Lead Asahi Linux Developer Quits Days After Leaving Kernel Maintainer Role (theregister.com) 63
Hector Martin has resigned as the project lead of Asahi Linux, weeks after stepping down from his role as a Linux kernel maintainer for Apple ARM support. His departure from Asahi follows a contentious exchange with Linus Torvalds over development processes and social media advocacy. After quitting kernel maintenance earlier this month, the conflict escalated when Martin suggested that "shaming on social media" might be necessary to effect change.
Torvalds sharply rejected this approach, stating that "social media brigading just makes me not want to have anything at all to do with your approach" and suggested that Martin himself might be the problem. In his final resignation announcement from Asahi, Martin wrote: "I no longer have any faith left in the kernel development process or community management approach."
The dispute reflects deeper tensions in the Linux kernel community, particularly around the integration of Rust code. It follows the August departure of another key Rust for Linux maintainer, Wedson Almeida Filho from Microsoft. According to Sonatype's research, more than 300,000 open source projects have slowed or halted updates since 2020.
Torvalds sharply rejected this approach, stating that "social media brigading just makes me not want to have anything at all to do with your approach" and suggested that Martin himself might be the problem. In his final resignation announcement from Asahi, Martin wrote: "I no longer have any faith left in the kernel development process or community management approach."
The dispute reflects deeper tensions in the Linux kernel community, particularly around the integration of Rust code. It follows the August departure of another key Rust for Linux maintainer, Wedson Almeida Filho from Microsoft. According to Sonatype's research, more than 300,000 open source projects have slowed or halted updates since 2020.
Martin's post (Score:5, Informative)
Re:Martin's post (Score:5, Insightful)
Book length resignations always come off making the person look like a crying bitch.
A simple, "It's been great but time to move on to other things" would do and avoid the cry baby look.
If ya gotta go, go out like a pro.
Re: Martin's post (Score:3)
He volunteered to be used, now he wants to complain.
Pox on society (Score:2)
Re: (Score:3)
Re: (Score:2)
Or even "It used to be great, but now I hate the shape of your shadow so go EABOD" and move on.
That wall of text looks like a novel, and it's reasonable to assume that a novel is fictional.
Re:Martin's post (Score:4, Insightful)
Yea, I didn't read it but my take on this is exactly that. It seems like this guy had a long term plan of disrupting something... important. And now that Linus put his foot down it won't be possible to execute his long term goals of destroying whatever critical underpinning of Linux kernel development he was paid to destroy, so he's just trying desperately to blow up as much shit on his way out the door as possible.
Re: Martin's post (Score:2)
It sounds more like he's passed off that using social media to cancel people doesn't work anymore. Or at the very least, not as well as it used to.
Re: Martin's post (Score:2)
What do you mean? Social media these days cancels people all the time - usually if they say the wrong things like pro-DEI, or pro-LGBTQ+ or other things they get rapidly torn down. Or if you try to insult King Musk or the President. Or imply there is a Constitution or other such things.
So when do you expect to be cancelled given you've just done all of that on social media?
Re: (Score:2)
Book length resignations always come off making the person look like a crying bitch.
A simple, "It's been great but time to move on to other things" would do and avoid the cry baby look.
If ya gotta go, go out like a pro.
I dunno. I read the resignation letter and while take your point, I think there's value in calling out the perceived problems and frustrations that Martin described.
What he was able to say while on the way out the door is probably far more candid and informative than what he could have gotten away with while he was still part of the project. Also, preempting some of the inevitable rumor-mill bullshit could be considered a service.
Re: (Score:3, Insightful)
With some effort, I read his whole book.
I saw nothing of value that will change anything, inform anyone, make the world a better place, etc.
I've had real jobs I was unfairly fired from. I told my team thanks, best wishes and moved on. He should've done the same and just left quietly. Like a pro. Now his rant is forever on the net flagging him as non-pro. This wasn't even a real job. It was volunteer work. What's he crying about, exactly? Why is he so emotional about a volunteer job no longer being
Re: Martin's post (Score:2, Funny)
"The ranting cry baby thing never looks good."
Like your handle?
Re: (Score:2)
No, like silly bitches who are triggered by my silly handle which I chose specifically and intentionally to trigger and out those silly bitches.
But not you, you're not a silly triggered bitch worried about dumb internet handles, nope, no sirree, not you at all.
Re: (Score:2)
I know what country he is from and it doesn't matter what country anyone else is from if they're not American. You're all the same. The absolute refusal of most of you to give your country when asked is hilarious, though. As if what, I'd dox your entire country?
I get a laugh knowing I'm always there in your every thought. You login first thing when you wake to find my posts, you login in all day looking for me and the last thing you do at night is look for me. We both know it's true. You can't deny it
Re: (Score:2)
Yes and importantly it is always a waste of time. If they were not going to accept your opinion/vision/whatever while you were they, they certainly don't give a hoot now that you're gone.
The louder message is just to wordless take your ball and go home.
Re: (Score:2)
I actually use Perplexity to summarize the post...
Re: (Score:3)
It's not the length that's the issue, it's the content.
Complaining about donations tapering off when you fail to implement the exact features that people expect and want when they're donating to your project reeks of being completely out of touch with your fanbase. Of COURSE the people trying to pay you to make their machines work on Linux want all the features to work. OF COURSE they want working sound, all of their potential battery life and support for external displays... wtf did you expect?
I understand
Re: (Score:2)
When running a big project, there are a LOT of important decisions that need to be made, on a daily basis.
When a big project has two leaders, they will disagree on many of these decisions. It is inevitable.
So, a project must have only one leader. That's the only way to ensure a unity of vision. Maybe the vision isn't perfect, and maybe it isn't the best possible vision, but a unified "ok" vision is still worlds better than a fragmented hodge-podge of rival perfect visions.
So, at some point, the leader si
He can fork it if he doesn't like it (Score:5, Insightful)
That's what open source is supposed to be about.
I also agree with the big L that technical decisions should not be made based on who can raise a bigger social media shit storm.
Re: (Score:1)
If you had read the thread, you would have noticed that the entire problem here is that Linux is inherently unforkable.
There are no "forks". There is just Linus. You can't go on github and click the fork button. All you can do is maintain patches and try and keep them patchable on the current kernel, which requires constant work. And that is why people quit! It's just too much damn work...
Even if you could fork, then you run into issues such as the ones Hector described in hist post: Mesa only accepts patch
Re: He can fork it if he doesn't like it (Score:5, Insightful)
This is of course stupid entitled bullshit.
Forking never entitled you to have your patches accepted by the project you forked from.
Google has its own fork for Android. Stop telling us about how it cannot be done when it's being done.
Re: (Score:1)
You missed the point entirely, that's not at all what's being claimed by anybody. Go read it again more carefully, and while you're at it read the original post too, which precisely uses AOSP as an example.
Re: He can fork it if he doesn't like it (Score:2)
This is hardly unique to Linux. If you ever need to extend the fictionality of a library, you make it a lot easier on yourself if you submit it upstream so that you don't have to patch in your changes into new releases.
Sometimes maintainers are kind of dumb though. I still recall a case where a googler was (still is I think) maintaining a very security related module in golang and he was offered a patch to fix undefined behavior, and the maintainer still hasn't done anything with it. We use this module wher
Re: He can fork it if he doesn't like it (Score:2)
I read it. It is full of falsehoods. No wonder you love it.
Re: (Score:2)
Google has its own fork for Android.
It is worth pointing out that Google has significantly more resources than one developer, and even then it took them years to do something meaningful with the fork of the Linux kernel.
So no, you can't just fork something at will, and pointing to one of the world's largest mega corporations doing it as your primary example sort of helps make the opposite point you're going for.
Re: (Score:2)
Google has done meaningful things with their fork of the kernel from early days, like tampering with schedulers and adding support for devices. They've contributed some locking mechanisms needed for power management to mainline. There was no need for these guys to mainline everything. Google wasn't allowed to mainline some things because the kernel maintainers believed that they wouldn't stick around to maintain them. And lo! This guy is pulling a hissy fit, and won't be around to maintain things.
ANYONE can
Re: (Score:2)
Google has its own fork for Android.
We're talking about a couple of guys on a pet project, not Google. And we're also not talking about something like ZFS which can reasonably be maintained as a separate directory with a couple of changes to mainline here and there and some hooks (which is difficult enough to keep up-to-date).
Their project requires numerous upstream changes which the maintainers are afraid to implement because they'll be stuck with supporting it later on, plus they don't know rust. And if they do fork then a bunch of stuff st
Re: (Score:2)
Okay I'm going to read this dickhead's whole screed if I can stand it. Annnnnd I hate him by the first half of the second paragraph. Boo hoo hoo obvious thing happened. ...now he's surprised someone wants to know the temperature of the most important part of their computer... now he's overpersonal vagueposting. Now he's not going to tell us what his problem is with Rust because it's well-documented, but it's definitely Linus' fault.
Hmm, I found one thing to agree with him on, "the thin blue line" gives me t
Re: (Score:2)
Linux has become too large and too important to fork easily. It's not like forking a run-of-the-mill software project, Linux has thousands of devs.
If you cannot figure out how to work with those people, then you don't have a point mucking up those thousands' of devs work. I don't have an opinion on this C vs Rust thing, but I don't think a programming language is the problem here.
Re: (Score:1)
Re: (Score:3)
I thought you couldn't fork Linux on GitHub because it wasn't on GitHub, then I checked: https://github.com/torvalds/li... [github.com]
I didn't try to fork it, but the fork button was on the screen and not disabled.
So I don't have any idea why he claimed you couldn't fork it. Superficial evidence seems to indicate that you could. (And you've long been able to fork it from the main tree. That's certain, because many people have done so.)
Re: (Score:2)
Well ok, but all patches go through Linus, not via pull requests.
To be fair (Score:3)
These developers are asking for a lot of change that really doesnâ(TM)t need to happen - and historically never has.
Linus may have his faults, but heâ(TM)s not wrong on this or the Rust issue.
Re: (Score:2)
The issue here is the maintainer of Apple Silicon in Linx trying to runaround an LKML rust technical disagreement by social media brigading, i.e., bitching on social media, and trying to rile up a mob to create a shitstorm. That, Linus does not support.
Re: (Score:2)
I have a very different interpretation.
> an LKML rust technical disagreement
The point is that there's no technical disagreement. It's only a philosophical/religious disagreement about whether Rust should belong in the kernel or not. (Go see the original thread, the maintainer flat out says that much.)
My issue with this situation is Linus' inconsistency. Either Linus accepts Rust in the kernel, or he does not. Since he did accept Rust in the kernel, he shouldn't allow maintainers to boycott patches just f
Re: (Score:2)
Linux is where it is today by a steady keel making small adjustments along the way. And when the likes of Microsoft and others get involved those holding the keel on course need to beat off the Microsoft, and other devs, with a large orr so not to pollute and destroy what's put Linux where it is today.
LoB
Re: To be fair (Score:2)
The technical argument is over whether each driver should have its own C bindings if it uses another language, instead of having common bindings for all drivers. I'd think common bindings are a good idea, especially if you want to have type safety and overall more reliable code.
But the C purists argue that rust is too hard and they shouldn't have to touch it if they change their interfaces and they want the rest of the code to compile. Which I think is kind of dumb because that means they already want to ma
Rust is not ready yet! (Score:5, Insightful)
The issue with the Rust kernel people, is the same issue I have with Jr Developers and Jr Engineers. I understand you learned about a cool, shiny new framework / language. I understand that you used that shiny new framework to build a little, tiny, barely functional app. That doesn't mean it's ready to replace Angular in our Web App, or, C in the RTOS.
Frankly, when it comes to Rust, it's not even memory safe, so, basically, you want to inject a language into an existing code base, that has bad tooling, bad support, and doesn't meet its claim to fame? If you want a Rust version of the kernel, clone the kernel and make a Rust version.
Re: (Score:3)
Even Linus thinks Rust is ready to play a role in the Linux kernel. Just not as big a role as Martin thinks, not yet. Martin wants the rest of the kernel devs to take on a large and long-term maintenance obligation for something that hasn't yet proven itself. And when he tried to politic it the rest of the devs (including Linus) got annoyed.
Re: (Score:2)
And maybe the brigading that needs to happen is that all the rust supporters who constantly make noise about rust need to go help Hector do whatever it is he's trying to do and unable to keep up with. As others have mentioned, forking linux is a herculean effort, but it is being done. It may require a large number of people to accomplish successfully.
Or perhaps all they want to do is talk and social media, in which case this effort is not really valuable enough to invest in.
Re: (Score:3)
Re: (Score:2)
I don't know whether rust is ready or not, but I know that lots of developers aren't, and don't intend to learn rust to get ready. To me rust feels like programming with one hand tied behind your back. Even within a local routine only one instance of a pointer at a time can be write ready. This is wildly excessive. If I wanted that kind of environment I'd have been using Erlang or one of the functional languages. (Actually, Erlang would be easier to use.)
Re: (Score:2)
Pretty much (Score:2)
I have my "issues" with C, but C is what it is, and what it is, Rust isn't.
Joel Spolsky back in the day when people were reading his blog posts offered a software engineering guidline of "if it ain't broke, don't fix it because fixing it will make many more bugs."
In the Spolsky-esque traditions, the Linux kernel contains many shims and patches to work within the limitations of C, but if you change it over to something else, even portions of it, you lose all of what those shims and patches are doing.
Re: (Score:2)
Re: (Score:2)
Angular was a cool shiny new framework once.
Re: (Score:2)
Open Source activism is now a thing :o (Score:4, Informative)
Linus Torvalds: “How about you accept the fact that maybe the problem is you.”
Re:Open Source activism is now a thing :o (Score:5, Insightful)
All activists - be they political or technical or whatever (this disagreement seems particularly non-political), have to accept the fact the sometimes NOTHING will work, and you have to be ok with that.
You're not owed the change you seek. You can try to make a change and people may or may not agree with you. If they don't, that doesn't give you license to adopt increasingly more drastic measures until they accept it - that's how terrorists are born.
Sometimes things just don't go your way, and you have to be ok with that.
Re: (Score:2)
Re: (Score:2)
No - there can be mature and immature activists. No different than if I said "All adults have to accept that sometimes you have to forgo luxury expenses to pay bills."
You wouldn't say "Adult is just what we call financially irresponsible people now." would you? The statement that all of a group has to perform an action doesn't mean that many of them aren't already doing so.
Re: Open Source activism is now a thing :o (Score:1)
Often, yeah. Luigi Mangione was pissed off that no doctor would perform the surgery he wanted, and it had nothing to do with his ability to pay, insurance, etc. So he says to himself "I know, I'll kill a bean counter!" (his words) and did so. And then you've got a lot more activists who think that was the right course of action, even though with or without insurance companies, universal health care, or whatever the hell it is they want, the surgery is still not fucking happening in his case, even if he hadn
Safety first? (Score:3)
"having to handle various abusers and stalkers who harassed and attacked me and my family (and continue to do so)."
Safety first. Time to go.
Also, when it quiets down, perhaps consider if you need to change something about yourself to avoid such things.
Yes, I am a victim blamer. I believe that most people are useless and selfish twits. Those people aren't going to follow you around and attack you for coding mistakes. Something else much larger is going on, and those attackers THINK following you around is worth it.
Don't make it worth it. Live quietly.
Disagreements Are Not a Surprise (Score:2)
It shouldn't surprise anyone that there is some dissent and disagreement among a group of high-powered people, particularly when they are not being paid for their work. Open source projects die from this, just as they die from disinterest or from the exhaustion of the maintainers. The real surprise is that Linux has been able to continue and to improve for so many years already. It's a great gift that we all should appreciate, but it really is amazing that it has endured.
Re: (Score:2)
Linux's success isn't really that surprising when you think about who depends on it — everyone. Even Microsoft needs it. (Some might say, especially.) And now that it's gotten to this point, it is probably true that it can survive even the loss of Linus, though I certainly hope that day is far in the future — both for sentimental reasons, and because I think we all benefit from his continued involvement and influence.