4 ms·
Adopting Open Source Firmware Approach for Intel FSP
- userbinator 4y agoProvide more technical documentation This, absolutely this. I don't care at all if they keep their code closed-source; just release the docs and we'll do it ourselves. In fact, I'd say that's better than open-source, since the majority of open-source HW code seems to be more of a "poke this value into this register" thing that's not much better for understanding how to use the hardware than reading the disassembly. Sadly, I'm not so optimistic about this because Intel has been infected with people from the pro-DRM (and "security") crowd, who would very much like for computers to be controlled by them instead of the end-user, and thus there is functionality they would like to keep secret. I think Intel started closing up visibly in the early 90s with the infamous "Appendix H", and it's only gotten worse since then.
- marcodiego 4y agoThey could release just enough information to allow a FLOSS implementation of the FSP firmware sans DRM. Those of us who don't care about using DRM can live without it.
- nonrandomstring 4y agoAuditability is the key phrase. There are lots of things I'd like to see in silicon that would massively improve security, sustainability, reuse, resilience, privacy and a whole lot of other good things... but unless they can be verified via external software or with everyday test-bench equipment, nobody but major laboratories with decap and electron microscopy can know for sure. For one thing I'd like to see a physical jumper pin or one time fuse to hard disable non-core features like ME or TPM enclaves. Microprocessors that can be trusted by the user are essential to a viable civic cybersecurity model.