making a free bios for a present notebook x86 mainboard?
I have read about free bios. And talked with computer persons about free bios. I have gotten the impression, that you cannot know in advance if a free bios can be made. I have read about libreboot on wiki. It was possible to make a free bios for the thinkpad x60 because hardware manufactures handed over enough source software and documentation? Else it could not have been done?
Is it a matter of money to get a free bios for a present notebook mainboard? Is it possible to narrow down a couple of notebook x86 mainboards having the highest likelihood they could get a free bios if knowledge and resources were allocated? Or have all plausible efforts been made? Taking into account how many computer science departments US have, is none of them working on a free notebook mainboard bios? Can you xray a mainboard to provide documentation?
AIUI some organisation paid big money for the reverse engineering and
coding involved in porting coreboot (free bios) to the x60 because
they really wanted it for a big contract. And money along with dev
time & skills are the main limiting factors in why we don't have more
supported laptops / desktops.
RMS/FSF have said 'more reverse engineering please' to urge more devs
to do it, so things might change.
I have seen a video where he quite directly encourage corporate engineers to whistle blow about source software and documentation.
X60 is supported since a coreboot developer spent much time reverse
engineering one chip (needing expensive equipment). All other parts
were supported earlier, since the German government funded i945 support
(which used documentation from Intel and needed many months of work).
For a present notebook: see what newer Thinkpads are supported. Intel
ME firmware (needed by all their modern chipsets except for embedded
ones with other issues, on many used for AMT) practically cannot be
replaced. For AMD you need months of reverse engineering work to
replace graphics firmware (outside of the scope of boot firmware, but
needed for a free system; if you did it, newer Loongson computers would
be RYF-endorsable if sold with an appropriate distro); several other
blobs need work too. Both CPUs might need nonfree microcode updates
(known to be required on some Intel CPUs) which would prevent it from
being FSF-endorsable. ARM would be easier, but all ARM laptops use bad
SoCs which need blobs (Samsung Exynos), maybe Tegra Chromebooks will
change this (while they need other blobs). Making custom laptops with
appropriate ARM chips is probably beyond the scale at which RYF-endorsed
products could be sold.
Xraying boards is not usually done: it needs special rarely-available
equipment and the needed information can be obtained from hardware
registers or by observing what the proprietary boot firmware does. Some
developers remove chips from their boards and trace their connections,
it might help in some cases.
It's not computer science. I won't comment on what computer science
departments do.
Thanks. When I ask a question about this subject the replies are like Pandora's box. Things I do not know about appear and it is complicated. AMD would require reverse engineering on the graphic card and the boot software. And more. Difficult but not impossible? Loongson mainboards only require reverse engineering of the graphic card? The rest is prepared for free software? How is the performance of the loongson hardware you are referring to compared to x86 cpus? I should want to know, what graphic hardware the 3b mainboards get.
The free hardware i945 intel core duo is fast enough to run current gnulinux versions. New mainboards are not available. What are the obstacles in starting production again? Is intel controlling the licenses and they will not permit a production to start again?
Loongson 3A is slower than Core 2 Duo in my X60t. All 3B boards listed
at http://www.loongson.cn/devsys.php have AMD GPUs, same as for 3A.
Making custom laptops for 100% free systems is not practical unless you
can sell many thousands of them.