NVIDIA ConnectX-6 Dx Adapter Cards Firmware Release Notes v22.39.1002
NVIDIA ConnectX-6 Dx Adapter Cards Firmware Release Notes v22.39.1002

Known Issues

VF Network Function Limitations in SRIOV Legacy Mode

Dual Port Device

Single Port Device

127 VF per PF (254 functions)

127

VF Network Function Limitations in Switchdev Mode

Dual Port Device

Single Port Device

127 VF per PF (254 functions)

127

VF+SF Network Function Limitations in Switchdev Mode

Dual Port Device

Single Port Device

  • 127 VF per PF (254 functions)

  • 512 PF+VF+SF per PF (1024 functions)

  • 127 VF (127 functions)

  • 512 PF+VF+SF per PF (512 functions)

Known Issues

Internal Ref.

Issue

3464393

Description: PhyLess Reset is currently not supported.

Workaround: N/A

Keywords: PhyLess Reset

Discovered in Version: 22.39.1002

3525865

Description: Unexpected system behavior might be observed if the driver is loaded while reset is in progress.

Workaround: N/A

Keywords: Sync 1 reset, firmware reset

Discovered in Version: 22.39.1002

3547022

Description: When tx_port_ts is set to "true", due to a compensation mechanism in the Tx timestamp available in some hardware Rx timestamp errors, a symmetrical error and no clock offset occur when using the timestamps to synchronize the device clock.

This might also cause an error while using timestamps for delay measurements (e,g., delay measurements reported by a PTP daemon) and even negative delay measurements in some cases.

Workaround: N/A

Keywords: PTP path delay

Discovered in Version: 22.38.1002

3457472

Description: Disabling the Relaxed Ordered (RO) capability (relaxed_ordering_read_pci_enabled=0) using the vhca_resource_manager is currently not functional.

Workaround: N/A

Keywords: Relaxed Ordered

Discovered in Version: 22.37.1014

3449960

Description: In some cases, performance degradation might be experienced when wrong usage of extra engines dedicated to DMA is removed.

Workaround: N/A

Keywords: Performance

Discovered in Version: 22.37.1014

2878841

Description: Firmware rollback fails for the signature retransmit flow if the QPN field is configured in the mkey (as it only allows the given QP to use this Mkey) as the firmware rollback flow relies on an internal QP that uses the mkey.

Workaround: N/A

Keywords: Signature retransmit flow

Discovered in Version: 22.37.1014

3171699

Description: Occasionally, after a few toggles, link may not raise when changing the speed when in loopback mode.

Workaround: N/A

Keywords: Link speed, loopback

Discovered in Version: 22.37.1014

2444892

Description: PMA loopback feature is supported only with NRZ speeds .

Workaround: N/A

Keywords: PMA loopback, NRZ

Discovered in Version: 22.36.1010

2745023

Description: RDMA statistics for sent packets are not updated when RoCE traffic is running in a loopback on the same uplink.

Workaround: N/A

Keywords: RoCE

Discovered in Version: 22.35.2302

3266807

Description: PMA loop-back is not supported on PAM4 speeds.

Workaround: N/A

Keywords: Counters, CRC

Discovered in Version: 22.35.2302

3267506

Description: CRC is included in the traffic byte counters as a port byte counter.

Workaround: N/A

Keywords: Counters, CRC

Discovered in Version: 22.35.2302

3235397

Description: PCC force mode does not work if the link is raised after disabling DCQCN with PPCC.

Workaround: N/A

Keywords: PCC

Discovered in Version: 22.35.1012

3200779

Description: Changing dynamic PCIe link width is not supported.

Workaround: N/A

Keywords: PCIe

Discovered in Version: 22.34.1002

3033910

Description: BAR misses caused by a memory write/read actions are not reported in the AER and the device status.

Workaround: N/A

Keywords: BAR miss, AER

Discovered in Version: 22.34.1002

2169950

Description: When decapsulation on a packet occurs, the FCS indication is not calculated correctly.

Workaround: N/A

Keywords: FCS

Discovered in Version: 22.34.1002

3141072

Description: The "max_shaper_rate" configuration query via QEEC mlxreg returns a value translated to hardware granularity.

Workaround: N/A

Keywords: RX Rate-Limiter, Multi-host

Discovered in Version: 22.34.1002

3106146

Description: Live migration of MPV affiliated function pair is not supported when port numbers are changed. Each function should stay on the same port number as before migration.

Workaround: N/A

Keywords: MPV live migration

Discovered in Version: 22.34.1002

2870970

Description: GTP encapsulation (flex parser profile 3) is limited to the NIC domain.

Encapsulating in the FDB domain will render a 0-size length in GTP header.

Workaround: N/A

Keywords: GTP encapsulation

Discovered in Version: 22.34.1002

2937445

Description: A long linkup time can be seen 1/5 toggles when raising link in autoneg flow in ConnectX-6 Dx vs Ixia in 200G_4x.

Workaround: N/A

Keywords: AN, port toggling, Ixia

Discovered in Version: 22.33.1048

2850003

Description: Occasionally, when rising a logical link, the link recovery counter is increase by 1.

Workaround: N/A

Keywords: Link recovery counter

Discovered in Version: 22.33.1048

2825403

Description: When connecting NVIDIA Spectrum-3 devices and ConnectX-6 Dx devices with DAC MCP7F80-W002R26 while splitting to 8x with 50GbE per lane in force mode, effective BER may appear.

Workaround: N/A

Keywords: NVIDIA Spectrum-3, Cables, Split

Discovered in Version: 22.32.2004

2866931

Description: When the host powers up directly into the standby mode, the adapter may not handle WOL packets.

Workaround: N/A

Keywords: WOL packets

Discovered in Version: 22.32.1010

2864238

Description: VPD cannot be accessed after firmware upgrade or reset when the following sequence is performed:

  1. Upgrade to a new firmware and perform a cold reboot

  2. Downgrade to an old firmware

  3. Run fwreset

  4. Upgrade to a new firmware

  5. Run fwreset

Workaround: Run the upgrade or reset sequence as follow:

  1. Upgrade to a new firmware and perform a cold reboot

  2. Downgrade to an old firmware

  3. Run fwreset

  4. Upgrade to a new firmware

  5. Perform a cold reboot

Keywords: VDP

Discovered in Version: 22.32.1010

2863674

Description: Host management magic packet is not supported in Socket-Direct adapter cards' single PF per Numa mode.

Workaround: N/A

Keywords: Socket-Direct, single PF per Numa, host management, magic packet

Discovered in Version: 22.32.1010

2836032

Description: When using SW steering mlx5dv_dr API to create rules containing encapsulation actions in MLNX_OFED v5.5-1.x.x.x, the user should upgrade firmware to the latest version. Otherwise, the maximum number of encapsulation actions that can be created will be limited to only 16K, and degradation for the rule insertion rate is expected compared to MLNX_OFED v5.4-.x.x.x.x.

Workaround: N/A

Keywords: Encapsulation rules insertion rate, firmware upgrade, MLNX_OFED

Discovered in Version: 22.32.1010

2756866 / 2740651

Description: On rare occasions, following fast linkup (toggle link from the NIC side) a few effective errors might be seen in the first 20 seconds.

Workaround: Perform link maintenance to fix it so additional errors will not be seen afterwards.

Keywords: Link toggle, effective errors

Discovered in Version: 22.31.2006

-

Description: Downgrading to an older firmware version that does not support the new flash type is not supported. Doing so will result in burning process failure and unknown errors will be received.

The errors will be more informative in the next tools' version.

Workaround: N/A

Keywords: Burning tools, firmware downgrading, flash type

Discovered in Version: 22.31.2006

2667681

Description: As the Connection Tracking (CT) is not moved to SW state after receiving a TCP RST packet, any packet that matches the windows even after the RST is marked as a valid packets.

Workaround: N/A

Keywords: Connection Tracking

Discovered in Version: 22.31.1014

2607158

Description: When using more than 512 MSIX per function, the CPU PCIe Completion Timeout Value needs to be set to a value of 200us or higher.

Workaround: N/A

Keywords: Extended MSIX, Asymmetrical MSIX configuration, PF_NUM_PF_MSIX_VALID, PF_NUM_PF_MSIX

Discovered in Version: 22.31.1014

2577966

Description: Fast linkup is not supported when connecting to an Ixia switch.

Workaround: N/A

Keywords: Fast linkup

Discovered in Version: 22.30.1004

2446583

Description: On rare occasions, when both network devices are NVIDIA, PAM4 link will raise with several effective errors. These errors will not affect traffic once the link is up.

Workaround: Clear counters once the link is up

Keywords: Effective errors

Discovered in Version: 22.29.2002

2371060

Description: When Emulated PCIe Switch is enabled, and the OS does resource reallocation, the OS boot process might halt.

Workaround: N/A

Keywords: Emulated PCIe Switch

Discovered in Version: 22.29.1016

2297201

Description: Unable to complete migration when virtio device is in high traffic load (20/20 MPPS) as although vDPA hardware offload solution can support higher speed than the software solution, it needs to enable QEMU auto-converge to complete migration.

For further information see: https://wiki.qemu.org/Features/AutoconvergeLiveMigration

Workaround: Turn auto-converge on by adding "--auto-converge".

For example: virsh migrate --verbose --live --persistent gen-l-vrt-295-005-CentOS-7.4 qemu+ssh://gen-l-vrt-295/system --unsafe --auto-converge

Keywords: virtio, vDPA, live migration

Discovered in Version: 22.29.1016

2378593

Description: Sub 1sec firmware update (fast reset flow) is not supported when updating from previous releases to the current one. Doing so may cause network disconnection events.

Workaround: Use full reset flow for firmware upgrade/downgrade.

Keywords: Sub 1sec firmware update

Discovered in Version: 22.29.1016

2384965

Description: Eye-opening can cause effective errors on the port.

Workaround: N/A

Keywords: Eye-opening

Discovered in Version: 22.29.1016

2384849 / 2373640

Description: Phyless Reset functionality is not supported when updating firmware from v22.28.4000 (and below) to v22.29.1016 and higher.

Workaround: N/A

Keywords: Phyless Reset

Discovered in Version: 22.29.1016

2213356

Description: The following are the Steering Dump limitations:

  • Supported only on ConnectX-5 adapter cards

  • Requires passing the version (FW/Stelib/MFT) and device type to stelib

  • Re-format is not supported

  • Advanced multi-port feature is not supported – LAG/ROCE_AFFILIATION/MPFS_LB/ESW_LB (only traffic vhca <-> wire)

  • Packet types supported:

    • Layer 2 Eth

    • Layer 3 IPv4/Ipv6/Grh

    • Layer 4 TCP/UDP/Bth/GreV0/GreV1

    • Tunneling VXLAN/Geneve/GREv0/Mpls

  • FlexParser protocols are not supported (e.g AliVxlan/VxlanGpe etc..).

  • Compiles only on x86

Workaround: N/A

Keywords: Steering Bump

Discovered in Version: 22.29.1016

2365322

Description: When configuring adapter card's Level Scheduling, a QoS tree leaf (QUEUE_GROUP) configured with default rate_limit and default bw_share, may not obey the QoS restrictions imposed by any of the leaf’s ancestors.

Workaround: To prevent such a case, configure at least one of the following QoS attributes of a leaf: max_average_bw or bw_share

Keywords: QoS

Discovered in Version: 22.29.1016

2201468

Description: Running multiple resets ("mlxfwreset --sync=1") simultaneously is not functioning properly,

Workaround: Wait a few seconds until you run "mlxfwreset --sync=0".

Keywords: mlxfwreset, reset-sync, reset, sync

Discovered in Version: 22.28.1002

2089277

Description: The CRC is being removed despite using the keep_crc flag, and the byte count of the packet are counted without the CRC.

Workaround: N/A

Keywords: Decapsulated packets

Discovered in Version: 22.27.6008

2149437

Description: When the SLTP configuration is wrongly set, the “Bad status” explanation will not be presented (only error indication) to the user.

Workaround: N/A

Keywords: SLTP configuration

Discovered in Version: 22.27.6008

1895917

Description: On Dual-Port devices, and only after Rx buffer modification, resetting all Physical Functions over one port (through reboot / driver restart / FLR), while there are active Physical Functions over the second port (which caused the Rx buffer changes), will cause the Rx buffer default values to be restored, although not expected by the active Physical Function on the second port.

Workaround:

  • Re-apply the changes

  • Reset the functions from both ports together (driver restart / FLRs / reboot)

  • Power cycle or reset the firmware

Keywords: VoQ, Shared Buffer, Rx Bufffer, PFCC, PBMC, PPTB, SBCM, SBPM, SBPR, Rx buffer modifications

Discovered in Version: 22.27.2008

2120378

Description: Phyless Reset is not supported when using PAM4 mode.

Workaround: N/A

Keywords: Phyless, PAM4 mode, 200GbE

Discovered in Version: 22.27.2008

2071210

Description: mlxconfig query for the BOOT_INTERRUPT_DIS TLV shows a wrong value in the "current value" field.

Workaround: Use "next boot" indication to see the right value.

Keywords: mlxconfig

Discovered in Version: 22.27.1016

2063038

Description: PRBS is not functional when using Wedge switch.

Workaround: N/A

Keywords: PRBS

Discovered in Version: 22.27.1016

1796936

Description: 200GbE Optical cables in Auto-Negotiation mode work only in 200GbE speed.

Workaround: N/A

Keywords: Cables

Discovered in Version: 22.27.1016

2038821

Description: When running MH TCP, few packets are dropped every second due to no Receive WQEs.

Workaround: Use 4K RX queue size: ethtool -G <intf> rx 4096

Keywords: Performance, MH, WQE

Discovered in Version: 22.27.1016

-

Description: After programing firmware in LF, power-cycle must be recovered.

Workaround: N/A

Keywords: LF

Discovered in Version: 22.27.1016

2029716

Description: Software Reset does not work on ConnectX-6 Dx adapter cards.

Workaround: N/A

Keywords: Software Reset

Discovered in Version: 22.27.1016

© Copyright 2023, NVIDIA. Last updated on Nov 3, 2023.