I get why he didn’t too though. IMO this seems like a strategic decision to try and let it sort itself out first. Maybe I’m misjudging the situation, and I’m sure others here know Linus better,...
I get why he didn’t too though. IMO this seems like a strategic decision to try and let it sort itself out first. Maybe I’m misjudging the situation, and I’m sure others here know Linus better, but sometimes letting drama percolate for a while helps calm (or exhaust) things and allows for other ideas and leaders emerge/step up.
It can let everybody have a voice before stepping in and help you see who thinks what and how people react.
I’ve worked for some good leadership that operated this way on occasion and while the drama might be taxing, once the head honcho speaks on it - that’s it. All communication on the matter stops, but the heat potentially hasn’t dissipated.
Actually yeah… who’s next in the line of succession to run the project? I know Linus has handed off projects just fine in the past, like git, but I wonder if something more monumental like Linux...
Actually yeah… who’s next in the line of succession to run the project? I know Linus has handed off projects just fine in the past, like git, but I wonder if something more monumental like Linux would be harder.
Greg Kroah-Hartman The problem is he's about the same age, so it answers the "Linus gets hit by a bus suddenly one day" problem, but not really the "what do we when linus is old enough to want to...
Greg Kroah-Hartman
The problem is he's about the same age, so it answers the "Linus gets hit by a bus suddenly one day" problem, but not really the "what do we when linus is old enough to want to retire?" problem.
I can see in his view that "this is obvious stuff" so he simply had good faith the maintainers would sort this out. C-first, Rust bindings need to adapt around it. But if you don't want to work...
I can see in his view that "this is obvious stuff" so he simply had good faith the maintainers would sort this out. C-first, Rust bindings need to adapt around it. But if you don't want to work with the rust bindings you don't get a say on how the rust bindings implement (outside of your indirect control in maintaining the C interfaces).
And then the most obvious: no, you cannot stop people from using your code (in this context, to help implement the rust bindings). That's just a core principle of FOSS. I'm sure that's when he realized he needed to intervene.
Disclaimer: I have been vaguely following this situation, but do not fully understand all the specifics and therefore struggled to concisely articulate my thoughts on this. Linus seems a lot more...
Disclaimer: I have been vaguely following this situation, but do not fully understand all the specifics and therefore struggled to concisely articulate my thoughts on this.
Linus seems a lot more annoyed/aggressive in this email than I was expecting, but I have read very little of Linus' writing. I wonder if the forceful nature of this is partially due to Linus seeing some discussions occur after Marcan stepping down that Linux is only being handled by companies because individuals do not want to deal with the toxic culture without the financial incentive. I also recall Linus being quite supportive towards Asahi, so maybe that also influenced how forceful his writing comes across. However, I do feel his tone is warranted considering all the arguments he laid out, I guess it just kind of surprised me.
I think Linus' argument is good that it is problematic that maintainers who take the stance that they don't want to be involved with Rust for Linux to veto Rust rather than just themselves not maintain it. I think Linus' argument that if they change something in C, then it is up to the Rust developers to fix the broken parts of Rust code valid, whereas the C maintainers felt they would be responsible for it (or that was their public argument because they just didn't want Rust near their C code).
Linus recently had a private conversation with Hellwig after the drama started. Hellwig didn't divulge the entire contents of that conversation, but it strikes me as unlikely that Linus wouldn't...
Linus recently had a private conversation with Hellwig after the drama started. Hellwig didn't divulge the entire contents of that conversation, but it strikes me as unlikely that Linus wouldn't have made it clear already to Hellwig he was out of line while simultaneously telling him that the Rust code would be merged over Hellwig's objections.
So Linus being frustrated here makes sense. I think you've probably already heard of Linus's rants, but if not, you're in for a treat. His language before taking time to self-reflect (2018?) sure was something!
I think this says a lot: ... I saw a discussion on Reddit once about Linux and politics. A user basically stated that Linux IS politics. It sounds like from Linus's viewpoint, Hellwig has lost...
Linus seems a lot more annoyed/aggressive in this email than I was expecting
I think this says a lot:
So let me be very clear: if you as a maintainer feel that you control
who or what can use your code, YOU ARE WRONG.
...
So this email is not about some "Rust policy". This email is about a
much bigger issue: as a maintainer you are in charge of your code,
sure - but you are not in charge of who uses the end result and how.
I saw a discussion on Reddit once about Linux and politics. A user basically stated that Linux IS politics. It sounds like from Linus's viewpoint, Hellwig has lost sight of at least one of the things Linux and the FOSS movement stands for. Also his actions/words directly led to someone leaving the project. I do agree with the sentiment that maybe Marcan was a significant part of the problem, but still.
I'd imagine that in any healthy organization, sometimes you need to maintain order within your management structure (or whatever the equivalent is). Like in a political party, if you have a Member of Parliment or Congressman (whatever) that's saying stupid shit, they will absolutely get a tongue lashing. Its usually just not as public as this.
I don't think Linus could have been less forceful yet still delivered his points.
See our previous discussion on this for more background: https://tildes.net/~comp/1lz8/resigning_as_asahi_linux_project_lead (Admins I'm not sure if it's more appropriate to just post this as a...
Is there some missing contextual follow up? Without reference, this just seems like a fairly petty statement to be safely dismissed.
On Wed, 19 Feb 2025 at 22:42, Christoph Hellwig hch@infradead.org wrote:
The document claims no subsystem is forced to take Rust. That's proven
to be wrong by Linus. And while you might not have known that when
writing the document, you absolutely did when posting it to the list
Is there some missing contextual follow up? Without reference, this just seems like a fairly petty statement to be safely dismissed.
Rust bindings to DMA are being merged against Hellwig's wishes. These are part of the Rust subsystem, and Hellwig is under no obligiation to maintain compatibility with them. Hellwig still views...
Rust bindings to DMA are being merged against Hellwig's wishes. These are part of the Rust subsystem, and Hellwig is under no obligiation to maintain compatibility with them. Hellwig still views this as "DMA" being forced to accept Rust, while Linus and the R4L devs see the DMA Rust bindings as just another user of DMA, not part of the DMA subsystem.
I guess the issue is that the ‘user’ in this case is another kernel dev who are, in this case solely responsible for maintaining the rust bindings. I’m not a kernel dev but from what I know if you...
I guess the issue is that the ‘user’ in this case is another kernel dev who are, in this case solely responsible for maintaining the rust bindings.
I’m not a kernel dev but from what I know if you break a driver or something with your patch you are responsible to fix it. This kind of goes against that where you can submit a merge request that breaks compilation but someone else has to clean it up for you.
There’s nothing wrong with that approach but to my knowledge it’s different than how things worked before.
I believe this is still broadly true, but while R4L is still in the setup phase, the expectations have been relaxed somewhat. It's okay to temporarily break the Rust build (it would not be okay to...
I believe this is still broadly true, but while R4L is still in the setup phase, the expectations have been relaxed somewhat. It's okay to temporarily break the Rust build (it would not be okay to break a C consumer in this way), and if you do break the Rust build, you don't have to be the one to fix that - there is a floating R4L group of maintainers who will help out and fix things like that.
In the medium/long term, I believe the goal is to have a maintainer for every subsystem who is willing and able to keep the Rust bindings in sync. Right now, a number of subsystems are happy maintaining their own bindings, but in Hellwig's case, he has clearly stated that he doesn't want to maintain the Rust bindings, nor does he want to accept new maintainers to do that. For now this is fine - the Rust bindings can still be merged, and someone else will have to keep them up-to-date. In the medium term I suspect this topic will get revisited, particularly if Hellwig remains so opposed to Rust and continues trying to exert his power to block it.
Also do keep in mind that Linux release cycle starts with the phase of merging changes, then follows with multiple rounds of fixing whatever broke or someone overlooked. So it's not like the Rust...
Also do keep in mind that Linux release cycle starts with the phase of merging changes, then follows with multiple rounds of fixing whatever broke or someone overlooked.
So it's not like the Rust code is released broken.
It's a shame he hasn't said as much earlier. Could have avoided a lot of drama.
I get why he didn’t too though. IMO this seems like a strategic decision to try and let it sort itself out first. Maybe I’m misjudging the situation, and I’m sure others here know Linus better, but sometimes letting drama percolate for a while helps calm (or exhaust) things and allows for other ideas and leaders emerge/step up.
It can let everybody have a voice before stepping in and help you see who thinks what and how people react.
I’ve worked for some good leadership that operated this way on occasion and while the drama might be taxing, once the head honcho speaks on it - that’s it. All communication on the matter stops, but the heat potentially hasn’t dissipated.
He's also not going to be here forever and doesn't want Linux to be completely dependent on him stepping in to break up these fights.
I choose to believe Linus Torvalds will outlive us all out of sheer willpower to keep the Kernel project intact.
Hopefully. :-)
Actually yeah… who’s next in the line of succession to run the project? I know Linus has handed off projects just fine in the past, like git, but I wonder if something more monumental like Linux would be harder.
Greg Kroah-Hartman
The problem is he's about the same age, so it answers the "Linus gets hit by a bus suddenly one day" problem, but not really the "what do we when linus is old enough to want to retire?" problem.
I can see in his view that "this is obvious stuff" so he simply had good faith the maintainers would sort this out. C-first, Rust bindings need to adapt around it. But if you don't want to work with the rust bindings you don't get a say on how the rust bindings implement (outside of your indirect control in maintaining the C interfaces).
And then the most obvious: no, you cannot stop people from using your code (in this context, to help implement the rust bindings). That's just a core principle of FOSS. I'm sure that's when he realized he needed to intervene.
Disclaimer: I have been vaguely following this situation, but do not fully understand all the specifics and therefore struggled to concisely articulate my thoughts on this.
Linus seems a lot more annoyed/aggressive in this email than I was expecting, but I have read very little of Linus' writing. I wonder if the forceful nature of this is partially due to Linus seeing some discussions occur after Marcan stepping down that Linux is only being handled by companies because individuals do not want to deal with the toxic culture without the financial incentive. I also recall Linus being quite supportive towards Asahi, so maybe that also influenced how forceful his writing comes across. However, I do feel his tone is warranted considering all the arguments he laid out, I guess it just kind of surprised me.
I think Linus' argument is good that it is problematic that maintainers who take the stance that they don't want to be involved with Rust for Linux to veto Rust rather than just themselves not maintain it. I think Linus' argument that if they change something in C, then it is up to the Rust developers to fix the broken parts of Rust code valid, whereas the C maintainers felt they would be responsible for it (or that was their public argument because they just didn't want Rust near their C code).
Man, working in tech so long must have affected my judgement more than I thought. His email struck me as measured, polite, and clear.
Linus recently had a private conversation with Hellwig after the drama started. Hellwig didn't divulge the entire contents of that conversation, but it strikes me as unlikely that Linus wouldn't have made it clear already to Hellwig he was out of line while simultaneously telling him that the Rust code would be merged over Hellwig's objections.
So Linus being frustrated here makes sense. I think you've probably already heard of Linus's rants, but if not, you're in for a treat. His language before taking time to self-reflect (2018?) sure was something!
"Mauro, SHUT THE FUCK UP!"
I think this says a lot:
...
I saw a discussion on Reddit once about Linux and politics. A user basically stated that Linux IS politics. It sounds like from Linus's viewpoint, Hellwig has lost sight of at least one of the things Linux and the FOSS movement stands for. Also his actions/words directly led to someone leaving the project. I do agree with the sentiment that maybe Marcan was a significant part of the problem, but still.
I'd imagine that in any healthy organization, sometimes you need to maintain order within your management structure (or whatever the equivalent is). Like in a political party, if you have a Member of Parliment or Congressman (whatever) that's saying stupid shit, they will absolutely get a tongue lashing. Its usually just not as public as this.
I don't think Linus could have been less forceful yet still delivered his points.
See our previous discussion on this for more background: https://tildes.net/~comp/1lz8/resigning_as_asahi_linux_project_lead
(Admins I'm not sure if it's more appropriate to just post this as a comment there or make a new topic, so clarification would be appreciated)
The discussions around Rust for Linux are so much larger than marcan's involvement, so I think it's worth having a separate topic for this.
Is there some missing contextual follow up? Without reference, this just seems like a fairly petty statement to be safely dismissed.
Rust bindings to DMA are being merged against Hellwig's wishes. These are part of the Rust subsystem, and Hellwig is under no obligiation to maintain compatibility with them. Hellwig still views this as "DMA" being forced to accept Rust, while Linus and the R4L devs see the DMA Rust bindings as just another user of DMA, not part of the DMA subsystem.
Thank you kindly for the context, these chains can be hard to follow without being more involved.
I guess the issue is that the ‘user’ in this case is another kernel dev who are, in this case solely responsible for maintaining the rust bindings.
I’m not a kernel dev but from what I know if you break a driver or something with your patch you are responsible to fix it. This kind of goes against that where you can submit a merge request that breaks compilation but someone else has to clean it up for you.
There’s nothing wrong with that approach but to my knowledge it’s different than how things worked before.
I believe this is still broadly true, but while R4L is still in the setup phase, the expectations have been relaxed somewhat. It's okay to temporarily break the Rust build (it would not be okay to break a C consumer in this way), and if you do break the Rust build, you don't have to be the one to fix that - there is a floating R4L group of maintainers who will help out and fix things like that.
In the medium/long term, I believe the goal is to have a maintainer for every subsystem who is willing and able to keep the Rust bindings in sync. Right now, a number of subsystems are happy maintaining their own bindings, but in Hellwig's case, he has clearly stated that he doesn't want to maintain the Rust bindings, nor does he want to accept new maintainers to do that. For now this is fine - the Rust bindings can still be merged, and someone else will have to keep them up-to-date. In the medium term I suspect this topic will get revisited, particularly if Hellwig remains so opposed to Rust and continues trying to exert his power to block it.
Also do keep in mind that Linux release cycle starts with the phase of merging changes, then follows with multiple rounds of fixing whatever broke or someone overlooked.
So it's not like the Rust code is released broken.