5 ms·

The DMA infrastructure is core to drivers. Saying no to having a wrapper to use it in rust means every rust driver needs to reimplement it, which creates more w

by wbl 2y ago

The DMA infrastructure is core to drivers. Saying no to having a wrapper to use it in rust means every rust driver needs to reimplement it, which creates more work and issues.

arp242 2y ago
I am aware. Doesn't mean it's not an option, or even a bad idea. Or maybe there is a third option; I don't know. By the way: I don't agree with Hellwig, insofar I can judge things, I'm just saying his opinion is valid, and that "Linus agreed on Rust, so therefore we can merge this patch" is not really a valid argument.
- Tuna-Fish 2y ago
  It's just really, really dumb to both a) have rust drivers in the kernel and b) not merge this patch. It's just obviously stupid. If you start with the assumption of a), there are no valid technical challenges to merging it. It's just better for everyone. Before Hellwig put his foot down as "not merging because rust sucks", he made a series of technical arguments against the patch, which were all transparently bullshit. It was those arguments that really raised such a furor, instead of all the other ways some C devs have disdained rust in the kernel in the past, because they were obviously made in bad faith. And when he was called out for them, he just went full "no rust in kernel".
  - arp242 2y ago
    > Before Hellwig put his foot down as "not merging because rust sucks" He didn't say this at all. He explicitly and repeatedly said he has no problems with Rust as a language. And you can't just assert "there are no valid technical reasons". Just because you don't agree with the objections, or even think they're dumb, doesn't mean you can just dismiss them and start ascribing bad faith motives.
    - sdwr 2y ago
      It's pretty clear that the battle over rust is a power struggle. You can't take any of it at face value
      - arp242 2y ago
        So what are you saying? That it's okay to just invent quotes and spread misinformation about people, based on what you suspect their true thoughts might be?
        Ar-Curunir 2y ago
        You’re the one doing that by pretending that Hellwig is acting contrary to his own statements.
        tw1984 2y ago
        wait a second, I thought you are the one labelling people as "toxic" when completely ignoring the context that really matters.
    - cyberax 2y ago
      He literally called "Rust" a "cancer".
      - arp242 2y ago
        He literally did not. He literally explicitly said right there in the (in)famous "cancer" message that it didn't refer to Rust as a language.[1] "And I also do not want another maintainer. If you want to make Linux impossible to maintain due to a cross-language codebase do that in your driver so that you have to do it instead of spreading this cancer to core subsystems. (where this cancer explicitly is a cross-language codebase and not rust itself, just to escape the flameware brigade)." Stop spreading this kind of misinformation. And no, I don't think he came off very well here, but please, give it a good faith reading.[2] [1]: https://lwn.net/ml/all/20250128092334.GA28548@lst.de/ https://lwn.net/ml/all/20250128092334.GA28548@lst.de/ [2]: https://news.ycombinator.com/item?id=42977720 https://news.ycombinator.com/item?id=42977720
ww520 2y ago
Just curious. Why can't the wrapper be an independent library outside of the DMA infrastructure? It can still be used by all rust drivers. I think Hellwig is against moving the wrapper into the DMA project that he's forced to maintain it.
- loeg 2y ago
  No, Hellwig also Nack'd the patch that added the wrapper to an independent library outside of the core C DMA directory.
  - ww520 2y ago
    Why would he do that? He has no control over an independent library outside of C DMA? I thought he said he's fine with any drivers written in Rust. Just that the maintenance of such including any wrapper cannot fall into C DMA's lap.
    - loeg 2y ago
      > Why would he do that? It's a great question. I mean, my read of it is he hates the idea of Rust4Linux and is using his position to obstruct. > He has no control over an independent library outside of C DMA? Apparently not. > Just that the maintenance of such including any wrapper cannot fall into C DMA's lap. The patch he rejected did not add any code to C DMA, nor C DMA's directory (kernel/dma). Just: rust/bindings/bindings_helper.h | 1 + rust/kernel/dma.rs | 271 ++++++++++++++++++++++++++++++++ rust/kernel/lib.rs | 1 + (Nor does any Rust4Linux code add any maintenance burden to C -- C maintainers are allowed to break Rust code at will.)
      - saghm 2y ago
        This isn't even just your read of it; as other people in this thread have pointed out, he's pretty explicitly stated it at this point.
    - dralley 2y ago

by wbl 2y ago

The DMA infrastructure is core to drivers. Saying no to having a wrapper to use it in rust means every rust driver needs to reimplement it, which creates more work and issues.

arp242 2y ago

I am aware. Doesn't mean it's not an option, or even a bad idea. Or maybe there is a third option; I don't know. By the way: I don't agree with Hellwig, insofar I can judge things, I'm just saying his opinion is valid, and that "Linus agreed on Rust, so therefore we can merge this patch" is not really a valid argument.

Tuna-Fish 2y ago
It's just really, really dumb to both a) have rust drivers in the kernel and b) not merge this patch. It's just obviously stupid. If you start with the assumption of a), there are no valid technical challenges to merging it. It's just better for everyone. Before Hellwig put his foot down as "not merging because rust sucks", he made a series of technical arguments against the patch, which were all transparently bullshit. It was those arguments that really raised such a furor, instead of all the other ways some C devs have disdained rust in the kernel in the past, because they were obviously made in bad faith. And when he was called out for them, he just went full "no rust in kernel".
- arp242 2y ago
  > Before Hellwig put his foot down as "not merging because rust sucks" He didn't say this at all. He explicitly and repeatedly said he has no problems with Rust as a language. And you can't just assert "there are no valid technical reasons". Just because you don't agree with the objections, or even think they're dumb, doesn't mean you can just dismiss them and start ascribing bad faith motives.
  - sdwr 2y ago
    It's pretty clear that the battle over rust is a power struggle. You can't take any of it at face value
    - arp242 2y ago
      So what are you saying? That it's okay to just invent quotes and spread misinformation about people, based on what you suspect their true thoughts might be?
      - Ar-Curunir 2y ago
        You’re the one doing that by pretending that Hellwig is acting contrary to his own statements.
      - tw1984 2y ago
        wait a second, I thought you are the one labelling people as "toxic" when completely ignoring the context that really matters.
  - cyberax 2y ago
    He literally called "Rust" a "cancer".
    - arp242 2y ago
      He literally did not. He literally explicitly said right there in the (in)famous "cancer" message that it didn't refer to Rust as a language.[1] "And I also do not want another maintainer. If you want to make Linux impossible to maintain due to a cross-language codebase do that in your driver so that you have to do it instead of spreading this cancer to core subsystems. (where this cancer explicitly is a cross-language codebase and not rust itself, just to escape the flameware brigade)." Stop spreading this kind of misinformation. And no, I don't think he came off very well here, but please, give it a good faith reading.[2] [1]: https://lwn.net/ml/all/20250128092334.GA28548@lst.de/ https://lwn.net/ml/all/20250128092334.GA28548@lst.de/ [2]: https://news.ycombinator.com/item?id=42977720 https://news.ycombinator.com/item?id=42977720

ww520 2y ago

Just curious. Why can't the wrapper be an independent library outside of the DMA infrastructure? It can still be used by all rust drivers. I think Hellwig is against moving the wrapper into the DMA project that he's forced to maintain it.

loeg 2y ago
No, Hellwig also Nack'd the patch that added the wrapper to an independent library outside of the core C DMA directory.
- ww520 2y ago
  Why would he do that? He has no control over an independent library outside of C DMA? I thought he said he's fine with any drivers written in Rust. Just that the maintenance of such including any wrapper cannot fall into C DMA's lap.
  - loeg 2y ago
    > Why would he do that? It's a great question. I mean, my read of it is he hates the idea of Rust4Linux and is using his position to obstruct. > He has no control over an independent library outside of C DMA? Apparently not. > Just that the maintenance of such including any wrapper cannot fall into C DMA's lap. The patch he rejected did not add any code to C DMA, nor C DMA's directory (kernel/dma). Just: rust/bindings/bindings_helper.h | 1 + rust/kernel/dma.rs | 271 ++++++++++++++++++++++++++++++++ rust/kernel/lib.rs | 1 + (Nor does any Rust4Linux code add any maintenance burden to C -- C maintainers are allowed to break Rust code at will.)
    - saghm 2y ago
      This isn't even just your read of it; as other people in this thread have pointed out, he's pretty explicitly stated it at this point.
  - dralley 2y ago