Bug Fixes in this Firmware Version
Internal Ref. | Issue |
4823907 / NVbug 5742181 | Description: Fixed an issue where, in certain configurations with the ConnectX-8 PCIe switch enabled, downstream devices (including GPUs) might not be detected and could drop from the PCI bus, with GPU sensors/properties reporting nan. This was caused by the device not receiving the required PERST# assertion during initialization, and was seen only when PCIe settings were manually modified via mlxconfig (e.g., restricting link speed/width or ASPM on specific PCI buses). Note: On legacy firmware, additional configuration steps may still be required, as detailed below. If you cannot update the firmware immediately, you can restore device detection using one of the following options:
|
Keywords: ConnectX-8 PCIe, GPU, PERST# assertion | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4786813 | Description: Fixed an issue where the DPA kernel used unsafe ICM access during process creation/modification, which could cause the DPA kernel to hang during FLR. |
Keywords: DPA kernel, FLR | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4884739 | Description: Link failures may occasionally be observed at PAM4 speeds over optical interfaces in rare cases. |
Keywords: PAM4 speeds, optical interfaces | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4804664 / 4806969 | Description: Fixed an issue in the User Debugger “query caps” where it returned only the number of capabilities, not the capability bitmap. |
Keywords: User Debugger “query caps” | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4813862 / 4146077 | Description: Fixed an issue where CR dumps could time out when accessing xpl_top addresses across all three pcores. |
Keywords: CR dump | |
Detected in version: 40.47.1026 | |
| Fixed in Release: 40.48.1000 | |
4833440 | Description: Fixed an issue where the Virtio and NVMe EMU_MNG settings were exposed incorrectly, which could cause confusion when using mlxconfig. |
Keywords: Virtio and NVMe emulation, mlxconfig | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4756451 | Description: Fixed an issue where the PHY LED could show green during the initializing state when active speed was set to full speed. In IB mode, the initializing-state LED should be amber only. |
Keywords: PHY LED | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4768546 | Description: Fixed an issue where, on multi-PF-per-port systems, a PF FLR could impact the traffic bandwidth of other PFs on the same port. |
Keywords: PF FLR | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4705918 | Description: Fixed an issue where PTP could converge to an incorrect time/offset and report an inaccurate path delay. |
Keywords: PTP | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4657792 / NVbug 5567725 | Description: Fixed an issue where, in Flit Mode, the device could become unresponsive when receiving malformed or invalid traffic from a link partner. |
Keywords: Flit Mode | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4484662 | Description: Fixed an issue where mlxlink reported 0 values for SNR (media and host) due to incorrect local port mapping in firmware and an incorrect page number used by MFT. |
Keywords: mlxlink | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4621747 / 4792742 / 4794290 / NVbug 5502241 | Description: Fixed an issue where parallel accesses to the MCIA register could return incorrect data. In some hosts running |
Keywords: MCIA register | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4686284 / NVbug 5607036 | Description: Implemented IB extended port telemetry counters via the NSM Type 1 Get Port Telemetry Counters command, adding counters 19 and 20: |
Keywords: IB extended port telemetry counters, NSM | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4758174 / NVbug 5698200 | Description: Fixed a rare attestation certificate signature formatting issue by removing an unnecessary leading zero byte in the “r” or “s” value. |
Keywords: Attestation certificate signature format | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4532684 / 4635872 / 4794865 / 4794866 / 4794867 / NVbug 5385446 | Description: Fixed an issue by improving the ADP-RETX algorithm to avoid re-arming without performing a retransmission. |
Keywords: ADP-RETX algorithm | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4542516 / 4554220 | Description: Fixed an issue where, in certain Gen6 setups, RDMA READ bidirectional traffic required at least 5 QPs to reach full wire speed. |
Keywords: RDMA READ bidirectional traffic | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4554763 / 4808657 | Description: Fixed an issue affecting single-process, unidirectional RDMA READ to GPU memory (4 QPs, 128KB messages) by enabling |
Keywords: Zero Touch Tuning, mlxconfig | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4608214 | Description: Fixed an issue where probe packets might not be sent under heavy traffic. |
Keywords: PCC, ZTR_RTTCC, probe | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4450570 / 4780432 / 4780433 | Description: Fixed an issue where the root complex sent MCTP-over-PCI messages before a BDF was assigned, causing responses to be sent with BDF 0. The fix ensures that MCTP messages routed by ID are ignored until a valid BDF is assigned. |
Keywords: MCTP-over-PCI, BDF, MCTP messages | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4809134 / 4824635 | Description: Fixed an issue where the steering tables were not updated after enabling partial Spectrum-X capabilities (BTH.AR) via LLPD. |
Keywords: Steering tables, LLDP | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 | |
4797308 / NVbug 5706024 | Description: Fixed an issue where an intX message was sent with a Requester ID of 0, causing an ACS violation at the root port. The fix uses the correct BDF as the Requester ID instead of 0. |
Keywords: intX message | |
Detected in version: 40.47.1026 | |
Fixed in Release: 40.48.1000 |