host: process table, shell, coreutils and harness (T8, T9)

T8: sausage:process WIT (spawn, wait, kill), one store per process over the
shared root, /proc/processes, /run, pipeline and cancellation groups, Exec.
brush and 78 uutils commands; the shared cwd-init object sets the cwd from PWD
and keeps numeric exit codes. The corpus matches native bash.

T9: V8 (deno_core) runs pi-durable, pi-ai and chord; SQLite storage in
<dir>/host/harness.sqlite (WAL, synchronous=FULL); file tools on the VFS, bash
through Exec; resumes after SIGKILL.
This commit is contained in:
2026-11-06 16:34:34 +00:00
parent 5ce2a2f2bd
commit 57a47b069e
41 changed files with 87601 additions and 82 deletions
+147
View File
@@ -0,0 +1,147 @@
//! Rust operations for tools. File access uses the world's shared VFS layer.
use crate::exec::{Command, Exec, Options, Stream};
use deno_core::{OpState, op2};
use deno_error::JsErrorBox;
use sausage_vfs::{Layer, Open};
use serde_json::{Value, json};
use std::{cell::RefCell, io::Write, rc::Rc, sync::Arc};
pub(super) struct Environment {
pub root: Arc<dyn Layer>,
pub exec: Arc<dyn Exec>,
}
#[op2]
#[serde]
pub async fn op_file(
state: Rc<RefCell<OpState>>,
#[string] kind: String,
#[string] path: String,
#[serde] bytes: Vec<u8>,
) -> Result<serde_json::Value, JsErrorBox> {
let root = state.borrow().borrow::<Environment>().root.clone();
let result: std::io::Result<Value> = async {
match kind.as_str() {
"read" => {
let file = root.open(&path, Open { follow: true, ..Default::default() }).await?;
let size = file.stat().await?.size;
Ok(json!(file.read(0, size as usize).await?))
}
"write" => {
let file = root.open(&path, Open { follow: true, write: true, create: true, truncate: true, ..Default::default() }).await?;
let mut offset = 0;
while offset < bytes.len() {
let count = file.write(offset as u64, bytes[offset..].to_vec()).await?;
if count == 0 { return Err(std::io::ErrorKind::WriteZero.into()); }
offset += count;
}
file.sync().await?;
Ok(Value::Null)
}
"mkdir" => { root.mkdir(&path).await?; Ok(Value::Null) }
"stat" => {
let s = root.stat(&path, true).await?;
Ok(json!({"path": path, "name": path.rsplit('/').next().unwrap_or(""), "kind": if s.directory {"directory"} else {"file"}, "size": s.size, "mtimeMs": 0}))
}
_ => Err(std::io::ErrorKind::Unsupported.into()),
}
}.await;
Ok(match result {
Ok(value) => json!({"ok": true, "value": value}),
Err(e) => {
json!({"ok": false, "error": {"code": match e.kind() { std::io::ErrorKind::NotFound => "not_found", std::io::ErrorKind::PermissionDenied => "permission_denied", _ => "unknown" }, "message": e.to_string(), "path": path}})
}
})
}
struct Execution {
receiver: tokio::sync::Mutex<tokio::sync::mpsc::UnboundedReceiver<Value>>,
task: tokio::task::AbortHandle,
}
impl deno_core::Resource for Execution {
fn close(self: Rc<Self>) {
self.task.abort();
}
}
impl Drop for Execution {
fn drop(&mut self) {
self.task.abort();
}
}
#[op2]
pub fn op_exec(
state: &mut OpState,
#[serde] command: serde_json::Value,
#[string] cwd: String,
timeout: Option<f64>,
) -> Result<u32, JsErrorBox> {
let exec = state.borrow::<Environment>().exec.clone();
let command = if let Some(s) = command.as_str() {
Command::Shell(s.to_owned())
} else {
Command::Argv(
serde_json::from_value(command).map_err(|e| JsErrorBox::generic(e.to_string()))?,
)
};
let (sender, receiver) = tokio::sync::mpsc::unbounded_channel();
let task = tokio::spawn(async move {
let result = exec.exec(command, Options { cwd, timeout: timeout.map(std::time::Duration::from_secs_f64), ..Default::default() }, &mut |stream, bytes| { let _ = sender.send(json!({"stream": if stream == Stream::Stdout {"stdout"} else {"stderr"}, "bytes": bytes})); }).await;
let _ = sender.send(match result {
Ok(exit) => json!({"exitCode": exit}),
Err(e) => json!({"error": e.to_string()}),
});
});
Ok(state.resource_table.add(Execution {
receiver: tokio::sync::Mutex::new(receiver),
task: task.abort_handle(),
}))
}
#[op2]
#[serde]
pub async fn op_exec_next(
state: Rc<RefCell<OpState>>,
rid: u32,
) -> Result<serde_json::Value, JsErrorBox> {
let execution = state
.borrow()
.resource_table
.get::<Execution>(rid)
.map_err(|e| JsErrorBox::generic(e.to_string()))?;
let value = execution
.receiver
.lock()
.await
.recv()
.await
.unwrap_or(Value::Null);
if value.get("stream").is_none() {
let resource = state
.borrow_mut()
.resource_table
.take::<Execution>(rid)
.map_err(|e| JsErrorBox::generic(e.to_string()))?;
deno_core::Resource::close(resource);
}
Ok(value)
}
#[op2]
#[string]
pub async fn op_input() -> Result<Option<String>, JsErrorBox> {
use tokio::io::AsyncBufReadExt;
let mut reader = tokio::io::BufReader::with_capacity(1, tokio::io::stdin());
let mut line = String::new();
let n = reader
.read_line(&mut line)
.await
.map_err(|e| JsErrorBox::generic(e.to_string()))?;
Ok((n != 0).then_some(line))
}
#[op2(fast)]
pub fn op_output(#[string] text: &str) {
print!("{text}");
let _ = std::io::stdout().flush();
}
+209 -5
View File
@@ -1,8 +1,212 @@
//! The durable agent harness: V8 with pi-durable, harness storage, and the tools' environment.
//! Owned by task T9.
//! Durable V8 harness, host-side SQLite storage, and world tool operations.
mod environment;
use crate::{exec::Exec, world::World};
use deno_core::{JsRuntime, OpState, RuntimeOptions, op2};
use deno_error::JsErrorBox;
use rusqlite::{Connection, types::Value as SqlValue};
use serde_json::{Value, json};
use std::path::Path;
use std::sync::Arc;
/// Start the world at `world` with the harness and a console.
pub async fn run(_world: &Path) -> anyhow::Result<()> {
anyhow::bail!("the harness does not exist yet (T9)")
#[op2]
#[serde]
fn op_sql(
state: &mut OpState,
#[string] kind: String,
#[string] sql: String,
#[serde] params: Vec<serde_json::Value>,
) -> Result<serde_json::Value, JsErrorBox> {
let work = || -> Result<Value, Box<dyn std::error::Error>> {
if kind == "open" {
let db = Connection::open(sql)?;
db.execute_batch("PRAGMA journal_mode=WAL; PRAGMA synchronous=FULL;")?;
state.put(db);
return Ok(Value::Null);
}
if kind == "close" {
state.take::<Connection>();
return Ok(Value::Null);
}
let db = state.borrow::<Connection>();
if kind == "exec" {
db.execute_batch(&sql)?;
return Ok(Value::Null);
}
let params: Vec<SqlValue> = params
.into_iter()
.map(|v| match v {
Value::Null => SqlValue::Null,
Value::Number(n) => n
.as_i64()
.map(SqlValue::Integer)
.unwrap_or_else(|| SqlValue::Real(n.as_f64().unwrap())),
Value::String(s) => SqlValue::Text(s),
Value::Array(a) => {
SqlValue::Blob(a.into_iter().map(|n| n.as_u64().unwrap() as u8).collect())
}
_ => panic!("unsupported SQL parameter"),
})
.collect();
let mut stmt = db.prepare_cached(&sql)?;
if kind == "run" {
let changes = stmt.execute(rusqlite::params_from_iter(params))?;
return Ok(json!({"changes": changes, "lastInsertRowid": db.last_insert_rowid()}));
}
let names: Vec<String> = stmt.column_names().into_iter().map(str::to_owned).collect();
let mut rows = stmt.query(rusqlite::params_from_iter(params))?;
let mut result = vec![];
while let Some(row) = rows.next()? {
let mut obj = serde_json::Map::new();
for (i, name) in names.iter().enumerate() {
let v = match row.get::<_, SqlValue>(i)? {
SqlValue::Null => Value::Null,
SqlValue::Integer(n) => json!(n),
SqlValue::Real(n) => json!(n),
SqlValue::Text(s) => json!(s),
SqlValue::Blob(b) => json!(b),
};
obj.insert(name.clone(), v);
}
result.push(Value::Object(obj));
}
Ok(json!(result))
};
work().map_err(|e| JsErrorBox::generic(e.to_string()))
}
deno_core::extension!(
storage,
ops = [
op_sql,
environment::op_file,
environment::op_exec,
environment::op_exec_next,
environment::op_input,
environment::op_output
]
);
/// Start the console with the process implementation supplied by the host.
pub async fn run_with_exec(world: World, exec: Arc<dyn Exec>) -> anyhow::Result<()> {
// Deno's unsynchronized operations require a current-thread Tokio runtime.
tokio::task::spawn_blocking(move || {
tokio::runtime::Builder::new_current_thread()
.enable_all()
.build()?
.block_on(run_module(world, exec, include_str!("../../js/bundle.js")))
})
.await?
.map_err(|error: anyhow::Error| {
let mut text = error.to_string();
for key in ["PI_PROXY_BASE_URL", "PI_PROXY_API_KEY"] {
if let Ok(value) = std::env::var(key) {
if !value.is_empty() {
text = text.replace(&value, "[redacted]");
}
}
}
anyhow::anyhow!(text)
})
}
/// Evaluate a bundled harness module with the same storage and tool operations as the console.
/// The caller must use a current-thread Tokio runtime.
pub async fn run_module(world: World, exec: Arc<dyn Exec>, source: &str) -> anyhow::Result<()> {
let mut runtime = JsRuntime::new(RuntimeOptions {
extensions: vec![
deno_telemetry::deno_telemetry::init(),
deno_webidl::deno_webidl::init(),
deno_web::deno_web::init(
Arc::new(deno_web::BlobStore::default()),
None,
false,
Default::default(),
),
deno_crypto::deno_crypto::init(None),
deno_net::deno_net::init(None, None),
deno_fetch::deno_fetch::init(Default::default()),
storage::init(),
],
extension_transpiler: Some(std::rc::Rc::new(|name, source| {
if !name.ends_with(".ts") {
return Ok((source, None));
}
let parsed = deno_ast::parse_module(deno_ast::ParseParams {
specifier: deno_core::url::Url::parse(&name).unwrap(),
text: source.into(),
media_type: deno_ast::MediaType::TypeScript,
capture_tokens: false,
scope_analysis: false,
maybe_syntax: None,
})
.map_err(|e| JsErrorBox::generic(e.to_string()))?;
let emitted = parsed
.transpile(
&Default::default(),
&Default::default(),
&deno_ast::EmitOptions {
source_map: deno_ast::SourceMapOption::None,
..Default::default()
},
)
.map_err(|e| JsErrorBox::generic(e.to_string()))?
.into_source();
Ok((emitted.text.into(), None))
})),
..Default::default()
});
runtime
.op_state()
.borrow_mut()
.put(deno_permissions::PermissionsContainer::allow_all(Arc::new(
deno_permissions::RuntimePermissionDescriptorParser::new(sys_traits::impls::RealSys),
)));
runtime.execute_script("bootstrap.js", include_str!("../../js/bootstrap.js"))?;
let env: serde_json::Map<String, Value> = ["PI_PROXY_BASE_URL", "PI_PROXY_API_KEY"]
.into_iter()
.filter_map(|key| std::env::var(key).ok().map(|v| (key.to_owned(), json!(v))))
.collect();
runtime.execute_script("env.js", format!("process.env = {};", Value::Object(env)))?;
runtime
.op_state()
.borrow_mut()
.put(environment::Environment {
root: world.root,
exec,
});
runtime.execute_script(
"config.js",
format!(
"globalThis.storagePath = {};",
serde_json::to_string(&world.host_dir.join("harness.sqlite"))?
),
)?;
let specifier = deno_core::ModuleSpecifier::parse("file:///harness.js")?;
let module = runtime
.load_main_es_module_from_code(&specifier, source.to_owned())
.await?;
let result = runtime.mod_evaluate(module);
runtime
.with_event_loop_promise(result, Default::default())
.await?;
runtime.run_event_loop(Default::default()).await?;
Ok(())
}
struct Unavailable;
#[async_trait::async_trait]
impl Exec for Unavailable {
async fn exec(
&self,
_: crate::exec::Command,
_: crate::exec::Options,
_: &mut (dyn for<'chunk> FnMut(crate::exec::Stream, &'chunk [u8]) + Send),
) -> anyhow::Result<i32> {
anyhow::bail!("process execution is not wired yet")
}
}
/// Open a world and resume its durable root conversation.
pub async fn run(path: &Path) -> anyhow::Result<()> {
run_with_exec(World::open(path)?, Arc::new(Unavailable)).await
}
+546 -6
View File
@@ -1,8 +1,548 @@
//! The process table: WASI programs in their own stores, the process interface (spawn, wait, kill),
//! `/proc`, and the `Exec` implementation. Owned by task T8.
use std::path::Path;
//! Runs components in independent stores over the world root. Owns process groups, stdio, and Exec.
use crate::{
exec::{Command, Exec, Options, Stream},
world::World,
};
use sausage_vfs::{
Mounts,
adapter::{self, Adapter, View},
};
use std::{
collections::HashMap,
path::Path,
sync::{
Arc, Mutex,
atomic::{AtomicBool, AtomicU32, Ordering},
},
};
use tokio::{
io::{AsyncReadExt, DuplexStream},
sync::watch,
task::AbortHandle,
};
use wasmtime::{
Engine, Store,
component::{Component, Linker, Resource, ResourceTable},
};
use wasmtime_wasi::{
WasiCtx, WasiCtxBuilder, WasiCtxView, WasiView,
cli::{AsyncStdinStream, AsyncStdoutStream},
p2::{
DynInputStream, DynOutputStream,
bindings::CommandPre,
pipe::{AsyncReadStream, AsyncWriteStream},
},
};
/// Run a shell in the world at `world`, or one command with `command`. Returns the exit status.
pub async fn sh(_world: &Path, _command: Option<String>) -> anyhow::Result<i32> {
anyhow::bail!("the process table does not exist yet (T8)")
mod bindings {
wasmtime::component::bindgen!({
path: "wit", world: "program",
imports: { default: async | trappable },
with: {
"wasi:io": wasmtime_wasi::p2::bindings::io,
"sausage:process/host.child": super::Child,
},
});
}
use bindings::sausage::process::host::{Host, HostChild, Request};
/// A finite snapshot of process metadata exposed through /proc/processes.
#[derive(Clone, Debug)]
pub struct ProcessInfo {
pub parent: u32,
pub argv: Vec<String>,
pub cwd: String,
pub state: String,
}
pub type Processes = Arc<Mutex<std::collections::BTreeMap<u32, ProcessInfo>>>;
struct Entry {
command: u32,
group: u32,
cancelled: Arc<AtomicBool>,
abort: AbortHandle,
}
/// One runtime keeps compiled components warm and shares the world's namespace across stores.
pub struct Runtime {
world: World,
engine: Engine,
cache: Mutex<HashMap<Vec<u8>, CommandPre<State>>>,
entries: Mutex<HashMap<u32, Entry>>,
next: AtomicU32,
alive: Arc<AtomicBool>,
}
impl Drop for Runtime {
fn drop(&mut self) {
self.alive.store(false, Ordering::Relaxed);
}
}
struct State {
wasi: WasiCtx,
table: ResourceTable,
fs: Adapter,
runtime: Arc<Runtime>,
pid: u32,
command: u32,
group: u32,
}
impl WasiView for State {
fn ctx(&mut self) -> WasiCtxView<'_> {
WasiCtxView {
ctx: &mut self.wasi,
table: &mut self.table,
}
}
}
pub struct Child {
pid: u32,
group: u32,
code: watch::Receiver<Option<i32>>,
input: Option<DynOutputStream>,
output: Option<DynInputStream>,
error: Option<DynInputStream>,
}
struct Started {
pid: u32,
group: u32,
code: watch::Receiver<Option<i32>>,
input: DuplexStream,
output: DuplexStream,
error: DuplexStream,
}
impl Runtime {
pub fn open(path: &Path) -> anyhow::Result<Arc<Self>> {
Self::new(World::open(path)?)
}
pub fn new(world: World) -> anyhow::Result<Arc<Self>> {
let mut config = wasmtime::Config::new();
config.epoch_interruption(true);
let mut cache = wasmtime::CacheConfig::new();
cache.with_directory(world.host_dir.join("compile-cache"));
config.cache(Some(wasmtime::Cache::new(cache)?));
let engine = Engine::new(&config)?;
let alive = Arc::new(AtomicBool::new(true));
let ticking = alive.clone();
let clock = engine.clone();
std::thread::spawn(move || {
while ticking.load(Ordering::Relaxed) {
std::thread::sleep(std::time::Duration::from_millis(10));
clock.increment_epoch();
}
});
Ok(Arc::new(Self {
world,
engine,
alive,
cache: Mutex::new(HashMap::new()),
entries: Mutex::new(HashMap::new()),
next: AtomicU32::new(1),
}))
}
fn kill_command(&self, command: u32) {
for entry in self.entries.lock().unwrap().values() {
if entry.command == command {
entry.cancelled.store(true, Ordering::Relaxed);
entry.abort.abort();
}
}
}
fn kill_group(&self, group: u32) {
for entry in self.entries.lock().unwrap().values() {
if entry.group == group {
entry.abort.abort();
}
}
}
fn kill_pid(&self, pid: u32) -> anyhow::Result<()> {
let group = self
.entries
.lock()
.unwrap()
.get(&pid)
.map(|e| e.group)
.ok_or_else(|| anyhow::anyhow!("no live process {pid}"))?;
self.kill_group(group);
Ok(())
}
async fn start(
self: &Arc<Self>,
mut request: Request,
parent: u32,
command: u32,
inherited_group: u32,
) -> anyhow::Result<Started> {
anyhow::ensure!(!request.argv.is_empty(), "empty argv");
let cancelled = if parent == 0 {
Arc::new(AtomicBool::new(false))
} else {
self.entries
.lock()
.unwrap()
.get(&parent)
.map(|entry| entry.cancelled.clone())
.ok_or_else(|| anyhow::anyhow!("parent exited"))?
};
request.cwd = sausage_vfs::resolve("/", &request.cwd);
request.env.retain(|(name, _)| name != "PWD");
request.env.push(("PWD".into(), request.cwd.clone()));
let executable = sausage_vfs::resolve(&request.cwd, &request.program);
let file = self
.world
.root
.open(
&executable,
sausage_vfs::Open {
follow: true,
..Default::default()
},
)
.await?;
let bytes = file
.read(0, usize::try_from(file.stat().await?.size)?)
.await?;
let pre = {
let mut cache = self.cache.lock().unwrap();
if let Some(pre) = cache.get(&bytes) {
pre.clone()
} else {
let component = Component::new(&self.engine, &bytes)?;
let mut linker = Linker::new(&self.engine);
wasmtime_wasi::p2::add_to_linker_async(&mut linker)?;
// The library linker includes exit-with-code. Only Wasmtime's CLI gates it.
adapter::add_to_linker(&mut linker, |s: &mut State| View {
adapter: &mut s.fs,
table: &mut s.table,
})?;
bindings::sausage::process::host::add_to_linker::<
_,
wasmtime::component::HasSelf<_>,
>(&mut linker, |s| s)?;
let pre = CommandPre::new(linker.instantiate_pre(&component)?)?;
cache.insert(bytes, pre.clone());
pre
}
};
let pid = self.next.fetch_add(1, Ordering::Relaxed);
let command = if command == 0 { pid } else { command };
let group = match request.group {
Some(0) => pid,
Some(group) => group,
None if inherited_group != 0 => inherited_group,
None => pid,
};
request.env.push(("PID".into(), pid.to_string()));
request.env.push(("PPID".into(), parent.to_string()));
let (input, child_input) = tokio::io::duplex(8192);
let (output, child_output) = tokio::io::duplex(8192);
let (error, child_error) = tokio::io::duplex(8192);
let mut mounts = Mounts::default();
mounts.mount("/", self.world.root.clone(), false);
let mut wasi = WasiCtxBuilder::new();
wasi.args(&request.argv)
.envs(&request.env)
.initial_cwd(&request.cwd)
.stdin(AsyncStdinStream::new(child_input))
.stdout(AsyncStdoutStream::new(8192, child_output))
.stderr(AsyncStdoutStream::new(8192, child_error));
let mut store = Store::new(
&self.engine,
State {
wasi: wasi.build(),
table: ResourceTable::new(),
fs: Adapter::new(mounts),
runtime: self.clone(),
pid,
command,
group,
},
);
store.epoch_deadline_async_yield_and_update(1);
self.world.processes.lock().unwrap().insert(
pid,
ProcessInfo {
parent,
argv: request.argv,
cwd: request.cwd,
state: "running".into(),
},
);
let (tx, code) = watch::channel(None);
// Capture cleanup before the first poll, so an early abort still removes the entry.
let cleanup = Cleanup {
runtime: self.clone(),
pid,
tx,
};
// The start gate makes the entry visible before a guest can spawn children or exit.
let (gate, ready) = tokio::sync::oneshot::channel();
let task = tokio::spawn(async move {
let _cleanup = cleanup;
if ready.await.is_err() {
return;
}
let result = async {
let command = pre.instantiate_async(&mut store).await?;
match command.wasi_cli_run().call_run(&mut store).await {
Ok(Ok(())) => Ok(0),
Ok(Err(())) => Ok(1),
Err(e) => match e.downcast_ref::<wasmtime_wasi::I32Exit>() {
Some(exit) => Ok(exit.0),
None => Err(e),
},
}
}
.await;
_cleanup.tx.send_replace(Some(result.unwrap_or(125)));
});
{
let mut entries = self.entries.lock().unwrap();
entries.insert(
pid,
Entry {
command,
group,
cancelled: cancelled.clone(),
abort: task.abort_handle(),
},
);
// A spawn already in progress must observe a command-wide cancellation.
if cancelled.load(Ordering::Relaxed) {
task.abort();
} else {
let _ = gate.send(());
}
}
Ok(Started {
pid,
group,
code,
input,
output,
error,
})
}
}
struct Cleanup {
runtime: Arc<Runtime>,
pid: u32,
tx: watch::Sender<Option<i32>>,
}
impl Drop for Cleanup {
fn drop(&mut self) {
if self.tx.borrow().is_none() {
self.tx.send_replace(Some(137));
}
self.runtime.entries.lock().unwrap().remove(&self.pid);
self.runtime
.world
.processes
.lock()
.unwrap()
.remove(&self.pid);
}
}
impl Host for State {
async fn spawn(
&mut self,
request: Request,
) -> wasmtime::Result<Result<Resource<Child>, String>> {
let child = match self
.runtime
.start(request, self.pid, self.command, self.group)
.await
{
Ok(child) => child,
Err(e) => return Ok(Err(e.to_string())),
};
Ok(Ok(self.table.push(Child {
pid: child.pid,
group: child.group,
code: child.code,
input: Some(Box::new(AsyncWriteStream::new(8192, child.input))),
output: Some(Box::new(AsyncReadStream::new(child.output))),
error: Some(Box::new(AsyncReadStream::new(child.error))),
})?))
}
async fn kill(&mut self, pid: u32) -> wasmtime::Result<Result<(), String>> {
Ok(self.runtime.kill_pid(pid).map_err(|e| e.to_string()))
}
}
impl HostChild for State {
async fn pid(&mut self, child: Resource<Child>) -> wasmtime::Result<u32> {
Ok(self.table.get(&child)?.pid)
}
async fn group(&mut self, child: Resource<Child>) -> wasmtime::Result<u32> {
Ok(self.table.get(&child)?.group)
}
async fn stdin(
&mut self,
child: Resource<Child>,
) -> wasmtime::Result<Resource<DynOutputStream>> {
let stream = self
.table
.get_mut(&child)?
.input
.take()
.ok_or_else(|| wasmtime::format_err!("stdin already taken"))?;
Ok(self.table.push(stream)?)
}
async fn stdout(
&mut self,
child: Resource<Child>,
) -> wasmtime::Result<Resource<DynInputStream>> {
let stream = self
.table
.get_mut(&child)?
.output
.take()
.ok_or_else(|| wasmtime::format_err!("stdout already taken"))?;
Ok(self.table.push(stream)?)
}
async fn stderr(
&mut self,
child: Resource<Child>,
) -> wasmtime::Result<Resource<DynInputStream>> {
let stream = self
.table
.get_mut(&child)?
.error
.take()
.ok_or_else(|| wasmtime::format_err!("stderr already taken"))?;
Ok(self.table.push(stream)?)
}
async fn status(&mut self, child: Resource<Child>) -> wasmtime::Result<Option<i32>> {
Ok(*self.table.get(&child)?.code.borrow())
}
async fn wait(&mut self, child: Resource<Child>) -> wasmtime::Result<i32> {
let mut code = self.table.get(&child)?.code.clone();
loop {
if let Some(code) = *code.borrow_and_update() {
return Ok(code);
}
code.changed().await?;
}
}
async fn kill(&mut self, child: Resource<Child>) -> wasmtime::Result<()> {
self.runtime
.kill_pid(self.table.get(&child)?.pid)
.map_err(|e| wasmtime::format_err!("{e}"))
}
async fn drop(&mut self, child: Resource<Child>) -> wasmtime::Result<()> {
self.table.delete(child)?;
Ok(())
}
}
struct GroupGuard {
runtime: Arc<Runtime>,
group: u32,
}
impl Drop for GroupGuard {
fn drop(&mut self) {
self.runtime.kill_command(self.group);
}
}
/// An Arc runtime implements Exec so the harness can retain one warm process table.
#[async_trait::async_trait]
impl Exec for Arc<Runtime> {
async fn exec(
&self,
command: Command,
options: Options,
output: &mut (dyn for<'chunk> FnMut(Stream, &'chunk [u8]) + Send),
) -> anyhow::Result<i32> {
let deadline = options
.timeout
.map(|duration| tokio::time::Instant::now() + duration);
let argv = match command {
Command::Shell(script) => vec![
"/bin/sh".into(),
"--noprofile".into(),
"--norc".into(),
"-c".into(),
script,
],
Command::Argv(argv) => argv,
};
let mut env = options.env;
if !env.iter().any(|(k, _)| k == "PATH") {
env.push(("PATH".into(), "/bin".into()));
}
let start = self.start(
Request {
program: argv.first().cloned().unwrap_or_default(),
argv,
env,
cwd: options.cwd,
group: None,
},
0,
0,
0,
);
let mut child = if let Some(deadline) = deadline {
tokio::time::timeout_at(deadline, start)
.await
.map_err(|_| anyhow::anyhow!("command timed out"))??
} else {
start.await?
};
let _guard = GroupGuard {
runtime: self.clone(),
group: child.pid,
};
drop(child.input);
let run = async {
let (mut out_open, mut err_open) = (true, true);
let (mut out, mut err) = ([0; 8192], [0; 8192]);
while out_open || err_open {
tokio::select! {
n = child.output.read(&mut out), if out_open => { let n = n?; out_open = n != 0; if n != 0 { output(Stream::Stdout, &out[..n]); } }
n = child.error.read(&mut err), if err_open => { let n = n?; err_open = n != 0; if n != 0 { output(Stream::Stderr, &err[..n]); } }
}
}
loop {
if let Some(code) = *child.code.borrow_and_update() {
return Ok(code);
}
child.code.changed().await?;
}
};
if let Some(deadline) = deadline {
tokio::time::timeout_at(deadline, run)
.await
.map_err(|_| anyhow::anyhow!("command timed out"))?
} else {
run.await
}
}
}
/// Construct the harness's shared executor over the same open world.
pub fn exec(world: World) -> anyhow::Result<Arc<dyn Exec>> {
Ok(Arc::new(Runtime::new(world)?))
}
/// Run the world's shell. Noninteractive shell scripts stream directly to host stdio.
pub async fn sh(world: &Path, command: Option<String>) -> anyhow::Result<i32> {
let runtime = Runtime::open(world)?;
let command = command.ok_or_else(|| anyhow::anyhow!("interactive shell is not implemented"))?;
runtime
.exec(
Command::Shell(command),
Options {
cwd: "/".into(),
..Options::default()
},
&mut |stream, bytes| {
use std::io::Write;
match stream {
Stream::Stdout => {
let _ = std::io::stdout().write_all(bytes);
}
Stream::Stderr => {
let _ = std::io::stderr().write_all(bytes);
}
}
},
)
.await
}
+33 -5
View File
@@ -3,7 +3,7 @@
//! Layout of a world directory:
//! - `world/`: the virtual root `/` that every guest sees. Agents change everything in it.
//! - `host/`: host state outside the sandbox (harness storage, later the vault). No guest sees it.
use sausage_vfs::{Host, Layer, Mounts, Process};
use sausage_vfs::{Generated, Host, Layer, Memory, Mounts, Process};
use std::{
path::{Path, PathBuf},
sync::Arc,
@@ -15,18 +15,46 @@ pub struct World {
pub root: Arc<dyn Layer>,
/// The host state folder, `<dir>/host`.
pub host_dir: PathBuf,
/// Live process metadata shared with the process table.
pub processes: crate::process::Processes,
}
impl World {
/// Open (and create if missing) the world directory `dir`. The process table adds more mounts
/// (`/proc`, `/run`).
/// Open (and create if missing) the world directory `dir`, including `/proc` and `/run` mounts.
pub fn open(dir: &Path) -> anyhow::Result<Self> {
let root_dir = dir.join("world");
let host_dir = dir.join("host");
std::fs::create_dir_all(&root_dir)?;
std::fs::create_dir_all(&host_dir)?;
let mut mounts = Mounts::default();
mounts.mount("/", Arc::new(Host::new(&root_dir, Process::default())?), false);
Ok(World { root: Arc::new(mounts), host_dir })
mounts.mount(
"/",
Arc::new(Host::new(&root_dir, Process::default())?),
false,
);
let processes: crate::process::Processes =
Arc::new(std::sync::Mutex::new(Default::default()));
let snapshot = processes.clone();
let mut proc = Generated::default();
proc.insert("/processes", move || {
let mut text = String::from("PID\tPPID\tSTATE\tCWD\tARGV\n");
for (pid, info) in snapshot.lock().unwrap().iter() {
text.push_str(&format!(
"{pid}\t{}\t{}\t{}\t{}\n",
info.parent,
info.state,
info.cwd,
info.argv.join(" ")
));
}
Ok(text.into_bytes())
});
mounts.mount("/proc", Arc::new(proc), true);
mounts.mount("/run", Arc::new(Memory::default()), false);
Ok(World {
root: Arc::new(mounts),
host_dir,
processes,
})
}
}