I have a Mellanox ConnectX card MHGH29-XTC (HP OEM 483514-B21) which won't work on VMware ESXi 6.0.0 and Linux. It seems that the problem was failure firmware upgrade.
The card was originally come with firmware version 2.6.0, i have tried upgrade to latest 2.9.1 download from Mellanox website. But unfortunately it won't work, so i fallback the firmware to old version (to 2.6.0). The card was become normal again.
After do some research on internet, some guy said we could repackage the firmware by modifier the OEM configure file. I was follow the procedure, repack the firmware for version 2.9.1 and then burn to the card. During the process without any error, it was successfully burn the firmware to the card and verified by the "flint" command. After reboot the system, the ESX server (Linux also not work) was not able to detect the card. I am guessing that might firmware issue which i have tried before. So i was re-burn the firmware version 2.6.0 which is originally come with the HP card.
Unfortunately, no matter which firmware version i was re-burn onto the card. It was not functional again (No matter is in ESX mode / Linux mode or Windows mode). May i know is there a way could completely cleanup and reburn the firmware onto my OEM card, and make it functional again?
Please advise anything i could do to repair the card? Thanks a lot.
MHGH29-XTC (HP OEM 483514-B21)
Dual ports 4xDDR Infiniband Card
Tried Firmware Version
- [Work] v2.6.0 : Come with the HP OEM Card
- [Not Work] v2.9.1 : Download from Mellanox website (fw-25408-2_9_1000-MHGH29-XTC_A1.bin)
- [Not Work] v2.9.1 : Download the source from Mellanox Website, repacked with customized Configure File.
Reference website "Custom Firmware for Mellanox OEM Infiniband Cards - WS2012 RDMA"
- [Not Work] v2.6.0 : Re-burn the original HP firmware come with the card
- [Not Work] v2.6.0 : Download the HP OEM Firmware from Mellanox website
# lspci
0000:04:00.0 Serial bus controller: Mellanox Technologies MT26418 [ConnectX VPI - 10GigE / IB DDR, PCIe 2.0 5GT/s] [vmnic2]
# mst start
Module mst loaded successfully
# mst status
MST devices:
------------
/dev/mt26418_pci_cr0
/dev/mt26418_pciconf0
* Without execute the "clear_semaphore" command, the card will never response for operation including query after the failed OEM firmware upgrade.
# flint -d /dev/mt26418_pciconf0 query
-E- Cannot open Device: /dev/mt26418_pciconf0. Can not obtain Flash semaphore (62). You can run "flint -clear_semaphore -d <device>" to force semaphore unlock. See help for details.
* Execute "clear_semaphore" command.
# flint -clear_semaphore -d /dev/mt26418_pci_cr0
Warning: Taking flash lock even though semaphore is set.
Now is able to query the card and response for the command.
# flint -d /dev/mt26418_pciconf0 query
Image type: FS2
FW Version: 2.6.0
Device ID: 26418
Description: Node Port1 Port2 Sys image
GUIDs: 00237dffff930fb4 00237dffff930fb5 00237dffff930fb6 00237dffff930fb7
MACs: 000000000000 000000000001
VSD:
PSID: HP_09D0000008
# flint -d /dev/mt26418_pciconf0 hw query
HW Info:
HwDevId 400
HwRevId 0xa0
Flash Info:
Type M25Pxx
TotalSize 0x200000
Banks 0x1
SectorSize 0x10000
WriteBlockSize 0x10
CmdSet 0x80
# flint -d /dev/mt26418_pciconf0 v
FS2 failsafe image. Start address: 0x0. Chunk size 0x80000:
NOTE: The addresses below are contiguous logical addresses. Physical addresses on
flash may be different, based on the image start address and chunk size
/0x00000038-0x00001033 (0x000ffc)/ (BOOT2) - OK
/0x00001034-0x0000489f (0x00386c)/ (BOOT2) - OK
/0x000048a0-0x0000521f (0x000980)/ (Configuration) - OK
/0x00005220-0x00005263 (0x000044)/ (GUID) - OK
/0x00005264-0x00005387 (0x000124)/ (Image Info) - OK
/0x00005388-0x000062a7 (0x000f20)/ (DDR) - OK
/0x000062a8-0x0000f1d3 (0x008f2c)/ (DDR) - OK
/0x0000f1d4-0x00046fc3 (0x037df0)/ (DDR) - OK
/0x00046fc4-0x00058b93 (0x011bd0)/ (DDR) - OK
/0x00058b94-0x0007987f (0x020cec)/ (DDR) - OK
/0x00079880-0x0007b55b (0x001cdc)/ (DDR) - OK
/0x0007b55c-0x0007e30f (0x002db4)/ (DDR) - OK
/0x0007e310-0x0007e827 (0x000518)/ (Configuration) - OK
/0x0007e828-0x0007e873 (0x00004c)/ (Jump addresses) - OK
/0x0007e874-0x0007e9f7 (0x000184)/ (FW Configuration) - OK
-I- FW image verification succeeded. Image is bootable.
# mlxfwmanager
Querying Mellanox devices firmware ...
Device #1:
----------
Device Type: ConnectX
Part Number:
Description: HP-EAST EAGLE IB DDR G2 R5 KIT
PSID: HP_09D0000008
PCI Device Name: /dev/mt26418_pci_cr0
Versions: Current Available
FW 2.6.0000 N/A
Status: No matching image found
Device Configuration file
# flint -d /dev/mt26418_pciconf0 dc
;; Generated automatically by iniprep tool on Mon Dec 15 18:55:56 IST 2008 from ./custom_eagle_ddr_gen2_dpdp.prs
;;
;; PRS FILE FOR Eagle
;; $Id: custom_eagle_ddr_gen2_dpdp.prs,v 1.1 2008/12/14 16:25:34 yarong Exp $
[PS_INFO]
Name = 483514-B21
Description = HP-EAST EAGLE IB DDR G2 R5 KIT
[ADAPTER]
PSID = HP_09D0000008
pcie_gen2_speed_supported = true
adapter_dev_id = 0x6732
[HCA]
hca_header_device_id = 0x6732
hca_header_subsystem_id = 0x6732
dpdp_en = true
[IB]
mellanox_ddr_ib_support = true
[PLL]
core_f = 44
core_r = 27
[FW]