Re: Rust in the kernel

[Akib Azmain Turja]

jbranso@dismail.de writes:

> July 4, 2022 1:36 PM, "Akib Azmain Turja" <akib@disroot.org> wrote:
>
>> Ludovic Courtès <ludo@gnu.org> writes:
>> 
>>> Hi!
>>> 
>>> Leo Famulari <leo@famulari.name> skribis:
>> 
>> The effort to use the Rust programming language within the Linux kernel
>> is progressing and may be realized in the next few months:
>> 
>> https://lwn.net/SubscriberLink/899182/6c831b90eaee015e
>> https://www.memorysafety.org/blog/memory-safety-in-linux-kernel
>> 
>> Within Guix, we'll need to adapt our kernel build processes in order to
>> support this.
>> 
>> Although I help with updating and configuring the kernel builds, I won't
>> be able to participate in the "Rust in the kernel" effort for Guix.
>>> Understood…
>> 
>> So, interested volunteers should begin organizing :)
>>> Yup!
>>> 
>>> Now, concretely, how long will it take before key parts of the kernel
>>> are written in Rust? Hopefully a long time, no? Per the article above,
>>> it’s starting small, with Rust usage in well-defined locations.
>>> 
>>> This is not to say that we shouldn’t start organizing, but rather that
>>> we still have a bit of time ahead.
>>> 
>>> (During that time, interested readers can also take a stab at improving
>>> support for the Hurd, which relies on that revolutionary technology
>>> called “address spaces” to ensure Memory Safety™ among other things!)
>>> 
>>> Ludo’.
>> 
>> "Address spaces"! What's that? Sorry for asking without searching the
>> internet first, but the Hurd designers are so creative that a few
>> understand the concepts and join the community, so there is a little
>> chance (if any) that I'll find any useful information on that.
>
> From the Hurd wiki: https://www.gnu.org/software/hurd/advantages.html
>
> The Hurd is built in a very modular fashion. Other Unix-like kernels
> (Linux, for example) are also modular in that they allow loading
> (and unloading) some components as kernel modules, but the Hurd goes
> one step further in that most of the components that constitute the
> whole kernel are running as separate user-space processes and are thus
> using different address spaces that are isolated from each other.
> This is a multi-server design based on a microkernel. It is not
> possible that a faulty memory dereference inside the TCP/IP stack
> can bring down the whole kernel, and thus the whole system, which
> is a real problem in a monolithic Unix kernel architecture.
>
> Some visual explantions:
>
> https://en.wikipedia.org/wiki/Microkernel#/media/File:OS-structure.svg
>
> The Hurd is on the right in this image.

Thanks, now I understand Ludo' was saying about virtual address space,
achieved using paging.
>
> Essentially, if your fileserver somehow gets hacked, the attacker
> cannot magically access your TCP/IP stack, because your TCP/IP is not
> in the some "software zone" as your fileserver. So microkernels like
> the Hurd are usually considered more secure and better designed
> than monolithic kernels like Linux. However, monolithic kernels
> will usually be faster than microkernels.

I know microkernels are theorically slow due to the heavy use IPC.  But
is it really impossible for well written microkernel to beat a well
written monolithic kernel?  L4 is super-fast, is it still slower than
Linux?

>
>> --
>> Akib Azmain Turja
>> 
>> This message is signed by me with my GnuPG key. It's fingerprint is:
>> 
>> 7001 8CE5 819F 17A3 BBA6 66AF E74F 0EFA 922A E7F5

-- 
Akib Azmain Turja

This message is signed by me with my GnuPG key.  It's fingerprint is:

    7001 8CE5 819F 17A3 BBA6  66AF E74F 0EFA 922A E7F5

signature.asc
Description: PGP signature