I wanted to set up some old Optiplexes with Coreboot for a Windows XP LAN party. I ended up going through several motherboards with bad flashes, both of Core and Libreboot
Eventually I rolled up in their IRC and basically asked what the hell I was doing wrong.
"oh yeah we disabled [some feature]. You need to dump a bunch of firmware blobs first and then add those to the build"
Even after doing that though I found I was able to boot SeaBIOS (but nothing else!) and Windows would similarly just immediately crash and burn trying to load in any capacity. Linux worked "fine"
More annoyingly it was both slower than the vendor firmware and SeaBIOS would also hang on an LSI SAS card I had inserted (I guess it was trying to run the card OptROM?) which was annoying because I wanted to use the machine to flash the card to "IT mode" (aka JBOD mode)
* For laptops that require extensive disassembly, once you realize that one day you might have to recover a bad software-based flashing, requiring disassembly again, you might be willing to cut a hole in the laptop: https://www.neilvandyke.org/coreboot/#cutouts
* Something like an IBM/Lenovo Hardware Maintenance Manual is a big win. Lately I even label removed screws with the page or section number of the manual that described their removal. https://www.neilvandyke.org/coreboot/#intro
* The flaky in-circuit reading&writing of the flash chip with hardware probes ranges from frustrating to maddening, for software person me. For example, I had an "assembly line" for Coreboot-ing seven ThinkPad X200 units, but in-circuit reading and writing failed different ways, multiple times, with pretty much every unit. And I fried a rare-ish variant of the ThinkPad T60 motherboard (I guess when current took a path it really wasn't supposed to).
* Had I only needed to do one laptop, afterwards I'd think that was rough, but that it would go much smoother next time, once I knew what I was doing. Not so. I never knew what I was doing. I did about 10 laptops, and each one was a pain. Maybe an EE/CE would do it much better, with better technique, or better tools?
* I really like the idea of open source boot firmware (and I hate UEFI complexity and dodginess). Now, if I could just get Coreboot and a good TrackPoint keyboard on Purism/Framework/System76/etc....
> * [flagged] "It's not that bad" they told me – Coreboot (lilysthings.org) | 42 points by todsacerdoti 2 hours ago | unvote | flag | hide | past | favorite | 11 comments*
Still a bit bummed that pcengines closed their doors, By far my favorite computer vendor. I miss their little routers and am still looking for a good alternative.
https://doc.coreboot.org/payloads.html says you can use grub as a payload; does that work?
There were two downsides, Windows 11 cannot sleep, or rather it sleeps but cannot wake up anymore, and you cannot adjust permanently the default drive (I believe you can if you reflash).
For people using Linux it works perfectly, and accessing the chip is quite easy, you don’t need a full disassemble.
I DO NOT WANT TO ONLY HAVE MBR BOOTING
Incidentally, I'd consider MBR boot an advantage compared to UEFI. As Linus says, it's "this other Intel brain-damage".
There's also "MBR the partitioning scheme", which does have a limit of 2^32 sectors or 2TB, but that's not exactly the same as "MBR the boot method".
So I buy a cheap clip and programmer kit... and get nowhere, nothing is working right and I have no experience with this sort of thing so don't know what to expect. thus it sat on the healing bench for a year. I chanced on an article that having a good clip was critical for this sort of thing and decided to try again after buying a recommended clip. and the infernal thing worked first try.
So while I still have no idea what I am doing, I am now a firm believer in having a good clip. anyhow, here is the one that worked for me.
https://www.pomonaelectronics.com/products/test-clips/soic-c...