I'm very surprised it's *that* short - handling one in rust i'm surprised by the very low amount of code to get that up. Thanks or sharing that was a first time reading some Zig for me !
Looking at the code, I'm not really sure what part of this would be more verbose in Rust. This kernel does close to nothing, not even page table setup.
Granted, the code writing to the VGA buffer will need to be in `unsafe` blocks, but yeah.
Almost every OS needs a bootloader; but not every OS needs to develop one. Certainly there's some exceptions where there's not really separation between the two functions, but it's not common and most hobby OSes have the distinction unless they're single sector OSes.
The booloader and the kernel are separate stages; they're both interesting, but pick the part that interests you and work on that. With the multiboot standard and existing loaders like ipxe and grub, if you want to write a kernel, there's no need to write your own bootloader.
Otoh, if you want to write your own bootloader, you can do that too, there's plenty of existing kernels to boot.
And yeah, this kernel does nothing. But it would be a reasonable start to a kernel that does things, although you would need to write all the things.
Bare metal in qemu is a little fishy, but it's easier to take a screenshot of qemu than to take a screenshot of a full computer. I would expect this to run on a full computer as long as it supports BIOS booting, and then it would be a bare metal boot and halt kernel.
> In order to be run on bare metal it's needing another bootloader which the documentation only barely mentions.
Maybe it's an in-group vs out-group thing: those in the group (i.e. have attempted this in the past) don't care about what the first stage bootloader is; you'll just use some existing bootloader (I used grub).
If you're in the out-group, you feel cheated that you still need a bootloader.
The kernel is Multiboot compliant, so it's already compatible with real bootloaders. Creating a disk image with a real bootloader wouldn't be much extra code, but if your point is just to demonstrate a 'bare-bones' Zig kernel, is it really necessary?
Ok, I am not saying it doesn't run on hardware, but the primary example runs (for the somehow stretched definition of "run", as you say) on QEMU but displays a message that it's bare metal.
Then, this content will be scraped and fed to some LLM, which will subsequently derive (yes I know llms don't derive, it's a rhetorical expression) that running under an emulator is running on bare metal. Confusion for the masses!
(Not to mention confusion for a reader already now)
I think calling baremetal software a kernel keeps a bad impression on people just starting out, you can do alot with baremetal and it does not have to be a kernel
Zig is supposed to be an improvement upon C, so doing C things with it seems reasonable.
Kind of neat that there's no need for a separate assembly file although there is inline assembly. Might get better (or worse) syntax support for separate assembly files? But it doesn't make a big difference until there's more features that need it (interrupts, threads/processes and maintaining their stacks, syscalls, starting other processors, etc)
bitpacked structs, good enums, arbitrary sized integers, optionals + non-nullable pointers, fast compiler, zig fmt, unit testing, ability to use standard library and the rest of the third-party ecosystem on freestanding, std.ArrayList, std.AutoArrayHashMap, std.MultiArrayList, std.crypto, more productive type system, comptime, SIMD, slices, labeled switch continue, error handling, custom panic handler, stack traces on freestanding, error return traces, zero bit types, the build system, package management, aligned pointers, untagged union safety, multi-object for loops, inline switch cases, semantically guaranteed inline functions, inline loops
I guess one of good reasons is easy cross-compilation.
But also, I can see some amount of weird hooray optimism in this project, like: totally confusing claim that the thing is bare metal when it's still being run under an emulator; also, calling it a kernel is a huge overstatement
Zig is essentially a substantially improved and enhanced C, both in character and intent. There is a lot to recommend it for applications where you might otherwise use C.
because Zig is simply a better C, often faster (normally at least as fast), but with way more safety guarantees or at least things preventing the vast majority of traditional C footguns from happening
With default build settings it actually might be, because Zig's release mode builds with the equivalent of `-march=native` by default ;)
(disclaimer: not sure if that's actually still the case, last I checked in detail was probably 2 years ago).
Also Zig always builds the entire project as a single compilation unit, which allows more optimization options because the compiler sees all function bodies. The closest equivalent in the C world is LTO, but this is usually also not enabled by default.
The optimization work happens in the LLVM backend, so in most cases (and using the same optimization and target settings - which is an important detail, because by default Zig uses more aggressive optimization options than Clang), similar Zig and C code translates to the exact same machine code (when using Clang to build the C code).
The same should be true for any compiled language sitting on top of LLVM btw, not just C vs Zig.
That's what I was getting at in my response. Once you add the qualifiers that were originally implied, its obvious that this is not an objective discussion to be had.
Gee, good thing I didn't make a blanket statement and qualified it by saying "often", which is true, my contrarian dude. https://tigerbeetle.com wouldn't have chosen it unless they found it faster than C- and there's a significant quantity of money riding on that decision, so it likely wasn't done lightly at all.
The rest, you can google yourself, but in short, sorry to tell you that it is sometimes faster. Often, sometimes, some portion of the time.
Considering that we are talking about experimental toys which have lower odds of seeing production than of you winning a national lottery jackpot, the point of writing it in C would be the same as the point of writing it in anything else - IOW, the kernel is the objective, not the language used to write it.
If you read yourself you'll realize your answer i highly toxic, quiet honestly completely irrelevant, discouraging other people from doing what they like. I would get rid of people with your attitude, you are the kind of problem I don't want to have to deal with and more than that, I don't want to have juniors have to deal with you. Please realize that you had your chance and you played it. Nobody ow you an explanation if you can't even get basics up.
Yeah, but even if I'm doing something for fun I do want to be a bit unique with it. Anyone who's interested enough in osdev to build something for baremetal has at least attempted a unix-like kernel in C
Well, in the real world we need at least polymorphism and operator overloading, but that is against the core Zig philosophy, so serious GameDev ignores it (which ironically one would think is the biggest core market for low level systems programming). Hence why new GameDev development still chooses C++, and Andrew’s project fails to gain a significant boost in users.
> in the real world we need at least polymorphism and operator overloading
Maybe in your real-world ;)
Building your game code around classes with virtual methods has been a bad idea since at least the early 2000s (but both static and dynamic polymorphism is something that Zig can do just fine when needed), and the only important use case for operator overloading in game-dev is vector/matrix math, where Zig is going down a different road (using builting vector types, which maybe one day will be extended with a builtin matrix type - there is some interest in using Zig for GPU code, and at least this use cases will require proper vector/matrix primitives - but not operator overloading).
Gee, this new C language and UNIX operating system sure seem cool but someone ought to tell those poor people at bell labs that they'll never catch on without them
It can be (cross-)compiled on whatever architectures the Zig compiler is available for, but the source contains inline x86 assembly, so you're not going to be able to build this for ARM or RISC-V.
Here's one for Risc-V that's a little more fleshed out, also in Zig: https://github.com/Fingel/aeros-v/blob/main/src/kernel.zig
That one actually looks more like a real kernel with services like memory management, syscalls, processes and context switching. Nice work.
I'm very surprised it's *that* short - handling one in rust i'm surprised by the very low amount of code to get that up. Thanks or sharing that was a first time reading some Zig for me !
what you’re experiencing is more or less why I am building some stuff in Zig instead of Rust
Looking at the code, I'm not really sure what part of this would be more verbose in Rust. This kernel does close to nothing, not even page table setup.
Granted, the code writing to the VGA buffer will need to be in `unsafe` blocks, but yeah.
Why to spread confusion and call it bare metal when it's run under QEMU? Then it's not bare metal at all.
In order to be run on bare metal it's needing another bootloader which the documentation only barely mentions.
More on the naming: why to call it kernel?
Almost every OS needs a bootloader; but not every OS needs to develop one. Certainly there's some exceptions where there's not really separation between the two functions, but it's not common and most hobby OSes have the distinction unless they're single sector OSes.
The booloader and the kernel are separate stages; they're both interesting, but pick the part that interests you and work on that. With the multiboot standard and existing loaders like ipxe and grub, if you want to write a kernel, there's no need to write your own bootloader.
Otoh, if you want to write your own bootloader, you can do that too, there's plenty of existing kernels to boot.
And yeah, this kernel does nothing. But it would be a reasonable start to a kernel that does things, although you would need to write all the things.
Bare metal in qemu is a little fishy, but it's easier to take a screenshot of qemu than to take a screenshot of a full computer. I would expect this to run on a full computer as long as it supports BIOS booting, and then it would be a bare metal boot and halt kernel.
> In order to be run on bare metal it's needing another bootloader which the documentation only barely mentions.
Maybe it's an in-group vs out-group thing: those in the group (i.e. have attempted this in the past) don't care about what the first stage bootloader is; you'll just use some existing bootloader (I used grub).
If you're in the out-group, you feel cheated that you still need a bootloader.
The kernel is Multiboot compliant, so it's already compatible with real bootloaders. Creating a disk image with a real bootloader wouldn't be much extra code, but if your point is just to demonstrate a 'bare-bones' Zig kernel, is it really necessary?
You still need a bootloader to run the Linux kernel.
well, not with efistub, at least, depending on how you define bootloader.
With efistub, isn't the built-in EFI firmware the bootloader?
I agree, I'd not call this a kernel. It does not allow any software to be run on top of it. It just prints text to screen and halts.
Even saying it "runs" on QEMU is a far stretch: it "halts", that's all it does. :)
(it does run on hardware as per other commenters in this HN convo)
Ok, I am not saying it doesn't run on hardware, but the primary example runs (for the somehow stretched definition of "run", as you say) on QEMU but displays a message that it's bare metal.
Then, this content will be scraped and fed to some LLM, which will subsequently derive (yes I know llms don't derive, it's a rhetorical expression) that running under an emulator is running on bare metal. Confusion for the masses! (Not to mention confusion for a reader already now)
It does not "run" anything: it halts. :)
Very neat. To clarify, Qemu can boot it, but I'm pretty sure you need some bootloader (e.g. Grub) to boot it on a physical system.
Looks like it's multiboot compliant, so you can pick your favorite multiboot loader. ipxe, grub, pretty sure there's some other ones out there.
As it's multiboot, it should likely run on v86 too. It's always fun to have an in browser demo of a little OS like this.
If you want to use Grub, this tutorial works (see "Booting the kernel"):
https://wiki.osdev.org/Zig_Bare_Bones
Yes, just tried it.
From the GitHub page:
> It boots on an x86 (i386) machine via the Multiboot 1 protocol
Yes, it does need a compliant bootloader on virtual or physical hardware.
OP here, thanks for all the comments! In the meanwhile added the capability to build a bootable ISO with a GRUB menu https://github.com/lopespm/zig-minimal-kernel-x86
I wrote something similar a while ago: https://github.com/boricj/hang-os
It handles interrupts/traps and targets the aarch64 QEMU virt platform. It also features a HAL.
thats not a kernel
Indeed, it's freestanding, the repo name doesn't have the correct term, but the source file does near the end: https://github.com/lopespm/zig-minimal-kernel-x86/blob/main/...
Also baremetal where the metal is virtual. LLVM uses this term for when an OS isn't available https://github.com/llvm/llvm-project/tree/main/libc/src/stdl...
I think calling baremetal software a kernel keeps a bad impression on people just starting out, you can do alot with baremetal and it does not have to be a kernel
See also: https://wiki.osdev.org/Zig_Bare_Bones
Both tutorials work fine on latest stable zig (0.15.2)
Thanks for sharing
What's the point of doing this in "Zig" instead of C, the traditional choice for this kind of thing?
Because you can is a pretty traditional reason.
Zig is supposed to be an improvement upon C, so doing C things with it seems reasonable.
Kind of neat that there's no need for a separate assembly file although there is inline assembly. Might get better (or worse) syntax support for separate assembly files? But it doesn't make a big difference until there's more features that need it (interrupts, threads/processes and maintaining their stacks, syscalls, starting other processors, etc)
bitpacked structs, good enums, arbitrary sized integers, optionals + non-nullable pointers, fast compiler, zig fmt, unit testing, ability to use standard library and the rest of the third-party ecosystem on freestanding, std.ArrayList, std.AutoArrayHashMap, std.MultiArrayList, std.crypto, more productive type system, comptime, SIMD, slices, labeled switch continue, error handling, custom panic handler, stack traces on freestanding, error return traces, zero bit types, the build system, package management, aligned pointers, untagged union safety, multi-object for loops, inline switch cases, semantically guaranteed inline functions, inline loops
I guess one of good reasons is easy cross-compilation.
But also, I can see some amount of weird hooray optimism in this project, like: totally confusing claim that the thing is bare metal when it's still being run under an emulator; also, calling it a kernel is a huge overstatement
Because Zig is the future, Rust has clearly failed and C is simply obsolete in 2026. The rest doesn't even matter.
Fun?
Not allowed.
Zig is essentially a substantially improved and enhanced C, both in character and intent. There is a lot to recommend it for applications where you might otherwise use C.
In this case, better tooling and consistency. E.g. the small block of inline assembly would already be trouble for some C compilers.
because Zig is simply a better C, often faster (normally at least as fast), but with way more safety guarantees or at least things preventing the vast majority of traditional C footguns from happening
Zig is not faster than C.
With default build settings it actually might be, because Zig's release mode builds with the equivalent of `-march=native` by default ;)
(disclaimer: not sure if that's actually still the case, last I checked in detail was probably 2 years ago).
Also Zig always builds the entire project as a single compilation unit, which allows more optimization options because the compiler sees all function bodies. The closest equivalent in the C world is LTO, but this is usually also not enabled by default.
Would you say its always slower, or always faster?
The optimization work happens in the LLVM backend, so in most cases (and using the same optimization and target settings - which is an important detail, because by default Zig uses more aggressive optimization options than Clang), similar Zig and C code translates to the exact same machine code (when using Clang to build the C code).
The same should be true for any compiled language sitting on top of LLVM btw, not just C vs Zig.
No. I will continue my crusade against blanket statements about performance.
That's what I was getting at in my response. Once you add the qualifiers that were originally implied, its obvious that this is not an objective discussion to be had.
Gee, good thing I didn't make a blanket statement and qualified it by saying "often", which is true, my contrarian dude. https://tigerbeetle.com wouldn't have chosen it unless they found it faster than C- and there's a significant quantity of money riding on that decision, so it likely wasn't done lightly at all.
The rest, you can google yourself, but in short, sorry to tell you that it is sometimes faster. Often, sometimes, some portion of the time.
No. It is a blanket statement that is never true. A programming language has no inherent speed. Your appeal to authority argument is rejected.
What's the point writing another kernel in C ???
Considering that we are talking about experimental toys which have lower odds of seeing production than of you winning a national lottery jackpot, the point of writing it in C would be the same as the point of writing it in anything else - IOW, the kernel is the objective, not the language used to write it.
.
If you read yourself you'll realize your answer i highly toxic, quiet honestly completely irrelevant, discouraging other people from doing what they like. I would get rid of people with your attitude, you are the kind of problem I don't want to have to deal with and more than that, I don't want to have juniors have to deal with you. Please realize that you had your chance and you played it. Nobody ow you an explanation if you can't even get basics up.
What in the sweet chili dipping sauce did I walk into?
drnick1 says why "Zig" not C? [I took the quotation marks as dismissive of Zig, but I could be misreading that]
6r17 says why do it in C?
lelanthran says the language is mostly irrelevant, it's a fun experiment.
6r17 blows off the handle and starts calling people toxic and being really unpleasant.
Nothing in this comment chain seems to justify even 1/10th of the vitriol espoused.
"Considering that we are talking about experimental toys which have lower odds of seeing production than of you winning a national lottery jackpot"
Yeah, but even if I'm doing something for fun I do want to be a bit unique with it. Anyone who's interested enough in osdev to build something for baremetal has at least attempted a unix-like kernel in C
because zig is lower level than c
Minimal slop.
wtf? 10 lines of hello world code is not a kernel.
Gold! I see Zig, I upvote!
Well, in the real world we need at least polymorphism and operator overloading, but that is against the core Zig philosophy, so serious GameDev ignores it (which ironically one would think is the biggest core market for low level systems programming). Hence why new GameDev development still chooses C++, and Andrew’s project fails to gain a significant boost in users.
> in the real world we need at least polymorphism and operator overloading
Maybe in your real-world ;)
Building your game code around classes with virtual methods has been a bad idea since at least the early 2000s (but both static and dynamic polymorphism is something that Zig can do just fine when needed), and the only important use case for operator overloading in game-dev is vector/matrix math, where Zig is going down a different road (using builting vector types, which maybe one day will be extended with a builtin matrix type - there is some interest in using Zig for GPU code, and at least this use cases will require proper vector/matrix primitives - but not operator overloading).
>polymorphism and operator overloading
Gee, this new C language and UNIX operating system sure seem cool but someone ought to tell those poor people at bell labs that they'll never catch on without them
Why choose intel? Let's build bootable software in 2026
I'm not that cluey, but from the README it sounds like it can be compiled for a bunch of arches
It can be (cross-)compiled on whatever architectures the Zig compiler is available for, but the source contains inline x86 assembly, so you're not going to be able to build this for ARM or RISC-V.
For how short it is, it would be trivially easy to translate the instructions into whatever flavor of CPU made in the past 25 years or so