| Crates.io | bootloader |
| lib.rs | bootloader |
| version | 0.11.14 |
| created_at | 2018-01-28 15:39:52.582736+00 |
| updated_at | 2026-01-13 18:26:01.446451+00 |
| description | An experimental x86_64 bootloader that works on both BIOS and UEFI systems. |
| homepage | |
| repository | https://github.com/rust-osdev/bootloader |
| max_upload_size | |
| id | 48654 |
| size | 128,513 |
An experimental x86_64 bootloader that works on both BIOS and UEFI systems. Written in Rust and some inline assembly, buildable on all platforms without additional build-time dependencies (just some rustup components).
You need a nightly Rust compiler with the llvm-tools-preview component, which can be installed through rustup component add llvm-tools-preview.
To use this crate, you need to adjust your kernel to be bootable first. Then you can create a bootable disk image from your compiled kernel. These steps are explained in detail below.
If you're already using an older version of the bootloader crate, follow our migration guides.
Our basic example showcases an OS that boots a minimal kernel using bootloader.
To combine your kernel with bootloader and create a bootable disk image, follow these steps:
bootloaderbootloader_api crate in your kernel's Cargo.toml.#![no_std] and #![no_main].fn kernel_main(boot_info: &'static mut bootloader_api::BootInfo) -> !. The function name can be arbitrary.
boot_info argument provides information about available memory, the framebuffer, and more. See the API docs for bootloader_api crate for details.entry_point macro to register the entry point function: bootloader_api::entry_point!(kernel_main);
_start entry point symbol for it. (If you use a linker script, make sure that you don't change the entry point name to something else.)&'static bootloader_api::BootloaderConfig to the entry_point macro. For example, you can require a specific stack size for your kernel:
const CONFIG: bootloader_api::BootloaderConfig = {
let mut config = bootloader_api::BootloaderConfig::new_default();
config.kernel_stack_size = 100 * 1024; // 100 KiB
config
};
bootloader_api::entry_point!(kernel_main, config = &CONFIG);
cargo build --target x86_64-unknown-none. You might need to run rustup target add x86_64-unknown-none before to download precompiled versions of the core and alloc crates.entry_point macro, the compiled executable contains a special section with metadata and the serialized config, which will enable the bootloader crate to load it.kernel subdirectory.os crate at the top level
$ cargo init --bin
# in Cargo.toml
[workspace]
resolver = "3"
members = []
$ cargo new kernel --bin
kernel crate as a build-dependency:
# in Cargo.toml
[build-dependencies]
kernel = { path = "kernel", artifact = "bin", target = "x86_64-unknown-none" }
Enable the unstable artifact-dependencies feature:
# .cargo/config.toml
[unstable]
bindeps = true
Experimental features are only available on the nightly channel:
# rust-toolchain.toml
[toolchain]
channel = "nightly"
targets = ["x86_64-unknown-none"]
std::process::Command to invoke the build command of your kernel in the build.rs script.build.rs build script in the os crate. See our disk image creation template for a more detailed example.
std::env::var_os("CARGO_BIN_FILE_MY_KERNEL_my-kernel")bootloader::UefiBoot and/or bootloader::BiosBoot to create a bootable disk image with your kernel.main.rs function. For example, run them with QEMU.See our disk image creation template for a more detailed example.
This project is split into three separate entities:
bootloader_api library with the entry point, configuration, and boot info definitions.
entry_point macro will encode the configuration settings into a separate ELF section of the compiled kernel executable.entry_point macro of the bootloader_api library.bootloader library to create bootable disk images that run a given kernel. This library is the top-level crate in this project.
build.rs.Licensed under either of
at your option.
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.