Eth0 fix missing for meta-fbtp build on Portwell Neptune Board

[twlevin]

I have a Portwell Neptune board that came with fbtp-2168080 loaded on it. When I downloaded the current facebook/openbmc, recompiled for meta-fbtp, and loaded on the board, I get the following error during initialization and no Ethernet.

Configuring network interfaces... [ 27.902809] Found NCSI NW Controller at (0, 0) [ 30.399163] NCSI error: Unknown Mezz Vendor! [ 31.059169] NCSI: MAC 00:11:22:33:44:55: [ 39.559173] Using NCSI Network Controller (0, 0)

The fbtp-2168080 originally on the board has the following text during startup: eth0 mac fixup ethaddr= 72:42:FD:E8:BF:7F use u-boot mac addr Setup crond...starting Busybox Periodic Command Scheduler: crond... done. done. crond[384]: crond (busybox 1.24.1) started, log level 8 Starting mTerm console server...done. Remove poweroff command from .bash mTerm_log: mTerm: daemon started mTerm_log: mTerm: daemon started INIT: Entering runlevel: 5 Configuring network interfaces... [ 27.330000] bwdbg03 - rtl8211 config aneg done.

When I contacted Portwell about the issue, I received the following response: So when we were developing the kit with FB, they built the images for us Later they had found an issue with Ethernet(NCSI) ports, so made another version to fix it(eth fix) FB was in charge of the BMC image side development, not sure if they put this on repository as you mentioned Unfortunately, source code of the image was never release to us But you may find some resources on the Github forum or post questions for it, or start a new image build

Is anyone in the community aware of this issue and where I might find the fix? Does facebook still support this build and are they aware of this particular fix? I have attached the full boot logs for reference. originalPortwellBoot.log NoEthernetErrorBootLog.txt PortwellEth0Issues.docx

[sandeept8]

Hi @twlevin, Apologies if this is not relevant for you anymore, Did you find any fix for this issue ? Do you still use the openBMC board as of today ?