anon (anon.delete@this.mailinator.com) on July 10, 2021 4:14 am wrote:
> cqwrteur (euloanty.delete@this.live.com) on July 9, 2021 9:56 pm wrote:
> > I just read the upcoming Rust code for Linux Kernel from Github. Maybe I am stupid.
> > What I see is just unsafe everywhere. All Rust functions are unsafe.
>
> Does the word unsafe trigger you? Rust under the keyword unsafe is still dramatically more
> safe than C. It does not turn off any of the safety features of rust, it just adds additional
> operations that cannot be checked by the compiler so need to be checked by humans.
>
> In this case, the unsafe operation is using ptr.write() (which is identical to c "*ptr = ...") to initialize
> a low-level primitive. Completely unchecked pointer access is obviously unsafe, because you could be doing
> anything. However, it is easy to see from the code that this particular use is indeed safe.
>
> There will be a lot of use of unsafe in the scaffolding that will be written to support
> rust in the kernel, because all low-level primitives fundamentally need it. The idea is
> that the drivers written on top of that scaffolding would not be using a lot of unsafe.

If it does not bring any safety, what's the point of using Rust? Add dependency to make kernel compilation harder? Or Adding binary bloat or other runtime bloat like rust fmt? Or just because Linus Torvalds believes using C or C++ violates human rights now?

Do not tell me you are waiting for Rust front-end GCC, you will be disappointed since Rust is an implementation, not a standard like C or C++ does.