upgrade to 25.7.2 from 25.7.1 killed my wan interface (intel i210)

[lebowski]

I feel pretty stupid. And to think that, at the time when i installed all this cabling, i invested in premium cables, cat7 sftp pimf with shielded connectors, made everything by hand, pulled parts of the cable through cableducts in my walls, used premium shielded cat6 wall sockets, because i wanted to make sure.... THIS wouldn't happen. And yet it did. And i have to say, the cable that probably was causing issues, i can't see anything wrong with it, it just doesnt work. So now, to be extra sure, i placed a temporary cable which soon will be replaced by a brand new factory made cat7 sftp pimf cable (has been ordered, is on it's way). I don't want to have this misery again.

I yet have to check the link you send me for the 700mb archive in the previous post. Do you think it still would be a good idea to try and update firmware for my i210-at controllers? Or is it as they say: "if it works, don't touch it"?

[BrandyWine]

I feel pretty stupid. And to think that, at the time when i installed all this cabling, i invested in premium cables, cat7 sftp pimf with shielded connectors, made everything by hand, pulled parts of the cable through cableducts in my walls, used premium shielded cat6 wall sockets, because i wanted to make sure.... THIS wouldn't happen. And yet it did. And i have to say, the cable that probably was causing issues, i can't see anything wrong with it, it just doesnt work. So now, to be extra sure, i placed a temporary cable which soon will be replaced by a brand new factory made cat7 sftp pimf cable (has been ordered, is on it's way). I don't want to have this misery again.

I yet have to check the link you send me for the 700mb archive in the previous post. Do you think it still would be a good idea to try and update firmware for my i210-at controllers? Or is it as they say: "if it works, don't touch it"?

You need a cable tester that can do near/far x-talk testing, continuity tests, etc.

As for i210 NVM updating? It's like many warnings that come with updates, "if you are not trying to fix a specific issue.....".

Can the update make things better? Possibly. I just not sure what metrics you would be measuring. I also lean on the other side of that fence, if they made a new NVM image then why did they make it, they didn't do the effort just for fun, etc.

I also am wary of the china made stuff, the maker of the device could have loaded in their own NVM image, and nobody really knows what that code is. I think the NVM loader tool allows you to extract out a copy of the EEPROM code, from there you could look at it in hex editor or some diff tools to see if it's real Intel code or has been modified (comparing same NVM versions as example). From what I have seen, the bin files are highly padded, way more EEPROM space than actual code, so this gives plenty of room to place more code into the bin and load it in.

This is a big battle for say OPNsense who tries to help support via community forums. Unvalidated hardware is a nightmare to deal with, and here we have a gazillion people using all sorts of hardware along with VM's, and then everyone comes here to complain. If "you" want validation then buy an official OPNsense device. It's that simple.

All that said, I guess the community tries to help the community, but many don't have the skillset to dive in and look around and then fix when fixing is needed.

[lebowski]

I completely agree about the cheap chinsese stuff; let alone being it a possible fire hazard, one does not know which code has been "programmed" in it's various firmwares. This is why i'm so happy with my us made supermicro appliance. I'm wondering myself if i should go ahead and update the nic's firmwares. As you said, they don't release firmware updates for nothing. On the other hand, a professional brand like supermicro would release these firmware update packages when it would be needed i guess. I'm a novice in this, but i guess they know what they're doing? And there isn't an update package for my appliance, which makes me wonder if there is anything to gain except a nasty brick of my nics making my beautiful appliance unuseable.

About fixing stuff: i'm no *nix expert. The great thing about opnsense in my opinion is that it is usable for novice/medium experienced users, not only hardcore network engineers, in contrary to (for example) mikrotik, which can't be used if your not some toughguy network engineer (is my experience). Opnsense gives you the robustness of a bsd operating system and the user friendlyness of a "working out of the box" product.

I think i will contact supermicro and ask them how they feel about updating their nics on my appliance. Maybe they have a custom build update package for internal use, which they care to share?

[BrandyWine]

[1] igb1: EEPROM V3.16-0 eTrack 0x800004d9

It's was on 3.22 now? Hmmm, that does not seem to align with your output post #17. Your EEPROM looks like 3.16, or it was.
Also, I do believe v3.30 or higher is available from somewhere.

Maybe the tool did do upgrade to 3.22. What does "dmesg | grep igb" show you now?

Also to note, integrator's like SM, Asrock, ASUS, HP Dell, etc etc etc, don't care to build firmware upgrades for flashable devices unless they really need to. Some like Dell and others will provide flash images when security issues demand it. Sometimes the integrator builds the firmware to suit their needs or restrictions, other times a device manufacturer has full-featured firmware. As example, some years ago I had a DVD-rw drive in my PC, but the model I had could not rw Blueray. Well, turns out it was only firmware change that was needed, I exported my existing firmware and applied a mod, now it can rw blueray (had to do it this way because the firmware is very specific to some things in the drive). The maker of dvd drive sells the blueray-enabled device for more money, same drive, different words on the box, different firmware.

The fun world of firmware and drivers.

Yes it was on 3.22 . The screenshot which i made (and failed to upload) shows 3.22 being the decimal version, and 3.16 the hex version. This is a bit confusing. The tool showed for both nics "Status: update not available". And then "Tool execution completed with the following status: Device not found." (although before this sentence it shows the both i210 nics).

I perhaps am wrong, and you are right.

I ran the nvmupdate64e to get inventory, and it shows my NVM for i226 as
"NVM Version            : 2.23(2.17)"

23dec = 17hex

I do now notice the NVM bin files are named using the hex value.