> Dave Airlie just announced in the Maintainers Summit that the DRM subsystem is only ""about a year away"" from disallowing new drivers written in C and requiring the use of Rust.
wow
jeroenhd 9 hours ago [-]
When the C absolutist maintainers fought for control over the ability to keep Rust out of their ballpark, I didn't expect the reverse to happen.
Still, I think it makes a lot of sense. Completely new GPU drivers are quite rare and the macOS drivers from Asahi are a showcase proving that Rust and GPU drivers work together well. If there's any subcomponent switching to Rust-first for new contributions, it makes sense for it to be the one that had already been proven to be Rust-compatible.
pantalaimon 4 hours ago [-]
I would envision to see some more GPU drivers from Chinese companies like MooreThreads
imcritic 9 hours ago [-]
Asahi project looks barely alive, almost abandoned. I know that their explanation of low activity is that they are being active elsewhere, supposedly pushing all their work upstream, but this has been happening for months and they don't give any reports about their progress, so I'm worried it will all die soon. And given that the project barely brought some Linux compatibility for m1 and m2 hardware and no prospects for bringing similar compatibility for newer generations - I fear it all will be kinda useless in the end.
c0balt 7 hours ago [-]
> this has been happening for months and they don't give any reports about their progress
They inarguably have slowed down, but this should be expected as the project matures. It has also inevitably now faced the time when new generations of contributors are needed as existing ones retire/ move to other projects.
charcircuit 4 hours ago [-]
>as the project matures
How can it be mature if it can't even boot on newer MacBooks. The slowness does not seem to be due to running out of impactful work that needs to be done.
GeekyBear 4 hours ago [-]
The new leadership team blogged last year that their priority would be on upstreaming their existing work.
> Our priority is kernel upstreaming. Our downstream Linux tree contains over 1000 patches required for Apple Silicon that are not yet in upstream Linux. The upstream kernel moves fast, requiring us to constantly rebase our changes on top of upstream while battling merge conflicts and regressions. Janne, Neal, and marcan have rebased our tree for years, but it is laborious with so many patches. Before adding more, we need to reduce our patch stack to remain sustainable long-term...
Where do the M3 and M4 fit in? Until upstreaming and CI progress, the core team cannot prioritize new hardware.
I think the majority of that upstreaming work (that isn't on hold until the kernal is ready for the Rust graphics driver to land) has happened and additional features like DP alt mode for USB C have been demoed.
The next update from the team should land on their blog after 6.19 ships
jeroenhd 7 hours ago [-]
Activity has died down as a result of general Linux kernel developer drama, petty in-fighting, and other factors, but that doesn't change the results they did produce during their most prolific phase so far.
Without proper support from upstream like AMD, Intel, and Qualcomm (to some extent) are doing, Linux will never work as well on Apple's hardware as it does on normal hardware.
kryptiskt 9 hours ago [-]
Wasn't it just a couple of weeks ago that they started supporting M3? That smells like progress to me.
mathfailure 7 hours ago [-]
One can start working on creation of a teleportation device. Doesn't mean we have it.
> The document claims no subsystem is forced to take Rust
monocasa 2 hours ago [-]
Dave Airlie is saying that the subsystem maintainers are themselves choosing to move to Rust.
ceteia 2 hours ago [-]
But what about this statement that Linus wrote:
> That's been made clear pretty much from the very beginning, that nobody is forced to suddenly have to learn a new language, and that people who want to work purely on the C side can very much continue to do so.
If any C developer developed drivers in C previously for the DRM subsystem, they might in the future be forced to learn Rust.
cwillu 6 minutes ago [-]
Nothing is happening “suddenly”, and it remains true that “people who want to work purely on the C side can very much continue to do so”.
rjsw 9 hours ago [-]
Means that other platforms need to allow Rust in the kernel too in order to use future drivers.
8 hours ago [-]
saidinesh5 7 hours ago [-]
What do you mean other platforms?
Also they can just expose c bindings to these rust libraries no?
rjsw 7 hours ago [-]
The old drivers are mostly dual GPL or MIT licenced, they have been used in all the BSD variants.
hexo 8 hours ago [-]
that is so ridiculous.
not_your_vase 1 hours ago [-]
What is not clear for me, is the purpose of this project. What I could find is that it wants to be a drop-in replacement of PanVK. But since PanVK exists, I fail to see the point. Is PanVK similar to xserver, that is not salvageable? Or is it just RiiR? That's also a fine reason, but I'd expect that more for personal hobby projects.
tialaramex 9 hours ago [-]
> One simply cannot deploy a driver that [...] crashes and takes the user's work with it.
Somebody needs to tell whoever wrote the drivers in the PC where I'm writing this.
GZGavinZhao 9 hours ago [-]
Can't wait to write a Rust driver for my eink tablet <3
AndrewDucker 10 hours ago [-]
Interesting to see the building blocks come together. I hope that they can lay foundations that last.
Aldipower 10 hours ago [-]
Tyr is a Danish metal band. Period. :-)
robert_foss 9 hours ago [-]
I thought Tyr was the Norse god of War & Justice.
Considering that the Mali GPUs were developed by ARM Norway, and this driver is Just, I would say this is one aptly named driver.
pinkmuffinere 4 hours ago [-]
T-Y-R is also the root in semitic languages (eg, Hebrew and Arabic) related to flying! Maybe not on purpose, but I really like that incidental connection, given the combined reputation of rust and GPU operations for being fast.
Edit: apparently T-Y-R is not a root relating to flight in Hebrew! Maybe other Semitic languages, Im not sure. In Arabic it certainly is
afdbcreid 3 hours ago [-]
As a Hebrew speaker I cannot understand how you came into this conclusion. The closest I can think of is ת-י-ר, which is the root of being in a trip.
pinkmuffinere 2 hours ago [-]
Oh sorry, i thought it was also in Hebrew but it looks like it is not. I would expect the same root to show up in other Semitic languages, but at least in Arabic it’s
ط ي ر
afdbcreid 1 hours ago [-]
ChatGPT claims that the same or similar root does exist in the meaning "bird" or "to fly" in a lot of other Semitic languages. Interestingly, it also claims that there is some correspondence in Hebrew, in the noun תור (tor) that represents a specific kind of bird (turtledove).
einpoklum 1 hours ago [-]
Indeed, in Hebrew, תור (تور, Tor) is the word for turtle-dove.
MisterTea 9 hours ago [-]
Technically they are from the Faroe Islands. Great band, seen them live many times.
MilanTodorovic 9 hours ago [-]
Faroese actually
Rendered at 00:34:22 GMT+0000 (Coordinated Universal Time) with Vercel.
wow
Still, I think it makes a lot of sense. Completely new GPU drivers are quite rare and the macOS drivers from Asahi are a showcase proving that Rust and GPU drivers work together well. If there's any subcomponent switching to Rust-first for new contributions, it makes sense for it to be the one that had already been proven to be Rust-compatible.
This seems a bit exaggerated, their latest progress report is barely two months old: https://asahilinux.org/2025/12/progress-report-6-18/
They inarguably have slowed down, but this should be expected as the project matures. It has also inevitably now faced the time when new generations of contributors are needed as existing ones retire/ move to other projects.
How can it be mature if it can't even boot on newer MacBooks. The slowness does not seem to be due to running out of impactful work that needs to be done.
> Our priority is kernel upstreaming. Our downstream Linux tree contains over 1000 patches required for Apple Silicon that are not yet in upstream Linux. The upstream kernel moves fast, requiring us to constantly rebase our changes on top of upstream while battling merge conflicts and regressions. Janne, Neal, and marcan have rebased our tree for years, but it is laborious with so many patches. Before adding more, we need to reduce our patch stack to remain sustainable long-term...
Where do the M3 and M4 fit in? Until upstreaming and CI progress, the core team cannot prioritize new hardware.
https://asahilinux.org/2025/02/passing-the-torch/
I think the majority of that upstreaming work (that isn't on hold until the kernal is ready for the Rust graphics driver to land) has happened and additional features like DP alt mode for USB C have been demoed.
The next update from the team should land on their blog after 6.19 ships
Without proper support from upstream like AMD, Intel, and Qualcomm (to some extent) are doing, Linux will never work as well on Apple's hardware as it does on normal hardware.
https://asahilinux.org/docs/platform/feature-support/m3/
What do you see as progress here? Nothing is supported, everything is "to be announced" (i.e. unsupported).
https://lkml.org/lkml/2025/2/20/2066
> The document claims no subsystem is forced to take Rust
> That's been made clear pretty much from the very beginning, that nobody is forced to suddenly have to learn a new language, and that people who want to work purely on the C side can very much continue to do so.
If any C developer developed drivers in C previously for the DRM subsystem, they might in the future be forced to learn Rust.
Also they can just expose c bindings to these rust libraries no?
Somebody needs to tell whoever wrote the drivers in the PC where I'm writing this.
Considering that the Mali GPUs were developed by ARM Norway, and this driver is Just, I would say this is one aptly named driver.
Edit: apparently T-Y-R is not a root relating to flight in Hebrew! Maybe other Semitic languages, Im not sure. In Arabic it certainly is
ط ي ر