First Flight of Kestrel, the FPGA OpenPOWER-Based BMC, and Arctic Tern

lol, read only jumper on the fpga object code eprom... boy that was a close one - we just narrowly avoided the horror of user controlled hardware.

The end product will be still FPGA and will be end-user programmable. This is a core differentiator for some of Raptor's client base, and I do not mean only individual RYF hardware fans - the design means that the system is also fully auditable by owners including ability to rebuild all firmware images from scratch.

The FPGA on the motherboard which handles some of the glue logic is also user programmable this way.

> ...I'm pretty sure that as an end user I would not feel comfortable...

The fact that you have access to all the information you need to do the job should alleviate any concern. It is a night and day difference from blindly hacking on hardware from a hostile manufacturer. Did you actually mean "I don't want to"?

>Looks cool, but I'm pretty sure that as an end user I would not feel comfortable reprogramming the board management controller of a $5000 dollar piece of equipment.

I don't think the point is that they expect you to start tearing into the code, it's more so that you can feel comfortable that if you have any questions at all about the code running on it, you can rebuild from source and reflash to verify yourself.

I also may be underselling the ability of this to harm your system, but I'm not sure what you'd possibly do that wouldn't be reversible. Bricking the BMC itself means replacing the BMC (although if you can flash it I think it'd be difficult to COMPLETELY brick) - maybe a motherboard if they happen to fully integrate it but that doesn't look like the plan at this point looking at the size of the add-on board.

> Looks cool, but I'm pretty sure that as an end user I would not feel comfortable reprogramming the board management controller of a $5000 dollar piece of equipment.

Why would it be any different than flashing a firmware image to a $5000 box?

Probably people like me: willing to put their money where their mouths are when it comes to security and actual hardware ownership. Obviously that is a vanishingly small market - as demonstrated by the fact that a third of the comments here are from me. There is a very interesting benefit from truly open hardware ownership that I hadn't anticipated, and is kind of hard to explain to anyone who hasn't experienced it themselves: reverse engineering is no longer a time-sink. I'd grown so accustomed to having to fight through the manufacturer's hostile anti-owner practices, while doing work that they benefit from for free... it kinda blew my mind when I hit a speedbump on my Blackbird - and was able to address it by referring to an actual schematic included with the board. Usually that step would have taken hours of scouring translations of archived Russian TV repair sites and Chinese DHT search engines - and rarely yielded anything beyond mislabeled zips containing boobytrapped boardviewer trojans. So for anyone who has ever tried to fix something, and found themselves fighting hardware that they supposedly own - figure that massive benefit into the price. But that too is a small market.