banger/internal/daemon
Thales Maciel 3805b093b4
roothelper: tie kill/signal authorization to banger-launched firecracker
validateFirecrackerPID was a substring check on /proc/<pid>/cmdline:
"contains 'firecracker'". Good enough to refuse init/sshd/the test
binary, but on a shared host where multiple users run firecracker
the helper would happily SIGKILL someone else's VM. The owner-UID
daemon could weaponise the helper as an arbitrary "kill any
firecracker on this box" primitive.

Replace the substring gate with two stronger acceptance modes:

  * Cgroup match (the supported path): /proc/<pid>/cgroup contains
    bangerd-root.service. systemd assigns every direct child of the
    helper unit into that cgroup at fork; the kernel keeps it there
    for the process's lifetime, so no daemon-UID code can forge it.
    Other users' firecracker processes live in different cgroups
    (user@<uid>.service, foreign service slices) and fail this
    check. Also robust across helper restarts: KillMode=control-group
    on the unit kills children when the service goes down, so an
    "orphan banger firecracker in some other cgroup" is rare by
    construction.

  * --api-sock fallback: cmdline carries `--api-sock <path>` with
    the path under banger's RuntimeDir. Covers the legacy direct
    (no-jailer) launch path, and gives daemon reconcile a way to
    clean up the rare orphan that lands outside the service cgroup
    after a hard helper crash.

Tried /proc/<pid>/root first — pivot_root semantics make jailer'd
firecracker read its root as "/" from any namespace, so the symlink
is useless as a banger-managed fingerprint. Cgroup is the right
signal.

Also added a signal allowlist: priv.signal_process now rejects
anything outside {TERM, KILL, INT, HUP, QUIT, USR1, USR2, ABRT}
(case-insensitive, with or without SIG prefix). STOP/CONT, real-time
signals, and numeric forms are refused — the helper running as root
must not be a generic "send arbitrary signal to my pid" primitive.
priv.kill_process is unaffected (it always sends KILL).

Tests: validateSignalName covers allowlist + numeric/STOP/RTMIN
rejection; extractFirecrackerAPISock pins the three flag forms
(--api-sock VAL, --api-sock=VAL, -a VAL); pathIsUnder gets a small
table; existing TestValidateFirecrackerPID still rejects PID 0,
PID 1, and the test process itself. Doctor's non-system-mode test
gained a t.TempDir-backed install path so it stops being
environment-dependent on machines that happen to have
/etc/banger/install.toml.

Smoke at JOBS=4 still green — every banger-launched firecracker
sails through the cgroup match.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-28 16:00:41 -03:00
..
dmsnap test: cover imagemgr + dmsnap helpers 2026-04-28 15:13:49 -03:00
fcproc roothelper: tighten input validation across privileged RPCs 2026-04-28 14:39:41 -03:00
imagemgr test: cover imagemgr + dmsnap helpers 2026-04-28 15:13:49 -03:00
opstate coverage: medium batch — hostnat runner, store guest-sessions, daemon helpers 2026-04-18 18:03:37 -03:00
workspace seams: move the last four package globals onto instance fields 2026-04-22 12:07:14 -03:00
ARCHITECTURE.md daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
autopull_test.go daemon: build a work-seed during image pull, refresh doctor check 2026-04-23 20:24:10 -03:00
capabilities.go daemon: surface previously-swallowed errors at warn 2026-04-26 22:30:51 -03:00
capabilities_test.go daemon: doctor passes vm dns when banger itself owns the port 2026-04-26 18:57:27 -03:00
concurrency_test.go daemon: build a work-seed during image pull, refresh doctor check 2026-04-23 20:24:10 -03:00
daemon.go roothelper: tighten input validation across privileged RPCs 2026-04-28 14:39:41 -03:00
daemon_test.go roothelper: tighten input validation across privileged RPCs 2026-04-28 14:39:41 -03:00
daemon_testing_test.go test: add newTestDaemon harness + options 2026-04-22 17:45:43 -03:00
dispatch.go daemon: extract StatsService sibling; shrink VMService's surface 2026-04-23 15:46:59 -03:00
dispatch_test.go daemon: thread per-RPC op_id end-to-end 2026-04-26 22:13:44 -03:00
dns_routing.go daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
dns_routing_test.go seams: move the last four package globals onto instance fields 2026-04-22 12:07:14 -03:00
doc.go daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
doctor.go doctor: surface security-posture drift in banger doctor 2026-04-28 14:58:34 -03:00
doctor_test.go roothelper: tie kill/signal authorization to banger-launched firecracker 2026-04-28 16:00:41 -03:00
fake_firecracker_test.go remove vm session feature 2026-04-20 12:47:58 -03:00
fastpath_test.go daemon: build the work disk fresh instead of cloning the seed file 2026-04-26 20:42:10 -03:00
guest_ssh.go remove vm session feature 2026-04-20 12:47:58 -03:00
host_network.go daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
image_seed.go daemon: serialise concurrent image/kernel pulls + atomic-rename seed refresh 2026-04-27 17:24:11 -03:00
image_service.go daemon: tighten concurrency around pulls, cleanup, and handle persistence 2026-04-27 19:32:43 -03:00
images.go daemon: tighten concurrency around pulls, cleanup, and handle persistence 2026-04-27 19:32:43 -03:00
images_helpers_test.go coverage: medium batch — hostnat runner, store guest-sessions, daemon helpers 2026-04-18 18:03:37 -03:00
images_pull.go daemon: build a work-seed during image pull, refresh doctor check 2026-04-23 20:24:10 -03:00
images_pull_bundle_test.go daemon: build a work-seed during image pull, refresh doctor check 2026-04-23 20:24:10 -03:00
images_pull_test.go daemon: build a work-seed during image pull, refresh doctor check 2026-04-23 20:24:10 -03:00
kernels.go daemon: tighten concurrency around pulls, cleanup, and handle persistence 2026-04-27 19:32:43 -03:00
kernels_test.go daemon split (6/n): extract wireServices + drop lazy service getters 2026-04-21 15:55:28 -03:00
lifecycle_flow_test.go test: end-to-end VMService lifecycle flow harness 2026-04-22 17:55:04 -03:00
logger.go daemon: thread per-RPC op_id end-to-end 2026-04-26 22:13:44 -03:00
logger_test.go seams: move the last four package globals onto instance fields 2026-04-22 12:07:14 -03:00
nat.go daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
nat_capability_test.go daemon: persist tap device on VM.Runtime so NAT teardown survives handle-cache loss 2026-04-23 14:21:13 -03:00
nat_test.go vm state: split transient kernel/process handles off the durable schema 2026-04-19 14:18:13 -03:00
open_close_test.go daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
preflight.go daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
privileged_ops.go firecracker: adopt firecracker-jailer for VM launch (Phase B) 2026-04-28 14:38:07 -03:00
runtime_assets.go daemon split (4/5): extract *VMService service 2026-04-20 20:57:05 -03:00
snapshot.go daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
snapshot_test.go daemon split (6/n): extract wireServices + drop lazy service getters 2026-04-21 15:55:28 -03:00
ssh_client_config.go daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
ssh_client_config_test.go daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
sshd_config_test.go daemon: delete flattenNestedWorkHome and normaliseHomeDirPerms 2026-04-23 18:33:06 -03:00
stats_service.go daemon: thread per-RPC op_id end-to-end 2026-04-26 22:13:44 -03:00
stats_service_test.go daemon: extract StatsService sibling; shrink VMService's surface 2026-04-23 15:46:59 -03:00
tap_pool.go daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
vm.go firecracker: adopt firecracker-jailer for VM launch (Phase B) 2026-04-28 14:38:07 -03:00
vm_authsync.go daemon: split owner daemon from root helper 2026-04-26 12:43:17 -03:00
vm_create.go daemon: tighten concurrency around pulls, cleanup, and handle persistence 2026-04-27 19:32:43 -03:00
vm_create_ops.go daemon: thread per-RPC op_id end-to-end 2026-04-26 22:13:44 -03:00
vm_create_test.go model: validate VM names as DNS labels at CLI + daemon 2026-04-23 14:06:40 -03:00
vm_disk.go system: mkfs work disks with lazy_itable_init + lazy_journal_init 2026-04-26 21:32:57 -03:00
vm_handles.go daemon: tighten concurrency around pulls, cleanup, and handle persistence 2026-04-27 19:32:43 -03:00
vm_handles_test.go daemon: persist teardown fallbacks and reject unsafe import paths 2026-04-23 16:21:59 -03:00
vm_lifecycle.go daemon: sync guest over ssh before stop to preserve workspace writes 2026-04-27 15:41:32 -03:00
vm_lifecycle_steps.go firecracker: adopt firecracker-jailer for VM launch (Phase B) 2026-04-28 14:38:07 -03:00
vm_lifecycle_steps_test.go daemon: extract startVMLocked into step runner with per-step rollback 2026-04-23 15:34:34 -03:00
vm_locks.go Move subsystem state/locks off Daemon into owning types 2026-04-15 15:58:33 -03:00
vm_service.go daemon: thread per-RPC op_id end-to-end 2026-04-26 22:13:44 -03:00
vm_set.go daemon: thread per-RPC op_id end-to-end 2026-04-26 22:13:44 -03:00
vm_test.go roothelper: tighten input validation across privileged RPCs 2026-04-28 14:39:41 -03:00
workspace.go feat(vm): add vm exec command with workspace dirty detection 2026-04-26 23:53:45 -03:00
workspace_rejection_test.go tests: targeted coverage for doctor, workspace rejections, and nat capability 2026-04-22 12:58:12 -03:00
workspace_service.go daemon: thread per-RPC op_id end-to-end 2026-04-26 22:13:44 -03:00
workspace_test.go feat(vm): add vm exec command with workspace dirty detection 2026-04-26 23:53:45 -03:00