Asahi Linux Lead Developer Hector Martin Resigns From Linux Kernel (theregister.com) 83
Asahi lead developer Hector Martin, writing in an email: I no longer have any faith left in the kernel development process or community management approach.
Apple/ARM platform development will continue downstream. If I feel like sending some patches upstream in the future myself for whatever subtree I may, or I may not. Anyone who feels like fighting the upstreaming fight themselves is welcome to do so.
The Register points out that the action follows this interaction with Linux Torvalds.
Hector Martin: If shaming on social media does not work, then tell me what does, because I'm out of ideas.
Linus Torvalds: How about you accept the fact that maybe the problem is you. You think you know better. But the current process works. It has problems, but problems are a fact of life. There is no perfect. However, I will say that the social media brigading just makes me not want to have anything at all to do with your approach. Because if we have issues in the kernel development model, then social media sure as hell isn't the solution.
Apple/ARM platform development will continue downstream. If I feel like sending some patches upstream in the future myself for whatever subtree I may, or I may not. Anyone who feels like fighting the upstreaming fight themselves is welcome to do so.
The Register points out that the action follows this interaction with Linux Torvalds.
Hector Martin: If shaming on social media does not work, then tell me what does, because I'm out of ideas.
Linus Torvalds: How about you accept the fact that maybe the problem is you. You think you know better. But the current process works. It has problems, but problems are a fact of life. There is no perfect. However, I will say that the social media brigading just makes me not want to have anything at all to do with your approach. Because if we have issues in the kernel development model, then social media sure as hell isn't the solution.
Re:Five Alarm Fire (Score:5, Informative)
Hector Martin didn't like the technical process for patch submission so threw in "shaming on social media" as an approach... that did not win any friends.
Re:Five Alarm Fire (Score:4, Interesting)
Re:Five Alarm Fire (Score:5, Interesting)
> A good process, though, shouldn't fall apart if you don't have "friends". The whole thing is very childish on both sides.
Hector pushed what he believed was the way forward for Rush, his Apple patches, and Linux as a whole. His positions did not garner much support. Other longtime kernel devs and maintainers disagreed with him. Hector tried to brigade on social media (and not for the first time). Linus himself points out this is counterproductive. Hector ragequites (well kind of -- he's still actively contributing as his anime girl Vtube persona).
If I have to go with Linus, warts and all, who has managed what is one of the most important and influential coding project of all time, or someone ragequitting after being called out for bad behavior, I'll stick with Linus.
Besides, look at it this way. People quit Linux dev all the time. When is the last time one contributor pretend quitting made Slashdot headlines?
Hector is, yet again, playing the social media rage bait game.
Re:Five Alarm Fire (Score:5, Funny)
Regardless of how you feel about Hector, what you say about his company is what you say about society.
Re: Five Alarm Fire (Score:2)
That joke will be 20% less funny if the stupid tariffs ever kick in.
Re: (Score:2)
This is the first Rush joke I have ever seen here. My faith in humanity is restored.
Re: (Score:1)
Linux kernel process has accumulated a rather large portion of toxic maintainers over the years. What's worse is, a significant portion of maintainers are technically incompetent, which makes them even more political and toxic in defending their meal tickets.
Case in point: how the fuck was the legendary Andrew Tridgell hounded out of kernel development? With the complicity and active participation of some otherwise respected journalists, what's more. There's actually a long list of such perverse hatchet job
Re: (Score:2)
Case in point: how the fuck was the legendary Andrew Tridgell hounded out of kernel development? With the complicity and active participation of some otherwise respected journalists, what's more. There's actually a long list of such perverse hatchet jobs on decent developers in favor of the old boys club.
What happened there? I remember, only in the barest terms, the BitKeeper snafu in the mid-00s, and that's what came up in a quick search.
Re: (Score:1)
Linus got vindictive about being forced to do the right thing by Tridge and the LWN editor joined in on hounding Tridge out of the community. For perspective, did you ever wonder where the name "git" came from? Yes, Linus knew he was doing the wrong thing but did it anyway. Obviously atoned pretty well for his sins... but not completely.
Re: Five Alarm Fire (Score:2)
Should the process coddle him when he makes enemies?
Re: (Score:2)
Should your analysis be a bit deeper?
Re: Five Alarm Fire (Score:3)
No. The situation is clear. He's whiny and he went to the court of public opinion and lost.
Re: (Score:1)
I assure you, your view is shallow. Way shallow. Thought better of you?
Re: (Score:2)
Read this [lkml.org].
Re: (Score:2)
That mail describes Free Software working as intended. It's not the job of the kernel team to build speculative new infrastructure on someone else's whim. Dr. Greg has the source code, and he alleges he has a base of users ready to start using what he's creating.
Then the correct way of moving forward is to maintain an out of kernel repo with a patch set, and start using it.
Once it's been in use, shown to be useful, and gains traction, is when it's time to lobby for having it included in the kernel. Not befo
Re:Five Alarm Fire (Score:5, Insightful)
> First they gave the BCachefs dev a "timeout" for not being "sensitive" enough (meaning users are denied bug-fixes; collective punishment)
If it's something users are relying upon, the issue should be that there's only one dev for the entire subproject, not that they were kicked out for being a jackass. The LF should be able to remove devs that are a problem, and if a dev falls under a bus, is arrested for murder, or goes insane, that shouldn't cause problems for Linux users. The issue here is that a project that has only one developer is in the official main kernel source tree, but I don't know how prevalent that issue is.
> then they kicked out the Russians
The "racist rant" is something I can't comment on, I haven't seen it, and experience suggests that it rarely is what it's claimed, but Russia is under heavy sanctions and Linux's legal status would be in jeopardy if they continued to work with Russia. It sucks, but those are the laws. The LF didn't have any choice here.
> then they kicked out the Chinese
Haven't actually seen that, and I can't find any evidence for it online, but if it's been talked about one assumes that if this happened this is also related to the various actions the American and other governments are taking to reduce Chinese influence in technologies after several high profile accusations (right or wrong) that they're spying using technology they've modified and shipped to target countries.
So it's entirely possible at some point, like Russia, the LF may be legally obliged to remove Chinese developers from their team, and may have discussed the issue.
Either way, not a "five alarm" fire.
Re: Five Alarm Fire (Score:5, Insightful)
[citation needed]
Re: Five Alarm Fire (Score:4, Interesting)
Hell, I have some issues with some folks about it too, but anyone whining about "diversity" or "politics" just reminds me of "ethics in video game reviews". Bad faith bullshit that will (a) get you ignored by normal people and (b) ensure you never get laid.
Re: (Score:3)
These people have hyperfocus on whatever the their pundits tell them to focus on. If next year it will be that carrots are bad then all they'll talk about on slashdot is carrots, no matter what tfa is about. Its really that simple. What's funny is that they think they're super independent minded when its so obvious to everyone else how they're total slaves to their thought-masters.
Re: (Score:2)
We all know they claim to be "critical thinkers" and their definition of "critical thinker" seems to be what the rest of us define as "mindless followers".
If someone says they are a "critical thinker" that is a big warning sign,
Re:Five Alarm Fire (Score:5, Insightful)
Can you give any specific examples? Because you're arguing I "spun" something (despite in one case pointing out the criticism was wrong but that the underlying issue was serious - LF allowing kernel features to rely on a single developer) but you haven't given a counter argument. If the LF is a dumpster file, I'd love to see evidence of that.
Right now the best I can point to is:
- Lack of action over the schism on improving the technologies used to develop the kernel. But even then that doesn't impact end users beyond maybe reducing the team's ability to prevent bugs as the kernel grows.
- Lack of resources for managing things like WiFi drivers yet an unwillingness to create stable ABIs or user space interfaces that would drastically reduce the need for first-party resources in the first place.
- The aforementioned issue where BcacheFS was allowed in the kernel despite having only one dev responsible for it.
These are issues, but they hardly suggest a "Five Alarm Fire". I bet both Mac OS X and Windows have far more dysfunction, based upon knowing how corporate America works.
Re: (Score:3)
it is a dumpster fire and your attempt at spin fails. We now have a kernel with politics and social agenda being forced on the developers and users
Show us some proof of this if you can. Actual proof, not your emotional rantings.
Re: (Score:2)
The LF should be able to remove devs that are a problem
You obviously have not the slightest clue how the kernel development organization works. Shame on the equally clueless readers who modded up your clueless screed.
Re: Five Alarm Fire (Score:1)
Re:Five Alarm Fire (Score:5, Interesting)
Putting aside responses on the situations you describe that others are responding to, LF != Linux Kernel.
LF is largely an industry marketing gimmick that tosses money Linus' way to get some credibility, but otherwise uses it to confer some of that "Linux Coolness" to whatever the sponsors want.
The reason for their disinterest in helping those projects is not some 'woke' distraction, it's because the corporate sponsor overlords have other initiatives they want to get the focus, and their marketing interests are not advanced by those projects.
Re:Five Alarm Fire (Score:5, Interesting)
The Russian devs were kicked out because they were associated with sanctioned entities and the entire rest of the world outside of the SinoRusso sphere has sanctioned them. There was no "racist rant", there was an entirely polite reply when everyone was well aware of why, but then chose to make a bunch of rants.
This is in fact a huge Rust bullshit rant. We're unfortunately having to look for another language ourself for embedded development because ultimately the Rust Community has an ACTUAL five alarm fire going on all the time doing stuff like this with Social Media brigading.
Re: (Score:2, Insightful)
your world seems to limit itself to a very tiny portion of the planet composed of the united states and their vassal countries
Re: Five Alarm Fire (Score:2)
The whole world except for Russia and China is tiny?
Re: (Score:1)
Re: (Score:3, Interesting)
"... a schism to an equity kernel and a meritocratic kernel..."
I like how you think these are inherently different. Tells us who you are.
Re:Five Alarm Fire (Score:5, Insightful)
Re: (Score:3)
Sorry, is this when all processes run as root? (Because privilege is bad, I hear?)
Re: (Score:1)
Not telling the truth (Score:5, Informative)
Hector has been a drama queen for much of the time that he's been working on his Linux projects. Oddly, this seems to be a theme that runs with the Rust community across multiple domains. In any case Hector's behavior has been criticized before by kernel devs, and each and every time he goes to social media to try to drum up attacks against the people who disagree with him. This time he looped in Linus and, in return, got LIGHTLY criticized [lkml.org].
Also, the not telling the truth part -- he has continued to contribute under his anime girl Vtube persona "Asahi Lina." So yet again, he's going for drama over substance.
Linus on Hector:
On Thu, 6 Feb 2025 at 01:19, Hector Martin wrote:
>
> If shaming on social media does not work, then tell me what does,
> because I'm out of ideas.
How about you accept the fact that maybe the problem is you.
You think you know better. But the current process works.
It has problems, but problems are a fact of life. There is no perfect.
However, I will say that the social media brigading just makes me not
want to have anything at all to do with your approach.
Because if we have issues in the kernel development model, then social
media sure as hell isn't the solution. The same way it sure as hell
wasn't the solution to politics.
Technical patches and discussions matter. Social media brigading - no
than\k you.
Linus
Re:Not telling the truth (Score:5, Funny)
This time he looped in Linus and, in return, got LIGHTLY criticized.
I'd say that's heavy criticism, and one Linus managed to make without a single swear word.
Re:Not telling the truth (Score:4, Informative)
"LIGHTLY" in the context of Linus! For Linus, that message is gentle treatment with kid gloves! He's mellowed.
Re: (Score:2)
Re: (Score:2)
Re: (Score:1)
Whether the criticism is light or heavy, depends on how closely you associate things like social media, and the expectation that there should never be any problems to deal with, with the developer in question. If those things are a core part of his identity, then he just got roasted hard. If social media is something he was messing around with one afternoon on a whim and he made one complaint once about a problem with the process, then
Re:Not telling the truth (Score:4, Insightful)
When everyone around you is a jerk, sometimes it's not everyone else that are the jerks.
Re: (Score:2)
Linus > Hector QED
Re: (Score:2)
I was under the impression that Linus was not employed to do "Linux" stuff at Transmeta, but other "stuff".
Re: (Score:3)
> Transmeta employed Linus even though Linux was not in any way a primary operating system for that company's products. I was under the impression that Linus was not employed to do "Linux" stuff at Transmeta, but other "stuff".
Correct. Like a lot of companies during that period, hiring "top talent" got them a lot of street cred even while they were in stealth mode (hiring smart people also meant that when the company pivoted (and during that period, many companies pivoted), you could still use the peoples talent). While his role did change, he was initially hired to work on the code morphing code. He took another approach, and got them farther faster.
Re: no shit (Score:2)
You're right, Linux is not what it was.
Linux WAS a hobby project.
Linux NOW underpins the world's computing resources, the Internet, computing research...
If it were the same as it was, it couldn't do that.
Re: (Score:2)
Actually, the Linux that underpins the majority of the world's computers, including effectively all supercomputers, most cell phones, the majority of web servers, doorbells, tvs sets, adapter plugs, etc etc... that Linux is the very same Linux that hobbyists used to install from floppy disks back in the day. Just continuously improved. Don't believe me? The records are right there in every Linux kernel repo, even the very early stuff before there even was Git. Read the change notes and see for yourself.
Re: no shit (Score:2)
Most of the code has changed, so no. Very much no.
Re: (Score:2)
Sorry, no. The code was continuously evolved. And actually, the basic structure never changed at all. Mostly just adding more detail. In _very_ rare occasions, large subsystems were changed. In almost no cases was anything completely deleted. The vast amount of code churn is in the drivers, and core pretty much continues to work the way it did from very early. Yes, we fixed the fucked up memory manager, but even that was a process of continuous evolution.
Re: no shit (Score:2)
Total bullshit. Huge parts of the kernel have been completely rewritten, some of them multiple times. Whole subsystems have been deprecated and replaced.
Re: (Score:3)
Today pretty much all of Linux and the larger ecosystem is being driven by large corporations like IBM, Microsoft, Google etc. -- the Linux Foundation is not a foundation in a traditional sense of an independent eleemosynary organization, rather it's basically a way for these large companies to collude and make Linux into what they want it to be. Those companies don't give a shit about getting Linux to run on Apple's computers and so projects like Apache get the stepchild treatment.
That is weird take on the situation. How about this take: Linux is mature after decades of development and many, many organizations rely on it. Some of them are corporations, and many of them contributed to Linux over the decades. It is not a good idea to change a bunch of things that break other things because the install base is affected.
In this case, Asahi Linux is specifically meant for Apple Silicon which is not supported by Apple. Asahi Linux is not for x86 Apple devices. Yes you are right that the
Re: (Score:2)
Linux used to have a stable branch (even numbered releases) and an unstable branch (odd numbered releases). Now it doesn't. It seems to me that things like bcachefs and the asahi stuff should be in an unstable branch until they are mature enough to go in the mainline kernel, and not having that option means that linux is going to stagnate.
Re: (Score:3)
Professional reply from Linus (Score:5, Informative)
_______________
On Thu, 6 Feb 2025 at 01:19, Hector Martin wrote:
If shaming on social media does not work, then tell me what does,
because I'm out of ideas.
How about you accept the fact that maybe the problem is you.
You think you know better. But the current process works.
It has problems, but problems are a fact of life. There is no perfect.
However, I will say that the social media brigading just makes me not
want to have anything at all to do with your approach.
Because if we have issues in the kernel development model, then social
media sure as hell isn't the solution. The same way it sure as hell
wasn't the solution to politics.
Technical patches and discussions matter. Social media brigading - no
than\k you.
Linus
_______________
I guess that's all that needs to be said right there.
Re: (Score:2)
That is all that needs to be said in response to Hector's whining "I'm tired" rant:
https://lkml.org/lkml/2025/2/6... [lkml.org]
It sounds like his ego needs to take a nap.
Re: (Score:2)
It's not in doubt that Linus is a bit of a tech bully and is not above leveraging his popularity to create what can charitably be called mugging situations. Not the worst abuser out there by a long country mile. But still, no angel either.
Not saying that Linus is wrong in this case, but I am saying... don't make the mistake of joining in a mugging just because Linus is leading it.
Re: (Score:2)
Re: (Score:2)
Right, mobbing is the better technical term. Though when just one geek bully leverages their popularity for a personal attack then mugging fits the description as well. Sadly, a frequent pattern in the kernel development community.
Re: (Score:2)
Re: (Score:2)
In the kernel developer context, this is the act of attacking a developer personally and robbing them of their reputation.
Re: (Score:2)
Hector is the problem. (Score:2)
Is it a Rust attitude? (Score:5, Interesting)
With the two developers walking out recently, it has me thinking enthusiastic Rust developers have a difficult time accepting restrictions in a large project like the Linux kernel. They seem to forget Rust does not have an ISO standard yet to provide a stable point of reference for code and compilation to be included elsewhere.
I wouldn't like to add another language to my code if I can't rely on it doing the same thing at compile time after updates to the language.
Say what you will about C and its shortcomings, but to me it looks like the most stable language available.
Re: (Score:1)
I'm not part of the kernel development world, but I did read the entire thread that sparked this all off.
There was a patch sent up for some rust code.
One of the maintainers decided to use that patch to take a hard stance on the Linux kernel being C only, not a specific dig on Rust.
The person who submitted the patch tried very professionally to find a solution and address any technical objections on the thread.
Some other people also chimed in, very professionally.
There were some strong words (not directed at
Re: (Score:2)
However brigading on social media was not the answer of course.
Re: (Score:2)
I agree that using the word cancer was too much, but I don't think there would have to be a complaint if the current maintainer had nothing to do with the patch. It seems there'd still be an impact from changes to the C API that would affect the Rust side at code level and knowledge. I looked at the patch but have no experience in the kernel code to judge.
Re: (Score:3)
Yes, Slashdot covered it this week - Hellwig calling Rust integration a cancer.
https://linux.slashdot.org/sto... [slashdot.org]
Just for context, this current discussion's summary doesn't mention why Hector might have rage-quit.
Re: (Score:2)
He quit because he couldn't deal rationally with technical arguments against his emotionally loaded position.
The kernel is much better without having emotionally focused persons like these precious snowflakes involved.
See how you can write that in a way that doesn't put your post at 0? (Okay mods, please don't put this at 0 too...)
Re: (Score:2)
Christoph is a great guy in general and a key contributor, but doesn't always get it right. I'm viewing this whole tempest as largely a get off my lawn situation.
Re: (Score:2)
For software projects though, the solution is trivial. Fork the project, and make your own big impact changes in your own tree. Go nuts, redesign the whole thing, rename i
Why must there be all this crying out? (Score:2)
MAY or MAY NOT (Score:1)
Glad he worded it that way. "MAY" is a stupid word. It implies NOTHING. Everything in the world is a MAY and if it's not then it's a MAY NOT.
I may have upset you with this analysis. Maybe I didn't.
Perhaps you don't like slashdot. Perhaps you do.
We may all die tomorrow. We may not.
Comprende, amigos, "MAY" is a useless word. Between MAY, MAY NOT, PERHAPS, POSSIBLY, POSSIBLY NOT it's a way of saying NOTHING while pretending to say something.
Maybe I'm right. Maybe not. Perhaps only time will tell. Per