Hacker News Reader

Comment by jeffrallen | Hacker News Reader

Wish we were talking about making Fil-C required for apt, not Rust...

Those seems to be independent issues. Fil-C is about the best way to compile/run C code.

Rust would be about what language to use for new code.

Now that I have been programming in Rust for a couple of years, I don't want to go back to C (except for some hobby projects).

I agree. The main advantage of Fil-C is compatibility with C, in a secure way. The disadvantages are speed, and garbage collection. (Even thought, I read that garbage collection might not be needed in some cases; I would be very interested in knowing more details).

For new code, I would not use Fil-C. For kernel and low-level tools, other languages seem better. Right now, Rust is the only popular language in this space that doesn't have these disadvantages. But in my view, Rust also has issues, specially the borrow checker, and code verbosity. Maybe in the future there will be a language that resolves these issues as well (as a hobby, I'm trying to build such a language). But right now, Rust seems to be the best choice for the kernel (for code that needs to be fast and secure).

> disadvantages are speed, and garbage collection.

And size. About 10x increase both on disk and in memory

  $  stat -c '%s %n' {/opt/fil,}/bin/bash
  15299472 /opt/fil/bin/bash
   1446024 /bin/bash

  $ ps -eo rss,cmd | grep /bash
  34772 /opt/fil/bin/bash
   4256 /bin/bash

How does that compare with rust? You don't happen to have an example of a binary underway moving to rust in Ubuntu-land as well? Curious to see as I honestly don't know whether rust is nimble like C or not.

I don't have numbers, but Rust is also terrible for binary size. Large Rust binaries can be improved with various efforts, but it's not friendly by default. Rust focuses on runtime performance, high-level programming, and compile-time guarantees, but compile times and binary sizes are the drawback. Notably, Rust prefers static linking.

My impression is - rust fares a bit better on RAM footprint, and about as badly on disk binary size. It's darn hard to compare apples-to-apples, though - given it's a different language, so everything is a rewrite. One example:

Ubuntu 25.10's rust "coreutils" multicall binary: 10828088 bytes on disk, 7396 KB in RAM while doing "sleep".

Alpine 3.22's GNU "coreutils" multicall binary: 1057280 bytes on disk, 2320 KB in RAM while doing "sleep".

Fil-C is slow.

There is no C or C++ memory safe compiler with acceptable performance for kernels, rendering, games, etc. For that you need Rust.

The future includes Fil-C for legacy code that isn’t performance sensitive and Rust for new code that is.

What does that have to do with apt?

Enough of it is performance sensitive that Fil-C is not an option.

Fil-C is useful for the long tail of C/C++ that no one will bother to rewrite and is still usable if slow.

How is apt performance sensitive?

Apt has been painfully slow since I started using Debian last millennium, but I suspect it's not because it uses a lot of CPU, or it would be snappy by now.

It parses formats and does TLS, I’m assuming it’d be quite bad. I don’t think you can mix and match.

stuff that talks to "the internet" and runs as "root" seems like a good thing to build with filc.

It probably uses OS sandboxing primitives already.

In normal operation, apt has to be able to upgrade the kernel, the bootloader, and libc, so it can't usefully be sandboxed except for testing or chroots.

No, that doesn't follow. That only means the networking and parsing functions can't be sandboxed in the same process that drops new root-owned files. C and C++ services have been using subprocesses for sandboxing risky functionality for a long time now. It appears Apt has some version of this:

https://salsa.debian.org/apt-team/apt/-/blob/main/apt-pkg/co...

That's true; you can't usefully sandbox apt as a whole, but, because it verifies the signatures of the packages it downloads, you could usefully sandbox the downloading process, and you could avoid doing any parsing on the package file until you've validated its signature. It's a pleasant surprise to hear that it already does something like this!

How slow? In some contexts, the trade-off might be acceptable. From what I've seen in pizlonator's tweets, in some cases the difference in speed didn't seem drastic to me.

Yeah, I would happily run a bunch of my network services in this. I have loads of services that are public-facing doing a lot of complex parsing and rule evaluation and are mostly idle. For example my whole mailserver stack could probably benefit from this. My few messages an hour can run 2x slower. Maybe I would leave dovecot native since the attack surface before authentication is much lower and the performance difference would be more noticeable (mostly for things like searches).

You may be aware that one of the things Bernstein is famous for is revolutionizing mailserver security.

I imagine Apt is usually IO constrained?

That's my guess, yeah

Also, Fil-C's overheads are the lowest for programs that are pushing primitive bits around.

Fil-C's overheads are the highest for programs that chase pointers.

I'm guessing the CPU bound bits of apt (if there are any) are more of the former

No, Rust is awful for game development. It's not really what it was intended for. For one, all the graphics API are in C, so you would have to use unsafe FFI basically everywhere.

I wish, we will have something like Fil-C as an option for unsafe Rust.

Fil-C works because you recompile the whole C userspace. Unsafe Rust doesn't do that... and for many practical purposes you probably want to touch the non-safe-version of the C userspace.

Still, it's all LLVM, so perhaps unsafe Rust for Fil-space can be a thing, a useful one for catching (what would be) UBs even [Fil-C defines everything, so no UBs, but I'm assuming you want to eventually run it outside of Fil-space].

Now I actually wonder if Fil-C has an escape hatch somewhere for syscalls that it does not understand etc. Well it doesn't do inline assembly, so I shouldn't expect much... I wonder how far one needs to extend the asm clobber syntax for it to remotely come close to working.

at the bottom of the turtle stack, there's a yolo-c libc that does some syscall stuff:

> libyoloc.so. This is a mostly unmodified [musl/glibc] libc, compiled with Yolo-C. The only changes are to expose some libc internal functionality that is useful for implementing libpizlo.so. Note that libpizlo.so only relies on this library for system calls and a few low level functions. In the future, it's possible that the Fil-C runtime would not have a libc in Yolo Land, but instead libpizlo.so would make syscalls directly.

but mostly you are using a fil-c compiled libc:

> libc.so. This is a modified musl libc compiled with Fil-C. Most of the modifications are about replacing inline assembly for system calls with calls to libpizlo.so's syscall API.

That links here: https://github.com/pizlonator/fil-c/blob/deluge/filc/include...

Quotes from: https://fil-c.org/runtime

Unsafe Rust actually has a great runtime analyzer: Miri. It's very easy to just run `cargo +nightly miri test` in your project to get some confidence in the more questionable choices along the way.

doesnt it only work on x86_64?

Reply on news.ycombinator.com