@mangeurdenuage Isn't the whole of point of Coreboot to have a FOSS microcode initializer on applicable PCs? What are they doing with proprietary code?
@SuperSnekFriend The original point of coreboot was to have free software init for computers, but the project has regressed into "FOSS" and is proprietary software galore.
Even if hardware can go without it (like all CPUs really), they go include proprietary microcode updates and configure things assuming proprietary microcode updates have been applied.
Any intel CPU after 2008 and any AMD CPU after 2015 no longer inits without proprietary software that has a signature on it to handcuff you and prevent you from replacing it with free software.
For CPUs after 2008, intel offers proprietary "reference init binaries", which go init the hardware and does RAMinit and god knows what and then jumps back to the specified target.
AMD has announced they are offering the same, except it appears that AMD will be doing the GPLv2 infringement rather than the coreboot developers.
AMD previously did release partial source code of AGESA, so raptorcs was able to add fully free support for KGPE-D16 (although they didn't bother to remove unused proprietary microcode array encoded in the sources), although they had to write RAMinit from scratch and add comments back, as all of them were stripped out, as old AMD didn't even provide full source code.
@pista Yes, certain non-AMD64 computers have free bootloaders and don't use coreboot.
Firmware is socketed ROM, which cannot be bricked with software - you flashed the bootloader software.
Coreboot is designed to carry out the crufty x86 or AMD64 init sequence and generally on architectures like Aarch64, you would use a free version of u-boot instead.
There is always pmon2000, the bootload from “the RMS laptop” (Lemote Yeeloong 8089).
You could actually test out new builds by loading it via existing PMON in ram and running it. Of course, actually flashing that was putting a revolver to your head. Bricked firmware more often than not.