NVIDIA BlueField-2 DPU Firmware Release Notes v24.35.2000 LTS

Bug Fixes History

Internal Ref.

Issue

3023205

Description: Fixed an issue that resulted in:

  • data corruption

  • NVMF queue getting stuck, following backend controller timeout and RNR NAK exceeded syndrome on initiator

due to mistakes in access to the internal database NVMF Backend Controller as a result of wrong data in the database.

Keywords: data_corruption

Discovered in Version: 24.33.1048

Fixed in Release: 24.34.1002

3101645

Description: Fixed an issue that caused some requests to get stuck in the pacer, thus not allowing detach NVMF namespace command to progress when the pacer was configured to use byte in flight limitation mechanism and NVMf backend controller timeout happened under traffic.

Keywords: CMD queue

Discovered in Version: 24.33.1048

Fixed in Release: 24.34.1002

3021669

Description: Added a new NVconfig parameter “MULTI_PCI_RESOURCE_SHARE” to support modes that allow choosing the utilization of the card's resources on each host in Socket-Direct / Multi host setup.

Keywords: Performance

Discovered in Version: 24.33.1048

Fixed in Release: 24.34.1002

2665773

Description: Added 50 Usec delay during PML1 exit to avoid any PCIe replay timer timeout.

Keywords: PCIe. PML1

Discovered in Version: 24.33.1048

Fixed in Release: 24.34.1002

3134894

Description: Fixed an issue where set_flow_table_entry failed when aso_flow_meter action was used.

Keywords: ASO Flow Meter, FW Steering

Discovered in Version: 24.33.1048

Fixed in Release: 24.34.1002

3039007

Description: Enabled Multi-Host RX Rate-limiter configuration via the QEEC mlxreg and the max_shaper_rate field.

Keywords: RX Rate-Limiter, Multi-host

Discovered in Version: 24.33.1048

Fixed in Release: 24.34.1002

3059379

Description: Added "Command Unsupported" response code in cases when running the MCTP control command "Get Vendor Defined Messages Supported", and there were no supported VDMs.

Keywords: MCTP control command

Discovered in Version: 24.33.1048

Fixed in Release: 24.34.1002

2994292

Description: Fixed a race condition occured between the duplicate read and QP commands (2RST, 2ERR and Destroy) in the signature that caused the command to hang.

Keywords: Race condition

Discovered in Version: 24.33.1048

Fixed in Release: 24.34.1002

3068927

Description: Added support for dynamic MSI-X when in SmartNIC mode.

Keywords: Dynamic MSI-X

Discovered in Version: 24.33.1048

Fixed in Version: 24.34.1002

3110286

Description: Fixed an issue where vPort counters had wrong values.

Keywords: vPort counters

Discovered in Version: 24.33.1048

Fixed in Version: 24.34.1002

2887966

Description: A hardware issue in the illegal_flowq that raises even when there is drop, results in the adapter card getting stuck during
high scale traffic.

Keywords: Performance

Discovered in Version: 24.33.1048

Fixed in Version: 24.34.1002

3056461

Description: Creating a Channel Service Object with bad parameters that lead to command rollback, results in the command getting stuck.

Keywords: Open SNAPI

Discovered in Version: 24.33.1048

Fixed in Version: 24.34.1002

Internal Ref.

Issue

2702118

Description: Changed the policy of VDPA queue number capability.

  • When the devices count <= 8, the VDPA queue number in cap is 256

  • When the device count >=32, the VDPA queue number in cap is 64

  • When the devices count is in 9~31, the VDPA queue number in cap is 128. Here the devices counts all port functions configured in mlxconfig, including pf, vf and sf.

Keywords: VDPA

Discovered in Version: 24.32.1010

Fixed in Version: 24.33.1048

2991255

Description: Fixed an issue that caused the host driver to hang when the received packet was bigger than the received buffer (according to the device's MTU) as the device reported a packet length bigger than the received buffer length.

Keywords: virtio net RX packet, RX buffer

Discovered in Version: 24.32.1010

Fixed in Version: 24.33.1048

3021734

Description: Fixed an issue that caused BMC to fail to detect the PCIe device when using the MCTP-over-PCIe protocol.

Keywords: BMC, MCTP-over-PCIe protocol, PCIe

Discovered in Version: 24.32.1010

Fixed in Version: 24.33.1048

2802943

Description: Implemented SLD detection code. Surprise Down Error Reporting Capable value was changed from 1 to 0 in boards where the downstream perst was not controlled thus causing SLD detection not to function properly.

Keywords: SLD detection, Surprise Down Error Reporting Capable

Discovered in Version: 24.32.1010

Fixed in Version: 24.33.1048

2513453

Description: Fixed rare lanes skew issue that caused CPU to timeout in Rec.idle.

Keywords: PCIe

Discovered in Version: 24.32.1010

Fixed in Version: 24.33.1048

2932436

Description: Optimized the virtio data path to reach line speed for Tx bandwidth.

Keywords: VDPA, virtio full emulation

Discovered in Version: 24.32.1010

Fixed in Version: 24.33.1048

2907707

Description: Fixed a configuration issue which flipped the MSB of Partition Key field in CNP packets and led to P_KEY mismatch between CNP packets and regular packets.

Keywords: Partition Key, PKEY, CNP, ECN

Discovered in Version: 24.32.1010

Fixed in Version: 24.33.1048

© Copyright 2023, NVIDIA. Last updated on May 23, 2023.