[coreboot] Removal of AGESA - to be or NOT to be? A story of G505S

[Alex G.]

On 11/07/2015 06:12 AM, Vladimir wrote:
> That is why Francis suggested adding it to coreboot LTS Candidates list,
> not to Libreboot list - because, while Lenovo G505S seems to be a great
> coreboot LTS Candidate , it is still not ready for Libreboot.

LTS requirements were meant to weed in hardware that is truly free. We
failed at that. libreboot succeeded where we failed. If it's not in
libreboot, it doesn't qualify for the LTS list.

Also, I doubt Francis appreciates people using his name and the
libreboot name in a way which makes it seem like he, or the libreboot
project endorses blobbed hardware. The libreboot project does not
endorse the Lenovo G505s.

> Which is a real pity, because there are no AMD laptops supported by Libreboot

And there most likely won't be. I did the bringup for the first AMD
laptop ever to be supported by coreboot. The g505 is almost an identical
copy of my work. I used AGESA, and spent over half the time debugging
and fixing AGESA to make it suitable for the task. I am very intimate
with the details of that hardware, and why the chances of it ever
hitting libreboot are infinitesimal.

> yet...

Nope. I've tried. You need linux to use the integrated GPU without a
VBIOS in order for this to work to libreboot standards. The maintainers
of the radeon driver have made it very clear that they will have none of
that. In plain english, the coreboot image will have to bundle the VBIOS
in order for the machine to work as intended.

> In any case, Lenovo G505S matches all the requirements for it to
> be considered among the Coreboot LTS Candidates list :
> 
> * RYF-certifiable - [V] check

Nope. See above.

> * Sturdy - [V] check

Really? I was going to buy one a few years back. It did not feel sturdy
at all.

> * Long shelf life - [V] check

Less than three years left? I think the bar here is closer to ten years.

> * Cool factor - [V] check

The low-end lenovo machines are anything but cool, but then this is a
subjective matter, so I'll let you win this point.

> Actually it was you who came up with these requirements :
> 
> http://www.coreboot.org/User_talk:MrNuke/LTS_Candidates#Requirements:

Yup. Yours truly.

> Sorry, but it is a really bad idea to treat LTS Candidate as a second
> grade citizen

The LTS idea died over a year ago because there wasn't any hardware to
match the requirements. The libreboot project took those ideals and
whatever hardware they support is essentially LTS.

If you want hardware that is true in spirit to the LTS requirements,
then please head over to http://minifree.org/ .

> Lenovo is open to requests of motherboards (as well as requests for
> [...]
> sale tax), so any physical person could get them even at 1pcs quantities ;)

Do you have a link to Lenovo's policy on such?

> Undoubtedly, Francis Rowe

I will stop trying to prevent people from tarnishing his or libreboot's
reputation by suggesting he or the libreboot project endorses blobbed
hardware. Just pointing out that they don't endorse it.

> please keep in mind that coreboot is GNU GPL project, and

Thank you master for enlightening me! Let's see... coreboot is _not_ a
GNU project. It is licensed to you under the terms of the GPLv2.

> you have transferred your G505S code to public domain

I did? I was pretty certain I reserved all rights and licensed it to
other people (you included). Maybe GPLv2+ got lost somewhere in translation.

> Imagine, what if someone would suddenly decide to revert/branch off all
> his significant contributions to a Linux kernel? People simply would not
> let it happen! So I doubt that you can do anything you want with your
> important commits to coreboot repository, be it a branch or anything
> else as significant,

> without considering the opinion of other coreboot people

There are coreboot users, coreboot developers, and coreboot developers
whom I have worked with in the past, and whom have earned my trust and
respect. The mostly care of the opinions and thoughts of the latter,
while also being sensitive to the needs of the middle, and trying to
accommodate the former.

But you'll have to find someone else to support you in your endeavors
with the code in question. You can't require that I continue to support
you. I'm on the Tim Pearson bandwagon now, with native init, and much
cleaner and easier to integrate code.

> First of all, it would be interesting to see a detailed report
> describing a process of how you came up with such a round value ( $1300> )

electricity costs : $50/month * 12 months = $600
Flying in for maintenance:
unpaid workday: $400 + plane ticket cost $520 = $920
replacement hardware costs $170 x 3 = 510
Total: $2030
60% if that is $1218, so I was off by $82. sue me.

Note that that doesn't include vehicle rental costs, gas costs,
transportation to/from airport, etc.

> It is not that fair for you to ask $1300 from a single person

My asking for payment for my running costs was a rhetorical statement on
the issue of making efficient use of resources, and not a bill.

> you could set up a crowdfunding at Crowd Supply

I don't mind paying the bills. What I do mind, is people taking that for
granted and demanding that I support them out of my own pocket, then
calling me out and throwing baseless accusations at me when I suggest
making more efficient use of the limited resources we have.

> Actually I have expressed my desire to participate in AGESA -> Native
> Init porting,

Gerrit is open for business 24/7.

> but as I said before I can't do it all alone because I don't have enough experience.

The only person who has the needed experience to port to that codebase
is Tim Pearson. Everyone else is still learning.

> In my earlier messages I already told what sense I am putting behind
> these words.

> Master branch will be receiving all the love: important bug
> fixes as well as great new features,

That doesn't mean that AGESA boards in master tree will receive the
same. That part of the codebase has matured and stabilized, even if that
means "don't touch it because it's so fragile".

> Meanwhile, nobody is going to constantly track check and copy all these commits to this separate
> branch, 

That's already happening (or not happening) today. If I add a new
feature and part of the codebase requires major rework to accomodate
that feature, I just make it optional. A lot of other developers operate
that way. Making changes of global nature is extremely difficult.

> which means it would be abandoned.

There are still plenty of users of that codebase that I suspect the
branch would be very much alive for years.

> I am not going to stay silent and watch how my
> precious toys are broken and are taken away.

And that's the problem of perception. coreboot is not a toy.

> Your logical separation (regardless of how we call it) - it WILL
> negatively affect a support for alive boards! I cannot stress it enough:

I think you're stressing quite well.

> Your recent proposals (in case if they would be accepted) will
> negatively affect many boards, including the alive boards - such as
> G505S, a coreboot LTS candidate. If you were expecting the praises of
> approval and kind words for such intentions - well, I am sorry to bust
> your expectations

Every developer expects some god damn respect for their work and
involvement, especially from users with little to no contributions.

[mrnuke at nukepro coreboot]$ git log |grep quickcracktime -c
0

Alex