vhaudiquet 2e6c84b65b fix(supervise): report engine failures instead of guest exit codes
Only the guest init can mint a guest code: it reports the payload's
code through the control file and the VMM exits the child with it. A
signal on the machine child therefore means the VMM or the engine
died; mapping it to 128+n masqueraded an OOM-killed VMM as the
payload having died by that signal (T3).

125 and 126 keep their two-sided reading: the child bails with 125 on
a boot failure and the init refuses with 126 on a failed share or a
missing workdir, but a payload may genuinely exit either code — the
control file is the arbiter, and a run that never reached the payload
is an error naming the cause.
2026-10-30 03:44:55 +01:00
2026-10-29 21:47:14 +01:00
2026-10-30 01:04:24 +01:00

écrou logo

ecrou

a small microVM runtime based on libkrun

Try ecrou in the terminal

Installation

ecrou is available as a strictly confined snap:

sudo snap install ecrou
sudo snap connect ecrou:kvm

Usage

$ ecrou run -it ubuntu
root@ecrou-vm-m2glhl:/# 

Embed ecrou into your project

ecrou is both a library and a tiny CLI frontend around it. The library is embeddable into any project, allowing to use microVMs easily.

Rust

ecrou is a Rust library, which can just be called:

use ecrou::vm::RunOptions;
use ecrou::NoProgress;

fn main() -> ecrou::Result<std::process::ExitCode> {
    let code = ecrou::vm::run(
        &RunOptions {
            image: "ubuntu:resolute".into(),
            command: vec!["uname".into(), "-a".into()],
            ..RunOptions::default()
        },
        &NoProgress,
    )?;
    Ok(std::process::ExitCode::from(code.clamp(0, 255) as u8))
}
Languages
Rust 99.4%
Ruby 0.6%