Jason Self at
A sad summary of a discussion in #libreboot on Freenode today:
AMD has been contributing source code to coreboot since 2011 but now they have stopped. AMD has started to release it as binary-only from now on, and the coreboot developers are accepting the submissions: http://review.coreboot.org/#/c/6274/ (It has +2, indicating it will be merged.) When they were submitting in source form I hear:
- there are parts that are for newer hardware (not yet released) that AMD has to remove before submitting to coreboot
- there are parts that have to have comments removed before submitting to coreboot
- there are other parts that also have to be removed for "other reasons"
(But otherwise, what they were submitting was free software.)
That "cleaning" process above, as the coreboot devs say, costs AMD a "lot of money" in terms of engineering time and AMD has ceased releasing source code for new AGESA versions, as a way of cutting costs, so any AGESA source that you see in coreboot is prior to AMD's decision to no longer release it.
coreboot has an opportunity to correct AMD's behavior. AMD relies on coreboot for a lot of their hardware, so coreboot developers should have at least some influence over them. If they refuse to accept new submissions from AMD now, then it creates a precedent: either AMD reverses the decision (releases source again) or coreboot recieves no additional support from AMD in the future. By accepting blobs now, they are saying to AMD that it's ok: this makes it harder to get them to reverse the decision in the future, as time goes by.
AMD has been contributing source code to coreboot since 2011 but now they have stopped. AMD has started to release it as binary-only from now on, and the coreboot developers are accepting the submissions: http://review.coreboot.org/#/c/6274/ (It has +2, indicating it will be merged.) When they were submitting in source form I hear:
- there are parts that are for newer hardware (not yet released) that AMD has to remove before submitting to coreboot
- there are parts that have to have comments removed before submitting to coreboot
- there are other parts that also have to be removed for "other reasons"
(But otherwise, what they were submitting was free software.)
That "cleaning" process above, as the coreboot devs say, costs AMD a "lot of money" in terms of engineering time and AMD has ceased releasing source code for new AGESA versions, as a way of cutting costs, so any AGESA source that you see in coreboot is prior to AMD's decision to no longer release it.
coreboot has an opportunity to correct AMD's behavior. AMD relies on coreboot for a lot of their hardware, so coreboot developers should have at least some influence over them. If they refuse to accept new submissions from AMD now, then it creates a precedent: either AMD reverses the decision (releases source again) or coreboot recieves no additional support from AMD in the future. By accepting blobs now, they are saying to AMD that it's ok: this makes it harder to get them to reverse the decision in the future, as time goes by.
Vasco Dias, Yutaka Niibe, Mike Linksvayer, testbeta and 3 others likes this.
Kete Foy, Claes Wallin (韋嘉誠), maiki, maiki and 3 others shared this.
@moggers87@microca.st Yes, blobs in coreboot is "nothing new", all routine and totally acceptable, right? (Not.) Anyway, going from free to non-free represents a regression, IMO. It is unfortunate that some people dismiss it.
Does coreboot have an explicit rationale for this? I see http://www.coreboot.org/Binary_situation "While we aim for a 100% free boot process, recent developments (and
general unwillingness by some hardware companies to provide
specifications) make it hard to achieve" but this does not actually explain their thinking. Compare with Mozilla, which has explicitly said (I'm paraphrasing) that they're doing non-free formats and DRM because if they don't they believe they'll lose market share, and in order to have any leverage they need market share. At least that's a calculation that one can argue about.