Bug Fixes in This Version

4520687

Description: HBN container may hang in init-sfs during container restart if TCP sessions on master VRF devices are not cleaned up properly

Keywords: Hang; container

Reported in version: 2.9.4

N/A

Description: N/A

Keywords: N/A

Reported in version: N/A

4659210

Description: The system incorrectly permits users to configure the internal CPU model to "Separated" mode, even though this mode is deprecated and no longer supported.

Keywords: Modes of operation

Reported in version: 4.9.4

4924239

4817141

Description: The RShim USB device may intermittently disappear from the DPU BMC, causing operations that rely on it to fail with a Failed to enable BMC rshim error.

Keywords: RShim USB; out-of-band update

Discovered in version: 4.9.4

4929721

4848119

Description: A BlueField-2 UEFI boot-time regression added approximately 20 seconds to system startup.

Keywords: BlueField-2; boot time; UEFI

Discovered in version: 4.9.4

4982730

Description: After a host power cycle, (particularly when ENABLE_BMC_WAIT=1 is configured), the UEFI boot progress is not completely logged in the BMC SEL. The expected System boot initiated event is missing, and an incorrect event message is logged instead.

Keywords: System Event Log; UEFI boot progress

Discovered in version: 4.9.4

4983438

Description: During BlueField boot, the BlueField BIOS/UEFI may fail to acquire a DHCP IP address from the BMC for Redfish communication. This would cause the BIOS to bypass the Redfish step and to ignore any configuration changes in Redfish.

Keywords: DHCP; BIOS; UEFI

Discovered in version: 4.9.4

5000684

Description: After upgrading a BFB image, UEFI boot entries may disappear, and the BootOrder may change in the efibootmgr output.

Keywords: EFI; BFB upgrade

Discovered in version: 4.9.4

5030858

Description: When running the doca-installer tool with the --show-running-fw flag, the utility may intermittently fail to read the CEC firmware component version, displaying N/A in the output list.

Keywords: CEC firmware; version reading error

Discovered in version: 4.9.4

5040557

Description: After a BFB update and system power cycle, the host mlx5_core driver may fail to initialize.

Keywords: BlueField; boot timeout

Discovered in version: 4.9.4

5041514

Description: Timing dependencies during boot can lead to the BMC misinterpreting SEL boot logs.

Keywords: Logging

Discovered in version: 4.9.4

5031932

4234285

Description: During BlueField boot, the BlueField BIOS/UEFI may fail to acquire a DHCP IP address from the BMC for Redfish communication. This would cause the BIOS to bypass the Redfish step and to ignore any configuration changes in Redfish.

Keywords: DHCP; BIOS; UEFI

Discovered in version: 4.9.4

4789661

Description: When frequently polling RShim misc information during a BFB image installation, the RShim USB_TIMEOUT may exceed the expected 40-second maximum limit. The high polling frequency causes the interface to appear persistently busy, which delays the timeout enforcement.

Discovered in version: 24.10-LTSU4

4811729

Description: The BMC event log may continuously record spurious PLDM_Sensor_1_100 warning messages. These messages do not indicate a genuine thermal issue or hardware failure. Additionally, this behavior may be accompanied by intermittent core dumps of the pldmd service within the BMC. These warnings are benign and can be safely ignored.

Discovered in version: 24.10-LTSU4

4850556

Description: Performing a power reset of the DPU via the DPU-BMC may result in partial IP address information being displayed in redfish/v1/Systems/Bluefield/Ethernetinterfaces/. This issue is cosmetic and does not impact network functionality.

Discovered in version: 24.10-LTSU4

4860548

4826162

Description: On systems equipped with multiple BlueField-3 DPUs, the doca-installer --compare command may intermittently fail to retrieve the CEC firmware version from one of the cards.

Discovered in version: 24.10-LTSU4

4895085

Description: When enabling Livefish mode from the DPU-BMC using ipmitool, the command may return a timeout error. This response is erroneous; the DPU successfully transitions into Livefish mode despite the error message.

Discovered in version: 24.10-LTSU4

4910247

Description: When operating BlueField in NIC mode, the UEFI may intermittently fail to retrieve Redfish credentials from the BMC following a host power cycle. This results in the DPU-BMC RF credentials not found error being recorded in the rshim and DPU console logs.

Discovered in version: 24.10-LTSU4

4910277

4875901

Description: When the ENABLE_BMC_WAIT flag is enabled, the system may intermittently fail to retrieve BMC Redfish credentials during boot following a host power cycle. When this occurs, the RShim logs will incorrectly report DPU-BMC RF credentials not found.

Discovered in version: 24.10-LTSU4

4969205

Description: When the ENABLE_BMC_WAIT flag is active, the BMC SEL may intermittently fail to record the complete UEFI boot progress following a host power cycle. This is a cosmetic issue and does not affect the functionality of the DPU.

Discovered in version: 24.10-LTSU4

4970594

Description: The BMC root partition may become entirely full due to excessive data accumulation in the /var/cper directory. When the partition reaches maximum capacity, critical BMC services (e.g., entity-manager) crash. This causes hardware monitoring commands like ipmitool sdr elist to fail with an Invalid data field in request error.

Discovered in version: 24.10-LTSU4

4970770

Description: The BMC Manager Redfish endpoint (/redfish/v1/Managers/Bluefield_BMC) may become unresponsive and return a TimedOut error. This issue is accompanied by continuous ipmb-host 1-1011: ERROR: Response timed out during ipmb detection messages in the console logs and severe delays (multiple minutes) when accessing the "System Configuration" page in the DPU UEFI menu.

Discovered in version: 24.10-LTSU4

4976717

Description: Starting a secondary BMC task (e.g., a log dump) before the monitoring logic attaches to an initiated firmware update can cause the monitor to track the wrong task, resulting in a BMC reboot failure.

Discovered in version: 24.10-LTSU4

5017608

Description: During BFB installations via Redfish, the task state may change to "Exception" before the specific error message is appended to the HTTP response payload. This results in incomplete error logs on the initial poll following a failure.

Discovered in version: 24.10-LTSU4

5017618

Description: The set_emu_params.sh script attempts to load the mlxbf_ptm driver, leading to continuous module load failures and log spam.

Discovered in version: 24.10-LTSU4

5030858

Description: When running the doca-installer tool with the --show-running-fw flag, the utility may intermittently fail to read the CEC firmware component version, displaying N/A in the output list.

Discovered in version: 24.10-LTSU4

4902171

Description: Fixed an issue where a slow module response to the NCSI GetModuleSerialData command could incorrectly trigger a dead IRISC watchdog assert. The firmware now handles this waiting state correctly, so slow module responses no longer cause a false watchdog hang indication.

Keywords: NCSI

Discovered in Version:

32.43.4100

Fixed in Release:

32.43.8016

4832172 / 4819986

Description: Fixed an issue where the server could occasionally become stuck during live migration or while updating the Virtio emulation device.

Keywords: virtio emulation, live migration, live update

Discovered in Version:

32.43.4100

Fixed in Release:

32.43.8016

4934276 / 4867111

Description: Fixed an issue where RoCEv2 TX traffic could be dropped when running the DOCA spcx_cc application due to stale PCC steering extension entries.

Keywords: spcx_cc cc_steering_ext

Discovered in Version:

32.43.4100

Fixed in Release:

32.43.8016

4780301 / 4683339

Description: Fixed an issue where QPs established before loading DOCA PCC could exhibit inconsistent algorithm-selected behavior between ports in LAG mode after DOCA PCC is loaded.

Keywords: Congestion Control, DOCA PCC

Discovered in Version:

32.43.2402

Fixed in Release:

32.43.8016

4980956 / 4947006

Description: Fixed two rare race conditions in the PCC steering extension when the feature is enabled via path migration, empty RTT, or a custom header.

Under rare timing conditions during QP creation and destruction, these issues could lead to traffic disruption or resource leaks.

Keywords: PCC steering extension

Discovered in Version:

32.43.4100

Fixed in Release:

32.43.8016

4874764 / 4657767

Description: Fixed an issue where repeatedly writing NVCONFIG TLVs could cause excessive NV_DATA partition swaps during garbage collection. This rapid cycling could accelerate flash wear (end-of-life at 100,000 erases) and potentially render the device inoperable.
Firmware now avoids unnecessary physical writes by returning OK when the requested configuration already exists in flash, and increases the maximum supported NV_DATA partition swaps from 100,000 to 200,000.

Keywords: NVCONFIG TLVs

Discovered in Version:

32.43.4100

Fixed in Release:

32.43.8016

4986752 / 4849553

Description: Fixed an issue where running RDMA traffic with CRC signature offloading in a lossy network could result in a LOCAL PROTECTION ERROR on the responder side or a CRC mismatch on the requester side during packet-loss retransmission.

Keywords: RDMA, CRC

Discovered in Version:

32.43.2566

Fixed in Release:

32.43.8016