Libogc (Wii homebrew library) discovered to contain code stolen from RTEMS

Copyright laws (in sane countries) have (varying amounts of) exceptions for reverse engineering pieces that are required for compatibility/interoperability.

Whether this applies to the Nintendo SDK… no clue, ask your lawyer ;). (i.e.: was there an alternative option to using RE'd pieces of the Nintendo SDK?)

It makes sense from a perspective/perception of: with the Nintendo SDK, [if] there wasn't really a choice or an alternative. With the RTEMS code there was.

> On the other hand, the Wii's no longer commercially relevant so I doubt Nintendo will do anything about it either to him or to the DevKitPro project.

They've been going after ROM sites. Going after libogc or Marcan would fit their recent MO.

I think it is really, really unwise to put libogc on blast like this. It draws Nintendo's attention, which is never good. It would be better to reach out privately, and tell them you'll publically call them out unless they add missing credit and/or remove the offending code. Which for all I know, might have been retired already.

> AFAIK it's software RE work, and nothing done in the console hacking scenes is truly cleanroom at all

There's a wide gradient of how much effort people put into reverse engineering consoles in a legal way vs. just copying code straight from their decompiler and slapping an open source license on it. libogc is very much on the "didn't even try" side of that gradient, it's been known since pretty much forever, and even their documentation is straight up copied from Nintendo's SDKs for part of their libraries.

What's new here is discovering that even the parts people thought were developed "fresh" and not just straight-up asm2c'd from Nintendo are actually stolen from other open source projects in a way that tries to conceal the origin of the code.

Whether you'll find that "more morally reprehensible" or not will largely depend on your personal morals, but clearly for some people that seems to be the case...

> To me this looks like a bad attempt at exposing dirty laundry in bad faith, which is not too surprising coming from him.

It seems odd that you would complain about the messenger, here, since it seems you don't actually dispute the message.

> Commit 3ba50ec which Marcan is complaining about was pushed in 2008 and didn't just delete attrib specifically, but all (?) VCS comments indiscriminately.

So it's OK that they did something wrong because they did everything wrong?

> there's no point to it either as Nintendo can knock&talk and/or send strongly worded letters when they please, legality be damned

There's very much a point to it (when you're building an emulator or tooling, rather than e.g. romhacks where it's unavoidable), because if you carefully stay entirely above board, you can burn those strongly worded letters, make DMCA counter-notices, and otherwise rely on the fact that both emulation and reverse-engineering are legal.

Hector just doesn't have enough drama going on in his life after resigning (twice!) from the LKML, eh?

Nonetheless, I respect him for calling out FOSS license violations, even if he's done it in the most drama-pilled way.

> 1. Commit 3ba50ec which Marcan is complaining about was pushed in 2008 and didn't just delete attrib specifically, but all (?) VCS comments indiscriminately. The file was barely touched since

FWIW -- I think this is the commit being talked about: https://github.com/m87h/libogc/commit/3ba50ecd4134ef37a0f18f...

Looks honestly totally innocent ...

>I don't know much about the Wii scene

It shows. It's an open secret to everyone in the Wii scene that libogc is based on proprietary Nintendo code.

>but what I do know is that libctru (3ds)/libnx (Switch) don't have that drama nor made the mistakes made in libogc

Because WinterMute is not behind them.

which Marcan is complaining about

"That prick again?" Not surprised at all. He's been trying to stir shit up for a long time, and best ignored as a troll.

I suspect that it was neither for clout nor circumvention, but ignorance and people doubling down on that ignorance. If you are not specifically bathed in the norms of the FOSS community, GPL is kind of an unintuitive concept. It's a copyright license that forces you to disclaim most of the benefits of copyright protection. If you're coming from a piracy or game modding scene, where copyright is a thing you wipe your ass with, even the bare minimum of GPL compliance is going to seem like a waste of time at best and someone else trying to butt in on your project at worst.

Think about how many pirates do piracy because they think copyright is unethical, versus how many of them are data hoarders, or just want shit for free, or are reselling shady IPTV boxes on eBay. The former two groups are FOSS-adjacent, but the latter two do not care. Then keep in mind how basically any free shit tends to be almost immediately abused by children with an Internet connection and no access to payment rails.

Homebrew scenes seem like a candidate for doing things "the right way", but culturally they're a lot closer to piracy scenes than anyone wants to admit, at least in front of a court.

Note the mention that libogc also copies code from the official Nintendo SDK, which is proprietary.

I would guess one of three cases:

- They didn't want to respect the GPL, because they thought their library would be less popular if it were GPLed. (Many homebrew projects don't want to be fully Open Source because they want to hold back some special sauce, either to slow down efforts by the console vendor to stop them, or to differentiate themselves from other homebrew projects for clout. So someone building a foundational library for homebrew on a platform might want to, legitimately or otherwise, avoid presenting themselves as GPLed.)

- They didn't want to respect the GPL because they couldn't, because they were also pulling in proprietary code they weren't supposed to be using anyway.

- They didn't care because they were already ripping off the Nintendo SDK so why not rip off an Open Source project too. For instance, they just pointedly didn't care about copyright at all, which is a very different position than just not caring about code being proprietary.

(I can respect the position of "we're ignoring the copyright on this old game, so that we can do some awesome modding/romhacking", which is very different than ignoring Open Source licenses and failing to even give credit. I don't see the former as hypocrisy; it's just "we should be able to hack on anything". Console game modders / romhackers / etc tend to have a huge amount of respect for the original game and its authors, and give due credit, even if they're technically violating copyright.)

What was the nature of the stolen(infringed really) code? Because a naive first glance show that they were distributing source code from a project that requires that you distribute source code.... shrugs, what's the problem here?

So was it removing license comments from the files?

[dead]

Being active doesn't matter, the copyright holders still hold the copyright.

How much they profit off the stolen portion is also questionable, and open source licenses weren't meant to extort money but to grant us rights to the code. What they should do is add attributions and fix their licensing (libogc needs to be GPLv2), or remove the code. Willingly, yesterday.

RTEMS is under active development and is running around the solar system right now :)

I'd lawyer up and throw the book at them.

Litigation is expensive.

Yeah, hobbies can be expensive and sure litigation could be someone’s hobby…nothing wrong with that and maybe worth having lawyers on retainer if that’s your jam.

But in this case, there’s probably not a business case…and being a civil matter, there’s no book-‘em Danoh book to throw. Just normal squeezing blood from turnips…which again, might be someone’s hobby at least in theory.

Note that reverse engineering does not have to be clean room, poking at hardware without ever seeing any software. In many places, poking at the proprietary software, decompiling it, tracing it, and so forth is fine despite what unenforcable EULA's might suggest. What is not okay is taking the binaries or decompiled source verbatim and re-distributing it.

Note that e.g. copyright does not apply to decompiled source code (the original authors did not write the decompiled source, unlikely assets taken verbatim - maybe that's where the arguments you mention stem from - although note that there may be regional regulatory differences). Instead, the things that might cause issues are things like the enforceable parts of the software license, any enforceable patents on the functionality, or enforceable platform license restrictions for applications built based on decompiled source.

> One can easily find a bazillion of "github repos" that distribute what is evidently directly decompiled game code with minimal cleanup. Bonus points if they also claim it is OK as long as the game art is not distributed, which in addition to being wrong is disrespectful to developers as a whole.

I'm sorry, this is supposed to be a bad thing?

> How much "reverse engineering" these days really is clean room and how much of it is just ripping off proprietary software?

In Nintendo console hacking scenes? None at all, there is no point to it, going through the hassle of doing cleanroom as an individual is wasted effort.

Though, the spectrum between copy-pasting HexRays output verbatim and rewriting things yourself is fairly large.

Just because some people steal software doesn't mean that Nintedo's behaviour isn't also bad. It's not an either-or situation.

> How much "reverse engineering" these days really is clean room and how much of it is just ripping off proprietary software?

I did mine clean room. When I reverse engineered my laptop's features, I intercepted the proprietary software's communications with the hardware, compiled my findings into a whole bunch of notes and then wrote my own free software to do the same thing based on those notes.

> But when the Nintendo copyright czar wakes up, they're the bad guys...

They are always the bad guys. Copyright owners are monopolists. Copyright as a whole should be abolished. I don't care what the so called "pirates" are doing, they are always less morally wrong than eternal copyright monopolists who rob us of our public domain rights and turn perfectly good computers into locked down digital fiefdoms where we are serfs.

20 year old games you grew up with? Give me a break. These companies have all made their fortunes multiple times over. This "intellectual property" should already be in the public domain by all reasonable accounts. God forbid Nintendo be unable to sell you the exact same Mario ROM for the 10th time though. We're all going to be long dead before our culture returns to us. That means it effectively never will.

i dont know of wii homebrew but in the nds space there is BlocksDS which is not part of devkitpro, and in gba space you can get (experimental) toolchains from wf-pacman (wonderful toolchain). sadly, the nintendo homebrew space is ruled by devkitpro, and i spent some time trying to find an alternative to it, but that doesnt seem to exist right now.

just realized you asked for something different, and my comment above is on the wrong subject. either way, i dont know of any big library/toolchain/sdk for the wii not from dkP, so the chances of a loader different from hbc and not based on libogc are small, sadly.

> The OP cites these (https://github.com/devkitPro/libogc/blob/52c525a13fd1762c103... and https://github.com/atgreen/RTEMS/blob/2f200c7e642c214accb7cc...), but that's hardly a smoking gun. The function is trivial, just filling in some struct fields.

Yeah, I'm with Marcan 90% of the time, and in my view Marcan is more likely than not right that that function is derived from the RTEMS function, but in my view there's still reasonable doubt. That is to say, purely based on the evidence linked, I only agree that it's probable that the code is copied and disagree with Marcan's claim that it's "not possible" for the implementation to be non-infringing.

The fact that some of the identifiers are similar raises the biggest suspicions. "__lwp_stack_isenough" vs. "_Stack_Is_enough" is suspicious because I'd probably call that "sufficient_stack_available" or something like that. "LWP_STATES_DORMANT" vs. "STATES_DORMANT" is also suspicious because the normal OS term would be "sleeping". Still, the function logic is different enough that, even with that evidence, one could plausibly claim that only the headers or interface were copied and the implementation was clean-room, which is non-infringing per Oracle v. Google.

In US legal system terms, I'd say that a preponderance of the evidence shows that the code is a derivative work (i.e. it's more likely than not that the code was copied at some point), but that there's still reasonable doubt (i.e. a reasonable person could plausibly believe otherwise).

FWIW, whether you agree with the accusation or not, it isn't anonymous. The commit history makes it obvious that it's marcan (Hector Martin) making the accusations.

Whether it's really worth all of the hooplah or not is going to be up to taste. I think it's pointless to just not explicitly credit RTEMS personally, but I suspect the real point of doing this is probably in large parts just to distance themselves from the reverse engineered libogc code.

I'm not an expert in embedded systems, but I do work on them at a very low level, and I can't be sure 100% of my code would differential from the RTEMS code base any more than libogs's. That doesn't mean they didn't do it - but the concepts behind Real Time Operating Systems in general are well known, and nearly standardized.

For completeness sake; another possibility is that both took from the same source.