’s description of the issue:  "Under complex microarchitectural conditions, the processor may experience a Caching Home Agent Table of Requests (TOR) timeout machine check exception."  This version of the System ROM will update internal processor firmware during boot.  This internal processor firmware cannot be downgraded even if the System ROM is downgraded. Using processors with different version of internal firmware is not supported.

Addressed an issue where the speed of Kioxia EDSFF drive might be downgraded from PCIe Generation 5 to Generation 4.

Known Issues:

None