Move the supported systemd path to two services: an owner-user bangerd for orchestration and a narrow root helper for bridge/tap, NAT/resolver, dm/loop, and Firecracker ownership. This removes repeated sudo from daily vm and image flows without leaving the general daemon running as root. Add install metadata, system install/status/restart/uninstall commands, and a system-owned runtime layout. Keep user SSH/config material in the owner home, lock file_sync to the owner home, and move daemon known_hosts handling out of the old root-owned control path. Route privileged lifecycle steps through typed privilegedOps calls, harden the two systemd units, and rewrite smoke plus docs around the supported service model. Verified with make build, make test, make lint, and make smoke on the supported systemd host path.
190 lines
6.3 KiB
Go
190 lines
6.3 KiB
Go
package fcproc
|
|
|
|
import (
|
|
"context"
|
|
"errors"
|
|
"log/slog"
|
|
"os"
|
|
"path/filepath"
|
|
"testing"
|
|
"time"
|
|
)
|
|
|
|
// scriptedRunner is a minimal Runner that records every call and
|
|
// plays back a pre-scripted sequence of (name, args, out, err)
|
|
// steps. Failing to match or running past the script fails the
|
|
// test. Mirrors the pattern from internal/daemon/snapshot_test.go
|
|
// but lives here because fcproc is a leaf package — it can't import
|
|
// its parent's test helpers.
|
|
type scriptedRunner struct {
|
|
t *testing.T
|
|
runs []scriptedCall
|
|
sudos []scriptedCall
|
|
}
|
|
|
|
type scriptedCall struct {
|
|
matchName string // empty for RunSudo (sudo has no distinct name arg)
|
|
matchArgs []string // nil means "don't care"
|
|
out []byte
|
|
err error
|
|
}
|
|
|
|
func (r *scriptedRunner) Run(_ context.Context, name string, args ...string) ([]byte, error) {
|
|
r.t.Helper()
|
|
if len(r.runs) == 0 {
|
|
r.t.Fatalf("unexpected Run(%q, %v)", name, args)
|
|
}
|
|
step := r.runs[0]
|
|
r.runs = r.runs[1:]
|
|
if step.matchName != "" && step.matchName != name {
|
|
r.t.Fatalf("Run name = %q, want %q", name, step.matchName)
|
|
}
|
|
return step.out, step.err
|
|
}
|
|
|
|
func (r *scriptedRunner) RunSudo(_ context.Context, args ...string) ([]byte, error) {
|
|
r.t.Helper()
|
|
if len(r.sudos) == 0 {
|
|
r.t.Fatalf("unexpected RunSudo(%v)", args)
|
|
}
|
|
step := r.sudos[0]
|
|
r.sudos = r.sudos[1:]
|
|
return step.out, step.err
|
|
}
|
|
|
|
// TestWaitForPathReturnsDeadlineExceededWhenSocketNeverAppears pins
|
|
// the timeout branch of waitForPath. If this drifts, every callsite
|
|
// that wraps it (EnsureSocketAccess on the firecracker API +
|
|
// vsock sockets) loses its bounded wait.
|
|
func TestWaitForPathReturnsDeadlineExceededWhenSocketNeverAppears(t *testing.T) {
|
|
missing := filepath.Join(t.TempDir(), "never-created.sock")
|
|
start := time.Now()
|
|
err := waitForPath(context.Background(), missing, 150*time.Millisecond, "api socket")
|
|
elapsed := time.Since(start)
|
|
|
|
if !errors.Is(err, context.DeadlineExceeded) {
|
|
t.Fatalf("err = %v, want wrapped context.DeadlineExceeded", err)
|
|
}
|
|
if !contains(err.Error(), "api socket") {
|
|
t.Fatalf("err = %v, want label 'api socket' in message", err)
|
|
}
|
|
// Timeout should fire close to the configured budget, not zero
|
|
// (tight-loop regression) and not way over (missing select
|
|
// regression). The 100ms poll tick plus the initial stat makes
|
|
// the lower bound noisy; check we at least waited a tick.
|
|
if elapsed < 90*time.Millisecond {
|
|
t.Fatalf("returned after %s; waitForPath exited before its timeout budget", elapsed)
|
|
}
|
|
}
|
|
|
|
// TestWaitForPathReturnsOnceSocketAppears pins the happy path:
|
|
// when the file materialises mid-wait, the function returns nil
|
|
// without having to walk to its deadline.
|
|
func TestWaitForPathReturnsOnceSocketAppears(t *testing.T) {
|
|
socketPath := filepath.Join(t.TempDir(), "will-appear.sock")
|
|
go func() {
|
|
time.Sleep(50 * time.Millisecond)
|
|
_ = os.WriteFile(socketPath, []byte{}, 0o600)
|
|
}()
|
|
if err := waitForPath(context.Background(), socketPath, 2*time.Second, "api socket"); err != nil {
|
|
t.Fatalf("waitForPath: %v", err)
|
|
}
|
|
}
|
|
|
|
// TestWaitForPathRespectsContextCancellation pins the ctx.Done()
|
|
// branch — a canceled request must not be blocked by the poll
|
|
// interval.
|
|
func TestWaitForPathRespectsContextCancellation(t *testing.T) {
|
|
missing := filepath.Join(t.TempDir(), "never.sock")
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
go func() {
|
|
time.Sleep(30 * time.Millisecond)
|
|
cancel()
|
|
}()
|
|
err := waitForPath(ctx, missing, 5*time.Second, "api socket")
|
|
if !errors.Is(err, context.Canceled) {
|
|
t.Fatalf("err = %v, want context.Canceled when ctx is cancelled mid-wait", err)
|
|
}
|
|
}
|
|
|
|
// TestEnsureSocketAccessChmodFailureBubbles verifies the chmod step
|
|
// fails fast before any ownership handoff. Once chown runs, the
|
|
// bounded helper no longer owns the socket and can't tighten its mode
|
|
// without CAP_FOWNER, so the order matters.
|
|
func TestEnsureSocketAccessChmodFailureBubbles(t *testing.T) {
|
|
socketPath := filepath.Join(t.TempDir(), "present.sock")
|
|
if err := os.WriteFile(socketPath, []byte{}, 0o600); err != nil {
|
|
t.Fatalf("WriteFile: %v", err)
|
|
}
|
|
|
|
chmodErr := errors.New("sudo chmod failed")
|
|
runner := &scriptedRunner{
|
|
t: t,
|
|
sudos: []scriptedCall{{err: chmodErr}},
|
|
}
|
|
mgr := New(runner, Config{}, slog.Default())
|
|
|
|
err := mgr.EnsureSocketAccess(context.Background(), socketPath, "api socket")
|
|
if !errors.Is(err, chmodErr) {
|
|
t.Fatalf("err = %v, want chmod error", err)
|
|
}
|
|
// chown must not have been attempted.
|
|
if len(runner.sudos) != 0 {
|
|
t.Fatalf("chown was attempted after chmod failed: %d sudo calls left", len(runner.sudos))
|
|
}
|
|
}
|
|
|
|
// TestEnsureSocketAccessChownFailureBubbles verifies the ownership
|
|
// handoff still surfaces errors after chmod succeeds.
|
|
func TestEnsureSocketAccessChownFailureBubbles(t *testing.T) {
|
|
socketPath := filepath.Join(t.TempDir(), "present.sock")
|
|
if err := os.WriteFile(socketPath, []byte{}, 0o600); err != nil {
|
|
t.Fatalf("WriteFile: %v", err)
|
|
}
|
|
|
|
chownErr := errors.New("sudo chown failed")
|
|
runner := &scriptedRunner{
|
|
t: t,
|
|
sudos: []scriptedCall{
|
|
{}, // chmod succeeds
|
|
{err: chownErr}, // chown fails
|
|
},
|
|
}
|
|
mgr := New(runner, Config{}, slog.Default())
|
|
|
|
err := mgr.EnsureSocketAccess(context.Background(), socketPath, "api socket")
|
|
if !errors.Is(err, chownErr) {
|
|
t.Fatalf("err = %v, want chown error", err)
|
|
}
|
|
}
|
|
|
|
// TestEnsureSocketAccessTimesOutBeforeTouchingRunner pins the
|
|
// ordering contract: if waitForPath never sees the socket, the
|
|
// sudo commands must not run. Running chown/chmod against a
|
|
// non-existent path would just noise the logs.
|
|
func TestEnsureSocketAccessTimesOutBeforeTouchingRunner(t *testing.T) {
|
|
missing := filepath.Join(t.TempDir(), "never.sock")
|
|
runner := &scriptedRunner{t: t} // no scripted calls — any runner invocation fails the test
|
|
mgr := New(runner, Config{}, slog.Default())
|
|
|
|
// EnsureSocketAccess's waitForPath has a hardcoded 5s timeout,
|
|
// and we can't inject a shorter one without widening the API.
|
|
// Use a short context instead — cancellation short-circuits
|
|
// waitForPath via the ctx.Done() branch.
|
|
ctx, cancel := context.WithTimeout(context.Background(), 150*time.Millisecond)
|
|
defer cancel()
|
|
|
|
err := mgr.EnsureSocketAccess(ctx, missing, "api socket")
|
|
if err == nil {
|
|
t.Fatal("EnsureSocketAccess: want error when socket never appears")
|
|
}
|
|
}
|
|
|
|
func contains(s, sub string) bool {
|
|
for i := 0; i+len(sub) <= len(s); i++ {
|
|
if s[i:i+len(sub)] == sub {
|
|
return true
|
|
}
|
|
}
|
|
return false
|
|
}
|