https://lore.kernel.org/lkml/20250208204416.GL1130956@mit.edu/ I know we programmers aren't exactly known for tact but if a dumbass like me knows that's beyond the fucking pale to say I would damn...
I'll let you in a secret. The maintainers are not "all-powerfui". We are the "thin blue line" that is trying to keep the code to be maintainable and high quality.
I know we programmers aren't exactly known for tact but if a dumbass like me knows that's beyond the fucking pale to say I would damn well hope an MIT-educated person knows that. Dear god.
Eh, I see that phrase and my initial recall is to the British tv show from the 90s. Im not from America or Europe so hardly exposed to the nuances of why such a phrase is bad form. Ok, maintainers...
Eh, I see that phrase and my initial recall is to the British tv show from the 90s.
Im not from America or Europe so hardly exposed to the nuances of why such a phrase is bad form.
The "thin blue line" is a term that typically refers to the concept of the police as the line between law-and-order and chaos in society.
Ok, maintainers of order make's sense on the surface.
In recent years, the symbol has also been used by the Blue Lives Matter movement in the United States, which aims to show solidarity with the police, and a number of far-right movements in the U.S., particularly after the Unite the Right rally in 2017.
Regardless of my feelings about Rust, it seems like the crux of the issue was that some Linux maintainers were doing their best to sabotage the work of the Rust for Linux team, despite whatever...
Regardless of my feelings about Rust, it seems like the crux of the issue was that some Linux maintainers were doing their best to sabotage the work of the Rust for Linux team, despite whatever earlier reassurances there were.
Perhaps the folks on the Asahi Linux side of things could have handled their communications better in response, but I know I would've been quite cross to know that despite reassurances, I had in essence wasted an enormous amount of my own time and energy on something that never had a chance to be merged - time that I could have spent doing something more productive, and more importantly time that I can't get back.
I hope folks on the Linux side consider this the next time they pontificate about how difficult it is to find maintainers.
I'm not sure how much I'd say this is a problem from the Rust team in general. Marcan wasn't part of this PR, he came in guns blazing partway through. His advice to his colleague was fine but he...
I'm not sure how much I'd say this is a problem from the Rust team in general. Marcan wasn't part of this PR, he came in guns blazing partway through. His advice to his colleague was fine but he also went on the attack a bit unnecessarily. Yes the DMA maintainer was aggressively combative but he also wasn't a necessary approval. They could have just ignored his protest any continued on with the PR work. Linus makes the final call so that's who to ensure is on board.
It's a mess overall because Linus hasn't really made a firm commitment, so people are forced to read his tea leaves. It certainly sucks for the Rust team as they have few assurances that their efforts will be fruitful but there's no other legitimate competition for C replacement in the works for Linux. So it's not far fetched either. This is going to be a long long project and the Rust maintainers need to be ready for it to take time.
I don’t know why but got somewhat annoyed reading the post from the guy and I don’t even know them. Is this the guy that asked for someone to be banned because used the the word “cancer”, after...
I don’t know why but got somewhat annoyed reading the post from the guy and I don’t even know them. Is this the guy that asked for someone to be banned because used the the word “cancer”, after that tried to start “crusade” against said guy in social media, then Linus appeared to say that this approach is stupid, and I agree is stupid, then deleted their social media accounts?
I think that's missing a lot of nuance, but yeah, that sounds like Marcan? IMO: I'm annoyed by both sides. If a coworker called something I'd worked on for years "cancer", then moved to block any...
I think that's missing a lot of nuance, but yeah, that sounds like Marcan?
IMO: I'm annoyed by both sides. If a coworker called something I'd worked on for years "cancer", then moved to block any contributions I was trying to make, I'd want nothing to do with them. For an organization which is literally slowly dying (aging out) due to its community being so toxic that no one -- except for corporations -- will put up with them, facilitating all of this heated bickering is absurd.
You shouldn't be. The reason Rust people were angry is pretty simple. They asked whether they can have a shared module (piece of reusable code) that multiple Rust drivers could use in order to...
I'm annoyed by both sides.
You shouldn't be. The reason Rust people were angry is pretty simple. They asked whether they can have a shared module (piece of reusable code) that multiple Rust drivers could use in order to avoid duplicating the code wrapping the DMA subsystem. Having multiple copies of said wrapper code would be much harder to maintain, much hard to fix bugs in and so on.
And the maintainer said "no". He literally barred them from having a shared piece of code, forcing them to duplicating the necessary code all over the place indefinitely. With no good reason. Just because he can.
Nah, I think I can be annoyed at the Rust leadership folks leading in such a way that they burn out and take a decent chunk of their project’s credibility with them. I am, however, hundreds of...
Nah, I think I can be annoyed at the Rust leadership folks leading in such a way that they burn out and take a decent chunk of their project’s credibility with them. I am, however, hundreds of times more annoyed at the Linux kernel devs for their behaviour and lack of foresight which could ultimately doom the project to corporate ownership within my lifetime.
(edit) To clarify: getting Rust into the kernel in any capacity has never been a technical challenge. The issue has always — clearly — been the maintainers. Not understanding that, and failing to have a plan for dealing with it, is annoying. That said, Rust + an acceptable community is (imo) the best proposal anyone has come forward with for keeping the Linux kernel owned by humans rather than megacorps, so the kernel maintainers should be seeing the writing on the wall and helping see it through.
That article is over sensationalized, I remember reading that message and he did not say that “his work is cancer” is okay being sensitive and attached to your work specifically when you love...
That article is over sensationalized, I remember reading that message and he did not say that “his work is cancer” is okay being sensitive and attached to your work specifically when you love something but everything afterwards and before was already stupid. Both sides on that discussion were stupid to be honest and I kinda understand both sides, just that the way they have been communicating their intentions can be misunderstood because I think they are just use to express their ideas in a different manner.
Basically one side is afraid that they are not going to be able to maintain something because they aren’t experts on the language and if the people working on one side just disappear, that thing is going to be unmaintained and the other side are sure that they going to be there long term and they can handle but need help from the other.
Both sides have good points, but those good points are being hidden by what it looks stubbornness and lack of communications skills
Yup, agreed. I think the kernel maintainers have more to lose, and should be taking efforts like this more seriously (Rust was purpose built to enable more people to contribute to complex,...
Both sides have good points, but those good points are being hidden by what it looks stubbornness and lack of communications skills
Yup, agreed. I think the kernel maintainers have more to lose, and should be taking efforts like this more seriously (Rust was purpose built to enable more people to contribute to complex, non-GC’d codebases), but marcan has made some missteps too.
As perhaps a separate note, it also sounds like the stress of being a public figure in charge of a popular product in the general population was also causing him a lot of stress (eg getting stalked, dealing with overwork), which couldn’t have helped matters.
Yeah, that part that you mentioned probably contributed to marcan reaction. I just hope that he can just take a time off and return to do what he likes. Because I also understand how tiring it can...
Yeah, that part that you mentioned probably contributed to marcan reaction. I just hope that he can just take a time off and return to do what he likes. Because I also understand how tiring it can be being in that position, that not only is there a bunch of people asking something from you and also facing setbacks to do it by something external to you
He did call the project a cancer. It was indirect but very clear. He was saying that any dual language project is a cancer which this one is. He tried to cover it by saying he likes Rust in...
He did call the project a cancer. It was indirect but very clear. He was saying that any dual language project is a cancer which this one is. He tried to cover it by saying he likes Rust in isolation but that he would never support Rust in Linux.
Indeed. From what I can recall, he mentioned that introducing more than one language to a project the new one can be cancer and blah blah. It’s exaggerated I believe, but he got a point, have you...
Indeed. From what I can recall, he mentioned that introducing more than one language to a project the new one can be cancer and blah blah. It’s exaggerated I believe, but he got a point, have you tried to review something that you don’t know anything about?
It’s harder for those maintainers because they are being introduced to a new language that they probably have no interest, they have to check something they don’t understand what it does. Yes, they could learn it, but it’s going to take time, time that they won’t be able to spent doing what they are doing. And it’s something just to learn a language to be able to make a simple application and a different thing is become good enough to maintain a kernel. And yes, know the kernel is a multi language project now, but for the longest time it wasn’t. So this is going to cause some friction, specially this being rust and not a different language.
I would like for them to take a different approach in this situation, and try to understand each other’s position. I don’t care that much about rust, but since it has been accepted into the project everyone should work for this to be an easier transition
I understand that multi language sucks. But the maintainers aren't being asked to understand Rust. The only thing that's going into the kernal is typed interfaces. They have no logic, just...
I understand that multi language sucks. But the maintainers aren't being asked to understand Rust. The only thing that's going into the kernal is typed interfaces. They have no logic, just descriptions. They're also relatively easy to read for non-Rust developers. The C devs only break the build if they change the definitions of their C interfaces in a breaking fashion. It's already an issue that have to resolve since they need to update all instances that use their code when they make these changes, the Rust interfaces are just another consumer.
Overall I think this is a good thing for Linux. It requires the C devs to document their breaking changes more meaningfully. This helps all consumers, not just the Rust team.
Once in the kernel there is no rust side anymore, is going to be just the kernel. And if something breaks, they are not going to be able to help fix it, they would need to wait for people with...
Once in the kernel there is no rust side anymore, is going to be just the kernel. And if something breaks, they are not going to be able to help fix it, they would need to wait for people with knowledge in the language to assist. If someone in C fuck up, then they do have the ability to fix it, I suppose this is other concern that they expressed, don’t remember if on this thread or different one.
I know that is expected that they are going to be long term in the project, but things happen, and as I said I understand both sides but honestly they could have handled this better
Perhaps I am confused, but hellwig is a maintainer for dma and he gave an explicit nack on the patches for dma. Isn't that basically blocking? EDIT: it seems I am confused, the code wasn't in the...
Perhaps I am confused, but hellwig is a maintainer for dma and he gave an explicit nack on the patches for dma. Isn't that basically blocking?
EDIT: it seems I am confused, the code wasn't in the dma subtree. Stupid all round.
Yeah they want to use his module so that's why they added him to the email. It was a courtesy and a request of "hey did we misinterpret anything?". He had no actual authority as to whether the PR...
Yeah they want to use his module so that's why they added him to the email. It was a courtesy and a request of "hey did we misinterpret anything?". He had no actual authority as to whether the PR can be merged.
I haven't followed the project or the developers past the first year, but if Marcan says he's been stalked etc lately I believe him and can't help but feel sorry he's reached a breaking point on a...
I haven't followed the project or the developers past the first year, but if Marcan says he's been stalked etc lately I believe him and can't help but feel sorry he's reached a breaking point on a project he cared so much about.
The other thing I believe is that kernel devs on lkml will never cease to act in bad faith and get away with it by demonstrating that they can become the biggest possible pain to deal with.
https://lore.kernel.org/lkml/20250208204416.GL1130956@mit.edu/
I know we programmers aren't exactly known for tact but if a dumbass like me knows that's beyond the fucking pale to say I would damn well hope an MIT-educated person knows that. Dear god.
Eh, I see that phrase and my initial recall is to the British tv show from the 90s.
Im not from America or Europe so hardly exposed to the nuances of why such a phrase is bad form.
Ok, maintainers of order make's sense on the surface.
Ah, I see.
Tiny bit of additional context, the maintainer that made the comment is American and should be well aware of the comparisons his phrasing carries.
Regardless of my feelings about Rust, it seems like the crux of the issue was that some Linux maintainers were doing their best to sabotage the work of the Rust for Linux team, despite whatever earlier reassurances there were.
Perhaps the folks on the Asahi Linux side of things could have handled their communications better in response, but I know I would've been quite cross to know that despite reassurances, I had in essence wasted an enormous amount of my own time and energy on something that never had a chance to be merged - time that I could have spent doing something more productive, and more importantly time that I can't get back.
I hope folks on the Linux side consider this the next time they pontificate about how difficult it is to find maintainers.
I'm not sure how much I'd say this is a problem from the Rust team in general. Marcan wasn't part of this PR, he came in guns blazing partway through. His advice to his colleague was fine but he also went on the attack a bit unnecessarily. Yes the DMA maintainer was aggressively combative but he also wasn't a necessary approval. They could have just ignored his protest any continued on with the PR work. Linus makes the final call so that's who to ensure is on board.
It's a mess overall because Linus hasn't really made a firm commitment, so people are forced to read his tea leaves. It certainly sucks for the Rust team as they have few assurances that their efforts will be fruitful but there's no other legitimate competition for C replacement in the works for Linux. So it's not far fetched either. This is going to be a long long project and the Rust maintainers need to be ready for it to take time.
I don’t know why but got somewhat annoyed reading the post from the guy and I don’t even know them. Is this the guy that asked for someone to be banned because used the the word “cancer”, after that tried to start “crusade” against said guy in social media, then Linus appeared to say that this approach is stupid, and I agree is stupid, then deleted their social media accounts?
I think that's missing a lot of nuance, but yeah, that sounds like Marcan?
IMO: I'm annoyed by both sides. If a coworker called something I'd worked on for years "cancer", then moved to block any contributions I was trying to make, I'd want nothing to do with them. For an organization which is literally slowly dying (aging out) due to its community being so toxic that no one -- except for corporations -- will put up with them, facilitating all of this heated bickering is absurd.
You shouldn't be. The reason Rust people were angry is pretty simple. They asked whether they can have a shared module (piece of reusable code) that multiple Rust drivers could use in order to avoid duplicating the code wrapping the DMA subsystem. Having multiple copies of said wrapper code would be much harder to maintain, much hard to fix bugs in and so on.
And the maintainer said "no". He literally barred them from having a shared piece of code, forcing them to duplicating the necessary code all over the place indefinitely. With no good reason. Just because he can.
Nah, I think I can be annoyed at the Rust leadership folks leading in such a way that they burn out and take a decent chunk of their project’s credibility with them. I am, however, hundreds of times more annoyed at the Linux kernel devs for their behaviour and lack of foresight which could ultimately doom the project to corporate ownership within my lifetime.
(edit) To clarify: getting Rust into the kernel in any capacity has never been a technical challenge. The issue has always — clearly — been the maintainers. Not understanding that, and failing to have a plan for dealing with it, is annoying. That said, Rust + an acceptable community is (imo) the best proposal anyone has come forward with for keeping the Linux kernel owned by humans rather than megacorps, so the kernel maintainers should be seeing the writing on the wall and helping see it through.
That article is over sensationalized, I remember reading that message and he did not say that “his work is cancer” is okay being sensitive and attached to your work specifically when you love something but everything afterwards and before was already stupid. Both sides on that discussion were stupid to be honest and I kinda understand both sides, just that the way they have been communicating their intentions can be misunderstood because I think they are just use to express their ideas in a different manner.
Basically one side is afraid that they are not going to be able to maintain something because they aren’t experts on the language and if the people working on one side just disappear, that thing is going to be unmaintained and the other side are sure that they going to be there long term and they can handle but need help from the other.
Both sides have good points, but those good points are being hidden by what it looks stubbornness and lack of communications skills
Yup, agreed. I think the kernel maintainers have more to lose, and should be taking efforts like this more seriously (Rust was purpose built to enable more people to contribute to complex, non-GC’d codebases), but marcan has made some missteps too.
As perhaps a separate note, it also sounds like the stress of being a public figure in charge of a popular product in the general population was also causing him a lot of stress (eg getting stalked, dealing with overwork), which couldn’t have helped matters.
The whole situation sucks. Everyone loses, now.
Yeah, that part that you mentioned probably contributed to marcan reaction. I just hope that he can just take a time off and return to do what he likes. Because I also understand how tiring it can be being in that position, that not only is there a bunch of people asking something from you and also facing setbacks to do it by something external to you
He did call the project a cancer. It was indirect but very clear. He was saying that any dual language project is a cancer which this one is. He tried to cover it by saying he likes Rust in isolation but that he would never support Rust in Linux.
Indeed. From what I can recall, he mentioned that introducing more than one language to a project the new one can be cancer and blah blah. It’s exaggerated I believe, but he got a point, have you tried to review something that you don’t know anything about?
It’s harder for those maintainers because they are being introduced to a new language that they probably have no interest, they have to check something they don’t understand what it does. Yes, they could learn it, but it’s going to take time, time that they won’t be able to spent doing what they are doing. And it’s something just to learn a language to be able to make a simple application and a different thing is become good enough to maintain a kernel. And yes, know the kernel is a multi language project now, but for the longest time it wasn’t. So this is going to cause some friction, specially this being rust and not a different language.
I would like for them to take a different approach in this situation, and try to understand each other’s position. I don’t care that much about rust, but since it has been accepted into the project everyone should work for this to be an easier transition
I understand that multi language sucks. But the maintainers aren't being asked to understand Rust. The only thing that's going into the kernal is typed interfaces. They have no logic, just descriptions. They're also relatively easy to read for non-Rust developers. The C devs only break the build if they change the definitions of their C interfaces in a breaking fashion. It's already an issue that have to resolve since they need to update all instances that use their code when they make these changes, the Rust interfaces are just another consumer.
Overall I think this is a good thing for Linux. It requires the C devs to document their breaking changes more meaningfully. This helps all consumers, not just the Rust team.
Once in the kernel there is no rust side anymore, is going to be just the kernel. And if something breaks, they are not going to be able to help fix it, they would need to wait for people with knowledge in the language to assist. If someone in C fuck up, then they do have the ability to fix it, I suppose this is other concern that they expressed, don’t remember if on this thread or different one.
I know that is expected that they are going to be long term in the project, but things happen, and as I said I understand both sides but honestly they could have handled this better
Perhaps I am confused, but hellwig is a maintainer for dma and he gave an explicit nack on the patches for dma. Isn't that basically blocking?
EDIT: it seems I am confused, the code wasn't in the dma subtree. Stupid all round.
Yeah they want to use his module so that's why they added him to the email. It was a courtesy and a request of "hey did we misinterpret anything?". He had no actual authority as to whether the PR can be merged.
I haven't followed the project or the developers past the first year, but if Marcan says he's been stalked etc lately I believe him and can't help but feel sorry he's reached a breaking point on a project he cared so much about.
The other thing I believe is that kernel devs on lkml will never cease to act in bad faith and get away with it by demonstrating that they can become the biggest possible pain to deal with.
Companion post: https://asahilinux.org/2025/02/passing-the-torch/